Genetic Mapping of the HLA1 Locus Causing Hybrid Lethality in Nicotiana Interspecific Hybrids

Hybrid lethality, a postzygotic mechanism of reproductive isolation, is a phenomenon that causes the death of F1 hybrid seedlings. Hybrid lethality is generally caused by the epistatic interaction of two or more loci. In the genus Nicotiana, N. debneyi has the dominant allele Hla1-1 at the HLA1 locus that causes hybrid lethality in F1 hybrid seedlings by interaction with N. tabacum allele(s). Here, we mapped the HLA1 locus using the F2 population segregating for the Hla1-1 allele derived from the interspecific cross between N. debneyi and N. fragrans. To map HLA1, several DNA markers including random amplified polymorphic DNA, amplified fragment length polymorphism, and simple sequence repeat markers, were used. Additionally, DNA markers were developed based on disease resistance gene homologs identified from the genome sequence of N. benthamiana. Linkage analysis revealed that HLA1 was located between two cleaved amplified polymorphic sequence markers Nb14-CAPS and NbRGH1-CAPS at a distance of 10.8 and 10.9 cM, respectively. The distance between these markers was equivalent to a 682 kb interval in the genome sequence of N. benthamiana.


Introduction
Species are often reproductively isolated from each other. This mechanism, called reproductive isolation, has been an interesting theme in evolutionary biology because it plays an important role in preventing gene flow between species [1]. However, reproductive isolation is an obstacle that should be resolved for plant breeding. Although many studies have reported the prevention or bypass of reproductive isolation [2][3][4], challenging cases are still often encountered.
Developmental defects of seedlings such as hybrid lethality, hybrid necrosis, and hybrid weakness are often observed in F 1 [5][6][7] or later generations (hybrid breakdown) [8,9] in many plant species. These phenomena are caused by epistatic interactions of two or more loci as explained by the Bateson-Dobzhansky-Muller model [10]. Many of the identified causal genes for seedling developmental defects encode disease resistance proteins [11][12][13][14]. In other cases, seedling developmental defects were also reported to be caused by duplicate genes such as histidinol-phosphate amino-transferase gene [15] or photosynthetic gene plastid transcriptionally active chromosome 14 [16].
In the genus Nicotiana, hybrid lethality that causes death of F 1 hybrid seedlings is often observed in interspecific crosses including those involving cultivated species N. tabacum [17,18]. Many types of hybrid lethality are temperature sensitive and suppressed at elevated temperatures ranging from 32 • C to 36 • C [19][20][21]. Several mechanisms which explain how hybrid lethality occurs have been revealed: programmed cell death with apoptotic features [22,23], generation of reactive oxygen species [24], and ethylene [25], and autophagy [26] are observed during hybrid lethality. Considering that some types of disease resistance responses are temperature sensitive [27][28][29] and involve programmed cell death [30], the results from the experiments on hybrid lethality in Nicotiana have suggested that disease resistance-related responses might be associated with hybrid lethality. Several genes and proteins also relevant to disease resistance have been identified in the process of hybrid lethality [31][32][33][34]. However, the causal genes for hybrid lethality in Nicotiana are poorly understood.
Hybrid lethality in Nicotiana is thought to be caused by epistatic interactions between alleles at different loci [18,35]. Previous studies have demonstrated that the Q chromosome of the S genome in N. tabacum has gene(s) that cause hybrid lethality [17,20,36,37]. Recently, the N. tabacum hybrid lethality 1 (NtHL1) gene encoding a protein containing a coiled-coil, nucleotide-binding site, and leucine-rich repeat was identified in N. tabacum [38]; this gene might correspond to the hypothesized gene(s) of the epistatic interaction on the Q chromosome. For the counterpart of the Q chromosome gene(s), we identified the hybrid lethality A1 (HLA1) gene that causes hybrid lethality by interaction with the hypothesized gene(s) on the Q chromosome [39]. Wild species in Nicotiana section Suaveolentes, N. debneyi, has the dominant allele Hla1-1 at the HLA1 locus; we obtained F 1 hybrids between N. debneyi and N. fragrans in which the latter also belongs to section Suaveolentes and possesses the non-causal allele hla1-2 at the HLA1 locus. The F 1 hybrids were self-fertile, and F 2 seeds were obtained [39]. Therefore, the HLA1 locus can be mapped to a specific region of a chromosome. However, the chance of crossover depends on whether chromosome pairing occurs. In interspecific hybrids, chromosome pairing is often disturbed due to differences in the genomes or chromosomes of both parents [40,41].
In the present study, we aimed to map the position of the HLA1 locus. Linkage analysis was carried out using the F 2 population derived from the cross N. debneyi × N. fragrans with several DNA markers. Because little genomic or marker information was available for the parental species N. debneyi and N. fragrans (both allotetraploids), we initially used random amplified polymorphic DNA (RAPD) and amplified fragment length polymorphism (AFLP) markers [42,43] that do not require previous sequence information, as well as simple sequence repeat (SSR) markers developed in N. tabacum [44]. In many cases, disease resistance-related genes have been identified as causal genes for hybrid lethality [11][12][13][14]38]. Therefore, it is reasonable to assume that HLA1 might be a disease resistance-related gene. One parent, N. debneyi, shows several disease resistances, some of which have been transferred to N. tabacum [45][46][47]. We used two reported sequence-characterized amplified region (SCAR) markers, SUBC180.251 [48] and Mil275 [49], linked to blue mold resistance in N. tabacum that was originally transferred from N. debneyi. To develop further DNA markers, we also used the draft genome sequence of N. benthamiana [50], a species belonging to the section Suaveolentes and closely related to N. debneyi and N. fragrans [51,52]. Additionally, we investigated chromosome pairing in the F 1 hybrid between N. debneyi and N. fragrans. Here, we report the map position of the HLA1 locus in the linkage map derived from the F 2 population of the cross N. debneyi × N. fragrans.

