Ectopic expression of a bamboo SVP-like gene alters flowering time and floral organs in Arabidopsis thaliana

The Short Vegetative Phase (SVP) gene is a key regulator for floral transition and development. Although SVP-like genes have been identified and characterized in many plant species, their orthologs in bamboo have not been characterized. In this study, one SVP homolog was isolated from lei bamboo based on the P. edulis transcriptome database and designated as PvSVP2. Phylogenetic analysis showed that PvSVP2 was closely related to rice OsMADS47. Expression analysis revealed that PvSVP2 was widely expressed in different tissues but significantly in vegetative tissues. Moreover, it has higher transcript levels in the late stages of flower development. Overexpression of PvSVP2 in Arabidopsis thaliana caused the early flowering and abnormal floral morphologies. Further, the qRT-PCR analysis showed that the genes regulating flowering time (FT and SOC1) and flower development (AP1, AP3 and PI) expressions significantly increased in transgenic A. thaliana lines and correlated with PvSVP2 expression. The subcellular location of PvSVP2 in both onion epidermal cells and A. thaliana protoplast was localized in the nucleus and cytomembrane. Through yeast two-hybrid and BIFC assays, we identified that PvSVP2 interacts with PvMADS56 (a SOC1 homolog) and PvVRN1 (an AP1 homolog). These results suggested that PvSVP2 may play an essential role in the flowering process of bamboo by regulating the PvMADS56 and PvVRN1. Besides, we purified and obtained the PvSVP2 recombinant protein by prokaryotic inducement. Our data will provide information to understand the characterization and function of PvSVP2 and be beneficial to illustration the molecular mechanism of bamboo flowering. PvSVP2, a bamboo SVP homolog, might cause early flowering and abnormal floral organs by interacting with PvMADS56 and PvVRN1.


Introduction
The floral transition is a complex process and significantly affects both plant fitness and crop yield. Therefore studying the underlying mechanisms is vital for increasing agricultural productivity (Bouché et al. 2015). The Short Vegetative Phase (SVP) gene is a crucial flowering repressor and regulated by the autonomous, thermosensitive and gibberellin pathways (Andrés et al. 2014;Hartmann et al. 2000;Lee et al. 2013;Li et al. 2008;Marín-González et al. 2015). Although SVP is widely expressed in plant stems and leaves, it can delay the floral transition by regulating integrators of flowering time signals SOC1, FT and FLC genes (Li et al. 2008;Jang et al. 2009;Mateos et al. 2015). The integrators integrate various environmental and endogenous signals to control the transition from vegetative to reproductive development (Bouché et al. 2015;Amasino and Michaels 2010;Parcy 2005). In addition, SVP also interacts directly or indirectly with AP1, AP3, PI and SEP3 to affect the development of floral organs (Gregis et al. 2006;de Folter et al. 2005). Therefore, SVP not only controls the flowering time but also affects flower development.
The role of SVP genes in flower development and flowering time regulation, is consistent with in most of the studies. In A. thaliana, overexpression of SVP genes leads to late flowering and svp mutant plants were early flowering (Bouché et al. 2015;Li et al. 2008;Masiero et al. 2004). Ectopic expression of PtSVP from trifoliate orange in A. thaliana causes late flowering and additional trichomes and floral defects (Li et al. 2010). Overexpression of two Medicago SVP genes causes floral defects and delayed flowering in A. thaliana (Jaudal et al. 2013). In kiwifruit, four SVP (SVP1, SVP2, SVP3 and SVP4) homologs have been identified. Among them, SVP1 and SVP3 act repressors of flowering and the overexpression studies identify that these genes was able to rescue the A. thaliana svp-41 mutant (Wu et al. 2011). These data show that SVP-like genes have the conserved roles in dicot species in floral development and flowering time regulation. In Antirrhinum, INCOMPOSITA (an SVP homolog) controls prophyll development, floral meristem identity and flowering time (Masiero et al. 2004).
In contrast, SVP homologs from monocots mainly regulate flower development instead of flowering time. In barley (Hordeum vulgare), ectopic expression of Barley MADS1 (BM1) and BM10 (SVP homologs) inhibit floral development and cause floral reversion (Trevaskis et al. 2006). There are three SVP-like genes (OsMADS22, OsMADS47 and OsMADS55) in rice (Lee et al. 2008a). Further, transgenic rice overexpressing OsMADS22 have abrupt floral morphogenesis (a disorganized palea, an elongated glume and a two-floret spikelet) (Sentoku et al. 2005). Heterologous expression of OsMADS22 and OsMADS47 in A. thaliana only causes the alternation of flower development but not complement for the flowering phenotypes of svp mutant (Fornara et al. 2008). However, overexpression of OsMADS55 delays the flowering time of wild-type plants and rescues the early flowering phenotype of svp mutant in A. thaliana (Lee et al. 2012). The above results indicated that SVP homologs in plants were mulfuctional genes along with conservative and divergent characteristics.
Bamboo is the woody monocot that belongs to the grass family Poaceae (Hou et al. 2021). However, bamboo is semelparous and owns many peculiar habits in flowering (Biswas et al. 2016;Lin and Mao 2007). Bamboo flowering has a long vegetative phase and is unpredictable. Further, the flowering intervals vary from a few years to several decades (Sharma et al. 2014;Yao et al. 2020;Zheng et al. 2020). The flowering is often followed by the death of the flowered clumps (Lin and Mao 2007;Hou et al. 2020). The gregarious bamboo flowering results in substantial economic loss and ecological crisis in many tropical and subtropical regions. The mechanism of bamboo flowering is still unknown, even if this phenomenon has been recorded and studied for a long time (Biswas et al. 2016;Dutta et al. 2018Dutta et al. , 2021Jiao et al. 2019;Liu et al. 2016a, b, c;Liu et al. 2020;Zhang et al. 2018;Zheng et al. 2017). The SVP-like genes from monocots and dicots play an important role in the regulation of flowering. However, the molecular characteristics of bamboo SVP-like genes and their functions are still unclear. In this study, an SVP-like gene from lei bamboo (Phyllostachys violascens) was cloned and characterized by analyzing the overexpression plants of transgenic A. thaliana. Our study will lay a foundation to understand the mechanism of bamboo flowering.

