Introduction

Perennial ryegrass (Lolium perenne L.) is one of the most widely used temperate forage grasses because it has good forage quality, palatability, and aftermath tillering (Thomas et al. 2003). However, its poor tolerance to abiotic stress (especially low winter hardiness, including freezing tolerance and snow mold resistance) limits its use under harsh environmental conditions (Abe 1986). For example, in Hokkaido, Japan, where the snow cover period exceeds 4 months and the lowest temperature goes down below −20 °C in winter, perennial ryegrass may suffer severe damage during winter; therefore, the use of meadow fescue (Festuca pratensis Huds.) instead of perennial ryegrass is often recommended.

Lolium and Festuca species can be hybridized to generate Festulolium, and their homoeologous chromosomes pair and recombine (Jauhar 1975; Thomas et al. 1994). Introgression breeding, that is, the transfer of segments of Festuca chromosome(s) into Lolium by backcrossing, could improve winter hardiness of Lolium while preserving its good forage characteristics (Thomas et al. 2003). Several studies have reported physical or genetic mapping of Festuca-derived genetic regions introgressed in Lolium involved in winter hardiness, especially freezing tolerance (Grønnerød et al. 2004; Kosmala et al. 2006, 2007; Humphreys et al. 2007). These studies used selected plants with high winter hardiness or freezing tolerance and their progeny; therefore, they may not necessarily have analyzed the effects of the introduction of Festuca chromosomal segments into the Lolium genome comprehensively. Quantitative trait loci (QTLs) for these traits were searched for over the whole genomes using linkage mapping populations of L. perenne (Yamada et al. 2004) and F. pratensis (Alm et al. 2011). However, such analysis in Festulolium is difficult because F1 diploid hybrids between Festuca and Lolium are generally sterile. Using an association mapping approach, Bartos et al. (2011) analyzed the relationship between polymorphisms in DArT markers and freezing tolerance in the high- and low-freezing-tolerance genotypes of Festulolium, and identified several markers significantly associated with freezing tolerance. However, it remained unclear whether the detected associated polymorphisms were between Lolium and Festuca or within each species.

Extensive comparative genomics studies have revealed the genomic relationship (genomic structure and colinearity) between forage grasses (Lolium and Festuca) and model crop grasses (e.g., rice, barley, and Brachypodium) (King et al. 2007; Pfeifer et al. 2013; Byrne et al. 2015). Therefore, genomic information available for model grasses can be used for forage grasses, for which genomic information is insufficient. For genetic analysis of Festulolium, intron markers showing polymorphisms between Lolium and Festuca orthologs were developed from sequences of single copy rice gene (Tamura et al. 2009, 2012). Genomic positions of these markers could be estimated using comparative genomic information; this approach has produced a set of markers randomly distributed across the genome.

The aim of this study was to identify the genomic regions involved in the effects of F. pratensis chromosome introgression into L. perenne on winter hardiness-related traits. Based on the hypothesis that survey on the association between presence or absence of Festuca allele and winter hardiness over the genome of the Festuca/Lolium introgression populations could identify the causal regions; first, we comprehensively screened diploid introgression populations with the L. perenne background and introgressed F. pratensis chromosome segments using DNA markers developed from rice genomic information. Next, we conducted linkage analysis using a backcross population with F. pratensis introgression in this region. This study revealed that the introduction of Festuca genomic regions either negatively or positively affects winter hardiness traits in the Lolium background.

