- Email: [email protected]
Effect of nisin on heat injury and inactivation of Salmonella enteritidis PT4
International Journal of Food Microbiology 43 (1998) 7–13
Effect of nisin on heat injury and inactivation of Salmonella enteritidis PT4 1
I.S. Boziaris, L. Humpheson , M.R. Adams* School of Biological Sciences, University of Surrey, Guildford, Surrey, GU2 5 XH, UK Received 1 May 1998; accepted 27 May 1998
Abstract The ability of heat injury to confer sensitivity to nisin in a Gram negative pathogen was investigated. Injury and inactivation kinetics of Salmonella enteritidis PT4 in the presence of nisin were determined in media, liquid whole egg and egg white using cultural methods and capacitance monitoring to detect injury. Addition of nisin in concentrations from 500 IU / ml to 2500 IU / ml in the heating menstruum caused a reduction of required pasteurisation time of up to 35%, principally as a result of its effect on cells suffering damage during heating. In egg white and liquid whole egg the organism’s heat susceptibility was greater than in nutrient broth, particularly in egg white which contained no fat and had an alkaline pH. The effect of nisin on heat susceptibility was however less pronounced than in nutrient broth due to its interaction with protein and fat. Though nisin did not enhance the lethality of heat processes, injury is more severe in egg white containing nisin, presumably as a result of its interaction with antimicrobial factors in egg white. 1998 Elsevier Science B.V. All rights reserved. Keywords: Nisin; Pasteurisation; Thermal injury; Salmonella
1. Introduction Nisin is a bacteriocin produced by Lactococcus lactis subsp. lactis which exhibits a broad spectrum of inhibitory activity against Gram-positive organisms including bacterial spores. It is not generally active against Gram-negative bacteria, yeasts and fungi, although Gram negatives do show nisin-sen*Corresponding author. Tel.: 1 44 1483 300 800; fax: 1 44 1483 300 374; E-mail [email protected] 1 Present address: SAC Auchincruive, Ayrshire, Scotland KA6 5HW, UK.
sitivity when their outer membrane permeability is affected by treatment with chelators such as EDTA, citrate and phosphates (Stevens et al., 1991, 1992; Cutter and Siragusa, 1995; Vaara, 1992). Physical treatments can also affect outer membrane permeability. Tsuchido et al. (1985) have reported its destruction in E. coli by heating and Kalchayanand et al. (1992) found that sublethally injured Gramnegatives produced by freezing and heating are sensitive to nisin. It is therefore possible that sublethal injury to Gram negatives during pasteurisation processes could induce sensitivity to nisin. Inclusion of nisin could contribute to a reduction of heating
0168-1605 / 98 / $19.00 1998 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 98 )00083-X
8
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
time, helping to reduce costs, conserve heat sensitive characteristics of a product and help meet consumer demands for safe but less processed foods. Egg products are an obvious potential application; a recent survey in the United States reported 48% of unpasteurized liquid egg samples to be contaminated with salmonella (Hogue et al., 1997). In this paper we report a study on thermal injury and the heat sensitivity of Salmonella enteritidis in nutrient media, liquid whole egg and egg white in the presence of nisin.
2. Materials and methods
2.1. Organism Salmonella enteritidis phage type 4 P167807, supplied by Division of Enteric Pathogens, Central Public Health Laboratory, London UK, was stored frozen in bead vials (Protect; Technical Service Consultants Ltd, Heywood, Lancashire, UK) at 2 708C and resuscitated to 10 9 cfu / ml in 10 ml Nutrient Broth (NB; Unipath, Basingstoke, Hampshire, UK) at 378C for 24 h. All the microbiological media used were supplied by Oxoid (Basingstoke, Hampshire, UK), unless otherwise stated.
2.2. Nisin Nisin was supplied in a purified form (5 3 10 7 IU / g) by Aplin and Barrett, Beaminster, Dorset, UK. Solutions of nisin were prepared in 0.02 N HCl (pH 2) and were stored at 48C. The activity of nisin was determined by horizontal agar plate diffusion according to the Fowler et al. (1975) bioassay. The assay was carried out for nisin in nutrient broth (NB), egg white (EW) and liquid whole egg (LWE). In NB, samples were taken at the beginning and the end of the heat treatment, diluted as appropriate in 0.02 N HCl and 0.1 ml of each sample transferred into the wells of the assay medium. In the case of EW and LWE, 1 ml samples were taken at the beginning and the end of the thermal treatment acidified with concentrated HCl to pH 2, boiled for 5 min, cooled rapidly, adjusted to 10 ml with HCl 0.02 N and centrifuged at 1000 3 g for 20 min. The aqueous supernatant was filtered in GF /A Whatman glass fibre paper (Whatman, Maid-
stone, Kent, UK), diluted as appropriate with 0.02 N HCl and 0.1 ml of each sample used in the bioassay. At least three replicates for both standards and samples were carried out.
2.3. Heat challenge Resuscitated cultures were diluted 10-fold in maximum recovery diluent (MRD) for the inoculation of pre-warmed NB at 378C (50 or 100 ml) to give an initial suspension of | 1 to 10 cfu / ml. All broths were incubated statically at 378C for 22 h61 h and immediately centrifuged (1000 3 g for 15 min at 208C). Previous determination of a growth curve under these conditions had shown the culture to be well into the stationary phase at this stage.
