The isolation and identification of Bacillus velezensis ZN-S10 from vanilla (V. planifolia), and the microbial distribution after the curing process

The market value of vanilla beans (Vanilla planifolia) is constantly increasing due to their natural aroma and flavor properties that improve after a curing process, where bacteria colonization plays a critical role. However, a few publications suggest that bacteria play a role in the curing process. Hence, this study aimed to isolate Bacillus sp. that could be used for fermenting V. planifolia while analyzing their role in the curing process. Bacillus velezensis ZN-S10 identified with 16S rRNA sequencing was isolated from conventionally cured V. planifolia beans. A bacteria culture solution of B. velezensis ZN-S10 (1 mL of 1 × 107 CFU mL−1) was then coated on 1 kg of non-cured vanilla pods that was found to ferment and colonize vanilla. PCA results revealed distinguished bacterial communities of fermented vanilla and the control group, suggesting colonization of vanilla. Phylogenetic analysis showed that ZN-S10 was the dominant Bacillus genus member and narrowly correlated to B. velezensis EM-1 and B. velezensis PMC206-1, with 78% and 73% similarity, respectively. The bacterial taxonomic profiling of cured V. planifolia had a significant relative abundance of Firmicutes, Proteobacteria, Cyanobacteria, Planctomycetes, and Bacteroidetes phyla according to the predominance. Firmicutes accounted for 55% of the total bacterial sequences, suggesting their colonization and effective fermentation roles in curing vanilla.


Sample preparation
Green, matured vanilla (V.planifolia) bean samples were collected from a local farm in Pingtung, Taiwan (22°25′41.8″N120°32′29.0″E).A 1 kg vanilla bean pod pack was washed with sterilized water and then blanched in hot water (~ 80 °C) for 3 min.The blanched vanilla pods were instantly placed in cooling boxes with paper tissues to let the pods dry out.According to Xu et al. 18 , this "killing" stage inhibits the vegetative growth of the beans while allowing enzyme activation and initiating color change (from yellow/green to dark brown) and the flavor or aroma development of the final cured vanilla bean.

Curing of vanilla bean pods
The blanched vanilla beans were allowed to naturally ferment in a cool freezer; wherein the samples were stored in plastic boxes and wrapped in tissue paper and stored at 10 ℃.It should be noted that the tissue paper was changed every 2 days to avoid mold formation.Occasionally, cool-air drying and humidification were also done once a week with a lamina flow and a humidifier, respectively.Additionally, as the beans dried and showed yellow to brown color changes owing to phenol browning and lipid oxidation that occur simultaneously with the enzymatic hydrolysis of vanillin precursors 22 .Thus, brown to dark brown vanilla bean pods without molds were recognized as the best-quality cured samples and were used for experimentation.

Bacteria isolation and purification
A batch of the brown-dry cured vanilla bean pods was cut into 0.3-0.5 cm pieces and ground into 0.5 g powder.Sterile distilled water was then added in 15 mL tubes with the vanilla powder, then 20 min agitation in ultrasound (Ultrasonic Bath Delta Model DC 150-H, Takashi, Japan) and stored at 20 ℃ overnight, following the method of Chen et al. 28 .A de Man, Rogosa, and Sharpe (MRS) agar medium 35 were modified and prepared on Petri dishes, maintained at pH 5.7, and autoclaved for 15 min.The altered and selective MRS media was composed of dextrose of 15 g L −1 , yeast extract (5 g L −1 ), 5 g L −1 of sodium acetate, potassium phosphate of 2 g L −1 , magnesium sulfate of 0.1 g L −1 , and 1.00 mL drop of Tween 80.A supernatant of 100 µL was inoculated on MRS agar and then spread on the Petri dishes for oven incubation at 37 ℃ for 16 h.After the incubation, the bacteria were selected after three consecutive plates according to shape, color, and size to obtain pure isolates.The re-streaking method on the MRS agar was used as a purification technique for the bacteria colonies.Hence, the pure isolates were selected for identification and were stored in 20% glycerol at − 80 ℃ for further experiments.

