The Heterologous Expression of a Chrysanthemum nankingense TCP Transcription Factor Blocks Cell Division in Yeast and Arabidopsis thaliana

Both the presence of, and the important contribution to growth and development made by TCP transcription factors, have been established in various plant species. Here, a TCP4 homolog isolated from Chrysanthemum nankingense was shown to be more strongly transcribed in the diploid than in the autotetraploid form of the species. CnTCP4 was shown to encode a member of the class II TCP family and to be transcribed most strongly in the leaf and ligulate flowers. Its transcription was found to be substantially inhibited by spraying the plant with the synthetic cytokinin 6-benzylaminopurine. The transient expression of CnTCP4 in onion epidermal cells showed that its product localized to the nucleus, and a yeast one hybrid assay suggested that its product had transcriptional activation ability. The constitutive expression of CnTCP4 in fission yeast suppressed cell proliferation, inducing the formation of longer and a higher frequency of multinuclated cells. Its constitutive expression in Arabidopsis thaliana reduced the size of the leaves. The presence of the transgene altered the transcription of a number of cell division-related genes. A yeast one hybrid assay identified a second TCP gene (CnTCP2) able to interact with the CnTCP4 promoter. A transient expression experiment in Nicotiana benthamiana leaves showed that CnTCP2 was able to activate the CnTCP4 promoter. Like CnTCP4, CnTCP2 was shown to encode a member of the class II TCP family, to be transcribed most strongly in the leaf and ligulate flowers, and to be suppressed by exogenous 6-benzylaminopurine treatment. The CnTCP2 protein also localized to the nucleus, but had no transcriptional activation ability. Its constitutive expression in A. thaliana had similar phenotypic consequences to those induced by CnTCP4.


Introduction
Development in a multicellular organism requires precise control over cell proliferation and expansion. The various phases of the cell cycle each depend largely on the activity of cyclins and cyclin-dependent kinases (CDKs) [1], with a contribution in plants also being made by members of the TCP transcription factor family [2][3][4]. The characteristic feature of the TCPs is a 59 residue stretch, which forms the basic helix-loop-helix (bHLH) motif central to the proteins' ability to bind to DNA and to direct protein-protein interactions [5]. Both the Arabidopsis thaliana and rice genomes harbor over 20 genes encoding a TCP [6]. Based on variation in the TCP domain sequence, the TCPs have been divided into two distinct classes, referred to as class I/PCF/TCP-P and class II/TCP-C [7,8].
While the former act to promote cell division, the latter repress it [6,9]. The products of the A. thaliana class I genes AtTCP14 and AtTCP15 both drive cell division in young internodes [10], while the consequence of knocking out the class I genes AtTCP9 and AtTCP20 is to inhibit cell division and to promote cell expansion [11]. The product of the class II gene AtTCP4 represses cell proliferation in the A. thaliana leaf [12]. AtTCP4 activates AtGIS, so reducing its presence results in an increase in trichome branching [3].
The diploid species Chrysanthemum nankingense is a near relative of the commercially important ornamental species C. morifolium [13,14]. The product of the TCP I gene DgBRC1 represses the formation of lateral branching [15], that of CmCYC regulates the growth of ray florets [16], while that of the gene CmTCP14 suppresses the size of the organ [17]. As yet, no attempt has been made to characterize the complement of TCP genes harbored by the C. nankingense genome. An analysis of an autotetraploid C nankingense showed that its leaves were larger than those of its diploid progenitor [18]. The experiments reported here were designed to reveal the transcriptional behavior of CnTCP4 in both diploid and autotetraploid forms of the species, and to study the phenotypic effect of heterologously expressing it in both fission yeast and A. thaliana.

