The University of New South Wales School of Chemical Engineering and Industrial Chemistry Food Science and Technology
ECOLOGY OF TOXIGENIC BACILLUS SPECIES IN RICE PRODUCTS
By
Mi Hwa Oh
BSc, MSc, Duksung Women’s University (Korea)
A thesis submitted to The University of New South Wales as fulfillment of the requirements for the degree of
DOCTOR OF PHILOSOPHY
G February, 2006 DECLARATION
I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, nor material which to a substantial extent has been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis.
I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.
______
Mi Hwa Oh TABLE OF CONTENTS
TABLE OF CONTENTS i ACKNOWLEDGEMENTS vii LIST OF PRESENTATIONS AND PAPERS ix ABSTRACT x LIST OF TABLES xii LIST OF FIGURES xiv
CHAPTER 1 INTRODUCTION 1
CHAPTER 2 LITERATURE REVIEW 2.1 Taxonomy of the genus Bacillus 7 2.2 Ecology of Bacillus species 10 2.3 Control measures 11 2.4 Detection, isolation and biochemical properties of Bacillus species 12 2.5 Centrifugation-plating 18 2.5.1 Enzymes and surfactant 18 2.5.2 Application of centrifugation-plating method 19 2.6 Methods for strains typing of Bacillus species 20 2.6.1 Restriction fragment length polymorphism (RFLP) and pulsed-field gel electrophoresis (PFGE) 22 2.6.2 Ribotyping 22 2.6.3 Randomly amplified polymorphic DNA (RAPD) 23 2.6.4 REP-PCR 24 2.7 Food poisoning associated with Bacillus species 25 2.7.1 Food poisoning associated with B. cereus 25 2.7.2 Food poisoning associated with other Bacillus species 26 2.8 Toxins of Bacillus cereus associated with food poisoning 28 2.8.1 B. cereus diarrhoeal toxin 28 2.8.1.1 Characteristics and properties of diarrhoeal toxin 28 2.8.1.2 Detection methods for diarrhoeal toxin 30 2.8.2 B. cereus Emetic toxin 31 2.8.2.1 Chemical structure and properties of emetic toxin, cereulide 31
G 2.8.2.2 Detection methods for cereulide 33 2.8.2.3 Limitations of current bioassays for cereulide and the future 36 2.8.3 Toxins of other Bacillus species 37 2.9 Capillary electrophoresis 39 2.9.1 Theory and advantages of capillary electrophoresis (CE) 39 2.9.2 Application of capillary electrophoresis (CE) 41
CHAPTER 3 DEVELOPMENT AND APPLICATION OF CENTRIFUGATION- PLATING METHOD FOR STUDY OF THE BIODIVERSITY OF BACILLUS SPECIES IN RICE PRODUCTS
3.1 Introduction 43
3.2 Materials and methods 45 3.2.1 Viability studies 45 3.2.1.1 Preparation of Bacillus cultures 45 3.2.1.2 Preparation of enzymes and surfactant 45 3.2.1.3 Effect of enzyme and surfactant on the viability of Bacillus species 46 3.2.2 Recovery of inoculated B. cereus by centrifugation-plating 47 3.2.2.1 Preparation of rice samples 47 3.2.2.2 Recovery of B. cereus inoculated into raw rice samples 48 3.2.2.3 Recovery of B. cereus inoculated into raw rice homogenate fluids 48 3.2.2.4 Recovery of Bacillus species in cooked rice samples 49 3.2.3 Application of centrifugation-plating method 50 3.2.3.1 Rice samples 50 3.2.3.2 Detection and enumeration of Bacillus species on raw rice 50 3.2.3.3 Detection and enumeration of Bacillus species in cooked rice 51 3.2.4 Spread plating method for detection of Bacillus in foods 51 3.2.5 Identification of Bacillus species by partial sequencing 52 3.2.5.1 Phenol-chloroform extraction of DNA 52 3.2.5.2 PCR amplification for partial sequencing of 16S rDNA 53 3.2.5.3 Sequence analysis 54
3.3 Results and discussion 54 3.3.1 Viability studies of Bacillus species in model food system (medium) 54 3.3.1.1 Effect of dilution medium 54 3.3.1.2 Effect of amylase and surfactant on the viability of Bacillus species 55
G 3.3.2 Recovery of inoculated B. cereus by centrifugation-plating 57 3.3.2.1 Recovery of inoculated B. cereus from raw rice 57 3.3.2.2 Recovery of inoculated B. cereus from cooked rice treated with amylase and surfactant Tween 80 59 3.3.3 Application of the centrifugation-plating method for examination of Bacillus species in rice products 62 3.3.3.1 Detection and enumeration of Bacillus on raw and parboiled rice products 62 3.3.3.2 Detection and enumeration of Bacillus on cooked rice products 70
3.4 Conclusions 78
CHAPTER 4 DEVELOPMENT AND APPLICATION OF MICELLAR ELECTROKINETIC CHROMATOGRAPHY TO ANALYZE BACILLUS CEREUS EMETIC TOXIN, CEREULIDE
4.1 Introduction 81
4.2 Materials and methods 82 4.2.1 Bacillus strains 82 4.2.2 Maintenance of cultures 82 4.2.3 Purification and analysis of B. cereus emetic toxin 84 4.2.3.1 Preparation of B. cereus emetic toxin 84 4.2.3.2 Purification of cereulide 84 4.2.3.3 RP-HPLC analysis 85 4.2.3.4 NMR analysis 86 4.2.4 Analysis of cereulide by cell cytotoxicity (MTT) assay 86 4.2.4.1 Growth of cells and media 86 4.2.4.2 Preparation of cereulide 86 4.2.4.3 MTT cell cytotoxicity assay 87 4.2.4.3.1 Chemicals 87 4.2.4.3.2 MTT assay procedure 87 4.2.5 Development and application of capillary electrophoresis (CE) 88 4.2.5.1 Chemicals 88 4.2.5.2 Sample preparation 89 4.2.5.2.1 Valinomycin and purified cereulide 89
G 4.2.5.2.2 Semi-purified samples 89 4.2.5.2.3 Rice samples 89 4.2.5.3 Development of MEKC procedure for cereulide 90 4.2.5.3.1 Equipment 90 4.2.5.3.2 MEKC procedures 90
4.3 Results and discussion 91 4.3.1 Purification and analysis of B. cereus emetic toxin for development of CE 91 4.3.1.1 Purification and analysis of B. cereus emetic toxin 91 4.3.1.2 NMR analysis 93 4.3.2 Determination of biological activity and concentration of cereulide by colorimetric MTT assay 95 4.3.3 Development and application of MEKC analysis 95 4.3.3.1 Method development 95 4.3.3.2 Application of MEKC method 101 4.3.3.2.1 Reproducibility and calibration curve 102 4.3.3.2.2 Semi-purified sample analysis 103 4.3.3.2.3 Rice sample analysis 109
4.4 Conclusions 111
CHAPTER 5 DIVERSITY AND, POTENTIAL AND ACTUAL TOXIGENICITY OF BACILLUS STRAINS IN RICE PRODUCTS
5.1 Introduction 115
5.2 Materials and methods 117 5.2.1 Typing of isolated Bacillus strains by RAPD-PCR 117 5.2.1.1 DNA preparation (crude DNA extract) 117 5.2.1.2 RAPD-PCR primers 117 5.2.1.3 Conditions of RAPD-PCR and gel electrophoresis 118 5.2.1.4 Gel analysis (data analysis) 119 5.2.2 Identification and confirmation of Bacillus species 119 5.2.2.1 Identification of Bacillus species by partial sequencing and B. cereus gyrase B gene (gyrB) analyses 119
G 5.2.2.1.1 Identification of Bacillus species by partial sequencing of 16S 119 5.2.2.1.2 PCR detection of gyrB) 119 5.2.2.2 Identification and confirmation of Bacillus species with the API 50CH 120 5.2.3 PCR analysis for detection of enterotoxin and emetic toxin gene 121 5.2.3.1 PCR conditions for detection of enterotoxin genes 121 5.2.3.3 PCR condition for detection of emetic toxin gene 122 5.2.4 Enzyme immunoassay for detection of enterotoxin and emetic toxin 124 5.2.4.1 Detection of diarrhoeal enterotoxin 124 5.2.4.2 Detection of emetic toxin 124 5.2.4.3 Reading and interpretation of results 125 5.2.5 Analysis of emetic toxin by micellar electrokinetic electrophoresis (MEKC) 125
5.3 Results and discussion 126 5.3.1 Typing of Bacillus isolated by RAPD-PCR 126 5.3.1.1 Optimisation of RAPD-PCR analysis 126 5.3.1.2 Diversity of Bacillus isolates 127 5.3.1.2.1 B. cereus, B. thuringiensis and B. mycoides 128 5.3.1.2.2 B. subtilis/B. mojavensis and B. amyloliquefaciens 137 5.3.1.2.3 B. licheniformis 143 5.3.1.2.4 B. pumilus 145 5.3.1.2.5 B. megaterium 147 5.3.1.2.6 B. sphaericus/B. fusiformis 147 5.3.1.2.7 Paenibacillus strains 148 5.3.2 Toxigenicity of Bacillus isolates 150 5.3.2.1 B. cereus group 150 5.3.2.2 B. subtilis group 154 5.3.2.3 Other Bacillus species and Paenibacillus species 160
5.4 Conclusions 166
CHAPTER 6 CONCLUSIONS AND RECOMMENDATIONS 6.1 Conclusions 169 6.2 Recommendations 175
G CHAPTER 7 BIBLIOGRAPHY 177
APPENDICES Appendix 1 209 Appendix 2 210 Appendix 3 211
G ACKNOWLEDGEMENT
I must say that I have gone through many life-changing experiences during the years I spent for this research. It has been few tough years but there are people close to me who have supported me through those years and I would like to use this opportunity to thank each and every one of them.
First of all and most important, I would like to give my sincerest thanks to my supervisor, Professor Julian Cox. His enthusiasm and guidance has been an inspiration for my studies not to mention his help and encouragement. I sincerely do feel that I am so fortunate to meet someone who has been a great friend as well as being an excellent supervisor. Second, I would like to express my sincerest thanks to Professor Ken Buckle who has given me tremendous help and sound encouragement for the completion of my thesis during the absence of my supervisor. Third, I would like to give my sincerest respect and thanks to Professor Graham Fleet. His down to earth attitude towards studies has become a role model for me that I hope will shape my future endeavors as an academic.
I would like to give special thanks to Professor Young Tae Ko, Professor Gun Hee Kim for being my mentors during these years and also to Dr Jayashree Arcot, Dr Gillian
Heard and others who have shown special interests in my studies.
To the technical and administration support staff, who have been invaluable help during my experiments, Camillo, Eileen, Yvone and Zbigniew, I thank you all for your great support.
G Again, and definitely not the least important, I would like to share this moment of joy with my friends, Melissa, Sung-Sik, Sung-Sook, Rag, Pat, Ai Lin, Peter, Maria, Hugh and Lidia. As we have toiled together for our own dreams, I hope for their success and fulfillment.
I must thank my family, my mum and dad, James, Michael, Jung Mi, Liz and Seung
Hyun, Jin Suk. Without their love, support and encouragement, this thesis would not have come into being. I would like to say that I love them dearly, and I thank God above all and I hope my work will bring glory to his name.
Lastly, I would like to acknowledge the support from UNSW, who bestowed on me a university postgraduate research scholarship (UPRS), and the Australian Research
Council (ARC) and Tecra International who provided funds and materials for this project.
Sunrise in Sydney is most beautiful.
MiHwa Oh.
G LIST OF PRESENTATIONS AND PAPERS
Mi-Hwa Oh, Julian M. Cox. (Accepted) Quantitative analysis of the Bacillus cereus emetic toxin, cereulide, by micellar electrokinetic chromatography. Journal of
Chromatography B.
Toh, M., Oh, M.H., Neilan, B.A., Moffitt, M.C., Raftery, M., Marquis, C.P., Henrichsen,
L. and Cox, J.M. (Submitted) The Production of the Emetic Toxin Cereulide, by Other
Bacillus species. Applied and Environmental Microbiology.
Mi-Hwa Oh, Julian M. Cox. (2003) “Analysis of the Bacillus cereus emetic toxin, cereulide, by micellar electrokinetic chromatography”. Poster presented at 11th
Australian Food Microbiology Conference, Noosaville, Queensland, Australia, March
26-28, 2003
Mi-Hwa Oh, Julian M. Cox. (2003) “Analysis of the Bacillus cereus emetic toxin, cereulide, by micellar electrokinetic chromatography”. Poster presented at the Rapid
Methods and Automation in Microbiology Workshop, Kansas State University,
Manhattan, Kansas, USA, July 13-20, 2003.
Mi-Hwa Oh, Julian M. Cox. (2004) “Quantitative analysis of the Bacillus cereus emetic toxin, cereulide, by micellar electrokinetic chromatography”. Poster presented at Annual
Meeting and International Symposium by The Korean Society of Food Science and
Nutrition, Jeju Island, Korea, November 17-19, 2004.
G ABSTRACT
Bacillus cereus is the most prevalent pathogenic Bacillus species found in foods, causing food spoilage and two types of toxin-mediated food poisoning known as the diarrhoeal and emetic syndromes. Other Bacillus species, particularly B. subtilis, B. licheniformis, B. brevis, B. pumilus and B. thuringensis, have also been recognised as food poisoning bacteria of increasing concern, with reports of outbreaks of diarrhoeal or emetic food poisoning.
This study involved a systematic ecological investigation of Bacillus species isolated from rice products, commonly associated with Bacillus emetic food poisoning, using cultural and molecular methods. A centrifugation-plating method, more sensitive than the conventional spread plating method, was developed and used to determine the occurrence and biodiversity of Bacillus species in rice, a well known source of B. cereus.
Eight different Bacillus species, B. cereus/B. thuringiensis, B. mycoides, B. subtilis/B. mojavensis, B. licheniformis, B. pumilus, B. sphaericus/B. fusiformis and B. megaterium, as well as Paenibacillus species, identified by partial rDNA sequencing, were isolated from raw (uncooked) and cooked rice products.
The diversity of the isolates at the subspecies (strain) level was investigated using the
RAPD-PCR typing technique, which proved to be useful for differentiating strains of bacilli, revealing broad diversity among the strains. Generally, different genotypes were found in raw and cooked rice, with some isolates of the same RAPD pattern found in both raw and cooked rice.
G The toxigenic potential of Bacillus isolates were also determined by molecular and immunological analysis as well as an MEKC method, developed in this study for quantitative analysis of the emetic toxin, cereulide. The results revealed that most isolates from the B. cereus group were potentially or actually toxigenic and some isolates were able to produce both diarrhoeal and emetic toxins. Other Bacillus species outside the B. cereus group were also shown to produce cereulide.
G LIST OF TABLES
2.1 Number of species assigned to the genus Bacillus in different editions of Bergey’s Manual up to 1974 8 2.2 Bacillus type strains 9 2.3 Incidences of B. cereus in raw and cooked rice 11 2.4 Selective and diagnostic system of B. cereus isolation media 15 2.5 Identification of Bacillus species of public health significance 17 2.6 Overview of genetic fingerprinting techniques for typing of Bacillus species 21 2.7 Enterotoxins produced by B. cereus 28 3.1 Source and characteristics of amylases 46 3.2 Preparation of treatments for viability studies of Bacillus species 46 3.3 Effect of amylase and Tween 80 treatments on the viability of Bacillus cells as determined by plating onto TSA and PEMBA 56 3.4 Recovery rate of the centrifugation-plating method to the enumeration of Bacillus cereus inoculated onto the surface of raw rice samples 58 3.5 Recovery rate of the centrifugation plating method to the enumeration of Bacillus cereus inoculated onto cooked rice samples 61 3.6 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus on raw rice samples 66 3.7 Number (proportions, %) of raw rice with different populations of Bacillus species 72 3.8 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus in cooked rice samples 74 3.9 Number (proportions, %) of cooked rice with different populations of Bacillus species 77 4.1 Bacillus strains used in this study 83 4.2 Screening results for semi-purified emetic toxin from Bacillus spp. using CE 106 5.1 Primers used in this study 118 5.2 Primers used for detection of diarrhoeal enterotoxins and emetic toxin genes 123 5.3 Identification results and the RAPD-types of identified B. cereus, B. thuringiensis and B. mycoides strains isolated from rice 134 5.4 Identification results and the RAPD-types of identified B. subtilis/B. mojavensis strains isolated from rice 140 5.5 Identification results and the RAPD-types of identified B. licheniformis strains isolated from rice 144 5.6 Identification results and the RAPD-types of identified B. pumilus
G strains isolated from rice 146 5.7 Identification results and the RAPD-types of identified strains of other bacilli isolated from rice 149 5.8 Toxigenicity of strains in B. cereus group isolated from rice 155 5.9 Toxigenicity of Bacillus species isolated from rice 162 5.10 Proportions of potential and actual toxigenic Bacillus species 166
Appendix 1 Samples of raw rice examined for Bacillus species by the centrifugation-plating method and the spread plating method 208
Appendix 2 Samples of cooked rice examined for Bacillus species by the centrifugation- plating method and the spread plating method 209
G LIST OF FIGURES
2.1 Isolation and identification of B. cereus 16 2.2 Structure of cereulide and valinomycin 31 2.3 Electron micrographs of HEp-2 cells 33 2.4 Thin sections of the middle piece of a boar spermatozoon exposed to cell extracts of B. cereus 4810/72 and ATCC 14579T 34 2.5 Principle of the MTT test 35 3.1 Procedure for the detection and enumeration of Bacillus species 52 3.2 Colony development on PEMBA plates spread inoculated with centrifugated sediments of homogenized raw and cooked rice samples and incubated at 30ºC for 48 h 64 4.1 Plate set-up for MTT assay 88 4.2 HPLC analysis of the standard valinomycin 92 4.3 HPLC analysis of the purification product 92 4.4 NMR spectrum of cereulide 94 4.5 Influence of borate and phosphate buffer 96 4.6 Influence of buffer pH on valinomycin separation 99 4.7 Influence of SDS concentration on valinomycin separation 100 4.8 Peak of cereulide 101 4.9 Calibration curve for cereulide 103 4.10 Cereulide produced by Bacillus species grown on tryptic soy agar plates for 3 days 108 4.11 MEKC analysis of ‘non-producing’ strains 109 4.12 Cereulide produced by B. cereus 7401 strain grown on tryptic soy agar plates and cooked rice for 3 days 110 4.13 MEKC analysis of the outbreak sample of emetic B. cereus food poisioning 111 5.1 ELISA procedure for detection of diarrhoeal and emetic toxins 125 5.2 Cluster analysis of RAPD-PCR profiles obtained with primer OPA 3 for strains in the B. cereus group 132 5.3 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for strains in the B. cereus group 133 5.4 Cluster analysis of RAPD-PCR profiles obtained with primer OPA 3 for B. subtilis/B. mojavensis and B. amyloliquefaciens strains 138 5.5 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. subtilis/B. mojavensis and B. amyloliquefaciens strains 139
G 5.6 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. licheniformis strains 145 5.7 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. megaterium strains 147 G Appendix 3 2 NMR spectrum of cereulide 1 in [ H6]DMSO (600MHz) 210
G
G G Chapter 1
Introduction G The genus Bacillus comprises a very large and diverse group whose members are widespread in nature. They are important in the food industry and also to consumers because of the association of some species with food poisoning and food spoilage.
There are several reasons for the problems caused by Bacillus species in the food industry. First, they are difficult to eliminate from foodstuffs and food processing systems due to formation of spores which are ubiquitous and highly resistant to adverse conditions such as heat, dehydration, and other physical stresses (Kamat et al., 1989;
Andersson et al., 1995; Larsen and Jørgensen 1999). Second, the spores, as well as older cells, are hydrophobic and have adhesive characteristics, which facilitate their attachment to the surface of processing equipment, with subsequent formation of biofilms (Andersson et al., 1995; Granum and Rönner, 1998; Peng et al., 2001b;
Heyndrickx and Scheldeman, 2002). Third, both spores and vegetative cells appear to have become increasingly tolerant or resistant to environmental stressors, enabling them to survive under conditions or through treatments generally considered to stop growth or to inactivate all viable microorganisms (Heyndrickx and Scheldeman, 2002).
Bacillus cereus is the most prevalent pathogenic Bacillus species found in food and other industrial processes (Pirttijärvi et al., 1999). Strains of this bacterium cause food spoilage during production, storage and distribution (Huis in’t Veld, 1996), and food poisoning by gastrointestinal intoxication as well as a variety of local and systemic infections such as endophthalmitis, endocarditis, meningitis, periodontitis, osteomyelitis, wound infection, and septicemia (Kortiranta et al., 2000, Schoeni and Wong, 2005).
In relation to foodborne illness, B. cereus is associated with two distinct types of gastrointestinal disorders, the diarrhoeal and emetic syndromes. The first and better understood is the diarrhoeal syndrome, which results in diarrhoea and abdominal pain 8
G Y to 16 h after ingestion of contaminated foods, while the other is the emetic syndrome, which is characterised by nausea and vomiting within 1 to 5 h of ingestion (Beattie and
Williams, 2000). The infective dose of diarrhoeal B. cereus in food poisoning ranges from 104 to 1011 cfu per g (or mL) (Granum and Lund, 1997) and the level of B. cereus in foods incriminated in incidents of emetic food poisoning has ranged from 103 to 1010 cfu per g (or mL) (Lund, 1990). It is generally considered that any food with a population exceeding 104 to 105 cells or spores per g of B. cereus may not be safe for consumption (Beattie and Williams, 2000; Granum, 2002), though the number is often exceeded in a wide range of foods that are actually consumed, without incident
(Notermans and Batt, 1998; Häggblom et al., 2002). This may reflect the wide variation of pathogenic potential and overall diversity of B. cereus strains (Granum, 2002;
Häggblom et al., 2002).
While several studies have indicated that only a minority of B. cereus isolates may produce the emetic toxin, cereulide, (Mikami et al., 1994; Agata, et al., 1996; Pirttijärvi et al., 1999), Yokoyama et al. (1999) suggested that more than 90% of food poisoning caused by B. cereus is of the emetic type in countries of the Far East. This may be partly due to wide consumption of rice, which is a well-known food vehicle for the emetic toxin. The stability of cereulide to heat, extremes of pH and proteases, also could contribute to the greater occurrence of the emetic disease as compared to the diarrhoeal type of the disease. The diarrhoeal enterotoxins have been shown to be sensitive to low pH and to the action of proteolytic enzymes such as trypsin and pepsin (Shinagawa et al., 1991), such that it is at least partially degraded during passage through the human alimentary tract (Christiansson, 1993).
The diarrhoeal enterotoxins have been studied extensively and a number of assays
G Z developed for their detection whereas relatively little was known regarding the emetic toxin due, at least in part, to a lack of reliable and rapid detection methods (Agata et al.,
1995b; Andersson, et al., 1998), although this is rapidly changing. The bioassays that are currently used for measuring cereulide give an approximate titre of toxicity, but not an accurate concentration (Andersson et al., 1998; Finlay et al., 1999; Mikami et al.,
1994). Therefore, the dose of cereulide causing illness in humans is not known and safety limits for cereulide in food thus cannot be indicated, such as those that have been set for the mycotoxins aflatoxin B1, ochratoxin A, and the trichothecenes (Häggblom et al., 2002).
Most food poisoning incidents attributed to Bacillus species are associated with B. cereus, but the role of other Bacillus species as food poisoning agents is being increasingly recognised, with recent epidemiological evidence linking B. licheniformis,
B.subtilis, B. pumilus, and B. thuringiensis with incidents or outbreaks (Drobniewski,
1993). The B. subtilis group (B. licheniformis, B. subtilis and B. pumilus) has also been associated with a range of clinical conditions as well as food spoilage such as ropy bread, while B. licheniformis has also been associated with serious systemic disease, such as septicemia, peritonitis and ophthalmitis (Salkinoja-Salonen et al.,1999).
Several studies have suggested that the diarrhoeal enterotoxins produced by Bacillus species other than B. cereus are similar to those toxins well characterised for B. cereus
(Hansen and Hendriksen, 2001; Perani et al., 1998), and diarrhoeal enterotoxins have been found to be produced by B. thuringiensis, B. anthracis, B. mycoides, B. subtilis and
B.lichenifomis when tested using established cell culture and immunoassay detection methods for these toxins (Beattie and Williams, 1999; Fletcher and Logan, 1999). More recently, Salkinoja-Salonen et al. (1999) reported that strains of B. licheniformis
G [ associated with food poisoning produced a toxin with properties extremely similar to those of B. cereus emetic toxin, cereulide. Further, From et al. (2005) reported that
Bacillus species outside the B. cereus group produced a putative emetic toxin which is extremely heat-stable and Taylor et al. (2005) also demonstrated that heat-stable toxin production by strains of Bacillus cereus, Bacillus firmus, Bacillus megaterium, Bacillus simplex and Bacillus licheniformis. However, it is still unclear that the heat-stable emetic toxin produced by other Bacillus species is identical to cereulide, or merely similar.
This project aims to systemically examine the microbial ecology of Bacillus species in rice products, with which B. cereus emetic food poisoning is most frequently associated.
This study will focus on both the toxic potential and the diversity of Bacillus species associated with rice products. The biodiversity of Bacillus species will be investigated using a combination of the conventional detection (dilution and plating) method and the centrifugation-plating method, developed for the sensitive detection and enumeration of
Bacillus species. The high sensitivity of this method is expected to lead to an improved detection of a variety of Bacillus strains in rice products, as many samples usually contain only low levels of Bacillus species. This study will examine the applicability of capillary electrophoresis to the rapid and quantitative detection of the emetic toxin, cereulide, and employ that method in studying a range of reference Bacillus species as well as bacilli isolated from rice products. Finally, the diversity and the toxigenicity of isolates from rice products will be determined using molecular and immunological methods. This includes use of a recently developed PCR method for detection of a gene associated with cereulide production and a prototype ELISA for detection of cereulide.G
G \ Chapter 2
Literature review
G This thesis examines the ecology of toxigenic Bacillus species associated with rice products. Key studies associated with these topics are presented in the introduction and discussion sections of subsequent experimental chapters. This chapter provides a background of the genus Bacillus and a detailed review of the literature describing the toxins produced by bacilli associated with food poisoning.
G 2.1 Taxonomy of the genus Bacillus The genus Bacillus comprises a very large and diverse group whose members are either aerobic or facultatively anaerobic, Gram-positive rods that are capable of forming endospores and usually produce catalase (Shinagawa, 1990; Granum and Lund, 1997).
In the last decade, following the application of numerical taxonomic methods and molecular techniques to bacterial systematics, at least seven new genera from Bacillus were established, including Alicyclobacillus (Wisotzkey et al., 1992), Paenibacillus
(Ash et al., 1993; validated Collins et al., 1994), Aneurinibacillus (Shida et al., 1996),
Brevibacillus (Shida et al., 1996), Virgibacillus (Heyndrickx et al., 1998),
Gracilibacillus (Wainö et al., 1999) and Salibacillus (Wainö et al., 1999). Although some species have been moved (or will be moved) to these new genera, the organisms of significance to the public health microbiologist remain in the genus Bacillus (Jenson and Moir, 2003).
The numbers of species in each Bacillus group fluctuated between the 1st edition (1923) and the 8th edition of Bergey’s Manual of Determinative Bacteriology (1974) (Table
2.1) due to heterogeneity within the genus. These numbers changed again in Bergey’s
Manual of Systematic Bacteriology (1986), in which a further 40 species were
G ^ recognised, with another 27 incertae sedis among the more than 200 species of Bacillus.
Some of the validated species are shown in Table 2.2.
Table 2.1 Number of species assigned to the genus Bacillus in different editions of Bergey’s Manual up to 1974
Bergey’s Manual Year Number of species 1st edition 1923 75 2nd edition 1925 75 3rd edition 1930 93 4th edition 1934 95 5th edition 1938 146 6th edition 1948 33 7th edition 1954 25 8th edition 1974 Group I: 22, Group II: 26 (Source: Gordon, 1981; Berkeley, 2002)
In the 8th edition of Bergey’s Manual of Determinative Bacteriology (Gibson and
Gordon, 1974), 22 species of Bacillus, including B. cereus, were classified in Group I of the genus (those members with central or terminal ellipsoid or cylindrical spores which do not distend the sporangium), and were widely accepted as distinct species, whereas the 26 members of Group II had received little attention (Berkeley, 2002). Of the Group
I members, the B. cereus group and the B. subtilis group have received widespread recognition as causative agents of food-borne illness. The Bacillus cereus group comprise six recognised species including B. cereus, B. anthracis, B. thuringiensis, B. mycoides, B. pseudomycoides and B. weihenstephanensis (Jensen et al., 2003). Studies of the sequences of 16S rRNA revealed that B. cereus, B. anthracis, B. mycoides and B. thuringiensis, referred to as the B. cereus group, are very closely related, more than 99% similar (Ash et al., 1991a, b). A later study (Helgason et al., 2000b) agreed with this consideration, except that B. mycoides may be a separate species, based on analysis
G _ using multilocus enzyme electrophoresis (MEE). Earlier DNA-DNA hybridisation studies of the B. cereus group (Kaneko, et al., 1978; Seki et al., 1978) also revealed a high degree of similarity between species but some inconsistencies were noted with respect to overall relatedness of species and thus it may not be appropriate to place them in the same species (Jenson and Moir, 2003). In fact, extensive biochemical, physiological and morphological studies of the genus failed to find characters that would consistently differentiate these four species (Priest et al., 1988), whereas members of the B. subtilis group, which are B. subtilis, B. licheniformis and B. pumilus, are accepted as distinct, easily differentiated species (Jenson and Moir, 2003).
Table 2.2 Bacilli type strains Type strains Type strains Type strains B. anthracis B. pasteurii P. alvei B. thuringiensis B. azotoformans P. polymyxa B. mycoides B. macquariensis P. larvae B. circulans B. globisporus Br. brevis B. cereus B. popilliae Br. laterosporus B. insolitus B. firmus 1)A. acidocaldarius B. lentus B. pumilus 2)A. xylanus B. pantothenticus B. popilliae V. pantothenticus B. alcalophilus B. firmus G. stearothermophilus B. megaterium B. subtilis S. inulinus B. sphaericus B. badius B. marinus B. licheniformis B. lentimorbus B. coagulans B., Bacillus; P., Paenibacillus; Br., Brevibacillus; 1)A., Alicyclobacillus; 2)A., Amphibacillus; V., Virgibacillus; G., Geobacillus; S., Sporolactobacillus (Source: Claus and Berkeley, 1986; Kämpfer; 2002)
G ` 2.2 Ecology of Bacillus species
Bacillus species are found in a number of ecological niches, due to the resistance of their endospores to environmental stress, as well as their long-term survival under adverse conditions, but it is generally accepted that the primary habitat of endospore- forming bacilli is the soil (soil contains 105 – 106 spores of B. cereus/g).
Therefore, raw foods of plant origin are a major source of B. cereus in food and the presence of Bacillus species in plant foods is the result of soil contamination rather than a specific association between the microorganism and plants (Priest, 1989). Cereal products and various types of peas and beans have been found to be frequently contaminated with B. cereus (Blakey and Priest, 1980; Rosenkvist and Hansen, 1995).
Raw rice is a well known source of B. cereus, with most samples containing the bacterium, but usually at low levels (Table 2.3) (Jenson and Moir, 2003). Both raw and pasteurised milk is also often contaminated with B. cereus (Ahmed et al., 1983;
Griffiths, 1992). In a study of high temperature-short time (HTST) pasteurised milk in
Germany, psychrotrophic B. circulans was isolated with the highest frequency, while B. cereus, B. licheniformis, and B. coagulans were found to be predominant in milk in
Scotland (Surtherland and Murdoch, 1994; Mayer et al., 1999).
The presence of B. cereus in processed foods stems from contamination from raw materials or is associated with the ability of spores to survive heat treatment processes.
In the case of bread, though not a major source of B. cereus, spoilage may be due to B. subtilis, B. licheniformis or other Bacillus species (Rosenkvist and Hansen, 1995;
Sorokulova et al., 2003), and boiled rice or, more commonly, fried rice may contain high numbers of B. cereus (Table 2.3) (Jenson and Moir, 2003). Ninety-five percent of B.
GXW cereus emetic syndrome outbreaks have been associated with the consumption of boiled or fried rice from Cantonese-style restaurants and thus food poisoning due to the emetic toxin has been termed the ‘Chinese-restaurant syndrome’ (McElroy et al., 2000).
Table 2.3 Incidences of B. cereus in raw and cooked rice Rice Incidence Count per gram Reference Raw rice 6/13 1×102 – 1×103 Lee and Chang (1980) Raw rice 11/16 <1×102 – 2×105 Lee et al. (1995) Raw rice 6/6 2×101 – 6×102 Kamat et al. (1989) Boiled rice 25/252 1×102 – 1×105 Gilbert and Parry (1977) Boiled rice 4/4 1×103 – 4×104 Karmat et al. (1989) Fried rice 49/204 1×102 – 1×105 Gilbert and Parry (1977) Fried rice 7/61 1×103 – 1×106 Schiemann (1978) (Source: Jenson and Moir, 2003)
2.3 Control measures
Since Bacillus species are widespread in nature, there are few foodstuffs that are free of these species. However, generally low numbers of Bacillus species and/or their spores will not cause problems unless spores are able to germinate and significant growth occurs. Therefore, control of Bacillus species in food relies on inhibiting spore germination and, most importantly, preventing the growth of vegetative cells in cooked, ready-to-eat foods by proper cooking/processing procedures and food handling practices.
To prevent growth of Bacillus species or greatly reduce its rate, foods should be kept at a temperature higher than 60ºC (Lund, 1990; Jenson and Moir, 2003) or, if it is going to be stored, it should be cooled rapidly to a temperature below 8 – 10ºC (Lund, 1990).
However, for refrigeration to be more effective, foods should be maintained at or below
5ºC, as some psychrotrophic strains can grow down at 4 – 5ºC (Dufrenne et al., 1994).
GXX 2.4 Detection, isolation and biochemical properties of Bacillus species
The only well-established detection methods for a Bacillus species of interest to the food microbiologist are for B. cereus (Jenson, 2000). Present detection and enumeration methods for B. cereus (Figure 2.1) rely primarily on direct plating. A number of selective-differential plating media for the detection and enumeration of B. cereus in foods have been developed (Table 2.4), the major criteria for detection and identification of B. cereus on these plating media including haemolytic activity, lecithinase activity (egg yolk reaction), since most strains of the B. cereus group possess lecithinase activity, and/or inability to ferment mannitol.
In mannitol egg yolk polymyxin (MYP) agar, a widely used medium, the single egg yolk diagnostic system of egg yolk agar (EYA) was expanded to a double diagnostic system by the addition of a mannitol/phenol red combination (van Netten and Kramer,
1992). However, due to the high peptone content (1%) of the medium, sporulation is often poor and re-streaking of suspect colonies may be necessary to confirm colony morphology and sporulation, thus increasing analysis time (Jenson and Moir, 2003).
Those limitations, as well as poor differentiation of mannitol-utilising organisms from B. cereus on MYP, and also on Kim and Goepfert (KG) agar, led to the development of polymyxin egg-yolk mannitol bromothymol blue agar (PEMBA) or B. cereus selective medium (Holbrook and Anderson 1980). PEMBA, the standard medium in Australia at the time of performance of this study, allows detection and enumeration of a low numbers of B. cereus against a high background flora. It contains polymyxin B to inhibit growth of Gram-negative bacteria, mannitol/bromothymol blue and egg yolk to identify suspect colonies of B. cereus; and low levels of peptone to promote sporulation.
GXY More recently, a methylene blue Mg2+/Zn2+ glucose polymyxin agar (RVC) agar was developed as a less expensive alternative to MYP and KG (van Netten and Kramer,
1992). This medium differs from others such as MYP and KG mainly in the use of a lower concentration of polymyxin B. However, for analysis of food samples, the use of a higher concentration of polymyxin B was recommended (van Netten and Kramer,
1992). Thus, RVC may not be a cheaper alternative, representing just another of a range of selective-differential plating media for B. cereus. Chromogenic agar media have been developed to give more specificity by including chromogenic substrates, such as 5- bromo-4-chloro-3-indoxyl-myo-inositol-1-phosphate or 5-bromo-4-chloro-3-indoxyl-ȕ- glucosidase, and greater selectivity by inclusion of ceftazidime (Peng et al., 2001a), but these media, due to the cost of these substrates, are currently too expensive to use routinely.
At present, none of the selective-differential media for B. cereus can differentiate between B. cereus and other members of the B. cereus group, such as B. mycoides, B. thuringiensis (Schraft and Griffiths, 1995) though most isolates of B. mycoides are differentiated easily on agar media because they characteristically produce rhizoid colonies (Jenson, 2000).
While typically the method of choice, the direct plating method may not be sufficient to use for foods contaminated with low numbers of B. cereus or injured cells. Thus, enrichment is needed to isolate the organism. The enrichment method for B. cereus
(Figure 2.1) was developed with a selective medium containing polymyxin B, which suppresses Gram-negative organisms. However, this enrichment procedure is not specific for B. cereus; other Bacillus species as well as other Gram-positive bacteria can
GXZ grow. Furthermore, it will not generate quantitative data because of the obvious variation in the population dynamics of enrichment.
The limitations could be overcome with a centrifugation-plating method. The principle of this method is very simple. Microbial cells are readily sedimented by centrifugation at 3,000 – 10,000 × g within minutes and centrifugation is a simple, efficient and inexpensive physical method to concentrate and enrich microbial cells (Cox and Fleet,
2003). Thus, centrifugation is a physical form of enrichment method that could assist in determining true population and diversity data. It can also overcome much of the uncertainty and unreliability associated with culture enrichment methods (Hawa et al.,
1984; Fleet et al., 1991; Mossel et al., 1991; Sharpe, 2000; Cox and Fleet, 2003)
GX[ Table 2.4 Selective and diagnostic system of B. cereus isolation media Medium Basal medium Selective system Diagnostic system Confirmation test 1. Bovine/horse blood agar (BA) None Haemolysis, Sugar fermentation, Colony morphology microscopy 2. Columbia base blood Blood agar base Polymyxin B (500 IU-on Haemolysis, polymyxin B agar (CBP-A) surface of 20ml plates) Colony morphology 3. Egg yolk agar (EYA) Nutrient medium None Egg yolk reaction (blood agar base no 2) 4. Bacillus cereus/ Clostridium Peptone beef extract None Egg yolk reaction, perfringens agar (BCP) mannitol/bromocresol purple 5. Citrate egg yolk polymyxin B Nutrient agar Polymyxin B (50 IU/mL), Egg yolk reaction, lithium chloride agar (CELP) Lithium chloride (0.5%) Phenol red 6. Kendall’s Bacillus cereus Nutrient agar None Egg yolk reaction, medium (BCM) mannitol/bromocresol purple 7. Mannitol egg yolk polymyxin Sporulation medium Polymyxin B (100 IU/mL) Egg yolk reaction, Microscopy, VP, glucose B agar (MYP) mannitol/phenol red and xylose fermentation, anaerobic growth, gelatin hydrolysis, nitrate reduction 8. Kim and Goepfert agar (KG) 0.1% peptone Polymyxin B (100 IU/mL) Egg yolk reaction, Serology by fluorescent- mannitol/phenol red labelled anti-exosporium sera 9. Polymyxin B egg-yolk 0.1% peptone Polymyxin B (100 IU/mL) Egg yolk reaction, Microscopy for spores and Mannitol bromothymol blue mannitol/bromothymol blue fat globules agar (PEMBA) 10. Polymyxin B egg-yolk 0.1% peptone Polymyxin B (100 IU/mL) Egg yolk reaction, mannitol bromocresol purple mannitol/bromocresol purple agar (PEMPA) 11. Mannitol egg-yolk Trypticase/tryptone Polymyxin B (50 IU/mL) Egg yolk reaction polymyxin B agar (RVC) 12. Chromogenic agar Protease peptone Polymyxin B (100 IU/mL) Phosphatidylinositol activity Ceftazidime (20 mg/L) (Source: van Netten and Kramer, 1992; Peng et al., 2001a; Jenson and Moir, 2003) GX\ Sample (Suspected foods, faeces, vomitus and swabs)
Direct plating Enrichment Enumeration of culture culture B. cereus
Media: BHI with MYP agar polymyxin B KG agar 10% homogenate PEMBA agar (32-35 C, 18-24 h) Inoculate (0.1 ml) etc.
