Inhibitory Spectra and Modes of Antimicrobial Action of Gallotannins from Mango Kernels (Mangifera indica L.)

Abstract

This study investigated the antimicrobial activities and modes of action of penta-, hexa-, hepta-, octa-, nona-, and deca-O-galloylglucose (gallotannins) isolated from mango kernels. The MICs and minimum bactericidal concentrations (MBCs) against food-borne bacteria and fungi were determined using a critical dilution assay. Gram-positive bacteria were generally more susceptible to gallotannins than were Gram-negative bacteria. The MICs of gallotannins against Bacillus subtilis, Bacillus cereus, Clostridium botulinum, Campylobacter jejuni, Listeria monocytogenes, and Staphylococcus aureus were 0.2 g liter⁻¹ or less; enterotoxigenic Escherichia coli and Salmonella enterica were inhibited by 0.5 to 1 g liter⁻¹, and lactic acid bacteria were resistant. The use of lipopolysaccharide mutants of S. enterica indicated that the outer membrane confers resistance toward gallotannins. Supplementation of LB medium with iron eliminated the inhibitory activity of gallotannins against Staphylococcus aureus, and siderophore-deficient mutants of S. enterica were less resistant toward gallotannins than was the wild-type strain. Hepta-O-galloylglucose sensitized Lactobacillus plantarum TMW1.460 to hop extract, indicating inactivation of hop resistance mechanisms, e.g., the multidrug resistance (MDR) transporter HorA. Carbohydrate metabolism of Lactococcus lactis MG1363, a conditionally respiring organism, was influenced by hepta-O-galloylglucose when grown under aerobic conditions and in the presence of heme but not under anaerobic conditions, indicating that gallotannins influence the respiratory chain. In conclusion, the inhibitory activities of gallotannins are attributable to their strong affinity for iron and likely additionally relate to the inactivation of membrane-bound proteins.

Polyphenol-rich plant extracts often exhibit antimicrobial activity (25, 34), which allows applications of antimicrobial polyphenolic compounds in food (1, 49). The chemical diversity of polyphenols (9) is matched by diverse inhibitory spectra and modes of action of individual compounds or different classes of polyphenolics. However, with the exception of epigallocatechin and some phenolic acids (23, 42), information on the inhibitory spectra and modes of action of purified phenolic compounds remains scarce. Knowledge of the inhibitory spectra of polyphenolic compounds, as well as their modes of action, is of importance to understand their role in plant ecology, to allow the targeted development of their application as preservatives in the food and pharmaceutical industries, and to enable the preparation of plant extracts with standardized antimicrobial activities.

Gallotannins are hydrolyzable tannins composed of a glucose core esterified with gallic acid residues. Mango (Mangifera indica L.) contains gallotannins with degrees of galloylation ranging from 4 to 12 (6, 15, 16). Gallotannin-rich plant extracts and penta-O-galloylglucose exhibit antibacterial activity (1, 10, 25) which is remarkably selective. Escherichia coli, Pseudomonas, and lactic acid bacteria were resistant to mango extracts, whereas a majority of other Gram-positive bacteria were sensitive (25). The antibacterial activity of mango kernel extract could be attributed to gallotannins (15), and the use of high-speed countercurrent chromatography (HSCCC), a method for preparative purification of gallotannins, enabled the determination of the antibacterial activities of purified compounds (16).

The complexation of iron contributes to the inhibitory activity of gallotannins (10, 15). Iron is needed in a variety of essential biological processes, and only a few bacterial species grow in the absence of iron (8, 27). Many bacteria secrete siderophores, low-molecular-weight carriers with high affinity for Fe³⁺ (31), to sequester iron from the growth substrate. Cognate membrane receptors bind the ferrisiderophore to the bacterial envelope, where the metal ion undergoes a reductive separation from the ligand and becomes biologically available (31). Siderophores are composed of phenolic acids such as 2,3-dihydroxybenzoic acid (DHBA) or the catechol 2,3-dihydroxybenzoyl (DHB)-serine (48), as well as its cyclic trimer, enterobactin (33). Gallotannins have a high affinity for Fe²⁺ and Fe³⁺ in a galloylation-dependent manner (16) and can thus be considered structural and functional analogues to siderophores. Because a majority of Gram-positive bacteria are more susceptible to gallotannins than are Gram-negative bacteria (15, 25, 45), the Gram-negative bacterial outer membrane, a barrier against hydrophobic and large hydrophilic compounds, may constitute a second mechanism of resistance.

The property to precipitate proteins is eponymous for tannins and may be involved in their modes of action. Larger and more hydrophobic gallotannins bind proteins more effectively than do smaller gallotannins (4). Due to their size and charge, gallotannins are unlikely to penetrate bacterial membranes and can thus be expected to preferentially interact with cell wall proteins or membrane proteins. Flavonoid compounds such as epigallocatechin gallate (EGCG) were found to interact with membranes (23) as well as membrane proteins (21).

Although literature data on the inhibitory spectra of gallotannins as well as their interactions with ions and proteins provide clues on the mechanisms responsible for their antimicrobial activity, data on their modes of antimicrobial action are unavailable. It was therefore the aim of this study to investigate the inhibitory and bactericidal spectra of purified gallotannins. Gallotannin activity was compared to those of mangiferin, a bioactive compound isolated from mango, and EGCG, a related compound for which literature data on the inhibitory spectrum and mode of action are available.

MATERIALS AND METHODS

Bacterial strains, media, and culture conditions.

