Global overview of the risk linked to the Bacillus cereus group in the egg product industry: identification of food safety and food spoilage markers

Authors

  • C. Techer,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • F. Baron,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • L. Delbrassinne,

    1. Food, Medicines, and Consumer Safety, Scientific Institute of Public Health (WIV-ISP), Brussels, Belgium
    Search for more papers by this author
  • R. Belaïd,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • N. Brunet,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • A. Gillard,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • F. Gonnet,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • M.-F. Cochet,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • N. Grosset,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • M. Gautier,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • M. Andjelkovic,

    1. Food, Medicines, and Consumer Safety, Scientific Institute of Public Health (WIV-ISP), Brussels, Belgium
    Search for more papers by this author
  • V. Lechevalier,

    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    Search for more papers by this author
  • S. Jan

    Corresponding author
    1. Equipe Microbiologie de l'Œuf et des Ovoproduits (MICOV), Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, Rennes, France
    • Correspondence

      Sophie Jan, Agrocampus Ouest, INRA, UMR1253 Science et Technologie du Lait et de l'Œuf, 65 rue de Saint Brieuc, 35042 Rennes, France.

      E-mail: sophie.jan@agrocampus-ouest.fr

    Search for more papers by this author

Abstract

Aims

To evaluate the food safety and spoilage risks associated with psychrotrophic Bacillus cereus group bacteria for the egg product industry and to search for relevant risk markers.

Methods and Results

A collection of 68 psychrotrophic B. cereus group isolates, coming from pasteurized liquid whole egg products, was analysed through a principal component analysis (PCA) regarding their spoilage and food safety risk potentials. The principal component analysis showed a clear differentiation between two groups within the collection, one half of the isolates representing a safety risk and the other half a spoilage risk.

Conclusions

Relevant risk markers were highlighted by PCA, that is (i) for the food safety risk, the presence of the specific 16S rDNA-1m genetic signature and the ability to grow at 43°C on solid medium and (ii) for the spoilage risk, the presence of the cspA genetic signature.

Significance and Impact of the Study

This work represents a first step in the development of new diagnostic technologies for the assessment of the microbiological quality of foods likely to be contaminated with psychrotrophic B. cereus group bacteria.

Introduction

Of the 6·4 billion eggs consumed in 2009 in the European Union, more than one-third were consumed in the form of egg products, that is eggs removed from their shells (Magdelaine et al. 2011). Liquid, frozen and dried egg products are widely used by the foodservice industry and may be used as ingredients in sensible food stored at refrigerated temperatures, such as prepared mayonnaise and desserts. Nevertheless, eggs and egg products are the major cause of food poisonings in France; between 2006 and 2008, they were implicated in 15% of the cases of confirmed food poisonings, Salmonella being identified as the main causative agent (76%) (Delmas et al. 2010). Contamination by Salmonella, which remains the major food safety issue in this sector, is now under the control of breeding surveillance and of effective pasteurization processes in France. Nevertheless, the low heat treatments usually applied to liquid egg products, owing their high thermosensitivity, are inefficient on spore-forming bacteria, such as bacteria belonging to the Bacillus cereus group (Granum and Lund 1997; Kotiranta et al. 2000). Bacillus cereus group bacteria can contaminate the egg product processing chain and survive the low heat treatments because of their ubiquitous and spore-forming characteristics. Their growth is a putative source of both (i) food safety problems, due to their ability to produce enterotoxins and/or the emetic toxin (cereulide) and (ii) economic losses for the sector (Svensson et al. 2007), due to their food spoilage potential through various enzymatic activities, that can be expressed at low temperature in the case of psychrotrophic strains. A particular attention should be paid to the microbial quality of egg products which constitute sensitive food products intended to mass catering and/or to immuno-suppressed people.

As a relevant point for further assessing the risk in the sector, we previously demonstrated the ability of a Bacillus weihenstephanensis strain isolated from a spoiled liquid egg product to produce toxins, even at low temperature (Baron et al. 2007). A collection of B. cereus group bacteria coming from pasteurized liquid whole egg products was subsequently built, and their genotypic and phenotypic features related to growth and cytotoxic potentials, in situ (i.e. in liquid whole egg), were analysed at various temperatures (Jan et al. 2011). Supported by these results, additional relevant features have been chosen to be analysed in this study, in order to present a global overview of the food safety and spoilage risks linked to the B. cereus group in the sector of egg product manufacturing.

Although suitable methods are already available for the detection and the enumeration of the B. cereus group in routine laboratories, there is a lack of tools for the assessment of the sanitary and/or spoilage risks potentially posed by these bacteria. In particular, the sector of production of powder infant milk formula production might benefit from new researches in this area, since the enumeration of B. cereus group bacteria is required by the European regulation in this food product (Commission Regulation (EC) No 1441/2007).

With regards to the sanitary risk, the study of Cadot et al. (2010) highlighted specific marker candidates for the distinction between pathogenic and nonpathogenic B. cereus group strains. The level of food safety risk appeared to be dependent on the affiliation of the strain to one of the seven phylogenetic groups described by Guinebretière et al. (2008, 2010). Namely, the highest risk was ascribed to the group III, a moderate risk to the groups VII, IV and II, a lower risk to the group V and a very low risk to the group VI. However, the search for these markers is currently difficult at the industrial level, since it implies gene expression quantification (Cadot et al. 2010) or gene sequencing (Guinebretière et al. 2008, 2010).

Controlling the spoilage issue will help to reduce the waste of products along the food chain and consequently save a considerable amount of food resources among the current global agricultural production (Foresight 2011). Food producers have to engage in the reduction of waste products, especially when they are involved in the downstream processes, which are known as having the highest environmental impact in the production chains. It appears that the food spoilage risk due to well-known enzymatic activities of the B. cereus group may be more accurately considered. This risk has mainly been assessed by targeting enzymatic activities in milk powders, UHT milk or in some vegetables foods, that is foods not requiring storage at refrigerated temperature (Braun et al. 1999; De Jonghe et al. 2010; Nemeckova et al. 2010; Ivy et al. 2012; Schmidt et al. 2012). In the sector of refrigerated pasteurized liquid foods, there is a lack of valuable methods for evaluating the spoilage risk due to the B. cereus group, such as dairy and egg products.

In addition to providing a global overview of the sanitary and spoilage risks linked to the B. cereus group in the egg product industry, the present study offers new insights for the development of pertinent controls, that should be based on relevant markers which are related to each type of risk (sanitary and/or spoilage).

Materials and methods

Characteristics of the bacterial collection

The collection comprised 68 isolates obtained after enrichment at 10°C of samples of liquid pasteurized whole egg products provided by six French egg breaking companies (C1–C6) at consecutive warm and cold seasons. The collection was partially characterized in a previous work, regarding genetic signatures (mesophilic 16S rDNA-1m signature and psychrotrophic 16S rDNA-2p and cspA signatures), growth and cytotoxic activities at various temperatures in optimum medium (Brain Heart Infusion, Yeast Extract—BHI-YE) and in whole egg (Jan et al. 2011; white columns in Table 1).

