Overexpression of PvCO1, a bamboo CONSTANS-LIKE gene, delays flowering by reducing expression of the FT gene in transgenic Arabidopsis
In Arabidopsis, a long day flowering plant, CONSTANS (CO) acts as a transcriptional activator of flowering under long day (LD) condition. In rice, a short day flowering plant, Hd1, the ortholog of CO, plays dual functions in respond to day-length, activates flowering in short days and represses flowering in long days. In addition, alleles of Hd1 account for ~ 44% of the variation in flowering time observed in cultivated rice and sorghum. How does it work in bamboo? The function of CO in bamboo is similar to that in Arabidopsis?
Two CO homologous genes, PvCO1 and PvCO2, in Phyllostachys violascens were identified. Alignment analysis showed that the two PvCOLs had the highest sequence similarity to rice Hd1. Both PvCO1 and PvCO2 expressed in specific tissues, mainly in leaf. The PvCO1 gene had low expression before flowering, high expression during the flowering stage, and then declined to low expression again after flowering. In contrast, expression of PvCO2 was low during the flowering stage, but rapidly increased to a high level after flowering. The mRNA levels of both PvCOs exhibited a diurnal rhythm. Both PvCO1 and PvCO2 proteins were localized in nucleus of cells. PvCO1 could interact with PvGF14c protein which belonged to 14–3-3 gene family through B-box domain. Overexpression of PvCO1 in Arabidopsis significantly caused late flowering by reducing the expression of AtFT, whereas, transgenic plants overexpressing PvCO2 showed a similar flowering time with WT under LD conditions. Taken together, these results suggested that PvCO1 was involved in the flowering regulation, and PvCO2 may either not have a role in regulating flowering or act redundantly with other flowering regulators in Arabidopsis. Our data also indicated regulatory divergence between PvCOLs in Ph. violascens and CO in Arabidopsis as well as Hd1 in Oryza sativa. Our results will provide useful information for elucidating the regulatory mechanism of COLs involved in the flowering.
Unlike to the CO gene in Arabidopsis, PvCO1 was a negative regulator of flowering in transgenic Arabidopsis under LD condition. It was likely that long period of vegetative growth of this bamboo species was related with the regulation of PvCO1.
KeywordsCONSTANS (CO) Flowering time Functional divergence Flowering regulation Bamboo Phyllostachys violascens
Flowering locus T
Heading date 3a
Open reading frame
Rapid amplification of cDNA end
Reverse transcription PCR
The transition from a vegetative phase to a reproductive phase is an important developmental switch in plants. This transition is controlled by several environmental and endogenous conditions [1, 2]. Different plant species have various mechanisms to regulate this process  and many, such as grasses, have distinct flowering habits.
Bamboo is one of the most important non-timber forests and belongs to the Poaceae. Unlike other plants in this family, such as rice, maize, and wheat, the flowering time of bamboo appears to be random. Some species have prolonged vegetative growth lasting decades before flowering and death. One such species is Phyllostachys heterocycla, a woody bamboo that has ecological, economic and cultural value . Another economically important species, Ph. violascens, belongs to the same genus and has very similar genetic background with Ph. heterocycla. In this species, those elder plants at the age of 6 years would usually be harvested for gain yield of bamboo shoots. However, compared with Ph. heterocycla, the flowering pattern of Ph. violascens is variable. Its different individuals can flower at different times during the year. Some plants flower twice and more, some only once, some never flower even when they were harvested [5, 6]. Some young plants without leaves grow poorly but still flower and then die. There are some individual plants flower every year. Flowering duration can be 60 to 90 d. Many researchers have attempted to explain the factors controlling flowering. These factors include nutrition, climate, stress, and molecular mechanisms [7, 8]. Currently, studies on the molecular mechanism of bamboo flowering have focused on transcriptome sequencing and expression of genes involved in the developmental stages of flowering [4, 5, 9, 10, 11, 12, 13, 14], while reports on the genes involved in floral induction are rare [6, 15, 16, 17, 18]. Peng et al.,  reported that repeat insertions in the regulatory region of most homologs encoding CONSTANS (CO), might result in low gene expression in Ph. heterocycla. And the CONSTANS (CO) gene was originally isolated as a photoperiodic floral promoter.
