Species-specific quantification of probiotic lactobacilli in yoghurt by quantitative real-time PCR




Lactobacilli strains with probiotic effects have been widely used in dairy products such as yoghurts as well as in food additives and pharmaceuticals. Despite their successful commercial application, the current species identification and quantification methods of the genus Lactobacillus are time-consuming and labour-intensive.

Methods and Results

To fulfil the requirements of a robust quality management, we have developed a quantitative real-time PCR assay based on the heat shock protein 60 gene (hsp60) for accurate identification and quantification of five commercially important Lactobacillus species. The developed assay allows an unambiguous species-specific detection of the species Lactacidophilus, Lact. brevis, Lact. delbrueckii subsp. bulgaricus, Lact. helveticus and Lact. reuteri from bacterial cultures as well as directly from dairy products for instance yoghurt.


With the assay, we were able to specifically detect lactobacilli strains down to 105 CFU ml−1 directly from yoghurt, which is a sufficient detection limit as commercial products usually contain 106–1012 CFU ml−1 of probiotic strains.

Significance and Impact of the Study

The real-time PCR assay developed here might become a convenient tool enabling an accurate, fast and sensitive detection of probiotic lactobacilli commercially used in food.


Members of the genus Lactobacillus are Gram positive, nonmotile, nonsporulating, acid and aerotolerant facultative anaerobes and homofermentative bacteria with a low G+C content (Maier and Olek 2002). Naturally, lactobacilli occur in habitats such as wine, milk, meat, fruits, vegetables and cereal grains and have been used in food fermentation processes for centuries (Zhang et al. 2011; Tanguler and Erten 2012). In addition, some species are used as starter or adjunct cultures controlling food fermentation in dairy products (yoghurt and cheese), fermented vegetables (sauerkraut, olives, pickles), fermented meat (salami, sausages) or sourdough bread (Coeuret et al. 2003; Margolles et al. 2009; Zhang et al. 2011). Nevertheless, members of the genus Lactobacillus are present in mucosal surfaces, the intestinal tract, the vagina and the oral cavity of humans and animals (Tannock 2004; Morita et al. 2008). A variety of lactobacilli strains have been massively used commercially exploiting over the last years as they are believed to possess probiotic features. Strains are utilized in manufacturing fermented food from raw agricultural material, for example milk, meat, vegetables and cereals (Stiles 1996). There is a wide range of probiotic products available on the market partially explicitly claiming health benefits due to their possession of probiotic bacterial strains. Probiotics are defined as ‘live micro-organisms that, when administered in adequate amounts, confer a health benefit on the host’ (FAO/WHO Food and Agriculture Organization and World Health Organization, 2001) (Reid and Bocking 2003). For quality management reasons and for being in line with the European Health Claims Regulation (EC 2007), a diagnostic tool is needed to identify and quantify probiotic strains used in food. Currently, despite their economic impact, most of the assays that are used to identify lactobacilli are classical microbiological methods, which are often time-consuming, not easy to standardize and sometimes error-prone (Margolles et al. 2009). These methods are including morphology, Gram staining and biochemical tests such as fermentation of carbohydrates, growth at different temperatures and different salt concentrations (Margolles et al. 2009). Especially morphology screening seems problematic for differentiation as it is known that lactobacilli display different morphotypes within the same species (Ivanova et al. 2008).

Therefore, a fast and reliable identification tool is needed based on genomic features to identify probiotic bacteria at species level. Conventional PCR approaches which are widely used for bacterial identification lack in the quantification of bacterial loads and numbers (Berthier and Ehrlich 1998; Song et al. 2000; Kwon et al. 2004). Recently, the matrix-assisted laser desorption ionization-time-of-flight mass spectrometry (MALDI-TOF MS) is fast and reliable identification technique as much as genome-based approaches, however, has a lack in quantification capability (Dušková et al. 2012).

Although as probiotic strains have to be used in an adequate amount offering health beneficial properties, their quantification remains crucial to assess the quality of a product containing probiotics (Karapetsas et al. 2010). As an example, the number of probiotic bacterial cells is needed to induce an immune defensive benefit suppressing allergic and autoimmune disease when being used in an amount of 106 to 108 CFU ml−1 (Robinson 1987; Shah 2000).

