Salmonella enterica Serovar Typhimurium and Listeria monocytogenes Acid Tolerance Response Induced by Organic Acids at 20{degrees}C: Optimization and Modeling

An acid tolerance response (ATR) has been demonstrated in Listeriamonocytogenes and Salmonella enterica serovar Typhimurium inresponse to low pH poised (i.e., adapted) with acetic or lacticacids at 20°C and modeled by using dynamic differentialequations. The ATR was not immediate or prolonged, and optimizationoccurred after exposure of L. monocytogenes for 3 h at pH 5.5poised with acetic acid and for 2 h at pH 5.5 poised with lacticacid and after exposure of S. enterica serovar Typhimurium for2 h at pH 5.5 poised with acetic acid and for 3 h at pH 5.5poised with lactic acid. An objective mechanistic analysis ofthe acid inactivation data yielded estimates of the durationof the shoulder (t_s), the log-linear decline (k_max), and themagnitude of a critical component (C). The magnitude of k_maxgave the best agreement with estimates of conditions for optimumATR induction made from the raw data.

INTRODUCTION

Top

Introduction

Many food-borne pathogenic bacteria exhibit stress responses,which enhance their survival in adverse environmental conditions.One stress commonly encountered in foods is an acidic environment,where enhanced survival can involve induction of an acid toleranceresponse (ATR). The ATR is defined as the resistance of cellsto low pH when they have been grown at moderately low pH orwhen they have been exposed to a low pH for some time (5). Listeriamonocytogenes and Salmonella enterica serovar Typhimurium exhibitan ATR when exposed to mildly acidic pH (7, 14). The ATR isgrowth phase specific (4, 15), with distinct responses occurringin both logarithmic and stationary phases, and it requires thede novo synthesis of acid-shock proteins (6, 8, 18). The ATRconfers cross-resistance to other stresses such as heat, sodiumchloride, and ethanol (16, 17) and there is some evidence thatit may increase bacterial virulence (9, 18). The ATR, therefore,impacts on both predictive modeling and risk assessment approachesto microbiological food safety.

Most studies of the ATR have been conducted at temperaturesof 30 or 37°C (6, 14, 18) and have typically used mineralacids, such as hydrochloric acid, as the acidulant. Food-bornepathogenic bacteria are more commonly exposed to temperaturesof 20°C or below and to weak organic acids, such as lacticor acetic acids, either as by-products of bacterial metabolismin fermentation processes or as deliberate additions to foodsas preservatives. As a result, the ATR expressed in responseto these conditions is of most relevance to food. In addition,previous studies have typically measured survival of cells atan arbitrary time as an indicator of the ATR. This ignores certainfeatures of inactivation kinetics, such as the shoulder durationand rate of inactivation, that may be indicators of the mechanismsinvolved in the ATR.

The study reported here determined optimum conditions for theinduction of an ATR in S. enterica serovar Typhimurium and L.monocytogenes at 20°C by using acetic or lactic acids asacidulants. The experimental temperature of 20°C was chosento reflect what could possibly occur in products that are temperatureabused. Lower temperatures can offer some preservative effect;therefore, studying the response at 20°C (i.e., room temperature)allowed the worse case scenario to be studied. Acetic and lacticacids were the chosen acidulants since these acids are commonlyencountered by microorganisms within food products. We thenapplied a dynamic mathematical model to the data to obtain anestimate of parameters associated with the acid inactivationkinetics. This provided an objective approach to the quantificationof the ATR, and we discuss here the use of such a model as ameans to identify key features associated with the response.

MATERIALS AND METHODS

Top

Materials and Methods

Bacteria.
L. monocytogenes strain EGD-e, as sequenced by Glaser et al.(11), was obtained from J. Vazquez-Boland, Departmento PatolgiaAnimal I, Universidad Complutense, Madrid. S. enterica serovarTyphimurium strain SL1344 was provided by Catherine Lee, Departmentof Microbiology and Molecular Genetics, Harvard Medical School.

