MicroRNA156 and its targeted SPL genes interact with the photoperiod, vernalization, and gibberellin pathways to regulate wheat heading time

TL;DR

This study shows how microRNA156 and SPL genes interact with other pathways to control wheat heading time, offering new tools to improve wheat adaptation.

Contribution

The paper introduces new dominant miR156-resistant SPL alleles that accelerate wheat heading time under various conditions.

Findings

Reduced miR156 or increased SPL gene expression leads to earlier wheat heading time.

SPL proteins interact with DELLA and SQUAMOSA proteins to regulate flowering pathways.

Dominant rSPL alleles accelerate heading time and are available for public use.

Abstract

Heading time has a large impact on adaptation to different environments and crop productivity. In this study, we characterized the effect of the endogenous age pathway on heading time and its interactions with the photoperiod and vernalization pathways in the leaves of tetraploid wheat (Triticum turgidum ssp. durum). Plants with reduced levels of microRNA156 or increased expression of its downstream targets, the SQUAMOSA PROMOTER BINDING PROTEIN‐LIKE genes SPL3, SPL4, and SPL13 exhibited accelerated heading time, with stronger effects under suboptimal inductive conditions. Earlier heading was associated with the upregulation of miR172 and flowering‐promoting genes VRN1, FUL2, and FT1 and the downregulation of flowering‐repressing genes AP2L1 and VRN2. Additionally, we uncovered complex interactions among SPL, SQUAMOSA (VRN1 and FUL2), and DELLA proteins that modulate wheat heading time. We showed that DELLA proteins, which are negative regulators in the gibberellic acid pathway, can interact with SPL proteins reducing their ability to induce flowering. We also discovered previously unknown interactions between SQUAMOSA and DELLA proteins in wheat that compete with the DELLA‐SPL interactions, likely reducing DELLA's ability to repress SPL3 and SPL4 activity. Since SPL3 and SPL4 directly promote VRN1 and FUL2 transcription, these interactions generate a positive regulatory feedback loop that accelerates wheat heading time. Finally, we developed dominant miR156‐resistant alleles rSPL3, rSPL4, and rSPL13 that accelerate wheat heading time under both optimal and suboptimal inductive conditions. These publicly available genetic resources can be used to fine‐tune heading time and improve wheat adaptation to changing environments.

Wheat heading time is critical for adaptation to diverse environments. We generated dominant mutations for the SPL3, SPL4, and SPL13 genes that accelerate heading time. Different combinations of these mutations can be used to modulate heading time and improve wheat adaptation to changing environments.

Figures

miR156‐regulated SPL3, SPL4, and SPL13 promote flowering time in spring wheat.(A–C) Schematic diagrams showing gene structure, conserved SBP domain in green, and miR156 target site in red. Mutations and edits in the sequences are indicated with red letters (A) rSPL‐A4 synonymous mutation in Kronos mutant line K2942. (B) rSPL‐A3 and rSPL‐B3 mutations in Cadenza (CAD1995 and CAD1033) (C) CRISPR‐induced deletions in miR156 binding sites of SPL‐A13 (21 bp) and SPL‐B13 (17 bp) located in the 3′ UTR. The miR156 target sequence is highlighted in yellow, and the PAM site is underlined.(D–K) Days to heading under long (LD). (D) Target mimicry against miR156 (MIM156 n = 15) versus wild‐type (WT n = 7) (E) rSPL‐A4 (n = 13) versus WT (n = 11). (F) rSPL‐A3 (n = 5) versus WT (n = 10). (G) rSPL‐B3 versus WT (both n = 12). (H) rSPL‐A13 (n = 11) versus WT (n = 9). (I) rSPL‐B13 versus WT (both n = 11). (J) rSPL‐A4 and rSPL‐A13 single and double mutants compared to WT (all n = 12). (K) Triple resistant lines combining rSPL‐B3 rSPL‐A4 rSPL‐A13 (n = 15) versus WT (n = 14).(L) Comparison between plants carrying WT, double rSPL‐A4/A13, and triple rSPL‐B3/A4/A13 resistant alleles at head emergence. Error bars represent SEM. P values are from t‐test (ns = not significant, *P < 0.05, P < 0.01, *P < 0.001). Raw data and statistical analyses are available in Data S4.

Effect of Ubi::rSPL‐A13‐HA on gene expression in the leaves and heading time. Transcript levels in the third leaves are expressed relative to ACTIN as endogenous control using the Delta Ct method.(A–C) Transcript levels of (A) SPL13, (B) VRN1, (C) FT1 in the third leaf of Ubi::rSPL‐A13‐HA and a sister line without the transgene (wild‐type, WT).(D) Days to heading and (E) leaf number. Means of the transgenic plants were compared with the WT using Dunnett tests. ns, not significant; *P < 0.05, P < 0.01, *P < 0.001. Raw data and statistical analyses are available in Data S6.

Interactions between the age and the vernalization pathways in wheat leaves.(A) Box plot showing days to heading of wild‐type (WT) and MIM156 plants grown without vernalization (NV) and with 30 or 42 days vernalization.(B, C) Expression levels of VRN1 (B) and VRN2 (C) determined in leaves of WT and plants carrying MIM156 without vernalization or after 30 days vernalization followed by 2 weeks at room temperature (RT).(D–F) Expression levels of FT1 (D), VRN1 (E), and VRN2 (F) in leaves of WT and MIM156 without vernalization and after 42 days vernalization followed by 2 weeks at RT.(G–I) Transcript levels of SPL3 (G), SPL4 (H), and SPL13 (I) in leaves of spring and winter Kronos plants grown under long days (LD). L, fully expanded leaf; W, week. W1 = L1, W5 = L7, W10 = L10–11, W12 = L13–14, W14 = L15–16, W16 = L18–19, and W18 = flag leaves L20–23.(J–L) Box plots showing expression levels of SPL3 (J), SPL4 (K), and SPL13 (L) in L7 of NV winter Kronos plants, and L7 of winter Kronos and vrn1‐null mutant plants vernalized for 8 weeks and then moved to RT for 2 weeks (8wV+2wRT). Transcript levels were quantified by qRT‐PCR using ACTIN as endogenous control. Error bars represent SEM. (A–F) P values are from t‐test (ns, not significant; *P < 0.05, P < 0.01, *P < 0.001). (J–L) Different letters over the bars indicate significant differences in Tukey tests (P < 0.05). Raw data and statistics are available in Data S10.

Genetic interactions between the age and gibberellin pathways in wheat leaves.(A) Plant height (n = 21–23).(B) Days to heading (n = 16–18).(C) Leaf number (n = 14–17). Genotypes: wild‐type (WT), MIM156, rht‐B1‐null, and double mutants MIM156 rht‐B1‐null.(D–G) Expression analysis of VRN1 (D), FUL2 (E), FT1 (F), and VRN2 (G) in the third leaves of WT, MIM156, rht‐B1‐null, and double mutants MIM156 rht‐B1‐null. Expression was determined by qRT‐PCR using ACTIN as an endogenous control. Error bars represent SEM. ns, not significant; *P < 0.05, P < 0.01, and *P < 0.001 for Dunnett tests versus WT. Raw data and statistics in Data S11.

Interactions between DELLA and SQUAMOSA proteins VRN1 and FUL2.(A) Yeast‐two‐hybrid assays showing interactions between DELLA and both VRN1 and FUL2 MADS‐box proteins and negative controls.(B) Confirmation of DELLA—SQUAMOSA interactions by coimmunoprecipitation (co‐IP) assays in Nicotiana benthamiana. Co‐IP experiments were performed using GFP‐tagged magnetic beads. DELLA‐GRASS and the GFP empty vector were detected with the anti‐GFP antibody, while VRN1 and FUL2 were detected with the anti‐FLAG antibody. The GFP and FLAG empty vectors served as the negative control.(C) Alpha‐galactosidase activity Y3H assays showing VRN1 and FUL2 competition (as third protein) for the interaction between the DELLA‐GRAS domain (activation domain, AD) and SPL proteins SPL3 and SPL4 (DNA binding domain, BD). Lack of autoactivation of the DELLA‐GRAS domain and SPL3 and SPL4 proteins is shown in Figure S11a,b. *P < 0.001. Raw data for the Y3H α‐galactosidase activity assays are available in Data S13.

Working model for the role of the age pathway in the regulation of wheat heading time.(A) The upper panel represents the leaves and the lower panel the shoot apical meristem (SAM) and reproductive tissues. The dotted line from FT1 indicates transport of the protein from leaf to SAM through the phloem. Blue arrows indicate promotion and red lines ending in perpendicular lines indicate repression (dotted blue arrows indicate tentative effects). VRN1 and FUL2 interact physically with the DELLA protein, reducing its ability to repress SPL3 and SPL4 protein activity. The three SPL proteins promote the expression of VRN1, generating a positive feedback regulatory loop (circular arrows) that accelerates wheat heading time. VRN1, VRN2, FT1 form a separate positive feedback regulatory loop that promotes heading.(B) Schematic representation of the expression profiles of SPL3, SPL4, SPL13, VRN1, AP2L genes and miR172 in the leaves of non‐vernalized (NV) winter wheat with very late heading time (18 weeks).

