Correspondence: John A. Juvik, tel. +217 333 1966, fax +217 244 6342, e-mail: firstname.lastname@example.org
Plants from the genus Miscanthus are potential renewable sources of lignocellulosic biomass for energy production. A potential strategy for Miscanthus crop improvement involves interspecific manipulation of ploidy levels to generate superior germplasm and to circumvent reproductive barriers for the introduction of new genetic variation into core germplasm. Synthetic autotetraploid lines of Miscanthus sacchariflorus and Miscanthus sinensis, and autoallohexaploid Miscanthus x giganteus were produced in tissue culture from oryzalin treatments to seed- and immature inflorescence-derived callus lines. This is the first report of the genome doubling of diploid M. sacchariflorus. Genome doubling of diploid M. sinensis, M. sacchariflorus, and triploid M. x giganteus to generate tetraploid and hexaploid lines was confirmed by stomata size, nuclear DNA content, and chromosome counts. A putative pentaploid line was also identified among the M. x giganteus synthetic polyploid lines by nuclear DNA content and chromosome counts. Comparisons of phenotypic performance of synthetic polyploid lines with their diploid and triploid progenitors in the greenhouse found species-specific differences in plant tiller number, height, and flowering time among the doubled lines. Stem diameter tended to increase after polyploidization but there were no significant improvements in biomass traits. Under field conditions, M. x giganteus synthetic hexaploid lines showed greater phenotypic variation, in terms of plant height, stem diameter, and tiller number, than their progenitor lines. Production of synthetic autopolyploid lines displaying significant phenotypic variation suggests that ploidy manipulation can introduce useful genetic diversity in the limited Miscanthus germplasm currently available in the United States. The role of polyploidization in the evolution and breeding of the genus Miscanthus is discussed.
Polyploidy is a common genomic feature in all eukaryotes and is particularly prominent in plants. It is broadly classified into auto- and allopolyploidy, which are formed by doubling the diploid genome within a species and from hybrids between species, respectively. Polyploidization, together with hybridization, plays an important role in plant evolution and speciation (Soltis & Soltis, 2009). In contrast to the gradual evolutionary process whereby new species evolve from isolated populations, new species of plants can emerge abruptly via polyploidization. Differences in chromosome number provide instant reproductive isolation, limiting gene flow between newly formed polyploids (neopolyploids) and their progenitors (Tate et al., 2005; Otto, 2007). Furthermore, physiological and morphological changes due to polyploidization may alter reproductive biology of neopolyploids (Thompson & Lumaret, 1992).
Polyploidization can alter cytogenetic, genetic, and epigenetic characteristics of organisms resulting in phenotypic variation among neopolyploid plants, which would be a target for natural selection. Genome duplication increases cell volume by increasing genome size and has been associated with delayed development of polyploid plants. Larger genomes require more time for cell replication and tend to display slower cell growth rates (Bennett & Leitch, 2005). A slower cell cycle can result in delayed and/or prolonged flowering (Ramsey & Schemske, 2002) by postponing the termination of apical growth which can lead to a longer period of vegetative growth (Salas Fernandez et al., 2009).
The genome of newly formed polyploid plants usually undergoes extensive genetic and epigenetic changes. Genomes of neopolyploids are usually unstable and experience rapid repatterning (Wendel, 2000) often beginning in the first generation after polyploidization (Levy & Feldman, 2004). The extensive and rapid genomic rearrangements are likely due to sequence rearrangements, homoeologous recombination, and sequence elimination (Adams & Wendel, 2005; Otto, 2007). The heritable epigenetic changes, such as DNA methylation, histone modification, and RNA interference can also alter gene expression (Wolffe & Matzke, 1999; Liu & Wendel, 2003; Doyle et al., 2008). In addition, the activation or suppression of transposable elements is considered an important component of the evolution of polyploid genomes (Matzke & Matzke, 1998). These genetic and epigenetic changes can result in phenotypic variation.
Miscanthus species are perennial C4 grasses and have a base chromosome number of x = 19, with nominally diploid (2n = 2x = 38) and tetraploid (2n = 4x = 76) accessions. One of these species, Miscanthus x giganteus (2n = 3x = 57) is a candidate for dedicated bioenergy crop production due to its high biomass productivity and capacity to capture greenhouse gases by sequestering carbon in underground rhizomes (McLaughlin & Walsh, 1998). While M. x giganteus is considered an excellent bioenergy crop due to its high biomass productivity and cold tolerance (Heaton et al., 2010), it has limited genetic diversity with very few different genotypes available in Europe and the United States (Greef et al., 1997). Introducing new variability in M. x giganteus is constrained by the sterility associated with its triploid genome. The fertile, putative parental species of M. x giganteus, Miscanthus sinensis, and Miscanthus sacchariflorus are also potential candidates as bioenergy crops in that they can produce substantial biomass in certain environments (Clifton-Brown et al., 2001) and show extensive genetic variability in their native ranges in eastern Asia (Jakob et al., 2009; Sacks et al., 2012).
