Potential function of CbuSPL and gene encoding its interacting protein during flowering in Catalpa bungei

Background “Bairihua”, a variety of the Catalpa bungei, has a large amount of flowers and a long flowering period which make it an excellent material for flowering researches in trees. SPL is one of the hub genes that regulate both flowering transition and development. Results SPL homologues CbuSPL9 was cloned using degenerate primers with RACE. Expression studies during flowering transition in “Bairihua” and ectopic expression in Arabidopsis showed that CbuSPL9 was functional similarly with its Arabidopsis homologues. In the next step, we used Y2H to identify the proteins that could interact with CbuSPL9. HMGA, an architectural transcriptional factor, was identified and cloned for further research. BiFC and BLI showed that CbuSPL9 could form a heterodimer with CbuHMGA in the nucleus. The expression analysis showed that CbuHMGA had a similar expression trend to that of CbuSPL9 during flowering in “Bairihua”. Intriguingly, ectopic expression of CbuHMGA in Arabidopsis would lead to aberrant flowers, but did not effect flowering time. Conclusions Our results implied a novel pathway that CbuSPL9 regulated flowering development, but not flowering transition, with the participation of CbuHMGA. Further investments need to be done to verify the details of this pathway.


Background
Flowers allow flowering plants to have a broader evolutionary relationship and extend their ecological niche so that they can dominate the terrestrial ecosystem. Flowering is extremely important for the development of perennial woody plants and for improving the economic value of plants. However, due to complex genomes and other objective characteristics of perennial woody plants, research on the flowering process in perennial woody plants remains limited. Catalpa bungei is valuable as both a timber and an ornamental tree [1]. "Bairihua", which is a natural variety of C. bungei, has been characterized for its especially short juvenile period, large number of flowers and long flowering period. The flowering period of "Bairihua" is approximately 15 days, and its accumulative flowering period reaches 100 days, which is very rare for woody plants (http://www.forestry.gov.cn/). "Bairihua" provides an excellent opportunity to evaluate the flowering process of woody plants.
In an effort to study the molecular mechanism of the flowering process in "Bairihua", we evaluated whether there were protein interactions involved in SPL regulation during the flowering process. As a first step to address this question, we isolated and characterized the SPL9 orthologous gene from "Bairihua" and performed native in planta gene expression analysis. The putative function of this gene was then tested by ectopic expression experiments. This gene showed similar expression patterns and the ability to induce flower organ development and early flowers in Arabidopsis. The findings of this study indicated that CbuSPL9 is functionally conserved. An architectural TF, CbuHMGA, was found via screening CbuSPL9interacting proteins. CbuHMGA is involved in the floral organ development of "Bairihua"', but not in the regulation of flowering time. These results provide a molecular basis for studying the molecular mechanism of "Bairihua" flowering and provide a research direction for the study of the floral transition of perennials.

Plant materials
C. bungei is a perennial tree that typically flowers over a 30-day flowering period. However, "Bairihua", the new variety of C. bungei, was found in Henan Province, China, and confirmed by the National Forestry and Grassland Administration (http://www.forestry.gov.cn/). The flowering period of a single flower is approximately 15 days, and its accumulative flowering period reaches 100 days, which is very rare for woody plants. From January 15 to April 2, 2017, we collected the first round of axillary buds of FB and LB varieties. Samples were collected every 10 days during the dormant and germinating periods. Since the floral transition of "Bairihua" was completed within 7 to 10 days, samples were collected every day during the floral transition period, and the samples were collected every 5 days during the reproductive growth stage. The samples used for RNA extraction were washed with distilled water, frozen immediately in liquid nitrogen, and stored at − 80°C. Samples for histological analysis were fixed in a formalin:glacial acetic acid:70% ethanol (5:5:90 vol.; FAA) solution under a vacuum for at least 24 h. All the plant material collections in our study were complied with national guidelines. The field experiment we made were in accordance with local legislation. The voucher specimens were deposited in Research Institute of Forestry, Chinese Academy of Forestry. Dr. Wenjun Ma and Dr. Junhui Wang undertook the formal identification of the samples.

