Terminal Shoot Maturity Orchestrates LcSPLs–LcFT1 Signaling to Dictate Low-Temperature Floral Induction in Litchi (Litchi chinensis Sonn.)

1. Introduction

Litchi (Litchi chinensis Sonn.), a prominent evergreen fruit tree species of the family Sapindaceae, has a long history of cultivation in tropical and subtropical regions [1,2]. However, the sustainable development of the litchi industry is severely hampered by alternate bearing, a phenomenon primarily driven by poor or erratic floral induction [3,4,5]. As a species characterized by terminal flowering, litchi undergoes floral bud differentiation in response to a complex interplay of environmental and physiological factors [2]. Although environmental cues particularly chilling exposure during winter are prerequisite for floral induction, the physiological status of the bearing mother shoots serves as the controllable core determining flowering success [6,7,8]. Field observations consistently demonstrate that trees with immature young flushes or ungreened red leaves prior to winter chilling fail to initiate floral buds, ultimately leading to yield loss.

Shoot maturity provides both the structural and metabolic foundation for successful floral induction. Carbohydrate accumulation, particularly the dynamics of starch and soluble sugars in leaves and small twigs, is indispensable for the emergence of panicle primordia (traditionally recognized as the "whitish millet" tage) [8,9,10]. Immature leaves exhibit insufficient carbon allocation and an inability to perceive or transduce low-temperature signals. Nevertheless, precise physiological criteria for defining shoot maturity remain lacking, and the molecular mechanism explaining why young leaves inhibit floral induction remains a long-standing unresolved question in litchi biology.

At the molecular level, plant floral transition is regulated by an integrated genetic network. The SQUAMOSA PROMOTER BINDING PROTEIN-LIKE (SPL) transcription factor family, regulated by miR156, plays a pivotal role in the age-dependent pathway governing vegetative-to-reproductive phase change[11,12]. Concurrently, upon sensing chilling signals, florigen (FT, FLOWERING LOCUS T) synthesized in mature leaves translocates long-distance through the phloem to the shoot apical meristem, where it interacts with FD to activate downstream floral meristem identity genes such as AP1 and LFY [13,14,15]. Although several key flowering related genes including LcFT, LcFLC, LcLFY, and LcSPLs have been identified in litchi [8,16], how LcSPL genes modulate the floral responsiveness of terminal shoots at distinct maturity stages remains poorly understood.

To address these knowledge gaps, this study systematically analyzed the expression patterns of key floral induction genes, with a primary focus on the LcSPL family, in leaves of the latest shoot flush across different maturity stages. The findings aim to elucidate the physiological and molecular mechanisms by which shoot maturity governs floral transition, thereby providing theoretical insights and practical strategies for mitigating alternate bearing and achieving precise flowering management in woody perennial fruit crops.

2. Materials and Methods

2.1. Plant Materials and Treatments

To evaluate low-temperature-induced flowering, uniform and vigorous three-year-old air-layered 'Guiwei' litchi potted trees were selected. Four trees with the terminal flush at the expanding stage and four at the mature stage were subjected to low-temperature treatment (15/10°C, day/night) in a temperature-controlled growth room for 60 days, followed by transfer to ambient conditions (25/20°C, day/night) to promote floral initiation. Each tree served as an independent biological replicate using a randomized sampling design. Leaf samples from the terminal flush were collected at seven time points: 0, 1, 7, 20, 40, and 60 days during low-temperature treatment, as well as at the "whitish millet" stage. Collected leaves were immediately flash-frozen in liquid nitrogen and stored at -80°C for subsequent assays.

For the simulated low-temperature induction experiment on juvenile plants, 20 uniform and vigorous one-year-old litchi seedlings were transferred into an artificial climate chamber and maintained under low temperature (15/10°C, day/night) for 60 days, before being moved to ambient temperature (25/20°C, day/night). Leaves of two distinct maturity levels—apical tender leaves and basal older leaves were harvested at seven time points: 0, 1, 7, 20, 40, 60, and 68 days after treatment initiation. Sample collection and preservation procedures strictly followed the protocol described for adult potted trees.

2.2. Floral Biology Statistics of Litchi

The emergence time of the "whitish millet" stage, percentage of flowering shoots, percentage of pure flowering shoots, as well as inflorescence length and width were recorded for terminal shoots at different maturity stages.

