Evaluation of Flea Beetle (Phyllotreta Spp.) Resistance in Spring and Winter-Type Canola (Brassica Napus )

Evaluation of Flea Beetle (Phyllotreta spp.) Resistance in Spring and Winter-Type Canola (Brassica napus)

Julian R. Heath

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Doctor of Philosophy in Plant Agriculture

Guelph, Ontario, Canada

ABSTRACT

EVALUATION OF FLEA BEETLE (PHYLLOTRETA SPP.) RESISTANCE IN SPRING AND WINTER-TYPE CANOLA (BRASSICA NAPUS L.)

Julian R. Heath Co-Advisors: University of Guelph, 2017 Drs. Laima Kott and Istvan Rajcan

This thesis is an investigation into the understanding of flea beetle (Phyllotreta spp.) resistance in spring-type and winter-type canola quality Brassica napus L.. Canola is one of the world’s most widely grown oilseed crops and an economically important crop in Western Canada. Genetic resistance to this common pest would add to the available tactics for integrated pest management of flea beetles in canola. The purpose of this research was to better understand the interactions of the flea beetle with canola at key feeding times in the life cycle of the flea beetle and identify genetic components related to flea beetle herbivory on canola seedlings. The objectives were to: 1) investigate seasonal effects of flea beetle herbivory on both spring-type and winter-type canola; and 2) identify quantitative trait loci (QTL) for flea beetle herbivory in two winter-type doubled haploid (DH) populations using simple sequence repeat (SSR) markers and single nucleotide polymorphic (SNP) markers. Year and seasonal effects were noted but overall trends amongst entries were similar concluding that flea beetle feeding patterns did not change throughout its life cycle. Spring-type and winter-type germplasm reacted similarly under flea beetle feeding. As such, there does not appear to be any novel resistance mechanisms that evolved as a result of divergent growth habit types in canola. Seven QTL were identified over the

two DH populations studied. Linkage group (LG) N13 had multiple QTL identified. The remaining

QTL were located on LG N04, N06, N15 and N17. It is unknown as to what mechanisms these QTL are associated with. The results of this thesis provide insight into flea beetle-canola plant interactions and identify some genetic areas of interest related to flea beetle herbivory in canola.

ACKNOWLEDGEMENTS

I would like to express my gratitude to all those who have supported me during this nine-year journey. Firstly, thank you to my co-advisors Dr. Istvan Rajcan and Laima Kott for your patience, support and encouragement throughout this process. This has been a challenging undertaking and without the support and strong encouraging words when needed, this may not have been successful. It has been a pleasure learning from you.

I would like to thank my advisory committee of Drs. Ron Fletcher (deceased), Rebecca Hallett and Jay Patel. Sadly, Dr. Fletcher passed prior to seeing this completed but his insight with the material helped steer the project into what it became. Dr. Hallett provided a varying point of view coming from the insect world and provided fantastic feedback and direction throughout the process. Finally, thank you to Dr. Patel for whom I have had the pleasure of working with over the past 16 years. It was your guidance and support that got me started on this journey and helped me see it through to the end. Your guidance over the years for this research and beyond has been greatly appreciated.

Thank you to the members of the examining committee for taking the time to review the manuscript and provide constructive feedback; Dr. Clarence Swanton (chair), Dr. Sally Vail (external examiner), Dr. Ali Navabi, Dr. Jay Patel and Dr. Istvan Rajcan.

I wish to thank DuPont Pioneer for supporting me in this process. It has been a long journey and there have been many people that have assisted me over the years at various stages. A huge thank you to everyone that played a role – big or small, as without you, I would not have been able to do this.

Special thanks to all my friends and family for understanding the time and commitment to work and school.

Finally, thank you to my wife Chrissie and daughter Ellie for the love and support and tolerating me during the highs and lows of this journey.

LIST OF ABBREVIATIONS

AAFC Agriculture and Agri-Food Canada ANOVA Analysis of variance B Brassica Bt Bacillus thuringiensis CRISPR Cluster regularly interspaced short palindromic repeats C.V. Coefficient of variation DH Doubled haploid EPN Entomopathogenic nematodes F-Value F test statistics FBSC Flea beetle score GWAS Genome wide association study IST Insecticidal seed treatment LG Linkage group(s) LOD Logarithm/Likelihood of odds QTL Quantitative trait loci (locus) RCBD Randomized complete block design R² Coefficient of determination SD Standard deviation SE Standard error SNP Single nucleotide polymorphism SSR Simple sequence repeats TALEN Transcription activator-like effector nuclease ZFN Zinc finger nuclease

TABLE OF CONTENTS Abstract ...... ii

Acknowledgements ...... iv

List of Tables ...... ix

List of Figures ...... xii

List of Figures (Appendix) ...... xvii

General Introduction ...... 1

Chapter 1 - Literature Review ...... 6

1.1. The life history and biology of Phyllotreta cruciferae and P. striolata ...... 7 1.1.1. Description and life history of Phyllotreta species – P. cruciferae and P. striolata ...... 7

1.1.2. Flea beetle management ...... 9

1.2. Host plant selection and behavior of Phyllotreta spp. on Brassicaceae ...... 14 1.2.1. Host plant selection ...... 15

1.2.2. Host plant acceptance ...... 19

1.2.3. Chemical cues ...... 23

1.2.4. Physical cues ...... 27

1.2.5. Cost of resistance ...... 30

1.3. Flea beetle (Phyllotreta spp.) responses to host plant secondary metabolites and morphology 31 1.3.1. Glucosinolates ...... 31

1.3.2. Cyanogenic glycosides ...... 37

1.3.3. Alkaloids ...... 38

1.3.4. Terpenoids ...... 38

1.3.5. Phenolics ...... 40

1.3.6. Other plant hormones, amino acids, proteinase inhibitors, etc...... 41

1.3.7. Herbivore mechanisms of handling secondary metabolites ...... 43

1.3.8. Herbivore mechanisms of handling morphological defense ...... 44

1.4. Identifying the most promising plant chemical or physical feature to investigate in the development of flea beetle resistance in canola...... 45 1.4.1. Current works of interest ...... 46

1.4.2. Transgenic or molecular-based resistance ...... 47

1.5. Summary ...... 49 1.6. Hypothesis...... 51 Chapter 2 - Evaluation of Winter-type Brassica napus Germplasm for Genetic Diversity in Response to Flea Beetle Herbivory in Typical and Atypical Planting

Windows ...... 55

2.1. Abstract ...... 55 2.2. Introduction ...... 56 2.3. Materials and Methods ...... 59 2.3.1. Plant material ...... 59

2.3.2. Test sites and experimental designs ...... 60

2.3.3. Phenotypic data - scoring of plants for flea beetle herbivory ...... 61

2.3.4. Statistical analysis ...... 62

2.4. Results and Discussion ...... 63 2.4.1. Results ...... 63

2.4.2. Discussion...... 65

2.5. Conclusion ...... 68 Chapter 3 - Evaluation of Spring-type Brassica napus Germplasm for Genetic Diversity in Response to Flea Beetle Herbivory in Typical and Atypical Planting

Windows ...... 79

3.1. Abstract ...... 79

vii

3.2 Introduction ...... 80 3.3. Materials and Methods ...... 82 3.3.1. Plant material ...... 82

3.3.2. Test sites and experimental designs ...... 83

3.3.3. Phenotypic data - scoring of plants for flea beetle herbivory ...... 84

3.3.4. Statistical analysis ...... 85

3.4 Results and Discussion ...... 85 3.4.1. Results ...... 85

3.4.2. Discussion...... 87

3.5 Conclusion ...... 89 Chapter 4 – Genetic Mapping of QTL for Resistance to Flea Beetle (Phyllotreta spp.) in Two Winter-type B. napus Doubled Haploid Populations...... 100

4.1 Abstract ...... 100 4.2 Introduction ...... 101 4.3. Materials and Methods ...... 105 4.3.1 Plant material and experimental design ...... 105

4.3.2. Molecular markers and genotyping ...... 109

4.3.3. Linkage map construction ...... 111

4.4 Results and Discussion ...... 112 4.4.1. Phenotyping ...... 112

4.4.2. Construction of genetic maps ...... 116

4.5 Conclusion ...... 122

Chapter 5 – General Discussion and Conclusions ...... 140

References ...... 145

Appendix ...... 181

viii

LIST OF TABLES

Table 2.1. Summary of experiments, dates, locations, experimental design, entry numbers and notes collected for all winter-type canola B. napus trials examined for flea beetle injury from 2009 to 2011…..…………………………………………………………………………………………….…………69

Table 2.2. Mean scores for flea beetle herbivory 14 days after planting on 15 winter-type canola lines evaluated from 2009 to 2011 in Alloa (ALL) ON, Belfountain (BEL) ON and Edmonton (EDM) AB. SAS PROC GLM statistics Type III ANOVA summary on the lower part of the table by experiment. Flea beetle score (FBSC): Scale of 0-9 where 0 = plant dead and 9= less than 10% flea damage on cotyledons.………………………………………...... 70

Table 3.1. Summary of experiments, dates, locations, experimental design, entry numbers and notes collected for all spring-type B. napus trials examined for flea beetle injury from 2009 to 2011………..……………………………………………………………………………………………………..…………..91

Table 3.2. Mean scores for flea beetle herbivory 14 days after planting on 15 spring-type canola lines evaluated from 2009 to 2011 in Acton (ACT) ON, Alloa (ALL) ON, Belfountain (BEL) ON and Edmonton (EDM) AB. SAS PROC GLM statistics Type III ANOVA summary on the lower part of the table by experiment. Flea beetle score (FBSC): Scale of 0-9 where 0 = plant dead and 9= less than 10% flea damage on cotyledons…………………..……….…………..92

Table 3.3. Mean scores for flea beetle herbivory (FBSC) 21 days after planting on 20 spring-type Brassicaceae varieties evaluated from 2009 to 2011 in Acton (ACT), Alloa (ALL), Belfountain (BEL), ON and Edmonton (EDM), AB. SAS PROC GLM statistics Type III ANOVA summary on the lower part of the table by experiment. FBSC: Scale of 1-9 where 0 = plant dead and 9= no damage.………………………………………………………………………………………..….…..94

Table 4.1. Threshold values (LOD scores) for all traits in each population. Estimated with permutation analysis after 1000 iterations and significance level of 0.05.……………..……..125

Table 4.2. Summary of statistics for DH Population J10-02 flea beetle feeding in arenas……….125

Table 4.3. Summary of statistics for Population J10-11 flea beetle feeding in arenas……..……..125

Table 4.4. Summary of statistics for Populations J10-02 and J10-11 flea beetle feeding in 2015 field trial near Saskatoon……………………………………………………………………………………………….126

Table 4.5. Marker distribution by linkage group (LG) and marker type (SSR or SNP) for flea beetle damage for indoor and outdoor screening on two winter-type Brassica napus DH populations…………………………………………………………………………………………………………………..127

Table 4.6. Major and minor QTL summary in the two DH mapping populations for flea beetle herbivory in winter-type canola……………………………………………………………………………………128

LIST OF FIGURES

Figure 1.1. Common flea beetles found on canola in Western Canada; a) adult crucifer flea beetle, Phyllotreta cruciferae Goeze. and b) adult striped flea beetle, Phyllotreta striolata Fabricius. (diagram adapted from Knodel and Olson (2002))………………………………………………………….52

Figure 1.2. Life cycle of the crucifer flea beetle. (diagram adapted from Knodel and Olson (2002)). …………………………………………………………………………………………………………………………………..…..52

Figure 1.3. The wounding response. Generalized overview of the plant wounding response and signaling molecules that modulate it. The pathways necessary for both local and systemic induction of insecticidal proteins are shown. Abbreviations: ABA, abscisic acid; SA, salicyclic acid. (diagram adapted from Ferry et al., 2004)………….…………………………………………………..53

Figure 1.4. Summary of the biosynthetic pathway and stress-induced metabolite production. The basic metabolic pathway is drawn in the circle and the stimuli and the compounds increased (+) or decreased (−) as a result of these are listed outside (diagram adapted from Jahangir et al., 2009)……………………………………………………………………………………………………………………..54

Figure 2.1. Canola cotyledons with levels of flea beetle damage varying from less than 10% (panel a) to 100% (panel j) damage to both cotyledons (adapted from Soroka and Underwood 2011)……………………………………………………………………………………………………………………………….74

Figure 2.2. Canola seedlings with varying levels of flea beetle damage: Panels a to f represent first leaf stage; Panels g to i represent two leaf stage (adapted from Soroka and Underwood 2011)……………………………………………………………………………………………………………………….……..75

Figure 2.3. Graph showing average flea beetle feeding scores on winter-type canola lines at 14d post planting by entry for each year. Error bars indicate standard error of the mean for each

xii

entry over trials in that year (2009=3; 2010=7; 2011=1). Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. ANOVA indicated no interaction between year and the entry (=0.05) but do show insect pressure differences by year (=0.05).…………………………………………………………………………………………………………………………76

Figure 2.4. Graph showing flea beetle feeding scores on winter-type canola lines at 14d post planting by entry for spring and fall planting in 2009 (Belfountain) and 2010 (Alloa and Belfountain). Error bars indicate standard error of the mean for each entry over three trials per season. Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. ANCOVA indicated no interaction between season and entry (=0.05) but showed significant differences between seasons (=0.05)………….……………………………………………….77

Figure 2.5. Graph showing average combined flea beetle feeding scores on winter-type canola lines from eleven field trials (2009 to 2011). Flea beetle injury score (FBSC) as a scale of 0 = dead or no plant to 9 = less than 10% damage. Error bars indicate standard error of the mean for each entry. Family grouping indicate within and among family variation for FBSC. Resistant checks (46A65-treated and Ace-Sinapis alba) have greater resistance than other varieties evaluated; S. alba was significantly more resistant than family 3 ( =0.05) and Cutlass (=0.1)…………………….………………………………………………………………………………………….78

Figure 3.1. Graph showing average flea beetle feeding scores on spring-type canola lines at 14d post planting by entry for each year. Error bars indicate standard error of the mean for each entry over trials in that year (2009=4; 2010=7; 2011=1). Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. ANOVA indicated no interaction between year and the entry (=0.05) but do show insect pressure differences by year (=0.05)………………………………………………………………………………………………………………………….97

Figure 3.2. Graph showing seasonal averaged feeding scores at 14d post planting by entry. ANCOVA indicated no interaction between season and entry and showed no differences

xiii

between seasons. Flea beetle injury score (FBSC) is a scale of 0 to 9 where 0=full plant injury or death and 9 is less than 10% damage on cotyledon…………………………………………………….98

Figure 3.3. Graph showing average combined flea beetle feeding scores on spring-type canola lines from twelve field trials (2009 to 2011). Flea beetle injury score (FBSC) as a scale of 0 = dead or no plant to 9 = less than 10% damage. Error bars indicate standard error of the mean for each entry. Family grouping indicate within and among family variation for FBSC. Resistant checks (46A65-treated and Ace-Sinapis alba) have greater resistance than other varieties evaluated; no family or checks were significantly greater than any other check or family comparison (=0.05)……………………………………………………………………………………………99

Figure 4.1. Indoor experimental arena used for flea beetle scoring. Twenty-five entries were screened at a time with 23 experimental entries plus two controls, Ace and 46A65. Fifty flea beetles were added to each arena for a 24-hour period to cotyledons planted 7 days prior……………………………………………………………………………………………………………………………..129

Figure 4.2. Visual representation of feeding bite versus test bite ……………..…………………………130 Figure 4.3a. Crepe paper seed tape where glue stick was used to glue 5 canola seeds 2.5cm apart to the seed tape with 20cm between plots. 4.3b. Planting of crepe paper seed tape in 2015. 4.3c. Trial site after scoring once plants were at the 2-4 leaf stage (plant stage in image above about 3 weeks after scoring)……………………………………………………………….130-131

Figure 4.4. Distribution of flea beetle feeding bites (A), test bites (B) and total bites (C) in arenas for Population J10-02 with a normal distribution line graph overlaid.……………………………132

Figure 4.5. Distribution of flea beetle bites (A), test bites (B) and total bites (C) in arenas for Population J10-11 with normal distribution line graph overlaid ……………………………………132

xiv

Figure 4.6. Distribution of number of bites in DH Population J10-02 including parental lines 1147- 03 and 1147-01 and B. napus check 46A65 evaluated in indoor arena feeding study…….133

Figure 4.7. Distribution of number of bites in DH population J10-11 Including B. napus check 46A65 evaluated in indoor arena feeding study……………………………………………………….……133

Figure 4.8. Populations J10-02 and Population J10-11 BLUEs distribution from 2015 field flea beetle scores. From left to right, red bar is Ace (resistant check), two yellow bars are resistant parents (1147-13 and 1147-03, respectively), next red bar is 46A65 with final yellow being susceptible parent (1147-01). Progeny from both populations are interspersed along the entire length.………………………………………………………………………………………………….134

Figure 4.9. Graph showing the LOD score profile for DH Population J10-02 with the traits BLUEs*, average injury score*, feeding bites**, test bites** and total bites** in field and indoor screening. (*=field based score, **=indoor based score). Lower graph shows the additive effects of the traits at each allele………….……………………………………………………………………….135

Figure 4.10. Graph showing the LOD score profile for DH Population J10-11 with the traits BLUEs*, average injury score*, feeding bites**, test bites** and total bites** in field and indoor screening. (*=field based score, **=indoor based score). Lower graph shows the additive effects of the traits at each allele……….………………………………………………………………………….136

Figure 4.11. Genetic linkage map and the locations of QTL for flea beetle damage in winter-type Brassica napus DH Population J10-02 using SSR and SNP markers. There are 17 linkage groups identified using JoinMap (Van Ooijen, 2006)and DuPont Pioneer’s genetic map and are represented by vertical bars. Marker names are listed to the right of the linkage groups with the position in centimorgans (cM) listed to the left side. The four identified QTL (two overlapping on N04, one on N06 and one on N13) associated with flea beetle injury were indicated by red bars within the linkage group.……………………………………………………………..137

Figure 4.12.a and b. Genetic linkage map and the locations of QTL for flea beetle damage in winter-type Brassica napus DH Population J10-11 using SSR and SNP markers. There are 19 linkage groups identified using JoinMap (Van Ooijen, 2006) and DuPont Pioneer’s genetic map and are represented by vertical bars. Marker names are listed to the right of the linkage groups with the position in centimorgans (cM) listed to the left side. The four identified QTL associated with flea beetle injury were indicated by red bars within the linkage groups N13 N15 (thicker black line at ~137cM) and N17…………………………….138-139

xvi

LIST OF FIGURES (APPENDIX)

Figure A4.1. – Linkage groups from Population J10-02 showing the position of QTL as highlighted in red. a.) at approximately 16cM; b.) at approximately 31cM; c.) at approximately 189cM……………………………………………………………………...... 181

Figure A4.2. – Linkage groups from Population J10-11 showing the position of QTL as highlighted in red. a.) at approximately 35cM and 118cM; b.) at approximately 75cM..182

xvii

GENERAL INTRODUCTION

Canada’s canola industry is worth more than $26.7 billion annually (Canola Council of Canada,

2016). Canola, primarily Brassica napus L., is the third most important oilseed crop worldwide after soybean and cotton for oil meal and vegetable oil after soybean and oil palm (FAOSTAT,

2017). Within Canada, 18.4 million tonnes of canola seed was produced in 2016 on 8.3 million planted hectares (20.4 million acres)(Canola Council of Canada, 2016).

Flea beetles are one of the first pests to attack newly emerged canola seedlings. Flea beetles are found throughout the Northern Great Plains of North America (Burgess, 1977a). There are eight flea beetle species known to attack Brassicas such as canola/oilseed rape (B. napus), mustard

(Brassica juncea) and rapeseed (non-canola quality Brassica napus) (Burgess, 1977a). Of these eight, two are predominant: Phyllotreta cruciferae Goeze, also known as the crucifer flea beetle, and P. striolata Fabricius, also known as the striped flea beetle (Burgess, 1977a). Flea beetles have been estimated to cause about a 10% reduction in yield of open-pollinated varieties (Lamb and Turnock, 1982). Currently, public institutions (Agriculture and Agri-Food Canada (AAFC)-

Lethbridge, Beaverlodge, AB and Saskatoon, SK and Manitoba Agriculture, Carman MB) are validating this on current hybrid varieties (Wist, pers. comm.). At current canola seed pricing

(Winnipeg Exchange, February futures 2017) of approximately CA$500/tonne, flea beetles may be causing $900 million dollars in yield loss alone in Western Canada although most reports use

Knodel and Olson's (2002) figures of $300 million dollars. The first 20 days after emergence is

1 the most crucial period for canola development, as flea beetle feeding has been shown to reduce yield potential even at low pressure during that time (Bracken and Bucher, 1986). Damage occurs when adults feed on cotyledons and stems of seedlings, resulting in reduced photosynthetic capability, wilting, or host plant mortality (Westdal and Romanow, 1972). This feeding continues as the plant grows but is less detrimental as the plant is able to compensate for damaged tissues

(Gavloski and Lamb, 2000). During later growth stages, the newly emerged adult flea beetles feed on the pods and peduncles which may result in uneven maturity, premature shattering, shrivelled seed and increased seed chlorophyll (green seed) but the economic impact is minimal relative to the feeding on early plant growth stages (Lamb, 1984; Knodel and Olson, 2002).

The most common control method for flea beetles is the use of insecticidal seed treatment (IST) during planting, where 99.5% of acres planted use IST (Sekulic and Remple, 2016). Over the years, efficacy of seed treatments has improved and the most active ingredients are now effective against flea beetles up to 40 days post-emergence (Sekulic and Remple, 2016). Flea beetle feeding is required for the seed treatment to be effective as the insecticide is systemic within the plant and must be ingested by the herbivore to work. Plants do eventually outgrow the damage by the flea beetle, however, under warm dry conditions, flea beetle activity may exceed the plant growth and the rate at which the active ingredient is translocated by the plant to adequately inhibit flea beetle activity (Sekulic and Remple, 2016). A foliar insecticide application may be warranted when the economic threshold is reached; when an average of 25% of the surface area of cotyledons and the first true leaves has been injured (Knodel and Olson, 2002). With current

2 social-political concerns over pesticide use, seed treatments, although preferred over foliar applications, are facing increasing scrutiny or outright bans in certain jurisdictions.

A number of studies have been conducted with the goal of improving agronomic practices to reduce flea beetle damage. Intercropping studies and trap crops have shown reductions in flea beetle feedings and subsequent populations (Tahvanainen and Root, 1972; Altieri and Schmidt,

1986; Garcia and Altieri, 1992; Bohinc and Trdan, 2013; Metspalu et al., 2014). Cromartie (1975) and Kareiva (1983) both found that as plot size increased, the frequency of the two species of flea beetles (P. cruciferae and P. striolata) also increased, which is in agreement with Knodel and

Olson (2002) who reported that crop rotations did not provide effective management of flea beetles due to their strong flying ability and wide distribution of flea beetles. Dosdall and

Stevenson (2005) determined that flea beetle pressure can be reduced by planting canola in the fall instead of in the spring, with increased seeding rates and using Brassica napus varieties rather than B. rapa. Knodel and Olson (2002) recommend early planting, i.e. from April to mid-May, to reduce damage due to flea beetles. Other effective strategies for cultural control of flea beetles include using zero-tillage rather than conventional tillage (Cromartie, 1975; Dosdall et al., 1999), using large seeds instead of small seeds (Bodnaryk and Lamb, 1991a; Elliott et al., 2008) and increasing row spacing from 10cm to 30cm (Dosdall et al., 1999). Although agronomic practices can be altered to reduce flea beetle pressure in modern farm practices, these practices alone are not sufficient to prevent economic loss (Dosdall and Stevenson, 2005).

Over the past few decades, much research has been done to understand the relationship between members of the Brassicaceae and Phyllotreta spp. pests, but little gain has been made in developing genetic resistance in Brassicaceae. Interactions between flea beetles and their host plants have been studied by many investigators and the general consensus on flea beetle resistance in Brassicas appears to be multi-genic resistance with numerous small effects from different mechanisms. Studies by Bodnaryk and Lamb (1991b), Lamb et al. (1993) and others

(Brandt and Lamb, 1993) have showed that genetic variation is present in host plants based on the feeding responses of flea beetles to different hosts. Gavloski et al. (2000), Soroka and

Grenkow (2013) and Metspalu et al. (2014) reported varying degrees of susceptibility to flea beetle feeding within the Brassicaceae.

Biotechnology strategies have been investigated. Gruber et al. (2006) and Soroka et al. (2011) have identified/developed transgenic enhanced-trichome resistance to defend against flea beetle feeding. Transgenic chemical resistance within plants includes Bt (Bacillus thuringiensis- derived natural insecticide) (Gatehouse, 2008) and secondary metabolite modifications such as cyanogenic glycosides (Tattersall et al., 2001; Kristensen et al., 2005). New biotechnologies such as cluster regularly interspaced short palindromic repeats (CRISPR/CRISPR-associated (Cas)), transcription activator-like effector nuclease (TALENs) and zinc finger nuclease (ZFNs) technology may provide other techniques to enhance insect resistance in plants (Lusser et al., 2012; Chen and Lin, 2013; Tamiru et al., 2015)

In this thesis, Brassica napus spring-type and winter-type breeding germplasm in response to damage from flea beetle herbivory was evaluated. Two doubled haploid winter-type populations were created and evaluated for flea beetle herbivory damage. These populations were used to identify molecular markers that tag QTL associated with flea beetle herbivory. This study provides insight into the flea beetle feeding habits, the potential to move resistance genes between winter and spring-type Brassicas, and confirms previous reports of genetic variation within the Brassicas.

Results from this thesis will provide insight into how flea beetle feeding behaviour along with possibly identifying potential pathways and mechanisms associated with flea beetle resistance.

CHAPTER 1 - LITERATURE REVIEW

“Brassicas: Oil-, food- and fodder-bearing crops with small seeds; that can grow as fast as Wisconsin rapid cycling brassicas; can grow as big as ornamental plants; can be seen as vast green fields of vegetable crops or as oilseed crops with fields of bright yellow flowers; lead to the production of economically important agricultural products; used as food for humans and animals; are important as valuable renewable bioenergy resources; are huge reservoirs of plant innate defences; show multiple defence responses in response to stresses; possess anticancer properties; hold ample potential for pest management” (Ishita Ahuja from (Ahuja et al., 2010).

Canada’s canola industry is worth more than $26.7 billion annually (Canola Council of Canada,

2016). Canola, primarily Brassica napus L., is the third most important oilseed crop worldwide after soybean and cotton for oil meal and third for vegetable oil after soybean and oil palm

(FAOSTAT, 2017). Within Canada, canola competes with wheat as the most valuable field crop, resulting in 18.4 million tonnes produced in 2016 on 8.3 million planted hectares (20.4 million acres)(Canola Council of Canada, 2016). The canola industry has set a goal of at least 8.1 million hectares (20 million acres) of production, with average yields of 21 bushels per hectare (52 bushels per acre) for production of 26 million tonnes annually (Canola Council of Canada, 2016).

Seed yield gains can be achieved in various ways, whether through genetic or agronomic improvements. Pest management, particularly of flea beetles (Phyllotreta spp.) (Coleoptera:

Chrysomelidae: Alticinae), is one area that could help to meet those targets. Flea beetles have been estimated to cause about a 10% reduction in yield (Lamb and Turnock, 1982). Over the past few decades, much research has been undertaken to understand the relationship between members of the Brassicaceae and Phyllotreta spp. pests but little gain has been made in developing host plant genetic resistance, as defined by Painter (1951), in Brassicaceae.

1.1. THE LIFE HISTORY AND BIOLOGY OF PHYLLOTRETA CRUCIFERAE AND P. STRIOLATA

1.1.1. Description and life history of Phyllotreta species – P. cruciferae and P. striolata

Flea beetles are found throughout the Northern Great Plains of North America. There are eight predominant flea beetle species known to attack canola, mustard and rapeseed. Of these eight, two are predominant, Phyllotreta cruciferae Goeze also known as the crucifer flea beetle and P. striolata Fabricius., also known as the striped flea beetle (Burgess, 1977a). These two species are the focus of this thesis. Flea beetles are thought to have been introduced to North America from Eurasia with the crucifer flea beetle first reported in the United States in 1923

(Bonnemaison, 1965) and in Aggasiz, BC in the early 1920’s (Burgess, 1977a) while the striped flea beetle is thought to have been introduced into the United States in the 1700s (Bain and

LeSage, 1998).

1.1.1.1. Description

Flea beetles belong to the Coleoptera: Chrysomelidae family and are named for their ability to jump quickly when disturbed. Phyllotreta cruciferae adults are 2 to 3mm, dorsally flat, elongate oval, black with a bright blue lustre and have enlarged hind femoras (thighs) (Figure 1.1a.). The eleven antennae segments are similar in both sexes with the fifth segment not much different from the sixth (Burgess, 1977a). In spring, one to four eggs, about 0.4mm long by 0.2 mm wide, oval and light yellow, are deposited near the bases of host plants. Mature larvae are approximately 3mm, white to very light brown with a copper-brown head and anal plate and are slender with small legs. Larvae feed on roots and root hairs and pupate in soil, emerging in late summer. Phyllotreta striolata adults are 2 to 2.5mm, dorsally flat, elongate oval, black and

7 have enlarged hind femoras (Figure 1.1b.). Each elytron, which is a modified hardened forewing, has a distinctive pale yellow stripe that is wavy along its outside margin and curves towards the middle near the ends of the elytra (Tansey, 2007). The yellow stripes do not reach the posterior elytral margins, distinguishing P. striolata from another flea beetle, P. robusta (Burgess, 1977a).

As in P. cruciferae, there are eleven antennae segments but the fifth antennal segments of males are enlarged (Balsbaugh and Hays, 1972). The egg laying, larvae development and colorings are similar to P. cruciferae, (Burgess, 1977a) making species identification at the larval stage very challenging. The two species are found to coexist in most areas with similar host preferences and requirements although P. cruciferae has been the predominant species

(Burgess, 1977a), recently P. striolata is becoming more predominant (Tansey et al., 2008).

1.1.1.2. Life cycle

Both major species of flea beetles have one life cycle per calendar year (Figure 1.2.), however, if the conditions are right, two cycles may be possible (Westdal and Romanow, 1972). Adult flea beetles overwinter primarily in leaf litter and surrounding shelterbelts (Burgess, 1977a, 1981;

Wylie, 1979). They emerge in the spring, becoming active at 14oC. Flea beetle activity is greatest in the spring when the weather is sunny, warm and dry. Cool, damp conditions can reduce the feeding intensity of the beetles and aid plant growth to the point where they can withstand the feeding damage (Burgess, 1977a). The characteristic type of flea beetle injury to plants consists of small holes or pits in the epidermis of leaves. Damage occurs when adults feed on cotyledons and stems of seedlings, resulting in reduced photosynthetic capability, wilting, or host plant mortality (Westdal and Romanow, 1972). This feeding continues as the plant grows but is less

8 detrimental as the plant is able to compensate (Gavloski and Lamb, 2000). The first 20 days after emergence is the most crucial period for canola development, as flea beetle feeding has been shown to reduce yield potential even at low pressure during that time (Bracken and Bucher,

1986).

After emergence in the spring, flea beetles mate, lay their eggs in or on the soil and then die in late June or early July. The larvae develop in the soil and have been shown to feed on the roots of the host plants (Burgess, 1977a). Damage by larvae has been shown to be significant (Bracken and Bucher, 1986) but has not been studied extensively. Pupation occurs in late June into July with the new generation of adults emerging in late July and August (Westdal and Romanow,

1972). The new generation of flea beetles is known to feed on green, maturing Brassica crops, removing the epidermis of the stems, leaves, and pods, thus stunting the growth of the seeds

(Burgess, 1977a). This late season damage is not usually significant but may lend the host plant to increased susceptibility to secondary infection.

1.1.2. Flea beetle management

Flea beetle management involves several aspects of a typical pest management strategy. These include cultural, biological, chemical and genetic control of the pest in interest, in this case, flea beetles.

1.1.2.1. Cultural control

Modern day agriculture is a monoculture cropping system where single species cropping dominates. Gone are the days of mixed cropping or small scale farms with varied crops. The monoculture cropping system of modern farming may be partially responsible for the increasing issue of pests in crops. In a natural or more diversified environment, ecological predictions would predict herbivores to decrease, while predators are predicted to increase (Root, 1973; Hooks and

Johnson, 2003). Work by Garcia and Altieri (1992) and Tahvanainen and Root (1972) supported this theory reporting that flea beetles migrates from a diculture crop to a monoculture crop. They attributed this finding to crop architecture, claiming the companion crop creates unfavorable conditions for the flea beetle. Intercropping studies have shown reductions in flea beetle feedings and subsequent populations (Tahvanainen and Root, 1972; Altieri and Schmidt, 1986;

Garcia and Altieri, 1992). Cromartie (1975) and Kareiva (1983) both found that as plot size increased, the frequency of the two species of flea beetles (P. cruciferae and P. striolata) also increased, which is in agreement with (Knodel and Olson (2002) who also reported that crop rotations did not provide an effective management option for flea beetle control due to the strong flying ability and wide distribution of flea beetles.

The movement of insects to more favourable neighbouring hosts is part of the “push-pull” strategy of pest management first coined by Pyke et al. (1987) as referenced by Cook et al. (2007).

In the push-pull strategy, deterrent plants are interplanted within the crop with preferred trap crops planted in the surrounding area (Cook et al., 2007). Trap crops have shown some effectiveness in reducing flea beetle damage within a crop (Altieri and Schmidt, 1986; Bohinc and

Trdan, 2013; Metspalu et al., 2014). As flea beetles migrate from the nearby grassland areas into the crops, they find suitable host plant requirements and do not venture further (Root, 1973).

Stinner et al. (1983) suggests that more specialized phytophagous species generally have lower dispersal rates, however, Knodel and Olson (2002) report flea beetles have strong flying abilities.

This suggests that flea beetles are capable of readily moving to a primary host area but once the hosts are located, they do not leave.

The preceding observation supports Root's (1973) resource concentration hypothesis theory well. He suggested that generalists have more varied dietary requirements than specialist herbivores. As a result, generalists are more mobile in searching for food. On the other hand, specialist herbivores tend to aggregate in a more localized area. This is seen in flea beetles where mating, oviposition and adult and larval feeding all occur on the same plant species or group of species (Burgess, 1977a).

Agronomic practices can be altered to reduce flea beetle pressure in modern farm practices.

Dosdall and Stevenson (2005) determined that flea beetle pressure can be reduced by planting in the fall instead of in the spring, increased seeding rates and using Brassica napus varieties versus B. rapa. Knodel and Olson (2002) recommend early planting from April to mid-May to reduce damage due to flea beetles. Other effective strategies for cultural control of flea beetles include using zero-tillage versus conventional tillage (Cromartie, 1975; Dosdall et al., 1999), using large seeds instead of small seeds (Bodnaryk and Lamb, 1991a; Elliott et al., 2008) and using wide row spacing (Dosdall et al., 1999).

1.1.2.2. Biological control

There are limited biological control methods for flea beetles. Currently, none are considered feasible for modern farming practices. The primary option for biological control involves the encouragement of natural predators. Natural enemies of flea beetles include lacewing larvae

(Chrysopa carnea), big-eyed bugs (Geocoris bullatus), two-lines collops (Collops vittatus), the western damsel bug (Nabis alternates) and the northern field cricket (Gryllus pennsylvanicus) along with parasitic wasps such as Microtonus vittate. Two predatory heteropterans, Podisus maculiventris (Say) and Nabicula americolimbata (Carayon) were recorded feeding on the adults of P. striolata, but data on their control potential are lacking (Burgess, 1977b, 1980, 1982;

Culliney, 1986). Due to the rapid emergence of the flea beetle population in the narrow window during the spring, natural enemies are unable to have an impact on flea beetle control (Wylie et al., 1984; Knodel and Olson, 2002).

