Molecular cytogenetic characterization and phylogenetic analysis of four Miscanthus species (Poaceae)

Abstract Chromosomes of four Miscanthus (Andersson, 1855) species including M. sinensis (Andersson, 1855), M. floridulus (Schumann & Lauterb, 1901), M. sacchariflorus (Hackel, 1882) and M. lutarioriparius (Chen & Renvoize, 2005) were analyzed using sequentially combined PI and DAPI (CPD) staining and fluorescence in situ hybridization (FISH) with 45S rDNA probe. To elucidate the phylogenetic relationship among the four Miscanthus species, the homology of repetitive sequences among the four species was analyzed by comparative genomic in situ hybridization (cGISH). Subsequently four Miscanthus species were clustered based on the internal transcribed spacer (ITS) of 45S rDNA. Molecular cytogenetic karyotypes of the four Miscanthus species were established for the first time using chromosome measurements, fluorochrome bands and 45S rDNA FISH signals, which will provide a cytogenetic tool for the identification of these four species. All the four have the karyotype formula of Miscanthus species, which is 2n = 2x = 38 = 34m(2SAT) + 4sm, and one pair of 45S rDNA sites. The latter were shown as strong red bands by CPD staining. A non-rDNA CPD band emerged in M. floridulus and some blue DAPI bands appeared in M. sinensis and M. floridulus. The hybridization signals of M. floridulus genomic DNA to the chromosomes of M. sinensis and M. lutarioriparius genomic DNA to the chromosomes of M. sacchariflorus were stronger and more evenly distributed than other combinations. Molecular phylogenetic trees showed that M. sinensis and M. floridulus were closest relatives, and M. sacchariflorus and M. lutarioriparius were also closely related. These findings were consistent with the phylogenetic relationships inferred from the cGISH patterns.


