International Journal of Food Microbiology 180 (2014) 88–97

Contents lists available at ScienceDirect

International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro

Cultivation-independent analysis of microbial communities on Austrian raw milk hard cheese rinds Elisa Schornsteiner a, Evelyne Mann a, Othmar Bereuter b, Martin Wagner a, Stephan Schmitz-Esser a,⁎ a b

Institute for Milk Hygiene, Department for Farm Animals and Veterinary Public Health, University of Veterinary Medicine, 1210 Vienna, Austria Landwirtschaftskammer Vorarlberg, 6900 Bregenz, Austria

a r t i c l e

i n f o

Article history: Received 5 December 2013 Received in revised form 24 March 2014 Accepted 6 April 2014 Available online 13 April 2014 Keywords: 16S rRNA cloning 18S rRNA cloning Cheese rind bacteria Cheese rind yeast Raw milk hard cheese

a b s t r a c t “Vorarlberger Bergkäse” (VB) is an Austrian artisanal hard cheese produced from raw cow's milk. The composition of its rind microbiota and the changes in the microbial communities during ripening have not previously been investigated. This study used 16S and 18S rRNA gene cloning and sequencing to characterize the bacterial and fungal communities of seven pooled cheese rind samples taken in seven different ripening cellars of three Austrian dairy facilities. A total of 408 clones for 16S and 322 clones for 18S rRNA gene libraries were used for taxonomic classification, revealing 39 bacterial and seven fungal operational taxonomic units (OTUs). Bacterial OTUs belonged to four different phyla. Most OTUs were affiliated to genera often found in cheese, including high numbers of coryneforms. The most abundant OTU from 16S rRNA gene libraries showed highest similarity to Halomonas. Young cheese rinds were dominated by Actinobacteria or Proteobacteria, particularly by Halomonas and Brevibacterium aurantiacum, while Staphyloccocus equorum was most abundant in old cheeses. The most abundant 18S rRNA OTU had highest similarity to the filamentous fungus Scopulariopsis brevicaulis. Pairwise correlation analyses revealed putative co-occurrences between a number of OTUs. It was possible to discriminate the different cheese rind microbiota at the community-level by facility affiliation and ripening time. This work provides insights into the microbial composition of VB cheese rinds and might allow the processing- and ripening conditions to be improved to enhance the quality of the product. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Cheese production is characterized by dynamic changes in the complex cheeses microbiome, which consists of both bacteria and fungi. An important period in cheese production is the ripening process, during which the cheese surface microbial community establishes itself and matures, facilitating the development of the cheese's organoleptic and textural properties. Bacteria and fungi on the cheese surface contribute to ripening due to their proteolytic and lipolytic activities and the production of volatile sulfur compounds as well as of ammonia. The metabolic activity of both yeasts and bacteria on the cheese surface, which either naturally develop during contact with air, or are intentionally inoculated during the ripening process, are of great importance (Irlinger and Mounier, 2009; Montel et al., 2014). From the perspective of microbial ecology, the cheese surface and core are fundamentally different environments; the latter becomes anaerobic during ripening, whereas the surface remains aerobic. Particularly surface-ripened cheeses, which are defined by additional ripening from the cheese ⁎ Corresponding author at: Institute for Milk Hygiene, University of Veterinary Medicine Vienna, Veterinärplatz 1, 1210 Vienna, Austria. Tel.: + 43 1 25077 3510; fax: +43 1 25077 3590. E-mail address: [email protected] (S. Schmitz-Esser).

http://dx.doi.org/10.1016/j.ijfoodmicro.2014.04.010 0168-1605/© 2014 Elsevier B.V. All rights reserved.

surface to the interior, develop their flavor in this period. Surfaceripened cheeses can be divided into bacteria-ripened (e.g. Limburger, Munster, Tilsit, and Appenzeller) and mold-ripened (e.g. Camembert, Brie) cheeses (Brennan et al., 2002; Dolci et al., 2009; Mounier et al., 2005). Due to extensive washing of the cheese surfaces with brine during ripening, bacteria-ripened cheeses are also known as washedrind cheeses and develop a red-brown glistening appearance. At the beginning of the ripening process lactate, produced by lactic acid bacteria, is metabolized by yeasts, primarily Debaryomyces hansenii and Geotrichum candidum, resulting in an increase in the pH of the cheese surface (Bockelmann et al., 2005; Irlinger and Mounier, 2009; Montel et al., 2014). These processes and the metabolism of yeast growth factors facilitate the growth of coagulase-negative staphylococci such as Staphylococcus equorum and salt-tolerant bacterial communities – mainly coryneforms – such as Corynebacterium spp., Arthrobacter spp., Micrococcus spp., Brevibacterium spp. and Brachybacterium spp. Finally, bacteria and yeasts cover the entire surface of the cheeses. Coryneform bacteria are particularly abundant in cheese rinds (Delbes et al., 2007; Gori et al., 2013; Maoz et al., 2003; Rea et al., 2007). In addition, Gram-negative bacteria such as Halomonas spp., Vibrio spp., and Hafnia alvei have been detected on cheese surfaces (Brennan et al., 2002; Coton et al., 2012; Feurer et al., 2004a; Gori et al., 2013; Irlinger and Mounier, 2009; Ishikawa et al., 2007; Maoz et al., 2003; Mounier et al., 2005; Rea

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

et al., 2007). Austrian Vorarlberger Bergkäse (VB) with a protected designation of origin (PDO) is an artisanal long-ripened hard cheese produced in a long-established cheese producing area in the western part of Austria (Vorarlberg). Cheese production and cheese treatment relies exclusively on traditional techniques including the use of raw cow's milk, the addition of a defined combination of starter cultures, and the brining of cheese wheels either in a brine bath or by dry salting surface treatment (Table 1). Except for the treatments that allow the development of the characteristic yellow-orange rind, no other treatment is applied during the ripening process. Based on market demand and customer requirement, ripening takes at least three months and may last for up to 18 months. Highly similar types of hard cheese, e.g. “Tiroler Bergkäse PDO” or “Allgäuer Bergkäse PDO” are produced in neighboring regions of the Alps. There have been a number of studies of smear- and mold-ripened soft or semi-soft cheeses. However, the structure of microbial communities of raw milk hard washed-rind cheeses such as VB has received only limited attention (Irlinger and Mounier, 2009; Jany and Barbier, 2008; Montel et al., 2014; Ndoye et al., 2011; Quigley et al., 2011). Some recent studies have explored the community structure of cheeses and cheese rinds by highthroughput pyrosequencing (Alegria et al., 2012; Masoud et al., 2011; O'Sullivan et al., 2013; Quigley et al., 2012) but, the cheeses investigated differed from VB. Despite the advantages in terms of throughput and sequencing depth of the novel sequencing technologies, we used rRNA cloning and Sanger sequencing to provide the longest possible sequences and thus the best possible taxonomic assignments, as only few rRNA data are currently available on VB and similar cheeses. We have examined the diversity of bacteria and fungi in rind samples of Austrian VB cheese and the alterations in species abundances associated with different ripening times and production facilities. We hypothesized that VB cheese rinds harbor a high bacterial diversity and that there are significant community shifts with longer ripening times. Based on 16S and 18S rRNA gene cloning and sequencing, we characterized the microbial rind communities from seven pooled cheese rind samples belonging to three cheese producing facilities.

