DEVELOPMENTAL RESPONSES TO POLYCOMB COMPLEX MUTATIONS AND PLANTING DENSITY IN HEXAPLOID WHEAT By Selena Lopez A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Plant Breeding, Genetics and Biotechnology – Crop and Soil Sciences – Master of Science 2021 ABSTRACT DEVELOPMENTAL RESPONSES TO POLYCOMB COMPLEX MUTATIONS AND PLANTING DENSITY IN HEXAPLOID WHEAT By Selena Lopez Hexaploid wheat (Triticum aestivum) has undergone two hybridization events resulting in severe genetic bottlenecks, particularly in the D genome contributed by the diploid progenitor species, Aegilops tauschii. One method to regain this genetic diversity is through interspecific hybrids with diploid progenitors. This requires embryo rescues to recover progeny due to endosperm failure of the hybrids. Previous studies in Arabidopsis and rice have demonstrated that mutations in imprinted Polycomb Repressive Complex 2 (PRC2) genes medea (MEA), fertilization independent endosperm (FIE), and fertilization independent seed 2 (FIS2) can, or potentially can, restore endosperm function to interploidy hybrids. Homologs of MEA, FIE, and FIS2 were identified in the wheat A, B and D genomes. Pyramids of double and triple mutants were developed using a combination of Kompetative allele specific PCR (KASP) markers and amplicon sequencing. Single, double, and triple mutants at MEA, FIE, and FIS2 associated with the three subgenomes were hybridized with Ae. tauschii to identify mutant combinations that may restore endosperm function. The wheat breeding program at Michigan State University utilizes a modified bulk breeding method to rapidly advance populations from the F2 to F4 stage under greenhouse conditions. To determine the influence of planting density on yield components and plant architecture in greenhouse settings, two experiments were carried out using two spring and winter wheat. Experiments with both spring and winter wheat demonstrated that increased planting density reduces plant height, spikelets per spike, grain length and grain width. TABLE OF CONTENTS LIST OF TABLES .......................................................................................................................... v LIST OF FIGURES ....................................................................................................................... vi CHAPTER I .....................................................................................................................................1 INTRODUCTION ...........................................................................................................................1 Modern wheat ......................................................................................................................2 Aegilops tauschii (D genome)..............................................................................................3 Endosperm formation and Ae. tauschii introgression ..........................................................4 Polycomb repressive complex 2 ..........................................................................................5 PRC2 genes ..........................................................................................................................6 Paternally expressed genes ..................................................................................................7 Imprinting ............................................................................................................................7 KASP marker usage .............................................................................................................8 Amplicon sequencing...........................................................................................................9 Planting density..................................................................................................................10 BIBLIOGRAPHY ..........................................................................................................................11 CHAPTER II..................................................................................................................................16 DEVELOPMENTAL CHANGES IN POLYCOMB REPRESSIVE COMPLEX 2 (PRC2) MUTANTS ....................................................................................................................................16 Abstract ..............................................................................................................................17 Introduction ........................................................................................................................17 Materials and methods .......................................................................................................21 Identification of Arabidopsis PRC2 orthologs in wheat and rice ..........................21 Plant materials .......................................................................................................22 Plant growth conditions .........................................................................................22 Pyramiding PRC2 mutations .................................................................................26 Cycle 1 .......................................................................................................26 Cycle 2 .......................................................................................................26 Cycle 3 .......................................................................................................26 Cycle 4 .......................................................................................................27 Genotyping .............................................................................................................27 Amplicon sequencing .............................................................................................27 Phylogenetic trees ..................................................................................................31 Variant calling .......................................................................................................34 Cadenza mutants and Ae. tauschii hybridizations .................................................34 Results ................................................................................................................................38 Cycle 1, 2, and 3 genotyping with KASP markers .................................................38 Amplicon sequencing Cycle 3 ................................................................................40 Amplicon sequencing Cycle 4 ................................................................................41 Hybridizing single, double and/or triple mutant homeoallele(s) to Ae. tauschii ...42 iii Discussion ..........................................................................................................................45 BIBLIOGRAPHY ..........................................................................................................................49 CHAPTER III ................................................................................................................................53 DEVELOPMENTAL CHANGES IN WINTER AND SPRING WHEAT IN VARIOUS PLANTING DENSITIES ..............................................................................................................53 Abstract ..............................................................................................................................54 Introduction ........................................................................................................................54 Materials and methods .......................................................................................................56 Experimental design...............................................................................................56 Winter wheat ..........................................................................................................56 Spring wheat ..........................................................................................................58 Data collection .......................................................................................................58 Data analysis .........................................................................................................59 Results ................................................................................................................................59 Plant height ............................................................................................................60 Spikelets per spike ..................................................................................................62 Mature spikelets per spike .....................................................................................63 Spikes per pot .........................................................................................................65 Total seed weight ...................................................................................................66 Average seed weight ..............................................................................................68 Seed count ..............................................................................................................69 Seed length .............................................................................................................71 Seed width ..............................................................................................................73 Seeds per spike .......................................................................................................74 Seeds planted vs. number of spikes ........................................................................76 Planting density vs line ..........................................................................................77 Discussion ..........................................................................................................................78 Traits that decreased as planting density increased ..............................................79 Traits that increased as planting density increased ..............................................79 Traits that followed no trends as planting density increased ................................79 Planting density vs line ..........................................................................................80 Future implications ................................................................................................81 BIBLIOGRAPHY ..........................................................................................................................82 iv LIST OF TABLES Table 2.1: SNP, zygosity, SNP type, sift score, line names, gene names, and cDNA location of SNP for mutant donor parent ........................................................................................................24 Table 2.2: Gene components and properties of mutant donor parents ..........................................25 Table 2.3: KASP primer information.............................................................................................30 Table 2.4: Amplicon sequencing primers .....................................................................................32 Table 2.5: Grep sequences used for read 1 Cycle 3 samples. Reverse complement of these sequences was used for read 2 ......................................................................................................32 Table 2.6: Grep sequences used for Cycle 4 samples ....................................................................33 Table 2.7: Amplicon sequencing primer products used as reference genome in variant calling pipeline...........................................................................................................................................36 Table 2.8: Cycle 2 KASP genotyping results from selected offspring. ? indicates KASP results that were labeled “unknown” .........................................................................................................38 Table 2.9: Cycle 3 KASP genotyping results. Genotypes missing a subgenome had inconclusive KASP results. ** indicates KASP markers were not available for FIE B and D genome mutations ........................................................................................................................................................39 Table 2.10: Cycle 3 amplicon sequencing genotypic distribution .................................................40 Table 2.11: Cycle 4 genotyping results..........................................................................................41 Table 2.12: Cycle 4 chi square results ..........................................................................................42 Table 2.13: Caryopses yield from single, double, and/or triple mutant Cadenza lines hybridized with Ae. tauschii. Wild Cadenza x wild Cadenza and wild Cadenza x Ae. tauschii hybridizations were made as controls ....................................................................................................................43 Table 2.14: Average seeds per spike of the different genotypes of MEA, FIE, and FIS2 resulting from hybridizations between Ae. tauschii ......................................................................................44 Table 3.1: Seed density treatments ................................................................................................56 Table 3.2: P-values of independent variables planting density, line, and the interaction between planting density and line. * indicates significance at P<0.05, ** indicates significance at P<0.01 ........................................................................................................................................................59 Table 3.3: Adjusted R squared values for planting density and line. Bolded values indicate adjusted R square values that were greater for line than planting density for traits that were significant for both planting density and line. Underlined values indicate traits that were significant for line only and had larger adjusted R squared values than planting density .............78 v LIST OF FIGURES Figure 1.1: