Biodiversity and anti-listerial activity of surface microbial consortia from limburger, reblochon, livarot, tilsit and gubbeen cheese.

Kluyveromyces marxianus and Kluyveromyces lactis are two closely related yeasts isolated from dairy products, particularly from some fermented milks such as kefir and from many cheese. It is well known that these yeast cannot be easily differentiated using physiological or molecular tests. During this project, it was not possible to differentiate the two species by phenotypic investigations (physiological tests and FT-IR spectroscopy) or by the use of PCR primers targeting the lac4 gene. We developed a protocol using the RAM-PCR technique which allowed the differentiation of the two type strains of K. marxianus (CBS 712T) and K. lactis (CBS 683T) and showed that all Livarot isolates belonged to K. lactis. Isolates belonging to the same species showed the same profile, differing from the other. These results are going to be disseminated by the mean of a scientific paper, which have to be prepared (some additional results have to be obtained before). No benefits are expected. End-users of these results are the scientific community working in particular in mycology, but also cheese producers and producers of commercial strains of microorganisms for the smear cheese industry. This technique was used with others primers to differentiate other yeast species from wine. The primer used in this work allowed also the differentiation of nine other species from cheese.

Approximately 160 coccus-shaped isolates from three different batches of cheese (or cheesemaking facilities) and three different stages of ripening (early middle and late) of 5 European cheeses are deposited in our R-collection (Research collection, Laboratory of Microbiology, University of Ghent). Gram staining, motility test, cell and colony morphology, catalase test and oxidase test were performed on all isolates. A set of over 60 reference strains was included for identification by rep-PCR. The set of reference strains included the genera Staphylococcus, Micrococcus, Macrococcus, Kocuria and Rothia. Each species was represented by several strains. The reference set was chosen accordingly to the most common species present on red smear cheeses. DNA of cocci was prepared using a lysostaphin and a phenol-chloroform step. PCR was performed according to Versalovic et al. (1994, Molecular and Cellular Biology. 5: 25-40) with the primer (GTG)5. Band patterns were analyzed using the Pearson coefficient and the UPGMA dendrogram type with the Bionumerics software (Applied Maths). If a cheese isolate did not cluster with any of the reference strains, a partial 16S sequence analysis was performed in order to assign the isolate to a species group. All isolates were identified either to species or species-group level confirming rep-PCR as a rapid tool for identification. The most common species present on the cheeses were Staphylococcus cohnii/equorum (between which rep-PCR is not able to distinguish), and Staphylococcus saprophyticus. Two representatives of each strain were deposited in the BCCM culture collection for research purposes, for a total of 51 staphylococci and other coccus-shaped bacteria. The project enabled the construction of a comprehensive database on various coccus-shaped species, improving the process of identification of such isolates. Hence, the BCCM/LMG Bacteria Collection can offer a service for identification and molecular typing of cocci isolated from foods.

Over 400 coryneform isolates from three different batches of cheese (or cheesemaking facilities) and three different stages of ripening (early middle and late) of 5 European cheeses are deposited in our R-collection (Research collection, Laboratory of Microbiology, University of Ghent). Gram staining, motility test, cell and colony morphology, catalase test and oxidase test were performed on all isolates. A set of over 120 reference strains was included for identification by rep-PCR. The set of reference strains includes the genera Arthrobacter, Brevibacterium, Brachybacterium, Corynebacterium, Clavibacter, Curtobacterium, Leucobacter, Mycetocola, Microbacterium and Rhodococcus. Each species was present with several strains. The reference set was chosen accordingly to the most common species present on red smear cheeses. DNA of coryneforms was prepared with a mutanolysin and a phenol-chloroform step. PCR was performed according to Versalovic et al. (1994, Molecular and Cellular Biology. 5: 25-40) with the primer BOXA1R. Band patterns were analysed using the Pearson coefficient and the UPGMA dendrogram type with the Bionumerics software (Applied Maths). If a cheese isolate did not cluster with any of the reference strains, a partial 16S sequence analysis was performed in order to assign the isolate to a species group. All isolates were identified either to species or species-group level confirming rep-PCR as a rapid tool for identification. The most common species present on the cheeses were Arthrobacter arilaitensis and Brevibacterium aurantiacum, two newly described species, Corynebacterium casei, Corynebacterium variabile and Microbacterium gubbeenense. Two representatives of each strain were deposited in the BCCM culture collection for research purposes, for a total of 122 coryneforms. The project enabled the construction of a comprehensive database on various coryneform species, improving the process of identification of such isolates. Hence, the BCCM/LMG Bacteria Collection can offer a service for identification and molecular typing of coryneforms.