89

concentration of brine only at scheduled intervals and no ripening cultures were added (Table 1). VB ripens for at least three months before it is ready for sale but it is often ripened for six months or longer. The final cheese wheels have a diameter of 50 to 55 cm and a weight of approximately 30 kg. The external appearance of a ready-for-sale product – depending on the ripening time – is a yellow to orange cheese rind with a hard cheese body texture. Although slightly different dairy technologies are used during cheese production and ripening, the cheeses from all three facilities are sold as VB PDO with the main organoleptic and textural differences attributable to differences in ripening time. 2.2. Cheese rind sampling Seven pooled cheese rind samples (A1, A2, A3, B1, B2, C1, C2) from three different cheese producing facilities A, B and C consisting of two to three different ripening cellars in Vorarlberg, Austria were taken in July 2012. For each pooled sample, 25 to 30 different cheese wheels within each ripening cellar were sampled by scraping cheese rinds with sterile scalpels. The entire surface of each cheese wheel at the end of the ripening stage was sampled by a single person. Samples were stored on ice at 4 °C for transport to the laboratory. 2.3. DNA extraction from cheese rind samples Six DNA extractions were performed for each of the pooled samples. The samples were homogenized before DNA extraction. DNA was extracted from 200 mg of pooled cheese rind with the PowerSoil™ DNA Isolation kit (MoBio Laboratories, Carlsbad, California, USA), used in accordance with the manufacturer's instructions. The DNA concentration was measured with a Qubit® fluorometer (Invitrogen, Vienna, Austria). 50 μl of each DNA sample was pooled and precipitated with 3 M sodium acetate and ethanol (96%). The DNA pellet was resuspended in 10 mM Tris–HCl (pH 8) and stored at −20 °C. 2.4. 16S and 18S rRNA gene amplification

2. Materials and methods 2.1. Cheese production VB cheeses were produced from a single morning milking of raw cow's milk. A defined combination of starter cultures (Streptococcus thermophilus, Lactococcus delbrueckii ssp. lactis and Lactobacillus casei) was added to the preheated milk before calf's rennet was added. Curd was cut by using a cheese harp while temperature was gradually raised to 48–50 °C. After a short rest in whey, curd was collected, molded and pressed to drip off residual whey. Cheese wheels were stored in 20–22% brine for 48 h to 60 h before cheeses were stored in the ripening cellars. Except in ripening cellar B1, in which dry salting was used during the first week of ripening, cheese wheels were treated with a distinct

To detect the bacterial microflora of cheese rind samples, 16S rRNA gene PCR was performed using the primers 616 F (5′-AGA GTT TGA TYM TGG CTC-3′, E. coli 16S rRNA positions 8 to 27) (Juretschko et al., 1998) and 1492R (5′-GGY TAC CTT GTT ACG ACT T-3′, E. coli 16S rRNA positions 1492 to 1510) (Lane, 1991). Each PCR reaction was performed in a final volume of 50 μl, containing 0.2 pmol/μl of each primer, 0.8 mM dNTP-mix (TaKaRa, Saint-Germain-en-Laye, France), 1×Ex Taq Buffer (TaKaRa, Saint-Germain-en-Laye, France), 0.025 U TaKaRa Ex Taq HS (TaKaRa, Saint-Germain-en-Laye, France), 1 μl DNA template and DEPC-treated water (Thermo Scientific, Vienna, Austria). 16S rRNA gene PCR was performed under the following conditions: initial denaturation at 95 °C for 5 min, 25 cycles at 94 °C for 40 s, annealing at 52 °C for 40 s, elongation at 72 °C for 1 min and a final elongation at 72 °C for 7 min.

Table 1 Properties of the ripening cellars, cheese treatment and starter cultures. Facility

Ripening cellar

Sample designation

Ripening age [months]

Temperature [°C]

Humidity [%]

Cheese treatment with brine

Brine concentration [%]

Starter cultures for cheese production

A

A1 A2 A3 B1 B2 C1 C2

A1 A2 A3 B1 B2 C1 C2

0–4 0–4 4–18 0–6 6–10 6–14 0–6

13.5 13.5 10 13 13 11 13

96 96 95–96 93–94 93–94 93 94

Daily Daily Weekly 2–3 times a week Weekly Weekly 3–4 times a week

20 20 10 10 15 2–3 22

Streptococcus thermophilus, Lactococcus delbrueckii ssp. lactis, Lactobacillus casei

B C

Streptococcus thermophilus, Lactococcus delbrueckii ssp. lactis, Lactobacillus casei Streptococcus thermophilus, Lactococcus delbrueckii ssp. lactis, Lactobacillus casei

Brine washing of cheese surfaces was performed in all cellars. Dry salting of cheese surfaces was only applied in ripening cellar B1 (during day 0 to 6; each side three times with 45 to 50 g NaCl per side). Starter cultures were obtained from the Federal Institute for Alpine Dairying, Rotholz, Austria.

90

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

To amplify fungi, 18S rRNA gene PCR was performed using the primer combination EK528F (5′-CGG TAA TTC CAG CTC C-3′) and U1391R (5′-GGG CGG TGT GTA CAA RGR-3′) amplifying the hypervariable V3 to V7 regions in the 18S rRNA gene (Edgcomb et al., 2011). The final reaction volume of 50 μl contained: 0.2 pmol/μl of each primer, 0.8 mM dNTP-mix (TaKaRa, Saint-Germain-en-Laye, France), 1× Ex Taq Buffer (TaKaRa, Saint-Germain-en-Laye, France), 0.025 U TaKaRa Ex Taq HS (TaKaRa, Saint-Germain-en-Laye, France), 1 μl DNA template and DEPC-treated water (Thermo Scientific, Vienna, Austria). 18S rRNA gene amplification was carried out under the following PCR conditions: initial denaturation at 95 °C for 5 min, followed by 28 cycles at 94 °C for 40 s, annealing at 56 °C for 40 s, elongation at 72 °C for 1 min and a final elongation at 72 °C for 7 min. PCR products were visualized on a 1.5% agarose gel and stained with SYBR Safe DNA Gel Stain (Invitrogen, Vienna, Austria). Gels were run in 1×TBE-buffer and photographed using Bio Rad Gel Doc 2000 (Bio-Rad Laboratories, Vienna, Austria). All PCR products were stored overnight at 4 °C until further processing. 2.5. Cloning of 16S and 18S rRNA genes Before cloning, PCR amplicons were purified with the GenElute PCR Clean-Up Kit (Sigma-Aldrich, Vienna, Austria), used in accordance with the manufacturer's instructions. Purified PCR products were ligated into the pSC-A-amp/kan PCR cloning vector using the StrataClone PCR Cloning Kit (Agilent Technologies, Vienna, Austria) and transformed into competent cells following the manufacturer's instructions (Agilent Technologies, Vienna, Austria). One gene library was created for each ripening cellar for 16S rRNA and 18S rRNA amplicons. 2.6. 16S and 18S rRNA gene sequencing and 16S rRNA full-length sequencing For 16S rRNA gene libraries, 70 randomly chosen clones per library were sequenced using the primer Univ 1390R (5′-GAC GGG CGG TGT GTA CA-3′), yielding approximately 900 to 1000 bp high-quality sequences per clone. For 18S rRNA gene libraries, 50 clones per library were randomly chosen for sequencing using the primer M13F (5′-TGT AAA ACG ACG GCC AG-3′), yielding 900 to 1000 bp high quality sequences per clone. Sequencing was performed at Microsynth (Balgach, Switzerland). As cloning of 16S rRNA genes was performed with PCR products of approx. 1500 bp, and based on the analysis of the 16S rRNA gene sequencing, near full-length (approx. 1490 bp) 16S rRNA sequences were determined for one representative clone for the most common OTUs 1–25. 2.7. Sequence analyses All sequences were analyzed using mothur (Schloss et al., 2009) with the following parameters: minimum sequence length of 650 bp, and the maximal number of ambiguities set to five. Chimeric sequences were excluded with “chimera.uchime”. The remaining sequences were assigned to a reference taxonomy – the SILVA SSURef 102 reference database (Pruesse et al., 2007) – using either a version for bacterial 16S rRNA gene sequences or a version for eukaryotic sequences (both implemented in mothur) and applying the RDP na ve Bayesian rRNA classifier (confidence threshold = 80%). Sequences were clustered into operational taxonomic units (OTUs) using a distance limit of 0.01 (99% similarity) for 16S rRNA gene libraries and 0.03 (97% similarity) for 18S rRNA gene libraries. This resulting OTU classification was used for all further analyses on all taxonomic levels. ACE and Chao 1 (Chao, 1984) species richness estimates, Simpson evenness index (Simpson, 1949), parametric and non-parametric Shannon diversity index (Shannon, 1948) and rarefaction curves were calculated for all samples with mothur. Heatmaps were created using JColorGrid (Joachimiak et al., 2006).

