Genome-wide identification of the TCP gene family in Chrysanthemum lavandulifolium and its homologs expression patterns during flower development in different Chrysanthemum species

TCP proteins, part of the transcription factors specific to plants, are recognized for their involvement in various aspects of plant growth and development. Nevertheless, a thorough investigation of TCPs in Chrysanthemum lavandulifolium, a prominent ancestral species of cultivated chrysanthemum and an excellent model material for investigating ray floret (RF) and disc floret (DF) development in Chrysanthemum, remains unexplored yet. Herein, a comprehensive study was performed to analyze the genome-wide distribution of TCPs in C. lavandulifolium. In total, 39 TCPs in C. lavandulifolium were identified, showing uneven distribution on 8 chromosomes. Phylogenetic and gene structural analyses revealed that ClTCPs were grouped into classes I and II. The class II genes were subdivided into two subclades, the CIN and CYC/TB1 subclades, with members of each clade having similar conserved motifs and gene structures. Four CIN subclade genes (ClTCP24, ClTCP25, ClTCP26, and ClTCP27) contained the potential miR319 target sites. Promoter analysis revealed that ClTCPs had numerous cis-regulatory elements associated with phytohormone responses, stress responses, and plant growth/development. The expression patterns of ClTCPs during capitulum development and in two different florets were determined using RNA-seq and qRT-PCR. The expression levels of TCPs varied in six development stages of capitula; 25 out of the 36 TCPs genes were specifically expressed in flowers. Additionally, we identified six key ClCYC2 genes, which belong to the class II TCP subclade, with markedly upregulated expression in RFs compared with DFs, and these genes exhibited similar expression patterns in the two florets of Chrysanthemum species. It is speculated that they may be responsible for RFs and DFs development. Subcellular localization and transactivation activity analyses of six candidate genes demonstrated that all of them were localized in the nucleus, while three exhibited self-activation activities. This research provided a better understanding of TCPs in C. lavandulifolium and laid a foundation for unraveling the mechanism by which important TCPs involved in the capitulum development.