Currently, time-consuming plating methods are used to validate the amount of bacteria. As an alternative, we developed a culture-independent molecular-based method using real-time PCR. This method should enable a fast identification and quantification of probiotic lactobacilli in food and might be applied in quality management and regulatory purposes. Quantitative real-time PCR represents an established tool for species-specific designation and quantification within a short time and readily available in many laboratories (Schaad and Frederick 2002).

In the present study, we are using the heat shock protein (hsp60) target region to identify and quantify different Lactobacillus species. This region has already been successfully used for identification approaches of other bacterial species (Goh et al. 1996; Espy et al. 2006; Park et al. 2012). Additionally, restriction fragment length polymorphism (RFLP) method is used to identify members of the genus Lactobacillus using the same target gene (Espy et al. 2006; Blaiotta et al. 2008). During this study, other target genes were utilized for species identification utilizing real-time PCR [23S-5S rRNA intergenic region (Moller et al. 2008) and rpoA (Kim et al. 2012; Oguntoyinbo and Narbad 2012)]. However, as we obtained the best results using hsp60 region, further development of the assay will be based on this target region.

Based on the commercial importance of lactobacilli in food, most common probiotic species were included in the real-time PCR assay. Examples of commercial products are LC1® yoghurt (Nestlé, Frankfurt/Main, Germany) containing c. 8 billion living Lactobacillus johnsonii LA1® cells (Gupta and Abu-Ghannam 2012) or Actimel® (Danone, Haar, Germany) with up to 10 billion viable Lactobacillus casei (Gupta and Abu-Ghannam 2012). Lactobacillus acidophilus is a probiotic strain (Ng et al. 2011) used in Proviact Bifidus pur (HMI GmbH, Everswinkel, Germany) yoghurt (Ouwehand et al. 2002) and in probiotic cereals, for example YogActive® (A & V 2000 Inc., Montréal, Canada). We also added Lactobacillus delbrueckii subsp. bulgaricus to the panel of tested strains – a probiotic species traditionally used as starter culture for instance in Ayran (Garmo AG, Stuttgart, Germany) (Simsek et al. 2007) and in commercial products such as Activia® (Danone, Haar, Germany).

Materials and methods

Lactobacilli strains and isolates

In this study, reference strains were used and their origin is listed in Table 1. Furthermore, we isolated strains from commercially available yoghurts mentioned in Table 1. A total of 77 different strains of the genera Lactobacillus (Lact.), Bifidobacterium (Bif.) and other probiotic species were used as positive and negative controls confirming species specificity of the developed primer pairs (primer pairs: Table 2, species: Table S1). Species of the genera Bifidobacterium, Lactococcus and Streptococcus were chosen as they are often used as probiotics.

Table 1. Reference Lactobacillus strains and field isolates obtained from yoghurt in this study
  1. IMT, Institut für Mikrobiologie & Tierseuchen strain collection number; DSM, Deutsche Sammlung von Mikroorganismen strain collection number.

Lactobacillus acidophilus LA-11, DGCC 9353, IMT 22354Danisco, Kantvik, Finland
Lactobacillus brevis Lbr-35, DGCC 9912, IMT 22350Danisco, Kantvik, Finland
Lactobacillus brevis DSM 20054, IMT 24651Institute of Food Science and Biotechnology, University of Hohenheim, Germany
Lactobacillus brevis DSM 6235, 6/91, IMT 25568Institute of Meat Hygiene and Technology, Freie Universität Berlin, Germany
Lactobacillus delbrueckii subsp. bulgaricus‘Zott’ 992a v. 4/88, IMT 21376Institute of Meat Hygiene and Technology, Freie Universität Berlin, Germany
Lactobacillus helveticus ‘SANOFI’ v. 9/89, IMT 21377Institute of Meat Hygiene and Technology, Freie Universität Berlin, Germany
Lactobacillus reuteri K1 D287 Botazzi v. 2/91, IMT 21493Institute of Meat Hygiene and Technology, Freie Universität Berlin, Germany
Sample isolates
 Lactobacillus acidophilus IMT 30260Proviact Bifidus pur® yoghurt
 Lactobacillus delbrueckii subsp. bulgaricusIMT 30257Activia® yoghurt
Table 2. Lactobacillus species-specific primers based on hsp60 genea
Primer pairSpeciesPrimerbTargetSequence (5′ to 3′)cAccession-No.dPCR annealing temperature and timeSize of amplicon (bp) Quantification and species specific amplification, detection limit
  1. a

    hsp60, heat shock protein 60.

