Analysis of bacterial and fungal communities in Marcha and Thiat, traditionally prepared amylolytic starters of India

Marcha and thiat are traditionally prepared amylolytic starters use for production of various ethnic alcoholic beverages in Sikkim and Meghalaya states in India. In the present study we have tried to investigate the bacterial and fungal community composition of marcha and thiat by using high throughput sequencing. Characterization of bacterial community depicts phylum Proteobacteria is the most dominant in both marcha (91.4%) and thiat (53.8%), followed by Firmicutes, and Actinobacteria. Estimates of fungal community composition showed Ascomycota as the dominant phylum. Presence of Zygomycota in marcha distinguishes it from the thiat. The results of NGS analysis revealed dominance of yeasts in marcha whereas molds out numbers in case of thiat. This is the first report on microbial communities of traditionally prepared amylolytic starters of India using high throughput sequencing.

In present study, we attempted to profile the microbial community composition of marcha and thiat, traditionally prepared ethnic starter cultures of India using targeted amplicon sequencing. We selected two different traditionally prepared amylolytic starter cultures from two regions in India, marcha (Fig. 1a) from Sikkim (www. sikkim.gov.in) and thiat (Fig. 1b) from Meghalaya (www.megtourism.gov.in). Marcha is prepared from soaked rice with some wild herbs ((Plumbago zeylanica, Buddleja asiatica and Vernonia cinerea), ginger and red dry chili, 1-2 % of previously prepared marcha powder as an inoculum, crushed in a wooden mortal by wooden pestle, mised and dough are made into round to flatted cakes of different size and shape. Cakes are covered with fern fronds (Glaphylopteriolopsis erubeseens), fermented at room temperature for 24 h, sun dried for 3-5 days and are used as amylolytic starters for production of cereal-based ethnic fermented beverages such as kodo ko jaanr, bhaati jaanr, raksi, etc. 2 During thiat preparation, soaked glutinous rice is grinded with leaves and roots of wild plant Amomum aromaticum, 1-2% of old thiat, mixed and made into a dough by adding water. Flat to round balls are made and fermented for 1-3 days. The freshly prepared thiat balls are sun dried for 3-5 days. It is used to ferment alcoholic beverage locally called kiad in Meghalaya 2 . Fermentation process involved in preparation of these starters is unconditional and may harbor both bacterial and fungal communities as consortia. Therefore, we aimed to explore the bacterial and fungal (filamentous molds and yeasts) communities in marcha and thiat. This is the first report on complete microbial community profile of traditionally prepared amylolytic starters of India using NGS technique.

Results
Characterizing microbial diversity. High throughput sequencing and quality trimming of 16S rRNA and ITS gene yielded ~0.85and ~0.29 million quality reads in both marcha and thiat, respectively, which was used for subsequent analysis. Taxonomic assignment of sequences with the reference database resulted into 5,015 operational taxonomic units (OTUs). The average Good's coverage of both the samples of marcha and thiat for 16Sr RNA amplicon sequencing was found to be 99.08% ± 0.1% (mean ± SD) whereas for ITS region was recorded as 87.5% ± 17.6% (mean ± SD) indicating majority of the diversity was captured.