2.3.1. Preparation of eggs prior to heat challenge Locally purchased eggs were washed in a detergent solution and immersed for a few seconds in industrial methylated spirit to kill the microorganisms on the shell and allowed to air-dry. They were then broken and the contents (egg white or liquid whole egg) collected in a sterile beaker. Then 25 ml samples were transferred into 50-ml plastic tubes (Bibby Sterilin, Staffordshire, UK). The pH of both EW and LWE was measured with a BDH Gelplus, double junction, flat tip, electrode (BDH, Merck, Poole, Dorset, UK). 2.3.2. Heat challenge in NB Centrifuged cell pellets were resuspended in 1 ml prewarmed at 378C NB. Nisin solution or 0.02 N HCl (0.2 ml) was added to NB already heated at the temperature of 558C in order to give the required final nisin concentration. The heating menstruum was held at the temperature of 558C in a plugged 100 ml conical flask in a thermostated water bath 5560.058C and inoculated with 1 ml of the prepared suspension. The flask contents were stirred via a magnetic follower, propelled by a custom-made 12 V d.c. submersible stirrer operating at 60 rev. / min to minimise vortex formation. Temperature regulation was provided by a Haake DC-1 circulator heater (Fisons, Loughborough, Leicestershire, UK). Heating menstruum temperature was measured using a NAMAS certified probe and digital indicator (Pt 100
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
probe and Series 268 indicator; Anville Instruments, Camberley, Surrey, UK).
2.3.3. Heat challenge in egg white and liquid whole egg With EW and LWE, nisin solution or HCl 0.02 N (0.2 ml) was added to 25 ml of EW or LWE in order to give the required final concentration and vortexed for 10 s. One ml of the prepared cell suspension was added and the mixture was vortexed again for 10 s. One ml of the inoculated mixture was placed in each of the 75 3 12 mm test tubes closed with hydrophobic cotton. They were subsequently placed into a water bath at 558C. A temperature probe (Pt 100 probe and Series 268 indicator; Anville Instruments) was inserted into the geometric centre of the liquid in a control tube containing 1 ml of uninoculated EW or LWE. 2.4. Enumeration of the cells With NB, a 1 ml sample was taken every 5 min, diluted as appropriate in maximum recovery diluent (MRD; 0.85% NaCl, 0.1% bacteriological peptone) and enumerated as 0.1 ml spread plates on Nutrient and XLD agar. Plates were incubated for 48 h at 378C. With EW and LWE (every 1 and 3 min, respectively) the contents of one of the test tubes was cooled immediately to below 378C by adding 3 ml of MRD at room temperature and the suspension transferred to a sterile universal vial. The test tube was rinsed (3 3 2 ml MRD) and the washings transferred to the sterile universal to give the first dilution. Samples were then diluted as appropriate and plated (0.1 ml) onto NA and XLD medium.
9
Purpose Plus Medium (BGPPM, bioMerieux). The detection time, DT, was recorded after incubation of the Bactometer wells at 378C. Cells were also enumerated by plating 0.1 ml of the same dilution used to inoculate the Bactometer well, on NA and XLD and incubating it at 378C for 48 h. A calibration graph of log(cfu / ml) on NA against DT was first plotted by using non heat-stressed Salmonella enteritidis cells.
2.6. Calculation of D values For NB and LWE, D values were calculated from the linear regression of log survivors against time. With EW, the short duration of the heat process meant that the bacterial population was very low by the time the medium reached the 558C. It was therefore necessary to take the heating-up time into account when calculating the D values. Accordingly D values were derived from the number of pasteurisation units at 558C (P55 ) required to produce a 4 log reduction in survivors. One P55 corresponds to a process equivalent in terms of its lethality to 1 min at 558C. These were calculated from the survivor curves and lethal rate curves derived from time temperature data taken during the heat treatment. With the z value of S. enteritidis equal to 48C (Humpheson et al., 1998), the lethal rate (Lv) at any time t is equal to: Lv 5 10 (T255) / 4 , where T is the temperature at time t. The area under the Lv curve between times t 1 and t 2 which correspond to 7 log cfu / g and 3 log cfu / g survivors respectively, was calculated by using FigP (Version 2.5, FigP Software Corp., Durham, NC, USA). Each experiment was carried out at least in triplicate.
2.5. Capacitance measurements
3. Results
Capacitance measurements were made with the Bactometer Model 120SC (Bactomatic, Inc. USA) with a Model 123 central data processor (bioMerieux). The heat challenge procedure was the same as described previously. Samples were taken every 5 min, diluted in MRD to give a population less than 10 5 cfu / ml and 0.1 ml of the diluted samples (in duplicate) was placed into the Bactometer disposable module wells (bioMerieux Vitek, Inc. 99052) containing 1 ml of Bactometer General
3.1. Injury and inactivation during heating Sublethal injury in survivors during heating at 558C was assessed using two methods: the difference between counts on selective (XLD) and non-selective (NA) media and the extension of capacitance detection time for survivors relative to that predicted from the calibration curve for the same number of uninjured cells. Both methods showed the initial population to be
10
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
uninjured (zero extension of detection time and identical counts on NA and XLD) and the extent of injury to increase during heating as indicated by the increasing difference between the colony counts on XLD and NA and the predicted and observed detection times (Fig. 1). The actual extent of injury can be represented by the extension of detection time and by the difference between counts on non-selective and selective media as a proportion of the total population ((NA 2 XLD) / NA) (Fig. 2). Extension of
Fig. 1. Nutrient agar and XLD counts and predicted and observed capacitance detection times of Salmonella enteritidis at 558C in nutrient broth. d, NA counts; s, XLD counts; j, predicted DT; h, observed DT.
