Tài liệu Bacterial community dynamics and metabolite changes in myeolchi-aekjeot, a korean traditional fermented fish sauce, during fermentation

International Journal of Food Microbiology 203 (2015) 15–22 Contents lists available at ScienceDirect International Journal of Food Microbiology journal homepage: www.elsevier.com/locate/ijfoodmicro Short communication Bacterial community dynamics and metabolite changes in myeolchi-aekjeot, a Korean traditional fermented fish sauce, during fermentation Se Hee Lee, Ji Young Jung, Che Ok Jeon ⁎ Department of Life Science, Chung-Ang University, Seoul 156-756, Republic of Korea a r t i c l e i n f o Article history: Received 13 November 2014 Received in revised form 7 February 2015 Accepted 25 February 2015 Available online 3 March 2015 Keywords: Fermented fish sauce Myeolchi-aekjeot Anchovy Bacterial community dynamics Metabolites Tetragenococcus a b s t r a c t Myeolchi-aekjeot (MA) is a Korean traditional fish sauce, made by fermenting salted [approximately 25% (w/v)] anchovies. Three sets of MA samples, S-MA, M-MA, and L-MA, were prepared using small (5–8 cm), medium (8–10 cm), and large (10–13 cm) anchovies, respectively, and their bacterial communities and metabolites were investigated for 280 days. Bacterial community analysis using pyrosequencing revealed that, in S-MA, the initially dominant genera, including Phychrobacter, Photobacterium, and Vibrio, disappeared rapidly and Salinivibrio, Staphylococcus, and Tetragenococcus/Halanaerobium appeared sequentially as the major populations. In contrast, in M-MA and L-MA, the initially dominant genera were maintained relatively well during the early fermentation period, but eventually Tetragenococcus became predominant without the growth of Halanaerobium. The changes in the bacterial community occurred more quickly in MA prepared with smaller anchovies than in those prepared with larger anchovies. Metabolite analysis using 1H NMR showed that amino acids, glycerol, acetate, and lactate rapidly increased in all MA samples during the early fermentation period. Amino acids increased more quickly and then decreased after reaching their maximum level in S-MA, while they increased continually until the end of fermentation in L-MA. This suggests that the complete fermentation of L-MA may require more time than that for S-MA. A correlative analysis between bacterial communities and metabolites revealed that the increase in acetate, butyrate, and putrescine in S-MA was associated with the growth of Halanaerobium, which may be a useful indicator of anchovy sauce quality. © 2015 Elsevier B.V. All rights reserved. 1. Introduction Fish sauces are clear brown liquids with a salty taste and a distinctive fish flavor that are generally made by spontaneous long-term fermentation (more than 6 months) of salted whole small fish (e.g., anchovy and sand eel) (Fukui et al., 2012; Lopetcharat et al., 2001). Myeolchi-aekjeot (MA), a Korean traditional fermented fish sauce, is made by a long-term fermentation of salted anchovies. It has been known that anchovy (called myeolchi in Korean, Engraulis japonicus) has an average life span of 1.5 years and grows up to 15 cm in length and contains different amounts of protein, lipid, and carbohydrate depending on their size, which suggests that MA fermentation may be different depending on the size of anchovies. The naturally occurring fermentation of fish sauces without starter cultures leads to the growth of various halophilic or halotolerant microbes due to their high salt conditions (Fukui et al., 2012; Udomsil et al., 2010). So far, various bacteria including Achromobacter, Bacillus, ⁎ Corresponding author at: Department of Life Science, Chung-Ang University, 84, HeukSeok-Ro, Dongjak-Gu, Seoul 156-756, Republic of Korea. Tel.: +82 2 820 5864; fax: +82 2 825 5206. E-mail address: [email protected] (C.O. Jeon). http://dx.doi.org/10.1016/j.ijfoodmicro.2015.02.031 0168-1605/© 2015 Elsevier B.V. All rights reserved. Halomonas, Micrococcus, Brevibacterium, Halobacterium, Vibrio, Flavobacterium, Staphylococcus, and Tetragenococcus species have been identified in fish sauces by mainly culture-dependent approaches despite the presence of many uncultivable microbes (Fukui et al., 2012; Guan et al., 2011; Lopetcharat et al., 2001; Taira et al., 2007). Culture-independent approaches based on clonal analysis of 16S rRNA genes have been also applied to investigations of microbial communities of fish sauces (Chuon et al., 2014; Kim and Park, 2014), but these approaches have limitations in ascertaining microbial community dynamics during fermentation as they tend to be time-consuming, laborious, and hence relatively low throughput. Therefore, highthroughput pyrosequencing has been extensively applied to better understand the microbial community dynamics of fermented