Characterization of a TCP4 Homolog in C. nankingense
The length of the cloned full cDNA sequence was 1812 nt, of which 1203 nt represented the coding sequence. CnTCP4 was therefore predicted to be a 400 residue protein, which included an atypical basic-helix-loop-helix domain (Figure 1c). The sequence was related to the class II set of TCP proteins, as its most closely related sequence was the product of AtTCP4 (Figure 1d). CnTCP4 transcript was abundant in both the leaf and ligulate florets of C. nankingense (Figure 1e). A qRT-PCR experiment revealed that the abundance of CnTCP4 transcript was greater in the leaf of the diploid form of C. nankingense than in the leaf of the autotetraploid form ( Figure 1a). The autetraploid plants, on the other hand, produced larger leaves than the diploid plants ( Figure 1b). CnTCP4 was significantly downregulated when plants were sprayed with the synthetic cytokinin 6-benzylaminopurine; the abundance of its transcript decreased marginally over the first 4 h following treatment, but fell by at least twofold over the period 8-24 h post treatment ( Figure 2). When the p35S::GFP-CnTCP4 fusion construct was transiently expressed in onion epidermal cells, the ensuing GFP signal was concentrated in the nucleus, while the introduction of a p35S::GFP transgene resulted in a cell-wide distribution of GFP activity (Figure 3a). The upshot of a yeast one hybrid experiment intended to reveal the gene's transcriptional activity was that yeast cells harboring either pCL1, pGBKT7-CnTCP4, or pGBKT7-CnTCP4-F3 were all able to grow on SD/-His/-Ade and to metabolize X-α-Gal, while cells harboring either the negative control plasmid pGBKT7, pGBKT7-CnTCP4-F1, or pGBKT7-CnTCP4-F2 all failed to show any galactosidase activity (Figure 3b,d). The conclusion was that the full length CnTCP4 was transcriptionally active, and that residues 201-300 were required for its transcriptional activation.
qRT-PCR experiment revealed that the abundance of CnTCP4 transcript was greater in the leaf of the diploid form of C. nankingense than in the leaf of the autotetraploid form ( Figure 1a). The autetraploid plants, on the other hand, produced larger leaves than the diploid plants ( Figure 1b). CnTCP4 was significantly downregulated when plants were sprayed with the synthetic cytokinin 6-benzylaminopurine; the abundance of its transcript decreased marginally over the first 4 h following treatment, but fell by at least twofold over the period 8-24 h post treatment ( Figure 2).  When the p35S::GFP-CnTCP4 fusion construct was transiently expressed in onion epidermal cells, the ensuing GFP signal was concentrated in the nucleus, while the introduction of a p35S::GFP transgene resulted in a cell-wide distribution of GFP activity (Figure 3a). The upshot of a yeast one hybrid experiment intended to reveal the gene's transcriptional activity was that yeast cells harboring either pCL1, pGBKT7-CnTCP4, or pGBKT7-CnTCP4-F3 were all able to grow on SD/-His/-Ade and to metabolize X-α-Gal, while cells harboring either the negative control plasmid pGBKT7, pGBKT7-CnTCP4-F1, or pGBKT7-CnTCP4-F2 all failed to show any galactosidase activity (Figure 3b,d). The conclusion was that the full length CnTCP4 was transcriptionally active, and that residues 201-300 were required for its transcriptional activation.

The Heterologous Expression of CnTCP4 Affected Cell Division of Fission Yeast
Yeast is used as a model system for studying gene function [19,20], and AtTCP4 protein has been studied in yeast cells [2]. So we investigated the function of CnTCP4 gene using the yeast system. CnTCP4, when driven by the NMT1 promoter, was induced in a medium lacking thiamine ( Figure 4a). Under inductive conditions, the proliferation of cells heterologously expressing CnTCP4 was strongly inhibited, while in a noninductive medium, transformants harboring either CnTCP4 or an empty pESPM plasmid both grew freely (Figure 4b,c). When cultured under inductive conditions, the length of cells harboring the CnTCP4 transgene was some 2.5 fold greater than that of those harboring the empty vector (Figure 4d,e), and some of those expressing CnTCP4 were multiseptated. Inspection of DAPI stained transformed cells revealed morphological differences between those expressing CnTCP4 and those harboring the empty vector ( Figure 5a): the former produced many cells harboring two nuclei or multiple nuclei (Figure 5b). The conclusion

