Application of Lactic Acid Bacteria for Enhanced Food Safety of Cambodian Fermented Small Fish (Pha-ork Kontrey)

Abstract

“Pha-ork Kontrey” is a Cambodian traditional fermented small fish and is characterized by a low salt concentration and short fermentation period. Suitable starter culture strains from different fermented fish were isolated in an effort to control the growth of contaminating bacteria in the food. Five log CFU/g of Escherichia coli or Staphylococcus aureus was spiked into Pha-ork samples before fermentation. Cultures of isolated Lactobacillus plantarum strains (8 to 9 log CFU/g) were spiked to the samples at the same time. Growth of the inoculated L. plantarum showed rapid increases in lactic acid concentration. The number of spiked E. coli or S. aureus was significantly reduced in 48 h at room temperature (28–32 °C) in the presence of starter culture (9 log CFU/g), whereas 120 h fermentation was required without it. The use of starter culture did not significantly change the quality of Pha-ork Kontrey in term of appearance and color; however, for odor, taste, and overall, the use of starter culture at 9 log CFU/g was significantly better than other treatments. To suppress the growth of spiked E. coli and/or S. aureus into Pha-ork Kontrey, 3-day fermentation with isolated L. plantarum (9 log CFU/g) is recommended.

Introduction

Fermentation not only extends the shelf-life of fish but also enhances the taste, aroma, texture, nutritional value and other favorable properties of foods. Some fermented foods are reported to have health-promoting activities (Parvz et al., 2006; Borresen et al., 2012). The acid fermentation process is performed mainly by lactic acid bacteria (LAB). LAB converts carbohydrates to organic acids (mainly lactic acid) under anaerobic conditions, resulting in a slightly sour food. In the case of Sikhae (a Korean fermented fish), Leuconostoc (L.) mesenteroides and Lactobacillus (L.) plantarum were identified as major acid-forming bacteria. They contribute to generate an acceptable flavor and taste during fermentation of the product (Rhee et al., 2011). Pramuan et al. (2010) reported the isolation of L. plantarum IFRPD P15 and Lactobacillus reuteri IFRPD P17 from Thai naturally fermented fish, Plaa-som. These bacteria were used as a mixed starter culture to suppress and eliminate E. coli and other pathogenic bacteria. An L. plantarum strain and Lactobacillus pentosus strain isolated from Malaysian fermented fish, Pekasam, showed broad antimicrobial activity towards pathogenic bacteria as well as potential probiotic properties (Muryany et al., 2017). An Indonesian study revealed that from sixty strains isolated from an indigenous fermented fish, Chao, two strains of L. plantarum and Pediococcus acidilactici exhibited proteolytic activities (Matti et al., 2019).

Pha-ork Kontrey is a traditional fermented fish that has long been consumed in Cambodia. Pha-ork Kontrey is popularly served with a selection of raw vegetables such as cucumber, eggplant, or winged bean. Its natural lactic acid fermentation proceeds by mixing degutted small fish with a low concentration of salt (approx. 4–5%) for one day, and is then marinated with sugar, ground roasted rice and ground galangal (optional). The mixture is kept for an additional three to five days, depending on the temperature and the taste and flavor, before use as a ready-to-eat food. Pha-ork Kontrey is processed similarly to Plaa-som, a Thai fermented fish, and Pa-som, a Laos fermented fish (Marui et al., 2014). Because these foods are low-salt or lightly salted fermented foods and undergo spontaneous fermentation, bacterial contaminants from the raw materials or unhygienic practices can easily survive or grow during the fermentation process. In traditional fermentation, naturally contaminating bacteria from the environment, which attach to the surface of the equipment, act as the starter of fermentation. In some cases, a small amount of previously produced fermented product is used as a starter (Sven-Olof, 2008). Cambodian fermented fish products are produced in small and medium processing units. Further, food hygiene awareness by producers remains limited, and standards for equipment and processing facilities is lacking. There is a lack of potable water for washing and contamination with pathogenic bacteria during natural fermentation can occur; thus, the fermented fish products can pose a health hazard. Common pathogenic bacteria found in the early stage of low-salt fish fermentation are E. coli and Staphylococcus spp. (Saithong et al., 2010; Marui et al., 2014). Forty-five percent of fermented meat and fish in retail foods purchased in Thailand were contaminated with E. coli (Ananchaipattana et al., 2012). Further, E. coli was found in 10% of lightly salted fermented vegetables purchased from Phnom Penh, Cambodia (Chrun et al., 2017). Gadaga et al. (2004) reported that Bacillus cereus, E. coli and S. aureus were the most commonly encountered pathogens found in African fermented foods.

