当前位置:首页 >> 信息与通信 >>

FTIR-based polyphasic identification of lactic acid bacteria isolated from traditional

Food Microbiology 28 (2011) 76e83

Contents lists available at ScienceDirect

Food Microbiology
journal homepage: www.elsevier.com/locate/fm

FTIR-based polyphasic identi?cation of lactic acid bacteria isolated from traditional Greek Graviera cheese
John Samelis a, *, Anne Bleicher b,1, Céline Delbès-Paus c, Athanasia Kakouri a, Klaus Neuhaus b, Marie-Christine Montel c
a b c

National Agricultural Research Foundation, Dairy Research Institute, Katsikas, 45221 Ioannina, Greece Technische Universit?t München, ZIEL, Abt. Microbiologie, Weihenstephaner Berg 3, 85356 Freising, Germany Unité de Recherches Fromagères, URF-545 INRA, 20 C?te de Reyne, F-15000 Aurillac, France

a r t i c l e i n f o
Article history: Received 3 May 2010 Received in revised form 16 August 2010 Accepted 17 August 2010 Available online 21 August 2010 Keywords: Graviera cheese Lactic acid bacteria Polyphasic taxonomy FTIR RFLP SSCP

a b s t r a c t
This study used a combination of phenotypic, physical (Fourier Transformed Infra-Red [FTIR] spectroscopy) and molecular (RFLP and SSCP analysis of 16S rRNA genes) methods to identify the lactic acid bacteria (LAB) ?ora present in traditional Greek Graviera cheese after ?ve weeks of ripening. A total of 300 isolates collected from high dilution plates of TSAYE (incubated at 30  C), M-17 (22  C) and M-17 (42  C) agar media were clustered by FTIR and then representative strains of each cluster were crossidenti?ed blindly by all methods. Based on their FTIR spectra, 282 isolates were LAB grouped in 28 clusters. The LAB species identi?ed and their prevalence in the cheese samples were: Lactobacillus casei/ paracasei (68.8%), Lactobacillus plantarum (19.5%), Streptococcus thermophilus (8.9%), Enterococcus faecium (2.1%), and Lactococcus lactis (0.7%). Also, Staphylococcus equorum (11 isolates), Corynebacterium sp. (5 isolates) and Brevibacterium sp. (1 isolate) were recovered from TSAYE. Comparative identi?cation results showed that phenotypic and molecular methods were in mutual agreement as regards the LAB species identi?ed. The present polyphasic identi?cation approach based on rapid FTIR screening of 10-fold more isolates than a previous classical identi?cation approach allowed or improved detection of few sub-dominant species; however the predominant LAB species in the cheese samples were the same with both approaches. ? 2010 Elsevier Ltd. All rights reserved.

1. Introduction Polyphasic taxonomy has been recognized as a consensus approach to bacterial systematics, and is particularly useful for identi?cation and classi?cation of lactic acid bacteria (LAB) involved in food-associated ecosystems (Vandamme et al., 1996). Various culture-dependent, phenotypic, including chemotaxonomic, and genotypic methods have been applied in different combinations for studying LAB ecology and diversity in raw milk and traditional cheeses made mainly from raw milk (Fitzsimons et al., 1999; Bizzarro et al., 2000; De Angelis et al., 2001; Weinrichter et al., 2001; Bouton et al., 2002; Callon et al., 2004; Georgieva et al., 2008). In recent years, DNA-based, culture-independent methods,

* Corresponding author. Tel.: ?30 2 651 094789; fax: ?30 2 651 092523. E-mail address: jsam@otenet.gr (J. Samelis). 1 Present address: Leibniz Institute for Natural Product Research and Infection Biology e.V., Hans-Kn?ll-Institut, Beutenbergstr. 11a, 07745 Jena, Germany. 0740-0020/$ e see front matter ? 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.fm.2010.08.009

such as Single Strand Conformation Polymorphism (SSCP), are used as powerful tools for evaluating microbial (LAB) communities in foods (Giraffa and Neviani, 2001), including raw milk (Callon et al., 2007) and traditional cheeses (Randazzo et al., 2002; Duthoit et al., 2003; Dolci et al., 2008; Martin-Platero et al., 2009). Despite their existing limitations, culture-independent methods have attracted signi?cant attention because culture-dependent ones, either phenotypic or genotypic, are laborious and timeconsuming requiring bacterial isolations, puri?cation and maintenance of puri?ed stock isolates prior to identi?cation. The greater the number of isolates, the more comprehensive and reliable a microbiological ecology study is, and most culture-dependent methods are hampered because of this requirement. Besides, by the classical phenotypic identi?cation methods, isolates belonging to bacterial (LAB) genera or species of high phenotypic heterogeneity may be misidenti?ed or unidenti?ed requiring chemotaxonomic or molecular approaches for con?rming and/or resolving their taxonomy (Vandamme et al., 1996; Moschetti et al., 2001). Thereby, any analytical tool that could simultaneously provide rapid and

J. Samelis et al. / Food Microbiology 28 (2011) 76e83


robust screening plus identi?cation of large amounts of isolates would be useful; one such tool is Fourier Transform Infra-Red (FTIR) spectroscopy. Since bacterial FTIR spectra give a global picture of whole cellular components (fatty acids, proteins and polysaccharides plus nucleic acids), they have been proposed as a link between phenotypic and genotypic methods (Amiel et al., 2001). The capacity of FTIR spectroscopy for identi?cation and classi?cation of dairy LAB, even for the discrimination of closely related species, such as Streptococcus thermophilus from Streptococcus salivarius, or Lactobacillus casei/paracasei from Lactobacillus zeae and Lactobacillus rhamnosus, has been proven (Amiel et al., 2000, 2001). The evolution of Lactococcus strains during ripening of Brie cheese has also been followed using FTIR (Le?er et al., 2000). Based on FTIR spectra, Lactobacillus isolates from homemade cheeses, Swiss cheeses with different Lactobacillus adjunct strains added, and pathogenic from non-pathogenic species of Listeria and Staphylococcus have been discriminated in more recent studies (Lamprell et al., 2006; Savic et al., 2008; Rebuffo-Scheer et al., 2008; Chen et al., 2009). Graviera is the most popular traditional Greek hard cooked cheese (Litopoulou-Tzanetaki and Tzanetakis, 2007). Several varieties are produced in different regions of Greece from ewes’, ewes’ mixed with goats’, or cows’ milk, and three of them produced in the islands of Crete and Naxos and the mountain area of Agrafa have PDO status (Anonymous, 2004). Recently, we evaluated the microbiological quality and safety of Graviera cheeses traditionally manufactured at semi-industrial plant scale from thermized milk with addition of a product-speci?c commercial starter culture (CSC). All cheeses were stable and hygienically safe in compliance with the current European Union microbiological regulatory criteria (European Commission, 2007). Several challenge studies showed that neither Listeria monocytogenes nor enterotoxigenic Staphylococcus aureus strains could grow in the cheese core during ripening and/or on the cheese surface post-ripening (Giannou et al., 2009a; Samelis et al., 2009b, 2009c). A more recent study (Samelis et al., 2010) indicated that the microbial ?ora of Graviera cheeses after ?ve weeks of ripening was dominated by non-starter (NSLAB) mesophilic lactobacilli, Lb. casei/paracasei (67.5%) and Lactobacillus plantarum (16.3%). Conversely, S. thermophilus, Lactococcus lactis and Leuconostoc sp. included in the CSC were isolated from the ripened cheeses at frequencies as low as 3.8%, 0.6% and 1.9%, respectively. Enterococcus faecium (9.4%) and Enterococcus durans (0.6%) were isolated among the main LAB ?ora from two cheese batches; in general, enterococci were present in all batches at 10- to 100-fold lower populations than mesophilic lactobacilli (Samelis et al., 2010). However, the above results on microbial composition of ripened Graviera cheeses were based on the isolation of a limited number of isolates (10 from each agar medium per batch) followed by their identi?cation by phenotypic methods and criteria only. The present study was therefore undertaken to validate previous phenotypic identi?cation data on the LAB ?ora composition of traditional Graviera cheese and recover additional species potentially underlying the dominant species identi?ed by Samelis et al. (2010). For this purpose, a polyphasic approach was applied to Graviera cheeses selected on the basis of their low LAB species phenotypic diversity. Compared to our previous classical approach (Samelis et al., 2010), this study used ten-fold more cheese isolates, which were ?rst clustered by FTIR. Next, isolates representing FTIR clusters were identi?ed by restriction fragment length polymorphism (RFLP), SSCP, and biochemical methods. Further, the FTIR spectra of the cheese isolates were compared to a spectrum library to obtain FTIR identi?cation. All methods were applied blindly within the laboratories participating in this study in order to evaluate the relevance and mutual agreement between methods, and the capacity of each method to accurately identify different LAB species present in Graviera cheese.