Bacterial strains used in this study, as well as their origins and culture conditions, are listed in Table 1. Strains were grown in Luria-Bertani (LB) broth (5 g liter⁻¹ glucose, 5 g liter⁻¹ yeast extract, and 5 g liter⁻¹ NaCl); brain heart infusion (BHI) broth (Becton, Dickinson and Co., Mississauga, ON, Canada); de Man, Rogosa, Sharpe (MRS) broth (Becton); or M17 broth (Oxoid, Nepean, ON, Canada) as indicated in Table 1.

TABLE 1.

Strains and culture conditions

Organism	Origin and/or referencea	Growth conditions
Bacillus subtilis FAD 110	Ropy bread (37)	LB, MRS, or BHI; aerobic or anaerobic, 37°C
Bacillus amyloliquefaciens FAD 82	Ropy bread (37)	LB, aerobic, 37°C
Bacillus cereus ATCC 14579		LB, aerobic, 37°C
Staphylococcus warneri FUA 3136	Vaginas of dairy cows	LB, aerobic, 30°C
Staphylococcus aureus ATCC 6538		LB, aerobic, 30°C
Listeria monocytogenes ATCC 7644		LB, aerobic, 30°C
Pediococcus acidilactici FUA 3072	Sausage	MRS, anaerobic, 30°C
Lactococcus lactis MG1363	20	MRS or M17, aerobic or anaerobic, 30°C
Lactobacillus plantarum TMW1.460	19	MRS, anaerobic, 30°C
Enterococcus faecalis ATCC 33186		MRS, anaerobic, 30°C
Pseudomonas fluorescens FUA 1232	Chicken meat	LB, aerobic, 30°C
Pseudomonas fluorescens ATCC 13525		LB, aerobic, 30°C
Campylobacter jejuni FUA 1220	Chicken meat	BHI, microaerophilic, 42°C
Escherichia coli AW 1.7	Beef (3)	LB, aerobic, 37°C
Escherichia coli TMW 2.497	14	LB, aerobic, 37°C
Escherichia coli ECL 13086	ETEC	LB, aerobic, 37°C
Escherichia coli ECL 13795	ETEC	LB, aerobic, 37°C
Escherichia coli ECL 13998	ETEC	LB, aerobic, 37°C
Escherichia coli ECL 14408	ETEC	LB, aerobic, 37°C
Escherichia coli FUA 1062	Vaginas of dairy cows, SLT1	LB, aerobic, 37°C
Escherichia coli FUA 1064	Vaginas of dairy cows, SLT2	LB, aerobic, 37°C
Clostridium botulinum 62A	Group I, BotNT type A	MRS, anaerobic, 37°C
Clostridium botulinum A6	Group I, BotNT type A	MRS, anaerobic, 37°C
Clostridium botulinum 2B	Group II, BotNT type B	MRS, anaerobic, 37°C
Clostridium botulinum 368B	Group I, BotNT type B	MRS, anaerobic, 37°C
Clostridium botulinum DB2	Group II, BotNT type B	MRS, anaerobic, 37°C
Clostridium botulinum ATCC 13983	Group I, BotNT type B	MRS, anaerobic, 37°C
Clostridium botulinum 17B	Group II, BotNT type B	MRS, anaerobic, 37°C
Salmonella enterica serovar Typhimurium, lipopolysaccharide (LPS) and siderophore mutantsb
LT2	Wild type (41)	LB, aerobic, 37°C
SL3770	Smooth (36)	LB, aerobic, 37°C
SL3749	Ra (36)	LB, aerobic, 37°C
SL3750	Rb2 (36)	LB, aerobic, 37°C
SL3748	Rb3 (36)	LB, aerobic, 37°C
SL3769	Rd1 (36)	LB, aerobic, 37°C
SL3789	Rd2 (36)	LB, aerobic, 37°C
SA1355	Smooth (41)	LB, aerobic, 37°C
SA1377	Re (41)	LB, aerobic, 37°C
TA2442	ent-1 (33)	LB, aerobic, 37°C
TA2443	ent-7 (33)	LB, aerobic, 37°C
Mucor plumbeus		MRS, aerobic, 25°C
Aspergillus niger		MRS, aerobic, 25°C
Penicillium spp.		MRS, aerobic, 25°C

^aGroup I, proteolytic strains; group II, nonproteolytic strains, BotNT, botulinum neurotoxin; ETEC, enterotoxigenic E. coli.

^bReceived from the Salmonella Genetic Stock Center (SGSC), Department of Biological Sciences, University of Calgary, Calgary, Alberta, Canada.

Isolation of purified gallotannins from mango.

Penta-, hexa-, hepta-, octa-, nona-, and deca-O-galloylglucose were purified as described previously (16). In brief, gallotannins were extracted from ground, defatted mango kernels by liquid-liquid extraction and isolated by HSCCC. The identity and purity of gallotannin preparations were verified by liquid chromatography-mass spectrometry (LC-MS) (16). Stock solutions of 10 g liter⁻¹ in water were sterilized by filtration and stored at −70°C. Epigallocatechin gallate and mangiferin were obtained from Sigma (St. Louis, MO), dissolved in 80% (vol/vol) acetone to a concentration of 5 g liter⁻¹, and sterilized by filtration.

SEM.

An overnight culture of Bacillus subtilis in LB broth was washed twice with 0.85% (wt/vol) saline and resuspended in saline. Hepta-O-galloylglucose was added to a concentration of 0.2 g liter⁻¹. Samples were incubated overnight at 37°C, washed in saline, and fixated for 100 min in a solution containing 2% glutaraldehyde in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8.5 mM Na₂HPO₄, 2 mM KH₂PO₄; pH 7.2). Cells were washed in PBS, dehydrated through a graded ethanol series, immersed in hexamethyldisilazane, and air dried at room temperature. Samples were coated with gold and used for scanning electron microscopy (SEM) (30). Control samples without hepta-O-galloylglucose were prepared by the same protocol.

