Proteomic and high-throughput analysis of protein expression and microbial diversity of microbes from 30- and 300-year pit muds of Chinese Luzhou-flavor liquor

Food Research International 75 (2015) 305–314

Contents lists available at ScienceDirect

Food Research International journal homepage: www.elsevier.com/locate/foodres

Proteomic and high-throughput analysis of protein expression and microbial diversity of microbes from 30- and 300-year pit muds of Chinese Luzhou-flavor liquor Qi Zheng a,b,1, Bairong Lin a,1, Yibin Wang a, Qiuping Zhang a, Xinxin He a, Ping Yang b, Jun Zhou b, Xiong Guan a,⁎, Xiaohong Huang a,⁎ a b

Key Laboratory of Biopesticide and Chemical Biology, Ministry of Education, Fujian Agriculture and Forestry University, Fuzhou 350002, People's Republic of China Lu Zhou Lao Jiao Co., Ltd., Luzhou, Sichuan 646000, People's Republic of China

a r t i c l e

i n f o

Article history: Received 25 November 2014 Received in revised form 22 May 2015 Accepted 17 June 2015 Available online 20 June 2015 Keywords: Luzhou-flavor Chinese liquor iTRAQ 16S rDNA Proteomics Microbial diversities Chemical compounds studied in this article: Methane (PubChem CID: 297) Butyric acid (PubChem CID: 264) Caproic acid (PubChem: 8892)

a b s t r a c t Luzhou-flavor liquor is fermented based on the metabolism of special microbial communities in pit. In this study, total proteins and DNAs of microbes from 30- and 300-year pit muds were firstly extracted. Meanwhile, an efficient approach for protein extraction with increased protein content was optimized. iTRAQ-based proteomic was then applied to investigate the aroma-forming functional protein expression of microbes from the samples. Furthermore, high-throughput sequencing of 16S rDNA was employed to reveal microbial diversity. We comparatively identified 63 proteins of aroma-forming functional microbes in these samples, and found that 59 of these proteins were highly expressed in the 300-year pit mud. Those aroma-forming functional proteins were found to be involved in methanogenesis, as well as the formation of caproic acid and butyric acid during the liquor fermentation. High-throughput sequencing revealed that the microbes most commonly found in both samples were members of phylum Firmicutes (by 97% sequence similarity), both of which, along with another common Methanobacterium, were important components of aroma-forming functional colonies in the pit muds for the brewing of Chinese liquor. The findings in this study afford us new insight into the different protein expression levels and microbial communities in two pit muds. © 2015 Elsevier Ltd. All rights reserved.

1. Introduction Luzhou-flavor liquor is one of the foremost traditional distilled liquors in China. Many famous Luzhou-flavor liquors, such as Luzhou Laojiao, Jiannanchuan and Wuliangye, are all fermented from grains such as sorghum and rice in earth cellars. The production of desired liquor is related to either pit mud or pits. The pit mud is the main substance contributing to liquor fermentation. The fermentation refers to a complex metabolic process in which a variety of microbes colonizing the pits or pit mud engage in three phase boundaries of solid, liquid and gas (Luo, Liu, Tian, Liang, & Xiang, 2011). After long-term cultivation and diversification of pit mud microbes, the species were enriched and slowly formed a microecological bacterial system dominated by methanogens

⁎ Corresponding authors at: Key Laboratory of Biopesticide and Chemical Biology, Ministry of Education, Fujian Agriculture and Forestry University, Fuzhou 350002, People's Republic of China. E-mail addresses: [email protected] (X. Guan), [email protected] (X. Huang). 1 These authors contributed equally to this work.

http://dx.doi.org/10.1016/j.foodres.2015.06.029 0963-9969/© 2015 Elsevier Ltd. All rights reserved.

and caproic acid forming clostridia (Deng, Tang, & Zhang, 2010). The composite aroma produced by the metabolism would further lead to rich aroma of Luzhou-flavor liquor. Along with different fermentation stages, the bacteria population structure changes as well; thus, the quality of pit mud is changed year after year. Therefore, it is essentially important to elucidate the correlation between the protein expression of microbes from pit mud of different ages and the production of Luzhou-flavor liquor, which would be helpful for the development of Chinese liquor industry. Systematic identification and quantification of the proteome require high-throughput molecular biology techniques. Isobaric Tags for Relative and Absolute Quantitation (iTRAQ) is a technique launched by Applied Biosystems Inc. in 2004 (Wang, Dai, & Tu, 2010). The technique combines multidimensional liquid chromatography and tandem mass spectrometry, which was widely used for proteomics investigation (Wu, He, & Jiang, 2013). The iTRAQ technique could be employed for relative or absolute quantification of either four or eight different protein samples simultaneously (Liu et al., 2012). iTRAQ is a powerful method which could be applied to differential proteomics research for profiling protein expression changes, in particular of protein detection and identification in a variety of pit mud

