The presence of this large duplicated 1BS-1RS colinear region in both T-9 and 1B+40 was particularly surprising given their different chromosomal configurations. To understand the structure of these chromosomes, we first determined the borders of the 1RS region in T-9 and 1B+40 by mapping the exome capture reads at high stringency to the combined CS RefSeq v1.1 chromosome 1B and 1RSAK58. The first 1RS border proximal to the 1BS insertion in T-9 and T-21 was located approximately between 3,096,772 and 3,151,733 bp in the 1RSAK58 genome and it was conserved in 1RSRW . The second 1RS border distal to the 1BS insertion in 1B+40 was located approximately between 10,071,332 and 10,079,335 bp in the 1RSAK58 genome . Analyses of the 1RS and 1BS border regions suggest that the breakpoints in both T-9 and 1B+40 involved orthologous regions in the 1BS and 1RS genomes , which is expected since these lines were generated using the ph1b mutation that promotes homoeologous recombination . A more detailed analysis of the 1RS/1BS break point region in 1B+40 revealed the presence of a 0.7–1.2 Mb secondary inversion in the border of the 1RS segment nested within the large 13.875 Mb inversion.This explains why the crossover in this region did not generate a dicentric chromosome and acentric fragment,nft hydroponic which is expected from a crossover event within an inverted region . Except for this small secondary inversion, the rest of the genes in the distal 1RS segment are in the opposite orientation to the order of the genes in the 1BS segment, explaining the duplication of the genes in this region.
The analysis of the break point region in T-9 failed to reveal any obvious secondary inversion , but we cannot rule out the possibility of small inversions affecting a few genes since we do not have the complete genome of wheat cultivar Pavon, which is the source of the 1BS segment. We hypothesize that, similarly to what we observed in 1B+40, a small secondary inversion within the large 13.875 Mb inversion may have generated a small region with a common centromere–telomere orientation facilitating the origin of the T-9 and T-21 recombinant chromosomes. As in 1B+40, we predict that most genes in the 1BS and 1RS segments at both sides of the break point point in T-9 and T-21 are duplicated and in an inverted order . Regardless of the mechanism that generated the T-9 and 1B+40 chromosomes, the crossover within the 4.8 Mb of the overlapping 1BS region that originated the 1RSRW chromosome is expected to generate a duplication of the 1RS segment between 3.1 and 10.1 Mb. This duplication is clearly visible in Supplemental Figure S4 , where we plotted the ratios between the reads per kilobase per gene for 1RSRW/1RS from the NimbleGen exome capture experiment vs. the position in 1RSAK58. Since the 1BS segment inserted in 1RSRW isorthologous to the duplicated 1RS region, all genes in the 1BS segment are triplicated . As expected, the borders of the 1RSRW duplicated region coincide with the borders of the 1BS–1RS breakpoints in T-9 and 1B+40 . The 1RSRW/1RS ratios between the normalized number of reads from the exome capture were significantly higher in the region inside the duplication than in the region outside the duplication . These results support the chromosome models presented in Figure 2a.To test if these changes in gene dosage affect root length, we generated a radiation mutant population for 1RSWW. Among the mutagenized 1RSWW plants, we identified only one line that carried a deletion in the critical region.
To control for possible background mutations in the phenotypic experiments, we generated two sister lines, designated 1RSWW-del-8 and 1RSWW-del-10, using independent back crosses. For the exome capture, these two lines served as replicates in the determination of the deletion borders . To study the border of the deletion within the 1BS segment, we compared the ratios 1RSRW/1BS with the 1RSWW-del-8/10/1BS ratios. We observed a significant drop in the 1RSWW-del-8/10/1BS ratios between the normalized numbers of reads per kb between 7,995,986 and 8,116,677 bp that was not observed in 1RSRW . This deletion border was supported by significantly higher ratios in the 1BS region between 4.8 and 7.9 Mb than in the region between 8.2 and the end of the 1BS segment at 9.6 Mb in the deletion lines relative to the 1RSRW control with a complete 1BS segment . To determine if the radiation deletion extended into the 1RS segment of 1RSRW and to determine its size, we analyzed the ratios of reads/kb in 1RSRW/1RS, 1RSRW-del-8/1RS, and 1RSRW-del-10/1RS along the 1RSAK58. These analyses showed that the deletion was within the 1RS duplicated region and that the border of the 1RS deletion was between 6,010,958 and 6,422,733 bp in 1RSAK58 . This border was supported by significantly smaller ratios in the deleted 1RS region than between 6.4 and 10 Mb in the deletion lines but not in the 1RSRW control . Interestingly, since the genes in 1RS and 1BS regions are in inverted orientation around the break point, all genes deleted in the 1RS region have orthologues in the deleted 1BS region . Wheat genes TraesCS1B02G017200 and TraesCS1B02G017300 were deleted in 1RSWW-del-8/10 but their rye orthologues were outside the deleted region . Because of the 1RSWW-del-8/10 deletion, the 1RS region between 3.1 and 6.0 Mb is present in one copy in 1RSWW-del-8/10 and in two copies in 1RSRW, which also carries an additional 1BS orthologous copy.