Introduction
The genus Miscanthus (Andersson, 1855), belonging to the tribe Andropogoneae of family Poaceae, is a tall perennial grass with C4 photosynthesis (Stewart et al. 2009). It includes 14-20 species and has been considered as one of the most promising highyield fiber-based energy crops (Christian et al. 2009, Brosse et al. 2012. China is a genetic center of diverse Miscanthus germplasm. Four Miscanthus species, M. sinensis (Andersson, 1855), M. floridulus (Schumann & Lauterb, 1901), M. sacchariflorus (Hackel, 1882) and M. lutarioriparius (Chen & Renvoize, 2005), are most widely distributed. These have high biomass yield and are prone to interspecific hybridization, which lead to high genetic diversity (Liu et al. 2013, Zhao et al. 2017.
M. sinensis has been already sequenced (Miscanthus sinensis v7.1 DOE-JGI, https://phytozome.jgi.doe.gov/) and is an important species for comparative genomics. Therefore, it is necessary to investigate the chromosomes of M. sinensis and other Miscanthus species and reveal their genomic homology. It will provide a reference for further development of specific probes based on the M. sinensis genome sequence for chromosomal localization in Miscanthus and related genera. M. floridulus is similar to M. sinensis in morphology. M. lutarioriparius is a native Miscanthus species of China (Chen and Renvoize 2006, Sheng et al. 2016, Yang et al. 2019. Some scholars have published it as a variant or subspecies of M. sacchariflorus (Liu 2009, Sun et al. 2010, Lu 2012, Hu 2015 because of their high similarity in morphology. In a word, the interspecific relationships of the four Miscanthus species are complex and their origins are unclear Renvoize 2006, Tamura et al. 2016). Until now, taxonomic studies on M. sinensis, M. floridulus, M. sacchariflorus and M. lutarioriparius have been carried out by using morphological features (Chae et al. 2014, Yook et al. 2014, nuclear DNA content (Chae et al. 2014, Sheng et al. 2016, molecular markers (Chae et al. 2014, Yook et al. 2014, Tang et al. 2015, fluorescence in situ hybridization (FISH) , and spectroscopy (Jin et al. 2017). However, none of them could distinguish the four species from each other unequivocally. Karyotype, the characterization of a genome at the chromosomal level, is a valuable tool for species identification and evolution analysis (Silva et al. 2018). However, it is difficult to perform accurate karyotype analysis in Miscanthus species because of their little chromosomal differentiation and lack of distinct chromosomal landmarks. Chromosome banding by Giemsa staining, fluorochrome staining or FISH with repetitive DNA sequences can provide additional characteristics to discriminate the chromosomes in the cell complement (Filion 1974, Sumner 1990, Koornneef et al. 2003.
Fluorochrome banding techniques use fluorescent dyes preferentially binding to GC-or AT-rich DNA sequences to display different classes of heterochromatin on chromosomes (Sumner 1990). Among the techniques used, combined PI (propidium iodide) and DAPI (4', 6 diamino-2-phenylindole dihydrochloride) staining (called CPD staining) can reveal simultaneously GC-rich and AT-rich regions along chromosomes with high precision and repeatability (Peterson et al. 1999, She et al. 2006, 2015. FISH with 5S and 45S rDNA probes have been widely applied in plants to determine the number and location of rDNA sites, and to provide effective markers for chromosome identification. Moreover, information on evolutionary relationships between species can be provided by comparing rDNA distribution characteristics between closely related species (Moscone et al. 1999, de Moraes et al. 2007, She et al. 2015. The combination of chromosome morphology, fluorochrome bands and FISH signals can be employed to construct molecular cytogenetic karyotype. It can reveal chromosome-level genome organization of a plant species, investigate the evolutionary relationships among related species, and integrate genetic and physical maps , She et al. 2015. Cytogenetic studies in the four Miscanthus species were so far primarily restricted to chromosome counts and conventional karyotype descriptions (Chramiec-Głąbik et al. 2012, Chae et al. 2014. FISH has been applied in diploid M. sinensis and tetraploid M. sacchariflorus, but their molecular cytogenetic karyotypes have not been established as yet .
The internal transcribed spacer (ITS) regions of 45S rDNA have been used extensively for determining phylogenetic relationships at interspecific or intraspecific level because of its relatively high rate of mutation (Álvarez and Wendel 2003, Hao et al. 2004, Capua et al. 2017. Another direct method for examining genome relationships is comparative genomic in situ hybridization (cGISH), in which the labelled total genomic DNA of one species is hybridized to the chromosomes of another species without competitive DNA (Zoller et al. 2001). It generates hybridization signals in regions of conserved repetitive DNA sequences. Therefore, it can be used to identify the evolutionary relationships between species within a genus (Wolny and Hasterok 2009, She et al. 2015. So far, the phylogenetic relationships among the four Miscanthus species were mainly carried out at the morphological, cellular and molecular levels (Chae et al. 2014, Yook et al. 2014. In previous studies, the ITS sequence was used to assess the phylogeny of the four Miscanthus species and the species of the Saccharum complex and other related genera (Hodkinson et al. 2002b, Chen et al. 2007, Liu et al. 2010, and the genome relationship between diploid M. sinensis and tetraploid M. sacchariflorus was examined by FISH with rDNA, genomic DNA and Saccharum centromeric repeats . However, there has been no study on the phylogenetic relationships of all the four species by combining molecular cytogenetic characterization with ITS sequence analysis.
In the current study, well spread mitotic metaphase chromosomes of four Miscanthus species were prepared using the modified flame-drying method. Chromosomes were characterized using sequential CPD staining and FISH with 45S rDNA probe. Detailed molecular cytogenetic karyotypes of these species were established using combined data of chromosome measurements, CPD bands, DAPI bands and 45S rDNA FISH signals. Meanwhile, cGISH was carried out to detect the homology of repetitive DNAs among these species, and a comparative sequence analysis of the ITS regions in these species was also conducted. The data were collected and evaluated to gain insight about the phylogenetic relationships among the four Miscanthus species.

