Registered designation of origin areas of fermented food products defined by microbial phenotypes and artificial neural networks

Lactobacillus panisapium sp. nov., from honeybee Apis cerana bee bread

A novel facultatively anaerobic, Gram-stain-positive, non-motile, non-spore-forming, catalase-negative bacterium of the genus Lactobacillus, designated strain Bb 2-3^T, was isolated from bee bread of Apis cerana collected from a hive in Kunming, China. The strain was regular rod-shaped. Optimal growth occurred at 37 °C, pH 6.5 with 5.0 g l⁻¹ NaCl. The predominant fatty acids were C_18 : 1ω9c, C_16 : 0 and C_19 : 0 iso. Respiratory quinones were not detected. Seven glycolipids, three lipids, phosphatidylglycerol and diphosphatidylglycerol were detected. The peptidoglycan type A4α l-Lys-d-Asp was determined. Strain Bb 2-3^T was closely related to Lactobacillus bombicola DSM 28793^T, Lactobacillus apis LMG 26964^T and Lactobacillus helsingborgensis DSM 26265^T, with 97.8, 97.6 and 97.0 % 16S rRNA gene sequence similarity, respectively. A comparison of two housekeeping genes, rpoA and pheS, revealed that strain Bb 2-3^T was well separated from the reference strains of species of the genus Lactobacillus. The average nucleotide identity between strain Bb 2-3^T and the type strains of closely related species was lower than the 95–96 % threshold value for delineation of genomic prokaryotic species. The G+C content of the genomic DNA of strain Bb 2-3^T was 37.4 mol%. On the basis of phenotypic, chemotaxonomic and phylogenetic analyses, strain Bb 2-3^T is proposed to represent a novel species of the genus Lactobacillus, for which we propose the name Lactobacillus panisapium sp. nov. The type strain is Bb 2-3^T (=DSM 102188^T=ACCC 19955^T).

The incongruent gelatinase genotype and phenotype in Enterococcus faecalis are due to shutting off the ability to respond to the gelatinase biosynthesis-activating pheromone (GBAP) quorum-sensing signal

The concomitant presence of a complete fsr quorum-sensing system and gelE-sprE operons in Enterococcus faecalis is known to be essential for the detection of gelatinase activity. However, there are reports of the absence of gelatinase activity despite the presence of complete fsr and gelE loci. In order to understand this incongruence between genotype and phenotype we sequenced fsr and gelE loci of the E. faecalis LN68 strain, which was previously found to carry both operons but to lack gelatinase activity. Of the 59 nucleotide differences detected compared with the gelatinase-positive V583 strain, we found a nonsense mutation (a premature STOP codon) predicted to truncate the ATPase sensor domain of the FsrC protein, responsible for sensing and transducing the signal from the quorum-sensing molecule. Strain LN68 was highly affected in the expression of the gelE and sprE genes, further supporting the lack of Fsr-dependent gelE induction. When we constructed a V583 mutant with the same premature stop mutation in the fsrC gene the resulting strain was no longer able to degrade gelatin. We conclude that the reduced ability to transduce the quorum-sensing signal of the prematurely truncated FsrC protein is sufficient to explain the negative gelatinase phenotype. As the incongruent genotype and phenotype is detected in natural isolates, we believe that the silencing of the quorum-sensing system Fsr may be beneficial for some E. faecalis strains.

Enterococci from artisanal dairy products show high levels of adaptability

Enterococci are ubiquitous organisms able to promote both health (fermented food/probiotics) and illness (human/animal infections). Disturbingly, several enterococcal species commonly found in artisanal cheeses, such as Enterococcus faecalis and E. faecium, are being increasingly established as causes of infection, posing a problem for food safety. In this study enterococci from ewe's milk and cheese were compared to clinical and reference strains by growth in media simulating environmental colonization and infection sites: 2YT, BHI, skim milk, urine and rabbit serum at different pHs, NaCl concentrations and temperatures. Growth curves were obtained with Microbiology Workstation Bioscreen C and used to calculate relative indexes--RIs--(based on absorbance, lag phase and specific growth rate) for each strain and environmental condition. Similar or higher RIs were obtained for food strains growing in infection-related environments when compared to clinical ones, revealing their ability to adapt and grow in these conditions. A dendrogram built using Pearson's correlation coefficient and a PCA analysis clustered the strains regardless of their origin or species allocation, suggesting a strain-specific mode of growth and a high environmental adaptability of enterococcal strains. These evidences turn essential the evaluation of strains to be used as starters or probiotics.

