Plant heights of parents and F2 plants
The plant height of JAM with 102.33 ± 0.58 cm was higher than that of DPW with 81.67 ± 1.20 cm. The F2 plant heights exhibited four phenotypes including dwarf, two semi-dwarf, and tall types (Fig. 1A), which varied from 40 to 163 cm (Fig. 1B). The plant numbers of dwarf, semi-dwarf, and tall types met an expected Mendelian segregation ratio 1 : 6 : 9 (χ2 = 5.37, p < 0.05) (Fig. 1B), suggesting that two dwarfing genes, Rht22 in JAM and Rht-B1b in DPW, contributed to the dwarfism. The plant height of JAM was lower than that of GM with 153.33 ± 0.79 cm (Fig. 1C). Their F2 plants exhibited semi-dwarf and tall phenotypes with 67-164 cm (Fig. 1D). The segregation ratio of semi-dwarf : tall was 1 : 3 (χ2 = 2.43, p < 0.05) (Fig. 1D), indicating that Rht22 as a single gene in JAM contributed to the semi-dwarfism.
High-density genetic map
A total of 171 lines of F7 RIL_DJ and their parents were sequenced. After alignment against the reference genome of wild emmer, 1,512,154,735, 645,735,510, and 332,609,682 clean reads were obtained from RIL_DJ, DPW and JAM, respectively, which individually represented the average sequencing depth with 11.35, 7.94 and 6.85 times (STable 2).
Totally, 230,562 SLAF markers of aa × bb between DPW and JAM were produced (STable 3). After quality control in RIL_DJ, 8091 high quality SNP markers were obtained to construct a genetic map (Table 1). Out of 8091 markers, 1022 skeleton markers (unique loci) were clustered into 14 linkage groups with approximately equal number between the A (507) and B genomes (515). Two translocation linkage groups, 1AS/7AL and 7AS/1AL, were constructed (Fig. 2). The genetic map covered 2201.36 cM with the A genome 1238.48 cM and the B genome 962.88 cM. The average density of skeleton markers was 2.15 cM with 73 markers on each chromosome; the largest gap between two adjacent skeleton markers varied from 8.58 (4B) to 34.19 cM (1B) (Table 1). For individual chromosome, the genetic length spanned from 36.46 (7B) to 242.82 cM (5A); the number of skeleton markers varied from 23 (7B) to 120 (7A); and the average length between two adjacent markers ranged from 1.16 (4B) to 3.63 cM (6A). For seven homologous groups, the length of the individual group ranged from 221.74 (group 7) to 447.24 cM (group 5), and the number of skeleton markers varied from 118 (group 6) to 187 (group 5). A total of 229 (9.57%) non-recombinant chromosomes were investigated among 171 × 14 = 2394 RIL chromosome combinations (Table 2). The number of non-recombinant chromosomes varied from 2 (3B) to 49 (4B), and B genome with 137 was larger than A genome with 92. The correlation coefficient between the genetic and the physical positions of the skeleton markers ranged from 0.39 (1A-1A)-0.93 (7B), indicating that the marker order on this genetic map was highly consistent with that in the wheat physical map (Fig. 3; STable 4).
QTL of plant height
Two QTLs, named as Qph.sicau-4B and Qph.sicau-7A, were detected in both F7 and F8 RIL_DJ population. Qph.sicau-4B was mapped to the position 28.49-30.39 Mb of chromosome 4B with LOD scores ranged from 77.42 to 79.07, which explained 63.89-79.07% of the phenotypic variation (Table 3). This region contains a legendary dwarfing gene Rht-B1b. Thus, its candidate gene is Rht-B1b, which contributes to the semi-dwarfism of DPW (Chai et al. 2021). Qph.sicau-7A was mapped to the position 0.33-4.46 Mb of chromosome 7A with LOD scores ranged from 43.09 to 51.19; it explained 22.06-30.50% of the phenotypic variation (Table 3). Since Qph.sicau-7A was a single QTL underlying plant height on chromosome 7A, it was the candidate region of Rht22.
