In Vitro Anti-Helicobacter pylori Activity of the Probiotic Strain Bacillus subtilis 3 Is Due to Secretion of Antibiotics

Abstract

A limited number of antibiotics can be used against Helicobacter pylori infection, and resistance jeopardizes the success of treatment. Therefore, a search for new agents is warranted. The use of probiotics to enhance gastrointestinal health has been proposed for many years, but the scientific basis of the prophylactic and therapeutic actions of probiotics has not yet been clearly delineated. Probiotic strain Bacillus subtilis 3, whose safety has previously been demonstrated, is known to have antagonistic properties against species of the family Enterobacteriaceae. In the present study, it was also found to inhibit H. pylori. The anti-H. pylori activity present in the cell-free supernatant was not related to pH or organic acid concentration. It was heat stable and protease insensitive. At least two antibiotics, detected by thin-layer chromatography (R_f values, 0.47 and 0.85, respectively) and confirmed by high-performance liquid chromatographic analysis, were found to be responsible for this anti-H. pylori activity. All H. pylori strains tested were sensitive to both compounds. One of these compounds was identified as amicoumacin A, an antibiotic with anti-inflammatory properties. MICs for H. pylori determined in solid and liquid media ranged between 1.7 and 6.8 μg/ml and 0.75 and 2.5 μg/ml, respectively. The underestimation of MICs determined in solid medium may be due to physicochemical instability of the antibiotic under these test conditions. An additive effect between amicoumacin A and the nonamicoumacin antibiotic against H. pylori was demonstrated.

Helicobacter pylori infection is the major cause of chronic gastritis and peptic ulcer disease (16) and is a risk factor for gastric cancer in humans (8). This infection is extremely common throughout the world, and its prevalence increases with age and lower socioeconomic status.

Combinations of several drugs are now widely used for the eradication of H. pylori. Numerous clinical trials have indicated that eradication of H. pylori by treatment that includes bismuth or antisecretory drugs combined with antibiotics leads to healing of gastritis and drastically decreases the rate of peptic ulcer relapse. However, a limited number of antibiotics can be used, and resistance jeopardizes the success of treatment (15, 19). Therefore, a search for new antimicrobial agents is warranted.

Several authors have previously reported that certain probiotic bacteria, such as Lactobacillus spp., exhibit inhibitory activity against H. pylori in vitro and in vivo (5, 17). The term “probiotic” was first used by Parker (21) to describe “organisms and substances which contribute to the intestinal microbial balance.” This definition has since been narrowed to focus on normal microflora that have a beneficial effect in the prevention and treatment of a diverse spectrum of gastrointestinal disorders such as antibiotic-associated diarrhea and bacterial and viral diarrhea (including diarrhea caused by Shigella spp., Salmonella spp., rotavirus, and others) (6, 23). Currently, the most well-studied probiotics are the lactic acid bacteria, particularly lactobacilli, and to a lesser extent, bifidobacteria. Other organisms used or considered for use as probiotics in humans or animals include various Bacillus spp., Bacteroides spp., Propionibacterium spp., and fungi (2, 23, 26). Characterization and study of the mechanisms of action of these microorganisms as probiotics lag significantly behind characterization and study of those of the lactic acid bacteria.

Probiotic strain Bacillus subtilis 3 is of interest because of its safety for use in humans (20) and its antagonistic effect in vitro and in vivo against different human bacterial pathogens such as Campylobacter spp. (26). Recent reports have suggested that the antimicrobial activity of Lactobacillus sp. probiotics involves production of secreted compounds, such as organic acid, hydrogen peroxide, as well as various antibiotics or bacteriocins (1, 5, 12, 25, 27). Therefore, the aim of the present study was to determine the antagonistic activity of probiotic strain B. subtilis 3 against H. pylori in vitro and the cause of any inhibitory activity.

MATERIALS AND METHODS

Bacteria.

All except three of the bacterial strains used as test cultures (see Tables 1 and 2) were grown in Mueller-Hinton (MH) medium at 37°C under aerobic conditions. H. pylori, Campylobacter jejuni, and Bacteroides fragilis strains, however, were cultured on MH agar (Difco Laboratories, Detroit, Mich.) supplemented with 10% defibrinated horse blood (MH&B) and were grown for 48 to 72 h under microaerobic (for H. pylori and C. jejuni) or anaerobic (for B. fragilis) conditions.