2.3. Quantification of Leaf Carbohydrates

Soluble sugar profiles in litchi leaves were determined following a protocol adapted from Yang [17]. Briefly, cryogenic liquid nitrogen was applied to homogenize frozen leaf samples. 0.3 g portion of the pulverized tissue was suspended in 6 mL of 90% (v/v) ethanol, followed by a 20-min thermal incubation in an 80°C water bath. The resulting slurry was centrifuged at 4,000 rpm for 10 min under ambient conditions. The remaining residue underwent an identical re-extraction step. Supernatants from both extractions were pooled and concentrated to dryness using a rotary evaporator (RapidVap, Labconco, Kansas City, MO, USA).

The dried extract was reconstituted in double-distilled water, then clarified via centrifugation at 13,000 × g for 10 min. To eliminate impurities, the supernatant was passed through a Sep-Pak® C18 cartridge (1 cc/100 mg; Waters Corp., Milford, MA, USA). Sugar separation and quantification were carried out on an Agilent 1200 HPLC system (Agilent Technologies, Waldbronn, Germany) integrated with a Transgenomic CARB Sep Coregel 87C column accompanied by a matching guard cartridge. The analytical column and the refractive index detector were held at 80°C and 40°C, respectively. Chromatographic runs were executed with an injection volume of 10 μL using deionized water as the mobile phase at a constant flow rate of 0.4 mL min⁻¹. Target sugars were identified by aligning retention times with authentic standards and quantified against corresponding calibration curves. For each time point, all litchi tissue samples were analyzed in at least three independent biological replicates.

2.4. Real-Time PCR Analysis

Genes associated with energy metabolism were retrieved from the litchi genome database (http://121.37.229.61:82/). Specific primer pairs were synthesized using Primer 5.0 software (Supplementary Table 1). Quantitative real-time PCR was performed on an ABI 7500 Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). The thermocycling conditions comprised an initial denaturation at 95°C for 5 min, followed by 40 cycles of denaturation at 95°C for 10 s, annealing at 55°C for 30 s, and elongation at 72°C for 30 s, with a final melting curve analysis to confirm reaction specificity. All reactions were executed in at least three independent biological replicates, with data presented as the mean ± standard error (SE). Target gene expression levels were normalized to LcActin (GenBank accession no. HQ615689), a well-validated reference gene exhibiting stable expression across litchi cultivars and under abiotic stresses [18]. Relative transcript abundances were quantified using the 2^–∆∆CT method [19].

2.5. Dual-Luciferase Reporter Assay

The transcriptional activity of LcSPLs on the LcFT1 promoter was evaluated using a dual-luciferase system according to Hellens [20]. The CDS of LcSPLs and LcFT1 promoter were cloned into pGreen II 0029 62-SK (effector) and pGreen II 0800-LUC (reporter) vectors, respectively. Agrobacterium tumefaciens (GV3101) cells carrying the constructs were prepared in induction buffer (10 mM MES, 10 mM MgCl₂, 200 μM acetosyringone, pH 5.6) to OD₆₀₀ = 0.8. Mixtures of effector and reporter strains were co-infiltrated into leaves of 3-week-old Nicotiana benthamiana, with empty effector vector co-infiltrations as negative controls. After 3 days of growth at 23°C, relative luciferase activities (LUC/REN) were measured using a GloMax 20/20 Luminometer (Promega, USA). Assays included three biological replicates and were repeated twice independently.

2.6. Ectopic Overexpression of Litchi Genes in Arabidopsis

To generate overexpression constructs, the full-length coding sequences (CDS) of LcSPL1/3/9/10 were inserted into the pCAMBIA1302 vector via homologous recombination. The validated recombinant plasmids were subsequently transformed into Agrobacterium tumefaciens strain GV3100. Arabidopsis thaliana ecotype Columbia-0 (Col-0) plants were transformed using the floral dip method. Transgenic T1 seeds were surface-sterilized and selected on Murashige and Skoog (MS) agar medium supplemented with hygromycin. Seven-day-old resistant seedlings were transplanted into a soil mixture (nutrient soil:vermiculite:perlite = 4:1:1, v/v/v) and maintained in a growth chamber at 23°C under long-day conditions.