Introduction
Chrysanthemum ×morifolium Ramat., a member of the Asteraceae family (Anthemideae), has been under cultivation in China for more than 3,000 years because of its high economic and ornamental values. These species have attractive ornamental traits with colorful and various types of flowers. C. ×morifolium exhibits a prominent ornamental trait in the form of a typical radiate capitulum, commonly referred to as a pseudanthium. This structure comprises central disc florets (DFs) and peripheral ray florets (RFs), making it one of the most notable features of the plant (Bello et al., 2013;Ibanez et al., 2017). Moreover, DFs and RFs display distinct variations concerning gender differentiation, tissue structure, organ fusion, and pigment distribution (Stevens et al., 2001;Li et al., 2010;Bello et al., 2013;Wen et al., 2019b). Investigations on transcriptomics and gene functions have provided many candidate genes potentially responsible for the differences between the two floret types (Broholm et al., 2008;Chapman et al., 2012;Tahtiharju et al., 2012;Juntheikki-Palovaara et al., 2014;Huang et al., 2016;Liu et al., 2016;Zhao et al., 2016;Bello et al., 2017;Zhang et al., 2017;Chen et al., 2018;Wen et al., 2019a;Shen et al., 2021). However, the mechanisms underlying the development of the two floret types are still uncertain.
Transcription factors (TFs) play vital roles in flower development and their temporal and spatial expression differences are usually closely related to flower morphology (Doebley, 1993;Doebley and Lukens, 1998). TCPs are responsible for a variety of plant growth processes, including shoot branching (Aguilar-Martinez et al., 2007;Martin-Trillo et al., 2011), plant height regulation (Daviere et al., 2014;Shi et al., 2016), leaf development (Palatnik et al., 2003;Crawford et al., 2004), flower development (Luo et al., 1996;Feng et al., 2006;Tahtiharju et al., 2012;Garces et al., 2016;Huang et al., 2016;Bello et al., 2017;Chen et al., 2018;Wen et al., 2019a;Shen et al., 2021), fruit (Leng et al., 2019) and seed development (Takeda et al., 2006;Nag et al., 2009), and other biological and abiotic stress response (Viola et al., 2016;Guan et al., 2017). Conducting a genome-wide analysis of TCPs in C. lavandulifolium holds significant importance in gaining deeper in si gh ts i n to t he m o le c u la r m e c h a n i s m s r e g u l a t i n g flower development.
TCP proteins are plant-specific TFs responsible for several processes of plant growth/development by modulating cell proliferation and differentiation (Cubas et al., 1999;Lopez et al., 2015). These proteins are characterized by a TCP domain, containing a conserved 59-residue-long basic helix-loop-helix (bHLH) structure at the N-terminus. This domain is known for its roles in DNA binding, protein-protein interaction, and facilitating nuclear localization of the protein (Danisman et al., 2013). TCPs were classified as two subfamilies, distinguishing them based on dissimilarities in their TCP domains: class I (known as PCF or TCP-P) and class II (or referred to as TCP-C). The most obvious distinction between the two subfamilies is that class I members lack 4 amino acids in the TCP domain (Martin-Trillo and Cubas, 2010). The class II TCPs are subdivided into 2 subclades (CYC/TB1 and CIN) in accordance with the differences between the TCP domains. The former subclade has eight members in Arabidopsis, out of which five are modulated by miR319 (Palatnik et al., 2003). Some class II members comprise an arginine-rich motif (R domain) that is responsible for the facilitation of protein-protein interaction (Cubas et al., 1999;Martin-Trillo and Cubas, 2010). Phylogenetic analysis indicates that the CYC/TB1 clade is subdivided into 3 types of genes that expanded through duplication within the group: CYC1, 2, and 3. Taking into account that two classes of TCP proteins possess distinct yet partially overlapping DNA binding sites with the consensus sequence GGNCCC (Kosugi and Ohashi, 2002), previous research has indicated that class I and II TCP proteins are responsible for synergistic and antagonistic biological interactions (Li, 2015), which usually occurs between the same class of TCP transcription factors.
Class I TCP family members enhance cell proliferation and growth, however class II members suppress these processes (Herve et al., 2009;Li et al., 2019). Studies have shown that some class I TCP TFs have been found to participate in various hormonesignaling pathways and play crucial roles in regulating inflorescence stem elongation (Daviere et al., 2014), gynoecium development (Lucero et al., 2015), seed germination (Resentini et al., 2015), filament elongation of stamen (Gastaldi et al., 2020) and flowering (Lucero et al., 2017) in Arabidopsis. Compared to class I members, the functions of the majority of class II TCPs have been extensively studied and understood. For instance, Arabidopsis CIN-TCPs are associated with the regulation of flowering time and floral organ development (Kubota et al., 2017;Li et al., 2019), and recent advances in the function and regulation of CIN-like TCPs are well described (Lan and Qin, 2020). It is well known that TB1 participates in the determination of maize axillary meristem fate (Doebley et al., 1997) and CYC controls the asymmetry of petals in Antirrhinum flower (Luo et al., 1996). In Arabidopsis, AtTCP12 and AtTCP18 are two homologs of TB1 that are responsible for inhibiting bud outgrowth (Aguilar-Martinez et al., 2007). Moreover, AtTCP18 is responsible for the suppression of the early floral transition in axillary meristems by binding with florigen proteins (Niwa et al., 2013). AtTCP1 is the homologous gene of CYC, which affects plant growth/ development by regulating the expression levels of DWARF4, a brassinosteroid (BR) biosynthesis gene (Guo et al., 2010), and also affects leaf elongation growth (Koyama et al., 2010). Additionally, CYC-like genes have been shown to regulate floral symmetry in many species, such as Primulina heterotricha (Gao et al., 2008;Yang et al., 2012), Torenia fournieri (Su et al., 2017), Saintpaulia ionantha (Hsu et al., 2018), Gerbera hybrida (Elomaa et al., 2018;Zhao et al., 2020), Chrysanthemum (Huang et al., 2016;Chen et al., 2018;Shen et al., 2021) and so on.
Recently, TCPs have been discovered in various plant species, and the number of TCPs varies among different plants. For example, Arabidopsis, Solanum lycopersicum, Oryza sativa, Populus euphratica, P. trichocarpa, Malus domestica and Prunus mume have 24, 30, 28, 33, 36, 52 and 19 TCPs, respectively (Martin-Trillo and Cubas, 2010;Parapunova et al., 2014;Xu et al., 2014;Ma et al., 2016;Zhou et al., 2016). However, the TCP gene family in C. lavandulifolium remains relatively unexplored, despite its significance as an origin species of cultivated chrysanthemum (Dai et al., 2002). This plant also serves as an excellent model material for investigating the development of RFs and DFs in Chrysanthemum. Recognizing the crucial roles of TCPs in plant growth/development, we undertook an extensive examination of the TCP gene family in C. lavandulifolium. Herein, we identified 39 TCPs from the C. lavandulifolium genome and executed a comprehensive analysis, including chromosomal position, phylogenetic relationship, conserved motifs, gene structures, and miRNA target sites. To clarify their functions in flower development, we further analyzed the expression patterns of ClTCPs during capitulum development and in two different florets, as well as the key ClTCPs were tested by qPCR in Chrysanthemum species. Additionally, the subcellular localization characteristics and the selfactivation activity in yeast of key TCP proteins were also examined. This study not only provided a basis for a comprehensive understanding of TCPs in Chrysanthemum, but also laid a foundation for revealing the mechanism of TCPs regulating the capitulum development.

Plant materials and growth conditions
The C. lavandulifolium G1 line, a diploid relative wild species for C. ×morifolium Ramat., which has achieved genomic sequence (Wen et al., 2022), was the main material for gene family analysis and screening key ClTCPs. Other Chrysanthemum samples, including C. indicum (Ci), C. vesticum (CVW), C. ×morifolium 'C27', 'D91', 'G70', 'T3', and C. aromaticum (SN) were chosen to detect the expression profiles of target genes in DFs and RFs. All plant materials were planted and managed under conventional field conditions in the chrysanthemum germplasm nursery at Beijing Forestry University (Beijing, China).