TABLE 1.

MICs of the purified amicoumacin A produced by B. subtilis 3 for different H. pylori strains

H. pylori strain	MIC (μg/ml) in:
	Liquid mediuma	Solid mediumb
1382	2.5	6.8
1079	1.0	3.4
AM A900211	1.0	3.4
Azm1 P079	0.75	1.7
9008 704F	2.0	3.4
Az 1099	0.75	3.4
DG 1021	1.0	3.4
SS1	1.0	3.4
AA9002 11	1.5	1.7
1470	2.0	3.4
1289	2.5	6.8
815C3	1.0	1.7
704RG 9001	1.0	3.4
704AD 9001 211	0.75	1.7
1060	0.75	1.7
SM 1028 Az	2.0	3.4
A 12211	0.75	1.7
Azm A9002 11	1.0	3.4
NC 012 455	2.0	3.4
AT 9008 704F 9	2.5	6.8
J99	2.0	6.8

^aDulbecco modified Eagle's medium supplemented with 10% fetal bovine serum. 

^bMH&B. 

TABLE 2.

MICs of purified amicoumacin A produced by B. subtilis 3 for different pathogenic and nonpathogenic intestinal microorganisms

Strain tested	MICsa of amicoumacin A (μg/ml)
B. fragilis LMBA 19031	>25.00
C. jejuni LMBA 2568	1.25
E. faecium LMBAb 25347	0.75
E. faecium LMBA 27323	1.00
E. coli LMBA 3547	>25.00
E. coli LMBA 20684	10.00
Lactobacillus acidophilus ATCC 4356T	>25.00
L. acidophilus LMBA 456	>25.00
Lactobacillus plantarum ATCCc 8014	>25.00
Lactococcus lactis ATCC 11454	>25.00
L. lactis LMBA 374	25.00
S. flexneri LMBA 12225	0.75

^aMIC of amicoumacin A was determined in appropriate liquid medium. 

^bLMBA, Laboratoire de Microbiologie et Biochimie Appliquée, ENITA de Bordeaux, France. 

^cATCC, American Type Culture Collection. 

Probiotic strain B. subtilis 3, originally isolated from animal fodder (IMV National Academy of Sciences, Kiev, Ukraine), is one of the basic components of the commercial probiotic Biosporin (Biofarm, Dniepropetrovsk, Ukraine). It was grown in a liquid starch medium composed of starch (10 g/liter), peptone (5 g/liter), and NaCl (0.5 g/liter) for 72 h at 28°C in an aerobic atmosphere on a rotary shaker at 175 rpm.

All strains were stored at −70°C in brucella broth (Difco Laboratories) containing 10% (vol/vol) glycerol.

Antimicrobial testing. (i) Spotting method.

B. subtilis 3 was inoculated as a spot (approximately 5 to 10 mm) on the surface of a starch agar (1.2% [wt/vol]) plate. Controls consisted of uninoculated starch agar plates. After 72 h at 28°C, the bacteria were killed by exposure to chloroform vapor for 15 min, and the plates were overlaid with MH&B containing 24- to 48-h-old H. pylori strains at 10⁸ CFU/ml (adjusted by turbidimetry). All plates were incubated for 24 to 48 h at 37°C in an appropriate atmosphere. The antagonistic activity was detected as the presence of a growth inhibition zone around the spot.

(ii) Assay for antibacterial substance production.

Assessment of the anti-H. pylori activities of culture supernatants and step purification fractions was performed by an agar well diffusion method (5) with MH&B. The inoculum of H. pylori strains was approximately 10⁸ CFU/ml. Aliquots of 100 μl of these suspensions were inoculated onto agar medium. The wells (diameter, 6 mm) were then cut into each plate. Serial dilutions of each test solution were made, and the wells were filled with 100 μl of native or diluted solution. The plates were incubated at 37°C in a microaerobic atmosphere for 24 to 48 h. The inhibition zones of H. pylori were then measured.

The titer of anti-H. pylori was defined as the reciprocal of the highest dilution showing inhibition of H. pylori and was expressed in activity units (AU) per milliliter.

(iii) Viability of H. pylori SS1.