Tables

Table 1. Effect of rSPL alleles on days to heading (DTH) and leaf number (LN) relative to wild‐type alleles in three experiments. Raw data are available in supplemental Data S4, S5, S7, and S8

Exp.	Genotype	Photoperiod	Diff. DTH	P a	Diff. LN	P b
1	rSPL‐A3	LD	−1.90	*	−0.70	*
3	rSPL‐A3	LD	−3.18	*	−0.82	*
1	rSPL‐B3	LD	−1.83	***	−0.43	*
2	rSPL‐B3	LD	−3.40	***	0.00	ns
3	rSPL‐B3	LD	−2.14	*	−0.29	*
2	rSPL‐B3	SD	−3.03	*	0.00	ns
1	rSPL‐A4	LD	−1.94	**	−0.50	*
2	rSPL‐A4	LD	−1.33	ns	−0.33	ns
3	rSPL‐A4	LD	−2.88	***	−0.88	***
1	rSPL‐A4	SD	−6.41	***	.	.
2	rSPL‐A4	SD	−4.58	**	−0.70	*
1	rSPL‐A13	LD	−3.06	***	−0.67	*
2	rSPL‐A13	LD	−0.37	ns	0.50	ns
3	rSPL‐A13	LD	−2.65	*	0.15	ns
2	rSPL‐A13	SD	−4.21	*	−0.69	*
1	rSPL‐B13	LD	−2.36	***	.	.
3	rSPL‐B13	LD	−2.89	**	−0.25	ns
1	rSPL‐A4/A13	LD	−5.75	***	−1.33	***
2	rSPL‐A4/A13	LD	−2.83	**	−0.17	ns
2	rSPL‐A4/A13	SD	−6.43	**	−0.86	**
1	rSPL‐B3/A4/A13	LD	−6.55	***	−1.79	***
2	rSPL‐B3/A4/A13	LD	−3.83	***	−1.00	**
2	rSPL‐B3/A4/A13	SD	−6.34	**	−1.00	***
			Correlation DTH vs LN		0.74

Table 2. Effect of rSPL‐resistant alleles on the expression of main flowering genes under long (LD, 16 h light, orange) and short days (SD, 8 h light, light blue). In the significance levels, the yellow highlight indicates higher expression in the resistant allele than in the wild‐type (WT), whereas the blue highlight indicates lower expression in the resistant allele than the WT (P < 0.10). Raw data and statistical tests are available in (Data S9)

Leaf	Ph.P.	Exp.	Gene	Higher in mutant				Lower in mutant
				miR172	FT1	FUL2	VRN1	VRN2	AP2L1
L3	LD	1	rSPL‐ A4	ns	**	ns	ns	ns	ns
L5	LD	1	rSPL‐ A4	*	**	**	ns	ns	NA
L5	LD	2	rSPL‐ A4	ns	***	***	ns	**	**
L6	LD	2	rSPL‐ A4	*	ns	***	**	**	ns
L5	SD	2	rSPL‐ A4	ns	**	**	*	***	ns
L7	SD	2	rSPL‐ A4	*	ns	*	*	0.079	0.061
L9	SD	2	rSPL‐ A4	ns	ns	**	ns	ns	ns
L3	LD	1	rSPL‐ A3	ns	ns	ns	**	*	0.063
L5	LD	1	rSPL‐ A3	ns	ns	ns	ns	ns	ns
L5	SD	2	rSPL‐ A3	ns	ns	ns	ns	ns	0.097
L7	SD	2	rSPL‐ A3	ns	ns	ns	*	ns	ns
L9	SD	2	rSPL‐ A3	0.055	ns	ns	ns	ns	ns
L3	LD	1	rSPL‐ B3	ns	ns	0.072	0.066	ns	ns
L5	LD	2	rSPL‐ B3	ns	ns	*	ns	0.053	**
L6	LD	2	rSPL‐ B3	ns	ns	*	**	0.063	ns
L5	SD	2	rSPL‐ B3	ns	ns	ns	*	ns	ns
L7	SD	2	rSPL‐ B3	ns	ns	ns	ns	ns	ns
L9	SD	2	rSPL‐ B3	ns	ns	ns	ns	ns	ns
L3	LD	1	rSPL‐ A13	ns	ns	ns	ns	0.089	*
L5	LD	1	rSPL‐ A13	ns	ns	ns	ns	ns	ns
L5	LD	2	rSPL‐ A13	ns	ns	ns	ns	ns	ns
L6	LD	2	rSPL‐ A13	ns	ns	ns	ns	ns	ns
L5	SD	2	rSPL‐ A13	ns	ns	ns	ns	**	ns
L7	SD	2	rSPL‐ A13	ns	*	*	ns	*	ns
L9	SD	2	rSPL‐ A13	*	*	ns	*	ns	ns
L3	LD	2	rSPL‐B13	ns	ns	ns	ns	ns	ns
L5	LD	2	rSPL‐B13	ns	ns	**	ns	*	ns
L6	LD	2	rSPL‐B13	ns	*	***	0.068	*	ns
L5	SD	2	rSPL‐B13	ns	*	ns	*	***	ns
L7	SD	2	rSPL‐B13	ns	**	**	ns	**	ns
L9	SD	2	rSPL‐B13	**	ns	*	ns	ns	ns

Full text

INTRODUCTION

Wheat (Triticum aestivum) provides more than one fifth of the calories and protein consumed by the human population (FAOSTAT, 2023) and is important for global food security. Continuous increases in wheat grain yield are required to feed a growing human population in a rapidly changing environment. An important determinant of wheat adaptation to these changes is the adjustment of the reproductive phase to the time of the year with optimal conditions for reproductive success. The precise timing of the transition from the vegetative to the reproductive phases and of spike emergence (heading time) is critical to maximizing wheat grain yield.

Natural variation in heading time is mainly associated with genes regulating vernalization (long exposures to cold temperatures) and photoperiod (length of the day–night cycle) requirements (Beales et al., 2007; Distelfeld et al., 2009; Fu et al., 2005; Wilhelm et al., 2009; Yan et al., 2006; Yan, Helguera, et al., 2004; Yan, Loukoianov, et al., 2004). Wheat cultivars are classified as winter or spring types based on their vernalization requirements and as photoperiod‐sensitive (PS) or photoperiod‐insensitive (PI) based on the difference in heading time between long (LD) and short days (SD).

Winter wheat is sown in the fall, when flowering is prevented by high expression levels of the VERNALIZATION2 (VRN2) gene, a LD repressor of the florigen‐encoding gene FLOWERING LOCUS T1 (FT1, also known in wheat as VRN3) (Yan et al., 2006; Yan, Loukoianov, et al., 2004). The long exposure to cold temperatures during winter results in the upregulation of VERNALIZATION1 (VRN1) in the shoot apical meristem (SAM), which promotes the transition to the reproductive phase (Oliver et al., 2009; Yan et al., 2003). VRN1 is also upregulated in the leaves, where it prevents the upregulation of VRN2 in spring by directly binding to its promoter (Chen & Dubcovsky, 2012; Deng et al., 2015). As day length increases during spring, VRN2 is repressed by VRN1, and the PHOTOPERIOD1 (PPD1) gene induces the expression of FT1 in the leaves. The FT1 protein is then transported to the SAM via the phloem, where it further upregulates VRN1 expression and promotes gibberellin (GA) biosynthesis, thereby accelerating spike development and inducing stem elongation (Pearce et al., 2013). A positive feedback loop between VRN1, VRN2, and FT1 in the leaves ensures an irreversible commitment to flowering after vernalization (Distelfeld et al., 2009; Loukoianov et al., 2005).

Wheat heading time is also regulated by the endogenous age pathway (Debernardi et al., 2022). Micro RNA156 (miR156), which is expressed at high levels in juvenile tissues and declines with plant age, plays a central and conserved role in regulating the transition between the juvenile and adult stages during the vegetative growth phase (Lawson & Poethig, 1995; Poethig, 2009; Wu et al., 2009; Wu & Poethig, 2006). Constitutive expression of miR156 leads to a prolonged juvenile phase and late flowering, while reducing miR156 activity by expressing artificial target mimics (MIM156) reduces juvenile traits and accelerates flowering (Chuck et al., 2007; Debernardi et al., 2022; Schwab et al., 2005; Wang, Wang, et al., 2015; Wu et al., 2009; Xie et al., 2006).

In Arabidopsis, miR156 affects plant development by controlling the expression of SQUAMOSA PROMOTER BINDING PROTEIN (SBP)‐LIKE (SPL) genes (Wu & Poethig, 2006). Members of this gene family encode plant‐specific transcription factors (TFs) that bind to regulatory regions of MADS‐box genes within the SQUAMOSA clade (Huijser et al., 1992; Klein et al., 1996) and promote flowering in Arabidopsis (Yu & Wang, 2020). SPL genes also promote the transition to the reproductive phase by inducing the expression of miR172 (Hyun et al., 2016; Wang, 2014; Wang et al., 2009; Wang, Sun, et al., 2015; Wu et al., 2009). The age‐dependent induction of miR172 results in the downregulation of its targeted APETALA2‐LIKE (AP2L) genes, which function as repressors of the flowering transition. In wheat and other plant species, ectopic expression of miR172 or knock‐out mutants of AP2L genes accelerate flowering, whereas transgenic expression of a target mimicry against miR172 (MIM172) or miR172‐resistant versions of AP2L genes (rAP2L) delay flowering (Aukerman & Sakai, 2003; Debernardi et al., 2022; Lee et al., 2014). The expression of miR172 is also regulated by GIGANTEA in both Arabidopsis (Jung et al., 2007) and wheat (Li et al., 2024), linking the endogenous and photoperiod pathways.