Plant material and DNA extraction
Twenty-four Miscanthus accessions comprised of 6 M. sinensis, 6 M. floridulus, 6 M. sacchariflorus and 6 M. lutarioriparius were selected from different provinces of China and planted in the Miscanthus germplasm nursery located at the Hunan Agricultural University (Table 1). All the materials were used for ITS sequence analysis. Meanwhile, for CPD staining, rDNA FISH, karyotype analysis and cGISH, No. 03 (M. sinensis), No. 10 (M. floridulus), No. 16 (M. sacchariflorus) and No. 21 (M. lutarioriparius) were used. Imperata cylindrica (Beauvois, 1812) was included as an outgroup for the ITS phylogenetic analysis, and its sequences (JN407505.1) were obtained from GenBank Database.
Total genomic DNA (gDNA) was extracted from fresh leaf tissue using the cetyltrimethylammonium bromide (CTAB) method described by Murray and Thompson (1980). The quality and concentration of DNA were measured by 1% agarose gel electrophoresis and a microplate reader (BioTek Instruments Inc, Winooski, USA).

Chromosome preparation
Mitotic metaphase chromosomes were prepared by using the root tips according to the procedure described by She et al. (2006). The actively growing root tips were collected from potted plants and treated with saturated α-bromonaphthalene for 1.5 h at 28 °C, fixed in 3:1 (v/v) methanol/glacial acetic acid for at least 12 h at room temperature, and then stored at 4 °C until use. The fixed root tips were then washed in double distilled water and citrate buffer (0.01 mM citric acid-sodium citrate, pH 4.5) for 10 min each and incubated in a mixture of 2% cellulase R-10 (Yakult Pharmaceutical Industry, Tokyo, Japan), 2% pectolyase Y-23 (Yakult Pharmaceuticals), and 2% macerozyme R-10 (Sigma-Aldrich, Steinhem, Germany) in citric acid buffer at 28 °C for 2.5~3 h. Root tips were transferred to a glass slide along with the fixative and dissected using fine-pointed forceps. Finally, the slides were dried above a flame. Good preparations were selected by Olympus BX60 phase contrast microscope, and then stored at -20 °C.

Staining with CPD and DAPI
The CPD staining followed the procedure described in She et al. (2006). Chromosome preparations were treated with RNase A and pepsin then stained with a mixture of 3 μg/ml DAPI and 0.6 μg/ml PI (both from Sigma-Aldrich) in a 30% (v/v, using double-distilled water as solvent) solution of Vectashield H-1000 (Vector Laboratories Burlingame, USA). Preparations were examined under an Olympus BX60 epifluorescence microscope equipped with a CoolSNAP EZ CCD camera (Photometrics, Tucson, USA). The CCD camera was controlled using Ocular software (Molecular Devices, Sunnyvale, USA). Photographs were taken using a green excitation filter for PI and a UV excitation filter for DAPI. DAPI and PI grey-scale images of the same plate were merged to produce a CPD image. The final images were optimized for contrast and background using Adobe PHOTOSHOP CS8.0.