Determination of MICs.

The inhibitory activities of penta-, hexa-, hepta-, octa-, nona-, and deca-O-galloylglucose; epigallocatechin gallate; and mangiferin were determined with indicator strains listed in Table 1. The MIC was determined by using a critical-dilution assay (17). Acetone was completely evaporated in the airflow of a sterile bench. Twofold serial dilutions of gallotannin, epigallocatechin gallate, and mangiferin were inoculated with an overnight culture of indicator strains to a cell count of about 10⁷ CFU ml⁻¹ and incubated overnight. Unless noted otherwise, MICs were defined as the lowest concentrations of the substances that inhibited the growth of microbial strains. In cases where turbidity caused by bacterial growth was obscured by precipitation of medium components with gallotannins, pH changes were measured by the addition of bromocresol green, and the MIC was defined as the lowest concentration of the substances that inhibited the acidification of the medium. To determine the inhibitory effect of gallotannins on germination of Clostridium botulinum endospores, spores were activated for 15 min at 55°C and 75°C for nonproteolytic and proteolytic strains, respectively, and media were inoculated to 3 × 10⁴ CFU ml⁻¹. Germination was visually determined after 48 h of incubation.

Bactericidal activity of hepta-O-galloylglucose and estimation of MBCs.

Serial 2-fold dilutions of penta-, hexa-, hepta-, octa-, nona-, and deca-O-galloylglucose; epigallocatechin gallate; and mangiferin with the respective growth media were prepared as described above and inoculated with indicator strains to a cell count of about 10⁷ CFU ml⁻¹. After overnight incubation, 10-μl aliquots of each dilution, corresponding to about 10⁵ cells of the initial inoculum, were transferred to 100 μl of fresh medium and incubated again overnight. The minimum bactericidal concentration (MBC) was defined as the lowest concentration preventing growth. Failure to grow upon subculturing with 10% inoculum corresponds to a reduction of the initial cell counts by at least 4 log units through the bactericidal activity of gallotannins. To validate the method, the bactericidal activity of hepta-O-galloylglucose against B. subtilis and Staphylococcus aureus, as well as the inhibitory effect toward Pseudomonas fluorescens, was assayed in PB buffer (50 mM H₂KPO₄ adjusted to pH 6.5 with NaOH). Cells from overnight cultures were washed and resuspended with the same buffer containing hepta-O-galloylglucose at the concentrations corresponding to their MBCs (0.3 g liter⁻¹ for B. subtilis and S. aureus) or MIC (3.3 g liter⁻¹ for P. fluorescens). Cells suspended in PB buffer served as control. Cultures were incubated at 37°C (B. subtilis) and 30°C (S. aureus and P. fluorescens). Cell counts were determined after 0, 1, 2, 4, 8, 12, and 24 h of incubation by surface plating of 10-fold serial dilutions on LB agar.

Influence of Fe³⁺, Mn²⁺, Mg²⁺, Ca²⁺, l-cysteine, Tween 80, and MRS broth ingredients on inhibitory activity of hepta-O-galloylglucose.

LB media were supplemented with ferric chloride, manganese sulfate, magnesium sulfate, calcium chloride, or l-cysteine to concentrations ranging from 0.026 mM to 3.3 mM. Serial 2-fold dilutions of hepta-O-galloylglucose in the respective media were prepared, and the MIC of hepta-O-galloylglucose was determined with S. aureus as an indicator strain. In a similar way, the effects of Tween 80 and the MRS broth ingredients peptone, yeast extract, beef extract, ammonium citrate, dextrose, sodium acetate, manganese sulfate, and dipotassium phosphate were determined in M17 broth with Lactococcus lactis as indicator organism. The compounds were added to M17 medium to concentrations matching their concentrations in MRS broth or exceeding their concentrations in MRS broth 4-fold.

Influence of tannins on the multidrug resistance (MDR) transporter HorA.

The influence of gallotannins on the HorA activity in Lactobacillus plantarum was tested using ethidium bromide (EB; Sigma) or Hoechst 33342 (Invitrogen, Burlington, ON, Canada) as a substrate (40, 46). To determine the effects of gallotannins on hop resistance in L. plantarum TMW1.460, an isomerized hop extract containing 30% iso-α-acids (Yakima Chief Inc., Sunnyside, WA) was added to MRS broth to concentrations ranging from 5.2 to 670 mg liter⁻¹ and the MICs of hepta-O-galloylglucose in these media were determined.

Effects of hepta-O-galloylglucose, oxygen, and heme on growth and metabolism of Lactococcus lactis MG1363.

L. lactis MG1363 was incubated in M17 medium in the presence of 0 or 2 g liter⁻¹ hepta-O-galloylglucose and in MRS medium in the presence of 0, 2, or 3 g liter⁻¹ hepta-O-galloylglucose. Cultures were grown with addition of hemin (4 ml liter⁻¹ of a solution of 5 g liter⁻¹ hemin [Sigma] in 0.05 N NaOH) or in the absence of hemin. Media (1.5 ml) were inoculated with overnight cultures of L. lactis to a cell count of about 10⁷ CFU ml⁻¹ and incubated with or without aeration. Aerated cultures were shaken at 250 rpm in 15-ml tubes; anaerobic cultures were incubated without agitation in 2-ml tubes. Samples were taken after 0, 2, 4, 8, 12, and 24 h.

