Microencapsulation of Probiotic Cells for Food Applications

Transcription

1 This article was downloaded by: [Bayerische Staatsbibliothek] On: 16 February 2012, At: 08:10 Publisher: Taylor & Francis Informa Ltd Registered in England and Wales Registered Number: Registered office: Mortimer House, Mortimer Street, London W1T 3JH, UK Critical Reviews in Food Science and Nutrition Publication details, including instructions for authors and subscription information: Microencapsulation of Probiotic Cells for Food Applications Thomas Heidebach a, Petra Först a & Ulrich Kulozik a a ZIEL Research Center for Nutrition and Food Science, Institute for Food Process Engineering and Dairy Technology, Technische Universität München, Weihenstephan Weihenstephaner Berg 1, 85354, Freising-Weihenstephan, Germany Available online: 14 Feb 2012 To cite this article: Thomas Heidebach, Petra Först & Ulrich Kulozik (2012): Microencapsulation of Probiotic Cells for Food Applications, Critical Reviews in Food Science and Nutrition, 52:4, To link to this article: PLEASE SCROLL DOWN FOR ARTICLE Full terms and conditions of use: This article may be used for research, teaching, and private study purposes. Any substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any form to anyone is expressly forbidden. The publisher does not give any warranty express or implied or make any representation that the contents will be complete or accurate or up to date. The accuracy of any instructions, formulae, and drug doses should be independently verified with primary sources. The publisher shall not be liable for any loss, actions, claims, proceedings, demand, or costs or damages whatsoever or howsoever caused arising directly or indirectly in connection with or arising out of the use of this material.

2 Critical Reviews in Food Science and Nutrition, 52: (2012) Copyright C Taylor and Francis Group, LLC ISSN: / online DOI: / Microencapsulation of Probiotic Cells for Food Applications THOMAS HEIDEBACH, PETRA FÖRST, and ULRICH KULOZIK ZIEL Research Center for Nutrition and Food Science, Institute for Food Process Engineering and Dairy Technology, Technische Universität München, Weihenstephan Weihenstephaner Berg 1, 85354, Freising-Weihenstephan, Germany The addition of microencapsulated probiotic cells to food products is a relatively new functional food concept. Most of the published scientific research in this field is not older than ten years. However, the technological background reaches back to the 1980s, where lactic acid bacteria were microencapsulated within the concept of the so-called immobilized cell technology (ICT). Target applications of ICT were continuous fermentation processes and improved biomass production. The methods adopted from immobilized cell technology were applied for the microencapsulation of probiotics, often optimized towards specific requirements associated with the protection of probiotic cells in food applications. However, there are still significant hurdles with respect to currently available methods for probiotic cell microencapsulation. This is mainly due to the fact that important characteristics of microcapsules based on ICT appear to be in conflict with the requirements arising from an application of probiotic microcapsules in food products, with particle size and inappropriate matrix characteristics being the most prominent ones. Based on this situation the aim of this review is to give a critical overview of the current approaches regarding the microencapsulation of probiotic cells for food applications and to report on emerging developments. Keywords ANNOTATIONS Functional foods, immobilization, entrapment, Lactobacillus, Bifidobacterium ICT: Immobilized cell technology CFU: Colony forming units EY: Encapsulation yield Probiotic microcapsules: Microcapsules that contain probiotic cells as core-material NGYC: medium non-fat milk, glucose, yeast-extract, and cysteine medium MRS: medium de Man, Rogosa, Sharpe medium INTRODUCTION An important trend in the food industry in recent years is the demand for health promoting foods from which the concept of functional foods emerged. The term has been Address correspondence to Thomas Heidebach, ZIEL Research Center for Nutrition and Food Science, Institute for Food Process Engineering and Dairy Technology, Technische Universität München, Weihenstephan Weihenstephaner Berg 1, 85354, Freising-Weihenstephan, Germany. thomas.heidebach@wzw.tum.de 291 coined to describe foods fortified with ingredients capable of producing health benefits (Stanton et al., 2001). In this context, the addition of living probiotic microorganisms to food is a prominent way to create functional foods (Rodgers, 2008). Various health related properties of different probiotic strains are well documented and living lactic acid bacteria with probiotic activity are believed to play a beneficial role in the ecosystem of the human intestinal tract (Jia et al., 2008; Naidu et al., 1999; Tamime et al., 2005). Recently, genome-based studies started to provide insights regarding the mechanistic functions of probiotics in the ecosystem of the human gut (Ventura et al., 2009). However, the loss of bioactivity, that is the loss of living probiotic cell numbers during processing, storage, and gastrointestinal transit caused by various stress factors is an important issue and has to be avoided (Mattila-Sandholm et al., 2002; Shah, 2000; Siuta-Cruce and Goulet, 2001). Hence the protection of living probiotic cells became an important issue. In this context, microencapsulation is the most prominent technique for providing a protective environment for microorganisms under adverse conditions (Augustin, 2003; Champagne et al., 2005; Ross et al., 2005; Parada and Aguilera, 2007). Microencapsulation consists of coating or entrapment of a core material into capsules in the size range of a few micrometers up to a few millimeters (Kirby, 1991).

