Genome-Wide Identification, Evolution, and Expression Analysis of the TCP Gene Family in Rose (Rosa chinensis Jacq.)

Abstract

Roses have not only high ornamental and economic values but also cultural importance worldwide. As a plant-specific transcription factor gene family, the TCP (TEOSINTE BRANCHED 1, CYCLOIDEA, PROLIFERATING CELL FACTOR LAND 2) genes have been indicated to be involved in various aspects of plant biological processes, such as leaf morphogenesis and senescence, lateral branching, flower development, stress response and hormone signaling. Currently, TCP genes have been identified and analyzed in many plants, yet there is no systematic analysis in Rosa chinensis. Here, we identified 16 RcTCP genes from R. chinensis genome, which were unevenly distributed in five out of all seven chromosomes. Phylogenetic and structural analyses showed that RcTCP family could be classified into two classes, I (namely PCF) and II, and class II genes can be further divided into CIN and CyC/TB1 subclasses. The different classes of TCP genes were showed to have undergone different evolutionary processes, and genes in the same branch shared similar motifs, gene structures and conserved structural domains. Promoter analysis showed that RcTCPs had many cis-acting elements that are mainly associated with plant growth and development, plant hormones and abiotic/biotic stress responses. Furthermore, the expression levels of RcTCPs under vegetative and reproductive growth and drought stress treatments were analyzed based on public RNA-seq dataset, and it was shown that RcTCPs exhibited serious tissue-specific expression, with most of them dominantly expressed in flowers, leaves and stems, with high levels of expression at different stages of flower and bud differentiation, particularly during petal formation and gametophyte development. The high inducement of seven RcTCP genes from PCF class in drought stress indicated their important roles in biological processes against drought stress. Our results provide valuable information for the evolution and functional characterization of TCP genes in R. chinensis.

1. Introduction

TCP transcription factors belong to a plant-specific gene family, which play significant roles in the regulation of plant growth and development and stress response [1]. In 1999, the TCP family was first defined as a new transcription factor family with few members [2]. This family was identified from four unrelated genes encoding proteins and named by their initials: TB1 (TEOSINTE BRANCHED 1), CyC (CYCLOIDEA), PCF1 (PROLIFERATING CELL FACTORS 1) and PCF2 (PROLIFERATING CELL FACTORS 2) [3]. An atypical basic helix-loop-helix (bHLH) structure forms at the N-terminus of the TCP domain, which is highly conserved [2]. TCP proteins are divided into two subfamilies, class I and class II. Class I, the TCP-P subfamily, is also called PCF subfamily. Class II, called TCP-C subfamily, includes CYC/TB1, and CIN subclades. There is a difference between the two classes in that the TCP domain in class I has four amino acids deletion compare to that in class II [4]. Additionally, the members of CYC/TB1 specifically contain a hydrophilic α-helix (R domain) structure which is rich in polar amino acids, while it does not exist in other classes [5]. The CIN is widespread in plants, but CyC/TB1 only occurs in angiosperms. TCPs have been characterized to be involved in a variety of growth-related processes, including embryonic growth, leaf development, branching, floral organ morphogenesis, pollen development, germination, senescence, circadian rhythms, cell cycle regulation, and hormone signaling [6]. However, the majority of the functional features were only from model species, such as Arabidopsis and rice. Currently, even the genome-wide analysis of TCP gene family have been performed for many species, such as in Arabidopsis thaliana (24 TCPs) [3], soybean (Glycine max) (54) [7], cotton (Gossypium hirsutum) (74) [8], tomato (Solanum lycopersicum) (30) [9], grapevine (Vitis vinifera) (15) [10], plum (Prunus mume) (19) [11], but not in roses yet.