Histological analysis
For histological analysis, the samples were immersed in FAA fixative and placed under vacuum at 4°C. Samples were dehydrated in gradient ethanol and then embedded in paraffin. Ten-mm-thick sections (RM2255 Fully Automated Rotary Microtome; Leica, Germany) were stained with Safranine O and fast green FCF (Sigma-Aldrich, USA). The slices were observed and photographed using a Leica DM 6000B fully automated upright microscope (Leica Microsystems GmbH, Wetzlar, Germany).
Cloning CbuSPL9 and CbuHMGA sequences from "Bairihua" RNA from "Bairihua" was extracted from the buds of the mutant using an RNA extraction kit (TaKaRa), and contaminating DNA was removed with RNase-free DNase I (TaKaRa). One or two micrograms of total mRNA template was added to Oligod (dT)18 primer and reversetranscribed into single-stranded cDNA by M-MLV RTase (TaKaRa). The full-length cDNA of CbuSPL9 was cloned via 3′-RACE and 5′-RACE by using the (TaKaRa) according to the manufacturer's instructions. In 3′-RACE, the CbuSPL9 gene-specific forward primers F1/F2 were designed based on published and aligned SPL sequences from NCBI (http://blast.ncbi.nlm.nih.gov/ Blast.cgi). F1 nested immediately upstream of F2. The 3′-cDNA synthesis primer was provided in a kit. The PCR products were cloned into the PMD18-T vector and sequenced. CbuSPL9s were identified using BLAST. In 5′-RACE, the gene-specific reverse primers R1/R2 (Additional file 1) were designed based on sequences from 3′-RACE. R2 nested immediately upstream of R1. The PCR products were cloned into the PMD18-T vector and sequenced.
Yeast two-hybrid (Y2H) screening of the "Bairihua" cDNA library The equal amounts of "Bairihua" buds (Collected the first round of axillary buds of "Bairihua") from each period (Dormant period, Germination period, Floral transition period and Reproductive growth period) were pool together. Total RNA from the "Bairihua" buds from four periods, was mixed to provide the mRNA sample. A yeast library was constructed by Oebiotech (Shanghai, China) via cloning the full-length cDNA library from mRNAs sample into the pGADT7 vector. CbuSPL9 was inserted into the pGBKT7 vector. Screening of the yeast library was performed using pGBKT7-CbuSPL9. The corresponding primers are listed in Additional file 1. Transformed cells were grown on SD medium supplemented with Trp. The transformants were screened on supplemented SD medium lacking Leu, Trp, His and Ade and supplemented with X-a-Gal and Aureobasidin A. Plates were incubated at 30°C for 48-72 h and photographed. The positive clones were verified with both pGBKT7 vector consensus primers. Positive control mating was as follows: pGADT7-T in Y2HGold and pGBK-53 in Y187. Negative control mating was as follows: pGADT7-T in Y2HGold and pGBKT7-Lam in Y187.

Subcellular localization
To validate subcellular localization, the full-length coding sequences (without the stop codon) of CbuSPL9 and CbuHMGA were amplified from RNA of "Bairihua" buds by RT-PCR. The PCR products of CbuSPL9 and CbuHMGA were ligated to the vector pCAMBIA1304 using the Seamless Assembly Cloning Kit (CloneSmarter, Beijing, China) to construct the CbuSPL9/CbuHMGA-GFP fusion genes driven by a CaMV35S promoter. pCAMBIA1304-GFP was used as a positive control. The transient expression vectors CbuSPL9-GFP, CbuHMGA-GFP and GFP-HDEL were injected into the leaf lower epidermal cells of Nicotiana tabacum L. using Agrobacterium transformation. The transformed cells were incubated for 2 days. The leaves were removed, cut into squares and immersed in PBS buffer containing 1 g mL − 1 DAPI to stain the nuclei. The transient expression of the CbuSPL9/ CbuHMGA-GFP fusion proteins was observed under an UltraVIEW VoX 3D Live Cell Imaging System Spinning Disk confocal laser scanning microscope (PerkinElmer, Waltham, MA, USA). The wavelength of excitation used was 488 nm for GFP and 405 nm for DAPI.

