**4. Mannitol and sorbitol production**

Sugar alcohols are hydrogenated carbohydrates widely used in the food industry as sugar replacers. Mannitol and sorbitol are used as food additives due to their sweetening effect (about half as sweet as sucrose) and low calorie content. They are also used in the food and pharmaceutical industries due to their technological properties, such as texturing agents, humectants, softeners and color stabilizers. In nature, mannitol is found in some plants, algae and mushrooms, and sorbitol is found in many fruits and vegetables. Those polyols are also produced by fungi, yeast and bacteria, where they play several roles in carbon storage and protection during osmotic and oxidative stresses. Industrial production of most sugar alcohols is performed by catalytic reduction of sugars with hydrogen gas and nickel at high temperature and pressure, for which highly pure sugar substrates and costlychromatographic purification steps are required. Regardless the limitations of this chemical method, it is until now the only process able to assume the high market demand of sorbitol and mannitol, estimated to be thousands of tons per year. However, processes using

In addition to rational methods of metabolic engineering, lactic acid production has also been enhanced by a combination of classical strain improvement methods (nitrosoguanidine and ultraviolet mutagenesis) with whole-genome shuffling by recursive protoplast fusion. In this way, shuffled strains derived from an industrial strain of *Lactobacillus* have been isolated, and they produce threefold more lactic acid than the wild type at pH 4.0 (Patnaik *et al.*, 2002). Shuffled *L. rhamnosus* strains with improved tolerance to glucose and enhanced Llactate production has also been obtained (Yasuda *et al.*, 2008). In the same way, a fusant derived from *Lactobacillus delbueckii* able of growing at low pH and utilizing starch from cassava bagasse was obtained and it produced large amounts of L-lactic (John *et al.*, 2008).

Diacetyl and acetoin are important compounds of buttery flavor in fermented foods and are used as additives in the food industry. Both compounds are derived from pyruvate, which is converted to -acetolactate by the action of -acetolactate synthase or acetohydroxyacid synthase. Then, acetoin is formed by the activity of -acetolactate decarboxylase on acetolactate and diacetyl results from a non-enzymatic oxidative decarboxilation of acetolactate (Figure 2). Most metabolic engineering approaches to produce diacetyl/acetoin by fermentation have been developed in the model LAB *L. lactis*, in which strains that divert an important part of pyruvate flux towards the production of -acetolactate have been constructed (Hugenholtz *et al.*, 2000; Lopez de Felipe *et al.*, 1998). *ilvBN* genes, encoding acetohydroxyacid synthase from *L. lactis*, have been expressed from the lactose operon in *L. casei*, an organism which shows marginal production of diacetyl/acetoin, resulting in increased diacetyl formation (Gosalbes *et al.*, 2000). In addition, to enhance diacetyl/acetoin production, the amount of pyruvate available for IlvBN was increased by blocking pyruvate alternative pathways in *L. casei*. Thus, the *L. casei* strain that expresses the *ilvBN* genes was mutated in the *ldh* gene and in *pdhC*, encoding the E2 subunit of the pyruvate dehydrogenase enzyme. The introduction of these mutations resulted in an increased capacity to synthesize diacetyl/acetoin from lactose fermentation in whey permeate (1400

Sugar alcohols are hydrogenated carbohydrates widely used in the food industry as sugar replacers. Mannitol and sorbitol are used as food additives due to their sweetening effect (about half as sweet as sucrose) and low calorie content. They are also used in the food and pharmaceutical industries due to their technological properties, such as texturing agents, humectants, softeners and color stabilizers. In nature, mannitol is found in some plants, algae and mushrooms, and sorbitol is found in many fruits and vegetables. Those polyols are also produced by fungi, yeast and bacteria, where they play several roles in carbon storage and protection during osmotic and oxidative stresses. Industrial production of most sugar alcohols is performed by catalytic reduction of sugars with hydrogen gas and nickel at high temperature and pressure, for which highly pure sugar substrates and costlychromatographic purification steps are required. Regardless the limitations of this chemical method, it is until now the only process able to assume the high market demand of sorbitol and mannitol, estimated to be thousands of tons per year. However, processes using

**3. Diacetyl and acetoin production**

mg/l at pH 5.5) (Nadal *et al.*, 2009).

