Overcoming Hybrid Lethality Induced by Chromosomal Instability in an Interspecific Hybrid of Genus Nicotiana


 Hybrid lethality is a type of reproductive isolation in which hybrids die before maturation, due to the interaction between the two causative genes derived from each of the hybrid parents. The interspecific hybrid of Nicotiana suaveolens x Nicotiana tabacum is a model plant for studies of hybrid lethality. In this cross, most hybrid seedlings die, but rare individuals grow normally and mature. Separately, a technique for producing mature hybrids by artificial culture has been developed. However, the mechanism by which hybrids overcome lethality, either spontaneously or by artificial culture, remains unclear. In the present study, we found that some hybrids that overcome lethality, either spontaneously or by artificial culture, lack the distal part of the Q chromosome, a region that includes the gene responsible for lethality. Quantitative polymerase chain reaction results suggested that the distal deletion of the Q chromosome, detected in some hybrid seedlings that overcome lethality, is caused by reciprocal translocations between homoeologous chromosomes. These results indicated that the chromosomal instability during meiosis of the amphidiploid N. tabacum and during artificial culturing of hybrid seedlings is involved in overcoming hybrid lethality in interspecific hybrids of the genus Nicotiana.


Introduction
Hybrid lethality is a type of postzygotic reproductive isolation in which fertilization between different species or populations occurs, but the resulting hybrids do not mature. In higher plants that experience hybrid lethality, fertilization is successful, but the hybrid embryos die before germination, or the hybrid seedlings show lethal symptoms such as browning, withering, and yellowing after germination. Hybrid lethality has been reported in a wide range of plant species, including wheat 1 , rice 2 , and cotton 3 . This phenomenon is a signi cant obstacle that restricts the genetic resources available for use in plant crossbreeding.
It is known that hybrid lethality also occurs in various combinations of interspeci c crosses in the genus Nicotiana 4 . Hybrid seedlings of Nicotianasuaveolens (2n = 2x = 32, genome constitution SuSu) and Nicotianatabacum (2n = 4x = 48, SSTT) germinate normally, but show type-II lethality, such that the hypocotyl browns and eventually dies a few days after germination (DAG) 5 . It has been reported that the amount of transcript of PAL (encoding phenylalanine ammonia-lyase), which is an immune responserelated gene, is increased in hybrid seedlings that show lethality 5 . We also have con rmed that the accumulation of protein aggregates promotes programmed cell death via autophagy in cultured cells of N.suaveolens x N.tabacum hybrids that exhibit lethality 6,7 .
Hybrid lethality is caused by the interaction of two complementary genes derived from each hybrid parent 8,9 . In plants, R genes encoding proteins that recognize effectors derived from a pathogen during a disease response have been reported as causative genes for hybrid lethality [10][11][12] . In hybrid lethality within the genus Nicotiana, Nt6549g30, a kind of NBS-LRR-type R gene of N. tabacum, has been reported as the gene responsible for the type-II lethality expressed in hybrid seedlings from crosses between N. tabacum and any of nine wild species, not including N. suaveolens, of Nicotiana section Suaveolentes 11,13 . On the other hand, it has been con rmed that the locus responsible for lethality in the N. suaveolens x N. tabacum cross is located on the Q chromosome belonging to the S genome of N. tabacum 14,15 .
Previous studies have reported con icting information regarding the chromosomal location of Nt6549g30. Ma 11 reported that Nt6549g30 is present on the H chromosome. Other work has suggested that the H chromosome belongs to the T genome 16 . On the other hand, the SSR marker of Linkage group No. 11, a gene that is detected on the Q chromosome of the S genome 17 , also was detected on the H chromosome 18 . Additionally, Ma 11 observed that the H chromosome belongs to the S genome. Therefore, we infer that the H chromosome in the reports of Hancock et al. 18 and Ma 11 is the same as the Q chromosome in the report of Tezuka et al. 17 . If this interpretation is valid, we consider it highly likely that Nt6549g30 is involved in the lethality of the N. suaveolens x N. tabacum cross.
