New insights in mechanisms of bacterial inactivation by carvacrol

Authors


Correspondence

Diego García-Gonzalo, Departamento de Producción Animal y Ciencia de los Alimentos, Facultad de Veterinaria, Universidad de Zaragoza, C/ Miguel Servet, 177, 50013, Zaragoza, Spain.

E-mail: Diego.Garcia@unizar.es

Abstract

Aims

To study the mechanism of bacterial inactivation by carvacrol and the influence of genetic and environmental factors in its antimicrobial activity.

Methods and Results

In general, bacterial inactivation by carvacrol was higher in the Gram-positive Listeria monocytogenes than in the Gram-negative Escherichia coli and at acidic pH. At pH 4·0, 25 μl l−1 of carvacrol for 5 h inactivated 1 and more than 5 log10 cycles of E. coli and L. monocytogenes populations, respectively. Genetic and environmental factors also influenced cell resistance to carvacrol: rpoS and sigB deletion decreased carvacrol resistance in E. coli and L. monocytogenes, respectively; a heat shock induced a phenomenon of cross-protection to carvacrol treatments. Repair of sublethal injuries in cell envelopes suggested that carvacrol targets lipid fractions and proteins of these structures. This result was corroborated by attenuated total reflectance infrared microspectroscopy analysis.

Conclusions

This study shows critical genetic and environmental factors, such as rpoS or sigB and heat shocks, and reveals new microbial structures involved in the mechanism of bacterial inactivation by carvacrol.

Significance and Impact of the Study

A better understanding of the mechanisms of microbial inactivation is of great relevance to design more appropriate carvacrol treatments with high antimicrobial effects.

Introduction

Essential oils extracted from aromatic plants are attracting much attention from the food industry because of their consideration as GRAS (generally recognized as safe) compounds (FDA 2011). Essential oils extracted from spices, aromatic herbs, fruits and flowers are mixtures of multiple compounds, including carvacrol (Burt 2004). Carvacrol is an isoprenyl phenol whose antimicrobial activity in food preservation treatments has been studied by many scientists (Kim et al. 1995; Karatzas et al. 2001; Roller and Seedhar 2002; Kisko and Roller 2005; Ait-Ouazzou et al. 2011). A proper understanding of the mechanisms of inactivation is of great relevance when designing more appropriate carvacrol treatments with high antimicrobial effects.

Cell envelopes are considered as a major target of carvacrol. Different studies have shown that carvacrol disintegrates outer membrane properties (Helander et al. 1998; Di Pasqua et al. 2007), induces morphology modifications (La Storia et al. 2011), causes cell permeabilization, dissipation of pH gradients, leakage of inorganic ions and loss of membrane potential (Ultee et al. 1999, 2002; Lambert et al. 2001; Veldhuizen et al. 2006; Xu et al. 2008) and perturbs lipid fractions of bacterial cytoplasmic membranes (Cristani et al. 2007). Furthermore, bacterial susceptibility experiments to carvacrol-like compounds indicated that hydrophobicity, the presence of a free hydroxyl group and a delocalized system allowing proton exchange are key factors in the mechanism of inactivation by carvacrol (Ultee et al. 2002; Ben Arfa et al. 2006). All these studies have revealed critical events and structures affected by carvacrol and that the mechanism of inactivation of carvacrol is related to cell envelopes and the ability of carvacrol to permeabilize and depolarize the cytoplasmic membrane. However, genetic and environmental factors influencing the mechanisms of inactivation by carvacrol and its precise targets have not been accurately described.

On the one hand, microbial cells encounter multiple stresses during food preservation treatments, such as heat shocks, that can induce cross-resistance to subsequent treatments (Wesche et al. 2009). Cross-resistance phenomena as a consequence of a sublethal stress are regulated, among others, by alternative sigma factors, such as σS (encoded by rpoS; Vidovic et al. 2012) in Escherichia coli and σB in Listeria monocytogenes (encoded by sigB; Chung et al. 2006; Chaturongakul et al. 2008). Moreover, σS and σB also modify bacterial resistance to food preservation technologies, including essential oils (Wemekamp-Kamphuis et al. 2004; O'Byrne and Karatzas 2008; Somolinos et al. 2010b; Ait-Ouazzou et al. 2012). Thus, the study of the influence of cross-resistance and both sigma factors in carvacrol bacterial resistance would be of great interest to designers of food preservation treatments. This study was performed with stationary phase cells because bacterial resistance to food preservation technologies uses to be maximum at this growth stage (Mañas and Pagán 2005). In addition, as rpoS and sigB levels are highly expressed in stationary phase cultures, these conditions allow comparing resistance of wild-type and deletion mutant cells to evaluate the role of these factors on bacterial resistance to carvacrol.

On the other hand, a better description of microbial components involved in the mechanism of inactivation by carvacrol would help in identifying the changes and modifications of microbial structures, as a previous step to design efficient combined processes.

The objectives of this work were as follows: (i) to study E. coli and L. monocytogenes inactivation by carvacrol, describing the effect of the pH of the treatment medium and the treatment time; (ii) to determine the influence of alternative sigma factors σS and σB and the effect of a previous heat shock on bacterial resistance to carvacrol; (iii) to evaluate the biosynthetic requirements to repair sublethal injuries caused by carvacrol in bacterial envelopes to identify affected structures and (iv) to identify the bacterial components affected by carvacrol through ATR-IRMS spectroscopy.

Materials and methods

Micro-organisms and growth conditions

The strains used were E. coli BJ4 (Krogfelt et al. 1993) and its isogenic ΔrpoS mutant BJ4L1 (Krogfelt et al. 2000), and L. monocytogenes EGD-e (Murray et al. 1926) and its isogenic deletion mutant ΔsigB (Chatterjee et al. 2006). One glass bead from a −80°C stock culture was transferred into a tube containing 5 ml of tryptic soy broth (Biolife, Milan, Italy) with 0·6% yeast extract added (Biolife; TSBYE), which was incubated overnight at 37°C (E. coli strains) or 30°C (L. monocytogenes strains) to enable the comparison of these data with previous studies (Wemekamp-Kamphuis et al. 2004; Somolinos et al. 2010a; Ait-Ouazzou et al. 2011, 2012; Espina et al. 2011). With these subcultures, 250-ml Erlenmeyer flasks containing 50 ml of TSBYE were inoculated to a concentration of 104 CFU ml−1. These flasks were incubated under agitation (130 rpm; Selecta, mod. Rotabit, Barcelona, Spain) at the appropriate temperature (see above), for 24 h, until the stationary growth phase was reached. As reported previously (Krogfelt et al. 2000; Hain et al. 2008), growth patterns of wild-type and mutant cells were similar.

