Action of multiple intra-QTL genes concerted around a co-localized transcription factor underpins a large effect QTL

Abstract

Sub-QTLs and multiple intra-QTL genes are hypothesized to underpin large-effect QTLs. Known QTLs over gene families, biosynthetic pathways or certain traits represent functional gene-clusters of genes of the same gene ontology (GO). Gene-clusters containing genes of different GO have not been elaborated, except in silico as coexpressed genes within QTLs. Here we demonstrate the requirement of multiple intra-QTL genes for the full impact of QTL qDTY12.1 on rice yield under drought. Multiple evidences are presented for the need of the transcription factor 'no apical meristem' (OsNAM12.1) and its co-localized target genes of separate GO categories for qDTY12.1 function, raising a regulon-like model of genetic architecture. The molecular underpinnings of qDTY12.1 support its effectiveness in further improving a drought tolerant genotype and for its validity in multiple genotypes/ecosystems/environments. Resolving the combinatorial value of OsNAM12.1 with individual intra-QTL genes notwithstanding, identification and analyses of qDTY12.1has fast-tracked rice improvement towards food security.

Figure 1. Effect of qDTY_12.1 on yield under drought.

(A) Additive effect of qDTY_12.1 QTL+ and QTL- lines for grain yield over Vandana under varying severity of drought stress in six experiments conducted over three seasons at IRRI. (B) Grain yield data for the parental lines Vandana and Way Rarem and two NILs, 481-B and 283-B, from 11 trials over three years. NILs consistently outperformed the recipient parent Vandana for yield under drought.

Figure 2. Panicle and root branching.

(A) Under reproductive stage drought, 481-B exhibited increased numbers of secondary branches, total spikelets and filled spikelets in the panicles as a route to increased yield compared to Vandana. Similarly root branching was also increased in 481-B under PEG 8000 simulated water deficit as seen here as well as in the field (Figure S7). (B) IR64-Tr-E2 plants overexpressing the OsNAM_12.1 recapitulated similar traits of panicles and roots under drought but with reduced effect on yield compared to 481-B.

Figure 3. Identification of gene-based high density markers and candidate genes.

(A) Schematic representation of the qDTY_12.1 region for advancement in marker saturation. 1. Five markers (blue peaks) used for the original QTL detection. 2. A further 3 markers (orange peaks) for fine mapping detected two sub-QTLs. 3. Nine more gene-based markers (red) used in MCMC analysis detected 4 sub-QTLs. The original QTL was reduced from 3.6 Mb to 1.8  Mb to 1.5 Mb successively. (B) Genes used as gene-based markers (dotted) by a strategy outlined in Figures S8, S9 and Table S2. Grey shade indicates putative OsNAM_12.1 targets. Final mapping shifted the QTL peaks directly to four putative candidate genes (beige) based on results in C. (C) Heat map style representation of the result that higher the number of Way Rarem alleles, more improved the yield. Raw data represented for yield (Y-axis, right; kg ha⁻¹) in recombinant lines (Y-axis, left) tested for 14 markers (X-axis, top) under well watered (WW) and drought stress (DS) conditions. Alleles are indicated by colors: yellow = Vandana; blue = Way Rarem; light blue = heterozygous, and white = unknown.

Figure 4. Root analysis of the transgenic plants.

(A) 3D scatterplot for total root traits in Vandana, Way Rarem, 481-B, and transgenic Vandana overexpressing OsNAM_12.1. Data was collected with ImageJ root analyzer ((http://imagej.nih.gov) on 2 week old plants with 8 to 15 plants per genotype under well watered (WW) and drought stress (DS) conditions. Statistical analysis and graph plotting were performed using R software (http://CRAN.R- project.org/package=scatterplot3d). Multiple stalks of a single color represent mean values for multiple samples within a ‘genotype’, which represent different standard deviations within the group for the three root traits. (B) Panel showing that plants mutated for the particular genes led to higher lateral root growth than the WT. (C) Quantitative difference in total root length of each mutant as percent increase over WT. Up to five seedlings from pots were sampled for each mutant and WT.