In some lethal cross combinations between N. tabacum and wild Nicotiana species, including N. suaveolens x N. tabacum, viable hybrid seedlings (i.e., hybrid seedlings that overcome lethality) appear spontaneously at a certain frequency. Hancock et al. 18 detected hybrid seedlings that overcame lethality at a frequency of about 1.1 x 10 -3 in a N. tabacum x N. africana cross. Those authors also con rmed that the SSR marker at the distal part of Linkage group No. 11 was not detected in approximately 47% of such seedlings, suggesting the loss of the distal segment of the H chromosome (Q chromosome). This missing chromosomal end region contains Nt6549g30 (reference 11). However, the cause of such highfrequency loss of a chromosome terminus in hybrid seedlings remains unknown.
It has been reported that reciprocal translocation can occur between homoeologous chromosomes during meiosis in allopolyploid plant species such as Brassica napus 19 , coffee 20 , and interspeci c hybrids of the genus Lilium 21 . In N. tabacum, reciprocal translocation also has been reported to occur between homoeologous chromosomes, leading to loss of the N gene, a kind of R gene 22 . Based on these observations, we hypothesized that hybrid seedlings that overcome lethality following a cross between N. tabacum and wild species occur by formation of mutated gametes in which a distal segment of the Q chromosome, including Nt6549g30, is replaced with the homologous region of the homoeologous Q' chromosome by reciprocal translocation during meiosis (Hypothesis 1).
Since hybrid lethality is an obstacle to cross-breeding, various methods have been employed in attempting to produce hybrid plants that overcome lethality. In intraspeci c and interspeci c hybrids of wheat, hybrid plants that overcome lethality have been obtained by arti cial culture of hybrid embryos 23 , by proline treatment during fertilization and hybrid embryogenesis 24 . In N. suaveolens x N. tabacum, when hybrid seedlings were cultured in a medium containing a high concentration of cytokinin, vigorous shoots were formed at the stem bases of the seedlings, and regenerated plants that overcame lethality were obtained by rooting of these shoots 25 . However, the mechanism whereby plants that overcome lethality are produced by arti cial culturing remains unclear.
In in vitro cultured explants, it is known that reactive oxygen species (ROS) are generated by oxidative stress induced by medium components (plant hormones and salts) and culture environment 26 . On the other hand, ROS are thought to cause chromosome breakage 27 . Based on the above observations, we hypothesized that the production of regenerated plants that overcome lethality by culturing hybrid seedlings of N. suaveolens x N. tabacum in cytokinin-supplemented medium results from the appearance of cells lacking the distal part of the Q chromosome during in vitro culture (Hypothesis 2).
Elucidating the causes of interspeci c hybrids that overcome lethality is expected to reveal one aspect of the mechanism by which new species form by overcoming reproductive isolation. This information also is expected to contribute to the establishment of techniques for overcoming hybrid lethality, thereby leading to an expansion of the genetic resources available for cross-breeding. In the present study, we tested the above two hypotheses in an attempt to clarify the mechanisms whereby hybrid plants overcome lethality. First, we used the polymerase chain reaction (PCR) to assess the presence or absence of Nt6549g30 and the distal part of the Q chromosome in seedlings and regenerated plants, derived from the N. suaveolens x N. tabacum cross, that overcame hybrid lethality. Next, we used quantitative PCR (qPCR) to con rm reciprocal translocation between the Q chromosome region where Nt6549g30 resides and the homologous region of the homoeologous Q' chromosome.

Results
Acquisition of hybrid seedlings that overcome lethality A total of 15,476 seeds obtained from the cross of N. suaveolens x N. tabacum were sown, from which 12,943 seeds germinated. Starting shortly after germination, most of the seedlings from this cross showed lethal symptoms, such as browning of hypocotyls and roots along with yellowing of leaves, but 16 seedlings did not show any lethal symptoms at 20 days after sowing ( Fig. 1A-C). These seedlings were transferred to half-strength Murashige and Skoog medium (0.5x MS) in a plant box (Fig. 1D). The Randomly Ampli ed Polymorphic DNA-PCR (RAPD-PCR) method was used to evaluate the hybridity of viable seedlings; gel electrophoresis of the products showed that 12 of the 16 surviving seedlings yielded all amplicons from the two parents (i.e., RAPD bands speci c to N. suaveolens as well as those speci c to N. tabacum) (Fig. 2, Table S1). Among the remaining 4 viable seedlings, the s14-5 seedling lacked one of the three N. tabacum-speci c RAPD bands, the one corresponding to the product obtained with the OPA-15 primer. The other 3 seedlings (s14-9, s14-10, s17-3) yielded only the N. suaveolens-speci c RAPD bands.
