Development of Chinese chestnut whiskey: yeast strains isolation, fermentation system optimization, and scale-up fermentation

In this study, we used Chinese chestnut as the main raw material to develop a novel type of whiskey. First, 16 yeasts were isolated and identified for producing aroma using olfactory plate assay. Of these, we screened nine yeast strains based on their fermentation capacity, aroma profile, and sensory evaluation. The results demonstrated the combination of strains HN006 (Saccharomyces cerevisiae) and HN010 (Wickerhamomyces anomalus) provided satisfactory wine fermentation with an interesting flavor profile, as strain HN010 was highly aromatic and had elevated sensory scores with comparatively low ethanol yield, while strain HN006 had a poor flavor profile but produced the largest amount of ethanol. Subsequently, we co-cultured strains HN006 and HN010 to optimize the fermentation system. The results revealed the following optimum parameters: a mixed inoculum of 6% (v/v) at an HN006/HN010 ratio of 1:2 (v/v), a raw material ratio of 5:3:2 (chestnut: malt: glutinous rice), and yeast extract concentration of 6 g/L. Additionally, this fermentation system was successfully scaled-up to a 1000 L pilot-scale system. The results of this study showed that strains HN006 and HN010 could be used as alternatives for whiskey fermentation, as well as provided a generalized experimental scheme to assess other microorganisms.


Introduction
Chinese chestnut, one of the major nuts in China, belongs to the genus Castanea Miller and the family Fagaceae and is also known as "the king of dry fruit" (Zhang et al. 2018a, b). Its edible kernels are popular in China and are rich in both macronutrients (starch, protein, and fat) and micronutrients (minerals, such as iron, calcium, phosphorus, and several types of vitamins, as well as trace minerals). Modern pharmacological studies have shown that the Chinese chestnut nourishes the stomach and the spleen, improves the functioning of kidneys, strengthens the tendons, activates blood circulation, and is used for the prevention and treatment of hypertension and coronary heart disease (Zhang et al. 2014).
Chinese chestnut is a cultivated plant with a history dating back approximately 3000 years in China, and ever since, there has been an annual increase in the production of chestnuts (Zhang et al. 2018a, b). However, the phenomenon of "difficult to sell chestnut" still exists in many planting areas in China, which has reduced the motivation of the chestnut farmers. Additionally, the seasonal availability of Chinese chestnut fruits along with their high water content (approximately 50% on a wet basis) limits their storage due to germination, mildew, and insect infestation (Sheng and Yu 2015). Currently, the main processing methods of Chinese chestnuts include baking, cooking, or frying with sugar (tang chao li zi), which is not beneficial economically. Thus, there is an urgent need to identify and develop novel processing methods for using Chinese chestnuts to produce highvalue products, such as Chinese chestnut wine.
Whiskey is one of the four famous distilled wines in the world and is popular internationally. The most famous whiskey producing countries are Scotland, Ireland, the USA, Japan, and Canada. However, there are few reports on whiskey production in China. The high starch content of the Chinese chestnut makes it suitable for whiskey brewing. However, Chinese chestnut whiskey (wine) is still scarce in the domestic and foreign markets.
The sensory attributes of whiskey result from compounds that originate from the raw materials, the metabolites of yeast, and maturation process. Among them, the yeast make a special contribution to the flavour quality of whiskey, and the selection of microorganisms is critical for the formation of aroma, which is one of the key attributes affecting consumer acceptance (Du et al. 2019).
Therefore, this study focused on developing a novel whiskey using the Chinese chestnut as the main raw material. We isolated and investigated the potential of specific yeast strains for effective Chinese chestnut wine fermentation. Additionally, we used the screened strains to optimize the fermentation process based on high ethanol yield and sensory scores. Furthermore, we used a 1000 L pilot-scale system to scale-up the fermentation to industrialize the process of Chinese chestnut wine production. The results of this study would provide supportive evidence to develop a Chinese chestnut-based novel whiskey.

