Antibacterial activity and mode of action of a commercial citrus fruit extract

Authors


Correspondence

Avelino Álvarez-Ordóñez, Infectious Diseases and Epidemiology Unit, Department of Animal Health, Veterinary Faculty, University of León, Campus de Vegazana s/n, 24071-León, Spain. E-mail: aalvo@unileon.es

Abstract

Aims

This study addresses the antibacterial activity and mechanism of action of BIOLL, a commercial extract obtained from citrus fruits.

Methods and Results

Strong activities with minimum inhibitory concentrations (MIC) ranging from 10 ppm (for some Brachyspira hyodysenteriae strains) to 80 ppm (for various Salmonella enterica and Escherichia coli strains) were observed. Membrane integrity tests and Fourier transform infrared (FT-IR) spectroscopic analyses were performed to shed light on the effects caused on molecular structure and composition. Physical effects, with formation of pores and leakage of intracellular components, and chemical effects, which were dependent on the bacterial species, were evident on cellular envelopes. Whereas for S. enterica and E. coli, changes were focused on the carboxylic group of membrane fatty acids, for B. hyodysenteriae, the main effects were found in polysaccharides and carbohydrates of the cell wall.

Conclusions

The great antibacterial activity shown by BIOLL and its proposed dual physico-chemical mode of action, with species-specific cellular targets, show its attractiveness as an alternative to antibiotics.

Significance and impact of the study

Antibiotic resistance is becoming a serious problem. Our study characterizes a novel antimicrobial extract, which could represent an alternative to antibiotics for treatment or prevention of bacterial infectious diseases.

Introduction

Antibiotic-resistant bacteria represent a major concern worldwide. The possibility of developing bacterial resistant populations when using antibiotics as growth promoters in animal production has led to the EU and USA ban on the use of antibiotics as feed additives on farm animals (Casewell et al. 2003; Gaggìa et al. 2010; Hashemi and Davoodi 2011). Therefore, intense research efforts focused on the search for new alternatives to prevent intestinal infectious diseases, including use of probiotics, prebiotics and feed additives, have been made in the last decade. Special attention has been paid to the antimicrobial activity of diverse plant oil extracts and their components, which have been reported to show great inhibitory effects against pathogenic bacteria, yeasts, fungi and viruses (Burt 2004; Alviano and Alviano 2009; Reichling et al. 2009; Solórzano-Santos and Miranda-Novales 2012). Among the potential candidates, citrus fruit extracts are generating great interest because their components have been reported to show not only antimicrobial effects (Nannapaneni et al. 2008; O'Bryan et al. 2008; Bevilacqua et al. 2010; Frassinetti et al. 2011; Settani et al. 2012), but also other beneficial biological activities – for example anti-inflammatory properties (Manthey et al. 2001; Cushnie and Lamb 2005).

BIOLL is a natural extract obtained from Citrus paradisi (grapefruit), Citrus reticulata (mandarin), Citrus aurantium subsp. bergamia (bergamot) and Citrus sinensis (sweet orange) and commercialized by Grupo Omega (Spain) to supplement animal feed. Indeed, it is especially recommended in swine and poultry production at dosages of 100–500 g per ton of feed (granulated presentation) or 50–350 ml per 103 l of drinking water (liquid presentation) as a supplement indicated to increase productivity and feed conversion and to control digestive processes and gastrointestinal pathogenic bacteria. This extract is rich in flavonoids (1·74 g/100 mg) and citric acid (1·87 g/100 mg), compounds with known antibacterial activity. Among those flavonoids with antibacterial activity present in BIOLL, the most numerous are quercetin (1972 ppm), naringenin (833 ppm) and naringin (54 ppm). However, the extract also contains other compounds with antibacterial activity, such as ascorbic acid, hesperidin, rutin and saponins (Grupo Omega, personal communication).

The antibacterial activity and mode of action of BIOLL are not completely characterized yet. The aim of this study was to test its efficacy to inactivate or to inhibit the growth of several pathogenic bacteria of interest for animal health, that is, S. enterica subsp. enterica, Escherichia coli and Brachyspira hyodysenteriae. Various strains of these important animal pathogens, most of them isolated from swine finishing farms, were included in the study. In addition, to elucidate its mechanisms of action, membrane integrity tests and Fourier transform infrared (FT-IR) spectroscopic analyses were performed on treated cells to shed light on the effects caused by this citrus fruit extract on the molecular structure and composition, with special focus on the cellular envelopes.