Selecting suitable strains for starter cultures to shorten the fermentation process and to address hygiene issues has been studied (Hurtado et al., 2010; Ying et al., 2016). LAB have been used for this purpose, since they are recognized as safe based on their long history of application in food and beverage fermentation (Caplice and Fitzgerald, 1999; Ray, 1992). The starter cultures contain much higher levels of suitable bacteria for fermentation compared to unsuitable bacteria at the initial stage of fermentation. Successive fermentation is achieved by the dominant growth of LAB, which suppresses the growth of undesirable spoilage or pathogenic bacteria by a rapid reduction in pH. In addition, certain LAB produce bacteriocins and other compounds, e.g., ethanol, which can suppress the growth of unsuitable contaminating bacteria. Hence, the use of such a starter culture is thought to enhance the microbial safety of food.

The objective of this study was to evaluate the growth suppressing activity of a selected L. plantarum starter culture against spiked E. coli or S. aureus (surrogate pathogenic bacteria) during the Pha-ork Kontrey fermentation process. The quality of the laboratory-prepared Pha-ork Kontrey using a starter culture was also examined.

Materials and Methods

Isolation of E. coli and S. aureus from spoiled fish    Twenty-nine small fish (Coilia lindmani) samples were collected from a local fish market in Phnom Penh, Cambodia. They were kept for 16 to 20 h at 28 to 30 °C, and then E. coli and S. aureus strains were isolated by the methods reported by Chrun et al. (2017). Eosin methylene blue (EMB) agar or egg-yolk mannitol salt (EYMS) agar (Oxoid Co., Ltd., UK) was used to isolate E. coli or S. aureus, respectively. Bacterial isolates were identified based on colony characterization on agar plates, microscopic morphology and biochemical characteristics. API 20E (bioMérieux, Marcy-l'Étoile, France) or API Staph (bioMérieux) was used to confirm E. coli or S. aureus, respectively. Strains exhibiting the same morphological and biochemical characteristics as described in “Bergey's Manual of Determinative Bacteriology 9th Edition” (Hensyl, 2000) were considered to be E. coli or S. aureus. We identified E. coli and S. aureus based mainly on the following keys characteristics: E. coli is a Gram-negative short rod. It is oxidase reaction negative and catalase reaction positive; indole production and methyl red reaction positive; Voges-Proskauer reaction, urea hydrolysis, arginine dehydrogenase production and use of citrate negative. It ferments D-glucose, D-mannitol and L-arabinose but not inositol and D-amygdalin. S. aureus is a catalase positive Gram-positive cocci. It is susceptible to lysis by lysostaphin but not by lysozyme. It produces acid from D-glucose, D-Fructose, D-mannose, maltose, D-trehalose and sucrose but does not ferment xylitol and raffinose. Other biochemical characteristics were used for identification by accessing the APIWEB online software (bioMérieux). A mixture of the isolated four E. coli (E1 to 4) or four S. aureus (S1 to 4) was used for the following experiments.

Preparation of E. coli and S. aureus inoculants    For each of the 4 isolated E. coli (E1 to 4) or S. aureus (S1 to 4) strains, a loop of frozen culture was transferred to a test tube containing 10 mL of tryptic soy broth (TSB; Oxoid Co., Ltd.) and incubated at 37 °C for 24 h. Bacterial cells were harvested by centrifugation (4 000 rpm, 10 min., 20 °C), and then diluted with sterile 0.85% saline water. Saline suspensions of the four E. coli or S. aureus strains were adjusted to an absorbance of 0.1 at 600 nm or equilibrated to 0.5 McFarland Standard (8 log CFU/mL) (Cappuccino and Welsh, 2019) using a spectrophotometer (Ultrospec 3330 pro; Amersham Biosciences, Ltd., Buckinghamshire, UK). Afterwards, the volume of the mixture of each strain suspension was calculated for inoculation to a final concentrations of 5 log CFU/g in the Pha-ork Kontrey.

