Whey is a by-product of the cheese industry, which is generally disposed as waste especially in developing countries. Nowadays still 47% of the worldwide produced whey is discarded in rivers, lakes or loaded into the ground causing environmental contamination problems (Athanasiadis et al. Reference Athanasiadis, Parakevopoulou, Blenkas and Kiosseoglou2004). Whey is a source of vitamins, minerals and high quality proteins compared with caseins, egg and soy proteins (Dralić et al. Reference Dralić, Tratnik and Boźanić2005; Smithers, Reference Smithers2008). Currently, some of the produced whey is processed by various methods such as concentration, fractionation (ultrafiltration, microfiltration) and drying resulting in several valuable products such as lactose and whey protein concentrates.
The composition of whey varies according to the milk used, the methods of curd coagulation and the cheese produced; values are usually in the following range (g/l): 45–50 lactose, 6–8 soluble proteins, and 4–5 lipids (Gonzalez Siso, Reference Gonzalez Siso1996; Gernignon et al. Reference Gernignon, Schuck and Jeantet2010). The functional properties of whey proteins encompass physicochemical features such as solubility, whipping and foaming, emulsification and gelation that make it useful in food products (Ji & Hauque, Reference Ji and Hauque2003).
The major constituents of bovine whey proteins are β-lactoglobulin (β-lg) and α-lactalbumin (α-la), which constitute 55–60 and 15–20% of total proteins, respectively; serum albumin and immunoglobulins, lactoferrin, lactoperoxidase and various growth factors are present in lower concentrations. β-Lg is found in the whey fraction of cow's milk but is absent in human, rodents and lagomorphs milk (Iametti et al. Reference Iametti, Rasmussen, Frøkiær, Ferranti, Addeo and Bonomi2002). This protein is not degraded by pepsin and may cause allergenicity problems especially in children less than three years of age. Denaturation and proteolysis represent powerful tools for enhancing protein digestibility. Whey protein hydrolysates are physiologically better than intact proteins because their intestinal absorption appears to be more effective (Peñas et al. Reference Peñas, Préstamo, Baeza, Martínez-Molero and Gomez2006).
Lactic acid bacteria (LAB) are commonly used as starter cultures in the fermented food industry mainly due to their acidifying capacity but also to their metabolic activity on proteins, sugars and lipids, thus contributing to food digestibility and preservation as well as to the texture and sensory characteristics of the end product (Khalid & Marth, Reference Khalid and Marth1990; Wood, Reference Wood1997). Although LAB are usually considered to be weakly proteolytic, they do cause significant degree of proteolysis in multiple fermented dairy products (Pritchard & Coolbear, Reference Pritchard and Coolbear1993; Liu et al. Reference Liu, Bayjanov, Renckens, Nauta and Siezen2010). These bacteria possess a variety of proteolytic enzymes that enable them to utilize milk proteins as nitrogen source and to ensure their growth during fermentation (Matar et al. Reference Matar, Amiot, Savoie and Goulet1996). Several proteolytic enzymes have been identified for Streptococcus thermophilus and Lactobacillus delbrueckii subsp. bulgaricus; also, the whole proteolytic system of Lactococcus lactis and several lactobacilli species has been well characterized (Liu et al. Reference Liu, Bayjanov, Renckens, Nauta and Siezen2010; Do Carmo et al. Reference Do Carmo, da Silva, De Oliveira, Borges, De Carvalho and De Moraes2011).
The functional food market had a combined value of $16 billion worldwide in 2005 and nearly 43% of this market was entirely made up of fermented dairy products. Nowadays, the fermented beverage market represents a large and growing industry (Hozer & Kirmaci, Reference Hozer and Kirmaci2009). Whey-based lactic beverages constitute an emerging segment of nonconventional dairy products (Gallardo-Escamilla et al. Reference Gallardo-Escamilla, Kelly and Delahunty2007). Thus, the production of a functional beverage produced upon whey fermentation by LAB could be a good alternative for whey utilization to improve the whey nutritional and sensory characteristics in a natural and non-expensive way.
In this study sixty-four LAB strains belonging to species commonly used in the elaboration of fermented milks were evaluated for their technological characteristics as well as their capability to degrade whey proteins to select appropriate candidate strains for formulating whey-based drink starter cultures.
