Enterococci are naturally present in the gastrointestinal tract of humans and animals, in soil, surface waters, on plants and also in foods, especially food of animal origin, such as cheeses and fermented sausages (Franz et al. Reference Franz, Holzapfel and Stiles1999).
Enterococci are an essential part of the microflora of different artisanal European cheeses. They occur and grow in different types of cheeses produced from either raw or pasteurised milk (Franz, Reference Franz, Stiles, Schleifer and Holzapfel2003). Despite their historical bad reputation of faecal contaminants, nowadays enterococci are recognised as a normal part of the microbial population of traditional cheeses. Even more, enterococci play an important role in the ripening of some traditional cheeses and contribute to their specific taste and flavour. Enterococci have been successfully used as a significant component of the starter cultures for the production of different European cheeses (Giraffa, Reference Giraffa2003); furthermore, in some countries they are used as probiotics, since many strains are able to survive in the gastrointestinal tract and to produce antimicrobial substances, including bacteriocins (Franz et al. Reference Franz, Holzapfel and Stiles1999).
The identification of genus Enterococcus has always been problematic, since there are no definite phenotypic criteria available for clear discrimination of enterococci from other Gram-positive, catalase-negative, coccus-shaped bacteria (Domig et al. Reference Domig, Mayer and Kneifel2003). For reliable and fast identification, especially from sources with a heterogeneous microbial population, molecular-based methods are essential (Klein, Reference Klein2003). DGGE of PCR-amplified 16S rDNA fragments is currently one of the most used methods for studying complex bacterial communities (Muyzer, Reference Muyzer1999; Ercolini, Reference Ercolini2004; Temmerman et al. Reference Temmerman, Huys and Swings2004).
Selection of the appropriate primer pair is a crucial step for successful DGGE analysis. A lot of reports describe the use of one primer set for analyses of various food samples, but only a few reports are focused on using more and different primer pairs for analysing a particular food sample. The PCR primers used in our study were selected from published research describing DGGE analysis of microbial communities in DNA extracted directly from food samples such as cheeses, sausages and probiotic products. Six particular primer pairs were selected according to criteria as follows: different 16S rDNA regions amplification and suitability for DGGE analysis of bacterial population in fermented food/dairy samples. Selected primers were evaluated on DNA of 12 pure enterococci strains and 8 other LAB, most often found in cheeses (7 lactobacilli and 1 Lc. lactis). Primer sets were also DGGE analysed on DNA extracted directly from cheese samples and from enterococcal population harvested from cheese, respectively.
Materials and Methods
Bacterial strains, growth conditions and DNA extraction
Type strains of 12 enterococci were selected: Enterococcus faecalis LMG 7937T, Ent. faecium LMG 11423T, Ent. hirae LMG 6399T, Ent. durans LMG 10746T, Ent. pseudoavium LMG 11426T, Ent. avium LMG 10744T, Ent. malodoratus LMG 10747T, Ent. casseliflavus LMG 10745T, Ent. mundtii LMG 10748T, Ent. asini LMG 18727T, Ent. pallens LMG 21842T and Ent. villorum LMG 12287T; and eight other bacterial species often found in cheeses: Lactococcus lactis IM 128, Lactobacillus helveticus ATCC 15009T, Lb. rhamnosus ATCC 53103, Lb. acidophilus ATCC 4356, Lb. paracasei IM 307, Lb. brevis LMG 6906, Lb. plantarum NCDO 1193 and Lb. casei DSM 20011 (LMG: Belgian Co-ordinated Collections of Micro-organisms; ATCC: American Type Culture Collection; IM: Chair of Dairy Science, Biotechnical Faculty, Slovenia; NCDO: National Collection of Dairy Organisms; DSM: German Collection of Microorganisms and Cell Cultures). Enterococci and lactococci were cultured in M17 broth at 37°C or 30°C, respectively, and lactobacilli in MRS broth at 37°C. DNA was extracted using the Wizard® Genomic DNA Purification Kit (Promega, Wisconsin, USA).