Table 1. Characterization of the Bacillus cereus group collection (68 isolates) coming from pasteurized liquid whole egg products provided by six egg breaking companies (C1–C6) at two consecutive seasons (cold and warm): presence (1)/absence (0) of the cspA signature as defined by Francis et al. (1998); presence (1)/absence (0) of the 16S rDNA-1m and the16S rDNA-2p signatures as defined by Von Stetten et al. (1998); ability (1)/inability (0) to grow at 43 and 6°C on BHI-YE agar; cell concentration in liquid BHI-YE and in liquid whole egg after incubation of 3 ± 0·5 log CFU ml−1 at 30, 10 and 6°C (mean of three independent experiments in log CFU ml−1 ± standard deviation (SD)); percentage of cytotoxic activity in BHI-YE and in liquid whole egg after incubation of 3 ± 0·5 log CFU ml−1 for 18 h and 5 days at 30 and 10°C, respectively (underlined characters highlight conditions under which the isolates are considered cytotoxic, that is leading to a percentage of Caco-2 cell damage higher than 50%); absence (0) of the ces gene; detection of cereulide synthesis in liquid whole egg at 30 and 10°C (LC/MS2; ND, Not detected); onset of egg spoilage at 30 and 10°C, from an inoculum of 3 ± 0·5 log CFU ml−1 (mean of three independent experiments expressed in h ± SD and day ± SD); affiliation to one of the phylogenetic groups defined by Guinebretière et al. (2008, 2010). White columns refer to results already published in Jan et al. (2011). Grey columns refer to additional results of the present study. Bolded characters highlight several results particularly differing from the others inside the collection, as shown by the PC analysis (see the text)
NrOriginSeasonGenetic signaturesGrowth on agar mediumCell concentration in liquid medium (log CFU ml−1)Cytotoxic activity (%)Cereulide synthesisOnset of egg spoilagePhylogenetic group (*)
ces geneLC/MS2
cspA 16S rDNA-1m16S rDNA-2pBHI-YEBHI-YEEggBHI-YEEggEgg
43°C6°C30°C10°C6°C30°C10°C6°C30°C10°C30°C10°C30°C and 10°C30°C (h)10°C (d)
  1. UC, uncoagulated.

  2. *Phylogenetic group, according to Guinebretière et al. (2008, 2010); the cell concentration of 1·3 log CFU ml−1 corresponded to the detection limit for the enumeration by the micro-method (Baron et al. 2006).

