Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China

Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China

Accepted Manuscript Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China ...

2MB Sizes 0 Downloads 2 Views

Accepted Manuscript Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China

Feifei Gao, Jiluan Chen, Jing Xiao, Weidong Cheng, Xiaoji Zheng, Bin Wang, Xuewei Shi PII: DOI: Reference:

S0963-9969(19)30261-3 https://doi.org/10.1016/j.foodres.2019.04.029 FRIN 8410

To appear in:

Food Research International

Received date: Revised date: Accepted date:

7 October 2018 15 February 2019 13 April 2019

Please cite this article as: F. Gao, J. Chen, J. Xiao, et al., Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China, Food Research International, https://doi.org/10.1016/ j.foodres.2019.04.029

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Microbial community composition on grape surface controlled by geographical factors of different wine regions in Xinjiang, China Feifei Gaoa, Jiluan Chena, Jing Xiaob, Weidong Chenga, Xiaoji Zhenga, Bin Wanga,* & Xuewei Shia,*

RI PT

a. Food College, Shihezi University, Shihezi 832000, Xinjiang Uygur Autonomous Region, P. R. China;

b. College of Information Science and Technology, Shihezi 832000, Xinjiang Uygur Autonomous

SC

Region, P. R. China.

* Corresponding authors

NU

E-mail addresses: [email protected] (B. Wang); [email protected] (X. Shi)

AC

CE PT

ED

MA

Tel.: 86-0993-2058093

1

ACCEPTED MANUSCRIPT Abstract: Native microorganisms on wine grape surface contribute to regional wine characteristics and quality. The Xinjiang Uygur Autonomous Region in northwest China is one of the eight main wine-producing areas in China. To investigate the relationship between the microbial community structure of wine grape epidermises and environmental conditions, 16S rDNA and ITS sequences of 48 wine grape

RI PT

samples from four wine grape cultivars and six wine-growing regions in Xinjiang were sequenced, based on Illumina high-throughput sequencing technology. A total of 691 operational taxonomic units (OTUs) in 16 bacterial phyla and 349 OTUs in three fungal phyla were identified. Among them, Proteobacteria and Ascomycota were the

SC

predominant bacteria and fungi, respectively. Canonical correspondence analysis indicated that bacterial community diversity was largely related to altitude, latitude

NU

and longitude, while that of the fungi was closely related to altitude, dryness, frost-free period, latitude and longitude. Our results suggest that microbial community

MA

structure on the surface of wine grape is controlled primarily by geographical conditions.

ED

Keywords: Wine grape; Microbial community; geographical factors; Illumina

AC

CE PT

high-throughput sequencing

2

ACCEPTED MANUSCRIPT 1. Introduction Wine is an alcoholic beverage derived from grapes. Natural factors, such as climate, geology, soil and grape cultivars, remarkably contribute to wine quality by affecting grape features or microorganism communities (Karna L. Sacchi, Linda F. Bisson, & Adams, 2005; Mirás-Avalos et al., 2017). Especially, epiphytic

RI PT

microorganism on the surface of grape berries, including bacteria, yeast and filamentous fungi, and establish an intricate and kinetic microbe ecosystem (Y. Liu et al., 2017; Martins et al., 2014). It was previous reported that microorganisms

SC

colonized on the surface of grape berries, such as yeast, fungi and filamentous fungi, play a major role in crop health and contribute to wine quality as participants in

NU

winemaking process (Nisiotou, Rantsiou, Iliopoulos, Cocolin, & Nychas, 2011; Verginer, Leitner, & Berg, 2010). Thus, the microbial ecology on wine-grape surface

MA

is played strong attention in recent study.

Indeed, the structure and the composition of yeast cells greatly contribute to the sensory features of wine, especially wine fermented using spontaneous fermentation

ED

(Romano, Fiore, Paraggio, Caruso, & Capece, 2003). Moreover, the flavour of different wines can be established during winemaking via the complex roles of

CE PT

various microorganisms, particularly yeast (Christiaens et al., 2014). Some yeasts like Saccharomyces, Cryptococcus, Hanseniaspora and Rhodotorula genus have been proven to contribute to the composition, sensory properties and flavour of wine,

AC

forming a wine style (Ciani & Comitini, 2010; Hierro, Gonzalez, Mas, & Guillamon, 2006; Hong & Park, 2013). During spontaneous fermentation, epiphytic microorganisms always play vital roles in affecting wine quality or safety by producing a large amount of metabolites, such as secondary aroma and urethane (Cordero-Bueso et al., 2011). In addition, It was reported that the plant-associated microorganism have also been proven to have a positive interaction with their host plants, such as plant growth promotion and pathogen defence (Garbeva, van Veen, & van Elsas, 2004; Xie, Zhang, & Paul, 2009). Some strains such as Botrytis cinerea and Neofusicoccum parvum can devastate the fungal pathogens of grapevines (Haidar et 3

ACCEPTED MANUSCRIPT al., 2016). In recent studies, microbes on grape surface have been reported in large numbers. Furthermore, the technology of high-throughput sequencing has been widely applied to investigate microbial communities for grape, must or wine (Portillo & Mas, 2016; Li et al., 2018). Barata et al. (2012) have reported over 50 bacterial species identified

RI PT

on grape berries, mostly belonging to two groups: Firmicutes and Proteobacteria. 47 yeast species belonging to 22 different genera have been reported in the literature, including Aureobasidium, Auriculibuller, Brettanomyces, Bulleromyces, Candida, Cryptococcus,

Debaryomyces,

Hanseniaspora,

Issatchenkia,

Kluyveromyces,

Sporidiobolus,

Sporobolomyces,

SC

Lipomyces, Metschnikowia, Pichia, Rhodosporidium, Rhodotorula, Saccharomyces, Torulaspora,

Yarrowia,

Zygoascus,

and

NU

Zygosaccharomyces (Liu et al., 2017). It was reported that the microbial community on the grape berry surface changed with the ripening of grape, especially fermentative

MA

yeasts were detected at harvest and not in the first stage of grape growth (Renouf Claisse, & Lonvaud-Funel, 2005). Further, the methods of study microbial community

ED

were reviewed, which highlighted the contribution of high-throughput next generation sequencing and metagenomics for vineyard microbial ecology (Morgan, Toit, & Setati,

CE PT

2017).

However, the microbial community structure on the surface of wine grape might be influenced by some factors such as geography (Altitude, latitude and longitude), climate (rain, temperature, humidity, frost-free period), grape variety and viticultural

AC

practice (herbicides, fertilizers, pesticides, fungicides). The latest research has shown that climatic conditions play a key role in the nature of microbial communities in the environment (Castro, Classen, Austin, Norby, & Schadt, 2010; Wallenstein & Hall, 2011). Gayevskiy et al. (2012) reported that yeast communities and populations associated with vines and wines were affected by geographlic factors in New Zealand. Bokulich et al. (2014) demonstrated that regional, site-specific, and grape cultivar factors shape the fungal and bacterial consortia inhabiting wine-grape surface. It was reported that the differences of microbial community were the result of using a biodynamic and traditional agriculture approach (Guzzon et al., 2016). Moreover, it 4

ACCEPTED MANUSCRIPT has been proved that microbial community was subject of geographic patterns and agronomic management (Vitulo et al., 2018). Of course, using fungicides in grape growing process may kill some microbes, causing the indifference of microbial community (Comtini & Ciani, 2008; Milanovic, Comitini, & Ciani, 2013). Due to its unique geographical environment and climate, the Xinjiang Uygur

RI PT

Autonomous Region of northwest China is among eight main wine-producing areas in China. Although Zhang et al. (2017) has reported epiphytic microbial community on the surface of grape berries from Shacheng region of China,it is not clear that the epiphytic microbial community is influenced by several factors such as geography,

SC

climate, variety and agriculture. Moreover, that is rarely reported for Xinjiang region. In this study, the bacterial and fungal community structures of four wine grape

NU

cultivars from six climatic regions of China were researched using 16S rDNA/ITS (internal transcribed spacer) Illumina high-throughput sequencing method. It is further

MA

investigated whether geography, climate and variety factors have an effect on microbial communities of grape epidermises in different regions. This study provides

ED

significant scientific proof for classifying wine grape quality in different areas of Xinjiang and for evaluating wine-producing techniques and microbe resources.