306

Q. Zheng et al. / Food Research International 75 (2015) 305–314

microbes. The objective of this study was to determine the correlation between protein function and aroma compounds of liquor by comparing the protein expression levels in the 30- and 300-year pit muds of Luzhou-flavor liquor with iTRAQ technique followedby a bioinformatics analysis. 2. Materials and methods 2.1. Sampling and location of pit mud The pit mud samples were collected from two pits used for production of liquor by the Luzhou Laojiao Group (Luzhou, Sichuan, China, 105°270′86″ E, 28°53′15″ N; April 2014). Samples were taken from two locations (the wall and the bottom of the cellar) of three individual pits with 30 years (PM30) and 300 years (PM300) old, respectively. Each sample plot was divided into five sub-plots. All samples taken from the same year were mixed together. After sampling, the two pit mud samples (PM30 and PM300) were transferred to sterile polyethylene bags and stored at −80 °C for further use. 2.2. Protein extraction Four different extraction protocols were applied in this study. For each protocol, 1 g of pit mud was amended with 10% (w/w) polyvinylpolypyrrolidone (PVPP) by grinding with a pestle to prevent contamination from other substances, such as humic acid. The proteins in the samples were individually extracted with sodium-dodecyl sulfate (SDS) method (Keiblinger et al., 2012), NaOH method (Benndorf, Balcke, Harms, & von Bergen, 2007), SDS–phenol method (Keiblinger et al., 2012) and C/S–P–M method (Wang, Zhang, et al., 2011) as described previously. 2.3. SDS-PAGE and determination of protein concentration SDS-PAGE and determination of protein concentration were performed with previously described methods (Bastida, Hernández, & García, 2014; Laemmli, 1970). Briefly, for SDS-PAGE, proteins were dissolved into a buffer solution (20% of glycerol, 2% of 2-mercaptoethanol, 4% of SDS, 0.1 M Tris–HCl (pH 6.8), and 0.2% of bromophenol blue), and then incubated at 90 °C for 5 min, followed by loading on SDS gel (4% stacking gel, 10% separating gel). For determination of protein concentration, 20 μL of protein solution was mixed with 1 mL of Quick start Bradford reagent (Bio-Rad) and then incubated at room temperature for 10 min. Protein concentration was determined with a spectrophotometer (Thermo Electron Corporation Heliosα) at a wave length of 595 nm referring to a standard calibration curve produced with bovine serum albumin (BSA). 2.4. Protein digestion and labeling with iTRAQ reagents Trypsin digestion and iTRAQ labeling were performed according to the manufacturer's protocol (Applied Biosystems, Foster City, CA). Each sample with 100 μg of protein was further reduced, alkylated, and then digested at 37 °C with trypsin (mass spectrometry grade; Promega, Madison, WI) overnight. Finally, the samples were labeled with iTRAQ™ reagents (Applied Biosystems). The iTRAQ labeling of proteins were listed as follows: PM30 were labeled with tags 113, PM300 were labeled with tags 114. Two labeled digests were then mixed and dried.

exchange fraction was dried and then dissolved into buffer C (5% acetonitrile, 0.1% formic acid). Finally, the samples were analyzed on Qstar Elite (Applied Biosystems). Briefly, peptides were separated on a reversed-phase column (ZORBAX 300SB-C18 column, 5 μm, 300 Å, 0.1 ∗ 15 mm; Micromass) using a 20AD HPLC system (Shimadzu). The HPLC gradient of buffer D (95% acetonitrile, 0.1% formic acid) was 5–35% which was added to buffer C at a flow rate of 0.2 μL/min for 65 min. Survey scans were acquired from m/z 350–1800 with up to four precursors selected for MS/MS from m/z 100–2000 using a dynamic exclusion of 30 S. The ratios of peak areas of the iTRAQ reporter ions reflect the relative abundances of the peptides, indirectly reflecting the proteins in the samples. Larger, sequence-information-rich fragment ions were also produced under these MS/MS conditions, referring to the identity of the protein. 2.6. Data analysis The software used for data acquisition was Analyst QS 2.0 (Applied Biosystems). The software used for protein identification and quantitation was ProteinPilot™ 4.5 software (Software Revision Number: 1656; Applied Biosystems) with the integrated Paragon™ search algorithm (Revision Number: 4.5.0.0, 1654; Applied Biosystems). The data analysis parameters were set as follows: Sample type: iTRAQ (peptide labeled); Cys alkylation: Methanethiosulfonate (MMTS); Digestion: Trypsin; Instrument: QSTAR Elite; Species: Bacteria, Fungi and Archaea; ID Focus: Biological modifications; Database: Uniprot Bacteria database (downloaded September 2014, 69,369,645 sequences), Uniprot Archaea database (downloaded September 2014, 865,138 sequences), Uniprot Fungi database (downloaded September 2014, 3,178,344 sequences); Search Effort: Thorough; Max missed cleavages: 2; User Modified Parameter Files: No; Bias Correction: Auto; and Background Correction: Yes. Identified proteins were grouped by the software to minimize redundancy. All peptides used for the calculation of protein ratios were unique to the given protein or proteins within the groups. The peptides similar to other isoforms or proteins of the same family were ignored. The protein confidence threshold cutoff was set as 1.3 (unused ProtScore) with at least one peptide with 95% confidence. The biological process and molecular function classification of these proteins were annotated using the Gene Ontology (GO) database. 2.7. Extraction of total DNAs and Illumina 16S rDNA V4 library preparation The extraction of total DNAs was performed with a method described previously (Wang, Zhang, et al., 2011). To amplify the V4 region of 16S rDNA gene, a pair of universal bacterial primers developed by Caporaso et al. (2010) was employed in PCR reactions: a forward primer (515f) GTG CCA GCM GCC GCG GTA A and a reverse primer (806r) GGA CTA CHV GGG TWT CTA AT. In this study, the reverse primer was modified by adding a 6-bp error-correcting barcode, which is unique to each sample and serves as a multiplexing marker. PCR reactions were performed using NEB Phusion PCR Master mix with GC buffer(New England Biolabs, Ipswich, MA, USA)following the manufacturer's instructions, and then the amplified products were loaded on 1% agarose gel and stained with ethidium bromide. For gel purification that was performed with an E.Z.N.A.TM Gel Extraction Kit (OMEGA Bio-tekInc., Norcross, GA, USA). Once DNA concentration and quality were determined, a multiplexing sample was created by pooling equimolar ratios of purified ~300-bp amplicons from the individual samples. 2.8. Illumina MiSeq sequencing and data analysis