In the United States, M. sinensis was first introduced from Japan in the late 1800s (Quinn et al., 2010) and various Miscanthus accessions have been introduced thereafter, mainly as ornamentals. Efforts are underway to add diverse Miscanthus genotypes to the current germplasm pool in the United States with accessions from Korea, Japan, and China, but restrictions associated with importation has been a major barrier and slowed germplasm acquisition (Jakob et al., 2009). Currently at University of Illinois Urbana-Champaign (UIUC), more than 101 accessions of Miscanthus have been collected (mostly ornamentals) and planted both in the field and in the greenhouse. Within this collection, most are of M. sinensis with only a few diploid and tetraploid M. sacchariflorus and M. x giganteus accessions (Chae, 2012).
Potential strategies for Miscanthus crop improvement include the resynthesis of new triploid M. x giganteus genotypes by conventional hybridization between tetraploid M. sacchariflorus and diploid M. sinensis, intra- and interspecific hybridization among diploid compatible Miscanthus species followed by successive selection, and by manipulation of ploidy levels to circumvent reproductive barriers (Heaton et al., 2010). Although prevalent in Japan (Nishiwaki et al., 2011), few tetraploid accessions of M. sacchariflorus, one of the presumed parents of M. x giganteus, are currently available in the United States. Therefore, producing tetraploid M. sacchariflorus plants would be desirable to resynthesize new genotypes of triploid M. x giganteus germplasm. The production of triploid plants can also circumvent invasive issues associated with the two fertile seed-bearing parental species (Quinn et al., 2010).
Certain diploid Miscanthus species can produce fertile seed from interspecific hybridization and efforts to produce such hybrids have been made for decades. In Japan, hybrid populations from crosses between M. sinensis and M. floridulus (Adati & Shiotani, 1962) and between M. sinensis varcondensatus and M. tinctorius (Hirayoshi et al., 1959) were produced. Recently, several populations from crosses between M. sinensis and M. sacchariflorus have been produced by Mendel Biotechnology, Inc. (Hayward, CA, USA) and by our research group at UIUC (Sacks et al., 2012). There is also potential to generate progeny from crosses between plants with different ploidy levels. Triploid Miscanthus hybrids were produced from conventional crosses in Japan (Hirayoshi et al., 1960) and recently, we have generated new triploid M. x giganteus plants from the cross of diploid M. sinensis by tetraploid M. sacchariflorus (Chae, 2012).
Whole genome duplication via polyploidization is thought to have occurred in the genus Miscanthus in the process of or after divergence from the closely related Saccharinae clade (Swaminathan et al., 2012) fewer than 3 million years ago (Paterson et al., 2010). Also, the genus Miscanthus contains accessions with varying ploidy levels in the wild (Hodkinson & Renvioze, 2001). Genetic and epigenetic changes associated with genome doubling in Miscanthus could result in phenotypic variation for biomass characteristics and delayed flowering time. The objective of this study was to produce synthetic polyploids from clonal accessions of three Miscanthus species to compare cytogenetic and phenotypic differences among polyploid lines and their progenitor genotypes.
Materials and methods
To define terms used in this article, a ‘line’ refers to plants generated from the same callus that represents an independent event of genome doubling. A ‘regenerated’ and ‘oryzalin treated and regenerated’ line refers to control plants regenerated from callus without and with oryzalin treatment, respectively, and have the same ploidy level as the diploid and triploid progenitors. A ‘synthetic polyploid’ represents plants regenerated from callus treated with oryzalin and having a doubled genome compared with the corresponding progenitor. Thus, synthetic autopolyploids from diploid M. sinensis and M. sacchariflorus are tetraploids. In M. x giganteus, synthetic polyploids include both hexaploid and putative pentaploid lines unless otherwise specified.