Culture media.
Stock cultures of L. monocytogenes were maintained on nutrientagar (Oxoid) slopes stored at 0°C. Slopes were subculturedmonthly by culture to fresh nutrient agar slopes which wereincubated at 25°C for 1 day prior to storage at 0°C.Stock cultures of S. enterica serovar Typhimurium were storedat -80°C in tryptone soya broth (Oxoid) plus glycerol (30%[wt/vol]). This culture was used to inoculate nutrient agarslopes monthly, which were incubated at 25°C for 1 day andsubsequently stored at 0°C. Liquid growth medium consistedof Trypticase soy broth (Baltimore Biological Laboratory) plus1% (wt/vol) glucose (TSBG) adjusted to pH 7.0 with HCl and sterilizedby filtration through a cartridge filter unit containing a 0.2-µm-pore-sizemembrane filter (Kleenpak, Pall, United Kingdom).

Preparation of inocula.
Bacteria were grown successively at 25°C for 24 h and thenat 20°C for 24 h. The resultant population of ca. 10⁹ viablecells ml^-1 was diluted in peptone salt dilution fluid (13) toproduce a suspension that contained the required number of viablebacteria.

Acid tolerance.
Sterile TSBG (pH 7.0) (100 ml) was transferred to sterile, 250-mlErlenmeyer flasks, which were closed with a cover that allowedgaseous exchange. Approximately 10⁴ viable cells ml^-1 were addedto two parallel flasks, and these were incubated at 20°Cfor 15 h ±1 h on an orbital platform shaker (IKA-Labortechnik,Esslab, United Kingdom) rotating at 120 rpm to maximize aeration.This resulted in cultures that were in the mid-logarithmic phaseof growth (data not shown).

The pH of one culture was then adjusted, by the addition ofeither 1 M lactic acid or 1 M acetic acid, to a range of adaptationpH values and reincubated as described above for between 1 and6 h. The final concentration of acids required to adjust culturesof L. monocytogenes to pH 5.0 and 5.5 were 0.03 M acetic orlactic acid and 0.02 M acetic or lactic acid, respectively.The final concentrations of acetic acid and lactic acid requiredto adjust cultures of S. enterica serovar Typhimurium to pH5.0, 5.5, and 5.8 were 0.02 M acetic or lactic acid, 0.018 Macetic or lactic acid, and 0.015 M acetic or lactic acid, respectively.These adaptation pHs were chosen since they included pH valuesused for these organisms in previous studies (pH 5.5 and 5.8)(4, 7, 14, 18), as well as a value (pH 5.0) representative ofthe mildly acidic environment organisms may encounter withincertain foods. After the required incubation time, cells wereremoved from the medium by filtration through a 0.22-µm-pore-sizemembrane filter (Millipore). These cells were resuspended in100 ml of fresh TSBG (pH 7.0) by vortexing to remove the cellsfrom the membrane filter. This flask was adjusted to pH 3.0± 0.1 by using 1 M HCl and then reincubated as describedabove for up to 12 h. At intervals throughout incubation a samplewas removed, and the number of viable bacteria was determinedby spreading 50 µl of this suspension onto the surfaceof duplicate plates of plate count agar (Oxoid CM325) by usinga spiral plate maker (Spiral Systems, Inc.). Plate count agarplates were incubated at 35°C for 24 h before enumeration.

Cells were also removed by filtration from the parallel controlculture (pH 7.0) incubated as described above. These cells wereresuspended in 100 ml of fresh TSBG (pH 7.0), the pH was adjustedto pH 3.0 ± 0.1, and the numbers of viable cells weredetermined as described above. All experiments were performedat least in duplicate.