2. Materials and methods 2.1. Graviera cheeses Two Graviera cheeses (T and R) derived from one commercial production run (batch A) after ?ve weeks of ripening were selected for the polyphasic identi?cation study among four batches previously analyzed by phenotypic methods (Samelis et al., 2010). Both cheeses were produced in a local semi-industrial plant (Pappas Bros., Filippiada, Epirus) from thermized (63  C for 30 s) ewes’/ goats’ (90:10) milk. A product-speci?c CSC containing S. thermophilus, Lc. lactis subsp. lactis, Lc. lactis subsp. lactis var. diacetylactis and Leuconostoc strains of natural origin (GR02, Mo?n Alce Group, Novara, Italy) was added to the cooled milk before rennet addition and curdling, as described by Samelis et al. (2009b, 2010). After standard cheese cooking, molding, pressing, brining and draining operations, Graviera-T (GR-T) cheese was ripened in the plant’s ripening room at temperatures of 17e19  C and relative humidity (RH) of 90e92% with manual monitoring of the air ventilation, whereas Graviera-R (GR-R) cheese was ripened in a controlled pilot ripening room under constantly monitored conditions of temperature (17.5 ? 0.1  C), RH (96.0 ? 2.3%) and air ventilation (continuous at 1.5 m/s) (Samelis et al., 2010).

2.2. Colony isolation procedure for FTIR analysis A total of 50 colonies from one high dilution plate of each of TSAYE (incubated at 30  C), M-17 (incubated at 22  C) and M-17 (incubated at 42  C) agar media (Samelis et al., 2010) were isolated for FTIR analysis. All isolation media were purchased from LAB M (Bury, UK). The colonies were collected randomly from the agar surface with the aid of pre-sterilized tooth sticks. No isolations from MRS agar (LAB M) plates incubated at 30 or 45  C were made for this study because those colonies had grown inside the agar layer after pour plating of the cheese samples, and thus, it was dif?cult to pick 50 of them from single plates. Since bacterial FTIR spectra are known being in?uenced by culture medium and incubation conditions, two different standardized experimental protocols were applied to the cheese isolates to con?rm comparability with reference spectra in the FTIR library. Speci?cally, isolates from TSAYE plates, considered to represent the total mesophilic cheese ?ora, were streaked on casein soy (CASO) agar (Merck, Darmstadt, Germany) incubated at 30  C aerobically, whereas all isolates from M-17 agar plates, considered to represent the mesophilic (22  C) and thermophilic (42  C) cheese LAB ?ora, were streaked on APT agar (Merck) incubated at 34  C anaerobically. In the above manner, 300 colonies in total (150 from GR-T and 150 from GR-R) were collected. During previous routine isolation procedures 10-fold less colonies (5 colonies from each medium ? 3 media ? 2 cheeses ? 30 colonies in total) were isolated from the duplicate agar plates of the same cheese samples and subjected to biochemical identi?cation (Samelis et al., 2010). Speci?cally, the number of colonies grown on high (?6) dilution TSAYE, M-17/22  C and M-17/42  C agar plates used for the FTIR isolation procedure were 108, 71 and 66, and 174, 85 and 50 for GR-T and GR-R samples, respectively. Thus, by picking 50 colonies from each plate, the FTIR method was actually based on 28.7 and 46.3%, 58.8 and 70.4%, and 75.8 and 100.0% of the microbial (LAB) populations on TSAYE, M-17/ 22  C and M-17/42  C plates, respectively. Conversely, by random picking of ?ve colonies only from each agar plate, the corresponding percentages by the classical method had been as low as 3.4 and 3.7% (TSAYE), 4.4 and 5.0% (M-17/22  C) and 6.1 and 10.0% (M-17/42  C) for GR-T and GR-R cheeses, respectively.