Hybridizations that formed hexaploid wheat ................................................................3 Figure 2.1: Pyramid diagram for Cycles 1-4 .................................................................................29 Figure 2.2: KASP segregation for mea-A, mea-B, mea-D, fie-A, and fis2-D in Cycle 3 genotyping. Blue squares represent homozygous mutant samples, green triangles represent heterozygous samples, orange circles represent wild type samples, and red X’s represent unknown genotypes. Samples in black circles contain wild controls ............................................39 Figure 2.3: Phylogenetic trees of MEA, FIE, and FIS2 genes ......................................................44 Figure 3.1: Average plant height of spring wheat and winter wheat for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error .............60 Figure 3.2: Spikelets per spike of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Spikelets per spike include both mature spikelets with seeds and immature spikelets without seeds. Letters are LSD values for best fitted model. Bars are standard error ....62 Figure 3.3: Mature spikelets per spike of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error ................................................................................................................................................63 Figure 3.4: Spikes per pot of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for the best fitted model. Bars are standard error ...............65 Figure 3.5: Total seed weight of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error .............66 Figure 3.6: Average seed weight of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error.........68 Figure 3.7: Seed count of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Black dots on plot represent mean seed count for each planting density ................69 Figure 3.8: Seed length of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error .....................71 Figure 3.9: Seed length of spring wheat lines for planting densities 2 to 39.5 seeds/10cm2. Black dots on plot represent mean seed count for each planting density .................................................72 Figure 3.10: Seed width of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for the best fitted model. Bars are standard error ...............73 Figure 3.11: Seeds per spike of spring wheat and winter wheat lines for planting densities of 2 seeds/10cm2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error ................................................................................................................................................74 vi Figure 3.12: Seeds planted vs number of spikes for spring wheat and winter wheat lines for planting densities 32 to 640 seeds per pot. Letters are LSD values for best fitted model. Bars are standard error ................................................................................................................................76 vii CHAPTER I INTRODUCTION 1 Modern wheat Wheat (Triticum aestivum, 2n=6x=42, AABBDD) is a food source that billions of people rely on for calories and nutrition. In 2019, 35,468 hg/ha of wheat was harvested from 2.159e8 hectares worldwide (FAOSTAT 2020). Wheat accounts for 20% of protein and 28% of calories consumed in people’s diets (Velu et al., 2016). As the second most produced cereal in the world, wheat is a major component in the agricultural market (Velu et al., 2016). Modern day wheat is an allohexaploid consisting of an A, B, and D genomes (AABBDD), which contain a total of 17Gb of DNA sequences. The A genome is most related to Triticum urartu (2n = 2x = 14), the B genome is a close relative of Aegilops speltoides (2n = 14x, SS), and the D genome is Aegilops tauschii (2n=2x=14, DD) or goat grass (Ling et al., 2013); these genomes were merged over two separate hybridization events (Figure 1.1). The first hybridization event occurred 300,000-500,000 years ago between the A and B genome forming Triticum turgidum (AABB) (Huang et al., 2002). Tetraploid wheat (T. turgidum, AABB) is used in modern agriculture for pasta as well as breeding synthetic hexaploid wheat (Zhang et al., 2008). The second hybridization occurred between tetraploid wheat (AABB) and Ae. tauschii (DD) to form modern hexaploid wheat, T. aestivum (AABBDD) (Mcfadden and Sears 1946). 2 Figure 1.1: Hybridizations that formed hexaploid wheat. Aegilops tauschii (D genome) Ae. tauschii (2n=2x=14), or goat grass, has a center of origin in Iran containing a mixture of subspecies and lineages of Ae. tauschii (Singh et al. 2019). The subspecies recognized as most related to the D genome of hexaploid wheat, Ae. tauschii ssp. strangulata, ranges from Transcaucasia to eastern Caspian Iran (Wang et al., 2013). Studies have indicated that other lines outside of this subspecies, including ssp. tauschii, are also significantly related to the D subgenome of hexaploid wheat (Dvorak et al., 1998). Two lineages of Ae. tauschii exist: Ae. tauschii ssp. tauschii belonging in lineage 1 (L1) and Ae. tauschii ssp. strangulata belonging in lineage 2 (L2) (Mizuno et al. 2010). These lineages most likely diverged based on geographic location. Genetic analysis determined that L2 accessions generally originate in western regions, and L1 accessions generally originate in eastern regions such as Afghanistan and Pakistan (Takumi et al., 2008). 3 Only a few plants of Ae. tauschii hybridized with wheat, resulting in a very narrow genetic background of the wheat D genome (Lagudah et al., 1991). By analyzing the geographic and genetic diversity of Ae. tauschii, desired traits from specific Ae. tauschii accessions can be introgressed into hexaploid wheat. Genes conferring disease or insect resistance to leaf rust, stem rust, yellow rust, Hessian fly, greenbug, cereal leaf beetle, root lesion nematode, karnal bunt, soil borne mosaic virus, and powdery mildew have been successfully introgressed to hexaploid wheat from Ae. tauschii (Miranda et al., 2006; Olson et al., 2013; Cox et al., 2017; Saluja et al., 2018). In recent years, Pm58, a powdery mildew resistance gene, was successfully introgressed into hexaploid wheat from Ae. tauschii (Wiersma et al., 2017). Future genes of interest can be exploited in Ae. tauschii, such as temperature adaptations in the differing altitudes of L2 (Singh et al., 2019). L1 accessions are also a valuable source of genetic diversity as L1 only contributes 2.7% introgression from Ae. tauschii in United States winter wheat lines (Singh et al., 2019). Endosperm formation and Ae. tauschii introgression Double fertilization is an essential process for most angiosperms. Two male gametes are necessary: one to combine with the female egg cell, and the other to combine with the two polar nuclei. The latter forms the endosperm portion of the seed. Endosperm is the primary food source for the germinating embryo. The resulting combination of two polar nuclei with the sperm is a triploid (3n) endosperm. This 2:1 ratio of maternal and paternal genetic material is required for proper formation of endosperm. There are two methods of introgressing Ae. tauschii into hexaploid wheat: direct hybridization and synthetic hexaploid wheat. In direct hybridizations, the endosperm ratio is severely off balance from the 2:1 maternal to paternal dosage in the endosperm. A hexaploid 4 wheat and Ae. tauschii direct hybridization would have a 6:1 maternal to paternal dosage resulting in failed endosperm formation and eventual seed abortion. Research involving these direct hybridizations typically employ embryo rescue and subsequent tissue culture of the rescued young embryo (Shen et al., 2011). Synthetic hexaploid wheat is created by hybridizing T. turgidum (AABB) with Ae. tauschii (DD) and then undergoes spontaneous or induced chromosome doubling (Cox et al., 2017). Synthetic hexaploid wheat lines are often spring type and non-free threshing (Cox et al., 2017). Additionally, changes can occur in the genome structure and/or copy numbers of genes in synthetic hexaploids (Li et al., 2018). Polycomb repressive complex 2 A major mechanism responsible for endosperm formation is Polycomb Repressive Complex 2 (PRC2), one of several polycomb group complexes (Holec and Berger, 2012). These complexes were first discovered through research in segmentation of Drosophila melanogaster (Lewis, 2004). Later research determined homologs of PRC2 are also found in mammals, angiosperms, and some unicellular organisms (Margueron and Reinberg, 2011). PRC2 is not entirely conserved as some yeast species do not have the complex (Margueron and Reinberg, 2011). PRC2 catalyzes trimethylation of histone H3 on lysine 27 (H3K27me3) allowing epigenetic activity to occur in select vegetative and reproductive development (Weinhofer et al., 2010). Homology searches using online resources were carried out between Arabidopsis PRC2 genes and wheat. Based on these searches, it was found that the wheat PRC2 complex consists of three genes: fertilization independent endosperm (FIE), fertilization independent seed (FIS2), and Medea (MEA); these genes are maternally expressed (Köhler et al., 2003a). 5 PRC2 genes Several studies have researched the actions of PRC2 genes. One found that increasing maternal dosage in Arabidopsis can bypass endosperm failure (Kradolfer et al., 2013). The study analyzed MEA and FIS2 mutations. Both MEA and FIS2 mutants formed enlarged seeds, some of which were viable and can be germinated. Imprinting regulation involves both FIS2 and MEA (Köhler and Makarevich, 2006). These imprinted genes are silenced differently in Arabidopsis. MEA is silenced by H3K27 trimethylation from PRC2 and FIS2 is silenced through DNA methylation or, more specifically, methyltransferase 1 (MET1) (Jullien et al., 2008). The methylation FIS2 experiences is also controlled by the retinoblastoma pathway as it represses MET1 when the female gamete is forming (Jullien et al., 2008). Medea (MEA) interacts with DNA methyltransferase (MET1) to repress autonomous endosperm in Arabidopsis (Schmidt et al., 2013). MET1 is not orthologous to wheat, therefore altering the function of MEA alone may impose autonomous endosperm formation in wheat. MEA is known to down-regulate maternal alleles during fertilization and silences paternal alleles during seed formation (Baroux et al., 2006). Specifically, MEA represses expression of the PHERES1 (PHE1) MADS box gene as well as other genes that are expressed during the first stages of endosperm formation (Jarillo et al., 2009). Both MEA and FIE associate with the promoter of PHE1, meaning PHE1 expression is likely epigenetically regulated by PRC2 proteins (Köhler et al., 2003b). MEA will bind to its own promoter prior to fertilization (Baroux et al., 2006). This binding might not include any other PRC2 proteins, due to unchanged MEA levels in FIS2 and FIE mutants. FIE proteins are in a WD repeat protein family (Nallamilli et al., 2013). The WD repeats form a beta propeller structure, a circular structure that creates surface loops used as the basis of 6 protein complex formation (Yadegari et al., 2000). This family promotes protein-protein interactions. FIE functions in seed development, grain filling, and may also function in male gametophyte development (Nallamilli et al., 2013). In Rice, the FIE protein contains a retinoblastoma protein (RB) region (Nallamilli et al., 2013). Paternally expressed genes Paternally expressed genes (PEGs) commonly have disrupted expression in interspecific hybrids (Kirkbride et al., 2015). These disruptions are due to methylation, MEA, and ploidy differences, and result in seed abortion. In Arabidopsis, crosses between a diploid female and a hexaploid male resulted in upregulation of PEGs and PHE1 and aborted seed (Kirkbride et al., 2015). Upregulation is also found in interspecific hybrids. The same phenotype is observed between crosses of a hypomethylated parent and a normal diploid plant (Adams et al., 2000). Altering PEG expression has been shown to produce viable seed in interploidy crosses in Arabidopsis (Wolff et al., 2015). Mutations in genes PEG2 and PEG9 resulted in decreased gene expression and subsequent endosperm cellularization. Seed size can be altered by PEGs due to increased ploidy in uneven ploidy crosses (FitzGerald et al., 2008). For example, a tetraploid male crossed with a diploid female will likely form larger seed than a tetraploid female crossed with a diploid male. This is, however, only if the ploidy differences are not too great in that seed formation does not occur. Imprinting Parents face a dilemma when determining the best way to maximize their offspring’s survival: size of offspring versus number of offspring (Smith and Fretwell, 1974). It is an 7 evolutionary trade off determined in a tug of war between the male and female parents. The female parents tend to distribute resources evenly among offspring to increase the numbers of offspring (Haig, 2013). The male parents try to increase resources for the embryo and increase its size. One of the reasons the male parent would want a larger embryo size is because there is a possibility of other sibling seeds being fertilized by a different male plant (Haig and Westoby, 1989). Imprinting is evolutionarily conserved, but the mechanisms and genes may differ. Yang et al. (2018) found that genes controlling imprinting are not well conserved between A. thaliana and monocots. Additionally, they found that a large percentage of imprinted genes in diploid or tetraploid wheat are also imprinted in hexaploid wheat. This is not surprising, as diploid and tetraploid wheat are ancestors to hexaploid wheat. These results are useful in understanding that imprinting, although it is widespread, it is not well conserved outside of related species. KASP marker usage Kompetitive allele-specific PCR (KASP) has allowed breeders to determine the genotypic diversity of plant germplasm quickly and affordably. These markers can be used for trait assessment ranging from disease resistance to improving eating and cooking quality (Meade et al., 2019; Yang et al., 2019). KASP markers determine if a single nucleotide polymorphism (SNP) and indels are present via fluorescence-based genotyping technology (Semagn et al., 2014). Other single-marker methods to visualize SNPs have also been used; cleaved amplified polymorphic sequences (CAPS), allele-specific PCR (AS-PCR), and re-sequencing (Rasheed et al., 2016). KASP markers are preferred over others due to their cost efficiency, high throughput, and reduce labor inputs. These markers require a forward primer containing the desired SNP, a 8 reverse common primer, isolated DNA, and KASP master mix. All of these components undergo multiple rounds of PCR to allow amplification of the desired sequence region, as well as the attachment of fluorescently labelled primers in the KASP master mix. These fluorescent primers allow sequences with the targeted SNP to be visualized. Although KASP markers are great for visualizing SNPs, there is potential for off-target amplification. Amplicon sequencing Amplicon sequencing allows researchers to sequence a small, known section of DNA approximately 200-500bp long. This type of sequencing has been implemented in polyploid species due to a lack of homoallele specificity when using markers (Balázs and Cowling, 2010). Homoallele specificity is crucial when isolating a specific mutation, such as a SNP, insertion, or deletion. Without this specificity, regions that do not include the desired mutation will be amplified. These off-target regions cloud marker results by indicating the mutation is present when it is not. Additionally, it is difficult to determine if a sample is homozygous or heterozygous for the desired mutation with these off-target regions. Amplicon sequencing removes this uncertainty by allowing researchers to directly compare the sequences to the reference genome and determine if the anticipated mutation is present. Amplicon sequencing, along with RNA-seq, has made locating mutations in complex polyploid genomes more accurate (Li et al., 2019). 