Segregation of the Hybrid Lethality Phenotype in the F 2 Population
After the test crosses with N. tabacum, crossed seeds were obtained from 81 of 125 F 2 plants derived from the cross N. debneyi × N. fragrans; seeds could not be obtained in crosses using the remaining 44  2.2. Construction of a Linkage Map Using RAPD, AFLP, and SSR Markers RAPD, AFLP, and SSR markers were initially used to map the HLA1 locus. In RAPD analysis, 385 bands were detected using 80 random primers; 78 (20.3%) bands showed polymorphism between N. debneyi and N. fragrans (Figure 1a,b; Table 1). In AFLP analysis, 2327 (44.8%) of 5191 bands detected using 256 primer pairs showed polymorphism between the parents (Figure 1c,d; Table 1). These polymorphic bands were assumed to be dominant markers.

Segregation of the Hybrid Lethality Phenotype in the F2 Population
After the test crosses with N. tabacum, crossed seeds were obtained from 81 of 125 F2 plants derived from the cross N. debneyi × N. fragrans; seeds could not be obtained in crosses using the remaining 44 F2 plants.

Construction of a Linkage Map Using RAPD, AFLP, and SSR Markers
RAPD, AFLP, and SSR markers were initially used to map the HLA1 locus. In RAPD analysis, 385 bands were detected using 80 random primers; 78 (20.3%) bands showed polymorphism between N. debneyi and N. fragrans (Figure 1a,b; Table 1). In AFLP analysis, 2327 (44.8%) of 5191 bands detected using 256 primer pairs showed polymorphism between the parents (Figure 1c,d; Table 1). These polymorphic bands were assumed to be dominant markers.  For SSR analysis, 112 bands were detected using 66 primer pairs; 22 and 12 bands were specific to N. debneyi and N. fragrans, respectively, and 14 bands were codominant. Therefore, although SSR markers were generally codominant, only seven markers were detected as codominant in the present study. Totally, 41 markers were detected as polymorphic ( Figure 1e, Table 1).
Of the above-identified polymorphic markers, all 78 RAPD markers, 111 AFLP markers, and 13 SSR markers were used to construct a linkage map. Linkage analysis resulted in only two major (LG1 and LG2) and 24 minor linkage groups (LG3 to LG26). Within the minor linkage groups, map distances of LG3 to LG8 were comparatively longer than others ( Figure S1). Many markers tended to cluster in some regions. The HLA1 locus was mapped between the codominant SSR marker PT30138a and the AFLP marker E40M58.600 in LG1.

