Logo of aemPermissionsJournals.ASM.orgJournalAEM ArticleJournal InfoAuthorsReviewers
Appl Environ Microbiol. Mar 2007; 73(6): 1882–1891.
Published online Jan 26, 2007. doi:  10.1128/AEM.01716-06
PMCID: PMC1828836

Monitoring Bacterial Communities in Raw Milk and Cheese by Culture-Dependent and -Independent 16S rRNA Gene-Based Analyses[down-pointing small open triangle]


The diversity and dynamics of bacterial populations in Saint-Nectaire, a raw-milk, semihard cheese, were investigated using a dual culture-dependent and direct molecular approach combining single-strand conformation polymorphism (SSCP) fingerprinting and sequencing of 16S rRNA genes. The dominant clones, among 125 16S rRNA genes isolated from milk, belonged to members of the Firmicutes (58% of the total clones) affiliated mainly with the orders Clostridiales and the Lactobacillales, followed by the phyla Proteobacteria (21.6%), Actinobacteria (16.8%), and Bacteroidetes (4%). Sequencing the 16S rRNA genes of 126 milk isolates collected from four culture media revealed the presence of 36 different species showing a wider diversity in the Gammaproteobacteria phylum and Staphylococcus genus than that found among clones. In cheese, a total of 21 species were obtained from 170 isolates, with dominant species belonging to the Lactobacillales and subdominant species affiliated with the Actinobacteria, Bacteroidetes (Chryseobacterium sp.), or Gammaproteobacteria (Stenotrophomonas sp.). Fingerprinting DNA isolated from milk by SSCP analysis yielded complex patterns, whereas analyzing DNA isolated from cheese resulted in patterns composed of a single peak which corresponded to that of lactic acid bacteria. SSCP fingerprinting of mixtures of all colonies harvested from plate count agar supplemented with crystal violet and vancomycin showed good potential for monitoring the subdominant Proteobacteria and Bacteroidetes (Flavobacteria) organisms in milk and cheese. Likewise, analyzing culturable subcommunities from cheese-ripening bacterial medium permitted assessment of the diversity of halotolerant Actinobacteria and Staphylococcus organisms. Direct and culture-dependent approaches produced complementary information, thus generating a more accurate view of milk and cheese microbial ecology.

Microbial diversity in raw milk is considered essential to the sensory richness and variety of traditional cheeses (3, 11, 13). However, some members of these complex communities may also be responsible for cheese flavor defects (7) or may constitute a health risk (10). Being able to characterize the microbial communities in raw milk and follow the dynamics of the entire populations throughout the cheese-making and ripening processes is therefore critical. The bacterial community in raw milk and cheese has been determined by cultivation followed by molecular identification of isolates (1, 2, 6, 28, 31). Culture-independent studies based on molecular fingerprinting techniques such as denaturing gradient gel electrophoresis (DGGE), temporal temperature gradient electrophoresis (TTGE), and single-strand conformation polymorphism (SSCP) have been carried out to evaluate bacterial diversity in raw milk (23), the cheese core (9, 14, 15, 34, 39), or the cheese surface (17, 38). The cheese-making process aims at favoring the growth and activity of lactic acid bacteria (LAB). The addition of selected LAB starter cultures ensures safety through rapid acidification. As expected, LAB were subsequently found to be predominant on bacterial fingerprints from the cheese core. Streptococcus thermophilus and thermophilic lactobacilli were dominant in the DGGE patterns of Ragusano cheese (39), and Streptococcus bovis and Lactococcus lactis were dominant in the patterns of Pecorino Siciliano cheese (40). Ogier et al. (33), using direct TTGE and DGGE analyses, also found that the microbial community in Saint-Nectaire cheese cores was dominated by S. thermophilus and L. lactis. Indeed, the fingerprints produced by such techniques as SSCP, DGGE, and TTGE using general bacterial primers reveal only the most dominant populations of a bacterial community. In the presence of one or several largely dominant populations, these methods do not permit investigation of the diversity of less dominant populations (37). In food fermentation microbial communities, which are generally believed to harbor a large fraction of culturable species (30), this limitation may be partly overcome by applying fingerprinting techniques to subcommunities capable of growing on different culture media. This approach was previously used to explore the diversity of culturable bacteria in cheese (16, 35, 40) and other fermented foods, such as capers (36), but was focused mainly on LAB. The aim of the present study was to evaluate the relevance of direct versus culture-based approaches combining SSCP fingerprinting and 16S rRNA gene sequencing to monitor the diversity and dynamics of bacterial populations in Saint-Nectaire, a semihard cheese made from raw milk. To identify the dominant bacterial populations of raw milk, the diversity of 16S rRNA genes in a clone library constructed from DNA isolated from milk and in bacterial colonies isolated from various culture media was analyzed. In an attempt to reveal the fate of subdominant populations in cheese, the diversity of bacterial cell suspensions harvested from culture media allowing the growth of non-LAB bacteria was investigated by SSCP fingerprinting.


Milk and cheese samples.