Materials and Methods
Microorganisms
The strains belonging to the species Lb. acidophilus (13); Lb. delbrueckii subsp. bulgaricus (15), Lb. casei (2), Lb. paracasei subsp. paracasei (13), Lb. rhamnosus (4) and Str. thermophilus (17) used in this work were obtained from the Culture Collection of Centro de Referencia para Lactobacilos (CERELA), San Miguel de Tucumán, Argentina. Cultures were stored at −20°C in 10% (w/v) sterile reconstituted skim milk containing 0·5% (w/v) yeast extract, 1·0% (w/v) glucose, and 10% (v/v) glycerol. The assayed strains were chosen as they belong to LAB species commonly used in the production of fermented milks.
Preparation of reconstituted whey powder
The composition (g/100 g product) of the whey powder (WP) used in this study (kindly provided by DANONE S.A., Buenos Aires, Argentina) was the following: fat, 2·0; protein, 10·0; moisture, 3·0; lactose, 76·5; and sodium salts, 0·40–0·45. Whey (WP) was reconstituted to two different concentrations: 10% (w/v, RW10) and 16% (w/v, RW16) with distilled water, the former one was used by Dralić et al. (Reference Dralić, Tratnik and Boźanić2005) to grow LAB while 16% (w/v) was chosen considering the low protein concentration of WP. The pH of the RW was adjusted to 6·7 with 2 m NaOH and the obtained solutions were pasteurized at 80°C for 30 min (Fiahlo et al. Reference Fiahlo, Martins, Donval, Leitáo, Ridout, Jay, Morris and Sá-Correia1999) and stored at 4°C until used (no longer than one week).
Fermentation conditions
Strains were activated in MRS (Britania, Buenos Aires, Argentina) and transferred twice in RW10 or RW16, respectively, prior to experimental use; 16-h old cultures were used as inoculums at 2% (v/v). Fermentations were performed statically in sealed bottles containing 200 ml RW10 or RW16 solution and incubated at 37°C for 24 h. Samples were aseptically withdrawn at 0, 6 and at 24 h incubation. Cell viability was determined by plating appropriate dilutions of the cultures in MRS agar (MRS plus 1·5%, w/v agar). Plates were incubated at 37°C for 48 h and colony-forming units CFU/ml were determined and expressed as log CFU/ml. Decrease in pH was followed with a digital pH meter (Altronix TPX 1, New York, USA). Results of pH and CFU/ml were expressed as the difference between values obtained at the beginning (initial counts were in the order of 106 CFU/ml) and the end of fermentation (6 or 24 h). The strains were cultured in duplicate and analyses were done in triplicate.
Proteolytic activity
Proteolytic activity by LAB was measured at 6 and 24 h incubation by the o-phthaldialdehyde (OPA) test (Church et al. Reference Church, Swaisgood, Porter and Catignani1983) measuring the increase in optical density at 340 nm relative to the control using a VERSAmax™ Tunable Microplate reader (Sunnyvale, CA, USA). The OPA solution contained: 2·5 ml 20% (w/v) SDS, 25 ml 100 mm sodium tetraborate, 40 mg OPA (previously dissolved in 1 ml methanol), 100 μl 2-mercaptoethanol and distilled water up to a 50 ml final volume. Fermented samples were incubated with 0·75 m trichloroacetic acid (ratio sample to TCA=1:3) at 4°C for 30 min and centrifuged (5000 g, 10 min). A 10 μl supernatant aliquot of this mixture was added to 0·2 ml OPA reagent and incubated at room temperature for 5 min until reading at an optical density of 340 nm in the microplate spectrophotometer. Proteolytic activity was arbitrarily expressed as mm leucine using a standard curve of L-leucine (BDH Chemicals Ltd Poole, England). Proteolytic activity was expressed as the difference between the free-amino group values (using Leu as standard) at the beginning and the end of fermentation (0 and 6 or 24 h).