Measurement of bacterial inactivation by carvacrol

Carvacrol, as with many plant essential oil constituents, is practically immiscible in water. As solvents and detergents might decrease the antimicrobial effect of essential oils (Burt 2004) and might also damage cell envelopes, these substances were not used in these experiments. As described by Friedman et al. (2002), carvacrol suspensions in citrate–phosphate buffer (McIlvaine buffer) at pH 7·0 (27·09 g l−1) and pH 4·0 (23·85 g l−1; Dawson et al. 1974) were prepared by vortex shaking (Genius 3, Ika, Königswinter, Germany). Cells from stationary phase cultures were added at final concentrations of 2·107 CFU ml−1 to citrate–phosphate buffers at pH 7·0 or pH 4·0, with or without carvacrol at concentrations of 25, 50, 100 and 200 μl l−1. Buffer pH was not modified by the addition of carvacrol. Carvacrol treatments were carried out at 20°C for 24 h. Samples were taken at regular intervals (10 min, 1, 3, 5 h and 24 h), and survivors were enumerated. These treatment conditions were chosen according to the previously published data (Espina et al. 2010, 2011; Somolinos et al. 2010b; Ait-Ouazzou et al. 2011). Untreated cells were insensitive to incubation in citrate–phosphate buffers at pH 7·0 or pH 4·0 for 24 h at 20°C (data not shown).

Survival counts

After treatments, samples were adequately diluted in 0·1% w/v peptone water (Biolife) containing 1% v/v Tween 80 (Biolife) as a neutralizer. Next, 0·1 ml of each samples was pour plated onto tryptic soy agar (Biolife), with 0·6% of yeast extract added (TSAYE) as a nonselective recovery medium. The plates were incubated for 24 h at 37°C (E. coli strain) or 48 h at 30°C (L. monocytogenes strain). Longer incubation times did not influence survival counts (data not shown). After incubation, colony forming units (CFU) were counted with an improved image analyser automatic counter, as previously described (Condón et al. 1996). Inactivation was expressed in terms of the extent of reduction in log10 counts after any treatment.

Detection of sublethal injury

The method used to determine the occurrence of sublethal carvacrol damage consisted of plating survivors after treatments in two culture media: a nonselective one, which allows cells to repair sublethal damage and recover, and a selective one, in which survivors are not capable of repairing their injuries and finally do not multiply (Mackey 2000). The simultaneous use of nonselective and selective recovery media provides information about the protection mechanisms with respect to environmental factors, either based on a modification of the intrinsic resistance of bacterial cells or on the capacity of bacterial cells to repair sublethal damage.

To evaluate cytoplasmic and outer membrane damage after carvacrol treatments, sodium chloride and bile salts were added to a nonselective recovery medium, respectively. The loss of resistance to sodium chloride and bile salts is likely to be multifactorial. However, sensitivity to sodium chloride implies a loss of osmotic functionality of the cytoplasmic membrane, whereas sensitivity to bile salts reflects the loss of outer membrane integrity and/or impairment of multidrug efflux systems by the loss of proton motive force (Thanassi et al. 1997; Mackey 2000). We have used the terms injury/damage as convenient shorthand to indicate loss of functions associated in part with outer or cytoplasmic membranes.

After carvacrol treatments, samples were plated onto TSAYE, TSAYE with 3% (E. coli strains) or TSAYE with 6% (L. monocytogenes strains) sodium chloride (Fisher Scientific, Loughborough, United Kingdom) added (TSAYE-SC), and onto TSAYE with 0·35% (E. coli strains) bile salts (Biolife) added (TSAYE-BS) as previously described (Ait-Ouazzou et al. 2012). These levels of sodium chloride and bile salts were determined as the maximum noninhibitory concentrations for stationary phase untreated cells (data not shown). Plates containing selective media were incubated for 24 h longer than those containing nonselective medium. Longer incubation times did not influence survival counts (data not shown). The number of sublethally injured cells was expressed as the difference between the survival counts (CFU ml−1) on nonselective medium (TSAYE) and the survival counts on selective media (TSAYE-SC and TSAYE-BS) after treatments.

Heat shock

One 5-ml aliquot of stationary phase suspension was immersed in a thermostated water bath (mod. Digiterm; Selecta, Barcelona, Spain) at 45°C and held at this temperature for 2 h (E. coli strains) (Somolinos et al. 2008) or 1 h (L. monocytogenes strains) (Somolinos et al. 2010a). After this heat shock, cell suspensions were centrifuged at 6000 g for 5 min and resuspended in the buffer for carvacrol treatments. No inactivation or sublethal damage was detected after the heat shock treatment.

Measurement of biosynthetic requirements for the repair of sublethally injured cells

After a treatment with 200 μl l−1 of carvacrol at an initial inoculum size of 2·107 CFU ml−1 of E. coli BJ4 at pH 7·0 for 3 h, treated cells were incubated into sterile TSBYE as a repair medium: alone or with the addition of sodium azide (2000 mg l−1), chloramphenicol (50 mg l−1), cerulenin (70 mg l−1) or penicillin G (10 mg l−1; García et al. 2006) as inhibitors of energy, protein, lipid and peptidoglycan synthesis, respectively, and were maintained at 37°C for 4 h.

In the case of L. monocytogenes EGD-e, treated under the same conditions described for E. coli BJ4 (see above), cells were incubated in sterile TSBYE as a repair medium: alone or with sodium azide (50 mg l−1), chloramphenicol (100 mg l−1), cerulenin (70 mg l−1) or penicillin G (1 mg l−1) added (Somolinos et al. 2010a), and they were incubated at 30°C for 4 h.

Samples were taken at intervals, and viable counts were evaluated on the nonselective and selective media, as described previously. These antibiotic concentrations were the minimum inhibitory growth concentrations for native cells.

