Pyrosequencing-Based Analysis of the Bacterial Community During

Full Paper Pyrosequencing-based analysis of the bacterial community during fermentation of Alaska pollock sikhae: traditional Korean seafood (Received July 2, 2014; Accepted August 13, 2014) Hyo Jin Kim,1 Min-Jeong Kim,1 Timothy Lee Turner,2,3 and Myung-Ki Lee1,* 1 Fermentation Research Center, Korea Food Research Institute, Gyeonggi-Do, Seongnam-Si 463‒746, Republic of Korea 2 Institute for Genomic Biology, University of Illinois at Urbana-Champaign, Urbana, Illinois 61801, USA 3 Department of Food Science and Human Nutrition, University of Illinois at Urbana-Champaign, 1206 West Gregory Dr., Urbana, Illinois 61801, USA

fermented foods were designed for long-term storage, to We analyzed the bacterial community of Alaska supplement various nutrients, and to provide featured pollock sikhae, a traditional Korean food made by flavors. Jeotgal and sikhae are both traditional fermented natural fermentation with Alaska pollock, utilizing seafoods made for long-term storage (Rhee et al., 2011). pyrosequencing. We fermented the Alaska pollock While jeotgal contains a relatively high concentration of salt sikhae at two different temperatures (10°C and [generally 20‒30% (w/w)], sikhae uses a low concentration 20°C). Before fermentations, the bacterial commu- of salt [＜7%] (Cha et al., 2004; Jung et al., 2013a, b; Guan nity was varied. After fermentations, however, Lac- et al., 2011; Lee et al., 2014). In contrast to jeotgal, sikhae tobacillus sakei became dominant. The Alaska pol- contains grains that provide abundant carbon sources to lock sikhae sample before fermentations contained microbes. Gajami-sikhae, Alaska pollock sikhae, and squid only 2% L. sakei, but the sample on day 6 of fermen- sikhae are the three most popular Korean sikhae. In addition tation at 10°C comprised 74% L. sakei (90% at to salt and fish (Alaska pollock), Alaska pollock sikhae is 20°C). In addition, we observed a reduction in the composed of cooked grains, white radish, red pepper powder, composition of unpreferred bacterial species for malt, garlic, and ginger. Alaska pollock sikhae is mainly foods after fermentation. The composition of unpre- divided into Kangwondo- and Kyoungsangdo-style recipes. ferred bacterial species was more than 30% of total While Kyoungsangdo-style recipes use dried Alaska pollock, reads in samples before fermentation and decreased undried fish are utilized for the Kangwondo-style recipes. to less than 0.2% after fermentation. This result Despite its importance as a Korean fermented food, there is suggested that the fermentation of Alaska pollock limited research regarding sikhae and our understanding of sikhae can be beneficial for food safety. Alaska pol- the fermentation mechanisms of sikhae is largely incomplete. lock sikhae might be a favorable habitat for L. sakei. Novel bacterial strains were isolated from sikhae, mainly Our study is the first report illustrating the altera- from gajami-sikhae (Kim et al., 2009; Jung et al., 2012a; tion of the bacterial community of Alaska pollock Shin et al., 2011; Park et al., 2013, 2014). Alishewanella sikhae during fermentation utilizing pyrosequenc- jeotgali MS1T was isolated from gajami-sikhae and its ing analysis. genome was fully sequenced (Kim et al., 2009; Jung et al., 2012a). Shin and co-workers isolated Leutobacter celer Key words: Alaska pollock sikhae; bacterial com- NAL101T from gajami-sikhae (Shin et al., 2011). They munity analysis; Lactobacillus sakei; pyrosequenc- suggested that the gajami-sikhae habitat might not be the ing analysis optimal growth condition for the NAL101T strain (Shin et al., 2011). Park and colleagues isolated Lactobacillus plantarum LG42 from gajami-sikhae and showed its antiobesity Introduction effect on high-fat-diet-induced obese mice (Park et al., 2013, 2014). Recently, we characterized the bacterial community Numerous fermented foods are popular in Korea. Korean of eight gajami-sikhae products and isolated beneficial

