Upper and Lower Case Letters to Be Used

Isolation, characterization and genome sequencing of a soil-borne

Citrobacter freundii strain capable of detoxifying trichothecene mycotoxins

Rafiqul Islam

A Thesis

Presented to

The University of Guelph

In Partial Fulfilment of Requirements

for the Degree of

Doctor of Philosophy

Plant Agriculture

Guelph, Ontario, Canada

ISOLATION, CHARACTERIZATION AND GENOME SEQUENCING OF A SOIL-

BORNE CITROBACTER FREUNDII STRAIN CAPABLE OF DETOXIFIYING

TRICHOTHECENE MYCOTOXINS

Rafiqul Islam Advisors:

University of Guelph, 2012 Dr. K. Peter Pauls

Dr. Ting Zhou

Cereals are frequently contaminated with tricthothecene mycotoxins, like deoxynivalenol (DON, vomitoxin), which are toxic to humans, animals and plants. The goals of the research were to discover and characterize microbes capable of detoxifying DON under aerobic conditions and moderate temperatures. To identify microbes capable of detoxifying

DON, five soil samples collected from Southern Ontario crop fields were tested for the ability to convert DON to a de-epoxidized derivative. One soil sample showed DON de-epoxidation activity under aerobic conditions at 22-24°C. To isolate the microbes responsible for DON detoxification (de-epoxidation) activity, the mixed culture was grown with antibiotics at 50ºC for 1.5 h and high concentrations of DON. The treatments resulted in the isolation of a pure

DON de-epoxidating bacterial strain, ADS47, and phenotypic and molecular analyses identified the bacterium as Citrobacter freundii. The bacterium was also able to de-epoxidize and/or de-acetylate 10 other food-contaminating trichothecene mycotoxins. A fosmid genomic DNA library of strain ADS47 was prepared in E. coli and screened for DON detoxification activity. However, no library clone was found with DON detoxification activity. The whole genome of ADS47 was also sequenced, and from comparative genome analyses, 10 genes (characterized as reductases, oxidoreductases and deacetylase) were identified as potential candidates for DON/trichothecene detoxifying enzymes. Bioinformatic analyses showed that the major animal pathogenic locus of enterocyte effacement (LEE) operon genes absent from the ADS47 genome. Strain ADS47 has the potential to be used for feed detoxification and the development of mycotoxin resistance cereal cultivars.

DEDICATIONS

The Ph.D. thesis is dedicated to the departed souls of my mother MOZIBUN NESSA and

father-in-law DR. M.A. AHAD MIAH

iii

ACKNOWLEDGEMENTS

Foremost, I am deeply indebted to my advisors Prof. K. Peter Pauls and Dr. Ting Zhou for the continuous support of my Ph.D. study; for their endurance, inspiration, enthusiasm and mentoring. Their continued guidance has helped me in all aspects of my research and in writing of this dissertation. I sincerely express my gratitude to them for their co-operation from the beginning to conclusion of the thesis and many other areas. This will guide me in my future professional career.

My thesis committee guided me through all these years. I would like to express my gratitude to Prof. Paul Goodwin for his insightful comments and advice on my thesis research.

Special thanks to Dr. Chris Young for teaching me chemistry, chemical structures and interpreting liquid chromatography ultraviolet mass spectra data. I am sincerely grateful to

Dion Lepp for helping in analyzing the bacterial genome and sharing his knowledge and valuable suggestions on writing the genomic chapter.

I would like to express my appreciation to staff members of Dr. Ting Zhou and Prof.

K. Peter Pauls who supported me in many aspects during my PhD study. I am thankful to

Jianwei He, Xiu-Zhen Li, Honghui Zhu for collections and preparation of soil samples and

LC-UV-MS analyses of numerous samples. Special thanks to Dr. David Morris Johnston

Monje for sharing his knowledge, discussions and camaraderie at coffee breaks.

I wish to acknowledge the Natural Sciences and Engineering Research Council of

Canada (NSERC) for awarding me a prestigious scholarship for my Ph.D. program and

Dean’s Tri-Council Scholarship (University of Guelph), which made my life easier during this study. I also would like to convey my thanks to Agriculture and Agri-Food Canada/Binational

Agricultural Research and Development Fund (AAFC/BARD) project for providing me financial support.

Mike Bryan, a very supportive person, helped provide solutions to my software issues that improved my exposure and caliber at conference awards competitions. Several student competition awards honored and inspired me to work hard and lead to the successful completion of my thesis research. I would like to convey my appreciation to the judges of different student award committees who recognized my work. They are: Society for Industrial

Microbiology and Biotechnology conference 2010 (USA); Food Safety Research Forum-

Ontario Ministry of Agriculture and Rural Affairs annual meeting 2010 (Canada);

International Mycological Conference 2009 (UK); and Southwestern Ontario Regional

Association of Canadian Phytopathological Society annual meeting 2009-2011 (Canada).

I would like to express my very special thanks to Prof. Golam Hafeez Kennedy, my brother-in-law, who is like my elder brother. In many aspects I am indebt to him for his cooperations which aided me to achieve my current status.

Lastly, and most importantly, I wish to give my special thanks to my wife Mimi and lovely daughter Racy, whose endless love, sacrifice, support and encouragement enabled me to cross the tough long road and made possible to achieve my goal and aspirations. Racy, once you told me that one of your desires for me is to finish my Ph.D., which is done now. I know you want to be a scientist, which is a great ambition. I wish you all the best in your future aims Racy.

TABLE OF CONTENTS

DEDICATIONS…………………………………………………………………………...…iii

ACKNOWLEDGEMENTS…………………………………………………………………iv

TABLE OF CONTENTS……………………………………………………………………vi

LIST OF TABLES…………………………………………………………………………...ix

LIST OF FIGURES………………………………………………………………………...... x

LIST OF ABBREVIATIONS……………………………………………………………...xiii

1.0. GENERAL INTRODUCTION AND LITERATURE REVIEW …………………….1

1.1. Background………………………………………………………………………………..1

1.2. Plant pathogenesis and trichothecene production…………………………………………3

1.3. Biosynthesis of food contaminating TC…………..……………………………………….6

1.4. Occurrence of TC mycotoxins in food and feed ………………………………………...11

1.5. Significance of deoxynivalenol and trichothecenes……………………………………...16

1.6. Toxic effects of deoxynivalenol and trichothecenes……………………………………..19

1.7. Cellular and molecular mechanisms of DON toxicity in animals and plants…….……...22

1.8. Strategies to control Fusarium head blight and trichothecenes...………………………..26

1.9. Isolation of DON/TC de-epoxidating micro-organisms and genes………………….…...28

1.10. The bacterial species Citrobacter freundii……………………………………………...30

1.11. Hypotheses and objectives………………………………………………………..…….31

2.0. AEROBIC AND ANAEROBIC DE-EPOXIDATION OF MYCOTOXIN

DEOXYNIVALENOL BY BACTERIA ORIGINATING FROM AGRICULTURAL

SOIL...... 33

2.1. Abstract ………………………………………………………………………………….33

2.2. Introduction ……………………………………………………………………………...34

2.3. Materials and methods…………………………………………………………………...37

2.4. Results …………………………………………………………………………………...45

2.5. Discussion………………………………………………………………………………..62

2.6. Acknowledgements………………………………………………………………………64

3.0. ISOALTION AND CHARACTERIZATION OF A NOVEL SOIL BACTERIUM

CAPABLE OF DETOXIFYING ELEVEN TRICHOTHECENE MYCOTOXINS..65

3.1. Abstract ………………………………………………………………………………….65

3.2. Introduction ……………………………………………………………………………..66

3.3. Materials and methods…………………………………………………………………...68

3.4. Results …………………………………………………………………………………...75

3.5. Discussion………………………………………………………………………………..96

3.6. Acknowledgements……………………………………………………………………..100

4.0. CONSTRUCTION AND FUNCTIONAL SCREENING OF GENOMIC DNA

LIBRARY FOR ISOLATION OF DON DE-EPOXYDIZING GENE(S) FROM

BACTERIAL STRAIN ADS47……………………...... 102

4.1. Abstract ………………………………………………………………………………...102

4.2. Introduction …………………………………………………………………………….103

4.3. Materials and methods………………………………………………………………….104

vii

4.4. Results ………………………………………………………………………………….112

4.5. Discussion………………………………………………………………………………118

4.6. Acknowledgements……………………………………………………………………..121

5.0. GENOME SEQUENCING, CHARACTERIZATION AND COMPARATIVE

ANALYSIS OF A TRICHOTHECENE MYCOTOXIN DETOXIFYING

CITROBACTER FREUNDII STRAIN ADS47…………………………………….....122

5.1. Abstract ………………………………………………………………………………...122

5.2. Introduction …………………………………………………………………………….123

5.3. Materials and methods………………………………………………………………….125

5.4. Results ………………………………………………………………………………….129

5.5. Discussion………………………………………………………………………………162

5.6. Acknowledgements……………………………………………………………………..168

6.0. CONCLUSION………………………………………………………………………...171

6.1. Summary ……………………………………………………………………………….171

6.2. Research contributions …………………………………………………………………174

6.3. Future work……………………………………………………………………………..175

7.0. REFERENCES………………………………………………………………………...176

viii

LIST OF TABLES

Table 1-1. Fusaria produced type-A and type-B trichothecene mycotoxins…………………..9

Table 1-2. Worldwide occurrence of deoxynivalenol (DON) in animal feeds…………….....12

Table 1-3. Occurrence of DON in Canadian grains, grains foods and beers…………………13

Table 1-4. Legislative maximum tolerated levels of trichothecene mycotoxins in Canada….14

Table 1-5. In vitro and in vivo synergistic effects of different food contaminating mycotoxins ……………………………………………………………...... 18

Table 3-1. Microbial transformation of type-A and type-B tricthothecene mycotoxins after 72 h of incubation under aerobic conditions and at room temperature………..………..95

Table 4-1. Analysis of Fosmid DNA library for deoxynivalenol (DON) de-epoxidation

capability ……..…………………………………………………………………………….117

Table 5-1. General features of C. freundii strain ADS47 genome…...…….………………..132

Table 5-2. Genome comparisons between strain ADS47 and five Citrobacter species...... 141

Table 5-3. Functional categorized gene abundances among the six Citrobacter genomes…145

Table 5-4. Unique genes of C. freundii strain ADS47 ..…..………………………………...156

Supplemental Table S5-1. List of unique genes of strain ADS47, gene IDs and their functional roles………………………………………………………………………………169

LIST OF FIGURES

Figure 1-1. Structures of trichothecene mycotoxins produced by Fusaria…………………….7

Figure 1-2. Cellular and molecular mechanisms of deoxynivalenol toxicity in eukaryotic organisms…..……………………………………………………………………..23

Figure 2-1. Mechanism of microbial DON de-epoxidation…………………………………..36

Figure 2-2. LC-UV-MS chromatographic analysis of DON and its de-epoxidated product dE-DON obtained by culturing with selected soil microbes………………………………….46

Figure 2-3. Time course of relative biotransformation of DON to dE-DON by the initial soil suspension culture in MSB medium……………………………………………………...49

Figure 2-4. Microbial de-epoxidation of DON after five days of incubation….……………..51

Figure 2-5. Analysis of microbial population reduction by terminal restriction fragment length polymorphism (T-RFLP)……………………………………………………………...53

Figure 2-6. Microbial genera identified in the enriched DON de-epoxidizing microbial culture…………………………………………………………………………………………56

Figure 2-7. Effects of culture conditions on DON to dE-DON transformation by the enriched microbial culture…………………………………………………………………....58

Figure 2-8. Microbial DON to dE (de-epoxy)-DON transformation activity under aerobic and anaerobic conditions……………………………………………………………………...61

Figure 3-1. Colony color, shape and size of the mycotoxin degrading bacterial strain

ADS47 …….………………………………………………………………………………….76

Figure 3-2. Dendrogram based on 16S-rDNA sequences for strain ADS47 and 11 reference Citrobacter species……….………………………………………………………...77

Figure 3-3. Effects of DON concentration on bacterial growth and de-epoxidation activities………………………………………………………………………………………79

Figure 3-4. Bacterial growth and DON de-epoxidation capabilities under aerobic and anaerobic conditions ……...………………………………………………………………….82

Figure 3-5. Effects of media on bacterial growth and DON de-epoxidation activity....…...... 85

Figure3-6. Effects of temperature on microbial growth and DON de-epoxydation ………...87

Figure 3-7. Influence of pH on microbial growth and DON de-epoxydation capability…...... 89

Figure 3-8. Effects of sodium azide (SA) on bacterial DON de-epoxidation activity………..91

Figure 3-9. Determination of DON to deepoxy-DON transformation by bacterial culture filtrate and cell extract…………………………………..…………………………………….92

Figure 3-10. Determination of microbial de-epoxidation and de-acetylation of a type-A

trichothecene mycotoxin (HT-2 toxin) by LC-UV-MS……...………………………………94

Supplemental Figure S3-1. Flow chart depicted the procedures of enrichment and isolation of DON de-epoxidation bacterial strain ADS47……………………….…….……………...101

Figure 4-1. Schematic overview of the construction and functional screening of Fosmid

DNA library..………………………………………………………………………………..105

Figure 4-2. Genetic map of copy controlled Fosmid vector (pCC1FOS)………………...…108

Figure 4-3. Preparation of 40 kb size insert DNA from a DON de-epoxidizing bacterium...113

Figure 4-4. Production of genomic DNA library in E. coli EPI300-T1 host strain…………114

Figure 4-5. Confirmation of the Fosmid DNA library construction in E. coli ….………… 116

Figure 5-1. Co-linearity between C. freundii strain ADS47 genome and C. koseri …..……129

Figure 5-2. Circular genome map of C. freundii strain ADS47 ...…………………………..133

Figure 5-3. Size distribution of the predicted genes identified in C. freundii strain

ADS47……… ...... 135

Figure 5-4. Functional classification of the predicted coding sequences (CDS) of

C. freundii strain ADS47 .…………………………..…………………………………...... 137

Figure 5-5. Phylogeny of the trichothecene detoxifying C. freundii strain ADS47 ...... 139

Figure 5-6. Comparisons of coding sequences between C. freundii strain ADS47 and five Citrobacter species……….……………………………………………………………143

Figure 5-7. Determination of the locus enterocyte effacement (LEE) operons in

Citrobacter species by comparing C. rodentium …………….……………………………..151

Figure 5-8. Locations of unique genes of C. freundii strain ADS467..………..……………157

Figure 5-9. Unique genomic regions and associated genes in strain ADS47 compared to five Citrobacter species…………..………………………………………………………159

xii

LIST OF ABBREVIATIONS

APCI= atmospheric pressure chemical ionization DAS= diacetoxyscirpenol

A/E= attaching and effacing dH2O= distilled water

A= aerobic conditions DOGT1= UDP-glucosyltransferase

ABC= ATP-binding cassette DON= deoxynivalenol

ACT= artemis comparative tools dsRNA= double stranded RNA

AN= anaerobic conditions dE= de-epoxy

3-ADON= 3-acetyl-deoxynivalenol dE-DON= de-epoxy DON

15-ADON= 15-acetyl-deoxynivalenol E. coli= Escherichia coli

ATA= alimentary toxic aleukia ER=endoplasmic reticulum bp= base pair FAO= Food & Agriculture Organization

BGI= Beijing genomics institute FM= full medium

BLAST= basic local alignment search tool FHB= Fusarium head blight

BRIG= blast ring image generator FUS= fusarenonX

Bw= body weight gDNA= genomic DNA

C= carbon GC= guanine cytosine

CDS= coding sequence GER= Gebberella ear rot

CFIA= Canadian food inspection agency GPI= glycosylphosphatidylinositol

CHEF= clamped homogeneous electric fields h= hour

CMB= corn meal broth Hck= hematopoietic cell kinase

CVTree= composition vector tree Kb= kilobase

xiii

Kg= kilogram m/z= mass ion ratio

LC-UV-MS= liquid chromatography-ultraviolet- N= nitrogen mass spectrometry Na-Cl= sodium–clorine

LEE= locus of enterocyte effacement NADH= (reduced) nicotinamide adenine

LD50= lethal dose of 50% animal dinucleotide

LB= Luria Bertani NAD(P)= nicotinamide adenine

LMP= low melting point dinucleotide phosphate

MAPK= mitogen-activated protein kinase NCBI= national center for biotechnology

µg= microgram information mg= milligram NEO= neosolaniol mL= milliliter ng= nanogram

µL= microliter NIV= nivalenol

µM= micromole OD600nm= optical density at 600 min= minute nanometer wavelength

Mb= megabases OMAFRA= Ontario Ministry of

MCS=multiple cloning site Agriculture and Rural Affairs

MSA= MSB+ agar Paa= phenylacetic acid

MSB= mineral salts broth PCR= polymerase chain reaction

MM= minimal medium PDB= potato dextrose broth

MSC= mineral salts + 0.3% Bacto Peptone ppm= parts per million

MS-DON= mineral salts + 200 µg/mL DON PFGE= pulsed field gel electrophoresis mRNA= messenger RNA PKR= protein kinase

MSY= mineral salts + yeast extract

xiv

PNDO= pyridine nucleotide-disulfide TCA= tricarboxylic acid cycle oxidoreductase TSB= tryptic soy broth

QTL= quantitative trait locus TF= transcription factor

RAST= rapid annotation subsystem technology tRNA= transfer RNA

RDP= ribosomal database project TIC= total ion count

Rpl3= 60S ribosomal protein L3 TRAP= tripartite ATP-independent rRNA= ribosomal RNA periplasmic s= second T-RFLP=terminal restriction fragment

SA= sodium azide length polymorphism

SDR= short chain dehydrogenase/reductase TRI101= trichothecene 3-O-

SE= soil extract acetyltransferase

SIM= selected ion monitoring YEB= yeast extract broth

SOAP= short oligonucleotide analysis package VER= verrucarol sp.= species VKOR= vitamin K epoxide reductase sRNA= small RNA w/v= weight by volume

TC= trichothecene

CHAPTER 1

1.0. GENERAL INTRODUCTION AND LITERATURE REVIEW

1.1. Background

According to the Food and Agriculture Organization (FAO) of the United Nations, more than 25% of the world’s food crops are contaminated with mycotoxins every year, causing serious adverse effects on livestock industry, crop production and international trade

(Boutrif and Canet 1998, Desjardins 2006, Binder et al. 2007, Surai et al. 2010). The negative effects of mycotoxins on the consumers are diverse, including organ toxicity, mutagenicity, carcinogenicity, neurotoxicity, teratogenicity, modulation of the immune system, reproductive system effects, and growth retardation. Mycotoxins are toxic secondary metabolites produced by various fungal species during crop colonization in field and/or post harvest conditions

(FAO 1991, Barrett 2000). Aspergillus, Penicillum and Fusarium are the predominant fungal genera that accumulate economically important mycotoxins, such as aflatoxins, zearalenone, trichothecenes, fumonisins, ochratoxins and ergot alkaloids in agricultural commodities.

Humans and animals are principally exposed to mycotoxins through consumption of contaminated food/feed and animal products, such as milk, meat and eggs (Fink-Gremmels

1989, Hollinger and Ekperigin 1999).

Trichothecenes (TCs) produced by a variety of toxigenic Fusarium species are probably the most food and feed contaminating mycotoxins in the temperate regions of

America, Europe and Asia (Creppy 2002, Yazar and Omurtag 2008). The fungal pathogen accumulates various type-A (e.g., T-2 toxin, HT-2 toxin) and type-B trichothecenes (e.g.,

1 deoxynivalenol, nivalenol and their derivatives) in cereals. Among them type-B trichothecenes, particularly deoxynivalenol (DON) and/or its derivatives appear to be the most commonly found mycotoxins in grain and grain products worldwide, impairing human health and productivity of farm animals (Placinta et al. 1999, Sherif et al. 2009). A number of human and animal disease outbreaks, associated with TC contaminated grain food and feed consumption, have been reported in Asia (Japan, India and central Asia), Europe, New

Zealand and United States of America (USA) (Pestka and Smolinski 2005, Desjardins 2006).

A severe outbreak of Alimentary Toxic Aleukia (ATA) mycotoxicosis killed 100,000 people in the Orenburg region of Russia (Eriksen 2003). In 1996, losses of over $100 million were estimated in Ontario because of costs related to managing DON (Schaafsma 2002).

Historically, T-2 toxin, DON and nivalenol (NIV) have likely been used as a biological warfare in Laos, Kampuchea and Afghanistan, and there is concern because of the potential use of these compounds for bioterrorism (Wannemacher et al. 2000, Tucker 2001).

Despite the extensive efforts that have been made to prevent TC contamination, with the current varieties and production practices, its presence in agricultural commodities is likely unavoidable. There are no crop cultivars currently available that are completely resistant to Fusarium head blight/Gibberella ear rot (FHB/GER) diseases. Chemical or physical methods cannot safely remove TC mycotoxins from grain completely, or they are not economically viable (Varga and Tóth 2005). Therefore, an alternative strategy to address the issue is of great importance (Awad et al. 2010, Karlovsky 2011).

1.2. Plant pathogenesis and trichothecene production

Cereals infested by the ascomyceteous fungi may cause Gibberella ear rot (GER) in maize (Zea mays L.) and Fusarium head blight (FHB; also known as ‘scab’) of small grain cereals [wheat (Triticum aestivum L.), barley (Hordeum vulgare L.)] and other small grain cereals (Goswami and Kistler 2004). As many as 17 fungal species have been reported to be associated with FHB/GER, which include F. graminearum Schwabe [telemorph: Gibberella zeae (Schwein.) Petch], F. culmorum (W.G. Smith) Sacc., F. sporotrichioides Scherb., F. poae (Peck) Wollenw., Microdochium nivale (Fr.) Samuels & I.C. Hallett and F. avenaceum

(Fr.) Sacc. (Osborne and Stein 2007, Audenaert et al. 2009). The two species, F. graminearum and F. culmorum, are the principal species causing crop damage. Among them,

F. graminearum is the widely adapted and most damaging species across the world including

North America (Dyer et al. 2006). F. culmorum tend to be the main species in the cooler parts of North, Central and Western Europe (Wagacha and Muthomi 2007).

The fungi overwinter mainly on crop residues as saprophytes, where they produce asexual (micro- and macro-conidia) and/or sexual (ascospores) fruiting structures, which act as the initial source of inoculums in the subsequent crop season (Sutton 1982, Wagacha and

Muthomi 2007). Ascospores are primarily dispersed by air (Trail et al. 2002) and conidia move by rain-splash (Paul et al. 2004). After they land on maize silks and florets of small grain cereals infection occurs through natural openings (e.g., stomata), or occasionally, by direct penetration. The pathogens can invade maize kernels through wounds caused in the husks by insects or birds. Infection may occur throughout the crop growing season, but disease incidence and severity is facilitated by warm temperatures (>20°C), high humidity

(>80%) and frequent precipitation that coincides with anthesis (Osborne and Stein 2007,

Wagacha and Muthomi 2007). The Fusarium pathogen initially infects as biotroph, which becomes a necrotroph as the disease progresses. After penetration, Fusaria produce different extracellular cell wall degrading enzymes that trigger infection and hyphal progression towards the developing ear/kernel where further colonization occurs and development of

FHB/GER diseases (Voigt et al. 2005, Miller et al. 2007, Kikot et al. 2009). FHB disease is characterized by a bleached appearance around the middle of the head. Severely infected grains have a pinkish discoloration and can be shriveled. In maize, the disease produces a reddish–brown discoloration of the kernels and progresses from the tip to the bottom of affected ears.

The fungi associated with FHB and GER contaminate cereals by accumulating various host-nonspecific TC mycotoxins (Brennen et al. 2007). TCs have typically been associated with cereals grown in temperate climatic conditions, including Canada (Doohan et al. 2003).

DON production by F. graminearum and F. culmorum is maximal at warm temperatures (22-

28˚C), high grain moisture contents (>17%) and humid conditions (>70% RH) (Martins and

Martins 2002, Hope et al. 2005, Llorens et al. 2006). Grain can be contaminated with TCs in the field as well as in storage, if conditions for fungal growth are optimal and grain contains high moisture (>17%) (Larsen et al. 2004, Birzele et al. 2002).

Populations of FHB/GER causing pathogens are dynamic. Genetic variability between and within Fusarium species appears to have occurred to allow for their adaptation to new agro-climates (Desjardins 2006, Fernando et al. 2011). Three strain specific Fusarium chemotypes, NIV, 3-ADON (3-acetyl-deoxynivalenol) and 15-ADON (15- acetyl- deoxynivalenol) have been identified, which correlate with the geographical distributions

(Starkey et al. 2007, Zhang et al. 2007, Audenaert et al. 2009). The NIV chemotype, mainly

4 produced by F. culmorum, accumulates NIV and 4-NIV (Fusarinon X), is predominant in

Asia, Africa and Europe (Lee et al. 2001). The 3-ADON chemotype, which synthesizes DON and 3-ADON, is the dominant strain in Europe/Far East (Larsen et al. 2004). In North

America, 15-ADON is the main chemotype which produces DON and 15-ADON (Goswami and Kistler 2004). Significant Fusarium population shifting was observed in the past and in recent years (Kant et al. 2011). A dynamic change from F. culmorum to F. gramninearum, and DON to NIV producing chemotyes occurred in North West European countries. The 3-

ADON chemotype, which is uncommon in North America, was found in North Dakota and

Minnesota (Goswami and Kistler 2004). F. pseudograminearum, which historically occurs in

Australia, Africa and North West Pacific America, has been found in Western Canada (Larsen et al. 2004, Kant et al. 2011). In Manitoba, a replacement of 15-ADON chemotype by the more phytoxic 3-ADON type was observed in recent years (Fernando et al. 2011). All together, the FHB/GER associated pathogen complex is highly dynamic, which not only reduces crop yield, but also decreases grain quality by accumulating various mycotoxins.

1.3. Biosynthesis of food contaminating TC

TCs are fungal secondary metabolites that were initially identified in Trichotheceium roseum. Thus, the name of this class of compounds follows the name of the genus,

Trichotheceium (Freeman and Morrison 1949). There have been 180 TCs identified in various fungal species belonging to the genera Fusarium, Myrothecium, Stachybotrys, Trichoderma and Thrichothecium (Eriksen 2003, Eriksen et al. 2004). The FHB/GER associated TCs are mainly synthesized by different Fusaria, and are characterized as tricyclic sesquiterpenoids with a double bond at position C9-10 and a C12-13 epoxide ring.

(a) (b)

(c)

Figure 1-1. Structures of trichothecene mycotoxins produced by Fusaria. (a) General structure of trichothecenes. R1 = H, OH, OiValeryl, keto; R2 = H, OH; R3= OH, OAc; R4 =

H, OH, OAc; R5= OH, OAc. (b) Type-A trichothecene mycotoxin. (c) Type-B trichothecene mycotoxin, which contain a carbonyl functional group at C8 position (Young et al. 2007).

The fungal metabolites are small in size (molecular weight 250-550), heat stable and non-volatile molecules (Pestka 2007, Yazar and Omurtag 2008). Food contaminating TCs are divided into two groups, type-A and type-B (Figure 1-1b and 1c). Type-A is characterized by the lack of a ketone function at C8 position, and include T-2 toxin, HT-2 toxin, diacetoxyscirpenol (DAS) and neosolaniol (NEO), which are mainly produced by F. sporotrichioides and F. poae (Creppy 2002, Quarta et al. 2006). The most frequently found

TC belongs to type-B, which contains carbonyl functions at C8 positions, DON, NIV and their acetylated derivatives. The two types are further differentiated by the presence or absence of functional hydroxyl (-OH) and acetyl (CH3-CO-) groups at C-3, C-4, C-7, C-8, and

C-15 positions (Table 1-1).

The syntheses of various TCs are species/strain specific and are genetically controlled, but influenced by environmental conditions (Desjardins 2006, Paterson and Lima 2010). Until now, 15 Fusaria catalytic genes and regulators involved in TC productions have been identified (Merhej et al. 2011). These biosynthetic genes are located at three loci on different chromosomes. The core trichothecene (TRI) gene cluster is comprised of 12 genes localized in a 25-kb genomic region (Kimura et al. 2001). The remaining three genes are encoded by the TRI1-TRI16 and TRI101 loci (Gale et al. 2005, Alexander et al. 2009). Fusaria produce

TC following a number of oxygenation, isomerization, acetylation, hydroxylation and esterification reaction steps (Desjardins 2006, Desjardins and Proctor 2007, Kimura et al.

2007, Merhej et al. 2011). In brief, farnesyl pyrophosphate is initially cyclized into trichodiene by TRI5, which codes for trichodiene synthase. In the next four steps, isotrichotriol is produced by TRI4 (a multifunctional cytochrome P450) via C-2 hydroxylation and 12,13 epoxidation, followed by two hydroxylation reactions.

Table 1-1. Fusaria produced type-A and type-B trichothecene mycotoxins. Molecular weight

(MW) and acetyl (-R), hydroxyl (-OH) and ketone (=O) functions at different carbon (C) position of the tricyclic rings. A-types: T-2 toxin, HT-2 toxin, DAS (diacetoxyscirpenol),

NEO (neosolaniol), VER (verrucarol). B-type: DON (deoxynivalenol), 3-ADON (3-acetyl- deoxynivalenol), 15-ADON (15-acetyl-deoxynivalenol), NIV (nivalenol) and FusX

(fusarenon X) (Young et al. 2007).

Trichothecenes MW R1/C3 R2/C4 R4/C7 R3/C15 R5/C8

T-2 toxin 466 OH OCOCH3 H OCOCH3 OCOCH2CH(CH3)2

HT-2 toxin 424 OH OH H OCOCH3 OCOCH2CH(CH3)2

Type-A DAS 366 OH OCOCH3 H OCOCH3 H

NEO 382 OH OCOCH3 OH OCOCH3 H

VER 266 H OH H OH H

DON 296 OH H OH OH =O

3-ADON 338 OCOCH3 H OH OH =O

Type-B 15-ADON 338 OH H OH OCOCH3 =O

NIV 312 OH OH OH OH =O

FusX 354 OH OCOCH3 OH OH =O

The next two non-enzymatic reactions results in isotrichodemol. Eventually, the trichothecene skeleton is produced via the formation of a C-O bond between the C-2 oxygen and C-11. Subsequent acetylation at C-3 by TRI101, hydroxylation at C-15 by TRI1 and catalytic activity of TRI3 leads to calonectrin production. The pathway described above is common for both type-A (e.g., T-2 toxin, HT-2 toxin) and type-B (e.g., DON, NIV and their acetylated derivatives) TCs.

The following steps are species/strain specific. In DON producing F. graminearum strains, calonectrin is converted to either 3-ADON or 15-ADON followed by DON production by the catabolic activities of TRI1 and TRI8. The biosynthesis of 3-ADON or 15-ADON is controlled by variation in the esterase coding sequences of TRI8. F. graminearum produces

NIV from calonectrin via synthesis of 4-NIV, followed by NIV by TRI1 and TRI8 gene actions. NIV can be produced either from DON by hydroxylation at C-4, or via transformation of calonectrin to 3,15-diacetoxyscirpenol (3,15-DAS), followed by addition of a ketone at C-8 position by the activity of the TRI7 and TRI13 genes. 3,15-DAS is the principal substrate for synthesis of type-A TCs by F. sporotrichioides, the main type-A producer. The biosynthetic pathway of T-2 toxin involves C-4 acetylation of 3,15-DAS by the

TRI7 gene product and TRI8 triggered by TRI1. 3,15-DAS is also the substrate for HT-2 toxin synthesis. In the absence of TRI7 expression, HT-2 toxin is produced through hydroxylation of 3,15-DAS and isovalerate addition to C-8 position (Foroud and Eudes 2009).

1.4. Occurrence of TC mycotoxins in food and feed

TC producing Fusaria are adapted to variable climatic conditions and agricultural ecosystems. As a consequence, this class of mycotoxins, particularly DON, is frequently detected in foods and feeds including breakfast cereals, bakery products, snack foods, beer and animal feed derived from cereals in most part of the world (Pittet 2001, Desjardins 2006,

Lancova et al. 2008). A two-year survey during 2003-2005, conducted by Romer Lab Inc. revealed unacceptable levels of DON in animal feeds/feed raw materials (such as maize, wheat, wheat bran) in Asia and Oceania (10.62 µg/g), central Europe (8.02 µg/g), Mid-West

USA (5.5 µg/g), northern Europe (5.51 µg/g) and southern Europe (3.03 µg/g) (Binder et al.

2007) (Table 1-2).

Quite high levels of DON contamination (19 µg/g) were detected in north Asian and

Chinese feed/feed materials. Recently, five-year on-farm studies in Switzerland showed that

74% of the wheat grain and straw samples were contaminated with high levels of DON

(average 5 µg/g). The amount is higher than the European maximum limit of 1.25 µg/g

(Vogelgsang et al. 2011).

The occurrences of TCs in Canadian agricultural commodities are indicated in Table

1-3. Seventeen surveys conducted during the 1980s and 1990s showed up to 100% DON incidence in Ontario wheat and maize samples with an 18% average incidence (Scott 1997). A three-year survey recently (2006-08) conducted by the Ontario Ministry of Agriculture and

Rural Affairs (OMAFRA) detected high levels of DON in spring (maximum 26 µg/g) and winter wheat (maximum 29 µg/g) samples (Ana Matu, personal communication). The DON levels were much higher than Health Canada’s maximum limit of 2 µg/g for uncleaned soft wheat for staple foods (Table 1-4).

Table 1-2. Worldwide occurrence of DON in animal feed. Feed and feed raw material samples were collected from animal farms/feed production sites during 2003-2005. The survey was conducted by Romer Labs Inc. (http://www.romerlabs.com).

Country/ Sample Incidence Maximum Mean DON Survey Source region type (%) DON content content period (µg/g) (µg/g) North 0.71 18.99 0.92 Asia/China South East 0.14 1.52 0.16 Asia South Asia 0.70 10.62 0.49 Binder et Oceania Animal 2003-05 al. 2007 Australia feed 0.41 1.86 0.30 Northern 0.70 5.51 0.59 Europe Central 0.66 8.02 0.57 Europe Southern 0.52 3.03 0.30 Europe

Table 1-3. Occurrence of DON in Canadian grains, grain foods and beers. Infant cereal foods: oat, barley, soy and multi-grain based foods.

