Three Rice NAC Transcription Factors Heteromerize and Are Associated with Seed Size

Abstract

NACs are plant-specific transcription factors (TFs) involved in multiple aspects of development and stress. In rice, three NAC TF encoding genes, namely ONAC020, ONAC026, and ONAC023 express specifically during seed development, at extremely high levels. They exhibit significantly strong association with seed size/weight with the sequence variations located in the upstream regulatory region. Concomitantly, their expression pattern/levels during seed development vary amongst different accessions with variation in seed size. The alterations in the promoter sequences of the three genes, amongst the five rice accessions, correlate with the expression levels to a certain extent only. In terms of transcriptional properties, the three NAC TFs can activate and/or suppress downstream genes, though to different extents. Only ONAC026 is localized to the nucleus while ONAC020 and ONAC023 are targeted to the ER and cytoplasm, respectively. Interestingly, these two proteins interact with ONAC026 and the dimers localize in the nucleus. Trans-splicing between ONAC020 and ONAC026 results in three additional forms of ONAC020. The transcriptional properties including activation, repression, subcellular localization and heterodimerization of trans-spliced forms of ONAC020 and ONAC026 are different, indicating toward their role as competitors. The analysis presented in this paper helps to conclude that the three NAC genes, which are associated with seed size, have independent as well as overlapping roles during the process and can be exploited as potential targets for crop improvement.

Keywords: NAC; association analysis; rice; seed development; transcriptional properties.

FIGURE 1

Domains of the three NAC TFs and the different transcript fusion forms of ONAC020. All the proteins have a NAC domain (with subdomains A–E) and a transcriptional regulation region (TRR). ONAC020 and ONAC026 have a putative DLN repressor motif in subdomain B; and all these proteins have a putative NAC repression domain (NARD) in subdomain D. The transcript fusion or trans-splicing between ONAC020 and 026 results in three additional proteins (ONAC020.B, 020.C, and 020.D), with insertions/deletions as marked, with respect to ONAC020.A. The color legend indicating the domains is shown at the bottom of the figure.

FIGURE 2

Relative expression levels of three seed-specific NAC genes, namely, ONAC020, ONAC026 and ONAC023, in five rice accessions with varying seed weights. The accessions are (A) Sonasal (SN), (B) Pusa Basmati 1 (PB1), (C) indica Rice 64 (IR64), (D) Nipponbare (NB) and (E) Large Grain Rice (LGR), arranged according to increasing weights. The relative expression for each gene, in each accession has been studied in all the five stages of seed development (S1–S5), as mentioned in the color legend in (A).

FIGURE 3

Alignments of the promoter sequences of (A)ONAC020, (B)ONAC026, and (C)ONAC023 in five accessions of rice, namely SN, PB1, IR64, NB, and LGR. The deletions have been shown by a black line. The SNPs among the accessions have been written in their corresponding positions (Supplementary Tables S3 and S4). Seed-specific elements with changes amongst different accessions (Supplementary Tables S5 and S6) have been marked with colored triangles, as per the color legend. The promoter elements identified on the positive strand have been marked on the upper side and those on the negative strand, on the lower side. In longer stretches of deletion, in ONAC026 and ONAC023, repeatedly occurring elements in 300 and 75 bp stretches, respectively, have been represented only once by their respective triangles.

FIGURE 4

Association of NAC genes with grain weight. (A) represents significance of the sequence variants identified for the NAC genes, for their association with grain weight trait by the Manhattan plot. The relative density of the sequence variants identified from the NAC genes are taken on the X-axis with their -log₁₀(P) used for scanning the significant trait associated sequence variant on the Y-axis. (B) Validation of the strong association of the InDel in the URR of ONAC026 gene with the grain weight trait, where DRR and URR represent the downstream and upstream regulatory region respectively, ATG and TAG correspond to the translation start and stop codons respectively and TRR stands for the transcriptional regulatory region. The InDel clearly differentiates the six haplotypes into low grain weight and medium/high grain weight groups.

FIGURE 5

Transcriptional ability of the NAC TFs by (A) transactivation and (B) transrepression assays of the three NAC proteins, and the proteins generated by trans-splicing of ONAC020/ONAC026 (i.e., ONAC020.B–D). The graphs show the relative β-galactosidase activity of the NAC TFs by the ONPG assay on the Y-axis. The vector pGBKT7 is the negative control for transactivation assays in (A) and the reconstituted GAL4 TF (rGAL4) is the positive control for transrepression assays in (B). The standard error bars have been shown. The values which are significant with a P-value cut off ≤0.05 and ≤0.005 are indicated by a single and double asterisks, respectively.

FIGURE 6

Subcellular localization of NAC-YFP fusion proteins. (A) represents the localization images of the three main NAC TFs and the transcript fusion forms of ONAC020. The left column indicates the localization image of the YFP fusion protein at 514 nm. The middle column is the merger of bright field, fluorescence image and DAPI stained nucleus while the right column shows an enlarged image. (B) represents the co-localization of ONAC020 with endoplasmic reticulum (ER) marker. The protein names have been mentioned and the scale bar indicates 50 μm.

FIGURE 7

Interactions amongst the NAC proteins by BiFC. The left column indicates the localization image of the reconstituted YFP protein at 514 nm. The middle column is the merger of bright field and fluorescence images while the right column shows an enlarged image of the nucleus. The nucleus has been stained by DAPI dye. The interacting protein names have been mentioned and the scale bar indicates 50 μm.