Entomopathogenic nematodes (EPNs) may be potential alternatives for the control of soil- dwelling pests (including larval stage of flea beetles) because of their ability to actively search for their hosts (Xu et al., 2010). Although this would not reduce feeding on newly emerged cotyledons on spring planted crops, it may reduce populations of flea beetles for subsequent years and reduce flea beetle pressure for fall planted crops. Antwi and Reddy (2016) report EPNs are suitable alternatives to conventional chemical seed treatment and can complement the effects of conventional chemical seed treatments once conventional chemicals are no longer effective.

1.1.2.3. Chemical control

Chemical control is the primary control defense against flea beetles in North America, with more than 99.5% of the Canadian canola acres treated with systemic insecticidal seed treatments

(Sekulic and Rempel, 2016). These seed treatments last up to 40 days. After that, foliar post- emergent sprays are used. Post-emergence sprays applied to plants after some exposure to flea beetle damage are generally less effective in preventing yield loss than in-furrow or seed treatments that provide continuous protection during and after germination (Bracken and

Bucher, 1986). Today’s options are primarily neonicotinoids where older technology is pyrethrin- based. Recent reports (Tansey et al., 2008, 2009) have shown a species differentiation in effectiveness of neonicotinoid seed treatments. With increasing environmental concerns and bans (Health Canada, 2016; Ontario, 2016) along with the persistence of flea beetle pressure, alternate, non-chemical control would be preferred, although seed treatments are relatively easy to apply and highly effective for the time being.

1.1.2.4. Genetic control

Genetic control of flea beetles has been studied over the past several decades, however, little progress has been made in developing a commercial variety. The interaction between the flea beetle and the host plant has been studied by many and the general consensus appears to be multigenic resistance with numerous small effects from different mechanisms. Studies by

Bodnaryk and Lamb (1991b) and Lamb et al. (1993) and others (Brandt and Lamb, 1993) show that genetic variation is present in the genetic pool of host plants based on the feeding responses

13 to the hosts but significant or chemical replacement levels of flea beetle resistance has yet to be realized (Patel, pers. comm.).

Transgenic options have also been investigated. Current work by Gruber et al. (2006) and Soroka et al. (2011) involves transgenic trichome enhanced resistance to defend against flea beetle feeding. Chemical resistance within plants using transgenic means includes Bt (Bacillus thuringiensis-derived natural insecticide) (Gatehouse, 2008) and secondary metabolite modifications such as cyanogenic glycosides (Tattersall et al., 2001; Kristensen et al., 2005).

1.2. HOST PLANT SELECTION AND BEHAVIOR OF PHYLLOTRETA SPP. ON BRASSICACEAE

Insects go through several phases in finding a host plant as described by Schoonhoven et al.

(2005). Firstly, they must search for the host. This requires the insect to move towards and contact or at least remain near a host plant. Ultimately, it must select an appropriate host. This involves multiple sensory cues and short-term memory of the insect to recall previous choices in order to decide which host to accept. Prolonged feeding and/or ovipositing indicates acceptance of the host plant.

Herbivorous insects use multiple mechanisms and strategies to determine host plant selection.

The strategies or mechanisms depend upon the insect’s geography, ultimate goal and options available to it. For long range searches (greater than 1m), visual and olfactory cues direct an insect towards a host plant (Schoonhoven et al., 2005). When closer than one metre, short range mechanisms are used to determine the plant’s suitability as a host. This includes evaluation of

14 physical and/or chemical properties of the host plant by the insect, including preference to specific plant geographies such as upper and lower leaf surfaces (Matsuda, 1988; Reifenrath et al., 2005). The interaction between the chemicals elicited by the plant and the insect herbivore affect the attractiveness of the host plant to the herbivore, its feeding habits and ovipositing

(egg-laying) (Schoonhoven et al., 2005; Hilker and Meiners, 2011). These same cues may also attract or deter predators of the herbivorous insects (Cusumano et al., 2015) or trigger host defense mechanisms as in the case of herbivores laying eggs on the plants (Hilker and Fatouros,

2015). Host preference and fecundity are expected to be correlated to reproductive performance, resulting in the females likely being more sensitive to host plant characteristics

(Jallow et al., 2004). Generally, female herbivores consume more foliage due to the high protein dietary requirements associated with ovipositing (Schoonhoven et al., 2005). The ratio of male and female flea beetle feeding on Brassicas in the wild is poorly reported. In the end, the insect’s ultimate goal is to reproduce, therefore, insects will adjust the breeding strategy to ensure success. This may include delayed egg laying, reduced or increased egg deposition (Awmack and

Leather, 2002).

1.2.1. Host plant selection

Host plant selection behaviours vary depending upon the environment and physiological stage of the herbivore. Different life stages of the same insect may require different host plant species as their dietary needs differ (Schoonhoven et al., 2005). The need for host plant selection therefore varies as a result of not only the different life stages of the herbivore but also the

15 environmental changes throughout the herbivore’s life cycle. For herbivores to survive, they must be capable of locating appropriate food.

1.2.1.1. Orientation

The process of locating a suitable host requires two steps; searching and contact testing

(Schoonhoven et al., 2005). Generally, insects show a “programmed behaviour” that is stereotypical of the species (Schoonhoven et al., 2005). Henderson et al. (2004) were able to determine such a pattern in flea beetle feedings in contact with possible hosts. In order to find a host though, many insects use visual and olfactory stimuli (Schoonhoven et al., 2005). Insects fly in random patterns until the odour plume of the plant of interest is located. Once in the plume, the insect remains in the odour stream in a relatively straight pattern until it overshoots the odour source, at which time it returns to the random zig-zag pattern to search for the lost odour plume (Beck and Schoonhoven, 1980; Chapman, 1998). Crucifer flea beetles have been reported to move randomly within and between host patches with little change due to the concentration of host odour (Bernays and Chapman, 1994). Random activity is an efficient search strategy under many circumstances and may assist when host plant signals are weak and host plant suitability is variable (Bartumeus and Catalan, 2009).

During the random activity, flea beetles may be using both olfactory cues plus visual cues

(Schoonhoven et al., 2005). Using these stimuli, the insect is directed towards or away from the host plant. The ease of locating a host plant may depend on the physical environment as well and not just the host characteristics. When host populations are large, the hosts may be easier to

16 locate due to shear volume in optical and olfactory stimuli. When individual plants are evaluated from a larger group of similar species, less herbivory is observed as there is more food to go around and resistance may appear better. The opposite situation results when few plants are available. Although harder to find, once found, host plants tend to suffer more as they are hosting more herbivores, even though the herbivore population may be unchanged (Pimentel, 1961).

1.2.1.2. Optical

Optical stimuli are considered fixed and fast moving (speed of light) but confined to a small area

(not detected from great distances). They are generally unaffected by fluctuating weather conditions but are subject to environmental influence (Schoonhoven et al., 2005). Prokopy and

Owens (1983) conclude four important roles of visual stimuli in host-plant selection. Stimuli include shape, spectral quality, detailed dimensional or pattern characteristics and finally the role of background composition (soil, sky, etc.). From a distance, insects may be able to determine the shape of possible hosts as silhouettes against a contrasting background. This may be newly emerged seedlings in a cultivated field in the case of flea beetles. Secondly, within metres of the potential host plant, the spectral quality of the plant becomes the predominant cue. At this point, the herbivores are trying to identify the best host amongst a number of similar possible hosts.

Using spectral reflectance, insects may be able to determine the landing position on the host or select a plant with a better nutritional value than a similar neighbouring plant. Third, at close range (less than one metre), the detailed dimensional or pattern characteristics of host or non- host plants are not discernible except for a few plant species. This has been shown in grasshoppers (Mulkern, 1969) but not reported for flea beetles. Finally, the role of background

17 composition (soil, sky, etc.) plays an important but poorly understood role in host-plant selection.

The larval stage of flea beetles takes place in soil so soil composition may play a role in host plant selection too.

Many studies have involved optical recognition. Light contains many factors such as intensity, spectral composition and polarization, however, they are generally unaffected by air movement but is fast moving (can be detected from a distance) (Schoonhoven et al., 2005). Although plants are generally green, reflected light may differ amongst different plant species due to different leaf surface composition such as trichomes and wax crystals and also due to various biotic (age, nutrition) and abiotic (density, incident light intensity, background) factors (Prokopy and Owens,

1983; Schoonhoven et al., 2005). The ability to distinguish wavelengths has been shown in a number of species, including those in Coleoptera (Chapman, 1998). To be able to accomplish this, separate retinula cells containing photopigments with maximum sensitivity to light of different wavelengths are needed. All insects have a visual pigment in the green range of the spectrum

(maximum 490 to 540nm) with the range of absorption from wavelengths below 400nm (UV region) to maximum 600nm (orange). Most insects have two additional pigments with maximum sensitivity to wavelengths in the UV and blue regions (Chapman, 1998). Wavelengths for flea beetles have not been reported but preference has been shown for whites, yellows and greens

(Beck and Schoonhoven, 1980; Prokopy and Owens, 1983; Al-Doghairi, 1999). These are colours that may be attractive in the late season when the new adults emerge. They are also the basic colours for foliage, in general, beginning with the cotyledons, when new plant material is starting to emerge.

1.2.1.3. Olfactory

Complimentary to optical stimuli, olfactory cues involve chemical stimuli. These are subject to air movement and can move rather slowly in still conditions. Under natural conditions, odour concentrations are highly variable and can move great distances as packets of odour

(Schoonhoven et al., 2005). For chemical detection, insects rely on sensillia located on the antennae for long range host plant recognition (Mitchell, 1988). In the Brassicas, allyl isothiocyanates have been identified as a strong attractive cue for flea beetles and is commonly used in flea beetle traps (Mitchell, 1988). However, the accuracy of using one chemical versus multiple chemicals has been shown by several authors. Vincent and Stewart (1984) compared flea beetle sweeps to allyl isothiocynate baited traps and noted discrepancies as to which flea beetle species where prevalent. Depending on whether a physical net or baited traps were used,

Vincent and Stewart (1984) observed the ratio of different species changed depending on the trapping technique used. Vincent and Stewart (1984) attributed this difference to varied host specificity among the different flea beetle species. It has also been noted that female flea beetles are more responsive to chemicals (Vincent and Stewart, 1984). This sex-type difference is not unique to flea beetles and has also been reported in Diabrotica virgifera v virgifera (Andersen and Metcalf, 1986).

1.2.2. Host plant acceptance

Once a host has been located, it must be accepted. This requires the herbivore to physically contact the host plant. For host plant acceptance, flea beetles use a combination of physical and

19 chemical cues. Locomotion is generally restricted to a small area (Schoonhoven et al., 2005), at which time, the herbivore uses various sensilla to measure mechanical, olfactory and hygro/thermal stimuli (Ritcey and McIver, 1990).

Henderson et al. (2004) under artificial experimental conditions called the first phase of searching

“acclimation”. When searching for food, brief periods of antennal waving was soon combined with tarsal tapping and tarsal rubbing on the cotyledon surface. Brief periods of grooming or rest interrupted antennal and tarsal movements. Once the antennae contact the plant tissue, the flea beetle has moved from this acclimation stage to the stimulation phase. The stimulation phase for flea beetles also involves complex combinations of antennal waving, tarsal tapping and tarsal rubbing, with frequent but brief interruptions of grooming (Henderson et al., 2004). Sensilla are located on the tarsi and antennae so the movements described are likely ways of enhancing olfactory detection.

1.2.2.1. Receptors

Sensilla are simple sensory receptors consisting of one cell or a few cells. The term sensilla refer to the basic structural and functional unit of cuticular mechanoreceptors and chemoreceptors.

The sensilla are composed of a cuticular structure, the neuron or neurons, the associated sheath cells with cavities they enclose and the structures they produce (Chapman, 1998).

There are different types of olfactory sensillia that respond to different classes of chemicals.

Numerous small pores allow entry of the chemical stimuli into the olfactory sensillia (Chapman,

1998). These are located primarily on the antennae of all insects along with various other organs such as the maxillary and labial palps and the genitalia (Chapman, 1998). Depending on the architecture of the antennae, more sensillia may be present, thus allowing for detection of more odours (Chapman, 1998). In P. cruciferae, antennal olfactory sensilla appear to be used for host plant locating and recognition rather than mating as both the male and female antennal architecture appears alike (Ritcey and McIver, 1990). Depending on the location of the sensilla, different stimuli may be recognized and the same sensilla may detect both chemical and physical stimuli (Chapman, 1998).

Contact chemoreceptors (taste) are used by the insect to determine specific key components of food or contact pheromones. Contact chemoreceptors, in varying concentrations, may be found on all parts of the insect’s body except the mandibles and within the gut. Major concentrations are found on various mouthparts. As well, the antennae, especially the tip, harbour a large number of chemoreceptors too (Chapman, 1998). Chemoreceptors may be specific to certain components such as glucosinolates (Stadler et al., 1995). Concentrated on the mouthparts and antennae, contact chemoreceptors are located on various insect parts but are not found in the gut (Chapman, 1998). Contact chemoreceptors are sensitive to different chemical types (salt, sugar, amino acid or deterrent) and may be sensitive to different subgroups of chemical types depending on the receptor location. This enables the insect to determine a bigger picture of the stimuli it is dealing with (Chapman, 1998).

As mentioned earlier, various mechanoreceptors and chemoreceptors are used to detect stimuli via contact with the plant leaf surface and may promote settling (Southwood, 1996). Once a host plant has been “short-listed”, physical attributes come into play prior to feeding or ovipositing.

Insects use mechanoreceptors to determine the surface structures. Insects are looking for the optimal pubescence, whether relatively glabrous or heavily pubescent, particular shapes, crevices, or for soil ovipositing insects, they are concerned with soil particle size (Beck and

Schoonhoven, 1980). Flea beetles of canola would be concerned with both leaf and soil types as they use both at various times in their life cycle.

Finally, host plants must pass one last test before acceptance. Many insects determine the quality of the plant by probing the plant surface with their antennae or tarsi but without biting into the tissue (Roessingh et al., 1997; Isidoro et al., 1998; Schoonhoven et al., 2005). Special insect movements may be performed to help intensify the chemical stimulation the insects are probing for (Beck and Schoonhoven, 1980). If after brief periods of test biting (less than 18 seconds) there is insufficient chemosensory stimulation, flea beetles have been shown to return to the acclimation or stimulation stages. Flea beetles that feed for longer than 18 seconds do not return to previous stages. It is suggested that sufficient chemosensory information has been collected during the stimulation stage and the potential host plant passes (Henderson et al., 2004).

1.2.2.2. Host plant quality

It has been shown that insect growth and reproduction are positively correlated with nitrogen content of the food (Scriber and Slanksky Jr., 1981). Plant quality plus the allelochemic

22 composition are important factors influencing host acceptance and ultimately herbivore performance. Finally, biotic and abiotic stresses such as drought can alter the nutrient and allelochemic composition of plants and thereby affect larval performance (Schoonhoven et al.,

2005).

1.2.2.3. Experimental terminology

The world of research introduces other terms associated with locating a suitable host plant. They are preference and recognition. Preference is a term used in choice assays. The host is not

“found” but given the choice, one host is preferred over another and is therefore acceptable. It may not be the accepted choice in an unrestricted environment, hence the difference between accepted and acceptable host plant. Recognition is also associated with acceptance. In this case, a familiar host is recognized indicating a neural mechanism. This recognition is thought to be genetic and may play a role in host selection in cases where an adult insect feeds on a particular plant and the resulting larvae “recognize” the host plant (Schoonhoven et al., 2005). As a researcher, one must be aware of such biases as they may not be seen in a less controlled setting.

1.2.3. Chemical cues

The most important factor for insect-host plant determination is considered to be plant chemistry

(Matsuda, 1988). As previously discussed, insects use olfactory stimuli to locate a potential host plant but also to stimulate biting, probing and ovipositing once the insect is in physical contact with the plant (Beck and Schoonhoven, 1980). Stimuli such as wind direction and visual targets along with odour direct the insect to the potential host (Beck and Schoonhoven, 1980).

Difficulties arise when polyculture environments are encountered. Visser (1986) suggests plant odours overlap and as a result, the herbivore may not be able to distinguish it hosts suitability, thereby not selecting it. This may explain results by Garcia and Altieri (1992) where single species environments are more susceptible to herbivory than mixed environments. Kareiva (1985) noted that movements of flea beetles to target Brassicaceae host declined as the distance from the host increased, with a greater effect for P. striolata versus P. cruciferae. Even at small distances of 4m, flea beetles were not able to locate the host plant very well, indicating that olfactory cues may not be as great of a stimulus as generally thought.

1.2.3.1. Behavourial stimuli

It has been widely suggested that the same chemical or combination of chemicals in such plants like the Brassicaceae may act as both attractants and repellants or deterrents and stimulants, depending on whether the herbivore is a generalist or a specialist (Renwick, 2002). Dethier et al.

(1960) list five classifications of stimuli – attractants, arrestants, repellents, stimulants and deterrents. Of these five, the first three are olfactory whereas the latter two are gustatory (Beck and Schoonhoven, 1980). Miller et al. (2009) recently updated these classifications based on new information regarding stimuli. They replace the term “stimuli” with “locomotor initiator” and

“locomotor stimulant” depending upon the level of kinetic motion. They have now added the terms “engagent” and “disengagent” referring to the chemicals effecting locomotion. Dethier et al.'s (1960) original classifications are expanded below.

i. Attractants

Attractants elicit a behavioural response towards the source (Dethier et al., 1960). Renwick

(2002) lists isothiocyanates as an attractant for specialist insects. Working on a range of wild and cultivated Brassica species, Cole (1996) found a strong link between the intrinsic rate of increase of B. brassicae and a combination of four glucosinolates.

ii. Arrestants

Arrestants are agents that cause insects to aggregate but to cease locomotion (Dethier et al.,

1960). This may or may not be the same as an attractant. They can be chemical but also visual, mechanical and thermal (McGovran, 1969).

iii. Repellants

Repellants cause an insect to make behavioural movements away from the source (Dethier et al.,

1960; Visser, 1986). Repellants have been reported in several species (Visser, 1986) including high concentrations of sinalbin to P. cruciferae (Bodnaryk and Rymerson, 1994). These chemical cues likely occur in short range situations (Visser, 1986).

iv. Stimulants

Stimulants elicit feeding, phagostimulation, oviposition, etc. (Dethier et al., 1960). Schoonhoven et al. (2005) indicate the number of secondary plant metabolites from non-host plants that act as a stimulant is increasing, especially for some phenolic acids and flavonoids. For flea beetles, glucosinolates is considered a stimulant (Renwick, 2002). Along with glucosinolates, other metabolites are being reported as necessary for flea beetle recognition of the host plants. For horseradish flea beetles to recognize its host, glucosinolates alone were not enough. Nielsen et al. (1979) report flavonoids and other sugar moieties play an important role in host plant

25 recognition. This combination of glucosinolates and flavonoids has been shown to increase feeding preferences in specialist caterpillars of plants in the Brassicaceae (Hopkins et al., 2009).

This has also been seen with some flea beetle species (Larsen et al., 1982). Allyl isothiocyanates used in baited traps trapped more female than male versus unbaited traps where the male to female ratio was 1:1. This led Vincent and Stewart (1984) to suggest allyl isothiocyanates are not only feeding stimulant but may also be ovipositing stimuli too. Similar results with baited traps were reported by Wylie (1981). However, Pivnick et al. (1992); Vincent and Stewart (1984) observed greater effect with P. cruciferae than P. striolata indicating some special differences in chemical stimulants.

v. Deterrents

Deterrents inhibit feeding or oviposition, where absent would be acceptable (Dethier et al.,

1960). Deterrents, both physical and chemical, play an important role in host plant selection.

Chapman (1974) reviews the wide spectrum of secondary compounds that act as feeding deterrents. Physical deterrents will be discussed later. Jermy (1966) demonstrated that host- plant rejection by various insects is due to the presence of deterrents or feeding inhibitors.

Glucosinolates have been considered a primary deterrent for generalist herbivores (Renwick,

2002) but are not a deterrent to specialists such as the flea beetles. Reported flea beetle deterrents are cardenolides (Nielsen, 1978; Nielsen et al., 1979), flavonoids ((Matsuda, 1976) as cited by Nielsen et al. (1979)) and saponins (Nielsen et al., 2010a; b). Cucurbitacins and cardiac glycosides have been identified by Nielsen (1978) and Nielsen et al. (1977) as deterrents to some flea beetle species.

1.2.3.1. Pheromones

Pheromones are chemicals produced by the herbivore and not the host plant as the previous five behavioural stimuli were. There appears to be increasing evidence of pheromones being produced by flea beetles. In particular, male-specific sesquiterpenes are being produced by male flea beetles (Peng and Weiss, 1992; Peng et al., 1999; Bartelt et al., 2001; Beran et al., 2011).

These are found in several flea beetle species and appear to be equally sensed by both sexes. As a result, they are likely being used to locate nearby feeding and ovipositing spots (Bartelt et al.,

2001).

On any given plant, multiple chemical stimuli may be active. From species to species or plant to plant, the chemicals involved may result in different behavioural outcomes. This includes the effect of neighbouring plants, particularly on multi-species environments. Tahvanainen and Root

(1972) have shown that non-host plants contain enough olfactory cues to repulse herbivores from otherwise host plants.

1.2.4. Physical cues

Physical cues for host selection vary depending on the insect-plant relationship. Visual cues such as plant tissue colour have been observed in several insect-plant relationships (Schoonhoven et al., 2005). Other physical cues can be deterrents. These include surface waxes and trichomes

(Eigenbrode and Espelie, 1995). Both of these deterrents interfere with insect attachment, ovipositing, feeding and movement in other cases it has been shown to help phytophagous insects (Eigenbrode and Espelie, 1995).

1.2.4.1. Plant tissue colour

In the case of flea beetles, studies by Al-Doghairi (1999) concluded that flea beetles are attracted to the colours Saturn green, Saturn yellow and white significantly more so than the transparent control. However, at the cotyledon stage, these colours are not available indicating that perhaps visual stimuli are not as important cues for host detection than others. Spectral analysis does not seem to have been reported in the literature at either the cotyledon or true leaf stage.

1.2.4.2. Trichomes

The role of trichomes in plant defense is multifaceted. They interfere with plant-insect contact, produce toxic compounds and can produce chemical exudates to interfere with insect movement

(Howe and Jander, 2008). Electroantennogram studies on trichome-protected Brassicas indicate that trichome protected plants are as stimulating as low-density or no trichome plants. This suggests trichomes and no other factors such as chemical repellents or feeding deterrents are responsible for this protection (Palaniswamy et al., 1997; Traw and Dawson, 2002a; b). Studies by Traw and Dawson (2002a) conclude that flea beetles do not induce a defense response unlike

Pieris which resulted in altered trichome and glucosinolate composition. Therefore, they conclude trichomes play a role in induced defences but vary depending on the herbivore.

However, they also concluded that flea beetles are not as sensitive to the health of the plant versus Pieris showing no significant change in mass or feeding but did have higher mortality on unhealthy plants. Gruber et al. (2006) has introduced two trichome regulatory genes from

Arabidopsis into canola. As a result, trichome density on canola seedling leaves and stems

28 increased 1000-fold. This increase in trichome density has resulted in reduced flea beetle feeding on canola plants as well as observations of antixenosis and antibiosis resistance when exposed to diamondback moths (Plutella xylostella) (Alahakoon et al., 2016). The elevated trichome numbers are found only on stems and leaves, thus the cotyledons remain trichome-free, an area highly susceptible to flea beetle feeding.

1.2.4.3. Waxy leaf surfaces

Waxy surfaces can both reduce and increase herbivory, depending upon the herbivore

(Eigenbrode, 2004). Wax surfaces can interfere with insect attachment for feeding therefore reducing herbivory. They can also interfere with natural predators of herbivores, therefore allowing increased herbivory. In Brassicaceae, waxy leaves have been shown to reduce flea beetle feeding and alter the feeding pattern where it has been shown to be associated with the CC genome (Anstey and Moore, 1954; Bodnaryk, 1992a). Waxes have been shown to contain a mixture of long-chain aliphatic components primarily including alkanes, alkyl esters, aldehydes, primary alcohols, and fatty acids along with, cyclic compounds such as phenolics, flavonoids, and triterpenoids can be present (Jetter and Riederer, 1996; Hare, 2011). Determining if the chemical stimuli are part of the wax layer has been a challenging quest. Improved methodologies indicate that surface waxes do not contain glucosinolates that are attractive to some insects (Reifenrath et al., 2005). They further suggest that glucosinolates may be released through stomata. This would coincide well with increased flea beetle feedings during warmer weather when stomata are open. However, it appears that the chemical ratio and composition may be system specific.

1.2.5. Cost of resistance

Insect resistance does not come free. There is always a cost associated with expenditures; herbivore defense being one of them. The mechanism of resistance is determined by the frequency and severity of attack. The costs of resistance or tolerance may result in sub-optimal outcomes in the “better” plant such as reduced yield or may have no notable effects at all, depending on the characteristics measured (Smith, 2004). The costs of trichome and glucosinolates have been studied in Arabidopsis by Züst et al. (2011), where they reported that herbivory reduced plant growth rate and delayed flowering. It should be noted that most studies with herbivores are single species or single type of herbivory in monoculture. Generalist and specialist herbivores react differently and often co-exist. Selection pressures from these different herbivores and how they interact may confound the results seen in the whole community when compared to controlled experiments (van der Meijden, 1996)

Not only do defense mechanisms impose cost on the host plant, herbivores that have counter adaptations to the defense systems play a price too. Insects have been shown to perform better in controlled laboratory conditions likely due to the lack of physical and chemical plant defense factors (Beck and Schoonhoven, 1980). Faster growth with larger weight and improved fecundity and longevity has been seen on several phytophagous Lepidoptera species when reared on culturing media as opposed to host plant tissue (Beck, 1974). The success of rearing insects on artificial medium suggests that physical cues play a minor role in food recognition and are thus more important in ovipositing (Beck and Schoonhoven, 1980).

1.3. FLEA BEETLE (PHYLLOTRETA SPP.) RESPONSES TO HOST PLANT SECONDARY METABOLITES AND MORPHOLOGY

Brassica production worldwide is challenged by the many pests which attack it. Several specialist insects such as Phyllotreta cruciferae Goeze, Pieris rapae Linnaeus, Plutella maculipennis Curtis and Psylloides chrysocephala Linnaeus prefer Brassica species as the host plants (Bonnemaison,

1965; Lamb, 1989; Giamoustaris and Mithen, 1995; Dosdall and Mason, 2010). As sessile organisms, plants have adapted varying defense mechanisms to keep the herbivores at bay. Such strategies include chemical and physical barriers such as induction of defense proteins like phytoalexins, release of volatiles to attract natural predators of the herbivores, production of toxic secondary metabolites and altered tissue surfaces such as wax layers and trichomes (Mello and Silva Filho, 2002). These defense mechanisms can be classed as constitutive, inducible, induced, direct and indirect defences acting to either make the potential host plant unsatisfactory (antixenosis) or acceptable but with deleterious effects such as toxification or reduced digestibility, therefore reducing herbivore development. These multiple defense mechanisms exist, function and express themselves in conjunction or opposite each other (Ahuja et al., 2010) so they may, in fact, be a defense system against some pests and a stimulus for another (Hopkins et al., 2009).

1.3.1. Glucosinolates

The family Brassicaceae is renowned for production of a specific secondary metabolite referred to as glucosinolates (anionic glucosides) (Fahey et al., 2001). First associated with the sharp taste of mustard, the chemical properties of glucosinolates and their breakdown products

(isothiocyanates or mustard oils) have been found to have fungicidal, bacteriocidal, nematocidal and allelopathic properties as well as cancer chemoprotective attributes (Fahey et al., 2001).

Over 120 different glucosinolates have been identified with at least 30 found in Brassica species

(Fahey et al., 2001; Bellostas et al., 2007; Hopkins et al., 2009). Glucosinolates contain sulphur and nitrogen units and can be either acyclic or aromatic in structure (Schoonhoven et al., 2005).

Numerous studies show that glucosinolates play a key role in communication between the plant and the insect whether as a defense mechanism or a feeding and ovipositing stimuli (Matsuda,

1988; Mithen et al., 1995; Bartlet et al., 1996; Moyes et al., 2000; Halkier and Gershenzon, 2006;

Hopkins et al., 2009; Bohinc et al., 2013). Intact glucosinolates alone may provide resistance to insect herbivores (Kim and Jander, 2007) in particular against the generalist insects (Li et al.,

2000), but the response is enhanced when the plants are damaged. Glucosinolates are stored in the vacuole of cells. When plant cells are ruptured, glucosinolates are hydrolyzed by the myrosinase enzyme; a thioglucosidae found within special myrosinase cells located throughout the plant (Rask et al., 2000). As a result, glucose and sulphate is released, resulting in the production of pungent toxics such as isothiocyanates, nitriles, and oxazolidinethiones (Wittstock and Halkier, 2002; Bones and Rossiter, 2006).

Further evidence glucosinolates play an important role in plant defense is based on the varying amounts of glucosinolates within a plant at different ontogenetic stages, in different plant organs and various environmental conditions (Velasco et al., 2007; Badenes-Perez et al., 2014). In seedlings, cotyledons have the highest glucosinolate levels whereas roots have higher

32 concentrations than above ground tissue. Of the above ground tissue, the highest glucosinolate levels are found in strategically important tissue such as the reproductive structures including flowers and seeds (Brown et al., 2004; Hopkins et al., 2009). Clossais-Besnard and Larher (1991) showed the changes in glucosinolate concentration and types at various stages of early plant development. The glucosinolate profile changes rapidly as the seed germinates. During the first several days, the glucosinolate profile of the new seedling is intermediate between seeds and vegetative tissues vegetative. Once vegetative tissue begins to emerge, total glucosinolate content increases and then the ratio of the different glucosinolates changes from a seed profile to vegetative one (Brown et al., 2004). However, Brown et al. (2004) noted that glucosinolate concentrations decline during seed germination and leaf senescence. Mithen (1992) reported that low glucosinolate Brassicas (canola quality) are no more susceptible to pests and pathogens than high glucosinolate (non-canola quality) Brassicas, therefore, indicating the potential to manipulate leaf glucosinolates without compromising seed quality. The sulphur nutritional status of Brassica juncea plants were evaluated by Aghajanzadeh et al. (2014) to determine if and how sulphur affects the type and composition of glucosinolates. Results from Aghajanzadeh et al.

(2014) may provide insight on the role of myrosinase activity and ultimately a role in plant defense. Bodnaryk and Lamb (1991b) concluded high concentrations of sinalbin in young cotyledons of S. alba were repellent to P. cruciferae, but lower concentrations in older leaves did not offer any protection against either flea beetle species – P. cruciferae nor P. striolata. Hopkins et al. (1998) reported more than 90% of the primary glucosinolates in S. alba lines consisted of sinalbin. There was little difference amongst the lines for flea beetle damage nor concentration levels of sinalbin in the cotyledons and the proportion of sinalbin declined at later growth stages,

33 especially in the “low” breeding lines (Mithen et al., 1995). Nielsen et al. (2001) reported differing feeding responses in two flsea beetles species using transgenic Arabidopsis thaliana modified for elevated glucosinolate levels.

There are three types of glucosinolates classified based on the amino acid precursor (also referred to as R-side chain) – indole (tryptophan precursor), aliphatic (methionine precursor) and aromatic (tyrosine or phenylalanine precursor) (Giamoustaris and Mithen, 1995). The different classes are glucosinolates are synthesized in different ways and function differently (Parkin et al.,

1995; Gigolashvili et al., 2007; Hirai et al., 2007) with some shared enzymes (Kim and Jander,

2007). Indole glucosinolates are common and subject to environmental conditions more so than the other glucosinolate types (Velasco et al., 2007). Bodnaryk and Rymerson (1994) and Bodnaryk

(1992b) noted that indole glucosinolate content increased in plants exposed to flea beetle as well as when jasmonic acid was applied. Indole glucosinolates therefore appear to be more of a feeding and/or ovipositing stimuli (Agerbirk et al., 2009). However, Traw and Dawson (2002b) reported no change in leaf glucosinolates after 12 hours to 7 days for P. cruciferae feeding.

Aliphatic glucosinolates have been shown to be highly heritable with little environmental influence (Kushad et al., 1999). Although little work has been done on above and below ground glucosinolates, they do appear to be independent and can be selected for independently from one another (Kirkegaard et al., 2001) and thus prove to be a possible tool for reducing flea beetle larval growth.

Numerous studies have determined that plant defense may be induced as a result of herbivore damage or other physical damage with long lasting effects (Bartlet et al., 1996; Poelman et al.,

2008) or on different plant parts (Van Dam et al., 2005). This includes production of primary and secondary metabolites as well as altered nutritional quality (Van Dam et al., 2005; van Dam and

Oomen, 2008). Glucosinolates and other secondary metabolites can be deployed to needed areas via phloem transport (Chen et al., 2001) or de novo synthesis. Several studies indicated that the genetic expression increases in response to herbivore injury and that glucosinolate synthesis occurs rapidly (Mewis et al., 2006). Flea beetles would be one of the first, if not the first herbivore to feed on developing Brassica plants so induced defense would be possible.

The glucosinolate defense requires the use of the myrosinase enzyme to create the toxic secondary metabolites. Myrosinase and glucosinolate levels within members of the Brassicas have been observed to be proportional (Li and Kushad, 2004; Bones and Rossiter, 2006). Many reviews have covered the biological and ecological significance of the glucosinolate-myrosinase system (Rask et al., 2000; Bones and Rossiter, 2006; Borgen et al., 2010).

Bodnaryk and Palaniswamy (1990) report glucosinolate levels in cotyledons do not determine feeding levels in flea beetles. They also suggested that reduction or elimination of glucosinolates would not be an effective means of protecting seedlings from flea beetle damage. Further examination by Bodnaryk (1997) of S. alba and B. juncea concluded that flea beetle resistance was independent of glucosinolate concentration. Lines with extremely high sinalbin concentrations and those lines with low cotyledon glucosinolates both showed flea beetle

35 resistance. Soroka and Grenkow (2013) also reported no differences in canola quality versus non- canola quality for flea beetle herbivory on Brassica napus, B. rapa, B. juncea and B. carinata.

When looking at glucosinolate levels in a wide range Brassicaceae, they observed decreased quantities of hydroxybenzyl and butyl glucosinolates in preferred canola-quality S. alba lines and increased levels of hydroxybutenyl glucosinolates compared with levels in condiment S. alba lines. This further shows the complexity of the role of glucosinolates in Brassiaceae and highlights potential sources of resistance.

Studies by Li et al. (2000) and Mitchell-Olds et al. (1996) reported lower feeding rates of specialist herbivores on plants with high myrosinase activity. This indicates that glucosinolates themselves are not the defense mechanism but the defense may be limited by the enzymatic capacity of myrosinase. Myrosinase activity was measured in several Brassica species by Henderson and

McEwen (1972). They determined myrosinase levels in Sinapis alba was greater than either

Brassica napus or B. rapa. Sinapis alba has long been regarded as a source of flea beetle resistance and this may be one of the contributing factors. This in turn suggests that varieties with increased myrosinase activity, whether by traditional or transgenic breeding methods, could provide the necessary flea beetle resistance sought. Borgen et al. (2012) were able to demonstrate aphid feeding preference was related to myrosinase activity in Brassica napus.

Wallace and Eigenbrode (2002) reported decreasing levels of glucosinolates and myrosinase as the seedlings aged, but noted that the level of myrosinase declined slower than the glucosinolates suggesting active retention of myrosinase activity as young cotyledons expanded.

This finding was further supported by the work by Barth and Jander (2006).

