Storage and Propagation Characteristics of Miscanthus X Giganteus by Cassandra Doll Downey

Home , Assimilation (biology), Panicoideae, Potato, Yam (vegetable)

Storage and Propagation Characteristics of Miscanthus x giganteus by Cassandra Doll Downey

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Master of Science in Plant Agriculture

Guelph, Ontario, Canada © Cassandra Downey, September, 2018 ABSTRACT

STORAGE AND PROPAGATION CHARACTERISTICS OF MISCANTHUS X GIGANTEUS

Cassandra Doll Downey Advisor: University of Guelph 2018 Dr. Andrew Maxwell Phineas Jones

Miscanthus x giganteus Anderss. exhibits favourable characteristics as a dedicated second- generation biomass feedstock. Propagation of this sterile species occurs mainly through micropropagation or rhizome cuttings, resulting in high establishment costs. This thesis sought to investigate the improvement of storage duration for field-harvested rhizomes, prolongment of callus regeneration potential through impairment of the phenylpropanoid biosynthetic pathway, and the induction of microrhizomes through the modification of media compositions. It was found that plant growth from rhizomes yielded similar results after spring and autumn harvests, plantlet regeneration was achieved from calli after 12 months in culture when media was supplemented with 2-aminoindan-2-phosphonic acid, and microrhizomes were successfully induced over a 10- week incubation period in liquid media supplemented with 8% sucrose, 6-benzylaminopurine, and 1-naphthaleneacetic acid. These findings may be used to improve conservation, reduce establishment costs, and allow for a greater supply of Miscanthus biomass in the marketplace.

Acknowledgements

I would like to begin by acknowledging my advisor, Dr. Andrew Maxwell Phineas Jones, for providing me with the opportunity to pursue my MSc. His guidance, understanding, and encouragement for me to engage in activities and projects both within and outside the scope of my thesis was valuable and greatly appreciated. I would also like to acknowledge my co-advisor, Dr. Praveen Kumar Saxena, for allowing me with the chance to gain valuable experience working with the Gosling Research Institute for Plant Preservation before beginning my MSc; Dr. Bill Deen, who helped with revising this thesis by incorporating his expertise in biomass crop production; Dr. Mahendra Thimmanagari of OMAFRA for supporting this research, providing valuable insight into biomass cultivation, and agreeing to be a member on my defence committee; and Dr. Ralph Martin for agreeing to chair my defence exam. I also would like to show my appreciation to our funding partner, BioFuelNet, for their considerate financial support.

I am thankful for David Smith and the team at All Weather Farming Inc. for their generosity, depth of knowledge, and enthusiasm for this project. I would also like to thank Henk Williams and the Elora Research Station for providing us with additional planting material and support; Dr. Xin Hu at NSF International for helping us process samples in a timely manner; Dr. Michelle Edwards for her guidance in statistical analyses; Dr. Gale Bozzo for permitting us to use his cold storage and freezer space whenever we were in a bind; and Tara Israel for helping me coordinate my defence exam and complete necessary administrative tasks.

I would like to express my gratitude for the Gosling Research Institute for Plant Preservation and the fantastic students and staff I’ve had the pleasure of working alongside. You are all extremely talented scientists who have inspired me to continually pursue truth. I am especially grateful for Bob Nichols, who was always willing to answer my many questions; Dr. Mukund Shukla, who took the time to teach me various tissue culture techniques and experimental design; Dr. Elena Popova, who familiarized me with cryopreservation technologies; Dr. Sherif Sherif, who encouraged me to take part in worthwhile teaching assistantships; and Lauren Erland, who offered her metabolite quantification services to us without grievance, even when she was overwhelmed with other requests and commitments.

The Department of Plant Agriculture has a great sense of community, and I would like to thank the Plant Agriculture Social Committee for allowing me to be more involved with this body. Also,

iii an abundance of thanks to Dr. Jay Subramanian for allowing me to TA his tissue culture course during my MSc so I could extend my passion for plant tissue culture to others.

I am forever grateful for my parents, Glen and Karen Downey, who modelled sacrifice and exceptional work ethic to provide me with the opportunities to fulfill my dreams; for my many friends and family who continue to love and support me; and for my significant other who perfectly balances me out.

And above all else, I acknowledge and praise God.

1 Thessalonians 5:16-18.

Table of Contents 1: Introduction……………………………………………………………………………………..1 1.1 Summary………………………………………………………………………………1 1.2 Hypotheses and objectives……………………………………………………………..5 2: Literature Review……………………………………………………………………………….6 2.1 Miscanthus spp………………………………………………………………………...6 2.1.1 Miscanthus ancestry and taxonomy………………………………………….6 2.1.2 Aboveground physiology and morphology………………………………….6 2.1.3 Rhizome physiology and morphology……………………………………….7 2.2 Seedless triploid propagation………………………………………………………...... 9 2.2.1 Storage and establishment properties………………………………...………9 2.2.2 Storage and carbohydrate dynamics………………………………………..11 2.3 Micropropagation…………………………………………………………………….13 2.3.1 Fundamentals of micropropagation………………………………………...13 2.3.2 Miscanthus micropropagation………………………………...……………13 2.3.3 Miscanthus micropropagation challenges………………………...………..14 2.3.4 Phenylpropanoid biosynthetic pathway and in vitro culture…………...…...14 2.4 Microrhizomes……………………………………………………………………….16 2.4.1 Introduction……………………………………………………………...…16 2.4.2 Carbohydrate requirements………………………………………………...17 2.4.3 Carbohydrate and plant growth regulator interactions………………...... 18 2.4.4 Synthetic seeds…………………………………………………………..…21 3: Storage and propagation of rhizomes from five Miscanthus x giganteus genotypes grown in southwestern Ontario…………………………………………………………………………….24 Abstract…………………………………………………………………………………..24 3.1 Introduction…………………………………………………………………………..24 3.2 Materials and methods ……………………………………………………………28 3.2.1 Harvest – Port Ryerse……………………………………………………....28 3.2.2 Harvest – Elora…………………………………………………...………...32 3.2.3 Sample preparation…………………………………………………...…….34 3.2.4 Cold storage conditions…………………………………………………….36 3.2.5 Greenhouse conditions……………………………………………………..37 3.2.6 Planting conditions…………………………………………………………38 3.2.7 Growth parameters…………………………………………………………38 3.2.8 Total moisture content……………………………………………………...39 3.2.9 Sample preparation………………………………………………………....40 3.2.10 Total starch quantification………………………………………………...40 3.2.11 Reducing and non-reducing sugar quantification………………………....41 3.2.12 Experimental design and statistical analysis………………………………42 3.3 Results………………………………………………………………………………..43 3.3.1 Rhizome viability……………………………………………………….….43 3.3.2 Culm emergence…………………………………………………………....44 3.3.3 Tiller height………………………………………………………………...47 3.3.4 Tiller number…………………………………………………………….....50 3.3.5 Leaf number…………………………………………………………….….52 3.3.6 Stem node number……………………………………………………….....55

3.3.7 Stem diameter……………………………………………………...……….58 3.3.8 Chlorophyll content………………………………………………………...61 3.3.9 Moisture content………………………………………………………...….61 3.3.10 Total starch content……………………………………………………….63 3.3.11 Total soluble carbohydrate content………………………………………..66 3.3.12 Total soluble carbohydrates consisting of D-glucose…………………...... 68 3.3.13 Total soluble carbohydrates consisting of sucrose………………………...71 3.3.14 Total soluble carbohydrates consisting of D-fructose………………..…....74 3.4 Discussion…………………………………………………………………………....76 3.5 Conclusions…………………………………………………………………………..81 3.6 Acknowledgements………………………………………………………………..…81 4: Improving regeneration capacity of Miscanthus x giganteus ‘M161’ calli through inhibition of the phenylpropanoid biosynthetic pathway…………………………………………..…………..82 Abstract……………………………………………………………………………..……82 4.1 Introduction…………………………………………………………………………..82 4.2 Materials and methods…………………………………………………………….….85 4.2.1 Plant material……………………………………………………………….85 4.2.2 Callus induction and multiplication………………………………………...85 4.2.3 Callus morphology assessment……………………………………………..86 4.2.4 Development of regenerants and embryo-like structures…………………...86 4.2.5 Soluble phenolic content…………………………………………………...87 4.2.6. Regeneration and plantlet formation……………………………………….88 4.2.7. Experimental design and statistical analysis…………………………..…...88 4.3 Results………………………………………………………………………………..88 4.3.1 Callus morphology frequencies…………………………………………….88 4.3.2 Regenerant number…………………………………………………….…...95 4.3.3 Soluble phenolic content…………………………………………………...98 4.3.4 Plantlet development……………………………………………….………98 4.4 Discussion……………………………………………………………………….…...99 4.5 Conclusions…………………………………………………………………………102 4.6 Acknowledgements…………………………………………………………………102 5: In Vitro Induction and Encapsulation of Miscanthus x giganteus Anderss. Microrhizomes…...... 103 Abstract…………………………………………………………………………………103 5.1 Introduction…………………………………………………………………………103 5.2 Materials and methods……………………………………………………………....105 5.2.1 Culture initiation…………………………………………………………..105 5.2.2 Microrhizome induction………………………………………………..…106 5.2.3 Physiological assessment………………………………………………....106 5.2.4 Growth capacity…………………………………………………………..107 5.2.5 Microrhizome isolation…………………………………………………...108 5.2.6 In vitro growth…………………………………………………………….108 5.2.7 Ex vitro growth…………………………………………………………....108 5.2.8 Microrhizome cold storage ability………………………………………...109 5.2.9 Synthetic seed production……………………………………………...….109 5.2.10 Full plantlet cold storage capacity………………………………………..110

5.2.11 Experimental design and statistical analysis……………………………..111 5.3 Results………………………………………………………………………………111 5.3.1 Shoot number…………………………………………………..…………111 5.3.2 Plantlet weight…………………………………………………………….112 5.3.3 Physiological assessment…………………………………………..……..115 5.3.4 Root formation and shoot node development………………………..….....118 5.3.5 Relative fresh weights……………………………………………….……120 5.3.6 Microrhizome growth capacity………………………………………..…..121 5.3.7 Synthetic seed growth capacity………………………………………..…..124 5.3.8 Cold storage – full plantlets……………………………………………….125 5.4 Discussion…………………………………………………………………………..125 5.5 Conclusions…………………………………………………………………………131 5.6 Acknowledgements…………………………………………………………………132 6: Overall Discussion…………………………………………………………………………...132 6.1 Research contributions……………………………………………………………...132 6.2 Research limitations……………………………………………………………...…133 6.3 Conclusion……………………………………………………………………….….135 7: References……………………………………………………………………………………136

vii

List of Figures Figure 3.1 Stages of machine-specialized Miscanthus rhizome harvest by All Weather Farming Inc. in Port Ryerse, ON. A) Altered potato harvester attached to a tracker in a freshly-cut Miscanthus field; B) rhizomes were removed from the ground with a backhoe loader and placed in the harvester. A series of rotating blades allowed excess soil and stones to separate from the rhizomes and fall through the bottom, while rhizomes deposited at the rear; C) unprocessed rhizomes after exiting the harvester; D) primary wash of samples for removal of soil; E) streamlined root removal using circular saw blades; F) fully-processed rhizomes cut to approximately 6 mm in length…………………………………………………………………....31 Figure 3.2 Average monthly air temperature (including monthly highs and lows) recorded at the Delhi Climate Station (located near Port Ryerse, ON) in M. x giganteus rhizome harvest years: a) 2015 and b) 2016…………………………………………………………………………………32 Figure 3.3 Average monthly air temperature (including monthly highs and lows) recorded at the Elora Research Station (Elora, ON) in M. x giganteus rhizome harvest years: a) 2015 and b) 2016………………………………………………………………………………………………33 Figure 3.4 Stages of Miscanthus growth. A) Pairs of processed rhizomes after autumn 2015 harvest (from left to right: ‘Illinois’, ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’); B) newly-developed rhizomes with immature (IM) and mature (M) dormant buds after six weeks in greenhouse conditions; C) rhizome with healthy, thick, white roots after six weeks in greenhouse conditions; D) five-week old Miscanthus stem (measuring 700-800 mm in height) grown from a rhizome piece after cultivation in greenhouse conditions for six weeks……………………………………………….36 Figure 3.5 Representation of severe Miscanthus ‘Illinois’ rhizome contamination after various times in storage at 3°C. A) Otherwise visually-healthy rhizome coated in soft, white, fibrous material; B) discoloured rhizome exhibiting soft-rot symptoms……………………………….....37 Figure 3.6 Notable Miscanthus aerial growth traits cultivated from rhizome pieces and incubated in greenhouse conditions for six weeks. (A) Extravaginal and (B) intravaginal tiller growth patterns in 6-inch pots; C) exposed stem node; D) expanded leaf blade with formed ligule adjunct to stem………………………………………………………………………………….…………...39 Figure 3.7 Average rhizome viability (%) of samples after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November/December 2015 and April 2016 and stored at 3°C for up to 140 days. Each bar is represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…..…………………………..….44 Figure 3.8 Average emergence speed (days) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) pooled and b) individual genotypes. Individual data points and bars are represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test….……………………………....45 Figure 3.9 Average emergence speed (days) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON

viii

(‘Amuri’ and ‘Nagara’) in April 2016 and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…...………………………………………………….….46

Figure 3.10 Average emergence speed (days) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015 and April 2016 and stored at 1 or 3°C for up to 140 days. Data are categorized by a) harvest season and b) genotype. Each data point represents 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……..……………………….…..47

Figure 3.11 Average tiller height (mm) of tallest tiller of samples grown for six weeks in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’) in May 2015 and stored at 0 and 3°C over 126 days. Each data point represents 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test………………………………………………………………………………………………..48

Figure 3.12 Average tiller height (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) pooled and b) individual genotype averages over the experimental period. Individual data points and bars represent 64 and 80 ± SE samples, respectively. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.….……………………………………………………………….………….49

Figure 3.13 Average tiller height (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’) in December and November 2016, respectively, and stored at 1°C for up to 140 days. Each data point represents 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey’s HSD test…………….……………………50

Figure 3.14 Average tiller number (#) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘UK’ and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples…..……………………………………...51 Figure 3.15 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) individual and b) pooled genotypes. Individual bars and data points represent 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……………..…………………...53 Figure 3.16 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) individual and b) pooled genotypes. Individual

ix bars and data points represent 80 and 64 ± SE samples, respectively. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…..…..…………….54 Figure 3.17 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April and November/December 2016 and stored at 3°C for up to 140 days. Each data point is represented by 64±SE samples….……………………………………....55 Figure 3.18 Average stem node number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples….…………………….56

Figure 3.19 Average stem node number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘UK’ and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples…………………………………...57

Figure 3.20 Average stem node number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November/December 2015 and April 2016 and stored at 3°C for up to 140 days. Measurements are categorized by a) pooled and b) individual genotypes. Individual data points and bars are represented by 80±SE samples…………………………………………..58

Figure 3.21 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November/December 2015 and April 2016 and stored at 3°C for up to 140 days. Measurements are categorized by a) pooled and b) individual genotypes. Individual data points and bars are represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…………………………...……..60

Figure 3.22 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April and November/December 2016 and stored at 3°C for up to 140 days. Each data point is represented by 64±SE samples…...……………………………………..61

Figure 3.23 Average rhizome MC (%) of samples stored at 3°C over 140 days. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…………………………………………………………………...... 63

Figure 3.24 Average starch content [%, w/w (DW basis)] of rhizomes stored at 0 and 1°C over 105 or 140 days, respectively. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’) in May/April 2015/2016 and November/December 2015/2016. Each data point is represented by 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to

Tukey's HSD test. *Timepoint is measured by 21-day (spring 2015 only) and 28-day intervals.…………………………………………………………………………………...……..65

Figure 3.25 Average starch content [%, w/w (DW basis)] of rhizomes stored at 1 or 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in a) April and b) November/December 2016. Each data point is represented by 32±SE samples….………………………………………………………………………………………..66

Figure 3.26 Total soluble carbohydrate concentration (mg g-1 DW) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in December and November 2016, respectively. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test….………………………………………………………………...68

Figure 3.27 Percentage of total soluble carbohydrates consisting of D-glucose (%) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test….………………………………………………………………………………………70

Figure 3.28 Average percentage (%) of total soluble carbohydrate concentration consisting of D- glucose in rhizomes stored at 1 and 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in a) April and b) November/December 2016. Each data point is represented by 32±SE samples…..………………………………..…….71

Figure 3.29 Percentage of total soluble carbohydrates consisting of sucrose (%) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each data point is represented by 16 samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. S.E. 8.8001……………………………………………………………………………………….73

Figure 3.30 Average percentage (%) of total soluble carbohydrate concentration consisting of sucrose in rhizomes stored at 1 and 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in a) April and b) November/December 2016. Each data point is represented by 32±SE samples…...……………………………………..74

Figure 3.31 Percentage of total soluble carbohydrates consisting of D-fructose (%) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……...…………………………………………………………………………………..76

Figure 4.1 Representative M. x giganteus ‘M161’ callus morphologies. Calli were assessed using a dissecting microscope and were putatively classified as: a) root forming; b) shoot forming; c) somatic embryo forming (K2); and d) compact yellow/green (K1). Leaf primordia (e) and differentiated shoot-like structures (f) formed on M. x giganteus ‘M161’ calli. Roots generated on

M. x giganteus ‘M161’ calli; the root-cap is identifiable by the production of red/purple pigmentation and lack of root hairs (g), and anthocyanin spots are noticeable on the remainder of the callus (h). Scale bars represent 1 mm…………….…...…………………………………..…..90 Figure 4.2 Proportion of shoot-, root-, and somatic embryo-forming calli over time. Data are presented by 0 (a and b), 1 (c and d), 10 (e and f), 100 (g and h), and 1000 (i and j) µM AIP treatments grouped by 9.0 (a, c, e, g, and i) and 11.3 (b, d, f, h, and j) µM 2,4-D…….…….…...... 92 Figure 4.3 Proportion of K1, K2, K3, and browning callus over time. Data are presented by 0 (a and b), 1 (c and d), 10 (e and f), 100 (g and h), and 1000 (i and j) µM AIP treatments grouped by 9.0 (a, c, e, g, and i) and 11.3 (b, d, f, h, and j) µM 2,4-D……………………….…….…………...94 Figure 4.4 Average number of shoots (a), roots (b), and somatic embryo-like structures (c) developing on calli cultured on different levels of AIP and averaged over time…..…………...... 96 Figure 4.5 Average number of a) shoots, b) roots, and c) somatic embryos per callus at specific culture times (S.E. 1.1484 and 6.8712 for shoot and somatic embryo number, respectively)…….97 Figure 4.6 Regenerated calli in rooting medium (a) and after culture in liquid tillering medium (b) after approximately 28 days of incubation ...………………………………………………..……99 Figure 5.1 Total shoot number (#) of MR induction treatments over the duration of 10 weeks. Each data point represents the mean of 30 samples (six treatments consisting of four replicates each), repeated twice. Because plants were assessed non-destructively at weeks 0 and 5, conservative counts were made…...………………………………………………………………………..………………113 Figure 5.2 Average weight (g) of full plantlets used for MR induction over the duration of 10 weeks. Each data point represents the mean of four replicate plantlets, repeated twice. The analysis was performed using a gaussian distribution and predicted values on the data scale were plotted 2 using a proc nlin polynomial model predictor: Y3%suc = -0.0397(time) + 0.5519(time) + 0.6858, 2 2 2 pseudo R = 0.69; Y8%suc = -0.0871(time) + 1.0123(time) + 0.7441, pseudo R = 0.35; Y10%suc = - 2 2 2 0.1539(time) + 1.6983(time) + 0.8107, pseudo R = 0.46; Y8%suc+PGR = -0.0527(time) + 2 2 0.8606(time) + 0.7031, pseudo R = 0.15; Y10%suc+PGR = -0.0407(time) + 0.6195(time) + 0.6781, 2 2 2 pseudo R = 0.59; Y3%suc+PGR1 = -0.0462(time) + 0.6537(time) + 0.6804, pseudo R = 0.47; and 2 2 Y3%suc+PGR2 = -0.0567(time) + 0.9036(time) + 0.6478, pseudo R = 0.20. Because plants were assessed non-destructively at weeks 0 and 5, conservative counts were made……………..…....114 Figure 5.3 Average plantlet weight (g) of MR induction treatments over the duration of 10 weeks. Each data point represents the mean of 28 samples (seven treatments consisting of four replicates each), and both trials are exhibited. The analysis was performed using a reciprocal gamma distribution model and predicted values on the data scale were plotted using a proc nlin quadratic 2 2 model predictor: Ytrial 1 = -0.0831(time) + 1.0798(time) + 0.7688, pseudo R = 0.41; Ytrial 2 = - 0.0531(time)2 + 0.7202(time) + 0.6455, pseudo R2=0.57. Means followed by the same letter at each timepoint are not significantly different (P<0.05) according to Tukey's HSD mean separation test……...………………………………………………………………………………….……115 Figure 5.4 Stages of MR isolation and regrowth. MR isolation from plantlets incubated in MR induction media after 10 weeks (A); microscopic images of MRs with immature buds (B) and young tiller (C); plantlet growth after 8 weeks of incubation of isolated MRs in tillering medium

xii

(D); field-harvested rhizomes after pruning, washing, and removal of most roots (E); and cross- section of processed field rhizome (F)…….…………………………………………………….117 Figure 5.5 Plantlet incubated in MR treatment 8% sucrose. Roots developed in liquid media after 10 weeks of culture (A) and healthy, isolated roots (B)………………………………………....119 Figure 5.6 RFW of shoot, root, and MR tissue after 10 weeks of incubation in MR induction treatments. Each bar represents the mean of four replicate plantlets, repeated twice. Means of each tissue followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test…..………………………………………………………………..….121 Figure 5.7 Figure 4.7 Ex vitro plantlet growth from MRs isolated from induction medium samples and transferred directly to the mist bed. Photos were taken six weeks after planting. (A) germinated tiller in the mist bed, (B and C) harvested plantlets with shoot and root structures attached…..……………………………………………………………………………………..123

Figure 5.8 Stages of synthetic seed production. Healthy, green buds were assessed on isolated MRs using a dissecting microscope (A), and MRs were cut into small sections containing at least one bud. Sections were then encapsulated in 3% (w/v) sodium alginate and 1.0 % (w/v) calcium chloride dihydrate solution (B). Samples were then allowed to dry in a flow bench overnight (C). Early (D) and late (E) bud emergence from the matrix after 4 and 8 weeks in tillering medium were observed using a dissecting microscope. Full plantlet development from synthetic seeds were observed after 8 weeks in tillering media (F)…………………………………………..………..125

xiii

List of Tables

Table 3.1 Origins of five M. x giganteus genotypes used in the current study…………………….29

Table 3.2 General flowering and senescence times, stand ages, and other pertinent information for the five M. x giganteus genotypes used in the current study……………………………………...30 Table 3.3 Average percentage (%) of total soluble carbohydrate concentration consisting of D- fructose in rhizomes stored at 1 and 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April and November/December 2016. Each value is represented by 32±SE samples…...……….………….76

Table 5.1 Change (%) in plantlet shoot number and weight over the duration of the MR induction period. Each mean represents four replicate plantlets, repeated twice. Because plants were assessed non-destructively at weeks 0 and 5, conservative shoot number counts were made. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…...……………………………………………………………………………………….…113 Table 5.2 Total MR number (#), average MR node number (#), average MR length (mm), average MR FW yield per plantlet (g), and average MR DW yield per plantlet (g) after 10 weeks of incubation in MR induction media. Each value represents the mean of four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……...…………………………………………………………...116 Table 5.3 Total shoot, root, and MR MC (%) after 10 weeks of incubation in MR induction media. Each mean represents four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test……..………………………………………………………………………………………..118 Table 5.4 Average root length (mm), total root number (#), average shoot length (mm), average shoot node number (#), average shoot chlorosis (%), and average chlorophyll content (mg m-2) after 10 weeks of incubation in MR induction media. Each mean is represented by four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…………………………………………………………………..120 Table 5.5 (A) Total bud number at weeks 4 and 8 (#), and change in bud number (%) over the duration of the incubation period of all MR treatments in tillering medium. Each mean represents 32 MRs, repeated twice. Analysis was conducted using an RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. (B) Total bud number at weeks 4 and 8 (#), and change in bud number (%) over the duration of the incubation period of MR treatments in tillering medium after CS at 3°C. Each mean is represented by 32 MRs, repeated twice. Analysis was conducted using a split-plot design, with freezer representing the main plot and either plates (CS-C and CS-D) or flasks (CS-control and CS-ABA) representing subplots. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……...……………………………………….122

xiv

List of Appendices Appendix 3.1 Type III test for the significance of main effects and their interactions (‘temperature’, ‘storage time’, and ‘temperature x storage time’) on plant growth traits for ‘Illinois’ after spring 2015 harvest and storage at 0 or 3°C (α=0.05)…..………………………………………………151 Appendix 3.2 Type III test for the significance of main effects and their interactions (‘genotype’, ‘storage time’, and ‘genotype x storage time’) on plant growth traits for ‘Illinois’ (autumn 2015 and spring 2016 only), ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’ after autumn 2015, spring 2016, and autumn 2016 harvests and storage at 3°C (α=0.05)……….…………………………………….152 Appendix 3.3 Type III test for the significance of main effects and their interactions (‘genotype’, ‘storage time’, and ‘genotype x storage time’) on plant growth traits for ‘Illinois’, ‘UK’, ‘BC’, and ‘Amuri’ after autumn 2015 (‘Illinois’ and ‘UK’ only), spring 2016, and autumn 2016 harvests and storage at 1°C (α=0.05)…...………………………………………………………………….….153 Appendix 3.4 Type III test for the significance of main effects and their interactions (temperature, storage time, and temperature x storage time) on rhizome physiological traits for ‘Illinois’ after spring 2015 harvest and storage at 0 or 3°C (α=0.05)…..…………………………….…………154 Appendix 3.5 Type III test for the significance of main effects and their interactions (genotype, storage time, and genotype x storage time) on rhizome physiological traits for ‘Illinois’ (autumn 2015 and spring 2016 only), ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’ after autumn 2015, spring 2016, and autumn 2016 harvests and storage at 3°C (α=0.05)……..…………………………………..155 Appendix 3.6 Type III test for the significance of main effects and their interactions (genotype, storage time, and genotype x storage time) on rhizome physiological traits for ‘Illinois’, ‘UK’, ‘BC’, and ‘Amuri’ after autumn 2015 (‘Illinois’ and ‘UK’ only), spring 2016, and autumn 2016 harvests and storage at 1°C (α=0.05)……..………………….………………………………….156 Appendix 3.7 Parameters estimates results for the change in emergence speed of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05)…….……………………………………..157 Appendix 3.8 Parameter estimates results for the change in emergence speed of ‘Illinois’ harvested in spring 2016 and stored at 3⁰C (α=0.05)…...…………………………………………………..157 Appendix 3.9 Parameter estimates results for the change in emergence speed of ‘UK’ harvested in spring 2016 and stored at 3⁰C (α=0.05)……...…………………………………………………..157 Appendix 3.10 Parameter estimates results for the change in emergence speed of ‘Illinois’ harvested in autumn 2015 and stored at 1⁰C (α=0.05)….………………………………………..157 Appendix 3.11 Parameter estimates results for the change in emergence speed of ‘UK’ harvested in autumn 2015 and spring 2016 and stored at 1⁰C (α=0.05)…..………………………………..157 Appendix 3.12 Parameter estimates results for the change in tiller height of ‘Illinois’ harvested in spring 2015 and stored at 3⁰C (α=0.05)…...……………………………………………………..158 Appendix 3.13 Parameter estimates results for the change in tiller height of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05)….……………………………………………………..158

Appendix 3.14 Parameter estimates results for the change in tiller height of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)….……………………………………………………..158 Appendix 3.15 Parameter estimates results for the change in tiller height of ‘Amuri’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)….……………………………………………………..158 Appendix 3.16 Parameter estimates results for the change in tiller height of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)….……………………………………………………..158 Appendix 3.17 Parameter estimates results for the change in tiller height of ‘Illinois’, ‘UK’ ‘BC’, and ‘Amuri’ harvested in spring 2016 and stored at 1 or 3⁰C (α=0.05)…...…………………….159 Appendix 3.18 Parameter estimates results for the change in tiller number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)…….……………………………………………….159 Appendix 3.19 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2015 and stored at 3⁰C (α=0.05)…….………………………………………………….159 Appendix 3.20 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)….……………………………………………………..159 Appendix 3.21 Parameter estimates results for the change in leaf number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)….……………………………………………………..159 Appendix 3.22 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)….…………………………………………………..…159 Appendix 3.23 Parameter estimates results for the change in leaf number of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)….…………………………………………………..…160 Appendix 3.24 Parameter estimates results for the change in stem node number of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05)…….…………………………………..…160 Appendix 3.25 Parameter estimates results for the change in stem node number of ‘Amuri’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)…….………………………………….….160 Appendix 3.26 Parameter estimates results for the change in stem node number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)…….………………………………….…160 Appendix 3.27 Parameter estimates results for the change in stem node number of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)…….………………………………………………..160 Appendix 3.28 Parameter estimates results for the change in stem diameter of ‘Amuri’ harvested in spring 2016 and stored at 3⁰C (α=0.05)…...…………………………………………………..161 Appendix 3.29 Parameter estimates results for the change in stem diameter of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)….…………………………………………………..161 Appendix 3.30 Parameter estimates results for the change in stem diameter of ‘UK’ harvested in autumn 2015 and spring 2016 and stored at 3⁰C (α=0.05)…..…………………………………...161

xvi

Appendix 3.31 Parameter estimates results for the change in chlorophyll content of ‘Amuri’ harvested in spring 2016 and stored at 1⁰C (α=0.05)…..………………………………………...161 Appendix 3.32 Parameter estimates results for the change in chlorophyll content of ‘BC’ harvested in spring 2016 and stored at 1⁰C (α=0.05)….…………………………………………………....161 Appendix 3.33 Parameter estimates results for the change in total NSC of ‘UK’ harvested in autumn 2016 and stored at 3⁰C (α=0.05)…….…………………………………………………..162 Appendix 3.34 Parameter estimates results for the change in percent total soluble carbohydrates consisting of D-glucose of ‘UK’ harvested in autumn 2015 and stored at 1⁰C (α=0.05)………..162 Appendix 3.35 Parameter estimates results for the change in total NSC of ‘UK’ harvested in autumn 2016 and stored at 1⁰C (α=0.05)…….…………………………………………………..162 Appendix 3.36 Parameter estimates results for the change in percent total soluble carbohydrates consisting of D-glucose of ‘UK’ harvested in spring 2016 and stored at 1⁰C (α=0.05)…..……..162 Appendix 3.37 Parameter estimates results for the change in percent total soluble carbohydrates consisting of sucrose of ‘BC’ harvested in spring 2016 and stored at 1⁰C (α=0.05)…...………..162

Appendix 3.39 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 0 and 3°C over 126 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’) in May 2015. Each value is represented by 32±SE samples………………………………………………………………………………164 Appendix 3.40 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively. Measurements were taken 6 weeks after planting. Each value is represented by 80±SE samples………………………………………………………………………………….165 Appendix 3.41 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively. Each value is represented by 80±SE samples………………………………………………………………………………166 Appendix 3.42 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015. Measurements were taken 6 weeks after planting. Each

xvii data point is represented by 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test………………………………...167 Appendix 3.43 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015. Each data point is represented by 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test……….168 Appendix 3.44 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016. Each value is represented by 80±SE samples…………169 Appendix 3.45 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016. Each value is represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test…………………………………………………………………………….170 Appendix 3.46 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each value is represented by 64±SE samples…………………………171 Appendix 3.47 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each value is represented by 64±SE samples…………………………172 Appendix 3.48 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively. Each value is represented by 64±SE samples………………………………………………………………………………………….173 Appendix 3.49 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and

xviii

‘Nagara’) in December and November 2016, respectively. Each value is represented by 64±SE samples………………………………………………………………………………………….174 Appendix 3.50 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in December and November 2016, respectively. Each value is represented by 64±SE samples………………………………………………………………………………………….175 Appendix 3.51 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in December and November 2016, respectively. Each value is represented by 64±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test……………………………………………………………...176 Appendix 3.52 Average tiller height (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015 and April 2016 and stored at 1 or 3°C for up to 140 days. Each data point is represented by 64±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test….…………………………………………………………….…177 Appendix 3.53 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016 and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test……..……………………….…177 Appendix 3.54 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Each bar is represented by 80±SE samples…...……………………..….178 Appendix 3.55 Average stem diameter (mm) of the tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015 and April 2016 and stored at 1 or 3°C for up to 140 days. Measurements are categorized by a) harvest season and b) individual genotypes. Individual data points and bars are represented by 64 and 160 ± SE samples, respectively. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…..………………………….…179

Appendix 3.56 Total soluble carbohydrate content (mg g-1 DW) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test………………………………...180

xix and Elora, ON (‘Amuri’) in April and November/December 2016. Each bar is represented by 80±SE samples…………………………………………………………………………….……181 Appendix 4.1 Average GAE (µg g-1 DW) values of calli cultured on mediums supplemented with various AIP levels (S.E. 149.03). Each bar represents the mean of 50 calli over two 2,4-D levels (10 callus samples over five replications). Bars labeled with the same letter are not significantly different (P<0.05) according to Tukey's HSD test……...………………………………….……182 Appendix 4.2 Total MC (%) of calli used for soluble phenolic content measurements. Analyzes were conducted separately for AIP and 2,4-D. Each mean represents 50 and 100 calli (10 and 25 callus samples over five replications, respectively). Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test…...…………………...... 182

Appendix 5.1 Total shoot number (#) of plantlets destined for destructive measurements over the duration of the MR induction period. Because mean differences were significant between trials, both experimental trials are exhibited. Each mean represents four replicate plantlets, and the analysis was conducted using a RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test……………………………..183 Appendix 5.2 Total weight (g) of plantlets destined for destructive measurements over the duration of the MR induction period. Because mean differences were significant between trials, both experimental trials are exhibited. Each mean represents four replicate plantlets, and the analysis was conducted using a RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test………………………………………183 Appendix 5.3 Average shoot length (mm), chlorophyll content (mg m-2), and shoot chlorosis (%) after 10 weeks of MR induction treatment with (CS-control and ABA) and without (8%sucrose+PGR) four weeks of 3°C CS. Each mean represents four replicate plantlets (repeated twice), and the analysis was conducted using a RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test……..…….…..183 Appendix 5.4 Weight (g) and weight change (%) of full plantlets after four weeks of 3°C CS. The analysis was conducted using a split-plot design (repeated twice), with freezer acting as the main plot (replicated twice) and flasks acting as the subplots (replicated four times). Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test…...……………………………………………………………………..………..184 Appendix 5.5 Raw data of various MR treatments that developed full plantlets in vitro after 8 weeks of culture in tillering media. Due to sparse germination and development of plantlets, statistical analysis was not conducted on data…...…………………………………….………...184 Appendix 5.6 Raw data of 8%sucrose+PGR treatment that developed full plantlets ex vitro after 6 weeks of culture in a mist bed. Due to sparse germination and development of plantlets, statistical analysis was not conducted on data…...………………………………………………………...184 Appendix 5.7 Growth assessment of MRs after four weeks of 3°C CS. CS-D (A and B) and CS-C (C and D) samples after 8 weeks of incubation in tillering media; healthy, intact bud on MR (E); tillering media containing MRs that were dried (blue labels) and not dried (green labels) after fours weeks of CS (F)…….…………………………………………………………………………...185

List of Abbreviations and Nomenclature 2,4-D………………………………………………………………2,4-dichlorophenoxyacetic acid ABA…………………………………………………………………………………..Abscisic acid AGPase…………………………………………………………..ADP-glucose pyrophosphorylase AFLP…………………………………………………...Amplified fragment length polymorphism AIP………………………………………………………………2-aminoindan-2-phosphonic acid AMG……………………………………………………………………………Amyloglucosidase BAP…………………………………………………………………………..6-benzylaminopurine CEEDS™………………………………………Crop Expansion Encapsulation & Drilling System CRD………………………………………………………………..Completely randomized design CS……………………………………………………………………………………...Cold storage DW……………………………………………………………………………………...Dry weight DRP………………………………………………………………………..Direct rhizome planting EP…………………………………………………………………………………..Embryoid plant F……………………………………………………………………………………..Dilution factor F-6-P…………………………………………………………………………Fructose-6-phosphate F-C…………………………………………………………………………………Folin-Ciocalteu FK……………………………………………………………………………………..Fructokinase FW…………………………………………………………………………………….Fresh weight GA3………………………………………………………………………………...Gibberellic acid GAE………………………………………………………………………...Gallic acid equivalents ‘G’ x ‘E’………………………………………………………………‘Genotype’ x ‘Environment’ G-6-P…………………………………………………………………………Glucose-6-phosphate GHG…………………………………………………………………….………….Greenhouse gas HK…………………………………………………………………….………………..Hexokinase HSD…………………………………………………………………..Honest significant difference IAA…………………………………………………………………………….Indole-3-acetic acid IBA…………………………………………………………………………...Indole-3-butyric acid ISSR…………………………………………………………………..Inter simple sequence repeat ITS………………………………………………………………………Internal transcribed spacer K1……………………………………………………………..Compact white, embryogenic callus K2…………………………………………………Nodular yellow/green, non-embryogenic callus K3………………………………………………………………...Friable, non-embryogenic callus KT…………………………………………………………………………………………..Kinetin MC………………………………………………………………………………..Moisture content MR…………………………………………………………………………………..Microrhizome MS…………………………………………………………………………...Murashige and Skoog MT………………………………………………………………………………………Microtuber NAA………………………………………………………………………..Naphthaleneacetic acid NADP+………………………………….Nicotinamide adenine dinucleotide phosphate (oxidized) NADPH………………………………….Nicotinamide adenine dinucleotide phosphate (reduced) NSC…………………………………………………………………...Non-structural carbohydrate NUE………………………………………………………………………..Nitrogen use efficiency PAL………………………………………………………………….Phenylalanine ammonia lyase PCR……………………………………………………………………..Polymerase chain reaction PGI……………………………………………………………………..Phosphoglucose isomerase

xxi

PGR…………………………………………………………………………Plant growth regulator PMS………………………………………………………………………..Potato minituber sprout R2……………………………………………………………………..Coefficient of determination RAPD………………………………………………………Random amplified polymorphic DNA RCBD…………………………………………………………Randomized complete block design RDW…………………………………………………………………………...Relative dry weight RFW………………………………………………………………………….Relative fresh weight SOC…………………………………………………………………………….Soil organic carbon SuSy………………………………………………………………………………Sucrose synthase T…………………………………………………………………………………………Time point TIBA……………………………………………………………………..2,3,5-triiodobenzoic acid UDP-glucose………………………………………………………….Uridine diphosphate glucose v/v………………………………………………………………………………Volume by volume WUE…………………………………………………………………………..Water use efficiency w/v……………………………………………………………………………….Weight by volume

xxii

1: Introduction 1.1 Summary

Crops cultivated for bioenergy and biofuel production is a prevalent trend in many developed nations (eg. Sweden, United Kingdom, Germany, and the United States) for progressive replacement of non-renewable heat, fuel, and electricity sources (Clifton-Brown et al. 2004; Heaton et al. 2008). It is estimated that biomass feedstocks will eventually be able to deliver 80- 90% reductions in greenhouse gas (GHG) emissions compared to fossil fuels (El Kasmioui and Ceulemans 2013), aide in preserving native ecosystems (Haughton et al. 2009) and boost local economies by encouraging jobs related to green energy (Heaton et al. 2004).

Annual demand for ethanol in Canada is estimated at approximately 3.1 billion litres, with approximately two-thirds of the supply being allocated to 2% and 5% renewable-content blends for diesel and gasoline, respectively (Government of Canada 2012; Dessureault 2015). These mandates were implemented by the Government of Canada in 2010 in an effort to reduce GHG emissions by 17% between 2005 and 2020, and to stimulate local job creation (Government of Canada 2012); however, because of the stringent demands for national ethanol production in conjunction with unstable oil prices, the country is obligated to import around 20% of their annual supply from the United States (Dessureault 2015). Currently, the main sources for bioethanol in Canada are first-generation feedstocks corn (77%) and Triticum spp. (wheat) (23%) (Dessureault 2015). Issues arising from our current production system include: land and feedstocks suitable for food being converted to fuel; unstable national energy security; and the economic infeasibility of using corn – with an energy balance (energy input:output) of 1.5 (McCalmont et al. 2017) – as a biofuel feedstock when more suitable candidate species are available.

Perennial second-generation biofuel feedstocks are a popular alternative to annual first-generation sources due to their high lignocellulosic cell wall compositions (Le Ngoc Huyen et al. 2010), potential for growth on land not suitable for food production (Lewandowski 1998), and improvement of soil richness and ecological fitness (Semere and Slater 2007). One of the top candidate species in this category is Miscanthus x giganteus Andress. (M. x giganteus) who’s stems can be harvested for bioethanol production, heat generation via combustion (Heaton et al. 2004), paper and pulp, bioplastics, and bedding (Engbers and Deen 2013). The life-cycle of M. x giganteus can also be beneficial for soil organic carbon (SOC) remediation (Matthews and Grogan

2001), phytoremediation of soil-bound cadmium (Arduini et al. 2004), and improvement of flora and fauna biodiversity (Haughton et al. 2009; Semere and Slater 2007).