The Heterologous Expression of CnTCP4 Affected Cell Division of Fission Yeast
Yeast is used as a model system for studying gene function [19,20], and AtTCP4 protein has been studied in yeast cells [2]. So we investigated the function of CnTCP4 gene using the yeast system. CnTCP4, when driven by the NMT1 promoter, was induced in a medium lacking thiamine (Figure 4a). Under inductive conditions, the proliferation of cells heterologously expressing CnTCP4 was strongly inhibited, while in a noninductive medium, transformants harboring either CnTCP4 or an empty pESPM plasmid both grew freely (Figure 4b,c). When cultured under inductive conditions, the length of cells harboring the CnTCP4 transgene was some 2.5 fold greater than that of those harboring the empty vector (Figure 4d,e), and some of those expressing CnTCP4 were multiseptated. Inspection of DAPI stained transformed cells revealed morphological differences between those expressing CnTCP4 and those harboring the empty vector ( Figure 5a): the former produced many cells harboring two nuclei or multiple nuclei (Figure 5b). The conclusion was that the expression of CnTCP4 suppressed cell proliferation via its effect on both cell division and cell expansion. was that the expression of CnTCP4 suppressed cell proliferation via its effect on both cell division and cell expansion.

The Effect of Constitutively Expressing CnTCP4 in A. thaliana
Two independent constitutive expressors of CnTCP4 in A. thaliana were selected among the T 3 progeny bred from the original transformed Col-0 material ( Figure 6a). Their vegetative growth was severely compromised compared to wild-type plants ( Figure 6b); although they formed a higher number of rosette leaves (Figure 6c), both their leaf length ( Figure 6d) and width ( Figure 6e) were reduced. The size of the cells formed in the epidermis of the lower leaf surface was larger than in wild-type plants (Figure 6f,g). The effect of the transgene on the transcription of a range of cell cycle marker genes was summarized in Figure 7. The abundance of CYCA1;1, CYCA3;1, CYCA3;2, CYCB2;4, CDKB1;2, CDKB2;2, CDKD;2, and CDKD;3 transcripts were lower in the transgenic plants than in wild type, and vice versa for KRP5 and E2Fc, there was no effect on the transcript abundance of either CDKC;1 or CDKG;2.

The Effect of Constitutively Expressing CnTCP4 in A. thaliana
Two independent constitutive expressors of CnTCP4 in A. thaliana were selected among the T3 progeny bred from the original transformed Col-0 material ( Figure 6a). Their vegetative growth was severely compromised compared to wild-type plants ( Figure 6b); although they formed a higher number of rosette leaves (Figure 6c), both their leaf length ( Figure 6d) and width ( Figure 6e) were reduced. The size of the cells formed in the epidermis of the lower leaf surface was larger than in wild-type plants (Figure 6f,g). The effect of the transgene on the transcription of a range of cell cycle marker genes was summarized in Figure 7. The abundance of CYCA1;1, CYCA3;1, CYCA3;2, CYCB2;4, CDKB1;2, CDKB2;2, CDKD;2, and CDKD;3 transcripts were lower in the transgenic plants than in wild type, and vice versa for KRP5 and E2Fc, there was no effect on the transcript abundance of either CDKC;1 or CDKG;2.

CnTCP2 Interacted with the CnTCP4 Promoter
Multiple potential TCP binding sites were identified within the CnTCP4 promoter fragment

CnTCP2 Interacted with the CnTCP4 Promoter
Multiple potential TCP binding sites were identified within the CnTCP4 promoter fragment ( Figure S1), and various ethylene-responsive, gibberellin-responsive, and light-responsive elements were predicted by PLACE software (Table 1). When the CnTCP4 promoter fragment was subjected to a yeast one hybrid assay, a sequence designated CnTCP2 was identified as a putative interacting transcription factor. The interaction was confirmed by testing the CnTCP2 ORF sequence in an independent yeast one hybrid assay featuring the same CnTCP4 promoter fragment (Figure 8a). In addition, a transient assay conducted in N. benthamiana leaves showed that leaves infiltrated with a pCnTCP4::LUC transgene generated less LUC activity than did leaves co-infiltrated with both pCnTCP4::LUC and p35S::CnTCP2 (Figure 8b). Table 1. An in silico analysis of the regulatory element content of the CnTCP4 promoter.

Motif
No.