BiFC analysis
To confirm and visualize the interaction between CbuSPL9 and CbuHMGA in protoplasts from Populus trichocarpa, a BiFC assay was performed based on split EYFP. EYFP was fused to the C-terminus of CbuSPL9 and the Nterminus of CbuHMGA, resulting in CbuSPL:EYFP C and CbuHMGA:EYFP N . A positive EYFP signal indicates an interaction of EYFP C and EYFP N due to the heterodimerization of CbuSPL9 with CbuHMGA. CbuSPL:EYFP C was cotransfected with CbuHMGA:EYFP N and H2A: mCherry into protoplasts. The transient expression of the CbuSPL9/CbuHMGA-GFP fusion proteins was observed under an UltraVIEW VoX 3D Live Cell Imaging System Spinning Disk confocal laser scanning microscope (Perki-nElmer, Waltham, MA, USA).

Biolayer interferometry assay
CbuSPL9 was cloned into pGEX6P-1 as a C-terminal GST-tagged construct, and the construct was confirmed by sequencing. CbuHMGA was cloned into pET28a as an N-terminal 6His-tagged construct, and the construct was confirmed by sequencing. The protein was purified using the procedure for EMCV-3C and RV-3C as described above. Real-time interactions between CbuHMGA and CbuSPL9 were monitored with an Octet QK (Forte-Bio) that is based on BLI. BLI was used to determine dissociation constants (K D ) as well as the on-and off-rate (k on and k off ) for HIS-CbuHMG binding to GST-CbuSPL1.

Bioinformatic analysis
Multiple sequence alignment and phylogenetic analysis were performed using MEGA6.0. After alignment, the evolutionary history was calculated using the neighborjoining (NJ) method. The tree was inferred from 1000 bootstrap replicates to show the evolutionary history of the genes. The MEME online tool (http://meme-suite. org/tools/meme) was used to identify the motifs of the CbuSPL9 protein. MEME was run locally with the following parameters: number of repetitions = any and maximum number of motifs = 20. All candidate interacting protein sequences were examined by the domain analysis program SMART (Simple Modular Architecture Research Tool) (http://smart.embl-heidelberg.de/). CbuSPL9 was submitted to the psRNATarget server (http://plantgrn.noble.org/psRNATarget/) with an expectation value < 3. CbuSPL9 containing no more than four mismatches and G/U pairs within the complementary regions were considered miRNA targets.

RNA extraction and quantitative real-time PCR
Total RNA was isolated from "Bairihua" buds at different developmental stages. The purity and quality of RNA was checked by NanoDrop8000 (Thermo Fisher Scientific, Waltham, MA, USA) and analyzed by gel electrophoresis. First-strand cDNA synthesis was carried out with~1 μg RNA using the SuperScript III reverse transcription kit (Invitrogen) and random primers according to the manufacturer's instructions. Primers were designed using Primer 3 online. The melting temperature of the primers was 60°C, and the amplicon lengths were 100-200 bp. All primers are listed in Additional file 1. Real-time qRT-PCR was performed on a Roche Light-Cycle 480 Real-Time PCR System (Roche Applied Science, Germany) using a SYBR Premix Ex Taq™ Kit (TaKaRa, Dalian, China) according to the manufacturer's instructions using a 20 μl reaction volume (10 μl of 2× SYBR Premix, 2 μl of cDNA templates as prepared above and 1 μl of each specific primer to a final concentration of 200 nM) [43]. Relative expression levels were calculated using the 2 -ΔΔCt method. Actin was used as an internal control, and each reaction was conducted in triplicate [1]. The stem expression values were set to 1. U6 were amplified as an endogenous control [44]. All the primers are shown in Additional file 1. We test the correlation of expression (CEG) between CbuSPL9/miR156 by using the Pearson correlation coefficient. The Pearson correlation coefficient was calculated by COR() using average relative expressions of three replicates in R. A Pearson correlation is a number between − 1 and 1. 0 represents no correlation, negative value is negative correlation, positive value is positive correlation [45].

