Ehd1, a B-type response regulator in rice, confers short-day promotion of flowering and controls FT-like gene expression independently of Hd1

Abstract

Two evolutionarily distant plant species, rice (Oryza sativa L.), a short-day (SD) plant, and Arabidopsis thaliana, a long-day plant, share a conserved genetic network controlling photoperiodic flowering. The orthologous floral regulators-rice Heading date 1 (Hd1) and Arabidopsis CONSTANS (CO)-integrate circadian clock and external light signals into mRNA expression of the FLOWERING LOCUS T (FT) group floral inducer. Here, we report that the rice Early heading date 1 (Ehd1) gene, which confers SD promotion of flowering in the absence of a functional allele of Hd1, encodes a B-type response regulator that might not have an ortholog in the Arabidopsis genome. Ehd1 mRNA was induced by 1-wk SD treatment, and Ehd1 may promote flowering by inducing FT-like gene expression only under SD conditions. Microarray analysis further revealed a few MADS box genes downstream of Ehd1. Our results indicate that a novel two-component signaling cascade is integrated into the conserved pathway in the photoperiodic control of flowering in rice.

Figure 1.

Genetic analysis of Ehd1 and Hd1.(A) Major quantitative trait loci (QTLs) identified in 89 recombinant inbred lines derived from a cross between T65 and Nipponbare. Columns are chromosomes. Average of days to heading of each genotype and F values of single point QTL analyses at the Hd1 and Ehd1 loci are shown to the right of the rice linkage map. (B) Photograph of 80-day-old plants of NIL(Ehd1-gla) (left) and T65 (right) grown under SD (9L:15D) condition. (C) Days to heading of NIL(Ehd1-gla) and T65 under SD (10L:14D) and LD (15L:9D) conditions. (D) Allelic variation in Hd1. The genomic sequences of the functional Nipponbare and defective T65 alleles are shown schematically. The open triangle and asterisk indicate an insertion and a nucleotide substitution, respectively. T65 has a 1901-bp insertion in exon 2 that results in a premature stop codon (TAG) ahead of the CCT motif. P1, P2, and P3 are primers used in E and G. (E) mRNA levels of Hd1 in Nipponbare and T65. RT–PCR was performed with the combination of primers P1 and P2. UBQ (ubiquitin) is shown as a control. (F) Days to heading of T65 and a T₂ transgenic line of T65 carrying the functional Hd1-Nip allele (T65 + Hd1). Similar results were obtained using independent transgenic lines. (G) NIL(Ehd1-gla) carries the defective hd1 allele from T65. T65-allele-specific bands were amplified with primers P1 and P3 from genomic DNA. Controls amplified with P1 and P2 are also shown.

Figure 2.

Map-based cloning of Ehd1.(A) High-resolution linkage map of Ehd1 generated by using a mapping population of >2500 plants. Ehd1 is located on the Clemson BAC clone OSJNBa0071K18 (GenBank accession no. AC027038). (B) Physical map of Ehd1. The CAPS markers (12–14, 13–15, 19–21, 23–25, and 26–28) used for the linkage analysis are indicated by bars. Genes predicted by the GENSCAN program (Burge and Karlin 1997) are shown by arrows. BamHI (11.5 kb) and KpnI (7.6 kb) genomic fragments from Kasalath were introduced into T65. (C) Days to heading of T₀ transformants under SD (10L:14D) condition. (Open bar) Empty vector; (solid bar) vector with the genome fragment. Only plants transformed with the 11.5-kb BamHI fragment showed promotion of flowering. (D) Days to heading of two independent T₂ lines homozygous for the Ehd1-Kas transgene (1–16, 13–10), NIL(Ehd1-gla) and T65 under SD (9L:15D) and LD (14.5L:9.5D) conditions.

Figure 3.

Structure of Ehd1.(A) Gene organization of Ehd1. White boxes indicate exons. (B) Protein structure of Ehd1. (C) Comparison of receiver domains of Ehd1-Kas and representative Arabidopsis response regulators ARR1 (B-type; GenBank accession no. T51246), ARR4 (A-type; BAA34726), TOC1 (Pseudo-RR; AAF86252), and CheY (BAA15698). Residues identical to those in Ehd1 are in black boxes; similar ones are in gray. Asp (D₁ and D₂) and Lys (K) residues crucial for the phosphorelay function are conserved in Ehd1. (D) Alignment of GARP domains of Ehd1-Kas, Golden2 (AAG32325), ARR1 (T51246), and Psr1 (AAD55941). The position of amino acid variation in T65 (G→R) is indicated by an asterisk. Sequences are highlighted as in C. (E) Comparison of the DNA-binding ability of the GARP domains of Ehd1-gla and ehd1-T65. Amounts of fusion proteins in Escherichia coli extracts are shown at the top. Extra bands in the GST–GARP lanes may be partly digested products. The same extracts were used for gel-mobility shift assay (bottom). The arrowhead indicates the band corresponding to the retarded probe. (F) Phylogenetic comparison of B-type RRs between rice and Arabidopsis. ARR clones with MIPS (Munich Information Center for Protein Sequences) codes correspond to Arabidopsis B-type RRs. The others are EHD1 and B-type RRs from rice database information. B-type RRs corresponding to rice full-length cDNAs are indicated by AK accession numbers. Sequences prefixed with Contig are derived from the 93-11 rice genome sequences of the Beijing Genomics Institute.

Figure 4.

Expression analyses of Ehd1, RFT1, and Hd3a. (A,B) Ehd1 mRNA expression in NIL(Ehd1-gla) under SD (A) and LD (B) condition. (C,D) Expression of RFT1 (C) and Hd3a (D) in NIL(Ehd1-gla) (solid symbols) and T65 (open symbols) under SD condition. The ratios of the average expression levels of Ehd1, RFT1, and Hd3a to that of UBQ (ubiquitin) are plotted on the graphs. The average values were obtained from at least three real-time RT–PCR assays. Periods of light and darkness are indicated with white and black bars, respectively. (E) FT-like gene expression in T65 and T65 + Ehd1 (Fig. 2D, lines 1–16).

Figure 5.

Genes downstream of Ehd1 in RT–PCR assays. The same cDNA samples as used in Figure 4E were used as templates. (A) Clones up-regulated in T65 + Ehd1. (B) Clones down-regulated in T65 + Ehd1. (C) UBQ (ubiquitin) as a control. (D) Phylogenetic tree of MADS box genes in rice and Arabidopsis. All rice and Arabidopsis MADS box proteins belonging to AP1/CAL/FUL, SEP, and SOC1 clades in public databases plus AG and OsMADS3 as an outgroup are included in the tree. Accession numbers corresponding to full-length rice cDNAs are also shown. (E) RT–PCR of the remaining two rice MADS box genes, OsMADS15 and OsMADS20.