foods (Humblot and Guyot, 2009; Jung et al., 2013; Lee et al., 2014a; Roh et al., 2010; Sakamoto et al., 2011). Fish tissues are hydrolyzed by endogenous enzymes originated in the fish as well as exogenous enzymes derived from microbes during the fermentation of fish sauces (Jung et al., 2013; Yongsawatdigul et al., 2007). Therefore, studies of the microbial dynamics as well as the metabolite changes are indispensable to understand fermentation processes of fish sauces because the metabolites reflect collective phenotypic results of microbial communities and endogenous enzymes 16 S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 (Ercolini et al., 2011, 2013; Lee et al., 2014a). Studies on both the microbial communities and the metabolite changes during MA fermentation are very important to gain a better understanding of MA fermentation. However, to the best of our knowledge, to date, no simultaneous studies on the microbial community dynamics and metabolite changes during MA fermentation have been performed. 1 H NMR is a very comprehensive, relatively easy, and nondestructive method for the simultaneous analysis of multiple metabolites present in fermented foods. A combination of pyrosequencing and 1H NMR approach has been suggested to be one of the more powerful ways to better understand the relationships between the microbial community dynamics and metabolite changes during food fermentation (Jeong et al., 2013; Jung et al., 2013, 2014; Lee et al., 2014a, 2014b). Because fermented salted seafood is usually made by fermentation under highly salted [20–30% (w/w)] conditions, it has been hypothesized that archaea may contribute to salted seafood fermentation (Roh et al., 2010; Tapingkae et al., 2010). However, recent reports have shown that archaea may not play an important role in salted seafood fermentation (Jung et al, 2013; Lee et al., 2014a). Therefore, in this study, three sets of MA samples were prepared using anchovies of different sizes and a combination of pyrosequencing and 1H NMR was applied to investigate bacterial succession and metabolite changes during MA fermentation. 2. Material and methods 2.1. Preparation of myeolchi-aekjeot samples and analysis Three sets of myeolchi-aekjeot (MA) samples with approximately 25% (w/v) salt concentration were prepared in triplicate using anchovies (E. japonicus) of different lengths (5–8 cm, small (S); 8–10 cm, medium (M); 10–13 cm, large (L)) as described previously (Jung et al., 2013; Lee et al., 2014a). Because the average size of anchovies harvested from Korean waters depends on the fishing season, the three sets of MA samples were prepared independently at different times of the year. The crude protein, carbohydrate, and lipid contents of the anchovies were measured according to their standard analysis methods (MFDS, 2013). For the preparation of triplicated MA samples, fresh anchovies and solar salts (Shinan, Korea) were equally dispensed into three plastic containers to include 10-kg and 2.7-kg portions, respectively, and then 4 l of 25% (w/v) solar salt solution was added into each container to completely cover the anchovies. The triplicated three sets of MA samples (S-MA, M-MA, and L-MA) were incubated at 25 °C and their pH values were monitored. Four milliliters of supernatants (liquid fraction of MA) was intermittently sampled from each container and microorganisms were harvested by centrifugation (8000 rpm for 20 min at 4 °C). Microorganisms harvested from the triplicated samples were combined and stored at −80 °C until the bacterial community analysis. The centrifugation supernatants were stored separately at −80 °C for respective metabolite analysis. Bacterial abundances in MA samples were estimated using quantitative PCR (qPCR) according to a previously described method (Lee et al., 2014a). NaCl concentrations in MA samples were measured according to the Mohr method (AOAC, 2000). 2.2. Barcoded pyrosequencing and data processing Total genomic DNA extraction from MA samples, barcoded pyrosequencing of bacterial 16S rRNA genes, and data processing of sequencing reads were conducted according to the procedure described previously (Lee et al., 2012). Taxonomic classifications of high-quality reads were performed using the RDP naïve Bayesian rRNA Classifier 2.6 (Wang et al., 2007) at an 80% confidence threshold. Rarefaction analysis and calculation of operational taxonomic units (OTU), Shannon–Weaver and Chao1 richness indices, and evenness were performed using the RDP pyrosequencing pipeline (http://pyro.cme. msu.edu/) (Cole et al., 2014). To compare the bacterial communities among the MA samples, weighted hierarchical clustering analysis and principal coordinate analysis (PCoA) were performed according to the procedure described previously (Lee et al., 2012). 