**4. Mannitol and sorbitol production** 

bacteria and yeasts have demonstrated that biotechnological production may represent an efficient and cost-effective alternative to the chemical production.

The production of polyols by using genetically engineered LAB has been recently reviewed (Monedero *et al.*, 2010). Mannitol is a natural fermentation product in heterofermentative LAB, in which the NADH generated during sugar metabolism is regenerated by the

Fig. 2. Proposed pathways for sorbitol, mannitol, acetoin and diacetyl production by engineered lactic acid bacteria. [1] phosphoenolpyruvate: lactose phosphotransferase system, [2] phosphoenolpyruvate: glucose phosphotransferase system, [3] phosphogalactosidase, [4] glucokinase, [5] phosphoglucose isomerase, [6] sorbitol-6P dehydrogenase, [7] mannitol-1P dehydrogenase, [8] phosphofructokinase, [9] fructose 1,6 bisP aldolase, [10] galactose-6P isomerase, [11] tagatose-6P kinase, [12] tagatose-1,6DP aldolase, [13] triose-phosphate isomerase, [14] lactate dehydrogenase, [15]-acetolactate synthase, [16] acetohydroxyacid synthase , [17]-acetolactate decarboxylase, [18] pyruvate oxidase, [19] pyruvate dehydrogenase complex, [20] pyruvate-formate lyase. DHAP, dihydroxiacetone phosphate.

fermentations using whey permeate, a waste product from the dairy industry with high concentration of lactose, resulted in a conversion rate of 9.4% of lactose into sorbitol (De Boeck *et al.*, 2010). *L. plantarum* has also been metabolically engineered to produce sorbitol by constitutive overexpression of either *srlD1* or *srlD2* genes that encode S6PDH activities in a mutant strain deficient in LDH activity (Ladero *et al.*, 2007). Using non-growing or growing cells under pH control resulted in a very efficiency conversion rate of about 65% and 25%, respectively, of sugar into sorbitol. The different efficiencies were suggested to be

147

Some LAB produced EPS, which are extracellular polysaccharides, with important characteristics for the dairy industry. They are used to improve the rheological and textural properties of fermented foods. EPS have also potential as food additives and functional food ingredients. In this sense they are claimed to act as prebiotics in the intestine (Bello *et al.*, 2001) and to stimulate the immune system (Vinderola *et al.*, 2006). The synthesis of EPS in LAB starts at the glycolytic intermediate glucose-6P, which connects the anabolic pathways of biosynthesis of sugar nucleotides, the precursors of the EPS, and the catabolic pathways for obtaining energy through the glycolysis (Figure 1). Glucose-6P is converted to glucose-1P by the -phospoglucomutase (-Pgm) activity, and this sugar phosphate is further metabolized to UDP-glucose and UDP-galactose by the consecutively action of enzymes UDP-glucose pyrophosphorylase (GalU) and UDP-galactose 4-epimerase (GalE) (Boels *et al.*, 2001). Glucose-1P is also substrate for dTDP- glucose pyrophosphorylase to produce dTDPglucose, which will be further metabolized to dTDP-rhanmnose. Glucose, galactose and rhamnose are the principal sugars found in the EPS produced by LAB. The subsequent steps in the synthesis of EPS is the assembly of the monosaccharide repeating unit by specific glycosyltransferases, the polymerization of the repeating units and the secretion from the cell (Welman *et al.*, 2006). The enzymes that participate in these stages are encoded by genes that form part of *eps* gene clusters. Genetic engineering strategies could be applied to one or more of those stages involved in the EPS biosynthesis in order to increase the EPS production or to modify its composition, however, until now only strains of *L. lactis* and *Streptococcus thermophilus* have been genetically modified to enhance EPS biosynthesis. In *S. thermophilus* the modification of the levels of the GalU, PgmA and the Leloir route enzymes resulted in increased levels of EPS (Levander *et al.*, 2002). Homologous overexpression of a complete *eps* operon in *L. lactis* resulted in about fourfold increase in EPS production (Boels *et al.*, 2003). In *Lactobacillus* species there are no examples of metabolic engineering strategies aimed to produce EPS. In this species the synthesis of EPS has been improved by modifying the culture conditions, such as carbon source and pH. As well, chemically induced mutants of *Lactobacillus* species that produce higher amounts of EPS than the parental strain have been isolated. The synthesis of EPS by *L. casei* strain CRL 87 was improved by using galactose as carbon source at a controlled pH of 5.0, and the high EPS production was correlated with high activity level of the enzymes involve in the synthesis of UDP-sugars (Mozzi *et al.*, 2003). Similar approaches were applied for *L. helveticus* strain ATCC 15807, which produces a higher amount of EPS from lactose at pH 4.5 than at pH 6.2, which was correlated with higher levels of -Pgm activity (Torino *et al.*, 2005). A *L. delbrueckii* subesp. *bulgaricus* mutant with improved EPS production has been isolated, and it showed higher