Plating culture Plating culture (32-35 C, 18-24 h)
(32-35 C, 18-24 h) (32-35 C, 18-24 h)
B. cereus suspect colony
Grey-white colony B. cereus Mannitol (-) Presumptive Pure culture Egg yolk reaction (+) Estimate the number Non-selective: Blood agar of B. cereus Heart Infusion agar
Morphological test Confirmatory biochemical tests
Gram stain: Gram (+) Motility (+) Large rod VP reaction (+) Spore stain: Central (+) Strarch hydrolysis (+/-) Swelling (-) Citrate utilization (+) Toxin crystals (-) Nitrate reduction (+) Acid from: Glucose (+) Mannitol (-) Xylose (-) Identification Arabinose (-)
Typing tests Toxin detection
H-serotyping Enterotoxin Biotyping Emetic toxin Phage typing (biological and/or immunological tests)
Figure 2.1 Isolation and identification of B. cereus (Source: Shinagawa, 1990)
16 Isolates of bacilli can be identified by morphological studies based on Gram and spore
staining, and biochemical tests. Table 2.5 shows some of the physical and biochemical
characteristics of Bacillus species of public health concern.
Table 2.5 Identification of Bacillus species of public health significance B. B. B. B. B. B. B. thurin- licheni- cereus mycoides anthracis subtilis pumilus giensis formis Cell diameter >1.0Pm ++ + + - - -
Anaerobic glucose ++ + + -+ fermentation Nitrate reduction + + + + + + -
Voges-Proskauer + + d + + + + Tyrosine +d d d -- - decomposition Lysozyme sensitivity + + + + d d d
Mannitol fermentation - - - - + + +
Egg yolk reaction + d d + - - -
Motility + - + -
Rhizoid growth - + -
Haemolysis + + + -
Toxin crystals - - + - +, 90% or more of strains are positive; -, 90% or more strains are negative; d, 11-89% of strains are positive (Source: Claus and Berkeley, 1986; Jenson, 2000)
Miniaturised versions of traditional biochemical tests, embodied in API kits, Vitek cards
and Biolog plates, can make easier the performance of these laborious and time-
consuming traditional biochemical tests. However, these commercial identification
systems only allow the recognition of 30 – 34 species among the 123 valid species in
Bacillus and the seven new genera (Section 2.1) derived from it (Logan, 2002). Further,
X^G the reliability of these tests is, at times, questionable because ambiguous reactions may lead to misinterpretation. Therefore, these systems may not be sufficient to use alone for identification of bacteria.
2.5 Centrifugation-plating
Centrifugation is a simple and efficient physical approach to cell concentration and enrichment. The concentrated and sedimented cells can be resuspended in a small volume of diluent and then spread-plated onto an appropriate medium.
In analysis of food samples, centrifuged microbial cells may be co-sedimented with food particles, such that resuspension and plating of sedimented microbial cells cannot be performed easily. However, it is possible, through treatment with chemicals such as surfactants, digestion with hydrolytic enzymes, or physical separation, such as passage through a nylon mesh or sieve, to improve the separation of cells from food residue or minimise pelleted food residues during centrifugation (Cox and Fleet, 2003).
2.5.1 Enzymes and surfactant
Some food samples may need pre-treatments for successful application of the centrifugation-plating method in microbiological analysis. Enzymes can be used to digest food materials to facilitate separation or release of microbial cells, improving recovery and concentration on centrifugation (or filtration) and surfactants are used to disperse lipid globules in which microbial cells are trapped and, as with enzymes, to improve the efficiency of a physical separation (centrifugation or filtration) process.
X_G This was proven through the use of the proteolytic enzyme trypsin and the surfactant
Triton X-100 (Cousins et al., 1979; Pettipher et al., 1980), and the use of protease and
Tween 80 (Peterkin and Sharpe, 1980; Gnanou Besse and Lafarge, 2001) to improve the filterability of milk, milk products or cheese. The results demonstrated that these treatments increased the filterability of samples. Similarly, Entis et al. (1982) found trypsin alone was effective for many foods, with a surfactant (Tween 80) required to significantly improve filterability of high fat foods. It was also shown in some of these studies that treatment with a mild surfactant and a protease had no adverse effect on the total viable count of bacterial cells (Entis et al., 1982), the viability of eight common strains of bacteria (Peterkin and Sharpe, 1980) or Listeria monocytogenes (Gnanou
Besse and Lafarge, 2001). Therefore, it may be generally accepted that the use of proteolytic enzyme and mild surfactants such as Tween 80 can improve the filterability of foods. It follows that digestion or suspension of food matrix using such treatments will reduce the volume of pelleted material after centrifugation. However, there is no clear data on the effects of other types of enzymes such as amylases. Thus if other enzymes are to be used, their enhancement of the centrifugation process and influence on viability of bacteria cells would need to be studied.
2.5.2 Application of centrifugation-plating method
Several studies have demonstrated the application of this method. Hawa et al. (1984) used the method for enumeration of Salmonella on chicken carcasses and, as well as reporting numerical data, showed that the centrifugation method was more sensitive for detection of Salmonella than the conventional cultural standard method, detecting
Salmonella when the conventional enrichment and plating technique did not. A further study conducted by Morrison and Fleet (1985) showed that the method permitted
X`G detection of as few as 10 Salmonella cells per chicken carcass. More recently, the method was used for detection and enumeration of Listeria, in cheese (Uyttendaele et al., 2000) and chicken carcasses, cheese, minced meat, salad and milk (Jang, 2004).
Collectively, these studies showed that centrifugation can be applied successfully to analyse food samples and this simple and relatively rapid operation, centrifugation has great potential for application to the microbiological analysis of foods (Cox and Fleet
2003).
G 2.6 Methods for strain typing of Bacillus species
Monitoring food-borne pathogenic bacteria through typing at the subspecies (strain) level is an important part of investigative processes (Farber, 1996; Graves et al., 1999;
Wiedmann, 2002). Typing techniques that have good discrimination power are needed for epidemiological investigation of outbreaks of foodborne disease involving microbial pathogens, including Bacillus species, or tracking of environmental sources of contamination for management of food safety and quality.
In general, typing techniques can be divided into phenotypic and genotypic methods.
Traditionally, the relatedness of bacterial isolates has been determined mainly by phenotypic methods such as bio-, sero-, and phage typing, antibiotic susceptibility tests, immunoblotting, and multilocus enzyme- or polyacrylamide gel electrophoresis (Farber,
1996; Busch and Nitschko, 1999). However, conventional methods using phenotypic markers have several limitations (Farber, 1996; Farber et al. 2001). They are not available for all bacterial species, may not be reproducible due to unstable expression of markers under some environmental or culture conditions, may be difficult to standardise,
YWG and may lack discriminatory power.
On the other hand, molecular typing methods, based on DNA analysis, are more reproducible, have increased discriminatory power, are easier to interpret, and may provide speed of performance, technical simplicity, prospects for statistical data analysis, and ease of automation (Maslow et al., 1993; Farber, 1996). The molecular methods that have been developed in the last decade aimed originally at the discrimination of species, subspecies and even strains in order to perform epidemiological studies though they often provide information relevant to studies of general microbial populations (De
Vos, 2002). The most common genotypic typing methods used in food microbiology include chromosomal DNA restriction analysis, plasmid typing, ribotyping, pulsed-field gel electrophoresis (PFGE) and PCR-based methods such as randomly amplified polymorphic DNA (RAPD) analysis. Many of these molecular typing methods have been applied to isolates of Bacillus species, and those most commonly used are examined in this section. Some of the advantages and limitations of these methods are summarised in Table 2.6.
Table 2.6 Overview of genetic fingerprinting techniques for typing of Bacillus species Typing technique Advantages Limitations PFGE Very high discrimination power Expensive, slow to perform Ribotyping Can be automated, can be used Expensive, laborious, manually for bacterial identification, slow to perform good discrimination power RAPD Fast, simple, and cost-effective Reproducibility problems possible rep-PCR Fast, simple, and cost-effective Reproducibility problems possible (Source: Maukonen et al., 2003)
YXG 2.6.1 Restriction fragment length polymorphism (RFLP) and pulsed-field gel electrophoresis (PFGE)
RFLP and PFGE are the typing techniques that involve visualising fingerprints of the complete genome (De Vos, 2002). Since restriction endonuclease enzymes are very specific, complete digestion of a given DNA generates a reproducible pattern of DNA fragments and variations in the pattern of fragments is called RFLP (Farber, 1996).
PFGE, which is similar in principle to RFLP, was developed for the separation of high molecular weight DNA fragments generated using rare-cutting restriction enzymes (De
Vos, 2002).
These methods have been applied to the differentiation of species of bacilli such as B. sphaericus (Zahner et al., 1998) or Brevibacillus laterosporus (Zahner et al., 1999), and closely related Bacillus species such as those within the B. cereus group (Carlson et al.,
1994; Helgason et al., 2000a) and to assess genetic variability of B. anthracis and related species (Harrell et al., 1995). More practically, these methods have also been used for epidemiological purposes, such as the investigation of a pseudo-outbreak of B. cereus in a pediatric unit (Liu et al., 1997), or typing of Bacillus species isolated from food (Schraft et al., 1996).
2.6.2 Ribotyping
Ribotyping involves restriction enzyme digestion of chromosomal DNA, electrophoretic separation of fragments, transfer of those fragments from a gel onto a nitrocellulose or nylon membrane (Southern, 1975) and hybridisation with labelled rRNA genes (usually both 16S and 23S rRNA genes, or a whole rRNA operon containing 16S, 23S, and 5S rRNA genes and spacer regions) or a rDNA probe (Stull et al., 1988; Maukonen et al.,
YYG 2003). Ribotyping can be considered as an interesting approach to taxonomic and epidemiologic studies of Bacillus species because Bacillus sensu lato members contain
9 – 12 rRNA operons (De Vos, 2002). The Riboprinter, an automated ribotyping instrument, makes this method easy to perform, though the cost of equipment and consumable items makes use far too expensive in most cases.
This method has been applied to the investigation of strains of B. licheniformis from food poisoning (Salkinoja-Salonen et al., 1999), and discrimination of Bacillus weihenstephanensis from raw cow’s milk (Pacova et al., 2003), and aerobic mesophilic bacilli isolated from board and paper products containing recycled fibres (Suihko and
Stackebrandt, 2003). It has also been used to trace different pathotypes of B. thuringiensis (Akhurst et al., 1997). Several studies (Andersson et al., 1999; Shangkuan et al., 2000; Guillaume-Gentil et al., 2002) have compared or used ribotyping with other
DNA fingerprinting methods, and one study (Andersson et al., 1999) revealed that
RAPD-PCR is slightly more discriminatory than automated ribotyping for B. cereus isolates.
2.6.3 Randomly amplified polymorphic DNA (RAPD)
RAPD is a very simple and quick method in which arbitrary oligonucleotides are used as primers in PCR. RAPD-PCR differs from the normal PCR in two main respects
(Farber, 1996). First, the primers used are short (usually 9- or 10-mers or nucleotides), high in GC content, the sequences are chosen at random, and only one primer is used.
Second, the annealing temperature for RAPD is lower than for normal PCR. The short primer hybridises at random sites on the chromosome which may or may not be completely complementary in sequence and, if the sites are located sufficiently close
YZG together on complementary strands, an amplicon is generated, and the DNA fingerprints differ accordingly to the degree of relatedness of the strains investigated (Busch and
Nitschko, 1999; Cox and Fleet; 2003). This method has been commercialised by
Amersham-Pharmacia-Biotech (Sweden).
Even though there are many advantages to RAPD typing in terms of discriminatory power, ease of interpretation and performance (Maslow et al., 1993; Arbeit, 1995), there is major concern with the reproducibility of this method. However, some studies
(Meunier and Grimont, 1993; Farber and Addison, 1994; Neiderhauser et al., 1994) have shown that this method can be reproducible and works well when experimental conditions that could affect the variation in the RAPD fingerprint patterns are carefully controlled.
The RAPD PCR method has been applied to the discrimination and characterisation of
B. sphaericus (Woodburn et al., 1995), B. thuringiensis (Brousseau et al., 1993; Hansen et al., 1998) and thermophilic members of the Bacillus genus (Ronimus et al., 1997;
Zhang et al., 2002; Ronimus et al., 2003). It was even applied to discriminate B. anthracis from closely related Bacillus species such as B. cereus, B. thuringiensis and B. mycoides (Daffonchio et al., 1999) and identify rope spoilage strains of B. subtilis and B. licheniformis (Sorokulova et al., 2003).
2.6.4 REP-PCR
Rep-PCR is based on the observation that repetitive elements are randomly distributed in genomes of bacteria and relies on a primer derived from such repetitive sequences identified in many bacteria (Versalovic et al., 1991). Rep-PCR has been developed for
Y[G enterobacterial repetitive intergenic consensus (ERIC) and repetitive extragenic palindromic (REP) sequences (Cox and Fleet, 2003). The use of consensus primers may reduce stringency of PCR and this may negatively affect reproducibility of the method
(De Vos, 2002).
This method has been used to examine genetic diversity in B. sphaericus (Da Silva et al.,
1999; Miteva et al., 1999) and Paenibacillus larvae ssp. larvae (Alippi and Aguilar,
1998) and has been applied to the differentiation of B. sporothermodurans from other
Bacillus species in milk (Herman et al., 1998).
2.7 Food poisoning associated with Bacillus species
2.7.1 Food poisoning associated with B. cereus
Most food poisoning incidents attributed to Bacillus species are associated with Bacillus cereus. This bacterium is known to cause two different types of food poisoning, the diarrhoeal and the emetic syndromes. The diarrhoeal syndrome is characterised by diarrhoea 8 – 16 h after consumption of the food and is accompanied by abdominal pains and occasionally nausea, but seldom vomiting, and can be confused with illness caused by Clostridium perfringens (Kramer and Gilbert, 1989; Bennett et al., 1993;
Schoeni and Wong, 2005). Foods implicated include puddings and soups, green vegetables, mashed potatoes, sauces and proteinaceous foods such as cooked meat
(Kramer and Gilbert, 1989; Granum, 1994).
Symptoms of illness associated with ingestion of the emetic toxin are nausea and vomiting, but rarely diarrhoea, commencing 1 – 6 h after consumption (Kramer and
Y\G Gilbert, 1989; Schoeni and Wong, 2005). The illness can be confused with that caused by Staphylococcus aureus intoxication (Shinagawa, 1990; Bennett et al., 1993). Foods most commonly associated with emetic illness include farinaceous food such as cooked rice, noodles and pasta (Kramer and Gilbert, 1989; Granum and Lund, 1997).
The dominant B. cereus syndrome differs from country to country depending on their food culture. In Asian countries, such as Japan, the emetic syndrome is reported about ten times more frequently than the diarrhoeal syndrome, while in Europe and North
America the diarrhoeal type is most frequently reported (McKillip, 2000).
The risk of B. cereus food poisoning was highlighted in Australia in 2002, when 272 people were hospitalised due to emetic illness caused by B. cereus (Anon, 2002). The outbreak was associated with a meal consisting of rice, lamb and potatoes and B. cereus was isolated from the rice sample. This incident highlighted the importance of proper food preparation, in particular, maintaining food at the proper temperature. Storage at a growth-permissive temperature had allowed for growth and emetic toxin production.
Upon reheating, while the bacterium may have been eliminated, the toxin remained active. In food service settings it is common to cook large quantities of rice which may then be left at ambient temperature prior to reheating and service (Jenson and Moir,
2003).
2.7.2 Food poisoning associated with other Bacillus species
Among the non-B. cereus related reports, B. subtilis is most frequently associated with food poisoning, with meat dishes, pastry products, meat or seafood with rice dishes being the most commonly implicated foods, and numbers greater than 105 cfu/g are
Y]G required to cause illness (Lund, 1990; Shinagawa 1990; Granum and Baird-Parker,
2000). Several episodes of food poisoning by B. subtilis in Japan (Shinagawa, 1990) and the UK (Lund, 1990) have been described. The predominant symptom of B. subtilis food poisoning is vomiting in most cases and the illness has a short onset incubation period (Lund, 1990; Jenson and Moir, 2003).
Food poisoning due to B. licheniformis is most often associated with cooked meats and vegetables, with an incubation period of about 2 – 14 h, and diarrhoea the predominant symptom, with vomiting in about half the cases (Lund, 1990; Tatzel et al., 1994;
Granum and Baird-Parker, 2000). These characteristics are similar to those of B. cereus diarrhoeal syndrome. However, according to Salkinoja-Salonen et al. (1999), a partially-purified toxin of B. licheniformis isolated from foods (raw milk and industrially produced baby food) involved in food poisoning incidents showed physicochemical properties similar to those of cereulide, the emetic toxin of B. cereus.
Only a few cases of food poisoning associated with B. pumilus have been reported, from the UK, associated with meat dishes, scotch eggs, cheese sandwich, and canned tomato juice (Lund, 1990; Granum and Baird-Parker, 2000). There was usually an acute onset of symptoms of vomiting and in some cases diarrhoea (Lund, 1990; Jenson and Moir,
2003).
B. thuringiensis, which is closely related to B. cereus and used as a bioinsecticide, is reported to produce enterotoxins (Damgaard et al., 1996; Fletcher and Logan, 1999;
Rivera et al., 2000) and has also been reported to cause food poisoning (Jackson et al.,
1995). Thus, extensive use of this organism as an insecticide could potentially cause
Y^G serious problems in the food industry, as raw materials may carry a high load of potentially toxigenic organisms. It needs to be ensured that B. thuringiensis in use as a bioinsecticide should not carry enterotoxin genes and in turn, is unable to produce food- poisoning toxins. The Health and Consumer Protection Directorate-General of the
European Commission has already accepted that only non-toxin producing Bacillus species should be allowed to be utilised in animal nutrition (Granum, 2002) and such regulatory concern should be expressed in other countries not yet aware of the issue.
Reliable tests are needed to enable the toxigenicity of strains to be determined.
2.8 Toxins of Bacillus species associated with food poisoning
2.8.1 B. cereus diarrhoeal toxin
2.8.1.1 Characteristics and properties of diarrhoeal toxin
Although the number of enterotoxins and their properties have been the subject of debate for some time (Kramer and Gilbert, 1989; Granum, 1997) and some of them are still considered as only candidate enterotoxins (Schoeni and Wong, 2005), five different enterotoxins have been identified and characterised to date (Table 2.7). There are two distinct enterotoxin complexes, and three different single enterotoxic proteins, namely enterotoxin T (BceT), enterotoxin FM, and cytotoxin K (CytK).
A three-component (B, L1, and L2) haemolysin (Hbl) was purified and characterised
(Beecher and Wong, 1994, 1997), and suggested to be a primary virulence factor in B. cereus diarrhoea (Beecher et al., 1995). The three protein of the Hbl are transcribed from one operon (hbl) and the Northern blot analysis that had shown RNA transcripts of
Y_G 5.5 kb hblC (transcribing L2) and hblD (transcribing L1) are separated by 37 bp and encode protein of 447 aa and 384 aa, respectively (Ryan et al., Granum, 2002). L2 has signal peptide of 32 aa and L1a signal peptide of 30aa and the B-protein, transcribed from hblA, consists of 375aa, with sinnal peptide of 31aa (Heinrichs et al., 1993;
Granum, 2002). A non-haemolysin three-component enterotoxin (Nhe) was also characterised by Lund and Granum (1996) after it was purified from a B. cereus strain isolated after a large food poisoning incident in Norway in 1995. The nhe operon contains three open reading frames: nheA, nheB and nheC (Granum et al., 1999). The first two gene products were addressed as the 45-kDa and 39-kDa protein respectively and the last transcript from the nhe operon that is nheC had not been purified and its function is not known yet (Granum, 2002). The bacT gene of B. cereus encodes an enterotoxin protein with characteristics of diarrhoeal toxin, known as enterotoxin T
(BceT) (Agata, et al., 1995a). This toxin has not been implicated in any outbreaks of food poisoning to date, but approximately 43% of randomly selected B. cereus strains isolated from different food products possessed the bceT gene (Beattie and Williams,
2000). However, recent investigation of the bceT sequence demonstrated that the cloned bceT was created by an incidental joining of four DNA fragments during ligation and one of these fragments showed 93% homology to the open reading frame (ORF 101) located within the pathogenic island of Bacillus anthracis pXO1 virulence plasmid
(Hansen et al., 2003). Therefore, authors suggested that the original reported enterotoxic activity of the cloned bceT construct could be due to either the fusion gene or the fragment with homology to ORF 101 in pOX1. The role of enterotoxin FM in food poisoning and its properties are still unknown. The most recently discovered cytotoxin
K (CytK) was the cause of the symptoms in a severe outbreak of B. cereus food poisoning, which included production of bloody diarrhoea in several people and three
Y`G fatalities, in France in 1998 (Lund et al., 2000).
These enterotoxins are heat-labile, being inactivated by heating to 56ºC for 5 – 30 min,
stable between pH values of 4 and 11 (Jackson, 1991), and sensitive to proteolytic
enzymes (Jenson and Moir, 2003). The mode of action of the toxin(s) has been
suggested as reversing the absorption of fluid, Na+ and Cl, in the large intestine, and
causing malabsorption of glucose and amino acids, necrosis and mucosal damage
(Granum et al., 1993).
Table 2.7 Enterotoxins produced by B. cereus
Toxin Type Food poisoning Reference Haemolysin BL (Hbl) Protein, 3 components Probably Heinrichs et al. (1993); Beecher et al. (1995); Beecher and Wong (1997); Ryan et al. (1997)
Non-haemolytic Protein, 3 components Yes Lund and Granum (1996, enterotoxin (Nhe) 1997); Granum et al. (1999)
Enterotoxin T (BceT) Protein, 1 components ? Agata et al. (1995a)
Enterotoxin FM Protein, 1 components ?? Asano et al. (1997) (EntFM)
Cytotoxin K (CytK) Protein, 1 components Yes, 3 deaths Lund et al. (2000) (Source: Granum, 2002)
2.8.1.2 Detection methods for diarrhoeal toxin
Traditionally, detection of enterotoxin has relied upon in vivo testing procedures such as
mouse lethality, rabbit ileal loop, and the vascular permeability reaction in rabbit skin
(Glatz et al., 1974; Jackson, 1991; Jackson, 1993). There has also been the use of tissue
culture systems such as the McCoy cell assay, whereby the enterotoxin produces a
ZWG characteristic progressive destruction of cell monolayers (Jackson, 1991). Other cell cytotoxicity assays utilising different tissue culture cells (HeLa, CHO, Vero and HEL), have also been used for the detection of the diarrhoeal toxin (Christiansson, 1993;G
Fermanian et al., 1996). Originally these were based on indirect biological activity of the toxin.
More recently detection of these toxins has been based on immunological techniques and kits are available commercially. Two commonly used immunoassays have been found to detect different antigens. The diarrhoeal type B. cereus enterotoxin test kit
(RPLA kit; Oxoid) is specific to the HblC (L2) component while the Bacillus diarrhoeal enterotoxin visual immunoassay (BDE kit; Tecra) detects mainly the NheA (45-kDa) protein (Granum and Lund, 1997).
Molecular techniques for detecting potentially enterotoxigenic organisms are also available as the genes encoding enterotoxin T, cytotoxin K and the subunits of the complexes of haemolysin (Hbl) and the non-haemolytic enterotoxin (Nhe) have been identified. PCR methods have been used to detect the bacT gene and the gene encoding the B. subunit of haemolysin BL (Granum et al., 1996; te Giffel et al., 1997).
2.8.2 B. cereus emetic toxin
2.8.2.1 Chemical structure and properties of emetic toxin, cereulide
The B. cereus emetic toxin was identified as a cyclic dodecadepsipeptide and named cereulide by Agata et al. (1994). The chemical structure of cereulide [cyclo(L-Ɉ-Val-L-
Val- D-Ɉ-Leu-D-Ala-)3] is very similar to that of valinomycin [cyclo(L-Val-D-HylVa-D-
ZXG Val- L-Lac-)3], a commercially available antimicrobial agent (Figure 2.2).
Figure 2.2 Structures of cereulide and valinomycin (Source: Suwan et al., 1995)
Generally, cereulide is more resistant to adverse conditions than the enterotoxins.
Cereulide is an extremely stable molecule, able to withstand heating at 121ºC for 90 min (Lund, 1990), treatment with acid (pH 2 with HCl for 30 min), alkali (pH 12 with
NaOH for 30 min) and the action of proteinase (Mikami et al., 1994). It is a small molecular mass compound, less than 5,000 Da molecular weight, passing through an ultrafiltration membrane (Suwan et al., 1995), though in a methanol extract, but not in water (Andersson et al., 1998; Shinagawa et al., 1995) because of its hydrophobic nature.
Cereulide causes mitochondrial swelling and this consequently causes damage to the cell (Agata et al., 1995b). The biochemical basis of this swelling has been suggested variously as activation of phospholipase A2 (Gunter and Pfeiffer, 1990), permeability transition stimulated by Ca2+ and prooxidants (Zoratti and Szabo, 1995), or ionophoretic uptake of K+ driven by the transmembrane potential (Mitchell and Moyle, 1969). As cereulide acts as a K+-specific ionophore, like valinomycin (Kuse et al., 2000) the latter mechanism is the most probable (Mikkola et al., 1999).
ZYG 2.8.2.2 Detection methods for cereulide
Oral challenge in primates
This method relies on observation of emesis that is generally induced within an hour after feeding toxin-laden materials to animal models, such as rhesus monkeys
(Shinagawa et al., 1995) and rodents (Suncus murinus) (Agata et al., 1995b).
Traditionally, feeding tests on monkeys were the most reliable type of assay, but this type of assay is not feasible nowadays because of the high cost of the animals for tests, technical difficulties in dealing with animals (Shinagawa, 1993) and ethical considerations. Furthermore, there is a problem of specificity and reliability in this detection method as Staphylococcus aureus toxins cause similar signs and duration of illness. Additionally, false-negatives with feeding test may occur, with Melling and
Capel (1978) noting that some rhesus monkeys, when given a standard emetic B. cereus rice culture, appeared to be insensitive to the toxin.
HEp-2 vacuolation assay
Hughes et al. (1988) found that the vacuolation in HEp-2 cells exposed to culture supernatants from an emetic-syndrome B. cereus isolate correlated with activity of emetic toxin. Cereulide affects the proliferation of HEp-2 cells (a human laryngeal- carcinoma cell line, Figure 2.3) and inhibits the growth of these cells, indicating that cereulide is cytotoxic (Mikami et al., 1994).
ZZG Figure 2.3 Electron micrographs of HEp-2 cells (A) A control cell incubated with skim milk. (B) A cell incubated with cereulide (Source: Agata et al., 1995b)
This assay relies upon microscopic examination to assess the proliferation of HEp-2 cells or the inhibition of the growth of the cells and it may generate unreliable data, as the method is quite subjective and because the mitochondrial swelling which is used as the diagnostic marker for the presence of toxin is transient and easily missed (Finlay et al., 1999).
Boar sperm motility inhibition assay
Cereulide affects boar spermatozoa by acting as an ionophore, leading to loss of motility through blockage of oxidative phosphorylation in the mitochondria. The effect can be assessed subjectively by ordinary phase-contrast light microscopy, and more objectively by determining the exact proportion of highly motile sperm cells, using a computerised sperm motility analyser (Andersson et al., 1998).
Z[G Figure 2.4 Thin sections of the middle piece of a boar spermatozoon exposed to cell extracts of B. cereus 4810/72 (documented producer of emetic toxin) and ATCC 14579T (type strain) (A) Mitochondrial damage in the middle piece of a spermatozoon exposed to cell extracts of 2 mg (wet weight) of cells ml-1 of strain 4810/72. (B) Middle piece of a spermatozoon exposed to cell extracts of B. cereus ATCC 14579T (2 mg [wet weight] ml-1). Mitochondria of ordinary size with intact membranes are seen. Bars, 200nm. (Source: Andersson et al., 1998)
MTT cytotoxicity assay
The MTT assay was developed for detection of cereulide and is currently the most widely used method amongst bioassays. It uses the tetrazolium salt, 3-[4,5- dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide (MTT) as a material for assessing the toxicity of cereulide toward cells. The yellow, water-soluble MTT is converted to an insoluble purple formazan by metabolising cells (Beattie and Williams,
1999; Finlay et al., 1999). However, as cereulide adversely affect the metabolic status of living cells, it reduces or prevents formazan formation. Therefore, the level of cytotoxicity due to cereulide can be assayed by a simple colour change by using spectrophotometric measurement.
Z\G After 3 – 4 h of cell incubation with MTT formazan crystals develop in living and early apoptotic cells. Dead cells don’t produce formazan crystals.
Figure 2.5 Principle of the MTT test (Source: Klipski et al., 2000)
2.8.2.3 Limitations of current bioassays for cereulide and the future
To date, three bioassays have been developed to overcome the disadvantages of the oral challenge method, but these currently practiced assays also have several limitations.
First, all bioassays are laborious and time consuming. These methods include many steps and it takes more than 24 h to perform, after isolation of cereulide. Second, these assays depend on the use of cell cultures that are not convenient to use and they bring with them the problem of maintenance due to dealing with living cells. Most importantly, these methods generally do not allow for providing accurate or precise analysis that will generate quantitative data.
Therefore, it is necessary to develop rapid and reliable detection methods that will generate quantitative data for analysis of cereulide and examine the relationship between B.cereus growth and toxin production in food. While toxicologically relevant substances may be detected by screening procedures, such as immunoassays or receptor assays (at present, Tecra International is undertaking development of an immunoassay for detection of cereulide), there is a need to exploit analytical methods that generate
Z]G real quantitative data such as chromatographic techniques. Recently, the quantitative analysis of cereulide, using high-performance liquid chromatography-ion trap mass spectrometry (HPLC-MS) was reported (Häggblom et al., 2002).
Molecular techniques for detecting potentially emetic toxin-producing organisms are also available now because two recent studies (Ehling-Schulz et al., 2004; Toh et al.,
2004) reported the genetic detection of emetic B. cereus utilising the polymerase chain reaction (PCR). Ehling-Schulz et al. (2004) hypothesized that the emetic toxin, cereulide is synthesised enzymatically by a NRPS and a PCR based screening assay using degenerate primers was targeted to known non-ribosomal peptide synthetase
(NRPS) sequences. With this approach, they identified a genomic DNA fragment that was specifically present in emetic toxin producing B. cereus strains. It is proposed that cereulide is produced non-ribosomally via a thio-template mechanism, since it had characteristics in common with other similar peptides such as valinomycin (Agata et al.,
1994) and microcystin which are synthesised by NRPS systems (Perkins et al., 1990;
Arment and Carmichael, 1996; Toh et al., 2004). Hence, Toh et al. (2004) had also chosen the possible analogy known as NRPS as an approach to detect and characterise the gene responsible for cereulide production in their study.
If these methods prove to be reliable and specific to the emetic gene of B. cereus, it will be extremely useful in specific isolation (such as detecting toxigenic strains of Bacillus) and characterisation of microorganisms. Nevertheless, the results of PCR may not represent the real health threat because even though the gene is present, the toxin may not be produced or not to the level required for human intoxication. Therefore, in order to prevent toxin-mediated food poisoning, quantitative determination of a population of
Z^G cells and/or the toxin present is important.
2.8.3 Toxins of other Bacillus species
Until recently, there has been little or no distinctive information on the identity of the toxins formed by other Bacillus species although these species have been associated with outbreaks of foodborne illness (Beattie and Williams, 2000). Thus, present detection methods for toxins of other Bacillus species use the same methods for detection of B. cereus toxins such as immunological assays, the boar spermatozoa motility test and cell cytotoxicity assays.
Griffiths (1990) reported that the non-B. cereus shown to produce a diarrhoeal toxin
(detected by RPLA) include Bacillus circulans, B. lentus, B. pumilus, B. polymyxa and
B. carotarum and, according to Beattie and Williams (1999), isolates of B. mycoides, B. thuringiensis, B. subtilis, B. lentus, B. circulans, B.lichenifomis and B. laterosporus/cereus produced molecules detected by immunological assays (BDE and
RPLA) for the diarrhoeal toxin of B. cereus.
Beattie and Williams (1999) also revealed that some isolates of B. circulans, B. laterosporus, B. lentus, B. licheniformis, B. mycoides, B. subtilis, B. cereus, and B. thuringiensis were toxigenic to Chinese hamster ovary (CHO) cells using the MTT assay. Rowan et al. (2001) reported that cell-free culture supernatants of 90 and 81% respectively of 47 Bacillus isolates including B. licheniformis, some of B. subtilis, B. brevis, B. circulans, B. firmus, B. thuringiensis, B. polymyxa, B. pumilus and B. sphaericus showed cytoxicity in Hep-2 and Caco-2 epithelial cells.
G
Z_G Recently, From et al. (2005) reported that eight strains, assigned to four different species, B. subtilis, B. mojavensis, B. pumilus, or B. fusiformis, produced cytotoxic or putative emetic toxin, and one strain (B. mojavensis) produced both cytotoxin(s) and putative emetic toxin when they were analysed by a cytotoxicity assay on Vero cells and the boar spermatozoa motility assay. Taylor et al. (2005) also demonstrated that strains of Bacillus cereus, Bacillus firmus, Bacillus megaterium, Bacillus simplex and Bacillus licheniformis produced heat-stable toxin.
G Collectively, the results of these methods for detection of B. cereus toxins showed that other Bacillus species may produce toxins similar to that of B. cereus and may also present a potential hazard in food products.
G
2.9 Capillary electrophoresis
Capillary electrophoresis (CE) is a relatively new analytical technique that provides for simple and rapid analysis with highly efficient separation (Issaq, 2000). It continues to offer effective alternatives to analytical methodologies currently in use for detection, quantitation and characterisation (Flurer, 1999).
2.9.1 Theory and some advantages of capillary electrophoresis (CE)
Capillary electrophoresis can be broadly described as providing high-efficiency electrophoretic separation of samples in narrow bore columns, of between 20 and 150
Pm internal diameter. Traditional electrophoretic techniques offer a low level of automation, long analysis times and detection of the separated molecules using post- separation visualisation. They also give only poor separation of organic molecules due to Joule heating that is generated during separation processes at high voltage (Lee and
Z`G Heo, 1991; Altria, 1996).
Performing electrophoretic separation in capillaries, in contrast, offers the possibility of automated analytical equipment, fast analysis times and on-line detection of separated peaks. Furthermore, heat generated inside the capillary can be effectively dissipated through the walls of the capillary, allowing high voltages to be used to achieve rapid separation (Cancalon, 1995; Altria, 1996; Camilleri, 1998). CE can be operated under different modes that increase the efficiency of separation for different classes of molecules. These modes include capillary zone electrophoresis (CZE), that is based on separation according to differences in charge (and, to some extent, size), micellar electrokinetic capillary electrophoresis (MEKC) based on separation of neutral compounds using surfactant micelles, capillary gel electrophoresis (GCE) based on sieving of solutes of different size through a gel network, and capillary isoelectric focusing (CIEF) based on separation of zwitterionic solutes within a pH gradient (Altria,
1996).
CE also has several advantages over chromatographic methods such as gas chromatography (GC), high performance liquid chromatography (HPLC), thin-layer chromatography (TLC) and slab gel electrophoresis (SGE). It is applicable to the separation of widely different compounds using the same instrument and, in most cases the same column, while changing only the composition of the running buffer (Shihabi and Friedberg, 1997), not readily achievable using other separation techniques. In addition, CE has the highest resolving power of any liquid separation (Shihabi and
Friedberg, 1997). The amount of sample needed for CE analysis is minute, nanolitres of sample and microlitres of buffer, compared to other liquid separation techniques that
[WG require, as a minimum, microlitres of sample and millilitres of solvent. These factors reduce operating costs and also hazardous wastes, compared to other methods.
2.9.2 Application of capillary electrophoresis (CE)
With high efficiency and great selectivity, CE has been applied successfully in diverse of research areas including, pharmaceutical research, environmental research, forensic research, and biological research, including protein, peptide and amino acid analysis,
DNA sequencing, gene mutation research and other genetic applications. With the general exception of lipids, most food components can be analysed by CE (Cancalon,
1995; Larger et al., 1998; Siriamornpun et al., 2001; Lange et al., 2002; Cataldi et al.,
2003; Huang et al., 2003) and many of the applications of the MEKC technique relate to food constituents (Thompson et al., 1995; Watanabe, et al., 1998; Aucamp et al., 2000;
Cifuentes et al., 2001).
Both the general CZE mode, with some additives, and the MEKC mode of CE have been applied to separation of hydrophobic compounds (Weinmann et al., 1994; Dong et al., 1997; Fürtös-Matei et al., 2000; Oudhoff et al., 2003). For example, MEKC has been used to analyse polycyclic aromatic hydrocarbons (Brüggemann and Freitag, 1995;
Fu et al., 1998) and Idei et al. (1998) reported that highly hydrophobic peptides could be analysed by this method. CE methods, using CZE and MEKC modes, have been evaluated for their suitability in systematic toxicological analysis with a test set of 25 barbiturates (Boone et al., 1999). Thus CE shows promise for the quantitative analysis of cereulide, a highly hydrophobic peptide.
G
[XG G Chapter 3
Development and application of a centrifugation- plating method for study of the biodiversity of Bacillus species in foods G
G G G G G G G G G G G G G G G G G G G G G G G G 3.1 Introduction G
G
Bacillus species are ubiquitous spore-formers and, as such, the endospores are able to withstand harsh environmental conditions and many processing methods. The majority of these Bacillus species are harmless contaminants. The major concern in the food industry is the outgrowth of pathogenic Bacillus species which have the potential to cause a food poisoning incident. Bacillus cereus is the most prevalent pathogenic
Bacillus species found in food and other industrial processes (Pirttijärvi et al., 1999).
However, other species in the genus, including B. subtilis, B. licheniformis, B. brevis, B. pumilus (Shinagawa, 1990; Lund, 1990; Salkinoja-Salonen, 1999) and B. thuringiensis
(Jackson et al., 1995; Beattie and Williams, 1999), have been implicated as potential agents of food poisoning.
Methods to identify toxigenic Bacillus species are important tools for a food industry that is always conscious of consumer food safety. Research on detection methods has traditionally centred on B. cereus, which is well known as a causative agent in foodborne disease. Conventional plating and biochemical methods for detection and enumeration of Bacillus species were developed mainly for B. cereus, and present limitations for ecological studies of Bacillus species. First, if the sample is contaminated with a low number of cells (<100 cfu/g), the low sensitivity of the method may not be able to detect the Bacillus cells. Second, because it detects bacilli from a small proportion of a sample, it may not represent all types of Bacillus species that are present in the total sample.
