Assessing frost tolerance in white lupin and pea: a methodological study to implement a high-throughput phenotypic platform

Summary in the original language of the document

White lupin (Lupinus albus L.) and pea (Pisum sativum L.) are cool-season pulses of particular interest for reducing the European imports of high-protein feedstuff and improving the soil fertility and agriculture sustainability. Autumn sowing has rising interest for northern Italy to reduce the increasing risk of terminal drought, but frost tolerance remains a key target trait because the trend towards milder winters may emphasize the destructive effect of sudden frosts due to insufficient plant cold acclimation. The increasing year-to-year climatic variability hinders the field-based selection for frost tolerance, indicating the high potential interest of selection under controlled conditions to enable applicability, uniformity, and replicability. The main objectives of this work were (i) to identify for each species the optimal temperature for high-throughput phenotyping selection, and (ii) to assess the consistency of plant mortality responses across platform and field conditions, and (iii) to perform a GWAS analysis aimed to investigate the genetic architecture of the white lupin frost tolerance and the opportunity for its genomic prediction. A methodological study included 11 genotypes per species, covering a wide range of winter survival according to field data, which underwent 4 levels of freezing treatments: -7, -9, -11, and -13°C. Each of four replicates per genotype included 10 pregerminated seeds transplanted into a peat substrate that underwent 10 days of growth at 22.5°C, 15 days of hardening at 4°C, 3 hours of cooling at -3°C, 4 hours of treatment, 6 days of recovery at 4°C, and 15 days of regrowth at 15/20°C. We assessed mortality ratio at the end of the regrowth, and biomass injuries through a visual score based on the amount of necrosis after both recovery and regrowth and mortality. For each accession, we computed the lethal temperature correspondent to 50% of mortality (LT50) using a binomial GLM with probit link function, as well as ANOVA for genotype and treatment comparisons. A subsequent study investigated the frost tolerance of 144 white lupin genotypes at -11°C and its relationship with nearly 10,000 SNP marker data obtained via ApekI-based genotyping-by-sequencing. Pea displayed higher frost tolerance than white lupin, with average LT50 values of -12.8°C (ranging from -11.6 to -14.5°C) versus -11.0°C (ranging from -10.0 to -12.0°C). The treatments that maximized the genotype mortality variation were -13°C for pea -11°C for lupin, whose mortality ratios highly correlated with biomass injuries (0.98 for pea and 0.97 for lupin), and LT50 (0.91 for pea and 0.94 for lupin). We observed a high consistency of genotype frost tolerance responses across phenotyping platform and field-based evaluation. The subsequent evaluation of 144 white lupin lines at -11°C revealed a wide survival range (from zero to 97%) and provided preliminary indications on the genetic architecture of the trait and the accuracy of its genomic prediction.