R. Bennour Miled
1, L.Guillier
2, S. Neves
1, J-C. Augustin
3, P. Colin
4and N. Gnanou-Besse
11Agence française de sécurité sanitaire des aliments, Afssa Laboratoire d’Etudes et de Recherches sur la Qualité des Aliments et des Procédés agro-alimentaires. Afssa LERQAP, 23 Avenue du Général de Gaulle, 94706 Maisons Alfort cedex, France. (r.miled@afssa.fr; n.besse@afssa.fr).
2Agence française de sécurité sanitaire des aliments, Afssa Direction de l’Evaluation des Risques Nutritionnels et Sanitaires (Afssa DERNS, 27-31 Avenue du Général Leclerc, 94701 Maisons Alfort cedex, France).
3Ecole Nationale Vétérinaire d’Alfort (7 Avenue du Général de Gaulle, 94706 Maisons Alfort cedex, France).
4 Ecole Supérieure de Microbiologie et de Sécurité Alimentaire (ESMISAB, Plouzane, France).
Abstract
Cells of six strains of Cronobacter were submitted to a dry stress and stored for 2.5 months at ambient temperature. The individual cell lag time distributions of recovered cells were characterised at 25°C and 37°C in non selective broth. The individual cell lag times were deduced from the times for cultures issued from individual cells to reach an optical density threshold. In parallel, growth curves for each strain at high contamination levels were determined in the same growth conditions. In general, the extreme value type II distribution was the most effective to describe the 12 observed distributions of individual cell lag times.
Recently, an innovating model which allowed to characterise individual cell lag time distribution from populational growth parameters was developed for other food-borne pathogenic bacteria such as L. monocytogenes. We verified the applicability of this model to Cronobacter by comparing the mean and the standard deviation of individual cell lag times to populational lag times observed with high initial concentration experiments, and then by deducing the theoretical cell lag times distributions from the observed mean and the standard deviation of cell lag times. We also validated the model in realistic conditions by studying growth in powdered infant formula decimally diluted in Buffered Peptone Water (BPW), which represents the first enrichment step of the standardised detection method for Cronobacter.
Keywords : Enterobacter sakazakii, Cronobacter, powdered infant food formula, individual cell lag time, pooling, growth.
Introduction
Enterobacter sakazakii recently known as Cronobacter (Iversen et al., 2008) is considered as an opportunistic pathogen and has been implicated in outbreaks causing meningitis or bacteraemia, especially in neonates and infants with mortality rates of 20 to 50%
(Anonymous, 2006a). In most cases, powdered infant formula (PIF) has been identified as the source of infection.
In PIF, contamination levels are extremely low and generally much lower than 1 cfu per 100g (Anonymous, 2006a, 2008). Mistakes in biberon-preparation practices, such as improper holding temperatures, may lead to a critical cell level, and the occurrence of the infection. In such conditions of very low contamination levels, individual cells variability can have an important impact on the pathogen growth.
Knowing how long-term presence in PIF, and subsequent stress, affect the variability of single-cell lag times is extremely important in assessing the risk of cell recovery and growth in reconstituted milk or in enrichment broth, where low numbers of stressed cells of pathogenic bacteria may be distributed among PIF samples. Recently, a model which allowed to characterise individual cell lag time distribution from populational growth parameters has been developed for L. monocytogenes (Guillier and Augustin, 2006, 2008).
The first objective of the present work was to verify the applicability of this model to Cronobacter submitted to a dry stress for different regrowth conditions and strains. The second objective was to study single-cell lag times in a realistic condition such as the first enrichment of the standardised detection method (ISO/TS 22964, Anonymous, 2006b) which is performed in non selective Buffered Peptone Water (BPW). The log count distribution, or vertical distribution (D'Arrigo et al., 2006) was applied to estimate the distribution of the single-cell lag times in BPW. This study also allowed us to evaluate the impact of pooling samples on Cronobacter growth and detection. Indeed, to reduce analytic cost and heaviness, common practice in food industry consists in pooling samples at constant dilution rate, in order to perform a single pre-enrichment and subsequent analysis. Consequences on Cronobacter detection are not established.
