An Ethylmethane Sulfonate Mutant Resource in Pre-Green Revolution Hexaploid Wheat

Mutagenesis is a powerful tool used for studying gene function as well as for crop improvement. It is regaining popularity because of the development of effective and cost efficient methods for high-throughput mutation detection. Selection for semi-dwarf phenotype during green revolution has reduced genetic diversity including that for agronomically desirable traits. Most of the available mutant populations in wheat (Triticum aestivum L.) were developed in post-green revolution cultivars. Besides the identification and isolation of agronomically important alleles in the mutant population of pre-green revolution cultivar, this population can be a vital resource for expanding the genetic diversity for wheat breeding. Here we report an Ethylmethane Sulfonate (EMS) generated mutant population consisting of 4,180 unique mutant plants in a pre-green revolution spring wheat cultivar ‘Indian’. Released in early 1900s, ‘Indian’ is devoid of any known height-reducing mutations. Unique mutations were captured by proceeding with single M2 seed from each of the 4,180 M1 plants. Mutants for various phenotypic traits were identified by detailed phenotyping for altered morphological and agronomic traits on M2 plants in the greenhouse and M3 plants in the field. Of the 86 identified mutants, 75 (87%) were phenotypically stable at the M4 generation. Among the observed phenotypes, variation in plant height was the most frequent followed by the leaf morphology. Several mutant phenotypes including looped peduncle, crooked plant morphology, ‘gritty’ coleoptiles, looped lower internodes, and burnt leaf tips are not reported in other plant species. Considering the extent and diversity of the observed mutant phenotypes, this population appears to be a useful resource for the forward and reverse genetic studies. This resource is available to the scientific community.


Introduction
Functional genomic resources are important to determine function of the genes particularly that control agronomic traits. Gene alteration through mutagenesis is an important approach to generate such functional resources. Ethylmethane Sulfonate (EMS) has been the mutagen of choice in forward genetics approach to generate mutagenized populations in a wide array of species. EMS is an alkylating agent that causes mainly point mutations due to the addition of an alkyl group on guanine resulting in transition of GC to AT. EMS induces mutations creating allelic versions of genes resulting in phenotypes of varying intensity [1]. The forward genetic approach enables identification of agronomically desirable phenotypes that can be exploited in breeding programs. In wheat, the forward genetics approach has been successfully used to identify mutants of agronomic importance including imidazolinone tolerance [2], waxiness, and grain hardness [3] (http://wheat.pw.usda.gov/GG2/Triticum/wgc/2008/).
A number of publicly available mutant populations have been developed in various cultivars of bread wheat (Triticum aestivum L.) but their utility has been limited to only few traits with discernible phenotypes and to genes with relatively well-characterized functions. Examples of such genes include waxy in elite cultivars Express [4], Cadenza [5], and Ventura [6]; puroindolines in Australian soft-textured grain cultivar QAL2000 [6] and Alpowa [7]; Wheat Kinase Start (WKS) 1, WKS2, and Starch Branching Enzyme (SBE) IIa in breeding line UC1041+Gpc-B1/Yr36 [8], photoperiod (Ppd)-D1, Rubisco activase A and B [9], and starch synthase II [5]. Depending upon the EMS concentration, detection approach, and the target gene, the reported mutation frequency in hexaploid wheat ranged from 1/47 kb to 1/12 kb [6,9]. This frequency is expected to increase significantly with the adoption of more sensitive detection methods [10]. The available mutant populations are in the post-green revolution cultivars carrying 'reduced height' (rht) genes that may be lacking many useful alleles lost during selections leading to the green revolution. Further, with the changing environmental conditions and the ill-effects of the currently used rht dwarfing genes on a number of plant attributes including coleoptile and first leaf length, and primary and secondary roots and root biomass, it is important to develop a mutant resource in a non-rht genotype [11,12]. The objective of our study is therefore to generate and characterize an EMS mutant population in 'Indian', a tall hexaploid wheat cultivar released in early 1900s.

Plant material
A soft white spring hexaploid wheat cultivar Indian (CItr 4489) [13] with an average height of 103 cm, six tillers, and semi-awned spikes was used for mutagenesis. Being the tallest among the older cultivars in our world wheat collection [14], it lacks any of the known dwarfing mutations and possibly contains valuable alleles that got selected out during breeding over the decades.