Three batches of Reblochon cheese were produced in different farmhouses without addition of yeast or surface microflora starter. The predominance of Galactomyces Geotrichum on the surface of Reblochon cheese was found. Nevertheless, its intraspecies diversity was low since only three different clusters were obtained by similarity analysis of the FTIR spectra. Galactomyces Geotrichum appeared as a yeast well adapted to the biotope of this cheese. The current high occurrence of kluyveromyces lactis and its high diversity was not observed in the previous study on Reblochon cheese by Bärtshi et al., (1994). They rarely isolated this species (about 1%) in the five cheese factories investigates. Conversely, Debaryomyces hansenii was very less isolated than previously and only on cheeses from farmhouse. That is surprising because this species is generally recognised as a dominant species on numerous types of cheeses made in many countries on mould cheese as well as on smeared cheese. Toruloaspora delbrueckii was observed at the same low level in the two studies on Reblochon cheese. Other yeast species found here, especially Saccharomyces servazzi and Saccharomyces unisporus were not isolated in the previous work on Reblochon cheese. Satunispora was only found in Reblochon cheese made in farmhouse. The species was delineated using biomolecular tools. Finally, yeast flora found on Reblochon cheese in the course of years between factories, at least in farmhouse, have shown an original feature and high diversity. The microflora of Reblochon cheeses seems to be composed almost exclusively by coryneform bacteria. For the Arthrobacter genus, only A. arilaitensis was found at the end of ripening on cheese made in the three farmhouses. In the Brevibacterium linens group, two species were found: B. linens and much more abundant B. aurantiacum. Corynebacterium is the dominant genus since about half of identified bacterial isolates belong to this genus. C. variabile was the main species encountered on cheese whatever stage of repening or farmhouses. Mycetocola is a new genus of the Microbacterium family. To our knowledge it is the first time that this genus is isolated from cheese. The last genus which have shown a noticeable abundance was Microbacterium especially M. gubbeenense, found mostly at the end of ripening according to its sensitivity to low pH. Others genera were sporadically isolated such as Brachybacterium, Rothia, Staphylococcus. Biodiversity of yeasts and bacteria on the surface of Reblochon cheese appeared relatively large in comparison to others smear cheeses because commercial culture was inoculated on studied cheeses. The diversity of microflora found on this traditional cheese is involved, at least partially, in the sensorial diversity of Reblochon cheese and likely contributes to the production of non-uniform food.