Hemolysis test
The blood agar test was conducted with pure collected bacterial isolates to observe the safety of using these microbes for the food industry and animal use.The safety assessment was based hemolytic reaction recognized by lyses of blood cells on the blood agar 36 , hence the isolated bacteria colonies were tested with the hemolysis assay.In 1 L of distilled water, 40 g of a commercial agar-based medium (Infusion agar) (Himedia Laboratories Co., India), was suspended and then heated to boil at 121 ℃ at 15 lbs pressure.The agar medium was then allowed to cool and 5% v/v of sterile sheep-defibrinated blood was added and mixed well before pouring into Petri plates.The pure isolate bacteria were then streaked on these blood agar plates and then incubated at 37 ℃ for 16 h to observe the alpha, beta, and gamma properties of the bacteria.According to studies; Staphylococcus sp., Streptococcus sp., and Enterococcus sp. can be differentiated by the degree of hemolysis caused by their hemolysis 37,38 .Hence, the bacteria with beta (β) hemolysis 39 had been classified as complete hemolysis seen with the formation of a clear-transparent zone, and the plate with those properties was discarded in this study.Partial hemolysis was recognized with opaque-green zone formation on the blood agar plates, thus was also discarded in this study.The agar plates with notable growth on the blood agar medium characterized as gamma (γ) hemolysis were the targeted bacteria colonies and were used for identification and subsequent experiments of interest in this research.However, the bacteria with alpha (α) hemolysis were observed with partial hemolysis showing rupture or "digestion" with a lightened (yellow) and transparent area on the blood agar medium were discarded.Hence, based on the above-mentioned literature, this study performed a visual assessment of the hemolysis (α, β, γ) for food safety purposes.

Identification of the bacteria
A 16S rRNA gene sequencing technique 40 was used for the identification and classification of the pure bacteria isolates.The 16S rRNA molecular sequencing identification was performed by Mission Biotech Company (Taiwan), and arranged with ABI Big Dye Terminator v3.1.PCR mixtures were amplified by initially holding at 94 °C for 5 min.Besides, 40 cycles of denaturing at 94 °C for 30 s, annealing at 55 °C for 30 s, and extensions at 72 °C for 2 min 20 s.Then the reaction ended with a final extension at 72 °C for 5 min and a hold at 4 °C.In this study, the amplification of the 16S rRNA was F8 (5ʹ-AGA GTT TGATCMTGG CTC AG-3ʹ) and R1510 (5ʹ-CGG TTA CCT TGT TAC GAC TT-3ʹ) 41 primers, respectively.It should also be noted that the nucleotide sequences are obtainable on the National Center for Biotechnological Information (NCBI) website where the Basic Local Alignment Search Tool (BLAST) was used for phylogenetic analysis.

Phylogenetic tree analysis based on 16S rRNA gene sequences
The evolutionary data of the phylogenetic tree was analyzed using the Neighbor-Joining method 42,43 .With branch lengths in the equal units as those of the evolutionary distances used to deduce the phylogenetic tree, the ideal tree was drawn to scale (2.00).The evolutionary distances are expressed in base substitutions per site and were calculated using the Maximum Composite Likelihood approach 44 .Beside each internal node in the tree is the percentage of sites where at least 1 unambiguous base is present in at least 1 sequence for each descendent clade is shown next to each internal node in the tree.This analysis involved 10 nucleotide sequences.All ambiguous positions were removed for each sequence pair (pairwise deletion option).There was a total of 96 positions in the final dataset.Evolutionary analyses were conducted and extracted NCBI 16S rDNA sequences from the aforementioned genome sequences and aligned them using MEGA11 45 .The sequences were subjected prior to the NCBI database (https:// www.ncbi.nlm.nih.gov/) for a FASTA search as well as obtaining the accession numbers according to the sequences.

Fermentation of V. planifolia
A new batch of green matured vanilla bean pods (non-blanched) was sorted according to sizes of approximately 15 cm, considered the best quality.The pods (1 kg) were blanched for 2 min in hot water (~ 80 ℃) and then dried for 3 h.After the quick drying, the 1 kg bean pods were placed in a plastic box.Then liquid Bacillus velezensis ZN-S10 bacteria culture of 1 mL of 1 × 10 7 CFU mL −1 was sprayed on the vanilla pods and the plastic box was shaken to allow the bacteria to colonize and ferment the vanilla pods.The UV-Vis spectrometer (DU 640, Micro-Digital Co., Korea) was used to measure the OD value 600 of the liquid to obtain the desired concentrate for bacterial fermentation.These bacteria-fermented samples were considered as the treatment group, while the vanilla beans that were uncured and allowed to ferment naturally i.e. without the addition of any isolated bacteria; were categorized as the control group.