J. Samelis et al. / Food Microbiology 28 (2011) 76e83

2.3. FTIR analysis The microbial ?ora composition of the GR-T and GR-R cheeses was studied in detail using FTIR spectroscopy. Sample preparation to obtain reliable infrared spectra was carried out as described previously (Kümmerle et al., 1998; Oberreuter et al., 2002). Two or, if a spectrum did not match the quality criteria, more independent measurements were conducted using an HTS-XT FTIR spectrometer (Bruker, Karlsruhe, Germany). For data processing, OPUS software v.6 (Bruker) was used. The spectral distances of the isolates, re?ecting the similarity of two spectra by comparing the size of non-overlapping areas, were used to discriminate the isolates at the strain level. A subsequent cluster analysis was performed to visualize the abundance ratio of strains resulting in a dendrogram calculated according to the average linkage algorithm. For identi?cation purposes, the spectra were compared to a library containing 3287 spectra in the CASO database and 650 spectra in the APT database, respectively. 2.4. Biochemical identi?cation Representative isolates of each of the FTIR clusters were ?rst checked to be LAB and then identi?ed biochemically at species level; their identi?cations were compared with those of the corresponding 30 isolates from TSAYE, M-17/22  C and M-17/42  C media previously identi?ed by the classical method; isolates were checked for purity, maintained and subcultured for testing as described by Samelis et al. (2010). Biochemical LAB identi?cation was conducted according to established phenotypic criteria, and the methods described in previous studies (Giannou et al., 2009b; Samelis et al., 2009a, 2010). All tests were done in duplicate for each isolate. Non-LAB isolates representing FTIR clusters were subjected to microscopic examination and rapid testing for Gram strain, catalase and oxidase reactions only. 2.5. Molecular identi?cation The cheese LAB isolates representing all clusters formed using FTIR were subjected to molecular identi?cation also. The isolates were ?rst subjected to RFLP analysis of the 16S rRNA gene as described by Callon et al. (2007). They were assigned to a genus after in silico comparison of their restriction patterns to a pattern library from reference strains. Afterwards the species were determined by analyzing the variable V3 and V2 regions of 16S rRNA genes by SSCP and comparing the SSCP pro?le of the isolates with that of reference strains from the same genus as described by Duthoit et al. (2003). Representative non-LAB isolates were identi?ed by 16S rRNA gene analysis also. Staphylococcus isolates in particular were identi?ed on molecular level using rpoB gene sequencing (Drancourt and Raoult, 2002). 3. Results and discussion 3.1. Clustering and identi?cation of cheese LAB isolates by FTIR From 300 colonies in total isolated from TSAYE, M-17/22  C and M-17/42  C agar plates, 282 isolates (148 from GR-T and 134 from GR-R) were LAB clustered by FTIR and cross-identi?ed by FTIR, biochemical, RFLP and SSCP methods (Table 1). The remaining 17 colonies were non-LAB exclusively isolated from TSAYE agar plates (2 from GR-T and 15 from GR-R), whereas one GR-R isolate from M-17/42  C was not viable upon subculturing for FTIR analysis. All non-LAB isolates were grouped separately from the LAB ?ora using FTIR and were identi?ed using 16S rRNA or rpoB gene sequences, as described in paragraph 3.5 below.

Based on at least two independent FTIR measurements for each isolate, the 282 LAB isolates were grouped into 27 clusters, numbered separately for each isolation medium, plus one isolate, GR-R-46, which gave an “intermixed” spectrum between the M-17/ 42  C clusters 6 and 7 (Table 1). Next to the number of each cluster the corresponding number of isolates from GR-T, GR-R, and in total, and their presumptive identi?cation by FTIR, are presented. It should be noted that clusters with different numbers differ in their FTIR spectra from the other clusters formed within the isolates from the same isolation medium only, in Table 1. This is because, as expected, the FTIR spectra of the isolates were dependent on the culture medium and incubation conditions, i.e., anaerobically on APT vs. aerobically on CASO, which affected their metabolism and cellular component synthesis. FTIR analysis is principally based on the discrimination of bacteria in their “?ngerprint region”, a special window of infrared light absorption, where differences are most signi?cant. Consequently, the spectra of the 282 LAB isolates were clustered based on their differences, not their similarities. Since some isolates from TSAYE did not result in a “clear-cut” cluster, 2e3 representative strains were selected for biochemical and molecular identi?cation for clusters 7/8 and 13e15, respectively (Table 1). In general terms, FTIR spectroscopy was found to be a useful tool for rapid screening and clustering of the cheese LAB isolates as low as at strain level. Also, except for most LAB isolates from TSAYE which were misidenti?ed on CASO agar for reasons discussed below, the FTIR spectra provided accurate identi?cations at genus level. Further the FTIR cluster analysis was able to assign several strains of the same species to different clusters; for example, cluster 1 and 2 from M-17/42  C represented by strains TF1 and TF2. In some cases the FTIR spectra identi?cations at species level were unclear with the current database. However, this was quite acceptable because most of the intermixed identities referred to LAB belonging to closely related groups of species. In few cases the FTIR spectroscopy gave more than one possible identity at the genus level, i.e., for clusters represented by TF8, TF14, TF19, TF21e22 and TF24 isolates (Table 1). An adaptation of the spectra libraries to the isolation source (e.g., hard cheese) is mandatory for correct identi?cation results. 3.2. Identi?cation of Graviera cheese LAB isolates by molecular methods Comparative identi?cation results of the LAB isolates by molecular methods are shown in Table 1 also. The RFLP analysis of the isolates representing the FTIR clusters (TF1eTF29) was proven a powerful tool for their grouping, with few exceptions, at genus level. Comparison of the restriction patterns with reference strains resulted in eight different RFLP clusters. Cluster 1 (isolates TF1, TF2, TF9, TF10, TF17, TF21 and TF22), cluster 2 (TF3, TF4, TF15, TF27, TF28, TF29), cluster 5 (TF11, TF12, TF23) and cluster 6 (TF16, TF19) were presumably assigned to the genus Lactobacillus, and clusters 4 (TF7, TF18) and 7 (TF25, TF26) to the genus Enterococcus. Cluster 3 (TF5, TF6, TF8, TF14) was of doubtful identi?cation since the RFLP patterns of those isolates could be assigned to either S. thermophilus or Staphylococcus vitulinus. Also, cluster 8, which included TF24 only e a single GR-R isolate from TSAYE, was presumptively identi?ed as Lc. lactis or Lactobacillus delbruecki. Finally, isolate TF13 was not viable upon subculturing for molecular identi?cation (Table 1). Following identi?cation to genus by RFLP analysis, all representative LAB isolates were identi?ed to species using SSCP (V3 and V2 regions) analysis of 16S rRNA genes. All isolates of RFLP cluster 3, with doubtful identi?cation as S. thermophilus or Staph. vitulinus, were identi?ed as S. thermophilus by SSCP analysis. Also, isolate TF24 (RFLP cluster 8) was identi?ed as Lc. lactis. For all the other RFLP clusters, identi?cation at the genus level by RFLP was

Table 1 FTIR-based clustering of 282 lactic acid bacteria (LAB) isolates from traditional Greek Graviera cheese ripened under two ripening conditions,a and blind cross-identi?cation of 28 representative isolates by FTIR, classical and molecular methods. Isolation agar M-17/42  C FTIR cluster 1 2 3 4 6 GR-T isolates 34 0 10 0 0 GR-R isolates 21 7 7 1 5 Total isolates 55 7 17 1 5 Presumptive FTIR cluster identi?cation Lactobacillus paracasei/agilis/casei Lactobacillus paracasei/agilis/casei Lactobacillus plantarum/pentosus Lb. plantarum Streptococcus thermophilus Representative cluster isolates TF1 TF2 TF3 TF4 TF5 Biochemical identi?cation Lactobacillus casei (group I) Lactobacillus casei (group I) Lactobacillus plantarum Lactobacillus plantarum Streptococcus thermophilus RFLP cluster 1 1 2 2 3 Presumptive RFLP identi?cation Lactobacillus Lactobacillus Lactobacillus Lactobacillus Staphylococcus vitulinus, or Streptococcus thermophilus Staphylococcus vitulinus, or Streptococcus thermophilus Enterococcus Staphylococcus vitulinus, or Streptococcus thermophilus Lactobacillus Lactobacillus Lactobacillus Lactobacillus Non-viable/NA Staphylococcus vitulinus, or Streptococcus thermophilus Lactobacillus Lactobacillus Lactobacillus Enterococcus Lactobacillus SSCP identi?cation (V3 ? V2 region) Lactobacillus casei Lactobacillus casei Lactobacillus plantarum Lactobacillus plantarum Streptococcus thermophilus