The CO gene in Arabidopsis plays a critical role in control of flowering time by directly activating the expression of target genes including FLOWERING LOCUS T (FT) which encodes a florigenic protein . Overexpression of the CO gene accelerates Arabidopsis flowering regardless of photoperiod . CO gene mutation results in delayed flowering under long-day (LD) conditions, but has no effect on flowering time under short-day (SD) . CO encodes a B-box-type zinc-finger transcriptional factor with two B-box domains near the N-terminus and a CCT (CO, CO-like, and TOC1) domain near the C-terminus [21, 22, 23]. The B-box domain of CO is likely involved in protein-protein interactions and the CCT domain binds DNA [24, 25, 26]. The CO protein can bind to specific cis-elements in the FT promoter either by itself  or in a complex with CCAAT-binding factors [19, 26, 27] to regulate FT transcription. CO protein can also interact with specific 14–3-3 isoforms, 14–3-3 μ and ν proteins which belong to the family representing nodes of signal integration and cross talk, affecting photoperiodic flowering . In rice, a SD plant, the Hd1 gene, orthologous to CO, promotes flowering under SD conditions, but delays flowering in LD conditions [29, 30]. In addition, the mutant of Se1, allelic to Hd1, controlling photoperiod sensitivity, is also slightly later than its progenitor variety in heading date . The wheat TaHd-1 gene, also homologous to CO, can complement the function of rice Hd1: it also promotes heading under SD conditions, but delays it under LD conditions . Overexpression of LpCO (from Lolium perenne) leads to early flowering in Arabidopsis . The PnCO gene from Pharbitis nil can complement the co mutant of Arabidopsis . Overexpression of GmCOL1a, GmCOL1b, GmCOL2a and GmCOL2b from soybean rescued the late flowering phenotype of Arabidopsis co mutant . The alleles of Hd1 account for ~ 44% of the variation in flowering time in cultivated rice and sorghum, suggesting Hd1 plays an important role in flowering. Differences of CO gene expression are responsible for differences in flowering times . CO is important to many plant species including poplar (Populus spp) , but its function remains unknown in non-model systems such as woody perennial bamboo species.
Whether the COL genes in bamboo have the influence on the flowering time are unclear and it is unknown if COL gene functions in bamboo are similar to those in Arabidopsis. In the present study, two homologous CO genes, PvCO1 and PvCO2, were identified from Ph. violascens. Their expression patterns were analyzed and primary functions were characterized. The results give new insights into the understanding of the COLs genes involved in floral transition.
Phyllostachys violascens (Carriere) Riviere in this study were grown in the field under natural conditions on the campus of Zhejiang A&F University (30°14′N, 119°42′E). The mean annual temperature is 15.6 °C, with maximum and minimum temperatures of 41.7 °C and − 13.3 °C, respectively. The average length of sunshine in Lin’an is approximately 1,847 h per year. We chose those plants which flowered last year for sampling. Some of these plants flowered again from mid-March to mid-May and flowering lasts for 60 to 90 d.
To study expression of PvCOLs before, during, and after flowering, firstly, we sampled fully expanded, mature leaves from ten flowered plants at 5:00 pm on March 2, 2012. Ten days later, on March 12, we sampled leaves again from these ten flowered plants also at 5:00 pm. And then, we found 4 individual plants displayed flower bud and flowered again between on March 12 and 22, among which 3 were targeted for sampling. Thereafter, we collected leaves from these 3 flowering plants every 10 d until to May 31 at 5:00 pm because these three plants died between on May 31 and June 3. Day length increased from 11.5 h light on March 2 to 14 h light on May 31. Meanwhile, we immediately determined expression of target genes after collecting the leaf samples. Once the target gene was expressed, we also collected fully expanded, mature leaves from the same three flowering plants 8 times at 3 h intervals on March 30 (LD 12.5:11.5 h) and determined expression of target genes for diurnal expression analysis. The maximum and minimum temperatures were 15.7 °C and 13.3 °C on this day, respectively. We also sampled mature leaves, immature leaves, roots, stems and flower buds for determining the tissue-specific expression from 5 pm to 6 pm on April 13 (LD 13:11 h). The maximum and minimum temperatures on April 13 were 17.3 °C and 13.9 °C, respectively. All plant samples were stored at − 80 °C prior to further experiments.