Transient overexpression in Arabidopsis thaliana
Full-length CbuSPL9 and CbuHMGA were cloned into the binary vector pBI121 (BD Biosciences Clontech, USA) under the control of the cauliflower mosaic virus 35S promoter in the sense orientation. The transgenic plants were generated with the 35S:CbuSPL9 and 35S: CbuHMGA constructs via Agrobacterium tumefaciens GV3101 by using the floral dip method. Surfacesterilized T1 seeds were grown on a solid 0.5 × MS medium containing 30 μg mL − 1 hygromycin at 4°C for 2 days, which were then transferred to the greenhouse under long-day conditions (16 h light/8 h dark) at 22°C for 10 days. Subsequently, the seedlings were transplanted into soil. Phenotypes of the transgenic plants were observed in the T1 generation, and the overexpression of CbuSPL9 and CbuHMGA in the transgenic plants was confirmed by PCR genotyping (Additional files 2 and 3). For each construct, at least 10 transgenic lines with similar phenotypes were observed, and 3 of them were used for detailed analysis [1].

Flowering time measurement
Plants were grown on soil in the greenhouse under longday condition (16 h light/8 h dark) at 22°C. Flowering time was measured by counting the total number of rosette leaves and the number of days (from sowing to floral bud formation). The numbers of rosette leaves of 20 plants were counted and averaged. Data were classified with Win-Excel and analyzed via analysis of variance (ANOVA) using the SPSS (version 8.0, SPSS Inc., Chicago, IL, USA) statistical package. Comparisons between the treatment means were made using Tukey's test at a probability level of P < =0.05 [46,47].

Results
A flowering-related cDNA of SPL was cloning from "Bairihua" Full-length cDNAs of SPL from C. bungei were isolated by using homology-based cloning and RACE techniques. The isolated conserved sequences were extended by genome walking to acquire the full genomic sequences (exons and introns). One of the isolated sequences encoded a protein with a typical SPL protein structure, which included a highly conserved SBP-box domain bearing two zinc-binding sites and one bipartite nuclear localization signal [6,19,48] (Additional file 4). The first Zn-finger-like structure (ZN-1 in Fig. 1a) was C3H-type, and the second (Zn-2 in Fig. 1a) was C2HC-type. The nuclear localization sequence (NLS) is a highly conserved bipartite domain located at the C-terminus of SBP. Phylogenetic analysis showed that the isolated sequence clustered with AtSPL9 and AtSPL15 (Fig. 1b). Since it shared more sequence similarity with AtSPL9, we named the SPL homologous gene CbuSPL9. And CbuSPL9 was predicted to be a target of miR156 (Additional file 5), which is consistent with previous reports of SPL in Arabidopsis involving in flowering process .

Expression of CbuSPL9 during the flowering process
Intensive sampling was performed to investigate the changes in CbuSPL9 expression during the flowering process (Fig. 2). The determination of T1-T3 is based on the phenological of "Bairihua", the current state of the buds, and the morphology of the apical meristem. The buds in T1-T3 peroids is covered with a thick layer of hard bud scales, and its morphology has no significant differences to observed between the flowering buds (FBs) and the leaf buds (LBs). T4-T5 was the germination period. The germinating period of "Bairihua" generally begins in early march, when the buds gradually strip off the hard bud scale, and reveals the soft reddish brown tip.
During this period, the internal morphology was similar between the FBs and LBs. T6-T9 followed the short germination period and was the floral transition period, during which flower primordium and leaf primordium were developed in the FBs and LBs, respectively. Finally, T10-T12 was the reproductive growth period. The expression level of CbuSPL9 was significantly higher in the FBs than in the LBs, especially during the dormant period and the reproductive growth period. The classical model related flowering governed by microRNA 156 (miR156) and its targeted SPLs [35,49,50]. During flowering process of "Bairihua", the expression levels of miR156 was significantly higher in the LBs than in the FBs, and gradually decrease with age. Based on the overall trend, CbuSPL9 and miR156 exhibited negative correlations with r < − 0.8 (Additional file 6). Much higher levels of expression in FBs compared to LBs during T6-T12 periods was observed for AtSPL homologous target genes. The results showed that CbuAP1 and CbuLeafy are positive correlation with CbuSPL9 from T6 period to T12 period in FBs (Additional file 7). Overall, the expression study supported that CbuSPL9 was an SPL homologous gene and involved in flowering regulation.

