Properties of Antimicrobial Substances Produced by Bacillus amyloliquefaciens CJW15 and Bacillus amyloliquefaciens SSD8

Bacillus amyloliquefaciens CJW15와 SSD8이 만드는 항균물질들의 특성

Abstract

Two Bacillus strains, CJW15 and SSD8, with strong antibacterial activities were isolated from cheonggukjang. Both were identified as B. amyloliquefaciens strains after gene sequencing of rRNA and recA. CJW15 strongly inhibited the growth of B. cereus (ATCC14579), Listeria monocytogenes (ATCC19111), and Lactococcus lactis (ATCC11454). In comparison, SSD8 inhibited the growth of B. cereus (ATCC14579) and Enterococcus faecium (ATCC19953). The antibacterial activities of the two strains were not affected when exposed to a temperature of $100^{\circ}C$ for 15 min and were quite stable in acidic (pH 3) and alkaline (pH 12) pH conditions. Enzymatic treatments (trypsin, pepsin, proteinase K, and protease) had no effect on the activity of CJW15, but reduced the activity of SSD8 by half. Both isolates possess genes involved in the synthesis of lipopeptides (e.g. surfactin, fengycin, iturin, and iturin A), and genes encoding subtilin, a bacteriocin. Moreover, both isolates have fibrinolytic activities as well.

청국장에서 강력한 항균력을 지닌 2개의 Bacillus 균주들인 CJW15와 SSD8이 분리되었다. 16S rRNA와 recA 유전자들 염기서열 결정에 의해 두가지 균주 모두 B. amyloliquefaciens로 동정되었다. CJW15는 B. cereus ATCC14579, Listeria monocytogenes ATCC19111, Lactococcus lactis ATCC11454 들의 증식을 강력히 억제하며 SSD8은 B. cereus ATCC14579와 Enterococcus faecium ATCC19953 증식을 억제하였다. 두 균주들의 항균력은 $100^{\circ}C$, 15분 처리에도 감소하지 않았고 산성인 pH 3과 알칼리인 pH 12에서도 안정하였다. 트립신, 펩신, 프로테아제 K, 프로테아제 효소처리에 의해 CJW15 항균력은 변화가 없었지만 SSD8 항균력은 절반으로 감소하였다. 두 균주 공히 surfactin, fengycin, iturin, iturinA와 같은 lipopeptide 생합성 유전자들을 지니고 있고 subtilin과 같은 박테리오신 유전자들도 지니고 있다. 또 두 균주들은 혈전용해능을 지니고 있다.

Introduction

Bacillus subtilis and closely related species are known to produce a variety of antimicrobial substances and significant portions of their genomes are devoted to genes involved in the production of antimicrobial substances [15]. Antimicrobial substances produced by bacilli have been the subject of active research because they have potentials as novel biopreservatives preventing bacterial and fungal contaminations in foods. Although chemically synthesized compounds have been used as preservatives for food industry, their safety has been questioned and consumers prefer foods with no synthetic preservatives [2]. Substances classified as generally recognized as safe (GRAS) are the promising alternatives to chemical preservatives, and among them, bacteriocins produced by lactic acid bacteria (LAB) and bacilli are the primary candidates [3, 14]. Bacteriocins have some shortcomings such as low stabilities under high temperature and low or high pH environments, poor production yields, short duration times, and narrow inhibition spectra. Bacilli produce other types of antimicrobial substances such as lipopeptides in addition to bacteriocins. Surfactin, fengycin, and iturin are well-known lipopeptides and they possess antifungal, antibacterial, or antiviral activities [15]. More studies are required before use of lipopeptides or bacilli producing them as food preservatives is realized. Especially, not many studies have been done on the antimicrobials produced by B. amyloliquefaciens strains, compared to B. subtilis strains. In this work, 2 B. amyloliquefaciens strains were isolated from cheonggukjang and their antimicrobial activities were studied.