Starter culture preparation    Twenty-four LAB strains were isolated and identified from twenty-nine fermented fish samples, i.e., the commercially purchased Pha-ork Kontrey, according to the method reported by Inatsu et al. (2005). Isolated strains were identified by their microscopic morphology and biochemical characteristics using API 50 (bioMérieux). The obtained results were compared with the bacterial characteristics described in “Bergey's Manual of Determinative Bacteriology 9th Edition” (Hensyl, 2000). We identified L. plantarum based mainly on the following key characteristics: L. plantarum is a Gram-positive, non-spore forming rod. It is oxidase and catalase reaction negative. It ferments d-glucose, d-fructose, d-mannose, N-acetyl glucosamine and maltose but not l-xylose, dulcitol, inositol, inulin and xylitol. Other biochemical characteristics were used for identification by accessing the APIWEB online software (bioMérieux).

The culture pH of each of the isolated LAB was measured after 48-h incubation at 30 °C in de Man, Rogosa and Sharpe (MRS) broth (Difco Laboratory Inc., Franklin Lakes, NJ, USA). The pH was recorded with a digital pH meter (LAQUA F-74; Horiba Co., Ltd., Tokyo). The growth suppression activity of E. coli and S. aureus was measured using the microdilution method. We identified and isolated the strain that exhibited the highest activity against E. coli and S. aureus, which we named “L. plantarum D11” and L. plantarum S22, respectively.

The two L. plantarum strains (S22 and D11, ratio 1:1 in bacterial concentration) were selected as a starter culture for Pha-ork Kontrey fermentation, as they showed a rapid decrease in pH during MRS broth cultivation and inhibited the growth of E. coli and S. aureus on a laboratory scale. These two LAB strains were grown in MRS broth at 35 °C for 24 h. Each of the subcultures was harvested by centrifugation (4 000 rpm, 10 min), then washed with sterile 0.85% saline water twice. The mixture (ratio 1:1 in bacterial concentration) of washed LAB cells was used as the starter culture for the Pha-ork Kontrey fermentation experiments.

Pha-ork Kontrey preparation    All ingredients for Pha-ork Kontrey including the small fishes (Coilia lindmani) were purchased from local markets in Phnom Penh. Fishes used for each of the experimental replicates were purchased from different markets.

Sixty grams of fresh fish for each sample was dispended into sterilized glass jars after washing with tap water, and then mixed with 4% (w/w) salt, 2% (w/w) sugar and 20% (w/w) roasted rice flour. The above starter culture (8 or 9 log CFU/g) and 5 log CFU/g of surrogate pathogenic bacteria (E. coli or S. aureus) were inoculated into this mixture. The initial addition of starter culture was omitted for the control samples. The initial addition of E. coli or S. aureus was omitted from the samples used for sensory evaluation. The jars were kept at 35 °C for 120 h (5 days) for fermentation to proceed. Total acidity, pH, and bacterial numbers were determined every day for 120 h. Total acidity was determined by a titration method. Fermented fish samples (10 g) were diluted up to 100 mL with distilled water. The acid content in 10 mL of the diluted sample was determined by titration with 0.1 M NaOH by using 1% (w/v) phenolphthalein as an indicator (Dadzie and Orchard, 1997). The pH value of Pha-ork Kontrey was measured using a digital pH meter according to the method reported by Petrus et al. (2013). Homogenates were prepared by mixing 5 g of the samples with 10 mL of distilled water, and the pH was recorded with a digital pH meter.

