Quantitative evaluation of E. coli F4 and Salmonella Typhimurium binding capacity of yeast derivatives

Yeast derivatives such as yeast cell wall products and recently also yeast autolysate products are sold worldwide as feed supplements, some of which claim to bind enteropathogenic bacteria such as E. coli and Salmonella spp.. Although yeast cell wall products have been in the market for a long time, their binding capacity and the mechanism of pathogen binding is still not well understood. There is much confusion throughout the animal feed industry concerning the quality of yeast products. An important initial event in bacterial pathogenesis is the adherence of bacteria via their surface lectins to host intestinal cells. Infections are initiated only after the microorganism has first adhered to the host cell surface. If this initial adherence can be inhibited, so can the subsequent infection. This approach forms the basis of anti-adherence strategies with the most studied being receptor-analogs, which include oligosaccharides (Shoaf-Sweeney and Hutkins 2008).

According to the literature, mannan-oligosaccharides (MOS) are the substances in the yeast cell wall responsible for the binding of pathogens via mannose specific type-I fimbriae (e.g. Snellings et al. 1997; Sharon and Ofek 2000). However, questions have been raised of late as to whether the mannan quantity in the yeast cell wall is responsible for immobilizing enteric pathogens. A point made so far was the lack of quantitative analytical methods to determine the total amount of bacteria bound by the yeast cell wall (Applegate et al. 2010). Some qualitative test procedures such as the agglutination method are available to investigate the binding capacity of yeast products (e.g. Mirelman et al. 1980; Eshdat et al. 1981; Oyofo et al. 1989; Spring et al. 2000; Peng et al. 2001). The agglutination method mixes bacteria with a cell wall suspension followed by visual evaluation of binding. Critical issue with this procedure is the poor solubility of the yeast cell wall in aqueous dilutions. Furthermore visual evaluation of binding is highly subjective because it is not possible to determine whether the bacteria are actually bound or whether they are merely in proximity to the yeast. Thus reproducibility is not given. It is necessary to obtain the quantitative measurement of bacterial adhesion in order to evaluate the binding capacity and quality of yeast derivatives. The quantitative in vitro assay developed by Ganner et al. (2010) was used in the present study to investigate different yeast cell wall products (MOS) and yeast autolysates for their capacity to bind E. coli F4 and Salmonella Typhimurium. This method is based on a microtiter plate, the insolubility of the cell wall material, with its high molecular weight, can be used to coat the wells. The wells are allowed to be incubated with the test bacteria, unbound bacteria are rinsed out, and the growth rate of bound bacteria is subsequently quantified by optical density reading. The accuracy and reproducibility of binding is given by automated evaluation of optical density. However, in vitro assays may be criticized by product end users as not accurately reflecting in vivo responses. This may be true in some cases, but in others it may partially be a reflection of assay bias attributable to media selection, duration, or substrate concentration. Another limitation of this assay is that the binding mechanisms involved cannot be elucidated, for this purpose, studies on a molecular level and detailed characterization of cell wall components have to be examined in future research studies.

E. coli F4 and Salmonella Typhimurium were chosen for the present study as they are important pathogens in the animal sciences and are known to cause severe economic losses in livestock production. Furthermore, the mannan-, glucan- and protein content of the yeast products were determined and statistical evaluation for correlation was performed with IBM SPSS v19.

To collect more information on binding characteristics, a separate experiment was performed in addition to evaluating mannan- and glucan- content; eight different yeast strains were fermented in the laboratory, and subjected to autolysis and French press treatment, to investigate the potential influence of yeast strains and production methods.

The Mann Whitney U Test demonstrated an influence of the yeast autolysis on E. coli binding in contrast to French press treatment (P < 0.001). For Salmonella binding, no influence was noted due to autolysis or French press treatment (P = 0.27).