＊ Corresponding author: Dr. Myung-Ki Lee, Fermentation Research Center, Korea Food Research Institute, Gyeonggi-Do, Seongnam-Si 463‒746, Republic of Korea. E-mail: [email protected] None of the authors of this manuscript has any financial or personal relationship with other people or organizations that could inappropriately influence their work. 228 KIM et al. strains (Kim et al., 2014). The study indicated that the lactic adjusted to equimolar concentrations and sequencing was acid bacteria were dominant in the bacterial community of executed on a 454 GS Junior system (Roche, Basel, Switzer- the gajami-sikhae products. land) following the manufacturer’s instructions. Novel strains isolated from various fermented foods had Processing of sequenced data from the microbial community been characterized and their functionality as probiotics had in the Alaska pollock sikhae samples using pyrosequencing. also been investigated (Abbasiliasi et al., 2012; Shin et al., Sequences collected by pyrosequencing were sorted by a 2008; Wu et al., 2009; Chang et al., 2010; Park et al., 2013). unique barcode for each sikhae sample. Among the sorted Although research on an individual strain from a specific sequences, sequences less than the 300 bp read length or habitat is important, understanding the microbial communi- sequences whose average quality score was less than 25 bp ty present in the environment also provides valuable were removed. In addition, sequences were further edited by information. To estimate the microbial community composi- primer sequence trimming utilizing pairwise alignment and tion of the specific environment, numerous tools have been sequencing error correction. We employed the EzTaxon employed. Among them, next-generation sequencing (NGS) extended database for sequence identification by searching technologies such as pyrosequencing have gained attention the highest similarity among the top five hits of the BLASTN for microbial community analysis as tools that can overcome algorithm (Chun et al., 2007). UCHIME software was also issues in the conventional methods, such as limitations of employed for chimeric sequences and non-16S rRNA gene collecting information, fastidious culturability of some sequences that were further removed (Edgar et al., 2011). An microbes, and laborious processes (Lee et al., 2012; Jeong et operational taxonomic unit (OTU) was obtained by CD-HIT al., 2013; Mardis, 2008). In fact, numerous studies analyzing software with a 97% sequence identity cutoff (Li and the microbial community compositions in traditional Korean Godzik, 2006). foods have been performed utilizing pyrosequencing (Jung Statistical analysis to determine the richness and diversity et al., 2011, 2012b, 2013a, b; Nam et al., 2012b; Jeong et al., of species of each Alaska pollock sikhae sample. We 2013; Lee et al., 2014; Nam et al., 2012a; Mardis, 2008). performed statistical analysis using the Chao1 richness Our previous research revealed that Lactobacillus sakei index and Shannon diversity index with MOTHUR software was dominant in the gajami-sikhae microbiota (Kim et al., (Schloss et al., 2009). We obtained the Chao1 richness index 2014). In this study, we conducted pyrosequencing analysis and Shannon diversity index to estimate the richness and of the bacterial community in Alaska pollock sikhae made diversity of species of the different Alaska pollock sikhae with a Kangwondo-style recipe. Alteration of the bacterial samples. Fast UniFrac distance utilizing CLcommunityTM community compositions at two different temperatures software (ChunLab, Inc., Seoul, Korea) was used for princi- consistently indicated that L. sakei became prominently pal coordinate analysis (PCoA) to determine the differences dominant among the varied bacterial community. This result in the richness and diversity of bacterial communities among suggested that L. sakei might have the capability to suppress sikhae samples (Hamady et al., 2010). Pyrosequecing reads other species in the bacterial community or utilize nutrients reported in this study have been deposited in the EMBL of sikhae better than other bacterial species present. Sequence Read Archive database under the project accession number PRJEB6648 (http://www.ebi.ac.uk/ena/data/view/ Materials and Methods PRJEB6648).