1.3.2. Cyanogenic glycosides

Both glucosinolates and cyanogenic glucosides are derived from amino acids and have aldoximes as intermediates. Cyanogenic glucosides are found throughout the plant kingdom as opposed to glucosinolates found primarily within the Capparales (Rask et al., 2000). As the name suggests, cyanogenesis is the process by which hydrogen cyanide is released from these cyanide containing compounds (Gleadow and Woodrow, 2002). As a result, they are generally avoided by herbivores. Like glucosinolates, certain herbivores are attracted to cyanogenic glucosides and are therefore not toxic to all herbivores (Gleadow and Woodrow, 2002; Zagrobelny et al., 2004).

Similar to glucosinolates, cyanogenic glycosides are stored as inactive forms in plant vacuoles.

When the tissue is damaged, the cell membranes are broken, releasing the glycosides which mix with the enzymes in the cytoplasm to release the toxic by-product of hydrogen cyanide.

As a possible defense mechanism against flea beetles, transgenic studies have been conducted.

A gene for production in cyanogenic glycosides in sorghum was inserted into Arabidopsis. It produced the glucosinolate sinigrin and improved herbivore resistance (Nielsen et al., 2001).

Tattersall et al. (2001) transferred the entire pathway for synthesis of the tyrosine-derived cyanogenic glucoside dhurrin from Sorghum bicolor to Arabidopsis thaliana. As a result, plants were resistant to the flea beetle P. nemorum.

Although the option of transgenic herbivore resistance is possible, the chance of using cyanogenic glycoside resistance in canola would be minimal. Although levels of cyanogenic

37 glycosides could be regulated to be non-toxic for animal and human consumption, the public perception of adding a poison to a plant may be a tough public relations battle.

1.3.3. Alkaloids

Alkaloids are a class of secondary metabolites consisting of cyclic nitrogen-containing compounds with limited distribution among living organisms (Schoonhoven et al., 2005). Pyrrolizidine alkaloids, derived from amino acids, are strong feeding deterrents for most herbivores particularly generalists (Hartmann, 1999). This class of secondary metabolites includes well- known substances such as nicotine, caffeine, morphine, colchicine and strychnine (Schoonhoven et al., 2005). Alkaloids are produced in about 20% of angiosperms, therefore, as a possible flea beetle defense mechanism in Brassicas, they are not well studied. Brassica alkaloids, in the form of sinapine, are anti-nutritional compounds found only in seeds of canola and prevent the use of the seed protein for food and feed (Mailer et al., 2008). As well as glucosinolates, breeders are trying also to reduce sinapine in canola. However, with limited genetic variation, transgenic means of reduction are being pursued instead (Nair et al., 2000; Milkowski and Strack, 2010).

1.3.4. Terpenoids

Units of isoprene linked together in various ways with different types of ring closure make up the groups of terpenoids. Terpenoids vary in saturation and functional groups and are synthesized in one of two pathways (Schoonhoven et al., 2005). Terpenes and terpenoids are the primary constituents of the essential oils of many types of plants and flowers, particularly associated with the smell of coniferous trees, citronella and menthol to name a few. However, terpenoids form

38 another group of with possible defense functions. Various terpenoid classes, including latex, abeitic acid and limonoids, act as strong feeding deterrents (Schoonhoven et al., 2005).

Saponins are also part of the terpenoid group. They have been shown to interfere with insect growth and development (Schoonhoven et al., 2005; Nielsen et al., 2010b) as well as possessing antifungal properties (Osbourn, 1996a). More recently, saponins have been discovered in

Brassicas where they were not previously found. Nielsen et al. (2010b) report possible resistance/susceptibility of the flea beetle P. nemorum to the Brassica host Barbarea vulgaris may be associated with the role of saponin as a defense compound. (Kuzina et al., 2011) identified quantitative trait loci (QTL) for saponins in relation to flea beetle resistance, further strengthening the possible defensive role against flea beetles.

It has been determined by others, that the defense role of saponins is closely linked to a carbohydrate moiety (Osbourn, 1996a; b; Adel et al., 2000). As a result, alteration of the individual sugars may alter the defensive role. Dethier (1980) warns that the stimulus for sugar receptor on insects may be something other than carbohydrates but perhaps a specific amino acid. Nielsen et al. (2010) determined that hederagenin cellobioside was both the most abundant and the most active compound in the defense against flea beetles in Barbarea. Nielsen et al.

(2001) points out those results seen with P. nemorum are not necessarily those found with P. cruciferae.

In Brassicas, terpenoids may play an important role as an aggregation stimuli (Bartelt et al., 2001).

Their studies determined male-specific sesquiterpenes were a male pheromone. This aggregation pheromone attracts both sexes when the males are feeding on the host plants (Peng et al., 1999). These sesquiterpenes were found in several flea beetle species, however, the proportions were species specific suggesting the pheromonal role. Being sensed by the antennae of both sexes, this enabled the flea beetles to locate nearby like species for feeding and mating purposes (Bartelt et al., 2001).

1.3.5. Phenolics

Another class of compounds often involved with plant-insect interactions is phenolics. They consist of an aromatic ring with one or more hydroxyl groups, together with a number of other constituents (Schoonhoven et al., 2005). Phenolics are found within all plant species and include tannins, phenolic acid, flavonoids (the largest group of phenolics), anthocycanins, silicas and lignans (Schoonhoven et al., 2005; Despres et al., 2007). Both deterrent and stimuli properties can be found in phenolics which include one of the most potent insect deterrents; phaseolin (van

Loon, 1990; Onyilagha et al., 2004; Schoonhoven et al., 2005). Most phenolic studies involving insect-plant interactions deal with tree species. There is vast information regarding phenolics in plants yet there is very little done on plant-insect relationship especially with flea beetles and

Brassicas. Work by Nielsen et al. (1979) concluded that the flea beetle P. armoraciae used both glucosinolates and flavonoids for selection of its host plant. Together, these compounds are more stimulating than either compound alone.

1.3.6. Other plant hormones, amino acids, proteinase inhibitors, etc.

Several other plant chemicals and protein-based mechanisms can also be used by plants in the fight against herbivores. These include signaling compounds such as jasmonates (jasmonic acid), ethylene and salicylic acid as well as proteinase inhibitors, amino acids, peptides and sugars, all of which may be involved directly or indirectly in plant defense.

1.3.6.1. Signaling compounds

Signaling compounds have been reported by Mewis et al. (2006) to have primary roles in plant defense against chewing insects. The common theme of complex interactions of traits continues as not all signaling compounds work together (Mewis et al., 2006). Liang et al. (2006) and Mewis et al. (2006) concluded that jasmonate, salicylate, and ethylene signaling are involved with defense response in mutants of A. thaliana. However, depending on the environment, ethylene and salicylate signaling appeared to work in opposition to jasmonate. A similar conclusion was previously reached by Thaler et al. (1999). Beckers and Spoel (2006) suggest that jasmonic acid and salicylic acid are not independent from each other but work together, possibly with other compounds, in response to an encountered attack. Thaler et al. (2001) later showed jasmonate induced proteins were unfavourable for flea beetles. Jasmonates have also been associated with increased indole glucosinolates and therefore increased flea beetle resistance (Bartlet et al.,

1999). Bodnaryk and Rymerson (1994) induced plant defenses greatly using jasmonates resulting in noticeably altered cotyledon physiology.

1.3.6.2. Proteinase inhibitors

Another defense mechanism deployed by Brassica species is the use of proteinase inhibitors.

Proteinase inhibitors function by interfering with the proteolysis of proteins in the insect’s gut, ultimately reducing insect performance. (Broadway and Missurelli, 1990; Cipollini and Bergelson,

2000, 2001). Ingested protease inhibitors block protease activity and increase insect mortality by restricting the availability of essential amino acids (Dunse et al., 2010). In plants, proteinase inhibitors can be found in seeds, leaves, roots and tubers of plants (Broadway, 1989). Although not well characterized in Brassica species, proteinase inhibitors are environmentally and/or developmentally regulated in these tissues (Broadway and Missurelli, 1990). Proteinase inhibitors are constitutively produced and developmentally regulated as well as wound-inducible in Brassica napus and B. oleraceae (Bodnaryk and Rymerson, 1994).

Once again, certain herbivores have developed counter adaptations to be less affected by proteinase inhibitors (Broadway, 1995). Cipollini et al. (2003) was able to show variation in proteinase inhibitors in experiments where glucosinolate levels were constant. Serine proteinase inhibitors have been found at relatively high levels in many Brassica species (Broadway, 1989).

Cipollini et al. (2003) concluded that control and induction of levels of trypsin inhibitors varied genetically and, therefore, had the capacity to respond to future selection imposed by herbivores. Most research with proteinase inhibitors is done with hormonal or mechanical injury mimicking injury by Lepidopteron insects and may not be applicable to flea beetles. Cipollini and

Bergelson (2000) make note of the lack of studies done with proteinase inhibitors in Brassica species is due to the effort in understanding glucosinolates, especially in the case of flea beetles.

1.3.7. Herbivore mechanisms of handling secondary metabolites

Secondary metabolites such as glucosinolates have been shown to be a defense mechanism for generalist species and stimuli for specialist species (Giamoustaris and Mithen, 1995; van der

Meijden, 1996). For those that prefer these toxin-containing plants as their host, they must be able to handle the toxins somehow. Several different methods have been used by different insects for glucosinolates which enzymatic detoxification, excretion, sequestration and behaviour adaptation (Mainguet et al., 2000; Hopkins et al., 2009). In the specialist Plutella xylostella, the sulphate is removed from the glucosinolate rendering it inactive to the myrosinase (Ratzka et al.,

2002). Within the insect gut, glucosinolate sulphatase competes with myrosinase for glucosinolate substrates. As a result, myrosinase competes with glucosinolate sulphatase for glucosinolates by converting glucosinolates to an unusable form of which the sulphate by- product inhibits myrosinase (Ettlinger et al., 1961; Bones and Rossiter, 2006) therefore decreasing the activity level of myrosinase due to the reduced available substrate. Sulphatase has a broad range of action on a wide range of glucosinolates and is therefore a highly effective mechanism (Hopkins et al., 2009). Ratzka et al. (2002) also discovered in the specialist Pieris rapae a nitrile-specifier protein which converts aglycone to nitrile. This nitrile-specifier protein is not related to the plant epithiospecifier protein and allows the nitrile to be excreted with the feces

(Wittstock et al., 2004).

A unique myrosinase-glucosinolate system has evolved in the specialist cruciferous aphids

Brevicoryne brassicae and Lipaphis erysimmi and the sawfly Athalia rosae that plays an important

43 counter-defense function by sequestering the glucosinolates into microcrystalline (Bridges et al.,

2002; Husebye et al., 2005) or later excreted as one or more unidentified metabolites (Müller and Wittstock, 2005). Green peach aphids handle glucosinolates in yet another manner. They allow aliphatic glucosinolates to pass through the aphid gut intact, but indole glucosinolates are mostly degraded. Mixed function oxidases may also play a role in triggering detoxifying enzymes to handle the toxic metabolites (Brattsten et al., 1977; Ahmad, 1982). New omics technologies may allow better studies of how insects handle their food intake (Spit et al., 2012). In flea beetles, there is no reported documentation of how they handle the glucosinolates.

Insects may seek sub-optimal high anti-herbivory/nutritional plants even though they have negative growth effects. This behaviour may allow the herbivore a safer environment as it reduces natural enemy attack, reduced competition and provides access to plant parts with better nutritional offerings later (Hopkins et al., 2009).

1.3.8. Herbivore mechanisms of handling morphological defense

Recent advances in the battle for resistance to flea beetles in Brassicas have a connection to pubescence (Agerbirk et al., 2003; Gruber et al., 2006; Kuzina et al., 2011; Soroka et al., 2011).

Trichomes not only interfere with host plant selection by preventing the herbivore from locating a suitable spot but also may make it more vulnerable to natural enemies. Glandular trichomes also have the advantage of secreting toxic substances when disturbed, although this may not be beneficial in Brassicas for flea beetles as noted by Traw and Dawson (2002a; 2002b). Although proving promising defense mechanisms, the applicability to commercial canola crops may be

44 limiting as it only affects the leaf and stems but not the cotyledon. Work by Agerbirk et al. (2003),

Kuzina et al. (2011) and Nielsen et al. (2010) with a leaf mining Brassica flea beetle may be applicable to that of P. cruciferae and P. striolata but care must be taken as both P. cruciferae and P. striolata have soil-based larval stages and the mechanisms may be different as a result.

Also, the hairy canola work by Gruber et al. (2006) and Soroka et al. (2011) have not been able to enhance cotyledon protection. Cotyledons are the most susceptible growth stage so seed treatment would still be needed. Hairy canola would reduce or eliminate the need for additional protection once the seed treatment is exhausted.

1.4. IDENTIFYING THE MOST PROMISING PLANT CHEMICAL OR PHYSICAL FEATURE TO INVESTIGATE IN THE DEVELOPMENT OF FLEA BEETLE RESISTANCE IN CANOLA.

A large diversity of physical and chemical defenses impacting the physiology of insect herbivores is used in a back and forth battle between insects and plants where both plants and herbivores expend numerous costs for a better chance of survival (Simms and Rausher, 1987). As a result, a large number of defenses with additive or synergistic effects are used by plants to help protect from herbivory (Cipollini et al., 2003). Not only are the defense mechanisms used by plants not fully understood, neither is the insect-plant selection behaviour. Numerous theories have been developed in an attempt to explain how non-host plants interfere with host plant finding. Finch and Collier (2000) summarized these theories, which include physical or visual disruption of the host plant architecture, chemical masking of host plant odours or production of chemical repellants, influencing host plant physiology and Root's (1973) resource concentration hypothesis and natural enemies theories. Agrawal (2011) hypothesized that evolution of such

45 complex interactions were the result of a diverse set of herbivores requiring a diverse set of defense mechanisms, protection of survival mechanisms in both herbivores and plants and finally, their synergistic impacts. Matters are complicated further in the field as early induction of defense mechanisms by one species may reduce herbivory by another species at a later date

(Agrawal, 1998).

1.4.1. Current works of interest

Recent works of interest involve P. nemorum and Barbarea vulgaris (Badenes-Perez et al., 2014;

Christensen et al., 2014; Heimes et al., 2015). The larval stage of P. nemorum is a leafminer whereas most flea beetles have larval stages in the soil, where little work has been done due to the difficulty of making accurate observations. Renwick (2002) suggested possible flea beetle resistance may be possible to obtain from Barbarea vulgaris, wintercress. This species has shown resistance to the diamondback moth (Plutella xylostella) in North America and to the flea beetle Phyllotreta nemorum in Europe. The garlic mustard, Alliaria petiolata, has shown resistance in the USA to Pieris napi oleracea. The combination of a unique butenenitrile glycoside and a flavone glycoside appears to be the cause of resistance in garlic mustard which has evolved with the presence and pressure from the local insect species (Renwick, 2002; Bruce,

2014).

From an evolutionary standpoint, specialist herbivores may have evolved after generalist and omnivorous insects thus reducing the number of genes responsible for adapting to host plant deterrents (Dethier, 1954; Bernays, 1998). A limited number of genes may be favourable for

46 plant breeding purposes, yet the breeding for specialist resistance has not been so productive.

Resistance in P. nemorum in Barbarea vulgaris has been determined to be based on two sex- linked polymorphisms expressed by the pubescence of the host plant’s leaves (Nielsen, 1997).

Non-transgenic insect resistance has been accomplished in several crops such as wheat (wheat midge), barley (greenbug) and sorghum (greenbug and midge). In wheat, midge resistance was developed via oviposition deterrence and reduced hatch (Lamb et al., 2016). Greenbug resistance in barley is simply inherited by a dominant resistant gene Rsg1 (Azhaguvel et al.,

2014). QTL for greenbug resistance in sorghum have been linked with bio-type specific resistance or tolerance traits (Agrama et al., 2002). Resistance to flea beetles in canola may be possible with the genetic diversity within the Brassicas (Gavloski et al., 2000).

1.4.2. Transgenic or molecular-based resistance

A transgenic approach for flea beetle resistance in Brassicas has been reported as successful. The genes involved include proteinase inhibitors (Broadway and Missurelli, 1990; Girard et al., 1998), cyanogenic glycosides (Nielsen et al., 2001; Tattersall et al., 2001; Kristensen et al., 2005) as well as trichome upregulation (Gruber et al., 2006; Soroka et al., 2011). These genetic constructs have been derived from other plant sources as opposed to the bacteria-based Bt construct, which is common in other insect defense transgenic approaches (Stewart Jr et al., 1996; Christou et al.,

2006). Different proteinase inhibitors are available allowing for manipulation of the regulation and ratio of the different proteinase inhibitors (Hilder et al., 1987). Effective herbivore control may, therefore, require co-expression of multiple proteinase inhibitors and could be targeted to

47 particular herbivores without harming another. No matter which mechanism of defense is enhanced, one must be careful not to create an unfavourable product. For example, reduction of the hydrolysis of glucosinolates rather than lowering concentrations of glucosinolates may affect the suitability of the meal for ruminants (Borgen et al., 2010).

The defense mechanisms of plants involve many pathways and compounds, which is a complex process (Figure 1.3.). Today’s technology allows researchers to examine these processes closer and modify individual pathways to enhance or remove a compound. Genome editing tools such as TALEN, zinc finger nucleases and CRISPR/Cas9 may be used to manipulate these pathways without transgenes and address social concerns surrounding the use of transgenic traits (Kim and

Kim, 2014; Chandrasegaran and Carroll, 2016). Beyond that, the so-called ‘-omics’ technology applications (genomics, transcriptomics, proteomics, metabolomics, physionomics, phenomics)

(Borem and Fritsche-Neto, 2014; Van Emon, 2016) on both the insect and host may prove to be valuable tools in determining what mechanism is triggered, when and where (Boerjan et al.,

2012; Spit et al., 2012). By further understanding the interactions of the host plant and the herbivore, determining which products or pathways are up-regulated or down-regulated may indicate which genes are involved, when and how to regulate them.

An area that is lacking in the study of flea beetle resistance involves the mechanisms for handling toxic metabolites such as glucosinolates. Beran et al. (2014) have determined that P. striolata sequesters intact glucosinolates from host plants but uses its own myrosinase to breakdown glucosinolates into degradation products. The larval stages of numerous Lepidoptera species

48 have been studied and mechanisms are well understood. Perhaps due to the small larvae size and their soil bound nature, this work would be challenging but likely beneficial. Use of model species has its limits and may not be transferable as noted by Nielsen et al. (2001).

A better understanding of the pathways of seed desiccation and germination for improving resistance as defensive traits, without impact on the quality of the seed is needed. A number of current resistance strategies focus on true leaves whereas cotyledons are as susceptible to herbivory as well. Pathways and metabolites may need to be altered at the seed development stage. Early seedling vigour is the other important area for improving flea beetle resistance.

Enhanced seedling vigour would involve cold and moisture stress adaptations to allow the seedling to outgrow the herbivore damage in most cases without chemical control.

1.5. SUMMARY

With all the research done on flea beetles, improved flea beetle control has been made and further improvements are possible. The widespread distribution of this herbivore across a range of Brassicas important to human consumption makes flea beetle research widely applicable and worthwhile. Enhanced genetic flea beetle resistance within Brassicas is being studied and developed. It will be a matter of time before these improvements, whether traditional or transgenic in origin, become commercially viable. In the meantime, improvements in agronomic practices, including chemical control, continue to be the best options.

Due to the complexity of the interactions of so many physical and chemical stimuli and defense mechanisms, developing a “silver bullet” for plant resistance to herbivory is highly unlikely. As seen in Figure 1.4., when a plant is stressed whether by biotic or abiotic means, multiple defense mechanisms are activated. These mechanisms may be a defensive reaction in some cases and a stimulus in others. All indications point to a multi-faceted defense mechanism for increased resistance to a number of herbivorous insects; flea beetles being one. Of course, a transgenic approach may make the task easier, however, it may inadvertently allow for a new pest to take advantage.

In summary, today’s technology can provide further insight into host-insect relationships, the pathways involved in both the host and the herbivore and eventually identify genes that could be manipulated to confer plant resistance to flea beetle herbivory. While many studies investigated individual components of host-insect relationship, the complex interactions amongst these components are very difficult, if not impossible, to replicate in an artificial environment. Using genetic based techniques, such as identification of markers associated with

QTL, one may be able to locate genomic regions associated with many interaction factors. Using a combined genomics and metabolomics approach in their quest to understand the interactions between winter cress (Barbarea vulgaris) and flea beetle (Phyllotreta nemorum), Kuzina et al.

(2011) reported deviating host plant types, which differed in flea beetle resistance, saponin and glucosinolate profiles, as well as leaf pubescence. A genetics-based approach is taken in this dissertation. Once QTL are identified, the next step would be to incorporate them into a breeding

50 program. Determining what pathway or defense mechanism they are involved in would be beneficial to the scientific community.

1.6. HYPOTHESIS

The research presented in this thesis was based on two hypotheses; first, that flea beetle herbivory does not vary between spring-type and winter-type canola, and secondly, that resistance to flea herbivory is under the control of multi-genic mechanisms. The objectives of this thesis were to: 1) investigate whether or not flea beetle herbivory differed between spring-type and winter-type canola; 2) determine if flea beetle had different feeding preferences as newly emerged adults compared to adults preparing to overwinter; and 3) identify SSR and SNP markers that tag QTL for resistance to flea beetle herbivory.

a. a. b. b.

Figure 1.2. Life cycle of the crucifer flea beetle (Phyllotreta cruciferae (Goeze)). (diagram adapted from Knodel and Olson (2002)).

CHAPTER 2 - EVALUATION OF WINTER-TYPE BRASSICA NAPUS GERMPLASM FOR GENETIC DIVERSITY IN RESPONSE TO FLEA BEETLE HERBIVORY IN TYPICAL AND ATYPICAL PLANTING WINDOWS

2.1. ABSTRACT

Current practices of flea beetle (Phyllotreta spp.) control in Brassica napus rely heavily on seed treatments. With growing concerns regarding safety of seed treatments on non-target and beneficial insect populations, genetic resistance would be beneficial for a more balanced integrated pest management strategy. Currently, none of the registered Brassica napus L. canola varieties exhibit measurable resistance to flea beetle injury. Young plants of 14 winter-type B. napus breeding lines were evaluated for resistance to flea beetle feedings. These 14 lines came from four breeding families where at least one of these lines exhibited noticeably reduced damage compared to sister lines in a breeding nursery. Subsequent studies indicated natural genetic variation within B. napus for flea beetle antixenosis, which could contribute to the development of canola varieties with high levels of flea beetle resistance under a dedicated breeding strategy. Furthermore, data indicated that host plant resistance did not vary between newly emerged adult flea beetles in the fall and the overwintered adults in the spring, suggesting that favourable genes identified in winter-type material could be used to confer resistance on spring-type material and vice versa even though the flea beetle life cycle is different for each habit’s critical young development stage. Data also suggested adult feeding preferences were not unique when newly emerged adults were compared to overwintered adults and therefore the adults of either type can used to evaluate feeding damage.

2.2 INTRODUCTION

Approximately 8.3 million hectares (20 million acres) of canola are grown each year in Western

Canada (Canola Council of Canada, 2016) producing about 18.4 million tonnes of grain annually in 2016 (Canola Council of Canada, 2016). The western Canadian canola crop is spring planted, usually from late April to early June. It is during this time that flea beetles, predominantly

Phyllotreta cruciferae (Goeze) and Phyllotreta striolata (F.) (Coleoptera: Chrysomelidae), emerge from overwintering sites in surrounding field grasses, hedgerows and bushes in search for food and a place to lay their eggs (Burgess, 1977a). Canola is susceptible to flea beetle feeding from cotyledon emergence from the soil until the four leaf stage (BBCH=14 (Lancaster et al, 1991)), at which point, plants can usually tolerate damage (Gavloski and Lamb, 2000; Knodel and Olsen,

2002). Damage to the cotyledons and first true leaves can reduce yields by up to 10% (Lamb and

Turnock, 1982). Damage can also delay plant development and cause uneven plant height and delayed maturity as well as affect the chlorophyll content (Bodnaryk and Lamb, 1991;

Lamb,1982). A yield reduction of approximately 5% is estimated to result from larval feeding on roots with larval densities of 0. 16/cm2 soil surface area (Bracken and Bucher, 1984). Results of previous studies indicate that above and below ground plant resistance is needed to minimize yield loss due to flea beetle herbivory. Flea beetles may cause a total economic impact exceeding

CA$1 billion annually in North America through yield loss and costs of mitigated action against flea beetles (Lamb and Turnock, 1982; FAOSTAT, 2017; Winnipeg Exchange, February futures

2017).

There are over 4000 species of flea beetles worldwide affecting numerous plant species from vegetables, field crops and weeds (Konstantinov and Vandenberg, 2010). One generation per year is common in the Canadian prairies (Knodel and Olsen, 2002), although multiple generations have been reported in some places such as Ontario (Kinoshita et al., 1979). In early spring, the flea beetles emerge as overwintered adults in search of food and lay eggs at the base of the plants

(Figure 1.2.). The eggs develop into larvae during the summer and feed on the roots, before emerging as new adults later in the summer. Late season feeding is usually not considered to impact yield, however, at high densities, feeding on the outer layer of seedpods and other green plant tissue can result in early maturity, pod shatter and high green seed content (Burgess, 1977;

Knodel and Olsen, 2002).

The FAO (Food and Agriculture Organization of the United Nations) defines integrated pest management as:

“Integrated Pest Management (IPM) means the careful consideration of all available pest control techniques and subsequent integration of appropriate measures that discourage the development of pest populations and keep pesticides and other interventions to levels that are economically justified and reduce or minimize risks to human health and the environment. IPM emphasizes the growth of a healthy crop with the least possible disruption to agro-ecosystems and encourages natural pest control mechanisms (FAO, 2015).”

A good IPM program would include several control tactics such as physical, cultural, chemical, biological and genetics, to reduce flea beetle damage. These tactics could include avoidance

(agronomic practices), reduction (genetic tolerance or resistance and chemical control) and pest removal (chemical control). As there are no significant flea beetle resistant or tolerant varieties of canola available (Gavloski et al., 2000), most canola varieties are planted with a seed treatment, primarily of the neonicotinoid class. Significant losses due to flea beetles are reported

57 when a suitable seed treatment and/or foliar application is not applied (Knodel et al., 2008).

Unfortunately, seed treatments are only effective as long as the target pest does not overcome the insecticide’s mode of action. In the case of flea beetles, a shift in population make-up is occurring as a result of neonicotinoid pesticide use, where P. striolata is becoming more prevalent over P. cruciferae (Tansey et al., 2009). Agronomic practices such as tillage, crop rotations, early plantings, row spacing, higher seeding rates and larger seed size are all techniques that could be applied to reduce flea beetle injury (Dosdall et al., 1999; Dosdall and

Stevenson, 2005; Elliott et al., 2008).

Within the Brassicaceae, a number of studies have been conducted to determine if there is natural genetic variation present for flea beetle resistance (Gavloski et al., 2000; Lamb, 1988;

Lamb et al., 1993). Various levels of resistance within Brassicas (inter-specific and intra-specific) have been observed and shown to be genetic and heritable (Brandt and Lamb, 1993; Putnam,

1977). These studies have shown the difficulty of studying resistance to flea beetles and indicate resistance is present within the Brassica species but not at a commercially effective level.

Transgenic options that have been investigated include Bt toxins (Payne and Michaels, 1995) and increased trichome density (Gruber et al., 2006) but are not readily available. However, native genes resistance is preferred due to political and social considerations in some areas that are not accepting of genetically modified organisms along with the high cost of deregulation.

The objective of this chapter is to evaluate 14 winter-type doubled haploid Brassica napus canola breeding lines for flea beetle resistance under field conditions and determine whether resistance

58 is affected by seasonal differences between May and late August plantings. By planting twice in a growing season, it may be possible to determine whether feeding habits changed according to adult age by exposing young canola to overwintered adults or newly emerged adults. This study provides insight into flea beetle feeding habits, the potential to move resistance genes between winter and spring habit Brassica spp., and confirms previous reports of genetic variation within

Brassica species.

2.3. MATERIALS AND METHODS

2.3.1. Plant material

The genotypes used in this study were winter-type B. napus doubled haploid (DH) experimental lines developed by the University of Guelph. These lines were initially selected from a previous breeding experiment where the seed treatment failed and responses to flea beetle herbivory were documented. Prior to this study, initial breeding varieties were formally screened to reduce and finalize the number of entries for this flea beetle injury response study (data not shown).

Pedigrees of this selected material resulted in four families where three families have three entries of varied response to flea beetle and a fourth family with five entries. A fifth family comprised of check varieties and included the B. napus variety 46A65 (a previous open- pollinated commercial canola variety grown during the mid to late nineties in Western Canada) as well as key varieties from related species adapted to Western Canada, specifically: Ace

(Sinapis alba), Cutlass (B. juncea) and Tobin (B. rapa). Previous work by Gavloski et al. (2000) and

Palaniswamy et al. (1997) indicated varying levels of resistance to flea beetle feedings amongst different Brassica species, hence the inclusion of representative cultivars in this study as checks;

59 where S. alba and 46A65 with seed treatment were considered more resistant than the other three Brassica spp.. With the exception of 46W09, only spring-type (non-vernalizing) checks were used as they were readily available and would serve as common checks to the paired spring-type experiment (see Chapter 3).

2.3.2. Test sites and experimental designs

Over the course of this study, 11 trials were planted in three locations; Alloa, ON (43.697050,

-79.870154) and Belfountain, ON (43.791724, -80.000064) and Edmonton, AB (53.430290,

-113.541054), from 2009 to 2011. Edmonton was selected in order to get more exposure to the striped flea beetle (P. striolata), although sticky trap captures on site indicate the population was predominantly composed of crucifer flea beetle (P. cruciferae), as in Ontario (data not shown).

Trials were planted in May and late August to early September, two planting dates seven to ten days apart were planted for each trial in order to coincide with periods of abundant natural flea beetle populations.

In the spring of 2009, the trials were conducted as split plot experiments. The whole plots were completely randomized and were comprised of two seed treatments - with or without Helix®

(10.3% thiamethoxam, 1.24% difenoconazole, 0.39% metalaxyl-M, 0.13% fludioxonil - registration number 26637 – Syngenta Crop Protection Canada Inc, Guelph, ON) applied at a rate of 15mL kg-1. Whole plots were replicated three times. Within each whole plot, 24 genotypes

(subplots) were randomized. The field layout in spring 2009 consisted of two treatments, three replications, with each replicate consisting of two ranges by 12 rows; overall four ranges by 36

60 rows. The fall planting in 2009 was four ranges by six rows wide per replication, with replicates side by side. All trials in 2010 and 2011 were conducted as RCBD with three replicates where each replicate was arranged as two ranges by 10 rows and had the same entries in both years.

Replicates were side-by-side. Table 2.1. shows locations, experimental design, number of entries, planting date and data collected.

All locations were prepared using conventional agronomic practices. Prior to planting, all sites were tilled and incorporated with pre-emergent herbicide and fertilizer. All sites were planted with a Hege 1000 nursery planter (Wintersteiger, Salt Lake City, UT) with double disc openers.

Seeding rate was 120 seeds per row with row length of 4.5m trimmed to three metres. Row spacing in 2009 was 30cm, and 50cm in years 2010 and 2011. Attempts were made for two plantings of each trial approximately seven to ten days apart. For each pair of planting dates, the trial with the best visual appearance for discriminating flea beetle herbivory responses was evaluated, except in Edmonton 2010 and Alloa Fall 2010 where both sets were collected.

2.3.3. Phenotypic data - scoring of plants for flea beetle herbivory

Plants were scored for flea beetle herbivory at 14 days post planting and when possible 21 days post planting as well. A visual score was noted using a scale from 0 to 9 where nine was no feeding and one was plant death/no tissue remaining. In 2009 and 2011 plus part of 2010, average scores were assigned. In 2010, 15 single plants were scored individually and averaged for a row score.

For 21d scores in fall 2009 and spring 2011, cotyledons were scored separately from the true leaves. Visual keys (Figs. 2.1. and 2.2.), based on percentage of feeding, were used in determining rating scores. These percentages converted to a 1 to 9 scale by increments of 10% as follows: less

61 than 10% damage was a score of 9, 10-20% damage was a score of 8 and so on until more than

80% damage scored a value of 1. Plant death was scored as zero; where stem bites resulting in plant death were included here and not separated from cotyledon surface damage. When entire rows were assessed, the average score was estimated; for example, if all plants appeared to be a five (40% to 50% damage), a score of 5 was assigned to it. If plants varied evenly between seven

(20% and 30% damage) and three (60% to 70% damage), a score of 5 was also assigned.

2.3.4. Statistical analysis

The response variable (flea beetle feedings at 14d and 21d) was analyzed using PROC GLM (SAS

Institute, 2012) and data were tested for normality using PROC UNIVARIATE (SAS Institute, 2012).

Homogeneity of variance was examined via the residuals using PROC REG (SAS Institute, 2012).

Where single plants were scored, row averages (calculated using Microsoft Excel®) were used in the analysis. For the combined analysis, initially, only trials with acceptable homogeneity of variance and C.V.s less than or equal to 25% were included, however, all trials were ultimately included in analysis as results were not affected greatly. The untreated main blocks from the split plot design in spring 2009 were taken and combined with the other RCBD trials which did not have seed treatment applied. Statistical analyses included ANOVA, and MEANS comparisons using PROC GLM (SAS Institute, 2012) and R (R Core Team, 2015). Pair-wise genotype comparisons were examined using LSD at = 0.05.

2.4. RESULTS AND DISCUSSION

2.4.1. Results

Average flea beetle herbivory scores for each entry at 14 and 21 days after planting are presented in Tables 2.2. and 2.3., respectively. Approximately one-third of the locations had overall low flea beetle pressure (Flea Beetle score (FBSC): 7-9), half had moderate flea beetle pressure (FBSC: 4-

7) and two locations (2010-Alloa, Spring and Fall plantings) had very high pressure (FBSC: less than 4). Tables 2.2. and 2.3. also show the ranking of the experimental entries. With the exception of B. juncea, the checks were similarly positioned by rank in 14 days vs 21 days post-planting.

There were some changes in ranks of experimental entries between 14 days and 21 days post- planting for flea beetle injury where 1147-01 switched 1147-02 rank order, 1147-06 improved with time). There was less than 1.5 units difference between the means of the entries with the least feeding and those with the greatest feeding pressure.

Table 2.4. show the statistical values for the split plot field experiments in 2009, 14 and 21 days after planting at two locations. At Alloa and Belfountain, both 14 days and 21 days post planting indicated significant ( = 0.05) differences for seed treatment, replicates and entry. There were significant interactions ( = 0.05) observed between entry and treatment at Belfountain 14 days only.

Table 2.5. show the statistical values for the combined nine field experiments in 2009 to 2011 at

14 days post planting. In Edmonton 2010 and Alloa Fall 2010 where both planting were scored, only the better one based on means, and C.V. were included, otherwise, all experiments were

63 included. All main factors and 2-way interactions had significant values (=0.05). Three-way and four-way interactions were not significant and dropped from the model.

There appeared to be significant differences for all main factors and 2-way interactions. This can be seen in Figures 2.3., 2.4., and Table 2.5.. More complex interactions were not significant and removed from the model. By family, family three was more susceptible (mean=5.79) than the other three families whereas family 1 was more resistant (mean=6.23) overall with respect to flea beetle feeding. Check S. alba was significantly more resistant than family 3 (mean=5.79;

=0.05) and B. juncea-Cutlass (mean=5.65; =0.1) (Figure 2.5.).

