Introduction

Numerous studies have been focused on analysis of somaclonal variation in vitro (Larkin et al. 1984; Lee and Phillips 1988; Pershina et al. 2003). The initial reports were concerned with determining changes in the number, structure, and morphology of the chromosomes, such as chromosome banding patterns (Larkin et al. 1984; Lee and Phillips 1988), and later studies examined variations in genomic DNA caused by mutations, the loss and rearrangement of genes, gene silencing and transposons or reverse transposons in subculture (Shaked et al. 2001; Polanco and Ruiz 2002; James and Reiner 2007). However, little was known about the content and genotype of cultured cells of allohexaploid wheat (Triticum aestivum L.) that had been grown for different periods of time.

Wheat (T. aestivum L., 2n = 42) is an important cereal. The genetic variability for some important traits, such as high quality, diseases and stress resistance, is limited in the cultivated wheat germplasm. Related or distant species of wheat are an important reservoir of useful genes (Liu et al. 2005). There is thus an urgent need to broaden the wheat gene pool by introgressing genes for such traits from diverse cereal species. This kind of alien introgression line created can increase the genetic variation in wheat species. However, the sexual route with low crossability in achieving this limited the transfer of such genes (Wang et al. 2004; Liu et al. 2005). Somatic hybridization provides the possibility of overcoming sexual incompatibility of wheat with remote cereals (Xia et al. 2003). This technique has been used with some success as a means of introgressing genes from related grass and cereal species (Xia et al. 2003; Wang et al. 2005; Zhou and Xia 2005). However, when the donor species was only distantly related to wheat, fusions often failed to progress beyond undifferentiated calli or albino plants (Xiang et al. 1999; Li et al. 2001; Yue et al. 2001). In an attempt to overcome the difficulty of regenerating hybrids between distantly related species, a three-parent fusion system has been elaborated (Xiang et al. 2003a, 2004; Xu et al. 2003), based on two complementary wheat cell lines T1 and T2. The former is a non-totipotent cell line growing in long-term cell suspension, and the latter are totipotent cells taken from embryogenic calli (Xiang et al. 2003b). The third “parent” is an exotic donor species, the protoplasts of which are irradiated before fusion. Using this system, Xiang et al. (2003a) were able to regenerate a set of putative wheat/oat (Avena sativa L.) somatic hybrids and employed a combination of cytological and genetic assays to demonstrate the presence of a variable number of wheat/oat recombinant chromosomes. However, the contents and genotypes of the two wheat cell lines and their somatic fusion products with oat were unknown. This was critical for understanding the mechanism of the regeneration of hybrid plants.

In this work, our goal was to investigate the chromosome elimination and variation of T1 and T2, as well as the relationship between the genetic complementation of T1, T2, oat and somatic hybrid plant regeneration.

Materials and methods

Plant materials

Wheat protoplasts derived from two types of cell lines T1 and T2 of T. aestivum L. cv. Jinan 177, and oat protoplasts came from embryogenic calli A. sativa L. cv. Mapur. The T1 were harvested from a non-regenerable long-term (>10 years) cell suspension, and the T2 from 2-year-old embryogenic calli with regeneration frequency of 75% (number of calli which can differentiate to plants/number of calli analysed, %) (Xiang et al. 2003a). The oat cell line was 1.5-year-old embryogenic calli (Xiang et al. 2003a). The T1 and T2 lines were subcultured on MB liquid and solid media (Xia and Chen 1996) with 2.0 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D). Protoplasts were isolated from T1 suspension after 3 days of subculture, T2 and oat embryogenic calli were isolated 7 days after subculture according to the method of Xia and Chen (1996). After washing in 0.6 M mannitol and 5 mM CaCl2, oat protoplasts were transferred onto 3.5 cm Petri dishes in a thin layer and irradiated with UV light at an intensity of 300 μW/cm2 for 1 min before they were fused with T1 and T2 (Xiang et al. 2003a). T1 and T2 protoplasts were combined at the ratio of 1:1 and mixed in equal volumes with UV-treated protoplasts of oat. The fusion was carried out according to the improved PEG method (Xia and Chen 1996). When the fusion clones grew to small cell lines of 1.5–2.0 mm in diameter, they were transferred to the MB medium with 1 mg/L 2,4-D for proliferation. After subculture for 1–2 months, proliferating calli from the cell lines were moved to the MB medium containing 0.5 mg/L IAA and 0.5 mg/L zeatin for regeneration. Regenerated plantlets were transferred to seedling-strengthening MB medium containing 1–2 mg/L multi-effect triazal (MET) and 0.5 mg/L NAA for strengthening and rooting. All the regenerated plants from four hybrid clones resembled wheat, but only plants of clone no. 94-1 can grow into soil, and then tassel normally (Xiang et al. 2003a). In this experiment, a lager number of roots from hybrid plants of clone no. 94-1 in different growth stages were analysed. Accessions of Triticum urartu Thum, Aegilops speltoides Tausch and Aegilops tauschii (Cross.) Schmal were kindly provided by the Quality and Resource Institute of the Agriculture Science Academy of China.