EMS mutagenesis
Each of the initial experiment to optimize EMS concentration was carried out using 200 seeds of cultivar Indian. EMS concentrations of 0, 10,20,30,40, and 50 mM were tested to select 40 mM concentration (see results) to carry out the subsequent experiment using 600 gm seed of cultivar Indian.
For mutagenesis, 200 seeds were soaked in 40 ml of de-ionized water (3,000 ml for 600 gm of seed) in a flask that was kept on a benchtop shaker (Barnstead Labline-MaxQ 3,000), for four hours with gentle shaking at room temperature, with hourly change of water. After four hours, the water was replaced with the appropriate concentration of EMS solution and the flasks were kept on the shaker for 18 hours. The EMS solution was removed and the seeds were washed twice with 100 mM sodium thiosulphate solution for 30 minutes each. Seeds were dried for 24 hours on tissue papers at room temperature before planting at a very high density (to facilitate growth of only single tillers) in large plastic trays with holes containing the professional growing mix (Sunshine LC1 mix). The plants in the trays were grown in the greenhouse (Wheat Growth Facility, WSU) with 16 hours light, and 22°C temperature during the day and 18°C during the night in 2010. The main tiller of each of the surviving 4,180 M 1 plants was individually harvested and thrashed. In 2011, single M 2 seed from each of the M 1 plants was planted in the greenhouse. The M 3 seed from each of the surviving M 2 plants was bulk harvested and planted in six feet rows at the Spillman Agronomy Farm, Pullman, WA in 2011. For the selected 86 mutants, single plant showing the best mutant phenotype was harvested. For the remaining population, M 4 seed of each line was bulk harvested and stored at -20°C. The M 4 generation of the selected mutants was grown in three feet rows at the Spillman Agronomy Farm, Pullman, WA in 2012.

Greenhouse and field evaluation of the mutant population
Detailed phenotypic data at the tillering and maturity stages was collected on 1,902 of the 4,180 greenhouse grown M 2 plants of the mutant population. At the tillering stage, plants were scored for the plant architecture, number of tillers, and leaf and stem morphology. Data on plant height and tiller uniformity was collected at maturity along with the number of internodes, and stem and spike morphology. The remaining 'fertile' 1,561 M 2 plants from the population were grown in 2013 for seed multiplication.
With the objective to: a) study the stability of the mutant phenotype; and b) multiply the seed; approximately 60 M 3 seeds of each of the 1,902 M 2 plants were planted in field. Four rows were planted in each plot with a row-to-row and plot-to-plot spacing of one foot. Each row represented progeny of a single M 2 plant. For comparison, control cultivar Indian was planted after every 19 mutant rows. Starting at the seedling stage, M 3 plants were evaluated throughout the growing season for phenotypic differences in comparison to the wild type Indian. Leaf related phenotypes including leaf spots, leaf width, and leaf color were more obvious at the tillering stage. Tiller number, peduncle, stem, spike, and spikelet related morphologies were scored during later stages of the plant development. Data on plant height was collected on all the 1,902 plants of this population. The selected M 3 mutants were grown again in three-feet rows in 2012 with the similar field layout to study stability of the mutations. M 4 plants were evaluated similarly throughout the growing season for phenotypic differences in comparison to the wild type Indian. The remaining 1,561 of the 4,180 M 3 plants were seedincreased by growing in three-feet rows with a similar field layout in 2014.