Other observations noted within the trials included less feeding damage on plants that were clustered compared to ones more evenly dispersed. Feedings were worse when row consisted of few plants or large gaps with few plants. Also noted were changes in responses to flea beetle damage. B. juncea had a different reaction to flea beetle herbivory. Necrosis appeared surrounding the feeding holes left by the flea beetle. This became more prominent as the plants grew. As other highly damaged Brassicas appeared to senesce the leaves, B. juncea appeared to hold onto the leaves but contained the damage via localized necrosis. Damaged cotyledons senesced more readily than non-damaged as noted in 2009 experiments where seed treatments limited herbivory compared to the undamaged.

2.4.2. Discussion

Flea beetle studies have proven to be challenging to conduct in the field (Lamb, 1988; Lamb et al., 1993) which was confirmed in this study. The study was conducted over three years, in two

Canadian provinces for a total of 11 experiments. Two plantings were attempted in most cases, however, various factors such as low insect pressure, poor weather conditions, etc. did not allow for a second planting or notes collected in all cases. When a second planting was available, only the visually better of the two experiments was evaluated in this study. Natural population pressure was greatest in 2010 (Figure 2.3.) but resulted in more variability in feeding scores as indicated by the greater C.V. scores (Tables 2.2. and 2.3.). Under extreme insect population numbers, whether too little or too much herbivore pressure, discrimination amongst entries for flea beetle damage was difficult. The narrow spectrum of resistance being evaluated makes evaluations difficult with varied herbivore pressure; further adding to the complexity. With lower insect pressure, flea beetle feedings were low and very few differences were noted as reported earlier (Lamb, 1988). The combined data set resulted in significant differences in year, location, replicate and entry. Significant year and location effect indicated that flea beetle resistance in plants was highly influenced by the environment. The balance between ideal insect pressure and high-quality phenotype evaluations are challenging under natural population conditions, especially when there is a narrow range of variation to observe. Accurate phenotyping and experimental design becomes more important in analysis as to reduce error. With small differences, low statistical power becomes a concern (Button et al., 2013).

This study included 14 winter-type B. napus genotypes in addition to several checks; one representative variety each of B. rapa, B. juncea and S. alba. Various levels of resistance within

Brassicaceace [inter-specific (non-B. napus) and intra-specific (B. napus)] have been documented by others (Brandt and Lamb, 1993; Gavloski et al., 2000; Putnam, 1977). As expected, treated check 46A65 and S. alba were among the more resistant whereas check B. juncea was more susceptible. Greatest differentiation appeared amongst Brassica species and as a result, three non-B. napus checks along with one B. napus (treated and untreated with seed treatment) were included to validate findings from Lamb (1988) and Palaniswamy and Lamb (1992), and determine the level of resistance evaluated to that previously reported. In this study, greater differences were noted for inter-specific than intraspecific comparisons. Lamb et al. (1993) also noted high variability issues, especially within the row of each genotype when trying to identify de novo antixenosis. This was observed while evaluating the trial in 2009. Clustering of plants had less damage than plants spread further apart. This observation was used to change the data collection in subsequent trials to be on a single plant basis where possible. Palaniswamy and

Lamb (1992) noted that feeding patterns may change from cotyledons to true leaves, which was confirmed in the current study, particularly in the B. juncea check, Cutlass. The cotyledons of B. juncea developed a localized necrotic response to flea beetle damage. As the plant grew, the necrotic surrounding holes were more noticeable on the cotyledons. B. juncea leaves are more pubescent and as a result, may have deterred flea beetles better than without pubescence as the true leaves developed. Pubescence regulation is one strategy that was investigated by Gruber et al. (2006).

This study was paired with spring-type material (as described in Chapter 3.), therefore, material from both projects was planted twice in a growing season; once in early spring and a second in late summer to early fall. For winter-type material, the highly susceptible cotyledon growth stage would occur in late summer to early fall as the flea beetles are preparing for overwintering. This would represent natural insect-host relationship as winter canola is planted in the fall and not the spring. The material was also planted in the spring when spring-type material is typically planted. In this case, flea beetles would be emerging from overwintering sites, looking for food and a location to oviposit, which is atypical timing of expression of resistance for winter-type canola. By comparing the feeding damage in both plantings, it was concluded that host plant preference by flea beetles are not significantly different between newly emerged adults in the fall compared to emerging overwintered adults in the spring (Figure 2.4.). As noted by Lambdon

(2003) and McCloskey and Isman (1995), feeding preferences and plant reactions may change as the plants mature but when the flea beetles are exposed to the same host growth stage, there appears to be no difference, whether spring or fall planted. This may indicate that any attraction or deterrent strategy by the host plant may be growth stage dependent if the feeding preferences by the flea beetles do not vary. A similar pattern was reported in Chapter 3 for spring-type B. napus. As such, the results may suggest spring-type and winter-type B. napus have evolved similarly in regards to the response to flea beetles. Therefore, this also may suggest that there are likely few or no unique mechanisms at play for flea beetle resistance within B. napus.

2.5. CONCLUSION

In this study, fourteen entries from four breeding families represented a very small proportion of the genetic variation within winter-type B. napus. By evaluating of a larger number of lines, greater levels of resistance or tolerance may have been found, as defined by Painter (1951). This study is in agreement with previous reports indicating that inter-specific crosses may be required to transfer resistance genes from related species into the crops of interest. Without seed treatment, none of the B. napus entries evaluated had the resistance level expressed by S. alba.

The mechanisms involved in flea beetle antixenosis in canola is very complex for two reasons; first, two biological entities are interacting at various stages in their respective life cycle, and secondly, as described in Chapter 1, many traits from both the insect and host plant are involved with resistance and host identification. The potential to move genes not only between related

Brassica species but also between different habits of B. napus increases the opportunities to achieve the goal of increased flea beetle resistance (Painter, 1951) in canola for Western Canada.

Variation within a family also indicated that with a large enough population and selection intensity, an improved flea beetle resistant B. napus line could be developed via transgressive segregation over cycles of breeding. However, the strong environmental influence on the trait as exhibited by the highly significant year, location and replicate effects, indicated that care must be taken for experimental design and data collection.

Number Experimental Notes Trial Season Year Location of Planting Date Design Collected Entries 1 Spring 2009 Alloa Split Plot 24 May 25 †, †† 2 Spring 2009 Belfountain Split Plot 24 May 25 †, †† 3 Fall 2009 Belfountain RCBD 24 August 26 †, §, §§ 4 Spring 2010 Edmonton1 RCBD 20 May 18 † 5 Spring 2010 Edmonton2 RCBD 20 June 1 † 6 Spring 2010 Alloa RCBD 20 May 30 †, ‡ 7 Spring 2010 Belfountain RCBD 20 May 19 †, ‡ 8 Fall 2010 Alloa1 RCBD 20 August 18 †, ‡ 9 Fall 2010 Alloa2 RCBD 20 August 28 †, ‡ 10 Fall 2010 Belfountain RCBD 20 August 11 †, ‡ †, §, §§, 11 Spring 2011 Edmonton RCBD 20 May 24 ¶, # † and †† are flea beetle herbivory scores by row 14 days and 21 days post-planting respectively; § and §§ cot and leaf scores separate for 21 days post-planting respectively; ‡ is where the notes were collected as single plants and averaged for a row score, whereas other rows were evaluated as single averaged observation; # is the number of plants per square meter; ¶ is plant height measured from the soil surface to upper most point

2009† 2010 2011

Genotype/ Spring Fall Spring Fall Spring FBSC Overall Family SE Variety ALL BEL BEL ALL BEL EDM1 EDM2 ALL1 ALL2 BEL EDM Mean Rank 1147_01 1 7.67 6.33 8.33 1.36 6.76 4.67 6.33 3.49 6.11 7.51 8.00 6.05 1.05 13 1147_02 1 n/a 5.67 8.33 2.40 6.42 6.67 6.67 4.18 6.18 7.63 9.00 6.27 1.09 5 1147_03 1 n/a 6.33 8.67 2.29 6.04 7.00 7.33 3.65 6.42 7.37 9.00 6.38 1.11 4 1147_04 2 7.67 6.00 8.00 1.71 6.27 5.67 6.67 3.71 6.69 7.10 8.00 6.10 1.06 11¶ 1147_05 2 n/a 6.67 8.33 2.18 4.96 5.67 6.00 3.56 6.27 7.23 8.67 5.91 1.03 15 1147_06 2 8.00 6.33 8.00 1.31 4.64 5.33 6.33 3.16 6.13 7.44 8.67 5.94 1.03 14 1147_07 3 7.67 6.00 8.33 1.29 4.22 5.67 6.33 3.07 5.71 6.73 9.00 5.76 1.00 18 1147_08 3 7.67 6.33 8.33 1.02 5.91 3.67 7.67 3.65 5.16 7.13 8.33 5.82 1.01 16 1147_09 3 n/a 6.00 8.33 0.98 3.84 6.33 7.67 4.04 6.00 7.17 8.00 5.79 1.01 17 1147_10 4 7.00 6.67 8.67 1.02 6.07 6.00 7.00 3.96 5.80 7.40 8.33 6.14 1.07 9¶ 1147_11 4 7.67 6.67 8.00 1.13 5.24 5.33 6.00 2.84 5.13 7.27 8.33 5.68 0.99 19 1147_12 4 8.67 6.67 7.67 1.04 5.56 6.67 6.67 3.82 4.91 7.30 9.00 6.15 1.07 9¶ 1147_13 4 8.67 6.67 7.33 2.07 6.64 5.67 6.33 2.89 5.98 7.43 8.67 6.18 1.08 7¶ 1147_14 4 7.33 6.67 8.00 1.51 4.98 7.33 7.00 4.33 5.24 7.37 8.67 6.19 1.08 6 46W09 5 7.00 6.33 8.00 1.91 5.40 6.00 7.67 4.18 5.93 7.27 8.67 6.18 1.08 7¶ Cutlass(BJ)§ 5 8.67 7.00 8.33 0.91 4.67 6.33 7.00 1.60 2.91 6.73 8.33 5.65 0.98 20 Tobin(BR)§ 5 8.67 7.33 9.00 1.42 5.44 6.67 6.67 n/a n/a 7.07 n/a 6.49 1.13 3 Ace(SA)§ 5 9.00 8.33 9.00 3.91 7.80 6.33 8.00 4.64 6.67 7.17 9.00 7.26 1.26 1 46A65(t)§ 5 n/a n/a 8.75 1.69 5.18 6.00 6.67 5.73 7.24 7.47 9.00 7.14 1.37 2 46A65(u)§ 5 8.17 7.63 8.67 5.73 7.84 6.00 7.00 4.09 6.62 7.40 8.67 6.10 1.17 11¶ FBSC Mean 7.99‡ 6.73‡ 8.21‡ 1.84 5.69 5.95 6.85 3.72 5.85 7.29 8.60 Expt C.V. 8.21 11.59 8.25 22.69 19.09 21.83 15.05 30.07 12.18 5.42 5.59 R2 0.59 0.52 0.34 0.92 0.64 0.39 0.39 0.50 0.75 0.47 0.53 Total Entries 25 25 24 20 20 20 20 20 20 20 20 Pr>F (model) 0.0004 0.0034 0.5533 <.0001 0.0009 0.3444 0.3588 0.0620 <.0001 0.7923 0.0310 Pr>F (rep) 0.0007 0.0007 0.7625 <.0001 0.0049 0.5741 0.0636 0.0904 0.0007 0.2518 0.0113 Pr>F (entry) 0.0029 0.0277 0.4839 <.0001 0.0026 0.2984 0.5430 0.0843 <.0001 0.8626 0.0871 † full experiment was Split Plot with treated and untreated main plots. Data used for this summary is only the untreated RCBD component with all entries. Some entries are not included in table shown but overall statistics include hidden entries. ‡ FBSC mean is average of entries shown on table, entries within experiment not shown on table not included in this

70 calculation. ¶ where rankings were equal, all entries assigned highest rank, next unique rank continued assigned based on rank without unique scores. § BJ=Brassica juncea, BR=Brassica rapa, SA=Sinapis alba, t=seed treatment, u=untreated (no seed treatment).

Table 2.3. Mean scores and summary statistics for flea beetle herbivory 21 days after planting on 15 winter-type canola lines evaluated from 2009 to 2011 in Alloa (ALL) ON, Belfountain (BEL) ON and Edmonton (EDM) AB. Flea beetle score (FBSC): Scale of 0-9 where 0 = plant dead and 9= less than 10% flea damage on cotyledons or first true leaves. 2009† 2011 Spring↨ Fall Spring FBSC Mean Over Genotype/ Family EDM- EDM- (cot/ SE all Variety ALL BEL BEL-cot BEL-leaf cot leaf first Rank score) 1147_01 1 7.33 6.67 7.33 6.33 8.80 8.63 7.53 0.29 5 1147_02 1 n/a 6.67 5.33 7.00 8.80 8.77 6.60 0.57 19 1147_03 1 n/a 6.33 7.33 7.33 8.73 8.57 7.47 0.33 6 1147_04 2 8.33 7.67 4.67 8.00 8.63 8.60 6.74 0.51 14¶ 1147_05 2 n/a 7.67 5.67 7.33 7.60 8.53 6.64 0.37 18 1147_06 2 7.67 n/a 7.67 6.67 8.90 8.27 7.73 0.34 3 1147_07 3 7.00 6.33 4.67 7.67 8.30 8.37 6.66 0.48 17 1147_08 3 8.00 7.00 6.00 7.33 8.73 8.07 7.18 0.39 9 1147_09 3 n/a 6.67 6.67 7.00 8.17 8.67 6.94 0.39 11¶ 1147_10 4 7.00 7.00 6.00 8.00 8.33 8.63 7.00 0.30 10 1147_11 4 7.00 7.33 5.33 8.00 8.10 8.57 6.94 0.49 11¶ 1147_12 4 7.00 7.67 5.00 7.33 8.50 8.73 7.21 0.53 8 1147_13 4 8.00 7.67 3.67 7.00 8.60 8.17 6.90 0.65 13 1147_14 4 8.00 6.00 4.67 7.67 8.30 8.93 6.74 0.50 14¶ 46W09 5 7.00 7.67 8.00 7.67 8.20 8.77 6.72 0.45 16 Cutlass(BJ)§ 5 5.67 4.00 4.00 6.67 8.76 8.37 7.94 0.29 2 Tobin(BR)§ 5 6.33 5.67 5.33 7.33 n/a n/a 6.00 1.06 20 Ace(SA)§ 5 8.33 8.00 6.67 6.33 8.40 8.70 8.43 0.15 1 46A65(t)§ 5 n/a n/a 6.33 7.67 8.97 8.87 7.65 0.61 4 46A65(u)§ 5 7.96 7.41 5.67 7.00 8.73 8.80 7.35 0.44 7 FBSC Mean 7.47‡ 6.99‡ 5.88‡ 7.27‡ 8.08 8.15 Expt CV 3.86 6.31 22.30 11.10 5.85 6.45 R2 0.99 0.98 0.59 0.45 0.96 0.95 Total Entries 25 25 26 26 20 20 Pr>F (model) 0.00 0.0017 0.0013 0.0968 <.0001 <.0001 Pr>F (rep) <.0001 0.0001 0.3320 0.3311 0.0618 0.4937 Pr>F (entry) <.0001 0.0002 0.0010 0.0933 <.0001 <.0001 † 2009 full experiment was Split Plot with treated and untreated main plots. Data used for this summary is only the untreated RCBD component with all entries. Some entries are not included in table shown but overall statistics include hidden entries. ↨ 21d FBSC collected by visual full row evaluation and not separated into cotyledon and true leaf scores. ‡ grand mean is average of entries shown on table, entries within experiment not shown on table not included in this calculation. § BJ=Brassica juncea, BR=Brassica rapa, SA=Sinapis alba, t=seed treatment, u=untreated (no seed treatment).

Table 2.4. Split plot analysis using PROC ANOVA (SAS Institute) of flea beetle herbivory in winter-type canola lines after 14 and 21 days post-planting at Alloa and Belfountain, ON in 2009. Insecticidal seed treatment† was whole plot (with or without). Alloa 14 days post planting 21 days post planting Effect DF Type III SS MS Pr > F DF Type III SS MS Pr > F Total 125 80.30 125 106.15 Rep 2 3.04 1.52 0.0115 2 3.41 1.71 0.0146 Trt† 1 18.67 18.67 <.0001 1 5.89 5.89 0.0002 Error (1) 2 6.05 3.02 2 9.38 4.69 Entry 20 17.43 0.87 0.0009 20 50.64 2.53 <.0001 Trt*Entry 20 9.37 0.47 0.1217 20 6.20 0.31 0.6948 Error (2) 80 25.75 0.32 80 30.62 0.38 R2=0.68 C.V.=6.79 R2=0.71 C.V.=8.14

Belfountain 14 days post planting 21 days post planting Effect DF Type III SS MS Pr > F DF Type III SS MS Pr > F Total 167 229.95 167 198.85 Rep 2 5.25 2.63 0.0146 2 3.87 1.93 0.0243 Trt 1 102.15 102.15 <.0001 1 2.15 2.15 0.0411 Error (1) 2 5.08 2.54 2 13.30 6.65 Entry 24 24.99 1.04 0.0283 24 109.89 4.58 <.0001 Trt*Entry 24 24.23 1.01 0.0361 24 12.22 0.51 0.4582 Error (2) 114 68.25 0.60 114 57.42 0.50 R2=0.70 C.V.=10.29 R2=0.71 C.V.=9.98 † (trt) Helix ® liquid seed treatment is a Syngenta Crop Protection Canada Inc product that contains the insecticides and fungicides thiamethoxam, difenonconazole, metalaxyl-M and S-isomer and fludioxonil

Table 2.5. PROC GLM (SAS Institute) ANOVA of flea beetle herbivory after 14 days post-planting on winter-type canola lines for nine combined experiments conducted at Alloa and Belfountain, ON and Edmonton, AB in the spring and fall of 2009 to 2011. Sum of Mean Source DF Squares Square F-Value Pr>F Model 126 2173.54 17.25 26.74 <.0001 rep(loc) 6 13.44 2.24 3.47 0.0024 location 2 396.53 198.27 307.32 <.0001 season 1 545.85 545.85 846.07 <.0001 location*season 1 99.33 99.33 153.97 <.0001 year 2 763.64 381.82 591.82 <.0001 location*year 1 395.54 395.54 613.09 <.0001 entry 19 36.16 1.90 2.95 <.0001 location*entry 38 70.23 1.85 2.86 <.0001 season*entry 19 40.94 2.15 3.34 <.0001 year*entry 37 43.28 1.17 1.81 0.0034 Error 361 232.90 0.65 Corrected Total 487 2406.44 R2=0.90 C.V.=12.44 Root MSE=0.8032 FBSC Mean=6.45

Figure 2.1. Canola cotyledons with levels of flea beetle damage varying from less than 10% (panel a) to 100% (panel j) damage to both cotyledons (adapted from Soroka and Underwood 2011).

Figure 2.2. Canola seedlings with varying levels of flea beetle damage: Panels a to f represent first leaf stage; Panels g to i represent two leaf stage (adapted from Soroka and Underwood 2011).

3 Flea Flea InjuryBeetleScore (FBSC) 2

Entries 2009 2010 2011

Figure 2.3. Graph showing average flea beetle feeding scores on winter-type canola lines at 14d post planting by entry for each year. Error bars indicate standard error of the mean for each entry over trials in that year (2009=3; 2010=7; 2011=1). Flea beetle injury score as a scale of 0 = dead or no plant remaining to 9 = less than 10% damage. ANOVA indicated significant interaction between year and the entry (=0.05) and by year (=0.05) (see Table 2.5.).

10 9 8 7 6 5 4 3 2

Flea Flea InjuryBeetleScore 1

0 0=full 0=full plantinjury/death; 9= visible no plantinjury

Entries

Spring Fall

Figure 2.4. Graph showing flea beetle feeding scores on winter-type canola lines at 14d post planting by entry for spring and fall planting in 2009 (Belfountain) and 2010 (Alloa and Belfountain). Error bars indicate standard error of the mean for each entry over three trials per season. Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. ANOVA indicated significant interaction between season and entry (=0.05) and between seasons (=0.05)(See Table 2.5.).

3 Flea Flea InjuryBeetleScore (FBSC)

Entries Family 1 Family 2 Family 3 Family 4 Checks

Figure 2.5. Graph showing average combined flea beetle feeding scores on winter-type canola lines from eleven field trials (2009 to 2011). Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. Error bars indicate standard error of the mean for each entry. Family grouping indicate within and among family variation for FBSC. Resistant checks (46A65- treated and Ace-Sinapis alba) have greater resistance than other varieties evaluated; S. alba was significantly more resistant than family 3 (=0.05) and Cutlass (=0.1).

CHAPTER 3 - EVALUATION OF SPRING-TYPE BRASSICA NAPUS GERMPLASM FOR GENETIC DIVERSITY IN RESPONSE TO FLEA BEETLE HERBIVORY IN TYPCIAL AND ATYPICAL PLANTING WINDOWS

3.1. ABSTRACT

To meet the goals of feeding a growing population, ongoing improvements in genetics and in agronomic practices need to be delivered. Flea beetle damage is a major concern for canola growers in Western Canada where most canola is planted with insecticidal seed treatments

(ISTs). With growing concerns regarding the possible impact of ISTs on non-target and beneficial insect populations and resistance in target insect species to pesticides, genetic resistance would be beneficial for a more balanced and integrated pest management strategy. Although there are reports of some genetic insect resistance within Brassica species, currently, none of the registered Brassica napus L. canola varieties exhibit measurable resistance to flea beetle

(Phyllotreta spp.) injury. Fifteen spring-type Brassica napus canola breeding lines were evaluated in 2009, 2010 and 2011 for flea beetle resistance under field conditions to determine whether feeding preference of newly emerged adult flea beetles and overwintered adult flea beetles differed by planting spring-type canola lines in both spring (typical planting window) and late summer (atypical planting window). Data indicated that feeding preference did not vary throughout the flea beetle life cycle, indicating that favourable genes in spring-type canola could be transferred to winter-type canola and vice versa even though the flea beetle life cycle is different for each habit’s critical young development stage. A small number of breeding lines

consisting of six families with two to four breeding lines per family plus a group representing B. juncea, B. rapa, B. napus (with or without IST) and Sinapis alba were evaluated. Response to flea beetle herbivory varied within B. napus and amongst the related brassicas. This supports literature that indicates natural genetic variation within B. napus for flea beetle resistance, which could be further combined into novel flea beetle resistant varieties through a dedicated breeding strategy.

3.2 INTRODUCTION

As reported in Chapter 2, canola is a major crop in Western Canada. Flea beetles, predominantly

Chrysomelidae: Phyllotreta cruciferae (Goeze) and Phyllotreta striolata (F.), are a major canola pest with total economic impact through yield loss and mitigation hundreds of millions of dollars annually in North America (Lamb and Turnock, 1982; FAOSTAT, 2017).

One life cycle per year for flea beetles is common in the Canadian Prairies (J. Knodel and Olsen

2002) although multiple cycles have been thought to occur in places such as Ontario (Kinoshita et al. 1979). In early spring, the flea beetles emerge as overwintered adults in search of food and lay eggs at the base of the plants (Figure 1.1.). The eggs hatch and the larvae develop during the summer and feed on the roots before pupating then emerging as new adults later in the summer.

Late season feedings are deemed not significant although high densities can strip the outer layer of seedpods and any other young green plant material resulting in early maturity, pod shatter and high green seed content (Burgess 1977; Knodel and Olsen 2002).

As there are no flea beetle resistant varieties of canola available (Patel, pers. comm., Gavloski et al. 2000), most commercial canola is planted with a seed treatment, primarily a neonicotinoid.

Significant losses due to flea beetles are reported when a suitable seed treatment and/or foliar application is not applied (Knodel et al. 2008). Unfortunately, seed treatments are only effective as long as the target pest does not overcome the pesticide’s mode of action. In the case of flea beetles, a shift in population make-up as a result of neonicotinoid use has been reported (Tansey et al. 2009). Other agronomic practices such as tillage, crop rotation, early planting, row spacing, higher seeding rates and using larger seed size are all techniques that could be applied to reduce flea beetle injury (Dosdall et al. 1999; Dosdall and Stevenson 2005; Elliott et al. 2008). Seed treatments along with agronomic practices plus genetic resistance are keys to a good integrated pest management (IPM) system. In the end, it is the farmer’s ease of use and availability that dictates what practices are adopted. Genetic resistance or tolerance is the simplest option if available.

As reported in Chapter 1, a number of studies have been conducted to determine if there is natural genetic variation present for flea beetle resistance within Brassica species (Gavloski et al.

2000; Lamb 1988; Lamb et al. 1993). The levels reported are not high enough to replace pesticide use. Transgenic options have been explored, but are not yet commercially available; these include Bt toxins (Payne and Michaels 1995) and increased trichome density (Gruber et al., 2006).

However, native genes are preferred due to development and regulatory costs as well as political and social considerations in some jurisdictions.

The objective of this chapter is to evaluate 15 spring-type Brassica napus canola breeding lines for flea beetle resistance under field conditions and determine whether resistance is affected by seasonal differences between spring and late summer plantings. By exposing young canola plants to flea beetle populations at differing flea beetle life stages, changes in feeding habits were evaluated during the flea beetle life cycle (emerging overwintering versus overwintering preparation) or not. This study provides insight into flea beetle feeding habits on spring canola, the potential to move resistance genes between winter and spring B. napus, and confirms previous reports of genetic variation within and between a number of Brassica species.

3.3. MATERIALS AND METHODS

3.3.1. Plant material

The varieties used in this study were spring-type (non-vernalizing) B. napus doubled haploid (DH) experimental lines developed by DuPont Pioneer. These lines were initially selected from a previous breeding experiment where the seed treatment failed and responses to flea beetle herbivory were documented (data not shown). In 2009, ninety breeding lines from nineteen families were selected. Within those families, two to eight breeding lines with either good or poor flea beetle scores were selected for further studies. Fifteen breeding lines, from six of the nineteen families, with consistent flea beetle scores from the previous two trials, were further evaluated in 2010 and 2011 for flea beetle herbivory. Five check varieties evaluated in each trial included the open-pollinated B. napus variety 46A65 (a former Western Canada

Canola/Rapeseed Recommending Committee (WCC/RRC) registration standard) along with related Brassicaceae species – Ace (Sinapis alba), Cutlass (B. juncea) and Tobin (B. rapa). Previous

work by Gavloski et al. (2000) and Palaniswamy et al. (1997) indicated various levels of resistance to flea beetle feedings in various Brassicaceae. Representative cultivars were included in this study as checks. Spring-type checks were used in this study as they were more readily available and provide common checks in this and the paired winter-type experiment (see Chapter 2).

3.3.2. Test sites and experimental designs

Over the course of this study, twelve trials were planted in four locations – Alloa, ON (43.697050,

-79.870154), Belfountain, ON (43.791724, -80.000064), Acton, ON (43.651602, -79.976759) and

Edmonton, AB (53.430290, -113.541054) from 2009 to 2011 (Table 3.1.). Edmonton was selected in an attempt to evaluate differences in flea beetle population, targeting the striped flea beetle

(P. striolata), although numbers in the sticky traps on site indicate the population was predominantly composed of crucifer flea beetle (P. cruciferae), as in Ontario (data not shown).

Trials were planted in May and late August to early September, with attempts to have two planting dates seven to ten days apart for each trial to align most susceptible plant growth stage

(cotyledon and first and second) with abundant flea beetle populations.

In the spring of 2009, the trials were conducted as split plot experiments with 71 entries and 3 check varieties. The whole plots were based on seed treatment - with or without Helix® (10.3% thiamethoxam, 1.24% difenoconazole, 0.39% metalaxyl-M, 0.13% fludioxonil – PCP registration number 26637–Syngenta Crop Protection Canada Inc, Guelph, ON) applied at 15mL kg-1.

Genotypes made up the subplots. Each whole plot consisted of three randomized blocks, where all six blocks (three replicates x two treatments) were completely randomized. The field layout in

spring 2009 consisted of four ranges per block by 27 rows for a trial. Blocks were placed side by side to be along the treeline that bordered the field. The fall planting in 2009 was four ranges by

25 rows wide per block, with blocks side by side or end to end, depending on the location. All trials in 2010 and 2011 were conducted as RCBD with three blocks where each block was a replicate and each block was arranged as two ranges by 10 rows. Blocks were side-by-side. There were 20 entries in total for each block in 2010 and 2011. The Table 3.1. shows the locations, experimental design, number of entries, planting date and data that was collected.

All locations were prepared using conventional agronomic practices. Prior to planting, all sites were tilled and fertilized and pre-emergent herbicide (Bonanza 480TM – Trifluralin; PCP registration number 28289 – UAP Canada, Dorchester, ON) that was soil-incorporated. All sites were planted with a Hege 1000 nursery planter with double disc openers (Wintersteiger, Salt

Lake City, UT). Target seeding rate was 120 seeds per row with row length of 4.5m seeded then trimmed to three metres. Row spacing was 30cm in 2009 and 50cm in years 2010 and 2011.

Attempts were made for two plantings of each trial approximately seven to ten days apart. For each pair of planting dates, the trial with the best visual appearance for discriminating flea beetle herbivory responses was evaluated.

3.3.3. Phenotypic data - scoring of plants for flea beetle herbivory

As described in Chapter 2, plants were scored for flea beetle herbivory at 14 days post planting, as well as 21 days post planting, where possible. Rows were either scored on a scale of 0-9, as a single visual average or the average of 15 individual plants as noted in Table 3.1.

3.3.4. Statistical analysis

The response variable (flea beetle feedings at 14d and 21d) was analyzed using PROC GLM (SAS

Institute 2012) and data was tested for normality using PROC UNIVARIATE (SAS Institute 2012).

Homogeneity of variance was examined the by using the residuals in PROC REG (SAS Institute

2012). Where single plants were scored, row averages (calculated by Microsoft Excel®) were used in the analysis. For the combined analysis, the common 20 entries from 2009, 2010 and 2011 were analyzed together. The untreated IST main blocks from the split plot design in spring 2009 were taken and combined with the other untreated IST RCBD trials. Only trials with acceptable homogeneity of variance and C.V.s less than or equal to 25 were included, reducing the combined analysis to nine trials from twelve. Statistical analyses included ANOVA and MEANS comparisons using PROC GLM (SAS Institute 2012) and R (R Core Team, 2015). Pair-wise genotype comparisons were examined using LSD at  = 0.05.

3.4 RESULTS AND DISCUSSION

3.4.1. Results

The means and ANOVA results for 14 days and 21days post-planting have been represented in

Tables 3.2. and 3.3., respectively. When evaluating site flea beetle pressure, approximately half of the locations had low flea beetle pressure (FBSC: 7-9), whereas the other half had moderate flea beetle pressure (FBSC: 4-7) and one location (2010-Alloa, Spring) had very high pressure

(FBSC: 1.7). ANOVA models fit best with the moderate to higher flea beetle pressures but the trial’s C.V. also increased with the flea beetle pressure indicating increased variability (also seen in Chapter 2). Tables 3.2. and 3.3. also show the ranking of the experimental entries. As expected,

treated check 46A65 and S. alba were among the more resistant whereas the B. rapa and B. juncea check varieties were more susceptible. With the exception of B. rapa, the checks were similarly positioned by rank in 14 days vs 21 days post-planting. The experimental lines showed ranking as expected in all of the six families, with at least one line more resistant to flea beetles than at least one other within its family (Figure 3.3.). No check was significantly different than any other family (=0.05). Figure 3.1. shows the year-by-year comparisons for each entry, which was significant in that 2010 was different compared to 2009 and 2011 (Table 3.5.; =0.05). There was no significant entry-by-year interaction (Table 3.5.; =0.05), which indicated no genotype- by-environment interaction. Table 3.5. and Figure 3.2. indicated significant seasonal effects

(=0.05).

Split plot ANOVA results for 2009 trials at 14 and 21 days post planting are presented in Table

3.4. In both Alloa and Belfountain reps were significant (=0.05) for the 21d scores but not at

14d. Treatment was significant (=0.05) at three of four evaluations with a 21d trial at

Belfountain being the exception. Entries were significant (=0.05) at 21d but not 14d at both locations. Treatment-by-entry was not significant (=0.05) in either trial or date in 2009.

Table 3.5. shows the ANOVA results for flea beetle herbivory 14 days post planting with nine of the twelve trials from 2009 to 2011 combined into a single analysis. Alloa and Edmonton from spring 2010 were not included due to C.V. greater than 25%. Year, season and location were significant (=0.05) and year (=0.1). There were no significance differences for replicates nested within location and the two-way interactions of location-by-entry, season-by-entry nor year-by-

entry interactions. There were no significant differences (=0.05) between any check variety to any of the families. Neither the entry by year interaction (Figure 3.1.; =0.05) nor season (Figure

3.2.; =0.05) were significant on feeding scores, although both were notable effects.

3.4.2. Discussion

As noted in chapter 2, flea beetle studies have proven challenging to conduct in the field (Lamb

1988; Lamb et al. 1993). To address the challenges, this study was conducted for a total of 12 experiments over three years, in two Canadian provinces. As natural insect populations were being relied upon, two plantings were attempted to optimize more ideal herbivore pressure. In most cases, however, for reasons such as low insect pressure and unfavourable weather conditions, a second planting or evaluation was not possible in all cases. When a second planting was available, only the better of the two experiments was scored. Natural population pressure was greatest in 2010 but resulted in higher C.V. (Tables 3.2. and 3.3.) as a result of more variable environment; likely due to higher concentration of feedings in certain areas of the field and not uniformly distributed. The precision of the test to distinguish resistance levels in genotypes may have been reduced due to excess or uneven insect feeding pressure. With lower insect pressure, feeding scores were low and very few differences were noted as reported earlier (Lamb 1988).

The combined data set resulted in moderate C.V. and significant differences (=0.05) noted for entry, year and location. Two trials were not included in the combined dataset due to greater than 25% C.V. values for the trial (Table 3.5.).

This study included spring-type B. napus genotypes plus several spring-type Brassicaceae checks.

Various levels of resistance within Brassicaceae studied (inter-specific and intra-specific) have been documented by others (Brandt and Lamb 1993; Gavloski et al. 2000; Putnam 1977).

Greatest differentiation appeared among Brassica species and as a result, three non-B. napus checks along with one B. napus (treated and untreated with seed treatment) were included to assist in determination of a range of resistance to flea beetle herbivory in spring canola. In this study, greater differences were noted for inter-specific comparisons versus within-species. This result is consistent with Lamb (1988) and Palaniswamy and Lamb (1992). Lamb et al. (1993) also noted high variability issues, especially within the row of each genotype when trying to identify de novo antixenosis. Palaniswamy and Lamb (1992) noted that feeding patterns may change from cotyledons to true leaves. This was observed in the current study, particularly in B. juncea cultivar

Cutlass as well as entries Pioneer3 and Pioneer4 where these sister lines flipped ranking scores at 14 and 21 days post planting (Tables 3.2. and 3.3.).

The study began as a split plot experiment in two locations in 2009. The results showed that the insecticide was effective and that the efficacy of the seed treatment did lessen as time progressed as noted in 21d compared to 14d (Table 3.4.). The seed treatment comparison was originally included as a positive check to be able to compare plant response to more severe herbivory while having a treated reference of plots for comparison. In the end, it was determined this was not an effective use of time and was abandoned for subsequent experiments.

This study was paired with winter-type material (as described in Chapter 2.), therefore, material from both projects was planted twice in a growing season; once in early spring and a second time, in late summer to early fall. For spring-type material, the highly susceptible cotyledon growth stage would be present in the spring when overwintering flea beetles are looking for food prior to ovipositing, representing a natural pest-host relationship. This material was also planted in the late summer to early fall when winter-type material is typically planted. In this case, flea beetles would be preparing to overwinter. This would not represent the typical pest/host relationship for spring-type canola. By comparing the feeding damage in both planting, it was concluded that adult flea beetle feeding preference, as either a newly emerged adult or an overwintered adult, were not significant in this study. As noted by Lambdon (2003) and McCloskey and Isman (1995), feeding preferences and plant reactions may change as the plants mature. In this experiment where flea beetles are exposed to the same host growth stage, there appears to be no difference, whether the crop was spring or fall planted. Therefore, it can be hypothesized that any resistance mechanisms may be similar between spring-type and winter-type B. napus and those sources of resistance could be transferred between the two plant habits. On the other hand, this may also mean that there may be few mechanisms involved in flea beetle resistance.

