Cytogenetics of Festulolium (Festuca ! Lolium Hybrids)

Cytogenetics and Plant Bleeding

Cytogenet Genome Res 120:370–383 (2008) DOI: 10.1159/000121086

Cytogenetics of Festulolium (Festuca ! Lolium hybrids)

a b a D. Kopecký A.J. Lukaszewski J. Doležel

a Laboratory of Molecular Cytogenetics and Cytometry, Institute of Experimental Botany, Olomouc b (Czech Republic); Department of Botany and Plant Sciences, University of California, Riverside, CA (USA)

Accepted in revised form for publication by M. Schmid, 29 September 2007.

Abstract. Grasses are the most important and widely cul- ted detailed studies of intergeneric chromosome recombi- tivated crops. Among them, ryegrasses (Lolium spp.) and nation and karyotyping of Festulolium cultivars. These fescues ( Festuca spp.) provide high quality fodder for live- tools were also invaluable in revealing the origin of poly- stock, are used for turf and amenity purposes, and play a ploid fescues, and facilitated the development of chromo- fundamental role in environment protection. Species from some substitution and introgression lines and physical map- the two genera display complementary agronomic charac- ping of traits of interest. Further progress in this area will teristics and are often grown in mixtures. Breeding efforts require the development of a larger set of cytogenetic mark- to combine desired features in single entities culminated ers and high-resolution cytogenetic maps. It is expected that with the production of Festuca ! Lolium hybrids. The so the Lolium – Festuca complex will continue providing op- called Festuloliums enjoy a considerable commercial suc- portunities for breeding superior grass cultivars and the cess with numerous cultivars registered all over the world. complex will remain an attractive platform for fundamental They are also very intriguing from a strictly cytogenetic research of the early steps of hybrid speciation and interac- point of view as the parental chromosomes recombine free- tion of parental genomes, as well as the processes of chro- ly in hybrids. Until a decade ago this phenomenon was only mosome pairing, elimination and recombination. known in general quantitative terms. The introduction of Copyright © 2008 S. Karger AG, Basel molecular cytogenetic tools such as FISH and GISH permit-

Open landscape is usually covered by grasses, and by nity purposes, in sport and urban contexts where their aes- learning that their color is green, humans are aware of their thetic value cannot be overestimated. Last but not least, existence since early childhood. Not many people know, grasses play a fundamental role in soil conservation and en- however, that grasses are one of the most important culti- vironment protection (Wang and Ge, 2006). vated plants on Earth, with the total area of grassland esti- Due to their value, grasses are often preferred over other mated to be twice that of cropland (Jauhar, 1993). There are crops. For example, in spite of environmental conditions many important uses of grasses: they are widely used for amenable to production of cereals, corn, sugar beet and po- pasture and play a key role in providing high quality fodder tato, nearly 40% of the total acreage in the Czech Republic for livestock. Some grass species are used for turf and ame- is covered by grassland (Černoch et al., 2003) while in the United Kingdom, 75% of the total agricultural land is de- voted to grass production (Thomas et al., 2003). Out of many grass species, ryegrasses (Lolium spp.) and fescues This work was supported in part by the Czech Science Foundation (grant award (Festuca spp.) predominate. In Europe, 23% of the grassland 521/07/P479). area (52 million ha) is covered by just two ryegrass species, Request reprints from D. Kopecký Lolium multiflorum Lam. and Lolium perenne L. The an- Institute of Experimental Botany, Sokolovská 6 nual production of ryegrass seed in Europe exceeds 45,000 CZ–772 00 Olomouc (Czech Republic) telephone: +420 585 205 857; fax: +420 585 205 853 tons with the value of 160 million € (Lübberstedt et al., e-mail: [email protected] 2003).

Fax +41 61 306 12 34 © 2008 S. Karger AG, Basel Accessible online at: E-Mail [email protected] 1424–8581/08/1204–0370$24.50/0 www.karger.com/cgr www.karger.com Table 1. Chromosome number and DNA content of some ryegrass and fescue species

Species Common name Chromosome number/genome 1C DNA contenta pg DNA Referenceb

Lolium perenne Perennial ryegrass 2n = 2x = 14 2.08 1 Lolium multiflorum Italian ryegrass 2n = 2x = 14 4.10 2 Festuca pratensis Meadow fescue 2n = 2x = 14; AA 2.20 3 Festuca scariosa 2n = 2x = 14; BB 2.68 4 Festuca glaucescensc 2n = 4x = 28; BBCCe 4.28 5 Festuca mairei 2n = 4x = 28; BBDDe 3.95 5 Festuca arundinacea Tall fescue 2n = 6x = 42; AABBCCe 6.05 6 Festuca gigantead Giant fescue 2n = 8x = 56; AAXXYYZZe 7.23 5

a DNA content of one non-replicated chromosome complement (holoploid genome size). For the terminology on genome size see Greilhuber et al. (2005). b Key to references: 1 = Evans et al. (1972), 2 = Schifino and Winge (1985), 3 = Olszewska and Osiecka (1982), 4 = Harper et al. (2004), 5 = Seal AG. (pers. comm. 1979 in: Bennett et al. 1982), 6 = Ceccarelli et al. (1992). c Sometimes referred as F. arundinacea var. glaucescens and F. arundinacea var. fenas. d Hexaploid cytotypes were reported by some authors (i.e. Morgan et al., 1988). e Genomic constitution of allopolyploid species has only been considered by Bowman and Thomas (1976), Thomas et al. (1997) and Harper et al. (2004); further analyses are needed to prove it.

Whereas the genus Lolium consists of eight species, the Chromosome number and morphology genus Festuca is diverse and comprises almost 500 species (Clayton and Renvoize, 1986). The genus has been subdi- While all eight species comprising genus Lolium are dip- vided into six sections: Ovinae , Bovinae , Sub-bulbosae , loid (2n = 2x = 14), ploidy levels vary greatly in the genus Variae , Scariosae and Montanae (Hackel, 1882). In a recent Festuca (Table 1), ranging from diploid (i.e. F. pratensis ; classification, Clayton and Renvoize (1986) divide the genus 2n = 2x = 14), through tetraploid (i.e. F. arundinacea var. Festuca into nine subgenera: Hesperochloa , Xanthochloa , glaucescens (= F. glaucescens); 2n = 4x = 28), hexaploid (i.e. Drymanthele , Schedonorus , Subulatae , Subuliflorae , Obtu- F. arundinacea var. genuina (= F. arundinacea); 2n = 6x = sae , Festuca and Helleria . Of this plethora of species mainly 42), octoploid (i.e. F. arundinacea var . atlantigena forma F. pratensis and F. arundinacea , which belong to the Bovinae pseudo-mairei Lit. and Maire; 2n = 8x = 56), decaploid (i.e. section (subgenus Schedonorus ), are used in agricultural F. arundinacea var . letourneuxiana St. Yves; 2n = 10x = 70) production. to dodecaploid (i.e. F. summilusitana; 2n = 12x = 84). Only P e r e n n i a l r y e g r a s s (L. perenne) is the single most impor- a few studies dealt with the origin of polyploids (i.e. Hum- tant grass species and is used for grazing, hay and turf. This phreys et al., 1995). Nevertheless, all polyploid Festuca spe- species is native to Central and South Europe, North-West cies described herein are believed to be allopolyploid (Ta- Africa and South-West Asia, and comprises annual and bi- ble 1 ). Based on the widespread occurrence of polyploidy, ennial types. It is widespread in most temperate regions, Malik and Thomas (1966) considered Festuca more ancient including Australia and New Zealand, where it was intro- than Lolium . duced by the first settlers. Italian ryegrass (L. multiflorum ) Malik and Thomas (1966) were among the first to is a close relative of perennial ryegrass with similar geo- karyotype Festuca and Lolium. In ryegrasses (including graphic distribution, and is widely used as fodder plant for L. perenne and L. multiflorum), they identified three pairs hay and grazing. Meadow fescue (Festuca pratensis Huds.) of submedian chromosomes with secondary constrictions is a perennial grass of Europe and South-West Asia, which (sites of the nucleolar organizing regions, NORs), one large has been introduced into North America. Tall fescue ( Fes- median pair and three smaller pairs of submedian chro- tuca arundinacea Schreb.) is a perennial widely distributed mosomes. More recently, Thomas et al. (1996) confirmed over Europe, North-West Africa and temperate Asia and the presence of NORs on three pairs of L. multiflorum was also introduced into North America. The present re- chromosomes after fluorescence in situ hybridization view focuses on cytogenetics of grass species which are used (FISH) with the 45S rDNA probe. They also detected an in breeding of improved cultivars, including interspecific additional seventh site of 45S rDNA in L. perenne . This and intergeneric hybrids, and the main attention is given to site co-localized with a Giemsa band located proximally Festuca ! Lolium hybrids. Only scant attention is given to on the long arm of chromosome 5 (Thomas, 1981; Thomas other species belonging to both genera. et al., 1996). Interestingly, this seventh site of 45S rDNA was undetectable in our analyses (Kopecký et al., unpublished).

