Enzymatically Structured Emulsions in Simulated Gastrointestinal

Article pubs.acs.org/Langmuir

Enzymatically Structured Emulsions in Simulated Gastrointestinal Environment: Impact on Interfacial Proteolysis and Diffusion in Intestinal Mucus Adam Macierzanka,*,† Franziska Böttger,† Neil M. Rigby,† Martina Lille,‡ Kaisa Poutanen,‡ E. N. Clare Mills,†,§ and Alan R. Mackie† †

Institute of Food Research, Norwich Research Park, Colney Lane, Norwich NR4 7UA, United Kingdom VTT Technical Research Centre of Finland, P.O. Box 1000, FI-02044, Finland

‡

S Supporting Information *

ABSTRACT: Fundamental knowledge of physicochemical interactions in the gastrointestinal environment is required in order to support rational designing of protein-stabilized colloidal food and pharmaceutical delivery systems with controlled behavior. In this paper, we report on the colloidal behavior of emulsions stabilized with the milk protein sodium caseinate (Na-Cas), and exposed to conditions simulating the human upper gastrointestinal tract. In particular, we looked at how the kinetics of proteolysis was affected by adsorption to an oil−water interface in emulsion and whether the proteolysis and the emulsion stability could be manipulated by enzymatic structuring of the interface. After cross-linking with the enzyme transglutaminase, the protein was digested with use of an in vitro model of gastro-duodenal proteolysis in the presence or absence of physiologically relevant surfactants (phosphatidylcholine, PC; bile salts, BS). Significant differences were found between the rates of digestion of Na-Cas cross-linked in emulsion (adsorbed protein) and in solution. In emulsion, the digestion of a population of polypeptides of Mr ca. 50−100 kDa was significantly retarded through the gastric digestion. The persistent interfacial polypeptides maintained the original emulsion droplet size and prevented the system from phase separating. Rapid pepsinolysis of adsorbed, non-cross-linked Na-Cas and its displacement by PC led to emulsion destabilization. These results suggest that structuring of emulsions by enzymatic cross-linking of the interfacial protein may affect the phase behavior of emulsion in the stomach and the gastric digestion rate in vivo. Measurements of ζ-potential revealed that BS displaced the remaining protein from the oil droplets during the simulated duodenal phase of digestion. Diffusion of the postdigestion emulsion droplets through ex vivo porcine intestinal mucus was only significant in the presence of BS due to the high negative charge these biosurfactants imparted to the droplets. This implies that the electrostatic repulsion produced can prevent the droplets from being trapped by the mucus matrix and facilitate their transport across the small intestine mucosal barrier.

1. INTRODUCTION If we are to address the increase in diet related health disorders such as obesity, type-2 diabetes, cardiovascular disease, etc., and enhance effectiveness of drug delivery systems, then we must improve our fundamental understanding of physicochemical interactions between colloidal foods/pharmaceuticals and the human gastrointestinal tract (GIT). For example, the fight against obesity requires knowledge-based approaches to designing foods that can help to reduce calorie intake. There are a number of different approaches that can be used to invoke hormone-mediated effects resulting in inducing satiety, delaying the gastric empting, increasing the gastrointestinal transit time, and reducing calorie intake.1−8 For that reason, the production of colloidal foods with the ability to modulate the way that nutrients are released in the GIT has been advocated as a promising strategy for weight management and offers the possibility of sustainable reduction in food intake.9,10 A wide range of modern processed foods and pharmaceutical products contain colloidal dispersions of solid or liquid fats © 2012 American Chemical Society

stabilized by surface-active proteins. Therefore, engineering interfacial properties of proteins is becoming of considerable interest because of its potential for controlling both proteolysis and lipolysis of emulsions.11−13 Controlling the rates and locations of protein and lipid hydrolysis is important because of their potential to impact physiological responses. For example, the creaming and layering of fat from an acid-unstable emulsion in the stomach can lead to the aqueous phase being emptied into the small intestine first. This can then lead to a delay in detection of the energy-dense fat by the intestinal epithelium and can subsequently increase appetite scores after the meal as reported by several authors.14−17 Conversely, sedimentation of nutrients during the gastric digestion can impact gastric emptying. In vitro, such destabilization has recently been shown for full fat milk.18 There are also other mechanisms Received: October 26, 2011 Revised: October 22, 2012 Published: November 20, 2012 17349

dx.doi.org/10.1021/la302194q | Langmuir 2012, 28, 17349−17362

Langmuir

Article

Table 1. (a) Characteristics of the Initial Na-Cas Stabilized Emulsions (30% oil, w/w; pH 6.5) and (b) Their Concentrations in the Digestion Experimentsa (a) characteristics CNa‑Cas (mg/mL)

d4,3 (μm)

d3,2 (μm)

SSA (m2/mL)

ζ-potential (mV)

% adsorptionb

adsorbed Na-Cas coverage (mg/m2)

1.67 5.00

4.22 ± 0.44 3.83 ± 0.30

3.25 ± 0.29 3.02 ± 0.10

0.57 ± 0.05 0.60 ± 0.02

−26.3 ± 1.6 −28.9 ± 1.2

93.6 ± 1.4 45.1 ± 2.9

2.8 ± 0.2 3.5 ± 0.1

16.67

3.62 ± 0.25

2.90 ± 0.16

0.62 ± 0.03

−30.4 ± 0.9

17.0 ± 1.5

4.3 ± 0.2

(b) concnc CNa‑Cas 1 mg/mL, 18% oil (w/w) CNa‑Cas 3 mg/mL, 18% oil (w/w); or CNa‑Cas 1 mg/mL, 6% oil (w/w) CNa‑Cas 10 mg/mL, 18% oil (w/w)

a

CNa‑Cas, concentration of Na-Cas in emulsion; d4,3, volume mean droplet diameter; d3,2, surface area mean droplet diameter; SSA, specific surface area. bBased on the OPA method. cDiluted for digestion experiments to the concentrations given.

of adsorbed Na-Cas and emulsions produced with this protein. A change in the pattern of protein digestion may alter colloidal phase behavior of the emulsion in the stomach and the small intestine, which ultimately might influence both satiety and food intake. We have concentrated on proteolysis in the work presented. It is very likely that lipolysis of emulsified triglycerides must be preceded either by pepsinolysis of the adsorbed protein, in order to allow lipolytic enzymes to adsorb onto the oil−water interface, or by displacement of the protein by physiological surfactants.30−32 Therefore, the lipolysis and effect of its products on emulsion stability has been omitted in this work. The study aims to expand our knowledge about physicochemical and biophysical aspects of protein digestion before using this type of model emulsions in human studies. The work described in this paper has been organized in the order to monitor behavior of emulsions in simulated gastroduodenal conditions while increasing complexity of the digestion model. First, we looked at whether presenting NaCas in emulsion (adsorbed protein) and cross-linking with TG had any effect on the proteolysis as compared to the protein digested in solution and/or without structuring with TG. Next, we focused on the influence of physiological surfactants PC and BS on the cross-linked protein digestion pattern and the emulsion stability, and compared the results with the data obtained for the non-cross-linked systems. Finally, we discuss the impact of the enzymatic structuring and the digestion conditions on the diffusion of emulsion droplets through intestinal mucus. The study is schematically summarized in the Supporting Information, SI (Figure S1).

whereby satiety is increased by delivering nutrients, particularly fat, to the distal ileum and invoking the so-called “ileal brake”.10 Delivery of fat directly into the ileum was shown to stimulate peptide YY (PYY) release and suppress appetite in healthy volunteers.19 Thus a successful physicochemical approach to structuring interfaces in colloidal foods could have important implications in nutrition by modulating rates of digestion and reducing intake of calories through increasing satiety. The last boundary between ingested foods/pharmaceuticals and the GIT mucosa is the mucus barrier. This highly complex viscoelastic medium has evolved to provide a robust barrier that can trap and immobilize potentially hazardous particulates such as bacteria but still allow the passage of nutrients to the epithelial surfaces.20−22 These conflicting properties are particularly important in the small intestine where the mucus layer is thinnest and the majority of nutrients absorption takes place. However, the rules governing this selective barrier function, particularly in relation to transport of particulates, remain largely unknown. Recent studies suggested that surface properties of model nanoparticles can impact on colloidal transport in the intestinal mucus.23 However, there is no published information regarding detailed physicochemical characterization of the transport of postdigestion food/ pharmaceutical particulates under physiological conditions of the intestine. In our recent studies,24 we showed for the first time that interfacial adsorption of bile salts to the surface of model microparticles and lipid droplets significantly enhances their diffusion in the intestinal mucus. However, more work is needed to determine whether the diffusion of fat droplets in the mucus depends on other factors, such as interaction with other endogenous biosurfactants (e.g., phospholipids) in the GIT, or initial structuring of the surface of emulsion droplets. In the work described here, we produced triglyceride oil-inwater emulsions stabilized by sodium caseinate (Na-Cas), a milk protein extensively used in various food and pharmaceutical applications because of its excellent binding, emulsifying, and foaming properties.25,26 In our approach, the protein was modified after emulsification by enzymatic cross-linking with microbial transglutaminase (TG). This enzyme cross-links peptides and proteins through an acyl transfer mechanism between glutamine and lysine residues. TG is widely used in foods and pharmaceuticals to strengthen or otherwise modify protein networks and improve desirable technological, sensory, and other functional properties.27−29 By modifying the structure of the Na-Cas stabilized emulsions, we aimed to alter the pattern and/or the rate at which the protein was digested under in vitro gastro-duodenal conditions mimicking physicochemical conditions of the human GIT. We focused on gaining fundamental physicochemical knowledge on how the digestion process could be controlled by targeted modification