Plant material and growth conditions
Lei bamboo samples used for gene cloning and expression analysis were collected in the Bamboo Garden of Zhejiang Agriculture and Forestry University. Wild-type A. thaliana Columbia-0 (Col-0) was used for gene transformation. Seeds were sterilized with 10% sodium hypochlorite (NaClO) for ten minutes, washed five times with ddH 2 O, and then sowed onto 1/2 MS petri dish. The A. thaliana seedlings were transplanted into the soil when they grew four cotyledons. All A. thaliana plants grow at 22 °C under a long photoperiod (16 h light /8 h dark) in a control growth chamber.

Isolation of PvSVP2 from Lei bamboo
Total RNA was extracted from the young leaf using Trizol reagent (Invitrogen, US). RNA quality and concentration were detected with 1% RNase-free agarose/ethidium bromide gels and spectrophotometer. First-strand cDNA was synthesized by using Reverse Transcriptase M-MLV (TAKARA Company). The sequence of OsMADS55 and OsMADS47 (SVP homolog from rice) ORF were blasted on the moso bamboo (P. edulis) transcriptome database  by BioEdit software and isolated the PvSVP1 (data not show in this paper) and PvSVP2 genes. Full-length sequences were amplified using gene-specific primers of PvSVP2 (Table S1). The PCR product was cloned into a pMD20-T vector (Takara Company) and then sequenced for further verification.

Plasmid construction and transformation of A. thaliana
To construct the overexpression vector, the ORF of PvSVP2 was ligated to the KpnI-SalI site of the pCAMBIA1302 vector under the control of cauliflower mosaic virus (CaMV) 35S promoter. The pCAMBIA1302-PvSVP2 fusion vector was transformed into Agrobacterium tumefaciens strain GV3101 cells and then were introduced into A. thaliana (ecotype Col-0) using the floral dip method. Transgenic seeds were survived in the medium containing 50 μg/ml hygromycin were transferred to soil and further confirmed by genomic PCR. Screened as above, positive T 3 generation of 35S::PvSVP2 transgenic A. thaliana were used for further study. The flowering time of A. thaliana was scored as rosette leaf number and days in which the main stem bolted 1 cm.