Streptococcus thermophilus


Streptococcus thermophilus


Streptococcus thermophilus

8 Isolate R46 in between clusters 6e7

2 0

0 1

2 1

Enterococcus faecium/mundti Streptococcus thermophilus or Leuconostoc mesenteroides


Enterococcus faecium (type B) Streptococcus thermophilus

4 3

Enterococcus faecium Streptococcus thermophilus J. Samelis et al. / Food Microbiology 28 (2011) 76e83

Sub-total M-17/22  C 1 2 3 4 5 6

50 29 0 2 2 0 2

49 9 22 1 2 1 6

99 38 22 3 4 1 8 Lactobacillus paracasei/agilis/casei Lactobacillus paracasei/agilis/casei Lactobacillus rhamnosus/casei Lactobacillus casei paracasei/acidophilus Lactococcus lactis Streptococcus thermophilus or Lactococcus lactis/raf?nolactis Lactobacillus plantarum/pentosus Lb. plantarum Lactobacillus plantarum/pentosus Enterococcus faecium/mundti Lactobacillus curvatus/graminis, or Leuconostoc mesenteroides Arthrobacter nicotovorans, or Vagococcus carniphilus Lactococcus lactis Vagococcus carniphilus, or Lactococcus lactis TF9 TF10 TF11 TF12 TF13 TF14 Lactobacillus Lactobacillus Lactobacillus Lactobacillus casei casei casei casei (group (group (group (group I) I) II) I) 1 1 5 5 NA 3 Lactobacillus Lactobacillus Lactobacillus Lactobacillus casei casei casei casei

Lactococcus lactis/garviae Streptococcus thermophilus

Non-viable/NA Streptococcus thermophilus

7 8 9 10 11 Sub-total TSAYE/30 C

11 1 1 2 0 50 35 1 0

7 0 0 0 2 50 24 3 1

18 1 1 2 2 100 59 4 1

TF15 TF16 TF17 TF18 TF19

Lactobacillus plantarum Lactobacillus plantarum Lactobacillus plantarum Enterococcus faecium(type C) Lactobacillus contaminated with catalase-positive cocci Lactobacillus casei (group I) Lactobacillus casei (group II) Lactococcus lactis

2 6 1 4 6

Lactobacillus plantarum Lactobacillus plantarum Lactobacillus plantarum Enterococcus faecium Lactobacillus casei

7/8 9 10

TF21, TF22 TF23 TF24

1, 1 5 8

Lactobacillus Lactobacillus Lactococcus lactis or Lactobacillus delbruecki Enterococcus Enterococcus Lactobacillus

Lactobacillus casei Lactobacillus casei Lactococcus lactis

11 12 13e15 Sub-total Total viable isolates

0 1 11 48 148

1 0 6 35 134

1 1 17 83 282

Vagococcus carniphilus Enterococcus hirae/faecalis Bacillus licheniformis

TF25 TF26 TF27, TF28, TF29

Enterococcus faecium (type A) Enterococcus faecium(type A) Lactobacillus plantarum

7 7 2,2,2

Enterococcus faecium Enterococcus faecium Lactobacillus plantarum

GR-T, Graviera-T cheese ripened in the commercial plant’s ripening room under fairly standardized environmental conditions; GR-R, Graviera-R cheese ripened in a pilot ripening room under constantly monitored environmental conditions.



J. Samelis et al. / Food Microbiology 28 (2011) 76e83 Table 2 Differentiating biochemical reactionsa between Enterococcus faecium, mesophilic Lactobacillus, and Lactococcus lactis and Streptococcus thermophilus isolates in comparison with their RFLP patterns and SSCP species identi?cation. Representative isolates and RFLP clusters in bold format indicate discrepancies between biochemical and molecular identi?cation. a All representative LAB isolates were lactose-positive, xylose-negative, and homofermentative; except for E. faecium, all were arginine-negative; except for S. thermophilus all grew at 15  C; only S. thermophilus and E. faecium grew well at 45  C; all fermented strongly cellobiose, fructose, galactose, maltose and ribose, except for S. thermophilus isolates.


con?rmed by SSCP analysis (Table 1). Graviera isolates represented ?ve LAB species commonly associated with cheese, particularly hard or semi-hard cheeses (Lindberg et al., 1996; Fitzsimons et al., 1999; Beresford et al., 2001; Mama et al., 2002; Mannu et al., 2002; Callon et al., 2004; Ostlie et al., 2005; Dolci et al., 2008). They were identi?ed by SSCP as Lb. casei (RFLP clusters 1 and 5), Lb. plantarum (RFLP clusters 1, 2 and 6), S. thermophilus (RFLP cluster 3), E. faecium (RFLP clusters 4 and 7), and Lc. lactis (RFLP cluster 8), Based on the comparison of RFLP and SSCP results, it was observed that Graviera cheese isolates of Lb. casei, Lb. plantarum and E. faecium were characterized by more than one RFLP patterns for each species, suggesting an intra-species genetic diversity which was not observed for S. thermophilus isolates. 3.3. Polyphasic identi?cation of E. faecium isolates of phenotypic and genotypic diversity Identi?cation using biochemical methods was in high relevance and mutual agreement with SSCP identi?cation of the LAB isolates to the species (Table 1). In particular, the intra-species diversity of E. faecium isolates shown by RFLP analysis (clusters 4 and 7) was re?ected in their phenotypes also, i.e., as differences in the respective cluster ability to ferment sucrose, trehalose and melezitose (Table 2). E. faecium is well recognized as a phenotypically and genomically very heterogeneous species (Devriese et al., 1993; Vancanneyt et al., 2002). Therefore, routine phenotypic procedures for discriminating E. faecium from other enterococcal species commonly present in milk and cheese, such as Enterococcus faecalis and E. durans, require to be based on key sugar fermentation reactions, i.e., L-arabinose, mannitol, sorbitol, raf?nose, and sucrose (Knudtson and Hartman, 1992; Giannou et al., 2009b; Samelis et al., 2009a, 2010). Fermentation of L-arabinose remains the key sugar reaction differentiating E. faecium from most other enterococcal species, particularly E. faecalis (Knudtson and Hartman, 1992). Indeed, phenotypic identi?cation of L-arabinose-positive enterococcal strains as E. faecium was constantly validated by SSCP, whereas FTIR did not discriminate E. faecium from Enterococcus mundti, Enterococcus hirae and E. faecalis (Tables 1 and 2). 3.4. Polyphasic identi?cation of Lb. casei group of isolates prevailing in ripened Graviera cheeses Based on the FTIR cluster analysis, it is evident that the most prevalent LAB phenotype in Graviera cheeses were the Lb. casei isolates of phenotypic group I, which coincided with RFLP cluster 1 (Table 1). All isolates of this group were typically sorbitol-positive (Table 2) and were identi?ed as Lb. casei by SSCP (Table 1). Compared to Lb. casei group I, the isolates of Lb. casei phenotypic group II formed another separate RFLP cluster 5, and were of minor presence in all Graviera cheeses (Table 1; Samelis et al., 2010). In accordance, the representative isolates TF11 and TF23 of Lb. casei group II were unable to ferment sorbitol (Table 2), as the entire group II previously did (Samelis et al., 2010). In fact, TF12 was the only “intruder” isolate in RFLP cluster 5 that fermented sorbitol (Tables 1 and 2). Samelis et al. (2010) emphasized that the predominant sorbitol-positive group I of Lb. casei matched the description of Lb. paracasei subsp. paracasei ATCC 27092T, whereas the sorbitol-negative isolates of group II resembled with Lb. casei. As, however, the neotype strain of Lb. casei subsp. casei ATCC 334 and Lb. paracasei are perfectly homologous, it has yet to be decided as to whether the two species can be united and the name “paracasei” be rejected (Dicks et al., 1996; Dellaglio et al., 2002). Amiel et al. (2001) reported different FTIR clusters for strains of Lb. casei and Lb. paracasei related to the type strain ATCC 334. Thus, molecular ?ndings on the Lb. casei group were in part con?rmed by FTIR