3.5 CONCLUSION

The number of breeding line entries in this study represented a very small proportion of the genetic variation within spring-type B. napus. As such, greater differences between resistance and susceptibility may be seen with larger populations and useful levels of resistance or tolerance may be identified. It may have been more useful to continue the initial study with a large

population to confirm the reduction in numbers was indeed based on variation in resistance and not more environmentally influenced. However, the results of this study are in agreement with previous reports indicating that inter-specific crosses may be required to transfer genes from related species into the crops of interest. The mechanisms involved in flea beetle antixenosis in canola is very complicated as one is working with two biological entities that are interacting at various stages in their respective life cycle. The potential to move genes not only between related

Brassica species but also different growth habits increases the opportunities to achieve the goal of increased flea beetle resistance (antixenosis) in Brassicaceae, particularly canola, for Western

Canada. The defense mechanism for flea beetle herbivory is likely multi-genic and biochemically complex with multiple pathways involved. If and once a pathway of significance is identified, manipulation of the pathway may provide enough resistance to be effective.

Table 3.1. Summary of experiments, dates, locations, experimental design, entry numbers and notes collected for all spring-type B. napus trials examined for flea beetle injury from 2009 to 2011.

Experimental Number Planting Notes Experiment Season Year Location Design of entries Date Collected

1 Spring 2009 Alloa Split Plot 71 25-May †, ‡ 2 Spring 2009 Belfountain Split Plot 71 25-May †, ‡ 3 Fall 2009 Belfountain RCBD 71 26-Aug †, ‡ 4 Spring 2010 Edmonton1 RCBD 20 18-May † 5 Spring 2010 Edmonton2 RCBD 20 1-Jun † 6 Spring 2010 Alloa RCBD 20 30-May † 7 Spring 2010 Belfountain RCBD 20 19-May † 8 Fall 2010 Alloa1 RCBD 20 18-Aug † 9 Fall 2010 Alloa2 RCBD 20 28-Aug † 10 Fall 2010 Belfountain RCBD 20 11-Aug † 11 Spring 2011 Edmonton RCBD 20 24-May †, ‡, §, ^, ¶, # 12 Fall 2009 Acton RCBD 24 26-Aug †, ‡ †, ‡, are flea beetle herbivory scores 14 days (cots only) and 21 days post-planting respectively, cot (§) and leaf (^) scores separate, # is the number of plants per square meter, ¶ is plant height measured from the soil surface to upper most point

1-Pioneer1 1 9.00 6.00 8.33 8.00 0.97 6.16 5.67 7.00 4.49 4.18 6.93 8.33 6.26 0.39 12¶

2-Pioneer2 1 8.00 6.67 8.00 9.00 0.73 6.42 5.67 6.67 4.60 5.38 6.89 8.00 6.34 0.37 9

3-Pioneer3 2 8.67 6.33 8.33 8.67 1.50 6.53 4.33 6.67 4.67 5.27 7.22 8.00 6.35 0.38 8

4-Pioneer4 2 8.67 5.67 6.67 8.67 1.27 7.62 5.00 7.00 3.67 4.64 6.64 8.00 6.13 0.37 17

5-Pioneer5 3 9.00 7.33 8.67 8.30 1.43 5.67 4.33 5.33 4.98 4.42 7.33 8.00 6.24 0.39 15

6-Pioneer6 3 8.00 7.00 8.33 8.00 1.50 7.87 5.67 7.00 5.49 5.02 6.73 8.00 6.55 0.33 6 ¶

7-Pioneer7 4 8.33 5.33 8.67 9.00 0.80 6.78 4.67 7.33 5.33 4.27 6.98 8.00 6.29 0.42 11

8-Pioneer8 4 8.00 7.67 8.33 8.33 0.87 6.71 4.67 6.00 4.78 4.87 6.54 7.67 6.20 0.38 16

9-Pioneer9 4 8.67 7.67 8.67 8.00 0.90 5.80 5.00 6.67 3.95 5.00 6.82 8.00 6.26 0.41 12¶

10-Pioneer10 5 8.00 5.67 8.00 8.00 1.57 6.13 4.33 6.67 3.87 5.58 7.09 8.00 6.08 0.36 19

11-Pioneer11 5 8.00 7.00 8.67 8.33 2.73 7.42 4.33 7.00 5.36 5.71 6.98 8.00 6.63 0.32 4

12-Pioneer12 6 8.33 7.33 7.67 9.00 1.87 7.20 5.67 7.33 4.35 5.18 6.96 7.67 6.55 0.35 6¶

13-Pioneer13 6 8.33 5.00 9.00 9.00 1.30 6.62 4.00 6.67 4.82 5.18 6.94 8.00 6.30 0.39 10

14-Pioneer14 6 8.67 6.00 8.00 8.33 0.97 5.64 4.00 6.00 4.00 3.89 7.24 7.67 5.92 0.39 20

15-Pioneer15 6 8.00 6.00 7.67 8.33 1.77 6.69 3.00 5.67 4.76 6.22 7.29 8.00 6.12 0.35 18

16-Cutlass(BJ)§ 7 9.00 7.67 8.67 9.00 1.00 6.31 6.67 7.33 2.22 2.89 6.31 8.00 6.25 0.47 14 17-Tobin(BR)§ 7 8.33 6.00 8.00 7.67 2.23 5.42 6.00 7.00 n/a n/a n/a 8.00 6.60 0.39 5

18-Ace(SA)§ 7 9.00 8.33 8.67 9.00 5.23 7.47 4.67 7.67 6.00 6.18 7.51 7.33 7.23 0.25 3

19-46A65(t)§ 7 n/a n/a 8.67 8.11 3.60 7.89 5.67 7.00 6.58 6.62 7.49 8.33 8.18 0.13 1 20-46A65(u)§ 7 n/a n/a 8.33 8.80 2.57 7.27 4.33 7.17 5.27 6.22 7.54 8.00 7.57 0.17 2

FBSC Mean 8.44‡ 6.69‡ 8.27‡ 8.50‡ 1.74 6.68 4.91 6.76 4.69 5.09 7.02 7.95

Expt C.V. 8.26 25.04 9.81 8.92 30.93 19.16 25.91 11.62 23.2 14.34 6.92 4.95 2 R 0.35 0.39 0.43 0.41 0.87 0.43 0.45 0.48 0.59 0.77 0.42 0.37

Total Entries 18 18 20 20 20 20 20 20 20 20 20 20

Pr>F (model) 0.5358 0.3647 0.2003 0.2667 <.0001 0.1847 0.1750 0.0894 0.0072 <.0001 0.2315 0.4828

Pr>F (rep) 0.4582 0.1191 0.1132 0.0230 0.2737 0.0151 0.1308 0.4321 0.0127 <.0001 0.8988 0.2699 Pr>F (entry) 0.5133 0.4813 0.2649 0.5382 <.0001 0.4411 0.2469 0.0747 0.0176 <.0001 0.1679 0.5101 † full experiment was Split Plot with treated and untreated main plots. Data used for this summary is only the untreated RCBD component with all entries. Some entries are not included in table shown but overall statistics include hidden entries. # indicate both and which plantings in that season were evaluated, remaining trials scored only one of two plantings. ‡ FBSC mean

is average of entries shown on table, entries within experiment not shown on table not included in this calculation. ¶ where rankings were equal, all entries assigned highest rank, next unique rank continued assigned based on rank without unique scores. § BJ=Brassica juncea, BR=Brassica rapa, SA=Sinapis alba, t=seed treatment, u=untreated (no seed treatment).

Table 3.3. Mean scores for flea beetle herbivory (FBSC) 21 days after planting on 20 spring-type Brassicaceae varieties evaluated from 2009 to 2011 in Acton (ACT), Alloa (ALL), Belfountain (BEL), ON and Edmonton (EDM), AB. SAS PROC GLM statistics Type III ANOVA summary on the lower part of the table by experiment. FBSC: Scale of 1-9 where 0 = plant dead and 9= no damage. 2009† 2011 Combined Spring Fall Spring FBSC Mean Overall Variety Family BEL- BEL- EDM- EDM- SE ALL BEL (cot/first Rank cot leaf cot leaf score)

1-Pioneer1 1 7.67 7.67 6.67 8.33 8.77 9.13 7.69 0.31 5

2-Pioneer2 1 7.67 6.33 7.00 7.67 8.57 8.63 7.39 0.27 13

3-Pioneer3 2 7.33 7.00 7.00 8.00 8.90 9.03 7.61 0.35 7

4-Pioneer4 2 8.00 7.33 4.67 7.00 8.90 8.70 7.23 0.61 18

5-Pioneer5 3 7.67 6.33 6.67 7.67 8.57 8.77 7.31 0.32 16

6-Pioneer6 3 7.33 7.00 7.33 7.67 8.57 8.93 7.56 0.24 9

7-Pioneer7 4 7.00 6.50 6.67 7.67 9.00 9.03 7.36 0.35 14¶

8-Pioneer8 4 7.33 7.00 7.00 8.00 8.80 9.07 7.53 0.27 12

9-Pioneer9 4 8.33 7.33 5.67 7.33 8.83 9.03 7.54 0.41 11

10-Pioneer10 5 7.67 7.33 6.67 7.33 8.83 8.77 7.63 0.36 6

11-Pioneer11 5 8.00 7.33 7.00 7.67 8.63 9.10 7.74 0.28 4

12-Pioneer12 6 7.67 7.00 7.00 7.33 8.67 9.20 7.58 0.25 8

13-Pioneer13 6 8.33 7.67 7.33 8.33 8.50 8.94 7.96 0.21 2

14-Pioneer14 6 8.67 6.67 7.00 7.33 8.87 8.63 7.80 0.33 3

15-Pioneer15 6 7.33 6.67 6.67 7.33 8.77 9.07 7.36 0.31 14¶

16-Cutlass-BJ 7 7.00 4.33 7.00 7.00 8.10 7.20 6.61 0.56 19

17-Tobin-BR 7 6.67 5.33 6.33 7.89 n/a n/a 4.73 0.84 20

18-Ace-SA 7 7.67 8.00 8.00 8.22 8.90 6.23 8.14 0.16 1

19-46A65t 7 8.17 7.68 7.11 7.80 9.00 9.20 7.26 0.18 17 20-46A65u 7 NA NA 6.57 7.48 9.00 8.73 7.55 0.08 10

FBSC Mean 6.94‡ 7.03‡ 6.75‡ 7.35‡ 8.75 8.71

Expt C.V. 11.34 9.66 15.74 12.03 5.33 7.73 2 R 0.59 0.58 0.47 0.42 0.29 0.65

Total Entries 75 75 75 75 20 20

Pr>F (model) <.0001 <.0001 0.0025 0.0307 0.7520 0.0009

Pr>F (rep) <.0001 <.0001 <.0001 <.0001 0.2773 0.8350 Pr>F (entry) 0.0002 <.0001 0.3198 0.1786 0.8042 0.0005 ** indicates trials where grand mean and associated statistics reported include all entries from that experiment whereas only select entries that were used over multiple years appear on the table

Table 3.4. Split plot analysis using PROC ANOVA (SAS Institute) of flea beetle herbivory in spring-type canola lines after 14 and 21 days post-planting at Alloa and Belfountain, ON in 2009. Insecticidal seed treatment† was main plot (with or without). Alloa 14 days post planting 21 days post planting Effect DF Type III SS MS Pr > F DF Type III SS MS Pr > F Total 448 183.91 447 217.96 Rep 2 0.40 0.20 0.5582 2 3.39 1.69 0.0059 Trt† 1 37.17 37.17 <.0001 1 11.52 11.52 <.0001 Error (1) 2 1.11 0.56 2 6.67 3.33 <.0001 Entry 74 20.34 0.27 0.8726 74 73.73 1.00 <.0001 Trt*Entry 74 23.69 0.32 0.6313 74 27.26 0.37 0.2310 Error (2) 295 101.19 0.34 294 95.38 0.32 R2=0.45 C.V.6.83 R2=0.56 C.V.=7.12

Belfountain 14 days post planting 21 days post planting Effect DF Type III SS MS Pr > F DF Type III SS MS Pr > F Total 447 5317.61 442 282.92 Rep 2 37.73 18.87 0.1715 2 11.29 5.64 <.0001 Trt 1 541.92 541.92 <.0001 1 0.05 0.05 0.7397 Error (1) 2 3.77 1.88 2 8.22 4.11 Entry 74 857.05 11.58 0.3074 74 106.83 1.44 <.0001 Trt*Entry 74 750.28 10.14 0.5877 74 31.89 0.43 0.4874 Error (2) 294 3126.86 10.64 289 124.92 0.43 R2=0.41 C.V.42.00 R2=0.56 C.V.=9.32 † (trt) Helix ® liquid seed treatment is a Syngenta Crop Protection Canada Inc product that contains the insecticides and fungicides thiamethoxam, difenonconazole, metalaxyl-M and S-isomer and fludioxonil

Table 3.5. PROC GLM (SAS Institute) ANOVA of flea beetle herbivory on spring- type Brassicaceae varieties after 14 days post-planting for nine combined experiments conducted at Alloa and Belfountain, ON and Edmonton, AB in the spring and fall of 2009 to 2011. Sum of Mean F- Source DF Squares Square Value Pr>F Model 126 926.41 7.35 6.55 <.0001 rep(loc) 6 2.18 0.36 0.32 .9246 location 2 22.56 11.28 10.05 <.0001 season 1 35.67 35.67 31.78 <.0001 location*season 0 0 year 2 71.01 35.50 31.63 <.0001 location*year 0 0 entry 19 30.62 1.61 1.44 0.1060 location*entry 38 33.80 0.89 0.79 0.8076 season*entry 19 24.62 1.30 1.15 0.2942 year*entry 38 49.96 1.31 1.17 0.2306 Error 390 437.72 1.12 Corrected Total 516 1364.13 R2=0.68 C.V.=15.48 Root MSE=1.06 FBSC Mean=6.84

10.00 9.00 8.00 7.00 6.00 5.00 4.00 3.00

2.00 Flea Flea Beetle Injury Score (FBSC) 1.00 0.00

2009 a 2010 2011 Entries

9.00 8.00 7.00 6.00 5.00 4.00 3.00 2.00 1.00

0.00 Flea Flea Beetle Injury Score (FBSC)

Entries spring fall

Figure 3.2. Graph showing flea beetle feeding scores on spring-type canola lines at 14d post planting by entry for spring and fall planting in 2009 (Belfountain) and 2010 (Alloa and Belfountain). Error bars indicate standard error of the mean for each entry over three trials per season. Flea beetle injury score as a scale of 0 = dead or no plant to 9 = less than 10% damage. ANOVA indicated no interaction between season and entry (=0.05) but showed significant differences between seasons (=0.05).

9 8 7 6 5 4 3 2

1 Flea Flea Beetle Injury Score (FBSC) 0

Entries

1-Pioneer1Family 1 3-Pioneer3Family 2 5-Pioneer5Family 3 7-Pioneer7Family 4 10-Pioneer10Family 5 12-Pioneer12Family 6 16-Cutlass-BJChecks

CHAPTER 4 – GENETIC MAPPING OF QTL FOR RESISTANCE TO FLEA BEETLE (PHYLLOTRETA SPP.) IN TWO WINTER-TYPE B. NAPUS DOUBLED HAPLOID POPULATIONS.

4.1 ABSTRACT

Based on indoor and outdoor evaluations of flea beetle herbivory, genetic maps of two doubled haploid (DH) winter-type Brassica napus populations (J10-02 and J10-11) were constructed using simple sequence repeat (SSR) markers as well as single nucleotide polymorphic (SNP) markers provided by DuPont Pioneer. The combined SSR and SNP markers used corresponded to 226 and

354 loci in Population J10-02and Population J10-11, respectively. Six unique quantitative trait loci

(QTL) were identified as significant with a number of major and minor QTL aligning between the two populations in relation to flea beetle herbivory on cotyledons of B. napus. Field evaluations were reported as percent damage per row of individuals, while indoor screening was measured on an individual plant basis with the number and types of bites recorded. Two bite types were documented where the test bite had only the surface tissue removed compared to full bites where all tissue between the upper and lower epidermis was consumed. In Population J10-11, linkage group (LG) N1 and N6 indicated a stronger peak for test bite compared to full bites. By recording bite types separately, it may allow for future investigation into mechanisms of resistance that play a role in attracting or repelling the flea beetle compared to those that allowed or inhibited continued feeding. The QTL identified here have not been previously reported and their functional role is unknown.

100

4.2 INTRODUCTION

Approximately 8.3 million hectares of canola are grown annually in Western Canada (Canola

Council of Canada, 2016). In 2016, about 18.4 million tonnes of canola oilseed was produced

(Canola Council of Canada, 2016). The western Canadian canola crop is spring planted usually from late April to early June. It is during this time that flea beetles, predominantly Phyllotreta cruciferae (Goeze) and Phyllotreta striolata (F.) (Coleoptera: Chrysomelidae: Alticinae), emerge from overwintering in surrounding field grasses, hedgerows and bushes in search of food and a place to lay their eggs (Burgess, 1977a). Canola is most susceptible from the time it emerges from the soil until the four leaf stage, at which point the plants can outgrow subsequent insect damage

(Gavloski and Lamb 2000; Knodel and Olsen 2002). Flea beetles damage the cotyledons and first true leaves and therefore can reduce yields by up to 10% (Lamb and Turnock 1982). Flea beetle damage can also delay plant development and cause uneven plant height and delayed maturity as well as affect the chlorophyll content (Bodnaryk and Lamb 1991; Lamb 1982). Ontogeny also plays a role in insect resistance as expression of resistance changes as the plants develop (Boege and Marquis, 2005). Bodnaryk and Lamb (1991b) reported that Sinapis alba has better flea beetle resistance due to increased tolerance and antixenosis which is greater at earlier growth stages versus B. napus. This early growth stage is most crucial to canola development as hundreds of millions of dollars are lost annually due to yield loss and insect control due to flea beetles (Lamb and Turnock, 1982; Knodel and Olsen, 2002).

101

Within Brassicaceae, a number of studies have reported that there is natural genetic variation present for flea beetle resistance (Gavloski et al. 2000; Lamb 1988; Lamb et al. 1993). Various levels of resistance (inter-specific and intra-specific) have been observed and shown to be genetic and heritable (Brandt and Lamb 1993; Putnam 1977). While these studies have shown the difficulty of studying resistance to flea beetles, they did indicate that resistance existed within the Brassicaceae but not at a commercially viable level (Patel, pers. comm.).

Brassica napus is one of many polyploid crop species; a group that includes Triticum aestivum

(wheat), Solanum tuberosum (potato) and Avena sativa (oats). Nearly 20 million years ago, the hybridization between diploid progenitors Brassica oleraceae (C genome, n = 9) and Brassica rapa

(A genome, n = 10) resulted in the allopolyploid species Brassica napus (AC genome, n = 19) (U,

1935; Arias et al., 2014). This hybridization resulted in numerous duplicated segments and homoeologous regions within B. napus. Consequently, discriminating between a) two homoeologous sequences and b) two nearly-identical homoeologous sequences is complex and difficult in polyploids such as B. napus (Kaur et al., 2012). Extensive breeding efforts in B. napus and its diploid progenitors have contributed to an ever expanding genetic and genome knowledge base of the diploid and amphidiploid genomes that constitute U’s triangle (U, 1935) and define the B. napus genome (Wang et al., 2011; Chalhoub et al., 2014; Parkin et al., 2014).

Molecular markers for identifying and tracking plant characteristics are widely used in plant breeding (Mammadov et al., 2012). These tools are particularly useful when phenotypic data is difficult to collect as genetically-based tools are not subject to changing environmental

102

conditions (Bazakos et al., 2017). Molecular markers also allow breeders and other scientists to determine plant traits prior to physically growing plants to collect this data phenotypically

(Mammadov, et al., 2012). As newer techniques become more efficient, the cost of molecular analysis declines, and it becomes much more cost effective to genotype than to phenotype traits; however, trait evaluations under uniform and representative conditions are essential for optimizing molecular tools (Vales et al., 2005). Furthermore, the resulting genetic knowledge can also be applied to identifying traits to monitor subsequently in progenies as breeding programs become more of an exact science and less of an art.

Genetic linkage maps have been commonly used by scientists to identify chromosomal regions that contain genes controlling simple traits (single gene) or quantitative traits using quantitative trait loci (QTL) to follow gene movement across populations and related species. The first

Brassica napus genetic map was developed by Landry et al. (1991) using restriction fragment length polymorphism (RFLP) markers. Since then, a large number of genetic maps have been created using new populations and traits along with the latest molecular techniques. Gali and

Sharpe (2012) provide an extensive summary of the molecular linkage studies of Brassica spp. crosses, mapping populations, markers systems, map details and the quantitative traits identified. Raman et al. (2013) developed a consensus map using diversity array technology

(DArT) markers.

Simple sequence repeats (SSRs) and single nucleotide polymorphisms (SNPs) are distributed throughout the genome. With next generation sequencing (NGS), large numbers of SSR and SNP

103

markers are being developed. SSRs and SNPs are now becoming commonly used in many genetic and genomic applications such as fine mapping of genes and association studies, (Xu et al., 2011;

Bus et al., 2012; Bazakos et al., 2017; Bird et al., 2017; Qu et al., 2017). SSRs are highly reproducible, codominant, specific, highly polymorphic and amenable to mid-throughput automation (Mammadov et al., 2012). In order to design specific primers, prior knowledge of the genomic sequence is necessary, and therefore it limits the use to economically important species

(Shehata et al., 2009). SNPs followed SSRs in the evolution of genomic technologies. SNPs are highly abundant, ubiquitous, biallelic in nature and amenable to high and ultra-high throughput automation (Mammadov et al., 2012). Detection of SNPs in polyploid species is more difficult due to the homoeologous chromosomes where useful SNPs discriminate within genomes and not between genomes (Ganal et al., 2009).

The objective of this chapter is to evaluate two winter-type doubled haploid Brassica napus canola populations for flea beetle resistance under field and indoor conditions to identify QTL associated with flea beetle feeding. Two bite types were documented in the indoor evaluation where the test bite had only the surface tissue removed compared to full bites where all tissue between the upper and lower epidermis was consumed. By recording bite types separately, it may allow for future investigation into mechanisms of resistance that play a role in attracting or repelling the flea beetle compared to those that allowed or inhibited continued feeding. By comparing the indoor bite types to field evaluations, the two methods can be compared to see effectiveness of the evaluation methods. This study provides insight into the flea beetle feeding habits and the identification of QTL associated with flea beetle herbivory on canola thus allowing

104

the potential to identify resistance genes or mechanisms involved with flea beetle feeding or antixenosis using the genetic variation within the Brassica napus.

4.3. MATERIALS AND METHODS

4.3.1 Plant material and experimental design

4.3.1.1. Plant materials

Two segregating winter-type doubled haploid (DH) populations were used in this study. The DHs were produced using proprietary modifications to doubled haploid production methodologies described originally by Chuong and Beversdorf (1985) and Swanson et al. (1987). The DHs were created by microspore culture from two F1 hybrids generated from the cross between two DH winter-type canola quality Brassica napus experimental breeding lines from the University of

Guelph. One population was between two sister breeding lines (1147-01 x 1147-03 = Population

J10-02) while the second was between two unrelated breeding lines (1147-01 x 1147-13 =

Population J10-11). The pedigree (Purdy et al., 1968) for 1147-01 and 1147-03 is

Pollen//PS60/PS93 and the pedigree for 1147-13 is PS60/92-81//Otrad15/3/Otrad16/2541. In previous studies (see Chapter 2), the level of resistance to flea beetle herbivory was determined for each parental line. In Population J10-02, one parent is more resistant compared to the other

(Table 2.2.). For the second population, Population J10-11, both parents were considered resistant at 14 days but unrelated as evaluated in Chapter 2. More than one hundred fertile DH lines were developed for each population.

105

4.3.1.2. Indoor experimental design

Doubled haploid entries were evaluated for flea beetle resistance using greenhouse arenas.

Arenas were modifications of the protocol described by Palaniswamy and Lamb (1992).

Screening was performed in a completely randomized block design as the arenas held only 25 entries at a time in a five by five configuration (Figure 4.1.). Common check entries, 46A65 and

Ace, were used in each arena. All entries were replicated up to ten times. The experimental unit was a single plant with two cotyledons. Flats were filled with ProMix BX (Premier-Plant Products) with 1-2 seeds per cell. Flats were planted as per the randomization and grown outside of the arenas. Plants were placed in a growth room at 20oC/18oC day/night temperatures with 16 hours photoperiod under 400W and 600W sodium lamps. At seven days post planting, cotyledons were fully open and first leaf not yet emerged, at which point, any double planted cells were thinned to one plant per cell. Once thinned, flats were added to the arenas and sealed. If both seeds did not germinate, that entry was scored as missing value. Flea beetles from natural populations in the surrounding fields were captured by sweep net and a manual aspirator.

Insects were used within 24 hours collection. Fifty flea beetles were added to each arena of 25 plants. After 24 hours, the flea beetles were removed and the cotyledons were scored immediately or placed in a coldroom at 5oC for upto 1 week until scoring could be completed.

This stopped plant growth and allowed for flea beetle bites to be counted without affecting data quality.

Three traits were scored:

a) number of test bites which were considered superficial injury,

106

b) number of feeding bites where no plant matter remained, and

c) total number of bites which were the sum of test bites (a) and feeding bites (b).

Each bite was scored as a single bite; where multiple bites may have been clustered, the circular patterns of each single bite could be distinguished so that all bites were counted with a high degree of confidence. Feeding bites were more challenging to score where plant matter was missing but best estimates were determined using test bites on that cotyledon for guidance. For example, if the bite appeared to be like the infinity symbol (∞) or the number 8, and the size was similar to two test bites, it likely represented two side-by-side bites and was scored as two bites.

Total bites were recorded as the sum of both test bites and feeding bites (Figure 4.2). Scores were analyzed by PROC GLM using SAS 9.4 software (SAS, 2012).

4.3.1.3. Outdoor experimental design

In September of 2013, a field experiment was performed near Fergus, Ontario to evaluate the two DH populations (Population J10-02 and Population J10-11) under field conditions. Two experiments were conducted; each devoted to a single DH population. The first experiment included 203 experimental entries from Population J10-02 plus the two parental lines and a resistant check Sinapis alba cv Ace. The second experiment consisted of 142 entries of

Population J10-11, both parental lines and a resistant check, S. alba cv Ace. Both experiments were planted in randomized complete block designs with three replications arranged side by side. No additional measure for spatial variation was performed. Experimental units consisted of one single row of 60 seeds planted with a Hege 1000 single row nursery planter. A general

107

assessment of each row was taken to record an overall percent damage per row after 14 days post-planting was recorded based on Figures 2.3. and 2.4. (Soroka, 2011).

In May of 2015 and 2016, field experiments were conducted near Saskatoon, SK, to evaluate the same two doubled haploid populations (Population J10-02 and Population J10-11) under field conditions in a single experiment with a four-replicate randomized complete block design with diagonal checks. Diagonals were set at 7% of total plots in a two over-three up layout. By combining the two DH populations into one RCBD experiment and using diagonal checks within, spatial variation was better addressed. Each entry consisted of up to five single plants per replicate. Five seeds spaced 2.5cm apart were glued onto crepe paper in 10cm row lengths using glue stick in the randomized plot design. The row length was 10cm with a 20cm gap between ranges (Figure 4.4a.). The trial was 40 ranges deep by 40 rows wide, with 30cm spacing between rows (Figure 4.4c.). Ten percent of the total plots were diagonal checks inserted into the experiment (Gilmour et al., 1997). The diagonal checks were three varieties (two B. napus –

46A65 and an experimental Pioneer inbred plus S. alba cv Ace). The checks were selected to attempt to cover a range of feeding resistance to flea beetle but skew the content to be more B. napus than other Brassicas. The crepe paper seed tape was manually planted about 1cm deep into a shallow rototilled site (Figure 4.4b.). Flea beetle damage was scored about 14 days after planting before the first leaf opened.

Scores were based on percentage of tissue damaged of eaten by the flea beetles (percent damage) per plant and averaged per single plot. Spatial analysis was conducted using a diagonal

108

checks embedded across the experimental design (Gilmour et al., 1997) and best linear unbiased estimators (BLUEs) were calculated using Pioneer proprietary software based on methods published by Smith et al. (2001a), Smith et al. (2001b, 2005), and Gilmour et al. (2016). Raw unadjusted data and BLUEs were used in the generation of maps and identifying QTL. This experiment was repeated in May 2016 in Saskatoon.

4.3.2. Molecular markers and genotyping

Two different DNA extraction methods were used, depending on which platform was used.

SSRs were analyzed prior to SNPs so separate tissue collections were conducted as DNA quality would deteriorate over the several months between runs.

4.3.2.1. DNA extraction-SSR

Genomic DNA was isolated from lyophilized leaf tissue collected from four three–week old plants of the parents. Eight 0.2 cm2 leaf disks per entry were placed in a single well of a 96-well plate and lyophilized. DNA was extracted from ground leaf tissue using a simplified CTAB method based on Doyle and Doyle (1990). Cool isopropanol was added to vials. The vials were gently mixed by inverting several times to precipitate the DNA. Samples were incubated at -20oC for 20 minutes.

Samples were then centrifuged at 4000 rpm for 20 minutes at 10oC to pelletize the DNA. The isopropanol was carefully removed so as not to disturb the pellet. A volume of 70% ethanol was added to the pellets to purify the DNA further. Samples were vortexed and incubated overnight at 4oC. Samples were centrifuged at 4000 rpm for 10 min and the ethanol was carefully removed.

109

The pellets were air dried overnight. The DNA pellets were rehydrated and dissolved in distilled water, then re-suspended overnight prior to quantification.

4.3.2.2. Generation of SSR data

SSR markers were developed in-house by DuPont Pioneer. SSRs were selected on the basis of their ability to discriminate among the parental lines in canola germplasm. PCR amplifications were performed in reactions containing genomic DNA buffer, dNTPs, fluorescent-tagged primers and DNA polymerase. PCR conditions consisted of an extended initial denaturing cycle, followed by 35 denaturing/annealing/extension cycles. PCR products were analyzed using a 3730 DNA

Analyzer fluorescence detecting capillary electrophoresis platform according to manufacturers’ instructions (Thermo Fisher Scientific Inc., Waltham, USA).

4.3.2.3. DNA extraction-SNP

Genomic DNA was isolated from crushed seeds. An extraction buffer was added to each sample and mixed using a mechanical shaker. Samples were placed in a 95oC oven for 30 minutes to incubate. Samples were cooled and then centrifuged for five minutes at 4000 rpm. Supernatant was transferred to storage containers containing neutralization buffer, sealed and stored at -20oC until used.

4.3.2.4. Generation of SNP data

SNP markers were developed in-house by DuPont Pioneer to meet the following criteria:

1) high level of polymorphism,

2) high level of repeatability in the SNP analysis,

110

3) widespread ability to discriminate among canola germplasm, and

4) eliminate or avoid hemi-SNPs in polyploids (B. napus is a dibasic allotetraploid).

The PCR reaction was conducted on Array TapeTM (LGC Douglas Scientific, London, UK) using neutralized DNA from storage containers along with a PCR Assay mix containing fluorescent probes and primer pair plus buffer. Sealed samples were then placed in a PCR water bath thermocycler for an initial denaturation cycle followed by 40 denaturing/annealing/extension cycles. Assay fluorescence was then read on a plate reader using optimal settings as per manufacturer’s instructions. Scored samples were compared to controls to validate and interpret.

4.3.3. Linkage map construction

Linkage map construction was first analyzed in the two DH populations using SSR markers selected with polymorphisms between the parental lines in Population J10-02 (1147-01 and

1147-03) and Population J10-11 (1147-01 and 1147-13). Due to sparse coverage, SNPs were added to the analysis using polymorphic parental SNPs and avoidance of hemi-SNPs. Genotypes and markers with non-segregating marker results were removed from analysis. Linkage analysis using the combined markers was performed using JoinMap 4.0 (Van Ooijen, 2006). Using the

Maximum Likelihood mapping algorithm, the threshold for goodness-of-fit was set to 5.0, a recombination frequency of <0.4 and minimum logarithm of odds (LOD) scores (Morton, 1955) of 2.0. All genetic distances were expressed in centimorgan (cM) as derived by the Kosambi function (Kosambi, 1944). Segregation of alleles in the DH populations was analyzed by a chi- square (χ2) test for “goodness-of-fit”. A 1:1 ratio was expected for Mendelian ratios in DH

111

populations. Markers that showed significant segregation distortion (=0.05) were removed and not used for linkage mapping. As a result, the markers were reduced in Population J10-02 from 247 to 226 and in Population J10-11 from 408 to 354.

QTL analysis was performed by composite interval mapping (CIM) (Zeng, 1994) using the

Windows version of QTL cartographer 2.5 software (Wang, et al., 2012). Backward regression analysis was applied for QTL detection. Cofactors were selected by the program using Model 6.

Background controls were set at five markers, window size set at 10.0 cM and probability for into and out set at 0.1. Significance thresholds to control the rate of Type 1 errors (false positives) at the 0.05 significance level were estimated on the basis of 1000 permutations using the procedure as described previously (Churchill and Doerge, 1994). Table 4.1. shows the threshold values calculated for each trait and population. Walk speed was set at 1.0 cM. The confidence interval of QTL was determined by 1-LOD intervals surrounding the QTL peak.

4.4 RESULTS AND DISCUSSION

4.4.1. Phenotyping

4.4.1.1. Indoor assay

Figures 4.5. and 4.6. show the distribution of the bite traits collected for each population and

Tables 4.2. and 4.3. show the statistical summaries for each DH population evaluated. Normal distribution curves in Figures 4.5. and 4.6. were calculated by SAS (SAS, 2012) based on sample mean and sample standard deviation. Results indicated a skewed distribution in DH populations

J10-02, skewed only due to no scores having less than zero bites. DH population J10-11 showed

112

normal distribution of scores in all three bite measurements. In DH Population J10-02, replicates and entries indicated at least one entry was significantly different ( = 0.05) when compared to other entries for resistance to flea beetle damage in all three feeding metrics (Table 4.2.). There was no significant interaction ( = 0.05) of replicate-by-entry so it was removed from the equation.

Table 4.3. shows the ANOVA summary for the three flea beetle bite measurements for Population

J10-11. There were significant differences observed in entries for both test bites and feeding bites but not total bites. Replicates were significant for test bites and total bites but not for feeding bites. Both populations exhibited high C.V.s in all bite measurements (greater than 65). The R- squared value for all three bite measurements in population J10-02 was approximately 0.25 whereas in population J10-11 it was between 0.35 to 0.46. Extreme rankings (best or worst for resistance) appear to be dominated by extremes in either test bites or feeding bites and not favouring one bite type. Further investigations may reveal if high test bites but low feeding bites is real and what mechanisms make them unfavourable after test bites compared to lines identified with high or low test bites and high feeding bites. Full bites would be more damaging to flea beetle herbivory although a high number of test bites under hot dry conditions can be very damaging to a canola crop as well.

