An accurate karyotype can incorporate physical measurements like total length and arm length ratios, but can also include landmarks such as heterochromatic knobs [46
], patterns of chromatin condensation [43
], and molecular features visualized by FISH [47
]. Chromosome identification is critical for cytological analyses, as well as subsequent studies in genomics, taxonomy, and the evolution of polyploidy, enabling an understanding of the relationship between visible landmarks and genetic or physical map features [48
]. To that end, the construction of a basic karyotype for switchgrass promises to facilitate genomic analyses. The somatic metaphase chromosomes of switchgrass are small, which may have limited examination of cytological features in earlier studies [11
]. With the use of sophisticated imaging and molecular techniques, we are now able to present the first comprehensive karyotype for switchgrass that quantitatively distinguishes each of the nine base chromosomes of this bioenergy crop.
Use of a dihaploid line of switchgrass (ALB280) significantly simplified the karyotyping process. Acetocarmine- and DAPI-stained chromosome spreads allowed for visual pairing of homoeologous chromosomes in ALB280 and produced a karyotype based on total and relative lengths as well as arm ratios. In our experiments, a single switchgrass root tip preparation yielded an average of 20 or more dividing cells (prophase to metaphase). Chromosome spreads often resulted in a high frequency of nuclei at the pro-metaphase stage of mitosis. Pro-metaphase chromosomes demonstrated a characteristic condensation pattern (CP) along their length, corresponding to the compactness of the chromatids, which was used to create a quantitative idiogram [50
]. This approach has also been useful in cytological analysis of Brassica
species, sugarcane, and rice [51
]. In conjunction with physical measurement data, CP data allowed us to unambiguously identify the small metacentric chromosomes of switchgrass.
Although morphological and CP data suggest a balanced karyotype in the dihaploid line ALB280 (2n
18), FISH data presented here indicate that the subgenomes have different repetitive DNA content at PviCentC and 5S rDNA loci. This finding is in general agreement with the highly differentiated genomes indicated by linkage map data in which tetraploid ecotypes demonstrate fewer than expected markers mapping across subgenomes, and complete or near complete disomic inheritance [15
]. It also agrees with the observation of 18 non-pairing univalents at diakinesis of meiosis in the dihaploid line [36
]. FISH signal data at these genetic loci may point toward allopolyploid evolution of the switchgrass genome. However, these data are also consistent with natural loss of gene content following a whole genome duplication within a single species (autopolyploidy). To gain a greater understanding of the origins of switchgrass polyploidy, further phylogenetic analyses of the Panicum
(s.s.) subgenus and/or genomic in situ hybridization (GISH) techniques should be used [1
Under a simple additive model, FISH signals for the 5S rDNA locus and the CentC locus would be expected to be present in 2, 4, and 8 copies in the dihaploid, tetraploid, and octoploid lines, respectively. Surprisingly, only one FISH signal for these loci was observed in the dihaploid ALB280. In addition, the switchgrass-specific centromere probe, PviCentC, hybridized to all chromosomes of both subgenomes, but demonstrated a stronger FISH signal on Chromosome 3 of a single subgenome (same locus as maize CentC probe). This discrepancy in signal strength, particularly for the universal centromere probe PviCentC, suggests that a greater copy number of this repeat is present in Chromosome 3 of one subgenome than in all other chromosomes. Alternatively, or in conjunction, homology to the PviCentC FISH probe sequence may be much higher in Chromosome 3 than in all other labeled centromeres. The variation we observe at rDNA and centromeric loci may also be a result of the allogamous habit and self-incompatibility of switchgrass [20
]. In outbreeding species of the genus Secale
, high levels of repeat DNA polymorphism between homologous chromosomes have been documented [57
]. Other outbreeding species in the Lolium
complex demonstrate variation at rDNA loci [58
], suggesting that hemizygosity in switchgrass may result from out-crossing. Also contributing to non-additive FISH signal data may be the high frequency of switchgrass aneuploids, particularly among octoploid cytotypes [18
], which can lead to large-scale genetic changes and parental genome imbalance.