2. MATERIALS AND METHODS 2.1. Materials. Food-grade sodium caseinate (Na-Cas; 90% protein) was obtained from DMV International (The Netherlands). Microbial transglutaminase (TG) was treated to remove maltodextrin from the commercial product Activa WM (Vesantti Oy, Helsinki, Finland) by cation-exchange chromatography as described previously.33 The enzyme activity was assayed according to Folk,34 using 0.03 M N-carbobenzoxy-L-glutaminylglycine and hydroxylamine as substrates at pH 6.0. The measured activity of the TG stock solution was 11 000 nkat/mL (1 nkat is defined as the amount of enzyme activity that converts 1 nmol of the substrate per second in the assay conditions33). The solution was stored at −20 °C prior to use. Medium-chain triglyceride oil (MCT, Miglyol 812S) was obtained from Sasol GmbH, Germany. The fatty acid composition of the oil was 59.1% C8:0, 39.8% C10:0, 0.6% C12:0, and 0.5% comprising C6:0, C16:0, and C18:1. Prior to use, surface-active impurities were removed from the oil by stirring with Florisil (Sigma, F9127; 2:1 w/w, 140 rpm, 30 min). After filtering, the oil was stored under nitrogen. The buffers and reagents used in this work were prepared with ultrapure water (resistivity 18.2 MΩ cm). 2.2. Emulsion Preparation and Characterization. Oil-in-water (O/W) emulsions were prepared with 30% (w/w) MCT oil, 0.15 M 17350


Langmuir

Article

NaCl aqueous solution (pH 6.5), and Na-Cas (1.67, 5.00, or 16.67 mg/mL emulsion). The emulsions were produced by passing a premix of oil and Na-Cas solution through an EmulsiFlex-B3 valve homogenizer (Avestin Inc., Ottawa, Canada). Four to eight passes (depending on the protein concentration; high → low) were made at 13.8 MPa. The emulsions produced are referred to as initial emulsions (IEs) throughout the paper. Table 1 shows their basic characteristics. The proportion of protein adsorbed was determined after centrifuging the emulsion at 13 000 rpm (11 600 × g) for 10 min and assaying the amount of nonadsorbed protein in the clear subnatant formed below the emulsion cream layer. The subnatant was carefully removed with a syringe and needle and filtered, then the protein concentration was determined by the OPA method.35 Droplet size distribution was determined with use of a LS-230 particle sizer (Beckman Coulter Ltd., High Wycombe, UK). The measurements were made for at least eight replicate emulsions and the size characteristics evaluated after treating emulsions with 1 mM sodium dodecyl sulfate (SDS) to avoid possible flocculation of droplets in the samples being analyzed. The droplet size measurements of the initial emulsions that were not treated with SDS were also done, and the results of the size distributions and the mean size values matched exactly those done after the treatment with SDS. The measurement in the presence of SDS was only done in order to follow the same procedure for measuring the size of droplets as for the emulsions that were finally obtained after the digestion experiments and often showed flocculation and coalescence as a result of digestion as descried further in the text. In the latter case, the use of SDS allowed for analyzing the size of individual droplets in postdigestion samples. The concentration of adsorbed Na-Cas per unit surface area of emulsion (Na-Cas coverage) was calculated from the amount of the adsorbed protein in individual initial emulsions and their specific surface area (SSA) values. The ζ-potential was measured with a NanoZS Zetasizer (Malvern Instruments Ltd., Malvern, UK). Prior to analysis, emulsions were diluted about 400 times with 0.15 M NaCl (pH 6.5) in order to maintain the original salt concentration and also the concentration of salt in the digestion experiments. Diluted emulsions were then injected into a DTS1060 folded capillary cell (Malvern Instruments Ltd.). Each sample was analyzed at least 20 times and the results displayed as a mean. Data are shown as the average values obtained from at least three emulsions prepared under the same conditions. 2.3. Transglutaminase Cross-Linking of Sodium Caseinate in Emulsion and in Solution. The emulsions were incubated with TG at the TG:Na-Cas ratio of 100 nkat/g for 20 h under moderate stirring (50 rpm, 40 °C, pH 6.5). Control emulsion samples were also incubated in the absence of TG. After incubation and further dilution, the emulsions were used in digestion experiments at Na-Cas concentrations of 1, 3, or 10 mg/mL and oil content 18% or 6% (w/w), as described below. The cross-linking and control (i.e., without TG) incubations were also carried out for Na-Cas in solution. The protein was dissolved in 0.15 M NaCl aqueous solution (pH 6.5) at a concentration of 1.67 mg/mL, and the incubations were repeated as above. After dilution, the protein solutions were used in digestion experiments at an Na-Cas concentration of 1 mg/mL, as described below. All incubations were performed at least in triplicate for each condition. 2.4. In Vitro Gastro-Duodenal Digestion. The proteolysis experiments were based on the method described elsewhere.30 The detailed procedure used in this work has been given in the SI. Briefly, the experiments were done in triplicate for emulsions with cross-linked and non-cross-linked Na-Cas. The digestions were carried out in the presence or absence of physiological surfactants, vesicular phosphatidylcholine (PC; prepared as described in the SI), introduced in the gastric compartment, and bile salts (BS, 7.4 mM) in the subsequent duodenal compartment, in order to evaluate the impact these surfactants might have on the emulsion stability and the digestion pattern. The PC was introduced to the gastric compartment in the form of vesicular dispersion in order to mimic physiological conditions of secreting phospholipids by the gastric mucosa to the aqueous environment of digestion in the stomach. The emulsion obtained after 60 min of simulated gastric digestion, and subsequent increase of the

pH value to 6.5 (gastric emulsion; GE), was analyzed as described below or transferred directly to the simulated duodenal compartment of digestion. The emulsion obtained finally after 30 min of the simulated duodenal digestion is referred to as a gastro-duodenal emulsion (GDE) throughout this paper. The digestions were carried out at various protein concentrations, 1, 3, and 10 mg/mL, and two concentrations of the oil phase (18% and 6%, w/w), as summarized in Table 1. Digestion experiments were also undertaken for Na-Cas dissolved and cross-linked in aqueous solutions (Na-Cas, 1 mg/mL in digestion) to provide comparative data regarding the digestibility of the protein prepared under contrasting conditions (i.e., in the absence of oil). Samples of Na-Cas in solution were treated during the digestion in the same way as the emulsion samples. During all experiments time-point samples were collected and analyzed as described below. 2.5. Characterization of Digestion Samples. Progress of proteolysis was evaluated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE), including densitometric analysis, and size-exclusion HPLC of the collected time-point samples.30,36 The procedures have also been described in detail in the SI. Droplet size and ζ-potential of GEs and GDEs were analyzed in the same way as described earlier for the IEs. If the protein digestion led to separation of the oil phase, droplet size and ζ-potential were measured for the “subphase” emulsion, i.e. the emulsion remaining below the separated oil layer. In ζ-potential characterization, the solution used for diluting GDEs obtained after digestion done in the presence of BS was additionally made up with these surfactants (7.4 mM) to reflect the BS composition and concentration used in the digestion experiments. The microstructure of emulsions was investigated by using an Olympus BX-60 microscope (Olympus, Japan) operated in a differential interference contrast (DIC) mode. The images were acquired with a ProgRes C10plus digital camera (Jenoptik, Germany) connected to ProgRes CapturePro 2.1 software (Jenoptik, Germany). Diffusion of GDE droplets through intestinal mucus was analyzed by multiple droplet tracking using confocal microscopy as described in our previous report.24 The detailed procedure has also been given in the SI. 2.6. Pig Small Intestine Mucus: Collection and Preparation. Porcine small intestines from freshly slaughtered animals were obtained from a local abattoir. Immediately after slaughter, the segment of the GI tract containing the whole small and large intestines was removed. Only the jejunal part of the small intestine was used in further procedures, within ca. 30 min from the slaughter. The intestinal lumen was gently washed with 67 mM phosphate buffer (pH 6.7) containing 0.02% w/v sodium azide and a mix of protease inhibitors (Roche Diagnostics GmbH, Mannheim, Germany; 1 tablet per 50 mL buffer) in order to remove any remaining debris, prevent bacterial growth, and inhibit proteolytic enzymes. Mucus was gently removed from the epithelial surface of the intestine with a plastic scraper (Corning, NY). The scraped mucus was transferred to 0.5-mL plastic tubes, frozen in liquid nitrogen, and stored at −80 °C. Aliquots were incubated for 30 min at room temperature prior to use in experiments. Freezing did not significantly alter rheological properties of the mucus.24 2.7. Interfacial Tension Measurements. Evolution of the MCT oil−water interfacial tension (γ) was determined via drop shape analysis, using a FTA200 drop shape analyzer (FTA, Portsmouth, VA). All measurements were made at 37 °C. The measuring cell was filled with 4.2 mL of Na-Cas solution (1 or 10 mg/mL 0.15 M NaCl, pH 6.5) and left for 5 min to bring temperature up to 37 °C. Afterward, a 13-μL oil droplet was produced in the protein solution with use of a Jshaped needle and the protein was allowed to adsorb onto the oil− water interface for 30 min. Subsequently, the temperature was raised to 40 °C and a small aliquot of TG stock was added to give a TG:NaCas ratio of 100 nkat/g. The measuring cell was covered with 0.5 mL of oil to prevent evaporation of water, and the system incubated for 20 h. Incubations for control experiments without TG were also performed. Afterward, the temperature was brought back to 37 °C, the protective oil layer removed, and the cell very carefully washed with 0.15 M NaCl, pH 6.5 (40 mL, 37 °C) to remove any 17351


