Background

The tribe Triticeae includes about 450 species, of which about 75% are polyploid [1, 2]. Since Löve [1] proposed that the species with the same genome or same genome combinations were classified into one genus, about 30 genera were recognized by most of the grass scientists [3,4,5,6,7,8,9]. Elymus sensu lato (Elymus s.l.) is the largest genus of Triticeae, and it contains seven basic genomes: St, H, P, W, Ns, Y, and Xm [3, 8, 10,11,12]. St genome is from Pseudoroegneria (Nevski) Löve, H genome is from Hordeum L., P genome is from Agropyron Gartn., W genome is from Australopyrum (Tzvelev) Löve, Ns genome is from Leymus Hochst. The origin of Y and Xm is still unknown [3, 7, 9, 13]. Based on the genome combinations, Elymus s.l. was further divided into ten genera, including Elymus sensu stricto (Elymus s.s.) (StH), Roegneria C.Koch (StY), Hystrix Moench (StH/NsXm), Stenostachys Turcz. (HW), Douglasdeweya C.Yen, J.L.Yang et B.R.Baum (StP), Kengyilia C.Yen et J.L.Yang (StYP), Campeiostachys Drobov (StYH), Anthosachne Steudel (StYW), Pascopyrum Á. Löve (StHNsXm), and Connorochloa Barkworth, S.W.L.Jacobs et H.Q.Zhang (StYWH) [5,6,7,8, 13,14,15,16]. Of which, due to the dominant effect of the genes of the St and H genomes, it is challenging to distinguish Campeiostachys from Elymus s.s. based on single or combined morphological characters [6, 8, 17]. Moreover, the genome composition of many polyploid species in Elymus s.s. and Campeiostachys is still unknown, resulting in the classification of many species in these two genera remains controversial [6, 8]. Although the genome composition of some species is determined, their biosystematics remains controversial due to their similar morphological features.

Elymus breviaristatus (Keng) Keng ex Keng f. and Elymus sinosubmuticus S. L. Chen is sympatric species mainly distributed on hillsides in Sichuan, Qinghai, and Ningxia, China [9, 18,19,20,21]. Morphologically, those two species are quite similar, and the only difference exists in their awn length. E. breviaristatus has short awn (2–5 mm), while E. sinosubmuticus possesses degenerated awn only 0-2 mm in length [9, 19, 22]. Overlap** geographical distribution and similar morphology, whether or not they are the same species is under controversy. Based on the morphological characteristics and the results of RPDA analysis, these two species were treated as independent biological species [8, 9, 19, 21, 23]. Zhang et al. [24] suggested that E. breviaristatus and E. sinosubmuticus were the same species by comparing the leaf anatomical characteristics.

The chromosome pairing behavior of hybrid F1 at meiosis metaphase can be used to indicate chromosome homology and evolutionary relationship between genus or species in Triticeae [25, 26]. Genomic in situ hybridization (GISH) can effectively examine the genome composition and chromosomal rearrangement of polyploid species [27,28,29,30,31,32]. Cytologically, E. breviaristatus and E. sinosubmuticus are allohexaploid (2n = 6x = 42) perennial wheatgrass [1, 8, 21, 33], but Mason-Gamer et al. [34] reported that E. breviaristatus is tetraploid with StH genome. Based on the genome analysis and GISH, Yang et al. [35] recognized that E. breviaristatus was a hexaploid with the StYH genome and treated it as Campeistachys breviaristata (Keng) Y.H.Zhou, H.Q.Zhang et C.R.Yang. However, the genome composition of E. sinosubmuticus has not been reported at present.

Cytological and phylogenetic analyses are practical tools to determine the genome composition and explore the interspecies and intergeneric relationships of the species in Triticeae [36,37,38,39]. Molecular phylogeny analysis based on the single- or low-copy nuclear genes is less susceptible to concerted evolution and can be a handy marker for polyploid phylogeny [40,41,42,43,44]. Furthermore, Petersen et al. [43] found a correspondence between DNA sequences of diploid donors and allopolyploids in Triticeae. Therefore, more and more single-copy nuclear genes have been used to determine the genome composition and phylogenetic relationship of Triticeae. Acc1 and DMC1 sequences have higher evolutionary rates and have been widely applied in the phylogenetic study of the genera of Triticeae, such as Triticum, Kengyilia, Leymus, Roegneria, Hystrix, etc. [44,45,46,47,48,49]. In the present study, GISH, single-copy nuclear genes, and artificial hybridization were used to investigate the genome composition of E. sinosubmuticus and explore the biosystematics relationships between E. breviaristatus and E. sinosubmuticus.