Roses are the most well-known and beloved ornamental plant genus worldwide, with not only high ornamental and economic value, but also cultural importance. Chinese old rose contributed a continuous flowering trait, double petals and its signature tea scent to modern roses [12]. Along with the genome release of R. chinensis cv. ‘Old Blush’, one of the important progenitors of modern roses, roses have been proposed to be an ideal model species for studying the molecular basis of woody plants [13]. Currently, although different studies have been conducted on the transcriptome, gene family functions, and genetic diversity of various genes and traits in roses [14], there is no comprehensive characterization of the TCP gene family in R. chinensis. Here, we performed genome-wide identification, evolution and functional characterizations of TCP gene family in R. chinensis. Our results provided insights into the evolution and functional profiles of TCP genes, and will be of benefit for the functional characterizations of individual TCP genes in roses.

2. Materials and Methods

2.1. Identification of RcTCPs Genes in R. chinensis

The sequences of R. chinensis genes were obtained from GDR (Genome Database for Rosaceae, https://www.rosaceae.org (accessed on 15 April 2022)). The sequences of A. thaliana TCP genes were downloaded from phytozome (https://phytozome-next.jgi.doe.gov/ (accessed on 15 April 2022)). Subsequently, the possible TCP proteins were retrieved from R. chinensis by using the BLAST Wrapper function in TBtools (https://github.com/CJ-Chen/TBtools (accessed on 16 April 2022)) with the sequences of AtTCP protein sequences as queries, and followed by verification of TCP domain existence and integrity using HMMER searching (http://www.ebi.ac.uk/Tools/hmmer/ (accessed on 16 April 2022)), SMART (https://smart.embl-heidelberg.de/ (accessed on 16 April 2022)), Pfam database (PF03634) (https://pfam.xfam.org/ (accessed on 17 April 2022)), and CDD/NCBI (https://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi (accessed on 17 April 2022)), and removing redundant and short sequences manually. Finally, all RcTCP genes were obtained for further analysis.

2.2. Chromosome Distribution and Synteny Analysis of RcTCP Genes

The RcTCPs were mapped to the chromosome based on the R. chinensis genome information obtained from the gff files and visualized by MG2C (http://mg2c.iask.in/mg2c_v2.0/ (accessed on 20 April 2022)). MCScanx (https://github.com/vidsvur/MCScanX-tutorial (accessed on 21 April 2022)) was used to analyze the gene collinearity and tandem duplication events between RcTCP genes. Tbtools was used to visualize the homology relationships of RcTCPs.

2.3. Physicochemical Properties and Subcellular Localization of RcTCP Genes

Protparam (http://expasy.org/tools/protparam.html (accessed on 25 April 2022)) was employed to predict molecular mass, isoelectric point, amino acid, average hydrophilic coefficient aliphatic index and instability index of RcTCP proteins. Cell-PLoc-2 (http://www.csbio.sjtu.edu.cn/bioinf/Cell-PLoc-2/ (accessed on 27 April 2022)) was used to predict the subcellular localization of RcTCP proteins.

2.4. Phylogenetic Analysis

The genome sequences of Prunus mume, Malus domestica, and Fragaria vesca were obtained from GDR. Phylogenetic trees were constructed using all protein sequences of TCP domain using MEGA11.0 (https://www.megasoftware.net/ (accessed on 8 May 2022)). All TCP proteins were aligned using ClustalW (https://www.genome.jp/tools-bin/clustalw (accessed on 8 May 2022)). The phylogenetic tree was constructed using the Maximum-likelihood (ML) method with 1000 bootstrap replicates.

2.5. Motif, Domain and Gene Structure Analysis

The conserved motifs of the TCP proteins were identified using MEME suite (http://meme-suit.org/tools/meme (accessed on 10 May 2022)). The exon-intron organizations were identified by comparing the coding sequences with their corresponding genomic sequences using the TBtools v1.0987663 software (https://github.com/CJ-Chen/TBtools (accessed on 11 May 2022)).

2.6. RcTCP Cis-Regulatory Element (CRE) Analysis

The promoters (2000 bp upstream sequences from ATG) of RcTCP genes were used to analyze the cis-acting elements using PlantCARE (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/ (accessed on 12 May 2022)). The results of the analysis were compiled by Adobe Illustrator 2021 software (https://www.adobe.com/products/illustrator/free-trial-download.html (accessed on 12 May 2022)), and a heatmap of cis-acting element copies in promoter region of RcTCP genes was visualized by TBtools.