Chromosome Pairing in F 1 Hybrid between N. debneyi and N. fragrans
Because many linkage groups were generated, we investigated chromosome pairing in the F 1 hybrid used to produce the F 2 generation to verify whether crossover can occur during meiosis of F 1 hybrids between N. debneyi and N. fragrans. A high frequency of bivalents was observed with an average of 0.6 univalents, 23.2 bivalents, and 0.4 trivalents per pollen mother cell ( Figure S2; Table S1).

Linkage Analysis of HLA1 Using Disease Resistance-Related Markers
When two SCAR markers linked to blue mold resistance were tested, SUBC180.251 showed no polymorphism between parents, whereas Mil275 was detected in N. debneyi and the F 1 hybrid but not in N. fragrans ( Figure S3a,c). SUBC180.251 was converted to the cleaved amplified polymorphic sequence (CAPS) marker SUBC180.251-CAPS, and this newly developed marker distinguished the genotypes of both parents and the F 1 hybrid ( Figure S3b). Both SUBC180.251-CAPS and Mil275 were mapped to LG2 and thus were not linked to the HLA1 locus.
We then used the draft genome sequence of N. benthamiana [50]. When we started the experiments, the N. benthamiana genome v0.4.4 was available. To identify resistance gene homologs from the genome sequence, a BLAST search was conducted using TIR-NBS-LRR resistance genes in Arabidopsis. Four disease resistance gene homologs were identified from the N. benthamiana protein and transcript sequences (Table S2). Four sequence-tagged site (STS) markers, NbRGH1 to NbRGH4, were developed based on the sequences of their homologs (Table S3). Among these markers, only NbRGH4 showed polymorphism in the parents and F 1 hybrid (Figure 2d). After the conversion of the remaining three markers to CAPS or derived CAPS (dCAPS) markers, NbRGH2-dCAPS and NbRGH3-CAPS could be used as codominant polymorphic markers (Figure 2b,c). For NbRGH1-CAPS, a lower band resulting from BsrI digestion was detected in N. debneyi as expected, but two bands were unexpectedly detected in N. fragrans ( Figure 2a); this might be due to the complexity of the N. fragrans genome because this species is an allotetraploid. Linkage analysis revealed that NbRGH1-CAPS was linked to the HLA1 locus at a distance of 10.9 cM (Figure 3). Two markers were mapped to other linkage groups: NbRGH3-CAPS to LG2 and NbRGH4 to LG3 ( Figure S1). No marker linked to NbRGH2-dCAPS was detected.
expected, but two bands were unexpectedly detected in N. fragrans ( Figure 2a); this might be due to the complexity of the N. fragrans genome because this species is an allotetraploid. Linkage analysis revealed that NbRGH1-CAPS was linked to the HLA1 locus at a distance of 10.9 cM (Figure 3). Two markers were mapped to other linkage groups: NbRGH3-CAPS to LG2 and NbRGH4 to LG3 ( Figure S1). No marker linked to NbRGH2-dCAPS was detected.   expected, but two bands were unexpectedly detected in N. fragrans (Figure 2a); this might be due to the complexity of the N. fragrans genome because this species is an allotetraploid. Linkage analysis revealed that NbRGH1-CAPS was linked to the HLA1 locus at a distance of 10.9 cM (Figure 3). Two markers were mapped to other linkage groups: NbRGH3-CAPS to LG2 and NbRGH4 to LG3 ( Figure S1). No marker linked to NbRGH2-dCAPS was detected.   Because NbRGH1-CAPS, involved in the Niben044Scf00009821 in the v0.4.4 scaffold, was linked to the HLA1 locus, additional markers were developed using the v1.0.1 scaffold released during the conduct of this study. The Niben044Scf00009821 scaffold was involved in the Niben101Scf06736 scaffold in v1.0.1. Based on the scaffold information, we developed two STS markers: Nb7 and Nb14 (Table S3). Nb7 could be used as a dominant polymorphic marker (Figure 4a). Although Nb14 showed no apparent polymorphism between parents in agarose gel electrophoresis, the CAPS marker Nb14-CAPS, developed based on genome sequences of the parents at Nb14, showed polymorphism (Figure 4b,c). Eventually, the HLA1 locus was located between Nb14-CAPS and NbRGH1-CAPS at a distance of 10.8 involved in the Niben101Scf06736 scaffold in v1.0.1. Based on the scaffold information, we developed two STS markers: Nb7 and Nb14 (Table S3). Nb7 could be used as a dominant polymorphic marker (Figure 4a). Although Nb14 showed no apparent polymorphism between parents in agarose gel electrophoresis, the CAPS marker Nb14-CAPS, developed based on genome sequences of the parents at Nb14, showed polymorphism (Figure 4b,c). Eventually, the HLA1 locus was located between Nb14-CAPS and NbRGH1-CAPS at a distance of 10.8 and 10.9 cM to HLA1, respectively ( Figure  3). The distance between these markers was equivalent to a 682-kb interval in the genome sequence of N. benthamiana.

