Sébastien Theil

Cheeses are fermented dairy products consumed worldwide. Their global diversity results from various local variables, including technological practices, as well as the metabolic activity of diverse microorganisms. In Europe, this typicity is exemplified by Protected Designation of Origin (PDO) cheeses, for which genetic diversity remains largely unexplored. Combining culturomics (n = 373 bacterial genomes) and metagenomic (n = 146 metagenomes), we performed a national-scale survey of the microbial diversity encompassing 44 French PDO cheeses. Taxonomic (bacteria, fungi and viruses) and functional profiling reveal a high diversity in the cheese rind, mainly driven by the cheese technology. We also reconstructed 1,119 bacterial metagenome-assembled genomes (MAGs) encompassing seven phyla, including Actinomycetota, Bacillota, Pseudomonadota and Bacteroidota. Using GTDB as a reference, we identified 221 MAGs encompassing 46 genera, as well as 44 bacterial isolate genomes encompassing eight genera, which represent potentially 81 new species (based on <95% ANI). These species were particularly numerous among the genera Halomonas, Psychrobacter and Brachybacterium. Similar results were observed when compared with the cFMD database. We combined our genomic and metagenomic datasets into a catalog of 26.2 million protein clusters, with 50% of these clusters remaining unassigned to a known function and taxonomy. We illustrated the potential of this resource by searching for methionine gamma-lyase (MGL), an enzyme playing a significant role in cheese flavor. This protein was predominantly found in Pseudoalteromonas, a potentially new MGL-producing genus, Serratia, Pseudomonas, Proteus and Hafnia, and its prevalence varied with cheese technology. Our study provides a substantial genomic resource for food microbiologists and cheesemakers to further explore the biotechnological potential of PDO cheese biodiversity.

Microbial communities offer considerable potential for tackling environmental challenges by improving the functioning of ecosystems. Top-down community engineering is a promising strategy that could be used to obtain communities of desired function. However, the ecological factors that control the balance between community shaping and propagation are not well understood. Dairy backslopping, which consists of using part of the previous production to inoculate a new one, can be used as a model engineering approach to investigate community dynamics during serial propagations. In this study, 26 raw milk samples were serially propagated 6 times each, giving rise to 26 community lineages. Bacterial community structures were analyzed by metabarcoding, and acidification was recorded by pH monitoring. The results revealed that different types of community lineages could be obtained in terms of taxonomic composition and dynamics. Five lineages reached a repeatable community structure in a few propagation steps, with little variation between the final generations, giving rise to stable acidification kinetics. Moreover, these stabilized communities presented a high variability of structure and diverse acidification properties between community lineages. Besides, the other lineages were characterized by different levels of dynamics leading to parallel or divergent trajectories. The functional properties and dynamics of the communities were mainly related to the relative abundance and the taxonomic composition of lactic acid bacteria within the communities. These findings highlight that short-term schemes of serial fermentation can produce communities with a wide range of dynamics and that the balance between community shaping and propagation is intimately linked to community structure. IMPORTANCE: Microbiome applications require approaches for shaping and propagating microbial communities. Shaping allows the selection of communities with desired taxonomic and functional properties, while propagation allows the production of the biomass required to inoculate the engineered communities in the target ecosystem. In top-down community engineering, where communities are obtained from a pool of mixed microorganisms by acting on environmental variables, a major challenge is to master the balance between shaping and propagation. However, the ecological factors that favor high dynamics of community structure and, conversely, those that favor stability during propagation are not well understood. In this work, short-term dairy backslopping was used to investigate the key role of the taxonomic composition and structure of bacterial communities on their dynamics. The results obtained open up interesting prospects for the biotechnological use of microbiomes, particularly in the field of dairy fermentation, to diversify approaches for injecting microbial biodiversity into cheesemaking processes.