Province Sample Incidence Maximum Multiple Survey Sources type (%) DON mycotoxins period contents samples (%) (µg/g) Ontario Spring 70 26.0 Yes 2006-08 OMAFRA1 Wheat Ontario Winter 49 29.0 Yes 2006-08 OMAFRA wheat Ontario Maize 78 14.0 Yes 2009-10 Limay-Rios and Schaafsma 2010 Ontario Maize - 100.0 Yes 2011 A&L Biologicals Canada Breakfast 46 0.94 43.0 1999-01 Roscoe et al. cereals 2008 Canada Animal 73 42.5 Yes 1990-10 Zeibdawi et feeds al. 2011 Canada Beers 58 0.052 5.0 1990s Scott 1997 Canada Infant 63 0.98 Yes 1997-99 Lombaert et Cereals al. 2003

1 Ontario Ministry of Agriculture and Rural Affairs

2 µg/mL

Table 1-4. Legislative maximum tolerated levels of trichothecene mycotoxins in Canada according to Canadian Food Inspection Fact Sheet, April 23, 2009. DON- deoxynivalenol,

DAS- diacetoxyscirpenol.

Products DON T-2 toxin HT-2 toxin DAS (µg/g) (µg/g) (µg/g) (µg/g) Uncleaned soft wheat for human 2 - - - consumption Cattle/poultry diets 5 <1 0.1 <1 Swine/dairy cattle feeds 1 <1 0.025 <2

High DON incidence (78%) was observed in maize samples collected from southwestern Ontario fields with moderate levels (<1 µg/g) of contamination during 2009 and

2010 (Limay-Rios and Schaafsma 2010). Some samples had high DON levels (14 µg/g). In the 2011 crop season, extremely high levels of DON (>100 µg/g) were found in some southern Ontario maize samples (A&L Biologicals, London, ON, Personal communication).

In a Canadian Food Inspection Agency (CFIA) 20 year survey (1990-2010) 73% of animal feed samples collected from feed mills, farms and retail stores across Canada were contaminated with DON and for more than 25% of the samples the mycotoxin levels were above the maximum limit (≥ 1 µg/g) (Zeibdawi et al. 2011). The highest level of DON (15

µg/g) detected in the 2006-2007 samples were much lower than the levels found in the 1999-

2000 samples, which contained a maximum of 42.5 µg/g DON. Importantly, this study reported a significant number of samples contaminated with multiple mycotoxins, DON, NIV,

3-ADON, 15-ADON, T-2 toxin, HT-2 toxin, DAS and NEO.

Earlier studies in the 1990s detected high levels of DON content in Eastern Canadian cereals, maize (maximum 17.5 µg/g), wheat (maximum 9.16 µg/g), barley (maximum 9.11

µg/g) and oats (maximum 1.2 µg/g) (Campbell et al. 2002). The study identified a significant number of samples (15-36.7%) that contained multiple mycotoxins and up to 33% of the wheat samples exceeded 1 µg/g DON. Low amounts but high incidences of TCs were found in cereal-based foods collected from Canadian retail stores. Sixty three percent of infant cereal foods, 46% of breakfast cereals and 58% of beers were contaminated with one or multiple mycotoxins (Scott 1997, Lombaert et al. 2003, Roscoe et al. 2008). The maximum amount of DON was found in infant foods (0.98 µg/g) followed by breakfast cereals (0.94

µg/g) and beers (0.05 µg/mL).

1.5. Significance of deoxynivalenol and trichothecenes

For most cereals, including wheat, maize, rice, barley, oats and rye the risk of contamination with TCs resulted in grain quality reduction and FHB/GER disease progression

(Bai et al. 2002, Jansen et al. 2005). The impacts of TC contamination in agricultural commodities are diverse, including: reduced livestock production, induced human illnesses and enhanced FHB disease severity, which may result in the reduction of crop yield. Exposure of farm animals (particularly pigs) to DON is a continuing threat to the livestock industries in

Canada, USA and Europe (D’Mello et al. 1999). Ingestion of TC-contaminated feeds generally causes reduction of feed intake, feed refusal, poor feed conversion, retardation of weight gain, increased disease susceptibility and reduced reproductive capacities (Binder et al.

2007). High doses of DON result in acute toxic effects in pigs (e.g., emesis), while low doses

(≥1 µg/mL) cause chronic effects (e.g., growth retardation in pigs) (Rotter et al. 1996,

Accensi et al. 2006). Reduction of growth and reproduction ability was observed in experimental rats exposured to much higher doses of TC (Sprando et al. 2005, Collins et al.

2006). Several animal disease outbreaks that occurred in different continents have been attributed to consumption of feed contaminated with DON, NIV and DAS (Desjardins 2006).

The consequences of human exposure to TCs include: liver damage, gastric/intestinal lesions and central nervous system toxicity, resulting in anorexia, nausea, hypotension/shock, vomiting, diarrhea and death (Pestka 2007). Human mycotoxicosis (illnesses and outbreaks) by consumption of type-A and type-B TC mycotoxins in Europe, New Zealand, Russia,

Japan, South America, India and Vietnam have been reported (Bhat et al. 1989, Yoshizawa

1993, Luo 1994, Desjardins 2006). Severe outbreaks of Alimentary Toxic Aleukia (ATA) associated with TC killed thousands of people in the Orenburg region of Russia during the

1930’s and 1940’s (Eriksen 2003). TCs (T-2, DON and NIV) were likely used for biological warfare in Laos, Kampuchea and Afghanistan and its potential application for bioterrorism is an additional concern associated with TC (Tucker 2001, Bennett and Klich 2003).

TCs are phytotoxic and may contribute to FHB disease severity in cereals (Proctor et al. 1995, Maier et al. 2006, Masuda et al. 2007). Studies demonstrated that DON caused wilting, chlorosis, inhibition of seedling growth and necrosis in cereals (Cossette and Miller

1995, Mitterbauer et al. 2004, Rocha et al. 2005). DON-mediated host cell damage and subsequent necrotrophic fungal spread has been shown to lead to FHB/GER disease development and higher kernel damage (Procter et al. 1995, Mitterbauer et al. 2004, Maier et al. 2005). This suggests that DON contributes to yield losses as well as to reductions in the market value of cereals (Osborne and Stein 2007). Severely infected grain may contain more than 60 µg/g DON, which is unacceptable to the milling and brewing industries (Okubara et al. 2002). Significant crop losses due to FHB disease epidemics and mycotoxin contamination have been reported in Ontario (Schaafsma 2002). FHB disease epidemics during 1998 to 2000 caused huge crop losses, resulting in farm failures in the northern Great Plains in the USA

(Windels 2000). In that period, FHB associated losses were estimated at US $ 2.7 billion in the region (Nganje et al. 2002).

Importantly, Fusaria infested grains are often co-contaminated with multiple mycotoxins, which further increase health risks as interactions between the mycotoxins may cause greater toxic effects than expected (Schollenberger et al. 2006). This implies that an apparently low level of contamination of a single mycotoxin can have severe effects on consumers because of synergistic effects among multiple co-occuring mycotoxins (Surai et al.

2010).

A number of studies demonstrated synergistic toxic effects among Fusarium mycotoxins on pigs, turkeys, lamb, chicks, as well as in in vitro Caco-2 cell and yeast bioassays (Table 1-5).

Table 1-5. In vitro and in vivo synergistic effects of different food contaminating mycotoxins.

DON-deoxynivalenol, DAS-diacetoxyscirpenol, FB1-fumonisins 1, NIV-nivalenol, ZEN- zearelenone. FA-fusaric acid, Afl-aflatoxin.

Mycotoxins source Interacting Experimental Effects Reference mycotoxins animal/cells Naturally contaminated DON / FA Pigs Inhibit growth Smith et al. cereals 1997 DON contaminated DON / FB1 Pigs Weight loss Harvey et al. wheat and FB1 1996 Culture materials T-2 / FB1 Turkey poults Weight loss/ oral Kubena et al. contained mycotoxins lesion 1997 Inoculated rice and DAS / Afl Lamb Weight reduction Harvey et al. pure mycotoxin 1995 Naturally DON DON / T-2 Chicks Reduced body Kubena et al. contaminated wheat weight significantly 1989 and pure T-2 toxin Pure mycotoxins DAS / Afl Chicks Reduced body Kubena et al. weight 1993 Pure mycotoxins T-2 /Afl Chicks Reduced weight/ Kubena et al. oral lesion 1990 Pure mycotoxins DON/ Human intestinal Increased DNA Kouadio et al. ZEN/FB1 Caco-2 cells fragmentations 2007 Pure mycotoxins DON / NIV Yeast bioassay Synergistic effects Madhyastha et al. 1994 Pure mycotoxins T-2 /HT-2 Yeast bioassay Synergistic effects Madhyastha et al. 1994

1.6. Toxic effects of deoxynivalnol and trichothecenes

As small molecules, TCs simply enter into cells and interact with different cell components, resulting in disruption of cellular functions and causing various toxic effects that include inhibition of protein synthesis and ribosomal fragmentation (Eudes et al. 2000,

Minervini et al. 2004, Pestka 2007). Simple absorption of DON into human intestinal Caco-2 cells and crossing human placenta barriers through passive diffusion has been reported

(Sergent et al. 2006, Nielsen et al. 2011).

TC may cause various toxicities in animals and plants including: apoptosis, immunotoxicity, chromatin condensation, disruption of cellular/mitochondrion membranes, disruption of gut tissues, inhibition of protein and ribosome synthesis, breakage of ribosome and cell damage. The toxicity of TC is influenced by structural variability (i.e. position, number and type of functional groups attached to the tricyclic nucleus and C-C double bond at

C9-10 position) (Desjardins 2006). Nevertheless, epoxide function at the C12-13 position is the key for TC cytotoxicity (Swanson et al. 1988, Eriksen et al. 2004).

Different in vitro and in vivo studies revealed that the method and period of exposure, amount of TC and types of cell/tissue and host species have significant effects on TC’s cytotoxicity (Scientific Committee for Food 2002, Rocha et al. 2005, Nielsen et al. 2009).

Administration of 10 mg/kg body weight (bw) NIV caused apoptosis in mice thymus, Peyer’s patch and spleen tissues (Poapolathep et al. 2002). DON and NIV at 8.87 mg/kg bw for 3 days a week for four weeks induced immunotoxicity in experimental mice, determined by estimating increased immunoglobulins (IgA) and decreased IgM levels (Gouze et al. 2007).

Exposure of mice to 1 mg/kg bw of DON induced skeletal malformations and reduced mating tendency, while significant resorption of embryos observed at the dose of >10 mg/kg bw

(Khera et al. 1982). At high acute doses (49/7.2 mg/kg bw), DON/NIV and type-A trichothecenes caused mice death because of intestinal haemorrhage, necrosis of bone marrow/lymphoid tissues and kidney/heart lesions (Yoshizawa and Morooka 1973,

Yoshizawa et al. 1994, Forsell et al. 1987, Ryu et al. 1988). In a separate study, 15-ADON showed more toxicity than DON when the mycotoxins were orally administered to mice

(Forsell et al. 1987).

Studies suggested that type-A TCs are more toxic in mammals than type-B, which means that functional variability at the C-8 position of TC has a significant role in toxicity.

The toxicity of various TC mycotoxins in animals can be ranked in the order of T-2 toxin>HT-2 toxin>NIV>DON≥3-ADON>15-ADON according to the results from human cell lines and mice model systems bioassays (Ueno et al. 1983, Pestka et al. 1987, Nielsen et al.

2009). Moreover, animal feeding assays demonstrated that monogastric animals are most sensitive to DON. The order of sensitivity of animals to DON is pigs>rodents>dogs>cats>poultry>ruminants (Pestka 2005). The reasons for the variations in

DON toxicity species might be due to differences in metabolism, absorption, distribution and elimination of the mycotoxin among animal species (Pestka 2007).

To my knowledge, the mechanism of TC uptake by plant cells has not been studied.

Nonetheless, cereals are exposed to TC by toxigenic Fusaria infection which may cause various toxicities, such as cell damage, wilting, chlorosis, necrosis, program cell death and inhibition of seedling growth in sensitive plants (e.g., wheat, maize and barley) (Cossette and

Miller 1995, Rocha et al. 2005, Boddu et al. 2007). As animals, TC inhibits protein synthesis in non-host A. thaliana (Nishiuchi et al. 2006). Immunogold assays showed that the cellular targets of DON in wheat and maize were cell membrane, ribosome, endoplasmic reticulum,

20 vacuoles and mitochondria. Modulation of cell membrane integrity and electrolyte losses were observed by direct DON interference in maize and wheat cell assays (Cossette and

Miller 1995, Mitterbauer et al. 2004). DON-mediated mitochondrion membrane alterations, disruption of host cells enzymatic activity, apoptosis and necrosis have been observed in cereals (Mitterbauer et al. 2004, Rocha et al. 2005). In contrast to animal, type-B is more toxic to plants than type-A. DON showed 8- and 13-fold phytotoxicity to wheat coleoptiles than T-

2 and HT-2 toxins, respectively (Eudes et al. 2000). By compilation of different studies, TC phytoxicity can be ordered as: DON≥3-ADON>T-2 toxin>DAS (Wakulinski 1989, McLean

1996, Eudes et al. 2000).

1.7. Cellular and molecular mechanisms of DON toxicity in animals and plants

The underlying cellular and molecular mechanisms of DON/TC toxicity are evolving.

Exposure to DON proceeds with the up and down regulation of numerous critical genes of eukaryotic organisms, which causes various negative effects on animal and plant organisms

(Figure 1-2).

One of the early events of DON toxicity occurs when the mycotoxin binds to the active ribosome, resulting in blocking protein translation through different routes, by damaging the ribosome (Shifrin and Anderson 1999), degrading or altering conformational changes of 28S-rRNA (Alisi et al. 2008, Li and Pestka 2008), and inducing expression of miRNAs. This results in down regulation of the ribosome associated genes (He et al. 2010a,

Pestka 2010). Both animal and plant cell assays demonstrated that DON bound to the peptidyl transferase (Rpl3) domain of the 60S subunit of ribosome and blocked protein chain elongation and termination at high DON concentrations, while peptide chain termination and ribosome breakage occurred at a low level of DON exposure (Zhou et al. 2003a).

Figure 1-2. Cellular and molecular mechanisms of deoxynivalenol toxicity in eukaryotic organisms. DON is simply absorbed by the cell and interacts with the cytoplasmic membrane

(CM), mitochondrion (MT) membrane and ribosome. Disruption of the CM and alteration of the MT membrane causes a loss of electrolytes and affects the function of the MT, respectively. DON binds to the active ribosomes, which leads to ribosome damage and induces micro RNAs (miRNA). Expression of miRNA impairs the synthesis of ribosome associated protein. The subsequent transduction and phosphorylation of the RNA-activated protein kinase (PKR) and hematopoietic cell kinase (Hck) activate the mitogen-activated protein kinase (MAPK) cascade. Phosphorylation of the MAPKs induces targeted transcription factors (TF), which results in ribotoxis stress response, upregulation of proinflammatory cytokines and dysregulation of neuroendocrine signaling (derived from

Pestka 2007 and 2010).

The consequence of DON interaction with ribosomes leads to cellular and molecular changes, which are complex and not fully understood. However, it is well-demonstrated that the activation of the mitogen-activated protein kinase (MAPK) and subsequent molecular actions play key roles in the cytotoxicity of TC. Upon binding to the ribosome, DON rapidly induced interactions between the protein kinases and the ribosomal 60S subunit as observed in mononuclear phagocytes (Bae and Pestka 2008). Ribosomal binding to DON activated the ribosome associated kinases, for instances, protein kinase R (PKR), hematopoietic cell kinase

(Hck) and P38 mitogen-activated protein kinase. The protein kinases are induced by double- stranded RNA (dsRNA) or possibly by the damaged ribosomal peptides (Iordanov et al. 1997,

Shifrin and Anderson 1999, Zhou et al. 2003b, Pestka 2007). Subsequent phosphorylation of kinases (e.g., PKR/Hck/P38) activates the MAPK cascade resulting in ribotoxis stress response, upregulation of proinflammatory cytokines and dysregulation of neuroendocrine signaling (Pestka and Smolinki 2005, Pestka 2010). Studies suggested that high doses of

DON induces ribotoxis stress response (apoptosis) preceded by immunosuppression and disruption of intestinal permeability/tissue functions (Yang et al. 2000, Pestka 2008). Further study observed stabilization of the mRNA and expression of transcription factors for critical genes, when cells were exposed to a low concentration of DON (Chung et al. 2003).

The MAPK pathway proceeds with upregulation of a broad range of proinflammatory cytokine genes and dysregulation of neuroendocrine signaling within the central and enteric nervous systems. Modulation of neuroendocrine signaling by DON increased the level of serotonin, resulted in vomiting, reduced food intake and growth retardation in experimental animals/cell lines (Prelusky and Trenholm 1993, Li 2007). Upregulation of proinflammatory cytokines and chemokines genes was determined in mice liver, spleen, kidney and lung

24 following in 1 h of exposure to DON. A subsequent shock-like effect was observed in mice because of a “cytokine storm” (Pestka 2008). Induction of proinflammatory cytokines also led to various cellular effects, on the central nervous system, interference of the innate immune system, and growth hormone signaling. These contribute to anorexia, growth retardation, aberrant immune response, and tissue dysfunction as demonstrated in experimental animal systems (Pestka 2010).

DON-mediated growth retardation in mice via suppression of growth hormones

(insulin-like growth factor 1 and insulin-like growth factor acid-labile subunit) has been reported by Amuzie et al. (2009). Disruption of gut tissues by DON/NIV and subsequent impairment of intestinal transporter activity was demonstrated in a human epithelial cell model (Marseca et al. 2002). The study proposed that alteration of gut tissue permeability may cause gastroenteritis induction and growth retardation in livestock animals. Growth retardation of livestock may also occur via dysregulation of neuroendocrine signaling- mediated induction of emesis and anorexia (Pestka et al. 2010).

The molecular mechanism of TC toxicity in plant systems has not been well documented. A recent study showed the upregulation of numerous defense related genes in wheat upon exposure to DON (Desmond et al. 2008). The study further suggested that DON- induced program cell death in wheat occurred through a hydrogen peroxide signaling pathway. This mechanism had not been observed in mammals.

1.8. Strategies to control FHB and trichothecenes

During the past decades, many pre- and post-harvest strategies have been applied to protect cereal crops from Fusarium infestation and to eliminate TC from contaminated agricultural commodities (FAO 2001, Munkvold 2003, Varga and Toth 2005, Kabak and

Dobson 2009). Some of the above crop protection techniques have been patented (He and

Zhou 2010). In spite of those efforts, TC contamination of cereals appears to be unavoidable using the currently available methods.

Pre-harvest control strategies involving cultural practices (e.g., crop rotation, tillage, early planting) and fungicide applications, partially reduced Fusarium infection and DON contamination in cereals (Dill-Macky and Jones 2000, Schaafsma et al. 2001). However, repeated application of fungicides may lead to the evolution of Fusarium strains resistant to fungicides with greater mycotoxin production ability (D’Mello et al. 1998). At post-harvest, reduction of grain moisture by artificial drying, storage at low temperatures (1-4°C) and proper ventilation of the storage room have shown to be effective in inhibiting fungal growth and mycotoxin production in infected grain (Midwest Plan Service 1987, Wilcke and Morey

1995). This technology has become less applicable, particularly in developing countries, because of the frequent observation required to maintain proper storage conditions (Hell et al.

2000).

Removal of Fusarium infected grains by physical methods (e.g., mechanical separation, density segregation, colour sorting etc.) was unable to eliminate TC completely

(Abbas et al. 1985, He et al. 2010b). This is partly because FHB symptomless kernels may contain a considerable amount of DON, which are not eliminated by physical separation procedures (Seitz and Bechtel 1985). Several studies showed that milling and food processing

26 steps were unable to completely remove DON from grain-based foods (Jackson and

Bullerman 1999, Placinta et al. 1999, Kushiro 2008). Detoxification of TC by chemical treatments (e.g., ozone, ammonia, chlorine, hydrogen peroxide, and sodium metabisulfite) has been demonstrated (Young et al. 1986, Young et al. 1987, Young et al. 2006). The problem with these techniques is that the application of such chemicals is not economically viable and it also has a negative impact on grain quality (Karlovsky 2011). Although the technique has been popular and showed effectiveness to certain mycotoxins, utilization of absorbents to inhibit mycotoxin absorption by the livestock gastrointestinal tract appeared to be ineffective for TC (Huwig et al. 2001).

The introduction of Fusarium resistant cultivars would be the best strategy to address the FHB and associated mycotoxin problems in cereals. However, breeding for cultivars resistant to FHB/GER is difficult and time consuming as natural resistance against Fusarium is quantitative in nature. FHB resistant quantitative trait loci (QTLs) have been identified in several cereal genotypes (e.g., Sumai 3, Ning 7840, Bussard) (Bernando et al. 2007, Golkari et al. 2009, Häberle et al. 2009, Yang et al. 2010). The 3B, 5A and 6B QTLs for type-I

(resistant to initial infection) and type-II (resistant to fungal spread in the kernel) resistances have been identified on different wheat chromosomes (Buerstmayr et al. 2009). Type-I

(Qfhs.ifa-5A) and type-II (Fhb1/Fhb2-3B) QTLs from Sumai 3 are being used in many breeding programs and pyramiding of the QTLs in spring wheat showed additional effects.

Nonetheless, the key genes involved in FHB resistance in cereals are yet to be identified. This makes it difficult to integrate the resistance genes into elite cultivars.

Cultivar development by direct incorporation of TC/DON resistance genes may be an alternative way of protecting grains from these mycotoxins contaminations. Many genes from distantly related species have been integrated into crop cultivars using a transgenic approach with great economic benefits (Brookes and Barfoot 2006). DON transforming trichothecene

3-O-acetyltransferase (TRI101, from Fusarium sporotrichioides) and UDP- glucosyltransferase (from Arabidopsis thaliana) genes expressed in cereal cultivars showed some DON/DAS detoxifying ability under in vitro and controlled conditions (Kimura et al.

1998, Okubara et al. 2002, Poppenberger et al. 2003, Ohsato et al. 2007). By contrast, no reduction of mycotoxins in either of the above cases was observed under field conditions

(Manoharan et al. 2006). Importantly, the transformed acetylated and glycosylated compounds could revert to the original substances or convert to even more toxic products

(Gareis et al 1990, Alexander 2008).

1.9. Isolation of DON/TC de-epoxidating micro-organisms and genes

An enzyme with the capability of de-epoxidizing DON/TC appears to be a potential source for completely removing TC from contaminated cereals (Karlovsky 2011). However, such trichothecene detoxifying genes do not exist in cereals and other plants (Boutigny et al.

2008). Because of this, several research groups have attempted to obtain microorganisms that encode DON/TC de-epoxidation genes. Mixed microbial culture capable of de-epoxydizing trichothecenes have been isolated from chicken gut (He et al. 1992), cow ruminal fluid (King et al. 1984; Swanson et al. 1987; Fuchs et al. 2002), pig gut (Kollarczik et al. 1994), and digesta of chicken (Young et al. 2007), rats (Yoshizawa et al. 1983), sheep (Westlake et al.

1989) and fish (Guan et al. 2009). All of the microbial cultures cited above showed de-

28 epoxydation capabilities only under anaerobic conditions, realtively high temperatures and at low/high pH, except microbes from fish intestine that showed efficient de-epoxidation at low temperatures and wide range of temperatures (Guan et al. 2009). Such enzymes active under anaerobic conditions, high temperatues and low pH may have potential applications as feed additives, and the eubacterial strain BBSH 797 is an animal feed additive product with the brand name Mycofix® by BIOMIN (He et al. 2010b). Biotechnological applications of anaerobically-active enzymes from the above microbial sources seem to be inappropriate under the aerobic and moderate conditions at which cereals plants are grown.

Most of these efforts have failed to isolate single DON/TC degrading microbes effective under aerobic conditions because of inappropriate microbial growth conditions under laboratory conditions. Earlier attempts were made to obtain TC de-epoxidizing aerobic bacteria, but the culture did not retain DON/NIV de-epoxidation activity after reculturing

(Ueno 1983, He et al. 1992). In addition, the co-metabolic nature of the microbial TC de- epoxidation reaction has limited the use of mycotoxins as the sole nutrient source for selective isolation of the target microorganisms. Several DON/TC de-epoxidation microbial strains, including Bacillus strain LS100 and Eubacterial strain BBSH 797 have been isolated from chicken gut and rumen fluids through multi-step enrichment procedures (Fuchs et al. 2002,

Yu et al. 2010). A DON to 3-keto DON transforming bactial strain was islated from soil by prolonged incubation with high concentration DON (Shima et al. 1997).

Microbial genes/enzymes with TC de-epoxidation activity are not known. Based on the microbial TC de-epoxidation activity of a mixed microbial culture, it has been proposed that ceratin reductases or oxidoreductases may catalyze such a biochemical reaction (Islam et al. 2012). With the advent of molecular techniques, many novel genes with

29 industrial/agricultural value have been cloned by reverse genetics (genomic DNA library preparation and functional screening of the library clones), forward genetics (transposon mutagenesis) and comparative genomics methods (Frey et al. 1998, Koonin et al. 2001,

Suenaga et al. 2007, Chung et al. 2008, Ansorge 2009). The tremendous advancement of next generation sequencing technology and bioinformatics tools may facilitate the identification of the microbial genes responsible for DON/TC detoxification (Ansorge 2009). Isolation of TC de-epoxidation genes may play a major role in producing safe foods and feeds derived from grains.

1.10. The bacterial species Citrobacter freundii

Citrobacter freundii is a bacterial species belonging to the family Enterobacteriace. It is a facultatively anaerobic and Gram-negative bacillus (Wang et al. 2000). The bacterium survives almost everywhere, including: animal/human intestines, blood, water and soils

(Wang et al. 2000, Whalen et al. 2007). C. freundii is known as an opportunistic human pathogen and causes a variety of respiratory tract, urinary tract and blood infections, and responsible for serious neonatal meningistis disease (Badger et al. 1999, Whalen et al. 2007).

Interestingly, this bacterium plays a positive environmental role through nitrogen recycling

(Puchenkova 1996). Strains of C. freundii also contribute to the environment by sequenstering heavy metals and degrading xenobiotic compounds (Kumar et al. 1999, Sharma and Fulekar

2009).

1.11. Hypotheses and objectives

The study was based on the following hypotheses: i. Certain microorganisms in DON contaminated agricultural soils are able to transform DON to de-epoxy DON (de-epoxidation) under aerobic conditions and moderate temperatures. ii. DON de-epoxidation may be achieved by a single microbial strain. iii. Microbial DON detoxifying genes and associated regulators are able to express and function in a closely related host organism. iv. DON detoxifying microbial strains carry genes for de-epoxidation activity that are absent from related species that have not evolved under DON selection pressure.

The objectives of this research were to: i. Select microbial cultures from soil having DON de-epoxidation activity under aerobic conditions and moderate temperatures by analyzing potential DON contaminated soils in various growth conditions, namely aerobic/anaerobic conditions, different growth media, temperature and pH. ii. Select for organisms with DON de-epoxidation activity from the mixed microbial culture by treating with antibiotics, heating, subculturing on agar medium, and incubating for a prolonged period in media with a high DON concentration. iii. Identify microbial de-epoxidation genes by producing a microbial genomic DNA library from the DON de-epoxidizing strain, followed by screening of the clones for DON de- epoxidation.

31 iv. Determine microbial de-epoxidation genes by sequencing of the genome and comparative analysis of the genome with closely related species and identification of putative DON detoxification genes.

CHAPTER 2

2.0. AEROBIC AND ANAEROBIC DE-EPOXIDATION OF MYCOTOXIN

DEOXYNIVALENOL BY BACTERIA ORIGINATING FROM AGRICULTURAL

SOIL

Published in World Journal of Microbiology and Biotechnology. 2012, 28:7-13.

2.1. Abstract

Soil samples collected from different cereal crop fields in southern Ontario, Canada were screened to obtain microorganisms capable of transforming deoxynivalenol (DON) to de-epoxy DON (dE-DON). Microbial DON to dE-DON transformation (i.e., de-epoxidation) was monitored by using liquid chromatography-ultraviolet-mass spectrometry (LC-UV-MS).

The effects of growth substrates, temperature, pH, incubation time and aerobic versus anaerobic conditions on the ability of the microbes to de-epoxidize DON were evaluated. A soil microbial culture showed 100% DON to dE-DON biotransformation in mineral salts broth (MSB) after 144 h of incubation. Treatments of the culture with selective antibiotics followed an elevated temperature (50ºC) for 1.5 h and subculturing on agar medium considerably reduced the microbial diversity. Partial 16S-rRNA gene sequence analysis of the bacteria in the enriched culture indicated the presence of at least Serratia, Citrobacter,

Clostridium, Enterococcus, Stenotrophomonas and Streptomyces. The enriched culture completely de-epoxidized DON after 60 h of incubation. Bacterial de-epoxidation of DON occurred at pH 6.0 to 7.5, and a wide array of temperatures (12 to 40ºC). The culture showed

33 rapid de-epoxidation activity under aerobic conditions compared to anaerobic conditions. This is the first report on microbial DON to dE-DON transformation under aerobic conditions and moderate temperatures. The culture could be used to detoxify DON contaminated feed and might be a potential source for gene(s) for DON de-epoxidation.

Keywords: biotransformation, de-epoxidation, deoxynivalenol, soil bacteria, vomitoxin

2.2. Introduction

Deoxynivalenol (DON, commonly known as ‘vomitoxin’) 3α, 7α, 15-trihydroxy-

12,13-epoxytrichothec-9-en-8-one, is a trichothecene produced by toxigenic Fusarium species causing Fusarium head blight (FHB) or Gibberella ear rot (GER) disease on small grain cereals and maize, respectively (Starkey et al. 2007, Foroud and Eudes 2009). DON contamination in cereals and maize is widespread and can lead to tremendous losses to the growers. It has also detrimental effects on human and livestock health (Pestka and Smolinski

2005, Rocha et al. 2005). The toxin interferes with the eukaryotic ribosomal peptidyl transferase (Rpl3) activity, resulting in human and animal mycotoxicosis (Ueno 1985, Bae and Pestka 2008). Common effects of DON toxicity in animals include diarrhea, vomiting, feed refusal, growth retardation, immunosuppression, reduced ovarian function and reproductive disorders (Pestka and Smolinski 2005, Pestka 2007). In plants, DON causes a variety of effects, viz. inhibition of seed germination, seedling growth, green plant regeneration, root and coleoptile growth (cited in Rocha et al. 2005). In addition, DON enhances disease severity in sensitive host plants (Maier et al. 2006).

Avoidance of DON contamination or reduction of toxin levels in contaminated grains remains a challenging task. Physical and chemical processes are either inappropriate or prohibitively expensive to use to decrease the DON content of agricultural commodities

(Kushiro 2008, He et al. 2010b). The use of Fusarium resistant crop varieties might be the best method to reduce DON contamination. However, breeding for FHB or GER resistant cultivars appears to be very difficult and time consuming as natural sources of resistance against Fusarium-caused diseases is quantitative in nature (Ali et al. 2005, Buerstmayr et al.

2009). As an alternative, expressing genes for enzymatic detoxification of DON in elite cultivars may be an effective strategy for protecting the value of feeds and foods (Karlovsky

1999). Heterologous expression of eukaryotic trichothecene 3-O-acetyltransferase and UDP- glucosyltransferase genes in cereals and Arabidopsis have resulted in reduced toxicity via acetylation and glycosylation of the DON (Poppenberger et al. 2003, Ohsato et al. 2007).

It has been proposed that enzymatic de-epoxidation of trichothecene toxins could be used to limit DON contamination (Zhou et al. 2008, Foroud and Eudes 2009). DON belongs to type B-trichothecenes, which contain an epoxide ring at C12-13 and the epoxide is highly reactive and mainly responsible for toxicity (Figure 2-1) (Swanson et al. 1988, Eriksen et al.

2004).

De-epoxydation

Deoxynivalenol De-epoxyDeoxynivalenol (DON) (dE-DON)

Figure 2-1. Mechanism of microbial DON de-epoxidation. Removal of the oxygen atom from the epoxide ring results in formation of a C=C double bond at C-12 position, which reduces cytotoxicity (Swanson et al. 1988, Eriksen et al. 2004).

Microbial DON de-epoxidation by single bacterial strains and by mixed cultures has been reported (King et al. 1984, Swanson et al. 1987, He et al. 1992, Kollarczik et al. 1994,

Fuchs et al. 2002, Young et al. 2007, Guan et al. 2009, Yu et al. 2010). However, all of these microbes originated from fish and other animals, and were strictly anaerobic. The recovered microbial cultures and single strains only showed de-epoxidation of trichothecenes at low

(≤15ºC) and high (>30°C) temperatures. Such enzymes might have limited applications for toxin decontamination in cereal grains. The aim of the current study was to isolate the microbes from soil capable of transforming DON to a de-epoxy metabolite under aerobic conditions and moderate temperatures.

2.3. Materials and methods

2.3.1. Soil samples

The upper layer (0-30 cm) of soils along with crop debris were randomly collected from cereal (e.g., maize, wheat, barley) crop fields during 2004-2006 in Southern Ontario,

Canada. Detailed of the crops and locations of the fields have been described in Jianwei He’s

Master Degree thesis (He 2007). Briefly, soil samples were collected in sterile plastic bags and transported to the laboratory. Samples were sieved to remove large roots, stones and stored at 4ºC until used. The samples were mixed with grounded mouldy corn (containing 95

µg/g DON) + F. graminearum and incubated at 22-24°C for six weeks. Five samples were prepared for DON de-epoxidation activity analysis by dispersing 5 g soil in 50 mL distilled water and storing at4°C until further use.

2.3.2. Culture media

Media used in the study were: nutrient broth (NB), Luria Bertani (LB), minimal medium (MM) (He 2007), six mineral salts + 0.5% bacto peptone (MSB) (He 2007), tryptic soy broth (TSB), corn meal broth (CMB) (Guan et al. 2009), potato dextrose broth (PDB), six mineral salts + 0.5% yeast extract broth (MSY) and soil extract (SE). MM was composed of

. sucrose-10.0 g, K2HPO4-2.5 g, KH2PO4-2.5 g, (NH4)2HPO4-1.0 g, MgSO4 7H2O-0.2 g,

. FeSO4-0.01 g, MnSO4 7H2O- 0.007 g, water- 985 mL, and MSB was composed of the same salts as MM but was supplemented with Bacto™ Peptone (0.5%, w/v) as the sole C source instead of sucrose. To prepare soil extract, 100 g of the composite soil was homogenized in 1

L water. The soil particles and plant debris were removed by passing the soil suspension through Whatman® grade 3 filter paper (Sigma-Aldrich, Oakville, Ontario, Canada). The

37 extract was autoclaved at 121°C for 30 min. Media and ingredients were purchased either from Fisher Scientific Inc. (Fair Lawn, New Jersey, USA), Sigma-Aldrich (Oakvile, Ontario,

Canada) or Difco (Sparks, Maryland, USA). Prior to autoclaving, pH of the broth media was adjusted to 7.0. Solid media were prepared by adding 1.5 % agar.