Whey protein degradation
Whey protein degradation, determined in 24 h-fermentation samples, was measured by Tricine SDS-PAGE (Schagger & Von Jagow, Reference Schagger and Von Jagow1987) and performed as follows: fermented whey samples (8 μl) were suspended in 5 μl Laemmli buffer and heated at 100°C for 5 min. The corresponding blanks (non-fermented, heated RW10 and RW16), and the molecular weight marker (Protein marker, Broad Range, 2–212 kDa, New England Biolabs Inc., Ipswich, MA, USA) were loaded, separately. Tricine SDS-PAGE were carried out on 17% polyacrylamide gels on vertical slab electrophoresis cells (BIO RAD Mini PROTEAN® 3 System, Hercules, CA, USA) for 4 h at 60 V. Coomassie brilliant blue R250 was used to visualize the bands. The degradation of whey protein fractions was evaluated by densitometric analysis of gels by using the QuantiScan software (BIOSOFT 1.5, USA). The percentage of protein hydrolysis was calculated with the values obtained by integrating the peak areas displayed by the software when measuring the band intensities. The molecular weight of the bands was calculated by comparison of the mobility of the bands of the molecular weight marker in the gel using the same software.
Statistical analyses
The obtained data was analysed using principal component analysis (PCA) with INFOSTAT/professional version 2007p. The evaluated variables were ΔpH, growth (Δlog CFU/ml), proteolytic activity (ΔmM Leu) and percentage of protein degradation (variables corresponded to degradation of each protein fraction in whey) in RW10 and RW16. Comparisons between values obtained for each strain were accomplished by Analysis of Variance (ANOVA) general linear model followed by Tukey's post-hoc test and P<0·05 was considered significant.
Results and Discussion
Whey preparation and fermentation
The heat treatment (80°C for 30 min) applied to RW was sufficient to pasteurize the whey medium as no further cell growth was detected in the RW when plating in MRS agar. To evaluate whether the strains grew better in whey powder reconstituted at 10 (RW10) or 16 (RW16)% (w/v) both concentrations were used.
Whey fermentation by LAB to produce new fermented beverages could be an interesting alternative to solve whey discard. In this work, growth and proteolytic activity of diverse LAB species, usually present in yogurt and fermented milk starters, were analysed for their development in whey.
All the assayed LAB strains were able to grow in both RW10 and RW16, although to different extents. The relative distribution of the assayed strains according to their cell growth, pH values, and proteolytic activity in RW10 and RW16 after incubation at 37°C for 6 h is shown in Fig. 1. The first principal component (PC 1) displayed most of the strain discrimination (34%) resulting in a horizontal dispersion of the data mainly due to the pH and growth variables, independently of the RW concentration used.
Principal component analysis combining the variables: Growth expressed as ΔCFU/ml (Growth RW10 and Growth RW16), pH (pH RW10 and pH RW16) and proteolytic activity expressed as ΔmM Leu (Prot Act RW10 and Prot Act RW16) of sixty-four LAB strains of Lb. casei (○), Lb. paracasei (●), Lb. rhamnosus (△), Str. thermophilus (▲), Lb. delbrueckii subsp. bulgaricus (□) and Lb. acidophilus (■) in RW10 and RW16, incubated at 37°C during 6 h. The strains were cultured in duplicate and analyses were done in triplicate. Circles represent duplicates of cultures grown in RW10 and RW16 incubated at 37°C during 6 h.
According to the obtained results, growth values (Δlog CFU/ml) were arbitrarily grouped into three categories: values in the range of 0·1–0·9, were considered low; 1–2, moderate, and 2–3, high. Forty-six percent of the total strains showed moderate growth, 35% grew strongly while only 19% grew less than 0·9 Δlog CFU/ml after 6 h of incubation at 37°C in either RW assayed.
The proteolytic activity of the studied strains was evaluated since the release of amino acids and small peptides increases the nutritional values of whey as it is known that amino acids and small peptides are better absorbed in the intestine than whole proteins (Peñas et al. Reference Peñas, Préstamo, Baeza, Martínez-Molero and Gomez2006). The proteolytic activity varied among the assayed LAB strains. High, moderate, and low proteolytic activities were arbitrarily defined as: negative (due to amino acid consumption by the microorganisms) and zero values, were considered low; values between 0·05 and 0·11 Δmm Leu, moderate; and between 0·12 and 0·65; high. In general, the strains belonging to the species Lb. casei, Lb. paracasei and Lb. rhamnosus showed moderate proteolytic activity (65% of the strains in RW10 and 82% in RW16) and low acidifying capacity with maximum ΔpH values of 0·5, and were grouped on the left and bottom side of Fig. 1. These species together with Lb. zeae are phylogenetically very close and are considered to belong to the Lb. casei group; the taxonomical classification of these microorganisms was not clear until 2005 (Svec et al. Reference Svec, Drab and Sedlacek2005).