Cytogenet Genome Res 120:370–383 (2008) 371 In contrast to the similarity of karyotypes in various spe- in F. arundinacea and speculated on a role of L. multiflorum cies of Lolium , Festuca species display larger karyotypic in the evolution of F. arundinacea after allopolyploidization variation. By analyzing karyotypes in 12 fescue species after and diversification from its progenitors. orcein staining, Malik and Thomas (1966) identified four Other phylogeny studies in polyploid species were based pairs of chromosomes with secondary constrictions in F. on FISH with probes for rDNA. Theoretically, an autotetra- pratensis and six pairs of such chromosomes in F. glauces- ploid should have two pairs of chromosomes with the same cens. However, lower numbers of NORs were detected by chromosomal distribution of rDNA loci as its diploid par- silver staining (Carnide et al., 1986) and by FISH with 45S ent. A different number and/or distribution of rDNA loci rDNA (Thomas et al., 1997b). might indicate allopolyploidy. Based on the number and po- Mitotic chromosomes of both ryegrass and fescue spe- sition of 45S rDNA and 5S rDNA loci, Thomas et al. (1997b) cies are similar in size and parental chromosomes cannot confirmed that F. pratensis and F. glaucescens were the probe easily distinguished in hybrids using simple chromo- genitors of F. arundinacea. The two chromosome pairs of F. some staining techniques. The similarities in chromosome pratensis (carrying either 5S rDNA or 45S rDNA) are pres- sizes across the two genera correspond to the similarities in ent also in F. arundinacea . However, the authors found dif- nuclear DNA content (see Table 1). The only conflicting references in the number and distribution of rDNA loci in F. port provides DNA content of L. multiflorum (Schifino and glaucescens and in the genome of F. glaucescens that is pres- Winge, 1985) approximately twice that of L. perenne (Evans ent in hexaploid F. arundinacea . This suggests some genome et al., 1972) and F. pratensis (Olszewska and Osiecka, 1982). modifications either in F. glaucescens or in F. arundinacea . As the chromosomes of all three species are similar in size, A similar loss of rDNA loci of diploid progenitors was re- this discrepancy could be due to the analysis of tetraploid ported in hexaploid oat (Leggett and Markhand, 1995) and rather than a diploid cytotype of L. multiflorum . polyploid Scilla autumnalis (Vaughan et al., 1993). After analyzing the number and distribution of rDNA loci in tetraploid F. glaucescens and F. mairei , Thomas et al. (1997b) as- S y s t e m a t i c s o f Festuca and Lolium based on sumed that both have one common genome while the other chromosome studies one has undergone only minor changes. Also hexaploid F. arundinacea and F. gigantea seem to share one common ge- Eurasia, with a frequent occurrence of diploid and tetra- nome, which is considered to be of the F. pratensis origin ploid Festuca species is considered the region of origin of (Thomas et al., 1997b). Festuca genus and the center of its primary diversification Harper et al. (2004) analyzed karyotypes of diploid Fes- (Dubcovsky and Martínez, 1992). Genus Lolium (and sev- tuca species from sections Montanae ( F. altissima, F. eral others, i.e. Vulpia and Micropyrum ) is pictured as a de- drymeja, F. donax and F. lasto) and Scariosae ( F. scariosa ). rived group of Festuca (Clayton and Renvoize, 1986). Based Based on the distribution of rDNA loci, they considered F. on the sequence analysis of the ITS (internal transcribed scariosa a progenitor of F. mairei and F. glaucescens and F. spacer) and chloroplast trn region, Catalan et al. (2004) altissima as good progenitor candidates for polyploid spe- found close relationship of both groups (genus Lolium and cies of the Bovinae section. FISH with rDNA probes and Festuca subgen. Schedonorus). However, Darbyshire (1993) GISH were also used for phylogenetic studies of the Iberian proposed to transfer four species of the Festuca subgenus Peninsula fescues (Loureiro et al., 2007). Diploid species F. Schedonorus ( F. pratensis , F. arundinacea , F. mazzettiana , henriquesii was found to be a good candidate in evolution F. gigantea) and interspecific hybrid ( L. perenne ! F. pra- of F. ampla , as it had the same genomic distribution of tensis) to genus Lolium . From the new molecular analyses, rDNA loci. However, GISH did not confirm this hypoth- it appears that Festuca subgen. Schedonorus is far more esis. distant from Lolium and should remain in genus Festuca The use of rDNA loci in evolutionary studies may be lim- (Catalan et al., 2004). ited by their variability. We have observed three chromo- Soon after the first use of the in situ hybridization with some pairs with the 45S rDNA hybridization signals in dip- DNA probes (genomic in situ hybridization, GISH) in the loid L. perenne (data not published) while Thomas et al. Festuca ! Lolium hybrids (Thomas et al., 1994), the meth- (1996) reported an additional seventh site. Similar differ- od was applied to study the evolution of polyploid Festuca ences were found among plants of F. glaucescens and F. species. F. arundinacea is a hexaploid species widely distrib- arundinacea, where different numbers of 5S rDNA loci were uted over the world and used as a forage and turf grass. Jau- observed (Thomas et al., 1997b; our unpublished results). har (1993) suggested its allopolyploid origin with three dip- Recently, Loureiro et al. (2007) found differences in the loid progenitor genomes involved, A, B and C. The A ge- number of rDNA loci among the plants of several polyploid nome was known to be donated by F. pratensis , but the fescue species (e.a. F. summilusitana , F. duriotagana , F. ru- origin of the B and C genomes remained unknown until a bra ssp. litoralis ). The question is if these differences repre- GISH analysis of Humphreys et al. (1995). This analysis sent an intraspecific variation or are due to variability in unequivocally revealed that allotetraploid F. glaucescens FISH protocol. While a failure of FISH to detect weak sites (BBCC) participated in the evolution of allohexaploid F. of hybridization cannot be excluded, a mobility of rDNA arundinacea (AABBCC). Furthermore, using GISH, Paša- loci has been described in some species (Schubert and kinskiene et al. (1998) detected chromatin of L. multiflorum Wobus, 1985; Dubcovsky and Dvorak, 1995).

372 Cytogenet Genome Res 120:370–383 (2008) Hybrids between Festuca and Lolium Table 2. Fertility of hybrid progeny in relation to the ploidy of parental species L. perenne and L. multiflorum establish rapidly and are highly palatable, nutritious, and digestible. On the other Parent 1 (ploidy) Parent 2 (ploidy) Fertility of F1 hybrid progeny hand, Festuca species are generally more persistent and tol- Lolium (2x) F. pratensis (2x) sterile erate low levels of nutrients and extreme abiotic stresses Lolium (2x) F. pratensis (4x) sterile such as winter freezing and summer drought (Thomas et al., Lolium (4x) F. pratensis (2x) fertile 2003). These complementary attributes of both genera mo- Lolium (4x) F. pratensis (4x) fertile tivate breeders to combine their agronomically important Lolium (4x) F. arundinacea (6x) fertile Lolium (2x) F. arundinacea (6x) male sterile, low female fertile characters in intergeneric hybrids. The species of the two Lolium (4x) F. glaucescens (4x) fertile genera hybridize in nature (Borrill, 1975; Farragher, 1975), Lolium (4x) F. mairei (4x) fertile but their diploid hybrids are sterile. Similarly, only sterile plants were produced experimentally after crossing diploid Lolium and Festuca species (Jauhar, 1975). Two strategies have been used to bypass the problem of hybrid sterility: polyploidization or introgression breeding. arundinacea are morphologically similar to F. arundinacea , In the polyploidization approach, sterile diploid F1 hybrids but have superior yield characteristics. Two cultivars ‘Lofa’ are colchicine treated to produce allopolyploids and thus and ‘Bečva’ were developed from a backcross to L. multiflo- restore fertility, or synthetic (induced) tetraploid parental rum and are successfully used for silage, temporary mead- plants are used for initial crosses. In both cases fertile tetra- ows and pastures (Fojtík, 1998). Some Festulolium cultivars ploids are obtained. In introgression breeding, parental were also developed from L. perenne ! F. pratensis hybrids, plants of different ploidy levels are crossed and the hybrid and include two tetraploid (‘Prior’ and ‘Duo’) and one dip- progeny is backcrossed several times to one of the parents. loid (‘Matrix’) cultivars. These stocks are similar to those This process was used successfully with L. multiflorum ! originating from L. multiflorum ! F. pratensis and are F. arundinacea hybrids, where diploid L. multiflorum plants mostly used for forage. At present there are no registered were used for the initial cross and tetraploid F1 hybrids were cultivars developed from L. multiflorum ! F. glaucescens backcrossed to F. arundinacea (Buckner et al., 1977; Thom- hybrids. However, the strain ‘99-01’ produced at French Na- as and Humphreys, 1991; Fojtík, 1994). The fertility of the tional Institute for Agricultural Research (INRA) (M. Ghes- hybrid progenies relative to the ploidy levels of parental quiere, pers. comm.) is close to registration. It combines a plants used for initial crosses is summarized in Table 2 . high yield of L. multiflorum with resistance to water stress The F1 hybrids that display agronomically attractive and to diseases of F. glaucescens . combinations of parental characteristics are used for selec- The ability to produce intergeneric hybrids between Fes- tion and successive stabilization of hybrid genomes by em- tuca and Lolium and the commercial success of Festulolium ploying intercross or backcross strategies. Some of the re- cultivars stimulated research on the structure and behavior sulting germplasm has been registered as cultivars and are of their genomes. These studies revealed extraordinary currently referred to under a generic term ‘Festulolium’ plasticity of the hybrid genomes and the ability of homoeol- (Dijkstra and de Vos, 1975). The first Festulolium cultivars ogous chromosomes, originating even from different gen- were developed at the Institute of Grassland and Environ- era, to pair and recombine at ease. This work was largely mental Research (IGER), Aberystwyth, Wales more than 30 driven by progress in cytogenetics, and relied on the appli- years ago (Lewis et al., 1973), from hybrids L. perenne ! F. cation of molecular cytogenetics in particular, which en- pratensis (cv. ‘Prior’) and L. multiflorum ! F. pratensis (cv. ables identification of parental genomes in hybrids, identi- ‘Elmet’). Soon after, a successful cultivar ‘Kenhy’ was devel- fication of particular chromosomes and evaluation of their oped by Dr. Buckner in the USA after backcrossing L. mul- behavior, including meiotic pairing, recombination and tiflorum ! F. arundinacea hybrids to F. arundinacea . This elimination. high yielding cultivar has been used for a long time instead or in mixtures with pure F. arundinacea cultivars (Buckner et al., 1977). M o l e c u l a r c y t o g e n e t i c s Stimulated by successful pioneering efforts, over 30 Fes- tulolium cultivars have been registered worldwide, most of Chromosome studies in Festuca , Lolium and their hy- them in central Europe (i.e. Fojtík, 1994; Zwierzykowski et brids used to be hindered by difficulties in discriminating al., 1998b). The majority were produced from reciprocal tet- individual parental chromosomes in hybrids, tracing their raploid hybrids between L. multiflorum and F. pratensis . behavior and identifying their structural changes. With the These cultivars display good yield characteristics, improved exception of some specific cases, differences in chromo- persistency and good winter hardiness. Another group of some size were too minor to be of any use, and there were cultivars was developed from hybrids of L. multiflorum (2x) no patterns of constitutive heterochromatin to serve as C- ! F. arundinacea (6x). Tetraploid F1 hybrids thus obtained bands in karyotyping. So, while chromosome behavior on were backcrossed either to F. arundinacea (6x) or to L. mul- the whole genome level has been worked in considerable tiflorum (4x). The cultivars developed from backcross to F. precision in a very wide range of hybrids, the behavior and

