- Email: [email protected]
Seasonal variances in bacterial microbiota and volatile organic compounds in raw milk
International Journal of Food Microbiology 267 (2018) 70–76
Contents lists available at ScienceDirect
International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro
Seasonal variances in bacterial microbiota and volatile organic compounds in raw milk
T
⁎
Beata Nalepaa, , Magdalena Anna Olszewskaa, Lidia Hanna Markiewiczb a
Department of Industrial and Food Microbiology, Faculty of Food Science, University of Warmia and Mazury in Olsztyn, pl. Cieszyński 1, 10-726 Olsztyn, Poland Department of Immunology and Food Microbiology, Institute of Animal Reproduction and Food Research, Polish Academy of Sciences, ul. Tuwima 10, 10-748 Olsztyn, Poland b
A R T I C L E I N F O
A B S T R A C T
Keywords: Raw milk Microbiota VOCs PCR-DGGE HS-GC Seasonal variation
The aim of this study was to define the composition of microbiota and the volatile organic compounds (VOCs) in samples of raw milk collected for 22 months between 2012 and 2014 originated from north-eastern region of Poland. The results revealed that the VOCs profile changed with respect to the season of milk collection, and milk collected in autumn was characterized by a higher content of acetic acid (C2), propionic acid (C3) and valeric acid (C5), whereas spring was characterized by a frequent presence of acetone (Ac), ethanol (Et) and ethyl acetate (EtAc). Bacterial species composition changed considerably within the tested period and some bacterial species/groups occurred seasonally, e.g. L. helveticus (summer), L. casei (winter). The results show usefulness of the applied techniques (PCR-DGGE and HS-GC) and data analysis (PCA, correlation coefficients) methods in characterizing the raw milk quality intended for dairy production.
1. Introduction The aim of all technological treatments applied in food production is to make sure that the customer is safe and to obtain products of the highest quality. In the dairy industry, the microbiological quality of a final product is determined by the quality of its main material - raw milk. The hygienic requirements (Commission Regulation (EC) No 1662/, 2006) define only an acceptable number of bacteria in 1 cm3 of raw milk, not specifying the qualitative content of its microbiota, while those are some biochemical groups or microorganism species that play a crucial role in determining the quality and shelf life of milk products. Lactic acid bacteria (LAB), such as Lactococcus, Lactobacillus, Enterococcus may cause the sourness of milk (Jarosińska et al., 2014) or may function as non-starter lactic acid bacteria (NSLAB) (Settanni and Moschetti, 2010). Coliforms cause organoleptic changes and early cheese blowing (Jarosińska et al., 2014). Psychrotrophic bacteria (Pseudomonas, Aeromonas, Bacillus) and the enzymes they produce lead to unfavourable physicochemical and organoleptic changes. By causing protein decomposition, proteolytic enzymes alter milk texture (thickening, gelification) and change taste and smell (e.g., putrid, bitter). By hydrolysing milk fat, lipolytic enzymes cause a rancid, soap-like, tallowy taste and smell of milk (Adamiak et al., 2015; Hantsis-Zacharov and Halpern, 2007). By surviving thermal processing, spore-forming bacteria (Bacillus, Clostridium) may cause milk spoilage, including
⁎
rotting, a bitter aftertaste, a rancid taste, or late cheese blowing (Klijn et al., 1995). Thus, the microorganisms that occur in milk are of decisive significance in determining properties of a final product. However, performing the examinations aimed at identifying all bacterial groups and species in the material with standard methods is relatively laborious and time-consuming. Therefore, it is becoming more frequent to apply the tools of molecular biology combined with modern instrumental methods. The molecular methods (e.g., PCR-DGGE) enable a precise identification of species diversity for the microorganisms in a studied sample (Franciosi et al., 2009; Nalepa and Markiewicz, 2017) with no cultivation needed. The instrumental techniques, e.g., HeadSpace - Gas Chromatography (HS-GC) (Ayad et al., 1999) make it possible to trace metabolic changes performed by those microorganisms, e.g., identifying the content of volatile organic compounds (VOCs) produced. The presence of bacteria in raw milk and the VOCs they produce determines the taste and smell of a final product, especially ripened cheeses. Therefore, the aim this study was to identify microbiota and the VOCs in the pooled raw milk. Differences and some potential links between those data have been determined in order to evaluate the quality of raw milk intended for dairy production.
Corresponding author. E-mail address: [email protected] (B. Nalepa).
https://doi.org/10.1016/j.ijfoodmicro.2017.12.024 Received 23 August 2017; Received in revised form 15 November 2017; Accepted 21 December 2017 Available online 24 December 2017 0168-1605/ © 2017 Elsevier B.V. All rights reserved.
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
result of thermostating of 5 cm3 of milk at 70 °C for 40 min in 22-mm tightly closed vials. Then, the sample was pressurized for 1 min and injected into the column (split ratio of 2:1) in 0.08 min. The temperatures of the needle and the transfer line were 100 °C and 120 °C, respectively. The volatile compound separation was performed on the HPINNOWAX column (60 m × 1.00 μm × 0.537 mm) by Agilent Technologies, USA, applying the following thermal gradient: 40 °C (5 min) → ΔT 10 °C/min → 220 °C (5 min). The temperature of the injector and the FID was 230 °C. The following gases were applied for the analysis: helium (5 cm3/min), synthetic air (400 cm3/min), and hydrogen (40 cm3/min) with the pressure of 130 kPa in the column (MikšKrajnik, 2013; internal study).
2. Materials and methods 2.1. Raw milk samples The pooled raw milk samples from a dairy plant in the Region of Warmia and Mazury in north-eastern Poland were studied. The samples were collected into sterile vessels on a monthly basis for the period of twenty-two months: from September 2012 to July 2014. A period from July to September was consider summer, months from October to December were autumn, January to March were winter and April to June were consider spring. 2.2. DNA isolation
2.6. Raw milk microbiota biodiversity The total DNA from the raw milk samples was obtained using the Genomic Mini AX FOOD kit (A@A Biotechnology, Gdańsk, Poland) in accordance with the manufacturer's instructions. The isolated DNA was stored at −80 °C for further analyses.
The DGGE profiles for the raw milk samples were compared with the markers that had been constructed in previous research (Nalepa and Markiewicz, 2017) and they included 24 referential strains: P. freudenreichii ssp. shermanii DSM 4902, L. lactis ssp. lactis DSM 4366, L. mesenteroides DSM 20346, P. thoenii DSM 20276, L. acidophilus DSM 9126, L. plantarum ATCC 8014, L. brevis DSM 1267, L. casei ATCC 334, L. delbrueckii DSM 20080, L. fermentum DSM 200052, L. helveticus DSM 20075, E. coli ATCC 8739, E. aerogenes ATCC 13048, E. cloacae ATCC 13047, C. freundii ATCC 8090, E. faecalis ATCC 29212, B. subtilis ATCC 6051, C. butyricum ATCC 10702, C. tyrobutyricum ATCC 2637, C. perfringens ATCC 13124, L. monocytogenes ATCC BAA-751, S. thermophilus ATCC 19258, S. xylosus ATTC 29971, and S. aureus ATCC 43300. Basing on that, the presence of particular bacterial species was detected and assigned scores from 0 to 4 depending on their band brightness. Also, the 1D analysis of electrophorograms was performed using the Doc-It LS Image Analysis application (UVP Ltd., UK). The data obtained in that way were used to calculate the Shannon–Wiener index that defines biodiversity and is expressed by the following formula: (Sienkiewicz, 2010),
2.3. Polymerase chain reaction (PCR) The reaction mixture (25 μL) contained: 20 mM of Tris-HCl (pH 8.4), 50 mM of KCl, 3 mM of MgCl2, 50 μM of dNTPs, 5 pM of every primer, 1.25 U of Taq polymerase (Fermentas), and 20–40 ng of the DNA. The following primers were used: U968-GC (5′-CGC CCG GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGGG AAC GCG AAG AAC CTT AC-3′) and L1401-r (5′-CGG TGT GTA CAA GAC CC-3′) amplifying a fragment of region V6-V8. Amplification was carried out in the MJ Mini Gradient Thermal Cycler (BioeRad, Poland). The PCR profile was as follows: initial denaturation at 94 °C for 5 min followed by 35 cycles of: denaturation at 94 °C for 30 s, annealing at 56 °C for 30 s and extension at 68 °C for 40 s. Final extension at 68 °C for 7 min. (Randazzo et al., 2010). The PCR product with the estimated size of 450 bp was analyzed by electrophoresis on 1% agarose gel in 0.5 × TBE buffer in the MultiSub Choice system (Clever Scientific Ltd., UK).
S
H ′ = − ∑ pi ln pi i=1
2.4. Denaturing gradient gel electrophoresis (DGGE)
where pi is the proportion of individuals in species i to the total number of individuals in the community.
The PCR products were analyzed using the electrophoresis in a gradient of denaturing agents. The electrophoresis was conducted in the 8% polyacrylamide gel (acrylamide:bisacrylamide 37.5:1) with the gradient of denaturing agent (urea) ranging from 35% to 57.5%. The electrophoresis was carried out in the 0.5 × TAE buffer at the temperature of 60 °C under the voltage of 85 V for 16 h in the DCode Universal Mutation System (BioRad, Poland) (Randazzo et al., 2010). On each gel a marker set was run to enable the identification of DGGE bands (Nalepa and Markiewicz, 2017). The gels were stained in the SybrGreen I (1,10,000) solution for 15 min and then they were archived using G-Box (Syngen, Poland).
Seasonal variations in the VOC profiles of raw milk and associations between the VOC profiles and the microbial compositions of the collected samples were determined in the principal component analysis (PCA). The direction and strength of these correlations were determined basing on the values of the Pearson correlation coefficient (r) at the significance level of p < 0.05. The data were processed in Statistica v. 12.5 (StatSoft Inc., Tulsa, USA).