Fifteen surviving seedlings (with the exception of s17-3) were potted, of which 13 owered (Fig. 1E). Eleven of the owered seedlings were judged to be hybrid seedlings that overcame lethality, given that these seedlings showed morphologies intermediate between those of the parental species in terms of ower color and shape, leaf size and shape, and plant posture (Fig. 1F, Table S2). The other two seedlings (s14-9 and s14-10) showed N. suaveolens-type ower color and plant posture, consistent with the results of the RAPD analysis. For 3 of the 15 surviving seedlings (s17-1, s17-2, and s18-1), transcripts of immune response-related genes (Table S3) were present (in the young true leaves) at levels much lower than those detected in the cotyledons of lethal seedlings at 6 DAG (Fig. 3). Based on these results, the 11 seedlings that showed morphologies intermediate between their parents were judged to be hybrids that had overcome lethality. By this assessment, the frequency of appearance of hybrid seedlings that overcame lethality was 8.5 x 10 -4 .
Acquisition of regenerated hybrid plants that overcome lethality by in vitro culturing of hybrid seedlings All 35 seedlings of N. suaveolens x N. tabacum cultured in 0.5x MS medium containing 6benzylaminopurine (BAP), a synthetic cytokinin, showed lethal symptoms shortly after germination. However, approximately one month later, many green shoots were observed to have formed at the bases of the stems of the seedlings (Fig. 4A). Shoots were cut from multiple plants; of 18 randomly selected shoots that were transplanted into 0.5x MS medium, 14 were able to root (Fig. 4B). The shoots that did not root vitri ed and eventually died. When the hybridity of 12 regenerated plants that grew normally without vitri cation were evaluated by the RAPD method, 10 individuals yielded all amplicons from the two parents (i.e., RAPD bands speci c to N. suaveolens as well as those speci c to N. tabacum) ( Table  S1, Fig. 2). In the remaining two regenerated plants (r18-1 and r19-4), one of the two N. tabacum-speci c RAPD bands (that obtained with the OPA-11 primer) was not seen (Table S1).
There was no difference in morphology among the 12 regenerated plants; of 4 of these plants (r18-1, r18-4, r18-5, and r18-7) that were potted, all achieved owering (Fig. 4C). All owered individuals showed morphologies intermediate between those of the two parents in terms of ower color and shape, leaf size and shape, and plant posture (Fig. 4D, Table S2). As seen for the primary seedlings that overcame lethality, the abundance of transcripts of immune response-related genes in young true leaves of the individual owered regenerated plants was notably lower than that in the cotyledons of lethal seedlings at 6 DAG ( Fig. 3). Based on these results, these regenerated plants were judged to be hybrids that that had overcome lethality.
PCR con rmation of deletion of the distal part of the Q chromosome and Nt6549g30 PCR was performed using a primer pair capable of speci cally amplifying the SSR and Nt6549g30 markers present in the DNA of the Q chromosome belonging to the S genome of N. tabacum (Fig. 5). In lethal hybrid seedlings, PCR ampli cation products (amplicons) were detected with all SSR primer pairs and Nt6549g30 primer pairs (Table 1). On the other hand, in the 5 of 8 individual hybrid seedlings that overcame lethality and in 3 of 12 individual viable regenerated hybrid plants that overcame lethality, no amplicon was detected for reactions using the primer pairs (PT30342, PT30365, and PT52778) targeting three SSRs in the DNA sequence of the distal part of the Q chromosome. These amplicons also were not detected from these 8 individuals when using the two primer pairs capable of amplifying Nt6549g30. In addition, in 6 of these 8 individuals, some other SSRs were not ampli ed, but the location and number of the missing SSRs differed among the individuals.
Genome analysis was performed using Genotyping by Random Amplicon Sequencing-Direct (GRAS-Di) technology 28 for the seedlings of hybrid parents (N. suaveolens and N. tabacum), two lethal hybrid seedlings (samples No. 1 and 3), and one hybrid seedling that overcame lethality (s17-2). In GRAS-Di, DNA sequence analysis of the amplicons obtained with 63 random primers and mapping to reference sequences detected 187 amplicons derived from Q-chromosome DNA in N. tabacum and lethal hybrid seedlings (Table S4). On the other hand, in s17-2, amplicons derived from DNA sequences located from bp 1 to 81360673 on the Q chromosome were detected. However, four amplicons (AMP0045831, AMP0067003, AMP0062028, and AMP0042233) derived from DNA sequences located between bp 81360673 and 81497158 on the Q chromosome were not detected.

