Purpose: Higher alcohol is a by-product of the fermentation of wine, and its content is one of the most important parameters that affect and are used to appraise the final quality of Chinese rice wine. Ammonium compensation is an efficient and convenient method to reduce the content of higher alcohols, but the molecule mechanism is poorly understood. Therefore, an iTRAQ-based proteomic analysis was designed to reveal the proteomic changes of Saccharomyces cerevisiae to elucidate the molecular mechanism of ammonium compensation in reducing the content of higher alcohols. Methods: The iTRAQ proteomic analysis method was used to analyze a blank group and an experimental group with an exogenous addition of 200 mg/L (NH ) HPO during inoculation. The extracted intracellular proteins were 4 2 4 processed by liquid chromatography-mass spectrometry and identified using bioinformatics tools. Real-time quantitative polymerase chain reaction was used to verify the gene expression of differentially expressed proteins. Results: About 4062 proteins, including 123 upregulated and 88 downregulated proteins, were identified by iTRAQ- based proteomic analysis. GO and KEGG analysis uncovered that significant proteins were concentrated during carbohydrate metabolism, such as carbon metabolism, glyoxylate, and dicarboxylate metabolism, pyruvate metabolism, and the nitrogen metabolism, such as amino acid synthesis and catabolism pathway. In accordance with the trend of differential protein regulation in the central carbon metabolism pathway and the analysis of carbon metabolic flux, a possible regulatory model was proposed and verified, in which ammonium compensation facilitated glucose consumption, regulated metabolic flow direction into tricarboxylic acid, and further led to a decrease in higher alcohols. The results of RT-qPCR confirmed the authenticity of the proteomic analysis results at the level of gene. (Continued on next page) * Correspondence: email@example.com Guidong Huang and Hong Ren contributed equally to this work. Department of Food Science, Foshan University, Foshan 528231, China Guangdong Engineering Research Center for Traditional Fermented Food, Foshan 528231, China Full list of author information is available at the end of the article © The Author(s). 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. Huang et al. Annals of Microbiology (2021) 71:4 Page 2 of 13 (Continued from previous page) Conclusion: Ammonium assimilation promoted by ammonium compensation regulated the intracellular carbon metabolism of S. cerevisiae and affected the distribution of metabolic flux. The carbon flow that should have gone to the synthesis pathway of higher alcohols was reversed to the TCA cycle, thereby decreasing the content of higher alcohols. These findings may contribute to an improved understanding of the molecular mechanism for the decrease in higher alcohol content through ammonium compensation. Keywords: Higher alcohols, Proteomics, Ammonium compensation, Saccharomyces cerevisiae, iTRAQ Introduction chain amino acid, decarboxylation, and reduction of α- Chinese rice wine (CRW), a national unique and trad- ketone acids (Pires et al. 2014; Zhong et al. 2019). The itional wine that has a long history in China, is an alco- amount of higher alcohols produced in the pathway of holic beverage that is commonly considered to contain branched-chain amino acid synthesis pathway (Ehrlich molecules with a nutraceutical and pharmaceutical inter- pathway) accounted for 25% of the total amount of est, thus attracting a great deal of attention (Xie 2008; higher alcohols (Hazelwood et al. 2008). When nitrogen Gao et al. 2018; Wang et al. 2020). The final aroma of sources are insufficient, synthetic precursors of higher CRW constitutes of hundreds of flavor-active com- alcohols mainly come from ketonic acids, which are pounds produced during brewing, including higher alco- largely synthesized from carbohydrate metabolic path- hols, esters, volatile acids, and aldehydes (Pires et al. ways such as the Embden-Meyerhof-Parnas pathway 2014). The main higher alcohols that have a relatively (EMP) and tricarboxylic acid cycle (TCA), accounting high concentration are isoamylol (66.8-299.5 mg/L), iso- for 75% of the total amounts (Yan et al. 2017). Accord- butanol (73-238.8 mg/L), n-propyl alcohol (24.3-131.4 ing to the relationship between higher alcohols and ni- mg/L), and β-phenethyl alcohol (35.3-103.7 mg/L) dur- trogen nutrition, combined with the synthesis pathway ing different treatment fermentation processes of CRW under different conditions, nitrogen compensation or ni- (Yuan et al. 2018). CRW has a total higher alcohol con- trogen deficiency was considered as an effective means tent of 205-700 mg/L, is 1.2-fold higher than that of to control the content of higher alcohols (Gutierrez grape wine (260-299 mg/mL), and 4.5-fold higher than et al. 2013). Therefore, a simple and feasible way to re- that of beer (55-94 mg/mL) (Sun et al. 2020). duce the content of higher alcohols is through nitrogen Higher alcohols, which are important flavor compo- compensation, in which the nitrogen level in the fermen- nents, are related to intoxication and hangovers (Fang tation system is regulated. et al. 2018). Excessive intake of higher alcohol causes a Nitrogen supplementation to the fermentation system hangover accompanied by nausea and dizziness (Xie not only affects yeast growth and fermentation kinetics et al. 2018). At a concentration above 400 mg/L, higher but also adjusts and controls the non-volatile and vola- alcohols influence the flavor and taste of wine, destroy- tile compound composition, including higher alcohols, ing its quality (Stribny et al. 2016; Longo et al. 2020). At which show an inverse relationship to the nitrogen con- or below 300 mg/L, higher alcohols add a complex centration (Torrea et al. 2011). Various studies have in- aroma and a full-bodied taste to wine, thereby optimiz- vestigated the effects of nitrogen types on the ing the quality of CRW and harmonizing its organoleptic composition of higher alcohols, optimizing the addition properties (Zhong et al. 2019). Therefore, the content of level and process condition of nitrogen compensation higher alcohols affects the quality of CRW and repre- (Zhong et al. 2019). An appropriate amount of inorganic sents one of the important indexes to evaluate the qual- nitrogen sources can promote the growth of yeast, while ity of wine (Zheng et al. 2020). An essential step is to decreasing the content of higher alcohols (Huang et al. moderately lower the content of higher alcohols to 2018; Stéphanie et al. 2018). Ammonium hydrogen maintain the flavor and improve the final quality of phosphate, a food-grade additive, has the best effect on CRW. the reduction of the content of higher alcohols, which The main synthesis pathway of higher alcohols is the verifies that the nitrogen source compensation strategy Ehrlich pathway related to amino acid catabolism and has a certain