Use of artificial neural networks and a gamma-concept-based approach to model growth of and bacteriocin production by Streptococcus macedonicus ACA-DC 198 under simulated conditions of Kasseri cheese production

Growth of and bacteriocin production by Streptococcus macedonicus ACA-DC 198 were assessed and modeled under conditions simulating Kasseri cheese production. Controlled fermentations were performed in milk supplemented with yeast extract at different combinations of temperature (25, 40, and 55 degrees C), constant pH (pHs 5 and 6), and added NaCl (at concentrations of 0, 2, and 4%, wt/vol). The data obtained were used to construct two types of predictive models, namely, a modeling approach based on the gamma concept, as well as a model based on artificial neural networks (ANNs). The latter computational methods were used on 36 control fermentations to quantify the complex relationships between the conditions applied (temperature, pH, and NaCl) and population behavior and to calculate the associated biokinetic parameters, i.e., maximum specific growth and cell count decrease rates and specific bacteriocin production. The functions obtained were able to estimate these biokinetic parameters for four validation fermentation experiments and obtained good agreement between modeled and experimental values. Overall, these experiments show that both methods can be successfully used to unravel complex kinetic patterns within biological data of this kind and to predict population kinetics. Whereas ANNs yield a better correlation between experimental and predicted results, the gamma-concept-based model is more suitable for biological interpretation. Also, while the gamma-concept-based model has not been designed for modeling of other biokinetic parameters than the specific growth rate, ANNs are able to deal with any parameter of relevance, including specific bacteriocin production.

Vancomycin-susceptible dairy and clinical enterococcal isolates carry vanA and vanB genes

A total of 109 enterococcal isolates from dairy food products and from human and dog infections, isolated in Portugal, and 26 type and reference strains of the genus Enterococcus were screened for vancomycin resistance. MIC values, both for vancomycin and teicoplanin, were determined. The genetic relatedness of isolates carrying either vanA and/or vanB was determined using Pulsed Field Gel Electrophoresis. For vanA carrying isolates, transposon Tn1546 was partially mapped using PCR. None of the 59 dairy isolates was resistant to vancomycin. Among the 50 clinical isolates, only one, carrying vanB, behaved as resistant, with a MIC value of 256 microg/mL. The type and reference strains used were susceptible both to vancomycin and teicoplanin. vanA was found in 37% of the dairy isolates and 40% of the clinical isolates. vanB was only detected in 18% of the clinical, both human and dog, isolates. PCR partial mapping of Tn1546 revealed 23 different patterns among 42 isolates. Some patterns were shared between dairy and clinical isolates. Using Pulsed Field Gel Electrophoresis six groups of isolates were found to be genetically undistinguishable and grouping was found to be geographically and location specific/related. No genetic relatedness was found between isolates from dairy, human and veterinary sources. These results show that an incomplete and/or unfunctional Tn1546 element may explain the absence of resistant behaviour in the studied isolates, even when vanA gene is present. Moreover, the work reported shows that both clinical (human and animal) and dairy isolates have been in contact with VanA genotype of resistance and suggest that dissemination of vanA gene has been through transposable elements, like Tn1546, and not by clonal dissemination of a resistant strain. Therefore, a national strategy should be implemented to survey both vancomycin resistance and its genetic dissemination.