Cytogenet Genome Res 120:370–383 (2008) 373 fate of individual chromosomes was not (Jauhar, 1993). The Nejad et al. (2002) is suitable for large-scale studies. Ko- situation changed dramatically with the advent of molecu- pecký et al. (2005a, 2006) demonstrated its utility in Festuca lar cytogenetics and the development of protocols for fluo- ! Lolium hybrids. rescence in situ hybridization (FISH) and genomic in situ Despite their very high pairing affinity, the genomes of hybridization (GISH). FISH is a molecular cytogenetic Festuca and Lolium genera are distant enough at the DNA method that enables visualization of specific DNA sequenc- level so that their chromosomes can be readily discrimi- es on chromosomes deposited on a microscope slide nated by GISH (Table 3, Fig. 1). Characterization of genome (Schwarzacher, 2003). GISH is a modification of FISH where composition by GISH may play an important role in select- total genomic DNA of one species is labeled and used as a ing potentially useful hybrids during early stages of breed- probe in interspecific hybrids (Schwarzacher et al., 1989). ing as well as in registration of new Festulolium cultivars. This enables identification of the parental genomes and Moreover, the high frequency of intergeneric recombina- analyses of recombination between them. Whereas the tion and relative ease by which the inter-chromosomal original protocols for GISH and FISH were time-consum- transfers can be detected by GISH in Festuca ! Lolium hy- ing and laborious (Schwarzacher and Heslop-Harrison, brids make the method an attractive tool for introgression 2000), a simplified GISH protocol developed by Masoudi- breeding and physical mapping of traits of interest.

Table 3. The results of GISH in Lolium and Festuca species most frequently used in Festulolium breeding (based on the results of Pašakinskiene et al., 1998 and our own unpublished results)

Probe Fluorescent labeling of mitotic chromosomes L. perenne L. multiflorum F. pratensis F. glaucescens F. arundinacea (2n = 2x = 14) (2n = 2x = 14) (2n = 2x = 14) (2n = 4x = 28) (2n = 6x = 42)

L. perenne 14 chromosomes 14 chromosomes 2 chromosomes none 14 chromosomes with numerous with single faint band strong ‘GISH bands’ L. multiflorum 14 chromosomes 14 chromosomes 2 chromosomes none 14 chromosomes with numerous with single faint band strong ‘GISH bands’ F. pratensis none none 14 chromosomes none 14 chromosomes F. glaucescens none none none 28 chromosomes 28 chromosomes F. arundinacea none none none none 42 chromosomes

Fig. 1. Identification of parental chromosomes in F1 hybrid L. mul- Fig. 3. Determination of genomic constitution in Festulolium cv. tiflorum (2x) ! F. arundinacea (6x) using GISH on mitotic metaphase ‘Achilles’ ( L. multiflorum ! F. pratensis ) using GISH on mitotic meta- spread. Labeled genomic DNA of F. glaucescens hybridized with 14 phase chromosomes. Parental chromosome can be easily identified. chromosomes (green color). Labeled genomic DNA of F. pratensis hy- Note the high frequency of intergenomic exchanges. bridized with seven chromosomes (red color). Non-labeled genomic Fig. 4. Genomic constitution of two regenerants with 2n = 28 (a) DNA of L. multiflorum was used as blocking DNA and the chromo- and 2n = 27 (b) obtained from in vitro cultured anthers of L. multiflo- somes were counterstained by DAPI (blue color). Due to the absence of rum ! F. pratensis hybrid (2n = 4x = 28). Both plants were assumed to any labeled probe hybridization, the seven chromosomes L. multiflo- originate from reduced gametes and undergo spontaneous polyploidi- rum appear as blue. zation. 14 (a) and 13 (b) pairs of chromosomes with identical patterns Figs. 2–5. The use of genomic in situ hybridization in L. multiflo- of the parental chromatin were observed (the pairs are marked by let- rum ! F. pratensis hybrids. Genomic DNA of F. pratensis was labeled ters). The monosomic chromosome is marked by an arrow (b). by digoxigenin and detected by anti-DIG-FITC (yellow-green color), Fig. 5. The analysis of L. multiflorum/F. pratensis (2n = 4x = 28) and genomic DNA of L. multiflorum was used as blocking DNA (no monosomic substitution line by GISH on mitotic metaphase chromo- label). The chromosomes were counterstained by DAPI (red pseudo- somes. The monosomic chromosome of F. pratensis is clearly visible color). In Fig. 2 biotin-labeled probe for 45S rDNA was also used and (green color). detected by streptavidin-Cy3 (blue pseudocolor). Fig. 6. Development of cytogenetic markers to identify chromo- Fig. 2. Chromosome pairing during meiotic metaphase I in Festu- somes of F. pratensis (2n = 14). Chromosomes 2 and 3 can be recognized lolium cv. ‘Spring Green’ ( L. multiflorum ! F. pratensis ) using GISH according to the position of 5S and 45S rDNA loci, other three chro- and FISH. Pairing of non-hom(oe)ologous chromosomes resulted in mosomes can be identified using BAC clones that localize exclusively two quadrivalents (arrows), in which only two chromosomes carry 45S to single loci. Sites of probe hybridization were detected using FITC rDNA. Note that quadrivalents of homo(eo)logous chromosomes are (yellow-green signals) and Cy3 (red signals). The chromosomes were expected to carry four 45S rDNA loci (one per each chromosome). counterstained by DAPI (blue color). Assignment of BAC clones to genetic linkage groups is in progress as well as the search for cytogenetic markers to identify the remaining two chromosomes of F. pratensis . Scale bar in each photo represents 10 ␮ m .

374 Cytogenet Genome Res 120:370–383 (2008) 1 2

5 6

Cytogenet Genome Res 120:370–383 (2008) 375 Meiosis in Festuca ! Lolium hybrids (Evans and Davies, 1985). In contrast, Kopecky and Lu- kaszewski (unpublished data) used the tools of cytology and Despite the obvious DNA differences between the chro- directly examined the MI pairing in various combinations mosomes of Festuca and Lolium they readily pair in meiosis of single chromosome introgressions from F. pratensis into (Jauhar, 1975) and recombine (Zwierzykowski et al., 1998b) tetraploid L. multiflorum . These introgressions were mono- suggesting that either the chromosome pairing control sys- somic, disomic, complete chromosome plus its transloca- tem is highly permissive, or that DNA sequences involved in tions, or disomic complete chromosome plus its transloca- chromosome recognition and pairing did not diverge. And tion. Interestingly, only in the disomic condition, when two so, in diploid F. pratensis ! L. multiflorum hybrids (FpLm), pairs of homologous chromosomes were present and com- meiosis seems to be almost regular, with a high frequency of peting for pairing partners (one pair from F. pratensis and bivalents (Jauhar, 1975). Despite regular meiosis, diploid hy- one pair from L. multiflorum) any statistically significant brids are male and female sterile. The reason for this is not deviation from random pairing toward homologous pairing known. In triploid (FpLmLm) and tetraploid hybrids (FpFp- was detected, but even that was small. Given that several dif- LmLm), trivalents and quadrivalents are frequently formed. ferent chromosomes were examined and the genetic back- However, the frequency of trivalents depends on the genom- ground in all combinations was the same, the effect of any ic constitution of the hybrid, being 3.61 per pollen mother pairing control system or differences in ‘pairing rec ognition cell (PMC) in LmLmFp and 2.18 in LmFpFp. Conversely, the sites’ as postulated by Jauhar (1975) can be discounted. number of chiasmata per PMC is higher in the LmFpFp than High MI pairing rate in the hybrids results in high re- in the LmLmFp hybrid (Jauhar, 1975). The author explained combination rate of individual chromosomes and entire ge- the preferential pairing of Festuca chromosomes by their nomes. These hybrid genomes in the Lolium – Festuca com- larger size and a higher number of ‘pairing recognition sites’. plex appear to be very plastic. The parental genomes recom- The presence of two doses of Festuca chromosomes would bine readily in the early hybrid generations; chromosome then leave the Lolium chromosomes at a synaptic disadvan- elimination and aneuploidy are frequent. It has been report- tage. However, the average DNA content per chromosome ed that in addition to pairing of homologous and homoeol- in both species is quite similar (0.293 pg DNA in L. multi- ogous chromosomes, pairing of non-hom(oe)ologous chro- florum versus 0.319 pg DNA in F. pratensis per non-repli- mosomes also takes place and apparently leads to recombi- cated chromosome; Table 1 ) suggesting a minor difference nation. For example, such non-homoeologous chromosome in chromosome size. Thus, the reason for preferential pair- pairing was observed in triploid hybrids L. multiflorum ! ing of Festuca chromosomes in triploid L. multiflorum ! F. donax (LmLmFd) and L. multiflorum ! L. temulentum F. pratensis hybrids remains obscure. Even with high pair- (Morgan, 1990; Thomas, 1995). Cao et al. (2000) found one ing, there are clear differences in the affinity of individual quadrivalent in a triploid F1 L. perenne ! F. mairei hybrid genomes. In triploid hybrids L. multiflorum ! F. pratensis (LpFmFm), which indicated non-hom(oe)ologous pairing. (LmLmFp) and L. multiflorum ! F. glaucescens (LmLmFg) However, duplications or other structural differences be- it was obvious that the L. multiflorum genome had a higher tween the Festuca and Lolium genomes could potentially ac- affinity to that of F. pratensis than to the F. glaucescens , as count for quadrivalent formation in F1. Using GISH and the frequency of trivalents was almost twice as high in FISH on mitotic chromosome spreads, Kopecký et al. (2006) LmLmFp as in LmLmFg hybrid and conversely, there was a demonstrated recombination between non-homoeologous higher frequency of univalents in LmLmFg than in LmLm- chromosomes in six out of 28 euploid individuals of L. mul- Fp (Morgan et al., 2001; Kosmala et al., 2006b). tiflorum ! F. pratensis cultivars. Almost 4% of all recombi- Metaphase I pairing preferences of chromosomes in wide nation events appeared to be of non-homoeologous nature. hybrids have been used for long as a measure of their phylo- Preferential pairing was found not only between ho- genetic affinity, to the point where numerical methods of moeologous chromosomes of different genomes, but also quantifying such affinity have been developed, for triploids, among non-homologous chromosomes within one genome. tetraploids and various aneuploid combinations (Alonso The results of Thomas et al. (1988) obtained in triploid hy- and Kimber, 1981; Kimber and Alonso, 1981; Kimber, 1983). brids of L. multiflorum ! L. perenne (LmLmLp) indicated Based on these methods genome affinities in various genera that some chromosomes may have a greater potential to pair have been reexamined and at times revised (Kimber, 1983). preferentially than others. By analyzing meiosis in triploid The same methods have been used in grass hybrids. As it was L. multiflorum plants, Thomas (1995) determined chromo- stated above, wide hybrids in the Lolium – Festuca complex some 2 to have fewer partner exchanges than other chromo- are unique in the capacity of their chromosomes to pair with somes. Armstead et al. (2001) explained different recombi- homologous or homologous-like frequencies despite obvi- nation frequencies of two different chromosomes of F. pra- ous major differences at the DNA level. It has, therefore, tensis substituted into diploid L. perenne by differences in been of interest to grass cytogeneticists if in wide hybrids chromosome affinity, different innate recombination fre- any preference of homologous over homoeologous pairing quencies, by the sampling effect, or environmental and/or existed. In the absence of better tools at that time, this ques- genetic factors. tion has been examined in an interspecific Lolium hybrid by The reports on non-homologous pairing always have to a segregation study with the conclusion that a strong prefer- be treated with some degree of caution. Crossing over that ence for homologous over homoeologous pairing existed is a prerequisite for chiasmate MI pairing requires a certain

