Cytological Evaluations of Advanced Generations of Interspecific Hybrids between Allium cepa and Allium fistulosum Showing Resistance to Stemphylium vesicarium

Interspecific crossing is a promising approach for introgression of valuable traits to develop cultivars with improved characteristics. Allium fistulosum L. possesses numerous pest resistances that are lacking in the bulb onion (Allium cepa L.), including resistance to Stemphylium leaf blight (SLB). Advanced generations were produced by selfing and backcrossing to bulb onions of interspecific hybrids between A. cepa and A. fistulosum that showed resistance to SLB. Molecular classification of the cytoplasm established that all generations possessed normal (N) male−fertile cytoplasm of bulb onions. Genomic in situ hybridization (GISH) was used to study the chromosomal composition of the advanced generations and showed that most plants were allotetraploids possessing the complete diploid sets of both parental species. Because artificial doubling of chromosomes of the interspecific hybrids was not used, spontaneous polyploidization likely resulted from restitution gametes or somatic doubling. Recombinant chromosomes between A. cepa and A. fistulosum were identified, revealing that introgression of disease resistances to bulb onion should be possible.


Introduction
Stemphylium leaf blight (SLB), caused by Stemphylium vesicarium (Wallr.) Simmons, is a serious fungal disease of bulb onion (Allium cepa L.) occurring worldwide, which may result in 100% losses of the bulb crop [1]. SLB has also been reported in garlic [2], asparagus [3], sunflowers [4], pear [5], radish [6], and tomato [7]. Research has been undertaken to identify sources of resistance to SLB in Allium species and determine its genetic basis [8]. Allium fistulosum L. is a source of desirable traits for improvement of the bulb onion [9] and shows resistance to Stemphylium vesicarium which may be conditioned by a single dominant gene [8]. In 2001, hybrids between A. cepa and A. fistulosum were produced to initiate introgression of resistance to SLB into the bulb onion [8]. Although the first interspecific hybrids between A. fistulosum and A. cepa were obtained in 1935 [9], hybrids are highly sterile, and numerous attempts to transfer beneficial traits from A. fistulosum into the bulb onion have not been successful. The first fertile advanced backcross plants between A. cepa and A. fistulosum were obtained by Hou and Peffley [10]; however, no commercial cultivar of the bulb onion has ever been developed with a desirable trait from A. fistulosum.
Genomic in situ hybridization (GISH) is a reliable technique for the identification of parental genomes at the chromosomal level and for the monitoring of alien gene introgression via interspecific hybrids [11][12][13]. However, the technique is often not used in breeding because of its labor intensity and low-throughput. Molecular markers are an effective method for tracking of valuable traits allowing acceleration of the breeding process. However, molecular markers provide information only across chromosomal regions in linkage disequilibrium with the target trait. A huge number of molecular markers are required to analyze chromosomal introgression from one species to another. GISH provides information at the chromosome level and does not require any knowledge about genome sequences and can confirm copy numbers of homologous chromosomes in polyploid hybrids, which is difficult using molecular markers.
In this paper we present the results of a cytological study of advanced generations derived from hybrids between A. cepa × A. fistulosum that show resistance to Stemphylium vesicarium.

Plant Material
Seeds of parental species (A. cepa and A. fistulosum) and advanced generations derived from interspecific hybrids [8] were provided by the Asian Vegetable Research and Development Center in Taiwan (Table 1). A landrace ('Tuwel') of Allium fistulosum from Indonesia (accession AF468) was used as a male parent and A. cepa cultivar 'Rouge de Tana' (TA207) as the female parent to produce the interspecific hybrid, which was self-pollinated to the S 5 generation (AVON1275). Backcross generations were produced using AVON1275 and the open-pollinated onion cultivar 'Arka Niketan' (AC464), developed by the Indian Institute of Horticultural Research. After backcrossing to Arka Niketan, embryos were rescued [14] and resulting plants were evaluated for Stemphylium leaf blight (SLB) resistance after artificial inoculation (described below). Backcross progenies from the same family that showed resistance to SLB were intercrossed. Plants were grown in pots in greenhouses at RGAU-MTAA (Russia) or the University of Wisconsin (USA).

Chromosome Preparations
Mitotic chromosomes were prepared from young root meristems using the squash method according to Khrustaleva and Kik [15] with slight modifications. Young root tips were pretreated overnight with an aqueous saturated solution of 1-bromnaphtalene at 4 • C, fixed in 3:1 (v/v) ethanol-acetic acid for at least 1 h at room temperature and then stored at −20 • C, rinsed four times in distilled water, and finally incubated in 10 mM citrate buffer (pH 4.8) containing 0.1% (w/v) Pectolyase Y-23 (Kikkoman, Tokyo, Japan), 0.1% (w/v) Cellulase Onozuka R-10 (Yakult Co. Ltd., Tokyo, Japan) and 0.1% (w/v) Cytohelicase (Sigma-Aldrich, France) for 50 min at 37 • C. The macerated root tips were spread by dissecting and squashing in a drop of 45% acetic acid.