Sample preparation and fermentation of Alaska pollock Results sikhae. We prepared Alaska pollock sikhae following a Kangwondo-style recipe. Firstly, 5 kg of Alaska pollock Characteristics of the sequenced data for Alaska pollock fillet were completely washed with 3% salt water. The sikhae samples utilizing pyrosequencing analysis washed fish fillet was chopped and mixed with 1.05 kg of Alterations in the bacterial community compositions of cooked millet, 2.72 kg of shredded radish, 800 g of red Alaska pollock sikhae during the fermentation under differ- pepper powder, 10 g of green onion, 8 g of garlic, 5 g of ent temperatures (10°C and 20°C) were monitored by ginger, 1 g of salt, and 21 g of malt. We collected samples for high-throughput pyrosequencing of the partial 16S rRNA the bacterial community analysis and fermented the Alaska gene. The sequence length of the Alaska pollock sikhae pollock sikhae for further analysis under two different samples was distributed between 300 and 530 bp (Fig. 1). temperature conditions (10°C and 20°C). We collected The total reads for each Alaska pollock sample ranged from samples of the two different temperature conditions on days 6,138 to 20,072 (Table 1). The number of OTUs in five 3 and 6 of fermentation in the laboratory. different Alaska pollock sikhae samples was from 97 to 313. DNA extraction and amplification of samples from Alaska Although the control displayed a small number of valid pollock sikhae. Genomic DNAs from the collected Alaska reads (1,676), the number of OTUs was the second most pollock sikhae samples were extracted with a FastDNA among the samples (Table 1). SPIN extraction kit (MP Biomedicals, Santa Anna, CA). We amplified V1, V2, and V3 variable regions of 16S rRNA The alteration of the bacterial community compositions of using barcoded fusion primers as previously described Alaska pollock sikhae under different fermentation conditions (Chakravorty et al., 2007; Hur et al., 2011). The PCR for the The phylum-level bacterial composition analysis by DNA amplification was performed as previously described pyrosequencing showed that the control (the sample before (Kim et al., 2014). Additionally, PCR purification was done fermentation) consisted of six phyla (Firmicutes, Proteobac- using a QIAquick PCR Purification Kit (QIAGEN, Limburg, teria, Actinobacteria, Bacteroidetes, Fusobacteria, and Netherlands). For pyrosequencing, amplicon libraries were Deinococcus-Thermus) (Fig. 2a). However, the samples Alaska pollock sikhae microbiota 229

Table 1. Statistical parameters of the barcoded pyrosequencing data sets from Alaska pollock sikhae samples collected at different temperatures and fermentation times. Sample Total reads Valid reads OTUs Chao 1 Shannon

Control 6,138 1,676 292 415.68 4.82 3 day-10°C 20,072 14,822 313 478.15 2.52 6 day-10°C 10,310 7,587 98 127.08 1.95 3 day-20°C 7,989 4,847 148 201.71 2.83 6 day-20°C 12,211 9,048 97 126.00 1.78

resident in the Alaska pollock sikhae. As the fermentation progressed, the proportion of genera including lactic acid bacteria was significantly increased (Fig. 2b). The Alaska pollock sikhae sample on day 3 of fermentation at 20°C contained 87% Lactobacillus and 9% Weissella. A dramatic increase in the genera including lactic acid bacteria was also observed in the sample on day 6 of fermentation at 20°C (Fig. 2b). The amount of Lactobacillus genera (98%) became overwhelming, but the amount of Weissella genera was negligible (＜1%). Compared to the fermentation at 20°C, the genus-level composition of the bacterial community after 3 days of fermentation at 10°C was varied. The sample Fig. 1. Distribution of the sequence length (bp) of the pyrosequenc- collected after 3 days of fermentation at 10°C comprised ing data. Lactobacillus (59%), Weissella (12%), Psychrobacter (15%), Each line represents the distribution of the sequence length of the Pantoea (5%), Enterobacter (2%), Leuconostoc (1%), and sample collected from denoted fermentation conditions. The y-axis in- other genera whose proportions were less than 1%. The dicates the percentage of the counts per total counts of the assigned sample. proportions of Lactobacillus and Weissella were increased up to 76% and 22%, respectively, on day 6 of fermentation at 10°C (Fig. 2b). The dominance of the genera including collected after 3 days of fermentation at 10°C contained four Lactobacillus, Leuconostoc, and Weissella is consistent with phyla (Firmicutes, Proteobacteria, Actinobacteria, and the result observed in the previous study discussing the Bacteroidetes). In addition, the samples collected after 3 bacterial community analysis for the gajami-sikhae samples days of fermentation at 10°C showed a significant increase in (Kim et al., 2014). the phylum Firmicutes and decrease in the phylum Proteo- The species-level compositions of the bacterial communi- bacteria compared to the control (Fig. 2a). The increase in ties of the Alaska pollock sikhae samples showed a similar the Firmictues and the decrease in the Proteobacteria trend to the genus-level