Growth of L. lactis was judged by determination of the optical density at 600 nm. To quantify metabolites, cells were removed by centrifugation, perchloric acid was added to a concentration of 7%, and the precipitate was removed by centrifugation. Analysis was performed with a high-pressure liquid chromatography (HPLC) series 1200 system (Agilent Technologies, Mississauga, ON, Canada). Metabolites were separated on an Aminex HPX-87H column (Bio-Rad, Mississauga, ON, Canada) and eluted at 70°C with 5 mM H₂SO₄ at a flow rate of 0.4 ml min⁻¹. Analytes were quantified with UV (210 nm) and refractive index (RI) detectors. To investigate tannase activity, L. plantarum was grown at 30°C for 24 h in MRS medium in the presence of 0.1 g liter⁻¹ of hepta-O-galloylglucose. The concentration of hepta-O-galloylglucose in cell-free culture supernatant was determined by LC-MS analysis (16) before and after incubation.

Statistical analysis.

MICs, MBCs, and cell counts are expressed as means ± standard deviations of at least three independent experiments. Growth of L. lactis and L. plantarum in the presence of hepta-O-galloylglucose is reported as means ± standard deviations of duplicate independent experiments. Statistical significance was determined by Student's t test using SigmaPlot software (Systat Software, Inc., San Jose, CA), and P was <0.05.

RESULTS

Inhibitory spectra of gallotannins: effects of the degrees of galloylation and growth conditions.

Gallotannins inhibited microbial growth with MICs ranging from less than 0.1 g liter⁻¹ for S. aureus and Listeria monocytogenes to more than 3.3 g liter⁻¹ for Pediococcus acidilactici (Table 2). The MIC of epigallocatechin gallate was generally equal to or higher than the MICs of gallotannins, and mangiferin did not exhibit inhibitory activity at a concentration of 1.7 g liter⁻¹. The effects of growth conditions were evaluated with B. subtilis. In aerobic culture, B. subtilis grown in MRS broth was less sensitive to tannins than were cultures in LB broth. Remarkably, incubation of B. subtilis in MRS broth under anaerobic conditions resulted in resistance to gallotannins.

TABLE 2.

MICs and MBCs of galloylglucose, epigallocatechin gallate, and mangiferin^d

Organism (growth condition[s])	Degree of galloylation or comparatora	MIC (g liter−1)	MBC (g liter−1)
Bacillus subtilis (LB)b	5	0.2 ± 0.1	0.4 ± 0.4
	6	0.1 ± 0.0	0.4 ± 0.4
	7	0.1 ± 0.0	0.3 ± 0.1
	8	0.1 ± 0.0	0.3 ± 0.1
	9	0.1 ± 0.0	0.2 ± 0.1
	10	0.3 ± 0.1	0.9 ± 0.7
	E	0.6 ± 0.2	1.0 ± 0.6
	M	>1.7	>1.7
Bacillus subtilis (MRS, aerobic)b	5	3.3 ± 0.0	3.3 ± 0.0
	6	3.3 ± 0.0	>3.3
	7	1.7 ± 0.0	>3.3
	8	1.3 ± 0.6	>3.3
	9	1.3 ± 0.6	>3.3
	10	1.3 ± 0.6	>3.3
	E	1.3 ± 0.6	>1.7
	M	>1.7	>1.7
Bacillus subtilis (MRS, anaerobic)b	5	>3.3	>3.3
	6	>3.3	>3.3
	7	>3.3	>3.3
	8	>3.3	>3.3
	9	>3.3	>3.3
	10	NDe	ND
	E	>1.7	>1.7
	M	>1.7	>1.7
Bacillus cereusb	5	ND	ND
	6	<0.1 ± 0.0	0.2 ± 0.1
	7	<0.1 ± 0.0	0.1 ± 0.0
	8	0.1 ± 0.0	0.1 ± 0.1
	9	<0.1 ± 0.0	0.1 ± 0.0
	10	ND	ND
	E	0.1 ± 0.1	0.2 ± 0.1
	M	>1.7	>1.7
Staphylococcus aureusc	5	ND	ND
	6	<0.1 ± 0.0	0.2 ± 0.0
	7	<0.1 ± 0.0	0.3 ± 0.1
	8	<0.1 ± 0.0	0.3 ± 0.1
	9	<0.1 ± 0.0	0.3 ± 0.1
	10	ND	ND
	E	0.2 ± 0.2	0.3 ± 0.1
	M	>1.7	>1.7
Listeria monocytogenesc	5	ND	ND
	6	<0.1 ± 0.0	0.4 ± 0.0
	7	<0.1 ± 0.0	0.4 ± 0.0
	8	<0.1 ± 0.0	0.8 ± 0.0
	9	<0.1 ± 0.0	0.6 ± 0.2
	10	ND	ND
	E	0.5 ± 0.3	1.1 ± 0.5
	M	>1.7	>1.7
Pediococcus acidilacticic	5	ND	ND
	6	>3.3	>3.3
	7	>3.3	>3.3
	8	>3.3	>3.3
	9	>3.3	>3.3
	10	ND	ND
	E	>1.7	>1.7
	M	>1.7	>1.7
Lactococcus lactisc (MRS)	5	ND	ND
	6	3.3 ± 0.0	>3.3
	7	3.3 ± 0.0	>3.3
	8	3.3 ± 0.0	>3.3
	9	3.3 ± 0.0	>3.3
	10	ND	ND
	E	>1.7	>1.7
	M	>1.7	>1.7
Pseudomonas fluorescens ATCC 13525c	5	ND	ND
	6	1.4 ± 0.5	1.7 ± 0.0
	7	2.5 ± 0.9	2.8 ± 0.9
	8	2.8 ± 0.9	3.0 ± 0.7
	9	2.8 ± 0.9	1.7 ± 0.0
	10	ND	ND
	E	>1.7	>1.7
	M	>1.7	>1.7
Clostridium botulinum A6b	5	ND	ND
	6	ND	0.1 ± 0.1
	7	ND	0.2 ± 0.1
	8	ND	0.5 ± 0.4
	9	ND	0.3 ± 0.1
	10	ND	ND
	E	ND	>1.7
	M	ND	>1.7