3 292 T. HEIDEBACH ET AL. In food systems microencapsulation can have various aims. A prevalent objective is to protect the core material from degradation by reducing its reactivity to environmental conditions (Gibbs et al., 1999; Schrooyen et al., 2001). This is mainly achieved by control of the mass transfer between the core and the external environment by using the shell material as a physical barrier (Champagne and Fustier, 2007; Desai and Park, 2005; Kailasapathy, 2002; Lopez-Rubio et al., 2006). A possible design commonly used for encapsulation of microbial cells involves the so-called matrix capsule, where the living cells as core material are embedded and immobilized randomly in a continuous matrix, which often is a hydrogel (Desai and Park, 2005). Hence, the terms immobilization and encapsulation are used as synonyms in most reported works about the microencapsulation of probiotics (Anal and Singh, 2007; Krasaekoopt et al., 2003). The motivation for microencapsulation of living probiotic cells is to decrease the unavoidable drop of living cell numbers from the first addition of the probiotic concentrate to the food, until they reach their final destination in the human gut. Along with this, a complete release of the probiotic cells from the capsule into the human gut should be ensured, because colonization of the gastrointestinal tract is seen as an important requirement to exert probiotic effects (Naidu et al., 1999). Furthermore, the capsules must be sufficiently small to avoid a negative sensorial impact on the functional food product they have been added to. However, until today many approaches regarding probiotic encapsulation have significant flaws when it comes to delivering these key features for food applications. As a result there are still many apparent technological hurdles associated with the currently available solutions for the microencapsulation of probiotic cells. Many of them can be explained by the fact that the currently used probiotic encapsulation techniques and matrix-materials emerged from the technological background of immobilized cell technology (ICT). However, some important characteristics of microcapsules based on ICT appear to be in contrast to the requirements arising from adding microcapsules containing probiotic cells (later on referred to as probiotic microcapsules) directly to a food product, with particle size and inappropriate matrix characteristics being the most prominent ones. Based on this perception this paper critically reviews the current approaches for microencapsulation of probiotics for food applications and resulting future perspectives. IMMOBILIZED CELL TECHNOLOGY (ICT) ICT is applied in biotechnology and involves the entrapment of living cells in spherical gel beads (Kailasapathy, 2002). It was successfully used in different areas of fermentation, with the main utilization of this technology settled in the dairy industry (De Giulio et al., 2005), in applications such as continuous inoculation of milk for yogurt or cheese making, lactic acid production, and optimized biomass production within the matrix of the beads (Champagne et al., 1994; Lacroix et al., 2005; Prevost et al., 1985). At the time when protection of probiotics became an issue in food applications, ICT-methods for the entrapment of lactic acid bacteria and its applications in the dairy industry were already well studied and established. Therefore, it occurred as an immediate near-in solution for the emerging problems of probiotic cell protection to use already existing ICT-techniques. However, besides physical protection of the entrapped cells, some of the main objectives of ICT-based applications differ from the above-mentioned targets of probiotic encapsulation, as outlined in the following section. Capsule Features Capsule Size The desired size of capsules for ICT-applications is dependent on the required cell growth, the mechanical strength, and the separation characteristics of the capsule (Lacroix et al., 2005). It was found that beads with diameters below 1 mm can result in separation problems in continuous inoculation processes, such as clogging of the filter when the capsules are removed from the substrate after the fermentation (Champagne et al., 1994). Furthermore, reducing the size of the gel beads below 1 mm may result in mechanical instability during longterm continuous fermentation (Audet et al., 1992). Therefore, in ICT applications, spheres with size ranges between 1 and 3 mm are preferably used. Gel Network Density The production of cell biomass within spherical gel-matrices is mainly controlled by diffusion and mass-transfer phenomena. Therefore, non-uniform cell growth in the colonized microcapsules results in the formation of a high cell density region near the capsule surface, leading to a 20- to 30-fold higher cell concentration than in the center of the gel (Champagne et al., 1994). A sometimes undesired leakage of cells into the surrounding medium occurs once the matrix space in the gels has been fully occupied and the gel then breaks due to mechanical stresses resulting from cell growth (Champagne et al., 1992; Klinkenberg et al., 2001). Therefore, a low density gel network, which provides sufficient space for the production of concentrated biomass within the polymer-gel, is a possible way to circumvent the problem of cell leakage (Park and Chang, 2000). Biopolymers, such as alginate, gellan-gum, xanthan, - carrageenan, locust-bean-gum, or mixtures thereof are suitable and most commonly used in ICT for dairy applications, because of their ability to easily build hydrogels at low concentrations of about 1% in aqueous solutions (Burey et al., 2008; Champagne et al., 1994). Since the gel beads are removed from the dairy product after incubation, the non-dairy origin of such polymers is not a major issue.