High experimental variation may be due a number of reasons around cotyledon analysis. It was observed that depending on seed size, cotyledons were not equally sized. As a result, larger cotyledons may have higher counts of bites compared to a line with equal genetic resistance but

113

smaller cotyledon size. Percent area affected may be a better way to score the feeding pressure to eliminate the absolute surface size or report the numbers are a ratio to surface area. Another observation was the relative time a seed takes to emerge and have fully opened cotyledons across all entries. Seed quality may delay or enhance seed vigour and alter comparisons. Both of these observations were addressed where possible but due to limited seed quantities, standard seed size and equal sized seeding emergence was not possible in most entries. Cells were double seeded and thinned prior to flea beetle exposure to try and standardize the cotyledon size and reduce the number of missing samples. Other protocols by Bodnaryk et al. (1994), Pachagounder et al. (1998), and Soroka and Grenkow (2013) planted several individual pots per entry and selected the most uniform cotyledons for exposing to and evaluating flea beetle feeding. Check seed was adjusted for seed size to be more equal to tested entries, especially noted in the S. alba seed lots which was generally a larger seed and more vigorous of a plant relative to B. napus entries.

Both Populations showed transgressive segregation within the populations as both sets of populations had progeny beyond the parental lines for bite traits (Figures 4.6. and 4.7.). In both cases, the check variety 46A65 was in the middle of the overall feeding distribution, with fewer feeding bites but more test bites than surrounding neighbours with similar total bites. Figures

4.6. and 4.7. also show the variability in feeding bites to test bites, which may indicate that different mechanisms of resistance or attraction are active. The concern with test bites is the inability to distinguish between full test bites or interrupted feeding that may occur at the end of the 24-hour exposure period. As reported by Henderson et al. (2004), initial test bites under

114

favourable conditions results in continued feeding. Types of bites were used to determine if there was any antixenosis based on ingestion. Antixenosis has been noted amongst various species of

Brassica and varietal differences within species (Bodnaryk and Lamb, 1991; Lamb et al., 1993).

Full tissue removal was noted more around the perimeter of the cotyledon and was consistent with reports by Soroka (2011). Variation within species has been reported by Gavloski et al.

(2000) and Giamoustaris and Mithen (1995) with glucosinolate concentration and ratios possibly playing a role whereas Bodnaryk and Palamiswamy (1990) suggested glucosinolates have a limited effect on feeding preference within Brassicas. Henderson et al. (2004) suggested the presence of repellent or the absence of stimulatory volatile phytochemicals on the leaf surface may also play a role. Ultimately, feeding by the flea beetle depends on whether or not sufficient stimuli are present to attract or repel the herbivore (Henderson et al., 2004). These factors have not yet been determined, but genetic variability appears to exist within the Brassicas.

4.4.1.2. Outdoor assay

In 2013, there was severe variation in natural feeding pressure from one corner of the field versus the opposing corner. For this reason, data has not been presented as the data was considered invalid as this field variation could not be accounted for. This is one of the challenges of working with a wild population of insects with limited possibilities to control feeding pressure.

Entries were tested again in 2015 using a different approach due to seed restrictions and concerns about spatial variation as seen in 2013. As noted in Figure 4.9, the distribution of feedings in the outdoor evaluation showed a range of feeding scores. Transgressive segregation

115

was noted as with the indoor assay as there were progeny beyond the levels seen in either parent in both cases. In 2016, the experiment was repeated, however, due to some emergence issues and high flea beetle pressure, a poor establishment resulted in the site being discarded.

Table 4.4. shows the ANOVA summaries for both Population J10-02 and J10-11 in the 2015 field trial. In both populations, replicates showed significant differences but no significant differences within the populations. The variation due to replicates may be explained by the sampling method. Two teams were assigned two replicates each. There were no significant differences between the replicates each team scored but the two teams were not calibrated with each other as well as they should so slightly different scores were assigned by each group.

4.4.2. Construction of genetic maps

Individual genetic maps of the two DH populations Population J10-02 and Population J10-11 were constructed separately using software JoinMap 4.0 (Van Ooijen, 2006) and winQTL cartographer

(Wang et al., 2012). This resulted in 17 linkage groups (LG) in Population J10-02 and 19 linkage groups in Population J10-11. Based on DuPont Pioneer’s high density map, the linkage groups were named using conventional LG nomenclature for Brassica napus (Parkin et al., 1995). In

Population J10-02, two smaller linkage groups were consistent with parts of the Pioneer linkage group N2, hence were named N2a and N2b. Based on the DuPont Pioneer map, there is approximately 54cM between LG N2a and N2b. This gap maybe explained by the lack of polymorphic markers in the region. As sister lines were the parental lines, a lower number of

116

polymorphic markers were identified in this population. The two DH mapping populations

Population J10-02 and Population J10-11 provided map lengths of 976.1cM and 1631cM respectively with marker densities of 4.5cM and average density of 5.1cM between genetic loci

(Table 4.5.). The reduced map lengths and marker density differ to Raman et al. (2013) where their consensus map suggests a map length of 1987.2cM with an average marker density of

1.46cM between markers based on 1359 markers and Piquemal et al. (2005) with map length of

2619cM and average of 7.36cM between 363 genetic loci. The results reported here are within the range reported for other studies (Piquemal et al., 2005; Raman et al., 2013). However, there was no surprise that the sister parental line Population J10-02 had fewer polymorphisms compared to Population J10-11 with unrelated parents which likely resulted in the reduced map length.

Table 4.6. and Figures 4.9. and 4.10. show tabular and graphical representations of the LOD score profile for each of the two DH populations. The graphs have two parts; the upper part with the

LOD scores graphically represented and the lower part with additivity effect graphs which indicate which parent contributed the allele. The three indoor bite scores were included along with two field row scores; noted as % field damage (unadjusted) and spatially adjusted % field damage scores as BLUEs. Peaks greater than the LOD score threshold, as noted in Table 4.1., were considered significant (=0.05) for that trait and population. As the five traits listed are all measures of flea beetle feeding, the trends are generally overlapping, whether they were significant or not. In the case of the indoor assay, it may be possible to determine if test bites versus feeding bites trigger different sensory mechanisms in the flea beetle to make the host

117

plant more or less desirable. From there, future studies may determine how the feeding trigger may be controlled at a particular locus.

In Population J10-02, four peaks appear to be significant as seen in Figure 4.9. and Table 4.6. The

R2 values are based on the proportion of variance explained by the significant QTL at a test site with the estimated parameters (Wang et al., 2012). For linkage group N4, both outdoor screenings overlap and are significant, explaining 13% and 17% of the phenotypic variation observed (R2=0.13 and 0.17, respectively) with a second lower but non-significant peak (R2=0.11) on the same linkage group. The additive effect contributed to by the 1147-03 parent (negative values because this parent shows resistance with lower % damage). The second non-significant peak had an additive effect contributed to by parent 1147-01 (positive values) as noted by the lower additivity effects graph below the main logarithm of odds (LOD) score graph in Figure 4.9.

Linkage group N6 has peaks with additive effect due to parent 1147-01 for all indoor bite scores, however only full bites were significant (R2=0.11). Linkage group N13 has three peaks of interest, of which one is significant (R2=0.12) for BLUEs of percent feeding damage with additive effect coming from parent 1147-03. The other two LOD scores of interest are overlapping of feeding bites and total bites, both with additive effects due to parent 1147-01 (R2=0.09). In LG N14, a peak of interest for feeding bites was identified (R2=0.09), with allelic additive effect due to parent 1147-01. Significant peaks were identified as QTL and highlighted on the genetic linkage maps shown in Figure 4.11. Close-ups of the individual linkage groups are in the supplemental figures (Appendix-Figure A4.1).

118

In Population J10-11, four significant peaks were identified using the thresholds listed in Table

4.1. Several peaks of interest with LOD scores below the thresholds listed in Table 4.1. but greater than 2.0 were noted as seen in Figure 4.10. and Table 4.6.. In LG N13, there are two significant peaks. The observed values contribute to a significant peak (R2=0.13) attributed to parent 1147-

13 with smaller non-significant peak (R2=0.12) of interest for BLUEs nearby. Towards the other end of LG N13 is a significant peak (R2=0.11) for the full bites trait from parent 1147-13. The feeding bite trait shows a peak within the bites peak but it is not significant (R2=0.12). Linkage group 15 shows a significant QTL (R2=0.12) for full feeding bites. Linkage group N17 has a significant peak (R2=0.16) for the full bites trait from parent 1147-13. Significant peaks were identified as QTL and highlighted on the genetic linkage maps shown in Figure 4.12a. and 4.12b. and Table 4.6.. Close-ups of the individual linkage groups are in the supplemental figures

(Appendix-Figure A4.2).

In Population J10-11, other peaks of interest greater than 2.0 LOD score threshold that were not significant include one on LG N3 for feeding bites (R2=0.07). Linkage group N5 has a non- significant peak (R2=0.15) for total number of feeding bites. LG N7 showed a double peak for the feeding bites trait (R2=0.10 and 0.18). LG N11 has non-significant peaks of interest for the total bites trait (R2=0.08) with a smaller bites peak within it. LG N13 has a small peak (R2=0.15) for the

BLUEs field trait. A non-significant peak (R2=0.08) was identified in LG N14 for percent field damage, with a second non-significant peak (R2=0.09) further along the chromosome for test bites. Test bites show a non-significant peak on N18 (R2=0.11). The non-significant peaks of interest would require further studies to determine if there is any significance to be gathered

119

from those regions, particularly those that were just below the threshold or where all traits were peaking in the same area and are narrow. As host plant response to flea beetle herbivory is a complex trait, it is not surprising to see so many smaller peaks that are masked due to nearby stronger QTL effect or challenges associated with experiment design or phenotyping.

Between the two populations, significant peaks did not match well. However, LG N13 was a common chromosome where significant peaks and several non-significant peaks of interest were identified from either population. Linkage groups, such as N4, N7, N13 and N14, minor and major peaks in either population did overlap. As the final population sizes were small and the final number of markers relatively small, these peaks need to be investigated further. Small QTL are best analyzed with larger sample sizes where n=100 (standard recommendation). Increasing the sample size, n=300, is superior but beyond n=300 is not significantly better for preliminary genetic mapping (Vales et al., 2005). Larger sample sizes are recommended for high-resolution mapping with techniques available to reduce the sample size without jeopardizing mapping quality (Churchill and Doerge, 1994; Collard et al., 2005; Mammadov et al., 2012). Data quality is essential for optimizing estimates of QTL number, effect and interactions (Vales et al., 2005). Flea beetle feedings are challenging and subject to high levels of variation as the plant does not have the time to compensate compared to traits evaluated at a later plant growth stage.

In an attempt to understand the data better, the 2015 field experiment was set up as one large

RCBD design with a separate repeated three entry check experiment embedded within for spatial adjustment. The same dataset was used to calculate best linear unbiased estimates (BLUEs) and

120

RCBD averages with the difference being whether of not diagonal checks were included. As mentioned previously by Vales et al. (2005), data quality is essential for valid molecular tools. It would be expected that the same data when analyzed two ways should yield similar results. In most cases, this was observed, however, the magnitude of the differences varied. As a result, some QTL may or may not have been considered significant if only one technique was used to analyze the data and identify QTL.

Beyond this study, QTL for insect herbivory in canola are limited. Seven QTL related to diamondback moth herbivory, were reported by Asghari et al. (2009). One QTL was reported by

Lee et al. (2014) as being associated with cabbage seedpod weevil. Two QTL associated with root maggot feeding were reported by Ekuere et al. (2005). Several other herbivory response QTLs have been reported in related species such as Arabidopsis spp. and Barbarea spp., primarily associated with glucosinolates (Agerbirk et al, 2003; Feng et al., 2012; Zhao and Meng, 2003).

Current flea beetle work also includes the related Brassica species Barbarea vulgaris (Kuzina et al., 2011; Nielsen et al., 2010). It should be noted the flea beetle P. nemorum which feeds on

Barbarea spp. is most damaging as a tunneling larva that feeds on the leaves of the Barbarea spp. plants versus the chewing damage of the canola flea beetle adult. Work by Gruber et al. (2012) compared transcriptomes of undamaged and flea beetle damaged B. napus cotyledons. They were able to identify that the transcriptome was moving toward a defensive state after damage by flea beetles. This resulted in activation or suppression of genes controlling stress responses, cell wall synthesis, primary and secondary metabolism pathways and transport. A defense

121

response by the plant involves a number of zinc finger proteins and calcium-dependent proteins.

Gruber et al. (2012) show the complexity of the response to flea beetle herbivory.

4.5 CONCLUSION

This is the first report of QTL for flea beetle feeding resistance in B. napus. Through several methods and attempts, six unique QTL were shown to be associated with flea beetle herbivory on B. napus winter-type material. By comparing outdoor and indoor screening, it was shown that most measured responses for the trait evaluated resulted in similar peaks. As all traits are related to flea beetle herbivory, one would expect a large amount of overlap in the results. Between the two populations, there were a number of similarities as well as differences. It is clear that flea beetle herbivory is a complex issue to tackle. This can be noted in Figures 4.6. and 4.7. where the proportion of test bites to feeding bites is not standard across all entries. Some entries exhibited more test bites and less full bites and vice versa. This alone, warrants further investigation to confirm the above findings and determine the underlying cause of flea beetle response in the B. napus plants upon the start of feeding.

There was a common parent between the two DH populations. As such, it would have been reasonable to have expected a relatively large commonality between the two populations, which wasn’t the case. Several reasons could be responsible for this result such as segregation distortion (Foissr et al., 1993), limited population sizes (Vales et al., 2005) and a limited number of polymorphic markers (Collard and Mackill, 2008).

122

Segregation distortion violates the laws of segregation. Mendelian genetics show a predictable transmission of alleles and formation of genotypes. Deviation from this law violates conventional

Mendelian genetics and renders analysis invalid (Lu et al., 2002). Segregation distortion has been reported particularly when non-adapted exotic germplasm is used (Kopisch-Obuch and Diers,

2006). Introgression of exotic germplasm usually carries non-target genes resulting in linkage drag or pleiotropic associations with the target genes (Kopisch-Obuch and Diers, 2006). The germplasm used in this study did not include any exotic germplasm but segregation distortion can still occur. Segregation distortion may not be evenly distributed throughout the genomes but can be concentrated on particular chromosomes or regions (Lu et al., 2002). Segregation distortion can result from environmental sensitivities particularly with the doubled haploid process where gametes from one parental type are favoured in the DH process (Rajcan et al.,

1999), where highly conserved regions accumulate a number of similar alleles. Furthermore, polyploidy may create duplicates of these conserved regions and not allow for polymorphic markers to be detected (Li et al., 2016). It has been suggested that there is a high degree of conserved variants from the ancestral genome that is shared amongst the Brassicaceae (Smooker et al., 2011). Lines and molecular markers showing segregation distortion were removed from analysis.

The number of polymorphic markers used in this study may be considered low. As B. napus is a allopolyploid crop, it is more complicated to identify polymorphic markers. Markers need to be able to discriminate within the genomes (useful) and not between the genomes (not useful)

(Ganal et al., 2009). A more saturated map may identify more QTL or reduce the length of the

123

QTL region. By increasing the number of markers and increasing the population sizes to 200

(Ferreira et al., 2006), would allow for a more saturated map and possibly more or stronger QTL.

Further studies into the molecular mechanisms of flea beetle herbivory could include fine- mapping of the QTL identified in this dissertation. Non-significant peaks of interest may be associated with other traits that make up the complex herbivore defense system. Further to that, transcriptomic and metabolomics studies to complement Gruber et al. (2012) would further enhance the understanding and use of flea beetle resistance in Brassica species.

It is not known as to what the genes or their function may be involved in the identified QTL regions. It is quite possible that they are involved in one or more of the defense mechanisms in play as described in chapter one. Some may be involved in general plant defense response to insects as well as pathogens, where as others may be strictly for flea beetle resistance in canola.

By identifying the several QTLs along with numerous minor QTLs, specific chromosomes and chromosomal regions can be targeted for further studies. By evaluating the feeding patterns more in depth, a better understanding of the pest may also provide more details on where and how to proceed in efforts to develop improved flea beetle resistance.

124

Table 4.1. Threshold values (LOD scores) for all traits in each population. Estimated with permutation analysis after 1000 iterations and significance level of 0.05. Population Population Trait J10-02 J10-11 BLUEs 2.8 3.4 % field damage 2.7 3.4 Feeding bites 2.7 3.1 Test bites 2.9 3.2 Total bites 2.7 3.3

Table 4.2. Summary of statistics for DH Population J10-02 flea beetle feeding in arenas.

Feeding Bite Test Bite Total Bites df MS Pr>F df MS Pr>F df MS Pr>F Model 247 829.88 <.0001 247 148.57 <.0001 247 1475.89 <.0001 Rep 9 9079.87 <.0001 9 987.82 <.0001 9 23.38 <.0001 Entry 238 517.43 0.0004 238 116.85 <.0001 238 1.36 0.0006 Error 1674 377.63 1698 60.13 1698 692.79 R2 0.245 0.264 0.237 C.V. 90.070 116.740 90.340

Table 4.3. Summary of statistics for Population J10-11 flea beetle feeding in arenas.

Feeding Bite Test Bite Total Bites df MS Pr>F df MS Pr>F df MS Pr>F Model 88 296.97 0.0012 88 141.87 0.0347 88 420.10 0.2498 Rep 2 359.23 0.1275 2 396.56 0.0226 2 1505.81 0.0192 Entry 86 293.50 0.0016 86 135.62 0.0600 86 390.16 0.3936 Error 178 172.44 178 102.39 178 372.59 R2 0.460 0.407 0.358 C.V. 67.412 100.850 65.403

125

Table 4.4. Summary of statistics for Populations J10-02 and J10-11 flea beetle feeding in 2015 field trial near Saskatoon, SK.

J10-02 J10-11 df MS Pr>F df MS Pr>F Model 219 989.95 <.0001 155 978.94 <.0001 Rep 3 38578.19 <.0001 3 29721.64 <.0001 Entry 216 430.51 0.2397 152 393.53 0.7523 Error 544 398.12 379 433.16 R2 0.500 0.480 C.V. 37.441 40.305

126

Table 4.5. Marker distribution by linkage group (LG) and marker type (SSR or SNP) for flea beetle damage for indoor and outdoor screening on two winter-type Brassica napus DH populations.

POPULATION J10-02 11-47-01 x 1147-03 (sister lines) Final population size for mapping: 80 entries

Grand Chromosome N1 N2a N2b N3 N4 N5 N6 N7 N8 N9 N10 N11 N12 N13 N14 N15 N16 N17 N18 N19 Total

# SSR Markers 4 0 0 4 5 2 5 0 7 1 2 2 3 5 1 1 0 2 0 0 44 # SNP 0 4 3 25 15 0 7 0 25 3 0 12 11 36 19 6 0 11 5 0 181 Markers Total Markers 4 4 3 29 20 2 12 0 32 4 2 14 14 41 20 7 0 13 5 0 226

LGL 45.0 17.3 11.9 129.7 62.3 1.5 112.7 n/a 74.5 8.1 6.7 61.5 45.5 206.1 81.9 35.6 n/a 51.0 24.8 n/a 976.1 MD 11.3 4.3 4.0 4.5 3.1 0.8 9.4 n/a 2.3 2.0 3.4 4.4 3.3 5.0 4.1 5.1 n/a 3.9 5.0 n/a 4.5 No. marker 1 1 0 4 1 0 2 n/a 2 0 0 1 2 5 3 1 0 0 1 n/a gaps >10cM Grand total 976.1 cM covered with 226 markers with average of 4.5cM between markers

POPULATION J10-11 1147-01 x 1147-13 (non sister lines) Final population size for mapping: 105 entries

Grand Chromosome N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N11 N12 N13 N14 N15 N16 N17 N18 N19 Total

# SSR Markers 10 3 8 3 5 5 0 3 2 0 1 3 5 5 4 0 4 2 1 64 # SNP 18 5 37 15 22 18 16 19 5 7 8 12 40 24 8 5 14 10 7 290 Markers Total Markers 28 8 45 18 27 23 16 22 7 7 9 15 45 29 12 5 18 12 8 354

LGL 123 88.6 132.8 51 85.7 122.3 101.5 65.9 32.9 29.6 52.5 89.5 154.2 155.3 137.2 9.1 131.5 40.1 28.5 1631 MD 4.4 11.1 3.0 2.8 3.2 5.3 6.3 3.0 4.7 4.2 5.8 6.0 3.4 5.4 11.4 1.8 7.3 3.3 3.6 5.1 No. marker 2 3 2 1 1 3 3 2 2 1 3 3 5 3 4 0 4 1 0 gaps >10cM Grand total 1631cM covered with 354 markers; with average of 5.1 cM between markers LGL: linkage group length; MD: average marker density in cM of combined markers

127

Table 4.6. Major (bold font) and minor (regular font) QTL summary in the two DH mapping populations for flea beetle herbivory in winter-type canola.

Location Length Additive Threshold Population Trait ~ LG^ (Peak) LOD† R2 ¶ cM+ effect ‡ cM # cM % damage in N04 16.64 13.87 4.15 -5.52 0.17 2.7 field BLUEs N04 16.64 4.97 3.21 -3.77 0.13 2.8 % damage in J10-02 N04 51.90 22.81 2.56 4.04 0.11 field Feeding 1147-01 x bites N06 31.50 22.05 2.87 4.24 0.11 2.7 1147-03 Feeding bites N13 65.38 10.43 2.39 3.89 0.09 Total bites N13 65.38 5.43 1.98 5.41 0.09 BLUEs N13 188.90 1.33 2.97 -5.78 0.12 2.8 Feeding bites N14 18.90 10.27 2.09 3.74 0.09 Feeding bites N03 116.07 2.03 2.05 -2.64 0.07 Total bites N05 63.55 1.00 2.50 4.40 0.15 Feeding bites N07 9.53 3.53 3.00 -3.29 0.10 Feeding bites N07 14.66 2.20 2.31 5.24 0.18 Total bites N11 36.83 1.40 2.13 3.31 0.08 BLUEs N13 7.05 7.04 2.04 3.82 0.15 % field J10-11 damage N13 34.91 2.00 3.67 -5.77 0.13 3.4 BLUEs N13 41.84 2.00 3.22 -3.18 0.12 Feeding bites N13 96.24 7.04 1.98 5.12 0.12 1147-01 x Feeding 1147-13 bites N13 118.35 0.60 3.12 -3.49 0.11 3.1 % field damage N14 94.68 1.40 2.31 -3.08 0.08 Test bites N14 127.59 4.00 2.23 2.03 0.09 Feeding bites N15 137.17 1.00 3.17 -3.59 0.12 3.1 Feeding bites N17 78.54 5.54 3.41 -4.00 0.16 3.1 Test bites N18 27.39 3.03 2.72 2.43 0.11 ~ Traits are based on flea beetle injury scores. % field damage was the average score of 5 single plants in a row in a 4 replicate RCBD trial in 2015; BLUEs were the spatially adjusted average scores from the 2015 field trial and then converted to a best linear unbiased estimate (BLUE); Feeding bites were scored based on the number of feeding holes where both the upper and lower tissue was removed; Test bites are the number of feedings were either the upper or lower epidermal tissue was removed; Total bites is the combined total of feeding bites and test bites. The three bite scores were from indoor evaluations. Peaks greater than the threshold are bolded.

128

^ LG is the linkage group designation for Brassica napus (n=19) + Length in centriMorgans (cM) is the QTL interval length that exceeded the threshold as estimated by permutation analysis of each trait using 1000 iterations. † LOD: log of the odds score. To convert likelihood ratio (LR) to LOD values, LOD=0.217(LR). ‡ Additive effect indicates which parental allele causes an increase in the trait value. Positive values indicate that the susceptible parent 1147-01 allele increases trait values, and negative values indicate that the resistant parent, either 1147-03 or 1147-13, allele increases trait values ¶ R2 indicated the percentage of phenotypic variance in the mapping population explained by the detected QTL # LOD scores of QTL listed exceed the estimated threshold listed. Blanks indicate QTL that ranged between 2.0 and the threshold and are considered minor (not significant) and peaks of interest.

129

Test Bites

Feeding Bite

Figure 4.2. Visual representation of feeding bite versus test bite

a .

Figure 4.3a. Crepe paper seed tape where glue stick was used to glue 5 canola seeds 2.5cm apart to the seed tape with 20cm between plots.

130

b .

c .

Figure 4.3b. Planting of crepe paper seed tape in 2015. 4.3c. Trial site after scoring once plants were at the 2-4 leaf stage (plant stage in image above about 3 weeks after scoring).

131

A B C A A A

. . . Percent

Number of Bites Figure 4.4. Distribution of flea beetle feeding bites (A), test bites (B) and total bites (C) in arenas for Population J10-02 with a normal distribution line graph overlaid.

A B C A A A

. . . Percent

Number of Bites Figure 4.5. Distribution of flea beetle bites (A), test bites (B) and total bites (C) in arenas for Population J10-11 with normal distribution line graph overlaid.

132

80 1147-01 70

50 Total 40 46A65 Test 30 Feed 1147-03 Flea Flea InjuryBeetleSores 20

10 Averageg Averageg NumberofBites Taken per Plant

0 Entries

Figure 4.6. Distribution of number of bites in DH Population J10-02 including parental lines 1147- 03 and 1147-01 and B. napus check 46A65 evaluated in indoor arena feeding study.

40 46A65 TestingTest Bites Bite 4 30 FeedingFeeding Bite Bites

20 Flea Flea InjuryBeetleSores

10 Averageg Averageg NumberofBites Taken per Plant 0 Entries

Figure 4.7. Distribution of number of bites in DH Population J10-11 including B. napus check 46A65 evaluated in indoor arena feeding study.

133

134

BLUEs Avg All reps Feeding Bites Test Bites Total Bites

N1 N2a N2b N3 N4 N5 N6 N8 N9 N10 N11 N12 N13 N14 N15 N17 N18

Figure 4.9. Graph showing the LOD score profile for DH Population J10-02 with the traits BLUEs*, average injury score*, feeding bites**, test bites** and total bites** in field and indoor screening. (*=field based score, **=indoor based score). Lower graph shows the additive effects of the traits at each allele. Feeding bites have all plant tissue removed within feeding area, test bites have only epidermal layer removed, total bites are total feeding and test bites. No peaks for N7, N16 and N19 due to the lack of polymorphic SSR and SNP markers .

135

BLUEs Avg All reps Feeding Bites Test Bites Total Bites

N1 N2 N3 N4 N5 N6 N7 N8 N9 N10 N11 N12 N13 N14 N15 N16 N17 N18 N19

Figure 4.10. Graph showing the LOD score profile for DH Population J10-11 with the traits BLUEs*, average injury score*, feeding bites**, test bites** and total bites** in field and indoor screening. (*=field based score, **=indoor based score). Lower graph shows the additive effects of the traits at each allele. Feeding bites have all plant tissue removed within feeding area, test bites have only epidermal layer removed, total bites are total feeding and test bites.

136

Figure 4.11. Genetic linkage map and the locations of QTL for flea beetle damage in winter-type Brassica napus DH Population J10-02 using SSR and SNP markers. There are 17 linkage groups identified using JoinMap (Van Ooijen, 2006)and DuPont Pioneer’s genetic map and are represented by vertical bars. Marker names are listed to the right of the linkage groups with the position in centimorgans (cM) listed to the left side. The four identified QTL (two overlapping on N04, one on N06 and one on N13) associated with flea beetle injury were indicated by red bars within the linkage group. There are no linkage maps for LG N7, N16 and N19 due to lack of polymorphic markers.

137

Figure 4.12a.

138

139

CHAPTER 5 – GENERAL DISCUSSION AND CONCLUSIONS

The purpose of this thesis was to further our understanding of the interaction of flea beetles on spring-type and winter-type canola and identify QTL associated with flea beetle herbivory. Flea beetles cost farmers hundreds of millions of dollars in pest control and yield loss in western

Canadian agriculture for this economically important crop (Canola Council of Canada, 2016). Flea beetles are currently controlled with seed treatments and foliar sprays as needed. With further pressure from the public regarding seed treatments including pesticides and restrictions on GMO technology in certain jurisdictions around the world, natural genetic variation for insect resistance is of interest. The objectives of this thesis were to: 1) evaluate Brassica napus winter- type (vernalization requiring) breeding germplasm in response to flea beetle herbivory; 2) evaluate Brassica napus spring-type (non-vernalizing) breeding germplasm in response to flea beetle herbivory; 3) identify molecular markers that tag QTL associated with flea beetle herbivory.

In the first study of winter-type lines, significant (=0.05) results for treatment and entry were found in the 2009 split-plot design trials (Table 2.4.). When all untreated RCBD trials were combined, significant (=0.05) values for entry, year and location were found (Table 2.5.). As expected, Ace (Sinapis alba) and seed-treated 46A65 checks ranked higher than all other entries for resistance to flea beetles (Figure 2.7.). Levels of herbivory were affected by the year but the year to year trend was generally the same. There was no significant difference between fall or spring planted trials either, indicating that the flea beetle life cycle does not change the flea

140

beetle feeding habits. As such, the possibility of moving resistance genes between spring and winter-type germplasm is plausible. It also indicates that the resistance mechanisms involved may be limited even though winter and spring-type canola have evolved independently due to geographies and vernalisation requirements. Making wide intraspecific crosses involving spring and winter-type B. napus germplasm, for flea beetle resistance, may be less desirable compared to interspecific crosses with other Brassicas, due to limited variation within B. napus as seen in

Tables 2.3 and 3.3. Variation within and amongst winter families and checks evaluated was observed, supporting previous reports of genetic variation within the Brassicaceae (Figure 2.7.)

(Lamb et al., 1991; Gavloski et al., 2000).

In the second study on spring-type lines, significant (=0.05) results for treatment and entry were found (Table 3.4.). When all untreated RCBD trials were combined, entry, year and location were considered significant (=0.05) (Table 3.5.). These results were similar to the winter-type results from the first study. Year (Figure 3.6.) and seasonal effects (Figure 3.7.) of spring versus fall planted trials were not significant (=0.05). In addition, variation among spring lines and checks evaluated was observed, which supported previous reports of genetic variation with the

Brassicaceae (Lamb et al., 1991; Gavloski et al., 2000) and also seen in Chapter 2.

In order to develop canola with improved flea beetle resistance, large populations under uniform exposure to flea beetles along with multiples plant breeding cycles would be required to capitalize on the natural genetic variation found within the Brassicas. Genetic variation is found within the species but greater gains could be possible using interspecific crosses. As resistance

141

appears to be a quantitative trait, it would prove challenging to use modern technologies such as

CRISPR-Cas without first identifying genes or pathways involved with flea beetle resistance.

Alternate options could include genome selection where the specific gene(s) or pathway are not actually identified but the strength of the genetic makeup for the desired trait such as flea beetle resistance, allows a package of genetic building blocks to move together. After several cycles, these building blocks can be identified as they are either gathered or removed in making a superior product better tolerant to flea beetle herbivory. Finally, whether or not the genetic variation within the Brassicaceae is great enough to obtain field scale resistance to flea beetle herbivory has yet to be shown.

The third study involved QTL mapping of flea beetle resistance. Two winter-type DH populations were evaluated under field and indoor conditions for flea beetle injury. Both populations exhibited transgressive segregation (Figures 4.7. and 4.8.) as expected. Population J10-02 was derived from a sister line cross whereas Population J10-11 was result of a cross with unrelated parents. Both populations shared a common parent 1147-01 and was considered as the susceptible source in these crosses. As Population J10-02 had sister line parents, the number of polymorphic markers available was fewer resulting in a modest total linkage map length (Table

4.4.). Both SSR and SNP markers were used to develop linkage maps and identify QTL regions associated with flea beetle herbivory. Six QTL were identified with three from each population

(Figures 4.12. and 4.13.). Of these, only linkage group N13 had a QTL from each population but not the same genomic region. The other QTL were located on N4, N6 and N16. N13 also had a second QTL identified in Population J10-11. The identified individual QTL explained 11% to 17%

142

of the damage associated with flea beetle herbivory. Along with the 6 QTL identified, there were several non-significant peaks of interest. Within populations, most peaks overlapped for all flea beetle traits measured, however, there were two peaks where test bites were much greater than other traits (Population J10-11, LGs N1 and N6 – Figure 4.11.). This may suggest a difference in attraction or repulsion of flea beetles to continue feeding. In Population J10-02, the BLUEs trait had a peak in N13 where the other traits did not have any indication of a peak (Figure 4.10.). This is a good example where analysis may provide false positives or negatives. It was to be expected for the BLUEs trait and the observed trait to provide similar results as the only difference was the method by which the feeding scores were analyzed.

The results of this thesis contribute to gaining a better understanding of the interaction between flea beetle and canola plants in the terms of flea beetle herbivory. The first two studies indicated that phenotyping of cotyledon resistance to flea beetle herbivory was season-dependent and that any genes identified in spring or winter-type germplasm may be effective in either habit type. It also indicated that novel resistance due to habit type may be limited within the B. napus germplasm but may be available in related species such as Sinapis alba. In addition, genomic regions associated with resistance to flea beetle herbivory have been identified for the first time.

The markers linked to QTL for flea beetle resistance may be used for marker assisted selection as part of a dedicated breeding strategy to enhance flea beetle resistance in canola.

In conclusion, further work is required to make meaningful gains in flea beetle resistance in canola. The QTL identified in this thesis are not known to be associated with any component

143

traits of flea beetle resistance, making it difficult to determine what pathways or mechanisms may be involved in the trait. Further confirmation of these QTL in other germplasm, particularly spring-type canola, would be a logical next step. In addition, it would be of interest to evaluate the response of canola to flea beetle herbivory using the “-omics” technology currently available.

Several reports suggested that activation or suppression of certain genes is key in insect resistance. Work by Gruber et al. (2012) suggested that the overall expression of cotyledon transcriptome is defensive; induced as a result of feeding. This defensive response affects stress responses, primary and secondary metabolite pathways, cell wall synthesis and transport.

Currently these responses are insufficient in the defense against flea beetles. Once a better understanding of the physiological response is achieved, it may allow for up- or down regulation of the associated genes to achieve the level of response required to overcome flea beetle herbivory. Proving that this is viable at a field level is also another challenge yet to be determined.

In the meantime, if flea beetle resistance is sought in canola, a dedicated program using genome- wide association study (GWAS) and predictions on populations showing transgressive segregation may be the most successful path forward. Due to the narrow range of resistance in the native population, use of either recurrent selection over several breeding cycles within B. napus or wide crossing beyond B. napus would be required.

144

REFERENCES

Adel, M.M., F. Sehnal, and M. Jurzysta. 2000. Effects of alfalfa saponins on the the moth

Spodoptera littoralis. J. Chem. Ecol. 26(4): 1065–1078.

Agerbirk, N., M. Ørgaard, and J.K. Nielsen. 2003. Glucosinolates , flea beetle resistance , and leaf

pubescence as taxonomic characters in the genus Barbarea (Brassicaceae). Phytochemistry

63: 69–80.

Agerbirk, N., M. De Vos, J.H. Kim, and G. Jander. 2009. Indole glucosinolate breakdown and its

biological effects. Phytochem. Rev. 8(1): 101–120.

Aghajanzadeh, T., M.J. Hawkesford, and L.J. De Kok. 2014. The significance of glucosinolates for

sulfur storage in Brassicaceae seedlings. Front. Plant Sci. 5: 704.

Agrama, H., G. Widle, J. Reese, L. Campbell, and M. Tuinstra. 2002. Genetic mapping of QTLs

associated with greenbug resistance and tolerance in Sorghum bicolor. TAG Theoretical and

Applied Genetics 104(8): 1373-1378.

Agrawal, A. A. 1998. Induced responses to herbivory and increased plant performance. Science

279(5354): 1201–1202.

Agrawal, A.A. 2011. Current trends in the evolutionary ecology of plant defence. Funct. Ecol.

25(2): 420–432.

Ahmad, S. 1982. Roles of mixed-function oxidases in insect herbivory. p. 41. In Visser, J.H., Minks,

A.K. (eds.), Proceedings of the 5th International Symposium of Insect-Plant Relationships.

Pudoc, Centre for Agricultural Publishing and Documentation, Wageningen, The

Netherlands.

