Autopolyploids in fodder grass breeding: induction and field performance

Doubling of chromosome set directly affects plant performance through increase of organ size, higher feeding value and increased resistance to adverse environmental factors. Therefore efficient methods of polyploid induction are needed in order to develop new varieties of naturally diploid fodder grass species. The efficiency of antimitotic agents as colchicine, amiprophos-methyl, trifluralin and oryzalin was compared in a series of tetraploid induction experiments in Lolium multiflorum , L. perenne and Festuca pratensis , while newly developed tetraploid plants were compared to standard tetraploid varieties in the field trials. Colchicine treatment proved to be the most efficient method for in vitro cultured embryos in comparison with the other agents. Induced tetraploids of F. pratensis produced higher dry matter and seed yield and could be used for the development of new varieties. Induced tetraploid plants of Lolium spp. were equal to the standard varieties in field trials, therefore they could be used as parental genotypes in crosses. Induced tetraploids of F. pratensis produced higher dry matter and seed yield and could be used for development of new variety.


Introduction
Polyploidy is defined as presence of three or more complete chromosome sets per cell nucleus. It is a common phenomenon in plant kingdom and is thought to be major driving force behind plant evolution (Soltis et al., 2009). Ancient events of polyploidization led to emergence of many cultural plant species. Polyploidization often seems to be associated with increased plant vigour, higher adaptability to environmental conditions, increased cell size and, consequently, increased organs although the latter one is not always the case as the number of cell division can be reduced. Other benefits of being polyploid include better salinity, drought or extreme temperature stress tolerance (reviewed in Sattler et al., 2016). Gene duplication in autopolyploids can cause transcriptomic changes leading to higher adaptability compared to diploids (del Pozo & Ramirez-Parra, 2015), therefore if genotype undergoing whole genome duplication possesses allelic variants of genes conferring increased resistance to freezing temperatures  or genes involved in regulation of plant architecture traits , highly beneficial effect in stress response and growth habit of induced tetraploid can be expected. All these characteristics are highly desirable in cultivated crop varieties, with the exception of crops cultivated for seeds as polyploids tend to have lower seed yield compared to their diploid counterparts. Therefore breeders have been searching for most efficient methods of artificial autopolyploid induction in economically important yet naturally diploid plant species. Polyploids can be induced by doubling of the chromosome number in somatic tissues (mitotic polyploidization) or by generating unreduced gametes (meiotic polyploidization) (Younis et al., 2014). First attempts at inducing agricultural polyploids were reported nearly eight decades ago (Blakeslee & Avery, 1937). Colchicine was used in pioneering experiments, and it remained as a very popular antimitotic agent up to this day. Various tech-niques were developed to induce mitotic chromosome doubling in different plant species, including treatment of shoots, leaves, buds, node sections, embryos and cell suspension cultures (Dhooghe et al., 2011). Colchicine was proven to be very useful in most of these methods as well as in many plant species, however many researchers postulated the need for new antimitotic drugs as colchicine application is both hazardous due to its' toxicity, and costly, because high concentrations of colchicine are needed. Attention was turned towards antimicrotubular herbicides due to their ability to disrupt cell mitotic division. Herbicides oryzalin, trifluralin and amiprophosmethyl (APM) were proven to be just as effective or even more so -and in much lower concentrations -compared to colchicine (Yemets & Blume, 2008). Despite tremendous progress has been made since first attempts, methods of autopolyploidy induction in vitro are still being modified as the whole procedure has to be adapted for each plants species individually, and even after elaboration of efficient technique high variability of polyploid yield remains due to differences in genotype response.
Ryegrass is one of most important grass for forage and lawn in the world. Two subspecies of Lolium multiflorum Lam.: italicum (Italian ryegrass) and multiflorum (Westerwolths ryegrass) are used for fodder and erosion control. Perennial ryegrass (Lolium perenne L.) is sawn for both fodder and lawn purposes (Humphreys et al., 2010). Meadow fescue (Festuca pratensis Huds.) is another important forage species in Northern countries, widely used in the mixtures (Fjellheim et al., 2007). Ryegrass species and meadow fescue are often used for development of hybrid festulolium. This hybrid is characterized by better stress tolerance (cold and drought) than ryegrass and improved fodder value compared to the fescue (Humphreys et al., 2004). Both ryegrass species and meadow fescue exist as diploid (2n = 2x = 14) in nature. Therefore mitotic polyploidization is used not only for development of higher yielding fodder cultivars but also for restoration of fertility in interspecific hybrids.
The aim of this study was to compare effectiveness of different in vitro mitotic polyploid induction methods in Lolium perenne, Lolium multiflorum and Festuca pratensis, and evaluate field performance of newly created tetraploids.