Materials
Chinese chestnuts and glutinous rice were purchased from a retail market in the Huoshan county, Anhui province of China, in October 2019. The malt was obtained from Zhengzhou Longhai beer Materials Co., Ltd (Zhengzhou, Henan province, China). The thermostable α-amylase and glucoamylase for saccharification were acquired from the Sunson Industry Group Co., Ltd. (China).

Isolation of superior yeast strain Initial screening for yeast strains
In a 250 mL flask, we mixed 5 g of Chinese strong flavour Daqu, which is a saccharifying and fermenting agent utilized in the brewing of Chinese strong flavour liquor and contains a variety of yeast species, into 100 mL of yeast extract peptone dextrose medium (YEPD; 5 g/L yeast extract + 50 g/L glucose + 10 g/L peptone). The mixture was stirred at 200 rpm at 30 °C for 24 h. The culture was serially diluted with sterile water and coated on the YEPD plates supplemented with 0.01% chloromycetin in duplicates. Next, 48-72 h later, we isolated colonies with varying morphologies and purified them, followed by inoculation on a YEPD agar plate to conduct a smelling or "sniffing" test for initial aroma screening following the method of Alicia et al. (2018).

Repetitive element sequence-based PCR (Rep-PCR) analysis and yeast identification
A DNA extraction kit was used to extract the genomic DNA of the isolated yeast strains (Sangon Biotech, Shanghai, China). The rep-PCR analysis was done using the primer (GTG) 5 (5′-GTG GTG GTG GTG GTG-3′) using a previously described method (Simonetta et al. 2016). A 1.2% agarose gel electrophoresis (AGE) was used to separate the amplicons in a 1 × TBE buffer (150 min, 120 V). The clusters were identified by analyzing the resultant dendrograms.
For identification of the yeast, the universal primer pairs NL1 and NL4 were used to sequence the strain of each rep-PCR fingerprint group (Yan et al. 2019a, b), followed by alignment to the 26S rRNA gene sequences in the GenBank database using the BLAST algorithm tool. The nucleotide sequences obtained in this study have been assigned GenBank Accession Nos. MW076944-MW076959, as indicated in Fig. 1.

Testing for the formation of volatile flavor compounds
Chinese chestnuts were peeled, cleaned, and steamcooked for 30 min. After naturally cooling to room temperature, the cooked chestnut kernels were mixed with water (1: 3 (m/v)) and pulped using a fruit mixer. Next, the mixtures were transferred into a bioreactor and saccharified using α-amylase (10 U/g chestnut kernel; pH 5.5, 90 °C, 30 min) and subsequently, glucoamylase (200 U/g chestnut kernel; pH 4.5, 60 °C, 6 h). Then, a double gauze was used to separate the liquid phase of the saccharified mixture and placed into a 2.5 L glass bottle, each with 2 L volume. The liquid phase of the saccharified mixture was sterilized at 113 °C for 10 min and subjected to ethanol fermentation. The conditions of the fermentation and distillation of Chinese chestnut wine have been listed in Sect. 2.3. Finally, the volatile compounds in each sample of the obtained raw wine were detected and quantified by solid-phase microextraction-gas chromatography/mass spectrometry (SPME-GC-MS).

Testing for fermentation ability of isolated strains
The seed colonies of each isolated yeast strain were cultured aerobically in liquid YEPD medium for 24 h at 30 °C with constant stirring at 200 rpm. Next, 5% (v/v) of these suspension cultures (approximately 1 × 10 7 CFU/ mL) were added to 250 mL Erlenmeyer flasks containing YEPD medium (200 mL) with 190 g/L glucose. The flasks had perforated silicon stoppers equipped with 0.45 mm filters (Merck Millipore, Italy), to release carbon dioxide and to prevent contamination. All the tests were carried out in triplicates and performed under static conditions at 30 °C. Post-fermentation, each fermentation broth was sampled to detect the remaining reducing sugar content, as well as the ethanol content.

Sensory evaluation of the flavor of yeast fermentation products
Section 2.6 describes the methods used for the evaluation of the sensory features, including smell, taste, flavor, acidity, and overall acceptability of each obtained distillate. Further testing was done using yeast strains with high total scores.