 

Materials and Methods

Isolation of Bacillus strains with antibacterial activities

Bacilli were isolated from cheonggukjang purchased at a local market in Jinju, Republic of Korea, in 2014. Ten gram of cheonggukjang was mixed with 90 ml of 0.1% peptone water, homogenized by Stomacher 80 (Seward, Worthing, UK), and serially diluted. Diluted samples were spreaded on Luria-Bertani (LB) (BD, Franklin lakes, NJ, USA) plates and incubated for 48 h at 37℃. Colonies showing typical Bacillus morphologies were saved and examined for the antimicrobial activity by agar spot method [5]. Bacillus cereus ATCC14579 was used as an indicator. Bacilli were grown in LB for 48 h at 37℃ and 1 μl culture was spotted on a LB plate. The plate was overlaid with LB top agar (0.7%, w/v) containing 50 μl of B. cereus ATCC14579 culture grown in LB for 48 h.

Identification of CJW15 and SSD8

Two isolates, CJW15 and SSD8, were grown in LB broth for 18 h and cells were recovered by centrifugation at 12,000 × g for 5 min. Cells were resuspended in 1 ml of lysis buffer (10 mM Tris-HCl, pH 8.0, 10 mM EDTA, 100 mM NaCl, 2% (w/v) SDS, and 400 μg/ml proteinase K (Sigma, St. Louis, MO, USA)). Total DNA was extracted by phenol and chloroform extractions after 30 min incubation at 55℃. Purified DNA was resuspended in small volume of sterile water after ethanol precipitation and 70% ethanol wash. The 16S rRNA genes were amplified using the following primers: bac-F (5'-CGGCGTGCCTAATACATGCAAG-3'), and bac-R (5'-GGCATGCTGATCCGCATTACTA-3') [6]. For the amplification of recA genes, following primers were used, recA-F (5'-TGAGTGATCGTCAGGCAGCCTTAG-3') and recA-R (5'-CYTBRGATAAGARTACCAWGMACCGC-3') [9]. PCR mixture (50 μl) consisted of 1 μl each primer, 1 μl template DNA, 5 μl dNTP (0.25 mM), 0.5 μl ExTaq DNA polymerase (Takara, Shiga, Japan), and 5 μl 10 × Taq DNA polymerase buffer. PCR was performed using a MJ Mini™ thermal cycler (BioRad, Hercules, CA, USA). The amplification conditions were as follows, initial denaturation at 95℃ for 5 min, followed by 30 cycles of 95℃ for 30 s, 58℃ for 30 s, and 72℃ for 30 s, and a final extension at 72℃ for 5 min [11]. The PCR products were purified using a FavorPrep gel/PCR purification mini kit (Favorgen Biotech., Ping-Tung, Taiwan). Nucleotide sequences of the amplified fragments were determined at Cosmogenetech (Seoul, Korea) and analyzed by BLAST (NCBI, Bethesda, MD, USA) [11].

Inhibition spectra of B. amyloliquefaciens CJW15 and B. amyloliquefaciens SSD8

B. amyloliquefaciens CJW15 and B. amyloliquefaciens SSD8 were grown in tryptic soy broth (TSB) (MB cell, Los Angeles, CA, USA) for 48 h at 37℃. Supernatants obtained by centrifugation (7,000 × g, 15 min at 4℃) were filtered using 0.45 μm membrane filters (Advantec MFS, Inc, Dublin, CA, USA). Filtered supernatant (FS) was used as the sample for the inhibition study after the pH was adjusted to 7. Inhibition spectra of B. amyloliquefaciens CJW15 and B. amyloliquefaciens SSD8 were examined using FS and by agar well diffusion method [10]. Indicator strains used in this study are shown in Table 1. Pediococcus pentosaceus NRRLB-14009, Leuconostoc messenteroides ATCC9135, and Enterococcus faecalis ATCC29212 were grown in MRS broth (Neogen, Lansing, MI, USA) statically at 30℃. Streptococcus thermophilus KFRI193 and Listeria monocytogenes ATCC19111 were incubated in M17 (BD) and brain heart infusion (BHI) (BD), respectively, at 37℃ with aeration. Other indicators were grown in LB with shaking (150 rpm) at 37℃ [8].