9 Planting density The wheat breeding program at Michigan State University has implemented a modified bulk method, Minibulk, that rapidly advances crosses from the F2 to F4 stage (Olson et al., 2021). The Minibulk process uses relatively high planting densities of 300 seeds at F2 and then 500 seeds at F3 in 8x8 inch square pots (Olson et al., 2021). Minimal research has been conducted for grain planting density in greenhouse settings. Cisar et al. (1982) examined planting densities of oats in greenhouse settings. The goal of the study was to determine the optimal planting density that would result in the greatest number of spikes with few or single seeds for the use of single seed descent (Cisar et al., 1982). Additionally, the oats were grown in stressed conditions without fertilizer to ensure few seeds will develop per spike (Cisar et al., 1982). This differs from the goals of the wheat breeding program at Michigan State University. 10 BIBLIOGRAPHY 11 BIBLIOGRAPHY Adams S, Vinkenoog R, Spielman M, Dickinson HG, Scott RJ (2000) Parent-of-origin effects on seed development in Arabidopsis thaliana require DNA methylation. Development 127: 2493–2502 Balázs E, Cowling WA (2010) Exploiting genome-wide association in oilseed Brassica species. Genome 53: 853–855 Baroux C, Gagliardini V, Page DR, Grossniklaus U (2006) Dynamic regulatory interactions of Polycomb group genes: MEDEA autoregulation is required for imprinted gene expression in Arabidopsis. Genes Dev 20: 1081–1086 Cisar G, Howey AE, Brown CM (1982) Optimal Population Density and Random vs. Selective Elimination of Genotypes Under a Modified Single-Seed Descent Method with Spring Oats1. Crop Sci 22: cropsci1982.0011183X002200030034x Cox TS, Wu J, Wang S, Cai J, Zhong Q, Fu B (2017) Comparing two approaches for introgression of germplasm from Aegilops tauschii into common wheat. Crop J 5: 355– 362 Dvorak J, Luo M-C, Yang Z-L, Zhang H-B (1998) The structure of the Aegilops tauschii genepool and the evolution of hexaploid wheat. Theor Appl Genet 97: 657–670 FAOSTAT (2020) Food and agriculture organization of the United Nations. Food Agric data. http://www.fao.org/faostat/en/#data FitzGerald J, Luo M, Chaudhury A, Berger F (2008) DNA Methylation Causes Predominant Maternal Controls of Plant Embryo Growth. PLOS ONE 3: e2298 Haig D (2013) Kin Conflict in Seed Development: An Interdependent but Fractious Collective. Annu Rev Cell Dev Biol 29: 189–211 Haig D, Westoby M (1989) Parent-Specific Gene Expression and the Triploid Endosperm. Am Nat 134: 147–155 Holec S, Berger F (2012) Polycomb Group Complexes Mediate Developmental Transitions in Plants. Plant Physiol 158: 35 Huang S, Sirikhachornkit A, Su X, Faris J, Gill B, Haselkorn R, Gornicki P (2002) Genes encoding plastid acetyl-CoA carboxylase and 3-phosphoglycerate kinase of the Triticum Aegilops complex and the evolutionary history of polyploid wheat. Proc Natl Acad Sci 99: 8133 12 Jarillo JA, Piñeiro M, Cubas P, Martínez-Zapater JM, Jarillo JA, Piñeiro M, Cubas P, Martínez-Zapater JM (2009) Chromatin remodeling in plant development. Int J Dev Biol 53: 1581–1596 Jullien PE, Mosquna A, Ingouff M, Sakata T, Ohad N, Berger F (2008) Retinoblastoma and Its Binding Partner MSI1 Control Imprinting in Arabidopsis. PLOS Biol 6: e194 Kirkbride RC, Yu HH, Nah G, Zhang C, Shi X, Chen ZJ (2015) An Epigenetic Role for Disrupted Paternal Gene Expression in Postzygotic Seed Abortion in Arabidopsis Interspecific Hybrids. Mol Plant 8: 1766–1775 Köhler C, Hennig L, Bouveret R, Gheyselinck J, Grossniklaus U, Gruissem W (2003a) Arabidopsis MSI1 is a component of the MEA/FIE Polycomb group complex and required for seed development. EMBO J 22: 4804–4814 Köhler C, Hennig L, Spillane C, Pien S, Gruissem W, Grossniklaus U (2003b) The Polycomb-group protein MEDEA regulates seed development by controlling expression of the MADS-box gene PHERES1. Genes Dev 17: 1540–1553 Köhler C, Makarevich G (2006) Epigenetic mechanisms governing seed development in plants. EMBO Rep 7: 1223–1227 Kradolfer D, Hennig L, Köhler C (2013) Increased Maternal Genome Dosage Bypasses the Requirement of the FIS Polycomb Repressive Complex 2 in Arabidopsis Seed Development. PLOS Genet 9: e1003163 Lagudah ES, Appels R, Brown AHD, McNeil D (1991) The molecular–genetic analysis of Triticum tauschii, the D-genome donor to hexaploid wheat. Genome 34: 375–386 Lewis EB (2004) A Gene Complex Controlling Segmentation in Drosophila. In HD Lipshitz, ed, Genes Dev. Cancer Life Work Edw. B Lewis. Springer US, Boston, MA, pp 205–217 Li A, Liu D, Yang W, Kishii M, Mao L (2018) Synthetic Hexaploid Wheat: Yesterday, Today, and Tomorrow. Engineering 4: 552–558 Li W, Riday H, Riehle C, Edwards A, Dinkins R (2019) Identification of Single Nucleotide Polymorphism in Red Clover (Trifolium pratense L.) Using Targeted Genomic Amplicon Sequencing and RNA-seq. Front Plant Sci 10: 1257 Ling H-Q, Zhao S, Liu D, Wang J, Sun H, Zhang C, Fan H, Li D, Dong L, Tao Y, et al (2013) Draft genome of the wheat A-genome progenitor Triticum urartu. Nature 496: 87 Margueron R, Reinberg D (2011) The Polycomb complex PRC2 and its mark in life. Nature 469: 343–349 Mcfadden ES, Sears ER (1946) The origin of triticum spelta and its free-threshing hexaploid relatives. J Hered 37: 81–89 13 Meade F, Byrne S, Griffin D, Kennedy C, Mesiti F, Milbourne D (2019) Rapid Development of KASP Markers for Disease Resistance Genes Using Pooled Whole-Genome Resequencing. Potato Res. doi: 10.1007/s11540-019-09428-x Miranda LM, Murphy JP, Marshall D, Leath S (2006) Pm34: a new powdery mildew resistance gene transferred from Aegilops tauschii Coss. to common wheat (Triticum aestivum L.). Theor Appl Genet 113: 1497–1504 Mizuno N, Yamasaki M, Matsuoka Y, Kawahara T, Takumi S (2010) Population structure of wild wheat D-genome progenitor Aegilops tauschii Coss.: implications for intraspecific lineage diversification and evolution of common wheat. Mol Ecol 19: 999–1013 Nallamilli BRR, Zhang J, Mujahid H, Malone BM, Bridges SM, Peng Z (2013) Polycomb group gene OsFIE2 regulates rice (Oryza sativa) seed development and grain filling via a mechanism distinct from Arabidopsis. PLoS Genet 9: e1003322–e1003322 Olson EL, Noble A, Forsberg L, Brisco-McCann E (2021) The 'Minibulk' system for advancing winter cereal breeding programs. Manuscript in preparation. Olson EL, Rouse MN, Pumphrey MO, Bowden RL, Gill BS, Poland JA (2013) Simultaneous transfer, introgression, and genomic localization of genes for resistance to stem rust race TTKSK (Ug99) from Aegilops tauschii to wheat. Theor Appl Genet 126: 1179–1188 Rasheed A, Wen W, Gao F, Zhai S, Jin H, Liu J, Guo Q, Zhang Y, Dreisigacker S, Xia X, et al (2016) Development and validation of KASP assays for genes underpinning key economic traits in bread wheat. Theor Appl Genet 129: 1843–1860 Saluja M, Kaur S, Bansal U, Subhash Chander Bhardwaj, Chhuneja P (2018) Molecular mapping of linked leaf rust resistance and non-glaucousness gene introgressed from Aegilops tauschii Coss. in hexaploid wheat Triticum aestivum L. Plant Genet Resour 16: 82–88 Schmidt A, Wöhrmann HJP, Raissig MT, Arand J, Gheyselinck J, Gagliardini V, Heichinger C, Walter J, Grossniklaus U (2013) The Polycomb group protein MEDEA and the DNA methyltransferase MET1 interact to repress autonomous endosperm development in Arabidopsis. Plant J 73: 776–787 Semagn K, Babu R, Hearne S, Olsen M (2014) Single nucleotide polymorphism genotyping using Kompetitive Allele Specific PCR (KASP): overview of the technology and its application in crop improvement. Mol Breed 33: 1–14 Shen X, Gmitter FG, Grosser JW (2011) Immature Embryo Rescue and Culture. In TA Thorpe, EC Yeung, eds, Plant Embryo Cult. Methods Protoc. Humana Press, Totowa, NJ, pp 75–92 Singh N, Wu S, Tiwari V, Sehgal S, Raupp J, Wilson D, Abbasov M, Gill B, Poland J (2019) Genomic Analysis Confirms Population Structure and Identifies Inter-Lineage Hybrids in Aegilops tauschii. Front Plant Sci 10: 9 14 Smith CC, Fretwell SD (1974) The Optimal Balance between Size and Number of Offspring. Am Nat 108: 499–506 Takumi S, Mizuno N, Okumura Y, Kawahara T, Matsuoka Y (2008) Two major lineages of Aegilops tauschii Coss. revealed by nuclear DNA variation analysis. Velu G, Guzman C, Mondal S, Autrique JE, Huerta J, Singh RP (2016) Effect of drought and elevated temperature on grain zinc and iron concentrations in CIMMYT spring wheat. J Cereal Sci 69: 182–186 Wang J, Luo M-C, Chen Z, You FM, Wei Y, Zheng Y, Dvorak J (2013) Aegilops tauschii single nucleotide polymorphisms shed light on the origins of wheat D-genome genetic diversity and pinpoint the geographic origin of hexaploid wheat. New Phytol 198: 925– 937 Weinhofer I, Hehenberger E, Roszak P, Hennig L, Köhler C (2010) H3K27me3 Profiling of the Endosperm Implies Exclusion of Polycomb Group Protein Targeting by DNA Methylation. PLoS Genet 6: e1001152 Wiersma AT, Pulman JA, Brown LK, Cowger C, Olson EL (2017) Identification of Pm58 from Aegilops tauschii. Theor Appl Genet 130: 1123–1133 Wolff P, Jiang H, Wang G, Santos-González J, Köhler C (2015) Paternally expressed imprinted genes establish postzygotic hybridization barriers in Arabidopsis thaliana. eLife 4: e10074 Yadegari R, Kinoshita T, Lotan O, Cohen G, Katz A, Choi Y, Katz A, Nakashima K, Harada JJ, Goldberg RB, et al (2000) Mutations in the FIE and MEA Genes That Encode Interacting Polycomb Proteins Cause Parent-of-Origin Effects on Seed Development by Distinct Mechanisms. Plant Cell 12: 2367 Yang G, Chen S, Chen L, Gao W, Huang Y, Huang C, Zhou D, Wang J, Liu Y, Huang M, et al (2019) Development and utilization of functional KASP markers to improve rice eating and cooking quality through MAS breeding. Euphytica 215: 66 Yang G, Liu Z, Gao L, Yu K, Feng M, Yao Y, Peng H, Hu Z, Sun Q, Ni Z, et al (2018) Genomic Imprinting Was Evolutionarily Conserved during Wheat Polyploidization. Plant Cell 30: 37–47 Zhang W, Chao S, Manthey F, Chicaiza O, Brevis JC, Echenique V, Dubcovsky J (2008) QTL analysis of pasta quality using a composite microsatellite and SNP map of durum wheat. Theor Appl Genet 117: 1361–1377 15 CHAPTER II DEVELOPMENTAL CHANGES IN POLYCOMB REPRESSIVE COMPLEX 2 (PRC2) MUTANTS 16 Abstract Hexaploid wheat (Triticum aestivum) has undergone two hybridization events resulting in severe genetic bottlenecks, particularly in the D genome contributed by the diploid progenitor, Aegilops tauschii. One method to regain this genetic diversity is through interspecific hybridizations with diploid progenitors, which requires embryo rescue to recover progeny due to endosperm failure in seed development. Previous studies in Arabidopsis and rice have demonstrated that mutations in imprinted Polycomb Repressive Complex 2 (PRC2) genes medea (MEA), fertilization independent endosperm (FIE), and fertilization independent seed 2 (FIS2) can, or potentially can, restore endosperm function to interploidy hybrids. Homologs of MEA, FIE, and FIS2 were identified in the wheat A, B and D genomes. Pyramids of double and triple mutants of these genes were developed using a combination of KASP markers and amplicon sequencing. Single, double, and triple mutants at MEA, FIE, and FIS2 were hybridized with Ae. tauschii to identify mutant combinations that may restore endosperm function. Introduction Wheat (Triticum aestivum, 2n=6=42, AABBDD) is a food source that billions of people rely on for calories and nutrition. Wheat accounts for 20% of protein and 28% of calories consumed in people’s diets (Velu et al., 2016). In 2019, 35,468 hg/ha of wheat were harvested from 2.159e8 hectares worldwide (FAOSTAT 2020). As the second most produced cereal in the world, wheat is a major component in the agricultural market (Velu et al., 2016). Wheat is an allohexaploid consisting of A, B, and D genomes. The A genome is most closely related to Triticum urartu (2n = 2x = 14), the B genome is a relative of Aegilops 17 speltoides (2n = 14, SS), and the D genome is Aegilops tauschii (2n=2x=14, DD); these genomes were merged over two separate hybridization events (Ling et al., 2013). The first hybridization event occurred 300,000-500,000 years ago between the A and B genome progenitors, which resulted in Triticum turgidum (AABB) (Huang et al., 2002). The second hybridization occurred between Triticum turgidum (AABB) and Ae. tauschii (DD) to form modern hexaploid wheat, T. aestivum (AABBDD) (Mcfadden and Sears, 1946). Only a few Ae. tauschii plants hybridized with Triticum turgidum, resulting in a narrow genetic background of the wheat D genome in hexaploid wheat (Lagudah et al., 1991). In an effort to increase genetic diversity, Ae. tauschii has been used as a novel source of genetic diversity in hexaploid wheat for insect, disease, and abiotic stress resistance (Lagudah et al., 1991). Genes conferring disease or insect resistance to leaf rust, stem rust, yellow rust, Hessian fly, greenbug, cereal leaf beetle, root lesion nematode, karnal blunt, soil borne mosaic virus, powdery mildew, and others have been introgressed into hexaploid wheat from Ae. tauschii (Miranda et al., 2006; Olson et al., 2013; Cox et al., 2017; Wiersma et al., 2017; Saluja et al., 2018). Finding novel genes that are well adapted to undesirable conditions, drought, disease, heat, etc., is an important aspect of wheat breeding. Two primary methods have been used to introgress Ae. tauschii into hexaploid wheat: synthetic hexaploid wheat production and direct hybridization. Synthetic hexaploid wheat is created by hybridizing Triticum turgidum with Ae. tauschii and then undergoing spontaneous or induced chromosome doubling (Cox et al., 2017). Synthetic hexaploid wheat lines are often spring type and non-free threshing (Cox et al., 2017). Additionally, changes can occur in the genome structure and/or copy numbers of genes in synthetic hexaploids (Li et al., 2018). 