  2. b

    For/FOR/F, Forward primer, Rev/REV/R, Reverse primer.

  3. c

    [X-X], region of the sequence on Accession-No. gene.

  4. d

    Accession-No., NCBI-Accession-No. in NCBI database (http://www.ncbi.nlm.nih.gov/).

  5. e

    ✓: species specific amplification was possible using this primer pair.

1 L. acidophilus






AF429668 64°C, 10 s191e, >105
2 L. brevis






AF429671 64°C, 10 s168✓, >104

L. delbrueckii

subsp. bulgaricus






AF429677 64°C, 10 s181✓, >105
4 L. helveticus






AF429683 63°C, 15 s268✓, >104
5 L. reuteri






AF429714 64°C, 8 s200✓, >105

We serially diluted 1 ml yoghurt from 10−1 to 10−8 with 0·9% NaCl (Roth, Karlsruhe, Germany). 200 μl of each serial dilution was plated on MRS broth [Roth (Danner et al. 2003)], LBS agar [BD, Heidelberg, Germany (Song et al. 2000)], COL and CHOC plates (Sarstedt, Nümbrecht, Germany). After incubation for 24–96 h (37°C, 5% CO2), colonies were subcultured and species identification was performed as described below.

Confirmation methods for species identification

Lactobacilli reference strains and the isolates from yoghurt were identified using classical microbiologic methods to assure species identity. These methods included morphology screening and biochemical tests (API 50 CHL; bioMérieux, Nürtingen, Germany). In addition, species identification was carried out by MALDI-TOF MS using a Microflex LT instrument, FlexControl 3.0 software and the BioTyper 3.0 database (Bruker Corporation, Billerica, MA, USA) (Murugaiyan et al. 2012).

All amplicons processed with the developed primers were additionally fully sequenced (LGC Genomics, Berlin, Germany) and results compared with public available sequences to assure specificity of the PCR products for every strain tested (http://www.ncbi.nlm.nih.gov/).

DNA extraction from pure lactobacilli cultures

We used a DNA extraction method modified from Walter et al. (2000) by replacing a zirconium beads step by ultrasonic sound (30 s pulsing and continuous ultrasonic sound, Sonifier Cell Disruptor B-30; Branson Sonic Power, Danbury, CT, USA). For this, pellets of the bacterial cells were resuspended in 10 mmol l−1 Tris/HCl (Roth) with 150 mmol l−1 NaCl (pH 8), and after ultrasonic sound treatment, the crude cell lysate was treated with phenol–chloroform–isoamyl alcohol [25 : 24 : 1 (Roth)]. DNA was quantified using a NanoDrop 1000 Spectrophotometer (Thermo Scientific, Dreieich, Germany) and adjusted to 10 ng μl−1 in TE buffer [10 mmol l−1 Tris (Roth), 1 mmol l−1 EDTA (Roth), pH 8·5) and stored at −20°C.

DNA extraction from yoghurt

DNA was isolated from yoghurt using the method described in Lick et al. (1996) with a minor modification. In brief, assuring an optimal dilution and distribution of all probiotic species within the yoghurt, we continuously mixed and stirred the product for 1 min and used 1 ml from the mixed material for DNA isolation. Ultrasonic sound was applied identical to the original protocol after treating the cells using mutanolysin. Following the ultrasonic treatment, samples were incubated at 37°C for 60 min, and after adding 20 μl proteinase K (20 mg ml−1; Roth) and 25% SDS (dissolved in distilled water; Sigma Aldrich, Taufkirchen, Germany) for another 10 min at 60°C, phenol–chloroform–isoamyl alcohol (25 : 24 : 1; Roth) was added to the crude cell lysate, and DNA was precipitated with ethanol (96%; Roth) and 1·10−1 of the total volume of sodium acetate (3 mol l−1, pH 5·2; Roth) followed by a storage for 30 min at −20°C. Following centrifugation (20 min at 15 000 g), the supernatant was discarded and the pellet was air-dried. Afterwards, the DNA pellet was dissolved in 100 μl TE buffer [10 mmol l−1 Tris (Roth), 1 mmol l−1 EDTA (Roth), pH 8·5).