376 Cytogenet Genome Res 120:370–383 (2008) minimum DNA homology (Shen and Huang, 1986); hence, vor of the Lolium genome (ranging from 54.8% in cv. ‘Felo- the critical issue is how ‘chromosome homology’ is defined. pa’ to 62.8% in cv. ‘Agula’) and a fairly large number of ho- Chromosome identification of ‘homologues’ in all grass moeologous crossovers took place, scattered all over chro- wide hybrids studied so far was cytological and not genetic. mosome length. Soon afterwards, Canter et al. (1999) In cytological identification there is an implicit assumption reported on the genomic constitution of cv. ‘Prior’ ( L. pe- of the absence of polymorphism, or heterozygosity, for renne ! F. pratensis ). Again, a prevalence of Lolium genome structural chromosome differences within a parental spe- over that of Festuca was evident (62.1% of Lolium to 37.9% cies and the absence of structural rearrangements. The of Festuca ) with at least 1.61 translocation breakpoints per above mentioned example of seven, but not six or eight, translocated chromosome. Similar observations were made chromosomes with identifiable 45S rDNA sites in L. perenne by Kopecky et al. (2006) with the only difference being the (Thomas et al., 1996) suggests that either the detection tech- number of complete Lolium chromosomes. Such a differ- niques may be at their resolution limits or that polymor- ence could be a result of a drift in genome balance in suc- phism for chromosome structure exists. If there is polymor- cessive generations of Festulolium hybrids in favor of the phism for the presence of such structural features of chro- dominant Lolium genome as described by Zwierzykowski et mosomes as the NOR sites, and these sites are used for al. (2006), or a sampling error, especially with a miniscule chromosome identification purposes in FISH, it may appear sample of Canter et al. (1999). as if non-homologous chromosomes were involved in chi- Our more recent studies (Kopecký et al., 2005a, 2006) asmate pairing even though, in fact, no such phenomenon revealed larger differences in the proportions of parental takes place. chromatin present in various types of Festulolium cultivars. Even a cursory examination of published photographs as While cultivars originating from the L. multiflorum ! F. well as studies on individual species (Karp and Jones, 1983) pratensis , L. perenne ! F. pratensis, L. multiflorum ! F. indicate that the species of the Lolium – Festuca complex glaucescens hybrids consistently display rather equal pro- have a strong tendency toward distal chiasmata. However, portions of parental genomes, in remaining cultivars the regions of chromosomes blocked from crossing over do not predominance of the L. multiflorum genome was evident appear to exist and individual events occur along the entire (Fig. 3). An almost equal proportion of parental genomes in chromosome length, but with varying frequencies. Hot- hybrid cultivars was also demonstrated by Momotaz et al. spots of recombination were described in diploid L. perenne (2004) using SSR markers in eight Lolium ! F. pratensis ! F. pratensis (Armstead et al., 2001) and in triploid L. mul- cultivars. On the other hand, all cultivars originating from tiflorum ! F. pratensis (Zwierzykowski et al., 1999a) hy- the L. multiflorum ! F. arundinacea hybrids show a strong brids. Such hot-spots in F. pratensis chromosome 3 were prevalence of one genome over the other, depending on the found at 18 pu (physical unit; complete length of chromo- direction of the backcross needed for fertility restoration some is 100 pu) from the telomere of the NOR arm and at (Buckner et al., 1977). Derivatives of the L. multiflorum ! 12 pu from the telomere of the non-NOR-bearing arm (King F. glaucescens hybrids display the lowest numbers of recom- et al., 2002). These hot-spots seem to be gene rich. The low- bined chromosomes among all Festulolium hybrids. As est level of recombination was detected in a region between similar numbers of generations are involved in the cultivar 45 and 75 pu, which also carries a NOR (King et al., 2002). development regardless of the nature of the starting hybrid, Similar results were reported by Zwierzykowski et al. (1999a) lower levels of recombination (as measured by the number who detected the lowest frequency of recombination around of recombined chromosomes) can be explained by a stron- the centromere and telomeres. ger affinity of the L. multiflorum genome to that of F. pratensis. A closer relationship of L. multiflorum with F. pratensis rather than to F. glaucescens was previously described Genomic constitution of the Festulolium cultivars by Humphreys and Ghesquière (1994) based on PGI/2 allele segregation and by Kopecký et al. (2005a) based on the fre- Until recently, the genomes of Festuloliums were only quency of the recombination events involving the parental assumed to be hybrid. This was based on the well known genomes in L. multiflorum ! F. arundinacea hybrids. fact of high meiotic affinity of the parental genomes and the In some Festulolium cultivars (i.e. ‘Kemal’, ‘Duo’ and presence of various characteristics from both parents. The ‘Matrix’), GISH failed to detect any Festuca chromatin and first report based on GISH screening that provides a fairly only a few segments were found in several plants of Festu- detailed picture of the genome constitution of a Festulolium lolium cv. ‘Bečva’ and ‘Lofa’ (Kopecký et al., 2005a, 2006). dates back only a decade (Zwierzykowski et al., 1998b). All As these cultivars display morphological characters inter- cultivars in that study were derived from L. multiflorum ! mediate to the parents, it may well be that the intergenomic F. pratensis and F. pratensis ! L. multiflorum hybrid (since introgressions in these cultivars are below the resolution publication, the materials have been registered under culti- limit of GISH. This limit was estimated to range from 5 kb var names ‘FL-I’ = cv. ‘Agula’, ‘FL-II’ = cv. ‘Felopa’, ‘FL-V’ = (Kosmala et al., 2003) to several Mb (Lukaszewski et al., cv. ‘Sulino’, and ‘LF-6’ = cv. ‘Rakopan’; Z. Zwierzykowski, 2005), and may depend to some extent on the location of the pers. comm.). The analyzed samples were not large, with introgressed segments relative to the origin of the probe only four to eight plants per cultivar tested, but it was clear used in screening. If indeed intergeneric introgressions are that the proportions of parental genomes were lightly in fa- present in these cultivars but are below the resolution limit