Discussion
Our study provides comprehensive microbial diversity analysis using deep sequencing approach of ethnic amylolytic starter from India. Quantitative differences were noted for the presence of bacterial and fungal taxa among marcha and thiat; which could be the consequence of differences in the preparation, incubation period and most importantly the type of preservations. Alpha diversity estimation of amylolytic starters marcha and thiat using species richness and non-parametric Shannon index suggested higher bacterial diversity in thiat while marcha shows the higher assemblage of fungal diversity with dominance of yeast phylum Ascomycota. Persistence of higher fungal diversity in marcha is determinant factor suggesting the higher acidic conditions of marcha; in contrast, higher bacterial diversity of thiat depicts the faster turnover from acidic to alkali with the presence of acid neutralizing bacterial taxa 32 .
Acetobacter, Fructobacillus, Lactococcus, Lactobacillus, Leuconostoc, Burkholderia, and Gluconacetobacter were the predominant bacterial genera in marcha. Higher proportion of Acetobacter was possibly due to its retention and enrichment during fermentation. We observed relatively lower proportion of Streptococcus and Lactococcus than Lactobacilli; as Lactobacilli have high acid tolerance over former two 33 . Though some species of Lactococcus have low acid tolerance, however, they could be isolated from raw milk and were found flourishing during the early stage of fermentation 24 . This supports the lower abundance of Lactococcus than Lactobacillus as seen in our samples. Another interesting observation was absence of Pediococcus in bacterial community profile which was otherwise present as a one of the dominant genus in earlier report by culture dependent methods in marcha 10,16 . Furthermore, since there is no earlier report on microbial composition based on culture dependent or culture-independent methods of thiat the present study describe microbial diversity of thiat using NGS method as its first report. Pantoea, Cronobacter, Escherichia, Shigella, Enterobacter, Citrobacter, Salmonella, Serratia, and Enterococcus depicts most dominant bacterial genera of thiat each comprised over 0.1% of total bacterial sequences. Significantly varied microbial composition among thiat and marcha is a clear indication of differences in amylolytic starters. Genus Enterobacter was also detected in Mexican alcoholic beverages speculated to originate from the bacterial contamination in raw milk and they subsequently decreased during the fermentation process 34 . The lactic acid bacterium such as Lb. plantarum seemed to be one factor for the good quality of the alcoholic beverages, as it can perform malolactic fermentation to decrease wine acidity 32 and also produces bacteriocins 35 .
Exploration of fungal diversity of ethnic amylolytic starters suggested higher abundance of yeast in marcha and thiat constitutes for 32.33-fold yeast to the filamentous molds. This observation was in coherence with the earlier report of culture-dependent studies showing the dominance of Mucor and Rhizopus genera of Mucorales in marcha 8 . Interestingly no filamentous molds were detected in marcha using the applied high throughout sequencing method; the exact reasons for the observed variation in the microbiota have not been identified. This may be due to lower abundance of molds, limited sample size and/or age of the sample and finally also due to inadequate cell lysis which may prevent the release of nucleases 36 . Our study was in accordance to the previous reports describing the exposure of cheese to different external environments such as manufacturing process; geographical region, etc have varied impact on the microbial composition of the end products 28 . Thus, we speculate that the factor of geographic environment including altitudes and climate play a more significant role over the manufacturing process in resulting in the different microbial compositions of the starter culture under study. Some other crucial factors that may affect the composition of microbial communities in fermented amylolytic starters are level of hygiene, quality of the glutinous rice, water, as well as the back slopping technique. In this study three dominant yeasts in marcha were Wickerhamomyces anomalus, Candida quercitrus and Kazachstania exigua, followed by Saccharomyces and Pestalotiopsis which also accompany the findings of ref. 21 by PCR-DGGE method. ITS gene sequences analysis of the thiat revealed the existence of Aspergillus penicillioides, Aureobasidium pullulans and Mucor circinelloides as the most dominant filamentous molds in thiat. At family level thiat shows Trichocomaceae, Dothioraceaeand Mucoraceaeas the major constituents of fungal community composition emphasizing the significant differences between thiat and marcha viz differences in starter substrates, preparations, inoculums, consortia, geography, hygiene, preservation technique, caloric values etc.
In the present study Ascomycota was dominant in starter cultures of India like in Korean and Chinese starters cultures, which was also reported earlier, based on NGS tools, in Korean alcoholic beverages 3 and in Chinese liquors 37 . We could also expect similar observation in case of marcha as it has higher abundance of lactic acid bacteria. Aspergullus oryzae has strong secretion of amylases including alpha-amylase, which may accelerate the degradation of grains and provide more nutrients for microbes in alcoholic fermentation 38 .
Amylolytic starter culture-making technology reflects the traditional knowledge of the ethnic Indian people on sub-culturing desirable inocula from previous batch to new culture using rice as base substrates by back-sloping method. This technique preserves the consortia of microbial community ranging from filamentous molds, yeasts and bacteria which were co-existed in traditionally prepared amylolytic and alcohol producing starters 7 , and also preserves vast biological genetic resources, otherwise, which may be forced to disappear. Fermented beverages produced by using amylolytic starters in India are generally mild-alcoholic (4-5%), sweet taste with several health benefits to the local consumers as high source of calories, some vitamins and minerals 2 . Ethnic fermented beverages and alcoholic drinks have the potential to grow beverage industry if proper scientific and technical support are applied to the existing indigenous practices of home based alcoholic fermentation.