^aDegree of galloylation: 5, penta-O-galloylglucose; 6, hexa-O-galloylglucose; 7, hepta-O-galloylglucose; 8, octa-O-galloylglucose; 9, nona-O-galloylglucose; 10, deca-O-galloylglucose; E, epigallocatechin gallate; M, mangiferin (epigallocatechin gallate and mangiferin were used for comparison).

^bMIC determined by measurement of optical density.

^cMIC determined by measurement of acidification.

^dCulture conditions were as indicated in Table 1 unless otherwise specified.

^eND, not determined.

The degree of galloylation generally did not affect the antibacterial activities of gallotannins (Table 2). Therefore, Table 3 shows the MICs and MBCs of hepta-O-galloylglucose only in comparison with those of epigallocatechin gallate. Data for other gallotannins are shown in Table S1 of the supplemental material. Virtually all Gram-positive strains, with the exception of lactic acid bacteria, were sensitive to tannins, and MICs were 0.1 g liter⁻¹ or less. The sensitivities of Gram-positive food-borne pathogens or toxigenic organisms, i.e., Bacillus cereus, C. botulinum, L. monocytogenes, and S. aureus, were comparable to those of other Gram-positive bacteria.

TABLE 3.

MICs and MBCs of hepta-O-galloylglucose and epigallocatechin gallate

Organismd	Hepta-O-galloylglucose		Epigallocatechin gallate
	MIC (g liter−1)	MBC (g liter−1)	MIC (g liter−1)	MBC (g liter−1)
Bacillus subtilis (BHI)a	NDc	2.5 ± 1.2	ND	ND
Bacillus amyloliquefaciensa	0.1 ± 0.0	0.2 ± 0.2	0.3 ± 0.1	0.6 ± 0.2
Staphylococcus warnerib	<0.1 ± 0.0	0.3 ± 0.1	0.1 ± 0.1	0.3 ± 0.1
Lactococcus lactis (M17)b	0.4 ± 0.2	3.3 ± 0.0	ND	ND
Lactococcus lactis (M17 plus Tween 80)b	3.3 ± 0.0	ND	ND	ND
Lactobacillus plantarumb	>3.3	>3.3	>1.7	>1.7
Enterococcus faecalisb	3.3 ± 0.0	>3.3	>1.7	>1.7
Pseudomonas fluorescens FUA 07/03b	1.7 ± 0.0	2.5 ± 1.2	>1.7	>1.7
Pseudomonas fluorescens ATCC 13525b	2.5 ± 0.9	2.8 ± 0.9	>1.7	>1.7
Campylobacter jejunia	<0.1 ± 0.0	<0.1 ± 0.0	0.6 ± 0.3	0.7 ± 0.8
Escherichia coli AW 1.7b	0.8 ± 0.0	0.8 ± 0.0	ND	ND
Escherichia coli TMW 2.497b	0.8 ± 0.0	2.2 ± 1.0	ND	ND
Enterotoxigenic E. coli ECL 13086b	0.6 ± 0.3	0.8 ± 0.0	ND	ND
Enterotoxigenic E. coli ECL 13795b	0.8 ± 0.0	0.8 ± 0.0	ND	ND
Enterotoxigenic E. coli ECL 13998b	0.6 ± 0.3	0.8 ± 0.0	ND	ND
Enterotoxigenic E. coli ECL 14408b	0.8 ± 0.0	1.3 ± 0.6	ND	ND
E. coli SLT1 FUA 1062b	0.7 ± 0.1	0.8 ± 0.0	ND	ND
E. coli SLT2 FUA 1064b	0.8 ± 0.0	1.7 ± 1.4	ND	ND
Clostridium botulinum 62Aa	ND	0.5 ± 0.4	ND	>1.7
Clostridium botulinum 2Ba	ND	0.4 ± 0.0	ND	>1.7
Clostridium botulinum 368Ba	ND	0.5 ± 0.4	ND	>1.7
Clostridium botulinum DB2a	ND	0.6 ± 0.3	ND	>1.7
Clostridium botulinum ATCC 13983a	ND	0.5 ± 0.4	ND	>1.7
Clostridium botulinum 17Ba	ND	0.5 ± 0.4	ND	>1.7
Salmonella enterica serovar Typhimuriumb	0.7 ± 0.2	2.8 ± 1.0	ND	ND
Mucor plumbeusa	0.8 ± 0.0	ND	>1.7	>1.7
Aspergillus nigera	>3.3	ND	>1.7	>1.7
Penicillium spp.a	>3.3	ND	>1.7	>1.7

^aMIC determined by measurement of optical density.

^bMIC determined by measurement of acidification.

^cND, not determined.

^dCulture conditions were as stated in Table 1 unless otherwise specified.