Modeling inactivation kinetics.
The numbers of viable bacteria in the culture medium adjustedto pH 3.0 ± 0.1 were modeled to provide a quantitativeand objective comparison of inactivation kinetics. The modelof Geeraerd et al. (10) was originally developed for modelingmicrobial inactivation during a mild heat treatment but haskey attributes important in the case of acid inactivation. Theseattributes include the ability to model a shoulder on the inactivationcurve, a log-linear decline in viable numbers and a tail. Itis cast as dynamic differential equations to allow time-varyingchallenges.

The shoulder of the death curve was modeled by introducing theconcept of a critical component (C, which may or may not bea real substance within or outside the cell) that decays immediatelyupon acid challenge. The amount of this component at the startof the challenge is a measure of the "physiological condition"of the cells and (together with the actual death rate) determinesthe length of the shoulder. Component C decays as follows:

(1)
The death rate was moderated by a factor, ${alpha}$ , castas:

(2)
where C_c is a dimensionlessconcentration, normalized by a Michaelis constant. In otherwords, C_c = C/K_c (i.e., C_c, behaves exactly as C in the firstequation).

The decay of the population of viable cells is assumed to followthe form:

(3)
where the instantaneousrate constant, k, is given by:

(4)
andwhere N_res represents a residual subpopulation of cells (i.e.,the tail of the population).

The static solution for the model was as follows:

(5)
where the shoulder time is given by:

(6)
The start of the tail is given by:

(7)
The tail portion models the residual proportionof the population (N_res) that is naturally resistant due toheterogeneity within a bacterial population (reviewed by Booth[2]). The present study was primarily concerned with acid inactivationof a population of acid-adapted cells in comparison with anunadapted population, and the small residual population wasnot under investigation. Additionally, because the residualpopulation was small, the error associated with its enumerationby the viable count procedure was great. This resulted in poorfitting of the model and thus, for the purposes of the presentstudy, a reduced model was developed that ignored the residualpopulation (N_res).

The static solution of this reduced model is cast as:

(8)
The model was fitted to the data by using thestatistical software package "R," and a nonlinear least-squaresmethod was used. The goodness of fit was determined both byvisual inspection and from the value of the residual sum ofsquares (not shown). The standard errors for the fitted parametersare shown in Table 1.

View this table:
[in this window]
[in a new window]
TABLE 1. Effects of pretreatment of L. monocytogenes and S. enterica serovar Typhimurium on the parameters of k_max and C_c (0)

RESULTS

Top

Results

Acid tolerance.
All data presented are means of at least duplicate experiments.An integral process in the measurement of the acid toleranceis the removal of cells from one culture onto a filter membraneand resuspension of these cells in fresh medium adjusted topH 3.0. Measurements showed that >90% of cells were recoveredfrom the filter and that cells exposed to either no organicacid, acetic acid, or lactic acid were recovered equally efficiently(see similar initial values in Fig. 1 and 2). The mid-logarithmic-phasecells of L. monocytogenes and S. enterica serovar Typhimuriumgrown at pH 7.0 lost viability rapidly when exposed to pH 3.0(i.e., large values of k_max in Table 1; representative dataare shown in Fig. 1 and 2). However, when these cells were firstincubated for between 1 and 6 h at pH 5.0, 5.5, or 5.8, adjustedby using either acetic or lactic acid, they demonstrated anenhanced tolerance to the acidic conditions and survived formuch longer (i.e., small values of k_max in Table 1; representativedata shown in Fig. 1 and 2). The cells had developed an ATRduring incubation at a mildly acidic pH.

View larger version (22K):
[in this window]
[in a new window]
FIG. 1. Survival of L. monocytogenes either unadapted ( ${circ}$ ) or after prior adaptation at pH 5.5 poised with lactic acid (A) or acetic acid (B) for 1 h (•), 2 h ( ${blacksquare}$ ), 3 h ( ${blacktriangleup}$ ), 5 h ( ${blacktriangledown}$ ), or 6 h ( ${diamondsuit}$ ). Datum points are the means of at least duplicate experiments, and the bars show the range.