CE PT

2. Materials and methods

2.1. Sample collection and cell treatment Though Xinjiang has a vast territory and a temperate continental climate, grape

AC

cultivation is mainly concentrated in northern Xinjiang. Firstly, five samples from different sites of each vineyard were collected, then mixing these samples in sterile environment, finally taking out 20 g from mixed samples as a testing sample and repeating 3 times for the same variety. A total of 48 wine grape samples, belonging to four different wine grape cultivars (Cabernet Sauvignon, Merlot, Italic Riesling and Cabernet Franc), were collected from six wine-growing regions in northern Xinjiang in northwest China, including Shanshan, Yanqi, Heshuo, Huoerguosi, Fukang and Manasi. All of the samples were collected using sterile 50-mL centrifuge tubes and were 5

ACCEPTED MANUSCRIPT rapidly stored at -20 °C. Among the samples, the wine grape cultivars, Cabernet Sauvignon, Merlot, Italian Riesling and Cabernet Franc, were numbered 1, 2, 3 and 4, respectively. While the cultural areas, Shanshan, Yanqi, Heshuo, Huoerguosi, Fukang and Manasi, were abbreviated to S, Y, H, G, F and M, respectively. For example, sample M1 was grapes of Cabernet Sauvignon sampled from Manasi.

RI PT

After thawing, 20 g of grape samples was washed with 40 mL of sterile water five times. The suspension was collected in a 250-mL Erlenmeyer flask and then filtered using a 0.22-μm filter. A filtered microorganism was used to extract microbial genome DNA.

SC

2.2. Extraction of genome DNA and PCR amplification

Power Water® DNA Isolation Kit (MOBIO firm in America) was used to extract

NU

total DNA of the microorganisms on grape surface (Goldberg et al., 2016). The DNA concentrations were evaluated by electrophoresis in a 1.5% agarose gel, and the

MA

extracted DNA was stored at -20 °C until further testing (Ubeda, Maldonado Gil, Chiva, Guillamon, & Briones, 2014). Then, 1 μL of genomic DNA was added to a

ED

centrifuge tube and diluted to 1 ng/μL using sterile water. A total of 48 diluted DNA samples were submitted to the Novegene company for 16S rDNA Amplicon

CE PT

Sequencing and Internal Transcribed Spacer (ITS) sequencing based on the Illumina Hiseq sequencing platform.

The variable region V4 of the 16S rDNA gene was selected for the construction of the bacterial community library for the Illumina sequencing. The specific primers, (5’-GTGCCAGCMGCCGCGGTAA-3’)

AC

515F

and

806R

(5’-GGACTACHVGGGTWTCT AAT-3’), were used to amplify the sequence of the 16S rDNA gene, while ITS5-1737F (5’-GGAAGTAAAAGTCGTAACAAGG-3’) and ITS2-2043R (5’-GCT GCGTTCTTCATCGATGC-3’) were used to amplify the fungal ITS sequence. PCR was performed in 20-μL reactions in triplicate, with each reaction tube containing 0.2 mM of each primer, 10 ng of template DNA (the diluted genomic DNA), 0.25 mM of dNTPs, 1×PCR reaction buffer, and 2 U of FastPfu DNA Polymerase. The following PCR condition was used for the 16S rDNA: 95 °C for 2 min, 95 °C 30 s, 55 °C 30 s and 72 °C 45 s for 30 cycles, and a final extension of 6

ACCEPTED MANUSCRIPT 72 °C for 10 min. The same PCR conditions were used for ITS, except that the second stage had 35 cycles (Hao, Song, Mu, Hu, & Xiao, 2016). Additionally, Phusion® High-Fidelity PCR Master Mix with GC Buffer from the New England Biolabs company, a high-performance and high-fidelity enzyme, was used for PCR to ensure amplification efficiency and accuracy (Heydenreich et al., 2017). Then, PCR products

RI PT

were mixed at the same volume of the 1X loading buffer (containing SYB green) were detected by electrophoresis in a 2.0% agarose gel, and samples with a bright main strip between 400–450 bp were chosen for further experiments. 2.3. PCR product purification and library preparation

SC

PCR products were mixed in equal density ratios according to the concentration of PCR products and then purified using a Qiagen Gel Extraction Kit (Qiagen

NU

company in Germany) (Zetsche et al., 2017). Sequencing libraries were generated using the TruSeq® DNA PCR-Free Sample Preparation Kit (Illumina, USA)

MA

following the manufacturer's recommendations, then index codes were added (Nones et al., 2014). The library quality was evaluated using the [email protected] 2.0 Fluorometer

ED

(Thermo Scientific) and an Agilent Bioanalyzer 2100 system (Kao et al., 2016; Zhai, Wang, Tan, & Cao, 2016). Finally, the library was sequenced on an Illumina

CE PT

HiSeq2500 platform and 250 bp paired-end reads were generated. 2.4. Bioinformatic analyses

The paired-end (PE) reads were obtained after accomplishing high-throughput sequencing and were assigned to samples based on their unique barcode and truncated

AC

by cutting off the barcode and primer sequence. PE reads were spliced using FLASH V1.2.7 (El-Ashram & Suo, 2017), filtrated according to the QIIME V1.7.0 quality control process, and termed the high-quality clean tags. The tags were compared to the reference database using the UCHIME algorithm to detect chimaera sequences, and then the chimaera sequences were removed (Li et al., 2017). Then, the effective tags were finally obtained. Sequences with greater than 97% similarity were assigned to the same operational taxonomic units (OTUs). The QIIME suite of programs was used to evaluate alpha diversity, including ACE, Chao1 richness, Shannon diversity, Goods-coverage, Simpson index and observed species (Marsh, O'Sullivan, Hill, Ross, 7

ACCEPTED MANUSCRIPT & Cotter, 2013). Rarefaction analysis was used to estimate sequencing depth. The differences in community structure of the different samples and groups were analysed using principal coordinate analysis (PCoA) based on weighted and unweighted calculations (Horn, Hempel, Verbruggen, Rillig, & Caruso, 2017; W. Liu et al., 2015). Flower chart and Venn diagrams were used to explore the common and specific OTU

The

microbial

communities

of

grape

RI PT

information among different samples or groups. epidermises

among

different

regions/cultivars were further compared using analysis of molecular variance (Amova), analysis of similarities (Anosim), Metastats and LDA EffectSize (LEfSe).

SC

Amova analysis using mothur software was performed to assess the significance of the microbial community structure among different regions (Wallenstein & Hall,

NU

2011). Anosim is a distribution-free method of multivariate data analysis to test whether an inter-group difference is significantly greater than an intra-group

MA

difference (Hollister, Schadt, Palumbo, James Ansley, & Boutton, 2010). Metastats analysis is a statistical method for identifying significantly different bacterial and

ED

fungal genera in different wine-growing regions (Yan et al., 2016). LEfSe is a software used to identify high-dimensional biometric identifiers and

CE PT

identify genomic characteristics (Hollister et al., 2010; Segata et al., 2011). In this study, it was used to estimate the effects of the abundance of each component on the microbial population difference among various samples. The correlation between environmental variables and microbial community composition was performed using

AC

canonical correspondence analysis (CCA) (Sterkenburg, Bahr, Brandstrom Durling, Clemmensen, & Lindahl, 2015). Spearman correlation analysis and variance partitioning analysis were used to identify the correlation between microbial community composition and environmental and climatic factors (J. Liu et al., 2014; Xiao et al., 2017). The significant correlation was further proved using a Mental test (Yang, Ma, Jiang, Wu, & Dong, 2016). After basic analysis, figures were drawn using packages in RStudio (version 2.15.3) (Hayden & Beman, 2016; Jia et al., 2016).

3. Results 8

ACCEPTED MANUSCRIPT 3.1 Diversity assessment A total of 48 wine grape samples of Cabernet Sauvignon, Merlot, Italian Riesling and Cabernet Franc were collected from vineyards in Shanshan, Yanqi, Heshuo, Huoerguosi, Fukang and Manasi of the Xinjiang area of northwest China. To evaluate the quality of the sequencing results, the raw sequences, raw tags, average length,

RI PT

sequencing error rate and effective tag percentage of the samples were analysed. In this study, raw paired-end (PE), effective tags, avglen (nt), Q20, Q30 and effective % in the samples were greater than 53,329, 52,267, 219, 99.22%, 98.46% and 89.86%,

parameters met the demands of further analysis.

SC

respectively (Supplementary Table S1). These results indicated that all of the

After eliminating chimeric sequences and mismatches, the total number of 16S

NU

rDNA reads obtained from the 48 samples was 3,245,424 (with an average of 67,613), which were clustered into 691 operational taxonomic units (OTUs) of at least 97%

MA

similarity in nucleotide identity (Supplementary Table S2). However, 2,242,800 (with an average of 46,725) ITS reads were clustered into 349 OTUs (Supplementary Table

ED

S3). The species accumulation boxplot also agreed with the accuracy of these data (Supplementary Fig. S1). When the sequences reached 20,000 in number, the

CE PT

rarefaction curves of most of the samples tended to be complanate, which indicated that a reasonable sequencing depth had been acquired, and further sequencing data might only produce a small number of new OTUs (Supplementary Fig. S2). The

S3).