4 MICROENCAPSULATION OF PROBIOTIC CELLS 293 ICT-Capsules in Food Applications If microencapsulated probiotic cells are to be employed in final food applications, capsule-features that are important for a successful ICT-application can have an adverse impact. A low matrix density, required for concentrated biomass production within the capsules, is in contrast to the proposed protective mechanism of microencapsulation, that is, the creation of a physical barrier against unfavorable external conditions. Moreover, it must be considered that, in contrast to ICTapplications, probiotic microcapsules for food applications are clearly intended to remain in the food product until consumption. As a consequence, the size of the capsules must be considered with respect to the sensorial impact on the food. The matrix properties and release characteristics of microcapsules from ICT were not designed and optimized with respect to requirements of capsules intended to pass through the human gastro-intestinal tract. Alginate, which is most commonly used for probiotic encapsulation, can be extracted from the cell walls of marine algae. While it serves as a carbohydrate source for various marine molluscs (Gacesa, 1992), it is indigestible for humans, and behaves much like a dietary fiber (Brownlee et al., 2005). To effectively use probiotic microcapsules in food products, it must be ensured that the hydrocolloids used as matrix-material not only provide the desired barrier effect under acidic ph conditions, but are also digestible and, therefore, release the probiotic cells into the human gut. It can therefore be concluded that most of the important characteristics of microcapsules based on ICT do not meet the requirements of probiotic microcapsules, that are to be added directly to a food product. Encapsulation Techniques for Probiotic Cells The vast majority of microcapsules produced for ICTapplications in dairy systems are generated by two methods, the extrusion technique and the emulsion technique. Biopolymers such as alginate, gellan gum, xanthan, carrageenan, locust bean gum or mixtures thereof are commonly used as gelation material, since low concentrated (0.75 4%) aqueous solutions of these polymers can undergo mild ionotrophic and/or thermal gelation. Consequently, in most of the studies on probiotic cell encapsulation for food applications these methods are applied, mostly using alginate or gellan-xanthan mixtures as gelling agent (Champagne et al., 1994; Doleyres and Lacroix, 2005; Krasaekoopt et al., 2003). In recent years, spray drying was also utilized to encapsulate probiotic cells as an alternative to the encapsulation methods based on ICT, as outlined in the section titled Spray Drying. Extrusion Technique The extrusion technique involves preparing an aqueous hydrocolloid solution, adding concentrated microorganisms to Figure 1 Microencapsulation by means of the extension technique. it, and extruding the hydrocolloid-cell-mixture through a nozzle that forms droplets that fall into a hardening solution (see Fig. 1). In case of the most commonly used sodium-alginate, gelation can be achieved by dropping the droplets into a CaCl2-solution (Krasaekoopt et al., 2003). The size of the resulting capsules depends on the diameter of the orifice, the distance between the outlet, and the hardening-solution, and the viscosity of the hydrocolloid-cell mixture (Anal and Singh, 2007). Since the extrusion method was readily available from ICT, it was used by many researchers for the microencapsulation of probiotic cells, despite the large bead size ranges of mm (Krasaekoopt et al., 2003). Emulsion Technique For probiotic cell encapsulation, the most suitable method concerning control and flexible adjustment of the resulting capsule size is the emulsion technique. In this method a small volume of the aqueous hydrocolloidcell-mixture (discontinuous phase) is emulsified into a larger volume of vegetable oil (continuous phase). Once a water-inoil emulsion has been formed, the dispersed hydrocolloid-cellmixture must be insolubilized to form small beads within the oil phase (Krasaekoopt et al., 2003). When alginate capsules are produced, the microcapsules are hardened by slowly adding CaCl2-solution to the emulsion while stirring (see Fig. 2). When the calcium solution gets into contact with the dispersed alginate phase, instantaneous gelling occurs. Thus, the gelation kinetic is inhomogeneous, which sometimes leads to capsules having irregular shape (Sheu and Marshall, 1993). The technique was first developed by Nilsson et al. (1983) as a general method for immobilization of sensitive living cells. The authors stated that by adjusting the speed of a magnetic stirrer

5 294 T. HEIDEBACH ET AL. Figure 2 Microencapsulation by means of the emulsifying technique. during the emulsifying process capsules with average diameters between 0.1 and 5 mm could be produced at that time. The major parameters to control the size of the capsules are similar to those that influence particle size formation in common emulsifying processes, that is, the energy input during emulsification, the addition of emulsifiers, and the viscosity ratio between the dispersed and the continuous phase. For probiotic cell encapsulation, the emulsifying step is often accomplished by means of a magnetic stir bar or a directly driven mechanical stirrer (Ding and Shah, 2009a). In these cases the shear forces and resulting particle size reduction is rather undefined. However, with the emulsion method capsule sizes below 100 µm can be achieved when sufficiently high. Spray Drying Probiotic cell concentrates often need to be stored over longer periods prior to food manufacture and ingestion (De Giulio et al., 2005; Su et al., 2007). Hence, probiotic microcapsules are sometimes usually dried after production. In case of hydrogel-based microcapsules generated by extrusion or emulsification processes, freeze drying is frequently used to dry the capsules in a subsequent step after production (Godward and Kailasapathy, 2003; Heidebach et al., 2010, Kim et al., 2008; Lahtinen et al., 2007, Lee et al., 2004; Reid et al., 2007). An alternative method to achieve capsule-building and drying in a single step is spray-drying. Spray drying is a routine process in the food industry to convert liquids into dry powders. In recent years, spray drying has been utilized to encapsulate probiotic cells with the intention of not just simply drying, but as an alternative to the encapsulation methods based on ICT. In this context mixtures of probiotic cell concentrates were spray dried with aqueous solutions of various polymers, such as modified starch (O Riordan et al., 2001), gum arabic (Desmond et al., 2002), gelatin (Lian et al., 2003), whey protein isolate (Picot and Lacroix, 2004), maltodextrin mixed with gum arabic (Su et al., 2007), and ß-cyclodextrin mixed with gum arabic (Zhao et al., 2008), and their ability to protect the probiotic cells against adverse conditions was investigated. The advantage of spray drying is its wide availability in the food industry and that often favored small capsules with average diameters below 100 µm are usually generated at comparably low costs. However, in contrast to microcapsules generated from freeze-dried hydrogels, microcapsules prepared by this method are water soluble in most cases. Therefore, the cells are early released and are no longer protected from adverse conditions during product storage in non-dried products and during gastrointestinal transit (Krasaekoopt et al., 2003). The comparably high temperatures and rapid dehydration during spray drying generally lead