influence on the formation of higher alco- the Harris pathway associated with glycometabolism (El- hols (Huang et al. 2018). However, till now, the mechan- Dalatony et al. 2019; Rollero et al. 2015; Yan et al. 2017). ism of nitrogen compensation in reducing the content of With ample nitrogen sources, the biosynthesis of higher higher alcohol in S. cerevisiae has not been studied in alcohols is the de novo synthesis of branched-chain CRW in depth. amino acid through the Ehrlich pathway by a sequence Proteins are the executors of cell-specific activities and of reactions that involve transamination of branched- functions, which are responsible for the diversity, Huang et al. Annals of Microbiology (2021) 71:4 Page 3 of 13 integrity, complexity, functionality, and phenotype of Strains and culture conditions life. Comparative proteomics is regarded as a very The original strain used in this study was fully genotype powerful tool to reveal biological regulatory mecha- Saccharomyces cerevisiae (S. cerevisiae) (available in the nisms and comprehend the complementary functions School of Food Science and Engineering, Foshan Univer- of thewhole proteomein aspecial stateand is an sity, China), which was provided by the Food Science important way to evaluate the effects and mechanisms and Biotechnology Research Center of JiangNan (Hua et al. 2016;Santosetal. 2016). The expression University. degree and activity of proteins or their interaction A single colony of S. cerevisiae was picked from a solid with other proteins command-specific functional pat- slant medium and inoculated into a 250-mL shaking tern of organellar, cellular, and organism levels, which flask that contained 100 mL YPD culture medium. The can reflect the regulation process of cells with differ- strains were cultivated at 30 °C at a speed of 120 rpm for ent conditions (Choudhary et al. 2019). Potentially 24 h. Subsequently, activated second-generation strain important proteins associated with carbohydrate and was cultured under the same condition with a 5% inocu- nitrogen metabolism and related regulatory proteins lum size for 10 h, and OD was determined after may be revealed through comparative proteomic stud- culture dilution. ies. Until now, many studies have investigated and identified a series of genes and proteins associated Experimental treatments with higher alcohol synthesis through gene knockout According to our previous research, 200 mg/L and overexpression (Dickinson 2003; Yoshimoto et al. (NH ) HPO is the optimal amount to be added for the 4 2 4 2001; Schoondermark-Stolk et al. 2005). However, few reduction of higher alcohols (Zhong et al. 2019). Two studies focused on the mechanism by which the con- groups were set up in this study to explore the mechan- tent of higher alcohols is reduced through nitrogen ism of ammonium compensation: the blank group (no compensation.Therefore,proteomicswas used to ammonium salt added) and the experimental group explore the altered pathways and regulatory effects (containing a final concentration of 200 mg/L during the ammonium compensation process to in- (NH ) HPO ), with three parallel samples in each group. 4 2 4 vestigate the molecular mechanism of higher alcohol (NH ) HPO was added during inoculation along with S. 4 2 4 content reduction. cerevisiae. The critical point of the logarithmic and sta- In this study, isobaric tags for relative and absolute tionary phase (cultured about 8 h) during fermentation quantitation (iTRAQ)-based quantitative proteomic was selected as the time node of intracellular proteome analysis was applied to compare the changes in the and qPCR study. DEP profiles (comparative proteomics) relative content of intracellular proteomics in S. cere- and mRNA expression levels (real-time quantitative visiae under different conditions (with or without the polymerase chain reaction, RT-qPCR) were compared in addition of ammonium salt). The intracellular pro- the control group (without nitrogen feeding) and the ex- teins that change significantly during the process of perimental group (with nitrogen feeding). ammonium compensation were identified, and then differentially expressed proteins (DEPs) associated Protein extraction, digestion, and iTRAQ labeling with the synthesis metabolism of higher alcohols were Cells that were cultured for 8 h to reach the steady state analyzed to gain insights into the regulatory mechan- were collected by centrifugation at 25,000×g for 10 min ism of nitrogen compensation in reducing the content at 4 °C. After supernatant removal, the cell pellets were of higher alcohols. washed twice with cold deionized water and suspended in a lysis buffer of 10 mM DTT, 2 mM EDTA, and 1 mM PMSF. Cells were disrupted by vortexing and tissue Materials and methods lyser with glass beads (50 Hz, 120 s) to extract the total Materials intracellular proteins, and then the samples were centri- Glucose, tryptone, agar, and TEAB were purchased fuged at 4 °C, 25,000×g for 20 min. The supernatant was from Sangon Biotech Co. Ltd. (Shanghai, China). Gly- then obtained. The solution was mixed with 10 mM cerin, (NH ) HPO , urea, acetone, and isopropanol DTT at 56 °C for 1 h, and then IAA was added to reach 4 2 4 were purchased from Sinopharm Chemical Reagent a final concentration of 55 mM and incubated in the Co., Ltd. (China). DTT, PMSF, EDTA, and aceto- dark for 45 min. The protein solution was placed in a re- nitrile (HPLC grade) were purchased from Shanghai frigerator at −20 °C for 30 min after being washed 5 Macklin Biochemical Co., Ltd. (Shanghai, China). times the volume of precooled acetone with vigorous Trypsin was obtained from Takara Co., Ltd. (Kyoto, stirring, and then the supernatant was eliminated by Japan). iTRAQ Reagent 8 Plex One Assay Kit was centrifugation (25,000×g, 4 °C, 15 min). Precipitation was purchased from SCIEX (Scientific Export). redissolved with lysis buffer, and insoluble material was Huang et al. Annals of Microbiology (2021) 71:4 Page 4 of 13 removed by centrifugation at 4 °C, 25,000×g for 15 min; iTRAQ8plex (N-term), and iTRAQ8plex (K) were de- the supernatant was the protein solution. Protein con- fined as fixed modification, while protein methionine centrations were determined by a modified Bradford oxidation and iTRAQ8plex (Y) were defined as variable protein assay kit (Sangon Biotech Co. Ltd) with BSA as a modifications. iTRAQ-based quantification was con- standard, as described by Bradford (Bradford 1976). ducted in MaxQuant. Protein quantitation required a For each sample, 100 μg protein samples were sub- protein to contain at least two unique peptides. All the jected with 5 μg trypsin digestion at 37 °C overnight, and proteins with an FDR less than 1% were then subjected the enzymatic peptide was desalinated by Strata X col- to downstream analysis, including GO, KEGG pathway, umn and then vacuumed. Peptides were tagged with GO enrichment analysis, KEGG pathway enrichment iTRAQ tags according to the manufacturer’s