Fujitoshi Yanagida, The Institute of Enology and Viticulture, University of Yamanashi, 1-13-1, Kitashin, Kofu, Yamanashi 400-0005, Japan (e-mail: firstname.lastname@example.org).
Aims: To isolate, characterize and identify bacteriocins from lactic acid bacteria in soil.
Methods and Results: Thirty-four acid-producing bacteria were isolated from 87 soil samples. Antibacterial activities were detected, and one strain, L28-1 produced a bacteriocin that was active against some Gram-positive bacteria. L28-1 was identified as Enterococcus durans by 16S rDNA sequence analysis and API50CHL. This bacteriocin did not lose its activity after autoclaving (121°C for 15 min), but was inactivated by protease K. The bacteriocin was purified by hydrophobic column chromatography, and Sep-Pak C18. Tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis showed that the partially purified bacteriocin contained numerous protein bands. Two bands that displayed antibacterial activities were c. 3·4 and 2·5 kDa in size. In this work, the 3·4-kDa bacteriocin was analysed with N-terminal amino acid and DNA sequencing and matrix-assisted laser desorption ionization-time of flight mass spectrometry analysis. The results indicated that the 3·4-kDa bacteriocin of Ent. durans L28-1 is a new natural enterocin variant.
Conclusions: Enterococcus durans L28-1 produced a new bacteriocin.
Significance and Impact of the Study: This study reports a novel bacteriocin that is produced by Ent. durans that has potential for use as a food preservative.
Bacteriocins produced by lactic acid bacteria have attracted special interest as potential alternative, safe food preservatives. Lactic acid bacteria have been used as food and feed preservatives for centuries, and bacteriocin-producing lactic acid bacteria could replace chemical preservatives for the prevention of bacterial spoilage or the outgrowth of pathogenic bacteria in food products (Daeschel 1989). Bacteriocins of lactic acid bacteria are defined as antimicrobial peptides, proteins or protein complexes that are active against Gram-positive bacteria, especially a set of genetically closely related species (Klaenhammer 1988). Examples of Gram-positive food-borne pathogens against which some lactic acid bacteria-produced bacteriocins are active include Listeria monocytogenes, Staphylococcus aureus, Bacillus subtilis and spores of Clostridium perfringens (Toit et al. 2000).
From 1998, we began isolating lactic acid bacteria from soil, and screening bacteriocin-producing strains from these isolates. We found that the coccus strain L28-1 showed antibacterial activity against Lactobacillus sakei JCM 1557T in the first screening. In later tests, the bacteriocin showed antibacterial activity against some Gram-positive bacteria, such as Ent. faecium MR006 (Onda et al. 2002), Enterococcus durans MY411 (Onda et al. 2002), Lactococcus lactis ssp. lactis GM005 (Onda et al. 2003), Lact. plantarum NRIC 1067T and Lact. animalis JCM5670T. Strain L28-1 was identified as Ent. durans with 16S rDNA sequence analysis and API50CHL. Bacteriocins produced by enterococci have been frequently reported, but little is known about bacteriocins produced by Ent. durans. A partial N-terminal amino acid sequencing analysis, from position 1 to 28, was performed, and at least three amino acids were found to differ between the L28-1 bacteriocin and enterocin B. In this paper, the production, purification and characterization of the L28-1 bacteriocin are described.
Materials and methods
Isolation and characterization of acid-producing strains
Eighty-seven soil samples were collected in Yamanashi and Nagano prefectures of Japan from 1998 to 2000. Soil samples were collected in the rhizospheres of trees, at a depth of c. 5–10 cm from the soil surface, using sterile spoons, and saved into clean bags. Two methods, direct spreading and accumulation with incubation, were used for the isolation of acid-producing bacteria. All samples were incubated under anaerobic conditions (BBLTM GasPakTM, H2 + CO2) at 30°C for 2–3 d. After incubation, acid-producing bacteria were isolated.
Acid-producing bacteria, picked up from the direct-spreading and accumulation methods, were re-incubated with fresh broth. These bacteria were tested with a Gram stain, and only Gram-positive bacteria were kept.
These acid-producing bacteria were initially identified through examination of the cell shape, lactic acid isomers, fermentation type, growth temperature, fermentation of carbohydrates and type of cell wall. Further, sugar fermentation reactions of the bacteriocin-producing strain were performed using API50CH test strips and 50CHL medium (BioMérieux, Marcy I'Etoile, France).
The GC content, 16S rDNA restriction fragment length polymorphisms (RFLP) (Ramos and Harlander 1990; Gurtler et al. 1991; Johansson et al. 1995) and 16S rDNA sequence (Sato et al. 2001) were also used to classify and identify these isolates. The 16S rDNA sequences determined in this study were manually aligned with previously published sequences of reference strains obtained from the database of the DNA Data Bank of Japan (DDBJ, http://www.ddbj.nig.ac.jp/) and software of Genetyx-Mac (Ver. 7.3; Genetyx Corporation, Tokyo, Japan).