One approach, performed for the enumeration of Salmonella in chicken carcasses
43 (Hawa et al., 1984; Fleet et al., 1991) and Listeria in chicken carcasses, minced meat and cheese (Uyttendaele et al., 2000; Jang, 2004), could be applied to overcome the limitations of conventional plating methods presented above. This method utilises centrifugation of food homogenates to sediment and concentrate the microorganisms, which are then spread inoculated onto a selective-differential agar medium to detect and enumerate presumptive colonies (Hawa et al., 1984; Morrison and Fleet, 1985).
The advantages of this centrifugation-plating method are that it is significantly more sensitive than the spread plating method. Theoretically, this is because it utilises the whole sample rather than using part of, or a dilution of the sample. However, if the pelleted and sedimented food material is too voluminous after centrifugation, it may be difficult to spread plate. It is thus necessary to decrease the volume of particulate matter from the food in order to apply the centrifugation-plating method successfully.
Previously, attempts to overcome this problem involved hydrolysis of food homogenates with proteolytic enzymes and/or dispersion with surfactants, such as
Tween 80 (Fleet et al., 1991; McElroy et al., 1995; Sharpe, 2000; Jang, 2004). Previous studies have demonstrated that protease and surfactant treatments do not affect total viable counts (Peterkin and Sharpe, 1980; Pettipher et al., 1980; Entis et al., 1982) or the viability of vegetative cells such as Salmonella (Fleet et al., 1991; Karalis, 2000) or
Listeria (Hale et al., 1990; Gnanou Besse and Lafarge, 2001; Jang, 2004). To date, there has been no published study investigating the effect of sample pre-treatment upon the viability of Bacillus organisms.
Therefore, this chapter investigates the effect of commercially available amylases, in conjunction with Tween 80, on the viability of Bacillus organisms, when using the
44 centrifugation-plating method for examination of the diversity of Bacillus species in raw and cooked rice products.
G
3.2 Materials and methods G
3.2.1 Viability studies
3.2.1.1 Preparation of Bacillus cultures
The reference strains B. cereus NC 7401, B. cereus TICC 1994, B. thuringiensis FDA 2,
B. licheniformis TICC 368 and B. subtilis TICC 15 were incubated with enzyme and
Tween 80 to test the viability of Bacillus species after exposure to these agents. Bacillus cells were prepared by inoculating one loopful of biomass from a 24 h culture on
Tryptone Soya Agar (TSA) (Oxoid, Hampshire, England) into 10 mL of Tryptone Soya
Broth (TSB) (Oxoid), that was incubated for 24 h at 30ºC. This culture was used as a source of cells for treatments with enzyme and Tween 80.
G
3.2.1.2 Preparation of enzymes and surfactant
Two commercially available amylases were examined. The sources of these enzymes and some of their characteristics are shown in Table 3.1. The enzymes were prepared as a 10% solution in sterile 1.0 M Tris buffer (Sigma, pH 7.2) before use. Bacillus species were not detected in either enzyme preparation when plated onto PEMBA or
TSA. Tween 80 (Sigma, Sydney, Australia) was prepared as a 10% solution in distilled water and autoclaved at 121ºC for 15 min.
45 Table 3.1 Source and characteristics of amylases Characteristics Amylase A Amylase B Source Sigma-Aldrich Fine Chemicals, Novozymes A/S, Castle Hill, NSW 1765 Krogshoejvej 36, Australia DK-2880 Bagsvaerd, Denmark Type Ȼ-Amylase Ȼ-Amylase derived from Bacillus globigii derived from (Bacillus licheniformis) Bacillus amyloliquefaciens
Properties pH optimum 6.9, pH optimum 5 – 7, temperature 20ºC optimum temperature 70 – 90ºC
Application One unit will liberate 1.0 mg of One unit is the amount of maltose from starch in 3 min enzyme which dextrinises under standard condition 5.26 g dry starch per hour under standard condition
3.2.1.3 Effect of enzyme and surfactant on the viability of Bacillus species
The Bacillus culture, prepared as described previously (Section 3.2.1.1) was diluted 1 in 10 with either Buffered Peptone Water (BPW, Oxoid) alone, or including amylase, and/or Tween 80 to a total volume of 10mL. The four treatments used are shown in
Table 3.2.
Table 3.2 Preparation of treatments for viability studies of Bacillus species Treatments Composition of treatments i) Control Culture (1 mL) + dilution medium (9 mL) ii) Amylase treatment Culture (1 mL) + dilution medium (8 mL) + amylase solution (1 mL) iii) Tween 80 treatment Culture (1 mL) + dilution medium (8 mL) + Tween 80 solution (1 mL) iv) Mixture of amylase Culture (1 mL) + dilution medium (7 mL) + Tween 80 and Tween 80 treatment solution (1 mL) + amylase solution (1 mL)
46 The mixtures were incubated separately in 10 mL test-tubes at ambient temperature for
1 h with Amylase A and at 40ºC for 1 h with Amylase B. After the incubation period, a 1 mL aliquot of each treatment was taken, serially diluted 1 in 10 with 9 mL of BPW and then examined for viable cell growth by spread inoculation (100 PL) onto plates of TSA and PEMBA. The plates were incubated at 30ºC for 24 h and colonies enumerated.
Viable cell counts were done in triplicate and the mean values calculated. The recovery was calculated by comparing the counts after enzyme and/or Tween 80 treatment to that of the control. Recovery (%) was determined as followed:
Number of colonies on plates after treatment Recovery(%) u100 Number of colonies on control plates
The t-test was used for statistical analysis of data of the viability tests, using the
SigmaStat 2.03 program.
G
G 3.2.2 Recovery of inoculated B. cereus by centrifugation-plating
3.2.2.1 Preparation of rice samples
After several raw (uncooked) rice samples had been tested, one that was found to be uncontaminated with viable microorganisms was chosen as the matrix for recovery tests of B. cereus in raw rice. Without the addition of any bacteria, this also represented the negative control for every recovery test experiment.
In the preparation of cooked rice samples, two volumes of water to one volume of raw rice were combined, then autoclaved in a glass beaker for 20 min at 121ºC to simultaneously sterilise and cook the rice (McElroy et al., 2000).
47 3.2.2.2 Recovery of B. cereus inoculated in raw rice samples
The raw rice sample (10 g) was transferred into a sterile Stomacher bag and inoculated with B. cereus NC 7401 at 10 – 20 cfu and 100 – 200 cfu per 10 g of rice. Inoculated samples were left at ambient temperature (20 – 25ºC ) for 30 min to allow Bacillus cells to attach and absorb to the surface of the raw rice, then mixed with 81 mL of sterile
BPW, and 9 mL of a 10% solution of Tween 80 in distilled water. The mixture was incubated at ambient temperature for 30 min to facilitate the dispersion of food homogenates by Tween 80. Following incubation, the mixture was homogenized by stomaching for 2 min and then centrifuged at 15,000 × g for 12 min at 4ºC using a
Beckman model J2-MC centrifuge in a sterile centrifugation bottle. The supernatant was carefully discarded, and the pelleted material resuspended in 1 mL of sterile BPW by vortex mixing. The entire suspended material was spread inoculated over the surface of a large (150 × 20 mm) plate of PEMBA (approximately 90 mL of agar in a plate), which was then incubated aerobically for 48 h at 30ºC.
3.2.2.3 Recovery of B. cereus inoculated into raw rice homogenate fluids
This experiment was carried out as described previously (Section 3.2.2.2) with the exception that Bacillus cells were inoculated into the raw rice homogenate, rather than directly onto the raw rice samples in order to assess the degree of absorption of Bacillus cells to rice. After the raw rice was homogenised by stomaching with BPW, diluted
Bacillus culture (1 – 2 × 10 or 1 – 2 × 102 cells) was inoculated into the homogenate.
The homogenate was analysed as described previously in Section 3.2.2.2.
48 3.2.2.4 Recovery of Bacillus species in cooked rice samples
A number of methods were assessed to determine recovery of B. cereus (NC 7401) inoculated into cooked rice samples:
G i) Cooked rice (10 g) was mixed with B. cereus cells as previously described in Section
3.2.2.2. The inoculated sample was left at ambient temperature for 30 min to allow cells to attach and adsorb. The inoculated cooked rice was mixed with 81 mL of BPW and 9 mL of a 10% solution of Tween 80 in a filter-Stomacher bag (John Morris Scientific,
Sydney Australia) and homogenised by stomaching for 2 min. The homogenate was then assayed for Bacillus by the centrifugation-plating method previously described in
Section 3.2.2.2.
ii) Cooked rice (10 g) was inoculated with B. cereus cells (Section 3.2.2.2) and mixed with 72 mL of sterile BPW, 9 mL of a 10% solution of Tween 80 in distilled water and
9 mL of a 10% solution of Amylase A in Tris buffer (pH 7.2). The mixture was then homogenised by stomaching for 2 min. It was then analysed for Bacillus by the centrifugation plating method as described in Section 3.2.2.2.
iii) Cooked rice (10 g) was mixed with cells of B. cereus described as Section 4.2.2.2.
The inoculated sample was left at ambient temperature for 30 min to allow cells to attach and adsorb. The inoculated cooked rice was mixed with 80 mL of BPW, 9 mL of a 10% solution of Amylase A in Tris buffer (pH 7.2) and 1 mL of a 10% solution of
Tween 80 in a filter-Stomacher bag. The mixture was incubated at ambient temperature for 30 min for the enzyme reaction and then stomached for 90 s. The homogenate was then assayed for Bacillus by centrifugation-plating method as described in Section
49 3.2.2.2.
Method (iii) was chosen for recovery tests from cooked rice samples using the centrifugation-plating method.
3.2.3 Application of the centrifugation-plating method
3.2.3.1 Rice samples
All raw and the majority of cooked rice samples were purchased from several food outlet shops to give a range of representative brands (Appendices 1 and 2). Some of the cooked rice samples were prepared in the laboratory through the cooking of raw rice in a commercially available rice cooker. Raw rice was stored at ambient temperature until used and cooked rice was kept at 4ºC for a period of no longer than 4 h before analysis.
Cooked rice samples, which had been chilled, were equilibrated to ambient temperature
(20 – 25ºC) before analysis.
3.2.3.2 Detection and enumeration of Bacillus species on raw rice
Raw rice samples (10 g) were homogenised with 81 mL of BPW and 9 mL of 10%
Tween 80 for 2 min and the homogenates were incubated at 35ºC for 30 min to allow any injured cells to resuscitate (Hawa et al., 1984; Jang, 2004) and for the release of cells by Tween 80. The homogenate was then centrifuged and the sedimented material resuspended as described in Section 3.2.2.2. The resuspended material was spread- inoculated over a large plate of PEMBA (150 × 20mm).
50 The inoculated plates were incubated aerobically at 30ºC and the total number of presumptive Bacillus colonies was enumerated at 24 and 48 h. Confirmation as Bacillus was determined according to phenotypic characteristics being different colony morphologies on the plate and rod shaped cells with an endospore, by phase-contrast microscopy (Olympus, Olympus Australia Pty. Ltd, NSW, Australia). Several presumptive colonies (2 – 3 colonies of each morphology type) were selected from each
PEMBA plate. The selected single colonies were subcultured (streaked) onto TSA and subcultured two more times to obtain pure cultures. Colonies representative of a pure culture were identified by partial sequencing of 16S rDNA (Section 3.2.5). The Bacillus isolates were maintained on TSA slants at 4ºC or in TSB with 20% glycerol at -80ºC.
G G 3.2.3.3 Detection and enumeration of Bacillus species in cooked rice
The detection and enumeration of Bacillus species in retail cooked rice samples were carried out as described previously (Method (iii) in Section 3.2.2.4) without the inoculation of Bacillus cells. Presumptive Bacillus colonies from each sample were counted and identified described as in Section 3.2.3.2.
3.2.4 Spread plating method for detection of Bacillus in foods
Samples were examined for Bacillus simultaneously by both the centrifugation plating method and the spread plating method for all potential Bacillus species based on
‘Enumeration of Bacillus cereus and other Bacillus species’, issued by the Standards unit (2002), Evaluation and Standard Laboratory, Health Protection Agency (London,
UK). The procedure is depicted in Figure 3.1.
51 Homogenise and dilute rice sample p Inoculate 0.1 mL of each dilution onto duplicate plates of PEMBA p Incubate aerobically at 30qC and examine after 24 and 48 h p Examine for absence of mannitol fermentation and presence of lecithinase (B. cereus/B. thuringiensis) Examine for typical (visual observation on plate) colony morphology (Bacillus species) p Enumerate typical colonies: total Bacillus species count and the B. cereus count p Confirm putative Bacillus identification with microscopy
G
Figure 3.1 Procedure for the detection and enumeration of Bacillus species
3.2.5 Identification of Bacillus species by partial sequencing
3.2.5.1 Phenol-Chloroform extraction of DNA
To obtain high purity DNA, a phenol-chloroform DNA extraction method (Cocolin et al., 2001) was used with slight modification, in that the incubation procedure with sucrose solution containing lysozyme was excluded because high yields of DNA of
Bacillus species could be obtained without this procedure. Cultures of Bacillus species were grown for 24 h in 10 mL of TSB at 30qC and a 1 mL aliquot containing Bacillus cells pelleted by centrifugation (12,000 u g for 10 min). The supernatant was decanted, and the cell pellet resuspended in 0.5 mL of cold breaking buffer, containing 470 PL of cold TE buffer (50 mM Tris-HCl and 5 mM EDTA, pH 8.0), 10 PL of 10% sodium
52 dodecyl sulphate (SDS) and 20 PL of 5 M NaCl. Silica beads (0.3 g of 0.1 mm zirconia,
BioSpect Product, Inc. Bartlesville, OK 74005) and 0.5 mL of phenol-chloroform- isoamyl alcohol (25:24:1) were then added and the combination shaken in a bead beater
(BioSpect Product, Inc. Bartlesville, OK 74005) for 50 s at maximum speed (50 rpm).
The mixture was then centrifuged at 12,000 u g for 10 min at 4qC and the aqueous phase collected. The organic extraction procedure was repeated once more, and then a
0.25 mL volume of isopropanol was added to the purified aqueous phase to assist in the precipitation of total genomic DNA. The tubes were kept at -20qC for 2 h then centrifuged at 12,000 u g for 15 min at ambient temperature. The supernatant was decanted to leave the DNA pellet and the pellet washed with 70% ethanol (Aldrich,
Sydney, Australia). It was centrifuged again at 12,000 u g for 10 min at ambient temperature, and the supernatant was decanted and any residual ethanol evaporated by air-drying. The DNA pellet was then dissolved in 50 PL of sterile Milli-Q water
(Millipore, Bedford, MA, USA)
.
3.2.5.2 PCR amplification for partial sequencing of 16S rDNA
PCR amplification of DNA from bacterial cultures was carried out as described by
Nübel et al. (1996) and Garbeva et al. (2001) using the universal primers 968F (5’-
AACGCGAAGAACCTTAC-3’) (Eubacteria V6) (Heuer et al., 1999) and 1401R (5’-
CGGTGTGTACAAGACCC-3’) (Eubacteria V9) (Nübel et al., 1996). PCR amplification was performed in a reaction volume of 50 PL containing: 1 u PCR buffer
(10 mM Tris-HCl [pH 8.3], 50 mM KCl), 0.2 PM of each primer, 200 PM of each dNTP,
1.5 mM MgCl2, 1.25 U of Taq DNA polymerase and 10 ng of extracted DNA. The PCR was performed using a DNA thermal cycler (Gene Amp PCR system 9700) (Applied
53 Biosystems, California, USA). The temperature cycles comprised initial denaturation at
94qC for 2 min, followed by 10 cycles at 94qC for 30 s, 50qC for 30 s, and 72qC for 45 s, followed by 20 cycles using the same temperature but with each successive cycle using a 5 s longer elongation time, then a final extension step at 72qC for 7 min. PCR amplicons were visualised by gel electrophoresis in 1% (w/v) agarose gels in TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA) and PCR products of interest were used for sequence analysis.
G 3.2.5.3 Sequence analysis
Sequencing was carried out with the ABI PRISM´ BigDyeTM Terminators v3.1 Cycle
Sequencing Kit (Applied Biosystems, Forster City, CA, USA) with the primer 968F as a sequence primer and submitted to Angela Higgins at the Automated DNA analysis facility, School of Biotechnology and Biomolecular Sciences, UNSW, for processing.
The NCBI-BLAST (Basic Local Alignment Search Tool) program was used to conduct genetic similarity searches against the GenBank sequence database following generation of sequencing data. The data were used when more than 97% of similarity was obtained.
G
G
3.3 Results and discussion G
3.3.1 Viability studies of Bacillus species in model food system (medium)
3.3.1.1 Effect of dilution medium
After several preliminary tests with peptone water and buffered peptone water, buffered peptone water was chosen as a dilution medium because it gave more protection to the
54 cells (5 – 8% more recovery, based on triplicate experiments).
3.3.1.2 Effect of amylase and surfactant on the viability of Bacillus species
The use of amylase was required, particularly for cooked rice, to catalyse the breakdown of starch in the rice slurries, to assist the release of bacilli and to facilitate plating after centrifugation. Two commercially available amylases (listed in Table 3.1) were tested for their suitability with regards to hydrolysing polysaccharides such as starch. Amylase A (Sigma) with Bacillus culture was incubated at ambient temperature according to the manufacture’s instructions as being the optimum temperature for effective enzyme activity. Amylase B (Novozyme) has maximum activity at 75qC but this is too high a temperature to apply to Bacillus species generally. Therefore, 40qC was chosen for digestion as 70% of the enzyme activity was retained while this temperature did not impact negatively on Bacillus species according to data from the supplier.
The survival (%) of Bacillus (two different strains of Bacillus cereus and strains of other
Bacillus species) after different treatments with the two amylases and Tween 80 is shown in Table 3.3. In most cases, treatment with enzyme and Tween 80 did not have a significant adverse effect on cell viability. However, the viability of emetic B. cereus and B. subtilis strains decreased significantly with amylase A and amylase B-Tween 80 treatments respectively on selective-differential agar (PEMBA) (P<0.05). The same decreases were not observed on non-selective media (nutrient agar, TSA). Although significantly different to the untreated controls, recoveries still exceeded 80%.
55 Table 3.3 Effect of amylase and Tween 80 treatments on the viability of Bacillus cells as determined by plating onto TSA and PEMBA
Strain Plating Recovery (%) medium Amylase A Amylase B Tween 80 Tween 80 Tween 80 Tween 80 (25ଇ) (40ଇ) + amylase A + amylase B B. cereus (emetic) TSA 90.2±7.6 94.2±4.5 106.8±6.1 91.1±9.7 96.7±16.1 98.0±12.2 PEMBA 89.3±2.1a 99.5±8.3 88.3±8.5 85.4±14.1 98.9±11.9 91.6±7.4 B. cereus (diarrhoeal) TSA 99.0±2.8 102.8±10.6 93.3±11.7 81.5±9.1a 94.1±8.3 102.9±7.8 PEMBA 93.5±9.2 92.2±8.4 95.9±9.7 89.8±9.6 94.9±3.2 101.2±2.6 B. thuringiensis TSA 93.7±11.4 94.3±6.1 89.8±9.2 97.6±2.6 96.9±5.5 93.9±9.7 PEMBA 102.6±12.4 89.3±17.9 103.9±13.2 91.2±9.3 96.6±11.9 96.6±17.3 B. subtilis TSA 101.6±13.9 95.0±9.8 99.9±4.7 99.2±1.7 97.3±5.0 100.1±3.6 PEMBA 97.3±3.8 99.9±8.6 95.1±3.6 95.7±4.5 96.7±3.0 89.8±0.3a B. licheniformis TSA 94.9±5.4 94.4±16.0 106.2±12.9 88.6±8.0 91.9±7.6 93.0±7.2 PEMBA 93.5±9.2 92.0±7.0 102.2±5.1 95.3±19.3 95.2±6.6 93.1±19.4 Recovery (%); mean of recovery (%) ± standard deviation a indicates counts were significantly different from the control at the 5% level, Control was taken as 100%. Data are the mean values of 3 independent experiments for all Bacillus strains.
56 Consequently, amylase A and Tween 80 were chosen for use in digesting food samples for the centrifugation-plating method. This was because this treatment did not exert a significant effect on the viability of B. subtilis on PEMBA and only one strain of B. cereus was significantly affected by the treatment of amylase A alone (P<0.05). In contrast, the viability of B. subtilis was significantly decreased using amylase B-Tween
80 (P<0.05).
G G 3.3.2 Recovery of inoculated B. cereus by centrifugation-plating
3.3.2.1 Recovery of inoculated B. cereus from raw rice
The recovery of 101 or 102 cfu/10 g of B. cereus emetic strain 7401 from raw rice was investigated using the centrifugation-plating method. Initially, the recovery of Bacillus cells in raw rice was examined with and without Tween 80 treatment and using a variety of stomaching times. According to these preliminary tests, the recovery of cells was low
(50 – 63% with five replicates) without Tween 80. However when the cells are inoculated into the homogenates just before centrifugation, the recovery of cells was significantly higher (89 – 94% with three replicates). This suggested a problem with detachment of cells from rice rather than sedimentation of cells in homogenates by centrifugation. Due to the hydrophobic nature of B. cereus spores, they adhere tightly to surfaces (Husmark and Rönner, 1993) and vegetative cells, especially cells in the late stationary growth phase, are also hydrophobic (Peng et al., 2001b). Therefore, in order to facilitate detachment of cells from the raw rice, samples were treated with Tween 80 for 30 min at ambient temperature. Additionally it was found that detachment of cells from raw rice samples was also affected by stomaching time. Increasing the stomaching time from one to two minutes increased the recovery of cells.
57 Table 3.4 depicts the recovery of B. cereus inoculated onto raw rice at 1 – 2 × 101 cfu and 1 – 2 × 102 cfu per 10 g of rice. The mean recovery (%) of inoculated B. cereus was
86.3 ± 9.1 % in raw rice samples with Tween 80 treatment and 2 min stomaching time.
The rate was higher with samples that were inoculated with 1 – 2 × 102 cells and the recovery approached 90%. This recovery rate is not achievable with the conventional dilution-spread plating method. Theoretically, spread plating will detect only 1 or 2 colonies, even if 1 mL of a 10-1 dilution of the sample, containing 1 – 2 × 102 cfu per 10 g of sample, is used (see Figure 3.1).
G Table 3.4 Recovery rate of the centrifugation-plating method to the enumeration of Bacillus cereus inoculated onto the surface of raw rice samples
Sample Cells inoculated (cfu/10g) Recovery (%) 1 1 – 2 × 102 75.3 2 1 – 2 × 102 88.0 3 1 – 2 × 102 87.2 4 1 – 2 × 102 100.2 5 1 – 2 × 102 91.3 6 1 – 2 × 102 92.3 7 1 – 2 × 102 100.7 8 1 – 2 × 101 80.0 9 1 – 2 × 101 75.0 10 1 – 2 × 101 85.7 11 1 – 2 × 101 88.8 12 1 – 2 × 101 71.8 13 1 – 2 × 101 93.7 14 1 – 2 × 101 78.6
Mean 86.3 ± 9.1 Samples, from 1 – 14, were inoculated with B. cereus NC 7401 strain. The mean of recovery (%) was 90.7 ± 8.7 at inoculation with 1 – 2 uG102 cells and 81.9 ± 7.8 at inoculation with 1 – 2 uG101 cells.
58 These experiments show that centrifugation plating can be used successfully to recover
Bacillus cells for study of the ecology of Bacillus species in raw rice samples. In particular, the main advantage presented is successful cell recovery even if the samples are contaminated with a low number of cells.
G
3.3.2.2 Recovery of inoculated B. cereus from cooked rice treated with amylase and the surfactant Tween 80
The recovery of B. cereus emetic strain 7401 from cooked rice samples was assessed using the centrifugation-plating method. For the cooked rice, an internal filter mesh within a sterile Stomacher bag was used to separate the cooked rice material from the liquid suspension. However, it was found that the suspension of cooked rice after stomaching still contained a significant amount of particulate matter after centrifugation.
A Stomacher bag with a fine filter mesh in it (K & R technologies, Frederick, USA) was also tried and the resulting sediment of cooked rice was still similar in volume to that using a normal filter mesh. Therefore, amylase treatment using D-amylase from two commercial suppliers (Sigma-Aldrich, NSW, Australia and Novozyme, Bagsvaerd,
Denmark) was assessed for reduction of cooked rice sediment. It was found that both amylases reduced cooked rice sediment, though its activity was not sufficient to produce a dramatic reduction. Of the two enzymes, amylase A (Sigma-Aldrich) was chosen because the treatment of amylase A-Tween 80 showed no significant impact upon the viability of all tested Bacillus species (P<0.05) (Section 3.3.1.2) and the amylase works at ambient temperature, whereas the optimum temperature of amylase B is 70 – 90ºC, the temperature required for activity of most Ȼ-amylases (Outtrup and Jørgensen, 2002).
59 Appropriate selection of enzyme/s and surfactant, which do not affect the viability of target organisms, has been shown to be critical to successful application of the centrifugation-plating method. Of the available surfactants, Tween 80 has been proven to be the most effective (Sharpe and Harshman, 1976; Emswiler et al., 1977; D'Aoust et al., 1982; Dickson, 1990; Hale et al., 1990; Gnanou Besse and Lafarge, 2001), so it was selected for this study.
As a result of the preliminary test with cooked rice, it was found that recovery was not affected by the use of Tween 80 at a concentration between 0.1% and 0.01% or by the duration of stomaching, whereas the volume of sedimented particles of cooked rice was greater when the stomaching time was longer. Therefore, the volume of Tween 80 and the stomaching time required was decreased for cooked rice samples, as compared with that for raw rice.
The recovery of B. cereus from cooked rice samples was higher than that for raw rice and slightly exceeded 100% (Table 3.5). Because the sample was prepared by autoclave, there was no possibility of release of indigenous bacilli from the sample of cooked rice.
Tween 80-mediated separation of cells in colony forming units (cfu) was also discounted as no difference was found between two samples that were tested with or without Tween 80. The use of higher concentration of Tween 80 was not shown to have any effect on recovery (Section 3.3.2.1). Therefore, this might be due to the growth of B. cereus, as the bacterium has a generation time of 20 – 30 min at 35qC (Notermans and
Batt, 1998) and some strains have generation times of 59 min in cooked white rice at
25qC (Penna et al., 2002). However, B. cereus 7401 was shown not to grow within 1 h incubation at ambient temperature in BPW, the conditions under which the initial
60 recovery tests were conducted. In addition, B. cereus NC 7401 did not grow during the
30 min incubation period for amylase treatment. However, if samples were incubated with amylase and Tween 80 for more than 1 h at ambient temperature, the recovery of B. cereus increased to around 200%, almost certainly due to the growth of B. cereus cells during the 30 min incubation in cooked rice for attachment and absorption of cells and the 1 h incubation in BPW medium for amylase digestion. Consequently, the incubation time with enzyme was limited strictly to 30 min.
Table 3.5 Recovery rate of the centrifugation plating method for the enumeration of Bacillus cereus inoculated onto cooked rice samples
Sample Cell inoculated (cfu/10g) Recovery (%) 1 1 – 2 × 102 117.0 2 1 – 2 × 102 100.0 3 1 – 2 × 102 116.6 4 1 – 2 × 102 101.3 5 1 – 2 × 102 102.9 6 1 – 2 × 102 92.9 7 1 – 2 × 102 102.5 8 1 – 2 × 101 120.0 9 1 – 2 × 101 106.0 10 1 – 2 × 101 98.5 11 1 – 2 × 101 93.6 12 1 – 2 × 101 130.0 13 1 – 2 × 101 121.5 14 1 – 2 × 101 124.0
Mean 109.0 ± 12.1 Samples, from 1 – 14, were inoculated with B. cereus NC 7401 strain. The mean recovery (%) was 104.7 ± 8.9 at 102 cells inoculation and 113.4 ± 13.9 at 101cells inoculation.
61 3.3.3 Application of the centrifugation-plating method for examination of Bacillus species in rice products
3.3.3.1 Detection and enumeration of Bacillus on raw and parboiled rice products
The centrifugation-plating method was applied to determine the biodiversity of Bacillus species in raw rice samples. Thirty samples of raw rice, and five samples of parboiled and dried rice, purchased from ten different retail stores in the Sydney region (Appendix
1), were tested.
PEMBA was used as the selective agar for the detection and enumeration of Bacillus species as it is the standard medium for detection of B. cereus in Australia at the time of performance of these experiments. As a selective agar, PEMBA is supplemented with polymyxin B sulphate solution (1000 units/mL) to inhibit the growth of Gram-negative bacteria and egg yolk emulsion to serve as a substrate for lecithinase activity of B. cereus. Fermentation of D-mannitol is a further differential system in the medium. B. cereus was confirmed to have resistance to the antibiotic, polymyxin B and other
Bacillus species including B. thuringiensis, B. mycoides, B. licheniformis, B. subtilis, B. polymyxa, B. alvei and B. brevis were also shown to be able to grow on PEMBA
(Holbrook and Anderson, 1980; Szabo et al., 1984; Jenson, 2000). In this study, the following Bacillus species, B. amyloliquefaciens, B. subtilis, B. licheniformis, B. circulans, B. megaterium, B. mycoides, B. pumilus and B. sphaericus were all shown to grow on PEMBA. These toxigenic or potentially toxigenic Bacillus species are target organisms of the current study, and therefore PEMBA was chosen as the medium for isolation of Bacillus species instead of a non-selective nutrient medium such as TSA, to give selectivity, thus preventing overgrowth of organisms other than bacilli.
62 Even though typical B. cereus colonies change the colour of PEMBA from green to blue because of a lack of mannitol fermentation, it was noted that the raw rice slurry may have interfered with the colour change and thus colony recognition and detection.
However, B. cereus colonies were still quite obvious, even though the colour change in the medium was not fully developed, because of the typical morphology of B. cereus colonies (crenate to fimbriate and 3 – 5 mm diameter), surrounded by a zone of precipitation (Figure 3.2). This development of colonies was observed consistently on
PEMBA plates inoculated with centrifuged rice pellet but not with inocula from conventional dilution and spread plating. In contrast to B. cereus, B. subtilis and B. licheniformis are able to ferment mannitol and thus the species produce flat colonies about 3 mm in diameter and green to grey-green in colour, as previously reported
(Jenson, 2000). Interestingly, most colonies of B. subtilis (identified by partial DNA sequencing) on PEMBA from the rice pellet did not appear to ferment mannitol, producing a blue colour on PEMBA (Figure 3.2). Unlike B. cereus, there was no lecithinase activity associated with the B. subtilis colonies. Confirmation and identification of some of the Bacillus isolates to species level were sometimes problematic, as they showed atypical characteristics for a number of the physiological and biochemical tests (Rowan et al., 2001). Rowan et al. (2001) found that the clinical isolates B. licheniformis KD1 and KD8 and B. pumilus produce lecithinase, even though these species were previously reported to be unable to produce this phospholipase.
63 G Presumptive colonies of Bacillus cereus
(A)
Figure 3.2 Colony development on PEMBA plates spread inoculated with centrifuged sediments of homogenised raw and cooked rice samples and incubated 30oC for 48 h (A) Raw rice sample; (B) Cooked rice sample
64 Presumptive Bacillus from PEMBA were confirmed and identified by subsequent partial
DNA sequencing as belonging to a number of Bacillus species or groups, including B. cereus/B. thuringiensis, B. subtilis/B. mojavensis, B. licheniformis, B. pumilus and B. sphaericus/B. fusiformis and other bacilli such as B. mycoides, B. megaterium and
Paenibacillus species. These results highlight the efficacy of the chosen medium,
PEMBA, in recovering a range of bacilli
As mentioned previously, the centrifugation-plating method is a highly sensitive approach to detection compared to the standard spread plating method. Bacillus cells were detected in 34 out of 35 samples with the centrifugation-plating method whereas only 13 samples were found to contain bacilli by the spread plating method. To increase the sensitivity of the spread plating method, 1 mL of the 10-1 dilution of the sample was inoculated onto five normal-sized (90 mm) PEMBA plates as a first dilution.
Table 3.6 shows the distribution of Bacillus species on raw rice samples. Long grain rice and medium grain rice were found to contain low numbers of B. cereus and B. subtilis.
Of the 35 samples examined, jasmine rice sample 11 and parboiled rice sample 34 were each found to contain five different Bacillus species. One sample of long grain rice contained Paenibacillus, identified as P. amylolyticus by partial sequencing.
65 Table 3.6 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus on raw rice samples
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 1 Long grain 10 34 - 11 2 ------2 Long grain 10 6 13 ------3 Long grain 3 1 3 2 - 1(Pa)g ------4 Long grain 6 2 3 ------5 Long grain 20 31 ------6 Long grain 6 ------7 Long grain 6 14 - - 2 ------8 Jasmine long grain - - 1 ------9 Jasmine long grain 1 24 - - 5 ------10 Jasmine rice 395 38 - - - - 30 - - - - - 11 Jasmine rice 525 (10) 102 49 3 - - 70 50 30 - 10 - 12 Bas. aro. long grain* TNTC 76 - 100 - - 440 60 - 70 - - 13 Basmati long grain 189 163 - 290 - - 14 40 30 - - aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species; g(Pa), Paenibacillus amylolyticus. - , Not detected; *Bas. aro. long grain, Basmati aromatic long grain
66 Table 3.6 Comparison of centrifugation plating and spread plating methods for the detection and enumeration of Bacillus on raw rice samples (continued)
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 14 Basmati long grain 15 34 - 9 ------15 Basmati long grain 366 117 - 18 - - 60 50 - - - - 16 Basmati wild rice 160 7 - 11 - 11 (Bf)h 100 10 - - - - 17 Wild long grain 4 1 - 2 - 2 (Pa)g ------18 Wild rice 7 5 - 15 ------19 Brown long grain 16 64 ------20 Medium grain 4 9 5 ------21 Medium grain 1 (1) 5 10 ------22 Meditation medium grain 2 23 - 5 - - - 10 - - - - 23 Brown medium grain 8 ------24 Risotto 13 (2) 2 39 - - - 10 - - - - - 25 Risotto 11 TNTC - - TNTC - - 2700 - - 1800 - 26 Risotto 7 20 - - 5 ------aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species; g(Pa), Paenibacillus amylolyticus; h(Bf), B. fusiformis/B. sphaericus - , Not detected
67 Table 3.6 Comparison of centrifugation plating and spread plating methods for the detection and enumeration of Bacillus on raw rice samples (continued)
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 27 Japanese style sushi rice 54 181 - 3 ------28 Glutinous rice ------29 Broken rice 18 12 ------30 Broken rice 47 10 - 2 - - 20 - - - - - 31 Parboiled rice 13 TNTC - TNTC - - - 6750 - 1000 - - 32 Parboiled rice 3 ------33 Parboiled rice TNTC TNTC - - - - 110 220 - - - - 34 Parboiled rice 5 182 26 98 - 9 (Bf)h -90-30-- 35 Parboiled rice - TNTC - TNTC -- - 100 - 980 30 - aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species; h(Bf), B. fusiformis/B. sphaericus - , Not detected
68 Aromatic raw rice samples such as jasmine rice and basmati rice were generally found to contain a greater number and variety of Bacillus than the plain rice varieties. More than half the aromatic varieties of raw rice harboured B. cereus at levels between 100 and 500 cfu/10 g. Wild rice, a black-coloured raw rice, was not found to contain a high number of microorganisms in terms of either other bacteria or Bacillus species, while brown rice (samples 19 and 23) harboured a high number of other bacteria even though the numbers of Bacillus species were low (8 – 80 cfu/10 g). Five samples of parboiled and dried rice (samples 31 – 35) were also tested. These samples were found to contain noticeably higher populations of Bacillus than non-processed (plain) raw rice.
In some cases, the population data between the centrifugation and spread plating methods did not correlate well. This is likely to be due to the use of different proportions of samples between the two methods. In the spread plating method, only small sub-samples of the dilutions were used and the population data are much more of an estimation than that from the centrifugation-plating method. Thus, the number of colonies from the spread plating method perhaps is inaccurate, derived from what might be a poorly representative sub-sample of the first dilution. This is further supported by the fact that, for some samples, no bacilli were obtained using the conventional spread- plating method while detected using the centrifugation-plating method. That is, the latter method detected bacilli that were obviously there, when the conventional spread plating method failed.
A perceived disadvantage of the centrifugation-plating method is that colonies on the plates cannot be enumerated when the food samples are significantly contaminated with other microorganisms (>103). However, the problem of excessive background flora, or
69 even excessively high numbers of bacilli, could be overcome by spreading smaller aliquots of the resuspended pellet over more than one plate.
Table 3.7 shows the populations of B. cereus and the other Bacillus species on raw rice samples. Of the 35 samples, 71.4% contained 1 – 102 cfu/10 g of B. cereus and 62.8% of samples contained the same numerical range of colonies of other Bacillus species.
Three of 35 samples (8.6%) were found to contain more than 104 cfu/10 g of Bacillus species.
Table 3.7 Number (proportions, %) of raw rice samples with different populations of Bacillus species
Bacillus Bacillus Plating method B. cereus (cfu/10 g of raw rice) species Centrifugation None 1 (2.9) 3 (8.6) Centrifugation 1 – 100 22 (62.8) 25 (71.4) Centrifugation 101 – 500 4 (11.4) 4 (11.4) Centrifugation 501 – 1000 3 (8.6) 1 (2.9) Spread1001 – 10000 2 (5.7) 2 (5.7) Spread>10000 3 (8.6) 0 (0) Total 35 (100) 35 (100)
3.3.3.2 Detection and enumeration of Bacillus on cooked rice products
Cooked rice samples were examined using both the centrifugation-plating and conventional spread plating methods, in order to determine the diversity of Bacillus species. Some of the samples were prepared in the laboratory, but most were purchased from retail outlets such as Asian restaurants and take-away food shops (Appendix 2).
70 Table 3.8 shows the detection and enumeration of Bacillus species from cooked rice samples using the spread plate method and the centrifugation-plating method combined with amylase and Tween 80 treatment. Overall, the diversity of Bacillus species found in cooked rice was not as great as that in raw rice.
Of the seven samples of plain steamed rice, three were cooked in the laboratory in a rice cooker and four were purchased from retail outlets. Of the laboratory-prepared samples,
1 and 2 contained B. cereus and B. licheniformis. The retail samples (4 – 7) contained B. cereus and/or B. subtilis. These results suggest that B. licheniformis, B. subtilis and B. cereus can survive the cooking process. B. subtilis and B. licheniformis are mesophilic bacilli like B. cereus, however, the spores of these species are more heat resistant than those of most B. cereus strains (Jenson and Moir, 2003). Also, some B. licheniformis strains were reported to be able to grow at 50 – 55ºC (Ronimus et al., 1997; Mendo et al., 2000). Therefore, the occurrence of B. licheniformis and B. subtilis in cooked rice may be cause for concern as both species have been associated with food poisoning
(Lund, 1990; Salkinoja-Salonen et al.,1999). However, there is the possibility that the
Bacillus isolates could come from other environmental sources through post-process contamination, due to their ubiquity in nature. This is particularly true for isolates from retail samples because laboratory cooked rice samples were prepared under essentially aseptic conditions. Many dairy scientists believe that contamination of pasteurized milk with Bacillus species results from survival of spores originally present in raw milk, though other literature (Crielly et al., 1994) suggests that post-pasteurization contamination might contribute more to Bacillus counts in pasteurized milk (Schraft et al., 1996).