Bacteria community analysis
The purified bacteria isolates were isolated for DNA extraction using a Qiagen DNeasy Plant Kit (TopGen Biotechnology Co., Ltd, Kaohsiung, Taiwan).The DNA sequencing model was an ABI 3730XL DNA Analyzer using the primers 907R and 16sF-1F with an ABI Big Dye Terminator v3.1 reagent.The sample and 400 µL API buffer were added and heated to 65 ℃ for 15 min.This procedure was repeated twice for 8 min.The tubes were placed on ice for 5 min.The samples were centrifuged at 14,000 rpm, at 25 ℃ for 5 min.The supernatant was discarded and the lower solution was resuspended with AW1 buffer (1.5 times the volume of supernatant).At this step, a min centrifuge at 8000 rpm was done at room temperature.50 µL of 40 ℃ warm water was added and allowed to settle for 10 min.This tube was then subjected to centrifugation at 14,000 rpm for 1 min.

Taxonomic profiling and Venn diagram construction
The microbial communities for each unique sample group were characterized according to their taxonomic profile using the relative microbial abundance.Only the bacterial OTUs of the phylum and genus levels were used for analysis in this research.Hence, to compare the genus and phylum levels between the bacterial communities for non-cured vanilla beans and bacteria-fermented (cured) samples were presented in a Venn diagram (in OTUs).

Principal component analysis (PCA)
Statistical analysis or chemo-metrics can be employed after the curing process to observe the vanilla microbiome.A multivariate statistical approach; principal component analysis (PCA) can be used to minimize the dimensionality of a set of observations made up of a lot of interconnected variables 46,47 .The PCA was used in this study as a technique for analyzing the bacterial microbial communities.This technique is an ordination of reduced variables that are projected at the initial set of variables into new axes (PCs).Hence, the first-hand axis (PC1) is calculated on another axis (PC2) as an eigenvalue of the correlation (or covariance) matrix which is attained from the multivariate matrix of the data set, so a restricted number of axes extracts the highest possible variance, and eigenvectors linked with eigenvalues that decide each PC direction 48,49 .

Taxonomic abundance profiling of the bacterial communities
According to the abundance of information on the bacterial communities at the taxonomic level, a histogram at the phylum level and a heat map analysis at the genus level based on the Bray-Curtis distance were constructed to calculate the frequency of related communities of non-treated vanilla samples and bacteria fermented group.The grouping of the related bacterial communities that may influence the curing process during the fermentation of V. planifolia was exhibited in a bubble plot and a circos presentation at phylum and genus levels, respectively.

Ethical statement
The authors would like to declare that vanilla bean pods (Vanilla planifollia) were harvested, green and matured with permission from Pingtung Lin Tsai Bian Association Farm, Taiwan (22°25′41.8″N,120°32′29.0″E).The website link of the farm is as follows https:// www.agrih arvest.tw/ archi ves/ 73970 confirming that the harvest of the vanilla was grown under the appropriate shade and cultivation media before it was harvested and used in this study.The farm also confirmed that the vanilla was following the national regulations of the Ministry of Agriculture in Taiwan (https:// eng.moa.gov.tw/); hence vanilla bean pods were grown following international guidelines.Authors do not have any conflict of interest.

Strain identification
The bacteria used in this study was isolated from vanilla beans (Vanilla planifolia).The isolation was done from previously naturally fermented and cured batches of vanilla bean pods in the laboratory.A 16S rRNA gene sequencing was conducted based on Johnson et al. 50methods owing to high accuracy compared to other molecular identification markers.For instance, at the species level of Acinetobacter identification 98.2% accuracy was reported with 16S rRNA gene sequencing, rpoB gene sequencing had 93.4%, while 77.2% was with gyrB multiplex PCR, and the VITEK 2 technique had 35.9% accuracy 51 .Mohkam et al. 52 noted that discerning Bacillus isolates with rpoB gene sequencing was sometimes unsuccessful.In this study, the isolated bacteria strain was Bacillus velezensis ZN-S10 acquiring a complete genome.The Bacillus genus is classified as a gram-positive bacterium, from the Bacillaceae family.It is rod-shaped and aerobic with the formation of endospores; hence, it showed whitish, round, and medium-sized colonies in this study, as reported in other research [53][54][55] .