18 Direct hybridization is possible between hexaploid wheat as the maternal parent and Ae. tauschii as the paternal parent. However, only the embryo develops in these hybrids, not endosperm. Thus, the embryo has to be rescued and subsequent tissue culture is necessary to generate viable offspring (Shen et al., 2011). The necessity of embryo rescue is potentially due to the imbalance of maternal to paternal dosages in double fertilization. Two male gametes are necessary in double fertilization: one to combine with the female egg cell, and the other to combine with two polar nuclei (Faure J.E., 2011). The latter forms the endosperm portion of the seed. The resulting combination of two polar nuclei with the one sperm cell is a triploid (3n) endosperm. This 2:1 ratio of maternal and paternal genetic material is required for proper formation of endosperm. In the direct hybridization procedure, the endosperm ratio is 6:1 maternal to paternal dosage gene dosage, which may cause the failure in endosperm development. A primary component of endosperm development is the Polycomb Repressive Complex 2 (PRC2), one of several polycomb group complexes (Holec and Berger, 2012). These complexes were first discovered through research in segmentation of Drosophila (Lewis, 2004). PRC2 catalyzes trimethylation of histone H3 on lysine 27 (H3K27me3) allowing epigenetic activity to occur in select vegetative and reproductive development (Weinhofer et al., 2010). PRC2 consists of several genes, three of which are fertilization independent endosperm (FIE), fertilization independent seed (FIS2), and medea (MEA) (Köhler et al., 2003). These genes are maternally expressed and are present among angiosperms including rice and wheat (Nallamilli et al., 2013). Several studies have researched the actions of PRC2 genes. Most notably, Kradolfer et al. (2013) found that increasing maternal dosage of mutant MEA or FIS2 in Arabidopsis can bypass endosperm failure in maternal tetraploid and paternal diploid crosses. These crosses produced 19 viable triploid seeds (Kradolfer et al., 2013). The maternal excess of mutant MEA or FIS2 formed endosperm at a slower pace of 10 days after pollination instead of the usual 6 days after pollination (Kradolfer et al., 2013). Other research has focused on the specific mechanisms of MEA, FIE, and FIS2. MEA is silenced by H3K27 trimethylation from PRC2 and FIS2 is silenced through DNA methylation or, more specifically, methyltransferase 1 (MET1) in Arabidopsis (Jullien et al., 2008). MEA interacts with DNA methyltransferase (MET1) to repress autonomous endosperm in Arabidopsis (Schmidt et al., 2013). Since MET1 is not orthologous to wheat, altering the function of MEA alone may impose autonomous endosperm formation in wheat. FIE proteins are in a WD repeat protein family (Nallamilli et al., 2013). The WD repeats form a beta propeller structure, a circular structure that creates surface loops used as the basis of protein complex formation (Yadegari et al., 2000). This family promotes protein-protein interactions. FIE functions in seed development, grain filling, and may also function in male gametophyte development (Nallamilli et al., 2013). In rice, the FIE protein contains a retinoblastoma protein region (Nallamilli et al., 2013). Based on these previous studies, it is expected that having one or several mutant copies of MEA, FIS2, or FIE genes in maternal hexaploid wheat may allow proper endosperm formation, and therefore viable seed, to occur in interploidy hybridizations of hexaploid wheat and Ae. tauschii. Since wheat is an allohexaploid comprised of three homoeologous genomes (AABBDD), three hypotheses can be tested. 1. One mutant homeoallele of MEA, FIS2, or FIE will restore viable endosperm formation in 6nmatX 2npat hybridizations. One mutant homeoallele of MEA, FIS2, or FIE will lead the central cell of the maternal hexaploid wheat to function as 4n, instead of 6n. The 4:1 20 maternal to paternal endosperm ratio may be close enough to the desired 2:1 maternal to paternal endosperm ratio to allow endosperm to form. 2. Two mutant homeoalleles of MEA, FIS2, or FIE will restore viable endosperm formation in 6nmatX 2npat hybridizations. Two mutant homeoalleles of MEA, FIS2, or FIE will lead the central cell of the maternal hexaploid wheat to function as 2n, instead of 6n restoring the appropriate 2:1 maternal to paternal endosperm ratio. 3. Three mutant homeoalleles of MEA, FIS2, or FIE will restore viable endosperm formation in 6nmatX 2npat hybridizations. Three mutant versions of MEA, FIS2, or FIE will result in autonomous endosperm formation. Components of each hypothesis are still ongoing; however, a pipeline has been established to readily identify one, two, and three mutant homeoalleles of MEA, FIS2, or FIE. In other words, single, double, and triple homozygous and heterozygous mutants. The results of this research will facilitate more efficient direct hybridizations between any hexaploid wheat and Ae. tauschii. Materials and methods Identification of Arabidopsis PRC2 orthologs in wheat and rice Arabidopsis PRC2 proteins MEA, FIE, and FIS2 were identified via arabidopsis.org. Arabidopsis PRC2 proteins identified are AT3G20740.1 for FIE, AT2G35670.1 for FIS2, and AT1G02580.1 for MEA. These proteins were then used in TBLASTn searches in the wheat genome (Gertz et al., 2006). Orthologous PRC2 genes identified in wheat were then used to search the EMS database John Innes Center. TBLASTn searches were also used to identify 21 orthologous PRC2 genes in rice. Rice PRC2 accessions identified and used for phylogenetic trees are AY456262.1 for FIE, AY321106.1 for FIS2, and AJ421722.3 for MEA. Plant materials Germplasm was obtained from the John Innes Center http://www.wheat-tilling.com/. These Cadenza-derived lines were treated with ethyl methanesulfonate (EMS) to create random mutations in the genome. Several Cadenza lines were identified that contained a SNP in MEA, FIS2, or FIE homeoalleles. A total of 9 lines were used in crosses to combine the mutations of different homeoalleles (Table 2.1). These lines contain a transcript altering mutation of either a non-sense SNP or missense SNP in the targeted homeoallele (Table 2.1). Missense mutations are also paired with a Sorting Intolerant From Tolerant (SIFT) score, or a prediction of the amino acid change impacting protein function (Ng and Henikoff, 2003). All Cadenza lines with missense mutations have a sift score less than 0.05, which is categorized as a deleterious substitution (Ng and Henikoff, 2003). Transcription information was gathered for all homeoalleles of MEA, FIS2, and FIE (Table 2.2). All homeoalleles except fie-B has expression in the endosperm according to expression atlas. Plant growth conditions Cadenza lines were germinated at 4.4C for five days on 200 M weight White Verigood Blotter paper (Anchor Paper Company, St. Paul, Minnesota) saturated with 1% Thiram (T24201, Sigma Aldrich, St. Louis, MO) and DI water solution. Seeds were transplanted to Gro Pro Premium Black Square pots (6 in. x 6 in. x 8 in.) with saturated Sure Mix potting media 22 (Michigan Grower Products Inc., Galesburg, MI). A subdue application was added after transplanting. Ae. tauschii seeds were germinated in clear plastic boxes with paper saturated with 1% Thiram solution for 6-10 weeks. 1% Thiram was periodically applied throughout vernalization. After vernalization, Ae. tauschii seedlings were transplanted the similarly to Cadenza lines. All germplasm was grown in greenhouse conditions with fertilizer (20-20-20 NPK) used weekly (approximately 188g / 5-gal H2O at 1:50 dilution). Fertilization may have occurred more often as needed. 23 Homeoallele Zygosity SNP type sift Line Previous ensemble Previous Current Current scores Name gene name ensembl ensembl gene cDNA cDNA name position position fis2-A Heterozygous AA 0.02 Cadenza Traes_5AL_0C459 1387 TraesCS5A02G 1643 change 1069 8F51.1 179600 fis2-B Heterozygous Stop NA Cadenza Traes_5BL_96AA 286 TraesCS5B02G 2265 codon 1722 D5F61.1 177400 fis2-D Heterozygous Stop NA Cadenza Traes_5DL_600B6 974 TraesCS5D02G 849 codon 1699 FCDC.1 184200 mea-A Homozygous Stop NA Cadenza Traes_7AS_D9C6 785 TraesCS7A02G 1087 codon 0092 AC2A0.2 128600 mea-B Heterozygous AA 0 Cadenza Traes_7BS_C79E 175 TraesCS7B02G 1963 change 1722 BD23E.1 028500 mea-D Heterozygous AA 0.02 Cadenza Traes_7DS_23A2 225 TraesCS7D02G 2291 change 1182 E8EBB.1 127400 fie-A Heterozygous Stop NA Cadenza Traes_7AL_7CC9 811 TraesCS7A02G 1562 codon 1611 69251.1 308300 fie-B Homozygous Stop NA Cadenza Traes_7BL_B7F78 487 TraesCS7B02G 650 codon 1379 475F.1 377900LC fie-D Homozygous Stop NA Cadenza Traes_7DL_DF97 484 TraesCS7D02G 938 codon 1456 791D5.1 305100 Table 2.1: SNP, zygosity, SNP type, sift score, line names, gene names, and cDNA location of SNP for mutant donor parent. 24 Homeoallele Protein Exons Coding Domains Transcript Translation Transcripts Paralogs exons and length length features (base (amino pairs) acids) fis2-A TraesCS5A02G179600.1 22 21 7 2471 638 1 3 fis2-B TraesCS5B02G177400.1 21 20 7 2620 619 4 3 TraesCS5B02G177400.2 23 22 7 2734 657 4 3 TraesCS5B02G177400.3 22 21 7 2674 637 4 3 TraesCS5B02G177400.4 20 20 7 2233 618 4 3 fis2-D TraesCS5D02G184200.1 15 14 3 1487 393 2 2 TraesCS5D02G184200.2 22 21 7 2521 638 2 2 mea-A TraesCS7A02G128600.1 16 16 21 2821 801 2 33 TraesCS7A02G128600.2 10 10 6 1548 515 2 33 mea-B TraesCS7B02G028500.1 15 15 21 2595 792 4 33 TraesCS7B02G028500.2 16 16 21 2625 802 4 33 TraesCS7B02G028500.3 11 11 8 1569 522 4 33 TraesCS7B02G028500.4 11 11 8 1611 536 4 33 mea-D TraesCS7D02G127400.1 16 16 18 3005 802 0 0 fie-A TraesCS7A02G308300.1 14 14 23 2092 473 0 0 fie-B TraesCS7B02G377900L 13 13 NA 1470 389 0 0 C.1 fie-D TraesCS7D02G305100.1 14 14 23 1841 485 0 0 Table 2.2: Gene components and properties of mutant donor parents. 25 Pyramiding PRC2 mutations There were four cycles in the crossing design (Figure 2.1). Primary mutants were used in pyramiding crosses. All crosses were made using the approach method (Curtis and Croy, 1958). Cycle 1: Cadenza mutant lines were genotyped using KASP markers. The first cycle consisted of crossing a Cadenza mutant of one homeolog with a Cadenza mutant for a different homeolog (Figure 2.1). Cycle 2: The Cycle 1 progeny were Cadenza double mutant lines that have two heterozygous mutations in two homeologs (Figure 2.1). The progeny was expected to follow a 100% heterozygosity rate for the two mutant homeologs. Progeny from Cycle 1 were genotyped using KASP markers to confirm heterozygosity for two mutations. At the end of the Cycle 2, double mutant lines with complementary mutations were crossed to generate Cycle 3. Cycle 3: Progeny from Cycle 2 were genotyped via amplicon sequencing. MEA and FIE homeologs were also genotyped using KASP markers. Genotyping identified lines heterozygous for either three heterozygous mutations in three homeologs (AaBbDd), or two heterozygous mutations in two homeologs and a homozygous mutation in one homeolog (AaBbdd, aaBbDd, or AabbDd) (Figure 2.1). AaBbDd genotypes were expected to occur at 12.5% for, and AaBbdd, aaBbDd, or AabbDd genotypes at 6.25% These target genotypes were selected and self-pollinated to generate Cycle 4. 26 Cycle 4: Progeny from Cycle 3 were genotyped via amplicon sequencing. Genotyping screened for homozygous Cadenza triple mutant lines (aabbdd) and homozygous Cadenza double mutant lines (aaB_dd, A_bbdd, and aabbD_). Cycle 3 progeny lines were expected to occur at 6.25% or 3.125% homozygous triple mutants (aabbdd). Genotyping KASP markers were obtained through the John Innes Center (Table 2.3). The location of the SNP for Cadenza1379 mutant fie-B was not optimal for creating a KASP marker. KASP genotyping was conducted at the beginning of each cycle, prior to any crossing or self- pollination. Amplicon sequencing Amplicon sequencing was conducted at the beginning of cycle 3 and cycle 4. Mutation donor lines and wild type Cadenza were used as controls. Primers were created with Primer1 for each SNP and included a Fluidigm CS1/CS2 universal oligomers at their 5' end. Product sizes ranged from 200bp to 300bp in length (Table 2.4). Samples were amplified using PCR protocol: 95°C (5 min), followed by 35 cycles of 95°C (45 sec) 60°C (45 sec), 72°C (90 sec), and elongation at 72°C (10 min). Size selection and primer dimer removal was conducted with Omega Bio-Tek’s Mag-Bind TotalPure NGS. Samples were then quantified, normalized, and sequenced on an Illumina MiSeq v2 in 2x250 bp paired ends. Cycle 3 samples had an average 27 number of read pairs per sample of approximately 58,020 and ranged from 10,858 to 98,574 reads. Cycle 4 had an average number of read pairs per sample of approximately 10,200 with 96% of samples falling within 6,750 to 13,850 read pairs. 28 Figure 2.1: Pyramid diagram for Cycles 1-4. 