71 Steamed rice, stored in the rice cooker in the laboratory for 3 days (sample 3), was also tested. The colour of the rice changed from white-cream to yellow-cream, and after 3 days, the dominant bacterium was B. cereus (3 ×102 cfu/10 g). This result represents that the association of Bacillus species with rice may lead not only to risk of food poisoning but also to spoilage of cooked rice. B. cereus is usually not a competitive microorganism but spores of B. cereus will have little or no competition from vegetative cells, especially those faster-growing Gram-negative bacteria, following heat treatment
(cooking) and high storage temperature (in cooker), which are adequate to kill or suppress non-sporulating, vegetative, and faster-growing bacteria (Granum, 2002).
These results also indicated that some strains of B. cereus (probably as spores) can survive the warming temperature of a rice cooker (above 60ºC) for at least a short period. Therefore, one of the measures for control of B. cereus that is recommended for preventing the growth of this bacterium by keeping the food above 60ºCG(Lund, 1990;
Jenson and Moir, 2003) should be examined more closely.
Products consisting of steamed rice mixed with other ingredients such as vegetables and meat (samples 8 – 14) were more heavily contaminated with Bacillus species than the samples of steamed rice alone (samples 1 – 2 and 4 – 7). Comparatively greater contamination with bacilli was found in both raw and cooked rice mixed with other ingredients, than in plain raw or cooked rice. There are possible reasons for this. First, these foods are more prone to contamination through mishandling, or the additional ingredients, particularly herbs, spices, seasonings and flavourings, may increase the load, as all have been identified as sources of B. cereus (van Netten and Kramer, 1992).
Second, other ingredients may provide nutrients such as protein sources that can support the growth of B. cereus (Morita and Woods, 1977) and possibly other Bacillus species
72 as well.
No Bacillus species were detected in organic food (sample 15), brown rice (samples 17
– 18) or sushi (samples 26 – 31), although these products were heavily contaminated with other microorganisms, such as Pseudomonas, Pantoea and Marinococcus species, which were identified by partial sequencing. Only one sushi sample (sample 32) was contaminated with 2 × 102 cfu/10 g of B. cereus.
Most fried rice samples (samples 19 – 25) were highly contaminated with Bacillus species as compared to other cooked rice samples. Sample 19 was contaminated with 2
× 103 cfu/g of mixed Bacillius species, including B. cereus, B. mycoides and B. subtilis, with 9 × 102 cfu/g of B. cereus only. Sample 23 yielded 2×103 cfu/g of B. subtilis and 3
× 103 cfu/g of mix of B. subtilis and B. licheniformis. Most fried rice samples were contaminated mainly with B. cereus, B. subtilis and B. licheniformis, which seemed better able to survive the cooking process, as observed during analysis of steamed rice.
In the fried rice, a classic vehicle associated with the emetic syndrome, moderate to high populations are of concern as the food is ready to eat. There is the possibility that the population of Bacillus cells was even higher in the plain rice used to prepare the fried rice, due to the germination of spores and the growth of cells. Toxin may have been produced and, while the majority of viable cells or spores may be destroyed during the final stage of preparation of the fried rice, the heat-stable toxin(s) may still remain in the rice.
73 Table 3.8 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus in cooked rice samplesG
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 1 Cooked rice (Lab) 1 - - 1 ------2 Cooked rice (Lab) 1 - - 1 ------3 Stored cooked rice 298 - - - - - 30 - - - - - 4 Cooked rice - 1 ------5 Cooked rice 10 ------6 Cooked rice 3 ------7 Cooked rice 11 24 ------8 CR* with veg. & meat** TNTC TNTC - - - - 300 40 - - - - 9 CR with veg. & meat - - - TNTC - - - - - 5950 - - 10 CR with veg. & meat 29 39 22 ------11 CR with ram 2 ------12 CR with pork 143 33 ------13 CR with beef TNTC TNTC - TNTC TNTC - 30 650 - 1150 50 - aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species - , Not detected; *CR, Cooked rice; **CR with veg. & meat, Cooked rice with vegetable and meat
74 Table 3.8 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus in cooked rice samples (continued)
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 14 CR* with chicken - TNTC TNTC - - - - 340 20 - - - 15 Organic food ------16 Risotto TNTC - - - - - 12000 - - - - - 17 Brown rice ------18 Brown rice ------
19 Fried rice TNTC (TNTC) TNTC - - - - 900 (960) 120 - - - - 20 Fried rice 87 62 19 ------21 Fried rice 1 330 - 34 - - - 70 - - - - 22 Fried rice 6 ------23 Fried rice - TNTC - TNTC - - - 2360 - 690 - - 24 Fried rice 54 TNTC - TNTC - - - 510 - 60 - - 25 Fried rice 33 TNTC - TNTC - - - 150 - 90 - - 26 Sushi with Chicken ------aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species - . Not detected ; *CR, Cooked rice
75 Table 3.8 Comparison of centrifugation-plating and spread plating methods for the detection and enumeration of Bacillus in cooked rice samples (continued)
Centrifugation plating method (cfu/10 g) Spread plating method (cfu/g) Sample Bc (Bm)a Bsb Bpc Bld Bmee Speciesf Bc Bs Bp Bl Bme Species 27 Sushi with Salmon ------28 Sushi with Seaweeds ------29 Sushi with Prawn ------30 Sushi with Salmon ------31 Sushi with Beef ------32 Sushi with Salmon 200 ------33 Comm. CR* with tomato 1 2 ------34 Comm. CR with egg - 14 ------35 Comm. CR with coriander 2 ------aBc (Bm), B. cereus/ B. thuringiensis (B. mycoides); bBs, B. subtilis; cBp, B. pumilus; dBl, B. licheniformis; eBme, B. megaterium; fSpecies, Bacillus species - . Not detected; *Comm. CR, Commercial cooked rice product
76 Commercial par-cooked rice products (samples 33 – 35), which can be consumed after only two minutes of cooking, also contained several colonies of B. cereus and B. subtilis
(2 – 14 cfu/10 g).
Table 3.9 summarises the populations of B. cereus and Bacillus species detected in cooked rice samples. The levels of B. cereus and Bacillus species in cooked rice samples were mostly in the range of 1 – 102 cfu/10 g, though five samples contained
Bacillus species exceeding 104 cfu/10 g. Generally, the populations of B. cereus were similar to those reported by other researchers (Table 2.3 in Section 2.2) but the proportions of cooked rice contaminated with the B. cereus (71.4%) were higher than in these previous reports (16.3%).
Table 3.9 Number (proportions, %) of cooked rice samples with different populations of Bacillus species
Bacillus Bacillus Plating method B. cereus (cfu/10 g of raw rice) species Centrifugation None 9 (25.7) 14 (40.0) Centrifugation 1 – 100 12 (34.3) 14 (40.0) Centrifugation 101 – 500 5 (14.3) 4 (11.4) Centrifugation501 – 1000 0 (0.0) 0 (0.0) Spread1000 – 10000 4 (11.4) 2 (5.7) Spread>10000 5 (14.3) 1 (2.9) Total 35 (100) 35 (100) G
G
G
G
77 3.4 Conclusions G
G
In this study, the centrifugation-plating concept has been applied as a sensitive detection and enumeration method for Bacillus species. In turn, preliminary ecological information on Bacillus species in rice products has been generated. The high sensitivity of this method has lead to improved detection of a variety of Bacillus species in rice samples, as compared to the spread plate method.
Some practical issues should be considered when applying the centrifugation-plating method. First, an appropriate and effective selective-differential plating medium is needed to detect target species because centrifugation sediments and concentrates all microbial cells. PEMBA, a selective-differential agar for B. cereus, was chosen for detection and enumeration of Bacillus species. Second, in some cases, the amount of particulate matter from the matrix is too voluminous for the sediment to be spread onto the agar medium. To overcome this problem, particularly for cooked rice samples, a
Stomacher bag with an inner mesh pouch along with amylase and surfactant treatment was used in this study.
Farinaceous foods such as raw and cooked rice samples, major vehicles of transmission of bacilli and their toxins, were examined by the centrifugation-plating method to obtain information on the diversity of Bacillus species in these products. Comparatively more diverse Bacillus species were found in raw rice than in cooked rice samples. Eight different Bacillus species, B. cereus/B. thuringiensis, B. mycoides, B. subtilis/B. mojavensis, B. licheniformis, B. pumilus, B. sphaericus/B. fusiformisGand B. megaterium, as well as Paenibacillus species were found in raw rice. Six of these species were found
78 in cooked rice samples, though B. sphaericus/B. fusiformis and Paenibacillus species were not. Of these bacilli, toxigenic or potentially toxigenic B. cereus/B. thuringiensis,
B. subtilis, B. licheniformis and B. pumilus were found most frequently in both raw and cooked rice. The presence of these species indicates potential risk of foodborne disease associated with rice products because most food poisoning incidences with regards to
Bacillus involve the B. cereus and the B. subtilis groups.
G
In summary, the centrifugation-plating method is a very useful technique, generating information on the populations and diversity of Bacillus species associated with raw and cooked rice products.G
79 Chapter 4
Development and application of micellar electrokinetic chromatography to analyse Bacillus cereus emetic toxin, cereulide 4.1 Introduction G
Cereulide, an emetic toxin associated with B. cereus, is a highly hydrophobic, heat- stable and acid/alkali-resistant cyclic dodecadepsipeptide. It exhibits potassium ionophore activity that affects mitochondrial function, which in turn may lead to cell death (Mikkola et al., 1999), and induces emesis in the host. The chemical structure of cereulide is similar to that of the commercially available antibiotic, valinomycin
(Section 2.8.2.1).
The relationship between toxin production and illness is poorly understood, at least in part due to the lack of reliable and quantitative methods for cereulide analysis. Several bioassays have been developed for detection of cereulide (Hughes et al., 1988;
Andersson et al., 1998; Finlay et al., 1999), though these do not generate real quantitative data. Further, they are expensive, time-consuming, labour intensive, inconvenient to perform due to dealing with living cells and require considerable expertise to perform. There appears to have been only one report (Häggblom et al.,
2002) of the quantitative analysis of cereulide, using high-performance liquid chromatography-ion trap mass spectrometry (HPLC-MS).
Capillary electrophoresis (CE) is a newer, alternative method to HPLC, offering great potential as a separation technique, as it is simple, rapid, highly efficient in separation, and cost effective (Boone et al., 1999; Issaq, 2000). Micellar electrokinetic chromatography (MEKC) is one of the modes of capillary electrophoresis that was first described by Terabe et al. (1984) for analysis of neutral hydrophobic molecules. MEKC utilizes the chemical equilibrium to modify analyte electro-migration and manipulate
G _X separation selectivity. Analytes partitioned between the aqueous phase and the micellar
(pseudo-stationary) one and, based on hydrophobic interaction with the micelles, neutral analytes or analytes with identical charge-to-mass ration can be separated, if they differ in hydrophobicity (Idei, et al., 1998). Several studies have demonstrated the analysis of hydrophobic compounds using MEKC (Idei et al., 1998; Fürtös-Matei et al., 2000).
In this chapter, a simple and rapid MEKC method for quantitative analysis of cereulide was developed and was applied to analysis of cereulide in model and real food samples to evaluate application of the method to industry.
4.2 Materials and methods G
4.2.1 Bacillus strains
The strains of Bacillus species used in this thesis and their origins are described in Table
4.1. Strains classed as emetic or diarrhoeal in this study were defined as such based on their histories, stemming from the association with either emetic or diarrhoeagenic food poisoning outbreaks, respectively.
4.2.2 Maintenance of cultures
The purity of Bacillus strains received from culture collections was checked by streaking for isolation on TSA. For short term culturing, strains of pure culture were grown on TSA, Nutrient Agar (NA, Oxoid) or Polymyxin pyruvate Egg-yolk Mannitol
Bromothymol blue Agar (PEMBA, Oxoid), and enrichment was performed in either
Brain Heart Infusion (BHI) broth (Oxoid), or TSB throughout this study. For long term
G _Y storage, strains were maintained as suspensions of cells in TSB with 20% (v/v) glycerol at -80ºC.
Table 4.1 Bacillus strains used in this study Speciesa Strain IDb (source)c Description S1 NC F (Agata) Emetic; Food poisoning S2 NC G (Agata) Emetic; Food poisoning S3 NC Y (Agata) Emetic; Food poisoning S4 NC 88F (Agata) Emetic; Faeces S5 NC 90T (Agata) Emetic; Faeces S6 NC 954 (Agata) Emetic; Food poisoning S7 NC 1078 (Agata) Emetic; Food poisoning S8NC 1128 (Agata) Emetic; Soil S9NC 1149 (Agata) Emetic; Vomitus S10NC 1219 (Agata) Emetic; Faeces S11NC 1237 (Agata) Emetic; Foods S12NC 1240 (Agata) Emetic; Foods S13NC 1246 (Agata) Emetic; Soil S14NC 1249 (Agata) Emetic; Soil S15NC 1260 (Agata) Emetic; Foods S16NC 1291 (Agata) Emetic; Foods S17NC 7401 (Agata) Emetic; Faeces S18F 47 (Andersson) Emetic; no case history S19F 4810/72 (Andersson) Emetic; no case history S20F 5881 (Andersson) Emetic; no case history S21N0075-95 (Granum) Diarrhoeal; Fpd; veg stew S22N0248-00 (Granum) Diarrhoeal; Fp; lasagne S23N0674-98 (Granum) Diarrhoeal; Fp; eggs S24N1104-98 (Granum) Diarrhoeal; Fp; fish soup S25N1113-01 (Granum) Diarrhoeal; Fp; roast veal S26N1230-88 (Granum) Diarrhoeal; Fp; oriental stew S27N1649-99 (Granum) Diarrhoeal; Fp; mashed swede S28N1651-00 (Granum) Diarrhoeal; Fp; caramel pudding S29N 1694-00 (Granum) Diarrhoeal; Fp; cream sauce S30NVH 200 (Granum) Diarrhoeal; Fp; dish with rice B. amyloliquefaciens S31TICC 376 (Tecra£) No origin history S32TICC 377 (Tecra£) No origin history B. circulans S33TICC 2715 (Tecra£) No origin history B. licheniformis S34TICC 18 (Tecra£) No origin history S35TICC 369 (Tecra£) No origin history S36TICC 370 (Tecra£) No origin history S37TICC 371 (Tecra£) No origin history S38TICC 372 (Tecra£) No origin history S39TICC 373 (Tecra£) No origin history S40TICC 375 (Tecra£) No origin history B. megaterium S41TICC 16 (Tecra£) No origin history B. mycoides S42FS (UNSW) No origin history B. pumilus S43UNSWMI001200 (Wilson) No origin history S44TICC 17 (Tecra£) No origin history
G _Z Table 4.1 Bacillus strains used in this project (continued) Speciesa Strain IDb (source)c Description B. sphaericus S45TICC 378 (Tecra£) No origin history S46TICC 379 (Tecra£) No origin history B. subtilis S47TICC 11 (Tecra£) No origin history S48TICC 12 (Tecra£) No origin history S49TICC 15 (Tecra£) No origin history S50TICC 745 (Tecra£) ATCC e 11774 S51UNSWMI030702 (Wilson) No origin history B. thuringiensis S52FDA 2 (Bennett) No origin history S53FDA 4 (Bennett) No origin history S54FDA 7 (Bennett) No origin history S55FDA 11 (Bennett) F9470-CDCf S56FDA 12 (Bennett) ATCCe 10792 S57 TICC 4100 (Tecra£) No origin history S58 TICC 4161 (Tecra£) No origin history S59 TICC 4162 (Tecra£) No origin history S60 TICC 4163 (Tecra£) No origin history aB., Bacillus; SH., Starch Hydrolysis bDesignated treatment ID number for use in experiments cNorio Agata, Nagoya City Public Health Research Institute, Nagoya, JAPAN; Maria Andersson, Department of Applied Chemistry and Microbiology, FIN-00014 University of Helsinki, Helsinki, Finland; Reginald Bennett, US Food & Drug Administration, USA; Per Einar Granum, Department of Food Hygiene, Norwegian College of Veterinary Medicine, Oslo, Norway; Paul Horwood, Department of Microbiology and Immunology, James Cook University, Townsville, Qld, Australia; Tecra £ International Pty Ltd., Frenchs Forest, Sydney, Australia; Max Wilson, School of Biotechnology and Biomolecular Sciences Culture Collection, University of New South Wales, Sydney, Australia. dFp., Food poisioning eATCC., American Type Culture Collection fF9470-CDC., F9470-Centers for Disease Control
4.2.3 Purification and analysis of B. cereus emetic toxin
4.2.3.1 Preparation of B.cereus emetic toxin
B. cereus strain NC7401 was grown on plates of Tryptone Soya Agar (TSA; Oxoid,
Melbourne, Australia) (300 plates per experiment) at 28qC for 72 h. Biomass from the plates was collected and suspended in methanol (Andersson et al., 1998).
4.2.3.2 Purification of cereulide G
The purification procedure was essentially as described previously (Andersson et al.,
G _[ 1998), with slight modification according to Toh et al. (2004). The methanol extract of
B. cereus 7401 was centrifuged at 10,000Gu g for 20 min at 4qC to remove cellular debris, and the supernatant was filtered through a 0.45 Pm pore membrane (Advantec
MFS Inc., Dublin, CA, USA). The filtrate was then evaporated to dryness at 35qC using rotary evaporation (Buchi, Switzerland) and the dried residue dissolved in methanol/water (90:10 v/v). The solution was injected into a pre-wet Sep-pak C18 cartridge (Waters, Rydalmere, NSW, Australia), which was washed with water/methanol
(80:20 v/v) before elution of cereulide with 100% methanol. The methanol eluate containing cereulide was dried by rotary evaporation, and the residue was dissolved in 1 mL acetonitrile/water (90:10 v/v) containing 0.1% TFA. This solution was subjected to high performance liquid chromatography (HPLC) analysis for cereulide.
4.2.3.3 RP-HPLC analysis
The final solution above was analysed by reversed phase HPLC (Waters Millennium,
USA) according to protocols essentially as described by Mikkola et al. (1999). The column used was Sephasil C8 SC 2.1 mm in diameter u 100 mm, 5 Pm (Amersham
Pharmacia Biotech, Australia). The elution was isocratic, using water/acetonitrile/ trifluoroacetic acid (5:95:0.1 v/v/v) at a flow rate of 100 PL/min. Absorbance at 215 nm
(A215) was used for detection. Valinomycin (Sigma, St. Louis, USA), a close structural analogue of cereulide, was utilised as a standard. For collection of a pure cereulide sample, the eluate corresponding to a retention time and peak approximating that for valinomycin was tapped and collected following detection and analysis using the Waters
Millennium Chromatography Manager Data Analysis Software (Waters Millennium,
USA).
G _\ 4.2.3.4 NMR analysis
The putative toxin collected was evaporated to dryness by rotary evaporation then dissolved in 0.5 mL of 100% DMSO-d6 (Merck & Co., USA). The collected sample was submitted to Dr Graham Ball (School of Chemistry, UNSW, Sydney, Australia) who kindly carried out the NMR analysis.
G G 4.2.4 Analysis of cereulide by cellular cytotoxicity (MTT) assay
4.2.4.1 Growth of cells and media
A stock culture of CHO-K1 cells (UNSW, Sydney, Australia) was cultured in 25mL of
Dulbecco’s Modified Eagle Medium (DMEM) (JRH Biosciences, Lenexa, KS, USA) supplemented with 10% foetal bovine serum (FBS) (Gibco-BRL life Technologies, New
York, NY, USA) at 37qC in a humid atmosphere containing 5% CO2/95% air. An antibiotic/antimycotic supplement (A5955, Sigma, St. Louis, USA) was added to inhibit bacterial and fungal growth. The cells were sub-cultured until a confluent monolayer was observed in a 75cm2 tissue culture T-flask (T75) under the microscope, then harvested after trypsin treatment and suspensions containing 5 u 104 cells/mL were prepared.
4.2.4.2 Preparation of cereulide
Purified cereulide (up to 95% purity, according to Section 4.2.3.4) was autoclaved at
121qC for 15 min, dissolved in the non-polar solvent, dimethyl sulphoxide (DMSO,
Sigma, St. Louis, USA) and serially diluted (two-fold dilutions). As the concentration of
DMSO from the sample should be below 0.5% in the cell medium to prevent discernible adverse effects upon the cells, cereulide dissolved in 50 PL of DMSO was made up to 1
G _] mL with PBS and this formed the stock solution from which further dilutions were made. CHO-K1 (100 PL per well) cultures in 96-well plates were exposed to 100 PL of cereulide solution from each dilution, with each dilution tested in triplicate.
4.2.4.3 MTT cell cytotoxicity assay
4.2.4.3.1 Chemicals
MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide, Sigma, St. Louis,
USA) was dissolved in phosphate saline buffer (PBS) at 5 mg/mL and the solution was filter-sterilised with a 0.2 Pm porosity filter.
4.2.4.3.2 MTT assay procedure
A 100 Pl volume of toxicant (cereulide) and negative control was dispensed into designated wells of a 96-well microtitre plate. The plate was set up as illustrated in
Figure 5.1. The wells were seeded aseptically with 5 u 104 CHO-K1 cells, and plates were incubated at 37qC in a humid atmosphere containing 5% CO2 for 72 h. After the incubation period, 20 PL of an aqueous MTT solution were added to each well, and the plate incubated for a further 4 h at 37qC in a humid atmosphere containing 5% CO2.
After incubation, the liquid medium in the wells was removed and 220 PL DMSO was added to each well to dissolve formazan crystals. The plate was placed onto an automated shaker for 30 min to help solubilisation.
The absorbance of each well was determined in an ELISA plate reader at 540 nm, using
620 nm as a reference. The toxic effect of cereulide on the CHO cell line was calculated from the equation OD = [1 (ODTest /ODControl)] u 100%, where OD is final optical
G _^ density, ODTest is OD of cereulide-treated cells, and ODControl is the OD of untreated control cells (Beattie and Williams 1999). Samples were considered to be toxic if the optical density of the test well was more than 20% lower than that detected in the negative control wells.
Control B B Treatment L L (in triplicate) A A N N Treatment K K (in triplicate)
Control G G G Figure 5.1 Plate set-up for MTT assayG
4.2.5 Development and application of capillary electrophoresis (CE)
4.2.5.1 Chemicals
All chemicals and solvents were of analytical or HPLC grade. Valinomycin, sodium dodecyl sulphate (SDS), dibasic sodium phosphate (Na2HPO4), sodium hydroxide
(NaOH), methanol (MeOH) and ortho-phosphoric acid were purchased from Sigma-
Aldrich (Sydney, Australia). Boric acid was purchased from Ajax Chemicals (Sydney,
Australia). Acetonitrile (MeCN) was obtained from LabScan (Bangkok, Thailand).
HPLC-grade distilled, deionised water was prepared with a Milli-Q system (Millipore,
Bedford, MA, USA).
G __ 4.2.5.2 Sample preparation
4.2.5.2.1 Valinomycin and purified cereulide
Valinomycin and purified cereulide solutions were prepared by dissolving in distilled water/methanol (1:1 v/v).
4.2.5.2.2 Semi-purified samples
The 60 reference Bacillus strains (Table 4.1) were grown as a lawn on TSA plated at
28qC for 72 h. Biomass from plates was collected, and suspended in methanol at 100 mg mL-1 (Andersson et al., 1998). After 2 h of extraction of putative emetic toxin, the methanol extracts were centrifuged at 3,000 u g for 10 min and supernatants were collected. The supernatant was filtered through a 0.45 Pm porosity disposable filter, then diluted using an equal volume of water.
4.2.5.2.3 Rice samples
A mixture of two volumes of water (100 mL) and one volume of raw rice (50 g) was autoclaved in a glass beaker for 20 min at 121qC to sterilise and cook the rice simultaneously (McElroy et al, 2000). The boiled rice (10 g) was then transferred to a sterile Stomacher bag (Interpath, Sydney, Australia) and inoculated with B. cereus 7401
(102 – 103 cfu/g of rice). The inoculated samples were incubated for 72 h at 28qC.
Following incubation, boiled rice (10 g) was extracted with an equal volume (w/v) of methanol (10 mL) and centrifuged at 3,000 u g for 10 min. The supernatant was filtered through a 0.45 Pm porosity filter membrane and stored at -20qC until used. The prepared sample was diluted in an equal volume of distilled water before use.
G _` 4.2.5.3 Development of MEKC procedure for cereulide
4.2.5.3.1 Equipment
MEKC analyses were performed using the BioFocus model 3000 capillary electrophoresis system equipped with an UV detector (Bio-Rad, Hercules, CA, USA).
Uncoated capillaries, 41 cm (36.4 cm effective length) and 59 cm (54.4 cm effective length), u 50 Pm I.D. (Beckman, Fullerton, CA, USA), were used. The longer capillary was used for method development and the shorter capillary was used for analysis with the final method.
4.2.5.3.2 MEKC procedures
The running buffer for MEKC comprised borate (20 mM – 100 mM) or phosphate buffer (50 mM – 100 mM), or a mixture of the two, containing SDS (50 mM – 100 mM).
The pH of the buffers was adjusted as necessary with 0.5 M NaOH for borate buffer, and phosphoric acid for phosphate buffer. Acetonitrile or methanol was used as an organic modifier in the buffer. Buffer solutions were filtered through 0.45 Pm membrane filters before use.
Samples were injected for 2 s using hydrodynamic mode, at 34.5 kPa, the standard working pressure of the instrument. Separation was carried out at a constant voltage of between 10 – 25 kV, with positive polarity. The analysis was performed at a capillary temperature between 20 and 30qC, with the sample compartment (carousel) maintained at 22ºC. Detection was investigated at wavelengths between 190 nm and 254 nm, with sensitivity setting of 0.02.
G `W New capillaries were first rinsed with 0.1 M NaOH for 10 min, incubated with 0.1 M
NaOH for 20 min, followed by rinses with water for 10 min and electrophoresis buffer for 20 min before first use. Between runs, the capillary was rinsed for 1 min with 0.1 M
NaOH, 1 min with distilled water and 3 min with running buffer.
G
4.3 Results and discussion G
G
4.3.1 Purification and analysis of B. cereus emetic toxin for development of CE
4.3.1.1 Purification and analysis of B. cereus emetic toxin Cereulide was purified using methanol extraction and RP-HPLC. Despite the fact that the method can yield cereulide at high purity, it is too laborious and time-consuming to provide purified cereulide sufficient for use in subsequent experiments. Therefore, several other purification methods were attempted. However, simple aqueous, methanol or pentane extraction, without further purification by chromatography, yielded cereulide of low purity, when checked using NMR analysis. Even though hydrophobic interaction chromatography (HIC) using a HiTrapTM HIC selection kit (Amersham Pharmacia
Biotech, Australia) was performed for further purification, the purity (<90%) was less than that obtained with RP-HPLC. Consequently, methanol extraction followed by RP-
HPLC preparation was used to obtain toxin of high purity.
Figures 4.2 and 4.3 show the valinomycin peak and the peak profile of the purification product from HPLC. Valinomycin (100 Pg/mL with 10 PL of injection volume) was detected at a retention time of 10 – 11 min (Figure 4.2) and, from the toxin extract, a significant peak was observed at a retention time similar to that for valinomycin; this
G `X peak was assumed to represent cereulide (Figure 4.3). The fractions of putative cereulide was collected and submitted for NMR analysis at the School of Chemistry,
UNSW.
Figure 4.2 HPLC analysis of the standard valinomycin
Figure 4.3 HPLC analysis of the purification product
G `Y Beside the putative cereulide peak at a retention time of 10.163 min, other peaks were also observed through the analysis. These peaks were considered to come from unknown hydrophobic compounds that were components of B. cereus cells, produced by B. cereus, or from the medium. The solvent peak was observed when a sterile preparation of acetonitrile/water (90:10) with 0.1% TFA was analysed using HPLC, at a retention time of approximately 4 min.
4.3.1.2 NMR analysis
Nuclear magnetic resonance (NMR) analysis of putative cereulide sample was performed to verify the compound isolated using HPLC as cereulide and determine its purity.
The isolate from HPLC was confirmed to be cereulide, at up to 95% purity, after comparing the NMR result to that of Suwan et al. (1995) (Figure 4.4). A 1H NMR spectrum at 600 MHz (Suwan et al., 1995, Appendix 3), shows 2 sets of a doublet NH signal (G 8.0-8.4), 4 alpha protons (G 4.2-5.1), and 7 sets of doublet methyl signals
(G0.7-1.4). The latter signals suggested four partial structures of alanine (Ala), valine
(Val), 2-hydroxy-3-methylbutyric acid (O- Val), and O-Leu. These amino acids occur in the sequence cyclo-(O-Leu-Ala-O-Val-Val)3 and this represents the subunit of cereulide
(Suwan et al., 1995). The degree of purity also compares well with the commercially available analogue, valinomycin (90% for Sigma, 95% for Calbiochem, Australia).
G `Z G
G G
̺
ଘ ̻G ଘ ̻ ଗ ଗG ̻G̺GଗG
G G G G Figure 4.4 NMR spectrum of cereulide
The sample was confirmed by NMR to be cereulide, as compared with literature (Appendix 3) (Suwan et al., 1995). Two other peaks belong to the solvent components of water (G 3.0) and acetonitrile (G 1.9); (Pouchart and Behnke, 1993); that were not completely evaporated. Other differences to literature are attributed to very similar compounds or impurities in the sample.
G `[ 4.3.2 Determination of biological activity and concentration of cereulide by colorimetric MTT assay
G The MTT assay was utilised to investigate the toxicity levels of samples against tissue culture cells in order to determine both quantitative and qualitative levels of the emetic toxin, cereulide.
The results of the MTT assay using spectrophotometric quantification showed a dilution-dependent correlation. The last well in which MTT was not reduced correlated with the minimum dose required for toxicity toward the cells. Agata et al. (1994), had found that the minimum dose required to cause toxic effects in HEp-2 cells was 5 ng/mL. In the comparison of seven different mammalian cell lines, it was found that Int
407, CHO and HEp-2 were equally sensitive to B. cereus emetic toxin (Szabo et al.,
1991). As HEp-2 cells were not readily available at UNSW, CHO cells were used to conduct the MTT assays and it was assumed, that if CHO and HEp-2 are equally sensitive, that likewise, the minimum dose required for toxicity in CHO cells would be
5 ng/mL.GThe CHO cell proved its effectiveness for crude cereulide in the MTT assay by the study of Beattie and Williams (1999). Known amounts of purified cereulide were used to generate a calibration curve for the MEKC method (Section 4.3.3.2.1).
4.3.3 Development and application of MEKC analysis
4.3.3.1 Method development
The MEKC method, developed through analysis of valinomycin as a reference compound, was directly transferable to the analysis of cereulide, as the two molecules
G `\ .. analysis (Haggblom et al., 2002), it was assumed that they would be behave similarly using capillary electrophoretic analysis, which proved to be the case.
Borate and phosphate buffers were evaluated as the running buffer, with borate buffer chosen as it provided the best resoultion. Use of phosphate buffer resulted in a longer retention time and a broader analyte peak (Figure 5.5).2
'$"$$
Valinomycin (A) Borate buffer &&"&%
%("(' Methanol
*"*( Absorbance (mAU)
!%"%)
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$
Migration time (min)
($"$$ (B) Phosphate buffer &-"(&
Methanol %,",( Valinomycin
Absorbance (mAU) ,"&)
!&"''
$"$$ )"$$ %$"$$ %)"$$ &$"$$ &)"$$ Migration time (min)
Figure 4.5 Influence of borate and phosphate buffer
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 59 cm [54.4 cm to the detector] X50μm, 25 C,15 kV; UV absorbance at 195 nm.
96 Furthermore, borate-phosphate buffer was inferior to borate buffer alone. Increase in the concentration of borate buffer, from 20 mM to 50 mM or 100 mM, did not improve resolution but increased both retention time and current and the use of a lower concentration buffer can aid to minimise the band broadening due to Joule heating
(Altria, 1996). Therefore, 20 mM borate buffer was selected for subsequent experiments.
The influence of pH was examined in the range 6.7 to 9.2, with a detectable signal only between 7.5 and 9.2. At pH 6.7, electro-osmotic flow (EOF) is not sufficiently strong to force the countermovement of micelles to eventually pass through the detector; low pH buffers generate low rates of EOF in the MEKC mode. At pH 7.5, retention times were longer and sample peaks were broader. At pH values of 8.5 and 9.2, the sample peak was sharp and the peak for methanol (a major non-target peak) was smaller at pH 8.5 than at 9.2 (Figure 4.6). If the non-target peak is too large, the resolution of the target peak can be decreased, a particular problem if the concentration of target analyte is low.
Consequently, a pH of 8.5 was chosen for the analysis.
The addition of 5% or 10% acetonitrile or methanol to the borate buffer as an organic modifier was also investigated because the use of such amendments can contribute to the extension of the migration time window, with improvement in the resolution for hydrophobic compounds as a result (Chen et al., 1995). However, these organic modifiers did not improve resolution dramatically but generated baseline noise and increased retention time. Thus, analysis was continued without addition of any organic modifier to the buffer.
SDS was selected as a surfactant as it functions as a pseudo-stationary phase, forming
G `^ micelles lagging behind the EOF (Cheng et al., 2005), and has proved useful for the separation of highly hydrophobic compounds (Idei et al., 1998; Cheung at al., 2001).
SDS concentration had a variable effect on peak shape, and pseudo-retention time of the analyte (Figure 4.7) at fixed pH values. Between 50 mM and 75 mM SDS, there was no significant difference in retention time and peak resolution, though 75 mM SDS enhanced sensitivity. At 100 mM SDS, sensitivity was better than at 75 mM SDS but retention time was extended and resolution was very poor. Therefore, 75 mM SDS was chosen, providing the best balance of retention time, resolution of peak and detection sensitivity of sample.
G
A constant voltage of 10, 15, 20 and 25 kV was applied during electrophoresis, with profound impact upon retention time. Higher voltage decreased retention time, but in excess of 20kV, baseline noise increased. Therefore, 15 kV was chosen as the running voltage for analysis with the longer capillary.
Temperature had some effect in the range of 20 – 30qC, though its effect was less than voltage upon analysis. Higher capillary temperature slightly accelerated analysis, but decreased resolution of the sample peak slightly in terms of peak shape and a clean base line. Therefore, the capillary was maintained at 25qC as a good compromise between migration time and resolution. The carousel temperature was gradually increased while the analysis progressed. Therefore, the sample compartment temperature was maintained at 22qC during analysis to prevent a negative effect on the buffer system.
G `_ &$"$$
(A)pH6.7 %("+)
-")$ Methanol
("&( Absorbance (mAU)
!%"$%
$"$$ *"$$ %&"$$ %,"$$ &("$$ '$"$$ Migration time (min)
'$"$$ (B) pH 7.5
&&"%- Valinomycin
%("', Methanol
*")+ Absorbance (mAU)
!%"%)
$"$$ )"$$ %$"$$ %)"$$ &$"$$ &)"$$ Migration time (min)
'$"$$
Valinomycin (C) pH 8.5 &&"&%
%("(' Methanol
*"*( Absorbance (mAU)
!%"%)
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
'$"$$
Valinomycin (D) pH 9.2 &%"*$ Methanol
%'"%-
Absorbance (mAU) ("+$
!'"*&
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
Figure 4.6 Influence of buffer pH on valinomycin separation
The electrophoresis buffer is 20 mM borate buffer with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 59 cm [54.4 cm to the detector] X 50 μm, 25 C, 15 kV; UV absorbance at 195 nm.
99 &$"$$
(A) 50mM SDS Methanol %(")*
-"%% Valinomycin
'"*+ Absorbance (mAU)
!%"++
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
&$"$$
(B) 75mM SDS %("'(
Methanol Valinomycin -"$+
'"*% Absorbance (mAU)
!%",*
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
'$"$$ Methanol (C)100mM SDS
&%"&$
%&"(% Valinomycin
'"*% Absorbance (mAU)
!)"%-
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
Figure 4.7 Influence of SDS concentration on valinomycin separation
The electrophoresis buffer is 20 mM borate buffer of pH 8.5. The electrophoretic conditions are: uncoated capillary of 59 cm [54.4 cm to the detector] X 50 μm, 25 C, 15 kV; UV absorbance at 195 nm.
100 Absorbance values for the sample peak were also investigated in the range 190 nm to
254nm. The sample peak could be distinguished between 190 nm and 220 nm, with the highest resoultion achieved between 190 nm and 200 nm, and 195 nm was selected as optimal. Although the cereulide peak was detected using HPLC at 210 nm (Agata et al.,
1994) or 215 nm (Mikkola et al., 1999), a shorter wavelength range proved more useful with MEKC. After development of the method using valinomycin, effectiveness in the analysis of cereulide was determined using purified cereulide (Figure 5.8), and a capillary of 54.4 cm effective length.
%$"$$ Methanol
*"$+
Cereulide '"+(
$"*& Absorbance (mAU)
!&")%
$"$$ ("$$ ,"$$ %&"$$ %*"$$ &$"$$ Migration time (min)
Figure 4.8 Peak of cereulide
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 59 cm [54.4 cm to the detector] X 50 mm, 25 C,15 kV; UV absorbance at 195 nm.
4.3.3.2 Application of the MEKC method
A shorter capillary (36.4 cm of effective length instead of 54.4 cm) was used for application of the method because it gave better detection sensitivity with purified
101 cereulide when the calibration curve was generated, the separation voltage could be decreased from 15kV to 12kV, and a single analysis required less than 10 min.
4.3.3.2.1 Reproducibility and calibration curve
Reproducibility of analysis was investigated and a calibration curve was obtained with purified cereulide, the purity confirmed by nuclear magnetic resonance (NMR) analysis
(Section 4.3.1.2). The biological activity of purified cereulide was checked with the cell cytotoxicity-based MTT assay (Section 4.3.2).
The reproducibility of the method is important for quantitative studies because electroosmotic flow and electrophoretic mobilities of ions is influenced by many factors associated with the sample matrix, such as pH, ionic strength, viscosity, dielectric constant and temperature (Markuszewski et al., 2003). The reproducibility was determined with six sequential runs of a standard solution containing 30 µg/ml of purified cereulide. The relative standard deviation (RSD) values of migration times and peak areas were 1.8 % and 4.2 %, respectively. The sample preparation and dilution procedure were critical to reproducibility during the analysis of cereulide.
A calibration curve, using peak area rather than peak height, was obtained with purified cereulide, with linearity between 2 and 30 µg/mL (Figure 4.9). The regression equation and correlation coefficient (r) were determined as y = 29641x + 58038 and 0.9932 (n =
4) respectively, with the SigmaPlot 8.0 program.
G XWY 120 y = 29641x + 58038
100 r = 0.9932
4 80
60 Peak area (x10 ) 40
20
0 0 5 10 15 20 25 30 35
Concentration (μg/ml)
Figure 4.9 Calibration curve for cereulide
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 41 cm [36.4 cm to the detector] x 50 mm, 25 C,12 kV; UV absorbance at 195 nm.
4.3.3.2.2 Semi-purified sample analysis
Cereulide production by reference cultures of Bacillus species (Table 4.1) was tested using MEKC. The emetic toxin, cereulide of Bacillus species was extracted by methanol (see method 4.2.5.2.2). Sixty of theses semi-purified samples were subjected to capillary electrophoresis and results are shown in Table 4.2.