Attenuated total reflectance infrared microspectroscopy (ATR-IRMS) with multivariate analysis

Escherichia coli BJ4 cells (initial concentration: 2 × 107 CFU ml−1) were treated for 24 h at 20°C in the absence or presence of 200 μl l−1 of carvacrol in citrate–phosphate buffers of pH 4·0 or 7·0. An aliquot of treated cell suspensions was centrifuged at 6000 g for 10 min at 4°C. Pellets were washed three times with 1 ml of 0·9% NaCl and centrifuged at 6000 g for 10 min. Pellets were placed onto grids of hydrophobic membrane (HGM; ISO-GRID, Neogen Corporation, Lansing, MI, USA) and dried out under laminar flow at room temperature for 1 h. Samples were analysed by IR equipment (Illuminate IR, Smiths detection, The Genesis Centre Science Park South Birchwood Warrington, UK) interfaced with mercury–cadmium–telluride photoconductive detector and equipped with a microscope with a motorized xy stage, 20× and 50× objectives, and slide-on attenuated total reflection (ATR) diamond objective (Smiths Detection).

The hydrophobic membranes were placed on the stage of the microscope, and a position of the microbial pellet was selected with the assistance of the microscope and live camera (Leica OM 2500, Module FT-IR; Renishaw plc, New Mills, Wotton-under-Edge, Gloucestershire, UK). The microscope was software controlled using Wire ver. 3.2 software (Renishaw plc).

Spectra were collected from 4000 to 800 cm−1 with a resolution of 4 cm−1. The spectrum of each sample was obtained by taking the average of 128 scans. Spectra were displayed in terms of absorbance obtained by rationing the single-beam spectrum against that of the air background. The spectrometer was completely software controlled by synchronize IR basic ver. 1.1 software (SensIR Technologies, Smiths Detection). Pirouette® multivariate analysis software (ver. 4.0, InfoMetrix, Inc., Woodville, WA, USA) was used to analyse the raw spectra of bacterial cells. The IR spectral data were mean centred, transformed to their second derivative using a 15-point Savitzky-Golay polynomial filter and vector-length normalized; sample residuals and Mahalanobis distance were used to determine outliers. Discriminating power was used to define the wavenumbers that have a predominant effect on the classification of bacterial cells when comparing two different samples. Spectra were obtained from three independent samples.

Statistical analysis

Experiments were carried out in triplicate on different working days. Inactivation was expressed in terms of the extent of reduction in log10 counts after every treatment. The error bars in the figures indicate the mean ± standard deviations from the data obtained from at least three independent experiments. Analyses of variance (P = 0·05) were calculated using spss software (SPSS, Chicago, IL, USA).

Results

Microbial inactivation by carvacrol

Listeria monocytogenes and E. coli were chosen as Gram-positive and Gram-negative representatives, respectively. Stationary phase cells at an initial concentration of 2 × 107 CFU ml−1 were treated by carvacrol. Figure 1 shows the inactivation levels reached with increasing carvacrol concentrations (0, 25, 50, 100 and 200 μl l−1) after 5 h in a population of wild-type L. monocytogenes EGD-e and E. coli BJ4 treated in buffer at pH 4·0 and 7·0. The higher the carvacrol concentration, the greater the inactivation levels reached: while 25 μl l−1 of carvacrol at pH 7·0 inactivated 1 log10 cycle of the initial population of L. monocytogenes cells in 5 h, 200 μl l−1 of carvacrol killed 3 log10 cycles of the same population during the same treatment time. Acidification of treatment medium increased the efficacy of carvacrol: inactivation after 5 h of L. monocytogenes with 25 μl l−1 of carvacrol was 4 log10 cycles higher when pH of the treatment medium was 4·0 instead of 7·0. The influence of duration treatment, carvacrol concentration and pH of the treatment medium was similar for E. coli inactivation by carvacrol. In general, E. coli cells were more resistant to carvacrol than L. monocytogenes at both treatment pHs. At pH 4·0, 50 μl l−1 of carvacrol for 5 h inactivated 1 and more than 5 log10 cycles of the initial cell population of E. coli and L. monocytogenes, respectively.

Figure 1.

Bacterial inactivation by carvacrol. Log10 of survival fraction of Listeria monocytogenes EGD-e (□,■) and Escherichia coli BJ4 (image_n/jam12028-gra-0004.png,image_n/jam12028-gra-0003.png) and treated for 5 h with carvacrol at different concentrations (25, 50, 100 and 200 μl l−1) in McIlvaine buffer pH 7·0 (□,image_n/jam12028-gra-0004.png) and 4·0 (■,image_n/jam12028-gra-0003.png). Survivors were recovered in nonselective medium (TSAYE). Data represent means ± standard deviations (error bars). The dotted line represents the detection limit.

Similar conclusions can be drawn from Tables 1 and 2, showing the influence of time, pH and carvacrol concentration on the inactivation of E. coli and L. monocytogenes in the studied range (0, 10 min, 1, 3, 5 h and 24 h; pH 4·0 and 7·0; 25, 50, 100 and 200 μl l−1 of carvacrol).

Table 1. Inactivation of Listeria monocytogenes by carvacrol. Log10 of survival fraction of L. monocytogenes EGD-e and ΔsigB cells after a carvacrol treatment (25, 50, 100 and 200 μl l−1) at different treatment times (10 min, 1, 3, 5 and 24 h): Carvacrol treatments were applied in McIlvaine buffers at pH 4·0 or 7·0. Survivors were recovered in nonselective medium (TSAYE) or selective medium added with sodium chloride (TSAYE-SC)
   TSAYETSAYE-SC
   25 μl l−150 μl l−1100 μl l−1200 μl l−125 μl l−150 μl l−1100 μl l−1200 μl l−1
  1. Data represent means ± standard deviations.