the result of a higher ATP demand for biomass production in growing cells.

**5. Exopolysaccharides (EPS) production** 

production of lactate and ethanol. However, in the presence of fructose, a mannitol dehydrogenase activity (MDH) can account for NADH recycling with the concomitant production of mannitol. Homofermentative LAB, which use the glycolytic pathway for sugars fermentation and lack MDH, are also able to produce mannitol under special circumstances (Figure 2). Mutants of *L. plantarum* and *L. casei* impaired in NADH regeneration by the lactate dehydrogenase were able to produce small amounts of mannitol from glucose due to a mannitol-1-P dehydrogenase (M1PDH) activity on fructose-6P (Ferain *et al.,* 1996; Viana *et al.,* 2005). M1PDH can recycle NADH rendering mannitol-1P that can be dephosphorylated to mannitol and excreted from the bacterial cell. M1PDH activity is generally low, because its gene (*mtlD*) is only induced by the presence of mannitol. Furthermore, mannitol is also a common carbon and energy source that can be fermented. Therefore, subsequent re-uptake and metabolism of the produced mannitol should be avoided. In bacteria, mannitol is usually taken up by a mannitol-specific phospho*enol*pyruvate: sugar phosphotransferase system (PTSMtl) which catalizes the simultaneous mannitol uptake and phosphorylation to mannitol-1P. *L. lactis ldh* mutants have been constructed which were deleted in *mtlA* and *mtlF*, encoding the EIICBMtl and EIIAMtl components of the PTSMtl, respectively, involved in mannitol uptake (Gaspar *et al.*, 2004). This resulted in strains unable to utilize mannitol which converted 33% of the fermented glucose into mannitol. In another approach, the M1PDH encoding gene from *L. plantarum* and a gene encoding a mannitol-1P phosphatase from the protozoan parasite *Eimeria tenella* were overexpressed by using the NICE system in an *L. lactis ldh* mutant (Wisselink *et al.*, 2005). This strategy avoided the main bottleneck in mannitol production: most mannitol was accumulated inside the cell as mannitol-1P, which could reach concentrations up to 76 mM in high density non-growing cells of an *L. lactis ldh* mutant (Neves *et al.*, 2000). In this new strain 50% of the glucose was converted to mannitol (maximum theoretical yield of 67%). Other alternatives comprise the expression of MDH genes from heterofermenters. The *mdh* gene from *L. brevis* was expressed in a *L. plantarum* strain deficient in both *ldhL* and *ldhD* genes, and resulted in an increase in mannitol synthesis from glucose (Liu *et al.*, 2005).