Gene expression analyses
Quantitative real-time PCR (qRT-PCR) was used to analyze the spatio-temporal expression of PvSVP2. Total RNA was extracted from different tissues, including young leaf, mature leaf, culm, rhizome, shoot or flower, and at three times (young leaf and flower) as described by Liu et al. (2016a, b, c) in flowering and non-flowering plants and healthy leaves of wild-type and T 3 homozygous transgenic plants. qRT-PCR reactions were performed in CFX96 Real-Time PCR Detection System (Bio-Rad, Germany) with SYBR Premix Ex Taq II mix (Takara). The PCR conditions were as follows: 95 °C for 3 min, followed by 40 cycles of amplification (95 °C for 10 s, 60 °C for 20 s). Reactions were performed in 20-µL mixtures consisting of 10 µL 2 × SYBR Premix Ex Taq II mix, 0.5 µL each forward or reverse primer, 1 µL cDNA template and 8 µL ddH 2 O. The PheUBC18 gene from moso bamboo was used as the control for normalization (Qi et al. 2013) for different tissues expression analyses. The primers are listed in Table S1. Further, FLC, FT, SOC1, AP1, AP3 and PI genes expressions were also analyzed by gene-specific primers (Table S1) in wild-type and T 3 homozygous transgenic plants. The actin1 was used as an internal control. The relative gene expression levels were calculated using the 2 −∆∆Ct method (Livak and Schmittgen 2001).

Subcellular localization
To determine the subcellular location, the coding sequences of PvSVP2 without TAG at the 3ʹ end was ligated to the BamHI-SalI site of the vector Cam35S-GFP to generate a 35S::PvSVP2-GFP fusion vector. The plasmid 35S::PvSVP2-GFP was transfected into onion epidermal cells by particle bombardment method (Wang et al.1988) and A. thaliana protoplasts by the polyethylene glycol (PEG) as described by Yoo et al.(2007). Meanwhile, the transient expression of the empty vector (Cam35S-gfp) was used as a positive control. The corresponding fluorescence images were taken using a confocal laser scanning microscope (LSM510, Zeiss, Germany).

Yeast two-hybrid assay
The coding sequences of PvSVP2, PvMADS56 and PvVRN1 were amplified with specific primers (Table S1) and cloned into the pGADT7 (AD) and pGBKT7 (BD) plasmids. All the constructs were verified by sequencing. According to the manufacturer's instructions, the prey pGADT7 vector and constructs AD-PvSVP2, AD-PvVRN1, and AD-PvMADS56 were transformed into yeast strain Y2H Gold, and the bait pGBKT7 vector and constructs BD-PvSVP2, BD-PvVRN1, BD-PvMADS56 were transformed into yeast strain Y187. The transformed yeast cells were incubated at 30 °C for 3-5 days on the selection medium SD/-Leu/X-α-gal and SD/-Trp/ X-α-gal for autoactivation. The cells that do not develop blue were confirmed as they do not contain autoactivation. The pGBKT7-53 + pGADT7-T and pGBKT7-Lam + pGADT7-T were positive and negative controls, respectively. The protein interactions were confirmed on SD/-Trp/-Leu/-His/-Ade/ X-α-gal medium.

Prokaryotic expression and purification of PvSVP2
The open reading frame of PvSVP2 was amplified using pMD19-PvSVP2 plasmid as the template and then linked to the BamHI-SalI site of the vector pET-HTT (pHTT) to form a PvSVP2-pHTT fusion vector. The positive PvSVP2-pHTT construct was transformed into E.coli Rossetta TM (DE3) cell for protein production. The cells were cultivated in Luria-Bertani (LB) media containing 1% tryptone, 0.5% yeast extract, 1% NaCl and 50 mg/ml kanamycin. 10 ml bacteria was further inoculated into 1L LB culture at 37 °C for 220r/m until the OD600 reached to 0.5-0.6. Isopropylthiogalactoside (IPTG) was added with a final concentration of 400 μmmol/L while no IPTG as a control. Then the bacteria were induced under 37 °C and 20 °C for 5 h and 12 h, respectively. The cells were centrifuged at 5000×g for 10 min. The collected precipitates were disrupted in lysis buffer with a 10 min sonication at amplitude of 35% power (on/off 8/6 s). Then the supernatants and precipitates were separately collected and were detected by 12% sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE).
Next, the PvSVP2-pHTT fusion protein was purified from the supernatant using the Dextrin Sepharose High performance (GE Company, Shanghai). The supernatant was added to 2 ml Nickel beads and rocked slowly on ice for 1 h after washing Nickel beads with the sterile distilled water and lyse buffer successively. The Nickel beads were then washed with 15 column volumes (CV) of lysis buffer with 40 mmol/L concentration. PvSVP2 protein was further eluted by 15 CV of lysis buffer, including 40 mmol/L imidazole. The eluents were collected and analyzed by 12% SDS.