Previous field studies demonstrated that cultivar Hahn carrying the standard 1RS.1BL translocation had longer roots, better access to water, and significantly higher grain yields than isogenic Hahn lines carrying the 1RSRW chromosome . Hydroponic studies confirmed that 2 wk after germination, the roots in Hahn-1RSRW showed a significant reduction in the elongation rate, altered gradients of reactive oxygen species, and the emergence of lateral roots close to the RAM . This earlier reduction in root growth rates in 1RSRW relative to 1RS was also observed in this study, even in experiments that showed variable overall root growth responses . We initially assumed that the 4.8 Mb 1BS segment in the 1RSRW chromosome arm was the result of a homologous recombination event between the overlapping 1BS segments of lines T-9 and 1B+40 and that, therefore, the 1BS wheat genes have replaced the orthologous 1RS rye genes.Given the known positive effect of the 1RS translocation on drought tolerance in wheat, we hypothesized that the lost 1RS genes were the cause of shorter roots in 1RSRW. However, the exome capture sequencing of 1RS and 1RSRW demonstrated that both the 1BS and its orthologous 1RS segment were still present in 1RSRW, disproving our original hypothesis. Our second hypothesis was that wheat genes present in the 4.8-Mb 1BS segment inserted in 1RSRW could be responsible for the shorter roots. However, the characterization of the Hahn-T-18 line, which carries a 17-Mb distal 1BS segment and has no identifiable duplications, provided evidence against this hypothesis. The roots of T-18 were slightly longer than those in 1RS at the initiation of the measurements but showed no significant differences in their root growth rates after that day . When the 1BS segment was combined with the 1RS segment in the Hahn-T-21 and Hahn-1B+40, the roots were significantly longer than the roots of 1RSRW and slightly, but not significantly, shorter than the roots in the control 1RS line . Taken together, these results provided conclusive evidence that the presence of the wheat genes in the 1BS segment alone was not responsible for the short roots 1RSRW and disproved our second hypothesis. Our third, and still current, hypothesis,nft system is that the change in gene dosage generated by the duplications of the 1BS and 1RS colinear regions was responsible for the arrest in the seminal root growth. The lack of differences in root growth rate between T-18 and 1RS between 9 and 28 d suggest that the genes in the 1BS segment are not responsible for the reduced growth rate in 1RSRW during the same period . The 1BS-1RS duplication in T-21 and 1B+40 resulted only in a minor decrease in growth rate relative to 1RS and their final root lengths were significantly longer than in 1RSRW . As T-21 tended to be shorter than 1B+40 in both experiments, we cannot rule out the possibility that their different proximal regions may contribute to modulate the effect of the 2R+2B duplication on root length. These results suggest that adding duplicated 1BS genes has a smaller effect on seminal root growth than adding more copies of the 1RS orthologues. The stronger effect of the 1RS segment was evident in plants heterozygous for the 1RSRW chromosome , which showed seminal root length intermediate to that of 1RS and 1RSRW . Based on this result, we hypothesize that the duplication of the 1RS region in 1RSRW is the main driver for shorter roots in this line, but we do not entirely discard the idea that the genes in the 1BS segment may also contribute to the reduced root growth when combined with additional 1RS orthologues. The dosage hypothesis was reinforced by the hydroponic experiments with the radiation-mutants 1RSWW-del-8 and 1RSWW-del-10 back crossed independently to both to Hahn- 1RSRW and Hahn-1RS. In the hydroponic experiments using the back cross lines segregating for the deletions and 1RSRW, the roots of the deletion lines were significantly longer than those of the sister lines carrying at least one 1RSRW chromosome.
By contrast, in the lines segregating for the deletions and the 1RS chromosome, we observed no significant differences in root length between the homozygous deletions and their sister lines carrying at least one 1RS chromosome . The four consecutive back crosses of 1RSWW-del-8 and 1RSWW-del-10 into 1RSRW and 1RS minimized the chances of a possibly confounding effect of independent deletionsin other chromosomes of the radiation mutants. However, they did not rule out the possibility of a confounding effect of a linked deletion in 1RS. Using the exome capture, we did find a linked missing 1RS region corresponding to the orthologous rye region replaced by the proximal wheat segment in homozygotes for the 1RSWW chromosome. We have previously shown that the proximal wheat segment has no effect on root length and confirmed this result in the hydroponic experiments presented in this study . The exome capture data also allowed us to determine the length of the 1RS deleted segment in the deletion lines and to establish that the 1BS and 1RS deletions include mostly orthologous genes . Therefore, the homozygous 1RSWW-del-8 and 1RSWW-del-10 lines are expected to lose two gene copies in 1BS and two in 1RS, changing the gene dosage from 4R+2B to 2R. This hypothesis explains the identical seminal root size observed in the 1RS and the homozygous deletion lines . One limitation of the exome capture assays is that they are closed systems and some genes are not included, which resulted in annotated genes with no reads. We eliminated those genes for the analysis used to delimit the borders of the 1RS–1BS recombination events or of the duplicated 1RS region . This likely resulted in a slight overestimate of the size of the candidate gene regions and the number of potential candidate genes.Once we established conservative borders of the 1BS and 1RS deleted regions in 1RSRW-del-8/10, we considered all the annotated genes in these regions as candidates regardless of their presence in the exome capture. The 1RSAK58 genome is very close to the 1RS present in our lines, so it probably provides a good representation of the rye candidate gene region. However, the CS RefSeq 1.1 used as 1BS reference is not identical to the 1BS Pavon segment, and therefore, we cannot rule out the possibility of genes present in Pavon that are not present in the wheat reference. Since the deletion mutants showed similar root length as the 1RS line, we decided to focus on the 14 genes expressed in roots that were deleted in both the 1RS duplication and the adjacent and orthologous 1BS insertion . Although the annotated functions of these genes based on conserved domains and homology will require further experimental validation, the list is useful to summarize their inferred functions and to provide a preliminary idea of potential candidate genes.