2.5. Off-line 2D LC–MS/MS The mixed peptides were fractionated by strong cation exchange chromatography on a 20AD HPLC system (Shimadzu) by using a polysulfoethyl column (2.1 ∗ 100 mm, 5 μm, 300 Å; the Nest Group Inc.) according to a previous study (Xiao et al., 2010). Each strong cation

Sequencing was conducted with a paired-end, 2 × 300 bp cycle running on an Illumina MiSeq sequencing system and MiSeq Reagent Nano Kit version 2 (500 cycle) chemistry. After initial QC processing, pairs of reads were merged using FLASH (V1.2.7, http://ccb.jhu.edu/ software/FLASH/) (Caporaso et al., 2011) and were demultiplexed for

Q. Zheng et al. / Food Research International 75 (2015) 305–314

each sample according to the unique sample barcode sequence. The Quantitative Insights into Microbial Ecology (QIIME, V1.6.0, http:// qiime.org/index.html) software package and UPARSE pipeline (Liu & Wang, 2007) were used to process the sequence reads, and custom Perl scripts were used to analyze within-sample alpha diversity. 3. Results 3.1. Comparison of protein extraction protocols from 30- and 300-year pit muds Compared with the other three protein extraction protocols, the bands of 30- and 300-year protein samples extracted with the C/S–P–M method in SDS-PAGE gels were clearer. Interestingly, much more low- and high-molecular weight proteins were obtained using this protocol (Fig. 1). The concentration of protein extracted with C/S–P–M method was 29.3 mg/mL from the 300-year pit mud and 25.9 mg/mL from the 30-year pit mud, respectively. The values were higher than those of SDS method (11.6 mg/mL and 11.1 mg/mL from the 300-year and the 30-year pit mud samples, respectively), NaOH method (18.9 mg/mL and 16.3 mg/mL from the 300-year and the 30-year pit mud samples, respectively) and SDS–phenol method (15.1 mg/mL and 11.6 mg/mL from the 300-year and the 30-year pit mud samples, respectively). Thus, the C/S–P–M method was chosen for protein extraction of pit mud in the following studies.

307

Nitrococcus mobilis displayed distinct functions. Consequently, we did not eliminate the redundancy of the proteins. We used either a high threshold of 1.2-fold or a low threshold of 0.8-fold to assess significant changes in protein expressions (p b 0.05). We found that the expressions of 114,029 proteins were significantly different, 99,169 proteins were classified into 165 groups which were highly expressed in PM300, while 14,860 proteins were classified into 94 groups which were expressed in high quantities in PM30. GO analysis revealed that the molecular functions of the proteins were mainly divided into 29 parts. 35% of these proteins were related to ion binding process, followed by oxidoreductase activity (Fig. 2A). In regard to biological process, biosynthetic process contributed the most (15%), followed by small molecule metabolic process of 11% (Fig. 2B). As for methanogens, Clostridium butyricum and Clostridium kluyveri were the main aroma-forming functional bacteria during the brewing of liquor. Thus, we focused on the proteins derived from the above microorganisms (Tables A.1, A.2 and A.3). We found that 63 aroma-forming functional proteins were involved in the production of high quality liquor (Table A.4). They were involved in the formation of flavor components of liquor such as butyric acid and caproic acid. GO analysis was performed on these 63 proteins (Fig. 2C and 2D). With respect to molecular function, 41% of proteins were related to oxidoreductase activity, followed by ion binding (33%), transferase activity (11%) and methyltransferase activity (7%) (Fig. 2C). With further respect to biological processes, 33% were related to the generation of precursor metabolites and energy, and 29% to the biosynthetic process (Fig. 2D).

3.2. Quantitative identification of pit mud proteins using iTRAQ

3.3. OTU picking and taxonomic assignment

To detect the proteome difference of pit mud between the 30-year and 300-year pit muds, a quantitative shotgun proteomics technique (iTRAQ) was employed in this study. We identified 163,844 peptides, which were classified into 305 protein groups. Consulting with Bacteria, Archaea and Fungi databases, we found that the 305 protein groups contain 135,930 proteins. It is well-known that various microorganisms inhabit the pit mud. The homologous proteins in different microbes may exhibit different functions (Wu & Mandrand-Berthelot, 1995). For instance, three homologous proteins, RNA polymerase-associated protein RapA in Psychrobacter aquaticus, methyl-accepting chemotaxis protein in Pseudomonas, and pyochelin synthetase F protein in

In order to obtain species information from the pit mud samples, clean sequencing tags were clustered into OTUs by 97% sequence similarity using the UPARSE software. As shown in Fig. 3A, PM30 had 190,268 total and 182,311 taxon tags that were higher than those in PM300 (73,297 and 64,028, respectively). They were all classified but few unique tags were presented in both samples. However, PM300 had more OTUs than that of PM30 (1164 vs. 1062). Meanwhile, a higher OTU density in PM300 than that in PM30 (1164/64,028; 1.8% vs. 1062/ 182,113; 0.6%) was observed. Among all OTUs built by UPARSE, only one OTU with the highest frequency of sequencing tag was used to assign taxonomy. Based on taxonomic assignment in each sample, the

Fig. 1. Proteins extracted by different protocols separated by SDS-PAGE. (A) Proteins extracted by C/S–P–M method; (B) pit mud proteins extracted by NaOH method, SDS–phenol method and SDS method.