Table 1.Degree of inhibition: +, 0.5 to 2 mm (moderate inhibitory activity); ++, 2 to 4 mm (strong inhibitory activity); +++, more than 4 mm (very strong inhibitory activity). Inhibition zone = (diameter of an inhibition zone in mm- diameter of a well)/2.

When the indicator culture reached the exponential phase, 100 μl of indicator culture was added to 10 ml of soft agar (0.7%, w/v) and the mixed soft agar was immediately overlayed on a LB plate. Wells were made on the overlayed plate using a Pasteur pipette and 50 μl FS (pH adjusted to 7.0) was added into each well. The plate was incubated for overnight before inhibition zons were examined [8]. The strength of antimicrobial activity was determined by two-fold serial dilution method as described previously [11]. The activity was defined as the reciprocal of the highest dilution that still produced a definite zone of inhibition and expressed as activity units (AU) per milliliter after multiplied by a conversion factor.

The effect of medium on the growth and the antibacterial activities of isolates

B. amyloliquefaciens CJW15 and SSD8 were grown overnight in LB and the cultures were used to inoculate (1%, v/v) each 250 ml of LB, TSB, NB, and BHI. Inoculated media were incubated for 96 h at 37℃ with shaking (150 rpm). At every 6 h, 4 ml of each culture was taken and the OD600 value and antibacterial activity were measured. The antimicrobial activity of FS was assayed by agar well diffusion method using B. cereus ATCC14579 as the indicator. All measurements were done in triplicates and the means were shown.

Stability of the antimicrobial substances

B. amyloliquefaciens CJW15 and SSD8 were grown in TSB for 60 h and the resulting FS was used as the sample for the stability test (Table 2). The pH of FS (3 ml) was adjusted to 3−11 using 1 N HCl or 1 N NaOH. After 2 h incubation at 25℃, the pH was readjusted to 7 and the remaining activity was measured by agar well diffusion method using B. cereus ATCC14579 as the indicator. Sample without pH adjustment was used as the control. FS (pH 7.0) was heated at 50, 60, 70, 80, 90 and 100℃ for 15 min. After cooled to room temperature, the remaining activity was measured. Sample without heat treatment was used as the control. Stability against enzyme treatment was examined for the following enzymes, protease (Sigma, P5147 in 5 mM phosphate buffer, pH 7.0), pepsin (P-6887 in 0.02 N HCl), trypsin (T-8918 in 10 mM phosphate buffer, pH 7.0), proteinase K (P2308 in 10 mM phosphate buffer, pH 7.0), lipase from porcine pancreas (L3126, in 50 mM Tris-HCl and 10 mM CaCl2 buffer, pH 7.0), carboxypeptidase B from porcine pancreas (C9584, in 50 mM sodium phosphate buffer, pH 7.6), α-amylase solution (A3306), and β-amylase solution (A7005). Each enzyme was added to FS at the concentration of 1 mg/ml. After 2 h at 37℃, the residual activity was examined as described above. Buffer without an enzyme was used as the control.

Table 2.Stability of antimicrobial substances upon pH, heat, and enzyme treatments.

Tricine SDS-PAGE and gel overlay

FS was prepared from 48 h culture in LB and concentrated as described below. FS was subjected to ammonium sulfate precipitation. Ammonium sulfate powder was very slowly added to each 200 ml FS from SSD8 and CJW15 to make a final concentration of 80% saturation. The mixture was stirred gently for 6 h and left in ice for overnight. The precipitant obtained from centrifugation (11,000 × g, 30 min) was dialyzed for 24 h using Tris-HCl buffer (20 mM, pH 7.0). Dialyzed sample was concentrated using Amicon ultra-15 centrifugal filter devices (10K device, Merck Millipore Ltd., Cork, Ireland) by centrifugation at 6, 000 × g for 30 min. Tricine-SDS-PAGE was carried out according to the method described by Schägger and von Jagow [13]. Concentrated samples were loaded on a Tricine SDS-PAGE gel with a 1 cm 10% gel between16% separation gel and the 4% stacking gel, duplicately. After electrophoresis, half of the gel was stained with coomassie brilliant blue R-250 in methanolwater-acetic acid (7:40:53, v/v) for 30 min, destained for 1 h in the solution of methanol-water-acetic acid (40:17:43, v/v) followed by overnight destaining in methanol-water-acetic acid (5:17:78, v/v). Another half of the gel was washed with distilled water for 4 h with water change every half hour, then the washed gel was overlaid with 25 ml of 0.7% (w/v) soft agar containing 1% of B. cereus ATCC14579 culture (OD600 = 0.4 − 0.8). The overlaid gel was incubated for overnight at 37℃ and the clear zone of inhibition was examined.

