Genome-Wide Analysis of the D-type Cyclin Gene Family Reveals Differential Expression Patterns and Stem Development in the Woody Plant Prunus mume

Cyclins, a prominent class of cell division regulators, play an extremely important role in plant growth and development. D-type cyclins (CYCDs) are the rate-limiting components of the G1 phase. In plants, studies of CYCDs are mainly concerned with herbaceous plants, yet little information is available about these genes in perennial woody plants, especially ornamental plants. Here, twelve Prunus mume CYCD (PmCYCDs) genes are identified and characterized. The PmCYCDs were named on the basis of orthologues in Arabidopsis thaliana and Oryza sativa. Gene structure and conserved domains of each subgroup CYCDs was similar to that of their orthologues in A. thaliana and O. sativa. However, PmCYCDs exhibited different tissue-specific expression patterns in root, stem, leaf, bud, and fruit organs. The results of qRT-PCR showed that all PmCYCDs, except PmCYCD5;2 and PmCYCD7;1, were primarily highly expressed in leaf buds, shoots, and stems. In addition, the transcript levels of PmCYCD genes were analyzed in roots under different treatments, including exogenous applications of NAA, 6-BA, GA3, ABA, and sucrose. Interestingly, although PmCYCDs were induced by sucrose, the extent of gene induction among PmCYCD subgroups varied. The induction of PmCYCD1;2 by hormones depended on the presence of sucrose. PmCYCD3;1 was stimulated by NAA, and induction was strengthened when sugar and hormones were applied together. Taken together, our study demonstrates that PmCYCDs are functional in plant stem development and provides a basis for selecting members of the cyclin gene family as candidate genes for ornamental plant breeding.