Karyotype analysis and in situ hybridization

Chromosome spreads were obtained from T1 and T2 cells and from the root tips of the regenerants and germinating seedlings, as described by Xiang et al. (2003b). Karyotypes of cv. Jinan 177 were derived from a sample of ten mitotic metaphase cell spreads. Karyotype classification followed the method of Sears (1969) and was based on the data for Chinese Spring wheat from Gill (1987). The parameters of the karyotypes were based on ten metaphase cell spreads. The arm ratios and the relative distances of the chromosomes from the hybrid cells selected were analysed. For in situ hybridization (ISH) purposes, genomic DNA of cv. Jinan 177, oat, T. urartu, Ae. tauschii, Ae. speltoides and the hybrid regenerants was isolated following Doyle and Doyle (1990). The pSc119.2 and pAs1 plasmids contain, respectively, B and D genome-specific repetitive sequences (Mukai et al. (1993). These were labelled for use as fluorescence ISH (FISH) probes with digoxigenin-11-dUTP, using a nick translation kit, according to the manufacturer’s instructions (Boehringer Mannheim, Germany), as described elsewhere (Wang et al. 2005). Total oat genomic DNA was labelled in the same way to generate a genomic ISH (GISH) probe. The GISH and FISH methodology followed Wang et al. (2005), the former using a ratio of 1:50 labelled oat genomic DNA to unlabelled wheat genomic DNA, and the latter 1:50 labelled pSc119.2 to unlabelled genomic (T. urartu + Ae. tauschii) DNA, or 1:50 pAs1 to unlabelled genomic (T. urartu + Ae. speltoides) DNA. In sequential GISH and FISH experiments (pAs1/pSc119.2 and GISH/pAs1/pSc119.2), the initial signal was rinsed before rehybridization with the subsequent probe.

Microsatellite (SSR) analysis

DNA extracted from T1 cells, T2 calli and leaf of somatic hybrids and parental lines (Doyle and Doyle 1990) was used as template for the analysis of allelic constitution at 101 SSR loci, following Röder et al. (1998). The marker set covered the centromeric and subtelocentric regions of all the chromosomes present in T1 and T2, and the locations selected based on the previous GISH, FISH and karyotype analysis of hybrid. The relative distance (%; the distance from the centromere to SSR locus checked/the length of the arm involved × 100%) was measured according to the genetical map of Röder et al. (1998).

Results

The chromosome content of T1 and T2 and cv. Jinan 177

Of the 138 T1 cells analysed, 119 (86.2%) contained 22–25 chromosomes (Fig. 1a), while 104/124 (83.9%) of the T2 cells had 34–38 chromosomes (Fig. 1b). Some T1 and T2 cells included telocentric chromosomes, acrocentric chromosomes and dicentric chromosomes (see Suppl. Table 1, Fig. 1a, b). One or two chromosome fragments were present in ~87% of both the T1 and T2 cells. The karyotype of cv. Jinan 177 conformed to that of standard bread wheat (Gill 1987; Suppl. Table 2, Fig. 1c).