Favourable physiological characteristics of C4 NADH-malic acid Miscanthus (Wang et al. 2008) -2 -1 includes high solar radiation efficiency and photosynthetic rates (20 μmol CO2 m s ) (Defra -1 2002), high water use efficiency (WUE) (22 g DM l H2O) (Beale et al. 1999) and nitrogen use efficiency (NUE) (Zub and Brancourt-Hulmel 2010), rapid growth rate (50 g m-2 day-1) (D. G. Christian and Haase 2001), and few pests known to exhibit negative consequences on overall plant growth (Covarelli et al. 2012). M. x giganteus has an energy balance of approximately 0.2, significantly more efficient than other bioenergy systems such as canola and corn (equal to or greater than 0.8) (McLaughlin and Walsh 1998), and results in a 61% higher harvestable biomass yield compared to corn due to its earlier emergence and high levels of photosynthetic activity late into the growing season (Dohleman and Long 2009).

When establishing M. x giganteus plantations using traditional propagation techniques, rhizomes are partitioned into specific sizes ranging between 3-20 cm in length, 2.6-80 g fresh weight (FW) per cutting (Pyter et al. 2007; Pyter et al. 2010; Xue et al. 2015) and planted at a soil depth of 5 to 10 cm (Pyter et al. 2010). Depending on the landscape and soil type, M. x giganteus rhizomes can be planted in densities between 10,000 (Jeżowski et al. 2011) and 40,000 (Bullard et al. 1995) plants ha-1. In Ontario, it has been demonstrated that M. x giganteus should be planted in early spring and allowed to grow for two- to three-years for crop establishment and optimal yield potential (Pyter et al. 2009).

Initial crop establishment has been identified as a critical step to maximize peak harvest yields, as demonstrated in studies conducted in Illinois and Poland (Heaton et al. 2008; Jeżowski et al. 2011). Findings from various field plots in Poland indicated that biomass per 30 Miscanthus plants experienced an above-ground yield increase of 0.14-0.86 kg between year one and two of growth (Jeżowski et al. 2011). In Illinois field trials, plants reached an average of 3-4 m in height by the end of the growing season, and 30 to 61 t ha-1 yields after the third year of establishment (Heaton et al. 2008). Following the establishment period, M. x giganteus plantations can maintain an optimal annual yield capacity for an additional 20-25 years (Xue et al. 2015), and Heaton et al. (2004) calculated that a net profit of approximately $2,900 ha-1 could be obtained with plantations by the 10th year of production.

M. x giganteus is a sterile triploid species with commercial propagation occurring vegetatively, primarily through rhizome cuttings (Lewandowski 1998), rhizome-derived plants (Xue et al. 2015), or micropropagated plantlets (Atkinson 2009). After taking specific parameters for farmers in different countries (Germany, Britain, Austria, and Canada) into consideration (cost of land, production materials, fixed machine costs and storage, and variable machine costs), direct rhizome planting (DRP) is currently regarded as the most cost-effective method for establishing M. x giganteus plantations (3,375 € ha-1), while rhizome-derived plants (4,400.80 € ha-1) and micropropagation methods (6,320 € ha-1) remain the most cost-intensive (Xue et al. 2015). DRP is also advantageous over micropropagated methods due to the formation of thicker shoots, and reduced water, ash, nitrogen, and potassium contents in the above-ground biomass, resulting in enhanced combustion quality (Lewandowski and Kicherer 1997). The morphological and physiological differences found between plants originating from DRP and micropropagation is hypothesized to be caused by an imbalance of nutrient stores between above-ground and below- ground tissues (Greef et al. 1994).

Using DRP, Miscanthus growers have either focused on rhizome multiplication for further propagation (40,000 plants ha-1) or for biomass production (commercial densities typically ranging from 12,000 to 16,000 plants ha-1) (Atkinson 2009). Rhizome harvesting, splitting, and replanting is achieved in situ, and can either occur in late autumn/early winter (Atkinson 2009) or in early spring after the last winter frost has occurred (Xue et al. 2015).

Currently, the average price for Miscanthus cultivation via DRP restricts widespread adoption by farmers. The average establishment price in Canada (at 20,000 plants ha-1) is $5.430 ha-1. The bulk of this cost originates from initial germplasm expenses (each propagule costing $0.15-0.54) and a lack of specialized harvesting machinery, which results in increased time and labour (Atkinson 2009).

To combat these obstacles, All Weather Farming Inc. (based out of Port Ryerse, Ontario) has engineered a harvester specified for Miscanthus rhizomes. With this technology, money will be saved in terms of labour and time, while harvesting and processing will be streamlined. The goal of applying this technology to Miscanthus rhizomes is to increase rhizome multiplication and reduce initial germplasm costs to a fraction of the current market price. To further optimize the

3 system, this study investigated preferable rhizome harvesting time, CS conditions, and carbohydrate dynamics that reflect maximal sprout emergence and plant vigor.

The advantage of late autumn/early winter rhizome harvest is based on having an extended window of opportunity for early spring planting. Storage of rhizomes over winter months can also be favourable for sales within the national marketplace. Due to vast differences in climates throughout Ontario (and Canada) over the winter – including average snowfall, daily temperatures, and date of the last frost of the season – these regions may experience different rhizome planting/shoot emergence dates (Rosser 2012). Rhizome establishment in the spring is critical for maximal harvestable stem yields and enough time and resources for the plant to allocate essential minerals and nutrients to the rhizome for overwintering.

Immediately after harvesting M. x giganteus rhizomes in late autumn, rhizomes must be washed, processed, and stored in cold storage (CS) conditions until the growing season commences. Currently, there have been no studies conducted on optimizing conditions for long-term CS of Miscanthus rhizomes. In Chapter 3 of this thesis, it is hypothesized that agronomic practices and physiological characteristics of Miscanthus rhizomes would help in determining optimal conservation protocols to allow for maximal plant regrowth when required. Investigation of these components will be critical when assessing favourable characteristics between candidate genotypes for commercial growth in Ontario.

Though micropropagation is not economically viable for M. x giganteus at the present time, this technology can be valuable for other reasons, such as the creation of certified, disease-free germplasm used for national and international trade and establishing nurseries, the conservation of elite genotypes, and a foundation for in vitro breeding. Though Miscanthus micropropagation has been initiated and established (Ślusarkiewicz-Jarzina et al. 2017), obstacles that have yet to be reported on include: regeneration of plantlets after long-term storage (investigated in Chapter 4); development of microrhizomes (MRs) for use as alternative propagules; and the ability of utilizing these tissues for the creation of synthetic seeds (assessed in Chapter 5). To tackle these challenges, a wide variety of tissue culture techniques will be applied to calli induced from field-grown immature inflorescences, established in vitro plantlets to induce MRs, and encapsulation of sectioned MRs into alginate beads. Furthermore, regeneration of shoots and emergence of culms will be assessed from these propagation methods.

2: Literature Review 2.1 Miscanthus spp. 2.1.1 Miscanthus ancestry and taxonomy

Miscanthus spp. consists of monocotyledonous, true grass species from the Poaceae family, subfamily Panicoideae, and tribe Andropogoneae (Clayton and Renvoize 1986; Hodkinson et al.

2002). The Andropogoneae tribe consists of many economically-important C4 food, feed, and fuel resource crops, including Saccharum officinarum L. (sugarcane), Sorghum bicolor L. (sorghum), and Zea mays L. (maize) (Hodkinson et al. 2002). Miscanthus is also a member of the subtribe Saccharinae (Głowacka 2011), made evident by the high sequence similarity and phylogenetic relationship to the Saccharum genus via comparison of sequences of the internal transcribed spacer (ITS) of nuclear ribosomal DNA, and plastid DNA trnL and trnL-F (Hodkinson et al. 2002). Currently, there are an estimated 14-20 Miscanthus species sanctioned by horticultural organizations globally (Cruz and Dierig 2015), with the majority of species originating from East Asia (Clifton-Brown et al. 2004) and Pacific Islands, ranging from temperate Siberian to tropical Polynesian regions (Cruz and Dierig 2015).

Amongst the Miscanthus genus, M. x giganteus has been deemed the highest-ranking candidate for terrestrial lignocellulosic biomass production (Clifton-Brown and Lewandowski 2000; Clifton- Brown et al. 2001) for a range of end-product formulations and uses (Engbers and Deen 2013). In previously reported literature, M. x giganteus has been categorized as Miscanthus x ogiformis Honda ‘Giganteus’ (Hansen and Kristiansen 1997), Miscanthus sinensis (Thunb.) Anderss. ‘Giganteus’ (Nielsen et al. 1993), and Miscanthus sacchariflorus var. bredidarbis (Cruz and Dierig 2015). M. x giganteus was first introduced into the European market from Japan in 1935 for ornamental purposes (Cruz and Dierig 2015). The origins of allotriploid M. x giganteus (3x = 57) is hypothesized to be from an interspecific cross between diploid Miscanthus sinensis (M. sinensis) (2x = 38) and tetraploid Miscanthus sacchariflorus (M. sacchariflorus) (4x = 76) (Linde-Laursen 1993).

2.1.2 Aboveground physiology and morphology

The diversity of stem physiologies and morphologies are evident within the Miscanthus genus and across genotypes. A single grass unit (metamer) “consists of a blade, sheath, nodes, internodes, axillary buds”, and (occasionally) adventitious roots (Janišová 2006). M. x giganteus exhibits

5 intermediate stem morphology, representing phenotypic characteristics of both natural parents. Diploid M. sinensis is characterized by a tufted (caespitose) phenotype with high stem density, few hairs observed on the stems, no branching culms, leaves exhibiting a scabrous margin (Adati and Shiotani 1962), and foliated stems during the winter (Moon et al. 2013). In contrast, tetraploid M. sacchariflorus exhibits tall culms which branch into stems or roots from culm nodes (Adati and Shiotani 1962). M. sacchariflorus is also characterized and differentiated from M. sinensis by low culm abundance (diffuse), hollow stems for adaptation to high moisture conditions, dense bristles along young leaves (Adati and Shiotani 1962), and defoliated stems in the winter (Moon et al. 2013).

Miscanthus spp. exhibit tough inflorescence rachis and pedicellate spikelet pairs, with the length of paired pedicles differing from one another (Hodkinson et al. 2002). Differences arise in M. sinensis, where inflorescence is generally characterized by the presence of awned spikelets, while awnless spikelets are typical of M. sacchariflorus (Moon et al. 2013).

Miscanthus spp. can develop either extravaginal or intravaginal tillers (differentiated by the development of tillers exiting through subtending leaf sheath or within the leaf sheath, respectively) (Brejda et al. 1989). Intravaginal tillering morphology is defined by the presence of vertical (apogeotrophic) rhizomes that are typically unbranched and determinate (sympodial), with the resulting tillers developing close in proximity to the previous years’ tiller (Brejda et al. 1989). The newly-form tillers eventually deviate from the mother tiller/main axis, eventually forming its own rooting system; however, these remain connected to the mother tiller through vascular systems (Janišová 2006).

2.1.3 Rhizome physiology and morphology

Belowground organs such as tubers (Solanum tuberosum), bulbs (Allium spp.), and rhizomes (Miscanthus spp., Zingiber spp., Curcuma spp.) are specialized tissues that allow for the storage of vital nutrients and minerals for perennial plants that endure periods of unfavourable growing conditions (Brejda et al. 1989; Clifton-Brown and Lewandowski 2000). Rhizomes are characterized as subterranean shoot/stem organs that can grow either apogeotrophically or horizontally (plagiotropic), though some grass species can demonstrate both rhizome growth types (eg. Panicum virgatum L.) (Brejda et al. 1989). Rhizomes are composed of nodal and terminal buds which exhibit meristematic activity and can either produce new rhizome tissue or differentiate

6 into shoots or roots (Atkinson 2009). Depending on the species, rhizome morphology can be very diverse. M. x giganteus rhizomes are characterized by thick, semi-trailing stems, and appear in an oval or round conformation (Xue et al. 2015). In contrast, diploid parent M. sinensis exhibit rhizomes that are thin-stemmed and generally short (relative to other Miscanthus species), while M. sacchariflorus develops thick and broad-trailing stems (Xue et al. 2015). Withers (2015) investigated the rhizome morphology of diploid, allotriploid, and tetraploid Miscanthus species and genotypes and correlated their growth habits to cold tolerance. The ability of Miscanthus to overwinter is partially based on aboveground tissues remaining green to uphold operational photosynthetic machinery until the first frost of the winter season (Friesen et al. 2014). Sympodial rhizomes develop upward to subsequently give rise to new aerial shoot(s), and new determinate rhizomes develop at their base when the shoot(s) reach maturity (Withers 2015). Indeterminate (monopodial) rhizomes grow laterally through the soil, branching at nodes, and give rise to shoots at axillary buds developed along the length of the rhizomes or at the apex terminal ends (Brejda et al. 1989). Spreading rhizomes are suggested to exhibit poor cold tolerance compared to clumping-type rhizomes (Clifton-Brown and Lewandowski 2000), though there is evidence to the contrary (Rosser et al. 2012).

During harsh conditions, rhizomes endure a period of endodormancy that begins during rhizome commencement and ceases during bud maturity and culm emergence (Beale and Long 1997). Lang (1987) defined tissue dormancy as any plant material containing a viable meristem that exhibits temporary growth suspension (even under favourable conditions), and is governed by genetic (Barling et al. 2013), biochemical (reduced mitochondrial respiration, manipulated concentrations of plant growth regulators (PGRs), and altered carbohydrate metabolism) (Davies et al. 2011; Płażek et al. 2011; Withers 2015), and environmental (eg. daily minimal temperatures, flowering, and stem senescence) factors (Jensen et al. 2013; Purdy et al. 2015). Regulating the timing of dormancy and maintaining tissue integrity for optimal regrowth at the beginning of the growing season is of great importance to enhance overall stem/rhizome yield, favourable physiological characteristics, and preservation of overwintering capacity.

Research focusing on Solanum tuberosum (potato) tuber genetics, biochemistry, and physiology has surpassed that of other plant storage organs (Alam 1992; Tognetti et al. 2013) and can be used as a reference to understand the physiology and development of Miscanthus rhizomes both in vivo

7 and in vitro. Rhizome-forming species such as Zingiber spp. (ginger) and Curcuma spp. (turmeric) can also provide valuable insight to the factors affecting MR production, storage, and acclimation to greenhouse/field conditions.

2.2 Seedless triploid propagation 2.2.1 Storage and establishment properties

Few studies have investigated in situ M. x giganteus rhizome storage (Pyter et al. 2010; Davies et al. 2011; Xue et al. 2015) and dormancy (Beale and Long 1997; Purdy et al. 2015) characteristics, and instead have focused on freezing tolerance and overwintering capacity (Lewandowski 1998; Clifton-Brown and Lewandowski 2000; Płażek et al. 2011; Rosser 2012; Friesen et al. 2014; Peixoto 2015; Withers 2015). To obtain lower establishment costs, propagule expenditure must be reduced for farmers to invest in cultivating Miscanthus; this can be achieved with streamlined rhizome harvest and planting techniques.

In Canada, rhizome harvest and planting times vary drastically due to location. Because of this variability, spring rhizome harvests allow minimal time for processing and replanting, as well as potential financial losses for farmers anticipating to sell germplasm to other regions of the country. If the harvest period in one area of the country transpires after the optimal planting period of another area in the spring, this will result in economic drawbacks for the producer. Optimal planting period is defined as the span of time which growth potential for biomass production is greatest over a growing season, and is dependent on two variables: 1) rhizome dormancy break and 2) favourable weather conditions (including soil moisture content, soil temperature, air temperature, etc.).

To overcome this obstacle, rhizome harvest and processing in autumn may be executed, and germplasm could then be stored at reduced temperatures to induce/maintain dormancy (Farage et al. 2006; Pyter et al. 2010) until required for distribution/planting. Rhizome storage studies have either been conducted after harvest in late autumn (December-January) (Davies et al. 2011) or in the spring after the ground has thawed (Pyter et al. 2010; Xue et al. 2015). To date, there has been no reported literature comparing storage and regrowth of rhizomes harvested in autumn verses those gathered in the spring, and whether the regrowth from these propagules collected during the two seasons yield similar values.

Particular non-destructive growth parameters of above-ground tissues emerging from rhizomes during establishment have been disclosed as predictive of expectations (such as plant height and yield) for stands in their third year (Clifton-Brown et al. 2002; Rosser 2012). (Rosser (2012) determined that culm height and basal stem circumference across 20 Miscanthus genotypes were associated with winter survival in the field, indicating that these criteria may be applied for measuring plant vigor. In a three-year experimental period using M. sinensis – one of the progenitors of M. x giganteus – analysis of trait-marker associations determined that characteristics including plant height, leaf length and width, culm circumference, internode number, and tiller number correlated to biomass yield (Nie et al. 2016). Though there was vast phenotypic variation across the 138 genotypes tested, plant height remained a consistently stable predictable trait, and tiller number, biomass FW, and biomass DW exhibited phenotypic coefficients of variation exceeding 40% (Nie et al. 2016). Moreover, Jeżowski et al. (2011) observed significant ‘genotype’ (‘G’) x ‘cultivation year’ interactions for a number of characteristics of field-grown Miscanthus in the first three years of growth, including stem and clump diameter, and ‘G’ x ‘cultivation year’ x ‘location’ interactions for plant height, tillering, and biomass yield. Though these studies confirm associations between easily quantifiable measurements at field establishment and potential biomass yields in the third year of growth, stands were established using micropropagated plants instead of rhizomes, and these two propagation methods have been documented to exhibit different establishment attributes (Lewandowski 1998).

For rhizomes to be stored effectively, temperatures and tissue MC must mimic favourable dormancy conditions. Pyter et al. (2010) stored processed rhizomes in the dark at 4°C and 75% relative humidity to maintain viability of germplasm and lower the risk of desiccation. Total MC remained consistent at around 65% when Miscanthus rhizomes were harvested in December/January and stored at 2 or -2°C, and this was not severely affected by time in storage or storage temperature (Davies et al. 2011). When M. x giganteus rhizomes were harvested in autumn from four European locations (Sweden, England, Germany, and Denmark), rhizome MC ranged from 75±0.8 to 82±1.0% (Clifton-Brown and Lewandowski 2000). After storage of rhizomes for up to four months in CS, no significant loss in viability was observed after planting (Pyter et al. 2010).

Xue et al. (2015) documented that rhizome size significantly influenced plant survival (75.5%), stem diameter (0.96 cm), shoot number (81), and biomass yield (2242.2 g m-2) for Miscanthus. Statistically, this factor was more critical for determining plant vigor than other accounted treatment factors, including genotype and harvest day. Over the duration of one growing season (establishment phase), rhizomes averaging less than 6 cm yielded lower biomass development than their larger counterparts (Xue et al. 2015).

A later spring harvest (April 20th, 2011) resulted in higher viability (82.5%) and shoot diameter than earlier dates (March 10th and 30th, 2011) (Xue et al. 2015). Though this information is useful for determining favourable harvest periods, storage potential of germplasm harvested on the same day was not addressed. Davies et al. (2011) undertook this question and found that DWs and culm number significantly increased with increasing time in CS. Over the span of the study, differences in culm emergence and rhizome viability remained minimal, providing preliminary evidence that conservation of this germplasm may be beneficial to farmers and the biofuel feedstock marketplace.

2.2.2 Storage and carbohydrate dynamics

Carbohydrates formed through photosynthesis may be utilized for a number of processes, including conversion to cellulose or hemicellulose to fulfill structural roles, preserved as soluble sugars, or converted to starch (functioning as transitory storage) (Purdy et al. 2015). Carbohydrate reserves and dynamics are critical for dormancy, growth, and respiration in plants (White 1973). More specifically, reserve carbohydrates are utilized for winter survival, re-emergence in the spring, replacement of aerial tissues after removal from predators, and compensation for insufficient photosynthetic activity (Caldwell et al. 1981; Clifton-Brown and Lewandowski 2000; Purdy et al. 2013, 2015). Major components of reserve carbohydrates ( also known as non- structural carbohydrates or NSCs) include reducing sugars (glucose and fructose), nonreducing sugars (sucrose), starches, and fructosans (McCarty 1938). In contrast to these compounds, hemicellulose and cellulose are classified as structural carbohydrates and are not easily hydrolyzed sugar reserves (Sullivan and Sprague 1943). Early research has shown that grasses of temperate origin mainly store sucrose and fructosans, while those of subtropical origin utilize sucrose and starch for reserves (Smith 1968). Though nutrient reserves also include proteins and nitrogenous compounds, carbohydrates seem to take preference in regard to respiration (White 1973). Since

10 hybrid M. x giganteus is the result of progenitors from both temperate and tropical lineage (Cruz and Dierig 2015), it is unclear as to which NSC(s) are most influential for determining rhizome storage properties.

It has been demonstrated that Miscanthus exhibits diurnal carbohydrate patterns (Purdy et al. 2013) and that carbohydrate remobilization from the stem to the rhizome is largely dependent on average daily low temperatures, such as those experienced in autumn (Purdy et al. 2015). The link between carbohydrate dynamics and rhizome storage in Miscanthus has only been previously investigated by Davies et al. (2011), while other studies have focused on assessing how carbohydrates associate with overwintering in the field (Clifton-Brown and Lewandowski 2000; Purdy et al. 2013, 2015; Withers 2015). Seasonal dynamics of other nutrients and dry matter aggregation in Miscanthus rhizomes have been studied by Christian and Haase (2001) and Beale and Long (1997), and fluctuations in these components are expected to follow similar patterns of accumulation and depletion as starch and NSCs during induction and release of dormancy. Rhizomes from M. x giganteus stands in Europe reached seasonal minimal nutrient and dry matter contents between May to July; thereafter, these tissues began to acquire nutrients, well before early flowering had been known to occur (mid- to late August in Ontario) (Beale and Long 1997; Christian and Haase 2001; Rosser 2012). Beale and Long (1997) also observed that aerial parts of the plant began to show decreases in nutrients in mid- to late summer in the UK, while these levels in the rhizome increase during this period in preparation for dormancy.

Like expected in Miscanthus, starch and sucrose gradually decreased – and reducing sugars increased – from the beginning of the growing season until June, followed by heightened reducing sugar levels in the leaves until September in greenhouse-cultivated blueberry (Vaccinium angustifolium Ait.) rhizome tissue (Townsend et al. 1968). Generally, blueberry rhizomes contained more starch than aerial tissues, while leaves accumulated more sucrose and reducing sugars (Kaur et al. 2012).

During six weeks of storage at either 2 or -2°C, Davies et al. (2011) observed that starch content in M. x giganteus rhizomes harvested in early spring did not diminish over time or between storage temperatures. Conversely, soluble sugar concentrations increased with time spent in storage, especially at -2°C. The lack of correlation between starch and soluble sugar levels in this study

11 indicated that increases in soluble sugars may have been derived from the hydrolysis of structural carbohydrates such as hemicelluloses (Davies et al. 2011).

Purdy et al. (2015) harvested Miscanthus rhizomes periodically from the field over the course of winter from several UK locations. Significant increases in Miscanthus rhizome starch was observed between November and January, except in one genotype who was cultivated in a location which experienced relatively warmer soil temperatures over winter months. Higher soil temperatures during the expectant dormant period presumably increased respiration rates, which caused a decline in carbohydrate reserves (Purdy et al. 2015). The researchers’ hypothesized that harvesting Miscanthus rhizomes after dormancy has been induced in November would be adequate for ensuring that the germplasm contained enough carbohydrate reserves for viability.

2.3 Micropropagation 2.3.1 Fundamentals of micropropagation

Micropropagation is the process of removing a small amount of somatic plant tissue from the donor plant, performing surface sterilization, and inoculating the tissue into a controlled growth environment (Lewandowski 1997). Growing plants in vitro has many advantages over in vivo systems, including production of certified disease-free germplasm for national and international exchange or establishing nurseries, rapid multiplication of elite genotypes that exhibit genetic integrity, performance of species manipulation and alternative breeding technologies (Mehrotra et al. 2007), and the study of plant processes without interference of unwanted artefacts (Vreugdenhil et al. 1998).

2.3.2 Miscanthus micropropagation

Both direct and indirect methods of micropropagation for Miscanthus have been established in the literature. The direct method of micropropagation – also categorized as in vitro tillering – involves bud development from axillary nodes and apical meristems (Lewandowski 1997) and is advantageous for breeding interests by conservation of genetic uniformity (Rambaud et al. 2013). Indirect micropropagation methods have mainly focused on somatic embryogenesis through immature inflorescence cultures (Lewandowski 1997; Petersen 1997; Kim et al. 2010; Głowacka et al. 2010; Kim et al. 2012; Gubišová et al. 2013; Perera et al. 2015; Ślusarkiewicz-Jarzina et al. 2017), which can be an asset for creating ease in genetic transformation for improvement in a

12 species (Rambaud et al. 2013). Though other explants have been used as source material (eg. leaf and root sections), contamination, tissue necrosis, tillering inability, and species/genotype specificity have prevented these explants from becoming more utilized in commercial applications (Holme and Petersen 1996; Petersen 1997).

Steps for indirect micropropagation include: the induction of embryogenic callus; plantlet regeneration; in vitro tillering; and in vitro or ex vitro rooting (Kim et al. 2012). Direct micropropagation differs in its sequence of events by excluding induction of embryogenic callus and plantlet regeneration, and instead must undergo “shoot induction” before the tillering phase. In some previous literature, regenerated/induced shoot clusters are exposed to rooting media before they undergo tillering (Kim et al. 2012).

2.3.3 Miscanthus micropropagation challenges

The major challenges surrounding successful Miscanthus plantlet propagation is browning/necrosis of explants (Gubišová et al. 2013) and hindered regeneration potential (H. S. Kim et al. 2010). To prevent Miscanthus explants from browning, pre-soaking the tissue in antioxidants (150 mg l-1 citric acid and/or ascorbic acid) and amending media with 50 mg l-1 L- cysteine–HCl, 50 mg l-1 polyvinylpyrrolidone 10, or 3 g l-1 active charcoal has previously been investigated (Lewandowski 1997; Gubišová et al. 2013). Of these compounds, the addition of 50 mg l-1 L-cysteine into media (Lewandowski 1997) was the only tested agent that could reduce browning significantly. Regarding regeneration potential, it has been demonstrated that M. x giganteus immature inflorescence cultures can only develop shoot-like structures up to four months of culture on semi-solid media (Kim et al. 2010). Though regeneration could occur for up to a year when calli were maintained in suspension culture (Holme et al. 1997; Kim et al. 2010), this system presents different challenges, including more frequent media replacement and cell subcultures (every 7-14 days compared to every 21-28 days for adherent cultures), increased risk of contamination, and the requirement for continuous agitation.

The findings made by Lewandowski (1997) suggests that browning and subsequent necrosis in Miscanthus nodes is primarily caused by an accumulation of oxidized phenols – characterized by a benzene ring bonded to a hydroxyl group – that generate quinones and water from the wounded explant as a defence response (Zaid 1987). When the plant tissue is wounded, toxic

13 compartmentalized phenols are released to the rest of the plant tissue and surrounding medium, visually resulting in colour changes and eventual cell death (Zaid 1987).

2.3.4 Phenylpropanoid biosynthetic pathway and in vitro culture

In response to this innate defence mechanism, L-cysteine functions as a nucleophilic agent toward quinones to yield a colourless adduct by “reducing o-quinones to their polyphenol precursors” (Ali et al. 2014). Miscanthus is favoured as a biofuel feedstock because of its high contents of cellulose and lignin in its cell walls (Le Ngoc Huyen et al. 2010), and lignin macrostructures contain abundant amounts of quinones, which can be high-yielding during lignin degradation (Zawadzki and Ragauskas 1999). The accumulation of phenols has been shown to induce browning in in vitro American elm (Ulmus americana L.), sugar maple (Acer saccharum), and Artemisia (Artemisia annua L.) (Jones et al. 2012; Jones and Saxena 2013), and have been shown to be critical in the regeneration potential of several species, including maize, wheat, sessile oak (Quercus petraea Liebl.), and alfalfa (Medicago sativa L.) (Lozovaya et al. 1996; Cvikrová et al. 1998, 1999, 2003; Hrubcová et al. 2000). Therefore, it is hypothesized that high quinone tissue concentrations could be reduced by regulating lignin and phenol biosynthesis and accumulation, subsequently improving tissue browning and regeneration.

Phenol biosynthesis has been significantly reduced in both in vitro American elm suspension cultures, and sugar maple and Artemisia callus cultures by supplementing 2-aminoindan-2- phosphonic acid (AIP) into the media (Jones et al. 2012; Jones and Saxena 2013). AIP works to competitively inhibit phenylalanine-ammonia lyase (PAL), which is the enzyme responsible for the first committed step of the phenylpropanoid biosynthetic pathway. PAL functions to convert the amino acid phenylalanine to ammonia and trans-cinnamic acid, which then leads to the creation of various phenols and quinones (MacDonald and D’Cunha 2007). In the Poaceae, resilience of the cell wall structure is primarily caused by compounds such as p-coumaric acid and ferulic acid (Hartley and Ford 1989), and hydroxycinnamic acids. Hydroxycinnamic acids produce hydroxycinnamoyl-CoAs that function to strengthen the cell wall and induce lignification for protection against wounding and possible microbial attack (Jones et al. 2012).

In Artemisia callus cultures, it was found through both visual inspection and analysis of total phenols that browning incidence decreased in a dose-dependent manner up to 10 μmol AIP (Jones and Saxena 2013). The addition of 100 μmol AIP into the medium resulted in nearly two-fold calli

14 fresh (1268.9±248.55 mg plate-1) and dry (116.3±22.78 mg plate-1) weights compared to the control, a significant reduction in browning on both media containing BA/1-naphthaleneacetic acid (NAA) (0.6±0.4) and 2,4-dichlorophenoxyacetic acid (2,4-D) (1.0±0.32), and reduced total tissue phenols (Jones and Saxena 2013).

The addition of 10 μmol AIP into callus induction and maintenance media has also been beneficial for increasing regeneration frequency of somatic embryos in sessile oak through the reduced accumulation of phenolic compounds, while high phenolic acid content was associated with non- converting embryos (Cvikrová et al. 1998). Additionally, supplementing 10 μM AIP into alfalfa culture mediums increased mitotic activity and incited cellular division (Hrubcová et al. 2000), both of which are pivotal functions for successful regeneration.

Therefore, inhibiting browning and improving embryogenic capacity for M. x giganteus calli on semi-solid MS media may be achieved by supplementing callus-inducing media with AIP. Addition of AIP in callus induction media may aid in reducing wounding response – characterized from the accumulation of oxidized phenols and quinones – from stressed immature inflorescence explants.

2.4 Microrhizomes 2.4.1 Introduction

Similar to M. x giganteus, sexual reproduction is limited/absent in both ginger and turmeric (Cousins and Adelberg 2008; Babu et al. 2016), and in situ rhizome propagation can be compromised by pathogen contamination, including rhizome rot (Pythium aphanidermatum), bacterial wilt (Ralstonia salanacearum), and soft rot (Meloidogyne spp.) (Sharma and Singh 1995; Covarelli, Beccari, and Tosi 2012; Babu et al. 2016). Rhizomes also multiply slowly (15-20 tons ha-1) (Balachandran et al. 1990), demonstrating the need for a disease-free propagation system.

In vitro MRs are small rhizomes produced from in vitro plants and have primarily been developed in liquid MS media (Cousins and Adelberg 2008) with increased sucrose levels (6-10%) and 6- benzylaminopurine (BAP) supplementation (maximum concentration around 35.2 µM) (Sharma and Singh 1995; Nayak 2000). Microtuber (MT) formation in potato requires similar culture conditions: for example, 8% sucrose supplemented with a cytokinin concentration of 44.0 µM (Sarkar et al. 2006). Unlike micropropagated plantlets, MRs have the capacity to be stored for long

15 periods of time under relatively cold conditions similar to field-grown rhizomes (Shirgurkar et al. 2001; Babu et al. 2016) and are considered “low maintenance” (and potentially less expensive) compared to other in vitro germplasm conservation methods. Harvested MRs can also be planted directly in field conditions without requiring an acclimation period (Shirgurkar et al. 2001; Babu et al. 2016).

In vitro MRs provide an attractive alternative propagation system for M. x giganteus because this technology has the potential for development of consistent traits across individual propagules regarding size, storage, and culm emergence timing. Though in vitro systems are generally more expensive than in vivo rhizome production methods, encapsulated MR sections possess the ability to obtain division rates of 1:30, while DRP averages at 1:10 (Xue et al. 2015).

2.4.2 Carbohydrate requirements

Exogenous sucrose application functions as a major energy source and provides a carbon skeleton for in vitro MRs and other belowground storage organs (Zheng et al. 2008). This fundamental principle is observed in in situ rhizomes and tubers, which store mainly starch and other NSCs for ample dormancy and overwintering (White 1973; Clifton-Brown and Lewandowski 2000; Davies et al. 2011; Purdy et al. 2015).

Rhizome physiology is relatively similar to other below-ground storage organs, such as tubers. At the subcellular level in potato tubers, starch is primarily located within amyloplasts, while soluble metabolites are stored in the vacuole (Wiltshire and Cobb 1996). During sprouting, starch is hydrolyzed and causes increases in the soluble sugar pool, allowing for remobilization of carbohydrates to buds, which subsequently results in culm emergence (Hariprakash and Nambisan 1996). Though the basic premise of carbohydrate action during tuberization and sprouting are conceptualized, the specific events mediating these operations have still not been fully elucidated (Hajirezaei et al. 2003).

Exogenous sucrose has regulatory effects on carbohydrate metabolism in in vitro plant systems. Potato tubers grown from axillary buds and incubated in 8% sucrose media demonstrated heightened activity of sucrose synthase (susy), fructokinase (FK), and ADP-glucose pyrophosphorylase (AGPase) after 10 days of culture, and these activities were greater in swollen than non-swollen sections of the explants (Vreugdenhil et al. 1998). The activities of these

16 enzymes were greatly hindered by the addition of 0.5 µM GA3 (which resulted in the production of stolon-like shoots) and reduction in sucrose (which caused normal shoot development). Moreover, the levels of NSCs (sucrose, glucose, and fructose) diminished with the onset of tuber formation. Reduced glucose levels could be explained by the increased activity of susy over invertases (each which produce uridine diphosphate glucose (UDP-glucose) and D-glucose, respectively), while the decline in fructose could be rationalized by heightened FK activity (Appeldoorn et al. 1999). In contrast to these reductions, rapid onset of starch accumulation coincided with tuberization. Heightened starch formation was hypothesized to be partly caused by the presence of AGPase; however, based on enzyme kinetic calculations, other starch synthesis pathways were expected to be involved in the tuber development process (eg. starch phosphorylase) (St-Pierre and Brisson 1995).

Quadratic saturation 310 D-optimal design was employed to model the most efficient concentrations of media components influencing ginger MR development in vitro (Zheng et al. 2008). The researchers’ discovered that at 8% sucrose, plant height significantly decreased compared to treatments with lower sucrose levels (4.7 and 3.4 cm at 2 and 8% sucrose, respectively); however, total rhizome weight per plant was highest at 8% sucrose (680 mg per plant). Similar trends have been observed with MR, leaf, and root relative DW (RDW) in turmeric incubated in MR induction media supplemented with 6% sucrose (compared to 2 and 4%) (Adelberg and Cousins 2007). Cousins and Adelberg (2008) calculated that to obtain a 1 g increase in tissue dry weight (DW) in turmeric, 1.8 g of sugar was required in the medium; excess sugar was hypothesized to be lost either through other plant processes or CO2 release.

Ginger MRs could be induced from in vitro rooted plantlets in MS basal medium supplemented with 90% sucrose and an incubation period of 90 to 100 days. These propagules could then be planted ex vitro in field conditions without the need for an acclimation stage (Babu et al. 2016). Sharma and Singh (1995) induced ginger MRs in liquid MS media (static culture) with 7.5% sucrose and 8 mg l-1 BAP for 50-60 days, which resulted in 18-20 rhizomes per flask. The average weight and bud number per rhizome in this study ranged from 73.8 to 459.3 mg and 1.4 to 2.9, respectively. After MRs had been harvested, 80% of the samples that were stored at room temperature for 2 months yielded roots and shoots when transferred to moist sand.

2.4.3 Carbohydrate and plant growth regulator interactions

The interplay of sucrose with various PGRs can have drastic effects on endogenous carbohydrate allocation and accumulation, especially in regards to MR formation, dormancy, and sprouting. Sucrose supplemented into MR induction media has been found to have significant effects on endogenous abscisic acid (ABA):gibberellic acid (GA3) ratios in ginger (Zheng et al. 2008). The dynamics of these PGRs aide in regulating the activities of amylase and invertase, which function to catalyze starch and sucrose hydrolysis, respectively (Zheng et al. 2004).

Zheng et al. (2008) observed that MR development was inhibited in ginger incubated on media supplemented with GA3, kinetin (KT), or NAA when sucrose concentration was established at 2%; however, these PGRs assisted MR induction when applied with 8% sucrose. Though GA3 exhibited the greatest influence over MR formation, significant interactions between GA3 and KT were also identified. It’s important to note that this observation does not apply to all systems; though GA3 has been shown to be advantageous for rhizome formation in some species/genotypes, it is documented to have inhibitory effects on potato tuberization, even when in the presence of other favourable tuber-inducing conditions (Okazawa 1967; Vreugdenhil et al. 1998).

Both GA3 and cytokinins encourage cell expansion in culture; however, high cytokinin concentrations may induce considerable branching of rhizomes, leading to nutrient competition and subsequent individual MR weight loss (Zheng et al. 2008). This observation was in agreement with Shirgurkar et al. (2001) who demonstrated that intermediate levels of BAP (between 4.4 and 35.2 µM) could lead to reduced MR development in turmeric. At 0 and 0.32 µM BAP, Cousins and Adelberg (2008) observed that turmeric MR formation was hindered, and that this was accompanied by abated root and shoot growth. Conversely, total plantlet and MR FW was significantly enhanced by BAP at 1 µM. In ginger cultures, 26.7 µM BAP supplemented into high- sucrose media demonstrated intermediate MR number per culture vessel (4) but exceptional MR yield (0.65 g) (Zheng et al. (2008).

Nayak (2000) discovered that MS liquid medium supplemented with 6% sucrose and 5 µM BAP produced the most mango ginger plantlets manifesting MR formation (73.3%), heaviest MRs (315.2 mg), maximum number of buds per rhizome (2.3), and second-highest germination percentage (63.6%) compared to all other treatments examined. In a different study, turmeric MR production was optimal in MS liquid medium with 8% sucrose and 5 µM BAP (Nayak 2002). Samples that produced MRs exhibited swelling at the base of their shoots (Nayak 2000)

18 accompanied by adventitious buds, and harvested MRs could be stored for up to six months in in vitro conditions (MS media with 4.5% sucrose, 0.1 mg l-1 BAP, and 0.8% bactoagar). Plantlets incubated in 3% sucrose failed to produce any rhizomes, even when paired with BAP and amended photoperiod (Nayak 2000).

Auxins (specifically indole-3-acetic acid (IAA)) were shown to have more influence over in vitro potato tuber growth than cytokinins (Borzenkova and Borovkova 2003); however, cytokinins are major PGRs involved in tuber initiation, possibly due to their innate ability to increase cell division rates (Zheng et al. 2008). Cytokinin importance was realized in potatoes, when increases of this class of PGRs correlated with 14C-photoassimilates and starch accumulation exclusively in immature tubers (Borzenkova et al. 1998). As tubers matured, cell division declined as cell expansion increased in the piths. This shift was associated with starch biosynthesis, and heightened IAA and ABA content. Moreover, cytokinin:ABA ratios were significantly higher in the cortex than in the pith, where cell division was dominant over cell expansion (Borzenkova and Borovkova 2003).

Despite results reported from Borzenkova and Borovkova (2003), involvement of auxins in the tuberization process have been especially inconclusive in the literature. Kolachevskaya et al. (2015) sought to elucidate this PGR’s role in in vitro potato plants through transformation of samples with a tryptophan-2-monooxygenase (IAM) gene (which functions to convert L- tryptophan to indole-3-acetamide in the IAA biosynthetic pathway) and a class I patatin (B33) promotor (employed for localized gene expression). The researchers’ found that the transgene was preferentially expressed in tubers (opposed to shoots) and showed sharp declines between weeks four and 12 of culture (cultivar-dependent). IAA levels correlated with transgene expression, and concentrations significantly exceeded those found in shoot tissue. Under short-day conditions and suboptimal and optimal sucrose concentrations (5 and 8%, respectively), tuberization began earlier (5-7 days) in the transgenic treatments than the control, and average MT weight typically exceeded control values as well. Interestingly, by the end of the culture period, the number of MTs between treatments were similar, suggesting endogenous auxin is important in the primary stages of tuber formation.

A member of the Poaceae that has been investigated for in vitro MR production – and that is more genetically similar to Miscanthus (Hodkinson et al. 2010) than ginger, turmeric, or potato – is

Bambusa bambos var. gigantea (Kapoor and Rao 2006). It has been demonstrated that shoot clusters were able to form MRs when cultured in MS basal media supplemented with 50 µM NAA, 2.5 µM BAP, and 5% sucrose (85% frequency), though the number of rhizomes per culture was relatively low (between two to six rhizomes). Increasing BAP in these cultures decreased MR and root formation; however, increased culm shoot growth. The addition of GA3 significantly affected the rate of MR induction (three instead of four to five weeks), MR node number, and emergence of tillers at both intermediate and terminal areas. In contrast, the addition of 0.1 µM 2,3,5- triiodobenzoic acid (TIBA) – a polar auxin transport inhibitor – into the media reduced the amount of cultures with MRs to 25%. This study was significant for demonstrating the importance of endogenous auxin availability on MR development in a member of the Poaceae, and these findings could be potentially more valuable than MR/MT induction studies focusing on species from plant families differing from M. x giganteus.