AC

Shannon-Wiener curve was also in accordance with this claim (Supplementary Fig.

The alpha diversity of the different samples at a 97% consistency threshold was calculated via the Chao1, ACE, Shannon, Simpson and coverage indices. The alpha diversity of bacteria was shown in Table 1. The Chao1 and ACE indices of the samples from the same region displayed slight differences. The value of the Heshuo (H), Yanqi (Y) and Huoerguosi (G) regions were higher than those of the Shanshan (S), Manasi (M) and Fukang (F) regions. These results showed that the bacterial community richness on the grapes from the H, Y and G regions was obviously higher than that of grapes from the S, M and F regions. Furthermore, the Shannon and 9

ACCEPTED MANUSCRIPT Simpson diversity indices, combining evenness and species richness, also indicated that the bacterial community diversity of grapes from the H, Y and G regions was remarkably higher than that of the grapes from the S, M and F regions. The coverage estimates (sequencing depth indices) were very high for all of the samples.

RI PT

Table 1. Abundance and diversity estimation of the 16S rDNA sequencing libraries from 48 wine grape samples. observed Sample

OTUs

shannon

simpson

chao1

ACE

PD whole tree

species 36

20.00±3.61

0.83±0.38

0.35±0.20

23.17±5.32

28.69±7.74

3.22±0.61

S2

38

19.00±3.46

0.34±0.20

0.09±0.05

27.61±3.91

29.35±6.38

2.99±1.90

M1

398

227.33±25.58

0.90±0.62

0.28±0.25

265.19±32.43

275.18±19.20

16.77±1.99

M2

47

22.67±3.51

1.02±0.35

0.40±0.21

29.71±4.75

36.87±6.89

3.43±0.60

M3

56

29.67±8.96

0.28±0.19

0.07±0.05

46.83±18.87

56.76±23.16

4.01±1.22

M4

38

17.00±6.93

0.15±0.05

0.04±0.01

19.55±9.21

22.07±10.54

2.54±1.10

F1

41

23.00±4.36

0.14±0.03

0.03±0.01

23.93±4.76

25.31±4.93

3.08±0.71

F2

51

28.00±3.61

0.43±0.11

0.11±0.03

32.90±4.43

37.42±4.37

4.01±0.88

F3

35

22.33±1.15

0.32±0.02

0.07±0.01

25.42±1.38

27.25±2.57

2.97±0.16

H1

441

254.33±65.96

0.71±0.49

0.16±0.14

291.98±59.47

310.92±56.93

19.38±3.75

H3

311

170.67±41.48

0.64±0.27

0.16±0.08

210.32±49.27

224.77±54.68

14.32±2.09

Y1

251

138.67±14.74

0.36±0.03

0.07±0.01

175.95±31.85

192.37±30.31

12.05±1.21

Y4

336

180.00±64.65

0.96±0.54

0.29±0.20

214.30±80.38

227.81±82.88

15.11±5.32

G1

297

157.33±34.00

0.66±0.56

0.22±0.25

201.90±44.27

207.78±34.47

14.07±3.89

G2

310

G4

392

NU

MA

ED

CE PT

SC

S1

168.00±60.02

0.97±0.61

0.31±0.24

220.04±69.80

229.80±72.26

14.27±4.71

225.67±37.81

1.13±0.44

0.36±0.20

382.18±194.29 356.00±126.38

18.46±1.44

Regarding fungi (Supplementary Table S4), the Chao1 and ACE values of the samples from the G, H, M and Y regions were distinctly higher than those from the S

AC

and F regions. These results suggested that the fungal community richness of the G, H, M and Y regions was obviously higher than that of the S and F regions. At the same time, the Shannon and Simpson diversity indices indicated that the fungal community diversity of the grapes from the G, H, M and Y regions was remarkably higher than that from the S and F regions. However, coverage estimates were very high for all of the samples. The principal coordinates analysis (PCoA) based on the unweighted Unifrac distance showed that the bacterial community composition flocked well compared to 10

ACCEPTED MANUSCRIPT that of the fungal community composition on grapes from different regions. These results indicated that the bacterial community was more sensitive to the natural environment and local climate from different regions than fungal community. The bacterial community composition of the grape surfaces from the S, M and F region mainly flocked to gather, while that of the grape surfaces from the H and G the

RI PT

regions was divided into two clusters. One of the clusters was similar to the Y region (Fig. 1A). Some outliers were also evident. For example, S1.2, S2.1 and F2.1 deviated from their main clusters. The contribution rate of PC1 is 35.1%, followed by PC2 (9.2%). At the same time, the fungal community composition on grape surfaces from

SC

the H region flocked together, while that from the other regions was scattered (Fig.

NU

1B). PC1 alone explains 17.64% of the variance, followed by PC2 at 9.2%. Amova showed that there was a significant difference among the F, G, H, M, S

MA

and Y fungal communities (Amova, p < 0.001), but the difference in the bacterial communities was not significant. The Anosim results indicated that the difference of

ED

bacterial communities between groups is greater than difference within group in F and G region (R = 0.183, p < 0.05). These groups are no exception, such as F and H (R =

CE PT

0.291, p < 0.05), F and S (R = 0.246, p < 0.05), F and Y (R > 0), S and Y (R > 0), and M and G (R > 0). Similarly, the difference among the F, G, H, M, S and Y fungal communities was substantially greater than the intra-group difference (Anosim, R > 0, p < 0.05).

AC

3.2. Taxonomic distribution of bacteria The OTUs of the samples from the different regions were further assigned to different taxa and their relative taxonomic richness was evaluated. In addition to some unknown groups, 36 bacterial classes were grouped into at least sixteen phyla (Supplementary Table S5). Among these phyla, proteobacteria strains were represented in all of the samples, and the relative abundance was higher than 98.55% of the total bacterial population in each group of samples. Furthermore, common and exclusive bacterial OTUs of the grape epidermises 11

ACCEPTED MANUSCRIPT were shown in Venn diagrams. Bacterial OTU distribution of grape epidermises in the F, G, H, M, S and Y cultural areas were as follows: 85 (12.3%), 540 (78.1%), 499 (72.2%), 430 (62.2%), 58 (8.4%) and 405 (58.6%), respectively (Supplementary Fig. S4). Bacterial OTU distribution of the different cultivar grapes in an area showed a significant difference. For example, the OTUs of the Cabernet Sauvignon, Merlot and

RI PT

Cabernet Franc cultivars in Huoerguosi were 297, 310 and 392, respectively, and 163 bacterial OTUs were common (Supplementary Fig. S4C). Similarly, bacterial OTU distribution of the same grape cultivar in different cultural areas was also compared. For instance, for the Cabernet Sauvignon cultivar, exclusive bacteria OTUs of grape

SC

epidermises in the F, G, H, M, S and Y cultural areas were 1, 40, 80, 45, 2 and 17, respectively, and 13 bacterial OTUs were common (Fig. 2). These results indicated

NU

that bacterial OTU distribution of grape epidermises was affected by the cultivation area and cultivar.

MA

Moreover, these common and exclusive bacteria OTUs truly shaped the bacterial community structure. Particularly, the top 10 of relative abundance for bacterial OTUs

ED

were showed by bacterial genera, including Pantoea, Pseudomor, Buchnera, Rhodococcus, Nitrosospira, Massilia, Aeromonas, Steroidobacter, Thermomonas and

CE PT

Pedobacter (Fig. 3). Furthermore, the differences of bacterial abundance were more significant in different regions. For example, the relative abundances of Pantoea genus were 96.3%, 79.6%, 91.3%, 76.6%, 85.1% and 88.9% in F, G, H, M, S and Y cultural areas, respectively. Therefore, the bacterial abundance was affected by

AC

cultivation areas. Proteobacteria

Sequence reads of the proteobacteria phylum could be classified into four classes, Alpha-, Beta-, Delta- and Gammaproteobacteria, as well as some unidentified and unclassified classes. Gammaproteobacteria had a relatively high number of reads, followed by that of Beta-, Alpha- and Deltaproteobacteria (Supplementary Table S5). These four classes were completely present in the G and M cultivation areas, whereas only Alpha-, Beta- and Gammaproteobacteria were present in the F, H, S and Y areas. Gammaproteobacteria was more dominant than Betaproteobacteria, followed by 12

ACCEPTED MANUSCRIPT Alphaproteobacteria and Deltaproteobacteria. Gammaproteobacteria were represented by sixteen orders and twenty-three families, while the Betaproteobacteria were represented

by

eight

orders

and

nine

families.