Fig. 2. The proportion of the injured cells and the extension of detection time, observed during heating at 558C in nutrient broth. s, extension of DT; d, (NA 2 XLD) / NA.
detection time (DT observed 2 DT predicted ) and proportion of injury curves both show a similar profile with the extent of injury increasing rapidly so that more than 99% of survivors displayed injury after 10 min heating, reflected in an extended lag period (detection time) of about 5 h (Fig. 2). After 15 min the extent of injury increased at a much slower rate. To determine whether the increasing injury during heating affected the nisin sensitivity of the organisms, thermal death curves were produced using nutrient broth containing different nisin concentrations. The mean D values determined from counts of total survivors (NA) and uninjured survivors (XLD) are presented as Fig. 3. The D value derived from XLD counts measures the combined rate of both injury and inactivation during heating and is therefore much shorter than the D value from NA counts which measure primarily the rate of inactivation. From Fig. 3, it is apparent that as nisin concentration increases, the NA-derived D values decrease by up to 35% while the XLD-derived D value decreased only slightly, but significantly (P , 0.05). This indicates that nisin is killing primarily those cells that have been sublethally injured by the heat treatment whereas uninjured cells are less affected. A saturation effect of nisin against injured cells is also apparent. Concentrations of nisin higher than 1500 IU / ml did not cause any significant additional effect. Analysis of variance showed that D values for S.
Fig. 3. Effect of nisin on the D value of S. enteritidis at 558C in nutrient broth (pH 7.3). j, NA counts; h, XLD counts. The error bars show the standard deviation.
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
11
enteritidis at nisin concentrations of 1500, 2000 and 2500 IU / ml did not differ significantly (P . 0.05) in contrast with the D values obtained at 0, 500, 1000 and 1500 IU / ml.
3.2. Heat injury and inactivation in egg white and liquid whole egg Egg white (EW) and liquid whole egg (LWE) are two heat sensitive products where a reduction in the time / temperature of pasteurisation or an increased lethality with existing protocols is desirable. The different experimental procedure required when determining survival of S. enteritidis in heated EW meant that D values were derived from the number of P55 units required for a reduction from 7 log cfu / ml to 3 log cfu / ml of S. enteritidis (see Section 2). The D 55 value in that case being the P55 divided by four (the log cfu / ml reduction). The D 55 against nisin concentration is plotted in Fig. 4. In LWE where, unlike NB, there was an appreciable heating up time (6 min), D values at 558C were calculated from survivors curves once the temperature had reached 558C. The linear region of the survivor curve extended over a 5 log reduction in survivors and had an r 2 . 0.99. The equivalent D values from XLD counts were also calculated and both are shown in Fig. 5 for different concentrations of nisin.
Fig. 4. Effect of nisin on D value of S. enteritidis in egg white (pH 9.0–9.3). j, NA counts; h, XLD counts. The error bars show the standard deviation.
Fig. 5. Effect of nisin on the D value of S. enteritidis at 558C in liquid whole egg (pH 7.5–7.8). j, NA counts; h, XLD counts. The error bars show the standard deviation.
D values in LWE were considerably longer than in EW presumably a result of the combined effect of the higher pH in EW (9.0 to 9.3 compared with 7.5 to 7.8 in LWE) and the protective effect of fat in the LWE. The reduction of D value and therefore heating time in the presence of nisin was significant in EW but not in LWE. The percentage reduction in D value in EW was not however as great as that achieved in nutrient broth; 500 IU / ml showed no reduction and for 1500 and 2500 IU / ml the reduction of D value in EW was 17.7 and 26.5%, respectively. Equivalent figures for NB were 29.9 and 35%, respectively. This can be ascribed to the lower activity of nisin at the pH of the egg white (9.0–9.3) and possible binding of nisin with egg white proteins which reduced its antibacterial activity. Nisin levels were measured before and after all heat treatments and found to be virtually unchanged suggesting that any binding was reversible. Though nisin was only seen to enhance the lethal effect of heat in EW at higher concentrations, it did produce appreciably greater rates of injury in survivors than were observed in nutrient broth. The D values derived from the XLD counts decreased more in EW with nisin than in NB with nisin. For example, at 2500 IU / ml the percentage of reduction in NB was only 16% in contrast to 26% in EW. This was not seen in LWE, where the XLD-derived D values did not differ significantly with increasing nisin concentration.
