Dr Medana Zamfir, Genetic Engineering Laboratory, Institute of Biology, Splaiul Independentei no. 296, 79651 Bucharest, Romania
M. ZAMFIR, R. CALLEWAERT, P.C. CORNEA, L. SAVU, I. VATAFU and L. DE VUYST.1999.Lactobacillus acidophilus IBB 801 produces a small bacteriocin, designated acidophilin 801, with an estimated molecular mass of less than 6·5 kDa. It displays a narrow inhibitory spectrum (only related lactobacilli but including the Gram-negative pathogenic bacteria Escherichia coli Row and Salmonella panama 1467) with a bactericidal activity. The antimicrobial activity of cell-free culture supernatant fluid was insensitive to catalase but sensitive to proteolytic enzymes such as trypsin, proteinase K and pronase, heat-stable (30 min at 121 °C), and maintained in a wide pH range. The proteinaceous compound was isolated from cell-free culture supernatant fluid and purified. Crude bacteriocin was isolated as a floating pellicle after ammonium sulphate precipitation (40% saturation) and partially purified by extraction/precipitation with chloroform/methanol (2/1, v/v). Further purification to homogeneity was performed by reversed phase Fast Performance Liquid Chromatography. The amino acid composition was determined. Amino acid sequencing revealed that the N-terminal end was blocked.
Lactic acid bacteria are traditionally used as starter cultures for the fermentation of foods and beverages because of their contribution to flavour and aroma development and to spoilage retardation ( Gilliland 1986). The preservative effect is mainly due to the acidic conditions that these bacteria create in food during their development, but they are capable of producing and excreting inhibitory substances other than lactic and acetic acid. These include hydrogen peroxide, ethanol, diacetyl, carbon dioxide, bacteriocin- or antibiotic-like substances, and bacteriocins ( De Vuyst & Vandamme 1994a). Bacteriocins are proteinaceous antibacterial compounds that inhibit Gram-positive bacteria, particularly closely related species ( Klaenhammer 1988; De Vuyst & Vandamme 1994b). Some of them are inhibitory towards food spoilage and food-borne pathogenic bacteria including Bacillus, Clostridium, Staphylococcus and Listeria. Therefore, bacteriocins of lactic acid bacteria are of particular interest because of their existing and potential applications as natural preservatives in foods ( Holzapfel et al. 1995 ; Delves-Broughton et al. 1996 ; Stiles 1996) and as genetic markers in food-grade cloning and expression systems ( Allison & Klaenhammer 1996; Platteeuw et al. 1996 ).
Based on their primary structure, molecular mass, heat stability and molecular organization, bacteriocins produced by lactic acid bacteria can be subdivided into four classes ( Klaenhammer 1993): class I, the lantibiotics ( Jack et al. 1995 , 1998; Sahl et al. 1995 ; Konings & Hilbers 1996); class II, the non-lantibiotic peptides ( Nes et al. 1996 ), which are divided into the subgroups IIa or pediocin-like bacteriocins with strong antilisterial activity, IIb bacteriocins whose activity depends on the complementary action of two peptides, and IIc sec-dependent secreted bacteriocins; class III, large, heat-labile protein bacteriocins; and class IV, bacteriocins claimed to consist of an undefined mixture of protein(s), lipid(s) and carbohydrate(s).
Recently, a bacteriocin from Lact. acidophilus IBB 801, a dairy strain with potential probiotic properties, was found ( Cornea et al. 1996 ). In this paper, the bacteriocin, designated acidophilin 801, is isolated, purified and partially characterized.
MATERIALS and METHODS
Bacterial strains and culture media
Lactobacillus acidophilus IBB 801, a bacteriocinogenic dairy strain, was used throughout this study ( Cornea et al. 1996 ). The other bacterial strains used in this study are listed in Table 1. Lactic acid bacteria were maintained in lyophilized form. Before experimental use, the cultures were propagated twice in de Man Rogosa Sharpe (MRS) broth ( de Man et al. 1960 ). The pathogenic indicator strains were maintained on nutrient agar slants and were sub-cultured in Luria Bertani (LB) broth ( Sambrook & Fritsch 1989) . Listeria strains were grown on Listeria enrichment broth base (LEB, Merck).
