Species Origin of Genomic Factors in Nicotiana nudicaulis Watson Controlling Hybrid Lethality in Interspecific Hybrids between N. nudicaulis Watson and N. tabacum L

Hybrid lethality is expressed at 28°C in the cross Nicotiana nudicaulis×N. tabacum. The S subgenome of N. tabacum has been identified as controlling this hybrid lethality. To clarify the responsible genomic factor(s) of N. nudicaulis, we crossed N. trigonophylla (paternal progenitor of N. nudicaulis) with N. tabacum, because hybrids between N. sylvestris (maternal progenitor of N. nudicaulis) and N. tabacum are viable when grown in a greenhouse. In the cross N. trigonophylla×N. tabacum, approximately 50% of hybrids were vitrified, 20% were viable, and 20% were nonviable at 28°C. To reveal which subgenome of N. tabacum was responsible for these phenotypes, we crossed N. trigonophylla with two progenitors of N. tabacum, N. sylvestris (SS) and N. tomentosiformis (TT). In the cross N. sylvestris×N. trigonophylla, we confirmed that over half of hybrids of N. sylvestris×N. trigonophylla were vitrified, and none of the hybrids of N. trigonophylla×N. tomentosiformis were. The results imply that the S subgenome, encoding a gene or genes inducing hybrid lethality in the cross between N. nudicaulis and N. tabacum, has one or more genomic factors that induce vitrification. Furthermore, in vitrified hybrids of N. trigonophylla×N. tabacum and N. sylvestris×N. trigonophylla, we found that nuclear fragmentation, which progresses during expression of hybrid lethality, was accompanied by vitrification. This observation suggests that vitrification has a relationship to hybrid lethality. Based on these results, we speculate that when N. nudicaulis was formed approximately 5 million years ago, several causative genomic factors determining phenotypes of hybrid seedlings were inherited from N. trigonophylla. Subsequently, genome downsizing and various recombination-based processes took place. Some of the causative genomic factors were lost and some became genomic factor(s) controlling hybrid lethality in extant N. nudicaulis.