Langmuir

Article

nonadsorbed protein, and with 0.15 M NaCl, pH 2.5 (15 mL, 37 °C) to bring pH down to 2.5. The oil droplet with an adsorbed protein layer prepared in such a way was used for further analyses with pepsin or vesicular PC under simulated gastric conditions. In the experiments mimicking the gastric proteolysis of the adsorbed protein, a 5 mg/mL stock solution of pepsin (0.15 M NaCl, pH 6.5) was injected into the cell to give the same concentration of the enzyme as in the digestion of emulsion. Real-time measurements of γ were carried out for 60 min. For the experiments with PC, the vesicular dispersion of this lipid (pH 2.5) was prepared as described in the SI text. PC was introduced into the measuring cell in amounts required to give concentrations ranging from 0.01 to 2 mg/mL. The PC concentration was increased stepwise at 10-min intervals. The evolution of γ was monitored in real time, 10 min before introducing PC and after each injection of the lipid dispersion (37 °C, pH 2.5).

a mixture of higher molecular weight protein material (Figures 1 and 2c,d), which could not be dissociated by SDS and reduction. This might have been a contamination from the whey fraction or, more likely, casein aggregates (including S−S bridged κ-casein molecules). It is known that caseins in Na-Cas solutions exist in various states of aggregation and complexes, depending upon the protein concentration and other factors.37 When not cross-linked, this protein material was quickly digested together with the monomeric casein, and after 10 min there were only peptides smaller than pepsin present in the gastric digestion mix (Figures 1 and 2c,d). The effect of interfacial adsorption and cross-linking was also analyzed by SDS-PAGE. Figure 2 shows the proteolysis patterns for Na-Cas samples at a concentration of 1 mg/mL. As previously observed,36 TG-cross-linked Na-Cas oligomers were evident from the smear of high molecular weight protein unable to enter the polyacrylamide gel (Figure 2a, control track). Cross-linking was more extensive for Na-Cas in solution than for adsorbed protein in the emulsion sample (6% oil w/ w), as some monomeric Na-Cas remained after the incubation of emulsion with TG (Figure 2b, control track). This is in agreement with our recent studies on the kinetics of Na-Cas cross-linking.36 The SDS-PAGE confirmed the results observed by HPLC for the 3 mg/mL samples (Figure 1). Despite less extensive cross-linking of Na-Cas after adsorption to the emulsion droplets and relatively rapid loss of the very high molecular weight material after 5 min gastric digestion, the smear of protein Mr ca. 50−100 kDa remained largely unaltered between 20 and 60 min of digestion (Figure 2b). After 60 min of pepsinolysis, the SDS-PAGE revealed two distinctive smear regions in the protein pattern: a range of low molecular weight peptides (Mr roughly 50 kDa was found to be retarded through the pepsin digestion and still present in the system after exposure to the enzyme for 60 min (Figure 1a). This was not observed for the protein cross-linked in solution, for which only a range of peptides smaller than pepsin was seen after 20 min of proteolysis (Figure 1b). The original Na-Cas contained 17352


Langmuir

Article

Figure 2. SDS-PAGE analysis (reducing conditions) of simulated gastric followed by duodenal digestion of Na-Cas (1 mg/mL) in (a, c) solution and in (b, d) emulsion (6% oil, w/w). Impact of the transglutaminase cross-linking of Na-Cas: (a, b) cross-linked or (c, d) non-cross-linked before the digestion. The frame in part b highlights evolution of a population of oligomers of Mr ca. 50−100 kDa. The frames in parts c and d show degradation of non-cross-linked Na-Cas. The symbols indicate the bands corresponding to (▼) pepsin, (○) chymotrypsin, (●) trypsin, and (◊) Bowman−Birk trypsin/chymotrypsin inhibitor. All digestions were done in the absence of phosphatidylcholine.

Figure 3. SDS-PAGE analysis (reducing conditions) of simulated gastric digestion of Na-Cas cross-linked in emulsion (18% oil, w/w). Impact of the Na-Cas concentration (1, 3, 10 mg/mL). The frames in the SDS-PAGE gels indicate evolution of a population of oligomers of Mr ca. 50−100 kDa. All digestions were done at a pepsin:Na-Cas ratio of 1:20, w/w, and in the absence of phosphatidylcholine. Densitometric analysis of the protein patterns is shown in the SI (Figure S2).

simulated gastric digestion (Figure 3). The most apparent difference between the emulsions was in the rate the crosslinked protein was broken down by pepsin, with the slowest kinetics observed for the emulsion containing 10 mg/mL NaCas. In non-cross-linked emulsions, Na-Cas was completely broken down within 10−20 min of the gastric digestion (Supporting Information, Figure S3). The cross-linking of Na-Cas in emulsion and the resulting persistence of the population of high molecular weight oligomers during pepsinolysis had a profound effect on emulsion stability. None of the cross-linked emulsions showed a significant change in the droplet size distribution during the gastric digestion (Figure 4a−c). However, the negative charge of droplets in these emulsions was reduced from ca. −26.3 to −30.4 mV in the IEs (Table 1) to ca. −21.9 to −25.5 mV in the GEs (Table 2), depending on the protein concentration. In contrast, rapid and complete pepsinolysis of Na-Cas in

The digestibility of Na-Cas stabilizing 18%, w/w, dispersed phase emulsions was also studied by SDS-PAGE at a range of protein concentrations (Table 1). Adsorbed protein varied from only 17% in the emulsion containing 16.67 mg/mL protein (10 mg/mL for digestion) to more than 93% for the emulsion with the lowest Na-Cas concentration of 1.67 mg/mL (1 mg/mL for digestion). Nevertheless, the 10 mg/mL emulsion had the highest interfacial protein load (Na-Cas coverage) of 4.3 mg/m2 as compared to 2.8 mg/m2 for the emulsion with the least protein concentration (Table 1). The range of surface coverage found here is in good agreement with previous work showing that stable emulsions could be formed from caseins at a surface coverage of 1 mg/m2 and that a saturated monolayer was formed at ca. 3 mg/m2.38 The protective effect of cross-linking against the pepsinolysis was observed for all three emulsions. Again, a population of oligomers with Mr ca. 50−100 kDa was retarded through the 17353


Langmuir

Article

Figure 4. Impact of the transglutaminase (TG) cross-linking and the Na-Cas concentration (1, 3, 10 mg/mL) on the structural evolution of Na-Cas stabilized emulsions (18% oil, w/w) during the simulated gastro-duodenal digestion carried out in the absence of phosphatidylcholine. The initial emulsion (IEs) were incubated with/without the TG and subsequently subjected to the 60-min gastric phase of digestion (yielding the gastric emulsions, GEs), followed by the 30-min duodenal phase of digestion with bile salts (yielding the gastro-duodenal emulsions, GDEs). The size distributions were determined after treating emulsions with 1 mM SDS to avoid possible flocculation of droplets in the analyzed samples.

Table 2. ζ-Potential (mV) of Emulsions Obtained after Simulated Gastric and Duodenal Proteolysis Carried out in the Presence or Absence of Physiological Surfactants, PC and BS, and Effect of the Transglutaminase Crosslinking of Na-Cas in Initial Emulsions before the Digestion Experimentsa cross-linked Na-Cas CNa‑Cas (mg/mL)

PC

1 3 10 1 10

− − − + +

GE −21.9 −25.5 −25.5 −6.1 −23.0

± ± ± ± ±

1.7 2.7 3.1 1.2 2.2

non-cross-linked Na-Cas

GDE (+BS)

GDE (-BS)

−49.8 −52.3 −46.6 −44.1 −42.2

−17.0 ± 1.2 na na −5.9 ± 1.5b na

± ± ± ± ±

3.6 1.7 0.8 1.0 2.3

GE −17.6 −21.1 −23.3 −5.7 −18.9

± ± ± ± ±

3.8 2.1 3.5 1.8 0.5

GDE (+BS)

GDE (-BS)

−50.2 −52.5 −45.5 −46.5 −44.4

−15.8 ± 1.0 na na −5.6 ± 0.8b na

± ± ± ± ±

1.9 1.5 4.7 1.6 4.1

a

All measurements were done at pH 6.5. PC, phosphatidylcholine; GE, gastric emulsion; GDE, gastro-duodenal emulsion; BS, bile salts. na, not analyzed. bThe value obtained for the subphase emulsion located below the separated oil layer.

emulsions not cross-linked before digestion resulted in coalescence of the oil droplet and broadening of the droplet size distributions toward larger particle sizes (Figure 4d,e), although this was not the case for the emulsion containing 10 mg/mL Na-Cas (Figure 4f), which was also completely broken down by pepsin to low molecular weight peptides within 20 min of the gastric digestion (Figure S3c, SI). Evidently, the concentration and surface activity of the peptides produced were sufficient to maintain the original droplet size. The gastric digestion of non-cross-linked Na-Cas in emulsions produced a more significant drop in the negative charge of droplets (to ca. −17.6 to −23.3 mV, depending on the protein concentration) compared to the cross-linked protein (Table 2). The influence of pepsinolysis on oil−water interfacial tension (γ) was investigated with use of drop shape analysis (Figure 5). Upon incubation with pepsin, γ gradually increased as the enzyme degraded the adsorbed Na-Cas to peptides with lower surface activity than the parent protein. For both Na-Cas concentrations studied (1 and 10 mg/mL), the cross-linking was found to delay the increase in γ. However, the difference between cross-linked and non-cross-linked systems was more pronounced for the protein layer produced with 10 mg/mL NaCas, where a significant delay before γ started to rise (lag time of ca. 7 min) was only seen after strengthening the layer by cross-linking. In the system produced with 1 mg/mL Na-Cas, a 2-min lag time was followed by increase of γ to ca. 12.5 mN/m within 8 to 15 min, depending on whether the protein was