2.4.4 Synthetic seeds

M. x giganteus commercial propagation options can be broadened through synthetic seed development, in which tissues are encapsulated in a matrix that enables them to be planted in a manner analogous to true seed. The fundamental objectives for synthetic seed production are to protect the encapsulated tissue during handling and storage and allow for bud emergence when regrowth initiates (Redenbaugh et al. 1986). The concept of the synthetic seed was theorized by Murashige (1977) and first practiced with Daucus carota (carrot) somatic embryos encased in a polyoxyethylene coating (Kitto and Janick 1982). Using this foundation, synthetic seed technology has been applied to a wide variety of species, including alfalfa (Redenbaugh et al. 1986), ginger (Babu et al. 2016), mango ginger (Curcuma amada) (Nayak 2002), bamboo (Dendrocalamus strictus) (Mukunthakumar and Mathur 1992), and potato (Sharma et al. 2007; Ghanbarali et al. 2016). Adequate systems have been achieved using a range of explant sources (somatic embryos, protocorm-like bodies, calli, shoot tips, axillary buds, microshoots, nodal segments, and hairy roots) (Gantait et al. 2015) and different synthetic seed states (hydrated versus desiccated) (Nieves et al. 2001).

Due to their bipolar nature and ability to simultaneously develop shoot and root structures, somatic embryos are the preferred propagule for synthetic seed production (Redenbaugh et al. 1986; Gantait et al. 2015). Though theoretically somatic embryos illustrate sought-out properties for

20 synthetic seed construction, this explant source faces obstacles in some species, such as potato (Ghanbarali et al. 2016). Some of these challenges include: low rates of induction; potential somaclonal variation due to the use of 2,4-D in induction media (Xu et al. 2004); genotype- specificity (Rambaud et al. 2013); prolonged induction periods; incapability to synchronize embryo maturity (Fiegert et al. 2000); and whether the species/genotype models direct or indirect somatic embryogenesis and conversion potential (Kim et al. 2010; Kim et al. 2012). When these challenges are realized, other explant sources may be considered and present alternative advantages, such as easier handling, cost-efficient production, and genetic stability based on the assessment of amplified fragment length polymorphisms (AFLPs) (Ghanbarali et al. 2016).

The most representative models for synthetic seed production using explant in vitro MR sections with buds are in vivo ginger microshoots (Sundararaj et al. 2010), mango ginger microshoots with rhizome portion attached (Banerjee et al. 2012), and potato minituber sprouts (PMS) (Ghanbarali et al. 2016). All explants used were anywhere from 2-5 mm in size, and all utilized sodium alginate

(3-4%) and CaCl2 (1-1.5%) for encapsulation (akin to seed endosperm) and matrix-formation (corresponding to seed coat), respectively.

Babu et al. (2016) used in vitro shoots and embryoids (with apical dome intact) 1-3 mm in size for ginger synthetic seed development. Meristematic tissue was encapsulated in MS basal medium with 4% sodium alginate, 2 M glycerol, and 0.4 M sucrose. Solidification of the beads occurred with a 20-minute incubation in 0.1 M CaCl2 solution. Thereafter, beads could be stored between six to eight months, with regrowth occurring in vitro in MS medium supplemented with 1 mg l-1 BAP and 0.5 mg l-1 NAA. Similar propagation systems have also been applied to mango ginger (Nayak 2002).

Synthetic seeds have been revealed to exhibit excellent durations of storage (upwards of six months), though this is dependent on gelling matrix, temperature, storage substrate (if any), and whether samples are hydrated or desiccated (Nieves et al. 2001; Sundararaj et al. 2010; Banerjee et al. 2012; Babu et al. 2016; Ghanbarali et al. 2016). Banerjee et al. (2012) presented that mango ginger microshoots could be stored for six months when Bavistin® was included in a porous luffa environment or in the encapsulation solution, and Sundararaj et al. (2010) revealed that encapsulated ginger microshoots had the capacity to be stored up to 12 weeks (13% conversion) when first dehydrated in a 0.25 M sucrose solution. Synthetic seed PMS samples subjected to CS

21 at 4°C could retain germination and plantlet conversion capacity, even when stored without a storage matrix (Ghanbarali et al. 2016), and Salvia officinalis L. (sage) shoot tips enclosed in 3% -1 alginate, 1.5% sucrose, and 0.25 mg l GA3 were able to produce plantlets with (33.3%) and without (63%) roots after 24 weeks of 4°C storage (Grzegorczyk and Wysokińska 2011).

To induce desiccation and dormancy by functioning to regulate the accumulation of storage carbohydrates, lipids, proteins, and cryoprotectants (eg. proline and polyamine), encapsulated sugarcane somatic embryos (encased in sodium alginate with MS basal medium, 90 g l-1 sucrose, -1 -1 -1 -1 2 mg l GA3, 0.5 mg l IAA, 6 mg l arginine, and 8 mg l glutamic acid) were treated with 3.8 µM ABA (Nieves et al. 2001). Nieves et al. (2001) treated a subsection of samples with sucrose (0.5 M over 24 hours) and ABA (desiccated to 60% water content) which resulted in the greatest survival rate (73%) compared to control treatments with and without dehydration (53 and 27%, respectively).

Although there are numerous factors that can significantly influence the production, storage duration, and subsequent regrowth of synthetic seeds, the current study focused on preliminary data regarding the development of M. x giganteus ‘M161’ synthetic seeds using immature in vitro MR bud explants. This was achieved through investigation of the innate capacity of sectioned MR buds to maintain dormancy and retain regrowth potential when encapsulated in 3% sodium alginate and 1% CaCl2 solution, void of additional nutrients, PGRs, or fungicides. To further represent Miscanthus in situ seed characteristics (of fertile species), synthetic seeds were also desiccated in a flow bench overnight (analogous to an innate quiescent state) (Christian 2012) before storage or growth assessment. Similar technology has already been made available to farmers (Crop Expansion Encapsulation & Drilling System™, abbreviated as CEEDS™) through New Energy Farms™; however, manufacturing details of this proprietary product have not been made public. According to New Energy Farms™, large-scale production of Miscanthus synthetic seeds will reduce planting weights up to 75%, reduce transportation fees of germplasm by 80%, allow maximal window of opportunity for planting, and create ease in planting compared to conventional DRP (New Energy Farms 2015).

1.2 Hypotheses and objectives Rhizome storage and dormancy

It was hypothesized that plant growth and vigor from M. x giganteus rhizomes would be dependent on harvest season, storage temperature, storage duration, and physiology (genotype, NSC content, and total starch content). The objectives of this study were to: a) compare various growth parameters of five M. x giganteus genotypes grown in southwestern Ontario after predetermined times in CS; b) quantitatively measure NSC and total starch content and surmise if these dynamics influenced rhizome dormancy; and c) investigate if harvest season had a significant effect on storage and growth factors.

Prolonged germplasm conservation

It was hypothesized that application of AIP would reduce total soluble phenolic content in callus, and subsequently improve regeneration potential and timeframe for M. x giganteus ‘M161’ calli cultured on semi-solid media. To test this hypothesis, Miscanthus calli derived from immature inflorescence explants were cultured on callus maintenance media supplemented with various concentrations of AIP over a period of 12 months. Both qualitative and quantitative measurements related to callus type, regeneration efficiency, and soluble phenolic content over time were investigated.

In vitro microrhizome development

It was hypothesized that sucrose, cytokinin, and auxin supplemented into in vitro Murashige and Skoog (MS) media would induce rhizome development in M. x giganteus ‘M161’, and that these propagules could be used as an alternative propagation system. The objectives of this study were to: a) investigate MR formation and overall plantlet vigor with a variety of MS liquid mediums supplemented with different concentrations of sucrose, BAP, and NAA; b) test MRs for their capacity to be used as propagules in in vitro and ex vitro conditions before and after CS treatment; and c) observe CS and subsequent growth parameters after being sectioned and encapsulated into alginate beads.

3: Storage and propagation of rhizomes from five Miscanthus x giganteus genotypes grown in southwestern Ontario Abstract

Lignocellulosic Miscanthus x giganteus Anderss. has been considered a promising second- generation biomass feedstock for biofuel and value-added product development. Its triploid nature renders it sterile, requiring that propagation must be achieved through vegetative means. Currently, the least expensive and most common method for large-scale cultivation is through rhizome cuttings ($0.15-0.54 per propagule) planted at densities ranging from 10,000 to 40,000 plants ha- 1, resulting in exorbitant establishment costs. To address this challenge, All Weather Farms Inc. recently developed a custom rhizome harvester that will automate much of the process and significantly reduce the cost per rhizome. However, the optimal harvest time and storage conditions for Miscanthus rhizomes are not well documented and need to be determined for commercial implementation. The objective of the current study was to evaluate the storage properties of several genotypes suitable for Ontario conditions harvested at different times to help guide commercial propagation. We hypothesized that plant growth and vigor from M. x giganteus rhizomes would be dependent on agronomic practices (harvest season, storage temperature, and storage duration) and plant physiology (genotype, NSC content, and total starch content). To test this, five M. x giganteus genotypes (‘M161’, ‘M114’, ‘M116’, ‘UK’, and ‘BC’) were evaluated for long-term cold storage (0/1°C and 3°C) of rhizome cuttings harvested from two locations in southwestern Ontario (Elora and Port Ryerse, ON) in spring 2015/2016 and autumn 2016/2017. Rhizomes harvested in mid-autumn and stored for 4-5 months at either 0/1 or 3°C yielded greenhouse-grown plants of similar performance to those harvested in the spring. Furthermore, ‘M114’ and ‘M116’ consistently outperformed the other genotypes in terms of viability, culm emergence, and tiller height. These findings may help in reducing initial establishment costs for M. x giganteus and allow for a greater supply of biomass for end-product development.

3.1 Introduction

Miscanthus is a rapid-growing, perennial tallgrass composed of high concentrations of cell wall phenolic acids and cellulose, making it a desirable second-generation biofuel and value-added product resource (Le Ngoc Huyen et al. 2010; Engbers and Deen 2013). Allotriploid Miscanthus x giganteus Anderss. (M. x giganteus) has been regarded as the most promising candidate species

24 in the genus for commercial non-food biofeedstock cultivation, due to its high annual biomass yields (Christian and Haase 2001; Dohleman and Long 2009), exceptional water- and nitrogen- use efficiencies (Beale et al. 1999; Zub and Brancourt-Hulmel 2010), ability to thrive in temperate climates (Clifton-Brown and Lewandowski 2000; Clifton-Brown et al. 2001; Rosser 2012; Friesen et al. 2014; Cruz and Dierig 2015; Peixoto 2015; Withers 2015), low risk of contracting disease which may severely affect its productivity during the growing season (Covarelli et al. 2012), and innate sterility securing reassurance that this species will not become invasive in foreign environments (Lewandowski 1998).

The principle apprehension for farmers deciding to cultivate Miscanthus is largely based on establishment costs, especially when producers cannot harvest and begin making revenue until the second or third year after planting (Atkinson 2009; Xue et al. 2015; Mangold et al. 2017). The majority of these expenditures are rooted in initial germplasm costs, which is also reflected in transportation and labour costs (New Energy Farms 2015). Considering that M. x giganteus is a sterile triploid species, nurseries dedicated to producing viable seeds that exhibit genetic integrity to elite genotypes have yet to be developed (Lewandowski et al. 2016; Mangold et al. 2017). The prime methods of establishment for this species are DRP (3,375 € ha-1), rhizome-derived plants (4,400.80 € ha-1), and micropropagation (6,320 € ha-1) (Xue et al. 2015).

Though DRP is currently the most cost-effective method of M. x giganteus cultivation, this approach encounters copious challenges. It is recommended that M. x giganteus is planted at densities of approximately 12,000-16,000 and 40,000 rhizomes ha-1 for biomass and germplasm production, respectively. With prices of individual propagules typically ranging from $0.15-0.54 (Atkinson 2009), the requirement to lower M. x giganteus germplasm in the marketplace is essential for persuading farmers to cultivate this species.

Conventionally, Miscanthus rhizomes are harvested in early spring after the ground has thawed (Pyter et al. 2010), usually leaving farmers with limited time to process and either re-plant or sell their stock before embarkment of the growing season. After planting, early spring culm emergence from rhizomes in the field is essential for both maximal biomass yields and sufficient rhizome filling for overwintering (Beale and Long 1997; Lewandowski 1998; Xue et al. 2015).

Successful rhizome field establishment can reach as high at 90% when favourable conditions are available; however, these values can be cut by almost half for large-scale operations (Xue et al. 2015). Reasons for this reduction include limited specialized rhizome harvesting machinery, inconsistencies in rhizome quality (encompassing but not limited to: disparate rhizome ages and sizes within stands; tissue damage caused by pest attack, frost, and dehydration; and mechanical damage during harvest and processing), seasonal conditions during harvest periods (eg. low air temperature and soil moisture content (MC)), agronomic practices after planting (eg. application and timing of liquid fertilizers, pesticides, etc.), and conservation of germplasm which will guarantee optimal viability and growth potential (Christian and Haase 2001; Atkinson 2009; Christian et al. 2009; Pyter et al. 2010; Zub and Brancourt-Hulmel 2010; Davies et al. 2011; Xue et al. 2015).

To allow for adequate rhizome harvesting and processing time before the growing season begins in the spring, harvests in autumn before the soil freezes may be an ideal alternative. This option is also intriguing for granting farmers greater flexibility for germplasm trade to locations that have different optimal planting periods, which may help in reducing propagule costs long-term. For this system to be presented as a reasonable contingency for producers, storage conditions must be acceptable enough to warrant growth from these propagules comparable to those harvested in the spring.

Based on LT artificial freezing tests, the average lethal temperature for M. x giganteus rhizomes is -3.4°C (Clifton-Brown and Lewandowski 2000), while retardation of growth without inducing full dormancy has been found to be between 7 to 9°C (Farage et al. 2006; Pyter et al. 2010). Pyter et al. (2010) demonstrated that M. x giganteus rhizomes stored for up to four months at 4°C maintained dormancy and viability when harvest commenced in late autumn. Though this study provided some insight into the feasibility of autumn harvest, this experiment was conducted in Illinois, USA and is not representative of southern Ontario climate. Moreover, only one storage temperature (4°C) and genotype (‘Illinois’) was investigated, and the values obtained from autumn harvests were not compared to those in the spring.

Aerial tissue growth from rhizomes planted in the establishment year have been demonstrated to be representative of biomass yields for stands in their third year of cultivation (Clifton-Brown et al. 2002; Rosser 2012). Rosser (2012) determined that culm height and basal stem circumference

26 across 20 Miscanthus genotypes were associated with winter survival, indicating that these criteria may be applied for measuring plant vigor. In a three-year experimental period using M. sinensis, trait-marker associations determined that characteristics including plant height, leaf length and width, culm circumference, internode number, and tiller number correlated to biomass yield (Nie et al. 2016). Furthermore, Jeżowski et al. (2011) observed significant ‘G’ x ‘cultivation year’ interactions for stem and clump diameter, and ‘G’ x ‘cultivation year’ x ‘location’ interactions for plant height, tillering, and biomass yield for M. x giganteus during the initial three years of growth.

Reserve carbohydrates are critical for dormancy and winter survival, re-emergence in the spring, replacement of aerial tissues after removal from predators, and compensation for insufficient photosynthetic activity in perennial species (Caldwell et al. 1981; Clifton-Brown and Lewandowski 2000; Purdy et al. 2013, 2015). Major components of reserve carbohydrates include reducing sugars (glucose and fructose), nonreducing sugars (sucrose), starches, and fructosans (McCarty 1938). Early research has shown that grasses of temperate origin mainly store sucrose and fructosans, while those of subtropical origin utilize sucrose and starch for reserves (Smith 1968). Since hybrid M. x giganteus is the result of progenitors from both temperate and tropical lineage (Cruz and Dierig 2015), it is unclear as to which NSCs are most influential for determining rhizome storage properties.

Over the course of six weeks in storage at either -2 or 2°C, Davies et al. (2011) observed that starch reserves from M. x giganteus rhizomes harvested in early spring did not change over time, while soluble sugars increased with storage duration. Purdy et al. (2015) concluded that Miscanthus rhizomes continued to accumulate starch between November and January in the field, well after stem senescence had occurred, indicating that rhizome harvest after dormancy had been induced would be acceptable for viability. However, whether the trends found in these studies materialize in long-term storage conditions, and whether these trends greatly differ after both autumn and spring harvests, have yet to be substantiated.

Based on previous literature, it is hypothesized that plant growth and vigor from M. x giganteus rhizomes will be dependent on CS temperature, CS duration, harvest season, genotype, reducing and nonreducing carbohydrate content, and total starch content. The objectives of this study were to: a) compare non-destructive aerial growth traits of five M. x giganteus genotypes grown in southwestern Ontario after predetermined times in CS; b) quantitatively measure rhizome NSC

27 and total starch content and surmise if these dynamics influenced rhizome dormancy, viability, and culm emergence rates; and c) investigate if harvest season had a significant effect on storage and growth factors. Verifying the factors that influence M. x giganteus rhizome storage and re- emergence can benefit farmers producing germplasm for sale, subsequently lowering the costs of propagules and encouraging more widespread cultivation of this biomass crop.

3.2 Materials and methods 3.2.1 Harvest – Port Ryerse Five M. x giganteus genotypes were selected for investigation of rhizome CS capacity. Origins and physiological differences of each of the genotypes are described in Tables 3.1 and 3.2, respectively. The first harvest was executed at All Weather Farming Incorporated at Port Ryerse, ON (42°47′ N 80°12′ W), and surrounding sites (within a 10-km radius) in spring 2015 (early May), and only included ‘M161’ (‘Illinois’). Rhizomes harvested in autumn 2015 (mid- November), spring 2016 (late April), and autumn 2016 (mid-December) also included the genotypes ‘M161’, ‘UK’, and ‘BC’. All rhizomes originating from this site were harvested with a mechanical harvester either before the first suspected frost of the season (autumn) or after the ground had thawed adequately (spring), based on the farmer’s discretion (refer to Figure 3.1). The confirmed origins of genotypes ‘UK’ and ‘BC’ are unknown; however, are suspected to originate from locations in the United Kingdom and British Columbia, Canada, respectively (farmers reported divergent rhizome morphologies, rhizome growth patterns, and aerial tissue growth in field conditions for ‘UK’ and ‘BC’) (personal communication). Previous to harvests, ‘M161’ plants were cultivated in primarily lacustrine heavy clay soil, while ‘UK’ and ‘BC’ plants were established in sandy soil (Presant and Acton 1984). Average monthly air temperatures at this location and during these harvest years are depicted in Figure 3.2.

Table 3.1 Origins of five M. x giganteus genotypes used in the current study. Genotype ‘M161’; ‘M114’; ‘M116’; ‘UK’ ‘BC’ ‘Illinois’ ‘Amuri’ ‘Nagara’ Year 1988a 1997b 1996 Unknownc ~1970d released Fertility 3x=57; sterile 2x=38; fertile 3x=57; sterile Undetermined* Progenitors Natural, (♂, 2x) M. (♂, 2x) M. Undetermined* interspecific sinensis (from sinensis (from cross: M. sinensis European European (diploid) x M. collection) x collection) x sacchariflorus (♀, 2x) M. (♀, 4x) M. (tetraploid) sacchariflorus sacchariflorus (Linde-Larson ‘Robustus’ ‘Robustus’ 1993) (selected by (from Nagara Karl Foerster region in in Potsdam, Japan) Germany) (Withers 2015) (Deuter 2011; Withers 2015) *Information obtained by personal communications. aInitial use of ‘M161’ as a research standard at the University of Illinois at Urbana-Champaign. This genotype was originally acquired from the Chicago Botanical Gardens (Heaton et al. 2010; Friesen et al. 2014). b“Year released” dates for ‘M114’ and ‘M116’ indicate crossing year (Deuter 2011). cOriginally from Tinplant in Germany and transferred to the United Kingdom (personal communications). dOriginally from Agassiz, British Columbia, Canada (personal communications).

Table 3.2 General flowering and senescence times, stand ages, and other pertinent information for the five M. x giganteus genotypes used in the current study. Genotype ‘M161’; ‘M114’; ‘M116’; ‘UK’ ‘BC’ ‘Illinois’ ‘Amuri’ ‘Nagara’ Flowering Mid-September August Early-to-late Early autumn* time (Anderson et al. (Aurangzaib autumn (Deuter 2011) 2012) 2011) Senescence Late senescence Early senescence Late senescence Undetermined* timing - after killing (Aurangzaib - after killing frost 2012) frost (Aurangzaib (Aurangzaib 2012) 2012) Stand age (years) at initiation of 4, 2014* 7, 2008* 4, 2014* 6, 2012* study and year of planting Miscellaneous Consistently More cold- More cold- These genotypes were higher tolerant than tolerant than categorized as M. aboveground 'M161' (NRCS 'M161' (NRCS sacchariflorus var. biomass yields 2011) 2011); stem Voucher by ITS analysis than other lodging (40- in 2018 (A & L Canada genotypes 80% across one Laboratories Inc., (Engbers and field) (Engbers London, ON, Canada) Deen 2013) and Deen 2013)

*Information obtained by personal communications.

Figure 3.2 Average monthly air temperature (including monthly highs and lows) recorded at the Delhi Climate Station (located near Port Ryerse, ON) in M. x giganteus rhizome harvest years: a) 2015 and b) 2016.

3.2.2 Harvest – Elora

Two additional genotypes – ‘M114’ (‘Amuri’) and ‘M116’ (‘Nagara’) – were supplied from the Elora Research Station, Elora, ON (43°38’ N, 80°24’ W), managed by the University of Guelph. The soil type characterized in this area is a Guelph silt loam soil classified as “a Grey Brown

Luvisol (Soil Classification Working Group 1998) or Albic Luvisol (Food and Agriculture Organization of the United Nations 2006)” (Withers 2015). All rhizomes from this location originated from plants established in plots used for previous research projects and were harvested manually in autumn 2015 (early December), spring 2016 (late April), and autumn 2016 (mid- November). Similar to the Port Ryerse site, samples were harvested before the first suspected frost of the season (autumn) or after the ground had thawed adequately (spring), based on the discretion of the research manager. Average monthly air temperatures at this location and during these harvest years are depicted in Figure 3.3.

Figure 3.3 Average monthly air temperature (including monthly highs and lows) recorded at the Elora Research Station (Elora, ON) in M. x giganteus rhizome harvest years: a) 2015 and b) 2016.

3.2.3 Sample preparation

Unprocessed rhizomes spent no more than seven days stored at 4°C after the initial harvest date. Harvested rhizomes were thoroughly washed with non-potable raw water, manually processed into single branches/clumps (dependent on rhizome morphology) of near-uniform size in situ (approximately 30 to 60 mm in length, depending on quality, quantity, and growth pattern of material available at each harvest date), and had roots and at least one terminal end removed (both

34 preferable if rhizomes exhibited an indeterminate growth pattern). Processed rhizomes selected for experimentation thereafter were based on visual inspection of health, which included absence of necrosis, discolouration, and severe mechanical damage. Prepared samples were washed vigorously once more with tap water and allowed to air dry for approximately 30 minutes before being put into either CS or greenhouse conditions (refer to Figure 3.4).

M IM

C) D)

Figure 3.4 Stages of Miscanthus growth. A) Pairs of processed rhizomes after autumn 2015 harvest (from left to right: ‘Illinois’, ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’); B) newly-developed rhizomes with immature (IM) and mature (M) dormant buds after six weeks in greenhouse conditions; C) rhizome with healthy, thick, white roots after six weeks in greenhouse conditions; D) five-week old Miscanthus stem (measuring 700-800 mm in height) grown from a rhizome piece after cultivation in greenhouse conditions for six weeks. 3.2.4 Cold storage conditions After preparation of samples, rhizomes of each genotype (n=50) were placed into separate plastic containers (Bella Storage Solution™ 5.2 l locking lid, 13.25” x 9.25” x 4.75”) (subplots) and put

36 into one of four (two at either 0 or 1°C, and two at 3°C) chest freezers (main plots), arranged in a split-plot design.

Only ‘M161’ was tested for CS capacity during the spring 2015 harvest (0 and 3°C). CS at 3°C was investigated for all five genotypes at autumn 2015, spring 2016, and autumn 2016 harvests. Because of severe contamination developing throughout the duration of the final storage trial, growth data for ‘M161’ (autumn 2016 harvest) stored at 3°C was excluded from analysis (contamination severity represented in Figure 3.5). The number of genotypes investigated for rhizome CS at 1°C varied between harvest dates due to inconsistency of available material; ‘M161’ and ‘UK’ were tested at 1°C storage at autumn 2015 harvest, while ‘M161’, ‘UK, ‘BC’, and ‘M114’ were examined at 1°C storage at spring 2016 and autumn 2016 harvests. All treatments were replicated four times, and completely randomized within each chamber.

A) B)

Figure 3.5 Representation of severe Miscanthus ‘Illinois’ rhizome contamination after various times in storage at 3°C. A) Otherwise visually-healthy rhizome coated in soft, white, fibrous material; B) discoloured rhizome exhibiting soft-rot symptoms.

A subsample (n=8) of rhizomes were removed from each container every 21 days over seven time points (T) (spring 2015 harvest; experimental duration of 126 days) or every 28 days for six time points (autumn 2015 onward; experimental duration of approximately 140 days for each trial). Half of these samples were used for greenhouse growth trials, while the remaining samples were destined for laboratory analysis.

3.2.5 Greenhouse conditions

All growth experiments were conducted in the Edmund C. Bovey greenhouses at the University of Guelph. Greenhouse compartments were split longitudinally, with one half receiving supplementary lighting from Philips® 400W (80 μmol m2 s−1) LED18 High Pressure Sodium

Bulbs (Philips Alto, USA) when ambient light was below 300 W m-2 for at least 20 minutes. Once supplementary lights were triggered, they remained on for at least one hour. The temperature was kept constant at 21°C, and photoperiod was set to a 16-hour day (06:00 to 22:00)/8-hour night (22:00 to 06:00) format.

3.2.6 Planting conditions

After preparing samples for CS, 16 rhizomes of each genotype were individually potted in Sunshine® Mix LA#4 (Sun-Gro Horticulture, Bellevue, WA, USA) at a depth of approximately 50 mm and placed in the greenhouse as a control (T=0). Plastic pots measuring 6 inches in diameter were used for all greenhouse growth experiments (The HC Companies, Inc., USA). Before planting at any planting time point, samples were measured for exact length (mm), FW (g), and node number. Due to the determinate growth patterns of ‘M114’ rhizomes, sample length was based off of visual identification of the primary rhizome branch.

Due to material availability, T=0 planting was conducted for ‘M161’ at spring 2015 and autumn 2015 harvests; T=0 planting was executed for all five genotypes at spring 2016 and autumn 2016 harvests. For all planting time points, samples were fertilized only on day one of planting with 20- 8-20 (N-P-K) fertilizer solution (Plant Products®, Leamington, Canada; 250 ppm N or 1.25 g l-1; pH adjusted to 6 with phosphoric acid). Watering with non-potable water commenced every two to three days as needed.

3.2.7 Growth parameters

Non-destructive measurements of plants included: culm emergence speed (days); rhizome viability (%); tiller number; tiller height (mm); leaf number; stem node number; stem diameter (mm); greenness based on leaf chlorophyll content (mg m-2); lodging (if any); and tiller growth pattern (extravaginal vs. intravaginal) (refer to Figure 3.6). Spring 2015 harvest growth parameters only included culm emergence speed, rhizome viability, tiller number, and tiller height.

Culm emergence speed (sprout measuring at least 2.5 cm) was recorded on a daily basis, while all other parameters were measured at six weeks after planting. Rhizome viability was calculated as the percentage of samples that resulted in at least one tiller emerging within six weeks after planting. The tallest tiller per rhizome sample was measured for height, leaf number, stem node number, stem diameter, and chlorophyll content. After straightening the shoot to be upright, height

38 was measured from the base of the stem to the tip of the culm. Leaves were characterized by tissue (longer than 5 mm) extending out from the stem at a defined ligule/collar. Stem diameter was measured approximately 2.5 cm above the base. Ten chlorophyll readings were taken from two randomly selected, mature, visually-healthy leaves, belonging to the tallest tiller per sample, using an Opti-Sciences™ CCM-300 chlorophyll fluorescence metre (Opti-Sciences Inc., Hudson, NH, USA).

A) B)

C) D)

Figure 3.6 Notable Miscanthus aerial growth traits cultivated from rhizome pieces and incubated in greenhouse conditions for six weeks. (A) Extravaginal and (B) intravaginal tiller growth patterns in 6-inch pots; C) exposed stem node; D) expanded leaf blade with formed ligule adjunct to stem.

3.2.8 Total moisture content

The remaining subsamples removed from CS at each planting time point were preserved for total MC and laboratory analysis. Briefly, the FW of four subsamples from each container were

39 recorded immediately after removal. Samples were then enclosed in aluminum foil and frozen in liquid nitrogen. Samples were kept at -80°C before lyophilization; thereafter, samples were stored at -20°C until use. DW was recorded after lyophilization, and total MC (%) was calculated: ((FW- DW)/FW) x 100.

3.2.9 Sample Preparation

Dried rhizome samples were ground into a fine powder for metabolite analysis. Briefly, samples were cut into relatively small pieces (approximately 10 mm in length) with pruners before being ground in an IKA® A11 Basic Analytical Mill (IKA® Works Inc., Wilmington, USA). Liquid nitrogen was added to the grinding container, and most was allowed to evaporate before grinding commenced. Usually, three to four pulses (duration of 30 seconds per pulse) yielded sufficient quality material for analysis.

3.2.10 Total starch quantification

Total starch content [%, w/w (DW basis)] was determined using a stepwise enzymatic assay kit (Megazyme International, Catalogue No: K-TSTA, Megazyme International Ireland Ltd., Bray, Ireland Ltd.) with slight modifications. Aqueous ethanol (80 % v/v) was added to samples before analysis to eliminate excess levels of D-glucose and maltodextrins. Briefly, starch was hydrolyzed into maltodextrins after the addition of thermostable α-amylase. D-glucose was then formed after maltodextrins were hydrolyzed by amyloglucosidase (AMG). The subsequent addition of glucose oxidase oxidized D-glucose to produce D-gluconate and one mole of hydrogen peroxide. It was with further reactions of hydrogen peroxide with peroxidase that produced a quinoneimine dye. The absorbance values produced by this final reaction was read colorimetrically at 510 nm. It is important to note that this “assay is specific for α-glucans (including starch, glycogen, phytoglucogen and non-resistant maltodextrins)” (Megazyme International, Catalogue No: K- TSTA, Megazyme International Ireland Ltd., Bray, Ireland Ltd.); however, the main advantage of this assay is that both thermostable α-amylase and AMG function at pH 5.0, which simplifies the procedure and reduces the risk of forming 4-α-glucopyranosyl-D-fructose (Crabb and Shetty 2003) which is resistant to hydrolysis of these enzymes.

Concisely, 80 % (v/v) aqueous ethanol was added to approximately 100 mg (exact DWs recorded) of each CS treatment replicate from all harvest seasons (spring 2015/2016 and autumn 2015/2016)

40 before initiating enzyme reactions, as outlined in the manufacturer’s instructions. Exactly 3 ml of Glucose Determination Reagent (GOPOD reagent) – containing p-hydroxybenzoic acid and sodium azide buffer; and glucose oxidase, peroxidase, and 4-aminoantipyrine – was added to diluted 0.1 ml sample extracts, standards, or sample blanks. Each sample extract, standard, and sample blank were added in duplicate to separate polypropylene tubes before addition of the GOPOD reagent. Thereafter, 0.2 ml of each reaction was added in duplicate to each well of a 96- well flat bottom microplate (Corning, Corning, USA) in a completely randomized design. D- glucose standard solution accompanied the assay kit, and a standard curve was constructed for each plate used (1000, 500, 250, and 125 µg ml-1). All sample and standard readings were corrected with blanks. Absorbance values were measured at 510 nm with a Synergy™ H1 microplate reader (Biotek, Winooski, USA).

3.2.11 Reducing and non-reducing sugar quantification

Sucrose, D-glucose, and D-fructose concentrations (mg g-1 DW) were quantified using a step-wise enzymatic assay kit (Megazyme International, Catalogue No: K-SURFG, Megazyme International Ireland Ltd., Bray, Ireland Ltd.) with slight modifications. Briefly, the concentration of D-glucose in a sample was calculated before and after sucrose hydrolysis with β-fructosidase (an invertase enzyme), and sucrose was determined by the difference of these values. D-glucose and D-fructose was then converted by hexokinase (HK) to glucose-6-phosphate (G-6-P) and fructose-6-phosphate (F-6-P), respectively. D-fructose concentration was then determined after the addition of phosphoglucose isomerase (PGI), which converted F-6-P to G-6-P, subsequently yielding gluconate-6-phosphate and nicotinamide adenine dinucleotide phosphate (NADPH). The amount of NADPH formed in each reaction was stoichiometric with the quantity of D-glucose and D- fructose present, and increased NADPH correlated to increased absorbance values. It is important to note that this assay is only semi-qualitative for sucrose content; β-fructosidase is not specific to sucrose, but also hydrolyzes low molecular weight fructans (McCleary and Blakeney 1999).

Briefly, approximately 100 mg (exact DWs recorded) of each CS treatment replicate from all harvest seasons (spring 2015/2016 and autumn 2015/2016) were diluted sufficiently with double- distilled water [dilution factor (F) of 1000] to yield absorbance values that fell within the range of the standards. Clear supernatant was then extracted and stored at -20°C until use. Succinctly, 10 μl aliquots of sample extracts, standards, or sample blanks were added to each well of a clear, 96-

41 well, flat bottom microplate (Corning, Corning, USA) in a completely random design. D-glucose plus D-fructose standard solution accompanied the assay kit, and a standard curve was constructed for each plate used (2, 1, 0.5, 0.25, 0.125, and 0.0625 mg ml-1). Sample extracts, standards, and blanks were replicated twice, and all sample and standard readings were corrected with blanks. Absorbance values were measured at 340 nm with a Synergy™ H1 microplate reader (Biotek, Winooski, USA). Due to the kit originally being prepared for a standard sized spectrophotometer, all reactions outlined by the manufacturer were reduced ten-fold. The calculated quantities of sucrose, D-glucose, and D-fructose for each sample were then pooled to yield total NSC concentrations (mg g-1 DW). The percentage of each sugar relative to the total NSC was also calculated for each sample (%).

3.2.12 Experimental design and statistical analysis

Greenhouse growth trials were subject to variance analysis using a mixed-model analysis of variance (ANOVA) using SAS® 9.4 software (SAS Institute Inc., Cary, NC, USA) arranged in a randomized complete block design (RCBD). All treatments consisted of four sampling units and were replicated four times.

CS trials were subject to variance analysis using a mixed-model analysis of variance (ANOVA) using SAS® 9.4 software (SAS Institute Inc., Cary, NC, USA) arranged in a split-plot design over two temperatures (either 0 or 1°C and 3°C). Freezers represented main plots (two levels) and replicate containers represented subplots (four levels).

Greenhouse growth trials and CS trials were repeated four times for ‘M161’ (spring 2015/16 and autumn 2015/16) and three times for the remaining genotypes (spring 2016 and autumn 2015/16). The type III main effects were ‘temperature’ (two levels), ‘storage time/duration/time point’ (seven levels), and their interaction (‘temperature x storage time/duration/time point’) (14 levels) for spring 2015, and ‘genotype’ (two, four, or five), ‘storage time/duration/time point’ (six levels), and ‘genotype x storage time/duration/time point’ (12, 24, or 30 levels) for autumn 2015, spring 2016, and autumn 2016 trials, with storage temperatures assessed individually for each trial.

Normal distribution of errors assumption was assessed using the Shapiro-Wilk statistic. According to Lund’s test of studentized residuals, no outliers were present throughout the data. In order to

42 normalize variances not following a Gaussian distribution, a variety of distribution and link functions were applied for each dependent variable.

Multiple-means comparisons were generated for all analyses using Tukey’s honest significant difference (HSD) test. Where linear or polynomial correlations were discernible over time, regression models and Pearson’s correlation coefficients were determined using PROC REG. Due to the abundance of data analyzed, only regression models with an R2 value of 0.50 or higher are reported (refer to Appendices 3.7-3.37). Type I and III error rates of α=0.05 were assigned for all analyses. The data presented in the results are from the original data scale.

3.3 Results 3.3.1 Rhizome viability

During the autumn 2015 trial, significant differences in rhizome viability were found between genotypes stored at 3°C, with ‘Amuri’ samples surviving at a rate of 87.1% over the experimental period. The lowest value was observed with ‘UK’ rhizomes, exhibiting a 52.2% survival rate. No differences in rhizome viability were noticed in samples stored at 1°C (data not shown).

Average rhizome viability for samples harvested in spring 2016 and stored at 3°C demonstrated genotype x storage duration interactions. In general, ‘Illinois’, ‘UK’, and ‘Amuri’ experienced heightened rhizome survival as time in storage persisted (37.7, 75.1, and 62.6% at time point 1 to 62.4, 78.4, and 93.8% at time point 5, respectively) (see Figure 3.7). In ‘BC’, survival dropped considerably between time points 1 (56.1%) and 5 (25.3%), and in ‘Nagara’, survival spiked at time point 3 (100%) before returning to levels similar to time point 1 (74.8%). Samples stored at 1°C during this period displayed genotype effects only, with ‘UK’ and ‘Amuri’ (75.1 and 86.2%, respectively) having higher values than ‘Illinois’ (45.6%) and ‘BC’ (42.5%).

Genotype effects were significant in samples harvested in autumn 2016, with the Elora-originating genotypes showing higher viability (60.9-80.2%) than the Port Ryerse rhizomes (9.2-20.6%). Genotype x storage duration significantly influenced this variable in samples stored at 1°C, with viability generally increasing from time point 1 to 3, then either declining slightly or leveling-off thereafter.

Figure 3.7 Average rhizome viability (%) of samples after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November/December 2015 and April 2016 and stored at 3°C for up to 140 days. Each bar is represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

3.3.2 Culm emergence

Rhizomes stored at 3°C after harvest in autumn 2015 demonstrated significant differences in culm emergence speeds between genotypes and over the storage period (refer to Figure 3.8). Both genotypes obtained from Elora, ON exhibited the fastest emergence speeds (11.8-13.3 days) while ‘UK’ and ‘BC’ from Port Ryerse, ON displayed the slowest (21.2-24.2 days). Collectively, emergence occurred more rapidly as storage time progressed after the autumn harvest. Culm emergence speed in the spring 2016 trial ranged from 5 to 18.2 days across all genotypes stored at 1 and 3°C, with the exception of ‘Illinois’ at time point 0 (31.3-31.4 days) (see Figures 3.9 and 3.10). In autumn 2016, samples stored at 3°C all showed similar trends of culm emergence speed, with rates of emergence occurring faster as storage time progressed (20.5-31.7 at time point 1 to 9-19.4 days by time point 5). These trends also resembled rhizomes stored at 1°C during this period.

b) Figure 3.8 Average emergence speed (days) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) pooled and b) individual genotypes. Individual data points and bars are represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

Figure 3.9 Average emergence speed (days) after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016 and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

3.3.3 Tiller height

Tiller height decreased in a linear fashion over storage duration for ‘Illinois’ in May 2015 from 740.5 to 351.4 mm (R2=0.63) (see Figure 3.11 and Appendix 3.12). The following autumn, significant genotype effects were realized, with ‘Nagara’ producing the tallest tillers averaged over

47 the trial period (986.7 mm) than the other genotypes (494.7-726.6 mm). No differences in tiller height were observed between 3°C ‘Amuri’ and ‘Nagara’ in autumn 2016, although these heights were shorter than the previous years’ (564.8-576.4 mm) (refer to Figure 3.12). In autumn 2016 rhizomes treated at 1°C, genotype x storage duration effects were significant, with the heights of all tested genotypes increasing from time points 1 to 5 (see Figure 3.13).

3.3.4 Tiller number

Tiller number differences between genotypes were only significant in autumn trials. ‘Amuri’ developed the most tillers (2.0) in 2015 from 3°C treatment compared to the remaining samples (ranging from 1.0-1.3). In 2016, genotype x storage duration significantly influenced these values (refer to Figure 3.14). Over the trial period for 3°C treatment, ‘Nagara’ displayed increasing tiller

50 number between the first (1.0) and final time point (2.7), whereas ‘Amuri’ remained consistent around 1 tiller throughout the trial. Increases were realized in ‘UK’ between time points 2 (0.5) to 4 (1.8), but sharply declined thereafter (0.29). For samples stored at 1°C, average tiller number peaked in both ‘UK’ (3.5) and ‘BC’ (3.3) at time points 4 and 5, respectively.

3.3.5 Leaf number

Leaf number per plant was significantly influenced by both genotype and storage duration in autumn 2015 (see Figure 3.15). For samples stored at 3°C, ‘Nagara’ produced the most leaves (4.2) while ‘BC’ produced the fewest (2.6). When genotypes were pooled and assessed over time, a gradual increase in leaf development was observed. These same trends in both genotype and storage duration factors were exhibited in samples harvested in autumn 2016 and stored at 3°C (see Figure 3.16). For samples stored at 1°C, genotype x storage duration factors were significant. Only genotype effects influenced leaf number in spring 2016; with samples stored at 3°C, Port Ryerse genotypes produced the most leaves (4.6-4.7), while Elora genotypes yielded the fewest (4.2). When genotypes were pooled and assessed over spring and autumn trials, leaf number remained fairly consistent over the storage duration after spring harvest, while leaf number increased as storage time progressed after harvest in autumn (refer to Figure 3.17).

b) Figure 3.15 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) individual and b) pooled genotypes. Individual bars and data points represent 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

Figure 3.16 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Measurements are categorized by a) individual and b) pooled genotypes. Individual bars and data points represent 80 and 64 ± SE samples, respectively. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

Figure 3.17 Average leaf number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April and November/December 2016 and stored at 3°C for up to 140 days. Each data point is represented by 64±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

3.3.6 Stem node number

Stem node number remained relatively consistent throughout spring trials but exhibited drastic changes after autumn harvests as storage times progressed (refer to Figures 3.18, 3.19, and 3.20). Only ‘Illinois’ demonstrated stem nodes early after harvest in autumn trials, while the remainder of genotypes only began developing nodes after time points 2 or 3.