12
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
4. Discussion Injured cells can repair and grow on non-selective media but may not be able to grow on selective media due to their increased sensitivity to selective agents (Ray, 1979, 1993). Clark and Ordal (1969) and Ray et al. (1971) reported that Salmonella typhimurium and Salmonella anatum cells injured by heating and freeze drying, respectively, were sensitive to deoxycholate, the selective agent used in xylose lysine deoxycholate (XLD) agar. The NA 2 XLD difference is therefore taken to represent the population that has been subject to injury affecting the cell outer membrane which increases their sensitivity to deoxycholate in XLD (Vaara, 1992). Injury may not however always render cells sensitive to a particular selective agent and capacitance monitoring offers an alternative instrumental technique, which measures the extended lag phase while injured populations repair. It therefore detects a broader range of injury (Mackey and Derrick, 1984; Alexandrou et al., 1995). This work further illustrates the usefulness of impedimetric methods for determining total injury in microbial populations. The similar profile of injury as assessed by impedimetry and differential counts suggests that the outer membrane was a prime site for thermal injury. The D values derived from counts on selective and non-selective media confirm that nisin is primarily inactivating those cells that have been heat injured; presumably those where injury is to the cell’s outer membrane (OM). Conformational alterations of the OM proteins and lipopolysaccharides which can take place during heating (Katsui et al., 1982) change the structure of the OM and its permeability (Tsuchido et al., 1985). Under these circumstances nisin can gain access to the plasma membrane where it exerts its effect. The saturation kinetics suggest that the number of sites where this can occur is limited so that increased concentrations of nisin have no additional effect once these sites are occupied. In food systems the effect of nisin was far less pronounced. Binding of nisin by lipids (particularly phospholipids) and protein has been identified as the reason for its decreased activity in other complex food systems such as meat (Delves-Broughton, 1990). Lipids are known to interfere with nisin activity (Daeschel, 1990) and their presence in LWE
was probably the cause of the more pronounced neutralisation seen with this product where concentrations as high as 2500 IU / ml did not give any statistically significant reduction in D value. The higher pH in the egg products, 7.5–7.8 in LWE and 9.0–9.3 in EW, compared with NB (7.3) could also reduce the effect since nisin is less stable as pH increases (Delves-Broughton, 1990). The observation that nisin caused more cell membrane injury in S. enteritidis on heating in EW than in NB is interesting and suggests a synergistic action with egg white components which make cells more sensitive to deoxycholate. Lysozyme is a possible candidate. Perhaps nisin itself interacts with the outer membrane allowing lysozyme increased access to the underlying peptidoglycan. Synergistic effects of nisin and lysozyme against Gram positive bacteria have been reported by Monticello (1989), though this has not been reported previously in Gram negatives. This requires further investigation. The results show that nisin can contribute to a reduction in pasteurisation time in heat sensitive products and in products such as egg white, the increased injury in survivors will render them more sensitive to other preservation hurdles that may be present.
Acknowledgements We would like to thank the State Scholarship Foundation (I.K.Y.) of the Hellenic Republic for the financial support of I.S.B.
References Alexandrou, O., Blackburn, C. deW, Adams, M.R., 1995. Capacitance measurement to assess acid-induced injury to Salmonella enteritidis PT4. Int. J. Food Microbiol. 27, 27–36. Clark, C.W., Ordal, Z.J., 1969. Thermal injury and recovery of Salmonella typhimurium and its effect on enumeration procedures. Appl. Microbiol. 18, 332–336. Cutter, C.N., Siragusa, G.R., 1995. Population reduction of gramnegative pathogens following treatments with nisin and chelators under various conditions. J. Food Protect. 58, 977– 983. Daeschel, M.A., 1990. Applications of bacteriocins in food systems. In: Bills, D.D., Kung, S. (Eds.), Biotechnology and Food Safety. Butterworth-Heinemann, Stoneham, MA, pp. 91–104.
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13 Delves-Broughton, J., 1990. Nisin and its use as a food preservative. Food Technol. 4, 100–112. Fowler, G.G., Jarvis, B., Tramer, J., 1975. The assay of nisin in foods. Soc. Appl. Bacteriol. Tech. Series 8, 91–105. Hogue, A.T., Ebel, E.D., Thomas, L.A., Schlosser, W., Bufano, N., Ferris, K., 1997. Surveys of Salmonella enteritidis in unpasteurised liquid egg and spent hens at slaughter. J. Food Protect. 60, 1194–1200. Humpheson, L., Adams, M.R., Anderson, W.A., Cole, M.B., 1998. Biphasic thermal inactivation kinetics in Salmonella enteritidis PT4. Appl. Environ. Microbiol., in press. Kalchayanand, N., Hanlin, M.B., Ray, B., 1992. Sublethal injury makes gram-negative and resistant gram-positive bacteria sensitive to the bacteriocins, pediocin AcH and nisin. Lett. Appl. Microbiol. 15, 239–243. Katsui, N., Tsuchido, T., Hiramatsu, R., Fujikawa, S., Takano, M., Shibasaki, I., 1982. Heat-induced blebbing and vesculation of the outer membrane of Escherichia coli. J. Bacteriol. 151, 1523–1531. Mackey, B.M., Derrick, C.M., 1984. Conductance measurements of the lag phase of injured Salmonella typhimurium. J. Appl. Bacteriol. 57, 299–308.
13
Monticello, D.J., 1989. Control of microbial growth with nisin / lysozyme formulations. European Patent Application No 89123445.2. Ray, B., 1979. Methods to detect stressed microorganisms. J. Food Protect. 42, 346–355. Ray, B., 1993. Sublethal injury, bacteriocins and food microbiology. ASM News 59, 285–291. Ray, B., Jezeski, J., Busta, F.F., 1971. Repair of injury in freezedried Salmonella anatum. Appl. Microbiol. 22, 401–407. Stevens, K.A., Sheldon, B.W., Klapes, N.A., Klaenhammer, T.R., 1991. Nisin treatment for inactivation of Salmonella species and other gram-negative bacteria. Appl. Environ. Microbiol. 57, 3613–3615. Stevens, K.A., Sheldon, B.W., Klapes, N.A., Klaenhammer, T.R., 1992. Effect of treatment conditions on nisin inactivation of gram-negative bacteria. J. Food Protect. 55, 763–766. Tsuchido, T., Katsui, N., Takeuchi, A., Takano, M., Shibasaki, I., 1985. Destruction of the outer membrane permeability barrier of Escherichia coli by heat treatment. Appl. Environ. Microbiol. 50, 298–303. Vaara, M., 1992. Agents that increase the permeability of the outer membrane. Microbiol. Rev. 56, 395–411.