Introduction
Hybrid lethality is a postzygotic barrier of reproductive isolation, which can occur in intergeneric, interspecific, and intraspecific crosses and causes death of some hybrids in Nicotiana species, Arabidopsis thaliana, tomato (Solanum pimpinellifolium6S. lycopersicum) and lettuce (Lactuca saligna6L. sativa) [1][2][3][4][5]. Bomblies and Weigel [6] concluded that hybrid lethality is a barrier to gene flow in Arabidopsis thaliana and tomato (Solanum pimpinellifolium6S. lycopersicum) and acts as an autoimmune-like response. Hybrid lethality is a mechanism playing a crucial role in the evolution of plants and is an important obstacle in plant breeding programs. In Nicotiana, hybrid lethality shows five specific phenotypes depending on the cross combination [7,8].
In crosses between 11 of the 20 Suaveolentes species and N. tabacum, the S subgenome of N. tabacum is involved in hybrid lethality at 28uC [18][19][20][21][22][23]. Furthermore, causative gene(s) of hybrid lethality in N. tabacum are encoded on the Q chromosome belonging to the S subgenome [24][25][26][27][28]. Based on these results, Tezuka [2] speculated that many species of section Suaveolentes share the same gene(s) triggering hybrid lethality by interaction with gene(s) on the Q chromosome. Recently, the first linkage map for N. tabacum was constructed by Bindler et al. [29] with SSR markers. Tezuka et al. [30] used this map to determine that the gene(s) responsible for hybrid lethality map to the region between SSR markers PT30342 and PT30365 of linkage group 11, corresponding to the Q chromosome.
Although investigations into the causative gene(s) in N. tabacum for hybrid lethality have been extensive, investigations on genomic factor(s) in wild species of section Suaveolentes have not. Although a single dominant gene, designated HYBRID LETHALITY A1 (HLA1), controls hybrid lethality in N. debneyi (a wild species of section Suaveolentes) by interaction with gene(s) on the Q chromosome of N. tabacum, the gene location is unknown [31].
Section Repandae exists in North America and consists of four allopolyploid species: N. nudicaulis, N. repanda, N. stocktonii and N. nesophila. Section Repandae is considered to have originated approximately 5 million years ago from hybridization between ancestors of extant species of Nicotiana section Sylvestres (N. sylvestris) and Trigonophyllae (N. obtusifolia, formerly known as N. trigonophylla), the maternal and paternal progenitors, respectively, and their evolutionary relationships are well known [17,32,33]. N. nudicaulis is sister to, and morphologically distinct from, the three remaining species, diverging from them 2-3 million years ago, whereas N. stocktonii and N. nesophila diverged from N. repanda one million years ago; section Repandae is genetically distant from N. tabacum, even though they share the same maternal progenitor, N. sylvestris [34,35].
In a previous study [36], hybrid lethality was observed in crosses of N. repanda6N. tabacum. Furthermore, Kobori and Marubashi [37] concluded that the T subgenome of N. tabacum encodes the causative gene(s) for hybrid lethality in the cross N. repanda6N. tabacum. On the other hand, the S subgenome of N. tabacum is related to hybrid lethality at 28uC in hybrids between N. nudicaulis and N. tabacum [38,39]. In contrast to research on the subgenome controlling hybrid lethality in N. tabacum, no studies of the subgenome of wild species involved in hybrid lethality have been reported to our knowledge. In previous studies [40,41], N. sylvestris (maternal progenitor of N. nudicaulis) was crossed with N. tabacum, and viable hybrids were produced. Here, we added the cross between N. trigonophylla (paternal progenitor of N. nudicaulis) and N. tabacum to infer the likely origins of genomic factor(s) in this wild species.
Apoptosis is a distinct type of cell death, and it is characterized in animals by condensation of chromatin, fragmentation of nuclei, cytoplasmic reduction, and fragmentation of DNA [42][43][44]. In hybrid seedlings of N. nudicaulis6N. tabacum expressing hybrid lethality, features of apoptosis such as chromatin condensation, nuclear fragmentation, and fragmentation of DNA were evident [38].
In the present study, we examined several phenotypes of hybrid seedlings in the cross N. trigonophylla6N. tabacum, and half of the hybrid seedlings showed vitrification. Vitrification frequently occurs and is a serious problem in plant micropropagation. Vitrified plants have translucent stems, brittle, water-soaked, elongated and curled leaves [45]. This report is the first to confirm that vitrified hybrid seedlings appear in interspecific crosses and that nuclear fragmentation is typically associated with vitrification. Because nuclear fragmentation also progresses during expression of hybrid lethality in the cross between N. nudicaulis and N. tabacum [38], we concluded that there is a relationship between vitrification and hybrid lethality. Furthermore, we found that the causative gene(s) in N. tabacum of vitrification are encoded on the S subgenome involved in hybrid lethality. These results also suggest that vitrification has a relationship to hybrid lethality. On the other hand, when N. sylvestris was crossed with N. tabacum, hybrid lethality was not observed [41,42]. Based on these results, we inferred that the genomic factor(s) in N. nudicaulis leading to hybrid lethality in the cross between N. nudicaulis and N. tabacum is likely derived from N. trigonophylla.

Plant Materials
The seeds of N. trigonophylla Dunal (2n = 24, TrTr), N. tabacum  Ovule Culture Because in conventional crossing between N. trigonophylla and N. tabacum, it is difficult to obtain hybrid seedlings, ovule culture was carried out 5 days after pollination and seedlings obtained as described by Chung et al. [46]. Flowers of N. trigonophylla used as the maternal parent were emasculated before flowering and pollinated with fresh N. tabacum pollen. Five days after pollination, flowers of N. trigonophylla were collected and the sepals, petals, and styles were removed. The ovaries were surface-sterilized with 70% ethanol for 30 s with 5% sodium hypochlorite for 10 min, then rinsed three times with sterilized water. The ovary walls were peeled to expose the placenta with intact ovules. All of the ovules were excised and cultured on Petri dishes containing 8 ml 1/2x MS medium [47] supplemented with 8% sucrose and 0.25% Gelrite, pH 5.8, at 28uC under continuous illumination (30 mmol photons s 21 m 22 ).

