Introduction
Alfalfa (Medicago sativa ssp. sativa) is a member of the Medicago sativa complex, which includes both the diploid (2n = 2x = 16) and the tetraploid (2n = 4x = 32) forms. The diploid members of the complex include M. sativa ssp. falcata, M. sativa ssp. caerulea and their natural hybrid, M. sativa ssp. hemicycla. The tetraploid subspecies in the complex include M. sativa ssp. sativa, M. sativa ssp. falcata and the tetraploid hybrid M. sativa ssp. varia (Quiros and Bauchan, Reference Quiros, Bauchan, Hanson, Barnes and Hill1988).
Hybridization between taxa of the M. sativa complex is possible even across the ploidy levels (Bingham and Saunders, Reference Bingham and Saunders1974). Gene transfer is more common from the diploid to the tetraploid genome and it usually happens by development of unreduced gametes. Both 2n egg and 2n pollen producers could be used for the direct gene transfer from the wild diploid relatives into the cultivated alfalfa by means of 2x–4x and 4x–2x crosses (Barcaccia et al., Reference Barcaccia, Tavoletti, Mariani and Veronesi2003). Based on more favourable genetic segregation ratios and more robust genetic mapping capabilities, identification of desirable alleles in the diploid germplasm is considerably easier than in the tetraploids (Şakiroğlu and Brummer, Reference Şakiroğlu and Brummer2011). Thus, the use of diploid alfalfa relatives as a source of alleles with higher breeding value is advantageous in alfalfa breeding.
Alfalfa is an important component of sustainable forage systems worldwide. Sensitivity of alfalfa to acidity and aluminium (Al) toxicity in soil limits its productivity and persistence. Yield losses associated with soil acidity are commonly prevented with application of lime fertilizers (Gomes et al., Reference Gomes, Borges, Neves and Fontes2002; Grewal and Williams, Reference Grewal and Williams2003), which are rather expensive (Acharya, Reference Acharya2014). Moreover, liming does not reduce soil acidity in the subsoil, which can substantially reduce the alfalfa yield. Therefore, the most desirable option to improve alfalfa production is the development of plants with Al tolerance (Reyno et al., Reference Reyno, Khu, Monteros, Bouton, Parrot and Brummer2013).
The lack of Al tolerance in primary alfalfa germplasm dictated the need for identifying genes or QTLs (quantitative trait loci) for Al tolerance in Medicago germplasm that could be transferred to cultivated alfalfa (Bouton, Reference Bouton1996; Dall'Agnol et al., Reference Dall'Agnol, Buton and Parrott1996). Genotypes with Al tolerance have been identified among the diploid M. sativa ssp. caerulea (Sledge et al., Reference Sledge, Bouton, Dall'Agnoll, Parrott and Kochert2002) and M. truncatula germplasm (Sledge et al., Reference Sledge, Pechter and Payton2005). Al-tolerant accession M. sativa ssp. caerulea was later used as the female parent in 2x–4x hybridization with M. sativa ssp. sativa (Khu et al., Reference Khu, Reyno, Han, Zhao, Bouton, Brummer and Monteros2013). Although extensive breeding efforts have been made in obtaining the acid-tolerant alfalfa, the progress has been modest in terms of their commercial use (Zhang et al., Reference Zhang, Humphries and Auricht2007; Bouton, Reference Bouton2012). According to the currently available data, only one commercial alfalfa cultivar was registered as specifically suitable for cultivation on acidic soil (Acharya, Reference Acharya2014).
The objectives of this study were: (i) obtaining the tetraploid offspring from 2x–4x M. sativa ssp. caerulea–M. sativa ssp. sativa crosses, by utilizing the occurrence of 2n gametes in natural populations of diploid M. sativa ssp. caerulea; and (ii) development of a breeding strategy for obtaining Al/acid tolerance in alfalfa, using M. sativa ssp. caerulea as a source of tolerance.