Table 1. Bacterial strains used in this study and their sensitivity to the bacteriocin produced by Lactobacillus acidophilus IBB 801
For bacteriocin production, the Lact. acidophilus IBB 801 strain was grown in MRS broth. Since the best inhibitory activity was found with Lact. helveticus 102 as indicator strain, it was used for detection and quantitative determination of the bacteriocin activity of Lact. acidophilus IBB 801. As solid media, Micro Assay Culture Agar (MACA; Difco) or MRS broth containing 1·5% agar (Oxoid) were used. For a soft agar top layer, MRS was supplemented with 0·7% agar.
Inhibitory spectrum, inhibitory effect and quantitative determination of bacteriocin activity
To establish the inhibitory spectrum of the bacteriocin, cell-free culture supernatant fluid adjusted to pH 6·5, derived from a Lact. acidophilus IBB 801 culture incubated for 12 h at 37 °C in MRS broth, was spotted onto indicator lawns of several lactic acid bacterium strains and other Gram-positive and Gram-negative bacteria ( Table 1) to observe possible growth inhibition. These lawns were prepared by propagating fresh bacterial cultures to an optical density of 0·45 (600 nm) and adding 200 μl of the cell suspension to 3·5 ml of overlay agar (top layer). Overlaid agar plates were incubated for 24 h at 37 °C.
To determine the inhibitory effect of acidophilin 801 towards a sensitive strain, potassium phosphate buffer (50 mmol l−1, pH 6·5) containing partially purified (see further) bacteriocin at a final concentration of 500, 250, 125 and 62·5 AU ml−1 was inoculated with an exponentially-growing culture of Lact. helveticus 102 (the most sensitive strain). Alternatively, partially purified bacteriocin at a final concentration of 500 AU ml−1 was added to MRS broth that was first incubated for 3 h with the sensitive indicator strain. Growth was followed by plate counting on MRS agar.
Bacteriocin activity was assayed quantitatively by an agar spot test ( De Vuyst et al. 1996a ). Briefly, serial twofold dilutions in 50 mmol l−1 potassium phosphate buffer of cell-free culture supernatant fluid containing bacteriocin were spotted (10 μl) onto fresh indicator lawns of Lact. helveticus 102. The activity was defined as the reciprocal of the highest dilution which demonstrated complete inhibition of the indicator lawn and was expressed in activity units (AU) per millilitre of culture medium.
Preliminary characterization of the bacteriocin
To characterize the bacteriocin, the producer strain was cultivated in MRS broth for 12 h at 37 °C. Cells were removed by centrifugation (13 000 g, 10 min, 4 °C) and the cell-free culture supernatant fluid was considered to be crude bacteriocin. First, the influence of the proteases trypsin, proteinase K and pronase on bacteriocin activity was tested. All enzymes were dissolved in 3 mmol l−1 potassium phosphate buffer (pH 7·5) and added to the crude bacteriocin sample (1600 AU ml−1) at a final concentration of 0·5 mg ml−1. After incubation for 1 h at 37 °C, the inhibitory activity was tested. Cell-free culture supernatant fluid was also treated with catalase (final concentration of 5 mg ml−1) dissolved in 50 mmol l−1 potassium phosphate buffer (pH 7·0) at 25 °C in order to completely eliminate possible inhibitory activity due to hydrogen peroxide. Finally, the heat sensitivity and pH stability were tested. The cell-free culture supernatant fluid (containing 1600 AU ml−1 bacteriocin activity) was heated at 70 °C for 15, 30, 45 and 60 min, at 100 °C for 15, 30 and 45 min, and at 121 °C for 30 min, before testing the activity. To test the influence of pH, the cell-free culture supernatant fluid was adjusted to a pH of 3·0, 4·0, 5·0, 6·0, 7·0, 8·0, 9·0 or 10·0 with HCl or NaOH, mixed, and allowed to stand for a few minutes before testing the inhibitory activity. The activity was checked as described above, except that either a top layer of MRS agar (0·7%), or of MRS agar (0·7%) containing 1·0 mg ml−1 (final concentration, m/v) of trypsin or pepsin, was used.