DNA extraction from bulk cells and cheese samples
DNA was extracted from five artisanal cheese samples, made from raw ewes' milk (cheeses 2 and 3) or cows' milk (cheeses 1 and 5). Cheese 4 was made from thermized cows' milk. Cheese samples (10 g) were homogenized with BagMixer® 400 (Interscience, St Nom, France) in 90 ml of 2% (wt/v) K2HPO4 (Merck KGaA, Darmstadt, Germany). For DNA extraction directly from cheese, 10 ml of homogenate was pelleted, resuspended in 1 ml H2O and extraction was performed according to Cocolin et al. (Reference Cocolin, Manzano, Cantoni and Comi2001). Prior DNA extraction from bulk cells, 100 μl of homogenate was plated on Citrate Azide Tween Carbonate (CATC) agar (Merck KGaA, Darmstadt, Germany) and incubated for 24 h at 37°C. Outgrown colonies were collected by rinsing with 1 ml of quarter-strength Ringer's solution and DNA from 1 ml of cell suspension was further extracted using the Wizard® Genomic DNA Purification Kit (Promega, Wisconsin, USA).
PCR primers and protocols
The primers used in this study are listed in Table 1. PCR was performed in final volume of 30 μl containing 1X Colorless GoTaqTM Reaction Buffer, GoTaqTM DNA Polymerase (Promega, Wisconsin, USA), 0·1 mm of each dNTP, 10 μm of primers and MgCl2 at the concentrations listed in Table 1. The amplification programme for primer pairs HDA1-GC/HDA2, V3f-GC/V3r, U968-GC/L1401 and Ec1055-GC/Ec1492 was initiated with denaturation at 95°C for 3 min, followed by 35 cycles of denaturation at 95°C for 30 s, annealing at 58°C for 30 s and elongation at 72°C for 40 s, and terminated with final elongation at 72°C for 5 min. For primer pair P1V1-GC/P2V1 the annealing was at 54°C for 30 s and elongation at 72°C for 30 s; and for primer pair Com1-GC/Com2Ph annealing was at 52°C for 30 s and elongation at 72°C for 1 min. PCR products were examined on 1·5% agarose gels, and the reaction mixtures yielding products of expected size were used for DGGE analysis.
PCR primers used in this study
† A GC clamp was added to each forward primer according to Walter et al. (Reference Walter, Tannock, Tilsala-Timisjarvi, Rodtong, Loach, Munro and Alatossava2000)
‡ In E. coli 16S rDNA gene
DGGE
DGGE was performed with D GENE system (Bio-Rad, California, USA) using 16 cm×16 cm×1 mm gels. DGGE electrophoresis was performed on 8% (v/v) polyacrylamide gels (acrylamide/bisacrylamide ratio 37:1 [Sigma-Aldrich Chemie GmbH, Steinheim, Germany]) with denaturating gradient 30 to 60% for primer pair P1V1-GC/P2V1, 30 to 50% for primer pairs HDA1-GC/HDA2 and V3f-GC/V3r and for the remaining primer pairs 40 to 55% (100% corresponds to 7 m urea [Sigma-Aldrich Chemie, Buchs, Switzerland] and 40% (v/v) formamide [Sigma-Aldrich Chemie GmbH]) in 1×TAE buffer. Gels were electrophoresed at constant voltage of 130 V for 5–6 h at 60°C, stained with SYBR safeTM (Invitrogen, California, USA) and photographed with UV illuminator.
Results and Discussion
DGGE analysis of pure strains
Six sets of primers, amplifying different 16S rDNA regions, were applied on DNA of 20 selected strains and 5 cheese samples. They all generated PCR fragments of expected length, which is different from Cocolin et al. (Reference Cocolin, Diez, Urso, Rantsiou, Comi, Bergmaier and Beimfohr2007) who reported a lack of PCR product for some of the tested strains when primer pair HDA1-GC/HDA2 was used. All PCR fragments were subjected to DGGE analysis and the results are shown in Figs. 1 and 2.
DGGE profiles of tested strains obtained with primers P1V1-GC/P2V1 (left), HDA1-GC/HDA2 (middle), and V3f-GC/V3r (right).
DGGE profiles of tested strains obtained with primers Com1-GC/Com2-Ph (left), Ec1055-GC/Ec1392 (middle), and U968-GC/L1401 (right).
