ELONGATED HYPOCOTYL 5a modulates FLOWERING LOCUS T2 and gibberellin levels to control dormancy and bud break in poplar

Abstract

Photoperiod is a crucial environmental cue for phenological responses, including growth cessation and winter dormancy in perennial woody plants. Two regulatory modules within the photoperiod pathway explain bud dormancy induction in poplar (Populus spp.): the circadian oscillator LATE ELONGATED HYPOCOTYL 2 (LHY2) and GIGANTEA-like genes (GIs) both regulate the key target for winter dormancy induction FLOWERING LOCUS T2 (FT2). However, modification of LHY2 and GIs cannot completely prevent growth cessation and bud set under short-day (SD) conditions, indicating that additional regulatory modules are likely involved. We identified PtoHY5a, an orthologs of the photomorphogenesis regulatory factor ELONGATED HYPOCOTYL 5 (HY5) in poplar (Populus tomentosa), that directly activates PtoFT2 expression and represses the circadian oscillation of LHY2, indirectly activating PtoFT2 expression. Thus, PtoHY5a suppresses SD-induced growth cessation and bud set. Accordingly, PtoHY5a knockout facilitates dormancy induction. PtoHY5a also inhibits bud-break in poplar by controlling gibberellic acid (GA) levels in apical buds. Additionally, PtoHY5a regulates the photoperiodic control of seasonal growth downstream of phytochrome PHYB2. Thus, PtoHY5a modulates seasonal growth in poplar by regulating the PtoPHYB2–PtoHY5a–PtoFT2 module to determine the onset of winter dormancy, and by fine-tuning GA levels to control bud-break.

ELONGATED HYPOCOTYL 5a regulates FLOWERING LOCUS T2, a key target of poplar dormancy induction, and mediates bud break by negatively regulating gibberellic acid levels in apical buds.

IN A NUTSHELL.

Background: Winter dormancy, a process where growth stops and plants acquire a dormant and freezing-tolerant state, is crucial for the survival of woody plants in boreal and temperate regions. The seasonal growth of perennial woody plants is controlled by dormancy entry and release, which are regulated by complex and precise mechanisms influenced by environmental cues. Dormancy entry is mainly induced by short photoperiods and is characterized by growth cessation and bud set. The central activator of flowering, FLOWERING LOCUS T (FT), is also the key target for winter dormancy induction. To satisfy distinct needs in dormancy or flowering, woody plants likely use different pathways than herbaceous plants use to regulate FT. Although a LATE ELONGATED HYPOCOTYL (LHY)–FT regulatory pathway analogous to that in herbaceous plants exists in woody plants, its efficacy in preventing dormancy in trees is limited.

Question: Do other regulatory mechanisms regulate winter dormancy in trees?

Findings: We determined that PtoHY5a, a poplar (Populus tomentosa) ortholog of Arabidopsis (Arabidopsis thaliana) ELONGATED HYPOCOTYL5 (HY5), mediates dormancy induction downstream of photoreceptor PHYB signaling. PtoHY5a directly and indirectly regulates the key target for winter dormancy induction, FT2, via two transcriptional modules. Moreover, PtoHY5a controls gibberellic acid levels in apical buds to inhibit bud break in poplar.

Next steps: The next step would be to explore PtoHY5a as a key target for genetic modification in trees. Genetic modification of PtoHY5a could be particularly useful when introducing trees to different latitudes; for instance, from northern regions to high-altitude southern areas, such as from the Tibetan Plateau to the South Asian Highlands, or vice versa. By understanding and manipulating PtoHY5a, we could potentially optimize tree adaptation to new environments, thereby enhancing their survival and productivity.

Introduction

The seasonal cycle of active growth and dormancy is a distinct feature of perennial plants in boreal and temperate regions, and winter dormancy represents one of the most basic adaptations to cope with cold environments (Cooke et al. 2012). When the trees perceive the environmental cues that herald the advent of winter, mainly daytime shortening below critical daylength, plants enter dormancy by ceasing vegetative growth and setting buds (Rohde et al. 2011; Singh et al. 2017). Trees have evolved a complex regulatory network for the perception of photoperiod signals, involving phytochrome photoreceptors, circadian clocks, and a systemic effector.

In poplar (Populus spp.) trees, there are three genes encoding phytochromes, PHYA, PHYB1, and PHYB2. Overexpression of PHYA prevented growth cessation in response to short-days (SDs), and the downregulation of PHYA resulted in earlier growth cessation and bud set in SD conditions (Kozarewa et al. 2010). It was suggested that PHYB2 is the associated quantitative trait locus controlling growth cessation and bud set in poplar (Frewen et al. 2000), and knockout of PHYB2 triggered earlier SD-induced growth cessation and bud set (Ding et al. 2021). PHYB1 knockout plants behave like the wild type (WT). These findings indicate that phytochromes have important roles in the photoperiod-regulated seasonal growth of poplar. The perception of light signals by photoreceptors also entrains the circadian clock, the molecular timer that allows plant metabolism to adjust to the 24-h daily cycle. As in Arabidopsis (Arabidopsis thaliana), the central clock homologs in poplar trees consist of LATE ELONGATED HYPOCOTYL1 (PtLHY1), PtLHY2, and TIMING OF CAB EXPRESSION1 (PtTOC1) (Takata et al. 2009; Ibáñez et al. 2010). PHYA downregulation in poplar modified the amplitude and phase of LHY expression, linking the phytochrome with the circadian clock (Kozarewa et al. 2010). Furthermore, the downregulation of LHY and TOC1 altered the phase and period of clock-driven gene expression and delayed growth cessation (Ibáñez et al. 2010). In particular, Ramos-Sánchez et al. (2019) observed that poplar LHY2 plays an important role in transmitting SD information for determining dormancy induction (Ramos-Sánchez et al. 2019). However, knockout of PHYB did not affect the expression of the central circadian clock genes under either long or SD conditions (Ding et al. 2021), suggesting the existence of another pathway independent of central circadian clock for SD signaling to induce tree dormancy.

Böhlenius et al. (2006) showed that the downregulation of the poplar ortholog of Arabidopsis FLOWERING LOCUS T (FT) triggered SD-induced dormancy in hybrid aspen (Populus tremula × tremuloides) (Böhlenius et al. 2006). FT2, one of the two closely related FT orthologs, has been shown to be the key target of SD signals in winter dormancy induction (Hsu et al. 2011). Accordingly, FT2 overexpression prevents the induction of growth cessation and bud set by SDs in hybrid aspen. In contrast, FT2 RNAi poplar lines showed premature growth cessation and bud set (Böhlenius et al. 2006; André et al. 2022). Some key factors for the downstream signaling of the FT2 have been identified in poplar, such as a tree ortholog of the floral meristem identity gene APETALA1 (AP1), Like-AP1 (LAP1), and AINTEGUMENTA-like 1 (AIL1) at the apical meristem, which regulate the expression of core cell cycle genes such as D-cyclins, leading to growth activation or cessation (Karlberg et al. 2011; Azeez et al. 2014).

To date, several pathways have been identified to regulate FT2 expression and dormancy induction. As in Arabidopsis, poplar CONSTANS (CO) is an upstream regulator of FT, and its downregulation suppresses the expression of FT, which leads to more sensitive to the SD signal, demonstrating an important role for the CO-FT2 module in dormancy induction (Böhlenius et al. 2006; Ding et al. 2016). Second, LHY2 inhibited the expression of FT2 by binding to the FT2 promoter, thereby regulating photoperiodic growth and dormancy through the LHY2–FT2 module (Ramos-Sánchez et al. 2019). Third, GIGANTEA (GI)-like genes (GIs) mediated seasonal growth cessation by directly regulating FT2 through a CO-independent pathway in Populus (Ding et al. 2018). Downregulation of LHY2 or overexpression of GIs delayed SD-induced dormancy to some extent.

Additionally, gibberellic acid (GA) plays a significant role in the photoperiodic regulation of growth cessation. Upon exposure to SD conditions, the levels of bioactive GAs decreased rapidly (Junttila and Jensen 1988). However, growth cessation in hybrid aspens under SD conditions can be delayed by the application of exogenous bioactive GAs or the overexpression of GA20 oxidase (GA20ox), which leads to increased GA levels (Eriksson et al. 2000). Nevertheless, the overexpression of GA20ox delayed growth cessation by being insensitive to FT2 signals in hybrid aspens, suggesting that there may be an FT2-independent regulation of SD-induced growth cessation through GAs (Eriksson et al. 2015). Intriguingly, recent research has proposed that poplar FT2 promotes shoot apex development and restricts internode elongation by regulating the GA 13-hydroxylation pathway and GA1 levels (Gomez-Soto et al. 2022). Therefore, these studies suggest that poplar FT2 acts upstream of GAs or synergistically with GAs to regulate the growth or cessation of apical buds (AB) in poplar.

However, neither modification of these regulated genes nor exogenous application of GAs is sufficient to completely prevent growth cessation and bud set under SDs, pointing to the presence of an additional SD-dependent FT2 regulatory pathway in poplar. Moreover, FT2 plays a dual role in flowering and seasonal growth of woody plants (Hsu et al. 2006). It thus seems reasonable to hypothesize that there are several pathways to fine tune FT2 in woody plants, so as to satisfy the distinct needs of dormancy and flowering. Here we demonstrated that PtoHY5a in poplar (Populus tomentosa Carr. Clone 741) plants, a tree ortholog of the Arabidopsis ELONGATED HYPOCOTYL5 (HY5), is a major mediator of dormancy induction by downstream signaling of the photoreceptor PHYB2. PtoHY5a is a positive regulator of FT2 and together they form a photoperiodic control transcriptional module that mediates control of growth cessation and bud set in poplar. We uncovered that PtoHY5a suppresses temperature-controlled bud break in poplar trees by negatively regulating GAs levels in AB. Moreover, PtoHY5a mediates growth of shoot apex and control of plant height by activating FT2 and regulating genes for GA deactivation and biosynthesis under long days (LDs). Thus, our results reveal a genetic network of photoperiodic control mediating seasonal growth in poplars, thus contributing to a deeper understanding of the molecular mechanisms underlying this process.

Results

PtoHY5a directly binds to and activates the FT2 promoter in poplar

To explore how FT2 is transcriptionally regulated in poplar, we employed the FT2 promoter (a 2,016 bp upstream sequence of the FT2 start codon) as bait against a P. tomentosa cDNA library in yeast one hybrid (Y1H) assays in vitro. We obtained several interesting candidates (Supplementary Table S1), including a homolog of Arabidopsis ELONGATED HYPOCOTYL5, PtoHY5a (Fig. 1A). Studies in herbaceous plants have shown that HY5 typically regulates transcription by directly binding to the ACE-box cis-element (ACGT) (Stracke et al. 2010; Norén et al. 2016; Guo et al. 2021). We found three ACE-box motifs in the PtoFT2 promoter (Fig. 1F). To verify the relationship between PtoHY5a and FT2, we carried out a dual-luciferase (Luc) assay by transiently co-expressing ProPtoFT2:LUC2 or ProPtoFT2(mutant):LUC2 with 35S:PtoHY5a in leaves of Nicotiana benthamiana and protoplasts of poplar leaves (Fig. 1B), respectively. We found that PtoHY5a enhanced the activity of LUC by 3-fold and 4.5-fold in poplar protoplasts and N. benthamiana leaves, respectively, compared to the control (the empty vector). However, there was no significant change in LUC activity compared to the control when the ACE motifs on the FT2 promoter were mutated (Fig. 1, C to E), suggesting that PtoHY5a can activate the expression of FT2. Next, we tested whether PtoHY5a can directly bind to these sites using the EMSA assay (see Materials and methods). Indeed, PtoHY5a could bind to the ACE-box motifs of the PtoFT2 promoter. When the motif was mutated, PtoHY5a could not bind to those mutant probes (Fig. 1F). Thus, our results demonstrated that the PtoHY5a can activate the expression of FT2 by binding to the ACE motifs of its promoter.