29 Homeoallele Wild Type Primer Alternative primer Common Primer fis2-A ACCATGTAATGATTGAGCA ACCATGTAATGATTGAGCAGG ATTTCCAGATATTTCTGCACC GGATC ATT AC fis2-B GCGTTTCCTTATGATTAAAC GCGTTTCCTTATGATTAAACTC GCCGCAGACATTCATGGT TCTGG TGA fis2-D ACCACGATTTTCCCCAACTT ACCACGATTTTCCCCAACTTA CCATCACATTTGGTTCCCTTTT G TAG mea-A TGCAAACTACATGGGTGTT TGCAAACTACATGGGTGTTCT GGTCCATCATCCATGCTGTC CTC T mea-B GCTGCTTCCCTATGAGAGA GCTGCTTCCCTATGAGAGATA TGTACCCCAGAATGATGTGGA TAAG AA mea-D AGGTTATGTTTGTAGCAGG AGGTTATGTTTGTAGCAGGCA TGCCTGCATTAATCCGCTCA CG fie-A GCGAAGGCAAAATCTATGT GCGAAGGCAAAATCTATGTGT CTAGCACCACCACCACATT GTG A fie-B NA NA NA fie-D GTTCGCTAACCTTTCATTGA GTTCGCTAACCTTTCATTGACT CCACCCTTCTGATATCTACCG CC Table 2.3: KASP primer information. 30 A systematic grep search around the targeted SNP was used as a general guide in selecting planting desired genotypes. Grep searching allowed a fast turnaround time between harvesting and planting. The Unix command zgrep -c was used to generate the number of reads containing desired SNP and the number of reads containing the wild type sequence in cycle 3 samples. Cycle 4 samples were first cleaned using Cutadapt to remove Fluidigm CS1/CS2 adapters. Reads were then paired using software PEAR before the Unix command zgrep -c was used to generate the number of reads containing the desired SNP and the number of reads containing the wild type sequence. Tables 2.5 and 2.6 list the sequences used in Cycle 3 and Cycle 4 grep search. Some Cycle 4 grep sequences differ from Cycle 3 grep sequences due to pooled samples causing false positive calls. Phylogenetic trees Phylogenetic trees were created in R studio using R package phylogram (Figure 2.3). The phylogenetic trees were originally generated using EMBl-EBI tool Clustal Omega and were recreated in R studio to generate higher quality images (Madeira et al., 2019). 31 Homeoallele Primer Left Primer Right primer Total bp Wild Mutant set bp from SNP SNP 5' to SNP fis2-A 1 CACACGGAATTTCAGCATTG GTGGGTGGTGAAAGATTGCT 208 165 G A fis2-B 1 AGGAGCTGGCGTTTCCTTAT GGATCACTGCTGCTTCAACA 281 33 G A fis2-D 1 AGGTCGCACCAAAGCTACAT CACATGGTCCACAGAGCAGT 225 178 C T mea-A 1 CAGCATTGGATTCCTTCGAT CTGTCCCAAGGTGATTGCTT 268 158 C T mea-B 1 TGGCAAGGATGACTGTCTTG AAACCGAGCTGACAAGGAGA 245 38 C T mea-D 1 TGCTTTTGCATGAGCTGAGT GTCAGGCGCGTAATGGTAAT 289 209 G A fie-A 1 ACATGCTCGTGGCTAATCGT CACCACCACCACATTCAGAG 242 143 G A fie-A 2 AGTTTCTGAGGATGGCCGTA CCGGGTAATTAGCACAGGAG 234 197 G A fie-B 1 CGGCACTAACTGTCCTTTCA TCTCCACGTATGGCCAAAAT 223 106 G A fie-D 1 GCCAGTTGTGGCATGGATA AGCACGATGCTGAGAGGATT 291 35 G A Table 2.4: Amplicon sequencing primers. Cycle 3 Homeoallele wild mutant fis2-A GATCCTGCTCAATCATTACATGGT AATCCTGCTCAATCATTACATGGT fis2-B GCGTTTCCTTATGATTAAACTCTGG GCGTTTCCTTATGATTAAACTCTGA fis2-D CAAGTTGGGGAAAATCGTGGT TAAGTTGGGGAAAATCGTGGT fie-A GCGAAGGCAAAATCTATGTGTG GCGAAGGCAAAATCTATGTGTA fie-B TGGATAATACTGTTAAAATCTG TGGATAATACTGTTAAAATCTA fie-D GGTCAATGAAAGGTTAGCGAA AGTCAATGAAAGGTTAGCGAA mea-A TGCAAACTACATGGGTGTTCTC TGCAAACTACATGGGTGTTCTT mea-B CTTATCTCTCATAGGGAAGCAGC TTTATCTCTCATAGGGAAGCAGC mea-D AGGTTATGTTTGTAGCAGGCG AGGTTATGTTTGTAGCAGGCA Table 2.5: Grep sequences used for read 1 Cycle 3 samples. Reverse complement of these sequences was used for read 2. 32 Cycle 4 Homeoallele wild mutant fis2-A GATCCTGCTCAATCATTACATGGT AATCCTGCTCAATCATTACATGGT fis2-B GCGTTTCCTTATGATTAAACTCTGG GCGTTTCCTTATGATTAAACTCTGA fis2-D CAAGTTGGGGAAAATCGTGGT TAAGTTGGGGAAAATCGTGGT fie-A GCGAAGGCAAAATCTATGTGTG GCGAAGGCAAAATCTATGTGTA fie-B CATTCATACTTCCGTTATATCAGGACTTCCA CATTCATACTTCCGTTATATCAGGACTTCCACCCTTC CCCTTCTGATATCTACCGAATTGCCAGTTGT TGATATCTACCGAATTGCCAGTTGTGGCATGGATAA GGCATGGATAATACTGTTAAAATCTGGTCA TACTGTTAAAATCTAGTCAATGAAAGGTTAGCAAAC ATGAAAGGTTAGCAAACCATTTCATCAATC CATTTCATCAATCTTGCTTCCTATTTACATTTTATTA TTGCTTCCTATTTACATTTTATTACTGCTGA CTGCTGATCACTGACATGGTTTGTTAA TCACTGACATGGTTTGTTAA fie-D CGGCACTAACTGTCCTTTCATTCATACTTCT CGGCACTAACTGTCCTTTCATTCATACTTCTGTTATA GTTATATCAGGACTTCCACCCTTCTGATATC TCAGGACTTCCACCCTTCTGATATCTACCGAATTGC TACCGAATTGCCAGTTGTGGCATGGATAAT CAGTTGTGGCATGGATAATACTGTTAAAATCTAGTC ACTGTTAAAATCTGGTCAATGAAAGGTTAG AATGAAAGGTTAGCGAACCATTTCATCAATCTTGCT CGAACCATTTCATCAATCTTGCTTCCTATTT TCCTATTTACATTT ACATTT mea-A TGCAAACTACATGGGTGTTCTC TGCAAACTACATGGGTGTTCTT mea-B CTTATCTCTCATAGGGAAGCAGC TTTATCTCTCATAGGGAAGCAGC mea-D AGGTTATGTTTGTAGCAGGCG AGGTTATGTTTGTAGCAGGCA Table 2.6: Grep sequences used for Cycle 4 samples. 33 Variant calling The primary method used to confirm genotypes was a BWA/bcftools pipeline (Yao et al., 2020; Danecek et al., 2021). Reads were first cleaned using cutadapt (version 2.4.Py3) with parameters -a ACACTGACGACATGGTTCTACA, -A TACGGTAGCAGAGACTTGGTCT, and -m 1. The reference genome was comprised of expected amplicon products (Table 2.7). The cleaned reads were aligned to the amplicon product reference genome using bwa (version 0.7.16a) mem to generate SAM output files. The SAM files were converted to BAM format via samtools (version 1.11) view with parameters -S and -b. The bam files were then sorted via samtools sort. The sorted files were used to generate a BCF file of genotype likelihoods using bcftools (version 1.11) mpileup with parameters -d 10000000 and -a FORMAT/AD. Variants were then called using bcftools call with parameters -Oz and -m. Bcftools view with parameter - Oz was used to convert the files to VCF format. VCF files were compressed using bcftools index. Finally, VCF files from all samples were merged into a single VCF file using bcftools merge with parameter -Ov. Cadenza mutants and Ae. tauschii hybridizations Hybridizations between Cadenza mutants and Ae. tauschii was done using the approach method (Curtis and Croy, 1958). The maternal parents, Cadenza mutants, were emasculated and covered for 3-6 days prior to hybridization. Afterwards, the emasculated heads were arranged to be slightly below Ae. tauschii spikes. Ae. tauschii spikes were either currently releasing viable pollen or about to release pollen. Dialysis tubing was hydrated and dried to a hollow tube shape. The tube contained the emasculated Cadenza heads and Ae. tauschii heads for each hybridization 34 and prevented cross contamination while also allowing light to pass through. Each Cadenza mutant line, and wild type Cadenza, had 3-10 spikes, or replicates, hybridized with Ae. tauschii. The exception was wild type Cadenza which had 15 spikes hybridized with Ae. tauschii. Seeds per spike were calculated within each line. Lines where then grouped into genotypes within MEA, FIE, or FIS2. 35 Homeoallele Primer Product (Wild) set fis2-A 1 CACACGGAATTTCAGCATTGAGAAGTTCTGCATGCTGTTTTGAATATAACACACCTGAAACA TTTTGATCCTTACCTTAGGTGAAGATTTATTTCTTACACGTGCTCATATCCCATTTCCGTTGT CTAAAATTTCCAGATATTTCTGCACCACGTGCTTCTGTTGATCCTGCTCAATCATTACATGGT AGCAATCTTTCACCACCCAC fis2-B 1 AGGAGCTGGCGTTTCCTTATGATTAAACTCTGGAACCACAGTCTATTAGATGCCCGCGCCAT GAATGTCTGCGGCACAATTCTTCAAGGCTACCAAAATGAAAGCTCGGACCCCAAGAAAATG TGAGTCGAGCATAGTGCGCTCATTATAATTTAAATCAAGTGGGTGTTTGGAGGAGGCCTTAG GAGCAGGCTAAAGAGAGGTGAAAGCTGCACCTGAGGCCATGGAGGACTATTCTCGATTCTA TTTATCGATCGGTTGTGTTGAAGCAGCAGTGATCC fis2-D 1 AGGTCGCACCAAAGCTACATTCATAATTCCAGACATCAAGAATTTATCAACCTCCCGAGCTT GCAACCTTAACATTATCCTTATCAGCTGCGGTATGATGTGCTTCATATGTGATGTTTATGGTT TTACATTTCTTCACCACCATCACATTTGGTTCCCTTTTTAGTTTCGGAAGGGCAAGTTGGGGA AAATCGTGGTGAACATAACTGCTCTGTGGACCATGTG mea-A 1 CAGCATTGGATTCCTTCGATAATCTCTTTTGTCGGCGATGCCTAGTATGCTCATCTATCAATC ATCATGTGATATATGTAATGCCTCTTGTTGCTGATGAATATTCTTCTCATCGTGACATTGGCA GGTTTTTTATTGCAAACTACATGGGTGTTCTCAAGATTTAATATTTCCTGTAAGTTTTGTTCC AGAATTATTTCCAAAATTACGTGTTTTTTCTTATTTGTGCCTTCTTCATACAGGCAGAAAAGC AATCACCTTGGGACAG mea-B 1 TGGCAAGGATGACTGTCTTGGGGAGTACACTGGGGAGCTTATCTCTCATAGGGAAGCAGCC AAGCGTGGACAGAGATATGACCGTGAGAATTCATCATTCCTTTTCAACTTAAATACTGAGGT ACGTGTGCAGAAATATACCAGTTACAGTATGCAAACATTCACGGTTACACGCGGGACTGAT CACCAGATTTAGCTATAGTGCTTTTGCATGAGCTGAGTTATTCTCCTTGTCAGCTCGGTTT mea-D 1 TGCTTTTGCATGAGCTGAGTTATTCTCCTTCTCAGCTCGGTTTATTCAAGAACCTGTATTTTA TATCTGATCCAGACGCTAGCTGTTGCTTATGCATACTCTTTGGACGCAGTTTGTTCTCGATGC CTTCAGGATGGGCAACAAGCTGAAGTTTGCCAACCATTCCCCTGATCCGAATTGCTATGCCA AGGTTATGTTTGTAGCAGGCAACCATAGGGTGGGCATATTCGCCAATGAGCGGATTAATGC AGGCAAGGAGATTTTTTATGATTACCATTACGCGCCTGAC Table 2.7: Amplicon sequencing primer products used as reference genome in variant calling pipeline. 36 Table 2.7 (cont’d) Homeoallele Primer Product (Wild) set fie-A 1 ACATGCTCGTGGCTAATCGTTTGCTCGTAAACGAACTCATCTACATGCACTTGATCAGGGTT AGATGAAATTCAAAAAGTATAACTCTTGTGGTTAATCGTTTGCTTGTAAACAGGCAACCGCG AAGGCAAAATCTATGTGTGGGAAGTGCAGGCGAGCCCTCCTGTGCTAATTACCCGGTAAGT TTAACCGAGCAAGTTCAGCTATTCCCAATTTCTCCTCCTCTGAATGTGGTGGTGGTG fie-A 2 AGTTTCTGAGGATGGCCGTACTTATACACGTGAAAAAACTGATATTTGTAGAAAACATGCTC GTGGCTAATCGTTTGCTCGTAAACGAACTCATCTACATGCACTTGATCAGGGTTAGATGAAA TTCAAAAAGTATAACTCTTGTGGTTAATCGTTTGCTTGTAAACAGGCAACCGCGAAGGCAA AATCTATGTGTGGGAAGTGCAGGCGAGCCCTCCTGTGCTAATTACCCGG fie-B 1 CGGCACTAACTGTCCTTTCATTCATACTTCCGTTATATCAGGACTTCCACCCTTCTGATATCT ACCGAATTGCCAGTTGTGGCATGGATAATACTGTTAAAATCTGGTCAATGAAAGGTTAGCA AACCATTTCATCAATCTTGCTTCCTATTTACATTTTATTACTGCTGATCACTGACATGGTTTG TTAATCTTGTCATAGAATTTTGGCCATACGTGGAGA fie-D 1 GCCAGTTGTGGCATGGATAATACTGTTAAAATCTGGTCAATGAAAGGTTAGCGAACCATTTC ATCAATCTTGCTTCCTATTTACATTTTATTACTGCTGATCACTGACATGGTTTGCTAACGTTG TCATAGAATTTTGGCCATACGTGGAGAAATCCTTTACATGGACTGACCTTCCATCAAAATTT CCAACAAAATTTGTCCAATTTCCGGTATGTTAAGTATCTCTCATCTTTGAACCTGTTATTATG CAAACTTTTAGTATTGTTCTTAATCCTCTCAGCATCGTGCT 37 Results Cycle 1, 2 and 3 genotyping with KASP markers KASP markers were used to identify homozygous mutant, heterozygous mutant, and wild type samples for each homeoallele of MEA, FIE and, FIS2 in Cycles 1, 2, and 3. Cycle 1 genotyping was used to either confirm homozygosity or select for homozygous individuals in heterozygous mutant donor parents. Cycle 2 genotyping identified F1 intercrosses that were heterozygous for two mutations (Table 2.8). Gene Cross Offspring Genotype Plants MEA AABBdd/AAbbDD AABbDd 4 aaBBDD/AABBdd AaBBDd 2 AaBB ? 1 aaBBDD/AAbbDD AaBbDD 2 FIE AABBdd/aaBBDD AaBBDd 4 AAbbDD/AABBdd AABbDd 2 AABBDd 5 FIS2 AAbbDD/AABBdd AABbDD 2 aaBBDD/AABBdd AABBDd 2 AAbbDD/aaBBDD AABbDD 2 aaBBDD/AABBdd AaBBDd 4 ? BBDd 2 AABBDd 2 Table 2.8: Cycle 2 KASP genotyping results from selected offspring. ? indicates KASP results that were labeled “unknown”. Cycle 3 KASP genotyping identified Cadenza triple mutant lines, or lines with either three heterozygous mutations in three homeologs (AaBbDd), or two heterozygous mutations in two homeologs and a homozygous mutation in one homeolog (AaBbdd, aaBbDd, or AabbDd) (Table 2.9). 38 Gene Genotype Plants MEA AaBbDd 3 AabbDd 1 AaBb* 1 FIE** Aa * * 16 FIS2 AaBbDd 4 Aa*Dd 1 Table 2.9: Cycle 3 KASP genotyping results. Genotypes missing a subgenome had inconclusive KASP results. ** indicates KASP markers were not available for FIE B and D genome mutations. Homeoallele specificity, or lack thereof, in KASP markers became apparent after genotyping Cycle 2 F1 lines. The mea-A, mea-B, mea-D, fie-A, and fis2-D KASP markers demonstrated clear segregation patterns between homozygous mutant, heterozygous mutant, and wild type samples (Figure 2.2). Figure 2.2: KASP segregation for mea-A, mea-B, mea-D, fie-A, and fis2-D in Cycle 3 genotyping. Blue squares represent homozygous mutant samples, green triangles represent heterozygous samples, orange circles represent wild type samples, and red X’s represent unknown genotypes. Samples in black circles contain wild controls. 39 The fie-B, fie-D, fis2-A, and fis2-B KASP markers did not demonstrate clear and consistent segregation patterns between homozygous mutant, heterozygous mutant, and wild type samples. Wild controls were often indistinguishable from homozygous mutant and heterozygous mutant samples. A KASP marker was not available for fie-B. Attempts were made to create a homeoallele specific fie-B KASP maker without success which led to an amplicon sequencing approach for genotyping. Amplicon sequencing Cycle 3 A subset of lines that were expected to be double or triple homozygous mutant from KASP were genotyped using amplicon sequencing. A total of 20 Cycle 3 lines were genotyped, along with the mutant donor parents and wild Cadenza. Cycle 2 progeny were expected to occur at 12.5% for heterozygous mutations in three homeologs (AaBbDd), and 6.25% for heterozygous mutations in two homeologs and a homozygous mutation in one homeolog (AaBbdd, aaBbDd, or AabbDd). Amplicon sequencing confirmed MEA genotypes and elucidated further information for FIS2 and FIE samples (Table 2.10). The target genotype (AaBbDd, AaBbdd, aaBbDd, or AabbDd) could not be obtained for FIS2. Genotype MEA FIE FIS2 AaBbDd 3 1 0 AabbDd 1 0 0 AaBbdd 1 0 0 AaBbDD 0 7 0 AaBBDd 0 0 1 AaBBDD 0 7 1 AABBDD 0 0 3 total 5 15 5 Table 2.10: Cycle 3 amplicon sequencing genotypic distribution. 40 Amplicon sequencing Cycle 4 A total of 509 Cycle 4 lines were genotyped, along with the mutant donor parents and wild Cadenza (Table 2.11). Desired Cycle 3 progeny were expected to occur at 3.125% if the parental genotype was heterozygous for all three homeologs (AaBbDd) or 6.25% if the parental genotype had heterozygous mutations in two homeologs and a homozygous mutation in one homeolog (AaBbdd, aaBbDd, or AabbDd). FIE samples sequenced had parental genotypes of the former (AaBbDd) while MEA samples sequenced had parental genotypes of the latter (AaBbdd and AabbDd). Even though there were no homozygous triple mutant FIE lines, this was within the range of the expected distribution (P>0.05). MEA homozygous triple mutants occurred at a slightly lower rate than expected, but only slightly at P=0.047. (Table 2.12). Homozygous triple mutant FIS2 lines were unable to be obtained. MEA FIE FIS2 aabbdd 2 0 0 AaBbDd 0 1 0 AabbDd 10 1 0 aaBbDd 0 5 0 aabbDd 9 0 0 aaBBdd 0 0 0 aaBBDd 0 0 1 AaBBDd 0 1 7 AaBBdd 1 0 0 AAbbdd 0 0 0 AABbDd 0 3 0 AABbdd 0 0 0 AAbbDd 6 0 0 aabbDD 16 0 0 AaBbDD 9 94 0 AabbDD 23 24 0 aaBbDD 5 46 0 AaBBDD 8 15 62 aaBBDD 3 6 17 AABbDD 7 58 0 Table 2.11: Cycle 4 genotyping results. 41 Table 2.11 (cont’d) MEA FIE FIS2 AAbbDD 10 18 0 AABBDd 0 0 3 AABBdd 0 0 0 wild 5 7 26 total 114 279 116 Genotype Observed occurrence Expected occurrence chi square value P-value MEA aabbdd 2 7.125 3.93 0.047 FIE aabbdd 0 1.656 1.709 0.191 Table 2.12: Cycle 4 chi square results. Hybridizing single, double, and/or triple mutant homeoallele(s) to Ae. tauschii In Cycle 1, single homeoallele mutant hybridizations yielded no viable endosperm for any homeoallele in any gene. After Cycle 4, single, double and/or triple mutant Cadenza lines were hybridized with Ae. tauschii (Table 2.13). Wild type endosperm development did not occur for any of these hybridizations, but caryopses were still able to be harvested. Due to lines having a variable number of spikes hybridized with Ae. tauschii (3-10 spikes per line), seeds per spike was calculated for each line (Table 2.13). Means were generated for all genotypes within each PRC2 gene (Table 2.14). No genotype had a higher seeds per spike count than the self-crossed wild Cadenza, however all averages were greater than the wild Cadenza hybridized with Ae. tauschii. FIS2 genotypes in particular had more seeds per spike than MEA and FIS2 genotypes. FIS2 genotype aaBBDd average seeds per spike was close to the self-crossed wild Cadenza seeds per spike (Table 2.14). 