Identification of TCPs in C. lavandulifolium
Genome sequences were obtained from the C. lavandulifolium database (Wen et al., 2022). To determine the TCPs in C. lavandulifolium, we used the Arabidopsis Information Resource (TAIR) database (http://www.arabidopsis.org) to download all AtTCP protein sequences as reference sequences and screened the C. lavandulifolium genome database using local BLAST comparisons. All ClTCP protein sequences were verified by the Swissprot database (https://www.sib.swiss/swiss-prot). We removed repeated sequences using Excel and further analyzed the obtained non-redundant ClTCP sequences. Subsequently, we calculated the molecular weights (MW) and isoelectric points (pI) of ClTCP proteins using the ExPasy website (https://web.expasy.org/ protparam/) and predicted the subcellular localization of the ClTCPs through WoLF PSORT (http://www.genscript.com/ psort/wolf_psort.html).

Phylogenetic analysis, conserved motif and gene structure identification
To explore the phylogenetic relationships of TCPs between C. lavandulifolium and Arabidopsis, we used MAFFT v7 (https:// mafft.cbrc.jp/alignment/server/index.html) to perform multiple sequence alignments and subsequently used MEGA X software to develop a phylogenetic tree according to the multiple alignments data using the Maximum Likelihood (ML) approach with 1000 replicates.
Conserved motifs of TCP proteins were discovered and evaluated by MEME (https://meme-suite.org/) (parameter settings: max. number of motifs 10; max. width of motifs 50; min. width of motifs 6), and the SMART (http://smart.emblheidelberg.de/) was employed to annotate the identified motifs. The TBtools toolkit (Chen et al., 2020) was used to assess the gene structures with the corresponding DNA sequences and CDSs at default parameter settings.

Chromosomal position and miR319 target site prediction
Based on positional information from the C. lavandulifolium genome, locations of TCPs on the C. lavandulifolium chromosomes were determined using the TBtools toolkit. To predict target sites for miR319, we used the psRNATarget online application (https:// www.zhaolab.org/psRNATarget/analysis?function=3) to analyze ClTCPs nucleotide sequences.

In silico promoter analysis
The 2,000 bp upstream of the ClTCPs coding regions were extracted from C. lavandulifolium genome data, which was regarded as the promoter sequences. We used the PlantCARE tool (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) for the identification of the putative cis-acting elements of ClTCP promoters.

Expression patterns of ClTCPs in various development stages and two different florets
We obtained the expression patterns of ClTCPs in six development stages of capitula (S1, S2, S5, S6, S9, and S10) and two different florets (R1, D1, R2, D2) from the previously reported RNA-seq data (Wen et al., 2019a;Wen et al., 2022). Data normalization was conducted according to the average expression value of each gene across all studied samples. Genes and their expression patterns were systematically grouped according to the average Person's metric. Finally, the heatmap of ClTCPs was constructed with TBtools (Chen et al., 2020).

Expression analysis of ClCYC2 genes by quantitative real-time PCR
The qRT-PCR was conducted to analyze the expression patterns of key differential ClTCPs in various Chrysanthemum species. Total RNA was isolated from the DFs and RFs at two development stages (R1, D1, R2, D2) using the RNAprep-Pure-Plant-Kit (Tiangen, China) and used for cDNA synthesis with the FastKing-RT-Kit (Tiangen, China), qRT-PCR analysis was conducted on a CFX96real-time-system (Bio-Rad, USA) using the SYBR-Premix-Ex-Taqkit (Takara, Japan). The primer pairs are presented in Supplementary Table 1. We used ClSAND as a reference gene (Qi et al., 2016;Pu et al., 2020;Lu, 2022).

Subcellular localization of ClCYC2 proteins
The cDNA of the ClCYC2a, ClCYC2b, ClCYC2c, ClCYC2d, ClCYC2e, and ClCYC2f without stop codons were amplified with the primer pairs (Supplementary Table 2) and recombined into pBI121-eGFP to generate six fusion constructs. Agrobacterium tumefaciens strain GV3101 was employed to introduce the recombinants and pBI121-eGFP (used as a control) into the leaves of Nicotiana benthamiana through infiltration. After 48 hours, the expression of GFP was examined under a confocal laser scanning microscope (Leica-TCS-SP8, Wetzlar, Germany).

Analysis of transactivation activity in yeast
To assess the transactivational activity of six ClCYC2 proteins in yeast, we employed the pGBKT7 vector (BD, Clontech, United States) for the experiments. PCR amplification of the entire ClCYC2 open reading frames was performed with the gene-specific primers pairs (Supplementary Table 3). These amplicons were then independently recombined into the pGBKT7 vector. Following this, we cotransformed the pGBKT7-ClCYC2 recombinant vectors, pGBKT7-53, and pGBKT7-lam, along with pGADT7 (AD, Clontech, United States) empty vectors, into the Y2HGold strain. The transformed strains were then grown in SD/-Leu-Trp and SD/-Ade-His-Leu-Trp media, and then incubated at 30°C for 3-5 days.