The overexpression of CbuSPL9 in Arabidopsis
As there is no available transformation system in C. bungei, CbuSPL9 was overexpressed in Arabidopsis (Columbia ecotype, col). The flowers from col were tetradynamous and had four petals distributed in cross type (Fig. 3aI). In contrast, the oe-spl9 transgenic plants showed aberrant flower organs (Additional file 8). The oe-spl9 transgenic lines exhibited flower organs with altered numbers and locations, such as shrunken petals, increased stamens, and overlapping petals (Fig. 3aII-VI). In addition to the evident changes in floral organ morphology, an acceleration in flowering time was observed in the oe-spl9 lines (Fig. 3b, Additional file 9). The col line initiated flowering when 14 rosette leaves were present (Fig. 3c). However, oe-spl9 possessed less than 9 rosette leaves at the time of bolting (Fig. 3d). This result indicated that the function of CbuSPL9 was conserved. The regulatory mechanism of CbuSPL9 in "Bairihua" might be similar to that in Arabidopsis.

Screening CbuSPL9-interacting proteins
As a TF, SPL-DNA interaction studies have been extensively performed. However, TFs can also fine-tune specific biological processes through protein interactions. We constructed a C. bungei yeast two-hybrid cDNA library to explore the proteins that interact with CbuSPL9. A total of 809 blue colonies representing potential positive clones were obtained on QDO/ Aba/X-a-Gal plates. The potential positive clones were subsequently tested by PCR for the library plasmid, and 406 clones were positive. The resulting PCR products were sequenced, and the sequences were aligned using the NCBI BLASTp search function. Finally, 12 interacting protein candidates were identified (Table 1). These predicted proteins included HMGA, aquaporins, bHLH48, GATA-related protein, PHD finger protein ALFIN-LIKE 4, heavy metalassociated protein, etc. Notably, even though floral transition was never observed in the leaf buds, the leaf buds collected in the period corresponding to floral transition are henceforth called F2 and R2 for convenience. Error bars indicate SD from three independent biological replicates. *Difference between flowering buds (black) and leaf buds (gray) is significant (Student's test; p < 0.05). **Difference between flowering buds (black) and leaf buds (gray) is highly significant (Student's test; p < 0.01) Cloning the HMGA gene Among these candidates, HMGA, which is an AT-hook rich protein, belongs to the most abundant nonhistone protein family in the nucleus. We cloned the full-length HMGA homologous gene via rapid amplification of cDNA ends. Phylogenic analysis revealed that this protein has a close relationship with AtHMGA. Therefore, this protein was renamed CbuHMGA. CbuHMGA encoded a H15 domain and 4 AT-hook motifs (Fig. 4a).
The expression analysis showed that the expression trend of CbuHMGA was similar to that of CbuSPL9 in the four periods in "Bairihua" (Fig. 4b). The expression of CbuHMGA was elevated in the dormant period, and the highest expression level was detected in the reproductive growth period. However, the increase in the expression of CbuHMGA before the reproductive growth period occurred slightly earlier than that of CbuSPL9. Notably, even though floral transition was never observed in the leaf buds, the leaf buds collected in the period corresponding to floral transition are henceforth called the floral transition period and reproductive period for convenience. Error bars indicate SD from three independent experiments. *Difference between flowering buds (black) and leaf buds (gray) is significant (Student's test; p < 0.05). **Difference between flowering buds (black) and leaf buds (gray) is highly significant (Student's test; p < 0.01)