Cytogenet Genome Res 120:370–383 (2008) 377 of GISH, analyses using appropriate molecular markers may of L. perenne ! F. pratensis hybrids (Kopecky et al., 2005b). be needed to confirm their hybrid status. However, a reliable Reduced vigor and plant size of diploids over tetraploids are marker system suitable for large-scale screening of Festulo- obvious advantages in this kind of breeding; the tetraploid lium at a respectable resolution is currently unavailable. So, level of the starting hybrids offered fertility of early genera- despite the resolution limits, GISH is still the method of tions (diploid hybrids are sterile, see above) and ample op- choice in the characterization of Festulolium genomes and portunity for intergenomic recombination. Over 100 an- detection of introgressed segments. As such it could be used drogenic progeny from L. multiflorum ! F. pratensis and in the registration process of Festulolium cultivars. F. pratensis ! L. perenne hybrids were analyzed by GISH (Kopecky et al., 2005b). Apart from detailed characterization of the breeding material, GISH was found invaluable to Gamete-derived progenies of Festuca ! Lolium determine the origin of each individual androgenic proge- hybrids ny, a perennial problem in androgenesis. Festulolium, by reacting so well with GISH and with its Pre- and post-zygotic selection is a concern in distant high recombination rates of parental genomes offers a hybridization efforts as it may dramatically limit the use of unique system for testing the origin of each regenerated hybrid germplasms by eliminating some or all potentially plant. If the chromosome constitution of the donor plant is useful gene combinations. These problems can be circum- known, the origin of every progeny can be established with vented by making use of gamete-derived haploid progenies. ease. A plant with the same chromosome number and con- Another advantage of the approach is quick detection and stitution as the donor plant was obviously regenerated from assessment of recessive alleles, a feat difficult to accomplish the somatic tissue of the donor. A plant with the same chro- in conventional breeding. Since the first successful in vitro mosome number as parent and pairs of chromosomes paint- haploid generation from cultured anthers (Nitzsche, 1970), ed by GISH in the same pattern is an obvious microspore- androgenesis has been used frequently for haploid and di- derived progeny that spontaneously doubled its chromo- haploid production in Lolium – Festuca complex (i.e. Rose some number during early development (Fig. 4 ). A plant et al., 1987; Zwierzykowski et al., 1998a, 1999b, Zare et al., with half the chromosome number of the donor is a haploid 1999; Lesniewska et al., 2001; Kopecky et al., 2005b). derived from the microspore. Any new translocation break- Early on, the regenerants were characterized only by points appearing among progeny that were not present in their chromosome numbers. Once GISH was in place, it the donor are the result of a single round of crossing over. provided a much more detailed picture not only as to the Thus, Kopecky et al. (2005b) were able to conclude that, general ploidy levels and possible aneuploidy, but also very among the screened progeny, all 27- and 28-chromosome detailed descriptions of the genomic and chromosome plants were products of spontaneous chromosome dou- structure of individual plants. In an early attempt, Hum- bling; 27-chromosome plants must have been produced by phreys et al. (1998) analyzed two androgenic regenerants somatic loss of a chromosome following the doubling event. from L. multiflorum ! F. arundinacea hybrids (2n = 5x = Progeny with ϳ 14 chromosomes were obvious haploids 35; LmLmFpFgFg), with high drought and freezing resis- generated from microspores. There was not a single in- tance. Both regenerants had 21 chromosomes with six to stance of a regenerant originating from the somatic anther eight chromosomes from each of the parental genomes. tissue. However, several plants were identified with such Crossover events involving all three parental genomes were strange patterns of probe hybridization that their origin detected but with only two individuals tested, they could could not be explained. not be quantified. In a similar study and using the same type Haploids (mostly 2n = 3x = 21) from anther culture of of starting material, Zwierzykowski et al. (1998a) observed pentaploid L. multiflorum ! F. arundinacea hybrids are at least one recombination event involving L. multiflorum mostly sterile, which hampers their use in breeding (Hum- and F. arundinacea genomes per regenerated plant. The dis- phreys et al., 1998). In contrast, Lesniewska et al. (2001) re- tribution of translocation breakpoints was interesting, in ported partial fertility of haploids produced from tetraploid that they appeared to be absent from the terminal and cen- F. pratensis ! L. multiflorum hybrids. Some fertility was tromeric portions of the arms and were frequent in intersti- also present among the haploids (2n = 2x = 14) for turf tial regions with non-homogeneous distribution. However, breeding (Kopecky et al., 2005b). Even if all anthers appear the apparent absence of recombination events in the proxi- indehiscent there is sufficient female fertility to produce mal and distal regions could be due to the resolution limit some BC1 progeny by open pollination to fertile diploids. of GISH. The progeny in turn can be used to produce new cultivars GISH analyses of haploid progenies (here understood as with fixed desirable characteristics (such as an abiotic stress plants with half of the chromosome number of the donor tolerance) in homozygous condition by chromosome dou- plant; depending on the ploidy level of the donors these can bling and/or backcrossing with Lolium (or Festuca ) for the be polyhaploids) produced by androgenesis in vitro were introgression breeding strategy. Genotypes with desirable also described in F. pratensis ! L. multiflorum and L. pe- genes can then be intercrossed to produce populations iso- renne ! F. pratensis hybrids (Lesniewska et al., 2001; Guo genic for the characters of interest. The homozygosity can et al., 2005). We used androgenesis to reduce the chromo- be also achieved by the intercross of androgenic plants with some numbers from tetraploid to diploid for turf breeding targeted genes (Thomas et al., 2003).

378 Cytogenet Genome Res 120:370–383 (2008) Physical mapping of introgressed genomic segments F. mairei were identified also by GISH (Chen and Sleper, 1999; Cao et al., 2000, 2003). A combination of high homoeologous recombination in Freezing tolerance genes were successfully introgressed the Festuca ! Lolium hybrids with the ability to discrimi- into Lolium species from F. pratensis and F. arundinacea . nate parental chromatin by GISH provides for a very unique Grønnerød et al. (2004) and Guo et al. (2005) reported on and powerful system to study introgressions from one spe- transfer of freezing tolerance genes from F. pratensis chro- cies to another. This approach has been used extensively in mosome 3 into L. perenne. Kosmala et al. (2006a) used FISH the development of ryegrasses resistant to abiotic and biotic and GISH to identify a segment of F. pratensis chromosome stresses. The first use of GISH in grasses focused on map- 2, which was introgressed into L. multiflorum and was lo- ping a segment carrying mutation of sid (senescence in- cated terminally on a non-NOR arm in two of the three duced degradation), which was introgressed from F. praten- most freezing-tolerant plants. Chromosome 4 of F. pratensis sis into L. multiflorum (Thomas et al., 1994). Mutation of with the translocated terminal segments of L. multiflorum the sid gene ( sidy ) hampers the enzymatic step which de- on both arms was detected in the third plant. stroys the green color by opening the tetrapyrrole macro- Apart from abiotic stresses, ryegrass cultivars are sus- cycle, for what the genotypes are called staygreen . A chro- ceptible to several pathogens (rusts, bacterial wilt, and hel- mosome segment carrying this gene was located at the dis- minthosporium) to which resistance exists among fescues. tal end of a pair of Festuca chromosomes. Subsequently it GISH was used to follow the transfer of chromatin carry- was shown that the segment of Festuca chromatin carrying ing pathogen resistance, albeit less frequently. For exam- the sid y gene was transferred to the short arm of chromo- ple, Roderick et al. (2003) developed introgression lines of some 6 of Lolium (Thomas et al., 1997a). Soon after, other L. multiflorum with F. pratensis segments responsible for studies described the localization of genes for resistance crown rust (Puccinia coronata) resistance. The intro- against abiotic stress (Table 4 ). gressed segment of F. pratensis responsible for the resis- The most heat- and drought-tolerant grass species are F. tance was identified using GISH to occupy the terminal glaucescens and, especially, F. arundinacea and F. mairei . region of L. multiflorum chromosome 5 (Armstead et al., This is not surprising as F. mairei and some ecotypes of F. 2006). arundinacea are adapted to the harsh conditions of the A complex nature of stress-tolerance traits seems to be mountains of Northern Africa (Humphreys et al., 1997). On reflected by physical position of the Festuca genome seg- the other hand, the most freezing tolerance and winter har- ments introgressed into Lolium. In derivatives of F. arundi- diness is known to exist in F. pratensis, whose range of dis- nacea – L. multiflorum, the introgressed segment carrying tribution reaches beyond the Arctic Circle. genes for drought resistance (originating from the F. praten- Drought resistance was introduced from fescue into rye- sis genome) was located in the middle of the non-NOR arm grass in several breeding programs (Humphreys and of Lolium chromosome 3 (Humphreys and Pašakinskiene, Thomas, 1993; Humphreys and Pašakinskiene, 1996; Les- 1996). Drought resistance in an F. glaucescens – L. multiflo- niewska et al., 2001; Kosmala et al., 2006a). Humphreys and rum introgression was associated with a segment of an F. Thomas (1993) introgressed quantitative trait loci (QTLs) glaucescens chromosome located at the terminus of the for drought tolerance from F. arundinacea to L. multiflo- NOR arm of Lolium chromosome 3 (Humphreys et al., rum . GISH with two diploid plants revealed a single seg- 2005). A comparable situation exists in freezing-tolerance ment of F. arundinacea chromatin located interstitially on traits, where at least two different chromosomes of F. pra- the long arm of the L. multiflorum chromosome 2 (Hum- tensis introgressed into L. multiflorum appear to affect the phreys and Pašakinskiene, 1996). Based on GISH with the survival rate under freezing conditions (Grønnerød et al., F. pratensis genomic DNA probe, the authors proposed that 2004; Kosmala et al., 2006a). The variability in chromosom- segment originated from the F. pratensis genome present in al location of loci affecting the same trait could be explained F. arundinacea . (Note that according to current chromo- by a high number of genes affecting the plant’s response to some nomenclature, the chromosome 2 of Humphreys and stress that are located on different chromosomes, differenc- Pašakinskiene (1996) is now chromosome 3 (Jones et al., es in the genetic make up of recipient lines (that is, presence 2002)). of a different set of loci that combined may produce the Introgression of drought-tolerance gene(s) from F. same phenotypic effect) that creates a situation where addi- glaucescens to L. multiflorum was reported by Morgan et al. tion of a different locus from the donor species creates a (2001) and Humphreys et al. (2005). According to GISH, the measurable phenotypic effect, and perhaps by non-ho- transferred drought tolerance gene(s) were located on chro- moeologous recombination as discussed earlier. As the mosome 3 of L. multiflorum. This was also confirmed by stress resistance genes are known to belong to multigene AFLP and STS markers. North African xeromorphic tetra- families and be quantitatively inherited, we are inclined to ploid F. mairei is another potential source of drought toler- favor the first two possibilities. Regardless of the conflicting ance genes. In spite of the larger phylogenetic distance of F. results from different studies, it is obvious that both GISH mairei to Lolium as compared with other fescues (Catalan and FISH enable the identification of the segments carrying et al., 2004), extensive homoeologous pairing necessary for desirable genes of one species when introgressed into an- introgression breeding was observed and L. perenne plants other species ( Table 4 ). with the segments carrying genes of drought tolerance from