2.3.3. Screening of soil samples

Frozen soil suspensions were thawed and vortexed to homogenize the microbial population. An aliquot (100 µL) of the suspension was immediately transferred to MSB, PDB and MM broth media. DON at 50 µg/mL was added to the microbial suspension cultures. The cultures were incubated at 22-24°C with continuous shaking at 180 revolutions per minute

(rpm). The extent of DON de-epoxidation was determined at 24 h, 48 h, 72 h, 96 h, 120 h and

144 h of incubation.

2.3.4. Determination of de-epoxidation product

Microbial DON de-epoxidation was measured according to Guan et al. (2009).

Samples were prepared by mixing equal volume of microbial culture and methanol, and allowed to stand for 30 min at room temperature. Microorganisms in the mixture were removed by passing them through a 0.45 µm Millipore filter (Whatman®, Florham Park, New

Jersey, USA). The filtrates were used for de-epoxidation analysis by a liquid chromatography- ultraviolet-mass spectrometry (LC-UV-MS). Identification and quantification of DON and its transformation product was achieved using a Finnigan LCQ DECA ion trap LC-MS instrument (ThermoFinnigan, San Jose, CA, USA). The equipment detects 30-2000 m/z

(molecular mass: ion ratio) compounds and able to determine 0.01 µ/mL DON. The HPLC

38 system was equipped with a Finnigan Spectra System UV6000LP ultraviolet (UV) detector and an Agilent Zorbax Eclipse XDB C18 column (4.6×150 mm, 3.5 μm). The binary mobile phase consisted of solvent A (methanol) and solvent B (water). The system was run with a gradient program: 25% A at 0 min, 25–41% A over 5 min, isocratically at 41% A for 2 min and 41–25% A over 1 min. There was a 3 min post-run at its starting condition to allow for the equilibration of the column. The flow-rate was set at 1.0 mL/min. UV absorbance was measured at 218 nm. Atmospheric pressure chemical ionization (APCI) positive ion mode was used for MS detection. A DON standard was used to adjust the instrument to its maximum response. The system was operated as follows: shear gas and auxiliary flow-rates were set at 95 and 43 (arbitrary units); voltages on the capillary, tube lens offset, multipole 1 offset, multipole 2 offset, and entrance lens were set at 30.00, −17.00, −3.70, −8.50, −35.00, and −42.00 V, respectively; capillary and vaporizer temperatures were set at 200 and 450°C, respectively; and the discharge needle current was set at 7 μA. Identification of mycotoxin was achieved by comparing their retention times, UV spectra and mass spectrum with DON and de-epoxy DON (Biopure, Union, MO, USA) as standards. Quantification of DON and product was based on peak areas related to their UV and MS calibration curves.

2.3.5. Biological DON de-epoxidation

To help identify the types of microorganisms causing the DON de-epoxidation, MSB medium was supplemented with either 100 µg/mL penicillin G or streptomycin sulfate for suppressing gram positive or negative bacterial growth and activities, respectively. Three replications of antibiotic-treated media were inoculated with 100 µL of microbial culture capable of de-epoxidizing DON. The antibiotic-treated cultures were incubated at 27°C spiked with 50 µg/mL DON. Controls were prepared without adding antibiotics in the microbial culture. Inhibition of DON de-epoxidation was analyzed after five day of incubation.

2.3.6. Reduction of microbial diversity by antibiotic treatments, heating and subculturing on agar medium

In order to reduce non-DON degrading microbial populations, the DON de- epoxidation capable cultures were sequentially treated with antibiotics having diverse modes of action, including: i) dichloramphenicol that kills various gram positive, Gram-negative and anaerobic bacteria, ii) trimethoprim that kills Gram-positive and negative bacteria, iii) penicillin that is lethal to Gram-positive bacteria and iv) cyclohexamide to kill filamentous fungi and yeasts. Antibiotics were purchased from Sigma-Aldrich. Antibiotic stock solutions

(10 mg/mL) were prepared with sterile water or methanol. Microbial cultures in MSB medium were treated with 100 µg/mL of each antibiotic followed by incubation of the culture at 50ºC for 1.5 hours to kill undesirded microbial species. After heating, cultures were treated with 50 µg/mL DON and incubated at 27ºC for five days. DON de-epoxidation activity of the microbial cultures was determined following the procedure described above. Cultures that

40 showed >99% de-epoxidation were considered for subsequent antibiotic and heat treatments.

Aliquots of the achieved culture were spread on MSB agar plates and grown at room temperature for few days. Colonies appeared on the plates were collected and incubated in

MSB medium supplemented with 50 µg/mL DON. De-epoxidation ability of the subcultured microbial cultures was determined after five days of incubation at 27°C by LC-UV-MS method.

2.3.7. Analysis of microbial population diversity

Terminal Restriction Fragment Length Polymorphism (T-RFLP) analysis

Initially microbial diversity was determined by terminal restriction fragment length polymorphism (T-RFLP) fingerprint analysis (Bruce and Hughes 2000). Genomic DNA from microbial cultures with and without enrichment was extracted using DNeasy Plant DNA

Extraction Mini Kit (Qiagen Inc., Mississauga, Ontario, Canada). 16S-rDNA was amplified by PCR using the HEX-labeled forward primer Bac8f (5'-agagtttgatcctggctcag-3') and unlabeled reverse primer Univ1492r (5'-ggttaccttgttacgactt-3') (Fierer and Jackson 2006).

Fifty microlitter PCR mixture for each sample was prepared having 1x HotStarTaq Master

Mix (Qiagen), 0.5 µM of each primer, 50 µg of BSA, and 50 ng of DNA. The 35 PCR cycles followed 60 s at 94°C, 30 s at 50°C, and 60 s at 72°C. After size verification by agarose-gel electrophoresis, 50 ng of purified PCR products was digested for 3 h in a 20 µL reaction mixture with 5 U of HhaI restriction enzyme (New England Biolabs, Ipswich, MA, USA).

The size and relative abundance of the labeled T-RFLP fragments were separated by capillary electrophoresis in an automated DNA sequencer equipped with fluorescence detector

(Applied Biosystem 3730 DNA Analyzer). The DNA fragments were analyzed with the Peak

Scanner™ Software v1.0 to create an electropherogram with the genetic fingerprints of the microbial community.

Analysis of V3-V5 region of the 16S-rDNA

To define the bacterial genera present in the enrichment culture, the 16S-rDNA was amplified using two sets of bacterial universal primers. Primarily, 16S-rDNA was PCR amplified using the primers 46f-5’gcytaacacatgcaagtcga3’ (Kaplan and Kitts 2004) and

1371r-5’gtgtgtacaagncccgggaa3’ (Zhao et al. 2007). PCR was performed in a total volume of

25 μL, which contained 0.6 μM of each primer, 1x of HotStarTaq® Master Mix (Qiagen Inc.) and 20-40 ng of DNA. The PCR cycles included the following steps: 95ºC (15 min); 35 cycles of 94ºC (20 s), 52ºC (15 s), 72ºC (1 min 30 s); and a 72ºC (7 min) extension step. The second pair of universal bacterial primers were 27f-5’agagtttgatcmtggctcag3’and 1492r-

5’ggttaccttgttacgactt3’ (Lane 1991). PCR was conducted with 25 cycles following the procedures of Tian et al. (2009). Amplification was carried out in a GeneAmp® PCR System

9700 (Applied Biosystems). Furthermore, the PCR products were used as templates to amplify the hypervariable V3-V5 region of the 16S-rDNA using primers 357f-

5’cctacgggaggcagcag3’ and 907r7 5’ccgtcaattcctttgagttt3’ (Yu and Morrison 2004).

After purification of the PCR products by QIAquick PCR purification kit, two amplicon libraries were prepared by ligating to pCR-TOPO vector (Invitrogen) and the library clones were transformed into TOPO10 competent cells according to Invitrogen’s TOPO cloning manual. The clone libraries were plated on LB plates containing 25 μg/mL kanamycin. Fifty colonies from each library were randomly picked up from the overnight plates and inserts were sequenced as described by Geng et al. (2006). PCR products were

42 sequenced with an automated ABI 3730 DNA analyzer (Applied Biosystems). The sequence similarities of the amplified V3-V5 region were compared with the ribosomal database project

(RDP) (http://rdp.cme.msu.edu/). Sequences with greater than 95% identity were assigned to the same bacterial genus.

2.3.8. Factors influencing DON de-epoxidation

DON de-epoxidation ability of the enriched microbial culture was investigated under different conditions as described below.

Maintaince of active culture

To retain microbial de-epoxidation activity, the active culture has been maintained in

MSB medium containing 50 µg/mL DON.

Media, growth temperature and pH

Overnight microbial cultures (100 μL) were added to 2 mL of NB, LB, MM, TSB,

MSB, CMB, MSY, PDB or SE broths. The pH of the media was adjusted to 7.0 and the experiments were conducted at 27°C temperature. The effect of temperature was studied by incubating the culture in MSB medium (pH 7) at 12, 20, 25, 30 and 40ºC. To determine the effects of pH, MSB medium pH was adjusted to 5.5, 6.0, 6.5, 7.0, 7.5 and 8.0. Media with different pHs were inoculated with aliquots of the overnight active culture and then incubated at 27ºC. The cultures were supplemented with 50 μg/mL DON before incubation. Three replications for each medium, temperature and pH condition were evaluated.

De-epoxidation under aerobic and anaerobic conditions

Three replicate microbial cultures of five mL each were grown in sterile 15 mL centrifuge tubes under aerobic conditions and in an anaerobic chamber filled with 95% CO2 and 5% H2 (COY Laboratory Products Inc., Ann Arbor, MI, USA). Three days after initiation, half (2.5 mL) of the cultures from anaerobic conditions was transferred to new centrifuge tubes and brought to aerobic conditions. Similarly, half of the aerobically grown cultures were placed into an anaerobic chamber. Cultures moved from anaerobic to aerobic conditions were shaken after opening the lids of the tubes in a laminar flow every eight hours intervals to maintain aerobic conditions in the culture. Lids of the culture tubes transferred to anaerobic conditions were kept loose to allow oxygen to escape from the culture. Two days after transferring to the new conditions, the cultures in both conditions were treated with 50 µg/mL

DON and incubated at 27°C. Every 12 h, 500 µL were removed from each tube and analyzed for microbial DON de-epoxidation activity by LC-UV-MS.

2.3.9. Data analysis

Data were processed using Microsoft Excel Program 2007. Statistical analysis was done using General Linear Model procedures in SAS version 9.1 (SAS Institute Inc., Cary,

NC, USA, 2002-03). The results were the average of three replications, stated as mean ± standard error. Differences between means were considered significant at p< 0.05 according to Tukey’s-HSD test.

2.4. Results

2.4.1. Analyses of DON to dE-DON biotransformation

A typical LC-UV-MS analysis where there is incomplete de-epoxidation of DON by a particular soil suspension culture is shown in Figure 2-2. Analysis by LC with UV detection and integration of peak areas showed 66.2% conversion of DON to dE-DON (Figure 2-2a).

Analysis LC with MS detection (expressed as total ion counts) indicated 64.3% conversion

(Figure 2-2b). LC analyses with UV and TIC (total ion count) detection are subject to interferences due to co-eluting contaminants, and therefore, integration of small peaks was very difficult making quantification inaccurate. MS detection by SIM (selected ion monitoring) can eliminate the influence of co-eluting contaminants. For example, the SIM LC chromatogram for DON (Figure 2-2c) was generated by using the major ions resulting from the APCI positive ion MS of DON (Figure 2-2e).

Similarly, the SIM LC chromatogram for dE-DON (Figure 2-2d) was derived from the

MS ions of dE-DON (Figure 2-2f). The signal to noise ratios (S/N) of peaks observed by SIM were at least 20 times greater than those detected by TIC because there is reduced likelihood of co-eluting peaks having the same ions. Integration of the SIM peaks revealed a 61.6% conversion. The three methods gave similar results, but the SIM LC results were considered to be the most accurate. Using that method, the conversion of DON to dE-DON by the mixed culture occurred most quickly at 48 h and was completed by 144 h (Figure 2-3).

Figure 2-2. LC-UV-MS chromatographic analysis of DON and its de-epoxidated product dE-

DON obtained by culturing with selected soil microbes. (a) LC-UV chromatogram of product mixture monitored at 218 nm; (b) LC-MS total ion chromatogram (TIC) of product mixture;

(c) selected ion monitoring (SIM) chromatogram of cultured DON at m/z 231, 249, 267, 279, and 297; (d) SIM chromatogram of cultured dE-DON at m/z 215, 233, 245, 251, 263, and 281;

(e) atmospheric pressure chemical ionization (APCI) positive ion mass spectrum of DON; (f)

APCI positive ion mass spectrum of dE-DON.

2.4.2. Screening of soil samples

Initially, samples originating from soils were analyzed in three types of culture media to support the growth of fungi, yeast and bacteria. However, cultures growing in PDB, which is a suitable substrate for many fungi and yeasts, did not permit transform DON to dE-DON

(data not shown by microorganisms). A define medium (i.e., MM) was also used in the sample screening assay. None of the soil samples showed de-epoxidation activity in MM

(data not shown). From five soil samples, one soil sample culture showed de-epoxidation of

DON in MSB. The culture in MSB converted >99% DON to dE-DON within 144 h (Figure 2-

3). Conversion began after a lag of 48 h and was linear thereafter with a rate of 1% per hour.

Figure 2-3. Time course of relative biotransformation of DON to dE-DON by the initial soil suspension culture in MSB medium. The culture contained 50 μg/mL DON at time 0. Culture with three replications was incubated at room temperature with continuous shaking at 180 rpm. The conversion of DON to dE-DON was analysed by LC-UV-MS at 24 h intervals. Blue and red lines are the percent conversion of DON and recovery of dE-DON, respectively.

2.4.3. Evidence of biological DON de-epoxidation

Treatment of active microbial culture with 100 µg/mL of streptomycin completely inhibited DON de-epoxidation. In contrast, the same concentration of penicillin G did not affect activity of the culture (Figure 2-4). Streptomycin is effective against Gram-negative bacteria, while penicillin G is lethal to Gram-positive bacteria. These experiments provided evidence that DON de-epoxidation may be caused by certain Gram-negative bacteria.

120

100

60 epoxydation (%) epoxydation

- 40

DON de DON 0 Penicillin Strepto Control

Antibiotics

Figure 2-4. Microbial de-epoxidation of DON after five days of incubation. The active DON de-epoxidation culture was treated with either 100 µg/mL penicillin G or streptomycin

(Strepto) in MSB. Cultures having three replications were added 50 µg/mL DON and incubated at room temperatures with continuous shaking at 180 rpm. For controls, cultures were without antibiotics. DON to dE-DON transformation was monitored after five day of incubation by LC-UV-MS method.

2.4.4. Reduction of the microbial diversity

Plating the suspension cultures onto MSB agar revealed colonies with a variety of sizes, morphologies and colours reflecting a mixed culture (data not shown). To reduce the complexity of the culture three steps of selection pressures were applied, namely treatment with antibiotics, heat shocks and subculturing on solid medium. The above indicated treatments did not significantly affect DON de-epoxidation of the microbial cultures (data not shown). T-RFLP profiles of the mixed culture before treatments showed six distinct peaks of about 110, 150, 200, 220, 340 and 420 bp restriction fragments, with the 110 bp peak being the predominant one (Figure 2-5a).

After treatment with antibiotics and heat, only three peaks at 110, 150 and 325 bp remained (Figure 2-5b), suggesting that there was a reduction in the microbial diversity without affecting DON de-epoxidation. After enrichments the bacteria species in the culture changed their growth patterns as the electropherogram produced different sizes of pick lengths compared to untreated culture.

(a)

0 100 200 300 400 500

32000 Relative fluorescence Relative

0 16S-rDNA fragment size (bp)

(b)

0 100 200 300 400 500

32000 Relative fluorescence Relative

16S-rDNA fragment size (bp)

Figure 2-5. Analysis of microbial population reduction by terminal restriction fragment length polymorphism (T-RFLP). (a) Restriction fragments of the 16S-rDNA of the untreated culture, and (b) fragmentation patterns of the enriched culture. Picks along the X-axis showed the fragmentation patterns of the 16S-rDNA which is related to the bacterial species diversity.

The length of picks (along the Y-axis) showed abundance of terminal restriction fragments related to the abundance of particular bacterial species/genera.

2.4.5. Microbial population structures in the enriched culture

Microbial diversity in the enriched culture was determined by analyzing the sequences of the V3-V5 region (~586 bp) amplified from the 16S-rDNAs of the microorganisms it contained. The sequence analysis showed the presence of Serratia, Citrobacter, Clostridium,

Enterococcus, Stenotrophomonas and Streptomyces in the enriched culture (Figure 2-6).

Out of 100 V3-V5 sequences, 54 sequences were found to be belonging to Serratia, the predominant species of the culture. The remaining sequences matched with 12

Citrobacter, 8 Clostridium, 14 Enterococcus, 8 Stenotrophomonas and 4 Streptomyces species.

60 54

20 12 14 8 8

4 Sequence frequency frequency Sequence 0

Microbial genera

Figure 2-6. Microbial genera identified in the enriched DON de-epoxidizing microbial culture. The bacterial genera were determined by comparing the hypervariable regions (V3-

V5) of the 16S-rDNA using ribosomal database project (http://rdp.cme.msu.edu/). X-axis showed the identified bacterial genera and Y-axis demonstrated the abundance of the V3-V5 sequences belonging to different bacterial genera.

2.4.6. Effects of media, growth temperature, pH and oxygen requirements on DON de- epoxidation

Of the nine media that were tested, DON de-epoxidation by the enriched microbial culture only occurred in NB and MSB (Figure 2-7a). It appeared that C and/or N concentrations in the media were crucial for the enzymatic activity because less or more than

0.5% Bacto Peptone in MSB reduced the microbial DON transformation ability.

DON de-epoxidation activity occurred over a wide range of temperatures from 12 to

40ºC (Figure 2-7b). The fastest rates of DON de-epoxidation occurred between 25 to 40ºC.

De-epoxidation was observed within a relatively narrow range of pH (6.0 to 7.5) with the most rapid DON biotransformation activity observed at pH 7.0 (Figure 2-7c). At pH 7.0 the enriched culture showed DON de-epoxidation activity 1.25 times faster than at pH 6.0 or

7.5. At pH 5.5 and 8.0, there was no observable DON de-epoxidation.

(a)

120

100

60 epxydation epxydation (%) - 40

DON de DON 0

Media

(b)

120

100

60 epoxydation (%) epoxydation - 40

20 DON de DON

C C C C C

° ° ° ° °

12 20 25 30 40

Temperature

(c)

120

100

60 epoxydation (%) epoxydation

- 40

DON de DON 0

5.5 6.0 6.5 7.0 7.5 8.0 pH

Figure 2-7. Effects of culture conditions on DON to de-epoxy DON transformation by the enriched microbial culture. For the media experiment (a), all the media were adjusted to pH 7 and the incubation temperature was 27°C. For the temperature experiment (b), MSB pH 7 was used. For the pH experiment (c), MSB at 27°C was used. Microbial cultures were inoculated with 50 μg/mL DON. De-epoxidation was determined after 72 h of incubation. Media: MSY

(mineral salts + 0.5% yeast extract), LB (Luria Bertani), TSB (tryptic soy broth), MSB

(mineral salts + 0.5% bacto peptone), SE (soil extract), CMB (corn meal broth), PDB (potato dextrose broth) and MM (minimal medium).

The enriched culture showed DON de-epoxidation activity under both aerobic and anaerobic conditions (Figure 2-8). The fastest transformation (100% de-epoxidation after 60 h of incubation) occurred in cultures (A) that were grown under aerobic conditions throughout the culture growth period. A medium rate of de-epoxidation (100% de-epoxidation after 72 h) was observed in cultures that were initially grown in either aerobic or anaerobic conditions followed by transferring to anaerobic or aerobic conditions (A+AN and AN+A), respectively.

The lowest rate occurred (100% de-epoxidation after 96 h) in cultures grown in continual anaerobic conditions (AN). The means comparison of microbial DON de-epoxidation efficiencies at 60 h of incubation showed that cultures continually grown in conditions A and

AN were statistically different. Biotransformation abilities of the cultures that were switched from one conditions to another (A+AN and AN+A) were statistically similar among each other but different from cultures that were only grown in conditions A or AN (p<0.05).

Figure 2-8. Microbial DON to dE (de-epoxy DON)-DON transformation activity under aerobic and anaerobic conditions. A = culture incubated and examined the activity only under aerobic conditions; A+AN = culture initially was grown in aerobic conditions and then transferred to anaerobic conditions; AN = cultures grown throughout the assay under anaerobic conditions; AN+A = culture initially grown under anaerobic conditions and then transferred to aerobic conditions. Differences of lower case letters are significantly different in de-epoxidation capability (after 60 h) among incubation conditions according to Tukey’s

HSD test (p < 0.05). The experiment was conducted with three replications.

2.5. Discussion

The current study describes a novel bacterial culture isolated from agricultural soil that was capable of transforming DON to de-epoxy DON under aerobic and anaerobic conditions at moderate temperatures. Because it has been speculated that DON translocates to soil from contaminated crops and DON that enters into the soil could be metabolized by microbes

(Völkl et al. 2004), soil samples for the current study were mainly collected from fields where cereal crops contaminated with DON were grown.

In our initial sample screening steps, upwards of 144 h of incubation were required to achieve >99% DON de-epoxidation. Such a long period may have been due to the presence of fast growing microorganisms that had suppressive effects on the growth or activity of the

DON transforming bacteria (Davis et al. 2005). The sequential use of different antibiotics with different modes of actions followed by thermal treatments and subculturing on solid medium resulted in a reduction in the number of bacterial species as determined by T-RFLP analysis. Analysis of V3-V5 region of the 16S-rDNA showed that five bacterial genera were present in the enriched culture. The culture showed higher DON de-epoxidation activity (only

60 h required for >99% biotransformation) compared to initial soil suspension culture that required 144 h to complete the de-epoxidation. This microbial culture could be used to isolate the DON de-epoxidizing gene(s) applying metagenomics approaches (Daniel 2005).

The culture showed complete de-epoxidation activity both in MSB and NB media.

However, we have found that the use of nutritionally rich NB medium enhanced non-target bacterial (e.g., Serratia lequifaciences) growth, but this did not occur in MSB, which is a comparably poor medium. To avoid the possible suppressive and detrimental effects of the fast growing bacteria on the target bacteria, MSB was used for enrichment and

62 characterization of the culture. An important characteristic of the DON de-epoxidation activity observed for the microbial culture in the current study is its activity at moderate temperature (25-30°C). In contrast, previously isolated cultures showed optimal activities at low (≤15°C) or high temperatures (37°C) (Guan et al. 2009). Also, the culture showed DON de-epoxidation activity at neutral pH range (pH 6.0-7.5) in contrast to previous reports, which found that microbial trichothecene de-epoxidation occurred at acidic and basic pHs (Swanson et al. 1987, He et al. 1992, Kollarczik et al. 1994, Young et al. 2007). However, a microbial culture from fish gut showed de-epoxidation under wide range of pHs (Guan et al. 2009).

Probably, the pH range reflects the environment in which the bacteria naturally grow which in this case is similar to the soil pH of a typical crop field in Southern Ontario of Canada. The advantage of the present finding is that DON de-epoxidation activity of the culture was optimal in a temperature range that matches the temperature range of activity for most arable plants. Therefore, it is reasonable to suggest that the gene(s)/enzyme(s) responsible for the activity have potential application for enzymatic detoxification of DON in cereals (Karlovsky

1999).

The current observation that bacterial DON de-epoxidation activity occurred under both aerobic and anaerobic conditions was interesting because DON de-epoxidation is a reductive chemical reaction that might be expected to be promoted by anaerobic conditions.

However, there are previous reports of aerobic de-epoxidation of eldrin/dieldrin by soil bacteria (Matsumoto et al. 2008), and Goodstadt and Ponting (2004) described aerobic reductive de-epoxidation by vitamin K epoxide reductase/phylloquinone reductase. A definitive determination of the oxygen sensitivity of the enzyme responsible for DON de- epoxidation will ultimately come from experiments with the purified protein.

In conclusion, this study reports on a novel bacterial source from soil that efficiently de-epoxidizes DON at moderate temperatures under both aerobic and anaerobic conditions.

This research may lead to the isolation of microbes and gene(s) encoding enzyme(s) capable of de-epoxidizing DON that could create transgenic plants with reduced DON contamination, thus improving food and feed safety and reducing negative effects of the toxin on public health.

2.6. Acknowledgements

The authors gratefully acknowledged the Natural Sciences and Engineering Research

Council of Canada (NSERC) for awarding a scholarship to Rafiqul Islam for PhD program.

The research was supported by the Agriculture and Agri-Food Canada/ Binational

Agricultural Research and Development Fund (AAFC/BARD) project. We are thankful for the contributions of Jianwei He, Xiu-Zhen Li and Honghui Zhu.

CHAPTER 3

3.0. ISOLATION AND CHARACTERIZATION OF A NOVEL SOIL BACTERIUM

CAPABLE OF DETOXIFYING ELEVEN TRICHOTHECENE MYCOTOXINS

3.1. Abstract

Deoxynivalenol (DON, vomitoxin) is one of the most commonly detected food and feed contaminating mycotoxins worldwide. In order to develop a control method effective under aerobic conditions and moderate temperatures, a bacterial strain, ADS47, was isolated capable of transforming DON to the less toxic de-epoxy DON. The bacterial strain is facultatively anaerobic, Gram-negative, rod shaped with an average cell size of 1.50 X 0.72

µm, and has peritrichous (multiple) flagella. The 16S-rRNA gene sequence had >99% nucleotide identity to Citrobacter freundii, a bacterium that belongs to the gamma subclass of proteobacteria. Strain ADS47 showed near 100% de-epoxidation of DON after 42 h of incubation in nutrient broth (NB) under aerobic conditions, moderate temperature (28˚C) and neutral pH (7.0). However, de-epoxidation could occur at 10-45˚C and pH 6.0-7.5. Under anaerobic conditions, near 100% de-epoxidation was observed after 72 h mainly during the stationary growth phase. Microbial DON de-epoxidation ability increased with increasing

DON concentrations in the culture media. Inhibition of bacterial de-epoxidation by sodium azide suggested that an active electron transport chain was required. In addition to DON, strain ADS47 was able to de-epoxidize and/or de-acetylate five type-A and five type-B trichothecene mycotoxins. This is the first report of a bacterial isolate capable of de- epoxidizing DON and trichothecene mycotoxins under aerobic conditions and moderate

65 temperatures. Strain ADS47 could potentially be used in bioremediation to eliminate DON and other trichothecene mycotoxins from contaminated food and feed.

Keywords: aerobic de-epoxidation, Citrobacter freundii, isolation, soil bacterium, trichothecenes

3.2. Introduction

Cereals are often contaminated with a variety of type-A and type-B trichothecene mycotoxins, produced by a number of toxigenic Fusarium species (Placinta et al. 1999,

Binder et al. 2007, Foroud and Eudes 2009). Trichothecenes are tricyclic sesquiterpenes with a double bond at the C9, 10 position and a C12, 13 epoxy group (Desjardins 2006). The two types are distinguished by the presence of a ketone at position C8 in type-B, which is absent in type-A (Yazar and Omurtag 2008). The type-B toxin, deoxynivalenol (DON), is the most frequently found mycotoxin in food and feed around the world (Larsen et al. 2004, Desjardins

2006, Starkey et al. 2007). Different Fusarium species associated with Fusarium head blight

(FHB) or Gibberella ear rot (GER) disease of small grain cereals and maize are mainly responsible for contaminating grains with DON and other type-A and type-B trichothecenes.

These can cause serious mycotoxicosis in humans, animals and plants (Ueno 1985, D’Mello et al. 1999, Bae and Pestka 2008). Common effects of DON/TC toxicity in animals include diarrhea, vomiting, feed refusal, growth retardation, immunosuppression, reduced ovarian functions/reproductive disorders and even death (Pestka and Smolinski 2005, Pestka 2007).

DON enhances FHB disease severity in wheat and maize (Proctor et al. 1995, Desjardins

1996, Maier et al. 2006).

Reduction of mycotoxin levels in contaminated agricultural products remains a challenging task. Current physical and chemical processes to decrease DON content are either ineffective or prohibitively expensive (Kushiro 2008, He et al. 2010b). However, the use of detoxifying microbes to biodegrade mycotoxins is an alternative (Awad et al. 2010,

Karlovsky 2011). The toxicity of DON is mainly due to the epoxide, and there is a significant reduction of DON toxicity (54%) when it is reduced to produce de-epoxy DON. Related mycotoxins also show reductions in toxicity when they are de-epoxidized, such as NIV (55%) and T-2 toxin (400%) (Swanson et al. 1988, Eriksen et al. 2004). Bacillus arbutinivorans strain LS100 and Eubacterial strain BBSH 797, originating from chicken intestine and rumen fluid, respectively, are capable of cleaving the epoxide ring of DON under anaerobic conditions and high temperature (37°C) (Fuchs et al. 2002, Yu et al. 2010). Strain BBSH 797, which can de-epoxidize DON, scirpentriol and T-2 triol trichothecene mycotoxins, has been used as an additive for poultry and swine diets (Karlovsky 2011). However, there are no reports of microbes able to de-epoxidze DON and other trichothecenes under aerobic conditions and at moderate temperatures.

To develop an effective method for eliminating DON and trichothecene mycotoxins from contaminated grains, soils from fields with FHB/GER infected crops were screened for

DON de-epoxidizing microbes (refer to chapter 2; Islam et al. 2012). One soil sample yielded a mixed microbial culture was partially enriched for bacteria with DON de-epoxidizing activity (chapter 2). The current study describes the purification and characterization of a novel DON de-epoxidizing bacterium from this culture.

3.3. Materials and methods

3.3.1. Microbial source

A mixed microbial culture originating from a screen of more than 150 separate soil samples collected from maize and wheat fields near Guelph, Ontario, Canada (Islam et al.

2012) was used for enrichment, purification and characterization of the DON de-epoxidizing bacterium.

3.3.2. Media and chemicals

The media used were: nutrient broth (NB), 1/10 strength NB (1/10 NB), minimal medium (MM) consisting of six mineral salts (MS) + 10.0g per liter sucrose (Islam et al.

2012), MS + 0.5% Bacto Peptone (MSB), MS + 0.3% Bacto Peptone (MSC), MS + 0.5% yeast extract (MSY), MS + 200 µg/mL DON (MS-DON), sterile water-soluble soil extract

(SE) (Islam et al. 2012), potato dextrose broth (PDB) and Luria Bertani (LB). Prepared media and media components were obtained from Fisher Scientific (Fair Lawn, New Jersey, USA),

Sigma-Aldrich (Oakville, ON, Canada) and/or Difco (Sparks, Maryland, USA). Media pH was adjusted as required prior to autoclaving. For solid media preparation, 1.5 % agar was added. DON and all trichothecene mycotoxins were purchased from Biomin (Romer Lab,

Union, MO, USA).

3.3.3. Isolation of the DON de-epoxidizing bacterial strain

Aliquots (20 µL) of the microbial culture partially enriched for DON de-epoxidizing ability (Islam et al. 2012) was transferred to 0.5 mL MSB (pH 7.0) in sterile 1.5 mL

Eppendorf tubes. After addition of 200 µg/mL pure DON, the culture was incubated at room

68 temperature with continuous shaking at 220 revolutions per minute (rpm). Seven to ten days later, 20 µL of the culture was transferred to fresh MSB supplemented with 200 µg/mL DON.

The culture was then incubated for 7 to 10 days under the above conditions. This procedure was repeated 12 times.

After three months of repeated transfers in broth with 200 µg/mL DON, 100 µL of the culture was serially diluted in 0.9 mL MSB containing 0.5% Tween®20 (Sigma Aldrich) to reduce bacterial cell aggregation (Mabagala 1997). The dilutions were vigorously shaken for ten minutes. A 100 µL aliquot from each dilution was spread on MSB + agar (MSA). The

MSA plates were incubated at room temperature until single colonies appeared.

Single colonies from plates were transferred to 1 mL MSB+ 50 µg/mL DON and incubated at room temperature with continuous shaking at 180 rpm under aerobic conditions.

After three days, the DON de-epoxidation ability was examined by mixing an equal volume of the bacterial culture and methanol and incubating for 30 min at room temperature to lyse the cells. Insoluble materials in the mixture were removed by passing it through a 0.45 µm

Millipore filter (Whatman®, Florham Park, New Jersey, USA). The level of de-epoxy DON

(dE-DON, also known as DOM-1) in the filtrate was determined by liquid chromatography- ultraviolet-mass spectrometry (LC-UV-MS) (Guan et al. 2009).

Individual colonies that showed de-epoxidation activity were serially diluted in

MSB+0.5% Tween®20, and plated onto MSA. Single colonies that appeared on the agar plate were incubated in MSB as described above and retested for DON de-epoxidation ability. To ensure that pure cultures were obtained, this was repeated 10 times and individual colonies were selected on MSA each time.

3.3.4. Identification of a DON de-epoxidizing bacterium

Strain ADS47 with DON de-epoxidation ability was obtained from a single colony after 10 serial transfers, and was grown on MSA at room temperature under aerobic conditions for 72 h. Colony color, shape, edge and elevation were observed. The cell morphology of the bacterium was examined with a transmission electron microscope at the

University of Guelph, Guelph, Ontario, Canada. Whole bacterial cells were placed directly on a formvar and carbon-coated 200-mesh copper grid. They were washed two times by touching the grids to a drop of clean dH2O, and the excess water was blotted off. The samples were negatively stained using 1% Uranyl Acetate (aq) for five seconds. Electron microscopy was performed on a Philips CM10 transmission electron microscope at 60kV under standard operating conditions (Philips Optics, Eindhoven, Netherlands). Gram staining was performed using the procedure described in the commercial gram stain kit from BD Canada

(Mississauga, Ontario, Canada).

Total bacterial DNA was extracted using a DNeasy Plant DNA Extraction Mini Kit

(Qiagen Inc.). The 16S-rDNA was PCR amplified using the primers 46f-

5’gcytaacacatgcaagtcga3’ (Kaplan and Kitts 2004) and 1371r-5’gtgtgtacaagncccgggaa3’

(Zhao et al. 2007). PCR was performed in 25 µL containing 0.6 µM of each primer, 1x of

HotStarTaq® Master Mix (Qiagen Inc.) and ~40 ng of DNA. The PCR cycles were: 95ºC (15 min); 35 cycles of 94ºC (20 s), 52ºC (15 s), 72ºC (1 min 30 s); and a 72ºC (7 min) extension step. Amplification was carried out in a GeneAmp® PCR System 9700 (Applied Biosystems).

PCR products were sequenced with an automated ABI 3730 DNA analyzer (Applied

Biosystems). The 16S-rRNA gene sequence similarity was performed by BlastN using the

Ribosomal Database Project (RDP) (http://rdp.cme.msu.edu/seqmatch/).