The major yeasts present on the surface of five smear cheeses: Limburger (Germany), Reblochon (France), Livarot (France), Tilsit (Austria) and Gubbeen (Ireland) have been identified by FTIR spectroscopy during this project. It has been shown that the main yeast species occurring on the cheeses were Debaryomyces hansenii, Geotrichum candidum, Kluyveromyces lactis and/or Kluyveromyces marxianus, Yarrowia lipolytica and, for Gubbeen, Clavispora lusitaniae. Our objectives were to construct specific PCR primers for identification of the yeasts listed above and to test when appropriate their ability for in situ detection of the yeasts from total DNA from Livarot cheese. In order to assess the selectivity of PCR primers to be designed, thirty-six yeast strains representative of 17 yeast species were selected to be studied as belonging to the more frequently described yeasts in smear cheeses. They were obtained from different culture collections or were commercial cultures used in cheese processing. Livarot curd and cheeses were used for in situ validation of the primers. Alignments of nucleotide sequences were conducted using Basic Local Alignment Search Tool in order to localize nucleotide sequences suitable for the selective DNA amplification for all strains of a given yeast species. Primers designed in this study were compared to each other by aligning the published sequences from the DNA sequence libraries (DDBJ, EMBL, GenBank) and their specificity was evaluated using DNA extracted from the other yeast species. A couple of primers were designed leading to DNA amplification from 7 reference and commercial strains of Kluyveromyces lactis and K. marxianus. An other couple of PCR primers were designed leading to specific DNA amplification from five Y. lipolytica reference strains. We designed specific primers leading to a single band of approximately 200bp, when used in PCR, for one out of two collection strains of Debaryomyces hansenii and for a commercial strain. Using two other primers, a 200 bp PCR product was obtained for Clavispora lusitaniae collection strains CBS 4413T and CBS 52992. We were also able to design two sequences specific for Geotrichum candidum and validated on reference strains and two commercial strains. Using these primers, Kluyveromyces lactis/marxianus, Debaryomyces hansenii, Geotrichum candidum and Yarrowia lipolytica were detected in situ as well in Livarot curd than in ripened cheese and in frozen smears. Primers designed for C. lusitaniae were also tested in situ although this species was not identified in Livarot cheese. As expected, no DNA amplification occurred. To validate primers specificity from a taxonomic point of view, they should be tested with species whose phylogeny is closely related to the species we intended to detect. In accordance with the objectives of the SCM project, primers specificity was only checked in a context of smear cheese microbiology. However, this work constitutes a first stage towards a better knowledge of the principal yeast species on the surface of smear ripened cheese. Further application for those primers is their use in real time PCR in order to quantify yeast populations and to follow dynamics during ripening. These results will be disseminated by the mean of a scientific paper which is currently in preparation. No benefits are expected. End-users of these results are the scientific community in particular for ecological studies, but also producers of commercial yeasts and sectors of the food industry where the yeasts listed above are useful or often encountered as contaminants. In particular Geotrichum candidum is widely used in the cheese industry as ripening agent, but it is also responsible for the degradation of fresh cheeses and fruit and vegetable juices. It proliferates in refrigerated cakes and in frozen branch vegetables. It attacks fresh fruit from the orchard as well as warehoused fruit, in which it produces aqueous, viscous or slimy rot. The specific detection of this species could be interesting in quality control. To our knowledge, one probe has already been described to detect G. candidum but the use of primers is easier. One couple of primers was described which efficiency was checked in the context of medical studies. This is the first time, to our knowledge, that specific primers were designed for studies in food microbiology.

Members of the genera Arthrobacter, Staphylococcus and Microbacterium are some of the dominant genus that were identified on the surface of Tilsit, Livarot, Limburger, Reblochon and Gubbeen cheeses. Full length 16S rDNA sequences of Arthrobacter, Microbacterium and Staphylococci. were downloaded from the RDP (http://rdp.cme.msu.edu/hierarchy/hb_intro.jsp), EMBL and NCBI (http://www.ncbi.nlm.nih.gov/Taxonomy/taxonomyhome.html/) databases for alignment using the Clustal W and Clustal X programs (Thompson et al 1994, 1997). Strength of the alignment was also checked by other manual alignment programs, notably PHYDIT: Phylogenetic Molecular Sequence Editor (Chun, J. 1995. Eighty two representative Staphylococcus sequences, 66 Microbacterium sequences and 54 Arthrobacter sequences were independently aligned. Using these alignments 16S rDNA-targeted PCR primers were designed from the conserved regions. Primers were compared to the aligned SSU_rRNA database of the RDP using the CHECK_PROBE utility (Maidak et al, 1997). Previously published Arthrobacter and Microbacterium (Kolloffel et al, 1997) genus specific primers were tested for comparison. Primers were designed for members of all three genera. Phylogenetic trees were constructed to identify adjacent genera to test for non-specificity. Results from in silico analysis showed that the primers were specific for the relevant genus. These primers proved to be more specific than the previously published primers, which targeted genera like Agrococcus. Previously published Arthrobacter.Micrococcus specific primers also showed non-specificity with many other genera.