Materials and methods

Bacterial strains and culture conditions

Strains of S. enterica, E. coli and B. hyodysenteriae used in this study, listed in Table 1, belonged to the Infectious Diseases Unit Collection (IDUC) of the University of León. Salmonella enterica and E. coli cultures were resuscitated and maintained on tryptic soy agar (TSA) plates. Test cultures were prepared by transferring an isolated colony from a plate into a test tube containing 10 ml of sterile brain heart infusion (BHI) followed by incubation at 37°C for 24 h in a shaking incubator. Brachyspira hyodysenteriae strains were grown on blood agar plates and subsequently on fastidious anaerobe agar (FAA) at 42°C in an anaerobic chamber. Test cultures were prepared by transferring the bacterial load from a FAA plate into a test tube containing 10 ml of sterile BHI supplemented with 10% foetal bovine serum, followed by incubation at 42°C for 48 h under anaerobic atmosphere in a shaking incubator. Salmonella enterica, E. coli and B. hyodysenteriae test cultures were subsequently used for all experiments performed.

Table 1. Bacterial strains and their susceptibility to BIOLL
IDUC Id.StrainOriginSourceaMinimum inhibitory concentrations (ppm)
  1. a

    All strains belonged to the Infectious Diseases Unit Collection (IDUC) of the University of León, Spain.

RS37Salmonella Typhimurium CECT443Spanish Type Culture CollectionIDUC40
RS38S. Typhimurium CECT 883Spanish Type Culture CollectionIDUC20
RS39Salmonella Enteritidis CECT 4300Spanish Type Culture CollectionIDUC40
RS40Salmonella Infantis CECT 700Spanish Type Culture CollectionIDUC80
RS41Salmonella Cholerasuis CECT 915Spanish Type Culture CollectionIDUC80
SP11S. Typhimurium DT104Swine Finishing farm isolateIDUC80
SP28Salmonella LondonSwine Finishing farm isolateIDUC80
SP36Salmonella RissenSwine Finishing farm isolateIDUC80
SP58S. Typhimurium DT104Swine Finishing farm isolateIDUC80
SP62Salmonella DerbySwine Finishing farm isolateIDUC80
EC58Escherichia coli (haemolytic strain)Swine Finishing farm isolateIDUC80
EC59E. coli (haemolytic strain)Swine Finishing farm isolateIDUC80
EC60E. coli (haemolytic strain)Swine Finishing farm isolateIDUC40
EC61E. coli (haemolytic strain)Swine Finishing farm isolateIDUC80
EC62E. coli (haemolytic strain)Swine Finishing farm isolateIDUC40
EC63E. coli (haemolytic strain)Swine Finishing farm isolateIDUC40
EC64E. coli (haemolytic strain)Swine Finishing farm isolateIDUC40
EC65E. coli (haemolytic strain)Swine Finishing farm isolateIDUC80
EC66E. coli (haemolytic strain)Swine Finishing farm isolateIDUC40
EC67E. coli (haemolytic strain)Swine Finishing farm isolateIDUC80
B78Brachyspira hyodysenteriae ATCC 27164American Type Culture CollectionIDUC10
B204B. hyodysenteriae ATCC 31212American Type Culture CollectionIDUC40
H2 B. hyodysenteriae Swine Finishing farm isolateIDUC20
H9 B. hyodysenteriae Swine Finishing farm isolateIDUC40
H40 B. hyodysenteriae Swine Finishing farm isolateIDUC20
H52 B. hyodysenteriae Swine Finishing farm isolateIDUC20
H76 B. hyodysenteriae Swine Finishing farm isolateIDUC20
H151 B. hyodysenteriae Swine Finishing farm isolateIDUC10
H363 B. hyodysenteriae Swine Finishing farm isolateIDUC10
H507 B. hyodysenteriae Swine Finishing farm isolateIDUC20
H549 B. hyodysenteriae Swine Finishing farm isolateIDUC40
H555 B. hyodysenteriae Swine Finishing farm isolateIDUC40
H583 B. hyodysenteriae Swine Finishing farm isolateIDUC40
H591 B. hyodysenteriae Swine Finishing farm isolateIDUC10