2.7. Collinearity Analysis

The syntenic relationships among the TCP genes of R.chinenisis and other species was investigated using MCScanX (https://github.com/vidsvur/MCScanX-tutorial (accessed on 18 May 2022)) with the default parameters. Oryza sativa genome sequences were obtained from Ensembl Genomes (https://plants.ensembl.org/Oryza_sativa/Info/Index (accessed on 16 May 2022)). Malus domestica genome sequences were obtained from GDR (https://www.rosaceae.org (accessed on 16 May 2022)).

2.8. Expression Patterns of RcTCP Genes

The expression levels of the RcTCPs were analyzed using transcriptome datasets that retrieved from ROSAseq web interface database [15] (http://iant.toulouse.inra.fr/R.chinensis (accessed on 17 May 2022)) and NCBI, including datasets of drought treatment (PRJNA486271, PRJNA663119) [16,17], flower organ development (PRJNA351281, PRJNA398090, PRJNA325324, PRJCA000258) [18,19,20,21]. The expression levels were illustrated based on the log2 transformed FPKM values using Kallisto (https://kallisto/download.html (accessed on 22 May 2022)) and visualized by TBtools (https://github.com/CJ-Chen/TBtools (accessed on 23 May 2022)).

2.9. Plant Material and PEG Treatments

The plants of R. chinensis ‘Old Blush’ were grown in the rose resource nursery of Nanjing Agricultural University. The stems with at least one node were cut and used as explants, and cultured on Murashige and Skoog (MS) medium (Duchefa Biochemie, Haarlem, The Netherlands) supplemented with 1.0 mg/L 6-benzyl aminopurine (6-BA), and 0.05 mg/L 1-naphthaleneacetic acid (NAA) for 30 d at 22 ± 1 °C, under a 16 h light/8 h dark photoperiod. The shoots were then transferred to half strength MS medium supplemented with 0.1 mg/L NAA for 25 d for rooting. Then, the plants were transferred to Hoagland’s nutrient solution and were grown in controlled conditions (25 °C, 40% relative humidity, and 200 μmol·m⁻²·s⁻¹) under LDs (16:8 h, light:dark) for 14 days. To simulate drought stress, the R. chinensis seedlings were transferred to Hoagland’s nutrient solution containing 20% w/v PEG6000 and cultured for 24 h, taking purity Hoagland’s nutrient solution as control. The treatments were repeated three times, and at least nine seedlings were used for each treatment. The samples were collected after treatment and immediately frozen in liquid nitrogen and stored at −80 °C until RNA isolation.

2.10. RNA Isolation and Real-Time Quantitative PCR Analysis

Total RNA of the samples were extracted using the BioTeke Quick RNA isolation Kit (Cat. #: RP3301, BioTeke Corporation, Beijing, China) and 1 μg of high-quality total RNA was reverse transcribed using the PrimeScriptTM RT reagent Kit (Cat. #: RR047A, TaKaRa, Dalian, China) according to the manufacturer’s instructions. Gene-specific primers for RcTCP genes were designed using GenScript (https://www.genscript.com/ (accessed on 3 June 2022)) and the IDT (http://sg.idtdna.com/scitools/Applications/RealTimePCR/ (accessed on 3 June 2022)) online server. RT-qPCR assays were carried out by the QuantStudio 6 Real Time PCR System (Thermo Fisher Scientifific, Carlsbad, CA, USA) using SYBR^® Premix Ex TaqTM (Tli RNaseH Plus) (catalog number: RR420A, TaKaRa) according to the manufacturer’s instructions. Expression levels of RcTCP genes were normalized by using RcGAPDH gene as reference, and the 2^−ΔΔCT method was used to calculate the relative expression of RcTCP genes [22]. Three biological replicates and three technical replicates were used for each experiment. The primers used in RT-qPCR are listed in Supplemental Table S1.