The sorbitol (*gut*) operon of *L. casei* contained the genes *gutCBA*, encoding the EII component of the sorbitol-specific PTS involved in sorbitol transport and phosphorylation, two regulatory genes, *gutR* and *gutM*, and the gene *gutF*, encoding a sorbitol-6P dehydrogenase (S6PDH) (Alcantara *et al.*, 2008). A recombinant strain of *L. casei* with the *gutF* gene integrated in the chromosome at the lactose operon produces sorbitol from fructose-6P by reversing the sorbitol catabolic pathway (Nissen *et al.*, 2005) (Figure 2). Resting cells of this strain synthesized small amounts of sorbitol from glucose, with a conversion rate of 2.4 %. Subsequent inactivation of *ldh1* gene, encoding the main LDH (Rico *et al.*, 2008) promoted an increment in the conversion rate (4.3 %), suggesting that the engineered route provides an alternative pathway for NAD+ regeneration. Once glucose was depleted, reutilization of the produced sorbitol by *L. casei* recombinant strains was avoided by deleting *gutB* gene that encodes the IIBC component of the PTSGut (De Boeck *et al.*, 2010). *L. casei* recombinant strains produced mannitol in addition to sorbitol and this polyol mixture was avoided by inactivation of the *mtlD* gene that encodes a M1PDH. The engineered *L. casei* strain (*lac::gutF ldh1 gutB mtlD*) produced sorbitol from lactose, the milk sugar, in non-growing cells or in growing cells under pH control. Fed-batch fermentations using whey permeate, a waste product from the dairy industry with high concentration of lactose, resulted in a conversion rate of 9.4% of lactose into sorbitol (De Boeck *et al.*, 2010). *L. plantarum* has also been metabolically engineered to produce sorbitol by constitutive overexpression of either *srlD1* or *srlD2* genes that encode S6PDH activities in a mutant strain deficient in LDH activity (Ladero *et al.*, 2007). Using non-growing or growing cells under pH control resulted in a very efficiency conversion rate of about 65% and 25%, respectively, of sugar into sorbitol. The different efficiencies were suggested to be the result of a higher ATP demand for biomass production in growing cells.

### **5. Exopolysaccharides (EPS) production**

152 Food Industrial Processes – Methods and Equipment

production of lactate and ethanol. However, in the presence of fructose, a mannitol dehydrogenase activity (MDH) can account for NADH recycling with the concomitant production of mannitol. Homofermentative LAB, which use the glycolytic pathway for sugars fermentation and lack MDH, are also able to produce mannitol under special circumstances (Figure 2). Mutants of *L. plantarum* and *L. casei* impaired in NADH regeneration by the lactate dehydrogenase were able to produce small amounts of mannitol from glucose due to a mannitol-1-P dehydrogenase (M1PDH) activity on fructose-6P (Ferain *et al.,* 1996; Viana *et al.,* 2005). M1PDH can recycle NADH rendering mannitol-1P that can be dephosphorylated to mannitol and excreted from the bacterial cell. M1PDH activity is generally low, because its gene (*mtlD*) is only induced by the presence of mannitol. Furthermore, mannitol is also a common carbon and energy source that can be fermented. Therefore, subsequent re-uptake and metabolism of the produced mannitol should be avoided. In bacteria, mannitol is usually taken up by a mannitol-specific phospho*enol*pyruvate: sugar phosphotransferase system (PTSMtl) which catalizes the simultaneous mannitol uptake and phosphorylation to mannitol-1P. *L. lactis ldh* mutants have been constructed which were deleted in *mtlA* and *mtlF*, encoding the EIICBMtl and EIIAMtl components of the PTSMtl, respectively, involved in mannitol uptake (Gaspar *et al.*, 2004). This resulted in strains unable to utilize mannitol which converted 33% of the fermented glucose into mannitol. In another approach, the M1PDH encoding gene from *L. plantarum* and a gene encoding a mannitol-1P phosphatase from the protozoan parasite *Eimeria tenella* were overexpressed by using the NICE system in an *L. lactis ldh* mutant (Wisselink *et al.*, 2005). This strategy avoided the main bottleneck in mannitol production: most mannitol was accumulated inside the cell as mannitol-1P, which could reach concentrations up to 76 mM in high density non-growing cells of an *L. lactis ldh* mutant (Neves *et al.*, 2000). In this new strain 50% of the glucose was converted to mannitol (maximum theoretical yield of 67%). Other alternatives comprise the expression of MDH genes from heterofermenters. The *mdh* gene from *L. brevis* was expressed in a *L. plantarum* strain deficient in both *ldhL* and *ldhD* genes, and resulted in an increase in mannitol