2.3. Metabolite analysis using 1H NMR and redundancy analysis Metabolite analysis in MA samples using 1H NMR and redundancy analysis (RDA) for bacterial community and metabolite changes using the vegan package (Oksanen et al., 2011) in the R programming environment (http://cran.r-project.org/) were conducted according to the procedure described previously (Jung et al., 2013). 2.4. Pyrosequencing data accession number The pyrosequencing data of the 16S rRNA genes are publicly available in the NCBI Short Read Archive (SRA) under accession no. SRX755990. 3. Results 3.1. Compositions of differently sized anchovies The crude protein, crude carbohydrate, and total lipid contents of anchovies of the three different size ranges used for the preparation of three sets of MA samples were measured in triplicate (Table 1). The average crude protein content ranged between approximately 17.6 and 19.7 g/100 g fresh anchovy. The crude carbohydrate content was higher in the large anchovies (approximately 1.1 g/100 g fresh anchovy) than in small and medium anchovies, both of which were approximately 0.6 g/100 g fresh anchovy. The total lipid content was quite different depending on the anchovy size. The total lipid contents of the large anchovies were approximately 6.0 g/100 g fresh anchovy, while those of small and medium anchovies were approximately 1.5 and 3.7 g/100 g fresh anchovy, respectively. 3.2. General features of MA fermentation The NaCl concentrations were nearly constant, approximately 24.8 ± 0.6% (w/v), in all MA samples during the entire fermentation period. The initial pH values of the MA samples were in the range of 5.75–6.01 and the pH values were relatively constant until 50 days of fermentation in all MA samples (Fig. 1A). The pH of the M-MA and LMA samples remained relatively constant until the end of fermentation, while that in S-MA increased rapidly to the highest value of approximately pH 6.5 at 170 days, and then decreased to approximately pH 6.2 (day 280). The initial bacterial 16S rRNA gene copy number in the MA samples was inversely related to anchovy size (Fig. 1B). The initial bacterial 16S rRNA gene copy numbers in S-MA, M-MA, and L-MA were approximately 1.2 × 108, 3.1 × 107, and 7.1 × 106 copies/ml, respectively. The copy number in S-MA increased starting from the early fermentation period without a lag phase, while that in the larger anchovies, particularly L-MA, was relatively constant during the early fermentation period. The copy numbers in M-MA and L-MA increased after approximately 60 and 80 days, respectively. The 16S rRNA gene copies were higher in MA samples prepared with smaller anchovies than in those prepared with larger anchovies during the entire fermentation period. The bacterial 16S rRNA gene copy number in S-MA increased to its highest value of approximately 9.5 × 109 copies/ml at 170 days, whereas the highest copy numbers observed in M-MA and L-MA were approximately 3.4 × 109 and 4.5 × 108 copies/ml, respectively. The bacterial 16S rRNA gene copy number in S-MA was approximately 2 orders of magnitude higher than that in L-MA. S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 17 Table 1 Compositions (%) of small, medium, and large anchovies used for the preparation of three myeolchi-aekjeot samples. Anchovies (fishing time) Range of length (cm) Range of weight (g) Average content ± SDa Moisture Total proteins Total carbohydrate Lipids Ash Small (Aug. 2011) Medium (Aug. 2011) Large (Jan. 2012) 5–8 8–10 10–13 2–5 5–10 10–16 74.9 ± 0.3 72.3 ± 0.3 72.3 ± 0.4 18.9 ± 0.2 19.7 ± 0.3 17.6 ± 0.2 0.6 ± 0.1 0.6 ± 0.1 1.1 ± 0.2 1.5 ± 0.1 3.7 ± 0.2 6.0 ± 0.2 4.1 ± 0.2 3.7 ± 0.2 3.0 ± 0.2 a The contents were measured in triplicate and SD represents standard deviation. 3.3. Bacterial diversity changes in MA samples during fermentation A total of 159,701 pyrosequencing reads for bacterial 16S rRNA genes were generated from 39 samples (three sets of MA samples × 13 samplings). After the removal of low-quality, chimeric, and Streptophyta (plant) 16S rRNA gene sequences, a total of 130,727 high-quality bacterial reads, with a 462-bp average read length and an average of 3,351 reads per sample, were obtained. The rarefaction curves showed that the bacterial diversities increased during the early MA fermentation period in all three sample sets (Supplementary Fig. S1). After their initial increases, the bacterial diversities decreased and approached their respective asymptotes as the fermentation progressed. The bacterial diversities increased again during the end of the fermentation period. The bacterial diversity changes occurred more quickly in the MA samples of smaller anchovies, especially in SMA, than those of larger anchovies. Although the number of sequencing reads affected statistical diversity indices including OTU, Shannon– Weaver, Chao1, and evenness, the diversity indices were also in good Fig. 1. Changes in pH (A) and total bacterial 16S rRNA gene copy numbers (B) during fermentation of myeolchi-aekjeot samples prepared with small (S-MA), medium (M-MA), and large (L-MA) anchovies. agreement with the rarefaction analysis results (Supplementary Table S1). 