Antibacterial activity of the commercial citrus fruit extract

A stock suspension of BIOLL (Grupo Omega; liquid presentation, Arganda del Rey, Spain) was prepared at 3200 ppm and used for all tests. Susceptibility testing was performed by the broth dilution method using 48-well tissue culture plates (Iwaki, Tokyo, Japan). Briefly, twofold serial dilutions of the citrus fruit extract, with concentrations ranging from 160 to 1·25 ppm, were carried out in BHI (supplemented with 10% foetal bovine serum in the case of B. hyodysenteriae). Afterwards, the 48-well plate was prepared by dispensing into each well 250 μl of the test growth medium and 250 μl of the test bacterial cell suspension standardized with 0·5 McFarland turbidity standard and diluted to a final level of ~5 × 105 cells ml−1. Positive control wells (containing 250 μl of bacterial suspension and 250 μl of BHI without citrus fruit extract supplementation) and negative control wells (containing 500 μl of BHI) were included. The minimum inhibitory concentration (MIC) was determined for each strain as the lowest concentration which prevented visible growth after incubation in a shaking incubator at 37°C for 24 h (Senterica and E. coli) or 42°C for 96 h (B. hyodysenteriae). The absence of contamination was confirmed for B. hyodysenteriae by phase-contrast microscopy. Aliquots from wells not showing visible growth were plated onto solid media to confirm whether those concentrations were bactericidal. MIC50 and MIC90 values, defined as the lowest concentrations that inhibited the growth of 50 and 90% of total isolates tested, were estimated for each bacterial species as the median and the 90th percentile values, respectively, of the MICs observed for all bacterial isolates tested.

To study the bactericidal activity of the citrus fruit extract, inactivation kinetic experiments were performed for some strains of Senterica and E. coli. In these experiments, 1 ml of stationary-phase cultures was centrifuged at 11 000 g for 3 min at 4°C. The supernatant liquid was discarded, and the cellular pellet was resuspended in 1 ml of BHI supplemented with different concentrations of the citrus fruit extract. After incubation for 90 min at room temperature, survival was monitored. Tenfold serial dilutions were produced in sterile phosphate-buffered saline (PBS), and suitable dilutions were plated in duplicate on TSA. Viable cell counts were enumerated following incubation of TSA plates at 37°C for 48 h (longer incubation times did not show any influence on the count). Inactivation experiments were performed in triplicate using three different fresh cultures.

Fourier transform infrared spectroscopic analysis

Control cells and cells exposed to different (sub)lethal concentrations of the citrus fruit extract for 90 min were harvested by centrifugation at 11 000 g for 3 min at 4°C and suspended in 50 μl of PBS, placed (15 μl) in a ZeSn window and stove-dried (15 min, 50°C). Infrared spectra were obtained with a FT-IR spectroscope (Perkin-Elmer 2000 FT-IR, Waltham, MA, USA) equipped with a DTGS detector. Measurements were recorded over the wavelength range of 3500–700 cm−1 with an interval of 1 cm−1. The spectral resolution was 4 cm−1. The final spectra were achieved averaging 20 scans. FT-IR experiments were performed in triplicate.

A software application developed for the Perkin-Elmer environment was used for transformation, including normalization (0 setting with absorption at 1800 cm−1; 1 setting at maximal absorption, located around 1650 cm−1), smoothing and first derivative. After transformation, spectra were recorded in ASCII format and processed (Maradona 1996; Mouwen et al. 2005).

The spectral data were subjected to multivariate statistical methods [hierarchical cluster analysis (HCA) and factor analysis (FA)] to separate spectra into different classes. Pearson's product moment correlation coefficient was used to measure the similarity between spectra, and strain clustering was achieved using Ward's algorithm. All of the analyses (calculation of coefficients, joining of variables, canonical analysis and graphical display) were carried out with the Statistica for Windows, version 7.0, program (Statsoft Inc., Tulsa, OK, USA).