View larger version (19K):
[in this window]
[in a new window]
FIG. 2. Survival of S. enterica serovar Typhimurium either unadapted ( ${circ}$ ) or after prior adaptation at pH 5.5 poised with lactic acid (A) and acetic acid (B) for 1 h (•), 2 h ( ${blacksquare}$ ), 3 h ( ${blacktriangleup}$ ), 4 h ( ${blacktriangledown}$ ), or 5 h ( ${diamondsuit}$ ). Datum points are means of at least duplicate experiments, and the bars show the range.

Modeling inactivation kinetics.
Comparison of acid inactivation data (such as those given inFig. 1 and 2) can include simple comparisons of the numbersof cells surviving at an arbitrary time point. However, fittingthe model of Geeraerd et al. (10) to the data enabled objectiveestimation of the parameters of the inactivation curve (forexample, see Fig. 3). We used the reduced form of the model(equation 8) to estimate the parameters of the critical componentC_c(0) and log-linear decline (k_max). From these parameters canbe calculated an additional parameter, the magnitude of shouldertime (t_s). The goodness of fit was determined both by visualinspection and by the value of the residual sum of squares (datanot shown). Data shown were from model fits of acceptable qualityboth visually and statistically. The parameters of k_max andC_c(0) are shown in Table 1, and the duration of the shouldertime (t_s) is shown in Fig. 4.

View larger version (26K):
[in this window]
[in a new window]
FIG. 3. Survival of unadapted cells of L. monocytogenes EGD-e. Points are experimental data. The line is the model fitted to these data. Parameters estimated by the model are also indicated. These include the initial cell concentration N₀ (in CFU per milliliter), the length of the shoulder t_s (in minutes), the time to the start of the tail t_res (in minutes), and the slope of the log linear region of the curve (the rate constant k_max [per hour]).

View larger version (21K):
[in this window]
[in a new window]
FIG. 4. (A) Effect of adaptation time on the duration of the shoulder (t_s) shown by L. monocytogenes EGD-e during exposure to pH 3.0 ± 0.1. Adaptation conditions were acetic acid at pH 5.5 (•), acetic acid at pH 5.0 ( ${blacktriangleup}$ ), lactic acid at pH 5.5 ( ${blacksquare}$ ), or lactic acid at pH 5.0. ( ${diamond}$ ). (B) Effect of adaptation time on the duration of the shoulder (t_s) shown by S. enterica serovar Typhimurium during exposure to pH 3.0 ± 0.1. Adaptation conditions were acetic acid at pH 5.5 (•), lactic acid at pH 5.8 ( ${blacktriangleup}$ ), lactic acid at pH 5.5 ( ${blacksquare}$ ), acetic acid at pH 5.8 ( ${blacktriangledown}$ ), or lactic acid at pH 5.0. ( ${diamondsuit}$ ).

As expected, the values of k_max were large for unadapted populationsof either L. monocytogenes or S. enterica serovar Typhimurium,indicating that the numbers of viable cells were reduced rapidlyupon exposure to pH 3.0.

In the case of L. monocytogenes, k_max was decreased in populationsexposed to either acetic or lactic acids at any pH during adaptationfor 1 h. Subsequent exposure to pH 5.0 poised with (i.e., adaptedwith) either acid resulted in an increase in k_max. At pH 5.5poised with acetic acid the k_max decreased to a minimum valueafter exposure for 3 h, and at pH 5.5 poised with lactic acidthe k_max decreased to a minimum value after exposure for 2 h.

In the case of S. enterica serovar Typhimurium at pH 5.5 poisedwith acetic acid the k_max decreased to a minimum value afterexposure for 2 h, and at pH 5.5 poised with lactic acid thek_max decreased to a minimum value after exposure for 3 h. Thek_max did not show such marked minima during exposure to pH 5.8poised with either acid.

The values of C_c(0) differed with adaptation time during exposureof both organisms to all conditions investigated. Large standarderrors were also associated with these parameter estimates.No clear pattern emerged of the value of C_c(0) with respectto adaptation conditions.