308

Q. Zheng et al. / Food Research International 75 (2015) 305–314

number of tags was counted at all levels of classification (Fig. 3B) and the relative abundance was calculated at the level of phylum (Fig. 3C). As shown in Fig. 3B, the sequence tag was most abundant at the level of genus in both PM30 and PM300 (171,982 vs. 38,173 and 94.3% vs. 59.6%, respectively) and it was more abundant at the level of family in PM300 (16,693, 26.0%) than that of PM30 (6388, 3.5%). As shown in Fig. 3C, there were 11 phyla in the two samples, of which Firmicutes was most abundant in both samples (N 99% in PM30 and ~ 75% in

PM300). In addition, relative abundant Euryarchaeota (~ 4%) and Bacteroidetes (~2%) were also observed in PM300. We then performed statistical analysis on classification trees based on the top 20 species with most abundant sequence tags. As shown in Fig. 4A, the microbes in PM30 were mostly classified into two kingdoms: Archaea and Bacteria. Archaea accounted for a quite low percentage (0.06%) in all species and almost all of them were methanobacteria of the phylum Euryarchaeota in the Archaea kingdom. In the Bacteria

Fig. 2. Gene ontology (GO) assignment of proteins in pit mud. (A) Molecular function of the entire proteins: 35% were related to ion binding; (B) biological process (15%) of the entire proteins: biosynthesis was the dominant feature in biological process category; (C) biological process of aroma-forming functional proteins: 33% were related to generation of precursor metabolites and energy, followed by biosynthetic process (29%); (D) molecular function of aroma-forming functional proteins: oxidoreductase activity (41%) was the dominant feature in the molecular function category, followed by ion binding (33%).

Q. Zheng et al. / Food Research International 75 (2015) 305–314

C

D Fig. 2 (continued).

kingdom, however, Firmicutes was most abundant and accounted for as high as 92.9% of all species. This finding is summarized in Fig. 3C. Fig. 4B shows that the microbes in PM300 were also mostly classified into the two main kingdoms — Archaea and Bacteria which accounted for about 54.0% of all species (43.7% of Bacteria and 8.8% of Archaea). Although Archaea was less abundant than Bacteria in PM300, it was much higher than that of PM30 (8.8% vs. 0.06%). This finding is also shown in Fig. 3C. 3.4. Within-sample alpha diversity The within-sample alpha diversities in PM30 and PM300 are illustrated by curves created with the abundance of sequencing tags against the three metrics (rarefaction curve, Chao1 index, and Shannon index) computed with the UPARSE software (Fig. 5A–C). The rarefaction curve (Fig. 5A) suggests a near full coverage of microbial communities regarding PM30 and PM300 samples, with sequencing depth around 60,000. The Chao1 index curve (Fig. 5B) demonstrates a high level of microbial diversity at each sample; and the Shannon index curve (Fig. 5C) reveals a significantly higher diversity index of PM300 than that of PM30. 4. Discussions The aroma components in Luzhou-flavor liquor are generated during the entire fermentation process, including decomposition of

309

raw materials via microbes and/or enzymes, metabolism of microbes, as well as the complex interactions between all kinds of metabolites from microbes and raw materials (Fan & Qian, 2005; Shen, 2005; Zhuang, 2007). The interactions in microbes in the pit mud are crucial for solid state fermentation of Chinese liquor. There was no report in microbial field in respect to protein expression of pit muds so far. We benchmarked our approach by describing the metaproteomes of pit muds with different ages. Pit mud provides a suitable habitat for the growth of brewing microbiota. Some special communities have been cultivated for more than 20 years in order to produce high quality liquor (Ding, Wu, Huang, Li, & Zhou, 2014). As is well known, the older the pit mud is, the higher is the quality of the Chinese liquor (Zheng et al., 2013). We chose two pit muds with different ages and investigated the correlation of their protein expression and quality of liquor, i.e., flavor components. We employed high-throughput sequencing of 16S rDNA to elucidate the microbial diversities of the two samples, which allows us to compare the difference between species and abundance of microorganisms in pit muds. Compared with other genomics approaches such as transcriptomics, metaproteomics has the advantage of retrieving the actual and stable gene products which could be further investigated (Kolmeder et al., 2012). iTRAQ is a gel-free technique that uses isotope-coded covalent tags to quantify proteins from different samples in a single experiment (Han et al., 2013; Pütz, Stephanie, Boehm, Stiewe, & Sickmann, 2012). Due to its relatively high reproducibility and sensitivity, iTRAQ was widely employed in quantitative proteomics studies (Martyniuk, Alvarez, & Denslow, 2012). In this study, we optimized one desirable approach from four methods to extract proteins for pit muds and demonstrated the feasibility of iTRAQ technique for the analysis of mixed microbe samples. To increase the efficiency and throughput of our study, we selected proteins derived from aroma-forming functional microbes in pit muds and focused our attention to the proteins associated with the liquor quality (i.e., flavor components). We demonstrated that the iTRAQ analysis was efficient to reveal the difference of protein expression in pit muds with different ages. The iTRAQ coupled to 2D LC–MS/MS analysis identified 135,930 proteins of 305 protein groups, of which 114,029 proteins displayed significantly different abundance between PM30 and PM300. 63 of them were involved in the formation of flavor components. These proteins were mainly derived from methanogens, C. butyricum and C. kluyveri and 59 of these proteins were highly expressed in PM300 (Table A.4). Methanogens play an important role in fermentation that leads to production of high quality liquor (Zhang & Liang, 1996). We identified 44 proteins that belong to methanogens, of which 40 proteins were highly expressed in PM300 sample (Table A.4). These 44 proteins were derived from the genera Methanoculleus, Methanobrevibacter, Methanoplanus, Methanotorris, Methanosarcina, Methanolobus, Methanothermobacter, Methanobacterium and Methanomethylovorans, respectively. Methanoculleus and Methanoplanus are affiliated with the family Methanomicrobiaceae of the order Methanomicrobiales. Methanobrevibacter, Methanothermobacter and Methanobacterium belong to the family Methanobacteriaceae of order Methanobacteriales. Methanosarcina, Methanolobus and Methanomethylovorans belong to the family Methanosarcinaceae of the order Methanosarcinales. Methanotorris is affiliated to the family Methanocaldcoccaceae of the order Methanococcales. Methanococcales and Methanobacteriales are hydrogenotrophic methanogens, used H2 or formate as the reducing agent, and some species could use secondary alcohols as electron donors (Liu & Whitman, 2008). Methanomicrobiales could utilize CO2 and H2 or acetate to produce methane (Fu & Xin, 2009). Furthermore, the order Methanosarcinaceae could not only use CO2 and H2 or acetate to produce methane, but also utilize one-carbon compounds such as methylamines, methanol and methyl thiols for methanogenesis (Galagan et al., 2002). These biological processes promote the reactions