Material and methods
Plant material
Accession PI 464724 of wild diploid alfalfa M. sativa ssp. caerulea that has shown tolerance to Al (Sledge et al., Reference Sledge, Bouton, Dall'Agnoll, Parrott and Kochert2002; Narasimhamoorthy et al., Reference Narasimhamoorthy, Bouton, Olsen and Sledge2007) was used as the female parent. Plants of PI 464724 were grown in the experimental field of the Institute of Field and Vegetable Crops in Novi Sad, Serbia, surrounded by the tetraploid M. sativa ssp. sativa. Plants were put in open pollination, aimed at utilizing the production of 2n unreduced gametes in the diploid M. sativa ssp. caerulea in field conditions and favouring the 2x–4x crosses. One M. sativa ssp. caerulea plant survived and produced 75 seeds that were harvested and sown in the greenhouse during the winter.
Cytogenetic analysis
Fertility of plants was estimated by examining male (size and vitality of pollen grains) and female (number of ovules per ovary) gametophytes. Pollen vitality was determined according to the method of Atlagić et al. (Reference Atlagić, Terzić and Marjanović-Jeromela2012), while the number of ovules was determined as in Dattee (Reference Dattee1972). Pollen grains and ovule number were observed with the stereo microscope Stemi 2000 and photographed with the digital camera Power Shot G5. Light microscope (Amplival 30-G048C, Carl Zeiss, Jena; ocular 12.5 × 16 and objective 16/0.40) was used for pollen viability analysis. Microphotographs were made with the video camera CCD Sony DXC151AP, video card and software Win-Fast PVR2. The size of pollen grains was estimated using light microscope. Microphotographs were made with the digital camera (Canon Power shot A80) and the data were analysed using the AxioVision LE software, Rel.4.3.
DNA analysis
Total genomic DNA was extracted from the leaves of 11 potential hybrid plants and M. sativa ssp. caerulea PI 454756 plant, according to the protocol of Somma (Reference Somma, Querci, Jermini and Van den Eade2004). Eleven SSR primer pairs were used to confirm the hybrid nature of the offspring (Table 1).
SSR primers used to confirm the hybrid nature of M. sativa ssp. caerulae × M. sativa ssp. sativa offspring
LG, linkage group; SSR55TD, SSR touchdown PCR profile with annealing at 55°C.
PCR was carried out in a 25-μl reaction volume containing 2.5 µl buffer, 0.2 mM of each dNTP, 0.8 µM of each primer (Metabion international AG, Steinkirchen, Germany), 1.5 units of Taq polymerase (Thermo Scientific, Waltham, MA USA) and 50 ng of DNA. PCR reactions were carried out in the Tpersonal PCR thermocycler (Biometra GmbH, Göttingen, Germany) using the following ‘touchdown’ PCR profiles: an initial denaturation of 2 min at 94°C, followed by a cycle with denaturation at 94°C for 30 s and extension at 72°C for 50 s. The annealing temperature was decreased 0.8°C per cycle in subsequent cycles from 55°C until reaching 47°C and was kept constant for the remaining 30 cycles. PCR ended with a final extension at 72°C for 5 min. Exceptions were the PCR protocol for primers AfCa1, AFCT45, AFCT 60 and MTIC299. These protocols consisted of an initial denaturation at 94°C for 3 min, followed by 40 cycles at 94°C for 30 s, 50°C (MTIC299) or 52°C for 40 s and 72°C for 40 s, with final elongation at 72°C for 5 min. PCR products were separated on the 3% Metaphore agarose (Lonza Group Ltd, Basel, Switzerland) gels containing 0.005% ethidium bromide and visualized under the UV light. The band size was calculated by comparison with GeneRuler 50 bp DNA Ladder (Thermo Scientific, Waltham, MA USA ).