Figure 1.

PtoHY5a activates PtoFT2 by directly binding to its promoter. A) Y1H analysis of the interaction between PtoHY5a and the promoter of PtoFT2. SD/−Leu/AbA, synthetic medium lacking leucine, supplemented with 0, 100 and 200 ng·mL⁻¹ aureobasidin A (AbA), as indicated. B) The effector, reporter, and reference constructs used in the luciferase activity tests are depicted schematically. PtoHY5a under the control of the 35S promoter (Pro35S) was used as the effector. The firefly luciferase gene LUC driven by the PtoFT2 promoter (ProPtoFT2) and the Renilla luciferase gene REN under the control of the 35S promoter were used as the reporter and internal control, respectively. The ProPtoFT2-m indicates the promoter of PtoFT2 harboring the mutated ACE-box (AAAA). C) Luciferase activity assay showing that PtoHY5a binds to the promoter region of PtoFT2 was measured in N. benthamiana leaves, PtoHY5a activate PtoFT2 promoter activities. The color bar indicates the intensity of luciferase activity. Replacing the ACE-box with AAAA in PtoFT2 promoters resulted in the loss of transcriptional regulation by PtoHY5a on target promoters. D) Quantitative analysis of relative luciferase activity of the experimental materials in N. benthamiana leaves. Three randomly selected fields from three individual N. benthamiana plants per group were used for counting. E) Quantitative analysis of relative luciferase activity of the experimental materials in protoplasts from P. tomentosa. (C to E) The ProPtoFT2 indicates the promoter of PtoFT2 gene. The ProPtoFT2-m indicates the promoter of PtoFT2 harboring the mutated ACE-box (AAAA). The relative luciferase activities are normalized to the Renilla luciferase activity (Relative LUC:REN activity). F) EMSA analysis showed that PtoHY5a directly binds to the ACE-box of the PtoFT2 promoter in vitro. The black ellipse indicates the position of the ACE-box, p1 to p3 are EMSA probes. Recombinant PtoHY5a was isolated from E. coli cells, and DNA-binding experiments using PtoFT2 and PtoFT2 mutant probes were conducted. As negative controls, purified proteins from empty vectors were employed. Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on two-way ANOVA followed by Fisher's LSD test.

We analyzed the FT2 promoter of five poplar species with genome sequences (P. tomentosa, Populus trichocarpa, Populus euphratica, Populus tremula × tremuloides (Ptotrx T89), and Populus alba) and found that all of them contained three to five unequal ACE-box motifs (Supplementary Fig. S1), at least two of them within the active regulatory region of −500 bp upstream of the FT2 start codon. These results suggest that the HY5a–FT2 module may be a generic regulatory module in Populus.

Evolutionarily conserved PtoHY5a is ubiquitously expressed and localizes to the nucleus

HY5 is highly conserved in different plant species according to phylogenetic analyses (Supplementary Fig. S2A). It contains three conserved motifs, including a bZIP domain which is responsible for binding target DNA, a COP1 interaction core sequence (VPE/DG), and a consensus casein kinase II phosphorylation site (Supplementary Fig. S2B). Interestingly, we observed that gene duplication had occurred in Populus (both in P. tomentosa and P. trichocarpa) resulting in the presence of two HY5 homologous genes, HY5a and HY5b, with a high similarity of 94.6% (Supplementary Fig. S2C).

We then tested in which tissues PtoHY5a and PtoHY5b were expressed by RT-qPCR. We observed that PtoHY5a and PtoHY5b were ubiquitously expressed in all tissues of WT poplars, but PtoHY5a was highly expressed in young leaves (YL), mature leaves (ML) and in the AB. In contrast PtoHY5b was highly expressed in ML (Supplementary Fig. S3A). Assessing seasonal expression patterns indicated that PtoHY5a was mainly expressed in leaves during spring. After abundant expression in spring, PtoHY5a continued to be expressed at lower levels in leaves until mid-winter. Conversely, PtoHY5b transcripts in leaves were abundant during winter (Supplementary Fig. S3B). In the AB, both PtoHY5a and PtoHY5b showed a more stable seasonal expression pattern, except for a relatively high expression of PtoHY5b in winter (Supplementary Fig. S3B). These results suggest that the accumulation of PtoHY5a and PtoHY5b is temporally and spatially separated.

To test whether PtoHY5b can also activate PtoFT2 expression like PtoHY5a, we performed a dual-luciferase (Luc) assay by transiently co-expressing ProPtoFT2:LUC2 with Pro35S:PtoHY5b in leaves of N. benthamiana and protoplasts of poplar leaves, respectively (Supplementary Fig. S3C). The results showed that PtoHY5b could not activate the expression of FT2 in either N. benthamiana leaves or poplar protoplasts (Supplementary Fig. S3, D and E). Moreover, the expression level of endogenous FT2 did not show significant changes in protoplast cells with transiently overexpressing PtoHY5b compared to the control group, but significantly increased in protoplast cells with transiently overexpressing PtoHY5a (Supplementary Fig. S3F), indicating that there may be differences in the transactivation of FT2 by PtoHY5a and PtoHY5b. Based on the direct activation of FT2 by PtoHY5a (Fig. 1, A to F) and their highly similar seasonal expression patterns of PtoHY5a and poplar homolog of FT2 (Hsu et al. 2011), we subsequently conducted a more systematic study of the function of PtoHY5a.

We further confirmed PtoHY5a expression in roots, stems, and leaves using GUS staining, with PtoHY5_Pro:GUS reported lines (Supplementary Fig. S4A). From the cross-sections of the roots, stems, and petioles, we observed that PtoHY5a was mainly expressed in the epidermis of roots, in all tissues of stems, and in the epidermis and vascular tissues of petioles (Supplementary Fig. S4, B to G). We examined PtoHY5a subcellular localization by co-transforming the PtoHY5a-GFP fusion plasmid and the pSAT4A-Pro35S:HMGB1-mcherry nuclear marker plasmid into rice (Oryza sativa) protoplasts (Supplementary Fig. S4H). A strong nuclear signal was detected in PtoHY5a-GFP plants, which is consistent with the function of a bipartite NLS at residues 89 to 109 in PtoHY5a. Having characterized PtoHY5a, the next step was to investigate its functions.

PtoHY5a affects vegetative growth and rescues the early flowering phenotype of the Arabidopsis hy5 mutant

To address the functions of PtoHY5a in Populus, we generated transgenic poplars overexpressing PtoHY5a (PtoHY5a-OE) (Supplementary Fig. S5, A and B) and also knockout mutants (PtoHY5a-KO) using CRISPR/Cas9 technology (Supplementary Fig. S6, A to E). PtoHY5a-OE lines displayed growth inhibition under normal conditions, including shorter plant height, smaller leaf lamina size, and less biomass than the WT plants (Supplementary Fig. S7, A to H). Similarly, we also observed the low growth phenotype in transgenic Arabidopsis lines heterologously expressing PtoHY5a (Supplementary Fig. S8, A to C). In contrast, PtoHY5a-KO lines showed higher plant height, larger leaf lamina size, and more biomass than the WT (Supplementary Fig. S7, A to H).

We further investigated PtoHY5a function in light signaling using A. thaliana hy5 mutants. The hy5 showed an elongated hypocotyl under a constant light (Osterlund et al. 2000), whereas the PtoHY5a-OE/hy5 complementation lines showed the expected short hypocotyls, similar to the WT (Col-0), and rescued the hy5 deficient phenotype under constant light (Supplementary Fig. S8, D to G). There was no significant difference in growth among genotypes in the dark, suggesting that the PtoHY5a can restore the light-signaling pathway in A. thaliana. In addition, our results showed that heterologous expression of PtoHY5a in hy5 mutants rescued the early flowering phenotype, that was recently reported in A. thaliana under long days (LDs) (Chu et al. 2022). At the same time, Arabidopsis PtoHY5a-OE transgenic lines showed a delayed flowering phenotype relative to both Col-0 and hy5 plants (Supplementary Fig. S8, H and I). These observations indicated that overexpression of PtoHY5a results in prolonged vegetative growth that delays flowering.

PtoHY5a suppresses SD-induced growth cessation, bud set, and chilling-induced bud burst

As PtoFT2 can be regulated by PtoHY5a in poplar, it is necessary to understand whether PtoHY5a is involved in the seasonal growth of trees. Gene expression analyses indicated that PtoHY5a displayed diurnal expression patterns peaking at the end of the day (ZT16) in LDs (Fig. 2A). However, its diurnal expression pattern was significant changes, reaching its peak at dawn (ZT24) in SDs (Fig. 2B). Immunoblotting experiment also showed that HY5a highly accumulated during the day and substantially decreased at night (Fig. 2, D and E). In addition, PtoHY5a transcript was downregulated by SDs in leaves (Fig. 2C). This suggests that PtoHY5a may be involved in those seasonal growth processes that are regulated by the photoperiod.

Figure 2.

Effects of PtoHY5a expression changes on the bud set in poplar. A, B) Diurnal expression analysis of PtoHY5a in leaves of poplar grown under LD conditions (16 h:8 h, light:dark; LD^16h, 22 °C) (A) and SD conditions (8 h:16 h, light:dark; SD^8h, 22 °C) (B). C) Expression analysis of PtoHY5a in LD^16h and SD^8h treatments. ML of the same leaf position at different treatment times were taken for the experiment at ZT14 (Zeitgeber Time). Its expression in LD^16h was used as a control. Leaf tissues were collected for RNA extraction followed by RT-qPCR assays. PtoUBQ was used as a reference gene. Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test. D, E) Western blots showed PtoHY5a protein expression level in LD^16h(D) and SD^8h(E) conditions. Samples were harvested at the indicated times and subjected to immunoblot analysis. Samples were analyzed by immunoblotting with anti-PtoHY5a antibody and anti-Actin antibody. Actin served as a loading control. Gray and white boxes indicate night and day, respectively. F) Apical buds of WT, PtoHY5a-OE (PtoHY5a overexpressing lines 4, 5 and 6), and PtoHY5a-KO (PtoHY5a knockout lines 5 and 7) plants after 4 and 6 wk of SD^8h conditions (arrows indicate apical buds that have stopped growing). Scale bars indicate 1 cm. G) Bud set score of WT, PtoHY5a-OE, and PtoHY5a-KO plants after transfer from LD^16h to SD^8h conditions. H) Cumulative shoot elongation in WT, PtoHY5a-OE, and PtoHY5a-KO plants after transfer from LD^16h to SD^8h conditions. Data shown are the mean values from five individual plants of each line. Error bars indicate mean ± Sd and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test.