to a deterioration of the cells, resulting in significant losses of living cells and a diminished resistance against unfavorable environmental conditions (Meng et al., 2008). It was shown that the survival of probiotic cells during spray drying increases with decreasing outlet air temperature (Ananta et al., 2005; Gardiner et al., 2000; Lian et al., 2002; To and Etzel, 1997). The chosen outlet air temperature is therefore often a compromise between the required residual water content and the probiotic cell survival, which is often as low as 1 10% in sufficiently dried powders (Desmond et al., 2002; Gardiner et al., 2000; Lian et al., 2002; Picot and Lacroix, 2004; Wang et al., 2004; Zhao et al., 2008). ASSESSMENT OF CURRENT METHODS FOR PROBIOTIC ENCAPSULATION Encapsulation Yield One of the main reasons for the application of biopolymers within ICT is the mild ionotrophic gelation suitable for a virtually loss-free entrapment of living microbial cells (Kailasapathy, 2002; Poncelet et al., 1992). This is in contrast to the abovementioned rather low probiotic survival generally found during probiotic encapsulation by spray-drying. By using the original extrusion method known from ICT, encapsulation yields (EY) of 100% were achieved for the encapsulation of various probiotic cells in large microcapsules

6 MICROENCAPSULATION OF PROBIOTIC CELLS 295 (Krasaekoopt et al., 2004; 2006; Kushal et al., 2006; Leverrier et al., 2005; Sun and Griffiths, 2000; Talwalkar and Kailasapathy, 2003; Urbanska et al., 2007). The EY is usually calculated by comparing the probiotic colony forming units (CFU) per gram dry matter of the initial polymer-cell-solution versus the generated microcapsules. The EY is therefore a combined parameter that describes the survival of viable cells and the efficacy of entrapment during the encapsulation procedure. The main reason for an EY below 100% is mainly probiotic cell damage due to detrimental conditions caused by the encapsulation process itself, such as heating, shear stress, or the application of concentrated solutes. Furthermore, a physical loss of cells into the hardening solution during the encapsulation process can appear in significant numbers. It should also be noted that a disintegration process is required to measure the concentration of living cells in the microcapsules. In the case of alginate-based capsules, dissolution of the capsules can be easily achieved by gently shaking them in a phosphate-buffer solution (Sheu and Marshall, 1993). In contrast, for capsules based on irreversible gelation, mechanical disintegration methods are often required. An incomplete disintegration as well as detrimental influences of the disintegration process can shift the found EY towards lower levels (Annan et al., 2008). Therapeutic Minimum and Core Load In contrast to capsules used for ICT, probiotic microcapsules for food applications are generally not intended to be propagated via a fermentation process that is accompanied by cell growth within the capsules. Entrapped cell growth preferentially takes place near the capsule surface because of better nutrient availability (Audet et al., 1992). This is undesirable in view of the physical protective effect of microencapsulation, being most effective within the core of the capsule. In probiotic foods a concentration of CFU probiotic cells per gram or ml of the resulting product has been suggested as an effective or therapeutic minimum (Agrawal, 2005; Champagne et al., 2005). Therefore, a high initial EY, that is, a high core load with living probiotic cells after the encapsulation, is required to match the therapeutic minimum in the food, especially at a preferably low capsule addition ratio. However, when ICT-methods were modified to produce capsules with physical characteristics suitable to meet the requirements of probiotic encapsulation in foods, that is, small capsules sizes or the use of more suitable matrix materials such as proteins, several problems that are associated with having high EY values may arise. Technological Challenges Arising from Probiotic Encapsulation Hydrocolloid-Based Microcapsules Besides spray-drying, the creation of microcapsules with diameters below 100 µm can also be achieved by modification of ICT methods. In most studies, the impact on the resulting EY was not assessed. However, some authors found that the required modifications led to undesirably low EY, as outlined below. Capela et al. (2007) encapsulated various probiotic strains in a 3% alginate solution by means of the emulsification technique using a magnetic stirrer system, resulting in microcapsules with an average diameter of 381 µm. Aqueous solutions of hydrocolloids used as precursors for encapsulation in ICT, such as alginate, carrageenan, gellan, or xanthan can have very high viscosities even at low concentrations. Hence, due to a high viscosity ratio between the dispersed and the continuous phase (mostly vegetable oil), a high energy input is necessary to produce sufficiently small microcapsules. An additional high-shear step during the emulsifying process was applied by means of an Ultra-Turrax, a Silverson mixer, or a high-pressure homogenizer in the study of Capela et al. (2007), to reduce the capsule size below 100 µm. The resulting EY differed significantly between strains and the method of homogenization. For Lactobacillus casei, Lactobacillus acidophilus, and Bifidobacterium longum, EY of 30% and less than 5% were found, respectively. Lactobacillus rhamnosus had an all over higher EY of 5% and 65%. The authors stated that individual probiotic strains may vary in their sensitivities to mechanical or thermal stresses caused by homogenization processes leading to a low survival of cells during the encapsulation process. In a study by Ding and Shah (2009a) a high shear process was applied using an Ultra-Turrax or microfluidizer to reduce the capsule-size of alginate-based capsules generated by the emulsifcation technique. For both devices it was found that the content of living cells from each of the eight different, individually encapsulated strains within the capsules gradually decreased from approximately 0.5 up to 3.5 log cycles CFU with increasing energy input during the emulsion process. Relatively low EY were also found by Cui et al. (2000), when small alginate microcapsules with sizes between 5 and 200 µm were prepared by spraying an aqueous mixture of alginate and Bifidobacteria into a CaCl2 solution using an air-driven atomization device. The authors found an EY of only 12.3%. The low EY was explained by the loss of cells in the surrounding aqueous CaCl2-solution during the gelation step. Apparently this was caused by the high surface-to-volume ratio, compared to large capsules that were produced by the standard dropping method. Protein-Based Microcapsules An alternative strategy of using different matrix-materials compared to ICT-technology involves the application of protein solutions as precursor material for the capsule matrix. On the one hand, small microcapsules can be produced by the emulsion technique with less effort due to the good emulsifying properties and the rather low viscosity of food protein solutions, leading to a lower shear stress and resulting in a higher EY (Heidebach et al., 2009a). Furthermore, in contrast to the commonly used

