Effect of using mycotoxin-detoxifying agents in dairy cattle feed on natural whey starter biodiversity

118
Natural whey cultures (NWC) are undefined multiple-strain bacterial starter communities that can be affected by even small changes along the entire dairy chain. We applied a multidisciplinary approach to investigate how the addition of 2 mycotoxin-detoxifying agents [sodium smectite and lignocellulose-based material (B1); leonardite and betaine (B2)] to cow diets modified the microbiota of the NWC in manufacture of a Grana-like cheese. Microbiological and flow cytometry analyses showed that the content and viability of lactic acid bacteria (LAB) and the total whey microbiota were not affected by the detoxifying agents, and Streptococcus thermophilusLactobacillus helveticus, and Limosilactobacillus fermentum were the dominant taxa. Random amplified polymorphic DNA-PCR fingerprinting and metagenomic analysis highlighted differences in the bacterial community of the NWC and in the relative abundance of Bacteroidetes that increased when B1 and B2 were included in the diet. Two of 6 St. thermophilus biotypes were detected only in control samples; conversely, none of the Lb. helveticus biotypes found in control samples were isolated from B1 and B2. In vitro tests showed that the 2 binders did not significantly affect the development of St. thermophilus, but they stimulated the growth of Lb. helveticus strains recovered only from B1 and B2 NWC. The addition of binders in cow feed can affect the LAB biotypes present in NWC.

Key words

INTRODUCTION

Food and feed contamination with mycotoxins represents a global issue of increasing importance. These secondary metabolites, produced by toxigenic fungi, reach humans and animals by different routes, including contaminated crops (cereals and vegetables) and carryover from animal feedstuffs to animal products (

). For example, aflatoxin B1, when ingested by a lactating cow through contaminated feed, is bio-transformed into the toxic and carcinogenic metabolite aflatoxin M1 (AFM1), which excreted into milk. The amount of AFM1 present in milk is generally only about 1 to 2% of the amount of ingested aflatoxin B1, although in high-yielding dairy cows, this percentage can be higher, reaching values >6% (

). Given that AFM1 represents a potential risk for public health, several countries have regulated its maximum residue level in milk: the United States, China, and Brazil have set a maximum residue level for AFM1 of 0.5 µg/kg (

;

;

), and in Europe, milk concentrations of AFM1 must not exceed 0.05 µg/kg (

). In addition, regulatory limits for cheese have been introduced by some countries, including the Netherlands (200 ng/kg), Austria, Italy, Switzerland (250 ng/kg), China, and the United States (500 ng/kg) (

). Different authors have shown that the AFM1 concentration in cheese is related to different factors, including cheese type and technology factors, such as the amount of water removed during cheese-making and ripening (

).

Dairy products play an important role in Western diets because they contain fundamental supplements and nutrients for all age groups (

). Cheese production is usually based on fermentation processes carried out by starter cultures used to generate lactic acid–enabling gel syneresis, whey expulsion, and curd formation. Natural whey cultures (NWC) are undefined cheese starters obtained by the traditional back-slopping procedure (

) and that play a key role in cheese specificity, uniqueness, and development of sensory characteristics. Specifically, NWC ensure rapid and complete lactose fermentation and contribute to casein proteolysis and to aroma and flavor development through AA catabolism (

). Currently, NWC are used in the manufacture of different European long-ripened cheeses (Grana Padano, Parmigiano Reggiano, Comté, and Gruyère), and their core microbiome mainly comprises Lactobacillus helveticusLactobacillus delbrueckiiLimosilactobacillus fermentum (formerly known as Lb. fermentum), and Streptococcus thermophilus (

;

;

). The relationships among these lactic acid bacteria (LAB) species and their diversity can be affected by many biotic and abiotic drivers, such as bacteriophages, environmental conditions, and changes in whey and cheese-making parameters (

;

). A recent study demonstrated that milking equipment cleaning procedures can modify the microbial composition of NWC and, consequently, the core microbiome of the cheese (

). In recent years, it has become common practice to add inert sorbents to dairy cattle diets to reduce bioavailability of mycotoxins (

;

;

).

Although the effects of these compounds on raw milk composition and its yield have described (

;

), no information is available about their effects on the microbiota of dairy products.

Raw milk contains a complex microbiota derived from a variety of sources such as an animal’s teats, dairy equipment, and environmental factors (

). Microorganisms present in the environment are transferred to raw milk during and after milking through feces that adheres to the udders (

). Detoxifying agents that may exhibit antimicrobial action (

) can be blended into the feed, pass through the digestive tract of animals, and then be discharged with feces (

). Our hypothesis is that these agents could alter the fecal microbiota and, consequently, the biodiversity of raw milk and whey. Therefore, in this study, we investigated the effect of 2 mycotoxin-detoxifying agents on the microbiota of NWC used in production of a Grana-like cheese.

MATERIALS AND METHODS

The study was authorized by the Italian Health regulations that pertain to the accommodation and care of animals used for experimental and other scientific purposes (authorization n. 296/2019-PR issued on April 9, 2019).

Experimental Design and NWC Collection

The 2 mycotoxin-detoxifying agents considered in this study were provided by New Feed Team S.r.l. (Lodi, Italy). These agents—sodium smectite mixed with a lignocellulose-based material (named B1) and leonardite mixed with betaine (named B2)—were previously selected for their ability to adsorb in vitro several mycotoxins, including aflatoxin B1 (

). Thirty clinically healthy dairy cows in mid lactation were used in experiments conducted in the Eco-Farm Carpaneta experimental center (San Giorgio Bigarello, Mantova, Italy). The animals were divided into 3 groups, each consisting of 10 animals. The first (control group, CTRL) was fed a TMR; the second group (B1) received the CTRL diet (TMR) with the addition of B1 (115 ± 1 g/cow per day corresponding to 5 g/kg of DM in the TMR), and the third group (B2) was fed the TMR with the addition of B2 (115 ± 1 g/cow per day corresponding to 5 g/kg of DM in the TMR). The trial consisted of 2 periods, an initial period for adaptation (7 d) followed by the experimental period (7 d), and it aimed to evaluate the potential effect of the B1 and B2 additives on NWC microbiota. Two separate tests were conducted sequentially: the adaptation period of CTRL and B1 was conducted from d 0 to 7 and the experimental period from d 8 to 14. In the second trial, the adaptation period of the CTRL and B2 groups was conducted from d 15 to 21, and the experimental period from d 22 to 28 (Figure 1).

Figure thumbnail gr1
Figure 1Flowchart of the analyses performed on the natural whey cultures and lactic acid bacteria (LAB) strains isolated from whey cultures. RAPD = randomly amplified polymorphic DNA; CTRL = control diet; B1 and B2 = diets with mycotoxin-detoxifying agents. Light gray indicates the microbiological analysis performed on whey samples and LAB strains, medium gray indicates the genotypic LAB characterization, dark gray indicates the technological characterization of whey samples, black indicates the flow cytometry analysis, and white indicates the genotypic analysis conducted directly from natural whey cultures.
During the experimental period, cows were milked twice a day and the milk collected was used to produce a Grana-like cheese. The NWC were obtained from spontaneous fermentation of the whey from the cheese-making performed on the previous day, using identical fermenters for CTRL whey and for B1 and B2 wheys at a controlled temperature (45–46°C, for 19 h). The NWC were collected daily for analysis in the last 4 d of each experimental week; whey samples intended for microbiological and cytofluorimetric analysis were cooled to 4°C and analyzed within 12 h, whereas NWC for next-generation sequencing microbiota analysis were immediately frozen and then transferred to the laboratory. A total of 4 NWC and their respective controls were analyzed for both B1 and B2 agents. Figure 1 depicts the flowchart of the NWC analysis process for this study.