The

majority

of

the

Gammaproteobacteria reads belonged to the orders Aeromonadales, Enterobacteriales, Pseudomonadales and Xanthomonadales. Furthermore, Enterobacteriales dominated followed

by

Xanthomonadales

and

Aeromonadales

RI PT

Pseudomonadales,

(Supplementary Fig. S5). The four orders were present in the G, H, M, S and Y regions. Notably, Oceanospirillales substituted for Aeromonadales in the F area. Most Betaproteobacteria reads were affiliated with the orders Burkholderiales and

SC

Nitrosomonadales, in which the former possessed a higher relative abundance than the latter. The orders Nitrosomonadaceae and Oxalobacteraceae were present in the G, H,

NU

M and Y areas, but the Betaproteobacteria was only represented by the order Burkholderiales in the F and S areas. Enterobacteriaceae was a dominant family in

MA

Enterobacteriales orders (Supplementary Table S3). Burkholderiales were principally represented by the families Alcaligenaceae, Burkholderiaceae, Comamonadaceae and

ED

Oxalobacteraceae. The relative abundance of many families varied greatly among the F, G, H, M, S and Y areas (Fig. 4A).

CE PT

3.3. Taxonomic distribution of fungi Seventeen classes were classified into at least three phyla (Ascomycota, Basidiomycota

and

(Supplementary

Zygomycota),

Table

S6).

as

Among

well these

as

some

phyla,

unidentified

Ascomycota

groups

dominated

AC

Basidiomycota, followed by Zygomycota. These phyla were represented in all grape epidermises. The Ascomycota had the largest number of reads for each grape epidermis, together constituting more than 70% of the total fungi population in each group of samples. Furthermore, the common and exclusive fungal OTUs of the grape epidermises were shown in Venn diagrams. Fungal OTU distribution of grape epidermises in the F, G, H, M, S and Y cultural areas were as follows: 164 (47.0%), 228 (65.3%), 174 (49.9%), 253 (72.5%), 156 (44.7%) and 163 (46.7%), respectively (Supplementary Fig. 6). Regarding different grape cultivars in the same area, the difference in fungal 13

ACCEPTED MANUSCRIPT OTU distribution was obviously apparent. For example, the OTU of the Cabernet Sauvignon, Merlot, Italian Riesling and Cabernet Franc cultivars in Manasi was 176, 139, 127 and 169, respectively, and 79 fungal OTUs were common (Supplementary Fig. 5A). Similarly, the fungal OTU distribution of the same grape cultivars in different cultural areas was also compared (Fig. 5). For example, for the Cabernet

RI PT

Sauvignon cultivar, the number of exclusive fungal OTUs of grape epidermises in the F, G, H, M, S and Y cultural areas was 4, 27, 10, 33, 12 and 16, respectively, and 64 fungal OTUs were common (Fig. 5A). These results indicated that fungal OTU distribution of grape epidermises was affected by the cultivation area and cultivar.

SC

However, these common and exclusive fungal OTUs truly shaped fungal community structure. In particular, fungal OTUs of the top 10 abundance were

NU

showed by fungal genera included Aureobasidum, Alternaria, Botrytis, Rhodotorula, Cryptococcus, Hanseniaspora, Mucor, Fusarium, Chaetopyrena, and Cladosporium.

MA

The differences of fungi abundance were more significant in the different regions. For example, the relative abundance of Aureobasidum genus was 48.8%, 45.4%, 13.3%,

ED

36.7%, 12.4% and 7.4% in the F, G, H, M, S and Y cultural areas, respectively. Thus, fungal abundance was affected by different cultivation areas (Fig. 6).

CE PT

Differential features were identified at the OTU level. Six cultivation areas were used as the class of the subjects. LEfSe revealed 6, 9, 6, 11, 13 and 5 fungal clades on samples from the F, G, H, M, S and Y regions, respectively, as well as statistically significant differences in fungal communities among the six areas (Fig. 7). The most

AC

differentially abundant fungal taxa in the F, G, H, M, S and Y areas belonged to the Dothideomycetes

(Ascomycota),

Microbotryomycetes

(Basidiomycota),

Leotiomycetes (Ascomycota), Sordariomycetes (Ascomycota), Tremellomycetes (Basidiomycota) and Incertae sedis Ascomycota (Ascomycota), respectively (Fig. 6A). The overrepresented clades of the M area also included Choanephora (Mucorales, Zygomycota) and Fusarium (Hypocreales, Sordariomycetes), which were different from those of the F (Aureobbasidium, Dothioraceae, Dothideomycetes), G (Rhodotorula, Sporidiobolales, Microbotryomycetes; Cladosporium, Capnodiales, Dothideomycetes), H (Botrytis, Helotiales, Leotiomycetes), S (Cryptococcus, 14

ACCEPTED MANUSCRIPT Tremellales, Tremellomycetes; Hanseniaspora, Saccharomycetales, Saccharomycetes) and Y areas (Chaetopyrena, Ascomycota). Ascomycota Sequence reads of Ascomycota could be classified into nine classes, including 24 orders, 51 families and 104 genera, in addition to some unknown classes. The

RI PT

Dothideomycetes, Leotiomycetes, Saccharomycetes, Sordariomycetes, Incertae sedis Ascomycota and Eurotiomycetes classes were present together in six areas. Among the classes, Dothideomycetes represented the Ascomycota in all of the samples. Of course, other classes with less richness on grape epidermises were also incompletely

SC

present in all cultivation regions. For example, both the Orbiliomycetes and Pezizomycetes classes were present in the G and M areas, but only the

NU

Taphrinomycetes class was uniquely shown in the S area (Fig. 7B). The mean relative abundance of Leotiomycetes in the H area (29.55%; Fig. 6A) was significantly higher

MA

than that in the F (7.59%), G (8.04%), M (1.97%), S (2.43%) and Y areas (11.19%) (Amova, p < 0.001), as was that of the Saccharomycetes (S = 6.11%, M = 0.09%, F =

abundance

of

the

ED

0.12%, H = 0.22%, Y = 0.05%, G = 0.07%; p < 0.05). In contrast, the mean relative Sordariomycetes,

Incertae

sedis

Ascomycotawas

and

CE PT

Eurotiomycetes was considerably lower in each area. The dominating Dothideomycetes orders were Dothideales, Pleosporales and Capnodiales. The abundance of Capnodiales and Dothideales in the F, G and M areas

AC

were significantly higher than that in the H, S and Y areas, which was in contrast to Pleosporales. Helotiales and Erysiphales were the predominant Leotiomycetes orders, while Saccharomycetales and Hypocreales separately dominated Saccharomycetes and Sordariomycetes. There were 4, 5, 5, 5, 4 and 5 orders in the F, G, H, M, S and Y areas, respectively. The relative abundance of the prevalent orders in different grape epidermises was extremely distinct. For example, Helotiales was found in each area, but had 29.54% abundance in H, followed by lower values in the Y (11.18%), G (7.85%), F (7.59%), S (2.41%) and M (1.87%) areas.

15

ACCEPTED MANUSCRIPT There were 4, 5, 6, 5, 5 and 6 dominating families in the F, G, H, M, S and Y areas, respectively. The relative abundance of the same family varied greatly for different grape surfaces (Fig. 8). For example, the mean relative abundance of Dothioraceae (Dothideales) was significantly higher in the F (50.26%), G (45.33%) and M (36.77%) areas than in the H (14.01%), S (13.02%), and Y (8.28%) areas (p <

RI PT

0.05; Fig. 8A and Supplementary Fig. S7D), which was in contrast to that of Incertae sedis Pleosporales (Pleosporales; H = 5.07%, Y = 4.31%, S = 3.95%, M = 0.66%, F = 0.46%, G = 0.20%; Fig. 8A).

SC

Basidiomycota

NU

Seven classes were identified, and 4, 7, 3, 4, 5 and 5 classes were separately found in the F, G, H, M, S and Y areas. Basidiomycota was mainly represented the by

MA

Microbotryomycetes and Tremellomycetes classes, which were completely present in each region (Fig. 6A). Atractiellomycetes, Cystobasidiomycetes, Exobasidiomycetes, Ustilaginomycetes, and Incertae sedis Basidiomycota were of extremely low

ED

abundance. The Microbotryomycetes class could be classified into three families belonging to two orders and some unclassified groups. Similarly, Tremellomycetes represented

by

four

CE PT

was

orders

and

four

families.

The

predominant

Microbotryomycetes order was Sporidiobolales, and its relative abundance was significantly higher in the G area (11.70%) than in the S (1.32%), Y (0.88%), M (0.55%), F (0.46%), and H (0.44%) areas (P < 0.05, Supplementary Figs. S7E).