recommen- analysis, cluster analysis, and subcellular localization dations. Each sample was dissolved with an addition of analysis. The sequence data of the target proteins were 0.5 M TEAB and then transferred to the correct iTRAQ bundled in batches, which were retrieved from the Uni- reagent conducted with 50 μL acetone to be incubated at ProtKB database. The retrieved sequences were searched room temperature for 2 h. against the SwissProt database by using NCBI BLAST. The top 10 blast results were retrieved for each query Mass spectrometry and proteomic data analysis sequence and finally loaded into Blast2GO to analyze C18 column (5 μm 4.6 × 250 mm Gemini C18) was used the related KEGG pathways. to desalt the peptides in each sample, and subsequently concentrated by vacuum centrifugation and reconsti- Real-time quantitative PCR analysis tuted in buffer A (2% acetonitrile, 0.1% formic acid). To Total RNA was extracted from the same batch of sam- conduct the MS experiments on a Q-Exactive HF X ples as the proteomic analysis and qualified by using an (Thermo Scientific, USA) by applying a DDA model, the RNA Extraction Kit (Catalog No. 15596018, Invitrogen samples were loaded onto a Trap column to desalt, TRizol, USA). The concentration and purity of RNA flowed into the C18 column (75 μm 4.6 × 250 mm) samples were determined through spectrophotometric (Thermo Scientific, USA) in buffer A (2% acetonitrile, analysis by considering the absorbance ratio at 260/280 0.1% formic acid), and then separated with a linear gra- nm. After genomic DNA was removed, RNA (1 μg) was dient of buffer B (98% acetonitrile, 0.1% formic acid). reverse transcribed to cDNA in a 10 μL reaction mixture The full MS scan (350-1500 m/z) was performed in the by using a Bestar® SybrGreen qPCR Master Mix Kit positive ion mode at a resolution of 60,000 (at 100 m/z). (DBI-2043 Bioscience, Germany). The cDNA was quan- The top 20 peak which intensity exceeds 10,000 and tified by RT-qPCR reactions on an ABI7300 (Thermo most intense 2-5 charged ions were selected, and the dy- Fisher, USA). Reactions were performed in a 20 μL sys- namic exclusion time was set at 30 s. tem that contained 10 μL Bestar® SybrGreen mix, 0.4 μL forward primer (10 μM), 0.4 μL reverse primer (10 μM), Data quality control, protein function annotation, and 1 μL cDNA, and 8.2 μL ddH O. The RT-qPCR proced- quantitative analysis ure was set as follows: initial denaturation at 95 °C for Quality control of the data was conducted to determine 10 min and followed by 40 cycles of 95 °C for 15 s, 60 °C whether the data were qualified during the search and for 34 s, and 72 °C for 30 s. At the end of the amplifica- identification. The final reliable protein identification re- tion cycle, the melting curves of the PCR amplicons sults could be obtained after all the proteins were were obtained at 60-98 °C. Relative quantification of the screened, with a false discovery rate (FDR) less than 1%. expression level of each gene was normalized according The Gene Ontology (GO), Kyoto Encyclopedia of Genes to the β-actin gene (ACT1) expression. The primers used and Genomes (KEGG), Cluster of Orthologous Groups in the experiment can be found in the “Additional files” of proteins (COG), and other databases were used to an- section (Table S1). RT-qPCR was used to confirm the notate the identified proteins. authenticity and accuracy of the proteomic analysis results. Sequence database search and data analysis The MaxQuant software suite (version 2.3.02) was used Result and discussion to analyze the MS data, which were searched against the In our previous study, we found that the content of UniProtKB database (Database_info: uniprot_Saccharo- higher alcohols could be reduced by adding ammonium myces_cerevisiae 190220; Database 27355 sequences). salt to the medium during the fermentation process of For peptide identification, a mass tolerance of 20 ppm CRW. (NH ) HPO is a suitable inorganic nitrogen 4 2 4 was permitted. This search was conducted with the en- source, which leads to a 23.99% decrease in the total zymatic cleavage rule of trypsin with a maximum of two higher alcohols from 205.58 to 156.26 mg/L; specifically, missed cleavage sites allowed. Carbamidomethyl (C), isobutyl alcohol, isoamyl alcohol, and β-phenethyl Huang et al. Annals of Microbiology (2021) 71:4 Page 5 of 13 alcohol decreased by 25.39%, 21.00%, and 12.89%, re- To further understand the functions, all identified pro- spectively (Huang et al. 2018; Zhong et al. 2019). These teins and the DEPs between experimental groups and findings proved the feasibility of the nitrogen compensa- control groups were annotated according to different tion strategy and confirmed the variety of compensated categories, including GO terms and KEGG pathways. nitrogen sources. However, the regulatory mechanism of Detailed information of all identified proteins was listed inorganic nitrogen in decreasing the content of higher in Additional file Table S2. GO is a major bioinformatics alcohols was not investigated. initiative to unify the representation of gene and gene product attributes across all species, and it can flexibly annotate homologous gene and protein sequences in Quality control of the proteome data multiple organisms based on biological process, molecu- In this iTRAQ quantification project, a total of 946,482 lar function, and cellular component (Ashburner et al. spectrums were generated, and 27,973 peptides and 2000). In this study, GO annotated 3068 identified pro- 4062 proteins were identified with 1% FDR. Mass spec- teins, and the GO analysis result of the identified and trometry (MS)data must be subjected to quality control DEPs was shown in Fig. 1a. Under the biological process checks, including peptide length distribution, protein category, 2234 identified proteins and 102 DEPs were in- mass distribution, unique peptide number, protein volved in the metabolic process, and 12 identified coverage, and CV distribution in replicate. A detailed proteins and 4 DEPs were involved in the nitrogen quality control report is shown in Figure S1. iTRAQ utilization. In the molecular function category, 1778 technology can identify proteins with different molecular identified proteins and 79 DEPs had binding activities, weights, and our result showed that about 12.5% of the and 1428 identified proteins and 48 DEPs had catalytic identified proteins have a molecular weight of more than activities. In the cellular component, 2879 identified pro- 100 kDa, and less than 1% have a molecular weight of 0- teins and 142 DEPs were cell-related proteins; 967 iden- 10 kDa. In this study, the range of peptide length was 5- tified proteins and 50 DEPs were membrane related 25 amino acids, according to the detection limit of MS. proteins. As can be seen from the bar chart, the GO The number of protein unique peptides identified in the term enrichment analysis indicated the highest number yeast was correlated to the reliability of the correspond- of DEPs with binding