Plant materials, explant tissues, and tissue culture media
Seeds (Jelitto Staudensamen GmbH, Schwarmstedt, Germany) or immature inflorescence tissues from eight genotypes of M. sinensis, M. sacchariflorus, and M. x giganteus (Table 1) were sterilized by immersion in 0.5% NaOCl solution for 10 min and rinsed three times with sterilized double distilled (dd) H2O. Immature inflorescences (2–4 cm long) were taken from the terminal nodes of culms following removal of sheath leaves just before flag leaves became visible. Sterilized seeds and immature inflorescence explants (2–3 mm) were then transferred to callus induction medium consisting of MS basal salts and MS vitamins (Murashige & Skoog, 1962) with 13.6 mm of 2,4-dichlorophenoxyacetic acid (2,4-D), 0.44 mm of 6-benzylaminopurine (BAP), 2.88 gL−1 of proline, 30 gL−1 of sucrose, and 750 mg L−1 of MgCl2·6H2O (Petersen, 1997). All media were supplemented with 2 g L−1 PhytagelTM agar (Sigma-Aldrich, St Louis, MO, USA), and adjusted to pH 5.5 prior to autoclaving. Cultures were incubated in darkness at 27 ± 2 °C and subcultured at 2-week intervals for the first month followed by 3-week intervals thereafter.
Table 1. Survival of calli after oryzalin treatment and plant regeneration of Miscanthus accessions and their ploidy levels
No. callus treated/survived (%)
No. plants regenerated
Di- or triploids
Synthetic polyploids (%)
M. x giganteus
Production of synthetic autopolyploids
Three months after culture initiation, calli were treated with 10 μm of oryzalin (Sigma-Aldrich) in liquid callus maintenance medium having the same composition as the callus induction medium aforementioned excluding BAP. The concentration and exposure duration of oryzalin treatment was determined based on the results of Yu et al. (2009). Filter-sterilized oryzalin solutions were added after autoclaving the medium. 1–2 mm pieces of calli from each genotype were transferred into 50 mL of liquid callus maintenance medium in a 250 mL Erlenmeyer flask. The flasks were shaken at 14,600 g at 27 ± 2 °C in the dark for 36 h. After treatment, calli were transferred to solid callus maintenance medium and subcultured at 2-week intervals. Forty days after oryzalin treatment, callus survival rate was recorded (Table 1) and calli were transferred to regeneration medium consisting of MS basal salts and MS vitamins with 1.3 mm of NAA, 22 mm of BAP, 20 gL−1 of sucrose, and 750 mg L−1 of MgCl2·6H2O. Regeneration media was supplemented with 3 gL−1 of PhytagelTM agar, and adjusted to pH 5.5 before autoclaving. The growth conditions were 16 h of cool white fluorescent light (40 micro Einsteins s−1 m−2) at 27 ± 2 °C. Calli were maintained for 70 days in regeneration medium, and regenerated shoots were transferred into MS basal medium (Murashige & Skoog, 1962) and maintained for 40–60 days for root induction. Plantlets with roots were transferred to plastic pots (700 cm3) containing Metro-Mix 500 (Sun Gro Horticulture, Canada Ltd., Vancover, British Columbia) in the greenhouse at UIUC. The temperature of the greenhouse was maintained at 28 ± 2 °C day (14 h)/22 ± 2 °C night (10 h), with supplemental lighting provided from 06:00 to 20:00 hours if light intensities fell below 2670 micro Einsteins s−1 m−2.
Determination of plant ploidy levels
Three to 4 months after transplanting in the greenhouse, regenerated plants were screened for putative polyploids by measuring leaf guard cell length. Leaves from all oryzalin-treated plants and corresponding regenerated progenitors were collected for comparison of guard cell length using microphotography of epidermal impressions as described by Rayburn et al. (2009). The plants possessing average stomata lengths of 20% or larger than progenitors were considered putative synthetic polyploids (Yu et al., 2009). These lines were then subjected to flow cytometric analysis to measure their nuclear DNA (nDNA) content using the modified protocol of Yu et al. (2009). Ploidy levels were confirmed by comparing the peaks in the nDNA histogram between those of the progenitor (external standard) and putative polyploid lines. Briefly, young leaf tissue (1 cm2) from newly emerging shoots from both a treated and control plant was chopped in a petri dish containing 10 mL extraction buffer consisting of 13% (v/v) hexylene glycol, 10 mm Tris-HCl (pH 8.0), and 10 mm MgCl2 with 200 μL of 25% Triton X. The samples were filtered through a 50 μm nylon mesh (Partec GmbH, Gorlitz, Germany) into a labeled test tube and kept on ice throughout. Following filtration, samples were centrifuged for 25 min at 300 × g at 4 °C. The supernatant was then aspirated, and nuclei are resuspended in 300 μL of propidium iodide (PI; Sigma-Aldrich). Then the solution was transferred to a 1.5 mL microcentrifuge tube and incubated for 20 min at 37 °C. After incubation, 300 μL of PI salt was added to each sample. Samples are then briefly vortexed, placed on ice, and stored at 4 °C for at least 1 h. Nuclei were analyzed using flow cytometer Model LSRII (BD Biosciences, San Jose, CA, USA; Flow Cytometry Facility at the University of Illinois-Keck Biotechnology Center). The excitation wavelength was set at 488 nm and a 570 nm emission filter was used. A minimum of 20 000 nuclei per sample were analyzed.
