Whole-Genome Sequencing and RNA-Seq Reveal Differences in Genetic Mechanism for Flowering Response between Weedy Rice and Cultivated Rice

Abstract

Flowering is a key agronomic trait that influences adaptation and productivity. Previous studies have indicated the genetic complexity associated with the flowering response in a photoinsensitive weedy rice accession PSRR-1 despite the presence of a photosensitive allele of a key flowering gene Hd1. In this study, we used whole-genome and RNA sequencing data from both cultivated and weedy rice to add further insights. The de novo assembly of unaligned sequences predicted 225 genes, in which 45 were specific to PSRR-1, including two genes associated with flowering. Comparison of the variants in PSRR-1 with the 3K rice genome (RG) dataset identified unique variants within the heading date QTLs. Analyses of the RNA-Seq result under both short-day (SD) and long-day (LD) conditions revealed that many differentially expressed genes (DEGs) colocalized with the flowering QTLs, and some DEGs such as Hd1, OsMADS56, Hd3a, and RFT1 had unique variants in PSRR-1. Ehd1, Hd1, OsMADS15, and OsMADS56 showed different alternate splicing (AS) events between genotypes and day length conditions. OsMADS56 was expressed in PSRR-1 but not in Cypress under both LD and SD conditions. Based on variations in both sequence and expression, the unique flowering response in PSRR-1 may be due to the high-impact variants of flowering genes, and OsMADS56 is proposed as a key regulator for its day-neutral flowering response.

Keywords: Oryza sativa; RNA-Sequencing; days to heading; genetic interaction; photosensitivity; red rice.

Figure 1

Genome-wide distributions of unique SNPs and InDels, and structural variations overlapping with DTH QTLs along with differentially expressed genes in PSRR-1. (A) DTH QTL segments identified from Gramene database (https://archive.gramene.org/qtl/) (accessed on 5 May 2020) [29]: grey bands, and DTH QTLs of weedy rice from literature [24,30,31,32]: red bands. (B) Gene distribution across chromosome (green bars). (C) Scatter plot of differentially expressed genes (DEGs). Blue dots and red dots represent upregulated and downregulated DEGs (|log2 fold change| > 1.5 and p-adj < 0.05) in all combined pair-wise conditions, respectively. (D) Unique SNP density of PSRR-1 versus 3K Rice Genome (RG) dataset at 10-kb window (red bars). (E) Unique InDel density of PSRR-1 versus 3K RG dataset at 10-kb window. (F) Structural variations in PSRR-1 (green lines).

Figure 2

Unique high-impact SNPs, InDels, and structural variants (SVs) associated with QTLs for agronomically important traits mined from Q-TARO database (http://qtaro.abr.affrc.go.jp/) (accessed on 28 October 2020) [33]. Blue, red, yellow, and green bars correspond to deletion, insertion, SNPs, and SVs, respectively.

Figure 3

Differentially expressed gene (DEG) profiles of PSRR-1 and Cypress under different day length conditions. (A) Venn diagram showing overlapping of DEGs between PSRR-1 and Cypress (CPRS) under long-day (LD) and short-day (SD) conditions with pair-wise comparisons: LD PSRR vs. LD CPRS, LD PSRR vs. SD CPRS, and SD PSRR vs. SD CPRS (|log2 fold change| ≥ 1.5 and padj < 0.05). (B) Bar graphs showing number of DEGs from different pair-wise comparisons along with the regulation pattern.

Figure 4

Flowering genes with unique SNP, InDels, and SVs, and the regulation pattern under different day length conditions. Green-filled boxes indicate variants within the gene. Blue and red-filled boxes indicate the regulation pattern (up- and downregulated, respectively) (|log2 fold change| ≥ 1.5 and padj < 0.05) in pairwise comparisons of day length conditions between long day (LD) and short day (SD) of PSRR-1 (PSRR) and Cypress (CPRS).

Figure 5

Overlapping of flowering-related genes with days to heading (DTH) QTLs. Upregulated and downregulated genes are shown in blue and red fonts, respectively. Vertical green bars represent published DTH QTLs from Gramene (https://archive.gramene.org/qtl/) (accessed on 5 May 2020) [29] and vertical red bars represent QTLs from earlier studies involving weedy rice [24,30,31,32].

Figure 6

Relative expression levels of flowering-related genes in Cypress and PSRR-1 under short-day (SD) and long-day (LD) conditions. (A–F) Real-time quantitative reverse transcription PCR results of flowering genes using Ubq5 as the internal control and Cypress as the reference sample. Bar graph depicts mean relative quantity ± standard deviation. Asterisk (*) indicates significance by two-tailed Student’s t-test assuming unequal variances; * p < 0.05, ** p < 0.01, and *** p < 0.001.

Figure 7

Semiquantitative RT-PCR amplification of OsMADS56 in Cypress and PSRR-1 under short-day (SD) and long-day (LD) conditions.

Figure 8

A simplified model for flowering in cultivated rice (Cypress) and weedy rice (PSRR-1) under long-day (LD: blue arrow lines) and short-day (SD: red arrow lines) conditions. In Cypress, RFT1/Hd3a signals downstream genes (OsMADS14 and OsMADS15) to induce flowering. Under LD and SD, Hd1 and Ehd1 regulate RFT1/Hd3a, while OsMADS56 acts as an upstream regulator of Ehd1, RFT1, and Hd3a. In weedy rice, OsMADS56 is the major flowering activator under LD, while RFT1/Hd3a switches back to control flowering during SD. The role of OsMADS56 during SD is still unclear, indicated by ‘?’.