Effect of nisin on heat injury and inactivation of Salmonella enteritidis PT4 1
I.S. Boziaris, L. Humpheson , M.R. Adams* School of Biological Sciences, University of Surrey, Guildford, Surrey, GU2 5 XH, UK Received 1 May 1998; accepted 27 May 1998
Abstract The ability of heat injury to confer sensitivity to nisin in a Gram negative pathogen was investigated. Injury and inactivation kinetics of Salmonella enteritidis PT4 in the presence of nisin were determined in media, liquid whole egg and egg white using cultural methods and capacitance monitoring to detect injury. Addition of nisin in concentrations from 500 IU / ml to 2500 IU / ml in the heating menstruum caused a reduction of required pasteurisation time of up to 35%, principally as a result of its effect on cells suffering damage during heating. In egg white and liquid whole egg the organism’s heat susceptibility was greater than in nutrient broth, particularly in egg white which contained no fat and had an alkaline pH. The effect of nisin on heat susceptibility was however less pronounced than in nutrient broth due to its interaction with protein and fat. Though nisin did not enhance the lethality of heat processes, injury is more severe in egg white containing nisin, presumably as a result of its interaction with antimicrobial factors in egg white. 1998 Elsevier Science B.V. All rights reserved. Keywords: Nisin; Pasteurisation; Thermal injury; Salmonella
1. Introduction Nisin is a bacteriocin produced by Lactococcus lactis subsp. lactis which exhibits a broad spectrum of inhibitory activity against Gram-positive organisms including bacterial spores. It is not generally active against Gram-negative bacteria, yeasts and fungi, although Gram negatives do show nisin-sen*Corresponding author. Tel.: 1 44 1483 300 800; fax: 1 44 1483 300 374; E-mail [email protected] 1 Present address: SAC Auchincruive, Ayrshire, Scotland KA6 5HW, UK.
sitivity when their outer membrane permeability is affected by treatment with chelators such as EDTA, citrate and phosphates (Stevens et al., 1991, 1992; Cutter and Siragusa, 1995; Vaara, 1992). Physical treatments can also affect outer membrane permeability. Tsuchido et al. (1985) have reported its destruction in E. coli by heating and Kalchayanand et al. (1992) found that sublethally injured Gramnegatives produced by freezing and heating are sensitive to nisin. It is therefore possible that sublethal injury to Gram negatives during pasteurisation processes could induce sensitivity to nisin. Inclusion of nisin could contribute to a reduction of heating
0168-1605 / 98 / $19.00 1998 Elsevier Science B.V. All rights reserved. PII: S0168-1605( 98 )00083-X
8
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
time, helping to reduce costs, conserve heat sensitive characteristics of a product and help meet consumer demands for safe but less processed foods. Egg products are an obvious potential application; a recent survey in the United States reported 48% of unpasteurized liquid egg samples to be contaminated with salmonella (Hogue et al., 1997). In this paper we report a study on thermal injury and the heat sensitivity of Salmonella enteritidis in nutrient media, liquid whole egg and egg white in the presence of nisin.
2. Materials and methods
2.1. Organism Salmonella enteritidis phage type 4 P167807, supplied by Division of Enteric Pathogens, Central Public Health Laboratory, London UK, was stored frozen in bead vials (Protect; Technical Service Consultants Ltd, Heywood, Lancashire, UK) at 2 708C and resuscitated to 10 9 cfu / ml in 10 ml Nutrient Broth (NB; Unipath, Basingstoke, Hampshire, UK) at 378C for 24 h. All the microbiological media used were supplied by Oxoid (Basingstoke, Hampshire, UK), unless otherwise stated.
2.2. Nisin Nisin was supplied in a purified form (5 3 10 7 IU / g) by Aplin and Barrett, Beaminster, Dorset, UK. Solutions of nisin were prepared in 0.02 N HCl (pH 2) and were stored at 48C. The activity of nisin was determined by horizontal agar plate diffusion according to the Fowler et al. (1975) bioassay. The assay was carried out for nisin in nutrient broth (NB), egg white (EW) and liquid whole egg (LWE). In NB, samples were taken at the beginning and the end of the heat treatment, diluted as appropriate in 0.02 N HCl and 0.1 ml of each sample transferred into the wells of the assay medium. In the case of EW and LWE, 1 ml samples were taken at the beginning and the end of the thermal treatment acidified with concentrated HCl to pH 2, boiled for 5 min, cooled rapidly, adjusted to 10 ml with HCl 0.02 N and centrifuged at 1000 3 g for 20 min. The aqueous supernatant was filtered in GF /A Whatman glass fibre paper (Whatman, Maid-
stone, Kent, UK), diluted as appropriate with 0.02 N HCl and 0.1 ml of each sample used in the bioassay. At least three replicates for both standards and samples were carried out.
2.3. Heat challenge Resuscitated cultures were diluted 10-fold in maximum recovery diluent (MRD) for the inoculation of pre-warmed NB at 378C (50 or 100 ml) to give an initial suspension of | 1 to 10 cfu / ml. All broths were incubated statically at 378C for 22 h61 h and immediately centrifuged (1000 3 g for 15 min at 208C). Previous determination of a growth curve under these conditions had shown the culture to be well into the stationary phase at this stage.