Discussion
In the genus Nicotiana, only limited interspecific crosses have been used to produce segregating populations available to construct linkage maps, for example, the cross between N. plumbaginifolia and N. longiflora [53]. Both N. debneyi and N. fragrans have 48 chromosomes and are closely related to each other [51,52]. Furthermore, a high frequency of bivalents was observed in the F1 hybrid between N. debneyi and N. fragrans, suggesting that crossover would occur during meiosis. As a result, the HLA1 locus could be mapped to linkage group LG1 (Figures 3 and S2).
Compared to RAPD analysis, AFLP analysis amplified a greater number of bands per primer (5191/256 in AFLP against 385/80 in RAPD) and yielded a higher percentage of polymorphisms (Table 1). However, both types of DNA markers tended to cluster in some regions. In this regard, clustering of AFLP markers has often been reported [54]. SSR markers that were developed in N. tabacum also showed a higher percentage of polymorphisms than RAPD markers (Table 1). Although several SSR markers were dominantly scored; this might be because N. tabacum SSR markers amplified only single bands even in N. tabacum and showed somewhat low transferability to other Nicotiana species: 72% transferability for N. africana, 41% for N. benthamiana, and 50% for N. suaveolens [44,55].
Several markers deviated from expected Mendelian ratios were observed ( Figure S1). Segregation distortion might be caused by several reproductive isolation mechanisms [56,57]. Therefore, some reproductive isolation mechanisms might exist between N. debneyi and N. fragrans.
Because N. debneyi and N. fragrans possess 48 chromosomes, it is expected that the saturated linkage map consists of 24 linkage groups when constructed using the F2 population from the cross N. debneyi × N. fragrans. In the present study, however, the linkage map consisted of only two major (LG1 and LG2) and 24 minor linkage groups