Enumeration of bacteria    The number of bacteria in samples was enumerated by a serial dilution method. For microbial counts, 10 g of each sample was homogenized in 90 mL of sterile 0.85% saline water, and serial 10-fold dilutions were prepared. MRS agar (Oxoid Co., Ltd.), Chromocult coliform agar (EMD Millipore Co., Ltd., Germany) and EYMS agar (Oxoid Co., Ltd.) were used for enumeration of L. plantarum, E. coli and S. aureus, respectively. The surface plating method was employed for all bacteria (E. coli, S. aureus and LAB). The pour plating method was used to enumerate the homogenized samples without dilution for increased detection sensitivity. In this case, the results obtained from the pour plating method were used when 19 or fewer colonies were found on the surface plated agar plates. An 0.1-mL aliquot of a homogenized sample or diluted sample was used for surface plating. A 1-mL aliquot of the homogenized sample was used for pour plating. Five presumptive colonies each of L. plantarum, E. coli and S. aureus were confirmed with API50/API 50 CHL medium, API 20E, and API Staph (bioMérieux Co., Ltd.), respectively. When we found suspected colonies, we tested one or two more strain(s) additionally.

Sensory evaluation    The fermented fish products without spiked E. coli/S. aureus, in parallel with the treated samples, were subjected to sensory evaluation of appearance, color, odor, taste, and overall acceptability by using a 9-point hedonic scale from 1 (dislike extremely) to 9 (like extremely) as described by Munoz and King (2007). The samples were labeled randomly with three-digit numerical codes and served to a 15-member panel. The volunteer Cambodian panelists were between 20 and 25 years old and were highly knowledgeable about small fermented fish. The panelists were trained to evaluate the sensory characteristics of foods. The panelists were instructed to rinse their mouths with water before tasting each sample.

Statistical analysis    The experiments were performed in triplicate, using three samples per treatment. The mean and standard deviation (SD) of the obtained 9 sets of bacterial numbers were calculated after conversion to logarithmic values. The mean and SD values of the other obtained data were calculated without logarithmic conversion. Statistical analysis was carried out using a one-way ANOVA at a significance level of 0.05. The LSD test was used for multiple comparisons and was performed using the SPSS program (v22.0 for MS-Windows; IBM Corp., USA).

Results and Discussion

Changes in pH and total acidity profile during fermentation    The changes in pH and total acidity during Pha-ork Kontrey fermentation are showed in Fig. 1. The initial concentration of spiked E. coli was 5 log CFU/g in all samples. The starter culture, which consisted of the two selected L. plantarum strains isolated from Pha-ork Kontrey, was inoculated to the samples (8 or 9 log CFU/g) except for the control. Prior to fermentation, the pH of all samples ranged from 6.6 to 6.7. The pH was reduced and total acidity was increased concomitant with the increase in initial LAB concentration (p < 0.05) at 24 and 48 h fermentation. In samples containing 9 log CFU/g of starter culture, the pH decreased to 3.3 at 24 h and then remained stable. In samples containing 8 log CFU/g of starter culture, pH decreased to 3.8 or 3.5 after 24 or 96 h fermentation, respectively. In comparison, the uninoculated sample exhibited a significantly slower pH decrease until 120 h fermentation (p < 0.5). However, there was no significant difference in pH among all samples after 120 h fermentation (p > 0.05).

Fig. 1.

The change of pH (Column) and total acidity (Line) during Pha-ork Kontrey fermentation (E. coli was inoculated).

The columns in black, gray, and white color represent the pH of Pha-ork Kontrey inoculated E. coli at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively. The lines with marker ○, △, and □ represent total acidity of Pha-ork Kontrey inoculated E. coli at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively.

Exchanging E. coli with S. aureus in the above experiments gave similar results to that shown in Fig. 1. The Pha-ork Kontrey sample containing starter culture exhibited a significantly (p < 0.05) faster decrease in pH than the uninoculated control sample (Fig. 2). The pH of the sample started with 9 log CFU/g LAB decreased from 6.8 to 3.3 in 24 h. The pH of the sample started with 8 log CFU/g LAB was around 4.0 after 24 h, and reached 3.6 at 48 h. It took 120 h to reach a similar final pH without the use of starter culture. Maristela et al. (2008) reported similar results, in which the use of starter culture (L. plantarum) produced a rapid drop in pH during sausage fermentation.

Fig. 2.

The change of pH (Column) and total acidity (Line) during Pha-ork Kontrey fermentation (S. aureus was inoculated).

The columns in black, gray, and white color represent the pH of Pha-ork Kontrey inoculated S. aureus at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively. The lines with marker ○, △, and □ represent total acidity of Pha-ork Kontrey inoculated S. aureus at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively.