Figure 3.18 Average stem node number (#) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from a) Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and b) Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively, and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples.

3.3.7 Stem diameter

‘Illinois’ and ‘Nagara’ displayed the largest stem diameters (5.4 mm) at 3°C storage in autumn 2015 compared to the other genotypes (3.2-4.0 mm). These figures were similar to those obtained in spring 2016 (3°C treatment) (see Figure 3.21). In autumn 2016, both ‘Amuri’ and ‘Nagara’ (3.2- 3.7 mm) surpassed ‘UK’ and ‘BC’ stem diameter (1.4-1.7 mm). At 1°C, genotype x storage

58 duration interactions influenced this variable. Overall, each genotype developed larger stem diameters as time progressed throughout the trial.

In spring 2016 at 1°C, genotypes could be pooled and assessed over the trial period. In brief, stem diameter gradually reduced over time (4.5 to 3.5 mm) (refer to Figure 3.22).

3.3.8 Chlorophyll content

Only marginal changes in chlorophyll content were observed throughout all trials. Apart from rhizomes stored at 3°C in spring 2016 displaying significant storage duration effects, all other harvest periods exhibited genotype x storage duration effects. Chlorophyll content typically ranged from 320.9-640.3 mg m-2.

3.3.9 Moisture content

During the spring 2015 trial, no significant differences in total MC were found between storage temperatures or over time. MC ranged from approximately 28 to 61%.

Differences in the total MC content of rhizomes stored at 1 and 3°C could be attributed to genotype x storage duration interactions during the autumn 2015 trial; however, the only considerable values were found with ‘BC’ at time point 1 (61.5%) and ‘Illinois’ at time point 2 (31.5%) in 3°C (see Figure 3.23). When comparing ‘Illinois’ and ‘UK’ at both 1 and 3°C, MC declined slightly for both (approximately 54% at time point 1 to 49% at time point 5), with ‘Illinois’ containing marginally higher values.

Spring 2016 showed that ‘UK’ and ‘Amuri’ contained the highest (57.8-65.4%) and lowest (46- 50.8%) total MC when stored at 3°C, respectively. All other samples remained relatively constant throughout storage. For samples stored at 1°C, MC was significantly affected by storage duration, though little significant variation was observed (41.8-54.9%).

Regarding rhizomes stored at 1 and 3°C in autumn 2016, genotype x storage duration interactions affected total MC, with ‘Amuri’ and ‘Nagara’ generally displaying less MC (approximately 40- 50%) than ‘BC’ and ‘UK’ (55-65%).

3.3.10 Total starch content

During the spring 2015 trial, no statistical differences were realized in total starch content between storage treatments or over time. However, samples stored at 0°C demonstrated about 5% less total starch content over the first 63 days. The values calculated for ‘Illinois’ rhizomes during this trial ranged from 13.2 (time point 6) to 23.8%, w/w (DW basis) (time point 2).

Statistical differences in total starch content over time and between genotypes were realized during the autumn 2015 trial with samples stored at 3°C. Relatively high starch levels were found in ‘Illinois’, ‘BC’, and ‘UK’ genotypes at 28 days of storage [25.4-32.2%, w/w (DW basis)]; these values gradually declined as storage time persisted [‘BC’ displayed the lowest values of the Port Ryerse samples, with total starch content reaching a menial 10.6%, w/w (DW basis)], although a slight increase was exhibited by ‘Illinois’ rhizomes at time point 5 [24.8%, w/w (DW basis)]. Closer to storage initiation, ‘Amuri’ and ‘Nagara’ contained between 13.5 and 17.2%, w/w (DW basis) starch content. The levels in ‘Amuri’ remained relatively consistent until time point 3 [16.3%, w/w (DW basis)], which was followed by a gradual decline in starch content until time point 5 [10.8%, w/w (DW basis)]. ‘Nagara’ presented a dynamic increase in starch levels at time point 3 [26.2%, w/w (DW basis)], ensued by a slight reduction by the end of the storage period [21.1%, w/w (DW basis)]. When genotypes were pooled, total starch content ranged from 22.8 (time point 1) to 16.5%, w/w (DW basis) (time point 5). When these trends are compared to the 1°C storage temperatures (‘Illinois’ and ‘UK’ genotypes only), there were no significant differences between genotypes or storage duration; however, both ‘Illinois’ and ‘UK’ experienced similar drops in total starch content at 56 days of storage. When these two genotypes were pooled, total starch content was lowest at time points 2 and 5 [approximately 15%, w/w (DW basis)] and highest at time point 1 [28.8%, w/w (DW basis)].

When total starch content was assessed in samples harvested in spring 2016 and stored at 3°C, these trends were similar to those observed in autumn 2015 at 3°C (see Figure 3.24). After genotype pooling, these values ranged from 14.5 (time point 5) to 25.2%, w/w (DW basis) (time point 1). In spring 2016 samples stored at 1°C, total starch content trends for Port Ryerse genotypes also reflected those seen in autumn 2015, 3°C. However, the trends for ‘Amuri’ differed from those observed in autumn 2015; total starch levels remained constant between 10 and 13.6%, w/w (DW basis), with slight reductions occurring after 56 days in storage.

Total starch content for autumn 2016 samples stored at 3°C peaked at time point 4 [27.8%, w/w (DW basis)] and 2 [24.9%, w/w (DW basis)] for ‘UK’ and ‘BC’, respectively; however, there were no significant differences between these genotypes or over the storage duration. The starch levels in ‘Nagara’ remained constant over time [highest value of 24.6%, w/w (DW basis) at time point

5], while ‘Amuri’ values dipped at time point 4 [8.6%, w/w (DW basis)] and peaked at the end of the storage period [31.9%, w/w (DW basis)].

The differences found in total starch content in the autumn 2016, 1°C treatment trial could be explained by individual genotype and storage duration factors; ‘Illinois’ and ‘UK’ produced the highest starch levels overall [20.8-22%, w/w (DW basis)] while ‘Amuri’ developed the lowest [10.9%, w/w (DW basis)] (refer to Figure 3.25). When genotypes were pooled and analyzed over time, total starch content declined in a linear fashion from 22.8%, w/w (DW basis) in time point 1 to 14.9%, w/w (DW basis) in time point 5.

3.3.11 Total soluble carbohydrate content

Though there were no statistical differences in total soluble carbohydrate concentration between samples in the spring 2015 trial, these values ranged from 247.87 (time point 5) to 371.67 (time point 6) mg g-1 DW. In autumn 2015 samples stored at 3°C, total soluble carbohydrate

66 concentrations were highest in ‘UK’ and ‘BC’ genotypes (251.65-263.62 mg g-1 DW) throughout the duration of the experimental trial, while ‘Nagara’ exhibited the lowest values (120.64 mg g-1 DW). Although storage time did not significantly influence these values, total soluble carbohydrate concentration ranged from 183.47 (time point 3) to 201.79 mg g-1 DW (time point 1) when genotypes were pooled. Regarding samples stored at 1°C, total soluble carbohydrate concentrations differed significantly between genotypes and over storage duration. ‘Illinois’ carbohydrate levels were statistically lower than ‘UK’ in time points 1 and 2; however, both had similar levels thereafter. Overall, values at 1°C storage diverged from 120.42 at time point 4 to 436.64 mg g-1 DW at time point 2.

Total soluble carbohydrate concentrations showed considerable differences between genotypes and storage temperatures in spring 2016. Generally, 1°C treatments resulted in higher total soluble carbohydrates than 3°C for ‘UK’ (417.39 and 271.91 mg g-1 DW, respectively) and ‘Illinois’ (232.65 and 177.02 mg g-1 DW, respectively). ‘BC’ demonstrated higher levels in 3 (333.17 mg g- 1 DW) than 1°C (240.37 mg g-1 DW) storage, while ‘Amuri’ showed no apparent differences between treatments (approximately 141-152 mg g-1 DW). Overall, ‘UK’ and ‘BC’ displayed the highest values at both treatments, while ‘Amuri’ and ‘Nagara’ had the lowest. When genotypes were pooled, total soluble carbohydrate concentrations ranged from 203.48 (time point 2) to 239.05 mg g-1 DW (time point 5). In 1°C storage, total soluble carbohydrate concentrations remained consistently higher in ‘UK’ than the remaining genotypes over the experimental trial (379.7 to 486.43 mg g-1 DW). At time point 2, ‘Illinois’ experienced a sharp decline in soluble carbohydrate levels (352.81 to 131.61 mg g-1 DW) while ‘BC’ and ‘Amuri’ showed comparable increasing trends (231.48 to 300.13 and 129.18 to 170.26 mg g-1 DW, respectfully). Levels of each genotype were then maintained until the end of storage.

In autumn 2016 samples stored at 3°C, ‘UK’ and ‘BC’ exhibited the highest total soluble carbohydrate values at time point 1 (447.87 and 471.7 mg g-1 DW, respectively). ‘Amuri’ and ‘Nagara’ values peak at time points 2 (190.75 mg g-1 DW) and 3 (314.88 mg g-1 DW), respectively, before either gradually declining (‘Amuri’) or remaining constant (‘Nagara’) until the end of the experimental trial. In brief, ‘Amuri’ samples contained the lowest total soluble carbohydrate concentration. The 1°C storage treatment demonstrated that ‘UK’ at time point 1 had the highest total soluble carbohydrate values (641.93 mg g-1 DW), followed by ‘BC’ (526.71 mg g-1 DW) and

‘Illinois’ (433.77 mg g-1 DW). These genotypes followed very similar trends of carbohydrate decline between time points 2 and 3 (215.38-256.37 mg g-1 DW), with levels that persisted until the end of storage. ‘Amuri’ exhibited overall lower carbohydrate concentrations and did not experience a significant reduction in levels throughout the experimental trial (165.05-220.09 mg g-1 DW) (refer to Figure 3.26).

3.3.12 Total soluble carbohydrates consisting of D-glucose

No statistical differences were found between storage temperatures or storage duration for ‘Illinois’ samples harvested in spring 2015 and tested for D-glucose content. Briefly, D-glucose content ranged from 13.8 (time point 5) to 18.6% (time point 3).

D-glucose levels showed considerable differences over time and between genotypes in autumn 2015 samples stored at 3°C. ‘BC’ and ‘Nagara’ levels remained relatively constant over the storage period. ‘Illinois’ and ‘UK’ experienced sharp peaks in D-glucose content by the second time point, slight reductions by the third time point, then gradual increases until the end of storage. In contrast, ‘Amuri’ showed a sharp decline in D-glucose at time point 2, increases at time point three, then

68 maintained those levels until the end of the experimental period. After pooling genotypes, values spanned from 23.8 (time point 3) to 28.5% (time point 2). At 1°C, D-glucose levels displayed a consistent downward trend in ‘UK’ over the duration of storage (46% at time point 1 to 11.7% at time point 5). ‘Illinois’ initially had lower levels than ‘UK’ (23.7%); however, maintained higher levels beginning at time point 2 (43.5% at time point 2 to 27.7% at time point 5). In samples harvested in spring 2016 and stored at 3°C, D-glucose ranged from 13% (time point 3) to 21.4% (time point 2) when genotypes were pooled. No significant differences were realized in this trial. For samples stored at 1°C, ‘UK’ and ‘BC’ experienced overall decreases in D-glucose content over the storage duration (concluding with 8.9-14.1%), though began with the highest levels (31.7 to 41.9%). ‘Illinois’ and ‘Amuri’ demonstrated peak D-glucose levels at time point 2 (26.6 and 18.7%, respectively), reductions at time point 3 (12.4 and 7.2%, respectively), but then returned to levels at time point 5 similar to those at time point 1 (24.8 and 17.5%, respectfully) (refer to Figure 3.27). D-glucose concentration generally displayed an upward trend throughout the storage period for ‘BC’ and ‘UK’ (24.6-28.2% at time point 1 to 31.8-47.1% at time point 5) after autumn 2016 harvest and storage at 3°C. ‘Nagara’ experienced increased D-glucose levels between time points 2 to 4 (21.7-26.4%) before returning to levels resembling those at the start of the trial (17.4%), while ‘Amuri’ displayed declining D-glucose between time points 2 to 4 (6.5-18%). For samples stored at 1°C, differences in D-glucose content could be explained by genotype variations; ‘UK’ (33.6%) and ‘Amuri’ (18.1%) exhibited the greatest and lowest values, respectfully (see Figure 3.28).

3.3.13 Total soluble carbohydrates consisting of sucrose

The percentage of total soluble carbohydrates consisting of sucrose ranged from 48.9 (time point 3) to 60.7% (time point 6) in spring 2015 trial. These means were not significantly different and were not heavily influenced by storage temperature or storage duration.

In samples harvested in autumn 2015 and stored at 3°C, sucrose levels increased slightly in ‘BC’ and remained constant between time points 3 to 5. ‘Illinois’ and ‘UK’ displayed considerable declines in sucrose at time point 2, then showed increases to similar levels found at time point 1. Conversely, both ‘Amuri’ and ‘Nagara’ exhibited significant sucrose content increases at time points 2 and 3, respectfully, with reductions occurring immediately thereafter. When genotypes were pooled, sucrose content ranged from 38.9 (time point 2) to 56.3% (time point 3). Sucrose levels in ‘Illinois’ stored at 1°C exhibited a sharp decline at time point 2 (17%) which increased to levels similar to time point 1 thereafter (43.8-52%). ‘UK’ had reduced sucrose levels during the first two time points (12.5-12.6%), then experienced a significant increase that persisted until the end of the experimental period (55.9-69.5%). No significant differences in sucrose content were found in samples harvested in spring 2016 and stored at 3°C. Overall, sucrose content values ranged from 53.9 (time point 2) to 71.6% (time point 3). In 1°C treated samples, sucrose content resembled the inverse of D-glucose trends found in this trial, with ‘Illinois’ and ‘Amuri’ starting with the highest (61.6 and 78.2%, respectfully) and ‘UK’ and ‘BC’ initiating with the lowest (26.8 and 16.7%, respectfully) sucrose content (see Figure 3.29). Variation of means was significantly affected by genotype in autumn 2016 samples stored at 3°C, with ‘Amuri’ and ‘Nagara’ experiencing higher sucrose concentrations (54.6-63%) than ‘UK’ and ‘BC’ genotypes (32.1-35.4%) (see Figure 3.30). At 1°C treatment, no significant differences were found for sucrose content. After pooling genotypes, sucrose content values ranged from 31.5 (time point 3) to 49% (time point 4).

3.3.14 Total soluble carbohydrates consisting of D-fructose

The percentage of soluble carbohydrate content consisting of D-fructose ranged from 21.6 (time point 1) to 33.1% (time point 2) in ‘Illinois’ rhizome samples harvested in spring 2015 and stored at either 0 or 3°C. No significant differences were actualized in this trial.

D-fructose levels were significantly different between genotypes and throughout the storage duration in samples harvested in autumn 2015 and stored at 3°C. The genotypes obtained from Port Ryerse showed trends similar to those assessing D-glucose content. ‘Amuri’ D-fructose levels were relatively elevated between time points 3 to 5, while ‘Nagara’ displayed noticeable declines in levels until time point 3, with gradual increases thereafter. After pooling genotypes, D-fructose content spanned from 19.9 (time point 3) to 32.7% (time point 2). In 1°C samples, peaks in both ‘Illinois’ and ‘UK’ occurred at time point 2 (39.8-50.8%). In general, D-fructose levels were lowest at time point 5 (23.1%) and highest at time point 2 (45.3%) for rhizomes conserved at 1°C. In spring 2016 samples stored at 3°C, D-fructose content ranged from 15.3% (time point 3) to 24.9% (time point 2), with no considerable differences between means. For rhizomes stored at 1°C, D-fructose levels peaked at time point 2 for ‘Illinois’ and ‘UK’, and showed sharp declines at time point 3 (35.7-39.7% and 17-17.8%, respectfully). ‘BC’ exhibited a consistent downward trend over the storage duration (50.2% at time point 1 to 17.5% at time point 5). ‘Amuri’ peaked at time point 4 and levels remained constant until the end of storage (see Figure 3.31). In autumn 2016, D-fructose mean variations could be explained by genotype factors when stored at 3°C (refer to Table 3.3); ‘UK’ and ‘BC’ had higher levels (36.2-37.1%) than Elora genotypes (21.5-24.3%). At 1°C storage treatment, D-fructose concentrations differed significantly in ‘BC’ samples at time point 1 (50.8%), ‘UK’ at time point 2 (45.7%), and ‘Amuri’ rhizomes at time point 5 (11.5%).

Table 3.3 Average percentage (%) of total soluble carbohydrate concentration consisting of D- fructose in rhizomes stored at 1 and 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April and November/December 2016. Each value is represented by 32±SE samples. D-fructose (% total soluble carbohydrate concentration) Genotype 1°C 3°C ‘Illinois’ 23.28±1.8703 - ‘UK’ 28.45±1.8965 21.76±1.9428 ‘BC’ 22.03±1.8857 19±1.8651 ‘Amuri’ 15.9±1.8793 16.81±1.862

3.4 Discussion

The current demand to diversify product and energy resources is an undeniable reality, and M. x giganteus has overwhelming support in the scientific literature for its social and environmental benefits as a leading second-generation biomass feedstock. However, to secure this species’ future in biomass cultivation on a commercial-scale, the economic challenges rooted in establishment costs must be resolved. Examining the storage potential of the most utilized and least expensive

76 available propagule source for M. x giganteus - rhizome cuttings - and determining if the factors influencing conservation and successful regrowth from these propagules could be ascertained, was performed in the current study. The results from this study may provide insight into prospective optimization of M. x giganteus preservation, which may contribute to a greater supply of germplasm in the market and reduced establishment costs.

Three- to six-way interactions commonly arose when attempting to compare individual tested variables between storage temperature, harvest season, and/or year. The extent of variation described in this study is not unheard of in Miscanthus, where coefficients of variation for tiller number, biomass FW, and biomass DW had exceeded 40% over a three-year field trial in parent M. sinensis (Nie et al. 2016), and where ‘genotype x cultivation year’ and ‘genotype x cultivation year x location’ interactions had been previously reported by Jeżowski et al. (2011) for a number of characteristics during the first three years of growth in M. x giganteus.

To better understand the data, genotypes and storage times investigated at each storage temperature and harvest season were analyzed. What was discovered from this approach was that there was very little consistency in which factors were most influential for each tested trait (refer to Appendices 5.1-5.6). For example, tiller height was the only growth parameter assessed that demonstrated significant changes over the storage duration in the spring 2015 trial; however, this trait was shown to be significantly affected by genotype in autumn 2015, and genotype and storage duration in autumn 2016 in samples stored at 3ºC. To complicate the results further, tiller height for samples stored at 1ºC were most significantly impacted by storage duration and genotype x storage duration in autumn 2015 and autumn 2016, respectively.

When variables were assessed across harvest trials, very few patterns arose in the data. No type III fixed effects were demonstrated for: tiller and stem node number during spring 2015 and/or 2016 trials from samples stored at either 1 or 3ºC; stem node number during autumn 2015 and 2016 trials from samples stored at 1ºC; and the percentage of total soluble carbohydrates consisting of D-glucose, sucrose, and D-fructose during spring 2015 and 2016 trials from samples stored at 3ºC. Conversely, ‘genotype x storage duration’ interactions were significant for: chlorophyll content during autumn 2015 and 2016 trials from samples stored at 3ºC; emergence speed during autumn 2015, spring 2016, autumn 2016 trials from samples stored at 1ºC; and stem diameter during autumn 2015 and 2016 trials from samples stored at 1ºC.

As expected, average culm emergence speed accelerated as storage time progressed during both autumn trials, while spring values remained relatively unchanged over time. These results cohere to what was predicated by Barling et al. (2013) who conducted transcriptome analysis of autumn and spring Miscanthus rhizomes; rhizomes harvested in late fall demonstrated higher frequencies of transcripts associated with dormancy and seed maturation, while those harvested in the spring exhibited a higher number of transcripts related to PGR biosynthesis/signalling, cell wall development, and root production. Specifically, spring rhizomes were found to have a significant enrichment of transcripts affiliated with jasmonic acid signalling, and exogenous supplementation of this PGR has been shown to induce shoot formation in subterranean tubers in vitro (Koda and Kikuta 1991). Additionally, Miscanthus innate rhizome dormancy is maintained over late autumn and winter until favourable growing conditions are met in the spring (Xue et al. 2015). During the autumn trials, ‘Amuri’ and ‘Nagara’ genotypes had emerged faster after planting than their Port Ryerse counterparts. Furthermore, genotypes originating from Elora also outperformed those from Port Ryerse in terms of viability, tiller height, tiller number (3°C only), leaf number, and stem diameter. Regarding reserve carbohydrate content, ‘Amuri’ exhibited consistently lower levels of total starch, total soluble carbohydrates, and relative D-glucose and D-fructose concentrations, while its relative sucrose levels surpassed all other genotypes. Though confounding factors complicate the analysis, better performance of Elora-originating genotypes could be attributed to: 1) strictly genotype effects; 2) different endemic soil types (Guelph silt loam soil in Elora and either a heavy clay or sandy soil in Port Ryerse, ON) (Presant and Acton 1984; Soil Classification Working Group 1998); 3) latitude (43°38’ N, 80°24’ W and 42°47’ N, 80°12’ W for Elora and Port Ryerse, ON, respectively); and 4) lake effects caused by the vicinity of Lake Erie to the Port Ryerse Miscanthus plots (Sousounis and Fritsch 1994).

Of the five genotypes assessed for rhizome storage and growth, ‘Amuri’ was the only accession to develop rhizomes with a tussock/indeterminate form, while ‘Illinois’, ‘UK’, ‘BC’, and ‘Nagara’ produced rhizomes with determinate growth patterns (Withers 2015). The tussock morphology is characteristic of what is formed in diploid parent M. sinensis, while M. x giganteus and M. sacchariflorus generally develop rhizomes characterized by a spreading nature. Clifton-Brown and Lewandowski (2000) attributed the tussock rhizome growth from M. sinensis to advantageous cold tolerance, and this may partially-explain the exceptional performance of ‘Nagara’ throughout the study; however, this is only speculative.

For commercial-scale production and trade of Miscanthus to thrive in Canada, rhizome cuttings that can survive and establish after various lengths of time in storage is critical. Both ‘Nagara’ and ‘Amuri’ demonstrated exceptional rhizome viability throughout all trials (ranging from 60.9 to 100%) regardless of storage temperature. In contrast, ‘Illinois’ and ‘BC’ exhibited lower rhizome survival, while ‘UK’ demonstrated intermediate levels. The reasons for these differences in survivability between genotypes may be caused by a number of factors, including but not limited to respiration rates, ethylene production, and accumulation of toxic phenolics (Hamadina and Craufurd 2015; Purdy et al. 2015; Dai et al. 2016); all of which warrants further examination.

Tiller height, leaf number, and stem node number gradually increased as storage time progressed after autumn harvests, and either stagnated or declined with storage time during spring trials. These results are consistent with the trends found in sprouting speed, suggesting that growth rates do not vary considerably after culm emergence has begun.

Carbohydrate reserves are critical for non-photosynthetic storage organs and allow for important plant processes such as dormancy, respiration, and shoot re-emergence (White 1973). When assessing general trends across genotypes over a storage duration, spring-harvested samples demonstrated minimal fluctuations in total starch, increasing total soluble carbohydrates and sucrose content, and decreasing reducing sugars. Theoretically, these trends would show the opposite effect in autumn-harvested samples; however, this was not the case. The expected carbohydrate trends for autumn trials were realized with total soluble carbohydrates (declining), D-glucose (increasing), and D-fructose (only ‘Illinois’ and ‘Amuri’ exhibited rising values over time). During autumn trials, total starch levels declined slightly and sucrose levels remained relatively consistent over time.

The lack of correlation between starch and total soluble carbohydrates in this study was also reported on by Davies et al. (2011) who observed that starch levels in M. x giganteus rhizome tissue did not deplete significantly as storage time progressed; however, total soluble carbohydrate content increased during this period. It was hypothesized that this increase may have been caused by the hydrolysis of structural carbohydrate sources (eg. hemicelluloses), but this has yet to be substantiated.

The trends observed in the current study are not in agreement with those accounted for in other species that develop below-ground storage organs, such as greenhouse-cultivated blueberry.

Townsend et al. (1968) demonstrated that from the start of the growing season until June, starch and sucrose decreased while reducing sugars increased. Another unexpected result obtained from our study was that while ‘Amuri’ outperformed the Port Ryerse genotypes in regards to viability and faster culm re-emergence, it demonstrated the overall lowest starch content. Clifton-Brown and Lewandowski (2000) found that Miscanthus species with relatively higher rhizome starch content and lower reducing sugar levels (M. sinensis) were better adapted for overwintering survival in European field trials. Although ‘Amuri’ had the lowest starch content, it also displayed the lowest reducing sugar concentrations in both harvest seasons, suggesting that these carbohydrates may be more influential than starch in maintaining and releasing dormancy in Miscanthus rhizomes. To further support this theory, ‘UK’ and ‘BC’ demonstrated intermediate starch contents, but the overall greatest concentrations of reducing sugars over each trial; this may explain their poor growth in greenhouse conditions compared to the remaining genotypes (Dai et al. 2016).

Viola et al. (2001) had reported that the timing of certain developmental processes in potato tubers were correlated with apoplastic and symplastic phloem unloading of sucrose; reducing sugars produced through the hydrolysis of apoplastic sucrose by cell wall-bound acid invertase are utilized for growth and development, while the accumulation of carbohydrate reserves occurs by the filling of symplastic sucrose into intracellular plastids (Godt and Roitsch 2006). Genotypes ‘UK’ and ‘BC’ exhibited the greatest carbohydrate changes over storage duration and harvest seasons and manifested these trends more explicitly than the remaining genotypes. For instance, sucrose levels increased while reducing sugar content decreased in these genotypes as time progressed during the spring 2016 trial; as sucrose accumulated, culm emergence speeds and rhizome viability diminished (followed by the inevitable loss of overall plant vigor). Hajirezaei and Sonnewald (1999) had demonstrated a similar phenomenon in potato tubers; hindering sucrose mobility by reducing cytosolic pyrophosphate in potato tubers resulted in increases in sucrose, UDP-glucose, and fructose-6-bisphosphate concentrations, a reduction in starch accumulation, unaltered respiration rates, and sprouting inhibition. Future investigation into the compartmentalization of sucrose and reducing sugars in the rhizome tissue at distinct developmental processes may allow for a more complete picture of rhizome dormancy induction, maintenance, and release in different Miscanthus genotypes.

3.5 Conclusions

The current study was able to provide support for genotype differences in rhizome dormancy and storage potential, physiological growth characteristics, and carbohydrate dynamics in five M. x giganteus genotypes grown in southwestern Ontario. Although these differences were not further verified by a formal ‘G’ x ‘environment’ (‘E’) x ‘G x E’ experimental design, genotype differences of Ontario-cultivated Miscanthus accessions had been previously reported by Rosser (2012) and Withers (2015) regarding overwintering, first year survival, and flowering.

Based on the information obtained in this study, rhizome harvest can be conducted in either spring or autumn without consequential loss of rhizome viability, especially in ‘Amuri’ and ‘Nagara’ genotypes. Considering this point, spring-harvested rhizomes tend to sprout faster and grow taller than their autumn-collected counterparts (within the first six weeks of planting) during conventionally-preferable establishment periods in southwestern Ontario; however, this requirement may change depending on where the rhizomes from each harvest will be cultivated.

It is important to indicate that while minute differences in Miscanthus greenhouse performance were detected from rhizomes stored at 0/1 or 3℃, prolonged storage of rhizomes should likely be managed using 1℃ conditions. This conjecture is predicated on the anecdotal observations of ice crystal formation detected on rhizome tissue collected in 0℃ conditions, and disease symptoms appearing more severe in germplasm stored at 3℃ than at 1℃.

3.6 Acknowledgements

The authors wish to thank Kevin Piunno, James Nicholson, and Scott Belton for their assistance with germplasm collection and processing, All Weather Farming Inc. and the Elora Research Station for supplying rhizome material, and Dr. Michelle Edwards for aiding with statistical analyses. The authors are also grateful for our funding partner BioFuelNet for financially supporting this research. The funders had no role in the design of the study, data collection and analysis, decision to publish, or preparation of the manuscript.

4: Improving regeneration capacity of Miscanthus x giganteus ‘M161’ calli through inhibition of the phenylpropanoid biosynthetic pathway

Abstract

In previous studies, regeneration rates of Miscanthus x giganteus from calli cultured on semi-solid media significantly declined after four months of culture, presenting problems in germplasm conservation and use as an alternative propagation system. Due to the species’ lignocellulosic nature, it was hypothesized that the accumulation of phenolic compounds in calli may be responsible for inhibiting regeneration. The current study aimed at optimizing regeneration of Miscanthus calli by culturing it in the presence of AIP, a competitive inhibitor of phenylalanine ammonia lyase (PAL), to reduce the biosynthesis of phenolics. Embryogenic calli was cultured on media supplemented with 9.0 or 11.3 μM 2,4-D and 0, 1, 10, 100, or 1000 μM AIP. Every 28 days for seven months, calli were visually classified based on morphology and regeneration rate. Over the duration of the study, regeneration of shoot-like structures was consistently highest in calli cultured on 11.3 μM 2,4-D supplemented with 10 or 100 μM AIP (13-58.3%), and plantlet development from calli cultured on all levels of AIP demonstrated normal growth and morphology. Total soluble phenolic content of calli decreased in a dose-dependent manner from 2242.34 µg g- 1 dry weight in the control to 1569.71 µg g-1 dry weight in AIP-treated callus. Our data indicate that inhibiting PAL in Miscanthus cultures increases regeneration rates, extends the period in which the callus is competent, and that normal plantlets are produced from this process.

4.1 Introduction

Miscanthus x giganteus Anderss. (M. x giganteus) is a fast-growing, temperate-adapted member of the Poaceae (Hodkinson et al. 2002). The potential of M. x giganteus as a sustainable, non-food, lignocellulosic feedstock for bioethanol (Lewandowski 1998) and value-added product (Engbers and Deen 2013) development has been previously demonstrated. Its perennial nature allows for high yields of aboveground biomass for 20-25 years depending on location and agronomic practices (Xue et al. 2015). After three years of initial establishment, annual biomass yields can range between 31-61 tonnes per hectare (ha-1), significantly greater than other biomass crops such as maize or switchgrass (Panicum virgatum) (Heaton et al. 2008). Net profit for yields can amount to as much as $2,900 ha-1 annually ten years after initial establishment (Heaton et al. 2004).

Despite these advantages, adoption of M. x giganteus for commercial production is limited, in large part due to establishment costs. M. x giganteus is a seedless triploid (3x=57) that can be propagated using DRP or micropropagation with estimated costs of approximately $5,430 and $9,800 ha-1, respectively (Atkinson 2009; Xue et al. 2015). While micropropagation using current techniques is more expensive that DRP, plantlets derived from somatic embryos (embryoid plants – EP) exhibited greater shoot lengths and tiller number per plant (51 and 19.1 for DRP; 57 and 47.1 for EP, respectively) in the first year of field establishment in Germany (Lewandowski 1998). In regard to morphological characteristics, the number of branches over 15 cm long and number of bamboo-like shoots were greater in EP (7.0 and 11.6, respectively) than DRP (3.9 and 5.5, respectively). In addition to these characteristics, EP produced twice as much shoot biomass and their rhizomes were 135% heavier than DRP in the first year of planting. Further, in vitro technologies can facilitate genetic improvement of this sterile species through mutagenesis or genome editing (Mehrotra et al. 2007), production of certified disease-free germplasm for national and international exchange (Taşkın et al. 2013), rapid multiplication of “elite” genotypes that exhibit genetic integrity (Rambaud et al. 2013), and aid in the study of plant processes without interference from unwanted artefacts (Vreugdenhil et al. 1998). As such, in vitro technologies offer many advantages for Miscanthus, but work is needed to improve the technology.

Indirect regeneration of various species of Miscanthus have been successfully developed and have mainly focused on somatic embryogenesis from immature inflorescences (Lewandowski 1997; Petersen 1997; Kim et al. 2012; Gubišová et al. 2013; Perera et al. 2015; Ślusarkiewicz-Jarzina et al. 2017). Somatic embryo development in M x giganteus is typically induced with 2,4-D (Kim et al. 2012) and low levels of the cytokinin BAP (Lewandowski 1998). A major challenge in M. x giganteus micropropagation is the loss of competence during prolonged callus culture. Kim et al. (2010) attempted to improve regeneration of M. x giganteus plantlets from immature inflorescence cultures. Six weeks after initiation of calli, the majority were classified as yellow/white compact with shoot-like structures (41±4%), followed by K2 (22±2.1%) and K3 (37±3.7%) calli. Regeneration frequency of calli was greatest for shoot-forming and embryogenic-like calli after one (0.93) and four months (0.9) of culture on solid callus medium, respectively. However, regeneration was not achieved from callus maintained for more than four months (Kim et al. 2010).

Like other grass species, such as Sorghum bicolor (L.) Moench, regenerative capacity of calli maintained on callus maintenance media over long periods of time may be hindered by an accumulation or change in phenolic acid content (Liu et al. 2015), leading to recalcitrance and/or toxic effects to plant tissues (Zaid 1987), modification of endogenous phytohormone levels (Březinová et al. 1996), control of dormancy in somatic embryos, and influence over organogenesis (Cvikrová et al. 1998). Changes in phenolic content may be due to an upregulation of the phenylpropanoid biosynthetic pathway, or degradation of lignin macrostructures in response to wounding (Zawadzki and Ragauskas 1999). In M. x giganteus stems, Le Ngoc Huyen et al. (2010) demonstrated that cell walls were largely composed of ester-linked p-coumaric and ferulic acids (3:1). Regeneration of plantlets from M. x giganteus callus culture could be hindered due to formation of premature lignin from these soluble phenolic compounds in callus cells and/or somatic embryos.

Phenol and lignin biosynthesis has been shown to be significantly reduced in in vitro American elm suspension cultures, sugar maple callus, and Artemisia callus cultures by the addition of AIP into the culture media (Jones et al. 2012; Jones and Saxena 2013). AIP competitively inhibits PAL, the enzyme responsible for first committed step of the phenylpropanoid biosynthetic pathway that converts phenylalanine to ammonia and trans-cinnamic acid (MacDonald and D’Cunha 2007). In A. annua callus cultures, it was found through both visual inspection and total phenolic content that tissue browning decreased in a dose-dependent manner up to 10 μM AIP, and that 100 μM AIP significantly reduced total tissue phenolics (Jones and Saxena 2013). In addition, total soluble phenolic content has been demonstrated to be correlated to the amount of lignin found in somatic embryos and calli in vitro (Cvikrová et al. 2003), and embryo growth/development was negatively correlated with total phenolic content (Malá et al. 2000). Similar results have been observed in a number of species in vitro, including maize, sessile oak, and alfalfa (Lozovaya et al. 1996; Cvikrová et al. 1998, 1999, 2003; Hrubcová et al. 2000).

The objective of the current study was to optimize the regeneration capacity of M. x giganteus calli after prolonged conservation on semi-solid media. It was hypothesized that application of AIP would reduce total soluble phenolic content in callus, and subsequently improve regeneration potential and timeframe. The findings of this research may help to improve germplasm

84 conservation, maintain disease-free stocks, and create a foundation for future in vitro breeding technologies.

4.2 Materials and Methods 4.2.1 Plant material

Immature inflorescences (5.0-15.0 cm in length) were collected from field-grown M. x giganteus ‘Illinois’ (‘M161’) at All Weather Farming Incorporated (Port Ryerse, Ontario, Canada) (42°47′ N 80°12′ W) on September 15, 2015. Through visual inspection using a modified Biologische Bundesantalt, Bundessortenamt and CHemische Industrie (BBCH) scale, the majority of plants grown at this location were in principal growth stage 4: booting; specifically, varying between stages 43 to 47 (Tejera and Heaton 2017). Stems were cut beneath the most distal (youngest) node with sheath intact. Segments were wrapped in moist paper towel and kept in darkness at 4°C for five days. Thereafter, all but two-to-three surrounding leaves were removed from the immature inflorescences. Immature inflorescences were rinsed under running tap water for five minutes, soaked in 70% ethanol for one minute, then rinsed with sterile, deionized water thrice. Immature inflorescences were then immersed in 20% commercial bleach (12.5% sodium hypochlorite; Chlorox®) with approximately 0.1% tween 20 (Fisher Scientific Company, Ottawa, Canada) for 20 minutes with gentle shaking, followed by rinsing five times with sterile, deionized water. Explants were then aseptically dissected and divided into pieces of approximately 5.0 mm for callus induction.

4.2.2 Callus induction and multiplication

Callus induction and maintenance medium consisted of MS basal salts and vitamins (Phytotechnology; Shawnee Mission, KS, USA) (Murashige and Skoog 1962) supplemented with 30 g l-1 sucrose, 11.3 μM 2,4-D, 0.5 mg l-1 BAP, 1000 mg l-1 L-proline, 300 mg l-1 casein hydrosylate, and 7 g l-1 agar (Kim et al. 2012). All PGRs and L-proline were purchased from Sigma-Aldrich® (Sigma-Aldrich, St. Louis, MO, USA); casein hydrosylate from PhytoTechology Laboratories®; and agar from Fisher Scientific®. Callus induction medium was adjusted to pH of 5.7 before autoclaving at 121°C and 21 psi for 20 minutes. Ten explants were placed on each plate for callus induction (100 x 15 mm; Fisher Scientific®). Calli were subcultured and multiplied monthly (approximately every 28 days) for six months. In the fifth month of culture, calli measuring 3.0-8.0 mm in diameter were inoculated on one of five callus maintenance media.

Media were composed as described previously, and supplemented with either 0, 1, 10, 100, or 1000 μM AIP. In the sixth month of culture, half the calli from each of these treatments were placed on callus maintenance media modified with reduced 2,4-D (9.0 μM) and 0, 1, 10, 100, or 1000 μM AIP. Calli were incubated in the dark in a growth room at 24±2°C for the duration of the experiment. At each subculture time point, visibly necrotic/atrophied tissue weas gently removed from calli.

4.2.3 Callus morphology assessment

At each subculture (beginning after the initial six months of culture initiation and multiplication), calli were visually assessed for morphology using the following rating system: 1) “compact white” calli (K2), characterized by a smooth, white surface (Petersen 1997) generally accompanied by the appearance of “embryogenic-like” structures. 2) “Yellow/green” calli (K1), previously described as nodular and moderately soft. In previous grass culture studies, K1 calli types are generally root- forming (Morrish et al. 1987); however, in this study, the presence of calli resembling K1 types void of roots was reported. 3) “Friable” calli (K3), identified by a soft, watery appearance (Kim et al. 2010) usually lacking embryogenic capacity (Morrish et al. 1987), and 4) “browning” calli. It is important to note that browning calli could have been classified as any of the other callus morphologies; however, tissue necrosis due to any number of stressors would make its original classification indiscernible (representations of different callus morphologies may be found in Figure 4.1 a-d).

4.2.4 Development of regenerants and embryo-like structures

Along with these observations, the following characteristics were recorded: formation of “embryo- like” structures; development of “root-like” structures; and the presence of “shoot-like” structures (representations may be found in Figure 4.1 e-h). The frequency of each callus type and accompanied characteristics were calculated as follows: callus type/characteristic (%)=(number of each callus type or characteristic/total number of calli plate-1)x100 (Kim et al. 2010).

The number of root-, shoot-, and embryo-like structures per callus from each plate were assessed using a dissecting microscope. Analysis of regenerant structures were conducted between nine and 13 months of culture, while analysis of embryo-like structures began at 11 months of culture. Since embryo-like structures typically developed in clusters, and quantified without destruction of the

86 calli, conservative counts of somatic embryos were conducted from the visible surface of the calli. Individual leaves on calli constituted distinct shoot-like structures. For the duration of the AIP treatment period, six calli were placed on each media with five replications each. Care was taken to group calli of similar morphologies on the same plate and randomize this among treatments.

4.2.5 Soluble phenolic content

Soluble phenolic content was measured in calli at months seven to eleven for AIP treatments supplemented with both 2,4-D concentrations. At each subculture, one callus (with necrotic/atrophied tissue removed) was taken from each treatment replicate (five replicates per treatment) and pooled in a 15 ml centrifuge tube (Fisher Scientific®). Sample FW (g) from each treatment was recorded before flash-freezing and lyophilisation. After lyophilisation, sample DW (g) was recorded, total MC (%) was calculated [((FW-DW)/FW)x100], and samples were finely ground. Extraction solvent was added to each tube (80% methanol v/v) such that the tissue to solvent ratio (w/v) was 1:10. The tubes were vortexed and placed in a sonicating water bath (Branson 3510, Danbury, CT, USA) for 30 minutes. The tubes were then removed and centrifuged for 10 minutes at 21.1 g. The supernatant from each sample was then transferred into a new microcentrifuge tube.

Soluble phenolic content was assessed using a modified Folin-Ciocalteu (F–C) colorimetric assay. A standard curve was constructed using gallic acid (Sigma-Aldrich, Oakville, Canada) to obtain gallic acid equivalent (GAE) values at concentrations of 1000, 500, 250, 125, 62.5, and 31.25 µg ml-1 (Folin and Ciocalteu 1927). In brief, 10 μl aliquots of sample extracts, standards, or sample blanks were added to each well of a 96-well flat bottom microplate (Corning, Corning, NY, USA). A volume of 100 μl of 1:10 water:F-C phenol reagent (MP Biomedicals, Santa Ana, CA, USA) was added to each well and the plate was incubated for 5 minutes before adding 80 μl of aqueous

0.25 M Na2CO3 (Sigma-Aldrich®). The plate was then incubated in the dark for 1 hour before the absorbance values at 725 nm were measured with a Synergy H1 microplate reader (Biotek, Winooski, VT, USA). All sample and standard readings were corrected with blanks, and all samples, standards, and blanks were replicated in triplicate.

