A GWAS study highlights significant associations between a series of indels in a FLOWERING LOCUS T gene promoter and flowering time in white lupin (Lupinus albus L.)

Abstract

Background: White lupin (Lupinus albus L.) is a high-protein Old World grain legume with remarkable food and feed production interest. It is sown in autumn or early spring, depending on the local agroclimatic conditions. This study aimed to identify allelic variants associated with vernalization responsiveness, in order to improve our knowledge of legume flowering regulatory pathways and develop molecular selection tools for the desired phenology as required for current breeding and adaptation to the changing climate.

Results: Some 120 white lupin accessions originating from a wide range of environments of Europe, Africa, and Asia were phenotyped under field conditions in three environments with different intensities of vernalization, namely, a Mediterranean and a subcontinental climate sites of Italy under autumn sowing, and a suboceanic climate site of France under spring sowing. Two hundred sixty-two individual genotypes extracted from them were phenotyped in a greenhouse under long-day photoperiod without vernalization. Phenology data, and marker data generated by Diversity Arrays Technology sequencing (DArT-seq) and by PCR-based screening targeting published quantitative trait loci (QTLs) from linkage map and newly identified insertion/deletion polymorphisms in the promoter region of the FLOWERING LOCUS T homolog, LalbFTc1 gene (Lalb_Chr14g0364281), were subjected to a genome-wide association study (GWAS). Population structure followed differences in phenology and isolation by distance pattern. The GWAS highlighted numerous loci significantly associated with flowering time, including four LalbFTc1 gene promoter deletions: 2388 bp and 2126 bp deletions at the 5' end, a 264 bp deletion in the middle and a 28 bp deletion at the 3' end of the promoter. Besides LalbFTc1 deletions, this set contained DArT-seq markers that matched previously published major QTLs in chromosomes Lalb_Chr02, Lalb_Chr13 and Lalb_Chr16, and newly discovered QTLs in other chromosomes.

Conclusions: This study highlighted novel QTLs for flowering time and validated those already published, thereby providing novel evidence on the convergence of FTc1 gene functional evolution into the vernalization pathway in Old World lupin species. Moreover, this research provided the set of loci specific for extreme phenotypes (the earliest or the latest) awaiting further implementation in marker-assisted selection for spring- or winter sowing.

Keywords: Flowering; Flowering locus T; GWAS; Indel; Promoter; QTL; Vernalization.

Fig. 1

Summary of population structure analysis of 262 white lupin genotypes. The panels show STRUCTURE diagrams under different K-values (A), mean total number of growing degree days (GDDs) from sowing to flowering observed in 5 environments (B), and geographic localization of germplasm samples (C). Environments are as follows: spring sowing in controlled conditions with an absolute lack of vernalization (Greenhouse 2020 and 2021), autumn sowing in field conditions with strong vernalization (Lodi) and moderate vernalization (Sanluri) and spring sowing in field conditions with mild vernalization (Saint Sauvant). In total, 6765 markers were used for population structure analysis. Different colors mark major clusters for different K-grouping scenarios. Clustering at K8 is indicated by vertical lines and numbers from 1 to 8. GDDs were visualized by scale from blue (minimum value for the environment) through white (mean value) to red (maximum value)

Fig. 2

The principal component analysis (PCA) showing separation of major clusters of white lupin genotypes formed at K8. In total, 6765 markers were used for population structure analysis

Fig. 3

Results of genome-wide association study (GWAS) in white lupin diversity panel. The K = 5 was selected as the representative number of clusters for population structure. In total, 6765 markers obtained for 262 genotypes were analyzed. Only markers that revealed significant associations in at least three environment × trait combinations are shown. Flowering time was observed in 5 environments: spring sowing in controlled conditions with absolute lack of vernalization (Greenhouse 2020 and 2021) and in field conditions - autumn sowing with strong (Lodi) or moderate vernalization (Sanluri), and spring sowing with mild vernalization (Saint Sauvant). Trait abbreviations are as follows: GDDs, the number of growing degree days; Days, the number of calendar days; Hours, the total number of photoperiod hours; VF, the cumulative vernalization effectiveness of daily temperature. All traits were counted from the sowing date until the start of flowering. Markers were sorted from the negative to the positive effect of an alternative allele on the number of days to flowering. Colors indicate the impact and direction of effects according to minimum (blue), zero (white) and maximum (red) values provided on the scale. Asterisk (*) indicates Benjamini-Hochberg false discovery rate (FDR)-adjusted P-value in the following scheme: ***, p < 0.0001; **, 0.0001 ≤ p < 0.001; *, 0.001 ≤ p ≤ 0.05