145

Ahuja, I., J. Rohloff, and A.M. Bones. 2010. Defence mechanisms of Brassicaceae : implications

for plant-insect interactions and potential for integrated pest management . A review.

Agron. Sustain. Dev. 30: 311–348.

Al-Doghairi, M. 1999. Pest management tactics for the western cabbage flea beetle (Phyllotreta

pussilla (Horn) on Brassica crops. (unpublished doctoral dissertation) Colorado State

University, Fort Collins, USA.

Alahakoon, U., J. Adamson, L. Grenkow, J. Soroka, P. Bonham-Smith, and M. Gruber. 2016. Field

growth traits and insect-host plant interactions of two transgenic canola (Brassicaceae) lines

with elevated trichome numbers. Can. Entomol. 148(5): 603–615.

Altieri, M.A., and L.L. Schmidt. 1986. Population trends and feeding preferences of flea beetles

(Phyllotreta cruciferae Goeze) in collard wild mustard mixtures. Crop Prot. 5(3): 170–175.

Andersen, J.F., and R.L. Metcalf. 1986. Identification of a volatile attractant for Diabrotica and

Acalymma spp. from blossoms of Cucurbita maxima Duchesne. J. Chem. Ecol. 12(3): 687–

699.

Anstey, T.H., and J.F. Moore. 1954. Inheritance of glossy foliage and cream petals in green

sprouting broccoli. J. Hered. 45(1252): 39–41.

Antwi, F.B., and G.V.P. Reddy. 2016. Efficacy of entomopathogenic nematodes and sprayable

polymer gel against crucifer flea beetle (Coleoptera: Chrysomelidae) on canola. J. Econ.

Entomol. 109(4): 1706–1712.

Arias, T., M.A. Beilstein, M. Tang, M.R. McKain, and J.C. Pires. 2014. Diversification times among

Brassica (Brassicaceae) crops suggest hybrid formation after 20 million years of divergence.

Am. J. Bot. 101(1): 86–91.

146

Asghari, A., A.A. Fathi, and S.A. Mohammadi. 2009. QTL analysis for diamondback moth

resistance in canola (Brassica napus L .). International Journal of Plant Production 3(3):29-

34.

Awmack, C.S., and S.R. Leather. 2002. Host plant quality and fecundity in herbivorous insects.

Annu. Rev. Entomol. 47(1): 817–844

Azhaguvel, P., D. Mornhinweg, D. Vidya-Saraswathi, J.C. Rudd, K. Chekhovskiy, M. Saha, T.J. Close,

L.S. Dahleen, and Y. Weng. 2014. Molecular mapping of greenbug (Schizaphis graminum)

resistance gene Rsg1 in barley. Plant Breed. 133(2): 227–233.

Badenes-Perez, F.R., J. Gershenzon, and D.G. Heckel. 2014. Insect attraction versus plant defense:

Young leaves high in glucosinolates stimulate oviposition by a specialist herbivore despite

poor larval survival due to high saponin content. PLoS One 9(4): 39–42.

Bain, A., and L. LeSage. 1998. A late seventeenth century occurence of Phyllotreta striolata

(Coleoptera: Chrysomelidae) in North America. Can. Entomol. 130: 715–719.

Balsbaugh, E.U.J., and K. Hays. 1972. Leaf Beetles of Alabama (Coleoptera: Chrysomelidae).

Auburn, AL.

Bartelt, R.J., B.W. Zilkowski, A.A. Cossé, U. Schnupf, K. Vermillion, and F.A. Momany. 2001. Male-

specific sesquiterpenes from Phyllotreta flea beetles. J. Nat. Prod. 27(12): 2397–2423.

Barth, C., and G. Jander. 2006. Arabidopsis myrosinases TGG1 and TGG2 have redundant function

in glucosinolate breakdown and insect defense. Plant J. 46(4): 549–562.

Bartlet, E., G. Kiddle, I. Williams, and R. Wallsgrove. 1999. Wound-induced increases in the

glucosinolate content of oilseed rape and their effect on subsequent herbivory by a crucifer

specialist. Entomol. Exp. Appl. 91(1): 163–167.

147

Bartlet, E., R. Mithen, and S.J. Clark. 1996. Feeding of the cabbage stem flea beetle Psylliodes

chrysocephala on high and low glucosinolate cultivars of oilseed rape. Entomol. Exp. Appl.

80: 87–89.

Bartumeus, F., and J. Catalan. 2009. Optimal search behavior and classic foraging theory. J. Phys.

Math. Theor. 42.

Bazakos, C., M. Hanemian, C. Trontin, J.M. Jiménez-Gómez, and O. Loudet. 2017. New strategies

and tools in quantitative genetics: How to go from the phenotype to the genotype. Annu.

Rev. Plant Biol. 68(1): 435–455.

Beck, S.D. 1974. Theoretical aspects of host plant specificity in insects. p. 290–311. In Maxwell,

E.G., Harris, F.. (eds.), Proceedings of the Summer Institute on Biological Control of Plant

Insects and Diseases. University Press of Mississippi, Jackson.

Beck, S.D., and L.M. Schoonhoven. 1980. Insect behaviour and plant resistance. p. 115. In

Maxwell, F.G., Jennings, P.R. (eds.), Breeding Plants Resistant to Insects. John Wiley & Sons,

New York.

Beckers, G.J.M., and S.H. Spoel. 2006. Fine-tuning plant defence signalling: Salicylate versus

jasmonate. Plant Biol. 8(1): 1–10.

Bellostas, N., J.C. Sorensen, and H. Sorensen. 2007. Profiling glucosinolates in vegetative and

reproductive tissues of four Brassica species of the U-triangle for their biofumigation. J. Sci.

Food Agric. 87: 1586–1594.

Beran, F., I. Mewis, R. Srinivasan, J. Svoboda, C. Vial, H. Mosimann, W. Boland, C. Büttner, C.

Ulrichs, B.S. Hansson, and A. Reinecke. 2011. Male Phyllotreta striolata (F.) produce an

aggregation pheromone: Identification of male-specific compounds and interaction with

148

host plant volatiles. J. Chem. Ecol. 37(1): 85–97.

Beran, F., Y. Pauchet, G. Kunert, M. Reichelt, N. Wielsch, H. Vogel, A. Reinecke, A. Svato , I. Mewis,

D. Schmid, S. Ramasamy, C. Ulrichs, B.S. Hansson, J. Gershenzon, and D.G. Heckel. 2014.

Phyllotreta striolata flea beetles use host plant defense compounds to create their own

glucosinolate-myrosinase system. Proc. Natl. Acad. Sci. 111(20): 7349–7354.

Bernays, E.A. 1998. Evolution of feeding behavior in insect herbivores. Bioscience 48(1): 35–44.

Bernays, E.A., and R.F. Chapman. 1994. Behavior: the impact of ecology and physiology. p. 166–

205. In Host-Plant Selection by Phytophagous Insects. Chapman & Hall, New York.

Bird, K.A., H. An, E. Gazave, M.A. Gore, J.C. Pires, L.D. Robertson, and J.A. Labate. 2017.

Population structure and phylogenetic relationships in a diverse panel of Brassica rapa L.

Front. Plant Sci. 8(3)1–12.

Bodnaryk, R.P. 1992a. Leaf epicuticular wax an antixenotic factor in Brassicaceae that affects the

rate and pattern of feeding of flea beetles Phyllotreta cruciferae Goeze. Can. J. Plant Sci. 72:

1295–1303.

Bodnaryk, R.P. 1992b. Effects of wounding on glucosinolates in the cotyledons of oilseed rape

and mustard. Phytochemistry 31(8): 2671–2677.

Bodnaryk, R.P. 1997. Will low-glucosinolate cultivars of the mustards Brassica juncea and Sinapis

alba be vulnerable to insect pests ? Can. J. Plant Sci. 77: 283–287.

Bodnaryk, R.P., and R.J. Lamb. 1991a. Influence of seed size in canola, Brassica napus L. and

mustard, Sinapis alba L., on seedling resistance against flea beetles, Phyllotreta cruciferae

(Goeze). Can. J. Plant Sci. 71: 397–404.

Bodnaryk, R.P., and R.J. Lamb. 1991b. Mechanisms of resistance to the flea beetle, Phyllotreta

149

cruciferae (Goeze), in mustard seedlings, Sinapis alba L. Can. J. Plant Sci. 71(1411): 13–20.

Bodnaryk, R.P., R.J. Lamb, and K.A. Pivnick. 1994. Resistance of hybrid canola (Brassica napus L.)

to flea beetle (Phyllotreta spp.) damage during early growth. Crop Prot. 13(7): 513–518.

Bodnaryk, R.P., and P. Palaniswamy. 1990. Glucosinolate levels in cotyledons of mustard, Brassica

juncea L. and rape, B. napus L. do not determine feeding rates of flea beetle, Phyllotreta

cruciferae (Goeze). J. Chem. Ecol. 16(9): 2735–2746.

Bodnaryk, R.P., and R.T. Rymerson. 1994. Effect of wounding and jasmonates on the physico-

chemical properties and flea beetle defence responses of canola seedlings, Brassica napus

L. Can. J. Plant Sci. 74: 899–907.

Boege, K., and R.J. Marquis. 2005. Facing herbivory as you grow up: The ontogeny of resistance

in plants. Trends Ecol. Evol. 20(8): 441–448.

Boerjan, B., D. Cardoen, R. Verdonck, J. Caers, and L. Schoofs. 2012. Insect omics research coming

of age. Can. J. Zool. 90: 440–455.

Bohinc, T., I.J. Košir, and S. Trdan. 2013. Glucosinolates as arsenal for defending Brassicas against

cabbage flea beetle (Phyllotreta spp.) attack. Zemdirbyste-Agriculture 100(2): 199–204.

Bohinc, T., and S. Trdan. 2013. Sowing mixtures of Brassica trap crops is recommended to reduce

Phyllotreta beetles injury to cabbage. Acta Agric. Scand. Sect. B - Soil Plant Sci. 63(4): 297–

303.

Bones, A.M., and J.T. Rossiter. 2006. The enzymic and chemically induced decomposition of

glucosinolates. Phytochemistry 67(11): 1053–1067.

Bonnemaison, L. 1965. Insect pests of crucifers and their control. Annu. Rev. Entomol. 10: 233–

256.

150

Borem, A., and R. Fritsche-Neto (Eds). 2014. Omics in plant breeding. Wiley-Blackwell.

Borgen, B.H., I. Ahuja, O.P. Thangstad, B.I. Honne, J. Rohloff, J.T. Rossiter, and A.M. Bones. 2012.

“Myrosin cells” are not a prerequisite for aphid feeding on oilseed rape (Brassica napus) but

affect host plant preferences. Plant Biol. 14(6): 894–904.

Borgen, B.H., O.P. Thangstad, I. Ahuja, J.T. Rossiter, and A.M. Bones. 2010. Removing the mustard

oil bomb from seeds: Transgenic ablation of myrosin cells in oilseed rape (Brassica napus)

produces MINELESS seeds. J. Exp. Bot. 61(6): 1683–1697.

Bracken, G.K., and G.E. Bucher. 1986. Yield losses in canola caused by adult and larval flea beetles,

Phylotreta cruciferae (Coleoptera: Chrysomelidae). Can. Entomol. (118): 319–324.

Brandt, R.N., and J. Lamb. 1993. lmportance of tolerance and growth rate in the resistance of

oilseed rapes and mustards to flea beetles , Phyllotreta cruciferae (Goeze) (Coleoptera :

Chrysomelidae). Canadian Journal Plant Science 73: 169–176.

Brattsten, L.B., C.F. Wilkinson, and T. Eisner. 1977. Herbivore-plant interactions : Mixed-function

oxidases and secondary plant substances. Science 196(4296): 1349–1352.

Bridges, M., A.M.E. Jones, A.M. Bones, C. Hodgson, R. Cole, E. Bartlet, R. Wallsgrove, V.K.

Karapapa, N. Watts, J.T. Rossiter, S. Proceedings, B. Sciences, N. Jan, M. Bridgesl, A.M.E.

Jonesl, A.M. Bones, C. Hodgsonl, R. Cole, E. Bartlet, R. Wallsgrove, V.K. Karapapal, N. Watts,

and J.T. Rossiterl. 2002. Spatial organization of the glucosinolate-myrosinase system in

Brassica specialist aphids is similar to that of the host Plant. Proc. R. Soc. London B Biol. Sci.

269(1487): 187–191.

Broadway, R.M. 1989. Tryptic inhibitory activity in wild and cultivated crucifers. Phytochemistry

28(3): 755–758.

151

Broadway, M.R. 1995. Are insects resistant to plant proteinase inhibitors? J. Insect Physiol. 41(2):

107–116.

Broadway, R.M., and D.L. Missurelli. 1990. Regulatory mechanisms of tryptic inhibitory activity in

cabbage plants. Phytochemistry 29(12): 3721–3725.

Brown, J., J.P.M.C. Caffrey, D.A. Brown, B.L. Harmon, and J.B. Davis. 2004. Yield reduction in

Brassica napus , B . rapa , B . juncea , and Sinapis alba caused by flea beetle (Phyllotreta

cruciferae (Goeze) (Coleoptera : Chrysomelidae )) infestation in Northern Idaho. J. Econ.

Entomol. 97(5): 1642–1647.

Bruce, T.J.A. 2014. Glucosinolates in oilseed rape: Secondary metabolites that influence

interactions with herbivores and their natural enemies. Ann. Appl. Biol. 164(3): 348–353.

Burgess, L. 1977a. Flea beetles (Coleoptera: Chyrsomelidae) attacking rape crops in the Canadian

prairie provinces. Can. Entomol. 32(620): 21–32.

Burgess, L. 1977b. Geocoris bullatus, an occasional predator on flea beetles (Hemiptera:

Lygaeidae). Can. Entomol. 109: 1519–1520.

Burgess, L. 1980. Predation on adults of the flea beetle Phyllotreta cruciferae by lacewing larvae

(Neuroptera: Chysopidae). Can. Entomol. 112(7): 745–746.

Burgess, L. 1981. Winter sampling to determine overwintering sites and estimate density of adult

flea beetle pests of rape (Coleoptera: Chrysomelidae). Can. Entomol. 113: 441–447.

Burgess, L. 1982. Predation on adults of the flea beetle Phyllotreta cruciferae by the western

damsel bug, Nabis alternatus (Hemiptera: Nabidae). Can. Entomol. 114(8): 763–764.

Bus, A., J. Hecht, B. Huettel, R. Reinhardt, and B. Stich. 2012. High-throughput polymorphism

detection and genotyping in Brassica napus using next-generation RAD sequencing. BMC

152

Genomics 13(1): 281.

Button, K.S., J.P.A. Ioannidis, C. Mokrysz, B.A. Nosek, J. Flint, E.S.J. Robinson, and M.R. Munafò.

2013. Power failure: why small sample size undermines the reliability of neuroscience. Nat.

Rev. Neurosci. 14(5): 365–376.

Canola Council of Canada. 2013. The Economic Impact of Canola on the Canadian Economy.

Oxford.

Canola Council of Canada. 2016. 2016 Annual Report - Industry Inspired. Winnipeg.

Health Canada. 2016. Proposed re-evaluation decision PRVD2016-20 Imidacloprid. Ottawa.

Chalhoub, B., F. Denoeud, S. Liu, I.A.P. Parkin, H. Tang, X. Wang, J. Chiquet, H. Belcram, C. Tong,

B. Samans, M. Correa, C. Da Silva, J. Just, C. Falentin, C.S. Koh, I. Le Clainche, M. Bernard, P.

Bento, B. Noel, K. Labadie, A. Alberti, M. Charles, D. Arnaud, H. Guo, C. Daviaud, S. Alamery,

K. Jabbari, M. Zhao, P.P. Edger, H. Chelaifa, D. Tack, G. Lassalle, I. Mestiri, N. Schnel, M.-C.

Le Paslier, G. Fan, V. Renault, P.E. Bayer, A.A. Golicz, S. Manoli, T.-H. Lee, V.H.D. Thi, S.

Chalabi, Q. Hu, C. Fan, R. Tollenaere, Y. Lu, C. Battail, J. Shen, C.H.D. Sidebottom, X. Wang,

A. Canaguier, A. Chauveau, A. Berard, G. Deniot, M. Guan, Z. Liu, F. Sun, Y.P. Lim, E. Lyons,

C.D. Town, I. Bancroft, X. Wang, J. Meng, J. Ma, J.C. Pires, G.J. King, D. Brunel, R. Delourme,

M. Renard, J.-M. Aury, K.L. Adams, J. Batley, R.J. Snowdon, J. Tost, D. Edwards, Y. Zhou, W.

Hua, A.G. Sharpe, A.H. Paterson, C. Guan, and P. Wincker. 2014. Early allopolyploid evolution

in the post-Neolithic Brassica napus oilseed genome. Science 345 (6199): 950–953.

Chandrasegaran, S., and D. Carroll. 2016. Origins of programmable nucleases for genome

engineering. J. Mol. Biol. 428(5): 963–989.

Chapman, R.F. 1974. The chemical inhibition of feeding by phytophagous insects: A review. Bull.

153

Entomol. Res. 64: 339–363.

Chapman, R.F. 1998. The insects: Structure and function. Cambridge University Press, Cambridge.

Chen, H., and Y. Lin. 2013. Promise and issues of genetically modified crops. Curr. Opin. Plant Biol.

16(2): 255–260.

Chen, S., B.L. Petersen, C.E. Olsen, A. Schulz, and B.A. Halkier. 2001. Long-distance phloem

transport of glucosinolates in Arabidopsis. Plant Physiol. 127(1): 194–201.

Christensen, S., C. Heimes, N. Agerbirk, V. Kuzina, C.E. Olsen, and T.P. Hauser. 2014. Different

geographical distributions of two chemotypes of Barbarea vulgaris that differ in resistance

to insects and a pathogen. J. Chem. Ecol. 40(5): 491–501.

Christou, P., T. Capell, A. Kohli, J.A. Gatehouse, and A.M.R. Gatehouse. 2006. Recent

developments and future prospects in insect pest control in transgenic crops. Trends Plant

Sci. 11(6): 302–308.

Chuong, P. V., and W.D. Beversdorf. 1985. High frequency embryogenesis through isolated

microspore culture in Brassica napus L. and B. Carinata Braun. Plant Sci. 39(3): 219–226.

Churchill, G.A., and R.W. Doerge. 1994. Empirical threshold values for quantitative trait mapping.

Genetics 138(3): 963–971.

Cipollini, D.F., and J.M. Bergelson. 2000. Environmental and developmental regulation of trypsin

inhibitor activity in Brassica napus. J. Chem. Ecol. 26(6): 1411–1422.

Cipollini, D.F., and J. Bergelson. 2001. Plant density and nutrient availability constrain constitutive

and wound-induced expression of trypsin inhibitors in Brassica napus. J. Chem. Ecol. 27(3):

593–610.

Cipollini, D.F., J.W. Busch, K.A. Stowe, E.L. Simms, and J.O.Y. Bergelson. 2003. Genetic variation

154

and relationships of constitutive and herbivore-induced glucosinolates, trypsin inhibitors,

and herbivore resistance in Brassica rapa. J. Chem. Ecol. 29(2): 285–302.

Clossais-Besnard, N., and F. Larher. 1991. Physiological role of glucosinolates in Brassica napus .

concentration and distribution pattern of glucosinolates among plant organs during a

complete life cycle. J. Sci. Food Agric. 56: 25–38.

Cole, R. 1996. Abiotic induction of changes to glucosinolate profiles in Brassica species and

increased resistance to the specialist aphid Brevicoryne brassicae. Entomol. Exp. Appl. 80:

228–230.

Collard, B.C.Y., M.Z.Z. Jahufer, J.B. Brouwer, and E.C.K. Pang. 2005. An introduction to markers,

quantitative trait loci (QTL) mapping and marker-assisted selection for crop improvement:

The basic concepts. Euphytica 142(1–2): 169–196.

Collard, B.C.Y., and D.J. Mackill. 2008. Marker-assisted selection: an approach for precision plant

breeding in the twenty-first century Marker-assisted selection: an approach for precision

plant breeding in the twenty-first century. Phil Trans R Soc B 363(8): 557–572.

Cook, S.M., Z.R. Khan, and J. a Pickett. 2007. The use of push-pull strategies in integrated pest

management. Annu. Rev. Entomol. 52: 375–400.

Cromartie, W.J.J. 1975. The effect of stand size and vegetational background on the colonization

of cruciferous plants by herbivorous insects. J. Appl. Ecol. 12(2): 517–533.

Culliney, T.W. 1986. Predation on adult Phyllotreta flea beetles by Podisus maculiventris

(Hemiptera: Pentatimidae) and Nabicula americolimbata (Hemiptera: Nabidae). Can.

Entomol. 118: 731–732.

Cusumano, A., B.T. Weldegergis, S. Colazza, M. Dicke, and N.E. Fatouros. 2015. Attraction of egg-

155

killing parasitoids toward induced plant volatiles in a multi-herbivore context. Oecologia

179(1): 163–174. van Dam, N.M., and M.W. Oomen. 2008. Root and shoot jasmonic acid applications differentially

affect leaf chemistry and herbivore growth. Plant Signal Behav 3(2): 91–98.

Van Dam, N.M., C.E. Raaijmakers, and W.H. Van Der Putten. 2005. Root herbivory reduces growth

and survival of the shoot feeding specialist Pieris rapae on Brassica nigra. Entomol. Exp.

Appl. 115(1): 161–170.

Despres, L., J.P. David, and C. Gallet. 2007. The evolutionary ecology of insect resistance to plant

chemicals. Trends Ecol. Evol. 22(6): 298–307.

Dethier, V.G. 1954. Evolution of feeding preferences in phytophagous insects. Evolution 8(1): 33–

54.

Dethier, V.G. 1980. Evolution of receptor sensitivity to secondary substances with special

reference to deterrents. Am. Nat. 115(1): 45–66.

Dethier, V.G., B.L. Browne, and C.N. Smith. 1960. The designation of chemicals in terms of the

responses they elicit from insects. J. Econ. Entomol. 53(1): 134–136.

Dosdall, L.M., M.G. Dolinski, N.T. Cowle, and P.M. Conway. 1999. The effect of tillage regime, row

spacing, and seeding rate on feeding damage by flea beetles, Phyllotreta spp. (Coleoptera:

Chrysomelidae), in canola in central Alberta, Canada. Crop Prot. 18: 217–224.

Dosdall, L.M., and P.G. Mason. 2010. Key pests and parasitoids of oilseed rape on canola in North

America and the importance of parasitoids in integrated management. p. 167–213. In

Williams, I.H. (ed.), Biocontrol-based integrated management of oilseed rape pests. 1st ed.

Springer, Dordrecht.

156

Dosdall, L.M., and F.C. Stevenson. 2005. Managing flea beetles (Phyllotreta spp.) (Coleoptera:

Chrysomelidae) in canola with seeding date, plant density, and seed. Agron. J. 97: 1570–

1578.

Doyle, J.J., and J.L. Doyle. 1990. A rapid total DNA preparation procdeure for fresh plant tissue.

Focus 12: 13–15.

Dunse, K.M., Q. Kaas, R.F. Guarino, P.A. Barton, D.J. Craik, and M.A. Anderson. 2010. Molecular

basis for the resistance of an insect chymotrypsin to a potato type II proteinase inhibitor.

Proc. Natl. Acad. Sci. U. S. A. 107(34): 15016–15021.

Eigenbrode, S.D. 2004. The effects of plant epicuticular waxy blooms on attachment and

effectiveness of predatory insects. Arthropod Struct. Dev. 33(1): 91–102.

Eigenbrode, S.D., and K. Espelie. 1995. Effects of plant epicuticular lipids on insect herbivores.

Annu. Rev. Entomol 40: 171–194.

Ekuere, U.U., L.M. Dosdall, M. Hills, A.B. Keddie, L. Kott, and A. Good. 2005. Identification,

mapping, and economic evaluation of QTLs encoding root maggot resistance in Brassica.

Crop Sci. 45(1): 371–378.

Elliott, R.H., C. Franke, and G.F.W. Rakow. 2008. Effects of seed size and seed weight on seedling

establishment, vigour and tolerance of Argentine canola (Brassica napus) to flea beetles,

Phyllotreta spp . Can. J. Plant Sci. 88(1): 207–217.

Van Emon, J.M. 2016. The omics revolution in agricultural research. J Agric Food Chem 64(1): 36–

44.

Ettlinger, M.G., G.P. Dateo, B.W. Harrison, T.J. Mabry, and C.P. Thompson. 1961. Vitamin C as a

coenzyme: the hydrolysis of mustard oil glucosides. Proc. Natl. Acad. Sci. U. S. A. 47: 1875–

157

1880.

Fahey, J.W., A.T. Zalcmann, and P. Talalay. 2001. The chemical diversity and distribution of

glucosinolates and isothiocyanates amoung plants. Phytochemistry 56: 5–51.

FAOSTAT. 2017. Rapeseed. Available at http://www.fao.org/faostat/en/#data/QC (verified 4

March 2017).

Feng, J., Y. Long, L. Shi, J. Shi, G. Barker, and J. Meng. 2012. Characterization of metabolite

quantitative trait loci and metabolic networks that control glucosinolate concentration in

the seeds and leaves of Brassica napus. New Phytol. 193(1): 96–108.

Ferreira, A., M.F. da Silva, L. da Costa e Silva, and C.D. Cruz. 2006. Estimating the effects of

population size and type on the accuracy of genetic maps. Genet. Mol. Biol. 29: 187–192.

Ferry, N., M.G. Edwards, J.A. Gatehouse, and A.M.R. Gatehouse. 2004. Plant-insect interactions:

Molecular approaches to insect resistance. Curr. Opin. Biotechnol. 15(2): 155–161.

Finch, S., and R.H. Collier. 2000. Host-plant selection by insects - A theory based on

“appropriate/inappropriate landings” by pest insects of cruciferous plants. Entomol. Exp.

Appl. 96(2): 91–102.

Foisset, N., R. Delourme, M.O. Lucas, and M. Renard. 1993. Segregation analysis of isozyme

markers on isolated microspore-derived embryos in Brassica napus L. Plant Breeding 110:

315–322.

Gali, K.K., and A.G. Sharpe. 2012. Molecular linkage maps: Strategies, resources and

acheivements. p. 85–129. In Edwards, D., Batley, J., Parkin, I., Kole, C. (eds.), Genetics,

Genomics and Breeding of Oilseed Brassicas. 1st ed. Science Publishers, Boca Raton, FL.

Ganal, M.W., T. Altmann, and M.S. Röder. 2009. SNP identification in crop plants. Curr. Opin.

158

Plant Biol. 12(2): 211–217.

Garcia, M.A., and M.A. Altieri. 1992. Explaining differences in flea beetle Phyllotreta cruciferae

Goeze densities in simple and mixed broccoli cropping systems as a function of individual

behavior. Entomol. Exp. Appl. 62(3): 201–209.

Gatehouse, A.G. 2008. Biotechnological prospects for engineering insect-resistant plants. Plant

Physiol. 146: 881–887.

Gavloski, J.E., and R.J. Lamb. 2000. Compensation by cruciferous plants is specific to the type of

simulated herbivory. Environ. Entomol. 29(6): 1273–1282.

Gavloski, J.E., U. Ekuere, A. Keddie, L. Dosdall, L. Kott, and A. G. Good. 2000. Identification and

evaluation of flea beetle (Phyllotreta cruciferae) resistance within Brassicaceae. Can. J. Plant

Sci. 80: 881–887.

Giamoustaris, a, and R. Mithen. 1995. The effect of modifying the glucosinolate content of leaves

of oilseed rape (Brassica napus ssp. oleifera) on its interaction with specialist and generalist

pests. Ann. appl. Biol 126: 347–363.

Gigolashvili, T., B. Berger, H.P. Mock, C. Müller, B. Weisshaar, and U.I. Flügge. 2007. The

transcription factor HIG1/MYB51 regulates indolic glucosinolate biosynthesis in Arabidopsis

thaliana. Plant J. 50(5): 886–901.

Gilmour, A.R., B.R. Cullis, A.P. Verbyla, S. Journal, E. Statistics, A.R. Gilmour, B.R. Cullis, and A.P.

Verbyla. 1997. Accounting for natural and extraneous variation in the analysis of field

experiments. J. Agric. Biol. Environ. Stat. 2(3): 269–293.

Girard, C., M. Le Metayer, B. Zaccomer, E. Bartlet, I. Williams, M. Bonade-Bottino, M.-H. Pham-

Delegue, and L. Jouanin. 1998. Growth stimulation of beetle larvae reared on a transgenic

159

oilseed rape expressing a cysteine proteinase inhibitor. J. Insect Physiol. 44(3–4): 263–270.

Gleadow, R.M., and I.A.N.E. Woodrow. 2002. Constraints on effectiveness of cyanogenic

glycosides in herbivore defense. J. Chem. Ecol. 28(7): 1301–1313.

Government of Canada. 2015. Table 1. Preliminary estimates of principal field crop areas.

Statisitics Canada: http://www.statcan.gc.ca/daily-quotidien/150630/t0.

Gruber, M.Y., S. Wang, S. Ethier, J. Holowachuk, P.C. Bonham-Smith, J.J. Soroka, and A. Lloyd.

2006. "HAIRY CANOLA” – Arabidopsis GL3 induces a dense covering of trichomes on Brassica

napus seedlings. Plant Mol. Biol. 60: 679–698.

Gruber, M., L. Wu, M. Links, B. Gjetvaj, J. Durkin, C. Lewis, A. Sharpe, D. Lydiate, and D. Hegedus.

2012. Analysis of expressed sequence tags in Brassica napus cotyledons damaged by crucifer

flea beetle feeding. Genome 55(2): 118–133.

Halkier, B.A., and J. Gershenzon. 2006. Biology and Biochemistry of Glucosinolates. Annu. Rev.

Plant Biol. 57(1): 303–333.

Hare, J.D. 2011. Ecological role of volatiles produced by plants in response to damage by

herbivorous insects. Annu. Rev. Entomol. 56: 161–180.

Hartmann, T. 1999. Chemical ecology of pyrrolizidine alkaloids. Planta 207(4): 483–495.

Heimes, C., J. Thiele, T. van Molken, and T.P. Hauser. 2015. Interactive impacts of a herbivore and

a pathogen on two resistance types of Barbarea vulgaris. Oecologia 117: 441–452.

Henderson, A.E., R.H. Hallett, and J.J. Soroka. 2004. Prefeeding behavior of the crucifer flea

beetle, Phyllotreta cruciferae, on host and nonhost crucifers. J. Insect Behav. 17(1): 17–39.

Henderson, H.M., and T.J. McEwen. 1972. Effect of ascorbic acid on thioglucosidases from

different crucifers. Phytochemistry 11(11): 3127–3133.

160

Hilder, V.A., A.M.R. Gatehouse, S.E. Sheerman, R.F. Barker, and D. Boulter. 1987. A novel

mechanism of insect resistance engineered into tobacco. Nature 300: 160–163.

Hilker, M., and N.E. Fatouros. 2015. Plant responses to insect egg deposition. Annu. Rev. Entomol.

60: 493–515.

Hilker, M., and T. Meiners. 2011. Plants and insect eggs: How do they affect each other?

Phytochemistry 72(13): 1612–1623.

Hirai, M.Y., K. Sugiyama, Y. Sawada, T. Tohge, T. Obayashi, A. Suzuki, R. Araki, N. Sakurai, H.

Suzuki, K. Aoki, H. Goda, O.I. Nishizawa, D. Shibata, and K. Saito. 2007. Omics-based

identification of Arabidopsis Myb transcription factors regulating aliphatic glucosinolate

biosynthesis. Proc. Natl. Acad. Sci. U. S. A. 104(15): 6478–83.

Hooks, C.R.R., and M.W. Johnson. 2003. Impact of agricultural diversification on the insect

community of cruciferous crops. Crop Prot. 22(2): 223–238.

Hopkins, R.J., N.M. Van Dam, and J.J.A. Van Loon. 2009. Role of glucosinolates in insect-plant

relationships and multitrophic interactions. Annu. Rev. Entomol 54: 57–83.

Hopkins, R.J., B. Ekbom, and L. Henkow. 1998. Glucosinolate content and susceptibility for insect

attack of three populations of Sinapis alba. J. Chem. Ecol. 24(7): 1203–1216.

Howe, G.A., and G. Jander. 2008. Plant immunity to insect herbivores. Annu. Rev. Plant Biol.

59(1): 41–66.

Husebye, H., S. Arzt, W.P. Burmeister, F. V. Härtel, A. Brandt, J.T. Rossiter, and A.M. Bones. 2005.

Crystal structure at 1.1 Å resolution of an insect myrosinase from Brevicoryne brassicae

shows its close relationship to β-glucosidases. Insect Biochem. Mol. Biol. 35(12): 1311–1320.

Isidoro, N., E. Bartlet, J. Ziesmann, and I.H. Williams. 1998. Antennal contact chemosensilla in

161

Psylliodes chrysocephala responding to cruciferous allelochemicals. Physiol. Entomol. 23(2):

131–138.

Jahangir, M., I.B. Abdel-Farid, H.K. Kim, Y.H. Choi, and R. Verpoorte. 2009. Healthy and unhealthy

plants: The effect of stress on the metabolism of Brassicaceae. Environ. Exp. Bot. 67(1): 23–

33.

Jallow, M.F.A., J.P. Cunningham, and M.P. Zalucki. 2004. Intra-specific variation for host plant use

in Helicoverpa armigera (Hubner) (Lepidoptera: Noctuidae): implications for management.

Crop Prot. 23: 955–964.

Jermy, T. 1966. Feeding inhibitors and food preference in chewing phytophagous insects.

Entomol. Exp. Appl. 9(1): 1–12.

Jetter, R., and M. Riederer. 1996. Cuticular waxes from the leaves and fruit capsules of eight

Papaveraceae species. Can. J. Bot. 74(3): 419–430.

Jones, E.S., H. Sullivan, D. Bhattramakki, and J.S.C. Smith. 2007. A comparison of simple sequence

repeat and single nucleotide polymorphism marker technologies for the genotypic analysis

of maize (Zea mays L.). Theor. Appl. Genet. 115(3): 361–371.

Kareiva, P. 985. Finding and losing host plants by Phyllotreta : Patch size and surrounding habitat.

Ecology 66(6): 1809–1816.

Kareiva, P. 1983. Influence of vegetation texture on herbivore populations: Resource

concentration and herbivore movement. p. 259. In Denno, R.F., McClure, M.S. (eds.),

Variable Plants and Herbivcores in Natural and managed Systems. Academic Press, New

York.

Kaur, S., M.G. Francki, and J.W. Forster. 2012. Identification, characterization and interpretation

162

of single-nucleotide sequence variation in allopolyploid crop species. Plant Biotechnol. J.

10(2): 125–138.

Kim, J.H., and G. Jander. 2007. Myzus persicae (green peach aphid) feeding on Arabidopsis

induces the formation of a deterrent indole glucosinolate. Plant J. 49(6): 1008–1019.

Kim, H., and J.S. Kim. 2014. A guide to genome engineering with programmable nucleases. Nat

Rev Genet 15(5): 321–334.

Kinoshita, G B, Svec H J, Harris, C R, Mcewen, F.L. 1979. Biology of the crucifer flea beetle,

Phyllotreta cruciferae (Coleoptera: Chyrsomelidae), in southwestern Ontario. Can. Entomol.

111: 1395–1407.

Kirkegaard, J.A., G.J. Rebetzke, and R.A. Richards. 2001. Inheritance of root glucosinolate content

in canola. Aust. J. Agric. Res. 2(7): 745–753.

Knodel, J., and D.L. Olsen. 2002. Crucifer flea beetle: biology and integrated pest management in

canola. North Dakota State Univ. Coop. Ext. Serv Publ. E123.