4.2.6 Regeneration and plantlet formation

At six months of culture, excess calli from each 11.3 μM 2,4-D treatment measuring 3.0-8.0 mm in diameter were inoculated on regeneration medium described by Kim et al. (2012). Regeneration medium consisted of MS basal salts and vitamins supplemented with 30 g l-1 sucrose, 5 mg l-1 BAP, 1 mg l-1 2,4-D, 1000 mg l-1 L-proline, 300 mg l-1 casein hydrosylate, and 7 g l-1 agar. Regeneration medium was adjusted to pH 5.8 before autoclaving at 121°C at 21 psi for 20 minutes. Six to seven calli were placed on each plate and incubated in the dark for one month. After regeneration, calli with shoots were transferred to either root induction [MS basal salts, 30 g l-1 sucrose, 0.1% activated charcoal (w/v), 7 g l-1 agar, pH 5.8], or tillering medium [MS basal salts and MS vitamins, 30 g l-1 sucrose, 5 mg l-1 BAP, 0.1 mg l-1 indole-3-butyric acid (IBA), 0.45 mg l-1 IAA, pH 5.7] for plantlet development (Kim et al. 2012).

4.2.7 Experimental design and statistical analysis

Data obtained from non-destructive visual assessments [‘callus morphology frequencies’ (seven months), ‘number of regenerants’, ‘phenolic content’, and ‘total MC’(five months), and ‘number of embryo-like structures’ (three months)] were subject to variance analysis conducted using a mixed-model repeated measures analysis of variance (ANOVA) using PROC GLIMMEX in SAS® 9.4 software (SAS Institute Inc., Cary, NC). The experiment was constructed in a factorial design and arranged in a completely randomized design (CRD). Each treatment was replicated over five plates, each containing six samples. A logit link function with beta distribution was used for assessing callus morphology frequencies, and an identity link with lognormal distribution was used to analyze the number of regenerants and somatic embryos. For phenolic content and total MC, each analyzed treatment was composed of five individual calli from five replicate plates.

A type I and III error rate of α=0.05 was assigned for all analyses.

4.3 Results 4.3.1 Callus morphology frequencies

The percentage of calli producing shoot-like structures was significantly affected by interactions between sample timing and AIP concentration. The data demonstrated trends of decreasing shoot- like regeneration as culture time progressed in all individual treatments. With the exception of media lacking AIP in the first month of the experiment (66.7±6.811%), the greatest frequency of

88 calli developing shoot-like structures was observed on media supplemented with 10 µM AIP (peaking in the third month of culture with 58.3%). In regard to the proportion of calli producing shoot-like structures compared to root and somatic embryo-like tissues, 100 µM AIP supplemented with 11.3 µM 2,4-D exhibited the highest value by the third culture month (73.3%).

a) b)

c) d)

e) f) g) h) Figure 4.1 Representative M. x giganteus ‘M161’ callus morphologies. Calli were assessed using a dissecting microscope and were putatively classified as: a) root forming; b) shoot forming; c) somatic embryo forming (K2); and d) compact yellow/green (K1). Leaf primordia (e) and differentiated shoot-like structures (f) formed on M. x giganteus ‘M161’ calli. Roots generated on M. x giganteus ‘M161’ calli; the root-cap is identifiable by the production of red/purple pigmentation and lack of root hairs (g), and anthocyanin spots are noticeable on the remainder of the callus (h). Scale bars represent 1 mm.

The frequency of calli with root-like regenerants (%) were significantly affected by sample timing and AIP concentration. The highest frequency of rooting calli was observed on medium supplemented with 10 µM AIP (45.4±4.663%), followed by 1 and 100 µM treatments (24.5±3.705 and 25.2±3.467%, respectively). The lowest and highest AIP levels resulted in frequencies

90 between 17 and 19% when values were averaged over all subculture timepoints. Compared to calli producing other distinctive structures, 1000 µM AIP supplemented with 11.3 µM 2,4-D accounted for 60-63.3% root-forming calli by month six and seven of culture.

Three-way interactions between AIP, 2,4-D, and sampling time were observed with somatic embryo-forming calli. When the proportion of somatic embryo calli were compared to shoot- and root-forming calli, higher values were generally observed in AIP treatments supplemented with 9.0 than 11.3 µM 2,4-D (refer to regenerant proportions in Figure 4.2).

a) b)

c) d)

e) f)

g) h)

i) j)

Figure 4.2 Proportion of shoot-, root-, and somatic embryo-forming calli over time. Data are presented by 0 (a and b), 1 (c and d), 10 (e and f), 100 (g and h), and 1000 (i and j) µM AIP treatments grouped by 9.0 (a, c, e, g, and i) and 11.3 (b, d, f, h, and j) µM 2,4-D.

Differences in the frequencies of both K1 and K2 calli between subcultures were significantly affected by three-way interactions between sampling time, AIP concentration, and 2,4-D concentration. When the proportion of these callus types were compared to the percentage of calli producing browning or K3 calli, higher K2 frequencies were generally observed in AIP treatments supplemented with 9.0 µM 2,4-D while greater K1 values were associated with 11.3 2,4-D. Frequencies of K3 (0-30.7%) and browning (4.0-12.0%) calli were not significantly affected by sample timing, AIP concentration, or 2,4-D concentration. The frequencies of these callus types remained lower in proportion to K1 and K2 callus types over the duration of the experiment (refer to callus morphology proportions in Figure 4.3).

a) b)

c) d)

e) f)

g) h)

i) j)

Figure 4.3 Proportion of K1, K2, K3, and browning callus over time. Data are presented by 0 (a and b), 1 (c and d), 10 (e and f), 100 (g and h), and 1000 (i and j) µM AIP treatments grouped by 9.0 (a, c, e, g, and i) and 11.3 (b, d, f, h, and j) µM 2,4-D.

4.3.2 Regenerant number

AIP concentration significantly influenced the average number of shoot-like structures per shoot- forming callus, with 10 µM AIP exhibiting the most shoots per callus (4.18±0.5464). This was followed by 100 and 1000 µM treatments (3.84 and 3.23, respectively), with 0 and 1 µM AIP developing the fewest (2.12 and 1.55, respectively) (refer to Figure 4.4 for mean regenerant numbers averaged over time and Figure 4.5 for means at each timepoint).

The average number of root-like regenerants demonstrated significant interactions between sampling time and AIP concentration. The highest values for average number of root regenerants were observed at the fifth sampling time in treatments 1 and 100 µM AIP, and the fourth sampling time in 1000 µM AIP (17.25 and 13.23±2.444, and 18.83±3.5655, respectively). Trends for 1, 100, and 1000 µM AIP treatments demonstrated average root number increasing as time progressed throughout five months of culture, while 0 and 10 µM treatments remained relatively unchanged during this period.

Though the average number of somatic embryos ranged from 14.93 (1000 µM AIP) to 33.7 (100 µM AIP), no observable trends were detected with these levels when analyzed over the culture period.

Figure 4.4 Average number of shoots (a), roots (b), and somatic embryo-like structures (c) developing on calli cultured on different levels of AIP and averaged over time.

Figure 4.5 Average number of a) shoots, b) roots, and c) somatic embryos per callus at specific culture times (S.E. 1.1484 and 6.8712 for shoot and somatic embryo number, respectively).

4.3.3 Soluble phenolic content

Total phenolic content (µg g-1 DW) was negatively-associated with increasing AIP concentrations, with the highest content being observed at 1 µM (2242.34±149.03 µg g-1 DW) and the lowest at 1000 µM AIP (1569.71 µg g-1 DW) (refer to Appendix 4.1).

Total MC (%) was significantly affected by both 2,4-D and AIP concentrations; however, there were no interactions between these two factors (P=0.1893) (refer to Appendix 4.2). 11.3 μM 2,4- D demonstrated greater MC than the 9.0 μM treatment (85.9 and 84.4±0.4879%, respectively). In addition, MC was lowest in 10 (84.0±0.7715%) and highest in 1000 (87.3%) µM AIP. There were no significant differences among 0, 1, and 100 µM AIP treatments (84.6, 85.0, and 84.9%, respectively)

4.3.4 Plantlet development

After calli were maintained on callus maintenance media for six months, some from each AIP treatment (cultured on media supplemented with 11.3 µM 2,4-D) were transferred to regeneration medium. There were no significant differences in regeneration characteristics after calli were cultured on regeneration medium for one month. Regenerants transferred to root- or shoot- inducing media developed healthy roots and/or shoots with no abnormal morphological traits (representations may be observed in Figure 4.6). Shoot multiplication rates ranged from six to 18 per explant (data not shown).

a) b)

Figure 4.6 Regenerated calli in rooting medium (a) and after culture in liquid tillering medium (b) after approximately 28 days of incubation.

4.4 Discussion

Regeneration of M. x giganteus through somatic embryogenesis and/or shoot organogenesis from immature inflorescences has previously been developed (Lewandowski 1997; Głowacka et al. 2010; Kim et al. 2010; Kim et al. 2012; Gubišová et al. 2013). While this approach offers opportunities for genetic improvement and large-scale plant propagation of this sterile triploid, several problems have been encountered which limits its application. Specifically, callus has been reported to lose its regenerative capacity within four months of culture (Kim et al. 2010), presenting issues in applying this technology. In the current study, AIP – a potent inhibitor of the phenylpropanoid pathway – was found to increase regeneration in this system over extended culture periods.

The modified F-C assay (Singleton and Rossi 1965) used in this study is a rapid, non-specific colorimetric technique which functions to quantify the amount of readily-oxidized phenolic substances (Singleton et al. 1999) within biological material. The results found in this study supported our hypothesis as soluble phenolic content was significantly affected by AIP levels, with the lowest values detected in tissue cultured on mediums with 1000 μM AIP, and the highest levels from calli incubated with 1 µM treatment (1569.71 and 2242.34 µg g-1 DW, respectively). GAE levels decreased in a dose-dependent manner (2141.47, 2091.86, and 1731.78 µg g-1 DW) between the remaining treatments (0, 10, and 100 µM, respectively); however, the differences among these

99 values were not statistically significant. The capacity for shoot regeneration of M. x giganteus calli was achieved after long term culture on semi-solid callus maintenance media of both 2,4-D concentrations tested. During the experiment, embryogenic calli (K2) could be maintained as well. The results presented here are an improvement from the findings reported by Kim et al. (2010), who experienced complete loss of shoot-regenerative and embryogenic-development capacity after four months of culture on semi-solid media. This system will allow for relatively long term, low-maintenance, effective callus culture system for M. x giganteus. In addition to these findings, we were able to demonstrate that regenerants from M. x giganteus calli were able to develop into plantlets on both tillering and rooting mediums, with no noticeable morphological abnormalities after long-term culture.

The frequency of different callus types and regenerative structures was highly variable among and within treatments, and the proportions changed over the duration of the experiment. This has previously been observed in other grasses, such as sugarcane (Fitch and Moore 1990). Regarding regenerant development, embryo-like structures were identified by a smooth, rounded architecture, which appeared either white or green in colour. These structures varied in shape, with clusters of embryo-like structures represented by a conglomerate of spherical (globular)-, cone (heart)-, and torpedo-shaped somatic embryos. While histological analysis to verify their bipolar structure was not conducted, previous studies using a similar system demonstrated that they exhibited a bipolar configuration with a discernable epidermis, characteristic of somatic embryos (Ślusarkiewicz- Jarzina et al. 2017).

Calli that developed roots were generally nodular and semi-soft in appearance and occurred alongside anthocyanin spots like what has been previously reported (Petersen 1997). These structures were glabrous and displayed a red/purple pigmentation where the calyptra would be located. Conversely, shoot-like structures (determined by the presence of green buds/primordia and light green, translucent leaves) mainly formed on compact white and friable callus. Leaves could be ascertained by the presence of trichomes observable on all parts of the tissue, including the leaf tip. The most pertinent data for achieving longer term culture was the frequency of calli developing shoot-like structures. Though trends for all treatments showed reductions in the percentage of regenerative calli over time, shoot regeneration was sustained for at least seven months after initiation of the experiment. Independent of auxin levels, all AIP treatments had the

100 capacity to develop shoot-like structures, and subsequently produced full plantlets. Though callus maintenance medium absent of AIP exhibited the highest amount of shooting calli in the first month of the experiment (66.7%), this treatment only produced between 10-30% for the remainder of the trial. 10 µM AIP yielded the greatest improvement for shoot induction (averaging as much as 58.3% by the third sampling time). By the end of the experiment, the percentage of shooting- calli remained at 40% in the presence of 100 µM AIP, significantly higher than all other treatments at this time (ranging from 6.8-18.3%).

Previous authors have found that supplementing 10 μM AIP in the medium of alfalfa cultures increased mitotic activity and prompted cellular division (Hrubcová et al. 2000). This has also been demonstrated in sessile oak cultures where increased total phenolic acid content was positively correlated with non-converting embryos and decreased phenolic content through application of AIP yielded embryos that more readily regenerated into plantlets (Cvikrová et al. 1998). Likewise, in the current study, over time AIP increased the capacity for regeneration in M. x giganteus, However, regeneration of calli at the highest level, 1000 μM AIP, was greatly hindered compared to calli cultured on 10 and 100 μM AIP. This was accompanied by a greater total MC in calli cultured with 1000 μM AIP (87.3%). Similar side effects caused by drastic inhibition of the phenylpropanoid biosynthetic pathway have been documented previously and include increased incidence of hyperhydricity caused by reduced lignification, and reduced lignin production for cell wall-formation and development of shoots (Cvikrová et al. 2003). This could be of particular significance for grass species as phenylpropanoids are major components of their cell walls and play important roles in cell growth/development and intercellular adhesion (Jones and Saxena 2013).

Another deleterious effect of high AIP levels is the hindrance of root formation and growth, as demonstrated in Secale cereale (Reuber et al. 1993). In contrast, average root regeneration peaked in calli incubated on 1000 μM AIP at the fourth sampling time (18.8) and continued to produce relatively high quantities into the final sampling month in this study (9.05); however, as expected, the frequency of calli with root-like structures was more strongly represented in samples cultures with 11.3 μM than 9.0 μM 2,4-D. As with shoot-like structures, root-like structures were induced for at least a year in culture, though investigation of root culture – with roots derived from calli – requires further examination in M. x giganteus.

101

It has been demonstrated in sugarcane that embryogenic calli contain over 50% less GAE concentrations than non-embryogenic calli (Neves et al. 2003). This finding may explain why a higher proportion of calli cultured on 10 and 100 μM AIP exhibited K2 formation than lower AIP treatments. Because of the high variability of callus morphologies between treatments and over time, no statistically significant trends could be ascertained.

4.5 Conclusions

The results presented here demonstrate that regeneration capacity of M. x giganteus can be maintained for relatively long-term culture on semi-solid callus maintenance media, and that regeneration can be enhanced with the addition of 10 and 100 μM AIP. AIP functions to inhibit the phenylpropanoid biosynthetic pathway, and has many roles in mediating phytohormone levels, cell wall development, and auxin metabolism (Lee 1980; Hrubcová et al. 2000; Cvikrová et al. 2003). Though soluble phenolic content decreased in a dose-dependent manner in regards to AIP, a more targeted approach at specific phenolic content changes, including cell wall-bound phenols, and accounting for potential interfering compounds, could contribute to better understanding the regeneration capacity for this species. The findings of this study will help to develop options for alternative propagation methods through somatic embryogenesis/synthetic seeds and help to develop a foundation for in vitro breeding strategies.

4.6 Acknowledgements

The authors wish to thank Abhishek Chattopadhyay for his training and assistance with determining soluble phenolic content of the samples, All Weather Farming Inc. for supplying explant material, and Dr. Michelle Edwards for aiding with statistical analyses. The authors are also grateful for our funding partner BioFuelNet for financially supporting this research. The funders had no role in the design of the study, data collection and analysis, decision to publish, or preparation of the manuscript.

102

5: In Vitro Induction and Encapsulation of Miscanthus x giganteus Anderss. Microrhizomes

Abstract

Miscanthus x giganteus Anderss. demonstrates favourable characteristics for cultivation as an advanced biofuel feedstock. However, because of its innate sterility, proliferation is commonly performed using costly vegetative methods, making field establishment cost prohibitive. Techniques for the induction of microrhizomes and production of synthetic seeds has not yet been established. The current study aimed to investigate microrhizome induction of in vitro plantlets using liquid Murashige and Skoog basal media supplemented with various concentrations of sucrose (3, 8, and 10%), BAP (2.5 or 26.5 μM), and 1-naphthaleneacetic acid (0.6 or 50 μM), as a potential alternative propagation system. Microrhizomes were successfully induced, with the most effective treatment on 8% sucrose, 26.5 μM 6-benzylaminopurine, and 0.6 μM 1- naphthaleneacetic acid, producing an average of 46.4±3.1193 microrhizomes weighing an average of 1.01±0.06674 g per plantlet after a 10-week incubation period. Microrhizomes from this treatment were harvested and kept either fully intact or sectioned into small (2-3.0 mm), medium (3-4.0 mm), or large (4.0-4.5 mm) fragments (with at least one bud affixed) for encapsulation in

3% sodium alginate and 1.0% CaCl2 solutions (w/v) for synthetic seed production. Samples were then tested for short-term CS (3°C) and growth in ex vitro and in vitro conditions. Additionally, full plantlets after 10 weeks of induction treatment were tested for CS with and without abscisic acid (3.8 μM) pre-treatment, and regrowth was assessed thereafter. Green bud development was greatest in microrhizomes isolated from plantlets stored at low temperatures (control and abscisic acid-treated) after four weeks in vitro (10.8-11.3). In vitro plantlet conversion was greatest in synthetic seeds produced using large explants (9.4±0.1805%) after eight weeks of culture. Limited ex vitro growth was observed from microrhizomes planted in the mist bed for a six-week period, and germination of synthetic seeds was not realized in these conditions due to fungal attack. Findings from this study will help in establishing a foundation for Miscanthus microrhizome development and synthetic seed production, allowing for alternative propagation and conservation.

5.1 Introduction

Miscanthus x giganteus Anderss. (M. x giganteus) is a C4-photosynthetic, fast-growing, lignocellulosic tallgrass belonging to the Poaceae family (Hodkinson et al. 2002). Its high yields and ability to grow in relatively cold climates (such as parts if northern Europe and Canada) make

103 it a potential bioenergy feedstock in these areas (Clifton-Brown and Lewandowski 2000; Clifton- Brown et al. 2001; Rosser 2012). This species’ rhizomatous and deep-rooting nature also allows it to be grown on land not suitable for food crop production (Heaton et al. 2004), helping to conserve nutrient-rich locations for human and livestock food utilization. When grown commercially for biomass production, M. x giganteus stands can be managed for 10-25 years (Xue et al. 2015) at a density of approximately 10,000 (Jeżowski et al. 2011) to 20,000 rhizomes ha-1, with shoot harvest occurring after the second or third year of establishment (Atkinson 2009).

M. x giganteus (3x=57) is a sterile triploid and putative hybrid between diploid M. sinensis (2x=38) and tetraploid M. sacchariflorus (4x=76) (Linde-Laursen 1993). As such, it is propagated through vegetative means, including DRP, rhizome-derived plugs, and micropropagated plantlets, with the bulk of establishment costs arising from propagules (Xue et al. 2015). According to Xue et al. (2015), seed-propagation of Miscanthus is cheaper than rhizome or micropropagated plants (€414.30, 2,400.00, and 5,120.00 ha-1, respectively), but this approach is limited to diploid M. sinensis and tetraploid M. sacchariflorus. Though fertile species are available for biomass cultivation, the challenges of these species include a higher chance of invasiveness in non-native environments (Boersma and Heaton 2014), geographical restrictions for flowering (Jensen et al. 2013), and potential loss of favourable traits and crop consistency due to outcrossing (Lewandowski et al. 2016). As such, M. x giganteus propagation is currently limited to rhizome cuttings (Atkinson 2009), collar pieces (Mangold et al. 2017), and micropropagation through shoot organogenesis or somatic embryogenesis (Lewandowski 1998).

An alternate approach that has not been reported for M. x giganteus is the development of in vitro MRs. The induction of MRs and MTs has been achieved in a variety of other species such as turmeric, mango ginger, ginger, Bambusa bambos var. gigantea (bamboo), and the most studied system, potato (Tognetti et al. 2013). The formation of MRs and MTs in these species is influenced by a variety of factors including nutrient concentration (Murashige and Skoog 1962), sucrose levels, exogenous cytokinin/auxin levels, GA3, duration of incubation, photoperiod, temperature, and cultivar/genotype (Garner and Blake 1989; Kapoor and Rao 2006; Cousins and Adelberg 2008; Zheng et al. 2008; Kolachevskaya et al. 2015). Based on the induction of these structures in other species, it is likely possible to induce MRs in M. x giganteus by manipulating these factors.

104

An additional avenue for M. x giganteus commercial propagation is the production of synthetic seeds, in which tissues are encapsulated in a matrix that enables them to be planted in a manner analogous to true seed. The fundamental goals for successful synthetic seed production are to protect the encapsulated tissue during handling and storage and allow for bud emergence when regrowth commences (Redenbaugh et al. 1986). In most synthetic seed processes, the tissue is encapsulated in calcium alginate (artificial endosperm), then encased in a matrix (artificial seed coat). The matrix can have nutrients, PGRs, and anti-microbial compounds added to it (Ghanbarali et al. 2016) to regulate dormancy/storage, germination, disease control, and plantlet vigor. While a form of synthetic seed for Miscanthus has been commercially developed under the trade name CEED™ (Crop, Expansion, Encapsulation and Delivery System), this is a proprietary product and the production methods or tissues used have not been published.

The objectives of this study were to: 1) establish a protocol for MR formation in M. x giganteus, 2) test MRs for their capacity to be used as propagules in vitro and ex vitro before and after CS treatment, and 3) observe growth parameters and CS capacity of MRs after being sectioned and encapsulated into alginate beads. The results of this study lay the foundation to further develop MRs as an alternate propagation strategy for this important biofuel crop.

5.2 Materials and methods 5.2.1 Culture initiation

Plantlets were initiated from regenerative calli as previously described (Kim et al. 2012). Plantlets were maintained in liquid tillering medium composed of full-strength MS basal salts with vitamins (Phytotechnology, Shawnee Mission, KS, USA), 30 g l-1 sucrose, 22.1 µM BAP, 5.14 µM IAA, and 0.984 µM IBA (Kim et al. 2012). pH was adjusted to 5.7 using 0.1 M NaOH and HCl before autoclaving at 121°C and 21 psi for 20 minutes. All PGRs were purchased from Sigma-Aldrich® (Sigma-Aldrich, St. Louis, MO, USA). Before autoclaving, tillering medium (approximately 20 ml) was dispensed in 125 ml Erlenmeyer flasks. Shoot clusters (two to five shoots) were subcultured approximately every four to five weeks, and two to four clusters were cultured in each flask. At each subculture timepoint, chlorotic and vitrified shoots, and atrophied tissue at the crown, were gently removed.

105

5.2.2 Microrhizome induction

Individual shoot clusters were used for MR development and optimization. The following treatments were investigated in full-strength MS basal liquid culture with:

3% sucrose (‘3%sucrose’);

3% sucrose supplemented with 26.5 µM BAP and 0.6 µM NAA (‘3%sucrose+PGR1’);

3% sucrose supplemented with 2.5 µM BAP and 50 µM NAA (‘3%sucrose+PGR2’);

8% sucrose (‘8%sucrose’);

8% sucrose with 26.5 µM BAP and 0.6 µM NAA (‘8%sucrose+PGR1’);

10% sucrose (‘10%sucrose’); and

10% sucrose with 26.5 µM BAP and 0.6 µM NAA (‘10%sucrose+PGR1’).

PGR concentrations (26.5 µM BAP and 0.6 µM NAA) were selected based on optimized MR induction of ginger (Zheng et al. 2008). An additional combination of PGRs (50 µM NAA and 2.5 µM BAP) was selected based on optimized MR development of bamboo (Kapoor and Rao 2006) when sucrose concentrations were set at a standard 3%. pH was adjusted to 5.8 and autoclaved as described previously. Before autoclaving, 20 ml of media was dispensed into 250 ml Erlenmeyer flasks. One shoot cluster was placed into each flask, and each treatment was replicated four times. Before MR initiation, shoot cluster FW and tiller number were recorded. Treatments were arranged in an RCBD and incubated for 10 weeks before analysis. Media were replaced five weeks into the experiment, and shoot clusters were recorded for FW and tiller number (after pruning of vitrified, chlorotic, and atrophied tissue). Change in shoot number and plantlet weight (%) was calculated between weeks 0 and 5, and weeks 5 and 10 of the incubation period. All treatments were incubated in a growth room at 24±2°C and 16-hour photoperiod (40 μmol m2 s−1) provided by cool-white fluorescent lamps (Philips Canada, Scarborough, ON).

5.2.3 Physiological assessment

Destructive and non-destructive measurements were conducted ten weeks after subculturing the plantlets. Immediately after plantlet removal, they were rinsed with deionized water and gently dried with paper towel. Chlorophyll measurements (fifteen per plantlet) were taken at random

106 locations from healthy leaves (leaves characterized by tissue extending from a ligule/collar) with an Opti-Sciences™ CCM-300 (mg m-2). Plantlets were then separated into individual tillers (with leaves attached), roots (if present), and MRs. The amounts of individual tillers (tillers characterized as shoots longer than 5 mm), percentage of chlorotic tillers (number of tillers with observed chlorosis of the stem/sheath divided by the total number of tillers per sample), and number of shoot nodes, roots, and MRs were recorded. Average lengths of tillers, roots, and MRs per sample were assessed with ImageJ® (ImageJ, U.S. National Institutes of Health, Bethesda, Maryland, USA), and MR node number was inspected using a dissecting microscope. Subsequently, total tiller, root, and MR FWs per sample were recorded before freezing in liquid nitrogen, and samples were stored at -80°C until lyophilization. Thereafter, DWs were discerned and ‘total MC (%)’, ‘total plantlet FW’, ‘MR relative FW (RFW)’, ‘shoot RFW’, and ‘root RFW’ were calculated:

[((FW - DW) / FW) x 100];

(FWshoot + FWroot + FWMR);

(FWMR / FWtotal plant);

(FWshoot / FWtotal plant); and

(FWroot / FWtotal plant), respectively.

Prior to each destructive measurement, atrophied tissues were gently removed.

5.2.4 Growth capacity

The treatment that yielded the most promising results for MR production (quantity, fresh and DWs, lengths, etc.) was used in subsequent experiments for determining in vitro and ex vitro bud development and shoot emergence. Twenty shoot clusters were transferred from tillering medium to MR induction medium ‘8%sucrose+PGR1’. Medium preparation, treatment initiation, incubation, and medium replacement were carried out as described previously, with treatments being arranged in a CRD. Ten weeks after incubation, four plantlets were randomly selected for each growth experiment. The first four plantlets were destined for destructive measurements as described in ‘5.2.3 Physiological assessment’.

107

5.2.5 Microrhizome isolation

Dissection of MRs was done in aseptic conditions. MRs were isolated by gently removing atrophied tissue surrounding the material, paying close attention to preserve rhizome tissue integrity and any green buds. After isolation, samples were washed thoroughly with sterile, deionized water, and dabbed dry with sterile Whatman® filter paper. Immediately after dissection of MRs from four replicate plantlets, MRs were placed into either: 1) tillering media (10 ml in 125 ml Erlenmeyer flasks) (MS basal medium supplemented with 8% sucrose, 26.5 µM BAP and 0.6 µM NAA) (‘8%sucrose+PGR’); 2) potting soil ex vitro; 3) petri dishes for CS (‘CS’) at 3°C (‘CS- C’); or 4) a flow bench for an hour of desiccation before CS at 3°C (‘CS-D’). Four replicates were used per experiment, and each replicate contained eight MRs.

5.2.6 In vitro growth

For in vitro tillering, four flasks were used per replicate (two MRs per flask), and samples were weighed prior to induction. Samples were arranged in an RCBD and kept in growth room conditions as described previously. Samples were checked for either green bud emergence (buds characterized as 5 mm or less) using a dissecting microscope, and sprouted tillers every four weeks for two months. Medium was replaced at four weeks after experiment initiation, and atrophied tissue was removed from samples before inspection at four and eight weeks after culture initiation. For samples with sprouted tillers at each of these timepoints, samples were aseptically removed, gently washed with sterile, deionized water, and dabbed dry with Whatman® filter paper. Samples were weighed and observed for plantlet conversion/germination rate (%), tiller number, tiller length (mm) and leaf number (of longest shoot), and average chlorophyll content (mg m-2) (fifteen readings per plant).

5.2.7 Ex vitro growth

Concurrently, ex vitro MR sprouting was observed in greenhouse conditions in a mist bed. Samples were planted approximately 5 mm deep in autoclaved Sunshine® Mix LA#4 (Sun-Gro Horticulture, Bellevie, WA, USA) potting soil. The experiment was arranged in an RCBD using a plastic 16 x 8 cell tray (each cell measuring 2.45 cm3). Each tray was arranged into four replicates (tray separated into quarters, lengthwise) and each replicate contained eight MRs (eight cells with one sample planted into each cell). Samples were fertilized with 20-8-20 (N-P-K) fertilizer solution

108

(Plant Products®, Leamington, Canada; 250 ppm N or 1.25 g l-1; pH adjusted to 6.0 with phosphoric acid) immediately after planting. Samples were observed until emergence (checked every three to four days). After six weeks, samples were measured for plantlet conversion/germination rate (%), tiller number, tiller height (mm) and leaf number (of tallest tiller), average chlorophyll content (fifteen readings per plant), root number, and full plant FW.

5.2.8 Microrhizome cold storage ability

The remaining MRs were assessed for CS ability. Briefly, the FW of 64 MRs were recorded (eight at a time); half of the samples were placed into petri dishes (100 x 15 mm; Fisher Scientific, Canada) and sealed with Parafilm® M, while the remaining samples were allowed to dry aseptically in a flow bench for one hour (‘CS-C’ and ‘CS-D’, respectively). The FW of these samples was recorded after drying and placed in petri dishes as described above. Petri dishes were stored in a chest freezer maintained at 3°C for four weeks. The experiment was arranged in a split- plot design, with two freezers acting as main plots, and petri dishes acting as sub-plots. MRs stored without drying and with drying were replicated four times each. After four weeks of storage, half of the samples from each dish were transferred to tillering medium, and the remaining samples were planted in the mist bed. Experiment initiation, observations, and observation time points were executed as previously described in ‘5.2.6 In vitro growth’ and ‘5.2.7 Ex vitro growth’.

5.2.9 Synthetic seed production

Four replicate plantlets were used for synthetic seed production. Aseptically-isolated MRs were cut into transverse sections (2-3 mm long) containing at least one immature bud (still presumed dormant) using a dissecting microscope. Care was taken not to crush the samples or tear the rhizome epidermis. Immediately after cutting, samples were transferred into autoclaved 3% sodium alginate (Sigma®) solution. Individual samples in sodium alginate solution were then pipetted with an autoclaved, 1000 µM micropipette tip with a small, medium, or large opening

(approximately 2.0-3.0, 3.0-4.0, and 4.0-4.5 mm, respectively) into autoclaved 1.0% CaCl2 (w/v) (Sigma-Aldrich, St. Louis, USA) solution. Samples were incubated for 20 to 25 minutes with occasional gentle stirring before being removed and washed thoroughly with sterile, deionized water (Ghanbarali et al. 2016). Synthetic seeds were then allowed to dry overnight in the flow bench before subsequent use (Janick et al. 1989).

109

Treatments were divided into ‘small’, ‘medium’, and ‘large’ artificial seeds. Before subjecting artificial seeds to in vitro or ex vitro growth conditions, samples were soaked in sterile, deionized water for approximately six hours. A third of the samples were used directly for in vitro growth in tillering medium; each treatment consisted of four replicate flasks, with each flask containing two samples. Treatments were arranged in an RCBD in a growth room for two months as described in ‘In vitro growth’. Medium was replaced four weeks after culture initiation and growth characteristics were recorded as illustrated previously in ‘5.2.6 In vitro growth’. “Bud emergence” was characterized by buds emerging from the matrix wall (Ghanbarali et al. 2016). Another third of the samples were planted ex vitro in the mist bed as outlined in ‘5.2.7 Ex vitro growth’, with each replicate containing six samples; two of each treatment. The remaining samples were destined for CS treatment as described in ‘5.2.8 Microrhizome cold storage ability’. Each treatment consisted of four replicate dishes containing four samples. Subsequent growth treatments included ‘CS-small’, ‘CS-medium’, and ‘CS-large’. After four weeks of storage, half of the samples were moved to in vitro growth, and the remaining were planted in the mist bed.

5.2.10 Full plantlet cold storage capacity

The remaining four plantlets were tested for CS capacity, and if prior incubation with a dormancy- inducing PGR – ABA – would affect subsequent MR growth in vitro. Briefly, MR induction medium was removed with a sterile pipette. Full plantlets were aseptically cut in half and weighed, and half of these samples were then placed into sterile 250 ml flasks (eight samples total). Four of the samples were placed into 3°C chest freezers (two samples per freezer), while the rest were incubated with filter-sterilized 0.5 mg l-1 (3.2 µM) ABA, pH 6.0 (Janick et al. 1989; Kendall et al. 1993; Pospíšilová et al. 1998; Zenkteler and Bagniewska-Zadworna 2005) for 20 minutes (‘CS- control’ and ‘CS-ABA’, respectfully). The ABA solution was then removed with a sterile pipette and plantlets were arranged in chest freezers as described above.

After four weeks, plantlets were removed from CS and FW was recorded. Each treatment was then aseptically cut in half, with one half destructively measured as outlined in ‘5.2.3 Physiological assessment’, while the remaining samples were put towards in vitro MR growth, as described in ‘5.2.6 In vitro growth’.

110

5.2.11 Experimental design and statistical analysis

Treatments and means comparisons were subject to variance analysis using a mixed-model analysis of variance (ANOVA) in SAS® 9.4 (SAS Institute Inc., Cary, NC, USA). Experiments were arranged in either an RCBD or split-plot fashion. All analyses were subject to repeated measures analysis, and each treatment was replicated four times.

In order to normalize variances not following a Gaussian distribution, a variety of distribution and link functions were applied for each dependent variable. Multiple-means comparisons were generated for all analyses using Tukey’s HSD test. Regression equations, Pearson’s correlation coefficients, and significance were determined using PROC NLIN for plantlet weight assessment over the MR induction period.

A type I and III error rate of α=0.05 was assigned for all analyses. The data presented in the results are from the original data scale.

5.3 Results 5.3.1 Shoot number

Shoot number and plantlet weight for plants exposed to various induction media at weeks 0, 5, and 10 (individual weights of plant parts ‘shoots’, ‘roots’, and ‘MRs’ were assessed instead of ‘intact plantlet’ weight at week 10 of treatment) are presented in Figure 5.1. Percentage change in shoot number and weight between these sample dates were also calculated (refer to Table 5.1).

There were no significant differences among treatments in shoot quantity and plantlet weight at week 0 of the experiment; however, there were significant trial effects. The trends observed between treatments in both trials were similar, supporting near-uniformity among the samples used in each trial and consistent effects of the treatments on plantlet growth. Because of this observation, only one set of data is presented.

Over time, significant differences in both variables were observed among treatments. At week 5, shoot quantity was greatest in samples with PGRs added to the media, especially in 8%sucrose+PGR1 and 3%sucrose+PGR2 (17 and 14.4, respectively). The fewest shoots were observed in treatments void of PGRs (7.8-8). By week 10, there were significant differences between 8%sucrose+PGR1 and 3%sucrose+PGR2 (49.5 and 57.5, respectively) and the remainder

111 of the treatments (ranging from 10.2 to 31). Shoot number with plantlets subjected to 10%sucrose alone were not statistically different from 10%sucrose+PGR1 and 3%sucrose+PGR1, and 3%sucrose treatment yielded the fewest tillers. Significant treatment and trial effects were observed for shoot number at five weeks, and significant treatment, trial, and treatment x trial effects were found for shoot number at 10 weeks.

Significant treatment effects were observed for percentage shoot change between both weeks 0 and 5, and weeks 5 and 10. 3%sucrose+PGR2 resulted in the greatest increase between weeks 0 and 5, followed by 8%sucrose+PGR1 and 3%sucrose+PGR1 (228.1, 159.1, and 128.4%, respectively). From weeks 5 to 10, 8%sucrose demonstrated the greatest shoot number increase (249.7%). In both measurement periods, treatments void of PGRs exhibited the smallest shoot number increases (reaching as low at 20.3% in 3%sucrose).

5.3.2 Plantlet weight

Full plantlet weight at week 5 did not follow the same pattern as shoot number, though this variable was significantly influenced by both treatment and trial effects. 10%sucrose produced the heaviest samples, followed by 8%sucrose, 8%sucrose+PGR1, and 3%sucrose+PGR2; 3%sucrose resulted in the lightest samples. Plantlet weights at week 10 were significantly influenced by treatment, trial, and treatment x trial interactions. By week 10, 8%sucrose+PGR1 and 3%sucrose+PGR2 accumulated the most biomass (4.04 and 4.02 g, respectively); 3%sucrose, 8%sucrose, and 10%sucrose accumulated the least (2.36, 2.16, and 2.41 g, respectively).

Regression analysis of plantlet weight over the 10-week period revealed trial, time, and treatment x time interaction effects. Treatment weight trends can be found in Figures 5.2 and 5.3.

There were no significant differences in percentage weight change between weeks 0 and 5 (increases ranged from 288.1% in 3%sucrose to 569.6% in 10%sucrose+PGR1). Weight change between weeks 5 to 10 displayed significant escalation in 8%sucrose+PGR1, 3%sucrose+PGR2, and 3%sucrose+PGR1 (10.9, 10.7, and 4.4%, respectively), and reductions in the remaining treatments (10%sucrose demonstrated a weight reduction of 37.9%).

112

Figure 5.1 Total shoot number (#) of MR induction treatments over the duration of 10 weeks. Each data point represents the mean of 30 samples (six treatments consisting of four replicates each), repeated twice. Because plants were assessed non-destructively at weeks 0 and 5, conservative counts were made.

113

Figure 5.2 Average weight (g) of full plantlets used for MR induction over the duration of 10 weeks. Each data point represents the mean of four replicate plantlets, repeated twice. The analysis was performed using a gaussian distribution and predicted values on the data scale were plotted 2 using a proc nlin polynomial model predictor: Y3%suc = -0.0397(time) + 0.5519(time) + 0.6858, 2 2 2 pseudo R = 0.69; Y8%suc = -0.0871(time) + 1.0123(time) + 0.7441, pseudo R = 0.35; Y10%suc = - 2 2 2 0.1539(time) + 1.6983(time) + 0.8107, pseudo R = 0.46; Y8%suc+PGR = -0.0527(time) + 2 2 0.8606(time) + 0.7031, pseudo R = 0.15; Y10%suc+PGR = -0.0407(time) + 0.6195(time) + 0.6781, 2 2 2 pseudo R = 0.59; Y3%suc+PGR1 = -0.0462(time) + 0.6537(time) + 0.6804, pseudo R = 0.47; and 2 2 Y3%suc+PGR2 = -0.0567(time) + 0.9036(time) + 0.6478, pseudo R = 0.20. Because plants were assessed non-destructively at weeks 0 and 5, conservative counts were made.

114

2 Ytrial 1 = -0.0831(time) + 1.0798(time) + 0.7688

Pseudo R2 = 0.41 a

ab bc 1 c

2 Ytrial 2 = -0.0531(time) + 0.7202(time) + 0.6455

Pseudo R2=0.57

d d

Figure 5.3 Average plantlet weight (g) of MR induction treatments over the duration of 10 weeks. Each data point represents the mean of 28 samples (seven treatments consisting of four replicates each), and both trials are exhibited. The analysis was performed using a reciprocal gamma distribution model and predicted values on the data scale were plotted using a proc nlin quadratic 2 2 model predictor: Ytrial 1 = -0.0831(time) + 1.0798(time) + 0.7688, pseudo R = 0.41; Ytrial 2 = - 0.0531(time)2 + 0.7202(time) + 0.6455, pseudo R2=0.57. Means followed by the same letter at each timepoint are not significantly different (P<0.05) according to Tukey's HSD mean separation test.

5.3.3 Physiological assessment

MR quantities and characteristics harvested from induction treatments are presented in Table 5.2, and representative rhizome samples extracted from in vitro and in vivo plants are displayed in Figure 5.4. MR quantity per plantlet was greatest in 8%sucrose+PGR1 (46.4), followed by 3%sucrose+PGR2 (37.6), and the fewest MRs were harvested from 3%sucrose, 8%sucrose, and 10%sucrose (12.4, 12.8, and 14.3, respectively). Average MR node number (2.7-4.3) and lengths (5.3-7 mm) remained relatively consistent between treatments.

115

Treatment and treatment x trial effects were observed with average MR FW yield per plantlet, while average MR DW yield per plant was influenced by separate treatment and trial factors. Greatest MR FW yield was recorded in 8%sucrose+PGR1 (1.01 g) followed by 10%sucrose+PGR (0.68 g). Treatments without supplemented PGRs experienced less yield (0.21-0.25 g). Regarding total MR DW yield, the greatest amounts were observed in 8%sucrose+PGR1 and 10%sucrose+PGR1 (0.27 and 0.21 g, respectively), and 3%sucrose and 8%sucrose yielded the lightest DWs (0.039 and 0.066 g, respectively). No significant differences were found in average individual MR weights between treatments (data not shown). When total MR MC was calculated, 3% sucrose samples with and without PGRs demonstrated the greatest values (81.4-83.6%) compared to the remaining treatments (outlined in Table 5.3).

