Diversity of lactic acid bacteria in vegetable-based and meat-based fermented foods produced in the central region of Vietnam

The diversity of lactic acid bacteria (LAB) in naturally fermented foods produced in Hue, a city in the central region of Vietnam, was examined. From local markets, a total of 25 samples of three vegetable-based fermented products, specifically dua gia (bean sprouts), dua cai (cabbage), and mang chua (bamboo shoots), and two meat-based fermented products, specifically nem chua (uncooked pork) and tre (cooked pork) were obtained. The LAB diversity was assessed by quantitative real-time polymerase chain reaction (PCR) and qualitative denaturing gradient gel electrophoresis. Lactic and acetic acid contents were greater in meat-based products than in vegetable-based products. Major LAB species found in vegetable-based products (Lactobacillus plantarum, Lactobacillus fermentum, and Lactobacillus helveticus) were different from those identified in meat-based products (Pediococcus pentosaceus, Weissella cibaria, and Lactococcus lactis). The total bacterial population was approximately 109−10 copies/g regardless of the food item, with the proportion of Lactobacillus spp. determined to be from 78% (dua cai) to 94% (nem chua).


Introduction
Most East Asian fermented foods are non-dairy products including various raw materials such as cereals, soybeans, fruits, and vegetables, as well as fish and other marine products. These foods are produced largely in households or on a small scale using local indigenous microbiota; this is different from western countries in which fermented foods are industrially produced on a large scale using selected starter cultures [1,2].
Vietnam. Because nem chua (uncooked pork) and tre (cooked pork) are different regarding the use of heat processing before fermentation, these two food items, consumed in the central region, are worth comparing. Likewise, most findings regarding LAB microbiota were derived from de Man, Rogosa, Sharpe (MRS) isolates; hence, more diversified microbiota could be observed if analysis was independent of plate-culture. In this study, three vegetable-based and two meat-based fermented foods were obtained from local markets and households in Hue, a city in the central region of Vietnam, and LAB diversity was assessed by quantitative real-time polymerase chain reaction (PCR) and qualitative denaturing gradient gel electrophoresis (DGGE). Differences among retailers and households were also examined.

Samples of fermented foods
Twenty-five samples with five products each of dua gia (bean sprouts), dua cai (cabbage), mang chua (bamboo shoot), nem chua (uncooked pork), and tre (cooked pork) were purchased from 12 shops at six local markets in Hue.

Chemical analyses
Dry matter content was determined by drying the materials in an oven at 60 °C for 48 h. The lactic acid, short chain fatty acid, and alcohol contents were determined from water extracts using an ion-exclusion polymeric high-performance liquid chromatography method with refractive index detection [5]. A portion of the water extracts was passed through a 0.20 μm filter, and 10 μL of the filtrate was injected into an ICSep COREGEL-87H column (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) containing a cation exchange polymer in the hydrogen ion form. The mobile phase was 0.004 mol/L sulphuric acid, and the flow rate was 0.6 ml/min at 60 °C.

DGGE
Samples were added to a 10 volume of sterilized phosphate buffered saline (pH 7.4). Extraction involved shaking the samples vigorously for 10 min and obtaining microbial pellets by centrifugation at 8000× g for 15 min. Bacterial DNA was purified using a commercial kit (Blood and Tissue Kit; Qiagen, Germantown, MD, USA), according to the manufacturer's instructions. PCR was employed to amplify a variable (V3) region of the bacterial 16S rRNA gene using primers and PCR protocols as previously described [6].
The PCR products were separated according to their sequences using a DCode Universal Mutation Detection System (Bio-Rad Ltd., Tokyo, Japan). The samples were applied directly to 100 g/L (w/v) polyacrylamide gels that had a denaturing gradient ranging from 25% to 50%; these gradients had been prepared using 7 mol/L urea and 400 mL/L formamide respectively, as 100% denaturants. Electrophoresis was performed at a constant voltage of 150 V for 12 h at 60 °C.
After electrophoresis, the gels were stained with SYBR Green (Cambrex Bio Science Inc., Rockland, ME, USA) and photographed under UV illumination. Select bands were excised from DGGE gels, and extracted DNA was amplified by PCR, followed by cloning into the pTAC-1 vector.
The resulting plasmids were transformed into Escherichia coli strain DH5α competent cells (DynaExpress TA cloning kit; BioDynamics Laboratory Inc., Tokyo, Japan). The DNA base sequences were analysed using an ABI PRISM® 3130 sequencer (Applied Biosystems Inc., Foster City, CA, USA) and Basic Local Alignment Search Tool (BLAST) searches were performed against the GenBank database. The rRNA gene sequences determined in this study have been deposited in the DNA Data Bank of Japan (Accession Numbers LC205657-LC205687).