3. Results

3.1. Identification and Characteristics of RcTCP Genes

To identify TCP genes in R. chinensis, we firstly retrieved all candidate genes using the BLAST Wrapper function in TBtools; 17 candidate genes were obtained. Then, HMMER searching, SMART, Pfam database (PF03634), and CDD/NCBI were further employed to verify the existence and integrity of TCP domains within the candidate genes, the genes with integrity of TCP domain less than 50% and short sequences were removed manually. Finally, all 16 RcTCP genes were obtained and named according to chromosome positions for further analysis (

Table 1. Physicochemical properties and subcellular localization of RcTCP proteins.

Protein Name	Gene ID	Molecular Mass (Da)	Isoelectric Point	Amino Acid (aa)	Average Hydrophilic Coefficient	Aliphatic Index	Instability Index	Subcellular Localization Prediction
RcTCP01	RC1G0213700	39,668.36	8.40	360	−0.510	70.19	50.85	Nucleus.
RcTCP02	RC1G0312100	32,831.23	8.57	307	−0.391	66.81	60.89	Nucleus.
RcTCP03	RC1G0531500	34,056.20	6.60	328	−0.226	76.13	47.93	Nucleus.
RcTCP04	RC2G0577900	39,859.44	9.17	360	−0.632	64.78	53.90	Nucleus.
RcTCP05	RC2G0599700	39,158.12	6.62	355	−0.800	63.46	55.30	Nucleus.
RcTCP06	RC2G0685400	32,176.69	9.27	304	−0.648	65.89	57.49	Nucleus.
RcTCP07	RC4G0096800	27,591.84	8.82	262	−0.445	63.82	41.51	Nucleus.
RcTCP08	RC4G0266300	32,732.35	9.08	290	−0.064	80.41	44.72	Nucleus.
RcTCP09	RC4G0400400	49,609.69	9.45	444	−0.837	58.90	48.24	Nucleus.
RcTCP10	RC4G0435200	44,618.17	6.57	423	−0.598	60.28	56.15	Nucleus.
RcTCP11	RC5G0134300	47,923.63	8.42	441	−0.826	57.82	49.47	Nucleus.
RcTCP12	RC5G0279600	48,250.47	6.68	438	−0.807	57.56	54.65	Nucleus.
RcTCP13	RC7G0042700	45,529.80	6.81	426	−0.734	54.39	64.76	Nucleus.
RcTCP14	RC7G0061200	42,654.47	8.93	384	−0.685	65.76	51.74	Nucleus.
RcTCP15	RC7G0073800	52,440.83	7.04	466	−0.869	59.27	52.95	Nucleus.
RcTCP16	RC7G0209000	40,894.44	6.40	393	−0.474	60.38	62.10	Nucleus.

). Furthermore, the amino acid number, molecular weight (MW), isoelectric point (PI), average hydrophilicity coefficient, lipid coefficient, instability coefficient and subcellular localization of the 16 RcTCP proteins were analyzed (Table 1), and it showed that a total of 262 (RcTCP07)-466 (RcTCP15) amino acids make up the RcTCP protein, with an average of 373.8 amino acids. The molecular weight ranged from 27.59 kDa (RcTCP07) to 52.44 kDa (RcTCP15). The isoelectric point (PI) ranged from 6.4 (RcTCP16) to 9.45 (RcTCP09). The average hydrophilic coefficients of all proteins were less than zero, indicating that RcTCP proteins are hydrophilic. The aliphatic index ranged from 54.39 to 80.41. Each RcTCP protein has an instability index greater than 40, indicating its instability. Subcellular localization predictions showed that all RcTCP proteins were located in the nucleus.