We conducted a detailed study under controlled-growth conditions to investigate the function of PtoHY5a during SD-induced growth cessation and bud set. When transferred from LD^16h to SD^8h, PtoHY5a-KO plants displayed a hypersensitive phenotype that started growth cessation after only 3 wk and bud setting after 4 wk (Fig. 2, F to H). In contrast, the PtoHY5a-OE plants showed a hyposensitivity phenotype that started growth cessation after 5 wk and bud setting after 6 wk. The WT plants ceased to grow and started to set buds after PtoHY5a-KO but before PtoHY5a-OE (Fig. 2, F to H). These results suggest that PtoHY5a suppresses SD-induced growth cessation and bud set.

We also evaluated the function of PtoHY5a in bud break of poplar, a key trait for dormancy release, by simulating warm spring temperature conditions. The WT, PtoHY5a-OE, and PtoHY5a-KO plants that had set buds under the SD^8h treatment for 10 wk and were then cold treated for 3 mo. These plants were then returned to LD^16h at 25 °C for 2 wk (Fig. 3A), and we scored the timing of bud break (Fig. 3C). PtoHY5a-KO plants flushed 8 d after returning to a LD photoperiod with warm temperature and 4 d earlier than the WT; but the flushing of PtoHY5a-OE plants was delayed by 4 d (Fig. 3B). Furthermore, the WT showed initial bud burst after 10 d and shoot growth started after 20 d (scores 2 and 5, in Fig. 3, C and D, respectively). In contrast, the same stages occurred 16 and 26 d later in PtoHY5a-OE plants, and 6 and 18 d earlier in PtoHY5a-KO plants (Fig. 3, C and D). In addition, the PtoHY5a-KO plants grown outdoors also showed early bud break compared to the WT poplar plants, while bud break was substantially delayed in PtoHY5a-OE plants (Fig. 3E). Hence, these results suggest that PtoHY5a represses chilling-induced bud burst.

Figure 3.

PtoHY5a mediates in photoperiodic control of bud break. A) Apical buds of WT, PtoHY5a-OE (PtoHY5a overexpressing lines 4, 5 and 6), and PtoHY5a-KO (PtoHY5a knockout lines 5 and 7) plants after 3 mo of cold, SD conditions (8 h:16 h, light:dark; SD^8h, 4 °C) were shifted to warm, LD conditions (16 h:8 h, light:dark; LD^16h, 22 °C). Scale bars indicate 1 cm. B) Dynamics of bud break in PtoHY5a-OE and PtoHY5a-KO plants compared with WT. Timing and degree of bud break following 3 mo of chilling (4 °C) and subsequent bud break under LD and warm temperatures. Values are means of five biological replicates. C) Classification of the different apex stages during bud break used to monitor growth reactivation. D) Bud break score of WT, PtoHY5a-OE, and PtoHY5a-KO plants after transfer from cold, SD conditions (SD^8h, 4 °C) to warm, LD conditions (LD^16h, 25 °C). E) Bud breaking phenotypes of 2-yr-old WT, PtoHY5a-OE (lines 4, 5, and 6) and PtoHY5a-KO (lines 5 and 7) plants growing in the field in March 20. Scale bars indicate 5 cm. Data shown are the mean values from five individual plants of each line. Error bars indicate mean ± Sd and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test.

PtoHY5a positively regulates the expression of FT2, LAP1 and AIL1 in poplar

We previously showed that PtoHY5a can activate the expression of PtoFT2 by binding to the PtoFT2 promoter (Fig. 1, A to F). To further confirm that PtoHY5a affects growth cessation and bud set by regulating FT2, we examined the expression of FT2, LAP1, and AIL1 in PtoHY5a-OE and PtoHY5a-KO plants under LD^16h conditions. The results from RT-qPCR indicated that the expression level of PtoFT2 increased in PtoHY5a-OE lines, and decreased in PtoHY5a-KO lines under either LD or SD conditions (Fig. 4, A and B). The downstream genes of PtoFT2, PtoLAP1, and PtoAIL1, also exhibited similar expression pattern in apical buds (Supplementary Fig. S9, A to D).

Figure 4.

Overexpression of PtoFT2 suppresses bud set in PtoHY5a-KO under SDs. A) Relative expression levels of PtoFT2 in leaves of WT, PtoHY5a-OE (PtoHY5a overexpressing lines 4, 5 and 6), and PtoHY5a-KO (PtoHY5a knockout lines 5 and 7) lines. Poplar plants were grown under LD conditions (16 h:8 h, light:dark; LD^16h, 22 °C) for 4 wk, and ML of the same leaf position were taken for the experiment at ZT10 (Zeitgeber Time). B) Comparative expression analysis of PtoFT2 in leaves of WT, PtoHY5a-OE, and PtoHY5a-KO plants grown under LD^16h conditions, and 5 wk after transfer to SD conditions (8 h:16 h, light:dark; SD^8h, 22 °C). ML of the same leaf position at different treatment conditions were taken for the experiment at ZT10 (Zeitgeber time). Error bars indicate mean ± Sd from three biological replicates and asterisks indicate statistical differences (**P < 0.01; ***P < 0.001; Student's t-test). C, D) Diurnal expression analysis of PtoFT2 in leaves of WT, PtoHY5a-OE and PtoHY5a-KO plants grown under LD^16h(C) and SD^8h(D) conditions. Leaf tissues were collected for RNA extraction followed by RT-qPCR assays. PtoUBQ was used as a reference gene. Error bars indicate mean ± Sd and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test. E, F) Enrichment of a DNA fragment in the PtoFT2 promoter containing an ACE-box and quantified by ChIP-qPCR. The position diagram of the primer sets (#1 to #3) on the promoter used in ChIP-qPCR assays are shown above. NC indicates the location of the DNA fragment in the 3′ UTR region, with no ACE-box as a negative control. Chromatin from leaves of HA-PtoHY5a-OE and WT was isolated using anti-HA (E) and anti-PtoHY5a (F) antibody, respectively. IgG was used as a negative control. ChIP-purified DNA was used to perform ChIP-qPCR and expression values are represented as the percentage of input DNA. Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on two-way ANOVA followed by Fisher's LSD test. G) Apical buds of WT, PtoFT2-OE (PtoFT2 overexpressing line), PtoHY5a-KO/PtoFT2-OE (Transgenic lines overexpressing PtoFT2 in PtoHY5a-KO-L7 mutant poplar, 1, 2 and 3), PtoHY5a-OE (PtoHY5a overexpressing line 6), and PtoHY5a-KO (PtoHY5a knockout line 7) plants after 6 wk of SD^8h conditions (arrows indicate apical buds that have stopped growing). Scale bars indicate 1 cm. Bud set score (H) and height increment (I) of WT, PtoFT2-OE, PtoHY5a-KO/PtoFT2-OE (lines 1, 2 and 3), PtoHY5a-OE (line 6), and PtoHY5a-KO (line 7) plants after transfer from LD^16h to SD^8h conditions. Data shown are the mean values from five individual plants of each line. Error bars indicate mean ± Sd and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test.

Moreover, PtoFT2 displayed a diurnal expression pattern in the WT plants, with the highest expression at Zeitgeber time 12 h (ZT12) under LDs (Fig. 4, C and D), however, the peak expression of PtoFT2 appeared 8 h earlier in PtoHY5a-OE plants, and was delayed by 4 h in PtoHY5a-KO plants under LDs (Fig. 4C; Supplementary Fig. S10D). Under SDs, the expression of PtoFT2 in WT and PtoHY5a-KO plants did not exhibit a clear diurnal expression and oscillation rhythm (Fig. 4C; Supplementary Fig. S10E), but the PtoFT2 still exhibited a periodic oscillation rhythm in PtoHY5a-OE plants, with two high expression peaks at Zeitgeber time 4 and 12 h (ZT4 and ZT12). The changed expression patterns of PtoFT2 under LDs and SDs in the PtoHY5a-OE and PtoHY5a-KO lines, further suggested that they are controlled by PtoHY5a.

To further examine the ability of PtoHY5a to bind to the PtoFT2 promoter in vivo, ChIP-qPCR assay was performed in HA-PtoHY5a overexpression lines and WT using anti-HA and anti-PtoHY5a antibodies, respectively. The result showed that both HA-PtoHY5a and native PtoHY5a bound to the promoter of PtoFT2 through the ACE-box (Fig. 4, E and F). Taken together, these results further supported the notion that HY5a controls SD-induced dormancy in poplar by directly and positively regulating the expression of FT2.

Overexpression of PtoFT2 suppresses the premature growth cessation and early bud set phenotype of PtoHY5a-KO plants

Previous research showed that FT2 overexpression in poplar suppresses SD perception, subsequently leading to failure to cease growth and bud set in response to SDs (Böhlenius et al. 2006; Hsu et al. 2011). We hypothesize that the earlier growth cessation of PtoHY5a-KO plants may be due to their inability to activate PtoFT2 expression in SDs, the major target of dormancy induction. To address this question, we overexpressed the PtoFT2 (PtoFT2-OE) in PtoHY5a-KO-L7 mutant poplar plant (Supplementary Fig. S10, A to C). We then compared growth cessation and bud set among WT, PtoHY5a-OE-L6, PtoHY5a-KO-L7, PtoFT2-OE-L1 and PtoHY5a-KO/PtoFT2-OE lines (L1, L2 and L3) after SD^8h treatment.

PtoHY5a-KO-L7 plants ceased growth after 2 wk, set buds after 3 to 4 wk, and displayed earlier growth cessation and bud set relative to the WT plants. However, PtoHY5a-KO/PtoFT2-OE plants continued to grow, formed new leaves, and did not form buds, even after 6 wk of SDs when PtoHY5a-OE-L6 plants had already set buds (Fig. 4, G to I). Consistently, PtoFT2-OE-L1 plants showed premature flowering in juvenile, and did not set bud (Fig. 4G; Supplementary Fig. S11A). Interestingly, plants overexpressing PtoHY5a flowered earlier in the second year after transplanting outdoors (Supplementary Fig. S11B). Thus, our results indicate that PtoFT2 overexpression suffices to suppress the defect in SD-induced premature growth cessation and bud set resulting from loss of PtoHY5a function.