Primer pair P1V1-GC/P2V1, amplifying V1 region of 16S rDNA, generated multiple bands for all of the tested strains, which is in agreement with previous reports (Cocolin et al. Reference Cocolin, Innocente, Biasutti and Comi2004, Reference Cocolin, Diez, Urso, Rantsiou, Comi, Bergmaier and Beimfohr2007). Since comparison of 16S rDNA sequences of enterococcal population showed the highest heterogeneity within the V1 region, attempts were made to find a reason for the appearance of multiple bands. Additional bands with lower intensity could be PCR artefacts that depend upon PCR conditions (Ogier et al. Reference Ogier, Son, Gruss, Tailliez and Delacroix-Buchet2002). Therefore, some modifications were made to the PCR programme and the products amplified with different PCR conditions were subjected to DGGE gel. First the temperature influence on the PCR programme was tested by lowering the annealing temperature from 56 to 52°C in intervals of 2 deg C every three cycles with the additional 25 cycles performed at 50°C for 1 min. Denaturing step was at 95°C for 1 min and extension was at 72°C for 1 min. Initial denaturation and final elongation steps were as described in materials and methods. However, no significant change on DGGE gel was observed. Then, the influence of MgCl2 (1·5, 2·0, 2·5 and 3·0 mm) and DNA (dilution 1:10 and 1:100) concentrations were tested, but again no differences were observed on DGGE gel. Further, the influence of different PCR primers on DGGE was examined, since slightly modified primers and/or primer position can alter the melting behaviour of DNA fragments (Wu et al. Reference Wu, Hayes, Osinga, Mulder, Looman, Buys and Hofstra1998). First, modified reverse primer was used for PCR reaction, because original primer differs by two nucleotides from enterococcal 16S rDNA sequences found on NCBI homepage (www.ncbi.nlm.nih.gov). The new primer sequence was 5′-TTA CCC ACG TGT TAC TCA CC-3′, nevertheless, no change on DGGE gel was observed, which is in agreement with the findings of Boutte et al. (Reference Boutte, Grubisic, Balthasart and Wilmotte2006). Second, several newly constructed and applied forward and reverse primers, amplifying V1 variable 16S rDNA region, resulted in some changes on DGGE gel, but still multiple bands for single species were present. A possible explanation for these multiple bands phenomena could be the presence of heterogeneous copies of rRNA operons (von Wintzingerode et al. Reference von Wintzingerode, Gobel and Stackebrandt1997; Muyzer, Reference Muyzer1999; Randazzo et al. Reference Randazzo, Torriani, Akkermans, de Vos and Vaughan2002; Florez & Mayo, Reference Florez and Mayo2006) since enterococci possess between five and six rRNA operons (Gurtler, Reference Gurtler1999; Zarazaga et al. Reference Zarazaga, Tenorio, Del Campo, Ruiz-Larrea and Torres2002). For DGGE analysis of the enterococcal population, primers P1V1-GC/P2V1 seem to be more convenient for analysing pure cultures than whole communities since multiple bands for single species could lead to overestimation of the number and variety of bacteria present.
DGGE analysis of fragments amplified with primer pairs HDA1-GC/HDA2, V3f-GC/V3r and Com1-GC/Com2-Ph revealed single band for most of the analysed species. Even if a multiple band pattern appeared, a dominant reference band was always recognisable. In some cases, bands of different bacterial species migrated to the same position on the gel which was expected since there was high sequence similarity; however, it is possible to differentiate between migration distances of bacterial groups. Several authors observed co-migration of bands derived from different species (Walter et al. Reference Walter, Tannock, Tilsala-Timisjarvi, Rodtong, Loach, Munro and Alatossava2000; Ercolini et al. Reference Ercolini, Moschetti, Blaiotta and Coppola2001a; Fasoli et al. Reference Fasoli, Marzotto, Rizzotti, Rossi, Dellaglio and Torriani2003; Ogier et al. Reference Ogier, Lafarge, Girard, Rault, Maladen, Gruss, Leveau and Delacroix-Buchet2004; Fontana et al. Reference Fontana, Vignolo and Cocconcelli2005b). Therefore, DGGE with primer pairs HDA1-GC/HDA2, V3f-GC/V3r and Com1-GC/Com2-Ph can be used for initial screening of possible identities of bacterial isolates and then identity could be confirmed with species-specific PCR (Walter et al. Reference Walter, Tannock, Tilsala-Timisjarvi, Rodtong, Loach, Munro and Alatossava2000). In this way, isolates could be tested only with a couple of primer pairs instead of a complete primer collection. DGGE fingerprints derived with primers V3f-GC/V3r were almost identical to fingerprints derived with primers HDA1-GC/HDA2. These results were somehow expected, since analysis of 16S rDNA showed that both primer pairs amplify the same region.