Investigating the diversity of a given species could give clues for the development of autochthonous starter cultures. However, few studies have focused on the intraspecies diversity of Lactobacillus delbrueckii strains, a technologically important lactic acid bacterium for the dairy industry. For this reason, Lactobacillus delbrueckii strains from the Saint-Nectaire Protected Designation of Origin (PDO) area were isolated and characterized. Genetic diversity was determined based on core genome phylogenetic reconstruction and pangenome analysis, while phenotypic assessments encompassed proteolysis and volatile compound production potential. A total of 15 L. delbrueckii ssp. lactis unique new strains were obtained. The genetic analysis and further proteolytic activities measurement revealed low variability among these Saint-Nectaire strains, while substantial genetic variability was observed within the L. delbrueckii ssp. lactis subspecies as a whole. The volatile compound profiles slightly differed among strains, and some strains produced volatile compounds that could be of particular interest for cheese flavor development. While the genetic diversity among Saint-Nectaire strains was relatively modest compared to overall subspecies diversity, their distinct characteristics and pronounced differentiation from publicly available genomes position them as promising candidates for developing autochthonous starter cultures for cheese production.

Streptococcus thermophilus is of major importance for cheese manufacturing to ensure rapid acidification; however, studies indicate that intensive use of commercial strains leads to the loss of typical characteristics of the products. To strengthen the link between the product and its geographical area and improve the sensory qualities of cheeses, cheese-producing protected designations of origin (PDO) are increasingly interested in the development of specific autochthonous starter cultures. The present study is therefore investigating the genetic and functional diversity of S. thermophilus strains isolated from a local cheese-producing PDO area. Putative S. thermophilus isolates were isolated and identified from milk collected in the Saint-Nectaire cheese-producing PDO area and from commercial starters. Whole genomes of isolates were sequenced, and a comparative analysis based on their pan-genome was carried out. Important functional properties were studied, including acidifying and proteolytic activities. Twenty-two isolates representative of the diversity of the geographical area and four commercial strains were selected for comparison. The resulting phylogenetic trees do not correspond to the geographical distribution of isolates. The clustering based on the pan-genome analysis indicates that isolates are divided into five distinct groups. A Kyoto Encyclopedia of Genes and Genomes (KEGG) functional annotation of the accessory genes indicates that the accessory gene contents of isolates are involved in different functional categories. High variability in acidifying activities and less diversity in proteolytic activities were also observed. These results indicate that high genetic and functional variabilities of the species S. thermophilus may arise from a small (1,800 km²) geographical area and may be exploited to meet demand for use as autochthonous starters.

High-throughput sequencing of two lettuces showing virus-like symptoms in France provided evidence of infection by members of the family Secoviridae. One plant (JG1) had a complex mixed infection that involved, among others, a novel waikavirus (lettuce waikavirus 1) and two isolates of a sequivirus related to lettuce mottle virus (LeMoV). The second lettuce plant (JG2) was singly infected by LeMoV. Complete genomic sequences were obtained for all four isolates and, in addition, near complete genome sequences were obtained for other LeMoV or LeMoV-related isolates (from French cultivated and wild lettuces and from a Brazilian cultivated lettuce) and for two isolates of another family Asteraceae-infecting sequivirus, dandelion yellow mosaic virus (DaYMV). Analysis of these genomic sequences allows the proposal of tentative genome organization for the various viruses and clarification of their phylogenetic relationships. Sequence and host range comparisons point to significant differences between the two sequivirus isolates identified in the JG1 plant and LeMoV isolates from France and Brazil, suggesting they belong to a novel species for which the name lettuce star mosaic virus is proposed.

The supplementation of animal feed with microbial additives remains questioning for the traditional or quality label raw milk cheeses with regard to microbial transfer to milk. We evaluated the effect of dietary administration of live yeast on performance and microbiota of raw milk, teat skin, and bedding material of dairy cows. Two balanced groups of cows (21 primiparous 114 ± 24 DIM, 18 multiparous 115 ± 33 DIM) received either a concentrate supplemented with Saccharomyces cerevisiae CNCM I-1077 (1 × 10¹⁰ CFU/d) during four months (LY group) or no live yeast (C group). The microbiota in individual milk samples, teat skins, and bedding material were analysed using culture dependent techniques and high-throughput amplicon sequencing. The live yeast supplementation showed a numerical increase on body weight over the experiment and there was a tendency for higher milk yield for LY group. A sequence with 100% identity to that of the live yeast was sporadically found in fungal amplicon datasets of teat skin and bedding material but never detected in milk samples. The bedding material and teat skin from LY group presented a higher abundance of Pichia kudriavzevii reaching 53% (p < 0.05) and 10% (p < 0.05) respectively. A significant proportion of bacterial and fungal ASVs shared between the teat skin and the milk of the corresponding individual was highlighted.