Activity and expression of a virulence factor, gelatinase, in dairy enterococci

In order to understand the virulence potential of dairy enterococci, 35 isolates from raw ewe's milk and traditionally fermented cheeses, identified as Enterococcus faecalis, Enterococcus faecium, Enterococcus durans, Enterococcus dispar and Enterococcus hirae, were screened for their capacity to produce gelatinase and for the presence of the genes gelE, sprE, fsrA, fsrB and fsrC. Studies correlating gelatinase production with maintenance and subculture of the isolates in the Laboratory environment, and growth in different media were performed. These studies were conducted with two dairy isolates identified as E. faecalis and E. durans, and one clinical isolate, E. faecalis OG1-10. RT-PCR was used for detection of transcripts of the above mentioned genes. Results demonstrated that the virulence factor gelatinase is disseminated among the genus Enterococcus and that dairy isolates are capable of producing gelatinase at comparable levels with clinical isolates, although this capacity is easily lost during conservation by freezing in the laboratory. Therefore, gelatinase production potential of dairy enterococci may be underestimated. The gene gelE was present in all studied isolates. The same was observed for the fsr operon, either complete or incomplete, revealing that the gelatinase genetic determinants, so far only described in E. faecalis, are a common trait in the genus. This work describes for the first time the detection of the complete Fsr-GelE operon in other species than E. faecalis, namely E. faecium and E. durans. The loss of expression of this virulence factor under laboratory culture conditions correlated with the loss of one or more genes of the regulatory fsr operon, although the gene gelE was maintained, demonstrating that a complete fsr operon is required for a positive GelE phenotype. Independent of the detection of any gelatinase activity, if both gelE and the complete fsr operon are present in the cell, all genes are transcribed, as revealed by RT-PCR, suggesting that regulation of gelatinase activity can also be post-transcriptional. The silent behavior of gelE was only observed in E. faecalis, but not in E. durans, suggesting different modulation mechanisms of gelatinase activity in these two species. Overall, these findings reopen the issue of food safety of enterococci and reinforce the need to further study the mechanisms responsible for triggering the virulence factor gelatinase in non-pathogenic enterococcal environmental isolates.

Use of unsupervised and supervised artificial neural networks for the identification of lactic acid bacteria on the basis of SDS-PAGE patterns of whole cell proteins

Conventional multivariate statistical techniques (hierarchical cluster analysis, linear discriminant analysis) and unsupervised (Kohonen Self Organizing Map) and supervised (Bayesian network) artificial neural networks were compared for as tools for the classification and identification of 352 SDS-PAGE patterns of whole cell proteins of lactic acid bacteria belonging to 22 species of the genera Lactobacillus, Leuconostoc, Enterococcus, Lactococcus and Streptococcus including 47 reference strains. Electrophoretic data were pre-treated using the logistic weighting function described by Piraino et al. [Piraino, P., Ricciardi, A., Lanorte, M. T., Malkhazova, I., Parente, E., 2002. A new procedure for data reduction in electrophoretic fingerprints of whole-cell proteins. Biotechnol. Lett. 24, 1477-1482]. Hierarchical cluster analysis provided a satisfactory classification of the patterns but was unable to discriminate some species (Leuconostoc, Lb. sakei/Lb. curvatus, Lb. acidophilus/Lb. helveticus, Lb. plantarum/Lb. paraplantarum, Lc. lactis/Lc. raffinolactis). A 7x7 Kohonen self-organizing map (KSOM), trained with the patterns of the reference strains, provided a satisfactory classification of the patterns and was able to discriminate more species than hierarchical cluster analysis. The map was used in predictive mode to identify unknown strains and provided results which in 85.5% of cases matched the classification obtained by hierarchical cluster analysis. Two supervised tools, linear discriminant analysis and a 23:5:2 Bayesian network were proven to be highly effective in the discrimination of SDS-PAGE patterns of Lc. lactis from those of other species. We conclude that data reduction by logistic weighting coupled to traditional multivariate statistical analysis or artificial neural networks provide an effective tool for the classification and identification of lactic acid bacteria on the basis of SDS-PAGE patterns of whole cell proteins.

