Interactions between SQUAMOSA and SHORT VEGETATIVE PHASE MADS-box proteins regulate meristem transitions during wheat spike development

Abstract

Inflorescence architecture is an important determinant of crop productivity. The number of spikelets produced by the wheat inflorescence meristem (IM) before its transition to a terminal spikelet (TS) influences the maximum number of grains per spike. Wheat MADS-box genes VERNALIZATION 1 (VRN1) and FRUITFULL 2 (FUL2) (in the SQUAMOSA-clade) are essential to promote the transition from IM to TS and for spikelet development. Here we show that SQUAMOSA genes contribute to spikelet identity by repressing MADS-box genes VEGETATIVE TO REPRODUCTIVE TRANSITION 2 (VRT2), SHORT VEGETATIVE PHASE 1 (SVP1), and SVP3 in the SVP clade. Constitutive expression of VRT2 resulted in leafy glumes and lemmas, reversion of spikelets to spikes, and downregulation of MADS-box genes involved in floret development, whereas the vrt2 mutant reduced vegetative characteristics in spikelets of squamosa mutants. Interestingly, the vrt2 svp1 mutant showed similar phenotypes to squamosa mutants regarding heading time, plant height, and spikelets per spike, but it exhibited unusual axillary inflorescences in the elongating stem. We propose that SQUAMOSA-SVP interactions are important to promote heading, formation of the TS, and stem elongation during the early reproductive phase, and that downregulation of SVP genes is then necessary for normal spikelet and floral development. Manipulating SVP and SQUAMOSA genes can contribute to engineering spike architectures with improved productivity.

Figure 1

Figure 2

Selected VRT2 and SVP1 mutations and crosses with vrn1 and ful2 mutants, and transgenic UBI_pro:VRT2 plants. A–B, Locations of the selected mutations in VRT2 (A) and SVP1 (B) are shown in the gene structure diagram. Thick horizontal lines represent introns. Exons are represented by rectangles, with those in orange encoding the MADS domain and those in green encoding the conserved K domain. Ca = hexaploid wheat Cadenza and K = tetraploid wheat Kronos. GT is the mutated splice site. For both genes, the A genome mutants are indicated above the gene structure diagram and the B genome mutants below. C, Reference map of the crosses used to generate vrt2 and svp1 loss-of-function mutants and the higher-order mutants used in this study: 1. Interaction between a transgene with constitutive VRT2 expression (UBI_pro:VRT2) with the ful2 mutant. 2. Complementation of vrt2 with UBI_pro:VRT2. 3. Generation of a vrt2 svp1 double mutant. 4. Interactions between vrt2 and vrn1 ful2 mutants generated by Li et al. (2019). Vrn1 = heterozygous for Vrn-A1. The symbol “/*N” indicates the number of crosses to Kronos recurrent parent performed to reduce background mutations.

Figure 3

Effects of individual and combined vrt2 and svp1 mutants on important agronomic traits. A–F, Visual and quantitative phenotypes of WT, vrt2, svp1, and vrt2 svp1 mutant plants. A, Plants 80 days after planting. Bar = 10 cm. B, Spikes (note the axillary spikelet in the first node of vrt2 svp1). Bar = 1 cm. C, Days to heading. D, Leaf number. E, SNS. F, Plant height (cm). C–F, The number of plants analyzed is indicated below the genotypes. *P < 0.05, ***P < 0.001 for differences with WT using Dunnett’s test (Supplemental Data Set S2). Box-plot features are explained in the “Statistical analyses” section of “Materials and methods.”

Figure 4

Axillary inflorescences in vrt2, svp1, and vrt2 svp1. A, Comparison of internodes in WT and vrt2 svp1 (−1 is the node below the peduncle, and −3 is the most basal node). B–D, Detail of the three nodes in WT. E–J, Nodes in vrt2 svp1. E, Spikelet in node −1. F, Dissection of the spikelet showing glumes and three florets. G, Dissection of floret one (red star in (F)) showing normal floral organs. H, Spikelet in node −2. I, Axillary spike in node −3 surrounded by a bract. J, Same axillary spike without the bract showing lateral spikelets. K, Axillary spike in vrt2 surrounded by a bract. L, Axillary spike in svp1 surrounded by a bract. M, Proportion of plants with axillary spikes or spikelets in each of the three nodes below the spike (n = 7). Green = axillary meristem (AM) absent or not developed, Orange= axillary meristem developed into a spike or spikelet. N, The yellow arrow points to an axillary spike emerging from its subtending leaf in vrt2 svp1. Bars in (A) = 5 cm, (B–L) = 2 mm.