Identification and chromosomal location of TCPs in C. lavandulifolium
The C. lavandulifolium genome database offers the references for the genome-wide detection of its candidate genes (Wen et al., 2022). In order to identify the TCPs in C. lavandulifolium, we performed a BLAST search against the C. lavandulifolium genome database using all the Arabidopsis TCP protein sequences as the query. A total of 39 sequences were obtained. Including one sequence isolated from the transcriptome (Wen et al., 2022), there were a total of 40 sequences obtained, which confirmed the occurrence of the TCP domain using the Swissprot database, indicating there were at least 40 TCPs in C. lavandulifolium. Among these genes, 14 TCPs with R domain were identified ( Figure 1). In addition, The TCP proteins displayed diversity in their characteristics, including variations in length, MW, and pI. The protein lengths varied between 138 and 435 amino acids, the MW ranged from 15.14 to 46.37 kDa, and the range of pI was 5.51-10.46 pH. According to WoLF PSORT predictions, the majority of ClTCP proteins showed nuclear localization, while a few were found in the cytoplasm and chloroplast (Table 1).
We located a total of 39 ClTCPs on chromosomes ( Figure 2), but they were unevenly distributed across the C. lavandulifolium genome. There were no genes identified on Chr 3. Among them, members of the CYC/TB1 subfamily were mapped to Chr 2, 4, 6, 7, and 8, respectively. Especially, four ClTCPs (ClCYC2c, ClCYC2d, ClCYC2e, and ClCYC2f) which belong to the CYC2 subclade formed a gene cluster on Chr 7, and subfunctionalization might have occurred in these genes throughout the evolutionary process of the C. lavandulifolium genome (Wen et al., 2022). While the multi-copy genes of ClCYC2a-1, ClCYC2a-2, and ClCYC2a-3 formed a gene cluster on Chr 6. Two segmentally duplicated gene pairs (ClTCP5 and ClTCP6, ClTCP20 and ClTCP21) were located on Chr 9. Multiple sequence alignments of TCP proteins in C. lavandulifolium and analysis of the potential miR319-targeted sites. (A) TCP domain alignment for the ClTCP proteins. The conserved amino acids were shaded in grey, and the bHLH regions were marked. (B) R domain alignment of the class II subfamily ClTCP members. Multiple sequence alignments were performed with Genedoc. (C) Alignment of potential target sites for miR319 (aligned in reverse). Target sites were taken from the coding sequences of ClTCPs, and the mature miR319 sequence was obtained from miRBase (http:// www.mirbase.org/).

Phylogenetic analyses and classification of ClTCPs
To investigate the phylogenetic and evolutionary relationships between the C. lavandulifolium TCPs and Arabidopsis TCPs, a phylogenetic tree was established using the complete sequences of 40 TCP proteins from C. lavandulifolium and 24 TCP proteins from Arabidopsis. According to the previous classification of TCPs in Arabidopsis, phylogenetic analysis and multiple sequence alignments all showed that ClTCP proteins were divided into two categories: class I and II, which contain 16 and 24 genes, respectively ( Figures 1A, 3). It was clear that class II proteins possess an additional four amino acids in the basic domain compared to class I members ( Figure 3A). The R domains were identified at the C-terminus of all ClTCP proteins belonging to class II CYC/TB1, as well as one member from the CIN class ( Figure 3B). The phylogenetic analysis indicated that class II was subdivided into two subclades, CIN (11) and CYC/TB1 (13). Then CYC/TB1 is further divided into three branches: CYC1, 2 and 3. In C. lavandulifolium, the numbers of CYC1, CYC2 and CYC3 protein members were 3, 9 and 1, respectively. Compared with Arabidopsis, the Class II TCPs in C. lavandulifolium were significantly expanded, especially the CYC/TB1 subclade with 13 members, and 9 TCPs belonging to the CYC2 branch. Studies on the CYC2-like genes shown that they regulate the floret morphology in G. hybrida, Helianthus annuus, Senecio vulgaris and C. ×morifolium (Tahtiharju et al., 2012;Garces et al., 2016;Huang et al., 2016). The phylogenetic tree suggested that particular genes could have experienced evolution specific to each species.

Conserved motif and gene structure identification
The conserved motif and gene structures of ClTCPs were examined to reveal the structural and evolutionary characteristics of the TCP proteins in C. lavandulifolium. By utilizing the complete sequences of the ClTCP proteins, a phylogenetic tree was constructed, effectively categorizing the ClTCP proteins into three distinct subgroups ( Figure 4A). Ten conserved motifs were identified by the MEME online tool to gain a further understanding of the variety of motif compositions among ClTCPs. The results indicated that all ClTCP proteins had highly conserved TCP domains (motif 1). Almost all of the CYC/TB1 subclade proteins and one CIN protein hit the conserved R domains (motif 3) ( Figure 4B). In contrast, most CIN proteins contained the N-terminal TCP domains of motif 7. Furthermore, motif 2 was only found in PCF, supporting previous research (Wei et al., 2016;Lei et al., 2017) that certain motifs present in a specific subgroup can play particular roles in those genes.
The intron-exon organization of nearly all ClTCPs was relatively conserved, 28 of the 39 ClTCPs had no introns, while the remaining 11 genes had one to three introns. Except for ClTCP8, ClTCP10, and ClTCP14, we discovered that no TCP-P (PCF) type genes included introns. Sequences of TCP-C genes in C. lavandulifolium had more introns compared to TCP-P genes. Within the TCP-C genes, five of them contained an intron, two contained two introns, and one contained three introns. In summary, the intron-exon distribution patterns of most ClTCPs within the same subfamily are similar, which reinforces the subclade classification and supports the understanding of their evolutionary relationships ( Figure 4C).