Probe DNA labeling
A 45S rDNA clone containing a 9.04-kb tomato 45S rDNA insert (Perry and Palukaitis 1990) were used as probe to localize the 18S-5.8S-26S ribosomal RNA gene. The DNAs (45S rDNA and gDNAs) of the four Miscanthus species were labeled with biotin-16-dUTP and digoxigenin-11-dUTP, respectively, using the Nick Translation Kit (Roche Diagnostics, Mannheim, Germany). lutarioriparius chromosomes were performed, respectively. The slides previously stained or hybridized were washed twice for 15 min each in 2 × SSC, dehydrated through an ethanol series (70%, 90% and 100%, 5 min each) and then used for hybridization. The in situ hybridization and detection were performed as described by She et al. (2006). Briefly, 40 μl of the hybridization mixture, which contained 20 μl 20% dextran sulfate, 1 μl ssDNA, 16 μl hybridization buffer (HB50, containing 50% deionized formamide and 50 mM sodium phosphate, pH 7.5) and 3 μl labeled DNA (final concentration 100-150 ng/slide), was added to each slide and covered with a 24 × 50 mm glass coverslip. Chromosomes and probe were denatured together on ThermoBrite S500-24 (Abbott Molecular, USA) at 80 °C for 3 min, and then were incubated at 37 °C for 24 h. Post-hybridization washing was performed in 0.1 × SSC two times for 15 min each at 42 °C, followed by rinsing in 2 × SSC three times for 5 min each at 42 °C and in TN buffer (containing 100 mM Tris-HCl and 150 mM NaCl, pH 7.5) for 5 min at room temperature. Hybridization signals were detected after incubating the slides with 100 μl TNB buffer (0.5% Roche blocking reagent in TN buffer) for 30 min at 37 °C, and followed by rinsing in TN buffer for 1 min at room temperature. The biotin-labeled 45S rDNA was detected using Fluorescein Avidin D (Vector Laboratories). The digoxigenin-labeled gDNA was detected using Anti-digoxigeninrhodamine (Roche Diagnostics). The specific steps were as follows: 100 μl of 1% Fluorescein Avidin D or Anti-digoxigenin-rhodamine, diluted with TNB buffer, was added to each slide and covered with a glass coverslip, and then were incubated at 37 °C in dark for 1 h. Afterwards, the coverslip was removed and rinsed with TN buffer three times for 5 min each in dark. Slides were counterstained with 3 μg/ml DAPI in a 30% solution of Vectashield H100 and subsequently examined under an epifluorescence microscope equipped with the CCD camera as mentioned above. Observations were made using a UV, blue and green excitation filters for DAPI, fluorescein, and rhodamine, respectively. Grey-scale images were digitally captured and merged by the Ocular software. The final images were adjusted with Adobe PHOTOSHOP CS8.0.

Karyotype analysis
For each species, five well-spread metaphase plates were measured using Adobe PHO-TOSHOP CS8.0 to obtain the chromosome relative lengths (RL; % of haploid complement), arm ratios (AR = long arm/short arm), chromosome length ratio (longest chromosome length / shortest chromosome length), size of the fluorochrome band (expressed as a percentage of the karyotype length) and the percentage distance from the rDNA site to the centromere (di = d × 100/a; d = distance of the centre of the rDNA sites from the centromere; a = length of the corresponding chromosome arm) (Greil-huber and Speta 1976). In addition, the total length of the haploid complement (TCL; i.e. karyotype length) was measured using the five metaphase cells with the highest degree of chromosome condensation. The arm ratio was used to classify the chromosomes according to the system described by Levan et al. (1964). Karyotype asymmetry was determined using the mean centromeric index (CI), the intrachromosomal asymmetry index (A1), the interchromosomal asymmetry index (A2) (Zarco 1986), the ratio of the length of all long arms in the chromosome set to the total chromosome length in set (As K%) (Arano 1963), the asymmetry index (AI) (Paszko 2006), and the categories of Stebbins (Stebbins 1971). The chromosomes were arranged in order of decreasing lengths. Idiograms were drawn based on chromosomes measurement data, fluorochrome bands and 45S rDNA FISH signals.

DNA sequences and phylogenetic analyses
Each DNA sequence was spliced by bi-directional sequencing. Then, the similarity searches were performed using the NCBI (National Centre for Biotechnology Information), BLAST network service. Sequences were aligned with CLUSTAL W program. The MEGA 7.0 software (Kumar et al. 2016) was used for sequence analyses (estimating percentage of the G + C content, variable sites and parsimony informative sites). ITS sequences of I. cylindrical was used as the outer group plant, and phylogenetic analyses were carried out using the neighbour joining (NJ) and maximum parsimony (MP) methods. In the ITS phylogenetic tree, the confidence of each branch was tested using bootstrap (Felsenstein 1985), each performing 1 000 cycles to evaluate the systematic significance and reliability of each branch.