DGGE separation of amplicons generated with primer pair Ec1055-GC/Ec1392 performed fuzzy bands for all enterococcal strains and for Lc. lactis which is because of the presence of multiple melting domains (Wu et al. Reference Wu, Hayes, Osinga, Mulder, Looman, Buys and Hofstra1998). This phenomenon was also confirmed with sequence analysis using WinMeltTM software (Bio-Rad Laboratories, California, USA), a program for analysing the melting profile of DNA sequence. Analysis showed that in enterococcal strains primer pair Ec1055-GC/Ec1392 amplifies the fragment with an interior low melting domain consecutively leading to the production of fuzzy bands (Wu et al. Reference Wu, Hayes, Osinga, Mulder, Looman, Buys and Hofstra1998). DGGE analysis of tested strains obtained with primers U968-GC/L1401 produced double bands (next to the dominant band one weaker band is present). Our results revealed that the use of this primer pair is not optimal for distinguishing bacterial species belonging to genus Enterococcus from other tested bacteria, since the migration distances of different bacterial species are very close together.
DGGE analysis of DNA extracted directly from cheese samples
In order to define which primers are suitable for direct analysis of enterococci in cheese samples, DGGE analysis of 5 cheese samples was performed. DNA was extracted directly from cheese samples as well as from bulk cells harvested from CATC, and PCR amplifications with all six primer pairs were performed. DGGE fingerprints of cheese samples, with addition of marker composed from Ent. faecalis and Ent. faecium, two enterococcal species most frequently found in cheese, are shown in Figs. 3 and 4.
DGGE profiles of cheese samples and bulk cells obtained with primers P1V1-GC/P2V1 (left), HDA1-GC/HDA2 (middle), and V3f-GC/V3r (right). M-marker composed of E. faecalis (A) and E. faecium (B) amplicons.
DGGE profiles of cheese samples and bulk cells obtained with primers Com1-GC/Com2-Ph (left), Ec1055-GC/Ec1392 (middle), and U968-GC/L1401 (right). M-marker composed of E. faecalis (A), E. faecium (B) and Lc. lactis (C).
DGGE analysis of cheese samples with primer pairs HDA1-GC/HDA2 and V3f-GC/V3r (Fig. 3) exhibited similar fingerprints, composed from 4 to 9 clear bands. The presence of band at the same position as Ent. faecalis was detected only on fingerprint for cheese 1, while no band was present at the same position as Ent. faecium. Using other primer pairs, bands at positions typical for Ent. faecalis and Ent. faecium were not detected. Several authors reported the lack of bands corresponding to enterococci on DGGE gel from fermented products, although the presence of enterococci was confirmed with other methods (Ercolini et al. Reference Ercolini, Moschetti, Blaiotta and Coppola2001b; Fontana et al. Reference Fontana, Sandro Cocconcelli and Vignolo2005a & Reference Fontana, Vignolo and Cocconcellib; Florez & Mayo, Reference Florez and Mayo2006). In contrast, Coppola et al. (Reference Coppola, Blaiotta, Ercolini and Moschetti2001) and Randazzo et al. (Reference Randazzo, Torriani, Akkermans, de Vos and Vaughan2002 & Reference Randazzo, Vaughan and Caggia2006) were able to show the presence of enterococci with DGGE. Poor detection of enterococci from cheese could be due to their presence in low numbers, insufficient DNA extraction or preferential PCR amplification. Bacterial species could be detected in mixed populations if they are present in more than 104 CFU/g of cheese/food sample (Cocolin et al. Reference Cocolin, Manzano, Cantoni and Comi2001; Ercolini et al. Reference Ercolini, Moschetti, Blaiotta and Coppola2001b; Temmerman et al. Reference Temmerman, Scheirlinck, Huys and Swings2003; Fontana et al. Reference Fontana, Vignolo and Cocconcelli2005b) or when they represent at least 1% of the total microbial population (Muyzer et al. Reference Muyzer, de Waal and Uitterlinden1993; Fasoli et al. Reference Fasoli, Marzotto, Rizzotti, Rossi, Dellaglio and Torriani2003). In the tested cheeses enterococci were present in more than 105 CFU/g (determined by plate count on CATC media); with the exception of cheese 4 with only 103 CFU/g. Therefore, in most tested cheeses enterococci are present in numbers above detection limit. Another