Fig. 1
figure 1

Mitotic metaphase spreads of T1 (a), T2 (b), cv. Jinan 177 (c) and the karyotype of cv. Jinan 177 (d). Telocentric and subtelocentric chromosomes indicated by an arrow, dicentric chromosomes by an arrowhead

Full size image

The genome content of T1 and T2 cells

FISH preparations based on hybridization with pSc119.2 and pAs1 succeeded in defining the B and D genome content of T1 and T2. From a sample of 136 T1 cells, 82 (~60%) contained seven B genome, and 83 (~61%) contained eight D genome chromosomes (Table 1, Fig. 2a, b); similarly, of the 95 T2 cells analysed, ~60% had ten B and ~54% 12 D genome chromosomes (Table 1, Fig. 2c, d). The range in the number of B and D genome chromosomes in T1 was, respectively, 6–8 and 7–9, and in T2, 9–11 and 10–12. Chromosomes 2B, 3B, 7B, 3D and 6D were present in most T1 and T2 cells, but 4B, 5B and 4D were rare. Chromosome abnormalities were present in, respectively, 27 and 39% of the T1 B and D genome chromosomes and 31 and 36% of T2 chromosomes (Table 2). Therefore, many eliminations and rearrangements including duplication of B- and D-genome chromosomes in T1 and T2 presented in the calli subculture. Chromosome arm ratios and relative lengths were used to infer the A genome content of T1 and T2 cells. Comparing the statistics of karyotype data of T1, T2 with cv. Jinan 177, it was revealed that A genome chromosomes of T1 and T2 were distributed mainly in 4A and 5A, whereas 3A was the lowest. About 61% of T1 cells contained nine A genome chromosomes, while ~57% of the T2 cells contained 12 (Table 1, Fig. 2a, b). The range in A genome chromosome number in T1 was 7–11, and in T2 11–15. Non-standard A genome chromosomes were present in 22% of T1 and 16% of T2 cells (Tables 1, 2).

Table 1 Frequency of pSc119.2 and pAs1 hybridization signals in T1 and T2 cell lines
Full size table
Fig. 2
figure 2

FISH profile of the mitotic metaphase chromosomes. a and b B and D genome chromosomes of T1 identified by their pSc119.2 and pAsI patterns. c and d B and D genome chromosomes of T2 identified by their pSc119.2 and pAsI patterns. Scale bars 10 μm

Full size image
Table 2 Chromosome constitution of the T1 and T2 lines
Full size table

Distinguishing between DNA deletion and modification in T1 and T2

Simple sequence repeats genotyping was employed to distinguish between DNA deletion and modification in T1 and T2 as a result of cell culture. Of the 58 loci assayed, 36 exposed genetic polymorphism between T1 and T2 and T (cv. Jinan 177), and the remainder were non-informative (Table 3, Fig. 3). Of the 36 informative markers, 13 showed same patterns among T, T1 and T2, 15 between T1 and cv. Jinan 177 and 25 between T2 and cv. Jinan 177 (Table 3). The rates of DNA deletion were 55.6% (20/36) in T1, such as in Xgwm160 and Xgwm428 (Fig. 3), and 19.4% (7/36) in T2, such as in Xgwm335 (Fig. 3) and Xgwm273, and 16.7% (6/36) in both T1 and T2 (Table 3). This indicated that the rate of DNA deletion of T1 was higher than that of T2. The ratio of DNA modification was 8.3% (3/36) in T1 with examples such as Xgwm389 (Fig. 3) and Xgwm160 (Fig. 3). In T2 the ratio of DNA modification was 13.9% (5/36), with examples of Xgwm389, Xgwm335 and Xgwm428 (Table 3, Fig. 3). B-genomic DNA was eliminated at a faster rate than D-genomic DNA, which was deleted faster than A-genomic DNA in T1 and T2. Anyway, the rate of DNA deletion was greater in T1 than in T2, while the rate of DNA modification was greater in T2 than in T1 (Table 3). Therefore, there were many instances of DNA variation at different positions in T1 and T2, and within the calli of subcultures of wheat.