1C1Warm101017·9 ± 0·37·0 ± 0·35·8 ± 0·17·8 ± 0·48·0 ± 0·26·8 ± 0·4 52·2 ± 8·6 0·0 ± 014·4 ± 6·425·6 ± 6·40ND20·3 ± 6·48·0 ± 1·0VI
2C2Warm101017·9 ± 0·36·4 ± 0·25·5 ± 0·28·4 ± 0·28·2 ± 0·17·0 ± 0·32·9 ± 6·20·0 ± 016·6 ± 7·013·2 ± 3·80ND24·0 ± 0·06·7 ± 1·5VI
3C2Warm101018·2 ± 0·47·0 ± 0·55·8 ± 0·18·0 ± 0·18·2 ± 0·16·6 ± 0·827·8 ± 0·20·0 ± 05·1 ± 6·46·5 ± 1·90ND21·0 ± 4·47·3 ± 0·6VI
4C3Warm101018·3 ± 0·26·9 ± 0·45·8 ± 0·38·2 ± 0·18·5 ± 0·17·2 ± 0·0 96·4 ± 1·6 0·0 ± 019·1 ± 8·77·3 ± 0·00ND16·0 ± 3·66·0 ± 1·7VI
5C3Warm101018·2 ± 0·26·8 ± 0·25·8 ± 0·28·3 ± 0·28·3 ± 0·57·5 ± 0·1 97·3 ± 0·9 3·4 ± 0·143·1 ± 0·613·5 ± 3·80ND15·3 ± 2·16·7 ± 1·2VI
6C3Warm101018·3 ± 0·26·9 ± 0·45·7 ± 0·38·3 ± 0·28·4 ± 0·57·1 ± 0·211·6 ± 6·10·0 ± 016·7 ± 8·80·0 ± 7·30ND20·7 ± 4·97·7 ± 0·6VI
7C5Warm101018·0 ± 0·37·3 ± 0·76·2 ± 0·27·9 ± 0·48·1 ± 0·36·4 ± 0·70·0 ± 00·0 ± 016·0 ± 7·910·7 ± 9·30ND19·3 ± 0·66·0 ± 1·7VI
8C6Warm101018·3 ± 0·26·9 ± 0·15·5 ± 0·18·0 ± 0·48·1 ± 0·37·3 ± 0·2 93·9 ± 2·9 0·0 ± 0 78·8 ± 6·9 0·0 ± 4·90ND17·0 ± 5·05·7 ± 1·2VI
9C6Warm101018·2 ± 0·17·2 ± 0·36·1 ± 0·38·2 ± 0·28·2 ± 0·17·2 ± 0·20·0 ± 022·6 ± 0·525·8 ± 8·05·4 ± 7·60ND11·7 ± 0·67·7 ± 3·1VI
10C3Warm101018·1 ± 0·17·1 ± 0·35·9 ± 0·18·2 ± 0·37·9 ± 0·27·4 ± 0·227·4 ± 9·39·8 ± 0·210·5 ± 2·63·3 ± 2·40ND11·7 ± 1·25·0 ± 0·0VI
11C3Warm101018·2 ± 0·26·9 ± 0·26·1 ± 0·48·2 ± 0·48·2 ± 0·37·5 ± 0·719 ± 4·41·3 ± 0·07·9 ± 6·114·6 ± 3·60ND21·7 ± 2·17·0 ± 1·7VI
12C5Warm101018·1 ± 0·26·8 ± 0·35·9 ± 0·18·4 ± 0·37·7 ± 0·77·2 ± 0·2 96·3 ± 1·5 0·0 ± 06·4 ± 5·20·0 ± 00ND18·0 ± 1·08·0 ± 1·0VI
13C3Warm101018·6 ± 0·37·3 ± 0·46·1 ± 0·38·2 ± 0·38·3 ± 0·37·5 ± 0·1 91·8 ± 4·0 14·6 ± 0·336·9 ± 5·510·5 ± 3·40ND15·0 ± 2·65·7 ± 1·2VI
14C3Cold101018·5 ± 0·26·9 ± 0·35·8 ± 0·38·4 ± 0·08·3 ± 0·27·4 ± 0·10·0 ± 08·5 ± 0·20·0 ± 8·17·1 ± 3·40ND14·7 ± 4·66·3 ± 0·6VI
15C3Cold101018·1 ± 0·36·9 ± 0·35·7 ± 0·28·0 ± 0·48·4 ± 0·26·8 ± 0·20·0 ± 00·0 ± 00·2 ± 6·60·0 ± 8·70ND15·3 ± 6·77·7 ± 0·6VI
16C2Cold101018·5 ± 0·26·5 ± 0·25·2 ± 0·28·4 ± 0·28·3 ± 0·26·5 ± 0·30·0 ± 015·3 ± 0·30·0 ± 4·90·0 ± 8·00ND17·3 ± 3·24·3 ± 0·6VI
17C4Cold101018·1 ± 0·27·3 ± 0·55·8 ± 0·78·0 ± 0·37·8 ± 0·16·6 ± 0·40·0 ± 031·4 ± 0·68·3 ± 5·19·3 ± 7·90ND30·0 ± 7·99·7 ± 1·2VI
18C6Cold101017·9 ± 0·37·2 ± 0·25·8 ± 0·08·4 ± 0·27·9 ± 0·17·0 ± 0·3 94·4 ± 2·7 92·9 ± 1·9 92·3 ± 1·8 0·0 ± 00ND14·3 ± 2·35·3 ± 0·6VI
19C6Cold101018·4 ± 0·46·8 ± 0·26·0 ± 0·77·5 ± 0·88·0 ± 0·07·0 ± 0·1 96·3 ± 2·4 0·0 ± 024·5 ± 6·60·0 ± 00ND16·0 ± 3·65·3 ± 0·6VI
20C6Cold101018·4 ± 0·26·7 ± 0·55·9 ± 0·28·4 ± 0·08·2 ± 0·27·0 ± 0·40·0 ± 00·0 ± 00·0 ± 7·20·0 ± 00ND15·7 ± 4·05·7 ± 0·6VI
21C6Cold101018·8 ± 0·27·3 ± 0·15·9 ± 0·38·1 ± 0·38·5 ± 0·17·3 ± 0·3 97·8 ± 1·7 20·6 ± 0·47·2 ± 4·721·7 ± 2·10ND22·7 ± 12·25·7 ± 0·6VI
22C6Cold101018·5 ± 0·37·0 ± 0·45·8 ± 0·28·3 ± 0·38·1 ± 0·17·4 ± 0·1 97·7 ± 0·5 7·0 ± 0·119·3 ± 4·011·4 ± 8·10ND17·0 ± 4·65·3 ± 0·6VI
23C6Cold101018·5 ± 0·47·4 ± 0·45·7 ± 0·48·2 ± 0·38·0 ± 0·17·3 ± 0·2 54·9 ± 2·4 0·0 ± 013·5 ± 7·50·0 ± 00ND18·7 ± 5·15·3 ± 0·6VI
24C6Cold101018·3 ± 0·17·5 ± 0·55·5 ± 0·18·6 ± 0·78·2 ± 0·27·1 ± 0·20·0 ± 00·0 ± 00·0 ± 1·81·2 ± 5·00ND15·7 ± 4·06·3 ± 0·6VI
25C4Cold101018·1 ± 0·16·6 ± 0·45·6 ± 0·47·9 ± 0·58·0 ± 0·06·8 ± 0·521·1 ± 0·039·8 ± 0·80·0 ± 7·30·0 ± 00ND21·7 ± 2·16·7 ± 0·6VI
26C4Cold101018·1 ± 0·27·0 ± 0·36 ± 0·28·2 ± 0·18·3 ± 0·16·8 ± 0·10·0 ± 00·0 ± 05·1 ± 2·90·0 ± 00ND22·0 ± 3·55·7 ± 0·6VI
27C5Cold101018·4 ± 0·26·7 ± 0·26·3 ± 0·18·3 ± 0·17·1 ± 1·37·5 ± 0·61·1 ± 3·40·0 ± 04·5 ± 2·20·0 ± 00ND11·7 ± 0·616·3 ± 0·6VI
28C5Cold101018·0 ± 0·37·0 ± 0·44·4 ± 0·28·2 ± 0·17·8 ± 0·45·1 ± 0·10·0 ± 031·6 ± 0·60·0 ± 7·90·0 ± 00ND30·3 ± 10·77·0 ± 1·7VI
29C6Cold101018·4 ± 0·27·4 ± 0·55·9 ± 0·27·7 ± 0·48·2 ± 0·27·1 ± 0·426·4 ± 8·10·0 ± 012·8 ± 1·20·0 ± 00ND23·7 ± 0·66·3 ± 0·6VI
30C6Cold101018·1 ± 0·26·5 ± 0·15·8 ± 0·38·1 ± 0·18·1 ± 0·17·1 ± 0·20·0 ± 00·0 ± 022·2 ± 2·916·2 ± 5·40ND23·3 ± 1·25·0 ± 0·0VI
31C6Cold101017·9 ± 0·06·7 ± 0·46·0 ± 0·48·1 ± 0·48·1 ± 0·46·4 ± 0·342·2 ± 8·114·4 ± 0·322·3 ± 5·30·2 ± 11·20ND24·0 ± 0·07·7 ± 1·5VI
32C4Cold101018·1 ± 0·27·1 ± 0·35·7 ± 0·18·0 ± 0·38·3 ± 0·37·0 ± 0·30·0 ± 00·0 ± 04·5 ± 8·70·0 ± 00ND21·7 ± 4·05·3 ± 0·6VI
33C1Warm011018·3 ± 0·26·9 ± 0·43·1 ± 3·28·5 ± 0·28·0 ± 0·11·8 ± 0·919·9 ± 2·00·0 ± 0 97·3 ± 0·3 0·0 ± 8·30ND11·7 ± 0·69·3 ± 0·6II
34C6Warm011018·2 ± 0·26·8 ± 0·42·2 ± 1·58·2 ± 0·27·6 ± 0·42·1 ± 1·3 96·1 ± 0·8 0·0 ± 0 96·5 ± 0·5 5·0 ± 7·20ND11·7 ± 0·611·7 ± 0·6II
35C3Warm011018·2 ± 0·36·4 ± 0·22·4 ± 1·18·2 ± 0·37·8 ± 0·02·4 ± 1·026·6 ± 1·40·0 ± 032·9 ± 7·48·3 ± 2·00NDUCUCII
36C5Warm011018·1 ± 0·17·0 ± 0·23·1 ± 0·68·0 ± 0·78·0 ± 0·14·4 ± 0·3 94·4 ± 1·2 0·0 ± 0 98·0 ± 0·2 0·1 ± 2·40ND11·7 ± 0·69·7 ± 1·2II
37C3Cold011018·4 ± 0·37·5 ± 0·42·1 ± 1·48·4 ± 0·28·0 ± 0·52·0 ± 0·6 93·3 ± 4·3 0·0 ± 0 96·3 ± 0·4 0·0 ± 00ND11·7 ± 0·69·3 ± 0·6II
38C6Warm011118·0 ± 0·16·8 ± 0·3<1·38·5 ± 0·17·2 ± 0·31·8 ± 0·9 92·6 ± 1·5 0·0 ± 0 91·3 ± 9·9 0·0 ± 00ND11·7 ± 0·611·7 ± 0·6II
39C6Warm011117·8 ± 0·57·0 ± 0·32·3 ± 1·78·1 ± 0·17·8 ± 0·01·9 ± 0·9 95·7 ± 1·7 0·0 ± 0 95·3 ± 3·6 8·6 ± 6·20ND11·7 ± 0·611·7 ± 0·6II
40C6Warm011118·4 ± 0·27·0 ± 0·41·4 ± 0·28·4 ± 0·36·3 ± 1·72·4 ± 0·9 83·9 ± 4·8 0·0 ± 0 95·9 ± 1·8 0·9 ± 1·80ND14·3 ± 2·5 15·3 ± 1·2 II
41C3Warm011118·2 ± 0·27·1 ± 0·31·7 ± 0·78·2 ± 0·37·7 ± 0·23·2 ± 0·6 93·6 ± 4·1 0·0 ± 0 93·8 ± 1·6 5·8 ± 0·30NDUCUCII
42C5Warm011118·1 ± 0·27·1 ± 0·42·1 ± 1·48·1 ± 0·37·7 ± 0·21·6 ± 0·6 89·8 ± 5·1 0·0 ± 0 96·3 ± 1·0 1·9 ± 1·80ND11·7 ± 0·610·3 ± 1·5II
43C5Warm011117·8 ± 0·17·1 ± 0·32·3 ± 1·18·2 ± 0·17·7 ± 0·1<1·3 91·2 ± 5·6 0·0 ± 0 97·8 ± 0·6 1·0 ± 5·00ND11·7 ± 0·611·0 ± 1·0II
44C5Warm011118·1 ± 0·36·8 ± 0·42·4 ± 1·28·2 ± 0·37·7 ± 0·82·4 ± 1·0 94·0 ± 1·9 0·0 ± 0 93·9 ± 7·0 2·7 ± 4·60ND11·7 ± 0·611·0 ± 1·0II
45C5Warm011118·1 ± 0·26·7 ± 0·32·4 ± 1·28·4 ± 0·47·6 ± 0·21·7 ± 0·8 91·0 ± 3·3 3·8 ± 0·1 97·8 ± 0·9 3·0 ± 6·10ND11·7 ± 0·611·7 ± 1·5II
46C6Warm011118·2 ± 0·26·8 ± 0·52·2 ± 1·58·1 ± 0·47·7 ± 0·22·7 ± 1·2 86·3 ± 5·8 17·1 ± 0·3 96·7 ± 1·1 2·0 ± 1·80ND11·7 ± 0·612·7 ± 2·1II
47C6Warm011118·2 ± 0·47·1 ± 0·52·1 ± 0·78·0 ± 0·77·9 ± 0·23·0 ± 1·522·0 ± 2·916·6 ± 0·3 93·2 ± 1·7 5·9 ± 3·60NDUCUCII
48C2Cold011118·4 ± 0·27·6 ± 0·4<1·38·3 ± 0·27·3 ± 0·52·2 ± 0·8 71·4 ± 7·1 6·9 ± 0·1 96·3 ± 1·1 0·0 ± 00ND13·3 ± 1·512·7 ± 1·2II
49C1Cold011118·2 ± 0·27·2 ± 0·4<1·38·4 ± 0·17·5 ± 0·2<1·3 90·3 ± 6·0 0·0 ± 012·6 ± 2·50·0 ± 00ND12·0 ± 0·012·0 ± 1·0II
50C1Cold011118·5 ± 0·37·2 ± 0·21·9 ± 1·18·2 ± 0·27·4 ± 0·1<1·3 95·2 ± 1·4 0·0 ± 00·0 ± 6·07·0 ± 6·30ND14·7 ± 3·212·0 ± 1·0II
51C5Cold011118·5 ± 0·16·9 ± 0·42·1 ± 1·58·5 ± 0·28·1 ± 0·13·6 ± 0·438·5 ± 3·40·3 ± 0·0 95·1 ± 3·9 5·1 ± 8·90ND11·7 ± 0·69·7 ± 1·5II
52C5Cold011118·6 ± 0·25·9 ± 0·41·7 ± 0·78·4 ± 0·26·5 ± 0·21·7 ± 0·5 96·2 ± 4·0 7·5 ± 0·2 95·4 ± 0·6 4·1 ± 3·90ND11·7 ± 0·6 24·0 ± 0·0 VI
53C6Warm011108·4 ± 0·27·2 ± 0·11·9 ± 0·58·5 ± 0·27·6 ± 0·22·8 ± 0·3 94·0 ± 2·1 0·0 ± 0 96·5 ± 0·7 0·0 ± 00ND11·3 ± 0·611·0 ± 1·7IV
54C2Warm011108·3 ± 0·26·7 ± 0·41·7 ± 0·78·3 ± 0·07·2 ± 0·3<1·3 86·3 ± 1·5 0·0 ± 0 97·6 ± 0·5 0·0 ± 00ND11·7 ± 0·6 13·0 ± 1·0 II
55C2Warm011108·3 ± 0·16·3 ± 0·41·9 ± 1·08·4 ± 0·26·9 ± 0·31·8 ± 0·9 96·5 ± 0·4 0·0 ± 0 97·8 ± 0·3 7·0 ± 1·90ND11·7 ± 0·6 24·0 ± 0·0 IV
56C3Warm011108·4 ± 0·26·3 ± 0·32·6 ± 0·78·5 ± 0·27·3 ± 0·31·9 ± 1·0 69·9 ± 5·5 0·0 ± 0 97·1 ± 1·2 11·1 ± 1·60ND11·7 ± 0·6 24·0 ± 0·0 IV
57C3Warm011108·6 ± 0·26·2 ± 0·32·1 ± 0·78·5 ± 0·27·2 ± 0·22·9 ± 0·2 96·9 ± 0·5 0·0 ± 0 96·3 ± 0·6 10·7 ± 4·00ND11·7 ± 0·6 24·0 ± 0·0 IV
58C3Warm011108·0 ± 0·46·3 ± 0·21·7 ± 0·48·4 ± 0·26·0 ± 0·82·3 ± 0·9 97·1 ± 0·7 1·0 ± 0·0 97·9 ± 0·3 2·1 ± 6·20ND11·7 ± 0·6 18·7 ± 0·0 IV
59C2Warm011108·3 ± 0·17·0 ± 0·32·7 ± 1·58·4 ± 0·47·8 ± 0·14·0 ± 0·2 94·3 ± 2·5 15·8 ± 0·3 94·7 ± 2·4 14·6 ± 8·70ND11·7 ± 0·69·0 ± 0·0II
60C4Cold011108·6 ± 0·36·9 ± 0·31·6 ± 0·58·3 ± 0·17·6 ± 0·6<1·3 95·4 ± 1·4 0·0 ± 0 96·3 ± 1·3 0·6 ± 4·80ND11·7 ± 0·68·3 ± 0·0II
61C3Cold011108·2 ± 0·37·2 ± 0·51·7 ± 0·78·3 ± 0·17·8 ± 0·11·6 ± 0·6 67·7 ± 5·2 0·0 ± 011·7 ± 7·50·0 ± 00ND12·0 ± 0·024·0 ± 0·0III
62C3Cold011108·5 ± 0·06·7 ± 0·11·8 ± 0·58·3 ± 0·16·1 ± 0·3<1·3 97·8 ± 1·2 23·0 ± 0·520·7 ± 7·70·0 ± 00ND12·0 ± 0·0 15·0 ± 1·7 V
63C6Cold011108·2 ± 0·07·1 ± 0·41·6 ± 0·48·2 ± 0·37·5 ± 0·01·7 ± 0·8 97·0 ± 0·6 1·0 ± 0·0 96·7 ± 0·7 0·0 ± 00ND12·0 ± 0·012·3 ± 0·6II
64C6Warm011008·2 ± 0·26·6 ± 0·21·4 ± 0·28·1 ± 0·27·4 ± 0·2<1·3 95·2 ± 3·8 0·0 ± 0 91·2 ± 4·3 0·0 ± 00ND12·3 ± 0·610·3 ± 0·6II
65C3Warm011008·1 ± 0·06·9 ± 0·3<1·38·4 ± 0·27·1 ± 0·1<1·310·6 ± 4·40·0 ± 0 97·6 ± 0·4 10·0 ± 1·60ND11·7 ± 0·611·3 ± 0·6II
66C5Cold011008·3 ± 0·26·9 ± 0·3<1·38·3 ± 0·27·2 ± 0·1<1·30·0 ± 05·8 ± 0·122·9 ± 1·90·0 ± 00ND11·7 ± 0·615·5 ± 0·7II
67C2Warm101118·2 ± 0·46·7 ± 0·35·8 ± 0·18·0 ± 0·37·9 ± 0·37·1 ± 0·9 97·0 ± 0·3 0·0 ± 0 97·1 ± 0·8 16·3 ± 4·20ND20·3 ± 3·26·7 ± 1·2II
68C4Cold101118·1 ± 0·17·1 ± 0·56·2 ± 0·37·9 ± 0·67·8 ± 0·26·3 ± 0·80·0 ± 00·0 ± 00·0 ± 8·24·4 ± 4·90ND36·0 ± 0·09·3 ± 0·6II