Anti-listerial yeasts: A screening method to test for anti-listerial activity in yeast was developed by Partner 7. The reproducibility of this method was confirmed by investigations done by Partners 1, 2, 3 and 7. As a test set 5 inhibitory and 5 non-inhibitory yeasts from previous investigations of Partner 7 and three strains each from Livarot, Tilsit, Gubbeen and Limburger cheese were chosen. The inhibitory potential of these 22 yeast strains was tested against six Listeria monocytogenes strains, WSLC 1001, 1364, 1685, 1416, 1211 and 1039. The anti-listerial activity was evaluated using a score-based system developed by Partner 7. Yeast strains with a strong and moderate anti-listerial activity retained this property. The order of the scores of all four partners were similar and Listeria strains showed differences in sensitivity. Under the best inhibitory yeasts were the top five yeasts from previous experiments done by Partner 7, followed by yeasts isolated from Livarot and Tilsit cheese. According to the results of the four partners the method was considered to be reproducible and reliable. Further screenings of yeasts from the different project cheeses were done by the individual partners. A total of 58 yeasts out of 1944 tested inhibited at least for of the selected L. monocytogenes indicator strains used in these investigations. Afterwards Partners 1, 2, 3 and 8 validated inhibitory yeasts by different methods. Partner 2 tested the 5 inhibitory yeasts from Partner 7 as well as 5 inhibitory yeasts from Tilsit against two L. monocytogenes indicator strains using a cheese agar. A clear inhibition was observed compared to the control. The other three partners used cheeses for their validation. The results obtained were different. Partner 1 observed inhibition at some time points but not at others. Investigations done in two replicates were not reproducible. Partner 3 obtained inhibition with some yeasts. Differences in individual experiments were most likely due to the concentrations of yeasts used. One yeasts was tested in detail but no inhibition was obtained. In investigations on cheeses of Partner 8 Listeria grew to a lesser extend with highly inhibitory yeasts than with less inhibitory yeasts. Possible explanations for the differences between the results on cheese agar and cheese were the level of yeasts used, the time of addition of Listeria and yeast, the contact between bacteria and yeast, the porosity of the cheese surface and the interactions between other surface microorganisms.

There was a steady progression of micro-organisms in each stage of ripening. In Limburger cheese the dominant groups identified were Arthrobacter arilaitensis and Brevibacterium linens at all stages of ripening from each batche. In Livarot cheese diversity is lost in the late stage of ripening from each of the three different plants. Most complex profiles were found in Reblochon cheese. Brevibacterium spp is found to dominante in Gubbeen cheese at all stages of ripening. Overall profiles in each cheese were different and more complex in the later stages of ripening. Different groups were identified by DGGE in each stage of ripening, a comparison of these groups was made with identifications of isolates.

DNA was extracted from 3mm surface areas of each of the cheeses. A nested PCR approach was used with actinobacterial specific primers (Stach et al, 2003) in order to amplify a 650 base pair fragment of the 16S rRNA gene. Total community DNA amplification was carried out with a reaction mixture consisting of 1µl of 20pmol/µl of S-C-Act-235-a-S-20 (5�-CGC GGC CTA TCA GCT TGT TG-3�), 1µl of 20 pmol/µl S-C-Act-878-a-A-19 (5�-CCG TAC TCC CCA GGC GGG G-3�) primers, 1µl of 1.25 pmol/µl of dNTP�s (Sigma), 5µl of 10X NH4 + buffer (Sigma), 1.5µl of 50mM of MgCl2 0.25µl of Taq polymerase (Sigma) and 1µl of genomic DNA. The final volume was made up to 50µl with sterile distilled water. DNA amplification were performed using a Perkin Elmer thermal cycler (Applied Biosystems, USA_ Amplification was done using touchdown PCR. PRC products were quantified by electrophoresis on a 1.5% (wt/vol) agarose gel stained with ethidium bromide. Parallel DGGE was performed essentially as described previously by Muyzer et al, 1993 using a Biorad Dcode universal mutation detection system TM. Band patterns were compared and specific bands cut out, amplified and sequenced. Sequences were assigned to genera and species on the basis of similarity to 16S sequences in the RDP database. Community composition was compared with the identification n of cheese isolates. Both uncultured and members belonging to the class Actinobacteria in any stages of ripening. In general, more groups were identified by DGGE when compared to the identificatio of cultivable strains. Some of the groups which were not identified could be attributed to the bias in DGGE or cultivation techniques. A number of different groups not previously isolated from cheese were detected by DGGE. All of the new taxa highlighted in studies of the culturablestrains were not detected by DGGE. It is clear from the data that both culture independent (DGGE) and culture independent (isolation) methods are needed to reveal the full extent of bacterial diversity on the surface of smear ripened cheeses.