Analysis of Fertility
Pollen morphology was studied using an acetocarmine staining method [16] from flowering plants from AVON 1275, AVON 1290, AVON 1291, AVON 1501, and AVON 1502. The other hybrid lines did not flower under our greenhouse conditions. We analyzed five flowers per accession when flowers were at peak anther maturity. Two anthers from each flower were squashed in 1% acetocarmine on separate microscope slides and analyzed using phase−contrast microscopy with ocular ×12 and objective ×10 magnification.

Genomic DNA Isolation and Probes Preparation
Genomic DNA was isolated from young leaves of the parental species (A. cepa TA207and AC464, and A. fistulosum AF468) according to the protocol of Rogers and Bendich [17]. Genomic DNA of A. fistulosum was sonicated to 1-3 kb fragments and used for the labeled probe preparation. DNA from A. cepa was sonicated to 200-400 bp fragments and used as the blocking DNA. Probe DNA was labeled with digoxigenin (DIG)-11-dUTP by nick-translation (Roche, Diagnostics Gmbh, Mannheim, Germany).

Microscopy, Image Analysis, and Karyotyping
GISH preparations were visualized using a fluorescent microscope, Zeiss Axio Imager (Carl Zeiss MicroImaging, Jena, Germany), conjugated with black−white sensitive digital camera Axiocam, and with Axio Vision software v. 4.6.3. The final optimization of images was performed using Adobe Photoshop (Adobe Inc., San Jose, CA, USA). Karyotype analysis and identification of individual chromosomes with fluorescent signals were performed according to bulb onion nomenclature [18] and previously published karyotypes of closely related Allium species [19]. The relative position of the recombination site on the relevant chromosome arm was a ratio between its arm length and distance from centromere to the recombination point.

Cytoplasmic Evaluations
Genomic DNA was isolated from the pooled leaf tissue of all accessions and cytoplasms were classified using high-resolution melting (HRM) of an indel in the chloroplast accD gene [20]. Controls included previously isolated genomic DNA from A. cepa (N-cytoplasmic inbred B1750B and S-cytoplasmic B1750A), A. fistulosum, and A. galanthum.

Stemphyllium Screening
The AVON accessions listed in Table 1 were developed in fields under natural heavy disease pressure at the World Vegetable Center in Shanhua, Taiwan. Seven of the accessions (AVON 1275, 1278, 1282, 1283, 1284, 1290, and 1291), the cultivar 'Green Banner' (A. fistulosum), and four cultivars of bulb onion were evaluated for SLB in greenhouses/mist chambers at Seminis Vegetable Seeds in Deforest, WI. Plants were grown in Fafard 4P ® mix (Sun Gro Horticulture, Agawam, MA, USA) in trays with 5-cm cells in a greenhouse at 24 to 27 • C with 12 h lights. Four replications of 18 plants of each accession were sprayed three times at 8, 9, and 10 weeks after sowing with 1 × 10 5 conidia of S. vesicarium per mL in 0.01% tween-20. Plants were then placed into a mist chamber to maintain leaf wetness for 48 h, and then returned to the greenhouse. Eight days after the third inoculation, disease severity in each replication was visually rated by four individuals using a scale of 1 to 5, where 1 = no symptoms to 5 = severe leaf blight or dead plants. Disease severity ratings were averaged over replications. RStudio was used for statistical analyses [21]. One-way ANOVA was calculated based on four replications of each accession, and the least significant differences were calculated using RStudio.

Stemphylium Leaf Blight Evaluations
During development, the interspecific hybrid, derived accessions (Table 1), and cultivars of A. fistulosum and A. cepa were planted in the field at the World Vegetable Center (Shanhua, Taiwan) and subjected to natural disease pressure by S. vesicarium. In these fields, AVON 1275 and derived accessions (Table 1), as well as the original A. fistulosum parent (AF468), appeared resistant to SLB, while all cultivars of A. cepa appeared susceptible. In greenhouse evaluations, there were significantly different (P < 0.001) disease severity reactions (DSRs), which ranged from 1.3 to 2.3 for A. fistulosum and derived accessions, whereas cultivars of A. cepa had significantly higher DSRs ranging from 2.8 to 4.6 ( Table 2). Because A. fistulosum, AVON 1275, and many derived lines had significantly higher resistance to SLB than A. cepa, resistance from A. fistulosum appears to show some level of dominance over susceptibility of the bulb onion (Table 2).