Table 5.2 Total MR number (#), average MR node number (#), average MR length (mm), average MR FW yield per plantlet (g), and average MR DW yield per plantlet (g) after 10 weeks of incubation in MR induction media. Each value represents the mean of four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. Total MR Average MR Average MR Average MR FW Average MR DW number per node number length (mm) yield per plantlet yield per plantlet plantlet (#) (#) (g) (g) 3%sucrose 12.4c 3.4a 6.2 ± 0.3463a 0.21 ± 0.02814c 0.039d 8%sucrose 12.8c 4.3a 6.4 ± 0.2288a 0.24 ± 0.02814c 0.066cd 8%sucrose+PGR1 46.4a 4.3a 7 ± 0.0.3018a 1.01 ± 0.06674a 0.27a 10%sucrose 14.3c 4.2a 6.3 ± 0.02862a 0.25 ± 0.02814c 0.1c 10%sucrose+PGR1 25.8bc 5.4a 6.2 ± 0.2117a 0.68 ± 0.09303ab 0.21b 3%sucrose+PGR1 21.5c 2.7a 5.3 ± 0.8643a 0.47 ± 0.09942bc 0.096c 3%sucrose+PGR2 37.6ab 3.1a 5.9 ± 0.3439a 0.55 ± 0.03606b 0.092c S.E. 3.1193 0.7663 - - 0.00965

116

A) B)

C) D)

E) F)

Figure 5.4 Stages of MR isolation and regrowth. MR isolation from plantlets incubated in MR induction media after 10 weeks (A); microscopic images of MRs with immature buds (B) and young tiller (C); plantlet growth after 8 weeks of incubation of isolated MRs in tillering medium (D); field-harvested rhizomes after pruning, washing, and removal of most roots (E); and cross- section of processed field rhizome (F).

117

There were significant treatment and treatment x trial effects for total shoot FW, and treatment and trial effects for total shoot DW. Total shoot FW was greatest in samples subjected to 10 weeks of 3%sucrose+PGR2, ensued by 8%sucrose+PGR1 (3.14 and 3.03 g, respectively). Treatments with the lowest shoot FWs were 3%, 8%, and 10% sucrose void of PGRs (1.92, 1.45 and 1.62 g, respectively). Total shoot DW was greatest in 8%sucrose+PGR1 (0.71 g) and 10%sucrose+PGR1 (0.60 g). The lowest values were observed in samples incubated in 3%sucrose, 8%sucrose, and 3%sucrose+PGR1 (0.31, 0.33, and 0.30 g, respectively). Total shoot MC was significantly influenced by treatment effects and was most prevalent in 3% sucrose with and without PGRs (83.6-87.8%) compared to the remaining treatments (70.6-76%).

Table 5.3 Total shoot, root, and MR MC (%) after 10 weeks of incubation in MR induction media. Each mean represents four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. MC (%) Shoot Root MR 3%sucrose 83.6±0.3176b 73.9±4.789a 83.6a 8%sucrose 75.9±1.555c 56.7±1.493b 66.4c 8%sucrose+PGR1 75.8±0.5123c 0 72.3b 10%sucrose 74.3±1.529cd 49.2±0.2975c 66.6c 10%sucrose+PGR1 70.6±0.4554d 0 66.3c 3%sucrose+PGR1 85.6±0.9202ab 38±0.3332d 81.4a 3%sucrose+PGR2 87.9±0.2993a 0 83.8a S.E. - - 1.2064

Significant differences were found among treatments regarding average shoot lengths and average chlorophyll content, while no significant treatment differences were observed in percentage shoot chlorosis. Generally, treatments without PGRs (3, 8, and 10% sucrose) exhibited longer shoots (107.4-120.5 mm) than the remaining treatments (54.2-70.7 mm). Average chlorophyll content was highest in 8%sucrose (406.83 mg m-2) and lowest in 8%sucrose+PGR1 (347.27 mg m-2), and there were no significant differences between the other treatments (310.4-377.99 mg m-2) (details provided in Table 4.4).

5.3.4 Root formation and shoot node development

Shoot nodes only developed in plantlets incubated in 3%sucrose, 8%sucrose, 10%sucrose, and 3%sucrose+PGR1 treatments. These treatments also yielded white, visually healthy roots, with the exception of 3%sucrose+PGR1, which only produced roots during the second trial. Average shoot node number was greatest in 3%sucrose and 10%sucrose (1.2 and 1.4, respectively), and total root

118 number was significantly affected by treatment and treatment x trial effects, with the most roots being observed in 8%sucrose+PGR1 (59.8) (represented in Figure 5.5). Average root length was impacted mainly by treatment effects. The shortest roots were produced in 3%sucrose+PGR1 (24.3 mm during trial 2) and the longest in 8%sucrose and 10%sucrose plantlets (65-67.9 mm) (detailed in Table 5.4).

The greatest total root FWs and DWs were observed in 8 and 10%sucrose treatments (0.47-0.55 and 0.21-0.28 g, respectively). Total root MC was influenced by treatment, trial, and treatment x trial factors. These values were greatest in 3%sucrose (73.9%) and lowest in 10%sucrose (49.2%) (see Table 5.3).

119

Table 5.4 Average root length (mm), total root number (#), average shoot length (mm), average shoot node number (#), average shoot chlorosis (%), and average chlorophyll content (mg m-2) after 10 weeks of incubation in MR induction media. Each mean is represented by four replicate plantlets, repeated twice. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. Average Total Average shoot Average Average Average root length root length (mm) shoot node shoot chlorophyll (mm) number number (#) chlorosis content (mg g-2) (#) (%) 3%sucrose 0 ± 45.4b 32.3b 114.4 ± 5.217a 1.2ab 373.37ab 0.03757a 8%sucrose 0.35 ± 65a 59.8a 120.5 ± 9.3244a 0.3b 406.83a 0.06257a 8%sucrose+PGR1 0.32 ± 0c 0c 64 ± 6.8663ab 0c 347.27b 0.2422a 10%sucrose 0.8 ± 67.9a 53.9ab 107.4 ± 7.0977a 1.4a 366.04ab 0.6017a 10%sucrose+PGR1 0.34 ± 0c 0c 54.2 ± 6.2458b 0c 310.4ab 0.2369a 3%sucrose+PGR1 70.7 ± 0.24 ± 0c 0c 0.07b 342.72ab 21.0882ab 0.06135a 3%sucrose+PGR2 0.13 ± 0c 0c 57.5 ± 8.077b 0c 377.99ab 0.06294a S.E. 5.3532 6.4975 - 0.3552 - 28.2009

5.3.5 Relative fresh weights

RFWs for each tissue type (shoots, roots, and MRs) were calculated for each treatment at the end of the 10-week incubation period (see Figure 5.6), and treatment effects were demonstrated for all measurements. MR RFW was greatest in 10%sucrose+PGR1 (0.34) and 8%sucrose+PGR1 (0.26), and lowest for treatments not supplemented with PGRs (0.092-0.11). Shoot RFW was greatest in 3%sucrose+PGR2 (0.87) and 3%sucrose (0.81). 8 and 10% sucrose samples had root RFWs of 0.23-0.24, while 3%sucrose and 3%sucrose+PGR1 exhibited values of 0.096 and 0.033, respectively.

120

Figure 5.6 RFW of shoot, root, and MR tissue after 10 weeks of incubation in MR induction treatments. Each bar represents the mean of four replicate plantlets, repeated twice. Means of each tissue followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test.

5.3.6 Microrhizome growth capacity

MRs harvested and used to assess CS displayed statistically-similar initial average weights (0.035- 0.04 g). After one-hour drying of treatment (CS-D), significant weight differences were observed (0.011 g in CS-D compared to 0.035-0.04 g for CS-C and 8%sucrose+PGR). Moreover, significant differences in weight also occurred in CS-C and CS-D after CS treatment (0.015 and 0.0076 g, respectively), and these values were considerably different from each other and 8%sucrose+PGR treatment. When CS MR treatments were compared without 8%sucrose+PGR, the percentage of weight change from before to after CS was far greater in CS-C (-56.6%) than CS-D (-29.5%). When average weights were assessed over all MRs before induction into tillering media (8%sucrose+PGR, CS-C, CS-D, CS-control, and CS-ABA), significant treatment differences were observed; CS-D had the lightest MRs (0.0072 g), followed by CS-C (0.014 g). All other treatments did not demonstrate significant weight differences from each other (0.029-0.036 g).

121

The total number of green buds formed on MRs in tillering medium was recorded at 4 and 8 weeks after culture initiation (details presented in Table 5.5). At week 4, MRs from CS-control and CS- ABA treatments developed the greenest buds (12 and 11.3, respectively). 8%sucrose+PGR developed 10.1 green buds, while CS-C produced approximately half of the CS-control and CS- ABA treatments (6.4). At week 8, the number of green buds decreased; however, no significant differences were observed among treatments 8%sucrose+PGR, CS-control, CS-ABA, and CS-C (7.5, 7.6, 6.3, and 2.6, respectively). The percentage change in total bud development from weeks 4 to 8 was significantly different in CS-ABA (-86.4±19.7882%), and no statistical differences occurred between the remaining treatments (-33.5 to -72%). CS-D treatment did not produce any viable buds at 4 and 8 weeks of culture and was excluded from the analysis.

Table 5.5 (A) Total bud number at weeks 4 and 8 (#), and change in bud number (%) over the duration of the incubation period of all MR treatments in tillering medium. Each mean represents 32 MRs, repeated twice. Analysis was conducted using an RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. (B) Total bud number at weeks 4 and 8 (#), and change in bud number (%) over the duration of the incubation period of MR treatments in tillering medium after CS at 3°C. Each mean is represented by 32 MRs, repeated twice. Analysis was conducted using a split-plot design, with freezer representing the main plot and either plates (CS-C and CS-D) or flasks (CS-control and CS-ABA) representing subplots. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. (A) 8%sucrose+PGR CS-C CS-D CS-control CS-ABA S.E. Week 4 (#) 10.1ab 6.4b 0a 12a 11.3a 1.0926 Week 8 (#) 7.5±1.0609a 2.6±0.5956a 0a 7.6±1.0966a 6.3±0.9557a - % change -13.7ab -55.1c 0a -38.4bc -42.6c 6.8229

(B) 8%sucrose+PGR CS-C CS-D CS-control CS-ABA S.E. Week 4 (#) - 6±0.7716b 0a 11.3±1.448a 10.8±1.3957ab - Week 8 (#) - 1.6b 0a 7.3a 6.2a 1.126 % change - -63.6c 0a -37.3b -42.4b 4.7011

Few in vitro and ex vitro MRs germinated and produced plantlets after 8 and 6 weeks of incubation, respectively (see Appendix 5.5 for in vitro data, Appendix 4.6 for ex vitro data, and Figure 5.7 for representative ex vitro plantlet formation).

122

123

5.3.7 Synthetic seed growth capacity

Synthetic seed physical representations are depicted in Figure 5.8. The total number of buds developed from synthetic seeds after 4 and 8 weeks in tillering medium was only assessed in one trial as no buds developed in the second trial. The number of buds ranged between 1 and 3.3 at week 4, and between 0.5 and 1.5 at week 8 of culture. No significant differences resulted from explant size or CS treatment; however, the most buds developed from large and CS-large samples (3.3 and 2.7 at 4 weeks of culture, respectively). Unlike intact MRs, in vitro germination occurred in almost all synthetic seed treatments; however, ex vitro samples failed to germinate. Percentage germination was observed mainly from large explants without (9.4%) and with (6.3%) CS treatment. No germination was observed with CS-small samples. All other treatments displayed no significant differences (3.1%).

Regarding plantlet physiology, no significant treatment differences were found with leaf number (of tallest tiller) (0.3-2.8) or average chlorophyll content (300.37-441.4 mg m-2). Tiller number was greatest from medium-sized explants (20) compared to remaining samples (1-7), and tiller length was highest in large explant samples (89.8 compared to 3.6-16.1 mm). When entire plantlet weight was considered, large explants with (1.28 g) and without (1.85 g) CS yielded the most biomass.

124

A) B) C)

D) E) F) Figure 5.8 Stages of synthetic seed production. Healthy, green buds were assessed on isolated MRs using a dissecting microscope (A), and MRs were cut into small sections containing at least one bud. Sections were then encapsulated in 3% (w/v) sodium alginate and 1.0 % (w/v) calcium chloride dihydrate solution (B). Samples were then allowed to dry in a flow bench overnight (C). Early (D) and late (E) bud emergence from the matrix after 4 and 8 weeks in tillering medium were observed using a dissecting microscope. Full plantlet development from synthetic seeds were observed after 8 weeks in tillering media (F).

5.3.8 Cold storage – full plantlets

After CS treatment, full plantlets were assessed for various physiological variables. Significant differences were found between both CS treatments (control and ABA) and 8%sucrose+PGR1 regarding average shoot length (41.4, 41.3, and 56 mm, respectively), average chlorophyll content (263.07, 265.73, and 365.85 mg m-2, respectively), and percentage shoot chlorosis (0.095% in 8%control+PGR1). Despite these variations, no significant differences were found in total MR quantity, average MR node number, average MR length, and shoot and MR FW (data not shown).

5.4 Discussion

Exploiting Miscanthus as a biofuel crop large-scale has been hampered by the high costs of field establishment, and these costs are primarily associated with the propagation of this seedless triploid species. Current approaches include rhizome cuttings, crown divisions, and micropropagation through shoot organogenesis and somatic embryogenesis. This study represents the first report of an alternative propagation system using the in vitro induction of MRs.

125

Cytokinins function to promote cell division and lateral bud growth while auxins function to organize meristems, maintain apical dominance, and promote root formation (Gaspar et al. 1996). However, in these experiments treatments 8%sucrose+PGR1 (high cytokinin:auxin) and 3%sucrose+PGR2 (low cytokinin:auxin ratio) both resulted in prolific shoot proliferation. Though NAA generally induces root development and helps in the initial stages of tuber/rhizome formation, increased shoot production in these samples could be an indirect effect of MR development. This observation is in agreement with Kapoor and Rao (2006), who reported that a high NAA:BAP ratio in MR induction media caused premature MR formation, and that culm growth from MRs (instead of from axillary shoot nodes) occurred when BAP concentration was increased.

Plantlets subjected to different treatments displayed characteristics similar to distinct lifecycle stages in in situ perennial grasses. The broad lifecycle of perennial forage grasses are as follows: 1) germination; 2) vegetative-leaf development; 3) elongation-stem elongation; 4) reproductive- floral development; and 5) seed development and ripening (Moore et al. 1991). At week 0 of the MR induction period, visual evaluation implied that all plantlets were in the vegetative-leaf development stage. By week 10, plants cultured on high sucrose (8% and 10%) without PGRs, or on 3% sucrose with and without 26.5 µM BAP and 0.6 µM NAA, had completed stage 2 and entered stage 3 of development (elongation-stem elongation). These were the only treatments that developed shoot nodes, which typically occurs during the ‘elongation’ stage of growth, and ends at the ‘boot’ stage (upper leaf sheath enclosing immature inflorescence) (Moore et al. 1991; Tejera and Heaton 2017). Likewise, rooting mainly occurred in media supplemented only with sucrose (3, 8, and 10%), and with an intermediate sucrose level (3%) supplemented with a high cytokinin:auxin ratio. Previous studies have found that increased sucrose concentrations can aid in adventitious root induction (Gibson 2005). In Asparagus officinalis L. (asparagus), the percentage of rooted minicrowns increased from under 20% to over 90% when incubated in media supplemented with 2 and 6% sucrose, respectively (Conner and Falloon 1993). The same study also found that the addition of other sugars in equimolar amounts to sucrose (ie. mannitol, raffinose, mannose, etc.) induced a similar response, suggesting a role for osmotic factors. However, these roots were stunted, thin, and visually unhealthy, implying the nutrient-providing effects of sucrose alongside its osmotic characteristics.

126

High sucrose media can also reduce shoot vitrification by regulating the osmotic balance in plant tissues (Gibson 2005). In the present study, the role of sucrose as an osmotic regulator was evident, since shoots harvested from 3% sucrose treatments (3%sucrose, 3%sucrose+PGR1, and 3%sucrose+PGR2) resulted in greater shoot, root, and MR MCs, and shoot RFWs compared to samples incubated in high-sucrose media (with and without PGRs). Conversely, MR RFW was greatest in high sucrose media with PGRs, and lowest in sucrose treatments without PGRs, though these differences were not statistically different. Culturing on 3% sucrose treatments with PGRs resulted in intermediate MR RDWs, suggesting a role for exogenous cytokinin and auxin application (alongside sucrose concentration) for sufficient MR formation.

Cousins and Adelberg (2008) observed that MR dry matter continued to accumulate while leaves and roots exhibited atrophy in turmeric, supporting the theory of remobilization of nutrients from aerial tissues to storage organs (Purdy et al. 2015). Shoot chlorosis was assessed in samples to determine if senescence of aerial tissues was correlated to increased MR biomass. Senescence is a controlled process where the chlorophyll in aerial structures is degraded and nutrients released from the organelles are remobilized to sink tissues (ie. below-ground storage organs) for dormancy and overwintering (Himelblau and Amasino 2001). This process also plays an important role in carbohydrate dynamics in plants and occurs in situ before killing frost in some Miscanthus genotypes (Rosser 2012; Purdy et al. 2015). In this study, differences in shoot chlorosis was not observed among treatments; however, this may be due lifecycle stage (senescence generally occurs after flowering in Miscanthus) (Clifton-Brown et al. 2001). Incubation of samples for a longer period of time, or introduction of amended temperature and/or photoperiod, may help induce senescence and determine if this can influence MR yields.

Shirgurkar et al. (2001) demonstrated that shoot size affected turmeric MR weight, with “large” shoots (15 cm length) producing the heaviest MRs (1.06 g). In this study, plantlets with longer tillers formed lighter MRs, but treatments with greater shoot FW and shoot number were associated with greater MR FW yield and MR number.

Based on overall plantlet growth and MR quantity and quality, 8%sucrose+PGR1 was selected for further MR assessment. Investigation of MR regrowth and evaluating the feasibility to use them as synthetic seeds was conducted. Examination of CS capacity, in vitro growth, and ex vitro growth was done to evaluate this as an alternative propagation system. MRs that were subjected to

127 desiccation before CS (CS-D) did not develop viable buds in culture or ex vitro; however, browning and atrophy also did not occur in these samples, while it was evident in the controls. In agreement with Cousins and Adelberg (2008), tissue atrophy may be a necessary process for allocating valuable resources to storage organs during formation or regrowth stages.

While MRs were successfully induced and appeared healthy, bud development and plantlet conversion of MRs were relatively low in this study. This may have been due to a number of factors including endodormancy (Leclerc et al. 1995), inconsistent bud development and inability to synchronize stages of bud formation (Seran 2013), tissue damage during harvest, and rhizome maturity. During MR harvest, samples varied in regard to colour (green to beige) and sturdiness; however, these traits were not formally evaluated.

There have been contradictory findings on the dormancy of in vitro potato MTs. Some reports found that in vitro MTs lack a dormant period and sprout prematurely in culture, while other studies have found that they do exhibit dormancy lasting anywhere from one to seven months (Leclerc et al. 1995). Leclerc et al. (1995) concluded that in vitro potato MTs have dormancy periods that are positively correlated with the patterns of dormancy observed in each cultivars’ in vivo conditions. For example, cv. ‘Kennebec’ exhibited an in vivo dormant period of approximately 6 weeks, while its in vitro dormant period lasted anywhere from 12.1 to 15 weeks (>250 and <250 mg MTs, respectively). Moreover, the researchers’ demonstrated that MT size (under and over 250 mg FW) and incubation length on induction media significantly affected dormancy length, but PGRs in induction media beforehand did not significantly affect MT dormancy. Though dormant states were observed in potato MTs, a separate study found that turmeric germination (78.2%) in shoot multiplication media was independent of MR size and weight (Shirgurkar et al. 2001).

The best plantlet conversion was observed in MRs harvested from whole plants following a 4- week cold treatment (with or without ABA). Full plantlets stored in the cold lost considerably less weight (-4.9 to -13.3%) than isolated MRs at the same temperature (-29.5 to -56.6%), suggesting that the MRs lost significant MC when exposed to relatively cold temperatures without an insulation matrix. Further, average shoot length and chlorophyll content was lower in CS-control and CS-ABA plantlets (41.3-41.4 mm and 263.07-265.73 mg m-2, respectively) after CS than in plantlets without CS treatment (56 mm and 365.85 mg m-2, respectively). This suggests that full

128 plantlets subjected to 3°C conditions were prompted to reallocate photosynthate and nutrients from aerial tissues to the developing MRs similar to in situ dormancy induction (Beale and Long 1997). In field conditions, daily minimal temperatures strongly influenced aerial tissue senescence, and senescence was correlated to reduced leaf chlorophyll content (Purdy et al. 2015). Conversely, the same study also observed that rapid cooling at night negatively influenced leaf starch metabolism, which may be detrimental to the allocation of soluble carbohydrates to rhizomes. Future research may consider subjecting in vitro plantlets to a wider range of storage temperatures and photoperiods to improve MR formation and conversion rates.

Plantlet conversion was low for MRs cultured in vitro and planted ex vitro. The greatest conversion was observed from MRs produced by plants stored at 3°C with ABA treatment and subsequently cultured in tillering medium (37.5%). For ex vitro analysis, plantlet conversion was only observed from the control (8%sucrose+PGR). Obstacles in plantlet conversion could be attributed to endodormancy, MC, tissue integrity, pathogen attack, and planting substrates. Optimizing growth potential from MRs should be investigated using pretreatments (eg. GA3, cytokinins, fungicides, commercial biostimulants, etc.), different pre-storage conditions (temperature, photoperiod, controlled relative humidity), and amended mediums/soil types (ie. plantlet conversion was achieved in ginger MRs planted in moist sand) (Sharma and Singh 1995).

The current study used 3% sodium alginate solution in combination with 1.0% CaCl2 for the production of encapsulated MR beads. These concentrations combined with a 20-25 incubation time in CaCl2 were selected based on synthetic seed production in other systems (Gantait et al. 2015). Encapsulated MRs that remained viable after CS and/or in vitro culture remained green/beige in colour, while non-viable samples generally turned brown and atrophied, similar to what has been observed in encapsulated carrot somatic embryos (Janick et al. 1989). Large explants (4-4.5 mm) with and without CS treatment exhibited higher rates of germination than the other treatments, with CS-small demonstrating no germination/plantlet conversion. Larger explants also performed better in encapsulated potato mini-tuber sprout (PMS) samples

(Ghanbarali et al. 2016). Larger explants encapsulated in 3% sodium alginate and 1-1.5% CaCl2 matrix culture medium (consisting of MS salts, 5 mg l-1 BAP, 10 mg l-1 NAA, and 300 mg l-1 activated charcoal, pH 5.7) was sufficient for adequate regrowth speed and rate after four weeks

129 of culture (0.76 and 61%, respectively) (Ghanbarali et al. 2016). In the same study, larger explants regrew at a rate between 39 and 89%, while smaller explant rate peaked at 39%.

Depending on the explant source and species, desiccation of synthetic seeds can have advantages over hydrated samples, such as reversible quintessence leading to longer periods of storage (Nieves et al. 2001). The present study dehydrated samples in a flow bench overnight before either subjecting them to CS in petri dishes without substrate, or submerging them in sterile, deionized water for six hours before planting them in vitro and ex vitro. Carrot somatic embryos (both encapsulated and non-encapsulated) exhibited greater survival at high humidity, suggesting that drying rate during the desiccation process is crucial for viability (Janick et al. 1989). This was also reported in ginger microshoots encapsulated in alginate solution, where samples that were dried in a flow bench for even one hour (total MC of 88.3%) experienced a plantlet conversion percentage of 66.7%, compared to 93.3% germination in samples that had not undergone drying (total MC of 92.1%). By hour five of drying, total MC of the synthetic seeds had reached 25.2%, and germination had been reduced to 36.7% (Sundararaj et al. 2010).

Dehydration by incubation in sucrose may be a better alternative to air-drying because the accumulation of non-soluble carbohydrates in plant tissues has been reported to be an adequate coping mechanism for stress – including cold and dehydration – by regulating protein and membrane stabilization (Zhu et al. 2006). The addition of sucrose and glucose in gelling solutions (3 and 4%, respectively) aided in the storage of encapsulated Fragaria x ananassa (strawberry) and Rubus spp. (raspberry) shoot tips at 4°C for upwards of nine months (Lisek and Orlikowska 2004). Additionally, there was better regrowth of ginger microshoots dehydrated by incubation in 0.25 M sucrose (16-hour treatment) subject to four weeks of 25°C storage (73.3%) compared to fully-hydrated synthetic seeds (53.3%) (Sundararaj et al. 2010). Additional studies investigating M. x giganteus synthetic seed development should focus on plantlet conversion rate after: desiccation to various MCs and the utilization of different methods of dehydration; treatment with other protective agents (eg. serotonin, L-cysteine, ascorbic acid, etc.) for improved growth; and encapsulation of different tissues (ie. somatic embryos, mini-sprouts, root tips, etc.).

The current study investigated growth potential of synthetic seeds after four weeks of 3°C CS. As reviewed by Banerjee et al. (2012), storage temperatures for synthetic seeds has been a priority over storage substrates. Reduction in regrowth rate after CS treatment in the present study loosely

130 reflected those of Ghanbarali et al. (2016) where plantlet conversion declined with CS. Furthermore, none of the synthetic seeds planted in the mist bed germinated after six weeks of incubation. Ghanbarali et al. (2016) found that optimal regrowth substrates for PMS synthetic seeds were MS media void of PGRs, and coco peat (50-94 and 33-61%, respectively), while soil mixture substrate (similar to what was used in the present study) only yielded a growth rate between 17 and 28%. Nyende et al. (2003) also reported 0% germination of encapsulated potato shoot tips void of carbendazim when planted ex vitro, due to excessive moisture loss and simultaneous attack by soil fungi. Germination increased to approximately 100% when carbendazim was supplemented in the capsules in combination with MS preculture (2-3 weeks). Adopting these approaches for encapsulated M. x gigeanteus MRs, specifically evaluating pretreatments (eg. MS preculture, cold-shock, GA3, cytokinins, biostimulants, etc.), growth substrates (eg. sand, coco peat, etc.), planting depths, and fungicide application (ie. carbendazim, Bavistin®, and rose Bengal) may enable this technology to be developed for commercial application (Nyende et al. 2003; Banerjee et al. 2012; Ghanbarali et al. 2016).

5.5 Conclusions

This is the first report investigating the induction of in vitro M. x giganteus MR development and encapsulation. The genotype selected for this research – ‘M161’ or ‘Illinois’ – is characterized as an “elite” variety and has been the standard for Miscanthus research at the University of Illinois (Withers 2015). While further research is needed to optimize this system to improve plantlet conversion, this represents a promising alternative propagation system for the species.

M. x giganteus is a sterile triploid species that can only be propagated currently through expensive vegetative means (DRP, rhizome-derived plugs, and micropropagation), with the majority of costs being attributed to germplasm expenses. An alternative propagation system for this candidate bioenergy crop has the potential to lower germplasm costs through increasing division rate, retaining “elite” genotypes, optimizing low-maintenance storage conditions, creating certified, disease-free propagules for national and international sale, and providing consistency in storage and regrowth quality. Our study demonstrated that development of in vitro MRs and creation of synthetic seeds capable of regrowth is possible with M. x giganteus, and that cost-effective propagules are promising with further optimization of this technology.

131

5.6 Acknowledgements

The authors wish to thank Dr. Mukund Shukla for his insight on various plant tissue culture techniques and experimental design; Abhishek Chattopadhyay, Scott Belton, and James Nicholson for their help in maintaining shoot cultures; All Weather Farming Inc. for supplying explant material; the Gosling Research Institute for Plant Preservation for allowing use of their laboratory space and supplies; and Dr. Michelle Edwards for her guidance on data analysis.

6: Overall Discussion 6.1 Research contributions

The findings presented in this thesis summarize a variety of propagation and germplasm conservation methods available for M. x giganteus both in situ and in vitro. The information contained in these studies may be of particular importance to producers looking to expand their role in the biomass feedstock marketplace through prolonged storage and multiplication of rhizome cuttings (Chapter 3), in vitro preservation of genotypes that have the potential to be manipulated via unconventional breeding strategies (Chapter 4), and companies offering propagules which occupy less space with reduced chance of disease incidence to farmers (Chapter 5).

As found in numerous studies, M. x giganteus rhizome cuttings served as an acceptable propagule source for an assortment of accessions, including the ‘Illinois’ genotype (Clifton-Brown and Lewandowski 2000; Pyter et al. 2010; Davies et al. 2011; Xue et al. 2015), with little loss in viability after exposure to CS conditions. There were minor correlations found between rhizome carbohydrate concentrations and regrowth after predetermined times in storage; however, the genotypes with highest (‘UK’ and ‘BC’) and lowest (‘Amuri’) levels of reducing sugars associated with inferior and superior greenhouse growth, respectively. Future research delving into ‘G’ x ‘E’ x ‘G x E’ factors and interactions influencing these parameters should be conducted, such as the design executed by Rosser (2012).

As verified by studies by Kim et al. (2010) and Kim et al. (2012), utilizing immature inflorescences as an explant source for the development of M. x giganteus embryogenic and organogenic calli has many benefits, including the creation of disease-free germplasm stock and a foundation for in vitro breeding initiatives. Though these previous studies were able to provide the basis for this

132 system, calli maintained on semi-solid media would lose their regeneration potential only after four months in culture. The current study (Chapter 4) was able to prolong regeneration for up to one year when calli were maintained on semi-solid callus maintenance media supplemented with AIP. AIP functions to inhibit the PAL enzyme, which facilitates the rate-limiting step of the phenylpropanoid biosynthetic pathway (Zoń and Amrhein 1992). Future optimization of this in vitro conservation system should include the verification of genetic stability over culture duration (Rambaud et al. 2013), the quantification of individual phenolic compounds over time, the possible interactions between PGRs and phenols, and the compartmentalization of these compounds (eg. accumulation in the callus mass vs. somatic embryo) (Lozovaya et al. 1996; Cvikrová et al. 1998).

Because M. x giganteus is a sterile triploid, it is unable to produce viable seed and must be propagated through vegetative means. In order for this species to thrive as a second-generation biomass feedstock, and depending on the end-goal of the customers, a variety of propagation methods for producers to offer their clients would be advantageous (Xue et al. 2015). Similar to what has been demonstrated in ginger, turmeric, and potato, this study (Chapter 5) was able to develop in vitro intact and encapsulated MRs in M. x giganteus. Compared to traditional rhizome cuttings harvested from greenhouse or field conditions, MRs can be certified disease-free, be produced with reduced mechanical damage and improved consistency, and can occupy a fraction of the space and resources required for transport and storage. Further research should aim at optimizing the media composition and environmental conditions necessary for greater MR yield and quality.

6.2 Research limitations

The current study intended to reflect harvest, processing, storage, and growing practices that would be implemented by farmers (Chapter 3). However, concluding palpable interpretations from the results of this study was largely impractical due to the considerable extent of confounding factors. These factors included the following: 1) different stand establishment locations for distinct genotypes (Port Ryerse vs. Elora, ON); 2) distinct harvest times within and between trials; and 3) differing harvest methods at each location (specialized machinery vs. manual harvest).

Regarding location, genotype stands differed in age, were established in different soil types and at various densities, were managed with distinct agronomic practices, and were subjected to different environmental conditions. When considering harvest times, the removal of Port Ryerse and Elora

133 genotypes differed by at least one week in spring 2016, and by approximately one month in autumn 2015 and 2016. Between trials, rhizomes were harvested later in autumn 2016 compared to autumn 2015, due to an unseasonably-warm autumn. Furthermore, the different harvest methods and soil types between the two sites potentially influenced the amount of mechanical damage experienced by the bulk of the germplasm, as well as the quality of samples. Samples obtained from Port Ryerse were harvested over a relatively large area of land comprising many stands, while those collected from Elora could only be gathered from a few stands in border rows.

Moreover, sample availability was dependent on the discretion of the farmer and research manager, which resulted in a reduction in genotype(s) obtained for the lower storage temperature treatment (0 or 1°C) in most trials. In addition to these considerations, the ‘Illinois’ samples in the autumn 2016 trial stored at 3°C were removed from analysis due to severe contamination, and a spring harvest season that included all five genotypes could not be repeated due to time restraints.

In the final trial of the storage experiment, virtually all ‘Illinois’ rhizomes stored at 3°C displayed symptoms of pathogenic attack. These symptoms included discolouration of the rhizome epidermis and cortex (ranging from burnt orange to maroon), increased moisture accumulation, loss of tissue rigidity (soft-rot) (Beccari et al. 2010), putrid odour, and emergence of white mycelium on or within the epidermis, cortex, and pith. Though pest attack in Miscanthus rhizomes can be provoked by agents like wireworm larvae and microtus (Caslin et al. 2010), fungal species (including Fusarium avenaceum, Fusarium oxysporum, and Mucor hiemalis) (Covarelli et al. 2012) were the alleged suspects in this experiment based on the symptoms displayed. The extent of the disease in these samples was so intense that culm emergence was completely hindered in the greenhouse, which resulted in samples being excluded from the analysis. It is interesting to note that ‘Illinois’ rhizomes of the same origin stored at 1°C developed milder symptoms than those in 3°C conditions and continued to display culm emergence and seemingly normal growth after planting. This observation suggests that lower storage temperatures may aide in obstructing the growth and disease severity of some pathogens in M. x giganteus germplasm, though this warrants further investigation.

Although the regenerative potential of M. x giganteus calli could be prolonged on semi-solid media for extended periods of time (relative to previous studies) (Chapter 4), sections of somatic embryos or roots were not taken and verified histologically (Ślusarkiewicz-Jarzina et al. 2017). Also,

134 because Miscanthus is monocotyledonous, the stages of embryo development (proembryo, globular, scutellar, and coleptilar) were not apparent without fixing and staining. Future investigation into the preservation of Miscanthus callus should also assess genetic stability through methods such as inter simple sequence repeat (ISSR), AFLP, and/or random amplified polymorphic DNA (RAPD) polymerase chain reactions (PCR) (Qin et al. 2013).

Intact and encapsulated MRs were successfully induced and viable from in vitro M. x giganteus plantlets (Chapter 5). However, only a select few media compositions were tested for MR development. A greater range of basal media, carbohydrate, and PGR concentrations, as well as different PGR combinations, should be investigated in the future for better optimization of MR yield and quality. Furthermore, it has been shown in other species that changing incubation duration, temperature, and photoperiod can all have dramatic effects on MR development (Balachandran et al. 1990; Kapoor and Rao 2006).

Only one encapsulation method was tested on in vitro MR sections in this study; amending the concentrations and incubation times of sodium alginate and CaCl2 have been shown to improve explant viability in other species and warrants further investigation in Miscanthus. In explants such as potato minitubers, the addition of MS basal salts, antioxidants, and fungicides supplemented in the encapsulation medium improved regrowth and plantlet vigor (Gantait et al. 2015; Ghanbarali et al. 2016).

6.3 Conclusion

Though M. x giganteus is a sterile triploid and fails to produce viable seed, many alternative propagation methods exist for this species, allowing for improvements of this crop to be tackled from multiple avenues. The current study had only explored three of these avenues: in situ rhizome cuttings (Chapter 3), embryogenic and organogenic calli (Chapter 4), and in vitro intact and encapsulated MRs (Chapter 5). These methods also double as conservation systems, and still have much room for improvement; however, the findings presented in this thesis provide a foundation for the optimization of M. x giganteus production, conservation, and a starting point for a potential reduction of germplasm costs in the marketplace.

135

7: References

Adati, S., and I. Shiotani. 1962. The Cytotaxonomy of the Genus Miscanthus and Its Phylogenetic Status. Bull. Fac. Agric. Mie. Univ. 25: 1–14.

Adelberg, J.W., and M.M. Cousins. 2007. Development of Micro- and Minirhizomes of Turmeric, Curcuma longa L., in Vitro. Acta Hort. 756: 103–108.

Alam, S.M.M. 1992. Morphological and Physiological Changes During Dormancy Release of Tubers of Potato. Guelph, ON, Canada: University of Guelph.

Ali, H.M., El-Gizawy, A.M., El-Bassiouny, R.E.I., and M.A. Saleh. 2014. Browning Inhibition Mechanisms by Cysteine, Ascorbic Acid and Citric Acid, and Identifying PPO-Catechol- Cysteine Reaction Products. Journal of Food Science and Technology 52 (6): 3651–3659.

Anderson, E., Arundale, R., Maughan, M., Oladeinde, A., Wycislo, A., and T. Voigt. 2011. Growth and agronomy of Miscanthus x giganteus for biomass production. Biofuels 2: 167-179.

Appeldoorn, N.J.G., de Bruijn, S.M., Koot-Gronsveld, E.A.M., Visser, R.G.F., Vreugdenhil, D., and L.H.W. van der Plas. 1999. Developmental Changes in Enzymes Involved in the Conversion of Hexose Phosphate and Its Subsequent Metabolites during Early Tuberization of Potato. Plant, Cell & Environment 22: 1085–1096.

Arduini, I., Masoni, A., Mariotti, M., and L. Ercoli. 2004. Low Cadmium Application Increase Miscanthus Growth and Cadmium Translocation. Environmental and Experimental Botany 52: 89–100.

Atkinson, C.J. 2009. Establishing Perennial Grass Energy Crops in the UK: A Review of Current Propagation Options for Miscanthus. Biomass and Bioenergy 33: 752–759.

Aurangzaib, M. 2012. Performance evaluation of nine varieties of Miscanthus in Iowa. Graduate Theses and Dissertations. 12690. http://lib.dr.edu/etd/12690. Babu, K.N., Samsudeen, K., Divakaran, M., Pillai, G.S., Sumathi, V., Praveen, K., Ravindran, P.N., and K.V. Peter. 2016. Protocols for In Vitro Propagation, Conservation, Synthetic Seed Production, Embryo Rescue, Microrhizome Production, Molecular Profiling, and Genetic Transformation in Ginger (Zingiber officinale Roscoe.). In Protocols for In Vitro Cultures and Secondary Metabolite Analysis of Aromatic and Medicinal Plants, Second Edition, edited by S.. Jain, 1391: 403–426. New York, NY: Springer New York.

Balachandran, S.M., Bhat, S.R., and K.P.S. Chandel. 1990. In Vitro Clonal Multiplication of Turmeric (Curcuma spp.) and Ginger (Zingiber officinale Rosc.). Plant Cell Reports 8: 521–524.

Banerjee, S., Singh, S., Pandey, H., Pandey, P., and L.U. Rahman. 2012. Conservation and Storage of Curcuma amada Roxb. Synseeds on Luffa Sponge Matrix and RAPD Analysis of the Converted Plantlets. Industrial Crops and Products 36: 383–388.

136

Barling, A., Swaminathan, K., Mitros, T., James, B.T., Morris, J., Ngamboma, O., Hall, M.C., Kirkpatrick, J., Alabady, M., and A.K. Spence. 2013. A Detailed Gene Expression Study of the Miscanthus Genus Reveals Changes in the Transcriptome Associated with the Rejuvenation of Spring Rhizomes. BMC Genomics 14: 864.

Beale, C.V., and S.P. Long. 1997. Seasonal Dynamics of Nutrient Accumulation and Partitioning in the Perennial C4-Grasses Miscanthus × giganteus and Spartina cynosuroides. Biomass and Bioenergy 12: 419–428.

Beale, C.V., Morison, J.I.L., and S.P. Long. 1999. Water Use Efficiency of C4 Perennial Grasses in a Temperate Climate. Agricultural and Forest Meteorology 96: 103–115.

Beccari, G., Covarelli, L., Balmas, V., and L. Tosi. 2010. First Report of Miscanthus × giganteus Rhizome Rot Caused by Fusarium Avenaceum, Fusarium Oxysporum and Mucor Hiemalis. Australasian Plant Disease Notes 5: 28–29.

Boersma, N.N., and E.A. Heaton. 2014. Propagation Method Affects Miscanthus×giganteus Developmental Morphology. Industrial Crops and Products 57: 59–68.

Borzenkova, R.A., and M.P. Borovkova. 2003. Developmental Patterns of Phytohormone Content in the Cortex and Pith of Potato Tubers as Related to Their Growth and Starch Content. Russian Journal of Plant Physiology 50: 119–124.

Borzenkova, R.A., Sobyanina, E.A., Pozdeeva, A.A., and M.Y. Yashkov. 1998. Effect of Phytohormones on Starch-Synthesizing Capacity in Growing Potato Tubers. Russian Journal of Plant Physiology 45: 557–566.

Brejda, J.J., Moser, L.E., and S.S. Waller. 1989. Rhizome and Tiller Development of Three Nebraska Sandhills Warm-Season Grasses. Prairie Pioneers: Ecology, History and Culture: Proceedings of the Eleventh North American Prairie Conference 22: 211–215.

Březinová, A., Holík, J., Zažímalová, V., and J. Malá. 1996. Somatic Embryogenesis in Oak (Quercus robur L.), from Molecular Mechanisms to the Plant: An Integrated Approach. Plant Physiology and Biochemistry, no. Special Issue: 31.

Bullard, M.J., Heath, M.C., and P.M.I. Nixon. 1995. Shoot Growth, Radiation Interception and Dry Matter Production and Partitioning during the Establishment Phase of Miscanthus sinensis ‘Giganteus’ Grown at Two Densities in the UK. Annals of Applied Biology 126: 365–378.

Caldwell, M.M., Richards, J.H., Johnson, D.A., Nowak, R.S., and R.S. Dzurec. 1981. Coping with Herbivory: Photosynthetic Capacity and Resource Allocation in Two Semiarid Agropyron Bunchgrasses. Oecologia 50: 14–24.

Caslin, B., Finnan, J., and A. McCracken. 2010. Miscanthus Best Practice Guidelines. Teagasc and the Agri-Food and Bioscience Institute.

137

Christian, D.G., and E. Haase. 2001. Agronomy of Miscanthus. In Miscanthus for Energy and Fibre, 21–45. The Cromwell Press, UK: Earthscan.

Christian, D.G., Yates, N.E., and A.B. Riche. 2009. Estimation of Ramet Production from Miscanthus×giganteus Rhizome of Different Ages. Industrial Crops and Products 30: 176–178.

