Introduction

Improving crop yield in a sustainable and efficient manner is an unavoidable global objective needed in order to face the increasing demand and consumption of the human population. Regarding wheat, only a small part of the diversity available is being used for this purpose; therefore, unlocking genetic resources and broadening the genetic basis are a great challenge (Longin and Reif 2014). Wheat genetic diversity may be efficiently exploited through heterosis (hybrid vigor) displayed in hybrids. Hybrid production requires blocking self-fertilization, which can be achieved by different means, including genetic male sterility, photo-thermo-sensitive male sterility, chemical agents, or cytoplasmic male sterility (Chen and Liu 2014; Kim and Zhang 2018). Cytoplasmic male sterility (CMS) is caused by the genomic conflict between the mitochondrial and the nuclear genomes, which results in the production of non-functional pollen. Sterility can be reverted by nuclear genes known as restorer-of-fertility (Rf) genes (Bohra et al. 2016). These strategies have been successfully applied in other important crops for hybrid production but not in wheat, where hybrid programs represent a minor fraction of the overall production (Singh et al. 2014; Whitford et al. 2013).

The CMS source msH1 in bread wheat (Triticum aestivum L.; 2n = 6x = 42; AABBDD) was first described by Martín et al. (2008b). The CMS system uses the cytoplasm of Hordeum chilense accession H1 as a source of male sterility, while restoration of fertility is associated with the addition of the short arm of chromosome 6Hch (6HchS) from this accession. In further work, lines with new reorganizations of 6HchS chromosome, also restoring fertility, were obtained. First, the common wheat line T650, harboring the double translocation T6HchS·6DL in H. chilense cytoplasm, was developed and its fertility restoration ability was confirmed (Martín et al. 2009). Later, a new acrocentric recombined chromosome including fragments from 1HchS and 6HchS was identified (Castillo et al. 2014) as a new source of fertility restoration in the msH1 system.

Compared with hexaploid bread wheat, durum wheat (Triticum turgidum; 2n = 4× = 28; AABB) has a narrower genetic base due to its tetraploid nature. Broadening its genetic base may be achieved by using alien species, through whole chromosome additions, substitutions, and translocations or transference of small segments, among others (Gupta 2016). Also, hybrid breeding in durum wheat has shown to have high potential, but a cost-effective system of hybrid seed production is still essential for the development of hybrid varieties (Gowda et al. 2010). Furthermore, durum wheat may show more heterotic potential compared to common wheat due to its lower ploidy level. A putative system to develop a CMS source in durum wheat has been investigated by using the alloplasmic combination of Triticum longissimum cytoplasm with durum wheat nucleus, which results in non-viable progeny. The combination of the two genes scsti and Vi is necessary not only to restore male fertility but also to recover seed viability and plant vigor (Maan 1992; Maan et al. 1999; Simons et al. 2003).

In this work, we aim to determine the viability of the msH1 system as a new source of CMS for the production of hybrids in durum wheat. Adapting the msH1 system to durum wheat requires the availability of three lines: (1) a male sterile mother line (alloplasmic durum wheat in H. chilense cytoplasm), (2) a fertile maintainer line (euplasmic durum wheat), and (3) a fertile restorer line (euplasmic durum wheat with fertility restorer ability). The objectives of this work were to develop the alloplasmic and restorer lines and to determine their potential in the development of a hybrid durum wheat system.

Material and methods

Plant material

The plant material used for this study is shown in Table 1. Line T622 is a ‘Langdon’ durum wheat where the chromosome pair 6A is substituted by the pair 6D from common wheat (Joppa and Williams 1988). Line T650 is an alloplasmic common wheat ‘Chinese Spring’ in H. chilense cytoplasm with the Robertsonian translocation of the short arm of chromosome 6HchS from H. chilense with the long arm of chromosome 6DL in homozygosis (Martín et al. 2009). HT47 (AABBHchHch) is an hexaploid tritordeum (Martin and Sanchez-Monge 1982), an amphiploid derived from the cross of line H8 from H. chilense (female parent) with durum wheat MAGH72 (male parent) from the crossing block of CIMMYT 1986-87, obtained by duplication of the hybrid with colchicine. This hexaploid tritordeum obtained by using H. chilense has recently been given the species name of ×Tritordeum martinii (Pujadas Salvá 2017).