Material and methods
Research was carried out at the Institute of Agriculture, Lithuanian Research centre for Agriculture and Forestry.

Plant material
The germplasm consisted of 16 ryegrass and 7 fescue diploid genotypes ( Table 1). All of the breeding lines used in the experiment were developed locally.
Embryos of Westerwolths, Italian and perennial ryegrass and meadow fescue were excised from mature seeds following surface sterilization with 50% sulfuric acid for 15 min, 70% ethanol for 2 min. and 50% commercial bleach containing up to 5 % of sodium hypochlorite for 7 min. Thereafter embryos were cultivated in 10 mL test tubes on Murashige-Skoog basal medium (Sigma, 2/3 of salts) containing 0.2 mg of indoleacetic acid (IAA), 0.25 mg kinetin, 1 mg thiamin, 0.1 g pyridoxine, 30 g sucrose and 7 g agar per liter of medium. Germinated embryos (8-14 mm long) were treated with antimitotic agents for 4 hours, then rinsed with sterile distilled water and placed on fresh medium. The experiment was carried out in three replicates, 734 germinated embryos of Westerwolths ryegrass, 420 embryos of Italian ryegrass, 525 embryos of perennial ryegrass and 730 embryos of meadow fescue were treated with each antimitotic agent in total, 100-105 per each genotype. Embryos were kept in test tubes (24˚C, 16 h light) for 45-60 days, then transferred to peat substrate. Chromosomes were counted in root tip meristem cells, aceto-carmine solution (3%) was used as staining agent. Five or more root tips were sampled from each plant, chromosomes were counted in at least 5 clear metaphase plates in each root tip. The plants with less than five diploid metaphase plates in total were classified as tetraploid.
The soil of the experimental fields was Endocalcari -Epihypogleyic Cambisols (CMg-p-w-can), characterised by a homogeneous texture, pH KCl 7.3-7.0, humus content 1.9-2.2%, available P 2 O 5 206-270 mg/kg and K 2 O 101-154 mg/kg. In the autumn of each year of use phosphorus and potassium fertilisers (P 60 K 90 ) were applied. Nitrogenous fertilisers (N 150 ) were applied each year of herbage use in several applications: in spring N 60 , and N 45 after cuts.

Statistical analyses
Statistical reliability of research data was assessed with the lowest significant difference (LSD05, LSD01).
Computer software STATISTICA 7 (StatSoft Inc., USA) was used for the statistical analysis.

Post-treatment survival and tetraploid induction efficiency of different mitotic inhibitors
In this experiment, concentration of oryzalin used proved to be the most toxic, the lethal effect of both colchicine and APM treatment was similar while trifluralin treatment was the least toxic ( Table 2). The response of different species was also diverse with Westerwolths ryegrass being the most recalcitrant for polyploidy induction. The proportions of induced tetraploid plants after treatment with different mitotic inhibitors in mature embryo culture was analysed in a separate experiment. Again, oryzalin was the most efficient in tetraploid induction (Table 3). However extremely low survival of plants regardless of species indicates that application of this mitotic inhibitor needs adjustments. Trifluralin application induced lowest mean proportion of tetraploid plants, only treatment of perennial ryegrass embryos yielded high induced tetraploid proportion (81.3%) when this antimicrotubular chemical was used (Table 3). The mean efficiency of colchicine was the highest (58.2%), it proved to be the most suitable agent for polyploidization of both ryegrasses and meadow fescue if we take into account the surviving plants. The variation in the tetraploid induction efficiency (cv%) between the treated populations in each species ranged from 17.6 to 91.3 (Table 4). Again the mean variation of colchicine efficiency (21.6-53.3%) was lower compared to other antimitotic agents.