Farm productions of Saint-Nectaire, a noncooked, registered designation of origin, semihard cheese variety made from raw cow's milk and produced in the medium-altitude mountainous area of Massif Central, France, were studied. Nine cheese-making batches from three farms designated F1, F2, and F3, which were sampled three times at 3-month intervals (in winter, spring, and summer), were analyzed. Milk protein content was approximately 31 g/liter, and fat content ranged between 27.6 and 39.5 g/liter. The raw milk was inoculated with a commercial starter culture (Streptococcus thermophilus, Lactococcus lactis) immediately after the milking. The milk was coagulated by adding calf rennet (27.5 to 30 ml/100 liters). After coagulation, the curd was cut into small pieces (around 0.5 cm) and gently stirred, and then it was gathered, placed into molds, and pressed under 3-bar pressure for 10 min to remove whey. Thirty to 40 grams of salt per kg was placed on the surface of each 1.7-kg cheese, and pressing was completed at 23 to 26°C for 6 h. The cheeses were ripened at 10 to 12°C for 28 days. Samples were taken aseptically from the raw milk prior to the addition of the starter culture and from the cheese on day 1 and also after 28 days. At each sampling point, three cheeses from the same production batch were sampled using a sterile cheese trier, the rind was discarded (thickness, 5 mm), and the three core subsamples (around 10 g each) were mixed together by grinding. All samples were stored at −20°C for further bacteriological and molecular analyses.

Microbiological analyses of milk and cheese samples and collection of isolates and cell mixtures.

Cheese samples were emulsified in sterile phosphate buffer (20 mM KH2PO4-0.01 M K2HPO4, pH 7.5) and blended in a Stomacher lab blender (Seward Medical, London, United Kingdom) for 4 min. Both the milk and the cheese suspensions were diluted in Ringer's solution, and appropriate dilutions were spread in triplicate on M17 agar (42) and incubated at 42°C for 48 h on the cheese-ripening bacterial medium (CRBM) described by Denis et al. (12), which contains 50 g/liter NaCl, 9 mg/liter natamycin, and 40 mg/liter nalidixic acid. Samples were also incubated at 25°C for 5 days on plate count agar supplemented with 1% milk (PCAM) (32) and on plate count agar supplemented with 1% milk, 5 mg/liter vancomycin, and 5 mg/liter crystal violet as inhibitors of gram-positive bacteria (PCAI). Both PCA media were incubated at 30°C for 48 h. All media were purchased from Biokar Diagnostics (Pantin, France). Within 24 h after cells were counted, duplicate plates containing each medium and 30 to 100 colonies were selected. One milliliter of 4 M guanidine thiocyanate-0.1 M Tris (pH 7.5)-1% N-lauroylsarcosine solution was spread on the first plate, all the colonies present on the surface of the agar were suspended with a rake, and the resulting cell suspension was collected and stored at −20°C for further molecular analyses. For F2 milk and cheese, individual colonies were also picked from the second plate. One specimen of each colony morphotype was picked, and additional isolates were picked randomly among nonconfluent colonies up to a total of 20 to 50 isolates per plate. Colonies were restreaked for purification prior to direct PCR amplification. All isolates were further identified by 16S rRNA gene sequencing.

Extraction of total bacterial DNA from milk and cheese and plate count cell suspensions.

One-milliliter milk samples or 1-g cheese samples were first mixed with 1 ml of 4 M guanidine thiocyanate-0.1 M Tris (pH 7.5) and 125 μl of 10% N-lauroylsarcosine. The cheese suspension was homogenized by grinding it with a stainless steel bead (1-cm diameter) in a stainless steel container (Kurt Retsch GmbH, Haan, Germany), using a reciprocating shaker (MM200, Kurt Retsch GmbH) for 2 min at 30 Hz. DNA extraction from cell suspensions harvested from agar plates started directly at this stage. A 2-ml tube containing 250 μl of the suspension was filled with 200 mg of zirconium beads (0.1-mm diameter), 100 μl of 20% sodium dodecyl sulfate solution, 400 μl of 0.1 M phosphate buffer (pH 8.0), 400 μl of 50 mM sodium acetate containing 10 mM EDTA (pH 6.0), and 400 μl of phenol-chloroform-isoamyl alcohol (25:24:1) (pH 8.0). The tube was heated at 80°C for 2 min in a water bath, chilled on ice for 2 min, and shaken for 2 min at room temperature in the reciprocating shaker. Both the shaking and heating steps were performed a second time. Total bacterial DNA was then purified by a phenol-based method as previously described (14).

Bacterium-targeted SSCP analysis-PCR.

DNA isolated from milk or cheese or plate count cell suspension was subjected to SSCP analysis-PCR. The target DNA amplified was variable region V3 of the 16S rRNA gene, which corresponds to a 200-bp fragment (E. coli 16S rRNA gene positions 330 to 533) (5). The primers, corresponding to conserved sequences bordering the variable regions, were w49 (5′-ACGGTCCAGACTCCTACGGG-3′) and w34 (5′-TTACCGCGGCGTGCTGGCAC-3′). Primer w34 was labeled with the 5′-fluorescein phosphoramidite NED (Applied Biosystems). All primers were synthesized by Applied Biosystems. The amplification reaction mixture contained 1 μl of DNA, 1× PCR buffer, a 200 μM concentration of each deoxynucleoside triphosphate, 2 mmol MgCl2 liter−1, 0.5 μmol of each primer liter−1, and 1.25 U Pfu Ultra DNA polymerase (Stratagene, La Jolla, CA). The PCR amplification was performed under the following conditions: 3 min of initial denaturation at 96°C; 25 cycles of denaturation (30 s at 96°C), annealing (30 s at 61°C), and extension (30 s at 72°C); and a final extension at 72°C for 10 min. Amplification products were analyzed on 1.5% (wt/vol) agarose gels.

SSCP electrophoresis and fingerprint data analysis.