2.7. Data Processing and Statistical Analysis

Experimental data were processed using Microsoft Excel, and figures were generated with Origin 2021 software (OriginLab Corp., Northampton, MA, USA). Biological flowering parameters were subjected to one-way analysis of variance (ANOVA). Differences among treatment means were determined using Duncan’s multiple range test at a significance level of P < 0.05 via SPSS 21.0 (IBM Corp., Armonk, NY, USA). Linear relationships between variables were evaluated using Pearson’s correlation analysis (P < 0.05).

3. Results

3.1. Terminal Shoot Maturity Influences Litchi Floral Induction

Litchi inflorescences originate from terminal shoot apical meristems (SAMs), which also govern recurrent vegetative flushes. To evaluate how terminal shoot maturity conditions floral development, potted litchi trees possessing terminal flushes at either the leaf-expanding or mature stage (Figure 1A, D) were exposed to a chilling regime (15/10 °C, 12 h light/12 h dark). Following 60 days of cold induction, the trees were returned to ambient temperature (25 °C). Shoots that were mature prior to treatment achieved floral initiation within 8 days (Figure 1 E) and ultimately developed full panicles (Figure 1 F). Conversely, shoots treated during the expansion stage failed to initiate flowering (Figure 1 B, C).

To evaluate the influence of shoot maturity on litchi floral induction, the flowering characteristics of terminal shoots at the expanding and mature stages were quantified following low-temperature treatment (Table 1). The time required to reach the "whitish millet" stage showed no statistically significant difference between the expanding and mature. However, shoot maturity markedly enhanced both flowering efficiency and inflorescence morphology. The flowering rate of mature terminal shoots reached 84.62%, which was significantly higher than that of expanding shoots 14.51%. Notably, while expanding shoots failed to form pure flowering shoots, mature shoots exhibited a pure flowering rate of 62.22% (P < 0.01). Furthermore, the inflorescence dimensions were significantly superior in mature shoots compared to expanding ones, displaying nearly double the length and more than twice the width.

3.2. Dynamic Changes in Sugar Accumulation During Low-Temperature Floral Induction

To clarify the metabolic differences between shoot maturity levels during floral induction, dynamic changes in primary soluble sugars (sucrose, glucose, fructose, quebrachitol and total sugar) were monitored across a 68-day chilling period (Figure 2). Overall, mature terminal shoots maintained significantly higher levels of sugar accumulation compared to expanding shoots throughout the low-temperature treatment. Sucrose was the primary soluble sugar, in mature shoots, sucrose and total sugars peaked rapidly at 7 d, remaining significantly higher than in expanding shoots from 7 d to 60 d (P < 0.05). By 68 d, both groups declined to similarly low levels. Glucose in mature shoots peaked at 7 d and significantly exceeded expanding shoots at 20 d and 40 d. Fructose remained generally low, though mature shoots showed significantly lower levels than expanding shoots at 40 d.

3.3. Expression Analysis of LcTPS Genes in Leaves of the Terminal Shoot at Different Maturity

To evaluate the involvement of trehalose-6-phosphate signaling in litchi floral transition, temporal expression profiles of nine LcTPS family members were tracked in mature and expanding terminal shoots during chilling treatment (Figure 3). Distinct transcription kinetics were observed between the two maturity stages. In mature shoots, LcTPS1, LcTPS2, and LcTPS5 functioned as early responders, undergoing rapid up-regulation within 1 d of chilling and maintaining significantly higher expression levels than in expanding shoots throughout cold exposure (P < 0.05). Likewise, LcTPS3 and LcTPS4 displayed persistently elevated transcript abundance in mature shoots across the entire 68-day induction period, whereas LcTPS3 remained near baseline levels in expanding shoots. Conversely, LcTPS7 was enriched in expanding shoots, showing significantly higher expression at 20, 40, and 68 d, while LcTPS8 maintained comparable transcript levels across maturity stages. Crucially, at the onset of the "whitish millet" stage, a vital physiological marker of morphological flower bud differentiation, mature shoots exhibited significantly elevated transcript levels of LcTPS1, LcTPS3, and LcTPS4 relative to expanding shoots. This preferential accumulation of LcTPS1, LcTPS3, and LcTPS4 transcripts in mature shoots closely aligns with their high flowering competence, indicating that the sustained activation of these core genes serves as a pivotal molecular driver promoting litchi floral induction.