Comparative karyotyping
The karyotype measurement data for the four Miscanthus species are listed in Suppl. material 1, Tables S1-S4. The general karyotype features and parameters for the four Miscanthus species are listed in Table 2. Representative mitotic chromosomes, karyotypes showing the fluorescent bands and ideograms are shown in Figs 1-3.
All the four Miscanthus species had diploid chromosome number of 2n = 2x = 38. The mitotic metaphase chromosomes with a mean chromosome length 3.59 μm for M. sacchariflorus and 4.53 μm for M. floridulus. The total length of the haploid complement (TCL) ranges from 68.15 μm to 86.13 μm, and the mean centromeric index (CI) of the complements varied slightly between 44.00 ± 4.97 and 44.81 ± 4.28. In contrast, M. floridulus has exhibited the large variation in chromosome length, whereas M. sinensis has displayed the large variation in centromeric index.
The karyotype formulas of the four Miscanthus species were same, composed of 34 metacentric (m) chromosomes and 4 submetacentric (sm) chromosomes with a secondary constriction located on the long arms of chromosome 1, namely 2n = 2x = 38 = 34m(2SAT) + 4sm. All the karyotypes of the four species studied fell into the categories 2B of Stebbins (1971). The ranges for Romero Zarco's (1986) asymmetry indices were as follows: A1 = 0.11-0.20 and A2 = 0.24-0.52. The As K% of Arano (1963) ranged from 55.25 to 55.85, and Paszko's (2006) asymmetry index (AI) ranged from 2.37 to 4.47. RRL, CI, A1 and As K% have shown close similarity among species. In contrast, TCL, A2, and AI have displayed relatively large variation among species. According to the AI values, the karyotype of M. sacchariflorus is the most symmetrical and that of M. lutarioriparius is the most asymmetrical among the four species.

Fluorescence banding patterns
After CPD staining, slightly different fluorochrome banding patterns were observed among the four Miscanthus species (Fig. 1-3; Table 3). The red CPD bands were recorded in all species, whereas blue DAPI bands were found only in M. sinensis and M. floridulus. KF, karyotype formula; TCL, total length of the haploid complement (i.e. karyotype length); RRL, ranges of chromosome relative length; CI, mean centromeric index; A1 and A2, intra-chromosomal asymmetry index and inter-chromosomal asymmetry index, respectively; As K%, ratio of the length of all long arms in chromosome set to total chromosome length; AI, karyotype asymmetry index.    Results showed that only one pair of CPD bands in M. sinensis, M. sacchariflorus and M. lutarioriparius had occurred in the secondary constrictions on the long arms of chromosome 1, and were co-localized with the 45S rDNA-FISH hybridization sites (called rDNA CPD bands; Fig. 1A, E, G). There were three CPD bands in the M. floridulus: two bands correspond to the secondary constriction on the long arms of chromosome 1; the other band was a non-rDNA CPD band with weaker fluorescence, occurring in the pericentromeric region of a homologue of chromosome pair 15 (Fig. 1G). The rDNA CPD bands in M. sinensis and M. lutarioriparius were similar in size and intensity on the two homologous chromosomes, while those in M. floridulus and M. sacchariflorus displayed heterozygosity, the band on one chromosome was large and bright, whereas the band on the other homologue was small and weak. The CPD bands of M. sinensis, M. floridulus, M. sacchariflorus and M. lutarioriparius accounted for 0.93%, 1.11%, 0.90% and 0.75% of the karyotype length, respectively.
M. sinensis showed seven blue DAPI bands (Fig. 1A, 2A, 3A): two pairs of weak bands occurred in the centromeric regions of chromosomes 10 and the pericentromeric regions of the long arm of chromosome 12, three relatively strong bands occured on chromosome 19. Among the DAPI bands on pair 19, two were located in the centromeric regions of both homologues, and one occured in the proximal region of the long arm of one homologue. M. floridulus had shown only one pair of weak DAPI bands in the pericentromeric regions on the chromosomes 19 (Fig. 1G, 2B, 3B). The DAPI bands of M. sinensis and M. floridulus accounted for 1.94% and 0.45% of the karyotype length, respectively.