The SSCP-PCR products were analyzed on an ABI Prism 310 genetic analyzer (Applied Biosystems, Foster City, CA) as previously described (14). The nondenaturing polymer matrix used was 5.6% GeneScan polymer (Applied Biosystems)-10% glycerol-1× buffer with EDTA (Applied Biosystems). The buffer was 1× buffer with EDTA (Applied Biosystems)-10% glycerol. The fluorescence signal was analyzed using the GeneScan analysis software (Applied Biosystems). The patterns to be compared were aligned using the internal standard, Genescan 400-ROX (Applied Biosystems). To analyze the different profiles, relative peak areas (Pi) were calculated as aiai, where ai is the area of peak i and Σai is the sum of the peak areas from the whole SSCP pattern. Only peaks with an ai of >700 and Pi of >1% of the total profile area were taken into account to estimate pattern richness (number of peaks) and for subsequent multivariate statistical analyses. Tests to evaluate the reproducibility of patterns had been conducted previously (14).

Milk DNA clone library, sequencing of 16S rRNA genes of clones, individual isolates obtained from culture media, and identification of SSCP peaks.

The dominant bacterial populations of raw milk were identified both by a culture-independent approach based on the construction of a 16S rRNA gene clone library from DNA isolated from milk and plasmid insert sequencing and by a culture-dependent approach based on sequencing of the 16S rRNA genes of individual isolates recovered from milk on the different culture media. To generate the clone library, the complete 16S rRNA gene was amplified from total DNA directly obtained from F2 raw milk as previously described (14). The PCR product was ligated into the pCR4Blunt-TOPO vector and transformed into Escherichia coli TOP10 OneShot cells as specified by the manufacturer (Invitrogen, Carlsbad, CA). Plasmid inserts were amplified by PCR with the universal plasmid primers T3 (5′-ATTAACCCTCACTAAAGGGA-3′) and T7 (5′-TAATACGACTCACTATAGGG-3′) as specified by the manufacturer (Invitrogen). For both amplified inserts and individual isolates, the 450 bases of the 5′ end of the 16S rRNA gene, comprising the V3 region, were sequenced with the BigDye Terminator cycle sequencing ready reaction kit with AmpliTaq DNA polymerase FS (Applied Biosystems) and an ABI Prism 310 genetic analyzer (Applied Biosystems). Each sequence obtained was identified by comparison with the closest sequences available in databases (GenBank and RDP) (25).

To identify 16S rRNA genes liable to correspond to peaks on complex SSCP patterns directly obtained from milk or cheese DNA, amplified plasmid inserts were subjected to a second nested SSCP analysis-PCR. Peaks of plasmid inserts were then compared with patterns obtained from milk or cheese DNA. Similarly, to identify peaks in SSCP patterns of the bacterial cell mixtures harvested from culture plates, SSCP patterns of individual isolates obtained from each culture medium were compared with patterns obtained from cell mixtures collected on the same medium.

Statistical analysis.

The differences in microbial count (numbers of CFU/ml, log transformed) and in richness of SSCP pattern between each farm were tested using analysis of variance and a subsequent Fisher's least-square difference test. Principal-component analyses were performed from the relative peak area data set. All statistical analyses were performed with Statistica software (version 6; Statsoft, Tulsa, OK).


Monitoring bacterial communities from milk and cheese by direct molecular methods.

To identify the dominant bacterial populations of raw milk, a clone library was constructed from the complete 16S rRNA genes amplified from DNA isolated from the milk collected from farm F2 during the winter period. Among the 125 clones analyzed, we identified 61 operational and taxonomic units (OTUs), as defined by 97% sequence similarity or greater. Table Table11 shows the closest relative species names assigned to each clone by comparison with sequences from the GenBank database. Members of the Firmicutes (low-G+C-content gram-positive bacteria) were distributed over 37 OTUs (57.6% of the total clones). Among them, Clostridiales constituted the most frequent bacterial order, representing 24% of the total clones. Six of the Clostridiales clones were affiliated with Clostridium lituseburense (97% sequence similarity), and two other clones were identified as Clostridium glycolicum, but the other 22 could not be assigned to a species. Other dominant members of the Firmicutes belonged to the Lactobacillales (16%) and Bacillales (7.2%) orders, with the most frequent OTUs being affiliated with Turicibacter sanguinis, Lactococcus lactis, Streptococcus dysgalactiae, Lactobacillus casei, and Facklamia sp. Members of the Proteobacteria were distributed over nine OTUs (21.6% of the clones), with dominant sequences being affiliated with Ralstonia pickettii. Members of the Actinobacteria (high-G+C-content gram-positive bacteria) were distributed over 11 OTUs (16.8% of the clones), with Arthrobacter arilaitensis and Corynebacterium confusum being the dominant species. Finally, members of the Bacteroidetes were represented by four OTUs (4% of the clones).

Phylogenetic affiliations of 125 cloned sequences obtained after cloning 16S rRNA genes isolated from DNA in raw milk from farm F2

The diversity and dynamics of bacterial communities from milk and cheese in the three farms F1, F2, and F3 were investigated through SSCP analysis based on direct DNA extraction followed by PCR amplification of the V3 region of 16S rRNA genes. rRNA genes liable to correspond to peaks in SSCP patterns were identified by comparison with SSCP peaks of clones. Some peaks probably resulted from the coelution of two or more different sequences, as can be seen from the fact that different species were assigned to the same peak numbers (Table (Table11).

Total numbers of peaks in SSCP patterns from milks and cheeses are summarized in Table Table2.2. The data presented are the averages of results for three batches sampled at 3-month intervals. Milks from the three farms produced complex SSCP patterns with up to 21 peaks. The total number of peaks was significantly higher in farm 2 milk than in farm 1 and farm 3 milk. Principal-component analysis of peak ratios (Fig. (Fig.1)1) showed that F2 milks from the three sampling periods clustered together and differed from F1 and F3 milks by their higher ratios for peaks corresponding to Clostridiales (SSCP peaks 1 to 3, Table Table1),1), C. confusum and Nocardioides dubius (peak 7), Arthrobacter psychrolactophilus (peak 10), uncultured rumen bacterium (peak 10), and Enterobacter agglomerans (peak 10) and for two peaks to which none of the clone inserts comigrated (peaks 9 and 12). In contrast, SSCP analysis of cheese samples on days 1 and 28 produced identical patterns composed of a single peak corresponding to L. lactis and S. thermophilus sequences (peak 19).