The results showed that B. cereus diarrhoeal strains and Bacillus species other than B. cereus also produced cereulide putatively, yielding a peak from MEKC analysis with the same retention time and shape as highly purified cereulide (Section 4.3.1). B. cereus emetic strains produced cereulide in the ranges of 60 - 227 μg/mL except one, NC 1078
(42 μg/mL). Three of the most widely used emetic strains, NC 7401, NC Y and F
103 4810/72, produced the greatest amount of putative cereulide among twenty reference emetic strains (S1NC F – S20F5881). Of ten diarrhoeal B. cereus strains (S21N0075-95 –
S30NVH200), four did not produce a detectable amount of putative cereulide while six produced it, but at a far lower concentration than that produced by emetic B. cereus strains.
Strains of other Bacillus species, including B. megaterium, B. pumilus and B. subtilis, produced putative cereulide at levels similar to that produced by emetic B. cereus. Of these three species, B. pumilus and B. subtilis have been associated with outbreaks of food poisoning involving an emetic syndrome (Beattie and Williams, 2000; Jenson and
Moir, 2003). For B. licheniformis and B. thuringiensis, some strains produced putative toxin at less than 2 Pg/mL or none at all. Likewise, one of each strain of B. amyloliquefaciens, B. mycoides and B. sphaericus also produced putative cereulide.
These results indicate that toxin production among Bacillus is sometimes associated widely with species, as many B. cereus and B. subtilis isolates produced cereulide, while only specific strains of B. licheniformis, B. thuringiensis, B. amyloliquefaciens, B. mycoides and B. sphaericus produced toxin, and at much lower levels.
Collectively, the results of MEKC analysis of the sixty Bacillus strains revealed that the other Bacillus species as well as diarrhoeal B. cereus strains also are capable of producing cereulide-like molecules. Such capability was suggested recently after MTT assay and PCR analysis of eighty reference Bacillus strains (Toh et al. submitted), including the sixty strains analysed using MEKC in the current study. Toh et al.
(submitted) used HPLC-MS analysis of representative strains, including emetic (S17NC
7401) and diarrhoeal (S26N 1230-88) B. cereus stains, B. circulans (S33TICC 2715), B.
G XW[ megaterium (S41TICC 16), and B. thuringensis (S53FDA 4), as well as B. licheniformis
(S34TICC 18) as a ‘non-producing’ strain. That analysis showed that all strains except B. licheniformis (‘non-producing’ strain) produced cereulide though much less than that produced by the emetic B. cereus strain, NC 7401. The MEKC results in the current study correlate very well qualitatively with those HPLC-MS data, although the amount of cereulide produced by the strains differed. This can be explained by separate and different preparation methods for cereulide extraction, a different standard curve for quantification, and usage of different equipment for measuring the target analyte.
Among other Bacillus species, production of emetic toxin was also suggested recently by From et al. (2005). After testing 333 Bacillus species, they found eight strains that were toxic toward HEp-2 cells and, using LC-MS, identified what appeared to be a cereulide-like molecule, though different in size to cereulide. However, both the studies of From et al. (2005) and Toh et al. (submitted) provide evidence supporting the fact that other Bacillus species also produce emetic toxin, though whether these organisms produce cereulide only (according to the finding of Toh et al., submitted) or whether another form of emetic toxin exists (according to the finding of From et al., 2005) remains to be established. Taylor et al. (2005) also demonstrated that heat-stable toxin production by strains of Bacillus cereus, Bacillus firmus, Bacillus megaterium, Bacillus simplex and Bacillus licheniformis.
G XW\ Table 4.2 Screening results for semi-purified emetic toxin from Bacillus spp. using CE
Speciesa Strain IDb Cereulide (Pg/mL) B. cereus S1 NC F (emetic strain) 89 S2 NC G (emetic strain) 92 S3 NC Y (emetic strain) 177 S4 NC 88F (emetic strain) 60 S5 NC 90T (emetic strain) 64 S6 NC 954 (emetic strain) 170 S7 NC 1078 (emetic strain) 42 S8NC 1128 (emetic strain) 95 S9NC 1149 (emetic strain) 80 S10NC 1219 (emetic strain) 73 S11NC 1237 (emetic strain) 83 S12NC 1240 (emetic strain) 60 S13NC 1246 (emetic strain) 115 S14NC 1249 (emetic strain) 112 S15NC 1260 (emetic strain) 100 S16NC 1291 (emetic strain) 84 S17NC 7401 (emetic strain) 227 S18F 47 (emetic strain) 92 S19F 4810/72 (emetic strain) 188 S20F 5881 (emetic strain) 112 S21N0075-95 (diarrhoeal strain) 24 S22N0248-00 (diarrhoeal strain) <2c S23N0674-98 (diarrhoeal strain) <2 S24N1104-98 (diarrhoeal strain) 19 S25N1113-01 (diarrhoeal strain) 28 S26N1230-88 (diarrhoeal strain) 24 S27N1649-99 (diarrhoeal strain) 29 S28N1651-00 (diarrhoeal strain) 24 S29N 1694-00 (diarrhoeal strain) <2 S30NVH 200 (diarrhoeal strain) <2 B. amyloliquefaciens S31TICC 376 <2 S32TICC 377 87 B. circulans S33TICC 2715 83 B. licheniformis S34TICC 18 <2 S35TICC 369 52 S36TICC 370 77 S37TICC 371 26 S38TICC 372 15 S39TICC 373 41 S40TICC 375 <2 B. megaterium S41TICC 16 165 B. mycoides S42FS 54 S43UNSWMI 001200 <2 B. pumilus S44TICC 17 149
G XW] Table 4.2 Screening results for semi-purified emetic toxin from Bacillus spp. using CE (continued)
Speciesa Strain IDb CereulideGOPg/mL) B. sphaericus S45TICC 378 63 S46TICC 379 9 B. subtilis S47TICC 11 131 S48TICC 12 212 S49TICC 15 122 S50TICC 745 40 S51UNSWMI030702 104 B. thuringiensis S52FDA 2 44 S53FDA 4 44 S54FDA 7 20 S55FDA 11 <2 S56FDA 12 <2 S57 TICC 4100 <2 S58 TICC 4161 44 S59 TICC 4162 37 S60 TICC 4163 41 aB., Bacillus bDesignated treatment ID number for use in experiments c< 2, less than 2 Pg/mL
Figure 4.10 shows cereulide produced by emetic B. cereus F 4810/72, diarrhoeal B. cereus N 1230-88 and B. subtilis TICC 12, while Figure 4.11 shows no production by diarrhoeal B. cereus N 0674-98 and B. licheniformis TICC 18 after analysis using
MEKC.
G XW^ '$"$$
(A) &&"$$ Cereulide
%'"-- Absorbance (mAU) )"--
!&"$%
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$ Migration time (min)
&$"$$ (B)
%'",, Cereulide
+"+% Absorbance (mAU) %")+
!(")+
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$
Migration time (min)
&$"$$ (C) Cereulide %%"%*
,"'% Absorbance (mAU) &"(+
!'"'+
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$ Migration time (min)
Figure 4.10 Cereulide produced by Bacillus species grown on tryptic soy agar plates for3days
(A) Emetic B. cereus F 4810/72 (B) B. subtilis TICC 12 (C) B. megaterium TICC 16
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 41 cm [36.4 cm to the detector] X50μm, 25 C,12 kV; UV absorbance at 195 nm.
108 &$"$$ (A)
%("$-
,"%- Absorbance (mAU) &"&,
!'"*'
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$ Migration time (min)
'$"$$ (B)
&&"$$
%'"--
Absorbance (mAU) )"--
!&"$%
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$ Migration time (min)
Figure 4.11 MEKC analysis of 'non-producing' strains
(A) Diarrhoeal B. cereus strain N 0674-98 (B) B. licheniformis TICC 18
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are: uncoated capillary of 41 cm [36.4 cm to the detector] X50μm, 25 C,12 kV; UV absorbance at 195 nm.
4.3.3.2.3 Rice sample analysis
Cereulide production during growth of B. cereus NC 7401 in both TSA and cooked rice was monitored successfully using the MEKC method. A detectable amount of cereulide was produced after one day of incubation, with the concentration increasing markedly after 3 days of incubation (Figure 4.12). These results are in agreement with those of .. Haggblom et al. (2002) who indicated that cereulide production commences after the end of logarithmic growth or the early stage of stationary phase (approximately 24 h), and cereulide accumulates during the stationary phase of growth of emetic B. cereus.
109 &$"$$ Cereulide (A)
%("+*
-")'
Absorbance (mAU) ("&$
!$"-(
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$ Migration time (min)
+"$$ (B)
)"&& Cereulide
'"()
Absorbance (mAU) %"*+
!$"%$
$"$$ &"$$ ("$$ *"$$ ,"$$ %$"$$
Migration time (min)
Figure 4.12 Cereulide produced by B. cereus 7401 strain grown on tryptic soy agar plates and cooked rice for 3 days
(A) Grown on tryptic soy agar plates (B) Grown on cooked rice
The electrophoresis buffer is 20 mM borate buffer of pH 8.5 with 75 mM SDS. The electrophoretic conditions are; uncoated capillary of 41 cm [36.4 cm to the detector] X50μm, 25 C,12 kV; UV absorbance at 195 nm.
Cooked rice samples, associated with an outbreak of emetic B. cerues food poisoning in
Melbourne, Australia (Anon, 2002) were also examined. Of the three samples analysed, the first (rice with meat dish) showed production of cereulide (Figure 4.13). The peak assigned as that of cereulide was first tentatively confirmed by adding a known amount of valinomycin into the sample, with co-elution observed. The same semi-purified extract, prepared for MEKC analysis, showed activity in the MTT cell cytotoxicity assay.
110 %("*, !%"&- &$"$$ ("$' -"'* $"$$ esreto eeld ocnrto a endfiutbcseo,a es npr,a part, in least at of, becasue difficult been has concentration cereulide of Measurment foods. suspected of consumption before the illness prevent of could possibility toxin emetic for methods detection Reliable poisoning. food emetic the as of well detection as Therefore, toxin 1990). emetic (Lund, min 90 is for cereulide, C toxin, 121 the withstanding as heat-stable, remains, highly still poisoning food emetic of risk the treatment, heat toxin Lund, emetic and though (Granum even food Moreover, in 1997). performed is toxin emetic whereas host the of intestine cereus Bacillus Conclusions 4.4 h lcrpoei odtosae notdcplayo 1c 3. mt h detector] the to cm [36.4 SDS. cm mM 41 75 of with capillary 8.5 X50 uncoated pH are: of buffer conditions borate electrophoretic mM The 20 is buffer electrophoresis The iue4.13 Figure Absorbance (mAU) μ ,2 ,0k;U bobnea 9 nm. 195 at absorbance UV kV; C,10 25 m, ECaayi fteotra apeemetic sample outbreak the of analysis MEKC irheltxni ugse ob rdcdmil ntesmall the in mainly produced be to suggested is toxin diarrhoeal '"$$ .cereus B. elcuti motn nmnmsn h ikof risk the minimising in important is count cell irto ie(min) time Migration 111 .cereus B. *"$$ Cereulide el r lmntdfo od by foods from eliminated are cells .cereus B. odpoisoning food -"$$ %"$ %)"$$ %&"$$ lack of suitable methods. While there is a lack of data on the amount of toxin needed to induce emesis in humans, reliable and quantitative analysis methods for cereulide will assist in gaining this information.
A capillary electrophoresis method, employing the principles of MEKC, has been developed that allows the quantitative analysis of the B. cereus emetic toxin, cereulide.
MEKC gives results quickly and simply, and holds advantages over HPLC, through simple sample handling and preparation, less sample required and lower cost. All these advantages of MEKC advance its application in the quantitative analysis of emetic toxin, cereulide. Purified cereulide was used to generate a calibration curve for the MEKC method and this appears to be the first study that actually uses cereulide instead of valinomycin in establishing a quantitative analytical method.
To demonstrate the applicability of the MEKC method, sixty reference Bacillus strains and a rice sample inoculated with an emetic B. cereus strain were analysed. B. cereus emetic strains produced the most cereulide whereas diarrhoeal strain produced much less. This indicates that some B. cereus strains can produce both emetic and diarrheoal toxins and might be classified as emetic and/or diarrhoeal strains according to their capability to produce toxin. This interpretation can be supported by the study of Hansen and Hendriksen (2001) who showed that the reference B. cereus emetic strain, F
4810/72, produced a positive amplicon with PCR reaction for detection of the Nhe complex genes which is one of the diarrhoeal enterotoxin complexes. This emetic-type strain also showed positive reaction with an immunoassay for detection of diarrhoeal toxin (Tecra BDE-VIA). The results of MEKC analysis revealed that other Bacillus
G XXY species, as well as diarrhoeal B. cereus strains, also are capable of producing cereulide- like molecules.
One of the main advantages of the MEKC method is simple sample preparation. In this study, a simple methanol extraction was applied and the extract subjected to CE directly.
Following analysis of an inoculated rice sample, food poisoning outbreak samples were tested by MEKC analysis.
In summary, a simple and rapid MEKC method for detection and quantification of cereulide was developed and the method was successfully applied to both semi-purified toxin extracts of cultures and food samples.
G
G XXZ Chapter 5
Diversity and, potential and actual toxigenicity of Bacillus strains in rice products
G 5.1 Introduction G
The genus Bacillus includes members that demonstrate a wide diversity with respect to their physiology and ecological niche to DNA sequence and gene regulation. The species of most interest tend to be known for their pathogenicity and are closely linked genetically (Schoeni and Wong, 2005). Therefore, typing techniques that have good discrimination power to strain level, and reliable and rapid detection methods for
(potentially) toxigenic strains are needed to trace pathogenic strains among the diversity of strains of Bacillus species and to investigate the relationship between these strains and food-borne outbreaks.
Of the molecular typing methods discussed in Chapter 2, the randomly amplified polymorphic DNA (RAPD) PCR based method was chosen for application to Bacillus isolates from raw and cooked rice samples as it has several advantages such as good discrimination power, ease of interpretation and performance (Maslow et al., 1993;
Arbeit, 1995). Although RAPD is sensitive to annealing temperature, reproducibility has been achieved in many previous studies (Farber and Addison, 1994; Neiderhauser et al.,
1994; van Leeuwen et al 1996; Truong et al., 2000; Dautle et al., 2002) and the RAPD method has been successfully applied to the differentiation, characterisation and identification of Bacillus species (Brousseau et al., 1993; Woodburn et al., 1995;
Ronimus et al., 1997; Hansen et al., 1998; Zhang et al., 2002).
Traditionally, detection of B. cereus toxins has relied upon in vivo testing procedures
(Jackson, 1991, 1993), however, these methods are time-consuming and labor intensive to perform. Thus, immunological assays were developed for the detection of
115 enterotoxin(s). Two different immunoassays are commercially available which are the B. cereus enterotoxin-reversed passive latex agglutination (BCET-RPLA) test (Oxoid) and the sandwich ELISA or Bacillus diarrhoeal enterotoxin visual immunoassay (BDE-VIA) test (Tecra). Although immunoassays exist for the detection of diarrhoeal toxins, development of an immunoassay for emetic toxin was considered difficult due to its low antigenicity (Mikami et al., 1994). However, Tecra´ International is developing such an assay, with a prototype currently available. G
An approach for assessing the potential toxigenicity of Bacillus strains is the detection of specific, unique DNA sequences of genes of enterotoxic and emetic toxins. PCR- based methods have been developed for detection of the genes encoding diarrhoeal toxins (Granum et al., 1996; Asano et al., 1997; Granum et al., 1999; Yuan et al., 2002), while very recently, two studies (Ehling-Schulz et al., 2004; Toh et al., 2004) have reported PCR-based detection of genes associated with cereulide production.
This chapter reports the diversity and the toxicity (toxigenicity) of strains isolated from raw and cooked rice samples, determined using molecular, immunological and chromatographic methods. This includes use of a recently developed PCR method for detection of a gene associated with cereulide production, as well as a prototype ELISA and the capillary electrophoresis method reported earlier in this thesis (Chapter 5) for detection of cereulide.
116 5.2 Materials and Methods G
5.2.1 Typing of isolated Bacillus strains by RAPD-PCR
5.2.1.1 DNA preparation (crude DNA extract)
A DNA extraction method from pure cultures was adapted from the method of Ronimus et al. (1997). Isolates from rice samples (Section 3.2.3.1 in chapter 3) were streaked on
TSA plates and incubated at 30ºC for 20 – 24 h. Two or three well-isolated colonies were taken using a sterile disposable loop and suspended in 1.0 mL of 50 mM Tris-HCl,
100 mM Nacl, 20 mM EDTA, pH 8.0. This cell suspension was centrifuged at 4,000佴g for 10 min at ambient temperature and the pellet was resuspended in 0.1 – 1 mL of
Milli-Q water (Millipore, Bedford, MA, USA), with the amount used reflecting pellet size. The cells were lysed by boiling for 15 min, transferred to ice immediately, then centrifuged at 12,000佴g for 10 min to remove cell debris. The supernatant was carefully aspirated and brought to a standard absorbance at 260 nm (A260) and 280 nm
(A280) using Milli-Q water to obtain an absorbance (A260 = 0.3) of DNA. The purity of the DNA was checked by determining the A260/A280 ratio using a spectrophotometer
(UV-1601 Spectrophotometer Shimadzu Corporation, Japan).
5.2.1.2 RAPD-PCR Primers
Nine arbitrary 10-mer primers of the series OPR, OPA and OPG (Table 5.1) were tested for the amplification of random segments of genomic DNA of Bacillus species. After screening of these primers with reference Bacillus strains, selected primers were used for application to all Bacillus isolates from rice products.G
117 Table 5.1 Primers used in this study. Primers Primer sequence %GC References OPR 2 5’-CACAGCTGCC 70.0 Ronimus et al. (1997) OPR 13 5’-GGACGACAAG 60.0 Ronimus et al. (1997) OPR 16 5’-CTCTGCGCGT 70.0 Ronimus et al. (1997) OPA 2 5’-TGCCGAGCTG 70.0 Zahner et al. (1999) OPA 3 5’-AGTCAGCCAC 60.0 Artiushin & Minion (1996) OPA 17 5’-GACCGCTTGT 60.0 Ankarloo et al. (2000) OPG 5 5’-CTGAGACGGA 60.0 Daffonchio et al. (1998) OPG 8 5’-TCACGTCCAC 60.0 Daffonchio et al. (1998) OPG 16 5’-AGCGTCCTCC 70.0 Daffonchio et al. (1998)
5.2.1.3 Conditions of RAPD-PCR and gel electrophoresis
The RAPD-PCR assay was performed in a reaction volume of 25 PL containing: 4.0 mM MgCl2, 2.5 PL 10×buffer (1× in final concentration), 200 PM dNTP, 1.25 U
AmpliTaq, 1 PM of each primer and 2.5 PL of diluted boiled-cell supernatant (Ronimus et al., 1997). The PCR cycles consisted of an initial denaturing step at 94qC for 4 min, followed by 40 cycles of DNA denaturation at 94qC for 1 min, primer annealing at 35qC for 1 min, and DNA extension at 72qC for 2 min, then a final extension step at 72qC for
5 min (PC-960 air cooled thermal cycler, Corbett research, Australia) (Daffonchio et al.,
1998). A negative control or blank was included in every PCR experiment and consisted of the PCR reaction mix with the exclusion of any DNA to assess the presence of any contamination in the PCR reagent pre-mix. RAPD reactions (10 PL) were electrophoresed in a 1.5% (w/v) agarose gel (ICN biomedicals Inc., Ohio, USA). The gel, which was immersed in TBE buffer, was run at 90 V for 150 min. AmpliSizeTM
Molecular Ruler 50 – 2,000 base pairs (bp) ladder was used as a molecular weight
118 marker (Bio-Rad, Sydney, Australia). After migration of DNA bands, the gel was stained with ethidium bromide for 30 min and destained for 30 min, and then photographed on a UV transilluminator (Sigma, Sydney, Australia).
5.2.1.4 Gel analysis (data analysis)
The RAPD profiles (band pattern) in photographs were scanned, and the collected data were normalized and further processed using Gelcompar 4.0 software (Applied Maths,
Kortrijk, Belgium). Generation of the cluster analysis in a dendrogram was based on
Dice’s similarity coefficient (SD) and the un-weighted pair group method using arithmetic averages (UPGMA). The RAPD-PCR products greater than 2000 bp were considered unreliable (Ronimus et al., 1997), and they were not included in the analysis.
G G 5.2.2 Identification and confirmation of Bacillus species
5.2.2.1 Identification of Bacillus species by partial sequencing and B. cereus gyrase
B gene (gyrB) analysis
The DNA extraction for partial sequencing and PCR analysis of B. cereus gyrase B gene
(gyrB) was carried out as described in Section 3.2.5.1 (Chapter 3).
5.2.2.1.1 Identification of Bacillus species by partial sequencing
Partial 16S rDNA sequencing for identification of the Bacillus isolates was performed as described in Section 3.2.5.2.
5.2.2.1.2 PCR detection of gyrB
PCR amplification of the gyrase B (gyrB) gene of B. cereus was carried out as described
119 by Yamada et al. (1999) except only half the reaction volume was used. The PCR primers used for differentiation of B. cereus from B. thuringiensis were BC1 (5’-
ATTGGTGACACCGATCAAACA-3’) and BC2r (5’-TCATACGTATGGATGTTATTC-
3’). Reaction volumes of 50 PL contained 50 ng of genomic DNA, 200 PM of each dNTPs, 2.5 U of Taq DNA polymerase and primers at 0.4 PM each in reaction buffer
(10 mM Tris-HCl; 1.5 mM MgCl2; 50 mM KCl; pH 8.3). Negative control reaction mixtures with no added template DNA were included in every experiment. The amplification of B. cereus-specific fragments (365 bp) was performed using 30 cycles, each consisting of 60 s at 94ºC, 90 s at 58ºC, and 150 s at 72ºC, with a final extension step at 72ºC for 7 min. After DNA amplification, PCR products were analysed by gel electrophoresis in 1% (w/v) agarose gels in TBE buffer, stained with ethidium bromide, and visualized under UV transillumination.
5.2.2.2 Identification and confirmation of Bacillus species with the API 50CH
The API 50CH kit (bioMérieux, Australia), supplemented with the API 20E kit
(bioMérieux, Australia), were used for confirmation and identification of Bacillus species. B. cereus strain 7401 and representative Bacillus isolates from rice products were streaked onto TSA plates and incubated at 30ºC for 18 h. For the API 50 CH test, several well-isolated colonies were mixed in API 50CHB/E medium until the turbidity of suspension was equivalent to a McFarland turbidity standard of 2 with an ATB®
Densitometer. The test strips (tubes) were filled with the inoculated API 50CHB/E medium and incubated for 24 h (· 2 h) and 48 h (· 6 h) at 29ºC · 2ºC according to the manufacturer’s instructions. The API 20E was also inoculated, incubated and read according to manufacturer’s instructions. The results were analyzed with the ATB Plus
(bioMérieux) computer program to identify isolates to species level.
120 5.2.3 PCR analysis for detection of enterotoxin and emetic toxin gene
DNA for detection of enterotoxin and emetic toxin genes was extracted as described in
Section 3.2.5.1.
5.2.3.1 PCR conditions for detection of enterotoxin genes
For the detection of the B component of haemolysin BL (hblA) and enterotoxin BceT
(bceT) genes, primer mixture A (Table 5.2) was prepared and used in a multiplex PCR
(Yuan et al., 2002). Primer set B (Table 5.2) was used for detection of the nheA gene for the isolates of the B. cereus group that did not harbour any of the genes encoding hemolysin BL or enterotoxin T (Ghelardi et al, 2002). Information of PCR primer sequences, locations and PCR product size are given in Table 5.2.
The concentration of multiplex reaction components such as primers, dNTPs, MgCl2, template DNA, Taq DNA polymerase was optimised to yield three different bands, representing hemolysin BL and enterotoxin T genes. Finally, the PCR with primer mixture A was performed in 25 PL reaction volumes containing 50 ng of genomic DNA,
0.2 PM of each of primers HblA 1, HblA2 (for hblA), B 1F and B 7R (for HBL B- component gene), 0.1 PM of primer ET F and ET R (for bceT), 0.2 mM of each dNTP, 2 mM MgCl2, 2.5 U of Gold-Taq DNA polymerase in reaction buffer. The PCR mixture
(50 PL) for detection of nheA gene contained 50 ng of genomic DNA, deoxynucleoside triphosphates at a concentration of 0.2 mM each, 2.5 U of Taq DNA polymerase and 0.2
PM each primer in supplied buffer (10 mM Tris-HCl; 1.5 mM MgCl2; 50 mM KCl; pH
8.3). Negative control reaction mixtures with no added template DNA were included in every experiment. The PCR was performed using a DNA thermal cycler of Gene Amp
PCR system 9700 (Applied Biosystems, California, USA). The temperature cycling 121 profiles for primer sets A and B are given in Table 5.2. PCR products (2 – 4 »L) were run on 1% (w/v) agarose gels and visualised by staining with ethidium bromide.
5.2.3.2 PCR condition for detection of emetic toxin gene marker
The primer set and the PCR conditions for detection of the emetic toxin gene marker were as described by Toh et al. (2004). The details of emetic toxin gene marker-specific primers, BEF and BER (primer set C) and reaction conditions are given in Table 5.2.
PCR amplification was performed in a reaction volume of 30 PL containing 3 PL 25
-1 mM MgCl2, 3 PL 10 u buffer, 2 PL 2 mM dNTPs, 0.1 PL 5.5 U P Taq polymerase, 1
PL each of the forward and reverse primer in 19 PL Milli-Q water, and 1 PL of DNA
(100 ng) was included in the reaction volume. Negative control reaction mixtures with no added template DNA were included in every experiment. PCR products were analysed by gel electrophoresis in 1% (w/v) agarose gels in TBE buffer and visualised by staining with ethidium bromide.
122 Table 5.2 Primers used for detection of genes of diarrhoeal enterotoxins and emetic toxin gene marker Primer Sequence (5’-3’) Product PCR amplification Mixture (bp) Initial denaturation Main cycle Final extension
A HblA1 GCTAATGTAGTTTCACCTGTAGCAAC 874 95qC for 60 s HblA2 AATCATGCCACTGCGTGGACATATAA 95qC for 10min 55qC for 60 s 30 cycles 72qC for 10min B 1F ACGAACAATGGAGATACGGC 1066 72qC for 180 s B 7R ATTTTTGTGGAGTAACAGTTTCTAC ET F TTACATTACCAGGACGTGCTT 428 ET R TGTTTGTGATTGTAATTCAGG
B nheA 344 S TACGCTAAGGAGGGGCA 499 95qC for 5min 94qC for 15 s 72qC for 7 min nheA 843 A GTTTTTATTGCTTCATCGGCT 55qC for 45 s 30 cycles 72qC for 120 s
C BE F ACTTAGATGATGCAAGACTG 850 94qC for 2 min 94qC for 10 s 72qC for 7 min BE R TTCATAGGATTGACGAATTTT 50qC for 20 s 30 cycles 70qC for 80 s
123 5.2.4 Enzyme immunoassay for detection of enterotoxin and emetic toxin
5.2.4.1 Detection of diarrheoal enterotoxin
Sample preparation and use of the Tecra Bacillus diarrhoeal enterotoxin (BDE) Visual
Immunoassay kit (NSW, Australia) were performed according to manufacturer’s instructions. Samples were prepared by inoculating a loopful of biomass from a 24 h culture on TSA into 10 mL of BHI, incubated for 16 – 18 h at 37ºC. After incubation, samples were centrifuged for 10 min at 3000 × g and the supernatants were collected and used as samples for detection of enterotoxin after addition of the sample additive supplied with the ELISA kit. The ELISA protocol is shown in Figure 5.1.
5.2.4.2 Detection of emetic toxin
The procedure for the cereulide enzyme immunoassay was based on instructions provided by Tecra (unpublished) with slight modification, in that cultures were incubated for 48 h instead of 24 h, as cereulide production commences after the end of logarithmic growth (approximately 24 h) (Häggblom et al., 2002) Samples were prepared by inoculating a loopful of biomass of Bacillus cells from 24 h culture on TSA into 10 mL of TSB. Following 48 h of incubation in a shaking incubator (150 rpm) at
30ºC, samples were centrifuged for 10 min at 3000 × g and the supernatants were collected. The collected supernatants were filtered through 0.45 Pm disposable filters to remove residual bacteria and mixed with an equal volume of methanol. The ELISA was performed as for the BDE ELISA (Figure 5.1).
124 Add 200 ȝL of sample (including test and controls) into an ELISA well p Cover the wells with plastic cling wrap film to minimise evaporation and incubate for 2 h at 35 – 37qC p Empty the wells and wash four times with washing solution p Add 200 ȝL of conjugate to the wells and incubate covered for 1 h at ambient temperature p Empty the wells and wash five times with washing solution p Add 200 ȝL reconstituted substrate to the wells and incubate at 25qC Incubate for minimum of 30 min, and then read the result
G
G Figure 5.1 ELISA procedure for detection of diarrhoeal and emetic toxins
5.2.4.3 Reading and interpretation of results
The absorbance of samples was read at 405 nm with the reference wavelength at 490 nm using an ELISA plate reader (AD 340C reader, Beckman, Sydney, Australia) blanked against air. A positive value was interpreted as an absorbance 0.2.
G G 5.2.5 Analysis of emetic toxin by micellar electrokinetic electrophoresis (MEKC)
Cultures of selected representative Bacillus isolates were prepared as samples for
MEKC analysis as described in Section 4.2.5.2.2 and MEKC analysis was carried out as described in Section 4.2.5.3.2.
125 5.3 Results and discussion G
5.3.1 Typing of Bacillus isolates by RAPD-PCR
5.3.1.1 Optimisation of RAPD-PCR analysis
Of the nine primers tested, OPR13 and OPA3, which produced diverse and reproducible polymorphic band patterns, with reference Bacillus strains including species of B. cereus, B. thuringiensis, B. mycoides, B. subtilis, B. licheniformis, B. pumilus, B. megaterium, B. circulans, B. sphaericus, were chosen for RAPD-PCR analysis of all the
Bacillus isolates from rice samples.
The effect of concentration of primer was examined with primer OPR13 on the same fifteen representative strains. Three different primer concentrations, 0.4, 1 and 2 µM, were examined, the use of 1 and 2 µM showing better discrimination power with a greater number of fragments produced, than when 0.4 µM was used. However, there was no discernible difference in terms of the number or intensity of fragments when 1 or 2 µM was tested with the reference Bacillus strains. Therefore, in the current study, a primer concentration of 1 µM was chosen. This result differed slightly from the study of
Ronimus et al. (1997), in which increasing the primer concentration from 0.2 to 2 µM increased both the number of fragments produced and their intensity. They also reported that the intensity of bands increased remarkably between 0.2 and 0.5 µM and steadily as the primer concentration was raised from 0.5 to 1.0, 1.5 and 2.0 µM with no further increase above 2.0 µM.
To determine the consistency and reproducibility of the RAPD methodology, RAPD-
126 PCR with 1 µM concentration of OPR 13 was performed in triplicate (3 PCR runs) with the same fifteen Bacillus representative strains. There was almost no deviation in RAPD patterns between the triplicate reactions, though some minor variation in intensity of the amplicons was observed.
5.3.1.2 Diversity of Bacillus isolates
All 351 isolates confirmed as bacilli, obtained using the centrifugation-spread plating method (Chapter 3), were tested using the optimised RAPD protocol to examine diversity. After analysis of the isolates with both OPR 13 and OPA 3, the isolates were clustered based on the RAPD patterns obtained with OPR 13 as it gave better discrimination than the OPA 3 primer. This was particularly true for species in the B. cereus group, for which 32 RAPD profiles were generated with OPR 13 whereas only
20 were obtained with OPA 3 (Table 5.3). Therefore, the primer OPR 13 was further selected between two primers and the primer OPA 3 was used for only comparison purpose (Tables 5.3 – 5.7).
Of the 351 isolates analysed using RAPD-PCR, if several had the same morphology
(Chapter 3) as well as the same RAPD pattern (100% similarity as assigned using the
Gelcompar 4.0 program), one or two (sometimes several) representative strains were subjected to partial 16S rDNA sequencing. BLAST analysis of partial sequences revealed that all isolates showed high similarity (mostly 97 –G100%) to the genera
Bacillus or Paenibacillus and, in total, nine different species were found from rice products (Tables 5.3 – 5.7). The 29 representative isolates (and one reference strain of B. cereus NC 7401) were also subjected to biochemical testing using the API systems and all but one (one of Paenibacillus isolate, Section 5.3.1.2.7) were successfully identified
127 as Bacillus species or Paenibacillus species and the results generally correlated well with the results of partial rDNA sequencing (Tables 5.3 – 5.7).
5.3.1.2.1 B. cereus, B. thuringiensis and B. mycoides
A total of 85 isolates form raw rice and 41 isolates from cooked rice were classified, based on the RAPD analysis, into 32 groups of B. cereus/B. thuringiensis and B. mycoides (Table 5.3). Representative strains including at least one strain from each group were subjected to partial sequencing. Partial sequencing was not able to differentiate B. cereus and B. thuringiensis as BLAST analysis showed that isolates share similarity with sequences for both species. Therefore, these isolates were assigned to B. cereus or B. thuringiensis after detection of the gyrase B gene (gyrB) of B. cereus.
Sequencing results revealed that none of the isolates was B. weihenstephanensis or B. pseudomycoides but one strain was very similar (97%) to B. cereus/ B. thuringiensis as well as to B. anthracis. Thus, a motility test was performed with this isolate as a confirmation test and the result showed that the isolate was motile, a characteristic of B. cereus/B. thuringiensis. Finally, this isolate was classified as B. thuringiensis on the basis of PCR analysis of the gyrase B gene (gyrB). Six isolates (five from raw rice and one from cooked rice) were allocated to B. mycoides by partial sequencing and biochemical properties. Though BLAST analysis of partial sequencing reveled that these isolates had the same similarity to B. cereus and B. mycoides, B. mycoides can be differentiated from other species in the B. cereus group based on its rhizoid growth. A motility test was performed with these six isolates and the results supported identification as B. mycoides because all isolates were non-motile.
Finally, 87 isolates from rice products were classified based on the RAPD-PCR analysis
128 with primer OPR 13 into 19 groups of B. cereus and 33 isolates were allocated into 9 groups of B. thuringiensis (Table 5.3). Six isolates of B. mycoides showed four different
RAPD profiles with primer OPR 13 (Table 5.3) and this species separately clustered into one group from other species (Figure 5.3). RAPD-PCR with primer OPA 3 generated only 20 different profiles in the B. cereus group (Table 5.3) and B. mycoides isolates were not clustered together (Figure 5.2).
RAPD-PCR typing showed that isolates in this Bacillus group were very heterogeneous, as previously reported for B. cereus (Brousseau et al., 1993; Andersson et al., 1999) and
B. thuringiensis (Adelaida et al., 2003). However, some of the RAPD patterns were common among B. cereus and B. thuringiensis isolates from different samples (for example, MC Lab-1 and MC Ram-1, and MR 14-11 and MR 31-1). Comparatively more diverse RAPD patterns generated with OPR 13 were observed among the B. cereus isolates from raw rice (13 RAPD profiles) rather than from cooked rice (8 RAPD profiles) and among these RAPD patterns, 2 patterns (patterns 3 and 10) were found in isolates from both raw and cooked rice. These results suggest that the cooking procedure used for the preparation of cooked rice may be selective for certain RAPD types and other types may derive from post-process contamination with Bacillus cells or spores. These results suggest that the cooking procedure used for the preparation of cooked rice may be selective for certain RAPD types. Other types may have derived from other spores belonging to other types in raw rice which are activated for germination by heating and also may have derived from post-cross contamination with
Bacillus cells or spores.
An earlier study (Gilbert and Parry 1977) revealed that the H1 serotype was prevalent
129 among B. cereus isolates from cooked rice associated with the emetic type of B. cereus food poisoning, and this serotype was rarely found in raw rice samples. The RAPD pattern for the B. cereus MC Lab-1 strain, shown to survive cooking (Section 3.3.3.2,
Chapter 3), may indicate that this RAPD pattern represents the heat-tolerant B. cereus strain (spores) though same pattern was not found among isolates from raw rice samples.
Unlike B. cereus isolates, there were no isolates of B. thuringiensis or B. mycoides with the same RAPD pattern in raw and cooked rice.
B. cereus and B. thuringiensis cannot be differentiated using biochemical test systems.
Among phenotypic tests, these species can be differentiated only by the observation of toxin crystals in B. thuringiensis under the microscope but often it is hard to identify the toxin crystals, resulting in misidentification. Comparative studies of the sequence of
16S rRNA genes suggested that the four members of the B. cereus group, including B. cereus, B. thuringiensis, B. mycoides and B. anthracis, are so closely related that they could be considered one species (Ash et al., 1991a, 1991b). A later study (Helgason et al., 2000b) agreed with this consideration, except that B. mycoides may be a separate species, based on analysis using multilocus enzyme electrophoresis (MEE). A recent analysis has tried to overcome the apparent taxonomical uniformity presented by studies of ribosomal RNA genes (Yamada et al, 1999). The authors cloned and sequenced gyrB genes from B. cereus, B. thuringiensis and B. anthracis. In spite of minor differences in
DNA sequence they designed PCR primer that differentiated between strains of these organisms. Also, earlier DNA-DNA hybridisation studies of the B. cereus group
(Kaneko, et al., 1978; Seki et al., 1978) also revealed a high degree of similarity between species but some inconsistencies were noted with respect to overall relatedness of species and thus it may not be appropriate to place them in the same species (Jenson
130 and Moir, 2003). In particular, B. mycoides appeared to comprise two distinct homology groups (Nakamura and Jackson, 1995) and these can also be differentiated from other species in the B. cereus group based on their rhizoid growth even though there are some exceptions.
B. thuringiensis is known to share marked homology with B. cereus (Seki et al., 1978;
Nakamura, 1994). The results of RAPD-PCR analysis in the current study reinforce the close relationship of these two species. While B. mycoides isolates clustered into one group with primer OPR 13, distinct from B. cereus and B. thuringiensis, some B. thuringiensis isolates were closer to some B. cereus isolates than others of B. thuringiensis. However, the reliability of the PCR analysis for detection of the gyrase B gene (gyrB) might also affect the results in terms of differentiation of B. cereus from B. thuringiensis. One of the reference cultures of B. thuringiensis yielded a gyrB amplicon for B. cereus, identifying it as B. cereus, when it was pre-tested with a collection of reference cultures (five each of B. cereus, B. thuringiensis and other Bacillus species) in the current study. Therefore, this method may not perfectly differentiate B. cereus from
B. thuringiensis, though it was considered previously as useful in the differentiation of
B. cereus and B. thuringiensis (Yamada, et al., 1999). However, there is also a possibility that the culture previously identified B. thuringiensis was classified incorrectly.