AC

Tremellales, dominating the Tremellomycetes order, was obviously richer in the S (15.80%) area than in the Y (6.58%), H (0.79%), M (0.38%), G (0.37%) and F (0.13%) areas (P < 0.05, Supplementary Figs. S7F). Sporidiobolales and Tremellales were respectively controlled by Incertae sedis Sporidiobolales and Incertae sedis Tremellales, and their abundance distribution corresponding with that of class (Fig. 6A). 3.4. Cultivars and environmental conditions shaping microbial community

16

ACCEPTED MANUSCRIPT To elucidate the relationship between growing region, cultivar, climate, and microbial biogeography, climatic and geographic data for the six growing regions in 2016 were collected from the Statistical Yearbook of the Xinjiang Uygur Autonomous Region, Meteorological Data Center of China Meteorological Administration (http://data.cma.cn)

and

the

Altitude

Information

Inquiry

Network

M

S

RI PT

Y

175 10.64

216 97.84

184 42.58

SC

(http://haiba.qhdi.com) (Table 2).

462 86.22 44.3 19.97

381 90.12 42.87 27.17

1059 86.57 42.06 21.47

Table 2. The environment conditions of collecting sample site. G

H

Frost-free Dryness

172 8.32

181 5.92

187 41.35

Altitude Longitude Latitude Average temperature

552 87.99 44.16 21.6

800 80.41 44.21 21.69

1094 86.86 42.27 20.22

NU

F

MA

Climatic conditions

The former analysis showed that alpha diversity might be correlated with environmental factors in the six regions. Then, Spearman correlation analysis was

ED

used to further investigate the correlation between microbial species richness (alpha diversity) and environmental factors. The ACE, chao1, observed species and Goods

CE PT

coverage indices of bacteria were significantly affected by altitude (A), frost-free (FR) period and longitude (LO). LO was negatively correlated with the ACE, chao1 and observed species indices, while similarly the Goods coverage index was also correlated with A. In contrast, the ACE, chao1 and observed species indices were

AC

significantly correlated with A, whereas the Goods coverage index was negatively correlated with A. Regarding the other indices, the correlation was not significant (Fig. 9A).

The relationship between bacterial community composition and environmental factors in each area was explored using canonical correspondence analysis (CCA). The CCA indicated that A, LA and LO showed significant relationships with bacterial taxonomic composition (Fig. 10A). The first two axes explained 87.85% of the taxonomic information. Further variance partitioning of the CCA suggested that env2 17

ACCEPTED MANUSCRIPT (including the A, V, LA and LO factors) was the most important factor affecting bacterial taxonomic composition. Furthermore, variance partitioning of the CCA showed that 15.16% of the total variability in the bacterial community composition was explained by the environmental variables (Fig. 11A). Similarly, alpha diversity of fungi was also affected by these factors. For example,

RI PT

LO had a negative effect on the ACE, chao1 and observed species indices, in contrast to the Shannon, Simpson and Goods coverage indices. However, degree of dryness (D) was positively correlated with the Shannon, Simpson and Goods coverage indices (Fig. 9B). The canonical correspondence analysis (CCA) also showed that the A, D,

SC

FR, LA and LO factors had significant relationships with fungal taxonomic composition (Fig. 10B). The first two axes explained 59.88% of the taxonomic

NU

information. In addition, further variance partitioning of the CCA suggested that env2 (including the A, V, LA and LO factors) may be the most important factor affecting

MA

fungal taxonomic composition. Furthermore, variance partitioning of the CCA showed that 34.73% of the total variability in the fungal community composition was

ED

explained by the environmental variables (Fig. 11B). Likewise, a Mantel test indicated that the fungal community composition was significantly correlated to A,

CE PT

LO, LA and V rather than D, FR and T, which was consistent with the overlapping areas in the variance partitioning analysis.

4. Discussion

AC

Microbial community on wine-grape surface has been deep studied due to its importance for wine characteristic, style and quality, and it is affected by various factors, such as geography, climate and viticultural practices (e.g. herbicides, fertilizers, pesticides, fungicides). Even in the Uighur Autonomous Region of Xinjiang, one-sixth of the total area of China, one of the main winemaking regions in China, microbial community on wine-grape surface is different, causing different wine style in each wine sub-region. In this study, microbial community composition on wine-grape surface have been studied with 48 wine-grape samples, originating from six different regions of Xinjiang, as well as the associations between geography 18

ACCEPTED MANUSCRIPT and microbial community have been further explored. Microbial community on wine-grape surface has been heavily report in previous study. Portillo Mdel et al. (2016) reported bacterial diversity on the surface of two grape varieties (Grenache and Carignan) grape berries based on high-throughput sequencing and detected up to 14 bacterial phyla with Firmicutes, Proteobacteria and

RI PT

Actinobacteria being the most abundant. Moreover, it was demonstrated that the bacterial communities on grape surface were predominated by Proteobacteria followed by Firmicutes, Actinobacteria, Acidobacteria and Bacteroidetes (Catia Pinto, 2014; Portillo & Mas, 2016). Similarly, five fungal phyla with Ascomycota,

SC

Basidiomycota, Chytridiomycota, Glomeromycota and Blastocladiomycota was recovered on sauvignon blanc grape berry from the Marlborough region on the South

NU

Island of New Zealand by 454-sequencing, and Basidiomycota was the most diverse and abundant, followed by Ascomycota (Morrison-Whittle, Lee, & Goddard, 2017).

MA

In this study, sixteen bacterial phyla have been detected in each sample, while three fungal phyla including Ascomycota, Basidiomycota and Zygomycota have also been

ED

found on the surface of grape berry. Proteobacteria (98.55%) and Ascomycota (70%) were the most abundant bacterial and fungal phylum, respectively. Furthermore, the

CE PT

bacterial community was mainly dominated by Pantoea followed by Pseudomor in this research, and the relative abundance of mainly bacteria genus had differences in six wines of Xinjiang. Similarly, the bacteria of Pantoea and Pseudomor genus were reported in Vitis vinifera phyllosphere (Singh et al., 2018). The top ten of fungal

AC

genera were Aureobasidum, Alternaria, Botrytis, Rhodotorula, Cryptococcus, Hanseniaspora, Mucor, Fusarium, Chaetopyrena and Cladosporium, which was previously reviewed in other study (Barata, Malfeito-Ferreira, & Loureiro, 2012; Morgan, du Toit, & Setati, 2017). Overall, it was obvious that the bacterial and fungal community composition on grape surface showed differences in six wine region of Xinjiang, resulting from geography and environment of growing grapevine. Epiphytic microorganisms on wine-grape surface play a vital role in grape health and quality, which resolutely affect the winemaking process. It was reported that Aureobbasidium pullulans or Sporidiobolus pararoseus on grape berries can act as 19

ACCEPTED MANUSCRIPT effective indigenous antagonists against microorganisms known to have negative effects on wine quality like Botrytis cinerea (Kecskemeti, Berkelmann-Lohnertz, & Reineke, 2016). Botrytis cinerea are devastating the fungal pathogens of grapevines (Haidar et al., 2016). Particularly, A. pullulans and C. allicinum were found on grape berry from Fukang region, but B. cinerea grew on samples of Heshuo region in this

RI PT

study (Fig. 7). This result might comfirm that the quality of Fukang’s grape berries were superior to Heshuo in 2016. Apart from this, some species with significant difference were found in other regions, such as Manasi (Choanephora infundibulifera and Fusarium proliferatum), Shanshan (Cryptococcus albidus, Hanseniaspora and

Alternaria

alternate),

Huoerguosi

(Rhodotorula

SC

opuntia

glutinis

and

Cladosporium cladosporioides) and Yanqi (Chaetopyrena penicillate). Usually, starter

NU

cultures of Saccharomyces cerevisiae are used in wine fermentation, but it is not detected in this study. This might be explained by previous study that the microbial

MA

community on the grape berry surface changed with the ripening of grape, especially fermentative yeasts were detected at harvest and not in the first stage of grape growth

ED

(Renouf, Claisse, & Lonvaud-Funel, 2005). The yeast of R. glutinis as a source of pigments and carotenoids might have an effect on wine color and nutrition (Aksu &

CE PT

Eren, 2007; Hernández-Almanza et al., 2014). Consequently, the wine of Huoerguosi region might have more unique color and nutrition than one of other regions. However, it was rarely reported that these fungal species including C. infundibulifera, F. proliferatum, A. alternate, C. cladosporioides and C. penicillate were found on

AC

grape surface. Thus, it is still not clear whether these fungi have effect on grape health or winemaking. Considering bacteria, Lactic acid bacteria like Oenoccocus and Pediococcus genus have been confirmed contributing to wine quality (Liu et al., 2017), but these are not detected in this study. Yet, Pantoea and Pseudomor, being able to induce systemic resistance in grapevine against B. cinerea (Aziz et al., 2015), were found as dominated bacterial genus on the surface of wine grape in this study. It has been frequently reported that microbial community on wine-grape surface is dependent on several factors, such as geography (Gayevskiy & Goddard, 2012), climate (Bokulich, Thorngate, Richardson, & Mills, 2014; Hannah et al., 2013; 20