PtoHY5a negatively regulates PtoLHY2 expression in poplar

In Arabidopsis, the clock genes LHY, TOC1, CIRCADIAN CLOCK-ASSOCIATED1 (CCA1), and EARLY FLOWERING 3 (ELF3) are downstream target genes for HY5, but no changes in their transcription levels were detected in the hy5 mutant (Lee et al. 2007). In poplar, the downregulation of LHY led to altered phase and period of clock-controlled gene expression, delayed SD-induced growth cessation, bud set and chilling-induced bud burst (Ibáñez et al. 2010). To explore whether PtoHY5a also regulates dormancy induction by regulating the clock gene LHY2, we examined the expression levels of LHY2 in PtoHY5a-OE and PtoHY5a-KO lines. As shown in Fig. 5, A to C, PtoLHY2 was downregulated in PtoHY5a-OE plants and upregulated in PtoHY5a-KO plants, relative to the WT, indicating that LHY2 can be negatively regulated by PtoHY5a in poplar. This response has not yet been observed in hy5 mutants of Arabidopsis and other herbaceous plants. In addition, PtoLHY2 in poplar displayed a clear circadian rhythm, with a peak expression of PtoLHY2 at dawn (ZT24) in LDs, whereas its peak expression in SDs was advanced by 4 h, at ZT20 (Fig. 5, B and C), as compared with LDs. Its diurnal pattern was not substantially altered in PtoHY5a-OE and PtoHY5a-KO plants compared to the WT (Fig. 5, B and C).

Figure 5.

PtoHY5a affects the expression of clock genes PtoLHY2.A) Expression analysis of PtoLHY2 in the WT, PtoHY5a-OE (PtoHY5a overexpressing lines 4, 5, and 6), and PtoHY5a-KO (PtoHY5a knockout lines 5 and 7) plants. Poplar plants were grown under LD conditions (16 h:8 h, light:dark; LD^16h, 22 °C) for 4 wk, ML of the same leaf position were taken for the experiment at ZT0 (Zeitgeber Time). Error bars indicate mean ± Sd from three biological replicates and asterisks indicate statistical differences (**P < 0.01; ***P < 0.001; Student's t-test). B, C) Diurnal expression analysis of PtoLHY2 in leaves of WT, PtoHY5a-OE, and PtoHY5a-KO plants grown under LD^16h(B) conditions and SD conditions (8 h:16 h, light:dark; SD^8h, 22 °C) (C). Leaf tissues were collected for RNA extraction followed by RT-qPCR assays. PtoUBQ was used as a reference gene. Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) according to one-way ANOVA followed by a Fisher's LSD test. D) The effector, reporter and reference constructs used in the luciferase activity tests are depicted schematically. REN, Renilla luciferase; LUC, firefly luciferase. E) Luciferase activity assay showing that PtoHY5a binds to the promoter region of PtoLHY2 was measured in N. benthamiana leaves. The color bar indicates the intensity of luciferase activity. F) Quantitative analysis of relative luciferase activity of the experimental materials in N. benthamiana leaves. Three randomly selected fields from three individual N. benthamiana plants per group were used for counting. G) Quantitative analysis of relative luciferase activity of the experimental materials in protoplasts from P. tomentosa. In (E) to (G), The ProPtoLHY2 indicates the promoter of PtoLHY2 gene. The ProPtoLHY2-m indicates the promoter of PtoLHY2 harboring the mutated ACE-box (AAAA). Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on two-way ANOVA followed by Fisher's LSD test. H) The position diagram of the probe and primer on the promoter used in EMSA and ChIP-qPCR assays. The position diagram of the primer sets (#1 and #2) on the promoter used in ChIP-qPCR assays are shown above. NC indicates the location of the DNA fragment in the 3′ UTR region, with no ACE-box as a negative control. I) EMSA analysis showed that PtoHY5a directly binds to the ACE-box of the PtoLHY2 promoter in vitro. The black ellipse indicates the position of the ACE-box, p1 and p2 are EMSA probes. Recombinant PtoHY5a was isolated from E. coli cells, and DNA-binding experiments using PtoLHY2 and PtoLHY2 mutant probes were conducted. As negative controls, purified proteins from empty vectors were employed. J, K) Enrichment of a DNA fragment in the PtoLHY2 promoter containing the ACE-box and quantified by ChIP-qPCR. Chromatin from leaves of HA-PtoHY5a-OE and WT was isolated using anti-HA (J) and anti-PtoHY5a (K) antibody, respectively. IgG was used as a control antibody. ChIP-purified DNA was used to perform ChIP-qPCR and expression values are represented as the percentage of input DNA. Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on two-way ANOVA followed by Fisher's LSD test.

To demonstrate the regulation of PtoHY5a on PtoLHY2, we employed a LUC assay by transiently co-expressing ProPtoLHY2:LUC2 or ProPtoLHY2(mutant):LUC2 with Pro35S:PtoHY5a in leaves of N. benthamiana and poplar protoplasts, respectively (Fig. 5D). Our result showed that PtoHY5a caused a 2.5- and 3-fold reduction in LUC activity relative to the control (the empty vector) in leaves of N. benthamiana and poplar protoplasts, respectively (Fig. 5, E to G). There was no significant change in LUC activity compared to the control when the ACE motifs on the PtoLHY2 promoter were mutated, confirmed that PtoHY5a can inhibit the expression of LHY2 (Fig. 5, E to G). Furthermore, we found four ACE-box motifs in the PtoLHY2 promoter by sequence analysis (Fig. 5H). EMSA and ChIP-qPCR experiments suggested that ACE-box motif on the PtoLHY2 promoter can be directly bound by PtoHY5a (Fig. 5, I to K). Thus, these results demonstrated that the PtoHY5a can inhibit the expression of PtoLHY2 by binding to ACE-box motif of its promoter. Our results reveal that PtoHY5a indirectly regulates FT2 through directly regulating PtoLHY2.

PtoHY5a acts as the transcription factor downstream of PtoPHYB2 that regulates seasonal growth in poplar

The photosensor PHYB2 plays a key role in photoperiod perception and in mediating SD-induced dormancy induction in poplar (Ingvarsson et al. 2006; Ding et al. 2021). PtoPHYB2 showed a diurnal expression pattern, with a peak expression at ZT4 under LDs (Fig. 6A), its expression level was significantly reduced in leaves under SDs (Fig. 6C). Under SDs, PtoPHYB2 did not exhibit a clear diurnal expression pattern and its oscillation amplitude was substantially attenuated, further indicating the regulatory effect of photoperiod on PtoPHYB2 (Fig. 6B). Previous studies indicated that the downstream signaling of PHYB can be positively regulated by HY5 in A. thaliana (Osterlund et al. 2000; Lee et al. 2007; Sakuraba et al. 2018). We thus explored the regulatory relationship between PtoPHYB2 and PtoHY5a by using PtoPHYB2-OE and PtoPHYB1/PHYB2-RNAi plants (Supplementary Fig. S12, C and E). Similar to previous observation on the growth phenotype in PHYB-RNAi aspen (Ding et al. 2021), PtoPHYB2-OE lines exhibited delayed SD-induced growth cessation and bud set, whereas PtoPHYB1/2-RNAi lines showed earlier SD-induced growth cessation (Supplementary Fig. S12, A and B).

Figure 6.

Overexpression of PtoPHYB2 does not affect bud formation in PtoHY5a-KO plants during SDs. A, B) Diurnal expression analysis of PtoPHYB2 in leaves of poplar under LD conditions (16 h:8 h, light:dark; LD^16h, 22 °C) (A) and SD conditions (8 h:16 h, light:dark; SD^8h, 22 °C) (B). C) Expression analysis of PtoPHYB2 in LD^16h and SD^8h treatments. ML of the same leaf position at different treatment times were taken for the experiment at ZT4 (Zeitgeber Time). Error bars indicate mean ± Sd from three biological replicates and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test. D) Expression of PtoHY5a in the WT, PtoPHYB2-OE (PtoPHYB2 overexpressing lines 3, 9, and 10), and PtoPHYB1/2-RNAi (PtoPHYB1/2 RNA co-interference lines 3, 6, and 9). Poplar plants were grown under LD^16h for 4 wk, ML of the same leaf position were taken for the experiment at ZT4 (Zeitgeber Time). Error bars indicate mean ± Sd from three biological replicates and asterisks indicate statistical differences (*P < 0.05; ***P < 0.001; Student's t-test). E) Apical buds of WT, PtoHY5a-KO (PtoHY5a knockout line 7), PtoHY5a-KO/PtoPHYB2-OE (Transgenic lines overexpressing PtoPHYB2 in PtoHY5a-KO-L7 mutant plant, 15, 16, and 17), PtoPHYB2-OE (PtoPHYB2 overexpressing line 10), and PtoPHYB1/2-RNAi (PtoPHYB1/2 RNA co-interference line 9) plants after 3 and 4 wk under SD^8h conditions (arrows indicate apical buds that have stopped growing). Scale bars indicate 1 cm. F) The relative expression levels of PtoPHYB2 in WT, PtoHY5a-KO (line 7), PtoHY5a-KO/PtoPHYB2-OE (lines 15, 16 and 17), PtoPHYB2-OE (line 10), and PtoPHYB1/2-RNAi (line L9) plants. Leaf tissues were collected for RNA extraction followed by RT-qPCR assays. PtoUBQ was used as a reference gene. Error bars indicate mean ± Sd from three biological replicates and asterisks indicate statistical differences (*P < 0.05; ***P < 0.001; ns, not significant; Student's t-test). G) Bud set score of WT, PtoHY5a-KO (line 7), PtoHY5a-KO/PtoPHYB2-OE (lines 15, 16, and 17), PtoPHYB2-OE (line 10) and PtoPHYB1/2-RNAi (line 9) plants after transfer from LD^16h to SD^8h conditions. Data shown are the mean values from five individual plants of each line. Error bars indicate mean ± Sd and different lowercase letters indicate significant differences (P < 0.05) based on one-way ANOVA followed by Fisher's LSD test.

We also found that PtoPHYB2 could regulate the expression of PtoHY5a at the transcriptional level, as PtoHY5a expression was upregulated in the PtoPHYB2-OE lines, and significantly downregulated in the PtoPHYB1/2-RNAi lines (Fig. 6D). Thus, we employed further genetic approach to confirm the regulatory relationship between PtoPHYB2 and PtoHY5a. We overexpressed PtoPHYB2 in PtoHY5a-KO-L7 mutant plant and compared SD-induced growth cessation and bud set (Fig. 6, D to F). PtoHY5a-KO-L7 and PtoHY5a-KO/PtoPHYB2-OE lines (L15, L16, and L17) both showed earlier growth cessation and bud set compared to WT and plants, but little difference was observed between PtoHY5a-KO and PtoHY5a-KO/PtoPHYB2-OE lines under SD^8h (Fig. 6, E to G). This demonstrates that PtoHY5a-KO can eliminate the delay of growth cessation and bud set in PtoPHYB2-OE under SDs. Therefore, PtoHY5a acts as the transcription factor downstream of PtoPHYB2 that regulates seasonal growth.