Fibrinolytic activity of B. amyloliquefaciens CJW15 and SSD8

B. amyloliquefaciens CJW15 and SSD8 were cultivated in TSB (1%, v/v) for 84 h 37℃. At 12 h intervals, 5 ml of culture was taken out and centrifuged at 4,000 × g for 10 min. FS was used as the sample for the fibrinolytic activity measurement by fibrin plate method [7]. Fibrinogen (Sigma) solution (0.5%, w/v) in 10 ml of 1 M phosphate-buffered saline (PBS) was mixed with an equal volume of 2% (w/v) agarose solution and then poured onto a plate, which previously evenly distributed with 0.1 ml of thrombin solution (100 NIH units/ml, Sigma). The plate was left at room temperature for 2 h to allow a fibrin clot to form. Holes were made by a capillary glass tube and then 10 μl of FS was added into the hole. On the same plate, plasmin was added at different units (3, 6, 9, 45 mU). The plate was incubated at 37℃ for 8 h before measuring the sizes of the lysis zones. The size was converted into plasmin units (U) by comparing it to those from plasmin standards. A standard curve showing the relationship was obtained [7]. Supernatant from B. subtilis HK 176 was used as a positive control [7].

Detection of genes encoding antimicrobial substances using PCR

Total genomic DNA was extracted from B. amyloliquefaciens strains. PCR was performed to detect genes involved in the synthesis of surfactin, fengycin, iturin, iturin A, subtilin, and subtilosin A. Primers used were shown in Table 3. PCR was carried out using a MJ Mini™ PCR machine (BioRad). The reaction mixture (50 μl) consisted of 1 μl of template DNA, 1 μl of each primer (10 μM), 5 μl of dNTPs (0.25 mM), and 1 μl of ExTaq DNA polymerase (Takara). The cycling conditions were as follows: 95℃ for 4 min, 40 cycles of 94℃ for 30 sec, 43−59℃ for 30 sec, 72℃ for 30 sec, and a final extension at 72℃ for 2 min. PCR products were visualized on an agarose gel.

Table 3.1+, a PCR product with the expected size was observed. 2−, a PCR product with the expected size was not observed.

Inhibition mode of antibacterial substances

FS from B. amyloliquefaciens CJW15 and SSD8 was obtained from 96 h culture in TSB and further concentrated by using Amicon Ultra-15 centrifugal filter devices (Merck Millipore Ltd.) at 4,000 × g for 40 min, respectively. The final activity was 3,200 AU/ml for both FS. LB broth (20 ml) was 1% inoculated with overnight culture of B. cereus ATCC14579. Inoculated broth was grown until it reached exponential phase of growth (OD600 = 0.64), then each 5 ml of culture was transfered into 100 ml conical flasks. At the same time, 5 ml (16,000 AU) of concentrated FS sample was added to each flask. Five ml of TSB was added to a flask as a negative control. Conical flasks were shaken at 37℃ and the viable cell numbers of each sample were measured at every 3 h for the next 12 h period.