Table 1 Description of the plant material used in this study
Full size table

All plants were grown in greenhouse conditions maintaining a 0–30 °C temperature without supplemental night light.

Viability and fertility scoring

Viable plants were considered those that survived the plantlet stage. Seed set was used as the criterion for assessing male fertility or sterility. Fertility was scored by counting total number of grains per total number of flowers in the spikes. Plants with at least one grain were considered as fertile.

Cytological observations

For chromosome counting, root tips of 1-cm length were collected from germinating seeds and pre-treated for 4 h at 25 °C in an aqueous colchicine solution (0.05%). They were fixed in 3:1 solution of absolute ethanol:glacial acetic acid (v/v) and stained by the conventional Feulgen technique.

Genome in situ hybridization

Chromosome preparation and genome in situ hybridization (GISH) were carried out as described previously (Rey et al. 2018). Aegilops tauschii was used as a probe to label wheat D genome. H. chilense and A. tauschii genomic DNA were labeled with biotin-16-dUTP and digoxigenin-11-dUTP, using the biotin-nick translation mix and the DIG-nick translation mix, respectively (Sigma, St. Louis, MO, USA) according to the manufacturer’s instructions. Images were taken using a Leica DM5500B microscope equipped with a Hamamatsu ORCA-FLASH4.0 camera and controlled by Leica LAS X software v2.0.

Molecular marker selection

Leaf tissue was harvested from plantlets for DNA isolation. Genomic DNA was extracted following the CTAB protocol with slight modifications (Murray and Thompson 1980).

All markers used for selection were first tested with parent and control lines to confirm their location and utility in the genetic background used. Marker Bmac316 was used for selection of the short arm of chromosome 6Hch (Ramsay et al. 2000; Rodríguez-Suárez and Atienza 2012). Absence of 6HchL was corroborated with marker K03014 (Hagras et al. 2005; Nasuda et al. 2005). Primer pair FEH_1310-F/FEH_1685-R (Zhang et al. 2008) amplifies three fragments of different sizes which served us to identify 6AS, 6BS, and 6DS arms. Chromosome arm 6DL was tagged with Xgdm98 (Pestsova et al. 2000; Röder et al. 1998), and chromosome arm 6AL with marker Xgpw3029 (http://wheat.pw.usda.gov/cgi-bin/graingenes). Cytoplasm origin was confirmed with the chloroplast marker ccSSR4 (Martín et al. 2008a).

All PCR amplifications were carried out using MyTaq™ DNA polymerase (Bioline, London, UK) following manufacturer’s instructions with an annealing temperature of 60 °C. Amplification products were resolved in agarose gels and visualized with Safeview™ Nucleic Acid Stain (NBS Biologicals Ltd, Cambridgeshire, England).

Results

Obtaining a durum wheat alloplasmic line

To transfer the cytoplasm from H. chilense into durum wheat, repeated substitution backcrosses were performed. As cytoplasm is maternally inherited in Triticeae species (Kihara 1951), the tritordeum line HT47 (AABBHchHch) harboring the cytoplasm from H. chilense line H8 was used as female parent in a cross with durum wheat ‘Don Pedro’. The resulting hybrid was backcrossed to durum wheat three times until all H. chilense chromosomes from the amphiploid were eliminated. Somatic chromosome counting was used to follow this chromosome elimination process. The cytoplasm was also checked with molecular marker ccSSR4 (Martín et al. 2008a), confirming that it was maternally inherited as expected. A durum wheat (2n = 4× = 28) alloplasmic line (cytoplasm from H. chilense) was finally obtained and named (H8) T711 (Fig. 1), being completely male sterile under different environmental conditions (growth chamber, greenhouse, and open field).

Fig. 1
figure 1

Breeding scheme followed to obtain the alloplasmic and restorer durum wheat lines and to evaluate the T6HchS·6DL translocation effect and the restoration ability of the 6HchS chromosome. Chromosome arms in bold mean positive amplification of the molecular markers used for their identification; x 3BC means three backcrosses to durum wheat

Full size image

Transferring the T6HchS·6DL translocation into durum wheat

The common wheat alloplasmic line T650 was used to transfer the T6HchS·6DL translocation into T622, the ‘Langdon’ durum wheat substitution line 6D (6A). Reciprocal crosses were made, being only self-fertile the progeny obtained by using the euplasmic line T622 as the female parent. After several cycles of selfing and selection for the presence of 6HchS in consecutive years, a plant with 28 chromosomes and harboring the T6HchS·6DL translocation in hemizygosis was identified and named T622T650-13. To reach this point, all progenies generated in each selfing step were cytological screened by somatic chromosome counting, selecting for chromosome number close to 2n = 28. Molecular markers Bmac316, FEH_1310-F/FEH_1685-R, and Xgdm98 were used to track 6HchS, 6DS, and 6DL, respectively. The multiband pattern shown by FEH_1310-F/FEH_1685-R served us to tag the 6AS arm and Xgpw3029 for 6AL.