pH 4·0L. monocytogenes EGD-e10 min−0·19 ± 0·02−0·29 ± 0·01−0·50 ± 0·25−1·72 ± 0·26−0·53 ± 0·35−0·85 ± 0·57−2·20 ± 0·42<−5·00
1 h<−5·00<−5·00<−5·00<−5·00<−5·00<−5·00<−5·00 
L. monocytogenes EGD-e ΔsigB10 min−0·35 ± 0·21−1·39 ± 0·30−2·04 ± 0·35−3·25 ± 0·37−0·67 ± 0·16−2·25 ± 0·35−3·10 ± 0·42<−5·00
1 h<−5·00<−5·00<−5·00<−5·00<−5·00<−5·00<−5·00 
pH 7·0L. monocytogenes EGD-e10 min−0·19 ± 0·01−0·22 ± 0·04−0·25 ± 0·04−0·30 ± 0·04−0·24 ± 0·02−0·26 ± 0·09−0·33 ± 0·07−0·39 ± 0·09
1 h−0·38 ± 0·17−0·55 ± 0·06−0·58 ± 0·03−0·60 ± 0·08−0·56 ± 0·16−0·72 ± 0·04−0·92 ± 0·11−0·96 ± 0·19
3 h−0·71 ± 0·03−0·81 ± 0·03−1·05 ± 0·07−0·79 ± 0·25−0·84 ± 0·05−0·96 ± 0·05−1·30 ± 0·11−3·58 ± 0·75
5 h−0·74 ± 0·05−0·94 ± 0·03−1·24 ± 0·08−2·85 ± 0·62−1·15 ± 0·18−1·30 ± 0·16−1·77 ± 0·614·39 ± 0·55
24 h−0·87 ± 0·02−1·07 ± 0·18−2·44 ± 0·21<−5·00−1·26 ± 0·09−1·34 ± 0·36−3·09 ± 0·33<−5·00
L. monocytogenes EGD-e ΔsigB10 min−0·12 ± 0·02−0·14 ± 0·01−0·14 ± 0·02−0·16 ± 0·02−0·13 ± 0·04−0·14 ± 0·02−0·16 ± 0·01−0·17 ± 0·04
1 h−0·32 ± 0·20−0·39 ± 0·01−0·39 ± 0·01−0·52 ± 0·04−0·62 ± 0·04−0·68 ± 0·10−0·68 ± 0·14−0·69 ± 0·14
3 h−0·52 ± 0·04−0·56 ± 0·06−0·71 ± 0·10−0·84 ± 0·05−0·68 ± 0·03−−0·82 ± 0·08−1·13 ± 0·04−2·34 ± 0·20
5 h−0·58 ± 0·04−0·77 ± 0·04−0·88 ± 0·11−1·21 ± 0·30−0·95 ± 0·07−1·27 ± 0·07−1·77 ± 0·64−4·25 ± 0·35
24 h−0·66 ± 0·06−0·75 ± 0·07−1·56 ± 0·65<−5·00−1·06 ± 0·18−1·30 ± 0·47−2·42 ± 0·23<−5·00
Table 2. Inactivation of Escherichia coli by carvacrol. Log10 of survival fraction of E. coli BJ4 and ΔrpoS cells after a carvacrol treatment (25, 50, 100 and 200 μl l−1) at different treatment times (10 min, 1, 3, 5 and 24 h) in McIlvaine buffers at pH 4·0 or 7·0. Survivors were recovered in nonselective medium (TSAYE), selective medium added with sodium chloride (TSAYE-SC) or selective medium added with bile salts (TSAYE-BS)
   TSAYETSAYE-SCTSAYE-BS
   25 μl l−150 μl l−1100 μl l−1200 μl l−125 μl l−150 μl l−1100 μl l−1200 μl l−125 μl l−150 μl l−1100 μl l−1200 μl l−1
  1. Data represent means ± standard deviations.