Near full-length OTUs were classified to type strains using the Ribosomal Database Project (RDP) website (Cole et al., 2014). The chi-square test was used to determine significant changes (p b 0.05) in ripening cellars and calculated by using Microsoft Excel (Microsoft, Seattle, USA). To illustrate community-based shifts between ripening cellars, discriminant analysis was performed and pairwise-correlations were calculated (JMP Pro, SAS Institute, North Carolina, USA). Comparison of community structure was made with Cramer-von Mises test statistic (Libshuff implemented in mothur). To compare 16S rRNA gene sequences, local BlastN (Camacho et al., 2009) searches were undertaken using sequences from three different studies (Gori et al., 2013; Mounier et al., 2005, 2009) as a query (n = 44) against a database of near full-length 16S rRNA sequences retrieved in this study. 2.8. Calculation of phylogenetic trees The phylogenetic relationships of selected OTUs were calculated with ARB (Ludwig et al., 2004) using the SILVA 111 SSU Ref NR release (Quast et al., 2013) applying 50% conservation filters for the respective phyla and the maximum parsimony and RaxML treeing algorithms included in ARB. All trees were calculated with 1000 × bootstrapping. One near full-length representative sequence from each OTU was used for treeing. 2.9. Accession numbers The 16S and 18S rRNA gene sequences obtained have been submitted to EMBL/GenBank and are available under the accession numbers HG795415 to HG795822 and HG795823 to HG796144, respectively. 3. Results 3.1. Sequence analysis reveals diverse bacterial communities in VB cheese rinds A total of 490 clones from 16S rRNA gene libraries and 350 clones from 18S rRNA gene libraries were sequenced. After quality control and removal of chimeric sequences, 408 sequences from 16S rRNA gene libraries (average length: 1040 bp) and 322 sequences from 18S rRNA gene libraries (average length: 922 bp) remained. The number of clones per sample is shown in Table S1. Taxonomic classification revealed 39 bacterial, six fungal and one metazoan OTU. Rarefaction curves and Good's coverage were calculated for 16S and 18S rRNA gene libraries (Fig. 1, Supplementary Tables S1, S2). For all gene libraries, Good's coverage and rarefaction curves suggest that the diversity present was sufficiently covered (Fig. 1). For 18S rRNA gene libraries low species richness and diversity as well as unevenly distributed communities were observed (Supplementary Table S2). 3.2. Analysis of combined 16S and 18S rRNA gene libraries The most abundant OTU in the 16S rRNA gene clones was OTU 1 (Halomonas boliviensis; relative abundance: 18.9%), followed by OTU 2 (Brevibacterium aurantiacum; relative abundance: 16.4%) and OTU 3 (Staphylococcus equorum; relative abundance: 13.5%). No Enterobacteriaceae – possible spoilage microorganisms – and no lactobacilli or lactococci were detected (Supplementary Table S3). Similarly, none of the bacteria added as starter cultures were found in the rinds. Of the 18S rRNA OTUs, six belong to fungi or yeasts and one to metazoa. The most abundant OTUs had highest similarity to: Scopulariopsis brevicaulis (OTU 1, relative abundance: 39.4%), Pyxidiophora arvernensis (OTU 2, relative abundance: 25.8%) and Debaryomyces hansenii (OTU 3, relative abundance: 18.3%) (Supplementary Table S4). The phylogeny of all OTUs and their relative abundance in the ripening cellars are shown in Fig. 2.

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

B 20

Number of OTUs

16

A1 A2 A3 B1 B2 C1 C2

6 5

Number of OTUs

A

91

12 8 4

4

A1 A2 A3 B1 B2 C1 C2

3 2 1

0

0 10

20

30

40

50

60

Number of sequences

10

20

30

40

50

60

Number of sequences

Fig. 1. Diversity of bacterial and fungal communities in cheese rind samples. Rarefaction curves for 16S rRNA gene libraries (0.01 distance, A) and 18S rRNA gene libraries (0.03 distance, B) were calculated for each sample.

To reveal possible co-occurrence patterns among bacterial OTUs, pairwise correlations were calculated for the 25 most abundant OTUs (Supplementary Fig. S1). The analysis showed strong correlations (r N 0.8) between a number of OTUs: e.g. OTU 5 (Advenella kashmirensis), OTU 11 (Yaniella halotolerans), and OTU 16 (Pusillimonas). A strongly negative correlation was found between OTU 2 (B. aurantiacum) and OTU 15 (Leucobacter); weaker (r N 0.6) negative correlations of OTU 2 were found with OTU 14 (Citricoccus zhacaiensis), OTU 16 (Pusillimonas) and OTU 20 (Leucobacter).

3.3. Communities in the cheese rinds from different facilities and with different ripening times exhibit distinct abundance patterns 3.3.1. Comparison between cheese producing facilities Within each sample between ten and 19 and three to five OTUs were found for 16S and 18S rRNA gene libraries, respectively (Supplementary Tables S1, S2). Eight (OTU 1, 2, 3, 4, 6, 7, 8 and 14) OTUs for 16S rRNA gene libraries, and four OTUs (OTU 1, 2, 4 and 6) for 18S rRNA gene libraries were found in all three cheese producing facilities. The shared OTUs correspond to 70% of all clones in the 16S rRNA gene libraries, whereas the four shared OTUs from 18S rRNA gene libraries correspond to 79% of all clones. The overall structures of the 18S rRNA gene communities differ significantly between the three facilities (Libshuff, p b 0.05, data not shown). Principal coordinate analysis for 16S rRNA gene libraries revealed that the structure of the cheese communities differ according to the facility affiliation (Fig. 3A). The examination of 16S rRNA gene libraries showed that cheese rinds from facility A harbored significantly increased ratios of OTU 2 (B. aurantiacum) and OTU 9 (Psychrobacter aquaticus), while the 18S rRNA showed OTU 3 (D. hansenii) significantly increased (1.5, 13.5, and 7.5 fold changes, respectively). In cheese rinds from facility B, OTU 1 (H. boliviensis) and OTU 7 (B. conglomeratum) were significantly increased compared to the other facilities (1.5 and 2.7 fold change), whereas the 18S rRNA OTU 1 (Scopulariopsis brevicaulis) decreased significantly by 2.8 fold and OTU 2 (Pyxidiophora arvernensis) increased significantly (3.3 fold change). 16S rRNA gene libraries showed OTU 5 (A. kashmirensis) and 18S rRNA gene libraries showed OTU 4 (Nectria mariannaeae) to have increased significantly in facility C (10.4 and 6.5 fold change) (Fig. 4, Supplementary Table S5 and S6). Four phyla (Actinobacteria, Proteobacteria, Firmicutes and Bacteroidetes) were found in all three cheese producing facilities, although their relative abundances varied. While cheese rinds from facility A were dominated by Actinobacteria, with a relative abundance of 55.5%, the most common phylum in rinds from facility C was Proteobacteria (45.4%). Rinds from facility B were dominated by both Actinobacteria (40.5%) and Proteobacteria (37.1%). While Firmicutes

was more or less equally represented, Bacteroidetes was the least abundant phylum in all facilities (Fig. 5A).