Membrane integrity tests

Assessment of propidium iodide intake by flow cytometry

Bacterial cultures grown until stationary phase and exposed to different concentrations of the citrus fruit extract for 90 min were subsequently diluted in PBS to achieve a final cellular concentration of ~108 cells ml−1. Afterwards, propidium iodide (PI) (Molecular Probes, Life Technologies, Madrid, Spain; Invitrogen, Life Technologies) was added (final PI concentration of 0·1% v/v), and the mix was incubated in the dark at room temperature for 10 min. Prior to the analysis of the samples by flow cytometry, the cell suspension was centrifuged at 11 000 g for 3 min at 4°C, and the cellular pellet was suspended in PBS. Flow cytometry experiments were carried out using a CyAn-adp flow cytometer (Beckman Coulter, Brea, CA, USA). Samples were excited using a 488-nm air-cooled argon-ion laser. The instrument was set up with the following configuration: forward scatter (FS), side scatter (SS) and red fluorescence (613/20 nm) for PI. The results were collected on logarithmic scale. The cell population was selected by gating in a FS vs SS dot plot, which allowed for the exclusion of aggregates and cell debris. Fluorescence histograms were represented in single-parameter histograms. Data were analysed with Summit version 3.1 software (Cytomation, Fort Collins, CO, USA).

Measurement of cellular leakage

Aliquots of 3 ml of cell cultures exposed to different concentrations of the citrus fruit extract during 90 min were filtered through a 25-mm-diameter, 0·22-μm-pore-size Millex-GS syringe filter (Millex-GS, Millipore, Madrid, Spain). The presence of nucleic acids in the filtrate was checked by measuring the absorbance at 260 nm (Uvikon 810 spectrophotometer, Kontron analytical, Munich, Germany).

Results

The activity of a commercial citrus fruit extract against 34 strains belonging to three bacterial species of interest for animal health, Senterica, E. coli and B. hyodysenteriae, was assessed by means of the broth dilution method (Table 1). The citrus fruit extract was able to inhibit bacterial growth at very low concentrations. Thus, the minimum inhibitory concentrations (MICs) ranged from 20 to 80 ppm for Senterica (MIC50 of 80 ppm; MIC90 of 80 ppm), from 40 to 80 ppm for E. coli (MIC50 of 60 ppm; MIC90 of 80 ppm) and from 10 to 40 ppm for B. hyodysenteriae (MIC50 of 20 ppm; MIC90 of 40 ppm). It is worth noting that for all strains studied, the minimal inhibitory concentration resulted to be bactericidal (data not shown).

To further investigate the antibacterial activity and mode of action of the extract, three Salmonella Typhimurium strains with different susceptibilities were selected (strain RS38, with a MIC of 20 ppm, and strains SP11 and SP58, with a MIC of 80 ppm). Inactivation experiments were performed after exposure of the strains to different concentrations of the citrus fruit extract (5, 20, 80, 160 ppm) for 90 min (Fig. 1a). Whereas no significant reductions in cell numbers were observed with 5 and 20 ppm, the exposure to 80 ppm gave rise to a 2·6-log-cycle reduction for strain RS38, the most sensitive one. In the presence of 160 ppm, the log-cycle reductions achieved for strains RS38, SP11 and SP58 were 4·4, 2·0 and 2·2, respectively.

Figure 1.

Antibacterial activity of BIOLL against Salmonella Typhimurium strains. (a) Reduction in cell numbers (in log CFU ml−1) for S. Typhimurium strains RS38 (black), SP11 (grey) and SP58 (white) after exposure to different concentrations of BIOLL for 90 min. The averages of three independent experiments ± standard deviations are shown. (b) Transformed representative FT-IR spectra (1500–1200 cm−1) of S. Typhimurium RS38 nontreated cells (grey continuous line) and S. Typhimurium RS38 cells exposed for 90 min to 5 ppm (grey broken line), 20 ppm (black continuous line), 80 ppm (black broken line) or 160 ppm (black dotted line) of BIOLL. (c) Scatter plot obtained from factor analysis of the w3 (1500–1200 cm−1) spectral region, showing the distribution of control and BIOLL-treated samples of S. Typhimurium strains RS38, SP11 and SP58. Three replicates of each experimental condition are represented. (d) Increase in OD260 of cell-free filtrates of S. Typhimurium strains RS38 (black), SP11 (grey) and SP58 (white) after exposure to different concentrations of BIOLL for 90 min. The averages of three independent experiments ± standard deviations are shown.

