Genomic Characterization of Dairy Associated Leuconostoc Species and Diversity of Leuconostocs in Undefined Mixed Mesophilic Starter Cultures

Abstract

Undefined mesophilic mixed (DL-type) starter cultures are composed of predominantly Lactococcus lactis subspecies and 1-10% Leuconostoc spp. The composition of the Leuconostoc population in the starter culture ultimately affects the characteristics and the quality of the final product. The scientific basis for the taxonomy of dairy relevant leuconostocs can be traced back 50 years, and no documentation on the genomic diversity of leuconostocs in starter cultures exists. We present data on the Leuconostoc population in five DL-type starter cultures commonly used by the dairy industry. The analyses were performed using traditional cultivation methods, and further augmented by next-generation DNA sequencing methods. Bacterial counts for starter cultures cultivated on two different media, MRS and MPCA, revealed large differences in the relative abundance of leuconostocs. Most of the leuconostocs in two of the starter cultures were unable to grow on MRS, emphasizing the limitations of culture-based methods and the importance of careful media selection or use of culture independent methods. Pan-genomic analysis of 59 Leuconostoc genomes enabled differentiation into twelve robust lineages. The genomic analyses show that the dairy-associated leuconostocs are highly adapted to their environment, characterized by the acquisition of genotype traits, such as the ability to metabolize citrate. In particular, Leuconostoc mesenteroides subsp. cremoris display telltale signs of a degenerative evolution, likely resulting from a long period of growth in milk in association with lactococci. Great differences in the metabolic potential between Leuconostoc species and subspecies were revealed. Using targeted amplicon sequencing, the composition of the Leuconostoc population in the five commercial starter cultures was shown to be significantly different. Three of the cultures were dominated by Ln. mesenteroides subspecies cremoris. Leuconostoc pseudomesenteroides dominated in two of the cultures while Leuconostoc lactis, reported to be a major constituent in fermented dairy products, was only present in low amounts in one of the cultures. This is the first in-depth study of Leuconostoc genomics and diversity in dairy starter cultures. The results and the techniques presented may be of great value for the dairy industry.

Keywords: cheese; comparative; dairy; differentiation; diversity analysis; genomics; leuconostoc; starter cultures.

Figure 1

Bacterial counts for five starter cultures A–E on MRS and MPCA supplemented with vancomycin to select for Leuconostoc. The counts are the mean of three separate extractions made from the same culture batch and the error bar indicates the standard deviation. The blue bars represent the bacterial counts on MPCA, while the orange bars represent the bacterial counts on MRS. The Y-axis is cut at 1,0E+06 for better readability.

Figure 2

Pan- and core-genome estimation. The estimation is made by including genomes one by one, matching the genetic content from each genome, with the growing pan- and the decreasing core-genome. Homologous genes are clustered together in orthologous groups. If the genomes included in the estimation are sufficiently distant from each other with regards to phylogeny, more than one orthologous group can exist for the same gene. The cut-off for this is set by the inflation value in the Markov Cluster Algorithm (MCL), for our dataset the inflation value was set to 1.5. The genetic content was curated for significantly divergent singletons, likely to be the product of erroneous assembly or annotation. The final pan-genome was estimated at 4415 orthologous groups, while the core-genome was estimated at 638 orthologous groups.

Figure 3

Differentiation of 59 Leuconostoc genomes using the pan-genome of 4415 OGs. Hierarchal clustering of genomes clearly separated Leuconostoc species and subspecies. Moreover, the high sensitivity of the method produced twelve robust Leuconostoc lineages annotated on the right side of the figure. Four lineages of Ln. mesenteroides (colored orange), three lineages of Ln. cremoris (colored blue), four lineages of Ln. pseudomesenteroides (colored green), and one linage of Ln. lactis (colored purple) are shown. (^*) The Ln. cremoris TIFN8 genome was excluded from further analysis because the genome data contained a high number of fragmented genes and redundant sequences. The heatmap was generated with R using the heatmap.2 function included in the Gplots package supplemented by the Dendextend package.

Figure 4

Composition of leuconostocs in five starter cultures A–E using targeted-amplicon sequencing of the eno gene.