Characterization of isolated pure colonies and hemolysis test findings
This study found that Bacillus velezensis ZN-S10 colonies were effectively isolated from vanilla bean pods (V.planifolia) when grown with an overnight culture in MRS agar plates and incubated at 37 ℃ for 16 h.As observed in Fig. 1, the colonies were fairly spread with off-white colonies on the surface of MRS agar Petri dishes after three consecutive plates (MRS subculture), hence the single pure colonies (Fig. 1A) were found.Researchers in other studies have reported that the MRS medium can be used as a selective medium supporting the growth of Lactobacilli and Bacillus bacteria strains, due to the vitamins, carbon, and nitrogen found in sodium acetate and ammonium citrate of the MRS composition 56,57 .Hereafter, miscellaneous microorganisms, especially bacteria in the environment are eliminated while aiming for specific edible bacteria for vanilla food processing.
The hemolysis test characterized by Savardi et al. 58 as alpha, beta, and hemolysis was conducted with blood agar.The isolated bacteria (B.velezensis ZN-S10) exhibited gamma hemolysis (Fig. 1B) exhibiting no hemolysis since the bacteria did not absorb blood agar.The absence of hemolysis showed that the pure isolate was safe for vanilla fermentation use in both handling, and application, hence ZN-S10 can also be edible for human consumption.
According to this study, 11 strains were isolated from traditionally cured vanilla bean pods.A 16S rRNA gene sequencing was done and identified more than 98.5% partial genomes of 10 Bacillus species.Bacillus velezensis ZN-S10 was the only strain with a complete genome sequence, at 100% identity, thus it was selected as the appropriate strain to investigate the capability of bacterial strains in the effective colonization and fermentation of vanilla bean pods.The accession numbers of the isolated bacteria strain from V. planifolia used in this study, can be obtained from the NCBI website (https:// www.ncbi.nlm.nih.gov/) for the nucleotide sequences.The accession number and sequence length (in bp) of the ZN-S10 strain have been presented in the Supplementary Studies have shown that the vanillin content of high-quality vanilla ranges between 1.7 and 3.6% for V. planifolia and 1.0-2.0%for V. tahitensis 30,60,61 owing to the indirect influence of microorganisms.Hence, the bacterial genome of Bacillus velezensis ZN-S10 with the potential of increasing vanillin in V. planifolia was investigated in this study.

The phylogenetic tree of the evolutionary relationship of taxa
In this study, the phylogenetic tree was constructed based on the 16S rRNA gene sequences, where the accession numbers of the isolated strain and related sequences were obtained from the NCBI web database (https:// www.ncbi.nlm.nih.gov/).Bacterial phylogenetic analyses are frequently carried out to investigate the evolutionary relationships among distinct bacterial species and genera 62 .Hence, we performed the comparison of the Bacillus velezensis strains acquired from non-cured vanilla beans employing the Neighbor-Joining method to show the phylogenetic relationships of the strains (Fig. 2).Yeh et al. 6 reported that microbial coating with Bacillus subtilis subsp.subtilis on Vanilla planifolia yielded fermentation products and volatile components that contributed better aromatic profile of vanilla.Hence, with future studies, the B. velezensis strain ZN-S10 used for fermentation was found distinct from other members of the Bacillus group which might alter the aromatic characteristics of vanilla.Hence, based on the constructed phylogenetic tree (Fig. 2), B. velezensis ZN-S10 were the major bacteria that colonized the vanilla bean pods, and the strain was used for fermentation at the curing process for a new batch of non-cured vanilla bean pods.The strain ZN-S10 accounted for the bootstrap value of 79%, with lesser differences of similar isolate sequences as neighbor strains of B. velezensis EM-1, PMC206-1, and C5 strains with bootstrap values of 73%.The Bacillus genus exhibited that the colonizing microbes were involved the coating of the vanilla pods with ZN-S10 strain, based on the 16S rRNA as illustrated by the phylogenetic tree.Other studies have shown that during the curing of Vanilla planifolia Andrews, Bacillus isolates (such as Bacillus vanillea XY18, B. subtilis XY11, and B. subtilis XY20), can colonize the vanilla beans 19,21 .