Materials and methods

Plant materials

Three populations (IPA12, IPA12-24, and IPA18-2) of diploid L. perenne with introgressed F. pratensis chromosome segments were generated as described in Fig. 1. The triploid F1 hybrid plants A12 and A18 with high pollen fertility were selected from progeny of a cross between a diploid F. pratensis ‘Makibasakae’ plant as a maternal parent and a tetraploid L. perenne ‘Pokoro’ plant as a paternal parent. BC1 progeny of A12 and A18 were generated by crossing with diploid L. perenne ‘Yatsugatake D12’ plants as maternal parents. Nuclear DNA content was estimated by using flow cytometry (PA, Partec, Munster, Germany); DNA was stained with DAPI (CyStain UV Precise P High Resolution DNA Staining Kit, Partec). Among the BC1 plants derived from A12, those with a nuclear DNA content corresponding to a diploid were named the ‘IPA12’ population (n = 47). All A18 BC1 progeny (n = 5) had a nuclear DNA content corresponding to a triploid. The BC1 plants A12-24 (generated from A12) and A18-2 (generated from A18) with a nuclear DNA content corresponding to a triploid were crossed with ‘Yatsugatake D12’ plants as maternal parents. BC2 progeny of A12-24 and A18-2 with a nuclear DNA content corresponding to a diploid were designated ‘IPA12-24’ (n = 79) and ‘IPA18-2’ (n = 77). For the linkage analysis of the winter hardiness QTL on chromosome 7, a BC2 population (n = 186) derived from the A12-103 plant (from IPA12), which carried a F. pratensis chromosome 7 segment, was generated by backcrossing with a ‘Yatsugatake D12’ plant as a maternal parent. Five plants from ‘Yatsugatake D12’ used in the five backcrossing are different genotypes, respectively. Flowers of all maternal parents were emasculated and enclosed in bags.

Fig. 1
figure 1

A schematic diagram of the breeding process of introgression populations used in this study. Lp and Fp mean Lolium perenne and Festuca pratensis, respectively. Boxes indicate populations, and the others are individuals. All Lp ‘Yatsugatake D12’ individuals used for backcrossing were different genotypes. Ploidy level was estimated from the nuclear DNA content

Full size image

Genotyping and linkage mapping

To investigate the introgression and segregation of F. pratensis alleles, DNA markers polymorphic between L. perenne and F. pratensis were used for genotyping. For introgression mapping, 84 Lolium/Festuca intron-flanking markers (Supplementary material 1) were used to genotype 203 individuals. Many of these markers were previously described by Ishikawa et al. (2007, 2009) and Tamura et al. (2009, 2012); primer sequences for some markers were modified (Supplementary material 1). Eight markers were newly developed using the previously described method (Tamura et al. 2009, 2012). Genotyping was performed as in Tamura et al. (2009, 2012). Markers were named based on the Institute for Genomic Research (TIGR) locus names of corresponding single-copy rice genes (Supplementary material 1). The chromosomal location of each marker was previously determined using seven Lolium/Festuca monosomic substitution plants (Tamura et al. 2009; Harper et al. 2011) or estimated using L. perenne GenomeZipper data (Byrne et al. 2015).

For the segregation mapping of a F. pratensis segment of chromosome 7, ten Lolium/Festuca intron-flanking markers and six L. perenne EST-derived SSR markers located on L. perenne linkage group 7 (Studer et al. 2010) were used. The presence or absence of a F. pratensis allele was regarded to indicate a heterozygote or homozygote, respectively. Linkage analysis was performed using the program JoinMap version 4.0 (Van Ooijen 2006) with the population type code “BC1”. Marker order and genetic distance were calculated using a regression mapping algorithm with Kosambi’s mapping function.

Genomic in situ hybridization (GISH)

The experiments were performed as described in Kubota et al. (2015).