Table 3 SSR profiles of cv. Jinan 177 and T1 and T2 cell lines
Full size table
Fig. 3
figure 3

SSR profiles of cv. Jinan 177 (T) and cell lines T1 and T2, based on primers directed to loci on wheat chromosome arms 1BL, 3BS, 4AL, 5BL and 7DL. a, b: cv. Jinan 177 allele; n novel allele

Full size image

Introgression of oat chromatin into wheat

About 71% of the cells of the hybrid regenerants contained 46–48 chromosomes (mean 47.2), with ~58% containing oat chromosome introgression segments distributed on six wheat chromosomes. Based on sequential ISH experiments, it was possible to identify 12–16 of the B genome and 11–13 of the D genome chromosomes (Table 4, Fig. 4a, b). By comparing their karyotype with that of cv. Jinan 177, 17 A genome chromosomes and three chromosomal abnormalities were detected (Table 4). Almost the whole wheat genome was represented in the hybrid nuclei, along with duplicated, deleted or re-arranged chromosomes. The GISH/FISH analysis identified the presence of oat chromatin on chromosomes 3B and 4B (Fig. 4a, c), and 3D, 6D and 7D (Fig. 4b, c). A further introgression site was identified indirectly on chromosome 4A (Fig. 4c). The arm locations of the six introgression sites were 4AL, 3BS, 4BL, 3DS, 6DL and 7DL (Table 5, Fig. 4). An analysis of the arm ratios of the introgressed chromosomes, and the relative distances between the centromere and the introgression break point are given in Table 5.

Table 4 Chromosome content of a bread wheat/oat somatic hybrid derivative
Full size table
Fig. 4
figure 4

Sequential FISH and GISH preparations of metaphase chromosomes of a bread wheat/oat somatic hybrid derivative. FISH profiling to identify B (a) and D (b) genome chromosomes. c GISH profile to identify chromosome segments introgressed from oat. Scale bars 10 μm

Full size image
Table 5 Localization of introgressed oat chromatin in a wheat/oat somatic hybrid derivative
Full size table

Of the 58 SSR loci used to genotype the hybrid regenerants, 32 were amplified in all of the samples. Of the 32 informative markers, only 14 (43.8%) identified differences between hybrid plants and their parents. The alleles from both parents were present at nine of above 14 loci, only the oat allele was amplified at two loci, while at the remaining three loci, the alleles were biparental and novel (Table 6). The relative distance (%) between the centromeres and the SSR loci were measured. SSR bands were located on 4AL, 3BS, 4BL, 3DS, 6DL and 7DL chromosome arms, and the relative distances (%) of fragments ranged from 3.10 to 76.36% (Table 6), in agreement with the results of sequential GISH and FISH (Table 5). Oat alleles were also amplified by primers targeting SSR loci on chromosome arms 3AS, 3AL, 6AL and 7AL, none of which were identified by GISH as sites of introgression (Table 6). This observation is suggestive of the occurrence of many sub-microscopic introgression events.

Table 6 SSR position on the wheat chromosome of hybrid and the amplification marker
Full size table

Discussion

By using the pSc119.2 with high repetitive sequences of B-genome from rye (Secale cereale L.), and pAs1 with insertion of repetitive sequences of D-genome from Ae. tauschii for two-colour FISH, Mukai et al. (1993) established the ideogram of B- and D-genomes and a pair of 4A chromosomes of Chinese Spring wheat. Using N-, C-banding, GISH, FISH and SSR markers in combination with karyotypes data, heterogeneric chromatin in many wheat hybrids were localized into the wheat chromosomes (Jiang et al. 1993; Nagy et al. 2002; Malysheva et al. 2003; Silkova et al. 2006). Our previous result had also confirmed that FISH analysis with pSc119.2 and pAs1, in combination with karyotype data and GISH, could differentiate all of the A-, B- and D-genome chromosomes of the cultivar wheat Jinan 177 and localize small donor chromosomes in somatic hybrid of Jinan 177 with Agropyron elongatum (Wang et al. 2005). Here, we have extended this analysis to the three-cell fusion system (Xiang et al. 2003a), which has uncovered chromosome content and genotype of two wheat cell lines and of their somatic hybrid plant with oat.