Figure 5

In situ hybridization analysis of VRT2 and SVP1 in developing Kronos inflorescences. A–H, WT Kronos, early spike development stages before TS formation. A, D, E, H = early DR; B, F = DR and C, G = TS. I–L, Kronos vrn1 ful2 mutant inflorescence with lateral VMs. The developmental stage is equivalent to TS in WT, but without TS. J and L are amplified regions from I and K, respectively. A–D and I–J, VRT2. E–H and K–L, SVP1. D and H, Control sense probes for VRT2 and SVP1, respectively. A–H, Triticum monococcum probes. I–L, Kronos probes. Primers for the probes are described in Supplemental Table S4. Bars = 500 μM.

Figure 6

Phenotypic characterization of Kronos lines constitutively expressing VRT2. Three independent UBI_pro:VRT2 events with weak (T#8), intermediate (T#4), and strong phenotypes (T#2). A, Days to heading. B, Stem length (without spike). C, SNS. D, Spikelet density (spikelet number/spikelet length in cm). E, Glume 1 length. F, Lemma 1 length. A–F, The number of plants analyzed is indicated below the genotypes. *P < 0.05, **P < 0.01, *** P < 0.001 in Dunnett’s test versus WT control (Supplemental Data Set S2). Box-plot features are explained in the “Statistical analyses” section of “Materials and methods.”

Figure 7

Spikes and spikelets changes in Kronos lines constitutively expressing VRT2. Three independent events with weak (T#8), intermediate (T#4), and strong phenotypes (T#2). A, Spike phenotype. B, Aligned glumes showing difference in length. C, Basal “spikelet” from UBI_pro:VRT2 event T#2. D, Dissection of the basal “spikelet” shows a determinate branch with multiple spikelets. E, Detail of spikelets 1 and 2 in (D), each with glumes, florets and an elongated rachilla. Bars = 1 cm. g = glume, fl= floret and sp= spikelet.

Figure 8

Relative expression of wheat flowering genes in developing spikes at the TS stage of UBI_pro:VRT2 transgenic lines T#8, T#4, and T#2 and sister lines without the transgene (WT). A, VRT2 (transgenic plus endogenous transcripts). B, A-class MADS-box genes VRN1 and FUL2. C, B-class MADS-box genes PI1 (∼OsMADS4) and AP3-1 (∼OsMADS16). D, C-class MADS-box genes AG1 (∼OsMADS58) and AG2 (∼OsMADS3). E, E-class MADS-box genes SEP1-2 (∼OsMADS1), SEP1-4 (∼OsMADS5), SEP1-6 (∼OsMADS34), SEP3-1 (∼OsMADS7) and SEP3-2 (∼OsMADS8). A–E, Graphs are based on four biological replicates (each replicate is a pool of 6–8 developing spikes at the TS stage). *P < 0.05, **P < 0.01, ***P < 0.001 in Dunnett’s test versus the WT control (Supplemental Data Set S2). Expression was determined by qRT-PCR using ACTIN as endogenous controls and normalization relative to the WT (WT = 1). Box plot features are explained in the “Statistical analyses” section of “Materials and methods.”