Authenticity Assessment of Dairy Products

The authenticity of dairy products has become a focal point, attracting the attention of scientists, producers, consumers, and policymakers. Among many others, some of the practices not allowed in milk and milk products are the substitution of part of the fat or proteins, admixtures of milk of different species, additions of low-cost dairy products (mainly whey derivatives), or mislabeling of products protected by denomination of origin. A range of analytical methods to detect frauds have been developed, modified, and continually reassessed to be one step ahead of manufacturers who pursue these illegal activities. Traditional procedures to assess the authenticity of dairy products include chromatographic, electrophoretic, and immunoenzymatic methods. New approaches such as capillary electrophoresis, polymerase chain reaction, and isotope ratio mass spectrometry have also emerged alongside the latest developments in the former procedures. This work intends to provide an updated and extensive overview since 1991 on the principal applications of all these techniques together with their advantages and disadvantages for detecting the authenticity of dairy products. The scope and limits of different tools are also discussed.

Antimicrobial resistance profiles of dairy and clinical isolates and type strains of enterococci

The susceptibility to 30 antimicrobial agents was determined by the disk diffusion method for a collection of 172 enterococcal strains, including 96 isolates from dairy sources, 50 isolates of human and veterinary origin, and 26 reference strains from 24 different enterococcal species. Results were analysed by hierarchic numerical methods to cluster strains and to group antimicrobials according to similarity profiles. Resistance to 17 of the 30 antimicrobials showed to be correlated, leading to four groups reflecting the mode of action: quinolones (ofloxacin, enrofloxacin, ciprofloxacin and norfloxacin); macrolides (erythromycin, spiramycin), phenicols (cloramphenicol) and tetracyclins (tetracycline, oxytetracyclin); aminoglycosides (gentamicin, kanamycin) and lincosamides (clindamycin); penicillins (amoxicillin, ampicillin, penicillin G, piperacillin) and carbapenems (imipenem). Overall, the genus Enterococcus behaved as resistant to lincomycin, colistin, polimixin B and, with a few exceptions in dairy isolates, to methicillin. In general, all isolates were susceptible to vancomycin, cloramphenicol and fusidic acid. Clusters containing only dairy isolates were susceptible to the majority of antimicrobials tested, as opposed to clusters constituted only by clinical enterococcal isolates. Among the clinical isolates, 62% were highly multiresistant. Low level gentamicin resistance was found to be associated with clinical enterococci. Among dairy isolates, those that clustered with clinical isolates were both resistant to gentamicin and identified as Enterococcus faecalis. Resistance to macrolides, quinolones, penicillins and imipenem was found to be associated also with clinical environments, mainly with multiresistant isolates, contrary to what is generally agreed as a characteristic of the genus. Veterinary clinical isolates were mainly grouped with the multiresistant clinical human isolates. The 26 reference enterococcal strains were distributed in clusters with different antibiotic resistance profiles and were mainly clustered with dairy isolates.

Microbiological characterization of artisanal Montasio cheese: analysis of its indigenous lactic acid bacteria

The aim of this study was to investigate the dynamics of the microflora during Montasio cheese ripening, with specific reference to some characteristics of biotechnological interest. Nine batches of Montasio cheese produced in different plants were analyzed. Streptococcus thermophilus was the predominant species throughout the whole ripening period of Montasio cheese. Enterococci were also frequently present. This microbial group resulted probably from milk, and its proportion decreased rapidly during ripening. The most acidifying microbial species was S. thermophilus, while the most proteolytic strains belonged to the genera Enterococcus. A high degree of phenotypic diversity occurred within the microbial species.

