To isolate a Pin1-like gene from P. violascens, the amino acid sequences of Pin1 homologs from grass family plants were compared and the primers from the conserved regions were designed. Then a 300-bp fragment of the PvPin1 gene was amplified from P. violascens. Using gene-specific primers, a 744-bp cDNA sequence of Pin1-like was isolated from P. violascens by using 3′ and 5′ rapid amplification of cDNA ends (RACE) and designated as PvPin1. DNA sequencing analysis showed that the 744-bp cDNA contained a complete open reading frame (ORF) encoding a polypeptide of 122 amino acids. Based on the cDNA sequence of PvPin1, a 3078-bp genomic DNA sequence was cloned. Comparison of the genomic DNA sequence and ORF sequences showed that PvPin1 had two exons (248 and 121 bp) and one intron (2709 bp) (Figure 1A), which was same as Pin1At (Landrieu et al., 2000) and Pin1-like in rice (NCBI Oryza sativa Japonica Group Annotation Release 101). Amino acid sequence alignment revealed that the PvPin1 protein had the theoretical values of 7.97/13148.7 pI/Mw. Secondary structure analysis with SOPMA indicated that the putative PvPin1 protein contained an alpha helix (38.52%), a beta turn structure (16.39%), and a random coil (38.52%). Sequence comparison of the PvPin1 protein with its homologs in other plants showed that the catalytic core was well conserved and contained only a PPIase domain with four additional amino acids (Figure 1B).

The sequences of Pin1-like from more than 16 plant species were downloaded from NCBI. Phylogenetic comparison of the PvPin1 protein with homologs in other plants species showed that PvPin1 belongs to the monocots clade and is closely related to Pin1-like proteins from O. sativa, O. brachyantha, Dichanthelium oligosanthes, Zea mays, Aegilops tauschii, and Brachypodium distachyon, especially Pin1-like in P. edulis, which is the affinis species of P. violascens, having the highest identity (94.26%) with PvPin1 (Figure 1C).

RT-qPCR was used to characterize the expression pattern of PvPin1 in young and mature leaf, floral bud, culm, bamboo shoot, and rhizome tissue in flowering bamboo plants. Although PvPin1 transcripts were detectable in almost all tested organs, its expression was highest in young leaves and lowest in floral bud (Figure 2A).

Also, we used RT-qPCR to detect the temporal expression of PvPin1 in young bamboo leaves at different flowering stages from March 15 to April 12. PvPin1 expression peaked on March 22, before flowering. Although the PvPin1 transcript level increased significantly in young leaves of flowering bamboo plants at the early stage, it declined in leaves during flower development (Figure 2B).

To determine whether the expression patterns of PvPin1, Pin1At, and PePin1 (Pin1-like in P. edulis) were associated with the regulation of their promoters, we compared and analyzed their promoter sequences. An upstream 1500-bp sequence of PvPin1’s start codon was cloned, and the same length promoter sequences of Pin1At and PePin1 were downloaded from the NCBI database. On sequence alignment, the promoter sequence of PvPin1 shared 39.09% and 60.43% similarity with those of Pin1At and PePin1, respectively. The potential cis-acting regulatory elements were predicted by using PlantCARE. The typical CAAT-box and TATA-box core elements and other elements involving in light-responsive (LTR), MeJA-responsive (CGTCA-motif), endosperm expression (Skn-1_motif), and anaerobic responsive (ARE) were commonly found in these three promoters (Figure 3). However, the ABRE and TGACG-motif cis-acting elements, which are regulated by ABA and MeJA, respectively, were specific to the promoter sequences of bamboo (PvPin1 and PePin1).

P. violascens plants were treated with ABA and MeJA because the ABRE and TGACG-motif elements were specific to the promoter sequences of bamboo (PvPin1 and PePin1). The mRNA level of PvPin1 in leaf was significantly lower with ABA and MeJA than mock treatment (Figure 4), which suggested that PvPin1 can respond to ABA and MeJA.