131 Figure 5.2 Cluster analysis of RAPD-PCR profiles obtained with primer OPA 3 for strains in the B. cereus group
132 Figure 5.3 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for strains in the B. cereus group
133 Table 5.3 Identification results and the RAPD-types of identified B. cereus, B. thuringiensis and B. mycoides strains isolated from rice RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns B. cereus MR 6-1 17(4) 2 99%, B. cereus/B. thuringiensis MR 7-4 6(1) 2 99%, B. cereus/B. thuringiensis MR 9-1 7(13) 20 99%, B. cereus/B. thuringiensis 99.8%, B. cereus 1 MR 13-2 19(1) 19 99%, B. cereus/B. thuringiensis MR 13-6 4(4) 3 98%, B. cereus/B. thuringiensis 93.6%, B. cereus 1 MR 14-2 12(9) 12 97%, B. cereus/B. thuringiensis MR 15-3 14(2) 6 98%, B. cereus/B. thuringiensis MR 19-1 13(15) 13 97%, B. cereus/B. thuringiensis 94.5%, B. cereus 1 MR 22-1 10(3) 4 97%, B. cereus/B. thuringiensis MR 24-1 15(2) 6 98%, B. cereus/B. thuringiensis MR 30-1 16(4) 6 97%, B. cereus/B. thuringiensis 96.9%, B. cereus 1 MR 31-6 1(1) 4 98%, B. cereus/B. thuringiensis MR E-7 3(3) 4 98%, B. cereus/B. thuringiensis MC 6-1 3(8) 6 97%, B. cereus/B. thuringiensis 91.3%, B. cereus 1 MC 15-4 5(3) 5 97%, B. cereus/B. thuringiensis MC 21-2 10(2) 7 98%, B. cereus/B. thuringiensis MC 24-2 9(1) 2 98%, B. cereus/B. thuringiensis MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples
134 Table 5.3 Identification results and the RAPD-types of identified B. cereus, B. thuringiensis and B. mycoides strains isolated from rice (continued) RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MC 27-3 8(1) 2 99%, B. cereus/B. thuringiensis MC Ivan-1 2(2) 6 98%, B. cereus/B. thuringiensis MC Lab-1 18(5) 14 99%, B. cereus/B. thuringiensis 93.6%, B. cereus 1 MC Beef-1 11(3) 6 99%, B. cereus/B. thuringiensis MC Ram-1 18(-a) 14 97%, B. cereus/B. thuringiensis B. thuringiensis MR 6-5 25(1) 18 97%, B. cereus/B. thuringiensis MR 10-4 24(4) 11 98%, B. cereus/B. thuringiensis MR 14-6 23(5) 15 99%, B. cereus/B. thuringiensis 91.3%, B. cereus 1 MR 14-11 26(7) 18 97%, B. cereus/B. thuringiensis MR 31-1 26(-b) 18 99%, B. cereus/B. thuringiensis MR cr1-6 27(1) 18 97%, B. cereus/B. thuringiensis MC p4-8 28(2) 10 97%, B. cereus/B. thuringiensis 99.9%, B. cereus 1 MC 20-2 22(4) 8 97%, B. cereus/B. thuringiensis MC 21-3 21(7) 8 99%, B. cereus/B. thuringiensis MC 21-8 21(-c) 8 99%, B. cereus/B. thuringiensis MC 22-9 20(2) 1 98%, B. cereus/B. thuringiensis MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples a Same isolates as for MC Lab-1; b Same isolates as for MR 14-11; c Same isolates as for MR MC 21-3 135 Table 5.3 Identification results and the RAPD-types of identified B. cereus, B. thuringiensis and B. mycoides strains isolated from rice (continued) RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns B. mycoides MR 14-1 29(1) 16 97%, B. cereus/B. mycoides MR 20-6 32(1) 9 97%, B. cereus/B. mycoides MR 21-1 30(3) 16 98%, B. cereus/B. mycoides MC 21-1 31(1)17 97%, B. cereus/B. mycoides 73.9%, B. mycoides MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples
136 5.3.1.2.2 B. subtilis/B. mojavensis and B. amyloliquefaciens
A total of 137 isolates identified as B. subtilis/B. mojavensis and B. amyloliquefaciens were differentiated into 32 RAPD patterns with primer OPR 13 as well as primer OPA 3
(Table 5.4). Unlike with the B. cereus group, primer OPR 3 also showed good discrimination power with B. subtilis/B. mojavensis. Twenty-seven different RAPD patterns with primer OPR13 were observed with B. subtilis/B. mojavensis isolates from raw rice whereas B. subtilis/B. mojavensis isolates from cooked rice yielded only 7 different patterns. One B. amyloliquefaciens isolate had a different RAPD pattern with primer OPR 13 to that from B. subtilis/B. mojavensis isolates. The analysis of RAPD profiles showed that the B. subtilis/B. mojavensis isolates in rice products were quite diverse, similar to the isolates in the B. cereus group. Sarkar et al. (2002) also found that the B. subtilis isolates from spontaneously fermented soybeans (Indian Kinema) and locust beans (African Soumbala) showed a high degree of heterogeneity.
B. subtilis and B. mojavensis could not be differentiated by partial rDNA sequencing but these isolates were obviously clustered into two groups by RAPD pattern with primer
OPR 13 (Figure 5.5). Therefore, each clustered group may represent each species. The similarity of RAPD profiles with primer OPR 13 between B. subtilis isolates was between 29 and 80% and between 31 and 86% for stains grouped into clusters I and II, respectively, and 24% between the two clusters (Figure 5.5). Unfortunately, with primer
OPA3, isolates were not clustered into the same groups that were generated with primer
OPR 13. The RAPD similarity was low (between 5 and 85%) with OPA3 (Figure 5.4).
The isolates were assigned to B. subtilis/B. mojavensis based on partial rDNA sequencing as they could not be identified using the API systems. B. subtilis and B.
137 mojavensis could not be differentiated at the species level without additional methods such as DNA-DNA hybridisation (Slepecky and Hemphill, 1992) or partial gyrA gene sequence analysis (Chun and Bae, 2000). Further studies were not performed for the differentiation of these two species in this study. Only one isolate (MC P-1) from cooked rice was identified as B. amyloliquefaciens/B. subtilis by partial rDNA sequencing, with 99% similarity from BLAST analysis (Table 5.4). This isolate were identified and confirmed as B. amyloliquefaciens with further support by analysis of the
API systems.
Figure 5.4 Cluster analysis of RAPD-PCR profiles obtained with primer OPA 3 for B. subtilis/B. mojavensis and B. amyloliquefaciens strains
138 Figure 5.5 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. subtilis/B. mojavensis and B. amyloliquefaciens strains
139 Table 5.4 Identification results and the RAPD-types of identified B. subtilis/B. mojavensis strains isolated from riceG RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MR 5-7 12(2) 14 99%, B. subtilis/B. mojavensis MR 7-2 13(1) 28 99%, B. subtilis/B. mojavensis MR 8-4 10(3) 31 97%, B. subtilis/B. mojavensis MR 8-5 19(4) 30 99%, B. subtilis/B. mojavensis MR 9-3 11(5) 31 97%, B. subtilis/B. mojavensis 95.9%, B. subtilis MR 10-2 24(8) 29 99%, B. subtilis/B. mojavensis MR 10-3 24(- a) 29 98%, B. subtilis/B. mojavensis MR 10-6 8(6) 31 98%, B. subtilis/B. mojavensis MR 13-1 17(4) 31 97%, B. subtilis/B. mojavensis MR 13-9 25(2) 6 99%, B. subtilis/B. mojavensis MR 16-2 1(1) 9 97%, B. subtilis/B. mojavensis MR 16-5 2(5) 16 99%, B. subtilis/B. mojavensis 98.3%, B. subtilis MR 16-6 2(-b) 24 98%, B. subtilis/B. mojavensis MR 17-1 20(2) 22 99%, B. subtilis/B. mojavensis MR 17-6 26(2) 25 100%, B. subtilis/B. mojavensis MR 18-1 31(1) 23 99%, B. subtilis/B. mojavensis MR 19-3 32(3) 7 98%, B. subtilis/B. mojavensis MR 20-1 22(4) 11 99%, B. subtilis/B. mojavensis MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples a Same isolates as for MR 10-2; b Same isolates as for MR 16-5 140 Table 5.4 Identification results and the RAPD-types of identified B. subtilis/B. mojavensis strains isolated from rice (continued) RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MR 22-2 25(- c) 6 97%, B. subtilis/B. mojavensis MR 22-3 28(1) 3 98%, B. subtilis/B. mojavensis MR 24-2 6(1) 2 98%, B. subtilis/B. mojavensis MR 28-2 23(10) 1 99%, B. subtilis/B. mojavensis 96.0%, B. subtilis MR 30-4 8(-d) 2 96%, B. subtilis/B. mojavensis 97.4%, B. subtilis MR 30-5 4(2) 12 97%, B. subtilis/B. mojavensis MR 31-4 3(1) 15 99%, B. subtilis/B. mojavensis MR 31-7 23(-e) 11 98%, B. subtilis/B. mojavensis MR E-2 21(5) 4 98%, B. subtilis/B. mojavensis MR E-3 21(-f) 4 98%, B. subtilis/B. mojavensis MR T-1 27(1) 4 99%, B. subtilis/B. mojavensis MR T-5 9(6) 27 98%, B. subtilis/B. mojavensis MR cr1-1 14(7) 21 100%, B. subtilis/B. mojavensis MR cr1-2 7(1) 31 98%, B. subtilis/B. mojavensis MR cr2-9 14(-g) 21 98%, B. subtilis/B. mojavensis MR cr2-10 16(1) 19 97%, B. subtilis/B. mojavensis 96.4%, B. subtilis MR cr3-7 4(2) 12 98%, B. subtilis/B. mojavensis MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples c Same isolates as for MR 13-9; d Same isolates as for MR 10-6; e Same isolates as for MR 28-2; f Same isolates as for MR E-3;Gg Same islates as for MR cr1-1G 141 Table 5.4 Identification results and the RAPD-types of identified B. subtilis/B. mojavensis and B. amyloliquefaciens strains isolated from rice (continued)
RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MC 4-2 25(1) 8 99%, B. subtilis/B. mojavensis MC 4-3 18(3) 26 96%, B. subtilis/B. mojavensis 95.7%, B. subtilis MC 10-1 20(17) 22 98%, B. subtilis/B. mojavensis MC 15-2 12(19) 5 97%, B. subtilis/B. mojavensis MC 22-3 12(- h) 18 97%, B. subtilis/B. mojavensis MC 22-7 12 (1) 13 99%, B. subtilis/B. mojavensis MC 25-3 30(1) 10 97%, B. subtilis/B. mojavensis MC 27-1 29(1) 3 98%, B. subtilis/B. mojavensis 93.1%, B. subtilis MC Ivan-3 5(1) 20 99%, B. subtilis/B. mojavensis MC Beef-4-1 12(- h) 32 97%, B. subtilis/B. mojavensis MC Pork-1 12(- h) 32 98%, B. subtilis/B. mojavensis MC Chicken-2 12(- h) 32 99%, B. subtilis/B. mojavensis MC Chicken-3 12(- h) 32 98%, B. subtilis/B. mojavensis 98.3%, B. MC P-1 15(2) 16 99%, B. amyloliquefaciens/B. subtilis amyloliquefaciens MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples h Same isolates as for MC 15-2
142 5.3.1.2.3 B. licheniformis
Thirty-four isolates, yielding 12 RAPD patterns with primer OPR 13 and 9 RAPD patterns with primer OPA 3, were identified as B. licheniformis by partial 16S rDNA sequencing and the API systems (Table 5.5). With the OPR 13 primer, isolates from raw and cooked rice yielded nine and four RAPD patterns respectively, with only one common to both types of rice. RAPD-PCR profiles of isolates in this species were illustrated in Figure 5.6.
Similar to the B. cereus group, B. licheniformis isolates with the same RAPD profile were retrieved from different cooked rice samples. For example, MC 23-3 and MC Lab-
2 were isolated from retail cooked rice and steamed rice cooked in the laboratory using a rice cooker, respectively. The RAPD profile of these isolates may represent more heat- tolerant B. licheniformis strains because the latter isolate was shown to survive after cooking. The RAPD profiles generated with OPA 3 were the same for one of the raw rice isolates (MR 7-15) and cooked product isolates MC 23-3, MC Beef-8 and MC Lab-
2 (Table 5.5). These data suggest that the cooking process may select certain RAPD types, although these four isolates were further divided into two different RAPD types using the OPR 13 primer (Table 5.5). Similar result was reported by Sorokulova et al.
(2003). They found the same RAPD pattern of B. licheniformis isolates in different flour or ropy bread samples. They also found that some B. licheniformis strains survived the
o baking process and produced rope in baked bread during storage for 2 days at 37 C. G
143 Table 5.5 Identification results and the RAPD-types of identified B. licheniformis strains isolated from rice RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MR 7-15 3(3) 8 97%, B. licheniformis 99.9%, B. licheniformis MR 17-3 5(1) 4 96%, B. licheniformis 99.9%, B. licheniformis MR 18-3 1(1) 6 99%, B. licheniformis MR 19-5 6(8) 5 98%, B. licheniformis MR 23-8 12(1) 3 95%, B. licheniformis MR 25-3 9(2) 9 97%, B. licheniformis MR 28-3 11(3) 7 97%, B. licheniformis MR T-3 7(1) 1 97%, B. licheniformis MR E-8 8(2) 2 97%, B. licheniformis MC p4-5 10(4) 7 97%, B. licheniformis 99.9%, B. licheniforims MC p5-5 10(-a) 7 97%, B. licheniformis MC 23-3 2(6) 8 97%, B. licheniformis MC Lab-2 2(-b) 8 97%, B. licheniformis MC Beef-8 3(1) 8 99%, B. licheniformis MC Beef-9 4(1) 9 97%, B. licheniformis MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples a Same isolates as for MC p4-5; b Same isolates as for MC 23-3
144 Figure 5.6 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. licheniformis strains
5.3.1.2.4 B. pumilus
A group of 31 isolates yielded a diversity (11) of RAPD patterns with primer OPR 13, yet partial rDNA sequencing revealed greater than 98% similarity between 13 representative isolates and the reference B. pumilus (Table 5.6). Some isolates from the same sample showed different RAPD patterns with both primers. For example, MR 7-6 and MR 7-10 showed different RAPD patterns with the primer OPR 13 (11 and 8, respectively) as well as with the primer OPA 3 (6 and 1, respectively). This result indicates that a sample can be contaminated with more than one strain. None of the same RAPD patterns were observed among the isolates from raw and cooked rice.
145 Table 5.6 Identification results and the RAPD-types of identified B. pumilus strains isolated from rice RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns MR 7-6 11(2) 6 99%, B. pumilus MR 7-10 8(6) 1 98%, B. pumilus MR 8-1 6(8) 1 99%, B. pumilus MR 8-14 6(-a) 1 98%, B. pumilus MR 10-7 5(2) 2 98%, B. pumilus MR 14-14 3(4) 8 97%, B. pumilus 99.9%, B. pumilus MR 19-11 4(1) 7 98%, B. pumilus MR 20-3 7(2) 3 98%, B. pumilus MR 21-3 3(-b) 8 99%, B. pumilus 99.9%, B. pumilus MR r4-2p 1(1) 9 98%, B. pumilus MR cr2-7 9(2) 4 99%, B. pumilus MC 22-6 10(2) 5 97%, B. pumilus 99.9%, B. pumilus MC Chicken-7 2(1) 10 98%, B. pumilus MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples a Same isolates as for MR 8-14; bSame isolates as for MR 14-14
146 5.3.1.2.5 B. megaterium
The isolates most closely related to B. megaterium by 16S rDNA sequence analysis yielded eight different RAPD patterns (with OPR 13, Table 5.7), with the 42 – 80%
RAPD similarity between the isolates within the species (Figure 5.7).
Figure 5.7 Cluster analysis of RAPD-PCR profiles obtained with primer OPR 13 for B. megaterium
5.3.1.2.6 B. sphaericus/B.fusiformis
On the basis of partial 16S rDNA sequence analysis, seven isolates, previously clustered into three RAPD patterns, were affiliated with B. sphaericus and/or B. fusiformis (Table
5.7). There was 16S rDNA sequence similarity between isolates MR 23-4 and both B. sphaericus and B. fusiformis. The API systems could not differentiate these two species because two species (B. sphaericus and B. fusiformis) was identified as B. sphaericus in
API tests with a negative reaction for all carbohydrate tests. These bacteria are negative for the various physiological tests based on carbohydrate metabolism, which introduces problems for traditional taxonomists (Gordon et al., 1973).
147 The RAPD patterns of isolate MR 23-6 (identified as B. fusiformis) shared greater similarity with MR 2-1-2 (identified as B. fusiformis) than with isolate MR 23-4 (B. sphaericus/B.fusiformis by sequencing) even though the isolates of MR 23-6 and MR
23-4 came from the same sample.
5.3.1.2.7 Paenibacillus strains
Based on partial 16S rDNA sequencing, two isolates from raw rice were identified as
Paenibacillus amylolyticus (Table 5.7). However, one of the two isolates was identified as P. polymyxa (97.6%) by the API systems and the other isolate remained unidentified by those systems. This was the only case in the current study in which the API systems failed to provide a useful identification, with the closest match, at 56.8%, being B. circulans (Table 5.7).
At the species level, most Paenibacillus are globally distributed (Priest, 1993) and it has been reported that P. polymyxa is found in pasteurised milk and milk products (Griffiths and Philips, 1990; Tatzel et al., 1994). Generally, Paenibacillus is considered a harmless bacillus but opportunistic infection of humans and animals may be caused by some species in the genus (Fritze, 2002). Notably, P. alvei and P. laterosporus, and some other species such as P. larvae ssp. larvae, P. popilliae and P. lentimorbus are pathogenic toward honeybees or some Coleoptera larvae (Fritze, 2002).
148 Table 5.7 Identification results and the RAPD-types of identified strains of other bacilli isolated from rice RAPD Partial 16S API test Strain OPR 13 OPA 3 Closest similarity, % Closest similarity, % Patterns (No. of isolates) Patterns B. megaterium MR 5-1 8(2) 1 97%, B. megaterium 99.9%, B. megaterium 2 MR 14-13 1(1) 6 98%, B. megaterium MR 15-8 3(1) 5 97%, B. megaterium MR 17-8 4(1) 3 98%, B. megaterium MR 22-9 2(1) 5 97%, B. megaterium MR r2-2b 1(4) 6 98%, B. megaterium MR cr3-9 5(1) 4 98%, B. megaterium 99.7%, B. megaterium 2 MC 27-12 7(1) 2 97%, B. megaterium MC Beef-10 6(2) 2 97%, B. megaterium B. sphaericus /B. fusiformis MR 23-4 1(3) 1 97%, B. sphaericus/B. fusiformis MR 23-6 2(2) 2 97%, B. fusiformis 95.3%, B. sphaericus MC cr2-1-2 3(2) 3 97%, B. fusiformis 95.3%, B. sphaericus Paenibacillus 56.8%, B. circulans 2 MR 7-16 1(1) 1 99%, P. amylolyticus /38.1%, P. polymyxa MR 25-5 2(1) 1 98%, P. amylolyticus 97.6%, P. polymyxa MR, Strains isolated from raw rice samples; MC, Strains isolated from cooked rice samples
149 5.3.2 Toxigenicity of Bacillus isolates
A group of 140 isolates from raw and cooked rice, including representatives of all
RAPD types, were tested for their potential toxigenicity and actual toxin production. To investigate toxigenic or potentially toxigenic strains of bacilli, genes responsible for the emetic and diarrhoeal toxins were analysed. For detection of diarrheoal enterotoxin genes, multiplex PCR encoding genes for haemolysin (Hbl) and enterotoxin T (BceT) were used and the PCR amplification conditions optimised in this study. In contrast to well-established PCR detection methods for these diarrhoeal toxins (Hbl and BceT),
PCR-based detection for a gene associated with cereulide production was developed very recently (Toh et al., 2004) and this method was used for detection of this putative gene among isolates from rice products.
The production of toxins by isolates was determined using ELISA (Tecra) for both diarrhoeal and emetic toxins. Of the 140 isolates, 28 were also subjected to analysis using capillary electrophoresis, developed for quantitative analysis of cereulide (Chapter
4), and used for comparative purposes with the prototype ELISA.
5.3.2.1 B. cereus group
The results of PCR analysis as well as ELISA analysis showed that most of the B. cereus and B. thuringiensis isolates from rice products were toxigenic (Table 5.8). All of the B. cereus isolates and all but two of the B. thuringiensis isolates gave a positive result with the enterotoxin ELISA. However, for four isolates each of B. cereus and B. thuringiensis, the PCR did not yield the haemolysin (Hbl) and enterotoxin T (BceT) amplicons. As discussed in Chapter 2, five different enterotoxins have now been associated with the diarrhoeal syndrome of B. cereus, in contrast to the single emetic toxin.
150 Among those five enterotoxins, haemolysin (Hbl) and enterotoxin T (BceT) have been the primary focus of detection of encoding genes using PCR. The tripartite hemolysin
BL (Hbl) has been well established as a diarrhoeal enterotoxin among the enterotoxins, or enterotoxin candidates (Beecher et al., 1995; Schoeni and Wong, 2005). The single component protein, BceT, was identified in early studies as an enterotoxin (Agata et al.,
1995a) and its enterotoxic activity has been demonstrated with several detection methods (Jenson and Moir, 2003) and targeted by PCR assay.
The eight isolates of B. cereus and B. thuringiensis that did not harbour the genes for
Hbl and BceT, were further analysed by PCR for detection of the nheA gene, as the
Tecra enterotoxin ELISA kit detects primarily NheA (45-kDa), a component of the Nhe complex diarrhoeal toxin (Granum and Lund, 1997). Of those eight isolates, six yielded a nheA amplicon, though B. cereus MR 13-2 and B. thuringiensis MR 10-4 did not. B. thuringiensis (MR 10-4) gave a negative result in the ELISA, while B. cereus (MR 13-
2) was positive. There is a possibility that the enterotoxin immunoassay (Tecra) may not only detect NheA protein, as part of the non-haemolytic enterotoxin (Nhe). It has been considered that of the three component proteins of Nhe, the 39-kDa, 45-kDa and 105- kDa subunits, the Tecra immunoassay detects mainly the 45-kDa protein but also the
105 kDa component (Beattie and Williams, 2000; written before 1999). However,
Granum et al. (1999) found that the three proteins comprising Nhe are in fact NheA (45- kDa, more correct size is 41-kDa), NheB (39-kDa, more correct size is 39.8-kDa) and
NheC (36-kDa) (Granum, 2002; Jenson and Moir, 2003). Therefore, though the 105- kDa protein in Nhe is not involved directly with enterotoxin activity and is now identified as an apparent collagenase (Lund and Granum, 1999), this protein may react with the Tecra enterotoxin immunoassay. Interestingly, B. thuringiensis MR 14-6
151 yielded a nheA amplicon but was negative with the enterotoxin ELISA. The production of a fragment of a gene using PCR, while suggesting the ability to produce toxin, may not necessarily indicate toxin production will occur. According to Prüß et al. (1999), the enterotoxin gene hblA was found to be distributed widely amongst members in the B. cereus group. However, in certain strains of B. thuringiensis and B. mycoides, the gene was either incomplete or silent, as these strains did not demonstrate cytotoxic effects, a charactersistic induced by the enterotoxin gene product.
Analogous results were obtained when analysis for the emetic toxin was carried out.
MR 22-1 and MC 27-3 showed positive results with the cereulide PCR, but only MC
27-3 gave a positive result with the emetic ELISA. However, in this case, the PCR results for the isolates MR 22-1 and MC 27-3 were supported by MEKC analysis, as these isolates produced the greatest amount of putative cereulide amongst the B. cereus isolates tested using MEKC. With respect to analysis for the emetic toxin of B. cereus isolates, the results of the PCR and ELISA tests were not always in agreement and, in most cases, MEKC analysis showed agreement with the PCR.
Many of the B. cereus isolates produced both emetic and diarrhoeal enterotoxin(s), as shown by several detection methods (Table 5.8). Even though most studies with B. cereus isolates associated with food poisoning outbreaks classified them distinctly as either emetic or diarrhoeal types (Nishikawa et al., 1996), some exceptions have been reported, with some food poisoning incidents involving both vomiting and diarrhoea, possibly due to the production of both types of toxins (Granum and Lund, 1997). The fact that a single B. cereus strain could produce both type of toxins, confirmed with testing the reference emetic B. cereus strain NC Y and NC 7401, was confirmed with
152 field isolates in the current study. The two reference strains (B. cereus NC Y and NC
7401)Gproduced a large amount of emetic toxin, cereulide, confirmed by NMR analysis, and also showed a strong positive reaction with the enterotoxin ELISA and PCR for detection of the nheA gene. A similar finding was reported by Hansen and Hendriksen
(2001) with the reference B. cereus emetic strain, F 4810/72, which yielded an amplicon indicative of the presence of the gene encoding the Nhe complex, and also produced enterotoxin, as determined with an enterotoxin ELISA kit (Tecra BDE-VIA). Likewise,
Toh et al. (2004) reported the reference diarrhoeal B. cereus strain, N1230-88 to be positive in a PCR for a gene associated with cereulide production. Collectively, some B. cereus strains are indeed capable of producing both types of toxin and many B. cereus isolates from rice products were shown to be co-producers. A possible reason that cooked rice is associated mainly with the emetic type of food poisoning is, at least in part, that cereulide production is encouraged in a farinaceous food.G
B. thuringiensis MR 10-4 did not produce diarrhoeal toxin and also did not carry the genes (hblA, bceT and nheA) encoding the diarrhoeal toxins, but this isolate gave a positive result with the cereulide ELISA. Though B. thuringiensis has been used widely as an insecticide, concerns have been raised over its safety as strains are now well known to produce diarrhoeal toxin (Granum, 2002). Therefore, it should be confirmed that strains used as a bioinsecticide do not carry the enterotoxin genes and, more importantly, cannot produce the diarrhoeal toxin. However, some B. thuringiensis isolates in the present study were shown to produce the emetic toxin or indeed both diarrhoeal and emetic toxins, though MEKC analysis showed production of much smaller amounts of toxin by B. thuringiensis and diarrhoeal B. cereus strains than by emetic B. cereus strains (Section 4.3.2.2, Chapter 4). Also, some isolates of B.
153 thuringiensis (MR 6-5 and MC 22-9) were carrying the gene fragment specifically associated with the emetic toxin. Therefore, production of both diarrhoeal and emetic toxins by B. thuringiensis should be considered.
All B. mycoides isolates from raw rice showed positive results for diarrhoeal enterotoxin
(Table 5.8). Similar result was reported by other researcher (Prüß et al., 1999), not surprising given the similarity of genotype and phenotype between this species and B. cereus and B. thuringiensis. However, no food poisoning incidents involving this species have been reported. Interestingly, no B. mycoides isolates from cooked rice produced enterotoxins or cereulide, which may explain the lack of association with food poisoning, at least the emetic syndrome.
5.3.2.2 B. subtilis group
None of the isolates from the B. subtilis group (B. subtilis/B. mojavensis, B. licheniformis and B. pumilus) were enterotoxigenic, based on PCR or ELISA analysis
(Table 5.9). However, most of the isolates identified as B. subtilis/B. mojavensis and B. pumilus showed evidence of cereulide production. None of the isolates yielded a positive result with the cereulide-specific PCR but, interestingly, many gave positive results with the cereulide ELISA and MEKC analysis (Table 5.9).
154 Table 5.8 Toxigenicity of strains in B. cereus group isolated from rice Enterotoxin Emetic toxin RAPD patterns a,b Strain PCR CE with OPR 13 ELISAc PCR ELISAc hblA bacT nheA (Pg/ml) B. cereus MR 6-1 17 - - + +++ + ++ MR 7-4 6 - + ++ + ++ MR 8-2 7 - + ++ - - <2 MR 9-1 7 - + +++ - - MR 10-1 6 - + ++ + - 71 MR 13-2 19 - - - +++ - - MR 13-6 4 + - ++ - + <2 MR 14-2 12 - + + + ++ 80 MR 15-3 14 - + ++ + - 75 MR 19-1 13 + + ++ + ++ 97 MR 22-1 10 - + ++ + - 137 MR 24-1 15 - + ++ + - 88 MR 30-1 16 - + ++ - - MR 31-1 1 - + ++ + - MR E-7 3 - + ++ + - MC cr2-2-1 7 - + + + +++ 53 a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
155 Table 5.8 Toxigenicity of strains in B. cereus group isolated from rice (continued) Enterotoxina,b Emetic toxinc RAPD patterns Strain PCR CE with OPR 13 ELISA PCR ELISA hblA bacT nheA (Pg/ml) B. cereus (continued) MC 6-1 3 - + ++ + - MC 15-4 5 - + ++ + - MC 21-2 10 - + ++ + + MC 24-2 9 + + ++ + - MC 27-3 8 - - + +++ + + 172 MC Ivan-2 2 - + + - - MC Lab-1 18 - + ++ + - 44 MC Beef-1 11 - - + ++ - - MC Beef-3 11 - + ++ + - MC Ram-1 18 - + ++ + + B. thuringiensis MR 6-5 25 - + ++ + - 55 MR 10-4 24 - - - - - + MR 14-6 23 - - + - - - MR 14-11 26 - - + +++ - - MR 31-6 26 - + +++ - - a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
156 Table 5.8 Toxigenicity of strains in B. cereus group isolated from rice (continued) Enterotoxin a,b Emetic toxinc RAPD patterns Strain PCR CE with OPR 13 ELISA PCR ELISA hblA bacT nheA (Pg/ml) B. t (continued) MC cr1-6 27 - + +++ - - MC p4-8 28 - + + - - <2 MC 20-2 22 - + ++ - - MC 21-3 21 + + ++ - - MC 21-8 21 + + ++ - - MC 22-9 20 - - + +++ + + B. mycoides MR 14-1 29 - + + - - MR 20-6 32 - + + - - MR 21-1 30 - + + - - MR 21-1 31 - - - - - <2 a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
157 Other researchers also found that the results of PCR to determine toxic potential do not always correlate with toxin production when determining toxic potential and subsequent toxin expression of enterotoxic B. cereus (Mäntynen and Lindström, 1998; Phelps and
McKillip, 2002; Toh et al. 2004). According to Toh et al. (2004), who developed the cereulide PCR used in the current study, PCR-based assays designed to detect toxin genes may not always produce a positive signal when in fact the toxin is synthesised.
Diarrhoeal strains tested in that study were shown to be cereulide PCR-negative but
MTT-positive, suggesting production of cereulide. This may result from a mutation in the binding site of one of the primers (Prüß et al., 1999) or it may indeed reflect the absence of the responsible gene, in this case a non-ribosomal peptide synthetase (NRPS) gene (Toh et al, 2004). They hypothesised that cereulide is produced non-ribosomally by a large, multi-domain, enzyme complex and, based on supporting literature, the primers for detection of genes responsible for cereulide production were developed by targeting non-ribosomal peptide synthetases (NRPS). Originally, it was thought that the target sequence for the cereulide-PCR is unique to the B. cereus group. Although some strains of other Bacillus species, including B. circulans (S33TICC 2715) and B. megaterium (S41TICC 16) (in Table 4.1, Chapter 4), showed positive results from the cereulide PCR (performed by Toh), this could be explained by similarities between
NRPS genes in other Bacillus species (Toh et al., 2004) and those genes in B. circulans and B. megaterium may have greater similarity to the NRPS genes in B. cereus than those in B. subtilis and B. pumilus, which were shown to be emetic PCR-negative in this study.
A reference B. subtilis strain (S49TICC 15, in Table 4.1, Chapter 4) was shown to produce heat-stable emetic toxin using the MTT assay (performed by Toh) and also
158 shown to produce putative cereulide using MEKC analysis (Chapter 4), but this strain did not appear to harbour the gene associated with synthesis. This evidence supports the results from the present study, that some isolates of B. subtilis/B. mojavensis and B. pumilus could produce cereulide while being PCR-negative. Even though there is a possibility of a non-specific reaction in the MTT assay, heat-treated (121ºC for 15 min) supernatants should contain only highly heat-resistant toxin, most likely cereulide.
Furthermore, in this present study, a prototype immunoassay was used instead of the
MTT assay, the former considered to be highly specific for cereulide. There is also strong evidence that other Bacillus species can indeed produce cereulide, based on
HPLC-MS analysis with reference Bacillus species.
Fletcher and Logan (1999) found B. mycoides and B. thuringiensis strains tested positive for ‘B. cereus enterotoxin’ production with both immunoassay kits (BCET-
RPLA and BDE-VIA), suggesting that some of their extracellular metabolites are very similar to those of B. cereus. Some of the mismatched results between the MEKC and
ELISA method in the present study, could be explained in this way. For example, isolate
MR 16-2 was ELISA-negative yet MEKC-positive, suggesting production of a similar rather than an identical compound if the ELISA is truly specific for cereulide. While
MEKC analysis was shown to correlate well with HPLC-MS analysis (Toh et al. submitted), the possibility exists that extremely similar molecules may be produced by some bacilli. For example, the antibiotic valinomycin is very similar chemically and structurally to cereulide and exhibits similar though lower biological activity. Also, a recent study by From (2005) reported that a B. mojavensis isolate could produce a heat- stable toxin and their results suggest that this toxin is different from cereulide. Thus, while the substances produced by these other bacilli may indeed be toxin(s), similar but
159 not identical to cereulide, the specificity of the cereulide ELISA should still be investigated.
Some of the isolates in the B. subtilis group produced putative emetic toxin at levels similar to the levels of cereulide produced by emetic isolates of B. cereus, as shown by
MEKC analysis, and similar results were found with reference strains of B. subtilis and
B. pumilus (Section 4.3.2.2 in Chapter 4). These species, and in particular B. subtilis and
B. pumilus, have been associated with outbreaks of the emetic type of food poisoning, and have been shown to produce toxin(s) (Lund, 1990; Shinagawa 1990; Dahl, 2000).
These results highlight the importance of paying attention to these species with respect to food safety and public health, although they have, to date, received little attention when compared to the B. cereus group.
In the case of B. licheniformis, only some isolates from raw rice yielded a peak when analysed using MEKC and these results fit with the results of MEKC analysis for the reference strains of B. licheniformis (Chapter 4). In the current study, most B. licheniformis isolates were not potentially or actually toxigenic (Table 5.9), even though this bacterium has been associated with food poisoning incidents as well as spoilage
(ropy bread) (Salkinoja-Salonen et al., 1999; Sorokulova et al., 2003).
5.3.2.3 Other Bacillus species and Paenibacillus species
Bacillus species, other than those within the B. cereus and the B. subtilis groups, as well as isolates identified as Paenibacillus, were also tested for toxin production (Table 5.9).
None of the B. megaterium or Paenibacillus isolates from raw or cooked rice produced any toxins. Only one B. sphaericus/B. fusiformis isolate MR 23-4 gave a positive result
160 with the cereulide ELISA. B. fusiformis isolate MR 23-6 showed putative cereulide production with MEKC analysis though the cereulide ELISA was negative. A similar finding was shown with reference B. sphaericus strains in Chapter 4. Two reference strains of B. sphaericus putatively produced cereulide at concentrations of 62 µg/ml and
9 µg/ml respectively, determined by MEKC analysis (Section 4.3.2.2, Chapter 4).
B. sphaericus is, along with B. thuringiensis, one of the main entomopathogenic bacteria and is used widely for mosquito control. This bacterium synthesises parasporal crystal proteins similar to those of B. thuringiensis. B. sphaericus has been proven to have no toxic effect against blackflies and littels or no effect on non-target insects, and no effect on mammals (Priest, 2002). Yuan et al. (2002) reported that none of the B. sphaericus isolates tested carried the genes encoding the diarrhoeal toxins. However, B. sphaericus and B. fusiformis may need to be considered with respect to emetic toxin production. Not only did the reference strain produce putative cereulide, at a concentration of 62 µg/ml, but showed toxicity in the MTT assay toward CHO-K1 cells.
From et al. (2005) also reported that isolates shown to be able to produce emetic toxin were identified as B. fusiformis. Therefore, care will be needed to deal with these two species as well as the other (potentially) toxigenic species.
161 Table 5.9 Toxigenicity of Bacillus species isolated from rice Strains RAPD patterns Enterotoxin a,b Emetic toxinc with OPR 13 PCR ELISA PCR ELISA CE B. subtilis /B.mojavensis MR 5-7 12 - - - ++ MR 6-3 13 - - - + MR 7-2 13 - - - ++ MR 8-4 10 - - - - MR 8-5 19 - - - ++ MR 9-3 11 - - - +++ 122 MR 10-2 24 - - - +++ 128 MR 10-3 24 - - - ++ MR 10-6 8 - - - ++ MR 13-1 17 - - - - MR 13-9 25 - - - ++ MR 16-2 1 - - - - 71 MR 16-5 2 - - - + 146 MR 16-6 2 - - - - MR 17-1 20 - - - ++ MR 17-6 26 - - - + MR 18-1 31 - - - + MR 19-3 32 - - - - MR 20-1 22 - - - ++ MR 22-2 25 - - - - 91 MR 22-3 28 - - - - MR 24-2 6 - - - ++ MR 27-2 23 - - - ++ MR 28-2 23 - - - ++ MR 30-4 8 - - - + MR 30-5 4 - - - + MR 31-4 3 - - - - MR 31-7 23 - - - + MR E-2 21 - - - + MR E-3 21 ---+ a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
162 Table 5.9 Toxigenicity of Bacillus species isolated from rice (continued) Strains RAPD patterns Enterotoxina,b Emetic toxinc with OPR 13 PCR ELISA PCR ELISA CE B.s (continued) MR T-1 27 - - - ++ MR T-5 9 - - - + MR r2-1a 27 - - - - MR cr1-1 14 - - - ++ MR cr1-2 7 - - - + MR cr2-9 14 - - - ++ MR cr2-10 16 - - - ++ MR cr3-7 4 - - - + MC 4-2 25 - - - + MC 4-3 18 - - - + MC 10-1 20 - - - + MC 15-2 12 - - - + MC 22-3 12 - - - +++ MC 22-7 12 - - - ++ MC 25-3 30 - - - ++ MC 27-1 29 - - - +++ 80 MC Ivan-3 5 - - - ++ MC Beef 4-1 12 - - - + MC Pork-1 12 - - - + MC Chicken-2 12 - - - + MC Chicken-3 12 - - - ++ MC Chicken-4 12 - - - ++ MC Pa-1 15 - - - + B. licheniformis MR 7-15 3 - - - - MR 17-3 5 - - - + MR 18-3 1 - - - - MR 19-5 6 - - - + MR 23-8 12 - - - - MR 24-3 12 - - - - a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
163 Table 5.9 Toxigenicity of Bacillus species isolated from rice (continued) Strains RAPD patterns Enterotoxina,b Emetic toxinc with OPR 13 PCR ELISA PCR ELISA CE B. l (continued) MR 25-3 9 - - - - MR 28-3 11 - - - - MR T-3 7 - - - + 84 MR E-8 8 - - - - MC p4-5 10 - - - - MC p5-5 10 - - - - MC 23-3 2 - - - - MC 25-2 2 - - - - MC Lab-2 2 - - - - MC Beef-8 3 - - - - MC Beef-9 4 - - - - B. pumilis MR 7-6 11 - - - ++ 55 MR 7-10 8 - - - ++ MR 8-1 6 - - - ++ 113 MR 8-14 6 - - - ++ MR 10-7 5 - - - + MR 14-14 3 - - - - MR 19-11 4 - - - ++ MR 20-3 7 - - - ++ MR 21-3 3 - - - ++ 119 MR Cr2-7 9 - - - + MR R4-2p 1 - - - ++ MC 20-4 10 - - - + MC 22-6 10 - - - ++ 162 MC Chicken-7 2 - - - +++ B. megaterium MR 5-1 8 - - - - <2 MR 14-13 1 - - - - MR 15-8 3 - - - - a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
164 Table 5.9 Toxigenicity of Bacillus species isolated from rice (continued) Strains RAPD patterns Enterotoxina,b Emetic toxinc with OPR 13 PCR ELISA PCR ELISA CE B. me(continued) MR 17-8 4 - - - - MR 22-9 2 - - - - MR r1-5a 3 - - - - MR r2-2b 1 - - - - MR cr3-9 5 - - - - MC 27-12 7 - - - - MC Beef-10 6 - - - - B. sphaericus /B.fusiformis MR 23-4 1 - - - + MR 23-6 2 - - - - 37 MC cr2-1-2 3 - - - - Penibacillus MR 7-16 1 - - - - MR 25-5 2 - - - - <2
a b c +, detected; -, not detected; -, OD405 0.2; +, 0.2
G G Table 5.10 provides a summary of the toxigenic Bacillus isolates found in rice. Of the
26 B. cereus isolates tested, all produced enterotoxin and 96.2% of the isolates also
showed positive results with the PCR for detection of the diarrhoeal genes, while 73.1%
of the B. cereus isolates had positive results with the cereulide-specific PCR. In the case
of B. thuringiensis, a much lower proportion of isolates showed positive results with the
cereulide-specific PCR (18.2%) and the emetic ELISA (18.2%). Of the 53 B. subtilis
isolates and the 14 B. pumilus isolates, 83.0% and 92.9%, respectively, gave positive
results with the cereulide ELISA. None of the B. megaterium strains from rice samples
were toxigenic.