3.3.2. Comparison between young and old cheeses The results were analyzed with respect to ripening time, differentiating between short (A1, A2, B1, C2; ripening time up to six months) and long (A3, B2, C1; ripening for more than six months) ripening times (Table 1). In the 16S rRNA gene libraries, 19 out of 39 OTUs (92% of all clones) were shared between young and old cheeses. The community structure of 16S rRNA gene libraries differed significantly between old and young cheeses (Libshuff, p = 0.05, data not shown); similarly, a distinction between old and young cheeses could be visualized in principal coordinate analysis (Fig. 3B). OTU 1 (H. boliviensis) was highly abundant in young cheeses and was most frequently found in ripening cellar B1 (61.0%). Cheese rinds from cellars A1 and A2 were dominated by OTU 2 (B. linens, relative abundance: 35.0 and 34.5%). OTU 3 (S. equorum) and OTU 4 (Brevibacterium sp.) were most abundant in old cheeses. Short-ripened cheese rinds harbored a significantly increased (p b 0.05) ratio of OTU 1, OTU 2, OTU 5 and OTU 9 (H. boliviensis, B. aurantiacum, A. kashmirensis, P. aquaticus; 5.7, 2.1, 2.9 and 10.1 fold changes, respectively) and a significantly decreased ratio of OTU 3, OTU 4, OTU 7 and OTU 11 (S. equorum, Brevibacterium pityocampae, Brachybacterium conglomeratum, Y. halotolerans; 3.6, 5.8, 3.7 and 10.4 fold changes, respectively) compared to long-ripened cheeses (Fig. 4, Supplementary Table S5). Five out of seven 18S rRNA OTUs (87% of all clones) were shared between young and old cheeses (Fig. 4). Short-ripened cheeses harbored a significantly increased ratio of OTU 2 and OTU 3 (Pyxidiophora arvernensis, Debaryomyces hansenii; 1.1 and 3.4 fold changes), while OTU 4 (Nectria mariannaeae) could only be detected in short-ripened cheeses. OTU 1 and OTU 5 (Scopulariopsis brevicaulis and Arachnomyces glareosus) were significantly increased in long-ripened cheeses (2.8 and 8.8 fold change) (Fig. 4, Supplementary Table S6). Short-ripened cheeses were dominated by Actinobacteria or Proteobacteria. Within young cheeses, the highest abundance of Actinobacteria (43.3 and 67.3%) was observed in cheese rinds from ripening cellars A1 and A2. In contrast, cheese rinds from ripening cellars B1 and C2 were dominated by Proteobacteria (61 and 77.2%). The occurrence of Firmicutes and Bacteroidetes within the short-ripened cheeses was comparatively low. The long-ripened cheeses in cellar A3, B2 and C1 were dominated by Actinobacteria (56.9, 56.1 and 43.5%) and Firmicutes (22.4, 28.1 and 37.1%). Phyla within the long-ripened cheeses were more homogenously distributed among the three different ripening cellars (Fig. 5B), although the relative abundance of Proteobacteria was lower in old cheeses than in young cheeses.

92

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

Brachybacterium, for OTUs 2, 4, 17, and 19 which affiliated to the genus Brevibacterium and for OTUs 10 and 24 affiliating to Corynebacterium. The genus Leucobacter also contained more than one OTU: OTU 13, 15, and 20 (Supplementary Fig. S2). While some genera comprise relatively closely related OTUs (e.g. Advenella or Corynebacterium), Brevibacterium and Leucobacter in particular contained highly different OTUs.

3.4. Phylogenetic analysis of selected OTUs and considerable novel diversity in cheese rind samples To characterize in more detail the affiliation of some OTUs that belong to genera harboring more than one OTU, phylogenetic trees were calculated for OTUs 7, 22 and 25 which affiliated to the genus

Ripening cellars

6 ense OTU 4 rium daejeon OTU Sphingobacte aeae TU 1 riann 2 O ia ma ulis TU 5 Nectr vica is O TU 3 s bre O U en sis us OT ii en ns

ho

hn

ra

ern

arv

ha

os

re

es

yc

gla

m

es

yo

ar

yc

eb

om

D

m

riop

pula

iop

ac ar

Ac

Sco

xid

Py

Ar St

a ole lba r O erm ans TU OT 2 e n m au se ranti OT U 113 acum U Brevibac OT 19 terium au rantiacum U 17 OTU 2 Brevibacterium pityocampae OTU 4 teriu

ibac

Brev

Bre

vib

act

eriu

mp

ha

lla

nie Ya

Ne

ste

re

lot

nk on ia

s im

30% 45%

sis

en

rd

o ef

d

ar

ig

im

la el

1% 10%

m

oli

gis

en

s gin

sis n as iren on ve hm Ad sillim kas a ll 34 Pu sis e iren ven TU 16 ashm O TU 18 Ad lla k e n O U us e v quatic OT U 5 Ad acter a OT 9 Psychrob OTU iensis OTU 1 Halomonas boliv OTU 31 Halom onas variabilis OTU OT 12 Mari noba O U 28 P cter m se aritim O TU us O TU 27 Id udoalt TU 35 ero iom m ona ari arm 30 Idio in n iglus ar ma Ar t in a osa mb rin ca lico aa no la by ba ss ct a er lis iu m ha em ol yt ic um

e bil ria va e bil m ria riu a v cte m ba riu ne acte - tum ry b comnera Co ryne m lo o teriu g 10 U 24 C hybac ecium um fa OTTU rac acteri O U7B chyb tarium OT U 22 Bra rium alimen OT achybacte OTU 25 Br

er ell a mo sp. bili OT sO U Leuco bacte TU 7 r chro 6 miires istens OTU 1 3 Leucobacter chro miiresistens OTU 15 icius OTU 20 Leucobacter sals 0 U OT 4 ense TU 326 bbeen se O TU 2 3 um gu ri n te e c n ba O U3 4 bee Micro 1 T MG gub ium ei L ae O TU cter cas O in r s roba t u ic is a c M c s a o i n r e roc cu ai Ag Ko ac zh s u cc co tri Ci

aru

OTU 42 Psychroflexus tropicus OTU 3 Staphyl O ococcus OTTU 37 F equorum ackla O U mia O T 8 taba O TUU 21 Alkali TU 3 A cina bac salis teriu 39 8 A topo mk Pu topo stip apii sil sti es s p lim u e on s s icloa ca as uic lis so loac li ali s

A1 A2 A3 B1 B2 C1 C2

Relative abundance

Fig. 2. Relative abundance of OTUs in different ripening cellars. Distribution, taxonomy and relative abundance of all 16S and 18S rRNA gene libraries OTUs for all seven cheese rind samples are shown. The relative abundance of OTUs (%) is proportional to the size of colored circles. For more details, see Table S3.

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

A

B 4

4

facility C

longripened cheeses 2

PC 2 (20.6%)

2

PC 2 (28.4%)

93

0 facility B

-2

-2

-4

0

facility A

-4

-4

-2

0

2

4

PC 1 (29.5%)

shortripened cheeses

-4

-2

0

2

4

PC 1 (32.5%)

Fig. 3. Principal coordinate analyses (PCoA) of cheese rind sample of 16S rRNA gene OTUs. OTUs with more than one sequence assigned were considered as covariates and (A) facilities (• facility A, +facility B, ♦ facility C) or (B) ripening time (+short vs. ♦ long) as variables. PCoA plots are based on the weighted UniFrac distance matrix.