Bacteriocin isolation, purification to homogeneity, and amino acid analysis and sequencing
For bacteriocin isolation, the bacterial strain was cultivated for 12 h in 2 l of MRS broth at 37 °C. Cells were removed by centrifugation and the pH of the cell-free culture supernatant fluid was adjusted to 6·5. The supernatant fluid was then precipitated with ammonium sulphate (40% saturation) overnight at 4 °C with gentle stirring. A floating pellicle was formed. After centrifugation, the pellicle was collected and dissolved in 15 ml phosphate buffer (pH 6·5) and then extracted with 15 volumes of a mixture of chloroform/methanol (2/1, v/v). After l h at 4 °C, the white precipitate formed was centrifuged for 1 h at 13 000 g and resuspended in 3 ml ultrapure water.
To estimate the molecular mass of the bacteriocin, tricine-sodium dodecyl sulphate-polyacrylamide gel electrophoresis (tricine-SDS-PAGE) was carried out, according to the method of Schägger & von Jagow (1987). Polyacrylamide concentrations in the stacking gel and separating gel were 9·6% and 16·0%, respectively. Electrophoresis was conducted at a constant voltage of 30 V for 18 h. After the run, the gel was washed during 5 h with sterile ultrapure water that was replaced every hour. Finally, the gel was transferred to a MACA plate and overlaid with a top layer of soft MRS agar inoculated with the indicator strain, Lact. helveticus 102. A part of the gel was stained with copper(II)chloride, as described by Tessmer & Dernick (1990). Another part, containing the standard proteins (low-molecular-mass range Sigma Marker protein standards), was stained with Coomassie Brilliant Blue.
The final purification step was a reversed phase fast performance liquid chromatography (FPLC) run. The chloroform/methanol extract was injected into a 1 ml Pep RPC HR 5/5 C2/C18 column (Pharmacia, Uppsala, Sweden). As mobile phases, 10% isopropanol with 0·1% trifluoroacetic acid (TFA) as solvent A, and 100% isopropanol with 0·1% TFA as solvent B, were used. The bacteriocin was eluted with a step gradient from 10 to 100% of solvent B. The absorption was measured both at 210 and 280 nm. Fractions of 0·5 ml were collected at a rate of 0·5 ml min−1, and their activity was tested against Lact. helveticus 102 as indicator organism.
Amino acid composition analysis was carried out by hydrolysis of partially purified bacteriocin in 6 N HCl under vacuum at 110 °C for 24 h. The amino acids were derivatized with phenylisothiocyanate, and the PTC amino acids were separated by reverse-phase HPLC. The N-terminal amino acid sequencing was performed by automated Edman degradation with a model 477 A sequencer/model 120 A phenylthiohydantoin analyser (Applied Biosystems, Perkin Division, Foster City, CA, USA).
After 12 h of cultivation of Lact. acidophilus IBB 801 in MRS broth at 37 °C, the producer cells were removed, the pH of the cell-free culture supernatant fluid was adjusted to pH 6·5 and the inhibitory activity towards different bacterial strains was tested. The results are presented in Table 1. The inhibitory spectrum of Lact. acidophilus IBB 801 is quite narrow, being active particularly towards closely related lactobacilli, but it also includes two Gram-negative pathogenic strains (Escherichia coli Row and Salmonella panama 1467). In view of its highest sensitivity, Lact. helveticus 102 was further used for quantitative determinations of bacteriocin activity.
Preliminary characterization of crude bacteriocin
To determine the proteinaceous nature of the antimicrobial substance, the effect of some proteolytic enzymes (trypsin, pronase and proteinase K) on the inhibitory activity was tested. Also the influence of catalase was tested. Incubation of the samples for 1 h at 37 °C with these enzymes completely destroyed the antimicrobial activity, except for catalase, indicating that the inhibitory material was of a proteinaceous nature.