Characterization of the CnTCP2 Sequence
The full length CnTCP2 cDNA sequence comprised 2066 nt, of which 1329 nt was occupied by the ORF: its gene product was 442 residues long. The sequence was classified as a class II TCP, related more strongly to AtTCP2 than to any other A. thaliana TCP gene (Figure 1b). The sequence encoded an atypical bHLH domain (Figure 1c). The gene was abundantly transcribed in ligulate florets and leaves ( Figure S2a), and was strongly downregulated in plants sprayed with 6-benzylaminopurine ( Figure S2b). The gene product was directed to the nucleus ( Figure S3a), and lacked transcriptional activation activity ( Figure S3b). Analysis of the two independent constitutive expressors of CnTCP2 in A. thaliana selected among the T3 progeny bred from the original transformed Col-0 material (Figure 9a) showed that that presence of the transgene affected growth

Characterization of the CnTCP2 Sequence
The full length CnTCP2 cDNA sequence comprised 2066 nt, of which 1329 nt was occupied by the ORF: its gene product was 442 residues long. The sequence was classified as a class II TCP, related more strongly to AtTCP2 than to any other A. thaliana TCP gene (Figure 1b). The sequence encoded an atypical bHLH domain (Figure 1c). The gene was abundantly transcribed in ligulate florets and leaves ( Figure S2a), and was strongly downregulated in plants sprayed with 6-benzylaminopurine ( Figure S2b). The gene product was directed to the nucleus ( Figure S3a), and lacked transcriptional activation activity ( Figure S3b). Analysis of the two independent constitutive expressors of CnTCP2 in A. thaliana selected among the T 3 progeny bred from the original transformed Col-0 material (Figure 9a) showed that that presence of the transgene affected growth in the same way as did the constitutive expression of CnTCP4 (Figure 9b-g). The abundance of CYCA1;1, CYCB2;4, CDKB1;2, CDKD;2, and CDKD;3 transcripts were lower in the transgenic plants than in Col-0, and vice versa for CDKC;1, CDKG;2, KRP5, and E2Fc (Figure 9h). in the same way as did the constitutive expression of CnTCP4 (Figure 9b-g). The abundance of CYCA1;1, CYCB2;4, CDKB1;2, CDKD;2, and CDKD;3 transcripts were lower in the transgenic plants than in Col-0, and vice versa for CDKC;1, CDKG;2, KRP5, and E2Fc (Figure 9h). Columns headed by a different letter denote transcript abundances differing significantly (p < 0.05) from the transcript abundance measured in Col-0 plants.

Discussion
TCP transcription factors, featuring a bHLH motif that facilitates their ability to bind to DNA and to mediate protein-protein interactions, are unique to plants, and contribute to both their growth and development [6]. Both the CnTCP4 and CnTCP2 genes harbored by C. nankingense encode class II TCPs. Proteins of this type have been implicated in the regulation of several aspects of plant development, including the development of axillary meristems and the differentiation of lateral organs [3,21,22]. Antirrhinum majus cin mutants (CIN encodes a TCP transcription factor that promotes growth arrest, particularly at the leaf margin) develop larger leaves than do wild-type plants, and their leaves have an undulating edge due to excessive growth at the leaf margin [23].