3.4. Bacterial community changes in MA samples during fermentation The bacterial sequencing reads were classified at the phylum and genus levels. At the phylum level, Proteobacteria and Firmicute were predominant in all MA samples during the entire fermentation period (Supplementary Fig. S2), which is consistent with previous reported results for other fermented salted seafood (Jung et al., 2013; Lee et al., 2014a, 2014b). Proteobacteria was initially predominant, but was rapidly replaced with Firmicutes as the fermentation progressed. The replacements occurred earlier in the MA made with smaller anchovies, especially in S-MA, than larger anchovies. At the genus level, Photobacterium, Vibrio, Phychrobacter, unclassified Gammaproteobacteria, and unclassified Alteromonadales, which might be primarily derived from raw anchovies, were identified in all initial MA samples (day 0), although their relative abundances depended on the anchovy size (Fig. 2). In S-MA, the initially dominant genera disappeared rapidly within only 5 days, and Salinivibrio became the predominant genus, followed by Staphylococcus; after 30 days of fermentation, Tetragenococcus became predominant until the end of fermentation. Halanaerobium, the growth of which was trivial or not detectable in M-MA and L-MA, also became relatively predominant after 80 days of fermentation. In M-MA, Phychrobacter, which was a major genus in the initial samples, was still maintained as a predominant group until 60 days of fermentation, but Tetragenococcus became predominant after 80 days of fermentation. In L-MA, the initial bacterial community was relatively stable until 80 days of fermentation without predominance by a particular genus, and Tetragenococcus became predominant after 100 days of fermentation. In conclusion, the bacterial successions were quite different depending on the anchovy size during the early fermentation period, but eventually Tetragenococcus predominated regardless of the anchovy size. To more rigorously evaluate the results presented above for the bacteria community changes in the MA samples during the fermentation period, hierarchical clustering analysis and PCoA were performed using all high-quality sequencing reads. The bacterial communities in the MA samples during the early fermentation period were distinctly clustered depending on the size of the anchovy used to prepare the MA, but they were grouped more coherently during the end of fermentation (Supplementary Fig. S3A); this is consistent with the predominance of Tetragenococcus in all MA samples during the end of fermentation shown in Fig. 2. However, the S-MA samples were slightly more distantly clustered from those of M-MA and L-MA, suggesting that their corresponding bacterial communities were different during the end of fermentation, which also explains the presence of Halanaerobium in S-MA as shown in Fig. 2. The bacterial community changes shown in the bacterial classification (Fig. 2) and hierarchical clustering analysis (Supplementary Fig. S3A) during the fermentation were also confirmed by PCoA (Supplementary Fig. S3B). The PCoA results also show that the bacterial community changes occurred more rapidly in S-MA than in M-MA and L-MA and they progressed differently during the early fermentation period depending on the size of anchovy used to prepare the MA, but eventually became similar during the end of fermentation. The data plots showing the bacterial community changes shifted steadily 18 S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 Fig. 2. Bacterial taxonomic compositions at the genus level of myeolchi-aekjeot samples prepared with small (S-MA), medium (M-MA), and large (L-MA) anchovies during fermentation. Others are composed of the phyla or the genera, each showing a percentage of reads b2.0% of the total reads in all samples of each panel. from day 0 to day 60 in S-MA, while the data plots in M-MA and L-MA shifted very quickly after 60 and 80 days, respectively, which is consistent with the bacterial abundance profiles (Fig. 1B), the rarefaction curves (Supplementary Fig. S1), and the bacterial community changes (Fig. 2). The PCoA results also show that the S-MA samples were separated from M-MA and L-MA during the late fermentation period, as shown in the hierarchical clustering analysis (Supplementary Fig. S3A). 