Instead, the strains of Lb. delbrueckii subsp. bulgaricus and Lb. acidophilus were placed throughout Fig. 1 showing high and moderate proteolytic activities and acidifying capacities.
The majority (10 out of 17 strains) of the streptococci showed high acidifying capacity. On the contrary, these strains consumed amino acids and small peptides initially present in whey, resulting in a negative balance of these metabolites (between −0·01 and −0·35 Δmm Leu) after 6 h incubation in both RW media indicating that in general, the Str. thermophilus strains had low proteolytic activity and some of them consumed the amino acids and small peptides initially present in whey. In this respect, Letort et al. (Reference Letort, Nardi, Garault, Monnet and Juillard2002) reported that a cell wall-associated proteinase positive Str. thermophilus strain consumed amino acids during its growth in milk. Desmazeaud & de Roissart (Reference Desmazeaud, de Roissart, de Roissart and Luquet1994) stated that Str. thermophilus strains have complex amino acid requirements and their growth is influenced by the availability and active transport/exchange of free amino acids from the growth medium.
Despite the fact that the strains were originally grouped according to their species, variations of strains within the same species were observed. This finding allowed selecting strains exhibiting the best behaviour belonging to each species or group (as the case of the Lb. casei, Lb. rhamnosus and Lb. paracasei strains). Thus, 14 strains were selected considering their proteolytic activity, acidifying capacity and/or growth values (Table 1) as follows. From the group comprising Lb. casei, Lb. rhamnosus and Lb. paracasei the strain Lb. casei CRL 238 was included for its high proteolytic activity in RW16 and moderate growth while Lb. paracasei CRL 69 and CRL 203 were chosen for their moderate to high growth and proteolytic activity in both RW10 and RW16. The strains Lb. acidophilus CRL 636, CRL 639 and CRL 745 showed high OPA values (Table 1) and moderate to high growth in RW10 and RW16. Regarding Str. thermophilus, the strain CRL 804 was selected for its high growth and proteolytic activity in whey. Also, the strains CRL 728, CRL 808 and CRL 817 were chosen for their moderate to high growth in whey and their capacity to reduce pH. Lb. delbrueckii subsp. bulgaricus CRL 454 and CRL 656 were selected for their high growth in both RW10 and RW16, while the strains CRL 870 and CRL 861 were chosen because of their high growth in RW16. Also, the strains CRL 454, CRL 656 and showed high proteolytic activity while CRL 861 showed moderate activity.
Growth and proteolytic activity of selected LAB stains in whey incubated at 37°C during 6 h
Table shows means and sd of replicate cultures. nd=sd<0·01. Comparisons were done between strains of the same species (letters) († strains belonging to Lb. casei and Lb. paracasei were considered in the same group and compared between them) in the same RW concentrations. Values with different superscript were considered significantly different after ANOVA analyses and Tukey's post-hoc test
The behaviour of the selected fourteen strains was further analysed using a longer incubation period of 24 h. The PCA distribution of these strains for their growth in RW10 and RW16 at 6 and 24 h is shown in Fig. 2; additional information of the pat-hoc statistical analysis of the results is shown in Table 2. The results obtained after 6 h incubation (on the left side of Fig. 2) displayed lower values compared with those from 24 h (right side of Fig. 2) indicating that the strains continued to grow, to digest whey proteins and to acidify the whey media beyond 6 h incubation. After this period, the Lb. paracasei strains displayed the lowest growth and acidifying capacity remaining close to the centre of Fig. 2. Lb. casei CRL 238 showed better growth and higher proteolytic activity than the Lb. paracasei strains after 24 h incubation (Fig. 2, Table 2).