Figure 5. The time dependence of the Miglyol oil−water interfacial tension (γ) as identified from pendant drop shape analysis for the simulated gastric pepsinolysis of adsorbed Na-Cas (pH 2.5, 37 °C). Prior to the introduction of pepsin, the surface of the oil droplet was saturated with Na-Cas from either 1 or 10 mg/mL protein solutions, and subsequently the protein cross-linked with TG or incubated without TG at 40 °C (20 h, pH 6.5).

cross-linked with TG. This was followed by a slight decrease, implying a possible rearrangement in the protein material remaining at the interface. A similar effect was previously observed for pepsinolysis of adsorbed β-casein and βlactoglobulin.30 For both protein concentrations, the final γ observed after 60 min of the pepsinolysis was significantly 17354


Langmuir

Article

Figure 6. Impact of the transglutaminase (TG) cross-linking and Na-Cas concentration (1 or 10 mg/mL) on the structural evolution of Na-Cas stabilized emulsions (18% oil, w/w) during the simulated gastro-duodenal digestion carried out in the presence of phosphatidylcholine (PC). The initial emulsions (IEs) were incubated with/without the TG and subsequently subjected to the 60-min gastric digestion with PC (yielding the gastric emulsions, GEs), followed by 30-min duodenal digestion with bile salts (yielding the gastro-duodenal emulsions, GDEs). The micrographs show the structure of relevant GEs obtained for the Na-Cas concentration of 10 mg/mL (the scale bars correspond to 10 μm) immediately after the gastric digestion. The size distributions were determined after treating emulsions with 1 mM SDS to avoid possible flocculation of droplets in the analyzed samples.

Figure 7. SDS-PAGE analysis (reducing conditions) of simulated gastric followed by duodenal digestion of cross-linked Na-Cas in emulsion (18% oil, w/w), and in the presence of phosphatidylcholine (PC). Impact of the Na-Cas concentration: (a) 1 mg/mL or (b) 10 mg/mL. The black frame in part b highlights evolution of a population of oligomers of Mr ca. 50−100 kDa during the gastric digestion. The symbols indicate the bands corresponding to (□) PC, (▼) pepsin, (○) chymotrypsin, (●) trypsin, and (◊) Bowman−Birk trypsin/chymotrypsin inhibitor. All digestions were done at the same proteolytic enzymes:Na-Cas ratios. Densitometric analysis of the protein patterns is shown in the SI (Figure S5).

droplets (ζ-potential ca. −45.5 to −52.5 mV, depending on the initial protein concentration; Table 2). 3.2. Effect of Phosphatidylcholine on Digestion Progress and Emulsion Stability. Since phosphatidylcholine (PC) is present in the gastric fluid and mucosa39,40 is one of the major components of bile,41 and can modulate proteolysis and emulsion stability during simulated gastro-duodenal digestion,30,42 its influence on the proteolysis pattern and the emulsion behavior was also investigated. For the 1 mg/mL emulsion with cross-linked Na-Cas, inclusion of PC during digestion resulted in an increase in volume mean droplet diameter, d4,3, from 4.2 ± 0.4 μm in the IE to 50.9 ± 3.8 μm in the GE (Figure 6a). The destabilization was even more significant for the non-cross-linked emulsion (d4,3 = 81.0 ± 4.1 μm in the GE, Figure 6c). A large reduction in the negative charge of the GE droplets was also observed, which is likely to have been a result of displacement of the interfacial protein by zwitterionic PC. Thus the presence of PC resulted in the ζpotential increasing to ca. −6 mV, compared to −17.6 or −21.9 mV for the GEs in the absence of PC (Table 2; non-cross-

below the value recorded for the plain oil−water interface, suggesting that some of the peptides produced remained adsorbed at the interface. It was not the aim of this experiment to evaluate exact kinetics of similar changes in emulsions during digestion as the dynamics in evolution of γ seen for single pendant drop might be different from those to be expected in emulsions containing a large number of oil droplets, and hence very large interfacial area. The purpose of the interfacial measurements was to examine how the influence of pepsinolysis can be modified by TG cross-linking of the interfacial protein layer. Despite rapid degradation in emulsion of the remaining peptides by trypsin and chymotrypsin in the duodenal compartment of digestion (SI, Figure S4 +BS; data for 3 and 10 mg/mL Na-Cas not shown), the droplet size distribution of GDEs was preserved by the bile salts introduced to this compartment (Figure 4). The BS adsorbed to the droplets and might have partially or completely displaced the remaining peptides from the interface as indicated by the significant negative charge the surfactants provided to the emulsion 17355


Langmuir

Article

suspended in the protein solution and incubated with or without TG. The adsorption of Na-Cas caused reduction in γ from ca. 19 to 8.5 mN/m. After the incubation, nonabsorbed protein was removed and the pH reduced to 2.5. Washing of the measuring cell did not remove the adsorbed protein from the surface of the oil droplet as γ remained unchanged. After introduction of PC, γ was reduced as the more surface-active PC most likely displaced the protein from the interface (Figure 8). A similar effect was observed in our previous study for interfacial β-casein and β-lactoglobulin.30 The non-cross-linked Na-Cas layer produced at 1 mg/mL began to be displaced after increasing the concentration of PC to 0.1 mg/mL. Increasing the PC concentration to 0.5 mg/mL caused a decrease in γ to ca. 2 mN/m and detachment of the droplet. Although crosslinking of the protein made the interfacial layer resistant to 0.1 mg/mL PC, the increase to 0.5 mg/mL resulted in a rapid drop of γ. A much more significant protective effect of cross-linking was observed for the interfacial protein layer produced at 10 mg/mL Na-Cas, which was only displaced by PC after increasing the PC concentration to 1 mg/mL and only fully removed by 2 mg/mL PC. As in the pendant drop experiment on interfacial pepsinolysis (Figure 5), the single drop shape analysis data cannot be used directly to evaluate the kinetics of interactions in emulsions. However, these interfacial measurements with PC can serve as a good method for evaluating the effect of cross-linking of the adsorbed protein on its displacement by the surfactant. 3.3. Effect of Bile Salts on Digestion Progress and Emulsion Stability. After the gastric digestion, the emulsions were transferred to the duodenal phase of digestion including trypsin, chymotrypsin, and bile salts (BS). The remaining protein was further broken down by the enzymes to lower molecular weight peptides (Figure 7). The BS adsorbed to the oil droplets and imparted a significant negative charge (ca. −42 to −47 mV, depending on the protein concentration in emulsion; Table 2). The remaining interfacial peptides were very likely to have been displaced by the BS. The negative charge of the GDE droplets from experiments with PC was about 10% less than that for the GDEs obtained without PC (Table 2), suggesting that both PC and BS could form mixed films at the surface of emulsion droplets. The droplet size of GDEs did not increase compared to the GEs (Figure 6). In all the experiments described above, the BS were introduced to the simulated duodenal compartment of digestion to mimic physiological conditions. To investigate the role of the BS further, we also looked at how their absence would affect the properties of emulsion exposed to the duodenal conditions of digestion. Trypsin and chymotrypsin introduced to the digestion mix were found to hydrolyze any protein left in the emulsion after the gastric digestion, whether with or without PC, and after 30 min of duodenal digestion only low molecular weight peptides were detected by SDSPAGE as shown in the SI (Figures S4 and S6). The proteolysis pattern did not seem to be affected by the presence of the BS, implying that accessibility of the duodenal proteases to the protein oligomers/peptides was not significantly changed by interfacial adsorption or their displacement by PC and/or BS. However, the absence of BS had a profound effect on emulsion stability. Without the BS adsorbed onto the emulsion droplets, the destabilization of emulsion, already started during the gastric digestion due to pepsinolysis and/or displacement by PC, was continued in the duodenal compartment because of further degradation of any remaining interfacial protein by

linked or cross-linked emulsion, respectively). This probably accounts for the significant coalescence observed (Figure 6a,c). The presence of PC also altered the proteolysis pattern of the 1 mg/mL emulsion containing cross-linked Na-Cas (Figure 7a), with a lack of the persistent population of oligomers with Mr ca. 50−100 kDa observed above (Figure 3a). Thus, the pepsinolysis pattern shown in Figure 7a resembled the one seen for the pepsinolysis of Na-Cas in solution (Figure 2a), where no protein material larger than pepsin was left in the system after 20 min of gastric digestion. This confirms the interfacial protein displacement by PC during the simulated gastric digestion. Digestion of the 10 mg/mL non-cross-linked emulsion in the presence of PC resulted in droplet coalescence and a bimodal size distribution of the GE (Figure 6d). Displacement of adsorbed protein/peptides by PC led to a reduction in the negative charge of the droplets, the ζ-potential dropping to −18.9 mV (Table 2). However, digesting the 10 mg/mL crosslinked Na-Cas emulsion in the presence of PC, the droplet size and size distribution were not changed (Figure 6b), being equivalent to the 10 mg/mL emulsions digested in the absence of PC (Figure 4c,f). The ζ-potential of the GE was similar (−23.0 mV) to the counterpart GE obtained from the digestion without PC (−25.5 mV), Table 2. This suggests that crosslinking of the adsorbed protein significantly retarded displacement of interfacial protein by PC at this Na-Cas concentration, which is in marked contrast to the emulsions with 1 mg/mL Na-Cas. The improved resistance to displacement of the crosslinked interfacial protein at 10 mg/mL Na-Cas seems to be supported by the results obtained from SDS-PAGE (Figure 7b) where the proteolysis pattern matched that observed for the counterpart cross-linked emulsion digested in the absence of PC (Figure 3c); with a distinctive population of oligomers (Mr ca. 50−100 kDa) retarded through gastric digestion. Displacement of interfacial protein by vesicular PC under simulated gastric conditions was also studied with use of the pendant drop technique (Figure 8). A single oil droplet was