3.2. Chromosome Distribution and Synteny Analysis of RcTCP Genes

In order to gain more insight into the evolution of RcTCP genes, the distributions of RcTCP genes on chromosomes were surveyed. The results showed that RcTCP genes are unevenly distributed on R. chinensis chromosomes with only five out of seven chromosomes containing TCP genes (Figure 1A), and no putative TCP genes were found on chromosomes three and six, indicating their biased evolution profiles during whole genome evolution. Furthermore, gene duplication events were detected by using MCScanX 3-28-2013 software (Figure 1B). Six pairs of RcTCP genes were found to be evolved from segmental duplications, while no tandem duplication events were found. These results suggested that the evolution of the TCP gene family was highly associated with whole-genome duplication events during evolution in R. chinensis.

3.3. Phylogenetic Analysis of RcTCP Proteins

In order to explore the evolutionary and phylogenetic relationship between the TCP proteins of R. Chinensis and other known TCP proteins, a phylogenetic tree was constructed for 130 TCP protein sequences from R. chinensis, A. thaliana, P. mume, M. domestica and F. vesca using MEGA11.0 (Supplemental Figure S1). The phylogenetic tree showed that the TCP genes from all species can be divided into two major classes, class I, TCP-P/PCF, and class II, TCP-C, and in which, Class II was further divided into two subclasses, namely CyC/TB1 and CIN, according to previous study [6]. The phylogenetic tree of TCP genes from R. chinensis and A. thaliana were represented for classification of R. chinensis (Figure 2A). Interestingly, for each subclass, the copy number of TCP genes varied significantly (Figure 2B). The copy number of TCP genes in PCF and CIN subclass from apple (which suffered recent whole genome duplication event, WGD) is at least 2.2- and 4.3-fold higher than that of other Rosaceae species. However, the copy number of TCP genes in CyC/Tb1 is relatively well-conserved. This result further indicated that the evolution of TCP genes from PCF and CIN subclass is highly associated with whole-genome duplication events during evolution in Rosaceae, while it is not in CyC/Tb1. Furthermore, the relatively lower copy number of TCP genes from PCF and CIN subclass in R. chinensis suggests there might be gene loss during evolution.

3.4. Conservative Motifs, Structural Domains, Gene Structure Analysis

An analysis of the conserved motifs, domains, and gene structures of R. chinensis and A. thaliana was conducted to determine the structural differences between the TCP proteins of R. chinensis along with the phylogenetic tree (Figure 3A). Six conserved motifs were identified from R. chinensis and A. thaliana TCP proteins by using MEME online service. Among them, all TCP proteins share the conserved motif 1 which corresponds to the TCP domain. While motif 2 were found to be specific to PCF subclass, and motif 3 is specific to CIN and CyC/TB1 subclass, and motif 4 is very conserved within the CyC/TB1 subclass (Figure 3B). A high degree of similarity between protein motifs in a subfamily indicates that the subfamily division is consistent with the distribution of conserved motifs. A subfamily of PCF proteins is only represented by five proteins exhibiting motif 6, which indicates that some motifs play an important role in subfamily-specific function. As R. chinensis and A. thaliana have conserved motifs with similar compositions and positions, they may have similar structural characteristics and a similar function as well. Based on domain predictions, we defined these proteins as TCP proteins since they shared a common TCP protein domain (Figure 3C). The domain of TCP protein in CIN branch of class II is closer to that of TCP protein of class I, which means that CIN branch may be older than CyC/TB1 branch. Furthermore, the domain regions and some motifs could be approximately overlapped when the conserved motif and domain maps were combined, indicating that the conserved motifs of TCP proteins corresponded to the domains, and that TCP proteins had additional motifs in addition to the reported bHLH and R-domains.

Genetic structure analysis showed that almost all TCP genes in Arabidopsis and R. chinensis exhibit a highly conserved exon-intron organization, but with differences (Figure 3D). Interestingly, 19 TCP genes were found to be intronless, and most of the RcTCP genes show similar exon intron arrangement and number in the same subfamily, which supports the classification and evolutionary relationship of subclasses, indicating that the biological functions of these genes might be highly conserved.