2.3.1. Preparation of eggs prior to heat challenge Locally purchased eggs were washed in a detergent solution and immersed for a few seconds in industrial methylated spirit to kill the microorganisms on the shell and allowed to air-dry. They were then broken and the contents (egg white or liquid whole egg) collected in a sterile beaker. Then 25 ml samples were transferred into 50-ml plastic tubes (Bibby Sterilin, Staffordshire, UK). The pH of both EW and LWE was measured with a BDH Gelplus, double junction, flat tip, electrode (BDH, Merck, Poole, Dorset, UK). 2.3.2. Heat challenge in NB Centrifuged cell pellets were resuspended in 1 ml prewarmed at 378C NB. Nisin solution or 0.02 N HCl (0.2 ml) was added to NB already heated at the temperature of 558C in order to give the required final nisin concentration. The heating menstruum was held at the temperature of 558C in a plugged 100 ml conical flask in a thermostated water bath 5560.058C and inoculated with 1 ml of the prepared suspension. The flask contents were stirred via a magnetic follower, propelled by a custom-made 12 V d.c. submersible stirrer operating at 60 rev. / min to minimise vortex formation. Temperature regulation was provided by a Haake DC-1 circulator heater (Fisons, Loughborough, Leicestershire, UK). Heating menstruum temperature was measured using a NAMAS certified probe and digital indicator (Pt 100
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
probe and Series 268 indicator; Anville Instruments, Camberley, Surrey, UK).
2.3.3. Heat challenge in egg white and liquid whole egg With EW and LWE, nisin solution or HCl 0.02 N (0.2 ml) was added to 25 ml of EW or LWE in order to give the required final concentration and vortexed for 10 s. One ml of the prepared cell suspension was added and the mixture was vortexed again for 10 s. One ml of the inoculated mixture was placed in each of the 75 3 12 mm test tubes closed with hydrophobic cotton. They were subsequently placed into a water bath at 558C. A temperature probe (Pt 100 probe and Series 268 indicator; Anville Instruments) was inserted into the geometric centre of the liquid in a control tube containing 1 ml of uninoculated EW or LWE. 2.4. Enumeration of the cells With NB, a 1 ml sample was taken every 5 min, diluted as appropriate in maximum recovery diluent (MRD; 0.85% NaCl, 0.1% bacteriological peptone) and enumerated as 0.1 ml spread plates on Nutrient and XLD agar. Plates were incubated for 48 h at 378C. With EW and LWE (every 1 and 3 min, respectively) the contents of one of the test tubes was cooled immediately to below 378C by adding 3 ml of MRD at room temperature and the suspension transferred to a sterile universal vial. The test tube was rinsed (3 3 2 ml MRD) and the washings transferred to the sterile universal to give the first dilution. Samples were then diluted as appropriate and plated (0.1 ml) onto NA and XLD medium.
9
Purpose Plus Medium (BGPPM, bioMerieux). The detection time, DT, was recorded after incubation of the Bactometer wells at 378C. Cells were also enumerated by plating 0.1 ml of the same dilution used to inoculate the Bactometer well, on NA and XLD and incubating it at 378C for 48 h. A calibration graph of log(cfu / ml) on NA against DT was first plotted by using non heat-stressed Salmonella enteritidis cells.
2.6. Calculation of D values For NB and LWE, D values were calculated from the linear regression of log survivors against time. With EW, the short duration of the heat process meant that the bacterial population was very low by the time the medium reached the 558C. It was therefore necessary to take the heating-up time into account when calculating the D values. Accordingly D values were derived from the number of pasteurisation units at 558C (P55 ) required to produce a 4 log reduction in survivors. One P55 corresponds to a process equivalent in terms of its lethality to 1 min at 558C. These were calculated from the survivor curves and lethal rate curves derived from time temperature data taken during the heat treatment. With the z value of S. enteritidis equal to 48C (Humpheson et al., 1998), the lethal rate (Lv) at any time t is equal to: Lv 5 10 (T255) / 4 , where T is the temperature at time t. The area under the Lv curve between times t 1 and t 2 which correspond to 7 log cfu / g and 3 log cfu / g survivors respectively, was calculated by using FigP (Version 2.5, FigP Software Corp., Durham, NC, USA). Each experiment was carried out at least in triplicate.
2.5. Capacitance measurements
3. Results
Capacitance measurements were made with the Bactometer Model 120SC (Bactomatic, Inc. USA) with a Model 123 central data processor (bioMerieux). The heat challenge procedure was the same as described previously. Samples were taken every 5 min, diluted in MRD to give a population less than 10 5 cfu / ml and 0.1 ml of the diluted samples (in duplicate) was placed into the Bactometer disposable module wells (bioMerieux Vitek, Inc. 99052) containing 1 ml of Bactometer General
3.1. Injury and inactivation during heating Sublethal injury in survivors during heating at 558C was assessed using two methods: the difference between counts on selective (XLD) and non-selective (NA) media and the extension of capacitance detection time for survivors relative to that predicted from the calibration curve for the same number of uninjured cells. Both methods showed the initial population to be
10
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
uninjured (zero extension of detection time and identical counts on NA and XLD) and the extent of injury to increase during heating as indicated by the increasing difference between the colony counts on XLD and NA and the predicted and observed detection times (Fig. 1). The actual extent of injury can be represented by the extension of detection time and by the difference between counts on non-selective and selective media as a proportion of the total population ((NA 2 XLD) / NA) (Fig. 2). Extension of
Fig. 1. Nutrient agar and XLD counts and predicted and observed capacitance detection times of Salmonella enteritidis at 558C in nutrient broth. d, NA counts; s, XLD counts; j, predicted DT; h, observed DT.