limitation in analysing enterococci from cheese is insufficient DNA extraction. It is very difficult to extract pure DNA from complex samples like cheese, since natural constituents (such as proteins, lipids, salts) can persist until the end of extraction and are found in extract where they might act as inhibitors. Furthermore, not all species have the same sensitivity to lytic agents because of differences in cell wall organisation, particularly Gram-positive species (like enterococci) may not all lyse equally well (Ercolini, Reference Ercolini2004; Ogier et al. Reference Ogier, Lafarge, Girard, Rault, Maladen, Gruss, Leveau and Delacroix-Buchet2004). Therefore, we preliminarily tested 8 different protocols for DNA extraction from the cheese samples and finally the protocol of Cocolin et al. (Reference Cocolin, Manzano, Cantoni and Comi2001) was applied, as it yielded more enterococcal DNA than others (data not shown). Another possibility for lack of enterococcal detection in cheese samples is differential/preferential amplification of rDNA genes by PCR. Choice of primers, the varying mol percent G+C composition of 16S rDNA genes, genome size and number of different rRNA operons can affect the differential amplification. DNA-associated molecules which resist standard procedures during DNA purification could be another source of diminished amplification efficiency of 16S rDNA of Gram-positive bacteria (von Wintzingerode et al. Reference von Wintzingerode, Gobel and Stackebrandt1997).
DGGE analysis of DNA extracted from bulk cells
Propagation step of cheese samples on CATC media improved detection of enterococcal species. DGGE analysis of microbial population of cheeses grown on CATC media using primer pairs HDA1-GC/HDA2 and V3f-GC/V3r (Fig. 3) exhibited a band at the same position as Ent. faecalis for all cheese samples and a band at Ent. faecium position for four samples. It has to be pointed out, that Ent. mundtii migrated to the same position as Ent. faecalis, so the presence of a band at Ent. faecalis position could indicate the presence of either or both species. The same is significant for species Ent. faecium, Ent. hirae, and Ent. durans. Similar results were obtained using primer pair Com 1-GC/Com2-Ph (Fig. 4), besides additional band was present for all cheese samples at a position distinct from the positions of the tested enterococcal strains. Primer pair P1V1-GC/P2V1 (Fig. 3) generated several bands for all cheese samples. Since multiple bands arose from a single species, it was not possible to confirm the presence of either Ent. faecalis or Ent. faecium in tested samples. DGGE analysis revealed that primer pairs U968-GC/L1401 and Ec1055-GC/Ec1392 (Fig. 4) did not differentiate between Ent. faecalis and Ent. faecium; moreover, detection of either species is vague, since the bands are not sharp.
Conclusions
Our work strongly supports discriminatory potential of PCR-DGGE technique in analysis of enterococcal species. However, regarding the results of our study, it is not simple to find primers for clear differentiation and identification of related bacterial species and also of bacterial species belonging to different genera. Perhaps the selection of genus specific primers or primers for functional or technological genes would enable better discrimination of different enterococcal species. Beside primer selection numerous factors influence on DGGE results for instance protocol of DNA isolation, presence of bacterial population under detection limit, detection of non-viable cells, presence of multiple copies of rRNA operons. Therefore it is worth stressing that DGGE results must be interpreted with great caution.
Mostly DGGE analysis is combined with sequencing of excised bands which is not suitable for identification of related bacterial species since they often harbour the same nucleotide sequence. Besides, sequencing of bands would result in extended time of analysis which is not acceptable for control of food-processing where fast and reliable results are required. For reliable detection of enterococcal population from the food/dairy environment, where viable cells are the ones that count, a combination of both, culture-dependent and culture independent methods, is recommended.