activity, followed by catalytic activ- ing proteins. More unique peptides correspond to high ity in the molecular function category; the amount of protein reliability. Of the 4062 identified proteins, 1183 DEPs involved in cell, cell part, and organelle part is the proteins contained one unique peptide segment, and largest in the classification of cell components and those 2879 proteins (70% identified proteins) contained at least involved in cellular and metabolic processes were the two unique peptide segments (Fig. S1C). Therefore, the most numerous in S. cerevisiae’s biological processes. number of reliable proteins was relatively large, which The KEGG pathway is a collection of manually drawn indicated a high reliability of results. At the same time, pathway maps representing our knowledge on the mo- 1633 proteins (40.2% of 4062 all identified proteins) had lecular interaction and reaction networks for many me- a protein coverage distribution of 0-10%, and 943 pro- tabolite and signal transduction pathways (Yan et al. teins (23.2%) had a protein coverage distribution of 10- 2019). In this study, 1937 proteins were annotated to the 20%. Other proteins had a protein coverage distribution KEGG pathway database, and the KEGG analysis result at 20-100%, accounting for 33.6%, which meant the pro- of the identified proteins and DEPs were shown in Fig. tein coverage distribution was relatively wide (Fig. S1D). 1b. A total of 1935 identified proteins were involved in The coefficient of variation (CV) is the ratio of standard the metabolism pathway, accounting for 58%, of which deviation to the mean, which was used to evaluate the 785 identified proteins participated in global and over- reproducibility. A lower CV corresponded to good re- view maps and 862 of identified proteins (25.8% pro- producibility. More than 80% of the protein CV was less teins) were involved in genetic information processing. than 10%, thereby verifying the reliability of the iTRAQ Moreover, 372, 119, and 40 proteins were involved in data (Fig. S1E). cellular processes, environmental information process- ing, and organismal systems, accounting for 11.1%, 3.6%, Integrative analysis of the global proteomes and 1.2%, respectively. A total of 148 DEPs were anno- On the basis of the large-scale iTRAQ-label comparative tated in pathway entries. The amounts of DEPs involved proteomic method and statistical filtration (fold change in metabolism accounted for more than half (58.1%), in > 1.2, p value < 0.05), about 211 proteins were identified which global and overview maps accounted for 22.9%, as DEPs in the ammonium compensation group as followed by carbohydrate metabolism (10.1%), and compared with the control group among the quantified amino acid metabolism (6%). The genetic information proteins (123 upregulated and 88 downregulated processing pathway accounted for 20.2%, translation proteins) (Additional file Table S1). accounted for 9%, and the DEPs involved in cellular Huang et al. Annals of Microbiology (2021) 71:4 Page 6 of 13 Fig. 1 Classification of all identified proteins in ammonium compensation group and differentially expressed proteins (DEPs) compared to the control group. a GO analysis of all identified proteins and DEPs. All identified proteins and DEPs were classified by GO terms based on their cellular component, molecular function, and biological process. b KEGG analysis of all identified proteins and DEPs. All identified proteins and DEPs were classified by eight KEGG terms based on organismal systems, metabolism, human diseases, genetic information, processing, environmental, information processing, and cellular processes processes accounted for 15%. DEPs in S. cerevisiae be- differential expression. The red and green dots indi- tween the experimental group and control group are cate points of interest that indicate large-magnitude mainly involved in carbohydrate metabolism and amino fold changes and high statistical significance, respect- acid metabolism. ively. The red dots mean significantly upregulated proteins that passed the screening threshold, and the Protein differential analysis green dots mean significantly downregulated proteins. Figure 2a depicts the volcano plots of identified pro- Gray dots are nonsignificant DEPs. The DEPs were teins and differentially expressed proteins. The x-axis grouped based on their subcellular localizations (Fig. showed the logarithm of fold change (base 2), and the 2b). About 11 subcellular components were identified, negative logarithm of the p value was taken as the y- including 115 nuclear-localized DEPs accounted for axis (base 10), representing the probability of the 54.5%, 29 cytosol-localized DEPs (13.7%), and 20 DEPs. A p value less than 0.05 and a fold change mitochondria-localized DEPs (9.48%). Some DEPs in more than 1.2 are set as the significant threshold for other organelles occurred at multiple locations. Huang et al. Annals of Microbiology (2021) 71:4 Page 7 of 13 Fig. 2 Distribution of differentially expressed proteins (DEPs). a Volcano plot of the DEPs between experiment and control group. Colors of the scatter points indicate the proteins. Red dots indicate significantly upregulated proteins, green dots indicate significantly downregulated proteins, and gray dots indicate proteins with no significant difference. b Subcellular location of DEPs. Different colors represent different positions, and the value represents the number of proteins located at that position and the percentage of total differentially expressed proteins KEGG pathway enrichment analysis of DEPs the x-axis represented the rich factor, which was the ra- KEGG pathway enrichment analysis (El-Dalatony et al. tio of numbers of DEPs annotated to the corresponding 2019; Yan et al. 2017) was conducted based on the pathway to all identified proteins annotated to this path- KEGG database for DEP screening, and p value less than way. The enrichment level in each pathway increased 0.05 was set as the standard for significant enrichment with the ratio. The dot size in the figure represented the of DEPs. The specific number of up- and downregulated number of DEPs annotated to this pathway. The p value DEPs from each pathway annotation item was listed in is related to the intensity of the difference in DEPs. A Fig. 3a. Pyruvate was produced by the glycolysis pathway small p value corresponded to the great difference. The and then partially entered the TCA cycle. Thus, the dif- p value for carbon metabolism, glyoxylate and dicarbox- ferential proteins involved in these three pathways had ylate metabolism, and pyruvate metabolism were all less consistent up and down trends. In addition, multiple than 0.01. Therefore, these pathways were significantly differential proteins were found in the amino acid me- enriched for the DEPs. The DEPs involved in the pyru- tabolism pathway, such as glycine, serine, and threonine vate metabolism pathway accounted for a larger propor- metabolism; alanine, aspartate, and