Introduction
The organs of higher plants are developing constantly throughout their life cycle, relying on the flexible control of cell division and cell proliferation.G1/S and G2/M are two of the principal control points within the cell cycle progression [1,2].The mechanisms that control the G1/S transition in mammals and plants are functionally conserved and are characterized as D-type cyclin (CYCD)/retinoblastoma (Rb) pathways [3].CYCD and a cyclin-dependent kinase (CDK) form a CYCD-CDK complex that binds to the Rb protein through the LxCxE (x represents any amino acid) motif of CYCD to activate the phosphorylation of the Rb protein [4,5].Prior to activation, the Rb protein is bound to transcription factor E2Fs, which is released via phosphorylation of the Rb protein, thus activating G1/S phase transition [4][5][6].
CYCDs are the rate-limiting components through the G1 phase.In plants, the expression levels and activities of CYCDs are affected by the levels of hormones and carbohydrates, which exert a great impact on the cell division process in plants.Since the first plant CYCD was discovered in Arabidopsis thaliana [7], more CYCDs have been identified from various organisms.However, most of them were characterized from herbaceous species such as A. thaliana, tobacco, and rice [6,[8][9][10][11].At present, 10 genes encoding CYCDs have been identified in A. thaliana, which are divided into seven sub-families (CYCD1-CYCD7).Specifically, there are three members in the CYCD3 sub-family, two members in the CYCD4 subfamily, and one gene in each of the other five CYCD sub-families [12].CYCD1;1, CYCD2;1 and CYCD3;1 were initially screened from the genome of A. thaliana and were capable of restoring the screened phenotype of the yeast cyclin G1 mutant [7].Most CYCD proteins can interact with CDKA;1, and the over-expression of some CYCD genes can promote S phase transition [13,14], which indicates that the CDKA-CYCD complex can regulate the G1/S phase transition.In fact, expressing the AtCYCD2;1 gene of A. thaliana in tobacco can increase the overall growth rate of bud tips, but the final size of each organ is consistent with the wild type [15].Previous studies showed that some CYCD genes are specifically present in tissues or organs.The most obvious evidence is the transcript pattern of two AmCYCD3 genes in Antirrhinum majus L.; AmCYCD3b specifically presents in all dividing cells, while AmCYCD3a is not expressed in stem tip meristem, but is observed mainly in the primordial of lateral organs, which is similar to expression pattern of CYCD3 in the lateral organs of A. thaliana [16].Further detailed studies have shown that sucrose has a significant impact on the expression of CYCD2 and CYCD3 in the G1 phase in A. thaliana [17].Particularly, in sucrose-deficient cell cultures (in a still state), i.e., in the early G1 phase, the mRNA level of CYCD2 can be enhanced at 30 min after sucrose addition.The mRNA level of CYCD3 increases at a later time (4 h after sucrose addition), around the late G1 to S phase.The expression of CYCD2 and CYCD3 induced by sucrose indicates that gene expression is a direct response to sucrose, rather than an indirect result of cell growth and differentiation, which has been confirmed by adding a certain concentration of cycloheximide (CHX) into cell culture medium.CHX is a protein synthesis inhibitor, which can block new protein synthesis and cell cycle progression [17,18].The mechanism of CYCD3 and CYCD2 induction by exogenous signals involves protein phosphatase [19].Interestingly, the accumulation of CYCD2 mRNA is less sensitive to CHX than that of CYCD3, suggesting that sucrose-induction may be regulated by different protein phosphorylations.Different regulation patterns of CYCD2 and CYCD3 in A. thaliana are also confirmed by their response to cytokinin.When cytokinin was added to the cell culture medium, CYCD3 expression was induced; when cytokinin was removed, the expression level of CYCD3 was reduced, whereas the expression level of CYCD2 was not affected, regardless of whether hormone concentration was increased or decreased [17,18].Therefore, sucrose plays a dominant role in regulating CYCD3 expression compared with cytokinin.
Although CYCDs play an essential role in cell cycle progression, only limited plant cyclins have been characterized functionally.Until now, studies involving CYCD functional characteristics in plants have been limited to a handful of CYCD genes and have only been performed on herbaceous species, such as Arabidopsis and tobacco, yet little information is available about CYCD genes in woody trees.Due to the significant differences between herbaceous plants and woody plants in genome size and secondary growth, studies concerning cyclins in herbaceous plants cannot accurately reflect the function and role of cyclins in woody plants.Menges et al. [20] identified 22 poplar (Populus trichocarpa) CYCD genes by searching the P. trichocarpa genome sequence.Over-expression of wheat gene, TaCYCD2;1 in Arabidopsis up-regulated the transcript of vascular development-related genes, including the KNAT1, PHB, and PHV gene [21].Transferring the poplar gene, PtaCYCD1;2, in transgenic poplar seedlings result in smaller cell size, curved leaf and petioles, and tortuous stem phenotype [22].The above results speculate that some CYCDs may play an important role in regulating the development of stem and vascular morphogenesis in plants.
Prunus mume is a woody plant that has high ornamental and economic value and is widely distributed and used in landscape architecture in Asia.P. mume has acquired favorable ornamental characteristics, including colorful corollas, pleasant fragrance, and various types of flowers and architecture (including upright type, weeping type, and tortuous type).To date, very few genes related to architectural traits (i.e., tortuous trait) have been reported in woody plants.CYCDs, that are closely related to cell division and proliferation, were newly found to regulate the stem development and plant architecture of woody plants [23][24][25].In this study, we identified 12 CYCDs from the P. mume genome and performed detailed structure, evolutionary, and expression pattern analyses in different organs, as well as analyzing different responses to hormones and sucrose.Our study provides a basis for further study of the cyclin gene family and reserves candidate genes for ornamental plant breeding for architectural characteristics.

Gene Structure and Conserved Motif Analyses
Coding sequence (CDS) and genome sequences of PmCYCDs were gained from P. mume genome and used for the gene structure (exon/intron) analysis, using GSDS 2.0 [37].Conserved motifs of PmCYCDs were analyzed using the motif-based sequence analysis tools (MEME) [38], with the following specific parameters: The maximum numbers of motifs, 20; optimum motif width was set at 30-50.