42 PRC2 Gene Genotype Seeds per Spike MEA AabbDd 1.3 MEA aabb 3.333333 MEA aabb 1.833333 MEA aabb 1.166667 MEA aabb 6.666667 MEA aabbdd 5.666667 MEA aabb 1.333333 MEA bb 4.166667 MEA aabb 1 MEA aabb 0.666667 MEA bb 0.666667 MEA aabb 1.333333 MEA aabb 0.166667 MEA bb 1.666667 MEA aa 1.5 MEA aabb 4.5 MEA aabb 1.714286 MEA aabb 2 MEA bb 0.166667 MEA aa 4 MEA aabbdd 3.166667 MEA bb 0.857143 MEA aabb 2.333333 FIE bb 1.5 FIE aa 0.833333 FIE aaBb 1 FIE bb 2.333333 FIE aa 1.5 FIE aa 1.5 FIE aaBb 4 FIE aaBb 1 FIE aa 2.666667 FIE aaBb 2.5 FIE aaBb 0.5 FIE aaBb 3.571429 FIE bb 1.166667 FIE bb 1.333333 Table 2.13: Caryopses yield from single, double, and/or triple mutant Cadenza lines hybridized with Ae. tauschii. Wild Cadenza x wild Cadenza and wild Cadenza x Ae. tauschii hybridizations were made as controls. 43 Table 2.13 (cont’d) PRC2 Gene Genotype Seeds per Spike FIS2 AaBBDd 2.833333 FIS2 aaBBDd 6.666667 FIS2 aa 2.75 FIS2 AaBBDd 8 FIS2 aa 3.166667 FIS2 aa 7.166667 FIS2 aa 5.333333 Wild x Wild 8.666667 Wild x Ae. tauschii 0.6 PRC2 Gene Genotype Seeds per spike MEA aaBBDD 2.7 aabbDD 2.157 aabbdd 4.416 AabbDd 1.3 AAbbDD 1.5 FIE aaBBDD 1.625 aaBbDD 2.095 AAbbDD 1.58 FIS2 aaBBDD 4.604 aaBBDd 6.66 AaBBDd 5.416 Table 2.14: Average seeds per spike of the different genotypes of MEA, FIE, and FIS2 resulting from hybridizations between Ae. tauschii. Figure 2.3: Phylogenetic trees of MEA, FIE, and FIS2 genes. 44 Figure 2.3 (cont’d) Discussion KASP marker results can be split into two categories: 1. Clear and consistent segregation between wild, heterozygous, and homozygous mutant samples. 2. Little to no segregation between wild, heterozygous, and homozygous mutant samples. Markers in the first category have homeoallele specificity. Homeoallele specificity is consistent amplification of wild type and mutant type sequences in the homeoallele that marker is targeting. Markers in the first category most likely do not have homeoallele specificity and have off-target amplification. Polyploid species are particularly susceptible to off-target amplification with SNP markers due to similar homoeologous or paralogous sequences in the genome (Edwards et al., 45 2009). Similar issues arose when converting SNP chip arrays to KASP markers because of the presence of paralogs and homologs in polyploid crops like hexaploid wheat (Makhoul et al., 2020). Marker development and improvement was attempted for fie-B. A KASP marker was not available at John Innes Center for the targeted SNP of fie-B. Two attempts were made to create a KASP marker that was homeoallele-specific to the SNP of fie-B, but both attempts were unsuccessful. Accurate genotyping is necessary to distinguish homozygous and heterozygous mutant homeoalleles in lines after pyramiding cycles. Therefore, amplicon sequencing was employed to overcome the off-target amplification in fie-B, fie-D, fis2-A, and fis2-B KASP markers. MEA KASP results had consistent homeoallele specificity for all 3 homeoalleles. Due to this consistency, mea-A, mea-B, and mea-D were used to compare the effectiveness of amplicon sequencing in Cycle 3. Genotyping results were similar between KASP markers and the amplicon sequencing pipeline for MEA homeoalleles. The Unix command grep was used as an approximate guide for determining which lines would be planted in Cycle 4, and which lines would be hybridized with Ae. tauschii after Cycle 4. These grep searches contained the desired homeoallele SNP or wild type sequence. Grep searches closely matched the KASP and snpcalling pipeline genotyping results for MEA samples in Cycle 3. Unfortunately, the grep method had less accurate results for all other PRC2 genes in Cycle 3, and all PRC2 genes in Cycle 4. This, combined with inaccurate KASP markers, resulted in several lines being incorrectly labeled double and triple mutants. Most notably, the parent for fis2-B was mis-genotyped as homozygous mutant for fis2-B. The parental line Cadenza1722 is heterozygous for the desired SNP of homeoallele fis2-B, and accurate KASP genotyping wasn’t available for this line to determine segregating homozygous mutant fis2-B plants. The amplicon 46 sequencing pipeline was successful in genotyping pooled (Cycle 4) and non-pooled (Cycle 3) samples. It is necessary to have pooled samples to decrease sequencing costs, while maximizing the number of samples sequenced. Single, double, and triple mutant lines generated and identified in Cycle 4 were hybridized with Ae. tauschii (Table 2.13, Table 2.14). Although endosperm development did not resemble wild type endosperm development, caryopses did form from these hybridizations. It is possible that with continued pyramiding and genotyping using the amplicon sequencing pipeline, more triple and double homozygous mutants can be generated and potentially form viable wild type resembling endosperm when hybridized with Ae. tauschii. FIS2 is particularly promising. FIS2 mutants with the genotype aaBBDd produced an average 6.66 seeds per spike. FIS2 genotypes with more non-functional alleles of FIS2 (AaBbDd, aabbdd, or even aaBBdd) may out produce self-hybridized wild Cadenza. Work is underway to develop homozygous mutant double and triple mutants of FIE, FIS2, and MEA. Crossability, instead of endosperm development, is a potential phenotypic observation in this study. Crossability genes have been known to exist in wheat, most notably Kr1, Kr2, Kr3, and Kr4 (Manickavelu et al., 2009). Dominant crossability alleles have been known to repress production of fertile seeds from hybridizations between wheat and some Poaceae family members such as barley and rye (Alfares et al., 2009; Manickavelu et al., 2009). Most of these crossability genes are on the same chromosomes as the FIS2 homeoalleles: Kr1 on 5B, Kr2 on 5A, and Kr3 on 5D (Alfares et al., 2009). The mode of action is likely to be different than crossability genes, as dominant Kr alleles prevent the entry of pollen tubes into maternal tissue (Fedak and Jui, 1982). Hybridizations between wheat and Ae. tauschii can still create viable embryos, and therefore functioning pollen tubes to create those embryos. 47 The phenotype of these mutations may be influenced by the background mutations from the EMS lines. For example, the triple homozygous mutant lines of MEA may have a stronger impact on the phenotype of endosperm development when hybridized with Ae. tauschii, but the phenotype may be inhibited by the background mutations of the EMS lines. These lines will be backcrossed to the non-EMS parent and selected for heterozygous PRC2 mutations via the amplicon sequencing pipeline. The amplicon sequencing pipeline developed will have future applications in direct hybridization work including introgression into elite wheat cultivars. Research in other allopolyploid crops may also adapt this amplicon sequencing pipeline when experiencing off target amplification with markers. 48 BIBLIOGRAPHY 49 BIBLIOGRAPHY Alfares W, Bouguennec A, Balfourier F, Gay G, Bergès H, Vautrin S, Sourdille P, Bernard M, Feuillet C (2009) Fine Mapping and Marker Development for the Crossability Gene SKr on Chromosome 5BS of Hexaploid Wheat (Triticum aestivum L.). Genetics 183: 469–81, 1SI Cox TS, Wu J, Wang S, Cai J, Zhong Q, Fu B (2017) Comparing two approaches for introgression of germplasm from Aegilops tauschii into common wheat. Crop J 5: 355– 362 Curtis BC, Croy LI (1958) The Approach Method of Making Crosses in Small Grains1. Agron J 50: 49–51 Danecek P, Bonfield JK, Liddle J, Marshall J, Ohan V, Pollard MO, Whitwham A, Keane T, McCarthy SA, Davies RM, et al (2021) Twelve years of SAMtools and BCFtools. GigaScience. doi: 10.1093/gigascience/giab008 Edwards KJ, Reid AL, Coghill JA, Berry ST, Barker GLA (2009) Multiplex single nucleotide polymorphism (SNP)-based genotyping in allohexaploid wheat using padlock probes. Plant Biotechnol J 7: 375–390 FAOSTAT (2020) Food and agriculture organization of the United Nations. Food Agric data. http://www.fao.org/faostat/en/#data Faure J.E. (2011) Double Fertilization in Flowering Plants: Origin, Mechanisms and New Information from in vitro Fertilization. Fertil High Plants Mol Cytol Asp 79 Fedak G, Jui PY (1982) Chromosomes of chinese spring wheat carrying genes for crossability with betzes barley. Can J Genet Cytol 24: 227–233 Gertz EM, Yu Y-K, Agarwala R, Schäffer AA, Altschul SF (2006) Composition-based statistics and translated nucleotide searches: Improving the TBLASTN module of BLAST. BMC Biol 4: 41 Holec S, Berger F (2012) Polycomb Group Complexes Mediate Developmental Transitions in Plants. Plant Physiol 158: 35 Huang S, Sirikhachornkit A, Su X, Faris J, Gill B, Haselkorn R, Gornicki P (2002) Genes encoding plastid acetyl-CoA carboxylase and 3-phosphoglycerate kinase of the Triticum Aegilops complex and the evolutionary history of polyploid wheat. Proc Natl Acad Sci 99: 8133 Jullien PE, Mosquna A, Ingouff M, Sakata T, Ohad N, Berger F (2008) Retinoblastoma and Its Binding Partner MSI1 Control Imprinting in Arabidopsis. PLOS Biol 6: e194 50 Köhler C, Hennig L, Bouveret R, Gheyselinck J, Grossniklaus U, Gruissem W (2003) Arabidopsis MSI1 is a component of the MEA/FIE Polycomb group complex and required for seed development. EMBO J 22: 4804–4814 Kradolfer D, Hennig L, Köhler C (2013) Increased Maternal Genome Dosage Bypasses the Requirement of the FIS Polycomb Repressive Complex 2 in Arabidopsis Seed Development. PLOS Genet 9: e1003163 Lagudah ES, Appels R, Brown AHD, McNeil D (1991) The molecular–genetic analysis of Triticum tauschii, the D-genome donor to hexaploid wheat. Genome 34: 375–386 Lanning SP, Blake NK, Sherman JD, Talbert LE (2008) Variable Production of Tetraploid and Hexaploid Progeny Lines from Spring Wheat by Durum Wheat Crosses. Crop Sci 48: 199–202 Lewis EB (2004) A Gene Complex Controlling Segmentation in Drosophila. In HD Lipshitz, ed, Genes Dev. Cancer Life Work Edw. B Lewis. Springer US, Boston, MA, pp 205–217 Li A, Liu D, Yang W, Kishii M, Mao L (2018) Synthetic Hexaploid Wheat: Yesterday, Today, and Tomorrow. Engineering 4: 552–558 Ling H-Q, Zhao S, Liu D, Wang J, Sun H, Zhang C, Fan H, Li D, Dong L, Tao Y, et al (2013) Draft genome of the wheat A-genome progenitor Triticum urartu. Nature 496: 87 Madeira F, Park YM, Lee J, Buso N, Gur T, Madhusoodanan N, Basutkar P, Tivey ARN, Potter SC, Finn RD, et al (2019) The EMBL-EBI search and sequence analysis tools APIs in 2019. Nucleic Acids Res 47: W636–W641 Makhoul M, Rambla C, Voss-Fels KP, Hickey LT, Snowdon RJ, Obermeier C (2020) Overcoming polyploidy pitfalls: a user guide for effective SNP conversion into KASP markers in wheat. Theor Appl Genet 133: 2413–2430 Manickavelu A, Koba T, Mishina K, Sassa H (2009) Molecular characterization of crossability gene Kr1 for intergeneric hybridization in Triticum aestivum (Poaceae: Triticeae). Plant Syst Evol 278: 125–131 Mcfadden ES, Sears ER (1946) The origin of triticum spelta and its free-threshing hexaploid relatives. J Hered 37: 81–89 Miranda LM, Murphy JP, Marshall D, Leath S (2006) Pm34: a new powdery mildew resistance gene transferred from Aegilops tauschii Coss. to common wheat (Triticum aestivum L.). Theor Appl Genet 113: 1497–1504 Nallamilli BRR, Zhang J, Mujahid H, Malone BM, Bridges SM, Peng Z (2013) Polycomb group gene OsFIE2 regulates rice (Oryza sativa) seed development and grain filling via a mechanism distinct from Arabidopsis. PLoS Genet 9: e1003322–e1003322 51 Ng PC, Henikoff S (2003) SIFT: Predicting amino acid changes that affect protein function. Nucleic Acids Res 31: 3812–3814 Olson EL, Rouse MN, Pumphrey MO, Bowden RL, Gill BS, Poland JA (2013) Simultaneous transfer, introgression, and genomic localization of genes for resistance to stem rust race TTKSK (Ug99) from Aegilops tauschii to wheat. Theor Appl Genet 126: 1179–1188 Saluja M, Kaur S, Bansal U, Subhash Chander Bhardwaj, Chhuneja P (2018) Molecular mapping of linked leaf rust resistance and non-glaucousness gene introgressed from Aegilops tauschii Coss. in hexaploid wheat Triticum aestivum L. Plant Genet Resour 16: 82–88 Schmidt A, Wöhrmann HJP, Raissig MT, Arand J, Gheyselinck J, Gagliardini V, Heichinger C, Walter J, Grossniklaus U (2013) The Polycomb group protein MEDEA and the DNA methyltransferase MET1 interact to repress autonomous endosperm development in Arabidopsis. Plant J 73: 776–787 Shen X, Gmitter FG, Grosser JW (2011) Immature Embryo Rescue and Culture. In TA Thorpe, EC Yeung, eds, Plant Embryo Cult. Methods Protoc. Humana Press, Totowa, NJ, pp 75–92 Velu G, Guzman C, Mondal S, Autrique JE, Huerta J, Singh RP (2016) Effect of drought and elevated temperature on grain zinc and iron concentrations in CIMMYT spring wheat. J Cereal Sci 69: 182–186 Weinhofer I, Hehenberger E, Roszak P, Hennig L, Köhler C (2010) H3K27me3 Profiling of the Endosperm Implies Exclusion of Polycomb Group Protein Targeting by DNA Methylation. PLoS Genet 6: e1001152 Wiersma AT, Pulman JA, Brown LK, Cowger C, Olson EL (2017) Identification of Pm58 from Aegilops tauschii. Theor Appl Genet 130: 1123–1133 Yadegari R, Kinoshita T, Lotan O, Cohen G, Katz A, Choi Y, Katz A, Nakashima K, Harada JJ, Goldberg RB, et al (2000) Mutations in the FIE and MEA Genes That Encode Interacting Polycomb Proteins Cause Parent-of-Origin Effects on Seed Development by Distinct Mechanisms. Plant Cell 12: 2367 Yao Z, You FM, N’Diaye A, Knox RE, McCartney C, Hiebert CW, Pozniak C, Xu W (2020) Evaluation of variant calling tools for large plant genome re-sequencing. BMC Bioinformatics 21: 360 52 CHAPTER III DEVELOPMENTAL CHANGES IN WINTER AND SPRING WHEAT IN VARIOUS PLANTING DENSITIES 53 Abstract The wheat breeding program at Michigan State University utilizes a modified bulk breeding method to rapidly advance populations from the F2 to F4 stage under greenhouse conditions. To determine the influence