Discussion
TCP transcription factors, featuring a bHLH motif that facilitates their ability to bind to DNA and to mediate protein-protein interactions, are unique to plants, and contribute to both their growth and development [6]. Both the CnTCP4 and CnTCP2 genes harbored by C. nankingense encode class II TCPs. Proteins of this type have been implicated in the regulation of several aspects of plant development, including the development of axillary meristems and the differentiation of lateral organs [3,21,22]. Antirrhinum majus cin mutants (CIN encodes a TCP transcription factor that promotes growth arrest, particularly at the leaf margin) develop larger leaves than do wild-type plants, and their leaves have an undulating edge due to excessive growth at the leaf margin [23]. The A. thaliana JAW locus generates a microRNA, which guides the cleavage of several TCP mRNAs controlling leaf development, and its loss-of-function mutant exhibits a crinkly leaf phenotype [24]. The product of AtTCP3 is thought to be an important regulator of shoot lateral organ morphology [21], while the expression of a hyperactivated form of AtTCP4 dampens cell proliferation, producing plants with cup-shaped leaves of reduced size [25]. At the early stages of leaf development, the TIE1 protein has been shown to modulate leaf size and shape through its inhibition of TCP activity [26]. Here, it was demonstrated that the constitutive expression of CnTCP4 in A. thaliana had the effect of reducing leaf size, indicating that its product represses leaf development.
CIN-like TCP transcription factors are known to regulate growth though their interaction with phytohormone-associated pathways. For instance, AtTCP3 activates the auxin signaling repressor IAA3/SHY2 [27]. The expression of the rice gene OsTCP5 is influenced by both strigolactone and cytokinin [28]. Meanwhile AtTCP4 interacts with pathways associated with auxin, gibberellin, and abscisic acid [29]. It also promotes the synthesis of jasmonate via the upregulation of LOX2 [11,30]. In the present study, ethylene-, gibberellin-, jasmonate-and salicylic-responsive regulators were all identified within the CnTCP4 promoter sequence ( Table 1), suggestive of the participation of CnTCP4 in the phytohormone-directed regulation of plant growth, a conclusion supported by the transcriptional responsiveness of CnTCP4 to treatment with the synthetic cytokinin 6-benzylaminopurine ( Figure 2). Cytokinin 6-benzylaminopurine promoted cell division by repressing the expression of CnTCP4.
AtTCP4 has been implicated in coordinating cell division and cell differentiation in the developing leaf; since the downregulation of AtTCP4 has been shown to promote mitotic activity [24,31], the overexpression of this gene induces cells to exit the proliferation phase prematurely, resulting in the formation of smaller leaves [25]. Here, the constitutive expression of CnTCP4 in fission yeast caused the cells to elongate and to become multiseptated, while also suppressing cell division; its constitutive expression in A. thaliana reduced leaf size and expanded the size of the epidermal cells on the abaxial leaf surface. It showed a significant increase in average epidermal cell size in tcp20 mutant leaves, but no obvious size or shape alterations could be observed in the tcp20 mutant line, owing to the fact that the cell size effect was accompanied by a reduction in total cell number in the leaf [11]. Smaller leaf phenotypes could be observed in overexpressed CnTCP2/4 lines, indicating that the cell size effect is completely compensated for by a reduction of the total number of cells in the leaf. The suggestion is that these effects derive from the transgene product's suppression of cell proliferation, allied with its promotion of cell expansion. This notion is fully consistent with the observation that the tetraploid form of C. nankingense formed larger leaves than did the diploid form, while the abundance of CnTCP4 transcript was significantly lower (Figure 1b).
The timing of the onset of cell proliferation and expansion is correlated with the transition from cell division to the endocycle [32]. Cell division is orchestrated by a number of proteins, including cyclins, CDKs, and CDK inhibitors [33]. The activity of the CDK-cyclin complex is regulated by CDK-activating enzymes and CDK inhibitory proteins [34,35]. Here, all of AtCYCA1;1, AtCYCA3;1, AtCYCA3;2, AtCYCB2;4, AtCDKB1;2, AtCDKB2;2, AtCDKD;2, and AtCDKD;3 were downregulated in plants constitutively expressing CnTCP4, which may explain the observed reduction in leaf size induced by the transgene, because each of these genes acts to suppress cell division. The CDKCs are responsible for the phosphorylation of the C terminal domain of RNA polymerase II. According to Kitsios et al. (2008), the function of the plant CDKC-CycT complex may be similar to that of the mammalian CDK9/CycT complex of the positive transcription elongation factor b [36]. In A. thaliana, a loss of CDKC;2 activity promotes cell division [37]. The A. thaliana genome harbors two genes encoding G-type CDKs, one which complexes with CYCL1 to delay entry into the cell cycle [38]. AtKRP5 influences cell cycle progression at both the G1/S and G2/M transitions by inhibiting the kinase activity of the CYCD2/CDKB complex and reconstituted CYCD2-associated kinases [39]. AtE2FC is a further protein that regulates entry into the G1/S transition [40]. Here, the upregulation of AtCDKC;1, AtCDKG;2, AtKRP5, and AtE2Fc induced by the constitutive expression of CnTCP4 may have acted to inhibit cell division, thereby contributing to the reduction in leaf size observed in the transgenic plants.
BHLH proteins regulate their downstream targets by interacting with sequences within their promoter [41]. The AtTCP20 protein binds to the promoter of AtTCP9, thereby triggering its expression [11]. Based on the yeast one hybrid experiment and the transient expression assay in N. benthamiana leaves carried out here, the CnTCP2 protein was concluded as capable of binding to the CnTCP4 promoter. The constitutive expression of CnTCP2, as was also the case for CnTCP4, suppressed cell division and promoted cell expansion, which suggests the presence in C. nankingense of a TCP2-TCP4 pathway. CnTCP2 had no transcription activity, but the constitutive expression of CnTCP2 in A. thaliana also generated the same phenotype as CnTCP4 (Figures 6 and 9), it is possible that CnTCP2 binds to AtTCP4 and hence produces small plants.