In our analyses of 45S rDNA loci, pairs of telomeric FISH signals demonstrated a regular, additive pattern up the ploidy series. Among tetraploid cultivars, our data are consistent with those of Costich et al. [18
] in which all tetraploids analyzed (upland and lowland) demonstrated two pairs of telomeric 45S rDNA signals. However, in octoploid cultivars, Costich et al. [18
] describe a large amount of variation in size, number, and location of 45S rDNA signals. The variation in 45S rDNA signal intensity seen in our analysis of the octoploid cv. Caddo (see Figure i) may suggest rDNA repeat variation and/or differences in probe hybridization affinity. With only a single upland octoploid (cv. Caddo) analyzed with 45S rDNA in this study, our results likely demonstrate one of many chromosomal constitutions for switchgrass octoploids. Overall, FISH analyses of tetraploid and octoploid individuals support elimination of rDNA and centromere sequences and demonstrates that patterns of subgenome differentiation are broadly maintained.
Our data also demonstrate unique ecotype differences at 5S rDNA loci. Variations in FISH signal patterns between upland and lowland tetraploids (and between lowland tetraploids and upland octoploids) provide features that distinguish these taxonomic divisions within switchgrass. We hypothesize that these variations in rDNA loci are related to both the phenotypic and geographic distribution differences observed between switchgrass ecotypes. In Oryza
species, polymorphisms in the number, the chromosomal location, and the repeat length of rDNA loci have revealed species-specific and subgenome-specific FISH patterns [60
]. The authors suggest that specific inversions, rDNA amplification, and locus transposition may have occurred during the process of Oryza
evolution. Such a scenario may also be true for the divergence of switchgrass upland and lowland ecotypes. Classified cultivars of switchgrass are barely removed from the wild, and rapidly evolving rDNA loci are likely still undergoing change. The additive pattern of 5S rDNA loci seen between lowland tetraploids (2 signals) and upland tetraploids (4 signals) may be indicative of chromosomal rearrangements and rDNA changes that ultimately result in habitat adaptability differences between the two ecotypes. In addition, the conserved doubling pattern of 5S rDNA loci from upland tetraploid to upland octoploid further supports the maintenance of ecotype divergence across different cytotypes.
The presence of gene flow between switchgrass ecotypes and/or cytotypes has ramifications on the development and use of specific gene pools for switchgrass improvement. Therefore, knowledge of chromosome architecture and ploidy relationships is critical for cultivar development. Cytogenetic data presented here can be used to classify switchgrass plants according to ecotype and ultimately aid in identifying and isolating regionally adapted cultivars [61
]. Switchgrass improvement through trait identification and breeding for significant heterotic effects also warrants the maintenance of independent gene pools [63
]. Recent analyses of several putative upland and lowland accessions of switchgrass have identified the presence of natural hybrids between ecotypes as well as evidence of gene flow [22
]. Results demonstrated a mixture of both cytotypes and phenotypes within hybrid populations, suggesting a long history of gene flow. In addition, genetic marker data indicated that gene flow is bidirectional, from upland to lowland and from lowland to upland [23
]. A quantitative understanding of whole switchgrass chromosomes will help in distinguishing hybrid genotypes and aid in tracking genome sources during directed breeding programs. The use of FISH and GISH can identify translocations and/or introgression of new chromosome sources [25
], but can also be used to identify ancestral genomes that contribute to the evolution of polyploid species such as switchgrass [27
The development of a species karyotype, with unambiguous identification of individual chromosomes, is also critical for the integration of genetic and physical map data. Genetic maps of switchgrass have been constructed using SSR and STS markers [15
]. Hybridization of these genetically mapped markers to switchgrass chromosomes would lead to definitive assignment of linkage groups. Fluorescence detection of single- and low-copy sequences through the use of BAC clones as probes has proven highly successful in many species, including rice [67
], maize [68
], and sorghum [69
]. BAC libraries for switchgrass have recently been developed and may be utilized for the integration of linkage and physical maps through BAC-FISH probing. BAC sequences would also be useful in flow sorting of chromosome fractions for physical gene mapping and construction of chromosome-specific libraries [70
]. In this light, FISH technology will continue to be a valuable tool in understanding genome structure in switchgrass.