FIG. 1.
Principal-component analysis of peak ratios from SSCP patterns obtained after direct amplification of milk and cheese DNA. (A) Plot of peaks composing the patterns. Labels refer to SSCP peak numbers indicated in Table Table1.1. (B) Distribution ...
Numbers of peaks of SSCP patterns and microbial counts obtained after direct amplification and after cultivation on four culture media of milk or cheese DNA from farms F1, F2, and F3a

Monitoring bacterial communities from milk and cheese by culture-dependent molecular analyses.

The diversity of the dominant culturable bacteria from F2 milk and cheese on M17 medium, PCAM, CRBM, and PCAI was investigated. Table Table33 sums up the numbers of individual colonies analyzed and the dilutions from which they were recovered on each medium and shows the phylogenetic affiliations of the 126 isolates obtained from F2 milk and the 170 isolates obtained from F2 cheese on day 28. The 24 isolates retrieved from milk on M17 agar were composed of members of the Firmicutes phylum belonging to the Lactobacillales and Bacillales classes and of two members of the Actinobacteria class (Streptomyces, Brachybacterium). The 38 isolates retrieved from milk on PCAM plates were composed of members of the Actinobacteria (Microbacterium oxydans), Firmicutes (L. lactis), Alphaproteobacteria (Sphingomonas sp.), Gammaproteobacteria (Stenotrophomonas maltophilia, Luteibactor rhizovicinus, Psychrobacter faecalis, Moraxella osloensis), and Flavobacteriaceae (Chryseobacterium sp., Flavobacterium sp.). When isolates retrieved from cheese on M17 medium and PCAM were analyzed, only lactic acid bacteria (S. thermophilus, L. lactis, and L. casei) were found. In contrast, with PCAI, a large diversity was observed among milk isolates (44 isolates) as well as in cheese isolates (48 isolates). It was composed of members of the Flavobacteriaceae (Chryseobacterium sp., Flavobacterium sp.), members of the Firmicutes belonging to the Lactobacillales (Enterococcus faecalis, L. lactis, S. dysgalactiae, and Streptococcus parauberis), Alphaproteobacteria (clone NOS7.108WL), and Gammaproteobacteria (Klebsiella oxytoca, Klebsiella terrigena, Klebsiella trevisanii, S. maltophilia, L. rhizovicinus, P. faecalis, M. osloensis, “Pseudomonas reactans,” Enterobacter aerogenes), whose distributions were different in the milk and cheese. Finally, isolates retrieved from CRBM plates (20 from milk and 38 from cheese) belonged mainly to Staphylococcus (S. fleurettii, S. saprophyticus, S. vitulinus, S. haemolyticus, S. epidermidis, S. equorum) and Actinobacteria (Rothia sp., Brevibacterium linens, Brachybacterium sp., Kocuria rhizophila, Kocuria carniphila, A. arilaitensis, Corynebacterium flavescens) and were also differently distributed between milk and cheese. E. faecalis, Marinilactibacillus psychrotolerans, Bacillus pumilus, and Aerococcus viridans were also isolated on CRBM. In total, considering all culture media, 36 different OTUs were obtained from F2 milk and 21 OTUs from F2 cheese.

Phylogenetic affiliations of isolates recovered from milk and cheese from farm F2 on four culture media

The diversity of the culturable communities in milks and cheeses from the three farms was investigated by SSCP-PCR analysis of DNA isolated from bacterial cell suspensions collected from agar plates. The numbers of peaks in SSCP patterns of the culturable communities from milks and cheeses and the corresponding microbial counts are summarized in Table Table2.2. The data presented are averages for three batches sampled at 3-month intervals. Milk samples produced complex SSCP patterns comprising from 4 to 20 peaks depending on the culture medium and the farm. Levels of the culturable populations from milk on the different media were relatively similar, with at most a 1-log difference between counts on the different media. For all cheeses, an overall 2- to 5-log difference was observed between counts on PCAM or M17 medium and counts on CRBM or PCAI. Counts on M17 medium and PCAM increased markedly on day 1, reaching values around 8 log CFU/g, while SSCP analysis of cell mixtures collected from M17 medium and PCAM resulted in single-peak patterns corresponding to L. lactis and S. thermophilus sequences. In contrast, cell suspensions harvested from CRBM and PCAI plates produced multiple-peak patterns. Counts on CRBM gradually increased between days 0 and 28, reaching values around 5 log. At the same time, the SSCP pattern richness of communities culturable on CRBM did not change significantly from day 0 to day 28. Counts on PCAI slightly increased between days 0 and 1 and reached around 4.2 log at day 28, while the PCAI pattern richness significantly decreased from day 1 (Table (Table22).

Sequences corresponding to peaks in SSCP patterns of the cell mixtures harvested from each culture medium were identified by comparison with SSCP patterns of individual isolates from F2 milk and cheese collected from the same medium. Again, some peaks likely resulted from the coelution of two or more different sequences (Table (Table33).

Results of principal-component analysis of the peak ratios of PCAI patterns are shown in Fig. Fig.2.2. Axis 1 and axis 2 explained only 41.35% of the variance but showed milk patterns of the three farms clustered apart from cheese patterns. PCAI-culturable communities from milk were characterized by higher peak ratios for peaks corresponding to M. osloensis and Staphylococcus pasteuri (SSCP peak no. p12, Table Table3),3), S. maltophilia (p13), “P. reactans” (p14), and S. dysgalactiae (p15). Cheese PCAI patterns were distributed according to their ratios for peaks corresponding to Flavobacterium sp. (p3 and p8), Chryseobacterium sp. (p7 and p10), K. terrigena (p9), K. trevisanii (p11), E. faecalis (p11), L. rhizovicinus (p11), and L. lactis (p16), depending on time and farm.