Plant Materials
The accessions of diploid and autotetraploid C. nankingense used here were maintained by Nanjing Agricultural University's Chrysanthemum Germplasm Resource Preserving Centre. Plants were propagated by cutting and grown in a 1:1 (v/v) mixture of soil and vermiculite. Rooted seedlings were grown in a greenhouse maintained at 22 • C during the day and at a minimum of 15 • C during the night; the relative humidity ranged from 70% to 75%. A. thaliana (ecotype Col-0) plants were grown in a 1:1:1 (v/v/v) mixture of vermiculite, perlite, and soil under a 16 h photoperiod (80-100 µmol m −2 s −1 illumination), and a day/night temperature regime of 22 • C/18 • C.

Transcription Profiling of CnTCP4
Transcription profiling was performed using quantitative real-time PCR (qRT-PCR), based on the primer pair CnTCP4-QF/QR (sequences given in Table S1). Total RNA was extracted from C. nankingense tissue using the RNAiso reagent (TaKaRa, Tokyo, Japan) according to the manufacturer's instructions. Each qRT-PCR assay was represented by three technical replicates, and each sample by three biological replicates. The C. nankingense EF1α gene was used as the reference (primer sequences given in Table S1). The reactions were initially denatured (95 • C/2 min), then cycled 40 times through 95 • C/15 s, 55 • C/15 s, 72 • C/20 s. Relative transcript abundances were calculated using the 2 − Ct method [42].

Cloning and Sequencing of CnTCP4 cDNA
Diploid C. nankingense leaves were snap-frozen in liquid nitrogen, and total RNA isolated using the RNAiso reagent (TaKaRa, Tokyo, Japan), following the manufacturer's instructions. The first cDNA strand was synthesized from 1 µg total RNA using M-MLV reverse transcriptase (TaKaRa) according to the manufacturer's instructions. The primer pair CnTCP4-F/-R (sequences given in Table  S1) was designed to amplify a fragment of CnTCP4 based on a sequence represented in a transcriptome database [43]. The full length cDNA sequence was obtained by applying 5'-and 3'-RACE PCR. For the 3'-RACE reaction, the first cDNA strand was generated using the dT adaptor primer, following by a nested PCR based on the primer pairs CnTCP4-3'-F1/F2/F3 and the Adapter-R adaptor primer (sequences given in Table S1). For the 5' reaction, the nested PCR used the primers AAP and AUAP provided with a 5' RACE System kit v2.0 (Invitrogen, Carlsbad, CA, USA), along with the gene-specific primers CnTCP4-5'-F1/-F2/-F3 (sequences given in Table S1). The PCR products were purified using an AxyPrep DNA Gel Extraction kit (Axygen, Shanghai, China) and inserted into pMD19-T (TaKaRa) for sequencing. Finally, a gene-specific primer pair (CnTCP4-ORF-F/-R, sequences given in Table S1) was designed to amplify the gene's complete open reading frame (ORF) sequence.

Phylogenetic Analysis
A. thaliana TCP sequences were downloaded from the Arabidopsis Transcription Factors database [44], and combined with the newly acquired CnTCP4 sequence to perform a multiple alignment analysis based on ClustalW software [45]. The subsequent phylogenetic analysis was based on the neighbor-joining method, as implemented in MEGA v6 software [46]. The robustness of each dendrogram branch was estimated by means of a bootstrap analysis (1000 replicates).

Phytohormone Treatments
The diploid form of C. nankingense at the 6-8 leaf stage were chosen to assess the transcriptional response of CnTCP4 to 5.0 µM 6-benzylaminopurine treatment [47]. Control plants were sprayed with distilled water. The leaves were sampled both before the first spray treatment, and then again after 1, 2, 4, 8, 12, and 24 h. Each treatment was replicated three times. The leaf samples were snap-frozen in liquid nitrogen and stored at −70 • C.