Chromosome loss and variation in the T1 and T2 cell lines

Variation in chromosome number and the formation of chromosomal aberrations are common in cells cultured in vitro over many generations (Larkin et al. 1984; Song et al. 2000). Here we have shown that the T1 cell line has lost about 18, while T2 has lost only about eight chromosomes (Table 1, Fig. 2), as would be expected given that T1 has been in culture for a much longer period than T2. The evidence is that particular chromosomes tend to be more highly prone to loss, while others are lost rather rarely, as has been noted in other studies (Lee and Phillips 1988; Doğramaci-Altuntepe et al. 2001). The former group includes chromosomes 4A, 5A, 3B, 7B, 3D, 6D, and the latter chromosomes 3A, 6A, 4B, 5B and 4D (Table 2). As for the chromosomal abnormalities of T1 and T2, the frequencies in D-, B-groups were higher than that of the A-group, with the order of D- > B- > A-group in both T1 and T2 (Table 2).

Hybrid chromosome number and regeneration ability

Previous studies have indicated that the totipotency of a hybrid cell line was dependent on the number of chromosomes present, so those having a chromosome number close to the bread wheat somatic number of 42 tended to be more fit than those which were either hypo- or hyper-ploid (Xia et al. 1996; Xiang et al. 2003a, 2004). The majority of the regenerants from the wheat/oat fusion had a chromosome number in the range 46–48, far less than the sum of the three fusion parents’ chromosomes (93–108). Thus, there must have been a massive and rapid phase of chromosome elimination following the fusion, particularly involving the oat chromosomes (aided by the pre-fusion irradiation treatment), but also involving chromosomes inherited from T1 and T2.

Genetic complementation of T1, T2 and A. sativa with hybrid plant regeneration

When bread wheat is pollinated by either Hordeum bulbosum or maize, the pollen parent’s chromosomes are eliminated during the first few mitotic divisions of the hybrid zygote, a process which has been exploited for the production of dihaploids (Inagaki and Tahir 1990; Chen et al. 1999). Even in compatible wide crosses, which occur in nature and are maintained by polyploidization, a spectrum of genetic and cytological events leads to various modifications (Kashkush et al. 2003; Birchler et al. 2005; James and Reiner 2007; Ma and Gustafson 2008). Thus, it is unsurprising that such events also affect the somatic hybridization process. The T1 cell line has lost the capacity to regenerate, while T2 cells remain totipotent (Xiang et al. 2003a, 2004; Xu et al. 2003). T2 protoplasts cannot divide in vitro, while T1 cells readily form non-regenerable calli. In combination, these two lines are able to contribute both the ability to grow in vitro and totipotency. Thus, it is plausible to suggest that the loss of totipotency of the T1 cell line is due to the absence of a critical chromosome(s) and that this absence can be complemented in the fusion nucleus by the contribution of the T2 line genome, where the critical chromosome(s) is still represented. A similar argument can be made for the ability to divide in in vitro culture. In addition, GISH/FISH pattern indicated the introgression of oat chromatin to the 4AL, 3BS, 4BL, 3DS, 6DL and 7DL of the hybrid wheat (Table 5, Fig. 4). It is interesting that 10 SSR loci near to the introgression position showed the same profile between the hybrid and oat (Table 6). Therefore, we suggested that hybrid plant regenerated through genetic complementation of T1 and T2 and oat. Somatic hybrid derivatives have also been recovered from other combinations with wheat, including oat, foxtail millet and maize (Xiang et al. 2003a, 2004; Xu et al. 2003), which may be explained by the genetic complementation between the T1 and T2 cell lines and the donor species. But no hybrid progenies were produced in these plants from the “Triple parents” (Xiang et al. 2003a, 2004; Xu et al. 2003). In our early reports, hybrid cells from high capacity for regeneration suspension of wheat cv. Jinan 177 with wheatgrass DNA were fertile and heredity stable (Wang et al. 2005). Thus, we suggest that besides the genetic complementation, some other factors also affect the fertileness and genetic stability of these distant hybrid plants. This is worthy of further research (Fig. 5).

Fig. 5
figure 5

SSR profiles of cv. Jinan 177 (lane T), oat (lane O) and a bread wheat/oat somatic hybrid derivative (lane H), based on primers directed to loci on wheat chromosome arms 3AS, 4AL, 3BS, 4BL, 6DL and 7DL. Oat alleles indicated by an arrow, wheat alleles by a thin arrowhead, novel alleles by a full arrowhead

Full size image