3.4. Expression Analysis of LcSPL Genes in Leaves of the Terminal Shoot at Different Maturity

To evaluate the involvement of the microRNA156/SQUAMOSA PROMOTER BINDING PROTEIN-LIKE module in litchi flowering, temporal expression profiles of LcSPL family members were tracked in mature and expanding terminal shoots across the chilling period up to the "whitish millet" stage (Figure 4). Crucially, key flower-promoting genes LcSPL1, LcSPL3, LcSPL4, and LcSPL10, exhibited significantly higher transcript abundance in mature leaves compared to expanding ones at the critical "whitish millet" stage (P < 0.05). At this physiological landmark marking the onset of flower bud differentiation, these four genes were strongly up-regulated and reached peak or near-peak levels in mature shoots; specifically, LcSPL4 and LcSPL10 maintained elevated baseline expression throughout mid-to-late chilling, whereas LcSPL1 and LcSPL3 displayed a distinct late-stage accumulation precisely coinciding with "whitish millet" formation. This preferential enrichment of LcSPL1/3/4/10 in mature leaves strongly correlates with their high flowering competence, indicating that up-regulation of these specific LcSPL members at the "whitish millet" stage serves as a key molecular driver promoting litchi floral transition. In contrast, LcSPL6, LcSPL9, LcSPL13, LcSPL14, and LcSPL17 maintained substantially higher expression in expanding shoots across nearly all sampled time points, with prominent transcript peaks during early chilling, while LcSPL12 displayed a stage-specific shift from higher early expression in expanding shoots to significantly elevated levels in mature shoots during later stages, including at the "whitish millet" phase.

3.5. Expression Analysis of LcSPL Genes in Leaves of Seedling Plants at Different Maturity

Under simulated low-temperature floral induction conditions, members of the LcSPL gene family displayed distinct leaf-maturity-dependent expression dynamics in litchi seedlings (Figure 5). Crucially, LcSPL1, LcSPL4, and LcSPL6 maintained significantly higher transcript abundance in secondary older leaves than in tender young leaves during mid-to-late chilling, with LcSPL1 and LcSPL4 exhibiting peak expression at 20 d in older leaves. Conversely, LcSPL3, LcSPL10, and LcSPL13 showed sharp, transient expression peaks in tender young leaves at 0 d prior to cold exposure, followed by rapid down-regulation upon entering simulated chilling conditions. Meanwhile, LcSPL12 and LcSPL17 were progressively induced in both leaf types during late chilling, reaching maximum transcript levels at 60 d and 20 d, respectively (P < 0.05).

3.6. Expression Analysis of Key Genes Involved in Floral Induction in Leaves of the Terminal Shoot at Different Maturity

To elucidate the florigen-mediated regulation of litchi floral induction, the temporal expression profiles of LcFT1, LcFT2, LcTFL1, and LcFD were monitored in mature and expanding terminal shoots throughout chilling up to the "whitish millet" stage (Figure 6). Crucially, LcFT1 emerged as the primary molecular driver of litchi floral initiation. While both LcFT1 and LcFT2 remained at baseline levels during early chilling, LcFT1 exhibited a massive, predominant up-regulation in mature shoots starting at 20 d and reached an unprecedented peak at 60 d. Throughout mid-to-late chilling 20-68 d, LcFT1 transcript abundance in mature leaves was extraordinarily higher than in expanding shoots (P < 0.05), where it peaked at only approximately 10 fold; even at the "whitish millet" stage, LcFT1 maintained exceptionally elevated expression in mature shoots, directly aligning with their superior flowering competence. Although LcFT2 displayed a similar dynamic trend, peaking at over 1400 fold at 60 d, its expression magnitude remained substantially lower than that of LcFT1, while the florigen interaction partner LcFD maintained relatively stable expression overall, with significantly higher levels in mature shoots at 20 d and 60 d. In stark contrast to LcFT1, the floral repressor LcTFL1 remained persistently high in expanding shoots across the entire 68 d period while being almost completely suppressed in mature shoots, further emphasizing that the dramatic transcript enrichment of LcFT1 over LcTFL1 in mature leaves provides the decisive florigenic signal governing litchi floral induction.