Chromosome Complement of Amphidiploid Backcross Progenies
Amphidiploid AVON 1275 was backcrossed as the female parent with A. cepa. GISH analysis of AVON 1278 (S 5 BC 1 ) showed that the hybrids were allotetraploid with 16 chromosomes of A. cepa and 16 chromosomes of A. fistulosum (Figure 1b) and no recombinant chromosomes. Karyotype analysis (Figure 1b') showed the presence of complete diploid sets of both parental species: 2n = 4x = 32 (16C + 16F).
GISH analysis of AVON 1282, 1290, 1501, and 1502 (all S 4 BC 1 families) showed that the hybrids possessed 16 chromosomes of A. cepa and 16 chromosomes of A. fistulosum (Figure 1c-f) with no recombinant chromosomes. Karyotype analysis (Figure 1c'-f') showed the presence of complete diploid sets of both parental species: 2n = 4x = 32 (16C + 16F). These results indicate that progenies were not derived from backcrossing to bulb onion, but from self-pollination of the amphidiploid parent.

Backcross Progenies with Recombinant Chromosomes
GISH analysis of AVON 1291 (S 4 BC 1 ) showed that the hybrids were allotetraploid, possessing 16 chromosomes from A. cepa, 15 chromosomes from A. fistulosum, and one recombinant chromosome between A. cepa and A. fistulosum (2n = 4x = 32) (Figure 3a). Karyotype analysis (Figure 3a') showed the presence of complete diploid set of A. cepa (8 pairs of homologous chromosomes), 7 pairs of A. fistulosum homologous chromosomes, and one pair of A. fistulosum chromosome 2, which contained one recombinant chromosome with a centromeric region belonging to A. fistulosum and one non-recombinant chromosome (16C + 15F + 1F/C). The recombination site was located on the long arm of chromosome 2. The relative position of recombination site from the centromere was 83.6% ± 0.6 (SD, n = 8). GISH analysis of AVON 1504 (S 4 BC 1 ) showed that the hybrids were allotetraploids possessing 16 chromosomes from A. fistulosum, 14 chromosomes from A. cepa, and 2 recombinant chromosomes between A. cepa and A. fistulosum (2n = 4x = 32) (Figure 3b). Karyotype analysis (Figure 3b') showed the presence of complete diploid sets of A. fistulosum, 7 pairs of the A. cepa homologous chromosomes and one pair of recombinant chromosome 7 with the centromeric region from A. cepa (16F + 14C + 2C/F). The recombination site is located in the short arm of chromosome 7, and the relative positions of recombination sites on both chromosomes were the same at 49.7% ± 0.3 (SD, n = 5). These results clearly demonstrate recombination between chromosomes from A. cepa and A. fistulosum, supporting larger backcrossing efforts towards introgression of beneficial traits into the bulb onion.

Pollen Fertility
Pollen fertility of the advanced generations from interspecific hybrids between A. cepa and A. fistulosum was evaluated using acetocarmine stained pollen. The plants were male−fertile, with fertility ranging from 60 to 95% (Table 3). Plants from AVON 1291 which possessed one recombinant chromosome 2 showed 93% stainable pollen, revealing that the recombinant chromosome did not significantly reduce male fertility.

Cytoplasms of Plants Derived from Interspecific Crosses between A. cepa and A. fistulosum
Cytoplasms of plants from advanced generations from interspecific hybrids were determined using high-resolution melting (HRM) of an indel in the chloroplast accD gene [20]. This marker distinguishes normal (N) male−fertile cytoplasm from male−sterile (S) cytoplasm of onion, as well as the cytoplasms of A. fistulosum and A. galanthum (Figure 4). None of the accessions originating from the original interspecific cross possessed the cytoplasm of A. fistulosum. Most interspecific hybrids between A. cepa and A. fistulosum have been produced using A. fistulosum as the female parent, which limits introgression due to deleterious nucleo-cytoplasmic interactions [22]. In theory, chromosome regions carrying disease or pest resistance(s) in these accessions should avoid this deleterious interaction because they possess N cytoplasm of A. cepa.

