Functional and evolutionary characterization of the CONSTANS gene family in short-day photoperiodic flowering in soybean

Abstract

CONSTANS (CO) plays a central role in photoperiodic flowering control of plants. However, much remains unknown about the function of the CO gene family in soybean and the molecular mechanisms underlying short-day photoperiodic flowering of soybean. We identified 26 CO homologs (GmCOLs) in the soybean genome, many of them previously unreported. Phylogenic analysis classified GmCOLs into three clades conserved among flowering plants. Two homeologous pairs in Clade I, GmCOL1a/GmCOL1b and GmCOL2a/GmCOL2b, showed the highest sequence similarity to Arabidopsis CO. The mRNA abundance of GmCOL1a and GmCOL1b exhibited a strong diurnal rhythm under flowering-inductive short days and peaked at dawn, which coincided with the rise of GmFT5a expression. In contrast, the mRNA abundance of GmCOL2a and GmCOL2b was extremely low. Our transgenic study demonstrated that GmCOL1a, GmCOL1b, GmCOL2a and GmCOL2b fully complemented the late flowering effect of the co-1 mutant in Arabidopsis. Together, these results indicate that GmCOL1a and GmCOL1b are potential inducers of flowering in soybean. Our data also indicate rapid regulatory divergence between GmCOL1a/GmCOL1b and GmCOL2a/GmCOL2b but conservation of their protein function. Dynamic evolution of GmCOL regulatory mechanisms may underlie the evolution of photoperiodic signaling in soybean.

Figure 1. CO homologs in soybean.

(A) Phylogenetic analysis of full-length amino acid sequences of CO homologs of soybean (Gm), Arabidopsis (At), rice (Os) and Chlamydomonas reinhardtii (Cr). Numbers at nodes indicate the value of 1,000 bootstrap analyses. The domain structure of B-box 1 (red rectangles), B-box 2 (blue rectangles), CCT (yellow rectangles) and the introns (black triangle) of the genes are shown. Followed are the accession numbers of the proteins and their corresponding databases (in square brackets). (B) BLAST best hits of CO homologs in Arabidopsis (A. thaliana) and soybean (G. max). The BLAST best hits are shown by dotted lines with an arrow head, and the protein pairs of the reciprocal best hits are shown by lines with double-head arrows.

Figure 2. mRNA abundance of GmCOLs measured by RNA sequencing. RPKM values are displayed on the left.

The top panel shows GmCOLs in Clade I, the middle panel shows GmCOLs in Clade II, and the bottom panel shows GmCOLs in Clade III. SD is 10 hours light (6∶45–16∶45), LD is 16 hours light (6∶45–22∶45), and LD-SD is a shift from three weeks LD to 5 days SD. Samples are three representative time points: T1 (6∶30), T3 (14∶30) and T5 (22∶30).

Figure 3. Hierarchical clustering of GmCOLs based on their mRNA abundance under LD and SD.

GmCOL2a, GmCOL2b, GmCOL10a and GmCOL10b are excluded due to their extreme low abundance. To demonstrate the variation in the mRNA abundance, a rainbow color scheme was used in which red indicates high expression and blue indicates low expression.

Figure 4. mRNA abundance of GmCOL1a, GmCOL1b, GmCOL2a and GmCOL2b.

Circles and solid lines represent the data obtained by qRT-PCR, and gray bars represent the data obtained by RNA sequencing. Relative expression levels of qRT-PCR among the samples in each panel are shown on the left, and RPKM values of RNA sequencing are shown on the right. SD is 10 hours light (6∶45–16∶45) and LD is 16 hours light (6∶45–22∶45). Time points T1–T6 represent 6∶30, 10∶30, 14∶30, 18∶30, 22∶30 and 2∶30. Note that RNA sequencing samples contain three representative time points: T1, T3 and T5.

Figure 5. mRNA abundance of GmFT2a and GmFT5a.

Circles and solid lines represent the data obtained by qRT-PCR, and gray bars represent the data obtained by RNA sequencing. Relative expression levels of qRT-PCR among the samples in each panel are shown on the left, and RPKM values of RNA sequencing are shown on the right. SD is 10 hours light (6∶45–16∶45), LD is 16 hours light (6∶45–22∶45) and LD-SD is a shift from three weeks LD to 5 days SD. Time points T1–T6 represent 6∶30, 10∶30, 14∶30, 18∶30, 22∶30 and 2∶30. Note that RNA sequencing samples contain three representative time points: T1, T3 and T5.

Figure 6. Comparison of the 2.5GmCOL1a, GmCOL1b, GmCOL2a and GmCOL2b.

Empty arrows indicate the coding sequence (introns were excluded); lines indicate the upstream intergenic region. Conserved sequences between two sequences are evident from the detected diagonal dotted lines. The dot plot was created using the Nucleic Acid Dot Plots program, with the following parameters: window size: 11; mismatch limit: 0.

Figure 7. Overexpression of GmCOL1a, GmCOL1b, GmCOL2a and GmCOL2b rescued the late flowering phenotype of Arabidopsis co-1 mutant.

(A) The vertical axis indicates the number of T1 transgenic plants, and the horizontal axis indicates the number of rosette leaves. (B) Col-0 (left), co-1 mutant (middle) and co-1 mutant overexpressing GmCOL1a (right).