The viability of H. pylori Sydney strain 1 (SS1) in the presence of cell-free supernatant, solid-phase extract, or purified fractions of the antimicrobial substances was determined by the method described by Bernet-Camard et al. (3). Briefly, an 18-h-old H. pylori culture was resuspended in phosphate-buffered saline (PBS) and centrifuged at 5,500 × g for 5 min. The supernatant was discarded, and the bacteria were washed once with PBS and resuspended (approximately 10⁷ CFU/ml) in medium containing 50% (vol/vol) brain heart infusion broth (Difco Laboratories), 40% (vol/vol) Dulbecco modified Eagle's medium (Life Technologies, Cergy-Pontoise, France), and 10% (vol/vol) fetal bovine serum. Then, 4.5 ml of this suspension was incubated with 0.5 ml of a cell-free supernatant, solid-phase extract, or purified fraction of antimicrobial substances at 37°C under microaerobic conditions. In this experiment, the dried solid-phase extract or purified fractions were dissolved in 0.5 ml of starch medium and were used at a final concentration of 200 AU/ml. Initially and then at predetermined intervals, aliquots of the incubated suspension were removed, serially diluted, and plated onto MH&B to determine bacterial colony counts.

Physicochemical and enzymatic treatments.

For thermostability assay, lyophilized cell-free supernatant (20 mg/ml in 5 mM Tris HCl buffer [pH 7.0]) was submitted to heat treatment for 15, 30, 45, and 60 min at 100°C. For pH stability, the preparation (20 mg/ml) was incubated for 1 h at 37°C in appropriate buffer solutions (5 mM Tris or 5 mM CH₃COONa adjusted to different pH values with HCl or CH₃COOH, respectively). Sensitivity to enzymes was tested by incubation of the preparation (20 mg/ml in 5 mM Tris HCl buffer [pH 8.3]) for 1 h at 50°C for proteases or at 37°C for lipase, α-amylase, and lysozyme. The enzymes were used at final concentrations of 1 mg/ml. The anti-H. pylori activity was determined before and after each treatment by the agar well diffusion method.

Isolation and purification.

Cultures of B. subtilis 3 grown for 72 h in a starch liquid medium at 28°C in a volume of 8 liters were clarified by filtration through diatomaceous earth (Spendalite N; Spindal, Paris, France). The antimicrobial substances in cell-free filtrates were batch adsorbed on Amberlite XAD7 polyacrylic ester resin (Sigma-Aldrich Chimie, St. Quentin Fallavier, France). After washing of the resin with water and H₂O-isopropanol (80:20 [vol/vol]), antimicrobial compounds were eluted with H₂O-isopropanol (30:70 [vol/vol]) and were then concentrated under vacuum. The dry residue (solid phase extract) was dissolved in H₂O-acetonitrile (90:10 [vol/vol]) containing 0.1% trifluoroacetic acid (TFA). This solvent mixture was used as solvent A. The dissolved residue was loaded onto a Nova Pak HR C₁₈ (Waters Div., Millipore Corp., St-Quentin-Yvelines, France) semipreparative reverse-phase high-pressure liquid chromatography (HPLC) column (25 by 100 mm; particle size, 6 μm; pore size, 60 Å) equilibrated at a 18 ml/min flow rate with solvent A. Activity was eluted with a two-step linear gradient of acetonitrile containing 0.07% TFA (solvent B): from 0 to 40 min with 0 to 19% solvent B and then from 40 to 50 min with 19 to 100% solvent B.

Active fractions were pooled, freeze-dried, resuspended in solvent A, and then processed on the same system by using an 80-min linear gradient with from 0 to 20% solvent B. Active fractions (purified antimicrobial substance) were collected, freeze-dried, and stored at −80°C.

All purification steps were performed at room temperature with a 600 E HPLC system (Waters Division, Millipore Corp.); the absorbance of the eluate was detected at 210 nm.

TLC.

Thin-layer chromatography (TLC) analysis of cell-free supernatants and purified antimicrobial fractions was performed on precoated Kieselghur 60 plates (20 by 20 cm; Macherey Nagel, Duren, Germany) with chloroform-methanol-H₂O (65:25:4 [vol/vol/vol]) as a mobile phase. The bands on the TLC chromatogram were marked under UV light at 310 nm. The different bands were scraped separately, dissolved in 1 ml of H₂O-isopropanol (30%/70% [vol/vol]), vortexed, and centrifuged for 15 min at 6,500 × g. The anti-H. pylori activities of the extracts that were obtained were tested by the agar well diffusion method.