Discussion
GISH clearly distinguished chromosomes from A. cepa and A. fistulosum despite the presence of similar DNA sequences in their genomes, such as a subtelomeric tandem repeat showing 80% identity [23,24]. Additionally, retrotransposons are abundant in the Allium genomes [25] and are dispersed along entire chromosomes, which may contribute to cross-hybridization. GISH analyses of interspecific hybrids between A. cepa and A. fistulosum revealed that spontaneous chromosome duplication produced allotetraploids. These results agree with Levan [26], who reported polyploid F 2 plants from crosses between A. cepa and A. fistulosum without artificial chromosome duplication. S 5 plants of AVON1275 were allotetraploids with complete diploid chromosome sets from both parents and were used as the female parent for backcrossing with A. cepa. Most of the progenies from the backcross generations were allotetraploids (2n = 4x = 32) with complete diploid sets from both parental species (16C + 16F). Equal chromosome numbers from both parents indicates that progenies were from self-pollination and not from backcrosses to A. cepa. The same observation was reported by Budylin et al. [27] who completed GISH analyses of advanced generations of hybrids between A. cepa and A. fistulosum showing resistance to downy mildew. These authors revealed spontaneous polyploidization and absence of predominantly A. cepa chromosomes in backcross generations. The absence of backcrossing in diploid F 1 BC 3 hybrids between A. cepa and A. fistulosum was also reported by Hou and Peffley [10]. Our observations indicate that the amphidiploid plants of S 4 BC 1 lines AVON 1282, 1290, 1501, and 1502 were derived from the selfing of the amphidiploid female parent as opposed to backcrossing.
Plants from AVON 1283 (S 4 BC 1 ) and 1284 (S 3 BC 1 ) possessed a complete diploid set of chromosomes from A. cepa and seven pairs of A. fistulosum chromosomes, with one pair of chromosome 5 missing (16C + 14F − 2Ch5). Because both AVON 1283 and 1284 are resistant to SLB, this indicates that SLB resistance from A. fistulosum is not carried on chromosome 5.
Plants from AVON 1291 (S 4 BC 1 ) and AVON 1504 (S 4 BC 1 ) possessed recombinant chromosomes. In AVON1291, one recombinant chromosome 2 was observed and its homolog was non-recombinant chromosome 2 of A. cepa. It is likely that a crossing over event occurred between A. cepa and A. fistulosum chromosomes at the allotetraploid stage. Homeologous recombination has been detected in both recently formed natural polyploids and synthetic interspecific polyploids [28]. Recombination between chromosomes of A. cepa and A. fistulosum has been documented [10,13,29], and intergenomic pairing and recombination have been documented for allotetraploid hybrids of sour cherry (Prunus cerasus), Solanum lycopersicum × Solanum lycopersicoides, and for allopolyploids in asterids [30]. Two recombinant chromosome 7 were revealed in AVON 1504 with the centromeric region belonging to A. cepa. The size of the A. cepa region within both recombinant chromosomes 7 was the same, at 49.7% ± 0.3 (SD, n = 5), which indicates the presence of two sister chromatids that did not segregate during the second meiotic division and remained in the same gamete. This suggests that spontaneous polyploidization in the interspecific hybrids between A. cepa and A. fistulosum was caused by second division restitution, resulting in 2n-gamete formation. Alternatively, the presence of two identical recombinant chromosomes 7 could be the result of the somatic doubling of the complement in the zygote. The recombinant chromosome could have been formed during meiosis in the interspecific hybrid, either in a pollen mother cell or in a macrospore mother cell and underwent somatic doubling. Another interesting observation from the GISH analysis of AVON 1504 was the lack of genetic material from A. cepa in the distal region on the short arm of the recombinant chromosomes 7, which has been replaced with genetic material of A. fistulosum (Figure 3b,b',b"). It is known that 5S rDNA is located on the short arm of chromosome 7 in both species [31,32] and the lengths and sequences of the coding 5S rRNA gene segments are conserved among Allium species [33]. Based on this observation, we conclude that this part of the A. cepa genome can be replaced by A. fistulosum and function in the cytoplasm of A. cepa.
Spontaneous chromosome doubling in interspecific hybrids between A. cepa and A. fistulosum was observed in our materials and has been previously reported [26]. This could be caused by fusion of 2n gametes, what led to the formation of allotetraploid progeny possessing complete diploid sets from both parental species. However, the formation of restitution gametes is rare in A. cepa [34]. Even if we assume the formation of restitution gametes, the probability of fusion of 2n-egg with 2n-sperm is unlikely. An alternative hypothesis may be that facultative apomixes occurred, which has been reported for several species of genus Allium [35,36]. Spontaneous polyploidization could be also caused by cytomixis, which involves the transfer of nuclear material from one cell to another via cytomictic channels [37]. Cytomixis has been observed in the pollen mother cells of A. cepa and A. fistulosum [38,39].

Conclusions
Spontaneous polyploidization and no evidence of backcrossing were revealed in backcross progenies from interspecific hybrids between A. cepa and A. fistulosum. The mechanism of spontaneous polyploidization is not clear, and it will be of great interest to determine if it results from somatic or meiotic chromosome doubling. Clear evidence for recombination between the chromosomes of A. cepa and A. fistulosum supports continued research to introgress beneficial traits, such as SLB resistance, from A. fistulosum into the bulb onion.