Results

Meiosis and fertility of parentals and F1 hybrids

Five hybrids were obtained from the combination of E. breviaristatus × E. sinosubmuticus. We observed the chromosome pairing of PMCs at metaphase I (MI) of parents and hybrids (Table 1). Meiosis of E. breviaristatus and E. sinosubmuticus forming mostly ring bivalents, with an average of 21.00 and 20.92 bivalents per cell, respectively (Table 1; Fig. S1, see Additional file 1). The F1 hybrids of E. breviaristatus × E. sinosubmuticus was a hexaploid (2n = 42), showing an average of 0.50 univalents, 20.62 bivalents, 0.06 trivalent, and 0.02 quadrivalents (Table 1; Fig. S1, see Additional file 1). The chiasmata per cell were 37.70, with a c-value of 0.89, suggesting that they were genetically affinity species and had similar StYH genome constitution.

Table 1 Meiotic associations at metaphase I in pollen mother cells of parental species and their hybrids
Full size table

Pollen grains of parents (E. breviaristatus and E. sinosubmuticus) showed a high level of stainability, was 92.91% and 92.32%, respectively. The percentage of stained pollen grains of the hybrids was comparatively high at 53.90%. The seed setting rate of E. breviaristatus and E. sinosubmuticus were 89% and 87%, respectively. And the seed setting rate of their hybrids was 22.46%, indicating that the two species were highly affinities.

GISH analysis

To confirm the genome constitution of E. sinosubmuticus, root meristem cells that went through mitosis metaphase were collected for GISH. It showed that E. sinosubmuticus is a hexaploid with 42 chromosomes. Of which, 28 chromosomes were hybridized with the StY probe (from Roegneria ciliaris (Trin.) Nevski) when blocked by the H genome (from Hordeum bogdanii Wilensky) (Fig. 1a). And 14 chromosomes were hybridized with the H probe when blocked by the StY genome (Fig. 1b). Double-color GISH showed that 28 chromosomes were stained by the StY probe (in red), and 14 chromosomes were labeled by the H probe (in green) (Fig. 1c). In accordance, E. breviaristatus also contains 42 chromosomes, and 28 chromosomes displayed StY signals on the entire arm, and 14 chromosomes showed H signals (Fig. 1d, e, f).

Fig. 1
figure 1

GISH on somatic metaphase cells from root tips of Elymus sinosubmuticus and Elymus breviaristatus. a-c, E. sinosubmuticus. a, 14 chromosomes showed H genome single which from Hordeum bogdanii when blocked with StY genome which from Roegneria ciliaris. b, 28 chromosomes showed StY singles when blocked with H genome. c, 14 chromosomes showed H genome singles and 28 chromosomes showed StY singles when StY genome and H genome as probes. d-f, E. breviaristatus. d, 14 chromosomes showed H genome singles when blocked with StY genome. e, 28 chromosomes showed StY singles when blocked with H genome. f, 14 chromosomes showed H genome singles and 28 chromosomes showed StY singles when StY genome and H genome as probes. Scale bar equal 10 μm

Full size image

Phylogenetic analyses

The Acetyl-CoA carboxylase (Acc1) sequences length of E. sinosubmuticus ranged from 1421 to 1443 bp, and E. breviaristatus went from 1428 to 1441 bp. The Acc1 data matrix of sequences was analyzed based on maximum likelihood (ML) using the model TIM1 + I + G (-Ln likelihood = 8147.4309). The assumed nucleotide frequencies were A = 0.2555, C = 0.1794, G = 0.2116, T = 0.3535. The tree generated by Bayesian analysis was similar to ML analysis. All the Acc1 sequences were grouped into six clades (Fig. 2). The sequence from E. breviaristatus and E. sinosubmuticus were divided into clade I, clade III, and clade IV, respectively. Clade I contained the Pseudoroegneria, Elymus, Roegneria, and Campeiostachys species (BS = 100%, PP = 76%,). Clade III included in the Dasypyrum, Roegneria, and Campeiostachys species (BS = 100%, PP = 80%,). Clade IV grouped with the Hordeum, Elymus, and Campeiostachys species (BS = 100%, PP = 100%).

Fig. 2
figure 2

Maximum likelihood tree derived from Acc1 sequences data. The bold indicated sequences from Elymus sinosubmuticus and Elymus breviaristatus. The capital letters in brackets after the species name indicate the genome composition of the species, and the “?” indicate the genome composition of the species is unknown. The numbers above and below the branches indicate bootstrap values > 50% and Bayesian posterior probability values > 90%, respectively

Full size image

A total of 71 disrupted meiotic cDNA (DMC1) sequences were used for ML analysis, Bromus sterilis as the outgroup. TPM2uf + G as the best-fit model, -Ln likelihood = 5355.5355. The assumed nucleotide frequencies were A = 0.2576, C = 0.2120, G = 0.2085, T = 0.3220. The tree generated by Bayesian analysis was similar to ML analysis. The DMC1 sequences from E. breviaristatus and E. sinosubmuticus were divided into three clades (Fig. 3). In clade I, grouped with the diploid species (Pseudoroegneria), and tetraploid species (Elymus and Roegneria), and hexaploid species (Campeiostachys) (BS = 97%; PP = 74%). In clade II, their sequences grouped with the species of the genus Roegneria and Campeiostachys (BS = 99%, PP = 70%). In clade III, grouped with the diploid species (Hordeum) and the species of the genus Elymus and Campeiostachys (BS = 100%, PP = 59%). Clade IV and clade V grouped with the other diploid species in Triticeae (Fig. 3).