Bacterial Enumeration and LAB Identification and Typing

One milliliter of each NWC sample was serially diluted in sterile nonfat dry milk (10% wt/vol; Sacco S.r.l.) for bacterial counts. The LAB counts were determined in de Man, Rogosa, and Sharpe (MRS) agar (pH 5.4; Biolife Italiana) under anaerobic conditions (Anaerocult A, Merck), and in M17 agar containing lactose (0.5%, wt/vol; Biolife Italiana). Both media were incubated at 42°C for 72 and 48 h, respectively.
Colonies from countable plates of each medium, representing the dominant microbiota of the NWC samples, were randomly picked and subcultured in sterile nonfat dry milk (Sacco S.r.l.) at 42°C. The purity of the isolates was checked by streaking repeatedly on homofermentative heterofermentative differential agar (Biolife Italiana). After purification, the isolates were examined for cell morphology and catalase activity and then stored at −20°C in Litmus milk (Biolife Italiana).
Genomic DNA of the isolates was extracted from overnight cultures using the Microlysis kit (Aurogene) following the manufacturer’s instructions. Yield and purity of DNA were evaluated using the Infinite F200 PRO microplate reader (Tecan). The identification of isolates was carried out using a pentaplex PCR assay (Lb. delbrueckii ssp. lactisLb. delbrueckii ssp. bulgaricusLim. fermentumLb. helveticus, and St. thermophilus) previously described by

.

Random amplified polymorphic DNA (RAPD)-PCR analysis was applied to investigate the diversity and genetic relatedness within the identified strains using 3 primers (M13, D11344, and D8635;

). The resulting fingerprints were compared using the BioNumeric 5.0 software package (Applied Maths), using the unweighted pair group method with arithmetic averages (UPGMA). The biotype was defined at a minimum similarity of 90%. The diversity within the LAB isolates was measured by the Simpson (1 − D) and Shannon-Wiener (H′) indices. These indices were calculated based on RAPD-PCR patterns using scripts available in the BioNumerics 5.0 software package (Applied Maths;

). The Simpson index value of diversity ranges between 0 and 1, where a higher value indicates greater biodiversity; the Shannon-Wiener index value is generally between 1.5 and 3.5, and again, a higher value indicates greater biodiversity richness of the community (

).

Evaluation of Viability of Whey Samples

The viability and total bacterial counts of whey samples were determined by flow cytometry, according to International Organization for Standardization (ISO) standard method 19344 (

). Briefly, whey samples were diluted in PBS (NaCl 9 g/L, Na2HPO4 0.421 g/L, KH2PO4 0.144 g/L; pH 7.4), stained with 10 µL of propidium iodide (PI; 0.2 mmol/L in water) and 10 µL of SYTO 24 (0.1 mmol/L in water), and incubated in the dark for 15 min at 37°C. The staining method is based on dual nucleic acid staining with the cell-permeant dye SYTO 24 (fluorescent emission maximum at 515 nm), and cell-impermeant dye PI (fluorescent emission maximum at 620 nm). SYTO 24 permeates the membrane of total cells and stains the nucleic acids with green fluorescence; PI penetrates only bacteria with damaged membranes, reducing the SYTO 24 fluorescence when both dyes are present. Thus, live bacteria with intact cell membranes fluoresce bright green (measured in active fluorescence units, AFU), bacteria with slightly damaged membranes exhibit both green and red fluorescence (damaged cells), whereas bacteria with broken membranes fluoresce red (non-AFU). The samples were analyzed using an Accuri C6 Plus flow cytometer (BD Biosciences), equipped with a 488-nm laser beam excitation.

Acidification Activity of NWC Samples

To evaluate the effects of the mycotoxin-detoxifying agents on acidification activity of NWC, a multi-channel pH meter (Acidification Monitoring System and Analyzer, Star Ecotronics) equipped with combined pH electrodes (In Lab Power 51343110, Mettler-Toledo) was used. The NWC were inoculated (2% vol/vol) in reconstituted sterile nonfat dry milk (10% wt/vol; Sacco S.r.l.) and incubated at 42°C using a controlled water bath for 20 h. During fermentation, pH was automatically recorded at 10-min intervals. From the collected data, the following kinetic parameters were calculated: (1) maximum acidification rate (Vm; pH unit/h), (2) time to reach Vm (Tm, h), and (3) time needed to decrease the pH by 1 unit (TΔpH). All measurements were made in duplicate (

;

).

Detection of LAB Species in NWC Samples by Pentaplex PCR Analysis

Pentaplex PCR was performed to detect the presence of Lb. delbrueckii ssp. lactisLb. delbrueckii ssp. bulgaricusLim. fermentumLb. helveticus, and St. thermophilus in whey samples (

). The DNA was directly extracted from 1 mL of NWC according to

. All determinations were performed in duplicate.

RAPD-PCR Fingerprinting Analysis of Whey Samples

Randomly amplified polymorphic DNA-PCR analysis was performed to evaluate the heterogeneity of the microbial community of NWC as previously reported by

. The DNA was directly extracted from 1 mL of NWC according to

, and RAPD-PCR was applied to the total DNA of each whey sample using 4 primers: M13 (

) and OPA-4, OPA-13, and OPA-18 (

). The RAPD-PCR fingerprints were imported into BioNumeric 5.0 (Applied Maths) and analyzed using composite data set and UPGMA cluster analysis.

Metataxonomic Analysis

Genomic DNA Extraction, Library Preparation, and Sequencing

Five milliliters of each whey sample was centrifuged at 4°C for 5 min at 2,000 × g; DNA was then extracted using a method that combined a chaotropic agent, guanidium thiocyanate, with silica particles to ensure bacterial cell lysis and nuclease inactivation (

,

). Bacterial DNA was then amplified using primers (

) that target the V3-V4 hypervariable regions of the 16S rRNA gene. All PCR amplifications were performed as previously described by

.

Sequence Analysis

Demultiplexed paired-end reads from 16S rRNA gene sequencing were checked for quality with FastQC (

) and then joined into single reads using the C++ program SeqPrep (

). After joining, reads were filtered for quality based on (1) maximum of 3 consecutive low-quality base calls (Phred <19) allowed; (2) fraction of consecutive high-quality base calls (Phred >19) in a read over total read length ≥0.75; and (3) no “N”-labeled (missing/uncalled) bases. Reads that did not match all criteria were filtered out. The remaining reads were combined in a single FASTA file for the identification and quantification of operational taxonomic units (OTU). Reads were aligned against the SILVA closed reference sequence collection release 132, with 97% cluster identity (

;

), applying the Cd-hit clustering algorithm (

). A predefined taxonomy map of reference sequences to taxonomies was then used for taxonomic identification from the main taxa ranks to the genus level (domain, phylum, class, order, family, genus). By counting the abundance of each OTU, an OTU table was created and then grouped at each phylogenetic level; OTU with total counts <10 in <2 samples were filtered out. The whey core microbiota was identified by selecting OTU that were shared by at least 95% of the samples. All the above steps, except the FastQC quality check, were performed with the QIIME 1.9 open-source bioinformatics pipeline for microbiome analysis (

). The command lines and parameters that were used to process 16S rRNA gene sequence were adapted from

.