A phylogenetic tree was constructed by comparing the partial 16S-rDNA sequence of strain ADS47 with the 11 known Citrobacter species (Bornstein and Schafer 2006, Brenner et al. 1993). The Citrobacter strains were: C. freundii [FN997616], C. koseri [AF025372], C. amalonaticus [AF025370], C. farmeri [AF025371], C. youngae [AB273741], C. braakii

[AF025368], C. werkmanii [AF025373], C. sedlakii [AF025364], C. rodentium [AF025363],

C. gillenii [AF025367] and C. murliniae [AF025369]. Escherichia coli O157:H7 [AE005174] was used as an out group strain. The analysis was performed using distance matrix with a gap opening of 10.0 and a gap extension of 0.1 by a server based ClustalW2 program

(http://www.ebi.ac.uk/Tools/msa/clustalw2). A dissimilarity tree was created and viewed using a TreeView X software (Page 1996).

3.3.5. Effects of DON concentration, aeration, substrate, temperature, pH and antibiotic

To determine if DON concentration can affect the growth of the isolated bacteria, tubes containing 5 mL of NB or MSB with 20, 200 or 500 µg/mL of DON were inoculated with strain ADS47 that had been grown for 16 h with continuous shaking at 220 rpm and then

7 adjusted to cell density at 0.25 OD600nm (10 cells/mL). The cultures were incubated at 28°C with continuous shaking at 180 rpm, and 0.5 mL was removed every 12 h to measure

OD600nm. The culture (0.5 mL) was then analysed for DON de-epoxidation activity by LC-

UV-MS (Guan et al. 2009), as described above. Three replicate cultures were tested.

To examine the effect of aeration, five mL of NB or MSB containing 50 µg/mL DON were inoculated with an overnight culture of strain ADS47, prepared as described above, and incubated at 28˚C with continuous shaking at 180 rpm. One mL of culture was removed every six hours, and bacterial growth was determined by measuring cell density at OD600nm. At the

71 same time points, culture samples were analysed for DON to dE-DON transformation ability

(Guan et al. 2009). The experiment was repeated with MSB medium containing 50 µg/mL

DON under anaerobic conditions as described by Islam et al. (2012) with DON to dE-DON transformation ability being determined every 6 h. Three replicate cultures were prepared for each assay.

The effect of substrate on DON de-epoxidation ability was tested after growing

ADS47 for 72 h in NB, 1/10 NB, MM, MS-DON, MSB, MSC, MSY, SE, PDB or LB containing 50 µg/mL DON. For this assay, the media were maintained at pH7.0±0.2 and 28°C for 72 h. The effect of temperature was examined with MSB containing 50 µg/mL DON at 10,

15, 20, 25, 30, 35, 40, 45 or 50°C. The effect of pH on growth of strain ADS47 was examined with MSB adjusted to pH 5.5, 6.0, 6.5, 7.0, 7.5 or 8.0, and the cultures were incubated at 28°C for 72 h. Three replicate cultures were used for each assay.

Overnight cultures of ADS47 in MSB were treated with 0.01% and 0.001% (w/v) of an antibiotic, sodium azide (SA) (Sigma-Aldrich). Controls were prepared in MSB without

SA. Three replicate cultures were spiked with 50 µg/mL DON. DON to dE-DON transformation ability of the bacterial culture was analysed (Guan et al. 2009) after 72 h at

28°C with continuous shaking at 180 rpm. Three replicate cultures were tested.

3.3.6. Localization of DON to dE-DON transformation ability in the bacterium

Bacterial cell extracts were prepared using B-PER II Bacterial Protein Extraction

Reagent following the manufacturer’s instructions (PI 78260, Thermo Scientific, Rockford,

IL, USA). ADS47 was grown in 1.5 mL MSB for 48 h under the growth conditions described above. A culture (0.5 mL) was centrifuged at 5000 x g for 10 minutes, and the pellet was

72 collected. B-PER II reagent (2 mL) was added to the pellet, and the cells were lysed by pipetting the suspension up and down several times. The cell lysate was treated with 20 µL

Halt Protease Inhibitor Cocktail (Thermo Scientific, Rockford, IL, USA). Culture filtrate was prepared by passing 0.5 mL of the culture through a 0.22 µm Millipore filter (Fisher

Scientific). The remaining 0.5 mL culture was used as control. Cell lysate, culture filtrate and control were incubated with 50 µg/mL DON and incubated at 28°C. DON to dE-DON transformation ability of the cell lysate, culture filtrate and control were analysed after 72 h incubation (Guan et al. 2009). Three replicate cultures were tested.

3.3.7. Detoxification of additional trichothecene mycotoxins

Strain ADS47 was evaluated for its ability to degrade five type-A trichothecene mycotoxins (HT-2 toxin, T-2 toxin, T2-triol, diacetoxyscirpenol and neosolaniol) and five type-B trichothecene mycotoxins (nivalenol, verrucarol, fusarenon X, 3-acetyl- deoxynivalenol and 15-acetyl-deoxynivalenol). One and a half mL NB was added with 100

µg/mL of each mycotoxin and then inoculated with 100 µL of ADS47 culture grown overnight as previously described. The culture of NB + mycotoxin + ADS47 was incubated at

28°C with continuous shaking at 180 rpm, and DON to dE-DON transformation ability of the culture was analysed by LC-UV-MS (Guan et al. 2009). Five replicate cultures were tested for each trichothecene mycotoxin.

3.3.8. Statistical analysis

Data were processed using Microsoft Excel Program 2007. Statistical analysis was performed using General Linear Model procedure in SAS version 9.1 (SAS Institute, Cary,

NC, USA, 2002-03). Replications were averaged and stated as mean ± standard error. The mean differences were considered significant at p< 0.05 according to Tukey’s-HSD test.

3.4. Results

3.4.1. Isolation and identification of a DON de-epoxidizing bacterium

A DON de-epoxidizing bacterium, designated strain ADS47, was isolated after 12 transfers over three months, from cultures growing in MSB containing 200 µg/mL DON.

Strain ADS47 produced small to medium size, circular, convex and white colonies on MSA at room temperature under aerobic conditions (Figure 3-1a). The bacterium is rod shaped, and

1.50 x 0.72 µm in size (Figure 3-1b). It has numerous flagella that cover the surface of the cell

(peritrichous flagella) and is Gram-negative.

BlastN analysis of the 16S-rDNA sequence of strain ADS47 showed the most similarity to the eleven known Citrobacter species in the gamma proteobacteria of the

Enterobacteriaceae. In particular the 16S-rDNA nucleotide sequence of ADS47 was more than 99% identical to C. fruendii (Figure 3-2). The next most similar sequences were a cluster of C. werkmanii and C. murliniae, followed by more distantly related C. braakii, C. gilleni and C. youngae. A separate cluster was formed by C. rodentium, C. farmeri, C. sedlakii, C. koseri and C. amalonaticus.

(a) (b)

Figure 3-1. (a) Colony color, shape and size of the mycotoxin degrading bacterial strain

ADS47 on mineral salts agar (MSA) medium. (b) Transmission electron microscopic (TEM) image of the Citrobacter freundii strain ADS47. The bacterium showed multiple flagella

(peritrichous flagella) and numerous fimbriae around the cell surface.

Figure 3-2. Dendrogram based on 16S-rDNA sequences for strain ADS47 and 11 reference

Citrobacter species. An E. coli O157:H7 strain was considered as an out group species. The tree was constructed using a genetic distance matrix by ClustalW 2 and viewed by

TreeViewX software. The distance was converted to scale and shown by branch lengths (scale bar 0.01, which indicated the 16S-rDNA nucleotide sequence variability between the bacterial species). GenBank accession numbers for 16S-rDNA sequences of the species used for comparison are shown in materials and methods section (3.3.4).

3.4.2. Effects of DON concentrations on bacterial DON de-epoxidation

Growth of ADS47, aerobically, in MSB with 20, 200 and 500 µg/mL DON was not significantly different from that observed in control cultures without DON (Figure 3-3a).

Similarly, increasing DON concentrations did not affect bacterial growth in NB medium

(Figure 3-3b).

DON de-epoxidation activity determined as µg/mL DON converted/cell density

(OD600) was measured lowest in MSB+20 µg/mL DON, and highest in MSB+ 500 µg/mL

DON containing cultures (Figure 3-3c). The highest rate of dE-DON formation was 11.5 µg/h in ADS47 cultures in MSB with 500 µg/mL DON between 48 to 72 h. All of the DON was converted in the cultures containing 20 µg/mL DON and 200 µg/mL DON between 48-60 h and 60-72 h, respectively. In 200 µg/mL and 500 µg/mL of DON cultures, the rate of de- epoxidation increased significantly after 36 h.

(a)

1.6 Control 1.2 20 µg/mL

600nm 0.8 200 µg/mL OD 0.4 500 µg/mL

0.0 0 24 36 48 60 72 96

Incubation time (h)

(b)

(C)

Figure 3-3. Effects of DON concentration on bacterial growth and de-epoxidation activities.

Strain ADS47 was grown in MSB and NB media with three different DON concentrations: 20

µg/mL, 200 µg/mL and 500 µg/mL. Control cultures were prepared without DON. Cultures were grown at 28°C with continuous shaking at 180 rpm. (a) Microbial density (OD600 nm) in

MSB with different DON concentrations was measured at 24, 36, 48, 60, 72 and 96 h. (b)

Microbial growth (OD600 nm) in NB having different DON concentration was measured at 24,

48, 72 and 96 h. (c) Bacterial DON de-epoxidation activity in 20, 200 and 500 µg/mL DON containing cultures at different time points. DON de-epoxidation was measured in µg/mL per

OD600 nm. Three replicate cultures for each assay were analysed.

3.4.3. Influence of growth conditions and antibiotic on microbial DON de-epoxydation

Aerobic and anaerobic conditions

ADS47 grew more quickly uner aerobic conditions under MSB+50 µg/mL DON than in MSB+50 µg/mL DON under anaerobic conditions (Figure 3-4a). Aerobic growth was faster in NB+50 µg/mL DON compared to MSB+50 µg/mL DON.

However, the DON de-epoxidation activities, on a cell density basis, were very similar in NB+50 µg/mL DON compared to MSB+50 µg/mL DON (Figure 3-4b). Also, DON de- epoxidation activity of ADS47 under aerobic conditions in MSB+50 µg/mL DON was higher than in the same medium under anaerobic conditions (Figure 3-4b).

(a)

(b)

Figure 3-4. Bacterial growth and DON de-epoxidation capabilities under aerobic and anaerobic conditions. (a) Microbial growth under aerobic conditions in MSB and NB medium and anaerobic conditions in MSB medium. (b) Microbial DON de-epoxidation ability under aerobic conditions in MSB and NB media and anaerobic conditions in MSB medium.

Cultures media having 50 µg/mL DON were inoculated with overnight culture of strain

ADS47. Initial microbial concentration in the cultures was adjusted to 0.25 OD600nm. Cultures were incubated at 28°C having three replications. Bacterial growth at OD600nm and DON de- epoxidation in µg/mL per OD600nm were measured at six hour interval.

Media, temperature and pH

Growth of ADS47 in media containing 50 µg/mL DON under aerobic conditions at 72 h was three-fold higher in NB and LB than in SE, two-fold higher than in 1/10 NB and MSC and did not occur in MS-DON (Figure 3-5a). Growth in MM, MSB and MSY was intermediate.

Over the 72 h DON de-epoxidation per cell density was highest in MSB and MSY and approximately 25% lower in NB and LB and lower by 75% and 87% in 1/10 NB and MSC, respectively (Figure 3-5b). No DON de-epoxidation was observed in MS-DON, in which

DON was used as a carbon source. Bacteria grown in MM broth, which contains sucrose as a sole carbon source were unable to de-epoxidize DON. No DON de-epoxidation was observed in SE and PDB grown cultures.

Growth of ADS47 under aerobic conditions for 72 h in MSB+50 µg/mL DON was greatest between 25-40°C, but no growth occurred at 50°C (Figure 3-6a). DON de- epoxidation activity was greatest and equivalent between 20-40°C and lower at 10, 15 and

45°C temperatures (Figure 3-6b).

ADS47 growth under aerobic conditions for 72 h in MSB+50 µg/mL DON was limited to neutral pHs (i.e., 6.0 - 7.5) (Figure 3-7a). The transformation of DON to dE-DON occurred at the same pH range, and DON de-epoxidation activity was not significantly different between pH 6.0 to 7.5 (Figure 3-7b).

(a)

a a

b b b b c c d

(b)

a a

b b

c d

Figure 3-5. Effects of media on bacterial growth and DON de-epoxidation activity. (a)

Microbial growth (at OD600nm) on different culture media. Media pHs were adjusted to

7.0±0.2. (b) Bacterial DON to dE-DON transformation in µg/mL by OD600nm in different media after 72 h. The cultures having three replications were incubated for 72 h at 28°C and

% DON de-epoxidation was measured. The Media: NB = nutrient broth, 1/NB = 1/10 strength

NB, MM = minimal medium, MS-DON = six mineral salts + 200 µg/mL DON as sole C source. MSB = six mineral salts + 0.5% bacto peptone, MSC = six mineral salts broth + 0.3% bacto peptone, MSY = six mineral salts with 0.5% yeast extract, SE = soil extract, PDB = potato dextrose broth and LB = Luria Bertani. Means with the same label are not significantly different according to Tukey’s HSD test (p < 0.05).

(a)

(b)

Figure 3-6. Effects of temperature on microbial growth and DON de-epoxydation capability.

(a) Microbial growth (at OD600nm) at different temperatures after 72 h of incubation. (b)

Bacteria DON de-epoxidation in µg/mL by OD600nm in different temperatures. The most efficient temperature range (≥40 µg/mL DON de-epoxidation) is marked by dot lines. The experiment was conducted in MSB medium at pH 7.0±0.2 with three replications.

(a)

(b)

Figure 3-7. Influence of pH on microbial growth and DON de-epoxydation. (a) Microbial growth (OD600nm) in MSB at different pH. (b) Bacterial DON to de-epoxy-DON transformation in µg/mL by OD600nm in different pH. Dot lines indicated the efficient de- epoxidation (≥40 µg/mL DON conversion) pH range. Experiment was conducted at 28°C with three replications.

Antibiotic

Treatment of ADS47 at 72 h in MSB+50 µg/mL DON with sodium azide (SA), at

0.001% and 0.01% reduced DON de-epoxidation activity to 5% (2.5 µg/mL) and 0%, respectively. The control with no SA added was able to de-epoxidize DON completely

(100%) (Figure 3-8).

120

100

80 epoxydized - 60

20 % DON de DON % 0 SA1 SA2 C

Figure 3-8. Effects of sodium azide (SA) on bacterial DON de-epoxidation activity.

Microbial culture in MSB+50 µg/mL DON was treated with 0.01% (SA-1) and 0.001% (SA-

2) sodium azide. In control (C), no SA was added. The experiments with three replications were incubated for 72 h at 28°C and % DON de-epoxidation was measured by LC-UV-MS method.

3.4.4. Localization of bacterial DON de-epoxidation

A cell lysate prepared from ADS47 grown for 48 h in MSB transformed 20% (10

µg/mL) of DON added at 50 µg/mL to dE-DON within 72 h (Figure 3-9). In contrast, a cell- free culture filtrate did not show DON de-epoxidation activity.

120

100

epoxydized 60 -

20 % DON de DON % 0 Filtrate Extract Control

Figure 3-9. Determination of DON to de-epoxy DON transformation by bacterial culture filtrate and cell extract. The treatments having three replications were spiked with 50 µg/mL

DON and incubated at 28°C for 72 h. In control, strain ADS47 was incubated in MSB broth with 50 µg/mL DON. DON de-epoxidation activity in percentage was examined.

3.4.5. Bacterial detoxification (de-epoxidtaion and de-acetylation) of other trichothecene mycotoxins

Aerobic cultures of ADS47 grown for 72 h in NB+100 µg/mL HT-2 converted 74% of the toxin to de-epoxy HT-2 toxin, and 26% to a de-acetylated product (Figure 3-10 and Table

3-1).

In a culture treated with T2-toxin 46% was converted to de-epoxy T-2, 16% was converted to an unknown compound and 38% was not converted. In ADS47 culture with diacetoxyscirpenol (DAS), 61% was transformed into de-epoxy DAS and 39% was de- acetylated DAS (39%). Neosolaniol (NEO) was transformed to de-epoxy NEO (55%) and de-acetylated NEO (45%). T2-triol was transformed in to de-epoxy T2-triol (47%) and de- acetylated T2-triol (53%).

Aerobic cultures of ADS47 were also grown for 72 h in NB+100 µg/mL of one of five type-B trichothecenes. For nivalenol (NIV), 92% was de-epoxydized, while the remaining 8%

NIV was transformed to DON. 3-acetyl-deoxynivalenol (3-ADON) was transformed into three products, de-epoxy NIV (5%), de-epoxy DON (54%) and DON (41%). For 15- acetyldeoxynivalenol (15-ADON), 8% was transformed to de-epoxy 15-ADON and 92% to de-epoxy DON (dE-DON). FusarenonX (FUS) had 95% transformed into de-epoxy FUS, while 97% of verrucarol (VER) was transformed to de-epoxy VER. The remaining 5% FUS and 3% VER were not transformed.

Figure 3-10. Determination of microbial de-epoxidation and de-acetylation of HT-2 toxin by

LC-UV-MS. Bacterial culture in NB was added with 100 µg/mL HT-2 toxin. Bacterial HT-2 toxin transformation ability was examined after 72 h of incubation at 28°C. The upper panel illustrated mass spectrometry chromatograph of the de-acetylated (red peak) and de- epoxydized (blue peak) products. Arrows indicated the targeted acetyl (red) and epoxy (blue) functional groups of HT-toxin. The middle and lower panels showed the mass spectra of the de-acetylated (dA) and de-epoxidized (dE) compounds. Arrows on the upper panel indicated the targeted acetyl (red) and epoxy (blue) functional groups of HT-2 toxin. Red and blue arrows on the middle and lower panels pointed at the degradation sites of HT-2 toxin.

Table 3-1. Microbial transformation of type-A and type-B tricthothecene mycotoxins after 72 h of incubation under aerobic conditions at room temperature. DAS=diacetoxyscirpenol,

DON=deoxynivalenol, FUS=FusarenonX, 3AON=3-acetyl-deoxynivalenol and

15ADON=15-acetyl-deoxynivalenol.

Trichothecene Untransformed Transformed products of mycotoxin trichothecenes (%) trichothecenes (%) De- De- Other1 epoxy acetylate HT-2 toxin 0 74 26 - T-2 toxin 38 46 - 16 Type-A (unknown) DAS 0 61 39 - Neosolanol 0 55 45 - T2-triol 0 47 53 - DON 0 100 - - Nivalenol 0 92 - 8 (DON) Type-B 3-ADON 0 59 41 (DON) 15-ADON 0 100 - - FUS 5 95 - - Varrucarol 3 97 - -

1 Convertion of trichothecene mycotoxins to either DON or an unknown compound.

3.5. Discussion

The ADS47 strain of C. freundii was isolated in the present study by a prolonged incubation of a mixed microbial culture with a high concentration DON, since the bacterium efficiently de-epoxidized (detoxified) DON and number of other trichothecene mycotoxins under aerobic and anaerobic conditions at moderate temperatures and neutral pH. The total procedures for enrichment and isolation of the strain ADS47 is depicted in supplemental

Figure S3-1. The reductive de-epoxidation activity may be due to a cytoplasmic enzyme, which would be in agreement with a previous study that showed de-epoxidation of epoxyalkanes was catalyzed by a bacterial cytoplasmic flavoprotein (Westphal et al. 1998).

The ability of the currently isolated strain of C. freundii to biodegrade environmental pollutants, such as aminoaromatic acids and heavy metals (copper), is also likely mediated by cytoplasmic enzymes since these compounds need to be taken up by the cells before they are metabolized (Savelieva et al. 2004, Sharma and Fulekar 2009). In contrast, oxidative transformation of DON to 3-keto DON by microbial extracellular enzyme has been reported

(Shima et al. 1997).

Isolation of a trichothecene de-epoxidizing microorganism from a mixed microbial community is very challenging. Although trichothecenes are toxic to eukaryotes, prokaryotes appear to be resistant to these mycotoxins (He et al. 1992, Rocha et al. 2005). The present study showed that the bacteria grew well in media with high concentrations of DON (500

µg/mL). Bacterial de-epoxidation of DON involves the removal of an oxygen atom from the functional epoxy ring of the tricyclic DON structure (Yoshizawa et al. 1983). Such co- metabolic degradation does not provide the bacteria with any nutrients from the reacting compounds (Horvath 1972, Fedorak and Grabic-Galic 1991). This is supported by the results

96 of the current study, which showed that strain ADS47 was unable to grow in media with DON as a sole carbon source and increasing the concentration of DON did not increase ADS47 growth in MSB. Hence, DON could not be used as sole nutrient source in the mixed culture enrichment process.

Many previous attempts to obtain DON de-epoxidizing microbes have failed, particularly under aerobic conditions, because media or laboratory conditions were inappropriate to support growth or initially active cultures lost de-epoxidizing function after serial transfer (Karlovsky 2011).

Media composition appears to be a crucial factor for microbial trichothecene de- epoxidation activity (Völkl et al. 2004). Effective biotransformation of trichothecene mycotoxins by the fish digesta microbial culture C133 occurred in full media (FM) but a low level of de-epoxidation occurred in media having a single nitrogen source and no transformation activity was observed in the common bacterial medium, LB (Guan et al.

2009). Zhou et al. (2010) also showed that chitin de-acetylation by soil bacterial strains occurred more rapidly in yeast extract, compared to common bacterial media. Therefore, a screening strategy for microbial DON de-epoxidation ability should use different media in parallel, as was done in this study.

This study observed efficient de-epoxidation in common bacterial media (NB and LB) as well as in mineral salts medium containing yeast extract (MSY) or 0.5 % Bacto Peptone

(MSB). However, reduction of Bacto-Peptone to 0.3% (MSC) significantly reduced the bacterial de-epoxidation activity, suggesting that the activity is sensitive to the availability of protein and amino acids. The study found that medium containing sucrose as carbon source

(MM) and a commonly used medium PDB blocked microbial DON de-epoxidation, even

97 though bacteria grew in those growth media. These suggest that media containing high sugar contents had inhibitory effects on microbial de-epoxidation enzymatic activity. Previously, sugars have been found to have negative effects on the production and/or activity of bacterial enzymes (Saier and Roseman 1971).

Losses of microbial functions under laboratory conditions limited researchers to obtain economically important microbial genetic sources. The soil borne Curtobacterium sp. strain

114-2 efficiently de-epoxydized nivalenol, but unfortunately, the strain lost its activity in latter generations (Ueno et al. 1983). Instability of de-epoxidation activity of ADS47, observed in cultures grown on agar media without DON, resulted in the loss of DON de- epoxidation activity by more than 80% of the colonies after one generation. However, strain

ADS47 retained de-epoxidation activity after ten subcultures in MSB (data not shown).

The non-specific de-epoxidation of DON made it impossible to use DON as a sole substrate C source for selective isolation of ADS47. Thus, a number of techniques were utilized to select ADS47 and suppress other bacteria, including the use of a variety of antibiotics and culture at elevated temperatures (Islam et al. 2012). Selection of the isolate

ADS47 required repeated subculturing and prolonged exposure to media containing 200

µg/mL DON. After three months of selection with DON, culture that was obtained had a

DON de-epoxidation activity that was up to 30-fold higher than the mixed culture. As growth of ADS47 was not affected by 200 µg/mL DON, the prolonged incubation may have reduced populations of other bacteria that are more sensitive to DON. Enrichment and isolation of a

DON to 3-keto DON transforming bacterial strain E3-39 by prolonged incubation of a soil microbial culture with 200 µg/mL DON was reported by Shima et al. (1997).

This study demonstrated that microbial DON de-epoxidation occurred mainly at the stationary growth phase. A DON degrading strain of Citrobacter may have gained energy from DON by co-metabolic cleavage of the epoxy ring under less nutrient conditions.

Microbial use of energy from co-metabolic reaction has been suggested by Chiu et al. (2005).

The rate of de-epoxidation activity increased when ADS47 grew at higher concentrations of

DON. Taken together, a prolonged exposure to high concentration DON may have resulted in non-DON metabolizing microbial populations declining by DON toxicity, while facilitated the population of DON metabolizing bacterium by providing energy.

DON is a small molecule and is simply absorbed by cells (Sergent et al. 2006).

Microbial DON de-epoxidation occurred in intact and lysed bacterial cells. In addition, treating the ADS47 culture with SA, which blocks gram negative bacterial electron transport and activity, inhibited DON de-epoxidation activity (Lichstein and Malcolm 1943, Erecinska and Wilson 1981), implying that an active electron transport system required for microbial de- epoxidation reaction (He et al. 1992). The data suggest that a cytoplasmic epoxy reductase is involved in the DON de-epoxidation reaction in the ADS47 microbial strain that was isolated in the present study.

Since agricultural commodities are often contaminated with multiple fungal mycotoxins, the ability of bacteria, like ADS47, to degrade mycotoxins by several mechanisms may be an advantage. De-acetylation might be another mechanism of microbial mycotoxin detoxification, since toxicity of trichothecenes also relies on the acetyl functional groups (Desjardins 2006). Strain ADS47 showed efficient de-epoxidation and de-acetylation of several type-A and type-B trichothecene mycotoxins that are commonly detected in cereals in association with DON. In conclusion, the soil-borne gram-negative bacterium, C. freundii

99 strain ADS47 is capable of detoxifying trichothecene mycotoxins by both de-epoxidation and de-acetylation under aerobic and anaerobic conditions, neutral pH and a range of temperatures. ADS47 has great potential to detoxify DON contaminated animal feed.

Furthermore, isolation of novel bacterial de-epoxidation genes might be used to generate transgenic plants that can detoxify DON in contaminated seed. Alternatively, contaminated seed might be treated with manufacture de-epoxidation enzyme(s). These applications can improve the quality and safety of agricultural commodities that are prone to trichothecene mycotoxin contamination.

3.6. Acknowledgements

Thank to the Natural Sciences and Engineering Research Council of Canada (NSERC) for a scholarship to Rafiqul Islam. The work was funded by the Agriculture and Agri-Food Canada/

Binational Agricultural Research and Development Fund (AAFC/BARD) project. We appreciate Honghui Zhu for technical assistance.

100

Step 1: Screening of soil Obtained DON de-epoxidation capable microbial samples culture obtained by analyzing soil samples collected from cereal fields

Step 2: Optimization of Optimized the culture conditions, i.e. growth substrates, cultural conditions temperatures, pH and oxygen requirements

Step 3. Elimination of Serially diluted active culture was grown on agar media. Collected undesired microbial washed colonies having DON de-epoxidation activity species

Step 4. Removal of Active culture was sequentially treated with different mode of actions untargeted microbial antibiotics, heating cultures at 50°C and subculturing on agar media species

Step 5. High The partially enriched culture that retained DON de-epoxidation activity concentration was incubated with 200 µg/mL DON in MSB media with continuous DON enrichment shaking at 220 µg/mL for 7 d. The procedure was repeated 12 times

Step 6. Separation of The enriched culture was serially diluted and treated dilutes with 0.5 % bacterial cells and Tween 20. After vigorously vortex for ten min, aliquot cultures were spread dilution plating on different agar media plates to produce single colonies

Step 7. Single Colonies were tested for their DON de-epoxidation ability. Purified colony analysis the single colony via several times subculturing on agar medium

Supplemental Figure S3-1. Flow chart depicts the procedures of enrichment and isolation of

DON de-epoxidation bacterial strain ADS47. Steps 1 to 4 describe the optimization of

microbial DON to dE-DON transformation activity and removal of undesired microbial

species. Step 5 to 7 shows the culture enrichment by incubation with high concentration DON

and isolation of the single strain ADS47.

101

CHAPTER 4

4.0. CONSTRUCTION AND FUNCTIONAL SCREENING OF GENOMIC DNA

LIBRARY FOR ISOLATION OF DON DE-EPOXYDIZING GENE(S) FROM

CITROBACTER FREUNDII STRAIN ADS47

4.1. Abstract

Enzymatic activity-based screening of a genomic DNA library is a robust strategy for isolating novel microbial genes. A genomic DNA library of a DON detoxifying bacterium

(Citrobacter freundii strain ADS47) was constructed using a copy control Fosmid vector

(pCC1FOSTM) and E. coli EPI300-T1R as host strain. A total of 1250 E. coli clones containing large DNA inserts (~40 kb) were generated, which resulted in about 10-fold theoretical coverage of the ADS47 genome. The efficiency of E. coli transformation with the recombinant Fosmid vector was examined by restriction digestion of Fosmids extracted from

15 randomly selected library clones. The fragmentation patterns showed that all of the clones contained an 8.1 kb Fosmid backbone and ~40 kb of insert DNA, except one that had a small insert. This suggests that a DNA library was prepared with expected DNA inserts. The de- epoxidation ability of the library was analysed by incubating clones in Luria Bertani (LB) broth medium supplemented with 20 µg/mL DON with induction (by addition of L-arabinose) or without induction. None of the 1000 clones showed DON de-epoxidation activity. The potential DON de-epoxidation gene(s) of ADS47 may not have been expressed in E. coli or might have been expressed but lacked additional factors required for DON de-epoxidation.

102

Keywords: DNA library, fosmid vector, functional library screening, DON de-epoxidation activity.

4.2. Introduction

Microorganisms contain numerous valuable genes and biocatalysts, which may have significant roles in agricultural productivity (Jube and Borthakur 2007). Various techniques have been used to isolate novel genes from different organisms, including transposon mutagenesis (Frey et al. 1998, Tsuge et al. 2001), gene knockout (Winzeler et. al. 1999), whole genome sequencing (Krause et al. 2006), comparative genomics (Koonin et al. 2001) and suppression subtractive hybridization (Dai et al. 2010).

The genomic DNA library preparation and activity-based screening of the library is a relatively new technique, which has been efficiently used for isolation of microbial novel enzymes and compounds, such as lipases, proteases, oxidoreductases, antibiotic resistance genes, membrane proteins and antibiotics (Henne et al. 2000, Gupta et al. 2002, Knietsch et al. 2003, Gabor 2004, Suenaga et al. 2007, van Elsas et al. 2008). The strength of the technique is that genes from microbial sources could be obtained without prior sequence information, and not requiring culturing the organisms under laboratory conditions. Using this technique, an unknown gene or a set of genes could be obtained by determining particular phenotypes of the library clones. Eventually, the gene of interest from the positive clones is isolated by using a forward (e.g., mutation) or reverse genetics method (e.g., subcloning the positive clones into a microbial expression vector), followed by functional analysis of the subclones.

103

Although this strategy has been used effectively for isolating novel microbial genes, several factors need to be considered in order for the method to be successful. These factors include DNA quality, vector type (e.g., plasmid, cosmid, fosmid, bacterial artificial chromosome etc.), vector copy number, and host strains (Daniel 2005, Singh et al. 2009). The

Fosmid vector-based technique has been used successfully to clone bacterial genes, including xenobiotics degradation genes, antifungal gene clusters and antibiotic resistance genes from soil metagenomes (Hall 2004, Suenaga et al. 2007, Chung et al. 2008). This study reports the construction and activity-based screening of a Fosmid DNA library in an E. coli host with the aim of isolating a novel DON de-epoxidation gene from Citrobacter freundii strain ADS47

(Chapter 3).

4.3. Materials and methods

4.3.1. Genomic DNA library construction

The genomic DNA library was prepared using a Copy Control Fosmid Library

Production Kit (Epicentre Biotechnologies, Madison, Wisconsin, USA) following the manufacturer’s instructions as depicted in Figure 4-1.

104

Figure 4-1. Schematic overview of the construction and functional screening of a Fosmid

DNA library.

105

Citrobacter freundii strain ADS47 was grown in NB medium at 28°C with continuous shaking at 220 rpm. The high molecular weight bacterial genomic DNA was extracted from the overnight culture using the Gentra Puregene Yeast/Bacteria Kit (Qiagen Inc., Mississauga,

Ontario, Canada). The purity of the DNA was confirmed by determining OD260/280 at ~1.90 by a NanoDrop spectrophotometer. A 5 µL sample having ~2.5 µg high molecular weight DNA was sheared by passing through a 200-µL pipette tip and achieved a sufficient amount of

DNA about 40 kb in size. Blunt ended, 5’-phosphorylated insert DNA was generated by thoroughly mixing it with End-Repair Enzyme reagents on ice. The 20 µL mixture was incubated for 45 min at room temperature, followed by incubation at 70°C for 10 min to inactivate the End-Repair Enzyme activity.

The treated DNA was size fractionated by a Low Melting Point-Pulsed Field Gel

Electrophoresis (LMP-PFGE) on the clamped homogeneous electric fields (CHEF-DR III) system (Bio-RAD Laboratories) using 1% UltraPure Low Melting Point Agarose (Invitrogen).

The PFGE was conditioned according to manufacturer’s instructions, as follows: initial switch time-42 s, final switch time-7.61 s, run time 16 h 50 min, angle-120°, gradient- 6 V/cm, temperature-14°C, and running buffer-0.5% TBE. DNA size markers for the 42 kb Fosmid control DNA (Epicentre Biotechnologies) and low range Pulse Field Gel (PFG) markers (New

England Biolabs Inc.) were loaded adjacent to the DNA sample slots. After electrophoresis, a portion of the gel that contained DNA size markers was cut off. The gel was stained in SYBR

Safe DNA gel stain (Invitrogen) for 30 min, visualized and marked on the 23, 42 and 48.5 kb marker bands. A slice of PFGE gel containing 40 ± 10 kb size DNA was cut off with a sterile razor blade by comparing the marker sizes. DNA from the gel was recovered by digesting it with GELase enzyme (@ 1 U/100 µL melted agarose) in 1X GELase Buffer, followed by a

106 sodium acetate-ethanol precipitation. The DNA with the desired size was stored at 4°C for next day’s ligation to a cloning-ready Fosmid vector (pCC1FOSTM).

Ligation was performed in a 10 µL reaction mixture by thoroughly mixing 10X Fast-

Link Ligation Buffer, 10mM ATP, 0.25 µg of ~40 kb insert DNA, 0.5 µg Fosmid Copy

Control Vector (pCC1FOS) (Figure 4-2) and 2 U Fast-Link DNA lipase. The mixture was incubated for 4 h at room temperature, preceded by inactivating the ligation reaction by heating at 70°C for 15 min. The recombinant Fosmid vector was stored at -20°C for later use.