Cytogenet Genome Res 120:370–383 (2008) 379 Table 4. Introgression of various traits between Festuca and Lolium

Introgressed Genes introgressed Genes introgressed Method of detection Reference tolerance/resistance from to sid – delayed senescence F. pratensis L. multiflorum GISH, AFLP Thomas et al. (1994, 1997) Drought resistance F. arundinacea L. multiflorum GISH/FISH Humphreys and Pašakinskiene (1996) Drought resistance F. mairei L. perenne GISH, RFLP Chen et al. (1995) GISH, RFLP Chen and Sleper (1999) GISH, RFLP Cao et al. (2000, 2003) Drought resistance F. glaucescens L. multiflorum GISH/FISH, AFLP, STS Humphreys et al. (2005) Drought- and cold-tolerance F. arundinacea L. multiflorum GISH Humphreys et al. (1997) GISH, AFLP Skibinska et al. (2002) GISH, AFLP Kosmala et al. (2003) Drought- and cold-tolerance F. pratensis L. multiflorum GISH Lesniewska et al. (2001) GISH, AFLP Skibinska et al. (2002) GISH, AFLP Kosmala et al. (2003) Freezing tolerance F. pratensis L. multiflorum GISH/FISH Kosmala et al. (2006a) Freezing tolerance F. pratensis L. perenne GISH Grønnerød et al. (2004) GISH Guo et al. (2005) Crown rust F. pratensis L. multiflorum GISH Roderick et al. (2003) AFLP, STS Armstead et al. (2006)

Analyses of individual chromosomes section of individual chromosomes into smaller and smaller bins, depending on research objectives. The usefulness of A detailed analysis and manipulation of genomes is fa- introgression mapping is that it provides a number of mark- cilitated by the construction of genetic and physical maps of ers for specific regions of the F. pratensis genome. Using this individual chromosomes. The great potential of Festuca ! strategy, Armstead et al. (2006) identified a candidate gene Lolium hybrids with their high frequency of homoeologous responsible for the stay-green phenotype. recombination and easy discrimination of parental chro- We have initiated a project to develop a set of disomic matin in hybrids by GISH created a new strategy for map- introgressions of F. pratensis chromosomes into tetraploid ping, called ‘introgression mapping’ (King et al., 1998). The L. multiflorum ‘Mitos’. Whether this will provide a more authors selected three complete chromosome substitution stable research environment than monosomic introgres- lines from BC1 progeny of L. perenne ! F. pratensis plus sions into a diploid remains to be seen. The disomic condi- two additional chromosome arm substitutions. The avail- tion of the introgressed F. pratensis chromosomes does not ability of a full set of single chromosome substitution lines necessarily guarantee stability from generation to genera- would permit comprehensive studies of the genetic control tion because of their inherent ability to engage in homoeol- of desirable agronomic characters and the localization of ogous pairing, hence recombination. Nevertheless, it ap- the genes on genetic and physical maps. Such a linkage be- pears likely that each of the seven F. pratensis chromosomes tween physical and genetic maps in grasses has been dem- is already present as monosomic introgression in L. multi- onstrated by Armstead et al. (2001), who developed a pow- florum; at least three are already disomic (Kopecký et al., erful approach for studying genetic recombination and high unpublished; Fig. 5). The numbers are uncertain because, resolution mapping for dissection of specific characters. In with few exceptions of the rDNA-bearing chromosomes, their study they make use of two monosomic substitutions the Lolium – Festuca complex still lacks a reliable system of of F. pratensis chromosomes into diploid L. perenne and chromosome identification and only testcrosses and obser- their recombinant derivatives, and demonstrate that the re- vation of the meiotic pairing reliably confirm the identity of lationships between physical and genetic distances vary individual chromosomes. One possibility of resolving this along the length of a chromosome arm and the presence of bottleneck could be to select chromosome specific BAC recombination hot-spots. Similar results were reported for (Bacterial Artificial Chromosome) clones to be used as rye (Wanous and Gustafson, 1995) and barley (Laurie et al., probes for FISH. With this in mind, we have constructed a 1993). genomic BAC library from F. pratensis ‘Laura’ (Kopecký et Recently, King et al. (2007) described the development of al., unpublished). Other genomic BAC libraries were con- a full set of monosomic substitution lines L. perenne/ F. pra- structed from L. perenne and F. pratensis (Donnison et al., tensis (13 chromosomes of L. perenne plus one chromosome 2005; Farrar et al., 2007). So far, new cytogenetic markers of F. pratensis). Each of the monosomic substitution lines for three chromosomes of F. pratensis seem to be working, was backcrossed to L. perenne to develop a set of single chro- leaving only two fescue chromosomes that cannot be dis- mosome recombinant lines. The sets of such lines could be criminated from each other (Fig. 6 ). of infinite size, thus providing an excellent tool for the dis-

380 Cytogenet Genome Res 120:370–383 (2008) Table 5. The origin of names of some Festulolium cultivars

Name Hybrid Country of origin Cultivar name based on

Achilles LmFp CZE Greek Hero Perseus LmFp CZE Greek Hero Perun LmFp CZE Slavic God Agula FpLm POL Name of granddaughter of breeder Wojciech Jokś Sulino FpLm POL Dedicated to late Prof. Stanisław Sulinowski Felina LmFaFa CZE Festuca – Lolium Hykor LmFaFa CZE Kostrava rakosovita (tall fescue in Czech) Korina LmFaFa CZE Kostrava rakosovita (tall fescue in Czech) Lesana LmFaFa CZE Les (a forest in Czech) Fojtan LmFaFa CZE Dedicated to late Dr. Antonin Fojtik, a famous Czech grass breeder Bečva LmLmFa CZE River running close to Breeding station Hladke Zivotice Lofa LmLmFa CZE Lolium – Festuca

Conclusions and future perspectives In addition to attractive use in agriculture, recreation, sports and urban environment, Festuca ! Lolium hybrids The global importance and the potential of Festulolium provide a unique system to dissect the genomes of related hybrids are best demonstrated by the rising interest of farmers species. One may foresee the development of a complete se- and by a steady increase of the area on which Festulolium is ries of disomic chromosome substitution lines and intro- cultivated. The impact of Festulolium hybrids and progress in gression lines that will facilitate development of genetic and their breeding contrasts with a rather poor knowledge of their physical maps, development of markers from marker assist- genome structure and behavior. Consequently, their breeding ed selection and for gene cloning in both species. While there remains empirical. While this strategy may be sustainable for is nothing unique about such sets of alien introgressions as the time being, further improvement of Festulolium may re- they are common in cultivated crops, what is unique about quire advanced tools of genomics and biotechnology. the Lolium – Festuca complex is the ability of the homoeologs Effective application of advanced tools requires im- to pair and recombine with frequencies approaching those proved genome knowledge at the chromosome level. Quite of homologs. At the same time, the parental genomes can be surprisingly, even today it is not possible to identify all chro- easily discriminated by GISH. This means that every single mosomes during mitosis or meiosis, and let alone in the in- recombination event can be visualized and as such, easily terphase in any of the Lolium or Festuca species, including perpetuated ad infinitum for detailed analyses, as new tools those used in breeding. This obviously impairs the efforts arrive. This is an unprecedented system, not available in any to understand karyotype evolution and to follow the behav- other species. It may be therefore expected that cytogenetics ior of specific chromosomes in hybrid progenies. Although of Festuloliums will remain in the centre of attention of grass some general trends in the genome evolution in Festulolium breeders as well as scientists in fundamental research. have been observed, the understanding of the behavior of the parental genomes and their interactions in the hybrids is still fragmentary at best. Yet, considering the knowledge The story behind – the origin of Festulolium cultivar gained with GISH and FISH with the new cytogenetic mark- names ers, one may predict that detailed analysis of chromosome behavior will be possible in the near future. Obviously, this At times it is intriguing where the specific names come will open avenues for better characterization of breeding from. Although the grasses within the Lolium – Festuca com- materials and hopefully lead to the development of cost-ef- plex have been used for the benefit of humankind, their first fective breeding of improved Festulolium. designations were rather abusive. Festuca is a Latin word giv- Cytogenetic methods are generally considered time con- en by Pliny to a weed growing in barley. Similarly, Lolium was suming and tedious, and their throughput cannot cope with a name given by Virgil to a troublesome weed (http://delta- the numbers of breeding lines produced every year. Still, giv- intkey.com/grass/www/ident.htm). On the other hand, en the simplicity of the current GISH and FISH (Masoudi- breeders want to show the potential of their new cultivars Nejad et al., 2002) it most certainly could easily compete with and so they tend to choose boastful names (Table 5). any non-automated DNA marker protocol both in throughput and the total cost. Proper characterization of the new hybrid cultivars may have serious implications in the cultivar Acknowledgements registration process. On the other hand, an automated high through-put system for determination of genomic constitu- We thank Prof. Zbigniew Zwierzykowski for critical reading and valuable comments. tion in hybrid lines will require the use of molecular markers coupled with platforms for mass parallel screening.