 

Results and Discussion

Identification of CJW15 and SSD8

Two isolates showing strong antimicrobial activities, CJW15 and SSD8, were identified by 16S rRNA and recA gene sequencing. The 16S rRNA gene sequence (1,179 bp) from CJW15 was 100% identical with those from B. amyloliquefaciens, B. subtilis, and B. methylotrophicus (data not shown). recA gene sequence (758 bp) from CJW15 was 99% identical with those from Bacillus species. The 16S rRNA gene sequence (1,178 bp) from SSD8 was 100% identical with those of B. subtilis, B. amyloliquefaciens, and B. methylotrophicus (data not shown). recA sequence (702 bp) from SSD8 was 100% identical with those from B. amyloliquefaciens strains and 99% with other species. From these results, both CJW15 and SSD8 were identified as B. amyloliquefaciens. The Genbank accession numbers for the partial 16S rRNA and recA genes are KR095447 and KR095446, respectively, for B. amyloliquefaciens CJW15, and KR095443 and KR095442, respectively, for B. amyloliquefaciens SSD8.

Inhibition spectra of B. amyloliquefaciens CJW15 and SSD8

B. amyloliquefaciens CJW15, B. amyloliquefaciens SSD8, and B. subtilis EMD4 (control, a bacteriocin producer) inhibited the growth of B. cereus ATCC14579, B. amyloliquefaciens CH86-1 and B. subtilis 168 (Table 1). B. amyloliquefaciens CJW15 showed broader inhibition spectra than B. amyloliquefaciens SSD8, inhibiting the growth of L. monocytogenes ATCC19111, Lactobacillus delbrueckii ssp. lactis ATCC4797, and Lactococcus lactis ATCC11454 which were not inhibited by B. amyloliquefaciens SSD8. B. amyloliquefaciens SSD8 and B. subtilis EMD4 inhibited the growth of Enterococcus faecium ATCC19953, which was not inhibited by B. amyloliquefaciens CJW15. B. subtilis EMD4 showed broader inhibition spectra than two B. amyloliquefaciens isolates, inhibiting the growth of B. licheniformis ATCC4527, Streptococcus thermophilus KFRI193, Pediococcus pentosaceus NRRLB-14009, and Enterococcus faecalis ATCC29212. B. cereus is a common and important pathogen in foods including cheonggukjang and doenjang. B. amyloliquefaciens CJW15 and B. amyloliquefaciens SSD8 seem useful as starters for fermented soy foods such as cheonggukjang, where contamination of B. cereus is a constant threat [4].

Stability of the antimicrobial substances

FS from B. amyloliquefaciens CJW15 showed strong resistance to the proteolytic enzymes (trypsin, pepsin, protease, proteinase K, carboxypeptidase B), which indicated the non-proteinaceous nature of the antibacterial substance(s) (Table 2). The control, FS from B. subtilis EMD4, lost all the activity after proteinase K and protease treatments, and the result was expected because the inhibitory substance was a bacteriocin, BacEMD4 [11]. For the FS from B. amyloliquefaciens SSD8, half of the activity was lost by proteolytic enzymes except carboxypeptidase B. Lipase from porcine pancreas also reduced the activity to half. The results indicated a possibility that a bacteriocin or other substance sensitive to proteolytic enzyme or lipase was a part of the inhibitory substance(s). Antimicrobial activities of FSs from both strains were resistant against α-amylase and β-amylase. Antimicrobial activities were stable against extreme pH conditions. When the pH of FS was adjusted to different pH (3−11), the inhibiting activity was not affected. Inhibiting activities were also stable at high temperature, 100% of the activities were remained after 15 min exposure to 50−100℃. From these results, it was concluded that the dominant antibacterial substance in the FS from B. amyloliquefaciens CJW15 and SSD8 was not proteinaceous. High stability of an antimicrobial substance is a desirable property when the inhibitory substance or the producing strain is included into foods as a biopreservative to inhibit the growth of pathogens like B. cereus since many food processing steps include low pH or high temperature conditions.