Field experiment
Seeds of Lolium spp and F. pratensis tetraploid plants were collected and sown in the field trials. The average results of each species and the results of the best There is no consensus on the best method for tetraploid induction in fodder grasses. The popular method for ryegrass polyploidization is colchicine treatment of germinated seeds or young plants, with surviving success rate up to 40% of induced tetraploids plants, depending on colchicine concentration and treatment duration (Joshi & Verma, 2004;Nair, 2004;Pereira et al., 2014). In some studies treatment of dry seeds rather than germinated proved to be more efficient, where tetraploid induction rate in surviving perennial ryegrass plants was 64% (Dapkienė et al., 1999), yet the survival rate itself was low making this method applicable only when initial material is ample. Successful polyploidization results were obtained when germinated seeds of Dactylis polygama were treated with colchicine, while treatment of dry seeds was completely inefficient (Dabkevičienė et al., 2013). The data on application of in vitro embryo culture for plant polyploidy induction is scarce. Embryo culture was tested in cotton, however the authors found seed treatment to be more reliable due to rather complicated procedure of embryo in vitro cultivation in cotton (Omran & Mohammad, 2008). Treatment of germinated embryos with 7.5 mM colchicine in combination with low cultivation temperature and increased sucrose concentration in germination medium was the most efficient tetraploid induction method in perennial ryegrass (Pašakinskienė, 2000). tetraploid genotype were compared with the standard tetraploid variety. The percent of survival after winters of Italian and perennial ryegrass were significantly (p< 0.05) lower for all tetraploids compared with the standard varieties (Table 5). Winter damage of meadow fescue was very low regardless ploidy level of the plant. Significantly highest seed yield per plant was assessed for L. perenne (6.45 g, p< 0.05) and meadow fescue (28.58 g, p< 0.01) compared with the standard (5.53 and 17.27 g respectively) ( Table 6).
Average data of dry matter yield and other agrobiological traits of meadow fescue tetraploids showed the most statistically significant differences compared with standard, especially for plant height (Table 5 and  Table 6). The results in other species were similar to standards varieties.

Discussion
The success of autopolyploid induction in any plant species depends not only on the choice of mitotic inhibitor and application method, but also on responsiveness of particular genotype. High coefficients of variation were obtained in each grass species in this experiment, affirming that any method might yield rather different results when applied to distinct populations/cultivars.  Table 4. Tetraploid induction variation in Lolium multiflorum subsp. multiflorum (7 cultivars), L. multiflorum subsp. italicum (4 cultivars), L. perenne (5 breeding lines) and Festuca pratensis (2 cultivars, 5 breeding lines) after treatment with different mitotic inhibitors in culture of mature embryos.   Since 1990s when the attention was turned towards alternatives to colchicine, dinitroanilines (oryzalin and trifluralin) and phosphoroamidates (APM) have been tested in various plant species. Some researchers claim trifluralin and oryzalin are less lethal to treated plants than colchicine (Hansen & Andersen, 1996;Cheng et al., 2012). Oryzalin was the most effective antimitotic agent in cork oak double haploid production (Pintos & Manzanera, 2007) and fertility restoration of interspecific currant hybrids (Stanys et al., 2004), trifluralin application was the most efficient in Rosa chinensis minima polyploidization (Zlesak et al., 2005), APM was the best chromosome doubling agent for Allium cepa (Grzebelus & Adamus, 2004). Trifluralin was the least toxic agent for Lolium spp and Festuca pratensis alike in this experiment, yet, at our experiment conditions, it was also the least effective, while oryzalin killed almost all plants. Only APM efficiency was similar to colchicine, making it second-best choice in Lolium and Festuca polyploidization experiments.
Good overwintering is important for grasses, especially for ryegrass, in the Nemoral environmental zone which covers Southern Scandinavia and Baltic states. There is no consensus data in the literature for ploidy level influence on plant over-wintering. Humphreys (1991) indicated that tetraploids often are more tolerant to abiotic and biotic stresses in comparison to diploids. Whereas Sugiyama (2006) proposed that tetraploid genotypes of Italian and perennial ryegrass tended to show lower survival under freezing stress conditions than diploid as related to large shoot mass and high tissue water content of tetraploids. Lolium tetraploids suffered higher winter damage than diploids in this experiment, whereas winter survival rate of meadow fescues was close to 100%, which is the reason this species used for crossing with ryegrass to develop festulolium (Fjellheim et al., 2007).
The seed yield of grasses is a complex trait (Elgersma, 1990) and often negatively correlates with the dry matter yield (Bugge, 1987). It is generally stated, that herbage yield of tetraploid plants is higher compared to diploids (Burns et al., 2013;Kemesyte et al., 2017). Standards varieties of this study were tetraploid, therefore we cannot confirm or disprove these propositions. However all tetraploids populations in this study produced no significant difference of dry matter yield compared with registered varieties, and tetraploid meadow fescue line produced higher dry matter yield compared to the standard.
To conclude, considering all factors that define the efficiency of polyploid induction method (plant survival, tetraploid induction rate and variation between genotypes), colchicine in the present conditions, is proved to be the most effective antimitotic agent for Lolium multiflorum, L. perenne and Festuca pratensis. New tetraploid lines, developed during the experiment, were equal or superior to the registered cultivars.