7 296 T. HEIDEBACH ET AL. hydrocolloids, even highly concentrated aqueous solutions of most proteins have a relatively low viscosity. This facilitates the formation of microcapsules with dense gel network that provide a substantial buffering capacity, thereby supporting the idea of a protective barrier between the sensitive core material and the surrounding environment. However, the application of food protein based matrix materials as alternatives to commonly used polysaccharide-based matrices sometimes requires the modification or even establishment of novel encapsulation techniques, involving different gelation mechanisms during encapsulation. Annan et al. (2008) encapsulated Bifidobacterium adolescentis in small alginatecoated gelatine microcapsules, with average diameters of 50 µm. The capsules were produced by covalently cross-linking the gelatine-cell mixture with genipin, a non-toxic cross-linker from plants, during the emulsification process. An EY of only 41 43% was achieved. The authors stated that the strong physical stability of covalently cross-linked gels could have prevented a complete release of the cells leading to a lower EY. Hence, in this case it is not clear as to what extent a decrease of the EY is caused by the encapsulation process itself. Picot and Lacroix (2004) mixed various strains of Bifidobacteria separately with heat treated, denatured whey protein solutions at 40 C and then spray dried these mixtures to generate water-insoluble microcapsules with sizes of 3 75 µm. The authors found EY between 0.71 and 25.7%, depending on the heat tolerance of the strain. Yet, the authors concluded that cell damage caused by the relatively high shear levels and the thermal inactivation during the spray-drying process was a major drawback of the process. Reid et al. (2005) reported an EY of 22% during encapsulation by Ca2+-induced cold gelation via extrusion of a pre-heated whey protein solution that contained Lactobacillus ssp. in a concentrated CaCl2-solution. The authors stated that the exposure to concentrated CaCl2-solution during the gelation process may be responsible for the high mortality rates of the entrapped microorganisms during the encapsulation process. In a study of Weinbreck et al. (2010) water-insoluble microcapsules were created by spraying of pre-denaturated whey protein solution mixed with probiotic Lactobacillus rhamnosus onto core-particles in a fluidized bed coater. In this case a 103- fold decline in cell viability during encapsulation was found, which was attributed to cell damage during drying. An emulsion process based on enzymatic gelation of caseinate solutions to produce probiotic microcapsules was used by Heidebach et al. (2009b). With this method an EY of 70% and 93% was achieved for Lactobacillus paracasei and Bifidobacterium lactis, respectively. The high physical stability of covalently cross-linked gels could prevent a complete release of cells, and therefore be responsible for an EY of slightly less than 100%. Encapsulation of these strains into another protein based gel matrix, produced by enzymatic rennet gelation of a 35% skim-milk concentrate, led to a complete recovery of viable cells after the encapsulation process (Heidebach et al., 2009a). Optimization of ICT methods to match the specific requirements associated with the protection of probiotic cells in food applications is common practice (Kailasapathy, 2002). However, from the above-mentioned studies it can be seen that one of the most important features of ICT-capsules, namely the high initial EY can be lost during the above-mentioned modifications. The use of protein based hydrogels for the encapsulation of probiotic cells instead of polysaccharide biopolymers appears to be more promising to obtain small microcapsules if mild gelling mechanisms are applied, such as cold-induced gelation of whey concentrates or enzymatic gelation (Chen et al., 2006; 2003; Heidebach et al., 2009a; 2009b). Lipid-Based Microcapsules Aside from carbohydrates and proteins, the use of lipid-based encapsulation systems for the encapsulation of probiotics has not yet been well explored. Matrix-encapsulation can be achieved by mixing probiotic cells with molten fat and subsequent cooling. However, dispersion of probiotic cell concentrates in oil was reported to be a difficult technological task (Modler and Villa-Garcia, 1993; Picot and Lacroix, 2004). Premature melting of the capsules at elevated temperatures, like the ones in the human body must be considered. Because of possible separation problems it is likely that applications are limited to solid foods (Lahtinen et al., 2007). Studies revealed that encapsulation of probiotics in butterfat afforded no protective effect during storage in frozen yogurt (Modler and Villa-Garcia, 1993). Lahtinen et al. (2007) found only a slight protective effect when probiotics were encapsulated in cocoa butter during storage in fermented and non-fermented oat-drinks. Hence, fat-based microcapsules currently seem less suitable for probiotic encapsulation compared with polysaccharide- or protein-based microcapsules. IMPACT OF MICROCAPSULES ON PROBIOTIC SURVIVAL AND FOOD CHARACTERISTICS For food applications, creating microcapsules with sufficiently small average diameter is one of the most significant bottlenecks. On the one side, larger capsule diameters, and hence a higher volume-to-surface-ratio increases the likeliness of a protective effect (Anal and Singh, 2007). On the other side, the capsules must be small enough to not negatively impact the sensory properties of the food-product (Champagne and Fustier, 2007). This conflict of targets can lead to difficulties when it comes to application of probiotic microcapsules in food. From sensory studies dealing with the mouth feel sensation of particles in foods it can be concluded that large, hard, or sharp particles added in a high concentration to a low viscous medium produce a more rough, gritty, and unpleasant sensation, compared to small, soft, and spherical particles that are added at a lower concentration to a high viscous medium or gel (Engelen et al., 2005; Imai et al., 1995).