Figure 9

Effect of combined ful2 mutation and UBI_pro:VRT2 T#8 transgenic line on stem elongation and spike/spikelet development. A, Stem length (without spikes). B, SNS. C, Glume length in centimeter. D, Lemma length in centimeter. A–D, WT = homozygous Ful2 and not transgenic. T#8: weak constitutive transgenic UBI_pro:VRT2 line T#8. ful2 = loss-of-function mutant for ful-A2 and ful-B2 and not transgenic. T#8ful2 = homozygous ful2 and T#8 transgenic present. N = number of plants analyzed is indicated below the genotypes. *P < 0.05, **P < 0.01, ***P < 0.001 in Dunnett’s test (Supplemental Data Set S2). Box-plot features are explained in the “Statistical analyses” section of “Materials and methods.” E, Young spikes of WT Kronos, ful2, UBI_pro:VRT2 T#8, and combined UBI_pro:VRT2 T#8 – ful2. F, Combined UBI_pro:VRT2 ful2 basal spikelet transformed into a branch. G, Dissection of the basal spikelet converted into a branch showing lateral spikelets. H, Dissection of the spikelet marked with 1 in G. Bars = 1 cm.

Figure 10

Effect of combined Vrn1 ful2 and vrt2 mutations on spike and spikelet development. A, B, E, G–H, Vrn1 ful2. C, D, F, I vrt2 Vrn1 ful2. A and C, Young spikes. B and D, Dissection of basal spikelets. E and F, Older spikes. G–I Dissection of older spikelets. H, Detail of the third “floret” in G (red asterisk) that reverted to a spikelet with its own glumes. The inset in (H) shows a floret of this spikelet. I, Spikelet of the same age as in G from vrt2 Vrn1 ful2. Bars = 1 cm.

Figure 11

Y2H interactions between wheat SQUAMOSA, SVP, and SEP proteins. Wheat MADS-box proteins of the SQUAMOSA-clade are indicated by yellow boxes (VRN1, FUL2, and FUL3), proteins of the SVP-clade by green boxes (VRT2, SVP1, and SVP3) and proteins of the LOFSEP-clade by orange boxes (SEP1-2, SEP1-4, and SEP1-6). Positive interactions between SQUAMOSA and LOFSEP-clade proteins are shown with orange arrows, between SQUAMOSA- and SVP-clade proteins with green arrows, and between SVP- and LOFSEP-clade proteins in gray (the weak interaction between SVP3 and SEP1-6 is indicated by a dotted line). Black curved arrows indicate positive homodimerization. Interactions among SEP proteins were not analyzed.

Figure 12

Wheat VRT2 competes with LOFSEP proteins for interactions with SQUAMOSA proteins in yeast. A–C, Y3H assays were used to test the effect of VRT2 as a competitor, where (A) SEP1-2 (∼OsMADS1), (B) SEP1-4 (∼OsMADS5), and (C) SEP1-6 (OsMADS34) were expressed as DNA-binding domain fusions, and SQUAMOSA proteins VRN1/FUL2/FUL3 were expressed as activation domain fusions. The α-gal activity of the protein interactions in the absence of the competitor is shown in green box plots and in the presence of the competitor in orange box plots. Relative α-gal activity values for each interaction are the average of 12 replicates. **P < 0.01 and ***P < 0.001 (paired t tests, Supplemental Data Set S2). The insets show the α-gal activity for weaker interactions using different scales. Box-plot features are explained in the “Statistical analyses” section of “Material and methods.”

Figure 13

Working model of the role of SVP (VRT2 and SVP1), SQUAMOSA (VRN1, FUL2, and FUL3) and SEP genes on the regulation of wheat plant architecture. Bars on the left represent transcript levels of genes from the three MADS-box clades during wheat development. The three plant models represent the architecture of WT, vrn1 ful2 ful3 triple, and vrt2 svp1 double mutants. Green rectangles represent leaves and sheaths, black lines the spike rachis (circle end = determinate and arrow end = indeterminate). The green Xs represent repressed bracts in the spike and the red Xs repressed buds in the elongating nodes. The lower part of the plants represents vegetative growth (SVP genes only), the region between the two blue lines the elongation zone (central region, SVP + SQUAMOSA genes), and the region above the blue lines the developing spikes (SQUAMOSA + SEP genes). In vrn1 ful2 ful3, the lateral SMs regress to VMs, and the bracts are not suppressed. In vrt2 svp1, the axillary buds in the elongation zone are no longer repressed and develop into axillary spikes or spikelets.