The activity of some SPL TFs is also regulated through protein–protein interactions with DELLAs (Yu et al., 2012), which are negative regulators in the GA signaling pathway. In the absence of GA, DELLA proteins repress growth through physical interaction with their target proteins (Sun, 2010). GA promotes the degradation of DELLA by binding to a nuclear receptor, GIBBERELLIN INSENSITIVE DWARF1 (GID1) (Griffiths et al., 2006; Harberd et al., 2009; Sun, 2010). The roles of DELLA proteins and GA in the regulation of the SPL activity and the duration of the juvenile phase contribute to their effects on flowering time in Arabidopsis (Galvao et al., 2012; Hyun et al., 2016; Yu et al., 2012). In the absence of GA, DELLA can directly interact with SPL9 and interfere with its transcriptional activity (Yu et al., 2012). Increasing levels of GA with plant age promote DELLA degradation, thereby releasing SPL proteins that promote flowering. GA‐deficient mutants or mutants impaired in GA signaling show late flowering, whereas exogenous GA application in wheat accelerates heading time (Griffiths et al., 2006; Pearce et al., 2013; Wang, 2014).

We have recently shown that in spring wheat the miR156‐miR172‐AP2L pathway regulates FT1 expression in the leaves, whereas in winter wheat the regulation of AP2L1 expression is decoupled from the age‐dependent downregulation of miR156. In winter varieties, the induction of VRN1 by vernalization is required to repress AP2L1 in the leaves and promote flowering (Debernardi et al., 2022). However, the roles of wheat SPL genes in the endogenous age pathway and their interactions with the distinctive photoperiod and vernalization pathways of temperate grasses remain underexplored.

In this work, we identified three wheat SPL genes regulated by miR156 in the leaves and characterized their effects on heading time under optimal and suboptimal photoperiodic and vernalization conditions. We also explored the interactions among the endogenous age pathway and the photoperiod, vernalization, and GA pathways and discovered interactions between DELLA and SQUAMOSA proteins that have not been reported before. Finally, we developed dominant SPL alleles and discussed how they can be manipulated to fine‐tune wheat heading time and facilitate the development of wheat varieties better adapted to changing environments.

RESULTS

Identification of miR156‐targeted SPL genes in tetraploid wheat

We identified 19 SPL genes in each of the three genomes of hexaploid wheat (T. aestivum, genomes AABBDD) cultivar Chinese Spring (International Wheat Genome Sequencing Consortium, 2018) (RefSeq v1.1), 9 of which have complementary regions to miR156 (SPL2, SPL3, SPL4, SPL7, SPL13, SPL14, SPL16, SPL17, and SPL18; Figure S1a). Using an alignment of the conserved SBP domains from wheat, rice, and Arabidopsis SPL proteins (Figure S2), we generated a phylogenetic tree (Figure S3; Method S1), which is similar to previous studies (Cao et al., 2019, 2021; Chen et al., 2023; Gupta et al., 2023). SPL genes were named (Data S1) following a published phylogenetic‐based wheat nomenclature (Gupta et al., 2023).

Using previously published RNA‐seq data from multiple mature tissues (Figure S1b) and single‐molecule fluorescence in situ hybridization (smFISH) in the early transition of the SAM to an inflorescence meristem (IM; Figure S1c) (Xu et al., 2025), we found that among the nine SPL genes with miR156 binding sites (Figure S1a), four were expressed in mature leaves and developing spikes (SPL2, SPL3, SPL4, and SPL13), four mainly in spikes (SPL14, SPL16, SPL17, and SPL18), and one (SPL7) was almost undetectable in all tissues (Figure S1b). In the immature leaves, SPL3, SPL4, SPL13, and SPL14 showed higher expression than the other SPL genes (Figure S1c). Since the objective of this study was to characterize the effect of the wheat SPL genes on the regulatory gene network affecting heading time in the leaves, we initially focused on the four miR156‐regulated SPL genes expressed in mature leaves (SPL2, SPL3, SPL4, and SPL13).

To simplify the genetic analyses, we performed all our experiments in the tetraploid wheat (Triticum turgidum ssp. durum, genomes AABB) cultivar Kronos. Kronos carries the dwarfing allele Rht‐B1b (GA‐insensitive), the Vrn‐A1 allele for spring growth habit, and the reduced photoperiod sensitivity allele Ppd‐A1a that confers earlier heading under SD. We first validated the expression of SPL2, SPL3, SPL4, and SPL13 in the leaves by quantitative reverse transcription‐polymerase chain reaction (qRT‐PCR) using primers and methods described in Data S2 and Method S2. All four genes showed increased transcript levels in the leaves with plant age (Figure S4a–d) and altered expression in transgenic plants with increased or reduced levels of miR156. SPL2, SPL3, SPL4, and SPL13 were significantly down‐regulated in the leaves of plants overexpressing miR156 (Figure S4e–h), and upregulated in plants expressing a target mimicry against miR156 (MIM156; Figure S4i–l; Data S3). SPL2 transcript levels in leaves were >10‐fold lower than those of other SPL genes and were significantly upregulated in only two of the five MIM156 transgenic lines (Figure S4l). Based on these results, SPL2 was excluded from further studies.

Recessive mutations have limited effects in polyploid species, so we targeted the miR156 binding sites of SPL3, SPL4, and SPL13 to generate dominant alleles. Using a public database of sequenced EMS‐induced mutations in tetraploid wheat Kronos and hexaploid wheat Cadenza (Krasileva et al., 2017), we identified mutations within the miR156 target sites for SPL‐A4 in Kronos, and for SPL‐A3 and SPL‐B3 in Cadenza (Figure 1A,B). The mutations from Cadenza were transferred to Kronos by marker‐assisted backcrossing (see Method S3). We did not find EMS‐induced mutations within the miR156 binding site of SPL13, so we generated CRISPR‐induced mutations within this region for both SPL‐A13 and SPL‐B13 in Kronos (Figure 1C; Method S4). These selected mutations reduce the binding energy of miR156 (Figure 1A–C), resulting in miR156‐resistant alleles (henceforth, rSPL). All three rSPL alleles were expressed at higher levels than their respective wild‐type (WT) alleles (Data S3; Figure S5).

*miR156‐regulated SPL3, SPL4, and SPL13 promote flowering time in spring wheat.(A–C) Schematic diagrams showing gene structure, conserved SBP domain in green, and miR156 target site in red. Mutations and edits in the sequences are indicated with red letters (A) rSPL‐A4 synonymous mutation in Kronos mutant line K2942. (B) rSPL‐A3 and rSPL‐B3 mutations in Cadenza (CAD1995 and CAD1033) (C) CRISPR‐induced deletions in miR156 binding sites of SPL‐A13 (21 bp) and SPL‐B13 (17 bp) located in the 3′ UTR. The miR156 target sequence is highlighted in yellow, and the PAM site is underlined.(D–K) Days to heading under long (LD). (D) Target mimicry against miR156 (MIM156 n = 15) versus wild‐type (WT n = 7) (E) rSPL‐A4 (n = 13) versus WT (n = 11). (F) rSPL‐A3 (n = 5) versus WT (n = 10). (G) rSPL‐B3 versus WT (both n = 12). (H) rSPL‐A13 (n = 11) versus WT (n = 9). (I) rSPL‐B13 versus WT (both n = 11). (J) rSPL‐A4 and rSPL‐A13 single and double mutants compared to WT (all n = 12). (K) Triple resistant lines combining rSPL‐B3 rSPL‐A4 rSPL‐A13 (n = 15) versus WT (n = 14).(L) Comparison between plants carrying WT, double rSPL‐A4/A13, and triple rSPL‐B3/A4/A13 resistant alleles at head emergence. Error bars represent SEM. P values are from t‐test (ns = not significant, *P < 0.05, **P < 0.01, **P < 0.001). Raw data and statistical analyses are available in Data S4.

In summary, we identified three wheat SPL genes targeted by miR156 and expressed in the leaves, and generated five dominant miR156‐resistant alleles that were expressed at higher levels than their respective WT alleles.

Increased expression of SPL3, SPL4, and SPL13 accelerates heading time and reduces leaf number (LN)

We first explored the effect of MIM156 and SPL‐resistant alleles on heading time under LDs (Table 1). The MIM156 transgene led to the simultaneous upregulation of multiple SPL genes (Figure S4i–k) and accelerated heading by 5.2 days relative to the WT (Figure 1D). The individual rSPL alleles showed smaller effects on heading time (Figure 1E–I). The synonymous mutation in rSPL‐A4 did not alter the encoded protein but affected the binding energy of the miR156 (Figure 1A) and accelerated heading 2.1 ± 0.5 days (average ± SEM of three experiments) relative to the WT (Table 1; Figure 1E).