Field experiment with introgression populations

Plants of the three introgression populations and their parents grown in plastic nursery pots (5 cm × 5 cm × 5 cm) in a greenhouse were transplanted to the field nursery of the Hokkaido Agricultural Research Center, National Agriculture and Food Research Organization (HARC/NARO) in Sapporo, Japan (43°00′N, 141°42′E) on 16 July 2010. Three clones of each genotype were placed in 80 × 50 cm with a randomized block design. The environmental conditions of the 2010/2011 and the 2011/2012 winter are indicated in Supplementary material 2. The snow cover period in the second winter (2011/2012; 136 days) was longer than that in the first winter (2010/2011; 114 days). Cutting at a height of 7 cm above the ground was performed when the plant length reached 30–40 cm, followed by application of chemical fertilizer (21 kg N ha−1, 16 kg P2O5 ha−1 and 21 kg K2O ha−1). In 2011, the first cutting was conducted on 5 May and was followed by six cuttings until 28 September. In 2012, the first cutting was conducted on 18 May. Phenotyping was performed on three replicates before the first cutting in 2011, after that was performed on two replicates. Dry matter in the first cutting (early-spring dry matter) was measured after drying at 70 °C for 72 h. Winter hardiness was scored on a scale of 1–9 according to the vigor of sprouting after snowmelt (20 April 2011 and 2 May 2012), with 1 corresponding to no sprouting and 9 corresponding to very vigorous sprouting: Score 5 corresponded to moderate sprouting across the stub but with some injured parts. Plant vigor in fall (5 December 2010 and 10 December 2011) was evaluated as growth on a scale of 1–9, with 1 corresponding to very little growth and 9 corresponding to extensive growth: Score 5 was intended to be scored as median of all genotype. Dates of field operations were summarized in Supplementary material 3.

Field experiment with a Festuca pratensis chromosome 7 segregation population

For the phenotyping, of 186 genotypes used for linkage mapping, we randomly selected 52 of 55 recombined plants, 25 of 40 non-recombined heterozygous (Festuca/Lolium) plants and 21 of 91 of non-recombined L. perenne homozygous plants, respectively, because our nursery did not have enough area to investigate all genotypes, and the population have a significant segregation distortion with a lower frequency of F. pratensis alleles. Three clones of each genotype grown in a plastic nursery pots (5 cm × 5 cm × 5 cm) in a greenhouse were placed in 80 × 50 cm with a randomized block design at the field nursery of HARC/NARO on 13 September 2012. The environmental condition of the 2012/2013 winter is indicated in Supplementary material 2. The snow cover period was 151 days. Indexes of fall vigor, winter hardiness, and spring vigor were scored on 13 November, 23 April, and 13 May, respectively. Spring vigor was evaluated similar to fall vigor as described above. Overhead digital images of each plant were taken on 23 April. Using Photoshop CS6 (Adobe Systems, San Jose, CA, USA), the area of whole plant and its green parts were manually picked out from the background using the Magic Extractor tool and their pixels were counted. Proportion of green parts was calculated as a percentage ratio of the number of green part pixels to that of the whole plant. Heading date (ear emergence) was measured as days after 1 June 2013. Dry matter was measured at the heading stage as described above. Dates of field operations were summarized in Supplementary material 3.

Freezing tolerance test

Freezing tolerance of crown tissues was evaluated according to Moriyama et al. (1995) with some modifications. Three clones per genotype were transplanted in plastic pots (4 cm × 4 cm × 4 cm) on 23 July 2010 and were kept in a greenhouse. Plants were transferred outside at HARC/NARO on 8 October 2010. Freezing treatment was performed on 13–15 December 2010. Three clones were bulked, their tillers were divided, and leaves and roots were cut off. On average, seven crowns were wrapped together in a moist absorbent cotton sheet, covered with aluminum foil, and placed in a programmable freezer (LU-112, Tabai ESPEC, Osaka, Japan). After ice nucleation at −3.0 °C for 6 h, the temperature was reduced by 1 °C h−1 to −12.5 °C or −14 °C and then maintained for 3 h; after freezing, the crowns were transferred to 6 °C overnight. Crowns were transplanted into vermiculite in a plastic box and grown in a greenhouse for 4 weeks. Each divided tiller was scored, and average Larsen’s visual score was calculated for each genotype according to Kosmala et al. (2006) with some modifications: 1 = dead, no sign of leaf elongation; 2 = dead but leaves having previously elongated < 0.5 cm; 3 = dead, but leaves having previously elongated 0.5–2 cm; 4 = likely to die, but at least one new root elongated; 5 = likely to survive, but badly damaged; 6 = plant survived, but with severe damage to approximately 50% of leaves; 7 = plant survived but with visual signs of freezing injury; 8 = minimum freezing injury; 9 = no visible signs of injury.