? ? NT ? ?

? ? ?


? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? TF24 TF13 8 NA Arginine-negative Arginine-negative


? ? NT ? ?

? ? ?




? ? NT ? ?

? ? ?



? ? NT ? ?

? ? ?




? ? NT ? ?

? ? ?




? ? NT ? ?

? ? ?




? ? NT ? ?

? ? ?




? ? NT ? ?

? ? ?




Phenotypic group/type or atypical phenotypic feature

I II Contaminated culture Grows weakly at 45  C Homogenous group

Identical to plantarum main group

TF1, TF2, TF9, TF10, TF21, TF22 TF12 TF11, TF23 TF19 TF17 TF3, TF4, TF15, TF27, TF28, TF29 TF16

RFLP cluster

7 4 4


5 5 6 1 2


Lactococcus lactis/lactis Lactococcus lactis/garviae (by phenotypic method only) Streptococcus thermophilus/thermophilus

casei/casei-paracasei casei/casei casei/Lactobacillus sp. plantarum/plantarum plantarum/plantarum

Lactobacillus plantarum/plantarum

Enterococcus faecium/faecium Enterococcus faecium/faecium Enterococcus faecium/faecium

LAB species identi?cation by SSCP/phenotypic method

Lactobacillus casei/casei-paracasei

Lactobacillus Lactobacillus Lactobacillus Lactobacillus Lactobacillus


Homogenous group



TF5, TF6, TF8, TF14

Representative isolate/s

TF25, TF26 TF7 TF18









J. Samelis et al. / Food Microbiology 28 (2011) 76e83


which, in agreement with DNA-based studies (Dicks et al., 1996; Dellaglio et al., 2002; Vasquez et al., 2005), discriminated well between Lb. casei, Lb. zeae and Lb. rhamnosus (Amiel et al., 2001).

3.5. Identi?cation of non-LAB Graviera cheese isolates by FTIR and molecular methods As indicated, 2 out of 50 (4%) TSAYE isolates from GR-T and 15 out of 50 (30%) TSAYE isolates from GR-R cheese samples were nonLAB. The FTIR method grouped those non-LAB isolates in six clusters (Table 3) which were clearly apart from their counterpart LAB isolates (TSAYE clusters 7e15 in Table 1). Cluster 2 included isolates from both cheeses, which were identi?ed as Staphylococcus equorum by all methods. In fact, isolates of cluster 2 were the only nonLAB detected in GR-T cheese. RFLP analysis further showed that the representative isolate TF20 of cluster 2 was differentiated from the RFLP clusters of LAB in Table 1, whereas SSCP analysis of the 16S rRNA gene con?rmed identi?cation as Staph. equorum (Table 3). Clusters 1 and 3 included staphylococci also, which had been identi?ed as Staphylococcus saprophyticus and Staph. vitulinus, respectively, by FTIR; however, their rpoB-sequences were identical and suggested Staph. equorum for both clusters. Clusters 4, 5 and 6 were FTIR-identi?ed as Rhodococcus equi, Corynebacterium variabile, and Brevibacterium linens; however, 16S rRNA sequencing revealed Corynebacterium sp., C. variabile, and Brevibacterium sp., respectively (Table 3). Staphylococci were recently reported to constitute a considerable part (ca. 5e6 log cfu/g) of the micro?ora of Graviera cheese which was dominated by mesophilic LAB at levels above 8 log cfu/g on TSAYE, M-17 and MRS agar plates during ripening; in particular, the plant-ripened cheeses contained signi?cantly lower staphylococcal populations than their counterpart cheeses ripened in the pilot ripening room (Samelis et al., 2010). These ?ndings were con?rmed in the present study given that the relative (%) abundance of staphylococcal isolates on TSAYE was 4% and 18% for GR-T and GR-R cheese, respectively. Furthermore, this study indicated that Staph. equorum was the sole staphylococcal species isolated from both cheeses, and coryneform bacteria had a considerable presence (12%) on TSAYE plates of GR-R cheese (Table 3). Coryneforms are among the main bacteria grown on the surface of many cheeses (Beresford et al., 2001), and may be present in high numbers on non-edible rinds of hard cheeses like Graviera. Also, staphylococci have been reported as important members of the

surface micro?ora of Kefalograviera cheese (Siafaras et al., 2008), whereas they might be present at above 6-log levels close to rind of hard cheeses since they are microaerophilic. So, in this study, staphylococci and mainly corynebacteria were probably transferred to the core samples from the cheese surface upon insertion of the cork borers into the cheese mass during sampling. Both were more numerous on TSAYE plates of GR-R than GR-T cheeses probably due to the moister atmosphere in the pilot ripening room. 3.6. Complementary ef?cacy of polyphasic identi?cation methods Based on the results (Tables 1 and 3), FTIR spectroscopy seems likely to generate clusters within a species; consequently, FTIR alone is very unlikely to resolve dif?cult taxonomic debates, such as that of the Lb. casei/paracasei complex group. However, FTIR was very effective for a de-replication of isolates allowing a high throughput screening with subsequent clustering of 299 LAB and non-LAB isolates, and was in general agreement in regard to the identi?cation of 20 out of 29 (69%) representative isolates at the genus or species level. As indicated, some discrepancies for the LAB isolates from TSAYE were observed using FTIR (Table 1). These misidenti?cations are culture-medium dependent since those isolates were regarded as total mesophilic bacteria and were therefore grown on CASO agar; but the FTIR spectrum library for LAB necessitates APT agar. Such misidenti?cations can thus be recti?ed by simply subculturing CASO isolates on APT agar before measurements. In general aspects, the information provided by FTIR is useful and complementary to genomic information, and can be introduced in a polyphasic approach (Amiel et al., 2001), as it was successfully done in this study. Also, biochemical LAB identi?cation was reproducible and in mutual agreement with molecular (RFLP/SSCP) approaches. Generally, when carefully applied, phenotypic approaches are ideal to detect contaminated (e.g., TF19), or atypical (e.g., TF11, TF23, TF17, TF13 and TF24) LAB isolates (Table 2). 3.7. Comparison of microbial diversity of Graviera by polyphasic and classical approaches To facilitate the comparison between the FTIR-based polyphasic approach and the classical approach, the distribution (%) of the LAB species isolated from Graviera cheeses are summarized in Table 4, and in correlation with the isolation medium in Table 5. Overall, the