310

Q. Zheng et al. / Food Research International 75 (2015) 305–314

Fig. 3. OUT analysis and species annotation. (A) Classification and distribution of the sequence tags and operational taxonomic units; (B) constitution of the sequence tag at all levels of classification in the biological species; (C) relative abundance of the biological species at the level of phylum.

that produce organic acids such as acetic, butyric and caproic acids, which in turn improve the contents of various esters. The esters, in particular of ethyl caproate, are the main compounds that dominantly determine the quality and aromatic style of Luzhou-flavor liquor (Wang

et al., 2014). Proteins, such as methyl-coenzyme M reductase and mtrA exhibit different protein expressions in different genera of methanogens, however, with high expression trend. These 40 proteins (Table A.4) highly expressed in PM300 sample might result in high

Q. Zheng et al. / Food Research International 75 (2015) 305–314

Fig. 4. Classification tree of the biological species in (A) the PM30 and (B) the PM300 samples.

311

312

Q. Zheng et al. / Food Research International 75 (2015) 305–314

Fig. 5. Within-sample alpha diversity curves. (A) Rarefaction curve; (B) Chao1 curve; and (C) Shannon index.

production of organic acids and esters during the brewing of liquor, thus in turn result in generation of more flavor compounds. Through metagenome analysis, we found that the abundance of Euryarchaeota was extremely high in PM300 sample, even higher than that of Firmicutes (Fig. A.1). However, in PM30 sample, the abundance of Firmicutes was much higher than that of Euryarchaeota (Fig. A.1). Taxonomic analysis revealed that a much higher percentage of Methanobacterium (phylum Archaea) was observed in PM300 sample compared to that of PM30 sample (Fig. 3C), which was consistent with our iTRAQ results (Table A.4). The abundance of methanogens and protein expression in PM300 were both higher than those of PM30. Nevertheless, the relationship between abundance of methanogens and protein expression needs to be further investigated. Clostridium bacteria also play an important role in liquor fermentation (Zhang et al., 2005), contributing to produce aroma compounds such as acetic acid, butyric acid and caproic acid by utilizing starch (Zheng et al., 2013). C. butyricum and C. kluyveri are the major aromaforming Clostridium bacteria in pit muds. C. butyricum is an obligate anaerobic bacterium and its main products are butyric and acetic acid,

as well as CO2 and H2 deriving from various carbohydrates including glucose, fructose, disaccharides, lactose, xylose, n-butanol, sucrose, and fructose (Zhou et al., 2014). C. kluyveri utilizes ethanol to produce acetate, butyrate, caparoate and H2 under the synergistic methanogens (Bornstein & Barker, 1948). Moreover, butanol and hexanol are also formed at the end of the growth process (Seedorf et al., 2008). Acetic, butyric and caproic acids are important organic acids during the brewing of liquor. On one hand, acetic and butyric acids could be used by C. kluyveri as a substrate to produce caproic acid (Ding, Tan, & Wang, 2010); on the other hand, these acids could further react with ethanol to produce ethyl acetate, butyrate and caproate, which are the sources of aromatic flavor in Luzhou-flavor liquor (Hou et al., 2013). We identified 8 aroma-forming proteins in C. butyricum and 7 aromaforming proteins in C. kluyveri, which were highly expressed in PM300 sample (Table A.4). This result indirectly indicates that more aroma compounds would be formed. Furthermore, previous studies found that the concentration of flavor components were higher in old cellar than that in new one (Tao et al., 2014; Zheng et al., 2013), which was in consistence with our results that the protein expression of pit mud in 300-year cellar was higher than that of 30-year cellar.