Inoculum preparation

Before each experiment, frozen isolates (stored at −20°C in BHI (Scharlau, Barcelona; Spain)-YE (Merck, Darmstadt; Germany) with 5% (vol/vol) glycerol (Sigma–Aldrich, Saint Quentin Fallavier, France) were twice propagated in BHI-YE at 30°C for 18 h.

Growth at refrigerated temperature in optimum medium and in liquid whole egg

Each inoculum was transferred into liquid BHI-YE (optimum medium) or sterile liquid whole egg (prepared as described by Jan et al. 2011) at a concentration of 3 ± 0·5 log CFU ml−1. Incubation was carried out at 6°C for 14 days. Enumeration was made on BHI-YE agar (Merck) after 24 h of incubation at 30°C, using a plate-counting micro-method (Baron et al. 2006). Results were expressed as the means of three independent experiments.

Spoilage of liquid whole egg

Each inoculum was transferred into sterile liquid whole egg at a concentration of 3 ± 0·5 log CFU ml−1 and incubated either at 30 or at 10°C. Visual observations were made every hour at 30°C and everyday at 10°C until total visible spoilage, that is transformation of the sterile liquid whole egg product into a white slurry cream. Results were expressed as the onset of egg spoilage, namely the time (expressed in hours at 30°C and in days at 10°C) at which this particular state was detected. For each experiment, the absence of egg spoilage was checked in a negative control (sterile liquid whole egg incubated without previous inoculation). Results were expressed as the means of three independent experiments.