To elucidate the effects caused by the citrus fruit extract on the cellular structure and composition, bacteria exposed to these (sub)lethal conditions were subsequently analysed using FT-IR spectroscopy. The spectrum of S. Typhimurium nontreated cells was visually similar to the spectra reported in previous studies (Al-Qadiri et al. 2008; Álvarez-Ordóñez and Prieto 2010; Álvarez-Ordóñez et al. 2010). Strong absorptions were detected in the five spectral regions (w1, w2, w3, w4 and w5) that characterize the major cellular constituents. The spectra obtained for treated samples showed visible changes, even without any spectral transformation, in the w3 spectral region. With the aim of minimizing methodological variability and amplifying the chemically based spectral differences, the spectra were further processed – that is, normalization, smoothing and first-derivatization of spectra were carried out. The signal transformation thus achieved made differences between spectral features much more prominent. Figure 1b shows the w3 region of representative transformed spectra of strain RS38 after its exposure to different concentrations of the citrus fruit extract for 90 min. Concentrations exceeding the MIC caused striking changes in this region of the infrared spectrum, which were especially located around 1400 cm−1, frequency where vibrations from the carboxylic functional group (mainly C–O symmetric stretching of the deprotonated carboxylate group) of fatty acids are reflected (Jiang et al. 2004). Changes in other regions of the spectra were limited (data not shown). When strains SP11 and SP58 were tested, similar changes in FT-IR spectra were observed (data not shown). Multivariate analysis of the first-derivative spectra revealed significant differences among phenotypes within the spectral region w3 (Fig. 1c). A FA of this spectral region allowed us to discriminate between four groups of samples: (i) cells of strain RS38 exposed to 160 ppm, (ii) cells of strain RS38 exposed to 80 ppm, (iii) cells of strain RS38 exposed to 20 ppm and cells of strains SP11 and SP58 exposed to 160 ppm (iv) and spectra belonging to strains RS38, SP11 and SP58 exposed to lower concentrations of the citrus fruit extract.

The leakage of intracellular material after exposure to the different concentrations of the extract for 90 min was assessed by measuring the optical density at 260 nm (OD260) of cell-free filtrates (Fig. 1d). An increase in OD260 is indicative of leakage of intracellular nucleic acids and, consequently, reflects a loss in membrane integrity (Sampathkumar et al. 2003; Álvarez-Ordóñez and Prieto 2010). The increase in OD260 was dependent on the citrus fruit extract concentration and S. Typhimurium strain. Thus, it was more marked at high concentrations of the extract, with strain RS38 showing in all cases the highest OD260 values. Membrane integrity was also assessed by measuring the intake of PI, dye that enters the cell when membrane integrity is compromised and binds to intracellular nucleic acids (Fig. 2). For the three S. Typhimurium strains tested, <12% of cells were PI positive when exposed to concentrations of the extract below the MIC (Table S1; Fig. 2). However, with concentrations equal to the MIC, two populations of cells were evident: a first population with an intact membrane and a second population with damaged membrane (embracing 67–79% of total cells). With concentrations over the MIC, near 100% of cells were PI stained.

Figure 2.

Flow cytometry histograms (red fluorescence) showing the uptake of propidium iodide (PI) by Salmonella Typhimurium strains RS38 and SP11 after their exposure to different concentrations of BIOLL for 90 min. Undamaged cells are embraced in R3 gate. Damaged cells are embraced in R2 gate.

In a similar way, two E. coli strains (EC65, with a MIC of 80 ppm, and EC66, with a MIC of 40 ppm) were further studied after their exposure to different concentrations of the citrus fruit extract (20, 40, 80, 160 ppm) for 90 min. Inactivation experiments showed that whereas at low concentrations (20 and 40 ppm), no significant reductions in cell numbers were achieved, at higher concentrations similar reductions were observed for strains EC65 and EC66, with 0·3–0·6 and 3·1–3·5 log cycles of inactivation after exposure to 80 and 160 ppm, respectively (Fig. 3a). FT-IR analyses indicated that changes caused by the citrus fruit extract in FT-IR spectra were mainly located in the w3 spectral region, as happened to S. Typhimurium strains. Figure 3b shows the w3 region of representative transformed spectra of strain EC65 after its exposure to different concentrations of the extract for 90 min. As previously described for S.Typhimurium, the main spectral modifications were found around 1400 cm−1. A similar effect was evident for strain EC66 (data not shown). After a FA of the w3 spectral region, it was possible to discriminate four groups of samples: (i) cells of strain EC66 exposed to 160 ppm, (ii) cells of strain EC66 exposed to 80 ppm, (iii) cells of strain EC65 exposed to 160 ppm (iv) and spectra belonging to strains EC65 and EC66 exposed to lower concentrations (Fig. 3c). Leakage of intracellular content was also monitored for strains EC65 and EC66 (Fig. 3d). The increase in OD260 of cell-free filtrates observed after exposure to the citrus fruit extract was in all cases higher for strain EC66 and was more marked at high concentrations of BIOLL. When the intake of PI was evaluated, it was observed that at low concentrations (20 and 40 ppm), <12% of cells were PI stained, while exposure to 80 ppm gave rise to the presence of a numerous subpopulation with damaged membrane, which embraces 51% (strain EC65) and 76% (strain EC66) of total cells (Table S2; Fig. 4). When strains were exposed to 160 ppm, near 100% of cells showed a damaged membrane.