2.5. Detecting volatile organic compounds using gas chromatography
3. Results
Detection of selected volatile compounds: aldehydes (acetaldehyde), ketones (acetone, diacetyl, acetoin), alcohols (methanol, ethanol), esters (ethyl acetate, ethyl propionate, ethyl butyrate), fatty acids C2 - C7 (acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, valeric acid, isocaproic acid, caproic acid, heptanoic acid) was performed in duplicate using headspace gas chromatography (HS-GC) with the Headspace Turbomatrix 40 autosampler (Perkin Elmer, USA) and the Clarus 500 gas chromatograph (Perkin Elmer, USA) with a flame ionization detector (HS-GC-FID). The chromatograph was calibrated for a quantitative detection [μg/cm3] of the selected metabolites using some external standards. Calibration curves were made for every compound in the respective concentration scope. Equilibrium between the sample and the headspace was obtained as a
3.1. Raw milk microbiota
2.7. Statistical analysis
Upon comparing the DGGE profiles of the raw milk samples collected from September 2012 to July 2014, 3 to 12 bacterial species out of the 24 identified ones were detected (Table 1). The presence of 3 species was detected in 6 milk samples, 4 species were detected in 1 sample, 5 - in 2 samples, 6 - in 3 samples, 7 - in 5 samples, and 9 - in 2 samples. 8, 10, and 12 bacterial species were identified in the remaining 3 samples. The values of the Shannon-Wiener index (H′) (Table 1) correspond to the biological diversity of the collected raw milk samples. In 54.5% milk samples, the H′ indexes ranged from 1.3693 to 1.9782 (presence of 5–8 species) regardless of the analyzed season. In six (27%) milk samples the biological diversity was very low 71
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
(Fig. 1), as PCR-DGGE is a semi-quantitative method in which the amount of the DNA isolated from cells correlates with the amount of amplicons and their brightness. Applying that type of assessment, it has been determined that the most frequent bacteria included those of the P. thoenii and E. aerogenes species occurring in 15 of 22 samples (68%) followed by S. thermophilus/S. xylosus in 14 samples (63.6%). As for LAB, L. mesenteroides was the most frequent - in 13 samples (59.1%) and its presence was not common for neither of the analyzed seasons. In 7 samples (31.8%), L. brevis was identified, similarly as L. acidophilus. The presence of L. brevis was characteristic of autumn 2013 and winter 2014, and the L. acidophilus was present in the majority of the samples from summer 2012 to summer 2013. The presence of the L. lactis was identified in 5 samples (22.7%) and it was common for the summer/ autumn 2013 season. Out of the remaining LAB, L. plantarum (5 samples), L. casei (2 samples) and L. helveticus (2 samples) were identified sporadically and randomly. However, L. delbrueckii and L. fermentum were not detected in any of the examined samples (Fig. 1). Moreover, the fecal species (E. coli, E. aerogenes, E. cloacae, C. freundii and E. faecalis) were detected in 16 (72.2%) samples. E. aerogenes was the most frequent species as it was present in all the samples from the summer 2013 to summer 2014 season. Additionally, in the summer/autumn 2013 season, the presence of E. aerogenes is accompanied by occurrences of E. coli, E. cloacae, C. freundii and E. faecalis (Fig. 1). The sporeforming bacteria (B. subtilis, C. butyricum, C. tyrobutyricum and C. perfringens) were present in 17 (77.3%) samples. The presence of C. tyrobutyricum was identified in 9 samples (41%) of which 7 were from the summer 2012 to spring 2013 season. The presence of C. butyricum was characteristic of the samples from summer/autumn 2013 and that of C. perfringens for spring 2014. However, C. perfringens also occurred randomly in 6 other samples (Fig. 1). As for the remaining species, S. aureus was detected most often − in 6 samples (27.3%) from the summer 2012 to winter 2013 season (Fig. 1). When considering the results according to the collection years, it has been observed that all the samples of raw milk from 2012 contained L. mesenteroides, sporeforming C. tyrobutyricum, and S. aureus while the majority of samples from 2013 to 2014 contained fecal bacteria, such as E. aerogenes and
Table 1 The number of bacterial species at different times of years and the Shannon-Wiener indexes calculated from DGGE patterns of the raw milk from the 22-month period between 2012 and 2014. Seasona
Number of species
Shannon-Wiener index
Spring 2013 Spring 2013 Spring 2014 Spring 2014 Spring 2014 Summer 2012 Summer 2013 Summer 2013 Summer 2013 Summer 2014 Autumn 2012 Autumn 2012 Autumn 2012 Autumn 2013 Autumn 2013 Autumn 2013 Winter 2013 Winter 2013 Winter 2013 Winter 2014 Winter 2014 Winter 2014
3 6 7 4 3 7 12 3 9 5 6 6 7 10 9 7 5 8 3 5 3 8
0.5254 1.5193 1.8013 0.9294 0.9355 1.6593 2.2064 0.8757 2.0561 1.3693 1.4782 1.6383 1.7421 2.1649 2.1135 1.8258 1.4899 1.7605 0.8888 1.5403 0.8984 1.9782
a Spring: April, May, June. Summer: July, August, September. Autumn: October, November, December. Winter: January, February, March.
and then the H′ indexes ranged from 0.5254 to 0.9355 (presence of 3–4 species). Those were the samples collected mostly in the winter/spring season. In the four remaining samples (18.1%), however, the biodiversity was high. The values of the H′ indexes ranged from 2.0561 to 2.2064 (presence of 9–12 species). Those samples were collected in the summer-autumn season (Table 1). Bacterial species in milk samples were identified on the basis of DGGE band position. Band brightness was given a score of 0 to 4
Fig. 1. Presence of microorganisms in raw milk determined on the basis of PCR-DGGE patterns and band intensities in a scale from 0 (lack of a band) to 4 (the highest intensity). Samples of raw milk were collected in 22-month period between 2012 and 2014.
72
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
Fig. 2. The dendrogram showing a dissimilarity measure between the VOCs (A) and the PCA biplot of seasonal variables and volatile compounds data related to factor 1 (F1) and factor 2 (F2) (B). Ketones: Ac - acetone, Di - diacetyl, Acet - acetoin; Alcohols: Me - methanol; Et - ethanol; Esters: EtAc - ethyl acetate; Fatty acids: C2 - acetic acid, C3 - propionic acid, C4iso - isobutyric acid, C5iso - isovaleric acid, C5 - valeric acid, C6 - caproic acid; C7 - heptanoic acid.
Table 3 Pearson correlation coefficients (r) between selected microorganisms and those VOCs which were grouped into two clusters based the PCA (Fig. 2).
Table 2 The content (sum) of volatile organic compounds (VOCs) in the raw milk from the 22month period between 2012 and 2014. Seasona
Spring 2013 Spring 2013 Spring 2014 Spring 2014 Spring 2014 Summer 2012 Summer 2013 Summer 2013 Summer 2013 Summer 2014 Autumn 2012 Autumn 2012 Autumn 2012 Autumn 2013 Autumn 2013 Autumn 2013 Winter 2013 Winter 2013 Winter 2013 Winter 2014 Winter 2014 Winter 2014
Sum of VOCs (μg/cm3) Aldehydesb
Esters
Alcohols
Ketones
Fatty acids
Total
– – – – – – – – – – – – – – – – – – – – – –
0.20 – – 0.30 0.20 – – – – 0.10 – – – – – – – – – – – –
39.70 6.70 23.80 15.05 19.25 – 16.00 – – 90.05 – – – – – – – – – – – –
7.10 5.50 4.10 3.15 4.40 – 3.60 – 7.15 5.10 – – – – – – – – – – – –
46.60 92.70 23.35 46.70 46.60 260.35 46.70 40.80 178.15 77.95 117.05 159.20 203.35 900.85 471.35 91.60 – 425.10 107.65 – – 166.85
93.60 104.90 51.25 65.20 70.45 260.35 66.30 40.80 185.30 173.20 117.05 159.20 203.35 900.85 471.35 91.60 – 425.10 107.65 – – 166.85
VOCs (spring) Ac Et −0.26 −0.19 −0.18 −0.05 −0.01 −0.16 −0.35 −0.19
EtAc −0.17 −0.22 0.28 −0.34
−0.01 −0.17 0.32 0.27 0.39 0.25 0.29
−0.17 −0.27 −0.12 −0.20 0.61 −0.05 0.51
−0.07 0.01 0.34 0.27 0.20 0.20 0.00
Bacterial species E. cloacae L. plantarum C. butyricum S. thermophilus/S. xylosus E. coli L. brevis B. subtilis L. lactis L. mesenteroides E. aerogenes C. perfringens
VOCs (autumn) C2 C3 0.65 0.74 0.40 0.62 0.63 0.49 0.31 0.38
C5 0.48 0.61 0.28 0.16
0.23 0.13 0.23 0.30 0.05 −0.03 0.15
−0.11 0.02 0.23 0.16 −0.17 −0.05 0.12
0.15 0.26 0.08 0.18 −0.26 −0.06 −0.13
seasons, it was found out that the highest diversity of volatile compounds (however, in the lowest amounts) occurred in the spring seasons. Esters, ketones, alcohols, and fatty acids were identified in the milk samples collected in this season and their average amounts were 0.14, 4.85, 20.9, and 51.19 μg/cm3, respectively (Table 2). In the cold seasons (autumn and winter), only free fatty acids were detected. For the season of autumn 2012, C3, C4iso, C6, and C7 were identified in the amounts ranging from 117 to 203 μg/cm3 while for the autumn 2013 season - C2, C3, C5 and C6 (Supplementary Table 1) in the amounts ranging from 91 to 900 μg/cm3 (Table 2). No presence of the examined volatile compounds was detected in a half of the samples from the winter seasons (1 sample from winter 2013 and 2 samples from winter 2014). C2 and C3 in the amounts ranging from 107 to 425 μg/cm3 were identified in two samples from winter 2013 and in one sample from winter 2014, C3 had been detected in the amount of 166 μg/cm3 (Table 2). The highest diversity of the VOC contents was identified within the samples collected in the summer seasons. In two samples, the fatty acids only were present. The sample from summer 2012 contained the largest amount of fatty acids (260 μg/cm3), while one sample from summer 2013 contained them in the lowest amount (40 μg/cm3) and those were C3, C5izo, C5, C7, and C5izo, respectively (Table 2; Supplementary Table 1). Apart from the fatty acids, ketones (summer 2013), alcohols and ketones (summer 2013), as well as alcohols, ketones, and esters (summer 2014) were detected in very varied amounts in the remaining 3 samples collected in the summers. However, no acetaldehyde was detected in any of the examined milk samples (Table 2; Supplementary Table 1).
a Spring: April, May, June. Summer: July, August, September. Autumn: October, November, December. Winter: January, February, March. b Aldehydes: acetaldehyde. Esters: ethyl acetate, ethyl propionate, ethyl butyrate. Alcohols: methanol, ethanol. Ketones: acetone, diacetyl, acetoin. Fatty acids: acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, valeric acid, isocaproic acid, caproic acid, heptanoic acid.