Alpha and Beta Diversity Indices and OTU

The microbial diversity of whey was assessed within (α diversity) and across (β diversity) samples. All indices (α and β diversity) were estimated from the filtered and normalized (cumulative sum scaling,

) OTU table. In addition to the number of observed OTU directly counted from the OTU table, within-sample microbial richness, diversity, and evenness were estimated using the following indices: Chao1 and abundance-based coverage estimator (ACE) for richness; Shannon, Simpson, and Fisher’s α for diversity (

;

;

;

;

;

); and Simpson E and Pielou’s J (Shannon’s evenness) for evenness (

). The across-sample microbiota diversity was quantified by calculating Bray-Curtis dissimilarities (

). Details on the calculation of the mentioned α and β diversity indices can be found in

. Analysis of variance was used to test differences between CTRL and B1 and B2 samples in terms of α diversity and OTU abundances at various taxonomic levels. P-values < 0.05 were considered significant. For Bray-Curtis dissimilarities (β diversity), differences were tested nonparametrically using permutational ANOVA (PERMANOVA), with 999 permutations (

).

Software

Reads from 16S rRNA gene sequencing were processed using the QIIME 1.9 pipeline (

), which was also used to estimate most diversity indices. The ACE index and sample-base rarefaction were estimated using custom Python scripts (https://github.com/filippob/Rare-OTUsACE.git) and R (https://github.com/filippob/sampleBasedRarefaction). The plots were generated using the ggplot2 R package (

). Additional data handling and statistical analysis were performed in R (

).

Acidifying Activity of LAB Isolates and Growth Kinetics

One representative isolate for each biotype was chosen to evaluate the effect of the single agent on LAB acidifying activity. The strains were previously grown in sterile nonfat dry milk (Sacco S.r.l.) at 42°C for 18 h; then, 2% (vol/vol) of each fresh culture was inoculated in reconstituted milk supplemented with the mycotoxin-detoxifying agent (0.25%, wt/vol) to obtain an initial concentration of 105 cfu/mL. After inoculation, the samples were stirred for 20 min at room temperature to allow possible interactions between LAB and the B1 and B2 binders and then transferred to a water bath connected to an AMSA (Star Ecotronics) analyzer to monitor the acidification activity of LAB. The control consisted only of inoculated milk without addition of mycotoxin-detoxifying agents. Fermentations were performed in duplicate at 42°C for 20 h. As described above, TΔpH was used to describe the acidification kinetics of each isolate (

;

).

Growth kinetics of each microorganism were investigated throughout the milk fermentation, in the absence or presence of the B1 and B2 agents. The isolates were previously grown in sterile nonfat dry milk (42°C for 18 h) and then inoculated in reconstituted milk (2%) supplemented with B1 or B2 (0.25%, wt/vol). The maximum specific growth rate (μmax) was calculated for each isolate using the equation μmax = ln(X2 − X1)/(t2 − t1), where X2 and X1 are the counts (cfu/mL) at time t2 and t1, respectively. Generation time (GT) was calculated for each culture from the corresponding value of μmax (GT = ln2/μmax) (

). Streptococcus thermophilus isolates were enumerated on M17 agar 0.5% (wt/vol) lactose (Biolife Italiana), whereas Lb. helveticus were counted on MRS agar (pH 5.4; Biolife Italiana) under anaerobic conditions (Anaerocult A, Merck). Both media were incubated at 42°C for 48 and 72 h.

Generation time was also measured to determine the lowest concentration of B1 or B2 necessary to inhibit the LAB growth. Isolates (2%; about 107 cfu/mL) were inoculated in reconstituted milk supplemented with different concentrations of B1 and B2 ranging from 0.05 to 0.25% (wt/vol). The inoculated milk was stirred for 20 min at room temperature to allow the possible interaction between LAB and B1 and B2, and then transferred to a water bath at 42°C for 20 h. The control consisted of inoculated milk without either B1 or B2. All tests were performed in duplicate. The cut-off value was defined as the lowest concentration of agent that caused a GT increase of at least 0.5 h compared with the respective control.

Statistical Analysis

All data related to microbiological assays, acidifying activities, and growth kinetics are presented as means ± standard deviation (SD). Significant differences (P < 0.05) among the data were calculated by one-way ANOVA using Minitab ver. 14.13 (Minitab Inc.).

RESULTS AND DISCUSSION

Neither detoxifying agent (B1 or B2) significantly affected the level of LAB in NWC; the LAB content in MRS and M17 agar plates was similar in all samples analyzed: LAB counts in MRS ranged from 7.3 to 7.8 log10 cfu/mL and those in M17 ranged from 7.5 to 7.8 log10 cfu/mL (Table 1).

Table 1Microbial counts, flow cytometry, and lactic acid bacteria (LAB) species detected in whey samples and kinetic acidification parameters of reconstituted milk inoculated with natural whey cultures from cows fed control diets (CTRL) and diets with mycotoxin-detoxifying agents (B1 and B2)
Item Whey sample

(n = 4 each)

CTRL B1 CTRL B2
Microbial count

 LAB in MRS agar 7.7 ± 0.1 7.8 ± 0.1 7.5 ± 0.5 7.5 ± 0.7
 LAB in M17 agar 7.3 ± 0.2 7.4 ± 0.2 7.8 ± 0.1 7.6 ± 0.3
Flow cytometry

 Live cells 8.4 ± 0.9 8.8 ± 0.1 9.1 ± 0.2 9.1 ± 0.2
 Damaged cells 8.0 ± 0.3 8.2 ± 0.2 8.3 ± 0.2 8.4 ± 0.2
 Dead cells 8.8 ± 0.2 8.8 ± 0.0 8.6 ± 0.1 8.6 ± 0.2
 Total count 9.1 ± 0.1 9.2 ± 0.1 9.3 ± 0.1 9.3 ± 0.1
Kinetic parameters of acidification

 Vm (pH/h) 0.69 ± 0.01 0.66 ± 0.02 0.58 ± 0.13 0.60 ± 0.13
 Tm (h) 3.5 ± 0.2 3.5 ± 0.0 3.7 ± 0.2 3.9 ± 0.3
 TΔpH (h) 4.6 ± 0.1 4.6 ± 0.1 4.7 ± 0.2 4.8 ± 0.2
Pentaplex PCR
Streptococcus thermophilus + + + +
Limosilactobacillus fermentum + + + +
Lactobacillus helveticus + + + +
Lactobacillus delbrueckii ssp. lactis ­­– +
Lb. delbrueckii ssp. bulgaricus
1Data were expressed as means ± SD of the different parameters considered.
2 CTRL = control samples; B1 = B1 detoxifying agent; B2 = B2 detoxifying agent.
3 Microbial count results were expressed as log10 cfu/mL. MRS = de Man, Rogosa, and Sharpe.
4 Flow cytometry data were reported as log10 AFU/mL (live cells), log10 non-AFU/mL (dead cells), and log10 FU/mL (total count), where AFU = active fluorescence units and FU = fluorescence units.
5 Vm = maximum acidification rate; Tm = time to reach Vm; TΔpH = time needed to decrease the pH by 1 unit.
Analogous results were obtained studying the effect of B1 and B2 on the microbial population of the NWC by flow cytometry. Flow cytometry is a culture-independent methodology that, by using specific probes for different cell targets, supplies information about cell structure and physiological status within a bacterial population (