Chromosome counting of root tips
Chromosome numbers were counted for three progenitors and derived synthetic polyploid lines and one putative pentaploid line, Mxg5x-2 (Table 2). Root tips 1–2 cm in length were excised and soaked in 0.05% of 8-hydroxyquinoline for mitotic inhibition. After three hours, the root tips were rinsed in ddH2O for 5 min and stored in 3:1 (v/v) 100% ethanol/acetic acid. The roots were stored at room temperature for 4 days and then stored at 4 °C until use. Fixed root tips were rinsed in ddH2O, hydrolyzed in 5 N HCl for 45 min, and placed in Feulgen's stain for 2 h. Root tips were then rinsed in ddH2O and a drop of 1% acetocarmine was added to the root tip. A cover slip was placed over the tissue and gently tapped with a dissecting needle to disperse the tissue. The slide was then flamed over an alcohol burner, and direct pressure was applied to the slide. The slides were then viewed using an Olympus BX61 microscope (Olympus America Inc., Melville, NY, USA). Photographs of chromosome spreads were taken using an Olympus U-CMAD3 camera and chromosome counts conducted on the clearest preparations.
Table 2. 2C values of nuclear DNA content and chromosome numbers of regenerated and synthetic polyploid lines used for greenhouse investigation
Determination of 2C DNA contents in synthetic polyploid lines
The 2C values of nDNA in five genome doubled lines for each of M. sinensis ‘Grosse fontaine,’ M. sacchariflorus ‘Golf course,’ and M. x giganteus ‘Illinois’ and their diploid or triploid progenitors (Table 2) were measured by flow cytometry with sorghum as an internal standard using the modified protocol of Rayburn et al. (2009). The protocol is the same as described above with the replacement of the progenitor control with sorghum as the internal standard being cochopped with the Miscanthus lines. The 2C value of nDNA content of the sorghum line was calibrated at 1.74 pg using the maize genotype W-22 as a calibration standard which is reported to have 5.35 pg DNA/2C (McMurphy & Rayburn, 1991). Mean fluorescence of the Miscanthus G1 peak is divided by the fluorescence reading of the internal standard, multiplied by 1.74 pg/2C, and expressed in pg/2C nucleus. Since sample G2/G1 peak ratio is typically slightly less than the expected value of 2.0 (Wood & Todd, 1979; Watson, 1991), samples with sorghum G2/G1 peak ratios outside the range 1.94–2.03 were excluded. Sample target peaks in DNA histograms that were not symmetrical or where coefficients of variation exceeded 5% were also excluded. For each accession, 3–5 samples were examined and samples that did not meet these criteria were discarded until three acceptable replications (one leaf per replication) were recorded.
Phenotypic evaluations of synthetic polyploid lines
Phenotypic evaluations of synthetic polyploid lines were conducted in the greenhouse under conditions described above. The investigation was performed on fifteen synthetic polyploids and their three progenitors: M. sinensis ‘Grosse fontaine,’ M. sacchariflorus ‘Golf course,’ and M. x giganteus ‘Illinois’ (Table 3). Synthetic polyploid M. x giganteus ‘Illinois’ plants produced from Yu et al. (2009) were used in the evaluation with the addition of Mxg6x-20, a newly added line from the aforementioned tissue culture procedure (Tables 1 and 3). Each line was divided into single stems with rhizome portions (M. x giganteus and M. sacchariflorus) of equivalent size and weight and planted in 10 L pots on 7 October 2010. A randomized complete block design with four replications was used with each block containing 21 lines from the three species (one regenerated control line, one oryzalin-treated and regenerated control line, and five synthetic polyploid lines for each species). Phenotypic data were recorded every week from December 2010 to June 2011 for flowering time, plant height, stem diameter, and the number of tillers. Plant height was measured from ground to highest point and the stem diameter was taken at one third of mature plant height from the base.
Table 3. Phenotypic differences in number of tillers, plant height, flowering time, and stem diameter between the means of regenerated lines and five synthetic polyploid lines of the three Miscanthus species
t-tests were performed between calculated means from two progenitor lines and those from the synthetic polyploid lines of each species.
*, **, and *** indicate significant differences based on t-test at < 0.05, <0.01, and < 0.001, respectively.