Q. Zheng et al. / Food Research International 75 (2015) 305–314

High-throughput sequencing of 16S rDNA revealed that Clostridium and Lactobacillus in the phylum Firmicutes were the two main microbes in both PM30 and PM300 samples. This finding is consistent with the fact that only such microbe with extremely high adaptability could survive under such high level of alcohol, low oxygen and low pH environment (Wang et al., 2014). Lactobacillus is one of the aroma forming functional microbes during the liquor brewing (Zheng et al., 2012). It produces bacteriocin inhibiting the growth of spoilage organisms and spoilage (Lv et al., 2013). In addition, lactic acid could further react with ethanol to produce the ethyl lactate, which is one of the major flavor compounds of Luzhou-flavor liquor (Ding, Wu, Zhang, Zheng, & Zhou, 2014). However, accumulation of Lactobacillus restrains the production of caproic acid (Tao et al., 2014, Yao, Chen, Zhen, & Guo, 2010). In PM30, Clostridium accounted for 0.4% and Lactobacillus accounted for 90.4% of all species, whereas, in the PM300, Clostridium accounted for 2.7%and Lactobacillus accounted for 22.3% of all species, which was consistent with a previous study (Deng et al., 2012). The abundance of C. butyricum accounted for 0.025% in PM300 sample, in comparison of 0.004% in PM30 sample. The abundance of C. kluyveri was not observed by high-throughput sequencing of 16S rDNA in both samples. It was related to the fact that the data shown in classification trees were only the top 20 species with most abundant sequence tags, exclusive of the sequence tags of C. kluyveri. However, metaproteomics analysis revealed the existence of C. kluyveri in both samples. Consequently, a combination of iTRAQ and high-throughput could more precisely reveal the microbial communities in pit muds. iTRAQ-based proteomics revealed the existence of Fungi in both samples, such as Aspergillus, Saccharomyces, Wickerhamomyces and Pichia (data not shown), which was related to the addition of Daqu stater (Zheng et al., 2013). It is known that Daqu was used as inoculums for liquor brewing (Zheng et al., 2012), containing abundant microbes such as Aspergillus, Pseudomonas, Lactobacillus, Pichia, Wickerhamomyces and Saccharomyces (Wang, Gao, Fan, & Xu, 2011). Daqu was mixed with 9fold weight of raw material and stacked on the ground for 2–7 days, then put into pits and sealed for liquor fermentation (Chen, Wu, & Xu, 2014). During the long-term usage of pit, the microbes may exchange between the pit mud and Daqu during the liquor brewing. Consequently, the microorganisms in Daqu could significantly affect the microbial diversity in pit mud (Zheng et al., 2013). The microbiota in pit mud is interdependent and interacts with one another. For example, a previous study confirmed that methanogens and C. kluyveri are mutualistic (Tao et al., 2014). During the formation of caproic acids, hydrogen is also generated under anaerobic condition (Ding et al., 2010). The interspecies hydrogen transfer between caproic acid-producing and methanogens counteract the hydrogen pressure and makes the reaction towards generation of more caproic acids. The accumulation of butyric acid is conducive to generate caproic acid by C. kluyveri. These interactions of the microorganisms result in the unique flavor of Luzhou-flavor liquor.

5. Conclusions The aroma-forming functional proteins in 300-year pit mud were highly expressed with much higher content than that of 30-year pit mud, which was contributed to the production of more organic compounds during the fermentation of the 300-year pit mud. High-throughput sequencing of 16S rDNA revealed that aromaforming functional microorganisms like methanogens and Clostridium were higher in abundance in PM300 sample. This result was in accordance with the result of iTRAQ. The diversification of pit mud microbial composition, metabolic action of microorganisms and interactions between microorganisms were responsible for the production of more flavorful liquors. In general, our metaproteomics and high-throughput sequencing approach in this study provide us efficient tools to reliably identify pit mud microbial diversity and abundance.