ACCEPTED MANUSCRIPT Orduña, 2010) and viticultural practices (e.g. herbicides, fertilizers, pesticides, fungicides) (Cadez, Zupan, & Raspor, 2010; Canfora et al., 2018; Garcia et al., 2018). In this study, bacterial community composition showed significant relationship with altitude, latitude and longitude, while fungi had significant relationship with altitude, dryness, frost-free period, latitude and longitude (Fig. 10). It seems that microbial

RI PT

community was affected by geographic site of different wine region in Xinjiang. For example, microbial community composition in Heshuo and Yanqi regions is different from one in other regions, owing to the location in the south of Xinjiang, closing to the Tarim Basin and Bosten Lake, having higher altitude. Recent study also reported

SC

that bacterial diversity on grape surface was influenced by geographic patterns (Vitulo et al., 2018). However, the CCA results only explained 87.85% and 59.88%, which

NU

might verify previous study that microbial community composition on the surface of wine grape was influenced by other factors, such as grape varieties, vineyard

MA

management and viticultural practices (Cordero-Bueso, Arroyo, Serrano, Tello, et al., 2011; Cordero-Bueso, Arroyo, Serrano, & Valero, 2011; Masoni et al., 2017; Torrado-Agrasar,

Gonzalez-Barreiro,

Cancho-Grande,

&

ED

Noguerol-Pato,

Simal-Gandara, 2014). Additionally, it is further visible that the geographic factors

CE PT

(Altitude, Latitude, and Longitude) have a stronger effect on microbial community composition of grape surface than climatic factors (Dryness, Frost-free, temperature), according to variance partitioning analysis in this study. In the future, to accurately investigate the microbial diversity on wine-grape

AC

surface, a method used to indentify microbiota in species level is being looked forward to except currently cultivable and non-cultivable technologies. Further research is required to precisely reveal the correlation between the diversity on wine-grape surface and some factors, such as geography, climate, grape varieties, vintage, vineyard management and viticultural practices et al, determining which one is the most important factors. In addition, we need to expand sample sizes and sampling points to contribute to establishing microbial community fingerprint on wine-grape surface in different wine sub-regions of Xinjiang. 21

ACCEPTED MANUSCRIPT 5. Conclusions Our findings indicated that the microbial community structure on wine grape epidermises from different regions of Xinjiang was affected by environment factors. Dominant microbial community were ascertained by different relative abundance based on wine regions. Furthermore, the complexity of fungal community

RI PT

composition was obviously higher than that of bacterial community composition in all samples. The correlation between the microbial community structure and environment factors can be used to evaluate microbe resources on wine grape surface, regulate

SC

winemaking process conditions and show wine region characteristics based on different wine regions in Xinjiang. In conclusion, we clarified the relationships

NU

between the microbial communities and environmental factors, providing evidence that environmental factors strongly affect the microbial community composition on

MA

wine grape surface in Xinjiang.

ED

Acknowledgements

This work was supported by the National Natural Science Foundations of China (No.

CE PT

31560446 and 31500092), and the Project of Science and Technology, Xinjiang Production and Construction Corps (No. 2015AB016, 2016AB009 and 2016AD024). Author Contributions

F. Gao collected wine grape samples and environmental data, performed DNA

AC

extractions and PCRs, conducted the bioinformatic analysis of the Illumina sequencing data, compiled the figures and tables, and wrote the complete the manuscript. B. Wang and X. Shi. contributed advice and constructive critiques, reviewed the results and corrected language style in the final manuscript. J. Chen, J. Xiao, W. Cheng and X. Zheng together supervised and guided the research project. All authors reviewed the manuscript. Competing Interests: The authors declare that they have no competing interests.

22

ACCEPTED MANUSCRIPT References Aksu, Z., & Eren, A. T. (2007). Production of carotenoids by the isolated yeast of Rhodotorula glutinis. Biochemical Engineering Journal, 35 (2), 107-113. Aziz, A., Verhagen, B., Magnin-Robert, M., Couderchet, M., Clément, C., Jeandet, P., & Trotel-Aziz, P. (2015). Effectiveness of beneficial bacteria to promote systemic resistance of grapevine to gray mold as related to phytoalexin production in vineyards. Plant and Soil, 405 (1-2), 141-153. Barata, A., Malfeito-Ferreira, M., & Loureiro, V. (2012). The microbial ecology of wine grape berries.

RI PT

Int J Food Microbiol, 153 (3), 243-259. Bokulich, N. A., Thorngated, J. H., Richardsone, P. M., & Millsa, D. A. (2014). Microbial biogeography of wine grapes is conditioned by cultivar, vintage, and climate. Proceedings of the National Academy of Sciences, 111, E139-E148.

Cadez, N., Zupan, J., & Raspor, P. (2010). The effect of fungicides on yeast communities associated

SC

with grape berries. FEMS Yeast Res, 10 (5), 619-630.

Canfora, L., Vendramin, E., Felici, B., Tarricone, L., Florio, A., & Benedetti, A. (2018). Vineyard microbiome variations during different fertilisation practices revealed by 16s rRNA gene

NU

sequencing. Applied Soil Ecology, 125, 71-80.

Castro, H. F., Classen, A. T., Austin, E. E., Norby, R. J., & Schadt, C. W. (2010). Soil microbial Microbiology, 76 (4), 9.

MA

community responses to multiple experimental climate change drivers. American Society for Catia Pinto, D. P., Susana Sousa, Miguel Pinheiro, Conceic ˜o Egas, Ana C. Gomes. (2014). Unravelling the Diversity of Grapevine Microbiome. PLoS One, 9 (1), 12. Christiaens, J. F., Franco, L. M., Cools, T. L., De Meester, L., Michiels, J., Wenseleers, T., Hassan, B.

ED

A., Yaksi, E., & Verstrepen, K. J. (2014). The fungal aroma gene ATF1 promotes dispersal of yeast cells through insect vectors. Cell Rep, 9 (2), 425-432. Ciani, M., & Comitini, F. (2010). Non-Saccharomyces wine yeasts have a promising role in

CE PT

biotechnological approaches to winemaking. Annals of Microbiology, 61 (1), 25-32. Comtini, F., & Ciani, M. (2008). Influence of fungicide treatments on the occurrence of yeast flora associated with wine grapes. Annals of Microbiology, 58 (3), 5. Cordero-Bueso, G., Arroyo, T., Serrano, A., Tello, J., Aporta, I., Velez, M. D., & Valero, E. (2011). Influence of the farming system and vine variety on yeast communities associated with grape

AC

berries. Int J Food Microbiol, 145 (1), 132-139. Cordero-Bueso, G., Arroyo, T., Serrano, A., & Valero, E. (2011). Influence of different floor management strategies of the vineyard on the natural yeast population associated with grape berries. Int J Food Microbiol, 148 (1), 23-29. El-Ashram, S., & Suo, X. (2017). Exploring the microbial community (microflora) associated with ovine Haemonchus contortus (macroflora) field strains. Science Reports, 7 (1), 1-13. Garbeva, P., van Veen, J. A., & van Elsas, J. D. (2004). Microbial diversity in soil: selection microbial populations by plant and soil type and implications for disease suppressiveness. Annu Rev Phytopathol, 42, 243-270. Garcia, L., Celette, F., Gary, C., Ripoche, A., Valdés-Gómez, H., & Metay, A. (2018). Management of service crops for the provision of ecosystem services in vineyards: A review. Agriculture, Ecosystems & Environment, 251, 158-170. Gayevskiy, V., & Goddard, M. R. (2012). Geographic delineations of yeast communities and 23

ACCEPTED MANUSCRIPT populations associated with vines and wines in New Zealand. ISME J, 6 (7), 1281-1290. Goldberg, C. S., Turner, C. R., Deiner, K., Klymus, K. E., Thomsen, P. F., Murphy, M. A., Spear, S. F., McKee, A., Oyler-McCance, S. J., Cornman, R. S., Laramie, M. B., Mahon, A. R., Lance, R. F., Pilliod, D. S., Strickler, K. M., Waits, L. P., Fremier, A. K., Takahara, T., Herder, J. E., Taberlet, P., & Gilbert, M. (2016). Critical considerations for the application of environmental DNA methods to detect aquatic species. Methods in Ecology and Evolution, 7 (11), 1299-1307. Guzzon, R., Gugole, S., Zanzotti, R., Malacarne, M., Larcher, R., von Wallbrunn, C., & Mescalchin, E. (2016). Evaluation of the oenological suitability of grapes grown using biodynamic

RI PT

agriculture: the case of a bad vintage. J Appl Microbiol, 120 (2), 355-365.