Field evaluation for phenotypic variation among M. x giganteus lines was conducted on the Energy Bioscience Institute farm at UIUC. Five regenerated control triploid lines, five oryzalin-treated and regenerated control triploid lines, and eight synthetic hexaploid M. x giganteus lines (Yu et al., 2009) were utilized for this investigation. Transplants were prepared as described above and planted on 16th June 2010 in a randomized complete block design with three replications with 1.5 m spacing. A block contained 18 lines with three clones per each line (18 × 3 = 54 plants). Data on tiller number, plant height, and stem diameter were collected on 17th October 2011. Phenotypic data were collected as described above.
To estimate overall phenotypic effect of neopolyploidization, t-tests were performed on the phenotypic data between calculated means from two progenitor lines and those from five synthetic polyploid lines of each species (Table 3). Analysis of variation (anova, General Linear Model) was performed on the 2C value of nDNA content, and the phenotypic data among polyploids and progenitors with significant differences between mean values were determined by the Student's t-test. For field data for synthetic polyploids of M. x giganteus, the means of each line were grouped into three categories including regenerated triploid controls, oryzalin-treated regenerated triploid controls, and synthetic hexaploid lines, and box plots generated from the data. All statistical analyses were performed and box plots were generated using jmp software (sas Institute, Cary, NC, USA).
Calli induction and survival and shoot regeneration
All explants from the eight genotypes of the three species produced callion media with the same composition. Calli from each explant source were observed to proliferate after oryzalin treatments for 2 days. The survival rates of oryzalin-treated calli after 40 days of culture on solid callus maintenance medium differed depending on explant source and species. The survival rates of seed-derived calli were much higher than those of immature inflorescence-derived calli in M. sinensis (Table 1). Among immature inflorescence-derived calli, M. x giganteus ‘Illinois’ showed the highest and M. sinensis ‘Undine’ the lowest survival rates.
The number of regenerated shoots was counted at 70 days after culture on the regeneration medium. Approximately half of the calli transferred to the regeneration medium across all genotypes displayed necrosis and did not produce shoots. Shoot producing calli were usually compact and white embryogenic-like calli (Kim et al., 2010). Calli from M. sinensis ‘Undine,’ M. sacchariflorus ‘Blue stem,’ and ‘Robustus’ did not produce any regenerated shoots (Table 1).
Determination of ploidy level and nDNA contents of synthetic polyploids
The oryzalin-treated regenerated plants with average stomata size 20% or larger than their progenitor plants were selected as putative synthetic polyploid plants (Fig. 1). nDNA content of all selected plants was approximately double that of their progenitor counterparts. For oryzalin-treated regenerated plants from immature inflorescence-derived calli, polyploidization rates were 44.9% for M. sinensis ‘Grosse fontaine,’ 39.0% for M. sacchariflorus ‘Golf course,’ and 30.4% for M. x giganteus ‘Illinois’ (Table 1). One synthetic polyploid line per species was selected randomly for chromosome counting. Nuclei of synthetic polyploids from M. sinensis, M. sacchariflorus, and M. x giganteus were observed to contain 76, 76, and 114 chromosomes, respectively, confirming successful genome doubling in all three Miscanthus species as counts matched the expected number of chromosomes (Fig. 2 and Table 2).
Five independently produced, synthetic polyploid lines from each species were randomly selected, and 2C nDNA content of each was determined by flow cytometry as described above. The DNA histograms of regenerated control and synthetic polyploid lines of each Miscanthus species are shown in Fig. 3. Coefficient of variations (standard deviation/mean × 100) in nDNA content of all lines analyzed did not exceed 5% (data not shown). All of the synthetic polyploid lines from M. sinensis and M. sacchariflorus have the approximately doubled mean nDNA content of progenitor lines, 8.96 ± 0.09 pg and 10.71 ± 0.24 pg, respectively; however, a synthetic polyploid line from M. x giganteus had lower 2C nDNA content and chromosome numbers than expected (Table 2). This line (Mxg5x-02, Table 2) was selected for chromosome counting and was observed to harbor 95 chromosomes (Table 2, Fig. 2 g) and therefore is considered a putative pentaploid line (5x, 5 × 19 = 95 chromosomes).