3.5. Metabolite changes in MA during fermentation The metabolite analysis shows that amino acids were the major metabolites in the MA samples (Fig. 3). The concentrations of amino acids increased rapidly in all MA samples during the early fermentation period, but their overall profiles during the entire fermentation period were slightly different depending on the anchovy size. The concentrations of amino acids including arginine, aspartate, glutamate, glycine, and lysine decreased after reaching their maximum level during the end of fermentation in the MA prepared with smaller anchovies, especially in S-MA, while the concentrations of amino acids in L-MA increased continually until the end of fermentation, suggesting that more time may be required to accomplish the complete fermentation of L-MA than that required for the fermentation of S-MA. Glucose, glycerol, acetate, butyrate, lactate, and putrescine were also identified as the primary organic compounds during the MA fermentation (Fig. 4). The concentration of glucose, which might be derived from anchovy glycogen, increased quickly in L-MA during the initial fermentation period and its maximum level was much higher than that in SMA and M-MA, which is consistent with the high carbohydrate content of large anchovies as shown in Table 1. The glucose concentration gradually decreased in all MA samples after an initial increase (Fig. 4A). The concentration of glycerol, which might be derived from the hydrolysis of lipid, increased quickly during the early fermentation period in all MA samples (Fig. 4B). The glycerol concentration increased S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 19 Fig. 3. Changes in major amino acids and nitrogen compounds identified by 1H NMR in myeolchi-aekjeot samples prepared with small (S-MA), medium (M-MA), and large (L-MA) anchovies during fermentation. Data are presented as average values ± standard deviations in triplicate. until the end of fermentation in M-MA and L-MA, while it decreased in S-MA after 80 days. The glycerol concentration was significantly higher in the MA prepared with larger anchovies than smaller anchovies, which is consistent with the lipid content of the three size anchovy groups as shown in Table 1. The concentrations of acetate, which might be derived from the fermentation or hydrolysis of carbohydrates (glucose) or lipids, increased rapidly in all MA samples as the fermentation progressed. Interestingly, the acetate concentrations were higher in the MA prepared with smaller anchovies, although their carbohydrate (glucose) and lipid concentrations were much lower compared to those of the larger anchovies (Fig. 4C). The concentrations of butyrate were relatively low during the entire fermentation period, but slightly increased in S-MA during the end of fermentation (Fig. 4D). Lactate was identified in MA samples as the major organic acid (Fig. 4E). The lactate levels were relatively constant during the early fermentation period, but they began to increase during the middle fermentation period. The lactate increase occurred earlier in the MA prepared with smaller anchovies. Biogenic amines that are mainly produced by the microbial decarboxylation of amino acids or other nitrogen compounds in foods are important indicators of the quality of fermented seafoods (Halász et al., 1994). Putrescine was detected only in the S-MA samples (Fig. 4F) and its level increased rapidly after 45 days of fermentation. However, other biogenic amines including histamine, tyramine, and cadaverine were not detected in any MA samples. 3.6. Multivariate statistical analysis To statistically assess the changes in metabolites and bacterial abundances during MA fermentation, an RDA was performed on the basis of bacterial abundances at the genus level (Fig. 2) and metabolite concentrations (Figs. 3 and 4). The RDA triplot showed that the fermentation processes were different depending on the size of anchovy used for the preparation of MA, and the production of putrescine in S-MA might be related to the growth of Halanaerobium during the end of the fermentation period (Fig. 5). 4. Discussion The fact that anchovies of various sizes are used for the production of MA and the compositions of anchovies vary according to size (Table 1) suggests that fermentation properties including pH, bacterial abundance, and bacterial community and metabolite changes may be different depending on the size of anchovy used for the preparation of MA. However, to the best of our knowledge, no study has investigated microbial community dynamics and metabolite changes during the entire fermentation period in MA prepared with anchovies of different sizes. Bacterial community analysis revealed that in S-MA Salinivibrio, Staphylococcus, and Tetragenococcus/Halanaerobium appeared 20 S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 Fig. 4. Changes in major organic compounds [glucose (A), glycerol (B), acetate (C), butyrate (D), lactate (E), and putrescine (F)] identified by 1H NMR in myeolchi-aekjeot samples