Comparison and variation of bacterial communities in naturally fermented and treated vanilla
A Venn diagram (Fig. 3) was constructed for the number of common (uncured vanilla beans) and unique species (bacteria-fermented cured vanilla) in each group.All the sequences of each sample were compared with the NCBI RefSeq 16S rRNA database using EPI2ME v3.0 (Oxford Nanopore Technologies, UK).Hence, the number of variants was very similar to the mean expression of species in the mean groups; showing 230 OTUs amongst the naturally fermented vanilla beans and a very similar population of 233 OTUs in the treatment group.Between the two groups (control and treated vanilla), 230 OTUs were shared by both, while the non-treated pods had the majority of 253 OTUs in the bacterial communities.The shared OTUs between the naturally fermented vanilla (control) and B. velezensis ZN-S10 fermented samples suggested the successful colonization with microorganisms; hence effective fermented and cured vanilla beans.Similarly, Chen et al. 21reported that Bacillus sp.colonizes vanilla beans and is classified as β-d-glucosidase-producing bacteria.The bacterial community population was analyzed with 16 s rRNA gene sequencing presented in operational taxonomic units (OTUs) as shown in Fig. 3 (right).The results reported that the bacteria-fermented vanilla beans had a higher population (~ 390 OTUs) than the control as expected; hence showing the success of curing the beans with B. velezensis ZN-S10 used as a fermentation tool.Studies have also proved that bacteria communities are involved during the curing process of vanilla beans 18 .Hence, the occurrence of the ZN-S10 showed their role in the effective fermentation of V. planifolia during the curing process.

Principal component analysis of the bacterial communities
The microbial communities were profiled with principal component analysis (PCA) to visualize the differences between microorganisms obtained from naturally fermented vanilla and the beans that were treated with 1 mL of 1 × 10 7 CFU mL −1 of B. velezensis ZN-S10.As presented in Fig. 4, the microbial population accounted for 95.83% of PC 1 of the total variance in treated samples, while control samples showed that PC 2 accounted for 4.17% in the PCA score plot.The presence of endophytic bacteria has been shown on non-bacteria cured vanilla pods (control) and B. velezensis ZN-S10 treated samples, as also found in other studies.For instance, a rootcolonizing Bacillus amyloliquefaciens isolated by White et al. 63 , was also isolated in leaves, flowers, and pods of Vanilla planifolia.Likewise, researchers have reported that Colombian vanilla had a couple of endophytes on the plant even on harvested pods 64 .