Fig. 3
figure 3

Maximum likelihood tree derived from DMC1 sequences data. The bold indicated sequences from Elymus sinosubmuticus and Elymus breviaristatus. The capital letters in brackets after the species name indicate the genome composition of the species, and the “?” indicate the genome composition of the species is unknown. The numbers above and below the branches indicate bootstrap values > 50% and Bayesian posterior probability values > 90%, respectively

Full size image

Discussion

Elymus sinosubmuticus contains StStYYHH genome

In this study, genome analysis, GISH, and phylogenetic analyses indicate that E. sinosubmuticus is an allohexaploid with the StYH genome. Based on the genome combinations, the species with St, Y, H genomes should be classified into the genus of Campeiostachys [6]. Previously, E. sinosubmuticus was classified into the Elymus genus based on morphological traits and geographic distribution [8, 9, 21]. Phenotype is the co-consequence of genetics and environments. Some studies have shown that there are cryptic species (such as Roegneria panormitana (Parl.) Nevski and R. heterophylla (Bornm. ex Melderis) C. Yen, J. L. Yang and B. R. Baum) and cryptic genera (such as Elymus and Campeiostachys) in Triticeae [6, 8, 50]. The former has complete reproductive isolation, and the latter has different genome combinations. None of them can be distinguished morphologically. Therefore, for the Triticeae, especially for cryptic genera, it is not accurate to classify the species based only on morphological traits. Our study is reasonable to classify E. sinosubmuticus into the genus Campeiostachys based on the genome analysis, GISH, and phylogenetic analyses results.

Biosystematics relationships of E. breviaristatus and E. sinosubmuticus

There is still debate whether or not E. breviaristatus and E. sinosubmuticus are the same species [8, 23]. Karyotype analysis showed that those two hexaploid species belonged to type 2A [20]. From the leaf anatomical structure, the comparison of the leaf anatomical characteristics showed that the external morphology of E. breviaristatus and E. sinosubmuticus had little difference in leaf anatomy, and it was difficult to distinguish. Therefore, E. breviaristatus and E. sinosubmuticus were the same species, and E. sinosubmuticus should be a synonym for E. breviaristatus [24]. Conversely, Zhou et al. [23] based on the results of RPDA analysis, despite the close relationship between them, there was a certain degree of nucleotide sequence difference, and they were independent biological species. The morphological characteristics of E. breviaristatus and E. sinosubmuticus we observed were differing little. Both species are perennial tufted plants, culms erect. Leaf-sheaths glabrous, leaf-blades flat, margins ciliate. Spikes nodding or curved, with sparse remote spikelets, two spikelets on each rachis node, green or purple-tinged. Lemma is lanceolate and with five nerves. Palea is equal to lemma. Anthers yellow. The most significant difference is the length of the lemma awn, E. sinosubmuticus is only 0-2 mm, and E. breviaristatus is 2-6 mm. In addition, many types of interspecific variations were found in our field studies.

A high chromosome pairing frequency of hybrid F1 can indicate that the parents are closely related [51, 52]. A species represents an independent gene pool in the evolutionary system, and reproductive isolation is the only factor for forming such independent gene pools in organismal evolution [15]. Accordingly, reproductive isolation is the only standard for species identification. In our study, the hybrid F1 of E. breviaristatus and E. sinosubmuticus has a high-frequency bivalent at MI (mean value of 20.62), suggesting that the three genomes of E. breviaristatus and E. sinosubmuticus has high homology, and they are closely related. But the percentage of stained pollen grains of hybrids was 53.90%, and the seed setting rate was 22.46%. This suggests genetic differentiation between the two taxa, leading to a degree of reproductive isolation. Yang et al. [35] reported that E. breviaristatus was a hexaploid with the StYH genome and treated it as Campeiostachys breviaristata (Keng) Y. H. Zhou, H. Q. Zhang et C. R. Yang. Combined with morphological characteristics and the fertility of hybrids, E. sinosubmuticus should be classified into the genus Campeiostachys as the subspecies of E. breviaristatus and renominated as Campeiostachys breviaristata (Keng) Y. H. Zhou, H. Q. Zhang et C. R. Yang subsp. sinosubmuticus (S. L. Chen) Y. H. Zhou, H. Q. Zhang et L. Tan.