3.5. Promoter Cis-Regulatory Elements Analysis of RcTCP Genes

To further understand the function and regulatory mechanism of the RcTCP genes, the cis-acting elements within promoter region (2000 bp upstream sequences from ATG) of 16 RcTCP genes were analyzed (Figure 4, Supplemental Table S2). Out of the basic cis-acting elements, a large number of cis-acting elements in the promoter of RcTCP genes were found to be involved in light response, phytohormone response, plant growth and development and stress response. As shown, light-responsive elements were observed in all RcTCP genes, counting for a total of 182 times. The Auxin response element, the salicylic acid response element, the abscisic acid response element, the methyl jasmonate (MeJA) response element, and the gibberellin response element were found in the promoter regions of 6, 9, 13, 13 and 8 RcTCP genes, respectively, indicating that plant hormones play important roles in plant growth and development. Cis-acting regulatory elements involved in seed-specific regulation and endosperm expression were identified in the promoter regions of RcTCP01, RcTCP03 and RcTCP01, RcTCP16, respectively, indicating that some genes of the RcTCP family are involved in the regulation of seed growth and development. Furthermore, eight and three RcTCP genes contain cis-regulatory elements for plant meristematic tissue expression and circadian rhythm regulation, respectively.

Furthermore, several cis-acting elements associated with abiotic and biotic stresses have been identified in the promoter region of the RcTCP genes. It was found that 7, 6 and 10 genes contain cis-acting elements involved in defense and stress responses, low-temperature responsiveness and MYB binding site involved in drought-inducibility. In addition, 13 and 1 genes were observed for anaerobically induced and related stress elements involved in hypoxia-specific induction of enhancer-like elements. As a result, it appears that the RcTCP genes may be involved in the transmission of defense signals.

3.6. Collinearity Analysis of RcTCPs

During evolution, gene families may suffer gene gain or loss within different species after genome shuffling. In order to gain more insights into the evolutionary profiles of TCP genes, collinearity analysis was performed on TCP genes from R. chinensis, rice and apple. As shown in Figure 5, six pairs of collinearity events were found between R. chinensis and O. sativa TCP genes, accounting for 37.5% of the total RcTCPs, and 14 pairs of collinearity events were found between R. chinensis and M. domestica TCP genes, accounting for 87.5%. Accordingly, R. chinensis and M. domestica appear to be more evolutionarily conservative, which is in accordance with their phylogenetic relationships. Furthermore, by comparing with R. chinensis TCP genes, most of the TCP genes in M. domestica were duplicated, indicating their retention after whole-genome duplication events.

3.7. Expression Patterns of RcTCP Genes

To better understand the functional profiles of RcTCP genes in R. chinensis, transcript levels of RcTCP genes in different tissues were analyzed based on various public RNA-seq datasets (Supplemental Table S3). As shown in Figure 6, most of the RcTCP genes from the PCF subclass, and three out of five RcTCP genes from the CIN subclass were highly expressed or induced in majority of the tissues and treatments. In contrast, the expression levels of RcTCP genes from CyC/TB1 subclass were generally low, except RcTCP5 which was specifically highly expressed in dominant and active axillary buds. These results suggested the functional diversification of RcTCP between different subclasses. Interestingly, although RcTCP genes were barely expressed under drought treatments, three genes (RcTCP05, RcTCP11 and RcTCP12) from the CIN subclass and six genes (RcTCP06, RcTCP10, RcTCP13, RcTCP07, RcTCP02 and RcTCP16) from the PCF subclass were highly expressed under the condition of phosphorylation of RhPIP2; 1 (Figure 6B), which controls the transmembrane transport of water and plays a crucial role in petal expansion and stress response [23,24]. These results suggested that there are functional differences of RcTCP genes in the response to drought stress, and the CIN and PCF subclasses play an indirect role in the regulation of drought stress rather than being directly involved.