3.7. Transcriptional Activation and Repression of the LcFT1 Promoter by LcSPL Transcription Factors

To evaluate the transcriptional regulation of the core florigen gene LcFT1 by LcSPL transcription factors, a dual-luciferase reporter assay was performed in Nicotiana benthamiana leaves (Figure 7). Crucially, LcSPL1, LcSPL3, and LcSPL10 operated as key transcriptional activators of LcFT1, with LcSPL1 driving the highest transactivation level (P < 0.05), followed by LcSPL3 and LcSPL10. Conversely, LcSPL9 exhibited a significant repressive effect on the LcFT1 promoter, suppressing the relative LUC/REN ratio down to 0.63). Other evaluated members (LcSPL4, 6, 12, 13, 14, and 17) showed no significant regulatory effects relative to the control. These results demonstrate that LcSPL1, LcSPL3, and LcSPL10 promote litchi flowering through direct transcriptional activation of LcFT1, whereas LcSPL9 functions as a potential transcriptional repressor.

3.8. Effects of LcSPL1/3/4/9/10 on Flowering in Arabidopsis

To evaluate the functional roles of LcSPL1, LcSPL3, LcSPL9, and LcSPL10 in regulating flowering time, T2 transgenic Arabidopsis plants overexpressing these individual genes (35S::LcSPL) were analyzed alongside wild-type (WT, Col-0) controls (Table 3). Ectopic expression of LcSPL1, LcSPL3, and LcSPL10 significantly accelerated the floral transition, with 35S::LcSPL1, 35S::LcSPL3, and 35S::LcSPL10 lines flowering markedly earlier than WT plants. Consistently, these transgenic lines produced significantly fewer rosette leaves at flowering compared to WT, confirming that ectopic expression of LcSPL1, LcSPL3, or LcSPL10 confers a distinct early-flowering phenotype in Arabidopsis. In contrast, 35S::LcSPL9 plants exhibited no significant differences in flowering time and rosette leaf number at flowering relative to WT (P > 0.05).

Visual phenotypic observation confirmed marked differences in flowering time among the distinct overexpression lines (Figure 3). Ectopic expression of LcSPL1 (Figure 3A), LcSPL3 (Figure 3B), and LcSPL10 (Figure 3D) consistently conferred a distinct early-flowering phenotype; while wild-type (WT) controls remained strictly in the vegetative rosette stage without visible inflorescence bolting, the 35S::LcSPL1, 35S::LcSPL3, and 35S::LcSPL10 transgenic lines had already bolted, elongated main floral stems, formed open flower buds, and initiated this transition with notably fewer rosette leaves. In contrast, 35S::LcSPL9 overexpressing plants (Figure 3C) exhibited no observable difference in growth stage or flowering time relative to WT, remaining in the vegetative rosette phase without signs of premature bolting.

4. Discussion

Floral induction in perennial fruit trees is a pivotal developmental transition governed by a complex interplay between seasonal environmental signals and internal physiological competency [21,22,23,24]. In litchi, erratic floral initiation and alternate bearing pose severe challenges to sustainable production. Although winter chilling is a prerequisite environmental trigger, field management often encounters a major bottleneck: trees exhibiting immature terminal shoots during the chilling window fail to initiate flower buds (Figure 1), leading to vegetative flush emergence instead [2,25,26,27]. Despite its horticultural significance, the physiological and molecular framework explaining why shoot maturity conditions low-temperature responsiveness has remained elusive.

In this study, by integrating physiological assays, metabolic profiling, qRT-PCR expression dynamics, dual-luciferase reporter assays, and ectopic Arabidopsis transformation, we established that terminal shoot maturity orchestrates carbohydrate accumulation, T6P signaling, and an LcSPL1/3/10–LcFT1 transcriptional cascade to overcome LcTFL1 mediated floral repression, thereby conferring floral competence.

Carbohydrates act as essential energy sources and primary signals during floral transition [28,29,30]. Mature terminal shoots maintained significantly higher sucrose and total sugar levels than expanding shoots throughout chilling, establishing a high-energy state (Figure 2). Expanding shoots, acting as active carbon sinks, failed to accumulate equivalent reserves [8]. TPS acts as a primary sucrose sensor that regulates flowering [31,32,33,34]. LcTPS1, LcTPS2, and LcTPS5 acted as rapid cold-responders in mature leaves, maintaining elevated expression throughout chilling. Crucially, LcTPS1, LcTPS3, and LcTPS4 were strongly enriched in mature leaves at the "whitish millet" stage (Figure 3). This maturity-dependent activation of the T6P pathway directly aligns with the high flowering rate in mature shoots compared to expanding shoots, proving that carbohydrate saturation provides the metabolic permissive signal for floral transition.