Ten µL of the ligated DNA was thawed on ice and transferred to a pre-thawed

MaxPlax Lambda Packaging Extract tube containing 25 µL extract. By pipetting several times, the mixture was incubated at 30°C for lamda packaging. After 2 h of packaging, an additional 25 µL packaging extract was added into the mixture and incubated for a further 2 h at 30°C. The packaged phage particles were volumed to 0.5 mL by adding ~400 µL Phage

Dilution Buffer and 25 µL chloramphenicol.

Transformation was performed via transduction by mixing the packaged particles with the overnight cultured E. coli EPI300-T1R cells (1: 10 v/v ratio) in LB supplemented with 10 mM MgSO4 and 0.2% Maltose. Transformation was completed by incubation at 37°C for 1 h.

The transformed E. coli cells were spread on an LB agar plate containing 12.5 µg/mL chloramphenicol. The plates were incubated overnight at 37°C. Colonies that appeared on the plates were individually transferred to 96-well microtiter plates having 200 µL LB medium supplemented with 10% glycerol and 12.5 µg/mL chloramphenicol. The microtiter plates were incubated at 37°C for 24 h with continuous shaking at 180 rpm. The prepared library clones were stored at -80°C for further analysis and functional screening.

107

108

Figure 4-2. Genetic map of copy controlled Fosmid vector (pCC1FOSTM). The 8.13 kb

Fosmid vector backbone shows a multiple cloning site (MCS), chloramphenicol resistance selectable marker (CmR), replication initiation fragment (redF), inducible high copy number origin (oriV), replication initiation control (repE) and F-factor-based partitioning genes (ParA,

ParB, ParC). The vector contains a cos site for lamda packaging (not shown) and lacZ gene for blue/white colony selection. The vector map was provided by the Epicentre

Biotechnologies (http://www.epibio.com/item.asp?id=385).

109

4.3.2. Confirmation of library construction

The quality of the DNA library prepared was determined by examining 15 E. coli clones randomly selected. A 1.5 mL LB broth supplemented with 12.5 µg/mL chloramphenicol was inoculated by each of the transformed clones. To obtain higher Fosmid

DNA yield, 1 µL autoinduction solution (L-arabinose) to increase the copy number of the

Fosmid vector (Epicentre) was added to the cultures. After an overnight incubation of the cultures at 37°C with continuous shaking at 220 rpm, Fosmid DNA was extracted using a

Qiagen Large-Construct Kit, followed by restriction digestion with Not I enzyme. A 20 µL digestion mixture containing 0.5 µg Fosmid DNA, 1X digestion buffer and 0.5 U Not I was incubated at 37°C for 1.5 h. The partially digested DNA and a 42 kb DNA size marker

(Fosmid control DNA) were ran on 0.8% agrose gel for 1.5 h at 100 V. The linear Fosmid vector backbone (8.13 kb) and insert DNA fragments were visualized on the gel and the image was recorded.

4.3.3. Functional screening of the clones

Eppendorf tubes (1.5 mL) containing 500 µL LB supplemented with 12.5 µg/mL chloramphenicol were inoculated with a single clone (400 clones tested) or a combination of

12 library clones (600 tested) from microtiter plates stored at -80°C. One half of the cultures in each group (single or combined clones) were induced to increase the Fosmid copy number by adding a 1 µL (500X) autoinduction solution (contained L-arabinose) to the culture tubes.

L-arabinose induces the expression of a mutant trfA gene of the E. coli host strain, which leads to the initiation of Fosmid replication from a low-copy to a high-copy number origin of replication (oriV). The cultures were spiked with 20 µg/mL DON and incubated at 37°C for

110

72 h with continuous shaking at 180 rpm. The DON to de-epoxy DON transformation ability of the clone cultures was examined by a liquid chromatography-ultraviolet-mass spectrometry

(LC-UV-MS) method as described by Islam et al. (2012).

111

4.4. Results

4.4.1. Production of the DNA library

High molecular weight, intact (non-degraded) and good quality (OD600/280 ~1.90) genomic DNA was obtained from the DON de-epoxidizing bacterial strain ADS47 (data not shown). Pipetting of the DNA extract resulted in an adequate amount of insert DNA (>0.25

µg) of approximately 40 kb size as showed on a LMP-PFGE (Figure 4-3).

A DNA library in E. coli host was constructed using LMP-PFGE purified DNA. The library was comprised of 1250 E. coli transformants, which were selected on LB agar plates supplemented with chloramphenicol (Figure 4-4).

112

Figure 4-3. Preparation of a 40 kb size insert DNA from a DON de-epoxidizing bacterium.

Genomic DNA was extracted from the overnight culture of a DON de-epoxidizing

Citrobacter freundii strain ADS47. After fragmentation, the DNA was size fractionated by

Low Melting Point- Pulse Field Gel Electrophoresis (LMP-PFGE) with markers for the PFGE

(M1) and 42 kb Fosmid control DNA (M2). S1, S2, S3= different DNA samples extracted from bacterial strain ADS47. The red dots area indicates the section of the gel containing genomic DNA of the expected size (~40 kb).

113

Figure 4-4. Production of the genomic DNA library in E. coli EPI300-T1 host strain.

Approximately 40 kb long blunt ended genomic DNA from a DON de-epoxidizing bacterial strain ADS47 was ligated to a Copy Control Fosmid vector, pCC1FOS. The construct was packaged into lamda bacteriophage, followed by a transduction-mediated transformation of E. coli host cells. The clones, which appeared on the LB agar plate supplemented with 12.5

µg/mL chloramphenicol are expected to be transformed by the recombinant Fosmid vector containing insert DNA.

114

4.4.2. Conformation of the library construction

All of the clones tested acquired the Fosmid vector. Out of 15 clones examined, 14 received the recombinant Fosmid vector that contain the expected size insert DNA, as determined by an 8.1 kb vector backbone and at least one high molecular weight insert DNA band (~40 kb) on agarose gel (Figure 4-5). In contrast, one clone (lane 4) showed an 8.1 kb vector backbone and a low molecular weight band, which is also a fragment of the Fosmid vector. The result indicated that approximately 94% of the library clones were transformed with the recombinant vector.

4.4.3. Activity-based screening of the library

Out of 500 single copy number clones (200 individual and 300 combined clones), none showed DON to de-epoxy DON transformation activity (Table 4-1). An additional 500 high copy number clones (200 single and 300 combined clones) were unable to de-epoxidize

DON as well.

115

Figure 4-5. Confirmation of the Fosmid DNA library production in E. coli. Fosmid DNA collected from 15 clones was digested with Not I and run on 0.8% agarose gel at 100 V for 1.5 h. The presence of the vector backbone (8.1 kb) and the insert DNA fragment (~40 kb) in the clones was determined by comparing a 42 kb molecular size marker (M): Lanes (1-15) denote the recombinant Fosmid DNA extracted from 15 E. coli clones.

116

Table 4-1. Analysis of Fosmid DNA library clones for deoxynivalenol (DON) de-epoxidation capability*.

Type of clones Analysis of Number of Number of clones single/combined clones showed DON de- clones examined epoxidation activity Without induction Singly 200 0

(single copy number clones) Combination of 12 300 0

With induction (multi-copy Singly 200 0

number clones) Combination of 12 300 0

*De-epoxidation activity of the clones was examined alone or in combination of 12 clones without or with induction by adding L-arabinose. The clones grown in LB supplemented with

12.5 µg/mL and 20 µg/mL DON were analysed for DON to de-epoxy DON transformation ability by a liquid chromatography-ultra violet- mass spectrum (LC-UV-MS).

117

4.5. Discussion

While a Fosmid DNA library with the expected genome coverage of a DON de- epoxidizing bacterial genome was produced, the function-derived screening of the library clones did not identify any clones that showed DON de-epoxidation enzymatic activity. This is unlikely due to lack of coverage of the bacterial genome, as 1000 clones analyzed in the study should have provided eight fold coverage of the genome. Therefore, it was likely due to a lack of expression of the enzyme(s) for DON de-epoxidizing activity under the conditions used.

Thus far, there is only one report of heterologous expression of bacterial epoxide reductase gene in another bacterium. A high level of heterologous expression of a

Synechococcus sp. vitamin K epoxide reductase gene was achieved in E. coli with codon optimization and placing the gene under the control of a prokaryotic T7 promoter in an expression vector (Li et al. 2010). Therefore, heterologous expression of bacterial epoxide reducatse gene using its own promoter has not yet been demonstrated. The success of the method used in this chapter relies on an intact copy of the gene or operon being cloned, the selection of an appropriate cloning vector, a host organism that can recognize the cloned C. fruendii promoter, and a suitable library screening method for DON degradation by E. coli

(Mocali and Benedetti 2010).

The unbiased cloning efficiency of the Fosmid vector, its ability to contain large insert

DNA, high stability in E. coli as single copy number and autoinduction to increase copy number in host cells made it appear to be a suitable method for the present work (Kim et al.

1992, Lee et al. 2004). Library production with large insertions is a major advantage, as gene cluster involved in particular pathway could be trapped into single clone and their functions

118 could be analysed. The pathway of microbial DON de-epoxidation activity and the genes encoding the activity are not known. Therefore, it was reasonable to prepare a library with large inserts to ensure the possibility of cloning single as well as whole pathway genes of interest for functional screening. A further advantage of the method is that Fosmid clones are inducible, which allows detecting poorly expressed genes in a heterologous host by inducing high copy numbers (Daniel 2005).

One of the limitations of the method is that the functions of the target genes rely mostly on the host organism’s transcription and translation machineries. Therefore, selection of the host organism is critical for activity-based screening of library clones in heterologous hosts. A number of host strains, including E. coli, Streptomyces lividans, Rhizobium leguminosarum and Pseudomonas aeruginosa, have been used for functional screening activities from diverse organisms (Daniel 2005). The present work attempted to isolate C. freundii strain ADS47 genes encoding DON de-epoxidation enzymes. Both strain ADS47 and

E. coli belong to gamma proteobacteria, which suggest that E. coli would be an appropriate host organism for heterologous expression of ADS47 genes. Poor expression of heterologous genes could be another limitation of the method, which may be overcome by overexpression to increase the catalytic activity of less expressed genes. However, clones with DON de- epoxidation activity were not identified in this study even the Fosmid copy number was increased by induction.

In general, a large number of clones must be analysed to obtain the gene of interest using an activity-based screening technique (Lee et al. 2004, van Elsas et al. 2008). In this study, the constructed DNA library comprised of 1160 clones having approximately ~40 kb insert DNA, suggesting that the prepared library achieved ~10 fold coverage of the ADS47

119 genome (4.8 Mb), which ensured >99% probability of the presence of all DNA sequence in the library. In the produced library, three clones (lanes 3, 10 and 12) showed one intact band greater than 42 kb in size on the agarose gel. It is presumed that the Fosmid DNA of the three samples was not digested because of insufficient amount of restriction enzyme added to the digestion mixture. About half of the samples demonstrated multiple bands, suggesting that the insert DNA might have contained multiple Not I restriction sites.

However, not every heterologous gene may necessarily express or function in the E. coli host (Gabor et al. 2004). The functionality of the gene requires expression of the genes and proper folding of the protein, which may be influenced by environmental conditions (e.g., substrate, temperature and pH) and genetic factors. In this study, library clones were screened in growth conditions at which strain ADS47 performed the best de-epoxidation activity. The expression of genes may be controlled by the promoter, suitable ribosome binding site, and trans–acting factors (e.g., special transcription factors, inducers, chaperones, cofactors, proper modifying enzymes or an appropriate secretion system) (Gabor et al. 2004). Those factors have to be supplied by the cloned DNA fragments or host cells to allow any heterologous gene to be functional. Since DON de-epoxidation genes of strain ADS47 evolved in a distinct environmental condition, the expression and activity of the gene(s) might have required specific factors/regulators that were absent in E. coli. Even though the Fosmid based library screening technique facilitated isolation of many microbial genes, clones having de- epoxidation activity were not identified using such technique in this study.

120

4.6. Acknowledgements

We appreciate the Agriculture and Agri-Food Canada/Binational Agricultural

Research and Development Fund (AAFC/BARD) project for providing the financial support for this work. The authors thank Honghui Zhu for analyzing numerous samples by liquid chromatography ultraviolet mass spectrum method.

121

CHAPTER 5

5.0. GENOME SEQUENCING, CHARACTERIZATION AND COMPARATIVE

ANALYSIS OF CITROBACTER FREUNDII STRAIN ADS47

5.1. Abstract

The genome of a trichothecene mycotoxin detoxifying C. freundii strain ADS47, originating from soil, was sequenced and analyzed. Strain ADS47 contains a circular chromosome of 4,878,242 bp that encodes 4610 predicted protein coding sequences (CDSs) with an average gene size of 893 bp. Rapid Annotation Subsystems Technology (RAST) analysis assigned functions to 2821 (62%) of the CDSs, while 1781 (38%) CDSs were assigned as uncharacterized or hypothetical proteins. Comparative bioinformatics analyses were done for strain ADS47, C. freundii Ballerup 7851/39, Citrobacter sp. 30_2, C. youngae

ATCC 29220, C. koseri ATCC BAA-895 and C. rodentium ICC168. There were a significant number of common orthologous genes (2821) were found in the six Citrobacter genomes, implying a common evolutionary relationship. ADS47 is most closely related to the opportunistic human pathogen, C. freundii Ballerup. The human and mouse pathogenic species, C. rodentium and C. koseri are the most distantly related to ADS47, and non- human/animal pathogens, C. youngae and Citrobacter sp. 30_2 were closely related. ADS47,

C. koseri, C. youngae and Citrobacter sp. 30_2 lacked the clinically important locus for enterocyte effacement (LEE)-encoded type-III secretion systems, which are present in C. rodentium and C. freundii Ballerup. Strain ADS47 had 270 unique CDSs, many of which are located in the large unique genomic regions, potentially indicating horizontal gene transfer.

122

Among the unique CDSs nine were reductases/oxidoreductases and a de-acetylase that may be involved in trichothecene mycotoxin de-epoxidation and de-acetylation reactions.

Keywords: Citrobacter freundii, comparative genomics, genome sequencing, unique genes, reductases.

5.2. Introduction

According to the Food and Agriculture Organization (FAO), at least 25% of the world’s food crops are contaminated with mycotoxins, causing serious negative impacts on livestock and international trade (Boutrif and Canet 1998, Desjardins 2006, Binder et al.

2007). Trichothecenes (TCs), produced by a variety of toxigenic Fusarium species, are probably the most prevalent food and feed contaminating mycotoxins. It is estimated that infections by the fungus caused more than $1 billion in crop yield losses in USA and Canada alone (Windels 2000, Schaafsma et al. 2002, Yazar and Omurtag 2008). TC mycotoxins can have adverse affects on human health, including organ toxicity, mutagenicity, carcinogenicity, neurotoxicity, modulation of the immune system, negative reproductive system affects and growth retardation (D’Mello et al. 1999, Pestka et al. 2007). A strain of Citrobacter freundii,

ADS47, was isolated from soil in southern Ontario that is capable of detoxifying (i.e., de- epoxidation and de-acetylation) several TC mycotoxins, including deoxynivalenol, nivalenol and T2-toxin (Chapters 2-3).

Citrobacter freundii is a gram negative, motile, rod shaped, facultatively anaerobic bacterium belonging to the family of Enterobacteriaceae (Wang et al. 2000, Whalen et al.

2007). The opportunistic human pathogen is responsible for a number of infections, including

123 nosocomial infections of respiratory tract, urinay tract, blood, and causes neonatal meningitis

(Badger et al. 1999, Wang et al. 2000). C. freundii exists almost ubiquitously in soil, water, wastewater, feces, food and the human intestine. Biodegradations of tannery effluents, copper, aromatic amino acids by C. freundii strains isolated from water, mine tailings and soil, respectively have been previously reported (Kumar et al. 1999, Savelieva et al. 2004, Sharma and Fulekar 2009). Because of their biochemical capabilities, bacteria have played important roles in industrial biotechnology, including de-epoxidation and de-acetylation activities

(Demain and Adrio 2008, Plank et al. 2009, Zhao et al. 2010). However, the identification of genes for de-epoxidation activity against trichothecene mycotoxins is limited (Boutigny et al.

2008).

The advent of molecular techniques, such as next generation sequencing technologies and the related bioinformatics tools, has enabled the discovery of novel genes and the enzymes they encode in pathogenic and beneficial bacteria (van Lanen and Shen 2006,

Ansorge 2009, Zhao 2011). In the present study, the genome of mycotoxin-detoxifying strain

C. freundii ADS47 was sequenced and subjected to comparative genomic analyses with other

Citrobacter species to identify putative mycotoxin-detoxifying genes.

124

5.3. Materials and methods

5.3.1. Sequencing, quality control and assemble of ADS47 genome

Genomic DNA of the strain ADS47 was extracted using Gentra Puregene

Yeast/Bacteria Kit (Qiagen Inc., Mississauga, Ontario, Canada) and purity of the DNA was confirmed by determining OD 260/280 at ~1.90 using a Nanodrop spectrophotometer.

The genomic DNA was sent to be sequenced by the Illumina HiSeq 2000 sequencing analyzer at Beijing Genomics Institute (BGI), Hong Kong

(http://www.bgisequence.com/home/services/sequencing-services/de-novo-sequencing/). The sequencing workflow involved short read library preparation, cluster generation, sequencing and data analysis. Briefly, 3-5 ug of bacterial gDNA was fragmented to produce a 500 bp insert library by preparative electrophoresis through a 0.8% agarose gel. The fragmented

DNAs were blunt-ended by treating with a mixture of enzymes (T4 DNA polymerase, klenow fragment) and, subsequently, adapters were ligated to the blunt ends. Each of the DNA fragments was amplified to a cluster, which was performed in the flow cell by template hybridization and bridge amplification. Millions of clusters were then sequenced by sequencing-by-synthesis technology leading to base calls. The raw data was filtered to remove low abundant sequences from the assembly according to 15-mer frequencies. The low complexity reads, low quality (≤Q20) base calls and adapter contaminations (24 bp overlap between adapter and reads) were eliminated. The quality of the sequences was analysed by

GC content and depth correlative analysis (for GC biasness). K-mer analysis was performed using 15 bp of the short reads for heterologous sequence identification.

125

The cleaned short reads (approximately 90 bp) were self-assembled using the Short

Oligonucleotide Analysis Package (SOAP) denovo (http://soap.genomics.org.cn) (Li et al.

2008). The assembly errors were corrected by filling gaps and proofreading single base pairs using mapping information. The assembly produced 441 contigs with an N50 contig size of

29,222 bp (the largest contig size 114,839 bp) and 80 scaffolds with an N50 scaffold size of

325,726 bp (largest scaffold 662,452 bp) with an total sequence length of 4,878,242 bp, which constitutes 33-fold genome coverage. The scaffolds were ordered and oriented according to a reference genome, C. koseri ATCC BAA-895, the closest completely sequenced species of the newly sequenced bacterial genome, by alignment, with the MUMmer program v3.22

(http://mummer.sourceforge.net/) (Delcher et al. 2002). Genome coverage was determined by mapping the reads to the assembled genome with Bowtie software (Langmead et al. 2009). A preliminary genome annotation was performed to define the origin of replication by identifying colocalization of multiple genes, i.e., gyrB, recF, dnaN and dnaA (usually found near the origin of replication) and comparison with the C. koseri reference genome. Synteny analysis between the reference bacterium and assembled scaffolds was performed with the

MUMmer program.

5.3.2. Gene function annotation and classification

The genome of ADS47 was annotated by Rapid Annotation using Subsystems

Technology v4.0 (RAST), a server-based genome analysis system (http://rast.nmpdr.org/)

(Aziz et al. 2008). A circular genome map was constructed using CGView program

(http://stothard.afns.ualberta.ca/cgview_server/) (Stothard and Wishart 2005).

126

5.3.3. Molecular phylogenetic analysis

A robust phylogenetic tree was produced using Composition Vector Tree (CVTree) web-based program v2.0 (http://cvtree.cbi.pku.edu.cn/) (Qi and Hao 2004) by comparing whole genome amino acid sequences of strain ADS47, two completely sequenced Citrobacter genomes [C. rodentium ICC168 (NC_013716) and C. koseri ATCC BAA-8959 NC_009792)] and three partially sequenced Citrobacter genomes [C. freundii strain Ballerup 7851/39

(CACD00000000), Citrobacter sp. 30-2 (NZ_ACDJ00000000) and C. youngae ATCC 29220

(NZ_ABWL00000000)]. E. coli O157:H7 strain EDL933 (NC_002655) was used as an outgroup species. The tree was constructed from the generated pairwise distance matrix data using Neighbor-Joining Program in PHYLIP v3.69 package (Saitou and Nei 1987, Felsenstein

1989).

5.3.4. Comparative genomic analysis

Genome-wide comparative analyses between C. freundii ADS47, C. freundii strain

Ballerup 7851/39, Citrobacter sp. 30-2, C. youngae ATCC 29220, C. Koseri ATCC BAA-895 and C. rodentium ICC168 was carried out. The nucleotide sequence idendities were determined by pairwise alignments of the ADS47 genome with the five other Citrobacter genomes using MUMmer software (Delcher et al. 2002), and GC content, total CDSs, genome coverage by CDSs and number of stable RNAs were determined by RAST program.

Orthologous CDS groups among the total set of 29,079 CDSs derived from the six

Citrobacter genomes were determined using OrthoMCL software v2.02

(http://orthomcl.sourceforge.net/) based on all-against-all reciprocal BLASTP alignment of amino acid sequences (Li et al. 2003). Genes unique to each strain were identified as those

127

CDSs not included in any orthologous group of the five other Citrobacter genomes. BLASTP comparisons were visualized with Blast Ring Image Generator (BRIG) software v0.95

(http://sourceforge.net/projects/brig/) (Alikhan et al. 2011).

Protein sequences of the five Citrobacter genomes pooled from the GenBank data bases were compared with ADS47. A circular map was developed by BRIG software that showed the unique genomic regions/genes of ADS47. The unique genes were grouped according to their functional similarities previously identified by RAST software, and genomic regions possibly acquired by horizontal transfer were identified using RAST software.

128

5.4. Results

5.4.1. Assembly and organization of the ADS47 bacterial genome

Scaffolds derived from Citrobacter freundii strain ADS47 were ordered and oriented by alignment to Citrobacter koseri ATCC BAA-895, the most closely related genome that has been completely sequenced. The genomes of ADS47 and C. koseri displayed close synteny

(Figure 5-1).

Figure 5-1. Co-linearity between C. freundii strain ADS47 genome and C. koseri. Scaffolds derived from ADS47 genome sequences (Y- axis) were aligned and assembled by comparison with the closest completely sequenced genome, C. koseri ATCC BAA-875 (X- axis).

129

5.4.2. An overview of the ADS47 genome

General features of the C. freundii strain ADS47 are presented in Table 5-1. Some notable features are that strain ADS47 contains 4,881,122 bp, circular chromosome, with an average GC content of 52.02%. The genome contains 4,610 CDSs which comprise 87.84% of the whole genome. Annotation by RAST assigned 2,821 (62%) CDSs to known biological functions and remaining 1,789 (38%) genes were classified to unknown or hypothetical proteins.

The GC content in the coding regions is significantly higher (53.18%) than in the non- coding intergenic region (43.79%), which covers 12.15% (592,781 bp) of the genome. The intergenic region is comprised of 0.18% repeat sequences (total 8,999 bp), 5 (0.09%) transposons (4,806 bp), 0.25% stable RNA (12,091 bp) and the remaining ~11.5% is classified as regulatory and other functions. The ADS47 genome contains 62 tRNAs, 3 rRNAs and 34 copies of sRNAs. The average lengths of the tRNAs, rRNAs and sRNAs are

78.5, 885 and 134.41 bp, respectively. There are also 42 phage/prophage-related genes identified in the genome.

Figure 5-2 shows the distribution of the 26 categorized CDSs, rRNA, tRNA, GC content and GC skew. It appears that ADS47 genome contains equal quantity of genes in leading and lagging strands and showed occasional clustering (likely operons) of genes in the two strands. The genome contains three rRNAs (two large subunit and a small subunit) and 62 tRNAs, which are distributed across the genome. A consistent GC content was observed in the ADS47 genome, however, 7-10 low GC content (high black color peaks) genomic regions were found at locations: origin of replication, 600, 2000, 3000, 3300 3550, 3850, 4300 and

4800 kbp. This is indicative of possible horizontally transferred genes or insertions in those

130 genomic locations. GC skew denotes the excess of C over G in certain genomic locations.

Positive GC skew (+) (green color) indicates the skew value in the leading strand, while negative skew (-) (purple color) in the lagging strand. As in many bacterial genomes, the

ADS47 genome map shows the separation of positive and negative GC skew at the origin of replication (base pair position 1). The GC skew in the leading strand is highly skewed (high green color peak), indicating that possible horizontal gene transfer occurred into the ADS47 genome.

The CDSs ranged from ~100 bp to ~2,000 bp and have an average size of 893 bp

(Figure 5-3). The highest numbers of CDS (770) were 600-800 bp long. A considerable number of genes (220) were >2,000 bp in size. Some of the biological functions for the smallest genes (100-200 bp) are sulfur carrier protein, methionine sulfoxide reductase, ribonuclease P protein component, while the largest genes (>2,000 bp) are ferrous iron transport protein B, translation elongation factor G, pyruvate flavodoxin oxidoreductase and tetrathionate reductase.

131

Table 5-1. General features of the C. freundii strain ADS47 genome.

Feature Chromosome

Genome size 4,881,122 bp

Overall GC content 52.02 %

Number of genes 4610

Genes with assigned function (number and %) 2821 (62 %)

Genes without assigned function (number and %) 1789 (38 %)

Total length of protein coding sequences 4,285,461 bp

Genome coverage by coding sequences (%) 87.84 %

Average length of coding sequences 893 bp

GC content in the coding region 53.18 %

GC content in the non-coding region 43.79 %

Length and percentage of intergenic region 592,781 bp and 12.15%

Total repeat sequences and genome coverage 8,999 bp and 0.18 %

Total transposons sequences and genome coverage 4,806 and 0.09 %

Number, average length and genome coverage of tRNA 62 copies, 78.5 bp and 0.099 %

Number, average length and genome coverage of rRNA 3 copies, 885 bp and 0.067 %

Number, average length and genome coverage of sRNA 34 copies, 134,41 bp and 0.093 %

132

133

Figure 5-2. Circular genome map of C. freundii strain ADS47. Map was created with the

CGView software. From outside to inside, the first circle indicates the genomic position in kb.

Different functional categories of predicted coding sequences (CDS) are distinguished by colours with clockwise transcribed CDSs in the second circle and anticlockwise transcribed

CDSs in the third circle. Bars on the fourth and fifth circles indicate the rRNAs and tRNAs, respectively. Sixth and seventh circles show the coverage and GC content, respectively. The innermost circle demonstrates the GC skew + (green) and GC skew- (purple).

134

Figure 5-3. Size distribution of the predicted genes identified in C. freundii strain ADS47.

The X axis identifies the gene size categories (bp) and Y axis the abundance of genes.

135

5.4.3. Annotation and functional classification of ADS47 genome

RAST analysis identified predicted CDSs in the genome of Citrobacter strain ADS47, assigned functions and grouped the annotated genes into 26 functional subsystem categories out of 27. No CDS was found belonging to the photosynthesis category. A graphical presentation of the distribution of the CDSs into each of the 26 categories is given in

Figure 5-4.

A large portion (38%) of the CDSs belong to hypothetical or unknown functions.

Among the 62% known or predicted proteins, the highest numbers (835) belong to the carbohydrate category, followed by amino acids and derivatives (468) and cofactors/vitamins/prosthetic groups (324). A noticeable number of stress response (236) and respiration (198) categories genes were identified in the bacterial genome. The ADS47 genome encoded by 171 membrane transport genes, 113 virulence/defense genes and 126 genes involved in regulation/cell signaling categories. Two hundred and one CDSs were grouped into the miscellaneous category whose functions are different than the indicated

RAST categories. The RAST annotations further divided the categorized CDSs into subcategories based on similar functional roles, which is discussed in section 5.4.5.

136

Figure 5-4. Functional classification of the predicted coding sequences (CDS) of C. freundii strain ADS47. The annotation was performed by Rapid Annotation using Subsystems

Technology (RAST) program.

137

5.4.4. Molecular phylogenetic relationships

Genome-wide phylogenetic analysis, based on comparisons of whole genome amino acid sequences of two completely sequenced Citrobacter genomes (C. koseri and C. rodentium) and three partially sequenced Citrobacter genome sequences (C. freundii,

Citrobacter sp. and C. youngae), illustrates the relationships of ADS47 to five Citrobacter species (Figure 5-5).

C. youngae ATCC 29220, originating from human feces, and two pathogenic species

(C. koseri ATCC BAA-895 and C. rodentium strain ICC168) are progressively more distantly related to ADS47. The present analysis is consistent with the 16S-rRNA sequence based phylogenetic analysis (Chapter 3), which showed that the closest bacterial strain to ADS47 is

C. freundii (Figure 3-2). This confirmed the identification of strain ADS47 as the gamma proteobacterial species, C. freundii, made in Chapter 3.

138

Figure 5-5. Phylogeny of the trichothecene detoxifying C. freundii strain ADS47. The dendrogram was constructed by comparing the ADS47 genome (amino acid sequences) with five Citrobacter genomes and a E. coli strain (out group species) using the CVTree software.

The calculated pairwise distances matrix by Protdist program was transformed into a tree by neighbor-joining program. The dissimilarity tree was drawn to scale with branch lengths

(scale bar 0.02, indicated the amino acid substitutions among the genomes). The C. koseri, C. rodentium, C. freundii, Citrobacter sp., C. youngae and E. coli protein sequences were uploaded from GenBank (Benson et al. 2009).

139

5.4.5. Comparative genomics

Genome-wide comparative analyses between the genome of strain ADS47 and the five previously sequenced Citrobacter genomes showed that 87.2% of the total nucleotides between ADS47 and C. freundii Ballerup were identical. The values for the comparisons between C. freundii strain ADS47 were 82.12% for Citrobacter sp. 30-2, 76.37% for C. youngae, 59.81% for C. koseri, and 42.13% for C. rodentium (Table 5-2).

The two human and mouse pathogenic species, namely C. rodentium and C. koseri, and the bacteria originally isolated from human feces (C. youngae) have higher GC contents

(54.6 and 53.2%, respectively) compared to C. freundii strain ADS47, C. freundii strain

Ballerup and Citrobacter sp. 30_2 which have slightly lower GC contents (~52%). The two pathogenic species contain 22 copies of rRNA, which is significantly higher than the 9 found in C. freundii strain Ballerup, 6 in Citrobacter sp. 30_2 and 3 in C. freundii strain ADS47, but less than the 25 copies in C. youngae. The genome of C. freundii strain ADS47, C. freundii strain Ballerup, C. youngae and Citrobacter sp. 30-2 had a similar percentage of the genome comprising coding sequences (86.5-88%), C. koseri and C. rodentium genomes contained the highest (90%) and lowest (84%) genome coverage by coding sequences, respectively.

140

Table 5-2. Genome comparisons between strain ADS47 and five Citrobacter species. The base nucleotide sequence identities were performed using MUMmer DNAdff script program

(Delcher et al. 2002). The features (GC content, total CDS, percentage CDS and number of ribosomal RNA) were determined using RAST program (Aziz et al. 2008).

Genome feature Bacterial species C. C. freundii Citrobacter C. C. koseri C. freundii Ballerup sp. 30_2 youngae rodentium ADS47

Genome length 4,878,242 4,904,641 5,125,150 5,154,159 4,720,462 5,346,659

(bp)

Base similarity to NA 87.15 82.12 76.37 59.81 42.13

ADS47 (%)

GC (%) 52.02 52.18 51.6 52.65 53.0 54.6

Number of coding 4610 4685 4728 5278 4978 4800 sequences

Genome coverage 88.0 86.5 86.6 87.2 90.0 84.0 with coding sequences (%) rRNA copies 3 9 6 25 22 22

Total stable RNAs 99 81 78 105 120 142

141

The number of orthologs shared between ADS47 and other Citrobacter species is illustrated in Figure 5-6. ADS47 shares the smallest number of orthologous CDSs with C. rodentium and C. koseri, 3250 and 3341, respectively. The highest number of shared CDS were found with C. freundii Ballerup (3706) followed by C. youngae (3693) and Citrobacter sp. 30_2 (3676). A total of 2821 CDSs were found to be common to all of the six Citrobacter genomes analysed (data not shown). Nonetheless, significant diversity exists between C. freundii ADS47 and the other species, as a considerable number of unshared genes were identified from the comparisons with C. freundii Ballerup [979 (20%)] Citrobacter sp. 30_2

[1064 (23%)], C. youngae [1654 (31%)], C. koseri [1693 (34%)] and C. rodentium [1634

(34%)].

142

Figure 5-6. Comparisons of coding sequences between strain C. freundii ADS47 and five

Citrobacter species. Coding sequences of ADS47 were compared with C. freundii Ballerup,

Citrobacter sp. 30_2, C. youngae, C. koseri and C. rodentium genomes based on all-versus-all reciprocal FASTA comparison of at least 30% amino acid identity and >70% coverage of the total length using OrthoMCL program. A Venn diagram demonstrates the shared and unshared (indicated in the brackets) coding sequences.

143

Table 5-3 shows differences in the 26 RAST categorized genes between the six

Citrobacter genomes. Even though a low abundance of virulence/disease/defense category genes were found in the animal/human pathogenic bacteria of C. rodentium (85) and C. koseri

(92), significantly higher mycobacterium virulence genes (13) were identified in this two species. The genomes of the two closely related strains, C. freundii ADS47 and C. freundii

Ballerup lack this type of virulence gene. Instead, elevated numbers of genes for resistance to antibiotic and toxic compounds were found in C. freundii ADS47 (94), C. freundii Ballerup

(88) and Citrobacter sp. (110) compared to other three species. Copper homeostasis and cobalt-zinc-cadmium resistance genes were identified in strain C. freundii ADS47 (13 and 21, respectively) and closely related C. freundii Ballerup (12 and 22) and Citobacter sp. (18 and

22). Significantly lower numbers of these genes were present in C. rodentium (6 and 4) and C. koseri (7 and 8). C. rodentium and C. koseri lack mercuric reductase, mercuric resistance and arsenic resistance genes, perhaps because being animal/human pathogens, but these genes were found in the remaining four genomes. The C. koseri genome is richer in beta–lactamase genes (5) compared to the remaining genomes, which encodes 1-3 genes.

144

Table 5-3. Functional categorized gene abundances among the genomes of C. freundii

ADS47, C. freundii Ballerup, Citrobacter sp. 30_2, C. youngae, C. koseri and C. rodentium.

The genomes were classified into different functional categories with the RAST program, which classified the total CDSs.