The mutations in rSPL‐A3 (serine to phenylalanine) and rSPL‐B3 (synonymous mutation) miR156 binding sites were associated with higher SPL3 transcript levels (Data S3; Figure S5) and similar accelerations in heading time relative to their WT sister lines (average 2.5 ± 0.6 days in rSPL‐A3 and 2.5 ± 0.5 days in rSPL‐B3; Figure 1F,G; Table 1). The similar effects of the rSPL‐A3 and rSPL‐B3 alleles on heading time suggested that the amino acid change in rSPL‐A3 did not affect the function of the encoded protein.

The miR156 binding site of SPL13 is in the 3′ UTR, so the CRISPR‐induced indels in this region in rSPL‐A13 (21‐bp deletion) and rSPL‐B13 (17‐bp deletion) are not expected to affect the encoded proteins (Figure 1C). In two Cas9‐free populations segregating for either rSPL‐A13 or rSPL‐B13, plants carrying the resistant alleles showed higher SPL13 transcript levels (Figure S5) and headed 2.0 ± 0.8 days (rSPL‐A13; Figure 1H) and 2.6 ± 0.3 days (rSPL‐B13; Figure 1I) earlier than their respective WT controls.

We then intercrossed plants carrying the rSPL‐A4 and rSPL‐A13 alleles to study their combined effects. Plants carrying both rSPL‐A4 and rSPL‐A13 alleles headed on average 5.8 days earlier than the WT, whereas smaller differences were observed for sister plants carrying rSPL‐A4 (3.8 days) or rSPL‐A13 (3.3 days) alone (Figure 1J). The factorial anova for days to heading (DTH) showed highly significant effects for the individual resistant alleles (P < 0.001) and a marginally non‐significant interaction (P = 0.07; Data S4). These results suggest that the positive effects of the two alleles on DTH are mostly additive. Finally, we combined the three resistant alleles rSPL‐B3, rSPL‐A4, and rSPL‐A13, and observed that plants homozygous for all three mutations headed 6.6 days earlier than the sister lines without any resistant allele (Figure 1K; Data S4). A photograph comparing head emergence in plants carrying the combined resistant alleles and the WT is presented in Figure 1L. Similar photographs for the individual SPL‐resistant alleles and their corresponding WT sister lines are presented in Figure S6.

To determine which developmental phase was accelerated in the rSPL mutants, we measured LN, which is fixed at the time of the SAM transition to an IM. Plants carrying the single rSPL‐A3 or rSPL‐B3 alleles showed an average of 0.8 ± 0.1 and 0.24 ± 0.1 fewer leaves than the WT, respectively (Table 1; Figure S7a,b). Similar differences were observed in plants carrying homozygous rSPL‐A4 alleles (0.6 ± 0.2 leaves; Table 1; Figure S7d). The effects of rSPL‐A13 and rSPL‐B13 on LN were significant in only one of the four experiments conducted under LD (Table 1; Figure S7c,d).

A factorial anova for LN comparing the individual and combined rSPL‐A4 rSPL‐A13 alleles showed significant reductions in LN (P = 0.02) for both rSPL‐A4 (0.50 leaves) and rSPL‐A13 (0.67 leaves), and highly significant reductions (P < 0.001) for the combined alleles (1.3 leaves; Figure S7d; Table 1). The interaction between the two genes was not significant, indicating additive effects (Data S5). The triple mutant carrying the rSPL‐B3, rSPL‐A4, and rSPL‐A13 alleles had on average 1.4 ± 0.4 fewer leaves than the WT across two experiments performed under LD (Table 1; Figure S7e).

To further characterize the effect of SPL13, we generated transgenic Kronos plants overexpressing the SPL‐A13 coding region fused with a C‐terminal HA tag under the maize UBIQUITIN promoter (Method S4). This transgene does not include the 3′ UTR with the miR156 binding site and, therefore, is insensitive to miR156 activity (Ubi::rSPL‐A13‐HA). All five independent transgenic plants showed higher SPL13 expression levels than the non‐transgenic controls in the third leaf, but the differences were significant only for three of them (Figure 2A; Data S6). These same three transgenic plants showed significant increases in VRN1 expression (Figure 2B), but the differences for FT1 were not significant (Figure 2C). In agreement with the expression results, all five transgenic plants showed reduced DTH (average 2.1 days; Figure 2D) and LN (average 1 leaf; Figure 2E).

*Effect of Ubi::rSPL‐A13‐HA on gene expression in the leaves and heading time. Transcript levels in the third leaves are expressed relative to ACTIN as endogenous control using the Delta Ct method.(A–C) Transcript levels of (A) SPL13, (B) VRN1, (C) FT1 in the third leaf of Ubi::rSPL‐A13‐HA and a sister line without the transgene (wild‐type, WT).(D) Days to heading and (E) leaf number. Means of the transgenic plants were compared with the WT using Dunnett tests. ns, not significant; *P < 0.05, **P < 0.01, **P < 0.001. Raw data and statistical analyses are available in Data S6.

In summary, the increased SPL expression of the five rSPL alleles and Ubi::rSPL‐A13 was associated with relatively similar accelerations of heading times and reductions in LN under LD, and the effects of combined alleles were additive.

rSPL4 and rSPL13 have stronger effect on heading time under SD than under LD

Next, we compared the effect of the rSPL alleles on DTH and LN under both LD and SD (Table 1; Figure S8). For rSPL‐A4, we performed two experiments, which showed similar results: stronger differences in DTH between genotypes under SD (6.4 and 4.6 days) than under LD (1.9 and 1.3 days) (Figure S8a,b; Data S7). Factorial anovas including photoperiod and SPL4 alleles showed highly significant effects for both factors (P < 0.0001) and significant interactions in both experiments (P = 0.024 and 0.029, respectively; Data S7), confirming the stronger effect of the rSPL‐A4 allele on DTH under SD than under LD.

The rSPL‐A13 allele also showed a stronger effect on heading time under SD, which was supported by a significant genotype × photoperiod interaction in the factorial anova analysis (Figure S8c; Data S7). By contrast, the differences between rSPL‐B3 and its WT sister lines were similar under LD and SD (Figure S8d). As expected, the combined double rSPL‐A4/A13 and triple rSPL‐B3/A4/A13 resistant alleles also showed stronger effects on DTH under SD than under LD (Figure S8e,f; Data S7).

Surprisingly, the differences in LN between genotypes were not affected by photoperiod and showed similar values in SD and LD, both for the single and combined mutants (Table 1). These results suggested that the larger differences in DTH under SD relative to LD were associated mainly with changes in the spike differentiation and elongation phase rather than in the transition from the SAM to the reproductive stage. A relatively high correlation between the differences between genotypes in DTH and LN (R = 0.74) was observed when single and combined rSPL alleles were analyzed together (Table 1).

In summary, the differences between rSPL and WT alleles were larger under SD than under LD for DTH (for rSPL‐A4 and rSPL‐A13) but not for LN.

rSPL alleles upregulate flowering‐promoting genes and downregulate flowering‐repressing genes in leaves

To explore the molecular mechanisms underlying the earlier heading of the rSPL mutants relative to their sister WT lines, we compared the transcript levels of several flowering‐promoting and flowering‐repressing genes at different leaves using qRT‐PCR. We performed these comparisons in older leaves for plants grown under SD (L5, L7, and L9) than under LD (L3, L5, and L6) to account for the delayed transition to the reproductive phase under SD (Table 2). Despite some variability, an overall analysis of the significant comparisons in Table 2 revealed a consistent upregulation of flowering‐promoting genes (miR172, FT1, FUL2, and VRN1) and downregulation of flowering‐repressing genes (VRN2 and AP2L1) associated with the rSPL alleles (Data S9). These expression results are consistent with the earlier heading of the plants carrying the rSPL alleles.

Table 2: Effect of rSPL‐resistant alleles on the expression of main flowering genes under long (LD, 16 h light, orange) and short days (SD, 8 h light, light blue). In the significance levels, the yellow highlight indicates higher expression in the resistant allele than in the wild‐type (WT), whereas the blue highlight indicates lower expression in the resistant allele than the WT (P < 0.10). Raw data and statistical tests are available in (Data S9)

Comparisons between rSPL‐A4 and their WT sister lines showed the highest proportion of significant tests, with significant effects observed in all six genes. A smaller number of assays were significant for the rSPL3 alleles, and they were primarily concentrated on the FUL2 and VRN1 genes. This latter result suggested that these two MADS‐box genes may be the initial targets of the SPL3 proteins. Plants carrying the rSPL13 alleles showed a significant difference for all flowering genes, but AP2L1 was significant for only one comparison in L3. Significant differences for rSPL13 were concentrated in experiments performed under SD and were more frequent for rSPL‐B13 than for rSPL‐A13 (Table 2), which coincides with the higher expression levels of rSPL‐B13 in leaves (Figure S1b). The positive effect of rSPL13 on VRN1 expression was validated in Ubi::rSPL‐A13 transgenic plants ectopically expressing SPL‐A13 (Figure 2B).