Laboratory fermentation of Chinese chestnut wine
Additional file 1: Figure S1 shows the manufacturing process of Chinese chestnut whiskey raw wine. The Chinese chestnut kernel, malt, and glutinous rice were cleaned and soaked (glutinous rice was soaked for 2 h in 1.5 times water). Post-soaking, glutinous rice and Chinese chestnut kernels were steam-cooked for 30 min and cooled to room temperature. Then, water was added at a ratio of 1: 3 (m/v) and the mixture was saccharified by α-amylase (10 U/g raw material) at pH 5.5, 90 °C for 30 min; followed by saccharification with glucoamylase at pH 4.5, 60 °C for 6 h with the addition of the pre-crushed malt. Subsequently, a double gauze was used to separate the liquid phase of the saccharified mixture, followed by adjusting the pH to 5.0 and transferring the contents into a 5 L fermenter with 4 L of working volume. After sterilization at 113 °C for 10 min, the prepared sugar solution was subjected to wine fermentation by inoculation with pre-cultured yeast seeds. The temperature of the fermentation system was maintained at 30 °C.
Post-alcoholic fermentation, we immediately doubledistilled the fermented liquids in a 5 L Charentais potstill. When the alcohol concentration was less than 0.5% (w/v), we stopped the first distillation, resulting in an approximately 20% (v/v) distillate. During the second distillation, approximately one-third of the distillate was collected, resulting in a final alcohol concentration of 60% (v/v) in the final raw Chinese chestnut wine.

Optimization of Chinese chestnuts wine fermentation
We performed anaerobic Chinese chestnut wine fermentation as described previously for 6 days to optimize the fermentation process. At the end of the process, we collected samples of fermentation broth to determine the ethanol and residual sugar content, as well as to evaluate the overall sensory features of the obtained wine. We also evaluated the impact of different inoculum sizes on the production of Chinese chestnuts wine. Table 2 shows the settings used for this analysis. Additionally, we examined the impact of different raw materials on the brewing of chestnut whiskey. The Chinese chestnut kernels, malt, and glutinous rice were prepared in various ratios (Table 3) and were subjected to the wine fermentation process following the previously described procedure. Furthermore, we also tested different supplementation amounts of nitrogen nutrient (yeast extract). All experiments were conducted in triplicates, and the results were expressed as the average of the three repetitions.

The inoculum preparation for 1000 L pilot-scale fermentation
For inoculum preparation, the suspended cells of the obtained yeast from single colonies on YEPD agar plates were pre-cultured in 250 mL flasks containing 100 mL YEPD medium at 30 °C for 24 h. Then each of 20 mL cultures were transferred to five 500 mL flasks each containing 180 mL YEPD medium, and cultivated under the same conditions mentioned above, after which the entire volume of the five 500 mL-flasks was used to inoculate a 10 L fermentor (v/v, 10%). The culture broth from the 10 L fermentor was transferred to a 150L fermentor containing 100L of growth medium. Both the fermentation scales of 10 and 150 L were conducted at 30 °C, 300 rpm, and VVM of 4.0 for 24 h. Finally the culture broth from 150 L fermentor was used for the pilot-scale fermentation. The volumes to be inoculated were determined on the basis of cell counting under a light microscope with a Thoma Zeiss chamber.

L pilot-scale fermentation
We performed the pilot scaling-up experiments based on the optimization conditions derived from laboratory experiments in a pilot-scale plant built at Anhui Yingjia Distillery Co., Ltd., located in Huoshan county, Luan City, China. The scale-up fermentation was performed in a 1000 L fermenter with a liquid loading capacity of 900 L. Additional file 1: Figure S2 shows a schematic diagram of the experimental apparatus used in this study.

Analytical methods
During the fermentation process, we collected broth samples at specific time intervals, followed by 20-min centrifugation at 5000 g to remove the cellular debris. The 3, 5-dinitro-salicylic acid colorimetric method was used to determine the reducing sugar content (Miller 1959). The measurement of yeast cells and the ethanol content was performed following the method of Yan et al. (2013).