Con rmation of reciprocal translocation between homoeologous chromosomes by qPCR
For ampli cation of the homoeologous chromosome of the Q chromosome (the Q' chromosome) belonging to the T genome of N. tabacum, three sets of primer pairs capable of speci cally amplifying the SSRs present in the DNA sequences at both ends of the chromosome were selected (Fig. 5). Total DNA was extracted from 5 hybrid seedlings that overcame lethality and from 3 regenerated hybrid plants that overcame lethality, in which SSRs of terminal DNA of the Q chromosome were not detected, and from 3 lethal hybrid seedlings. The copy number of the PCR ampli cation region of each primer pair (normalized to a given amount of DNA) then was investigated by qPCR (Fig. 6). Speci cally, a copy number was determined for the region ampli ed by two separate sets of primer pairs (NtScfTN90_54683-2 and NtScfTN90_1535-4) designed based on the Q' chromosome's terminal DNA. i.e., sequences homologous to the Q chromosome terminal DNA, including the Nt6549g30 locus. Notably, in 4 of the 5 hybrid seedlings that overcame lethality, the copy number of these selected regions was approximately twice that in lethal hybrid seedlings. The remaining hybrid seedling (s17-1) showed a value that was similar to that seen in the lethal hybrid seedling. On the other hand, the copy numbers of the PT50790targeted domain (corresponding to the distal part of the Q' chromosome, at the end opposite to the Nt6549g30 locus ampli cation region) did not show a signi cant difference between lethal hybrid seedlings and hybrid seedlings that had overcome lethality. In addition, for the regenerated hybrid plants that overcame lethality, the copy number of each ampli cation region did not differ signi cantly from that of the lethal hybrid seedlings.

Discussion
This study sought to elucidate the mechanism by which individual hybrid plants overcome hybrid lethality, whether spontaneously or by arti cially propagation. To achieve this goal, we formulated and tested two (non-exclusive) working hypotheses based on previous ndings regarding the lethalityovercoming phenomenon of products of a N. suaveolens x N. tabacum cross. Hypothesis 1 was that seedlings that overcome lethality in the N. suaveolens x N. tabacum cross occur by formation of male gametes lacking the lethality-associated Nt6549g30 gene, which is lost as a result of reciprocal translocations, between homoeologous chromosomes, that occur during meiosis in N. tabacum. Hypothesis 2 focused on the observation that culturing of N. suaveolens x N. tabacum seedlings in the presence of cytokinin yields regenerated plants that overcome hybrid lethality at high frequency; we proposed that hybrid lethality is overcome as a result of the appearance of cells lacking the distal part of the Q chromosome during culturing.
To examine Hypothesis 1, we rst identi ed 16 surviving seedlings from 12,943 germinated N. suaveolens x N. tabacum seedling individuals. The results of RAPD-PCR and morphological observation demonstrated that some of the viable seedlings were indeed hybrids (possessing genomes from both parents) and showed extremely low expression levels of immune response-related genes, con rming these to be hybrid seedlings that had overcome lethality (Figs. 1, 2, 3, Tables 1, S1). Speci cally, 11 hybrid seedlings that had overcome lethality were obtained, and the frequency of appearance of hybrid seedlings that overcame lethality in this cross combination was calculated to be 8.5 x 10 -4 . This value was close to 1.1 x 10 -3 , which is the frequency of appearance of hybrid seedlings that overcame lethality in a previous study of the N. tabacum x N. africana cross 18 .
PCR results showed that 5 of 8 of the lethality-overcoming seedlings lacked the distal part of the Q chromosome, a region that contains Nt6549g30 (Table 1). When qPCR was performed using two sets of primer pairs designed to amplify DNA at the end of the Q' chromosome, the copy number of the ampli ed region from 4 of the 5 individuals that overcame lethality was approximately twice that of seedlings that exhibited lethality (Fig. 6). Therefore, in these 4 individuals, it appears that the end of the Q chromosome, including Nt6549g30, was replaced due to a reciprocal translocation between the Q chromosome and its homoeologous chromosome (Q'). These data support the existence of a lethality-overcoming mechanism consistent with our Hypothesis 1. Figure 7 shows a suggested model, based on our results, for the lethality-overcoming mechanism in N. suaveolens x N. tabacum. In this model, a hybrid seedling that overcomes lethality is produced by fertilization of a female gamete from N. suaveolens by a male gamete from N. tabacum; notably, the male gamete has lost Nt6549g30 due to reciprocal translocation, during meiosis, between the distal regions of the Q and Q' chromosomes. This model is consistent with the overcoming of lethality in N. tabacum x N. africana 18 and in other interspeci c crosses within the genus Nicotiana where N. tabacum serves as one parent.