NSLAB, L. brevis (Fig. 1). 3.2. Volatile organic compounds (VOCs) of raw milk Contents of selected volatile compounds in the groups of aldehydes, esters, ketones, and fatty acids C2 − C7 have been presented in Table 2 and Supplementary Table 1. The total content of volatile compounds in the examined samples of raw milk was 170.65 μg/cm3 on average. The highest content of VOCs was identified in samples collected in the autumn seasons. Their average content was 323.90 μg/cm3 while in the summer and winter seasons it was 145.19 and 116.60 μg/cm3 on average. The lowest contents of VOCs were detected in the milk samples from the spring seasons and they were 77.08 μg/cm3 on average (Table 2). When analyzing the VOC contents in particular groups and 73
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
collection (Beresford et al., 2001). Using RAPD-PCR, Ercolini et al. (2009) have shown that as for the fecal bacteria (of the Enterobacteriaceae family) those were Hafnia alvei, Serratia marcescens, Citrobacter freundii, E. coli, and Klebsiella that were most common in their raw milk samples while in our case those were E. aerogenes as well as, sporadically, C. freundii, E. coli, and E. cloacae. When considering the results seasonally, an increased incidence of coliforms (E. coli, E. cloacae, E. aerogenes, and C. freundii) and the L. lactis in the summer/ autumn 2013 season has been observed in our research. As for the LAB, those were L. brevis, L. plantarum, and L. acidophilus that occurred more frequently in the autumn/winter seasons. Seasonality in occurrences of some bacterial species in raw milk has also been shown by Czerwińska and Piotrowski (2011). When doing their research with culturing methods, they found out that in the summer/autumn seasons the number of bacteria of the Enterobacteriaceae family (including coliforms) in milk was higher than in other seasons and the number of lactococci was at its peak in autumn. As for the NSLAB, they identified the presence of L. casei in the winter/spring seasons. Similarly to our research, Ercolini et al. (2009) have also shown the presence of grampositive bacteria of the Bacillus and Staphylococcus in raw milk samples. Applying molecular methods, such as PCR-DGGE or next-generation sequencing (NGS), makes it possible to detect and identify not only dominant bacterial species/groups but also those that occur sporadically or in low-number populations. Using NSG, Escobar-Zepeda et al. (2016) have shown 3 dominant bacterial species in cheese made of raw milk. Those were L. plantarum, L. mesenteroides, and Weissella paramesenteroides, which made 80% of the whole population in cheese. The Aerococcus, Enterococcus, Lactococcus, and Staphylococcus made less than 10% of the population and 78 other bacterial genera made less than 1% which included Propionibacterium, Bifidobacterium, Corynebacterium, and others. By applying PCR-DGGE to the raw milk samples in our research, the domination of L. mesenteroides, Propionibacterium, and Staphylococcus has been identified. Bystroń et al. (2001) have detected the presence of staphylococci in 24% of their raw milk samples, of which the most frequently identified ones included S. aureus and S. xylosus, similarly to our research. Bacteria that occur in the material, particularly in raw milk, may cause its spoilage and worthlessness in production of other dairy products. It is caused by a possible growth of bacteria in milk, even during cold storage, and the release of proteolytic and lipolytic enzymes as well as volatile organic compounds (VOCs) that cause disturbances in the structure, texture, appearance, taste, and smell of a final product (Alothman et al., 2017). Thus, identifying VOC concentrations and occurrences of particular bacterial species may provide valuable information on the influence of microorganisms on the course of production processes and the quality of final products. Identification of VOC concentrations and their changes in raw milk or milk inoculated with single or mixed bacterial cultures has been performed in numerous studies (Alothman et al., 2017; Hettinga et al., 2008; ŁaniewskaTrokenheim et al., 2010; Toso et al., 2002; Villeneuve et al., 2013). When examining raw milk, Hettinga et al. (2008) have detected only 3 compound groups: aldehydes, ketones, and sulfur compounds. Free fatty acids (FFAs) occurred sporadically and in trace amounts. In our research in the milk collected in north-eastern Poland, the dominant class of volatile compounds includes free fatty acids C2 - C7, the highest concentration of which has been identified for the autumn seasons and the lowest one - for the spring seasons. Further on, they included alcohols, ketones, and esters in small concentrations; however, no presence of aldehydes was detected. Toso et al. (2002) have shown that in the milk collected in north-eastern Italy, the dominant class of compounds included ketones and then aldehydes, alcohols, carbohydrates, esters, sulfur compounds, and terpenes. When considering the obtained results in the seasonal aspect, the highest VOC diversity has been noticed in the spring seasons and the lowest one - in the autumn/winter seasons. However, autumn/winter seasons were characterized by a higher concentration of fatty acids which may result from the method of
3.3. PCA results A principal component analysis (PCA) has been conducted to determine if volatile organic compounds (VOCs) detected in the milk from the 22-month period between 2012 and 2014 relate with seasons of milk collection. As demonstrated by Fig. 2A first, a dissimilarity between acetic acid (C2), propionic acid (C3) and the remaining VOCs has been noticed. This corresponded to the C2 and C3 amounts of 52.8–437.1 μg/cm3 and 50.8–362.8 μg/cm3, respectively and those were the highest concentrations determined in milk in comparison to the rest of VOCs which ranged from the lowest level of 0.1 μg/cm3 to the highest level of 117,6 μg/cm3 (Supplementary Table 1). As shown on Fig. 2B, the PCA biplot mainly presents two clusters, I and II, which are a collection of spring and autumn samples, respectively. This PCA biplot comprised 37.77% of the total variance and based on the values of loading coefficients with F 1, the separation between C2, C3 and C5 vs. acetone (Ac), ethanol (Et) and ethyl acetate (EtAc) was revealed (Fig. 2B). C2, C3 and C5 are variables with the highest negative loading: − 0.48, − 0.67, and, − 0.42, respectively whereas the remaining VOCs are variables with the highest positive loading: 0.90, 0.75, and 0.69, respectively. As results, C2, C3 and C5 were attributed mainly to the autumn and the Ac, Et and EtAc to spring season of milk collection (Fig. 2B). In contrast, e.g., acetoin (Acet) and isovaleric acid (C5iso) are variables with the positive loading with F 2, however these coefficients were weak (0.34, 0.31, respectively). Besides, if they were detected, they were predominantly found in the summer samples. Additionally, based on Pearson correlation coefficients (r) microorganisms detected in milk samples were associated with certain VOCs. As shown in Table 3, differences in strengths and directions of relationship between VOCs vs. microorganisms have been revealed. It is worth noticing that, e.g.: E. cloacae is positively coupled to C2 (r = 0.65), C3 (r = 0.74) and C5 (r = 0.48), and all the coefficients are significant. Moreover, the r values revealed significant relationships between L. plantarum and C3 (r = 0.62), C5 (r = 0.61), as well as between C. butyricum vs. C2 (r = 0.63) and C3 (r = 0.49). Furthermore, the L. mesenteroides and C. perfringens may have contributed to EtAc (r = 0.61 and 0.51), since these correlations were moderate and significant. 4. Discussion Milk is the most important material in the dairy industry and its quality, particularly microbiological one, is of decisive significance for products made of it. In the case of raw milk, dairies plants determine the total bacteria count only that cannot exceed 1 × 105 cfu/cm3 (Commission Regulation (EC) No 1662/, 2006). All the raw milk samples examined for this study complied with that requirement (data not shown). However, as it has been shown by the PCR-DGGE analysis, the bacterial species composition in the raw milk varied significantly within 22 months of sample collection and some bacterial groups/species occurred seasonally (Fig. 1). Application of DGGE markers developed in previous research (Nalepa and Markiewicz, 2017) enabled a simple interpretation of the obtained DGGE profiles for the particular milk samples as well as an immediate identification of the species without the need to cut the bands out and sequencing. As for cocci, the presence of S. thermophilus and L. mesenteroides was identified most frequently in the milk. The propionate acid bacteria, P. thoenii were also common. Among NSLAB, L. brevis, L. acidophilus and L. plantarum were the most frequent ones. Using the same technique and primers, Aquilanti et al. (2013) have examined Italian artisan cheese made of raw milk. They have shown that the S. thermophilus, L. lactis and L. mesenteroides were the most frequent in their samples and as for the lactobacilli, those were species of the L. casei group. Other species of the NSLAB, such as L. acidophilus, L. helveticus, L. plantarum, L. brevis, and L. fermentum were detected less often. It is the material, which is raw milk, that is the source of those bacteria in fermented products and they get there mostly from the environment or the equipment used for milk 74
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
(PCR-DGGE and HS-GC) and data analysis (PCA, correlation coefficients) methods in characterizing the raw milk quality. The determined microbiological and VOCs profiles of raw milk may be of great importance in predicting possible disturbances in the dairy production and in planning of modifications of technological processes depending on the identified specific raw milk microbiota and the VOCs they produce. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ijfoodmicro.2017.12.024. a The correlation coefficients (r) written in bold are statistically significant (p ≤ 0.05). Ac - acetone; Et - ethanol; EtAc - ethyl acetate; C2 acetic acid; C3 - propionic acid; C5 - valeric acid. The range and strength of correlations: ± 0.00–0.15 - very low positive (negative) correlation; ± 0.16–0.29 - weak to low positive (negative) correlation; ± 0.30–0.49 - moderate to low positive (negative) correlation; ± 0.50–0.69 - moderate positive (negative) correlation; ± 0.70–0.89 strong positive (negative) correlation; ± 0.90–1.00 - very strong positive (negative) correlation. Shades of grey indicate only positive correlation coefficients (0.16 ≤ r ≤ 1.00).