Analyses of the volatile compounds
The HS-SPME-GC-MS was done to analyze the aroma profiles of the obtained wine samples, following a previously described method (Yan et al. 2019a, b). The volatile flavor compounds in the samples were identified by comparing the mass spectrum data with the NIST 05a MS database (Agilent Technologies, Inc., USA), and quantification was done based on the comparison of their peak intensity to that of the 2-octanol (internal reference).
Additionally, heatmap clustering was used to analyze the HS-SPME-GC-MS dataset to identify the differences in the fermentation profiles of different yeast strains. The analysis was done using R (programming language) Software (Annalisa et al. 2018).

Sensory evaluation
We evaluated the sensory characteristics of smell, taste, acidity, after-taste, and overall acceptability of the final Chinese chestnuts wine following a previously reported method (Yan et al. 2019a, b). Wine samples (20 mL) were labeled with a randomly assigned 3-digit code and simultaneously served in ISO tasting glasses covered with plastic Petri dishes at room temperature.

Preliminary screening of yeast strains
Post-morphological examination (both micro and macro), we identified 16 isolates that were growing on YEPD agar as yeast strains. The cluster analysis of the rep-PCR of these 16, based on a coefficient of similarity of 76%, resulted in eight clusters ( Fig. 1). We identified the isolates in each cluster by sequencing the D1/ D2 region of the 26S rRNA gene, followed by a BLAST search at GenBank (Fig. 1). Approximately 31.25% of the yeast isolates were found in cluster VII, which were identified as Wickerhamomyces anomalus (HN007, HN010), Wickerhamomyces mori (HN012, HN015), and Dabaryomyces hansenii (HN014). Approximately 18.75% of the yeast isolates were found in cluster V and shared 100% homology with GenBank sequences of Saccharomyces cerevisiae (HN006, HN009) and Saccharomyces kudriavzevii (HN016). The cluster I and VIII isolates were unambiguously identified as Hanseniaspora vineae (HN001, HN005) and Pichia kluyveri (HN002, HN004), respectively. The remaining four clusters (II, III, IV, and VI) were identified as Trichosporon asahii (HN003), Kluyveromyces lactis (HN008), Galactomyces geotrichum (HN013), and Yarrowia lipolytica (HN011), respectively, each showing 100% homology to GenBank sequences.
We performed the olfactory "sniff " test on the YEPD medium plates to preliminarily determine the aroma producing ability of the obtained 16 yeast strains. Table 1 presents the results of this analysis. We found that different genera of the isolates presented various aromas, differentiated by different intensities of sourness, fruity, floral, cheese-like, and cream-like smells. Saccharomyces cerevisiae (HN006, HN009) and Saccharomyces kudriavzevii (HN016) produced different intensities of alcoholic aroma in the YEPD agar plates. Both Dabaryomyces hansenii (HN014), and Yarrowia lipolytica (HN011) produced characteristic cheese-like smells with intermediate strength. Each of the following strains: Trichosporon asahii (HN003), Wickerhamomyces anomalus (HN007, HN010), Wickerhamomyces mori (HN012, HN015), Pichia kluyveri (HN002, HN004), and Kluyveromyces lactis (HN008) produced pleasant fragrance of pineapple fruit, apple and rose fruit, apple fruit, banana fruit, and strawberry fruit, respectively, and among which strain Wickerhamomyces anomalus (HN010) possessed the highest intensity. While the strain Hanseniaspora vineae (HN001, HN005) generated an unpleasant sour flavor, the strain Galactomyces geotrichum (HN013) produced a cream-like aroma, which was reported to have wide application in the fermentation of dairy products (Jacques et al. 2017).
Apart from presenting the overall aromatic profile of each tested strain, Table 2 also highlighted the differential behavior exhibited by strains belonging to the same species. We observed a variation in the total concentration of each type of volatile compounds based on different strains. The samples produced by non-Saccharomyces yeasts (HN002, HN003, HN008, HN010, HN011, HN013, and HN014) possessed higher amounts of volatile acids compared with those of Saccharomyces yeasts (HN006, HN009), especially the wine fermentation using the strain HN010 showed the highest concentration (15.721 μg/mL) of various acids, with the exception of butanoic acid, hexanoic acid, 2-methyl-butanoic acid, and nonanoic acid. Butanoic acid and hexanoic acid are generally present in the Chinese strong flavor liquor, and their high content can impart an unpleasant flavor to the wine (Aslankoohi et al. 2016). The presence of 2-methylbutanoic acid in the strains HN011 and HN014 imparts cheese-like and sweet characteristics (Marycarmen et al. 2020), which were consistent with the results of our preliminary olfactory "sniff " test. Nonanoic acid, detected only in strain HN009, had green and fat characteristics (Aslankoohi et al. 2016). No significant difference was observed in the type and concentration of volatile acids between HN006 and HN009.