Ces gene amplification by PCR

Frozen isolates were inoculated on nutrient agar plates and incubated at 30°C for 24 h. Half a colony was transferred into an Eppendorf tube containing 200 μl sterile water. The tubes were placed for 15 min in a water bath at 100°C. The detection of the ces cluster was tested with two different primer pairs, under the conditions described by Ehling-Schulz et al. (2004, 2005). The primers EM1F (5′-GACAAGAGAAATTTCTACGAGCAAGTACAAT-3′) and EM1R (5′-GCAGCCTTCCAATTACTCCTTCTGCCACAGT-3′) amplified a fragment of 635 bp from emetic B. cereus DNA (Ehling-Schulz et al.2004).The primers cesF1 (5′-GGTGACACATTATCATATAAGGTG-3′) and cesR2 (5′-GTAAGCGAACCTGTCTGTAACAACA-3′) amplified a fragment of 1271 bp from emetic B. cereus DNA (Ehling-Schulz et al. 2005). The final PCR mixture contained 12·5 μl of Maxima Hot Start Green PCR Master Mix (Fisher Scientific, Erembodegem-Aalst, Belgium), 1 μl of 50 ng DNA template, 0·25 μl of 100 pmol of each primer from the set and finally adjusted to 25 μl with water. Thermal cycling was carried out in a PCR thermocycler (Veriti96-Well Thermal Cycler; Applied Biosystems, Gent, Belgium) with the following run: a starting cycle of 5 min at 95°C, followed by 30 cycles of 15 s at 95°C, 30 s at 58°C and 30 s at 72°C, and a final extension of 8 min at 72°C. Ten microlitres of the PCR products were separated on 1·5% agarose gel (100 V, 45 min) in TBE 1X (Sigma Aldrich, Diegem, Belgium) together with the DNA Smart Ladder SF molecular weight marker (Eurogentec, Seraing, Belgium). Gels were stained with ethidium bromide (Sigma Aldrich) and digitized using a gel imager (Proxima, Illkirch, France) under UV light.

Cereulide production in whole egg

Each inoculum was transferred into sterile liquid whole egg at a concentration of 2% (vol/vol). After incubation at 30°C for 18 h or 10°C for 5 days, a one-third dilution of the medium was performed in BHI-YE. Bacterial suspensions were centrifuged at 1400 g at 4°C for 10 min, and the collected supernatants were re-centrifuged (3200 g, 4°C, 10 min). One millilitre of homogenized supernatant was evaporated to dryness under nitrogen flow. The residue was dissolved in 1 ml methanol, vortexed and centrifuged in Eppendorf tubes at 16 000 g for 10 min. The clear (filtrated on a 0·2 μm filter) supernatant was analysed by LC-MS2 analysis, using a LC-MS LCQ Deca-XP Plus ion trap instrument (ThermoFinnigan, Austin, TX, USA) as previously described (Delbrassinne et al. 2012a). Briefly, assessment of cereulide was performed in MS2 mode by monitoring the following m/z values: 1170·5, 1153·3 and 1125·3.

Two positive controls were included in both experiments at 30 and 10°C: one mesophilic strain (TIAC 1095) isolated from a Belgian emetic food poisoning event and one psychrotolerant strain (MC118) reported as a cereulide producer at temperatures below 10°C (Thorsen et al. 2006). A negative control, consisting in sterile liquid whole egg processed as described above, was injected to verify the absence of matrix interferences.

panC gene sequence analysis

Three hundred microlitres of a 25% (m/v) sterile suspension of Chelex beads (Grosseron, Saint-Herblain, France) prepared in sterile Milli-Q water (Sigma Aldrich, Saint Quentin Fallavier, France) was added to 5 ml of culture medium after twice overnight propagation of each frozen isolate in BHI-YE at 30°C. The mixture was vortexed and centrifuged at 5700 g for 7 min at 4°C. The cell pellet was resuspended in 200 μl sterile Milli-Q water (Sigma Aldrich, Saint Quentin Fallavier, France) and lysed by heating at 100°C for 10 min. After centrifugation at 5700 g for 7 min at 4°C, 150 μl supernatant was collected and re-centrifuged under the same conditions. One hundred microlitres of supernatant was used for panC gene amplification. The final PCR mixture contained 50 ng DNA template, 0·18 mM dNTPs (Eurogentec), 0·7 μM of each panC primers (5′-TYGGTTTTGTYCCAACRATGG-3′ as forward degenerated primer and 5′-CATAATCTACAGTGCCTTTCG-3′ as reverse primer), 53·6 U ml−1 AmpliTaq polymerase and 0·9× AmpliTaq buffer (Biolabs, Evry, France). Thermal cycling was carried out in a PCR thermocycler (iCycler optical module 584BR; Bio-Rad, Marnes-la-Coquette, France) with the following run: a starting cycle of 5 min at 94°C, followed by 30 cycles of 15 s at 95°C, 30 s at 55°C and 30 s at 72°C, and a final extension of 7 min at 72°C. The PCR products were separated on 1% agarose gel (80V, 30 min) in TBE 1X (Eurobio, Courtaboeuf, France) together with the DNA Smart Ladder SF molecular weight marker (Eurogentec). Gels were stained with ethidium bromide (Sigma Aldrich, Saint Quentin Fallavier, France) and digitized using a gel imager (Bioblock, Illkirch, France). The panC amplified fragment was identified according to its apparent length (651 bp). Purification and sequencing were performed by GATC Biotech AG (Konstanz, Germany) using the following primer: 5′-ATAATCTACAGTGCCTTTCG-3′. Assignation into the phylogenetic groups defined by Guinebretière et al. (2008) was carried out using the online tool (https://www.tools.symprevius.org/Bcereus/) (Guinebretière et al. 2010).

Principal component analysis (PCA)

Data were analysed by the FactoMine r package of the R 2.13.0 software (Husson et al. 2008). PCA transforms a set of putative correlated variables into new variables, which are mutually orthogonal (uncorrelated) linear combinations of the original variables. These new variables are called principal components (PC). Each PC is defined by the coefficients in the linear combination of the original variables. For PCA, quantitative values were used for cytotoxic activity, cell concentration and onset of egg spoilage. Plots of the first two PC scores represented the best 2D representation of natural variance in the data (Fig. 1). The variables forming an acute angle were positively correlated, whereas those forming an obtuse angle were negatively correlated. The variables showing the best correlations with the axes (correlation coefficient superior to 50%) were those nearest of the border of the designed correlation circle and represented variables which contributed to the axes. Each isolate was represented by a point whose coordinates corresponded to the scores contributing to the PC (Fig. 2). The variables corresponding to the different genetic signatures, the growth ability on solid medium at 43 and 6°C, the season, the industrial origin of the isolates, and their affiliation to a phylogenetic group were considered as qualitative variables. A one-way analysis of variance was carried out with the coordinates of the individuals on the PC explained by these qualitative variables. For each modality of each qualitative variable, the individuals were considered as significantly different from the mean for student t-test P-values lower than 5%. A hierarchical clustering was performed on the PC (HCPC function of factomineR package), in order to identify subsets of objects, that is clusters having similar characteristics within the whole collection.

Figure 1.

Correlation circle of the quantitative variables characterizing the psychrotrophic Bacillus cereus group collection coming from pasteurized liquid whole egg, according to (i) growth ability (conc.) and cytotoxic activity (cytotoxic act.) after growth in BHI-YE and in liquid whole egg at 30, 10 and 6°C; and (ii) onset of egg spoilage at 30 and 10°C. The percentages of variation explained by the principal components (PC1 and PC2) are indicated between brackets. Variables drawn with black lines correlated for at least 50% with PC1 or PC2, whereas those drawn with grey lines showed correlations of <50% with PC.

Figure 2.