Statistical analysis

Mean phenotypic values for each trait were compared between the presence and absence of a F. pratensis allele at each locus in the three bulked introgression populations by t test using JMP 9 (SAS Institute, Cary, NC, USA). To eliminate the background effect attributed to the different parents among three populations, values of the IPA12-24 and IPA18-2 populations were standardized by multiplication by the ratio of the mean value of IPA12 to the mean value for the respective population. Populations with no F. pratensis–derived introgressed allele at a locus were excluded from the comparison at that locus. Significance level was set at P (−log10) > 3.2 based on the Bonferroni correction at α = 0.05. QTL analysis using a population segregating for a F. pratensis–derived segment of chromosome 7 was conducted using a simple interval mapping procedure of MapQTL 6 (Van Ooijen 2009). Analysis was performed only for linkage group 7. Permutation tests (1000 iterations) were performed to determine the LOD threshold, and loci with a LOD score above the threshold (P < 0.05) were considered as significant QTLs.

Results

Genetic characterization of introgression populations

A total of 203 diploid progeny of triploid L. perenne/F. pratensis hybrids backcrossed with diploid L. perenne were genotyped using 84 intron-flanking markers randomly distributed across the genome. Among the 203 plants, 138 (68%) had a F. pratensis-specific allele at least in one locus: 41 of 47 (85%) in IPA12, 45 of 79 (57%) in IPA12-24, and 52 of 77 (68%) in IPA18-2 (Supplementary materials 4–6). The mean and maximum numbers of introgressed loci per plant were 6.8 and 22 in IPA12, 3.0 and 15 in IPA12-24, and 3.8 and 32 in IPA18-2. Based on the genotyping data, it was assumed that 41% (n = 84), 18% (n = 36), 6% (n = 13), 2% (n = 3) and 1% (n = 2) of the plants in the three populations had F. pratensis introgression(s) in one, two, three, four and five chromosome(s), respectively. Genomic in situ hybridization analysis with probes for genomic DNA of L. perenne and F. pratensis also revealed that several individuals had partial or whole chromosome(s) from F. pratensis (Supplementary material 7). Introgression frequency for each chromosome is shown in Table 1. Mean introgression frequency for all 84 loci was 8.1% in IPA12, 3.7% in IPA12-24, and 4.6% in IPA18-2 (Table 1). For some loci, the introgression frequency differed significantly among the three populations; in particular, the introgression frequency of chromosome 1 was much higher in IPA12 than in the other populations (P < 0.0001, Chi-square test; Table 1). The hybrid parents A12-24 and A18-2 had no F. pratensis allele at 7 and 19 loci, respectively (e.g., on chromosome 2 in A18-2); at such loci, backcross progeny naturally had no F. pratensis alleles (Supplementary materials 5, 6). Festuca pratensis-derived alleles were confirmed at all investigated loci except Os02g17870 on chromosome 6 (Supplementary materials 4–6); the latter locus was excluded from association analysis.

Table 1 Festuca pratensis allele frequency in the three introgression populations
Full size table

Winter hardiness and freezing tolerance in the introgression populations

Fall vigor index, winter hardiness index, and early-spring dry matter were evaluated over 2 years. After both winters, disease caused by snow molds, especially Typhula ishikariensis, was detected in most plants. Festuca pratensis and hybrid parents of the introgression populations showed superior winter hardiness and higher dry matter in early spring than L. perenne in both years (Table 2). Mean values of winter hardiness index and early-spring dry matter of the three introgression populations were lower than those of the corresponding hybrid parents and higher than those of L. perenne, except for the winter hardiness index after the first winter in IPA12-24 (Table 2). Only one plant died (i.e., winter hardiness index was 1.0) in IPA18-2 after the first winter; a total of 31 plants died after the second winter (7 in IPA12, 5 in IPA12-24, and 19 in IPA18-2). Introgression populations showed a wide range of winter hardiness, but only two plants showed greater winter hardiness than that of hybrid parents after the first winter (one in IPA12-24 and the other in IPA12-18), and no plant did after the second winter (Table 2).