Identification and evaluation of mutations over generations
The procedure to develop and evaluate mutant plants over generations is given in Fig 1. About 13% (278) of the 2,180 M 2 seeds either didn't germinate or produced nonviable plants. Viable plants (1,902) were compared to control Indian to identify the plants showing a mutant phenotype. Compared to control Indian, 453 of the 1,902 M 2 plants were putative mutants that showed phenotypic differences for the plant traits other than height and peduncle. Of these putative mutants, 155 showed uniform tillers compared to that of non-uniform tiller growth in control Indian and 141 showed sturdier stem (Table 1). With an objective to identify height mutations in pathways other than giberrellic acid (GA), the entire population was evaluated to select plants with height comparable to that of the currently grown wheat varieties. Compared to the average plant height of 103 cm (range of 87-113 cm) for the control Indian in the greenhouse, 733 plants showed plant height less than 87 cm, thus were selected as putative height mutants. Among these plants, 100 showed a height of 60 cm (dwarf) and 633 from 61 to 86 cm (semi-dwarf). 157 M 2 plants were taller than the control with height ranging from 114-137 cm. The remaining 1,012 M 2 plants showed plant height ranging from 87-113 cm, similar to that of the control Indian (Fig 2A).  Of M 3 progeny of the 1,902 M 2 plants grown in the field as head-rows, 1,690 survived. The progeny of remaining 212 either didn't germinate or were nonviable. A row was considered to be a phenotypic mutant if two or more plants showed the mutant phenotype in comparison to the wild-type Indian. Using this criterion, 240 M 3 rows were identified to be carrying phenotypic mutants for the traits other than height. Among these, leaf morphology including leaf shape, angle, and color, was the most frequent variation with 69 M 3 lines followed by 66 M 3 lines for early flowering ( Table 2). Leaf morphology included spots of varying sizes with a range in phenotypic severity (lesion mimic) (Fig 3B and 3C); leaf color of dark green, yellowgreen, silver-green, and purplish pigmentation (Fig 3); and upright leaf angle, curled, and crinkled leaves ( Table 2). Of the 37 mutants for spike morphology (Fig 4A-4G and Table 2), plants with longer awns, compared to medium awns in control Indian, were the most frequent type. Among mutant plants for stem morphology, sturdy stem was the predominant phenotype. The least frequent variations included iojap, uniculm, lazy, and bent nodes ( Table 2). Both flag leaf and peduncle showed variations for size and growth pattern (Fig 3I-3L). Among 240 mutants other than the height mutants, 59 showed more than one mutations either for leaf color, leaf shape, stem thickness, early flowering, bent nodes, awned/awnless spike, deformed spike, and uniculm.
Compared to the average height of 105 cm with a range of 90 to 125 cm for the wild-type Indian, significant variation was observed among M 3 mutants for the plant height in the field (Fig 2B). Height of 654 M 3 lines was less than 90 cm of which 32 were considered to be dwarf with an average height of 60 cm and the remaining 622 were semi-dwarf with plant height ranging from 61 to 89 cm. Plant height of the remaining 1,036 M 3 plants ranged from 90-125 cm similar to that of wild-type Indian.
The height of each mutant was compared between M 2 and M 3 generations. The control Indian has an average height of 103 cm [Coefficient of variation (CV) 9.3%] in M 2 and 105 cm (CV 8.4%) in M 3 generation. Considering average CV of 8.9% for control Indian, height of 662 M 2 mutants remained same in M 3 generation while 540 mutants showed decrease and 484 mutants showed increase in height. The M 4 progenies of 86 of the above-mentioned mutants selected for their obvious phenotypes were grown as rows in the field for phenotype stability. Thirty-five of these mutants were for the traits other than plant height including variation for spike, leaf, stem, and peduncle. Nine of these mutants were 'uniculm', showing only a single tiller. The remaining 42 mutants showed variation in plant height ranging from 33 to 87 cm including the agronomically desirable height range of 60 to 80 cm as observed for the popular wheat varieties grown in the same field ( Fig 5). Control Indian has average height of 105 cm (CV 2.6%) in M 4 generation. Considering the average CV of 5.5% of control Indian over M 3 and M 4 generation, 26 of these 42 mutants showed reduction in plant height, eight showed an increase, and the remaining eight mutants showed no change in height.

Cataloging of the stable mutations
Stability of mutations was evaluated by assessing the same mutations over generations to rule out the environmentally influenced phenotypes. From the selected 733 M 2 mutants with height <87 cm, 168 either did not germinate or were nonviable in M 3 generation but 258 showed height similar to that of the previous generation. For the remaining 307 putative mutants, the phenotype was not evident in M 3 generation and their height was similar to that of the wildtype Indian. Plant height of these 307 putative mutants during M 2 generation ranged from 40 to 86 cm. Therefore only 35% of the height mutants remained stable at M 3 generation. Stability of the mutant phenotype after M 3 generation was high. All selected 42 height mutants (see above) showed a height phenotype less than wild-type Indian in M 3 and M 4 ( Table 3).
Stability of the mutants other than that for height was also tested over generations. Of the 453 plants identified to be mutants during the M 2 generation, only 24 (5%) showed the same phenotype in M 3 generation and the remaining 95% behaved as control Indian. Seven of these 24 mutants along with additional 37 M 3 mutants selected on the basis of phenotypic traits and single tiller in the M 3 generation were further grown in the field to test for the stability in the M 4 generation. Of these, phenotype of 33 (75%) mutants was stable at this generation. Overall, 75 mutants showing reliable and stable phenotype including height over generations were selected (Table 4). These included lesion mimic, yellow green leaves, dark green leaves, anthocyanin pigmentation, early senescence, awned, grassy, upright leaves, and uniculm mutants (Figs 3, 4 and 6 and Table 2). However, to the best of our knowledge, several of the mutants identified during this study were novel and have not been reported in other plant species (Table 5). These include 'burnt leaf tips' (Fig 3E), looped peduncle (Fig 4I), crooked plant morphology (Fig 4L), looped lower internodes (Fig 6F), and 'gritty' coleoptiles ( Fig 6G). 'Lazy' mutant was also identified for the first time in hexaploid wheat ( Fig 6I).