Factor map of each individual of the 68 psychrotrophic Bacillus cereus group collection coming from pasteurized liquid whole egg, according to their cytotoxic activity (representative of the sanitary risk) and their growth and egg spoilage potentials at temperatures inferior or equal to 10°C (representative of the spoilage risk). The isolates located inside a dotted circle belong to the same cluster, as determined by the hierarchical cluster analysis performed after PCA. For each cluster, two subgroups were highlighted, indicated either with a black symbol (main subgroup) or with a grey symbol (secondary subgroup). The percentages of variation explained by the principal components (PC1 and PC2) are indicated between brackets.

Results

Search for cereulide synthesis, affiliation to a phylogenetic group, and study of the spoilage potential as new parameters for characterizing the B. cereus group collection coming from refrigerated liquid whole egg products

Table 1 presents the set of data related to several food safety and food spoilage features of a collection of 68 psychrotrophic B. cereus group isolates coming from the French egg product industry. These data refer to results already described in a precedent publication (Jan et al. 2011) and to new results from the present work.

The evaluation of the spoilage potential of the collection was assessed at medium temperature (30°C, Jan et al. 2011), and at refrigerated temperatures, that is 10°C (Jan et al. 2011) and 6°C (this work), through cell enumeration and detection of visible egg spoilage. The whole collection was shown to be able to grow at 30 and 10°C in liquid whole egg (Jan et al. 2011). The present study shows that 50% of the collection (34 isolates) was also able to grow at 6°C in liquid whole egg, reaching the same final level of population at both refrigerated temperatures (around 7·8 log CFU ml−1 at 6 and 10°C, Table 1). Under these conditions, the average final cell concentration was higher in liquid whole egg than in optimum medium (around 7·8 vs 6·9 log CFU ml−1 in liquid whole egg and in BHI-YE, respectively). Ninety-six percents of the collection (65 isolates) was also shown to be able to spoil liquid whole egg at 30 and 10°C, with onsets of egg spoilage varying from 11·3 to 36 h and from 5 to 24 days at 30 and 10°C, respectively. Only three isolates (isolates Nr35, Nr41 and Nr47) did not spoil liquid whole egg, neither at 30°C nor at 10°C, even after 2 months incubation (results of this study in Table 1).

The evaluation of the toxigenic potential of the collection, already assessed through cytotoxic assays on Caco-2 cells (Jan et al. 2011), was further investigated by the search for cereulide production, at the genetic level (search for the ces gene), and through cereulide quantification by LC-MS2. As already shown, 58·8% of the collection was cytotoxic, that is, expressed a percentage of Caco-2 cell damage superior to 50%, after growth in liquid whole egg at 30°C (results of Jan et al. 2011 in Table 1). Moreover, the percentage of cytotoxic isolates was higher in liquid whole egg than in optimum medium (only 45·6%). At refrigerated temperature, only one isolate (1·5%, Nr 18) was shown to be cytotoxic at 10°C, and exclusively in the optimum medium (results of Jan et al. 2011 in Table 1). The present study shows that none of the 68 isolates neither harboured the ces gene nor produced detectable amounts of cereulide in liquid whole egg, irrespective of the temperature tested (results of this study in Table 1). The evaluation of the toxigenic potential was further investigated by affiliating each isolate to one of the phylogenetic groups defined by Guinebretière et al. (2008, 2010). As shown in Table 1, 33 isolates (48·5%), 28 isolates (41%) and 5 isolates (7·5%) were assigned to the phylogenetic groups VI, II and IV, respectively, whereas only one isolate (1·5%) was affiliated to the group III and one isolate (1·5%) to the group V (results of this study in Table 1). None of the isolates were found to be affiliated to the most thermophilic phylogenetic B. cereus groups I and VII described by Guinebretière et al. (2008).

The food safety and food spoilage risks are negatively correlated

A principal component analysis (PCA) was carried out by using the quantitative variables corresponding to the set of data presented in Table 1. Significant correlations were found between several variables (Table 2). As an illustration, a positive correlation was shown between the cell concentration at 6°C in optimum medium and in liquid whole egg (Table 2). Inversely, the cell concentration at 6°C in liquid whole egg was shown to be negatively correlated to the cytotoxic activity in the same medium at 30°C (Table 2).

Table 2. Correlation probability matrix of quantitative variables characterizing the psychrotrophic Bacillus cereus group collection, that is cell concentration (conc.) in BHI-YE and in liquid whole egg at 30, 10 and 6°C; cytotoxic activity (cytotoxic act.) in BHI-YE and in liquid whole egg at 30 and 10°C; and onset of egg spoilage at 30 and 10°C
 Conc. BHI-YE 30°CConc. BHI-YE 10°CConc. BHI-YE 6°CConc. egg 30°CConc. egg 10°CConc. egg 6°CCytotoxic act. BHI-YE 30°CCytotoxic act. BHI-YE 10°CCytotoxic act. egg 30°CCytotoxic act. egg 10°COnset of egg spoilage 30°COnset of egg spoilage 10°C
  1. Significant correlations (R > |0·5|) are highlighted in grey.

Conc. BHI-YE 30°C1           
Conc. BHI-YE 10°C−0·011          
Conc. BHI-YE 6°C−0·080·181         
Conc. egg 30°C0·19−0·19−0·451        
Conc. egg 10°C−0·090·410·7−0·341       
Conc. egg 6°C−0·020·180·97−0·40·681      
Cytotoxic act. BHI-YE 30°C0·19−0·11−0·510·11−0·37−0·471     
Cytotoxic act. BHI-YE 10°C−0·110·140·200·050·21−0·031    
Cytotoxic act. egg 30°C−0·01−0·24−0·690·3−0·45−0·660·65−0·071   
Cytotoxic act. egg 10°C0·04−0·10·26−0·130·20·270·03−0·04−0·081  
Onset of egg spoilage 30°C−0·180·160·65−0·490·420·61−0·510·16−0·61−0·231 
Onset of egg spoilage 10°C0·15−0·45−0·670·41−0·75−0·680·33−0·180·44−0·08−0·461

Two PC (PC1 and PC2) were shown to explain 52·07% of the variance of the whole data (Fig. 1). The PC1 allowed classifying the correlated variables into two types of negatively correlated risks, that is a spoilage risk vs a food safety risk for the variables showing a positive vs a negative correlation coefficient with the PC1, respectively (Table 3). The spoilage risk was globally characterized by high cell concentration at low temperatures and fast onsets of egg spoilage at 10°C, whereas the food safety risk was characterized by a high cytotoxic activity at 30°C.

Table 3. Coefficients of correlation of the most contributory variables in the principal component analysis; cell concentration (conc.), cytotoxic activity (cytotoxic act.), liquid whole egg (egg)
VariablesPC1PC2
RP valueRP value
  1. R, coefficient of correlation of each variable to PC1 and/or PC2.

Conc. BHI-YE 6°C0·921·40 × 10−28  
Conc. egg 6°C0·901·80 × 10−25  
Conc. egg 10°C0·782·72 × 10−15  
Onset of egg spoilage 30°C0·765·85 × 10−14  
Conc. egg 30°C−0·533·07 × 10−06  
Cytotoxic act. BHI-YE 30°C−0·614·38 × 10−08  
Cytotoxic act. egg 30°C−0·767·21 × 10−14  
Onset of egg spoilage 10°C−0·791·10 × 10−15  
Conc. BHI-YE 10°C  0·711·36 × 10−11
Cytotoxic act. egg 10°C  −0·533 × 10−6

As shown in Table 3, the second principal component (PC2) was strongly correlated with cell concentration in optimum medium at 10°C and with cytotoxic activity in liquid whole egg at 10°C (Table 2). As only one qualitative variable, that is the affiliation to a phylogenetic group, allowed discriminating several isolates of the collection according to this second axis, the first axis (PC1) was mainly considered for further analysis.