The effect of synthetic polyploidization
The effect of polyploidization on phenotypic performance in synthetic polyploid lines differed among three Miscanthus species grown in the greenhouse. Simple t-tests between the means of progenitor and polyploid lines showed that the number of tillers (P <0.001) and plant height (P <0.01) were significantly different in M. sacchariflorus. The number of tillers (P <0.05) and flowering time (P <0.05) in M. sinensis, and stem diameter (P <0.05) in M. x giganteus were also significantly different between synthetic polyploid and progenitor lines (Table 3). Overall, the number of tillers was reduced in all synthetic polyploid lines. However, in M. sinensis and M. x giganteus, plant tiller numbers in the synthetic polyploid lines, GF4x-11 and Mxg6x-20, are comparable to that in progenitor lines, respectively (Table 4). Plant heights of synthetic polyploid lines of M. sinensis and M. x giganteus tended to be smaller than that of regenerated progenitors (Table 3); nevertheless, some synthetic polyploid lines displayed similar plant heights to regenerated progenitors (Table 4, Fig. 4b and c). Polyploidization did not affect flowering time in M. sacchariflorus, but delayed flowering in M. sinensis and M. x giganteus. Flowering in three synthetic polyploid lines of M. x giganteus was significantly delayed, from 2 to 17 weeks compared with their progenitors (Table 4). Although there were no clearly superior synthetic polyploid lines compared with regenerated lines in terms of biomass characteristics, the phenotypes of GF4x-11 and Mxg6x-20 are comparable to their diploid and triploid progenitors, M. sinensis and M. x giganteus, respectively (Table 4).
Table 4. Phenotypic variations in number of tillers, plant height, flowering time, and stem diameter of regenerated and synthetic polyploid lines in three Miscanthus species
Plant height (cm)
Stem diameter (mm)
Means levels within a column with a different letters are significantly different at P < 0.001.
Field evaluation of biomass accumulation traits in M. x giganteus confirms that genome doubling can generate phenotypic variation. As these lines are the first generation after genome doubling, oryzalin-treated and regenerated control lines (triploid) were also used to investigate the effect of the antimitotic agent on the phenotype under field conditions. Synthetic polyploid (hexaploid) lines of M. x giganteus showed greater phenotypic variation for all three traits investigated than the regenerated and oryzalin-treated and regenerated control lines (Fig. 5). The flowering time could not be investigated as most synthetic polyploid lines of M. x giganteus did not flower before senescence in the field due to the onset of chilling fall temperatures (Fig. 4c). The means of plant height and tiller number were significantly reduced in synthetic polyploid M. x giganteus, but the differences in plant height between synthetic polyploids and their progenitors were greater as most synthetic polyploid lines did not flower in the field, possibly due to slower vegetative growth. The mean stem diameter of synthetic polyploid lines increased slightly compared with their corresponding progenitor lines. These results were acquired from single year observation of a 2-year-old field plot after planting and thus may not represent the phenotypic differences in fully established plant stands. Future research will be conducted over multiple years with mature stands. The interpretation of these results is also somewhat limited as we used only 10 progenitors and 8 synthetic polyploid lines; however, the pattern of greater phenotypic variation among synthetic polyploids was consistent with greenhouse observation.
We produced synthetic polyploid plants of three Miscanthus species, which represents the first report of artificial production of tetraploid plants from diploid M. sacchariflorus. Tissue culture produced calli from two different explant tissues and from various genotypes of three Miscanthus species on the same medium composition. While calli survived and proliferated after oryzalin treatment, some genotypes did not regenerate shoots (Table 1). This indicates variation among genotypes in response to shoot regeneration media, suggesting different accessions require different media composition for shoot differentiation as has been observed in previous studies (Petersen et al., 2002; Wang et al., 2011). The high survival rates of calli after 10 μm of oryzalin treatment seen by Yu et al. (2009) was also observed in this experiment. Polyploidization rates ranged between 30.4% and 73.5% which are comparable to the highest polyploidization rates reported in other studies (Petersen et al., 2003; Yu et al., 2009; Głowacka et al., 2010).
Guard cell size is commonly used for estimating ploidy level in closely related species as it is often significantly larger in polyploids than in the diploid progenitors (Tate et al., 2005). The usefulness of guard cell length in distinguishing different ploidy levels has been demonstrated in other grass species, such as barley (Borrino & Powell, 1988), rye grass (Speckmann et al., 1965), Paspalum glaucescens (Gramineae; Paniceae) (Pozzobon & Valls, 2000), wheat (Khazaei et al., 2010), and even in fossilized plants (Masterson, 1994). In our study, the preliminary screening of oryzalin-treated plants by comparison of stomata size with their corresponding progenitors was successful as all selected plants were confirmed as synthetic polyploids by flow cytometry and chromosome counts (Figs 1–3).