Analysis of stable height mutants
The stable 42 height mutants in M 4 generation were analyzed for relative internode length in comparison to the plant height (Table 6). Percent internode was calculated as the contribution of each internode towards the total plant height. Compared to control Indian, the mutants differed significantly for the relative internode length (Table 6). About equal number of the height mutants showed an increase and decrease in internode 1 length. Relatively fewer plants showed a change in internode 2 or 3. Maximum number of mutants (25/42) showed reduction in internode 4. Twenty of these mutants along with two other mutants showed no visible 5 th internode.

Identification of mutations in pre-green revolution cultivar
Most of the mutant populations available in hexaploid wheat are developed in recently released cultivars carrying rht mutations. Because of the genetic uniformity created by the use of common rht mutations, and breeding and selection during the last few decades, the popular wheat varieties around the world have a narrow genetic base and probably are lacking valuable gene/allele combinations. Therefore, we focused on generating a mutant population in a pregreen revolution tall cultivar Indian released in the early 1900s.
In wheat, it is relatively easy to identify and recover mutations even in vital genes due to partial compensation by the homoeologs. In hexaploid wheat, about 86% of the genes show significant differences among homoeologs either for the expression level and/or for tissue specificity [15]. These results suggest that most of the hexaploid wheat homoeologs may have diverged to develop subtle differences in their functions thus making it possible to study effect of each copy individually. For some traits such as plant height, peduncle length, tiller number, plant maturity, awns, and leaf spots, a number of mutants showing a range of phenotypes from mild to severe were identified during the present study. These results might be due to complex genetic control of these traits involving multiple genes operating in complex interactions. Part of this phenotypic range is probably due to the involvement of multiple homoeologues corresponding to the underlying genes.

Succession of mutations over generations
Various concentrations of EMS have been successfully used in wheat to produce mutant populations but an ideal concentration is not known that induces maximum number of mutations per plant while maintaining a higher proportion of fertile plants. Genotypic differences for the response to EMS have also been reported. Furthermore, vendor-to-vendor and batch-to-batch differences in the potency of EMS have also been speculated. After testing a full range of concentrations, 40 mM was identified to be an ideal as it gave the highest proportion of mutant phenotypes. The 50 mM concentration resulted in a higher proportion of aberrant mutants but was not selected as the proportion of fertile mutants was severely reduced than the 40 mM concentration (data not shown).
Different tillers in M 1 plant are known to be chimeric for mutations and the main tiller is known to have the highest proportion of mutations [16]. Therefore, we planted the M 1 generation at a very high plant density to ensure only single tiller for each of the M 1 plants. Redundancy within the mutant population was avoided by taking a single M 2 seed from the main spikes of each of the M 1 plants to make the M 2 generation. In case of heterozygous mutations in a M 2 seed, only one-fourth of the M 3 seeds should be homozygous recessive for the mutation(s) thus the probability of losing a mutation in M 3 generation is small due to a large number of plants sown per mutant line in the field.
A large number of mutations observed during M 2 generation were not visible at M 3 generation were thus unstable. Only 5% (24) of the phenotypic and 35% (258) of the height M 2 mutants were stable at M 3 generation. Mutants identified in M 3 generation showed much

