Introduction

Yeast extracts (YE) are the water soluble portion of autolysed yeast cells. The beneficial effect of YE on the growth of lactic acid bacteria (LAB) is well established (Smith et al. 1975; Aeschlimann and von Stockar 1990). YE are an excellent source of the B-complex vitamins and are often used to supply these factors in bacteriological culture media (Difco 1984). They are also a reliable source of pep­tides and amino acids (Peppler 1982).

Amino acids and peptides stimulate the growth of LAB (van Boven and Konings 1986; Hemme et al. 1981; Juillard et al. 1995) and they have been shown to be partially re­sponsible for YE stimulation of the growth of lactic cultures (Benthin and Villadsen 1996). Peptides can be superior to free amino acids as a nitrogen sources for LAB because they

can provide the cell with amino acids in a form that can be utilized more efficiently (van Boven and Konings 1986).

Some commercial applications, microbiological media for example, require that YE produce clear solutions. One way of obtaining clear solutions is by filtering the YE. There is little information of the effect of filtration on the growth­promoting abilities of YE. Desmazeaud and Hermier (1972) separated peptide fractions by gel filtration of digested ca­sein and found that peptides with molecular weights in the range 1–2.5 kDa stimulated growth of Streptococcus thermophilus. Fractions of casein hydrolysates obtained by ultrafiltration (UF) have been used to supplement culture medium and to standardize milk cultures (St-Gelais et al. 1993). These results suggest that UF fractions of YE may have different growth-promoting abilities, but no informa­tion is available on the effect of UF on the biological value of YE with respect to LAB.

The aim of this study was to examine if UF of YE influ­ences their growth-promoting properties in various lactic acid cultures, and to provide a comprehensive view of the effect of ultrafiltration on the biological value of YE on lac­tic cultures. However, using only one strain and one YE product would not have enabled this. The literature shows that there is variability in both YE (Potvin et al. 1997) and lactic culture responses to growth-promoting compounds (St-Gelais et al. 1993; Valdez et al. 1985a). Therefore, in or­der to provide a comprehensive picture, five YE were stud­ied with nine strains of lactic cultures. Furthermore, two lots of each YE were used in order to reduce the effect of the YE lot.

Materials and methods

Commercial yeast extracts

Five commercial YE were obtained, and their suppliers are listed in Table 1. Since variability between lots has been reported (Potvin et al. 1997), two lots of each source were used. With each lot, two assays were carried out, for a total of 4 separate assays. Three YE were from bakers’ yeasts (A, B, E) while two were from brewers’ yeasts (C, D).

Ultrafiltration of yeast extracts

YE were suspended in deionized water to obtain a solution con­taining 10% (w/v) solids. This solution was pre-filtered on an 8 µm Whatman (No. 2 filter paper) membrane. The filtrate was further processed by UF, using a tangential filtration system (Minitan Fil­ter plates, Millipore, Bedford, U.S.A.), with membranes having molecular cutoffs of 10, 3, or 1 kDa. The system enabled the use of four 30 cm² membranes for a total filtration surface of 120 cm². A 500 mL solution of 10% YE was used, and the UF process was stopped after the recovery of 300 mL of filtrate. The yeast extracts filtrates (YEF) were concentrated by lyophilization in a LyoTech (Lyo San Inc., Lachute, Canada) at 24°C for 72 h and stored at – 20°C until used. Total and a-amino nitrogen content, total solids, turbidity, and growth-promoting properties of each YE and YEF were evaluated. All chemical analysis was done in duplicate for each lot tested and the results presented are the means of the two values obtained.

Chemical analyses of YE and YEF

a -Amino nitrogen

The a-amino nitrogen of the fractions was determined by titra­tion following reaction with formaldehyde (USP 1985). The 5% YEF solutions were adjusted at pH 7.0 with 0.1 M NaOH or 0.1 M HCl.
 

Total nitrogen determination

The total nitrogen determination was done using a FP-428 LECO apparatus (LECO Corporation, Saint Joseph, Mich.), oper­ated under the following conditions: sample size, 150 mg; oxida­tion furnace temperature, 900°C; oxidation standby temperature, 650°C; purge cycles, 3; minimum timeout, 30 s; comparator level, 1.00; loop select low range, flow constants at high; gases, oxygen 99.99% and helium 99.99%. The calibration standard was com­posed of 150 mg EDTA (No. 502–092, 9.56 ± 0.03% nitrogen, LECO Corporation, Saint Joseph, Mich.)
 

Turbidity

Turbidity of YE and YEF solutions were determined with a Orbico-Hellige turbidimeter (Model 965; Farmingdale USA). Hydrazine sulfate standards (VWR; West Chester Pa., USA) were used to calibrate the turbidimeter.
 

Water content

Water content of the YE, filtrates, and YEF powders were ob­tained by dry weights after an incubation at 105°C for 16 h.

Organisms and growth conditions

Lactobacillus plantarum EQ27, Lactobacillus casei EQ28, EQ85, EQ70, Lactobacillus acidophilus EQ57, Pediococcus acidilactici MA1815-M, and Pediococcus pentosaceus EQ44 were kindly provided by Lallemand-Equipharm SA (Aurillac, France). Lactococcus lactis subsp. cremoris Wg2L, a proteinase negative (Prt^-) culture was graciously provided by D. St-Gelais (Food Re­search and Development Center (FRDC), St-Hyacinthe, Canada) while Lactococcus lactis subsp. cremoris R2 Prt⁺ culture, was from our own FRDC collection.