Analysis of promoter cis-regulatory elements in ClTCPs
The cis-regulatory elements in the promoter sequences were examined to enhance the comprehension of gene functions and regulatory mechanisms of ClTCPs. The ClTCP gene promoter regions, which are a 2-kb segment of genomic DNA located upstream of the translation start site, were submitted to the PlantCARE database. Numerous cis-acting elements associated with phytohormone responses, stress responses, and plant growth/development were identified apart from the basic CAAT and TATA boxes ( Figure 5; Supplementary Table 4). Among them, the largest number of elements were the light-responsive, MeJAresponsive and abscisic acid-responsive elements. Additionally, the elements of auxin-responsive, stress-responsive, cell cycle regulation and meristem regulation were also found in their promoter regions of ClTCPs ( Figure 5).
A kind of cis-regulatory elements associated with endosperm expression (GCN4_motif) was determined in the promoter region of 8 ClTCPs, and a kind of elements associated with meristem expression (CAT-box) was determined in 18 ClTCPs (Supplementary Table 4). The circadian control element (circadian) and cell cycle regulation element (MSA-like) were observed in 10 and 3 ClTCPs, respectively. Furthermore, the flavonoid biosynthetic genes regulation (MBSI), and seed-specific regulation element (RY-element) were also detected in the ClTCPs promoter regions ( Figure 5; Supplementary Table 4).
In hormone-associated cis-acting elements, the MeJA-responsive element (CGTCA and TGACG motifs), the ABA-responsive element (ABRE) and the gibberellin-responsive element (TATC-box, GAREmotif and P-box) were found in the promoter regions of 28, 37 and 21 ClTCPs, respectively. The salicylic acid (SARE and TCA-element) and auxin-responsive element (TGA-element, AuxRE and AuxRR-core) were discovered in 12 and 18 ClTCPs, respectively. The promoter region of ClTCPs had an abundance of hormone-responsive elements, suggesting that hormones may be crucial regulators of plant growth/ development. Moreover, cis-elements associated with stress, such as low-temperature (LTR) responsiveness, drought-inducibility (MBS) and anaerobic induction (ARE) elements were also determined in the promoter regions of 10, 18 and 34 ClTCPs, respectively (  Chromosomal positions of TCPs on C. lavandulifolium chromosomes. The scale shows a chromosomal distance of 50 Mb. All 39 ClTCPs were localized to the chromosomes. Segmentally duplicated genes (EVM0004928.1_ClTCP5 and EVM0024248.1_ClTCP6, EVM0001252.1_ClTCP20 and EVM063727.1_ClTCP21) were marked with blue letters. Eleven ClTCPs belong to CYC/TB1 subclade, which are highlighted in red and located on the chromosomes.

Expression patterns of ClTCPs during capitulum development and in two different florets
The TCPs play vital roles in multiple progress during plant development. The previously reported RNA-seq data (Wen et al., 2019a;Wen et al., 2022) were applied to investigate the expression patterns of ClTCPs in six developmental stages of capitula (S1, S2, S5, S6, S9, and S10) and two different florets (R1, D1, R2, D2), to explore their functions in flower development. Through searching and blast alignment, a total of 40 TCP sequences were obtained, and the genes whose expression level FPKM value was 0 in all samples were removed. Finally, 36 and 35 ClTCPs corresponding to the two transcriptomes were used for drawing the expression heatmap (Supplementary Tables 5, 6).
The expression levels of TCPs differed in six developmental stages of capitula, 25 out of the 36 TCP genes were specifically expressed in flowers ( Figure 6A). The expression levels of EVM0029420 (ClCYC2d), EVM0028754 (ClCYC2a), EVM0014549 (ClCYC2f), and EVM0021074 (ClCYC2e) in CYC/TB1 subclade in S5 stage were significantly lower than those in S6 stage, and they expressed steadily and continuously in the subsequent S9 and S10 stages. This suggested that CYC2-like genes in C. lavandulifolium were expressed at the early stage of the capitula, and ultimately played a role in the ontogenetic differences between DFs and RFs. The expression level of EVM0034432 (ClCYC1c) in S5 stage was significantly upregulated compared to S6 stage and decreased continuously in S9 and S10 stages. Figure 6B showed the results of hierarchical clustering to further analyze the potential involvement of TCPs in the development of DFs and RFs. According to their expression patterns, the TCPs in C. lavandulifolium are categorized into three clusters. The TCPs in C. lavandulifolium were grouped into three clusters based on their expression profiles. Cluster 1 consisted of the genes whose expression levels in RFs were remarkably lower than those in DFs. Conversely, the genes in Cluster 2 whose expression levels in RFs were markedly elevated compared to DFs, and Cluster 2 is the largest group. In Cluster 3, the expression levels of ClTCPs, such as EVM0024907 (ClTCP22), EVM0022674 (ClTCP15), and EVM0027943 (ClTCP10) showed an increase at the late stage of two florets development. In CYC/TB1 subclade, EVM0047461 (ClCYC2c), EVM0029420 (ClCYC2d), EVM0014549 (ClCYC2f), EVM0021074 (ClCYC2e), EVM0028754 (ClCYC2a) and EVM0053853 (ClCYC2a) were detected to be highly expressed in RFs at two stages, but EVM0027950 (ClCYC1a) was detected to be highly expressed in DFs at two stages.