8 MICROENCAPSULATION OF PROBIOTIC CELLS 297 Table 1 Applications of encapsulated probiotics in yogurt Encapsulated strain 1 Matrix material and encapsulation technique Average size of capsules (µm) Storage time Increased survival due to microen-capsulation Reference Bifidobacterium bifidum and Alginate; emulsion not given 1 week 1 log cycle for each strain (Hussein and Kebary, 1999) Bifidobacterium infantis Bifidobacterium infantis Gellan-xanthan; extrusion weeks 1 log cycle (Sun and Griffiths, 2000) Two strains of Bifidobacterium κ-carrageenan; emulsion days 0.5 log cycles for each (Adhikari et al., 2000). longum strain Lactobacillus acidophilus and Alginate, co-encapsulation with weeks 0.5 log cycles for each (Sultana et al., 2000) Bifidobacterium infantis 2% resistant starch; emulsion strain Two strains of Bifidobacterium κ-carrageenan; emulsion days 1 log cycle for each strain (Adhikari et al., 2003) longum Two strains of Lactobacillus acidophilus Alginate, co-encapsulation with 1% resistant starch, coated with chitosan; extrusion weeks 3 log cycles for each strain (Iyer and Kailasapathy, 2005) Lactobacillus acidophilus, Bifidobacterium bifidum, and Lactobacillus casei Bifidobacterium lactis and Lactobacillus acidophilus Alginate, coated with chitosan; extrusion Alginate, co-encapsulation with 1% resistant starch; emulsion 1 If not stated otherwise, probiotic strains were separately microencapsulated While sugar crystals, that is, hard and irregular particles, can already be detected in various foods at sizes ranging from about 10 to 20 µm (Imai et al., 1999), soft spherical hydrogel microcapsules have a higher threshold level regarding a graininess detection. Hansen et al. (2002) reported a size below 100 µm as desirable to avoid having negative sensorial impacts of microcapsules in food. Application of Probiotic Microcapsules in Yogurt weeks 1 log cycle for each strain (Krasaekoopt et al., 2006) weeks Bifidobacterium lactis: 1 log cycle Lactobacillus acidophilus: 2 log cycles (Kailasapathy, 2006) Various fermented food products have already been supplemented with probiotic microcapsules. A major focus of most of these studies is an evaluation of the protective effect due to microencapsulation during product storage. Yogurt is the most extensively supplemented product so far, due to the fact that the probiotic activity found in yogurt is often rather low (Kailasapathy and Rybka, 1997; Shah, 2000). Table 1 gives an overview of the application of probiotic microcapsules in yogurt including the most important experimental conditions. From the studies shown in Table 1 it can be concluded that encapsulation of various probiotic strains in hydrocolloid gels enhances their survival during storage in yogurt at about 0.5 to 3 log cycles CFU. The protective effect is generally explained by limited diffusion of inhibitory substances such as metabolic products from the starter cultures, H2O2, lactic acid, and bacteriocin into the capsules (Krasaekoopt et al., 2006; Sun and Griffiths, 2000). Taken together with the rather low ph of 4.5 or below (Lourens-Hattingh and Viljoen, 2001), cell death from presence of oxygen has been discussed as one of the major factors for the low survival rates of probiotics in yogurt. Talwalkar and Kailasapathy (2003) showed that encapsulation in alginate hydrogels offers substantial protection for probiotics under aerobic conditions for several probiotic strains and could therefore be responsible for higher survival rates of encapsulated cells during storage in yogurt. In all cases, a protective effect could only be achieved by using capsules with sizes of mm. In this case, a negative sensorial impact of the capsules on the food product is most likely. This was confirmed by sensory evaluations of such products which was investigated in some studies mentioned in Table 1. In the study of Adhikari et al. (2000), sensory analyses showed that consumers preferred the yogurt containing free probiotic cells over the one containing probiotic microcapsules by ranking with a higher overall liking. It is not quite clear if the inferior sensory characteristics were due to off-flavor or grittiness. While the survival tests were performed with plain yogurt in the study of Adhikari et al. (2003), yogurt used for sensorial evaluation additionally contained 13% blackberry jam. Panellists detected a grainy structure for the yogurts containing probiotic microcapsules and a worse overall acceptability compared to yogurts containing free cells. Similarly, Kailasapathy (2006) reported that sensorial analyses revealed a slight grittiness for the yogurts containing microcapsules compared to the yogurts containing free cells. Based on the literature available so far, it appears that microencapsulation by using capsules based on ICT-methods results in increased probiotic survival during storage in yogurt. However, the addition of probiotic microcapsules also leads to inferior sensory attributes, compared to yogurt containing the respective probiotic cells in free form. While there is a direct relationship between large capsules sizes and graininess in yogurt, further changes in flavor due to more complex reactions between the hydrocolloid from the capsules and the yogurt matrix, as well as altered metabolic profiles of microorganisms caused by encapsulation, should be considered in future investigations.

9 298 T. HEIDEBACH ET AL. Table 2 Applications of encapsulated probiotics in cheese Encapsulated strain 1 Matrix material and encapsulation technique Average size of capsules (µm) Storage time and food matrix Increased survival due to microencapsulation Reference Bifidobacterium bifidum κ-carrageenan; extrusion Not given 6 months in cheddar cheese Bifidobacterium bifidum, Bifidobacterium infantis, and Bifidobacterium longum, encapsulated together Two different strains of Lactobacillus acidophilus, Bifidobacterium lactis, and Bifidobacterium infantis Lactobacillus acidophilus and Bifidobacterium lactis Bifidobacterium bifidum and Lactobacillus acidophilus Alginate; extrusion weeks in fresh cheese Alginate, co-encapsulation with 2% resistant starch; emulsion Alginate, co-encapsulation with 2% resistant starch; emulsion Alginate; extrusion (a) κ-carrageenan; emulsion (b) 1 If not stated otherwise, probiotic strains were separately microencapsulated Application of Probiotic Microcapsules in Cheese months in cheddar cheese Not given (a) (b) Next to yogurt, cheese is often used as a target of supplementation with probiotic microcapsules (Table 2). Table 2 shows that in the case of rennet cheese, encapsulation is not always useful, since physiological conditions in a hydrocolloid matrix of carrageenan or alginate were less favorable for probiotic cells compared to a milk protein based rennet-gel-matrix. This was expressed in higher cell counts of cheeses containing free cells, compared to those with encapsulated cells at the end of storage (Dinakar and Mistry, 1994; Godward and Kailasapathy, 2003; Kailasapathy and Masondole, 2005). Conflicting results found by Ozer et al. (2009) could be possibly explained by the high concentration of salt (12%) in the brine, rendering the cheese-matrix inappropriate for probiotic survival in this case. The sensory impact of microcapsules supplementation on the cheese was evaluated by Dinakar and Mistry (1994). Capsule sizes were not stated, but since the authors used the extrusion method, relatively large capsule diameters are likely to have been present. However, about 3% capsules within the cheese matrix did not alter the cheeses sensory. Accordingly, Ozer et al. (2009) and Gobbetti et al. (1998) report that cheeses containing encapsulated probiotics did not differ from cheeses containing free cells in terms of their sensory properties. In contrast to this, Godward and Kailasapathy (2003) found no difference in flavor between cheeses