Table 3 FTIR-based clustering of 17 non-LAB isolatesa from traditional Greek Graviera cheese ripened under two ripening conditions,b and their identi?cation by FTIR and selected molecular methods. Isolation agar TSAYE/30  C FTIR cluster 1 2 GR-T isolates e 2 GR-R isolates 3 4 Total isolates 3 6 Presumptive FTIR cluster identi?cation Staphylococcus saprophyticus Staphylococcus equorum Presumptive RFLP identi?cation NT Staphylococcus sp. (Isolate TF20; RFLP cluster 9). NT NT NT NT 16S rRNA sequencing identi?cation NT Staphylococcus equorum rpoB-sequence identi?cation Staphylococcus equorum Staphylococcus equorum

3 4 5 6 Total non-LAB isolates

e e e e 2

2 1 4 1 15

2 1 4 1 17

Staphylococcus vitulinus/pulvereri Rhodococcus equi Corynebacterium variabile Brevibacterium linens

NT Corynebacterium sp. Corynebacterium variabile Brevibacterium sp.

Staphylococcus equorum NT NT NT

NT, not tested. a Non-LAB isolates were exclusively isolated from TSAYE agar plates representing a minor part of the total mesophilic cheese ?ora (50 TSAYE isolates/cheese; 100 TSAYE isolates in total). The remaining 83 isolates representing the dominant LAB ?ora on TSAYE (clusters 7e15) are listed in Table 1. b GR-T, Graviera-T cheese ripened in the commercial plant’s ripening room under fairly standardized environmental conditions; GR-R, Graviera-R cheese ripened in a pilot ripening room under constantly monitored environmental conditions.


J. Samelis et al. / Food Microbiology 28 (2011) 76e83

Table 4 Comparison of the LAB ?ora compositiona of traditional Greek Graviera cheese as determined by FTIR-based polyphasic and classical phenotypic approaches. Isolation/Identi?cation approach Polyphasic Graviera-T cheese Streptococcus thermophilus Lactococcus lactis Lactobacillus casei (group I) Lactobacillus casei (group II) Lactobacillus plantarum Enterococcus faecium Total LAB isolates

Classical Graviera-R cheese 19 (14.2) 2 (1.5) 83 (61.9) 8 (6.0) 21 (15.7) 1 (0.7) 134 Graviera-T cheese e e 11 (73.3) e 4 (26.7) e 15 Graviera-R cheese 2 (13.3) e 10 (66.7) 2 (13.3) 1 (6.7) e 15

6 (4.0) e 98 (66.2) 5 (3.4) 34 (23.0) 5 (3.4) 148

Numbers of isolates are shown without parentheses. The numbers in parenthesis show the percentage (%) distribution of each of the LAB species in the two cheeses as determined by the two approaches.

Table 5 Comparison of the LAB ?ora compositiona of traditional Greek Graviera cheese as determined by FTIR-based polyphasic and classical phenotypic approaches in correlation with the LAB species distribution on the different isolation agar media. Isolation agar medium TSAYE Polyphasic Streptococcus thermophilus Lactococcus lactis Lactobacillus casei (group I) Lactobacillus casei (group II) Lactobacillus plantarum Enterococcus faecium Total LAB isolates

M-17/22  C Classical e e 7 (70.0) 2 (20.0) 1 (10.0) e 10 Polyphasic 8 (8.0) 1 (1.0) 60 (60.0) 9 (9.0) 20 (20.0) 2 (2.0) 100 Classical e e 9 (90.0) e 1 (10.0) e 10

M-17/42  C Polyphasic 17 (17.2) e 62 (62.6) e 18 (18.2) 2 (2.0) 99 Classical 2 (20.0) e 5 (50.0) e 3 (30.0) e 10

Total isolates Polyphasic 25 (8.9) 2 (0.7) 181 (64.2) 13 (4.6) 55 (19.5) 6 (2.1) 282 Classical 2 (6.7) e 21 (70.0) 2 (6.7) 5 (16.6) e 30

e 1 (1.2) 59 (71.1) 4 (4.8) 17 (20.5) 2 (2.4) 83

The sums of isolates from both cheeses (GR-T and GR-R) are shown without parentheses. The numbers in parenthesis show the percentage (%) distribution of each of the LAB species on each of the isolation media as determined by the two identi?cation approaches (polyphasic or classical).

microbial (LAB) communities in the cheeses were similar; however, it was evident that Lb. plantarum were more numerous in GR-T as opposed to S. thermophilus which were more numerous in GR-R (Table 4). Since this difference was prominent, two S. thermophilus strains had also been detected in GR-R by the classical approach; however, no S. thermophilus strain had been detected in GR-T cheeses of batch A (Table 4). In general, the polyphasic approach provided a more comprehensive view on the microbial ecology and biodiversity in GR-T and GR-R cheeses since it detected within its fraction of isolates several LAB species that had been undetectable within a 10-fold lower fraction of isolates by the classical approach, e.g., E. faecium and Lc. lactis in both cheeses, and Lb. casei group II and S. thermophilus in GR-T (Table 4). It seems therefore that subdominant members of the cheese LAB ?ora which comprised ca. 10% to even less than 1% of the colonies on TSAYE and M-17 agar plates could be detected using the polyphasic approach (Table 4). In parallel, the relative distributions of the main LAB species in accordance with the different isolation media and incubation conditions were better optimized by the polyphasic than the classical approach (Table 5). Finally, another sub-dominant part of the cheese micro?ora which comprised Staph. equorum and corynebacteria present on TSAYE plates only, could be detected by the FTIR-based isolation procedure. This non-LAB ?ora, which had an overall higher presence in GR-R than GR-T cheeses (Table 3), had also been undetectable by the classical approach (Samelis et al., 2010).