A putative pentaploid plant was obtained from oryzalin treatment to calli of triploid M. x giganteus ‘Illinois’. Odd ploidy level plants (triploids) were also observed in genome doubling of diploid M. sinensis by colchicine and oryzalin treatment (Petersen et al., 2003). Although the mechanism is not understood, genome doubling by antimitotic agents can result in aneuploidy and odd numbers of ploidy levels. This may be explained by residual antimitotic agent activity in cells in later cycles of mitosis after genome doubling. The concentration of residual antimitotic agent may vary among individual genome doubled cells and inhibit microtubule polymerization to various degrees in dividing cells, resulting in chromosomal aberration due to vagrant and/or laggard chromosomes in dividing cells (Sharma, 1990). The chromosomal aberration can result in a wide range in chromosome numbers in aneuploid cells and euploid cells containing either even or, by chance, odd ploidy levels. In successive generations, there may be tendency for the euploid genome cells to survive in a callus cell colony as aneuploidy may be deleterious to cell fitness due to changes in the copy number of structural genes (Torres et al., 2008) while euploid cells will maintain balanced genomes. As a result, a callus cell colony could be composed primarily of euploid cells, with either an even or odd ploidy level, although portion of odd ploidy calli should be very small. In previous work, only 9 triploids were obtained of 1,377 regenerated plants (0.7%) following colchicine and oryzalin treatments of diploid M. sinensis (Petersen et al., 2003).
Synthetic polyploid lines in our study differed from their corresponding progenitors in phenotypic traits, such as increased cell size (Fig. 1) and stem diameter (Table 3). Similar results were observed in colchicine-induced polyploid Miscanthus species, which showed increased cell size and stem diameter (Głowacka et al., 2010). Earlier studies in other species revealed common phenotypic changes associated with neopolyploidy: coarser, thicker, and larger leaves and larger reproductive organs (flowers and seeds) (Ramsey & Schemske, 2002 and references therein). The ‘gigas’ characteristics of neopolyploids such as increased cell size, larger leaves, enlarged reproductive organs, and robust stems are possibly due to an increased DNA content (Randolph, 1941).
Delayed flowering and reduced tiller number and plant height were observed in synthetic polyploids of all three Miscanthus species compared with their progenitors (Table 3), as was previously observed in M. sinensis (Głowacka et al., 2010). The delayed flowering, and reduced tillering and plant height are presumably due to delayed development in these polyploid lines. A positive correlation has been observed between genome size and the cell volume which is negatively correlated with the cell division rate (Bennett & Leitch, 2005). Polyploid cells are larger to accommodate larger genomes requiring more time to replicate with reduced cell cycle rates, thus delaying development and flowering (Noggle, 1946; Stebbins, 1971; TeBeest et al., 2012). However, the polyploid plants in our study have not yet experienced any cycles of sexual reproduction after genome doubling. In addition, our results were acquired from both a greenhouse trial and a single year observation of a 2-year-old field plot. Therefore, it is possible that the phenotypic differences may not be representative of fully established plant stands.
Allopolyploidy generally induces a greater variation in gene expression than autopolyploidy due to the combined effects of dosage changes and interactions between two genetically distinct genomes (Chen, 2007). The dramatic variation in flowering time evident in synthetic autoallopolyploids of M. x giganteus, which was not observed in the two synthetic autopolyploids of M. sinensis and M. sacchariflorus (Table 4) may be due to differential transcriptional or post-transcriptional gene regulation following allopolyploidization. Variation in gene transcriptional regulation may be due to either genetic or epigenetic changes. Genomic rearrangement in chromosomal regions where important flowering genes are located may be responsible for the flowering time variation in hexaploid M x giganteus lines. Nonreciprocal transposition of the FLOWERING LOCUS C (FLC) gene resulted in flowering time variation among resynthesized Brassica allopolyploids (Schranz & Osborn, 2000; Pires et al., 2004). Genomic rearrangements may also be associated with the activation of transposable elements. Polyploidization has been observed to activate transposable elements (Kashkush et al., 2002) generating indels within and/or near coding sequences as well as resulting in the breakage or rearrangement of chromosomes. In M. x giganteus, repetitive sequences related to either transposable elements or centromeric repeats comprise 95% of genome (Swaminathan et al., 2010). Epigenetic changes such as variation in DNA methylation and histone modification can also account for phenotypic variation in synthetic polyploids (Otto, 2007). Differential post-transcriptional regulation in allopolyploids by small RNAs (Ha et al., 2009; Ng et al., 2012) could also be associated with phenotypic variation among synthetic polyploid M. x giganteus lines.
The genus Miscanthus experienced an ancient genome duplication event (paleopolyploidization) and chromosome fusion after its divergence from closely related genera and before species diversification (Kim et al., 2012; Ma et al., 2012; Swaminathan et al., 2012). The variation in flowering time presumably due to polyploidy, generation of interploid hybrids (Nishiwaki et al., 2011), and/or the paleopolyploidization event in Miscanthus may help explain the wide geographical and latitudinal distribution of this genus. The differences in flowering time could create reproductive isolation of new polyploids from their progenitors (Hegarty & Hiscock, 2008) and allow neopolyploids to enter new ecological niches (Otto, 2007). The natural allopolyploid Arabidopsis suecica displays later flowering than the artificially generated allopolyploid (Wang et al., 2006), suggesting that phenotypic change related to allopolyploid events is subject to natural selection (Chen, 2007). Variation in the geographic distribution of diploid and tetraploid cytotypes of Anthoxanthum alpinum also appears to relate to differences in flowering phenology (Felber, 1988; Tate et al., 2005).