Real-time PCR assessment of bacterial populations
Amplification and detection of bacterial DNA were performed with KAPA SYBR® FAST qPCR Master Mix (Kapa Biosystems, USA) and a Mini Opticon real time PCR system (Bio-Rad Laboratories Inc., Tokyo, Japan). Primers targeting the 16S rRNA genes of the total bacterial population [7] and those targeting Lactobacillus spp. [8] were employed. Plasmid DNA, prepared with 16S rRNA genes of L. brevis (JCM 1059T), was used as a standard for qPCR analyses of both total bacteria and Lactobacillus spp. A three-step protocol was performed with an initial denaturation at 95 °C for 5 min followed by 35 cycles of denaturation at 95 °C for 15 s, annealing at 60 °C (total bacteria) or 58 °C (Lactobacillus spp.) for 20 s, and extension at 72 °C for 30 s. After the final qPCR cycle, melting curve analysis of the qPCR products was performed by heating from 60 °C to 95 °C, and the fluorescence signal was measured every 0.3 °C in the last 10 s to verify qPCR products.

Statistical analyses
Quantitative data were subjected to a one-way analysis of variance and a Tukey's multiple range test was used to examine the differences between food items (JMP ver. 11; SAS Institute, Tokyo, Japan). Qualitative DGGE band profiles were converted to a list of binary numbers using an image analysis system (Image J; NIH, Bethesda, MD, USA). A total of 39 bands were used in principal coordinate (PCoA) analysis on the Bray-Curtis similarities matrix using Primer version 7 with Permanova+ add-on software (Primer-E, Plymouth Marine Laboratory, Plymouth, UK).

Results
All samples were evaluated as being acceptable based on acidity and fermentation product composition (Table 1). Dry matter contents of vegetable-based fermented products were approximately 7 g/kg and those of meat-based products were 34-50 g/kg. Dua gia (bean sprouts), dua cai (cabbage), and nem chua (uncooked pork) had more lactic acid than acetic acid, whereas mang chua (bamboo shoot) and tre (cooked pork) showed the opposite trend. Ethanol was detected in one sample of dua gia (bean sprouts) at 3.9 g/kg but was not detected in all other samples. Mang chua had the lowest pH value among the five food items.
PCoA indicated that, although meat-based products were grouped together (except for one tre sample), vegetable-based products were classified into four small groups and additional separate clusters. Samples of mang chua (bamboo shoots) formed an independent group, whereas those of dua cai (cabbage) were seen as scattered plots. In addition to a clear difference between vegetable-based and meat-based products, significant microbiota diversity between households was shown for dua cai (cabbage).