PtoHY5a reduces GAs levels by regulating the expression of GA biosynthetic and catabolic enzyme genes in apical buds

Previous study indicated that dormancy release in winter and bud break in spring in poplar are related with GA levels (Rinne et al. 2011). To investigate whether the ProHY5a functions in bud break via the GAs pathway, we analyzed the transcript abundances of GA-deactivation and biosynthetic genes which were mainly expressed in the shoot apex (Katyayini et al. 2020), GA-deactivation genes including GA2ox1, GA2ox4, GA2ox5, GA2ox6, and GA2ox7 maintained higher expression levels in apical buds of PtoHY5a-OE lines that had been previously treated with cold temperature for 3 months (3MC) and subsequently returned to warm temperature for 2 weeks (2WW), while their expression was significantly downregulated in those of PtoHY5a-KO plants compared to the WT (Fig. 7A). Interestingly, GA2ox3 maintained higher expression levels in apical buds of both PtoHY5a-OE and PtoHY5a-KO lines at 3MC and 2WW compared to the WT (Fig. 7A).

Figure 7.

PtoHY5a is involved in GA metabolism during poplar bud breaking. A) Comparative expression analyses of GA deactivation genes GA2ox1, GA2ox3, GA2ox4, GA2ox5, GA2ox6, GA2ox7, and (B) GA biosynthesis genes GA3ox2, GA20ox2-1, GA20ox3, GA20ox5, GA20ox6, and GA20ox8 in apical buds of WT, PtoHY5a-OE (PtoHY5a overexpressing line 6), and PtoHY5a-KO (PtoHY5a knockout line 7) plants after 3 mo of SD and cold treatments (8 h:16 h at 4 °C, SD^8h; 3MC) and then after 2 wk under long-day and warm temperatures (16 h:8 h at 25 °C, LD^16h; 2WW) treatment. C to E) Schematic representation of GA metabolism. C) The endogenous levels of GA12, GA15, GA24, GA9, GA4, GA53, GA44, GA19, GA20, and GA1 in apical buds of WT, PtoHY5a-OE, and PtoHY5a-KO plants. D) The contents of GA7 and GA51 generated with GA9 as a precursor, and the content of GA34 generated with GA4 as a precursor. E) The contents of GA5 and GA29 generated with GA20 as a precursor, and the content of GA8 generated with GA1 as a precursor. Apical bud tissues were collected for RNA extraction followed by RT-qPCR assays. PtoUBQ was used as a reference gene. Error bars indicate mean ± Sd from three biological replicates and asterisks indicate statistical differences (*P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant; Student's t-test).

Conversely, GA biosynthetic genes GA3ox2, GA20ox2-1, GA20ox3, GA20ox5, GA20ox6, and GA20ox8 were drastically decreased in apical buds of PtoHY5a-OE plants at 2WW stage, whereas their expression was significantly upregulated in PtoHY5a-KO plants (Fig. 7B). Besides, the expression of GA-deactivated genes (with the exception of GA2ox3) was significantly reduced in apical buds at the 2WW stage compared to those at the 3MC stage (Fig. 7A), whereas the expression pattern of GA biosynthetic genes was opposite (Fig. 7B). In addition, we detected the content of GAs in the apical buds of WT and transgenic plants after 3MC of low temperature and 2WW of LD and warm environment. As shown in Fig. 7, C to E, the levels of active GAs (GA1 and GA4), GA precursors (GA9 and GA20), and metabolites (GA8 and GA34) increased significantly in PtoHY5a-KO apical buds, and decreased markedly in PtoHY5a-OE apical buds compared to the WT. These results indicate that PtoHY5a is involved in controlling bud break through regulating the GA biosynthetic and catabolic pathways.

Discussion

PtoHY5a negatively regulates induction of winter dormancy by activating FT2 and repressing LHY2 in poplar

The pathway for dormancy induction in photoperiod-sensitive woody plants shares one key factor with the floral induction pathway, the FT2 protein. Winter dormancy induction or normal growth in the growing season is dependent on the repression or activation of FT2, and woody plant flowering is dependent on FT2 accumulation to some high level. Photoperiod change is the major pathway for induction of winter dormancy in perennial woody plants (Horvath et al. 2003; Singh et al. 2017). Although the photoperiod pathway to regulate FT concentrates on the CO–FT module in Arabidopsis, woody plants must possess multiple pathways to fine tune FT2 to satisfy the distinct needs for dormancy or growth or flowering. Our results showed that PtoHY5a directly activates the expression of FT2 by binding to the ACE motifs of the FT2 promoter (Figs. 1, A to F and 4, E and F). Moreover, this study also demonstrated that PtoHY5a regulates the diurnal expression pattern of the FT2, and the FT2 in PtoHY5a-OE plants maintained periodic oscillation rhythm even under LDs (Fig. 4, C and D). Consequently, PtoHY5a-KO plants exhibited early growth cessation and bud set induced by SDs, while the PtoHY5a-OE showed delayed SD-induced growth cessation and bud set (Fig. 2, F to H). This result is consistent with the fact that FT2 overexpression and RNA interference separately prevent or accelerate, respectively, SD-induced growth cessation and bud set (Hsu et al. 2011; André et al. 2022).

Furthermore, our genetic complementation experiment identified that overexpressing PtoFT2 is sufficient to suppress the defect of PtoHY5a-KO mutant (Fig. 4, G to I). In addition, the early flowering phenotype of PtoHY5a-OE plants further confirmed the existence of the HY5-FT2 regulatory pathway in poplar (Supplementary Fig. S11, A and B). Recently, it has been suggested that HY5 in A. thaliana represses PHYTOCHROME-INTERACTING FACTOR 4 (PIF4) and CONSTANS-LIKE5 (COL5) by deacetylation through HISTONE DEACETYLASE9 (HDA9) and that it indirectly suppresses FT expression (Chu et al. 2022). However, our results reveal a pathway for FT regulation in plants, which is essential for the induction and establishment of photoperiod-dependent dormancy in perennial trees.

On the other hand, PtoHY5a indirectly regulates PtoFT2 transcription by repressing PtoLHY2. The effects of LHY2 as a negative regulator of FT2 transcription are well known (Ibáñez et al. 2010; Ramos-Sánchez et al. 2019). An earlier study had reported that HY5 homolog in Arabidopsis binds to promoter regions of both the morning-expressed circadian oscillator (CCA1/LHY) and the evening-expressed oscillator (TOC1/ELF4) (Lee et al. 2007). Recent studies further indicated that the CO–PSEUDO-RESPONSE REGULATOR 5 complex interacts with HY5 and regulate these circadian clock in Arabidopsis (Pham et al. 2018). However, the hy5 mutant in Arabidopsis did not significantly affect the transcription of those circadian clock genes (Lee et al. 2007; de los Reyes et al. 2023). Here, we report that PtoHY5a represses the expression of LHY2 by binding to the ACE-box motif of the LHY2 promoter (Fig. 5, A to I), showing the importance of HY5–LHY pathway in poplar species rather than Arabidopsis. In poplar, RNA interference of PttLHY2 led to altering the phase and period of clock-controlled gene expression and delayed growth cessation (Ibáñez et al. 2010). The accumulation of LHY2 induced under long nights repressed the expression of FT2 through binding to the FT2 promoter, leading to growth cessation under SDs (Ramos-Sánchez et al. 2019).

In this study, overexpression of PtoHY5a caused a delay in growth cessation and bud set (Fig. 2, F to H). This response was similar to the phenotype of the PtLHY2RNAi poplar (Ibáñez et al. 2010). Furthermore, PtoHY5a positive regulated the expression of LAP1 and AIL1 in apical buds (Supplementary Fig. S9, A to D), which acts downstream of FT2 in poplar (Karlberg et al. 2011). This further explains the function of PtoHY5a in growth cessation. Thus, our results indicate that PtoHY5a confers upregulation of PtoFT2 by two different ways in poplar: (i) by directly activating FT2 expression by binding to its promoter; and (ii) by indirectly upregulating the expression of FT2 through inhibiting of the expression of PtoLHY2 (although we did not evaluate here which of these two ways is more important to explain the change in PtoFT2 induced by PtoHY5a).

PtoHY5a regulates dormancy induction downstream of the photoreceptor PHYB signaling in poplar

HY5 is the central hub of the transcriptional network involved in photomorphogenesis. In Arabidopsis, either the photoreceptor PHYB or CRY transduce photoperiod signals to HY5 through suppression of COP1 activity, which is a potent inhibitor of HY5 function, and HY5 participates in photomorphogenesis by regulating the transcription of light-inducible genes (Gangappa and Botto 2016). In our study, PtoHY5a showed markedly diurnal expression pattern in LDs and SDs (Fig. 2, A to C), and had a distinct seasonal expression pattern (Fig. 2C; Supplementary Fig. S3B), which indicated that PtoHY5a may be involved in seasonal growth processes that are regulated by the photoperiod. Recent studies have shown that PHYB is involved in growth cessation induced by SDs in poplar (Ding et al. 2021). Here, we found that the expression of PtoHY5a was positively regulated by PHYB in PtoPHYB2 transgenic lines (Fig. 6D).

The fact that PtoHY5a-KO/PtoPHYB2-OE and PtoHY5a-KO showed similar phenotypes indicated that the delay in SD-induced bud dormancy for PHYB2 overexpression is fully reversed by the PtoHY5a knockout (Fig. 6, E to G). Therefore, our results reveal that PtoHY5a in poplar is an important regulator for SD-induced winter dormancy, which signal downstream of the photoreceptor PHYB in poplar plants. It was suggested that PHYB2 could promote the expression of FT2 through inhibiting its downstream target phytochrome-interacting factor 8 (PIF8) (Pham et al. 2018). The relationship of between PIF8 and HY5 in poplar needs to be further explored. Moreover, one PIF member (PIF) served as a cofactor of the COP1–SPA complex in the night to degrade the HY5 protein in Arabidopsis. Understanding how PIFs regulate HY5 in woody plants should be at the forefront of our research efforts.

PtoHY5a mediates dormancy breaking by regulating GA metabolism pathway

Dormancy breaking in perennial woody plants occurs in the spring, after endodormancy has been completed under low temperatures in the winter. Recent work has shown that the duration for dormancy release and bud flush was shorter for high latitude poplar ecotypes than for low latitude poplar ecotypes, when exposed to similar conditions of long days and warm temperatures. This indicated that higher latitude ecotypes flushed their buds faster to optimize the short duration of the growth season (Thibault et al. 2020). Our results additionally showed that PtoHY5a-KO lines flushed buds faster than the WT plants, while PtoHY5a-OE lines took longer periods of time (Fig. 3, A to E). This indicates that PtoHY5a can effectively regulate this process, potentially being one of the mechanisms underlying the differences in dormancy breaking with latitude.

Low temperature can induce the increase of GAs in the late stages of cold acclimation, and GA is required for the release of winter dormancy and the reactivation of shoot apical growth in trees (Rinne et al. 2011; Singh et al. 2018). In poplar, GA biosynthesis and deactivation enzymes have been extensively studied in different tissues and developmental stages, and their genes show specific expression patterns (Gou et al. 2011; Katyayini et al. 2020). Our results showed that PtoHY5a markedly upregulated the expression of key genes for GA deactivation in the apical buds, and significantly downregulated key genes for GA biosynthesis (Fig. 7, A and B). Consistently, GA levels were higher in the apical buds of PtoHY5a-KO lines, and lower in PtoHY5a-OE lines compared with WT (Fig. 7, C and D). These results suggest that PtoHY5a negatively mediates bud break by regulating the level of GAs in the apical buds.