Wild type (WT) and transgenic plants of Arabidopsis thaliana ecotype Columbia-0 (Col-0) were cultured in a room under ≈22 °C with LD (16 h light: 8 h dark) conditions. The light intensities is 200 umol/m .s.
DNA and RNA procedures
Total genomic DNA was extracted from young leaves of Ph. violascens by the CTAB method and total RNAs were extracted from the collected samples using Trizol reagent (Invitrogen, US). To remove any residual genomic DNA from the preparation, the RNA was treated with RNase-free DNase I according to manufacturer instructions (Qiagen, Valencia, CA, US). The first-strand complementary DNA (cDNA) was synthesized using the Super Script III kit (Invitrogen, US), according to the manufacturer manual.
A pair of degenerate primers (TOHLF1/TOHLR2) was designed according to the conserved sequence of CO homologous genes from rice, maize and wheat, and used to amplify the partial DNA and cDNA of PvCOL. Based on the partial DNA sequence of PvCOL, the primers 5SP1, 5SP2, 5SP3, 3SP1, 3SP2 and 3SP3 used for genome walking amplification and the primers GSP1, GSP2, GSP3 and GSP4 used for rapid amplification of cDNA end (RACE) were designed in order to obtain the 5’ and 3’ terminal sequences of the PvCOL genes. The PvCOL DNA sequence containing the encoding region was assembled by a combination of the conserved sequence and the 5’ and 3’ terminal sequences. To obtain the full-length cDNA and genomic DNA sequence of PvCOL, two pairs of specific primers, PvCO1F and PvCO1R for PvCO1, PvCO2F and PvCO2R for PvCO2, were designed based on the assembled sequence and used for amplification. Detailed information on all primers used is listed in Additional file 1: Table S1. All of the amplified fragments were gel purified, ligated into the pMD18-T vector, transformed into the DH5a competent cells, and sequenced.
Real-time PCRs were performed according to the procedures of Guo et al.,  and semi-quantitative PCR according to Putterill et al., . Annealing temperature and the cycles of PCR were adjusted according to the primers and target genes. Primers used (PvCO1qexpF and PvCO1qexpR for PvCO1, PvCO2qexpF and PvCO2qexpR for PvCO2, PvActinqexpF and PvActinqexpR for PvActin, AtFTF and AtFTR for AtFT, AtActinF and AtActinR for AtActin) in real-time PCR experiments and primers used (PvCO1expF and PvCO1expR for PvCO1, PvCO2expF and PvCO2expR for PvCO2, ActinF and ActinR for PvActin) in semi-quantitative PCR. PCR primers are listed in Additional file 1: Table S1. When PCR analyses were conducted using plasmid DNA harboring complete PvCO1 and PvCO2 cDNA as templates, no cross-amplification was detected. In this study, PvActin and AtActin genes were used as reference genes for normalization because they have stable expression pattern .
The open reading frame (ORF) of PvCOL cDNA was determined using the ORF Finder (https://www.ncbi.nlm.nih.gov/orffinder/) and translated into the corresponding amino acid sequence. The conserved domain was predicted using CD search (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). The predicted protein sequence alignments were performed via Clustalw, and the results of multiple sequence alignments were displayed by GENEDOC (http://iubio.bio.indiana.edu/soft/molbio/ibmpc/genedoc-readme.html). Phylogenetic analysis and statistical neighbor-joining bootstrap tests of the phylogenies were performed by MEGA version 5.0 (http://www.megasoftware.net/). Bootstraps with 1000 replicates for Poisson correction model were performed to assess node support. The information on 17 COL gene family members in Arabidopsis and 17 in rice was from Griffiths et al.  and Cockram et al. , respectively. The accession numbers for all these genes were listed in Additional file 2: Table S2.
To identify the COL genes in moso bamboo (Ph. heterocycla), we downloaded the genomic DNA sequence, predicted genes and protein sequences from Peng et al.,  (http://188.8.131.52/bamboo/down.php), constructed local blast database using BioEdit software, and then used rice COL protein sequences as queries to perform BLASTp search with the expectation (e)-value threshold of 1.0e− 30. The candidate proteins containing B-box domains and CCT domains were predicted via the NCBI-CDD (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi). To ensure that these candidate proteins actually belong to the COL gene family, we deleted the proteins lacking the B-box domain or the CCT domain.