Materials and Methods
Sample collection. Samples of sun-dried amylolytic starters marcha and thiat were collected immediately after the preparation from local people of Gangtok and Shillong in Sikkim and Meghalaya states of India, respectively. Dry samples were transferred to sterile containers, sealed, and stored at desiccator at room temperature for the further analysis.
Community DNA Extraction. The total community DNA was extracted using ProMega DNA kit (ProMega). 1g of amylolytic starter culture sample was suspended in lysis solution and incubated at 65 °C for 15 min. Subsequently, the RNA was eliminated from the cellular lysate by administering the RNase solution following incubation at 35 °C for 15 min. The residual proteins were removed by adding protein precipitation solution and centrifuged at max speed. Finally, the DNA was precipitated by adding isopropanol, which was purified with two washes of 70% ethanol. The quality of DNA was checked on 0.8% agarose gel and concentration was measured using Nano-DropND-1000 spectrophotometer (Nano Drop technologies, Willington, USA) as described by ref. 39. The DNA was stored at −20 °C until further processing. Amplicon sequencing. Amongst the nine hypervariable regions of bacterial 16S rRNA gene, we have targeted V4 hyper-variable region 40 to investigate bacterial diversity of marcha and thiat. The universal 16S rRNA gene primer sets F515 and 806R 41 was used for the amplification of V4 hyper-variable region. Similarly, fungal Internal Transcribed Spacer (ITS) II region was targeted for taxonomic profiling amylolytic starters, which was subjected to amplification using ITS1 and ITS2 primers. The library preparation of both the 16S rRNA and ITS gene amplicons were in accordance with the protocols of Illumina (USA). These amplicon libraries were further processed for sequencing using AMPure XT beads (Beckman Coulter Genomics, Danvers, MA, USA). The resultant product was screened with the LabChip GX (Perkin Elmer, Waltham, MA, USA) and with the Library Quantification Kit for Illumina (Kapa Biosciences, Woburn, MA, USA). Subsequently, the 16S rRNA and ITS gene library were sequenced on the Illumina MiSeq platform using 2x 250bp chemistry. The sequences obtained from high throughput sequencing effort were submitted to National Center for Biotechnology Information (NCBI) which are available under BioProject ID PRJNA376467.
Bioinformatics analysis. The raw sequences generated from MiSeq platform was assembled using FLASH tool (Fast Length Adjustment of Short reads) a Paired end assembler for DNA sequences 42 . The assembled reads were subjected to quality filtering using via Quantitative Insights into Microbial Ecology (QIIME) 1.8 42 . Sequence reads were assigned to bacterial and fungal operational taxonomic units (OTUs) by a closed reference-based OTU picking approach by using SILVA and UNITE reference databases, respectively. The OTU picking was carried out using UCLUST method with similarity threshold of 97% 43 . Taxonomic assignments were performed using RDP naïve bayesian classifier 44 . Alpha diversity indices like Chao, Shannon and Simpson were calculated via QIIME after rarefying all samples to the same sequencing depth 45, 46 . Data availability. The sequences obtained from high throughput sequencing effort, was submitted to NCBI which are available under Bio Project ID PRJNA376467.