glutamate; and mei- tion than others involved in metabolic pathways, and the osis in yeast. carbon metabolism pathway annotated the most DEPs. The metabolic pathway maps for significant enrich- Taken together, these results suggested that the DEPs ment of DEPs were presented in Fig. 3b. In this figure, between the ammonium salt compensation group and Huang et al. Annals of Microbiology (2021) 71:4 Page 8 of 13 Fig. 3 KEGG enrichment analysis of differentially expressed proteins (DEPs). a Statistical figure of up- and downregulation on pathway classification of DEPs. The specific numbers of up- and downregulated proteins were classified into specific metabolic pathways of KEGG, with red representing upregulated proteins and blue representing downregulated proteins. b Pathway statistical map of significant enrichment. The horizontal axis represents the ratios of the number of DEPs in the corresponding pathways to the number of proteins detected. Dot color represents the p value verified by a hypergeometric test. Dot size represents the numbers of DEPs. The DEP enrichment level in each pathway increased with ratio magnitude control group were mainly related to metabolic path- Differential proteins involved in carbohydrate metabolism ways, carbon metabolism, and pathways associated with In this study, differentially expressed proteins were in- cell growth. volved in several pathways, including carbohydrate metabolism (energy), amino acids biosynthesis and me- tabolism, regulation metabolism, and pathways involved Differential proteins associated with cell growth in cell meiosis, cell cycle, growth, and MAPK signaling. Nitrogen source is a nutritious material that provides ni- The most affected metabolism pathways that were in- trogen for the construction of proteins, nucleic acids, volved in the synthesis of higher alcohols were carbon and enzymes in cell metabolism. Compared with control, metabolism and pyruvate metabolism and key amino the yeast-specific growth rate increased with the exogen- acid metabolism in S. cerevisiae. ous addition of 200 mg/L (NH ) HPO during the log In carbon metabolism, two proteins related to the gly- 4 2 4 phase (Zhong et al. 2019). The main proteins associated colysis were significantly upregulated in the experimen- with cell growth, which were observed in GO and KEGG tal group. One protein was fructose-bisphosphate enrichment (Fig. 1) were involved in cell growth, mei- aldolase (EC:18.104.22.168-P14540) which catalyzed the cleav- osis, cellular component organization, or biogenesis (or- age reaction from β-D-fructose-1,6-bisphosphate (FBP) ganelle or part, membrane component), growth to dihydroxyacetone phosphate (DHAP) and regulation, and so on. Specifical differential proteins in- glyceraldehyde-3-phosphate (GAP) (Geiger et al. 2019). cluded zinc finger protein (Swiss-Prot Accession Num- This aldolase converted carbon six (C6) into two mol- ber: P32338), mannan endo-1,6-alpha-mannosidase ecule three-carbon (C3) compounds, which marked the (P36091), transcriptional regulatory protein (P34233), step-by-step entry to the second exoenergic stage of gly- aspartic proteinase precursor (P32329), putative metab- colysis. Another upregulated protein was phosphoglycer- olite transport protein (Q12407), and glucose transporter ate mutase (EC:22.214.171.124-Q12008) which converted 3- (A6ZT02). The proteomic results supported those of phosphoglycerate (3-PG) into 2-phosphoglycerate (2- previous studies, which showed that (NH ) HPO af- PG). These two enzymes, which were involved in the 4 2 4 fected growth. glycolysis pathway, were all upregulated, which signified that the glycolytic pathway was activated when an Huang et al. Annals of Microbiology (2021) 71:4 Page 9 of 13 appropriate amount of ammonium salt was added to the higher alcohols was the reduction of fusel aldehydes system for nitrogen compensation. The identified differ- into the corresponding higher alcohols. According to entially protein, phosphoenolpyruvate carboxykinase the literature, this step was almost catalyzed by ethanol (PEPCK) (EC:126.96.36.199-P10963), was an important en- dehydrogenase encoded by Adh1, Adh2, Adh3, Adh4, zyme related to a roundabout reaction in the gluconeo- and Adh5 or formaldehyde dehydrogenase encrypted by genesis pathway. The expression of PEPCK, which Sfa1 (Pires et al. 2014). This downregulation of alcohol converted oxaloacetate into phosphoenolpyruvate (PEP), dehydrogenase was encoded by Adh2, consistent with was significantly downregulated, which might result in the reported results. the inhibition of the gluconeogenesis pathway. Interestingly, two different malate dehydrogenases Differential proteins involved in nitrogen metabolism were found as DEPs involved in the TCA cycle and One identified DEP, carbamoyl-phosphate synthase (EC: pyruvate metabolism; one protein was upregulated (EC: 188.8.131.52-P07258, CPS), was an important enzyme related 184.108.40.206-P36013), whereas another was downregulated to the nitrogen assimilation pathway (Khoja et al. 2019). (EC:220.127.116.11-P22133)., The subcellular localization in CPS, which was identified in this study, referred to the the cell of the two proteins was examined to distinguish carbamoyl-phosphate synthase II that existed in cyto- the roles of these two malate dehydrogenase proteins. plasm, which catalyzed the biosynthesis of carbamoyl- The results showed that the expression level of the pro- phosphate with L-glutamine as a nitrogen source. tein in the mitochondria (EC:18.104.22.168-P36013) in- Carbamoyl-phosphate, which is a high-energy com- creased, whereas the level in the cytoplasm (EC: pound, can enter either pyrimidine metabolism or the 22.214.171.124-P22133) was decreased. The TCA cycle is pathway of arginine biosynthesis. The expression of CPS known to occur in mitochondria. Thus, the upregula- in cell cultured with nitrogen compensation was in- tion of malate dehydrogenase in the mitochondria indi- creased by 1.2-fold compared with that in cells of the cated that the TCA cycle was proceeding, that is, a control group. The expression level of CPS positively certain amount of oxaloacetic acid was accumulated regulated and promoted the absorption and assimilation and then, together with acetyl-CoA, the citric acid cycle of ammonium ions by consuming L-glutamine, which is completed. was produced by the glutamate-mediated assimilation One protein (EC:126.96.36.199-P37262) 6- pathway and enhanced its metabolic flow in the biosyn- phosphogluconolactonase, as an important enzyme that thesis of pyrimidine and purine. converts 6-phosphoglucono-δ-lactone into 6- Interestingly, the expression level of L-serine/L-threo- phosphogluconate in the oxidation stage of the pentose nine ammonia lyase (EC:188.8.131.52, 184.108.40.206-Q12008), phosphate (PP) pathway (Prabhu and Veeranki 2018), which catalyzed the conversion of serine/threonine into was significantly downregulated. The PP pathway, which corresponding pyruvate or 