Bioinformatics and statistical analyses
The amino acid sequences of related SVP-like proteins used in phylogenetic tree construction were retrieved from the NCBI database. Phylogenetic analysis was constructed in the MEGA 5.0 software using the Neighbor-Joining method with a bootstrap value of 1000 replications (Tamura et al. 2011). Primers were designed by the Vector NTI and Premier Primer 5 software. The software ProtParam in the ExPASy (http:// exPasy. org) was used to analyze the physical and chemical properties of the protein. The subcellular localization of the protein was predicted with WoLF PSORT (http:// www. gensc ript. com/ psort/ wolf_ psort. html). Statistical analysis was carried out using SPSS 21.0. Differences were analyzed with one-way ANOVA followed by Tukey's test. Significance was accepted at the level of p < 0.05 or p < 0.01.

Isolation and bioinformatics analysis of PvSVP2
We isolated two SVP-like genes PvSVP1 and PvSVP2, from P. violascens according to the P. edulis transcriptome database. However, only PvSVP2 data were shown in this paper. The PvSVP2 is 693 bp nucleotides long, and 230 amino acids possess a MADS-box and K-box (Fig. 1a). Moreover, many DNA binding sites and dimerization interfaces were found in the MADS-box of the PvSVP2 protein (Fig. 1a). Its theoretical molecular weight (MW) and the theoretical isoelectric point (pI) were 26.02 KDa and 7.76, respectively. Phylogenetic tree analysis showed that PvSVP2 were grouped with monocot Poaceae SVP-like proteins (Fig. 1b). Moreover, PvSVP2 clustered as a subgroup with BM1, ZMM20 and OsMADS47, and was more closely related to OsMADS47 from O. sativa, with 75.1% identity (Fig. 1c).

Expression analysis of PvSVP2
qRT-PCR analysis was carried out for detecting the transcript level of PvSVP2 in different tissues and developmental stages. The results showed that PvSVP2 was expressed in all tested tissues, including young leaf, mature leaf, culm, rhizome, shoot and flower (Fig. 2a). In flowering and nonflowering plants, the mRNA levels of PvSVP2 were higher in culm and rhizome, whereas lowest in the shoot. To determine whether PvSVP2 were related to flowering in P. violascens, we analyzed the expression in the leaf and flower of flowering plant (FL and FF), respectively and the leaf of the non-flowering plant (VL). These tissues were collected in three different stages T1, T2 and T3 as described by our previous publication, Liu et al. (2016a, b, c) ( The blooming stage when the anther was outcropped from palea (April 12)). The expression of PvSVP2 changed with the flowering stage from March 15 to April 12 (Fig. 2b). The mRNA level of PvSVP2 in FL increased first from T1 to T2, and then decreased to T3. In VL, the mRNA level of PvSVP2 was first reduced from T1 to T2 and then remained almost unchanged during T2-T3. In contrast to FL and VL samples, the mRNA level increased significantly from T1 and T2 in FF, then remained unchanged to T3. Moreover, the PvSVP2 expression in FF was lower at T1 and higher in T2 and T3 than VL and FL (Fig. 2b).

Ectopic expression of PvSVP2 in A. thaliana caused the early flowering and abnormal floral morphologies
To determine PvSVP2 function, we generated transgenic A. thaliana plants expressing PvSVP2 constitutively (35S::PvSVP2). Among eleven homologous T 3 transgenic A. thaliana lines, we selected three lines (line 1, line 28 and 42) for further analysis. Compared with the wild-type A. thaliana, plants carrying 35S::PvSVP2 vector (line1, line28, and line42) flowered significantly early by an average of 9.4 days ( Fig. 3a, b, p < 0.01). Further, the number of rosette leaves at the time of bolting for three 35S::PvSVP2 lines were less than the control plants with an average of 2.4 leaves (Fig. 3c, p < 0.05 or p < 0.01). The qRT-PCR analysis identified that the flowering time was correlated with the PvSVP2 expression level (Fig. 3d). For example, line 28 was the earliest flowering plant among the three lines studied, containing the highest transcript level of PvSVP2. Moreover, 35S::PvSVP2 transgenic lines also produced abnormal floral organs. As follow, the sepals of transgenic plants appeared to be small and leaf-like structures (Fig. 4b, c), which did not enclose and protect inner floral parts (Fig. 4e) and kept on until the  (Fig. 4f, g). Meanwhile, the petals also displayed the opened phenotype (Fig. 4a, b). Besides, the petals of 35S::PvSVP2 transgenic lines were initially green and became purple with the flowering process (Fig. 4b, c).