Effect of different culture media on the growth and antimicrobial activities of B. amyloliquefaciens strains

B. amyloliquefaciens CJW15 and B. amyloliquefaciens SSD8 were grown in different media (LB, NB, TSB, and BHI) and the growth (OD600) and the antibacterial activities were measured during growth. There were no big differences in terms of growth of both strains in four different media except both strains grew slightly better when grown in TSB and LB at the stationary growth stage (Fig. 1). In B. amyloliquefaciens CJW15, the antibacterial substance was produced during stationary phase (after 12 h) in LB, BHI, and TSB, and produced at the exponential phase in NB (after 6 h) (Fig. 1A). In B. amyloliquefaciens SSD8, the production of antibacterial substance started at stationary phase in TSB and BHI (after 12 h and 24 h, respectively) while it started during the exponential phase in NB and LB (Fig. 1B). The antibacterial activities of both strains increased along with the incubation time in all 4 media. For B. amyloliquefaciens CJW15, the highest activity was detected from TSB culture (320 AU/ml) after 60 h, and the highest activities were 160 AU/ml in the other 3 media. For B. amyloliquefaciens SSD8, the highest activity (160 AU/ml) was observed in LB (42 h), followed by TSB (48 h) and NB (60 h) cultures. BHI was the most unfavorable medium for the antibacterial activity and the highest activity was just 80 AU/ml. In conclusion, the best medium for the antibacterial activity of B. amyloliquefaciens CJW15 and SSD8 was TSB and LB, respectively.

Fig. 1.Growth and antibacterial activities of B. amyloliquefaciens CJW15 (A) and B. amyloliquefaciens SSD8 (B) in different growth media. -●-, OD600; -■-, antibacterial activity (AU/ml).

SDS-PAGE and gel overlay

FS from B. amyloliquefaciens CJW15 and SSD8 was concentrated by ammonium sulfate precipitation and further concentrated using Amicon Ultra-15 centrifugal filter. Eighty μg of concentrated samples were loaded on 16% Tricine SDS gels in duplicates. Although several bands were detected from SSD8 sample, the bands were not clearly separated from each other. CJW15 sample was more difficult to be resolved. When 20 μg of sample from B. amyloliquefaciens SSD8 was examined by activity detection method, two growth inhibition zones were observed below 6.5 kDa position (Fig. 2). The upper zone was around 3.5 kDa in size and the lower zone was below 1 kDa in size. But no matching bands were located at the corresponding positions on the coomassie stained gel. This might be due to either the low concentration of the inhibitory substance or poor stainability of the substance. Two inhibition zones with the same apparent sizes were also observed from CJW15 sample but the sizes of inhibition zones from CJW15 were larger than those from SSD8. It is likely that 2 different inhibitory substances are present in the supernatant of CJW 15 and SSD8.

Fig. 2.Tricine SDS-PAGE and gel overlay. 1. ultra low range molecular weight marker (M3546, Sigma); 2. SSD8, 80 μg; 3. CJW15, 80 μg; 4. SSD8, 20 μg on gel overlay; 5. CJW15, 20 μg on gel overlay.

Fibrinolytic activities of two strains

Fibrinolytic activities of B. amyloliquefaciens CJW15 and SSD8 increased and then decreased as the growth continued, which were quite different from the control, B. subtilis HK176 [7], a strain with strong fibrinolytic activity and its fibrinolytic activity increased continuously with incubation time (Fig. 3). Compared to the control, the activities of B. amyloliquefaciens strains were weaker. For both strains, the highest activity was observed at 48 h and the activity was 84.40 mU/μl for CJW15 and 107.44 mU/μl for SSD8.

Fig. 3.Fibrinolytic activity of B. amyloliquefaciens CJW15 and SSD 8. -■-, fibrinolytic activity of B. amyloliquefaciens CJW15; -●-, fibrinolytic activity of B. amyloliquefaciens SSD8; -○-, fibrinolytic activity of B. subtilis HK176 (control).