FIG. 2.
Principal-component analysis of peak ratios from SSCP patterns of milk and cheese subcommunities culturable on PCAI. (A) Plot of peaks composing the patterns. Labels refer to SSCP peak numbers indicated in Table Table3.3. (B) Distribution of milk ...

Principal-component analysis of the CRBM pattern peak ratio (Fig. (Fig.3)3) showed that milk patterns were characterized by higher ratios for peaks 8 and 13, corresponding to four Staphylococcus species (S. fleurettii, S. saprophyticus, S. vitulinus, and S. epidermidis). Patterns of 1- and 28-day cheeses were distributed according to their ratios for peaks corresponding to various members of the Actinobacteria (SSCP peak no. c2′, c3′, and c5′, Table Table3),3), E. faecalis (c11), and M. psychrotolerans (c16).

FIG. 3.
Principal-component analysis of peak ratios from SSCP patterns of milk and cheese subcommunities culturable on CRBM. (A) Plot of peaks composing the patterns. Labels refer to SSCP peak numbers indicated in Table Table3.3. (B) Distribution of milk ...

Comparison of direct and culture-dependent methods.

The distribution of the different OTUs recovered from the clone library obtained from direct isolation of DNA from the F2 milk and from the individual colonies obtained from the various culture media according to phylogenetic groups is shown on Fig. Fig.4.4. The diversity of the Gammaproteobacteria phylum and Staphylococcus genus was greater in the isolates than in the clone library, whereas the diversity of the Clostridiales order and other unclassified gram-positive bacteria and the Betaproteobacteria was greater in the clone library. Culturable species such as C. lituseburense or R. pickettii frequently found among clones were not isolated. Only nine species comprising LAB species (E. faecalis, L. lactis, L. casei) as well as S. dysgalactiae, A. arilaitensis, K. rhizophila, S. equorum, Enterobacter cloacae, and Chryseobacterium sp. were recovered by both direct and culture-dependent methods. Other species with population levels similar to those of these species on culture media, especially actinobacteria and staphylococci, were not detected among clones.

FIG. 4.
Distribution of OTUs recovered from a clone library obtained from milk DNA and individual isolates retrieved from milk on culture media, according to phylogenetic groups. OTUs were defined by 97% sequence similarity or greater. Black bars, OTUs found ...

On the basis of peak counts (Table (Table2),2), SSCP patterns obtained after direct amplification of milk DNA showed the greatest diversity, followed by patterns of milk communities culturable on PCAM, CRBM, M17 medium, and PCAI. As regards cheeses, only the culture-dependent approach permitted investigation of the diversity of some subdominant populations in addition to dominant LAB through the analysis of subcommunities culturable on PCAI medium and CRBM.


In an attempt to monitor the diversity of non-LAB in the cheese core by a culture-dependent fingerprinting technique, we adopted a strategy that takes advantage of some culture media which allow the growth of broad microbial groups.

The cultural approach is dependent on the ability of bacteria to grow on the media under the culturing conditions applied and on the number of isolates and selection methodology. In addition, in our study, the viability of bacteria may also have been affected by the storage of milk and cheese samples at −20°C. The advantage of SSCP analysis of cells harvested from culture plates is that it generates an image of the diversity of all the colonies growing on the plate. On the other hand, it is subject to bias due to differences in growth rates between the different species on the culture medium. The extent of the diversity that can be revealed on a given medium may be affected by the presence of a dominant population growing on that medium (15, 16, 30). It may also vary with the stage of the cheese-making process. Indeed, while there was no predominant group of bacteria in the raw milk, marked differences in population levels were observed in cheese, thus precluding detection of subdominant populations. Discrepancies between results of direct and culture-dependent inventories of raw-milk bacterial populations may also result from a number of biases, like differences in the levels of efficiency of cell lysis during DNA isolation from the food matrix, preferential PCR amplification prior to cloning, or interspecies 16S rRNA operon copy number heterogeneity (8). Hence, the prevalence of the Clostridiales among clones may partly result from the high ribosomal operon copy number encountered in species from this order (21), while the aerobic conditions used in culturing were not suitable for their isolation.

Caution should be taken when identifying bacterial populations using SSCP peak analyses due to possible coelution of different sequences. Coelutions were observed within a genus (Staphylococcus) and across bacterial phyla (Klebsiella trevisanii/Enterococcus faecalis). This has been noticed in previous studies with SSCP (17) and with different fingerprinting techniques (TTGE and DGGE [23, 34]). Therefore, it can be difficult to identify species reliably in complex and relatively balanced bacterial communities, such as those of raw milk, on the sole basis of a pattern database. This limitation can be partly overcome by using different sets of primers targeting various variable regions (V2, V3) or specific groups of bacteria (high-G+C-percent gram-positive bacteria) (14) or by focusing on culturable subcommunities, thus reducing the overall diversity and at the same time providing isolates that can be identified.