Reverse-phase HPLC.

Reverse-phase HPLC on a Lichrospher C₁₈ column (4 by 125 mm; particle size, 5 μm; pore size, 100 Å; Merck, Darmstad, Germany) was used to analyze solid-phase extracts and the purified substance. After sample injection, the column was eluted for 5 min at a flow rate of 1 ml/min with H₂O-acetonitrile (90:10 [vol/vol]) containing 0.1% TFA (solvent A) and was then eluted with a linear gradient with from 0 to 20% solvent B for 30 min. Antibiotic substances in the eluate were detected with a fluorescence detector (Waters Division, Millipore Corp.); the excitation wavelength selected was 312 nm, and emission was measured at 465 nm.

Physicochemical properties.

UV-visible antibiotic spectral data were recorded on a 930 Uvikon spectrophotometer (Bio-Tek Instruments, St. Quentin en Yvelines, France). ¹H and ¹³C nuclear magnetic resonance (NMR) spectra of antimicrobial compounds were acquired on a Bruker Spectrospin DRX 500 spectrometer (Bruker Canada Ltd., Milton, Ontario, Canada).

Molecular mass was determined by mass spectrometry in the electrospray mode with a SIEX API 100 spectrometer (Perspective Biosystem, Framingham, United Kingdom). Exact molecular mass determination was performed on a high-resolution chemical ionization mass spectrometer (JMS-AX505; JEOL USA Inc., Peabody, Mass.).

Determination of MICs.

The MIC of the purified antimicrobial substance for different bacterial strains were determined by a dilution method (18) in solid and/or liquid media containing different concentrations of purified antimicrobial substance. The bacterial inoculum was approximately 10⁷ CFU/ml. The MIC was considered to be the lowest concentration of antimicrobial substance that caused total inhibition of bacterial growth.

RESULTS

Inhibition of growth of H. pylori in vitro by B. subtilis 3 and its supernatant.

Of the 21 strains of H. pylori tested, all were inhibited by B. subtilis 3. When the cell-free supernatant of B. subtilis 3 was tested, the inhibitory zone for each H. pylori isolate ranged from 10 to 16 mm. The diameters of the inhibition zones increased significantly (15 to 25 mm) when the supernatant was concentrated approximately 10-fold.

Effects of heat, pH, and various enzymes on antimicrobial substance(s).

The antimicrobial compounds present in the cell-free supernatant were heat stable, and the level of activity remained unchanged after 60 min of incubation at 100°C. This antimicrobial supernatant was also active over a wide range of pH values, from 3 to 10. The activity decreased slightly (20%) at pH 11. Treatment with different enzymes (pepsin, trypsin, proteinase K, lipase, α-amylase, and lysozyme) did not affect the antimicrobial activity of the supernatant.

Study of antibiotic production by TLC analysis.

TLC of concentrated cell-free supernatant on silica gel plates yielded five major bands under UV light at 310 nm, with R_f values of 0.2, 0.43, 0.47, 0.8, and 0.85. These bands were submitted to an extraction procedure, and the activity of each band against H. pylori strains was measured. Only two of them, those with R_f values of 0.47 and 0.85, respectively, contained anti-H. pylori compounds.

Isolation and characterization.

The antimicrobial compounds present in the cell-free filtrate of the B. subtilis 3 culture were separated by solid-phase extraction and preparative reverse-phase HPLC assays. Two active fractions (Fractions a and b) were eluted (Fig. 1A); their R_f values measured by TLC analysis were 0.47 and 0.85, respectively.

FIG. 1.

Isolation of antimicrobial compounds. (A) Preparative reverse-phase HPLC separation of antimicrobial substances recovered by solid-phase extraction on B. subtilis 3 cell-free filtrate. The acetonitrile gradient used to elute the antibacterial compounds is indicated by the dashed line. Activity was detected in the fractions whose peaks are indicated by arrows: a, amicoumacin fraction (TLC R_f value, 0.47); b, nonamicouamcin antibiotic fraction (TLC R_f value, 0.85). (B) Rechromatography of peak a isolated in the first step (A) by analytical reverse-phase HPLC with the acetonitrile gradient, indicated by the dashed line: a, amicoumacin A; b, amicoumacin B; c, amicoumacin C. For experimental details, see Materials and Methods. OD, optical density.