Fig. 2. The proportion of the injured cells and the extension of detection time, observed during heating at 558C in nutrient broth. s, extension of DT; d, (NA 2 XLD) / NA.
detection time (DT observed 2 DT predicted ) and proportion of injury curves both show a similar profile with the extent of injury increasing rapidly so that more than 99% of survivors displayed injury after 10 min heating, reflected in an extended lag period (detection time) of about 5 h (Fig. 2). After 15 min the extent of injury increased at a much slower rate. To determine whether the increasing injury during heating affected the nisin sensitivity of the organisms, thermal death curves were produced using nutrient broth containing different nisin concentrations. The mean D values determined from counts of total survivors (NA) and uninjured survivors (XLD) are presented as Fig. 3. The D value derived from XLD counts measures the combined rate of both injury and inactivation during heating and is therefore much shorter than the D value from NA counts which measure primarily the rate of inactivation. From Fig. 3, it is apparent that as nisin concentration increases, the NA-derived D values decrease by up to 35% while the XLD-derived D value decreased only slightly, but significantly (P , 0.05). This indicates that nisin is killing primarily those cells that have been sublethally injured by the heat treatment whereas uninjured cells are less affected. A saturation effect of nisin against injured cells is also apparent. Concentrations of nisin higher than 1500 IU / ml did not cause any significant additional effect. Analysis of variance showed that D values for S.
Fig. 3. Effect of nisin on the D value of S. enteritidis at 558C in nutrient broth (pH 7.3). j, NA counts; h, XLD counts. The error bars show the standard deviation.
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
11
enteritidis at nisin concentrations of 1500, 2000 and 2500 IU / ml did not differ significantly (P . 0.05) in contrast with the D values obtained at 0, 500, 1000 and 1500 IU / ml.
3.2. Heat injury and inactivation in egg white and liquid whole egg Egg white (EW) and liquid whole egg (LWE) are two heat sensitive products where a reduction in the time / temperature of pasteurisation or an increased lethality with existing protocols is desirable. The different experimental procedure required when determining survival of S. enteritidis in heated EW meant that D values were derived from the number of P55 units required for a reduction from 7 log cfu / ml to 3 log cfu / ml of S. enteritidis (see Section 2). The D 55 value in that case being the P55 divided by four (the log cfu / ml reduction). The D 55 against nisin concentration is plotted in Fig. 4. In LWE where, unlike NB, there was an appreciable heating up time (6 min), D values at 558C were calculated from survivors curves once the temperature had reached 558C. The linear region of the survivor curve extended over a 5 log reduction in survivors and had an r 2 . 0.99. The equivalent D values from XLD counts were also calculated and both are shown in Fig. 5 for different concentrations of nisin.
Fig. 4. Effect of nisin on D value of S. enteritidis in egg white (pH 9.0–9.3). j, NA counts; h, XLD counts. The error bars show the standard deviation.
Fig. 5. Effect of nisin on the D value of S. enteritidis at 558C in liquid whole egg (pH 7.5–7.8). j, NA counts; h, XLD counts. The error bars show the standard deviation.
D values in LWE were considerably longer than in EW presumably a result of the combined effect of the higher pH in EW (9.0 to 9.3 compared with 7.5 to 7.8 in LWE) and the protective effect of fat in the LWE. The reduction of D value and therefore heating time in the presence of nisin was significant in EW but not in LWE. The percentage reduction in D value in EW was not however as great as that achieved in nutrient broth; 500 IU / ml showed no reduction and for 1500 and 2500 IU / ml the reduction of D value in EW was 17.7 and 26.5%, respectively. Equivalent figures for NB were 29.9 and 35%, respectively. This can be ascribed to the lower activity of nisin at the pH of the egg white (9.0–9.3) and possible binding of nisin with egg white proteins which reduced its antibacterial activity. Nisin levels were measured before and after all heat treatments and found to be virtually unchanged suggesting that any binding was reversible. Though nisin was only seen to enhance the lethal effect of heat in EW at higher concentrations, it did produce appreciably greater rates of injury in survivors than were observed in nutrient broth. The D values derived from the XLD counts decreased more in EW with nisin than in NB with nisin. For example, at 2500 IU / ml the percentage of reduction in NB was only 16% in contrast to 26% in EW. This was not seen in LWE, where the XLD-derived D values did not differ significantly with increasing nisin concentration.