Figure 8. Pendant drop shape analysis of the kinetics of displacement of adsorbed Na-Cas by phosphatidylcholine (PC) at pH 2.5, 37 °C. Effect of the transglutaminase cross-linking of Na-Cas and the initial concentration of the protein on changes in the oil−water interfacial tension (γ). Concentration of PC was increased stepwise in the measuring cell (the steps are indicated by the arrows). Prior to the introduction of the PC vesicular dispersion, the surface of the oil droplet was saturated with Na-Cas from either 1 (white symbols) or 10 mg/mL (black symbols) protein solutions, and subsequently the protein cross-linked with TG (squares) or incubated without TG (circles) at 40 °C (20 h, pH 6.5). 17356


Langmuir

Article

Figure 9. Effect of the physiological surfactants (bile salts, BS; phosphatidylcholine, PC) on the size distributions of gastro-duodenal emulsion (GDE) dropets. The GDEs were obtained after simulated gastric digestion (60 min) carried out in the presence or absence of PC, followed by the duodenal digestion (30 min) carried out in the presence or absence of BS; Na-Cas cross-linked or non-cross-linked with transglutaminase in emulsion before digestion (Na-Cas 1 mg/mL; 18% oil, w/w). If the protein digestion led to the separation of the oil phase, droplet size distribution was measured for the “subphase” emulsion, i.e., the emulsion remained below the separated oil layer. The size distributions were determined after treating the emulsions with 1 mM SDS to avoid possible flocculation of droplets in the analyzed samples.

modifications did not change interfacial properties of the emulsion droplets. The diffusion of GDEs in the intestinal mucus was monitored by confocal microscopy, and from trajectories of many droplets the evolution of ensemble mean-square displacement ⟨MSD⟩ and ensemble diffusivity ⟨Deff⟩ over time were quantified (Figure 10). The ⟨MSD⟩ values for GDEs containing adsorbed BS increased by 2 orders of magnitude over 120 s (Figure 10a), although a broad range of MSD was observed from trajectories of individual droplets (Figure 10b). Figure 10c shows that in the presence of BS absorbed onto emulsion droplets the diffusion did not change with time, indicating that the droplets were not trapped by the mucus and expressed normal diffusion over the time scale of the experiment. Neither the cross-linking of Na-Cas before digestion nor the presence of PC had a significant effect on the diffusion of emulsion droplets (Figure 10c +BS). Undoubtedly, the most profound impact was the presence of BS and the high negative charge these surfactants provided to the emulsion droplets when adsorbed to their surface during the duodenal digestion (Table 2 and Figure S7b, SI). For the GDEs without BS, and thus with low negative charge, most droplets were immobilized at the surface of the mucus layer (Figure 11b). Therefore small amounts of those GDEs were gently mixed with the mucus before the diffusion experiments, as explained in the SI, in order to enable monitoring interaction of the droplets with the mucus within this matrix. Nevertheless, in the absence of BS coating the GDE droplets, the ⟨MSD⟩ was 2 orders of magnitude smaller after 120 s than with BS (Figure 10a,b). As above, the MSD data were converted to Deff, and finally the ⟨Deff⟩ values were calculated. Droplets of the GDEs lacking adsorbed BS showed subdiffusive behavior indicated by a monotonic decrease of ⟨Deff⟩ (Figure 10c −BS), revealing that the transport of the majority of the droplets was hindered in the mucus over the time scale monitored.

trypsin and chymotrypsin. This led to increased coalescence and eventually separation of the oil phase (Figure 9). The droplet size distribution of emulsions obtained finally after the duodenal digestion was dependent on the structure of GE transferred to the duodenal conditions (i.e., it was dependent on the extent of destabilization observed in GE after the gastric digestion). This was in turn determined by whether the Na-Cas was modified by TG cross-linking before digestion and the pepsinolysis carried out in the presence or absence of PC. In the absence of BS during the final duodenal phase of digestion, the reduced negative charge observed for GEs entering this compartment of digestion was maintained or even further diminished in the corresponding GDEs finally obtained (Table 2). As mentioned above, in the presence of BS, the high negative charge imparted to the emulsion droplets by these biosurfactants prevented the droplets from coalescing under the duodenal conditions. 3.4. Diffusion of Gastro-Duodenal Emulsion (GDE) Droplets in Intestinal Mucus. The four GDEs produced from the cross-linked IEs (Figure 9a,b) were used to study the effect of the interfacial properties of emulsions obtained after in vitro digestions under different conditions (i.e., ±PC, ±BS) on the diffusion of emulsion droplets in the ex vivo porcine intestinal mucus. Additionally, two GDEs obtained from the non-cross-linked IEs (Figure 9c) were used as control samples in the study in order to provide comparative data regarding the diffusion of emulsions, which were not enzymatically structured before the digestion. Since all the GDEs exhibited broad ranges of droplet size after digestion (Figure 9), the size distribution was gravitationally narrowed as explained in the SI text. The narrow size distributions and the similar mean droplet sizes obtained for the emulsions (Figure S7a, SI; d4,3 = 2.5−2.6 μm) minimized the effect of polydispersity of droplets on their transport in the mucus and emphasized the impact of the interfacial composition of droplets. The emulsions were fluorescently stained with Nile Red, and their ζ-potential measured before adding to the mucus (Figure S7b, SI). The results matched those obtained for the original GDEs analyzed just after digestion (Table 2), showing that the above

4. DISCUSSION In this work, we have studied the effect of enzymatic structuring of Na-Cas stabilized emulsion on the pattern of simulated gastro-duodenal digestion of the protein, the interactions with 17357


Langmuir

Article

Figure 10. Diffusion of gastro-duodenal emulsion (GDE) droplets in the ex vivo intestinal mucus: (i) impact of the physiological surfactants, bile salts, BS (left- vs right-hand side column), and phosphatidylcholine, PC (shown in individual graphs), adsorbed onto the surface of droplets during the digestion, and (ii) effect of the cross-linking of NaCas stabilizing the initial emulsions before the digestion. (a) Ensemble mean-square displacements (⟨MSD⟩) as a function of time (Δt), (b) distributions of MSD values obtained for individual droplets at Δt = 120 s, and (c) ensemble diffusivities (⟨Deff⟩) as a function of Δt (n = 3 with 80−110 droplets per experiment).

Figure 11. Confocal micrographs of postdigestion emulsion droplets (i.e., GDEs) interacting with intestinal ex vivo mucus in the (a) presence or (b) absence of bile salts (BS) adsorbed onto the surface of the droplets. Top-right dark areas in both images are diluted GDEs where the transport of the oil droplets toward the mucus phase took place from (within 15-min incubation at 37 °C). The GDEs were obtained from cross-linked initial emulsions digested in the absence of phosphatidylcholine and in the (a) presence or (b) absence of BS, as explained in Material and Methods, and in the SI. The scale bars correspond to 15 μm.

physiological surfactants, and the resulting microstructural transformations of the emulsion as it passes through different stages of in vitro digestion. Cross-linking of Na-Cas in solution was found to delay gastric proteolysis with pepsin (Figures 1 and 2). A similar effect was recently observed for purified β-Cas43 and phaseolin, a bean storage protein,44 and is in general thought to be caused by reduced accessibility of proteolytic enzymes to proteins by formation of inter- and/or intramolecular covalent bonds between substrate amino acids and involving a number of protein molecules in formation of cross-linked protein aggregates.45,46 Resistance to proteolytic degradation depends on the type of substrate protein, the enzyme used, the extent of cross-linking, and the conditions of proteolysis. TG cross-links proteins through the formation of covalent bonds between glutamine and lysine residues.47,48 Pepsin preferentially cleaves at Phe, Leu, Tyr, and Trp in position P1 or P1′.49,50 Recently, Hamuro et al.51 carried out statistical analysis of data from 39 proteins (13 766 amino acid residues) digested with porcine pepsin. They found that the probability of pepsin cleavage at these residues was between 17% and 46%, depending on the residue position (i.e., P1 or P1′). The data are in agreement with previous studies by Powers et al.,52 which showed a cleavage probability of 24−51% in position P1 and 20−34% in position P1′ for the four residues. Both groups of researchers have also shown that for Met in position P1 the probability was

ca. 40%. It is worth mentioning that, according to Hamuro et al.,51 the cleavage probability at four other amino acids Cys, Glu, Asp, and Ala was in the range of 13−24% in position P1, with up to 10−20% probability for other amino acid residues, depending on their position in the protein. Although undoubtedly pepsin is very efficient in cleaving various proteins, the above showed that it cannot be considered as an enzyme with very specific cleavage site preferences. In four main constituent caseins of Na-Cas (β-, αs1-, αs2- and κ-casein (3:4:1:1))53 there are many cleavage sites that have the potential to be affected by TG cross-linking.54 If we focused on the five potentially preferred by pepsin, as mentioned above (i.e., Phe, Leu, Met, Tyr, and Trp), we can see that the majority of them are located in close vicinity to the TG substrate amino acids, glutamine and lysine (i.e., up to 3 amino acids apart from them), and hence the access of pepsin to the cleavage sites may be hindered by the cross-linking. We have calculated the percentage of these potentially preferred cleavage sites in individual caseins that might be affected by the cross-linking in terms of their accessibility for pepsin. The amino acid sequences of caseins with the substrate TG residues and potential preferred pepsin cleavage sites have been shown in 17358