Findings on the microbiome of the vanilla before and after fermentation with B. velezensis
The vanilla microbiome after fermentation with B. velezensis ZN-S10 was evaluated with high-throughput amplicon sequencing of the bacterial 16S rRNA gene showed a predominance of Firmicutes, Proteobacteria, Cyanobacteria, Planctomycetes, and Bacteroidetes, respectively (Fig. 5).These genera made up ∼95% of the total data set wherein at phylum level, a high dominance (≈ 80%) of Firmicutes with the subdominant of Proteobacteria, and a small fold of Cyanobacteria was observed (Fig. 5A) with the vanilla beans that were blanched and allowed to ferment naturally.Nevertheless, the vanilla samples fermented with isolated bacteria from previously naturally fermented vanilla batches also showed a high dominance of Firmicutes, (≈ 60%) which was slightly lower in abundance with control samples.The microbiome of treated samples was followed by Proteobacteria and Cyanobacteria which were a two-fold increase as compared to control samples (Fig. 5A).Researchers have also reported that Proteobacteria and Cyanobacteria were the common families found in the core of the vanilla microbiome 27 .Also, according to Xiong et al. 65 , the monoculture practice of growing vanilla beans contributes to the high dominance of Firmicutes due to low bacterial and fungal communities that occur during the continuous cropping of the vanilla beans.Furthermore, the relative abundances of Firmicutes, Actinobacteria, Bacteroidetes, and Basidiomycota are common phyla in vanilla plants especially when grown in a monoculture system.Hence, in this study, the results indicated that at a phylum level Firmicutes, Proteobacteria, and Cyanobacteria were the bacterial groups (Fig. 5A and B) with key influence in the curing of vanilla beans.The treated vanilla bean pods showed a higher relative abundance of Cyanobacteria and Proteobacteria as compared to the non-treated samples (Fig. 5B).Moreover, it should be noted Planctomycetes, and Actinobacteria were presented in the bacteriafermented vanilla although the communities in this group were in small abundance, rather than in the control samples (Fig. 5B).Notably, the discussed findings of the bacteria communities were found on the cured V. planifolia beans for both treated samples (ZN-S10 fermented) and non-treated pods.The availability of the Bacillus sp.might contribute to the aroma and flavor properties of the vanilla pods.Gu et al. 19 found that cured pods with B. vanillea and B. subtilis coating had elevated vanillin yielding better aromatic qualities of vanilla.However, it should be noted that different geographic locations of vanilla species vary in flavor and aroma attributes 26,66 .
According to the heat map analysis (Fig. 6), the vanilla microbiome at the genus level presented significant differences in the relative abundance across the bacteria-fermented samples (C3-1-9) and the control groups.Hence, an increase of Firmicutes, Proteobacteria, and Cyanobacteria was observed while Bacteroidetes, Verrucomicrobia, and Planctomycetes were reduced with a significant decrease of Actinobacteria on the microbiome of bacteria-treated vanilla samples.Similarly, Firmicutes were remarkably increased in the control group, and Proteobacteria and Cyanobacteria were reduced while significant decreases were observed in Bacteroidetes and Verrucomicrobia, with a slight increase in Planctomycetes and Actinobacteria in contrast with the microbiome of the treated vanilla.Carbajal-Valenzuela et al. 27 stated that vanilla plants' root and stem system can be accumulated with Actinobacteria, which can later be inhabited in post-harvested pods.Similarly, Ling et al. 67 studied root-grafted watermelon showed a high dominance of Firmicutes, Actinobacteria, and Cyanobacteria, Proteobacteria rather than noon-grafted watermelon.This suggests that the soil's vanilla microbial community might accumulate more in roots, but mostly important to the abundance of bacteria in the pods that survive during the blanching process.However, these findings require further research.
Figure 7 shows the bubble plot had four bacterial classes that include; Firmicutes, Cyanobacteria, Planctomycetes, and Proteobacteria among others were the most abundant in the microbiome of naturally fermented vanilla and samples fermented with B. velezensis.Hence, at the genus level, Priesta had a higher relative abundance of 0.5 for Firmicutes phylum on both control and treated vanilla bean samples.Studies have shown that the more abundant of Firmicutes with Actinobacteria in the bacterial phyla is commonly due to the production of secondary metabolites such as volatile compounds 68,69 .Therefore, the fermentation of vanilla with the most abundance of the bacteria phyla exhibited that volatile compounds were produced after the B. velezensis ZN-S10 inoculation on the pods.The results, owing to the bacteria phyla reported suggest that bacteria play a role in altering the aroma and flavor-volatile compounds in V. planifolia.
The distinct bacterial communities were presented in a circos plot at genus and phylum levels (Fig. 8) that was detected in naturally fermented vanilla (control) and the vanilla treated with B. velezensis for fermentation purposes during the curing process.We identified 5 different genera in both the control and treated vanilla samples.Hence, at the genus level, the treated vanilla samples had dominant Priesta (45%), while the naturally fermented vanilla had 55% dominance.Interestingly, at this level larger fractions of Pseudoxanthomonas (96%), Microcoleus (90%), Loriellopsis (95%), and Paenibacillus (97%), as compared to the control group (Fig. 8A).Nevertheless, the control vanilla microbiota was slightly supplemented with Microcoleus (10%), and Loriellopsis (5%); also showing a lower abundance of Paenibacillus (3%), Psedoxanthomans (4%), which was contrary to the microbiota of the treated vanilla.Therefore, the Priesta or Bacillus were presented as the predominant bacterial community at the genus level.It should also be noted that the Firmicutes were the most dominant phylum (Fig. 8B) as observed with higher relative abundance at the phylum level.Li et al. 70 found that pepper plants in allyl isothiocyanate fumigated soil had an increased relative abundance of Pseudoxanthomonas with a reduced abundance of Planctomycetes and Acinetobacter at genus level.The changes in the bacteria communities were similar to the significant differences between B. velezensis-treated vanilla and the control.The reports of these studies also exhibited that soil treatment should be monitored for soil bacteria and other microorganism communities.Thus, the variation of the bacteria genera in the vanilla beans might be affected by soil conditions.Moreover, studies with disease-resistant tobacco cultivar HB202 compared with a susceptible XY3 cultivar showed that the HB202 variety had higher bacteria communities in contrast to the XY3 variety 70 .Differences in vanilla bean varieties might support different microbial-network communities, especially relative abundances of bacteria genera after the curing process of vanilla.