A number of OTUs showed high (N 99%) similarity to their closest related reference type strains in RDP (e.g. OTU 2, 3, 5, 7, 10, 11 and 14). Interestingly, only few of them showed highest similarity to sequences derived from cheese samples, although most genera have been found to occur in cheese. Many OTUs (e.g. OTU 1, 9, 12 and 27) had highest similarity to sequences originating from saline and cold environmental samples. A considerable number of OTUs showed less than 97% similarity to their closest reference strains (e.g. OTU 4, 6, 9, 13, 15, 16, 20 and 21). Using the conservative species definition threshold of 97% similarity in 16S rRNA genes (and not the new threshold of 98.7 to 99%, (Stackebrandt and Ebers, 2006)), all these OTUs probably belong to novel species. However, formal description of new taxa would require additional analyses such as DNA–DNA hybridization or phenotypic characterization. To compare our results with those of other studies, we performed BlastN searches of the 16S rRNA gene sequences of isolates from cheese rinds from three studies of Danish, Irish and French samples from cheese surfaces (Gori et al., 2013; Mounier et al., 2005, 2009) against the 25 nearly full-length OTUs from our study. Interestingly, this analysis revealed a considerable degree of divergence between our study and the other three studies (Supplementary Table S7). Only four query sequences had more than 99% similarity to OTUs from our study, while all other query sequences had less than 99% similarity, suggesting species- to genus-level diversity of bacteria within the different cheese rinds. Unfortunately, most other studies of cheese rind surfaces, have not deposited 16S rRNA gene sequences in GenBank, so broader sequence-based comparisons are not possible. 4. Discussion 4.1. Community structure of rinds of VB cheese compared to those of other cheeses Similarly to previous studies (Maoz et al., 2003; Mounier et al., 2009; Quigley et al., 2013), we found Halomonas (OTU1) to be the most frequent OTU in short-ripened cheeses from all facilities (A, B, C). For some time the occurrence of Halomonas in dairy products was a matter for controversial discussion. Halomonas was suggested to be an indicator for hygienic deficiencies in cheese producing facilities (Maoz et al., 2003), while it has also been suggested that these bacteria, which originally originate in marine environments, are added via sea salt to cheese surfaces during washing and dry salting (Ishikawa et al., 2007). The

dominance of Halomonas in short-ripened cheese rinds is most likely due to more frequent washing with a higher brine concentration. The high abundance of Halomonas in VB cheese rinds suggests an important function during cheese ripening, particularly of young cheese. However, it is still unknown whether Halomonas contributes to the flavor or texture of VB or other cheeses. Coryneform bacteria including Brevibacterium spp., Corynebacterium spp., Arthrobacter spp., Microbacterium spp. and Brachybacterium spp. are important for cheese ripening due to their proteolytic activity and because they produce volatile sulfur compounds or ammonia (Eliskases-Lechner and Ginzinger, 1995). Similarly to other studies (Dolci et al., 2009; Feurer et al., 2004b; Gori et al., 2013; Maoz et al., 2003; Mounier et al., 2005; Rea et al., 2007), we found a high diversity and abundance of coryneforms (13 OTUs, 149 out of 408 clones) in VB cheese rinds. B. linens is an important flavor producer and responsible for the development of the typical red pigmentation of red-smear cheeses, which result from the production of a red carotenoid. It has been described as the most common organism on cheese surfaces (Eliskases-Lechner and Ginzinger, 1995; Rattray and Fox, 1999). A study published in 2004 split the species B. linens into four different species (B. linens, B. aurantiacum, B. antiquum and B. permense (Gavrish et al., 2004)); more recent studies have shown that B. aurantiacum and B. linens, intentionally inoculated on some red-smear, farmhouse or industrially-produced surfaceripened cheeses, almost completely disappear during ripening (Feurer et al., 2004b; Gori et al., 2013; Mounier et al., 2005; Rea et al., 2007). We found B. aurantiacum (OTU 2) to be dominant especially on shortripened cheese rinds of ripening cellars A1 and A2 but also – albeit to a lesser extent – in the long-ripened cheeses of ripening cellar A3. B. aurantiacum was less abundant in facility B and almost absent from facility C, both in short- and long-ripened cheeses. Nevertheless, Brevibacterium is the most abundant genus found in our study (five OTUs, 104 out of 408 clones). Particularly OTU 4 (B. pityocampae) increased significantly in old cheeses, suggesting an important function during ripening of old cheeses. Thus, bacteria of a single genus e.g. Brevibacterium OTUs 2, 4, 17, 19 and in the same habitat behave differently: while OTU 2 decreases in old cheeses, OTU 4 increases, suggesting that the two OTUs have different functions in cheese ripening. Several explanations have been proposed to account for the reduction of B. aurantiacum and B. linens during the ripening period, including the production of inhibitory substances by the surface community, a lack of growth nutrients produced by yeasts, inappropriate growth

94

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

Target OTU gene (number of clones)

Taxonomy

Relative abundance [%]

16S rRNA: similarity (%) to type strain, GenBank accession no.

Cheese producing facilities

18S rRNA: Best Blast Hits (GenBank nr, similarity [%]), GenBank accession no.

16S rRNA

18S rRNA

A

B

C

Ripening cellars young cheeses 0-4

0-4

A1

A2

0-6

B1

0-6

C2

Ripening cellars old cheeses 4-18

A3

6-10

B2

Relative abundance among all cellars [%]

6-14

C1

5

10

15

20

OTU 1 (77)

Halomonas boliviensis LC1 (98%), AY245449

OTU 2 (67)

Brevibacterium aurantiacum NCDO 739 (98%), X76566

OTU 3 (55)

Staphylococcus equorum ATCC 43958 (99%), AB009939

OTU 4 (28)

Brevibacterium pityocampae Tp12 (95%), EU484189

OTU 5 (23)

Advenella kashmirensis WT001 (99%), AJ864470

OTU 6 (19)

Sphingobacterium daejeonense TR6-04 (96%), AB249372

OTU 7 (19)

Brachybacterium conglomeratum JCM 11608 (98%), AB537169

OTU 8 (17)

Alkalibacterium kapii T22-1-2 (98%), AB294171

OTU 9 (14)

Psychrobacter aquaticus CMS 56 (94%), AJ584833

OTU 10 (12)

Corynebacterium variabile DSM 20132 (99%), AJ222815

OTU 11 (9)

Yaniella halotolerans YIM 70085 (99%), AY228479

OTU 12 (8)

Marinobacter maritimus CK47 (98%), AJ704395

OTU 13 (6)

Leucobacter chromiiresistens JG 31 (95%), GU390657

OTU 14 (6)

Citricoccus zhacaiensis FS24 (97%), EU305672

OTU 15 (5)

Leucobacter chromiiresistens JG 31 (96%), GU390657

OTU 16 (5)

Pusillimonas ginsengisoli DCY25T (92%), EF672088

OTU 17 (5)

Brevibacterium aurantiacum NCDO 739 (98%), X76566

OTU 18 (4)

Advenella kashmirensis WT001 (99%), AJ864470

OTU 19 (3)

Brevibacterium permense VKM Ac-2280 (98%), AY243343

OTU 20 (2)

Leucobacter salsicius M1-8 (96%), GQ352403

OTU 21 (2)

Atopostipes suicloacalis PPC79 (97%), AF445248

OTU 22 (2)

Brachybacterium faecium DSM 4810T (99%), X83810

OTU 23 (2)

Nesterenkonia alba CAAS 252 (95%), EU566871

OTU 24 (2)

Corynebacterium variabile DSM 20132 (98%), AJ222815

35 25 15 10 5

1

OTU 25 (2)

Brachybacterium alimentarium CNRZ 925 (99%), X91031

10

40

OTU 1 (127)

Scopulariopsis brevicaulis (99%), JN157617.1

OTU 2 (83)

Pyxidiophora arvernensis AFTOL-ID 2197 (98%), FJ176839.1

OTU 3 (59)

Debaryomyces hansenii (99%), AB628063.1

OTU 4 (40)

Nectria mariannaeae (99%), AB099509.1

OTU 5 (8)

Arachnomyces glareosus CBS 116129 (99%), FJ358341.1

OTU 6 (4)

Acarus immobilis UMMZ BMOC 00-1103-002 AD458 (99%), JQ000104.1

OTU 7 (1)

Starmerella sp. 2-1361 (99%), JX515985.1

20

30

90 70 50 30 10 5

0

0

Fig. 4. Relative abundance of the 25 most abundant 16S rRNA gene OTUs and of all 18S rRNA gene OTUs. The 25 most abundant OTUs from 16S rRNA and all seven OTUs from 18S rRNA gene libraries are listed together with heatmaps showing relative abundances in cheese rinds from different facilities (A, B, C) and in short- (A1, A2, B1, C2) and long-ripened cheeses (A3, B2, C1). Horizontal bar charts show the relative abundances of OTUs over all cellars sampled.