All 21 H. pylori strains tested were sensitive to both antibiotic compounds.

One of these active compounds was identified. The molecular formula of the active purified substance, C₂₀H₂₉N₃O₇, determined by high-resolution chemical ionization mass spectroscopy, (M+H)⁺ ion at m/z 424.2084 Da, and its UV-visible spectra (λ_max, 209, 244, and 312 nm) suggested that this substance was identical to amicoumacin A, an isocoumarin antibiotic (10). Its chemical structure was confirmed by ¹H NMR, ¹³C NMR, ¹H-¹H correlated NMR, ¹H-¹³C heteronuclear single-quantum-correlation NMR, and heteronuclear multiple-bond-correlation NMR spectral data analyses.

The presence of two other structurally related compounds (Fig. 1B, peaks b and c), was also detected during reverse-phase HPLC analysis. After mass spectrometry analysis, they were identified as amicoumacin B for peak b, (M+H)⁺ ion at m/z 425 Da, and as amicoumacin C for peak c, (M+H)⁺ ion at m/z 407 Da.

Kinetics of inhibition of H. pylori growth by antibacterial substances produced by B. subtilis 3.

Significant inhibitory effects were observed for the nonconcentrated supernatant, the solid-phase extract, and separated antimicrobial compounds, but none was observed for the starch medium control treatment (Fig. 2). After 2 h of contact with the cell-free supernatant (20 AU/ml), a 1-log decrease in the number of viable H. pylori cells was observed. A 6-h incubation with partially purified antimicrobial substances (solid-phase extract) containing 200 AU/ml led to killing of the bacteria. Incubation of H. pylori with each of the active compounds (amicoumacin A and the antibiotic with an R_f of 0.85) inhibited its growth, but there were some differences between the compounds in their H. pylori inhibition kinetics (Fig. 2).

FIG. 2.

Kinetics of in vitro anti-H. pylori SS1 activities of the antimicrobial substances produced by probiotic strain B. subtilis 3. a, cell-free supernatant (20 AU/ml); b, solid-phase extract (200 AU/ml); c, amicoumacin A (200 AU/ml); d, antibiotic with TLC R_f value of 0.85 (200 AU/ml); e, control. Each value shown is the mean ± standard error of the mean from three experiments. One AU was defined as the highest dilution that showed definite inhibition of H. pylori by the agar well diffusion method (see Materials and Methods).

MICs of amicoumacin A for H. pylori.

Of the three amicoumacins tested (amicoumacins A, B, and C), only amicoumacin A exhibited antibacterial activity.

The MICs of amicoumacin A for different H. pylori strains determined in solid and liquid media were in the range of 1.7 to 6.8 μg/ml and 0.75 and 2.5 μg/ml, respectively. The results are shown in Table 1.

The MICs of amicoumacin A were also determined for some standard strains of pathogenic and nonpathogenic intestinal bacteria (Table 2). This antibiotic was very active against the strains of Enterococcus faecium, Shigella flexneri, and C. jejuni tested, but it had no significant activity against the strains of lactic acid bacteria, B. fragilis, and Escherichia coli tested.

DISCUSSION

Recently, Lactobacillus sp. strains, as representatives of the normal microflora of the intestine, have been found to have an inhibitory effect on H. pylori (1, 5, 17). This effect is due to lactic acid and antibiotic production in vitro and in vivo (1, 5). A similar effect due to a probiotic strain of B. subtilis is reported for the first time. B. subtilis is a gram-positive, nonpathogenic, spore-forming organism normally found in the soil and the gastrointestinal tracts of some animals. The robustness of its spores is thought to enable passage across the gastric barrier, where a proportion of spores may germinate and colonize, albeit briefly, the intestinal tract (14).

The beneficial effects of certain Bacillus species (B. subtilis, Bacillus licheniformis, and Bacillus toyoi) on the balance of the gastrointestinal microflora of humans and animals are the basis of the rationale for their use as probiotics in the treatment and prevention of intestinal disorders (14, 26). These probiotics are available as pharmaceutical and food preparations in different European countries, Japan, Russia, Ukraine, and other countries (6, 14, 23, 26), even though little is understood about how these bacteria exert their prophylactic and therapeutic benefits.