Screening for bacteriocin-producing strains
Supernatants were obtained from these isolates in MRS (de Man, Rogosa and Sharpe) broth for about 14 h at 30°C. Two millilitres of soft agar (0·5% w/v) was seeded with 100 μl of an overnight culture of the indicator strain and used to overlay MRS agar plates. Antibacterial activity of each strain was determined by pouring the supernatant fluid into the hole made on the MRS agar plates. Plates were incubated overnight at 30°C and examined for clear zones of inhibition. In the first screening, antibacterial activity was detected with an agar-well diffusion assay using Lact. sakei JCM 1557T as the indicator strain. Some Gram-positive bacteria were also used as indicator strains in later tests.
Effect of enzymes and heat on bacteriocin activity
To evaluate heat stability, neutralized supernatant fluid of the bacteriocin-producing strain was incubated at 80°C for 30 min, 90°C for 30 min, 100°C for 30 min, 110°C for 15 min and 121°C for 15 min. Neutralized supernatant fluid was also treated with proteinase K (30 IU mg−1, 37°C for 5 h), and then bacteriocin activity was determined using an agar-well diffusion assay.
Partial purification of the bacteriocin
A culture of the bacteriocin-producing strain was grown for 14 h at 30°C, and its supernatant fluid was obtained by centrifugation at 5040 g at 4°C for 30 min using the SCR20B centrifuge with a RPR-10 rotor (Hitachi Ltd, Tokyo, Japan). The fluid was filtered through 0·45-μm membrane filters (Toyo Roshi Kaisha Ltd, Tokyo, Japan), adjusted to pH 4·2, and then cooled to 4°C and bathed in ice-water. This cell-free culture supernatant was brought to a final ammonium sulfate concentration of 80% saturation, by the addition of the salt, and stirred overnight at 4°C. The suspension was then centrifuged at 5040 g for 30 min at 4°C and the surface pellicles and bottom pellets were harvested and resuspended in ultrapure water (MilliQ; Millipore, Billerica, MA, USA).
The resuspended fluid was adjusted to pH 4·2 and added to a hydrophobic column of Phenyl-TOYOPEARL 650 mol l−1 (Tosoh, Tokyo, Japan). The column was equilibrated and washed in 50 mmol l−1 of sodium acetate buffer (pH 4·0), saturated with 20% ammonium sulfate. Elution was carried out with a linear gradient of 20–0% ammonium sulfate at a flow rate of 2·0 ml min−1. The protein concentration of the fractions was determined spectrophotometrically at 280 nm.
Bacteriocin fractions that showed high antibacterial activities were further purified from extraneous proteins by solid-phase extraction on Sep-Pak C18 cartridges (Waters, Milford, MA, USA) (Garver and Muriana 1994). Eluted bacteriocin fractions were freeze-dried and stored at 4°C.
Molecular size approximation
The molecular size of the crude bacteriocin purified by Sep-Pak C18 cartridges was analysed by tricine sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Tricine-SDS-PAGE), following the method of Schagger and Von Jagow (1987). Bacteriocin size was estimated using RainbowTM coloured protein molecular weight markers (Amersham Biosciences, Piscataway, MA, USA), a low molecular weight marker consisting of seven proteins of known molecular weights (range 2·5–45 kDa). After electrophoresis, the gel was cut in two. Half of the gel was stained with Silver Stain Kit Wako (Wako, Osaka, Japan) for molecular weight determination. For direct detection of bacteriocin, the method of Bhunia and Johnson (1992), with some modifications, was used on the other half of the gel. The gel was washed, overlaid with MRS soft agar, seeded with the indicator strain Lact. sakei JCM1157T and incubated at 30°C overnight.
MALDI-time of flight MS and MS/MS analyses
Samples were prepared with methods described by Yanagida et al. (2000) and MALDI-TOF-MS spectra were recorded with MALDI-Qq-TOF MS/MS (Applied Biosystems, Lincoln Centre Drive Foster City, CA, USA). Results were searched with software of Mascot (Matrix Science, Boston, MA, USA).
N-terminal amino acid sequences analysis
Activity of the partially purified bacteriocin was confirmed on the SDS-PAGE gel, the gel was then blotted on polyvinyliden difluoride (PVDF)-membranes and stained with CBB R-250 (Wako). The objective bands were cut out and analysed using N-terminal amino acid analysis performed by Edman degradation on a protein sequencer (Protein sequencer 491; Applied Biosystems).
Thirty-four strains of acid-producing bacteria were isolated from 87 soil samples. Antibacterial activities of 34 isolates were detected, and strain L28-1 showed antibacterial activity to the indicator strain of Lact. sakei JCM1157T in an agar-well diffusion assay. Strain L28-1 was isolated from soil in the neighbourhood area of Kawaguchiko Lake (Yamanashi prefecture), and identified as Ent. durans using 16S rDNA sequencing analysis and the sugar fermentation tests (Devriese et al. 1993) of an API 50CHL identification kit (BioMérieux).
The cell-free culture supernatant of Ent. durans L28-1 was tested against the Gram-positive bacteria listed in Table 1. This culture supernatant lost its antibacterial activity completely after treatment with proteinase K, but maintained antibacterial activity after autoclaving (121°C for 15 min).