Primer design

Partial heat shock protein gene sequences of lactobacilli were retrieved from the NCBI database (http://www.ncbi.nlm.nih.gov/). In addition, we sequenced the heat shock protein gene regions of all lactobacilli reference strains and product isolates of our study listed in Table 1. Sequences were aligned with MegAlign® alignment suite (Lasergene DNAStar, Madison, WI, USA) using the ClustalW algorithm (Higgins et al. 1996). The alignments were used to identify similarities and differences in the sequences to design species-specific primer pairs within partial conserved regions. All designed primers were screened for biophysical similarities and dimer formations with BLAST algorithm (McGinnis and Madden 2004) and OligoAnalyzer 3.0 software (Integrated DNA Technologies, Coralville, IA, USA). The primers listed in Table 2 showed the best specificity of all evaluated oligonucleotides of this study and were therefore used afterwards. In addition, all primers were checked using the BLAST algorithm of the NCBI database (http://www.ncbi.nlm.nih.gov/), assuring their specificity for the hsp60 region of the mentioned species (McGinnis and Madden 2004).

Quantitative Real-time PCR

All specific primer pairs developed during this study were tested with DNA samples of different species and strains of the genera Lactobacillus and Bifidobacterium available as positive (containing 20 ng μl−1 reference DNA) and negative controls (Table S1).

Real-time PCR was performed using a LightCycler 480® (Roche Diagnostics, Mannheim, Germany) based on SYBR Green detection (SYBR Green I Master; Roche Diagnostics). Each sample contained 5 μl of DNA template, 10 μl Mastermix, 1 μl of each primer (10 μnol; Thermo Scientific) and 3 μl of PCR-grade water. A single initial denaturation step of 7 min at 95°C was followed by 40 cycles of 95°C for 1 min (denaturation), 63–64°C for 8 to 15 s (annealing) and 73°C for 30 s (elongation). The fluorescence signal was measured at the end of each 72°C elongation step. Melting curve analyses were performed automatically by continuous heating starting at 65–95°C. Each real-time PCR included three technical repeats, and the results were analysed using the Roche® LightCycler 480® software.

Determination of the real-time PCR detection limit using pure lactobacilli DNA

To verify the sensitivity of the assay and to compare real-time PCR results with colony-forming units (CFU) of bacteria obtained by plating, DNA was isolated from serial dilutions of lactobacilli, which were plated in parallel on agar. Therefore, liquid cultures (three technical repeats) of each strain in BHI medium were set to a concentration of 108 CFU ml−1 using 0·5 McFarland Standard in a Sensititre® Inoculator (Thermo Scientific) and diluted from 108 to 103 CFU ml−1. 200 μl of each dilution was plated on MRS, LBS, COL and CHOC agar plates (origin as described above) and incubated for 24–96 h (37°C; 5% CO2). Colonies were counted and used for extrapolating the colony-forming units (CFU ml−1). One millilitre of each serial dilution of the strains was used for parallel DNA isolation using the modified protocol from Walter et al. (2000).


Species-specific amplification of lactobacilli DNA

Initially, different target regions had been tested for the species-specific detection of members of the genus Lactobacillus. Target genes such as 16S rRNA or 16S-23S rRNA had been tested successfully by Walter et al. (2000) using denaturing gradient gel electrophoresis (DGGE) or ribotyping (Walter et al. 2000; Chagnaud et al. 2001). Unfortunately, 23S-5S rRNA region was successfully working, detecting genus level by PCR, however, failed on species-specific level by real-time PCR.

We were able to establish species-specific primer pairs to detect different strains of the species Lactobacillus acidophilus, Lact. brevis, Lact. delbrueckii subsp. bulgaricus, Lact. helveticus and Lact. reuteri based on the hsp60 region using annealing temperatures shown in Table 2. Using Lactobacillus DNA from pure cultures, all reference strains and product isolates from yoghurt showed CT values ranging from 14 to 28 threshold cycles. Unspecific signals started from the 32nd cycle on; however, they were easily distinguishable by melting curve analysis. The latter unspecific amplification signals were induced by small artefacts due to primer dimers, which were proven by agarose gel electrophoresis.

Conclusively, using CT value and the melting curve analysis allowed a distinct species-specific identification of all lactobacilli species which are detectable. The specificity of the assay was tested using product isolates from yoghurt of Lact. delbrueckii subsp. bulgaricus and Lact. acidophilus. Both species isolated from different products were assigned correctly to the respective Lactobacillus species using our assay. In parallel, we always confirmed these results with conventional methods and MALDI-TOF MS (data not shown).