The sorbitol (*gut*) operon of *L. casei* contained the genes *gutCBA*, encoding the EII component of the sorbitol-specific PTS involved in sorbitol transport and phosphorylation, two regulatory genes, *gutR* and *gutM*, and the gene *gutF*, encoding a sorbitol-6P dehydrogenase (S6PDH) (Alcantara *et al.*, 2008). A recombinant strain of *L. casei* with the *gutF* gene integrated in the chromosome at the lactose operon produces sorbitol from fructose-6P by reversing the sorbitol catabolic pathway (Nissen *et al.*, 2005) (Figure 2). Resting cells of this strain synthesized small amounts of sorbitol from glucose, with a conversion rate of 2.4 %. Subsequent inactivation of *ldh1* gene, encoding the main LDH (Rico *et al.*, 2008) promoted an increment in the conversion rate (4.3 %), suggesting that the engineered route provides an alternative pathway for NAD+ regeneration. Once glucose was depleted, reutilization of the produced sorbitol by *L. casei* recombinant strains was avoided by deleting *gutB* gene that encodes the IIBC component of the PTSGut (De Boeck *et al.*, 2010). *L. casei* recombinant strains produced mannitol in addition to sorbitol and this polyol mixture was avoided by inactivation of the *mtlD* gene that encodes a M1PDH. The engineered *L. casei* strain (*lac::gutF ldh1 gutB mtlD*) produced sorbitol from lactose, the milk sugar, in non-growing cells or in growing cells under pH control. Fed-batch

synthesis from glucose (Liu *et al.*, 2005).

Some LAB produced EPS, which are extracellular polysaccharides, with important characteristics for the dairy industry. They are used to improve the rheological and textural properties of fermented foods. EPS have also potential as food additives and functional food ingredients. In this sense they are claimed to act as prebiotics in the intestine (Bello *et al.*, 2001) and to stimulate the immune system (Vinderola *et al.*, 2006). The synthesis of EPS in LAB starts at the glycolytic intermediate glucose-6P, which connects the anabolic pathways of biosynthesis of sugar nucleotides, the precursors of the EPS, and the catabolic pathways for obtaining energy through the glycolysis (Figure 1). Glucose-6P is converted to glucose-1P by the -phospoglucomutase (-Pgm) activity, and this sugar phosphate is further metabolized to UDP-glucose and UDP-galactose by the consecutively action of enzymes UDP-glucose pyrophosphorylase (GalU) and UDP-galactose 4-epimerase (GalE) (Boels *et al.*, 2001). Glucose-1P is also substrate for dTDP- glucose pyrophosphorylase to produce dTDPglucose, which will be further metabolized to dTDP-rhanmnose. Glucose, galactose and rhamnose are the principal sugars found in the EPS produced by LAB. The subsequent steps in the synthesis of EPS is the assembly of the monosaccharide repeating unit by specific glycosyltransferases, the polymerization of the repeating units and the secretion from the cell (Welman *et al.*, 2006). The enzymes that participate in these stages are encoded by genes that form part of *eps* gene clusters. Genetic engineering strategies could be applied to one or more of those stages involved in the EPS biosynthesis in order to increase the EPS production or to modify its composition, however, until now only strains of *L. lactis* and *Streptococcus thermophilus* have been genetically modified to enhance EPS biosynthesis. In *S. thermophilus* the modification of the levels of the GalU, PgmA and the Leloir route enzymes resulted in increased levels of EPS (Levander *et al.*, 2002). Homologous overexpression of a complete *eps* operon in *L. lactis* resulted in about fourfold increase in EPS production (Boels *et al.*, 2003). In *Lactobacillus* species there are no examples of metabolic engineering strategies aimed to produce EPS. In this species the synthesis of EPS has been improved by modifying the culture conditions, such as carbon source and pH. As well, chemically induced mutants of *Lactobacillus* species that produce higher amounts of EPS than the parental strain have been isolated. The synthesis of EPS by *L. casei* strain CRL 87 was improved by using galactose as carbon source at a controlled pH of 5.0, and the high EPS production was correlated with high activity level of the enzymes involve in the synthesis of UDP-sugars (Mozzi *et al.*, 2003). Similar approaches were applied for *L. helveticus* strain ATCC 15807, which produces a higher amount of EPS from lactose at pH 4.5 than at pH 6.2, which was correlated with higher levels of -Pgm activity (Torino *et al.*, 2005). A *L. delbrueckii* subesp. *bulgaricus* mutant with improved EPS production has been isolated, and it showed higher

but is has been demonstrated that some Slp proteins display binding capacity to ECM proteins and possess immunoregulatory capabilities. In general the adhesive properties of the above extracellular proteins did not show a strict specificity for substrate binding and it is postulated that they may possess lectin-like characteristics that allow their binding to