In dairy cattle, teat disinfection at the end of milking is commonly applied to limit colonization of the milk by pathogenic microorganisms via the teat canal. The post-milking products used can irritate the teat skin and unbalance its microbial population. Our study aimed to assess the impact of different milking products on the balance of the microbial communities on the teat skin of cows and in their milk. For 12 weeks at the end of each milking operation, three groups of seven Holstein dairy cows on pasture received either a chlorhexidine gluconate-based product (G) or a hydrocolloidal water-in-oil emulsion (A), or no post-milking product (C). The composition of the bacterial and fungal communities on the teat skin and in the milk were characterized using a culture-dependent method and by high-throughput sequencing of marker genes to obtain amplicon sequence variants (ASVs). The individual microbiota on the cows’ teat skin was compared for the first time to that of a cow pool. In contrast to the milk, the post-milking treatment influenced the microbiota of the teat skin, which revealed a high microbial diversity. The water-in-oil emulsion appeared to slightly favour lactic acid bacteria and yeasts and to limit the development of undesirable bacteria such as Pseudomonas and Staphylococcus.

A previous study identified differences in rind aspects between Cantal-type cheeses manufactured from the same skimmed milk, supplemented with cream derived either from pasture-raised cows (P) or from cows fed with maize silage (M). Using an integrated analysis of multiomic data, the present study aimed at investigating potential correlations between cream origin and metagenomic, lipidomic and volatolomic profiles of these Cantal cheeses. Fungal and bacterial communities of cheese cores and rinds were characterized using DNA metabarcoding at different ripening times. Lipidome and volatolome were obtained from the previous study at the end of ripening. Rind microbial communities, especially fungal communities, were influenced by cream origin. Among bacteria, Brachybacterium were more abundant in P-derived cheeses than in M-derived cheeses after 90 and 150 days of ripening. Sporendonema casei, a yeast added as a ripening starter during Cantal manufacture, which contributes to rind typical aspect, had a lower relative abundance in P-derived cheeses after 150 days of ripening. Relative abundance of this fungus was highly negatively correlated with concentrations of C18 polyunsaturated fatty acids and to concentrations of particular volatile organic compounds, including 1-pentanol and 3-methyl-2-pentanol. Overall, these results evidenced original interactions between milk fat composition and the development of fungal communities in cheeses.

Adding massive amounts of lactic starters to raw milk to manage the sanitary risk in the cheese-making process could be detrimental to microbial diversity. Adjusting the amount of the lactic starter used could be a key to manage these adverse impacts. In uncooked pressed cheeses, we investigated the impacts of varying the doses of a lactic starter (the recommended one, 1×, a 0.1× lower and a 2× higher) on acidification, growth of Staphylococcus aureus SA15 and Shiga-toxin-producing Escherichia coli (STEC) O26:H11 F43368, as well as on the bacterial community patterns. We observed a delayed acidification and an increase in the levels of pathogens with the 0.1× dose. This dose was associated with increased richness and evenness of cheese bacterial community and higher relative abundance of potential opportunistic bacteria or desirable species involved in cheese production. No effect of the increased lactic starter dose was observed. Given that sanitary criteria were paramount to our study, the increase in the pathogen levels observed at the 0.1× dose justified proscribing such a reduction in the tested cheese-making process. Despite this, the effects of adjusting the lactic starter dose on the balance of microbial populations of potential interest for cheese production deserve an in-depth evaluation.

Bioinformatic tools for marker gene sequencing data analysis are continuously and rapidly evolving, thus integrating most recent techniques and tools is challenging. We present an R package for data analysis of 16S and ITS amplicons based sequencing. This workflow is based on several R functions and performs automatic treatments from fastq sequence files to diversity and differential analysis with statistical validation. The main purpose of this package is to automate bioinformatic analysis, ensure reproducibility between projects, and to be flexible enough to quickly integrate new bioinformatic tools or statistical methods. rANOMALY is an easy to install and customizable R package, that uses amplicon sequence variants (ASV) level for microbial community characterization. It integrates all assets of the latest bioinformatics methods, such as better sequence tracking, decontamination from control samples, use of multiple reference databases for taxonomic annotation, all main ecological analysis for which we propose advanced statistical tests, and a cross-validated differential analysis by four different methods. Our package produces ready to publish figures, and all of its outputs are made to be integrated in Rmarkdown code to produce automated reports.