Furthermore, to investigate the expression pattern of TCP genes in the growth and development of rose organs, transcriptome data from the database were used to analyze the expression profile of TCP genes during the differentiation to full opening of R. chinensis floral organs (Figure 6D). Similarly to the previous findings, the genes with higher expression in rose organs were mainly found in the PCF and CIN, while the CyC/TB1 genes were barely expressed. RcTCP16 was expressed at very high levels throughout the growth and development of the floral organ. During the period from meristematic development to full opening of flower organs, RcTCP10, RcTCP13, RcTCP07 and RcTCP02 of PCF subclasses have relatively high expression levels, but their expression levels decrease with the development of flower organs, suggesting their involvement in the early development of flower organs. The expression of RcTCP05 in the phloem tissue of a flower was barely detectable, but it was relatively high from the beginning of flower opening to full open discoloration. Both RcTCP11 and RcTCP12 were highly expressed in the tested tissues except for partially open flowers and fully open-pink flowers. In general, different subgroups of RcTCPs play a conservative and diverse role in the process of rose floral organ differentiation to their full opening.

3.8. Expression Pattern of RcTCP Genes under PEG Treatments

As described above, some RcTCP genes were showed to be involved drought stress. Seven RcTCP genes were then selected for futher expression validations under PEG (20% w/v PEG6000) treatments by using RT-qPCR method (Figure 7, Supplemental Table S4). Our results showed that all of the seven RcTCP genes were significantly induced by PEG treatments, especially after 24 h treatment, indicating that these seven RcTCP genes may play important roles in the response of drought stress in R. chinensis.

4. Discussion

The plant transcription factors play key roles in all stages of plant development, signaling pathways, stress responses and biosynthesis of bioactive metabolites [25]. TCPs are plant-specific transcription factors that participate in various biological processes such as leaf development, flower symmetry, flowering process, stem branching, and ultimately affects plant growth and development [6]. The genome-wide analysis of TCP genes has been reported in Arabidopsis [26], rice [27], apple [28], grape [29], and other species. Roses, known as the queen of flowers, not only have great ornamental value and economic value but also have cultural value, and which was proposed to be an ideal model species for studying the molecular basis of woody plants [13]. However, systematic and comprehensive analysis on TCP gene family in R. chinensis will facilitate the understanding of flower evolution in plants.

In this study, a total of 16 RcTCP genes were identified in R. chinensis, which are unevenly distributed in five out of seven chromosomes (Figure 1A), suggesting that origination of RcTCP genes might be bias to some kind of subgenomes. Synteny analysis revealed that six pairs of RcTCP genes (containing nine genes) were evolved from segmental duplications, and no tandem duplication events were detected, suggesting that the expansion of TCP gene family was contributed by segmental duplications and might be associated with whole-genome duplication events during evolution of R. chinensis. Phylogenetic analysis indicated that the 16 RcTCP genes could be divided into two subfamilies, namely class I, TCP-P/PCF, and class II, TCP-C, and in which, class II was further divided into two subclasses, namely CyC/TB1 and CIN, which was consistent with the previous studies in Arabidopsis [27], rice [27], tomato [9] and strawberry [5]. Interestingly, the copy number of TCP genes in PCF and CIN subclass from apple (which suffered recent whole genome duplication event, WGD) is much higher than that of other Rosaceae species. By contrast, the copy number of TCP genes in CyC/Tb1 is relatively conserved. This result suggested that there were differences of evolution pressures between those subclasses.

The gene structure analysis showed that RcTCP members from the same group or subgroup have similar motif composition and intron/exon organization but differ between different groups, and motif 2 is only present in class I and motif 3 is only present in class II, suggesting their important roles in diversification and functions of RcTCP gene during gene family evolution. Furthermore, the cis-acting element analysis showed that the cis-acting elements identified from the promoter regions of RcTCPs were classified into four categories. The first is the elements of plant growth and development, including circadian rhythm and endosperm-specific expression, such as circadian rhythm elements detected in RcTCP05, RcTCP07, and RcTCP08. The second is related to plant hormones such as ABA, SA and MeJA. These elements may be involved in the signaling process of hormones. The third is associated with responses to biotic stresses, such as drought, anaerobic and low temperature. The fourth is light-responsive elements, which were detected in all RcTCP genes. The above conclusions are mainly based on data analysis and further studies are needed to elucidate the regulatory mechanisms of TCP genes in R. chinensis.