149

Species of lactobacilli from intestinal origin (*L. plantarum*, *L. reuteri*, *Lactobacillus gasseri, L. acidophilus, L. johnsonii*) express surface proteins covalently anchored to the cell wall by a sortase-dependent mechanism with mucus binding ability (Msa and Mub proteins) (Boekhorst *et al.*, 2006). They are large multidomain proteins containing up to fifteen tandem copies of a mucin-binding domain (MucBP) and act as mannose-dependent adhesins with the capacity to aggregate *Saccharomyces cerevisiae* cells. The presence of these proteins in certain *L. reuteri* isolates correlates with the binding ability to mucus. Similarly, the presence in *L. plantarum* of *msa* genes is the sole requisite for mucus binding in this species, but domain composition and subtle amino acids changes in each specific Msa protein account for the diverse adhesion properties reported in different strains. In *L. rhamnosus* GG (LGG), a well characterized probiotic with established mucus adhesion properties, the product of the LGG\_02337 gene is the only protein encoded in the genome which contains four MucBP domains about 50 amino acids shorter in length compared to the large mucus binding proteins Msa and Mub. This protein is anchored at the bacterial surface and possesses *in* 

A genome analysis of several *L. rhamnosus* strains identified the *spaCBA* cluster as another trait responsible for mucus adherence in LGG. This cluster codes for the three components of pili structures similar to pili described in Gram-positive pathogens that can be identified by immune electron microscopy at the LGG surface (Kankainen *et al.*, 2009). SpaA is the major pilin protein that forms the pilus shaft, while SpaC and SpaB are ancillary pilus proteins which are present at the pilus tip or along the pilus structure and possess adhesive properties. Adhesion experiments with purified proteins, specific antibodies, and mutant construction have demonstrated that SpaC and SpaB are responsible for the mucus binding activity displayed by LGG. This is the first example of the presence of pili adhesive structures in a probiotic strain and exemplifies the adaptation of these bacteria to persist in

Some of the adhesion factors characterized in probiotic bacteria may be targets for strain engineering aimed to enhance bacterial adhesion. In addition, heterologous expression of well characterized adhesins from different sources can be envisaged. This can be useful to increase residence times at the gastrointestinal tract, enhance interactions with the mucosal immune system and promote competitive exclusion of pathogens by probiotics. Some probiotic strains like *L. casei* Shirota have been engineered to express a fibronectin binding domain from the Sfb protein of *Streptococcus pyogenes*, allowing this strain, which barely binds fibronectin, to bind this ECM substrate, fibrinogen and human fibroblasts (Kushiro *et al.*, 2001). However, to date most genetic engineering strategies aimed to increase lactic acid bacteria adhesion have been carried out in the model *L. lactis* species. This bacterium is not a

highly glycosylated ECM proteins and mucosal surfaces (Velez *et al.*, 2007).

**6.1.1 Specific proteins implicated in mucus binding** 

*vitro* binding activity to mucus (von Ossowski *et al.*, 2011).

**6.1.2 Engineered lactic acid bacteria with enhanced adhesion** 

host tissues.

amounts of GalU activity, glucose-6P and ATP than the parent strain. These characteristics suggest that GalU and -Pgm enzymes play important roles in the synthesis of high EPS production. The elevated concentration of ATP in the mutant indicated that the glycolysis influence the anabolic route of EPS biosynthesis (Welman *et al.*, 2006). A metabolic engineering strategy aimed to direct the carbon flux towards UDP-glucose and UDPgalactose biosynthesis was successfully applied in *L. casei*. The *galU* gene coding for GalU enzyme in *L. casei* strain BL23 was cloned under control of the inducible *nisA* promoter, and the resulting strain showed about an 80-fold increase in GalU activity, a 9-fold increase of UDP-glucose and a 4-fold increase of UDP-galactose (Rodriguez-Diaz & Yebra, 2011). *L. casei* strain BL23 does not produce EPS, hence it would be an adequate host for the production of heterologous EPS.