pH 4·0E. coli BJ410 min−0·14 ± 0·03−0·18 ± 0·01−0·25 ± 0·02−0·35 ± 0·18−0·19 ± 0·02−0·20 ± 0·03−0·31 ± 0·03−0·42 ± 0·23−0·30 ± 0·07−0·35 ± 0·07−0·41 ± 0·03−0·48 ± 0·30
1 h−0·21 ± 0·02−0·25 ± 0·01−0·33 ± 0·05−0·91 ± 0·41−0·30 ± 0·08−0·33 ± 0·09−0·48 ± 0·10−0·95 ± 0·60−0·48 ± 0·05−0·56 ± 0·14−−0·70 ± 0·19−1·55 ± 0·10
3 h−0·59 ± 0·05−0·68 ± 0·03−0·78 ± 0·03<−5·00−0·65 ± 0·30−0·77 ± 0·37−0·80 ± 0·16<−5·00−0·82 ± 0·19−1·01 ± 0·07−1·08 ± 0·11<−5·00
5 h−0·77 ± 0·25−1·16 ± 0·06−1·43 ± 0·03 −0·77 ± 0·11−1·25 ± 0·22−1·64 ± 0·27 −1·00 ± 0·25−1·63 ± 0·07−2·68 ± 0·11 
24 h−0·79 ± 0·13−1·59 ± 0·28−3·74 ± 0·32 −0·92 ± 0·43−3·60 ± 0·11<−5·00 −1·09 ± 0·15−3·84 ± 0·09<−5·00 
E. coli BJ4 ΔrpoS10 min−0·17 ± 0·02−0·19 ± 0·04−0·27 ± 0·03−0·37 ± 0·18−0·17 ± 0·03−0·25 ± 0·04−0·26 ± 0·04−0·58 ± 0·11−0·20 ± 0·07−0·35 ± 0·14−0·20 ± 0·26−0·72 ± 0·17
1 h−0·25 ± 0·02−0·26 ± 0·42−0·38 ± 0·03−1·07 ± 0·47−0·31 ± 0·11−0·38 ± 0·11−0·60 ± 0·21−2·13 ± 0·24−0·590 ± 24−0·64 ± 0·23−0·81 ± 0·28−2·68 ± 0·16
3 h−0·91 ± 0·31−1·05 ± 0·06−1·27 ± 0·38<−5·00−1·21 ± 0·08−1·11 ± 0·41−1·47 ± 0·31<−5·00−1·37 ± 0·02−1·52 ± 0·33−1·71 ± 0·41<−5·00
5 h−1·35 ± 0·08−1·32 ± 0·18−2·90 ± 0·15 −1·59 ± 0·27−1·70 ± 0·28−3·55 ± 0·15 −1·79 ± 0·42−2·07 ± 0·18−4·17 ± 0·30 
24 h−1·56 ± 0·16−2·48 ± 0·11<−5·00 −2·29 ± 0·13−3·79 ± 0·15<−5·00 −3·05 ± 0·14−3·58 ± 0·58<−5·00 
pH 7·0E. coli BJ410 min−0·14 ± 0·02−0·16 ± 0·01−0·17 ± 0·02−0·19 ± 0·02−0·14 ± 0·05−0·17 ± 0·03−0·18 ± 0·03−0·22 ± 0·04−0·28 ± 0·04−0·18 ± 0·03−0·23 ± 0·04−0·28 ± 0·04
1 h−0·23 ± 0·05−0·27 ± 0·10−0·38 ± 0·03−0·47 ± 0·18−0·29 ± 0·04−0·32 ± 0·19−0·49 ± 0·05−0·64 ± 0·21−0·39 ± 0·09−0·43 ± 0·19−0·78 ± 0·03−1·20 ± 0·41
3 h−0·19 ± 0·01−0·37 ± 0·10−0·49 ± 0·12−0·74 ± 0·14−0·41 ± 0·24−0·38 ± 0·31−0·80 ± 0·07−3·69 ± 0·41−0·45 ± 0·07−0·42 ± 0·12−0·88 ± 0·10−4·19 ± 0·43
5 h−0·30 ± 0·03−0·52 ± 0·17−0·54 ± 0·20−1·34 ± 0·18−0·45 ± 0·21−0·50 ± 0·43−0·81 ± 0·05−4·25 ± 0·35−0·53 ± 0·09−0·67 ± 0·25−1·14 ± 0·23<−5·00
24 h−0·49 ± 0·01−0·65 ± 0·07−1·28 ± 0·59<−5·00−0·54 ± 0·23−0·67 ± 0·11−1·84 ± 0·22<−5·00−0·68 ± 0·24−0·93 ± 0·61−2·64 ± 0·20 
E. coli BJ4 ΔrpoS10 min−0·15 ± 0·04−0·18 ± 0·03−0·19 ± 0·02−0·29 ± 0·01−0·18 ± 0·11−0·24 ± 0·05−0·26 ± 0·09−0·34 ± 0·08−0·15 ± 0·04−0·28 ± 0·06−0·32 ± 0·10−0·50 ± 0·13
1 h−0·32 ± 0·08−0·30 ± 0·04−0·47 ± 0·04−0·60 ± 0·07−0·32 ± 0·04−0·33 ± 0·09−0·55 ± 0·07−0·72 ± 0·30−0·51 ± 0·05−0·61 ± 0·05−0·99 ± 0·25−1·32 ± 0·08
3 h−0·31 ± 0·19−0·46 ± 0·20−1·00 ± 0·02−1·46 ± 0·25−0·36 ± 0·20−0·50 ± 0·17−1·14 ± 0·65−4·53 ± 0·13−0·62 ± 0·10−0·54 ± 0·23−1·56 ± 0·32<−5·00
5 h−0·35 ± 0·22−0·60 ± 0·25−1·34 ± 0·12<−5·00−0·38 ± 0·14−0·65 ± 0·46−1·44 ± 0·63<−5·00−0·76 ± 0·23−0·96 ± 0·13−1·64 ± 0·50 
24 h−0·61 ± 0·10−0·68 ± 0·16−1·53 ± 0·22 −0·64 ± 0·07−0·70 ± 0·35−1·95 ± 0·36 −0·90 ± 0·56−1·04 ± 0·22−2·03 ± 0·67 

Influence of sigB and rpoS on microbial resistance to carvacrol

Figure 2 compares the inactivation of stationary phase wild-type and mutant strains as a function of the pH of the treatment medium. Deletion of the general stress–response alternative sigma factor rpoS decreased E. coli resistance to carvacrol at every treatment medium pH (Fig. 2A; P < 0·05). However, deletion of sigB led to a decrease in L. monocytogenes carvacrol resistance only at pH 4·0 (Fig. 2B): sigB mutant cells treated by carvacrol at pH 7·0 were more resistant than the wild-type cells (P < 0·05).

Figure 2.

Influence of rpoS and sigB on bacterial inactivation by carvacrol and occurrence of sublethal injury. Log10 of survival fraction of Escherichia coli BJ4 wild type (WT) and ΔrpoS mutants (A) and Listeria monocytogenes EGD-e wild type (WT) and ΔsigB mutants (B) treated with carvacrol in McIlvaine buffers at pH 4·0 and 7·0. E. coli BJ4 wild type and ΔrpoS mutants were treated for 5 h with 100 μl l−1 of carvacrol at pH 4·0 and with 200 μl l−1 of carvacrol at pH 7·0. Listeria monocytogenes EGD-e wild type and ΔsigB mutants were treated with 200 μl l−1 of carvacrol for 10 min at pH 4·0 and for 5 h at pH 7·0. Survivors were recovered in nonselective medium (TSAYE) (□), selective medium added with sodium chloride (TSAYE-SC) (image_n/jam12028-gra-0004.png) or selective medium added with bile salts (TSAYE-BS) (■). Data represent means ± standard deviations (error bars). The dotted line represents the detection limit.

Tables 1 and 2 show that, although differences between wild-type and mutant cells were evident for treatments with 25 μl l−1 of carvacrol, they were higher with 100 and 200 μl l−1 of carvacrol.

Occurrence of sublethal injuries in cell envelopes by carvacrol

Figure 2 also shows that plating of survivors in selective media reduced the number of survival counts as a function of the treatment conditions. Therefore, carvacrol induced sublethal injuries in the cell envelopes of L. monocytogenes EGD-e and its ΔsigB mutant, and E. coli BJ4 and its ΔrpoS mutant at pH 4·0 and 7·0 (P < 0·05). For example, recovery of L. monocytogenes EGD-e in nonselective and selective medium after a 10-min treatment with 200 μl l−1 of carvacrol at pH 4·0 (Fig. 2B) showed a difference of more than 3 log10 cycles in survival counts. In other words, more than 99·9% of L. monocytogenes EGD-e survivors were sublethally damaged in the cytoplasmic membrane. Similarly, the comparison of E. coli survivors evaluated in both nonselective and selective recovery media after carvacrol treatments (Fig. 2A) showed that the cytoplasmic and the outer membrane were sublethally injured. Although survival counts of E. coli ΔrpoS cells in selective media were lower than in the wild-type cells (P < 0·05), the proportion of survivors with sublethal damages in their cytoplasmic membranes, estimated from the difference in survival counts between nonselective and selective media, was similarly independent of the presence of rpoS (Fig. 2A; P > 0·05).