Haidar, R., Deschamps, A., Roudet, J., Calvo-Garrido, C., Bruez, E., Rey, P., & Fermaud, M. (2016). Multi-organ screening of efficient bacterial control agents against two major pathogens of grapevine. Biological Control, 92, 55-65.

SC

Hannah, L., Roehrdanz, P. R., Ikegami, M., Shepard, A. V., Shaw, M. R., Tabor, G., Zhi, L., Marquet, P. A., & Hijmans, R. J. (2013). Climate change, wine, and conservation. Proceedings of the National Academy of Sciences, 110 (17), 6907-6912. Taxus rhizosphere

via

NU

Hao, D. C., Song, S. M., Mu, J., Hu, W. L., & Xiao, P. G. (2016). Unearthing microbial diversity of MiSeq high-throughput

amplicon

sequencing and isolate

characterization. Science Reports, 6, 1-14.

MA

Hayden, C. J., & Beman, J. M. (2016). Microbial diversity and community structure along a lake elevation gradient in Yosemite National Park, California, USA. Environ Microbiol, 18 (6), 1782-1791.

Hernández-Almanza, A., Cesar Montanez, J., Aguilar-González, M. A., Martínez-Ávila, C.,

ED

Rodríguez-Herrera, R., & Aguilar, C. N. (2014). Rhodotorula glutinis as source of pigments and metabolites for food industry. Food Bioscience, 5, 64-72. Heydenreich, F. M., Miljuš, T., Jaussi, R., Benoit, R., Milić, D., & Veprintsev, D. B. (2017).

CE PT

High-throughput mutagenesis using a two-fragment PCR approach. Sci Rep, 7, 11. Hierro, N., Gonzalez, A., Mas, A., & Guillamon, J. M. (2006). Diversity and evolution of non-Saccharomyces yeast populations during wine fermentation: effect of grape ripeness and cold maceration. FEMS Yeast Res, 6, 102-111. Hollister, E. B., Schadt, C. W., Palumbo, A. V., James Ansley, R., & Boutton, T. W. (2010). Structural and functional diversity of soil bacterial and fungal communities following woody plant

AC

encroachment in the southern Great Plains. Soil Biology and Biochemistry, 42 (10), 1816-1824.

Hong, Y. A., & Park, H. D. (2013). Role of non-Saccharomyces yeasts in Korean wines produced from Campbell Early grapes: potential use of Hanseniaspora uvarum as a starter culture. Food Microbiol, 34 (1), 207-214. Horn, S., Hempel, S., Verbruggen, E., Rillig, M. C., & Caruso, T. (2017). Linking the community structure of arbuscular mycorrhizal fungi and plants: a story of interdependence? ISME J, 11 (6), 1-12. Jia, H.-R., Geng, L.-L., Li, Y.-H., Wang, Q., Diao, Q.-Y., Zhou, T., & Dai, P.-L. (2016). The effects of Bt Cry1Ie toxin on bacterial diversity in the midgut of Apis mellifera ligustica (Hymenoptera: Apidae). Sci Rep, 6, 24664. Kao, C. M., Liao, H. Y., Chien, C. C., Tseng, Y. K., Tang, P., Lin, C. E., & Chen, S. C. (2016). The 24

ACCEPTED MANUSCRIPT change of microbial community from chlorinated solvent-contaminated groundwater after biostimulation using the metagenome analysis. J Hazard Mater, 302, 144-150. Karna L. Sacchi, Linda F. Bisson, & Adams, a. D. O. (2005). A review of the effect of winemaking techniques on phenolic extraction in red wines. American Journal of Enology and Viticulture, 56 (3), 197-206. Kecskemeti, E., Berkelmann-Lohnertz, B., & Reineke, A. (2016). Are Epiphytic Microbial Communities in the Carposphere of Ripening Grape Clusters (Vitis vinifera L.) Different between Conventional, Organic, and Biodynamic Grapes? PLoS One, 11 (8), e0160852. Li, J., Hu, W., Huang, X., & Xu, Y. (2018). Investigation of yeast population diversity and dynamics in

RI PT

spontaneous fermentation of Vidal blanc icewine by traditional culture-dependent and high-throughput sequencing methods. Food Res Int, 112, 66-77.

Li, J., Jiang, X., Li, G., Jing, Z., Zhou, L., Ke, Z., & Tan, Y. (2017). Distribution of picoplankton in the northeastern South China Sea with special reference to the effects of the Kuroshio intrusion

SC

and the associated mesoscale eddies. Sci Total Environ, 589, 1-10.

Liu, J., Sui, Y., Yu, Z., Shi, Y., Chu, H., Jin, J., Liu, X., & Wang, G. (2014). High throughput sequencing analysis of biogeographical distribution of bacterial communities in the black soils of

NU

northeast China. Soil Biology and Biochemistry, 70, 113-122.

Liu, W., Zheng, Y., Kwok, L.-Y., Sun, Z., Zhang, J., Guo, Z., Hou, Q., Menhe, B., & Zhang, H. (2015). High-throughput sequencing for the detection of the bacterial and fungal diversity in

MA

Mongolian naturally fermented cow's milk in Russia. BMC Microbiol, 15 (45), 1-12. Liu, Y., Rousseaux, S., Tourdot-Marechal, R., Sadoudi, M., Gougeon, R., Schmitt-Kopplin, P., & Alexandre, H. (2017). Wine microbiome: A dynamic world of microbial interactions. Crit Rev Food Sci Nutr, 57 (4), 856-873.

ED

Marsh, A. J., O'Sullivan, O., Hill, C., Ross, R. P., & Cotter, P. D. (2013). Sequencing-based analysis of the bacterial and fungal composition of kefir grains and milks from multiple sources. PLoS One, 8 (7), 1-11.

CE PT

Martins, G., Vallance, J., Mercier, A., Albertin, W., Stamatopoulos, P., Rey, P., Lonvaud, A., & Masneuf-Pomarede, I. (2014). Influence of the farming system on the epiphytic yeasts and yeast-like fungi colonizing grape berries during the ripening process. Int J Food Microbiol, 177, 21-28.

Masoni, A., Frizzi, F., Brühl, C., Zocchi, N., Palchetti, E., Chelazzi, G., & Santini, G. (2017). Management matters: A comparison of ant assemblages in organic and conventional vineyards.

AC

Agriculture, Ecosystems & Environment, 246, 175-183. Milanovic, V., Comitini, F., & Ciani, M. (2013). Grape berry yeast communities: influence of fungicide treatments. Int J Food Microbiol, 161, 7. Mirás-Avalos, J. M., Fandiño, M., Trigo-Córdoba, E., Rey, B. J., Orriols, I., & Cancela, J. J. (2017). Influence of irrigation on consumer acceptability of Albariño and Godello wines. LWT - Food Science and Technology, 85, 345-352. Morgan, H. H., Toit, M. d., & Setati, M. E. (2017). The Grapevine and Wine Microbiome: Insights from High-Throughput Amplicon Sequencing. Front Microbiol, 8, 15. Morrison-Whittle, P., Lee, S. A., & Goddard, M. R. (2017). Fungal communities are differentially affected by conventional and biodynamic agricultural management approaches in vineyard ecosystems. Agriculture, Ecosystems & Environment, 246, 306-313. Nisiotou, A. A., Rantsiou, K., Iliopoulos, V., Cocolin, L., & Nychas, G. J. (2011). Bacterial species 25

ACCEPTED MANUSCRIPT associated with sound and Botrytis-infected grapes from a Greek vineyard. Int J Food Microbiol, 145, 432-436. Noguerol-Pato, R., Torrado-Agrasar, A., Gonzalez-Barreiro, C., Cancho-Grande, B., & Simal-Gandara, J. (2014). Influence of new generation fungicides on Saccharomyces cerevisiae growth, grape must fermentation and aroma biosynthesis. Food Chem, 146, 234-241. Nones, K., Waddell, N., Wayte, N., Patch, A. M., Bailey, P., Newell, F., Holmes, O., Fink, J. L., Quinn, M. C. J., Tang, Y. H., Lampe, G., Quek, K., Loffler, K. A., Manning, S., Idrisoglu, S., Miller, D., Xu, Q., Waddell, N., Wilson, P. J., Bruxner, T. J. C., Christ, A. N., Harliwong, I., Nourse, C., Nourbakhsh, E., Anderson, M., Kazakoff, S., Leonard, C., Wood, S., Simpson, P. T., Reid,

RI PT

L. E., Krause, L., Hussey, D. J., Watson, D. I., Lord, R. V., Nancarrow, D., Phillips, W. A., Gotley, D., Smithers, B. M., Whiteman, D. C., Hayward, N. K., Campbell, P. J., Pearson, J. V., Grimmond, S. M., & Barbour, A. P. (2014). Genomic catastrophes frequently arise in esophageal adenocarcinoma and drive tumorigenesis. Nature Communication, 5, 1-21. Food Research International, 43 (7), 1844-1855.