conditions and competitive interactions between the surface colonizers, leading to a decrease of B. aurantiacum and B. linens and an overgrowth by other bacteria such as Arthrobacter spp. or Gram-negative bacteria (Feurer et al., 2004b; Goerges et al., 2008; Gori et al., 2013; Mounier et al., 2005). We found a strongly negative correlation between OTU 2 (B. aurantiacum) and OTU 15 (Leucobacter), suggesting that Leucobacter replace B. aurantiacum in the rinds of old VB cheeses, particularly in cellars B2 and C1. In addition, weaker (r N0.6) negative correlations of

OTU 2 with OTU 16 (Pusillimonas) and OTU 20 (Leucobacter) were found in cellars C2 and C1. Corynebacterium casei and C. variabile have been shown to be dominant on Danish surface-ripened cheeses (Gori et al., 2013). The presence of C. glutamicum on Fontina cheese rinds has been confirmed and the species has been suggested to be among the most important ripening bacteria, increasing in abundance towards the end of ripening (Dolci et al., 2009, 2013). Brine, personnel, and equipment have been

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

B

60

100

Relative abundance [%]

Relative abundance [%]

A

95

50 40 30 20 10 0

Actinobacteria Proteobacteria Firmicutes Bacteroidetes

80 60 40 20 0

A

B

C

Cheese producing facility

A1 0-4 months

A2 0-4 months

B1 0-6 months

Ripening cellars young cheeses

C2 0-6 months

A3 B2 4-18 months 6-10 months

C1 6-14 months

Ripening cellars old cheeses

Fig. 5. Relative abundance of bacterial phyla in cheese rinds. Relative abundances calculated by facility (A) and ripening cellars (B). Short-and long-ripened cheeses are marked by curly brackets.

suggested as possible sources for Corynebacterium (Brennan et al., 2002). We observed C. variabile (OTU 10 and OTU 24) in our samples but did not find them to be dominant, not even towards the end of ripening. The presence of Staphylococcus equorum and several other species of coagulase-negative staphylococci on the surfaces of long-ripened cheeses has been reported previously (Gori et al., 2013; Irlinger et al., 1997). S. equorum may contribute to the aromatic characteristics and the orange pigments of cheeses as a result of the production of extracellular proteolytic and lipolytic enzymes. Staphylococci have been reported to produce antibacterial substances, e.g. S. equorum strain WS 2733 produces a peptide antibiotic (macrococcin P1) on the surface of Raclette cheese, which inhibits the growth of L. monocytogenes (Carnio et al., 2000). We detected a significant increase of S. equorum (OTU 3) in long-ripened cheese rinds of all facilities as well as in younger cheeses of ripening cellar A1. Members of the genus Advenella belong to the Alcaligenaceae family of phylogenetically coherent and physiologically distinct Betaproteobacteria and have been isolated from various human and veterinary clinical samples as well as from soil samples (Gibello et al., 2009). We found Advenella kashmirensis in VB cheese rinds (OTU 5). This species has also been found in Munster cheese (Coton et al., 2012). Facility C had a high level of A. kashmirensis especially in the short-ripened cheeses of ripening cellar C2. Among the moderately abundant OTUs, OTU 6 shows the highest similarity to Sphingobacterium. Sphingobacteria have been described as a component of the flora of raw milk and the dairy environment previously (Hantsis-Zacharov and Halpern, 2007; Quigley et al., 2013; Schmidt et al., 2012), although their role in cheese ripening is unknown. OTU 8 shared high similarity with Alkalibacterium gilvum — a halophilic and alkaliphilic lactic acid bacterium recently described and isolated from European soft and semi-hard cheeses (Ishikawa et al., 2013). It might inhibit growth of L. moncytogenes (Roth et al., 2010). Psychrobacter spp. are frequently found in samples of milk and cheese samples (Deetae et al., 2009; Gori et al., 2013; Irlinger et al., 2012; Mounier et al., 2009) and contribute significantly to cheese ripening, being responsible for the production of aroma compounds and the sensory quality of the cheese (Deetae et al., 2009; Irlinger et al., 2012). We found a high abundance of P. aquaticus (OTU 9) in young cheese rinds of ripening cellar A1. Among the OTUs described in this study, some are affiliated to genera that have to our knowledge not yet been described in milk or dairy products, such as Yaniella (OTU 11), Atopostipes (OTU 21), and Nesterenkonia (OTU 23). It is unclear, whether and how these bacteria contribute to cheese ripening.

Interestingly, no lactobacilli or lactococci were found in the samples analyzed here. This is in accordance with a number of studies of cheese rinds (Brennan et al., 2002; Gori et al., 2013; Mounier et al., 2005, 2009), although other studies have found lactobacilli and lactococci to be abundant in cheese rinds (Dolci et al., 2009; Feurer et al., 2004a; Quigley et al., 2012). These differences might be explained by different sampling techniques, different dairy technologies, different primer specificities, or different approaches to cultivation.

4.2. Association of fungi with cheese rind samples The growth of yeasts has an important role in cheese ripening, relating particularly to the deacidification of cheese surfaces due to lactate assimilation and to the production of alkaline metabolites resulting in a subsequent increase of pH (Corsetti et al., 2001; Irlinger and Mounier, 2009). These physicochemical changes favor the development of less acid- and salt-tolerant bacteria. We found the filamentous fungus Scopulariopsis brevicaulis (OTU 1) and the yeasts Pyxidiophora arvernensis (OTU 2), Debaryomyces hansenii (OTU 3) and Nectria mariannaeae (OTU 4) to be the most abundant fungi in the cheese rind samples analyzed. This finding is consistent with recent results on the surfaces of Danish cheeses, where Debaryomyces hansenii and Scopulariopsis brevicaulis were found to be abundant (Gori et al., 2013). The occurrence of S. brevicaulis in raw milk and cheese has been confirmed by other recent studies (Lavoie et al., 2012; Ropars et al., 2012) and Debaryomyces hansenii has been found on Fontina and various Austrian cheeses (Dolci et al., 2009, 2013; Lopandic et al., 2006; Prillinger et al., 1999). We also observed Pyxidiophora- and Nectria-like OTUs that have not been previously described in cheese, although Nectria has been isolated from raw cow's milk (Lavoie et al., 2012). The results on fungi are necessarily preliminary to some degree and a more complex approach using isolates or different/additional sequence information, such as internal transcribed spacers (ITS) or 28S rRNA would be necessary for better taxonomic classification of the fungi in VB rinds. Interestingly, we detected Acarus immobilis (OTU 6) in four out of seven samples of cheese rinds, albeit at a low abundance. Mites are important for flavor development and ripening in certain cheeses that are traditionally ripened by mites, such as Mimolette (France, Acarus siro) and Milbenkaese (Germany, Tyrolichus casei) (Melnyk et al., 2010). It is conceivable that Acarus immobilis might have a part in the ripening of VB cheese.