Langmuir

Article

strengthening of the interfacial film through intermolecular cross-links might have made some parts of the structured casein oligomers highly inaccessible to pepsin over the time scale of the gastric digestion and account for their increased stability. The kinetics of digestion of cross-linked Na-Cas was found to be dependent on the concentration of the protein (Figure 3), decreasing with an increase in the Na-Cas content. This might be an effect of the amount of protein involved in the interfacial layer subjected to cross-linking before digestion (Table 1). After increasing the total concentration of Na-Cas from 1 to 10 mg/mL, the amount of adsorbed protein increased ca. 1.5-fold. In the latter case, the interfacial protein concentration of 4.3 mg/m2 suggests that the protein might have been organized in the form of multilayers or aggregates at the interface as a monolayer has been shown to have a surface coverage of ca. 3 mg/m2.38,55 Nevertheless, only 17% of the total protein amount was adsorbed. The remaining 83% was cross-linked in solution (e.g., the aqueous phase of emulsion), although one cannot rule out the possibility that some protein molecules remaining in solution between emulsion droplets were covalently linked to the interfacial protein layer during the incubation with TG, and hence further increased the amount of protein involved in the interfacial layer. This may account for the substantial decrease in the overall kinetics of proteolysis in the emulsion containing 10 mg/mL cross-linked Na-Cas, where 20−40 min pepsinolysis was required to reduce the size of cross-linked oligomers, so they could enter the electrophoresis gel (Figure 3c). In contrast, only 5 min was required if Na-Cas was cross-linked in emulsion at 1 mg/mL protein concentration (Figure 3a). The findings seem to be supported by the pendant drop experiments (Figure 5), where the cross-linking of the protein layer produced with 10 mg/mL Na-Cas was shown to significantly delay the increase in γ during the proteolysis of adsorbed protein. The difference between cross-linked and non-cross-linked systems was also most pronounced for this protein concentration. However, the kinetics observed for a single drop can only be used as a supporting result and not as a means of assessing kinetics of similar changes in emulsions. Regardless of differences in the digestion kinetics, the existence of the resistant Mr 50−100 kDa population of oligomers prevented the cross-linked emulsions from destabilization during the digestion. A similar effect was observed previously for emulsions stabilized with bovine β-lactoglobulin,30 where 40% of the interfacial protein was stable to in vitro gastric digestion, which was sufficient to maintain the original emulsion droplet size. The interfacial β-casein, investigated in parallel in that study, was however very susceptible to pepsin digestion and the emulsion produced with that protein collapsed within a few minutes under simulated gastric conditions. Poor stability to the digestion conditions was also observed in the present study for emulsions containing 1 or 3 mg/mL Na-Cas that was not cross-linked before digestion (Figure 4d,e). However, after increasing the protein concentration to 10 mg/mL, the original droplet size distribution was maintained (Figure 4f) despite the fact that the protein was not structured by TG and hence completely degraded by pepsin to low molecular weight peptides with kinetics and digestion pattern similar to those observed for the less concentrated systems (Figure S3, SI). This suggests that the peptides produced in the 10 mg/mL emulsion were present in sufficiently high concentration to form an interfacial layer that was able to prevent oil droplets from coalescing. Nevertheless, the layer was not resistant to displacement by

Figure S8 (SI). For the most abundant of them (i.e., at Phe or Leu) the percentage of sites affected by the cross-linking in β-, αs1-, αs2-, and κ-casein might be 78%, 50%, 100%, and 50%, respectively, for Phe, and 68%, 71%, 77%, and 63%, respectively, for Leu. Similarly, for the cleavage sites at the three other amino acids, the calculations give 60%, 100%, 75%, and 0%, respectively, for Met, 75%, 40%, 75%, and 67%, respectively, for Tyr, and 100%, 100%, 0%, and 100%, respectively, for Trp. This strongly suggests that the access of pepsin to the preferred cleavage sites can be hindered after TG cross-linking of Na-Cas. The smearing seen in all cross-linked samples during the digestion might have been caused by nonspecific cleavage of the cross-linked protein oligomers by pepsin. So, the resulting product was a mix of postdigestion oligomers/polypeptides with a wide range of molecular weight, appearing as a smear on SDS-PAGE. In contrast, in all noncross-linked samples more specific cleavage of protein might have taken place as many well-resolved bands of the resulting peptides were observed. This might have been because the access of pepsin was not hindered by the cross-linking, and/or because the peptides produced after digestion were not linked by additional covalent bonds. In this study we have shown that TG cross-linking of a foodgrade milk protein can render it more resistant to simulated human gastric digestion. More importantly, we found that presenting the protein in an emulsion and subsequently structuring the interface by TG cross-linking increased the resistance of the protein to pepsin leading to significantly reduced proteolysis of high molecular weight cross-linked oligomers (Mr ca. 50−100 kDa). This effect was not observed for the digestion of cross-linked Na-Cas in solution (Figures 1 and 2), so the protection must have been limited to the protein adsorbed and cross-linked at the oil−water interface. Another explanation might be that although pepsin is still very efficient in cleaving interfacial protein into smaller peptides, they appear much larger in the cross-linked protein. So, the TG-structured protein does not resist pepsinolysis but might be held in a larger structure by the cross-links and adsorption may offer additional protection of pepsin cleavage sites. The source of the increased resistance is difficult to assess because of the complexity of Na-Cas composition. This is further complicated by the fact that the constituent caseins express different specificity to the TG cross-linking, which is also significantly affected by their adsorption to the oil−water interface.36 Nevertheless, the results shown here are surprising as we observed recently36 that the rate of cross-linking of Na-Cas adsorbed to the emulsion droplets was slower and the crosslinking extent smaller as compared to the protein incubated with TG in solution. Thus, it is clearly not only the extent of cross-linking that is crucial in protecting the protein from hydrolysis, but also the physical state of the substrate (i.e., protein in solution or adsorbed). In that study,36 we suggested that in an emulsion, the majority of protein can be immobilized in the interfacial film, and for such a case the cross-linking could be hindered as some substrate glutamine and lysine residues might be located close to the oil phase and not accessible for TG. We showed that incubation of Na-Cas with TG in solution led to some intramolecular cross-linking as the oligomers formed were found running faster on SDS-PAGE than their counterparts formed from Na-Cas cross-linked at the oil−water interface, indicating that intermolecular cross-linking might prevail at the interface. Immobilization of the protein molecules at the oil−water interface caused by adsorption, and the 17359


Langmuir

Article

the significance of adsorbed BS in facilitating diffusion of particles across the mucus layer. In those studies we also showed that BS were largely unable to change the overall surface charge of nonflagellated bacteria E. coli (ζ-potential ca. −18 mV in the presence or absence of BS), which resulted in hindered diffusion of bacteria in the mucus despite the presence of BS in the system. The work presented here shows that interaction of GDE with the mucus was not affected by the initial structuring of the protein stabilizing the emulsion or by the exposure of the emulsion droplets to PC during the digestion, but was controlled by adsorption of BS to the emulsion droplets, which facilitated their diffusion. As stated above, the BS used in the simulated digestion model at the physiological-relevant concentration were capable of displacing the interfacial protein/peptides and dominated electrostatic properties of the emulsion droplets passing through the duodenal conditions, regardless of whether the cross-linking of the interfacial protein rendered the protein partially resistant to pepsinolysis and displacing by PC in the gastric compartment of digestion. The significance of the increased negative charge of emulsion droplets on facilitating their diffusion in the intestinal mucus is supported by recent findings of Crater et al.23 The researchers monitored diffusion of model 200-nm spherical particles, with well-defined surface chemistry, in native porcine intestinal mucus. The transport rate of the particles was directly proportional to the electrostatic properties of the particles, increasing with an increase in their negative surface charge. The anionic particles were found diffusing 20−30 times faster than cationic particles.