We observed a significant difference in the concentrations of esters in the wines produced using different cultures (Table 2). We found a significantly higher concentration of esters in the non-Saccharomyces culture wines compared with the Saccharomyces culture wines, which was probably attributed to its low lipase activity ( Table 2). The highest concentration of esters (16.456 μg/mL) was found in the wine prepared using HN010, which was 3.22 and 3.70 times higher than that in the Saccharomyces yeast strain of HN006, HN009,  respectively. Amongst all types of esters, the concentration of ethyl acetate was found to be the highest for each strain, ranging from 2.187 μg/mL (HN006) to 16.456 μg/mL (HN010). Ethyl esters, typically described to possess a "fruity and flower" aroma (Laura et al. 2019), were the largest ester-based compounds in the Chinese chestnut wine. We found extremely high amounts of ethyl acetate (sweet, fruity), phenylethyl acetate (rosy), and ethyl caproate (apple peel, fruity) in the Chinese chestnut wine produced by HN010 (Marycarmen et al. 2020), along with high-molecular-weight ethyl esters, such as ethyl laurate, ethyl palmitate, ethyl oleate, and ethyl linoleate, which were reported to promote the accumulation of the after-taste in the Chinese strong flavor liquor (Yan et al. 2019a, b). Higher alcohols, produced by yeast, originate either from the degradation of branched-chain amino acids or directly from sugar fermentation. We found that the higher alcohols constituted the major quantitative component in Saccharomyces (HN006, HN009)-fermented chestnut wine samples, and were significantly higher than those obtained by non-Saccharomyces-fermented wines ( Table 2). The wine samples brewed using the strain HN006 exhibited the highest concentration of total higher alcohols, which included isoamyl alcohol (3.987 μg/mL), 1-nonanol (0.254 μg/mL), 1-hexanol (0.323 μg/mL), 2,3-butanediol (0.914 μg/mL), isooctanol (0.165 μg/mL), 2-methyl-1-propanol (1.851 μg/ mL), enanthol (0.294 μg/mL), 3-ethoxy-1-propanol (0.754 μg/mL), 1-octen-3-ol (0.653 μg/mL), octanol (0.572 μg/mL), and phenylethyl alcohol (0.354 μg/mL). Of these, isoamyl alcohol, 2,3-butanediol, and phenylethyl alcohol have previously been reported to be present in fermented alcoholic beverages as characteristic flavor compounds, such as Chinese strong flavor liquor (Yan et al. 2019a, b).
Aldehydes are produced from amino acids either by Strecker degradation or by transamination, followed by decarboxylation, and they can be easily reduced to alcohols. We observed that the Saccharomyces-fermented wines possessed a significantly higher concentration of aldehydes compared with the non-Saccharomyces-fermented wines. This trend was consistent with the trends of higher alcohols (Table 2).
Additionally, we detected three aldehydes in the Chinese chestnut wine samples, and they were found to be most abundant in the strain HN006. Among these, 2-octanone, which imparts a cream-like flavor to the wine body (Yan and Dong 2019), was found in the Chinese chestnut wine samples. Also, 2-nonanone, which has a fruity and flowery note (Yan and Dong 2019), was found in each chestnut wine sample fermented using strains HN005, HN006, HN007, and HN008.
We created a heatmap representation to allow a rapid visual assessment of the similarities and differences in the volatile profile of the strains between different samples. Additionally, the acids, alcohols, and esters represented the major functional groups for all the yeast strains (Fig. 2). Figure 2 shows the differences in the relative abundance of compounds among different strains, revealing the presence of different flavor profiles in different strains. For instance, the Saccharomyces strains (HN006, HN009), produced high levels of ethyl esters, isoamyl alcohol, and 2-methyl-1-propanol, along with the non-Saccharomyces strains (HN004 and HN010). Notably, the strain HN010 showed relatively larger amounts of volatile compounds compared with other strains, which indicated that it might produce wine with a more harmonious wine aroma. These results also suggested that different yeast strains, especially non-Saccharomyces yeasts, imparted a specific influence on wine flavor. However, their fermentation characteristics are not yet fully understood, and further research is required to identify the wine fermentation capacity of each strain.