In summary, the expression results suggest that the rSPL3, rSPL4, and rSPL13 alleles accelerated heading time by upregulating flowering‐promoting genes and downregulating flowering‐repressing genes in the leaves.

The miR156/SPL pathway modulates vernalization requirement and heading time in winter wheat

To explore the interactions between the age and vernalization pathways, we introduced the MIM156 cassette into a winter Kronos line carrying a loss‐of‐function mutation in the Vrn‐A1 allele for spring growth habit (Chen & Dubcovsky, 2012). Plants subjected to different vernalization treatments were synchronized through sequential planting dates and were at a similar developmental stage at the time of removal from the cold treatment.

In the absence of vernalization, the MIM156 winter plants headed 59 days earlier than the control winter Kronos plants without the transgene (Figure 3A). However, this difference was reduced to 34 days after 30 days of vernalization and to 22 days after 42 days of vernalization. A factorial anova for heading time (Data S10) revealed highly significant effects of genotype (P < 0.001), vernalization treatment (P < 0.001), and a significant interaction (P = 0.0016; Figure 3A; Data S10) that reflects the increased effects of MIM156 on heading time under suboptimal flowering inductive conditions (Figure 3A).

*Interactions between the age and the vernalization pathways in wheat leaves.(A) Box plot showing days to heading of wild‐type (WT) and MIM156 plants grown without vernalization (NV) and with 30 or 42 days vernalization.(B, C) Expression levels of VRN1 (B) and VRN2 (C) determined in leaves of WT and plants carrying MIM156 without vernalization or after 30 days vernalization followed by 2 weeks at room temperature (RT).(D–F) Expression levels of FT1 (D), VRN1 (E), and VRN2 (F) in leaves of WT and MIM156 without vernalization and after 42 days vernalization followed by 2 weeks at RT.(G–I) Transcript levels of SPL3 (G), SPL4 (H), and SPL13 (I) in leaves of spring and winter Kronos plants grown under long days (LD). L, fully expanded leaf; W, week. W1 = L1, W5 = L7, W10 = L10–11, W12 = L13–14, W14 = L15–16, W16 = L18–19, and W18 = flag leaves L20–23.(J–L) Box plots showing expression levels of SPL3 (J), SPL4 (K), and SPL13 (L) in L7 of NV winter Kronos plants, and L7 of winter Kronos and vrn1‐null mutant plants vernalized for 8 weeks and then moved to RT for 2 weeks (8wV+2wRT). Transcript levels were quantified by qRT‐PCR using ACTIN as endogenous control. Error bars represent SEM. (A–F) P values are from t‐test (ns, not significant; *P < 0.05, **P < 0.01, **P < 0.001). (J–L) Different letters over the bars indicate significant differences in Tukey tests (P < 0.05). Raw data and statistics are available in Data S10.

Two weeks after removing plants from the different cold treatments, we collected leaf samples and measured the expression of several flowering genes. VRN1 transcripts were almost undetectable in the non‐vernalized plants but increased after 30 days of vernalization, with MIM156 plants showing higher expression levels than the control without the transgene (Figure 3B). The VRN2 transcripts decreased after the partial vernalization treatment in the MIM156 plants but not in the WT (Figure 3C). Plants exposed to 42 days vernalization showed higher FT1 and VRN1 and lower VRN2 transcript levels than those treated for 30 days, with significantly larger effects in the MIM156 plants than in the WT (Figure 3D–F). These expression profiles were consistent with the earlier heading of the plants carrying the MIM156 transgene relative to the control (Figure 3A) and the significant interaction between age and vernalization pathways in the regulation of heading time.

A previous study showed a sharp decrease in miR156 transcript levels in leaves of both spring and non‐vernalized winter Kronos plants during the first 5 weeks (Debernardi et al., 2022). Spring plants headed before the next sampling point, whereas the non‐vernalized winter plants continued to produce leaves, with the flag leaf emerging at 18 weeks. During this extended period, the winter plants showed a gradual downregulation of miR156 and upregulation of miR172 in the leaves and a delayed upregulation of VRN1 and FT1 until week 16 (Debernardi et al., 2022). Using the same RNA samples, we observed that transcript levels of both SPL4 and SPL13 increased to similar levels in spring and winter wheat during the first 5 weeks (Figure 3H,I) and then continued to increase in winter wheat. The initial SPL3 upregulation was delayed in winter Kronos but then increased progressively with age (Figure 3G). Both SPL3 and SPL4 transcript levels peaked at week 16, whereas SPL13 transcripts increased until week 18 (Figure 3G–I).

In a separate experiment, we compared SPL expression levels in RNA samples from the 7th leaf collected from either un‐vernalized winter Kronos plants or 8‐weeks‐vernalized plants after they were returned to room temperature for 2 weeks. No significant differences were detected for SPL3 or SPL4, but SPL13 showed significantly higher levels in the vernalized plants carrying a functional VRN‐B1 allele than in the vrn1‐null mutant plants (Figure 3J–L). This result suggested that VRN1 or some of its downstream targets contributed to the transcriptional upregulation of SPL13 and explained the effect of VRN1 on the upregulation of miR172 and the downregulation of AP2L1 reported in the same samples in a previous study (Debernardi et al., 2022).

In summary, the effects of miR156 on DTH were stronger in non‐vernalized than in vernalized plants, and reduced levels of miR156 reduced the vernalization requirement in winter wheat. In the absence of vernalization in winter wheat, SPL genes were upregulated with age.

SPL proteins interact with SQUAMOSA regulatory regions

To test if the positive effects of the rSPL3, rSPL4, and rSPL‐13 alleles on the transcriptional regulation of VRN1 and FUL2 (Table 2) were the result of direct physical interactions, we performed yeast‐one‐hybrid (Y1H; Method S5; Figure S9) and electrophoretic mobility shift assays (EMSA; Method S6; Figure S10). We selected three VRN1 and five FUL2 regulatory regions carrying putative SPL‐binding sites for the Y1H assays (Data S2). However, the VRN‐B1 and FUL‐A2 promoter regions including a putative SPL‐binding motif near the transcription start site exhibited strong activation by endogenous yeast TFs and were excluded from the Y1H assays (Data S2). Instead, we evaluated the FUL‐A2 promoter region (−399 to −374) using EMSA and found clear interactions with all three SPL proteins (Figure S10). The decreased intensity of the shifted DNA‐protein band upon the addition of non‐labeled DNA (cold‐probe) further validated the specificity of the interaction.

The orthologous promoter region of VRN‐A1 including one putative SPL‐binding site (−404 to −350; Data S2) was able to interact with SPL4 but not with SPL3 or SPL13 in the Y1H assays (Figure S9a). The region in the first intron of VRN‐B1 (+190 to +244) including a putative SPL‐binding site failed to interact with any of the three SPL proteins tested (Figure S9b). This region was not evaluated in the VRN‐A1 homeolog, which has a large deletion in the first intron (Fu et al., 2005).

In addition to the FUL‐A2 promoter region near the transcription start site, we evaluated upstream promoter regions in FUL‐A2 (−836 and −666; Figure S9c) and FUL‐B2 (−992 to −826; Figure S9d), both of which contain two putative SPL‐binding sites (Data S2). The FUL‐A2 region showed a weak interaction only with SPL4, whereas the similar FUL‐B2 region (93% identity) failed to interact with any of the SPL proteins (Figure S9c,d). One of the putative SPL‐binding sites is located closer to the border of the DNA bait in FUL‐B2 than in FUL‐A2 (Data S2) and may have reduced the ability of the FUL‐B2 bait to interact with SPL4. Finally, we evaluated a region in the first intron of FUL2 that included three putative SPL‐binding sites in FUL‐A2 (+2556 to +2652) and four in FUL‐B2 (+1346 to +1514; Data S2). Both the FUL‐A2 and FUL‐B2 regions showed interactions with SPL3 and SPL4 but not with SPL13 (Figure S9e,f).

In summary, only SPL4 interacted with the VRN1 promoter region (−404 to −350), whereas all three SPL proteins were able to interact with the homeologous FUL‐A2 promoter region. In addition, SPL3 and SPL4 (but not SPL13) interacted with a region from FUL2 first intron, suggesting functional differentiation among SPL proteins.

SPL interactions with DELLA proteins modulate their effects on heading time

We first confirmed that the DELLA‐SPL physical interactions previously reported in Arabidopsis (Hyun et al., 2016; Yu et al., 2012) were conserved in wheat using yeast‐two‐hybrid (Y2H) assays (Method S7). We used a truncated DELLA protein containing only the C‐terminal GRAS domain (RHT1‐GRAS), as the full‐length protein exhibits autoactivation. We detected positive interactions between RHT1‐GRAS and both SPL3 and SPL4, but not with the smaller SPL13 protein (Figure S11a,b). These Y2H interactions were validated by bimolecular fluorescence complementation (BiFC) using rice protoplasts in two independent experiments (Figure S12; Method S8) and by coimmunoprecipitation (co‐IP) assays in Nicotiana benthamiana via Agrobacterium‐mediated infiltration (Figure S13; Method S9). Taken together, these results confirmed that both SPL3 and SPL4 can interact with DELLA‐GRAS in planta.

