Pollution by heavy metals limits the area of land available for cultivation of food crops. A potential solution to this problem might lie in the molecular breeding of food crops for phytoremediation that accumulate toxic metals in straw while producing safe and nutritious grains. Here, we identify a rice quantitative trait locus we name cadmium (Cd) accumulation in leaf 1 (CAL1), which encodes a defensin-like protein. CAL1 is expressed preferentially in root exodermis and xylem parenchyma cells. We provide evidence that CAL1 acts by chelating Cd in the cytosol and facilitating Cd secretion to extracellular spaces, hence lowering cytosolic Cd concentration while driving long-distance Cd transport via xylem vessels. CAL1 does not appear to affect Cd accumulation in rice grains or the accumulation of other essential metals, thus providing an efficient molecular tool to breed dual-function rice varieties that produce safe grains while remediating paddy soils.

The Chinese genebank contains 23,587 soybean landraces collected from 29 provinces. In this study, a representative collection of 1,863 landraces were assessed for genetic diversity and genetic diVerentiation in order to provide useful information for eVective management and utilization. A total of 1,160 SSR alleles at 59 SSR loci were detected including 97 unique and 485 low-frequency alleles, which indicated great richness and uniqueness of genetic variation in this core collection. Seven clusters were inferred by STRUCTURE analysis, which is in good agreement with a neighbor-joining tree. The cluster subdivision was also supported by highly signiWcant pairwise F st values and was generally in accordance with diVerences in planting area and sowing season. The cluster HSuM, which contains accessions collected from the region between 32.0 and 40.5°N, 105.4 and 122.2°E along the central and downstream parts of the Yellow River, was the most genetically diverse of the seven clusters. This provides the Wrst molecular evidence for the hypotheses that the origin of cultivated soybean is the Yellow River region. A high proportion (95.1%) of pairs of alleles from diVerent loci was in LD in the complete dataset. This was mostly due to overall population structure, since the number of locus pairs in LD was reduced sharply within each of the clusters compared to the complete dataset. This shows that population structure needs to be accounted for in association studies conducted within this collection. The low value of LD within the clusters can be seen as evidence that much of the recombination events in the past have been maintained in soybean, Wxed in homozygous self-fertilizing landraces.