Cytogenet Genome Res 120:370–383 (2008) 381 Humphreys MW, Zare AG, Pašakinskiene I, Thom- as H, Rogers WJ, Collin HA: Interspecific ge- Alonso LC, Kimber G: The analysis of meiosis in Dubcovsky J, Dvorak J: Ribosomal-RNA multigene nomic rearrangements in androgenic plants de- hybrids. II. Triploid hybrids. Can J Genet Cytol loci – nomads of the Triticeae genomes. Genet- rived from a Lolium multiflorum ! Festuca

23: 221–234 (1981). ics 140: 1367–1377 (1995). arundinacea (2n = 5x = 35) hybrid. Heredity 80: Armstead IP, Bollard A, Morgan WG, et al: Genetic Dubcovsky J, Martinez AJ: Cytotaxonomy of the 78–82 (1998). and physical analysis of a single Festuca praten- Festuca spp from Patagonia. Canad J Botany- Humphreys J, Harper JA, Armstead IP, Humphreys sis chromosome segment substitution in Loli- Revue Canadienne de Botanique 70: 1134–1140 MW: Introgression-mapping of genes for um perenne . Chromosoma 110: 52–57 (2001). (1992). drought resistance transferred from Festuca Armstead IP, Harper JA, Turner LB, et al: Introgres- Evans GM, Davies EW: The genetics of meiotic arundinacea var. glaucescens into Lolium mul- sion of crown rust (Puccinia coronata) resistance chromosome pairing in Lolium temulentum ! tiflorum. Theor Appl Genet 110: 579–587 from meadow fescue (Festuca pratensis) into Lolium perenne tetraploids. Theor Appl Genet (2005). Italian ryegrass (Lolium multiflorum): genetic 71: 185–192 (1985). Jauhar PP: Chromosome relationships between Lo- mapping and identification of associated molec- Evans GM, Rees H, Snell CL, Sun S: The relation- lium and Festuca ( Gramineae ). Chromosoma

ular markers. Plant Pathol 55: 62–67 (2006). ship between nuclear DNA amount and the du- 52: 103–121 (1975). Bennett MD, Smith JB, Heslop-Harrison JS: Nucle- ration of the mitotic cycle. Chromosomes To- Jauhar PP: Cytogenetics of the Festuca - Lolium ar DNA amounts in angiosperms. Proc R Soc day 3: 24–31 (1972). complex. Relevance to Breeding. Monographs ! B-Biol Sci 216: 179–199 (1982). Farragher MA: On the occurrence of Festuca Lo- on Theor Appl Genet Vol. 18, pp 255 (Springer- Borrill M: Festuca L., in Stace CA (ed): Hybridiza- lium hybrids in the Irish flora. Ir Nat J 18: 208– Verlag, Berlin, 1993). tion and the Flora of the British Isles, pp 543– 211 (1975). Jones ES, Dupal MP, Dumsday JL, Hughes LJ, For- 547 (Academic Press, London 1975). Farrar K, Asp T, Lubberstedt T, et al: Construction ster JW: An SSR-based genetic linkage map for Bowman JG, Thomas H: Studies in Festuca . 8. Cy- of two Lolium perenne BAC libraries and iden- perennial ryegrass (Lolium perenne L.). Theor tological relationships between F. glaucescens tification of BACs containing candidate genes Appl Genet 105: 577–584 (2002). (2n = 28), F. mairei (2n = 28) and F. scoriosa for disease resistance and forage quality. Mol Karp A, Jones RN: Cytogenetics of Lolium perenne . (2n = 14). Z Pflanzenzuecht 76: 250–257 (1976). Breed 19: 15–23 (2007). Part 2. Chiasma distribution in inbred lines. Buckner RC, Burrus PB, Bush LP: Registration of Fojtík A: Methods of grass improvement used at the Theor Appl Genet 64: 137–145 (1983). Kenhy tall fescue. Crop Sci 17: 672–673 (1977). Plant Breeding Station Hladké Životice. Genet Kimber G: Genome analysis in the genus Triticum, Canter PH, Pašakinskiene I, Jones RN, Humphreys Pol 35A:25–31 (1994). in Proceedings of the 6th International Wheat MW: Chromosome substitutions and recombi- Fojtík A: Šlechtĕní a využití Festulolium v České Re- Genetics Symposium Kyoto, Japan, pp 23–28 ! nation in the amphiploid Lolium perenne publice. Proc National Conference Festuloli - (1983). Festuca pratensis cv Prior (2n = 4x = 28). Theor um – Achievements and Perspectives, Poznań, Kimber G, Alonso LC: The analysis of meiosis in Appl Genet 98: 809–814 (1999). Poland, November 1998, pp 27–32 (in Czech) hybrids. II. Tetraploid hybrids. Can J Genet Cy- Cao MS, Sleper DA, Dong FG, Jiang JM: Genomic (1998). tol 23: 235–254 (1981). in situ hybridization (GISH) reveals high chro- Greilhuber J, Doležel J, Lysák MA, Bennett MD: King IP, Morgan WG, Armstead IP, et al: Introgres- mosome pairing affinity between Lolium pe- The origin, evolution and proposed stabiliza- sion mapping in the grasses. I. Introgression of renne and Festuca mairei . Genome 43: 398–403 tion of the terms ‘genome size’ and ‘C-value’ to Festuca pratensis chromosomes and chromo- (2000). describe nuclear DNA contents. Ann Bot 95: some segments into Lolium perenne . Heredity Cao M, Bughrara SS, Sleper DA: Cytogenetic analy- 255–260 (2005). 81: 462–467 (1998). sis of Festuca species and amphiploids between Grønnerød S, Fjellheim S, Humphreys MW, et al: King J, Armstead IP, Donnison IS, et al: Physical Festuca mairei and Lolium perenne . Crop Sci 43: Application of AFLP and GISH techniques for and genetic mapping in the grasses Lolium pe- 1659–1662 (2003). identification of Festuca chromosome seg- renne and Festuca pratensis. Genetics 161: 315– Carnide V, Orellana J, Ribeiro MAMD: Nucleolar ments conferring winter hardiness in a Lolium 324 (2002). ! organizer activity in Lolium and Festuca .1. Lo- perenne Festuca pratensis population, in King J, Armstead IP, Donnison IS, et al: Introgres- lium multiflorum, Festuca arundinacea and Hopkins A, Wang ZY, Mian R, Sladge M, Back- sion mapping in the grasses. Chromosome Res

Lolium-Festuca hybrids. Heredity 56: 311–317 er RE (eds): Molecular Breeding of Forage and 15: 105–113 (2007). (1986). Turf, pp 81–86 (Kluwer Academic Publishers, Kopecký D, Lukaszewski AJ, Dolezel J: Genomic Catalan P, Torrecilla P, Rodriguez JAL, Olmstead Dordrecht 2004). constitution of Festulolium cultivars released RG: Phylogeny of the festucoid grasses of sub- Guo YD, Mizukami Y, Yamada T: Genetic charac- in the Czech Republic. Plant Breed 124: 454– tribe Loliinae and allies ( Poeae, Pooideae ) in- terization of androgenic progeny derived from 458 (2005a). ! ferred from ITS and trnL-F sequences. Mol Lolium perenne Festuca pratensis cultivars. Kopecky D, Lukaszewski AJ, Gibeault V: Reduction Phylogenet Evol 31: 517–541 (2004). New Phytol 166: 455–464 (2005). of ploidy level by androgenesis in intergeneric Ceccarelli M, Falistocco E, Cionini PG: Variation of Hackel E: Monographia Festucarum Europearum Lolium-Festuca hybrids for turf grass breeding. genome size and organization within hexaploid (Theodor Fischer, Kassel et Berlin 1882). Crop Sci 45: 274–281 (2005b). Festuca arundinacea . Theor Appl Genet 83: Harper JA, Thomas ID, Lovatt JA, Thomas HM: Kopecký D, Loureiro J, Zwierzykowski Z, Ghesqui- 273–278 (1992). Physical mapping of rDNA sites in possible dip- ere M, Dolezel J: Genome constitution and evo- Černoch V, Našinec I, Šrámek P: Share of grassland loid progenitors of polyploid Festuca species. lution in Lolium ! Festuca hybrid cultivars

on landscape forming in the Czech Republic. Plant Syst Evol 245: 163–168 (2004). (Festulolium). Theor Appl Genet 113: 731–742 Czech J Genet Plant Breed 39: 158–162 (2003). Humphreys MW, Ghesquiere M: Assessing success (2006). Chen C, Sleper DA: FISH and RFLP marker-assist- in gene-transfer between Lolium multiflorum Kosmala A, Skibinska M, Zwierzykowski Z, Hum- ed introgression of Festuca mairei chromo- and Festuca arundinacea. Euphytica 77: 283– phreys MW, Rapacz M, Joks W: Introgression somes into Lolium perenne . Crop Sci 39: 1676– 289 (1994). of genes for abiotic stress resistance from Fes- 1679 (1999). Humphreys MW, Pašakinskiene I: Chromosome tuca pratensis and F. arundinacea into Lolium Chen C, Sleper DA, West CP: RFLP and cytogenet- painting to locate genes for drought resistance multiflorum germplasm. Vortr Pflanzenzüchtg ic analyses of hybrids between Festuca mairei transferred from Festuca arundinacea into Lo- 59: 225–231 (2003). and Lolium perenne . Crop Sci 35: 720–725 lium multiflorum. Heredity 77: 530–534 (1996). Kosmala A, Zwierzykowski Z, Gasior D, Rapacz (1995). Humphreys MW, Thomas H: Improved drought re- M, Zwierzykowska E, Humphreys MW: GISH/ Clayton WD, Renvoize SA: Genera graminum: sistance in introgression lines derived from Lo- FISH mapping of genes for freezing toler- ! Grasses of the World. Kew Bull, Add Ser 13: 1– lium multiflorum Festuca arundinacea hy- ance transferred from Festuca pratensis to

389 (1986). brids. Plant Breed 111: 155–161 (1993). Lolium multiflorum . Heredity 96: 243–251 D a r b y s h i r e S J : R e a l i g n m e n t o f Festuca subgenus Humphreys MW, Thomas HM, Morgan WG, et al: (2006a). Schedonorus with the genus Lolium (Poaceae ). Discriminating the ancestral progenitors of Kosmala A, Zwierzykowska E, Zwierzykowski Z: Novon 3: 239–343 (1993). hexaploid Festuca arundinacea using genomic Chromosome pairing in triploid intergeneric Dijkstra J, de Vos ALF: Meiotic doubling of chro- in situ hybridization. Heredity 75: 171–174 hybrids of Festuca pratensis with Lolium multi-

mosome number in Festulolium. Euphytica 24: (1995). florum , revealed by GISH. J Appl Genet 47: 215– 743–749 (1975). Humpherys MW, Thomas HM, Harper J, et al: Dis- 220 (2006b). Donnison IS, O’Sullivan DM, Thomas A, et al: Con- secting drought- and cold-tolerance traits in struction of a Festuca pratensis BAC library for the Lolium-Festuca complex by introgression

map-based cloning in Festulolium substitution mapping. New Phytol 137: 55–60 (1997).

lines. Theor Appl Genet 110: 846–851 (2005).