SC

Orduña, R. M. d. (2010). Climate change associated effects on grape and wine quality and production. Portillo Mdel, C., Franques, J., Araque, I., Reguant, C., & Bordons, A. (2016). Bacterial diversity of

NU

Grenache and Carignan grape surface from different vineyards at Priorat wine region (Catalonia, Spain). Int J Food Microbiol, 219, 56-63.

Portillo, M. d. C., & Mas, A. (2016). Analysis of microbial diversity and dynamics during wine

MA

fermentation of Grenache grape variety by high-throughput barcoding sequencing. LWT Food Science and Technology, 72, 317-321.

Renouf, V., Claisse, O., & Lonvaud-Funel, A. (2005). Understanding the microbial ecosystem on the grape berry surface through numeration and identification of yeast and bacteria. Australian

ED

Journal of Grape and Wine Research, 11, 12.

Romano, P., Fiore, C., Paraggio, M., Caruso, M., & Capece, A. (2003). Function of yeast species and strains in wine flavour. Int J Food Microbiol, 86, 12.

CE PT

Segata, N., Izard, J., Waldron, L., Gevers, D., Miropolsky, L., Garrett, W. S., & Huttenhower, C. (2011). Metagenomic biomarker discovery and explanation. Genome Biology, 12, 1-18. Singh, P., Gobbi, A., Santoni, S., Hansen, L. H., This, P., & Péros, J.-P. (2018). Assessing the impact of plant genetic diversity in shaping the microbial community structure of Vitis vinifera phyllosphere in the Mediterranean. Frontiers in Life Science, 11 (1), 35-46. Sterkenburg, E., Bahr, A., Durling, M. B., Clemmensen, K. E., & Lindahl, B. D. (2015). Changes in

AC

fungal communities along a boreal forest soil fertility gradient. New Phytologist, 207 (4), 1145-1158.

Ubeda, J., Maldonado Gil, M., Chiva, R., Guillamon, J. M., & Briones, A. (2014). Biodiversity of non-Saccharomyces yeasts in distilleries of the La Mancha region (Spain). FEMS Yeast Res, 14 (4), 663-673. Verginer, M., Leitner, E., & Berg, G. (2010). Production of volatile metabolites by grape-associated microorganisms. J Agric Food Chem, 58 (14), 8344-8350. Vitulo, N., Jr., W. J. F. L., Calgaro, M., Confalone, M., Felis, G. E., Zapparoli, G., & Nardi, T. (2018). Bark and Grape Microbiome of Vitis vinifera: Influence of Geographic Patterns and Agronomic Management on Bacterial Diversity. Front Microbiol, 9, 12. Wallenstein, M. D., & Hall, E. K. (2011). A trait-based framework for predicting when and where microbial adaptation to climate change will affect ecosystem functioning. Biogeochemistry, 26

ACCEPTED MANUSCRIPT 109, 13. Xiao, E., Krumins, V., Xiao, T., Dong, Y., Tang, S., Ning, Z., Huang, Z., & Sun, W. (2017). Depth-resolved microbial community analyses in two contrasting soil cores contaminated by antimony and arsenic. Environ Pollut, 221, 244-255. Xie, X., Zhang, H., & Paul, P. (2009). Sustained growth promotion in Arabidopsis with long-term exposure to the beneficial soil bacterium Bacillus subtilis(GB03). Plant Signaling & Behavior, 4 (10), 948-953. Yan, D., Zhang, T., Su, J., Zhao, L. L., Wang, H., Fang, X. M., Zhang, Y. Q., Liu, H. Y., & Yu, L. Y. (2016). Diversity and Composition of Airborne Fungal Community Associated with

RI PT

Particulate Matters in Beijing during Haze and Non-haze Days. Front Microbiology, 7, 1-12. Yang, J., Ma, L., Jiang, H., Wu, G., & Dong, H. (2016). Salinity shapes microbial diversity and community structure in surface sediments of the Qinghai-Tibetan Lakes. Sci Rep, 6, 1-6. Zetsche, B., Heidenreich, M., Mohanraju, P., Fedorova, I., Kneppers, J., DeGennaro, E. M., Winblad,

SC

N., Choudhury, S. R., Abudayyeh, O. O., Gootenberg, J. S., Wu, W. Y., Scott, D. A., Severinov, K., van der Oost, J., & Zhang, F. (2017). Multiplex gene editing by CRISPR-Cpf1 using a single crRNA array. Nat Biotechnol, 35 (1), 6.

NU

Zhai, Y., Wang, W., Tan, H., & Cao, L. (2016). A New Approach to Analyzing Endophytic Actinobacterial Population in the Roots of Banana Plants (Musa sp., AAA). Journal of Biochemistry and Molecular Biology Research, 2 (3), 180-184.

MA

Zhang, S., Chen, X., Zhong, Q., Huang, Z., & Bai, Z. (2017). Relations among epiphytic microbial communities from soil, leaves and grapes of the grapevine. Frontiers in Life Science, 10(1),

ED

73-83.

AC

CE PT

Figure:

Figure 1. Comparison of microbial community on the surface of wine grape from different cultivation regions. Principal Coordinate Analysis (PCoA) based on unweighted unifrac distance was generated with OTUs (at 97% similarity) present in the different cultivation areas samples. (A) bacteria, (B) fungi. Principal Coordinate 27

ACCEPTED MANUSCRIPT

MA

NU

SC

RI PT

(PC). Different color shapes represent samples of different cultivation areas.

Figure 2. The common and exclusive bacterial OTUs on the surface of same

ED

grape variety from different wine regions. A: Cabernet Sauvignon; B: Merlot; C:

AC

CE PT

Italic Riesling; D: Cabenet Franc.

28

ACCEPTED MANUSCRIPT

NU

SC

RI PT

Figure 3. The relative abundance of bacteria at class (A/a) and genus (B/b) level.

Figure 4. The frequency distribution of bacterial family in six areas. The different

AC

CE PT

ED

MA

color intensities represent the relative bacterial abundance in each group.

29

ED

MA

NU

SC

RI PT

ACCEPTED MANUSCRIPT

CE PT

Figure 5. The common and exclusive fungal OTUs on the surface of same grape variety from different wine regions. A: Cabernet Sauvignon; B: Merlot; C: Riesling

AC

Italico; D: Cabenet Franc.

30

NU

SC

RI PT

ACCEPTED MANUSCRIPT

AC

CE PT

ED

MA

Figure 6. The relative abundance of fungi at class (A/a) and genus (B/b) level.

Figure 7. LEfSe results showing significant fungi on the surface of wine grape. The cladogram reports the taxonomic representation of statistically and biologically consistent differences between F, G, H, M, S and Y fungal communities. 31

SC

RI PT

ACCEPTED MANUSCRIPT

Figure 8. Heat maps showing fungal family frequency distribution in six areas.

CE PT

ED

MA

NU

The different color intensities represent the relative bacterial abundance in each groups.

AC

Figure 9. Spearman correlation analysis studying correlation between microbial species richness (alpha diversity) and environment factors. The different color intensities represent species richness. A, bacteria; B, fungi. *, significant; **, extremely significant. Frost-free period, FR; Degree of dryness, D; Altitude, A; Degree of longitude, LO; Degree of latitude, LA; Average temperature, T; Cultivar, V.

32

RI PT

ACCEPTED MANUSCRIPT

Figure 10. CCA analysis based on bacterial (A) and fungal (B) OTUs. The different shapes represent the sample groups in different environments or conditions;

SC

the arrows indicate the environmental and climatic factors; the angle between the species and the environmental and climatic factor represents the positive and negative

NU

correlation between the species and the environmental and climatic factors (acute

CE PT

ED

MA

angle: positive correlation; obtuse: negative correlation; right angle: no correlation).

Figure 11. Venn diagram of the variance partitioning analysis shows the relative

AC

effects of multiple variables on the composition of bacterial (A) and fungal (B) taxa. The impact factors of community composition are made up of env1 (including D, FR and T factors) and env2 (including A, V, LA and LO factors). The areas correspond to the amount of variance explained by each factor. Overlapping areas indicate shared variation of the parameter effect on community composition (A).

33

ACCEPTED MANUSCRIPT Highlights: 1. The microbial community structure on wine grape epidermises from different regions of Xinjiang was affected by environment factors. 2. Proteobacteria and Ascomycota were the predominant bacteria and fungi,

RI PT

respectively. 3. The complexity of fungal community composition was obviously higher than that of bacterial community composition in all samples.

4. Bacterial community diversity was largely related to altitude, latitude and

SC

longitude.

5. Fungal community diversity was closely related to altitude, dryness, frost-free

AC

CE PT

ED

MA

NU

period, latitude and longitude.

34