The strain B. subtilis 3 was isolated from animal fodder, and its species identification was confirmed by 16S rRNA gene sequencing (data not shown). This strain was selected as a potential probiotic because of its safety profile in humans (20) and its antagonistic properties against different pathogenic bacteria, including Campylobacter spp. (26). Members of the Bacillus genus are known to be producers of various biologically active substances (13, 18, 22). The results of the present study show that the cell-free supernatant of B. subtilis 3 is able to inhibit the growth of H. pylori and that this inhibitory effect is not due to organic acid production but is due to the production of antimicrobial substances which are secreted in the medium during incubation.

Using TLC analysis, we detected the presence of at least two antimicrobial compounds (R_f values, 0.47 and 0.85, respectively) with anti-H. pylori activity in the cell-free supernatant of B. subtilis 3. These results were confirmed during purification by reverse-phase HPLC.

Bacteria of the genus Bacillus are able to produce specific antimicrobial substances such as bacteriocins (7, 18). However, physicochemical and biochemical characterization of the active cell-free supernatant indicates that the antimicrobial compounds secreted by B. subtilis 3 do not have functionally important protein moieties because they are resistant to proteases and drastic treatment. Therefore, these characteristics exclude the possibility that these compounds are bacteriocins or exoenzymes and indicate that they are antibiotics.

It has been known for at least 40 years that B. subtilis and other Bacillus spp. secrete antibiotics. These antibiotics usually belong to the peptide and lipopeptide groups, the aminoglycoside group, and sometimes, groups of antibiotics with various original chemical structures (11, 18, 24).

One of the antimicrobial substances produced by B. subtilis 3 was identified. The physicochemical characterization of this active purified substance suggested that it was identical to amicoumacin A, an antibiotic of the isocoumarin group. An NMR spectral data analysis confirmed its identity as amicoumacin A. Two other structurally related compounds (amicoumacin B and C; Fig. 1B) were detected. These two compounds did not have significant antibacterial activities. The production of amicoumacin A as a major component of a mixture of amicoumacins A, B, and C has already been described in another Bacillus species: B. pumilus AI-77 (10).

Furthermore, it has been reported that amicoumacin A also exhibits anti-inflammatory properties, as well as anti-stress ulcer properties, both of which were detected in a rat model (10). However, its activity against H. pylori, which causes inflammation of the gastric epithelium in humans, was not previously reported. The underestimation of MICs determined in solid medium versus those determined in liquid medium may be due to physicochemical instability of amicoumacin A under these test conditions (9). The MICs recorded for this antibiotic were only moderately high, but the small size and excellent solubility (10) of the compound indicate that it should have a good diffusion into the gastric mucosa.

The MICs for some normal intestinal microflora were also determined. This antibiotic did not affect the growth of lactic acid bacteria, B. fragilis, or E. coli, but it inhibited E. faecium. Moreover, it was shown that the increase in the concentrations of antimicrobial substances in the different fractions issued from the purification process correlates with a decrease in bacterial viability (Fig. 2). The inhibitory activity of the nonconcentrated cell-free supernatant was moderate, probably due to the low concentration of antimicrobial substances present, while after concentration and partial purification (solid-phase extract fraction), a clear bactericidal effect was observed. When H. pylori was incubated separately with each of the two antibiotic fractions isolated, some differences in their anti-H. pylori activities were observed. The more rapid bactericidal effect of the mixture (solid-phase extract) indicates the presence of additive activity between these antibiotics. Viability did not decrease rapidly enough to be due to bacterial lysis.

In conclusion, the present results demonstrate that probiotic strain B. subtilis 3 exerts antagonistic activity in vitro against H. pylori. They also show that this activity is due to the production of at least two antibiotics with additive activity, including amicoumacin A. This compound, which belongs to the isocoumarin group of antibiotics, may also exert other properties such as anti-inflammatory and antitumoral activities (4, 10) and is therefore very promising because of its potential for use in the treatment of H. pylori infection. Determination of the structure of the nonamicoumacin antibiotic produced by probiotic B. subtilis 3 is in progress. The B. subtilis 3 strain as well as the antibiotics produced by it should be tested in vivo.