To confirm the candidate region of Rht22, SLAF-Seq-BSA was performed using two bulks (semi-dwarf and tall bulks) derived from F9 RIL_GJ population and their parents. The average plant height of semi-dwarf bulk was significantly lower than that of tall bulk (Fig. 4). Sequencing produced 99.88, 99.32, 205.50 and 198.87 GB clean data for GM, JAM, tall bulk and semi-dwarf bulk, respectively. Further analysis exploited 1,018,660 SNP and 178,278 InDel markers between two bulks, which were non-uniformly positioned onto 14 chromosomes (Fig. 5A and 5B). Association analysis mapped the candidate region of Rht22 on chromosome 7A with interval 0-1.97 Mb (Fig. 5). After integrating these results, the candidate region of Rht22 was confirmed and narrowed to the position 0.33-1.97 Mb of chromosome 7A.
Accurately genetic region of Rht22 and its candidate genes
To further narrow the genetic region of Rht22, 184 pairs of SSR and 10 pairs of KASP markers were designed at the position 0.33-4.46 Mb of chromosome 7A (STable 1). Out of these markers, 12 pairs of SSR and one pair of KASP makers were used to genotype in F8 and F9 RIL_DJ, F9 RIL_GJ, and F10 RHL_GJ (SFig. 1), and linked to Rht22 in the three populations (Fig. 6). Rht22 was re-localized between KASP marker Xbag295.s53 and SSR marker Xbag295.191 (Fig. 6), and its genomic interval was 0.53-1.48 Mb on chromosome 7A.
According to the annotated information of reference genome of Chinese Spring (v1.0 and v2.1), 18 candidate genes were investigated in this genomic region (Table 4). There are 10 functionally annotated genes and 8 unknown genes. The functionally annotated genes individually encode glycerophosphoryl diester phosphodiesterase-like 2 (GDPDL2), resistance gene analogues 3 (RGA3), flavonol 3-sulfotransferase, achilleol B synthase (F3S), increased DNA methylation 1 (IDM1), achilleol B synthase (ACBSY), serine carboxypeptidase-like 18 (SCPL18), S-nitrosylation regulated protein 1 (SRG1), putative U-box protein 12 (PUB12), WD repeat-containing protein 89 (WDR89), and trehalose-6-phosphate synthase 1 (TPS1).
Rht22-related DEG
To reveal the molecular mechanism of Rht22, two pairs of NIL, NIL_TJ1semi-dwarf (107.07 ± 0.09 cm) and NIL_TJ1tall (170.07 ± 0.09 cm), NIL_TJ2 semi-dwarf (104.70 ± 0.06 cm) and NIL_TJ2tall (167.27 ± 0.28 cm), were used for RNA-seq. A total of 135 DEGs were recommended as the semi-dwarfism related genes. Out of these genes, 96 DEGs were functionally annotated; 13 DEGs were significant up-regulation and other 83 DEGs were down-regulation in NIL_TJ semi-dwarf when compared with NIL_TJ tall (Fig. 7). These DEGs were mainly classed into three pathways including hormone synthesis and signal transduction (32), cell wall metabolism (17), and cell proliferation and cell expansion (47). In hormone synthesis and signal transduction pathway, 26 and 6 DEGs were significant down- and up-regulation in semi-dwarf line, respectively, which were mainly involved in auxin (IAA), Abscisic Acid (ABA), gibberellin (GA), and brassinolide (BR) related metabolic pathways. In cell wall metabolic pathway, 12 and 5 DEGs were significant down- and up-regulation in semi-dwarf line, respectively, which were mainly involved in the synthesis and metabolism of cellulose, hemicellulose, and pectin. In cell proliferation and cell expansion pathway, 45 and 2 DEGs were significant down- and up-regulation in semi-dwarf line, respectively, which were mainly involved in DNA replication, phragmoplast formation and arrangement, and microtubule arrays.
Cell number and size in the first internode
In the first internode, the average cell number in NIL_TJsemi-dwarf (31.00 ± 1.53) was significant lower than that in NIL_TJtall (46.00 ± 3.06) (Fig. 8A). The average cell length in NIL_TJtall (327.08 ± 6.77 μm/per cell) was significant shorter than that in NIL_TJsemi-dwarf (257.85 ± 5.63 μm/per cell), while the average cell width between NIL_TJtall (25.68 ± 1.00 μm/per cell) and NIL_TJsemi-dwarf (25.44 ± 0.90 μm/per cell) were similar (Fig. 8B). These results indicated that Rht22 decreases the cell number to reduce the length of first internode in JAM.