Discussion
In the genus Nicotiana, only limited interspecific crosses have been used to produce segregating populations available to construct linkage maps, for example, the cross between N. plumbaginifolia and N. longiflora [53]. Both N. debneyi and N. fragrans have 48 chromosomes and are closely related to each other [51,52]. Furthermore, a high frequency of bivalents was observed in the F 1 hybrid between N. debneyi and N. fragrans, suggesting that crossover would occur during meiosis. As a result, the HLA1 locus could be mapped to linkage group LG1 (Figure 3 and Figure S2).
Compared to RAPD analysis, AFLP analysis amplified a greater number of bands per primer (5191/256 in AFLP against 385/80 in RAPD) and yielded a higher percentage of polymorphisms (Table 1). However, both types of DNA markers tended to cluster in some regions. In this regard, clustering of AFLP markers has often been reported [54]. SSR markers that were developed in N. tabacum also showed a higher percentage of polymorphisms than RAPD markers (Table 1). Although several SSR markers were dominantly scored; this might be because N. tabacum SSR markers amplified only single bands even in N. tabacum and showed somewhat low transferability to other Nicotiana species: 72% transferability for N. africana, 41% for N. benthamiana, and 50% for N. suaveolens [44,55].
Several markers deviated from expected Mendelian ratios were observed ( Figure S1). Segregation distortion might be caused by several reproductive isolation mechanisms [56,57]. Therefore, some reproductive isolation mechanisms might exist between N. debneyi and N. fragrans.
Because N. debneyi and N. fragrans possess 48 chromosomes, it is expected that the saturated linkage map consists of 24 linkage groups when constructed using the F 2 population from the cross N. debneyi × N. fragrans. In the present study, however, the linkage map consisted of only two major (LG1 and LG2) and 24 minor linkage groups ( Figure S1). This might be because of the inadequate number of markers for linkage map construction and/or the clustering of many markers. Nevertheless, considering that markers were not distributed uniformly among all linkage groups, it is more likely that enough crossover events did not occur, although a high frequency of bivalents was observed in the F 1 hybrid. Therefore, using the F 2 population from the cross N. debneyi × N. fragrans, only limited positions of loci such as the HLA1 locus can be mapped through linkage analysis.
Based on the N. benthamiana v1.0.1 genome scaffold, nine genes were found in the region between Nb14-CAPS and NbRGH1-CAPS where the HLA1 locus was mapped (Table S4). Among these genes, three genes (Scf06736g04011.1, Scf06736g06007.1, and Scf06736g08001.1) might be associated with plant responses to pathogens based on the description available at the Sol Genomics Network. Therefore, these genes may be good candidates for HLA1. Nevertheless, we cannot exclude the possibility that another locus gene is responsible for HLA1 because different genes might exist between N. benthamiana and N. debneyi or N. fragrans in the concerned region due to possible chromosomal differences generated during the process of speciation. Fine mapping is required to definitively identify candidate genes for HLA1.

Plant Materials
An F 1 hybrid between N. debneyi (2n = 48; homozygous for Hla1-1) and N. fragrans (2n = 48; homozygous for hla1-2) obtained in a previous study [39] was used to produce the F 2 population. A total of 125 F 2 plants were used as mapping populations. Each F 2 plant was test-crossed with N. tabacum 'Samsun NN' (2n = 48) to estimate the genotype at the HLA1 locus. The seeds were aseptically sown on 1/2 MS medium [58] supplemented with 1% sucrose and 0.2% Gelrite (pH 5.8) and then cultured at 25 • C (16 h light/8 h dark, approximately 100 µmol m −2 s −1 ). At least 10 plants per each cross were observed for lethality. F 2 plants were determined to be homozygous for Hla1-1 when all the progeny obtained from the cross with N. tabacum were inviable, heterozygous when the progeny was segregated for viable and inviable (approximately 1:1), and homozygous for hla1-2 when all the progenies were viable.

DNA Extraction
Total DNA was extracted from the leaves of parental, F 1 , and F 2 plants using the cetyltrimethylammonium bromide (CTAB) method [59] and used for linkage analysis using DNA markers.

RAPD Analysis
RAPD analysis was carried out as described by Williams et al. [42] with some minor modifications. Briefly, 80 random 10-mer oligonucleotide primers (Kit A to D) were obtained from Operon Technologies (Alameda, CA, USA). Reaction mixtures contained 1× Robust Buffer (BioAcademia, Osaka, Japan), 0.2 mM each dNTP, 0.5 µM primer, 10 ng template DNA, and 0.5 U Taq DNA polymerase (BioAcademia) in a total volume of 10 µL. PCR amplification was performed using a TProfessional Basic Thermocycler (Biometra, Göttingen, Germany) programmed for 1 min at 94 • C for initial denaturation, followed by 45 cycles of 30 s at 94 • C, 1 min at 36 • C, 2 min at 72 • C, and a final extension of 3 min at 72 • C. PCR products were separated by electrophoresis in a 1.5% agarose gel in TBE buffer and stained with ethidium bromide to visualize the DNA bands. During the analysis, only intense and clear DNA bands were observed. Each RAPD marker was named as the primer and the subsequent number that indicated the estimated DNA fragment size.