The appropriateness of four culture media for monitoring populations besides LAB in raw cow's milks and cheeses from three farms was assessed. In this context, on the basis of 71 isolates, M17 agar incubated at 42°C offered insight into the diversity of culturable aerobic gram-positive bacteria (mainly with a low G+C percentage) when starter LAB were not highly predominant, as in milk. However, the growth of many additional species, among them, gram-negative bacteria, may be observed on this general-purpose medium if they are incubated at a lower temperature. From the analysis of 58 isolates, CRBM appeared to be useful for assessing the diversity of halotolerant actinobacteria and staphylococci in milk and cheese. However, E. faecalis can grow on CRBM, as observed by Denis et al. (12), and this medium also revealed the presence of Marinilactibacillus psychrotolerans in cheese. Although the addition of 5 mg/liter crystal violet (20) and 5 mg/liter vancomycin in PCAI was not sufficient to completely inhibit the growth of gram-positive bacteria, results from the analysis of 92 isolates suggest that this medium has good potential for investigating the diversity of the subdominant members of the Alpha- and Gammaproteobacteria and Flavobacteriaceae in milk and cheese.

Our results clearly confirmed the necessity to combine culture-dependent and -independent methods to generate a more accurate view of the microbial ecology of foodstuffs, as suggested by Ercolini et al. (16), Feurer et al. (17), and Miambi et al. (30). Direct and culture-dependent analyses showed that milk bacterial communities were highly diversified. Milk from farm 2 was chosen for a detailed analysis of 16S rRNA gene sequence diversity on the basis of its higher peak count in SSCP patterns, which suggested that it may harbor the greatest diversity. This analysis was performed only during the winter period. However, as can be seen from the principal-component analysis of SSCP patterns of DNA isolated from the milk, the dominant bacterial communities of raw milk from the three farms seemed to be rather stable whatever the sampling period. Hence, although the balance between bacterial populations likely varied with the seasons, the diversity of the dominant populations does not seem to have changed markedly.

Chryseobacterium and Flavobacterium spp. occur frequently in dairy products (20). New species were recently isolated from raw milk (Chryseobacterium joostei [18]) and from a lactic acid beverage (C. shigense [41]). Clone library analysis highlighted the occurrence of the Clostridiales. They were not affiliated with Clostridium species commonly associated with silage and defective cheese (24) but rather with C. lituseburense. This species was previously found to be predominant in cow manure and dairy wastewaters, along with Turicibacter sanguinis (29). In addition, Ralstonia sp. and Dietzia sp. have recently been found in oil field soil with high salinity (22). The culture-dependent approach highlighted the presence of subdominant populations in cheese core, such as coryneform bacteria, Marinilactibacillus organisms, and Chryseobacterium organisms. Coryneform bacteria are commonly isolated from the surfaces of smear-ripened cheeses (4, 31). Presumptive Brevibacterium, Arthrobacter, and Micrococcus strains were isolated from the rind of Saint-Nectaire-type cheese after 20 days of ripening (27). Marinilactibacillus psychrotolerans originally isolated from the marine environment (19) was recently found in French and German red-smear soft cheese surface communities (17, 26). Our results support those of Duthoit et al. (14), who found coryneform bacteria in the core of Salers cheese, another variety of uncooked, semihard cheese produced in Massif Central. However, although the rind was discarded, the possibility of contamination of the core sample by the cheese surface when sampling is carried out cannot be excluded.

Despite possible bias, SSCP fingerprinting of culturable communities on CRBM and PCAI was useful in overcoming the problem of LAB dominance in cheese and in a follow-up of the fate of subdominant microbial groups (Actinobacteria, Staphylococcus, Alpha- and Gammaproteobacteria, Flavobacteriaceae) from milk to cheese. As in most cheeses (14, 39), dramatic shifts in the composition of the microbial community were observed. The microbial species added as starters (S. thermophilus, L. lactis) soon out-numbered the other species arising from milk. However, considerable variation in the dynamics of subdominant populations between cheeses was observed.

In conclusion, both direct and culture-dependent approaches are sources of bias but produce complementary information on milk microbial ecology. The proposed method of investigating the diversity of subdominant non-LAB in cheese by fingerprinting of culturable communities is an interesting alternative to direct fingerprinting with general bacterial primers, by which these populations would be overlooked. They may be detected by specific PCR amplification, but this would require the use of multiple specific primers. However, group-specific primers may be useful for monitoring populations that were detected only by a culture-independent inventory of the milk bacterial community, such as members of the Clostridiaceae, Aerococcaceae (Facklamia), and Betaproteobacteria (Ralstonia). Microbial dynamics results from the combined influence of microbial interactions and multiple technological and environmental factors (pH, temperature, salinity) affecting and being affected in return by microorganism activities. Changes in biotic and abiotic conditions may be responsible for the limited increase in the counts of the populations, mainly of gram-negative bacteria, growing on PCAI medium during cheese manufacture and ripening and for the concomitant drop in richness of the corresponding SSCP patterns from day 1. Further work should be done to investigate the functions of persistent subdominant populations, such as of Chryseobacterium spp., in this system.


We are indebted to DATAR—Massif Central for financial support.

We also sincerely thank the Pole Fromager AOC Massif Central, the Saint-Nectaire cheese trade, and the farmers for their collaboration. We sincerely thank Nadia Chougui, Béatrice Desserre, and René Lavigne for their technical assistance and Cécile Callon for helpful discussion. Thanks are extended to Philip Rousseau-Cunningham and Kate James for proofreading of our English.


[down-pointing small open triangle]Published ahead of print on 26 January 2007.