Transcriptome Data Analysis
Transcriptional tissue-specific expression profile data were downloaded from the GEO at NCBI (http://www.ncbi.nlm.nih.gov/geo/)[39].RNA samples of five tissues (bud, leaf, root, stem, and fruit) were extracted from the P. mume (GEO No. GSE40162).PmCYCD genes in five organs with RPKM > 0 were collected, and the Log2 value of RPKM was converted.Heat maps were generated using HemI (1.0) software [40].

Plant, Tissues, and Phytohormone Treatment
To detect the expression level in different tissues, flowers were sampled from ten-year-old P. mume plants that were planted on the campus of Beijing Forestry University.Leaf buds, apex shoots, stems, young leaves, and roots were obtained from six-month-old seedlings.

RNA Extraction and qRT-PCR
The total RNA was extracted from samples, using MiniBEST Plant RNA Extraction Kit (TaKaRa, Dalian, China) and treated with DNase I (TaKaRa).The quantification of RNA, integrity assessment and purity were analyzed by using a NanoDrop 2000c spectrophotometer (Thermo-Scientific, Wilmington, DE, USA).A 500 ng aliquot of total RNA was reverse-transcribed to cDNA using PrimeScript RT reagent Kit with gDNAEraser (Perfect Real Time) (TaKaRa).Two microliters of cDNA was used as template according to SYBR Premix EX Taq II Kit (TaKaRa) and qRT-PCR was performed by a CFX96 Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA).The gene-specific primers were designed by RealTime qPCR Assay tool (https://sg.idtdna.com/scitools/Applications/RealTimePCR/) [41] and Pmactin and PmPP2A genes were selected as internal references.Gene expression levels were determined using the 2-delta-delta Ct method [42].Three biological and technical replicates for each reaction were performed to ensure reproducibility of the results.Statistical significance was determined by one-way analysis of variance (ANOVA).All the primers used in text for qRT-PCR are listed in Table S1.

Identification of Putative PmCYCDs
The cyclin-N (PF00134) and cyclin-C (PF02984) profiles were used to conduct a HMMER (Biosequence analysis using profile hidden Markov models) search against the protein database of P. mume, and 10 D-type cyclin from A. thaliana were used to perform a BLASTP search against P. mume cyclin protein sequences obtained from the HMMER search.Finally, a total of 12 highly confident CYCD genes were identified from P. mume and used for further analysis (Table 1).To investigate the classification of CYCD proteins in P. mume and their evolutionary relationships in P. mume, A. thaliana, and O. sativa, full-length CYCD proteins were used to construct a ML phylogenetic tree (Figure 1).CYCD proteins of P. mume were divided into six subfamilies that were assigned to different subgroups of homologous genes reported in Arabidopsis and rice: CYCD1, CYCD2/CYCD4, CYCD3, CYCD5, CYCD6, and CYCD7 (Table S2): There are two each of the CYCD1, CYCD2/D4, CYCD5, and CYCD6-type cyclins and one CYCD7-type cyclin, and three CYCD3-type cyclins.The characteristics of the 12 PmCYCD proteins were varied greatly.The length of protein sequences ranged from 265 to 398 amino acids, The MW and pI values ranged from 35.55 to 45.63 kDa, and 4.85 to 6.71, respectively (Table 1).Eleven PmCYCD genes are distributed in 5 of the 8 chromosomes (Chr), with Chr 1 and 2 having most of cyclins, and one PmCYCD (PmCYCD1;2) gene was located on anchored scaffolds (Table 1 and Figure S1).After the genetic divergence between A. thaliana and P. mume, there were four PmCYCD genes, showing evidence of inter-chromosomal duplications, while without intra-chromosomal duplications (Figure S1).The different locations of PmCYCDs on chromosomes may be one of the evidence of the existence of 11 different cyclin genes.Eleven PmCYCD genes are distributed in 5 of the 8 chromosomes (Chr), with Chr 1 and 2 having most of cyclins, and one PmCYCD (PmCYCD1;2) gene was located on anchored scaffolds (Table 1 and Figure S1).After the genetic divergence between A. thaliana and P. mume, there were four PmCYCD genes, showing evidence of inter-chromosomal duplications, while without intra-chromosomal duplications (Figure S1).The different locations of PmCYCDs on chromosomes may be one of the evidence of the existence of 11 different cyclin genes.