313

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.foodres.2015.06.029. Acknowledgment We wish to thank Dr. Brian McGarvey for proofreading this paper. The present study was jointly funded by grants from the Foundation for Studies on Pit Mud of Vinasse for Liquor by DGGE and PLFA (the project number: ychx00018 (2013); Luzhou, Sichuan, China), and the Funds for Researches by Leading Personnel in Higher Education Institutions (the project number: k8012012a (2012); Fuzhou, Fujian, China). References Bastida, F., Hernández, T., & García, C. (2014). Metaproteomics of soils from semiarid environment: functional and phylogenetic information obtained with different protein extraction methods. Journal of Proteomics, 101, 31–42. Benndorf, D., Balcke, G. U., Harms, H., & von Bergen, M. (2007). Functional metaproteome analysis of protein extracts from contaminated soil and groundwater. ISME Journal, 1(3), 224–234. Bornstein, B., & Barker, H. (1948). The nutrition of Clostridium kluyveri. Journal of Bacteriology, 55(2), 223–230. Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D., Costello, E. K., Fierer, N., Pena, A. G., Goodrich, J. K., Gordon, J. I., Huttley, G. A., Kelley, S. T., Knights, D., Koenig, J. E., Ley, R. E., Lozupone, C. A., McDonald, D., Muegge, B. D., Pirrung, M., Reeder, J., Sevinsky, J. R., Turnbaugh, P. J., Walters, W. A., Widmann, J., Yatsunenko, T., Zaneveld, J., & Knight, R. (2010). QIIME allows analysis of high-throughput community sequencing data. Nature Methods, 7(5), 335–336. Caporaso, J. G., Lauber, C. L., Walters, W. A., Berg-Lyons, D., Lozupone, C. A., Turnbaugh, P. J. , Fierer, N., & Knight, R. (2011). Global patterns of 16S rRNA diversity at a depth of millions of sequences per sample. Proceedings of the National Academy of Sciences of the United State of America, 108(Suppl. 1), 4516–4522. Chen, B., Wu, Q., & Xu, Yan (2014). Filamentous fungal diversity and community structure associated with solid state fermentation of Chinese Maotao-flavor liquor. International Journal of Food Microbiology, 179, 80–84. Deng, B., Shen, C. H., Shan, X. H., Ao, Z. H., Zhao, J. S., Shen, X. J., & Huang, Z. G. (2012). PCRDGGE analysis on microbial communities in pit mud of cellars used for different periods of time. Journal of the Institute of Brewing, 118(1), 120–126. Deng, Y., Tang, Y. R., & Zhang, W. X. (2010). Application of 16S–23S rRNA ITS-AFLP fingerprint in the analysis of prokaryotic microbial diversity in liquor pit mud. (in Chinese) Liquor Making Science & Technology, 189(3), 46–47. Ding, H. B., Tan, G. Y.\A., & Wang, J. Y. (2010). Caproate formation in mixed-culture fermentative hydrogen production. Bioresource Technology, 101(24), 9550–9559. Ding, X. F., Wu, C. D., Huang, J., Li, H., & Zhou, R. Q. (2014). Eubacterial and archaeal community characteristics in the man-made pit mud revealed by combined PCR-DGGE and FISH analyses. Food Research International, 62, 1047–1053. Ding, X. F., Wu, C. D., Zhang, L. Q., Zheng, J., & Zhou, R. Q. (2014). Characterization of eubacterial and archaeal community diversity in pit mud of Chinese Luzhou-flavor liquor by nested PCR-DGGE. World Journal of Microbiology and Biotechnology, 30, 605–612. Fan, W. L., & Qian, M. C. (2005). Headspace solid phase microextraction and gas chromatography–olfactometry dilution analysis of young and aged Chinese “Yanghe Daqu” liquors. Journal of Agricultural and Food Chemistry, 53(20), 7931–7938. Fu, L., & Xin, M. (2009). Ecological diversity and industrial application of methanogens. (in Chinese) Chinese Journal of Applied & Environmental Biology, 15(4), 574–578. Galagan, J. E., Nusbaum, C., Roy, A., Endrizzi, M. G., Macdonald, P., FitzHugh, W., Calvo, S., Engels, R., Smirnov, S., Atnoor, D., Brown, A., Allen, N., Naylor, J., Stange-Thomann, N., DeArellano, K., Johnson, R., Linton, L., McEwan, P., Mckernan, K., Talamas, J., Tirrell, A., Ye, W., Zimmer, A., Barder, R. D., Cann, I., Graham, D. E., Grahame, D. A., Guss, A. M., Hedderich, R., Ingram-Smith, C., Kuettner, H. C., Krzycki, J. A., Leigh, J. A., Li, W., Liu, J., Mukhopadhyay, B., Reeve, J. N., Smith, K., Springer, T. A., Umayam, L. A., White, O., White, R. H., de Macario, E. C., Ferry, J. G., Jarrell, K. F., Jing, H., Macario, A. J. L., Paulsen, I., Pritchett, M., Sowers, K. R., Swanson, R. V., Zinder, S. H., Lander, E., Metcalf, W. W., & Birren, B. (2002). The genome of M. acetivorans reveals extensive metabolic and physiological diversity. Genome Research, 12(4), 532–542. Han, Z., Sun, J., Zhang, Y., He, F., Xu, Y., Matsumura, K., He, L., Qiu, J., Qi, S., & Qian, P. (2013). iTRAQ-based proteomic profiling of the barnacle Balanus amphitrite in response to the antifouling compound meleagrin. Journal of Proteome Research, 12(5), 2090–2100. Hou, X. G., Wang, J. Y., Li, X. S., Hu, B. Y., Li, S. L., & Gao, Y. Y. (2013). Research progress in main functional microbes in Nong-flavor liquor pits. (in Chinese) Liquor-Making Science & Technology, 224(2) (96–101,106). Keiblinger, K. M., Wilhartitz, I. C., Schneider, T., Roschitzki, B., Schmid, E., Eberl, L., Riedel, K., & Zechmeister-Boltenstern, Z. (2012). Soil metaproteomics—comparative evaluation of protein extraction protocols. Soil Biology & Biochemistry, 54(15–10), 14–24. Kolmeder, C. A., de Been, M., Nikkilä, J., Ritamo, I., Mättö, J., Valmu, L., Salojärvi, J., Palva, A., Salonen, A., & de Vos, W. M. (2012). Comparative metaproteomics and diversity analysis of human intestinal microbiota testifies for its temporal stability and expression of core functions. PLoS One, 7(1) (e29913). Laemmli, U. K. (1970). Cleavage of structural proteins during assembly of head of bacteriophage-T4. Nature, 227(5259), 680–685.