Discussion
Although selected starters were not used in any fermented foods, LAB were predominant, and populations were identified levels at 10 9 copies/g in both vegetable-based and meat-based products. Similar levels of LAB (10 8-9 cfu/g) were observed for Vietnamese fermented foods, regardless of the region, using MRS plates [3,4]. Concentrations of dry matter and organic acids in nem chua (uncooked pork) were also similar to those reported by Yoshimura et al. [9]. However, the pH values of nem chua (uncooked pork) in this study were relatively high compared to those of other reports, wherein the pH value was found to decrease to as low as 3.8-4.1 after three to four days of fermentation [9,10]. The pH and organic acid content between nem chua (uncooked pork) and tre (cooked pork) were similar, indicating that the use of cooked pork did not greatly influence the fermentation process.
Tran et al. demonstrated that, although L. plantarum was predominant in nem chua (uncooked pork) produced in the north and south of Vietnam, P. pentosaceus was mainly isolated in nem chua (uncooked pork) from the central region [4]. Our study using products from the central region of Vietnam also indicated that P. pentosaceus was frequently found in nem chua (uncooked pork).
L. lactis was steadily detected in nem chua (uncooked pork) and tre (cooked pork) in this study. Although Nguyen et al. found L. lactis in nem chua (uncooked pork), the proportion was as small as 1/119 isolates [11]. Because MRS medium enriches for Lactobacilli growth, Lactococcus spp. could be underrepresented using plate-culture compared to the use of culture-independent analysis. Indeed, Nguyen et al. detected L. lactis in nem chua (uncooked pork) by DGGE despite the fact that no L. lactis was identified in MRS isolates [3].
According to Nguyen et al., who examined 21 samples and 881 isolates by MALDI-TOF-MS and pheS gene sequence analysis [11], LAB including L. fermentum (56%), L. pentosus (24%), and L. plantarum (17%) were prevalent in vegetable-based products. Although L. fermentum and L. plantarum were also found, no L. pentosus was observed in this study, probably because L. plantarum, L. paraplantarum, and L. pentosus are difficult to discriminate by 16S rRNA gene analysis [12]. Meanwhile, our finding that L. helveticus was frequently identified in vegetable-based products was an interesting addition to previous knowledge. Likewise, L. acetotolerans and L. pontis have never been detected in Vietnamese fermented foods. L. acetotolerans and L. pontis are regarded as inhabitants of fermented foods with high acetic acid content, such as sourdough [13]. Because mang chua (bamboo shoot) had the highest acetic acid content among vegetable-based products, its detection in two samples was not unexpected.
Although clear differences were seen in microbiota between vegetable-based and meat-based products, it cannot simply be ascribed to the difference of the ingredients; microbiota involved in fermentation might have been affected by differences in dry matter concentrations. Although Tran et al. and Nguyen et al. found that L. plantarum was a predominant LAB species in nem chua (uncooked pork) [4,11], no L. plantarum was identified in this product. Likewise, although Nguyen et al. isolated higher numbers of L. farciminis compared to that of L. plantarum from nem chua (uncooked pork) [11], L. farciminis was not detected in any samples in this study. Meanwhile, the finding of Nguyen et al. that L. fermentum, a prevalent LAB in vegetable-based products [11], was not identified in nem chua was in agreement with our results.
Our results that W. cibaria and W. paramesenteroides were often found in tre (cooked pork) were also of interest. Although PCoA indicated a similar group for nem chua (uncooked pork) and tre (cooked pork), bands for W. cibaria and W. paramesenteroides were more distinctive in tre (cooked pork). Nguyen et al. also detected W. cibaria and W. paramesenteroides in nem chua (uncooked pork), but the proportions were 2/119 and 1/119, respectively [11]. Because Weissella spp. are widely observed in fermented plant and animal foods [14], the distinct detection of Weissella spp. in tre (cooked pork) is difficult to explain. However, individuals in the central regions might have the opportunity to take in Weissella spp. through tre (cooked pork) consumption. The primer set for Lactobacillus spp. quantification by real-time PCR is able to detect Weissella spp., Leuconostoc spp., and Pediococcus spp. in addition to Lactobacillus spp.; hence, most LAB species detected by DGGE should have been appropriately quantified.
The presence of Acinetobacter sp. and Staphylococcus sp. in nem chua (uncooked pork) was also an interesting addition to current knowledge. Because the majority of previous studies have focused on the potential benefits of LAB, information on microbiota based on food hygiene is lacking. In addition to the screening and understanding of beneficial LAB species, research with the goal of promoting the quality of naturally fermented foods is required.

Conclusion
Microbiota analysis combining DGGE and real-time PCR was performed on vegetable-based and meat-based fermented foods produced in Hue, a city in the central part of Vietnam. Major LAB species found in dua gia (bean sprouts), dua cai (cabbage), and mang chua (bamboo shoots) were L. plantarum, L. fermentum, and L. helveticus, whereas those in nem chua (uncooked pork) and tre (cooked pork) were P. pentosaceus, W. cibaria, and L. lactis. Because the association between L. helveticus, W. cibaria, and L. lactis and these food products was newly depicted in this study, our results are a good addition to the current knowledge of fermented food products. Several indigenous LAB species could be used as starters to assure the hygienic quality of nem chua (uncooked pork) fermentation, which is regarded as unstable and occasionally disrupted by the development of undesirable microorganisms.