Brevibacterium-primer Brevibacteria include species with both biotechnological and clinical significance. In biotechnology they contribute considerably to cheese ripening. They also have been encountered increasingly in humans as opportunistic pathogens and have been isolated from wounds, blood, an ear discharge, and from patients with otitis, osteomyelytis, peritonitis, septicemia, and groin abscess (Funke et al., 1997; Wauters et al., 2001). The purpose of this study was to develop genus specific primers for Brevibacterium spp. in order to identify members of the genus Brevibacterium by a genotypic method. 16S sequences of all Brevibacterium type strains, additional Brevibacterium strains and 16S sequences of closely related species (negative control-strains) were aligned. Regions of similarity in Brevibacterium species were identified visually ensuring that the corresponding regions in the other species were dissimilar. These conserved regions were used to design genus-specific primers. The primers gave a fragment of approximately 236bp. when a PCR was run on Brevibacterium spp. while they gave no band with a non-Brevibacterium strain. Brevibacteria belong to the suborder Micrococcinae. The identification of this suborder includes, among others, the determination of the type of peptidoglycan, menaquinones, and fatty acids in the cell wall. Some of these analyses are labour-intensive, which means that only few isolates can be analysed. However, accurate and reliable identification can be achieved by the 16S primers described in this study.

A database for yeasts was constructed as an Excel file. It contains data from FTIR measurements as well as results of phenotypic investigations of representatives selected from the dereplicated strains from the different cheeses. The biochemical and compositional analyses of the different cheeses are also included in the database. A completing dendrogram of representatives of all five cheeses (Gubbeen: 42 and 45, Limburger: 43, Livarot: 46, Reblochon: 51 and Tilsit: 76) was prepared to study the relationships between the flora of the different smears. There was only one cluster which contained yeast spectra from different cheeses, i.e., Livarot, Tilsit and Limburger. Most often the clusters consisted of spectra from one cheese or from two different cheeses. This indicates a certain individuality of each cheese`s yeast flora. A similarity between yeast spectra from Livarot and Tilsit cheese were noticed. Six common clusters existed, three assigned to G. geotrichum, two to Yarrowia lipolytica and one to Candida catenulata. Usually only one or two common clusters of spectra from yeasts from two different cheeses were formed. Interestingly, the yeast spectra from the two French cheeses showed only one common cluster existed. In representatives from Tilsit and Reblochon a high number of spectra, 47 and 29, respectively, was unclustered. A reason for this might be the generally high diversity within these isolates. This resulted in smaller clusters and a lot of unclustered spectra in the cheese specific dendrograms, which were the basis for dereplication. Contrary to this were the results obtained for Limburger cheese. Since these isolates appeared to be very homogeneous the clusters in the individual Limburger dendrogram which was used for dereplication contained a lot of spectra and therefore several representatives were chosen from each cluster. The number of representatives, which were taken from one cluster increased by cluster size. On the basis of this it is explainable that only a low number of unclustered spectra from Limburger yeasts occurred in the complete dendrogram. Clusters in that completing dendrogram which only contained spectra from Livarot strains were mainly assigned to Galactomyces geotrichum. Strains from this species on Livarot cheese showed a high homogeneity. In spectra from Gubbeen yeasts a lot of separate clusters were formed in the complete dendrogram. However, also a lot of unclustered spectra were noticed. The unclustered spectra most often belonged to yeasts isolated from the first three batches of Gubbeen cheese in which a higher diversity appeared. The dereplicated yeast strains of all five cheeses are stored as glycerol stocks at 80°C. Representatives of them will be freeze-dried and deposited in the Weihenstephan Yeast Collection.