Expression analysis of ClCYC2 genes in different Chrysanthemum species
To verify the RNA-seq data and determine ClCYC2 genes, which belong to the class II TCP subclade, and might be responsible for the development of RFs and DFs, qRT-PCR was executed to assess their expression profiles in RFs and DFs during capitulum development. Notably, six ClCYC2 genes were differentially expressed between RFs and DFs at two developmental stages. Almost all of these genes were highly expressed in the RFs of C. lavandulifolium, whereas they showed almost no or very low expression levels in the DFs, which was similar to previous research (Huang et al., 2016;Chen et al., 2018;Wen et al., 2019a). Among them, the expression difference between ClCYC2c and ClCYC2d in the two florets was the most significant ( Figure 7).
Next, other seven Chrysanthemum samples from the chrysanthemum germplasm nursery (Figure 7) were used to determine whether these CYC2-like exhibited similar expression patterns in the RFs and DFs of Chrysanthemum species. Notably, the expression patterns of CYC2 genes between two florets of different Chrysanthemum samples were almost similar, which suggested that the functions of the CYC2 gene may be conserved in the development of two florets of Chrysanthemum. In conclusion, we regarded ClCYC2c and ClCYC2d as the key genes responsible for DFs and RFs development in C. lavandulifolium.

Subcellular localization and transactivation activity of ClCYC2 proteins
Generally, transcription factors (TFs) can regulate the candidate genes by interacting with specific cis-regulatory elements located in their promoter regions, which occur in the nucleus. In this study, according to WoLF PSORT predictions, most ClTCP proteins were anticipated to be located in the nucleus (Table 1). To examine this feature, the CDSs of the ClCYC2a, ClCYC2b, ClCYC2c, ClCYC2d, ClCYC2e, and ClCYC2f without stop codons were amplified with the primer pairs (Supplemental Table 2) and recombined into pBI121-eGFP to generate six fusion constructs, then the fusion proteins and pBI121-eGFP were expressed transiently in N. b enthamiana leaves via Agrobacterium-regulated transformation. The results showed that the control vector pBI121-eGFP had a strong fluorescence signal in the whole cell, and the distribution of the fluorescence signal in the cell was not specific. The fluorescence signals of the six ClCYC2 vectors were localized in the nucleus (Figure 8), which was similar to the results in Moso Bamboo (Liu et al., 2018), Vitis vinifera (Jiu et al., 2019) and Ipomoea batatas (Ren et al., 2021). The positive controls for each test to ensure the reliability of the results. These results indicated that six ClCYC2 proteins of C. lavandulifolium were nuclear proteins, and consistent with the general characteristic of nuclear localization of TFs. To explore the transactivational activity levels of six ClCYC2 proteins, the pGBKT7-ClCYC2 recombinant vectors (gene-specific primer pairs in Supplementary Table 3), the pGBKT7-53 (positive control) and pGBKT7-lam with pGADT7 empty vectors (negative control) were co-transformed into theY2HGold strain, respectively. All of these transformants grew well and exhibited obvious white colonies on the SD/-Leu-Trp medium (Figure 9). Only the yeast cells carrying ClCYC2c, ClCYC2e, ClCYC2f, and the positive group could grow well on the SD/-Ade-His-Leu-Trp medium. Whereas, the yeast cells harboring ClCYC2a, ClCYC2b, ClCYC2d, and the negative group did not proliferate (Figure 9). Hence, three out of six candidate Class II ClTCP constructs activated the levels of the His3 and Ade2 reporter genes, demonstrating their transactivational activities in the yeast strain. To screen the interacting proteins in the subsequent yeast two-hybridization system, the bait expression vectors with self-activating activities were inhibited by adding 3-AT. The results showed that the bait vectors of ClCYC2c, ClCYC2e, and ClCYC2f could inhibit their self-activating activities by adding 80 mM, 20 mM, and 10 mM 3-AT, respectively.

Discussion
TCP protein is a TF specific to plants, which is involved in different aspects of biological and physiological processes throughout plant growth/development (Manassero et al., 2013). TCPs have yet been discovered and characterized in various plants, including Arabidopsis (Riechmann et al., 2000;Li, 2015), M. domestica (Xu et al., 2014), Legume (Ling et al., 2020), Vitis vinifera (Jiu et al., 2019), P. mume (Zhou et al., 2016), Moso Bamboo (Liu et al., 2018) and C. nankingense (Tian et al., 2022). Nonetheless, no comprehensive analysis of TCPs in C. lavandulifolium, a prominent ancestral species of cultivated chrysanthemum has been undertaken. In the study, 40 ClTCPs were determined from C. lavandulifolium genome  Expression patterns of ClTCPs during capitulum development and in two different florets of C. lavandulifolium. Red and blue boxes show high and low expression levels, respectively. Green, black and red color of gene names belong to PCF, CIN, and CYC/TB1, respectively. S1, S2, S5, S6, S9, and S10 indicated the different development stages of capitula (Wen et al., 2022), L indicated reproductive-stage leaf; D1, D2, R1, R2 indicated disc and ray florets at two stages. Wu et al. 10.3389/fpls.2023.1276123 and transcriptome. Moreover, we performed a comprehensive analysis of the ClTCPs by assessing their phylogenetic relationships, chromosomal position, conserved motifs, gene structures, promoter analysis, expression patterns in different developmental stages and different flower types, and subcellular localization and transactivation activity of key ClTCP proteins. These findings offer a basis for further investigations on the functions of ClTCPs during flowering.

