Novel QTL for Lateral Root Density and Length Improve Phosphorus Uptake in Rice (Oryza sativa L.)

Abstract

The rice root system consists of two types of lateral roots, indeterminate larger L-types capable of further branching, and determinate, short, unbranched S-types. L-type laterals correspond to the typical lateral roots of cereals whereas S-type laterals are unique to rice. Both types contribute to nutrient and water uptake and genotypic variation for density and length of these laterals could be exploited in rice improvement to enhance adaptations to nutrient and water-limited environments. Our objectives were to determine how best to screen for lateral root density and length and to identify markers linked to genotypic variation for these traits. Using different growing media showed that screening in nutrient solution exposed genotypic variation for S-type and L-type density, but only the lateral roots of soil-grown plants varied for their lengths. A QTL mapping population developed from parents contrasting for lateral root traits was grown in a low-P field, roots were sampled, scanned and density and length of lateral roots measured. One QTL each was detected for L-type density (LDC), S-type density on crown root (SDC), S-type density on L-type (SDL), S-type length on L-type (SLL), and crown root number (RNO). The QTL for LDC on chromosome 5 had a major effect, accounting for 46% of the phenotypic variation. This strong positive effect was confirmed in additional field experiments, showing that lines with the donor parent allele at qLDC5 had 50% higher LDC. Investigating the contribution of lateral root traits to P uptake using stepwise regressions indicated LDC and RNO were most influential, followed by SDL. Simulating effects of root trait differences conferred by the main QTL in a P uptake model confirmed that qLDC5 was most effective in improving P uptake followed by qRNO9 for RNO and qSDL9 for S-type lateral density on L-type laterals. Pyramiding qLDC5 with qRNO9 and qSDL9 would be possible given that trade-offs between traits were not detected. Phenotypic selection for the RNO trait during variety development would be feasible, however, the costs of doing so reliably for lateral root density traits is prohibitive and markers identified here therefore provide the first opportunity to incorporate such traits into a breeding program.

Keywords: Crown root; L-type lateral roots; P uptake simulation; Root system architecture; S-type lateral roots.

Fig. 1

Scan of a root system excavated 41 DAS from a low-P field experiment (a) and classification of L-type and S-type lateral roots in relation to their parent root (b). The scan was obtained using an Epson Perfection V700 photo dual-lens scanner with top lighting, with the following settings: 600 dpi, 16-bit grayscale, positive film

Fig. 2

Frequency distribution of 98 BC1F5 lines for seven lateral root traits, including L-type density on the crown root (LDC), S-type density on the crown root (SDC), S-type density on L-type (SDL), Single S-type length on the crown root (SLC), single S-type length on L-type (SLL), and crown root number (RNO) from a low P field experiment in 2021 (Experiment 2)

Fig. 3

Association of shoot P content of the 98 lines of the QTL mapping population (experiment 2) with their Root Index calculated by multiplying the three most influential traits (RNO*LDC*SDL). Note: for better readability index values are given as 1/10 of their real value

Fig. 4

Confirmation of allelic effects at qLDC5 in the BC1F6 generation by measuring L-type lateral root density on root samples excavated from the low-P field in Tsukuba in 2022. Different letters indicate significant differences and “n” indicates the number of lines classified into each allelic group