GRAS-Di analysis of one hybrid (s17-2) that escaped lethality suggested that lethality was overcome by reciprocal translocation. Notably, an amplicon derived from the region extending to Q chromosomal DNA bp 81360673 was detected, but no amplicon derived from the DNA region beyond bp 81497158 was detected (Table S4). This result implied that the break-point of the reciprocal translocation detected in s17-2 is located within a region of about 136 kb, positioned between bp 81360673 and bp 81497158 on the Q chromosome.
For one (s17-1) of the ve seedlings that overcame lethality, qPCR analysis showed that the copy number of the ampli ed region was the same as that of the lethal seedlings (Fig. 6). This observation implicated deletion of the distal part of the Q chromosome in overcoming lethality in this individual. Speci cally, for s17-1, products that would be ampli ed by three primer pairs (PT52864, PT55075, and PT60178) were not detected; these fragments would span the SSR marker located near the center of the Q chromosome (at 62.25 cM) (Table 1). Therefore, reciprocal translocations between homoeologous chromosomes may be more likely to occur in the region near the end of the Q chromosome, while deletions of relevant loci may be more likely to occur in the region closer to the center of the chromosome. In heteroploid synthetic Brassica plants, both chromosomal deletions and reciprocal translocations (between homoeologous chromosomes) have been reported in synthetic progeny [29][30] . We postulate that the deletion at the distal part of the Q chromosome may be the result of aberrant segregation and cleavage of chromosomes following synapsis between homoeologous chromosomes during meiosis of N. tabacum, which is a heteroploid.
In 3 individuals (s14-1, s14-2, and s14-4) that overcame lethality, a number that represented approximately 38% of the 8 seedlings that overcame lethality and were subjected to PCR, amplicons were detected for all of the primer pairs (Table 1). Therefore, it appeared that these individuals did not lack the lethality-associated gene on the N. tabacum side (Nt6549g30) as a result of reciprocal translocation or deletion of the distal part of the Q chromosome end. In N. tabacum x N. africana, it has been suggested that about 37% of lethality-overcoming seedlings have intact Q chromosomes 18 . Both N. suaveolens and N. africana are included in the Suaveolentes section and are thought to be derived from amphidiploid progenitors 31 . Based on these results, we hypothesized that reciprocal translocation or deletion of the chromosome segment on which the lethality-associated gene is located also can occur in N. suaveolens and N. africana. Overcoming lethality in the seedlings of N. suaveolens x N. tabacum and N. tabacum x N. africana, in which no reciprocal translocation or deletion of the distal part of the Q chromosome was detected, may result from chromosomal mutations due to synapsis of homoeologous chromosomes. However, since the lethality-associated genes in N. suaveolens and N. africana are unknown and the genomic sequences of those species remain unpublished, this hypothesis cannot be tested at this time.
To examine our Hypothesis 2, seedlings of N. suaveolens x N. tabacum were cultured in a medium containing a high concentration of cytokinin, in an effort to obtain viable regenerated plants. The results of RAPD-PCR and morphological observation suggested the resulting regenerated plants had genomes from both parents, con rming that these individuals were indeed hybrids. These plants also showed strongly decreased expression levels of immune response-related genes. Therefore, we judged these individuals to be regenerated hybrid plants that had overcome lethality (Figs. 2, 3, 4, Tables 1, S1). In 3 (r18-1, r18-7, and r19-3) of the 12 regenerated plants that overcame lethality, PCR analysis showed that the distal part of the Q chromosome (the region containing Nt6549g30) had been lost, as seen for some of the seedlings that overcame lethality (Table 1). Additionally, qPCR of these three individuals showed that the copy number of the ampli ed region was the same as that of lethal seedlings (Fig. 6). These results suggested that the deletion of the distal part of the Q chromosome in these three individuals contributed to overcoming lethality, an inference consistent with our Hypothesis 2. Hypothesis 2 was based on the previous nding that ROS are generated in culture 26 and are involved in chromosomal cleavage 27 . In the future, it will be necessary to verify that ROS are involved in the deletion of the distal part of the Q chromosome. Such con rmatory experiments will require demonstrating ROS production in hybrid seedlings that are cultured in a medium containing a high concentration of cytokinin, and then showing that inhibition of ROS production prevents the emergence of viable shoots.