Owing to the high susceptibility of L. monocytogenes cells at pH 4·0, survival counts in selective media exceeded our detection limit (5 log10 cycles of inactivation). This circumstance did not allow us to study the effect of sigB in occurrence of sublethal damage by carvacrol in L. monocytogenes envelopes in detail.

Owing to big differences observed in bacterial resistance, specific data in our detection range were selected to best illustrate differences between wild-type and mutant cells at both pHs in Fig. 2. Comparison of survival counts in nonselective and selective media in Tables 1 and 2 shows that the higher proportion of sublethal injuries occurred at the more intense treatment conditions, such as 3, 5 and 24 h with 100 and 200 μl l−1 of carvacrol.

Influence of a heat shock prior to carvacrol treatment

Figure 3 shows the number of log10 cycles of inactivation after a treatment with carvacrol at pH 7·0 (Fig. 3A) and 4·0 (Fig. 3B). As a result of a previous heat shock, both L. monocytogenes and E. coli wild-type and mutant strains increased their resistance to carvacrol (P < 0·05). While a treatment for 24 h with 200 μl l−1 of carvacrol at pH 7·0 (Fig. 3A) inactivated more than 5 log10 cycles of the initial population of native cells, inactivation of heat-shocked cells before the carvacrol treatment was below 2 log10 cycles. Similar resistance between mutant and wild-type strains was detected under these conditions (P > 0·05).

Figure 3.

Influence of a heat shock on bacterial inactivation by carvacrol. Log10 of survival fraction of Listeria monocytogenes EGD-e wild type (WT) and ΔsigB mutants (□,image_n/jam12028-gra-0002.png) and Escherichia coli BJ4 wild type (WT) and ΔrpoS mutants (image_n/jam12028-gra-0004.png,image_n/jam12028-gra-0003.png) after treatments with 200 μl l−1 carvacrol. In McIlvaine buffer at pH 7·0 (A), treatment time was 24 h for both L. monocytogenes and E. coli. In McIlvaine buffer at pH 4·0 (B), L. monocytogenes cells were treated for 1 h and E. coli cells for 3 h. Cells were not shocked (□,image_n/jam12028-gra-0004.png) or heat shocked (image_n/jam12028-gra-0002.png,image_n/jam12028-gra-0003.png) at 45°C for 1 h (Lmonocytogenes) or 2 h (E. coli). Survivors were recovered in nonselective medium (TSAYE). Data represent means ± standard deviations (error bars). The dotted line represents the detection limit.

When the carvacrol treatment was applied at pH 4·0 (Fig. 3B), a previous heat shock of wild-type cells also increased their resistance (P < 0·05). However, owing to the low resistance of mutant strains, survival counts of native and heat-shocked cells exceeded our detection limit. Contrary to behaviour at pH 7·0, heat-shocked mutant cells did not show the same carvacrol resistance as heat-shocked wild-type cells (P < 0·05).

Comparison of Tables 1 and 2 with Table 3 in the range shows that higher levels of cross-protection were detected at longer treatment times.

Table 3. Influence of a heat shock prior to carvacrol treatments. Log10 of survival fraction of Escherichia coli BJ4 and ΔrpoS and Listeria monocytogenes EGD-e and ΔsigB cells after carvacrol treatments applied in McIlvaine buffers at pH 7·0 or 4·0 at a concentration of 200 μl l−1. Cells were heat shocked at 45°C for 2 h (E. coli strains) or 1 h (L. monocytogenes strains). Survivors were recovered in nonselective medium
TimeE. coli BJ4E. coli BJ4 ΔrpoSL. monocytogenes EGD-eL. monocytogenes EGD-e ΔsigB
pH 4·0
1 h−1·79 ± 0·14−1·08 ± 0·16−2·19 ± 0·23 <−5·00
2 h−1·85 ± 0·21−1·97 ± 0·22<−5·00 
3 h−2·15 ± 0·15<−5·00  
24 h−3·81 ± 0·44   
pH 7·0
  1. Data represent means ± standard deviations.

1 h−0·13 ± 0·09−0·08 ± 0·02−0·56 ± 0·06−0·79 ± 0·08
2 h−0·28 ± 0·24−0·09 ± 0·01−0·69 ± 0·09−0·84 ± 0·06
3 h−0·43 ± 0·46−0·79 ± 0·12−0·97 ± 0·16−1·16 ± 0·19
24 h−1·32 ± 0·20−1·11 ± 0·46−1·89 ± 0·14−2·12 ± 0·11

Biosynthetic requirements for repair of sublethal damage caused by carvacrol

Figure 4 shows the surviving fraction of L. monocytogenes EGD-e cells treated by carvacrol (200 μl l−1; 90 min; pH 7·0) immediately after treatment (time 0) and after a following incubation for 4 h at 20°C. At time 0, 50% of the initial population were dead, and more than 99% of the survivors were sublethally injured. After 240 min of incubation in TSBYE, the damaged population had repaired their membrane damages and formed colonies in the selective medium (Fig. 4). After incubation, no statistically significant differences (P > 0·05) were observed between survival counts evaluated in both nonselective and selective media. Therefore, all sublethally injured cells were repaired.

Figure 4.

Biosynthetic requirements for the repair of sublethal injury caused in Listeria monocytogenes EGD-e by exposure to carvacrol. Cells were exposed to 200 μl l−1 of carvacrol for 3 h at pH 7·0 and then inoculated to subsequent incubation for 4 h at 30°C in TSBYE (●,○) and into this medium containing penicillin G (△), chloramphenicol (▽), cerulenin (♢) and sodium azide (□). Survivors were recovered in nonselective medium (TSAYE) (●) and selective medium added with sodium chloride (TSAYE-SC) (○,□,△,▽,♢). UT, untreated. Data represent means ± standard deviations (error bars).

The addition of inhibitors of specific metabolic processes in the repair of sublethal membrane damage in EGD-e cells showed biosynthetic requirements needed regain resistance to NaCl in the recovery medium. As shown in Fig. 4, the presence of penicillin G in the repair medium did not prevent the repair of sublethal injuries, because the repair kinetics were identical to those in TSBYE without any inhibitor (P > 0·05). Therefore, the repair of these sublethal damages did not require peptidoglycan synthesis.