The amplicons of the PvCO1 and PvCO2 CDS regions were inserted at the 5’end of a GFP gene driven by the CaMV35S promoter. The region corresponding to the PvCO1 C-terminal (Met269-Phe384) containing the CCT domain was amplified from the plasmid harboring complete PvCO1 cDNA via PCR and fused to the 5’end of GFP. This fusion was called PvCO1 (Cterm). Transient expression of the GFP fusions in onion epidermal cells were performed as previously described . Then the onion epidermal cells were observed with a confocal laser scanning microscope.
Transformation of Arabidopsis
The ORF of PvCO1 and PvCO2 amplified from the plasmids harboring complete PvCO1 cDNA and complete PvCO2 cDNA were purified and inserted into the pCAMBIA 1301 vectors, respectively, in which the target genes were under the control of CaMV35s promoter. Then the recombinant vectors were transformed into the Agrobacterium strain GV3101, respectively. The transformed Agrobacterium strain was used to infect the WT Arabidopsis thaliana plants using floral dipping method . Transgenic Arabidopsis were screened on 1/2 Murashige and Skoog (MS) agar media containing kanamycin. Flowering time was measured in the T3 generation using lines homozygous from several independent transformation events.
Yeast two-hybrid assay (Y2H)
Both full-length ORF of PvCO1 and PvGF14c were cloned into the pGBKT7 BD vector and pGADT7 AD vector for the swapping experiment. The truncated PvCO1 fragments encoding the N-terminus region containing the two B-box domain (Met1-Leu150) was cloned into the pGBKT7 BD vector, and the other truncated PvCO1 encoding the C-terminal (Met269-Phe384), was also cloned into the pGBKT7 BD vector. Both pGBKT7 BD vector and pGADT7 AD vectors were co-transformed into the yeast strain AH109. The positive transformants were selected on SD- Leu-Trp agar medium and then transferred to SD-Trp-Leu-His-Ade agar medium to identify the interaction in yeast. The positive and negative controls were from the kits cited below. The Y2H was performed according to the BD Matchmaker Library Construction & Screening Kits instructions (Clontech, Palo Alto, CA, US).
The full-length ORF of PvCO1, the truncated PvCO1 fragments encoding the N-terminus region containing the two B-box domain (Met1-Leu150) and the other truncated PvCO1 encoding the C-terminal (Met269-Phe384) were cloned into pET28a vectors (His tag), respectively. PvGF14c was cloned into pGEX-4 T-1 vector (GST tag). GST- PvGF14c and His-PvCO1 proteins were expressed in E. coli strain Rosseta and purified with glutathione sepharose 4B (GE Healthcare). Equal amounts of GST- PvGF14c protein coupled to glutathione sepharose 4B and His-PvCO1 proteins were incubated in PBS Buffer. The beads were then washed with PBS buffer. Bound proteins were eluted in elution buffer (50 mM Tris-HCl, 10 mM reduced glutathione, pH 8.0), separated by 12% SDS–PAGE, transferred to a polyvinylidene difluoride (PVDF) membrane (BIO-RAD, USA), and immunoblotted with anti-His antibody or anti- GST (Abmart, China). After washing the membranes with PBS buffer containing 0.2% Tween (PBST), the membranes were incubated for 1 h with anti-mouse IgG conjugated to horseradish peroxidase (Proteintech, USA). Detection was performed using Clarity Western Chemiluminescence (ECL) Substrate (BIO-RAD, USA) and visualized using a ChemiDoc MP system (BIO-RAD, USA).
All statistical analyses were performed using SPSS v19.0 (SPSS Inc., Chicago, IL, US). The data from independent assays are analysed using analysis of variance (ANOVA, GLM procedure) and presented as the mean ± SD. Differences at P < 0.01 were considered highly significant.