Detection limit, identification and quantification of Lactobacillus isolates from yoghurt

To correlate amplification curves and CT values with colony-forming units and the actual number of bacteria present in one sample, we performed serial dilution experiments as described in the Methods section. This enables assumptions concerning the amount of bacteria present in a sample based on the CT value obtained by real-time PCR. By this the detection limit of the assay was observed between 104 and 106 CFU ml−1 of all lactobacilli tested (Fig. 1, Table 2).

Figure 1.

Identification of Lactobacillus delbrueckii subsp. bulgaricus by real-time PCR. (a) Amplification curves of Lact. delbrueckii subsp. bulgaricus (IMT 21376) reference strain DNA isolate with Lact. delbrueckii subsp. bulgaricus-specific primers (Primer 3, Table 2). 1. Lactobacillus delbrueckii subsp. bulgaricus (IMT 21376), 2. Lactobacillus johnsonii (IMT 21375), 3. Lactobacillus plantarum (IMT 21365), 4. Bifidobacterium bifidum (IMT 21113), 5. Bifidobacterium thermophilum (IMT 21778), 6. Lact. helveticus (IMT 21377), 7. Water control containing Lact. delbrueckii subsp. bulgaricus primers. (b) Melting curves shown as −d(f1) dT−1 derivations. Melting temperature of Lact. delbrueckii subsp. bulgaricus IMT 21376: 86°C. Experiments were repeated with three biological and three technical repeats.

Two commercially available yoghurts (Activia®; Danone and Proviact Bifidus pur®; HMI GmbH) were used to verify whether an identification and quantification is possible using our developed assay. Therefore, DNA was directly isolated from these products without prior cultivation. We were able to identify Lact. delbrueckii subsp. bulgaricus correctly by real-time PCR assay in the Activia® yoghurt sample, which is used as a starting culture running the fermentation process. Additional melting curve analyses (Fig. 2) revealed nearly similar melting curves of the ‘unknown’ yoghurt DNA sample and the serial dilutions of reference strain DNA from Lact. delbrueckii subsp. bulgaricus. These serial dilutions enabled us to estimate the number of Lact. delbrueckii subsp. bulgaricus present in yoghurt samples based on the correlating amplification curve of the DNA mixture we extracted from Activia® yoghurt to range between 107 and 108 CFU ml−1 yoghurt. We were able to detect the strain Lact. delbrueckii subsp. bulgaricus within the DNA mixture extracted from Activia®, which was mentioned on the product's package.

Figure 2.

Quantification of Lactobacillus delbrueckii subsp. bulgaricus in DNA mixture extracted from Activia® using real-time PCR. (a) Amplification curves of Lact. delbrueckii subsp. bulgaricus reference strains' serial dilution DNA isolates, DNA isolate from yoghurt and negative controls containing Lact. delbrueckii subsp. bulgaricus primers. 1. Lact. delbrueckii subsp. bulgaricus detected in Activia® DNA isolate, 2. Lact. delbrueckii subsp. bulgaricus (IMT 21376, 4·25 × 108 CFU ml−1), 3. Lact. delbrueckii subsp. bulgaricus (IMT 21376, 4·25 × 107 CFU ml−1), 4. Lact. delbrueckii subsp. bulgaricus (IMT 21376, 4·25 × 10CFU ml−1), 5. Lact. delbrueckii subsp. bulgaricus (IMT 21376, 4·25 × 105 CFU ml−1), 6. Lact. delbrueckii subsp. bulgaricus (IMT 21376, 4·25 × 10CFU ml−1), 7. Water control containing Lact. delbrueckii subsp. bulgaricus primers. (b) Melting curves of real-time PCR as −d(f1) dT−1 derivations. Melting temperature of Lact. delbrueckii subsp. bulgaricus IMT 21376 serial DNA isolates: 87°C. Melting temperature of L. delbrueckii subsp. bulgaricus amplificates of DNA mixture isolate origin from Activia® yoghurt: 86°C. Experiments were repeated with three biological and two technical repeats.

In Proviact Bifidus pur (HMI GmbH), we were able to detect Lact. acidophilus using species-specific primer pairs, although this species was not mentioned in the nutritional facts.


Herein, we describe a real-time PCR approach for the rapid detection and quantification of important probiotic lactobacilli using DNA directly isolated from yoghurt. Our screening for an appropriate target region resulted in targeting the heat shock protein region (hsp60).