To determine the critical SPL region required for the DELLA‐SPL interaction (Figure S11c), we generated two SPL‐A3 mutants encoding truncated proteins (Figure S11d). The DELLA‐GRAS domain interacted with the truncated SPL‐A3‐Δ157 protein but not with the shorter SPL‐A3‐Δ194 (Figure S11e), which is missing 37 amino acids present in SPL‐A3‐Δ157. These 37 amino acids are well conserved in wheat SPL4 and Arabidopsis AtSPL2, AtSPL10 and AtSPL11, all of which also interact with DELLA in Y2H assays (Yu et al., 2012). These results suggested that this 37‐amino acid region is important for the SPL‐DELLA interaction (Figure S11f).

We then tested if the physical interactions between SPL and DELLA proteins were associated with genetic interactions in a cross between a Kronos line carrying MIM156 and a Kronos line containing a deletion encompassing the GA‐insensitive Rht‐B1b allele, designated hereafter as rht‐B1‐null. This mutant line still has a functional Rht‐A1a allele and, therefore, is GA‐sensitive (Method S3). Analyses of the main effects in the F_2_ progeny showed that plants homozygous for the *rht‐B1‐*null allele were, on average, 19.4 cm taller than those homozygous for the Rht‐B1b allele (Figure 4A; Data S11). A smaller effect on plant height was observed in the plants carrying the MIM156 transgene, which were 5.5 cm taller than those without the transgene (Figure 4A; Data S11). By contrast, the effects of MIM156 on heading time (4.4 days earlier; Figure 4B) and LN (1.3 fewer leaves; Figure 4C) were stronger than those associated with the RHT1 alleles (1.7 days earlier and 0.6 fewer leaves).

*Genetic interactions between the age and gibberellin pathways in wheat leaves.(A) Plant height (n = 21–23).(B) Days to heading (n = 16–18).(C) Leaf number (n = 14–17). Genotypes: wild‐type (WT), MIM156, rht‐B1‐null, and double mutants MIM156 rht‐B1‐null.(D–G) Expression analysis of VRN1 (D), FUL2 (E), FT1 (F), and VRN2 (G) in the third leaves of WT, MIM156, rht‐B1‐null, and double mutants MIM156 rht‐B1‐null. Expression was determined by qRT‐PCR using ACTIN as an endogenous control. Error bars represent SEM. ns, not significant; *P < 0.05, **P < 0.01, and **P < 0.001 for Dunnett tests versus WT. Raw data and statistics in Data S11.

In addition to the main effects, we evaluated the interactions between MIM156 and RHT1 (Data S11). For all three traits, the effects of MIM156 were stronger in the GA‐sensitive background (*rht1‐*null, more DELLA) than in the GA‐insensitive background (Rht‐B1b), and the effects of the RHT1 alleles were stronger in the presence of MIM156 than in its absence (Figure 4A–C). Although the factorial anovas showed no statistically significant interactions for plant height, DTH, or LN (Data S11), they did reveal significant interactions for the expression levels of flowering genes VRN1, FUL2, FT1, and VRN2 in the third leaf (Figure 4D–G; Data S12). Consistent with previous results in Figure S4, plants carrying MIM156 also exhibit a highly significant upregulation of SPL3, SPL4, SPL13, and miR172 (Figure S14; Data S12). The expression results were consistent with the earliest heading of the lines combining MIM156 and *rht‐B1‐*null alleles and with the interactions for DTH and LN described above. The effects of MIM156 on the expression of these genes were stronger and more significant in the GA‐sensitive than in the GA‐insensitive backgrounds, and the effects of RHT1 were stronger and more significant in the presence of MIM156 than in its absence (Figure 4D–G; Data S12).

In summary, physical interactions were detected between DELLA and both SPL3 and SPL4 proteins, which were paralleled by significant genetic interactions between DELLA and MIM156 for the regulation of flowering genes. Taken together, these interactions suggested that the effects of the GA and miR156 pathways on heading time and LN are not fully additive.

VRN1 and FUL2 interact with DELLA and compete with the SPL‐DELLA interaction

Since wheat loss‐of‐function mutants vrn1 and ful2 show reduced plant height (Li et al., 2019), we explored the interactions between DELLA and both VRN1 and FUL2. The Y2H assays revealed positive interactions between the DELLA‐GRAS domain and both VRN1 and FUL2 proteins, with FUL2 showing stronger interactions than VRN1 (Figure 5A). These interactions were validated by co‐IP assays in N. benthamiana (Figure 5B; Method S9), confirming that VRN1 and FUL2 can interact with the DELLA‐GRAS domain in planta.

*Interactions between DELLA and SQUAMOSA proteins VRN1 and FUL2.(A) Yeast‐two‐hybrid assays showing interactions between DELLA and both VRN1 and FUL2 MADS‐box proteins and negative controls.(B) Confirmation of DELLA—SQUAMOSA interactions by coimmunoprecipitation (co‐IP) assays in Nicotiana benthamiana. Co‐IP experiments were performed using GFP‐tagged magnetic beads. DELLA‐GRASS and the GFP empty vector were detected with the anti‐GFP antibody, while VRN1 and FUL2 were detected with the anti‐FLAG antibody. The GFP and FLAG empty vectors served as the negative control.(C) Alpha‐galactosidase activity Y3H assays showing VRN1 and FUL2 competition (as third protein) for the interaction between the DELLA‐GRAS domain (activation domain, AD) and SPL proteins SPL3 and SPL4 (DNA binding domain, BD). Lack of autoactivation of the DELLA‐GRAS domain and SPL3 and SPL4 proteins is shown in Figure S11a,b. **P < 0.001. Raw data for the Y3H α‐galactosidase activity assays are available in Data S13.

We then used a yeast‐three‐hybrid system (Y3H; Method S10) to test if the presence of VRN1 or FUL2 expressed as the third protein could interfere with the previously described interaction between DELLA and SPL3 and SPL4 proteins (Figure S11). The quantitative alpha‐galactosidase activity assays confirmed the interactions between the DELLA‐GRAS domain and both SPL3 and SPL4 proteins and showed a significant decrease in the strength of the DELLA‐SPL interactions when either VRN1 or FUL2 was expressed as a third protein (Figure 5C). The presence of FUL2 and VRN1 weakened the DELLA‐SPL3 interactions by 60 and 48%, and the DELLA‐SPL4 interaction by 55 and 36%, respectively (Figure 5C; Data S13). These results are consistent with the weaker physical interaction between DELLA and VRN1 relative to DELLA‐FUL2 (Figure 5A).

In summary, DELLA physically interacted with SQUAMOSA MADS‐box proteins VRN1 and FUL2, and these interactions competed with the DELLA‐SPL interactions.

A model for the role of the leaf‐expressed SPLs in the regulation of wheat heading time

Figure 6 presents a working model that incorporates the proposed roles of SPL3, SPL4, and SPL13, as well as the interactions among the age, vernalization, photoperiod, and GA pathways in the regulation of wheat heading time. This model focuses on interactions in the leaves that converge on the upregulation of FT1.

Working model for the role of the age pathway in the regulation of wheat heading time.(A) The upper panel represents the leaves and the lower panel the shoot apical meristem (SAM) and reproductive tissues. The dotted line from FT1 indicates transport of the protein from leaf to SAM through the phloem. Blue arrows indicate promotion and red lines ending in perpendicular lines indicate repression (dotted blue arrows indicate tentative effects). VRN1 and FUL2 interact physically with the DELLA protein, reducing its ability to repress SPL3 and SPL4 protein activity. The three SPL proteins promote the expression of VRN1, generating a positive feedback regulatory loop (circular arrows) that accelerates wheat heading time. VRN1, VRN2, FT1 form a separate positive feedback regulatory loop that promotes heading.(B) Schematic representation of the expression profiles of SPL3, SPL4, SPL13, VRN1, AP2L genes and miR172 in the leaves of non‐vernalized (NV) winter wheat with very late heading time (18 weeks).

In winter wheat, the repression of the VRN1 chromatin is released by vernalization. VRN1 then promotes FT1 transcription in the spring by two positive regulatory feedback loops (indicated by circle arrows). The first feedback loop, which has been described in detail in the introduction, involves VRN1 repressing VRN2, the release of VRN2 repression of FT1, and FT1's positive regulation of VRN1 (Distelfeld et al., 2009). The second feedback loop involves the interactions between DELLA and the induced VRN1 and FUL2 proteins, which compete with the repressive DELLA interactions with SPL3 and SPL4 (Figure 5C). The SPL proteins then contribute to the transcriptional upregulation of VRN1 and FUL2 by direct binding to their regulatory regions (Table 2; Figures S9 and S10), and indirectly through their positive effect on FT1 induction (Figure 6). FT1 is also regulated by the photoperiod pathway through the LD activation of PPD1. The photoperiod pathway is also connected to the endogenous age pathway through the GI upregulation of miR172 (Li et al., 2024) (Figure 6).