The SSR profiles were scored based on the size (bp) of the fragments amplified across the tested plants, based on the visual inspection of electrophoretic patterns. Genetic distance between the plants was calculated using the equation from Nei (Reference Nei1978). Cluster analysis was performed using the unweighted pair group method with arithmetic mean (UPGMA) in PopGene software version 1.32 (Yeh et al., Reference Yeh, Yang and Boyle1997) and visualized using the TreeView software version 1.6.6. (Page, Reference Page1996). In addition, relationship among tested plants was examined by the Principal Coordinates Analysis (PCoA) implemented in the GenAlex 6.41 software package (Peakall and Smouse, Reference Peakall and Smouse2006), based on the genetic distance calculated for co-dominant markers.
Results
Eleven plants fully developed out of 75 seeds sawn in the greenhouse. The developed plants showed differences in their morphology: plants 2, 3, 9, 10 and 11 were similar to M. sativa ssp. caerulae, being short, with small leaves and thin branches (Fig. 1a), while plants 1, 4, 5, 6, 7 and 8 looked much like M. sativa ssp. sativa, with bigger habitus, larger leaves and thicker branches (Fig. 1b).
Morphology of M. sativa ssp. caerulae × M. sativa ssp. sativa offspring (a) similar to M. sativa ssp. caerulea and (b) similar to M. sativa ssp. sativa.
Cytological analysis showed that all 11 plants were fertile. Pollen viability (Fig. 2a) ranged from 21.45 to 97.09%, while the average number of ovules (Fig. 2b) per plant ranged from 8.80 to 12.29 (Table 2). The size of vital pollen grains varied: plants 2, 3, 9, 10 and 11 had smaller pollen, while plants 1, 4, 5, 6, 7 and 8 had larger pollen grains (Table 2).
Cytological analysis of M. sativa ssp. caerulae × M. sativa ssp. sativa offspring (a) vitality of pollen grains (vital – red, non-vital – green) and (b) ovules in ovary.
Pollen grain size and vitality, number of ovules per ovary in offspring of M. sativa ssp. caerulae × M. sativa ssp. sativa
Eleven SSR primer pairs were used to confirm the hybrid nature of M. sativa ssp. caerulae × M. sativa ssp. sativa offspring (Table 3). In 10 plants, the presence of three alleles was detected after amplification with at least one primer pair. Plant number 7 had four alleles after multiplication with primers AFCT 45, AFCT60 and MTIC 299. Four alleles were detected after multiplication with MTIC 299 in plants 1 and 3, as well as in plant 11 after multiplication with primer AfCa1. Amplifications with the tested primers in plant 10 resulted in only one or two amplification products.
Multiple alleles obtained by SSR primers in offspring of M. sativa ssp. caerulae × M. sativa ssp. sativa
Based on the cluster analysis, the hybrid plants were grouped into two major clusters. The first cluster consisted of plants 1, 4, 6, 7 and 8; plants 2, 3, 5, 9, 10 and 11 were in the second cluster, while M. sativa ssp. caerulea plant was separated (Fig. 3a).
Relationships between the M. sativa ssp. caerulae (12) and M. sativa ssp. caerulae × M. sativa ssp. sativa offspring (1–11) revealed by (a) UPGMA and (b) PCoA calculated from the SSR data.
PCoA was performed as a type of ordination method in order to check the results of the cluster analysis. The first principal coordinate accounted for 26.6% of the total variation and separated the offspring in two groups (Fig. 3b), identically as in the dendrogram. The second principal coordinate accounted for 21.42% of the total variation and indicated that the plants most related to M. sativa ssp. caerulea were 9, 10 and 11.
Discussion
Interspecific hybridization between cultivated species of alfalfa and their wild relatives offers a potential opportunity for introgression of desired traits from the wild to the cultivated forms (Barcaccia et al., Reference Barcaccia, Tavoletti, Mariani and Veronesi2003). Utilization of diploid alfalfa genetic resources is of special importance, since identification of desirable alleles in the diploid germplasm is considerably easier than in the tetraploids.
