Chapter one of this thesis describes the discovery and validation of a natural polymorphism in the A genome copy of the FLOWERING LOCUS T2 (FT2), a gene which increases spikelet number per spike (SNS) in wheat. A previous study indicated that loss-of-function mutations in FT2 increased SNS, but also decreased fertility, negating the positive effects on grain yield. Fortunately, we found a natural amino acid change in the FT-A2 protein at position 10 from aspartic acid (D) to alanine (A) (henceforth, D10A), that was associated with increased SNS and no negative effects on fertility. We used a high-density genetic map to delimit the candidate gene region to 28 genes, and then determined that among them only FT-A2 had a non-synonymous polymorphism (D10A) consistent in all studied mapping populations. We concluded that FT-A2 was likely the cause for the increased SNS and validated its effect in a hexaploid spring and winter wheat population.Investigating the frequency of the A10 allele, we found that while it was present in 56% of the common wheat accessions, it was present in less than 1% of the durum accessions analyzed. This rapid increase from durum wheat to common wheat suggests that the A10 allele had undergone positive selection in common wheat and may be a useful target for improving grain yield in durum and common wheat.
Chapter two of this thesis investigates the effects of the introgression of the FT-A2 A10 allele into four high-yielding durum wheat varieties. We found that all lines with the A10 allele had significantly higher SNS with no significant decreases in fertility. However, due to variety x FT-A2 interactions, grain number per spike (GNS) was significantly increased in only three of the four varieties. Unfortunately, the increases in SNS and GNS were offset by a significant decrease in kernel weight in all four varieties, resulting in no significant differences in spike yield or grain yield per plot. There was a significant negative correlation between GNS and thousand kernel weight (r = -0.42, P = 0.0006), which we hypothesize reflects source limitations in our environment that resulted in incomplete filling of the extra grains. Incorporation of the FT-A2 A10 allele into high bio-mass genotypes which are less source-limited and the evaluation of sister isogenic lines in different environments will be necessary to test the usefulness of this allele in durum breeding programs.
Chapter three of this thesis describes the identification and characterization of bZIPC1, a bZIP-containing transcription factor from the C-subfamily, which we identified utilizing a yeast-two hybrid screen with the FT2 protein as bait. Combined loss-of-functions mutations in bZIPC-A1 and bZIPC-B1 (bzipc1) in tetraploid wheat resulted in drastic reduction in SNS with a limited effect on heading date. Expression studies revealed that genes previously known to affect SNS were not significantly affected by the bzipc1 mutations, suggesting that bZIPC1 may affect SNS through a different pathway. Investigating the natural variation in the bZIPC-B1 region, revealed three major haplotypes (H1-H3), with the H1 haplotype showing significantly higher SNS, GNS, and spike yield than both the H2 and H3 haplotypes, as well as an increased frequency from the ancestral tetraploids to the modern durum and common wheat varieties. We developed markers of the two non-synonymous SNPs that differentiate the H1 haplotype from H2 and H3 so the H1 haplotype can be identified, introgressed, and deployed in wheat breeding programs.