AFLP Analysis
AFLP analysis was carried out as described by Vos et al. [43] with some modifications. Total DNA was double-digested with EcoRI and MseI and ligated to the EcoRI adaptor and the MseI adaptor. The preamplification reaction was carried out using EcoRI+A (E01) and MseI+C (M02) primers (Table S5). After the reaction, the PCR product was diluted 10-fold with sterilized distilled water and used as a template for the second amplification reaction. Primers with three selective nucleotides, 16 EcoRI+ANN (N indicates A, T, G, or C; E31 to E46) and 16 MseI+CNN (M47 to M62) primers (Table S5), were used for the second amplification reaction with all possible combinations (256 primer pairs). In both amplification reactions, KAPA Taq EXtra DNA polymerase (Kapa Biosystems, Wilmington, MA, USA) was used as the enzyme. PCR amplification was performed using the PC-818A Program Temp Control System (Astec, Fukuoka, Japan). The second PCR products were separated by electrophoresis on an 8% polyacrylamide gel and silver stained [60]. Each AFLP marker was named as the primer combination and subsequent number that indicated the estimated DNA fragment size.

SSR Analysis
An N. tabacum linkage map consisting of 24 linkage groups was constructed using SSR markers in N. tabacum [44]. Sixty-six SSR markers (Table S6) were selected to cover all linkage groups in N. tabacum and were detected as follows: Reaction mixtures contained 1× Robust Buffer (BioAcademia), 0.2 mM each dNTP, 0.2 µM of each primer, 10 ng template DNA, and 0.25 U Taq DNA polymerase (BioAcademia) in a total volume of 10 µL. PCR amplification was performed using a TProfessional Basic Thermocycler (Biometra) programmed for 3 min at 94 • C for initial denaturation, followed by 35 cycles of 30 s at 94 • C, 1 min at 50 • C, 5 s at 72 • C, and a final extension of 5 min at 72 • C. PCR products were separated by electrophoresis in a 3% agarose gel in TBE buffer and stained with ethidium bromide to visualize the DNA bands. If several loci were detected using one primer pair, they were distinguished by adding a lower-case letter to the end of the marker name.

DNA Markers Related to Disease Resistance
Two SCAR markers-SUBC180.251 [48] and Mil275 [49]-linked to blue-mold resistance in N. tabacum were used, assumed to be introduced from N. debneyi [48,49]. These markers were detected using conventional PCR.
The draft genome sequence of N. benthamiana [50] was released through the Solanaceae (Sol) Genomics Network [61] (https://solgenomics.net/, accessed on 27 September 2021). Because disease resistance-related genes were not identified based on the genome sequence at the time of initiation of this research, we searched for disease resistance gene homologs. One-hundred-and-four TIR-NBS-LRR proteins in Arabidopsis thaliana were identified based on the protein sequences (TAIR10_pep_20101214) obtained from The Arabidopsis Information Resource [62] (https://www.arabidopsis.org/, accessed on 27 September 2021) and were used as query sequences in BLASTP searches (e-value < 1 × 10 −5 ) against the N. benthamiana protein sequences (Niben.genome.v0.4.4.proteins.annotated.fasta), downloaded from the Sol Genomics Network. Based on the N. benthamiana proteins identified in the searches, the corresponding transcripts were identified from N. benthamiana transcript sequences (Niben.genome.v0.4.4.transcripts.annotated.fasta) downloaded from the Sol Genomics Network. To determine the genomic regions encoding the transcripts identified, the transcript sequences were used as query sequences in BLASTN searches against the N. benthamiana genome scaffolds (Niben.genome.v0.4.4.scaffolds.nrcontigs.fasta) downloaded from the Sol Genomics Network. Based on the genomic sequences of four disease resistance gene homologs identified in N. benthamiana (Table S2), four STS markers, NbRGH1 to NbRGH4, were developed (Table S3). Subsequently, two STS markers, Nb7 and Nb14, located on the same scaffold as NbRGH1, were developed based on the newly available N. benthamiana genome scaffolds (Niben.genome.v1.0.1.scaffolds.nrcontigs.fasta) from the Sol Genomics Network. Primers for each marker were designed using Primer3 software [63]. After conventional PCR amplification, the PCR products were separated by electrophoresis on a 1.5% agarose gel or 8% polyacrylamide gel (only NbRGH4).

Linkage Analysis
After determining the HLA1 genotype and scoring the genotypes at DNA markers in F 2 plants, each locus was tested for goodness of fit to the expected 3:1 (dominant) or 1:2:1 (codominant) ratios by a chi-square test at the 5% level of significance. The linkage map