In the present study, 11 hybrid plants were obtained through hybridization between the diploid M. sativa ssp. caerulae and the tetraploid M. sativa ssp. sativa. In terms of their morphology, the hybrid plants segregated in comparison with the parent species: plants 1, 4, 5, 6, 7 and 8 in the sativa type and plants 2, 3, 9, 10 and 11 in the caerulae type. The occurrence of characteristics from both progenitors in the hybrids indicates that the hybrid plants inherited genomes from both parental species used for crossing. Since the correct identification of hybrid recognition plant by morphological traits is often be difficult, the molecular approach may be useful for the assessment of the hybrid status. The SSR markers used in this study clearly identified the hybrid origin of the tested plants, identifying three or four alleles for certain SSR markers. In plant 10, all tested markers had one or two alleles, indicating that the plant was either the result of M. sativa ssp. caerulae self-pollination or the hybrid with still undetected part of the M. sativa ssp. sativa genome. The results of cluster and PCoA, based on SSR data, agreed with morphological observations of hybrid offspring, segregating plants in the caerulae type from plants in the sativa type. The only exception was plant 5 that was placed in the sativa type according to the morphological observations, but was put among the caerulae-type plants by both the cluster analysis and the PCoA. The UPMGA clearly indicated a separate position of M. sativa ssp. caerulae plant compared with hybrids, but PCoA indicated a different pattern. It is known that the cluster analysis groups individuals in such a way that most similar individuals are first grouped together and progressively similar group are merged (Hair et al., Reference Hair, Black, Babin, Anderson and Tatham2006). PCoA results in two- or three-dimensional scatter plot of individuals in such a way that the genetic distances among the individuals are reflected in the distance between the points in the plot (Mohammadi and Prasanna, Reference Mohammadi and Prasanna2003). Therefore, it is not unusual that results of PCoA has varying degree of effectiveness in capturing cluster structure which, in our case, resulted in the different placement of M. sativa ssp. caerulea plant on dendrogram and PCoA plot.
The presented results support the findings of Havananda et al. (Reference Havananda, Brummer and Doyle2011), who pointed out that there was a potential for gene flow in virtually all combinations of the Medicago subspecies both within and between ploidies. Our results are in agreement with the previous findings considering gene flow in the M. sativa gene pool between the diploid and the tetraploid subspecies (Bingham and Saunders, Reference Bingham and Saunders1974; Veronesi et al., Reference Veronesi, Mariani and Bingham1986), but they also indicate that this process can happen spontaneously in natural, field conditions. Cytogenetic analysis showed that the hybrid plants were fertile, which is essential for introgression of valuable traits from wild relatives into cultivated germplasm (Barcaccia et al., Reference Barcaccia, Tavoletti, Mariani and Veronesi2003).
Having in mind that M. sativa ssp. caerulea accession PI 464724 has shown tolerance to Al (Sledge et al., Reference Sledge, Bouton, Dall'Agnoll, Parrott and Kochert2002; Narasimhamoorthy et al., Reference Narasimhamoorthy, Bouton, Olsen and Sledge2007), it is to be expected that a certain amount of genes for Al tolerance was transferred to some hybrid plants. Therefore, the hybrid offspring was crossed with the tetraploid M. sativa ssp. sativa to produce F2 plants and possible Al-tolerant progenies. The resulting seed will be planted in the field in soil with low pH, where it will be tested for Al tolerance and agronomic performance in field conditions. The surviving plants will be further multiplied in the field for development of Al-tolerant alfalfa germplasm.
The information generated in this study can be used for alfalfa acid/Al tolerance improvement and understanding of alfalfa's genetic base, as well as for development of cost-reduced and efficient methods for creating improved alfalfa germplasm. Future results with strategies based on conventional and molecular marker breeding efforts could lead towards the development of more tolerant alfalfa cultivars and more prosperous utilization of this crop on Al-contaminated soils.
Acknowledgements
This research was funded by the Ministry of Education, Science and Technological Development of the Republic of Serbia, project BTR 31024. The authors would like to thank Germplasm Resources Information Network (USDA) for kindly providing the M. sativa ssp. caerulea accession PI 464724 and Dr Bojana Stanić for English language editing