12
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13
4. Discussion Injured cells can repair and grow on non-selective media but may not be able to grow on selective media due to their increased sensitivity to selective agents (Ray, 1979, 1993). Clark and Ordal (1969) and Ray et al. (1971) reported that Salmonella typhimurium and Salmonella anatum cells injured by heating and freeze drying, respectively, were sensitive to deoxycholate, the selective agent used in xylose lysine deoxycholate (XLD) agar. The NA 2 XLD difference is therefore taken to represent the population that has been subject to injury affecting the cell outer membrane which increases their sensitivity to deoxycholate in XLD (Vaara, 1992). Injury may not however always render cells sensitive to a particular selective agent and capacitance monitoring offers an alternative instrumental technique, which measures the extended lag phase while injured populations repair. It therefore detects a broader range of injury (Mackey and Derrick, 1984; Alexandrou et al., 1995). This work further illustrates the usefulness of impedimetric methods for determining total injury in microbial populations. The similar profile of injury as assessed by impedimetry and differential counts suggests that the outer membrane was a prime site for thermal injury. The D values derived from counts on selective and non-selective media confirm that nisin is primarily inactivating those cells that have been heat injured; presumably those where injury is to the cell’s outer membrane (OM). Conformational alterations of the OM proteins and lipopolysaccharides which can take place during heating (Katsui et al., 1982) change the structure of the OM and its permeability (Tsuchido et al., 1985). Under these circumstances nisin can gain access to the plasma membrane where it exerts its effect. The saturation kinetics suggest that the number of sites where this can occur is limited so that increased concentrations of nisin have no additional effect once these sites are occupied. In food systems the effect of nisin was far less pronounced. Binding of nisin by lipids (particularly phospholipids) and protein has been identified as the reason for its decreased activity in other complex food systems such as meat (Delves-Broughton, 1990). Lipids are known to interfere with nisin activity (Daeschel, 1990) and their presence in LWE
was probably the cause of the more pronounced neutralisation seen with this product where concentrations as high as 2500 IU / ml did not give any statistically significant reduction in D value. The higher pH in the egg products, 7.5–7.8 in LWE and 9.0–9.3 in EW, compared with NB (7.3) could also reduce the effect since nisin is less stable as pH increases (Delves-Broughton, 1990). The observation that nisin caused more cell membrane injury in S. enteritidis on heating in EW than in NB is interesting and suggests a synergistic action with egg white components which make cells more sensitive to deoxycholate. Lysozyme is a possible candidate. Perhaps nisin itself interacts with the outer membrane allowing lysozyme increased access to the underlying peptidoglycan. Synergistic effects of nisin and lysozyme against Gram positive bacteria have been reported by Monticello (1989), though this has not been reported previously in Gram negatives. This requires further investigation. The results show that nisin can contribute to a reduction in pasteurisation time in heat sensitive products and in products such as egg white, the increased injury in survivors will render them more sensitive to other preservation hurdles that may be present.
Acknowledgements We would like to thank the State Scholarship Foundation (I.K.Y.) of the Hellenic Republic for the financial support of I.S.B.
References Alexandrou, O., Blackburn, C. deW, Adams, M.R., 1995. Capacitance measurement to assess acid-induced injury to Salmonella enteritidis PT4. Int. J. Food Microbiol. 27, 27–36. Clark, C.W., Ordal, Z.J., 1969. Thermal injury and recovery of Salmonella typhimurium and its effect on enumeration procedures. Appl. Microbiol. 18, 332–336. Cutter, C.N., Siragusa, G.R., 1995. Population reduction of gramnegative pathogens following treatments with nisin and chelators under various conditions. J. Food Protect. 58, 977– 983. Daeschel, M.A., 1990. Applications of bacteriocins in food systems. In: Bills, D.D., Kung, S. (Eds.), Biotechnology and Food Safety. Butterworth-Heinemann, Stoneham, MA, pp. 91–104.
I.S. Boziaris et al. / International Journal of Food Microbiology 43 (1998) 7 – 13 Delves-Broughton, J., 1990. Nisin and its use as a food preservative. Food Technol. 4, 100–112. Fowler, G.G., Jarvis, B., Tramer, J., 1975. The assay of nisin in foods. Soc. Appl. Bacteriol. Tech. Series 8, 91–105. Hogue, A.T., Ebel, E.D., Thomas, L.A., Schlosser, W., Bufano, N., Ferris, K., 1997. Surveys of Salmonella enteritidis in unpasteurised liquid egg and spent hens at slaughter. J. Food Protect. 60, 1194–1200. Humpheson, L., Adams, M.R., Anderson, W.A., Cole, M.B., 1998. Biphasic thermal inactivation kinetics in Salmonella enteritidis PT4. Appl. Environ. Microbiol., in press. Kalchayanand, N., Hanlin, M.B., Ray, B., 1992. Sublethal injury makes gram-negative and resistant gram-positive bacteria sensitive to the bacteriocins, pediocin AcH and nisin. Lett. Appl. Microbiol. 15, 239–243. Katsui, N., Tsuchido, T., Hiramatsu, R., Fujikawa, S., Takano, M., Shibasaki, I., 1982. Heat-induced blebbing and vesculation of the outer membrane of Escherichia coli. J. Bacteriol. 151, 1523–1531. Mackey, B.M., Derrick, C.M., 1984. Conductance measurements of the lag phase of injured Salmonella typhimurium. J. Appl. Bacteriol. 57, 299–308.
13
Monticello, D.J., 1989. Control of microbial growth with nisin / lysozyme formulations. European Patent Application No 89123445.2. Ray, B., 1979. Methods to detect stressed microorganisms. J. Food Protect. 42, 346–355. Ray, B., 1993. Sublethal injury, bacteriocins and food microbiology. ASM News 59, 285–291. Ray, B., Jezeski, J., Busta, F.F., 1971. Repair of injury in freezedried Salmonella anatum. Appl. Microbiol. 22, 401–407. Stevens, K.A., Sheldon, B.W., Klapes, N.A., Klaenhammer, T.R., 1991. Nisin treatment for inactivation of Salmonella species and other gram-negative bacteria. Appl. Environ. Microbiol. 57, 3613–3615. Stevens, K.A., Sheldon, B.W., Klapes, N.A., Klaenhammer, T.R., 1992. Effect of treatment conditions on nisin inactivation of gram-negative bacteria. J. Food Protect. 55, 763–766. Tsuchido, T., Katsui, N., Takeuchi, A., Takano, M., Shibasaki, I., 1985. Destruction of the outer membrane permeability barrier of Escherichia coli by heat treatment. Appl. Environ. Microbiol. 50, 298–303. Vaara, M., 1992. Agents that increase the permeability of the outer membrane. Microbiol. Rev. 56, 395–411.