In 9 of the 12 regenerated plants that overcame lethality, PCR analysis showed that amplicons were detected for all primer pairs (Table 1). Thus, in these individuals, it appears that lethality was overcome by a mechanism other than the deletion of the distal part of the Q chromosome. In addition to recombination-associated events that result in chromosomal lesions, mutation by DNA methylation and by DNA base substitution and deletion is known to occur in plant tissue culture 26 . It also has been suggested that the supplementation of medium with high-concentration BAP induces mutation 32 .
Furthermore, certain R genes have been reported to show frequent DNA base substitutions and deletions during mitosis 33 . It is possible that these various genetic mutations occur in the lethality-associated genes of N. suaveolens and N. tabacum, resulting in the formation of viable shoots during tissue culture. It also is possible that ROS production under culture conditions induces deletion of the chromosomal region containing the lethality-associated gene in the N. suaveolens genome. To clarify these mutational mechanisms, it will be necessary to perform detailed sequence analysis of the Nt6549g30 locus in regenerated plants that overcome lethality, and to identify the lethality-associated chromosome and gene in the N. suaveolens genome.
In the present study, expression analysis of immune-response-related genes (e.g., PAL1, PRB1-like, LOX1, and PDF-like protein 1) was performed in the process of detecting seedlings and regenerated plants that overcame lethality. In A. thaliana, the PAL1 gene product is involved in the synthesis of salicylic acid 34 , and PRB1 encodes a PR-1-like protein expressed in response to ethylene and methyl jasmonate 35 . On the other hand, the LOX1 gene product has been implicated in jasmonic acid synthesis 36 , and PDF1.2 encodes a jasmonic acid-responsive defensin 36 . In the present work, the expression levels of these genes in seedlings and regenerated plants that overcame lethality were much lower than those in the lethal seedlings at 6 DAG (Fig. 3). These data strongly suggested that the seedlings and regenerated plants that overcome lethality lack the function of the lethality-associated gene that triggers the immune-response signal, a de ciency that presumably is due to some genetic variation.
Hybrid lethality is a type of reproductive isolation and a mechanism for maintaining species independence. On the other hand, the results of the present study suggest that chromosomal instability permits hybrids to overcome hybrid lethality, thereby escaping reproductive isolation. We postulate that this phenomenon may contribute to the birth of new species in the process of plant evolution. In addition, hybrid lethality has been reported in a wide range of crop species 9 and is a serious obstacle in hybrid breeding of crops. The ndings obtained in this study may lead to a technique for arti cially breaking reproductive isolation by inducing chromosomal instability.
In summary, we found that the loss of the distal part of the Q chromosome is involved in overcoming hybrid lethality in seedlings and in regenerated plants obtained by culturing hybrid seedlings produced by interspeci c Nicotiana crosses. We propose that loss of this factor is the result of reciprocal translocation and deletion between the Q chromosome and its homoeologous partner chromosome (Q') in hybrid seedlings, and of a distal deletion in the Q chromosome in regenerated plants. These ndings, which relate to the mechanism of breaking reproductive isolation, are likely to be important in the evolution of polyploid plant species and in the expansion of genetic resources available for cross-breeding.

Materials And Methods
Sowing hybrid seeds and obtaining viable seedlings Hybrid seeds of N. suaveolens x N. tabacum were obtained according to the method of Yamada et al. 37 .
The seeds were surface-sterilized by immersion in 70% ethanol for 30 seconds, followed by immersion in 5% sodium hypochlorite for 20 minutes. After washing three times with sterile water on a clean bench, seeds were sown in half-concentration MS medium (0.5x MS) that had been adjusted to pH 5.8 and supplemented with 1% sucrose and a gelling agent (0.2% GelRite ® (Wako) or 0.3% Phytagel ™ (Sigma)). The medium containing the seeds then was incubated at 28°C in an incubator with a photoperiod of 24 h. The number of seedlings that remained viable 20 days after sowing was determined.