RAST subsystem category Citrobacter species C. C. C. C. C. C. freundii freundii sp. youngae koseri rodentium ADS47 Ballerup Cofactors/vitamins/prosthetic groups/pigments 323 317 310 311 325 304 Cell wall and capsule 279 272 248 262 247 262 Virulence, disease and defense 113 107 134 100 92 85 Potassium metabolism 50 53 33 32 30 28 Miscellaneous 210 228 200 185 245 185 Phage/prophage/transposable 24 75 25 4 6 49 element, plasmid Membrane transport 170 156 171 195 171 172 Iron acquisition and metabolism 80 81 86 62 113 27 RNA metabolism 220 222 217 215 238 212 Nucleosides and nucleotides 119 112 110 111 120 114 Protein metabolism 249 280 269 298 287 254 Cell division and cell cycle 30 29 32 29 38 29 Motility and chemotaxis 133 124 90 147 143 133 Regulation and cell signaling 126 130 124 135 132 123 Secondary metabolism 21 22 20 21 0 0 DNA metabolism 145 167 165 155 147 152 Fatty acids, lipids and isoprenoids 121 149 122 121 134 133 Nitrogen metabolism 50 52 53 49 56 58 Dormancy and sporulation 3 3 3 4 3 3

145

Table 5-3. (continued)

Citrobacter species C. C. C. C. C. C. RAST subsystem category freundii freundii sp. youngae koseri rodentium ADS47 Ballerup Respiration 198 184 180 191 166 162 Stress response 236 235 170 177 179 174 Metabolism of aromatic 20 20 28 25 17 48 compounds Amino acids and derivatives 482 476 450 466 428 416 Sulfur metabolism 48 50 52 57 56 56 Phosphorus metabolism 50 49 43 42 43 28 Carbohydrates 837 803 708 731 733 737

146

Strains C. freundii ASD47 and C. freundii Ballerup have much more potassium metabolism category genes (50 and 53, respectively) than the other bacterial species, which code for 28-33 proteins. Larger numbers of genes associated with the glutathione regulated potassium efflux system were also identified in C. freundii ASD47 with 23 and C. freundii

Ballerup with 26 compared to the remaining four species that encode only five genes. All of the six genomes contained phage/prophage-related genes with the maximum identified in C. freundii Ballerup (75) followed by C. rodentium (49), Citrobacter sp. (25) and C. freundii

ADS47 (24). Lower numbers of phage/prophage-related genes were found in C. koseri (6) and

C. youngae (4).

The six Citrobacter genomes showed remarkable variability in their secretion systems.

The type-III secretion operon genes are only found in C. rodentium and C. freundii Ballerup.

Type-II secretion genes are predominant in C. rodentium (12) followed by ADS47 (9).

Type-IV secretion system is uncommon in Citrobacter species since it was only identified in Citrobacter sp., where it is represented by 19 genes. Variable numbers of type-VI secretion system genes are present in the five genomes, except C. koseri, which lacks such genes. C. youngae has the highest number (43) of type-VI secretion proteins. The other strains have a range of type-VI secretion proteins; in particular 27 were identified in C. freundii

Ballerup, 15 in C. freundii ADS47 and 8 in Citrobacter sp.. Type-VII (chaperone/usher) secretion system genes were found in all of the six Citrobacter species, with the largest (42) and smallest (13) numbers of these genes occurring in C. koseri and C. freundii Ballerup, respectively.

Although a relatively similar number of ABC transporters (37-42) were identified in the six genomes, C. freundii ADS47 contain the largest number of tripartite ATP-independent

147 periplasmic (TRAP) transporter genes (7). By contrast, the closely related C. freundii

Ballerup, lacks such transporter system. An identical number of iron acquisition and metabolism category genes were found in C. freundii ADS47 (80) and C. freundii Ballerup

(81), while the two pathogenic species, C. rodentium and C. koseri, code for the lowest (27) and highest (113) numbers of such genes. C. koseri is rich in siderophore related genes (64), but the remaining five genomes have smaller numbers of siderophore-associated genes (20-

22).

Variable numbers of motility and chemotaxis related genes were found in the

Citrobacter genomes, with the largest number occurring in C. youngae (147), followed by C. koseri (143), C. freundii ADS47 (133), C. rodentium (133), C. freundii Ballerup (124) and

Citrobacter sp. (90). The two pathogenic species (C. rodentium and C. koseri) encode a remarkably low number of quorum sensing genes (2). In contrast, C. youngae codes for the highest number of quorum sensing genes (19). A common and distinctive feature of the two pathogenic species is the lack of genes required for secondary metabolism. In contrast, the non-pathogenic species contain a common set of genes for degrading a range of phenylpropanoids and cinnamic acid (20-22).

The lowest number of genes related to phospholipids/fatty acids biosynthesis and metabolism were found in C. freundii ADS47 (66) compared to C. freundii Balllerup (94) and

C. koseri (98). The smallest numbers of nitrate/nitrite ammonification genes were found in

ADS47 (31) and its closest genome, C. freundii Ballerup (32). In contrast, the two pathogens encode the highest number of such proteins (39-40).

C. rodentium and C. koseri code for significantly low numbers of respiratory category genes (162-165), while C. freundii ADS47 contains the highest number of such genes (198).

148

Further analysis of the subcategory genes determined that the two gut species (C. youngae and

Citrobacter sp. 30_2) contain the largest numbers of genes encoding anaerobic reductases

(22) and identical numbers of formate dehydrogenases (17). The largest numbers of formate dehydrogenase genes were identified in C. freundii Ballerup (32) and C. freundii ADS47 (30), but the two pathogenic species differed from the other species by encoding the lowest number

(13) of c-type cytochrome biogenesis genes.

A higher number of aromatic compound metabolism category genes were found in the mouse pathogen, C. rodentium (48), compared to the human pathogen (C. koseri) which contain 17 genes. A significantly higher number of genes involved in central aromatic intermediate metabolism (26) were found in C. rodentium compared to the mouse pathogen

(C. koseri) and other genomes which only contain 6-8 such genes. The two C. freundii strains,

ADS47 and Ballerup, lacked the anaerobic degradation of aromatic compound

(hydroxyaromatic decarboxylase family) genes.

In spite of the abundance of stress response category genes, strain ADS47 has only a small number of desiccation stress genes (5) compared to the other species, which ranged from 9 to10. However, ADS47 and Ballerup code for higher number of genes related to oxidative stress (94 and 91, respectively) and cold shock (8 and 7, respectively) than the other species.

The smallest numbers of amino acids and its derivatives category genes were found in the two pathogens, C. rodentium and C. koseri (416 and 428, respectively), while the largest number of these genes were identified in strain ADS47 (482) and C. freundii Ballerup (476).

Despite the similar number of sulfur metabolism category genes determined across the six species, only ADS47 lacks taurine utilization genes. Strains ADS47 and Ballerup were found

149 to be rich in the phosphorous metabolism gene category (49-50), while C. rodentium contains a significantly decreased number of these category genes (28).

The bioinformatics analyses found the largest number of carbohydrate category genes in ADS47 (837) followed by Ballerup (803). The gut bacteria (Citrobacter sp. 30_2) codes for the least number of genes assigned to the carbohydrate category (708), while C. koseri, C. rodentium and C. youngae contained 731-737 in this category genes. The analyses also determined that there are relatively small numbers of TCA and Entner Doudoroff pathway genes in C. rodentium and C. youngae compared to the other four species and relatively large numbers of fermentation-related genes in ADS47 (81) and C. freundii (85) compared to the other four species.

A comparison of the pathogenicity operons [locus of enterocyte effacement (LEE)] of

C. rodentium with other five Citrobacter genomes revealed the absence of this virulence determinant gene region in the other four Citrobacter species, including ADS47 (Figure 5-7).

The analysis showed the presence of LEE operon genes in C. freundii strain Ballerup, as no gap was observed in that genomic location, while C. freundii ADS47, C. koseri, C. youngae and Citrobacter sp. 30_2 lacked that loci (7a). Further ananlysis showed alignments between

C. rodentium and E. coli LEE operon genes (Figure 7b). In contrast, no alignment of such genes was observed in ADS47 genome.

150

(a)

151

(b)

LEE operons

Figure 5-7. (a) Blast Ring Image Generator (BRIG) output based of Citrobacter rodentium

(NC_013716) compared against ADS47, C. freundii, C. koseri, C. youngae, E. coli and

Citrobacter sp. genomes. The targeted comparative genome analysis was performed based on locus of enterocyte effacement (LEE) operons of C. rodentium (Petty et al. 2011). The innermost rings show GC content (black) and GC skew (purple and green). The remaining rings show BLAST comparisons of the other Citrobacter and E. coli genomes against C. rodentium. The region shown in red is LEE operon genes. (b) The absence of LEE operons in the ADS47 genome comparing to C. rodentium and E. coli. Analysis was performed and visualized by Artemis Comparison Tool software v13.2

(http://www.sanger.ac.uk/resources/software/act/) (Tatusov et al. 2000, Craver et al. 2005).

152

5.4.6. Analysis of the unique genes in C. freundii ADS47

Two hundred and seventy genes unique to ADS47 were identified from a comparison of its genome to the other five Citrobacter species. In general, the unique genes are distributed throughout the ADS47 genome, but some clusters of unique genes occurred including eight unique genomic regions (Figure 5-8). Six large unique regions are located at

767,204-775,090 (~8 kb), 1,025,572-1,042,510 (17 kb), 2,798,289-2,807,217 (~9 kb),

3,045,220-3,055,912 (~11 kb), 3,930,366-3,943,075 (12 kb) and 4,022,195-4,032,444 (10 kb).

Out of 270 unique genes, 147 (54 %) are hypothetical proteins. The remaining 123

(46%) genes are grouped into reductases/oxidoreductases (9 genes), de-acetylase (1), hydrolases (2), transferases (4), sulfatases/arylsulfatases (3), bacteriocins (3), type VI secretion proteins (2), kinases (4), transporters (11), membrane/exported proteins (8), regulators (12), integrase/recombinases (5), phages/prophages/transposaes related genes (27) and miscellaneous (32), that includes DNA helicases, ligases, fimbrial proteins and chromosome partitioning genes (Table 5-4).

The unique regions encode various functional genes that may cause trichothecene de- epoxidation or de-acetylation reactions. The 8 kb unique regions contains six genes encoding glycosyltransferases and two copies of oxidoreductase-like coding sequences (CMD_0711 and CMD_0712), that showed 39% identity to the Marinobacter aquaeolei F420 hydrogenase/dehydrogenase beta subunit. The largest (17 kb) unique genomic region codes for 16 genes, including: dehydrogenase, hydrolase, bacteriocin (pyocin), integrase and a short chain dehydrogenase/reductase-like (SDR) (CMD_0979) gene (Figure 5-9a).

153

The SDR like gene is 49% amino acid similarity to an oxidoreductase of the phytopathogenic Dickeya dadantii, and the enzyme contains a domain with 77% homology to the Dickeya zeae NAD(P) binding site. The 9 kb unique genomic section contains 8 genes, including a NaCl symporter, a dehydratase (CMD_2679) with 86% homology to

Enterobacteriaceae bacterium dihydroxy-acid dehydratase and a putative oxidoreductase

(CMD_2683) having 37% similarity with oxopropionate reductase of Bacillus sp. The unique region also encodes a short chain dehydrogenase/reductase (SDR) (CMD_2684) that shows

85% identity to Enterobacteriaceae bacterium (Figure 5-9b). A unique thiol-disulfide and thioredoxin like reductase (CMD_3721) was identified in a 10 kb unique region that contains

13 genes (Figure 5-9c). This gene is 59% identical to Oligotropha carboxidovorans thiol- disulfide isomerase and thioredoxin. In addition, several reductases/oxidoreductase-like unique genes were identified in the ADS47 genome, including an oxidoreductase

(CMD_1024) with 27% identity to Canadidatus nitrospira NADH-quinone oxidoreductase membrane L. Another oxidoreductase-like gene identified in a different location of the

ADS47 genome (CMD_1393) shows 87% homology to Klebsiella pneumoniae deoxygluconate dehydrogenase contains a SDR active site. A putative nitrate reductase gene

(CMD_3184) has a low level of identity (27%) with an uncultured bacterial respiratory nitrate reductase. The ~11 kb unique region encodes sulfatases/arylsulfatses

(CMD_2898/CMD_2905) and related genes. Furthermore, ADS47 contains a unique glycosylphosphatidylinositol (GPI) anchor biosynthesis operon, comprised of four genes, including a N-acetylglucosaminyl-phosphatidylinositol de-acetylase (CMD_2431), which showed 78% similarity to E. coli de-acetylase (Figure 5-9e).

154

Other unique genes/predicted genes in the ADS47 genome include polysaccharide/phenazine biosynthesis genes (CMD_0710/CMD_1406), heavy metal kinase

(CMD_0989), two component heavy metal regulator (CMD_0990), nitrate reductase regulator

(CMD_3477), putative potassium/zinc transporters (CMD_0973/CMD_4600), sodium symporter (CMD_2678), (tripartite ATP-independent periplasmic transporters) TRAP transporters (CMD_3693, CMD_3694, CMD_3695, CMD_4337), glutamine/ribose ABC transporters (CMD_3176/CMD_3544), toxin transcription regulator (CMD_3464), fimbrial protein/chaperone (CMD_1695, CMD_2434, CMD_3277, CMD_3278) and ParB chromosome/plasmid partitioning proteins (CMD_4033/CMD_4034).

Majority of the unique genes and their distributions are shown in comparative circular genome maps (Figure 5-8).

155

Table 5-4. Unique genes of C. freundii strain ADS47. Genes were categorized according to the functional similarity. The unique genes were determined by comparing ADS47 with five sequenced Citrobacter species using OrthoMCL software.

.Unique gene classes Number

Total unique genes 270 Reductases /oxidoreductases 9

De-acetylase 1

Hydrolase 2

Transferase 4

Sulfatase 3

Bacteriocin 2

Type VI secretion 2

Kinase 4

Transporter 9

Membrane/exported protein 8

Regulator 12

Integrase/recombinase 5

Phage/prophage/transposae 27

Miscellaneous 34

Hypothetical protein 147

156

157

Figure 5-8. Locations of the unique genes of C. freundii strain ADS47. The unique genes were determined by comparing the total CDS of strain C. freundii ADS47 with C. freundii

Ballerup, Citrobacter sp. C. youngae, C. koseri and C. rodentium CDSs by OrthoMCL program. The map was developed by Blast Ring Generator (BRIG) software. First, second and third circles from inside to out side show the genome size, % GC and GC skew of strain

ADS47. Forward (Fwd) and reversed (Rev) transcribed CDSs of ADS47 are shown in green

(forth circle) and red (fifth circle), respectively. The next five rings represent the circular genome maps of five Citrobacter species, distinguished by different ring colours (C. rodentium as sky blue, C. koseri as dark blue, C. youngae as light blue, C. freundii Ballerup as yellow and Citrobacter sp. as green). The degree of orthologous gene identity between

ADS47 and other five genomes are shown by color intensity (light color= 75% gene identity, ash color= 50% gene identity). Gaps across five genome rings indicate the absence of genes/regions compared to ADS47. Unique genes are shown surrounding the map rings and the identities of some of the genes are given. Arrows indicate putative reductases/oxidoreductases and de-acetylase.

158

(a)

(b)

159

(c)

(d)

160

Figure 5-9. Unique genomic regions and associated genes in C. freundii ADS47 compared to five Citrobacter species. RAST server based analysis compared the unique genes/regions of

ADS47 with the publicly available databases. The upper panels demonstrate the large unique regions defined by yellow boxes and genes located in the genomic regions are indicated by arrows. The lower panels are closely related genes/operons found in distantly related species.

The unique regions that contain potential trichothecene de-epoxidation genes are shown in (a) short chain dehydrogenase/reductase (SDR) and dehydrogenase, (b) putative oxidoreductase, dehydratase and SDR, (c) thiol-disulfide isomerase and thioredoxin oxidoreductase and (d) unique glycosylphosphatidylinositol (GPI) biosynthesis operon having N-acetylglucosaminylphosphatidylinositol de-acetylase gene.

161

5.5. Discussion

The de novo sequenced ADS47 genome was oriented and assembled using a completely sequenced C. koseri genome as reference strain. The synteny analysis between

ADS47 and C. koseri showed very similar overall organization of the two genomes. The high preservation of genes among five Citrobacter species and ADS47 were emphasized by the close alignments observed by Blast Ring Image Generator (BRIG) outputs obtained from these genomes.

Genome-wide phylogenetic analyses, based on a large number of common CDSs

(2,821) observed within the six Citrobacter genomes confirmed the species identification based on 16S-rDNA sequencing of the trichothecene mycotoxin-detoxifying C. freundii strain

ADS47. The base sequence idendity was 87.15% between these two genomes. They shared

~80% CDS and had a similar GC content (52%). Citrobacter are Gram-negative, rod shaped and facultatively anaerobic bacteria belonging to the family of Enterobacteriaceae (Wang et al. 2000, Whalen et al. 2007). Strains ADS47 and Ballerup of C. freundii are both aerobic gram-negative bacilli that have numerous flagella and inhabit in soil (Cong et al. 2011, Islam et al. 2012). C. freundii is considered to have a positive environmental role by reducing nitrate to nitrite (Rehr and Klemme 1989), which is an important step in the nitrogen cycle.

C. freundii, C. rodentium and C. koseri are animal/human pathogenic species. C. rodentium is a natural mice pathogen causing hemorrhagic colitis and hemolytic uremic syndrome (Gareau et al. 2010), C. koseri causes human disease (Badger et al. 1999, Kariholu et al. 2009), particularly neonatal meningitis, and C. freundii is an opportunistic human pathogen, responsible for a various types of infections like respiratory tract, urinary tract and blood (Whalen et al. 2007). C. freundii is very common representing approximately 29% of

162 all opportunistic human infections (Hodges et al. 1978). Outside of animals, they can be found in soil, water, sewages, and different organs of diseased animals (Wang et al. 2000). These pathogens code for large numbers of mycobacterium virulence operons/genes, which were absent in the three other genomes. However, they have fewer genes involved in environmental stress, such as copper homeostasis genes and no mercuric/arsenic resistance and secondary metabolism category genes, unlike the other Citrobacter genomes. Nonetheless, considerable variability exists between the two pathogens correlates with C. rodentium causing disease in mice, C. koseri causing meningistis in human and C. freundii causing infections in different human organs.There were attaching and effacing (A/E) lesion forming (LEE) operons in C. rodentium and C. freundii, but not in C. koseri. Those genes have shown to be acquired by lateral gene transfer playing major roles in virulence of enteropathogenic and enterohemorrhagic E. coli and C. rodentium (Mundy et. al. 2005, Petty et al. 2010, Petty et al.

2011. The ~35.6 kb pathogenic LEE gene island is comprised of five operons (LEE1 to

LEE5) that code for ~40 virulence genes that include type-III secretion genes and associated effectors (Berdichevsky et al. 2005). In addition, an elevated number of type-II and -VI secretion proteins identified in C. rodentium which may act as additional virulence factors

(Petty et al. 2010). Because it lacks type-III secretion apparatus genes, it is likely that C. koseri uses a different strategy to cause human neonatal meningitis (Badger et al. 1999, Ong et al. 2010).

To adapt to different environmental niches or hosts, bacteria like Citrobacter species frequently gain/lose genes, increase repeat sequences and recombine genomic regions using different mechanisms (Parkhill et al. 2001, Petty et al. 2011). The existence of large numbers of phage/prophage/transposae related genes in the Citrobacter genomes, particularly in C.

163 freundii Ballerup, suggests that their genomes are unstable. For example, C. freundii Ballerup and C. koseri might have acquired LEE operons by horizontal transfer via phage mediated transpositions. Interestingly, although ADS47 is closely related to C. freundii Ballerup, it lacks the pathogen-related LEE gene island and mycobacterium virulence operons required for invasion and intracellular survival in macrophages (Gao et al. 2006). Instead, ADS47 has a larger number and variety of stress-responsive genes, such as efflux genes and bacteriocins

(pyocin, hemolysin) genes that provide protection against heavy metals, and antibiotics and increase their competitiveness against other bacteria. These findings suggest that the strain

ADS47 is not an opportunistic animal pathogen, but has evolved to be highly competitive in the soil micro-environment (Arnold 1998, Bengoechea and Skurnik 2000). Perhaps ADS47 evolved trichothecene degrading genes for adaptation to soils containing trichothecene mycotoxins contaminated soils (Gadd and Griffiths 1978, Sharma and Fulekar 2009, Han et al. 2011). Conversely, the absence or reduced number of stress-responsive genes in C. rodentium and C. koseri may imply that those two pathogenic species are not exposed to such stressful and xenobiotic-contaminated environments as ADS47.

This interpretation is supported by the observation that the two of the three animal/human pathogenic species contained high of rRNA copy number, but lower numbers were found in ADS47, C. freundii Ballerup and Citrobacter sp. 30_2. Bacteria maintain large number of rRNA for higher production of ribosome during high growth rates (Stevenson et al.

1997, Klappenbach et al. 2000), such as might occur during the establishment of an infection.

However, a large copy number of rRNAs may not be metabolically favourable for bacteria, such as ADS47 that evolved in a highly competitive environment, as that costs higher cellular energy (Hsu et al. 1994).

164

C. freundii Ballerup seems to be an exception in that it is similar to the other animal pathogens for containing virulent factors (i.e. LEE-operon genes), but more like the environmental species for encoding high number genes for stress, defence, heavy metal and toxic compounds resistances.

The comparison of the genes in C. freundii ADS47 with the unique ability to detoxify trichothecenes with genes in other Citrobacter species allowed approximately 270 unique genes to be identified in ADS47. Some of those genes may be involved in trichothecene de- epoxidation or de-acetylation. Currently, the pathways and microbial enzymes leading to trichothecene de-epoxidation and de-acetylation are unknown, but based on the required reaction, i.e. the elimination of an oxygen atom from epoxide moiety, it is likely that epoxide reductases or epoxide oxidoreductases are responsible for such biodegradation activity (He et al. 1992).

From database searches, three enzymes having epoxide reductase activity were found.

These are a vitamin K epoxide reductase (VKOR), a monooxygenase PaaABCE and pyridine nucleotide-disulfide oxidoreductase (PNDO), which reduce vitamin K epoxide, phenylacetyl-

CoA and epoxyalkanes, respectively (Swaving et al. 1996, Li et al. 2004, Teufel et al. 2012).

In mammals, VKOR plays important roles in vitamin K cycle by reducing vitamin K epoxide to vitamin K, and its homologues, that are found in plants, bacteria and archaea (Goodstadt

2004, Li et al. 2010). The microbial multi-component PaaABCE monooxygenase catalyzes degradation of variety of aromatic compounds or environmental contaminants via epoxidation of phenylacetic acid (Paa) and NADPH-dependant de-epoxidation of epoxy-Paa (Luengo et al. 2001, Teufel et al. 2012). PNDO is a unique type of NADP-dependent oxidoreductase identified in Xanthobacter strain Py2 that was obtained by propane enrichment procedure

165

(van Ginkel and de Bont 1986). However, no orthologs of these genes were found in ADS47 genome.

Nine ADS47-unique reductase/oxidoreductase-like genes are potential candidates for the trichothecene de-epoxidation enzyme. Genes belong to short chain dehydrogenase/reductase (SDR) family are structurally and functionally diverse and cause different enzymatic activities including xenobiotic degradation (Kavanagh et al. 2008). This is a promising candidate for the trichothecene de-epoxidation gene. Genes belong to dehydratase family catalyze reduction reactions by cleaving carbon-oxygen bonds of organic compounds that result in removal of oxygen in various biosynthetic pathways (Flint et al. 1993). This mechanism may be relevant to microbial trichothecene de-epoxidation reaction since the latter involves removal of the oxygen atom forming the epoxide in trichothecene mycotoxins

(Young et al. 2007).

Another ADS47-unique gene was for a nitrate reductase that may be a candidate de- epoxidation gene. Under stress conditions, such an enzyme causes reductive removal of an oxygen molecule (Morozkina and Zvyagilskaya 2007).

A fourth type of ADS47-unique gene is for a thiol-disulfide isomerase/thioredoxin like

(Trx) oxidoreductase. The thiol-disulfide isomerase and thioredoxin oxidoreductases contain a thioredoxin like domain with a CXXC catalytic motif, which catalyzes various biological processes that allow the organism to adapt to environmental stress and develop competence to sporulate (Scharf et al. 1998, Petersohn et al. 2001, Smits et al. 2005). The membrane- anchored Trx–like domain initiates the vitamin K de-epoxidation reaction by reducing the active motif (CXXC) of VKOR (Schulman et al. 2010).

166

The present study did not find any homologues of VKOR in ADS47 genome, but we identified two copies of structurally similar quinone reductase, protein disulfide bond formation (DsbB) gene (CMD_1044 and CMD_2641). Both VKOR and DsbB contain four transmembrane segments, an active site CXXC motif and cofactor quinone, and use a common electron transfer pathway for reducing substrates (Li et al. 2010, Malojcic et al.

2008). As VKOR, bacterial DsbB requires a Trx-like partner (e.g., DsbA) for activity. Further experiments are required to examine if DON de-epoxidation occurs by the catalytic activity of

DsbB coupled with a unique thiol-disulfide isomerase and thioredoxin oxidoreductase present in ADS47 genome.

The analysis also identified an ADS47-unique glycosylphosphatidylinositol (GPI) biosynthetic operon that is comprised of four genes, including an N- acetylglucosaminylphosphatidylinositol de-acetylase. The GPI membrane anchor protein synthesis is initiated by transferring N-acetylglucosamine from uridine diphosphate N- acetylglucosamine to phosphatidylinositol by N-acetylglucosaminyltransferase, an endoplasmic reticulum (ER) membrane-bound protein complex (Hong and Kinoshita 2009).

In the second step of the pathway, N-acetylglucosamine-phosphatidylinositol is de-acetylated to produce glucosaminyl-phosphatidylinositol by the catalytic activity of a zinc metalloenzyme, N-acetylglucosaminylphosphatidylinositol de-acetylase which is localized on the cytoplasmic side of the ER (Nakamura et al. 1997, Hong and Kinoshita 2009). It is possible that the identified unique N-acetylglucosaminylphosphatidylinositol de-acetylase gene of ADS47 may be involved in trichothecene de-acetylation activity.

167

Further analysis is needed to determine which of these candidate gene or genes may be involved in tricthothecene mycotoxin de-epoxidation and de-acetylation activities.

Identification of gene responsible for the trichothecene detoxification activities of C. freundii

ADS47 opens many potential applications for its use in agri-based industries to contribute to crop/animal production and food safety (Karlovsky 2011).

5.6. Acknowledgement

We gratefully acknowledge Agriculture and Agri-Food Canada/Binational

Agricultural Research and Development Fund (AAFC/BARD) project for providing the financial supports of the work.

168

Supplemental Table S5-1. List of unique genes of strain ADS47, gene IDs and their functional roles. The unique genes were determined by comparing the ADS47 genome with five sequenced Citrobacter species (C. freundii, Citrobacter sp., C. youngae, C. koseri and C. rodentium) using OrthoMCL (Li et al. 2003).

Unique gene Number Gene ID and functions classes Total unique 270 123 known functions, 147 hypothetical proteins genes Reductases 9 CMD_0711 coenzyme F420 hydrogenase/dehydrogenase beta subunit, /oxidoreductases CMD_0979 Short chain dehydrogenase/ reductase (SDR)/ Oxidoreductase SDR dehydrogenase, CMD_1024 NADH-quinone oxidoreductase membrane L, CMD_1393 oxidoreductase: SDR / putative, deoxygluconate dehydrogenase (NAD+ 3-oxidoreductase), CMD_2679 Dehydratase/dihydroxy acid dehydratase, CMD_2683 oxidoreductase: oxopropionate reductase, CMD_2684 short-chain dehydrogenase/reductase SDR, CMD_3184 nitrate reductase alfa subunit, CMD_3721 thiol-disulfide isomerase and thioredoxin oxidoreductase Deacetylase 1 CMD_2431 N-acetylglucosaminylphosphatidylinositol deacetylase superfamily Hydrolase 2 CMD_0991 Dienelactone hydrolase, CMD_2895 Putative glycosyl hydrolase Transferase 4 CMD_1640 D12 class N6 adenine-specific DNA methyltransferase, CMD_2432 Glycosyltransferase, CMD_2433 putative acetyltransferase, CMD_3023 adenine-specific DNA modification methyltransferase, Sulfatase 3 CMD_2898 Possible sulfatase/ arylsulfatase/arylsulfatransferase, CMD_2902 Arylsulfatase/sulfatase, CMD_2905 Putative sulfatase/arylsulfatase Bacteriocin 3 CMD_0982 Thermosatble hemolysin superfamily, CMD_0993 Pyocin S2 superfamily (bacteriocin that kills other bacterial strain), CMD_3705 S-type pyocin containing domain protein Type VI secretion 2 CMD_1802 EvpB family type VI secretion protein, CMD_2364 type VI effector Hcp1 Kinase 4 CMD_0989 Heavy metal Sensor kinase, CMD_1794 Polo kinase kinase (PKK) superfamily CMD_3124 predicted diacylglycerol kinase, CMD_4266 putative signal transduction kinase Transporter 11 CMD_0973 TrkA-N domain-containing protein (K uptake transporter), CMD_1407 Branched-chain amino acid transport system carrier protein, CMD_2678 Sodium Symporter, CMD_3176 glutamine ABC transporter, CMD_3544 Ribose ABC transport system, CMD_3693 TRAP-type C4- dicarboxylate transport system binding protein, CMD_3694 TRAP-type C4- dicarboxylate transport system binding protein, small permease component (predicted N-acetylneuraminate transporter), CMD_3695 TRAP-type C4- dicarboxylate transport system, large permease, CMD_4337 2,3-diketo-L- gulonate TRAP transporter small permease protein yiaM component, CMD_4338 2,3-diketo-L-gulonate TRAP transporter large permease protein yiaN, CMD_4600 putative transporter involved in Zinc uptake (YciA protein).

169

Supplmental Table S5-1. (continued).

Unique gene Number Gene ID and functions classes Membrane/ 8 CMD_0713 putative membrane protein wzy, CMD_0983 Putative exported exported protein protein precursor, CMD_0986: Putative transmembrane protein, CMD_0987 putative exported protein precursor, CMD_1458 putative exported protein, CMD_2903 Putative exported protein, CMD_2904 putative secreted 5'-nucleotidase, CMD_4291 (S) putative integral membrane export protein Regulator 12 CMD_0977 probable transcriptional regulator protein, CMD_0980: Putative transcription activator TenA family, CMD_0990 Two component heavy metal response transcription regulator, CMD_1766 Leucine responsive regulatory protein, CMD_2677 transcriptional regulator GntR, CMD_2909 putative arylsulfatase regulator, CMD_3279 Uncharacterized outer membrane usher protein yqiG precursor, CMD_3464 toxin transcription regulatory protein, CMD_3465 EAL regulator protein (signaling), CMD_3475: putative transcription regulator, CMD_3477 Fumarate and nitrate reduction regulatory protein (for anaerobic adaptation), CMD_3812 putative regulator from phage. Integrase 5 CMD_0785, CMD_0992, CMD_0994, CMD_4032, CMD_4424 /recombinase Phage /prophage 27 Phages: CMD_1608, CMD_1609, CMD_1610, CMD_1611, CMD_1619, /transposae related CMD_1631, CMD_1913, CMD_1914, CMD_1915, CMD_1917, genes CMD_1920, CMD_1921, CMD_1936, CMD_3351, CMD_3804, CMD_3805, CMD_3807, CMD_3808, CMD_4581, CMD_4609, CMD_4610. Transposae: CMD_0252, CMD_0398, Prophage: CMD_1645, CMD_3806, CMD_4425, CMD_4597 Miscellaneous 32 CMD_0095 ATP-dependent DNA helicase Rep, CMD_0656 LysA protein, CMD_0710 Polysaccharide biosynthesis family protein, CMD_0978 TPR repeat containing protein, CMD_0981 AMP dependant synthatase and ligase, CMD_0985 Cupin superfamily, CMD_0987: flavodoxin, CMD_1404 Putative endoribonuclease, CMD_1405 Putative PAC domain protein, CMD_1406 Phenazine (PhzF) biosynthesis protein, CMD_1695 Putative fimbrial protein, CMD_1793 HP/Chromosome segregation ATPase domain protein, CMD_1922 FRG uncharacterized protein domain, CMD_2206 L-fucose isomerase like protein, CMD_2434 CS12 fimbria outer membrane usher protein, "CMD_2680 HpcH/HpaI aldose/citrate lyase protein family, CMD_2681 Aldose -1 epimerase, CMD_2682 Gluconolaconase /LRE like region, CMD_2900 Putative cytoplasmic protein, CMD_3009 putative restriction endonuclease, CMD_3277 Putative yqi fimbrial adhesion, CMD_3278 Hypothetical fimbrial chaperone yqiH, CMD_3340 Pea-VEAacid family, CMD_3518 3-hydroxy-3-methylglutaryl CoA synthase, CMD_3801 putative reverse transcriptase, CMD_3840 beta- D-galactosidase, alpha subunit, CMD_4033 ParB family chromosome partitioning protein, CMD_4034 ParB/repB chromosome/plasmid partitioning protein, CMD_4035 RelA/SpoT protein, CMD_4532 putative DNA ligase -like protein, CMD_4541 Antitermination protein Q and CMD_4585 putative ATP dependent helicase

170

CHAPTER 6

6.0. CONCLUSION

6.1. Summary

Contamination of grain with mycotoxins including deoxynivalenol (DON) and other trichothecenes is becoming a limiting factor for cereal and animal production (Binder et al.

2007). To address the issue, this study obtained a novel microbial source for DON detoxification and determined candidate genes for the function, which may have potential applications to detoxify grain and grain-derived food/feed contaminated with DON and TC mycotoxins.

From a screen of soil samples from crop field, microorganism was isolated that is able to de-epoxidize the most frequently detected food/feed contaminating mycotoxin, DON. The effects of media compositions, temperature, pH and aerobic/anaerobic conditions, on microbial detoxification activity were assessed. From 150 analyzed soil samples, a mixed soil microbial culture that showed DON de-epoxidation activity under aerobic conditions and moderate temperature (~24°C) was obtained. This culture exhibited a slow rate of microbial

DON to de-epoxyDON transformation (de-epoxidation), possibly because of competition between de-epoxidizing and other microorganisms in the culture.

The mixed microbial culture was enriched for DON detoxifying bacteria by reducing the non-targeted microbial species through antibiotic treatments, heating and subculturing on agar medium and selecting colonies. The enriched culture had higher DON de-epoxidation activity than the initial culture. Because of low abundance of DON detoxifying microorganisms, further culture enrichments were required to increase the target microbial

171 population. Reculture of the partially enriched microbial community in nutritonally poor MSB medium with high level of DON increased the DON de-epoxidizing bacterial population.