2-oxobutanoate, significantly was mainly responsible for the reduction of NADPH, increased by 1.42-fold compared with that in the control ribose-5-phosphate, tetrose, and pentaglucose, was group, which seemed to contradict the experimental weakened with the decrease in the lactonase enzyme ex- phenomena. However, the α-keto acids produced by L- pression level. The expression level of two vital en- serine or L-threonine deamination had multiple meta- zymes—namely, isocitrate lyase (EC:220.127.116.11-P28240) and bolic pathways for utilization and metabolism, which malate synthase (EC:18.104.22.168-P30952 and EC:22.214.171.124- can enter TCA catabolism; synthesize other amino acids A6ZRW6)—that involved in glyoxylate and dicarboxylate such as L-leucine, L-Isoleucine, and L-valine; and even metabolism were all downregulated. The glyoxylic acid synthesize the corresponding higher alcohols through cycle is an anaplerotic pathway that used acetyl-CoA as oxidative decarboxylation. The higher alcohols synthe- raw material and provided carbon four (C4) acids, and sized by amino acid decarboxylation accounted for only oxaloacetate in the citric acid cycle. The downregulation 25% of the total higher alcohols (Huang et al. 2018; Pires of all three proteins was probably the consequence of et al. 2014), and the catalytic product of this enzyme, α- the reduced flux of glyoxylate cycle, that is, acetyl-CoA keto acids, has several metabolic pathways. Therefore, a was directly involved in the TCA cycle without supple- certain increase in the expression level of L-serine/L- menting the oxaloacetate. This phenomenon was threonine ammonia lyase had little impact on the contri- consistent with the different regulation of malate de- bution of higher alcohols. hydrogenase in different parts. The expression of 5-aminolevulinate synthase (EC: Another downregulated DEP related to the biosyn- 126.96.36.199-P09950), which was an upregulated protein in- thesis of higher alcohols was alcohol dehydrogenase volved in glycine, serine, and threonine, was higher in (EC:188.8.131.52-P00331) in cytoplasmic form, which was the experimental group than in the control group. It cat- directly repressed in the presence of propanol- alyzed the formation of 5-aminolevulinate and CoA from preferring glucose. The final step in the biosynthesis of succinyl-CoA and glycine (Bailey et al. 2020), which Huang et al. Annals of Microbiology (2021) 71:4 Page 10 of 13 Fig. 4 RT-qPCR analysis of seven selected genes regulated by ammonium compensation in S. cerevisiae. FBA1, fructose-bisphosphate aldolase encoding gene; GPM2, bisphosphoglycerate-dependent phosphoglycerate mutase encoding gene; MAE1, malate dehydrogenase encoding gene in the mitochondria; SOL2, 6-phosphogluconolactonase encoding gene; CPA1, carbamoyl-phosphate synthase encoding gene; HEM1,5- aminolevulinate synthase encoding gene; ADE12, adenylosuccinate synthetase encoding gene consumed the succinyl-CoA from the TCA cycle, pro- differential proteins related to nitrogen metabolism ac- moted TCA progression, and facilitated the release of cording to the proteomic results, such as the carbamoyl- CoA at the same time. Succinyl-CoA is the allosteric in- phosphate synthase encoding gene (CPA1), the 5- hibitor of the key regulatory enzyme citrate synthase at aminolevulinate synthase encoding gene (HEM1), and the beginning of the TCA cycle, and if its content was the adenylosuccinate synthetase encoding gene (ADE12). too high, then it would inhibit the activity of citrate syn- The expression level of each gene was normalized to the thase and TCA cycle. Therefore, the upregulated content expression of the act gene. The expression levels of these of 5-aminolevulinate synthase was conducive to the selected genes at the mRNA level were basically consist- smooth progress of the TCA cycle. ent with the proteomic analyses (Fig. 4). Notably, the Another protein associated with nitrogen metabolism fold change of the expression of some proteins at the was adenylosuccinate synthetase (EC:184.108.40.206-C8ZG13, mRNA level was different from that at the protein level. AdSS), which is involved in de novo biosynthesis and Take SOL2 as an example; the mRNA expression level transdeamination of purine nucleotide (Galina et al. of SOL2 decreased to 0.25-fold, but it decreased to 0.81- 2017), and was downregulated. It catalyzed the forma- fold of the protein expression level under the condition tion of adenylosuccinate from hypoxanthine nucleotide of ammonium compensation. This phenomenon was and aspartic acid. Then, the intermediate product adeny- normal because no strictly linear relationship existed be- losuccinate splits into adenine nucleotides and fumarate tween genes and proteins, that is, mRNA levels do not under the action of the cleavage enzyme, realizing the correspond to protein levels (Abbott 2001; Fields 2001). transformation of IMP and AMP. Possible regulatory model for higher alcohol reduction by Validation of gene expression by RT-qPCR ammonium compensation To verify the differential expression levels of some key Over the past few decades, proteome analysis has shown metabolic genes at mRNA level, RT-qPCR assay was outstanding potential in exploring the diversity, com- performed. Seven important proteins involved in carbon plexity, functionality, and phenotype of living organisms, metabolism and nitrogen metabolism encoding genes because proteins are often the implementers of specific were chosen for qPCR analysis with β-actin as the in- physiological functions and the mechanisms of change ternal reference. The important proteins involved in can be interpreted at the protein level. Therefore, to fully EMP pathway included fructose-bisphosphate aldolase understand the effect and mechanism of ammonium encoding gene (FBA1) and bisphosphoglycerate- compensation on the reduction of higher alcohol con- dependent phosphoglycerate mutase encoding gene tent and its association with many other cellular events, (GPM2). The protein involved in the TCA cycle was the a possible approach was to analyze the proteomic ana- malate dehydrogenase encoding gene in the mitochon- lysis of the ammonium compensation group with the dria (MAE1). The protein involved in the PP pathway control group. was selected as the 6-phosphogluconolactonase encoding The proteomic comparison analysis results showed gene (SOL2). The other three proteins were selected as that most proteins with a high degree of expression Huang et al. Annals of Microbiology (2021) 71:4 Page 11 of 13 Fig. 5 A proposed regulatory model illustrating reduced higher alcohol content by ammonium compensation in S. cerevisiae (only two key DEPs were marked) differences were involved in metabolism, such as carbo- (Fig. 5). Higher alcohols were synthesized via the Harris hydrate, amino acid metabolisms, genetic information and Ehrlich pathways, and ammonium compensation process, and cell process, thereby indicating that the sig- probably affected higher alcohol biosynthesis