Cloning and sequence analysis of two PvCOLs in Ph. violascens
To identify the CO homologous gene in the bamboo species, Ph. violascens, we performed a local blastn and blastp search of rice COL sequences against the Moso Bamboo Genome Annotation database (http://184.108.40.206/bamboo/down.php). Fifteen COL genes belonging to the COL gene family were obtained in this 31,987 functional annotation database. However, another two genomic DNA sequences very similar to Hd1 were screened only in draft genome sequence, and these two cDNA or protein sequences were not found in 31,987 functional annotation database. Therefore, we designed a pair of primers TOHLF1/TOHLR2 based on the conserved sequence of CO genes from rice, maize and wheat to generate these two CO homologous with genomic DNA as template from the bamboo species, Ph. violascens, using PCR amplification. Two DNA fragments (1,433 bp and 1,273 bp, respectively) were obtained with the pair of primers. Sequence analysis indicated that the two fragments showed high identity with rice Hd1 (orthologous to CO), named PvCO1 and PvCO2. We also performed blastn against the NCBI database, and obtained one COL (referred as COL1) sequence from Ph. heterocycla that had 98% identity with PvCO1 fragment sequence using PvCO1 as the query and another COL (named COL2) sequence from Dendrocalamus xishuangbannaensis that showed 92% identity with the PvCO2 fragment sequence using PvCO2 as the query.
PhCOL gene family members in Ph. violascens
In our previous study, the identities of homologous genes between Ph. heterocycla and Ph. violascens were found to be > 95%. Using PvCO1 as a query for local BLASTn analyses against the Moso Bamboo Genome DNA database, we obtained several COL genes. Among them, one (named PhCO1) showed 98.25% identity with PvCO1. Another gene (named PhCO2), having 97.99% identity with PvCO2, was obtained using the same method. This suggested that COL corresponding homologous genes in Ph. heterocycla and Ph. violascens had a close relationship. Using rice COL, PvCO1 and PvCO2 as queries, we found 19 COL gene members in the Ph. heterocycla genome. All of them contained B-box domain and CCT domain. Then, according to these sequences, we designed primers (Additional file 2: Table S2) to amplify these 19 corresponding COL gene members from Ph. violascens. The 19 PvCOL proteins (Additional file 3: Table S3) were divided into three groups based on the identities of amino acid sequences similar to Arabidopsis and rice COL proteins (Fig. 2, Additional file 4: Table S4). In addition, based on variations within the B-box region, the 19 COL proteins group into three clusters: the first group (PvCO1, PvCO2, PvCO3, PvCO4, PvCO7, PvCO8, PvCO9, PvCO10, PvCO14, PvCO15, PvCO19) has two B-box domains, the second (PvCO5, PvCO6, PvCO11, PvCO12, PvCO13) has one B-box and a second diverged B-box, the third (PvCO16, PvCO17, PvCO18) has one B-box (Additional file 5: Figure S1). In group one, two homologous pairs, PvCO1/PvCO9 and PvCO2/PvCO3/PvCO10, had the highest sequence similarity to rice OsA/Hd1. Another homologous pair, PvCO14/PvCO19 was highly homologous to rice OsE. The fourth pair, PvCO8/PvCO7, had the closest relationship to the rice OsD. In the second group, there was one pair, PvCO12/PvCO13 with the highest identity to OsP. These data suggest that COL tandem duplication may have occurred in the genome of the bamboo species, Ph. violascens, during species evolution.
Expression of PvCO1 and PvCO2 genes
To examine the temporal expression pattern of the two target genes, real-time PCR and RT-PCR analysis was performed using total RNA isolated from field-collected bamboo leaves. Fig. 3d and e showed that expression of both PvCO genes fluctuated. PvCO1 mRNA level increased after March 12, maintained a high level during flowering from March 22 to May 21, and sharply declined to the initial level after May 21. This suggested that the transcript of the PvCO1 gene was present during flowering period. PvCO2 mRNA abundance was detectable at low levels during the entire period from March 2 to May 11, and then quickly increased.
The circadian expression of the PvCO1 and PvCO2 genes was confirmed under natural conditions using total RNA isolated from leaves collected at different times within one day. The samples were taken 3 h apart, starting at 8:00 am and ending at 5:00 am. The PvCO1 gene expression level increased at dusk and maintained a high level throughout the night. However, the transcript accumulation of PvCO2 was higher in the morning than at other times (Fig. 3f and g). These expression pattern results suggested that PvCO1 might be associated with flowering in Ph. violascens.