Conclusions
In this research, Bacillus velezensis ZN-S10 was successfully isolated from Vanilla planifolia and was used for bacterial fermentation in green-matured non-cured vanilla beans.The bacteria communities were also analyzed with the NGS technique, where 16S rRNA gene sequencing to explore the genus and phylum level of the bacteria population involved.The annotations provided by the sequence data, multiple aspects of the development and biochemistry of vanilla beans could be understood to be involved in the colonization of non-cured vanilla beans, while microorganisms are important in fermentation for the curing process in V. planifolia.The 16S rRNA gene sequencing in this study was shown to be very crucial for the understanding of the bacterial ecology, abundance, and relationship taxonomy to improve the fermentation procedures done for the curing process to achieve highquality cured vanilla beans.The findings of the bacterial communities suggested that at the genus level Priesta were dominant; hence were involved in the colonization or fermentation of V. planifolia.Moreover, at the phylum level, the Bray-Curtis distance revealed a higher relative abundance of Firmicutes, hence these bacteria had an important role in the curing process of the vanilla beans.Further research on the cured vanilla beans cannot only be used to extract the bacteria responsible for fermentation but for natural vanilla extract as well as for the preparation of other vanilla products.This article proved that the isolated bacteria from the vanilla, B. velezensis ZN-S10 could control the bacterial communities during the fermentation process for the manufacturing of safe vanilla food products (Supplementary Information).

Figure 1 .Figure 2 .
Figure 1.Bacillus velezensis ZN-S10 colonies.The ZN-S10 pure isolate appearance showing the single colonies on the fourth streak region (circled) grown on MRS agar Petri-plate, incubated at 37 ℃ for 16 h (A); and the gamma hemolysis of the ZN-S10 strain (B) grown on blood agar.

Figure 3 .
Figure 3. Venn diagram (left) at shared and the unique OTUS of the vanilla microbiome and schematic diagram (right) of dilution curve results of naturally fermented (control) vanilla beans and Bacillus velezensis ZN-S10-treated samples.representing the control vanilla beans and Bacillus velezensis ZN-S10-treated samples.

Figure 4 .
Figure 4.The principal component analysis (PCA) of the bacterial communities obtained from naturally fermented vanilla (control) and B. velezensis ZN-S10-treated vanilla beans.

Figure 5 .
Figure 5.The bacterial abundance of the vanilla microbiome.(A) Histogram of the bacterial composition at its relative abundance (%) presented at the phylum level of the control group (naturally fermented vanilla pods) and Bacillus velezensis ZN-S10 fermented vanilla (C3-1-9).(B) The relative sample abundance at the phylum level was analyzed with the Bray-Curtis distance of non-treated vanilla samples (control) and B. velezensis ZN-S10 fermented (C3-1-9).

Figure 6 .
Figure 6.The heat map analysis of the vanilla microbiota composition at the genus level for Bacillus velezensis ZN-S10 fermented samples (C3-1-9) as compared with the control group.

Figure 7 .Figure 8 .
Figure 7.The taxonomic bubble plot of the bacteria in vanilla at the relative abundance at genus and phylum levels for the control and treatment groups.Circle sizes represent the relative abundance, and colors indicate the bacterial phylum groups.