The sensitivity of the inhibitory compound to heat treatment was tested ( Table 2). The inhibitory activity was not significantly altered by heat treatment. After 60 min at 70 °C, the activity was about 1000 AU ml−1 and after 45 min at 100 °C, the activity decreased only to 800 AU ml−1. A small amount of inhibitory activity was observed even after 30 min at 121 °C. These results suggest that the antimicrobial substance produced by Lact. acidophilus IBB 801 is strongly heat resistant.
Table 2. The effect of different physical and biochemical treatments on bacteriocin activity
Also, the effect of pH (3·0, 4·0, 5·0, 6·0, 7·0, 8·0, 9·0 and 10·0) on bacteriocin activity was tested in the presence and absence of trypsin and pepsin in the MRS top agar layer. All samples showed inhibitory activity only in the absence of protease in the top layer, indicating that the inhibition was not due to lactic acid and that the bacteriocin was active in a wide pH range. When a top layer with trypsin or pepsin was applied, the inhibitory activity was lost, indicating that the inhibitory material was indeed of a proteinaceous nature.
Purification of the bacteriocin produced by Lact. acidophilus IBB 801
The cell-free culture supernatant fluid (1600 AU ml−1) was precipitated with 40% ammonium sulphate saturation. The bacteriocin activity of the pellicle increased to 25 600 AU ml−1. After chloroform/methanol extraction, the activity of the precipitate, re-dissolved in ultrapure water, was 51 200 AU ml−1. This material was considered to be a partially purified bacteriocin preparation.
The partially purified bacteriocin was boiled for 5 min with one volume of sample buffer and then subjected to tricine-SDS-PAGE for 18 h. After 5 h of washing with sterile ultrapure water, the gel was overlaid with soft MRS agar containing Lact. helveticus 102 as sensitive indicator organism. After 24 h of incubation at 37 °C, a clear inhibition zone, corresponding to a molecular mass of less than 6500 Da, could be observed ( Fig. 1a). Whereas staining with Coomassie Brilliant Blue did not give a clear peptide band (result not shown), the copper(II)chloride staining procedure revealed a peptide of less than 6500 Da ( Fig. 1b).
The partially purified bacteriocin was further purified by reversed phase FPLC. The sample was diluted with two volumes of a mixture of four parts of solvent A (10% isopropanol + 0·1% trifluoroacetic acid) and one part of solvent B (100% isopropanol + 0·1% trifluoroacetic acid) and centrifuged (13 000 g, 15 min) before injection. A step gradient from 10 to 100% of solvent B was applied (see Fig. 2). Fractions of 0·5 ml were collected and their activity tested towards the indicator strain. The purified bacteriocin eluted at about 55% of isopropanol with a high absorption peak at 210 nm ( Fig. 2). Two fractions, corresponding with this peak, were active. They were pooled and analysed for amino acid composition and N-terminal amino acid sequencing. The amino acid composition revealed the presence of many hydrophobic amino acids, averaging 50% of the total amino acid content, as well as the presence of one unidentifiable residue. N-terminal amino acid sequencing was not successful, most probably because the N-terminal end was blocked.
Inhibitory effect of the bacteriocin produced by Lact. acidophilus IBB 801
The addition of partially purified bacteriocin to a Lact. helveticus 102 culture did not permit growth of this strain ( Fig. 3a). Moreover, after 1 h, the bacterial count was approximately 1–2 log cycles lower than in the control, and, after 5 h, it was approximately 1, 3 and 4 log cycles lower after addition, respectively, of 62·5, 250 and 500 AU ml−1. The addition of 500 AU ml−1 of partially purified acidophilin 801 to a Lact. helveticus 102 culture after 3 h of growth ( Fig. 3b) resulted in a reduction of colony counts by 3 log cycles within 4 h, compared with the control. These observations showed the concentration-dependent bactericidal effect of acidophilin 801 towards a bacteriocin-sensitive strain.