). As shown in Table 1, no significant differences were detected among the 3 subpopulations identified (live, damaged, and dead cells), and neither binder affected the total counts of the samples examined (∼9.0 log10 cfu/mL). This result could indicate a nontoxic activity of these compounds on the cheese whey microbiota overall. Values obtained from flow cytometry were ∼1.0 to 1.5 orders of magnitude higher than those obtained by the plate counting technique. Therefore, the differences between flow cytometry data and plate counting are likely due to the presence in the whey samples of viable but nonculturable cells that are not enumerated by the culture method (

) because lactobacilli and streptococci culturable in MRS and M17 are the main components of the whey microbiota (

;

).

No significant differences were observed in any of the kinetic parameters of acidification considered herein, indicating that the addition of detoxifying agents (B1 and B2) to the cow diet did not affect the acidifying activity of the whey cultures (Table 1). The lack of effects of these substances on the NWC technological parameters is noteworthy because they could affect the cheese-making process, such as curd formation and whey drainage.
The pentaplex PCR showed that all NWC contained Lb. helveticusLim. fermentum, and St. thermophilus, in agreement with previously performed studies that characterized Grana-like cheese NWC throughout culture-dependent and culture-independent methods (

;

;

). Lactobacillus delbrueckii ssp. lactis, one of the most abundant LAB species in NWC, was recovered only in B2 whey samples (Table 1). This result confirms the findings of recent studies (

;

), which indicate a drastic reduction or absence of Lb. delbrueckii ssp. lactis in Grana Padano whey cultures.

The RAPD-PCR fingerprinting profiles obtained from total DNA of whey samples allowed us to group the NWC into 2 clusters (Figure 2). Most samples in the first trial (B1 and respective CTRL) were grouped in the same cluster and showed high similarity (84.7%), whereas high variability was observed among samples of the second trial. In total, 16 common RAPD bands were detected in all whey samples, which can be considered to represent bacterial “core” community of NWC obtained under the stressful conditions imposed by the processing of whey (

).

Figure thumbnail gr2
Figure 2Band matching cluster analysis showing the relationship between B1 and B2 whey cultures and their respective controls (CTRL). Randomly amplified polymorphic DNA (RAPD) patterns were obtained using 4 primers (M13, OPA-4, OPA-13, and OPA-18).
The microbiota structure of whey starter samples, explored by 16S rRNA gene sequencing, was characterized by a total of 2,271,093 high-quality reads, with a mean of 43,674.9 reads per sample. Rarefaction curves suggested that the depth of coverage was sufficient to describe the biological diversity in the sampled whey starters. The α and β diversity metrics differed between groups: B1 and B2 treatments were compared with the corresponding controls (CTRL for B1 and B2). The B2 groups were significantly different for 4 α indices (Shannon, P < 0.017; Simpson, P < 0.006; equitability, P < 0.0012; Simpson E, P < 0.0009), showing the high level of complexity associated with this treatment (Supplemental Figure S1; https://zenodo.org/record/6671355#.YrCQStJBxH4;

); similar community structures were observed in β diversity for B1 and B2 treatments.

Firmicutes was the dominant phylum, accounting for more than 70% of the bacterial microbiota in all samples, with an average relative abundance of 84.5 and 78.8% for CTRL whey samples and 73.7 and 72.2% for B1 and B2 samples, respectively (Figure 3). In agreement with recent data on Grana Padano cheese, Lactobacillus was the most prevalent genus (77.7 and 68.9% in CTRL samples and 63.1 and 52.1% in B1 and B2, respectively) present in whey cultures, with Lb. delbrueckiiLb. hamsteriLb. helveticus, and Lim. fermentum being the most abundant species recovered in all samples (

;

). Bacteroidetes represented the remaining dominant microbiota, increasing from 5% (CTRL) to 13.3% for B1 and from 8.5% (CTRL) to 15.6% for B2 (Figure 3). An increase in Bacteroidetes in ruminal microbiota was also shown by

when dairy cows were fed a TMR containing bentonite clay. Bacteroidetes are present in raw milk and usually come from the bovine environment (feces and teat skin;

). These findings support our hypothesis that addition of detoxifying agents to feed can affect the ruminal and fecal microbiota and, consequently, the microbial composition of raw milk and whey starter. Several other genera of gram-negative microorganisms were detected (Figure 3).

Figure thumbnail gr3
Figure 3Bubble chart of relative abundances of all taxa (≥1%) in the microbiota of the whey samples. B1 (circles), B2 (triangles), CTRL for B1 (squares), CTRL for B2 (plus signs) experimental groups. Different colors indicate different taxa. The size of the symbols is proportional to the relative abundance of taxa per treatment.
To further investigate the effect of the 2 mycotoxin-detoxifying agents on NWC biodiversity, we collected 111 colonies with different morphologies from M17 (55 isolates) and MRS (56 isolates) agar plates. Eighteen isolates did not grow after plate isolation and were discarded, 45 were identified as St. thermophilus, 47 as Lb. helveticus, and 1 as Lb. delbrueckii ssp. lactis. These results were in accordance with those obtained from the pentaplex PCR and 16S rRNA gene sequencing.
We performed RAPD-PCR fingerprinting to establish the number of genotypically different biotypes among the isolates (Figure 4). Cluster analysis of RAPD patterns showed that during the first trial (B1), St. thermophilus and Lb. helveticus were each grouped in 4 biotypes, with a high similarity level (81.0 and 82.3% respectively; Figure 4A). A different scenario was observed in the second trial (B2), where the percentage homologies of St. thermophilus (3 biotypes) and Lb. helveticus strains (7 biotypes) were 88.6 and 63.4%, respectively (Figure 4B). As reported in Table 2, 4 St. thermophilus biotypes (ST1, ST3, ST5, and ST6) were recovered in both CTRL samples and in B1 or B2, whereas strains ST2 and ST4 were isolated only from the controls. In contrast, none of the Lb. helveticus biotypes found in CTRL samples were isolated from B1 and B2 NWC except for LH3. Only 2 strains, St. thermophilus ST4 and Lb. helveticus LH1, were detected in both trials (Table 2). These differences in LAB biotypes were also highlighted by the Simpson (1 − D) and Shannon-Wiener (Table 2) indices, which were higher in CTRL samples, confirming the lower diversity in B1 and B2 NWC (

).