Other putative target regions including the 23S-5S rRNA (intergenic spacer region) and rpoA (Berthier and Ehrlich 1998; Song et al. 2000; Kwon et al. 2004; Cebeci and Gurakan 2011) showed no sufficient species specification due to false-positive amplification signals of negative controls. Primers targeting those regions were therefore excluded from further development, and we concentrated our work on the heat shock protein region hsp60. Using the BLAST algorithm, the sequences of the used primers were confirmed to solely target the genome of lactobacilli and no other genes were found showing a comparable DNA sequence. This conserved region appears to be a suitable target for the identification of lactobacilli as previous studies obtained similar results for other species such as Staphylococcus aureus and Staphylococcus epidermidis or Legionella pneumophila and species of the genus Bifidobacterium (Goh et al. 1996; Blaiotta et al. 2008; Karapetsas et al. 2010; Sun et al. 2010; Junick and Blaut 2012; Park et al. 2012; Yu et al. 2012; de Boer et al. 2013).

One of the key advantages of this assay is its rapidity, as it allows species-specific identification of lactobacilli strains within 7 h without any prior cultivation. In contrast, culture-dependent techniques such as API 50 CHL stripes, conventional colony PCR plus sequencing or using MALDI-TOF MS need up to 96 h to identify the isolated bacteria to species level. However, additionally utilized methods confirm the specificity of the described quantitative real-time PCR assay. The sensitivity of the assay is sufficient as the detection limit measured (105 CFU ml−1) is adequate to quantify strains of starting cultures and advertised species fortified in yoghurt in a range of 106–108 CFU ml−1 to exert probiotic activity (Robinson 1987; Shah 2000). To quantify probiotic bacteria in food, a vigorous usage of ultrasonic sound system had a positive effect on the assays' sensitivity.

Ultrasonic sound usage for a duration of 30 s on the cells of Lactobacillus spp. strains in overnight cultures, serial dilutions and diluted yoghurt samples disrupts the cells spreading the DNA throughout the sample, thus leading to earlier CT values and smaller standard derivations (data not shown).

Besides interpretation of the amplification curves, melting curve analysis presents an important parallel approach to assure that the amplicon generated is species-specific and not an artefact due to primer dimers. However, melting curve analysis of pure lactobacilli DNA and DNA mixtures isolated from yoghurt revealed differences in the melting temperature of the amplicons up to 1°C (Fig. 2). Nevertheless, by sequencing the amplicons obtained from yoghurt samples, we were able to show that the sequences were specific for the targeted strain in the yoghurt thus no false positives (Fig. S1). A possible reason for the differences in the melting temperature might be the formation of heteroduplex complexes caused by using mixtures of DNA in comparison with the homoduplexes in pure DNA (Zhou et al. 2004; Mader et al. 2008).

In conclusion, this newly developed real-time PCR assay is detecting Lactacidophilus, Lact. brevis, Lact. delbrueckii subsp. bulgaricus, Lact. helveticus and Lact. reuteri with high specificity, sensitivity and without any false-positive signals from other lactobacilli from pure cultures or DNA mixtures extracted from yoghurt. In contrast to other identification tools (physiological testing, morphology), the real-time PCR system enables the user to identify and quantify detectable strains of the genus Lactobacillus within a day including the DNA extraction from yoghurt. Therefore, this real-time PCR assay might be a useful method for the detection of Lactobacillus ssp. strains in food for regulation and quality management purposes.


This project had been financed by the ZIM fond (Zentrales Innovationsprogramm Mittelstand, KF2267401MD9) of the Federal Ministry of Economics and Technology of the Federal Republic of Germany in cooperation with the IBMT Fraunhofer Institute (Potsdam-Golm, Germany) and the CONGEN Biotechnologie GmbH (Berlin-Buch, Germany). I also would like to thank the SFB 852 (Grant No. SFB852/1) for supporting this research activity. Special thanks are going to Dr. Ouwehand (Danisco, Finland), Prof. Dr. Schmidt (Universität Hohenheim, Germany) and Prof. emerit. Dr. Dr. h.c. Reuter (Freie Universität Berlin, Germany), Dr. Loh (Deutsches Institut für Ernährungsforschung, Germany), Dr. Vahjen (Freie Universität Berlin, Germany) and Prof. Dr. Schillinger (Max Rubner-Institut, Germany) for supporting this project with different probiotic strains (Table 1, Table S1).

Conflict of Interest

The authors declare no conflict of interest.