The collection is divided into two equal groups associated with each risk

The 68 isolates were projected as individuals on a factor map defined by the two PC (Fig. 1). A hierarchical cluster analysis function highlighted two clusters in the collection (surrounded in Fig. 2). The first cluster (cluster 1) was composed of 34 isolates associated with a high food safety risk at medium temperature and the second cluster (cluster 2) comprised 34 isolates associated with a high spoilage risk at refrigerated temperature. Each cluster was further divided into two subgroups. The cluster 1 was composed of the subclusters 1·1 and 1·2, containing 26 and 8 isolates, respectively. The isolates of the subcluster 1·2. (Nr40, Nr52, Nr54 to Nr58 and Nr62) mainly differed from the other members of the cluster 1 by delayed onsets of egg spoilage at 10°C (bolded in Table 1), in addition to a high pathogenic potential at 30°C. For these isolates, the time necessary to spoil egg was about 20 days vs an average value of 11 days for the isolates delimiting the cluster 1·1. In the cluster 2, two subclusters were also highlighted (subclusters 2·1 and 2·2). The isolate Nr18, delimiting a separate cluster by itself (subcluster 2·1), differed from the other members of the cluster 2 by its high cytotoxic activity in optimum medium at 10°C (bolded in Table 1).

Several genetic and phenotypic signatures allow discriminating both risks and groups

The collection was further analysed with respect to the qualitative variables described in Table 1, that is the genetic signatures, the growth ability at different temperatures on solid medium, the affiliation to a phylogenetic group and the season and origin of collection. For each qualitative variable significantly correlated to PC1, the estimated coordinates of each modality on PC1 are presented in Table 4, highlighting a significant correlation of the variable either to the spoilage risk (positive coordinates on PC1) or to the food safety risk (negative coordinates on PC1).

Table 4. Coordinates of each modality of the qualitative variables significantly correlated to PC1 (presence/absence of the cspA signature; presence/absence of the 16S rDNA-1m signature; ability/inability to grow at 43 and 6°C on BHI-YE agar, affiliation to one of the phylogenetic groups and origin (cold/warm season, egg breaking company (C1 to C6))
VariableModalityNumber of isolatesEstimated coordinateP value*Risk
  1. *Student t-test on the coordinates of each modality of the qualitative variables on the PC1. Significance was defined as a P value below 0·05, **according to Guinebretière et al. (2008,2010).

cspA Presence342·055·42 × 10−29Spoilage
Absence34−2·055·42 × 10−29Food safety
16S rDNA-1mPresence34−2·055·42 × 10−29Food safety
Absence342·055·42 × 10−29Spoilage
Growth at 6°CAbility541·621·46 × 10−07Spoilage
Inability14−1·621·46 × 10−07Food safety
Growth at 43°CAbility28−1·627·39 × 10−12Food safety
Inability401·627·39 × 10−12Spoilage
Phylogenetic group**II28   
III1   
IV5−1·822·51 × 10−03Food safety
V1   
VI333·416·69 × 10−08Spoilage
SeasonWarm37−0·610·023Food safety
Cold310·610·023Spoilage
OriginC14   
C28   
C317   
C461·950·018Spoilage
C512   
C621   

The genetic signatures, that is the cspA and 16S rDNA-1m signatures, clearly discriminated both risks and, consequently, the two clusters highlighted in Fig. 2. Namely, the presence of the cspA signature was exclusively associated with the spoilage risk at refrigerated temperatures (positive coordinate for the modality ‘presence’ in Table 4) and also with the cluster 2 presented in Fig. 2 (< 10−4). At the opposite, the presence of the 16S rDNA-1m signature was exclusively associated with the food safety risk at medium temperature (negative coordinate for the modality ‘presence’ in Table 4) and also with the cluster 1 presented in Fig. 2 (< 10−4). To a lower extent, such discrimination was also observed for the other qualitative variables. The ability to grow at 6 and at 43°C on solid medium was, respectively, associated with the spoilage risk (positive coordinate for the modality ‘ability to grow at 6°C’) and with the food safety risk (negative coordinate for the modality ‘ability to grow at 43°C’).

One phylogenetic group is associated with each risk

The phylogenetic group VI, which is the main group represented within our collection (48·5%), was significantly associated with the spoilage risk at refrigerated temperature (positive coordinate for the modality ‘group VI’ in Table 4). In contrast, the phylogenetic group IV, even if representing only 7·5% of the collection, was significantly associated with the food safety risk at medium temperature (negative coordinate for the modality ‘group IV’ in Table 4). The isolates belonging to the group II were not significantly associated with none of the risks. Finally, the number of isolates affiliated to the groups III and V was too low for the establishment of a significant relationship between their phylogenetic affiliation and their involvement in food safety issues under our experimental conditions.

Several genotypic and phenotypic criteria are reassessed

Several discordances were highlighted between our results and the genotypic and phenotypic criteria described by Guinebretière et al. (2008). The first discordance concerns the exclusive and required allocation of the cspA signature to the group VI. Namely, two isolates of our collection (isolates Nr67 and Nr68), which possessed the cspA signature, were assigned to the phylogenetic group II. Moreover, one isolate (isolate Nr52) which lacked the cspA signature was assigned to the phylogenetic group VI. This latter was able to grow at 6 and 43°C and was mainly associated with the food safety risk (Table 1 and Fig. 2). Discordance was also highlighted regarding the temperature growth ranges described by Guinebretière et al. (2008). Namely, the isolate Nr61, ascribed to the group III, was able to grow at 10°C under our experimental conditions, while this group was described as being able to grow exclusively at temperatures in the range of 15–45°C (Guinebretière et al. 2008). Moreover, out of 28 isolates belonging to the group II, 18 (64·2%) were able to grow at 43°C, that is 3°C higher than the upper temperature limit described by these authors for this group. Such discordances could be ascribed to the variability of the experimental conditions used to characterize the strains or to the probable high diversity of the strains coming from egg products (Jan et al. 2011).

The risk depends on the origin of the isolates

The isolates collected at the cold season were significantly associated with the spoilage risk at refrigerated temperature (positive coordinate for the modality ‘cold season’ in Table 4). Inversely, the isolates collected at the warm season were significantly associated with the food safety risk at medium temperature (negative coordinate for the modality ‘warm season’ in Table 4). Otherwise, the isolates collected in the egg product company 4 (C4) were significantly associated with a higher spoilage risk than the isolates collected in the other companies (positive coordinate for the modality ‘company C4’).

Discussion

This study offers a global analysis of the toxinogenic and spoilage potentials of a representative collection of psychrotrophic B. cereus group isolates originating from the French egg product industry. As a starting point for risk assessment, liquid whole egg was confirmed as representing an optimum medium, allowing the growth of these bacteria at concentrations above 7 log CFU ml−1, even at refrigerated temperatures (Table 1). According to Langeveld et al. (1996), this bacterial concentration represents the threshold value leading to spoilage and food safety issues in milk. It appears hence imperative to assess more deeply the risk linked to the growth of these bacteria in the egg products, regarding both safety and economical issues for the sector.

One of the major contributions of our global analysis is the demonstration of an inverse relationship between the food safety and the spoilage risks, at the individual level. To our knowledge, only few works have investigated such correlation at the strain level. De Jonghe et al. (2010) have reported the absence of cytotoxic activity of raw milk isolates exhibiting a high spoilage potential. Sutherland (1993) noticed that the spoilage of flavoured desserts was not always obvious at 21°C before significant growth of B. cereus group bacteria, while toxin production was systematically observed. According to the present study, the isolates appearing as the most pathogenic at medium temperature appear as the least spoiling at refrigerated temperature (Fig. 2). Therefore, our work refutes the generally recognized idea based upon the link between visible spoilage and food safety issues, especially at low temperature.