prepared with small (S-MA), medium (M-MA), and large (L-MA) anchovies during fermentation. Data are presented as average values ± standard deviations in triplicate. sequentially as the major populations with the rapid disappearance of the initially dominant genera as the fermentation progressed, while in M-MA and L-MA the initially dominant genera were relatively stable until Tetragenococcus appeared as the predominant genus (Fig. 2). Fig. 1B shows that the bacterial abundances in S-MA increased from the early fermentation period without a lag phase, while the bacterial abundances in M-MA and L-MA began to increase after approximately 60 and 80 days of fermentations, respectively, which is consistent with the predominance of Tetragenococcus. These results suggest that in S-MA the growth of Salinivibrio, Staphylococcus, and Tetragenococcus/Halanaerobium occurred evidently during the fermentation, whereas in M-MA and L-MA bacterial growths during fermentation was negligible except for Tetragenococcus. Furthermore, the growth of the initially dominant genera, Photobacterium, Vibrio, and Phychrobacter that may include some pathogenic strains (López et al., 2012) did not occur, although they remained dominant until Tetragenococcus became abundant. The formation of amino acids through the proteolysis of proteins during fish fermentation is a very important aspect of MA production because amino acids are closely related to the taste (umami) and flavor of fermented seafood (Mok et al., 2000; Özden, 2005). The concentration of amino acids increased rapidly even in M-MA and L-MA with low abundances and trivial growth of bacteria during the early fermentation period (Fig. 3), which is consistent with previous results using S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 21 Fig. 5. A redundancy analysis (RDA) showing correlations among myeolchi-aekjeot samples, relative bacterial abundances, and metabolite concentrations during fermentation of myeolchi-aekjeot samples prepared with small (S-MA), medium (M-MA), and large (L-MA) anchovies. Numbers beside the symbols represent the fermentation time (days). The directions and lengths of the dotted straight arrows indicate the influence of the bacterial population on the myeolchi-aekjeot samples. The thick arrows indicate the routes of myeolchi-aekjeot fermentation on the RDA triplot. fermentation of other types of salted seafood (Jung et al., 2013; Lee et al., 2014a and b). These results suggest that the increase in amino acids during the early MA fermentation might be more closely associated with endogenous enzymes than bacterial proteinases, which is consistent with previous reports showing that bacterial proteinases have low activities in high-salt conditions (Guan et al., 2011; Nam et al., 1998). However, it is clear that the bacterial populations had influenced the changes in other metabolites during the MA fermentation including glucose, acetate, lactate, and putrescine as shown in Fig. 4. Our previous studies showed that Halanaerobium is a potential indicator of putrefaction or over-fermentation of fermented salted seafood because members of Halanaerobium are responsible for the production of acetate, butyrate, and biogenic amines through the fermentation of monosaccharides, amino acids, and glycerol (Jung et al., 2013; Lee et al., 2014a; Lee et al., 2014b). The carbohydrate and lipid contents were lower in S-MA than in M-MA and L-MA (Table 1) and as a result, the concentrations of glucose and glycerol were also lower in S-MA than in M-MA and L-MA (Fig. 4A and B). However, the concentrations of acetate and butyrate in S-MA were higher during the end of fermentation than those in M-MA and L-MA (Fig. 4C and D), which might be related to the growth of Halanaerobium. Putrescine, which is formed through the decarboxylation of ornithine, increased rapidly in S-MA, which is in good agreement with the growth of Halanaerobium (Fig. 5). These results indicate that Halanaerobium might be responsible for the production of acetate, butyrate, and putrescine during MA fermentation and the growth of Halanaerobium could be considered an important indicator of anchovy sauce quality. Brown et al. (2011) reported that a Halanaerobium species, Halanaerobium hydrogenifirmans, harbors an ornithine decarboxylase gene, suggesting that Halanaerobium species may be responsible for the putrescine production. The pH increased in S-MA as the fermentation progressed (Fig. 1A), which might be related to the production of putrescine, suggesting that the pH profile is also a potential indicator of the production of biogenic amines. Previous studies have also shown that members of Tetragenococcus, which are halophilic lactic acid bacteria, have been detected in various fermented seafood as the major population and they play important roles in