Different types of commercially available yeast cell wall and autolysed yeast products were used. Due to a confidentiality agreement, the trade names and manufacturers have been omitted.
Autolysis and French press of cell wall fragments
Cell wall fragments were produced in the laboratory scale under standard conditions from the following yeast strains: Saccharomyces cerevisiae Bio 19 (DSMZ 1848), Klyveromyces marxianus Bio 21 (DSMZ 5420), Saccharomyces boulardii CAN 214, Saccharomyces boulardii HA 282, Saccharomyces boulardii HA 283, Trichosporon mycotoxinivorans Bio 328 (DSMZ 14153), Aureobasidium pullulans CF 10 (DSMZ 14940), Aureobasidium pullulans CF 40 (DSMZ 14941). Bio, CAN and HA belong to the BIOMIN Research Center culture collection.
Cell wall fragments were obtained by incubating the respective yeast strains in 1000 ml shaking flasks containing a 400 ml medium consisting of 1% soy peptone, 0.1% yeast extract, 0.2% magnesium sulfate heptahydrate, 0.2% ammonium sulfate, 1% saccharose and 2.5% glucose for 48 hours at 30°C and pH 5.0. The yeast cells were subsequently autolysed by incubating at pH 5 at a temperature of 50°C for 24 hours and centrifuged (17700 × g, 10 min) to separate cell wall fragments from the supernatant. Finally, the cell wall fragments were freeze dried (24 hours, -30°C/+30°C, 0.001 mbar) (Epsilon 2–90, Christ, Osterode, Germany). Autolysis was adapted according to Liu et al. (2008). The degree of autolysis was assessed by analysis of the supernatant fraction: cell rupture leads to increasing amounts of proteins and carbohydrates and dry matter in the supernatant. These parameters reached a maximum value after 24 hours. The progress of autolysis was also monitored by observing cells under the microscope (magnified 1000 times) and comparing them to cells from the original culture.
French press: Rupture took place in the French press with 1 cycle at 20,000 psi (=1,378.96 bar). Microscopic images are taken of the ruptured cells. Liquid supernatant was removed from the non-soluble cell walls by centrifugation (17700 × g, 10 min).
The quantitative adhesion of E. coli F4 (Bio 104 from BIOMIN Research Center culture collection) and Salmonella Typhimurium (Bio 99 from BIOMIN Research Center culture collection) on yeast products was determined by measuring the optical density of the culture solutions, which indicates the growth of the adhering bacteria. The growth rate depended on the number of adhering bacteria – with higher numbers bound resulting in faster growth and thus the bacteria entering earlier into the exponential phase.
The yeast derivatives were suspended in phosphate buffered saline (PBS) in a concentration of 0.1 mg/L, the wells of a 96-well plate were coated with a 100 μL yeast suspension per well and incubated for 16–18 hours at 4°C. The plate was washed three times with PBS Tween20 using a microplate washer (Tecan 5082 M8/4R Columbus Plus). Subsequently, the bacterial suspension of E. coli F4 or Salmonella Typhimurium grown in Tryptic Soy Broth (TSB) overnight was adjusted to OD 0.01 and added to the wells of the microplate. Bacteria were allowed to adhere for 60 minutes at 37°C and the plate was subsequently washed 6 times with a microplate washer/PBS Tween20 to remove non-adherent bacteria. The wells were filled with 200 μL of TSB and overlaid with one drop of paraffin- oil. Blank, growth controls, negative control and binding control were also assessed.
The microplate was placed in a microplate reader (Tecan Genious), incubated for 18 hours at 37°C and OD was measured at a wavelength of 690 nm. Data were recorded every 15 minutes.
Serial dilutions of E. coli or Salmonella were prepared and applied in parallel on the microplate and the agar plate in order to establish a linear regression between the number of counted bacteria on the agar plate (CFU/mL) and the time in hours when the adhering bacterial solution reached the exponential phase (optical density of 0.1 at 690 nm). Independent calibration curves were created for the genera E. coli and Salmonella. Subsequently, the number of bacteria bound to the yeast product was determined using linear regression.
The binding control consisted of Bovine Serum Albumin (BSA) and test bacteria but without adding yeast to determine the non-specific binding sites. The binding control had no growth until after 6 hours. Finally the binding control, the bacterial number, which adhered to the non-specific binding sites of the well, was subtracted from the calculated data to determine the final number of specifically adhering bacteria. The blank consisted of the yeast cell wall product and BSA but without adding test bacteria and included all the assay steps to determine the bacterial number of the yeast cell wall product. The blank had no growth until after 10 h. The media control (TSB) had no growth until after 18 h.
The calibration is valid only for the respective genera and at an optical density of 0.1 (Ganner et al. 2010).
Carbohydrate and protein analysis
For total mannan and glucan determination, yeast cell wall polysaccharides were hydrolyzed with 72% sulfuric acid. Carrez-precipitation was subsequently performed to remove interfering proteins and fats by precipitation as zinc (2+) and/or cyanoferrat (II) - complexes (Matissek et al. 1992). The cleared supernatants were analyzed for glucose and mannose by HPLC-RID using an ICSep ION-300 column (Transgenomic).
Total protein was determined by Total Kjeldahl Nitrogen (TKN, AOAC Official Method 954.01, 1954).
Carbohydrate analyses were conducted in single, protein analyses were conducted in duplicates.
Experimental data were analyzed using IBM SPSS, version 19. Due to non-normal distribution of data, Spearman correlation was used. The calculated coefficient of correlation is only relevant for a given significance value (P-value <0.05); the coefficients are classified as below. A positive Spearman correlation coefficient corresponds to an increasing monotonic trend between X and Y. A negative Spearman correlation coefficient corresponds to a decreasing monotonic trend between X and Y. The influence of treatment (autolysis or French press) was calculated with the non-parametric Mann Whitney U Test (data non-normal distributed).
Coefficient of correlation classification
Correlations were considered very low (r ≤ 0.2), low (0.2 < r ≤ 0.5), medium (0.5 < r ≤ 0.7), high (0.7 < r ≤ 0.9), very high (0.9 < r ≤ 1).