containing free or encapsulated probiotics. However, in this case, grittiness for cheeses containing capsules was detected (Table 2). Other Foods Containing Encapsulated Probiotics Table 3 shows results from studies with other foods, supplemented with probiotic microcapsules. Food matrices generally 7 weeks in feta cheese 90 days in white-brined cheese Higher cell survival in samples containing non-encapsulated cells (Dinakar and Mistry, 1994) No protective effect (Gobbetti et al., 1998) Higher cell survival in samples containing non-encapsulated cells Higher cell survival in samples containing non-encapsulated cells 2 log cycles for each strain for (a) and (b) (Godward and Kailasapathy, 2003) (Kailasapathy and Masondole, 2005) (Ozer et al., 2009) differ in their suitability as a carrier for probiotic cells. As an example, probiotic cells show a higher stability in frozen foods, such as ice-cream, compared to refrigerated foods, such as yogurt. From Table 3 it is apparent that in frozen products even very small microcapsules have the ability to protect probiotic cells during storage. It was reported by Sheu et al. (1993) and Homayouni et al. (2008) that the use of such small capsules avoids a negative impact on sensory impression. Also, Hansen et al. (2002) stated that such capsules were small enough to avoid a grainy structure in milk. Nevertheless, in contrast to milk containing free probiotic cells, a bitter, sharp off-flavor was detected by the panellists in samples containing encapsulated cells (Table 3). Hence, there is evidence that also small capsules with an average diameter of about 30 µm that are not detectable by sensory tests can protect probiotic cells in foods. However, it has not been proven yet, whether the application of such small microcapsules can lead to an improved survival during storage in yogurt without simultaneously inducing a negative sensorial effect. In some of the other studies presented in Table 3, sensory evaluation was performed as well. In the study of McMaster et al. (2005), sensory evaluation revealed no off-flavor or grittiness due to incorporated microcapsules. This can be explained by the low level of addition of only 0.04% capsules to the beverages. It should be noted though that the mouthfeel of Mahewu has been described as grainy per se and Amasi has a thick, smooth structure as stated by the authors. Muthukumarasamy and Holley (2006) added probiotic microcapsules to a sausage batter before fermentation, at a level of 1%. A sensory evaluation revealed that no difference concerning texture, flavor and overall acceptability was found between sausages containing free or encapsulated probiotics. In the study of Khalil and Mansour (1998), improved sensory properties were reported for mayonnaise containing encapsulated cells, with

10 MICROENCAPSULATION OF PROBIOTIC CELLS 299 Table 3 Applications of encapsulated probiotics in other foods Encapsulated strain 1 Matrix material and encapsulation technique Average size of capsules (µm) Storage time and food matrix Increased survival due to microencapsulation Reference Lactobacillus bulgaricus Alginate; emulsion 15 (a) 30 (b) 2 weeks in frozen ice-milk at 18 C Bifidobacterium bifidum and Bifidobacterium infantis Lactobacillus acidophilus and Bifidobacterium ssp Alginate; emulsion Not given 8 weeks in mayonnaise with a ph of 4.4. Alginate; emulsion Not given 12 weeks in fermented frozen dairy dessert with a ph of 4.5 at 18 C No protective effect for (a); doubled survival rate for (b) About 5 log cycles for each strain 2 3 log cycles for each strain (Sheu et al., 1993) (Khalil and Mansour, 1998) (Shah and Ravula, 2000) Bifidobacterium longum Alginate; emulsion 20 2 weeks in non-acidified milk 0.5 log cycles (Hansen et al., 2002) Bifidobacterium lactis Gellan-xanthan; extrusion 640 (a): 3 weeks in African beverage based on (a): 2 log cycles (b): No protective effect (McMaster et al., 2005a) fermented maize with a ph of 3.5 (Mahewu) (b): 3 weeks in African beverage based on fermented milk with a ph of 4.5 (Amasi) Lactobacillus reuteri Alginate; emulsion (a) Alginate; extrusion (b) 40 (a) (b) 27 days in dry fermented sausages 2 log cycles for (a) and (b) (Muthukumarasamy and Holley, 2006) Bifidobacterium lactis and Lactobacillus casei Alginate, co-encapsulation with 2% resistant starch; emulsion 1 If not stated otherwise, probiotic strains were separately microencapsulated respect to flavor and texture. However, since no information about the addition level and the capsule size was given, hardly any conclusion can be drawn (Table 3). These studies show that large hydrogel-capsules can be successfully applied in some types of foods without altering the sensory properties. Compatibility mainly depends on the physical characteristics of the surrounding food matrix. While for yogurt capsules with average sizes above 200 µm were shown to adversely affect the mouthfeel, gelled foods such as cheese and foods with a structure that is naturally associated with coarseness seem to be more suitable for the application of large microcapsules. SURVIVAL DURING GASTRIC TRANSIT To obtain a notable health effect from the ingestion and colonization of probiotic cells in the gut, microorganisms must survive transit through the low ph gastric environment, which is an even tougher challenge compared to surviving processing and product conditions. The strong acidic conditions in the human stomach as a natural barrier of the host considerably reduce the number of living probiotic cells (Naidu et al., 1999; Ross et al., 2005). This makes the gastric transit the most crucial hurdle with respect to probiotic survival in food applications (Agrawal, 2005). In some cases enteric polymers, originally developed for controlled release of drugs in medical applications, were used to microencapsulate probiotic cells. By using cellulose-acetatephthalate (Favaro-Trindade and Grosso, 2002; Rao et al., 1989), days in ice-cream at 20 C 2 log cycles for each strain (Homayouni et al., 2008) Sureteric R, and AcrylEze R (Liserre et al., 2007) or Eudragit R (Graff et al., 2008) impressive protective effects under simulated gastric conditions were found. However, these substances, suitable for medical applications, are not permitted in food products (O Riordan et al., 2001). All common encapsulation techniques for probiotics in food applications result in matrix capsules, where the living cells as core material are embedded and immobilized in a continuous hydrogel matrix. On this account, it can be presumed that the protective effect of encapsulated probiotics depend on the physical characteristics of the capsule matrix and the capsule size. Many other factors influence the survival rates of encapsulated probiotic cells during simulated gastric transit, especially the sensitivity of the microbial strain as such, and the composition of the gastric fluid. The survival of a certain strain of bacteria under acidic conditions generally depends on the ability to control the activity of the membrane-bound ion transport system that generates the proton motive force (Booth, 1985). The regulation capacity of the cytoplasmic ph under stongly acidic external conditions is believed to strongly depend on the activity of the H+-ATPase, which can widely differ from strain to strain (Matsumoto et al., 2004). In the majority of studies a low concentrated ( %) NaCl-solution, adjusted with HCl to the target ph was used to simulate gastric juice without pepsin, according to the US- Pharmacopeia (USP, 2008). The rather simple composition of the simulated gastric juice commonly used can be explained by the fact that the gastric bactericidal barrier in vivo is primarily ph-hydrochloric acid dependent, with other constituents of gastric juice contributing little, if any, detectable effect on the killing of microorganisms (Giannellra et al., 1972).