Recent study has shown that the catabolic and biosynthetic genes of gibberellins are downstream of FT2 in the shoot apex of poplar (Gomez-Soto et al. 2022). However, FT2 cannot participate in the release and bud break of dormancy due to the loss of FT2 expression in the chilling-induced dormancy release and bud break stage (Hsu et al. 2011). Furthermore, it was suggested that the double knockout of FT2a and FT2b in hybrid aspen does not affect dormancy release and bud flush under low temperature induction (André et al. 2022). Therefore, we propose that bud flush mediated by the HY5a-GA module is independent of the FT2-GA pathway in apical buds. A number of studies in herbaceous plant have also shown that HY5 directly activates the key GA catabolic enzyme genes GA2ox and suppresses the accumulation of GAs (Weller et al. 2009; Wang et al. 2019).

PtoHY5a mediates growth of shoot apex and control of plant height under LDs

We also observed that PtoHY5a negatively regulates poplar height and biomass under growth-promoting conditions (LDs) (Supplementary Fig. S7, A to H). Similarly, transgenic Arabidopsis lines overexpressing PtoHY5a also exhibited the retarded growth phenotype (Supplementary Fig. S8A). These findings are also consistent with the function of rice bZIP48 (HY5 homolog) as its overexpression resulted in a semidwarf phenotype due to the downregulating GA biosynthesis genes (Burman et al. 2018). In addition, pea (Pisum sativum) LONG1, a divergent ortholog of Arabidopsis HY5, activates a GA catabolic key enzyme gene GA2ox2, in reducing active GA levels during photomorphogenesis (Weller et al. 2009). FT2 has been shown to promote the growth and development of shoot apex by activating LAP1 and AIL1 and gibberellin pathway (Karlberg et al. 2011; Azeez et al. 2014). In Contrast, in leaves, FT2 tunes the GA pathway, limiting GA1 production and restricting internode elongation (Gomez-Soto et al. 2022). Consistently, the overexpression of FT2 in poplars also showed a severe dwarfing phenotype (Hsu et al. 2006) and sustained shoot growth in SDs (Fig. 4, G to I), further confirming a dual role of poplar FT2 in promoting shoot apex development and restricting internode elongation.

Interestingly, under growth-promoting conditions (LDs), the expression of key genes for GA deactivation was significantly increased in the apical buds of PtoHY5a-OE plants and significantly decreased in PtoHY5a-KO plants, whereas the key genes for GA biosynthesis showed exactly the opposite expression changes (Supplementary Fig. S13, A to C), suggesting that PtoHY5a inhibits GA biosynthesis in apical buds. This is consistent with the decrease in GA levels in the apical buds of PtoHY5a-OE at the 2WW stage (Fig. 7, A to E). In the leaves, the expression patterns of key genes for GA deactivation and some biosynthesis (e.g. GA20ox5 and GA20ox8) were similar to those in the apical buds (Supplementary Fig. S13, B and D), which is the agreement with the results of recent studies on FT2 control of gibberellin in poplar leaves (Gomez-Soto et al. 2022). These results further explained the dwarfing phenotype of PtoHY5a-OE plants.

In poplar, the FT2 forms a transcriptional complex with FDL1 (FLOWERINGLOCUS D LIKE 1) to activate LAP1 and its downstream target gene AIL1, thereby promoting shoot apex development (Azeez et al. 2014; Tylewicz et al. 2015). Our studies have demonstrated that PtoLAP1 and PtoAIL1 are positively regulated by PtoHY5a in shoot apex (Supplementary Fig. S9, A to D), and that PtoHY5a activates the expression of PtoFT2 (Figs. 1, A to F and 4, A and B). It has been shown that FT2 loss of function plants set buds and showed elongated internodes under growth promoting conditions (Gomez-Soto et al. 2022). Therefore, it is possible that PtoHY5a restricts plant height by regulating GA biosynthesis and deactivation in apical buds and leaves, and promotes shoot apex growth and development by upregulating LAP1, AIL1 in shoot apex through activation of FT2. Although GA is also involved in apical bud growth, the growth retardation of apical buds caused by GA deficiency of PtoHY5a-OE plants may be compensated by the high accumulation of FT2. Indeed, our results showed that PtoHY5a-OE and PtoFT2-OE transgenic plants maintained the growth of apical bud even under SDs (Figs. 2F and 4G). Of course, the respective roles of FT2 and GA in the regulation of shoot apex development and the molecular mechanisms of their interactions need to be further investigated in the future.

The function of the PHYB2-HY5a-FT2 and HY5a-GA modules on seasonal growth in trees

Previous studies, as well as our own results, have validated the versatile direct way for regulation of FT2, the key target for winter dormancy induction, involving LHY2, GI, and HY5a, which directly bind to FT2 promoter in poplar (Böhlenius et al. 2006; Ding et al. 2018; Ramos-Sánchez et al. 2019). In our study, we characterized the function of the PtoHY5a, and demonstrated that PtoHY5a acts as downstream target of PHYB2 to regulate seasonal growth of poplar. We further confirmed that the PHYB-HY5a-FT2 module mediates photoperiod signals and thus regulates growth cessation and bud set in SDs by using both gain- and loss-of-function, and genetic complementation approaches. Additionally, we found that PtoHY5a negatively regulates bud break in poplar through controlling gibberellin (GA) levels in apical buds. Thus, PtoHY5a modulates seasonal growth in poplar through regulation of onset of winter dormancy by PtoPHYB2–PtoHY5a–PtoFT2 module, and through regulation of bud break via GA tuning. PtoHY5a also play an important role in shoot apex growth and control of plant height as well as leaf enlargement in LDs through FT2 and GA, which need to be further explored.

Here, we propose a model for how PHYB, HY5a, and FT2 as well as GA form a genetic network regulating seasonal growth in poplar (Fig. 8), expanding from previously known results, and based on this study. According to this model, long photoperiods promote the activation of PHYB, which in turn increases the expression of HY5a. HY5a regulates FT2 directly by activating the expression of FT2 and indirectly by repressing the expression of LHY2 (which in turn also activates FT2 expression). Accumulated FT2 can move from the leaf to the shoot, resulting in the upregulation of LAP1 and AIL1 in shoot apex. Upregulation of LAP1 and AIL1 promotes the shoot apex growth. At the same time, HY5a regulates GA deactivation in leaves and apical buds by both FT2-dependent and FT2-independent ways, ultimately controlling plant height. Conversely, short photoperiods weaken the activation of PHYB which in turn represses HY5a expression. SDs weakens the expression of HY5a which in turn leads to the downregulation of FT2, which ultimately may cause growth cessation and bud set. In addition, PIF8, a downstream target of PHYB, suppresses the expression of FT2, although the mechanistic links are still being unraveled. Furthermore, HY5a suppresses the increase of GA levels by regulating the expression of GA biosynthetic and catabolic genes in the apical buds after winter chilling, and then regulates bud break and growth reactivation.

Figure 8.

A model for how PHYB2-HY5a-FT2 and HY5a-GA control seasonal growth in poplars. Long photoperiods promote the expression of HY5a by activating PHYB. The downstream target of HY5a is FT2, a key positive regulator of shoot apex development. Meanwhile, HY5a further upregulates the expression of FT2 by repressing LHY2, a clock oscillator that negatively regulates FT2 and CO. The accumulation of FT2 modulates the integrator AIL1 to promote active growth in the apex. At the same time, HY5a promotes GA deactivation in leaves and apical buds by both FT2-dependent and FT2-independent ways, ultimately controlling plant height. In contrast, short photoperiodic signals negatively regulate HY5a by inhibiting PHYB2. Declining HY5a expression leads to the downregulation of FT2 in shoot apex, which ultimately may cause growth cessation and bud set. Additionally, the HY5a inhibits bud break with chilling through reducing GAs levels in shoot apex. Dotted lines reflect linkages that need to be described further, whereas solid lines represent direct genetic interactions or proven impacts on development processes. Arrows indicate activation, whereas flat-ended arrows indicate repression.

Materials and methods

Plant material and growth conditions

Poplar (P. tomentosa Carr. Clone 741) plants were used as WT and as the biological basis for plant transformation (Horvath et al. 2003; Hu et al. 2005). Rapidly propagating seedlings of P. tomentosa were cultured in glass tissue culture bottles with a woody plant medium (WPM) (PhytoTech) (Yang et al. 2021). The seedlings were rooted in the tissue culture bottle for a month, and then transplanted into small boxes with fertilized soil. P. tomentosa plants were grown in greenhouses at 22 °C under a 16 h light photoperiod using white light illumination (300 μmol·m⁻²·s⁻¹).

Arabidopsis (A. thaliana) from a Columbia (Col-0) background were used as WT. The floral dip method was used to construct complementation of the hy5-1 mutant lines and Col-0 with PtoHY5a (Clough and Bent 1998). All Arabidopsis plants (Col-0, hy5-1, PtoHY5a-OE, PtoHY5a/hy5-1) were grown in pots in a lighted incubator at 22 °C under a 16 h light photoperiod using white light illumination (120 μmol·m⁻²·s⁻¹). N. benthamiana plants were grown in pots in a growth chamber at 24 °C under a 16 h light photoperiod using white light illumination (300 μmol·m⁻²·s⁻¹) for Dual-Luciferase (Dual-LUC) assays. The rice (O. sativa) seedlings used for subcellular localization were grown in darkness at 28 °C.

Generation of vector constructs and transformation

For constructing the overexpressing vector, the full-length cDNAs of PtoHY5a, PtoPHYB2, and PtoFT2 were amplified from the cDNA of P. tomentosa with specific primers, respectively (Supplementary Table S2). The PCR products were cloned into the plant expression vector pBI121. For PtoPHYB1/2-RNAi vector construction, the fragment of PtoPHYB1/2 was amplified from the cDNA of P. tomentosa using specific primers, cloned into the pSKint vector (Guo et al. 2003), and then transferred to the final destination vector pBI121.

For pYLCRISPR/Cas9P35S-PtoHY5a vector construction, the binary pYLCRISPR/Cas9 multiplex genome targeting vector system was used to edit the poplar genome to generate mutants of PtoHY5a. The online tool CRISPR-P 2.0 (http://cbi.hzau.edu.cn/CRISPR2/) was used to design the CRISPR/Cas9 target sites of PtoHY5a. Two guide RNAs (sgRNAs) (Supplementary Table S2) were designed to the second and third exon of the coding sequence based on the on-target score and to generate a complete loss-of-function allele, respectively. The target sites of the sgRNAs were confirmed in WT poplar by sequencing. The sgRNA was synthesized by Sangon, cloned into the entry vector pYLgRNA and then assembled into the destination vector pYLCRISPR/Cas9-DH/B (Ma et al. 2015).