Serial dilution plating and selecting single colonies eventually led to isolation of a DON- detoxifying bacterial strain, named ADS47.

Phenotypic and 16S-rDNA sequence analyses provided evidence that the bacterium is a strain of C. freundii. Strain ADS47 is a gram negative, rod-shaped, facultative anaerobe and has peritrichous flagella. Strain ADS47 was unable to use DON as sole carbon source, however increasing DON concentrations in growth media enhanced its DON metabolism efficiency. The pure bacterial culture was shown to be able to rapidly de-epoxidize DON under aerobic conditions, moderate temperatures and neutral pH. Strain ADS47 more quickly de-epoxidizes DON in aerobic conditions than the anaerobic conditions. In addition to DON, strain ADS47 showed capability to de-epoxidize and/or de-acetylate 10 different type-A and type-B TC mycotoxins, which often co-contaminant grain with DON.

To clone the DON detoxifying genes, a DON detoxifying bacterial DNA library in E. coli host was constructed, followed by screening of the library clones for DON detoxification activity. The study produced a library giving approximately 10 fold coverage of the genome of strain ADS47. Restriction digestion analysis determined that most of the clones contained

ADS47 DNA inserts. This indicates that a good DNA library was produced which had >99% probability of containing all the DNA of the ADS47 genome. However, activity-based screening of the library could not identify any clones with DON de-epoxidation ability. It is assumed that bacterial DON de-epoxidation genes evolved in a certain environmental niche, which may be required for the DON de-epoxidation to be expressed and active but was not supplied by the host E. coli cells.

172

An attempt was made to identify the bacterial genes encoding DON de-epoxidation and de-acetylation enzymes by sequencing the genome of strain ADS47 and comparing its sequence to the genomes of related bacteria that are not known to posses DON de-epoxidation activity. A RAST server based bioinformatics analysis determined that the 4.8 Mb circular genome of ADS47 codes for 4610 predicted CDS. A genome wide phylogenetic analysis determined that strain ADS47 was >87% identical to the pathogen C. freundii strain Ballerup.

The relationship to two other animal/human pathogenic species C. rodentium and C. koseri was more distant. This result is consistent with the 16S-rDNA based phylogenetic analysis.

It is assumed that the microbial DON detoxifying genes were acquired by horizontal gene transfer or other means during the adaptation of ADS47 in mycotoxin contaminated soil.

If it is correct, then this means that those newly acquired genes are not common in other related microbial species, and no other Citrobacter species is known to degrade DON. Hence, the study identified 270 unique genes to ASDS47 with a subset of 10 genes for reductases, oxidoreductases and a de-acetylase which were considered to be candidate genes for the enzymes that have DON detoxifying activity. Furthermore, a targeted comparative genome analysis showed that the mycotoxin detoxifying ADS47 genome does not harbour the clinically important LEE encoded type–III secretion system and mycobacterium virulence genes. This suggests that strain ADS47 is not likely to be an animal pathogenic bacterium, unlike C. freundii Ballerup which encodes the pathogenic relevance LEE operons.

173

6.2. Research contributions

The research made four major contributions which may aid in reducing the risks of animals and humans from DON/TC exposure. Firstly, the study identified a novel bacterial strain ADS47 that is able to detoxify DON and 10 related mycotoxins over a wide range of temperatures under aerobic and anaerobic conditions. The discovery has led to a patent application for strain ADS47, supporting the novelty and importance of the finding. This bacterial strain could be used to detoxify trichothecene mycotoxins in contaminated animal feed, thus reducing food safety risks.

Secondly, the study resulted in the development of a method for isolating DON de- epoxidizing microbes from soil as well as optimizing the conditions for bacterial DON/TC detoxification activity to increase the effectiveness/ usefulness of the strain.

Thirdly, the thesis work resulted in availability of the genome sequence of a trichothecene detoxifying microbe, which may ultimately lead to an understanding of the cellular/molecular mechanisms of DON/TC de-epoxidation and de-acetylation genes/pathways. The genome analysis resulted in identification of several candidate genes for

DON/TC de-epoxidation and de-acetylation activities. Genes, which show de-epoxidation activity, could be used to develop DON/TC resistant cereal cultivars. Fourthly, strain ADS47 lacks major animal pathogenicity genes (in particular the LEE-operons) as well as does not encodes mycobacterium virulence genes. Thus, strain ADS47 has a great potential to be utilized to detoxify cereal grain and may be used as a feed additive for livestock/fishery. All together, the reported work may have significant applications to produce and supply

DON/TC-free food and feed.

174

6.3. Further research

This research provides the basis for developing biological detoxification of DON and related mycotoxins. This goal might be achieved by using DON de-epoxidizing gene to develop resistant plants, enzymatic detoxification of contaminated feed and use of the bacteria as a feed additive. Follow-up studies may reach the above indicated goals. Because the bacterium showed de-epoxidation activity under aerobic/anaerobic conditions and under wide range of temperatures, it is likely to have an application for enzymatic detoxification of contaminated animal feed. Therefore, additional studies on bacterial efficiency to detoxify contaminated cereal and grain-derived feed need to be conducted. Although strain ADS47 does not encode major animal pathogenicity genes, the bacterium must be directly tested on animals to see if it could negatively affect animal digestive systems. If not, then the potential for ADS47 is considerable for its use as a feed additive for pig, poultry and fish. Animal feeding trials will also need to be performed to determine if DON detoxification is sufficient to be cost effective as a feed additive. In order to develop DON/TC resistant cereal cultivars, the gene responsible to DON degrading activity among the presently identified candidate genes (reductases, oxidoreductases and de-acetylase) will need to be expressed in other organisms, like E. coli, and a codon modified version adapted for expression in plants, like corn and wheat.

175

7.0. REFERENCES

Abbas HK, Mirocha CJ, Pawlosky RJ, Pusch DJ. 1985. Effect of cleaning, milling, and

baking on deoxynivalenol in wheat. Applied and Environmental Microbiology, 50:482–

486.

Accensi F, Pinton P, Callu P, Abella-Bourges N, Guel WJF, Grosjean F, Oswald IP. 2006.

Ingestion of low doses of deoxynivalenol does not affect haematological, biochemical, or

immune responses of piglets. Journal of Animal Science, 84:1935–1942.

Agrawal D, Mahapatra AK. 2005. Vertically acquired neonatal Citrobacter brain abscess–

case report and review of the literature. Journal of Clinical Neuroscience, 12:188–190.

Alexander NJ. 2008. The TRI101 story: engineering wheat and barley to resist Fusarium head

blight. World Mycotoxin Journal, 1:31-37.

Alexander NJ, Proctor RH, McCormick SP. 2009. Genes, gene clusters, and biosynthesis of

trichothecenes and fumonisins in Fusarium. Toxin Reviews, 28:198–215.

Ali ML, Taylor JH, Jie L, Sun G, William M, Kasha KJ, Reid LM, Pauls KP. 2005.

Molecular mapping of QTLs for resistance to Gibberella ear rot, in corn, caused by

Fusarium graminearum. Genome, 48:521-533.

Alikhan N-F, Petty NK, Zakour NLB, Beatson SA. 2011. BLAST Ring Image Generator

(BRIG): simple prokaryote genome comparisons. BMC Genomics, 12:402.

Alisi A, Spaziani A, Anticoli S, Ghidinelli M, Balsano C. 2008. PKR is a novel functional

direct player that coordinates skeletal muscle differentiation via p38MAPK/AKT

pathways. Cellular Signalling, 20:534–542.

176

Amuzie CJ, Shinozuka J, Pestka JJ. 2009. Induction of suppressors of cytokine signaling by

the trichothecene deoxynivalenol in the mouse. Toxicological Science, 111:277–287.

Ansorge WJ. 2009. Next generation DNA sequencing techniques. New Biotechnology,

25:195-203.

Arnold JW, Senter SD. 1998. Use of digital aroma technology and SPME GC-MS to compare

volatile compounds produced by bacteria isolated from processed poultry. Journal of the

Science of Food and Agriculture, 78:343–348.

Audenaert K, Van Broeck R, Bekaert B, De Witte F, Heremans B, Messens K, Hofte

M, Haesaert G. 2009. Fusarium head blight (FHB) in Flanders: population

diversity, inter-species associations and DON contamination in commercial winter

wheat varieties. European Journal of Plant Pathology, 125:445-458.

Awad WA, Ghareeb K, Bohm J, Zentek J. 2010. Decontamination and detoxification

strategies for the Fusarium mycotoxin deoxynivalenol in animal feed and the

effectiveness of microbial biodegradation. Food Additives and Contaminants, 27:510-520.

Aziz K, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, Formsma K, Gerdes S, Glass

EM, Kubal M, Meyer F, Olsen GJ, Olson R, Osterman AL, Overbeek RA, McNeil LK,

Paarmann D, Paczian T, Parrello1 B, Pusch GD, Reich C, Stevens R, Vassieva O,

Vonstein V, Wilke A, Zagnitko O. 2008. The RAST Server: Rapid Annotations using

Subsystems Technology. BMC Genomics, 9:75.

Badger JL, Stins MF, Kim KS. 1999. Citrobacter freundii invades and replicates in human

brain microvascular endothelial cells. Infection and Immunology, 67:4208–4215.

Bae HK, Pestka JJ. 2008. Deoxynivalenol induces p38 interaction with the ribosome in

monocytes and macrophages. Toxicological Science, 105:59-66.

177

Bai GH, Desjardins AE, Plattner RD. 2002. Deoxynivalenol-nonproducing Fusarium

graminearum causes initial infection, but does not cause disease spread in wheat spikes.

Mycopathologia, 153:91–98.

Barrett JR. 2000. Mycotoxins: of molds and maladies. Environmental Health Perspectives,

108:A20–A23.

Bengoechea JA, Skurnik M. 2000. Temperature-regulated efflux pump/potassium antiporter

system mediates resistance to cationic antimicrobial peptides in Yersinia. Molecular

Microbiology, 37:67–80.

Bennett JW, Klich M. 2003. Mycotoxins. Clinical Microbiology Reviews, 16:497–516.

Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Sayers EW. 2009. GenBank. Nucleic

Acids Research, 37:D26-31.

Berdichevsky T, Friedberg D, Nadler C, Rokney A, Oppenheim A, Rosenshine I. 2005. Ler is

a negative autoregulator of the LEE1 operon in neuropathogenic Escherichia coli. Journal

of Bacteriology, 187:349–357.

Bernardo A, Bai GH, Guo PG, Xiao K, Guenzi AC, Ayoubi P. 2007. Fusarium graminearum-

induced changes in gene expression between Fusarium head blight-resistant and

susceptible wheat cultivars. Functional and Integrative Genomics, 7:69–77.

Bhat RV, Beedu SR, Ramakrishna Y, Munshi KL. 1989. Outbreak of trichothecene

mycotoxicosis associated with consumption of mould-damaged wheat production in

Kashmir Valley, India. The Lancet, 333:35–37.

Binder EM, Tan LM, Chin LJ, Handl J, Richard J. 2007. Worldwide occurrence of

mycotoxins in commodities feeds and feed ingredients. Animal Feed Science and

Technology, 137:265-282.

178

Birzele B, Meier A, Hindorf H, Krämer J, Dehne H-W. 2002. Epidemiology of Fusarium

infection and deoxynivalenol content in winter wheat in the Rhineland, Germany.

European Journal of Plant Pathology, 108:667–673.

Boddu J, Cho S, Muehlbauer GJ. 2007. Transcriptome analysis of trichothecene-induced gene

expression in barley. Molecular Plant–Microbe Interactions, 20:1364–1375.

Borenshtein D, Schauer DB. 2006. The Genus Citrobacter. Prokaryotes, 6:90–98.

Boutigny A-L, Richard-Forget F, Barreau C. 2008. Natural mechanisms for cereal resistance

to the accumulation of Fusarium trichothecenes. European Journal of Plant Pathology,

121:411–423.

Boutrif E, Canet C. 1998. Mycotoxin prevention and control: FAO programmes. Revue de

Médecine Véterinaire, 149:681-694.

Brennan JM, Leonard G, Fagan B, Cooke BM, Ritieni A, Ferracane R, et al. 2007.

Comparison of commercial European wheat cultivars to Fusarium infection of head and

seedling tissue. Plant Pathology, 56:55–64.

Brenner DJ, Grimont PA, Steigerwalt AG, Fanning GR, Ageron E, Riddle CF. 1993.

Classification of Citrobacteria by DNA hybridization: designation of Citrobacter farmeri

sp. nov., Citrobacter youngae sp. nov., Citrobacter braakii sp. nov., Citrobacter

werkmanii sp. nov., Citrobacter sedlakii sp. nov., and three unnamed Citrobacter

genomospecies. International Journal of Systematic Bacteriology, 43:645-58.

Brookes G, Barfoot P. 2006. GM Crops: The first ten years-global socio-economic and

environmental Impacts. no. 36. ISAAA (International Service for the Acquisition of Agri-

Biotech Applications), Ithaca, New York.

179

Bruce KD, Hughes MR. 2000. Terminal restriction fragment length polymorphism monitoring

of genes amplified directly from bacterial communities in soils and sediments. Molecular

Biotechnology, 16:261–269.

Buerstmayr H, Ban T, Anderson JA. 2009. QTL mapping and marker-assisted selection for

Fusarium head blight resistance in wheat: A review. Plant Breeding, 128:1–26.

Campbell H, Choo TM, Vigier B, Underhill L. 2002. Comparison of mycotoxin profiles

among cereal samples from eastern Canada. Canadian Journal of Botany, 80:526–532.

Carver et al. 2005. ACT: the Artemis Comparison Tool. Bioinformatics, 21:3422-3423.

Carver et al. 2009. DNAPlotter: circular and linear interactive genome visualization.

Bioinformatics, 25:119-120.

Chen CY, et al. 2003. Comparative genome analysis of Vibrio vulnificus, a marine pathogen.

Genome Research, 13:2577–2587.

Chiu TC, Yen JH, Hsieh YN, Wang YS. 2005. Reductive transformation of dieldrin under

anaerobic sediment culture. Chemosphere, 60:1182–1189.

Chung YJ, Zhou HR, Pestka JJ. 2003. Transcriptional and posttranscriptional roles for p38

mitogen-activated protein kinase in upregulation of TNF-alpha expression by

deoxynivalenol (vomitoxin). Toxicology and Applied Pharmacology, 193:188–201.

Chung EJ, Lim HK, Kim J-C, Choi GJ, Park EJ, Lee MH, Chung YR, Lee S-W. 2008. Forest

soil metagenome gene cluster involved in antifungal activity expression in Escherichia

coli. Applied and Environmental Microbiology, 74:723-730.

Collins TFX, Sprando RL, Black TN, Olejnik N, Eppley RM, Hines FA, Rorie J, Ruggles DI.

2006. Effects of deoxynivalenol (DON, vomitoxin) on in utero development in rats. Food

and Chemical Toxicology, 44:747–757.

180

Cong Y, Wang J, Chen Z, Xiong K, Xu Q, Hu F. 2011. Characterization of swarming motility

in Citrobacter freundii. FEMS Microbiology Letters, 371:160-171

Cossette F, Miller JD. 1995. Phytotoxic effect of deoxynivalenol and Gibberella ear rot

resistance of corn. Natural Toxins, 3:383–388.

Creppy EE. 2002. Update of survey, regulation and toxic effects of mycotoxins in Europe.

Toxicological Letters, 127:19-28.

Dai JS, Wang D, Guerlebeck C, Laturnus S, Guenther Z, Lu SC, Ewers C. 2010. Suppression

subtractive hybridization identifies an autotransporter adhesin gene of E. coli IMT5155

specifically associated with avian pathogenic Escherichia coli (APEC). BMC

Microbiology, 10:236.

Daniel R. 2005. The metagenomics of soil. Nature Review Microbiology, 3:470-478.

Davis KER, Joseph SJ, Janssen PH. 2005. Effects of growth medium, inoculum size, and

incubation time on culturability and isolation of soil bacteria. Applied and Environmental

Microbiology, 71:826-834.

Delcher AL, Phillippy A, Carlton J, Salzberg SL. 2002. MUMmer 2.1, NUCmer, and

PROmer are described in "Fast Algorithms for Large-scale Genome Alignment and

Comparison”. Nucleic Acids Research, 30:2478-2483.

Demain AL, Adrio JL. 2008. Contributions of microorganisms to industrial biology.

Molecular Biotechnology, 38:41–455.

Deng W, Li Y, Vallance BA, Finlay BB. 2001. Locus of enterocyte efffacement from

Citrobacter rodentium: sequence analysis and evidence for horizontal transfer among

attaching and effacing pathogens. Infection and Immunity, 69:6323–6335.

181

Desjardins AE, Proctor RH, Bai G, McCormick SP, Shaner G, Beuchley G, Hohn TM. 1996.

Reduced virulence of trichothecene-nonproducing mutants of Gibberella zeae in wheat

field tests. Molecular Plant-Microbe Interactions, 9:775-781.

Desjardins AE. 2006. Fusarium mycotoxins: chemistry, genetics, and biology. The American

Phytopathological Society, St. Paul, Minnesota, USA, p 260.

Desjardins AE, Proctor RH. 2007. Molecular biology of Fusarium mycotoxins. International

Journal of Food Microbiology, 119:47-50.

Desmond OJ, Manners JM, Stephens AE, MacLean DJ, Schenk PM, Gardiner DM, Munn

AL, Kazan K. 2008. The Fusarium mycotoxin deoxynivalenol elicits hydrogen peroxide

production, programmed cell death and defence responses in wheat. Molecular Plant

Pathology, 9:435-445.

Dill-Macky R, Jones RK. 2000. The effect of previous crop residues and tillage on Fusarium

head blight of wheat. Plant Disease, 84:71–76.

D’Mello JPF, Macdonald AMC, Dijksma WTP. 1998. 3-Acetyl deoxynivalenol and esterase

production in a fungicide-insensitive strain of Fusarium culmorum. Mycotoxin Research,

14:9–18.

D’Mello JPF, Placinta CM, Macdonald AMC. 1999. Fusarium mycotoxins: a review of

global implications of animal health, welfare and productivity. Animal Feed Science and

Technology, 80:183-205.

Doohan FM, Brennan J, Cooke BM. 2003. Influence of climatic factors on Fusarium species

pathogenic to cereals. European Journal of Plant Pathology, 109:755–768.

182

Dyer RB, Kendra DF, Brown DW. 2006. Real-time PCR assay to quantify Fusarium

graminearum wild-type and recombinant mutant DNA in plant material. Journal of

Microbiological Methods, 67:534–542.

Erecinska M, Wilson DF. 1981. Inhibitors of cytochrome c oxidase. In: inhibitors of

mitochondrial function, edited by M. Erecinska and D.F. Wilson. Pergamon Press Ltd.,

Oxford. pp 145-160.

Eriksen GS. 2003. Metabolism and toxicity of trichothecene. Doctoral thesis. Swedish

University of Agricultural Sciences. Uppsala, Sweden.

Eriksen GS, Pettersson H, Lundh T. 2004. Comparative cytotoxicity of deoxynivalenol,

nivalenol, their acetylated derivatives and de-epoxy metabolites. Food Chemistry and

Toxicology, 42:619-624.

Eudes F, Comeau A, Rioux S, Collin J. 2000. Phytotoxicity of eight mycotoxins associated

with Fusarium in wheat head blight. Canadian Journal of Plant Pathology, 22:286–292.

FAO (Food and Agriculture Organization). 1991. Food Nutrition and Agriculture. Food for

the Future, volume 1. FAO, Rome.

FAO (Food and Agriculture Organization). 2001. FNP73, Manual of the application of

HACCP system in mycotoxin prevention and control. FAO Food and Nutrition paper no.

73, FAO, Rome.

Fedorak PM, Grbic-Galic D. 1991. Aerobic microbial cometabolism of benzothiophene and

3-methylbenzothiophene. Applied and Environmental Microbiology, 57:932-940.

Felsenstein J. 1989. PHYLIP - Phylogeny Inference Package (Version 3.2). Cladistics, 5:164-

166.

183

Fernando WGD, Amarasinghe C, Demoz B, Surujdeo‐Maharaj S, Gauthier V, Guerrieri G,

Al‐Taweel K, Brûlé‐Babel A, Tamburic‐Illincic L, Gilbert J. 2011. What do we know of

the Fusarium graminearum chemotypes, their epidemiology, genetics, and infection

process in Canada? In: Proc. 7th Canadian Workshop on Fusarium Head Blight, edited by

J Gilbert, A Tekauz, N Sweetland and K Slusarenko. Winnipeg, Manitoba, Canada, p 45.

Fierer N, Jackson RB. 2006. The diversity and biogeography of soil bacterial community.

Proceedings of the National Academy of Sciences, 103:626-631.

Fink-Gremmels J. 1989. The significance of mycotoxin assimilation for meat animals.

Deutsche Tierarztliche Wochenschrift, 96:360-363.

Flint DH, Emptage MH, Finnegan MG, Fu W, Johnson MK. 1993. The role and properties of

the iron-sulfur cluster in Escherichia coli dihydroxy-acid dehydratase. Journal of

Biological Chemistry, 268:14732–14742.

Foroud NA, Eudes F. 2009. Trichothecenes in cereal grains. International Journal of

Molecular Science, 10:147-173.

Forsell JH, Jensen R, Tai JH, Witt M, Lin WS, Pestka JJ. 1987. Comparison of acute

toxicities of deoxynivalenol (vomitoxin) and 15-acetyldeoxynivalenol in the B6C3F1

mouse. Food and Chemical Toxicology, 25:155–162.

Frey M, Stettner C, Gierl A. 1998. A general method for gene isolation in tagging approaches:

amplification of insertion mutagenised sites (AIMS). The Plant Journal, 13:717-721.

Freeman GG, Morrison RI. 1949. The isolation and chemical properties of trichothecin, an

antifungal substance from Trichothecium roseum Link. Biochemical Journal, 44:1–5.

184

Fuchs E, Binder EM, Heidler D, Krska R. 2002. Structural characterization of metabolites

after the microbial degradation of type A trichothecenes by the bacterial strain BBSH 797.

Food Additives and Contaminants, 19:379–386.

Gabor EM, Alkema WBL, Janssen DB. 2004. Quantifying the accessibility of the

metagenome by random expression cloning techniques. Environmental Microbiology,

6:879–886.

Gadd GM, Griffiths AJ. 1978. Microorganisms and heavy metal toxicity. Microbial Ecology,

4:303–317.

Gale LR, Bryant JD, Calvo S, Giese H. 2005. Chromosome complement of the fungal plant

pathogen Fusarium graminearum based on genetic and physical mapping and cytological

observations. Genetics, 171:985–1001.

Gao LY, Pak M, Kish R, Kajihara K, Brown EJ. 2006. A mycobacterial operon essential for

virulence in vivo and invasion and intracellular persistence in macrophages. Infection and

Immunity, 74:1757–1767.

Gareau MG, Wine E, Reardon C, et al. 2010. Probiotics prevent death caused by Citrobacter

rodentium infection in neonatal mice. Journal of Infectious Diseases, 201:81e91.

Gareis M, Bauer J, Thiem J, Plank G, Grabley S, Gedek B. 1990. Cleavage of

zearalenoneglycoside, a masked mycotoxin, during digestion in swine. Journal of

Veterinary Medicine, 37:236-240.

Geng L, Xin W, Haung DW, Feng G. 2006. A universal cloning vector using vaccinia

topoisomerase I. Molecular Biotechnology, 33:23-28.

185

Golkari S, Gilbert J, Ban T, Procunier JD. 2009. QTLspecific microarray gene expression

analysis of wheat resistance to Fusarium head blight in Sumai-3 and two susceptible

NILs. Genome, 52:409–418.

Goodstadt L, Ponting CP. 2004. Vitamin K epoxide reductase: homology, active site and

catalytic mechanism. Trends in Biochemical Sciences, 29:289-292.

Goswami RS, Kistler HC. 2004. Heading for disaster: Fusarium graminearum on cereal

crops. Molecular Plant Pathology, 5:515-525.

Gouze ME, Laffitte J, Printon P, Dedieux G, Galinier A, Thouvenot JP, Loiseau N, Oswald

IP, Galtier P. 2007. Effect of subacute oral doses of nivalenol on immune and metabolic

defence systems in mice. Veterinary Research, 38:635–646.

Guan S, He J, Young JC, Zhu H, Li X, Ji C, Zhou T. 2009. Transformation of trichothecene

mycotoxins by microorganisms from fish digesta. Aquaculture, 290:290–295.

Gupta R, Beg QK, Lorenz P. 2002. Bacterial alkaline proteases: molecular approaches and

industrial applications. Applied Microbiology and Biotechnology, 59:15–32.

Haberle J, Schweizer G, Schondelmaier J, Zimmermann G, Hart L. 2009. Mapping of QTL

for resistance against Fusarium head blight in the winter wheat population

Pelikan//Bussard/Ning8026. Plant Breeding, 128:27–35.

Hall BG. 2004. Predicting the evolution of antibiotic resistance genes. Nature Reviews

Microbiology, 2:430-435.

Han JI, et al. 2011. Complete genome sequence of the metabolically versatile plant growth-

promoting endophyte Variovorax paradoxus S110. Journal of Bacteriology, 193:1183–

1190.

186

Harvey RB, Edrington TS, Kubena LF, Elissalde MH, Corrier DE, Rottinghaus GE. 1995.

Effect of aflatoxin and diacetoxyscirpenol in ewe lambs. Bulletin of Environmental

Contamination and Toxicology, 54:325-330.

Harvey RB, Edrington TS, Kubena LF, Elissalde MH. Casper HH, Rottinghaus GE, Turk JR.

1996. Effects of dietary fumonisin B1-containing culture material, deoxynivalenol-

contaminated wheat or their combination on growing barrows. American Journal of

Veterinary Research, 57:1790-1794.

He J. 2007. Microbial strategies for management of Fusarium graminearum and

transformation of deoxynivalenol. Master of Science Thesis. University of Guelph,

Guelph, Ontario, Canada.

He K, Vines L, Pestka JJ. 2010a. Deoxynivalenol-induced modulation of microRNA

expression in RAW 264.7 macrophages-A potential novel mechanism for translational

inhibition. Toxicologist, 114:310.

He J, Zhou T, Young JC, Boland GJ, Scott PM. 2010b. Chemical and biological

transformations for trichothecene mycotoxins in human and animal food chains: A review.

Trends in Food Science and Technology, 21:67-76.

He J, Zhou T. 2010. Patented techniques for detoxification of mycotoxins in feeds and food

matrices. Rec Pat Food Nutr Agric., 2:96–104.

He P, Young LG, Forsberg C. 1992. Microbial transformation of deoxynivalenol (vomitoxin).

Applied and Environmental Microbiology, 58:3857-3863.

Hell K, Cardwell KF, Setamou M, Poehling H-M. 2000. The influence of storage practices on

aflatoxins contamination in maize in four agroecological zones in Benin, West Africa.

Journal of Stored Product Research, 36:365–82.

187

Henne A, Schmitz RA, Bomeke M, Gottschalk G, Daniel R. 2000. Screening of

environmental DNA libraries for the presence of genes conferring lipolytic activity on

Escherichia coli. Applied and Environmental Microbiology, 66:3113–3116.

Hodges GR, Degener CE, Barnes WG. 1978. Clinical significance of Citrobacter isolates.

American Journal of Clinical Pathology, 70:37–40.

Hollinger K, Ekperigin HE. 1999. Mycotoxicosis in food producing animals. The Veterinary

Clinics of North America Food Animal Practice, 15:133-165.

Hong Y, Kinoshita T. 2009. Trypanosome glycosylphosphatidylinositol biosynthesis. Korean

Journal of Parasitology, 47:197–204.

Hope RJ, Aldred D, Magan N. 2005. Comparison of environmental profiles for growth and

deoxynivalenol production by Fusarium culmorum and F. graminearum on wheat grain.

Letters in Applied Microbiology, 40:295–300.

Horvath RS. 1972. Microbial co-metabolism and the degradation of organic compounds in

nature. Bacteriological Reviews, 36:146–155.

Hsu D, Shih LM, Zee YC. 1994. Degradation of rRNA in Salmonella strains: a novel

mechanism to regulate the concentrations of rRNA and ribosomes. Journal of

Bacteriology, 176:4761–4765.

Huwig A, Freimund S, Kappeli O, Dutler H. 2001. Mycotoxin detoxification of animal feed

by different adsorbents. Toxicological Letters, 122:179–188.

Iordanov MS, Pribnow D, Magun JL, Dinh TH, Pearson JA, Magun BE. 1997. Ribotoxic

stress response: activation of the stress activated protein kinase JNK1 by inhibitors of the

peptidyl transferase reaction and by sequence-specific RNA damage to the sarcin/ricin

loop in the 28S rRNA. Molecular and Cellular Biology, 17:3373–3381.

188

Islam R, Zhou T, Young JC, Goodwin PH, Pauls KP. 2012. Aerobic and anaerobic de-

epoxidation of mycotoxin deoxynivalenol by bacteria originating from agricultural soil.

World Journal of Microbiology and Biotechnology, 28:7-13.

Jackson LS, Bullerman LB. 1999. Effect of processing on Fusarium mycotoxins. Advances in

Experimental Medicine and Biology, 459:243-261.

Jansen C, von Wettstein D, Schafer W, Kogel K-H, Felk A, Maier FJ. 2005. Infection patterns

in barley and wheat spikes inoculated with wild-type and trichodiene synthase gene

disrupted Fusarium graminearum. Proceedings of the National Academy of Sciences,

102:16892–16897.

Jube S, Borthakur D. 2007. Expression of bacterial genes in transgenic tobacco: methods,

applications and future prospects. Electronic Journal of Biotechnology, 10:452-467.

Kabak B, Dobson AD. 2009. Biological strategies to counteract the effects of mycotoxins.

Journal of Food Protection, 72:2006-2016.

Kant P, Reinprecht Y, Martin CJ, Islam R, Pauls KP. 2011. Integration of biotechnologies,

disease resistance, pathology and Fusarium. In: Comprehensive Biotechnology, edited by

Moo-Young M. Second Edition, 4:729-743.

Kaplan CW, Kitts CL. 2004. Bacterial succession in a petroleum land treatment unit. Applied

and Environmental Microbiology, 70:1777-1786.

Kaprelyants AS, Kell DB. 1996. Do bacteria need to communicate with each other for

growth? Trends in Microbiology, 4:237-42.

189

Kariholu U, Rawal J, Namnyak S. 2009. Neonatal Citrobacter koseri meningitis and brain

abscess. The Internet Journal of Pediatrics and Neonatology, 10

(http://www.ispub.com/journal/the-internet-journal-of-pediatrics-and-

neonatology/volume-10-number-1/neonatal-citrobacter-koseri-meningitis-and-brain-

abscess.html) (accessed on April 19, 2012)

Karlovsky P. 1999. Biological detoxification of fungal toxins and its use in plant breeding,

feed, and food production. Natural Toxins, 7:1–23.

Karlovsky P. 2011. Biological detoxification of the mycotoxin deoxynivalenol and its use in

genetically engineered crops and feed additives. Applied Microbiology and

Biotechnology, 91:491-504.

Kavanagh KL, Jornvall H, Persson B, Oppermann U. 2008. Medium and short-chain

dehydrogenase/reductase gene and protein families: the SDR superfamily: functional and

structural diversity within a family of metabolic and regulatory enzymes. Cellular and

Molecular Life Sciences, 65:3895–3906.

Khera KS, Whalen C, Angers G, Vesonder RF, Kuiper-Goodman T. 1982. Embryotoxicity of

4-deoxynivalenol (vomitoxin) in mice. Bulletin of Environmental Contamination and

Toxicology, 29:487–491.

Kikot GE, Hours RA, Alconada TM. 2009. Contribution of cell wall degrading enzymes to

pathogenesis of Fusarium graminearum: a review. Journal of Basic Microbiology, 49:

231–241.

Kim UJ, Shizuya H, Jong PJ, Birren B, Simon MI. 1992. Stable propogation of cosmid sized

human DNA inserts in a F factor based vector. Nucleic Acids Research, 20:1083-1065.

190

Kimura M, Kaneko I, Komiyama M, Takatsuki A, Koshino H, Yoneyama K, Yamaguchi I.

1998. Trichothecene 3-O-acetyltransferase protects both the producing organism and

transformed yeast from related mycotoxins, cloning and characterization of Tri101.

Journal of Biological Chemistry, 273:1654-1661.

Kimura M, Anzai H, Yamaguchi I. 2001. Microbial toxins in plant–pathogen interactions:

biosynthesis, resistance mechanisms, and significance. The Journal of General and

Applied Microbiology, 47:149–160.

Kimura M, Tokai T, Takahashi-Ando N, Ohsato S, Fujimura M. 2007. Molecular and genetics

studies of Fusarium trichothecene biosynthesis: pathways, genes and evolution.

Bioscience, Biotechnology and Biochemistry, 71:2105-2123.

King RR, McQueen RE, Levesque D, Greenhalgh R. 1984. Transformation of deoxynivalenol

(vomitoxin) by rumen microorganisms. Journal of Agricultural and Food Chemistry,

32:181–1183.

Klappenbach JA, Dunbar JM, Schmidt TM. 2000. rRNA operon copy number reflects

ecological strategies of bacteria. Applied and Environmental Microbiology, 66:1328–

1333.

Knietsch A, Waschkowitz T, Bowien S, Henne A, Daniel R. 2003. Metagenomes of complex

microbial consortia derived from different soils as sources for novel genes conferring

formation of carbonyls from short-chain polyols on Escherichia coli. Journal of Molecular

Microbiology and Biotechnology, 5:46–56.

Kollarczik B, Gareis M, Hanelt M. 1994. In vitro transformation of the Fusarium mycotoxins

deoxynivalenol and zearalenone by the normal gut microflora of pigs. Natural Toxins,

2:105–110.

191

Koonin EV, Wolf YI, Aravind L. 2001. Prediction of the archaeal exosome and its

connections with the proteasome and the translation and transcription machineries by a

comparative-genomic approach. Genome Research, 11:240–252.

Kouadio JH, Dano SD, Moukha S, Mobio TA, Creppy EE. 2007. Effects of combinations of

Fusarium mycotoxins on the inhibition of macromolecular synthesis, malondialdehyde

levels, DNA methylation and fragmentation, and viability in Caco-2 cells. Toxicon,

49:306–317.

Krause L, Diaz NN, Bartels D, Edwards RA, Pühler A, et al. 2006. Finding novel genes in

bacterial communities isolated from the environment. Bioinformatics, 22:281–289.

Kubena LF, Huff WE, Harvey RB, Phillips TD, Rottinghaus GE. 1989. Individual and

combined toxicity of deoxynivalenol and T-2 toxin in broiler chicks. Poultry Science,

68:622–626.