Table 2 Fall vigor index (FVI), winter hardiness index (WHI), early-spring dry matter (ESDM), and freezing tolerance of the Lolium–Festuca introgression populations and their parents [mean ± standard deviation (minimum–maximum)]
Full size table

Freezing tolerance was investigated in 152 plants arbitrarily selected from the three populations. The order of freezing tolerance in both (−12.5 and −14 °C) treatments was as follows: F. pratensis > hybrid parents > L. perenne parents (Table 2). Introgression populations showed a wide range of freezing tolerance, but the mean value was similar to that of L. perenne (Table 2). After −14 °C treatment, freezing tolerance of only three plants (two in IPA12 and one in IPA12-24) was higher than that of the hybrid parents.

All correlation coefficients among fall vigor index, winter hardiness index, and early-spring dry matter in both years were positive and significant (Table 3), especially those between winter hardiness index and early-spring dry matter (0.73 and 0.77 after the first and second winter, respectively). The correlation coefficient between freezing tolerance at −12.5 and −14 °C was 0.52. Correlation coefficients between freezing tolerance at −12.5 °C and winter hardiness index were small but significant in both years (0.24 in 2010/2011 and 0.17 in 2011/2012; Table 3). Freezing tolerance at −14 °C did not show significant correlation with any traits (Table 3).

Table 3 Pearson’s correlation coefficients among winter hardiness-related traits in the bulked Festuca/Lolium introgression population
Full size table

Association between introgressed F. pratensis alleles and winter hardiness

Mean values of the winter hardiness-related traits with or without F. pratensis allele were compared at 83 loci in the bulked introgression population (mean F. pratensis allele frequency, 5.9%; Table 1). Significant differences in the fall vigor index assessed in the first year were detected on chromosome 5, with negative effects of F. pratensis alleles (Table 4). After the first winter, significant negative effects of F. pratensis alleles were also detected for winter hardiness index on chromosomes 4 and 6, and for early-spring dry matter on chromosome 4 (Table 4). After the second winter, there was no significant difference for fall vigor index and winter hardiness index at any locus, but significant differences in early-spring dry matter were detected at Os06g13810 and Os10g25360 loci on chromosome 7, with positive effects of F. pratensis alleles (Table 4). For freezing tolerance, no significant difference at P (−log10) > 3.2 was detected at any locus. At the locus Os02g03260 on chromosome 6, the F. pratensis allele increased freezing tolerance, although this effect was not significant: P (−log10) = 2.2 (−12.5 °C) and 2.6 (−14 °C).

Table 4 Association between winter hardiness-related traits and the presence or absence of Festuca pratensis alleles in the bulked Lolium/Festuca introgression population
Full size table

Winter-hardiness QTLs on chromosome 7

To confirm the positive effect of F. pratensis alleles on chromosome 7 on winter hardiness, we performed QTL analysis using the progeny of A12-103, which had an introgression of a F. pratensis chromosome 7 segment including loci that had significant effects on early-spring dry matter, backcrossed with L. perenne. This backcross population showed significant segregation distortion, with a lower frequency of the F. pratensis alleles (average ratio of L. perenne/L. perenne:F. pratensis/L. perenne genotype = 1:0.53 [SD, 0.04], whereas the expected ratio was 1:1; P < 0.0001, Chi-square test). A genetic linkage map of the introgressed F. pratensis region in linkage group 7 was constructed; this map spanned 34.1 cM (Fig. 2). In the field experiment, snow mold disease was detected after the winter, and about 30% of the clones died. One week after the snow had melted, the average proportion of green parts (leaves that stayed green under the snow, or sprouts) in each plant was 1.8% (Table 5). Winter hardiness index, winter survival rate, the proportion of green parts, spring vigor index, and dry matter at heading showed high positive correlation (Table 6). Fall vigor index and heading date showed no significant correlation with winter hardiness-related traits, except for a negative correlation between fall vigor index and proportion of green parts (Table 6). QTLs for winter hardiness index, winter survival rate, and proportion of green parts were detected around the Os08g33630 locus with R 2 6.7–9.6 (Table 5); the LOD 1.0 support interval included Os10g25360, which was associated with early-spring dry matter in the introgression population mentioned above (Fig. 2). For all these traits, F. pratensis alleles had positive effects (Table 5). No QTLs were detected for spring vigor index, heading date, or yield at heading.