Enhancement of genetic diversity for wheat breeding
About 95% of the wheat grown in the world contains the same rht mutations that are associated with negative impact on seedling emergence and vigor, root biomass, coleoptile, and leaf width [12,11]. Some of these traits are directly linked with plants' ability to deal with abiotic stresses such as drought and temperature that are expected to become more prominent in the coming years. Therefore, one of our objectives was to identify different types of dwarfing mutations without the ill effects known to be associated with the rht mutations. Probably because of our focus, plant height was the most abundant class of mutations accounting for 57% of the total stable mutants identified during this study. In spite of the potential bias for selection, our results suggest that plant height is a complex trait controlled by many different genes. There were significant differences among mutants for the relative reduction in internode length thus it would be possible to select an ideal dwarf plant out of the mutant collection. About 50% of the stable height mutants were very dwarf with plant height ranging from 33 to 59 cm and 21 showed plant height in the range of currently grown wheat varieties thus might have an  agronomic value. These mutants are now being evaluated for their GA sensitivity in order to identify height mutants in a pathway other than GA, as GA is known to play an important role in abiotic stress tolerance.
Based on the replication of phenotype and the height of selected mutants in M 4 generation, we have identified 75 stable M 4 mutants. These mutants might serve as a useful resource for improving the plant architecture, agronomic performance, and genetic studies. Tiller number controls the number of spikes and is one of the key components in controlling wheat yield. A mutant in rice-Monoculm 1 results in single tiller [17]. Its orthologs in Arabidopsis and tomato regulate branching [18,19]. In Triticum monococcum subsp. monococcum, single tiller mutant named tiller inhibition (tin3) [20] and tin in hexaploid wheat resulted in increased number of kernels per spike [17]. We found a similar uniculm mutant in our mutant population that remained stable over generations. This mutant could serve as a useful resource to understand the underlying mechanism of axillary meristem initiation, bud development, and tiller elongation in wheat.
We have identified a mutant for tiller angle in wheat. The mutant showed hyper-gravitropic response and bends towards soil surface in one direction along with short plant height and the wild type spike length. Similar type of mutants known in other tillering grasses have prostrate growth habit but with tillers spreading in all the directions. Therefore, it is a useful mutant to study the physiological mechanism controlling directional manifestation of the tiller angle. A number of mutants for tiller angle due to the disruption of polar auxin transport and gibberellins biosynthesis have been identified in different plant species including rice, maize, Arabidopsis, and sorghum [21][22][23]. Additionally, we found a mutant with upright leaves as a potential candidate to modify the plant architecture. Erect leaf architecture reduces shading on the lower leaves and allows better distribution of light to the lower portion of the plant, thus enables a higher plant density [24]. Erect leaf trait has also been shown to be beneficial in maize and rice to increase grain yield [25,26]. Among the identified novel mutants, the physiological mechanisms behind stem bending only in restricted areas such as peduncle and lower internodes are of great importance to study. Stable nature of these mutants over generations eliminates the effect of external signals. Stem bending in plants is known to be regulated through auxin distribution that results in irregular cell division and elongation [27,28]. Therefore, the targeted bending of peduncle and lower internodes in the two newly identified mutants might involve the differential distribution of local auxin hormone. Coleoptile protects the first leaf during emergence until it comes through the soil surface and provides mechanical force for seedling emergence. The gritty coleoptile observed in our population consisting of thick layer of tissue with a white, rough, and sturdy phenotype is an important mutant to study the plant emergence under deep and compact soils.
Additionally, mutants found with phenotypes similar to some of the already identified mutants in other cultivars of wheat may represent an additional set of allelic versions of the corresponding genes. This population can also be a potential resource for physiological screening of diseases as we have phenotypically observed variations for rust resistance in the population (data not shown).

Detection of mutations through sequencing
In previously reported mutant populations, the frequency of mutations ranged from 1/47 kb to 1/12 kb based on few targeted genes in population size of 1,152 to 2,610 with an EMS concentration of 0.8 to 1% [4,9]. Using Genotyping-by-Sequencing (GBS), EMS induced mutations were identified in a subset of 21 mutants of this population [10]. With 21 mutants and two control Indian plants, 28% of wheat genic sequence was covered by the reads generated by GBS., A total of 14,130 mutational changes [Single Nucleotide Polymorphisms (SNPs) and Insertions/ Deletions (INDELs)] identified from 10,374 unigenes corresponding to 72 Mb genic region thus resulted in a frequency of one mutational change per 5 kb of the wheat genome. In addition to the accumulative frequency, a mutation frequency ranging from 1/2.5 kb to 1/6.4 kb was identified from individual mutant plants [10]. Sequence data from this population has been submitted to the National Center for Biotechnology Information (NCBI) Sequence Read Archive (http://www.ncbi.nlm.nih.gov/sra/?term=SRP056743). Wide range of phenotypic mutants and higher number of genomic sequence changes observed in this population establishes it as a valuable resource for forward and reverse genetic studies in wheat.

Conclusions
The EMS induced mutant population in pre-green revolution cultivar Indian is extensively evaluated and cataloged for different phenotypic traits to increase its utility and is available for the wheat community. Novel mutant phenotypes identified in this population will serve as a useful resource for both genetic and physiological studies in wheat. The mutants identified for desirable agronomic traits can be utilized to transfer these mutations to other cultivars.