Test Tube Pollination and Ovule Culture
Because no fertile seeds were obtained in situ from the cross N. trigonophylla6N. sylvestris or N. sylvestris6N. trigonophylla, and ovaries of N. trigonophylla are too small for test tube pollination and ovule culture, we carried out test tube pollination and ovule culture in the cross N. sylvestris6N. trigonophylla as described [48], and hybrid seedlings were obtained.

Interspecific Crosses of N. trigonophylla6N. tomentosiformis
Conventional crossing was carried out as follows: flowers of N. trigonophylla used as the maternal parent were emasculated before flowering and then pollinated with fresh N. tomentosiformis pollen. F1 seeds (N. trigonophylla6N. tomentosiformis) were soaked in 0.05% gibberellic acid (GA 3 ) solution for 30 min, sterilized with 5% sodium hypochlorite for 15 min and then rinsed three times with sterilized water. Sterilized F1 seeds were sown on Petri dishes containing 8 ml 1/2x MS medium [47] supplemented with 1% sucrose and 0.25% Gelrite, pH 5.8; the plates were maintained for seed germination at 28uC under continuous illumination (30 mmol s 21 m 22 ). Hybrid seedlings cultured at 28uC were all

Random Amplified Polymorphic DNA (RAPD) Analysis
Total DNA was extracted from the leaves of the hybrid seedlings using cetyltrimethylammonium bromide [49]. RAPD analysis was carried out as described [39] using 20 random 10-mer oligonucleotide primers (Kit OPE, OPF; Operon Technologies, Inc., Alameda, CA, USA). PCR amplification was performed using an Applied Biosystems 2720 Thermal Cycler (Applied Biosystems, Foster City, CA, USA) programmed for 3 min at 94uC for initial denaturation, followed by 60 cycles of 30 s at 94uC, 1 min at 28uC for seedlings of N. trigonophylla6N. tomentosiformis, 32uC for seedlings of N. trigonophylla6N. tomentosiformis, or 25uC for seedlings of N. sylvestris6N. trigonophylla, then 2 min at 72uC, and a final extension of 5 min at 72uC. PCR products were separated by electrophoresis on 1% agarose gels in TAE buffer and stained with ethidium bromide to visualize DNA bands.

Chromosome Analysis
Root tips of viable seedlings were pretreated with distilled water for 24 h at 4uC and with 2 mM 8-hydroxyquinoline for 4 h at 4uC, then fixed in a 3:1 mixture of ethanol and acetic acid overnight to determine chromosome numbers. The root tips were hydrolyzed in 1 N HCl for 9 min at 60uC, stained with Schiff's reagent and squashed in 45% acetic acid. The number of chromosomes in at least five root tip cells for each plant was counted under a BX51 light microscope (Olympus, Tokyo, Japan), and photographed using a DP70 automatic photomicrography system (Olympus).

Flow Cytometry
For cytometric analysis, nuclei were isolated from 100 mg leaves (except for midrib) of N. trigonophylla6N. tabacum or N. sylvestris6N. trigonophylla hybrid seedlings at 28uC; the leaves were chopped off and macerated in ice-cold buffer [50]. Four plants were analyzed for each type of sample. The macerated tissue was filtered through a 25 mm nylon mesh. Nuclei were collected from the filtrate by centrifugation (5 min at 3000 rpm and 4uC) and suspended in icecold buffer supplemented with 5 ml/ml DAPI for 1 min at 4uC. The DNA content of the isolated nuclei was analyzed by flow cytometry on a Cell Lab Quanta SC system (Beckman Coulter Inc., La Brea, CA, USA). Between 16,000 and 22,000 nuclei were counted.