bacterial compositions. The heat appeared to continue until day 6 at 10°C. Similarly, the map analysis in Fig. 3 shows the five most abundant species increase in the Firmictues and the decrease in the Proteobac- whose percentages among the bacterial community composi- teria were also observed in the bacterial compositions of the tion were more than 5% for at least one of the individual samples fermented at 20°C (Fig. 2a). samples. While Lactobacillus sakei showed only 2% of the The changes in the genus-level bacterial composition bacterial community in the control sample, L. sakei became under two different fermentation conditions were in agree ment dominant regardless of the fermentation conditions (Fig. 5). with those of the phylum-level compositions. The control Weissella hellenica was also relatively abundant in the sample consisted of 21 genera including Enterobacter various samples (Fig. 5). Interestingly, Psychrobacter (27%), Leuconostoc (9%), Acinetobacter (8%), Chryseo- arcticus, a bacterial species that was firstly isolated from a bacterium (7%), Arthrobacter (6%), Pantoea (5%), Pseudo- Siberian permafrost core (Hinsa-Leasure et al., 2013), (14%) monas (4%), and Lactobacillus (4%), among others (Fig. was included in the sample after 3 days of fermentation at 2b). This result indicated that various bacterial genera were 10°C. In addition, the UniFrac clustering relationship with

Fig. 2. Bacterial community compositions of Alaska pollock sikhae samples collected from different temperatures and fermentation times. Panel (a) displays the bacterial community compositions from the sikhae samples at the phylum-level, and panel (b) shows the genus-level. 230 KIM et al. the five most abundant bacterial species demonstrated that tions under assigned temperature conditions (10°C and the control was distinct from the other samples (Fig. 3). 20°C) allowed Alaska pollock sikhae to develop a bacterial We also monitored the alteration of composition of the community preferable for food safety. bacterial species that might not be preferable for foods. A recent study suggested there exists a link between obesity Statistical comparison of microbiota between Alaska and Enterobacter, and consumption of foods containing high pollock sikhae quantities of Enterobacter might be unhealthy (Fei and Table 1 shows the summary of statistical parameters of Zhao, 2013). The control included high amounts of Entero- the pyrosequencing data from five Alaska pollock sikhae bacter species such as Enterobacter cowanii (19.3%), samples. The Chao1 richness revealed that species richness Enterobacter hormaechei (1.8%), Enterobacter cancerogenus between the samples was slightly varied (478.15‒126.00). (1.1%), Enterobacter kobei (0.7%), and Enterobacter amnige- The Shannon diversity index of the control was 4.82, nus (0.6%) (Fig. 4). In addition, the control showed that several whereas those of the other samples were between 1.78 and other bacterial species could be opportunistic pathogens such 2.83. We implemented principal coordinate analysis (PCoA) as Klebsiella oxytoca, Leclercia adecarboxylata, Acineto- of the bacterial communities from the five Alaska pollock bacter baumannii, Escherichia vulneris, Pseudomonas sikhae samples to estimate the relationship of the bacterial oryzihabitans, and Stenotrophomonas maltophilia. The total communities among the samples (Fig. 5). The PCoA analysis amount of species potentially adverse to health was more showed that the Alaska pollock sikhae samples at similar than 30% in the control sample (Fig. 4). Importantly, the temperatures could be clustered, while the control was composition of the unpreferred bacterial species was signifi- distinct from the others. The sikhae sample after 3 days of cantly reduced after fermentation regardless of the tempera- fermentation at 10°C was slightly different from the other ture condition. Specifically, the composition of the sikhae samples after fermentation. This result was in unpreferred bacterial species was reduced to under 0.2%. To accordance with the UniFrac cluster analysis of the five understand the reason for the decrease in those unpreferred most abundant species (Fig. 4). bacterial species, we measured the pH of the fermented food. The values of pH were decreased on day three of fermentation at 10°C and 20°C, but negligible change was observed at 0°C. Interestingly, the value of pH at 20°C was more than 6, indicating that the pH may not be the sole reason for the decrease of the unpreferred bacterial species. Combined with the result showing the dominance of lactic acid bacteria after fermentation, this result suggested that the fermenta-

Fig. 4. The reduction in composition of unpreferred bacterial species of Alaska pollock sikhae during fermentation. (a) Alteration in the relative abundances of unprefered microbes for food safety in the Alaska pollock sikhae samples during fermentation at different temperatures. (b) Alteration of pH of the Alaska pollock sikhae samples during the fermentation at 0°C, 10°C, and 20°C for 15 days.