Gene Structure and MEME Analysis
To assess and increase insights into the diversity of gene structure, the exon/intron structures of PmCYCDs were constructed on the basis of an independent phylogenetic tree of PmCYCDs using complete protein sequences (Figure 2a).The number of exons/introns in the same subfamilies usually showed similarity, with most introns ranging from 3 to 5 (Figure 2b).For instance, the subfamily PmCYCD2/PmCYCD6 shared 5 introns, but the subfamily PmCYCD3 contains 3 introns, and PmCYCD5 contains 4 introns.Intron numbers in subfamily PmCYCD1 were variable, PmCYCD1;1 contained 3 introns and PmCYCD1;2 contained 5 introns.

Gene Structure and MEME Analysis
To assess and increase insights into the diversity of gene structure, the exon/intron structures of PmCYCDs were constructed on the basis of an independent phylogenetic tree of PmCYCDs using complete protein sequences (Figure 2a).The number of exons/introns in the same subfamilies usually showed similarity, with most introns ranging from 3 to 5 (Figure 2b).For instance, the subfamily PmCYCD2/PmCYCD6 shared 5 introns, but the subfamily PmCYCD3 contains 3 introns, and PmCYCD5 contains 4 introns.Intron numbers in subfamily PmCYCD1 were variable, PmCYCD1;1 contained 3 introns and PmCYCD1;2 contained 5 introns.To further clarify the specific conserved region of PmCYCDs, a total of nineteen unique and nonredundant motifs were detected and distributed by the MEME tool (Figure 2c).Motifs 2, 3, 4, 6, 7, and 8 were found in most of the PmCYCD genes, indicating that they participate in the composition of cyclin-N domain (motifs 2, 3, 4, and 6) and cyclin-C domain (motifs 7 and 8).The same subgroup of PmCYCDs shares similar motifs.In addition, some unknown motifs have also been found.Motif 9 was found mainly in the N-terminal regions and possibly the LxCxE region.In addition, two special motifs, 12 and 14, were found to be unique to the CYCD2 subfamily.The motif structure of the CYCD5 subfamily consisted of seven motifs, less than that in the other CYCD motif structures.

Core Conservative Region Analysis
Plant cyclins contain two highly conserved regions in the middle of the protein sequence, named cyclin-N, and cyclin-C domains, respectively.The cyclin-N domain is a region of approximately 130 aa, existing in all the cyclins, mainly located in the regions between the second exon and third exon (Figure S2).The end of cyclin-N domain contains a cyclin box, a short polypeptide sequence, which binds to the CDK for cyclin activity.All PmCYCDs contain the cyclin-N domain and the cyclin box (Figure 3 and Table S3).The cyclin box is located in the region between the first exon and the second exon.Cyclin-C domain is also one of the important, but not an essential components of cyclin, which is absent only in 1 (PmCYCD5;1) of the 12 PmCYCDs.To further clarify the specific conserved region of PmCYCDs, a total of nineteen unique and non-redundant motifs were detected and distributed by the MEME tool (Figure 2c).Motifs 2, 3, 4, 6, 7, and 8 were found in most of the PmCYCD genes, indicating that they participate in the composition of cyclin-N domain (motifs 2, 3, 4, and 6) and cyclin-C domain (motifs 7 and 8).The same subgroup of PmCYCDs shares similar motifs.In addition, some unknown motifs have also been found.Motif 9 was found mainly in the N-terminal regions and possibly the LxCxE region.In addition, two special motifs, 12 and 14, were found to be unique to the CYCD2 subfamily.The motif structure of the CYCD5 subfamily consisted of seven motifs, less than that in the other CYCD motif structures.