11 300 T. HEIDEBACH ET AL. However, aside from the application of various ph-values, sometimes additional substances have been added, leading to a wide variation of simulated gastric conditions as outlined in the following sections. In Vitro Investigations Alginate-Based Microcapsules Many researchers have investigated whether a protective effect of an alginate gel-matrix, generated by ionotrophic gelation of alginate solutions at various concentrations of 1% 3% during gastric transit can be achieved. An important prerequisite is structural integrity of the microcapsules during gastric transit. Several in vitro studies have shown that alginate microcapsules remain physically stable at low ph-values in gastric fluid (Allan- Wojtas et al., 2008; Annan et al., 2008; Cui et al., 2000; Hansen et al., 2002; Iyer et al., 2004; Martoni et al., 2007). However, with respect to the achievable protective effect, published data available so far are inconsistent, as outlined in Table 4. In some of the studies mentioned in Table 4, prebiotic resistant starch granules were co-encapsulated together with the microorganisms, a concept referred to as synbiotics. Prebiotics are non-digestible food ingredients that beneficially affect the host by selectively stimulating the growth and activity of one or a limited number of bacteria in the colon. Synbiotics are created by a combined application of pro- and prebiotics (Fooks et al., 1999; Rastall and Maitin, 2002). It is generally thought that the resistance of probiotics against the harsh ph-conditions in the human gastro-intestinal-tract can be enhanced by coupling it with a selective growth promoter (Gibson, 2004). Iyer and Kailasapathy (2005), found an increase in the protective effect, due to co-encapsulation with resistant starch, compared to probiotic cells that were encapsulated in alginate alone. The authors assumed that the water-insoluble starch corns could block the pores of the capsules and therefore abate diffusion of acid into the network. An important conclusion that can be drawn from the results presented in Table 4 is that no direct relationship between capsule size or alginate concentration and the respective protective effect can be found. Some authors further studied the influence of capsule size or alginate concentration on the protective effect under equal conditions (Table 5). From the results displayed in Table 4 and Table 5 it becomes clear that a general statement about the suitability of an alginate gel matrix as a protective barrier towards a low ph environment is not possible. Table 5 shows that in direct comparison larger capsule sizes as well as higher alginate concentrations tend to provide a better protective effect. In contrast, as illustrated in Table 4, no such relationship can be found. A possible explanation could be that the variation of experimental simulated gastric conditions and the strain dependent acid-sensitivity dominates over the general improvement of alginate encapsulated cells in comparison to free cells. From Table 4 it can be further seen that in general the highest increase in survival is achieved with capsules that were freezedried before the simulated gastric experiments. A possible explanation is that the addition of dried powders to simulated gastric juice instead of hydrated capsules requires certain time for rehydration of the capsules and could therefore lead to a delayed penetration of the capsules with acid. Furthermore, the addition of capsules to simulated gastric juice can generally lead to an increase in ph of the resulting mixture. Despite the fact that this effect can greatly affect survival rates, it is generally not considered in most studies on simulated gastric survival of encapsulated probiotics. This was extensively discussed by Saarela et al. (2006) as generally aflaw in experimental design, when probiotic cells were tested against acid stress. In case of freeze-dried capsules, the addition of the same amount of dry powder instead of hydrated capsules can result in a higher buffering capacity. This, in turn, leads to a higher resulting ph which may explain the outstanding results that have in some cases been obtained with respect to freezedried capsules. Standardized simulated gastric conditions would therefore allow a more meaningful comparison of encapsulation systems for probiotics. For the failure of a protective effect of alginate during incubation of hydrated microcapsules under simulated gastric conditions different explanations can be found. The porosity of the alginate gel allows the diffusion of H+-ions into the gel, thus affecting the cells (Trindade and Grosso, 2000). Hansen et al. (2002) suggested that the porosity of the alginate matrix is increased due to the presence of the bacteria during the gelation process. Le-Tien et al. (2004) encapsulated a ph-sensitive color indicator in 3 mm alginate beads from 1.5 or 2.5% alginate solutions. The authors found that after incubation in simulated gastric juice at ph 1.5 the internal capsule ph went below 2 after approximately 8 min incubation time, independent of the alginate concentration used. It was concluded that alginate gels have a limited buffering capacity, and the alginate gel structure should be seen as a highly porous hydrogel, that provides virtually no barrier effect against the diffusion of H+-ions into the gel. Coating of the Capsule-Matrix Coating deposits an additional membrane-layer on the capsule surface. This leads to an increase in mechanical strength and a more pronounced barrier function. This is typically achieved by immersing the hydrogel capsules into a solution of coating polymer. Coating of biopolymer capsules has been a well known technique from the field of ICT, with its original goal to slow and reduce the release of cells into the surrounding medium (Champagne et al., 1992). It is also widely applied in the field of artificial cell therapy (Prakash and Martoni, 2006). The highly porous alginate network found in bacteria-loaded microcapsules may explain the limited protective effect under simulated gastric conditions (Allan-Wojtas et al., 2008). Thus, it was investigated whether coating can increase the protective effect under such conditions. Poly-l-lysine and chitosan are