feeding the cows in winter. At that time in Poland, cows are kept indoors and are fed with hay and silages. Villeneuve et al. (2013) report that in the milk of cows fed with hay and silages, there is more fat and, thus, more fatty acids. On the other hand, high concentrations of FFAs in some of the summer samples may be attributed to the metabolic activities of bacteria present in the milk. Aquilanti et al. (2013) suggest that short- and medium-chain FFAs are formed in three biochemical pathways, which are lactose fermentation, proteolysis, and lipolysis triggered by the enzymes exerted to milk by microorganisms. That is confirmed in the research by Alothman et al. (2017) who have identified VOC profiles in milk inoculated with the P. fluorescens and Chryseobacterium bacteria. Presence of those bacteria in milk increased concentrations of alcohols, esters, acetaldehyde, as well as acetic and butyric acids. The research by Łaniewska-Trokenheim et al. (2010) has shown that concentrations of ethanol, ketones, and acetic acid was increased in milk inoculated with L. acidophilus and L. crispatus. Thus, it is clearly visible that the profile of volatile compounds in milk depends not only on the method of cow feeding (season) but also on the amount and metabolic activity of microorganisms that get into the material e.g. during the process of milk collection. Knowledge of milk microbiota compositions and profiles of volatile compounds may be of tremendous significance for prediction of the quality of a final product so that any errors in operation processes would be avoided. However, comparing enormous amounts of detailed information obtained every day by various research methods in every phase of production is simply unfeasible. Here, it is the development of statistics that turns out to be very helpful. Applying multidimensional statistical analyses, such as principal component analysis (PCA) or partial least squares regression (PLS), makes it possible to notice trends and dependencies between numerous and very varied parameters. Many authors have applied those methods successfully. Toso et al. (2002) have applied PCA to assess the influence of fodder types used in cow feeding on profiles of volatile compounds and a possibility of raw milk qualification on that basis. The analysis has shown that nine volatile compounds is of discriminating character and basing on their presence and concentrations in milk the authors were able to assess the type of food used in cow feeding. Gallardo-Escamilla et al. (2005) have used PCA and PLS to establish a dependency between VOCs produced by various species of microorganisms and taste and smell of fermented milks made of various types of whey. In our research, applying PCA and correlation coefficients (r) has made it possible to show the seasonality in occurrences of particular classes of volatile compounds, as well as to attempt to correlate VOC occurrences with the presence of particular bacterial species. The samples collected between 2012 and 2014 were mainly grouped into two clusters (Fig. 2B). The first one was made of the samples from all the spring seasons and was characterized by the presence of alcohols (Et), ketones (Ac), and esters (EtAc). The second cluster was formed by the majority of the autumn seasons that were characterized mostly by the presence of fatty acids like C2, C3, C5. The highest diversity and no grouping has been shown in the case of the summer samples as they cannot be grouped with any VOCs and other seasonal samples. It may mean, e.g., a broadly understood lack of recurrence in the material collection and/or diversity of weather conditions in warmer months resulting in a presence of microorganisms responsible for further differences in quality of final products made from it. The analysis of dependencies between the volatile compounds and the bacteria has shown that a connection is possible between particular compounds and particular bacterial species. For example, amounts of ethyl acetate (EtAc) identified in the spring seasons are correlated with occurrences of C. perfringens. Some crucial correlation coefficients between dependencies in occurrences of acids, such as acetic, propionic, or valeric ones, and the presence of E. cloacae, L. plantarum, or C. butyricum may be connected with the samples collected in the autumn season. Summarizing, our results show usefulness of the applied techniques
Acknowledgements This study has been supported by the National Science Center, Poland (grant no. N N312 484140). References Adamiak, A., Górska, A., Mróz, B., 2015. Psychrotrophic bacteria in raw milk and dairy products. Żywność. Nauka. Technologia. Jakość 4 (101), 36–48 (in Polish). Alothman, M., Lusk, K.A., Silcock, P., Bremer, P.J., 2017. Comparing PTR-MS profile of milk inoculated with pure or mixed cultures of spoilage bacteria. Food Microbiol. 64, 155–163. http://dx.doi.org/10.1016/j.fm.2017.01.001. Aquilanti, L., Santarelli, S., Babini, V., Osimani, A., Clementi, F., 2013. Quality evaluation and discrimination of semi-hard and hard cheeses from the Marche region (Central Italy) using chemometric tools. Int. Dairy J. 29 (1), 42–52. http://dx.doi.org/10. 1016/j.idairyj.2012.11.001. Ayad, E.H.E., Verheul, A., de Jong, C., Wouters, J.T.M., Smit, G., 1999. Flavour forming abilities an amino acid requirements of Lactococcus lactis strains isolated from artisanal and non-dairy origin. Int. Dairy J. 9, 725–735. http://dx.doi.org/10.1016/ S0958-6946(99)00140-5. Beresford, T.P., Fitzsimons, N.A., Brennan, N.L., Cogan, T.M., 2001. Recent advances in cheese microbiology. Int. Dairy J. 11, 259–274. http://dx.doi.org/10.1016/S09586946(01)00056-5. Bystroń, J., Kosek-Paszkowska, K., Molenda, J., 2001. Occurrence of enterotoxigenic staphylococci in raw milk. Med. Weter. 57 (9), 645–648 (in Polish). Commission Regulation (EC) No 1662/2006 of 6 November 2006 amending Regulation (EC) No 853/2004 of the European Parliament and of the Council laying down specific hygiene rules for food of animal origin. Czerwińska, E., Piotrowski, W., 2011. Potential sources of milk contamination influencing its quality for consumption. Rocznik Ochrona Środowiska 13, 635–652 (in Polish). Ercolini, D., Russo, F., Ferrocino, I., Villani, F., 2009. Molecular identification of mesophilic and psychrotrophic bacteria from raw cow's milk. Food Microbiol. 26, 228–231. http://dx.doi.org/10.1016/j.fm.2008.09.005. Escobar-Zepeda, A., Sanchez-Flores, A., Baruch, M.Q., 2016. Metagenomic analysis of a Mexican ripened cheese reveals a unique complex microbiota. Food Microbiol. 57, 116–127. http://dx.doi.org/10.1016/j.fm.2016.02.004. Franciosi, E., Settanni, L., Cavazza, A., Poznański, E., 2009. Biodiversity and technological potential of wild lactic acid bacteria from raw cows' milk. Int. Dairy J. 19, 3–11. http://dx.doi.org/10.1016/j.idairyj.2008.07.008. Gallardo-Escamilla, F.J., Kelly, A.L., Delahunty, C.M., 2005. Influence of starter culture on flavor and headspace volatile profiles of fermented whey and whey produced from fermented milk. J. Dairy Sci. 88 (11), 3745–3753. http://dx.doi.org/10.3168/jds. S0022-0302(05)73060-5. Hantsis-Zacharov, E., Halpern, M., 2007. Culturable psychrotrophic bacterial communities in raw milk and their proteolytic and lipolytic traits. Appl. Environ. Microbiol. 73 (22), 7162–7168. http://dx.doi.org/10.1128/AEM.00866-07. Hettinga, K.A., van Valenberg, H.J.F., van Hooijdonk, A.C.M., 2008. Quality control of raw cows' milk by headspace analysis. Int. Dairy J. 18, 506–513. http://dx.doi.org/ 10.1016/j.idairyj.2007.10.005. Jarosińska, A., Barłowska, J., Wolanciuk, A., Pastuszka, R., Barłowska, K., 2014. Chemical composition and microbiological quality of market milk supplied to five dairies in the Lublin region, taking into account the season of purchase. Roczniki Naukowe Polskiego Towarzystwa Zootechnicznego 10 (2), 47–56 (in Polish). Klijn, N., Nieuwenhof, F.F.J., Hoolwerf, J.D., van der Waals, C.B., Weerkamp, A.H., 1995. Identification of Clostridium tyrobutyricum as the causative agent of late blowing in cheese by species-specific PCR amplification. Appl. Environ. Microbiol. 61 (8), 2919–2924. Łaniewska-Trokenheim, Ł., Olszewska, M., Mikš-Krajnik, M., Zadernowska, A., 2010. Patterns of survival and volatile metabolites of selected Lactobacillus strains during
75
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
quality and provide health benefits. Food Microbiol. 27 (6), 691–697. http://dx.doi. org/10.1016/j.fm.2010.05.023. Sienkiewicz, J., 2010. Concepts of biodiversity - their dimensions and measurements in the light of the subject literature. Ochrona Środowiska i Zasobów Naturalnych 45, 7–29 (in Polish). Toso, B., Procida, G., Stefanon, B., 2002. Determination of volatile compounds in cows' milk using headspace GC-MS. J. Dairy Res. 69, 569–577. Villeneuve, M.P., Lebeuf, Y., Gervais, R., Tremblay, G.F., Vuillemard, J.C., Fortin, J., Chouinard, P.Y., 2013. Milk volatile organic compounds and fatty acid profile in cows fed timothy as hay, pasture, or silage. J. Dairy Sci. 96 (11), 7181–7194. http://dx.doi. org/10.3168/jds.2013-6785.
long-terms incubation in milk. J. Microbiol. 48 (4), 445–451. http://dx.doi.org/10. 1007/s12275-010-0056-3. Mikš-Krajnik, M., 2013. Calibration of the Clarus 500 Gas Chromatograph for Quantitative Identification Volatile Compounds of Microbiological Origin in Samples of Milk and Ripened Cheese. 1− 6 (internal study). Nalepa, B., Markiewicz, L.H., 2017. PCR-DGGE markers for qualitative profiling of microbiota in raw milk and ripened cheeses. LWT- Food Sci. Technol. 84, 168–174. http://dx.doi.org/10.1016/j.lwt.2017.05.057. Randazzo, C.L., Pitino, I., Ribbera, A., Caggia, C., 2010. Pecorino Crotonese cheese: study of bacterial population and flavour compounds. Food Microbiol. 27, 363–374. http:// dx.doi.org/10.1016/j.fm.2009.11.010. Settanni, L., Moschetti, G., 2010. Non-starter lactic acid bacteria used to improve cheese
76
Contents lists available at ScienceDirect
International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro
Seasonal variances in bacterial microbiota and volatile organic compounds in raw milk
T
⁎
Beata Nalepaa, , Magdalena Anna Olszewskaa, Lidia Hanna Markiewiczb a
Department of Industrial and Food Microbiology, Faculty of Food Science, University of Warmia and Mazury in Olsztyn, pl. Cieszyński 1, 10-726 Olsztyn, Poland Department of Immunology and Food Microbiology, Institute of Animal Reproduction and Food Research, Polish Academy of Sciences, ul. Tuwima 10, 10-748 Olsztyn, Poland b
A R T I C L E I N F O
A B S T R A C T
Keywords: Raw milk Microbiota VOCs PCR-DGGE HS-GC Seasonal variation
The aim of this study was to define the composition of microbiota and the volatile organic compounds (VOCs) in samples of raw milk collected for 22 months between 2012 and 2014 originated from north-eastern region of Poland. The results revealed that the VOCs profile changed with respect to the season of milk collection, and milk collected in autumn was characterized by a higher content of acetic acid (C2), propionic acid (C3) and valeric acid (C5), whereas spring was characterized by a frequent presence of acetone (Ac), ethanol (Et) and ethyl acetate (EtAc). Bacterial species composition changed considerably within the tested period and some bacterial species/groups occurred seasonally, e.g. L. helveticus (summer), L. casei (winter). The results show usefulness of the applied techniques (PCR-DGGE and HS-GC) and data analysis (PCA, correlation coefficients) methods in characterizing the raw milk quality intended for dairy production.