Materials and methods
Six strains of Cronobacter were used in this study: 4 strains belonged to different species of Cronobacter (C. malonaticus, C.muytjensii and C. turicensis) and 3 to the same species (C.
sakazakii). Strains were submitted to desiccation: Cronobacter strains grown in an equal mixture of BHI and sterile infant formula for 24 h at 37°C were freeze-dried using the CHRIST LOC-2M apparatus (Bioblock Scientific, Ile de France, Vanves cedex, France).
Contaminated powder was further 1 in 100 diluted in PIF intended for infants below 6 months of age (previously tested not contaminated with Cronobacter and with a very low level of total microflora), and stored for 2.5 months at ambient temperature before use. The individual cell lag time distributions were characterised at 25°C and 37°C in non selective Brain Heart Infusion (BHI) broth: individual cell lag times were deduced from the times for cultures issued from individual cells to reach an optical density threshold, by measuring optical density (OD) at 600 nm using an automated spectrophotometer (Bioscreen C reader). In parallel, growth curves for each strain at high contamination levels (100-1000 cfu/g) were determined in triplicate in BHI broth, at 25°C and 37°C. Growth was monitored by direct plating enumeration, and curves were fitted to the Baranyi model using MicroFit software (http://www.ifr.ac.uk/MicroFit/ ).
For validation purpose, the distribution of Cronobacter log counts at given times during first-age PIF pre-enrichment was applied to estimate the distribution of the single-cell lag times. 40 bags of 10g and 40 bags of 100g (mimicking a pooling of 10*10g samples) first-age PIF were prepared and homogenised in sterile BPW diluent (1 in 10 dilution). Each sample was inoculated with freeze-dried C.sakazakii type strain (ATCC 29544) at a contamination level of 4 cells per bag. All bags were incubated at 37°C and enumerated after 8h and 20h by plating on the chromogenic selective isolation agar “Enterobacter sakazakii Isolation Agar”
(ESIA). In parallel, growth of PIF background microflora and of high Cronobacter populations were monitored in the same conditions.
Results and discussion
Four statistical distributions were tested in this study to describe data sets of single-cell lag time: the Gamma distribution, the Weibull distribution, the Log-Normal distribution, and the Extreme Value type II distribution. In general, the Extreme Value type II distribution provided the best fit over the whole range of growth conditions and strains tested. Guillier and Augustin (2006) investigated the individual lag times of L.monocytogenes cells and showed that this distribution was also the best one.
The relationships between the standard deviations and the means of individual cell lag times and between the individual cell lag times and the population lag times, were in agreement with those observed by Guillier and Augustin (2006, 2008).
In the second part of this study, these relations were applied for validation purpose: at two times of enrichment procedure of a low number of cells, Cronobacter log counts were both measured and estimated from growth rate and individual lag times calculated from population
growth curves. After 8 hours of enrichment, the results obtained for 10g and 100g PIF samples showed good agreement between observed and predicted values only if the variability of N0, individual lag times and growth rate are taken into account for predictions (Figure 1). Significant differences with observed values were found if lag time variability was not considered, which confirmed that individual cell lag times variability has a major impact on growth.
Figure 1: Cumulative distribution of C. sakazakii observed log counts () after 8 hours (10g samples). Predicted populations: () N0, µmax and lagi considered as constant, (--) N0 and µmax
with variability and lagi constant, (...)N0, µmax and lagi with variability.
After 20 h, vertical distributions were not in agreement with the predicted values of log counts (Figure 2). This difference can be attributed to bacterial interactions. Indeed, for population growth curves, we observed a stop of Cronobacter growth when background flora reached its stationary phase (approximately after 9h enrichment).