taste and flavor enhancement of fish sauces during fermentation (Chen et al., 2006; Fukami et al., 2004; Kim and Park, 2014; Kobayashi et al., 2000; Kuda et al., 2012; Thongsanit et al., 2002; Udomsil et al., 2011). The concentration of lactate, which might be released from the muscles of raw anchovies, increased very quickly during the initial fermentation period (Fig. 4E), and increased again after the middle fermentation period, which is consistent with the growth of Tetragenococcus. Although the bacterial community changes were quite different depending on the anchovy sizes during the early fermentation period in this study, eventually Tetragenococcus became predominant regardless of anchovy sizes during the end of fermentation (Fig. 2). Therefore, this study also suggests that Tetragenococcus could be a good bacterial starter to improve the taste and flavor characteristics of myeolchi-aekjeot. This is the first study to investigate bacterial community dynamics and metabolite changes during the entire Korean fish sauce fermentation. However, additional studies regarding the relationships among microbial communities, metabolites, and sensory characteristics (taste, flavor, and food safety) and the effects of fermentation conditions (e.g., temperature, salt concentration) may be required for a better understanding of myeolchi-aekjeot fermentation, which will shed light on the production of safe and high-quality fish sauces. Acknowledgments This work was supported by the Technology Development Program for Agriculture and Forestry (3120023) of the Ministry for Agriculture, Food and Rural Affairs and the Cooperative Research Program for 22 S.H. Lee et al. / International Journal of Food Microbiology 203 (2015) 15–22 Agriculture Science & Technology Development (Project No. PJ00999302), RDA, Republic of Korea. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ijfoodmicro.2015.02.031. References AOAC, 2000. Official Methods of Analysis. 17th ed. Association of Official Analytical Chemists, Washington, D.C., USA. Brown, S.D., Begemann, M.B., Mormile, M.R., Wall, J.D., Han, C.S., Goodwin, L.A., Pitluck, S., Land, M.R., Hauser, L.J., Elias, D.A., 2011. Complete genome sequence of the haloalkaliphilic, hydrogen-producing bacterium Halanaerobium hydrogeniformans. J. Bacteriol. 193, 3682–3683. Chen, Y.S., Yanagida, F., Hsu, J.S., 2006. Isolation and characterization of lactic acid bacteria from dochi (fermented black beans), a traditional fermented food in Taiwan. Lett. Appl. Microbiol. 43, 229–235. Chuon, M.R., Shiomoto, M., Koyanagi, T., Sasaki, T., Michihata, T., Chan, S., Mao, S., Enomoto, T., 2014. Microbial and chemical properties of Cambodian traditional fermented fish products. J. Sci. Food Agric. 94, 1124–1131. Cole, J.R., Wang, Q., Fish, J.A., Chai, B., McGarrell, D.M., Sun, Y., Brown, C.T., Porras-Alfaro, A., Kuske, C.R., Tiedje, J.M., 2014. Ribosomal Database Project: data and tools for high throughput rRNA analysis. Nucleic Acids Res. 42, D633–D642. Ercolini, D., Ferrocino, I., Nasi, A., Ndagijimana, M., Vernocchi, P., La Storia, A., Laghi, L., Mauriello, G., Guerzoni, M.E., Villani, F., 2011. Monitoring of microbial metabolites and bacterial diversity in beef stored under different packaging conditions. Appl. Environ. Microbiol. 77, 7372–7381. Ercolini, D., Pontonio, E., De Filippis, F., Minervini, F., La Storia, A., Gobbetti, M., Di Cagno, R., 2013. Microbial ecology dynamics during rye and wheat sourdough preparation. Appl. Environ. Microbiol. 79, 7827–7836. Fukami, K., Funatsu, Y., Kawasaki, K., Watabe, S., 2004. Improvement of fish sauce odor by treatment with bacteria isolated from the fish-sauce mush (moromi) made from rigate mackerel. J. Food Sci. 69, 45–49. Fukui, Y., Yoshida, M., Shozen, K., Funatsu, Y., Takano, T., Oikawa, H., Yano, Y., Satomi, M., 2012. Bacterial communities in fish sauce mash using culture-dependent and -independent methods. J. Gen. Appl. Microbiol. 58, 273–281. Guan, L., Cho, K.H., Lee, J.H., 2011. Analysis of the cultivable bacterial community in jeotgal, a Korean salted and fermented seafood, and identification of its dominant bacteria. Food Microbiol. 28, 101–113. Halász, A., Baráth, Á., Simon-Sarkadi, L., Holzapfel, W., 1994. Biogenic amines and their production by microorganisms in food. Trends Food Sci. Technol. 5, 42–49. Humblot, C., Guyot, J., 2009. Pyrosequencing of tagged 16S rRNA gene amplicons for rapid deciphering of the microbiomes of fermented foods such as pearl millet slurries. Appl. Environ. Microbiol. 75, 4354–4361. Jeong, S.H., Jung, J.Y., Lee, S.H., Jin, H.M., Jeon, C.O., 2013. Microbial succession and metabolite changes during fermentation of dongchimi, traditional Korean watery kimchi. Int. J. Food Microbiol. 