Figure thumbnail gr4
Figure 4Unweighted pair group method with arithmetic averages–based dendrogram derived from the RAPD-PCR analysis of Streptococcus thermophilus (ST), Lactobacillus helveticus (LH), and Lactobacillus delbrueckii ssp. lactis (LDL) strains isolated from natural whey cultures collected in this study. (A) Biotypes isolated from the CTRL and B1 natural whey cultures (NWC); (B) biotypes isolated from CTRL and B2 NWC. RAPD = randomly amplified polymorphic DNA; CTRL = control diet; B1 and B2 = diets with mycotoxin-detoxifying agents. The gray triangles represent the number of strains belonging to the same biotype.
Table 2Streptococcus thermophilusLactobacillus helveticus, and Lactobacillus delbrueckii ssp. lactis isolates isolated from natural whey cultures and Simpson and Shannon indices relating to the 2 trials

 

Species Biotype CTRL samples B1 samples
No. of strains Day No. of strains Day
1 2 3 4 1 2 3 4
St. thermophilus ST1 5 X X 10 X X X X
ST2 2 X
ST3 4 X X 2 X
ST4 3 X X
Lb. helveticus LH1 3 X
LH2 3 X X
LH3 6 X X X 4 X X
LH4 5 X X
Total number of strains 23 24
Simpson index 0.85 0.76
Shannon-Wiener index 1.73 1.45
CTRL samples B2 samples
Day Day
Species 1 2 3 4 1 2 3 4
St. thermophilus ST4 2 X
ST5 6 X X X 8 X X X
ST6 4 X X 4 X X
Lb. helveticus LH1 6 X X X
LH5 4 X X
LH6 2 X
LH7 2 X
LH8 2 X
LH9 2 X
LH10 2 X
Lb. delbrueckii ssp. lactis LDL1 1 X
Total number of strains 22 23
Simpson index 0.86 0.80
Shannon-Wiener index 1.84 1.58
1 Samples: CTRL = control samples; B1 = B1 detoxifying agent; B2 = B2 detoxifying agent.
2 Values represent the number of strains of each biotype determined by randomly amplified polymorphic DNA (RAPD)-PCR analysis and day of isolation.
The dominant bacterial structure differed slightly by day, highlighting that the whey cultures are dynamic microbial communities whose composition may change over time (

). Despite this, our results indicated that the mycotoxin-detoxifying agents can affect the LAB biotypes present in NWC, especially with regard to Lb. helveticus. Our hypothesis is that these agents can shift the fecal microbiota and consequently the biodiversity of the raw milk and whey. Thus, in addition to investigating the microbial biodiversity of the different NWC, we evaluated their growth rate and metabolism through their acidifying activity to further explore whether the presence of the 2 additives could differentially inhibit or stimulate individual bacterial strains. The acidification capability of the different biotypes was evaluated in pH-controlled fermentation either in the presence (0.25% wt/vol) or absence of B1 and B2 compounds in reconstituted milk, taking in account the difficulty of growing strains isolated from NWC in synthetic medium. The percentage of detoxifying agent was chosen to be consistent with the concentration of B1 and B2 present in dairy cow feces.

Considering the strains isolated during the first trial, we noted that addition of B1 sped up the fermentation process, resulting in a reduction in TΔpH (Table 3). This effect was particularly evident in all biotypes recovered from the B1 NWC but was not significant in the strains isolated from CTRL samples (Table 3). In contrast, the presence of B2 significantly delayed fermentation in all biotypes considered (P < 0.05; Table 3).

Table 3Kinetic parameters of acidification of milk fermented by lactic acid bacteria strains (time needed to decrease the pH by 1 unit, TΔpH) and their generation time (GT), in the presence or absence of the 2 mycotoxin-detoxifying agents

 

Source of LAB biotypes

Biotype

TΔpH (h) GT (h) MIC (%)
M M+B M M+B
CTRL ST2 7.2 ± 0.0 6.7 ± 0.3 2.1 ± 0.0 2.2 ± 0.1 >0.25
ST4 7.0 ± 0.0 6.8 ± 0.1 2.1 ± 0.0 2.1 ± 0.0 >0.25
LH2 >20.0 >20.0 3.1 ± 1.0

4.6 ± 0.3

0.25
CTRL and B1 ST1 6.7 ± 0.0

5.5 ± 0.0

2.1 ± 0.0 2.0 ± 0.0 >0.25
ST3 8.8 ± 0.2

7.0 ± 0.0

2.1 ± 0.0 2.0 ± 0.0 >0.25
LH3 17.2 ± 0.2

15.0 ± 0.1

2.9 ± 0.1 2.9 ± 0.2 >0.25
B1 LH1 17.6 ± 0.0

15.2 ± 0.0

3.6 ± 0.5

2.5 ± 0.2

>0.25
LH4 18.7 ± 0.1

16.8 ± 0.1

2.3 ± 0.1

1.8 ± 0.0

>0.25
CTRL ST4 6.3 ± 0.1

10.7 ± 0.1

2.9 ± 0.1

4.9 ± 0.1

0.13
LH5 17.2 ± 0.2

18.3 ± 0.2

3.1 ± 0.1

4.3 ± 0.0

0.25
LH6 >20.0 >20.0 2.4 ± 0.0

3.4 ± 0.2

0.25
LH7 16.2 ± 0.1

>20.0

3.1 ± 0.3

4.6 ± 0.5

0.13
LH8 >20.0 >20.0 3.2 ± 0.0

5.0 ± 0.2

0.25
CTRL and B2 ST5 10.2 ± 0.0

>20.0

3.0 ± 0.0 3.1 ± 0.0 >0.25
ST6 10.2 ± 0.0

18.5 ± 0.0

3.3 ± 0.1 3.2 ± 0.0 >0.25
B2 LH1 14.8 ± 0.1

17.7 ± 0.1

3.8 ± 0.5

2.6 ± 0.1

>0.25
LH9 >20.0 >20.0 2.8 ± 0.1

2.1 ± 0.0

>0.25
LH10 18.7 ± 0.1

>20.0

3.2 ± 0.2

2.4 ± 0.1

>0.25
A,B Means with different letters within a row are significantly different (P < 0.05).
1 Data are expressed as means ± SD.
2 M = milk; M+B = milk supplemented with B1 or B2 (0.25%, wt/vol).
3 CTRL = control samples; B1 = B1 detoxifying agent; B2 = B2 detoxifying agent.
4 Biotypes of Streptococcus thermophilus (ST) and Lactobacillus helveticus (LH).
The effect of B1 and B2 on LAB biotype growth was also evaluated by GT, which is defined as the time needed to double the population of microbes in culture medium. This parameter can be used to evaluate the effect of an antimicrobial agent or other compound on microorganism growth (

). In this study, GT was calculated to assess the effect of B1 and B2 on LAB growth kinetics. As reported in Table 3, the addition of B1 to the culture medium did not significantly affect the growth of St. thermophilus biotypes. However, B1 slowed the growth of the only Lb. helveticus biotype isolated from the CTRL samples (LH2), whereas it did not affect the LH3 strain recovered from both CTRL and B1 whey cultures and significantly stimulated the growth of biotypes LH1 and LH4 isolated only from the B1 NWC (P < 0.05; Table 3). A similar scenario was observed using B2 (Table 3). The addition of B2 to the culture medium significantly (P < 0.05) increased the GT values of all strains recovered from CTRL samples and promoted the growth of the Lb. helveticus strains found in B2 whey cultures. In addition, B2 did not affect the growth of strains isolated from CTRL and B2 NWC. As reported in Table 3, we observed a poor correlation between the variation in acidifying activity and growth rate. These discrepancies were probably due to the composition of B1 and B2 and their buffering power.