FISH mapping of 45S rDNA
45S rDNA FISH showed that M. sinensis, M. floridulus, M. sacchariflorus and M. lutarioriparius had only one pair of 45S rDNA sites, which were located in the secondary constriction on the long arms of chromosome 1, and their percentage distances of 45S rDNA sites were 25.53 ± 1.17, 29.43 ± 1.12, 32.07 ± 0.49, 28.45 ± 0.89, respectively. The 45S rDNA sites of the four Miscanthus species corresponded to their respective CPD bands in both size and intensity, that is, the 45S rDNA signals of two homologues in M. sinensis and M. lutarioriparius were similar in size and intensity, while those in M. floridulus and M. sacchariflorus differed in size and intensity, displaying heterozygosity.

GISH signal patterns
The GISH results are shown in Fig. 4. Both self-GISH (sGISH; the genomic DNA of a species is applied to its own chromosomes) and cGISH generated hybridization signals in most regions of all chromosomes. Overall, the hybridization signals in the proximal and/or centromeric regions of the chromosomes were strong or very strong, while those in the proximal regions were relatively weak. In the GISHs to the M. sinensis chromosomes, the signals generated by M. floridulus gDNA (Fig. 4B) were stronger and more evenly distributed than those generated by M. sacchariflorus and M. lutarioriparius gD-NAs (Fig. 4C, D), and more similar to the sGISH signals of M. sinensis (Fig. 4A). The hybridization signals of M. floridulus gDNA to the chromosomes of M. sacchariflorus (Fig. 4F) and M. lutarioriparius (Fig. 4I) were weaker than both the sGISH signals of M. sacchariflorus (Fig. 4E) and M. lutarioriparius (Fig. 4H), and the cGISH signals of M. lutarioriparius gDNA to M. sacchariflorus chromosomes (Fig. 4G) .

Phylogeny analysis based on ITS
Each ITS1-5.8S-ITS2 sequences were compared to the published sequences of Miscanthus and its related species, and the boundaries of the spacer regions were confirmed. The length and other characteristics of each ITS1-5.8S-ITS2 are given in Table  4. The entire ITS sequence (ITS1-5.8S-ITS2) of I. cylindrica that was used as the outgroup species was 684 bp in length, and its GC content was 63.89%. Neighbour joining (NJ) and maximum likelihood (ML) phylogenetic trees were developed based on the entire ITS sequences. The NJ and ML trees were very similar (Fig. 5), and the four Miscanthus species were divided into two categories: (i) group I contained M. sinensis and M. floridulus, which resulted in 82% (NJ) and 94% (ML) bootstrap values; (ii) group II included M. sacchariflorus and M. lutarioriparius, with 83% and 84% bootstrap values in NJ and ML trees, respectively. It was worth noting that in each branch the accessions of one species were not separated from those of another species (Fig. 5).

Molecular cytogenetic karyotypes
In the present study, detailed karyotypes of M. sinensis, M. floridulus, M. sacchariflorus and M. lutarioriparius were established using a combination of chromosome measurements, fluorochrome bands and 45S rDNA FISH signals, which provided the primary molecular-cytogenetic characterization of the four Miscanthus energy plants for the first time.
Our results had shown that the molecular cytogenetic karyotypes were rather similar among the four Miscanthus species. For instance, their karyotype formula, the categories of Stebbins, and the number and location of the 45S rDNA sites were same and there were slight differences in RRL, CI, A1 and As K%. However, several differences in their karyotypes were recorded: (1) obvious differences in their TCL, A2 and AI. In particular, AI, which can more accurately reveal the heterogeneity of chromosome length and centromere index in karyotype (Paszko 2006), had significant differences among the four species. Therefore, the four Miscanthus species could be accurately distinguished from each other using the molecular-cytogenetic karyotypic data.
As our study revealed, the chromosome numbers of the four Miscanthus species were all 2n = 2× = 38, being consistent with those reported previously (Hodkinson et al. 2002a, Chramiec-Głąbik et al. 2012. The current karyotypes of the four Miscanthus species comprised  , Chramiec-Głąbik et al. 2012. The deviations in karyotype were probably mainly due to differences in the material analysed, and difficulty in accurately pairing homologous chromosomes and distinguishing chromosomes by using the classical staining technique as applied before.