Christian, E.J. 2012. Seed Development and Germination of Miscanthus sinensis. Graduate Theses and Dissertations, Ames, Iowa, US: Iowa State University.

Clayton, S.D., and S.A. Renvoize. 1986. Genera Graminum : Grasses of the World. London : H.M.S.O. http://www.press.uchicago.edu/ucp/books/book/distributed/G/bo9856139.html

Clifton-Brown, J.C., and I. Lewandowski. 2000. Overwintering Problems of Newly Established Miscanthus Plantations Can Be Overcome by Identifying Genotypes with Improved Rhizome Cold Tolerance. New Phytologist 148: 287–294.

Clifton-Brown, J.C., Lewandowski, I., Andersson, B., Basch, G., Christian, D.G., Kjeldsen, J.B., Jørgensen, U., et al. 2001. Performance of 15 Genotypes at Five Sites in Europe. Agronomy Journal 93: 1013–1019.

Clifton-Brown, J.C., Lewandowski, I., Bangerth, F., and M.B. Jones. 2002. Comparative Responses to Water Stress in Stay-Green, Rapid- and Slow Senescing Genotypes of the Biomass Crop, Miscanthus. New Phytologist 154: 335–345.

Clifton-Brown, J.C., Stampfl, P.F., and M.B. Jones. 2004. Miscanthus Biomass Production for Energy in Europe and Its Potential Contribution to Decreasing Fossil Fuel Carbon Emissions. Global Change Biology 10: 509–518.

Conner, A.J., and P.G. Falloon. 1993. Osmotic versus Nutritional Effects When Rooting in Vitro Asparagus Minicrowns on High Sucrose Media. Plant Science 89: 101–106.

Cousins, M.M., and J.W. Adelberg. 2008. Short-Term and Long-Term Time Course Studies of Turmeric (Curcuma longa L.) Microrhizome Development in Vitro. Plant Cell, Tissue and Organ Culture 93: 283–293.

Covarelli, L., Beccari, G., and L. Tosi. 2012. Miscanthus Rhizome Rot: A Potential Threat for the Establishment and the Development of Biomass Cultivations. Biomass and Bioenergy 46: 263–269.

Crabb, D.W., and J. Shetty. 2003. Improving the Properties of Amylolytic Enzymes by Protein Engineering. Trends in Glycoscience and Glycotechnology 15: 115–126.

Cruz, V.M.V., and D.A. Dierig. 2015. Industrial Crops - Breeding for BioEnergy and Bioproducts. Handbook of Plant Breeding. New York, NY: New York: Springer. //www.springer.com/gp/book/9781493914463.

138

Cvikrová, M., Binarová, P., Eder, J., Vágner, M., Hrubcová, M., Zoń, J., and I. Macháčková. 1999. Effect of Inhibition of Phenylalanine Ammonia-Lyase Activity on Growth of Alfalfa Cell Suspension Culture: Alterations in Mitotic Index, Ethylene Production, and Contents of Phenolics, Cytokinins and Polyamines. Physiologia Plantarum 107: 329–337.

Cvikrová, M., Malá, J., Eder, J., Hrubcová, M., and M. Vágner. 1998. Abscisic Acid, Polyamines and Phenolic Acids in Sessile Oak Somatic Embryos in Relation to Their Conversion Potential. Plant Physiology and Biochemistry 36: 247–255.

Cvikrová, M., Malá, J., Hrubcová, M., Eder, J., Zoń, J., and I. Macháčková. 2003. Effect of Inhibition of Biosynthesis of Phenylpropanoids on Sessile Oak Somatic Embryogenesis. Plant Physiology and Biochemistry 41: 251–259.

Davies, M.J., Longbottom, H., and C.J. Atkinson. 2011. Changes in Duration of Rhizome Cold Storage and Manipulation of the Growing Environment to Promote Field Establishment of Miscanthus giganteus. Biomass and Bioenergy 35: 4268–4279. Defra. 2002. Establishing Perennial Grass Energy Crops: A Review of Current Propagation Options with a Focus on Miscanthus. Defra. file:///C:/Users/cdown/Downloads/NF0439_9800_FRA%20(1).pdf.

Dessureault, D. 2015. Canada Biofuels Annual. USDA Foreign Agriculture Service. https://gain.fas.usda.gov/Recent%20GAIN%20Publications/Biofuels%20Annual_Ottawa _Canada_8-19-2015.pdf.

Deuter, M. 2011. US PP22,033 P2. Hayward, California, US: Mendel Biotechnology Inc. Dohleman, F.G., and S.P. Long. 2009. More Productive Than Maize in the Midwest: How Does Miscanthus Do It? Plant Physiology 150: 2104–2115.

El Kasmioui, O., and R. Ceulemans. 2013. Financial Analysis of the Cultivation of Short Rotation Woody Crops for Bioenergy in Belgium: Barriers and Opportunities. BioEnergy Research 6: 336–350.

Engbers, H., and B. Deen. 2013. Field-Scale Agricultural Biomass Research and Development Project Final Report. Agriculture and Agri-Food Canada. https://www.ontariosoilcrop.org/wp-content/uploads/2015/11/biomass_final_report- january2014-h.engbers__b.deen.pdf.

Farage, P.K., Blowers, D., Long, S.P., and N.R. Baker. 2006. Low Growth Temperatures Modify the Efficiency of Light Use by Photosystem II for CO2 Assimilation in Leaves of Two Chilling-Tolerant C4 Species, Cyperus longus L. and Miscanthus × giganteus. Plant, Cell & Environment 29 (4): 720–728.

Fiegert, A.K., Mix-Wagner, G., and K.D. Vorlop. 2000. Regeneration of Solanum tuberosum L. cv. Tomensa: Induction of Somatic Embryogenesis in Liquid Culture for the Production of ‘Artificial Seed’. Landbauforschung Völkenrode 50: 199–202.

139

Fitch, M.M.M., and P.H. Moore. 1990. Comparison of 2,4-D and Picloram for Selection of Long- Term Totipotent Green Callus Cultures of Sugarcane. Plant Cell, Tissue and Organ Culture 20: 157–163.

Folin, O., and V. Ciocalteu. 1927. On Tyrosine and Tryptophane Determinations in Proteins. Journal of Biological Chemistry 73: 627–650.

Food and Agriculture Organization of the United Nations, ed. 2006. World Reference Base for Soil Resources, 2006: A Framework for International Classification, Correlation, and Communication. 2006 ed. World Soil Resources Reports 103. Rome: Food and Agriculture Organization of the United Nations.

Friesen, P.C., Peixoto, M.M, Busch, F.A., Johnson, D.C., and R.F. Sage. 2014. Chilling and Frost Tolerance in Miscanthus and Saccharum Genotypes Bred for Cool Temperate Climates. Journal of Experimental Botany 65: 3749–3758.

Gantait, S., Kundu, S., Ali, N., and N.C. Sahu. 2015. Synthetic Seed Production of Medicinal Plants: A Review on Influence of Explants, Encapsulation Agent and Matrix. Acta Physiologiae Plantarum 37: 1–12.

Garner, N., and J. Blake. 1989. The Induction and Development of Potato Microtubers In Vitro on Media Free of Growth Regulating Substances. Annals of Botany 63: 663–674.

Ghanbarali, S., Abdollahi, M.R., Zolnorian, H., Moosavi, S.S., and J.M. Seguí-Simarro. 2016. Optimization of the Conditions for Production of Synthetic Seeds by Encapsulation of Axillary Buds Derived from Minituber Sprouts in Potato (Solanum tuberosum). Plant Cell, Tissue and Organ Culture 126: 449–458.

Gibson, S. 2005. Control of Plant Development and Gene Expression by Sugar Signaling. Current Opinion in Plant Biology 8: 93–102.

Głowacka, K. 2011. A Review of the Genetic Study of the Energy Crop Miscanthus. Biomass and Bioenergy 35: 2445–2454.

Głowacka, K., Balachandran, S.M., and Z. Kaczmarek. 2010. Impact of Colchicine Application during Callus Induction and Shoot Regeneration on Micropropagation and Polyploidisation Rates in Two Miscanthus Species. In Vitro Cellular & Developmental Biology - Plant 46: 161–171.

Godt, D., and T. Roitsch. 2006. The Developmental and Organ Specific Expression of Sucrose Cleaving Enzymes in Sugar Beet Suggests a Transition between Apoplasmic and Symplasmic Phloem Unloading in the Tap Roots. Plant Physiology and Biochemistry: PPB 44: 656–665.

Government of Canada, Environment and Climate Change Canada. 2012. Environment and Climate Change Canada - Federal Renewable Fuels Regulations: Blending Operations. March 9, 2012. https://www.ec.gc.ca/energie-energy/default.asp?lang=En&n=828C9342- 1.

140

Greef, J.M., Pude, R., Rennebaum, H., Schwarz, K.U., and W. Diepenbrock. 1994. Development of Above- and Underground Organs of Miscanthus x giganteus in Northern Germany. In Alternative Oilseed and Fibre Crops for Cool and Wet Regions of Europe, S. Hennick, L.J.M. van Soest, K. Pithan and L. Hof, 101–112. Brussels: E. Guyot S.A.

Grzegorczyk, I., and H. Wysokińska. 2011. A Protocol for Synthetic Seeds From Salvia officinalis L. Shoot Tips. Acta Biologica Cracoviensia Series Botanica 53: 80–85.

Gubišová, M., Gubiš, J., Žofajová, A., Mihálik, D., and J. Kraic. 2013. Enhanced in Vitro Propagation of Miscanthus×giganteus. Industrial Crops and Products 41: 279–282.

Hajirezaei, M., and U. Sonnewald. 1999. Inhibition of Potato Tuber Sprouting: Low Levels of Cytosolic Pyrophosphate Lead to Non-Sprouting Tubers Harvested from Transgenic Potato Plants. Potato Research 42: 353–372.

Hajirezaei, M.-R., Börnke, F., Peisker, M., Takahata, Y., Lerchl, J., Kirakosyan, A., and U. Sonnewald. 2003. Decreased Sucrose Content Triggers Starch Breakdown and Respiration in Stored Potato Tubers (Solanum tuberosum). Journal of Experimental Botany 54: 477– 488.

Hamadina, E.I, and P.Q. Craufurd. 2015. Changes in Free Phenolics Contents during Tuber Development, Dormancy and Sprouting in White Yam (Dioscorea rotundata Poir.). International Journal of Plant Research 5: 34-41.

Hansen, J., and K. Kristiansen. 1997. Short-Term in Vitro Storage of Miscanthus x ogiformis Honda ‘Giganteus’ as Affected by Medium Composition, Temperature, and Photon Flux Density. Plant Cell, Tissue and Organ Culture 49: 161–169.

Hariprakash, C.S., and B. Nambisan. 1996. Carbohydrate Metabolism during Dormancy and Sprouting in Yam (Dioscorea) Tubers: Changes in Carbohydrate Constituents in Yam (Dioscorea) Tubers during Dormancy and Sprouting. Journal of Agricultural and Food Chemistry 44: 3066–3069.

Hartley, R.D., and C.W. Ford. 1989. Phenolic Constituents of Plant Cell Walls and Wall Biodegradability. In Plant Cell Wall Polymers Biogenesis and Biodegradation, 399:137– 45. ACS Symposium Series American Chemical Society. Washington: American Chemical Society. https://pubs.acs.org/doi/abs/10.1021/bk-1989-0399.ch009.

Haughton, A.J., Bond, A.J., Lovett, A.A., Dockerty, T., Sünnenberg, G., Clark, S.J., Bohan, D.A., et al. 2009. A Novel, Integrated Approach to Assessing Social, Economic and Environmental Implications of Changing Rural Land-Use: A Case Study of Perennial Biomass Crops. Journal of Applied Ecology 46: 315–322.

Heaton, E.A., Dohleman, F.G, and S.P. Long. 2008. Meeting US Biofuel Goals with Less Land: The Potential of Miscanthus. Global Change Biology 14: 2000–2014.

141

Heaton, E.A., Long, S.P., Voigt, T.B., Jones, M.B., and J.C. Clifton-Brown. 2004. Miscanthus for Renewable Energy Generation: European Union Experience and Projections for Illinois. Mitigation and Adaptation Strategies for Global Change 9: 433–451.

Himelblau, E., and R.M. Amasino. 2001. Nutrients Mobilized from Leaves of Arabidopsis thaliana during Leaf Senescence. Journal of Plant Physiology 158: 1317–1323.

Hodkinson, T.R., Chase M.W., and S.A. Renvoize. 2002. Characterization of a Genetic Resource Collection for Miscanthus (Saccharinae, Andropogoneae, Poaceae) Using AFLP and ISSR PCR. Annals of Botany 89: 627–636.

Hodkinson, T.R., Chonghaile, G.N., Sungkaew, S., Chase, M.W., Salamin, N., and C.M.A. Stapleton. 2010. Phylogenetic Analyses of Plastid and Nuclear DNA Sequences Indicate a Rapid Late Miocene Radiation of the Temperate Bamboo Tribe Arundinarieae (Poaceae, Bambusoideae). Plant Ecology and Diversity 3: 109–120.

Holme, I.B., Krogstrup, P., and J. Hansen. 1997. Embryogenic Callus Formation, Growth and Regeneration in Callus and Suspension Cultures of Miscanthus x ogiformis Honda 'Giganteus' as Affected by Proline. Plant Cell, Tissue and Organ Culture 50: 203–210.

Holme, I.B., and K.K. Petersen. 1996. Callus Induction and Plant Regeneration from Different Explant Types of Miscanthus x ogiformis Honda ‘Giganteus.’ Plant Cell, Tissue and Organ Culture 45: 43–52.

Hongfei, D., Fu, M., Yang, X., and Q. Chen. 2016. Ethylene Inhibited Sprouting of Potato Tubers by Influencing the Carbohydrate Metabolism Pathway Journal of Food Science and Technology 53: 3166–3174.

Hrubcová, M., Cvikrová, M., Eder, J., Zoń, J., and I. Macháčková. 2000. Effect of Inhibition of Phenylpropanoid Biosynthesis on Peroxidase and IAA-Oxidase Activities and Auxin Content in Alfalfa Suspension Cultures. Plant Physiology and Biochemistry 38: 949–956.

Janick, J., Kitto, S.L., and Y.-H. Kim. 1989. Production of Synthetic Seed by Desiccation and Encapsulation” In Vitro Cellular & Developmental Biology-Plant 25: 1167–1172.

Janišová, M. 2006. Tiller Demography of Festuca Pallens Host (Gramineae) in Two Dry Grassland Communities. Polish Journal of Ecology 54: 201–213.

Jensen, E., Robson, P., Norris, J., Cookson, A., Farrar, K., Donnison, I., and J.C. Clifton-Brown. 2013. Flowering Induction in the Bioenergy Grass Miscanthus sacchariflorus Is a Quantitative Short-Day Response, Whilst Delayed Flowering under Long Days Increases Biomass Accumulation. Journal of Experimental Botany 64: 541–552.

Jeżowski, S., Głowacka, K., and Z. Kaczmarek. 2011. Variation on Biomass Yield and Morphological Traits of Energy Grasses from the Genus Miscanthus during the First Years of Crop Establishment. Biomass and Bioenergy 35: 814–821.

142

Jones, A.M.P., Chattopadhyay, A., Shukla, M., Zoń, J., and P.K. Saxena. 2012. Inhibition of Phenylpropanoid Biosynthesis Increases Cell Wall Digestibility, Protoplast Isolation, and Facilitates Sustained Cell Division in American Elm (Ulmus americana). BMC Plant Biology 12: 75.

Jones, A.M.P., and P.K. Saxena. 2013. Inhibition of Phenylpropanoid Biosynthesis in Artemisia annua L.: A Novel Approach to Reduce Oxidative Browning in Plant Tissue Culture. PLOS ONE 8: e76802.

Kapoor, P., and I.U. Rao. 2006. In Vitro Rhizome Induction and Plantlet Formation from Multiple Shoots in Bambusa bambos var. Gigantea Bennet and Gaur by Using Growth Regulators and Sucrose. Plant Cell, Tissue and Organ Culture 85: 211–217.

Kaur, J., Percival, D., Hainstock, L.J., and J.-P. Privé. 2012. Seasonal Growth Dynamics and Carbon Allocation of the Wild Blueberry Plant (Vaccinium angustifolium Ait.). Canadian Journal of Plant Science 92: 1145–1154.

Kendall, E.J., Kartha, K.K., Qureshi, J.A., and P.Chermak. 1993. Cryopreservation of Immature Spring Wheat Zygotic Embryos Using an Abscisic Acid Pretreatment. Plant Cell Reports 12: 89–94.

Kim, H.S., Zhang, G., Juvik, J.A., and J.M. Widholm. 2010. Miscanthus×giganteus Plant Regeneration: Effect of Callus Types, Ages and Culture Methods on Regeneration Competence. Global Change Biology Bioenergy 2: 192–200.

Kim, S., Da, K., and C. Mei. 2012. An Efficient System for High-Quality Large-Scale Micropropagation of Miscanthus × giganteus Plants. In Vitro Cellular & Developmental Biology - Plant 48: 613–619.

Kitto, S.L., and J. Janick. 1982. Polyox as an Artificial Seed Coat for Asexual Embryos. Horticulture Science 17: 448.

Koda, Y., and Y. Kikuta. 1991. Possible involvement of jasmonic acid in tuberization of Yam plants. Plant Cell Physiology 32: 629-633. Kolachevskaya, O.O., Alekseeva, V.V., Sergeeva, L.I., Rukavtsova, E.B., Getman, I.A., Vreugdenhil, D., Buryanov, Y.I., and G.A. Romanov. 2015. Expression of Auxin Synthesis Gene Tms1 under Control of Tuber-Specific Promoter Enhances Potato Tuberization in Vitro. Journal of Integrative Plant Biology 57: 734–744.

Lang, G.A. 1987. Endo-, Para-, and Ecodormancy: Physiological Terminology and Classification for Dormancy Research. American Society for Horticultural Science 22: 371–377.

Le Ngoc Huyen, T., Rémond, C., Dheilly, R.M., and B. Chabbert. 2010. Effect of Harvesting Date on the Composition and Saccharification of Miscanthus x giganteus. Bioresource Technology 101: 8224–8231.

143

Leclerc, Y., Donnelly, D.J., Coleman, W.K., and R.R. King. 1995. Microtuber Dormancy in Three Potato Cultivars. American Potato Journal 72: 215–223.

Lewandowski, I. 1997. Micropropagation of Miscanthus x giganteus. In Biotechnology in Agriculture and Forestry, Bajaj YPS, 39:239–255. New York, NY: High-tech and micropropagation VSpringer.

Lewandowski, I. 1998. Propagation Method as an Important Factor in the Growth and Development of Miscanthus×giganteus. Industrial Crops and Products 8: 229–245.

Lewandowski, I., Clifton-Brown, J., Trindade, L.M., van der Linden, G.C., Schwarz, K.-U., Müller-Sämann, K., Anisimov, A., et al. 2016. Progress on Optimizing Miscanthus Biomass Production for the European Bioeconomy: Results of the EU FP7 Project OPTIMISC. Frontiers in Plant Science 7: 1620.

Lewandowski, I., and A. Kicherer. 1997. Combustion Quality of Biomass: Practical Relevance and Experiments to Modify the Biomass Quality of Miscanthus x giganteus. European Journal of Agronomy 6: 163–177.

Linde-Laursen, I. 1993. Cytogenetic Analysis of Miscanthus ‘Giganteus’, an Interspecific Hybrid. Hereditas 119: 297–300.

Lisek, A., and T. Orlikowska. 2004. In Vitro Storage of Strawberry and Raspberry in Calcium- Alginate Beads at 4°C. Plant Cell, Tissue and Organ Culture 78: 167–172.

Liu, G., Gilding, E.K., and I.D. Godwin. 2015. A Robust Tissue Culture System for Sorghum [Sorghum bicolor (L.) Moench]. South African Journal of Botany 98: 157–160.

Lozovaya, V., Gorshkova, T., Yablokova, E., Zabotina, O., Ageeva, M., Rumyantseva, N., Kolesnichenko, E., Waranyuwat, A., and J. Widholm. 1996. Callus Cell Wall Phenolics and Plant Regeneration Ability. Journal of Plant Physiology 148: 711–717.

MacDonald, M.J., and G.B.D. D’Cunha. 2007. A Modern View of Phenylalanine Ammonia Lyase. Biochemistry and Cell Biology 85: 273–282.

Malá, J., Cvikrová, H., Březinová, A., Hrubcová, M., Eder, J., Vágner, M., and M. Cvikrová. 2000. Endogenous Contents of Phytohormones and Phenylpropanoids in Sessile Oak Somatic Embryos in Relation to Their Conversion Potential. Journal of Forest Science 46: 197– 204.

Mangold, A., Lewandowski, I., Xue, S., and A. Kiesel. 2017. ‘Collar Propagation’ as an Alternative Propagation Method for Rhizomatous Miscanthus. Global Change Biology Bioenergy 10: 186–198.

Matthews, R., and P. Grogan. 2001. Potential C-Sequestration Rates under Short-Rotation Coppiced Willow and Miscanthus Biomass Crops: A Modelling Study. Aspects of Applied Biology 65: 303-312.

144

McCalmont, J.P., Hastings, A., McNamara, N.P., Richter, G.M., Robson, P., Donnison, I.S., and J. Clifton‐Brown. 2017. Environmental Costs and Benefits of Growing Miscanthus for Bioenergy in the UK. Global Change Biology. Bioenergy 9: 489–507.

McCarty, E.C. 1938. The Relation of Growth to the Varying Carbohydrate Content in Mountain Brome. U.S. Department of Agriculture Technical Bulletin 598: 1-24.

McCleary, B.V., and A.B. Blakeney. 1999. Measurement of Inulin and Oligofructan. Cereal Foods World 44: 398–406.

McLaughlin, S.B., and M.E. Walsh. 1998. Evaluating Environmental Consequences of Producing Herbaceous Crops for Bioenergy. Biomass and Bioenergy 14: 317–324.

Mehrotra, S., Goel, M.K., Kukreja, A.K., and B.N. Mishra. 2007. Efficiency of Liquid Culture Systems over Conventional Micropropagation: A Progress towards Commercialization. African Journal of Biotechnology 6: 1484–1492.

Moon, Y.-H., Cha, Y.-L., Choi, Y.-H., Yoon, Y.-M., Koo, B.-C., Ahn, J.-W., An, G.-H., Kim, J.- K., and K.-G. Park. 2013. Diversity in Ploidy Levels and Nuclear DNA Amounts in Korean Miscanthus Species. Euphytica 193: 317–326.

Moore, K.J., Moser, L.E., Vogel, K.P., Waller, S.S., Johnson, B.E., and Jeffrey F. Pedersen. 1991. Describing and Quantifying Growth Stages of Perennial Forage Grasses. Agronomy Journal 83: 1073–1077.

Morrish, F., Vasil, V., and I.K. Vasil. 1987. Developmental Morphogenesis and Genetic Manipulation in Tissue and Cell Cultures of the Gramineae. Edited by J.G. Scandalios and E.W. Caspari. Advances in Genetics, Molecular Genetics of Development, 24: 431–499.

Mukunthakumar, S., and J. Mathur. 1992. Artificial Seed Production in the Male Bamboo Dendrocalamus strictus L. Plant Science 87: 109–113.

Murashige, T., and F. Skoog. 1962. A Revised Medium for Rapid Growth and Bio Assays with Tobacco Tissue Cultures. Physiologia Plantarum 15: 473–497.

Murashige, T. 1977. Plant Cell and Organ Cultures as Horticultural Practices. Acta Horticulturae 78: 17–29.

Nayak, S. 2000. In Vitro Multiplication and Microrhizome Induction in Curcuma aromatica Salisb. Plant Growth Regulation 32: 41–47.

Nayak, S. 2002. High Frequency in Vitro Production of Microrhizomes of Curcuma amada. Indian Journal of Experimental Biology 40: 230–32.

Neves, N., Segura-Nieto, M., Blanco, M.A., Sánchez, M., González, A., González, J.L., and R. Castillo. 2003. Biochemical Characterization of Embryogenic and Non-Embryogenic Calluses of Sugarcane. In Vitro Cellular & Developmental Biology - Plant 39: 343–345.

145

New Energy Farms. 2015. CEEDSTM. http://www.newenergyfarms.com/products/ceeds/.

Nie, G., Huang, L., Zhang, X., Taylor, M., Jiang, Y., Yu, X., Liu, X., Wang, X., and Y. Zhang. 2016. Marker-Trait Association for Biomass Yield of Potential Bio-Fuel Feedstock Miscanthus sinensis from Southwest China. Frontiers in Plant Science 7: 802.

Nielsen, J.M., Brandt, K., and J. Hansen. 1993. Long-Term Effects of Thidiazuron Are Intermediate between Benzyladenine, Kinetin or Isopentenyladenine in Miscanthus sinensis. Plant Cell, Tissue and Organ Culture 35: 173–179.

Nieves, N., Martinez, M.E., Castillo, R., Blanco, M.A., and J.L. González-Olmedo. 2001. Effect of Abscisic Acid and Jasmonic Acid on Partial Desiccation of Encapsulated Somatic Embryos of Sugarcane. Plant Cell, Tissue and Organ Culture 65: 15–21.

Nyende, A.B., Schittenhelm, S., Mix-Wagner, G., and J.-M. Greef. 2003. Production, Storability, and Regeneration of Shoot Tips of Potato (Solanum tuberosum L.) Encapsulated in Calcium Alginate Hollow Beads. In Vitro Cellular & Developmental Biology - Plant 39: 540-544.

Okazawa, Y. 1967. Physiological Studies on the Tuberization of Potato Plants. Journal of the Faculty of Agriculture, Hokkaido University 55: 267–336.

Peixoto, M.d.M. 2015. Low Temperature Performance of Leading Bioenergy Crops Utilizing the C4 Photosynthetic Pathway. PhD Thesis, University of Toronto.

Perera, D., Barnes, D.J., Baldwin, B.S., and N.A. Reichert. 2015. Mutagenesis of in Vitro Cultures of Miscanthus × giganteus Cultivar Freedom and Detecting Polymorphisms of Regenerated Plants Using ISSR Markers. Industrial Crops and Products 65: 110–116.

Petersen, K.K. 1997. Callus Induction and Plant Regeneration in Miscanthus x ogiformis Honda ‘Giganteus’ as Influenced by Benzyladenine. Plant Cell, Tissue and Organ Culture 49: 137–140.

Płażek, A., Dubert, F., Janowiak, F., Krępski, T., and M. Tatrzańska. 2011. Plant Age and in Vitro or in Vivo Propagation Considerably Affect Cold Tolerance of Miscanthus×giganteus. European Journal of Agronomy 34: 163–171.

Pospíšilová, J., Wilhelmová, N., Synková, H., Čatský, J., Krebs, D., Tichá, I., Hanáčková, B., and J. Snopek. 1998. Acclimation of Tobacco Plantlets to Ex Vitro Conditions as Affected by Application of Abscisic Acid. Journal of Experimental Botany 49: 863–869.

Presant, E.W., and C.J Acton. 1984. The Soils of the Regional Municipality of Haldimand-Norfolk. Vol. II. Ontario Institute of Pedology 57. Guelph, ON, Canada: Metropolitan Dade County Planning Department.

Purdy, S.J., Cunniff, J., Maddison, A.L., Jones, L.E., Barraclough, T., Castle, M, Davey, C.L., et al. 2015. Seasonal Carbohydrate Dynamics and Climatic Regulation of Senescence in the Perennial Grass, Miscanthus. BioEnergy Research 8: 28–41.

146

Purdy, S.J., Maddison, A.L., Jones, L.E., Webster, R.J., Andralojc, J., Donnison, I., and J. Clifton- Brown. 2013. Characterization of Chilling-Shock Responses in Four Genotypes of Miscanthus Reveals the Superior Tolerance of M. × giganteus Compared with M. sinensis and M. sacchariflorus. Annals of Botany 111: 999–1013.

Pyter, R., Heaton, E., Dohleman, F., Voigt, T., and S. Long. 2009. Agronomic Experiences with Miscanthus giganteus in Illinois, USA. In Biofuels, 41–52. Methods in Molecular Biology. Humana Press, Totowa, NJ.

Pyter, R., Voigt, T., Heaton, E., Dohleman, F., and S. Long. 2007. Growing Giant Miscanthus in Illinois. University of Illinois Extension, Urbana-Champaign.

Pyter, R.J., Dohleman, F.G., and T.B. Voigt. 2010. Effects of Rhizome Size, Depth of Planting and Cold Storage on Miscanthus x giganteus Establishment in the Midwestern USA. Biomass and Bioenergy 34: 1466–1470.

Qin, Y., Kabir, M.A., Wang, H.W., Lee, Y.H., Hong, S.H., Kim, J.Y., Yook, M.J., et al. 2013. Assessment of Genetic Diversity and Relationships Based on RAPD and AFLP Analyses in Miscanthus Genera Landraces. Canadian Journal of Plant Science 93: 171–182.

Rambaud, C., Arnoult, S., Bluteau, A., Mansard, M.C., Blassiau, C., and M. Brancourt-Hulmel. 2013. Shoot Organogenesis in Three Miscanthus Species and Evaluation for Genetic Uniformity Using AFLP Analysis. Plant Cell, Tissue and Organ Culture 113: 437–448.

Redenbaugh, K., Paasch, B.D., Nichol, J.W. Kossler, M.E., Viss, P.R., and K.A. Walker. 1986. Somatic Seeds: Encapsulation of Asexual Plant Embryos. Nature Biotechnology 4: 797– 801.

Reuber, S., Leitsch, J., Krause, G.H., and G. Weissenböck. 1993. Metabolic Reduction of Phenylpropanoid Compounds in Primary Leaves of Rye (Secale cereale L.) Leads to Increased UV-B Sensitivity of Photosynthesis. Zeitschrift Fuer Naturforschung. C, Biosciences 48: 749–756.

Rosser, B. 2012. Evaluation of Miscanthus Winter Hardiness and Yield Potential in Ontario. Guelph, ON, Canada: University of Guelph. http://hdl.handle.net/10214/3663.

Sarkar, D., Pandey, S.K., and S. Sharma. 2006. Cytokinins Antagonize the Jasmonates Action on the Regulation of Potato (Solanum tuberosum) Tuber Formation in Vitro. Vol. 87: 285.

Semere, T., and F.M. Slater. 2007. Ground Flora, Small Mammal and Bird Species Diversity in Miscanthus (Miscanthus x giganteus) and Reed Canary-Grass (Phalaris arundinacea) Fields. Biomass and Bioenergy 31: 20–29.

Seran, T.H. 2013. In Vitro Propagation of Ginger (Zingiber officinale Rosc.) through Direct Organogenesis: A Review. Pakistan Journal of Biological Sciences 16: 1826–1835.

Sharma, S.K., Bryan, G.J., Winfield, M.O., and S. Millam. 2007. Stability of Potato (Solanum tuberosum L.) Plants Regenerated via Somatic Embryos, Axillary Bud Proliferated Shoots,

147

Microtubers and True Potato Seeds: A Comparative Phenotypic, Cytogenetic and Molecular Assessment. Planta 226: 1449–1458.

Sharma, T.R., and B.M. Singh. 1995. In Vitro Microrhizome Production in Zingiber Officinale Rosc. Plant Cell Reports 15: 274–277.

Shirgurkar, M.V., John, C.K., and R.S. Nadgauda. 2001. Factors Affecting in Vitro Microrhizome Production in Turmeric. Plant Cell, Tissue and Organ Culture 64: 5–11.

Singleton, V.L., Orthofer, R., and R.M. Lamuela-Raventós. 1999. Analysis of Total Phenols and Other Oxidation Substrates and Antioxidants by Means of Folin-Ciocalteu Reagent. Methods in Enzymology, Oxidants and Antioxidants Part A, 299: 152–178.

Singleton, V.L., and J.A. Rossi. 1965. Colorimetry of Total Phenolics with Phosphomolybdic- Phosphotungstic Acid Reagents. American Journal of Enology and Viticulture 16: 144– 158.

Ślusarkiewicz-Jarzina, A., Ponitka, A., Cerazy-Waliszewska, J., Wojciechowicz, M.K., Sobańska, K., Jeżowski, S., and T. Pniewski. 2017. Effective and Simple in Vitro Regeneration System of Miscanthus sinensis, M. × giganteus and M. sacchariflorus for Planting and Biotechnology Purposes. Biomass and Bioenergy 107: 219–226.

Smith, D. 1968. Classification of Several Native North American Grasses as Starch or Fructosan Accumulators in Relation to Taxonomy. Grass and Forage Science 23: 306–309.

Soil Classification Working Group. 1998. The Canadian System of Soil Classification. Third Edition. Ottawa, ON: Agric. and Agri-Food Can. Publ. 1646 (Revised).

Sousounis, P.J., and J.M. Fritsch. 1994. Lake-Aggregate Mesoscale Disturbances. Part II: A Case Study of the Effects on Regional and Synoptic-Scale Weather Systems 75: 1793–1812.

St-Pierre, B., and N. Brisson. 1995. Induction of the Plastidic Starch-Phosphorylase Gene in Potato Storage Sink Tissue. Planta 195: 339–344.

Sullivan, J.T., and V.G. Sprague. 1943. Composition of the Roots and Stubble of Perennial Ryegrass Following Partial Defoliation. Plant Physiology 18: 656–670.

Sundararaj, S.G., Agrawal, A., and R.K. Tyagi. 2010. Encapsulation for in Vitro Short-Term Storage and Exchange of Ginger (Zingiber officinale Rosc.) Germplasm. Scientia Horticulturae 125: 761–766.

Taşkın, H., Baktemur, G., Kurul, M., and S. Büyükalaca. 2013. Use of Tissue Culture Techniques for Producing Virus-Free Plant in Garlic and Their Identification through Real-Time PCR. The Scientific World Journal 2013: 1–5.

Tejera, M.D., and E.A. Heaton. 2017. Description and Codification of Miscanthus × giganteus Growth Stages for Phenological Assessment.” Frontiers in Plant Science 8: 1–12.

148

Tognetti, J.A., Pontis, H.G., and G.M.A. Martinex-Noel. 2013. Sucrose Signaling in Plants - A World yet to Be Explored. Plant Signaling & Behavior 8: e23316.

Townsend, L.R., Hall, I.V., and L.E. Aalders. 1968. Chemical Composition of Rhizomes and Associated Leaves of the Lowbush Blueberry. American Society for Horticultural Science 93: 248–253.

NRCS. 2011. Planting and Managing Giant Miscanthus as a Biomass Energy Crop. USDA Technical Note No. 4. https://www.nrcs.usda.gov/Internet/FSE_DOCUMENTS/stelprdb1044768.pdf. Viola, Roberto, Roberts, A.G., Haupt, S., Gazzani, S., Hancock, R.D., Marmiroli, N., Machray, G.C., and K.J. Oparka. 2001. Tuberization in Potato Involves a Switch from Apoplastic to Symplastic Phloem Unloading. The Plant Cell 13: 385–398.

Vreugdenhil, D., Boogaard, Y., Visser, R.G.F., and S.M. de Bruijn. 1998. Comparison of Tuber and Shoot Formation from in Vitro Cultured Potato Explants. Plant Cell, Tissue and Organ Culture 53: 197–204.

Wang, D., Portis, A.R., Moose, S.P., and S.P. Long. 2008. Cool C₄ Photosynthesis: Pyruvate Dikinase Expression and Activity Corresponds to the Exceptional Cold Tolerance of Carbon Assimilation in Miscanthus × giganteus. Plant Physiology 148: 557–567.

White, L.M. 1973. Carbohydrate Reserves of Grasses: A Review. Journal of Range Management 26: 13-18.

Wiltshire, J.J.J., and A.H. Cobb. 1996. A Review of the Physiology of Potato Tuber Dormancy. Annals of Applied Biology 129: 553–569.

Withers, K.K. 2015. Morphological Adaptations and Membrane Stabilizing Mechanisms of Overwintering Miscanthus (Poaceae). Unpublished thesis, Guelph, ON, Canada: University of Guelph. http://atrium.lib.uoguelph.ca/xmlui/handle/10214/8708.

Xu, M., Li, X., and S.S. Korban. 2004. DNA-Methylation Alterations and Exchanges during in Vitro Cellular Differentiation in Rose (Rosa hybrida L.). Theoretical and Applied Genetics 109: 899–910.

Xue, S., Kalinina, O., and I. Lewandowski. 2015. Present and Future Options for Miscanthus Propagation and Establishment. Renewable and Sustainable Energy Reviews 49: 1233– 1246.

Zaid, A. 1987. In Vitro Browning of Tissues and Media with Special Emphasis to Date Palm Cultures--a Review. Acta Horticulturae, In vitro problems related to mass propagation of horticultural plants, Sept 1987: 561–566.

Zawadzki, M., and A. Ragauskas. 1999. Quantitative Determination of Quinone Chromophores in Isolated Lignins. Georgia Institute of Technology, IPST technical paper series.

149

Zenkteler, E., and A. Bagniewska-Zadworna. 2005. Ultrastructural Changes in Rhizome Parenchyma of Polypodium Vulgare during Dehydration with or without Abscisic Acid Pretreatment. Biologia Plantarum 49: 209–214.

Zheng, Y., Liu, Y., Ma, M., and K. Xu. 2008. Increasing in Vitro Microrhizome Production of Ginger (Zingiber officinale Roscoe). Acta Physiologiae Plantarum 30: 513–519.

Zheng, Y.Q., Liu, Y.M., and K. Xu. 2004. Effects of Sucrose Concentration on Growth and Endohormone Concentration Variations of Tube-Ginger. China Vegetables 2: 15–17.

Zhu, G.-Y., Geuns, J.M.C., Dussert, S., Swennen, R., and B. Panis. 2006. Change in Sugar, Sterol and Fatty Acid Composition in Banana Meristems Caused by Sucrose-Induced Acclimation and Its Effects on Cryopreservation. Physiologia Plantarum 128: 80–94.

Zoń, J., and N. Amrhein. 1992. Inhibitors of Phenylalanine Ammonia-Lyase: 2-Aminoindan-2- Phosphonic Acid and Related Compounds. Liebigs Annalen Der Chemie 1992: 625–628.

Zub, H.W., and M. Brancourt-Hulmel. 2010. Agronomic and Physiological Performances of Different Species of Miscanthus, a Major Energy Crop. A Review. Agronomy for Sustainable Development 30: 201–214.

150

Appendices

Appendix 3.1 Type III test for the significance of main effects and their interactions (‘temperature’, ‘storage time’, and ‘temperature x storage time’) on plant growth traits for ‘Illinois’ after spring 2015 harvest and storage at 0 or 3°C (α=0.05). Spring 2015 Trait Effect(s) F-value Pr>F Average rhizome viability (%) Temperature 0.08 0.8257 Storage Time 2.15 0.2388 Temperature x Storage Time 0.47 0.8105 Average emergence speed (days) Temperature 0.91 0.5144 Storage Time 3.11 0.1459 Temperature x Storage Time 2.63 0.1155 Average tiller height (mm) Temperature 0.08 0.8209 Storage Time 11.29 0.0173* Temperature x Storage Time 3.7 0.0556 Average tiller number (#) Temperature 0.11 0.7964 Storage Time 1.19 0.4547 Temperature x Storage Time 2.93 0.0930 *Significant at α=0.05.

151

Appendix 3.2 Type III test for the significance of main effects and their interactions (‘genotype’, ‘storage time’, and ‘genotype x storage time’) on plant growth traits for ‘Illinois’ (autumn 2015 and spring 2016 only), ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’ after autumn 2015, spring 2016, and autumn 2016 harvests and storage at 3°C (α=0.05). Autumn 2015 Spring 2016 Autumn 2016 Trait Effect(s) F-value Pr>F F-value Pr>F F-value Pr>F Average Genotype 55.4 0.0009* 107.58 0.0003* 77.6 0.0024* rhizome Storage Time 4.3 0.0934 8.16 0.0190* 0.19 0.9532 viability (%) Genotype x Storage 2.05 0.0229* 3.16 0.0002* 0.99 0.4811 Time Average Genotype 7.53 0.038* 15.19 0.0110* 6.03 0.0871 emergence Storage Time 0.3 0.8679 25.02 0.0015* 20.2 0.0025* speed (days) Genotype x Storage 1.32 0.2149 4.72 <0.0001* 4.66 0.0001* Time Average Genotype 25.2 0.0042* 2.17 0.24 23.5 0.0138* tiller height Storage Time 4.46 0.0883 1.02 0.49 8.01 0.0197* (mm) Genotype x Storage 0.75 0.7369 1.63 0.07 0.97 0.4934 Time Average Genotype 21.8 0.0056* 3.75 0.11 0.3 0.8224 tiller number Storage Time 0.69 0.6371 1.76 0.27 10 0.0122* (#) Genotype x Storage 0.74 0.741 1.42 0.14 178 <0.0001* Time Average Genotype 5.21 0.0694 2.09 0.25 1.59 0.4085 stem node Storage Time 9.44 0.0258* 0.8 0.59 5.32 0.1006 number (#) Genotype x Storage 1.19 0.2999 1.1 0.36 3 0.0475* Time Average Genotype 24.3 0.0046* 21.51 0.01* 0.13 0.9333 stem Storage Time 2.75 0.1754 34.04 0.0007* 8.44 0.0176* diameter Genotype x Storage 1.19 0.3005 0.99 0.4869 1.11 0.3823 (mm) Time Average leaf Genotype 7.05 0.0424 9.41 0.03* 16.1 0.0236* number (#) Storage Time 33.8 0.0024* 0.62 0.7 5.07 0.0497* Genotype x Storage 1.15 0.3343 1.57 0.08 1.06 0.4089 Time Average Genotype 0.72 0.6222 0.93 0.53 1.26 0.4279 chlorophyll Storage Time 10.6 0.0211* 110.3 <0.0001* 2.66 0.1529 content (mg Genotype x Storage 3.97 <0.0001* 0.74 0.78 4.21 0.0004* g-2) Time *Significant at α=0.05.