Generation time was also used to determine the lowest concentration of mycotoxin-detoxifying agent necessary to inhibit LAB growth. As shown in Table 3, B1 and B2 did not affect the growth of most St. thermophilus strains but did slow growth of all Lb. helveticus biotypes isolated from CTRL samples. Therefore, as for Lb. helveticus, growth of strains isolated from whey starter related to the use of the 2 binders was stimulated by the presence of the specific binder, whereas growth of strains isolated in the corresponding CTRL starter is significantly slowed.
Although the 2 mycotoxin-detoxifying agents exerted positive and negative effects on growth of individual biotypes, their exact mode of action remains unclear. Presumably, these activities are related to the mineral clays (smectite and leonardite, respectively) but they may also be due to the other compounds (polyphenols and humic acids) that make up B1 and B2 detoxifiers.

reported that a variety of physical and chemical processes can confer an antibacterial effect on clays. It is known that clay minerals can adhere by physical attraction to cell walls, hampering the passive and active uptake of essential nutrients and thus, changing the bacterial metabolism. Alternatively, clays may deprive bacteria of essential nutrients by adsorption. Moreover, environmental conditions can affect the potential bacteriostatic activity of these; specifically, at pH <5.0 or >9.0, many metals present in hydrated clay minerals become soluble, increasing the oxidation state (>400 mV), which affects bacterial defenses (

). Polyphenols from lignin and their metabolites can act as activators or inhibitors of bacterial growth. Studies conducted in humans have shown that these compounds can stimulate the development of Lactobacillus in the gut, thus changing the gut microbiota composition (

). Nevertheless, how these compounds affect bacterial development remains partly unknown but seems related to their chemical structure and concentration. Previous studies also reported that the effect of polyphenols on LAB growth was strain dependent (

), and the addition of these compounds to skim milk can improve the acidification activity of LAB strains (

). Humic acids present in B2 could also promote or hinder bacterial growth. At the present time, no literature is currently available on the effect of humic acids on LAB strains, because these compounds have been tested only against soil bacteria and pathogenic species. In this case, the chemical composition of the humic acids was considered a key factor determining the positive or negative effect on bacterial growth (

;

).

CONCLUSIONS

The 2 mycotoxin-detoxifying agents did not affect LAB counts or acidifying activity of the NWC but did affect the LAB biotypes present in NWC. We observed an increase in the relative abundance of Bacteroidetes when B1 or B2 was included in the cow diet, highlighting their effect on the gut microbiota. Detoxifying agents that are blended into feed are known to pass through the digestive tract of animals and then discharged with feces. Because different studies have revealed that cow feces greatly affect milk microbiota composition, our hypothesis is that these agents shift the fecal microbiota and, consequently, the microbial composition of the raw milk and whey starter. Despite the small number of strains from the NWC included in this study, our results warrant further investigation to better understand the complex relationship between the inclusion of mycotoxin-detoxifying agents in animal feed and the microbiota composition in whey starter. Changes in the whey microbiota can have negative consequences for the quality and sensory characteristics of the cheese.

ACKNOWLEDGMENTS

This work was supported by the project AGER 2 “Farm level interventions supporting dairy industry innovation (FARM-INN)” (grant no. 2017-1130). The authors have not stated any conflicts of interest.

REFERENCES

    • Anderson M.
    A new method for non-parametric multivariate analysis of variance.

    Austral Ecol. 2001; 2632-46

    • Andrew S.
    FastQC: A quality control tool for high throughput sequence data.
    http://www.bioinformatics.babraham.ac.uk/projects/fastqc

    Date: 2020
    Date accessed: November 20, 2020
  1. ANVISA. 2011. Agência Nacional de Vigilância Sanitária Dispõe sobre limites máximos tolerados (LMT) para micotoxinas em alimentos ANVISA, Brasília, DF (2011) Resolução RDC nº 7, de 18 de fevereiro de 2011.
    • Avantaggiato G.
    • Greco D.
    • D’Ascanio V.
    • Logrieco A.F.
    Advances and criticisms on the use of mycotoxin detoxifying agents.

    in: Mycotoxins in Food and Beverages. CRC Press2021122-153

    • Bertani G.
    • Levante A.
    • Lazzi C.
    • Bottari B.
    • Gatti M.
    • Neviani E.
    Dynamics of a natural bacterial community under technological and environmental pressures: The case of natural whey starter for Parmigiano Reggiano cheese.

    Food Res. Int. 2020; 129 (32036924)108860

    • Biscarini F.
    • Palazzo F.
    • Castellani F.
    • Masetti G.
    • Grotta L.
    • Cichelli A.
    • Martino G.
    Rumen microbiome in dairy calves fed copper and grape-pomace dietary supplementations: Composition and predicted functional profile.

    PLoS One. 2018; 13 (30496201)e0205670

    • Bray J.R.
    • Curtis J.T.
    An ordination of the upland forest communities of Southern Wisconsin.

    Ecol. Monogr. 1957; 27325-349

    • Cachon R.
    • Jeanson S.
    • Aldarf M.
    • Divies C.
    Characterisation of lactic starters based on acidification and reduction activities.

    Lait. 2002; 82281-288

    • Campagnollo F.B.
    • Ganev K.C.
    • Khaneghah A.M.
    • Portela J.B.
    • Cruz A.G.
    • Granato D.
    • Corassin C.H.
    • Oliveira C.A.F.
    • Sant’Ana A.S.
    The occurrence and effect of unit operations for dairy products processing on the fate of aflatoxin M1: A review.

    Food Control. 2016; 68310-329

    • Caporaso J.G.
    • Kuczynski J.
    • Stombaugh J.
    • Bittinger K.
    • Bushman F.D.
    • Costello E.K.
    • Fierer N.
    • Peña A.G.
    • Goodrich J.K.
    • Gordon J.I.
    • Huttley G.A.
    • Kelley S.T.
    • Knights D.
    • Koenig J.E.
    • Ley R.E.
    • Lozupone C.A.
    • McDonald D.
    • Muegge B.D.
    • Pirrung M.
    • Reeder J.
    • Sevinsky J.R.
    • Turnbaugh P.J.
    • Walters W.A.
    • Widmann J.
    • Yatsunenko T.
    • Zaneveld J.
    • Knight R.
    QIIME allows analysis of high-throughput community sequencing data.

    Nat. Methods. 2010; 7 (20383131)335-336

    • Caporaso J.G.
    • Lauber C.L.
    • Walters W.A.
    • Berg-Lyons D.
    • Lozupone C.A.
    • Turnbaugh P.J.
    • Fierer N.
    • Knight R.
    Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample.

    Proc. Natl. Acad. Sci. USA. 2011; 108 (20534432)4516-4522

    • Chambel L.
    • Sol M.
    • Fernandes I.
    • Barbosa M.
    • Zilhão I.
    • Barata B.
    • Jordan S.
    • Perni S.
    • Shama G.
    • Adrião A.
    • Faleiro L.
    • Requena T.
    • Peláez C.
    • Andrew P.W.
    • Tenreiro R.
    Occurrence and persistence of Listeria spp. in the environment of ewe and cow’s milk cheese dairies in Portugal unveiled by an integrated analysis of identification, typing and spatial-temporal mapping along production cycle.