1. Introduction The aim of all technological treatments applied in food production is to make sure that the customer is safe and to obtain products of the highest quality. In the dairy industry, the microbiological quality of a final product is determined by the quality of its main material - raw milk. The hygienic requirements (Commission Regulation (EC) No 1662/, 2006) define only an acceptable number of bacteria in 1 cm3 of raw milk, not specifying the qualitative content of its microbiota, while those are some biochemical groups or microorganism species that play a crucial role in determining the quality and shelf life of milk products. Lactic acid bacteria (LAB), such as Lactococcus, Lactobacillus, Enterococcus may cause the sourness of milk (Jarosińska et al., 2014) or may function as non-starter lactic acid bacteria (NSLAB) (Settanni and Moschetti, 2010). Coliforms cause organoleptic changes and early cheese blowing (Jarosińska et al., 2014). Psychrotrophic bacteria (Pseudomonas, Aeromonas, Bacillus) and the enzymes they produce lead to unfavourable physicochemical and organoleptic changes. By causing protein decomposition, proteolytic enzymes alter milk texture (thickening, gelification) and change taste and smell (e.g., putrid, bitter). By hydrolysing milk fat, lipolytic enzymes cause a rancid, soap-like, tallowy taste and smell of milk (Adamiak et al., 2015; Hantsis-Zacharov and Halpern, 2007). By surviving thermal processing, spore-forming bacteria (Bacillus, Clostridium) may cause milk spoilage, including
⁎
rotting, a bitter aftertaste, a rancid taste, or late cheese blowing (Klijn et al., 1995). Thus, the microorganisms that occur in milk are of decisive significance in determining properties of a final product. However, performing the examinations aimed at identifying all bacterial groups and species in the material with standard methods is relatively laborious and time-consuming. Therefore, it is becoming more frequent to apply the tools of molecular biology combined with modern instrumental methods. The molecular methods (e.g., PCR-DGGE) enable a precise identification of species diversity for the microorganisms in a studied sample (Franciosi et al., 2009; Nalepa and Markiewicz, 2017) with no cultivation needed. The instrumental techniques, e.g., HeadSpace - Gas Chromatography (HS-GC) (Ayad et al., 1999) make it possible to trace metabolic changes performed by those microorganisms, e.g., identifying the content of volatile organic compounds (VOCs) produced. The presence of bacteria in raw milk and the VOCs they produce determines the taste and smell of a final product, especially ripened cheeses. Therefore, the aim this study was to identify microbiota and the VOCs in the pooled raw milk. Differences and some potential links between those data have been determined in order to evaluate the quality of raw milk intended for dairy production.
Corresponding author. E-mail address: [email protected] (B. Nalepa).
https://doi.org/10.1016/j.ijfoodmicro.2017.12.024 Received 23 August 2017; Received in revised form 15 November 2017; Accepted 21 December 2017 Available online 24 December 2017 0168-1605/ © 2017 Elsevier B.V. All rights reserved.
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
result of thermostating of 5 cm3 of milk at 70 °C for 40 min in 22-mm tightly closed vials. Then, the sample was pressurized for 1 min and injected into the column (split ratio of 2:1) in 0.08 min. The temperatures of the needle and the transfer line were 100 °C and 120 °C, respectively. The volatile compound separation was performed on the HPINNOWAX column (60 m × 1.00 μm × 0.537 mm) by Agilent Technologies, USA, applying the following thermal gradient: 40 °C (5 min) → ΔT 10 °C/min → 220 °C (5 min). The temperature of the injector and the FID was 230 °C. The following gases were applied for the analysis: helium (5 cm3/min), synthetic air (400 cm3/min), and hydrogen (40 cm3/min) with the pressure of 130 kPa in the column (MikšKrajnik, 2013; internal study).
2. Materials and methods 2.1. Raw milk samples The pooled raw milk samples from a dairy plant in the Region of Warmia and Mazury in north-eastern Poland were studied. The samples were collected into sterile vessels on a monthly basis for the period of twenty-two months: from September 2012 to July 2014. A period from July to September was consider summer, months from October to December were autumn, January to March were winter and April to June were consider spring. 2.2. DNA isolation
2.6. Raw milk microbiota biodiversity The total DNA from the raw milk samples was obtained using the Genomic Mini AX FOOD kit (A@A Biotechnology, Gdańsk, Poland) in accordance with the manufacturer's instructions. The isolated DNA was stored at −80 °C for further analyses.
The DGGE profiles for the raw milk samples were compared with the markers that had been constructed in previous research (Nalepa and Markiewicz, 2017) and they included 24 referential strains: P. freudenreichii ssp. shermanii DSM 4902, L. lactis ssp. lactis DSM 4366, L. mesenteroides DSM 20346, P. thoenii DSM 20276, L. acidophilus DSM 9126, L. plantarum ATCC 8014, L. brevis DSM 1267, L. casei ATCC 334, L. delbrueckii DSM 20080, L. fermentum DSM 200052, L. helveticus DSM 20075, E. coli ATCC 8739, E. aerogenes ATCC 13048, E. cloacae ATCC 13047, C. freundii ATCC 8090, E. faecalis ATCC 29212, B. subtilis ATCC 6051, C. butyricum ATCC 10702, C. tyrobutyricum ATCC 2637, C. perfringens ATCC 13124, L. monocytogenes ATCC BAA-751, S. thermophilus ATCC 19258, S. xylosus ATTC 29971, and S. aureus ATCC 43300. Basing on that, the presence of particular bacterial species was detected and assigned scores from 0 to 4 depending on their band brightness. Also, the 1D analysis of electrophorograms was performed using the Doc-It LS Image Analysis application (UVP Ltd., UK). The data obtained in that way were used to calculate the Shannon–Wiener index that defines biodiversity and is expressed by the following formula: (Sienkiewicz, 2010),
2.3. Polymerase chain reaction (PCR) The reaction mixture (25 μL) contained: 20 mM of Tris-HCl (pH 8.4), 50 mM of KCl, 3 mM of MgCl2, 50 μM of dNTPs, 5 pM of every primer, 1.25 U of Taq polymerase (Fermentas), and 20–40 ng of the DNA. The following primers were used: U968-GC (5′-CGC CCG GGG CGC GCC CCG GGC GGG GCG GGG GCA CGG GGGG AAC GCG AAG AAC CTT AC-3′) and L1401-r (5′-CGG TGT GTA CAA GAC CC-3′) amplifying a fragment of region V6-V8. Amplification was carried out in the MJ Mini Gradient Thermal Cycler (BioeRad, Poland). The PCR profile was as follows: initial denaturation at 94 °C for 5 min followed by 35 cycles of: denaturation at 94 °C for 30 s, annealing at 56 °C for 30 s and extension at 68 °C for 40 s. Final extension at 68 °C for 7 min. (Randazzo et al., 2010). The PCR product with the estimated size of 450 bp was analyzed by electrophoresis on 1% agarose gel in 0.5 × TBE buffer in the MultiSub Choice system (Clever Scientific Ltd., UK).
S
H ′ = − ∑ pi ln pi i=1
2.4. Denaturing gradient gel electrophoresis (DGGE)
where pi is the proportion of individuals in species i to the total number of individuals in the community.
The PCR products were analyzed using the electrophoresis in a gradient of denaturing agents. The electrophoresis was conducted in the 8% polyacrylamide gel (acrylamide:bisacrylamide 37.5:1) with the gradient of denaturing agent (urea) ranging from 35% to 57.5%. The electrophoresis was carried out in the 0.5 × TAE buffer at the temperature of 60 °C under the voltage of 85 V for 16 h in the DCode Universal Mutation System (BioRad, Poland) (Randazzo et al., 2010). On each gel a marker set was run to enable the identification of DGGE bands (Nalepa and Markiewicz, 2017). The gels were stained in the SybrGreen I (1,10,000) solution for 15 min and then they were archived using G-Box (Syngen, Poland).
Seasonal variations in the VOC profiles of raw milk and associations between the VOC profiles and the microbial compositions of the collected samples were determined in the principal component analysis (PCA). The direction and strength of these correlations were determined basing on the values of the Pearson correlation coefficient (r) at the significance level of p < 0.05. The data were processed in Statistica v. 12.5 (StatSoft Inc., Tulsa, USA).