Fig. 2
figure 2

Winter hardiness-related QTLs in linkage group 7 in the backcross progeny of Lolium/Festuca introgression genotype A12-103. Bars 1.0-LOD support intervals, triangles LOD peaks. Loci significantly associated with early-spring dry matter in the Lolium/Festuca introgression populations are shown in bold

Full size image
Table 5 Values of and quantitative trait loci for winter hardiness-related traits in the population segregating for a Festuca pratensis-derived segment of chromosome 7 in the Lolium background
Full size table
Table 6 Pearson’s correlation coefficients among winter hardiness-related traits in the population segregating for a Festuca pratensis–derived segment of chromosome 7 in the Lolium background
Full size table

Discussion

In this study, we used association analysis of introgression populations followed by linkage analysis of a backcross population, and identified a QTL on F. pratensis chromosome 7 with a positive effect on winter hardiness in the field. We detected no QTLs for heading date or dry matter yield at the heading stage in the backcross population, which indicates that the identified QTL does not affect growth rate in spring. In the field nursery where this study was performed, snow depth was sufficient to prevent soil freezing during winter; this is consistent with the absence of correlation between winter-hardiness index and freezing tolerance in the introgression population. In this field nursery, a major factor of winter injury is snow mold disease (Sanada et al. 2007), which was caused mainly by T. ishikariensis in this study. Therefore, the QTL detected on chromosome 7 might be involved in snow mold resistance. The absence of an association between winter hardiness and F. pratensis chromosome 7 introgression in the first winter (2010/2011) is likely attributable to a shorter period of snow cover than in winters 2011/2012 and 2012/2013, because the duration of snow cover is the most critical factor for snow mold disease (Matsumoto and Hoshino 2013). Previously, no QTL for winter hardiness in the field was detected in linkage group 7 in L. perenne or F. pratensis (Yamada et al. 2004; Alm et al. 2011; Paina et al. 2016), although Alm et al. (2011) detect a frost tolerance QTL and a drought tolerance QTL on chr 7 in F. pratensis. It might be because these studies did not use intergeneric hybrids. Further studies of snow mold infection and comparative genomic studies would reveal the function of this QTL.

Triploid hybrids between F. pratensis and L. perenne showed winter hardiness and freezing tolerance clearly superior to those of L. perenne, although they were not as tolerant as F. pratensis. Winter hardiness of almost all backcross progeny of the triploid hybrids was lower than that of the parents, but the average was still higher than that of L. perenne. However, we identified only one QTL with the F. pratensis allele having a small positive effect in the association analysis. These data suggest that superior winter hardiness of F. pratensis compared to L. perenne is conferred by multiple small-effect QTLs, most of which were not detected in this study. A small number of genotypes with F. pratensis introgression (mean F. pratensis allele frequency, 5.9%) might have prevented the detection of QTLs with small effects. Analysis of backcross progeny of plants carrying introgressed segments of each F. pratensis chromosome (similar to the analysis performed in our study for chromosome 7) would be more precise because many genotypes with recombined Festuca alleles would be available. A high frequency of the allele introgressed from Festuca would also enable detection of epistatic effects among multiple introgressed regions, which were found in the populations used in this study, for winter-hardiness related traits. This study used DNA markers derived from ESTs homologous to rice genes distributed over the whole genome. GenomeZipper between L. perenne and model grasses reported by Byrne et al. (2015) shows that some regions of the L. perenne genome do not correspond to any rice genes, and additional markers would be needed to identify small-effect QTLs that might be located in such regions. In this study, we did not detect the QTLs for freezing tolerance on chromosomes 2 and 4 reported by Kosmala et al. (2006, 2007). In addition to the low introgression frequency mentioned above, this discrepancy might be due to the difference in the background genomes: L. multiflorum was used by Kosmala et al. (2006, 2007), whereas we used L. perenne. Generally, winter hardiness and freezing tolerance of L. multiflorum are inferior to those of L. perenne; therefore, the positive effect of Festuca introgression could be clearer in L. multiflorum than in L. perenne.