PvSVP2 overexpression altered the expressions of flowering-related genes in transgenic A. thaliana plants
It is well reported that SOC1, FT and FLC genes are involved in the regulation of flowering time (Li et al. 2008;Jang et al. 2009;Mateos et al. 2015). Therefore, to analyze the influence of PvSVP2 on the flowering time, transcript levels of FT, SOC1 and FLC genes were analyzed by qRT-PCR. Compared with the wild-type plants, in 35S::PvSVP2 transgenic plants, FT and SOC1 transcript levels were both increased, but FLC transcript levels displayed no noticeable change (Fig. 5). These results indicate that PvSVP2 overexpression causes early   Further, we also analyzed the AP1, AP3 and PI genes, which are essential for developing flower organs (Liu et al. 2009). The AP1, AP3 and PI expressions were highly upregulated in 35S::PvSVP2 transgenic plants compared with the wild-type plants.
The qRT-PCR results indicated that AP1, AP3 and PI expressions are positively correlated with PvSVP2 expression in transgenic lines (Fig. 5). These results indicated that overexpression of PvSVP2 in A. thaliana affects the development of the floral organ by modulating AP1, AP3 and PI expression.

PvSVP2 located in the nucleus and cytomembrane
We predicted that the PvSVP2 was located in the nucleus by the WoLF PSORT software. Further, we confirmed the subcellular location using the particle bombardment method using onion epidermal cells. Confocal microscope images revealed that the PvSVP2-GFP fusion protein was mainly localized in the nucleus and thinly localized in the cytomembrane of the onion epidermal cells. In contrast, the empty vector was uniformly distributed throughout the whole onion cell (Fig. 6a). To further confirm this, the PvSVP2-GFP fusion vector was also transformed into the A. thaliana protoplasts by PEG-mediated transformation. These results also confirm that PvSVP2-GFP was positioned in the nucleus and cytomembrane of the A. thaliana protoplasts (Fig. 6b).

Prokaryotic expression of PvSVP2
We used the prokaryotic expression system to analyze PvSVP2 protein solubility. The PvSVP2-pHTT construct was induced at 37 °C and 20 °C, respectively (Fig. 9a). SDS-PAGE analysis showed that a specific protein band of about 30 kDa was expressed in supernatant at 20 °C by IPTG induction (Fig. 9a, lane 5). However, the molecular mass of the fusion protein was 27 kDa including a 1 kDa His tag protein and a 26 kDa PvSVP2 protein. To further verify it, the soluble protein was further purified with a dextrin sepharose high performance Fig. 9b) and the protein sequence was examined by MALDI-TOF/TOF mass spectrometer analysis. As Fig. 9c shown, the generated peptides and the peptides derived from PvSVP2 protein sequence had high coverage rate, reach to 53%, indicating that the protein was PvSVP2-pHTT fusion protein.