96

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97

4.3. Conclusion and outlook The formation of the cheese rind microflora is an essential step in the ripening process of cheeses, with a pronounced impact on organoleptic properties such as flavor, color and texture. We show that the structure of cheese rind communities differed between facilities, independent of ripening time and describe significant shifts in the microbiota between short- and long-ripened cheeses. Parallel cultural studies would be needed to complement our culture-independent analysis to gain deeper insights into the microbial communities of VB. In general, our results for the VB hard cheeses are consistent with previous findings from soft and semi-soft cheeses. Our analyses suggest that the composition of the communities of cheese rinds from different cheese types is distinct, although a cheese rind core microbiota is present and has within-genus-level diversity. This suggests that certain “core” genera are present in cheese rinds irrespective of the type of cheese (e.g. Corynebacterium, Brevibacterium) but within the genera different species occur at variable abundance in rinds from different types of cheese. Different species of a genus may have similar or different functions during cheese ripening. These insights provide a first basis for the analysis of microbial communities in VB and highly similar types of hard cheese, such as “Tiroler Bergkäse PDO” or “Allgäuer Bergkäse PDO”, which are produced in neighboring regions of the Alps. Our results should be useful in improving processing and ripening conditions to enhance the quality and the consistency of the product. Some of the microorganisms present in the rinds of VB cheese might be used as indicators of ripening. Determination of the presence and relative abundance, as well as the effect on cheese quality, of putative ripening indicators might help to better control the ripening process. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijfoodmicro.2014.04.010. Acknowledgements The research was funded by the Landwirtschaftskammer Vorarlberg, Austria. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. We thank Dr. Alexander Tichy and Dr. Daniela Klein-Jöbstl for their support in statistical analysis. The help of Graham Tebb with English language is also greatly acknowledged. References Alegria, A., Szczesny, P., Mayo, B., Bardowski, J., Kowalczyka, M., 2012. Biodiversity in Oscypek, a traditional Polish cheese, determined by culture-dependent and -independent approaches. Appl. Environ. Microbiol. 78, 1890–1898. Bockelmann, W., Willems, K.P., Neve, H., Heller, K.H., 2005. Cultures for the ripening of smear cheeses. Int. Dairy J. 15, 719–732. Brennan, N.M., Ward, A.C., Beresford, T.P., Fox, P.F., Goodfellow, M., Cogan, T.M., 2002. Biodiversity of the bacterial flora on the surface of a smear cheese. Appl. Environ. Microbiol. 68, 820–830. Camacho, C., Coulouris, G., Avagyan, V., Ma, N., Papadopoulos, J., Bealer, K., Madden, T.L., 2009. BLAST+: architecture and applications. BMC Bioinforma. 10, 421. Carnio, M.C., Holtzel, A., Rudolf, M., Henle, T., Jung, G., Scherer, S., 2000. The macrocyclic peptide antibiotic micrococcin P(1) is secreted by the food-borne bacterium Staphylococcus equorum WS 2733 and inhibits Listeria monocytogenes on soft cheese. Appl. Environ. Microbiol. 66, 2378–2384. Chao, A., 1984. Nonparametric-estimation of the number of classes in a population. Scand. J. Stat. 11, 265–270. Cole, J.R., Wang, Q., Fish, J.A., Chai, B., McGarrell, D.M., Sun, Y., Brown, C.T., Porras-Alfaro, A., Kuske, C.R., Tiedje, J.M., 2014. Ribosomal Database Project: data and tools for high throughput rRNA analysis. Nucleic Acids Res. 42, D633–D642. Corsetti, A., Rossi, J., Gobbetti, M., 2001. Interactions between yeasts and bacteria in the smear surface-ripened cheeses. Int. J. Food Microbiol. 69, 1–10. Coton, M., Delbes-Paus, C., Irlinger, F., Desmasures, N., Le Fleche, A., Stahl, V., Montel, M.C., Coton, E., 2012. Diversity and assessment of potential risk factors of Gram-negative isolates associated with French cheeses. Food Microbiol. 29, 88–98. Deetae, P., Spinnler, H.E., Bonnarme, P., Helinck, S., 2009. Growth and aroma contribution of Microbacterium foliorum, Proteus vulgaris and Psychrobacter sp. during ripening in a cheese model medium. Appl. Microbiol. Biotechnol. 82, 169–177.

Delbes, C., Ali-Mandjee, L., Montel, M.C., 2007. Monitoring bacterial communities in raw milk and cheese by culture-dependent and -independent 16S rRNA gene-based analyses. Appl. Environ. Microbiol. 73, 1882–1891. Dolci, P., Barmaz, A., Zenato, S., Pramotton, R., Alessandria, V., Cocolin, L., Rantsiou, K., Ambrosoli, R., 2009. Maturing dynamics of surface microflora in Fontina PDO cheese studied by culture-dependent and -independent methods. J. Appl. Microbiol. 106, 278–287. Dolci, P., Zenato, S., Pramotton, R., Barmaz, A., Alessandria, V., Rantsiou, K., Cocolin, L., 2013. Cheese surface microbiota complexity: RT-PCR-DGGE, a tool for a detailed picture? Int. J. Food Microbiol. 162, 8–12. Edgcomb, V., Orsi, W., Bunge, J., Jeon, S., Christen, R., Leslin, C., Holder, M., Taylor, G.T., Suarez, P., Varela, R., Epstein, S., 2011. Protistan microbial observatory in the Cariaco Basin, Caribbean. I. Pyrosequencing vs Sanger insights into species richness. ISME J. 5, 1344–1356. Eliskases-Lechner, F., Ginzinger, W., 1995. The bacterial flora of surface-ripened cheeses with special regard to coryneforms. Lait 75, 571–583. Feurer, C., Irlinger, F., Spinnler, H.E., Glaser, P., Vallaeys, T., 2004a. Assessment of the rind microbial diversity in a farmhouse-produced vs a pasteurized industrially produced soft red-smear cheese using both cultivation and rDNA-based methods. J. Appl. Microbiol. 97, 546–556. Feurer, C., Vallaeys, T., Corrieu, G., Irlinger, F., 2004b. Does smearing inoculum reflect the bacterial composition of the smear at the end of the ripening of a French soft, redsmear cheese? J. Dairy Sci. 87, 3189–3197. Gavrish, E.Y., Krauzova, V.I., Potekhina, N.V., Karasev, S.G., Plotnikova, E.G., Altyntseva, O. V., Korosteleva, L.A., Evtushenko, L.I., 2004. Three new species of brevibacteria, Brevibacterium antiquum sp nov., Brevibacterium aurantiacum sp nov., and Brevibacterium permense sp nov. Microbiology 73, 176–183. Gibello, A., Vela, A.I., Martin, M., Barra-Caracciolo, A., Grenni, P., Fernandez-Garayzabal, J.F., 2009. Reclassification of the members of the genus Tetrathiobacter Ghosh et al. 2005 to the genus Advenella Coenye et al. 2005. Int. J. Syst. Evol. Microbiol. 59, 1914–1918. Goerges, S., Mounier, J., Rea, M.C., Gelsomino, R., Heise, V., Beduhn, R., Cogan, T.M., Vancanneyt, M., Scherer, S., 2008. Commercial ripening starter microorganisms inoculated into cheese milk do not successfully establish themselves in the resident microbial ripening consortia of a South German red smear cheese. Appl. Environ. Microbiol. 74, 2210–2217. Gori, K., Ryssel, M., Arneborg, N., Jespersen, L., 2013. Isolation and identification of the microbiota of Danish farmhouse and industrially produced surface-ripened cheeses. Microb. Ecol. 65, 602–615. Hantsis-Zacharov, E., Halpern, M., 2007. Culturable psychrotrophic bacterial communities in raw milk and their proteolytic and lipolytic traits. Appl. Environ. Microbiol. 73, 7162–7168. Irlinger, F., Morvan, A., El Solh, N., Bergere, J.L., 1997. Taxonomic characterization of coagulase-negative Stapyhlococci in ripening flora from traditional French cheeses. System. Appl. Microbiol. 20, 319–328. Irlinger, F., Mounier, J., 2009. Microbial interactions in cheese: implications for cheese quality and safety. Curr. Opin. Biotechnol. 20, 142–148. Irlinger, F., Yung, S.A.Y.I., Sarthou, A.S., Delbes-Paus, C., Montel, M.C., Coton, E., Coton, M., Helinck, S., 2012. Ecological and aromatic impact of two Gram-negative bacteria (Psychrobacter celer and Hafnia alvei) inoculated as part of the whole microbial community of an experimental smear soft cheese. Int. J. Food Microbiol. 153, 332–338. Ishikawa, M., Kodama, K., Yasuda, H., Okamoto-Kainuma, A., Koizumi, K., Yamasato, K., 2007. Presence of halophilic and alkaliphilic lactic acid bacteria in various cheeses. Lett. Appl. Microbiol. 44, 308–313. Ishikawa, M., Yamasato, K., Kodama, K., Yasuda, H., Matsuyama, M., Okamoto-Kainuma, A., Koizumi, Y., 2013. Alkalibacterium gilvum sp. nov., slightly halophilic and alkaliphilic lactic acid bacterium isolated from soft and semi-hard cheeses. Int. J. Syst. Evol. Microbiol. 63, 1471–1478. Jany, J.L., Barbier, G., 2008. Culture-independent methods for identifying microbial communities in cheese. Food Microbiol. 25, 839–848. Joachimiak, M.P., Weisman, J.L., May, B., 2006. JColorGrid: software for the visualization of biological measurements. BMC Bioinforma. 7, 225. Juretschko, S., Timmermann, G., Schmid, M., Schleifer, K.H., Pommerening-Roser, A., Koops, H.P., Wagner, M., 1998. Combined molecular and conventional analyses of nitrifying bacterium diversity in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. Appl. Environ. Microbiol. 64, 3042–3051. Lane, D.J., 1991. 16S/23S rRNA sequencing. Nucleic Acid Techniques in Bacterial Systematics. Wiley, New York, pp. 115–147. Lavoie, K., Touchette, M., St-Gelais, D., Labrie, S., 2012. Characterization of the fungal microflora in raw milk and specialty cheeses of the province of Quebec. Dairy Sci. Technol. 92, 455–468. Lopandic, K., Zelger, S., Banszky, L.K., Eliskases-Lechner, F., Prillinger, H., 2006. Identification of yeasts associated with milk products using traditional and molecular techniques. Food Microbiol. 23, 341–350. Ludwig, W., Strunk, O., Westram, R., Richter, L., Meier, H., Yadhukumar, Buchner, A., Lai, T., Steppi, S., Jobb, G., Forster, W., Brettske, I., Gerber, S., Ginhart, A.W., Gross, O., Grumann, S., Hermann, S., Jost, R., Konig, A., Liss, T., Lussmann, R., May, M., Nonhoff, B., Reichel, B., Strehlow, R., Stamatakis, A., Stuckmann, N., Vilbig, A., Lenke, M., Ludwig, T., Bode, A., Schleifer, K.H., 2004. ARB: a software environment for sequence data. Nucleic Acids Res. 32, 1363–1371. Maoz, A., Mayr, R., Scherer, S., 2003. Temporal stability and biodiversity of two complex antilisterial cheese-ripening microbial consortia. Appl. Environ. Microbiol. 69, 4012–4018. Masoud, W., Takamiya, M., Vogensen, F.K., Lillevang, S., Abu Al-Soud, W., Sorensen, S.J., Jakobsen, M., 2011. Characterization of bacterial populations in Danish raw milk cheeses made with different starter cultures by denaturating gradient gel electrophoresis and pyrosequencing. Int. Dairy J. 21, 142–148.