However, the presence of cerulenin, chloramphenicol or sodium azide inhibited the repair of most sublethally injured cells (Fig. 4; P < 0·05). After 240 min of repair in the presence of these inhibitors, <1% of the damaged population had repaired their damages. The sensitivity of most carvacrol-damaged cells (99% of survivors) to these inhibitors indicates the requirement of lipid and protein synthesis and energy production to regain the ability to repair the detected membrane damage.

We also evaluated the repair of damages caused by carvacrol (200 μl l−1; 90 min; pH 7·0) in the cytoplasmic and outer membranes of E. coli cells. The repair was also completed after 240 min without any inhibitor. As observed for L. monocytogenes, the addition of penicillin G did not prevent repair (Fig. 5). However, the addition of cerulenin, chloramphenicol or sodium azide did not demonstrate the same clear results as for L. monocytogenes. To study E. coli cell damage better, we used attenuated total reflectance infrared microspectroscopy (ATR-IRMS) combined with a multivariate analysis, soft independent modelling of class analogy (SIMCA).

Figure 5.

Biosynthetic requirements for the repair of sublethal injury caused in Escherichia coli BJ4 by exposure to carvacrol. Cells were exposed to 200 μl l−1 of carvacrol for 3 h at pH 7·0 and then inoculated to subsequent incubation for 4 h at 37°C in TSBYE (□) and into this medium containing penicillin G (■), chloramphenicol (image_n/jam12028-gra-0001.png), cerulenin (image_n/jam12028-gra-0002.png) and sodium azide (image_n/jam12028-gra-0004.png). Survivors were recovered in selective medium added with sodium chloride (TSAYE-SC) or selective medium added with bile salts (TSAYE-BS). Data represent means ± standard deviations (error bars).

Attenuated total reflectance infrared microspectroscopy

Our classification models, soft independent modelling of class analogy (SIMCA) obtained from derivatized infrared spectra (1900–800 cm−1), allowed for the classification of E. coli BJ4 samples according to the presence or absence of carvacrol for each pH. Figures 6A and 6B show the wavenumbers that had a predominant effect on the discrimination of carvacrol-treated and untreated cells at pH 4·0 and 7·0, respectively. The discriminating power of living and carvacrol-treated samples at pH 4·0 (Fig. 6A) showed two spectral bands at 1624 and 1395 cm−1, corresponding to changes in the amide I absorption band of β-sheet proteins (Barth 2007; Kong and Yu 2007), and in the symmetric stretching of COO groups in amino acids and/or fatty acids (Legal et al. 1991; Belfer et al. 2005; Parikh and Chorover 2006). At pH 7·0, a comparison of carvacrol-treated and untreated cells showed that the major discriminating bands were 1083 and 1215 cm−1 (Fig. 6B), corresponding to the symmetric and asymmetric stretching of P=O groups in phosphodiester bonds (Belfer et al. 2005; Yu and Irudayaraj 2005; Álvarez-Ordóñez et al. 2011).

Figure 6.

ATR-IRMS analysis of Escherichia coli after a carvacrol treatment. Soft independent modelling class analogy (SIMCA) of discriminating power of intact and carvacrol-treated (200 μl l−1) E. coli BJ4 cells at pH 4·0 (A) or 7·0 (B) of transformed attenuated total reflectance infrared microspectroscopy (ATR-IRMS) spectra (1900–800 cm−1) using a diamond crystal accessory in reflectance mode.

Discussion

In this study, the Gram-negative bacterium E. coli was more resistant to carvacrol than the Gram-positive L. monocytogenes, in agreement with other authors that have attributed this fact to the selectivity for hydrophobic compounds of Gram-negative bacteria outer membrane (Burt 2004). As reported previously, minimum inhibitory concentrations of carvacrol for E. coli O157:H7 and L. monocytogenes EGD-e were determined to be 0·2 and <0·2 mg ml−1, respectively (Ait-Ouazzou et al. 2011). Furthermore, carvacrol's antibacterial effect of carvacrol is enhanced at low pH (Rivas et al. 2010; Ait-Ouazzou et al. 2011). Higher hydrophobicity and lower dissociation of carvacrol at low pH would allow for its better dissolution in the lipid phase of the cell membranes (Juven et al. 1994), which have been proposed as the main target of carvacrol (Helander et al. 1998).

Despite its relevance to food preservation, the biological basis behind microbial inactivation and resistance to carvacrol has been little studied. Expression of the general stress–response alternative sigma factors σS (RpoS) and σB (SigB) is associated with increased resistance to multiple environmental stresses (Hengge-Aronis 2002; Chaturongakul et al. 2008; O'Byrne and Karatzas 2008), including antimicrobial compounds, such as citral (Somolinos et al. 2010b). However, the influence of these alternative sigma factors in bacterial resistance to carvacrol had not been studied before. Our results demonstrate that E. coli ΔrpoS cells were less resistant to carvacrol at both pH 4·0 and 7·0 than wild-type cells, showing a dependence on general stress–response alternative sigma factor. Surprisingly, sigB deletion decreased carvacrol resistance at pH 4·0, but increased it at pH 7·0. Further research is needed to understand the higher L. monocytogenes ΔsigB resistance at pH 7·0 in relation to wild-type cells. As a consequence, identification of conditions under which σS and σB are expressed could help in the design of food preservation treatments envisaging bacterial resistance to carvacrol under different treatment conditions.

The occurrence of sublethal injury in cell envelopes of both E. coli and L. monocytogenes wild-type and mutant cells after carvacrol treatments (Fig. 2) shows the implication of cell membranes in the mechanism of inactivation by carvacrol. Although Ait-Ouazzou et al. (2011) had reported sublethal damage in cell envelopes by carvacrol, this is the first study evaluating sublethal injuries caused by carvacrol in the cell envelopes in rpoS and sigB mutants, showing that deletion of these genes did not prevent bacterial capacity to repair these injuries when they were plated in a nonselective medium. Furthermore, a similar recovery in selective media for both L. monocytogenes strains after a carvacrol treatment at pH 7·0 indicated that cellular intrinsic resistance to carvacrol was independent of sigB presence (Fig. 2B). In general, the degree of sublethal injuries in the outer membrane was higher than in the cytoplasmic membrane, suggesting that damage occurred first on the outer membrane and later on the cytoplasmic membrane, subsequently leading to cell inactivation.