2.5. Detecting volatile organic compounds using gas chromatography
3. Results
Detection of selected volatile compounds: aldehydes (acetaldehyde), ketones (acetone, diacetyl, acetoin), alcohols (methanol, ethanol), esters (ethyl acetate, ethyl propionate, ethyl butyrate), fatty acids C2 - C7 (acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, valeric acid, isocaproic acid, caproic acid, heptanoic acid) was performed in duplicate using headspace gas chromatography (HS-GC) with the Headspace Turbomatrix 40 autosampler (Perkin Elmer, USA) and the Clarus 500 gas chromatograph (Perkin Elmer, USA) with a flame ionization detector (HS-GC-FID). The chromatograph was calibrated for a quantitative detection [μg/cm3] of the selected metabolites using some external standards. Calibration curves were made for every compound in the respective concentration scope. Equilibrium between the sample and the headspace was obtained as a
3.1. Raw milk microbiota
2.7. Statistical analysis
Upon comparing the DGGE profiles of the raw milk samples collected from September 2012 to July 2014, 3 to 12 bacterial species out of the 24 identified ones were detected (Table 1). The presence of 3 species was detected in 6 milk samples, 4 species were detected in 1 sample, 5 - in 2 samples, 6 - in 3 samples, 7 - in 5 samples, and 9 - in 2 samples. 8, 10, and 12 bacterial species were identified in the remaining 3 samples. The values of the Shannon-Wiener index (H′) (Table 1) correspond to the biological diversity of the collected raw milk samples. In 54.5% milk samples, the H′ indexes ranged from 1.3693 to 1.9782 (presence of 5–8 species) regardless of the analyzed season. In six (27%) milk samples the biological diversity was very low 71
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
(Fig. 1), as PCR-DGGE is a semi-quantitative method in which the amount of the DNA isolated from cells correlates with the amount of amplicons and their brightness. Applying that type of assessment, it has been determined that the most frequent bacteria included those of the P. thoenii and E. aerogenes species occurring in 15 of 22 samples (68%) followed by S. thermophilus/S. xylosus in 14 samples (63.6%). As for LAB, L. mesenteroides was the most frequent - in 13 samples (59.1%) and its presence was not common for neither of the analyzed seasons. In 7 samples (31.8%), L. brevis was identified, similarly as L. acidophilus. The presence of L. brevis was characteristic of autumn 2013 and winter 2014, and the L. acidophilus was present in the majority of the samples from summer 2012 to summer 2013. The presence of the L. lactis was identified in 5 samples (22.7%) and it was common for the summer/ autumn 2013 season. Out of the remaining LAB, L. plantarum (5 samples), L. casei (2 samples) and L. helveticus (2 samples) were identified sporadically and randomly. However, L. delbrueckii and L. fermentum were not detected in any of the examined samples (Fig. 1). Moreover, the fecal species (E. coli, E. aerogenes, E. cloacae, C. freundii and E. faecalis) were detected in 16 (72.2%) samples. E. aerogenes was the most frequent species as it was present in all the samples from the summer 2013 to summer 2014 season. Additionally, in the summer/autumn 2013 season, the presence of E. aerogenes is accompanied by occurrences of E. coli, E. cloacae, C. freundii and E. faecalis (Fig. 1). The sporeforming bacteria (B. subtilis, C. butyricum, C. tyrobutyricum and C. perfringens) were present in 17 (77.3%) samples. The presence of C. tyrobutyricum was identified in 9 samples (41%) of which 7 were from the summer 2012 to spring 2013 season. The presence of C. butyricum was characteristic of the samples from summer/autumn 2013 and that of C. perfringens for spring 2014. However, C. perfringens also occurred randomly in 6 other samples (Fig. 1). As for the remaining species, S. aureus was detected most often − in 6 samples (27.3%) from the summer 2012 to winter 2013 season (Fig. 1). When considering the results according to the collection years, it has been observed that all the samples of raw milk from 2012 contained L. mesenteroides, sporeforming C. tyrobutyricum, and S. aureus while the majority of samples from 2013 to 2014 contained fecal bacteria, such as E. aerogenes and
Table 1 The number of bacterial species at different times of years and the Shannon-Wiener indexes calculated from DGGE patterns of the raw milk from the 22-month period between 2012 and 2014. Seasona
Number of species
Shannon-Wiener index
Spring 2013 Spring 2013 Spring 2014 Spring 2014 Spring 2014 Summer 2012 Summer 2013 Summer 2013 Summer 2013 Summer 2014 Autumn 2012 Autumn 2012 Autumn 2012 Autumn 2013 Autumn 2013 Autumn 2013 Winter 2013 Winter 2013 Winter 2013 Winter 2014 Winter 2014 Winter 2014
3 6 7 4 3 7 12 3 9 5 6 6 7 10 9 7 5 8 3 5 3 8
0.5254 1.5193 1.8013 0.9294 0.9355 1.6593 2.2064 0.8757 2.0561 1.3693 1.4782 1.6383 1.7421 2.1649 2.1135 1.8258 1.4899 1.7605 0.8888 1.5403 0.8984 1.9782
a Spring: April, May, June. Summer: July, August, September. Autumn: October, November, December. Winter: January, February, March.
and then the H′ indexes ranged from 0.5254 to 0.9355 (presence of 3–4 species). Those were the samples collected mostly in the winter/spring season. In the four remaining samples (18.1%), however, the biodiversity was high. The values of the H′ indexes ranged from 2.0561 to 2.2064 (presence of 9–12 species). Those samples were collected in the summer-autumn season (Table 1). Bacterial species in milk samples were identified on the basis of DGGE band position. Band brightness was given a score of 0 to 4
Fig. 1. Presence of microorganisms in raw milk determined on the basis of PCR-DGGE patterns and band intensities in a scale from 0 (lack of a band) to 4 (the highest intensity). Samples of raw milk were collected in 22-month period between 2012 and 2014.
72
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
Fig. 2. The dendrogram showing a dissimilarity measure between the VOCs (A) and the PCA biplot of seasonal variables and volatile compounds data related to factor 1 (F1) and factor 2 (F2) (B). Ketones: Ac - acetone, Di - diacetyl, Acet - acetoin; Alcohols: Me - methanol; Et - ethanol; Esters: EtAc - ethyl acetate; Fatty acids: C2 - acetic acid, C3 - propionic acid, C4iso - isobutyric acid, C5iso - isovaleric acid, C5 - valeric acid, C6 - caproic acid; C7 - heptanoic acid.
Table 3 Pearson correlation coefficients (r) between selected microorganisms and those VOCs which were grouped into two clusters based the PCA (Fig. 2).
Table 2 The content (sum) of volatile organic compounds (VOCs) in the raw milk from the 22month period between 2012 and 2014. Seasona
Spring 2013 Spring 2013 Spring 2014 Spring 2014 Spring 2014 Summer 2012 Summer 2013 Summer 2013 Summer 2013 Summer 2014 Autumn 2012 Autumn 2012 Autumn 2012 Autumn 2013 Autumn 2013 Autumn 2013 Winter 2013 Winter 2013 Winter 2013 Winter 2014 Winter 2014 Winter 2014
Sum of VOCs (μg/cm3) Aldehydesb
Esters
Alcohols
Ketones
Fatty acids
Total
– – – – – – – – – – – – – – – – – – – – – –
0.20 – – 0.30 0.20 – – – – 0.10 – – – – – – – – – – – –
39.70 6.70 23.80 15.05 19.25 – 16.00 – – 90.05 – – – – – – – – – – – –
7.10 5.50 4.10 3.15 4.40 – 3.60 – 7.15 5.10 – – – – – – – – – – – –
46.60 92.70 23.35 46.70 46.60 260.35 46.70 40.80 178.15 77.95 117.05 159.20 203.35 900.85 471.35 91.60 – 425.10 107.65 – – 166.85
93.60 104.90 51.25 65.20 70.45 260.35 66.30 40.80 185.30 173.20 117.05 159.20 203.35 900.85 471.35 91.60 – 425.10 107.65 – – 166.85
VOCs (spring) Ac Et −0.26 −0.19 −0.18 −0.05 −0.01 −0.16 −0.35 −0.19
EtAc −0.17 −0.22 0.28 −0.34
−0.01 −0.17 0.32 0.27 0.39 0.25 0.29
−0.17 −0.27 −0.12 −0.20 0.61 −0.05 0.51
−0.07 0.01 0.34 0.27 0.20 0.20 0.00
Bacterial species E. cloacae L. plantarum C. butyricum S. thermophilus/S. xylosus E. coli L. brevis B. subtilis L. lactis L. mesenteroides E. aerogenes C. perfringens
VOCs (autumn) C2 C3 0.65 0.74 0.40 0.62 0.63 0.49 0.31 0.38
C5 0.48 0.61 0.28 0.16
0.23 0.13 0.23 0.30 0.05 −0.03 0.15
−0.11 0.02 0.23 0.16 −0.17 −0.05 0.12
0.15 0.26 0.08 0.18 −0.26 −0.06 −0.13
seasons, it was found out that the highest diversity of volatile compounds (however, in the lowest amounts) occurred in the spring seasons. Esters, ketones, alcohols, and fatty acids were identified in the milk samples collected in this season and their average amounts were 0.14, 4.85, 20.9, and 51.19 μg/cm3, respectively (Table 2). In the cold seasons (autumn and winter), only free fatty acids were detected. For the season of autumn 2012, C3, C4iso, C6, and C7 were identified in the amounts ranging from 117 to 203 μg/cm3 while for the autumn 2013 season - C2, C3, C5 and C6 (Supplementary Table 1) in the amounts ranging from 91 to 900 μg/cm3 (Table 2). No presence of the examined volatile compounds was detected in a half of the samples from the winter seasons (1 sample from winter 2013 and 2 samples from winter 2014). C2 and C3 in the amounts ranging from 107 to 425 μg/cm3 were identified in two samples from winter 2013 and in one sample from winter 2014, C3 had been detected in the amount of 166 μg/cm3 (Table 2). The highest diversity of the VOC contents was identified within the samples collected in the summer seasons. In two samples, the fatty acids only were present. The sample from summer 2012 contained the largest amount of fatty acids (260 μg/cm3), while one sample from summer 2013 contained them in the lowest amount (40 μg/cm3) and those were C3, C5izo, C5, C7, and C5izo, respectively (Table 2; Supplementary Table 1). Apart from the fatty acids, ketones (summer 2013), alcohols and ketones (summer 2013), as well as alcohols, ketones, and esters (summer 2014) were detected in very varied amounts in the remaining 3 samples collected in the summers. However, no acetaldehyde was detected in any of the examined milk samples (Table 2; Supplementary Table 1).
a Spring: April, May, June. Summer: July, August, September. Autumn: October, November, December. Winter: January, February, March. b Aldehydes: acetaldehyde. Esters: ethyl acetate, ethyl propionate, ethyl butyrate. Alcohols: methanol, ethanol. Ketones: acetone, diacetyl, acetoin. Fatty acids: acetic acid, propionic acid, isobutyric acid, butyric acid, isovaleric acid, valeric acid, isocaproic acid, caproic acid, heptanoic acid.