Fig. 3. The heatmap analysis illustrating the relative amount of the five most abundant species in bacterial community compositions from the different Alaska pollock sikhae samples. The percentages of the five most abundant species among the bacterial community composition are more than 5% for at least one of the individual samples. The heat map gradient represents relative abundance of the species for the assigned samples. The value (percentage of Fig. 5. Relative abundance and clustering analysis for the Alaska the species among the bacterial community) was denoted on the top of pollock sikhae samples collected from different temperature white (0.00) and black (91.67) gradients. The relationship of bacterial and fermentation time using the PCoA of pyrosequencing communities was described by implementing UniFrac clustering data. above the heatmap. The UniFrac distance scale bar is depicted at the The dark grey circle clustered the bacterial communities of Alaska top-right corner of the figure. pollock sikhae samples at 10°C and the light grey circle at 20°C. Alaska pollock sikhae microbiota 231

Discussion L. sakei has also been found in Korean fermented food such as kimchi and jeotgal (Roh et al., 2010; Kim and Chun, In our previous study, we examined the bacterial community 2005; Jung et al., 2011). By 16S rRNA gene analysis, Kim in the gajami-sikhae samples utilizing pyrosequencing and Chun (2005) showed the bacterial community structure analysis. Although the gajami-sikhae samples were made by in kimchi purchased from five major manufacturers. Among different manufacturers, it was consistent that the lactic acid various lactic acid bacteria in five different kimchi products, bacteria were dominant in the microbial community of compositions of L. sakei were found in the range of gajami-sikhae. Importantly, the analysis of the gajami- 11.1‒38.5% (Kim and Chun, 2005). Another group sikhae samples showed that L. sakei was the predominant monitored the change in the microbial community during a species among the bacterial community isolated from the 29-day fermentation process (Jung et al., 2011). They gajami-sikhae samples (Kim et al., 2014). Consistently, L. provided genome recruitment analysis representing complete sakei was predominant in the bacterial community of the genomic sequences (Jung et al., 2011). L. sakei was one of fermented Alaska pollock sikhae. This is intriguing because the two most highly recruited genomes; the other was the control sample before fermentation comprised varied Leuconostoc mesenteroides, which was one of the predomi- bacterial species. This result indicated that the environment nant species among the kimchi microbiota in aprior study of Alaska pollock sikhae was favorable for the L. sakei (Kim and Chun, 2005). Roh and co-workers (2010) explored strain. Another important observation is that the many other the archeal and bacterial community compositions in the species (including some unpreferred bacterial species for various jeotgal microbiotas through pyrosequencing. The foods) predominant in the control sample substantially study elucidated that the Lactobacillus and Weissella genera decreased. This suggested a possible negative effect of the were predominant in the jeotgal bacterial community (Roh sikhae environment on these species that were reduced. et al., 2010). Among various jeotgal samples, L. sakei was Notably, L. sakei is able to grow and survive at low tempera- observed in the jeotgal samples originating from roe of ture (4°C) and high salt ranging from 3% to 9% NaCl pollock and tripe of pollock, suggesting the possible sources (Chaillou et al., 2005). This psychrotrophic and salt tolerable of the L. sakei strain of Alaska pollock sikhae (Roh et al., characteristic of L. sakei may be beneficial to its residence in 2010). Previous research regarding Japanese fermented the sikhae habitat. In addition, an acidic condition (under pH foods also showed that L. sakei was observed in the microbi- 4.6) created by L. sakei (and other lactic acid bacteria) could al community (Matsui et al., 2010; Matsui et al., 2008; play a crucial role of the decrease of unpreferred