Types of Abnormal Hybrid Seedlings of N. trigonophylla6N. tabacum Observed
When the cross N. trigonophylla6N. tabacum was carried out via conventional cross-pollination, no hybrid seedlings were obtained. Ovule culture was needed, as described previously [46]. Therefore, flowers of N. trigonophylla were pollinated with fresh N. tabacum pollen. From 31 flowers, 8284 ovules were obtained 5 days after pollination and 92 began to germinate 1 month after pollination. To confirm the expression of hybrid lethality, all of these seedlings were left at 28uC (Table 1).
Of the vitrified 43 hybrid seedlings, 25 hybrid plants were vitrified after germination ( Figure 1A). Eight hybrid plants were tumorous after germination, and became vitrified after browning  ( Figure 1B). Seven vitrified hybrid plants coexisted with tumors after germination ( Figure 1C). Two hybrid plants were tumorous after germination, and then became vitrified ( Figure 1D). One hybrid plant was malformed after germination. It developed a tumor at the crown, and then became vitrified ( Figure 1E). Of the 24 nonviable hybrid seedlings, 20 were tumorous after germination and died 15 DAG ( Figure 1F). The others were malformed after germination, and died 7 DAG ( Figure 1G). On the other hand, of the 21 viable hybrid seedlings, six coexisted with tumors after germination ( Figure 1H).
Of the two tumorous hybrid seedlings, one had a tumor that developed at the crown of the hybrid seedling, which was yellowing ( Figure 1I). The other tumor developed at the crown of the seedling, which was browning and vitrified ( Figure 1J).
To determine whether the viable and vitrified seedlings were true hybrids, we carried out morphological observation, RAPD analysis and chromosomal analyses on the viable seedlings, and flow cytometry on the vitrified seedlings. Four randomly selected seedlings were cultivated in a greenhouse for 60 DAG. All of the selected seedlings grew to maturity and came into flower (Figure 2A). The mature hybrid plants displayed uniform morphological characteristics, with leaf and flower shapes that were intermediate in appearance between those of the parents ( Figure 2B-D). Chromosomal analyses of the hybrid plants revealed that each possessed 36 chromosomes, which is the sum of the number of haploid chromosomes of the parents ( Figure 2E). RAPD analysis using the primer OPE-14 was carried out on four viable seedlings. In these plants, several bands characteristic of both parents coexisted ( Figure 2F, arrows); these patterns indicated that they are true hybrids. Similar results, shown in Figure 2F, were also obtained using several other random primers. All of the viable hybrid seedlings from the cross N. trigonophylla6N. tabacum were male sterile.
Flow cytometric analysis showed that the vitrified plants were true hybrids (Figure 3). The DAPI fluorescence values of the peaks corresponding to the G1 phase of the cell cycle differed, with values increasing in the order N. trigonophylla, viable hybrids and N. tabacum, using the G1 peak of N. trigonophylla as the internal standard in this experiment. The G1 peak of a viable hybrid was intermediate between the values of each parent ( Figure 3B). The G1 peak values of vitrified seedlings were also intermediate between that of the parents. However, fluorescence values of these G1 peaks were closer to that of N. trigonophylla and were different from that of viable hybrid seedlings. Furthermore, subG1 peaks, with lower fluorescence values than that of G1 peaks, indicating nuclear fragmentation, appeared and the values for G1 peaks differed among vitrified hybrid plants ( Figure 3C-E, black arrows).

Types of Abnormal Hybrid Seedlings of N. sylvestris6N. trigonophylla Observed
Reciprocal crosses between N. trigonophylla and N. sylvestris were carried out using conventional cross-pollination, but no hybrid seedlings were obtained (Table 2). Therefore, test tube pollination and ovule culture were performed to overcome the crossing barrier.
Pollination of 129 placentas of N. sylvestris in vitro resulted in 634 enlarged ovules that were cultured at 28uC, and 147 hybrid seedlings germinated from these ovules. Vitrified, viable and nonviable hybrid seedlings were observed; there were 89 vitrified seedlings, 57 viable seedlings and one nonviable seedling (Table 2). Of the 89 vitrified seedlings, 32 coexisted with tumors ( Figure 4A and B). The growth of one nonviable hybrid seedling stopped at an early stage, and it died 60 DAG ( Figure 4C).
To confirm that these seedlings were true hybrids, morphological observation, RAPD analysis and chromosome analysis were carried out for the four viable seedlings, and flow cytometry was carried out for the vitrified seedlings. All of the viable hybrid seedlings grew to maturity and flowered ( Figure 5A). The leaf shape, flower shape and flower color of the viable hybrid seedlings were intermediate in appearance between that of the parents ( Figure 5B-D). Four hybrid plants, selected randomly, had 24 chromosomes ( Figure 5E), which is the sum of the number of haploid chromosomes of their parents, N. sylvestris (2n = 24) and N. trigonophylla (2n = 24). These hybrid plants were male sterile. N. sylvestris, N. trigonophylla, and hybrid plants were also analyzed using RAPD. Hybrid plants had several bands in common between the RAPD patterns of parents using several primers (data not shown). RAPD patterns obtained with the primer OPF-03 are shown in Figure 5F as an example.
We used flow cytometry to clarify the extent of hybridity of the vitrified seedlings. The G1 peak of a viable seedling was intermediate between that of its parents, with a fluorescence value of 220 ( Figure 6A-C), and the G1 peak of a vitrified seedling appeared at the same fluorescence values as that of a viable seedling ( Figure 6C and D). Based on these results, we proved that the vitrified seedlings were true hybrids. However, the number of nuclei included in the G2/M peak of vitrified hybrids was greater than that of viable hybrids ( Figure 6D, red arrow). In addition, another G2/M peak was expressed at a fluorescence value of 870 ( Figure 6D). These results signify that in the G2/M peak of the vitrified hybrid plants ( Figure 6D, red arrow), a G1 peak corresponding to an allotetraploid was included.
The subG1 peak (corresponding to fragmented nuclei) with lower fluorescence values than that of the G1 peaks also appeared for the cross N. trigonophylla6N. tabacum ( Figure 6D, black arrow). This characteristic suggests that nuclear fragmentation progressed with vitrification.