382 Cytogenet Genome Res 120:370–383 (2008) Laurie DA, Pratchett N, Devos KM, Leitch IJ, Gale Olszewska MJ, Osiecka R: The relationship between Thomas HM, Harper JA, Meredith MR, et al: Com- MD: The distribution of RFLP markers on 2C DNA content, life-cycle type, systematic po- parison of ribosomal DNA sites in Lolium spe- chromosome 2(2H) of barley in relation to the sition, and the level of DNA endoreplication in cies by fluorescence in situ hybridization.

physical and genetic location of 5S rDNA. The- nuclei of parenchyma cells during growth and Chromosome Res 4: 486–490 (1996).

or Appl Genet 87: 177–183 (1993). differentiation of roots in some monocotyle- Thomas H, Evans C, Thomas HM, et al: Introgres-

Leggett JM, Markhand GS: The genomic identifica- donous species. Biochem Physiol Pflanzen 177: sion, tagging and expression of a leaf senes-

tion of some monosomics of Avena sativa L-Cv 319–336 (1982). cence gene in Festulolium. New Phytol 137: 29– Sun-Ii using genomic in situ hybridization. Ge- Pašakinskiene I, Anamthawat-Jonsson K, Hum- 34 (1997a).

nome 38: 747–751 (1995). phreys MW, Paplauskiene V, Jones RN: New Thomas HM, Harper JA, Meredith MR, Morgan Lesniewska A, Ponitka A, Slusarkiewicz-Jarzina A, molecular evidence on genome relationships WG, King IP: Physical mapping of ribosomal et al: Androgenesis from Festuca pratensis ! and chromosome identification in fescue ( Fes- DNA sites in Festuca arundinacea and related

Lolium multiflorum amphidiploid cultivars in tuca) and ryegrass (Lolium ). Heredity 81: 659– species by in situ hybridization. Genome 40: order to select and stabilize rare gene combina- 665 (1998). 406–410 (1997b).

tions for grass breeding. Heredity 86: 167–176 Roderick HW, Morgan WG, Harper JA, Thomas Thomas HM, Morgan WG, Humphreys MW: De- (2001). HM: Introgression of crown rust (Puccinia cor- signing grasses with a future – combining the Lewis EJ, Tyler BF, Chorlton KH: Annual Report of onata ) resistance from meadow fescue (Festuca attributes of Lolium and Festuca. Euphytica

the Welsh Plant Breeding Station 1972, pp 34– pratensis) into Italian ryegrass (Lolium multi- 133: 19–26 (2003). 37 (1973). florum) and physical mapping of the locus. He- Vaughan HE, Jamilena M, Rejon CR, Parker JS,

Loureiro J, Kopecký D, Castro S, Santos C, Silveira redity 91: 396–400 (2003). Garridoramos MA: Loss of nucleolar-organiz- P: Flow cytometric and cytogenetic analyses of Rose JB, Dunwell JM, Sunderland N: Anther cul- er regions during polyploid evolution in Scilla

Iberian Peninsula Festuca spp. Plant Syst Evol ture of Lolium temulentum , Festuca pratensis autumnalis . Heredity 71: 574–580 (1993). 269:89–105 (2007). and Lolium ! Festuca hybrids.1. Influence of Wang ZY, Ge YX: Invited review: Recent advances Lübberstedt T, Schejbel Andreasen B, Bach Holm pretreatment, culture medium and culture in- in genetic transformation of forage and turf

P: Development of ryegrass allele-specific cubation conditions on callus production and grasses. In Vitro Cell Dev Biol Plant 42: 1–18

(GRASP) markers for sustainable grassland im- differentiation. Ann Bot 60: 191–201 (1987). (2006). provement – A new EU framework V project. Schifino MT, Winge H: Karyotypes and nuclear Wanous MK, Gustafson JP: A genetic map of rye

Czech J Genet Plant Breed 39: 125–128 (2003). DNA content of species of the Briza complex chromosome 1R integrating RFLP and cyto-

Lukaszewski AJ, Lapinski B, Rybka K: Limitations and some other genera of Poeae (Gramineae ). genetic loci. Theor Appl Genet 91: 720–726

of in situ hybridization with total genomic Rev Brazilian Genet 6: 245–259 (1985). (1995). DNA in routine screening for alien introgres- Schubert I, Wobus U: In situ hybridization con- Zare AG, Humphreys MW, Rogers WJ, Collin HA: ! sions in wheat. Cytogenet Genome Res 109: firms jumping nucleolus organizing regions in Androgenesis from a Lolium multiflorum

373–377 (2005). Allium . Chromosoma 92: 143–148 (1985). Festuca arundinacea hybrid to generate ex- Malik CP, Thomas PT: Karyotypic studies in some Schwarzacher T: DNA, chromosomes, and in situ treme variation for freezing-tolerance. Plant

Lolium and Festuca . Caryologia 19: 137–196 hybridization. Genome 46: 953–962 (2003). Breed 118: 497–501 (1999). (1966). Schwarzacher T, Heslop-Harrison P: Practical in Zwierzykowski Z, Lukaszewski AJ, Lesniewska A, Masoudi-Nejad A, Nasuda S, McIntosh RA, Endo situ Hybridization (Bios Scientific Publishers Naganowska B: Genomic structure of andro- TR: Transfer of rye chromosome segments to Ltd, New York 2000). genic progeny of pentaploid hybrids, Festuca wheat by a gametocidal system. Chromosome Schwarzacher T, Leitch AR, Bennett MD, Heslop- arundinacea ! Lolium multiflorum . Plant

Res 10: 349–357 (2002). Harrison JS: In situ localization of parental ge- Breed 117: 457–462 (1998a).

Momotaz A, Forster JW, Yamada T: Identification nomes in a wide hybrid. Ann Bot 64: 315–324 Zwierzykowski Z, Tayyar R, Brunell M, Lukaszew- of cultivars and accessions of Lolium , Festuca (1989). ski AJ: Genome recombination in intergeneric and Festulolium hybrids through the detection Shen P, Huang HV: Homologous recombination in hybrids between tetraploid Festuca pratensis

of simple sequence repeat polymorphism. Plant Escherichia coli: dependence on substrate and Lolium multiflorum . J Hered 89: 324–328

Breed 123: 370–376 (2004). length and homology. Genetics 112: 441–457 (1998b). Morgan WG: Chromosome pairing in triploid hy- (1986). Zwierzykowski Z, Lukaszewski AJ, Naganowska B, brids and amphidiploids involving Lolium and Skibińska M, Kosmala A, Humphreys MW, Zwier- Lesniewska A: The pattern of homoeologous

diploid Festuca species. Genome 33: 472–477 zykowski Z: Application of GISH and AFLP recombination in triploid hybrids of Lolium (1990). techniques for identification of Lolium-Festuca multiflorum with Festuca pratensis . Genome

Morgan WG, Thomas H, Lewis EJ: Cytogenetic introgressions. Cell Mol Biol Lett 7: 493–498 42: 720–726 (1999a). studies of hybrids between Festuca gigantea (2002). Zwierzykowski Z, Zwierzykowska E, Slusarkie- Vill and Lolium multiflorum Lam. Plant Breed Thomas HM: The Giemsa C-band karyotypes of six wicz-Jarzina A, Ponitka A: Regeneration of an-

101: 335–343 (1988). Lolium species. Heredity 46: 263–267 (1981). ther-derived plants from pentaploid hybrids of Morgan WG, King IP, Koch S, Harper JA, Thomas Thomas HM: Meiosis of triploid Lolium . 2. Dis- Festuca arundinacea ! Lolium multiflorum .

HM: Introgression of chromosomes of Festuca crepancies between the analyses of chromo- Euphytica 105: 191–195 (1999b). arundinacea var. glaucescens into Lolium mul- some configurations at metaphase-I in inverse Zwierzykowski Z, Kosmala A, Zwierzykowska E,

tiflorum revealed by genomic in situ hybridisa- autoallotriploid combinations. Heredity 75: Jones N, Joks W, Bocianowski J: Genome bal-

tion (GISH). Theor Appl Genet 103: 696–701 446–452 (1995). ance in six successive generations of the allotet- (2001). Thomas H, Humphreys MO: Progress and potential raploid Festuca pratensis ! Lolium perenne .

Nitzsche W: Obtaining haploid plants from Festu- of interspecific hybrids of Lolium and Festuca . Theor Appl Genet 113: 539–547 (2006). ca-Lolium crossbreeds. Naturwissenschaften J Agr Sci 117: 1–8 (1991). 57: 199–201 (1970). Thomas H, Morgan WG, Humphreys MW: The use of a triploid hybrid for introgression in Lolium species. Theor Appl Genet 76: 299–304 (1988). Thomas HM, Morgan WG, Meredith MR, Hum- phreys MW, Thomas H, Leggett JM: Identifica- tion of parental and recombined chromosomes in hybrid derivatives of Lolium multiflorum ! Festuca pratensis by genomic in situ hybridiza-

tion. Theor Appl Genet 88: 909–913 (1994).

Cytogenet Genome Res 120:370–383 (2008) 383