Evolutionary conservation and divergence of TCPs in C. lavandulifolium
In C. lavandulifolium, 40 TCPs were discovered in the genome and transcriptome, and their distributions on the chromosomes were not uniform. The number of ClTCPs in C. lavandulifolium is about double that of C. nankingense, V. vinifera, and Arabidopsis, which possess 23, 17, and 24 TCP members, respectively (Li, 2015;Jiu et al., 2019;Tian et al., 2022). The collinearity analysis of the V. vinifera and C. lavandulifolium genomes revealed a 1 vs. 3 syntenic association between them, providing compelling evidence for whole-genome triplication (WGT-1) in C. lavandulifolium (Wen et al., 2022). The WGT event was shared by Asteraceae species and may be responsible for the emergence of complex traits in the capitulum. Earlier research has established that CYC2 clade genes play a crucial role in regulating DFs and RFs development in Asteraceae species, including G. hybrida, H. annuus, S. vulgaris, and C. ×morifolium (Tahtiharju et al., 2012;Garces et al., 2016;Huang et al., 2016;Chen et al., 2018). We identified nine CYC2-like genes in C. lavandulifolium (Figure 3). It possesses a larger number of TCPs, amounting to nine members, which surpasses the 7 and 8 members identified in C. nankingense and H. annuus, respectively, but fewer than 25 members found in C. ×morifolium (Song et al., 2023). Besides, tandem duplication events occurred in the evolutionary process in C. lavandulifolium CYC2 genes, which led to CYC2a multicopy genes forming a gene cluster on chromosome 6. The duplication event likely facilitated the proliferation of CYC2c/2d/2e/2f, and subsequent evolutionary processes could have led to subfunctionalization of these genes within the C. lavandulifolium genome (Figure 2). The notable difference in the number of TCPs between C. lavandulifolium and C. nankingense might be due to the divergency time, that C. lavandulifolium and C. nankingense diverged from each other around 7.2 Mya (Wen et al., 2022). The previous result shows the class II TCPs in C. lavandulifolium were significantly expanded (Figure 3). Subcellular localization of six ClCYC2 proteins. Six ClCYC2-GFP fusion recombinants (ClCYC2a-GFP, ClCYC2b-GFP, ClCYC2c-GFP, ClCYC2d-GFP, ClCYC2e-GFP, and ClCYC2f-GFP) and pBI121-GFP as control were transiently expressed in N. benthamiana leaves and observed by a laser scanning confocal microscope 48 hours after injection. Bar=25 µm.
According to multiple sequence alignments and phylogenetic analyses, all 40 ClTCPs were grouped into 3 subgroups (Figures 1 and  3), which was in accordance with those in Arabidopsis (Li, 2015). Generally, TCPs within a clade have similar gene structures, chromosomal locations and motif distributions, suggesting that TCPs are evolutionarily conserved. For example, we found the highly-conserved TCP domains (motif 1) occurred in all ClTCP proteins (Figure 4), while motif 2 was only present in the class I PCF clade and motif 7 was detected in most of class II CIN proteins ( Figure 4B). The close evolutionary relationships are further supported by the fact that the motif compositions and gene structures of ClTCPs exhibit a high degree of consistency. Additionally, our analysis revealed the presence of an arginine-rich R domain at the C-terminus of all ClTCP proteins classified under class II CYC/TB1, and one member (EVM0024660) from CIN. R domain was estimated to increase protein-protein interactions, as do four members (AtTCP2,12,18,and 24) in Arabidopsis.
Subcellular localization indicated that six ClCYC2 proteins of C. lavandulifolium were nuclear proteins, but these six CYC2 members have different transcriptional activation activities, which may be affected by several factors, including the binding ability of the TCP transcription factors to the target genes, differences in gene structures and sequences, differences in expression regulatory mechanisms, and the involvement of other cofactors (Martin-Trillo et al., 2011;Tahtiharju et al., 2012;Valsecchi et al., 2013;Parapunova et al., 2014). Further exploration and experimental verification will help to fully understand the regulatory mechanisms of these key ClCYC2 genes.