In the case of L. monocytogenes, only adaptation at pH 5.5 poisedwith acetic acid for 3 h resulted in a value of t_s markedlygreater than that of the unadapted population. In contrast,the values of t_s shown by S. enterica serovar Typhimurium weregreater than unadapted cells when cells were exposed at pH 5.5poised with acetic acid for 1 h and for 4 h, at pH 5.5 poisedwith lactic acid for 1 to 3 h, at pH 5.8 poised with lacticacid for 1 h, and at pH 5.8 poised with acetic acid for 3 h.

The parameters in Table 1 and in Fig. 4 are indicative of theeffects of exposure to pH 3.0 ± 0.1 on the inactivationkinetics of the bacteria. Use of the parameters offered an objectivemeans to quantify the ATR. Scrutiny of the inactivation datapresented in Fig. 1 and 2 indicated that the survival of thepopulation varied with adaptation time. Of the objective parametersextracted by the model, the value of k_max best represented theeffect of time on adaptation.

Although minima were apparent in Table 1, graphical representationof this parameter was inevitably biased by the large valuesfor the unadapted population. We have normalized the value ofk_max by reference to unadapted cells. Accordingly, Fig. 5 showsthe ratio of the magnitude of the log-linear decline (k_max)of unadapted to adapted cells. This demonstrates that optimizationof the ATR occurred after exposure of L. monocytogenes for 3h at pH 5.5 poised with acetic acid and 2 h at pH 5.5 poisedwith lactic acid. Optimization of ATR in S. enterica serovarTyphimurium occurred after exposure for 2 h at pH 5.5 poisedwith acetic acid and after exposure for 3 h at pH 5.5 poisedwith lactic acid.

View larger version (22K):
[in this window]
[in a new window]
FIG. 5. (A) Ratio of k_max (unadapted) to k_max (adapted) calculated from the fitted model parameters for L. monocytogenes EGD-e grown at pH 7.0 and adapted with acetic acid at pH 5.5 (•), acetic acid at pH 5.0 ( ${blacktriangleup}$ ), lactic acid at pH 5.5 ( ${blacksquare}$ ), or lactic acid at pH 5.0 ( ${diamondsuit}$ ). (B) Ratio of k_max (unadapted) to k_max (adapted) calculated from the fitted model parameters for S. enterica serovar Typhimurium grown at pH 7.0 and adapted with acetic acid at pH 5.5 (•), acetic acid at pH 5.8 ( ${diamondsuit}$ ), lactic acid at pH 5.8 ( ${blacktriangleup}$ ), lactic acid at pH 5.5 ( ${blacksquare}$ ), or with lactic acid at pH 5.0 ( ${blacktriangleup}$ ).

DISCUSSION

Top

Discussion

An ATR has previously been determined in S. enterica serovarTyphimurium and L. monocytogenes at 37°C. However, whetherthis response is also present at lower temperatures is of greaterimportance to the food industry. The present study demonstratesthat exposure of aerobically grown L. monocytogenes and S. entericaserovar Typhimurium to mild acid treatment at 20°C by usingeither acetic or lactic acid results in an ATR. The ATR observedwas not immediate or prolonged but was optimized under a distinctset of conditions, the most important of which appeared to bethe adaptation time period and pH. This time phenomenon wasalso noted in previous studies with L. monocytogenes at 37°C(20). Characterization of the ATR under these conditions providesan opportunity to compare the ATR with that produced in coloniesunder immobilized conditions, which are more representativeof some food environments (12).

The adaptation time may be linked to maintenance of the internalpH (pH_i) of the cell. Previous studies have shown that aciddeath is a direct result of a lowered pH_i (7, 19). Acid-adaptedcells were better able to maintain their pH_i and, as a result,survived better in acidic environments. Phan-Thanh et al. (20)also noted that organic acids were more lethal at low pH thaninorganic acids. This appeared to be due to their ability toalter the pH_i of L. monocytogenes to a lower level than thatobserved with inorganic acids. Further evidence for the involvementof pH_i maintenance in acid tolerance was provided in the studiesof Foster and Hall (7) and Cotter et al. (3).