Testing for fermentation ability
We tested the fermentation performance of the obtained nine yeast strains to identify the optimal yeast strain for preparing the Chinese chestnut wine. Figure 3a, b show the effects of yeast strains on the concentrations of ethanol and reducing sugar during the fermentation of Chinese chestnut wine. We observed a slight increase in the concentration of ethanol for Saccharomyces strains (HN006, HN009) during the initial 24 h of fermentation, which increased significantly and reached the maximum value after 120 h, whereas the ethanol content produced by non-Saccharomyces strains (HN002, HN003, HN008, HN010, HN011, HN013, and HN014) got slowly accumulated throughout the fermentation, and their corresponding concentrations were lower than those produced by the Saccharomyces strains. Table 3 also summarizes the kinetic parameters of the ethanol fermentation with different strains. The strain HN006 generated the maximum concentration of ethanol (90.220 ± 0.874 g/L), had the highest ethanol productivity (0.752 ± 0.026 g/L/h), and sugar utilization (99.5%) after 120 h (Table 3). Also, this strain achieved the highest final ethanol yield of 0.467 ± 0.043 g/g. Thus, we observed that the strain HN006 showed the high fermentation performance of Chinese chestnut wine.

Sensory evaluation
A sensory evaluation was performed to identify the yeast strain that had the potential to produce a pleasant tasting wine. Sensory characteristics are vital markers to evaluate the wine quality, amongst which sensory appraisal is a commonly accepted method. A significant difference was observed in different quality attributes, such as aroma, after taste, wine taste, mouthfeel, and overall quality (Additional file 1: Figure S3). Among these wine samples, the wine prepared using strain HN010 was superior to those with other strains based on its highest scores for taste (7.37), aroma (7.76), mouth feel (7.64), after taste (6.90), and overall acceptability (7.12), which agreed with the result of aromatic profile of the volatile compounds determined by SPME-GC-MS. The Saccharomyces strains (HN006, HN009) received the lowest scores in the sensory analysis. Thus, this test identified and verified that the taste-profile of the Chinese chestnut wine made using the yeast strain HN010 was appreciated by the testers and had an affirmative effect on the overall sensory features.
Thus, the non-Saccharomyces strain, HN010, showed the highest amounts of volatile compounds and best sensory characteristics but with poor ethanol fermentation capacity, and the Saccharomyces strain, HN006, showed the highest efficiency in ethanol production and consumption of reducing sugar. Therefore, both these yeast strains were selected to be used as mixed inoculum for producing an improved version of the Chinese chestnut wine. Both strains were deposited in China Center of Industrial Culture Collection (CICC) and assigned the number of CICC1035 (HN006) and CICC1783 (HN010), respectively.