Principal component analysis combining the variables: Growth expressed as ΔCFU/ml (Growth RW10 and Growth RW16), pH (pH RW10 and pH RW16) and proteolytic activity expressed as Δ mM Leu (Prot Act RW10 and Prot Act RW16), and time in h (Time) of 14 strains of Lb. casei (○), Lb. paracasei (●), Str. thermophilus (▲) Lb. delbrueckii subsp. bulgaricus (□), and Lb. Acidophilus (■) in RW10 and RW16 incubated at 37°C during 6 and 24 h. The strains were cultured in duplicate and analyses were done in triplicate. Circles represent duplicates of cultures grown in RW10 and RW16 incubated at 37°C during 6 h and 24 h.
Behaviour of selected LAB strains in whey incubated at 37°C during 24 h
Table shows means and sd of replicate cultures. nd=sd<0·01. Comparisons were done between strains of the same species (letters) († strains belonging to Lb. casei and Lb. paracasei were considered in the same group and compared between them) or among all the assayed strains (numbers) in the same RW concentrations for the cases cited in the text. Values with different superscript were considered significantly different after ANOVA analyses and Tukey's post-hoc test
Within the Lb. acidophilus species, the most proteolytic strain was CRL 745 (right and bottom of Fig. 2), which also showed the lowest acidifying capacity in RW16. Conversely, Lb. acidophilus CRL 636 showed the lowest OPA values despite its greatest growth compared with the other strains of the same species (Table 2). Concerning the Lb. delbrueckii subsp. bulgaricus species, the strains CRL 656 and CRL 454 were the most proteolytic (Table 2).
In general, the Str. thermophilus strains showed similar OPA values at 6 and 24 h incubation (Tables 1 & 2), indicating that this species probably consumed the amino acids and small peptides present in whey for sustaining their growth.
Hydrolysis of whey proteins
Peptide availability due to hydrolysis of whey proteins by the selected (14) LAB strains was assessed because of the high nutritional quality of these proteins.
Proteolysis was analysed in RW10 and RW16 in 24 h incubation-samples as no degradation was observed before this period (data not shown). Tricine SDS-PAGE and hydrolysis percentages were determined by using the QuantiScan software. The obtained results are shown in Table 3 and dispersion of the data after statistical analysis is presented in Fig. 3. SDS-PAGE of whey protein hydrolysates obtained with the most representative strains namely Str. thermophilus CRL 804 and CRL 808, Lb. casei CRL 238 and Lb. acidophilus CRL 636 is shown in Fig. 4. On the whole, the majority of the assayed strains digested the protein fractions of 69 and 50 kDa the most; Str. thermophilus CRL 808 showed the maximum degradation percentage for the 50 kDa fraction (66·7% in RW10 and 47·8% in RW16, respectively) and was placed on the lower right side of Fig. 3. In contrast, Lb. casei CRL 203 and Lb. paracasei CRL 238 poorly hydrolysed this fraction and preferentially hydrolysed the 25 kDa fraction (left side of Fig. 3). These results were opposite to those found for Lb. paracasei CRL 69, which hydrolysed more the high molecular weight fractions of 80, 69 and 50 kDa (lower right side of Fig. 3). Other authors showed variation on the ability of different strains to hydrolyse the main milk fractions. Ohmiya & Sato (Reference Ohmiya and Sato1969) studied the ability of Lb. helveticus, Lactococcus lactis and Lb. delbrueckii subsp. bulgaricus strains to degrade milk proteins and found different casein and whey protein degradation degrees for strains of each species. In addition, Khalid & Marth (Reference Khalid and Marth1990) reported degradation of different protein fractions between strains of Lb. plantarum and Lb. casei.
a) Principal component analysis of the data obtained after hydrolysis of whey proteins of: 14 kDa (14 kDa RW10 and 14 kDa RW16), 18 kDa (18 kDa RW10 and 18 kDa RW16), 25 kDa (25 kDa RW10 and 25 kDa RW16), 50 kDa (50 kDa RW10 and 50 kDa RW16), 69 kDa (69 kDa RW10 and 69 kDa RW16) and 80 kDa (80 kDa RW10 and 80 kDa RW16) by the 14 selected LAB strains of Lb. casei (○), Lb. paracasei (●), Str. thermophilus (▲), Lb. delbrueckii subsp. bulgaricus (□) and Lb. acidophilus (■) in RW10 and RW16 incubated at 37°C, during 24 h. The strains were cultured in duplicate and analyses were done in triplicate. Circles represent duplicate cultures in RW10 and RW16 incubated at 37°C during 24 h.