Establishing a regeneration and genetic transformation system in bamboo is difficult (Zang et al., 2016, 2017). To examine the function of PvPin1 in regulating flowering, we overexpressed PvPin1 under control of the CaMV 35S promoter in the pCAMBIA1301 vector in Arabidopsis. Six independent lines in the homozygous T3 generation grown under greenhouse conditions were chosen for further analysis. 35S::PvPin1 transgenic Arabidopsis showed a significantly late flowering phenotype (Figures 5A,B).

To understand whether the phenotypic alteration of flowering time in transgenic Arabidopsis was related to the expression of PvPin1, we detected the expression of PvPin1 in six homozygous lines of 35S::PvPin1 Arabidopsis. RT-qPCR revealed a positive association between flowering time and the expression of PvPin1 in transgenic Arabidopsis (Figures 5B,C).

In Arabidopsis, SOC1 and AGL24 are important regulatory genes locating at the convergence of the multiple floral induction pathways. We determined the transcript levels of SOC1 and AGL24 in transgenic Arabidopsis by RT-qPCR. The transcript levels of SOC1 and AGL24 in 35S::PvPin1 transgenic plants were greatly decreased (Figure 5D). Hence, PvPin1 delayed the flowering time in Arabidopsis by downregulating the expression of SOC1 and AGL24.

To further examine its function, PvPin1 was transformed into O. sativa (Dongjing), a member of the same grass family as bamboo. We analyzed the flowering time in six independent lines in the homozygous T3 generation that were grown under field conditions. 35S::PvPin1 transgenic rice plants showed a significantly late-flowering phenotype (Figures 6A,B). Moreover, days to heading were positively associated with the expression of PvPin1 in transgenic rice (Figures 6B,C).

OsMADS50 and OsMADS56 are two SOC1 homolog genes in rice. Because SOC1 expression was markedly decreased in 35S::PvPin1 transgenic Arabidopsis plants, we determined the expression of OsMADS50 and OsMADS56 in transgenic rice lines 4, 5, and 6. OsMADS56 expression was greatly increased in these lines, with no significant difference in expression of OsMADS50 in comparison with wild-type rice (Figure 6D), so overexpression of PvPin1 inhibited flowering in transgenic rice by upregulating the expression of OsMADS56.

Pin1At localizes in both the nucleus and cytoplasm (Wang et al., 2010). We used the infection method (Escobar et al., 2003) to determine the subcellular localization of PvPin1 protein. The fusion protein PvPin1-GFP located in the nucleus and cytolemma of epidermal cells of tobacco (Nicotiana benthamiana), whereas as a control, the GFP protein distributed in the whole tobacco cells (Figure 7). The different localization between Pin1At and PvPin1 implies that they might have different function.

P. violascens samples were collected from the campus of Zhejiang Agriculture and Forestry University. Wild-type (ecotype Columbia) and transgenic plants of A. thaliana were cultivated in a controlled temperature room under 22°C with 16-h light/8-h dark. N. benthamiana was grown in a controlled temperature room under 28°C with 10-h light/14-h dark. Rice (O. sativa cv. Dongjing) plants are cultivated in the field of Lin’an (Zhejiang, China, north latitude 30°14′ and east longitude 119°42′).

Total RNA from P. violascens was isolated by using RNAiso Plus (Takara, Shiga, Japan), then Reverse Transcriptase M-MLV (Takara, Japan) was used to synthesize first-strand cDNA. A specific Pin1-like cDNA fragment (approximately 300 bp) was amplified by using the pair of primers (Pin1-1 and Pin1-2, Table 1), which were designed by comparing the amino acid sequences of Pin1 homologs from grass family plants including P. edulis (FP099633.1), O. rufipogon (CU406178.1), O. sativa (AK243434.1), Triticum aestivum (AK333419.1), and Zea mays (NM001157033.1). The 3′ end and 5′ partial cDNA of Pin1-like were isolated with the RACE kit (Invitrogen) by using gene-specific primers (3′-1 and 3′-2; 5′-1 and 5′-2, Table 1). Finally, the full-length ORF sequence was obtained by using the primers (ORF-F and ORF-R, Table 1) based on the known 5′ and 3′ sequences.