Subcellular Location of CnTCP4
The CnTCP4 ORF sequence (lacking the stop codon) was amplified with the primer pair CnTCP4-pENTR1A-F/R (sequences given in Table S1) using a Phusion High-Fidelity PCR kit (New England Biolabs, MA, USA). Both the amplicon and the pENTR™ 1A vector (Invitrogen) were restricted with BamHI and NotI, and the products ligated using T4 DNA ligase (TaKaRa). Following its validation by sequencing, the pENTR™1A-CnTCP4 fusion product was recombined with pMDC43 using the LR Clonase™ II enzyme mix (Invitrogen) to form the construct p35S::GFP-CnTCP4. The p35S::GFP-CnTCP4 transgene or an empty pMDC43 vector were transiently transformed into onion epidermal cells using a PDS-1000 particle accelerator (Bio-Rad, Hercules, CA, USA) according to the manufacturer's instructions. The epidermal peels were incubated in the dark at 22 • C for 16 h on Murashige and Skoog (MS) medium [48]. The expression of GFP was observed by confocal laser microcopy (Leica SP2, Wetzlar, Germany).

Yeast Transformation, Growth, and Observation
The CnTCP4 coding sequence was amplified using primer pair CnTCP4-XhoI-F/ CnTCP4-BamHI-R (sequences given in Table S1) and the resulting amplicon inserted into pESPM via its XhoI and BamHI cloning sites to form the construct pESPM-CnTCP4. This construct (and in parallel, as a control, an empty pESPM plasmid) was transformed into Schizosaccharomyces pombe Leu -SPQ-01 using a lithium acetate-mediated method, as recommended in a commercial protocol (Clontech), and transformants were selected on Edinburgh minimal medium (EMM) agar plates containing thiamine held at 30 • C. Selected cells were then cultured in liquid EMM containing thiamine held at 30 • C up to the mid-exponential phase, after which they were flushed three times with EMM lacking thiamine to de-repress the NMT1 promoter. The same transgenic mid-exponential cells were then streaked onto liquid EMM (either with or without thiamine) and incubated at 30 • C for 30 h, and cell densities were monitored by measuring the OD 600 . The same number of transformed mid-exponential yeast cells were also streaked onto solid EMM medium with or without thiamine and incubated at 30 • C for 60 h. CnTCP4 transcription was monitored using an RT-PCR assay (primer pair CnTCP4-XhoI-F/CnTCP4-BamHI-R, sequences given in Table S1), after total RNA had been isolated from the cells using a Fungi RNA kit (Hua Yue Yang, Beijing, China). The S. pombe Tubulin gene was used as the reference (primer sequences given in Table S1). Cells were stained in a 1 mg/mL solution of 4,6-diamidino-2-phenylindole (DAPI) and monitored by fluorescence microscopy (Zeiss Axioskop40, CarlZeiss, Jena, Germany). Mean cell lengths were compared using routines implemented in SPSS v17.0 software (SPSS Inc., Chicago, IL, USA).

A. thaliana Transformation and Transcriptional Profiling of Cell Cycle Marker Genes
The p35S::GFP-CnTCP4 construct was introduced into Agrobacterium tumefaciens strain EHA105, and from thence into A. thaliana Col-0 using the floral dip method [49]. Transformants were selected by culturing on half strength MS medium agar plates containing 20 µg/mL hygromycin, and were advanced to the T 3 generation. The zygosity of the transgene was identified using a RT-PCR assay based on the primer pair CnTCP4-ORF-F/R (sequences given in Table S1). The size of epidermis cells from the abaxial surface of the fifth leaf of ten independent 35-day-old T 3 plants was measured from phase contrast micrographs. The transcriptional behavior of selected cell cycle marker genes was derived from qRT-PCR assays, with the AtActin gene chosen as the reference (primer sequences given in Table S1). RNA was extracted from the fifth leaves of 28-day-old wild-type and transgenic plants using the RNAiso reagent (TaKaRa) according to the manufacturer's instructions.