the more surface-active PC introduced in the form of a vesicular dispersion to the digestion experiments to mimic phospholipids secreted by the gastric mucosa. As a result, a significant decrease in the negative charge of the emulsion droplets was observed, accompanied by flocculation and coalescence of droplets (Table 2, Figure 6d). In fact, even intact Na-Cas adsorbed to the oil droplet from the 10 mg/mL protein stock solution was found to be quickly displaced by PC after increasing the concentration of PC to 0.5 mg/mL as found from the pendant drop experiments (Figure 8). Only crosslinking of the interfacial layer produced at 10 mg/mL could make it more resistant to PC competing for the interface. The pendant drop experiment cannot reflect the kinetics of displacement in emulsion because the interfacial area occupied by adsorbed protein in emulsion and exposed to PC is far larger than in the single pendent drop analyzed. However, the interfacial measurements can serve as a good method for evaluating the effect of cross-linking of the protein on its resistance to displacement by PC under simulated gastric conditions. Indeed, the increase in protein concentration in emulsion to 10 mg/mL and its cross-linking was found to markedly improve stability of emulsion to digestion in the presence of PC (Figure 6b). Both the delay in displacement by PC and the increased resistance to pepsin may account for the observed stability of the emulsion containing 10 mg/mL crosslinked Na-Cas. The data shown in Table 2 highlight the effect of BS. Regardless of the emulsion preparation conditions (i.e., Na-Cas concentration, ±TG) and the gastric digestion conditions (i.e., ±PC), the duodenal BS provided extensive negative charge to the surface of the emulsion droplets and could partially or completely displace the remaining interfacial peptides. Thus, the duodenal digestion of the remaining protein material was likely to take place largely in the aqueous phase of emulsion. In our previous studies,30 BS have been shown to be very effective in displacing both the interfacial casein peptides and the undigested casein from the oil−water interface. By imparting sufficiently high negative charge to the droplets (ζ-potential, −42 to −52 mV), the BS stopped flocculation and prevented emulsions from further destabilizing over the duodenal digestion (Figure 9). Thus the BS exerted a similar effect here to that observed for other protein-stabilized emulsions, summarized in recent review papers.13,56 The high negative charge imparted to the droplets by adsorbed BS also changed the interactions between the droplets and the small intestine mucus. The increase in the electrostatic repulsion between the droplets and the mucus might be a key factor affecting the mobility of GDE droplets in the mucus. With BS adsorbed, the droplets exhibited normal diffusion whereas hindered diffusion and trapping of droplets by the mucus was observed for the weakly negatively charged GDE droplets obtained in the absence of BS (ζ-potential ranging from ca. −5.6 to −17 mV; Table 2). We have previously shown24 that the charge of mucus aggregates was not significantly affected by the BS (ζ-potential of −10.0 ± 1.4 or −10.9 ± 1.1 mV in the absence or presence of BS, respectively). This emphasizes a dominant role of the interfacial composition of emulsion droplets in their electrostatic interactions with the intestinal mucus. The tracking experiments on the emulsion droplets presented here were carried out in parallel to the experiments on the diffusion of emulsion produced with non-cross-linked Na-Cas as well as the diffusion of latex beads in the mucus.24 Those studies also highlighted

5. CONCLUSIONS The results presented here suggest that the TG cross-linking could be a promising strategy for modifying the behavior of protein-stabilized emulsions in the human GIT. The crosslinking of interfacial protein was shown to delay proteolysis and prevent the emulsion from destabilizing under simulated gastric conditions. This may have an influence on the phase behavior of emulsion in the stomach in vivo. Moreover, it may affect sensing of fat in the GIT and reduce digestion by gastric lipase. After passing from the gastric to the duodenal conditions, the interfacial properties of emulsion droplets were almost exclusively governed by the BS. The interfacial BS had a dominant role in determining diffusion of emulsion droplets through the small intestine mucus. All the above results suggest that the changes in protein accessibility and structure induced by adsorption to the oil/water interface and TG cross-linking make the protein and potentially the lipid less susceptible to hydrolysis. Despite this, it is clear that even the thickest crosslinked layer studied was susceptible to proteolysis and displacing by BS under simulated duodenal conditions. Thus, although it may be possible to modulate the gastric phase behavior of emulsion by protein cross-linking, the structuring of the interfacial layer is not likely to affect the transport of fat droplets through duodenal mucus toward the epithelium where detection of the nutrients can trigger secretion of appetite suppressing hormones. More work is needed in order to confirm this to be the case. The data presented here were derived from an in vitro model and may not reflect the in vivo conditions. Therefore, further in vitro and in vivo studies are necessary. However, in vivo studies that have been undertaken recently show that the Na-Cas stabilized emulsion cross-linked with TG can alter the postprandial metabolism and appetite responses in healthy young individuals.57 17360


Langmuir

■

Article

(12) McClements, D.; Decker, E.; Park, Y. Controlling lipid bioavailability through physicochemical and structural approaches. Crit. Rev. Food Sci. Nutr. 2009, 49, 48−67. (13) Mackie, A. R.; Macierzanka, A. Colloidal aspects of protein digestion. Curr. Opin. Colloid Interface Sci. 2010, 15, 102−108. (14) Marciani, L.; Faulks, R.; Wickham, M. S. J.; Bush, D.; Pick, B.; Wright, J.; Cox, E. F.; Fillery-Travis, A.; Gowland, P. A.; Spiller, R. C. Effect of intragastric acid stability of fat emulsions on gastric emptying, plasma lipid profile and postprandial satiety. Br. J. Nutr. 2009, 101, 919−928. (15) Marciani, L.; Wickham, M.; Singh, G.; Bush, D.; Pick, B.; Cox, E.; Fillery-Travis, A. J.; Faulks, R.; Marsden, C.; Gowland, P. A.; Spiller, R. C. Enhancement of intragastric acid stability of a fat emulsion meal delays gastric emptying and increases cholecystokinin release and gallbladder contraction. Am. J. Physiol.: Gastrointest. Liver Physiol. 2007, 292, G1607−G1613. (16) Marciani, L.; Wickham, M. S.; Bush, D.; Faulks, R.; Wright, J.; Fillery-Travis, A. J.; Spiller, R. C.; Gowland, P. A. Magnetic resonance imaging of the behaviour of oil-in-water emulsions in the gastric lumen of man. Br. J. Nutr. 2006, 95, 331−339. (17) Foltz, M.; Maljaars, J.; Schuring, E. A.; van der Wal, R. J.; Boer, T.; Duchateau, G. S.; Peters, H. P.; Stellaard, F.; Masclee, A. A. Intragastric layering of lipids delays lipid absorption and increases plasma CCK but has minor effects on gastric emptying and appetite. Am. J. Physiol.: Gastrointest. Liver Physiol. 2009, 296, G982−G991. (18) van Aken, G. A.; Bomhof, E.; Zoet, F. D.; Verbeek, M.; Oosterveld, A. Differences in in vitro gastric behaviour between homogenized milk and emulsions stabilised by Tween 80, whey protein, or whey protein and caseinate. Food Hydrocolloids 2011, 25, 781−788. (19) Maljaars, J.; Romeyn, E. A.; Haddeman, E.; Peters, H. P. F.; Masclee, A. A. M. Effect of fat saturation on satiety, hormone release, and food intake. Am. J. Clin. Nutr. 2009, 89, 1019−1024. (20) Cone, R. A. Barrier properties of mucus. Adv. Drug Delivery Rev. 2009, 61, 75−85. (21) Linden, S. K.; Florin, T. H. J.; McGuckin, M. A. Mucin dynamics in intestinal bacterial infection. PLoS One 2008, 3, e3952. (22) Linden, S. K.; Sutton, P.; Karlsson, N. G.; Korolik, V.; McGuckin, M. A. Mucins in the mucosal barrier to infection. Mucosal Immunol. 2008, 1, 183−197. (23) Crater, J. S.; Carrier, R. L. Barrier properties of gastrointestinal mucus to nanoparticle transport. Macromol. Biosci. 2010, 10, 1473−83. (24) Macierzanka, A.; Rigby, N. M.; Corfield, A. P.; Wellner, N.; Böttger, F.; Mills, E. N. C.; Mackie, A. R. Adsorption of bile salts to particles allows penetration of intestinal mucus. Soft Matter 2011, 7, 8077−8084. (25) Dickinson, E. Structure formation in casein-based gels, foams, and emulsions. Colloids Surf., A 2006, 288, 3−11. (26) Carrera Sánchez, C.; Rodríguez Patino, J. M. Interfacial, foaming and emulsifying characteristics of sodium caseinate as influenced by protein concentration in solution. Food Hydrocolloids 2005, 19, 407− 416. (27) Dickinson, E. Enzymatic crosslinking as a tool for food colloid rheology control and interfacial stabilisation. Trends Food Sci. Technol. 1997, 8, 334−339. (28) Motoki, M.; Seguro, K. Transglutaminase and its use for food processing. Trends Food Sci. Technol. 1998, 9, 204−210. (29) Fontana, A.; Spolaore, B.; Mero, B.; Veronese, F. M. Site-specific modification and PEGylation of pharmaceutical proteins mediated by transglutaminase. Adv. Drug Delivery Rev. 2008, 60, 13−28. (30) Macierzanka, A.; Sancho, A. I.; Mills, E. N. C.; Rigby, N. M.; Mackie, A. R. Emulsification alters simulated gastrointestinal proteolysis of β-casein and β-lactoglobulin. Soft Matter 2009, 5, 538−550. (31) Sarkar, A.; Goh, K. K. T.; Singh, R. P.; Singh, H. Behaviour of an oil-in-water emulsion stabilized by β-lactoglobulin in an in vitro gastric model. Food Hydrocolloids. 2009, 23, 1563−1569. (32) Sarkar, A.; Goh, K. K. T.; Singh, H. Properties of oil-in-water emulsions stabilized by β-lactoglobulin in simulated gastric fluid as

ASSOCIATED CONTENT

S Supporting Information *

Expanded description of the methods, and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*Tel: +44 1603 255242. Fax: +44 1603 507723. E-mail: Adam. [email protected] and [email protected]. Present Address §

School of Translational Medicine, Manchester Academic Health Science Centre, Manchester Institute of Biotechnology, University of Manchester, 131 Princess Street, Manchester M1 7DN, UK. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work at IFR was funded by the BBSRC through an Institute Strategic Programme Grant (research grant BB/ J004545/1), and the work in Finland was funded by the Academy of Finland through the TEPESS project. The authors are participants of the EU funded COST action INFOGEST (COST FA 1005).