Optimization inoculation sizes for chestnut whiskey brewing
We tested different inoculation volumes of strains HN006 and the strain HN010 to inoculate the fermentation mash at the beginning of the fermentation process of the Chinese chestnut wine, and their inoculums were pre-cultured separately as specified in Sect. 2.2. Table 4 lists the effects of the mixed inoculation volumes with different ratios of strains HN006 and HN010 on the concentration of ethanol and overall sensory scores. We observed an increase in the concentration of ethanol with an increase in the overall combined inoculation volume and Fig. 2 Heatmap illustrating the differences in the concentrations of volatile compounds in the samples of Chinese chestnut wine prepared with various yeast strains reached a maximum value at 6% (v/v). Further increase in the inoculation volume resulted in a decrease in ethanol concentration. There was no significant difference in ethanol concentration for different inoculation ratios of HN006 and HN010 at an inoculation size of 5%. Further modification in the size and ratio of this inoculation resulted in a decreased ethanol production. As for the overall sensory scores, it showed an increasing trend with an increase in the inoculation ratio of the strains HN010 to strain HN006 at various inoculation volumes below 6%; however, there was no significant difference in the overall sensory scores for various inoculation ratios as the inculcation size reached up to 12% (v/v). The highest overall sensory score was also achieved at the mixed inoculation size of 6% (v/v) and a HN006 to HN010 inoculation ratio of 1: 2 (v/v). Thus, we used this inoculum size in further studies.
Optimization the ratios of different raw material for chestnut whiskey brewing Table 5 illustrates the influence of raw materials with different ratios on the brewing of the Chinese chestnut whiskey. The ratio of chestnut whiskey to malt, and to glutinous rice in the medium had a crucial effect on the sensory score and the production of ethanol.
The data presented in Table 5 revealed an increase in the initial reducing sugar concentration of the saccharified mash with the total amounts of malt and glutinous rice. The ethanol concentration and overall sensory scores increased as the proportion of malt and glutinous rice to Chinese chestnut increased, and both reached a maximum value at the ratio of 5:3:2 (Chinese chestnut: malt: glutinous rice). All multiple-grain fermentations were superior to single-grain fermentation of Chinese chestnut, as shown by a higher overall sensory score and the ethanol content. Table 5 demonstrated that the optimum ratio of chestnut whiskey to malt to glutinous rice, and to corn was 5:2:2:1 (w: w: w: w), which is justified by the highest corresponding value of the sensory score (8.210 ± 0.300) and highest ethanol concentration (86.021 ± 2.470). This value will be used for further testing.   Optimization nitrogen nutrient additions for chestnut whiskey brewing Table 6 demonstrates the effect of the addition of various amounts of nitrogen nutrients on the fermentation of the Chinese chestnut wine. We observed an increase in the production of ethanol with an increase in the supplementation amounts of yeast extract. We obtained the maximum concentration of ethanol (92.587 ± 3.872 g/L) was obtained at the yeast extract concentration of 6 g/L, compared with the fermentation without the addition. As for the overall sensory scores of the obtained wine, we observed a gradual increase with an increase in the amounts of yeast extract added and reached the maximum of 8.910 ± 0.300 when 6 g/L yeast extract was supplemented in the fermentation system; however, there was no significant change with a further increase in the supplementation of the yeast extract. This phenomenon was attributed to the fact that the appropriate amount of yeast extract promoted aroma production, while an excess induced the growth of the strain but was disadvantageous to the formation of the end product. Thus, both the productivity and the sensory characteristics were improved at 6 g/L of yeast extract (Table 6). Further investigation is required to study the impact of the supplementation of nitrogen on the metabolic pathways involved in producing and regulating aroma, as well as to understand how nitrogen affects the process of yeast metabolism in alcoholic fermentation, which will promote an improved utilization of this nutrient to produce volatile compounds.