of planting density on yield components and plant architecture in greenhouse settings, two experiments were carried out using two spring and winter wheat. Planting densities ranged from 2 to 39.5 seeds per 10 cm2. All traits measured increased as planting density increased, decreased as planting density increased, or followed no trend as planting density increased. Experiments with both spring and winter wheat demonstrated that increased planting density reduces plant height and reduces the number spikelets per spike. Grain dimensions are also influenced with reduced grain length and width observed at higher planting densities. Introduction Wheat yield components demonstrate plasticity in response to the environment. Nutrients, water, light, soil volume, and other plants are influential environmental factors for wheat growth habit and yield. Agronomic practices that provide an environment without limitations to nitrogen and micronutrients result in higher grain yields (Black et al., 1947). However, modern agronomic practices require high density monocultures that may create competition for resources. Competitive high density planting environments can change plant architecture and in turn affect components of grain yield. Competition is often less severe between relatives, or kin, within a species due to kin recognition (Dudley and File, 2007; Caffaro et al., 2013). Some species, such as Lupinus angustifolius, do not have kin recognition and have similar competition 54 between kin and non-kin lines (Milla et al., 2009). Kin recognition has other limits, specifically under high density plantings in commercial crop production. High density plantings force roots to be in close proximity and compete for water and nutrients (Nord et al., 2011). This competition may lead to overallocation of resources to roots, or root over-proliferation, to the detriment of above ground biomass contributing to grain yield (Nord et al., 2011; Zhu et al., 2019). Gersani et al. (2001) demonstrated that non-competing Glycine max produce 30% more yield than Glycine max competing with kin Glycine max. Few grain planting density studies in greenhouse conditions have been conducted. Cisar et al. (1982) observed the impacts of planting density of oats in greenhouse settings. Cisar et al. (1982) found that as planting density increased, so did the number of spikes containing 1 or 0 seeds. These oats were grown in stressed conditions which intensified the competition induced changes in yield components (Cisar et al., 1982). Recently, data gathered from Olson et al. (2021) suggested that some yield components may be genotype dependent, while others genotype independent. As planting densities increase, it is expected that yield components, such as number of seeds per spike, will be altered in ways that may or may not be genotypic dependent. Characterizing the impact of planting density on wheat phenotypes, especially yield components, will allow growers and breeders to know the full extent of risks and benefits of a range of planting densities. In this study, two experiments were carried out, one using spring wheat and one using winter wheat varieties, to characterize the impacts of planting density on yield components and plant architecture under greenhouse conditions. Planting densities ranged from 2 to 39.5 seeds per 10 cm2. Phenotypes evaluated included plant height, spikelets per spike, 55 mature spikelet count, spikes per pot, total seed weight, average seed weight, seed count, seed length, seed width, seeds per spike, and seeds planted vs number of spikes. Materials and methods Experimental design Winter wheat and spring wheat studies were conducted in a Randomized Complete Block Design (RCBD) created with PRISM (Central Software Solutions, Orange City, IA). 11 planting density treatments were applied (Table 3.1) increasing from 2 to 39.5 seeds per 10 cm2, or 32 to 640 seeds per 413 cm2 pot. Each planting density treatment was replicated four times across two different winter and spring wheat varieties. Seeds for each density level were counted using a Seed Counter R-25 (DATA Technologies, Baku, Yasamal AZ). Seeds per square 10 cm2 Seeds per pot 2 32 4 64 8 128 12 192 16 256 20 320 24 384 28 448 31.5 512 35.5 576 39.5 640 Table 3.1: Seed density treatments. Winter wheat The winter wheat lines, Whitetail and DF-112-R, were vernalized in and placed in a Dart 22.9 cm x 24.1 cm x 7.6 cm ClearSeal® Hinged Lid Plastic Container (Dart Container 56 Corporation, Mason, MI) at 4°C for 8 weeks. Each box contained an 18.4 cm x 19.1 cm section of 200 M weight White Verigood Blotter paper (Anchor Paper Company, St. Paul, Minnesota) saturated with 1% Thiram (T24201, Sigma Aldrich, St. Louis, MO) and DI water solution. 32, 64, 128, 192, 256, 320, 384, 448, 512, 576, or 640 seeds were added to each salad box. A light layer of Thiram solution was added over the seeds to prevent fungal growth. Thiram solution was added as needed throughout vernalization to ensure the Blotter paper stayed saturated, but not over saturated. The salad boxes were under LED lighting from WLFA2 series Fixtures (Super Bright LEDs Inc., Earth City, MO) with an 8-hour photoperiod. After vernalization, the seed mat of each salad box was transplanted in 20.3 cm x 20.3 cm x 19.1 cm Gro Pro® Premium Square Black Plastic Pots (Hawthorne Gardening Co., Vancouver, WA) with saturated Sure Mix potting media (Michigan Grower Products Inc., Galesburg, MI). The seed mats were placed on top of the soil and were lightly watered from above within 20 minutes of transplanting. Overhead watering occurred every 20-30 minutes for at least 5 hours after transplanting. Subdue, a systemic fungicide, was applied to the soil after transplanting to prevent downy mildew. 2’x4’ greenhouse trays (Greenhouse Megastore, Danville, IL) were used with two 1/16” holes drilled for drainage. Flood irrigation was used for both watering and water- soluble fertilizer. Fertilizer (20-20-20 NPK) was used one week after planting (188g / 5-gal H2O at 1:50 dilution) and again once weekly until Feekes 11. Plants were watered as needed and progressed to daily as plants matured. Water was cutoff when 75% of all plants in each pot reached physiological maturity (Feekes 11). The experiment was carried out in early spring and plants were under a 22-hour photoperiod. 57 Spring wheat The spring wheat lines, Melba and Seahawk, were germinated in clear salad boxes at 40F for 5 days. Each box contained a section of 200 M weight White Verigood Blotter paper (Anchor Paper Company, St. Paul, Minnesota) saturated with 1% Thiram (T24201, Sigma Aldrich, St. Louis, MO) and DI water solution. 32, 64, 128, 192, 256, 320, 384, 448, 512, 576, or 640 seeds were added to each salad box. A light layer of Thiram solution was added over the seeds to prevent fungal growth. After 5 days, the germinated seeds were transplanted to 8” x 8” x 7.5” Gro Pro® Premium Square Black Plastic Pots (Hawthorne Gardening Co., Vancouver, WA) with saturated Sure Mix potting media (Michigan Grower Products Inc., Galesburg, MI). The seeds were lightly covered with saturated soil and watered every 20-40 minutes for several hours after planting. Subdue was applied to the soil after transplanting. Greenhouse conditions were similar to winter wheat however the experiment was carried out during mid-winter and the lighting was switched to Phillips LED lighting for a 22-hour photoperiod. Watering and fertilizer were carried out the same as winter wheat. Data collection Data was collected for several yield components. Grain length and grain width were measured using a caliper. Number of spikes per pot were counted after senescence. Seed count was collected using DATA Technologies Seed Counter R-25 plus. Spikelet count was collected for mature and total spikelets. Winter genotypes had several spikes with immature spikelets, or spikelets that do not contain seeds. 10 subsamples were taken for grain length, grain width, 58 individual grain weight, mature and total spikelets per spike, and height. Seeds per spike was calculated by dividing total number of spikes per pot over total seed count. Data analysis Data analysis including ANOVA, LSD, and adjusted R squared values were generated in R studio. All traits were initially screened with the full model version of ANOVA (planting density and genotype and interaction of planting density and genotype) (Table 3.2). Factors in the models that were indicated to be significant at P<0.05 were used to create the final model used for LSD. LSD was calculated using R package agricolae. All graphs were created using R package ggplot2. Results Spring Planting density Line Interaction Height 4.11E-12** 0.9032 0.0136* Spikelets per spike 3.48E-14** 0.5362 0.6223 Mature spikelets per spike 2.00E-16** 0.03241* 0.87185 Spikes per pot 2.00E-16** 0.147 0.9462 Total seed weight 0.3365 0.05197 0.77064 Average seed weight 2.41E-05** 6.07E-09** 0.6085 Seed count 0.9588 1.73E-06** 0.481 Seed length 0.0509 0.000692** 0.4538 Seed width 1.12E-05** 0.02347* 0.06101 Seeds per spike 2.20E-16** 3.66E-06** 0.5636 Seeds planted vs. Number of spikes 2.00E-16** 0.1441 0.41 Spikelets per spike 3.47E-16** 0.0286* 0.468 Table 3.2: P-values of independent variables planting density, line, and the interaction between planting density and line. * indicates significance at P<0.05, ** indicates significance at P<0.01. 59 Table 3.2 (cont’d) Winter Planting density Line Interaction Height 2.00E-16** 0.277 0.056 Spikelets per spike 3.47E-16** 0.0286* 0.468 Seed width 8.50E-12** 0.03839* 0.85565 Seeds per spike 2.20E-16** 0.002596** 0.27199 Seeds planted vs. number of spikes 2.00E-16** 0.1236 0.6642 Mature spikelets per spike 2.00E-16** 0.4623 0.987 Spikes per pot 2.20E-16** 0.04797* 0.00254** Total seed weight 0.0008054** 0.7317 0.91644 Average seed weight 6.15E-07** 0.002308** 0.4944 Seed count 0.3298 0.01902* 0.9495 Seed length 8.47E-06** 5.12E-09** 0.4446 Plant height Figure 3.1: Average plant height of spring wheat and winter wheat for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. 60 Figure 3.1 (cont’d) Both spring wheat and winter wheat plant height decreased as planting density increased. In spring wheat, height decreased significantly from 2 to 16 seeds/10cm2 and from 16 to 24 seeds/10cm2 (P<0.01). Mean plant height was highest at 188.6 cm at 2 seeds/10cm2 and dropped to 136.1 cm at 35.5 seeds/10cm2. Height was not significantly different for planting densities 16 to 39.5 seeds/10cm2. From 16 to 39.5 seeds/10cm2, mean plant height ranged from 155.75 cm to 136.11 cm. Interaction between planting density and line was significant for height in spring wheat (P<0.05). In winter wheat, height decreased significantly from 2 to 20 seeds/10cm2 (P<0.01). Mean plant height was highest at 56.63 cm at 2 seeds/10cm2 and dropped to 43.22 cm at 16 seeds/10cm2. Height was not significantly different in planting densities 24 to 39.5 seeds/10cm2. From 24 to 39.5 seeds/10cm2, mean plant height ranged from 39.43 cm to 36.36 cm. 61 Spikelets per spike Figure 3.2: Spikelets per spike of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Spikelets per spike include both mature spikelets with seeds and immature spikelets without seeds. Letters are LSD values for best fitted model. Bars are standard error. Both spring wheat and winter wheat spikelets per spike decreased as planting density increased. In spring wheat, spikelets per spike decreased significantly from 2 seeds/10cm2 to 8 seeds/10cm2 (P<0.01). Mean spikelets per spike ranged from 19.312 spikelets at 2 seeds/10cm2 62 to 11.38 spikelets at 39.5 seeds/10cm2. Spikelets per spike were not significantly different in planting densities 12 to 39.5 seeds/10cm2. In winter wheat, spikelets per spike decreased significantly from 2 to 16 seeds/10cm2 (P<0.01). Mean spikelets per spike ranged from 13.26 spikelets at 2 seeds/10cm2 to 9.25 spikelets at 39.5 seeds/10cm2. Spikelets per spike were not significantly different in planting densities 20 to 39.5 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from one another for spikelets per spike (P<0.05). Mature spikelets per spike Figure 3.3: Mature spikelets per spike of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. 63 Figure 3.3 (cont’d) Both spring wheat and winter wheat mature spikelets per spike decreased as planting density increased. In spring wheat, mature spikelets per spike decreased significantly from 2 seeds/10cm2 to 8 seeds/10cm2 (P<0.01). Mean mature spikelets per spike peaked at 15.4 spikelets at 2 seeds/10cm2 and were lowest at 6.8 spikelets at 39.5 seeds/10cm2. Spikelets per spike were not significantly different for planting densities 12 to 39.5 seeds/10cm2. Spring wheat lines Melba and Seahawk were significantly different from one another for mature spikelets per spike (P<0.05). In winter wheat, mature spikelets per spike decreased significantly from 2 seeds/10cm2 to 20 seeds/10cm2 (P<0.01). Mean mature spikelets per spike were highest at 11.66 spikelets at 2 seeds/10cm2 and lowest at 3.96 spikelets at 39.5 seeds/10cm2. Mature spikelets per spike were not significantly different for planting densities 24 to 39.5 seeds/10cm2. 64 Spikes per pot Figure 3.4: Spikes per pot of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for the best fitted model. Bars are standard error. Spikes per pot increased as planting density increased for both spring wheat and winter wheat. In spring wheat, spikes per pot increased significantly from 2 seeds/10cm2 to 20 seeds/10cm2 (P<0.01). Mean spikes per pot ranged from 38.37 spikes at 2 seeds/10cm2 to 279.25 65 spikes at 39.5 seeds/10cm2. Spikes per pot were not significantly different for planting densities 24 to 39.5 seeds/10cm2. In winter wheat, spikes per pot increased significantly throughout the range of the planting densities 2 seeds/10cm2 to 39.5 (P<0.01). Mean spikes per pot started at 30.5 spikes at 2 seeds/10cm2 and increased to 423.125 spikes at 39.5 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from one another for spikes per pot (P<0.05). There was also significant interaction between planting density and line (P<0.01). Total seed weight Figure 3.5: Total seed weight of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. 