Integrated root phenotypes for improved rice performance under low nitrogen availability

Abstract

Greater nitrogen efficiency would substantially reduce the economic, energy and environmental costs of rice production. We hypothesized that synergistic balancing of the costs and benefits for soil exploration among root architectural phenes is beneficial under suboptimal nitrogen availability. An enhanced implementation of the functional-structural model OpenSimRoot for rice integrated with the ORYZA_v3 crop model was used to evaluate the utility of combinations of root architectural phenes, namely nodal root angle, the proportion of smaller diameter nodal roots, nodal root number; and L-type and S-type lateral branching densities, for plant growth under low nitrogen. Multiple integrated root phenotypes were identified with greater shoot biomass under low nitrogen than the reference cultivar IR64. The superiority of these phenotypes was due to synergism among root phenes rather than the expected additive effects of phene states. Representative optimal phenotypes were predicted to have up to 80% greater grain yield with low N supply in the rainfed dry direct-seeded agroecosystem over future weather conditions, compared to IR64. These phenotypes merit consideration as root ideotypes for breeding rice cultivars with improved yield under rainfed dry direct-seeded conditions with limited nitrogen availability. The importance of phene synergism for the performance of integrated phenotypes has implications for crop breeding.

Keywords: IR64; ORYZA_V3; OpenSimRoot; functional-structural plant modelling; nitrogen acquisition; nodal roots; phene synergism; root system architecture.

Figure 1

Visualization of different morphological root phenes simulated in OpenSimRoot/Rice. Panel (a, c) depicts rice root system with shallow (30°, panel a) and steep (60° panel c) nodal root angle. Panel b depicts different root classes—nodal (red), L‐type laterals (blue), S‐type laterals on nodal (green) and on L‐type laterals (pink). Panel d depicts different axial roots in a root system without any laterals—primary (green), large (red) and small (blue) diameter nodal roots and nodal roots from tillers (pink). Panel e highlights soil nitrogen dynamics with IR64 root system at 30 DAG, wherein red to dark blue colour depicts highest to lowest nitrogen availability in the soil, respectively. DAG, days after germination [Color figure can be viewed at wileyonlinelibrary.com]

Figure 2

Shoot biomass of the reference phenotype (IR64) over 30 days with different initial soil nitrogen supply. (a) Simulated versus observed shoot growth under optimal and low N conditions. Obs1, 2 & 3 are the measured shoot dry weights for rice cv. IR64 over optimal and low N supply (Luquet et al., 2006); (b) simulated shoot growth at 30 DAG with varying N supply under rainfed dry directed seed condition over dry growing season. DAG, days after germination [Color figure can be viewed at wileyonlinelibrary.com]

Figure 3

Shoot biomass over (a) varied nodal root angle (b) nodal root diameter (c) nodal root number (d) L‐type lateral root branching density (e) S‐type lateral root branching density under four levels of soil nitrogen supply [Color figure can be viewed at wileyonlinelibrary.com]

Figure 4

Nitrogen uptake efficiency (μmol/[cm²CCday]) of the phenotype at 30 DAG with varying individual root phenes over different initial soil N supply (same as Figure 3). The N uptake efficiency is calculated by dividing the cumulative N uptake (μmol) over 30 days with the root surface area duration (${\mathrm{RSA}}_{\mathrm{dur}}$, cm²·day)—an integral of the root surface area over time. The latter was as the sum of the areas of a series of rectangles. The area of each rectangle is the product of the interval between two consecutive time points and the average of the root surface area at each of the two time points. ${\mathrm{RSA}}_{\mathrm{dur}}\approx {\sum }_{t=1}^{t=30}\frac{({\mathit{SA}}_t-{\mathit{SA}}_{t-1})}{2}\times A_t-A_{t-1}$ Where, t are the time points, $A_t$ is age of plant at time t, and ${\mathit{SA}}_t$ is root surface area at time t. This is equivalent to the trapezoid rule in numerical integration. Root competition and poor coincidences of roots and soil resource in space and time may decrease resource uptake per unit area. DAG, days after germination [Color figure can be viewed at wileyonlinelibrary.com]

Figure 5

Nitrogen content and root length density over soil depth at 30 DAG for root system with varying nodal root angle over (a, c) optimal (58 kg/ha) and (b, d) low (5.8 kg/ha) initial soil N supply. DAG, days after germination [Color figure can be viewed at wileyonlinelibrary.com]

Figure 6

Reference phenotype IR64 (a), synergistic (b), additive (c), and antagonistic (d) effects of 1024 simulated phene combinations on shoot dry weight under low N conditions. Actual responses (i.e., shoot biomass, blue) are greater than, equal to and less than the expected (red) for synergistic, additive and antagonistic phenotypes, respectively [Color figure can be viewed at wileyonlinelibrary.com]

Figure 7

k‐Proto clustering of 70 root phenotypes with threshold reduction in shoot dry weight (i.e., ≤45% [a]) in response to low soil N availability. Clustering led to eight distinct clusters, sequentially labelled as A–H, with varying proportion of different phene states (b‐f) [Color figure can be viewed at wileyonlinelibrary.com]

Figure 8

Visualization of representative root phenotypes corresponding to each of the eight identified Clusters (a–h) with the reference phenotype IR64 (i) at 30 DAG. Labels correspond with the cluster names [Color figure can be viewed at wileyonlinelibrary.com]

Figure 9

Synergistic benefits of phenotypes in each cluster. Actual and expected response (i.e., shoot biomass) of phenotypes in each cluster is highlighted along with their corresponding synergistic benefits [Color figure can be viewed at wileyonlinelibrary.com]

Figure 10

Predicted rice grain yield (with 14% moisture content) of representative root phenotypes corresponding to each of the eight identified Clusters (A–H) and the reference IR64 (I) over rainfed dry direct‐seeded conditions with (a) optimal (58–58–58 kg/ha) and (b) low (5.8–0–0 kg/ha) N fertilization with planting density of 20/m². The phenotypes were simulated over 33 years of historic and 80 years of future dry season weather data from the IRRI research station, Los Baños, Philippines [Color figure can be viewed at wileyonlinelibrary.com]