A bacteriocin, designated acidophilin 801, produced by Lact. acidophilus IBB 801, has been partially characterized. It is sensitive to several proteases and inhibition occurs under conditions which eliminate the effects of organic acids and hydrogen peroxide. Like most of the known bacteriocins produced by Lact. acidophilus strains, acidophilin 801 is a heat-stable and low-molecular-mass (less than 6500 Da) peptide bacteriocin. Acidophilin 801 displayed a concentration-dependent bactericidal effect towards a bacteriocin-sensitive strain (Lact. helveticus 102) without causing concomitant cell lysis of the indicator cells. Amino acid composition analysis of acidophilin 801 revealed a strongly hydrophobic peptide. However, N-terminal amino acid sequencing was not possible, probably because the N-terminal end was blocked. This may indicate the presence of modified amino acids, for instance as a result of post-translational modifications. It was further found that acidophilin IBB 801 activity is stable at a wide pH range (3·0–10·0). Taking all these biochemical characteristics into consideration, acidophilin 801 produced by Lact. acidophilus IBB 801 appears to belong to the class II lactic acid bacterium bacteriocins according to the classification of Klaenhammer (1993).
Strong hydrophobicity is a characteristic property of the class II peptide bacteriocins. For instance, the proportion of hydrophobic amino acids in the Lact. acidophilus group bacteriocins, namely acidophilin 801, gassericin A, lactobin A and lactacin F, amounts to 50·8%, 45·7%, 38·0% and 29·0%, respectively ( Tahara & Kanatani 1996; Tahara et al. 1996 ; Nissen-Meyer & Nes 1997). This strong hydrophobic character is one of the main reasons why bacteriocin purification is tedious and cumbersome. The number of chromatographic steps varies from three or more (e.g. acidocin LF221 A and B, Bogovic-Matijašic et al. 1998 ) to only one (this paper) after concentration of the bacteriocin from the cell-free culture supernatant fluid either via ammonium sulphate precipitation (e.g. lactacin F, Muriana & Klaenhammer 1991a) or through ammonium sulphate precipitation followed by chloroform/methanol extraction (this study). Whereas most of the bacteriocins could be purified to homogeneity by reversed phase high performance liquid chromatography with a C18 column, gassericin A could only be purified with a low-hydrophobic C4 column, due to its extremely hydrophobic nature. The method reported here shows that purification of extremely hydrophobic peptide bacteriocins can be easily performed with a C18 column, too, provided appropriate pre-treatment steps are taken.
The strain Lact. acidophilus IBB 801 produces a bacteriocin characterized by a limited inhibition spectrum, being active only against some lactobacilli. It is, however, remarkable that inhibition towards the Gram-negative, pathogenic bacteria E. coli Row and Salmonella panama 1467 also occurs. In general, bacteriocins from lactic acid bacteria are only active towards Gram-positive bacteria ( Klaenhammer 1988). When injured sub-lethally, Gram-negative bacterial cells or spheroplasts may also be sensitive ( Stevens et al. 1991 ; Kalchayanand et al. 1992 ). A very narrow inhibitory spectrum, as observed here with Lact. acidophilus IBB 801, seems to be common among bacteriocin-producing isolates from the Lact. acidophilus group. For instance, lactacin F from Lact. johnsonii VPI 11088 (previously classified as Lact. acidophilus 11088), displays bactericidal activity towards other lactobacilli (six species) and Enterobacter faecalis (Muriana & Klaenhammer 1991a). Also, acidocin J1229 ( Tahara & Kanatani 1996), acidocin 8912 ( Kanatani et al. 1992 ), acidocin J1132 ( Tahara et al. 1996 ), amylovorin L471 ( De Vuyst et al. 1996a ), crispacin A ( Tahara & Kanatani 1997) and lactobin A ( Contreras et al. 1997 ) only inhibit a few Lactobacillus spp. In contrast, acidocin A ( Kanatani et al. 1995 ), gassericin A ( Itoh et al. 1995 ) and acidocins LF221A and LF221B ( Bogovic-Matijašic et al. 1998 ) are active against closely-related lactic acid bacteria and food-borne pathogens including Listeria monocytogenes. Finally, acidocin B affects growth of only a few lactobacilli (Lact. fermentum, Lact. delbrueckii