According to Celik et al. (2010), organic acids produced by LAB are thought to be the main reason for this rapid decline in pH. The increase in acidity of samples with E. coli or S. aureus is shown in the lines of Fig. 1 or 2, respectively. The difference in inoculating bacteria did not affect the increase in acidity significantly (p > 0.05). Increasing the initial LAB concentration by adding LAB starter culture enhanced the increase in total acidity. The total acidity of the samples with 9 log CFU/g LAB increased significantly faster than those with 8 log CFU/g LAB at 48 h fermentation (p < 0.05). In contrast, the total acidity of control samples was significantly lower and the final total acidity at 120 h was also lower. According to Haddadin et al. (2005), L. plantarum is a suitable strain for starter cultures, as it shows greater resistance to acidity than other LAB such as Leuconostoc species and can produce high levels of organic acids.

Changes in number of E. coli and LAB during fermentation    The reduction in E. coli counts and changes in LAB counts are shown in Fig. 3. Based on colony morphology and the API kit results, all picked colonies were thought to be the same as the inoculated strains. After 24 h fermentation, the number of LAB in the samples with 8 or 9 log CFU/g LAB starter culture was increased to 9.7 or 10.1 log CFU/g, respectively. These values were significantly (p < 0.05) higher than the control sample (9.1 log CFU/g). At that time, no significant (p > 0.05) difference in the number of E. coli in the samples with 9 or 8 log CFU/g starter culture was observed. The number of E. coli in the control sample was about 1 log CFU/g higher than the above values at the same time points. No significant (p > 0.05) difference in LAB counts among the three kinds of samples was observed after 48 h fermentation. The values were maintained at around 9 to 10 log CFU/g at 96 h fermentation. Subsequently, there was a slightly reduction in LAB counts due to the accumulation of organic acids. The population of E. coli in the samples with 9 or 8 log CFU/g starter culture decreased sequentially from 5 log CFU/g to undetectable levels in 48–72 or 72–96 h when the pH value reached 3.3 and 3.5, respectively. The number of E. coli in the control samples also decreased slowly and reached an undetectable level after 120 h at pH 3.6. This faster decrease in E. coli in samples containing the starter culture could be attributed to the faster accumulation of lactic acid and other compounds, such as acetic acid and ethanol, following inoculation with L. plantarum (Erbaş et al., 2005; Daeschel et al., 1993).

Fig. 3.

Change of E. coli and LAB count during Pha-ork Kontrey fermentation.

The lines with marker ■, ▲, and ●, represent viable count of E. coli in Pha-ork Kontrey inoculated E. coli at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively. The lines with marker ○, △, and □ represent viable count of LAB in Pha-ork Kontrey inoculated E. coli at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively.

S. aureus and LAB counts during the fermentation process of Pha-ork Kontrey    The changes in S. aureus and LAB counts of samples are shown in Fig. 4. Based on colony morphology and the API kit results, all picked colonies were thought to be the same as the inoculated strains. The growth of LAB was similar to that shown in Fig. 3. The growth of LAB reached late log to stationary phase (around 10 log CFU/g) at 24 h fermentation when starter culture was used. In contrast, an additional 24 h fermentation was required to reach similar LAB counts in the control sample. After this fermentation period, the LAB counts in all samples remained stable by 96 h. Subsequently, counts decreased slowly after the 96 h fermentation period. Similar changes in LAB counts during the production of Tarhana, a fermented milk product produced in Southeast Europe and the Middle East, were also observed (Daglioglu et al., 2002; Settanni, 2011).

Fig. 4.

Change of S. aureus and LAB count during Pha-ork Kontrey fermentation.

The lines with marker ■, ▲, and ●, represent viable count of S. aureus in Pha-ork Kontrey inoculated S. aureus at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively. The lines with marker ○, △, and □ represent viable count of LAB in Pha-ork Kontrey inoculated E. coli at 5 log CFU/g without inoculated L. plantarum (control), with inoculated L. plantarum at 8 log CFU/g, and with inoculated L. plantarum at 9 log CFU/g respectively.