66 Figure 3.5 (cont’d) Spring wheat and winter wheat experiments exhibited different trends for total seed weight. In spring wheat, total seed weight did not significantly change as planting density increased. Spring seed weight was almost significantly different between lines (P=0.51). The lack of significance among planting densities is likely due to unusually low total seed weights of Seahawk at 20 and 24 seeds/10cm2. Mean total seed weight ranged from 29.38 g. at 39.5 seeds/10cm2 to 48.15 g. at 4 seeds/10cm2. In winter wheat, total seed weight decreased significantly from 4 seeds/10cm2 to 39.5 seeds/10cm2 (P<0.01). Several planting densities were not significantly different from one another: 2, 12, 16, 24, and 31.5 seeds/10cm2. Mean total seed weight ranged from 17.47 g. at 39.5 seeds/10cm2 to 29.23 g. at 4 seeds/10cm2. 67 Average seed weight Figure 3.6: Average seed weight of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. Both spring wheat and winter average seed weight decreased as planting density increased. In spring wheat, average seed weight decreased significantly from 2 seeds/10cm2 to 12 seeds/10cm2 (P<0.01). Average seed weight was highest at 0.052g at 2 seeds/10cm2 and decreased to 0.038 g. at 28 seeds/10cm2. Average seed weight was not significantly different for 68 planting densities 16 to 39.5 seeds/10cm2. Spring wheat lines Melba and Seahawk were significantly different from each other (P<0.01). In winter wheat, average seed weight decreased significantly from 2 seeds/10cm2 to 12 seeds/10cm2 (P<0.01). Average seed weight was highest at 0.04 g. at 2 seeds/10cm2 and lowest at 0.031 g. at 28 seeds/10cm2. Average seed weight was not significantly different for planting densities 16 to 39.5 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from each other (P<0.01). Seed count Figure 3.7: Seed count of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Black dots on plot represent mean seed count for each planting density. 69 Figure 3.7 (cont’d) Both spring wheat and winter seed count did not have significant changes as planting density increased. However, lines within spring and winter wheat were significantly different from each other (P<0.01 for spring and P<0.05 for winter). In spring wheat, mean seed count was higher for Melba at 1394.93 seeds where Seahawk’s mean seed count was 920.84 seeds. Winter wheat seed count was slightly higher for Whitetail than for DF112R. Mean seed count for Whitetail was 761.977 seeds and for DF112R was 657.068 seeds. 70 Seed length Figure 3.8: Seed length of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. 71 Figure 3.9: Seed length of spring wheat lines for planting densities 2 to 39.5 seeds/10cm2. Black dots on plot represent mean seed count for each planting density. Spring wheat and winter wheat had different trends for seed length. In spring wheat, seed length did not have any significant changes as planting density increased. However, spring wheat lines Melba and Seahawk were significantly different from each other (P<0.01). Seahawk seed length was on average larger than Melba. Seahawk seed length was on average 5.89 mm. Melba was on average of 6.035 mm. Winter seed length decreased slightly as planting density increased. In winter wheat, seed length decreased significantly from 2 seeds/10cm2 to 16 seeds/10cm2 (P<0.01). Mean seed length ranged from 6.04 mm. at 16 seeds/10cm2 to 6.728 mm. at 2 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from each other for seed length (P<0.01). 72 Seed width Figure 3.10: Seed width of spring wheat and winter wheat lines for planting densities 2 to 39.5 seeds/10cm2. Letters are LSD values for the best fitted model. Bars are standard error. Both spring wheat and winter seed width decreased as planting density increased. In spring wheat, seed width decreased significantly from 2 seeds/10cm2 to 28 seeds/10cm2 (P<0.01). Seed width ranged from 3.35 mm. at 28 seeds/10cm2 to 3.717 mm. at 2 seeds/10cm2. 73 Spring wheat lines Melba and Seahawk were significantly different from each other for seed width (P<0.05). In winter wheat, seed width decreased significantly from 2 seeds/10cm2 to 8 seeds/10cm2 (P<0.01). Seed width ranged from 3.148 mm. at 31.5 seeds/10cm2 to 3.61 mm. at 2 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from each other for seed width (P<0.05). Seeds per spike Figure 3.11: Seeds per spike of spring wheat and winter wheat lines for planting densities of 2 seeds/10cm2 to 39.5 seeds/10cm2. Letters are LSD values for best fitted model. Bars are standard error. 74 Figure 3.11 (cont’d) Both spring wheat and winter seeds per spike decreased as planting density increased. In spring wheat, seeds per spike decreased significantly from 2 seeds/10cm2 to 12 seeds/10cm2 (P<0.01). Mean seeds per spike ranged from 3.68 seeds at 39.5 seeds/10cm2 to 30.76 seeds at 2 seeds/10cm2. Spring wheat lines Melba and Seahawk were significantly different from each other for seeds per spike (P<0.01). In winter wheat, seeds per spike decreased significantly from 2 seeds/10cm2 to 16 seeds/10cm2 (P<0.01). Mean seeds per spike ranged from 1.34 seeds at 39.5 seeds/10cm2 to 21.18 seeds at 2 seeds/10cm2. Winter wheat lines DF112R and Whitetail were significantly different from each other for seeds per spike (P<0.01). 75 Seeds planted vs. number of spikes Figure 3.12: Seeds planted vs number of spikes for spring wheat and winter wheat lines for planting densities 32 to 640 seeds per pot. Letters are LSD values for best fitted model. Bars are standard error. Both spring wheat and winter seeds planted vs number of spikes increased as planting density increased. In spring wheat, seeds planted vs number of spikes increased significantly 76 from 192 to 576 seeds per pot (P<0.01). Seeds planted vs number of spikes ranged from -6.37 to 360.75. In winter wheat, seeds planted vs number of spikes increased significantly from 32 to 576 seeds per pot (P<0.01). Seeds planted vs number of spikes ranged from 1.5 to 216.875. Planting density vs line Most of the traits evaluated are significantly different (P<0.05) between the genotypes used in spring and winter wheat studies (Table 3.2). Adjusted R square values were generated to determine the proportion of variance for each trait that is explained by planting density and line (Table 3.3). For traits that were indicated to be significant for both planting density and line, only three had adjusted R square values that were greater for line than planting density (Table 3.3). The traits were mature spikelet count and average seed weight in spring wheat, and seed length for winter wheat. Traits that were indicated to be significant for just line had larger adjusted R squared values than planting density. The traits were seed count and seed length for spring wheat, and seed count for winter wheat. However, all adjusted R squared values were less than 0.24. 77 Spring Planting density Line Height 0.53200 -0.01100 Spikelets per spike 0.64200 -0.00990 Mature spikelets per spike 0.00528 0.69960 Spikes per pot 0.73000 -0.00440 Total seed weight 0.01862 0.03300 Average seed weight 0.20300 0.25600 Seed count -0.09160 0.24900 Seed length 0.07550 0.10670 Seed width 0.28700 0.02630 Seeds per spike 0.78570 0.03720 Seeds planted vs. number of spikes 0.86630 -0.00828 Winter Planting density Line Height 0.79300 -0.00906 Spikelets per spike 0.67080 0.00660 Mature spikelets per spike 0.83770 -0.01040 Spikes per pot 0.95900 -0.01010 Total seed weight 0.24940 -0.01040 Average seed weight 0.35970 0.05570 Seed count 0.02100 0.05537 Seed length 0.21800 0.24890 Seed width 0.57180 0.01090 Seeds per spike 0.92880 -0.00456 Seeds planted vs. number of spikes 0.83950 -0.00714 Table 3.3: Adjusted R squared values for planting density and line. Bolded values indicate adjusted R square values that were greater for line than planting density for traits that were significant for both planting density and line. Underlined values indicate traits that were significant for line only and had larger adjusted R squared values than planting density. Discussion The traits evaluated for spring and winter planting density studies followed one of three trends: decreasing as planting density increases, increasing as planting density increases, or followed no trend. 78 Traits that decreased as planting density increased Height, spikelets per spike, mature spikelets per spike, average seed weight, seed width, and seeds per spike all decreased as planting density increased for both spring and winter wheat. Total seed weight and seed length for winter wheat also decreased as planting density increased. These findings would be useful for breeders or growers if their goal is to maximize one, or several, of these traits. To maximize these traits, such as generating a higher number of seeds per spike, a lower planting density should be used in greenhouse conditions. Traits that increased as planting density increased Spikes per pot and seeds planted vs number of spikes increased as planting density increased for spring and winter wheat. As planting density increases, more spikes are produced but less of the seeds planted reach maturity. If the goal of the breeder is to maximize the number of spikes that reach maturity, the breeder must also be prepared to lose more seeds that would have otherwise reached maturity at a lesser planting density. The limiting factor would then be the seed count from the previous generation. Traits that followed no trend as planting density increased Total seed count followed no trend as planting density increased for both spring and winter wheat. Total seed weight for spring wheat also had no trend, however this was likely due to the extremely low seed weights of seahawk at 5 and 6 seeds/in2. The low seed weights for those samples were potentially due to the placement of those pots in the greenhouse. Increasing 79 or decreasing planting density will not produce significantly more or less seeds. If the goal of a breeder or grower is to maximize the number of seeds produced in greenhouse conditions, then they should use larger pots or increase the number of replications used. Planting density vs line Seed count does not significantly change as planting density increases, however seed count changes significantly between genotypes. This finding may have implications for breeding programs that utilizes a modified bulk breeding method to rapidly advance populations in early generations under greenhouse conditions. The wheat breeding program at Michigan State University (MSU) uses this method to rapidly advance populations from the F2 to F4 stage. The lowest seed counts occurred in DF112R, some of which were less than the maximum number of seeds used in MSU’s modified bulk breeding system (500 seeds at F3). This system may have to be adjusted for genotypes that produce less seed in greenhouse conditions, such as planting less seed in the F3 generation. These low seed counts occurred in the highest planting densities, 9 and 10 seeds per square inch or 576 and 640 seeds. The low seed counts may not occur often if the planting densities are lower. Adjusted R squared values were used to determine the proportion of variance for each trait that is explained by planting density and line. This was particularly useful in determining the impact of planting density and line for traits that were significant (P<0.05) for both independent variables. Of those traits, only three had adjusted R squared values greater for line than planting density. Average seed weight in spring wheat and seed length in winter wheat had slightly larger adjusted R square values for line than planting density. However, mature spikelets per spike in spring wheat had a far larger adjusted square value for line than planting density. This suggests that mature spikelets per spike in spring wheat may vary depending on the genotype. 80 Future implications Although the current use of the information gathered in these experiments are useful to plant breeding, there may be future applications towards vertical farming. Asseng et al. (2020) described how up to 1940 +/- 230 t/ha of grain can be harvested annually in a hectare indoor vertical facility. Using vertical farming for growing grain crops is currently less feasible due to energy inputs but may be a viable practice in the future (Asseng et al., 2020). Additional information regarding the impacts of genotypes and planting density should be gathered in future studies that include more lines in the experimental design. These experiments could have also been more standardized by being conducted in the same greenhouse during the same time of year. Understanding the impacts of planting density on plant architecture and yield will allow breeders and growers to make informed decisions for indoor growing practices. 81 BIBLIOGRAPHY 82 BIBLIOGRAPHY Asseng S, Guarin JR, Raman M, Monje O, Kiss G, Despommier DD, Meggers FM, Gauthier PPG (2020) Wheat yield potential in controlled-environment vertical farms. Proc Natl Acad Sci USA 117: 19131 Black CA, Nelson LB, Pritchett WL (1947) Nitrogen Utilization by Wheat as Affected by Rate of Fertilization. Soil Science Society of America Journal 11: 393–396 Caffaro MM, Vivanco JM, Botto J, Rubio G (2013) Root architecture of Arabidopsis is affected by competition with neighbouring plants. Plant Growth Regulation 70: 141–147 Cisar G, Howey AE, Brown CM (1982) Optimal Population Density and Random vs. Selective Elimination of Genotypes Under a Modified Single-Seed Descent Method with Spring Oats1. Crop Science 22: cropsci1982.0011183X002200030034x Dudley SA, File AL (2007) Kin recognition in an annual plant. Biology Letters 3: 435–438 Gersani M, Brown J s., O’Brien EE, Maina GM, Abramsky Z (2001) Tragedy of the commons as a result of root competition. Journal of Ecology 89: 660–669 Milla R, Forero DM, Escudero A, Iriondo JM (2009) Growing with siblings: a common ground for cooperation or for fiercer competition among plants? Proc Biol Sci 276: 2531–2540 Nord EA, Zhang C, Lynch JP (2011) Root responses to neighbouring plants in common bean are mediated by nutrient concentration rather than self/non-self recognition. Functional Plant Biol 38: 941 Olson EL, Noble A, Forsberg L, Brisco-McCann E (2021) The 'Minibulk' system for advancing winter cereal breeding programs. Manuscript in preparation. Zhu Y-H, Weiner J, Li F-M (2019) Root proliferation in response to neighbouring roots in wheat (Triticum aestivum). Basic and Applied Ecology 39: 10–14 83