predominance of mesophilic NSLAB, Lb. casei/paracasei and Lb. plantarum, over coccoid starter species S. thermophilus and Lc. lactis, and a remarkable presence of indigenous E. faecium was found, generally con?rming previous data on traditional European hard cheeses (Beresford et al., 2001; De Angelis et al., 2001; Callon et al., 2004; Dolci et al., 2008). Polyphasic identi?cation based on rapid FTIR screening of isolates improved detection of sub-dominant species while further gaining valuable information on the dominant LAB species. In conclusion, combining phenotypic and genotypic methods is very useful. However, bacterial identi?cation by FTIR, RFLP and SSCP methods requires a reference database. Nonrepresented isolates can be identi?ed by classical methods or based on 16S rRNA sequences (or rpoB for Staphylococcus) and added over time to build a custom database. Setting up such databases is a time-consuming prerequisite before these methods can be used routinely in the laboratory. Acknowledgements Funding for this study was provided by TRUEFOOD (Traditional United Europe Food), which is an integrated project ?nanced by the European Commission under the 6th RTD Framework (contract FOOD-CT-2006-016264). The information in this document re?ects only the authors’ views, and the European Community is not liable for any use that may be made of the information contained therein. References

4. Conclusions Comparative results of cheese isolates identi?cation showed that biochemical and molecular methods were in agreement on the bacterial (LAB) species occurring in ripened Graviera cheeses, while some culture-medium dependent discrepancies at the species or genus level were observed with the FTIR method. A high

Amiel, C., Mariey, L., Curk-Daubie, M.-C., Pichon, P., Travert, J., 2000. Potentiality of Fourier Transform Infrared spectroscopy (FTIR) for discrimination and identi?cation of dairy lactic acid bacteria. Lait 80, 445e459. Amiel, C., Mariey, L., Denis, C., Pichon, P., Travert, J., 2001. FTIR spectroscopy and taxonomic purpose: contribution to the classi?cation of lactic acid bacteria. Lait 81, 249e255. Anonymous, 2004. Cheeses of protected denomination of origin. In: Hellenic Code of Food and Beverages, Part A, vol. 2. National Chemistry Laboratory, Ministry of Finance, National Publishing Of?ce, Athens, Greece, p. 907.

J. Samelis et al. / Food Microbiology 28 (2011) 76e83 Beresford, T.P., Fitzsimons, N.A., Brennan, N.L., Cogan, T.M., 2001. Recent advances in cheese microbiology. Int. Dairy J. 11, 259e274. Bizzarro, R., Torri Tarelli, G., Giraffa, G., Neviani, E., 2000. Phenotypic and genotypic characterization of lactic acid bacteria isolated from Pecorino Toscano cheese. Ital. J. Food Sci. 12, 303e316. Bouton, Y., Guyot, P., Beuvier, E., Tailliez, P., Grappin, R., 2002. Use of PCR-based methods and PFGE for typing and monitoring homofermentative lactobacilli during Comte cheese ripening. Int. J. Food Microbiol. 76, 27e38. Callon, C., Millet, L., Montel, M.C., 2004. Diversity of lactic acid bacteria isolated from AOC Salers cheese. J. Dairy Res. 71, 231e244. Callon, C., Duthoit, F., Le Frileux, Y., De Crémoux, R., Montel, M.C., 2007. Stability of microbial communities of goat milk during a lactation year: molecular approaches. Syst. Appl. Microbiol. 30, 547e560. Chen, G., Kocaoglu-Vurma, N.A., Harper, W.J., Rodriguez-Saona, L.E., 2009. Application of infrared microspectroscopy and multivariate analysis for monitoring the effect of adjunct cultures during Swiss cheese ripening. J. Dairy Sci. 92, 3575e3584. De Angelis, M., Corsetti, A., Tosti, N., Rossi, J., Corbo, M.R., Gobbetti, M., 2001. Characterization of non-starter lactic acid bacteria from Italian ewe cheeses based on phenotypic, genotypic and cell wall protein analyses. Appl. Environ. Microbiol. 67, 2011e2020. Dellaglio, F., Felis, G.E., Torriani, S., 2002. The status of the species Lactobacillus casei (Orla-Jensen 1916) Hansen and Lessel 1971 and Lactobacillus paracasei Collins et al. 1989. Request for an opinion. Int. J. Syst. Evol. Microbiol. 52, 285e287. Devriese, L.A., Pot, B., Collins, M.D., 1993. Phenotypic identi?cation of the genus Enterococcus and differentiation of phylogenetically distinct enterococcal species and species groups. J. Appl. Bacteriol. 75, 399e408. Dicks, L.M.T., DuPlessis, E., Dellaglio, F., Lauer, E., 1996. Reclassi?cation of Lactobacillus casei subsp. casei ATCC 393 and Lactobacillus rhamnosus ATCC 15820 as Lactobacillus zeae nom. Rev., designation of ATCC 334 as the neotype of L. casei subsp. casei, and rejection of the name Lactobacillus paracasei. Int. J. Syst. Bacteriol. 46, 337e340. Dolci, P., Alessandria, V., Rantsiou, K., Rolle, L., Zeppa, G., Cocolin, L., 2008. Microbial dynamics of Castelmagno PDO, a traditional Italian cheese, with a focus on lactic acid bacteria ecology. Int. J. Food Microbiol. 122, 302e311. Drancourt, M., Raoult, D., 2002. rpoB gene sequence-based identi?cation of Staphylococcus species. J. Clin. Microbiol. 40, 1333e1338. Duthoit, F., Godon, J.J., Montel, M.C., 2003. Bacterial community dynamics during production of “Registered Designation of Origin” Salers cheese as evaluated by 16S rRNA gene SSCP analysis. Appl. Environ. Microbiol. 69, 3840e3848. European Commission, 2007. Commission Regulation (EC) No 1441/2007 of 5 December 2007 amending Regulation (EC) No. 2073/2005 on microbiological criteria for foodstuffs. Off. J. Eur. Union L322, 12e29. Fitzsimons, N.A., Cogan, T.M., Condon, S., Beresford, T., 1999. Phenotypic and genotypic characterization of non-starter lactic acid bacteria in mature Cheddar cheese. Appl. Environ. Microbiol. 65, 3418e3426. Georgieva, R.N., Iliev, I.N., Chipeva, V.A., Dimitonova, S.P., Samelis, J., Danova, S., 2008. Identi?cation and in vitro characterization of Lactobacillus plantarum strains from artisanal Bulgarian white brined cheeses. J. Basic Microbiol. 48, 234e244. Giannou, E., Kakouri, A., Matijai, B.B., Rogelj, I., Samelis, J., 2009a. Fate of Listeria sc monocytogenes on fully-ripened Greek Graviera cheese stored at 4, 12, or 25  C in air or vacuum packages: in situ PCR detection of a cocktail of bacteriocins potentially contributing to pathogen inhibition. J. Food Prot. 72, 531e538. Giannou, E., Lianou, A., Kakouri, A., Kallimanis, A., Drainas, C., Samelis, J., 2009b. Identi?cation and biopreservation potential of Enterococcus spp. isolated from fully ripened Graviera, a traditional hard Greek cheese. Ital. J. Food Sci. 21, 135e147. Giraffa, G., Neviani, E., 2001. DNA-based, culture-independent strategies for evaluating microbial communities in food-associated ecosystems. Int. J. Food Microbiol. 67, 19e34. Knudtson, L.M., Hartman, P.A., 1992. Routine procedures for isolation and identi?cation of enterococci and fecal streptococci. Appl. Environ. Microbiol. 58, 3027e3031. Kümmerle, M., Scherer, S., Seiler, H., 1998. Rapid and reliable identi?cation of foodborne yeasts by Fourier-transform infrared spectroscopy. Appl. Environ. Microbiol. 64, 2207e2214. Lamprell, H., Mazerolles, G., Kodjo, A., Chamba, J.F., Noel, Y., Beuvier, E., 2006. Discrimination of Staphylococcus aureus strains from different species of Staphylococcus using Fourier transform infrared (FTIR) spectroscopy. Int. J. Food Microbiol. 108, 125e129. Le?er, D., Lamprell, H., Mazerolles, G., 2000. Evolution of Lactococcus strains during ripening in Brie cheese by Fourier transform infrared spectroscopy. Lait 80, 247e254.