Fig. 4

Manhattan plots for genome-wide association study (GWAS) in the white lupin diversity panel. The panels show the cumulative number of growing degree days in Greenhouse 2020 (A), Greenhouse 2021 (B), Lodi (C), Sanluri (D), Saint Sauvant (E) and the cumulative vernalization effectiveness of daily temperature in Sanluri (F). K = 5 was selected as the representative number of clusters for population structure. In total 6765 markers and 262 genotypes were analyzed. P-values are presented on y axis and chromosome positions on x axis. Benjamini-Hochberg false discovery rate (FDR)-adjusted P-value < 0.05 corresponds to the threshold indicated on the graphs as a horizontal line. For the sake of clarity, names were shown only for a few most significant markers

Fig. 5

Allelic effects on the number of growing degree days (GDDs) and the cumulative vernalization effectiveness of daily temperature (VF) from sowing to flowering of white lupin for LalbFTc1 gene PCR markers: PR_36b (A), PR_42a (B), PR_58c (C) and PR_71d (D). R stands for the reference allele (0), whereas V for an alternative allele (2) – see Table 3. Flowering time was observed in 4 environments: spring sowing in controlled conditions with absolute lack of vernalization (Greenhouse 2020 and 2021) as well as in field conditions - autumn sowing with strong (Lodi) or moderate vernalization (Sanluri), and spring sowing with mild vernalization (Saint Sauvant). Diamonds indicate mean values

Fig. 6

Allelic effects on the number of growing degree days (GDDs) and the cumulative vernalization effectiveness of daily temperature (VF) from sowing to flowering of white lupin for seven DArT-seq markers, Chr02_2625514_D (A), Chr02_2625564_D (B), Chr13_12561729_D (C), 13913452_D (D), Chr16_366800 (E), Chr16_572706 (F) and Chr16_788665 (G), originating from the chromosome segments overlapping with previously reported major QTL regions. R stands for the reference allele (0), V for an alternative allele (2), and H for a heterozygote (1). Flowering time was observed in four environments: spring sowing in controlled conditions with absolute lack of vernalization (Greenhouse 2020 and 2021) as well as in field conditions - autumn sowing with strong (Lodi) or moderate vernalization (Sanluri), and spring sowing with mild vernalization (Saint Sauvant). Diamonds indicate mean values

Fig. 7

Allelic effects on the number of growing degree days (GDDs) and the cumulative vernalization effectiveness of daily temperature (VF) from sowing to flowering of white lupin for seven DArT-seq markers, Chr04_2175652 (A), Chr06_14434379 (B), Chr08_3090141_D (C), Chr08_3090075_D (D), Chr10_13080319 (E), Chr11_18778542 (F) and Chr25_4002891_D (G), tagging novel QTL regions. R stands for the reference allele (0), V for an alternative allele (2), whereas H for a heterozygote (1). Flowering time was observed in 4 environments: spring sowing in controlled conditions with absolute lack of vernalization (Greenhouse 2020 and 2021) as well as in field conditions - autumn sowing with strong (Lodi) or moderate vernalization (Sanluri), and spring sowing with mild vernalization (Saint Sauvant). Diamonds indicate mean values

Fig. 8

Linkage disequilibrium (LD) plots for white lupin genomic regions carrying significant markers localized in chromosomes Lalb_Chr14 (A), Lalb_Chr02 (B), Lalb_Chr08 (C), Lalb_Chr16 (D) and Lalb_Chr13 (E). The r² values between significant SNPs are shown. Red indicates high measures of LD, while blue indicates low LD.