Overexpression of PvCO1 delays the flowering time under LD conditions in Arabidopsis
As shown in Fig. 4b, PvCO1 transgenic lines flowered significantly later 10–15 d than WT plants. The PvCO2 transgenic line was similar to WT Arabidopsis. PvCOLs transcript levels were studied in WT Arabidopsis and several independent transgenic lines overexpressing PvCO1 or PvCO2 using the total RNA isolated from 14 d seedlings at ZT 4 h under LD. As expected, compared to WT, expression levels of PvCOLs in different transgenic lines were significantly increased. However, there was no correlation between the PvCOLs mRNA abundance and the flowering time (Fig. 4c and d). These results indicated that PvCO1 repressed flowering times in Arabidopsis under LD conditions and this might suggest the possibility that PvCO1 represses flowering in Ph. violascens.
PvCO1 negatively affect AtFT expression
PvCO1 can interact with 14–3-3(PvGF14c) protein
Both PvCO1 and PvCO2 localize to nucleus
Bamboo grown under natural conditions has a wide range of flowering times. Some species have a lengthy vegetative stage that may last more than 120 years while other species flower after only 1 year. Little is known about floral induction in bamboo or the genes involved in the process [6, 15, 16, 17, 42].
We identified and characterized two genes, PvCO1 and PvCO2 in the bamboo Ph. violascens. These genes are homologs of the CO in Arabidopsis and Hd1 in rice. Both PvCOL genes consisted of two exons and one intron (Fig. 1a). They shared two B-box domains containing typical zinc finger structures near the N-terminal and a conserved CCT domain near the C-terminal (Fig. 1b), suggesting that both encode transcriptional factors. The predicted protein sequences of PvCO1 and PvCO2 had low similarity with CO, but the two B-box domains and the CCT domain were highly conserved and they showed high similarity to Hd1. The identity between PvCO1 and PvCO2 was 70.91%. The alignment of both PvCOLs, Hd1, and CO indicates that the COL protein family has been conserved in bamboo (Fig. 1b, Additional file 5: Figure S1).
CO belongs to a gene family composed of 17 COL genes in Arabidopsis . There are similar numbers of COL genes in the genomes of rice, sorghum and foxtail millet . A total of 19 COL genes were identified in Ph. violascens, indicating that the CO gene family in bamboo also has many members. Based on the number and variation of the B-box, these 19 PvCOL proteins were classified into three groups. Group I contains two B-boxes, group II has a B-box and a second diverged B-box, and group III contains one B-box (Additional file 5: Figure S1). The second diverged B-box lacks C or H residues, or has a substitution of the conserved C or H residue. However, the CCT domain of these COL proteins shows high similarity among rice, sorghum, foxtail millet, and bamboo. Excluding the B-box and CCT regions, the remaining regions had high variation in the COL proteins among the four species. Phylogenetic analysis of the COL gene family in Arabidopsis, rice and bamboo also resulted in three groups (Fig. 2). All of the COL genes in Ph. violascens had corresponding genes in rice. In every group, there was a gene pair in bamboo corresponding to a single gene in rice. For example, PhCO14/ PhCO19 (92.49% identity) had highest homologies to rice OsE. Group I includes most of the genes known to have COL function in other species and contains 11 of the 19 PvCOLs. These results correspond with the multiple genome duplication events that have occurred in bamboo. Analysis of single-copy genes and gene families that contained 2–4 gene members show fewer single-member gene families and more two-member families in the Ph. heterocycla genome than in the rice genome . Collinearity investigation of orthologous genes between bamboo and rice indicated that a whole-genome duplication event occurred in bamboo . Similar to maize, there may have been a tetraploidization event(s) during bamboo evolution . In rice, most of the seventeen COL genes form paralogous gene pairs (OsC-OsD, OsE-OsF, OsK-OsL, OsM-OsN, OsO-OsQ, OsP-OsR) . However, phylogenetic trees showed that COL genes in Ph. violascens had greater homology compared to gene pairs from rice. This may indicate that the divergence of gene pairs of COL predated the bamboo/rice divergence.