Viable Hybrid Seedlings of N. trigonophylla6N. tomentosiformis
Results of conventional cross-pollination between N. trigonophylla and N. tomentosiformis are shown in Figure 7. Hybrid seedlings were obtained approximately 3 weeks after pollination. Hybrid seeds from N. trigonophylla6N. tomentosiformis crosses were sown in 1/2x MS medium and germinated at 28uC. The hybrid seedlings germinated well, and all of the hybrid seedlings were viable at 20 DAG. These seedlings grew to maturity and flowered ( Figure 7A). The morphological characteristics of hybrid plants were uniform and intermediate in appearance between those of the parents, including the leaf shape and flower shape ( Figure 7B-D). The chromosome number of the hybrid plants was 24, the sum of the number of haploid chromosomes of the parents ( Figure 7E). RAPD patterns for hybrid plants showed several clear bands characteristic of the parents using several primers, indicating that they are true hybrids (data not shown). RAPD patterns obtained with the primer OPE-14 are shown in Figure 7F.

Discussion
Vitrification is a morphological and physiological disorder affecting in vitro regenerated plants, and is promoted by several environmental factors including low gelling agent concentrations, liquid media, high relative humidity, high water potential in the medium, excess auxin or cytokinin, low agar concentration, high ammonium levels and high airtightness of the vessel [51][52][53][54][55][56]. However, vitrification occurring in interspecific crosses does not appear to have been reported previously. Though the focus of the paper was on hybrid lethality, a connection to vitrification was one of our findings. We observed vitrified hybrid seedlings from N. trigonophylla6N. tabacum when they were cultured in MS medium (in the absence of added plant growth regulators), and the number of vitrified hybrids was approximately 50% of the total (Table 1). In Nicotiana, hybrid lethality is classified into five types based on the external phenotype: Type I, browning of shoot apex and root tips; Type II, browning of hypocotyls and roots; Type III, yellowing of true leaves; Type IV, formation of multiple shoots; and Type V, fading of shoot color [7,8]. To our knowledge, of 26 species in section Suaveolentes, 22 have been crossed with N. tabacum and viability of hybrid seedlings has been reported. Among these crosses, hybrid lethality (Type II or V) is observed in the hybrid seedlings of 20 wild species and N. tabacum. In two other crosses, hybrid seedlings are viable [2]. However, we found that the morphology of hybrid plants of N. trigonophylla6N. tabacum is nonuniform ( Figure 8A). In several previous reports, it was reported that the phenotype of hybrid seedlings in reciprocal crosses was the same [8]. Since in crosses between section Suaveolentes species and N. tabacum, all hybrid seedlings either express hybrid lethality or do not, we speculate that N. trigonophylla has causative genomic factors affecting the phenotype of hybrid seedlings.
Similar observations of vitrification occurred in a cross between N. trigonophylla and N. sylvestris, but not in a cross of N. trigonophylla6N. tomentosiformis. These results indicate that the vitrification in the interspecific cross N. trigonophylla6N. tabacum is genetically controlled, and that the S subgenome of N. tabacum controls it ( Figure 8A).