Both molecular markers and chromosome number identified T622T650-13 as a nullisomic line for chromosome pair 6A, harboring the translocation T6HchS·6DL in hemizygosis with a single copy of chromosome 6D. In the self-progeny, three expected genotypes were identified: homozygous for the T6HchS·6DL translocation, heterozygous for T6HchS·6DL and 6D, and homozygous for 6D (6A). The new line with 2n = 28, homozygous for the T6HchS·6DL translocation and nullisomic for chromosome 6A, was named T855 (Fig. 1). GISH was carried out on T855 line to confirm the presence of two copies of the T6HchS·6DL translocation (Fig. 2). T855 was able to produce viable pollen and seed set was observed; however, fertility was very variable, probably due to the physiological difficulty of handling with a double translocation (see “Discussion”).

Fig. 2
figure 2

Genomic in situ hybridization to root-tip metaphase cells from line T855 (2n = 28; homozygous for the T6HchS·6DL translocation). H. chilense genomic DNA probe is shown in magenta. Aegilops tauschii genomic DNA probe, which labels the D genome, is shown in green. Blue DAPI staining shows the A and B chromosomes

Full size image

Fertility restoration ability of T6HchS·6DL in alloplasmic durum wheat

To test the fertility restoration ability of the translocation T6HchS·6DL in durum wheat background, line T855 was used as male parent to pollinate the male sterile line (H8) T711 previously obtained (Fig. 1). Durum wheat line T855 is homozygous for the T6HchS·6DL translocation and nullisomic for chromosome 6A. Line (H8) T711, used as the female parent, is the alloplasmic sterile durum wheat with the cytoplasm of the H. chilense line H8. As the cytoplasm is maternally inherited, the progeny of this cross should be all sterile unless the presence of 6HchS restores fertility.

The self-progeny obtained from this cross consisted of 34 plants (Online Resource 1). The genomic composition of the progeny was cytologically determined by chromosome counting. The majority of plants had 28 chromosomes (27 plants). Chromosomal abnormalities were also observed: an extra chromosome (five plants), two extra chromosomes (one plant), and 27 chromosomes plus an extra chromosome arm (one plant). These plants were discarded.

Plants with 28 chromosomes were grown in greenhouse conditions and were characterized with molecular markers. All plants harbored at least one copy of chromosome 6A coming from the parental line (H8) T711, confirming that it was a real cross. Molecular marker ccSSR4 was also used to confirm that the cytoplasm was that of H8. Segregation was observed for the molecular markers tagging 6HchS and 6DL. Table 2 shows the frequencies of the four observed types. For each type, total number of plants, number of viable plants (those surviving the plantlet stage), and number of fertile and sterile plants are shown.

Table 2 F2 segregation of the progeny obtained from the cross (H8) T711 × T855. Fertility and survival rates are calculated by grouping the plants for the presence (+) or absence (−) of chromosome 6HchS. Genotypes are inferred from molecular marker amplification patterns
Full size table

The 12 plants without the translocation (6HchS− 6DL−) were all sterile. The only plant characterized as 6HchS− 6DL+ (type II, Table 2) was also sterile. This plant may be a reorganization of chromosome arm 6DL, probably in translocation with 6BS or 6AS, but it was not further investigated. Considering all plants without chromosome 6HchS as a group, the fertility rate was 0, since none of the plants without 6HchS produced a single seed.

Conversely, taken together all plants harboring chromosome 6HchS, the fertility rate was 87.5%, being only sterile one of the plants harboring the translocation (type III, Table 2). Additionally, two plants characterized as 6HchS+ 6DL− (type IV) were obtained being both fertile. These results clearly show that only the presence of 6HchS gives rise to fertile plants and that no fertile plants are recovered without the 6HchS chromosome arm.