Bacteria and yeasts from six batches of cheese collected over a five and a half year period were isolated after ripening for 4, 16, 23 and 37 d in the case of Batches 1, 2 and 3 and after 4, 10 and 16 d in the case of Batches 4, 5 and 6. The cheese was deliberately smeared with commercial cultures 2-3d after manufacture. FTIR spectroscopy was used to identify the yeast and PFGE, using sma or spe as restriction enzymes and rep-PCR in the case of the bacteria. A total of 1151 strains of bacteria were isolated and 1083 strains were identified. Eighteen species from 9 genera including Agrococcus, Arthrobacter, Brevibacterium, Corynebacteriun, Curtobacterium, Halomonas, Microbacterium, Psychrobacter and Staphylococcus were identified. Numerically, the most important organism was Staphylococcus saprophyticus (317 isolates), followed by Corynebacterium casei (257 isolates), Brevibacterium aurantiacum (190 isolates) and Corynebacterium variabile (134 isolates), each of which were isolated from every batch of cheese, and Microbacterium gubbeenense (55 isolates) which was isolated from cheese batches 1 to 4. The number of isolates of many of these species varied significantly between batches, e.g., B. aurantiacum dominated Batch 1 and only 1 isolate was found in Batch 2, 3 in Batch 4 and 2 in Batch 5; C. case dominated Batches 2 and 3 and very small numbers of isolates were found in Batch 1. This species also increased in number as ripening progressed in Batches 4 and 5. This data implies that each batch of cheese had a unique flora, which was not surprising given that this cheese is made at farmhouse level. Sporadic isolation of other species occurred, and one new species, Agrococcus sp. was isolated. Very few isolates (<10) from the commercial cultures used to smear the cheese were identified and these were only isolated at the initial stage of ripening, implying that the commercial cultures do not implant on the cheese. A progression of Gram-positive bacteria occurred during ripening with staphylococci dominating day 4 of ripening and coryneforms increasing as ripening progressed A total of 828 strains of yeast were isolated from the 6 batches of cheese and all were identified by FTIR spectroscopy. Debaryomyces hansenii was by far the most dominant yeast (75% of isolates), and was found in high numbers in all batches of cheese, except Batch 2; Candida catenulata (19% of isolates) was found in relatively high numbers in Batches 2 and 3 and much smaller numbers in Batches 1 and 4 while Clavispora lusitaniae(7.5% of isolates) was found in Batches 1 and 2 with 1 isolate in Batch 3. In contrast to the bacteria, no evidence of a progression of yeasts was found during ripening.

Three batches of Tilsit-cheese were produced in different dairies and deliberately smeared with 2 different commercial cultures of Brevibacterium aurantiacum and in one case with Debaryomyces hansenii following a defined protocol starting with the young cheeses. Five, 14 and 28 days after brining the smear of the cheeses was scraped off and approximately 20, apparently different colonies of each, bacteria and yeasts, were isolated from the smears. The bacterial flora was analysed using a polyphasic approach including denaturating gel electrophoresis (DGGE)of DNA directly extracted from the smear and characterisation of 185 single isolates by phenotypic methods, rep-PCR, 16S-rDNA sequencing, PAGE of whole cell proteins and Fourier Transform Infrared (FTIR) spectroscopy. Despite the application of single strain smear cultures only from the early phase of batch 3 the smear starter was highly recovered (94% of all isolates) whereas recovery in batch 1 was just 40% and it was not isolated from batch 2. Beside Brevibacterium aurantiacum Corynebacterium variabile dominated the early phase of batch 1 but their number decreased during ripening while Microbacterium spp. increased. Batch 2 was dominated by Staphylococci in the early and middle stage of ripening. This genus was also frequently isolated from the middle phase of batch 1 and 3. Generally the complexity of the bacterial population increased with ripening and coryneform bacteria dominated the late phase flora in all three batches. The yeast flora was investigated by the identification of 184 isolates by FTIR spectroscopy and phenotypic characterisation and appeared to very homogenous in all batches with Debaryomyces hansenii as the dominating species throughout ripening. During the late ripening stage approximately 20% of the isolates were pellet forming varieties of Geotrichum candidum. Batch 3 was unique by the isolation of Pichia triangularis and Candida tenuis from the early and middle ripening stage. Concerning the identity of the dominant taxa isolated from tilsit surface our results are in agreement with those from previous publications. However, contrasting the results of Bockelmann et al. different population dynamics were obtained from products of different dairies and Corynebacterium spp. (Bockelmann 50-90%) did not predominate the population of any ripened cheese. The different populations obtained from the different products seem to be determined by various factors, including lactate fermentation, technological parameters, the conditions in the ripening cells and the in-house flora. On the other hand it is well documented that problems in the smear development allow the growth of contaminating pathogens. This implies that by the maintenance of optimal conditions during cheese-making and ripening a relatively controlled smear development may be achieved thereby minimizing this risk. Moreover, our pilot scale cheese trials demonstrated that by the application of more complex smear cultures, including Arthrobacter arilaitensis, Corynebacterium spp., Microbacterium gubbeenense and Brachybacterium alimentarium beside Brevibacterium aurantiacum and Debaryomyces hansenii reduced the ripening time of Tilsit and positively influenced texture and flavour of the cheeses. Relevant publications: Posters: Bonaiti C., Chamba J. F., Cogan T., Desmasures N., Gelsomino R., Goodfellow M., Guéguen M., Irlinger F., Larpin S., Scherer S., Sebastiani H., Swings J., Thomas A., Vancanneyt M.: Biodiversity and anti listerial activity of surface microbial consortia from Limburger, Reblochon, Livarot, Tilsit and Gubbeen cheeses Congrilait, Paris 24.-27. September 2002. N.Bora, H.Sebastiani, R.Gelsomino, M.Vancanneyt, J.Swings, M.Goodfellow, A.C.Ward: Polyphasic study of actinomycetes on European bacterial smear ripened cheeses by culture dependent and culture independent methods. International symposium on biology of actinomycetes ISBA, Melbourne December 1-5th 2003 M.Hohenegger, N.Bora, R.Gelsomino, S.Goerges, M.Goodfellow, J.Swings, S.Scherer and H.Sebastiani: Surface Microflora of Tilsit-Cheese IDF-Symposium on Cheese, Prague March 21.-25., 2004 paper: M.Hohenegger, N.Bora, R.Gelsomino, S.Goerges, M.Goodfellow, J.Swings, S.Scherer and H.Sebastiani: Surface Microflora of Tilsit-Cheese Antonie van Leeuwenhoek 2004 (submitted for publication).