Section Repandae formed from a single hybridization event between relatives of extant N. sylvestris and N. trigonophylla, the respective maternal and paternal progenitors, producing four allopolyploids: N. nudicaulis, N. repanda, N. stocktonii and N. nesophila [17,32,33]. When N. sylvestris was crossed with N. tabacum, viable hybrids were obtained ( Figure 8B). On the other hand, in the cross N. nudicaulis6N. tabacum, hybrid lethality was expressed and controlled by the S subgenome of N. tabacum ( Figure 8C); the genomic factor related to vitrification is also in the S subgenome ( Figure 8A). Based on these results, we inferred that the genomic factor in N. nudicaulis controlling hybrid lethality in the cross N. nudicaulis6N. tabacum is not related to the maternal progenitor, N. sylvestris, but is inherited from the paternal progenitor, N. trigonophylla. Although in the cross N. trigonophylla6N. tabacum, approximately 50% of hybrid seedlings showed vitrification, and hybrid lethality is expressed in the cross N. nudicaulis6N. tabacum, both N. nudicaulis and N. trigonophylla interact with the same subgenome of N. tabacum (the S subgenome) to control hybrid lethality or vitrification. Moreover, when we confirmed the hybridity of the vitrified seedlings of both N. trigonophylla6N. tabacum and N. sylvestris6N. trigonophylla using flow cytometry, we detected characteristic patterns of subG1 peaks with lower fluorescence values than observed for G1 peaks, indicating nuclear fragmentation ( Figures 3C-E and 6C, black arrows); the fluorescence of these subG1 peaks increased in the seedlings of N. nudicaulis6N. tabacum as they progressed through the stages of apoptotic cell death [38]. Thus, these results indicate that vitrification and hybrid lethality share the same hallmark of apoptosis, namely nuclear fragmentation. Furthermore, based on these results, we infer that vitrification is related to hybrid lethality.
Recent advances have highlighted the ubiquity of wholegenome duplication (polyploidy) in angiosperms; although global analyses of genome size reveal a trend towards DNA loss subsequent to polyploidy (genome downsizing), increases in genome size also occur [34,57]. In section Repandae, there is considerable departure from additivity in seven transposable element families, typically through deletion, especially for transposable elements derived from the N. obtusifolia parent [35]. Moreover, Renny-Byfield et al. [58] presented evidence that loss of high-copy-number repeats in the genome of N. nudicaulis, some of which are Ty3/Gypsy retroelements, contributed to a reduction of approximately 14% in genome size, similar to the 19% estimated using Markov chain Monte Carlo approaches. From these reports, we speculate that when N. nudicaulis was formed from the hybridization of N. sylvestris and N. trigonophylla approximately 5 million years ago, several genomic factors determining phenotypes of hybrid seedlings were inherited from N. trigonophylla. Subsequently, genome downsizing and various recombination-based processes took place. Some of the causative genomic factors were lost and some of them became genomic factor(s) controlling hybrid lethality in extant N. nudicaulis.
Currently, we are attempting to perform crosses of monosomic lines of N. tabacum and N. trigonophylla or N. nudicaulis to determine the chromosome(s) of N. tabacum responsible for vitrification of N. trigonophylla6N. tabacum and hybrid lethality of N. nudicaulis6N. tabacum. In the present study, we found that fragmentation of nuclei was accompanied by vitrification during progression of lethal symptoms in hybrids; therefore, we plan to use additional cytobiological approaches to determine whether apoptotic cell death accompanies vitrification.
Such an approach may provide important information revealing the relationship between vitrification and hybrid lethality. Furthermore, we expect these planned experiments will contribute to clarifying the origins of genomic factor(s) in N. nudicaulis controlling hybrid lethality in interspecific hybrids between N. nudicaulis and N. tabacum. This knowledge will likely provide a better understanding of hybrid lethality as a reproductive barrier, which would have broad applications in agriculture.