Genomic DNA was isolated by the modified CTAB method (Reichardt and Rogers, 1993) from leaves. Then a 2709-bp intron sequence of PvPin1 was obtained by using the primers ORF-F and ORF-R.

The sequence of PvPin1 ORF was used for a BLAST search in the transcript online database for P. edulis (affinis species of P. violascens) (Peng et al., 2013)¹. A sequence (ID: FP099633.1) that exists between PH01001300G0520 and PH01001300G0540 with the highest identity to PvPin1 was identified. Then the correlative genomic sequence in PH01001300 was extracted from the genome database of P. edulis (Peng et al., 2013) and used to design the primers (S1-F, S1-R; S2-F, S2-R, Table 1) for amplifying the promoter of PvPin1. A promoter sequence of 1500 bp was obtained from the DNA by using Nested PCR (Gundersen and Lee, 1996).

The RT-qPCR primers (qPCR-F and qPCR-R, Table 1) were designed by using the full-length ORF sequence of PvPin1. Here, PeUBC18 was used as the internal control gene (Qi et al., 2013; Liu et al., 2016b; Table 1) because of the close relationship between P. edulis and P. violascens. CFX96TM Real-Time PCR Detection System (Bio-Rad) and the SYBR Premix ExTaq II mix (Takara) were used for PCR amplification. The program was 95°C for 3 min, followed by 40 cycles of amplification (95°C for 15 s, 60°C for 30 s). Reactions were performed in 20-μl mixtures consisting of 10 μl 2× SYBR Premix Ex Taq II Mix, 0.5 μl each of forward or reverse primer, 1 μl cDNA template (50 ng/μl), and 8 μl double distilled H₂O (Liu et al., 2016b). The data were analyzed by the 2^-ΔΔC_t method (Livak and Schmittgen, 2001).

The full-length ORF for PvPin1 was cloned into the binary vector pCAMBIA1301 under the control of the Cauliflower mosaic virus (CaMV) 35S promoter. Recombinant vector was transferred into A. tumefaciens strain GV3101, then into Arabidopsis by the floral dip method (Clough and Bent, 1998). Transformants were screened in media with 50 μg/ml kanamycin. The same construct was also transformed into rice plants (Dongjing) mediated by A. tumefaciens strain EHA105 as described (Xu et al., 2017). Positive transgenic rice lines were confirmed by genomic PCR. The expression of SOC1 and AGL24 genes in transgenic Arabidopsis in six T3 lines and WT Arabidopsis, and the expression of OsMADS50 and OsMADS56 in transgenic rice in three T3 lines and WT rice were analyzed by real-time qPCR with gene-specific primers (Table 1) following the protocol in expression pattern of PvPin1 section “Expression Pattern of PvPin1.”

The full-length coding sequence without terminator codon (TAA) of PvPin1 was cloned into the CaMV 35S-GFP vector that allowed the system to generate a PvPin1-GFP fusion protein for investigating subcellular location in epidermal cells from tobacco (N. benthamiana) and the transient expression assay method (Escobar et al., 2003) was adopted. The tobacco epidermal cells were visualized on confocal laser scanning microscopy (LSM510, Zeiss, Germany).

A BLAST search in the NCBI database was used to obtain the protein sequences of Pin1-like. The phylogenetic tree was constructed by the neighbor-joining method with the parameter bootstrap (10,000 replicates) in MEGA 5.0. The software ProtParam from ExPASy² was used to obtain the molecular weights (MW) and theoretical isoelectric point (pI) of PvPin1 protein. PlantCARE (Lescot et al., 2002) was used to analyze cis-acting regulatory elements in the PvPin1 promoter.

Leaves of P. violascens seedlings were sprayed with ABA (100 μM), MeJA (100 μM), and water as a blank control once a day for 9 days. Every treatment was performed with three biological replicates. The PvPin1 transcript level was detected after treatment.