As shown in Fig. 3, no increase in E. coli during fermentation was observed in all samples, including the control sample. In contrast, S. aureus was able to grow during the initial fermentation phase, i.e., 24 h, in the control sample and the sample with an initial 8 log CFU/g LAB. The S. aureus counts in the sample with an initial 9 log CFU/g LAB were almost stable by 24 h fermentation. The S. aureus counts in all samples were reduced after that period. It took 72, 120 and 120 h to reach below the detection limit (2.3 log CFU/g) in the samples with 9 log CFU/g LAB, 8 log CFU/g LAB and the control sample, respectively. The decrease in S. aureus counts during fermentation was closely related to the decline in pH and increase in total acidity (Fig. 2 and Fig. 4). Compared to the case of E. coli contamination, the higher initial LAB counts were thought to be highly important in S. aureus contamination.

Sensory evaluation    The results of sensory evaluation of the prepared Pha-ork Kontrey samples after 72 h (3 days) fermentation are shown in Table 1. The numbers “9” and “1” correspond to the best and worst evaluation, respectively. No significant differences in appearance, color, odor, taste and overall acceptability between the control sample and the sample with 8 log CFU/g starter culture were found (p > 0.05). The odor, taste, and overall acceptability were significantly (p < 0.05) improved by using the 9 log CFU/g starter culture. This might be attributable to the accumulation of sufficient organic acids and other compounds produced by the selected starter culture strains.

Table 1. Sensory score of Pha-ork Kontrey after 72 h fermentation

	Appearance	Color	Odor	Taste	Overall
Control	7.3 ± 0.8 a	7.3 ± 0.6 a	4.2 ± 1.1 a	4.2 ± 1.1 a	5.6 ± 0.9 a
LAB 8	7.5 ± 0.7 a	7.3 ± 0.6 a	4.6 ± 0.9 a	5.3 ± 0.9 b	5.9 ± 1.0 a
LAB 9	7.5 ± 0.7 a	7.4 ± 0.8 a	7.6 ± 0.7 b	7.5 ± 0.7 c	7.3 ± 1.0 b

Mean of the same column with different superscripts (a, b, c) are significantly different (p < 0.05).

The abbreviations on the table means; Control : Natural fermentation, LAB 8 : Fish fermentation by adding L. plantarum with 8 log CFU/g without inoculated E. coli and S. aureus, and LAB 9 : Fish fermentation by adding L. plantarum with 9 log CFU/g without inoculated E. coli and S. aureus, respectively.

Conclusion

The contamination and growth of pathogenic bacteria, e.g., certain strains of E. coli and S. aureus, during the fermentation of foods is a practical food safety risk, especially in developing countries such as Cambodia. Further, under unsuitable fermentation conditions, food spoilage can also occur. To prevent the growth of undesirable bacteria during fermentation, a rapid decrease in pH (or increase in total acidity) is required (Herken and Con, 2012). Gao and Liu (2014) reported that increasing the initial counts of LAB by using a starter culture is effective in achieving this objective.

Saithong et al. (2010) demonstrated that the use of starter culture containing two LAB strains (L. plantarum IFRPD P15 and L. reuteri IFRPD P17) was effective in controlling the growth of E. coli in fermented fish (Plaa-som) and reduced the fermentation time without affecting sensory characteristics. Similarly, we evaluated the effectiveness of new starter culture strains (L. plantarum D11 and S22) for Pha-ork Kontrey production in this study.

Based on the results shown in Figures 1 to 4, decrease of E. coli tended to be faster than that of S. aureus, though no significant differences in the growth of inoculated L. plantarum and acidity were observed. Savitori et al. (2017) isolated LAB strains from Indian traditional pickles and their growth suppression activities against E. coli and S. aureus were compared. Agüero et al. (2020) examined the growth suppression activity of LAB strains during screening of starter strains for dry fermented sausage. The results of both experiments also showed that E. coli was much more susceptible than S. aureus to LAB. The difference in susceptibility is thought to be due to distinctions in the acid resistance of the strains (Davidson and Taylor, 2007).