All the constructs above were individually transformed into an Agrobacterium (Agrobacterium tumefaciens) strain EHA105. The WT P. tomentosa was transformed by the leaf disk method to generate the PtoHY5a-OE, PtoHY5a-KO, PtoPHYB2-OE, PtoPHYB1/2-RNAi, and PtoFT2-OE lines, respectively. The generated plant transformation vectors (PtoPHYB2-OE and PtoFT2-OE) were used for transformation of PtoHY5a-KO plants, as previously described to generate PtoPHYB2-OE/PtoHY5a-KO and PtoFT2-OE/PtoHY5a-KO lines, respectively (Fan et al. 2015).

GUS analyses

For GUS analyses, the promoter sequence of PtoHY5a (1,444 bp upstream of ATG) was amplified from P. tomentosa genomic DNA with gene-specific primers (Supplementary Table S2) using PrimerSTAR Max DNA Polymerase (R045A, Takara), and the PCR products were integrated into pBI121 to replace the 35S promoter. As a result, we constructed the plasmid pBI121-ProPtoHY5a:GUS, which was transformed into P. tomentosa using the Agrobacterium-mediated approach. The transgenic plants were employed in the GUS histochemical staining experiment, following previously published approaches (Gao et al. 2020).

Subcellular localization

For subcellular localization, the coding sequence of PtoHY5a without stop codon was amplified by PCR using primers (Supplementary Table S2). The amplified fragments were cloned into the expression vector pTEX-GFP to generate pTEX-Pro35S:PtoHY5a-GFP by fusing to GFP. The recombinant vector was transformed into protoplasts of etiolated seedling from O. sativa and laser confocal microscopy (TCS SP8; Leica Wetzlar, Germany) was then used to observe protoplasts, both a green fluorescent signal at 488 nm and a red fluorescent signal at 555 nm were detected (Chen et al. 2022).

Induction of growth cessation, bud set, and bud break scoring

Poplar plants grown in a growth chamber for 2 mo under LD conditions (LD^16h, at 25 °C) and then shifted to SD conditions (SD^8h, at 25 °C) for 10 wk. After the transition to SD conditions, growth cessation and buds set were counted and recorded as previously described (Rohde et al. 2011; Ding et al. 2021). To assess cold temperature-induced bud break, poplar plants after 10 wk of SD were transferred to a cold growth chamber with short days (SD^8h, 4 °C) for 3 mo and returned to warm long days (LD^16h, 25 °C). Bud burst was monitored for at least 4 wk, and bud break was scored as previously described (Conde et al. 2017), using between two and three independent and representative transgenic lines, and five individual plants of each line.

Sample collection and gene expression analyses

To assess the diurnal pattern of gene expression, we sampled leaves every 4 h from transgenic and WT poplars grown under LD^16h and SD^8h conditions for 2 mo, respectively. For expression analysis of genes in SD-induced growth cessation, leaves and apical buds of transgenic and WT plants at different treatment times were collected under LD^16h and SD^8h conditions, respectively. For bud break samples, shoot apical buds were collected at different stages of dormancy. For seasonal transcription analysis of PtoHY5a and PtoHY5b, 2-m-tall, 2-yr-old WT P. tomentosa trees located in Mianyang, CHN (31°32′5′N, 104°42′0″E) were collected, three independent replications of leaves and shoot apical buds were sampled for summer (July), autumn (October), winter (January), and spring (April). In January, leaves on trees that had faded but not yet shed were collected. Plant tissues were ground separately in liquid nitrogen. The Biospin Plant Total RNA Extraction Kit (BioFlux) was used to obtain total RNA. The AT341 reverse transcription kit (TransGen Biotech) was used to obtain cDNA. We then used the PerfectStart Green qPCR SuperMix (TransGen Biotech) to conduct RT-qPCR analysis, calculated using the 2^−ΔΔCt methods (Livak and Schmittgen 2001). The P. tomentosa UBIQUITIN (UBQ) gene was employed as an internal reference (Hu et al. 2010). All primers used in the RT-qPCR analysis are listed in Supplementary Table S3.

Y1H library screening assays

For Y1H screening, the bait construct included about 2016bp PtoFT2 promoter sequence inserted into the pAbAi plasmid. Total RNA extracted from apical buds and leaves of P. tomentosa was mixed to construct a cDNA library. Construction methods and tests of the cDNA library were performed according to manufacturer's instructions (Clontech, Shanghai, Code No. 634901). Subsequently, the library was then transformed into yeast (Saccharomyces cerevisiae) with a bait vector and screened in SD/−Leu medium containing 100 ng·mL⁻¹ aureobasidin A (AbA). The colonies were allowed to grow at 30 °C for 3 to 4 d. The positive clones were selected and identified by colony PCR, and then the plasmids were extracted and sequenced for the positive clones.

Isolation and transfection of protoplasts from P. tomentosa

The protoplasts of P. tomentosa leaves were isolated and transfected according to the method described by Wang et al. (2021). Briefly, YL of 3-mo-old poplar trees were taken, cut into thin strips with a blade, shaken in the dark for 5 h in an enzyme solution (0.6 m mannitol, 20 mm KCl, 20 mm MES, 1.5% (w/v) Cellulase R-10, 0.75% (w/v) Macerozyme R-10, 10 mm CaCl₂, 0.05 mm β-mercaptoethanol, and 0.1% (w/v) bovine serum albumin (BSA)) and protoplasts were filtered using a 200-mesh cell filter, washed twice with W5 [154 mm NaCl, 125 mm CaCl₂, 5 mm KCl, 2 mm MES, 5 mm D-glucose], centrifuged, and the supernatant were removed. The protoplasts were resuspended to 1 × 10⁵ per mL by adding MMg solution (0.6 m D-mannitol, 15 mm MgCl₂, 4 mm MES), and the plasmid containing target gene was transfected into the protoplasts. This was then mixed with an equal volume of PEG/Ca²⁺ solution (40% (w/v) polyethylene glycol 4000 (PEG 4000), 0.2 m mannitol, 0.1 m CaCl₂). The protoplasts were incubated for 50 min at 23 °C in the dark, and then washed twice with W5 solution, after removing the supernatant, 1 mL of W5 solution was added and incubated for 16 h at 23 °C in the dark.

Promoter activation assay

The promoter sequence of PtoFT2 and PtoLHY2 (1,600 bp upstream of ATG) were amplified using P. tomentosa genomic DNA with gene-specific primers (Supplementary Table S1) and using PrimerSTAR Max DNA Polymerase (R045A, Takara), and it was then cloned into pGreenII 0800-LUC to generate a reporter plasmid. The mutant ACE-box promoter sequence was synthesized (Beijing Genomics Institute) and cloned into the pGreenII 0800-LUC vector. The full-length CDS of PtoHY5a and PtoHY5b were separately amplified and cloned into pGreenII 62-SK as effector plasmids (Hellens et al. 2005). A. tumefaciens GV3101, containing the effector and reporter plasmids, respectively, were co-injected into the leaves of N. benthamiana. Co-transformed N. benthamiana plants, which had been previously cultured in the dark for 2 d, were assayed using the Dual-Luciferase Reporter Assay System (Promega, USA) and the Berthold Centro LB960 luminometer system for assessment of Firefly and Renilla luciferase signals. Meanwhile, the effector and reporter plasmids were co-transfected into P. tomentosa protoplasts as previously described, and luciferase quantification was measured using the Dual-Luciferase Reporter Assay System (Promega, USA) (Guo et al. 2012). The relative luciferase activities are normalized to the Renilla luciferase activity (Relative LUC:REN activity). All luciferase quantification assays were performed with three biological replicates.

Gene expression analysis in poplar protoplasts

The PtoHY5a and PtoHY5b coding regions were cloned into the pGreenII 62-SK destination vector, respectively. The recombinant plasmid was transformed into protoplasts isolated from young poplar leaves. After 16 h of cultivation, the protoplasts were collected for RNA extraction and RT-qPCR analysis (Liu et al. 2022).

Protein extraction and immunoblotting

Plant leaves were ground into a fine powder in liquid nitrogen, and then add extraction buffer (0.44 m sucrose, 1.25% (w/v) polyethylene glycol 400, 2.5% (w/v) Dextran T40, 20 mm, 0.5% Triton X-100, 5 mm DTT, 1 mm PMSF, 1% (v/v) protease inhibitor cocktail 100×). The mixture was incubated and subsequently filtered through a Miracloth membrane. Next, the filtrate was centrifuged at 2,000 × g for 15 min at 4 °C, and the resulting pellet was resuspended in 1 mL of extraction buffer. This was centrifuged at 1,500 × g for 15 min, the pellet resuspended, the centrifugation process was repeated until the pellet became transparent. We added three times the volume of the precipitated sample to the loading buffer, heated at 95 °C for 15 min, and then centrifuged at 13,400 × g for 15 min. We collected the supernatant which was to be used as the nuclear protein sample.

For the immunoblot analysis, a protein-specific polypeptide QEQTTSSIPANSLPS (aa 2 to 16) from the N-terminal of PtoHY5a (Supplementary Fig. S14A) was synthesized and injected into rabbits to obtain polyclonal antibodies by ABclonal Technology Co., Ltd. The A. tumefaciens GV3101 containing recombinant plasmids with PtoHY5a or PtoHY5b, driven by the 35S promoter with HA-tag, were injected into N. benthamiana leaves, respectively, and immunoblotting was performed using PtoHY5a antibodies (ABclonal) (diluted at 1:1,000) to determine antibody specificity (Supplementary Fig. S14B). Nuclear proteins were separated by 10% SDS-PAGE and then transferred to an Immobilon-P PVDF membrane (Millipore, IPVH00005) by wet transfer. The membranes were blocked with 5% nonfat milk in TBST buffer for 1 h at room temperature, and incubated with an anti-PtoHY5a (1:1,000 [v/v]) or anti-HA (1:5,000 [v/v], BBI Life Sciences, catalog no. D110004) antibody for 2 h, anti-Actin (1:5,000 [v/v], CWBIO, catalog no. CW0096M) was used as a loading control. The membranes were then washed three times with TBST buffer and incubated with the secondary antibody for 2 h. The Goat anti-mouse immunoglobulin G(IgG)–horse-radish peroxidase (1:6,000 [v/v], HRP; Thermo Fisher Scientific, catalog no. 31430) was used as the secondary antibody in the immunoblot assays with anti-Actin antibodies, and Goat anti-rabbit IgG-HRP (1:8,000 [v/v], Thermo Fisher Scientific, catalog no.31460) was used as the secondary antibody in the immunoblotting assays with anti-PtoHY5a and anti-HA antibodies. The membranes were washed three times with TBST buffer and visualized using a chemiluminescence detection kit (US EVERBRIGHT).

Electrophoresis mobility shift assays (EMSA)

EMSA were performed as previously described (Zhang et al. 2021). Briefly, the MBP-PtoHY5a protein was expressed in Escherichia coli and purified using MBP Sepharose beads. The DNA fragment (Supplementary Table S4) was incubated with MBP-PtoHY5a in EMSA binding buffer (25 mm HEPES-KOH, pH 8.0, 50 mm KCl, 1 mm dithiothreitol (DTT), and 10% glycerol). An Electrophoretic Mobility Shift Assay Kit with SYBR Green and SYPRO Ruby EMSA stains (E33075, Thermo Fisher) was then used to detect protein–DNA interactions.