The Lactobacillus and Pediococcus strains were grown at 37°C in MRS broth (BDH, Darmstadt, Germany), until their pH reached 4.6, which required approximately 8 h. The cultures were put on ice for 30 min to stop the acidification process. The lactococci strains were obtained following a 16h/22°C incubation in reconsti­tuted skim milk (11% solids). The milk was sterilized at 110°C for 10 min.

Mother cultures were prepared by mixing 20 mL of cell suspen­sions, freshly grown on MRS or milk, with 50 mL of 20% skim milk and 50 mL of a 20% glycerol solution (glycerol and milk were sterilized separately). The milk/glycerol cell suspensions were divided into 1 mL fractions which were added to sterile cryovials (Nalgene, Rochester, N.Y.) and stored at –70°C until used.

Media

For the testing of the various YE and YEF in spectrophotomet­ric assays, the base medium used for the growth of the lactobacilli and pediococci had the following composition, per L of medium: Glucose, 20 g; K2HPO4, 3 g; KH2PO4, 3 g; sodium citrate, 3 g; Tween 80, 1 g; MgSO₄·7H₂O, 0.2 g; MnSO₄·2H₂O, 0.05 g; and YEF at the set concentrations for the different strains. A glucose and MgSO4·2H2O solution was first prepared with compounds hav­ing twice the concentration (40 g/L and 0.4 g/L respectively) and autoclaved 15 min at 121°C. The rest of the medium was also pre­pared at a double concentration and sterilized the same way. Both solutions were mixed following sterilization.

The media used for the assays with lactococci was the same as for lactobacilli and pediococci, except for glucose which was sub­stituted by lactose. In these assays, it was not required to prepare separate 2× solutions prior to sterilization. The same media were used to test various >10 kDa retentates.

Previous studies have shown that the appropriate concentration of YE in the basal medium, used to evaluate the biological value of various YE, must be defined for each strain (Potvin et al. 1997; Champagne et al. 1999). Preliminary studies were thus carried out using the method of Champagne et al. (1999) to determine the ap­propriate concentrations for the determination of maximum growth ( max) or maximum optical density _(ODmax). Thus, YE or YEF were added at 5 g/L for lactococci, at 1 g/L for the three L. casei strains and at 2 g/L for the other cultures.

Growth-promoting properties evaluation of YE and YEF

An automated turbidimetric instrument (Bioscreen C, Lab­systems, Corp., Helsinki, Finland) was used to assess the effect of YE or YEF added to the base medium on the increase in turbidity of the cultures. The method described by Champagne et al. (1999) was used.

Sterile media having various YE and YEF were prepared in test tubes, inoculated at 1% (v/v) from a fresh culture, and 350- L samples were added aseptically to each microwell. Microplates were incubated for 24 h at 37°C for the lactobacilli as well as the pediococci, and at 22°C for the lactococci. The absorbency at 600 nm of each microwell was automatically recorded every 15 min. Prior to reading, the plates were shaken at low intensity for 5 s. Two replicates for each fractions were evaluated on Bioscreen for both YE lots tested. Thus, four separate assays were carried out.

Statistical analyses

Statistical analyses were performed using SAS Institute, Inc. (SAS 1989, Cary, N.C.) software. Analysis of variance was used with significance defined at P < 0.05. The biomass values in the spectrophotometric assay system were transformed into Ln values of biomass and plotted against time for the exponential growth part of the curve (Sigma Plot software, Jandel). The slope of the curve, giving the _µmax, was obtained by simple linear regression. The term fermentation time is employed to represent the time corresponding to the end of the exponential phase of the growth curve of the strains.

Results and discussion

Nitrogen content of the YE or YEF

The total nitrogen (total N) and a-amino nitrogen ( - amino N) contents of the original and filtered YE powders are shown in Table 1. The variance analysis suggests that the effect of filtration was not significant with respect to nitro­gen contents. Paired T tests show, however, that the total N contents of the retentates and the YEF 10 kDa products were indeed significantly (P < 0.05) different from the original YE. As well, regression analyses show that the systematic increases in a-amino N in the YEF powders as filter pore­size decreased were real. The data shows nevertheless that the effect of the YE source was much greater than that of UF on total and a-amino N contents.

Brewers’ YE had sightly lower total N content than bak­ers’ YE. In the original YE, the proportion of a-amino N in total N was approximately 50%, which suggests that there was an important hydrolysis of protein and peptides during the autolysis process. The source of YE affected the content of a-amino N of YE (Table 1).

The a-amino N content of the YEF powders increased as the molecular weight cutoffs of the UF membranes, used to filter the YE solutions, decreased (Table 1). On average,

YEF powders obtained from the 1 kDa unit contained 30% more a-amino N than in the original YE. This was expected, as filtration removed the high molecular weight compounds and increased the proportional content of the remaining pep­tides and amino acids. The proportion of a-amino N to total N becomes more important as the pore size of the filtration membranes are reduced, which suggests that most of the ni­trogen is probably composed of peptides and amino acids. These results are in accordance with technical data given by Ohly (1998), which cites the protein-related fraction of stan­dard salt-free YE powder to contain 35–40% amino acids, 40–60% short peptides and 2–5% oligopeptides/proteins having MW greater than 3 kDa.

Table 1. Nitrogen composition of yeast extracts and yeast extracts fractions.