Cloning of CnTCP4 Promoter
Genomic DNA was extracted from leaves of diploid C. nankingense using a modified CTAB method [50], and used as a PCR template (primed by CnTCP4-DNA-F/-R, sequences given in Table S1) to scan for the presence of introns in the coding sequence. The CnTCP4 promoter was cloned from the genomic DNA using an LA PCR TM in vitro Cloning kit (TaKaRa), according to the manufacturer's instructions, using the primer pair promoter-AP/CnTCP4-pro-SP1 (sequences given in Table S1). A second nested round of PCR was performed using 1 µL of the initial PCR as the template and the primer pair promoter-AP/CnTCP4-pro-SP2 (sequences given in Table S1). The final nested PCR used 1 µL of the second PCR as the template and the primer pair promoter-AP/CnTCP4-pro-SP3 (sequences given in Table S1). Primer pair CnTCP4-pro-F/R (sequences given in Table S1) was then used to amplify the CnTCP4 pro -1336 sequence. Functional elements of CnTCP4 promoter were predicted by PLACE (http://www.dna.affrc.go.jp/PLACE/signalup.html).

Yeast One Hybrid Assay
The primer pair CnTCP4-PHIS-F/R (sequences given in Table S1) was used to amplify the CnTCP4 pro -1336 sequence and the amplicon was inserted into the pHISi plasmid (Clontech), which contains the HIS3 reporter gene and a minimal HIS3 promoter, to generate the reporter construct pHISi-CnTCP4 pro -1336. The recombined plasmid was linearized by digestion with AflII, and the resulting fragment introduced into brewers' yeast (S. cerevisiae) strain YM4271 (Clontech) following the manufacturer's protocol. The cells were cultured on SD/-His/-Ura medium for the selection of cells harboring the reporter plasmid, and then cultured on SD/-His/-Ura medium containing different 3-amino-1,2,4-triazole (3-AT) concentrations to select optimal 3-AT concentration. The selected cells, along with a premade Y1H cDNA library, were transformed into S. cerevisiae strain YM4271, and transformants were selected on SD/-His/-Ura/-Leu medium containing 30 mM 3-AT.
A gene-specific primer pair (CnTCP2-ORF-F/-R, sequences given in Table S1) was designed to amplify the complete open reading frame sequence of CnTCP2. The CnTCP2 ORF sequence was inserted into the yeast expression plasmid pDEST-GADT7 (Clontech) to generate the effector plasmid pDEST-GADT7-CnTCP2 (CnTCP2-pENTR1A-F/R, sequences given in Table S1).
The pHISi-CnTCP4 pro -1336 reporter plasmid and the pDEST-GADT7-CnTCP2 effector plasmid were transformed into S. cerevisiae strain YM4271, and the transformants selected by culturing on a SD/-His/-Ura medium. Selected transgenic cells were then cultured on SD/-His/-Ura/-Leu medium containing 30 mM 3-AT to monitor the activation of the CnTCP4 promoter.

Transient Expression Assay in Nicotiana benthamiana Leaves
A transient assay was used in N. benthamiana leaves to investigate whether CnTCP2 can interact with the CnTCP4 promoter in planta. The primer pair CnTCP4-LUC-F/-R (sequences given in Table S1) was used to amplify the CnTCP4 promoter and the amplicon was ligated into pCAMBIA1381Z-Luc (Clontech) to generate the reporter construct pCAMBIA1381Z-Luc-CnTCP4-Promoter. Overnight culturezzs of Agrobacterium tumefaciens GV3101 harboring the reporter construct, the effector construct p35S::GFP-CnTCP2, or an empty P19 plasmid were suspended in infiltration buffer (2 mM NaH 2 PO 4 , 50 mM MES, 100 µM acetosyringone, and 0.5% Glc) in sufficient density to give an OD 600 of 0.2. Leaves of 42-day-old N. benthamiana plants were spot-infiltrated with the transgenic Ag. tumefaciens cells, following the Wang et al. (2014) protocol [51]. The experiment group was performed with cultures of P19 plus 35S::CnTCP2 and pCAMBIA1381Z-Luc-CnTCP4-Promoter at a ratio of 1:1:1. The control group were performed with P19 plus GV3101 at a ratio of 1:2, P19 plus GV3101 and 35S::CnTCP2 at a ratio of 1:1:1, P19 plus GV3101 and pCAMBIA1381Z-Luc-CnTCP4-Promoter at a ratio of 1:1:1. Luciferase was imaged 48-96 h after infiltration, following the Walley et al. (2007) protocol [52].