Culturing hybrid seedlings in medium containing a high concentration of cytokinin, and obtaining viable shoots After surface-sterilization as above, hybrid seeds of N. suaveolens x N. tabacum were sown in 0.5x MS that had been adjusted to pH 5.8 and supplemented with 2.0 mg/L 6-benzylaminopurine (BAP) (Sigma), 1% sucrose, and a gelling agent (0.2% GelRite ® or 0.3% Phytagel ™ ). The medium containing the seeds then was incubated at 28°C in an incubator with a photoperiod of 24 h. Viable shoots that differentiated from the stem base after incubation for more than one month from sowing were excised, transplanted to 0.5x MS without plant hormones, and rooted.

DNA extraction
Leaves or stems were collected from hybrid parents, lethal seedlings, viable seedlings, and viable regenerated plants. Total DNA was extracted from these tissues using the cetyltrimethylammonium bromide (CTAB) method according to Yamada et al. 37 , or using a DNeasy Plant Mini Kit (Qiagen) according to the manufacturer's protocol. Purity and concentration of DNA samples were determined using Nanodrop (Thermo Fisher Scienti c) and Qubit (Thermo Fisher Scienti c) instruments.

Con rmation of hybridity by RAPD-PCR
Hybridity was tested by RAPD-PCR with reference to Marubashi & Onosato 14 and Tezuka et al. 38 . Six PCR primers were used, including OPA-1, OPA-5, OPA-9, OPA-11, OPA-12, and OPA-15 (Operon). KAPA Taq Extra Hotstart Ready mix with dye (KAPA Biosystems) was used for PCR. The PCR reaction solution was prepared by combining 10 μL of 2x KAPA Taq Extra Hotstart Ready mix with dye, 2 μL of OPA primer at 10 μM, and 20 ng of DNA; the solution then was adjusted to 20 μL with PCR-grade water and mixed. The PCR reaction solution was subjected to initial denaturation at 95°C for 15 min, followed by 45 cycles of denaturation at 94°C for 1 min, annealing at 35°C for 2 min, and extension at 72°C for 3 min, using an Applied Biosystems 2720 Thermal Cycler (Life Technologies). The PCR products were electrophoresed (along with a commercial DNA size marker) on a 2% or 3% agarose gel supplemented with 0.01% Gel Red (Biotium) using TAE buffer as the electrophoresis buffer, and the gel was photographed with UV light to detect the PCR products.

Con rmation of hybridity by morphological observation
Viable seedlings and regenerated plants were transplanted into pots containing Super Soil Mix A (Sakata Seed), acclimated in a constant temperature room at 25°C for 16 h, and allowed to ower. Leaf and ower morphology and plant posture were compared with those of the parent strains.

Expression analysis
Total RNA was extracted from cotyledons of lethal seedlings, true leaves of viable seedlings, and regenerated N. suaveolens x N. tabacum plants using RNAiso Plus (Takara) according to the method of Shinozaki et al. 39 . Lethal seedlings were cultivated according to the method described above for "Sowing hybrid seeds and obtaining viable seedlings"; the sample consisted of the cotyledons collected from 30 to 50 individuals at the 6-DAG stage. Seedlings and regenerated plants that overcame lethality were cultivated according to the method described above for "Con rmation of hybridity by morphological observation", and young true leaves were sampled. Purity and concentration of RNA samples were determined using Nanodrop and Qubit instruments. A PrimerScript RT reagent Kit with gDNA Eraser (Takara) was used for synthesis of cDNA, and the initial template cDNA was adjusted to 10 ng/μL. The immune-response genes for N. tabacum, which have high homology with those of A. thaliana, were used as the targets for expression analysis (Table S5). Primer pairs targeting each gene were designed using Primer3Plus (https://primer3plus.com/) and NetPrimer (PREMIER Biosoft International; http://www.premierbiosoft.com/netprimer/) software programs (Table S5). For real-time quantitative reverse transcription (qRT) -PCR, a mixture corresponding to 5 µL of KAPA SYBR FAST Universal 2x qPCR Master Mix (KAPA Biosystems), 0.2 µL of 10 µM Forward primer, 0.2 µL of 10 µM Reverse primer, 3.6 µL of PCR-grade water, and 1 µL of cDNA solution was generated and distributed to each well of a 0.2-mL 48-well PCR plate. The PCR reaction solution was subjected to initial denaturation at 95°C for 10 min followed by 40 cycles of denaturation at 94°C for 10 s, and annealing and extension at 60°C for 30 s using an Eco Real-Time PCR System (Illumina). Finally, melting curve analysis was performed from 60°C to 95°C and the plate then was maintained at 4°C. The mRNA copy number of each gene in 10 ng of total RNA of each sample was determined by comparison to a standard curve. The mRNA copy number of RFC3, a housekeeping gene, was determined as an internal standard and used for normalization to determine relative transcript amounts. qRT-PCR analysis was performed in triplicate; the resulting data were used to calculate mean transcript levels.