Lindberg, A.M., Christiansson, A., Rukke, E.O., Eklund, T., Molin, G., 1996. Bacterial ?ora of Norwegian and Swedish semi-hard cheese after ripening, with special reference to Lactobacillus. Neth. Milk Dairy J. 50, 563e572. Litopoulou-Tzanetaki, E., Tzanetakis, N., 2007. Characteristics of Greek traditional cheeses: from tradition to science and knowledge. In: Proceedings of International Symposium on Historical Cheeses of Countries Around the Archipelago Mediterraneo, Thessaloniki, Greece, 6 to 8 December 2007, pp. 97e121. Mama, V., Hatzikamari, M., Lombardi, A., Tzanetakis, N., Litopoulou-Tzanetaki, E., 2002. Lactobacillus paracasei subsp. paracasei heterogeneity: the diversity among strains isolated from traditional Greek cheeses. Ital. J. Food Sci. 14, 351e362. Mannu, L., Riu, G., Comunian, R., Fozzi, M.C., Scintu, M.F., 2002. A preliminary study of lactic acid bacteria in whey starter culture and industrial Pecorino Sardo ewes’ milk cheese: PCR-identi?cation and evolution during ripening. Int. Dairy J. 12, 17e26. Martin-Platero, A.M., Maqueda, M., Valdivia, E., Purswani, J., Martinez-Bueno, M., 2009. Polyphasic study of microbial communities of two Spanish farmhouse goats’ milk cheeses from Sierra de Aracena. Food Microbiol. 26, 294e304. Moschetti, G., Blaiotta, G., Villani, F., Coppola, S., Parente, E., 2001. Comparison of statistical methods for identi?cation of Streptococcus thermophilus, Enterococcus faecalis, and Enterococcus faecium from randomly ampli?ed polymorphic DNA patterns. Appl. Environ. Microbiol. 67, 2156. Oberreuter, H., Seiler, H., Scherer, S., 2002. Identi?cation of coryneform bacteria and related taxa by Fourier-transform infrared (FT-IR) spectroscopy. Int. J. Syst. Evol. Microbiol. 52, 91e100. Ostlie, H.M., Eliassen, L., Florvaag, A., Skeie, S., 2005. Phenotypic and PCR-based characterization of the micro?ora in Pr?st cheese during ripening. Int. Dairy J. 15, 911e920. Randazzo, C.L., Torriani, S., Akkermans, A.D.L., de Vos, W.M., Vaughan, E.E., 2002. Diversity, dynamics and activity of bacterial communities during production of an artisanal Sicilian cheese as evaluated by 16S rRNA analysis. Appl. Environ. Microbiol. 68, 1882e1892. Rebuffo-Scheer, C., Dietrich, J., Wenning, M., Scherer, S., 2008. Identi?cation of ?ve Listeria species based on infrared spectra (FTIR) using macrosamples is superior to a microsample approach. Anal. Bioanal. Chem. 390, 1629e1635. Samelis, J., Lianou, A., Kakouri, A., Delbes, C., Rogelj, I., Matijai, B.B., Montel, M.C., sc 2009a. Changes in the microbial composition of raw milk induced by thermization treatments applied prior to traditional Greek hard cheese processing. J. Food Prot. 72, 783e790. Samelis, J., Giannou, E., Lianou, A., 2009b. Assuring growth inhibition of listerial contamination during processing and storage of traditional Greek Graviera cheese: compliance with the new European Union regulatory criteria for Listeria monocytogenes. J. Food Prot. 72, 2264e2271. Samelis, J., Giannou, E., Matijaic, B.B., Rogelj, I., 2009c. Fate of Staphylococcus aureus s in traditional Greek Graviera cheese manufactured with or without Lactococcus lactis M104, a nisin-producing raw milk isolate. In: Proceedings (Electronic form) of the EFFoST 2009 Conference on New Challenges in Food Preservation, Budapest, Hungary, 11 to 13 November 2009, p. 298 (P-329). Samelis, J., Kakouri, A., Pappa, E.C., Matijaic, B.B., Georgalaki, M.D., Tsakalidou, E., s Rogelj, I., 2010. Microbial stability and safety of traditional Greek Graviera cheese: characterization of the lactic acid bacterial ?ora and culture-independent detection of bacteriocin genes in the ripened cheeses and their microbial consortia. J. Food Prot. 73, 1294e1303. Savic, D., Jokovic, N., Topisirovic, L., 2008. Multivariate statistical methods for discrimination of lactobacilli based on their FTIR spectra. Dairy Sci. Technol. 88, 273e290. Siafaras, G., Hatzikamari, M., Litopoulou-Tzanetaki, E., Tzanetakis, N., 2008. Antibacterial activities of the surface micro?ora of Kefalograviera cheese. Food Control 19, 898e905. Vancanneyt, M., Lombardi, A., Andrighetto, C., Knijff, E., Torriani, S., Bjorkroth, K.J., Franz, C.M.A.P., Foulquie Moreno, M.R., Revets, H., De Vuyst, L., Swings, J., Kersters, K., Dellaglio, F., Holzapfel, W.H., 2002. Intraspecies genomic groups in Enterococcus faecium and their correlation with origin and pathogenicity. Appl. Environ. Microbiol. 68, 1381. Vandamme, P., Pot, B., Gillis, M., DeVos, P., Kersters, K., Swings, J., 1996. Polyphasic taxonomy, a consensus approach to bacterial systematics. Microbiol. Rev. 60, 407e438. Vasquez, A., Molin, G., Pettersson, B., Antonsson, M., Ahrne, S., 2005. DNA-based classi?cation and sequence heterogeneities in the 16S rRNA genes of Lactobacillus casei/paracasei and related species. Syst. Appl. Microbiol. 28, 430e441. Weinrichter, B., Luginbuhl, W., Rohm, H., Jimeno, J., 2001. Differentiation of facultatively heterofermentative lactobacilli from plants, milk and hard type cheeses by SDS-PAGE, RAPD, FTIR, energy source utilization and autolysis type. Lebensm. Wiss. Technol. 34, 556e566.