Expression analysis showed that the RcTCP genes from CyC/TB1 subclass were barely expressed in all detected samples, while RcTCP genes from PCF and CIN subclasses showed very active expression profiles in both of different plant tissues and conditions, suggesting their robust roles in regulations of plant developments and stress response. These results were inconsistent with previous studies, in which RcTCP genes from PCF subclass played a positive cell division regulatory role mainly in different biological processes such as seed germination, leaf and flower organ development, gametophyte development and senescence [3,30,31,32]. A gene that encodes a CIN-like TCP is involved in the morphogenesis of monolearies in Arabidopsis [33]. A single mutant of attcp5 produces wider petals than a wild-type plant [30]. Transient overexpression of FvTCP9 in strawberry fruit significantly promoted the expression of a series of genes related to fruit color and aroma metabolism [34]. PfTCP17 and PfTCP27 of Paulownia fortunei may be associated with resistance to bush disease and drought [33]. In this study, RcTCP05 and RcTCP11 of the CIN subclass and RcTCP10 and RcTCP13 genes of the PCF subclass showed high levels of expression in all plant tissues, suggesting that they might play important roles in the growth and development of rose and the regulation of stress. The RcTCP genes were barely expressed under drought treatment, while most RcTCP genes from PCF subclass and CIN subclass showed higher expression in the phosphorylation of RhPIP2; 1, an aquaporin from rose under drought stress conditions, suggesting that these genes might play indirect roles in the regulation of drought stress response in R. chinensis. Four RcTCP genes from PCF subclasses, RcTCP10, RcTCP13, RcTCP07, and RcTCP02, were expressed at high levels during floral organ differentiation to complete opening, and RcTCP05 from CIN subclass was expressed at a high level during the period from flower initiation to complete open discoloration. RcTCP11 and RcTCP12 were highly expressed in the tested tissues of all but partially open flowers and fully open pink flowers, and both genes were on chromosome five. These results suggested the important roles of RcTCP genes in the regulations of flower organ development.

Drought stress is a common problem in horticultural production, which affects many aspects of plant growth and development. Previous studies have shown that TCP gene plays regulatory roles in drought stress [34]. Overexpression of ZmTCP42 increased the inducibility of ABA or drought tolerance related genes, thereby improving drought tolerance in Zea mays under drought stress [35]. PEG-6000 is an ideal osmotic regulator, which can simulate drought stress by reducing the water potential of the solution to make the root system difficult to absorb water. In our study, there were differences between the PCR results and transcriptome data, perhaps due to the difference between drought and PEG treatment. The previous study has shown that drought stress significantly upregulates aquaporin expression [16]. The phosphorylation of aquaporins decreases with decreasing water potential, thereby reducing the permeability of the water. Based on the drought transcriptome analysis, RcTCP genes from PCF and CIN subclasses are highly induced when the water channel protein RhPIP2; 1 was overexpressed in phosphorylation state under drought conditions, speculating that RcTCPs might be involved in the regulation of water channel proteins via plant hormones and signal transduction, thus responding to the expression of drought stress in plants. In summary, RcTCP genes are mainly involved in plant development, hormonal processes and plant defense. PCF and CIN subclass genes have multiple biological functions, while the function of CyC/TB1 subclass genes is unclear.

5. Conclusions

In this study, 16 RcTCP genes were identified in the R. chinensis genome, and which were classified into class I (PCF class) and class II, where class II genes were further classified into CIN subclass and CyC/TB1 subclass. Subsequently, the analyses of phylogenetic, gene structure, synteny, cis-acting elements and expression analyses based on RNA-seq data and RT-qPCR method provide valuable information for understanding the evolution and functional profiles of TCP genes in R. chinensis.