The order of the decline over time for E. coli was 9 log CFU/g LAB > 8 log CFU/g LAB > uninoculated control (Fig. 3). The use of 8 log CFU/g LAB starter culture could not suppress the growth of S. aureus at 24 h fermentation, and 48 h longer than the 9 log CFU/g LAB starter culture sample was needed to reduce S. aureus counts to below the detection limit (2.3 log CFU/g) (Fig. 4). Since Pha-ork Kontrey is usually eaten after fermentation for 48 to 72 h, the concentration of the starter culture used in this work should be set at 9 log CFU/g. Under this condition, 5 log CFU/g of spiked E. coli and S. aureus was reduced to below the detection limit (Fig. 3 and Fig. 4, respectively), and the sensory parameters were the same or better than those of the control samples (Table 1). This study clearly demonstrates the performance and effectiveness of the new starter culture for Cambodian traditional fermented small fish, Pha-ork Kontrey.

Acknowledgements    This research was partially supported by Kirin Holdings Co., Ltd. (formerly Kirin Brewery Co., Ltd.), Tokyo, through the JICA-Kirin Fellowship program.

References

•  Agüero,  de L.N.,  Frizzo,  L.S.,  Ouwehand,  A.C.,  Aleu,  G., and  Rosmini,  M.R. (2020). Technological characterization of probiotic lactic acid bacteria as starter cultures for dry fermented sausages., Foods, 9, E596.
•  Ananchaipattana,  C.,  Hosotani,  Y.,  Kawasaki,  S.,  Pongsawat,  S.,  Bari,  M.L.,  Isobe,  S., and  Inatsu,  Y. (2012). Bacterial Contamination in Retail Foods Purchased in Thailand. Food Sci. Technol. Res., 18, 705-712.
•  Borresen,  E.C.,  Henderson,  A.J.,  Kumar,  A.,  Weir,  T.L., and  Ryan,  E.P. (2012). Fermented foods: Patent approaches and formulations for nutritional supplementation and health promotion. Recent Pat. Food Nutr. Agric., 4, 134-140.
•  Caplice,  E. and  Fitzgerald,  G.F. (1999). Food fermentations: Role of microorganisms in food production and preservation. Int. J. Food Microbiol., 50, 131-149.
•  Cappuccino,  J. and  Welsh,  C. (2019). “Microbiology: A Laboratory Manual 12th ed.” Pearson, New York.
•  Chrun,  R.,  Hosotani,  Y.,  Kawasaki  S., and  Inatsu,  Y. (2017). Microbiological hazard contamination in fermented vegetables sold in local markets in Cambodia. Biocontrol Sci., 22, 181-185.
•  Celik,  I.,  Isik,  F., and  Yilmaz,  Y. (2010). Chemical, rheological and sensory properties of tarhana with wheat bran as a functional constituent. Akademik Gıda, 8, 11-17.
•  Dadzie  B.K. and  Orchard  J.E. (1997). “Routine Post-harvest Screening of Banana/plantain Hybrids: Criteria and Methods.” Bioversity International, Wageningen. pp. 11-13.
•  Daeschel,  M.A. (1993). Applications and interactions of bacteriocins from lactic acid bacteria in foods and beverages. In “Bacteriocins of Lactic Acid Bacteria,” ed. by  D.G.  Hoover and  L.R.  Steenson. Academic Press, New York, pp. 63-91.
•  Daglioglu,  O.,  Muhammet,  A., and  Tuncay,  G. (2002). Effects of tarhana fermentation and drying methods on the fate of Escherichia coli O157:H7 and Staphylococcus aureus. Eur. Food Res. Technol., 215, 515-519.
•  Davidson,  P.M. and  Taylor,  T.M. (2007). Chemical Preservatives and natural antimicrobial compounds. In “Food Microbiology : Fundamental and Frontiers, 3rd ed.,” ed by  M.P.  Doyle and  L.R.  Beuchat. ASM press, Washington, D.C., pp. 713-745.
•  Erbaş,  M.C. (2005). Microbiological and chemical properties of Tarhana during fermentation and storage as wet-sensorial properties of Tarhana soup. Food Sci. Technol., 38, 409-416.