Localization of the CbuSPL9 and CbuHMGA proteins
We fused GFP at the C-terminus of CbuSPL9 and CbuHMGA and transformed them into Nicotiana benthamiana leaf epidermal cells to determine the localization of the CbuSPL9 and CbuHMGA proteins. 35S:GFP was used as a control. The GFP fluorescence in CbuSPL9-GFP and CbuHMGA-GFP was exclusively observed in the nucleus, whereas the fluorescence of GFP in the control was distributed throughout the entire cell (Fig. 5a). These results showed that CbuSPL9 and CbuHMGA were located in the nucleus. This result was in agreement with the prediction based on the protein structure.

Protein interaction analysis
Biolayer interferometry (BLI) [51] was used to determine dissociation constants (K D ) as well as the on-and offrate (k on and k off ) for HIS-CbuHMGA binding to GST-CbuSPL9 (Table 2). Five different concentrations of GST-CbuSPL9 (2381.00 nm, 1191.00 nm, 595.30 nm, 297.60 and 148.80 nm) were evaluated, and the K D was 311 nm for each tested concentration. These results suggested a strong interaction between CbuSPL9 and CbuHMGA (Fig. 5b).
To confirm these interactions, the full-length cDNA of CbuSPL9 was inserted into the vector pGBKT7 (BD-CbuSPL9) as bait, and the full-length cDNA of CbuHMGA was inserted into the vector pGADT7 (AD-CbuHMGA) as prey. Yeast strains containing AD-CbuHMGA and BD-CbuSPL9 were positive for X-α-gal activity when grown on synthetically defined (SD)/−Trp/−Leu/−His/−Ade medium (Fig. 5c). These results showed that CbuSPL9 interacted with HMGA in yeast. Finally, a bimolecular fluorescence complementation (BiFC) assay was performed. CbuSPL: EYFP C was cotransfected with CbuHMGA:EYFP N and H2A:mCherry into protoplasts. The signal of enhanced yellow fluorescent protein (EYFP) was colocalized with mCherry ( Fig. 5d). Collectively, we demonstrated that CbuSPL9 formed a heterodimer with CbuHMGA in the nucleus.

The overexpression of CbuHMGA in Arabidopsis
To characterize the potential function of CbuHMGA, we overexpressed CbuHMGA in Arabidopsis. The oe-hmga transgenic lines developed aberrant flowers (Fig. 6aI-VI). This mutant phenotype was similar to the phenotype of floral organs when CbuSPL9 was overexpressed (Additional file 10). However, flowering time was not affected in the oe-hmga transgenic lines compared to that in the col line (Additional file 11). The expression of endogenous AtSPL9 was thoroughly studied to further monitor flower development in the transgenic lines. Within the sampling period, the expression level of AtSPL9 continuously increased in oe-hmga and wild type. However, in oe-spl9, the expression of AtSPL9 reached a peak in T6 and then decreased (Fig. 6b). The shift in AtSPL9 expression in oe-spl9, but not in oe-hmga, further confirmed the observation that CbuHMGA cannot accelerate flower development, while CbuSPL9 can accelerate flower development.