Regarding survival rate, almost all plants without 6HchS (92.31%) reached the mature stage. However, only around the 57% of plants with 6HchS were able to surpass the plantlet stage.

Morphological traits were also evaluated in greenhouse conditions. Table 3 shows the mean values for plant height, number of spikes, and anthesis date of plants with and without 6HchS. Although slight differences can be observed between groups, they are not significantly different for plant height (p = 0.289), number of spikes (p = 0.163), or anthesis date (p = 0.364).

Table 3 Mean values and standard errors for the morphological traits: plant height, number of spikes, and anthesis date (DAS, days after sowing) of (H8) T711 × T855 F2 progeny, grouped by the presence (+) or absence (−) of 6HchS
Full size table

Effect of the T6HchS·6DL translocation in durum wheat background

To analyze the effect of the T6HchS·6DL translocation itself in euplasmic durum wheat (nucleus and cytoplasm from durum wheat), one heterozygous plant derived from the self-progeny of T622T650-13 (T622T650-13-1) was selfed and its progeny was analyzed (Fig. 1). A total of 46 plants were obtained, all of them with 28 chromosomes (Online Resource 2). The three expected genotypes were identified by using molecular markers as homozygous for the T6HchS·6DL translocation, heterozygous for T6HchS·6DL and 6D, and homozygous for 6D (6A). Total number of plants, number of viable plants (those surviving the plantlet stage), number of fertile and sterile plants, and survival and fertility rates within each class are shown in Table 4.

Table 4 Segregation of the self-progeny obtained from the euplasmic heterozygous line T622T650-13-1
Full size table

A Mendelian transmission of the translocated chromosome was observed (χ2 = 2.348, p = 0.309), indicating that gamete viability was not dependent on the presence or absence of the translocation; however, not all the plants reached the mature stage. The survival rate was the lowest for plants with the translocation in homozygosis. Besides, none of the four viable plants was fertile (see “Discussion”). Heterozygous plants showed the best fertility rate and a survival rate similar to that of plants with the substitution 6D (6A).

Discussion

The wild barley H. chilense has shown to have a high potential in wheat breeding. Due to its high compatibility with Triticum species, fertile and stable amphiploids can be obtained. This facilitates the transfer of H. chilense traits to wheat such as resistance to Septoria tritici, abiotic stress tolerance, endosperm storage proteins (Atienza et al. 2005; Martín et al. 2000), carotenoid content (Rodríguez-Suárez and Atienza 2012; Rodríguez-Suárez et al. 2014), sterilizing cytoplasms, and fertility restorer genes (Castillo et al. 2014; Martín et al. 2008a, 2009).

In the development of the alloplasmic durum wheat line, the strategy of repeated backcrosses of tritordeum with durum wheat was followed, instead of using alloplasmic wheat lines previously obtained in our group (Atienza et al. 2007; Martín et al. 2008b). In the H. chilense chromosome elimination process, some alloplasmic plants with remaining H. chilense chromosomes may be fertile or partially fertile, indicating the presence of a restorer of fertility. Indeed, the 6HchS chromosome and the acrocentric chromosome Hchac were both identified as restorers of fertility in the msH1 system by these means (Martín et al. 2008b, 2010). Moreover, new chromosome reorganizations, as the substitution of chromosome 6DS by 6HchS (Martín et al. 2009), arise in tritordeum when crossed with wheat (Cabo et al. 2014; Delgado et al. 2017). In the present work, tritordeum was used to obtain the alloplasmic durum wheat line hoping to induce spontaneous reorganizations involving 6HchS, which could be used as restorer lines. Although none of these reorganizations have been identified in this work, this is a promising strategy to follow up in the future in the search for new restorer lines.

The new alloplasmic durum wheat line (H8) T711 was completely sterile under different environmental conditions (growth chamber, greenhouse, and open field). The same occurs with the type I plants derived from the cross (H8) T711 × T855 (Table 2). These plants are expected to recover a complete genome dotation AABB, as ordinary durum wheat, but in an alloplasmic background. In both cases, plants harbor the cytoplasm from H. chilense line H8. In common wheat, several H. chilense cytoplasms have been used for the development of alloplasmic lines. Interestingly, not all the cytoplasms cause male sterility in common wheat. For instance, H. chilense lines H7 and H46 give rise to fertile alloplasmic common wheats (Atienza et al. 2007; Rodríguez-Suárez et al. 2011). Both H1, used as the sterilizing cytoplasm in the common wheat CMS msH1, and H8 belong to the same H. chilense subspecific taxa (Castillo et al. 2010; Patto et al. 2001) and are both useful for the development of alloplasmic lines in CMS systems in wheat.