Figure 2: Cumulative distribution of observed log counts after 8 hours () and 20 hours () of C. sakazakii in enrichment medium (10g samples). Predicted concentrations after
8 hours (- -) and 20 hours ().
To better explore this phenomenon, we used the same previous simulations but with a stop of the growth at 9h, when the total microflora attained its maximum concentration. Results showed that for 10g bags and for 100g bags we obtained a good correlation between observed and predicted values (Figure 3). Furthermore, for the 100g samples, the initial concentration
cdf
log
10(cfu/ml)
log
10(cfu/ml)
cdf
is weaker than for the 10g samples which emphasized the negative impact of pooling on detection.
Figure 3: Cumulative distribution of observed log counts of C. sakazakii in enrichment medium after 20 hours in 100g samples () and 10g samples (). Predicted log counts in
100g (- -) and 10g () bags, taking into account variability of individual lag times and microbial interaction.
Conclusion
Relationships established for L. monocytogenes between populational and individual parameters of growth were observed for Cronobacter. These relationships can thus be used for predictive modelling and risk assessment studies provided a consolidation of the data.
Relationships were validated for Cronobacter cells undergoing enrichment culturing in BPW for 8h and 20h provided that bacterial interactions are taken into account. This demonstrated the importance of validating the model before use especially in non-selective broth.
We also noticed a strong impact of pooling on the populations of Cronobacter reached at 20h which corresponds to the end of the pre-enrichment duration of the standard detection method. This effect can be explained by a combined effect of a weaker Cronobacter initial concentration in 100g and of a premature stop of the growth due to bacterial interactions.
Thus, from a practical point of view, pooling can have an effect on the sensitivity of the detection method.
References
Anonymous 2006a. Enterobacter sakazakii and Salmonella in powdered infant formula: meeting report.
Microbiological risk assessment series 10. World Health Organisation-Food and Agriculture Organisation of the United Nations, Geneva and Rome.
Anonymous 2006b. Milk and milk products-Detection of Enterobacter sakazakii. ISO/TS 22964:2006 and IDF/RM 210:2006. International Organization for Standardization, Geneva, Switzerland.
Anonymous 2008. Contamination microbienne des préparations lactées en poudres destinées aux nourrissons etpersonnes agées . Afssa, Maisons Alfort, France.
D'Arrigo M.,García de Fernando G.D., Velasco de Diego R., Ordóñez J.A., George S.M. and Pin C. (2006).
Indirect measurement of the lag time distribution of single cells of Listeria innocua in food. Applied and Environmental Microbiology 72, 2533-2538.
Guillier L. and Augustin J.-C. (2006). Modelling the individual cell lag time distributions of Listeria monocytogenes as a function of the physiological state and growth conditions. International Journal of Food Microbiology 111, 241-251.
Guillier L. and Augustin J.-C. (2008). Erratum to “Modelling the individual cell lag time distributions of Listeria monocytogenes as a function of the physiological state and the growth conditions, International Journal of Food Microbiology 111 (2006) 241-251”. International Journal of Food Microbiology 124,114.
Iversen, C., Mullane, N., Lehner, A., Mc Cardell, B., Tall B. D., Lehner A., Fanning S., Stephan R. and Joosten H.
(2008).Cronobacter gen. nov., a new genus to accommodate the biogroups of Enterobacter sakazakii, and
cdf
log
10(cfu/ml)
proposal of Cronobacter sakazakii gen. nov., comb. nov., Cronobacter malonaticus sp. nov., Cronobacter turicensis sp. nov., Cronobacter muytjensii sp. nov., Cronobacter dublinensis sp. nov., Cronobacter genomospecies 1, and of three subspecies, Cronobacter dublinensis subsp. dublinensis subsp. nov., Cronobacter dublinensis subsp. lausannensis subsp. nov. and Cronobacter dublinensis subsp. lactaridi subsp.
nov. International Journal of Systematic and Evolutionary Microbiology, 58, 1442-1447.