152

Appendix 3.3 Type III test for the significance of main effects and their interactions (‘genotype’, ‘storage time’, and ‘genotype x storage time’) on plant growth traits for ‘Illinois’, ‘UK’, ‘BC’, and ‘Amuri’ after autumn 2015 (‘Illinois’ and ‘UK’ only), spring 2016, and autumn 2016 harvests and storage at 1°C (α=0.05). Autumn 2015 Spring 2016 Autumn 2016 Trait Effect(s) F-value Pr>F F-value Pr>F F-value Pr>F Average Genotype 0.01 0.9324 31.05 0.01* 30.1 0.0097* rhizome Storage Time 5.3 0.0455* 1.32 0.38 3.63 0.0919 viability (%) Genotype x Storage Time 1.36 0.2463 1.44 0.16 3.56 0.0002* Average Genotype 0.01 0.9229 3.84 0.15 0.64 0.6389 emergence Storage Time 7.09 0.0254* 5.47 0.04* 19.5 0.0027* speed (days) Genotype x Storage Time 6.03 0.0001* 5.98 <0.0001* 4.27 0.0003* Average tiller Genotype 0.5 0.6078 4.03 0.14 0.94 0.5199 height (mm) Storage Time 7.95 0.02* 4.21 0.07 12.8 0.0071* Genotype x Storage Time 1.6 0.1669 1.25 0.26 2.32 0.0256* Average tiller Genotype 0.79 0.5371 1.37 0.4 1.88 0.3092 number (#) Storage Time 1.43 0.3535 0.81 0.59 1.46 0.3451 Genotype x Storage Time 1.76 0.1321 1.1 0.38 4.82 0.0001* Average stem Genotype 2.3 0.3709 0.56 0.68 3.9 0.1467 node number (#) Storage Time 1.11 0.4571 2.34 0.19 1.37 0.4152 Genotype x Storage Time 1.31 0.296 1.27 0.25 0.93 0.533 Average stem Genotype 34.2 0.1078 2.67 0.22 0.36 0.7912 diameter (mm) Storage Time 1.73 0.2808 13.01 0.01* 5.34 0.0448 Genotype x Storage Time 2.99 0.017* 1.61 0.1 2.82 0.0091* Average leaf Genotype 0.87 0.523 7.27 0.07 9.99 0.0453* number (#) Storage Time 19.9 0.0026* 1.57 0.32 28.8 0.0011* Genotype x Storage Time 1.83 0.1135 0.96 0.51 3.22 0.0006* Average Genotype 12.5 0.1754 0.13 0.9340 1.96 0.2976 chlorophyll Storage Time 10.3 0.0116* 45.84 0.0004* 2.1 0.2176 content (mg g-2) Genotype x Storage Time 1.07 0.3857 2.47 0.0069* 5.45 <0.0001* *Significant at α=0.05.

153

Appendix 3.4 Type III test for the significance of main effects and their interactions (temperature, storage time, and temperature x storage time) on rhizome physiological traits for ‘Illinois’ after spring 2015 harvest and storage at 0 or 3°C (α=0.05). Spring 2015 Trait Effect(s) F-value Pr>F Average rhizome FW (g) Temperature 0.04 0.8751 Storage Time 1.13 0.4489 Temperature x Storage Time 5.5 0.0012* Average rhizome DW (g) Temperature 0.21 0.7236 Storage Time 0.87 0.5566 Temperature x Storage Time 5.59 0.0011* Average rhizome MC (%) Temperature 0.09 0.8178 Storage Time 0.69 0.651 Temperature x Storage Time 8.38 <0.0001* Average starch content [%, w/w (DW Temperature 0.13 0.7805 basis)] Storage Time 0.91 0.5389 Temperature x Storage Time 7.26 0.0002* Total NSC content (µg g-1 DW) Temperature 61.59 0.0807 Storage Time 1.27 0.4012 Temperature x Storage Time 0.61 0.6898 Average D-glucose (% total soluble Temperature 1.17 0.4745 carbohydrate concentration) Storage Time 0.47 0.7868 Temperature x Storage Time 0.41 0.8347 Average sucrose (% total soluble Temperature 1.9 0.3994 carbohydrate concentration) Storage Time 0.9 0.5429 Temperature x Storage Time 1.19 0.344 Average D-fructose (% total soluble Temperature 1.53 0.4329 carbohydrate concentration) Storage Time 1.19 0.4267 Temperature x Storage Time 1.1 0.3869 *Significant at α=0.05.

154

Appendix 3.5 Type III test for the significance of main effects and their interactions (genotype, storage time, and genotype x storage time) on rhizome physiological traits for ‘Illinois’ (autumn 2015 and spring 2016 only), ‘UK’, ‘BC’, ‘Amuri’, and ‘Nagara’ after autumn 2015, spring 2016, and autumn 2016 harvests and storage at 3°C (α=0.05). Autumn 2015 Spring 2016 Autumn 2016 Trait Effect(s) F-value Pr>F F-value Pr>F F-value Pr>F Average rhizome FW Genotype 66.99 6E-04* 136.8 0.0002* 19.65 0.007* (g) Storage Time 2.07 0.249 12.61 0.0154* 16.8 0.009* Genotype x Storage 2.61 0.003* 1.66 0.0789 28.04 <0.0001* Time Average rhizome DW Genotype 55.6 0.0009* 172.17 0.0007* 60.66 8E-04* (g) Storage Time 12.56 0.016* 10.54 0.0212* 3.16 0.145 Genotype x Storage 4.99 <0.0001* 1.65 0.1074 2.12 0.023* Time Average rhizome MC Genotype 3.15 0.146 8.12 0.0596 36.78 0.002* (%) Storage Time 7.24 0.041* 7.25 0.0405* 8.27 0.032* Genotype x Storage 5.92 <0.0001* 0.65 0.7936 12.84 <0.0001* Time Average starch Genotype 1.8 0.408 10.43 0.0428* 21.71 0.006* content [%, w/w (DW Storage Time 4.03 0.103 4.14 0.0989 2.86 0.167 basis)] Genotype x Storage 0.09 0.986 0.97 0.4854 1.88 0.048* Time Total NSC content Genotype 11.24 0.019* 11.25 0.0189* 41.6 0.002* (µg g-1 DW) Storage Time 0.11 0.974 1.02 0.4932 3.63 0.12 Genotype x Storage 1.66 0.082 1.77 0.0553 1.91 0.046* Time Average D-glucose Genotype 6 0.055 5.56 0.0627 5.32 0.067 (% total soluble Storage Time 0.48 0.755 1.65 0.3189 2.65 0.184 carbohydrate Genotype x Storage 2.46 0.006* 0.98 0.4925 2.22 0.019* concentration) Time Average sucrose (% Genotype 6.83 0.045* 2.01 0.2573 17.1 0.009* total soluble Storage Time 2.54 0.194 3.37 0.1332 0.03 0.997 carbohydrate Genotype x Storage 4.33 <0.0001* 1.35 0.1947 0.8 0.669 concentration) Time Average D-fructose Genotype 9.23 0.027* 2.31 0.219 32.04 0.003* (% total soluble Storage Time 2.57 0.191 1.73 0.3034 1.05 0.483 carbohydrate Genotype x Storage 3.45 3E-04* 1.44 0.1508 0.81 0.652 concentration) Time *Significant at α=0.05.

155

Appendix 3.6 Type III test for the significance of main effects and their interactions (genotype, storage time, and genotype x storage time) on rhizome physiological traits for ‘Illinois’, ‘UK’, ‘BC’, and ‘Amuri’ after autumn 2015 (‘Illinois’ and ‘UK’ only), spring 2016, and autumn 2016 harvests and storage at 1°C (α=0.05). Autumn 2015 Spring 2016 Autumn 2016 Trait Effect(s) F-value Pr>F F-value Pr>F F-value Pr>F Average rhizome FW Genotype 99.69 0.0635 279.75 0.0004* 29.17 0.0101* (g) Storage Time 1.59 0.3319 6.65 0.0468* 3.51 0.1259 Genotype x Storage 0.48 0.7488 1.91 0.0546 2.57 0.0097* Time Average rhizome DW Genotype 84.87 0.0688 172.17 0.0007* 20.38 0.0169* (g) Storage Time 7.88 0.0352* 10.54 0.0212* 4.42 0.0895 Genotype x Storage 2.97 0.0409* 1.65 0.1074 3.18 0.0019* Time Average rhizome MC Genotype 4.33 0.2851 8.12 0.0596 40.4 0.0063* (%) Storage Time 182.27 <0.0001* 7.25 0.0405* 11.59 0.0179* Genotype x Storage 3.06 0.037* 0.65 0.7936 6.7 <0.0001* Time Average starch content Genotype 1.8 0.4075 10.43 0.0428* 22.34 0.0149* [%, w/w (DW basis)] Storage Time 4.03 0.1029 4.14 0.0989 8.27 0.0324* Genotype x Storage 0.09 0.9859 0.97 0.4854 1.19 0.314 Time Total NSC content (µg Genotype 3.38 0.3172 20.62 0.0166* 11.32 0.0383* g-1 DW) Storage Time 4.32 0.0927 0.24 0.9017 13.4 0.0138* Genotype x Storage 1.08 0.3945 2.27 0.0218* 2.3 0.0215* Time Average D-glucose (% Genotype 5.42 0.2584 1.83 0.3161 10.05 0.0449* total soluble Storage Time 4.46 0.0882 9.3 0.0265* 1.17 0.441 carbohydrate Genotype x Storage 7.07 0.001* 2.71 0.0069* 1.01 0.4515 concentration) Time Average sucrose (% Genotype 0.05 0.8538 2.53 0.2332 8.1 0.0597 total soluble Storage Time 13.94 0.0128* 9.73 0.0244* 1.38 0.382 carbohydrate Genotype x Storage 14.26 <0.0001* 3.61 0.0007* 1.14 0.3565 concentration) Time Average D-fructose (% Genotype 1.69 0.4172 2.9 0.2027 6.09 0.086* total soluble Storage Time 7.42 0.039* 4.64 0.0832 2.26 0.2245 carbohydrate Genotype x Storage 4.55 0.0089* 2.77 0.006* 1.43 0.1857 concentration) Time *Significant at α=0.05.

156

Appendix 3.7 Parameters estimates results for the change in emergence speed of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 27.77 2.5350 <0.0001 Storage time -9.87 1.9318 <0.0001 (Storage time)2 1.22 0.3159 0.0013 2 2 Regression model y=1.22x 2-9.87x+27.77, Adj. R =0.76

Appendix 3.8 Parameter estimates results for the change in emergence speed of ‘Illinois’ harvested in spring 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 27.60 2.7286 <0.0001 Storage time -12.91 2.5666 <0.0001 (Storage time)2 2.11 0.4927 0.0003 Regression model y=2.11x2-12.91x+27.6, Adj. R2=0.54

Appendix 3.9 Parameter estimates results for the change in emergence speed of ‘UK’ harvested in spring 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 15.64 0.9901 <0.0001 Storage time -4.48 0.9313 <0.0001 (Storage time)2 0.71 0.1788 0.0007 Regression model y=0.71x2-4.48x+15.64, Adj. R2=0.53

Appendix 3.10 Parameter estimates results for the change in emergence speed of ‘Illinois’ harvested in autumn 2015 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 25.29 1.5942 <0.0001 Storage time -2.61 0.5266 <0.0001 Regression model y=-2.61x+25.29, R2=0.53

Appendix 3.11 Parameter estimates results for the change in emergence speed of ‘UK’ harvested in autumn 2015 and spring 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 37.28 2.5128 <0.0001 Storage time -4.63 0.7576 <0.0001 Regression model y=-4.63x+37.28, R2=0.67

157

Appendix 3.12 Parameter estimates results for the change in tiller height of ‘Illinois’ harvested in spring 2015 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 927.82 71.2398 <0.0001 Storage time -130.06 19.7584 <0.0001 Regression model y=-130.06x+927.82, R2=0.63

Appendix 3.13 Parameter estimates results for the change in tiller height of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 650.69 87.9558 <0.0001 Storage time 112.28 26.5197 0.0005 Regression model y=112.29x+650.69, R2=0.50

Appendix 3.14 Parameter estimates results for the change in tiller height of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 221.00 55.6348 0.0006 Storage time 138.39 18.3756 <0.0001 Regression model y=138.39x+221.00, R2=0.72

Appendix 3.15 Parameter estimates results for the change in tiller height of ‘Amuri’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 296.72 56.1436 <0.0001 Storage time 91.47 18.5436 <0.0001 Regression model y=91.47x+296.72, R2=0.53

Appendix 3.16 Parameter estimates results for the change in tiller height of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 26.71 82.5854 0.7494 Storage time 148.69 27.2771 <0.0001 Regression model y=148.69x+26.71, R2=0.57

158

Appendix 3.17 Parameter estimates results for the change in tiller height of ‘Illinois’, ‘UK’ ‘BC’, and ‘Amuri’ harvested in spring 2016 and stored at 1 or 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 927.82 71.2398 <0.0001 Storage time -130.06 19.7584 <0.0001 Regression model y=-130.06x+927.82, R2=0.63

Appendix 3.18 Parameter estimates results for the change in tiller number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 1.14 0.2385 0.0001 Storage time -0.35 0.2243 0.1295 (Storage time)2 0.12 0.04306 0.0091 Regression model y=0.12x2-0.35x+1.14, Adj. R2=0.51

Appendix 3.19 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2015 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 2.49 0.3251 <0.0001 Storage time 0.44 0.09801 0.0003 Regression model y=0.44x+2.49, R2=0.50

Appendix 3.20 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 2.39 0.3603 <0.0001 Storage time 0.64 0.1190 <0.0001 Regression model y=0.64x+2.39, R2=0.57

Appendix 3.21 Parameter estimates results for the change in leaf number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 2.13 0.3214 <0.0001 Storage time 0.63 0.1062 <0.0001 Regression model y=0.63x+2.13, R2=0.62

Appendix 3.22 Parameter estimates results for the change in leaf number of ‘Amuri’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 1.95 0.3272 <0.0001 Storage time 0.68 0.1081 <0.0001 Regression model y=0.68x+1.95, R2=0.64

159

Appendix 3.23 Parameter estimates results for the change in leaf number of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 0.86 0.7271 0.2478 Storage time 1.21 0.2401 0.0001 Regression model y=1.21x+0.86, R2=0.50

Appendix 3.24 Parameter estimates results for the change in stem node number of ‘Nagara’ harvested in autumn 2015 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept -0.52 0.2633 0.0632 Storage time 0.53 0.07937 <0.0001 Regression model y=0.53x-0.52, R2=0.71

Appendix 3.25 Parameter estimates results for the change in stem node number of ‘Amuri’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept -0.15 0.2148 0.5018 Storage time 0.38 0.07093 <0.0001 Regression model y=0.38x-0.15, R2=0.56

Appendix 3.26 Parameter estimates results for the change in stem node number of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 0.86 0.7271 0.2478 Storage time 1.21 0.2401 0.0001 Regression model y=1.21x+0.86, R2=0.50

Appendix 3.27 Parameter estimates results for the change in stem node number of ‘BC’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 0.21 0.2891 0.4794 Storage time -0.45 0.2720 0.1131 (Storage time)2 0.17 0.05221 0.0044 Regression model y=0.17x2-0.45x+0.21, Adj. R2=0.59

160

Appendix 3.28 Parameter estimates results for the change in stem diameter of ‘Amuri’ harvested in spring 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 5.68 0.2336 <0.0001 Storage time -1.50 0.2197 <0.0001 (Storage time)2 0.20 0.04218 0.0001 Regression model y=0.20x2-1.50x+5.68, Adj. R2=0.79

Appendix 3.29 Parameter estimates results for the change in stem diameter of ‘Nagara’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 2.26 0.3611 <0.0001 Storage time 0.56 0.1193 0.0001 Regression model y=0.56x+2.26, R2=0.50

Appendix 3.30 Parameter estimates results for the change in stem diameter of ‘UK’ harvested in autumn 2015 and spring 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 6.45 0.5641 <0.0001 Storage time -1.61 0.4299 0.0016 (Storage time)2 0.20 0.07029 0.0129 Regression model y=0.20x2-1.61x+6.45, Adj. R2=0.64

Appendix 3.31 Parameter estimates results for the change in chlorophyll content of ‘Amuri’ harvested in spring 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 612.68 13.7805 <0.0001 Storage time -107.39 12.7583 <0.0001 (Storage time)2 15.39 2.4758 <0.0001 Regression model y=15.39x2-107.39x+612.68, Adj. R2=0.85

Appendix 3.32 Parameter estimates results for the change in chlorophyll content of ‘BC’ harvested in spring 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 565.12 10.9419 <0.0001 Storage time -73.24 10.3270 <0.0001 (Storage time)2 9.77 1.9983 0.0001 Regression model y=9.77x2-73.24x+565.12, Adj. R2=0.82

161

Appendix 3.33 Parameter estimates results for the change in total NSC of ‘UK’ harvested in autumn 2016 and stored at 3⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 657093 108983 <0.0001 Storage time -232997 83052 0.0122 (Storage time)2 27520 13581 0.0587 Regression model y=27520x2-232997x+657093, Adj. R2=0.51

Appendix 3.34 Parameter estimates results for the change in percent total soluble carbohydrates consisting of D-glucose of ‘UK’ harvested in autumn 2015 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 48.24 5.7537 <0.0001 Storage time -7.91 1.7348 0.0002 Regression model y=-7.91x+48.24, R2=0.54

Appendix 3.35 Parameter estimates results for the change in total NSC of ‘UK’ harvested in autumn 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 973413 126134 <0.0001 Storage time -392308 96122 0.0008 (Storage time)2 47849 15718 0.0073 Regression model y=47849x2-392308x+973413, Adj. R2=0.67

Appendix 3.36 Parameter estimates results for the change in percent total soluble carbohydrates consisting of D-glucose of ‘UK’ harvested in spring 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept 65.59 10.0602 <0.0001 Storage time -25.66 7.6665 0.0038 (Storage time)2 3.06 1.2536 0.0259 Regression model y=3.06x2-25.66x+65.59, Adj. R2=0.60

Appendix 3.37 Parameter estimates results for the change in percent total soluble carbohydrates consisting of sucrose of ‘BC’ harvested in spring 2016 and stored at 1⁰C (α=0.05). Variable Parameter estimates S.E. p-value Intercept -25.65 15.5103 0.1165 Storage time 39.73 11.8199 0.0037 (Storage time)2 -3.90 1.9328 0.0600 Regression model y=-3.90x2+39.73x-25.65, Adj. R2=0.74

162

Appendix 3.38 Average rhizome viability (%), culm emergence speed (days), tiller height (mm), and tiller number of rhizomes stored at 0 and 3°C over 126 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’) in May 2015. Measurements were taken 6 weeks after planting. Each value is represented by 32±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 6 Average 88 ± 68.9 ± 56.1 ± 75.5 ± 65.7 ± 53.4 ± 31.3 ± rhizome 10.0375 10.039 10.0401 10.0369 10.0392 10.0389 10.0384 viability Average culm 14.5 ± 15.9 ± 12 ± 13.8 ± 13.3 ± 9.2 ± 8.4 ± emergence 2.8476 2.8787 2.8628 2.8346 2.8637 2.881 2.8742 speed (days) Average tiller 792.8 ± 906.1 ± 488.5 ± 732.1 ± 486.9 ± 476.2 ± 275.2 ± height 115.06 115.08 115.07 115.07 115.07 115.07 115.07 (mm) Average 1.1 ± 1.2 ± 0.98 ± 1.2 ± 1.4 ± 1.1 ± 0.92 ± tiller 0.2449 0.4256 0.246 0.2445 0.2456 0.2456 0.2452 number

163

164

Appendix 3.40 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively. Measurements were taken 6 weeks after planting. Each value is represented by 80±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 28.1 ± 60.9 ± 64.3 ± 69.1 ± 68.13 ± 64 ± rhizome - 8.605 6.7402 6.7614 6.7441 6.7379 6.7554 viability Average culm 12.6 ± 9.7 ± 10 ± 11.1 ± 8.7 ± 8.8 ± emergence - 2.6536 0.6078 0.6096 0.6078 0.6078 0.608 speed (days) Average 292.1 ± 537.9 ± 572.8 ± 611.6 ± 830.2 ± 835.3 ± tiller height - 88.6914 69.0235 69.1035 69.024 69.0235 69.0249 (mm) Average 1.2 ± 1.3 ± 1.4 ± 1.4 ± 1.5 ± 1.2 ± tiller - 0.1705 0.1015 0.1016 0.1016 0.1015 0.1016 number Average 1 ± leaf 2.9 2.9 2.9 5 4.2 0.2462 0.5825 number Average 0.17 ± stem node 0.3 0.07 0.3 0.9 1.3 0.1621 0.2257 number Average stem 4.2 ± 4.3 ± 3.8 ± 4.9 ± 4.6 ± 2.3 - diameter 0.2617 0.2619 0.2619 0.2616 0.2624 (mm) Average chlorophyll 249.2 ± 502.3 ± 514.4 ± 523.9 ± 582.5 ± 505.2 ± - content 59.435 20.72 20.8201 20.7419 20.7023 20.8064 (mg m-2)

165

Appendix 3.41 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in November and December 2015, respectively. Each value is represented by 80±SE samples. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 50.7 43.9 53.9 55.3 48.4 1.8712 (%) Average starch content [%, 22.83 ± 16.76 ± 20.4 ± 16.52 ± 18 ± 1.1856 - w/w (DW 1.1856 1.1856 1.1856 1.2183 basis)] Total soluble carbohydrate 201.77 ± 185.72 ± 183.47 ± 191.57 ± 185.37 ± - concentration 24.400 24.208 25.825 24.414 25.341 (mg g-1 DW) D-glucose (% total soluble 28.47 ± 23.77 ± 24.31 ± 25.7 ± 25 ± 3.5149 - carbohydrate 3.5149 3.4547 3.5149 3.6358 concentration) Sucrose (% total soluble 48.69 ± 38.88 ± 56.29 ± 50.71 ± 49.46 ± - carbohydrate 4.7302 4.7303 4.6558 4.7304 4.8742 concentration) D-fructose (% total soluble 26.56 ± 32.72 ± 19.94 ± 25.51 ± 24.17 ± - carbohydrate 3.0273 3.0274 2.9756 3.0274 3.1255 concentration)

166

Appendix 3.42 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015. Measurements were taken 6 weeks after planting. Each data point is represented by 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 28.1 ± 43.9 ± 65.3 ± 62.6 ± 53.5 ± 40.7 ± rhizome 8.605 8.8064 8.6048 8.6065 8.606 8.606 viability Average culm 12.6 ± 26.8 ± 24.1 ± 21.9 ± 17.3 ± 12 ± emergence - 2.6536 2.654 2.6536 2.654 2.6539 2.6538 speed (days) Average 292.1 ± 356.5 ± 412.7 ± 412.8 ± 705.9 ± 432.1 ± tiller height - 88.6914 88.748 88.6893 88.748 88.7256 88.732 (mm) Average 1.2 ± 0.88 ± 1.3 ± 1.2 ± 1 ± 1.3 ± tiller - 0.1705 0.1706 0.1705 0.1704 0.1704 0.1705 number Average 1 ± 2.3 ± 3.2 ± 3 ± 4.9 ± 3.2 ± leaf - 0.5825 0.5826 0.5825 0.5826 0.5826 0.5826 number Average 0.17 ± 0.05 ± 0.057 ± 0.66 ± 0.48 ± stem node 0 - 0.2257 0.2557 0.2557 0.2558 0.2558 number Average stem 2.3 3.5 4.7 3.1 4.3 3.6 0.5209 diameter (mm) Average chlorophyll 249.2 ± 461.1 ± 506.8 ± 507.6 ± 607.9 ± 395.8 ± - content 59.435 59.4491 59.4354 59.4491 59.445 59.4439 (mg m-2)

167

Appendix 3.43 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015. Each data point is represented by 32±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 54.3 33 44 51.1 48.2 2.2425 (%) Average starch content 28.8 15.18 19.39 16.14 15.27 3.7624 [%, w/w (DW basis)] Total soluble carbohydrate 399.54 ± 436.64 ± 249.49 ± 204.15 ± 240.05 ± - concentration 43.823 41.949 24.962 30.802 46.793 (mg g-1 DW) D-glucose (% total soluble 34.85 ± 41.3 ± 24.86 ± 19.81 ± 19.7 ± - carbohydrate 4.3217 4.5195 4.3217 4.3217 4.4078 concentration) Sucrose (% total soluble 33.8 ± 14.77 ± 49.81 ± 53.93 ± 55.19 ± - carbohydrate 4.2036ab 4.5803b 4.2036a 4.2036a 4.377a concentration) D-fructose (% total soluble 31.36 ± 45.3 ± 25.34 ± 26.26 ± 23.07 ± - carbohydrate 3.4875ab 3.614a 3.4875ab 3.4875ab 3.5932b concentration)

168

Appendix 3.44 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016. Each value is represented by 80±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 63.7 ± 75.2 ± 64.9 ± 73.6 ± 66.3 ± 68.7 ± rhizome - 6.3739 6.3738 6.3739 6.3739 6.3725 6.3735 viability Average culm 5 ± 6.1 ± 5.1 ± 5 ± 4.6 ± 4.3 ± emergence - 0.2362 0.2361 0.2362 0.2362 0.236 0.2361 speed (days) Average 770.4 ± 834.6 ± 755.2 ± 778.3 ± 723.4 ± 689.9 ± tiller height - 58.707 58.7057 58.707 58.707 58.6913 58.7019 (mm) Average tiller 1.2 1.4 1.4 1.2 1.4 1.3 0.1199 number Average 4.3 ± 4.3 ± 4.6 ± 4.1 ± 4.6 ± 4.4 ± leaf - 0.2855 0.2854 0.2855 0.2855 0.2851 0.2854 number Average 0.8 ± 1.2 ± 1.3 ± 1 ± 1.3 ± stem node 1 ± 0.208 - 0.208 0.208 0.208 0.2079 0.208 number Average stem 4.5 ± 5 ± 4 ± 3.4 ± 3.5 ± 3.4 ± - diameter 0.2535 0.2535 0.2535 0.2535 0.5233 0.2534 (mm) Average chlorophyll 599.2 ± 361.4 ± 319.9 ± 436.6 ± 436.3 ± 428.2 ± - content 37.7121 37.7118 37.7121 37.7122 37.707 37.7108 (mg m-2)

169

Appendix 3.45 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016. Each value is represented by 80±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 57 51.8 56.2 53 53 1.8712 (%) Average starch content 25.15a 19.34ab 23ab 16.97ab 14.46b 1.3512 [%, w/w (DW basis)] Total soluble carbohydrate 210.79 ± 203.48 ± 206.64 ± 222.50 ± 239.05 ± - concentration 20.823 21.464 21.464 20.823 20.823 (mg g-1 DW) D-glucose (% total soluble 18.58 ± 21.41 ± 13.01 ± 15.88 ± 13.99 ± - carbohydrate 2.7347 2.8142 2.8142 2.7347 2.7347 concentration) Sucrose (% total soluble 62.79 ± 53.89 ± 71.62 ± 62.08 ± 68.52 ± - carbohydrate 4.5899 4.6559 4.6559 4.5899 4.5899 concentration) D-fructose (% total soluble 18.63 ± 24.87 ± 15.3 ± 22.04 ± 17.49 ± - carbohydrate 2.7361 3.5163 3.0425 3.6717 2.3547 concentration)

170

Appendix 3.46 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each value is represented by 64±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 58 ± 62.2 ± 71.4 ± 63.1 ± 55.1 ± 64.1 ± rhizome - 5.7802 5.7769 5.7732 5.7850 5.7799 5.7805 viability Average culm 5.2 ± 5.4 ± 4.8 ± 4.5 ± 4.2 ± 4.3 ± emergence - 0.3309 0.3309 0.3307 0.331 0.3309 0.3308 speed (days) Average 760 ± 578.8 ± 733.4 ± 855.8 ± 556.9 ± 612.7 ± tiller height - 57.2119 57.2116 57.2088 57.2148 57.2119 57.212 (mm) Average tiller 1.2 1.2 1.3 1.3 1.2 1.2 0.1206 number Average 4.3 ± 3.9 ± 5.1 ± 4.9 ± 3.9 ± 4.3 ± leaf - 0.3624 0.3624 0.3623 0.3625 0.3624 0.3624 number Average 0.85 ± 0.71 ± 1.1 ± 1.7 ± 0.9 ± 0.9 ± stem node - 0.2495 0.2496 0.2493 0.2498 0.2495 0.2495 number Average stem 4.8 4.5 3.9 3.5 2.9 3 0.3265 diameter (mm) Average chlorophyll 594.3 ± 247.9 ± 327 ± 428.5 ± 432.8 ± 419.5 ± - content 44.9764 44.9695 44.961 44.9874 44.9744 44.9746 (mg m-2)

171

Appendix 3.47 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April 2016. Each value is represented by 64±SE samples. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 53.4 41.8 54.9 53.2 51.6 1.7292 (%) Average starch content 22.24 13.54 22.16 19.71 19.3 2.1465 [%, w/w (DW basis)] Total soluble carbohydrate 299.97 ± 262.06 ± 249.97 ± 241.34 ± 238.34 ± - concentration 54.410 33.269 42.425 40.517 40.199 (µg g-1 DW) D-glucose (% total soluble 24.78 ± 25.98 ± 10.05 ± 15.63 ± 13.82 ± - carbohydrate 2.617 4.0137 1.744 2.9565 2.9182 concentration) Sucrose (% total soluble 45.65 ± 40.69 ± 68.79 ± 61.03 ± 67.32 ± - carbohydrate 4.5347ab 4.4001b 4.4001a 4.5347ab 4.4001a concentration) D-fructose (% total soluble 29.2 ± 33.33 ± 21.16 ± 23.15 ± 18.86 ± - carbohydrate 3.0653 2.983 2.983 3.0653 2.983 concentration)

172

Appendix 3.48 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 3°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively. Each value is represented by 64±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 38.8 ± 43.1 ± 42.3 ± 44.9 ± 41.5 ± 45.6 ± rhizome - 5.6556 5.6563 5.613 5.6444 5.593 5.601 viability Average culm 18 ± 19.6 ± 20.1 ± 13.7 ± 12.4 ± 12 ± emergence - 2.0979 2.0979 2.0957 2.0973 2.0943 2.0947 speed (days) Average 134.5 ± 309.5 ± 391 ± 525.5 ± 456.2 ± 646.1 ± tiller height - 69.3146 69.3224 69.0565 69.2454 68.8918 68.942 (mm) Average 0.61 ± 0.8 ± 1 ± 0.86 ± 1 ± 1.5 ± tiller - 0.1761 0.1761 0.1745 0.1757 0.1739 0.1742 number Average 1.2 ± 2.6 ± 3.2 ± 4 ± 3.2 ± 4.6 ± leaf - 0.5767 0.5768 0.5567 0.5734 0.5598 0.5634 number Average 0.07 ± 0.25 ± 0.73 ± 0.71 ± 1.6 ± stem node 0 - 0.2316 0.2316 0.2316 0.2315 0.2316 number Average stem 2.4 ± 4.1 ± 3.6 ± 4.4 ± - - - diameter 0.4032 0.4264 0.4417 0.4251 (mm) Average chlorophyll 276.7 ± 345.5 ± 411.7 ± 411.2 ± 318.2 ± 406.3 ± - content 42.2729 42.2709 40.9932 42.0179 41.0474 41.2823 (mg m-2)

173

Appendix 3.49 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively. Each value is represented by 64±SE samples. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 49.5 52 49.4 41 47.8 1.8712 (%) Average starch content 18.2 21.1 15.7 16.7 21.8 1.7695 [%, w/w (DW basis)] Total soluble carbohydrate 327.47 227.79 216.71 228.98 211.72 31.204 concentration (mg g-1 DW) D-glucose (% total soluble 21.7 20.1 24.6 22.6 23.3 3.0420 carbohydrate concentration) Sucrose (% total soluble 45.9 41.3 47.4 44.5 42.1 4.2871 carbohydrate concentration) D-fructose (% total soluble 32.4 32.4 28 26.7 22.1 3.1118 carbohydrate concentration)

174

Appendix 3.50 Average rhizome viability (%), culm emergence speed (days), tiller height (mm) and number, leaf number, stem diameter (mm) and node number, and chlorophyll content (mg m- 2) of rhizomes stored at 1°C over 140 days. Measurements were taken 6 weeks after planting. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in December and November 2016, respectively. Each value is represented by 64±SE samples. Timepoint (28-day intervals) 0 1 2 3 4 5 S.E. Average 41.3 ± 47.4 ± 59.8 ± 52.4 ± 44.3 ± 51 ± rhizome - 3.0344 4.0186 3.0331 4.2946 2.9033 3.6662 viability Average culm 14.9 ± 12.3 ± 17.8 ± 12.9 ± 10.7 ± 12.5 ± emergence - 1.4442 1.3594 1.2127 1.1193 1.8746 1.5838 speed (days) Average 431.5 ± 390.2 ± 578.2 ± 713.6 ± 455.1 ± 636.7 ± tiller height - 32.3356 42.2049 36.5263 51.6685 53.4362 50.0282 (mm) Average 0.8 ± 0.86 ± 1.2 ± 1.1 ± 0.98 ± 1.2 ± tiller - 0.0665 0.0665 0.0665 0.09191 0.1365 0.106 number Average 2.6 ± 2.6 ± 4.3 ± 4.3 ± 3.4 ± 4.7 ± leaf - 0.2491 0.3107 0.3015 0.2932 0.4361 0.3651 number Average 0.44 ± 0.39 ± 0.79 ± 1.3 ± 0.72 ± 1.1 ± stem node - 0.173 0.173 0.1951 0.1854 0.2219 0.1738 number Average stem 2.9 ± 3.3 ± 3.2 ± 2.4 ± 3.3 ± 2.9 ± 0.21 - diameter 0.3394 0.1988 0.218 0.2974 0.2663 (mm) Average chlorophyll 387.6 ± 325.3 ± 431.1 ± 437.3 ± 356.4 ± 420.7 ± - content 27.7933 38.0861 38.1545 31.3383 33.1387 33.6727 (mg m-2)

175

Appendix 3.51 Average MC (%), starch content [%, w/w (DW basis)], total soluble carbohydrate concentration (mg g-1 DW), and percentage (%) of total soluble carbohydrate concentration consisting of D-glucose, sucrose, and D-fructose of rhizomes stored at 1°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in December and November 2016, respectively. Each value is represented by 64±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Timepoint (28-day intervals) 1 2 3 4 5 S.E. Average MC 54.8 53.3 48 51.4 52.9 1.7292 (%) Average starch content 22.82 ± 20.59 ± 13.96 ± 14.86 ± 14.91 ± - [%, w/w (DW 1.5415 1.511 1.5412 1.511 1.5718 basis)] Total soluble carbohydrate 446.16 ± 268.43 ± 242.04 ± 188.37 ± 233.23 ± - concentration 23.442a 23.442ab 23.442b 22.609b 24.246b (mg g-1 DW) D-glucose (% total soluble 25.94 ± 23.92 ± 30.6 ± 21.09 ± 28.09 ± - carbohydrate 3.2356 3.235 3.2356 3.1225 3.3438 concentration) Sucrose (% total soluble 36.36 ± 42.33 ± 31.5 ± 49.01 ± 46.26 ± - carbohydrate 6.1517 6.1517 6.1517 5.9833 6.3221 concentration) D-fructose (% total soluble 38.09 ± 33.6 ± 37.65 ± 29.91 ± 25.72 ± - carbohydrate 3.6558 3.6557 3.6558 3.5731 3.7417 concentration)

176

Appendix 3.52 Average tiller height (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015 and April 2016 and stored at 1 or 3°C for up to 140 days. Each data point is represented by 64±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

Appendix 3.53 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in April 2016 and stored at 3°C for up to 140 days. Each data point is represented by 16±SE samples. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

177

Appendix 3.54 Average stem diameter (mm) of tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘UK’ and ‘BC’) and Elora, ON (‘Amuri’ and ‘Nagara’) in December and November 2016, respectively, and stored at 3°C for up to 140 days. Each bar is represented by 80±SE samples.

178

Appendix 3.55 Average stem diameter (mm) of the tallest tiller after six weeks of growth in greenhouse conditions. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’ and ‘UK’) in November 2015 and April 2016 and stored at 1 or 3°C for up to 140 days. Measurements are categorized by a) harvest season and b) individual genotypes. Individual data points and bars are represented by 64 and 160 ± SE samples, respectively. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test.

179

180

Appendix 3.57 Total soluble carbohydrate concentration (mg g-1 DW) of rhizomes stored at 1 and 3°C over 140 days. Rhizomes were harvested from Port Ryerse, ON (‘Illinois’, ‘UK’, and ‘BC’) and Elora, ON (‘Amuri’) in April and November/December 2016. Each bar is represented by 80±SE samples.

181

Appendix 4.1 Average GAE (µg g-1 DW) values of calli cultured on mediums supplemented with various AIP levels (S.E. 149.03). Each bar represents the mean of 50 calli over two 2,4-D levels (10 callus samples over five replications). Bars labeled with the same letter are not significantly different (P<0.05) according to Tukey's HSD test Appendix 4.2 Total MC (%) of calli used for soluble phenolic content measurements. Analyzes were conducted separately for AIP and 2,4-D. Each mean represents 50 and 100 calli (10 and 25 callus samples over five replications, respectively). Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD test. AIP (µM) 2,4-D (µM) Concentration 0 1 10 100 1000 9.0 11.3 Mean (%) 84.62 ab 84.95 ab 83.98 b 84.88 ab 87.31 a 84.43 B 85.87 A S.E. 0.7715 0.4879

182

Appendix 5.2 Total weight (g) of plantlets destined for destructive measurements over the duration of the MR induction period. Because mean differences were significant between trials, both experimental trials are exhibited. Each mean represents four replicate plantlets, and the analysis was conducted using a RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Weight (g) – initial Weight (g) – week 5 Weight (g) – week 10 Treatment Trial 1 Trial 2 Trial 1 Trial 2 Trial 1 Trial 2 8%sucrose+PGR 0.55a 0.64a 3.3a 5.5a 2.9b 3.4ab CS-control 0.47a 0.46a 2.9a 2.4a 3.6ab 4.5ab CS-ABA 0.47a 0.46a 2.9a 2.4a 3.7ab 4.9a S.E. 0.08511 0.4858 0.4428 0.3716

Appendix 5.3 Average shoot length (mm), chlorophyll content (mg m-2), and shoot chlorosis (%) after 10 weeks of MR induction treatment with (CS-control and ABA) and without (8%sucrose+PGR) four weeks of 3°C CS. Each mean represents four replicate plantlets (repeated twice), and the analysis was conducted using a RCBD. Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Average shoot Chlorophyll content Shoot chlorosis (%) length (mm) (mg m-2) 8%sucrose+PGR 56a 365.85a 0.095a CS-control 41.4b 263.07b 0b CS-ABA 41.3b 265.73b 0b S.E. 3.2382 9.3841 0.1123

183

Appendix 5.4 Weight (g) and weight change (%) of full plantlets after four weeks of 3°C CS. The analysis was conducted using a split-plot design (repeated twice), with freezer acting as the main plot (replicated twice) and flasks acting as the subplots (replicated four times). Means followed by the same letter are not significantly different (P<0.05) according to Tukey's HSD mean separation test. Weight (g) – after CS Weight change (%) – after CS Treatment Trial 1 Trial 2 Trial 1 Trial 2 CS-control 3.8a 3.8a -7a -13.3a CS-ABA 3.9a 4.2a -4.9a -9.6a S.E. 0.6265 4.9172

Appendix 5.5 Raw data of various MR treatments that developed full plantlets in vitro after 8 weeks of culture in tillering media. Due to sparse germination and development of plantlets, statistical analysis was not conducted on data. Trial # 1 2 Replicate # 3 4 1 1 2 Treatment 8%sucrose CS-C CS-ABA CS-ABA CS-control + PGR Germination (%) 12.5 12.5 25 12.5 12.5 Tiller number (#) 5 1 11.5 4 5 Leaf number (#)a 4 2 4 2 2 Tiller length (mm)b 118.8 45.4 55.6 38 90 Chlorophyll content (mg m-2)c 369.67 - 336.63 - 360.67 Plantlet weight (g) 0.9631 0.2714 1.05715 0.5747 0.5664 aLeaf number of longest tiller only. bTiller length of longest tiller only. cChlorophyll content averaged over 15 individual readings.

Appendix 5.6 Raw data of 8%sucrose+PGR treatment that developed full plantlets ex vitro after 6 weeks of culture in a mist bed. Due to sparse germination and development of plantlets, statistical analysis was not conducted on data. Trial # 1 2 Replicate # 2-1 1-1 2-1 3-1 4-1 4-2 Tiller number (#)a 1 1 1 2 2 1 Leaf number (#) 3 4 3 3 3 2 Tiller length (mm)b 48 83.3 113.7 35.1 31 71.1 Root number (#) 5 2 4 10 3 3 Stem diameter (mm) 0.92 1.32 0.98 1.03 0.68 1.17 Chlorophyll content (mg m-2)c 410.8 367.6 304.4 309.3 510.95 Plantlet weight (g) - 0.1397 0.1906 0.1662 0.2179 aLeaf number of longest tiller only. bTiller length of longest tiller only. cChlorophyll content averaged over 15 individual readings.

184

A) B)

C) D)

F) E) Appendix 5.7 Growth assessment of MRs after four weeks of 3°C CS. CS-D (A and B) and CS-C (C and D) samples after 8 weeks of incubation in tillering media; healthy, intact bud on MR (E); tillering media containing MRs that were dried (blue labels) and not dried (green labels) after fours weeks of CS (F). Scale bar is 1 mm.

185