bacterial Koyanagi et al., 2013). Matsui and colleagues (2010) showed species. Interestingly, the dominance of L. sakei appeared to that L. sakei was the most abundant species in Ayu-narezushi, precede the formation of the acidic condition, indicating the a Japanese fermented seafood, utilizing the16S rRNA gene presence of other factors contributing to the dominance of L. clone library method. In their later research, they highlighted sakei (Fig. 4). that L. sakei was predominant in the bacterial community of L. sakei was firstly isolated from fermented sausage samma-narezushi, another Japanese fermented seafood (McLeod et al., 2013; Urso et al., 2006). It has been widely (Matsui et al., 2008). In another pyrosequencing analysis used as a starter in fermented meat products for biopreserva- regarding Kaburazushi, Japanese medieval sushi, L. sakei tion (Castellano et al., 2012). L. sakei is also known as a species grew rapidly (up to 79% of total microbiota) during bacteriocin producer (Chaillou et al., 2005; Mathiesen et al., fermentation (Koyanagi et al., 2013). These results suggest- 2005; Vaughan et al., 2003; Rawlinson et al., 2002). Interest- ed that L. sakei is versatile and dominant in various East ingly, the genome of L. sakei has cognate immunity genes Asian fermented foods, especially seafoods such as Alaska against the corresponding bacteriocin (Moretro et al., 2005). pollock sikhae. These cognate immunity genes are able to protect the L. sakei strain from the bacteriocin produced by closely-related Acknowledgments species (Moretro et al., 2005; Chaillou et al., 2005). The various L. sakei strains produce different kinds of bacterio- This work was funded by a grant from Korea Food Research cins such as sakacin K (CTC494), sakacin P (LB16 and Institute (project no. E0131901). We thank Dr. Bong-Soo Kim at ChunLab for reading of the manuscript and valuable comments. MF1053), sakacin A (Lb706), and lactocin S (L45) (Nyquist et al., 2011). Some strains produce more than one kind of References sakacin such as sakacin P, sakacin T, and sakacin X for the L. sakei 5 strain, and sakacin P and sakacin Q for the MF1053 Abbasiliasi, S., Tan, J. S., Ibrahim, T. A., Ramanan, R. N., Vakhshiteh, strain (Vaughan et al., 2003; Nyquist et al., 2011). In addition, F. et al. (2012) Isolation of Pediococcus acidilactici Kp10 with some L. sakei strains (e.g. L. sakei 23K) have remnants of ability to secrete bacteriocin-like inhibitory substance from milk the bacteriocin gene with the cognate immunity genes products for applications in food industry. BMC Microbiol., 12, (Chaillou et al., 2005). Production of bacteriocin and the 260. Castellano, P., Aristoy, M. C., Sentandreu, M. A., Vignolo, G., and capability to protect itself from bacteriocin produced by Toldra, F. (2012) Lactobacillus sakei CRL1862 improves safety other competitors can be hugely beneficial for the L. sakei and protein hydrolysis in meat systems. J. Appl. Microbiol., 113, strains to survive the interactions between L. sakei and other 1407‒ 1416. competitors of the various habitats. This versatility of L. Cha, Y. J., Jeong, E. J., Kim, H., Cho, W. J., and Nam, G. J. (2004) Biological activity in traditional Alaska pollack sikhae during low sakei may be advantageous for inhabiting the Alaska pollock temperature fermentation. Biofactors, 22, 319‒ 321. sikhae ecosystem. The bacteriocin from L. sakei could also Chaillou, S., Champomier-Verges, M. C., Cornet, M., Crutz-Le Coq, be the reason for reduction in the unwanted bacterial species A. M., Dudez, A. M. et al. (2005) The complete genome sequence that were predominant in the control sample. of the meat-borne lactic acid bacterium Lactobacillus sakei 23K. 232 KIM et al.