E. coli strains, including enterotoxigenic E. coli (ETEC) and Shiga-like toxin-producing (SLT1 and SLT2) strains, were inhibited by concentrations ranging from 0.6 to 0.8 g liter⁻¹. Two strains of P. fluorescens and Salmonella enterica were more resistant, with MICs higher than 1.7 g liter⁻¹ and 0.7 g liter⁻¹, respectively, whereas Campylobacter jejuni was inhibited by concentrations of less than 0.1 g liter⁻¹. For the fungal strains tested, gallotannins inhibited growth of Mucor plumbeus only.

Bactericidal activity of hepta-O-galloylglucose against B. subtilis, S. aureus, and P. fluorescens and MBC against other indicator organisms.

Tannins exhibited bactericidal activities when the concentrations exceeded the MICs (Tables 2 and 3). MBCs were generally determined by a critical-dilution assay. In keeping with the method for determination of the MBC used in this study, cell counts of S. aureus and B. subtilis were reduced by about 3 orders of magnitude within 24 h upon challenge with hepta-O-galloylglucose at the level of the MBC (Fig. 1). In contrast, cell counts of P. fluorescens were not reduced by hepta-O-galloylglucose at the level of the MIC. Cell counts in samples incubated without galloylglucose remained unchanged (data not shown).

FIG. 1.

Cell counts of P. fluorescens (▪) incubated in PB with hepta-O-galloylglucose at the level of the MIC (3.3 g liter⁻¹) and S. aureus (•) and B. subtilis (▾) incubated in PB with hepta-O-galloylglucose at the level of the MBCs (0.3 g liter⁻¹). Control samples, in which the tannin solution was replaced with water, did not lose viability (data not shown). Results are shown as means ± standard deviations of three independent experiments.

MICs were not determined for strains of C. botulinum due to precipitation in the growth medium and because growth of C. botulinum could not be detected by use of a pH indicator. MBCs of vegetative cells of proteolytic (group I) and nonproteolytic (group II) C. botulinum strains ranged from 0.2 to 0.6 g liter⁻¹, and C. botulinum was among the most sensitive organisms evaluated. Remarkably, the MIC of epigallocatechin gallate, which was generally comparable to those of gallotannins with most other indicator organisms, was at least 10-fold higher for clostridia (Tables 2 and 3). The germination of endospores of all strains of C. botulinum was inhibited by hepta-O-galloylglucose at concentrations of less than 0.1 g liter⁻¹.

Morphology of B. subtilis treated with hepta-O-galloylglucose.

The cell morphology of B. subtilis was observed with SEM after treatment with hepta-O-galloylglucose (Fig. 2). Cells incubated with 0.2 g liter⁻¹ hepta-O-galloylglucose were elongated to a multiple of their original size. Further magnification did not reveal changes in the appearance of the cell surface (data not shown).

FIG. 2.

Scanning electron microscope picture of B. subtilis in 0.85% (wt/vol) saline (A) and with 0.2 g liter⁻¹ hepta-O-galloylglucose (B).

Effects of iron, other ions, and MRS medium components on the inhibitory activity of tannins.

The complexation of iron contributes to the antibacterial mode of action of gallotannins (10, 16). It was determined whether other divalent cations counteract the antibacterial activity of gallotannins in a similar way (Fig. 3). Supplementation of LB medium with iron decreased the antimicrobial activity of hepta-O-galloylglucose against S. aureus, but manganese, magnesium, calcium, and l-cysteine had no effect. Addition of 1 g liter⁻¹ Tween 80 to M17 medium increased the MIC of gallotannins against L. lactis to 3.3 g liter⁻¹ (Table 3). Other components of MRS broth that were added to M17 medium had no effect on the antimicrobial activity (data not shown).

FIG. 3.

Influences of different concentrations of iron (•), manganese (▾), magnesium (▵), calcium (▪), and l-cysteine (○) on the sensitivity of S. aureus toward hepta-O-galloylglucose. Results for magnesium and l-cysteine are displayed, but MIC data obtained with calcium are superimposed on them. MICs were determined by addition of bromocresol green. Color change and precipitation made the determination of the MICs for iron concentrations higher than 0.4 mM impossible. Results are shown as means ± standard deviations of three independent experiments.

Sensitivities of isogenic LPS and siderophore mutant strains of S. enterica to hepta-O-galloylglucose.

The susceptibilities of Gram-negative bacteria toward hepta-O-galloylglucose were assessed using a wild-type strain of S. enterica and sets of isogenic mutants with defects in their lipopolysaccharide (LPS; Table 1). The length in sugar residues of the polysaccharide chain decreases stepwise from the smooth chemotype (wild type) to the deep rough mutants. A schematic representation of the LPS structure of the mutants is provided by Nikaido (32). Additionally, two mutant strains blocked in the biosynthesis of the siderophore enterobactin (33) were tested for their sensitivities toward hepta-O-galloylglucose (Fig. 4). MICs of chemotypes Rd₂ and Re, which are truncated beyond the ethanolamine phosphate-heptose residue in the carbohydrate chain, were significantly lower than the MIC for the wild-type strain. MBCs were significantly lower for chemotypes Rb₃, Rd₁, Rd₂, and Re (Fig. 4). The MICs of hepta-O-galloylglucose against siderophore ent-1 and ent-7 mutants were comparable to that against the wild type, but a significant difference was observed in the MBCs (Fig. 4).

FIG. 4.

Inhibitory and bactericidal activities of gallotannins against LPS and siderophore mutants of Salmonella enterica. The inhibitory activity is expressed in MICs (black columns) and minimum bactericidal concentrations (MBCs; gray columns). Results are shown as means ± standard deviations of three independent experiments. Asterisks mark samples that are significantly different (P < 0.050) from the wild-type strain (LT2).