While some authors had studied the cross-resistance caused by carvacrol (Pol et al. 2001; Luz et al. 2012), modifications in fatty acid composition (Dubois-Brissonnet et al. 2011) and heat shock proteins induced by carvacrol (Burt et al. 2007), there was no literature about stresses inducing cross-resistance to carvacrol. Our study showed that heat-shocked cells increased their resistance to carvacrol, agreeing with previous results showing that application of a heat shock at sublethal temperature conditions induces cross-resistance to multiple stresses (Chung et al. 2006). As application of a heat shock prior to carvacrol treatment at pH 7·0 also protected mutant cells, expression of genes responsible for cell protection to carvacrol induced by heat shock would not require sigB and rpoS presence. Nevertheless, heat shock did not protect rpoS mutants to carvacrol at pH 4·0, probably because the proteins regulated by heat shock could not exert their protective effect to carvacrol at low pH. From a practical point of view, sublethal stresses before carvacrol treatment should be avoided to obtain a high efficacy of carvacrol.

As cell envelopes have been demonstrated to play a key role in the mechanisms of inactivation by carvacrol, we decided to study further the nature of sublethal damages to understand the affected structures by this compound. Repair of damaged cytoplasmic membrane in L. monocytogenes took 4 h, requiring newly synthesized proteins and lipids and the production of energy. Although energy production is a normal requirement for damage repair (Ray and Speck 1972; Chilton et al. 2001; García et al. 2006; Somolinos et al. 2010b; Ait-Ouazzou et al. 2012), the need for lipids points to the membrane as a vital structure in the mechanism of inactivation by carvacrol. The requirement for proteins could point either at membrane proteins affected by carvacrol or at proteins involved in the mechanism of damage repair. Transcriptome and/or proteome analysis of carvacrol-treated cells would assist in elucidating this requirement.

While repair of sublethal injuries in E. coli cell envelopes after 4 h was associated with energy production, our experiment did not show clear results with regard to other biosynthetic requirements involved in the repair process (Fig. 5). Probably, different biosynthetic requirements were necessary for the simultaneous repair of damage in both envelopes. Owing to the limitations of our plating technique in this case, to further study the damages caused in E. coli envelopes by carvacrol, we included an analysis by ATR-IRMS. This technique evaluates the biochemical composition of the bacterial cell constituents (Álvarez-Ordóñez et al. 2011), such as water, proteins, nucleic acids, fatty acids and polysaccharides, making it a suitable tool to analyse changes caused by carvacrol. In effect, our ATR-IRMS results allowed selecting two major discriminating bands at both pH 4·0 and 7·0 as mainly responsible for the differences between untreated and carvacrol-treated cells.

On the one hand, the observed 1624 cm−1 band at pH 4·0 corresponds to the amide I absorption band of β-sheet proteins (Barth 2007; Kong and Yu 2007); the band at 1395 cm−1 reflects the symmetric stretching of COO groups in amino acids and/or fatty acids (Legal et al. 1991; Belfer et al. 2005; Parikh and Chorover 2006). As β-barrel membrane proteins occur in the outer membranes of Gram-negative bacteria (Tamm et al. 2004), and repair or cytoplasmic membrane damage required protein synthesis, the main contribution to the discrimination between untreated and carvacrol-treated cells at pH 4·0 could come from affected outer membrane proteins that form membrane-spanning β-barrels. On the other hand, the discriminating bands found after carvacrol treatments at pH 7·0 were 1083 and 1215 corresponding to the symmetric (1083 cm−1) and asymmetric (1215 cm−1) stretching of P=O groups in phosphodiester bonds (Belfer et al. 2005; Yu and Irudayaraj 2005; Álvarez-Ordóñez et al. 2011). As the phosphodiester bonds are found in phospholipids located in the cytoplasmic membrane and in the inner leaflet of the outer membrane (Silhavy et al. 2010), carvacrol probably targets these structures at pH 7·0. As shown for other bacterial stress responses (Álvarez-Ordóñez and Prieto 2010), ATR-IRMS allowed for description of changes caused by the action of carvacrol. These results show a different mechanism of E. coli inactivation by carvacrol as a function of the treatment medium pH: while carvacrol at pH 4·0 would target outer membrane proteins, at pH 7·0, the phospholipids are the main affected structures.

This study has shown that the mechanisms of inactivation by carvacrol were influenced by different factors, such as pH of the treatment medium or microbial species, probably due to the presence of outer membrane. Bacterial inactivation increased at low pH. This effect had been attributed to different carvacrol properties, but the ATR-IRMS results indicate that structures affected by carvacrol were different depending on the treatment medium pH, targeting phospholipids at pH 7·0 and outer membrane proteins at pH 4·0. Therefore, not only hydrophobicity and dissociated fraction of carvacrol at low pH, but also changes in structure or physicochemical properties of cell envelopes because of pH could be responsible for the lower cell resistance at low pH. Genetic and environmental factors also influenced cell resistance to carvacrol. On the one hand, rpoS and sigB deletion decreased carvacrol resistance in E. coli and L. monocytogenes, respectively. On the other hand, a heat shock at sublethal conditions prior to carvacrol treatments induced a phenomenon of cross-resistance, decreasing cell susceptibility to this antimicrobial compound.

A proper understanding of the mechanisms of inactivation is of great relevance to more appropriate carvacrol treatments with high antimicrobial effects. For example, overexpression of σS and σB and sublethal stresses before carvacrol treatments should be prevented in the design of food preservation treatments. Furthermore, knowledge of carvacrol targets would be useful to design combined process with other food preservation technologies affecting damaged structures to obtain a synergistic effect.

Acknowledgements

This study was financially supported by the CICYT (Project AGL2009-11660), Gobierno de Aragón, Fondo Social Europeo and Departament Quimica of Universitat Rovira i Virgili and Spanish AECID, which provided A. Ait-Ouazzou with a grant to carry out this investigation. Our thanks to Paper-Check.Com, LLC. for their collaboration in English revision of this work.

Ancillary