Table 1. Inhibition spectra for synergistic action of bacteriocins produced by Enterococcus durans L28-1
The purification of the bacteriocin produced by L28-1 was performed by ammonium sulfate precipitation and hydrophobic column chromatography, as described in the methods. The highest protein concentration was found in the fractions eluted in 0% ammonium sulfate (see Fig. 3). Fractions of high concentration were further purified from extraneous proteins by solid-phase extraction on Sep-Pak C18 cartridges and then freeze-dried. Bottom pellets were resuspended in ultrapure water, and then the molecular size of the crude bacteriocin was analysed by Tricine-SDS-PAGE. After electrophoresis, half of the gel, stained with silver stain kit, showed that the partially purified bacteriocin contained in numerous protein bands (Fig. 1). Two specific bands that displayed antibacterial activities were c. 3·4 kDa and 2·5 kDa in size (Fig. 1). Two specific clear zones were also observed in the other half of gel, overlaid on the plate with MRS soft agar seeded with indicator strain of Lact. sakei JCM1157T (Fig. 1).
Both antibacterial bands were cut out from a PVDF-membrane and then applied to the protein sequencer after blotting. A partial N-terminal amino acid analysis of band 1 (c. 3·4 kDa) revealed the following 21-amino acid sequence: NH2-Glu-Asn-Asp-His-Arg-Met-Pro-Tyr-Glu-Leu-Asn-Arg-Pro-Asn-Asn-Leu-Ser-Lys-Gly-Gly-Ala-. Amino acid sequencing analyses were performed three times and the results were confirmed. Results from the MALDI-TOF-MS analysis confirmed the amino acid sequences from position 6 to 18 (data not shown). Furthermore, primers based on the partial amino acid sequence were constructed and the sequence of the bacteriocin gene was determined by sequencing PCR products obtained with the methods described by Foulquié Moreno et al. (2003) and Casaus et al. (1997). The DNA sequencing confirmed the results of the amino acid sequencing from position 14 to 21. Furthermore, amino acid sequences from position 22 to 28 were obtained (data not shown). The 28 amino sequences are shown in Fig. 2. Antibacterial band 2 (c. 2·5 kDa) was also analysed; however, there were no clear signals obtained in the N-terminal amino acid analysis. The MALDI-TOF-MS analysis of antibacterial band 2 was not performed in this work.
A large number of bacteriocins of lactic acid bacteria have been described over the last few years. Frequent reports have been made of bacteriocins of lactic acid bacteria in milk products, fermented foods (Onda et al. 2003) and feed, but studies of bacteriocins from soil-isolated lactic acid bacteria remain scarce. Bacteriocins produced by Ent. faecium have been frequently reported: enterocin L50A and B (Toit et al. 2000), enterocin B (Casaus et al. 1997), enterocin CCM 4231 (Lauková and Czikková 1998), enterocin ON-157 (Ohmomo et al. 2000) and enterocin Q (Cintas et al. 2000). Bacteriocins of Ent. faecalis have also been frequently reported. However, studies of bacteriocins from Ent. durans are rare. Enterococcus durans L28-1 was originally isolated from soil, and in our present study, we showed that it produces a new bacteriocin. In a previous study, Ent. durans MY411 isolated from Miso-paste products (Onda et al. 2002) was found to produce a bacteriocin against some Gram-positive bacteria. Similar to the MY411 bacteriocin, two antibacterial bands were observed in the tricine SDS-PAGE profiles of the purified bacteriocin fluid. The amino acid sequences of MY411 are still unknown, and so it is difficult to compare the difference between the two bacteriocins. Using the agar-well diffusion assay and Ent. durans MY411 as the indicator strain, the supernatant fluid of Ent. durans L28-1 showed antibacterial activity to strain MY411. The results suggest that Ent. durans L28-1 and MY411 produce different bacteriocins.
The partial amino acid sequences indicated the similarities and difference between the antibacterial band 1 of L28-1 bacteriocins and enterocin B. One amino acid difference in L28-1 was shown: tyrosine at position 8 (EntB: asparagine). Tyrosine at position 8 was confirmed not only with N-terminal amino acid analysis but also with the MALDI-TOF-MS. MALDI-TOF-MS is effective for peptides and proteins with molecular masses ranging from 0·5 to 30 kDa and has been used to determine the masses of purified class I and II bacteriocins (Hindréet al. 2003).
No corresponding protein sequence was found, while 96% homology to enterocin B was obtained in the database (http://www.ddbj.nig.ac.jp/). Thus, we strongly suggest that Ent. durans L28-1 produces a novel, natural enterocin variant, and have termed it durancin L28-1A (Fig. 2).
In conclusion, the evidence presented in this report supports that the soil-isolated Ent. durans L28-1 produces a novel bacteriocin similar to enterocin B (Fig. 2). However, the individual characteristics of the two bacteriocins, such as heat stability, effect of enzymes and inhibition spectra, remained unclear. Items described above and full amino acid sequences of both bacteriocins produced by Ent. durans L28-1 will be analysed in the future. We anticipate that the bacteriocins of soil-isolated lactic acid bacteria can be used as food preservatives in the future.
We thank Dr Tsutomu Takayanagi for technical assistance and helpful advice.