NSLAB, L. brevis (Fig. 1). 3.2. Volatile organic compounds (VOCs) of raw milk Contents of selected volatile compounds in the groups of aldehydes, esters, ketones, and fatty acids C2 − C7 have been presented in Table 2 and Supplementary Table 1. The total content of volatile compounds in the examined samples of raw milk was 170.65 μg/cm3 on average. The highest content of VOCs was identified in samples collected in the autumn seasons. Their average content was 323.90 μg/cm3 while in the summer and winter seasons it was 145.19 and 116.60 μg/cm3 on average. The lowest contents of VOCs were detected in the milk samples from the spring seasons and they were 77.08 μg/cm3 on average (Table 2). When analyzing the VOC contents in particular groups and 73
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
collection (Beresford et al., 2001). Using RAPD-PCR, Ercolini et al. (2009) have shown that as for the fecal bacteria (of the Enterobacteriaceae family) those were Hafnia alvei, Serratia marcescens, Citrobacter freundii, E. coli, and Klebsiella that were most common in their raw milk samples while in our case those were E. aerogenes as well as, sporadically, C. freundii, E. coli, and E. cloacae. When considering the results seasonally, an increased incidence of coliforms (E. coli, E. cloacae, E. aerogenes, and C. freundii) and the L. lactis in the summer/ autumn 2013 season has been observed in our research. As for the LAB, those were L. brevis, L. plantarum, and L. acidophilus that occurred more frequently in the autumn/winter seasons. Seasonality in occurrences of some bacterial species in raw milk has also been shown by Czerwińska and Piotrowski (2011). When doing their research with culturing methods, they found out that in the summer/autumn seasons the number of bacteria of the Enterobacteriaceae family (including coliforms) in milk was higher than in other seasons and the number of lactococci was at its peak in autumn. As for the NSLAB, they identified the presence of L. casei in the winter/spring seasons. Similarly to our research, Ercolini et al. (2009) have also shown the presence of grampositive bacteria of the Bacillus and Staphylococcus in raw milk samples. Applying molecular methods, such as PCR-DGGE or next-generation sequencing (NGS), makes it possible to detect and identify not only dominant bacterial species/groups but also those that occur sporadically or in low-number populations. Using NSG, Escobar-Zepeda et al. (2016) have shown 3 dominant bacterial species in cheese made of raw milk. Those were L. plantarum, L. mesenteroides, and Weissella paramesenteroides, which made 80% of the whole population in cheese. The Aerococcus, Enterococcus, Lactococcus, and Staphylococcus made less than 10% of the population and 78 other bacterial genera made less than 1% which included Propionibacterium, Bifidobacterium, Corynebacterium, and others. By applying PCR-DGGE to the raw milk samples in our research, the domination of L. mesenteroides, Propionibacterium, and Staphylococcus has been identified. Bystroń et al. (2001) have detected the presence of staphylococci in 24% of their raw milk samples, of which the most frequently identified ones included S. aureus and S. xylosus, similarly to our research. Bacteria that occur in the material, particularly in raw milk, may cause its spoilage and worthlessness in production of other dairy products. It is caused by a possible growth of bacteria in milk, even during cold storage, and the release of proteolytic and lipolytic enzymes as well as volatile organic compounds (VOCs) that cause disturbances in the structure, texture, appearance, taste, and smell of a final product (Alothman et al., 2017). Thus, identifying VOC concentrations and occurrences of particular bacterial species may provide valuable information on the influence of microorganisms on the course of production processes and the quality of final products. Identification of VOC concentrations and their changes in raw milk or milk inoculated with single or mixed bacterial cultures has been performed in numerous studies (Alothman et al., 2017; Hettinga et al., 2008; ŁaniewskaTrokenheim et al., 2010; Toso et al., 2002; Villeneuve et al., 2013). When examining raw milk, Hettinga et al. (2008) have detected only 3 compound groups: aldehydes, ketones, and sulfur compounds. Free fatty acids (FFAs) occurred sporadically and in trace amounts. In our research in the milk collected in north-eastern Poland, the dominant class of volatile compounds includes free fatty acids C2 - C7, the highest concentration of which has been identified for the autumn seasons and the lowest one - for the spring seasons. Further on, they included alcohols, ketones, and esters in small concentrations; however, no presence of aldehydes was detected. Toso et al. (2002) have shown that in the milk collected in north-eastern Italy, the dominant class of compounds included ketones and then aldehydes, alcohols, carbohydrates, esters, sulfur compounds, and terpenes. When considering the obtained results in the seasonal aspect, the highest VOC diversity has been noticed in the spring seasons and the lowest one - in the autumn/winter seasons. However, autumn/winter seasons were characterized by a higher concentration of fatty acids which may result from the method of
3.3. PCA results A principal component analysis (PCA) has been conducted to determine if volatile organic compounds (VOCs) detected in the milk from the 22-month period between 2012 and 2014 relate with seasons of milk collection. As demonstrated by Fig. 2A first, a dissimilarity between acetic acid (C2), propionic acid (C3) and the remaining VOCs has been noticed. This corresponded to the C2 and C3 amounts of 52.8–437.1 μg/cm3 and 50.8–362.8 μg/cm3, respectively and those were the highest concentrations determined in milk in comparison to the rest of VOCs which ranged from the lowest level of 0.1 μg/cm3 to the highest level of 117,6 μg/cm3 (Supplementary Table 1). As shown on Fig. 2B, the PCA biplot mainly presents two clusters, I and II, which are a collection of spring and autumn samples, respectively. This PCA biplot comprised 37.77% of the total variance and based on the values of loading coefficients with F 1, the separation between C2, C3 and C5 vs. acetone (Ac), ethanol (Et) and ethyl acetate (EtAc) was revealed (Fig. 2B). C2, C3 and C5 are variables with the highest negative loading: − 0.48, − 0.67, and, − 0.42, respectively whereas the remaining VOCs are variables with the highest positive loading: 0.90, 0.75, and 0.69, respectively. As results, C2, C3 and C5 were attributed mainly to the autumn and the Ac, Et and EtAc to spring season of milk collection (Fig. 2B). In contrast, e.g., acetoin (Acet) and isovaleric acid (C5iso) are variables with the positive loading with F 2, however these coefficients were weak (0.34, 0.31, respectively). Besides, if they were detected, they were predominantly found in the summer samples. Additionally, based on Pearson correlation coefficients (r) microorganisms detected in milk samples were associated with certain VOCs. As shown in Table 3, differences in strengths and directions of relationship between VOCs vs. microorganisms have been revealed. It is worth noticing that, e.g.: E. cloacae is positively coupled to C2 (r = 0.65), C3 (r = 0.74) and C5 (r = 0.48), and all the coefficients are significant. Moreover, the r values revealed significant relationships between L. plantarum and C3 (r = 0.62), C5 (r = 0.61), as well as between C. butyricum vs. C2 (r = 0.63) and C3 (r = 0.49). Furthermore, the L. mesenteroides and C. perfringens may have contributed to EtAc (r = 0.61 and 0.51), since these correlations were moderate and significant. 4. Discussion Milk is the most important material in the dairy industry and its quality, particularly microbiological one, is of decisive significance for products made of it. In the case of raw milk, dairies plants determine the total bacteria count only that cannot exceed 1 × 105 cfu/cm3 (Commission Regulation (EC) No 1662/, 2006). All the raw milk samples examined for this study complied with that requirement (data not shown). However, as it has been shown by the PCR-DGGE analysis, the bacterial species composition in the raw milk varied significantly within 22 months of sample collection and some bacterial groups/species occurred seasonally (Fig. 1). Application of DGGE markers developed in previous research (Nalepa and Markiewicz, 2017) enabled a simple interpretation of the obtained DGGE profiles for the particular milk samples as well as an immediate identification of the species without the need to cut the bands out and sequencing. As for cocci, the presence of S. thermophilus and L. mesenteroides was identified most frequently in the milk. The propionate acid bacteria, P. thoenii were also common. Among NSLAB, L. brevis, L. acidophilus and L. plantarum were the most frequent ones. Using the same technique and primers, Aquilanti et al. (2013) have examined Italian artisan cheese made of raw milk. They have shown that the S. thermophilus, L. lactis and L. mesenteroides were the most frequent in their samples and as for the lactobacilli, those were species of the L. casei group. Other species of the NSLAB, such as L. acidophilus, L. helveticus, L. plantarum, L. brevis, and L. fermentum were detected less often. It is the material, which is raw milk, that is the source of those bacteria in fermented products and they get there mostly from the environment or the equipment used for milk 74
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
(PCR-DGGE and HS-GC) and data analysis (PCA, correlation coefficients) methods in characterizing the raw milk quality. The determined microbiological and VOCs profiles of raw milk may be of great importance in predicting possible disturbances in the dairy production and in planning of modifications of technological processes depending on the identified specific raw milk microbiota and the VOCs they produce. Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ijfoodmicro.2017.12.024. a The correlation coefficients (r) written in bold are statistically significant (p ≤ 0.05). Ac - acetone; Et - ethanol; EtAc - ethyl acetate; C2 acetic acid; C3 - propionic acid; C5 - valeric acid. The range and strength of correlations: ± 0.00–0.15 - very low positive (negative) correlation; ± 0.16–0.29 - weak to low positive (negative) correlation; ± 0.30–0.49 - moderate to low positive (negative) correlation; ± 0.50–0.69 - moderate positive (negative) correlation; ± 0.70–0.89 strong positive (negative) correlation; ± 0.90–1.00 - very strong positive (negative) correlation. Shades of grey indicate only positive correlation coefficients (0.16 ≤ r ≤ 1.00).
