Natural variation of maize root hydraulic architecture underlies highly diverse water uptake capacities

Abstract

Plant water uptake is determined by the root system architecture and its hydraulic capacity, which together define the root hydraulic architecture. The current research aims at understanding the water uptake capacities of maize (Zea mays), a model organism and major crop. We explored the genetic variations within a collection of 224 maize inbred Dent lines and successively defined core genotype subsets to access multiple architectural, anatomical, and hydraulic parameters in the primary root (PR) and seminal roots (SR) of hydroponically grown seedlings. We found 9-, 3.5-, and 12.4-fold genotypic differences for root hydraulics (Lpr), PR size, and lateral root size, respectively, that shaped wide and independent variations of root structure and function. Within genotypes, PR and SR showed similarities in hydraulics and, to a lesser extent, in anatomy. They had comparable aquaporin activity profiles that, however, could not be explained by aquaporin expression levels. Genotypic variations in the size and number of late meta xylem vessels were positively correlated with Lpr. Inverse modeling further revealed dramatic genotypic differences in the xylem conductance profile. Thus, tremendous natural variation of maize root hydraulic architecture underlies a high diversity of water uptake strategies and paves the way to quantitative genetic dissection of its elementary traits.

Figure 1.

Frequency distribution and PCA of main RSA and hydraulic traits in a diversity panel of 224 genotypes. A) Frequency distribution of PR surface area (SA_PR). B) Frequency distribution of LR surface area (SA-LR_PR). C) Frequency distribution of PR hydraulic conductivity (Lp_r). D) Frequency distribution of SRN. E) PCA using the 4 parameters above. The figure shows the first PCA plane with position of the parameters. F) Cluster analysis of PCA to select contrasting genotypes. The 13 selected genotypes are highlighted on the graph, with indicated color code for the 4 clusters.

Figure 2.

RSA of a core subset of 13 genotypes. A) SRN among the indicated genotypes and clusters. B) Same analysis of PR and SR surface area. C) PCA analysis using RSA and hydraulics parameters of PR and SR. The figure shows the first PCA plane with position of the 7 parameters analyzed. D) Pearson correlation study between the SA_PR and SRN. Each dot represents the average values of one of 13 selected genotypes. In A and B), each of 4 clusters (see Fig. 1F) is represented by 3 adjacent genotypes, as indicated. Each bar represents the mean value (±Se) of 10 replicates.

Figure 3.

Characterization of hydraulics in PR and SR. A)Lp_r of the PR and first SR in the core subset of 13 selected genotypes. Each of 4 clusters is represented by 3 adjacent genotypes, as indicated. B) Detailed analysis of the Lp_r of PR and all SR of 5 further selected genotypes. Note that SR number varies between genotypes (Supplemental Fig. S4). C) Root hydraulic conductance (L₀) reconstructed for 5 selected genotypes. In A) and B), asterisks (*) indicate significant differences in Lp_r between the PR and the indicated SR of the same genotype (t-test, P-value <0.05). In A to C) each bar represents the mean value (±Se) of 10 replicates, 12 replicates, and 12 replicates, for all samples except Lo1124 (8 replicates), respectively.

Figure 4.

Inverse modeling analysis of PR hydraulics in 4 representative genotypes A) the figure shows variations of axial conductance (K) as a function of distance to root tip. The solid lines represent lowess fits done on K profiles of PR of B89 (red; n = 9), DK78010 (green; n = 7), EZ46 (blue; n = 10), and Lo1124 (purple, n = 7). The dashed lines delineate the corresponding 95% confidence intervals. The side panel is a detail of the main figure up to a distance to tip of 0.35 m. B) Averaged radial conductivity (k ± Se) of the indicated genotypes (same numbers of replicates as in (A). C) Heat map, in representative RSAs of indicated genotypes, of local radial water flow (µL s⁻¹ m⁻²) for a pressure of 0.1 MPa. Scale bar: 10 mm.

Figure 5.

Contribution of aquaporins to Lp_r. A) Percentage (±Se) of Lp_r inhibition by NaN₃ in the PR (green) and first SR (orange) of indicated genotypes (n = 8 to 10). Asterisks (*) indicate significant differences in Lp_r inhibition between the PR and SR of a same genotype (t-test, P-value <0.05). B) Natural variation of ZmPIP2;5 relative expression. C) Natural variation of ZmPIP1;1 relative expression. D) Natural variation of ZmPIP1;5 relative expression. In B to D), the expression level of each of the ZmPIPs in B73 roots was set to 1 and subsequently used to calculate the relative expression level of the same gene in the 12 other genotypes. Asterisks (*) indicate a significant difference in relative expression levels between the PR (green) and SR (orange) of a same genotype (t-test for normally distributed data; Kolmogorov–Smirnov test for nonparametric data, P-value <0.05). Error bars represent Se.

Figure 6.

Natural variation of root anatomical parameters. A) Representative cross-sections of PR of the 13 genotypes of the core subset. Scale bars = 200 µm. B) Mean cross-section areas (±Se; n = 10 to 12) of the PR and SR of the same 13 genotypes. C) Mean LMZ size (±Se; n = 10 to 12) of the same 13 genotypes. Asterisks (*) indicate a significant difference between PR and SR of a same genotype (Kolmogorov–Smirnov test, P-value <0.05).

Figure 7.

Relation between cross-section anatomical parameters and Lp_r. A) Correlation studies showing a positive correlation between Lp_r and the number of LMX. B) Positive correlation between LMX size and Lp_r. LMX size was calculated as the average radial height and width of individual LMX vessels. In A) and B), each dot represents the average values of one of 13 genotypes of the core subset.