Influence of tannins on hop resistance and the MDR transporter HorA.

MDR transporters are integral membrane proteins exporting toxic components from the cell. The effect of hop compounds on HorA activity and the sensitivity of L. plantarum TMW1.460 toward hepta-O-galloylglucose were quantified. Hop resistance of L. plantarum TMW1.460 is dependent on HorA, a MDR transporter of the ATP-binding cassette family of transport proteins (19, 40, 46; for a review, see reference 39). Interactions of tannins with the dyes ethidium bromide and Hoechst 33342, however, rendered direct measurement of HorA impossible (data not shown). An indirect indication of HorA inactivation by gallotannins was obtained by the combined effect of a hop extract and hepta-O-galloylglucose. A hundredfold increase of the concentration of iso-α-humulone resulted in a corresponding reduction of the MIC of hepta-O-galloylglucose against L. plantarum (Fig. 5). In keeping with previous observations, isomerized hop extract alone inhibited growth of L. plantarum at a concentration of about 200 mg liter⁻¹ (Fig. 5) (19).

FIG. 5.

MICs of hepta-O-galloylglucose toward L. plantarum (•) in the presence of various concentrations of hops. MICs were determined by addition of bromocresol green. The MIC of hop extracts in the absence of gallotannins was 200 mg liter⁻¹. Results are shown as means ± standard deviations of three independent experiments.

Influence of gallotannins on metabolism and respiration in L. lactis.

The inhibitory spectra of gallotannins as well as the response of B. subtilis to gallotannins in the absence or presence of oxygen (Tables 2 and 3) indicates a possible role of respiratory enzymes in bacterial sensitivity to gallotannins. The influence of oxygen and respiration on bacterial resistance to hepta-O-galloylglucose was evaluated with L. lactis MG1363, a conditionally respiring organism. Effects of culture conditions observed in M17 or MRS medium in the absence of tannins were in excellent agreement with previous reports (13) (Fig. 6; see also Fig. S1 in the supplemental material; data not shown). Homolactic fermentation of glucose to lactate was observed under anaerobic conditions and also in aerobic cultures; supplementation with heme allowed for the formation of functional cytochromes and conversion of lactate to acetoin after cofactor regeneration by the respiratory chain (13) (Fig. 6). Addition of 3 g liter⁻¹ of hepta-O-galloylglucose had only minor effects on the anaerobic metabolism of L. lactis in MRS medium, irrespective of the presence of heme (Fig. S1). However, metabolism of cultures grown aerobically in the presence of hepta-O-galloylglucose shifted from lactate to acetoin formation despite the absence of functional cytochromes, indicating that gallotannins shift cofactor regeneration through unknown mechanisms (Fig. 6). Differently from cultures grown in the absence of gallotannins, addition of heme to allow formation of functional cytochromes had virtually no effect on the metabolic end products, indicating that the respiratory chain of L. lactis is not functional when gallotannins are present (Fig. 6).

FIG. 6.

Metabolism of L. lactis grown aerobically in the presence of 3 g liter⁻¹ hepta-O-galloylglucose (A and C) or in the absence of hepta-O-galloylglucose (B and D) in MRS medium. Cultures were grown with addition of 20 mg liter⁻¹ hemin (A and B) or in the absence of hemin (C and D). Shown are the concentrations of glucose (○), lactate (▴), acetate (▪), and acetoin (♦). Results are shown as means ± standard deviations of two independent experiments.

DISCUSSION

This study evaluated the inhibitory spectra of highly purified gallotannins with various degrees of galloylation toward a broad range of bacteria and several fungi. Together with literature data on the interactions of gallotannins with proteins and ions, particularly iron, this detailed characterization of the inhibitory spectra enabled the evaluation of their modes of action through a combination with biochemical and physiological experimentation to determine the contributions of iron complexation and protein denaturation. Data on the modes of action of gallotannins will facilitate application development for their use as antimicrobials in food.

Penta-, hexa-, hepta-, octa-, nona-, and deca-O-galloylglucose exhibited a selective inhibitory action. Gram-positive bacteria were generally more sensitive toward gallotannins than were Gram-negative bacteria, in accordance with studies that used gallotannin-rich plant extracts or penta-O-galloylglucose (10, 15, 25, 45). The MICs of gallotannins toward most indicator strains were generally comparable to that of epigallocatechin gallate; however, C. botulinum was more sensitive to gallotannins. In addition to their inhibitory effect, gallotannins acted bactericidally. Bactericidal activity was previously reported for tannic acid (22) but not for galloylglucose.

In contrast to previous studies, this study accounted for differences in growth conditions for comparison of the sensitivities of indicator bacteria. MICs were strongly influenced by the presence of oxygen and medium compositions. The antibacterial activity toward B. subtilis was more than 10-fold higher in LB medium (aerobic) than in MRS medium (anaerobic). Tween 80, a component of MRS medium but not of the other media used in this study, increased the MICs of gallotannins toward L. lactis. However, medium composition alone cannot account for the different sensitivities, as C. botulinum was sensitive toward gallotannins when grown in MRS medium.