It is possible that the adaptation time is directly linked tothe ability of the cell to maintain its pH_i. The cell may beable to maintain pH_i and adapt at mildly acidic conditions butnot for a prolonged period. After a critical time, the pH_i maintenancesystem fails and cells become sensitive to the lethal effectsof the acid (i.e., the ATR is no longer observed when cellsare challenged at pH 3.0). This critical time may be the optimumadaptation time observed in the present study.

The present study has also found that adaptation times differedaccording to the acidulant used to adapt cells. For example,at pH 5.5 S. enterica serovar Typhimurium required an adaptationtime of only 2 h for the optimum ATR to be induced by usingacetic acid as the acidulant but 3 h when lactic acid was theacidulant. A difference was also observed in the case of L.monocytogenes. However, both organic acids undergo dissociationand, at the relevant pH values used, the protonated (undissociated)concentration was small. For example, in the case of culturesof S. enterica serovar Typhimurium adjusted to pH 5.5, the totalconcentration of 0.018 M acetic (pK 4.76) or lactic acid (pK3.86) required to poise the pH was equivalent to a concentrationof undissociated acetic or lactic acid of 3 or 0.4 mM, respectively.Differences in adaptation times between acidulants may be explainedby the differing ability of organic acids to alter the pH_i ofthe cell. Certain organic acids may enter into the cell moreeasily than others and therefore alter the pH_i of the cell morereadily. Acetic acid is lipid soluble, diffuses rapidly throughthe plasma membrane and has been shown to have a dramatic effecton the pH_i of a cell (21). In contrast, lactic acid demonstratessome lipid solubility and diffuses slowly through the membrane.Disruption of the cell pH_i does not appear to be its main modeof inhibition (21).

If the pH_i of the cell is reduced more readily with acetic acidthan with lactic acid, then the optimum adaptation time foracetic acid may also be reduced, and the optimum adaptationtime may differ according to the organic acid used to poisethe system. Differences observed between organisms may resultfrom differences in membrane structure.

Previous studies have shown that inorganic acid adaptation providesprotection against organic acid stress and vice versa (1, 19).We have demonstrated in the present study that organic acidadaptation provides protection against inorganic acid stressat 20°C (i.e., cross-protection also occurs at this lowertemperature).

Application of a reduced model derived from the model of Geeraerdet al. (10) enabled an objective mechanistic analysis of theinactivation data. The parameters of the duration of the shoulder(t_s), and the magnitude of the critical component [C_c(0)] provedin the present study to be inadequate descriptors of the optimizedATR. However, the magnitude of the log-linear decline (k_max)was a good indicator of adaptation. Such a mechanistic modelingapproach is worthwhile in that it enables objective study ofthe effect of changing conditions (adaptation time, acidulant)on the features of the ATR.

The present study has identified conditions to produce aerobicallygrown mid-logarithmic-phase cells of L. monocytogenes and S.enterica serovar Typhimurium that are acid tolerant. The ATRhas been shown to be transient and the conditions identifiedhere enable cells to exhibit maximal tolerance to acid shock(which we describe as optimization of the ATR) prior to thedecay of the ATR upon continued exposure to organic acids.

These cells are the focus of further studies to identify themagnitude and mechanisms of cross-protection of L. monocytogenesand S. enterica serovar Typhimurium to food-relevant inimicalconditions.

ACKNOWLEDGMENTS

We are grateful for financial support from the UK Biotechnologyand Biological Sciences Research Council.

FOOTNOTES

* Corresponding author. Mailing address: Institute of Food Research, Norwich Research Park, Colney, Norwich NR4 7UA, United Kingdom. Phone: 44(0)1603-255216. Fax: 44(0)1603-507723. E-mail: elizabeth.greenacre{at}bbsrc.ac.uk.

REFERENCES

Top