A shorter positive feedback loop has been proposed between VRN1 and FT1 based on the reduced expression of FT1 in the vrn1 mutants and the binding of VRN1 to the FT1 promoter in wheat (Tanaka et al., 2018) and barley (Deng et al., 2015). However, it is difficult to distinguish a direct effect of VRN1 on FT1 from its indirect effects through VRN2 and AP2L1 (Figure 6). A previous study demonstrated that VRN1 can upregulate FT1 in a *vrn2‐*null mutant background (Shaw et al., 2019). However, experiments testing the effect of VRN1 on heading time in a combined vrn2‐ap2l1‐null background are still pending. To reflect this uncertainty, we connected VRN1 to FT1 by a dotted arrow (Figure 6).

The requirement of VRN1 for flowering is not absolute (Chen & Dubcovsky, 2012) and *vrn1‐*null winter wheat Kronos plants eventually head after more than ~120 days at room temperature and LD conditions (Debernardi et al., 2022). The continuous increase of SPL transcript levels with age (Figures 3G–I and 6B) and the redundant roles of FUL2 in spikelet and floral development (Li et al., 2019) likely contribute to heading and flowering in the absence of VRN1. In summary, this model suggests that the SPL genes can regulate FT1 expression in wheat leaves by both the miR172‐AP2L and VRN1‐FUL2‐VRN2 pathways.

DISCUSSION

Distinctive characteristics of the flowering pathways in the temperate cereals

Wheat and other temperate cereals are a monophyletic group characterized by a major expansion and reorganization of their genomes and a common basic chromosome number seven (Kellogg, 2001). These grasses underwent major physiological changes to adapt to cold temperatures, including the development of a vernalization pathway that was not present in the tropical grasses. This pathway is centered on the VRN1 and VRN2 genes and involves the epigenetic de‐repression of VRN1 during vernalization (Liu et al., 2024; Oliver et al., 2009) and its resetting in embryos in the next generation (Niu et al., 2024). This pathway is different from the Arabidopsis vernalization pathway, which is centered on the FLC and FRI genes and involves the epigenetic repression of FLC during vernalization (Sung & Amasino, 2005).

The photoperiod pathways in wheat and Arabidopsis also show major differences. While CO is the main sensor of day length in Arabidopsis (Andres & Coupland, 2012), it only plays a minor role in wheat, where PPD1 is a central daylength sensor (Shaw et al., 2020). Although the photoperiod pathways of both species converge in the upregulation of FT1 in the leaves during LDs, in winter wheat LDs do not induce FT1 unless VRN1 is present in the leaves to suppress FT1 repressors VRN2 and AP2L1 (Figure 6). VRN2 is a grass‐specific FT1 repressor (orthologous to rice Ghd7) that is repressed by VRN1 after vernalization, a pathway that is absent in the tropical grasses (Dubcovsky et al., 2006; Yan, Loukoianov, et al., 2004).

The valuable seasonal information carried by the vernalization‐induced VRN1 protein has been integrated into different pathways in the temperate grasses. In winter wheat, for example, the frost tolerance and cold acclimation pathways are less responsive to cold temperatures when VRN1 is present than when it is absent, resulting in different temperature thresholds in fall and spring (Galiba et al., 2009). The induction of VRN1 also contributes to the downregulation of the FT1 transcriptional repressors VRN2 (Chen & Dubcovsky, 2012) and AP2L1 (Debernardi et al., 2022), thereby generating a more permissive environment for FT1 induction by the photoperiod pathway (Figure 6). The physical interactions of VRN1 and FUL2 with DELLA represent an additional innovation in the temperate grasses that, to the best of our knowledge, have not been reported before. Our Y3H experiments indicate that these interactions can compete with the DELLA‐SPL interactions (Figure 5C), which reduce SPL transcriptional activity in both Arabidopsis (Yu et al., 2012) and wheat (Figure 4D,E).

The VRN1/FUL2—DELLA—SPL physical interactions generate a connection between the endogenous age pathway and the vernalization pathway in wheat, providing a possible mechanistic explanation for the previously reported role of VRN1 in the repression of AP2L1 (Debernardi et al., 2022). In addition, these competitive physical interactions, together with the ability of SPL3, SPL4, and SPL13 to directly upregulate VRN1 and FUL2 transcription, generate a positive feedback loop that reinforces the commitment to flowering (Figure 6). The SPL genes can also contribute indirectly to the upregulation of VRN1 (and possibly FUL2) through the induction of miR172, repression of AP2L1, and induction of FT1, which can bind to the VRN1 promoter as part of the florigen activation complex and induce VRN1 transcription (Li et al., 2015).

Conserved functions between wheat and Arabidopsis SPL homologs

Despite the differences between wheat and Arabidopsis in the interactions between the age and other flowering pathways described above, the central endogenous age pathway is relatively well conserved in these two distant species. This conservation includes the downregulation of miR156 with age and its ability to regulate the SPL genes (Figure S4i–k).

Wheat SPL3 and SPL4 are related to Arabidopsis AtSPL2, AtSPL10 and AtSPL11 (Figure S3) and both are expressed at high levels in leaf primordia. However, only wheat SPL3 and SPL4 are also expressed at high levels in mature leaves (Figure S1b,c) (Xu et al., 2016). Resistant alleles for these SPL genes in wheat (Tables 1 and 2) and Arabidopsis (Yao et al., 2019) have been associated with the direct transcriptional activation of SQUAMOSA MADS‐box genes, reduced LN, and early flowering. In wheat, SPL4 can bind to a promoter region of VRN1, and both SPL3 and SPL4 physically bind to promoter and intron regions of FUL2 (Figures S9 and S10). In Arabidopsis, chromatin immunoprecipitation (ChIP) assays have shown that AtSPL10 binds to the promoters of AtAP1 and AtFUL but not to the first intron (Yao et al., 2019).

The wheat SPL13 gene corresponds to Arabidopsis AtSPL3, AtSPL4, and AtSPL5 (Figure S3), and all of them share a small size, absence of conserved domains other than the SBP domain, and a miR156 binding site in the 3′ UTR. The Kronos rSPL13 allele was associated with a significant acceleration of heading time and altered expression of flowering genes in the leaves, similar to results reported in hexaploid wheat (Gupta et al., 2023). The homologous AtSPL3 gene was expressed in both juvenile and mature leaves but the resistant allele showed no significant acceleration of flowering time (Xu et al., 2016) except when expressed under the 35S promoter (Wang et al., 2009; Wu & Poethig, 2006). In ChIP assays, AtSPL3 showed preferential binding to regions in the promoter and first intron of FUL and SOC1, and the promoter of AP1 (Wang et al., 2009; Yamaguchi et al., 2009). The wheat SPL13 homolog was also able to bind to promoter regions of FUL2 but not to the region in the first intron including multiple SPL‐binding sites (Figures S9 and S10).

Overall, we observed a relatively good conservation of the role of the SPL genes from these two clades in the acceleration of flowering by the direct upregulation of SQUAMOSA genes (Figures S9 and S10), the promotion of miR172, and the repression of AP2L genes (Table 2).

Effect of the age pathway under non‐inductive conditions

In this study, we show that the effect of the transgenic expression of MIM156 is stronger in non‐vernalized winter plants than in spring plants (Figure 1D), with partially vernalized plants showing intermediate effects (Figure 3A). This result is consistent with the increased differences in heading time observed under non‐inductive conditions between winter Kronos and its sister lines overexpressing miR172 or carrying combined ap2l1 ap2l5 mutations (Debernardi et al., 2022). These results indicate that genes in the endogenous age pathway can modulate the vernalization response in winter wheat.

We have previously shown that in winter wheat the downregulation of AP2L1 and the upregulation of FT1 were decoupled from the age‐dependent downregulation of miR156, and that the induction of VRN1 contributed to the upregulation of miR172, the repression of AP2L1, and the activation of FT1 to promote flowering under LD (Debernardi et al., 2022). In this study, we show that these effects can be mediated by SQUAMOSA‐DELLA protein interactions competing with the repressive DELLA‐SPL3/4 interactions and resulting in more active SPL proteins (Figure 6). In addition, VRN1 or some of its downstream targets seem to have a positive effect in the transcriptional regulation of SPL13 (Figure 3L), suggesting that VRN1 effects on the age pathway are mediated by its transcriptional and post‐transcriptional regulation of SPL genes.

In non‐vernalized winter Kronos plants, the transcript levels of SPL3, SPL4, and SPL13 gradually increased, reaching levels higher than those observed in heading spring wheat plants, which likely contributed to the ability of winter Kronos to head after a long time at non‐vernalizing temperatures (Figure 3G–I). This hypothesis was supported by the accelerated heading of non‐vernalized winter plants overexpressing MIM156, which showed a significant transcriptional upregulation of SPL3, SPL4, and SPL13 (Figure S4i–k). These results support the idea that SPL genes can accelerate heading time of non‐vernalized winter wheat.

This study also showed that SPL alleles can accelerate heading time under non‐inductive SD conditions, and that the differences between plants carrying WT and resistant alleles rSPL4 or rSPL13 were larger under SD than under LD (Table 1; Figure S8). The stronger effect of rSPL13 on heading time under SD was also supported by field experiments in hexaploid wheat, in which rSPL13 accelerated heading time by 11 days under fall‐planting and by 5 days under spring planting (Gupta et al., 2023). Stronger differences in heading time were also reported between Ubi::miR172 and MIM172 transgenic Kronos plants under SD (40 days) than under LD (12 days) (Debernardi et al., 2022).