Figure 3.

Antibacterial activity of BIOLL against Escherichia coli strains. (a) Reduction in cell numbers (in log CFU ml−1) for E. coli strains EC65 (black) and EC66 (white) after exposure to different concentrations of BIOLL for 90 min. The averages of three independent experiments ± standard deviations are shown. (b) Transformed representative FT-IR spectra (1500–1200 cm−1) of E. coli EC65 nontreated cells (grey continuous line) and E. coli EC65 cells exposed for 90 min to 20 ppm (grey broken line), 40 ppm (black continuous line), 80 ppm (black broken line) or 160 ppm (black dotted line) of BIOLL. (c) Scatter plot obtained from factor analysis of the w3 (1500–1200 cm−1) spectral region, showing the distribution of control and BIOLL-treated samples of E. coli strains EC65 and EC66. Three replicates of each experimental condition are represented. (d) Increase in OD260 of cell-free filtrates of E. coli strains EC65 (black) and EC66 (white) after exposure to different concentrations of BIOLL for 90 min. The averages of three independent experiments ± standard deviations are shown.

Figure 4.

Flow cytometry histograms (red fluorescence) showing the uptake of propidium iodide (PI) by Escherichia coli EC65 after its exposure to different concentrations of BIOLL for 90 min. Undamaged cells are embraced in R3 gate. Damaged cells are embraced in R2 gate.

In the case of B. hyodysenteriae, no inactivation experiments were performed, because this bacterial species does not form isolated colonies on agar plates, which hinders the enumeration of survivors. Nevertheless, for all strains assayed, the MIC was found to be bactericidal after plating in blood agar plates, as shown by the lack of haemolysis for samples exposed for long time periods to the MIC (data not shown). Performance of membrane integrity tests was not possible either. The outer membrane of B. hyodysenteriae is associated with the protoplasmic cylinder by unknown connections in a very loose manner, and therefore, the membrane is easily removed during gentle laboratorial manipulative procedures, such as centrifugation (Sellwood and Bland 1997). Thus, control cells unexposed to citrus fruit extract were already PI stained and showed high OD260 values after routine laboratorial manipulation (data not shown). Brachyspira hyodysenteriae strain B204 was selected for FT-IR spectroscopic analyses. Exposure of this strain to various concentrations of the citrus fruit extract (10, 20, 40, 80 and 160 ppm) for 90 min gave rise to noticeable changes in the FT-IR spectra. The characteristic effect above described for S. Typhimurium and E. coli in the w3 region of transformed spectra was less evident for treated B. hyodysenteriae samples (Fig. 5a). However, major spectral modifications occurred in the w4 region of samples exposed to concentrations exceeding the MIC (Fig. 5b). This region is informative mostly for the carbohydrates and polysaccharides of the cell wall (Álvarez-Ordóñez and Prieto 2012). A HCA of the w4 region from transformed spectra clearly discriminated the spectra obtained for samples exposed to concentrations over the MIC from those obtained at lower concentrations (Fig. 5c).

Figure 5.