subsp. bulgaricus), while it inhibits Brochothrix thermospacta, Clostridium perfringens and Listeria monocytogenes (ten Brink et al. 1994). Even acidophilucin A, produced by Lact. acidophilus LAPT 1060, and lactacin B, produced by Lact. acidophilus N2, belonging to the class III and IV bacteriocins, respectively, display a narrow spectrum of activity, and only lactobacilli (four and two species, respectively) were found to be sensitive ( Barefoot & Klaenhammer 1983; Toba et al. 1991b ). On the other hand, Juven et al. (1992) isolated a bacteriocin, designated LA-147, from a chicken intestinal Lact. acidophilus strain that showed inhibitory activity only against strains of Lact. lactis (formerly Lact. leichmanii) and not against several other species of Lactobacillus or other selected Gram-positive and Gram-negative bacteria. The other extremes are, for instance, acidolin ( Hamdan & Mikolajcik 1974) and acidophilin ( Shahani et al. 1977 ), both produced by Lact. acidophilus strains, that are active against lactic acid bacteria and a wide range of other Gram-positive and Gram-negative bacteria. However, the nature of these antimicrobial component(s) remains to be determined. Many of the Lact. acidophilus antimicrobial, low-molecular-mass, proteinaceous substances with wide inhibition spectra, including both Gram-positive and Gram-negative bacteria, have been classified as small, heat-stable, bacteriocin-like peptides (for a review, see De Vuyst & Vandamme 1994b). The fact that the antimicrobial nature is often unknown also holds true for well known probiotic Lact. acidophilus strains commercially applied. For example, the adhering human Lact. acidophilus LA1 strain inhibits cell attachment and cell invasion by enterovirulent bacteria, and produces antibacterial activity insensitive to proteases and active in vitro against a wide range of Gram-negative and Gram-positive pathogens, such as Staphylococcus aureus, Listeria monocytogenes, Salmonella typhimurium, Shigella flexneri, Klebsiella pneumoniae, Pseudomonas aeruginosa and Enterobacter cloacae ( Bernet-Camard et al. 1997 ). By contrast, no activity was observed against species of the normal gut flora, such as lactobacilli and bifidobacteria. However, the nature of the antimicrobial component(s) secreted by Lact. acidophilus LA1 remains to be determined. Another well studied Lactobacillus probiotic strain is Lact. rhamnosus GG (formerly Lact. casei GG). Like LA1, it produces, most probably, a non-proteinaceous substance with antimicrobial activity ( Silva et al. 1987 ) that acts against Salm. typhimurium to block adhesion to Caco-2 cells in vitro and protect against infection in mice ( Hudault et al. 1997 ). Finally, Lact. acidophilus LB, a human strain adhering to cultured human polarized intestinal cells, displays antagonistic activity in vitro against non-invasive and invasive enterovirulent pathogens, and in vivo against Salm. typhimurium, Campylobacter and Helicobacter pylori infections ( Coconnier et al. 1998 ). The LB strain secretes an antimicrobial substance(s) other than lactic acid that is heat stable and only moderately sensitive to enzymatic treatments. Moreover, several characteristics of the component(s) supporting the antimicrobial activity suggest that it could contain an unusual acidic amino acid present in a novel peptidic agent. It therefore has still to be confirmed whether bacteriocins from probiotic lactic acid bacteria are among the compounds responsible for inactivation of pathogens in intestinal tracts. Hence, it cannot be excluded that acidophilin 801 also belongs to a novel class of peptidic agents, given its antimicrobial activity towards two Gram-negative species and the difficulties of unravelling its amino acid sequence. Both bacteriocins and bacteriocin-like peptides will contribute to the inhibitory action towards harmful intestinal bacteria of (potential) probiotic lactic acid bacterium strains.
Part of the research presented in this paper was financially supported by the Copernicus Program of the Commission of the European Community (grant ERB-CIPACT 940160). RC and LDV also acknowledge their finances from the Research Council of the Vrije Universiteit Brussel and the Fund for Scientific Research, Flanders.