165 Table 5.10 Populations of potential and actual toxigenic Bacillus species Identified strains Number of Enterotoxin Emetic toxin tested PCR ELISA PCR ELISA isolates B. cereus 26 25 26 19 9 B. thuringiensis 11 10 9 2 2 B. mycoides 4 3 3 00 B. subtilis/B. mojavensis 53 0 0 044 B. licheniformis 17 0 0 03 B. pumilis 14 0 0 013 B. megaterium 10 0 0 00 Other bacilli 5 0 0 01 Total 140 38 38 21 72 G
G
5.4 ConclusionsG
In this chapter, RAPD was successfully applied to the differentiation of isolates from raw and cooked rice and thus it could be a useful typing technique to trace contamination sources with rapid and easy performance. The survey of rice food products described in Chapter 3 yielded 351 Bacillus isolates that were subjected to
RAPD analysis. B. subtilis/B. mojavensis and B. cereus isolates, the dominant genospecies in both raw and cooked rice, showed the greatest strain diversity among all of the isolates of bacilli. In general, different RAPD types were observed among the isolates from raw and cooked rice samples for all Bacillus species, with greater diversity in raw rice.
Toxin production and potential toxigenicity of the isolates were also investigated in this chapter. Generally, most B. cereus and B. thuringiensis isolated from rice products
166 harbour the genes responsible for the production of diarrhoeal toxins and they also produced toxin. As mentioned previously, interesting results were obtained when the preliminary tests of toxin production were performed with reference cultures of B. cereus. Some isolates of B. cereus denoted as diarrhoeal strains showed production of emetic toxin using MEKC analysis (Section 4.3.3.2.2, Chapter 4) and ELISA, while the emetic B. cereus strains NC Y and NC 7401 showed strong positive results using the enterotoxin ELISA, indicating at least some strains are capable of producing both toxins.
In the cases of B. subtilis and B. pumilus, most isolates affiliated with these species showed the production of a putative emetic toxin, most likely cereulide. Finally, it was noted that isolates with the same RAPD pattern were shown to have a similar profile of enterotoxigenicity but less correlation was observed in relation to emetic toxin.
In summary, this chapter showed that RAPD analysis could be used successfully to differentiate Bacillus isolates from rice products and reveal information about their strain diversity and ecology. Genotypic and phenotypic analysis showed that most of the isolates of the B. cereus group from rice products were toxigenic and some isolates were co-producers of emetic and diarrhoeal toxins. The other Bacillus species outside the B. cereus group were also capable of producing the emetic toxin, cereulide.
167 G G G G Chapter 6
Conclusions and recommendations 6.1 Conclusions
Members of the genus Bacillus encompass a great diversity of bacterial species and are widespread in environment due to their ubiquitous nature. Due to the resistance of spores to stressed conditions and the capability of vegetative cells to secrete a variety of enzymes or toxins, this genus is important not only in the production of commercially important biomolecules such as insecticides, antibiotics and enzymes, and some foods, but also as a source of spoilage or pathogenic organisms transmitted through foods and beverages (Kramer and Gilbert, 1992).
A well-known member of the genus,GBacillus cereus is a food poisoning bacterium causing both emetic and diarrhoeal disease. Fatality as a result of intoxication from emetic-type B. cereus food poisoning has been recorded (Mahler et al., 1997) and severe forms of the diarrhoeal syndrome have been reported, including a necrotic enteritis causing three deaths (Lund et al., 2000). Recently, other Bacillus speciesSG particularly B. subtilis, B. licheniformis, B. brevis, B. pumilus and B. thuringensis, have also been recognised as food poisoning bacteria.
This study has involved a ecological investigation of Bacillus species in rice products, a well known source of B. cereus, using a combination of cultural and molecular methods.
In addition to this, toxigenic properties of Bacillus isolates were evaluated with rapid molecular methods and immunoassays, and a novel chromatographic method, MEKC.
Initially, a novel centrifugation-plating method was developed for the investigation of biodiversity of Bacillus species in rice products. Development of the centrifugation-
G X]` plating method required hydrolysis and solubilisation of the food matrix, particularly the starch in cooked rice products, using a combination of amylase and Tween 80 treatment. Two commercially available Ȼ-amylases, originally derived from Bacillus, were evaluated for their suitability and it was found that amylase A and Tween 80 treatment did not exert a significant effect on viability of reference Bacillus species including B. cereus, B. thuringiensis, B. licheniformis and B. subtilis (P<0.05). The centrifugation-plating method, combined with Tween 80 treatment or treatment with amylase and Tween 80, was also evaluated for its recovery and enumeration of known populations of B. cereus inoculated onto or into rice products. Generally, 86 – 100% of the inoculated cells could be recovered.
There were many occasions where the centrifugation-plating method detected the presence of bacilli in foods, while the organisms remained undetected by the conventional spread plating method, as the former could detect much lower numbers of
Bacillus than the latter. Of the 35 raw rice samples, 33 samples were found to contain
Bacillus species with the centrifugation-plating method whereas only 14 samples were found to contain Bacillus species with the spread plating method. Similar results were observed for cooked rice samples. Of the 35 cooked rice sample, 25 samples were found to be contaminated with Bacillus species with the centrifugation-plating and only 12 samples with the spread plating.
Colonies of Bacillus on PEMBA agar plates, obtained using the centrifugation-plating and spread plating methods, were identified and confirmed as Bacillus species by partial rDNA sequencing, and API 50CH and API 20E systems. A greater diversity of bacilli was also found with the centrifugation-plating method when compared to the
G X^W spread plating method. With the centrifugation plating method, seven different Bacillus species including B. cereus/B. thuringiensis, B. mycoides, B. subtilis/B. mojavensis, B. licheniformis, B. pumilus, B. sphaericus/B. fusiformis and B. megaterium, as well as
Paenibacillus species, were found in raw rice and six Bacillus species were found in cooked rice samples, including B. cereus/B. thuringiensis, B. mycoides, B. subtilis/B. mojavensis, B. licheniformis, B. pumilus and B. megaterium. Up to five different
Bacillus species were found in some samples (for example, raw rice samples 3 and 34 in
Chapter 4). B. sphaericus/B. fusiformis and Paenibacillus were found only with the centrifugation-plating method. Of all the bacilli found, toxigenic or potentially toxigenic
B. cereus/B. thuringiensis, B. subtilis, B. licheniformis and B. pumilus were found most frequently, in both raw and cooked rice.
Fried rice has been shown to be most frequently associated with the emetic form of food poisoning. In this study, comparatively more contamination with Bacillus cells was found in the fried rice samples than in the other cooked rice samples. Two out of seven fried rice samples were contaminated with 2 – 3 × 103 cfu of mixed Bacillius species per gram of sample. Even this seemingly moderate population cannot be considered entirely safe for consumption as a previous study reported that incidents of emetic B. cereus food poisoning have been associated with populations that ranged from 103 to
109 cfu per g or ml (Lund, 1990) and both B. subtilis and B. cereus have been involved with emetic food poisoning. Proper heat treatments and storage conditions are critical to maintain the quality and safety of the product because reheating may not be sufficient to kill spores, which may cause infection, leading to diarrhoeal disease, or the generation of emetic toxin, cereulide, which is particularly heat-stable. The outbreak that occurred in Melbourne, Australia in 2002 (Section 2.7.1), caused by Bacillus cereus (isolated
G X^X from rice), with rapid onset of emesis also highlighted the importance of proper food preparation and temperature control, as 272 people became ill.
Epidemiological investigation of outbreaks of foodborne disease caused by Bacillus species has revealed that different genotypes and strains are responsible. This highlights the importance of strain diversity within species and the development of molecular methods as useful tools, not only for epidemiological studies, but also for determining and monitoring the diversity in food ecosystem.
G In this study, the diversity among the bacilli (Bacillus species and Paenibacillus species) isolated from rice products was also investigated at the subspecies or strain level with the rapid and effective molecular typing method, RAPD-PCR. Although many studies have examined the genotypic diversity among members of the B. cereus group, the ecological significance of the genotypic and phenotypic diversity of bacilli still remains largely unexplored (Garbeva et al., 2003). RAPD-PCR was found to be a useful method for differentiating the isolates of other bacilli. The results of RAPD-PCR revealed broad diversity in the strains and revealed some associations of isolates from the raw and the cooked rice, at the genotypic level.
A comparatively greater diversity of strains was observed among isolates from raw rice than those from cooked rice, and generally different RAPD profiles of Bacillus isolates were found in the raw (uncooked) and the cooked rice products with few common to both types of rice. These results indicate that an originally diverse set of strains exists in raw rice products, with some strains, presumably because their spores are more heat- tolerant than others, surviving the cooking processes. The other RAPD types found in
G X^Y cooked rice, which were few in general among all species, may have derived from environmental sources, such as cooking equipment, packaging, or other ingredients used in mixed products, all possible due to the ubiquity of bacilli in nature. Similar results were found by Gilbert and Parry (1977) when they serotyped B. cereus from raw and cooked rice samples. Multiple serotypes of B. cereus strains were found in raw rice including serotype H1, but most isolates from cooked rice were serotype H1. It would be of interest to serotype the B. cereus isolates from the present study.
In general, a high level of diversity was observed among isolates of Bacillus species.
However, in some cases, the same RAPD patterns were found among isolates from different food samples. As an example, B. cereus isolates from the cooked rice (MC
Lab-1) and the cooked rice mixed with beef (MC Beef-1) showed the same RAPD pattern (Chapter 6). Similar results were found for other Bacillus species including B. thuringiensis, B. subtilis/B. mojavensis and B. licheniformis. On the other hand, some isolates from a single species yielded different RAPD patterns even though they were isolated from one sample. This showed that one food sample could be contaminated with a diversity of strains. Collectively, these results highlight not only the diversity among the bacilli, but the discriminatory power of the RAPD PCR technique, which may serve as a useful epidemiological tool.
The potential and actual toxin production of bacilli isolated from rice products was investigated using molecular, immunological and chromatographic methods. This study also generated additional information that may be useful in epidemiological studies through determination of the incidence of toxigenic strains of different bacilli in rice products.
G X^Z Initially, in relation to the study of toxin production, a MEKC capillary electrophoresis method was developed for quantitative analysis of cereulide (Chapter 5). MEKC gives results quickly and simply, and holds advantages over HPLC, through simple sample handling and preparation, less sample required and lower cost. After calibration with purified cereulide was obtained, the applicability of the method to model and real food samples was evaluated with sixty reference Bacillus strains and rice samples inoculated with an emetic B. cereus strain (NC 7401). The results showed that the method could be successfully applied to both semi-purified toxin extracts of cultures and food samples.
This method was also used for analysis of samples associated with an outbreak of B. cereus emetic food poisoning in Melbourne, Austrailia. Additionally, screening results using MEKC analysis revealed that other Bacillus species could produce putative cereulide. Amongst those bacilli shown to produce putative cereulide, B. thuringiensis
(FDA 4), B. megaterium (TICC 16) and B. circulans (TICC 2715) were confirmed to produce cereulide after HPLC-MS analysis.
The several PCR analysis and toxin detection methods, including the MEKC method developed for cereulide, were used to examine the toxigenicity of Bacillus isolated from rice products. This included use of a recently developed PCR method for detection of a gene associated with cereulide production, as well as a prototype ELISA and the MEKC method. Study of the isolated bacilli revealed that most in the B. cereus group harboured one or more diarrhoeal enterotoxin gene(s) or a gene associated with cereulide production. These isolates were subsequently shown to express these genes and produce toxins, using ELISA and MEKC analysis. Some of the isolates in this group are co- producers of both diarrhoeal and emetic toxins. Results of analysis for Bacillus isolates using cereulide PCR and ELISA provided further evidence that Bacillus species other
G X^[ than B. cereus can produce the emetic toxin, cereulide. In addition, generally, isolates that have the same RAPD pattern were shown to have the similar profile of enterotoxigenicity, particularly in the B. cereus group.
6.2 recommendations
This study aimed to systematically examine the microbial ecology of Bacillus species in rice products. Under this aim, the study focussed on both the toxic potential and the diversity of Bacillus species associated with rice products. During these investigations, a novel centrifugation-plating method as a sensitive detection method for bacilli, and a novel capillary electrophoresis method for detection of cereulide were developed and applied. For analysis of strain diversity, a RAPD-PCR method was utilised, and some
RAPD profiles appear to represent more heat-resistant Bacillus strains. Earlier studies revealed that specific serotype (H1) may represent the more heat-tolerable B. cereus strains. Therefore, if the serotyping method could be applied to the Bacillus isolates from this study, more concrete results regarding the association bewteen heat-tolerant strains and specific RAPD types will be obtained, based on the results of two different typing methods. The results of the study of toxigenicity among Bacillus isolates demonstrated that many, particularly B. cereus isolates, from rice products were at least potentially toxigenic.
The putative cereulide produced by B. subtilis/B. mojavensis and B. pumilus isolates needs to be further analysed by HPLC-MS and NMR as some of these strains produced this putative toxin at levels similar to that produced by emetic B. cereus as revealed by
MEKC and cereulide ELISA analysis. If the putative toxin produced by these species
G X^\ was confirmed to be cereulide (or a toxin similar in nature to it), these two species may need to be considered more seriously as foodborne pathogens, alongside B. cereus. This is especially true, given that selected strains of Bacillus species, including diarrhoeal B. cereus, B. circulans, B. megaterium and B. thuringiensis, have already been confirmed to produce cereulide in a related study (Toh et al., submitted). The current study has shown that isolates of B. subtilis/B. mojavensis and B. pumilus may also produce cereulide. Further, there is a need to differentiate B. subtilis and B. mojavensis as with B. cereus and B. thuringiensis, in order to generate more specific information about the relationship between each of the two species and toxin production. While starch hydrolysis is not a perfect marker, it has proved useful in the differentiation of cereulide-producing and non-producing strains of B. cereus. It would be interesting to apply this simple biotyping method to isolates from the current study not only of B. cereus, but other bacilli, to further examine the correlation between starch hydrolysis and the production of cereulide or other closely related toxins.
In relation to the MEKC method, it should be used to further investigate the relationship between growth and toxin production in various systems, including culture media, but especially rice products. This is true not only for cereulide and B. cereus, but other, closely related toxins produced by other species. It is also recommended for future studies that a variety of samples needs to be tested before the method is used in routine analysis; this could not be achieved in this study due to a limited time frame.
G
G X^] Chapter 7
Bibliography G
Adelaida, M., Rivera, G. and Priest, F.G. (2003) Molecular typing of Bacillus thuringiensis serovars by RAPD-PCR. Systematic Applied Microbiology 26:254-261.
Agata, N., Mori, M., Ohta, M., Suwan, S., Ohtani, I. and Isobe, M. (1994) A novel dodecadepsipeptide, cereulide, isolated from Bacillus cereus causes vacuole formation in HEp-2 cells. FEMS Microbiology Letters 121:31-34.
Agata, N., Ohta, M. and Mori, M. (1996) Production of an emetic toxin, cereulide, is associated with a specific class of Bacillus cereus. Current Microbiology. 33:67-69.
Agata, N., Ohta, M., Arakawa, Y. and Mori, M. (1995a) The bceT gene of Bacillus cereus encodes an enterotoxic protein. Microbiology 141:983-988.
Agata, N., Ohta, M., Mori, M. and Isobe, M. (1995b) A novel dodecadepsipeptide, cereulide, is an emetic toxin of Bacillus cereus. FEMS Microbiology Letters 129:17-20.
Ahmed, A.A.H., Moustafa, M.K. and Marth, E.H. (1983) Incidence of Bacillus cereus in milk and some milk products. Journal of Food Protection 46:126-128.
Akhurst, R.J., Lyness, E.W., Zhang, O.Y., Cooper, D.J. and Pinnock, D.E. (1997) A 16S rRNA gene oligonucleotide probe for identification of Bacillus isolates from sheep fleece. Journal of Invertebrate Pathology 69:24-30.
Alippi, A.M. and Aguilar, O.M. (1998) Characterization of isolates of Paenibacillus larvae ssp. larvae from diverse geographical origin by the polymerase chain reaction and BOX primers. Journal of Invertebrate Pathology 72:21-27.
Altria, K.D. (1996) Fundamental of capillary electrophoresis theory. In: Capillary Electrophoresis Guidebook (eds. K.D. Altria). pp. 3-13. Humana, New Jersey.
Andersson, A., Rönner, U. and Granum, P.E. (1995) What problems does the food industry have with the spore-forming pathogens Bacillus cereus and Clostridium perfringens? International Journal of Food Microbiology 28:145-155.
G X^_ G
Andersson, A., Svensson, B., Christiansson, A. and Ronner, U. (1999) Comparison between automatic ribotyping and random amplified polymorphic DNA analysis of Bacillus cereus isolates from the dairy industry. International Journal of Food Microbiology 47:147-151.
Andersson, M.A., Mikkola, R., Helin, J., Andersson, M.C. and Salkinoja-Salonen, M. (1998) A novel sensitive bioassay for detection of Bacillus cereus emetic toxin and related depsipeptide ionophores. Applied and Environmental Microbiology 64:1338- 1343.
Ankarloo, J., Caugant, D.A., Hansen, B.M., Berg, A., Kolstø, A. and Lövgren, A. (2000) Genome stability of Bacillus thuringiensis subsp. israelensis isolates. Current Microbiology 40:51-56.
Anon (2002) Food Standards News, Australia. The newsletter of the Australia New Zealand Food Authority (ANZFA) 36:2
Arbeit, R.D. (1995) Laboratory procedures for the epidemiologic analysis of microorganisms. In: Manual of clinical microbiology (eds. P.R. Murray, E.J. Baron, M.A. Pfaller, F.C. Tenover and R.H. Yolken). pp. 190-204. ASM Press, Washington, D.C.
Arment, A.R. and Carmichael, W.W (1996) Evidence that microcystin is a thio-template product. Journal of Physiology 32:591-597.
Artiushin, S. and Minion, F.C. (1996) Arbitrarily primed PCR analysis of Mycoplasma hyopneumoniae field isolates demonstrates genetic heterogeneity. International Journal of Systematic Bacteriology 46(1):324-328.
Asano, S.I., Nukumizu, Y., Bando, H., Iizuka, T. and Yamamoto, T. (1997) Cloning of novel enterotoxin genes from Bacillus cereus and Bacillus thuringiensis. Applied and Environmental Microbiology 63(3):1054-1057.
G X^` G
Ash, C., Farrow, J.A.E., Dorsch, M., Stackebrandt, E. and Collins, M.D. (1991a) Comparative analysis of Bacillus anthracis, Bacillus cereus and related species on the basis of reverse transcriptase sequencing of 16S rRNA. International Journal of Systematic Bacteriology 41(3):343-346.
Ash, C., Farrow, J.A.E., Wallbanks, S. and Collins, M.D. (1991b) Phylogenetic heterogeneity of the genus Bacillus revealed by comparative analysis of small subunit- ribosomal RNA sequences. Letters in Applied Microbiology 13:202-206.
Ash, C., Priest, F.G. and Collins, M.D. (1993) Molecular identification of rRNA group 3 bacilli using a PCR probe test. Antonie van Leeuwenhoek 64:253-260.
Aucamp, J.P., Hara, Y. and Apostolides, Z. (2000) Simultaneous analysis of tea catechins, caffeine, gallic acid, theanine and ascorbic acid by micellar electrokinetic capillary chromatography. Journal of Chromatography A 876:235-242.
Beattie, S.H. and Williams, A.G. (1999) Detection of toxigenic strains of Bacillus cereus and other Bacillus spp. with an improved cytotoxicity assay. Letters in Applied Microbiology 28:221-225.
Beattie, S.H. and Williams, A.G. (2000) Detection of toxins. In: Encyclopedia of food microbiology (eds. R.K. Robinson, C.A. Batt and P.D. Patel). pp. 141-148. Academic Press, A Harcourt Science and Technology Company, London.
Beecher, D.J. and Wong, A.C.L. (1994) Improved purification and characterization of hemolysin BL, a hemolytic dermonecrotic vascular permeability factor from Bacillus cereus. Infection and Immunity 62:980-986.
Beecher, D.J. and Wong, A.C.L. (1997) Tripartite hemolysin BL from Bacillus cereus. Hemolytic analysis of component interaction and model for its characteristic paradoxical zone phenomenon. Journal of Biological Chemistry 272:233-239.
Beecher, D.J., Schoeni, J.L. and Wong, A.C.L. (1995) Enterotoxin activity of hemolysin
G X_W G
BL from Bacillus cereus. Infection and Immunity 63:4423-4428.
Bennett, R.W., Murthy, G., Kaylor, L., Cox, S. and Harmon, S.M. (1993) Biological characterization and serological identification of Bacillus cereus diarrhoeal factor. Netherlands Milk dairy Journal 47:105-120.
Berkeley, R.C.W. (2002) CHAPTER 1 Whither Bacillus? In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 1-7. Blackwell Science Ltd., UK.
Blakey, L.J. and Priest, F.G. (1980) The occurrence of Bacillus cereus in some dried foods including pulses and cereals. Journal of Applied Bacteriology 48:297-302. Boone, C.M., Franke, J., Zeeuw, R.A. and Ensing, K. (1999) Evaluation of capillary electrophoretic techniques towards systematic toxicological analysis. Journal of Chromatography A 838:259-272.
Brousseau, R., Saint-Onge, A., Prefontaine, G., Masson, L. and Cabana, J. (1993) Arbitrary primer polymerase chain reaction, a powerful method to identify Bacillus thuringiensis serovars and strains. Applied and Environmental Microbiology 59:114-119.
Brüggemann, O. and Freitag, R. (1995) Determination of polycyclic aromatic hydrocarbons in soil samples by micellar electrokinetic capillary chromatography with photodiode-array detection. Journal of Chromatography A 717:309-324.
Buchanan, R.L. and Schultz, F.J. (1994) Comparison of the Tecra VIA kit, Oxoid BCET-RPLA kit and CHO cell culture assay for the detection of Bacillus cereus diarrhoeal enterotoxin. Letters in Applied Microbiology 19:353-356.
Busch, U. and Nitschko, H. (1999) Methods for the differentiation of microorganisms. Journal of Chromatography B 722:263-278.
Camilleri, P. (1998) Capillary electrophoresis: Theory and practice. 2nd edn. CRC Press. Boca Raton, Florida.
G X_X G
Cancalon, P.F. (1995) Capillary electrophoresis: A useful technique for food analysis. Food Technology 6:52-58. Carlson, C.R., Caugant, D.A. and Kolsto, A.B. (1994) Genotypic diversity of Bacillus cereus and Bacillus thuringiensis strains. Applied and Environmental Microbiology 60:1719-1725.
Cataldi, T.R.I., Nardiello, D., Carrara, V., Ciriello, R. and De Benedetto, G.E. (2003) Assessment of riboflavin and flavin content in common food samples by capillary electrophoresis with laser-induced fluorescence detection. Food Chemistry 82:309-314.
Chen, N., Terabe, S. and Nakagawa, T. (1995) Effect of organic modifier concentrates on electrokinetic migrations in micellar electrokinetic chromatography. Electrophoresis 16:1457-1462.
Cheng, Y., Fan, L., Chen, H., Chen, X. and Hu, Z. (2005) Method for on-line derivatization and separation of aspartic acid enantiomer in pharmaceuticals application by the coupling of flow injection with micellar electrokinetic chromatography. Journal of Chromatography 1072:259-265.
Cheung, H.Y., Cheung, C.S. and Kong, C.K. (2001) Determination of bioactive diterpenoids from Andrographis paniculata by micellar electrokinetic chromatography. Journal of Chromatography A 930:171-176.
Christiansson, A. (1993) Enterotoxin production in milk by Bacillus cereus: a comparison of methods for toxin detection. Netherlands Milk Dairy Journal 47:79-87.
Chun, J. and Bae, K.S. (2000) Phylogenetic analysis of Bacillus subtilis and related taxa based on partial gyrA gene sequences. Antonie van Leeuwenhoek 78:123-127.
Cifuentes, A., Bartolomé, B. and Gómez-Cordovés, C. (2001) Fast determination of procyanidins and other phenolic compounds in food samples by micellar electrokinetic chromatography using acidic buffers. Electrophoresis 22:1561-1567.
G X_Y G
Claus, D. and Berkeley, R.C.W. (1986) Genus Bacillus 1872. In: Bergey’s Manual of Systematic Bacteriology (eds P.H.A. Sneath, N.S. Mair, M.E. Sharpe, J.G. Holt). Vol. 2. pp. 1105-11401139. Williams and Wilkins, Baltimore, MD.
Cocolin, L., Manzano, M., Cantoni, C. and Comi, G. (2001) Denaturing gradient gel electrophoresis analysis of the 16S rRNA gene V1 region to monitor dynamic changes in the bacterial population during fermentation of Italian sausages. Applied and Environmental Microbiology 67(11):5113-5121.
Collins, M.D., Lawson, P.A., Willems, A., Cordoba, J.J., Fernandez-Garayzabal, J., Garcia, P., Cai, J., Hippe, H. and Farrow, J.A. (1994) The phylogeny of the genus Clostridium: proposal of five new genera and eleven new species combinations. International Journal of Systematic Bacteriology 44:812-826.
Cousins, C.M., Pettipher, G.L., McKinnon, C.H. and Mansell, R. (1979) A rapid method for counting bacteria in raw milk. Dairy Industry International 44:27-39.
Cox, J.M. and Fleet, G.H. (2003) New directions in the microbiological analysis of foods. In: Foodborne microorganisms of public health significance (eds. A.D. Hocking). 6th edn. pp. 103-161. AIFST Inc. (NSW Branch) Food Microbiology Group, Sydney.
Crielly, E.M.N., Logan, N.A. and Anderton, A. (1994) Studies on the Bacillus flora of milk and milk products. Journal of Applied Bacteriology 77:256-263.
D’Aoust, J.Y., Maishment, C., Stotland, P. and Boville, A. (1982) Surfactants for the effective recovery of Salmonella in fatty foods. Journal of Food Protection 45:249-252.
Da Silva, K.R.A., Rabinovitch, L. and Seldin, L. (1999) Phenotypic and genotypic diversity among Bacillus sphaericus strains isolated in Brazil, potentially useful as biological control agents against mosquito larvae. Research in Microbiology 150:153- 160.
Daffonchio, D., Borin, S., Frova, G., Gallo, R., Mori, E., Fani, R. and Sorlini, C. (1999)
G X_Z G
A randomly amplified polymorphic DNA marker specific for the Bacillus cereus group is diagnostic for Bacillus anthracis. Applied and Environmental Microbiology 65(3): 1298-1303.
Daffonchio, D., Borin, S., Frova, G., Manachini, P.L. and Sorlini, C. (1998) PCR fingerprinting of whole genomes: the spacers between the 16S and 23S rRNA genes and of intergenic tRNA gene regions reveal a different intraspecific genomic variability of Bacillus cereus and Bacillus licheniformis. International Journal of Systematic Bacteriology 48:107-116.
Dahl, M.K. (2000) Bacillus subtilis In: Encyclopedia of food microbiology (eds. R.K. Robinson, C.A. Batt and P.D. Patel). pp. 135-141. Academic Press, A Harcourt Science and TechnologyTecnnology Company, San Diego, San Francisco, New York, Boston, London, Sydney, Tokyo.
Damgaard, P.H., Larsen, H.D., Hansen, B.M., Bresciani, J. and Jorgensen, K. (1996) Enterotoxin-producing strains of Bacillus thuringiensis isolated from food. Letters in Applied Microbiology 23:146-150.
Dautle, M.P., Ulrich, R.L. and Hughes, T.A. (2002) Typing and subtyping of 83 clinical isolates purified from surgically implanted silicone feeding tubes by random amplified polymorphic DNA amplification. Journal of Clinical Microbiology 40(2):414-421.
Day, T.L., Tatani, S.R., Notermans, S. and Bennett, R.W. (1994) A comparison of ELISA and RPLA for detection of Bacillus cereus diarrhoeal enterotoxin. Journal of Applied Bacteriology 77:9-13.
De Vos, P. (2002) CHAPTER 10 Nucleic acid analysis and SDS-PAGE of whole-cell proteins in Bacillus taxonomy. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 141-159. Blackwell Science Ltd., UK.
Dickson, J.S. (1990) Comparison of homogenization by blending or stomaching on the
G X_[ G recovery of Listeria monocytogenes from beef tissues. Journal of Food Science 55:655- 657.
Dong, M., Oda, R.P., Strausbauch, M.A., Wettstein, P.J., Landers, J.P. and Miller, L.J. (1997) Hydrophobic peptide mapping of clinically relevant heptathelical membrane proteins by capillary electrophoresis. Electrophoresis 18:1767-1774.
Drobniewski, F.A. (1993) Bacillus cereus and related species. Clinical Microbiology Reviews 6(4):324-338.
Dufrenne, J., Soentoro, P., Tatini, S., Day, T. and Notermans, S. (1994) Characteristics of Bacillus cereus related to safe food production. International Journal of Food Microbiology 23:99-109.
Ehling-Schulz, M., Fricker, M. and Scherer, S. (2004) Identification of emetic toxin producing Bacillus cereus strains by a novel molecular assay. FEMS Microbiology Letters 232:189-195.
Emswiler, B.S., Pierson, C.J. and Kotula, A.W. (1977) Stomaching vs blending. Food Technology 40:40-42.
Entis, P., Brodsky, M.H. and Sharpe, A.N. (1982) Effect of pre-filtration and enzyme treatment on membrane filtration of foods. Journal of Food Protection 45:8-11.
Farber, J.M. (1996) An introduction to the hows and whys of molecular typing. Journal of Food Protection 59:1091-1101.
Farber, J.M. and Addison, C.J. (1994) RAPD typing for distinguishing species and strains in the genus Listeria. Journal of Applied Bacteriology 77:242-250.
Farber, J.M., Gendel, S.M., Tyler, K.D., Boerlin, P., Landry, W.L., Fritschel, S.J. and Barrett, T.J. (2001) Molecular typing and differentiation. In: Compendium of methods for the microbiological examination of foods (eds. K. Ito). 4th edn. pp. 127-156.
G X_\ G
American Public Health Association, Washington, D.C.
Fermanian, C., Lapeyre, C., Fremy, J.M. and Claisse, M. (1996) Production of diarrhoeal toxin by selected strains of Bacillus cereus. International Journal of Food Microbiology 30: 345-358.
Finlay, W.J.J., Logan, N.A. and Sutherland, A.D. (1999) Semiautomated metabolic staining assay for Bacillus cereus emetic toxin. Applied and Environmental Microbiology 65:1811-1812.
Fleet, G.H., Karalis, T., Hawa, A. and Lukondeh, T. (1991) A rapid method for enumerating Salmonella in milk powders. Letters in Applied Microbiology 13:255-259.
Fletcher, P. and Logan, N.A. (1999) Improved cytotoxicity assay for Bacillus cereus diarrhoeal enterotoxin. Letters in Applied Microbiology 28:393-400. Flurer, C.L. (1999) Analysis of antibiotics by capillary electrophoresis. Electrophoresis 20:3269-3279.
Fritze, D. (2002) CHAPTER 8 Bacillus identification – traditional approaches. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 100-122. Blackwell Science Ltd., UK.
From, C., Pukall, R., Schumann, P., Hormazábal, V. and Granum, P.E. (2005) Toxin- producing ability among Bacillus spp. outside the Bacillus cereus group. Applied and Environmental Microbiology 71(3):1178-1183.
Fu, X., Lu, J. and Chen, Y. (1998) Separation of polycyclic aromatic hydrocarbons by micellar electrokinetic chromatography with aqueous short-chain alcohol solvent. Talanta 46:751-756.
Fürtös-Matei, A., Day, R., St-Pierre, S.A., St-Pierre, L.G. and Waldron, K.C. (2000) Micellar electrokinetic chromatography separations of dynorphin peptide analogs. Electrophoresis 21:715-723.
G X_] G
Garbeva, P., van Overbeek, L.S., van Vuurde, J.W.L. and van Elsas, J.D. (2001) Analysis of endophytic bacterial communities of potato by plating and denaturing gradient gel electrophoresis(DGGE) of 16S rDNA based PCR fragment. Microbial Ecology 41:369-383.
Garbeva, P., van Veen, J.A. and van Elsas, J.D. (2003) Predominant Bacillus spp. in agricultural soil under different management regimes detected via PCR-DGGE. Microbial Ecology 45:302-316.
Ghelardi, E., Celandroni, F., Salvetti, S., Barsotti, C., Baggiani, A. and Senesi, S. (2002) Identification and characterization of toxigenic Bacillus cereus isolates responsible for two food-poisoning outbreaks. FEMS Microbiology Letters 208:129-134.
Gibson, T. and Gordon, R.E. (1974) Bacillus. In: Bergey’s Manual of Determinative Bacteriology (eds. R.E. Buchanan and N.E. Gibbons). 8th edn. pp. 529-550. Williams and Wilkins, Baltimore, MD.
Gilbert, R.J. and Parry, J.M. (1977) Serotypes of Bacillus cereus from outbreaks of food poisoning and from routine foods. Journal of Hygiene 78:69-74.
Glatz, B.A., Spira, W.M. and Goepfert, J.M. (1974) Alterations of vascular permeability in rabbits by culture filtrates of Bacillus cereus and related species. Infection and Immunity 10:299-303.
Gnanou Besse, N. and Lafarge, V. (2001) Development of a membrane filtration method for enumeration of Listeria monocytogenes from soft cheese. Food Microbiology 18:669-676.
Gordon, R.E. (1981) One hundred and seven years of the genus Bacillus. In: The Aerobic Endosporeforming Bacteria (eds. R.C.W. Berkeley and M. Goodfellow). pp. 1- 15. Academic Press, London.
Gordon, R.E., Haynes, W.C. and Pang, C.H.N. (1973) The genus Bacillus. Agriculture
G X_^ G handbook 427. US Deparment of Agriculture, Washington, D.C.
Granum, P.E. (1994) Bacillus cereus and its toxins. Journal of Applied Bacteriology Symposium Supplement 76:61S-66S.
Granum, P.E. (1997) Bacillus cereus In: Food microbiology: Fundamentals and frontiers (eds. M. Doyle, L. Beuchat and T. Montville). pp. 327-336. ASM press, Washington, D.C.
Granum, P.E. (2002) CHAPTER 4 Bacillus cereus and food poisoning. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 37-46. Blackwell Science Ltd., UK.
Granum, P.E. and Baird-Parker, T.C. (2000) Bacillus spp. In: The microbiological safety and quality of food (eds. B. Lund, T.C. Baird-Parker and G.W. Gould). pp. 1029-1039. Aspen Publisher, Gaitherburg, MD.
Granum, P.E. and Lund, T. (1997) Mini review: Bacillus cereus and its food poisoning toxins. FEMS Microbiology Letters 157:223-228.
Granum, P.E. and Rönner, U. (1998) The adhesion of Bacillus cereus spores to epithelial cells might be an additional virulence mechanism. International Journal of Food Microbiology 39:93-99.
Granum, P.E., Andersson, A., Gayther, C., Giffel, M., Larsen, H., Lund, T. and O’Sullivan, K. (1996) Evidence for a further enterotoxin complex produced by Bacillus cereus. FEMS Microbiology Letters 141:145-149.
Granum, P.E., Brynestad, S and O’Sullivan, K. and Nissen, H. (1993) Enterotoxin from Bacillus cereus: production and biochemical characterisation. Netherlands Milk Dairy Journal. 47: 63-70.
Granum, P.E., O’Sullivan, K. and Lund, T. (1999) The sequence of the non-haemolytic
G X__ G enterotoxin operon from Bacillus cereus. FEMS Microbiology Letters 177:225-229.
Graves, L.M., Swaminathan, B. and Hunter, S.B. (1999) Subtyping Listeria monocytogenes. In: Listeria, Listeriosis, and Food safety (eds. E.H. Marth). 2nd edn. pp. 279-297. Marcel Dekker, Inc., New York.
Griffiths, M.W. (1990) Toxin production by psychrotrophic Bacillus spp. present in milk. Journal of Food Protection 53:790-792.
Griffiths, M.W. (1992) Bacillus cereus in liquid milk and other milk products. Bulletin of the International Dairy Federation 275: 36-39.
Griffiths, M.W. and Phillips, J.D. (1990) Incidence, source and some properties of psychrotrophic Bacillus spp. found in raw and pasteurized milk. Journal of the Society of Dairy Technology 43, 62-66.
Guillaume-Gentil, O., Scheldeman, P., Marugg, J., Herman, L., Joosten, H. and Heyndrickx, M. (2002) Genetic heterogeneity in Bacillus sporothermodurans as demonstrated by ribotyping and repetitive extragenic palindromic – PCR fingerprinting. Applied and Environmental Microbiology 68(6):4216-4224.
Gunter, T.E. and Pfeiffer, D.R. (1990) Mechanisms by which mitochondria transport calcium. The American Journal of Physiology 258:C755-C786.
Häggblom, M.M., Apetroaie, C., Andersson, M.A. and Salkinoja-Salonen, M.S. (2002) Quantitative analysis of cereulide, the emetic toxin of Bacillus cereus, produced under various conditions. Applied and Environmental Microbiology 68:2479-2483.
Hale, K.A., Doores, S. and Walsh, R.A. (1990) An enzyme/surfactant treatment and filtration technique for the retrieval of Listeria monocytogenes from ice cream mix. Food Structure 9:61-67.
Hansen, B.M. and Hendriksen, N.B. (2001) Detection of enterotoxin Bacillus cereus
G X_` G and Bacillus thuringiensis strains by PCR analysis. Applied and Environmental Microbiology 67(1):185-189.
Hansen, B.M., Damgaard, P.H., Eilenberg, J. and Pedersen, J.C. (1998) Molecular and phenotypic characterization of Bacillus thuringiensis isolated from leaves and insects. Journal of Invertebrate Pathology 71:106-114.
Hansen, B.M., Høiby, P.E., Jensen, G.B. and Hendriksen, N.B. (2003) The Bacillus cereus bceT enterotoxin sequence reappraised. FEMS Microbiology Letters 223:21-24.