Detection of Nt6549g30 and SSR markers on the Q chromosome
The nucleotide sequence of Nt6549g30 was obtained from Ma 11 , and gene-speci c primers Nt6549g30-1 and Nt6549g30-3 were designed using NetPrimer software (Fig. 5, Table S5). The SSR markers on the Q chromosome (Linkage group No. 11) were obtained from Bindler et al. 40,41 , and correspond to primer pairs that amplify products from N. tabacum but not from N. suaveolens, or those that generate amplicons of distinct sizes in the two species (Fig. 1, Table S5). These PCR ampli cations used the KAPA Taq Extra PCR Kit (KAPA Biosystems), with each 10-µL reaction comprising 2.0 μL of 5x KAPA Taq Extra Buffer, 0.6 μL of 25 mM MgCl 2 , 0.2 μL of 10 mM dNTP Mix, 0.5 μL of 10 μM Forward primer, 0.5 μL of 10 μM Reverse primer, 0.5 μL of 5 U/μL KAPA Taq Extra DNA Polymerase, 20 ng of DNA, and PCR-grade water to volume. Alternatively, the reactions used the KAPA TaqExtra Hotstart Ready mix with dye, with each 10-µL reaction comprising 5 μL of 2x KAPA TaqExtra Hotstart Ready mix with dye, 0.5 μL of 10 μM Forward primer, 0.5 μL of 10 μM Reverse primer, 20 ng of DNA, and PCR-grade water to volume. For detection of Nt6549g30, the PCR reaction solutions were subjected to initial denaturation at 94°C for 5 min, followed by 30 cycles of denaturation at 94°C for 0.5 min, annealing at 55°C for 0.5 min, and extension at 72°C for 1 min. For detection of SSR markers, the PCR reaction solutions were subjected to initial denaturation at 94°C for 3 min followed by 35 cycles of denaturation at 94°C for 0.5 min, annealing at 55°C for 0.5 min, and extension at 72°C for 1 min. The PCR products were electrophoresed and visualized as described above for RAPD-PCR products.

GRAS-Di
GRAS-Di analysis was performed under contract at Gene Bay Co., Ltd. Among the obtained GRAS-Di amplicons, those speci c to N. tabacum reference genomic sequence according to Edwards et al. 42 (ftp://ftp.solgenomics.net/genomes/Nicotiana_tabacum/edwards_et_al_2017/assembly/Nitab-v4.5_genome_Chr_Edwards2017.fasta), those speci c to the scaffolds included in the Q chromosome (Linkage group No. 11), and those with no mismatch were selected. For each individual, the marker with the number of sequenced reads judged to be 0 was de ned as "no detection", and the marker for which the number of sequenced reads was judged to be greater than 0 was de ned as "detection".
Analysis of the copy number of the SSR marker on the Q homoeologous chromosome The SSR marker on the homoeologous chromosome of the Q chromosome (Linkage group No. 13, Edwards et al. 42 ) from Bindler et al. 41 that failed to exhibit PCR ampli cation in N. suaveolens was selected and used for qPCR (Fig. 5, Table S5). In addition, Primer BLAST (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) was used to identify the scaffold (Sierro et al. 43 ) that mapped to the homoeologous partner of the Q chromosome and contained the SSR marker. Flanking sequences from this scaffold were used to design primers (thereby de ning DNA markers) that generated products from N. tabacum but not from N. suaveolens (Fig. 5, Table S5). Real-time q-PCR was performed according to the method described above in "Expression analysis" using (as templates) the total DNA of lethal seedlings, viable seedlings, and regenerated plants in which no SSR marker corresponding to the distal part of the Q chromosome was detected. These measurements were repeated three times; the resulting data were used to calculate mean values.