Conclusions

Elymus sinosubmuticus is allohexaploid wheatgrass, and the genome composition is StYH. Its morphological characteristics are very similar to E. breviaristatus. Simultaneously, E. sinosubmuticus and E. breviaristatus have a degree of reproductive isolation, and it is reasonable to treat E. sinosubmuticus as the subspecies of E. breviaristatus. Because E. breviaristatus was treated as Campeiostachys breviaristata by Yang et al. [35], therefore, E. sinosubmuticus should be renamed as Campeiostachys breviaristata (Keng) Y. H. Zhou, H. Q. Zhang et C. R. Yang subsp. sinosubmuticus (S. L. Chen) Y. H. Zhou, H. Q. Zhang et L. Tan.

Methods

Plant materials

In our study, Elymus breviaristatus and Elymus sinosubmuticus were collected from the field in Sichuan Province, China, and numbered ZY 17,004 and ZY 17,008 respectively. No permissions were necessary to collect seed samples. Yonghong Zhou and Haiqin Zhang identified the two plant materials. They were used for artificial hybridization, and the materials and F1 hybrids were cultivated in the greenhouse at Hongyuan, Sichuan. The voucher specimens of E. breviaristatus and E. sinosubmuticus were deposited in the Herbarium of Triticeae Research Institute of Sichuan Agricultural University, China (SAUTI). Apart from E. breviaristatus and E. sinosubmuticus, diploid species and relative polyploid species with different genome combinations (StY, StH, StYH) in Triticeae were also applied for phylogenetic analyses. The basic information about these sequences is listed in Additional file 2: Table S1.

Hybridization and meiotic analysis

The hybridization procedure is as follows: after 2–3 days of the emasculation of the female parent, repeat pollinations with the corresponding mature pollens of the male parent were carried out. The female parents were used a plastic bag to isolate the pollen throughout the whole process. In crossing combination, E. breviaristatus was used as the male parent when crossed with E. sinosubmuticus, and E. sinosubmuticus was used as the male when hybridized with E. breviaristatus. The chromosome pairing of pollen grains of hybrids and parents was examined after fixing by Carnoy's Fluid II for 24 h. The mean pairing frequency of hybrids and parents at MI is described by Kimber and Alonso [53]. Mature pollen of hybrids and parents were detected activity after staining with I2-IK solution.

Chromosome preparation and GISH

The roots were collected from adult plants, treated with nitrous oxide for three hours, fixed with 90% glacial acetic acid for 5 min, and kept with 70% alcohol. The chromosome was prepared by drop methods [54]. Using the CTAB method [55] extracted the total genomic DNA from leaves. DNA was labeled using DIG-Nick Translation Kit (Roche, Indianapolis, IN, USA). The green probes were labeled with fluorescein-12-dUTP, and the red probes were labeled with Texas-red-5-dCTP using the nick translation method. Hybridization procedure, detection, and visualization are referred to Han et al. [56]. For monochromatic GISH, the concentration ratio of probe genomic DNA and non-labeled blocking genomic DNA was 1:120 (ng/uL). For double-color GISH, the concentration ratio of probes was 1:1 (ng/uL). Images of GISH were captured by Olympus BX61 fluorescence microscopy (Japan). At least ten metaphase cells for each species were observed. Adobe Photoshop CS6 was used to proceed with the images.

Sequence amplification and phylogenetic analyses

The Acc1 and DMC1 sequences were amplified with primers listed in Table 2. All polymerase chain reactions were amplified in a 50 uL reaction mixture, containing 25 uL 1 × phanta mix buffer, 1 mM dNTP mix, 1 uL DNA polymerase (Vazyme, Nan**g, China), 10 uM of each primer, 200 ng of template DNA, and distilled deionized water to the final volume. Polymerase chain reaction (PCR) products were cloned into the 007VS vector (TSINGKE Biological Technology, Bei**g, China). At least 15 random independent clones were selected for sequencing by Sangon Biological Engineering and Technology Service Ltd. (Shanghai, China). DNA sequences were confirmed through BLAST nucleotide alignment on NCBI database. The multiple sequences were aligned, and manual adjustments were made using the ClustalX [57]. jModelTest 3.0 [58] was used to determine appropriate DNA substitution models and gamma rate heterogeneity. Phylogenetic analyses were conducted using the maximum-likelihood method in PhyML 3.0 [59] and Bayesian inference (BI) in MrBayes version 3.1.2 [60]. Statistical support for nodes in ML analysis was estimated by using 1000 fast bootstrap replicates. Bootstrap support (BS) value < 50% and posterior probabilities (PP) value < 90% was not included in figures.

Table 2 The primers used in this study
Full size table