■

REFERENCES

(1) Jones, K. L.; Doran, S. M.; Hveem, K.; Bartholomeusz, F. D.; Morley, J. E.; Sun, W. M.; Chatterton, B. E.; Horowitz, M. Relation between postprandial satiation and antral area in normal subjects. Am. J. Clin. Nutr. 1997, 66, 127−132. (2) Geliebter, A. Gastric distension and gastric capacity in relation to food intake in humans. Physiol. Behav. 1988, 44, 665−668. (3) Goetze, O.; Steingoetter, A.; Menne, D.; van der Voort, I. R.; Kwiatek, M. A.; Boesiger, P.; Weishaupt, D.; Thumshirn, M.; Fried, M.; Schwizer, W. The effect of macronutrients on gastric volume responses and gastric emptying in humans: A magnetic resonance imaging study. Am. J. Physiol.: Gastrointest. Liver Physiol. 2007, 292, G11−G17. (4) Holgate, A. M.; Read, N. W. Effect of ileal infusion of intralipid on gastrointestinal transit, ileal flow rate, and carbohydrate absorption in humans after ingestion of a liquid meal. Gastroenterology 1985, 88, 1005−1011. (5) Layer, P.; Holst, J. J.; Grandt, D.; Goebell, H. Ileal release of glucagon-like peptide-1 (GLP-1). Association with inhibition of gastric acid secretion in humans. Dig. Dis. Sci. 1995, 40, 1074−1082. (6) Welch, I.; Saunders, K.; Read, N. W. Effect of ileal and intravenous infusions of fat emulsions on feeding and satiety in human volunteers. Gastroenterology 1985, 89, 1293−1297. (7) Welch, I. M.; Sepple, C. P.; Read, N. W. Comparisons of the effects on satiety and eating behaviour of infusion of lipid into the different regions of the small intestine. Gut 1988, 29, 306−311. (8) Meyer, J. H.; Hlinka, M.; Tabrizi, Y.; DiMaso, N.; Raybould, H. E. Chemical specificities and intestinal distributions of nutrient-driven satiety. Am. J. Physiol. 1998, 275, R1293−R1307. (9) van Aken, G. A. Relating food emulsion structure and composition to the way it is processed in the gastrointestinal tract and physiological responses: What are the opportunities? Food Biophys. 2010, 5, 258−283. (10) Maljaars, P. W. J.; Peters, H. P. F.; Mela, D. J.; Masclee, A. A. M. Ileal brake: A sensible food target for appetite control. A review. Physiol. Behav. 2008, 95, 271−281. (11) Singh, H. A.; Ye, Q.; Horne, D. Structuring food emulsions in the gastrointestinal tract to modify lipid digestion. Prog. Lipid Res. 2009, 48, 92−100. 17361


Langmuir

Article

influenced by ionic strength and presence of mucin. Food Hydrocolloids 2010, 24, 534−541. (33) Lantto, R.; Puolanne, E.; Kalkkinen, N.; Buchert, J.; Autio, K. Enzyme-aided modification of chicken-breast myofibril proteins: effect of laccase and transglutaminase on gelation and thermal stability. J. Agric. Food Chem. 2005, 53, 9231−9237. (34) Folk, J. E. In Metabolism of Amino Acids and Amines; Tabor, H., Tabor, C. W., Eds.; Academic Press: New York, 1970; pp 889−894. (35) Church, F. C.; Swaisgood, H. E.; Porter, D. H.; Catignani, G. L. Spectrophotometric assay using o-phthaldialdehyde for determination of proteolysis in milk and isolated milk proteins. J. Dairy Sci. 1983, 66, 1219−1227. (36) Macierzanka, A.; Bordron, F.; Rigby, N. M.; Mills, E. N. C.; Lille, M.; Poutanen, K.; Mackie, A. R. Transglutaminase cross-linking kinetics of sodium caseinate is changed after emulsification. Food Hydrocolloids 2011, 25, 843−850. (37) Lucey, J. A.; Srinivasan, M.; Singh, H.; Munro, P. A. Characterization of commercial and experimental sodium caseinates by multiangle laser light scattering and size-exclusion chromatography. J. Agric. Food Chem. 2000, 48, 1610−1616. (38) Fang, Y.; Dalgleish, D. G. Dimensions of the adsorbed layers in oil-in-water emulsions stabilised by caseins. J. Colloid Interface Sci. 1993, 156, 329−334. (39) Lichtenberger, L. M.; Graziani, L. A.; Dial, E. J.; Butler, B. D.; Hills, B. A. Role of surface-active phospholipids in gastric cytoprotection. Science 1983, 219, 1327−1329. (40) Hills, B. A.; Butler, B. D.; Lichtenberger, L. M. Gastric mucosal barrier: hydrophobic lining to the lumen of the stomach. Am. J. Physiol. 1983, 244, G561−G568. (41) Antsaklis, G.; Lewin, M. R.; Sutor, D. J.; Cowie, A. G.; Clark, C. G. Gallbladder function, cholesterol stones, and bile composition. Gut 1975, 16, 937−942. (42) Mandalari, G.; Mackie, A. R.; Rigby, N. M.; Wickham, M. S.; Mills, E. N. C. Physiological phosphatidylcholine protects bovine βlactoglobulin from simulated gastrointestinal proteolysis. Mol. Nutr. Food Res. 2009, 53, S131−S139. (43) Monogioudi, E.; Faccio, G.; Lille, M.; Poutanen, K.; Buchert, J.; Mattinen, M.-L. Effect of enzymatic cross-linking of β-casein on proteolysis by pepsin. Food Hydrocolloids 2011, 25, 71−81. (44) Mariniello, L.; Giosafatto, C. V. L.; Di Pierro, P.; Sorrentino, A.; Porta, R. Synthesis and resistance to in vitro proteolysis of transglutaminase cross-linked phaseolin, the major storage protein from Phaseolus vulgaris. J. Agric. Food Chem. 2007, 55, 4717−4721. (45) Seguro, K.; Kumazawa, Y.; Ohtsuka, T.; Ide, H.; Nio, N.; Motoki, M. Kubota, ε-(γ-Glutamyl)lysine: hydrolysis by γ-glutamyltransferase of different origins, when free or protein bound. J. Agric. Food Chem. 1995, 43, 1977−1981. (46) Yasumoto, K.; Suzuki, F. Aspartyl- and glutamyl-lysine crosslinks formation and their nutritional availability. J. Nutr. Sci. Vitaminol. 1990, 36, S71−S77. (47) Griffin, M.; Casadio, R.; Bergamini, C. Transglutaminases: nature’s biological glues. Biochem. J. 2002, 368, 377−396. (48) Kashiwagi, T.; Yokoyama, K.; Ishikawa, K.; Ono, K.; Ejima, D.; Matsui, H.; Suzuki, E. Crystal structure of microbial transglutaminase from Streptoverticillium mobaraense. J. Biol. Chem. 2002, 277, 44252− 44260. (49) Keil, B. Specificity of proteolysis; Springer-Verlag: NewYork, New York, 1992; p 335. (50) Baker, L. E. New synthetic substrates for pepsin. J. Biol. Chem. 1951, 193, 809−819. (51) Hamuro, Y.; Coales, S. J.; Molnar, K. S.; Tuske, S. J.; Morrow, J. A. Specificity of immobilized porcine pepsin in H/D exchange compatible conditions. Rapid Commun. Mass Spectrom. 2008, 22, 1041−1046. (52) Powers, J. C.; Harley, A. D.; Myers, D. V. Subsite specificity of porcine pepsin. Adv. Exp. Med. Biol. 1977, 95, 141−157. (53) Farrell, H. M.; Jimenez-Flores, R.; Bleck, G. T.; Brown, E. M.; Butler, J. E.; Creamer, L. K.; Hicks, C. L.; Hollar, C. M.; Ng-Kwai-

Hang, K. F.; Swaisgood, H. E. Nomenclature of the proteins of cows’ milk − sixth revision. J. Dairy Sci. 2004, 87, 1641−1674. (54) Swaisgood, H. E. In Developments in Dairy Chemistry; Fox, P. F., Ed.; Applied Science Publishers: London, UK, 1982; Vol. 1; pp 1−59. (55) Mackie, A. R.; Mingins, J.; North, A. N. Characterisation of adsorbed layers of a disordered coil protein on polystyrene latex. J. Chem. Soc., Faraday Trans. 1991, 87, 3043−3049. (56) Maldonado-Valderrama, J.; Wilde, P. J.; Macierzanka, A.; Mackie, A. R. The role of bile salts in digestion. Adv. Colloid Interface Sci. 2011, 165, 36−46. (57) Juvonen, K. R.; Macierzanka, A.; Lille, M. E.; Laaksonen, D. E.; Herzig, K. H.; Mykkänen, H. M.; Niskanen, L. K.; Mackie, A.; Mills, C.; Poutanen, K. S.; Karhunen, L. J. Crosslinking with transglutaminase alters the postprandial metabolism and appetite responses of sodium caseinate stabilised emulsion in healthy young individuals. Infogest COST Scientific Meeting; La Croisic, France, 2011.

17362


Enzymatically Structured Emulsions in Simulated Gastrointestinal

Recommend Documents