feeding the cows in winter. At that time in Poland, cows are kept indoors and are fed with hay and silages. Villeneuve et al. (2013) report that in the milk of cows fed with hay and silages, there is more fat and, thus, more fatty acids. On the other hand, high concentrations of FFAs in some of the summer samples may be attributed to the metabolic activities of bacteria present in the milk. Aquilanti et al. (2013) suggest that short- and medium-chain FFAs are formed in three biochemical pathways, which are lactose fermentation, proteolysis, and lipolysis triggered by the enzymes exerted to milk by microorganisms. That is confirmed in the research by Alothman et al. (2017) who have identified VOC profiles in milk inoculated with the P. fluorescens and Chryseobacterium bacteria. Presence of those bacteria in milk increased concentrations of alcohols, esters, acetaldehyde, as well as acetic and butyric acids. The research by Łaniewska-Trokenheim et al. (2010) has shown that concentrations of ethanol, ketones, and acetic acid was increased in milk inoculated with L. acidophilus and L. crispatus. Thus, it is clearly visible that the profile of volatile compounds in milk depends not only on the method of cow feeding (season) but also on the amount and metabolic activity of microorganisms that get into the material e.g. during the process of milk collection. Knowledge of milk microbiota compositions and profiles of volatile compounds may be of tremendous significance for prediction of the quality of a final product so that any errors in operation processes would be avoided. However, comparing enormous amounts of detailed information obtained every day by various research methods in every phase of production is simply unfeasible. Here, it is the development of statistics that turns out to be very helpful. Applying multidimensional statistical analyses, such as principal component analysis (PCA) or partial least squares regression (PLS), makes it possible to notice trends and dependencies between numerous and very varied parameters. Many authors have applied those methods successfully. Toso et al. (2002) have applied PCA to assess the influence of fodder types used in cow feeding on profiles of volatile compounds and a possibility of raw milk qualification on that basis. The analysis has shown that nine volatile compounds is of discriminating character and basing on their presence and concentrations in milk the authors were able to assess the type of food used in cow feeding. Gallardo-Escamilla et al. (2005) have used PCA and PLS to establish a dependency between VOCs produced by various species of microorganisms and taste and smell of fermented milks made of various types of whey. In our research, applying PCA and correlation coefficients (r) has made it possible to show the seasonality in occurrences of particular classes of volatile compounds, as well as to attempt to correlate VOC occurrences with the presence of particular bacterial species. The samples collected between 2012 and 2014 were mainly grouped into two clusters (Fig. 2B). The first one was made of the samples from all the spring seasons and was characterized by the presence of alcohols (Et), ketones (Ac), and esters (EtAc). The second cluster was formed by the majority of the autumn seasons that were characterized mostly by the presence of fatty acids like C2, C3, C5. The highest diversity and no grouping has been shown in the case of the summer samples as they cannot be grouped with any VOCs and other seasonal samples. It may mean, e.g., a broadly understood lack of recurrence in the material collection and/or diversity of weather conditions in warmer months resulting in a presence of microorganisms responsible for further differences in quality of final products made from it. The analysis of dependencies between the volatile compounds and the bacteria has shown that a connection is possible between particular compounds and particular bacterial species. For example, amounts of ethyl acetate (EtAc) identified in the spring seasons are correlated with occurrences of C. perfringens. Some crucial correlation coefficients between dependencies in occurrences of acids, such as acetic, propionic, or valeric ones, and the presence of E. cloacae, L. plantarum, or C. butyricum may be connected with the samples collected in the autumn season. Summarizing, our results show usefulness of the applied techniques
Acknowledgements This study has been supported by the National Science Center, Poland (grant no. N N312 484140). References Adamiak, A., Górska, A., Mróz, B., 2015. Psychrotrophic bacteria in raw milk and dairy products. Żywność. Nauka. Technologia. Jakość 4 (101), 36–48 (in Polish). Alothman, M., Lusk, K.A., Silcock, P., Bremer, P.J., 2017. Comparing PTR-MS profile of milk inoculated with pure or mixed cultures of spoilage bacteria. Food Microbiol. 64, 155–163. http://dx.doi.org/10.1016/j.fm.2017.01.001. Aquilanti, L., Santarelli, S., Babini, V., Osimani, A., Clementi, F., 2013. Quality evaluation and discrimination of semi-hard and hard cheeses from the Marche region (Central Italy) using chemometric tools. Int. Dairy J. 29 (1), 42–52. http://dx.doi.org/10. 1016/j.idairyj.2012.11.001. Ayad, E.H.E., Verheul, A., de Jong, C., Wouters, J.T.M., Smit, G., 1999. Flavour forming abilities an amino acid requirements of Lactococcus lactis strains isolated from artisanal and non-dairy origin. Int. Dairy J. 9, 725–735. http://dx.doi.org/10.1016/ S0958-6946(99)00140-5. Beresford, T.P., Fitzsimons, N.A., Brennan, N.L., Cogan, T.M., 2001. Recent advances in cheese microbiology. Int. Dairy J. 11, 259–274. http://dx.doi.org/10.1016/S09586946(01)00056-5. Bystroń, J., Kosek-Paszkowska, K., Molenda, J., 2001. Occurrence of enterotoxigenic staphylococci in raw milk. Med. Weter. 57 (9), 645–648 (in Polish). Commission Regulation (EC) No 1662/2006 of 6 November 2006 amending Regulation (EC) No 853/2004 of the European Parliament and of the Council laying down specific hygiene rules for food of animal origin. Czerwińska, E., Piotrowski, W., 2011. Potential sources of milk contamination influencing its quality for consumption. Rocznik Ochrona Środowiska 13, 635–652 (in Polish). Ercolini, D., Russo, F., Ferrocino, I., Villani, F., 2009. Molecular identification of mesophilic and psychrotrophic bacteria from raw cow's milk. Food Microbiol. 26, 228–231. http://dx.doi.org/10.1016/j.fm.2008.09.005. Escobar-Zepeda, A., Sanchez-Flores, A., Baruch, M.Q., 2016. Metagenomic analysis of a Mexican ripened cheese reveals a unique complex microbiota. Food Microbiol. 57, 116–127. http://dx.doi.org/10.1016/j.fm.2016.02.004. Franciosi, E., Settanni, L., Cavazza, A., Poznański, E., 2009. Biodiversity and technological potential of wild lactic acid bacteria from raw cows' milk. Int. Dairy J. 19, 3–11. http://dx.doi.org/10.1016/j.idairyj.2008.07.008. Gallardo-Escamilla, F.J., Kelly, A.L., Delahunty, C.M., 2005. Influence of starter culture on flavor and headspace volatile profiles of fermented whey and whey produced from fermented milk. J. Dairy Sci. 88 (11), 3745–3753. http://dx.doi.org/10.3168/jds. S0022-0302(05)73060-5. Hantsis-Zacharov, E., Halpern, M., 2007. Culturable psychrotrophic bacterial communities in raw milk and their proteolytic and lipolytic traits. Appl. Environ. Microbiol. 73 (22), 7162–7168. http://dx.doi.org/10.1128/AEM.00866-07. Hettinga, K.A., van Valenberg, H.J.F., van Hooijdonk, A.C.M., 2008. Quality control of raw cows' milk by headspace analysis. Int. Dairy J. 18, 506–513. http://dx.doi.org/ 10.1016/j.idairyj.2007.10.005. Jarosińska, A., Barłowska, J., Wolanciuk, A., Pastuszka, R., Barłowska, K., 2014. Chemical composition and microbiological quality of market milk supplied to five dairies in the Lublin region, taking into account the season of purchase. Roczniki Naukowe Polskiego Towarzystwa Zootechnicznego 10 (2), 47–56 (in Polish). Klijn, N., Nieuwenhof, F.F.J., Hoolwerf, J.D., van der Waals, C.B., Weerkamp, A.H., 1995. Identification of Clostridium tyrobutyricum as the causative agent of late blowing in cheese by species-specific PCR amplification. Appl. Environ. Microbiol. 61 (8), 2919–2924. Łaniewska-Trokenheim, Ł., Olszewska, M., Mikš-Krajnik, M., Zadernowska, A., 2010. Patterns of survival and volatile metabolites of selected Lactobacillus strains during
75
International Journal of Food Microbiology 267 (2018) 70–76
B. Nalepa et al.
quality and provide health benefits. Food Microbiol. 27 (6), 691–697. http://dx.doi. org/10.1016/j.fm.2010.05.023. Sienkiewicz, J., 2010. Concepts of biodiversity - their dimensions and measurements in the light of the subject literature. Ochrona Środowiska i Zasobów Naturalnych 45, 7–29 (in Polish). Toso, B., Procida, G., Stefanon, B., 2002. Determination of volatile compounds in cows' milk using headspace GC-MS. J. Dairy Res. 69, 569–577. Villeneuve, M.P., Lebeuf, Y., Gervais, R., Tremblay, G.F., Vuillemard, J.C., Fortin, J., Chouinard, P.Y., 2013. Milk volatile organic compounds and fatty acid profile in cows fed timothy as hay, pasture, or silage. J. Dairy Sci. 96 (11), 7181–7194. http://dx.doi. org/10.3168/jds.2013-6785.
long-terms incubation in milk. J. Microbiol. 48 (4), 445–451. http://dx.doi.org/10. 1007/s12275-010-0056-3. Mikš-Krajnik, M., 2013. Calibration of the Clarus 500 Gas Chromatograph for Quantitative Identification Volatile Compounds of Microbiological Origin in Samples of Milk and Ripened Cheese. 1− 6 (internal study). Nalepa, B., Markiewicz, L.H., 2017. PCR-DGGE markers for qualitative profiling of microbiota in raw milk and ripened cheeses. LWT- Food Sci. Technol. 84, 168–174. http://dx.doi.org/10.1016/j.lwt.2017.05.057. Randazzo, C.L., Pitino, I., Ribbera, A., Caggia, C., 2010. Pecorino Crotonese cheese: study of bacterial population and flavour compounds. Food Microbiol. 27, 363–374. http:// dx.doi.org/10.1016/j.fm.2009.11.010. Settanni, L., Moschetti, G., 2010. Non-starter lactic acid bacteria used to improve cheese
76