164, 46–53. Jung, J.Y., Lee, S.H., Lee, H.J., Jeon, C.O., 2013. Microbial succession and metabolite changes during fermentation of saeu-jeot: traditional Korean salted seafood. Food Microbiol. 34, 360–368. Jung, J.Y., Lee, S.H., Jeon, C.O., 2014. Microbial community dynamics during fermentation of doenjang-meju, traditional Korean fermented soybean. Int. J. Food Microbiol. 185, 112–120. Kim, M.S., Park, E.J., 2014. Bacterial communities of traditional salted and fermented seafoods from Jeju island of Korea using 16S rRNA gene clone library analysis. J. Food Sci. 79, M927–M934. Kobayashi, T., Kimura, B., Fujii, T., 2000. Differentiation of Tetragenococcus populations occurring in products and manufacturing processes of puffer fish ovaries fermented with rice-bran. Int. J. Food Microbiol. 56, 211–218. Kuda, T., Izawa, Y., Ishii, S., Takahashi, H., Torido, Y., Kimura, B., 2012. Suppressive effect of Tetragenococcus halophilus, isolated from fish-nukazuke, on histamine accumulation in salted and fermented fish. Food Chem. 130, 569–574. Lee, H.J., Jung, J.Y., Oh, Y.K., Lee, S.S., Madsen, E.L., Jeon, C.O., 2012. Comparative survey of rumen microbial communities and metabolites across one caprine and three bovine groups, using bar-coded pyrosequencing and 1H nuclear magnetic resonance spectroscopy. Appl. Environ. Microbiol. 78, 5983–5993. Lee, S.H., Jung, J.Y., Jeon, C.O., 2014a. Microbial successions and metabolite changes during fermentation of salted shrimp (saeu-jeot) with different salt concentrations. PLoS ONE 9, e90115. Lee, S.H., Jung, J.Y., Jeon, C.O., 2014b. Effects of temperature on microbial succession and metabolite change during saeu-jeot fermentation. Food Microbiol. 38, 16–25. Lopetcharat, K., Choi, Y.J., Park, J.W., Daeschel, M.A., 2001. Fish sauce products and manufacturing: a review. Food Rev. Int. 17, 65–88. López, J.R., Navas, J.I., Thanantong, N., de la Herran, R., Sparagano, O.A.E., 2012. Simultaneous identification of five marine fish pathogens belonging to the genera Tenacibaculum, Vibrio, Photobacterium and Pseudomonas by reverse line blot hybridization. Aquaculture 324–325, 33–38. MFDS, 2013. Food code. Ministry of Food and Drug Safety. Cheongju, Korea. Mok, C.K., Lee, J.Y., Song, K.T., Kim, S.Y., Lim, S.B., Woo, G.J., 2000. Changes in physicochemical properties of salted and fermented shrimp at different salt levels. Korean J. Food Sc. Technol. 32, 187–191. Nam, E.J., Oh, S.W., Jo, J.H., Kim, Y.M., Yang, C.B., 1998. Purification and characterization of alkaline protease from saewoo-jeot, salted and fermented shrimp (Acetes japonicus). Korean J. Food Sci. Technol. 30, 82–89. Oksanen, J., Blanchet, F.G., Kindt, R., Legendre, P., Minchin, P.R., O'Hara, R.B., Simpson, G.L., Solymos, P., Stevens, M.H.H., Wagner, H., 2011. Vegan: community ecology package. R package version 1.17–12. http://cran.r-project.org/. Özden, Ö., 2005. Changes in amino acid and fatty acid composition during shelf-life of marinated fish. J. Sci. Food Agric. 85, 2015–2020. Roh, S.W., Kim, K.H., Nam, Y.D., Chang, H.W., Park, E.J., Bae, J.W., 2010. Investigation of archaeal and bacterial diversity in fermented seafood using barcoded pyrosequencing. ISME J. 4, 1–16. Sakamoto, N., Tanaka, S., Sonomoto, K., Nakayama, J., 2011. 16S rRNA pyrosequencingbased investigation of the bacterial community in nukadoko, a pickling bed of fermented rice bran. Int. J. Food Microbiol. 144, 352–359. Taira, W., Funatsu, Y., Satomi, M., Takano, T., Abe, H., 2007. Changes in extractive components and microbial proliferation during fermentation of fish sauce from underutilized fish species and quality of final products. Fish. Sci. 73, 913–923. Tapingkae, W., Tanasupawat, S., Parkin, K.L., Benjakul, S., Visessanguan, W., 2010. Degradation of histamine by extremely halophilic archaea isolated from high saltfermented fishery products. Enzym. Microb. Technol. 46, 92–99. Thongsanit, J., Tanasupawat, S., Keeratipibul, S., Jitikavanich, S., 2002. Characterization and identification of Tetragenococcus halophilus and Tetragenococcus muriaticus strains from fish sauce (nam-pla). Jpn. J. Lactic Acid Bacteria 13, 46–52. Udomsil, N., Rodtong, S., Tanasupawat, S., Yongsawatdigul, J., 2010. Proteinase-producing halophilic lactic acid bacteria isolated from fish sauce fermentation and their ability to produce volatile compounds. Int. J. Food Microbiol. 141, 186–194. Udomsil, N., Rodtong, S., Choi, Y.J., Hua, Y., Yongsawatdigul, J., 2011. Use of Tetragenococcus halophilus as a starter culture for flavor improvement in fish sauce fermentation. J. Agric. Food Chem. 59, 8401–8408. Wang, Q., Garrity, G.M., Tiedje, J.M., Cole, J.R., 2007. Na ve Bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Appl. Environ. Microbiol. 73, 5261–5267. Yongsawatdigul, J., Rodtong, S., Raksakulthai, N., 2007. Acceleration of Thai fish sauce fermentation using proteinases and bacterial starter cultures. J. Food Sci. 72, M382–M390.