Synthetic polyploids provide many opportunities for crop improvement in a conventional breeding program. Interspecific hybrids between distant taxa (different species or genera) are often sterile because of the failure of chromosomes to properly pair during meiosis. Genome doubling is a technique to restore the fertility of these interspecific hybrids by providing homologous chromosomal duplicates (Ramsey & Schemske, 2002). Fertile allopolyploids have been produced from genome doubling of sterile hybrids of related plant species (Thomas, 1993; Nimura et al., 2006). Sterility associated with triploidy can also be overcome by genome doubling. Restoration of pollen viability and seed set was observed from oryzalin-induced hexaploids derived from sterile triploid interspecific rose hybrids (Kermani et al., 2003). Loss of self-incompatibility after polyploidization events has been observed in other species (Miller & Venable, 2000) and could overcome self-incompatibility in Miscanthus species, opening opportunities to uncover genetic variability in this obligate outcrossing genus (Heaton et al., 2010). These opportunities can be applied to Miscanthus breeding programs, where natural (Sobral et al., 2010) and artificial intergeneric hybrids (Park et al., 2011) exist and where self-incompatibility presents barriers to gene flow and the generation of inbred lines.
The establishment of triploid M. x giganteus production fields requires vegetative propagation via rhizome divisions or tissue culture, which would be expensive relative to seed propagation if the latter option were made available (Lewandowski et al., 2000). For rhizome harvest, cleaning, separation, and replanting, it has been estimated that one hectare of mature M. x giganteus (3 or more years old) will provide sufficient rhizomes for the planting of approximately only ten hectares of new production (T. Voigt, personal communication). This is a significant constraint in the development of a bioenergy cropping system that will need to have the capacity to rapidly upscale to fulfill large production requirements. In contrast, seed-based propagation of Miscanthus should be much cheaper, scalable to meet production needs, and utilize sexual hybridization and selection to generate diverse and improved germplasm for commercial production. Individual culm inflorescences of Miscanthus accessions are known to produce hundreds or even thousands of seeds.
The impetus for the seed propagation of Miscanthus, however, impacts on its potential invasiveness. The putative parental species of M. x giganteus, M. sacchariflorus, and M. sinensis are not native to the United States and the former is designated as invasive in several states with the latter considered as putatively invasive (Quinn et al., 2010). The production of allotriploid seeds from crosses between diploid and tetraploid M. sinensis and M. sacchariflorus can circumvent invasive issues by generating sterile non-invasive plants. Our recently developed polyploids derived from M. sinensis and M. sacchariflorus could be used to broaden the genetic base of M. x giganteus, thereby allowing for further exploitation of interspecific heterosis. If generation of triploid and pentaploid seeds (from crosses between hexaploid M. x giganteus and tetraploid accessions of M. sinensis and M. sacchariflorus) is feasible on a scale amenable to commercial seed production, this could result in seed-propagated production of cultivars that are potentially sterile.
Genome doubling could be also used for circumventing invasive issues by delaying flowering time. Most of our hexaploids did not complete flowering under the environmental conditions found in Central Illinois although viable pollen was produced from these hexaploid lines in the greenhouse (data not shown). Late flowering in Miscanthus is associated with higher biomass yields (Clifton-Brown et al., 2001). Seeds of late flowering and fertile polyploids could be harvested in southern latitudes and planted in temperate areas where growing seasons are not long enough for the initiation or completion of the flowering process. Such a production system would reduce risks of Miscanthus invasion.
The synthetic polyploid lines from this study can also be used to study neopolyploidization; an advantage over natural polyploids from which their corresponding progenitors are often unknown. Investigating the association between transcriptional and post-transcriptional regulation on phenotypic variation could help to reveal the mechanisms underlying neopolyploidization in Miscanthus and other species. Studies are currently underway to detect changes in gene expression and investigate variation in small RNA mediated gene silencing between synthetic polyploids and their corresponding progenitors.
This work was supported by the Consortium for Plant Biotechnology Research (grant number GO12026), the U.S. Department of Energy, and the scholar exchange program of Gangneung-Wonju National University (PN 2009-0017). Research assistantship support for the first author came from the Energy Biosciences Institute at the University of Illinois.