Nat. Biotechnol., 23, 1527‒ 1533. pyrosequencing and 1H nuclear magnetic resonance spectrosco- Chakravorty, S., Helb, D., Burday, M., Connell, N., and Alland, D. py. Appl. Environ. Microbiol., 78, 5983‒ 5993. (2007) A detailed analysis of 16S ribosomal RNA gene segments Lee, S. H., Jung, J. Y., and Jeon, C. O. (2014) Effects of temperature on for the diagnosis of pathogenic bacteria. J. Microbiol. Methods, microbial succession and metabolite change during saeu-jeot 69, 330‒ 339. fermentation. Food Microbiol., 38, 16‒ 25. Chang, J. H., Shim, Y. Y., Cha, S. K., and Chee, K. M. (2010) Probi- Li, W. and Godzik, A. (2006) Cd-hit: A fast program for clustering and otic characteristics of lactic acid bacteria isolated from kimchi. J. comparing large sets of protein or nucleotide sequences. Bioinfor- Appl. Microbiol., 109, 220‒ 230. matics, 22, 1658‒ 1659. Chun, J., Lee, J. H., Jung, Y., Kim, M., Kim, S. et al. (2007) EzTaxon: Mardis, E. R. (2008) Next-generation DNA sequencing methods. A web-based tool for the identification of prokaryotes based on Annu. Rev. Genomics Hum. Genet., 9, 387‒ 402. 16S ribosomal RNA gene sequences. Int. J. Syst. Evol. Microbi- Mathiesen, G., Huehne, K., Kroeckel, L., Axelsson, L., and Eijsink, V. ol., 57, 2259‒ 2261. G. (2005) Characterization of a new bacteriocin operon in sakacin Edgar, R. C., Haas, B. J., Clemente, J. C., Quince, C., and Knight, R. P-producing Lactobacillus sakei, showing strong translational (2011) UCHIME improves sensitivity and speed of chimera coupling between the bacteriocin and immunity genes. Appl. detection. Bioinformatics, 27, 2194‒ 2200. Environ. Microbiol., 71, 3565‒ 3574. Fei, N. and Zhao, L. (2013) An opportunistic pathogen isolated from Matsui, H., Saka, E., Isobe, Y., and Narita, M. (2010) Comparison of the gut of an obese human causes obesity in germfree mice. ISME the bacterial community structures of Ayu-narezushi produced by J., 7, 880‒ 884. two different manufacturers. Biocontrol. Sci., 15, 63‒ 68. Guan, L., Cho, K. H., and Lee, J. H. (2011) Analysis of the cultivable Matsui, H., Tsuchiya, R., Isobe, Y., Maeda, H., and Narita, M. (2008) bacterial community in jeotgal, a Korean salted and fermented Diversity of the bacterial community found in samma-narezushi seafood, and identification of its dominant bacteria. Food Micro- (saury narezushi) revealed by the 16S rRNA gene clone library. biol., 28, 101‒ 113. Biocontrol. Sci., 13, 97‒ 102. Hamady, M., Lozupone, C., and Knight, R. (2010) Fast UniFrac: Facil- Mcleod, A., Brede, D. A., Rud, I., and Axelsson, L. (2013) Genome itating high-throughput phylogenetic analyses of microbial sequence of Lactobacillus sakei subsp. sakei LS25, a commercial communities including analysis of pyrosequencing and Phylo- starter culture strain for fermented sausage. Genome Announc., 1, Chip data. ISME J., 4, 17‒ 27. e00475-13. Hinsa-Leasure, S. M., Koid, C., Tiedje, J. M., and Schultzhaus, J. N. Moretro, T., Naterstad, K., Wang, E., Aasen, I. M., Chaillou, S. et al. (2013) Biofilm formation by Psychrobacter arcticus and the role (2005) Sakacin P non-producing Lactobacillus sakei strains of a large adhesin in attachment to surfaces. Appl. Environ. contain homologues of the sakacin P gene cluster. Res. Microbi- Microbiol., 79, 3967‒ 3973. ol., 156, 949‒ 960. Hur, M., Kim, Y., Song, H. R., Kim, J. M., Choi, Y. I. et al. (2011) Nam, Y. D., Lee, S. Y., and Lim, S. I. (2012a) Microbial community Effect of genetically modified poplars on soil microbial commu- analysis of Korean soybean pastes by next-generation sequenc- nities during the phytoremediation of waste mine tailings. Appl. ing. Int. J. Food Microbiol., 155, 36‒ 42. Environ. Microbiol., 77, 7611‒ 7619. Nam, Y. D., Park, S. L., and Lim, S. I. (2012b) Microbial composition Jeong, S. H., Jung, J. Y., Lee, S. H., Jin, H. M., and Jeon, C. O. (2013) of the Korean traditional food “kochujang” analyzed by a massive Microbial succession and metabolite changes during fermenta- sequencing technique. J. Food Sci., 77, M250‒ 256. tion of dongchimi, traditional Korean watery kimchi. Int. J. Food Nyquist, O. L., Mcleod, A., Brede, D. A., Snipen, L., Aakra, A. et al. Microbiol., 164, 46‒ 53. (2011) Comparative genomics of Lactobacillus sakei with Jung, J., Choi, S., Jeon, C. O., and Park, W. (2013a) Pyrosequencing- emphasis on strains from meat. Mol. Genet. Genomics, 285, 297‒ based analysis of the bacterial community in Korean traditional 311. seafood, ojingeo jeotgal. J. Microbiol. Biotechnol., 23, 1428‒ Park, J. E., Oh, S. H., and Cha, Y. S. (2013) Lactobacillus plantarum 1433. LG42 isolated from