E. Schornsteiner et al. / International Journal of Food Microbiology 180 (2014) 88–97 Melnyk, J.P., Smith, A., Scott-Dupree, C., Marcone, M.F., Hill, A., 2010. Identification of cheese mite species inoculated on Mimolette and Milbenkase cheese through cryogenic scanning electron microscopy. J. Dairy Sci. 93, 3461–3468. Montel, M.C., Buchin, S., Mallet, A., Delbes-Paus, C., Vuitton, D.A., Desmasures, N., Berthier, F., 2014. Traditional cheeses: rich and diverse microbiota with associated benefits. Int. J. Food Microbiol. 177C, 136–154. Mounier, J., Gelsomino, R., Goerges, S., Vancanneyt, M., Vandemeulebroecke, K., Hoste, B., Scherer, S., Swings, J., Fitzgerald, G.F., Cogan, T.M., 2005. Surface microflora of four smear-ripened cheeses. Appl. Environ. Microbiol. 71, 6489–6500. Mounier, J., Monnet, C., Jacques, N., Antoinette, A., Irlinger, F., 2009. Assessment of the microbial diversity at the surface of Livarot cheese using culture-dependent and independent approaches. Int. J. Food Microbiol. 133, 31–37. Ndoye, B., Rasolofo, E.A., LaPointe, G., Roy, D., 2011. A review of the molecular approaches to investigate the diversity and activity of cheese microbiota. Dairy Sci. Technol. 91, 495–524. O'Sullivan, D.J., Giblin, L., McSweeney, P.L., Sheehan, J.J., Cotter, P.D., 2013. Nucleic acidbased approaches to investigate microbial-related cheese quality defects. Front. Microbiol. 4, 1. Prillinger, H., Molnar, O., Eliskases-Lechner, F., Lopandic, K., 1999. Phenotypic and genotypic identification of yeasts from cheese. Antonie Van Leeuwenhoek 75, 267–283. Pruesse, E., Quast, C., Knittel, K., Fuchs, B.M., Ludwig, W., Peplies, J., Glockner, F.O., 2007. SILVA: a comprehensive online resource for quality checked and aligned ribosomal RNA sequence data compatible with ARB. Nucleic Acids Res. 35, 7188–7196. Quast, C., Pruesse, E., Yilmaz, P., Gerken, J., Schweer, T., Yarza, P., Peplies, J., Glockner, F.O., 2013. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41, D590–D596. Quigley, L., O'Sullivan, O., Beresford, T.P., Ross, R.P., Fitzgerald, G.F., Cotter, P.D., 2011. Molecular approaches to analysing the microbial composition of raw milk and raw milk cheese. Int. J. Food Microbiol. 150, 81–94.

97

Quigley, L., O'Sullivan, O., Beresford, T.P., Ross, R.P., Fitzgerald, G.F., Cotter, P.D., 2012. High-throughput sequencing for detection of subpopulations of bacteria not previously associated with artisanal cheeses. Appl. Environ. Microbiol. 78, 5717–5723. Quigley, L., O'Sullivan, O., Stanton, C., Beresford, T.P., Ross, R.P., Fitzgerald, G.F., Cotter, P.D., 2013. The complex microbiota of raw milk. FEMS Microbiol. Rev. 37, 664–698. Rattray, F.P., Fox, P.F., 1999. Aspects of enzymology and biochemical properties of Brevibacterium linens relevant to cheese ripening: a review. J. Dairy Sci. 82, 891–909. Rea, M.C., Gorges, S., Gelsomino, R., Brennan, N.M., Mounier, J., Vancanneyt, M., Scherer, S., Swings, J., Cogan, T.M., 2007. Stability of the biodiversity of the surface consortia of Gubbeen, a red-smear cheese. J. Dairy Sci. 90, 2200–2210. Ropars, J., Cruaud, C., Lacoste, S., Dupont, J., 2012. A taxonomic and ecological overview of cheese fungi. Int. J. Food Microbiol. 155, 199–210. Roth, E., Miescher Schwenninger, S., Hasler, M., Eugster-Meier, E., Lacroix, C., 2010. Population dynamics of two antilisterial cheese surface consortia revealed by temporal temperature gradient gel electrophoresis. BMC Microbiol. 10, 74. Schloss, P.D., Westcott, S.L., Ryabin, T., Hall, J.R., Hartmann, M., Hollister, E.B., Lesniewski, R.A., Oakley, B.B., Parks, D.H., Robinson, C.J., Sahl, J.W., Stres, B., Thallinger, G.G., Van Horn, D.J., Weber, C.F., 2009. Introducing mothur: open-source, platformindependent, community-supported software for describing and comparing microbial communities. Appl. Environ. Microbiol. 75, 7537–7541. Schmidt, V.S.J., Wenning, M., Scherer, S., 2012. Sphingobacterium lactis sp nov and Sphingobacterium alimentarium sp nov., isolated from raw milk and a dairy environment. Int. J. Syst. Evol. Microbiol. 62, 1506–1511. Shannon, C.E., 1948. A mathematical theory of communication. Bell Syst. Tech. J. 27, 623–656. Simpson, E.H., 1949. Measurement of diversity. Nature 163, 688. Stackebrandt, E., Ebers, J., 2006. Taxonomic parameters revisited: tarnished gold standards. Microbiol. Today 6–9 (November).