Linkage mapping identifies a non-synonymous mutation in FLOWERING LOCUS T (FT-B1) increasing spikelet number per spike

Abstract

Total spikelet number per spike (TSN) is a major component of spike architecture in wheat (Triticum aestivum L.). A major and consistent quantitative trait locus (QTL) was discovered for TSN in a doubled haploid spring wheat population grown in the field over 4 years. The QTL on chromosome 7B explained up to 20.5% of phenotypic variance. In its physical interval (7B: 6.37-21.67 Mb), the gene FLOWERING LOCUS T (FT-B1) emerged as candidate for the observed effect. In one of the parental lines, FT-B1 carried a non-synonymous substitution on position 19 of the coding sequence. This mutation modifying an aspartic acid (D) into a histidine (H) occurred in a highly conserved position. The mutation was observed with a frequency of ca. 68% in a set of 135 hexaploid wheat varieties and landraces, while it was not found in other plant species. FT-B1 only showed a minor effect on heading and flowering time (FT) which were dominated by a major QTL on chromosome 5A caused by segregation of the vernalization gene VRN-A1. Individuals carrying the FT-B1 allele with amino acid histidine had, on average, a higher number of spikelets (15.1) than individuals with the aspartic acid allele (14.3) independent of their VRN-A1 allele. We show that the effect of TSN is not mainly related to flowering time; however, the duration of pre-anthesis phases may play a major role.

Figure 1

Phenotypic distribution of the best linear unbiased estimations (BLUEs) of the investigated traits. Arrows indicate the phenotypes of the parental lines with parent A corresponding to TRI-10703 and parent B to TRI-5310.

Figure 2

Pearson’s correlation coefficient of the best linear unbiased estimations (BLUEs) of total spikelet number (TSN), spike length (SL), heading date (HD), and flowering time (FT). Significance of the correlations is indicated with ***P < 0.001.

Figure 3

Quantitative trait loci (QTL) map of total spikelet number (TSN), spike length (SL), heading date (HD), and flowering time (FT) in a DH-population of spring wheat grown over 4 years and their corresponding best linear unbiased estimations (BLUEs). The horizontal dashed lines indicate the corresponding logarithm of odds (LOD) threshold estimated at α = 0.05 with 1000 permutations to assess the QTL significance.

Figure 4

Allele-wise phenotypic distribution of the KASP marker designed for FT-B1 of the best linear unbiased estimations (BLUEs) for the investigated traits in dominant (A) or recessive (B) VRN-A1 background. Significance of the mean differences of the marker alleles obtained from a Wilcoxon rank-sum test.

Figure 5

Intraspecific, intragenomic and interspecific diversity for the FT-B1 mutation. All values are percentages. (a) Allele frequency across 135 wheat varieties sequenced with the Axiom marker AX-94810990 of the nucleotide substitution, with C translating to histidine (H) and G translating to aspartic acid (D). (b) Allele frequency for the corresponding amino acid position across 70 PEBP genes annotated in the wheat reference genome. (c) Allele frequency for the corresponding amino acid position across 377 homologs retrieved from the PROVEAN analysis across a large diversity of plants.

Figure 6

Sequence alignment of the 20 first amino acids of a subset of FT-B1 homologs including the homoeologues on chromosome 7A and 7D. The wheat reference sequence (Chinese Spring), the sequences of the two parental lines, and FT-B1 homologs in diverse species viz. Hordeum vulgare, Brachypodium distachyon, Oryza sativa, Zea mays, and Arabidopsis thaliana. The blue arrow indicates the position of the aspartic acid (D) to histidine (H) mutation present in the parent A and associated with elevated total spikelet number. The amino acids are color coded according to their side chain polarity with nonpolar (gold), polar (green), basic polar (blue), and acidic polar (red).