AUTHOR=Sirén Kimmo , Mak Sarah Siu Tze , Melkonian Chrats , Carøe Christian , Swiegers Jan Hendrik , Molenaar Douwe , Fischer Ulrich , Gilbert M. Thomas P. 

TITLE=Taxonomic and Functional Characterization of the Microbial Community During Spontaneous in vitro Fermentation of Riesling Must

JOURNAL=Frontiers in Microbiology

VOLUME=Volume 10 - 2019

YEAR=2019

URL=https://www.frontiersin.org/journals/microbiology/articles/10.3389/fmicb.2019.00697

DOI=10.3389/fmicb.2019.00697

ISSN=1664-302X

ABSTRACT=Although there is an extensive tradition of research into the microbes that underlie the winemaking process, much remains to be learnt. We combined the high-throughput sequencing (HTS) tools of metabarcoding and metagenomics, to characterise how microbial communities of Riesling musts sampled at four different vineyards, and their subsequent spontaneously fermented derivatives, vary. We specifically explored community variation related to three points: i) how microbial communities vary by vineyard; ii) how community biodiversity changes during alcoholic fermentation; and (iii) how microbial community varies in musts that successfully complete alcoholic fermentation versus those that become ‘stuck’ in the process. 
At the vineyard level, our metabarcoding data shows an overall impact, and difference in the abundance of Metschnikowia between vineyards 4 and 5. We also observed at these two vineyards an increase in non-Saccharomycetaceae fungal functions and a decrease in bacterial functions during the early fermentation stage. To investigate this further, we reconstructed draft genomes (bins), yielding 11 coherent bins and found that these vineyards shared yeast bins Hanseniaspora and Saccharomyces. Read recruitment and functional analysis revealed that during fermentation, a high abundance of Metschnikowia might serve as a biocontrol agent against bacteria, via a putative iron depletion pathway, and this in turn could help Saccharomyces dominate the fermentation.
With regards to alcoholic fermentation, we observed a general decrease in biodiversity in both the metabarcoding and metagenomic data. Unexpected Micrococcus behaviour was observed in vineyard 4 according to metagenomic analyses based on reference based read mapping. Analysis of open reading frames using this data showed an increase of functions assigned to unknown Actinobacteria in the end of fermentation. Therefore, we hypothesise that bacteria might sit-and-wait until Saccharomyces activity slows down. 
Lastly, our metabarcoding data enabled us to identify a relationship between stuck fermentations and Starmerella abundance. Given that chemical analysis indicated that although the stuck samples contained residual glucose, all fructose had been consumed, we hypothesise that fructophilic Starmerella dominated these fermentations instead of Saccharomyces. Overall our results showcase how metagenomic functional analysis offer possibilities to improve our insights into the wine alcoholic fermentation process, including highlighting microbe interactions.