Discussion
Flowering is a very complex process, a qualitative change in the life history of higher plants, and a central link in plant development. "Bairihua", a variety of the flowering perennial woody plant C. bungei, has a large amount of flowers and a long flowering period. "Bairihua" is an excellent material for evaluating the flowering process in trees.
SPLs play an important role in regulating flowering in many plants, most notably in Arabidopsis. In woody plants, SPLs are studied extensively, but limited information about flowering is known [25,26,28,52,53]. Here, we cloned the SBP-domain-encoding gene CbuSPL9 in C. bungei. Its conserved structure, flower developmentrelated expression trend and the correlations with miR156, CbuAP1 and CbuLeafy suggest that CbuSPL9 might have conserved SPL9 functions in C. bungei. Furthermore, heterogeneous overexpression of CbuSPL9 in Arabidopsis accelerated flower development and lead to aberrant flower organs.
TFs can affect biological processes by forming complexes with other TFs. However, most of the studies on SPLs have focused on their interactions with DNA motifs [36][37][38]. The proteins that could directly interact with SPLs and participate in flower regulation are largely unknown. For further insight, we screened the yeast hybridization library for CbuSPL9 in buds. Many flowering-related proteins were detected, for example, the GATA-related proteins [54][55][56][57], bHLH48 [58], FLA [59], and PHD finger protein ALFIN-LIKE [60]. Furthermore, a high mobility group (HMG) protein, CbuHMGA, was identified to interact directly with CbuSPL9. The BiFC assay and BLI assay further confirmed their interaction. The BiFC assay suggested that CbuSPL9 and CbuHMGA could form protein complexes in the nucleus.
ChuHMGA contains an H15 domain and 4 AT-hook motifs. This similar structural description was confirmed in most plant HMGA proteins, which encoded four four AT-hook motifs [61,62]. Additionally, the amino-terminal region of plant HMGA proteins shares remarkable homology with the DNA-binding domain of histone H1 [63,64]. ChuHMGA proteins have a highly conserved structure. HMGA is an architectural TF [65][66][67]. It regulates gene expression in vivo by controlling the formation of multiprotein complexes on the AT-rich regions of certain gene promoters [63,64,66,68,69]. To date, some plant HMGA proteins have been isolated, such as those in soybean, rice, maize and Arabidopsis [70][71][72][73][74][75]. And the AT-hook motif has also been reported that AT-hook proteins also affects the flowering process [71,75,76]. However, studies of the interacting proteins and the function of HMGA in woody plants are rare. In "Bairihua", the expression of CbuHMGA was elevated in the dormant period, and the highest expression level was detected in the reproductive growth period. The result was consistent with that in Arabidopsis. Although the oe-hmga transgenic lines did not show a flowering time phenotype, the floral organ mutation of oe-hmga was similar to that of the oe-spl9 transgenic lines. The HMGA proteins have an important role in biological processes and interact with different TFs [66,67]. Their intrinsic flexibility allows the HMGA proteins to participate in specific protein-DNA and protein-protein interactions that induce structural changes in chromatin and the formation of stereospecific complexes called 'enhanceosomes' on the promoter/enhancer regions of genes whose transcription they regulate. The chromatin structure changes affect the ability of TFs to bind with the promoter/enhancer regions [61,70,[76][77][78][79]. From the results in this study, we suggest that the interaction of CbuHMGA with CbuSPL9 might strengthen or weaken the binding ability of CbuSPL9 with the corresponding DNA sequences or downstream proteins, thus affecting the flowering process. Further experiments are needed to test this hypothesis. For example, overexpression one gene represents the PCR results in oe-hmga; represents the PCR results in oespl9; and represents the PCR results in col. Three independent biological replicates were performed, and each replicate was measured in triplicate. The error bars show the standard deviation of the results of three technical replicates in an other gene Arabidopsis mutant, to reveal pstream and downstream relationship of CbuSPL9 interacting with CbuHMGA during flowering in Caltalpa bungei.

Conclusion
"Bairihua" provides us with a valuable opportunity to gain a deeper understanding of the flowering process of woody plants. Our study showed that SPLs may have similar structures and regulatory mechanisms in perennial trees and Arabidopsis. Screening CbuSPL9-interacting proteins revealed not only additional proteins for research on the regulatory pathway of CbuSPL9 but also the CbuSPL9 interaction with CbuHMGA. oe-hmga showed a phenotype that affected floral organ development but did not change flowering time. This result suggests that the mechanism by which CbuSPL9 affects the flowering process is a complex process, and CbuHMGA is involved in the development of flower organs but not in the regulation of flowering time.