Discussion
The SVP genes play a crucial role in controlling flowering time and floral organ characteristics (Hartmann et al. 2000;Michaels et al. 2003). In this study, we isolated and characterized one SVP homolog, PvSVP2, from Lei bamboo. Phylogenetic analysis showed that PvSVP2 belonged to the same group with BM1, ZMM20 and OsMADS47 from monocot plants and was closely related to OsMADS47 (Fig. 1b, c). Previous studies indicate that SVP homologs are different and have three types of expression patterns. The first type is mainly expressed in vegetative tissues but not in reproductive tissues like AtSVP (Hartmann et al. 2000), BM1 (Trevaskis et al. 2006), OsMADS47 (Lee et al. 2012) and AcSVPs (Wu et al. 2011; the second type is expressed in reproductive tissues but not in vegetative tissues such as OsMADS22 (Lee et al. 2012); the third type is expressed both in vegetative and reproductive tissues such as OsMADS55 (Lee et al. 2012), INCOMPOSITA (Masiero et al. 2004), MtSVP (Jaudal et al. 2014), CmSVP (Gao et al. 2017), LoSVP (Tang et al. 2019), EiSVPs (Jiang et al. 2019), MiSVPs (Mo et al. 2021) and PavSVP (Wang et al. 2021). Our results identified that PvSVP2 belongs to the third type and was expressed in all vegetative and reproductive tissues such as leaf, culm, rhizome, shoot and flower (Fig. 2). These results showed that the expression pattern of SVP-like genes is various, suggesting that their function may be different in different species.
To identify its function, PvSVP2 was ectopically expressed in wild-type A. thaliana. To ensure that the observed phenotypic changes were solely caused by the overexpression of PvSVP2, qRT-PCR analyses further confirmed its transcript level using the specific primers distinguish from AtSVP in the selected transgenic lines, but not in the wild-type (Fig. 3d). These results suggested that the phenotypic changes of transgenic plants are due to the overexpression of PvSVP2.
Compared to the wild-type plants, the 35S::PvSVP2 transgenic plants displayed abnormal sepal and petal structures (Fig. 4). These phenotypic observations were consistent with most SVP-group MADS-box genes studied. It was worth noting that the petals of these transgenic plants became purple in color. Similar phenotype was observed in A. thaliana transgenic plants overexpressed with the OsMADS22 gene (Fornara et al. 2008). Despite this, PvSVP2 was closely related to OsMADS47 rather than OsMADS22. For SVP homologs, most of them mainly as a repressor in regulating flowering time, but recent studies find that a few have opposite function. For example, PfMADS16 (Polypogon fugax) and MiSVP2 overexpressing in A. thaliana displayed an early flowering phenotype (Mo et al. 2021;Zhou et al. 2020). Our results also showed that overexpression of PvSVP2 led to early flowering in transgenic A. thaliana lines (Fig. 3a-c). These results suggested that SVP-like genes have functional conservation and diversification.
Because the overexpression of PvSVP2 resulted in early flowering and abnormal organs, the transcript levels of related genes were analyzed in 35S::PvSVP2 plants. The qRT-PCR results showed that the FT and SOC1 gene expression was Fig. 7 Interactions between PvSVP2 and PvMADS56 or PvVRN1 in yeast two-hybrid experiment. Yeast containing dual vectors of pGBKT7-53 and pGADT7-T was used as the positive control, that of pGBKT7-Lam and pGADT7-T as the negative control. SD/-Leu/-Trp/-His/-Ade medium was used for clone selection purpose. Appearance of X-α-gal activity mirrors positive interaction  (Fig. 5). These results were consistent with overexpression studies of MiSVP2 and PfMADS16 in A. thaliana, which displayed an early flowering phenotype (Mo et al. 2021;Zhou et al. 2020). But unlike MiSVP2, the expression of FLC remained unaltered in 35S::PvSVP2 transgenic plants. Moreover, AP1, AP3 or PI expressions were increased in these transgenic plants (Fig. 5). In conclusion, PvSVP2 affected flowering time by regulating FT and SOC1 expression and flower development by regulating AP1, AP3 and PI expression in transgenic A. thaliana plants.
In A. thaliana, SVP can interact with SOC1 or AP1 protein to regulate flowering time and to establish floral meristem identity, respectively Lee et al. 2008b). In this study, PvSVP2 could interacted with PvMADS56 and PvVRN1 as demonstrated by the yeast twohybrid and BiFC assays (Figs. 7, 8). PvMADS56 (a SOC1 homolog) and PvVRN1 (an AP1 homolog) from P. violascens overexpressing in A. thaliana display not only abnormal floral organs but also caused early flowering (Liu et al. 2016a, b, c;Ma et al. 2016). Therefore, we proposed that PvSVP2 might affect flowering time and flower development by interacting with PvVRN1 and PvMADS56 in bamboo.
SVP and SOC1 are located in the nucleus and cytoplasm, respectively. However, SOC1-SVP heterodimers are translocated to the nucleus, suggesting SOC1 localization is affected by SVP (Lee et al. 2008b). Unlike SVP, PvSVP2 is not only located in the nucleus but also in the cytomembrane (Fig. 6). The same results are found for the subcellular location of PavSVP (Wang et al. 2021). Intriguingly, PvMADS56-PvSVP2 and PvVRN1-PvSVP2 heterodimers were translocated to the cytoplasm by transient assays, whereas PvMADS56 and PvVRN1 are also located in the nucleus (Fig. 8) (Liu et al. 2016a, b, c;Ma et al. 2016). These data suggested that PvSVP2 was different from SVP in signal transduction and needed a further study for confirmation. Besides, with the success of prokaryotic expression in this paper, these results will be helpful for further functional research of PvSVP2 in bamboo.

Conclusions
In the present study, we identified an SVP homolog PvSVP2 in lei bamboo. PvSVP2 was preferentially expressed in vegetative tissues. Overexpressing in A. thaliana led to early flowering and abnormal floral organs. Moreover, PvSVP2 could interact with PvVRN1 and PvMADS56 by the yeast two-hybrid and BIFC assays. Therefore, we speculated that the PvSVP2 was involved in flowering time and flower organ development by regulating PvVRN1 and PvMADS56 in bamboo.