gajami sik-hae inhibits adipogenesis in Jung, J., Chun, J., and Park, W. (2012a) Genome sequence of extracel- 3T3-L1 adipocyte. Biomed. Res. Int., 2013, 460927. lular-protease-producing Alishewanella jeotgali isolated from Park, J. E., Oh, S. H., and Cha, Y. S. (2014) Lactobacillus plantarum traditional Korean fermented seafood. J. Bacteriol., 194, 2097. LG42 isolated from gajami sik-hae decreases body and fat pad Jung, J. Y., Lee, S. H., Kim, J. M., Park, M. S., Bae, J. W. et al. (2011) weights in diet-induced obese mice. J. Appl. Microbiol., 116, Metagenomic analysis of kimchi, a traditional Korean fermented 145‒ 156. food. Appl. Environ. Microbiol., 77, 2264‒ 2274. Rawlinson, E. L., Nes, I. F., and Skaugen, M. (2002) LasX, a transcrip- Jung, J. Y., Lee, S. H., Lee, H. J., and Jeon, C. O. (2013b) Microbial tional regulator of the lactocin S biosynthetic genes in Lactoba- succession and metabolite changes during fermentation of saeu- cillus sakei L45, acts both as an activator and a repressor. Biochi- jeot: Traditional Korean salted seafood. Food Microbiol., 34, mie, 84, 559‒ 567. 360‒ 368. Rhee, S. J., Lee, J. E., and Lee, C. H. (2011) Importance of lactic acid Jung, M. J., Nam, Y. D., Roh, S. W., and Bae, J. W. (2012b) Unexpect- bacteria in Asian fermented foods. Microb. Cell Fact., 10 Suppl. ed convergence of fungal and bacterial communities during 1, S5. fermentation of traditional Korean alcoholic beverages inoculat- Roh, S. W., Kim, K. H., Nam, Y. D., Chang, H. W., Park, E. J. et al. ed with various natural starters. Food Microbiol., 30, 112‒ 123. (2010) Investigation of archaeal and bacterial diversity in Kim, M. and Chun, J. (2005) Bacterial community structure in kimchi, fermented seafood using barcoded pyrosequencing. ISME J., 4, a Korean fermented vegetable food, as revealed by 16S rRNA 1‒ 16. gene analysis. Int. J. Food Microbiol., 103, 91‒ 96. Schloss, P. D., Westcott, S. L., Ryabin, T., Hall, J. R., Hartmann, M. et Kim, H. J., Kim, M. J., Turner, T. L., Kim, B. S., Song, K. M. et al. al. (2009) Introducing mothur: open-source, platform-indepen- (2014) Pyrosequencing analysis of microbiota reveals that lactic dent, community-supported software for describing and compar- acid bacteria are dominant in Korean flat fish fermented food, ing microbial communities. Appl. Environ. Microbiol., 75, 7537‒ gajami-sikhae. Biosci. Biotechnol. Biochem., 78, 1611‒1618. 7541. Kim, M. S., Roh, S. W., Nam, Y. D., Chang, H. W., Kim, K. H. et al. Shin, M. S., Han, S. K., Ryu, J. S., Kim, K. S., and Lee, W. K. (2008) (2009) Alishewanella jeotgali sp. nov., isolated from traditional Isolation and partial characterization of a bacteriocin produced by fermented food, and emended description of the genus Alishe- Pediococcus pentosaceus K23-2 isolated from Kimchi. J. Appl. wanella. Int. J. Syst. Evol. Microbiol., 59, 2313‒ 2316. Microbiol., 105, 331‒ 339. Koyanagi, T., Nakagawa, A., Kiyohara, M., Matsui, H., Yamamoto, K. Shin, N. R., Kim, M. S., Jung, M. J., Roh, S. W., Nam, Y. D. et al. et al. (2013) Pyrosequencing analysis of microbiota in Kabura- (2011) Leucobacter celer sp. nov., isolated from Korean ferment- zushi, a traditional medieval sushi in Japan. Biosci. Biotechnol. ed seafood. Int. J. Syst. Evol. Microbiol., 61, 2353‒ 2357. Biochem., 77, 2125‒ 2130. Urso, R., Rantsiou, K., Cantoni, C., Comi, G., and Cocolin, L. (2006) Lee, H. J., Jung, J. Y., Oh, Y. K., Lee, S. S., Madsen, E. L. et al. (2012) Technological characterization of a bacteriocin-producing Lacto- Comparative survey of rumen microbial communities and metab- bacillus sakei and its use in fermented sausages production. Int. J. olites across one caprine and three bovine groups, using bar-coded Food Microbiol., 110, 232‒ 239. Alaska pollock sikhae microbiota 233

Vaughan, A., Eijsink, V. G., and Van Sinderen, D. (2003) Functional Wu, R., Wang, L., Wang, J., Li, H., Menghe, B. et al. (2009) Isolation characterization of a composite bacteriocin locus from malt and preliminary probiotic selection of lactobacilli from koumiss isolate Lactobacillus sakei 5. Appl. Environ. Microbiol., 69, in Inner Mongolia. J. Basic Microbiol., 49, 318‒ 326. 7194‒ 7203.