Harrell, L.J., Andersen, G.L. and Wilson, K.H. (1995) Genetic variability of Bacillus anthracis and related species. Journal of Clinical Microbiology 33:1847-1850.
Harrigan, W.F. (1998) CHAPTER 11 Reactions involving carbohydrate and other carbon compounds. In: Laboratory methods in food microbiology, 3rd edn. pp.107. Academic press, London.
Hawa, S.G., Morrison, G.J. and Fleet, G.H. (1984) Method to rapidly enumerate Salmonella on chicken carcasses. Journal of Food Protection 47:932-936.
Heinrichs, J.H., Beecher, D.J., Macmillan, J.M. and Zilinskas, B.A. (1993) Molecular cloning and characterization of the hblA gene encoding the B component of hemolysin BL from Bacillus cereus. Journal of Bacteriology 175:6760-6766.
Helgason, E., Caugant, D.A., Olsen, I. and Kosto, A.B. (2000a) Genetic structure of population of Bacillus cereus and B. thuringiensis isolates associated with peridontitis and other human infections. Journal of Clinical Microbiology 38:1615-1622.
Helgason, E., Økstad, O.A., Caugant, D.A., Johansen, H.A., Fouet, A., Mock, M., Hegna, I. and Kolstø, A.B. (2000b) Bacillus anthracis, Bacillus cereus, and Bacillus thuringiensis – One species on the basis of genetic evidence. Applied and Environmental Microbiology 66(6):2627-2630.
G X`W G
Herman, L., Heyndrickx, M. and Waes, G. (1998) Typing of Bacillus sporothermodurans and other Bacillus species isolated from milk by repetitive element sequence based PCR. Letters in Applied Microbiology 26:183-188.
Heuer, H., Hartung, K., Wieland, G., Karmer, I. and Smalla, K. (1999) Polynucleotide probes that target a hypervariable region of 16S rRNA genes to identify bacterial isolates corresponding to band of community fingerprints. Applied and Environmental Microbiology 65(3):1045-1049.
Heyndrickx, M. and Scheldeman, P. (2002) CHAPTER 6 Bacilli associated with spoilage in dairy products and other food. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 64-79. Blackwell Science Ltd., UK.
Heyndrickx, M., Lebbe, L., Kersters, K., De Vos, P., Forsyth, G. and Logan, N.A. (1998) Virgibacillus: a new genus to accommodate Bacillus pantothenticus (Proom and Knight 1950). Emended description of Virgibacillus pantothenticus. International Journal of Systematic Bacteriology 48:99-106.
Holbrook, R. and Anderson, J.M. (1980) An improved selective and diagnostic medium for the isolation and enumeration of Bacillus cereus in food. Canadian Journal of Microbiology 26:753-759.
Huang, H.-Y., Chiu, C.-W., Sue, S.-L. and Cheng, C.-F. (2003) Analysis of food colorants by capillary electrophoresis with larger-volume sample stacking. Journal of Chromatography A 995:29-36.
Hughes, S., Bartholomew, B., Hardy, J.C. and Kramer, J.M. (1988) Potential application of a HEp-2 cell assay in the investigation of Bacillus cereus emetic-syndrome food poisoning. FEMS Microbiology Letters 52:7-12.
Huis in’t Veld, J.H.J (1996) Microbiol and biochemical spoilage of foods: an overview. International Journal of Food Microbiology 33:1-8.
G X`X G
Husmark, U. and Rönner, U. (1993) The influence of hydrophobic electrostatic and morphologic properties on the adhesion of Bacillus spores. Biofouling 5:334-335.
Idei, M., SeprĘdi, J., GyĘrffy, E., Hollósy, F., Vadász, Zs., Mészáros, Gy., Teplán, I. and Kéri, Gy. (1998) Micellar electrokinetic chromatography of highly hydrophobic peptides. Analytica Chimica Acta 372:273-279.
Issaq, H.J. (2000) A decade of capillary electrophoresis. Electrophoresis 21:1921-1939. Jackson, S.G. (1991) Bacillus cereus. Journal – Association of Official Analytical Chemists 74(4):704-706.
Jackson, S.G. (1993) Rapid screening test for enterotoxin-producing Bacillus cereus. Journal of Clinical Microbiology 31:972-974.
Jackson, S.J., Goodbrand, R.B., Ahmed, R. and Kasatiya, S. (1995) Bacillus cereus and Bacillus thuringiensis isolated in a gastroenteritis outbreak investigation. Letters in Applied Microbiology 21:103-105.
Jang, S.S. (2004) Detection, enumeration, and diversity of Listeria species in foods using centrifugation-plating and cultural enrichment methods. PhD thesis. Department of Food Science and Technology. The University of New South Wales, Sydney.
Jensen, G.B., Hansen, B.M., Eilenberg, J. and Mahillon, J. (2003) The hidden lifestyles of Bacillus cereus and relatives. Environmental Microbiology 5(8): 631-640.
Jenson, I. (2000) Detection by classical cultural techniques. In: Encyclopedia of food microbiology (eds. R.K. Robinson, C.A. Batt and P.D. Patel). pp. 149-158. Academic Press, A Harcourt Science and Technology Company, London.
Jenson, I. and Moir, C.J. (2003) Bacillus cereus and other Bacillus species. In: Foodborne microorganismsmircroorganisms of public health significance (eds. A.D. Hocking). pp. 445-478. Southwood Press, Australia.
G X`Y G
Kampfer, P. (2002) CHAPTER 18 Whole-cell Fatty Acid Analysis in the Systematic of Bacillus and Related Genera (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 271-299. Blackwell Science Ltd., UK.
Kaneko, T., Nozaki, R. and Aizawa, K. (1978) Deoxyribonucleic acid relatedness between Bacillus anthracis, Bacillus cereus and Bacillus thuringiensis. Microbiology and Immunology 22:639-641.
Karalis, T. (2000) Cultural methods for the detection and enumeration of Salmonella in foods. PhD thesis. Department of Food Science and Technology. The University of New South Wales, Sydney.
Karmat, A.S., Nerkkar, D.P. and Nair, P.M. (1989) Bacillus cereus in some Indian Foods, incidence and antibiotic, heat and radiation resistance. Journal of Food Safety 10:31-41.
Klijn, N., Herman, L., Langeveld, L., Vaerewijck, M., Wagendorp, A.A., Huemer, I. and Weekamp, A.H. (1997) Genotypical and phenotypical characterization of Bacillus sporothermodurans strains, surviving UHT sterilisation. International Dairy Journal 7:421-428.
Klipski, K., Majewski, S., Kinsner, A. and Sladowski, D. (2000) Invittox on video: MTT test. http://www.ib.amwaw.edu.pl/home/dslade/video/mtt.html (data accessed: 03/03/2002)
Kortiranta, A., Lounatmaa, K. and Haapasalo, M. (2000) Epidemiology and pathogenesis of Bacillus cereus infections. Microbes and Infection 2(2):189-198.
Kramer, J.M. and Gilbert, R.J. (1989) Bacillus cereus and other Bacillus species. In: Foodborne bacterial pathogens (eds. M.P. Doyle). pp. 21-70. Marcel Dekker, Inc., New York.
Kramer, J.M. and Gilbert, R.J. (1992) Bacillus cereus gastroenteritis. In: Handbooks of Natural Toxins. (eds A.T. Tu). 7th edn. pp. 119-153. Marcel Dekker, Inc., New York.
G X`Z G
Kuse, M., Franz, T., Koga, K., Suwan, S., Isobe, M., Agata, N. and Ohta, M. (2000) High incorporation of L-amino acid to cereulide, an emetic toxin from Bacillus cereus. Bioorganic and Medicinal chemistry letters 10:735-739.
Lange, J., Thomas, K. and Wittmann, C. (2002) Comparison of a capillary electrophoresis method with high-performance liquid chromatography for the determination of biogenic amines in various food samples. Journal of Chromatography B 779:229-239.
Larger, P.J., Jones, A.D. and Dacombe, C. (1998) Separation of tea polyphenols using micellar electrokinetic chromatography with diode array detection. Journal of Chromatography A 799:309-320.
Larsen, H.D. and Jørgensen, K. (1999) Growth of Bacillus cereus in pasteurized milk products. International Journal of Food Microbiology 46:173-176.
Lee, K.J. and Heo, G.S. (1991) Application of home-made capillary zone electrophoresis system to the separation of organic molecules. Journal of Korean Chemical Society 35(3):219-225.
Lee, M.S. and Chang, D.S. (1980) Distribution and physiological characteristics of Bacillus cereus in rice and rice products. Bulletin of Korean Society. 13:163-171.
Lee, P.K., Buswell, J.A. and Shinagawa, K. (1995) Distribution of toxigenic Bacillus cereus in rice samples marketed in Hong Kong. World Journal of Microbiology and Biotechnology 11:696-698.
Liu, P.Y.F., Ke, S.C. and Chen, S.L. (1997) Use of pulsed field electrophoresis to investigate a pseudo-outbreak of Bacillus cereus in a pediatric unit. Journal of Clinical Microbiology 35:1533-1535.
Logan, N.A. (2002) CHAPTER 9 Modern methods for identification In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan
G X`[ G and P. De Vos). pp. 123-140. Blackwell Science Ltd., UK.
Lund, B.M. (1990) Food-borne disease due to Bacillus and Clostridium species. Lancet 336:982-986.
Lund, T. and Granum, P.E. (1996) Characterisation of a non-haemolytic enterotoxin complex from Bacillus cereus isolated after a foodborne outbreak. FEMS Microbiology Letters 141:151-156.
Lund, T. and Granum, P.E. (1997) Comparison of biological effect of the two different enterotoxin complexes isolated from three different strains of Bacillus cereus. Microbiology 143:3329-3336.
Lund, T. and Granum, P.E. (1999) The 105-kDa protein component of Bacillus cereus non-haemolytic enterotoxin (Nhe) is a metalloprotease with gelatinolytic and collagenolytic activity. FEMS Microbiology Letters 178:355-366.
Lund, T., De Buyser, M.L. and Granum, P.E. (2000) A new enterotoxin from Bacillus cereus that may cause necrotic enteritis. Molecular Microbiology 38:254-261.
Mahler, H., Pasi, A., Kramer, J.M., Schulte, P., Scoging, A.C., Bär, W. and Krähenbühl, S. (1997) Fulminant liver failure in association with the emetic toxin of Bacillus cereus. The New England Journal of Medicine 336(16):1142-1148.
Mäntynen, V. and Lindström, K. (1998) A rapid PCR-based DNA test for enterotoxic Bacillus cereus. Applied and Environmental Microbiology 64:1634-1639.
Markuszewski, M.J., Otsuka, K., Terabe, S., Matsuda, K. and Nishioka, T. (2003) Analysis of carboxylic acid metabolites from the tricarboxylic acid cycle in Bacillus subtilis cell extract by capillary electrophoresis using an indirect photometric detection method. Journal of Chromatography A 1010:113-121.
Maslow, J.N., Mulligan, M.E. and Arbeit, R.D. (1993) Molecular epidemiology:
G X`\ G
Application of contemporary techniques to the typing of microorganisms. Clinical Infection and Disease 17:153-164.
Maukonen, J., Mättö, J., Wirtanen, G., Raaska, L., Mattila-Sandholm, T. and Sarrela, M. (2003) Methodologies for the characterization of microbes in industrial environments: a review. Journal of Industrial Microbiology and Biotechnology 30:372-356.
Mayer, R., Eppert, I and Scherer, S. (1999) Incidence and identification of psychrotrophic (7˚C-tolerant) Bacillus spp. in German HTST pasteurized milk. Milchwissenschaft 54:26-30.
McElroy, A.P., Cohen, N.D. and Hargis, B.M. (1995) Evaluation of a centrifugation method for the detection of Salmonella enteritidis in experimentally contaminated chicken eggs. Journal of Food Protection 58:931-933.
McElroy, D.M., Jaykus, L. and Foegeding, P.M. (2000) Validation and analysis of modeled predictions of growth of Bacillus cereus spores in boiled rice. Journal of Food Protection 63(2):268-272.
McKillip, J.L. (2000) Prevalence and expression of enterotoxins in Bacillus cereus and other Bacillus spp., a literature review. Antonie van Leeuwenhoek 77:393-399.
Melling, J. and Capel, B.J. (1978) Characteristics of Bacillus cereus emetic toxin. FEMS Microbiology Letters 4:133-135.
Mendo, S.A.L.V., Henriques, I.S., Correia, A.C.M. and Duarte J.M.C. (2000) Genetic characterization of a new thermotolerant Bacillus licheniformis strain. Current Microbiology 40:137-139.
Meunier, J.-R. and Grimont P.A.D. (1993) Factors affecting reproducibility of random amplified polymorphic DNA fingerprinting. Research in Microbiology 144:373-379.
Mikami, T., Horikawa, T., Murakami, T., Matsumoto, T., Yamakawa, A., Murayama, S.,
G X`] G
Katagiri, S., Shinagawa, K. and Suzuki, M. (1994) An improved method for detecting cytostatic toxin (emetic toxin) of Bacillus cereus and its application to food samples. FEMS Microbiology Letters 119:53-58.
Mikkola, R., Saris, N.L., Grigoriev, P.A., Andersson, M.A. and Salkinoja-Salonen, M.S. (1999) Ionophoretic properties and mitochondrial effects of cereulide. European Journal of Biochemistry 263:112-117.
Mitchell, P. and Moyle, J. (1969) Estimation of membrane potential and pH difference across the cristae membrane of rat liver mitochondria. European Journal of Biochemistry 7:471-484.
Miteva, V., Selensk-Pobell, S. and Mitev, V. (1999) Random and repetitive primer amplified polymorphic DNA analysis of Bacillus sphaericus. Journal of Applied Microbiology 86: 928-936.
Morita, T.N. and Woods, M.J. (1977) Stimulation of Bacillus cereus growth by protein in cooked rice combinations. Journal of Food Science 42:1232-1235.
Morrison, G.J. and Fleet, G.H. (1985) Reduction of Salmonella on chicken carcasses by immersion treatments. Journal of Food Protection 48:939-943.
Mossel, D.A.A., Van Netten, P. and Pijper, M. (1991) A centrifugation/quadrant plate technique for the simplified differential-bacteriological examination of adequately heat- processed foods. Letters in Applied Microbiology 13:115-117.
Nakamura, L.K. (1994) DNA relatedness among Bacillus thuringiensis serovars. International Journal of Systematic Bacteriology 44:125-129.
Nakamura, L.K. and Jackson, M.A. (1995) Clarification of the taxonomy of Bacillus mycoides. International Journal of Systematic Bacteriology 45:46-49.
Neiderhauser, C., Höfelein, C., Allmann, M., Burkhalter, P., Lüthy, J. and Candrian, U.
G X`^ G
(1994) Random amplification of polymorphic bacterial DNA: evaluation of 11 oligonucleotides and application to food contaminated with Listeria monocytogenes. Journal of Applied Bacteriology 77:574-582.
Nishikawa, Y., Kramer, J.M., Hanaoka, M. and Yasukawa, A. (1996) Evaluation of serotyping, biotyping, plasmid banding pattern analysis, and HEp-2 vacuolation factor assay in the epidemiological investigation of Bacillus cereus emetic-syndrome food poisoning. International Journal of Food Microbiology 31:149-159.
Notermans, S. and Batt, C.A. (1998) A risk assessment approach for food-borne Bacillus cereus and its toxins. Journal of Applied Microbiology Symposium Supplement. 84: 51S-61S.
Nübel, U., Engelen, B., Felske, A., Snaidr, J., Wieshuber, A., Amann, R.I., Ludwig, W. and Backhaus, H. (1996) Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus polymyxa detected by TGGE. Journal of Bacteriology 178:5636-5643.
Oudhoff, K.A., Schoenmakers, P.J. and Kok, W.Th. (2003) Characterisation of polyethylene glycols and polypropylene glycols by capillary zone electrophoresis and micellar electrokinetic chromatography. Journal of Chromatography A 985:479-491.
Outtrup, H. and Jørgensen, S.T. (2002) CHAPTER 14 The importance of Bacillus species in the production of industrial enzymes. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 206-218. Blackwell Science Ltd., UK.
Pacova, Z., Svec, P., Stenfors, L.P., Vyletelova, M. and Sedlacek, I. (2003) Isolation of the psychrotolerant species Bacillus weihenstephanensis from raw cow’s milk. Czech Journal of Animal Science 48(2):93-96.
Peng, H., Ford, V., Framptom. E.W., Restaino, L., Shelef, L.A. and Spitz, H. (2001a) Isolation and enumeration of Bacillus cereus from foods on a novel chromogenic plating medium. Food Microbiology 18:231-238.
G X`_ G
Peng, J.S., Tasi, W.C. and Chou, C.C. (2001b) Surface characteristics of Bacillus cereus and its adhension to stainless steel. International Journal of Food Microbiology 65:105- 111.
Penna, T.C.V., Moraes, D.A. and Farjardo, D.N. (2002) The effect of nisin on growth kinetics from activated Bacillus cereus spores in cooked rice and in milk. Journal of Food Protection 65:419-422.
Perani, M., Bishop, A.H. and Vaid, A. (1998) Prevalence of E-exotoxin, diarrhoeal toxin and specific G-endotoxin in natural isolates of Bacillus thuringiensis. FEMS Microbiology Letters 160:55-60.
Perkins, J.B., Guterman, S.K., Howitt, C.L., Williams, V.E. and Pero, J. (1990) Streptomyces genes involved in biosynthesis of the peptide antibiotic valinomycin. Journal of Bacteriology 172;3108-3116.
Peterkin, P.J. and Sharpe, A.N. (1980) Membrane filtration of dairy products for microbiological analysis. Applied and Environmental Microbiology 39:1138-1143.
Pettipher, G.L., Mansell, R., McKinnon, C.H. and Cousins, C.M. (1980) Rapid membrane filtration - Epifluorescent microscopy technique for direct enumeration of bacteria in raw milk. Applied and Environmental Microbiology 39:423-429.
Phelps, R.J. and McKillip, J.L. (2002) Enterotoxin production in natural isolates of Bacillaceae outside the Bacillus cereus group. Applied and Environmental Microbiology 68(6):3147-3151.
Pirttijärvi, T.S.M., Andersson, M.A., Scoging, A.C. and Salkinoja-Salonen, M.S. (1999) Evaluation of methods for recognising strains of the Bacillus cereus group with food poisoning potential among industrial and environmental contaminants. Systematic and Applied Microbiology 22:133-144.
Pouchart, C.J., and Behnke, J. (1993) The Aldrich library of 13C and 1H FT NMR
G X`` G
Spectra: Ed. 1, Volume 1. (Aldrich Chemical Company Inc.: USA)
Priest, F.G (1993) Systematics and ecology of Bacillus. In: Bacillus subtilis and other Gram-positive bacteria pp. 3-16 American Society for Microbiology Press, Washington, D.C.
Priest, F.G. (1989) Isolation and identification of aerobic endospore-forming bacteria. In: Bacillus (eds. C.R. Harwood). pp. 27-56. Plenum Press, New York.
Priest, F.G. (1993) Systematics and ecology of Bacillus. In: Bacillus subtilis and other Gram-positive bacteria (eds. A.L. Sonenshein, J.A.H.R. Losick) pp. 3-16 American Society for Microbiology Press, Washington, D.C.
Priest, F.G. (2002) CHPATER 13 Bacillus sphaericus and its insecticidal toxins. In: Applications and systematics of Bacillus and relatives (eds. R. Berkeley, M. Heyndrickx, N. Logan and P. De Vos). pp. 195-205. Blackwell Science Ltd., UK.
Priest, F.G., Goodfellow, M. and Todd, C. (1988) A numerical classification of the genus Bacillus. Journal of General Microbiology 134:1847-1882.
Prüß, B.M., Dietrich, R., Nibler, B., Märtlbauer, E. and Scherer, S. (1999) The haemolytic enterotoxin HBL is broadly distributed among species of the Bacillus cereus group. Applied and Environmental Microbiology 65(12):5436-5442.
Rivera, A.M.G., Granum, P.E. and Priest, F.G. (2000) Common occurrence of enterotoxin genes and enterotoxicity in Bacillus thuringiensis. FEMS Microbiology Letters 190:151-155.
Ronimus, R.S., Parker, L.E. and Morgan, H.W. (1997) The utilization of RAPD-PCR for identifying thermophilic and mesophilic Bacillus species. FEMS Microbiology Letters 147:75-79.
Ronimus, R.S., Parker, L.E., Turner, N., Poudel, S., Rückert, A. and Morgan, H.W.
G YWW G
(2003) A RAPD-based comparison of thermophilic bacilli from milk powders. International Journal of Food Microbiology 85:45-61.
Rosenkvist, H. and Hansen, Å. (1995) Contamination profiles and characterisation of Bacillus species in wheat bread and raw materials for bread production. International Journal of Food Microbiology 26:353-363.
Rowan, N.J., Deans, K., Anderson, J.G., Gemmell, C.G., Hunter, I.S. and Chaithong, T. (2001) Putative virulence factor expression by clinical and food isolates of Bacillus spp. after growth in reconstituted infant milk formulae. Applied and Environmental Microbiology 67(9):3873-3881.
Ryan, P.A., Macmillan J.M. and Zilinskas, B.A. (1997) Molecular cloning and characterization of the genes encoding the L1 and L2 components of hemolysin BL from Bacillus cereus. Journal of Bacteriology 179:2551-2556.
Salkinoja-Salonen, M.S., Vuorio, R., Andersson, M.A., Kampfer, P., Andersson, M.C., Honkanen-Buzalski, T. and Scoging, A.C. (1999) Toxigenic strains of Bacillus licheniformis related to food poisoning. Applied and Environmental Microbiology 65(10):4637-4645.
Sarkar, P.K., Hasenack, B. and Nout, M.J.R. (2002) Diversity and functionality of Bacillus and related genera isolated from spontaneously fermented soybeans (Indian Kinema) and locust beans (African Soumbala). International Journal of Food Microbiology 77:175-186.
Schiemann, D. A. (1978) Occurance of Bacillus cereus and the bacteriological quality of Chinease “take out” foods. Journal of Food Protection 41:450-454.
Schoeni, J.L. and Wong, A.C.L. (2005) Bacillus cereus food poisoning and its toxins. Journal of Food Protection 68(3):636-648.
Schraft, H. and Griffiths, M.W. (1995) Specific oligonucleotide primers for detection of
G YWX G lecithinase-positive Bacillus spp. by PCR. Applied and Environmental Microbiology 61(1): 98-102.
Schraft, H., Steele, M., McNab, B., Odumeru, J. and Griffiths, M.W. (1996) Epidemiological typing of Bacillus spp. isolated from food. Applied and Environmental Microbiology 62(11):4229-4232.
Seki, T., Chung, C.K., Mikami, H. and Oshima, Y. (1978) Deoxyribonucleic acid homology and taxonomy of the genus Bacillus. International Journal of Systematic Bacteriology 28:182-189.
Shangkuan, Y.H., Yang, J.F., Lin, H.C. and Shaio, M.F. (2000) Comparison of PCR- RFLP, ribotyping and ERIC-PCR for typing Bacillus anthracis and Bacillus cereus strains. Journal of Applied Microbiology 89(3):452-462.
Sharpe, A.N. (2000) Detection of microorganisms in foods: Principles of physical methods for separation and associated chemical and enzymological methods of detection. In: The microbiological safety and quality of food (eds. G.W. Gould). vol. 2. pp. 1734-1760. Aspen publishers, Inc., Maryland.
Sharpe, A.N. and Harshman, G.C. (1976) Recovery of Clostridium perfringens, Staphylococcus aureus and molds by the Stomacher: effect of fat content, surfactant concentration and blending time. Canadian Institute of Food Science and Technology Journal 9:30-34.
Shida, O., Takagi, H., Kadowaki, K. and Komagata, K. (1996) Proposal for two new genera, Brevibacillus gen. nov. and Aneurinibacillus gen. nov. International Journal of Systematic Bacteriology 46:939-946.
Shihabi, Z.K. and Friedburg, M.A. (1997) Analysis of small molecules for clinical diagnosis by capillary electrophoresis. Electrophoresis 18:1724-1732.
Shinagawa, K. (1990) Analytical methods for Bacillus cereus and other Bacillus species.
G YWY G
International Journal of Food Microbiology 10:125-142.
Shinagawa, K. (1993) Serology and characterisation of toxigenic Bacillus cereus. Netherlands Milk Dairy Journal 47:89-103.
Shinagawa, K., Konuma, H., Sekita, H. and Sugii, S. (1995). Emesis of rhesus monkeys induced by intragastric administration with the HEp-2 vacuolation factor (cereulide) produced by Bacillus cereus. FEMS Microbiology Letters 130:87-90.
Shinagawa, K., Ueno, S., Konuma, H., Matsusaka, N. and Sugii, S. (1991) Improved methods for purification of an enterotoxin produced by Bacillus cereus. FEMS Microbiology Letters 80:1-6.
Siriamornpun, S., Wootton, M., Cox, J.M., Bekes, F. and Wrigley, C.W. (2001) Capillary electrophoresis of wheat gliadin proteins and its potential for wheat varietal identification using pattern matching software. Australian Journal of Agricultural Research 52:839-843.
Slepecky, R.A. and Hemphill, H.E. (1992) The genus Bacillus – nonmedical. In: The prokaryotes (eds. A. Balows, H.G. Trüper, M. Dworkin, W. Harder and K.-H. Schleifer). vol. II. pp. 1663-1696. Springer, New York.
Sorokulova, I.B., Reva, O.N., Smirnov, V.V., Pinchuk, I.V., Lapa, S.V. and Urdaci, M.C. (2003) Genetic diversity and involvement in bread spoilage of Bacillus strains isolated from flour and ropy bread. Letters in Applied Microbiology 37:169-173. Southern, E.M. (1975) Detection of specific sequences among DNA fragments separated by gel electrophoresis. Journal of Molecular Biology 98:503-517.
Standards Unit, Evaluations and Standards Laboratory. (2002) Standard operating procedure: Enumeration of Bacillus cereus and other Bacillus species. Specialist and Reference Microbiology Division, Health Protection Agency, London, UK.
Stull, T.L., LiPuma, J.J. and Edlind, T.D. (1988) A broad-spectrum probe for molecular
G YWZ G epidemiology of bacteria: ribosomal RNA. Journal of Infectious Disease 157:280-286.
Suihko, M.-L. and Stackebrandt, E. (2003) Identification of aerobic mesophilic bacilli isolated from board and paper products containing recycled fibres. Journal of Applied Microbiology 94:25-34.
Surtherland, A.D. and Murdoch, R (1994) Seasonal occurrence of psychrotrophic Bacillus species in raw milk, and studies on the interactions with mesophilic Bacillus spp. International Journal of Food Microbiology 21:279-292.
Suwan, S., Isobe, M., Ohtani, I., Agata, N., Mori, M. and Ohta, M. (1995) Structure of cereulide, a cyclic dodecadepsipeptide toxin from Bacillus cereus and studies on NMR characteristics of its alkali metal complexes including a conformational structure of the K+ complex. Journal of the Chemical Society Perkin Transactions 1 7:765-775.
Szabo, R.A., Speirs, J.I. and Akhtar, M. (1991) Cell culture detection and conditions for production of a Bacillus cereus heat-stable toxin. Journal of Food Protection 54(4):272- 276.
Szabo, R.A., Todd, E.C.D. and Rayman, M.K. (1984) Twenty-four hour isolation and confirmation of Bacillus cereus in foods. Journal of Food Protection 47:856-860.
Tatzel, R., Ludwig, W., Schleifer, K.H. and Wallnöfer, P.R. (1994) Identification of Bacillus strains isolated from milk and cream with classical and nucleic acid hybridisation methods. Journal of Dairy Research 61:539-535
Taylor, J.M.W., Sutherland, A.D., Aidoo, K.E. and Logan, N.A. (2005) Heat-stable toxin production by strains of Bacillus cereus, Bacillus firmus, Bacillus megaterium, Bacillus simplex and Bacillus licheniformis. FEMS Microbiology Letters 242; 313-317. te Giffel, M.C., Beumer, R.R., Granum, P.E. and Rombouts, F.M. (1997) Isolation and characterisation of Bacillus cereus from pasteurized milk in household refrigerators in the Netherlands. International Journal of Food Microbiology 34:307-318.
G YW[ G
Terabe, S., Otsuka, K., Ichikawa, K., Tsuchiya, A. and Ando, T. (1984) Electrokinetic separations with micellar solutions and open-tubular capillaries. Analytical Chemistry 56(1):111-113.
Thompson, C.O., Trenerry, V.C. and Kemmery, B. (1995) Micellar electrokinetic capillary chromatographic determination of artificial sweeteners in low-Joule soft drinks and other foods. Journal of Chromatography A 694:507-514.
Toh, M., Moffitt, M.C., Henrichsen, L., Raftery, M., Barrow, K., Cox, J.M., Marquis, C.P. and Neilan, B.A. (2004) Cereulide, the emetic toxin of Bacillus cereus, is putatively a product of nonribosomal peptide synthesis. Journal of Applied Microbiology 97:992-1000.
Toh, M., Oh, M.H., Neilan, B.A., Moffitt, M.C., Raftery, M., Marquis, C.P., Henrichsen, L. and Cox, J.M. (2005) The production of the emetic toxin cereulide, by other Bacillus species. Applied and Environmental Microbiology (submitted).
Truong, T.L., Menard, C., Mouton, C. and Trahan, L. (2000) Identification of mutans and other oral streptococci by random amplified polymorphic DNA analysis. Journal of Medical Microbiology 49:63-71.
Uyttendaele, M., Hoorde, I.V. and Debevere, J. (2000) The use of immuno-magnetic separation (IMS) as a tool in a sample preparation method for direct detection of L. monocytogenes in cheese. International Journal of Food Microbiology 54:205-212. van Leeuwen, W., Sijmons, M., Sluijs, J. Verbrugh, H. and van Belkum, A. (1996) On the nature and use of randomly amplified DNA from Staphylococcus aureus. Journal of Clinical Microbiology 34:2270-2277. van Netten, P. and Kramer, J.M. (1992) Media for the detection and enumeration of Bacillus cereus in foods: a review. International Journal of Food Microbiology 17:85- 99.
G YW\ G
Versalovic, J., Koeuth, T. and Lupski, J.R. (1991) Distribution of repetitive DNA sequences in eubacteria and application to fingerprinting of bacterial genomes. Nucleic Acids Research 19:6823-6831.
Wainö, M., Tindall, B.J., Schumann, P. and Ingvorsen, K. (1999) Gracilibacillus gen. nov., with description of Gracilibacillus halotolerans gen. nov., sp. nov.; transfer of Bacillus dipsosauri to Gracilibacillus dipsosauri comb. nov., and Bacillus salexigens to the genus Salibacillus gen. nov., as Salibacillus salexigens comb. nov. International Journal of Systematic Bacteriology 49:821-831.
Watanabe, T., Yamamoto, A., Nagai, S. and Terabe, S. (1998) Micellar electrokinetic chromatography as an alternative to high-performance liquid chromatography for separation and determination of phenolic compounds in Japanese spirituous liquor. Journal of Chromatography A 793:409-413.
Weinmann, W., Parker, C.E., Baumeister, K., Maier, C., Tomer, K.B. and Przybylski, M. (1994) Capillary electrophoresis combined with 252Cf plasma desorption and electrospray mass spectrometry for the structural characterization of hydrophobic polypeptides using organic solvents. Electrophoresis 15:228-233.
Wiedmann, M. (2002) Molecular subtyping methods for Listeria monocytogenes. Journal of AOAC International 85:524-531.
Wisotzkey, J.D., Jurtshuk, P., Jr., Fox, G.E., Deinhard, G. and Poralla, K. (1992) Comparative sequence analyses on the 16S rRNA (rDNA) of Bacillus acidocaldarius, Bacillus acidoterrestris, and Bacillus cycloheptanicus and proposal for creation of a new genus, Alicyclobacillus gen. nov. International Journal of Systematic Bacteriology 42:263-269.
Woodburn, M.A., Younsten, A.A. and Hilu, K.H. (1995) Random amplified polymorphic DNA fingerprinting of mosquito-pathogenic and non-pathogenic strains of Bacillus sphaericus. International Journal of Systematic Bacteriology 45:212-217.
G YW] G
Yamada, S., Ohashi, E., Agata, N. and Venkateswaran, K. (1999) Cloning and nucleotide sequence analysis of gyrB of Bacillus cereus, B. thuringiensis, B. mycoides, and B. anthracis and their application to the detection of B. cereus in rice. Applied and Environmental Microbiology 65(4):1483-1490.
Yokoyama, K., Ito, M., Agata, N., Isobe, M., Shibayama, K., Horii, T. and Ohta, M. (1999) Pathological effect of synthetic cereulide, an emetic toxin of Bacillus cereus, is reversible in mice. FEMS Immunology and Medical Microbiology. 24:115-120.
Yuan, Z., Hansen, B.M., Andrup, L. and Eilenberg, J. (2002) Detection of enterotoxin genes in mosquito-larvicidal Bacillus species. Current Microbiology 45:221-225.
Zahner, V., Momen, H. and Priest, F. (1998) Serotype H5a and H5b is a major clone within mosquito-pathogenic strains of Bacillus sphaericus. Systematic and Applied Microbiology 21: 162-170.
Zahner, V., Rabdinovitch, L., Suffys, P. and Momen, H. (1999) Genotypic Diversity among Brevibacillus laterosporus strains. Applied and Environmental Microbiology 65(11):5182-5185.
Zhang, Y.C., Ronimus, R.S., Turner, N., Zhang, Y. and Morgan, H.W. (2002) Enumeration of thermophilic Bacillus species in composts and identification with a random amplification polymorphic DNA (RAPD) protocol. Systematic and Applied Microbiology 25:618-626.
Zoratti, M. and Szabo, I. (1995) The mitochondria permeability transition. Biochimica et Biophysica acta 1241:139-176.
G YW^ Appendices Appendix 1 Table Samples of raw rice examined for Bacillus species by the centrifugation-plating method and the spread plating method
Sample Location of Purchase Brand Name 1 Long Grain White Rice Coles, Pagewood Uncle Ben’s 2 Long Grain Rice Franklins, Pagewood Premier 3 Premium White Long Grain Rice IGA, Kingsford Sun Rice 4 Premium White Long Grain Rice IGA, Kingsford Sun Rice 5 Long Grain Rice Woolworths, Pagewood Home Brand 6 Long Grain Rice Coles, Randwick Farmland 7 Organic White Long Grain Rice Coles, Maroubra Sun Rice 8 Jasmine Rice Coles, Pagewood Coles 9 Long Grain Jasmine Fragrant Rice IGA, Kingsford Koala 10 Jasmine Rice Coles, Maroubra Golden Phoenix 11 Extra Long Grain Jasmine Rice Coles, Maroubra Kumarnthong 12 Basmati Aromatic Long Grain Rice Woolworths, Pagewood Riviana 13 Basmati Rice Coles, Pagewood Coles 14 Pure Basmati Rice Coles, Randwick Tilda 15 Long Grain Pure Basmati Rice Coles, Maroubra Taj Premium 16 Basmati Wild Rice Coles, Randwick Tilda 17 White Long Grain Wild Rice IGA, Kingsford Sun Rice 18 Cambrian Wild Rice Coles, Randwick Cambrian 19 Mahatma Natural Brown Long Grain Rice Franklins, Pagewood Riviana 20 White Calrose Medium Grain Rice IGA, Kingsford Sun Rice 21 Medium Grain Rice IGA, Kingsford Sun Rice 22 Arborio Mediterranean Medium Grain Rice Woolworths, Pagewood Sun Rice 23 Brown Calrose Medium Grain Rice Coles, Pagewood Sun Rice 24 Italian Long Grain Arborio Rice Coles, Pagewood Belmonte 25 Arborio Risotto Rice Coles, Randwick Uncle Ben’s 26 Arborio Risotto Rice Coles, Pagewood Riviana 27 Japanese Style Sushi Rice Woolworths, Pagewood Sun Rice 28 Glutinous Rice Local Asian Grocery Golden Boy Shop, Randwick 29 A1 Super Broken Rice Local Asian Grocery Lion Shop, Randwick 30 A1 Super Broken Rice Local Asian Grocery Rose Shop, Randwick 31 Ezi Cooking Basmati Rice Coles, Maroubra Hoyt’s 32 Premium Parboiled Rice Coles, Pagewood Tastic 33 Oriental Fried Rice Woolworths, Pagewood Continental 34 Chicken Rice Woolworths, Pagewood Rice & Riso 35 Oriental Style Rice Woolworths, Pagewood Sun Rice
G YW` Appendix 2 Table Samples of cooked rice examined for Bacillus species by the centrifugation- plating method and the spread plating method
Sample Location of Purchase 1 Lab Cooked Rice Laboratory 2 Lab Cooked Rice Laboratory 3 Lab Cooked Rice Laboratory 4 Cooked Rice Indian Restaurant, Randwick 5 Cooked Rice Indian Takeaway Shop, Pagewood 6 Cooked Rice Chinese Restaurant, Kingsford 7 Cooked Rice with Sausage & Chicken Chinese Restaurant, UNSWa 8 Cooked Rice with Vegetables & Meat Chinese Takeaway Shop, Burwood 9 Cooked Rice with Vegetables & Meat Chinese Takeaway Shop, Randwick 10 Rice with Oyster Sauce Chinese Restaurant, UNSWa 11 Cooked Rice with Ram Indian Takeaway Shop, Randwick 12 Cooked Rice with Pork Chinese Takeaway Shop, UNSWa 13 Cooked Rice with Beef Chinese Takeaway Shop, Strathfield 14 Cooked Rice with Chicken Korean-Japanese Restaurant, Kingsford 15 Cooked Rice with Vegetables Chinese Takeaway Shop, UNSWa 16 Risotto Pasta Takeaway Shop, Pagewood 17 Brown Rice Pasta Takeaway Shop, Pagewood 18 Brown Rice Pasta Takeaway Shop, Randwick 19 Fried Rice Chinese Takeaway Shop, Randwick 20 Fried Rice Chinese Restaurant, Kingsford 21 Fried Rice Korean-Japanese Takeaway Shop, Pagewood 22 Fried Rice Chinese Takeaway Shop, Pagewood 23 Fried Rice Chinese Takeaway Shop, Burwood 24 Fried Rice Chinese Restaurant, Kingsford 25 Fried Rice Chinese Takeaway Shop, Randwick 26 Sushi with Chicken Korean-Japanese Takeaway Shop, Pagewood 27 Sushi with Salmon Japanese Restaurant, Randwick 28 Sushi with Seaweeds Korean-Japanese Takeaway Shop, Pagewood 29 Sushi with Prawn Korean-Japanese Takeaway Shop, Pagewood 30 Sushi with Salmon Korean-Japanese Takeaway Shop, Pagewood 31 Sushi with Beef Japanese Takeaway Shop, Randwick 32 Sushi with Salmon Japanese Takeaway Shop, Randwick 33 Egg Fried Rice Woolworths, Pagewood 34 Tomato Rice Woolworths, Pagewood 35 Thai Lime & Coriander Basmati Rice Woolworths, Pagewood aUNSW, The University of New South Wales, Sydney
G YXW Appendix 3G
G
G
G
1 2 Figure shows a H NMR spectrum (600 MHz) of cereulide in (CD3)2SO([ H6]DMSO) depicting 2 sets of doublet NH signal (G 8.0-8.4), 4 alpha protons (G 4.2-5.1) and 7 sets of doublet methyl signals (G 0.7-1.4) (Source: Suwan et al., 1995). G G
G YXX