Kubena LF, Harvey RB, Huff WE, Corrier DE, Phillips TD, Rottinghaus GE. 1990. Efficacy

of hydrated sodium calcium aluminosilicate to reduce the toxicity of aflatoxin and T-2

toxin. Poultry Science, 69:1078–1086.

Kubena LF, Harvey RB, Huff WE, Elissalde MH, Yersin AG, Phillips TD, Rottinghaus GE.

1993. Efficacy of hydrated sodium calcium aluminosilicate to reduce the toxicity of

aflatoxin and diacetoxyscirpenol. Poultry Science, 72:51–59.

Kubena LF, Edrington TS, Harvey RB, Buckley SA, Phillips TD, Rottinghaus GE, Casper

HH. 1997. Individual and combined effects of fumonisin B1 present in Fusarium

moniliforme culture material and T-2 toxin or deoxynivalenol in broiler chicks. Poultry

Science, 76:1239–1247.

192

Kumar RA, Gunasekaran P, Lakshmanan M. 1999. Biodegradation of tannic acid by

Citrobacter freundii isolated from a tannery effluent. Journal of Basic Microbiology,

39:161-168.

Kushiro M. 2008. Effects of milling and cooking processes on the deoxynivalenol content in

wheat. International Journal Molecular Science, 9:2127-2145.

Lane DJ. 1991. 16S/23S rRNA sequencing. In: Nucleic acid techniques in bacterial

systematic, edited by Stackebrandt E and Goodfellow M. Wiley J and Sons Inc., New

York, pp 115-175.

Lancova K, Hajslova J, Poustka J, Krplova A, Zachariasova M, Dostalek P, Saschambula L.

2008. Transfer of Fusarium mycotoxins and “masked” deoxynivalenol (deoxynivalenol-3-

glucoside) from field barley through malt to beer. Food Additives and Contaminants,

25:732–744.

Langmead B, Trapnell C, Pop M, Salzberg SL. 2009. Ultrafast and memory-efficient

alignment of short DNA sequences to the human genome. Genome Biology, 10:R25.

Larsen JC, Hunt J, Perrin I, Ruckenbauer P. 2004. Workshop on trichothecenes with a focus

on DON: summary report. Toxicological Letters, 153:1-22.

Lee T, Oh DW, Kim HS, Lee J, Kim YH, Yun SH, Lee YW. 2001. Identification of

deoxynivalenol- and nivalenol-producing chemotypes of Gibberella zeae by using PCR.

Applied and Environmental Microbiology, 67:2966–2972.

Lee S-W, Won K, Lim HK, Kim J-C, Choi GJ, Cho KY. 2004. Screening for novel lipolytic

enzymes from uncultured soil microorganisms. Applied and Environmental Microbiology,

65:720-726.

193

Li L, Christian J, Stoeckert J, David SR. 2003. OrthoMCL: Identification of Ortholog Groups

for Eukaryotic Genomes. Genome Research, 13:2178-2189.

Li T, Chang CY, Jin DY, Lin PJ, Khvorova A, Stafford DW. 2004. Identification of the gene

for vitamin K epoxide reductase. Nature, 427:541–544.

Li Y. 2007. Sensory signal transduction in the vagal primary afferent neurons. Current

Medicinal Chemistry, 14:2554–2563.

Li MX, Pestka JJ. 2008. Comparative induction of 28S ribosomal RNA cleavage by ricin and

the trichothecenes deoxynivalenol and T-2 toxin in the macrophage. Toxicological

Science, 105:67–78.

Li R, Li Y, Kristiansen K, Wang J. 2008. SOAP: short oligonucleotide alignment program.

Bioinformatics, 24:713-4.

Li W, Schulman S, Dutton RJ, Boyd D, Beckwith J, Rapoport TA. 2010. Structure of a

bacterial homologue of vitamin K epoxide reductase. Nature, 463:507-512.

Lichstein HC, Malcolm HS. 1943. Studies of the effect of sodium azide on microbic growth

and respiration. Journal of Bacteriology, 47:221–230.

Limay-Rios V, Schaafsma AW. 2011. 2010 Update on the development of an integrated

mycotoxin management system in Ontario corn. 30th Centralia Swine Research Update,

Kirkton Ontario, Canada.

Llorens A, Hinojo MJ, Mateo R, Gonzalez-Jaen MT, Valle- Algarra FM, Logrieco A, Jimenez

M. 2006. Characterization of Fusarium spp. isolates by PCR-RFLP analysis of the

intergenic spacer region of the rRNA gene (rDNA). International Journal of Food

Microbiology, 106:287–306.

194

Lombaert GA, Pellaers P, Roscoe V, Mankotia M, Neil R, Scott PM. 2003. Mycotoxins in

infant cereal foods from the Canadian retail market. Food Additives and Contaminants,

20:494–504.

Luengo JM, García JL, Olivera ER. 2001. The phenylacetyl-CoA catabolon: a complex

catabolic unit with broad biotechnological applications. Molecular Microbiology,

39:1434–1442.

Luo X. 1994. Food poisoning caused by Fusarium toxins. In: Proceedings of the Second

Asian Conference on Food Safety. ILSI Thailand, pp 129–136.

Mabagala RB. 1997. The effect of populations of Xanthomonas campestris pv. phaseoli in

bean reproductive tissues on seed infection of resistant and susceptible bean genotypes.

European Journal of Plant Pathology, 103:175–181.

Madhyastha MS, Marquardt RR, Abramson D. 1994. Structure–activity relationships and

interactions among trichothecene mycotoxins as assessed by yeast bioassay. Toxicon,

32:1147–1152.

Maier FJ, Miedaner T, Hadeler B, Felk A, Salomon S, Lemmens M, Kassner H, Schäfer W.

2006. Involvement of trichothecenes in fusarioses of wheat, barley and maize evaluated

by gene disruption of the trichodiene synthase (Tri5) gene in three field isolates of

different chemotype and virulence. Molecular Plant Pathology, 7:449–461.

Malojcic G, Owen RL, Grimshaw JP, Glockshuber R. 2008. Preparation and structure of the

charge-transfer intermediate of the transmembrane redox catalyst DsbB. FEBS Letters,

582:3301–3307.

195

Manoharan M, Dahleen LS, Hohn TM, Neate SM, Yu XH, Alexander NJ, et al. 2006.

Expression of 3-OH trichothecene acetyltransferase in barley (Hordeum vulgare L.) and

effects on deoxynivalenol. Plant Science, 171:699–706.

Maresca M, Mahfoud R, Garmy N, Fantini J. 2002. The mycotoxin deoxynivalenol affects

nutrient absorption in human intestinal epithelial cells. The Journal of Nutrition,

132:2723–2731.

Martins ML, Martins HM. 2002. Influence of water activity, temperature and incubation time

on the simultaneous production of deoxynivalenol and zearalenone in corn (Zea mays) by

Fusarium graminearum. Food Chemistry, 79:315–318.

Masuda D, Ishida M, Yamaguchi K, Yamaguchi I, Kimura M, Nishiuchi T. 2007. Phytotoxic

effects of trichothecenes on the growth and morphology of Arabidopsis thaliana. Journal

of Experimental Botany, 58:1617–1626.

Matsumoto E, Kawanaka Y, Yun SJ, Oyaizu H. 2008. Isolation of dieldrin- and endrin-

degrading bacteria using 1,2-epoxycyclohexane as a structural analog of both compounds.

Applied Microbiology and Biotechnology, 80:1095-1103.

McLean M. 1996. The phytotoxicity of Fusarium metabolites: an update since 1989.

Mycopathologia, 133:163–179.

Merhej J, Richard-Forget F, Barreau C. 2011. Regulation of trichothecene biosynthesis in

Fusarium: recent advances and new insights. Applied Microbiology and Biotechnology,

91:519–528.

Midwest Plan Service. 1987. Grain Drying, Handling, Storage Handbook. Ames, Iowa, Iowa

State University. 2nd edition.

196

Miller SS, Reid LM, Harris LJ. 2007. Colonization of maize silks by Fusarium graminearum,

the causative organism of Gibberella ear rot. Canadian Journal of Botany, 85:369-376.

Minervini F, Fornelli F, Flynn KM. 2004. Toxicity and apoptosis induced by mycotoxins

nivalenol, deoxynivalenol and fumonisin B1 in a human erythroleukemia cell line.

Toxicology In Vitro, 18:21–28.

Mitterbauer R, Poppenberger B, RaditschnigA, Lucyshyn D, Lemmens M, Glossl J, Adam G.

2004. Toxin-dependent utilization of engineered ribosomal protein L3 limits trichothecene

resistance in transgenic plants. Plant Biotechnology Journal, 2:329–340.

Mocali S, Benedetti A. 2010. Exploring research frontiers in microbiology: the challenge of

metagenomics in soil microbiology. Research Microbiology, 161:497-505.

Morozkina EV, Zvyagilskaya RA. 2007. Nitrate reductases: structure, functions and effect of

stress factors. Biochemistry, 72:1151–1160.

Munkvold GP. 2003. Cultural and genetic approaches to managing mycotoxins in maize.

Annual Review of Phytopathology, 41:99-116.

Mundy R, MacDonald TT, Dougan G, Frankel G, Wiles S. 2005. Citrobacter rodentium of

mice and man. Cellular Microbiology, 7:1697–1706.

Nakamura N, Inoue N, Watanabe R, Takahashi M, Takeda J, Stevens VL, Kinoshita T. 1997.

Expression cloning of PIG-L, a candidate N-acetylglucosaminyl-phosphatidylinositol

deacetylase. The Journal of Biological Chemistry, 272:15834–15840.

Nganje WE, Bangsund DA, Leistritz FL, Wilson WW, Tiapo NM. 2002. Estimating the

economic impact of a crop disease: the case of Fusarium head blight in U.S. wheat and

barley. National Fusarium Head Blight Forum Proceedings. East Lansing: Michigan State

University, pp 275–281.

197

Nielsen C, Lippke H, Didier A, Dietrich R, Martlbauer E. 2009. Potential of deoxynivalenol

to induce transcription factors in humanhepatoma cells. Molecular Nutrition and Food

Research, 53:479–491.

Nielsen JKS, Vikström AC, Turner P, Knudsen LE. 2011. Deoxynivalenol transport across

the human placental barrier. Food and Chemical Toxicology. 49:2046-2052.

Nishiuchi T, Masuda D, Nakashita H, Ichimura K, Shinozaki K, Yoshida S, et al. 2006.

Fusarium phytotoxin trichothecenes have an elicitor-like activity in Arabidopsis thaliana,

but the activity differed significantly among their molecular species. Molecular Plant–

Microbe Interactions, 19:512–520.

Ohsato S, Ochiai-Fukuda T, Nishiuchi T, Takahashi-Ando N, Koizumi S, Hamamoto H,

Kudo T, Yamaguchi I, Kimura M. 2007. Transgenic rice plants expressing trichothecene

3-O-acetyltransferase show resistance to the Fusarium phytotoxin deoxynivalenol. Plant

Cell Reports, 26:531-538.

Okubara PA, Blechl AE, McCormick SP, Alexander NJ, Dill-Macky R, Hohn TM. 2002.

Engineering deoxynivalenol metabolism in wheat through the expression of a fungal

trichothecene acetyltransferase gene. Theoritical and Applied Genetics, 106:74-83.

Ong CL, Beatson SA, Totsika M, Forestier C, McEwan AG, Schembri MA. 2010. Molecular

analysis of type 3 fimbrial genes from Escherichia coli, Klebsiella and Citrobacter

species. BMC Microbiology, 10:183.

Osborne LE, Stein JM. 2007. Epidemiology of Fusarium head blight on small-grain cereals.

International Journal of Food Microbiology, 119:103-108.

Page RDM. 1996. TREEVIEW: An application to display phylogenetic trees on personal

computers. Computer Applications in the Biosciences, 12:357-358.

198

Parkhill J, Dougan G, James KD, Thomson NR, Pickard D, Wain J, Churcher C, Mungall KL,

Bentley SD, Holden MT, et al. 2001. Complete genome sequence of a multiple drug

resistant Salmonella enterica serovar Typhi CT18. Nature, 413:848–852.

Paterson RRM, Lima N. 2010. How will climate change affect mycotoxins in food? Food

Research International, 43:1902–14.

Paul PA, El-Allaf SM, Lipps PE, Madden LV. 2004. Rain splash dispersal of Gibberella zeae

within wheat canopies in Ohio. Phytopathology, 94:1342-1349.

Pestka JJ, Lin WS, Miller ER. 1987. Emetic activity of the trichothecene 15-

acetyldeoxynivalenol in pigs. Food and Chemical Toxicology, 25:855–858.

Pestka JJ, Smolinski AT. 2005. Deoxynivalenol: Toxicology and potential effects on humans.

Journal of Toxicology Environmental Health, 8:39–69.

Pestka JJ. 2007. Deoxynivalenol: toxicity, mechanisms and animal health risks. Animal Feed

Science and Technology, 137:283-298.

Pestka JJ. 2008. Mechanisms of deoxynivalenol-induced gene expression and apoptosis. Food

Additives and Contaminants, 25:1128-1140.

Pestka J.2010. Toxicological mechanisms and potential health effects of deoxynivalenol and

nivalenol. World Mycotoxin Journal, 3:323–347.

Petersohn A, Brigulla M, Haas S, Hoheisel JD, Völker U, Hecker M. 2001. Global Analysis

of the General Stress Response of Bacillus subtilis Journal of Bacteriology, 183:5617-

5631.

199

Petty NK, Bulgin R, Crepin VF, Cerdeno-Tarraga AM, Schroeder GN, Quail MA, Lennard N,

Corton C, Barron A, Clark L, Toribio AL, Parkhill J, Dougan G, Frankel G, Thomson NR.

2010. The Citrobacter rodentium genome sequence reveals convergent evolution with

human pathogenic Escherichia coli. Journal of Bacteriology, 192:525–538.

Petty NK, Feltwell T, Pickard D, Clare S, Toribio AL, et al. 2011. Citrobacter rodentium is

an unstable pathogen showing evidence of significant genomic flux. PLoS Pathology,

7:e1002018. doi:10.1371/journal.ppat.1002018.

Pittet A. 2001. Natural occurrence of mycotoxins in foods and feeds: A decade in review. In:

Mycotoxins and phytotoxins in perspective at the turn of the millennium, edited by De

Koe WJ, Samson RA, Van Egmond HP, Gilbert J, Sabino M. W. J. de Koe. Wageningen,

Netherlands, pp 153–172.

Placinta CM, D’Mello JPF, Macdonald AMC. 1999. A review of worldwide contamination of

cereal grains and animal feed with Fusarium mycotoxins. Animal Feed Science and

Technology, 78:21–37.

Plank B, Schuh M, Binder EM. 2009. Investigations on the effect of two feed additives,

Biomin BBSH 797 and Mycofix Plus (R) 3.E, as detoxificants of DON contaminated feed

of piglets. Wien Tierarztl Monatssch, 96:55–71.

Poapolathep A, Ohtsuka R, Kiatipattanasakul W, Ishigami N, Nakayama H, Doi K. 2002.

Nivalenol-induced apoptosis in thymus, spleen and Peyer’s patches of mice. Experimental

and Toxicological Pathology, 53:441–446.

200

Poppenberger B, Berthiller F, Lucyshyn D, Sieberer T, Schuhmacher R, Krska R, Kuchler K,

Glössl J, Luschnig C, Adam G. 2003. Detoxification of the Fusarium mycotoxin

deoxynivalenol by a UDP-glucosyltransferase from Arabidopsis thaliana. Journal of

Biological Chemistry, 278:47905–47914.

Prelusky DB, Trenholm HL. 1993. The efficacy of various classes of anti-emetics in

preventing deoxynivalenol-induced vomiting in swine. Natural Toxins, 1:296–302.

Proctor RH, Hohn TM, McCormick SP. 1995. Reduced virulence of Gibberella zeae caused

by disruption of a trichothecene toxin biosynthetic gene. Molecular Plant-Microbe

Interactions, 8:593-601.

Puchenkova SG. 1996. Enterobacteria in areas of water along the Crimean Coast.

Mikrobiolohichnyĭ zhurnal. 58: 3-7.

Qi BW, Hao BL. 2004. Whole-proteome prokaryote phylogeny without sequence alignment:

a K-string composition approach. Journal of Molecular Evolution, 58:1–11.

Quarta A, Mita G, Haidukowski M, Logrieco A, Mule G, Visconti A, 2006. Multiplex PCR

assay for the identification of nivalenol, 3- and 15-acetyl-deoxynivalenol chemotypes in

Fusarium. FEMS Microbiological Letters, 259:7–13.

Rehr B, Klemme JH. 1989. Formate dependent nitrate and nitrite reduction to ammonia by

Citrobacter freundii and competition with denitrifying bacteria. Antonie van Leeuwenhoek

Journal of Microbiology, 56:311–321.

Rocha O, Ansari K, Doohan FM. 2005. Effects of trichothecene mycotoxins on eukaryotic

cells: A review. Food Additives and Contaminants, 22:369–378.

Roscoe V, Lombaert GA, Huzel V et al. 2008. Mycotoxins in breakfast cereals from the

Canadian retail market: a 3-year survey. Food Additives and Contaminants, 25:347–355.

201

Rotter BA, Prelusky DB, Pestka JJ. 1996. Toxicology of deoxynivalenol (vomitoxin). Journal

of Toxicology and Environment Health, 48:1–34.

Ryu JC, Ohtsubo K, Izurniyarna N, Nakamura K, Tanaka T, Yamarnura H, Ueno Y. 1988.

The acute and chronic toxicities of nivalenol in mice. Fundamental and Applied

Toxicology, 11:38–47.

Saier MH, Roseman S. 1972. Inducer exclusion and repression of enzyme synthesis in

mutants of Salmonella typhimurium defective in enzyme I of the phosphoenolpyruvate:

sugar phosphotransferase system. The Journal of Biological Chemistry, 247:972–975.

Saier MHJ, Tran CV, Barabote RD. 2006. TCDB: the transporter classification database for

membrane transport protein analyses and information. Nucleic Acids Research, 34:D181–

186.

Saitou N, Nei M. 1987. The neighbor-joining method: a new method for reconstructing

phylogenetic trees. Molecular Biology and Evolution, 4:406–425.

Savelieva O, Kotova I, Roelofsen W, Stams AJM, Netrusov A. 2004. Utilization of

aminoaromatic acids by a methanogenic enrichment culture and by a novel Citrobacter

freundii strain. Archives of Microbiology, 181:163–170.

Schaafsma AW, Tamburic-Ilinic L, Miller JD, Hooker DC. 2001. Agronomic considerations

for reducing deoxynivalenol in wheat grain. Canadian Journal of Plant Pathology, 23:279–

285.

Schaafsma AW. 2002. Economic changes imposed by mycotoxins in food grains: case study

of deoxynivalenol in winter wheat. Advances in Experimental Medicine and Biology,

504:271–276.

202

Scharf C, Riethdorf S, Ernst H, Engelmann S, Völker U, Hecker M. 1998. Theoredoxin is an

essential protein induced by heat shock and other stresses in Bacillus subtilis. Journal of

Bacteriology, 180:1869-1877.

Schatzmayr G, Zehner F, Taubel M, Schatzmayr D, Klimitsch A, Loibner AP, Binder EM.

2006. Microbiologicals for deactivating mycotoxins. Molecular Nutrition and Food

Research, 50:543-551.

Schollenberger M, Müller H-M, Rüfle R, Suchy S, Plank S, Drochner W. 2006. Natural

occurrence of 16 Fusarium toxins in grains and feedstuffs of plant origin from Germany.

Mycopathologia, 161:43–52.

Scientific Committee for Food. 2002. Opinion on Fusarium toxins. Part 6: Group evaluation

of T-2 toxin, HT-2 toxin, nivalenol and deoxynivalenol (available at: http://

www.europa.eu.int/comm/food/fs/sc/scf/out123_en.pdf).

Schulman S, Wang B, Li W, Rapoport TA. 2010. Vitamin K epoxide reductase prefers ER

membrane-anchored thioredoxin-like redox partners. Proceedings of the National

Academy of Sciences, 107:15027–15032

Scott PM. 1997. Multi-year monitoring of Canadian grains and grain-based foods for

trichothecenes and zearalenone. Food Additives and Contaminants, 14:333-339.

Seitz LM, Bechtel DB. 1985. Chemical, physical, and microscopal studies of scab-infected

hard red winter wheat. Journal of Agricultural and Food Chemistry, 33:373-7.

Sergent T, Parys M, Garsou S, Pussemier L, Schneider YJ, Larondelle Y. 2006.

Deoxynivalenol transport across human intestinal Caco-2 cells and its effects on cellular

metabolism at realistic intestinal concentrations. Toxicological Letters, 164:167–176.

203

Sharma J, Fulekar MH. 2009. Potential of Citrobacter freundii for bioaccumulation of

heavy metal–copper. Biology and Medicine, 1:7-14.

Sherif SO, Salama EE, Abdel-Wahhab MA. 2009. Mycotoxins and child health: the need for

health risk assessment. International Journal of Hygiene and Environmental Health,

212:347–368.

Shifrin VI, Anderson P. 1999. Trichothecene mycotoxins trigger a ribotoxic stress response

that activates c-Jun N-terminal kinase and p38 mitogen-activated protein kinase and

induces apoptosis. The Journal of Biological Chemistry, 274:13985–13992.

Shima J, Takase S, Takahashi Y, Iwai Y, Fujimoto H, Yamazaki M, Ochi K. 1997. Novel

detoxification of the trichothecene mycotoxin deoxynivalenol by a soil bacterium isolated

by enrichment culture. Applied and Environmental Microbiology, 63:3825–3830.

Singh J, Behal A, Singla N, Joshi A, Birbian N, et al. 2009. Metagenomics: Concept,

methodology, ecological inference and recent advances. Biotechnology Journal, 4:480–

494.

Smith TK, McMillan EG, Castillo JB. 1997. Effect of feeding blends of Fusarium mycotoxin-

contaminated grains containing deoxynivalenol and fusaric acid on growth and feed

consumption of immature swine. Journal of Animal Science, 75:2184–2191.

Smits WK, Dubois JYF, Bron S, van Dijl JM, Kuipers OP. 2005. Tricksy business:

transcriptome analysis reveals the involvement of thioredoxin in redox homeostasis,

oxidative stress, sulfur metabolism, and cellular differentiation in Bacillus subtilis. Journal

of Bacteriology, 187:3921–3930.

204

Sprando RL, Collins TFX, Black TN, Olejnik N, Rorie JI, Eppley RM, Ruggles DI. 2005.

Characterization of the effect of deoxynivalenol on selected male reproductive endpoints.

Food Chemistry and Toxicology, 43:623–635.

Starkey DE, Ward TJ, Aoki T, Gale LR, Kistler HC, Geiser DM, Suga H, Tóth B, Varga J,

O'Donnell K. 2007. Global molecular surveillance reveals novel Fusarium head blight

species and trichothecene toxin diversity. Fungal Genetics and Biology, 44:1191–1204.

Stevenson BS, Schmidt TM. 1997. Growth rate-dependent expression of RNA from plasmid-

borne rRNA operons in Escherichia coli. Journal of Bacteriology, 180:1970–1972.

Stothard P, Wishart DS. 2005. Circular genome visualization and exploration using

CGView. Bioinformatics, 21:537–539.

Suenaga H, Ohnuki T, Miyazaki K. 2007. Functional screening of a metagenomic library for

genes involved in microbial degradation of aromatic compounds. Environmental

Microbiology, 9:2289-97.

Surai P, Mezes M, Fotina TI, Denev SD. 2010. Mycotoxins in Human Diet: A Hidden

Danger. In: Modern Dietary Fat Intakes in Disease Promotion, Edited by F. De Meester et

al.. Nutrition and Health, DOI 10.1007/978-1-60327-571-2_18.

Sutton JC. 1982. Epidemiology of wheat head blight and maize ear rot caused by Fusarium

graminearum. Canadian Journal of Plant Pathology, 4:195.

Swaving J, De Bont JAM, Westphal A, De Kok A. 1996. A novel type pyridine nucleotide-

disulfide oxidoreductase is essential for NAD+- and NADPH-dependent degradation of

epoxyalkanes by Xanthobacter strain Py2. Journal of Bacteriology, 178:6644–6.

205

Swanson SP, Rood JHD, Behrens JC, Sanders PE. 1987. Preparation and characterization of

the deepoxy trichothecenes: deepoxy HT-2, deepoxy T-2 triol, deepoxy T-2 tetraol,

deepoxy 15-monoacetoxyscirpenol, and deepoxy scirpentriol. Applied and Environmental

Microbiology, 53:2821–2826.

Swanson SP, Helaszek C, Buck WB, Rood JHD, Haschek WM. 1988. The role of intestinal

microflora in the metabolism of trichothecene mycotoxins. Food Chemistry and

Toxicology, 26:823–829.

Tian F, Yu Y, Chen B, Li H, Yao Y-F, Guo X-K. 2009. Bacterial, archaeal and eukaryotic

diversity in Arctic sediment as revealed by 16S rRNA and 18S rRNA gene clone libraries

analysis. Polar Biology, 32:93-103.

Tatusov RL, Galperin MY, Natale DA, Koonin EV, et al. 2000. Artemis: sequence

visualization and annotation. Bioinformatics, 16:944-945.

Trail F, Haixin X, Loranger R, Gadoury D. 2002. Physiological and environmental aspects of

ascospore discharge in Gibberella zeae (anamorph Fusarium graminearum). Mycologia,

94:181-189.

Tsuge K, Ohata Y, Shoda M. 2001. Gene yerP, involved in surfactin self-resistance in

Bacillus subtilis. Antimicrobial Agents and Chemotherapy, 45:3566–3573.

Tucker 2001. The “yellow rain” controversy: Lesson for arms control compliance. The

Nonproliferation Review, 8:25-42.

Ueno Y. 1985. The toxicology of mycotoxins. Critical Reviews in Toxicology, 14:99-132.

Ueno Y, Nakayama K, Ishii K, Tashiro F, Minoda Y, Omori T, Komagata K. 1983.

Metabolism of T-2 toxin in Curtobacterium sp. strain 114-2. Applied and Environmental

Microbiology, 46:120–127.

206 van Elsas JD, Speksnijder AJ, van Overbeek LS. 2008. A procedure for the metagenomics

exploration of disease suppressive soils. Journal of Microbiological Methods, 75:515–522. van Ginkel CG, de Bont JAM. 1986. Isolation and characterization of alkene-utilizing

Xanthobacter spp. Archives of Microbiology, 145:403–407. van Lanen SG, Shen B. 2006. Microbial genomics for the improvement of natural product

discovery. Current Opinion in Microbiology, 9:252–260.

Varga J, Tóth B. 2005. Novel strategies to control mycotoxins in feeds: a review. Acta

Veterinaria Hungarica, 53:189-203.

Voigt CA, Schafer W, Salomon S. 2005. A secreted lipase of Fusarium graminearum is a

virulence factor required for infection of cereals. The Plant Journal, 42:364–375.

Vogelgsang S, Hecker A, Musa T, Dorn B, Forrer HR. 2011. On farm experiments over five

years in a grain maize/winter wheat rotation: Effect of maize residue treatments on

Fusarium graminearum infection and deoxynivalenol contamination in wheat. In: Proc.

7th Canadian Workshop on Fusarium Head Blight, edited by J Gilbert, A Tekauz, N

Sweetland and K Slusarenko. Winnipeg, Manitoba, Canada, p 41.

Völkl A, Vogler B, Schollenberger M, Karlovsky P. 2004. Microbial detoxification of

mycotoxin deoxynivalenol. Journal of Basic Microbiology, 44:147–156.

Wagacha JM, Muthomi JW. 2007. Fusarium culmorum: Infection process, mechanisms of

mycotoxin production and their role in pathogenesis in wheat. Crop Protection, 26:877-

885.

Wakulinski W. 1989. Phytotoxicity of the secondary metabolites of fungi causing wheat head

fusariosis (head blight). Acta Physiologiae Plantarum, 11:301–306.

207

Wang JT, Chang SC, Chen YC, Luh KT. 2000. Comparison of antimicrobial susceptibility of

Citrobacter freundii isolates in two different time periods. The Journal of Microbiology,

Immunology and Infection, 33:258-62.

Wannemacher RJ, Stanley L, Wiener MD. 2000. Trichothecenes mycotoxins. In. Medical

aspects of chemical and biological warfare, edited by Zajtchuk R. Washington, DC:

Department of the Army, pp 656-676.

Westphal AH, Swaving J, Jacobs L, de Kok A. 1998. Purification and characterization of a

flavoprotein involved in the degradation of epoxyalkanes by Xanthobacter Py2. European

Journal of Biochemistry, 257:160–168.

Whalen JG, Mully TW, Enlgish JC. 2007. Spontaneous Citrobacter freundii infection in an

immunocompetent patient. Archives of Dermatology, 143:124-5.

White 2007. The physiology and biochemistry of prokaryotes. Third edition. Oxford

University Press, New York, Oxford, p 628.

Wilcke WF, Morey RV. 1995. Natural-air corn drying in the upper Midwest University of

Minnesota. Minnesota Extension Service Publication, BU-6577-E, pp 20.

Windels CE. 2000. Economic and social impacts of Fusarium headblight: changing farms and

rural communities in the Northern Great Plains. Phytopathology, 90:17–21.

Winzeler EA, Shoemaker DD, Astromoff A, Liang H, Anderson K, Andre B, Bangham R,

Benito R, Boeke JD, Bussey H, et al. 1999. Functional characterization of the S.

cerevisiae genome by gene deletion and parallel analysis. Science, 285:901–906.

Yang GH, Jarvis BB, Chung YJ, Pestka JJ. 2000. Apoptosis induction by the satratoxins and

other trichothecene mycotoxins: relationship to ERK, p38 MAPK, and SAPK/JNK

activation. Toxicology and Applied Pharmacology, 164:149–160.

208

Yang Q, Yin GM, Guo YL, Zhang DF, Chen SJ, Xu ML. 2010. A major QTL for resistance

to Gibberella stalk rot in maize. Theoretical and Applied Genetics, 121:673–687.

Yazar S, Omurtag GZ. 2008. Fumonisins, trichothecenes and zearalenone in cereals.

International Journal of Molecular Science, 9:2062–2090.

Yoshizawa T, Morooka N. 1973. Deoxynivalenol and its monoacetate: new mycotoxins from

Fusaruium roseum and moldy barley. Agricultural and Biological Chemistry, 37:2933–

2934.

Yoshizawa T, Takeda H, Ohi T. 1983. Structure of a novel metabolite from deoxynivalenol, a

trichothecene mycotoxin, in animals. Agricultural and Biological Chemistry, 47:2133–

2135.

Yoshizawa T. 1993. Red-mold diseases and natural occurrence in Japan. In: Trichothecenes-

Chemistry, Biology and Toxicological Aspects, edited by Y. Ueno, Elsevier, Amsterdam,

Netherlands, p 195-209.

Yoshizawa T, Sakamoto T, Okamoto K. 1984. In vitro formation of 3’-hydroxy T-2 and 3’-

hydroxy HT-2 toxins from T-2 by liver homogenates from mice and monkeys. Applied

and Environmental Microbiology, 47:130-134.

Young JC, Subryan LM, Potts D, McLaren ME, Gobran FH. 1986. Reduction in levels of

deoxynivalenol in contaminated wheat by chemical and physical treatment. Journal of

Agricultural and Food Chemistry, 34:461–465.

Young JC, Trenholm HL, Friend DW, Prelusky DB. 1987. Detoxification of deoxynivalenol

with sodium bisulfite and evaluation of the effects when pure mycotoxin or contaminated

corn was treated and given to pigs. Journal of Agricultural and Food Chemistry, 35:259–

261.

209

Young JC, Zhu H, Zhou T. 2006. Degradation of trichothecene mycotoxins by aqueous

ozone. Food Chemistry and Toxicology, 44:417–424.

Young JC, Zhou T, Yu H, Zhu H, Gong J. 2007. Degradation of trichothecene mycotoxins by

chicken intestinal microbes. Food Chemistry and Toxicology, 45:136-143.

Yu Z, Morrison M. 2004. Comparisons of different hypervariable regions of rrs genes for use

in fingerprinting of microbial communities by PCR-denaturing gradient gel

electrophoresis. Applied and Environmental Microbiology, 70:4800-4806.

Yu H, Zhou T, Gong J, Young C, Su X, Li X, Zhu H, Tsao R, Yang R. 2010. Isolation of

deoxynivalenol-transforming bacteria from the chicken intestines using the approach of

PCR-DGGE guided microbial selection. BMC Microbiology, 10:182.

Zeibdawi A, Steele M, Savoie A. 2011. An overview of CFIA's mycotoxin monitoring

program in livestock feeds. In: Proc. 7th Canadian Workshop on Fusarium Head Blight,

edited by J Gilbert, A Tekauz, N Sweetland and K Slusarenko. Winnipeg, Manitoba,

Canada, p 25.

Zhang J-B, Li H-P, Dang F-J, Qu B, Xu Y-B, Zhao C-S, Lia Y-C. 2007. Determination of the

trichothecene mycotoxins chemotypes and associated geographical distribution and

phylogenetic species of the Fusarium graminearum clade from China. Mycological

Research, 111:967–975.

Zhao JS, Manno D, Thiboutot S, Ampleman G, Hawari J. 2007. Shewanella canadensis sp.

nov. and Shewanella atlantica sp. nov., manganese dioxide- and hexahydro-1,3,5-trinitro-

1,3,5-triazine-reducing, psychrophilic marine bacteria. International Journal of Systemic

Evolution and Microbiology, 57:2155–2162.

210

Zhao X-Q. 2011. Genome-based studies of marine microorganisms to maximize the diversity

of natural products discovery for medical treatments. Evidence-Based Complementary

and Alternative Medicine. Article ID 384572, doi:10.1155/2011/384572.

ZhaoY, Park RD, Muzzarelli RA. 2010. Chitin deacetylases: properties and applications.

Marine Drugs, 8:24–46.

Zhou HR, Lau AS, Pestka JJ. 2003a. Role of double-stranded RNA-activated protein kinase R

(PKR) in deoxynivalenol-induced ribotoxic stress response. Toxicological Science,

74:335–344.

Zhou HR, Islam Z, Pestka JJ. 2003b. Rapid, sequential activation of mitogen-activated protein

kinases and transcription factors precedes proinflammatory cytokine mRNA expression in

spleen of mice exposed to the trichothecene vomitoxin. Toxicological Sciences, 72:130–

142.

Zhou T, He J, Gong J. 2008. Microbial transformation of trichothecene mycotoxins. World

Mycotoxin Journal, 1:23-30.

Zhou G, Zhang H, He Y, He L. 2010. Identification of a chitin deacetylase producing bacteria

isolated from soil and its fermentation optimization. African Journal of Microbiology

Research, 4:2597-2603.

211