1. Andrighetto, C., E. Knijff, A. Lombardi, S. Torriani, M. Vancanneyt, K. Kersters, J. Swings, and F. Dellaglio. 2001. Phenotypic and genetic diversity of enterococci isolated from Italian cheeses. J. Dairy Res. 68:303-316. [PubMed]
2. Aquilanti, L., L. Dell'Aquila, E. Zannini, A. Zocchetti, and F. Clementi. 2006. Resident lactic acid bacteria in raw milk Canestrato Pugliese cheese. Lett. Appl. Microbiol. 43:161-167. [PubMed]
3. Beuvier, E., K. Berthaud, S. Cegarra, A. Dasen, S. Pochet, S. Buchin, and G. Duboz. 1997. Ripening and quality of swiss-type cheese made from raw, pasteurized or microfiltered milk. Int. Dairy J. 7:311-323.
4. Brennan, N. M., A. C. Ward, T. P. Beresford, P. F. Fox, M. Goodfellow, and T. M. Cogan. 2002. Biodiversity of the bacterial flora on the surface of a smear cheese. Appl. Environ. Microbiol. 68:820-830. [PMC free article] [PubMed]
5. Brosius, J., T. J. Dull, D. Sleeter, and H. F. Noller. 1981. Gene organisation and primary structure of a ribosomal DNA operon from Escherichia coli. J. Mol. Biol. 148:107-127. [PubMed]
6. Callon, C., L. Millet, and M. C. Montel. 2004. Diversity of lactic acid bacteria isolated from AOC Salers cheese. J. Dairy Res. 71:231-244. [PubMed]
7. Cocolin, L., N. Innocente, M. Biasutti, and G. Comi. 2004. The late blowing in cheese: a new molecular approach based on PCR and DGGE to study the microbial ecology of the alteration process. Int. J. Food Microbiol. 90:83-91. [PubMed]
8. Coenye, T., and P. Vandamme. 2003. Intragenomic heterogeneity between multiple 16S ribosomal RNA operons in sequenced bacterial genomes. FEMS Microbiol. Lett. 228:45-49. [PubMed]
9. Coppola, S., G. Blaiotta, D. Ercolini, and G. Moschetti. 2001. Molecular evaluation of microbial diversity occurring in different types of Mozzarella cheese. J. Appl. Microbiol. 90:414-420. [PubMed]
10. De Buyser, M. L., B. Dufour, M. Maire, and V. Lafarge. 2001. Implication of milk and milk products in food-borne diseases in France and in different industrialised countries. Int. J. Food Microbiol. 67:1-17. [PubMed]
11. Demarigny, Y., E. Beuvier, S. Buchin, S. Pochet, and R. Grappin. 1997. Influence of raw milk microflora on the characteristics of Swiss-type cheeses: II. Biochemical and sensory characteristics. Lait 77:151-167.
12. Denis, C., M. Gueguen, E. Henry, and D. Levert. 2001. New media for the numeration of cheese surface bacteria. Lait 81:365-379.
13. Duthoit, F., C. Callon, L. Tessier, and M. C. Montel. 2005. Relationships between sensorial characteristics and microbial dynamics in “Registered Designation of Origin” Salers cheese. Int. J. Food Microbiol. 103:259-270. [PubMed]
14. Duthoit, F., J. J. Godon, and M. C. Montel. 2003. Bacterial community dynamics during production of registered designation of origin Salers cheese as evaluated by 16S rRNA gene single-strand conformation polymorphism analysis. Appl. Environ. Microbiol. 69:3840-3848. [PMC free article] [PubMed]
15. Ercolini, D., P. J. Hill, and C. E. R. Dodd. 2003. Bacterial community structure and location in stilton cheese. Appl. Environ. Microbiol. 69:3540-3548. [PMC free article] [PubMed]
16. Ercolini, D., G. Moschetti, G. Blaiotta, and S. Coppola. 2001. The potential of a polyphasic PCR-dGGE approach in evaluating microbial diversity of natural whey cultures for water-buffalo Mozzarella cheese production: bias of culture-dependent and culture-independent analyses. Syst. Appl. Microbiol. 24:610-617. [PubMed]
17. Feurer, C., F. Irlinger, H. E. Spinnler, P. Glaser, and T. Vallaeys. 2004. 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. [PubMed]
18. Hugo, C. J., P. Segers, B. Hoste, M. Vancanneyt, and K. Kersters. 2003. Chryseobacterium joostei sp. nov., isolated from the dairy environment. Int. J. Syst. Evol. Microbiol. 53:771-777. [PubMed]
19. Ishikawa, M., K. Nakajima, M. Yanagi, Y. Yamamoto, and K. Yamasato. 2003. Marinilactibacillus psychrotolerans gen. nov., sp. nov., a halophilic and alkaliphilic marine lactic acid bacterium isolated from marine organisms in temperate and subtropical areas of Japan. Int. J. Syst. Evol. Microbiol. 53:711-720. [PubMed]
20. Jooste, P. J., and C. J. Hugo. 1999. The taxonomy, ecology and cultivation of bacterial genera belonging to the family Flavobacteriaceae. Int. J. Food Microbiol. 53:81-94. [PubMed]
21. Klappenbach, J. A., P. R. Saxman, J. R. Cole, and T. M. Schmidt. 2001. rrndb: the Ribosomal RNA Operon Copy Number Database. Nucleic Acids Res. 29:181-184. [PMC free article] [PubMed]
22. Kleinsteuber, S., V. Riis, I. Fetzer, H. Harms, and S. Muller. 2006. Population dynamics within a microbial consortium during growth on diesel fuel in saline environments. Appl. Environ. Microbiol. 72:3531-3542. [PMC free article] [PubMed]
23. Lafarge, V., J. C. Ogier, V. Girard, V. Maladen, J.-Y. Leveau, A. Gruss, and A. Delacroix-Buchet. 2004. Raw cow milk bacterial population shifts attributable to refrigeration. Appl. Environ. Microbiol. 70:5644-5650. [PMC free article] [PubMed]