Little information on the negative effects of Festuca introgression into Lolium is available, because materials used in previous studies often originated from selected genotypes with superior stress tolerance (Grønnerød et al. 2004; Kosmala et al. 2006, 2007). We detected a negative effect of the introgression of F. pratensis segments of chromosomes 4 and 6 on winter hardiness index after the first winter. In these loci, genotypes with F. pratensis alleles tended to show lower scores of fall vigor index in the year of transplantation (data not shown). Fall vigor index was positively correlated with winter hardiness in both years. Therefore, the negative effects of Festuca introgression on winter hardiness might be due to reduced plant vigor.

King et al. (2013) revealed the presence of directional selection pressure for both the transmission of F. pratensis chromosomes and L. perenne/F. pratensis recombination in the backcross progeny of monosomic substitution lines. They reported that 72 of 73 backcross progeny with F. pratensis chromosome 1 substitution in the L. perenne background had recombined L. perenne/F. pratensis chromosomes. We confirmed positive selection for the transmission of recombinant L. perenne/F. pratensis chromosome 1 in the IPA12 population but not in IPA12-24 or IPA18-2. The difference in the transmission frequency of F. pratensis alleles among the populations might be caused by cytoplasmic type of the triploid parent: the triploid parent of IPA12 has the F. pratensis cytoplasm, whereas those of IPA12-24 and IPA18-2 have the L. perenne one. In a triploid hybrid of F. pratensis and L. multiflorum, the frequency of intergeneric translocations was higher in the Festuca-derived cytoplasm than in the Lolium-derived one (Zwierzykowski et al. 1999; Naganowska and Zwierzykowska 2001). King et al. (2013) also reported neutral selection pressure for the transmission of recombinant L. perenne/F. pratensis chromosomes, including chromosome 7. However, in this study, the progeny of a plant carrying a fragment of F. pratensis chromosome 7 backcrossed with L. perenne showed significant segregation distortion with a lower frequency of the F. pratensis allele. Further investigation of selection pressure for the transmission of Festuca chromosomes in the Lolium background using various genotypes is needed.

Conclusion

In this study, unlike previous studies of Festulolium, association between the introgression and winter hardiness were comprehensively surveyed over the genome using introgression populations. As a result, we identified a QTL on chromosome 7 with a positive effect of the F. pratensis allele on winter hardiness in the L. perenne genomic background, although multiple QTLs with small effects seem to be involved in the improvement of winter hardiness-related traits in Festulolium in comparison with L. perenne. We also found negative effects of F. pratensis alleles on winter hardiness in some genomic regions. Although the improvement of the winter hardiness of Lolium by introgression of a few Festuca genomic regions might be limited, even a minor improvement in winter hardiness with preservation of the advantageous traits of Lolium (e.g., good forage quality and rapid initial growth) would be useful. In addition to the previously reported QTLs, QTL information obtained in this study would improve the selection efficiency. Meanwhile, for a major improvement in winter hardiness-related traits, the use of F1 hybrids such as amphidiploids is a better approach in Lolium/Festuca breeding (unpublished data). QTL information obtained here might also be useful for the fine tuning of winter hardiness-related traits in amphidiploid Festulolium breeding.