The potential roles of ClTCPs in plant growth and development
The function of a gene is closely correlated with its expression patterns (Xu et al., 2015). Herein, we determined the expression patterns of all ClTCPs during flower development by RNA-seq data reported previously (Wen et al., 2019a;Wen et al., 2022) and detailed expression patterns of key class II ClTCPs (ClCYC2) in different Chrysanthemum samples via real-time RT-PCR. The result indicated that these genes exhibited unique expression patterns during capitulum development and in different flower types. In detail, most ClTCPs were specifically expressed in flowers during capitulum development, and TCPs were categorized into three clusters in accordance with their expression patterns in DFs and RFs. The expression levels of CYC2 genes in RFs were upregulated compared to DFs, but the expression profiles of CYC1 and CYC3 group genes were different from those of CYC2 genes in C. lavandulifolium, which presumed that they had different functions during flower development. These suggested that TCPs in C. lavandulifolium were associated with the growth of capitula or two florets.
In earlier research (Sarvepalli and Nath, 2018), the CIN subclade TCPs were found to primarily regulate leaf development, floral organ development, and flowering time. In Arabidopsis, 5 CIN-like genes have been reported to undergo post-transcriptional regulation by microRNA319 (miR-319). Notably, the regulation of TCP4 by miR-319a is crucial for the development of stamens and petals (Nag et al., 2009). In C. lavandulifolium, four homologs of CIN subclade genes (ClTCP24, ClTCP25, ClTCP26 and ClTCP27) included the potential binding sites for miR319 ( Figure 1C). ClTCP24, ClTCP25 and ClTCP26 are upregulated in leaves, suggesting their potential roles in leaf development. The conservation of miR319 targeting may extend to other TCPs as well (Palatnik et al., 2003;Seki et al., 2020).
Moreover, promoter analysis shows that the promoter region of ClTCPs was found to contain a substantial number of hormone-responsive elements ( Figure 5), implying these genes might exert significant influences on hormone-responsive plant developmental processes. Previous reports have shown that TFs Transactivational analysis of ClCYC2 proteins in Y2HGold strain. The fusion constructs with pGADT7 AD empty, positive constructs (pGBKT7-53) and negative constructs (pGBKT7-lam) were co-transformed into Y2HGold strain and incubated on the SD/-Leu-Trp and SD/-Ade-His-Leu-Trp media. and phytohormones play important roles in determining flower size in Arabidopsis (Krizek and Anderson, 2013). Cytokinin has been demonstrated to influence the duration of cell division in developing floral organs (Bartrina et al., 2011). Considering the abundant hormone-responsive elements distributed in ClTCP gene promoters (Figure 5), we speculate that ClTCPs may serve as important regulators of hormone-induced alterations in plant development. Of course, this hypothesis needs to be further tested.

Contribution of CYC2 genes to the morphological differences of two florets in Chrysanthemum
Flower development in the angiosperm is a complex physiological and morphological process, which was determined by a set of TFs regulatory networks (Soltis et al., 2002). Much evidence has already suggested that TCPs, especially the CYC2 clade genes affect the morphology of capitula by regulating the reproductive organ development, inflorescence architecture and floral symmetry (Juntheikki-Palovaara et al., Huang et al., 2016;Chen et al., 2018;Elomaa et al., 2018).
Our investigation revealed the presence of 13 genes in C. lavandulifolium that belong to the CYC/TB1 subclade, CYC2 group has nine members with the largest number. Among them, six TCPs from CYC2 group, EVM0047461 (ClCYC2c), E V M 0 0 2 9 4 2 0 ( C l C Y C 2 d ) , E V M 0 0 1 4 5 4 9 ( C l C Y C 2 f ) , EVM0021074 (ClCYC2e), EVM0028754 (ClCYC2a) and KX161380.1 (ClCYC2b) were all highly expressed in the RFs, whereas they shown almost no or very low expression levels in DFs (Figure 7). And the difference in expression was universal in different Chrysanthemum species, which suggested the functions of the CYC2 gene may be conserved in the two florets development of Chrysanthemum, whereas certain members of the CYC/TB1 subclade experienced functional divergence subsequent to gene duplication events. Based on previous studies, it was hypothesized that CYC2c and CYC2d are mainly involved in RF identity and thus affect the development of capitulum, CYC2b and CYC2e tend to regulate petal length in RFs (Shen et al., 2021), and CYC2c can interact with the CYC2f promoter to modulate floral symmetry development (Yuan et al., 2020). Moreover, six CYC2 genes were expressed in C. lavandulifolium at the early capitulum developmental stages, which was consistent with the results of nucleic acid in situ hybridization of Gerbera and S. vulgaris that detected CYC2-like gene expression in the meristem and floret primordium at the early capitula development (Garces et al., 2016;Elomaa et al., 2018). This implied that CYC2 genes in C. lavandulifolium were expressed at the early stages of the capitula, and ultimately played roles in the ontogenetic differences between ray and disc florets.

Conclusion
In summary, the first genome-wide analyses of TCPs in C. lavandulifolium were conducted. 40 ClTCPs were discovered and distributed on 8 chromosomes. These ClTCPs were classified into two main groups in accordance with the phylogenetic analysis and structural characteristics. The ClTCP promoter sequences had many different kinds of cis-acting elements, suggesting that ClTCPs were controlled by a complex regulatory network. ClTCPs, especially those from class II CYC/TB1 clade, may be crucial for the two florets development in Chrysanthemum as indicated by their expression patterns and previous studies. Notably, six candidate ClCYC2 genes were higher expressed in RFs than those in DFs, which speculated that the CYC2 genes may be functionally conserved in two florets development in Chrysanthemum. The subcellular localization and transactivation activity analysis of six candidate TCP proteins were also examined in C. lavandulifolium. In short, our study increased the understanding of ClTCP functions during flower development in Chrysanthemum species. Based on this study, in the future we can validate the functions of key TCPs in native plants, and deeply study the interactions between the key ClTCPs and their upstream and downstream regulatory mechanisms, to resolve the molecular network of TCPs involved in the capitulum development of Chrysanthemum.

Data availability statement
The original contributions presented in the study are included in the article/Supplementary Material. Further inquiries can be directed to the corresponding author.

Funding
The authors declare financial support was received for the research, authorship, and/or publication of this article.