ChIP-qPCR assay

ChIP was performed as described by Deng et al. (2022). Chromatin extracts were obtained from the leaves of the PtoHY5a-OE and WT poplar. All samples were fixed with 1% (v/v) formaldehyde. The chromatin was sonicated with a Bioruptor system (Q800R2; Qsonica, Newtown, CT, USA) with 30% power setting for 10 s “ON” and 10 s “OFF” for 11 min to shear DNA to an average length of 200 to 800 bp. Chromatin was then immunoprecipitated by the addition of protein A agarose beads and rabbit antibodies against HA (D110004; BBI Life Sciences) or anti-PtoHY5a (ABclonal) and incubated at 4 °C overnight. After the incubation, the beads were washed in the following buffers: once in a low salt buffer (150 mm NaCl, 0.1% [w/v] SDS, 1% [v/v] Triton X-100, 2 mm EDTA, 20 mm Tris-HCl pH 8.0), once in a high salt buffer (500 mm NaCl, 0.1% [w/v] SDS, 1% [v/v] Triton X-100, 2 mm EDTA, 20 mm Tris-HCl pH 8.0), once in a LiCl buffer (250 mm LiCl, 1% [w/v] sodium deoxycholate, 1% [v/v] NP-40, 1 mm EDTA, 10 mm Tris-HCl pH 8.0), and twice in a TE buffer (1 mm EDTA, 10 mm Tris-HCl pH 8.0). ChIP assays were performed with three biological replicates. The crosslinking was then reversed by incubating at 65 °C overnight and then incubating with 50 µg·mL⁻¹ Proteinase K (Cell Signaling Technologies; #10012S) at 65 °C for 3 h. The immunoprecipitated chromatin was analyzed by qPCR using specific primer sets (Supplementary Table S4).

Gibberellin quantification

GA contents were determined by MetWare's (http://www.metware.cn/) AB Sciex QTRAP 6500 + LC-MS/MS platform. After the low temperature treatment, the apical buds were sampled after 2 wk under LD^16h and warm temperatures (16 h:8 h at 25 °C, LD^16h; 2WW). Freshly harvested plant samples were instantly frozen, ground (30 Hz, 1 min) to a fine powder with a bead mill (MM400; Retsch, Haan, Germany), and stored at −80 °C. Fifty milligram of sample was placed in a 2 mL microtube and mixed with 1 mL MeOH/H₂O/formic acid (15:4:1, v/v/v). An internal standard (IS) solution (10 ng·mL⁻¹) was added for quantification. The supernatant was moved to new microtubes and dried. To the residue, 500 μL H₂O (3.5% [v/v] formic acid) and 1 mL ethyl acetate was added. Samples were vortexed and centrifuged (5 min, 12,000 × g, 4 °C). The supernatant was transferred to a brown vial, and 500 μL ethylacetate was added for re-extraction. After centrifugation (5 min, 12,000 × g, 4 °C), the supernatants were combined and evaporated. The resulting residue was dissolved in acetonitrile (ACN), and 10 µL each of TEA and BPTAB were added. The solution was vortexed, heated at 90 °C for 1 h, and dried under nitrogen gas. The residue was re-dissolved in 100 µL ACN/H₂O (90:10, v/v) and filtered through a 0.22 µm membrane for subsequent LC-MS/MS analysis. The data acquisition instrument system mainly includes ultrahigh performance liquid chromatography (Ultra Performance Liquid Chromatography, UPLC) (ExionLC AD, https://sciex.com.cn/) and tandem mass spectrometry (Tandem Mass Spectrometry, MS/MS) (QTRAP 6500+, https://sciex.com.cn/).

Amino-acid sequence analysis and phylogenetic tree construction

The Protein BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/) was used to obtain putative orthologs of Arabidopsis HY5. The amino-acid secondary structure of PtoHY5a was predicted using the Simple Modular Architecture Research Tool (SMART) software program (http://smart.embl-heidelberg.de). We used MEGA software v.X and the Neighbor-Joining method to build a phylogenetic tree. To assess the reliability of the phylogeny, we conducted bootstrap analysis with 1,000 replicates (Kumar et al. 2018).

Statistical analyses

GraphPad Prism and SPSS software were used for statistical analyses. The experimental data were statistically analyzed using three or more averages. The significance of the differences between groups was determined by a two-tailed Student's t-test (*P < 0.05; **P < 0.01; ***P < 0.001; ns, not significant). For multiple comparisons, using one-way or two-way analysis of variance (ANOVA), ANOVA testing followed by Fisher's LSD test (P < 0.05). Statistical data are provided in Supplementary Data Set 1.

Accession numbers

Sequence data from this article can be found in Phytozome Database 13 under the following accession numbers: AtHY5 (AT5G11260), PtoHY5a (Potri.018G029500), PtoHY5b (Potri.006G251800), PtoPHYB1 (Potri.008G105200), PtoPHYB2 (Potri.010G145900), PtoFT2 (Potri.010G179700), PtoLHY2 (Potri.014G106800), PtoLAP1 (Potri.008G098500), PtoAIL1 (Potri.002G114800), PtGA2ox1 (Potri.001G378400), PtGA2ox3 (Potri.004G065000), PtGA2ox4 (Potri.008G101600), PtGA2ox5 (Potri.010G149700), PtGA2ox6 (Potri.011G095600), PtGA2ox7 (Potri.014G117300), PtGA3ox1 (Potri.001G176600), PtGA3ox2 (Potri.003G057400), PtGA20ox2-1 (Potri.002G151300), PtGA20ox3 (Potri.005G184400), PtGA20ox5 (Potri.007G103800), PtGA20ox6 (Potri.012G132400), PtGA20ox8 (Potri.015G134600), PtoUBQ (Potri.014G115100). Sequence data for the phylogenetic tree can be found in the NCBI database with accession numbers as follows: CsHY5 (XP_010419684.1), BpHY5 (AHY20043.1), MtHY5 (XP_013459310.1), ZjHY5 (XP_015885857.1), CmHY5 (NP_001284656.1), MnHY5 (XP_010110356.1), MdHY5 (BAM71071.1), PbHY5 (XP_009355719.1), FvHY5 (XP_004291469.1), FaHY5 (AKG58815.1), PmHY5 (XP_008219477.1), PpHY5 (ONI34365.1), TcHY5 (XP_007013841.2), DcHY5 (XP_017229054.1), CcHY5 (XP_006450470.1), VvHY5 (XP_010648648.1), JcHY5 (XP_012076602.1), RcHY5 (XP_002515537.1), EgHY5 (XP_010048982.1), PtrHY5a (XP_002324289.1), PtrHY5b (XP_002308656.1), CaHY5 (APD29065.1), NnHY5 (XP_010250037.1), GrHY5 (AIC64080.1), SiHY5 (XP_011081579.1), NaHY5 (XP_019265660.1), StHY5 (XP_006361723.1), SlHY5 (NP_001234820.1), PdHY5 (XP_008785002.1), AcHY5 (XP_020097860.1), ThHY5 (XP_010541629.1), RsHY5 (XP_018445811.1), BrHY5 (XP_009121971.1), NcHY5 (JAU18721.1), EsHY5 (XP_006399627.1).

Author contributions

Y.G., Z.C., investigation and writing; Q.F., T.L., P.S., H.D., S.L., P.Y., W.T., investigation and formal analysis; J.D., H.D., L.G.R., reviewing; V.R., reviewing and project administration; L.W., project administration; Y.Y., conceptualization, reviewing, project administration.

Supplementary data

The following materials are available in the online version of this article.

Supplementary Figure S1. ACE-box cis-element analysis of FT2 promoter in different Populus species.

Supplementary Figure S2. Phylogenetic and alignment analyses of HY5 proteins in poplar and other plants.

Supplementary Figure S3. Expression patterns of PtoHY5a and PtoHY5b and their functional analysis in binding to the FT2 promoter.

Supplementary Figure S4. Spatio-temporal expression patterns and subcellular localization analysis of PtoHY5a.

Supplementary Figure S5. RT-qPCR and immunoblots identifying the overexpression of PtoHY5a transgenic lines.

Supplementary Figure S6. Production of transgenic poplar with CRISPR/Cas9 knockout of PtoHY5a.

Supplementary Figure S7. Morphological phenotypes of PtoHY5a-OE and PtoHY5a-KO transgenic poplar.

Supplementary Figure S8. PtoHY5a influences poplar vegetative development and rescues the early flowering phenotype of Arabidopsis hy5 mutant.

Supplementary Figure S9. Expression levels of PtoLAP1 and PtoAIL1 in PtoHY5a-OE and PtoHY5a-KO transgenic plants.

Supplementary Figure S10. The expression analysis of PtoHY5a and PtoFT2 in WT and different transgenic lines.

Supplementary Figure S11. Phenotypes of early flowering in PtoFT2-OE and PtoHY5a-OE transgenic poplar trees.

Supplementary Figure S12. Changes in the expression level of PtoPHYB2 affect bud set in poplars induced by SDs.

Supplementary Figure S13. The expression analysis of key genes for GA deactivation and biosynthesis in apical buds and leaves of PtoHY5a transgenic plants under LDs.

Supplementary Figure S14. Specific detection of PtoHY5a antibody.

Supplementary Table S1. List of candidate genes for binding to PtoFT2 promoter screened by Y1H assays.

Supplementary Table S2. Primers for vector construction.

Supplementary Table S3. Primer sequences of related genes used in RT-qPCR.

Supplementary Table S4. Primer sequences of related genes used in EMSA and ChIP-qPCR.

Supplementary Data Set 1. Summary of statistical tests.

Funding

This research was supported by the National Natural Science Foundation of China (31770644, 42350610258, and 31901333).

Data availability

The data used in this article can be found in the article itself and its online supplementary materials.

Dive Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

• PHYB Gramene: AT2G18790

• PHYB Araport: AT2G18790

• PIF4 Gramene: AT2G43010

• PIF4 Araport: AT2G43010

• HY5 Gramene: AT5G11260

• HY5 Araport: AT5G11260

• PHYA Gramene: AT1G09570

• PHYA Araport: AT1G09570

• TOC1 Gramene: AT5G61380

• TOC1 Araport: AT5G61380

• GA3ox2 Gramene: AT1G80340

• GA3ox2 Araport: AT1G80340

• GIS Gramene: AT3G58070

• GIS Araport: AT3G58070

• CCA1 Gramene: AT2G46830

• CCA1 Araport: AT2G46830

• COP1 Gramene: AT2G32950

• COP1 Araport: AT2G32950

• GA2ox6 Gramene: AT1G02400

• GA2ox6 Araport: AT1G02400

• GA2ox7 Gramene: AT1G50960

• GA2ox7 Araport: AT1G50960

• ELF3 Gramene: AT2G25930

• ELF3 Araport: AT2G25930

• GA2ox3 Gramene: AT2G34555

• GA2ox3 Araport: AT2G34555

• AtHY5 Gramene: At5g11260

• AtHY5 Araport: At5g11260