Taken together, the vernalization and photoperiod studies show that wheat genes of the age‐dependent pathway have larger effects on heading time under non‐inductive conditions than under inductive conditions. These results suggest that the age pathway may serve as a backup system to ensure wheat reproductive development under suboptimal inductive environments, as previously suggested in Arabidopsis (Wang et al., 2009).

Interactions between the age and the GA pathways modulate wheat heading time

In Arabidopsis, GA signals that promote flowering under non‐inductive SD are mediated by the SPL‐SOC1 module (Jung et al., 2012). GA degrades the DELLA proteins, disrupting the physical interactions with AtSPL9 and AtSPL15 (homologs of wheat SPL14 and SPL17) and AtSPL2/10/11 (homologs of wheat SPL3 and SPL4), restoring SPL activity (Hyun et al., 2016; Yu et al., 2012). In addition, post‐transcriptional and post‐translational regulation of AtSPL15 integrate the age and GA pathways at the SAM to promote Arabidopsis flowering under non‐inductive conditions (Hyun et al., 2016).

Previous studies have also revealed important roles of GA on reproductive development in diploid wheat Triticum monococcum. Diploid wheat accessions carrying the PS allele Ppd1b failed to head under SD and did not respond to the addition of exogenous GA. However, accessions carrying the Vrn1g or Vrn1f alleles with promoter mutations in a CArG box close to the transcriptional start site showed increased transcript levels of Vrn1 under SD (Dubcovsky et al., 2006) and responded to GA applications with accelerated spike development and stem elongation (Pearce et al., 2013). These effects were abolished by the addition of the GA inhibitor paclobutrazol, documenting the importance of GA in the promotion of spike development and stem elongation in wheat (Pearce et al., 2013). Under natural conditions, the longer days of spring induce FT1, triggering the upregulation of VRN1 and GA biosynthetic genes, which together promote spike development and stem elongation (Pearce et al., 2013).

The physical interactions between DELLA and both SPL3 and SPL4 proteins (Figures S12 and S13) revealed a conserved link between the GA pathway and the endogenous age pathway. The connection between these two pathways was also supported by significant genetic interactions between MIM156 and Rht1 alleles for the expression levels of flowering genes VRN1, FUL2, FT1, and VRN2 (Data S12). These significant interactions revealed stronger effects of MIM156 on the expression of flowering genes and on heading time in the presence of the GA‐sensitive allele, which were likely mediated by reduced DELLA levels and increased SPL activity. These interactions also resulted in stronger effects of the RHT1 alleles on gene expression and heading time in the presence of MIM156. As a result of these interactions, plants combining MIM156 and GA‐sensitive alleles had the highest levels of flowering‐promoting genes and the lowest levels of flowering‐repressing genes (Data S12), a result that is consistent with their earliest heading and fewest leaves (Data S11).

We speculate that the elevated levels of DELLA proteins in wheat cultivars carrying GA‐insensitive Rht1b alleles may exacerbate the repressive effect of DELLA on SPL proteins, and that this effect can be compensated by the induction of VRN1 and FUL2 proteins and their interactions with DELLA. The presence of the SQUAMOSA proteins may reduce the ability of DELLA to repress the SPL proteins and ameliorate DELLA's negative effect on heading time. VRN1's contribution to the transcriptional upregulation of SPL13 (Figure 3L), which does not physically interact with DELLA, may provide an alternative mechanism to offset the accumulation of DELLA in semidwarf wheats carrying the GA‐insensitive Rht1b alleles.

In summary, we propose that the competitive interactions between DELLA, SQUAMOSA, and SPL proteins interconnect the GA, age, and vernalization pathways, and that they can ameliorate the negative effects of the increased DELLA accumulation on heading time in wheats carrying GA‐insensitive Rht1b alleles.

CONCLUSION AND PRACTICAL APPLICATIONS

Wheat is a young polyploid species with high levels of gene redundancy that mask the effects of most recessive mutations (Uauy et al., 2017). Therefore, the dominant rSPL mutations are particularly attractive for fine‐tuning heading time in polyploid wheats. In spring wheats like Kronos, the individual rSPL3, rSPL4, and rSPL13 alleles accelerate heading time by 1–3 days, but different alleles can be combined to accelerate heading time by 4–6 days. In winter wheats and fall‐planted spring wheats, which face a longer period of non‐inductive conditions, the differences are expected to be 1 or 2 days larger than in spring planted wheats. An interesting application of the rSPL alleles is to reduce the vernalization requirement in regions where winters become milder due to climate change.

The non‐transgenic nature of the rSPL3 and rSPL4 mutants makes them particularly attractive for wheat breeding applications because they are not constrained by governmental regulations or CRISPR intellectual property protections that could limit or delay their deployment. To facilitate the adoption of these genetic resources, we deposited the five individual resistant alleles in GRIN Global and made them immediately available without any restrictions (see accession numbers under Data Availability).

MATERIALS AND METHODS

The phylogenetic analysis of wheat (T. aestivum), rice (Oryza sativa), and Arabidopsis SPL proteins based on their SBP domains is described in Method S1. Total RNA and small RNA extraction and purification, as well as qRT‐PCR quantification protocols, are described in Method S2. Plant materials and mutants used in this study, and their growing conditions are described in Method S3. The methods used to develop the SPL13 CRISPR mutants and the transgenic lines overexpressing this gene are described in Method S4. The methods used to characterize DNA‐protein interactions included Method S5 for Y1H and Method S6 for electrophoretic mobility shift assays. Methods for protein–protein interaction, including Y2H, BiFC, Co‐IP, and Y3H, are described in Methods S7–S10, respectively. Statistical methods are summarized in Method S11.

AUTHOR CONTRIBUTIONS

QL, LZ, and JMD performed most of the experimental work. ZZ and CZ contributed to the partial vernalization experiments. CL contributed to the Y2H and Y3H experiments. QL, LZ, and JMD generated the first draft of the manuscript that was revised by all coauthors. JMD and JD developed and directed the project and supervised students and postdocs. JD performed statistical analyses, secured funding, administered the project, and was responsible for the final version of the manuscript.

CONFLICT OF INTEREST

The authors declare that they do not have any conflicts of interest.

Supporting information

Figure S1. miR156 binding sites in nine wheat SPL genes and their expression profiles. Figure S2. Alignment of SBP domains for Arabidopsis, rice, and wheat SPL proteins. Figure S3. Phylogenetic tree of SPL proteins based on the SBP domain alignment. Figure S4. Regulation of SPL3, SPL4, SPL13, and SPL2 expression by miR156. Figure S5. Expression of wild‐type and resistant alleles rSPL3, rSPL4, and rSPL13. Figure S6. Pictures of wild‐type, rSPL3, rSPL4, and rSPL13 mutants at heading time. Figure S7. Effect of rSPL3, rSPL4, rSPL13 on leaf number. Figure S8. Effect of rSPL3, rSPL4, rSPL13 on heading time under LD and SD. Figure S9. Yeast‐one‐hybrid (Y1H) assays. Figure S10. Electrophoretic mobility shift assay (EMSA). Figure S11. Interactions between DELLA and SPL proteins. Figure S12. Bimolecular fluorescent complementation (BiFC or split YFP). Figure S13. Co‐IP experiments showing interactions between DELLA, and both SPL3 and SPL4 proteins. Figure S14. Effects of miR156 and RHT‐B1 alleles on gene expression.

Methods S1. Phylogenetic analysis of SPL family. Method S2. Expression studies by qRT‐PCR. Method S3. Plant materials, mutants, and growth conditions. Method S4. Transgenic plants and wheat transformation. Method S5. Yeast‐one‐hybrid (Y1H) assays. Method S6. Electrophoretic mobility shift assays (EMSA). Method S7. Yeast‐two‐hybrid (Y2H) assays. Method S8. Bimolecular fluorescence complementation (BiFC). Method S9. Co‐IP assays. Method S10. Yeast‐three‐hybrid (Y3H) assays. Method S11. Statistical analyses.

Data S1. Wheat SPL nomenclature, gene IDs, and homology with rice SPLs. Data S2. Primers used in this study. Data S3. Effect of age, miR156, and mutations in miR156 binding site on SPLs expression. Data S4. Effect of rSPL alleles on heading time under long days. Data S5. Effect of single and combined rSPL3, rSPL4, and rSPL13 alleles on LN under LD. Data S6. Effect of Ubi::rSPL‐A13HA on gene expression, heading time, and leaf number. Data S7. Effect of rSPL alleles on heading time under long and short‐day conditions. Data S8. Leaf number in plants with single and combined rSPL alleles. Data S9. Effect of rSPL alleles on the expression of flowering genes. Data S10. Effect of age and vernalization on heading time and expression of flowering genes. Data S11. Interactions between MIM156 and DELLA for plant height, DTH and LN. Data S12. Interactions between MIM156 and DELLA for expression of flowering genes. Data S13. Yeast three‐hybrid (Y3H) assays.