Abstract

Spontaneous vegetable fermentations, with their rich flavors and postulated health benefits, are regaining popularity. Their usage by professional chefs is steadily growing worldwide, whereas interest has also increased at household level. In addition, fermented vegetable juices, such as fermented carrot juice, show a great potential as an alternative for the more commonly used dairy-based probiotic carriers. However, in general, their microbiology is still poorly understood, and hence raising concerns about food safety.
In this PhD thesis, the spontaneous fermentation process of carrot juice was studied. To this end, samples from 38 fermentations were collected through a citizen science initiative, termed Ferme Pekes, in addition to three well-controlled laboratory fermentations. An initial high relative abundance of unwanted Enterobacteriaceae was found, which shifted to a high relative abundance of Leuconostoc and Lactobacillus afterwards. Furthermore, lactic acid and mannitol were produced in high concentrations, as well as the biogenic amine cadaverine. These results showed that to avoid high numbers of Enterobacteriaceae and high biogenic amine concentrations, this fermented food product could benefit from the development of starter cultures. Therefore, the genomes of 54 lactic acid bacterial strains, isolated from these fermentations, were sequenced for further analysis. Phylogenetic analysis and comparison with the high number of publicly available Lactobacillus (Lb.) genome sequences showed that these strains belonged to eight different species. For two of these species, putative Lb. plajomi and Lb. mudanjiangensis, this PhD thesis was the first to describe their genome sequences. Several of the fermented carrot juice strains showed the genetic potential to produce either homo- or heteropolysaccharides. In addition, this study showed that, due to the presence of genes related to biogenic amine production, the usage of Lb. brevis strains as starter culture should be avoided. In contrast, the strains classified as Lb. plantarum, Lb. plajomi, and Lb. mudanjiangensis appeared to be good candidates in the light of starter culture development.
Since its first description, no other study provided additional characterization or reported the isolation of other strains of the Lb. mudanjiangensis species. However, four strains isolated from three different spontaneous carrot juice fermentations were members of this species. This species harbored one of the largest genomes and highest gene counts of the Lb. plantarum group. The species showed a broad repertoire of glycosyl hydrolases and a predicted capability to degrade cellulose, a fiber present in carrots. Furthermore, three of the four strains studied showed the presence of pili on scanning electron microscopy images, which were linked to conjugative gene regions.
Finally, the computational comparative genomics pipeline developed for the analysis of Lb. mudanjiangensis was applied to the economically important Lb. casei group. Classification of strains belonging to this phylogenetic group has previously been troublesome. The phylogenetic analysis showed that many Lb. casei strains were wrongly annotated in the National Center for Biotechnology Information Assembly database and should be reclassified as Lb. paracasei. Reclassification led to the discovery of at least one catalase gene in all of the members of Lb. casei, making it the first described catalase-positive species of the Lactobacillus genus. Finally, the use of average nucleotide identity values was proposed as a computational cheap way to distinguish the species Lb. casei and Lb. paracasei.
Date of Award4 Feb 2019
Original languageEnglish
SupervisorLuc De Vuyst (Promotor)

Cite this

'