    Int. J. Food Microbiol. 2007; 11652-63

    • Chao A.
    Nonparametric estimation of the number of classes in a population.

    Scand. J. Stat. 1984; 11265-270

    • Chao A.
    • Lee S.M.
    Estimating the number of classes via sample coverage.

    J. Am. Stat. Assoc. 1992; 87210-217

    • Chao A.
    • Yang M.C.K.
    Stopping rules and estimation for recapture debugging with unequal failure rates.

    Biometrika. 1993; 80193-201

    • Chhaya R.S.
    • O’Brien J.
    • Cummins E.
    Feed to fork risk assessment of mycotoxins under climate change influences—Recent developments.

    Trends Food Sci. Technol. 2021;

    • Cremonesi P.
    • Castiglioni B.
    • Malferrari G.
    • Biunno I.
    • Vimercati C.
    • Moroni P.
    • Morandi S.
    • Luzzana M.
    Technical Note: Improved method for rapid DNA extraction of mastitis pathogens directly from milk.

    J. Dairy Sci. 2006; 89 (16357279)163-169

    • Cremonesi P.
    • Ceccarani C.
    • Curone G.
    • Severgnini M.
    • Pollera C.
    • Bronzo V.
    • Riva F.
    • Addis M.F.
    • Filipe J.
    • Amadori M.
    • Trevisi E.
    • Vigo D.
    • Moroni P.
    • Castiglioni B.
    Milk microbiome diversity and bacterial group prevalence in a comparison between healthy Holstein Friesian and Rendena cows.

    PLoS One. 2018; 13 (30356246)e0205054

    • Cremonesi P.
    • Morandi S.
    • Ceccarani C.
    • Battelli G.
    • Castiglioni B.
    • Cologna N.
    • Goss A.
    • Severgnini M.
    • Mazzucchi M.
    • Partel E.
    • Tamburini A.
    • Zanini L.
    • Brasca M.
    Raw milk microbiota modifications as affected by chlorine usage for cleaning procedures: The Trentingrana PDO case.

    Front. Microbiol. 2020; 11 (33123103)564749

    • Cremonesi P.
    • Vanoni L.
    • Morandi S.
    • Silvetti T.
    • Castiglioni B.
    • Brasca M.
    Development of a pentaplex PCR assay for the simultaneous detection of Streptococcus thermophilus, Lactobacillus delbrueckii subsp. bulgaricus, L. delbrueckii subsp. lactis, L. helveticus, L. fermentum in whey starter for Grana Padano cheese.

    Int. J. Food Microbiol. 2011; 146 (21377750)207-211

    • D’Ascanio V.
    • Greco D.
    • Menicagli E.
    • Santovito E.
    • Catucci L.
    • Logrieco A.F.
    • Avantaggiato G.
    The role of geological origin of smectites and of their physico-chemical properties on aflatoxin adsorption.

    Appl. Clay Sci. 2019; 181105209

    • de Souza Oliveira P.
    • Aliakbarian B.
    • Casazza A.A.
    • Le Blanc J.
    • Perego P.
    • De Souza Oliveira R.
    Production of fermented skim milk supplemented with different grape pomace extracts: Effect on viability and acidification performance of probiotic cultures.

    PharmaNutrition. 2018; 664-68

    • De Souza Oliveira R.P.
    • Perego P.
    • Converti A.
    • De Oliveira M.N.
    Growth and acidification performance of probiotics in pure culture and co-culture with Streptococcus thermophilus: The effect of inulin.

    Lebensm. Wiss. Technol. 2009; 421015-1021

    • EFSA
    Opinion of the scientific panel on contaminants in the food chain on a request from the commission related to aflatoxin B1 as undesirable substance in animal feed.

    EFSA J. 2004; 391-27

    • Falardeau J.
    • Keeney K.
    • Trmčić A.
    • Kitts D.
    • Wang S.
    Farm-to-fork profiling of bacterial communities associated with an artisan cheese production facility.

    Food Microbiol. 2019; 83 (31202418)48-58

    • FDA
    Guidance for industry and other stakeholders: Toxicological principles for the safety assessment of food ingredients.

    Redbook. 2000; 31-286

    • Feeney E.L.
    • Lamichhane P.
    • Sheehan L.J.
    The cheese matrix: Understanding the impact of cheese structure on aspects of cardiovascular health–A food science and a human nutrition perspective.

    Int. J. Dairy Technol. 2021; 74656-670

    • Fisher R.A.
    • Corbet A.S.
    • Williams C.B.
    The relation between the number of species and the number of individuals in a random sample of an animal population.

    J. Anim. Ecol. 1943; 1242

    • Foglia C.
    • Allesina S.
    • Amoruso A.
    • De Prisco A.
    • Pane M.
    New insights in enumeration methodologies of probiotic cells in finished products.

    J. Microbiol. Methods. 2020; 175 (32621828)105993

    • Forbes B.A.
    • Sahm D.F.
    • Weissfeld A.S.
    Study Guide for Bailey and Scott’s Diagnostic Microbiology.

    Elsevier Health Sciences2016

    • Gallo A.
    • Rocchetti G.
    • Piccioli Cappelli F.
    • Pavone S.
    • Mulazzi A.
    • van Kuijk S.
    • Han Y.
    • Trevisi E.
    Effect of a commercial bentonite clay (smectite clay) on dairy cows fed aflatoxin-contaminated feed.

    Dairy. 2020; 1135-153

    • Greco D.
    • D’Ascanio V.
    • Treglia A.
    • Santovito E.
    • Zanelli L.
    • Avantaggiato G.
    Development and in vitro efficacy assessment of a new multi-mycotoxins adsorbing additive.

    Toxins (Basel). 2017; 9276

    • Huey B.
    • Hall J.
    Hypervariable DNA fingerprinting in Escherichia coli: minisatellite probe from bacteriophage M13.

    J. Bacteriol. 1989; 171 (2565332)2528-2532

    • ISO
    19344: Milk and milk products. Starter cultures, probiotics and fermented products. Quantification of lactic acid bacteria by flow cytometry.

    International Organization for Standardization2015

    • Jiang Y.
    • Ogunade I.M.
    • Pech-Cervantes A.A.
    • Fan P.X.
    • Li X.
    • Kim D.H.
    • Arriola K.J.
    • Poindexter M.B.
    • Jeong K.C.
    • Vyas D.
    • Adesogan A.T.
    Effect of sequestering agents based on a Saccharomyces cerevisiae fermentation product and clay on the ruminal bacterial community of lactating dairy cows challenged with dietary aflatoxin B1.

    J. Dairy Sci. 2020; 103 (31785878)1431-1447

    • John J.S.
    SeqPrep: Tool for stripping adaptors and/or merging paired reads with overlap into single reads.
    https://githubcom/jstjohn/SeqPrep

    Date: 2011
    Date accessed: November 20, 2020
    • Li W.
    • Godzik A.
    Cd-hit: A fast program for clustering and comparing large sets of protein or nucleotide sequences.

    Bioinformatics. 2006; 22 (16731699)1658-1659

    • Ma G.
    • Chen Y.
    Polyphenol supplementation benefits human health via gut microbiota: A systematic review via meta-analysis.