Considering the food safety issue, it has been suggested that the toxin gene expression is tightly regulated by various internal and environmental signals, including temperature (Ceuppens et al. 2011). The virulence of psychrotolerant B. weihenstephanensis species has been shown to be at least partially determined by temperature: at 15°C, this species was as virulent and expressed the same cytotoxic activity as B. cereus mesophilic strains (Stenfors Arnesen et al. 2011). Even if our collection may be considered as psychrotolerant, regarding its ability to grow at 10°C (Table 1), its harbouring of the psychrotrophic 16S rDNA-2p signature (Table 1) defined by Von Stetten et al. (1998) and its substantial belonging to the psychrotrophic groups VI and II described by Guinebretière et al. (2008) (Table 1), the isolates representing the highest risk correspond to the intermediate thermal types described by Von Stetten et al. (1999). These latter systematically possess the mesophilic 16S rDNA-1m signature and are significantly (i) unable to grow at 6°C and (ii) able to grow at 43°C (Table 4). Our statistical analysis also allowed significantly ascribing the pathogenicity of our collection to the phylogenetic group IV, one of the main mesophilic groups represented in our collection (Table 4) among the groups described by Guinebretière et al. (2008).

Apart from Stenfors and Granum (2001), who showed a lack of correlation between pathogenicity and the temperature growth range, our work confirms the assumption of other groups describing the mesophilic B. cereus bacteria as more pathogenic than the psychrotrophic ones (Pruss et al. 1999; Choma et al. 2000; Borge et al. 2001; Stenfors et al. 2002). Psychrotrophic B. cereus group strains are generally recognized as representing a low emetic risk in refrigerated foods (Carlin et al. 2006; Svensson et al. 2007). However, Thorsen et al. (2006) have highlighted two B. weihenstephanensis strains that could produce the emetic toxin at temperatures as low as 8°C. The consumption of chilled food products of extended durability, contaminated with emetic B. weihenstephanensis strains, could confer risk of emetic food poisoning (Thorsen et al. 2009). It appeared hence crucial to assess this risk in the sector of refrigerated liquid egg product manufacturing. Under our experimental conditions, a cereulide production was induced by positive controls at 30 and 10°C, pointing to a potential sanitary risk in case of contamination of the egg products with emetic B. cereus strains. However, none emetic strain has been isolated from the analysed liquid pasteurized egg products, which nuances the actual risk, and confirmed the low prevalence of emetic strains in food (Wijnands et al. 2006; Svensson et al. 2007; Hoton et al. 2009; Delbrassinne et al. 2012b).

Supported by a robust statistical analysis of a comprehensive collection of psychrotrophic B. cereus group isolates, our study represents the confirmation of previous assumptions, which had however never been questioned in the sector of egg product manufacturing. The psychrotrophic bacteria of the B. cereus group are revealed as representing a low sanitary risk. However, the prevalence of even small numbers of isolates representing a high cytotoxic risk should not be omitted. We hence recommend the egg product manufacturers to strictly respect a temperature of 4°C all along liquid egg product manufacturing, storage and delivery. The 16S rDNA-1m signature seems to be the most reliable marker of the cytotoxic risk, knowing that this genetic character (i) is exclusively associated with the food safety risk and (ii) is more exclusive than the other markers already described (Guinebretière et al. 2008, 2010; Cadot et al. 2010). In the case of liquid egg products constituting particularly sensitive foods, such as those intended to mass catering and/or to immuno-suppressed people, our findings could give the producers the possibility to develop a diagnostic tool ensuring the absence of pathogenic strains, through the control of the absence of the 16S rDNA-1m signature in their products, rather than detecting the B. cereus group itself. This diagnostic tool could represent a time- and cost-effective methodology, undoubtedly more accessible than the search for the markers described by Cadot et al. (2010), where gene expression analysis is carried out by reverse transcription real-time PCR, or by Guinebretière et al. (2008), where gene sequencing is required. Pending the development of this diagnostic tool, the search for growth ability at 43°C on solid medium, a variable that was also shown to be significantly associated with the food safety risk, would represent the minor additional cost of purchasing and operating an additional incubator in the laboratory.

Considering the food spoilage issue, the isolates representing the highest risk are those that may be considered as belonging to the pure psychrotrophic thermal type described by Von Stetten et al. (1999), that is systematically possessing the psychrotrophic 16S rDNA-2p and cspA signatures, and without the mesophilic 16S rDNA-1m signature (Table 4). These strains are significantly (i) able to grow at 6°C and (ii) unable to grow at 43°C (Table 4). According to Francis et al. (1998), they belong to the B. weihenstephanensis species. Our statistical analysis also significantly ascribed this population to the phylogenetic group VI (Table 4), described by Guinebretière et al. (2008) as the most strongly marked by the psychrotrophic character within the B. cereus group. As expected, a significant correlation was highlighted between the ability to spoil liquid whole egg and to grow at low temperature (Table 2). One could assume that the adaptation of B. cereus group strains to a range of temperatures concerns both growth and production of spoilage enzymes. However, this assumption does not match with the behaviour of three isolates of the collection (isolates Nr35, Nr41 and Nr47) which did not induce egg spoilage, whatever the time and temperature of incubation, while exhibiting the same growth potential as the spoilage isolates. The growth ability in the food is hence not sufficient to evaluate the spoilage potential of B. cereus group strains. Concerning the type of enzymatic activities, whole egg is rich in protein and lipid substrates. It is then reasonable to envisage that visible egg spoilage was due to the production of B. cereus group extracellular hydrolytic enzymes, such as proteases, lipases and lecithinases, already known to cause food spoilage even at low bacterial counts (De Jonghe et al. 2010). Further experiments are needed for exploring the enzymatic activities responsible for the visible spoilage detected under our experimental conditions.

Also supported by this robust statistical analysis of a comprehensive collection of psychrotrophic B. cereus group isolates, our study clearly demonstrates that psychrotrophic bacteria of the B. cereus group represent a high spoilage risk in the sector of production of refrigerated liquid egg products. Our work could help the producers of refrigerated pasteurized food products to control food spoilage, by providing the first step in the development of a molecular tool devoted to the search for the cspA signature as an exclusive marker. These developments could allow predicting the shelf-life of the food products and to adjust their use (i.e. low- or uncooked vs cooked foods) depending on their microbiological quality.

Finally, our global analysis represents a useful tool for determining the effect of environmental factors on the type of risk. The isolates of the warm season represented a higher food safety risk than the isolates collected during the cold season (Table 4), in accordance with a report showing a 37% increase in collective food poisoning events due to the B. cereus group during the summer 2009 in France, compared with the preceding winter (Anonymous 2011). Inversely, a higher spoilage risk was associated with the cold season (Table 4), confirming the work of Marchand et al. (2009) regarding the involvement of psychrotrophic Pseudomonas species in milk spoilage. The egg product company 4 (C4) was distinguishable from the others, due to the fact that it essentially hosted the species B. weihenstephanensis, already shown to be associated with a high spoilage risk and a low food safety risk (Table 4). This specificity, as well as the weak strain diversity observed in this company (Jan et al. 2011), might imply specific procedures, in terms of egg supplying, weighing of transformed eggs and/or hygiene practices.

In conclusion, this study shows that psychrotrophic B. cereus group bacteria mainly present a spoilage risk in the egg product industry by their ability to grow and to induce egg product spoilage at low temperature. At the opposite, they mainly exhibit diarrhoeal toxin production at optimal temperature (30°C) and do not present any emetic risk if a storage temperature of 4°C is strictly respected.

Considering the need for a better control of the B. cereus group in the sector of refrigerated pasteurized food products, this study provides new insights for the development of relevant real-time PCR diagnostic tools, targeting either a pathogenic risk or a spoilage risk, depending on which type of food they compose. Further research will be devoted into the study of the expression of spoilage enzymes at refrigerated temperatures, which may further help the understanding and the control of this major source of economical losses for the sector.

Acknowledgements

The authors would like to thank Jacques Lhermitte for helpful technical assistance. Prof. Jacques Mahillon is also acknowledged for providing the psychrotolerant B. cereus MC118 isolate.

Conflict of Interest

The authors have no conflicting financial interests.

Ancillary