#### **6. Improvement of probiotic properties**

#### **6.1 Adhesion to intestinal epithelial cells**

Adhesion of probiotic bacteria has been employed as a criterion for strain selection and, although it is not indispensable for some probiotic traits, it has positive effects on strain persistence at the gastrointestinal tract and in pathogen inhibition by displacement and competition for adhesion sites. Also, it has been suggested that the capacity to adhere to mucosal surfaces influences the cross-talk established between probiotic bacteria and host cells (Sanchez *et al.*, 2008; Velez *et al.*, 2007).

Probiotic strains have shown the ability to bind to intestinal epithelial cultured cells (e.g. Caco-2, HT-29), to mucus components and to proteins of the extracellular matrix (ECM) such as collagen, fibronectin or laminin. Although theses late molecules are not commonly found at the mucosal surface, they may be shed into the mucus or may be exposed in case of trauma or inflammation. They are common targets for pathogen adhesion during the process of infection and adhesion to them by probiotic bacteria can compete with pathogen binding. In contrast to the knowledge about adhesive factors in bacteria causing infectious diseases in humans and animals, the knowledge about adhesion mechanisms in probiotics is very limited. Some molecules from probiotics have been identified as responsible for adhesion, including lipoteichoic acid and exopolysaccharides. However, surface proteins are the major responsible for adherence. Typical surface adhesins from pathogenic bacteria with binding capacity to cultured cells and ECM components are not found in probiotics although it is hypothesized that they may share similar mechanisms for attachment. Similar to some pathogens, probiotic lactobacilli display on their surface *moonlighting* proteins which are in most cases of cytoplasmic location and are exported and retained on bacterial surfaces by yet unknown mechanisms. These include glycolytic enzymes such as enolase and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), molecular chaperones (DnaK, GroEL) and translational elongation factors (EF-Tu, EF-G) (Sanchez *et al.*, 2008; Velez *et al.*, 2007). These proteins have demonstrated binding ability to ECM proteins and epithelial cells and in some cases interfere with host pathways (plasminogen activation mediated by enolase from *Lactobacillus crispatus* and *Lactobacillus johnsonii* (Antikainen *et al.*, 2007) or immunoregulation by GroEL from *L. johnsonii* (Bergonzelli *et al.*, 2006). Other surface proteins with identified roles in adhesion are surface layer (S-layer) proteins or Slp. S-layers from lactobacilli such as *L. acidophilus* or *L. brevis* are formed by small basic proteins which form a crystalline matrix on the bacterial surface. The real function of the S-layer is uncertain but is has been demonstrated that some Slp proteins display binding capacity to ECM proteins and possess immunoregulatory capabilities. In general the adhesive properties of the above extracellular proteins did not show a strict specificity for substrate binding and it is postulated that they may possess lectin-like characteristics that allow their binding to highly glycosylated ECM proteins and mucosal surfaces (Velez *et al.*, 2007).

### **6.1.1 Specific proteins implicated in mucus binding**

154 Food Industrial Processes – Methods and Equipment

amounts of GalU activity, glucose-6P and ATP than the parent strain. These characteristics suggest that GalU and -Pgm enzymes play important roles in the synthesis of high EPS production. The elevated concentration of ATP in the mutant indicated that the glycolysis influence the anabolic route of EPS biosynthesis (Welman *et al.*, 2006). A metabolic engineering strategy aimed to direct the carbon flux towards UDP-glucose and UDPgalactose biosynthesis was successfully applied in *L. casei*. The *galU* gene coding for GalU enzyme in *L. casei* strain BL23 was cloned under control of the inducible *nisA* promoter, and the resulting strain showed about an 80-fold increase in GalU activity, a 9-fold increase of UDP-glucose and a 4-fold increase of UDP-galactose (Rodriguez-Diaz & Yebra, 2011). *L. casei* strain BL23 does not produce EPS, hence it would be an adequate host for the

Adhesion of probiotic bacteria has been employed as a criterion for strain selection and, although it is not indispensable for some probiotic traits, it has positive effects on strain persistence at the gastrointestinal tract and in pathogen inhibition by displacement and competition for adhesion sites. Also, it has been suggested that the capacity to adhere to mucosal surfaces influences the cross-talk established between probiotic bacteria and host