Tricine SDS-PAGE of whey proteins. Lanes, 1 MWM, Broad range molecular weight marker (New England Biolabs); 2 β-lg, β-lactoglobulin (ICN); 3, 5 and 7 non-fermented whey (control); 4, 6, 8 and 9 whey protein hydrolysis by Str. thermophilus CRL 808, Lb. casei CRL 238, Str . thermophilus CRL 804, and Lb. acidophilus CRL 636, respectively, in RW10 after 24 h-incubation.
Hydrolysis of whey proteins by selected LAB incubated at 37°C during 24 h
Table shows means and sd of replicate cultures. Comparisons were done between strains for the fractions compared in the text (letters) († strains belonging to Lb. casei and Lb. paracasei were considered in the same group and compared between them) or between fractions within the same strain (numbers) always for the same RW concentrations. Values with different superscript were considered significantly different after ANOVA analyses and Tukey's post-hoc test
Concerning α-la (14 kg/mol fraction), the second most abundant whey protein, Lb. acidophilus CRL 636 and Str. thermophilus CRL 728 showed the maximum hydrolysis percentages (up to 23%) in both RW media compared with all the studied strains and were placed on the upper right side of Fig. 3.
Special attention was taken into account for the breakdown of β-lg (18 kDa protein) since this protein is recognized as the major allergen of milk and it is known that hydrolysis of this protein may reduce its allergenicity (Ehn et al. Reference Ehn, Allmere, Telemo, Bengtsson and Ekstrand2005; Bu et al. 2010). On the other hand, whey proteins containing β-lg are added to bovine milk to resemble human milk when preparing infant formula; thus, whey protein hydrolysates obtained by LAB could be an interesting substrate in the formulation of hypoallergenic drinks (Pescuma et al. Reference Pescuma, Hébert, Rabesona, Drouet, Choiset, Haertlé, Mozzi, Font de Valdez and Chobert2011).
The strains which showed the highest hydrolysis percentages of β-lg were Lb. acidophilus CRL 636 (16–18%) and Lb. delbrueckii subsp. bulgaricus CRL 656 (12–14%) (Table 3) and were placed on the upper side of Fig. 3. Significant differences (P<0·05) in whey protein degradation between strains of the same species were found. Despite the fact that hydrolysis of β-lg by the studied LAB strains was not higher than 18%, some reports showed that allergenicity of this protein could be reduced after being poorly hydrolysed by LAB during fermentation (Bu et al. Reference Bu, Luo, Zhang and Chen2010; Pescuma et al. Reference Pescuma, Hébert, Rabesona, Drouet, Choiset, Haertlé, Mozzi, Font de Valdez and Chobert2011).
The extensive work done here with a large number of LAB strains and species together with the PCA applied allowed selecting strains of different species exhibiting good growth and proteolytic activity in whey. Lb. acidophilus CRL 636 was selected for its good capacity to hydrolyse β-lg and growth in whey; Lb. delbrueckii subsp. bulgaricus CRL 656 displayed high whey acidification and proteolytic activity, and the most interesting feature the highest hydrolysis of β-lg; and finally Str. thermophilus CRL 728 was able to hydrolyse β-lg and rapidly acidify the whey medium. Reconstitution of whey up to 16% (w/v) (RW16) was selected for further investigations since the selected strains Lb. delbrueckii subsp. bulgaricus CRL 656 and Str. thermophilus CRL 728 grew significantly better (P<0·05) at this RW concentration.
This work showed an extensive study on growth and proteolytic activity of LAB strains during whey fermentation. The results and the statistical analysis applied allowed the discrimination of different properties within and among species and the selection of appropriate strains to further formulate functional whey-based beverages or products.
The authors acknowledge the financial support of CONICET, ANPCyT and CIUNT from Argentina. We thank to Ing. Ricardo Weill from DANONE S.A., Argentina, for generously providing the whey powder used in this study. M. Pescuma is recipient of a post-doctoral fellowship from CONICET, Argentina.