In most flowering plants, the activity of CO and its orthologous genes are regulated by photoperiod and shows a circadian rhythm that varies among different species. In Arabidopsis, CO expression, and control of flowering, is regulated by the circadian clock [19, 20]. CO expression was modulated by the circadian clock and day length and it peaked twice (dawn and dusk) under LD conditions [20, 43] and accumulated mostly during the dark period under SD conditions . In rice, the Hd1 mRNA level was low at midday and highest during the night regardless of LD or SD conditions . Tomatoes (Lycopersicum) are a day-neutral species and the effect of day length on peak expression time of the TCOL1 and TCOL3 genes was similar to that of CO in Arabidopsis . In Populus trichocarpa, PtCO2 expressed in a diurnal pattern, peaking at the end of the day under LD conditions and having a low expression peak at night under SD conditions . In Ph. violascens, PvCO1 displayed a diurnal pattern with higher expression during the night than the day under a LD 12.5:11.5 h photoperiod. This diurnal expression pattern was not completely consistent with, but similar to the expression pattern of other CO orthologous genes, suggesting that light and the circadian clock modulated the peak of PvCO1. Our expression studies also show that PvCO1 mRNA abundance accumulated from 5:00 pm to 5:00 am, unlike the high expression of PvFT1 from 2:30 pm to 8:30 pm . The time lag between the expression of PvCO1 and PvFT1 suggests the possibility that the CO/FT regulatory module is not strongly conserved and that there are unidentified mechanisms necessary for PvFT1 induction in Ph. violascence. CO and its orthologous gene regulated flowering time at a low level of expression. They were detected only by RT-PCR and were not found in the libraries screened [21, 29]. Their transcripts were not tissue-specific and were present in most of tissues examined [21, 29, 46, 47, 48, 49, 50]. Corresponding cDNA sequences of the two PhCOL genes in moso bamboo, PhCO1 and PhCO2, were not found in the 31,987 protein-coding gene database (http://220.127.116.11/bamboo/), and both PvCOL genes were detected only by RT-PCR, demonstrating that their mRNA accumulation was very low. Despite high sequence similarity between PvCO1 and PvCO2 which had highest identity to OsA/Hd1, they showed highly diverse expression patterns. The transcripts of PvCO1 were detected in leaves and stems, exhibited higher expression during the flowering stage, than stages before and after flowering, and showed a daily pattern under natural conditions with higher expression in darkness than in daylight, similar to CO/Hd1 of Arabidopsis and rice. In contrast, PvCO2 was only found in leaves and very low during the flowering stages, its expression levels were higher in the morning than at other times under the same conditions. These data suggested that PvCO1 and PvCO2 genes were functional differentiation and PvCO1 had more relevance to the flowering process while PvCO2 might be a paralog of CO/Hd1.
Based on the amino acid sequence, COLs belong to the zinc finger gene family and act as transcription factors [21, 22, 51, 52]. They are localized in the nucleus and bind the promoter region of downstream target gene FT to activate FT transcription by itself or via complex formation [24, 27]. Either the B-box domain or CCT domain can interact with the other protein and then induce FT transcription . Song et al.  reported that through the B-box domain, the CO protein can partially regulate FT transcription by forming a complex with ASYMMETRIC LEAVES 1 (AS1) protein. The B-box domain can also interact with the TGA4 transcription factor [22, 54]. Through the CCT domain, CO interacts with HEME ACTIVATOR PROTEIN (HAP) trimetric transcription factor complex, which regulates FT expression , as well as the COP1-SPAs E3 ubiquitin ligase complex, and then stabilizes CO protein [55, 56]. CCT may also function as a nuclear-localization signal for protein transport [21, 22]. In addition, previous reports suggest that the 14–3-3 proteins μ and υ influence the flowering transition and can interact with CO protein in Arabidopsis . Our studies confirmed that PvCO1 was localized in the nucleus via CCT domain, and PvCO2 was also localized in the nucleus. PvCO1 could interact with 14–3-3 protein c (PvGF14c) through the B-box domain but not the CCT domain (Fig. 6). In the plant, 14–3-3 proteins could influence their binding partners at the spatiotemporal and subcellular levels as well as post-translational modification and stability . Whether the interaction between PvCO1 and PvGF14c could affect the PvCO1 stability and nuclear transport need ongoing studies.