Probiotic strains have shown the ability to bind to intestinal epithelial cultured cells (e.g. Caco-2, HT-29), to mucus components and to proteins of the extracellular matrix (ECM) such as collagen, fibronectin or laminin. Although theses late molecules are not commonly found at the mucosal surface, they may be shed into the mucus or may be exposed in case of trauma or inflammation. They are common targets for pathogen adhesion during the process of infection and adhesion to them by probiotic bacteria can compete with pathogen binding. In contrast to the knowledge about adhesive factors in bacteria causing infectious diseases in humans and animals, the knowledge about adhesion mechanisms in probiotics is very limited. Some molecules from probiotics have been identified as responsible for adhesion, including lipoteichoic acid and exopolysaccharides. However, surface proteins are the major responsible for adherence. Typical surface adhesins from pathogenic bacteria with binding capacity to cultured cells and ECM components are not found in probiotics although it is hypothesized that they may share similar mechanisms for attachment. Similar to some pathogens, probiotic lactobacilli display on their surface *moonlighting* proteins which are in most cases of cytoplasmic location and are exported and retained on bacterial surfaces by yet unknown mechanisms. These include glycolytic enzymes such as enolase and glyceraldehyde 3-phosphate dehydrogenase (GAPDH), molecular chaperones (DnaK, GroEL) and translational elongation factors (EF-Tu, EF-G) (Sanchez *et al.*, 2008; Velez *et al.*, 2007). These proteins have demonstrated binding ability to ECM proteins and epithelial cells and in some cases interfere with host pathways (plasminogen activation mediated by enolase from *Lactobacillus crispatus* and *Lactobacillus johnsonii* (Antikainen *et al.*, 2007) or immunoregulation by GroEL from *L. johnsonii* (Bergonzelli *et al.*, 2006). Other surface proteins with identified roles in adhesion are surface layer (S-layer) proteins or Slp. S-layers from lactobacilli such as *L. acidophilus* or *L. brevis* are formed by small basic proteins which form a crystalline matrix on the bacterial surface. The real function of the S-layer is uncertain

production of heterologous EPS.

**6. Improvement of probiotic properties 6.1 Adhesion to intestinal epithelial cells** 

cells (Sanchez *et al.*, 2008; Velez *et al.*, 2007).

Species of lactobacilli from intestinal origin (*L. plantarum*, *L. reuteri*, *Lactobacillus gasseri, L. acidophilus, L. johnsonii*) express surface proteins covalently anchored to the cell wall by a sortase-dependent mechanism with mucus binding ability (Msa and Mub proteins) (Boekhorst *et al.*, 2006). They are large multidomain proteins containing up to fifteen tandem copies of a mucin-binding domain (MucBP) and act as mannose-dependent adhesins with the capacity to aggregate *Saccharomyces cerevisiae* cells. The presence of these proteins in certain *L. reuteri* isolates correlates with the binding ability to mucus. Similarly, the presence in *L. plantarum* of *msa* genes is the sole requisite for mucus binding in this species, but domain composition and subtle amino acids changes in each specific Msa protein account for the diverse adhesion properties reported in different strains. In *L. rhamnosus* GG (LGG), a well characterized probiotic with established mucus adhesion properties, the product of the LGG\_02337 gene is the only protein encoded in the genome which contains four MucBP domains about 50 amino acids shorter in length compared to the large mucus binding proteins Msa and Mub. This protein is anchored at the bacterial surface and possesses *in vitro* binding activity to mucus (von Ossowski *et al.*, 2011).

A genome analysis of several *L. rhamnosus* strains identified the *spaCBA* cluster as another trait responsible for mucus adherence in LGG. This cluster codes for the three components of pili structures similar to pili described in Gram-positive pathogens that can be identified by immune electron microscopy at the LGG surface (Kankainen *et al.*, 2009). SpaA is the major pilin protein that forms the pilus shaft, while SpaC and SpaB are ancillary pilus proteins which are present at the pilus tip or along the pilus structure and possess adhesive properties. Adhesion experiments with purified proteins, specific antibodies, and mutant construction have demonstrated that SpaC and SpaB are responsible for the mucus binding activity displayed by LGG. This is the first example of the presence of pili adhesive structures in a probiotic strain and exemplifies the adaptation of these bacteria to persist in host tissues.