Yeast cell wall fractions and yeast autolysate products have been proposed to bind enteropathogenic bacteria and therefore to possess prophylactic properties in the gut for the control of selective pathogenic bacteria such as E. coli and Salmonella. There is much in the literature describing this adherence mechanism and mode(s) of action of yeast derivatives (Mourao et al. 2006; Oyofo et al. 1989; Sharon and Ofek 2000; Shoaf-Sweeney and Hutkins 2008; Snellings et al. 1997; Spring et al. 2000; Terre et al. 2007; Yang et al. 2008). However, no study has ever investigated the quantitative number of pathogens bound by different yeast derivatives. The influences of yeast strains and production treatments (autolysis or French press) have also not been investigated before in combination with pathogen binding. We consider this an important quality parameter for yeast cell wall products. Indeed, there was a lack of quantitative analytical methods to determine the total amount of bacteria bound by the yeast cell wall (Applegate et al. 2010). Yeast cell wall products are believed to act as anti-adhesive agents and are thus able to prevent intestinal bacterial attachment by providing alternative adhesion sites to enterobacteria which contain mannose-specific type I fimbriae such as E. coli and Salmonella spp. (Ofek et al. 1977). Shoaf-Sweeney and Hutkins (2008) defined the adhesion affinity by the valency of protein-oligosaccharide interactions. A single molecule would have a low affinity but the creation of individual protein subunits with numerous oligosaccharide binding sites assembled to a filamentous structure leads to increased target avidity. The present study demonstrates that mannan content is an important parameter for E. coli binding (r = 0.315, low Spearman correlation), and glucan content has a high correlation with Salmonella Typhimurium binding affinity (r = 0.754). According to the current results and further evaluating the available literature, we propose that there are many more unknown factors (apart from mannan and glucan content) influencing binding affinity of these pathogens.

In addition to the total amount of mannan and glucan, the three-dimensional structure of mannose chains within the yeast cell wall is another critical factor to consider. We observed that mannose chains have a strong influence on the binding behavior of different yeast cell wall products and are important for pathogen immobilization. This is supported by the observation that in contrast to yeast cell wall products (mannan content between 10-25%), a purified mannan standard (99%) showed no pathogen binding activity (Ganner et al. 2007, unpublished data). Indeed, quantitative, rather than qualitative assays are essential in order to evaluate the pathogen binding capacity of cell walls from different yeast sources and strains (Ganner et al. 2008). Ofek et al. (2003) explained that apart from adhesion-receptor interaction, hydrophobic and other non-specific interactions might be involved in the adhesion process.

Becker and Galetti (2008) investigated various food and feed components for alternative adhesion of intestinal E. coli and Salmonella. Yeast mannan-oligosaccharides, pumpkin, sesame seed extract, palm kernel extract and konjac gum were among the strongest binding matrices for enteropathogens. Adhesion matrices differ tremendously which gives rise to various interpretations and speculations on substantial binding factors while suggesting that unspecific and unknown binding properties are involved. Moreover, taking into account the diversity of natural products and biological variation therein, the binding capacity of fibrous food and feed components for different bacteria seems rather unpredictable (Becker and Galetti, 2008).

Our observation that mannan content is strain dependent (Table 4) agrees with other studies (Jaehrig et al. 2008; Klis et al. 2002; Molnar et al. 2004). However, in our experiment with different yeast strains fermented under standardized laboratory conditions, we could not obtain any statistically significant correlations between mannan or glucan content and pathogen binding. According to these results, the yeast strain itself has no influence on E. coli F4 or Salmonella Typhimurium binding. However, we could demonstrate that the downstream process autolysis has an influence on E. coli binding (P < 0.001). Giovani et al. (2010) studied the effect of yeast strain and fermentation conditions on the release of cell wall polysaccharides. Temperature, media and sugar composition influenced the release of cell wall polysaccharides significantly. These data underline and support our results on the relative influence of production processes and production conditions on binding capacity.

The binding mechanism between pathogens and the yeast cell wall can be speculated, as demonstrated by the autolysis production process, the mannan content and the glucan content. However, as these binding mechanisms are only partially understood, we propose that other factors such as the cell wall structure, strain diversity, structural surroundings-, and non-specific interactions might also play important roles in pathogen immobilization, in addition to the total amount of mannan or glucan.

Summarizing, the present study demonstrates a variation among different yeast products concerning their capability to bind E. coli F4 and Salmonella Typhimurium. Salmonella Typhimurium was bound one to two orders of magnitude more than E. coli F4. We could demonstrate a statistically significant correlation between mannan and glucan content and pathogen binding. Furthermore we could demonstrate a relative influence of the production process autolysis.

However, the binding mechanisms involved are still not fully understood and our results underline the complexity of the present subject. The recently suggested mannose specific type-I fimbriae play a role and furthermore, unknown mechanisms which might interact synergistically are probably involved. Biological variation and natural diversity make predictions even more difficult and comprehensive. Further research is necessary, e.g., on a molecular level and a detailed characterization of cell wall components and cell wall structure, to elucidate on the additional forces which might be involved in the binding mechanism between yeast derivatives and enteropathogenic bacteria. Finally, additional standardized in vivo experiments with challenged animals could also further validate the binding potential of yeast products and the study of the anti-adherence strategy.