    J. Funct. Foods. 2020; 66103829

    • Magurran A.E.
    Measuring biological diversity.

    in: McGill B.J. Magurran A. Biological Diversity: Frontiers in Measurement and Assessment. Blackwell Publishing2004105

    • Mallardi D.
    • Tabasso C.
    • Piemontese P.
    • Morandi S.
    • Silvetti T.
    • Biscarini F.
    • Cremonesi P.
    • Castiglioni B.
    • Pica V.
    • Stuknyte M.
    • De Noni I.
    • Amato O.
    • Liotto N.
    • Mosca F.
    • Roggero P.
    Inoculation of mother’s own milk could personalize pasteurized donor human milk used for feeding preterm infants.

    J. Transl. Med. 2021; 19 (34627277)420

    • Mancini A.
    • Rodriguez M.C.
    • Zago M.
    • Cologna N.
    • Goss A.
    • Carafa I.
    • Tuohy K.
    • Merz A.
    • Franciosi E.
    Massive survey on bacterial–bacteriophages biodiversity and quality of natural whey starter cultures in Trentingrana cheese production.

    Front. Microbiol. 2021; 12 (34194413)678012

    • MoH
    National Standard No. 2761–2011, Maximum residue level of mycotoxin in food-national regulations for food safety.

    Official Document of Ministry of Health (MoH) of China2011

    • Morandi S.
    • Battelli G.
    • Silvetti T.
    • Goss A.
    • Cologna N.
    • Brasca M.
    How the biodiversity loss in natural whey culture is affecting ripened cheese quality? The case of Trentingrana cheese.

    Lebensm. Wiss. Technol. 2019; 115108480

    • Morandi S.
    • Biscarini F.
    Effect of using mycotoxin-detoxifying agents in dairy cattle feed on natural whey starter biodiversity. In Journal of Dairy Science. Zenodo.
    • Morandi S.
    • Brasca M.
    • Lodi R.
    Technological, phenotypic and genotypic characterisation of wild lactic acid bacteria involved in the production of Bitto PDO Italian cheese.

    Dairy Sci. Technol. 2011; 91341-359

    • Morandi S.
    • Cremonesi P.
    • Capra E.
    • Silvetti T.
    • Decimo M.
    • Bianchini V.
    • Alves A.C.
    • Vargas A.C.
    • Costa G.M.
    • Ribeiro M.G.
    • Brasca M.
    Molecular typing and differences in biofilm formation and antibiotic susceptibilities among Prototheca strains isolated in Italy and Brazil.

    J. Dairy Sci. 2016; 99 (27236754)6436-6445

    • Moser A.
    • Schafroth K.
    • Meile L.
    • Egger L.
    • Badertscher R.
    • Irmler S.
    Population dynamics of Lactobacillus helveticus in Swiss gruyere-type cheese manufactured with natural whey cultures.

    Front. Microbiol. 2018; 9 (29670601)637

    • O’Sullivan O.
    • Cotter P.D.
    Microbiota of raw milk and raw milk cheeses.

    in: Cheese. Academic Press2017301-316

    • Paulson J.N.
    • Stine O.C.
    • Bravo H.C.
    • Pop M.
    Differential abundance analysis for microbial marker-gene surveys.

    Nat. Methods. 2013; 101200-1202

    • Quast C.
    • Pruesse E.
    • Yilmaz P.
    • Gerken J.
    • Schweer T.
    • Yarza P.
    • Peplies J.
    • Glöckner F.O.
    The SILVA ribosomal RNA gene database project: Improved data processing and web-based tools.

    Nucleic Acids Res. 2013; 41 (23193283)D590-D596

    • R Core Team
    A Language and Environment for Statistical Computing.

    R Core Team2018

    • Rodrigues R.O.
    • Rodrigues R.O.
    • Ledoux D.R.
    • Rottinghaus G.E.
    • Borutova R.
    • Averkieva O.
    • McFadden T.B.
    Feed additives containing sequestrant clay minerals and inactivated yeast reduce aflatoxin excretion in milk of dairy cows.

    J. Dairy Sci. 2019; 102 (31030928)6614-6623

    • Rossetti L.
    • Fornasari M.E.
    • Gatti M.
    • Lazzi C.
    • Neviani E.
    • Giraffa G.
    Grana Padano cheese whey starters: Microbial composition and strain distribution.

    Int. J. Food Microbiol. 2008; 127 (18620769)168-171

    • Shannon C.
    A mathematical theory of communication.

    Bell Syst. Tech. J. 1948; 27379-423

    • Simpson E.H.
    Measurement of diversity.

    Nature. 1949; 163688

    • Smith B.
    • Wilson J.B.
    A consumer’s guide to evenness indices.

    Oikos. 1996; 7670

    • Tabasco R.
    • Sánchez-Patán F.
    • Monagas M.
    • Bartolomé B.
    • Moreno-Arribas M.V.
    • Peláez C.
    • Requena T.
    Effect of grape polyphenols on lactic acid bacteria and bifidobacteria growth: Resistance and metabolism.

    Food Microbiol. 2011; 28 (21839384)1345-1352

    • Tikhonov V.V.
    • Yakushev A.V.
    • Zavgorodnyaya Y.A.
    • Byzov B.A.
    • Demin V.V.
    Effects of humic acids on the growth of bacteria.

    Eurasian Soil Sci. 2010; 43305-313

    • Verrillo M.
    • Salzano M.
    • Cozzolino V.
    • Spaccini R.
    • Piccolo A.
    Bioactivity and antimicrobial properties of chemically characterized compost teas from different green composts.

    Waste Manag. 2021; 120 (33290882)98-107

    • Weatherly M.E.
    • Pate R.T.
    • Rottinghaus G.E.
    • Roberti Filho F.O.
    • Cardoso F.C.
    Physiological responses to a yeast and clay-based adsorbent during an aflatoxin challenge in Holstein cows.

    Anim. Feed Sci. Technol. 2018; 235147-157

    • Wickham H.
    ggplot2: Elegant graphics for data analysis.

    Springer-Verlag2016

    • Wilkinson M.G.
    Flow cytometry as a potential method of measuring bacterial viability in probiotic products: A review.

    Trends Food Sci. Technol. 2018; 781-10

    • Williams L.B.
    Natural antibacterial clays: Historical uses and modern advances.

    Clays Clay Miner. 2019; 677-24

    • Xiong J.L.
    • Wang Y.M.
    • Zhou H.L.
    • Liu J.X.
    Effects of dietary adsorbent on milk aflatoxin M1 content and the health of lactating dairy cows exposed to long-term aflatoxin B1 challenge.

    J. Dairy Sci. 2018; 101 (30055921)8944-8953

    • Yilmaz P.
    • Parfrey L.W.
    • Yarza P.
    • Gerken J.
    • Pruesse E.
    • Quast C.
    • Schweer T.
    • Peplies J.
    • Ludwig W.
    • Glöckner F.O.
    The SILVA and “All-species Living Tree Project (LTP)” taxonomic frameworks.

    Nucleic Acids Res. 2014; 42 (24293649)D643-D648

    • Zago M.
    • Bardelli T.
    • Rossetti L.
    • Nazzicari N.
    • Carminati D.
    • Galli A.
    • Giraffa G.
    Evaluation of bacterial communities of Grana Padano cheese by DNA metabarcoding and DNA fingerprinting analysis.

    Food Microbiol. 2021; 93 (32912585)103613