DNA from 1414 cheese bacteria isolated on Plate Count Agar (PCA) containing 3%NaCl (w/v/) were dereplicated by Rep-PCR. The patterns were analysed by Bionumerics and cluster analysis. Representatives of the Rep-PCR clusters were selected for further molecular characterisation and identification.

For the production of the Limburger cheese investigated a commercial smear starter culture containing Debaryomyces hansenii, Geotrichum candidum that is the anamorph form of G. geotrichum, Brevibacterium linens and Arthrobacter nicotianae was added to the cheese milk. In this study the latter two smear organisms were identified as the recently described species B. aurantiacum (Gavrish et al. 2004) and A. arilaitensis (Irlinger et al. 2004), respectively, by rep-PCR. FTIR spectroscopy and additional cluster analysis did not show similarity between the yeast smear starters and any of the 469 yeasts isolated from the nine Limburger cheeses of this study which were identified as D. hansenii or G. geotrichum. The same result was obtained for the bacterial smear starters. In a dendrogram containing the spectra of 420 coryneforms isolated from all Limburger project cheeses the spectra of the starter bacteria formed separate clusters although the coryneform isolates were also identified as B. aurantiacum and A. arilaitensis. Results for the bacteria were confirmed by rep-PCR and PFGE. In both cases the patterns of the bacterial smear starters were different when compared to these of the cheese isolates. Apparently, the yeast and bacterial starters did not succeed in the cheese environments. Gavrish E.Y., Krauzova V.I., Potekhina N.V., Karasev S.G., Plotnikova E.G., Altyntseva O.V., Korosteleva L.A. and Evtushenko L.I. 2004. Three new species of Brevibacteria, Brevibacterium antiquum sp. nov., Brevibacterium aurantiacum sp. nov., and Brevibacterium permense sp. nov. Microbiology. 73: 176-183. Irlinger F., Bimet F., Delettre J., Lefevre M. and Grimont P.A.D. 2004. Two new coryneform species isolated from the surface of French cheeses are species of the genus Arthrobacter: Arthrobacter bergerii sp. nov. and Arthrobacter arilaitensis sp. nov. submitted.