Knodel, J.J., D.L. Olson, B.K. Hanson, and R.A. Henson. 2008. Impact of planting dates and

insecticide strategies for managing crucifer flea beetles (Coleoptera: Chrysomelidae) in

spring-planted canola. J. Econ. Entomol. 101(3): 810–821.

Konstantinov A S and Vandenberg N J. 2010. Guide to paleartic flea beetle genera.

http://www.sel.barc.usda.gov/coleoptera/fleabeetles/fleas.htm accessed: 13 March 2017.

Kopisch-Obuch, F.J., and B.W. Diers. 2006. Segregation at the SCN resistance locus rhg1 in

soybean is distorted by an association between the resistance allele and reduced field

emergence. Theor. Appl. Genet. 112(2): 199–207.

Kosambi, D.D. 1944. the Estimation of Map Distances From Recombination Values. Ann. Eugen.

163

12: 172–175.

Kristensen, C., M. Morant, C.E. Olsen, C.T. Ekstrøm, D.W. Galbraith, B.L. Møller, and S. Bak. 2005.

Metabolic engineering of dhurrin in transgenic Arabidopsis plants with marginal inadvertent

effects on the metabolome and transcriptome. Proc. Natl. Acad. Sci. U. S. A. 102(5): 1779–

1784.

Kushad, M.M., A.F. Brown, A.C. Kurilich, J.A. Juvik, B.P. Klein, M.A. Wallig, and E.H. Jeffery. 1999.

Variation of glucosinolates in vegetable crops of Brassica oleracea. J. Agric. Food Chem.

47(4): 1541–1548.

Kuzina, V., J.K. Nielsen, J.M. Augustin, A.M. Torp, S. Bak, and S.B. Andersen. 2011. Barbarea

vulgaris linkage map and quantitative trait loci for saponins, glucosinolates, hairiness and

resistance to the herbivore Phyllotreta nemorum. Phytochemistry 72(2–3): 188–198.

Lamb, J. 1984. Effects of flea beetles Phyllotreta spp. (Chrysomelidae: Coleoptrea), on the

survival, growth, seed yield and quality of canola, rape and yellow mustard. Can. Entomol.

(116): 269–280.

Lamb, R.J. 1988. Assessing the susceptibility of crucifer seedlings to flea beetle (Phyllotreta spp.)

damage. Can. J. Plant Sci. 68: 85–93.

Lamb, R.J. 1989. Entomology of oilseed Brassica crops. Annu. Rev. Entomol. 34(1): 211–229.

Lamb, R.J., McVetty, P.B.E., Palaniswamy, P., Bodnaryk, R.P., Jeong, S.E. 1993. Susceptibility of

inbred lines of oilseed rape, Brassica napus, to feeding damage by lhe crucifer flea beetle,

Phyllotreta cruciferae (Goeze) [Coleoptera: Chrysomelidae, and its inheritance. Can. J. Plant

Sci. 73: 615–623.

Lamb, R.J., P. Palaniswamy, and R.P. Bodnaryk. 1991. Selection of oilseed rape with resistance to

164

flea beetles. p. 280–285. In GCIRC 1991 Congress.

Lamb, R.J., M.A.H. Smith, I.L. Wise, and R.I.H. McKenzie. 2016. Resistance to wheat midge

(Diptera: Cecidomyiidae) in winter wheat and the origins of resistance in spring wheat

(Poaceae). Can. Entomol. 148(2): 229–238.

Lamb, R.J., and W.J. Turnock. 1982. Economics of insecticidal control of flea beetles (Coleoptera:

Chrysomelidae) attacking rape in Canada. Can. Entomol. 114: 827–840.

Landry, B.S., N. Hubert, T. Etoh, J.J. Harada, and S.E. Lincoln. 1991. A genetic map for Brassica

napus based on restriction fragment length polymorphisms detected with expressed DNA

sequences. Genome 34(4): 543–552.

Larsen, L.M., J.K. Nielsen, and S. H. 1982. Identification of 3-O-[2-O-(/3-D-Xylopyranosyl)+-D-

Galactopyranosyl] flavonoids in horseradish leaves acting as feeding stimulants for a flea

beetle. Phytochemistry 21(5): 1029–1033.

Lee, R.W.H., I.T. Malchev, I. Rajcan, and L.S. Kott. 2014. Identification of putative quantitative

trait loci associated with a flavonoid related to resistance to cabbage seedpod weevil

(Ceutorhynchus obstrictus) in canola derived from an intergeneric cross, Sinapis alba ×

Brassica napus. Theor. Appl. Genet. 127(2): 419–428.

Li, Q., S.D. Eigenbrode, G.R. Stringam, and M.R. Thiagarajah. 2000. Feeding and growth of Plutella

xylostella and Spodoptera eridania on Brassica juncea with varying glucosinolate

concentrations and myrosinase activities. J. Chem. Ecol. 26(10): 2401–2419.

Li, X., and M.M. Kushad. 2004. Correlation of glucosinolate content to myrosinase activity in

horseradish (Armoracia rusticana). J. Agric. Food Chem. 52(23): 6950–6955.

Li, C., Y. Li, Y. Shi, Y. Song, D. Zhang, E.S. Buckler, Z. Zhang, Y. Li, and T. Wang. 2016. Analysis of

165

recombination QTLs, segregation distortion, and epistasis for fitness in maize multiple

populations using ultra-high-density markers. Theor. Appl. Genet. 129(9): 1775–1784.

Liang, Y.S., H.K. Kim, A.W.M. Lefeber, C. Erkelens, Y.H. Choi, and R. Verpoorte. 2006. Identification

of phenylpropanoids in methyl jasmonate treated Brassica rapa leaves using two-

dimensional nuclear magnetic resonance spectroscopy. J. Chromatogr. A 1112(1–2): 148–

155. van Loon, J.J.A. 1990. Chemoreception of phenolic-acids and flavonoids in larvae of 2 species of

Pieris. J. Comp. Physiol. a-Sensory Neural Behav. Physiol. 166(6): 889–899. van Loon, J.J.A., A. Blaakmeer, F.C. Griepink, T.A. van Beek, L.M. Schoonhoven, and A. de Groot.

1992. Leaf surface compound from Brassica oleracea (Cruciferae) induces oviposition by

Pieris brassicae (Lepidoptera: Pieridae). Chemoecology 3(1): 39–44.

Lu, H., J. Romero-Severson, and R. Bernardo. 2002. Chromosomal regions associated with

segregation distortion in maize. Theor. Appl. Genet. 105(4): 622–628.

Lusser, M., C. Parisi, D. Plan, and E. Rodríguez-Cerezo. 2012. Deployment of new biotechnologies

in plant breeding. Nat. Biotechnol. 30(3): 231–239.

Mailer, R.J., A. McFadden, J. Ayton, and B. Redden. 2008. Anti-nutritional components, fibre,

sinapine and glucosinolate content, in Australian Canola (Brassica napus L.) meal. JAOCS, J.

Am. Oil Chem. Soc. 85(10): 937–944.

Mainguet, A.M., A. Louveaux, G. El Sayed, and P. Rollin. 2000. Ability of a generalist insect,

Schistocerca gregaria, to overcome thioglucoside defense in desert plants: Tolerance or

adaptation? Entomol. Exp. Appl. 94(3): 309–317.

Mammadov, J., R. Aggarwal, R. Buyyarapu, and S. Kumpatla. 2012. SNP markers and their impact

166

on plant breeding. Int. J. Plant Genomics 2012.

Matsuda, K. 1976. Flavonoids as feeding stimulants of the beetles attaching the polygonaceous

plants. Tohoku J. Agric. Res. 27: 115–121.

Matsuda, K. 1988. Feeding stimulants of leaf beetles. p. 41–56. In Jolivet, P., Petitpierre, E., Hsiao,

T.H. (eds.), Biology of Chrysomelidae. Kluwer Academic Publishers, Dordrecht, The

Netherlands.

McGovran, E.R. 1969. Principles of plant and animal pest control. National Academy of Sciences,

Washington. van der Meijden, E. 1996. Plant defence, an evolutionary dilemma: contrasting effects of

(specialist and generalist) herbivores and natural enemies. Entomol. Exp. Appl. 80(1): 307–

310.

Mello, M.O., and M.C. Silva Filho. 2002. Plant-insect interactions: an evolutionary arms race

between two distinct defense mechanisms. Brazilian J. Plant Physiol. 14(2): 71–81.

Metspalu, L., E. Kruus, A. Ploomi, I.H. Williams, K. Hiiesaar, K. Jõgar, E. Veromann, and M. Mänd.

2014. Flea beetle (Chrysomelidae: Alticinae) species composition and abundance in

different cruciferous oilseed crops and the potential for a trap crop system. Acta Agric.

Scand. Sect. B-Soil Plant Sci. 64(7): 572–582.

Mewis, I., J.G. Tokuhisa, J.C. Schultz, H.M. Appel, C. Ulrichs, and J. Gershenzon. 2006. Gene

expression and glucosinolate accumulation in Arabidopsis thaliana in response to generalist

and specialist herbivores of different feeding guilds and the role of defense signaling

pathways. Phytochemistry 67(22): 2450–2462.

Milkowski, C., and D. Strack. 2010. Sinapate esters in brassicaceous plants: Biochemistry,

167

molecular biology, evolution and metabolic engineering. Planta 232(1): 19–35.

Miller, J.R., P.Y. Siegert, F.A. Amimo, and E.D. Walker. 2009. Designation of chemicals in terms of

the locomotor responses they elicit from insects: an update of Dethier et al. (1960). J. Econ.

Entomol. 102(6): 2056–2060.

Mitchell, B.K. 1988. Adult leaf beetles as models for exploring the chemical basis of host-plant

recognition. J. Insect Physiol. 34(3): 213–225.

MitchellOlds, T., D. Siemens, and D. Pedersen. 1996. Physiology and costs of resistance to

herbivory and disease in Brassica. Entomol. Exp. Appl. 80(1): 231–237.

Mithen, R. 1992. Leaf glucosinolate profiles and their relationship to pest and disease resistance

in oilseed rape. Euphytica 63(1–2): 71–83.

Mithen, R., A.F. Raybould, and A. Giamoustaris. 1995. Divergent selection for secondary

metabolites between wild populations of Brassica oleracea and Its implications for plant-

herbivore interactions. Heredity 75(5): 472–484.

Morton, N.E. 1955. Sequential tests for the detection of linkage. Am. J. Hum. Genet. 7(3): 277–

318.

Moyes, C.L., H.A. Collin, G. Britton, and A.F. Raybould. 2000. Glucosinolates and differential

herbivory in wild populations of Brassica oleracea. J. Chem. Ecol. 26(11): 2625–2641.

Moyes, C.L., and A.F. Raybould. 2001. The role of spatial scale and intraspecific variation in

secondary chemistry in host-plant location by Ceutorhynchus assimilis (Coleoptera:

Curculionidae). Proc. Biol. Sci. 268(1476): 1567–73.

Mulkern, G.B. 1969. Behavioral influences on food selection in grasshoppers (Orthoptera:

Acrididae). Entomol. Exp. Appl. 12(5): 509–523.

168

Müller, C., and U. Wittstock. 2005. Uptake and turn-over of glucosinolates sequestered in the

sawfly Athalia rosae. Insect Biochem. Mol. Biol. 35(10): 1189–1198.

N, U. 1935. Genomic analysis in Brassica with special reference to the experimental formation of

B. napus and peculiar mode of fertilization. Jap J Bot 7: 389–452.

Nair, R.B., R.W.J.I. Joy, E. Kurylo, X. Shi, J. Schnaider, R.S.S. Datla, W.A. Keller, and G. Selvaraj.

2000. Identification of a CYP84 family of cytochrome P450-dependent mono-oxygenase

genes in Brassica napus and perturbation of their expression for engineering sinapine

reduction in the seeds. Plant Physiol. 123(4): 1623–1634.

Nielsen, J.K. 1978. Host plant discrimination within Cruciferae: Feeding responses of four leaf

beetles (Coleoptera: Chrysomelidae) to glucosinulates, cucurbitacius and cardenolides.

Entomol. exp. appl. 24: 41–54.

Nielsen, J.K. 1997. Genetics of the ability of Phyllotreta nemorum larvae to survive in an atypical

host plant, Barbarea vulgaris ssp. arcuata. Entomol. Exp. Appl. 82(1): 37–44.

Nielsen, J.K., M.L. Hansen, N. Agerbirk, B.L. Petersen, and B.A. Halkier. 2001. Responses of the

flea beetles Phyllotreta nemorum and P. cruciferae to metabolically engineered Arabidopsis

thaliana with an altered glucosinolate profile. Chemoecology 11(2): 75–83.

Nielsen, J.K., L.M. Larsen, and H. Søorensen. 1977. Cucurbitacin E and I in Iberis amara: Feeding

inhibitors for Phyllotreta nemorum. Phytochemistry 16(10): 1519–1522.

Nielsen, J.K., L.M. Larsen, and H. Sorensen. 1979. Host plant selection of the horseradish flea

beetle Phyllotreta armoraciae (Coleoptera: Chysomelidae): Identification of two flavonol

glycosides stimulating feeding in combination with glucosinolates. Entomol. Exp. Appl. 26:

40–48.

169

Nielsen, J.K., T. Nagao, H. Okabe, and T. Shinoda. 2010a. Resistance in the plant, Barbarea

vulgaris, and counter-adaptations in flea beetles mediated by saponins. J. Chem. Ecol. 36(3):

277–285.

Nielsen, N.J., J. Nielsen, and D. Staerk. 2010b. New resistance-correlated saponins from the

insect-resistant crucifer Barbarea vulgaris. J. Agric. Food Chem. 58(9): 5509–5514.

Ontario, Government of. 2016. Nicotenoid Regulations. Gov. Ontario Available at

https://www.ontario.ca/page/neonicotinoid-regulations (verified 27 November 2016).

Onyilagha, J.C., J. Lazorko, M.Y. Gruber, J.J. Soroka, and M.A. Erlandson. 2004. Effect of flavonoids

on feeding preference and development of the crucifer pest Mamestra configurata Walker.

J. Chem. Ecol. 30(1): 109–124.

Osbourn, A. 1996a. Saponins and plant defence — a soap story. Trends Plant Sci. 1(1): 4–9.

Osbourn, A. 1996b. Preformed antimicrobial compounds and plant defense against fungal attack.

Plant Cell 8(10): 1821–1831.

Pachagounder, P., R.J. Lamb, and R.P. Bodnaryk. 1998. Resistance to the flea beetle Phyllotreta

cruciferae (Coleoptera : Chrysomelidae) in false flax, Camelina sativa (Brassicaceae). Can.

Entomol. 130: 235–240.

Painter, R.H. 1951. Insect resistance in crop plants. The MacMillan Co., New York.

Palaniswamy, P, Lamb, R.J. 1992. Host preferences of the flea beetles Phyllotreta cruciferae and

P. striolata (Coleoptera: Chrysomelidae ) for crucifer seedlings. J. Econ. Entomol. 85(3): 743–

752.

Palaniswamy, P., R.J. Lamb, and R.P. Bodnaryk. 1997. Antibiosis of preferred and non-preferred

host-plants for the flea beetle, Phyllotreta cruciferae (Goeze) (Coleoptera: Chrysomelidae).

170

Can. Entomol. 129(1583): 43.

Parkin, I.A., C. Koh, H. Tang, S.J. Robinson, S. Kagale, W.E. Clarke, C.D. Town, J. Nixon, V.

Krishnakumar, S.L. Bidwell, F. Denoeud, H. Belcram, M.G. Links, J. Just, C. Clarke, T. Bender,

T. Huebert, A.S. Mason, J. Pires, G. Barker, J. Moore, P.G. Walley, S. Manoli, J. Batley, D.

Edwards, M.N. Nelson, X. Wang, A.H. Paterson, G. King, I. Bancroft, B. Chalhoub, and A.G.

Sharpe. 2014. Transcriptome and methylome profiling reveals relics of genome dominance

in the mesopolyploid Brassica oleracea. Genome Biol. 15(6): R77

Parkin, I.A.P., A.G. Sharpe, D.J. Keith, and D.J. Lydiate. 1995. Identification of the A and C genomes

of amphidiploid Brassica napus (oilseed rape). Genome 38: 1122–1131.

Payne, J.M., and T.E. Michaels. 1995. Bacillus thuringiensis isolates seclectively active against

certain coleopteran pests. Pat. US 5427786 A.

Peng, C., R.J. Bartelt, and M.J. Weiss. 1999. Male crucifer flea beetles produce an aggregation

pheromone. Physiol. Entomol. 24(1): 98–99.

Peng, C., and M.J. Weiss. 1992. Evidence of an aggregation pheromone in the flea beetle,

Phyllotreta cruciferae (Goeze) (Coleoptera: Chrysomelidae). J. Chem. Ecol. 18(6): 875–884.

Pimentel, D. 1961. Species diversitty and insect population outbreaks. Ann. Entomol. Soc. Am.

(54): 76–86.

Piquemal, J., E. Cinquin, F. Couton, C. Rondeau, E. Seignoret, I. Doucet, D. Perret, M.J. Villeger, P.

Vincourt, and P. Blanchard. 2005. Construction of an oilseed rape (Brassica napus L.) genetic

map with SSR markers. Theor. Appl. Genet. 111(8): 1514–1523.

Pivnick, K.A., R.J. Lamb, and D. Reed. 1992. Response of flea beetles, Phyllotreta spp., mustard

oils and nitriles in field trapping experiments. J. Chem. Ecol. 18(6): 863–873.

171

Poelman, E.H., C. Broekgaarden, J.J.A. Van Loon, and M. Dicke. 2008. Early season herbivore

differentially affects plant defence responses to subsequently colonizing herbivores and

their abundance in the field. Mol. Ecol. 17(14): 3352–3365.

Prokopy, R., and E. Owens. 1983. Visual detection of plants by herbivorous insects. Annu. Rev.

Entomol. 28: 337–364.

Purdy, L.H., W.Q. Loegering, C.F. Konzak, C.J. Peterson, and R.E. Allan. 1968. A proposed standard

method for illustrating pedigrees of small grain varieties. Crop Sci. 8(4): 405.

Putnam L G. 1977. Respsonse of four Brassica seed crop species to attack by the crucifer fela

beetle, Phyllotreta cruciferae. Can. J. Plant Sci. 57(3): 987–989.

Pyke, B., M. Rice, B. Sabine, and M.P. Zalucki. 1987. The push-pull strategy—behavioural control

of Heliothis. Aust. Cott. Grow. May-July: 7–9.

Qu, C., L. Jia, F. Fu, H. Zhao, K. Lu, L. Wei, X. Xu, Y. Liang, S. Li, R. Wang, and J. Li. 2017. Genome-

wide association mapping and Identification of candidate genes for fatty acid composition

in Brassica napus L. using SNP markers. BMC Genomics 18(1): 232.

Rajcan, I., K.J. Kasha, L.S. Kott, and W.D. Beversdorf. 1999. Detection of molecular markers

associated with linolenic and erucic acid levels in spring rapeseed (Brassica napus L.).

Euphytica 105: 173–181.

Raman, H., R. Raman, A. Kilian, F. Detering, Y. Long, D. Edwards, I.A.P. Parkin, A.G. Sharpe, M.N.

Nelson, N. Larkan, J. Zou, J. Meng, M.N. Aslam, J. Batley, W.A. Cowling, and D. Lydiate. 2013.

A consensus map of rapeseed (Brassica napus L.) based on diversity array technology

markers: applications in genetic dissection of qualitative and quantitative traits. BMC

Genomics 14(1): 277.

172

Rask, L., E. Andréasson, B. Ekbom, S. Eriksson, B. Pontoppidan, and J. Meijer. 2000. Myrosinase:

Gene family evolution and herbivore defense in Brassicaceae. Plant Mol. Biol. 42(1): 93–113.

Ratzka, A., H. Vogel, D.J. Kliebenstein, T. Mitchell-Olds, and J. Kroymann. 2002. Disarming the

mustard oil bomb. Proc. Natl. Acad. Sci. U. S. A. 99(17): 11223–11228.

Reifenrath, K., M. Riederer, and C. Müller. 2005. Leaf surface wax layers of Brassicaceae lack

feeding stimulants for Phaedon cochleariae. Entomol. Exp. Appl. 115(1): 41–50.

Renwick, J.A.A. 2002. The chemical world of crucivores: Lures, treats and traps. Entomol. Exp.

Appl. 104(1): 35–42.

Ritcey, G.M., and S.B. McIver. 1990. External morphology of antennal sensilla of four species of

adult flea beetles (Coleoptera: Chrysomelidae: Alticinae). Int. J. Insect Morphol. Embryol.

19(2): 141–153.

Roessingh, P., E. Stadler, R. Baur, J. Hurter, and T. Ramp. 1997. Tarsal chemoreceptors and

oviposition behaviour of the cabbage root fly (Delia radicum) sensitive to fractions and new

compounds of host-leaf surface extracts. Physiol. Entomol. 22(2): 140–148.

Root, R.B. 1973. Organization of a plant-arthropod association in simple and diverse habitats :

The fauna of collards (Brassica oleracea). Ecol. Monogr. 43(1): 95–124.

SAS. 2012. SAS Institute Inc. Cary NC..

Schoonhoven, L.M., J.J.A. van loon, and M. Dicke. 2005. Insect-Plant Biology. 2nd ed. Oxford

Press, Oxford.

Scriber, J.M., and F. Slanksky Jr. 1981. The nurtritional ecology of immature insects. Annu. Rev.

Entomol. 26: 183–211.

Sekulic, G., and C.B. Remple. 2016. Evaluating the role of seed treatments in canola/oilseed rape

173

production: Integrated pest management, pollinator health, and biodiversity. Plants 5(3):

32.

Shehata, A.I., H.A. Al-Ghethar, and A.A. Al-Homaidan. 2009. Application of simple sequence

repeat (SSR) markers for molecular diversity and heterozygosity analysis in maize inbred

lines. Saudi J. Biol. Sci. 16(2): 57–62.

Simms, E.L., and M.D. Rausher. 1987. Costs and benefits of plant resistance to herbivory. Am.

Nat. 130(4): 570–581.

Singh, N., D.R. Choudhury, A.K. Singh, S. Kumar, K. Srinivasan, R.K. Tyagi, N.K. Singh, and R. Singh.

2013. Comparison of SSR and SNP markers in estimation of genetic diversity and population

structure of Indian rice varieties. PLoS One 8(12): 1–14.

Smith, C. 2004. Plant resistance against pests: Issues and strategies. p. 147. In Koul, O., Dhaliwal,

G.S., Cuperus, G.W. (eds.), Integrated Pest Management: Potential, Constraints and

Challenges. CABI Publishing, Oxford.

Smith, Alison; Cullis, Brian; Gilmour, A. 2001a. The analysis of crop variety evaulation data in

Australia. Aust. New Zeal. J. Stat. 43(2): 129–145.

Smith, A., B.R. Cullis, and R. Thompson. 2001b. Analyzing variety by environment data using

multiplicative mixed models and adjustments for spatial field trend. Biometrics 57(4): 1138–

1147.

Smith, A.B., B.R. Cullis, and R. Thompson. 2005. The analysis of crop cultivar breeding and

evaluation trials: an overview of current mixed model approaches. J. Agric. Sci. 143(6): 449.

Smooker, A.M., R. Wells, C. Morgan, F. Beaudoin, K. Cho, F. Fraser, and I. Bancroft. 2011. The

identification and mapping of candidate genes and QTL involved in the fatty acid

174

desaturation pathway in Brassica napus. Theor. Appl. Genet. 122(6): 1075–1090.

Snowdon, R., W. Lühs, and W. Friedt. 2007. Oilseed Rape. p. 55–114. In Kole, C. (ed.), Genome

Mapping and Molecular Breeding in Plants. Springer-Verlag, Berlin.

Soroka, J.J. 2011. Estimating flea Beetle damage in canola. Canola Watch Available at

http://www.canolawatch.org/2011/05/09/estimating-flea-beetle-damage-in-canola/.

accessed: 21 January 2016.

Soroka, J., and L. Grenkow. 2013. Susceptibility of Brassicaceous plants to feeding by flea beetles,

Phyllotreta spp. (Coleoptera: Chrysomelidae). J. Econ. Entomol 106(6): 2557–2567.

Soroka, J.J., L.F. Grenkow, and R.B. Irvine. 2008. Impact of decreasing ratios of insecticide-treated

seed on flea beetle (Coleoptera: Chrysomelidae, Phyllotreta spp.) feeding levels and canola

seed yields. J. Econ. Entomol. 101(6): 1811–1820.

Soroka, J.J., J.M. Holowachuk, M.Y. Gruber, and L.F. Grenkow. 2011. Feeding by flea beetles

(Coleoptera: Chrysomelidae; Phyllotreta spp.) is decreased on canola (Brassica napus)

seedlings with increased trichome density. J. Econ. Entomol. 104(1): 125–136.

Southwood, T.R.E. 1996. Insect-plant relations: overview from the symposium. Entomol. Exp.

Appl. 80: 320–324.

Spit, J., L. Badisco, H. Verlinden, P. Van Wielendaele, S. Zels, S. Dillen, and J. Vanden Broeck. 2012.

Peptidergic control of food intake and digestion in insects. Can. J. Zool. 90: 489–506.

Stadler, E., J. A.A. Renwick, C.D. Radke and K. Sachdev-Gupta. 1995. Tarsal contact

chemoreceptor response to glucosinolates and cardenolides mediating oviposition in Pieris

rapae. Physiol. Entomol. 20(2): 175–187.

Stewart Jr, C.N., M.J. Adang, J.N. All, P.L. Raymer, S. Ramachandran, and W. A. Parrott. 1996.

175

Insect control and dosage effects in transgenic canola containing a synthetic Bacillus

thuringiensis cryIAc Gene. Plant Physiol. 112(1 996): 115–120.

Stinner, R., C. Barfield, J. Stimac, and L. Dohse. 1983. Dispersal and movement of insect pests.

Annu. Rev. Entomol. 28(1): 319–335.

Swanson, E.B., M.P. Coumans, S.C. Wu, T.L. Barsby, and W.D. Beversdorf. 1987. Efficient isolation

of microspores and the production of microspore-derived embryos from Brassica napus.

Plant Cell Rep. 6(2): 94–97.

Tahvanainen, J.O., and R.B. Root. 1972. The influence of vegetational diversity on the population

ecology of a specialized herbivore, Phyllotreta cruciferae (Coleoptera: Chrysomelidae).

Oecologia 10(4): 321–346.

Tamiru, A., Z.R. Khan, and T.J.A. Bruce. 2015. New directions for improving crop resistance to

insects by breeding for egg induced defence. Curr. Opin. Insect Sci. 9: 51–55.

Tansey, J.A. 2007. Species Page - Phyllotreta striolata. Entomol. Collect. Available at

http://entomology.museums.ualberta.ca/searching_species_details.php?s=6298# (verified

12 October 2016).

Tansey, J.A., L.M. Dosdall, and B.A. Keddie. 2009. Phyllotreta cruciferae and Phyllotreta striolata

responses to insecticidal seed treatments with different modes of action. J. Appl. Entomol.

133: 201–209.

Tansey, J.A., L.M. Dosdall, B.A. Keddie, and R.M. Sarfraz. 2008. Differences in Phyllotreta

cruciferae and Phyllotreta striolata (Coleoptera : Chrysomelidae) responses to neonicotinoid

seed treatments. J. Econ. Entomol. 101(1): 159–167.

Tattersall, D.B., S. Bak, P.R. Jones, C.E. Olsen, J.K. Nielsen, M.L. Hansen, P.B. Høj, and B.L. Møller.

176

2001. Resistance to an herbivore through engineered cyanogenic glucoside synthesis.

Science 293(5536): 1826–1828.

Thaler, J.S., A.L. Fidantsef, S.S. Duffey, and R.M. Bostock. 1999. Trade-offs in plant defense

against pathogens and herbivores: a field demonstration of chemical elicitors of induced

resistance. J. Chem. Ecol. 25(7): 1597–1609.

Thaler, J.S., M.J. Stout, R. Karban, and S.S. Duffey. 2001. Jasmonate-mediated induced plant

resistance affects a community of herbivores. Ecol. Entomol. 26(3): 312–324.

Traw, M.B., and T.E. Dawson. 2002a. Differential induction of trichomes by three herbivores of

black mustard. Oecologia 131(4): 526–532.

Traw, M.B., and T.E. Dawson. 2002b. Reduced performance of two specialist herbivores

(Lepidoptera: Pieridae, Coleoptera: Chrysomelidae) on new leaves of damaged black

mustard plants. Environ. Entomol. 31(4): 714–722.

Vales, M.I., C.C. Schön, F. Capettini, X.M. Chen, A.E. Corey, D.E. Mather, C.C. Mundt, K.L.

Richardson, J.S. Sandoval-Islas, H.F. Utz, and P.M. Hayes. 2005. Effect of population size on

the estimation of QTL: A test using resistance to barley stripe rust. Theor. Appl. Genet.

111(7): 1260–1270.

Van Ooijen, J.W. 2006. Joinmap ®4, Software for the calculation of genetic linkage maps in

experimental populations. Kyazma B.V., Wageningen, Netherlands.

Velasco, P., M.E. Cartea, C. González, M. Vilar, and A. Ordás. 2007. Factors affecting the

glucosinolate content of kale (Brassica oleracea acephala Group). J. Agric. Food Chem. 55(3):

955–962.

Vincent, C., and R.K. Stewart. 1984. Effect of allyl isothiocyanate on field behavior of crucifer-

177

feeding flea beetles (Coleoptera: Chrysomelidae). J. Chem. Ecol. 10(1): 33–39.

Visser, J. 1986. Host odor perception in phytophagous Insects. Annu. Rev. Entomol. 31(1): 121–

144.

Wallace, S.K., and S.D. Eigenbrode. 2002. Changes in the glucosinolate-myrosinase defense

system in Brassica juncea cotyledons during seedling development. J. Chem. Ecol. 28(2):

243–256.

Wang, S.C., C.J. Basten, and Z.-B. Zeng. 2012. Windows QTL Catographer 2.5. Department of

Statistics, North Carolina University, Raleigh, NC)

http://statgen.ncsu.edu/qtlcart/WQTLCart.htm. )

Wang, X., H. Wang, J. Wang, R. Sun, J. Wu, S. Liu, Y. Bai, J.-H. Mun, I. Bancroft, F. Cheng, S. Huang,

X. Li, W. Hua, J. Wang, X. Wang, M. Freeling, J.C. Pires, A.H. Paterson, B. Chalhoub, B. Wang,

A. Hayward, A.G. Sharpe, B.-S. Park, B. Weisshaar, B. Liu, B. Li, B. Liu, C. Tong, C. Song, C.

Duran, C. Peng, C. Geng, C. Koh, C. Lin, D. Edwards, D. Mu, D. Shen, E. Soumpourou, F. Li, F.

Fraser, G. Conant, G. Lassalle, G.J. King, G. Bonnema, H. Tang, H. Wang, H. Belcram, H. Zhou,

H. Hirakawa, H. Abe, H. Guo, H. Wang, H. Jin, I.A.P. Parkin, J. Batley, J.-S. Kim, J. Just, J. Li, J.

Xu, J. Deng, J.A. Kim, J. Li, J. Yu, J. Meng, J. Wang, J. Min, J. Poulain, J. Wang, K. Hatakeyama,

K. Wu, L. Wang, L. Fang, M. Trick, M.G. Links, M. Zhao, M. Jin, N. Ramchiary, N. Drou, P.J.

Berkman, Q. Cai, Q. Huang, R. Li, S. Tabata, S. Cheng, S. Zhang, S. Zhang, S. Huang, S. Sato, S.

Sun, S.-J. Kwon, S.-R. Choi, T.-H. Lee, W. Fan, X. Zhao, X. Tan, X. Xu, Y. Wang, Y. Qiu, Y. Yin, Y.

Li, Y. Du, Y. Liao, Y. Lim, Y. Narusaka, Y. Wang, Z. Wang, Z. Li, Z. Wang, Z. Xiong, and Z. Zhang.

2011. The genome of the mesopolyploid crop species Brassica rapa. Nat. Genet. 43(10):

1035–1039.

178

Westdal, P.H., and W. Romanow. 1972. Observations on the biology of the flea beetle, Phyllotreta

cruciferae (Coleoptera: Chrysomelidae). Manitoba Entomol. 6: 35–45.

Wittstock, U., N. Agerbirk, E.J. Stauber, C.E. Olsen, M. Hippler, T. Mitchell-Olds, J. Gershenzon,

and H. Vogel. 2004. Successful herbivore attack due to metabolic diversion of a plant

chemical defense. Proc. Natl. Acad. Sci. U. S. A. 101(14): 4859–4864.

Wittstock, U., and B.A. Halkier. 2002. Glucosinolate research in the Arabidopsis era. Trends Plant

Sci. 7(6): 263–270.

Wylie, H.G. 1979. Observations on distribution, seasonal life history, and abundance of flea

beetles (Coleoptera: Chrysomelidae) tht infest rape crops in Manitoba. Can. Entomol. 111:

1345–1353.

Wylie, H.G. 1981. Effects of collection methig on estimates of parasitism and sex ratio of flea

beetles (Coleoptera: Chrysomelidae) that infest rape crops in Manitoba. Can. Entomol.

113(8): 665–671.

Wylie, H.G., W.J. Turnock, and L. Burgess. 1984. Phyllotreta spp., Flea Beetles (Coleoptera:

Chrysomelidae). p. 73–76. In Kelleher, J.S., Hulme, M.A. (eds.), Biological Control

Programmes against Insects and Weeds in Canada 1969-1980. Commonwealth Agricultural

Bureaux, Slough, UK.

Xu, C., P. De Clercq, M. Moens, S. Chen, and R. Han. 2010. Efficacy of entomopathogenic

nematodes (Rhabditida: Steinernematidae and Heterorhabditidae) against the striped flea

beetle, Phyllotreta striolata. BioControl 55(6): 789–797.

Xu, P., X. Wu, B. Wang, Y. Liu, J.D. Ehlers, T.J. Close, P.A. Roberts, N.N. Diop, D. Qin, T. Hu, Z. Lu,

and G. Li. 2011. A SNP and SSR based genetic map of asparagus bean (Vigna. unguiculata

179

ssp. sesquipedialis) and comparison with the broader species. PLoS One 6(1).

Zagrobelny, M., S. Bak, A.V. Rasmussen, B. Jørgensen, C.M. Naumann, and B.L. Møller. 2004.

Cyanogenic glucosides and plant-insect interactions. Phytochemistry 65(3): 293–306.

Zeng, Z.B. 1994. Precision mapping of quantitative trait loci. Genetics 136(4): 1457–1468.

Zhao, J., and J. Meng. 2003. Detection of loci controlling seed glucosinolate content and their

association with sclerotinia resistance in Brassica napus. Plant Breed. 122(1): 19–23.

Züst, T., B. Joseph, K.K. Shimizu, D.J. Kliebenstein, and L. a Turnbull. 2011. Using knockout

mutants to reveal the growth costs of defensive traits. Proc. R. Soc. London B Biol. Sci.

278(1718): 2598–2603.

180

APPENDIX

a. c.

Figure A4.1. – Linkage groups from Population J10-02 showing the position of three unique QTL for flea beetle herbivory in winter-type canola as highlighted in red. a.) at approximately 16.64cM (trait: % damage in field and BLUEs – overlapping); b.) at approximately 31.5cM (trait: feeding bites); c.) at approximately 188.9cM (trait: BLUEs).

181

a. b.

Figure A4.2. – Linkage groups from Population J10-11 showing the position of four unique QTL for flea beetle herbivory in winter-type canola as highlighted in red. a.) at approximately 34.9cM (trait: % field damage) and 118.35cM (trait: feeding bites); b.) at approximately 137.17cM (trait: feeding bites); c.) at approximately 78.54cM (trait: feeding bites).

182