by fluctuat- nificant changes in these metabolic pathways and cellu- ing the flux of TCA and nitrogen metabolism. When am- lar events were obviously regulated by ammonium monium was added to the fermentation system, NH compensation. Central carbon metabolism (CCM), in- rapidly transformed into a form of organic nitrogen, cluding glycolysis, TCA cycle, and PP pathway, was re- namely, glutamate and glutamine, through the GDH sys- sponsible for the production of accessible energy, the tem and the GS/GOGAT ammonium assimilation system, composition of important small molecular compounds, respectively (Roca et al. 2003; Sieg and Trotter 2014). The and the creation of primary building blocks of other me- concentration of α-ketoglutarate was consumed as a re- tabolisms for biological life processes (Zhang et al. sult, thus enhancing the TCA pathway metabolic capacity 2010). However, the regulation of different pathways in- and promoting its continuous progress. Therefore, the volved in carbon metabolism was different in S. cerevi- metabolism flow and moving directions of pyruvate gener- siae with ammonium compensation. The EMP pathway, ated by the EMP pathway were affected. More pyruvate a main glycolysis pathway of yeast, is a common meta- entered the TCA cycle to provide energy and intermediate bolic pathway for aerobic or anaerobic respiration. The products for the activities of living organisms, thereby fur- experimental results showed that the key enzyme ex- ther limiting the conversion of pyruvate into α- pression level of the EMP pathway (fructose-bispho- ketoisocaproic acid and reducing the conversion flow to sphate aldolase and phosphoglycerate mutase) was the Harris pathway, and finally leading to a decrease in the upregulated, indicating that the EMP pathway was acti- content of higher alcohols. At the same time, the addition vated and the flow rate of this pathway and TCA were of ammonium salt consumed α-ketoglutarate through the enhanced. Other metabolic pathways, such as glyoxylic assimilation pathway of ammonium, which competitively acid cycle, gluconeogenesis, and PP pathway, were inhib- inhibited the ammonia transamination of amino acids and ited to some extent. reduced the content of higher alcohols produced by the Based on the synthesis mechanism of higher alcohols, a Ehrlich pathway. diagram of possible regulatory mechanisms in response to Our hypothesis was consistent with the results of the reduction of higher alcohol content with ammonium proteomic research, which showed that nitrogen assimi- compensation in S. cerevisiae was proposed and verified lation and the concentration of metabolites in the TCA Huang et al. Annals of Microbiology (2021) 71:4 Page 12 of 13 cycle were involved in the regulation of higher alcohol Author details Department of Food Science, Foshan University, Foshan 528231, China. biosynthesis. These regulation changes at different ex- Guangdong Engineering Research Center for Traditional Fermented Food, tents ultimately reduced the content of higher alcohols. 3 Foshan 528231, China. Guangdong Engineering Research Center for Safety Therefore, the synthesis of higher alcohols was coordi- Control of Food Circulation, Foshan 528231, China. Foshan Engineering Research Center for Brewing Technology, Foshan 528231, China. Foshan nated by multiple pathways. Engineering Research Center for Agricultural Bio-manufacturing, Foshan 528231, China. Conclusion Received: 19 August 2020 Accepted: 20 November 2020 In conclusion, proper ammonium compensation is a promising strategy to reduce the content of higher alco- hols in S. cerevisiae via adjusting and altering the meta- References bolic flux of pyruvate without affecting the alcohol Abbott A (2001) And now for the proteome. Nature 409(6822):747 Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski production rate. Ammonium compensation promotes K, Dwight SS, Eppig JT, Harris MA, Hill DP, Issel-Tarver L, Kasarskis A, Lewis S, yeast growth, facilitates glucose consumption, fluctuates Matese JC, Richardson JE, Martin R, Rubin GM, Sherlock G (2000) Gene TCA cycles and amino acid metabolism, channels pyru- Ontology: tool for the unification of biology. Nat Genet 25:25–29 Bailey HJ, Bezerra GA, Marcero JR, Padhi S, Foster WR, Rembeza E, Roy A, Bishop vate flux into the TCA cycle, and decreases the synthesis DF, Desnick RJ, Bulusu G, Dailey HA Jr, Yue WW (2020) Human of α-ketoisocaproic acid and the metabolic flux of the aminolevulinate synthase structure reveals a eukaryotic-specific Harris pathway. These findings may shed light on the autoinhibitory loop regulating substrate binding and product release. Nat Commun 11:2813–2825 molecular mechanism of higher alcohol synthesis in S. Bradford MM (1976) A rapid and sensitive method for the quantitation of cerevisiae. The up- or downregulated proteins could microgram quantities of protein utilizing the principle of protein-dye serve as targeted sites for genetic engineering or genome binding. Anal Biochem 72:248–254 Choudhary J, Singh S, Tiwari R, Goel R, Nain L (2019) An iTRAQ based comparative editing to decrease the content of higher alcohols. proteomic profiling of thermotolerant Saccharomyces cerevisiae JRC6 in response to high temperature fermentation. Curr Proteomic 16:289–296 Supplementary Information Dickinson JR (2003) The catabolism of amino acids to long chain and complex The online version contains supplementary material available at https://doi. alcohols in Saccharomyces cerevisiae. J Biol Chem 278:8028–8034 org/10.1186/s13213-020-01611-7. El-Dalatony MM, Saha S, Govindwar SP, Abou-Shanab RAI, Jeon B-H (2019) Biological conversion of amino acids to higher alcohols. Trends Biotechnol 37:855–869 Additional file 1: Table S1. The detail information of DEPs. Fang C, Du H, Jia W, Xu Y (2018) Compositional differences and similarities Additional file 2: Table S2. The detail information of all identified between typical Chinese Baijiu and western liquor as revealed by mass proteins. spectrometry-based metabolomics. Metabolites 9(1):2 Additional file 3: Figure S1. Distribution of protein mass (A), peptide Fields S (2001) Proteomics in genomeland. Science 291(5507):1221–1224 length (B), unique peptide number (C), protein coverage (D), and CV Galina L, Dalberto PF, Martinelli LKB, Roth CD, Pinto AFM, Villela AD, Bizarro CV, distribution in replicate (E). Table S1. The primers used in the Machado P, Timmers L, de Souza ON, Carvalho EMD, Basso LA, Santos DS experiment. (2017) Biochemical, thermodynamic and structural studies of recombinant homotetrameric adenylosuccinate lyase from Leishmania braziliensis. RSC Adv 7:54347–54360 Acknowledgements Gao X, Li B, Mei J (2018) Research progress in active components and functional Not applicable. properties of yellow rice wine. Brewing Technol 1:91–96 Geiger J, Doelker R, Salo S, Roitsch T, Dalgaard LT (2019) Physiological Authors’ contributions phenotyping of mammalian cell lines by enzymatic activity fingerprinting of Guidong