For EGCG, the biological activities were attributed to the presence of galloyl and gallic acid moieties of the catechin structures. EGCG was found to change membrane fluidity and cell morphology and to cause leakage of cytoplasmic material (reference 23 and references therein). Gram-positive S. aureus cells absorb more EGCG into their cell wall and aggregate its presence, while Gram-negative E. coli does not aggregate and absorbs less EGCG (23, 48). This was attributed to the strong repulsive negative charge of lipopolysaccharides on the surfaces of Gram-negative bacteria. Peptidoglycan but not lipopolysaccharide or dextran was found to block the bactericidal activity of EGCG (50). Because the cell wall of Gram-positive bacteria is composed of 30 to 50 layers of peptidoglycan, EGCG can bind more effectively to S. aureus than to E. coli. Yoda et al. (50) hypothesized that the binding of EGCG to peptidoglycan disrupts its function in osmotic protection, cell division, and cell wall biosynthesis. Inhibition of cell division, caused by binding of gallotannins to the peptidoglycan or inhibition of enzymes involved in cell separation, may account for the altered bacterial morphology of B. subtilis with SEM.

Iron counteracts the antibacterial activity of gallotannins (15). Iron binding of gallotannins but not their antibacterial activity was dependent on the number of galloyl groups in the molecule, with a larger capacity at lower degrees of galloylation (15, 16; this study). Growth of B. subtilis and S. aureus in media with gallotannins was restored by the addition of Fe²⁺ and Fe³⁺ but not with Ca²⁺, Mg²⁺, or Mn²⁺, indicating that mineral complexation by gallotannins is selective. Animals developed systems with comparable specificities, e.g., lactoferrin in milk and ovotransferrin in eggs, to restrain bacterial growth. Polyphenols might act as the plant counterparts to mammalian iron-complexing proteins to control bacterial growth in injured tissues.

Microorganisms form siderophores to grow in iron-deficient substrates. The relative iron binding efficiencies are 1.0 for tannic acid and 0.3 for siderophores produced by E. coli (12). Enterobactin binds Fe³⁺ with a very high affinity (K = 10⁵² M⁻¹) (31, 35). Enterobactin-defective mutants of S. enterica (33) were more sensitive toward gallotannins. Likewise, siderophore-defective mutants of Erwinia chrysanthemi were more sensitive to gallotannins than were the wild-type strains, and their growth could be restored by the addition of iron (29).

The strict anaerobic clostridia employ iron-sulfur (Fe-S) proteins as essential catalysts (5, 28), and iron complexation by gallotannins may thus explain their strong inhibitory activity against C. botulinum. Iron complexation also inhibits spore germination; supplementation of sporulation media with iron resulted in C. botulinum 113B spores that germinated more readily, and outgrowth was observed for a higher proportion of the population (26). As an exception in the microbial world, lactic acid bacteria do not rely on iron for their growth (8, 27). In many enzymes, iron is replaced by unique cofactors such as manganese (24, 44), and high intracellular manganese concentrations replace iron-dependent superoxide dismutase (2). The lack of iron-dependent metabolism in lactic acid bacteria may reflect an alternative evolutionary strategy to grow in substrates that are deficient in iron.

With the noticeable exceptions of lactic acid bacteria and Campylobacter jejuni, Gram-positive bacteria are more susceptible to tannins than are Gram-negative bacteria (15, 25, 45; this study), indicating a possible role of the outer membrane. In analogy to previous studies probing the role of the outer membrane in resistance to antimicrobial agents, LPS mutants of S. enterica (18, 47) were found to be more susceptible toward hepta-O-galloylglucose than was the wild-type strain. In addition to its properties as a physical barrier, the negatively charged outer membrane might limit binding of the negatively charged gallotannins through electrostatic repulsion.

According to their defining characteristic, gallotannins precipitate proteins. Although this study failed to provide direct evidence for the inactivation of bacterial proteins by gallotannins, several lines of evidence suggest that membrane proteins are a second target for gallotannins. Hepta-O-galloylglucose sensitized L. plantarum to hop iso-α-acids, which act as ionophores (43). Hop resistance in beer-spoiling lactobacilli is dependent on HorA activity (19, 46; for a review, see reference 39), and HorA inactivation in L. plantarum TMW1.460 eliminates hop resistance (19, 46). EGCG inhibits staphylococcal MDR transporters (21, 42). The effect of gallotannins on hop resistance of L. plantarum TMW1.460 thus suggests inhibition of the integral membrane protein HorA. Inactivation of other membrane proteins involved in hop resistance may contribute to this effect (39).

In L. lactis, gallotannins had no effect on anaerobic growth and metabolism. However, gallotannins induced a metabolic shift from lactate production toward acetoin production under aerobic growth conditions that was independent of functional cytochromes. The respiration induced through addition of heme (13) was inhibited by addition of gallotannins. In aerobic cultures grown in the presence of gallotannins, metabolisms of L. lactis were comparable when heme was added and when it was not added. These results, together with the observation that anaerobically grown B. subtilis and the nonrespiring lactic acid bacteria and bifidobacteria are gallotannin resistant, indicate that the respiratory chain is inhibited by gallotannins.

Aspergillus and Penicillium species produce tannase (7), which may account for their resistance to gallotannins. Intracellular tannase activity was also reported in L. plantarum (38), but L. plantarum TMW1.460 did not degrade gallotannins.

In conclusion, gallotannins exhibit inhibitory activity against a large number of bacteria at physiologically relevant concentrations (0.1 to 1 g kg⁻¹) (6, 11), and their activity is attributable to their affinity for iron. Membrane-bound proteins likely are additional targets of gallotannins. The elucidation of the inhibitory spectra and modes of action of purified gallotannins supports the development of their application as preservatives in food products. The inhibitory activity of gallotannins is counteracted by iron, which limits their application in meat products, but not by calcium, manganese, and magnesium. Gallotannins are compatible with the simultaneous use of lactic starter cultures or protective cultures, as lactic acid bacteria are resistant to gallotannins. Food applications of gallotannins, however, additionally have to account for beneficial or adverse effects on human health (11).