In other plant species, the expression and function of CO genes may be less conserved. CO can act as a inducer of floral transition in Arabidopsis, rice, potato, tomato, soybean and sugar beet [29, 34, 43, 58, 59, 60, 61]. It is unclear the extent to which CO function has been preserved in poplar. Bo¨hlenius et al.  reported that the CO2/FT1 regulon controls the onset of reproduction in poplar, whereas, Hsu et al.  indicated that overexpression of CO1 and CO2 singly or together did not alter normal reproductive onset of poplar. In long-term field trials, overexpression of CO1 was able to complement the Arabidopsis co-2 mutant under long days. None of the eight MtCOL genes in Medicago truncatula could rescue the late-flowering phenotype of co Arabidopsis . In contrast, the group I genes CO3, OsCO3/OsB and OsCOL4/OsD, group II gene OsCOL10, OsCOL13, OsCOL16, as well as the group III gene AtCOL9 inhibited flowering [64, 65, 66, 67, 68, 69]. Our data showed that PvCO1, but not PvCO2, regulated the flowering time by reducing the expression of FT in Arabidopsis, because overexpression of PvCO1 caused floral delay, and overexpression of PvCO2 had no influence on the flowering time of Arabidopsis under long day conditions. Phylogenetic analysis showed that PvCO1 and PvCO2 clustered together with Arabidopsis CO and rice Hd1. This suggests the possibility of PvCO2 evolving a novel function, having no role in flowering regulation, or acting redundantly with other flowering regulators in Arabidopsis. The results indicate that the functions of CO in regulating flowering time are complex and diverse. It is likely that the long period of bamboo vegetative growth is related to the flowering inhibition regulator of PvCO1.
Two COL genes, PvCO1 and PvCO2, from Ph. violascens were identified. Both genes had different expression patterns. The expression of PvCO1 was related to floral transition, but expression of PvCO2 was not. Levels of both PvCO1 and PvCO2 mRNA displayed a circadian pattern. Overexpression of PvCO1 delayed flowering in Arabidopsis, while overexpression of PvCO2 has no effect on Arabidopsis flowering time. The long period of vegetative growth of bamboo may be related to an inhibition regulator of PvCO1.
The authors thank Dr. J. Chen for providing meteorological data, Dr. Q.Y. Zeng for his critical reading of the manuscript and L Jin help writing some of the manuscript.
This work was supported by the following grants: National Natural Science Foundation of China (grant no. 30901155, 31270715), Natural Science Foundation of Zhejiang Province (grant no. Y307499). The Funding bodies were not involved in the design of the study and collection, analysis, and interpretation of data and in writing the manuscript.
Availability of data and materials
The datasets generated and analysed during the current study are available from the corresponding author on reasonable request.
XQG and GHX designed research; GHX and BJL performed research; HJC, WC and RYG analyzed bioinformatic data; ZYW analyzed interaction data; BZM analyzed intracellular localization data. All authors have read and approved the final manuscript.
Ethics approval and consent to participate
Consent for publication
The authors declare that they have no competing interests.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
- 7.Zheng YS, Gao PJ, Chen LG. A study on the physiological and biochemical character of flowering for Dendroclamopsis oldhami. Sci Silvae Sinicae. 2003;39:143–7.Google Scholar
- 8.Zhan AJ, Li ZH. The nutrition dynamic of N, P, K in umbrella bamboo (Fargesia murielae) before and after flowering. J Wuhan Bot Res. 2007;25:213–6.Google Scholar
- 14.Gao J, Ge W, Zhang Y, Cheng Z, Li L, Hou D, et al. Identification and characterization of microRNAs at different flowering developmental stages in moso bamboo (Phyllostachys edulis) by high-throughput sequencing. Mol Gen Genet. 2015:1–19.Google Scholar
- 36.Ye SW, Cai CY, Ren HB, Wang WJ, Xiang MQ, Tang XS, et al. An efficient plant regeneration and transformation system of ma bamboo (Dendrocalamus latiflorus Munro) started from young shoot as explant. Front Plant Sci. 2017;8:1298. https://doi.org/10.3389/fpls.2017.01298.CrossRefPubMedPubMedCentralGoogle Scholar
- 42.Zheng ZG, Yang XM, Fu YP, Zhu LF, Wei H, Lin XC. Overexpression of PvPin1, a bamboo homolog of PIN1-Type Parvulin 1, delays flowering time in transgenic Arabidopsis and rice. Front Plant Sci. 2017. https://doi.org/10.3389/fpls.2017.01526.
- 53.Tripathi P, Carvallo M, Hamilton EE, Preuss S, Kay SA. Arabidopsis B-BOX32 interacts with CONSTANS-LIKE3 to regulate flowering. Proc Natl Acad Sci U S A. 2017;3:114,172–7.Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.