Comparative study using type strains and clinical and food isolates to examine hemolytic activity and occurrence of the cyl operon in enterococci

The hemolytic ability, the presence of cyl genes, and the diagnostic accuracy of cytolysin molecular detection were investigated in the genus Enterococcus by using 164 strains from 20 different species (26 reference strains, 42 clinical isolates from human and veterinary origin, and 96 isolates from ewe cheese and milk). Hemolysis was assayed with sheep and horse erythrocytes and under aerobic or anaerobic conditions. Screening of cytolysin genes (cylL(L), cylL(S), cylM, cylB, and cylA) was performed with new specific primers and the anaerobic assay of beta-hemolysis was used as the "gold standard" for the evaluation of cyl gene-based PCRs. Since beta-hemolysis and cyl genes were found in 10 and 14 species, respectively, the hemolytic ability seems to be spread throughout the genus ENTEROCOCCUS: Beta-hemolysis was observed in 6 of 26 (23%) reference strains, 14 of 42 (33%) clinical isolates, and 6 of 96 (6%) food isolates. The presence of cyl genes was detected in 15 of 26 (58%) reference strains, 37 of 42 (88%) clinical isolates, and 67 of 96 (70%) food isolates. These data indicate a virulence potential in food isolates, reinforcing the need of their safety assessment. Analysis of phenotypic-genotypic congruence suggests a divergent sequence evolution of cyl genes and the effect of environmental factors in the regulation of cytolysin expression. Evaluation of the diagnostic accuracy of cytolysin molecular detection points to cylL(L)-based PCR and cylL(L)L(S)MBA-based PCR as the most reliable approaches. Nevertheless, the low sensitivity (46%) and gene variability indicated by our study strongly recommend the phenotypic assay for the assessment of hemolytic ability in enterococci, followed by the molecular screening of cyl genes in nonhemolytic strains to evaluate their virulence potential.

Virulence factors in food, clinical and reference Enterococci: A common trait in the genus?

The occurrence of several virulence traits (cytolysin, adhesins and hydrolytic enzymes) was investigated in a collection of 164 enterococci, including food and clinical isolates (from human and veterinary origin), as well as type and reference strains from 20 enterococcal species. Up to fifteen different cyl genotypes were found, as well as silent cyl genes. The occurrence of the cyl operon and haemolytic potential seems to be widespread in the genus. A significant association of this virulent trait with clinical isolates was found (p < 0.05). High levels of incidence were also observed for genes encoding surface adhesins (esp, efaA(fs), efaA(fm)), agg and gelE, irrespectively of species allocation and origin of strains. Although gelE behaves as silent in the majority of the strains, gelatinase activity predominates in clinical isolates, whereas lipase and DNase were mainly detected in food isolates pointing to their minor role as virulence determinants. No hyaluronidase activity was detected for all strains. Numerical hierarchic data analysis grouped the strains in three main clusters, two of them including a total of 50 strains with low number of virulence determinants (from 2 to 7) and the other with 114 strains with a high virulence potential (up to 12 determinants). No statistical association was found between virulence clusters and species allocation (p > 0.10), strongly suggesting that virulence determinants are a common trait in the genus Enterococcus. Clinical strains seem to be significantly associated with high virulence potential, whereas food, commensal and environmental strains harbour fewer virulence determinants (p < 0.01). A high level of relative diversity in virulence patterns was observed (Shannon's index varies from 0.95 to 1.0 among clusters), reinforcing the strain-specific nature of the association of virulence factors. Although a low risk seems to be associated with the use of enterococci in long-established artisanal cheeses, screening of virulence traits and their cross-synergies must be performed, particularly for commercial starters, probiotic strains and products to be used by high risk population groups.

A statistical procedure for the analysis of microbial communities based on phenotypic properties of isolates