24. Le Bourhis, A.-G., K. Saunier, J. Dore, J.-P. Carlier, J.-F. Chamba, M.-R. Popoff, and J.-L. Tholozan. 2005. Development and validation of PCR primers to assess the diversity of Clostridium spp. in cheese by temporal temperature gradient gel electrophoresis. Appl. Environ. Microbiol. 71:29-38. [PMC free article] [PubMed]
25. Maidak, B. L., J. R. Cole, C. T. Parker, G. M. Garrity, N. Larsen, B. Li, T. G. Lilburn, M. J. McCaughey, G. J. Olsen, R. Overbeek, S. Pramanik, T. M. Schmidt, J. M. Tiedje, and C. R. Woese. 1999. A new version of the RDP (Ribosomal Database Project). Nucleic Acids Res. 27:171-173. [PMC free article] [PubMed]
26. Maoz, A., R. Mayr, and S. Scherer. 2003. Temporal stability and biodiversity of two complex antilisterial cheese-ripening microbial consortia. Appl. Environ. Microbiol. 69:4012-4018. [PMC free article] [PubMed]
27. Marcellino, N., and D. R. Benson. 1992. Scanning electron and light microscopic study of microbial succession on Bethlehem St. Nectaire cheese. Appl. Environ. Microbiol. 58:3448-3454. [PMC free article] [PubMed]
28. Marino, M., M. Maifreni, and G. Rondinini. 2003. Microbiological characterization of artisanal Montasio cheese: analysis of its indigenous lactic acid bacteria. FEMS Microbiol. Lett. 229:133-140. [PubMed]
29. McGarvey, J. A., W. G. Miller, S. Sanchez, and L. Stanker. 2004. Identification of bacterial populations in dairy wastewaters by use of 16S rRNA gene sequences and other genetic markers. Appl. Environ. Microbiol. 70:4267-4275. [PMC free article] [PubMed]
30. Miambi, E., J. P. Guyot, and F. Ampe. 2003. Identification, isolation and quantification of representative bacteria from fermented cassava dough using an integrated approach of culture-dependent and culture-independent methods. Int. J. Food Microbiol. 82:111-120. [PubMed]
31. Mounier, J., R. Gelsomino, S. Goerges, M. Vancanneyt, K. Vandemeulebroecke, B. Hoste, S. Scherer, J. Swings, G. F. Fitzgerald, and T. M. Cogan. 2005. Surface microflora of four smear-ripened cheeses. Appl. Environ. Microbiol. 71:6489-6500. [PMC free article] [PubMed]
32. Nelson, F. E. 1940. The effect of the new standard milk agar on the plate count of dairy products. J. Bacteriol. 39:263-272. [PMC free article] [PubMed]
33. Ogier, J.-C., V. Lafarge, V. Girard, A. Rault, V. Maladen, A. Gruss, J.-Y. Leveau, and A. Delacroix-Buchet. 2004. Molecular fingerprinting of dairy microbial ecosystems by use of temporal temperature and denaturing gradient gel electrophoresis. Appl. Environ. Microbiol. 70:5628-5643. [PMC free article] [PubMed]
34. Ogier, J. C., O. Son, A. Gruss, P. Tailliez, and A. Delacroix-Buchet. 2002. Identification of the bacterial microflora in dairy products by temporal temperature gradient gel electrophoresis. Appl. Environ. Microbiol. 68:3691-3701. [PMC free article] [PubMed]
35. Ostlie, H. M., L. Eliassen, A. Florvaag, and S. Skeie. 2004. Phenotypic and PCR-based characterization of the microflora in Norvegia cheese during ripening. Int. J. Food Microbiol. 94:287-299. [PubMed]
36. Perez Pulido, R., N. Ben Omar, H. Abriouel, R. Lucas Lopez, M. Martinez Canamero, and A. Galvez. 2005. Microbiological study of lactic acid fermentation of caper berries by molecular and culture-dependent methods. Appl. Environ. Microbiol. 71:7872-7879. [PMC free article] [PubMed]
37. Peu, P., H. Brugere, A.-M. Pourcher, M. Kerouredan, J. J. Godon, J. P. Delgenes, and P. Dabert. 2006. Dynamics of a pig slurry microbial community during anaerobic storage and management. Appl. Environ. Microbiol. 72:3578-3585. [PMC free article] [PubMed]
38. Rademaker, J. L. W., M. Peinhopf, L. Rijnen, W. Bockelmann, and W. H. Noordman. 2005. The surface microflora dynamics of bacterial smear-ripened Tilsit cheese determined by T-RFLP DNA population fingerprint analysis. Int. Dairy J. 15:785-794.
39. Randazzo, C. L., S. Torriani, A. D. L. Akkermans, W. M. de Vos, and E. E. Vaughan. 2002. Diversity, dynamics, and activity of bacterial communities during production of an artisanal Sicilian cheese as evaluated by 16S rRNA analysis. Appl. Environ. Microbiol. 68:1882-1892. [PMC free article] [PubMed]
40. Randazzo, C. L., E. E. Vaughan, and C. Caggia. 2006. Artisanal and experimental Pecorino Siciliano cheese: microbial dynamics during manufacture assessed by culturing and PCR-DGGE analyses. Int. J. Food Microbiol. 109:1-8. [PubMed]
41. Shimomura, K., S. Kaji, and A. Hiraishi. 2005. Chryseobacterium shigense sp. nov., a yellow-pigmented, aerobic bacterium isolated from a lactic acid beverage. Int. J. Syst. Evol. Microbiol. 55:1903-1906. [PubMed]
42. Terzaghi, B. E., and W. E. Sandine. 1975. Improved medium for lactic streptococci and their bacteriophages. Appl. Microbiol. 29:807-813. [PMC free article] [PubMed]

Articles from Applied and Environmental Microbiology are provided here courtesy of American Society for Microbiology (ASM)
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...