Antibacterial activity of BIOLL against Brachyspira hyodysenteriae strains. (a) Transformed representative FT-IR spectra (1500–1200 cm−1) of B. hyodysenteriae B204 nontreated cells (grey continuous line) and B. hyodysenteriae B204 cells exposed for 90 min to 10 ppm (grey broken line), 20 ppm (grey dotted line), 40 ppm (black continuous line), 80 ppm (black broken line) or 160 ppm (black dotted line) of BIOLL. (b) Transformed representative FT-IR spectra (1200–900 cm−1) of B. hyodysenteriae B204 nontreated cells (grey continuous line) and B. hyodysenteriae B204 cells exposed for 90 min to 10 ppm (grey broken line), 20 ppm (grey dotted line), 40 ppm (black continuous line), 80 ppm (black broken line) or 160 ppm (black dotted line) of BIOLL. (c) Dendrogram obtained from hierarchical cluster analysis of the w4 (1200–900 cm−1) spectral region of control and BIOLL-treated B. hyodysenteriae B204 cells. Cluster analysis was performed with the Pearson product moment correlation coefficient and by the Ward's algorithm method. Three replicates of each experimental condition are represented.

Discussion

This study shows the great efficacy of BIOLL (a commercial citrus fruit extract) in both inhibiting the growth and killing three bacterial gastrointestinal pathogens of interest for animal health (Senterica, E. coli and B. hyodysenteriae). MIC, defined as the lowest concentration of the citrus fruit extract that avoids visible growth, was determined for a collection of strains through the broth dilution method. All the strains tested were very susceptible to the effect of the extract, with MIC ranging from 10 ppm (for some B. hyodysenteriae strains) to 80 ppm (for various Senterica and E. coli strains), very low concentrations in comparison with those previously described for several essential oils (Burt 2004). This strong activity may be associated with the presence within BIOLL of a wide range of compounds with antimicrobial properties, that is, ascorbic acid, citric acid, naringin, hesperidin, quercetin, rutin, naringenin and saponins (Grupo Omega, personal communication), which may show additive or synergic activities due to their action on different cellular targets. In addition, our experiments revealed that the citrus fruit extract acts at its MIC as a bactericidal agent against Senterica, E. coli and B. hyodysenteriae. However, inactivation experiments evidenced that short-term exposition (for up to 90 min) to the MIC caused low or no reductions in cell numbers, which shows that this bactericidal activity is time dependent. Thus, concentrations of the extract exceeding the MIC (2× or 4×) or longer exposition times were required to achieve significant reductions in bacterial populations. Nevertheless, although there were no immediate effects of the citrus fruit extract at its MIC on bacterial viability, clear effects in membrane integrity and composition were observed after 90-min exposures, as assessed by membrane integrity tests and FT-IR spectroscopic analyses (below described).

Although recent investigations have focused on the elucidation of the mechanism of action of different natural antimicrobials and essential oils, including citrus fruits oils (Xu et al. 2008; Fisher and Phillips 2009; Bouhdid et al. 2010; Devi et al. 2010; Lu et al. 2011; Lv et al. 2011; Silva et al. 2011; Muthaiyan et al. 2012), no definitive conclusions have been drawn yet. Considering the large number of different groups of chemical compounds present in most essential oils and, especially, in BIOLL, it is likely that their antibacterial activity was not due to one only specific mechanism but that several targets in the cell were affected (Skandamis et al. 2001; Carson et al. 2002). However, it has been speculated that the main cellular target is the membrane. Due to their lipophilic/hydrophobic character, essential oils partition in the lipids of bacterial cell membranes, which results in membrane disruption, membrane expansion, increased membrane fluidity and permeability, inhibition of respiration and alteration in ion transport processes (Trombetta et al. 2005). Moreover, certain components of essential oils also appear to act on cell proteins embedded in the cytoplasmic membrane through distortion of the lipid–protein interaction or through direct effects on hydrophobic regions of proteins (Knobloch et al. 1989; Juven et al. 1994; Sikkema et al. 1995).