Huang designed the experiments and revised the manuscript. key carbohydrate metabolic enzymes: a pilot and feasibility study. BMC Res Hong Ren analyzed the data and prepared the manuscript. Ali Wang Notes 12:682–688 contributed to the manuscript revision. Xianfeng Zhong reviewed and Gutierrez A, Chiva R, Beltran G, Mas A, Guillamon JM (2013) Biomarkers for approved the final manuscript, and other authors prepared the samples and detecting nitrogen deficiency during alcoholic fermentation in different performed the experiments. The author(s) read and approved the final commercial wine yeast strains. Food Microbiol 34:227–237 manuscript. Hazelwood LA, Daran J-M, Maris AJ, Pronk JT, Dickinson JR (2008) The Ehrlich pathway for fusel alcohol production: a century of research on Funding Saccharomyces cerevisiae metabolism. Appl Environ Microbiol 74:2259–2266 This work was supported by the National Natural Science Foundation of Hua Y, Wang S, Liu Z, Liu X, Zou L, Gu W, Hou Y, Ma Y, Luo Y, Liu J (2016) iTRAQ- China (No. 31660459), the Natural Science Foundation of Guangdong based quantitative proteomic analysis of cultivated Pseudostellaria Province (No. 2018A0303130275; No. 2020A1515011444). heterophylla and its wild-type. J Proteomics 139:13–25 Huang G, Shanshan XU, Liu H, Mao J, Zukang FU, Sun G, Zhou Z (2018) Selected Availability of data and materials inorganic nitrogen in reducing the contents of higher alcohols from Chinese All data included in this study are available upon request by contact with rice wine. J Food Sci Biotechnol 37:82–87 the corresponding author. Khoja S, Nitzahn M, Truong B, Lambert J, Willis B, Allegri G, Rufenacht V, Haberle J, Lipshutz GS (2019) A constitutive knockout of murine carbamoyl Ethics approval and consent to participate phosphate synthetase 1 results in death with marked hyperglutaminemia This article does not contain any studies with human participants or animals. and hyperammonemia. J Inherit Metab Dis 42:1044–1053 Longo R, Carew A, Sawyer S et al (2020) A review on the aroma composition of Consent for publication Vitis vinifera L. Pinot noir wines: origins and influencing factors. Crit Rev Food Not applicable. Sci Nutr. https://doi.org/10.1080/10408398.2020.1762535 Pires EJ, Teixeira JA, Brányik T, Vicente AA (2014) Yeast: the soul of beer’s Competing interests aroma—a review of flavour-active esters and higher alcohols produced by The authors declare that they have no conflict of interest. the brewing yeast. Appl Microbiol Biotechnol 98:1937–1949 Huang et al. Annals of Microbiology (2021) 71:4 Page 13 of 13 Prabhu AA, Veeranki VD (2018) Metabolic engineering of Pichia pastoris GS115 for enhanced pentose phosphate pathway (PPP) flux toward recombinant human interferon gamma (hIFN-gamma) production. Mol Biol Rep 45:961–972 Roca C, Nielsen J, Olsson L (2003) Metabolic engineering of ammonium assimilation in xylose-fermenting Saccharomyes cerevisiae improves ethanol production. Appl Environ Microbiol 69:4732–4736 Rollero S, Bloem A, Camarasa C, Sanchez I, Ortiz-Julien A, Sablayrolles JM, Dequin S, Mouret JR (2015) Combined effects of nutrients and temperature on the production of fermentative aromas by Saccharomyces cerevisiae during wine fermentation. Appl Microbiol Biotechnol 99:2291–2304 Santos RM, Nogueira FCS, Brasil AA, Carvalho PC, Leprevost FV, Domont GB, Eleutherio ECA (2016) Quantitative proteomic analysis of the Saccharomyces cerevisiae industrial strains CAT-1 and PE-2. J Proteomics 151:114–121 Schoondermark-Stolk SA, Maria T, John C, ter Schure EG, Theo VC, Verkleij AJ, Johannes B (2005) Bat2p is essential in Saccharomyces cerevisiae for fusel alcohol production on the non-fermentable carbon source ethanol. FEMS Yeast Res 5:757–766 Sieg AG, Trotter PJ (2014) Differential contribution of the proline and glutamine pathways to glutamate biosynthesis and nitrogen assimilation in yeast lacking glutamate dehydrogenase. Microbiol Res 169:709–716 Stéphanie R, Audrey B, Anne O-J, Carole C, Benoit D (2018) Fermentation performances and aroma production of non-conventional wine yeasts are influenced by nitrogen preferences. FEMS Yeast Res 18:1–11 Stribny J, Querol A, Pérez-Torrado R (2016) Differences in enzymatic properties of the Saccharomyces kudriavzevii and Saccharomyces uvarum alcohol acetyltransferases and their impact on aroma-active compounds production. Front Microbiol 7:1–13 Sun H, Liu S, Mao J et al (2020) New insights into the impacts of huangjiu components on intoxication. Food Chem 317:126420 Torrea D, Varela C, Ugliano M, Ancin-Azpilicueta C, Francis IL, Henschke PA (2011) Comparison of inorganic and organic nitrogen supplementation of grape juice -effect on volatile composition and aroma profile of a Chardonnay wine fermented with Saccharomyces cerevisiae yeast. Food Chem 127:1072–1083 Wang J, Yuan CJ, Gao XL, Kang YL, Huang MQ, Wu JH, Liu YP, Zhang JL, Li HH, Zhang YY (2020) Characterization of key aroma compounds in Huangjiu from northern China by sensory-directed flavor analysis. Food Res Int 134:109238 Xie GF (2008) Functional constituents and health function of Chinese rice wine. Liquor Making 35:14–15 Xie J, Tian X-F, He S-G, Wei Y-L, Peng B, Wu Z-Q (2018) Evaluating the intoxicating degree of liquor products with combinations of fusel alcohols, acids, and esters. Molecules 23:1–12 Yan N, Du Y, Liu X, Chu M, Shi J, Zhang H, Liu Y, Zhang Z (2019) A comparative UHPLC-QqQ-MS-based metabolomics approach for evaluating Chinese and North American wild rice. Food Chem 275:618–627 Yan N, Du Y, Liu X, Zhang H, Liu Y, Shi J, Xue SJ, Zhang Z (2017) Analyses of effects of α-cembratrien-diol on cell morphology and transcriptome of Valsa mali var. mali. Food Chem 214:110–118 Yoshimoto H, Fukushige T, Yonezawa T, Sakai Y, Okawa K, Iwamatsu A, Sone H, Tamai Y (2001) Pyruvate decarboxylase encoded by the PDC1 gene contributes, at least partially, to the decarboxylation of α-ketoisocaproate for isoamyl alcohol formation in Saccharomyces cerevisiae. J Biosci Bioeng 92:83–85 Yuan N, Guan GK, Wan Z, Xia PY, Liu Y, Wang T, Chen YF, Zheng N, Han QY (2018) Application of modified yeast in reducing higher alcohols content in yellow rice wine. Liquor Making Sci Technol 8(290):85–88 Zhang N, Gur A, Gibon Y, Sulpice R, Flint-Garcia S, McMullen MD, Stitt M, Buckler ES (2010) Genetic analysis of central carbon metabolism unveils an amino acid substitution that alters maize NAD-dependent isocitrate dehydrogenase activity. Plos One 5:1–10 Zheng N, Jiang S, He Y, Chen Y, Zhang C, Guo X, Ma L, Xiao D (2020) Production of low-alcohol Huangjiu with improved acidity and reduced levels of higher alcohols by fermentation with scarless ALD6 overexpression yeast. Food Chem 321:1–7 Zhong X, Wang A, Zhang Y, Wu Z, Wen H (2019) Reducing higher alcohols by nitrogen compensation during fermentation of Chinese rice wine. Food Sci Biotechnol 29:805–816 Publisher’sNote Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Annals of Microbiology – Springer Journals
Published: Jan 14, 2021