•  Gadaga,  T.H.,  Nyanga,  L.K., and  Mutukumira,  A.N. (2004). The occurrence, growth and control of pathogens in African fermented foods. Afr. J. Food Agric. Nutr. Dev., 4, 20-23.
•  Gao,  Y. and  Liu,  X. (2014). Bacteriocin-producing Lactobacillus sakei C2 as starter culture in fermented sausages. Food Control, 35, 1-6.
•  Haddadin,  J.S. (2005). Kinetic studies and sensorial analysis of lactic acid bacteria isolated from white cheese made from sheep raw milk. Pakistan J. Nutr., 4, 78-84.
•  Hensyl,  W.R. (ed.). (2000). “Bergey's Manual of Determinative Bacteriology, 9th ed.” Lippincott Wlliams and Wilkins, Misigan, pp. 179-180, 532, 566.
•  Herken,  E. and  Con,  A. (2012). Use of different lactic starter cultures in the production of tarhana. J. Food Proces. Preserv., 38, 59-67.
•  Hurtado,  A.,  Reguant,  C.,  Bordons,  A., and  Rozés,  N. (2010). Evaluation of a single and combined inoculation of a Lactobacillus pentosus starter for processing Cv. Arbequina natural green olives. Food Microbiol., 27, 731-740.
•  Inatsu,  Y.,  Bari,  M.L.,  Kawasaki,  S., and  Kawamoto,  S. (2005). Bacteria in traditional fermented vegetables produced in northern part of Vietnam. Jpn. J. Food Microbiol., 22, 103-111.
•  Maristela,  C. S. (2008). Lactobacillus plantarum AJ2 isolated from naturally fermented sausage and its effects on the technological properties of Milano-type salami. Ciênc. Tecnol. Aliment., 28, 709-717.
•  Marui,  J.,  Boulom,  S.,  Panthavee,  W.,  Momma,  M.,  Kusumoto,  K.I.,  Nakahara,  K., and  Saito,  M. (2014). Culture-independent analysis of the bacterial community during fermentation of pa-som, a traditional fermented fish product in Laos. Fish. Sci., 80, 1109-1115.
•  Muñoz,  A. and  King,  S.C. (eds.) (2007). “International consumer product testing across cultures and countries: General Principles.” ASTM International, Pennsylvania, p. 11.
•  Parvz,  S.,  Malik,  K.A.,  Kang,  S.A., and  Kim,  H.-Y. (2006). Probiotic and their fermented food products are beneficial for health. J. Appl. Microbiol. Rev., 100, 1171-1185.
•  Petrus,  Purnomo,  H.,  Suprayitno,  E., and  Hardoko. (2013). Physicochemical characteristics, sensory acceptability and microbial quality of Wadi betok a traditional fermented fish from South Kalimantan, Indonesia. Int. Food Res. J., 20, 933-939.
•  Ray,  B. (1992). The need for food bio preservation. In “Food Bio Preservatives of Microbial Origin,” ed. by  B.  Ray and  M.  Daeschel. CRC Press, Florida, pp. 1-23.
•  Rhee,  S.J.,  Lee,  J.-E., and  Lee,  C.-H. (2011). Importance of lactic acid bacteria in Asian fermented foods. Microb. Cell Fact., 10, 1-5.
•  Saithong,  P.,  Panthavee,  W.,  Boonyaratanakornkit,  M., and  Sikkhamondhol,  C. (2010). Use of a starter culture of lactic acid bacteria in plaa-som, a Thai fermented fish. J. Biosci. Bioeng., 110, 553-557.
•  Savitri,  M.,  Kumar,  V.,  Kumari,  A.,  Angmo,  K., and  Bhalla,  T.C. (2017). Isolation and characterization of lactic acid bacteria from traditional pickles of Himachal Pradesh, India. J. Food Sci. Technol., 54, 1945-1952.
•  Settanni,  L.T. (2011). Evolution of fermenting microbiota in tarhana produced under controlled technological conditions. Food Microbiol., 28, 1367-1373.
•  Sven-Olof,  E. (2008). Fermented food. In “Food Microbiology 6,” ed. by  E.  Sven-Olof. KTH - Biotechnology, Stockholm, p. 78.
•  Ying,  L.,  Xie,  X.X.,  Ibrahim,  S.A.,  Khaskheli,  S.G.,  Yang,  H.,  Wang,  Y.F., and  Huang,  W. (2016). Characterization of Lactobacillus pentosus as a starter culture for the fermentation of edible oyster mushrooms (Pleurotus spp.). LWT - Food Sci. Technol., 68, 21-26.