In our study, the effects of the commercial citrus fruit extract on whole cellular composition and membrane integrity were monitored. For the first purpose, FT-IR spectroscopy was used. FT-IR spectroscopy is a methodology able to provide nondestructive, rapid, relevant information on microbial systematics, useful for classification and identification. In addition, it has been recently proposed as a useful technique for the study of the mechanisms of death induction after bacterial exposure to different antimicrobial compounds and adverse environmental conditions, with special focus on cytoplasmic membrane composition and structure (Álvarez-Ordóñez et al. 2011). Our results evidence that FT-IR spectra of S. Typhimurium, E. coli and B. hyodysenteriae are greatly affected after exposition for 90 min to the citrus fruit extract and that the changes observed are dependent on the extract concentration. Thus, spectral modifications were only evident at lethal concentrations equal to (for some strains), or exceeding (in most cases), the MIC. Another interesting finding was the fact that spectral changes observed were species dependent. Whereas for S. Typhimurium and E. coli, the main effects on FT-IR spectra were located in the w3 spectral region (1500–1200 cm−1), for B. hyodysenteriae the w4 region (1200–900 cm−1) was more seriously affected. This suggests that the citrus fruit extract shows a broad-spectrum antibacterial activity based on the action of several of its constituents on different cellular targets. The spectra obtained for S. Typhimurium and E. coli treated cells evidence that for these bacterial species, the more striking alterations were observed around 1400 cm−1, frequency where vibrations from carboxylic functional groups (mainly C–O symmetric stretching of the deprotonated carboxylate group) are reflected (Jiang et al. 2004). These changes suggest that the citrus fruit extract alters the macromolecular structure of bacterial membranes, mainly affecting carboxylic groups of membrane fatty acids. On the other hand, alterations in this region of the spectra were minimal for B. hyodysenteriae, which showed major changes in the w4 spectral region, where ring vibrations of functional groups C–O–C and C–O from polysaccharides and carbohydrates of the cell wall are recorded (Álvarez-Ordóñez and Prieto 2012), which suggests the presence of damage or conformational/compositional alterations in some or all of the components of the cell wall.

The ability of the commercial citrus fruit extract to damage the bacterial membrane was evaluated for S. Typhimurium and E. coli cells by measuring the release of intracellular contents and the uptake of PI, dye that enters the cell when membrane integrity is compromised and binds to intracellular nucleic acids. The exposure to concentrations equal to or higher than the MIC gave rise to the leakage of nucleic acids, as shown by the increase in OD260 of cell-free filtrates, which was more marked at high extract concentrations. This leakage of intracellular components suggests that the commercial citrus fruit extract may cause formation of pores in the cytoplasmic membrane. Regarding PI uptake experiments, in general, samples exposed to the MIC showed a relevant subpopulation of cells with damaged membrane, while part of the population remained undamaged. On the other hand, near 100% of cells showed a damaged membrane when concentrations higher than the MIC were used. Thus, flow cytometric results also evidence the capability of the citrus fruit extract to disrupt the cell membrane, rendering it more permeable. These findings are in agreement with previous studies that demonstrated that different essential oils are able to disturb the bacterial membrane, leading to a loss of cytoplasmic material (Moosavy et al. 2008; Fisher and Phillips 2009; Bouhdid et al. 2010; Devi et al. 2010; Lv et al. 2011).

On the basis of our findings, it can be postulated that BIOLL has a dual mode of action. First, it shows a physical effect on bacterial membranes, with the formation of pores and a consequent leakage of intracellular components. This physical effect is already evident in most cases after short-time exposures to the MIC, when no reduction in bacterial numbers has been achieved yet, and therefore, it is a reversible damage that can be repaired after culturing under optimal conditions. However, bacterial exposition to concentrations exceeding the MIC causes a chemical effect on the cellular envelopes, which is dependent on the bacterial species, as shown by FT-IR spectroscopic results. Thus, whereas for S.Typhimurium and E. coli, the changes are focused on the carboxylic group of membrane fatty acids, for B. hyodysenteriae the main effects are found in polysaccharides and carbohydrates of the cell wall. Exposition to the MIC for longer time periods is likely to cause similar effects on FT-IR spectra. This chemical effect is already linked to a reduction in bacterial numbers, which suggests that the damage caused is at this point irreversible. The proposed mode of action, with dual physico-chemical effects on bacterial envelopes and species-specific cellular targets, may be associated with a decreased likelihood of bacteria developing resistances. This fact, added to the great antibacterial activity showed by BIOLL, with MIC values ≤80 ppm, shows the attractiveness of this commercial citrus fruit extract as an alternative to antibiotics for its inclusion as a feed additive in animal nutrition. Nevertheless, further research in clinical trials under field conditions is necessary to verify its efficacy without detrimental side effects.

Acknowledgements

This work was funded by ‘Ministerio de Economía y Competitividad’ (project AGL 2010-18804). We acknowledge the contribution of Grupo Omega. The authors declare no conflict of interests.

Ancillary