A novel statistical procedure for the analysis of microbial communities based on phenotypic properties of randomly collected isolates is presented and discussed. The procedure allows the representation of the microbial communities as a set of ellipses in a bidimensional graph. This representation is obtained by the following steps: (a) measurement of a set of binary phenotypic properties for n isolates belonging to k samples, each representing a different community; (b) repeated sampling by bootstrapping of the m samples, thus obtaining, for each community, i subsamples of j isolates; (c) calculation of the frequency of positive results for each test for each subsample; (d) calculation of the matrix of Euclidean distances between the k x i frequency vectors; (e) use of multidimensional scaling (MDS) to obtain a representation in two dimensions of the distance relationships between the frequency vectors; (f) plotting of the 95% confidence ellipses for the i frequency vectors of each of the k communities. By using both simple, synthetic microbial communities, and samples of lactic acid bacteria isolated from natural microbial communities (sourdoughs, compressed yeast, fermented sausages), it was demonstrated that the position and shape of the ellipses are clearly related to the composition of the community, while the relationship between the size of the ellipses and the phenotypical diversity of the community is less straightforward: while communities with very different diversity (measured with the Functional Evenness index and the mean taxonomic distance) had ellipses that were very different in size, there was no strict proportionality between the size of the ellipse and the diversity of the community. Nevertheless, the representation of microbial communities obtained by bootstrapping and multidimensional scaling appears to be superior to the more usual representation based on tabulation of the frequencies of isolates belonging to different clusters.

Predictive non-linear modeling of complex data by artificial neural networks

An artificial neural network (ANN) is an artificial intelligence tool that identifies arbitrary nonlinear multiparametric discriminant functions directly from experimental data. The use of ANNs has gained increasing popularity for applications where a mechanistic description of the dependency between dependent and independent variables is either unknown or very complex. This machine learning technique can be roughly described as a universal algebraic function that will distinguish signal from noise directly from experimental data. The application of ANNs to complex relationships makes them highly attractive for the study of biological systems. Recent applications include the analysis of expression profiles and genomic and proteomic sequences.

Comparison of statistical methods for identification of Streptococcus thermophilus, Enterococcus faecalis, and Enterococcus faecium from randomly amplified polymorphic DNA patterns

Thermophilic streptococci play an important role in the manufacture of many European cheeses, and a rapid and reliable method for their identification is needed. Randomly amplified polymorphic DNA (RAPD) PCR (RAPD-PCR) with two different primers coupled to hierarchical cluster analysis has proven to be a powerful tool for the classification and typing of Streptococcus thermophilus, Enterococcus faecium, and Enterococcus faecalis (G. Moschetti, G. Blaiotta, M. Aponte, P. Catzeddu, F. Villani, P. Deiana, and S. Coppola, J. Appl. Microbiol. 85:25-36, 1998). In order to develop a fast and inexpensive method for the identification of thermophilic streptococci, RAPD-PCR patterns were generated with a single primer (XD9), and the results were analyzed using artificial neural networks (Multilayer Perceptron, Radial Basis Function network, and Bayesian network) and multivariate statistical techniques (cluster analysis, linear discriminant analysis, and classification trees). Cluster analysis allowed the identification of S. thermophilus but not of enterococci. A Bayesian network proved to be more effective than a Multilayer Perceptron or a Radial Basis Function network for the identification of S. thermophilus, E. faecium, and E. faecalis using simplified RAPD-PCR patterns (obtained by summing the bands in selected areas of the patterns). The Bayesian network also significantly outperformed two multivariate statistical techniques (linear discriminant analysis and classification trees) and proved to be less sensitive to the size of the training set and more robust in the response to patterns belonging to unknown species.

Modeling of pH and acidity for industrial cheese production

A three-layer feedforward neural network was successfully used to model and predict the pH of cheese curd at various stages during the cheese-making process. An extended database, containing more than 1800 vats over 3 yr of production of Cheddar cheese with eight different starters, from a large cheese plant was used for model development and parameter estimation. Neural network models were developed with inputs selected among 33 quantitative and qualitative process variables for final pH of cheese, pH at cutting, and acidity at whey drawing-off and at pressing. In all cases, very high correlation coefficients, ranging from 0.853 to 0.926, were obtained with the validation data. A sensitivity analysis of neural network models allowed the relative importance of each input process variable to be identified. The sensitivity analysis in conjunction with a priori knowledge permitted a significant reduction in the size of the model input vector. A neural network model using only nine input process variables was able to predict the final pH of cheese with the same accuracy as for the complete model with 33 original input variables. This significant decrease in the size of neural networks is important for applications of process control in cheese manufacturing.