Regarding survival rates, viability of alloplasmic durum wheat plants seems not to be compromised by H. chilense H8 cytoplasm, neither in the (H8) T711 line nor in the type I and type II plants (Table 2). It is worth mentioning the incompatibility shown by other durum CMS systems, like that with the T. longissimum cytoplasm, where two nuclear genes (scsti and Vi) are needed to overcome this nuclear-cytoplasm interaction (Maan 1992; Maan et al. 1999). The durum wheat lines with H. chilense cytoplasm obtained in this work did not show any need of additional genes for improving nuclear-cytoplasm compatibility, only for restoring male fertility. Nevertheless, more H. chilense cytoplasms may be assayed, and agronomic traits should be compared, to find optimum alloplasmic male sterile lines for the development of a future CMS system in durum wheat.

The contrasting types I and III obtained from the cross (H8) T711 × T855 (Table 2) give information about the effect of the T6HchS·6DL translocation. Both groups are alloplasmic, being the only difference between them, the presence of this translocation: Type I plants are complete AABB durum wheats, whereas type III plants harbor the translocation in hemizygosis. The significant reduction of plant viability (Table 2) indicates that the translocation T6HchS·6DL has a deleterious effect in this durum wheat background. Indeed, no plants with the T6HchS·6DL translocation in homozygosis were obtained in this cross.

The presence of the translocation T6HchS·6DL has an adverse effect in plant fitness that also affects fertility and survival rates in the euplasmic plants analyzed (Table 4). Line T855 is able to produce viable pollen, but not in all circumstances. In the progeny from T622T650-13-1 (Table 4), the four plants homozygous for the T6HchS·6DL translocation were not able to produce viable pollen. In this genetic background, the negative effect of the translocation is more evident in terms of fertility and survival, due to their chromosome configurations which are, in addition, nullisomic for chromosome 6A.

Durum wheat has a lower buffering ability compared to hexaploid wheat. Differential impact of an alien translocation at the hexaploid and tetraploid levels is well known (Ceoloni and Jauhar 2006). The 2BS·2RL centric-break fusion translocation reduces plant vigor and fertility in homozygous compared to that in heterozygous carriers in durum wheat (Friebe et al. 1999). Similarly, no homozygous lines containing an entire 3RS arm harboring the Sr27 stem rust resistance gene could be obtained in durum wheat (McIntosh et al. 1995). Adverse selection was also suffered by the Ag. elongatum translocation carrying the alien Sr26 gene in a durum wheat background, being well tolerated by the hexaploid wheat (McIntosh et al. 1995).

Regarding restoration of fertility, the present work demonstrates the ability of chromosome 6HchS to restore male fertility of durum wheat alloplasmic lines, as it occurs in common wheat (Castillo et al. 2014, 2015; Martín et al. 2008b, 2009). In absence of this chromosome arm (types I and II, Table 2), no fertile plants are recovered. There is therefore potential for the use of this system in the production of hybrid durum wheat. Nevertheless, as discussed above, the translocation itself is detrimental for the plant fitness in this genomic background. Considering the low tolerance shown by durum wheat to the T6HchS·6DL translocation, particularly in homozygosis, other combinations need to be explored in order to be used in a restorer line. Future development of the system will be aimed at obtaining translocations of 6HchS with other chromosomes (6AL and 6BL), obtaining small introgressions of the 6HchS, and the addition of the acrocentric chromosome described for the msH1 system (Castillo et al. 2014). Also, different durum wheat genomic backgrounds will be used, and other H. chilense accessions will be tested as cytoplasm donor.

Final remarks

In summary, it has been demonstrated that H. chilense cytoplasm is useful for the development of alloplasmic male sterile lines in durum wheat. Apart from stable sterility, no other effects derived from an incompatible nucleus-cytoplasm interaction are evident. It has been also demonstrated that the chromosome arm 6HchS restores fertility in male sterile alloplasmic durum lines. The msH1 system can therefore be transferred to durum wheat, as both sterility and restoration are possible in absolute terms. However, more effort needs to be put on the search for new restorer lines with no deleterious effects in plant fitness.