Expression analysis revealed distinct functional diversification among LcSPL genes based on shoot maturity [8,16,35]. LcSPL1, LcSPL3, LcSPL4, and LcSPL10 were preferentially enriched in mature leaves of adult trees and older seedling leaves, reaching peak expression at the "whitish millet" stage (Figure 4). Conversely, LcSPL6/9/13/14/17 were predominantly expressed in expanding shoots and young seedling leaves (Figure 5). Ectopic overexpression in Arabidopsis confirmed these distinct roles: 35S::LcSPL1/3/10 transgenic lines displayed pronounced early-flowering phenotypes with reduced rosette leaf numbers, whereas 35S::LcSPL9 lines showed no acceleration of flowering (Figure 8). These results establish LcSPL1/3/10 as direct genetic promoters of phase change requiring shoot maturity for induction.

FT and TFL1 balance determines SAM fate during chilling [36,37,38,39,40]. In mature leaves, LcFT1 transcript levels surged dramatically under cold exposure and remaining high at the "whitish millet" stage, whereas expanding shoots exhibited lower LcFT1 induction alongside persistent elevation of the repressor LcTFL1 (Figure 6). Dual-luciferase reporter assays provided the direct mechanism linking maturity to florigen activation: LcSPL1/3/10 directly bound to and significantly activated the LcFT1 promoter (Figure 7), with LcSPL1 exerting the strongest transactivation, whereas LcSPL9 repressed LcFT1[8,41].

In summary, expanding shoots remain in a vegetative state due to low sugar reserves, low LcTPS1/3/4 expression, and high LcTFL1 and LcSPL9 levels. Shoot maturation triggers sucrose accumulation and T6P signaling, upregulating LcSPL1/3/10. These LcSPLs directly transactivate LcFT1 expression, shifting the LcFT1/LcTFL1 ratio to promote panicle initiation in response to winter chilling

5. Conclusions

In conclusion, this study provides pivotal physiological and molecular insights into how terminal shoot maturity dictates low-temperature floral induction in litchi. We demonstrate that immature terminal shoots represent a major physiological barrier to floral initiation due to deficient carbohydrate accumulation, impaired T6P signaling, and unfavorable transcriptional outputs. Specifically, terminal shoot maturation establishes a essential metabolic foundation characterized by the accumulation of soluble sugars and the sustained activation of LcTPS1/3/4 genes. At the genetic level, mature shoots selectively enrich flower-promoting transcription factors, among which LcSPL1, LcSPL3, and LcSPL10 act as critical positive regulators. Mechanistically, these LcSPLs directly bind to and transactivate the promoter of the primary florigen gene LcFT1, effectively tipping the LcFT1/LcTFL1 balance toward reproductive transition at the shoot apical meristem. In contrast, immature shoots maintain elevated expression of the floral repressor LcTFL1 and the transcriptional suppressor LcSPL9, suppressing LcFT1 transactivation and perpetuating vegetative growth even under favorable chilling conditions.

Overall, our findings uncover a integrated "Metabolism–LcSPLs–LcFT1" regulatory network by which shoot maturity conditions environmental responsiveness. These theoretical advancements not only resolve a long-standing question in litchi floral biology, but also offer practical guidance for horticultural practices aimed at accelerating terminal shoot maturation to mitigate alternate bearing and achieve reliable yield in perennial fruit crops.

Based on our physiological, biochemical, and molecular findings, we propose a hypothetical working model to illustrate how terminal shoot maturity conditions low-temperature responsiveness and dictates floral induction in litchi (Figure 9). In mature litchi leaves, winter chilling induces substantial sucrose accumulation and sustained transcription of LcTPS1/3/4, activating T6P signaling to establish a high-energy metabolic status. This physiological state selectively upregulates the flowering-promoting transcription factors LcSPL1, LcSPL3, and LcSPL10, which directly transactivate the promoter of the core florigen gene LcFT1 while repressing the immature shoot-enriched suppressor LcSPL9. The resulting systemic LcFT1 signal translocates to the SAM, tipping the LcFT1/LcTFL1 transcriptional balance to drive SAM morphological reorganization, form the "whitish millet" stage, and ultimately enable robust panicle emergence under subsequent ambient temperatures.