314

Q. Zheng et al. / Food Research International 75 (2015) 305–314

Liu, J. Y., Jiang, Y. J., Xie, B. G., Chen, B. Z., Liao, W., Deng, Y. J., & Liu, P. H. (2012). Analysis of the proteomics in Volvariella volvacea homokaryotic and heterokaryotic mycelia by iTRAQ-coupled 2-D LC–MS/MS. (in Chinese) Acta Edulis Fungi, 19(1), 7–12. Liu, G. Q., & Wang, X. L. (2007). Head-space gas chromatographic analysis for the volatile flavor compounds from Ganoderma lucidum submerged-cultured broth. (in Chinese) Mycosystema, 26(3), 389–395. Liu, Y., & Whitman, W. B. (2008). Metabolic, phylogenetic, and ecological diversity of the methanogenic archaea. Annals of the New York Academy of Sciences, 1125, 171–189. Luo, H., Liu, S., Tian, W., Liang, H. Z., & Xiang, W. L. (2011). Dynamic analysis of microbial ecological factors in the fermentation pit mud of Luzhou-flavor Chinese liquor. (in Chinese) China Brewing, 232(7), 108–112. Lv, X. C., Huang, R. L., Chen, F., Zhang, W., Rao, P. F., & Ni, L. (2013). Bacterial community dynamics during the traditional brewing of Wuyi Hong Qu glutinous rice wine as determined by culture-independent methods. Food Control, 34, 300–306. Martyniuk, C. J., Alvarez, S., & Denslow, N. D. (2012). DIGE and iTRAQ as biomarker discovery tools in aquatic toxicology. Ecotoxicology and Enviromental Safety, 76(2), 3–10. Pütz, S. M., Stephanie, M., Boehm, A. M., Stiewe, T., & Sickmann, A. (2012). iTRAQ analysis of a cell culture model for malignant transformation, including comparison with 2D-PAGE and SILAC. Journal of Proteome Research, 11(4), 2140–2153. Seedorf, H., Fricke, W. F., Veith, B. V., Brüggemann, H., Liesegang, H., Strittmatter, A., Miethke, M., Buckel, W., Hinderberger, J., Li, F., Hagemeier, C., Thauer, R. K., & Gottschalk, G. (2008). The genome of Clostridium kluyveri, a strict anaerobe with unique metabolic features. Proceedings of the National Academy of Sciences of the United States of America, 105(6), 2128–2133. Shen, Y. F. (2005). Key factors to the formation of flavor quality in liquor. (in Chinese) Liquor-Making Science & Technology, 26(11), 30–34. Tao, Y., Li, J., Rui, J., Xu, Z., Hu, X., Wang, X., Liu, M., Li, D., & Li, X. (2014). Prokaryotic communities in pit mud from different-aged cellars used for the production of Chinese strong-flavor liquor. Applied and Environmental Microbiology, 80(7), 2254–2260. Wang, C. D., Chen, Q., Wang, Q., Li, C. H., Leng, Y. Y., Li, S. G., Zhou, X. W., Han, W. J., Li, J. G., Zhang, X. H., & Li, Y. Z. (2014). Long-term batch brewing accumulates adaptive microbes, which comprehensively produce more flavorful Chinese liquors. Food Research International, 62, 894–901. Wang, L. Q., Dai, Y., & Tu, Z. G. (2010). iTRAQ labeling and biomarker discovery in comparative proteomic studies. (in Chinese) Chemistry of Life, 30(1), 135–140. Wang, H. Y., Gao, Y. B., Fan, Q. C., & Xu, Y. (2011). Characterization and comparison of microbial community of different typical Chinese liquor Daqus by PCR-DGGE. Letters in Applied Microbiology, 53, 134–140.

Wang, H. B., Zhang, Z. X., Li, H., He, H. B., Fang, C. X., Zhang, A. J., Li, Q. S., Chen, R. S., Guo, X. K., Lin, H. F., Wu, L. K., Lin, S., Chen, T., Lin, R. Y., Peng, X. X., & Lin, W. X. (2011). Characterization of metaproteomics in crop rhizospheric soil. Journal of Proteome Research, 10(3), 932–940. Wu, P., He, F. C., & Jiang, Y. (2013). Label-free methods in quantitative proteomics. Progress in Biochemistry and Biophysics, 40(3), 281–292. Wu, L. F., & Mandrand-Berthelot, M. A. (1995). A family of homologous substrate-binding proteins with a broad range of substrate specificity and dissimilar biological functions. Biochimie, 77, 744–750. Xiao, Z. F., Li, G. Q., Chen, Y. H., Li, M. Y., Peng, F., Li, C., Li, F., Yu, Y. H., Ouyang, Y. M., Xiao, Z. Q., & Chen, Z. C. (2010). Quantitative proteomic analysis of formalin-fixed and paraffin-embedded nasopharyngeal carcinoma using iTRAQ labeling, twodimensional liquid chromatography, and tandem mass spectrometry. Journal of Histochemistry & Cytochemistry, 58(6), 517–527. Yao, W., Chen, M., Zhen, D., & Guo, Y. (2010). Isolation of lactate-producing microbes from fermented grains of Luhou-flavor liquor and their effect on simulative solid-state fermentation. (in Chinese) Liquor Making, 37, 37–41. Zhang, H. S., & Liang, W. H. (1996). Study on methane bacteria and mechanisms in the fermentation of liquor. (in Chinese) Liquor-Making Science & Technology, 17(6), 18–19. Zhang, W., Qiao, Z., Shigematsu, T., Tang, Y., Hu, C., Morimura, S., & Kida, K. (2005). Analysis of the bacterial community in Zaopei during production of Chinese Luzhou-flavor liquor. Journal of the Institute of Brewing, 111(2), 215–222. Zheng, J., Liang, R., Zhang, L., Wu, C., Zhou, R., & Liao, X. (2013). Characterization of microbial communities in strong aromatic liquor fermentation pit muds of different ages assessed by combined DGGE and PLFA analyses. Food Research International, 54, 660–666. Zheng, X. W., Yan, Z., Han, B. Z., Zwietering, M. H., Samson, R. A., Boekhout, T., & Rober Nout, M. J. (2012). Complex microbiota of a Chinese “Fen” liquor fermentation starter (Fen-Daqu), revealed by culture-dependent and culture-independent methods. Food Microbiology, 31(2), 293–300. Zhou, X., Lu, X. H., Li, X. H., Xin, Z. J., Xie, J. R., Zhao, M. R., Wang, L., Du, W. Y., & Liang, J. P. (2014). Radiation induces acid tolerance of Clostridium tyrobutyrium and enhances bioproduction of butyric acid through a metabolic switch. Biotechnology for Biofuels, 7, 22–36. Zhuang, M. Y. (2007). Regulatory mechanisms for the formation of aromatic substances and fermentation skills in production of Chinese liquor. (in Chinese) Niang Jiu, 34(4), 109–113.