Two species of yeast, Debaryomyces hansenii and Galactomyces geotrichum, were found. D. hansenii dominated the flora with up to more than 90 % during the different stages of ripening in all three batches investigated, except for the early stage of Batch 2. In this case G. geotrichum dominated. While there was a slight decrease of D. hansenii during the ripening in the first and third batch this was reversed in the second batch. Within D. hansenii two different colony morphology types were found with type 2 occurring as a minority. The two types were distinguishable by their physiological reaction patterns. Furthermore their FTIR-clusters were clearly separated from each other. Generally, the yeast dendrogram resulting from the FTIR measurements consisted of 9 main clusters which were divided in 27 subclusters. Eighty-three % of the isolates identified as D. hansenii grouped in 6 of 17 subclusters and 66 % of G. geotrichum isolates were arranged in 3 of 10 subclusters. The remaining 17 and 34 %, respectively, appeared to be quite heterogeneous. The 469 yeasts were dereplicated to 43 isolates. Only two bacteria were found within the isolates of all three batches and both of these are newly described species, Brevibacterium aurantiacum (Gavrish et al. 2004) and Arthrobacter arilaitensis (Irlinger et al. 2004), A single rep-pattern was observed within A. arilaitensis and another one was found for all B. aurantiacum. A minority of isolates (n=5) was identified as Macrococcus sp. Furthermore, some Gram-negatives and bacilli occurred within the Catalase-positive flora. Within the early stages of Batch 1 and 2, B. aurantiacum dominated with up to about 80 %. In addition, Macrococcus sp. or A. arilaitensis occurred. On all the other cheeses of the different batches investigated A. arilaitensis dominanted. The proportion of B. aurantiacum generally ranged between 16 to 30 %. In the dendrogram depicting the similarity of 423 isolates according to their REP-pattern, 11 clusters became apparent which could be divided into 26 subclusters. 5 isolates were unclustered. Twenty-seven isolates failed to grow after freezing. Subsequently it was found that isolates identified as A. arilaitensis were arranged in 13 different subclusters and, in 4 of these, about 60 % grouped together. A total of 7 clusters belonged to B. aurantiacum, 4 of them contained around 80 % of the isolates. One cluster could be assigned to Macrococcus sp. and the remaining 5 clusters related to sporeformers, Gram-negatives and lactobacilli (LAB). On the basis of the clustering the 423 bacteria were dereplicated down to 43 isolates. The Limburger cheese investigated showed a low species diversity within the 469 yeast and 423 bacterial strains isolated. The intraspecies diversity of yeast and bacterial isolates reflected a high homogeneity as well. Only a small part of isolates identified as D. hansenii, G. geotrichum and A. arilaitensis, revealed a certain heterogeneity. Generally, an unexpected low biodiversity was observed on the Limburger cheese microorganisms. Gavrish E.Y., Krauzova V.I., Potekhina N.V., Karasev S.G., Plotnikova E.G., Altyntseva O.V., Korosteleva L.A. and Evtushenko L.I. 2004. Three new species of Brevibacteria, Brevibacterium antiquum sp. nov., Brevibacterium aurantiacum sp. nov., and Brevibacterium permense sp. nov. Microbiology. 73: 176-183. Irlinger F., Bimet F., Delettre J., Lefevre M. and Grimont P.A.D. 2004. Two new coryneform species isolated from the surface of French cheeses are species of the genus Arthrobacter: Arthrobacter bergerii sp. nov. and Arthrobacter arilaitensis sp. nov. submitted.

The microflora on the surface of Livarot cheese from three different dairies at three points during ripening was investigated. A total of 450 bacteria and 450 yeasts were isolated. Bacteria were identified phenotypically using Biotype 100 strips and dereplicated by rep-PCR. Representative strains were identified by BOX or (GTG)5-PCR. Yeast were identified by Fourier Transform-Infrared Spectroscopy. Physiological and molecular tests were done on representative yeast to confirm their identity. The bacterial flora consisted of 66% of Gram positive and 34 % of Gram negative isolates. The Gram positive bacteria belonged to Brevibacterium sp., Arthrobacter sp., Microbacterium sp., Corynebacterium sp. or Staphylococcus sp. and included the recently described species Brevibacterium aurantiacum, Arthrobacter arilaitensis and Microbacterium gubbeenense. In addition less common genera such as Brachybacterium sp. and Macrococcus sp. and new taxa close to Agrococcus sp. and Leucobacter sp. were also found. The diversity of the Gram negative flora was also high. Eleven different yeast species were found at the surface of the Livarot cheese. Geotrichum candidum dominated the yeast flora. Kluyveromyces lactis appeared only at the beginning of ripening and Yarrowia lipolytica was found later in the ripening. The dynamics of the dominant bacteria were different in the three cheeses since no spectrum of species was found in common at any time point. The dynamics of the dominant yeasts, at the species level, were simpler. A high degree of microbial diversity was observed on the surface of the Livarot cheese and each dairy had its own individual microbial flora. These results are going to be disseminated by the mean of a scientific paper, which is currently ready for submission. No benefits are expected. End-users of these results are the scientific community in particular dairy microbiologists, but also cheese producers and producers of commercial strains of microorganisms for the smear cheese industry.

Rezultaty

Udostępnij tę stronę Udostępnij tę stronę w mediach społecznościowych

Pobierz Pobierz zawartość strony