Pilot-scale fermentation
After the completion of the fermentation process at the laboratory scale, the optimized parameters were tested in a scaled-up system of 1000 L reactor to investigate the possibility of industrializing the production of the Chinese chestnut wine. The system was used for producing three batches of Chinese chestnut fermented wine, with an agitation speed of 100 rpm and at a temperature of 30 °C. The Fig. 4 demonstrates the typical profiles of total cell growth, reducing sugar consumption and ethanol formation over time in the 1000 L bioreactor. After the initial 12-h lag phase, we observed an increase in the process of both reducing sugar consumption and ethanol production. The maximal growth for strain HN006 (2.065 × 10 8 CFU/mL) and the strain HN010 (4.201 × 10 8 CFU/mL) was observed after 48 h and 36 h of fermentation, respectively. The highest concentration of ethanol was 190.56 g/L, with a corresponding yield of 0.487 g/g. Over a 3-d period, we successfully completed three runs using the co-culture fermentation. In these three experiments, we observed the key fermentation parameters: final ethanol concentration, volumetric ethanol productivity, and ethanol yield to be consistent. The fermentation performance was successfully demonstrated in the 1000 L pilot-scale bioreactor and would be suitable for the industrial co-culture fermentation Chinese chestnut wine.

Discussion
The flavour of distilled wine depends not only on the raw materials and manufacturing technique, but also on the characteristics of yeast and other microorganisms (Alicia et al. 2018). Yeasts are important and can be detected during the whole manufacturing process. Not only do they directly determine the fermentation rate, but they can also transform nutrients into a wide variety of volatile flavour compounds (Ayşe et al. 2018;Wang et al. 2019). Therefore, yeasts are one of the key factors affecting the flavour type and product quality of whiskey (Shi et al. 2019;Wu et al. 2012).
At present, most whiskeys are fermented by Saccharomyces cerevisiae. Although the single fermentation mode by Saccharomyces cerevisiae obtains high alcohol content, but it has the phenomenon of homogenization of whiskey flavor. Recently years, a large number of studies have shown that non-Saccharomyces cerevisiae can produce alcohols, esters, acids, alkanes, aromatic hydrocarbons, ketones and other metabolites in the fermentation process, which can enhance the flavour and enrich the taste of the wine, however, it normally showed poor ethanol fermentation capacity. Nowadays, co-fermentation of selected non-Saccharomyces yeasts with S. cerevisiae received more and more acceptance to avoid the defects of pure fermentation and spontaneous fermentation (Luan et al. 2018). Remarkably, the application of mixed cultures of Saccharomyces cerevisiae and non-Saccharomyces cerevisiae has been reported to make important contributions to the formation of aroma substances, and also with high efficiency of ethanol production (Yan et al. 2019a, b). Additionally, the use of different species of yeast has proved to improve the percieved flavour in wine (Alicia et al. 2018). Therefore, the selection of yeast strains is critical for the formation of wine aroma, which is one of the key attributes affecting consumer acceptance . Yuan et al. (2019) selected one excellent non-Saccharomyces cerevisiae strain from 6 aroma-producing yeast strains and co-cultured with Saccharomyces cerevisiae for Yali fruit wine fermentation and the results demonstrated that the mixed fermentation helped to enhance the aroma and improve the complexity and characteristics of the wine.
In present study, a novel whiskey using Chinese chestnut as the main raw material was developed. First, we isolated the yeast strains and investigated their potential to produce effective Chinese chestnut wine, and the results demonstrated that the strain HN010, which showed the maximum production of the volatile compounds and the best sensory characteristics, along with the strain HN006 that possessed the highest efficiency in ethanol production were adopted as the most promising yeasts as they combined satisfactory Chinese chestnut wine fermentation with an interesting flavor profile. Further optimization of the Chinese chestnut wine fermentation using a co-culture of strain HN006 and strain HN010 demonstrated that the optimum conditions for chestnut wine production were achieved at a mixed inoculum volume of 6% (v/v) with an HN006/HN010 inoculation ratio of 1:2 (v/v), a raw materials ratio of 5:3:2 (chestnut whiskey: malt: glutinous rice), and a yeast extract concentration of 6 g/L. Additionally, a scaled-up fermentation of the Chinese chestnut wine was successfully carried out in a 1000 L pilot-scale system. This study is the first report to explore the fermentation of Chinese chestnut wine using a mixed culture system. The co-culture of strain HN006 and strain HN010 showed potential application in the Chinese chestnut wine fermentation process. Additionally, the methods used in this study could provide a generalized experimental scheme to initiate large-scale studies for the identification of useful microbes for the fermentation of wine as well as other food products.