Detection of genes encoding antimicrobial substances using PCR

PCR results showed that both isolates possessed genes encoding lipopeptides and bacteriocins (Table 3). Genes amplified by PCR included those encoding surfactin, fengycin, iturin, and iturin A, which are lipopeptides. Sfp, one of the genes responsible for the production of surfactin, was not amplified from both isolates, but two other genes, srfAA and srf/Ich, were successfully amplified (Table 3). Genes encoding bacteriocins were also amplified. A gene encoding subtilin, a 32 amino acid pentacyclic lantibiotic, was detected from both CJW15 and SSD8. A gene encoding subtilosin A was detected only from CJW15. The presence of bacteriocin genes does not necessarily prove that these bacteriocins are actually produced. For B. amyloliquefaciens CJW15, antimicrobial activity was not decreased by protease treatment, and this indicated that either bacteriocin was not produced or not the major component for the observed antimicrobial activity. In case of B. amyloliquefaciens SSD8, the antimicrobial activity was decreased to half by protease treatments. This indicated a possibility that a bacteriocin, probably subtilin, might be an important component contributing to the antimicrobial activity. PCR results indicated that both strains might produce diverse antimicrobial substances and the major inhibitory substances are probably lipopeptides, considering the stability of inhibitory activity against high temperature and pH change. Further studies are necessary on the identification of inhibitory compounds produced by CJW15 and SSD8. It is known that many bacilli produce surfactin and depending upon each strain, the detailed list of inhibitory substances is variable [15]. Detection of genes by PCR from a strain is not enough to prove that the inhibitory substances encoded by the genes are actually produced. Mutations in genes occur and abolish the production of the inhibitory substances. Therefore, purification and identification of compounds are required to prove whether a specific inhibitory substance is actually produced. Compared to B. subtilis strains, not many studies have been done on the antimicrobial substances produced by B. amyloliquefaciens strains.

Inhibition mode of antimicrobial substance(s)

When the concentrated FS from B. amyloliquefaciens CJW15 (16,000 AU) was added into B. cereus culture in its exponential growth phase, the viable cells decreased from 6.83 to 5.71 log CFU/ml in the first 3 h (Fig. 4). Viable cell counts were 5.65, 5.74, and 5.46 log CFU/ml at 6, 9, 12 h (after FS addition), respectively. Meanwhile, in the culture with concentrated FS from SSD8, the viable cells decreased from 6.83 to 6.44 log CFU/ml in the first 3 h, and viable cell counts were 5.75, 5.89, 5.67 log CFU/ ml at 6, 9, 12 h (after FS addition), respectively. The number of B. cereus cells in a control (without FS) increased, reaching to 10.08 log CFU/ml at 12 h. The results indicated that the inhibitory substance(s) had bacteriocidal effect although the degree of reduction in the viable counts was not great when compared with BacEMD4 from B. subtilis EMD4. When BacEMD4 (12,800 AU) was added into B. cereus ATCC14579 culture, the viable count decreased from 8.3 to 4.4 log CFU/ml in 3 h and then decreased further to 1.8 log CFU/ml in 12 h [11]. When BacW42 (1,600 AU/ml) was added into actively growing Listeria monocytogenes ATCC 19111, viable cells decreased from 8.3 to 6.6 log CFU/ml after 6 h, and to 1.1 log CFU/ml after 12 h [8]. The small reduction caused by FS from B. amyloliquefaciens isolates might be an indication that the inhibitory substances are not bacteriocins.

Fig. 4.Inhibition mode of B. amyloliquefaciens CJW15 and SSD8 against B. cereus ATCC 14579 cells. -●-, log CFU/ml of B. cereus ATCC14579 culture without FS (control); -▼-, log CFU/ ml of B. cereus ATCC14579 culture with CJW15 sample (16,000 AU); -○-, log CFU/ml of B. cereus ATCC14579 culture with SSD8 sample (16,000 A).

More studies are required to find out the nature of the antimicrobial substance(s) produced by B. amyloliquefaciens CJW15 and SSD8 strains. Other properties of these 2 isolates such as sensory properties of soy products fermented with these isolates should be investigated before they are employed as starters for fermented soy foods or biopreservatives for foods. Like B. subtilis strains, B. amyloliquefaciens strains are known to produce many diverse groups of antimicrobials [15], and some of them might be useful as biopreservatives for foods.