Phylogenetic relationships of the four Miscanthus species
The research on the evolutionary relationship among the four Miscanthus species could not be conducted using genomic information because the genome sequences of M. floridulus, M. sacchariflorus and M. lutarioriparius were currently unavailable. Therefore, at present the combination of molecular cytogenetic analysis with rDNA ITS and chloroplast DNA sequence analysis was an effective phylogenetic analysis pathway (She et al. 2015). As mentioned above, the molecular cytogenetic karyotypes of the four Miscanthus species were very similar, indicating the high genome similarity and small genome differentiation among them. Furthemore, the cGISH signals were rather similar in intensity and distribution to the sGISH signals, further demonstrating the high similarity among the four Miscanthus genomes (Wolny and Hasterok 2009, She et al. 2015. However, the differences in molecular cytogenetic karyotypes and cGISH signals among the four species also provided valuable evolution information: (1) AT-rich heterochromatin appeared in both M. sinensis and M. floridulus, but did not emerge in both M. sacchariflorus and M. lutarioriparius; (2) The hybridization signals of M. floridulus gDNA to M. sinensis chromosomes were stronger and more evenly distributed than those of M. sacchariflorus and M. lutarioriparius gDNA, whereas the hybridization signals of M. floridulus gDNA to M. sacchariflorus chromosomes were weaker than the hybridization signals of M. lutarioriparius gDNA to M. sacchariflorus chromosomes. This information indicates that there was a close phylogenetic relationship between M. sinensis and M. floridulus, and between M. sacchariflorus and M. lutarioriparius; the former two species were relatively distant from the latter two species.
In our study, the phylogenetic tree based on the ITS sequences had shown that, M. sinensis and M. floridulus clustered into one branch, and M. sacchariflorus and M. lutarioriparius clustered into another branch. It was consistent with not only the above molecular cytogenetic results, but also the previous clustering results based on morphology, molecular markers and the ITS sequence (Hodkinson et al. 2002b, Chen et al. 2007, Sun et al. 2010, Chae et al. 2014, Clark et al. 2014, Ge et al. 2017). In addition, our clusering analysis revealed that the accessions of the two species in each branch were mixed and without distinct boundaries. These findings were consistent with the phylogenetic trees of Miscanthus and related genera inferred from ITS sequences (Hodkinson et al. 2002b, Chen et al. 2007, Liu et al. 2010, SSR markers (Ge et al. 2017), SNPs (Clark et al. 2014) and the dataset of genome size, ploidy level and genomic polymorphisms (Chae et al. 2014). To summarise, our clustering results have demonstrated that there was a very close phylogenetic relationship between M. sinensis and M. floridulus, and between M. sacchariflorus and M. lutarioriparius and they cannot be distinguished only based on the ITS sequences. However, as revealed in this study, the molecular cytogenetic karyotype analysis can effectively identify the four species.

Conclusion
Molecular cytogenetic karyotypes of M. sinensis, M. floridulus, M. sacchariflorus and M. lutarioriparius were established for the first time, which can effectively distinguish the four species. Molecular cytogenetic comparison revealed basic similarities and certain differences in genome organization among the four species. These findings will provide a reference for further development of specific probes based on M. sinensis genome sequence for chromosomal localization in the species of Miscanthus and related genera. The combined data of molecular cytogenetic and ITS sequence analysis indicated a close phylogenetic relationship between M. sinensis and M. floridulus, and between M. sacchariflorus and M. lutarioriparius, respectively. It can be concluded that former two species have relatively distant relationship compared with the latter two species.