Core Conservative Region Analysis
Plant cyclins contain two highly conserved regions in the middle of the protein sequence, named cyclin-N, and cyclin-C domains, respectively.The cyclin-N domain is a region of approximately 130 aa, existing in all the cyclins, mainly located in the regions between the second exon and third exon (Figure S2).The end of cyclin-N domain contains a cyclin box, a short polypeptide sequence, which binds to the CDK for cyclin activity.All PmCYCDs contain the cyclin-N domain and the cyclin box (Figure 3 and Table S3).The cyclin box is located in the region between the first exon and the second exon.Cyclin-C domain is also one of the important, but not an essential components of cyclin, which is absent only in 1 (PmCYCD5;1) of the 12 PmCYCDs.The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).
O-Glycosylation and N-glycosylation of PmCYCD proteins were separately predicted using glycosylation prediction software NetOGlyc 4.0 Server (http://www.cbs.dtu.dk/services/NetOGlyc/)[45] and NetNGlyc 1.0 Server (http://www.cbs.dtu.dk/services/NetNGlyc/)[46].The results showed that all 12 PmCYCD proteins contained many O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained one N-glycosylation site (N in NLS, Nuclear location signal) (Table S3).Proline-directed kinases, such as cyclin-dependent protein kinase 5 (CDK5) have been implicated in the phosphorylation of cell division.Conserved motifs S/TP, associated with proline-directed kinases in PmCYCDs, are mainly located at the C-terminal.Among them, the most conserved site (equivalent to Thr-191 in NtCYCD3;3) can be found in most CYCDs in P. mume, Arabidopsis, and rice (Figure S2).The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).

Gene Expression Analysis by Transcriptome
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).

Gene Expression Analysis by Transcriptome
C-glycosylation sites.
The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to the Retinoblastoma-related (RBR) protein and phosphorylates Rb protein to release E2Fs, thus activating cell transition from G1 phase to S phase.All PmCYCDs except PmCYCD1;1, PmCYCD6;1 and PmCYCD6;2 contain the LxCxE sequence (Figure 3 and Table S3).
Five non-contiguous conserved amino acids [arginine (R), aspartic acid (D), leucine (L), lysine (K), and glutamic acid (E)] have been confirmed to be essential for the structure of cyclins, which are also conserved in CYCDs [43].The result of multiple-sequence alignment, showed that 12 PmCYCDs contain all the above-mentioned conserved amino acids (Figure S2).
The PEST sequence is a short amino acid sequence region rich in Pro (P), Glu (E), Ser (S), and Thr (T) residues, which is involved in ubiquitin ligase and proteasome hydrolysis systems.The result of the prediction of the PEST motif of general D-type cyclins in A. thaliana and P. mume showed that all AtCYCDs contain at least one poor PEST domain (PESTfind score: >0), with the exception of AtCYCD4;2 [20,44].Four-twelfths of the PEST sequences were not predicted in P. mume (Table S3).

Gene Expression Analysis by Transcriptome
The expression level of PmCYCDs was examined in five different types of organ tissue: Root, stem, leaf, bud, and fruit (Table S4) with our previous transcriptome data (No.GSE40162).As shown in Figure 4, we found that PmCYCDs exhibited specific expressions in five different organs.Most of the genes presented relatively higher transcript level in the stem, except PmCYCD6;1 and PmCYCD2;2, PmCYCD2;1, PmCYCD3;2 and PmCYCD3;3 had high expression levels in five samples.PmCYCD1;1 and PmCYCD5;1 had higher expression levels in both the buds and stems, suggesting simultaneous involvement in the regulation of bud and stem development.The expression levels of the PmCYCD6;1 and PmCYCD7;1 genes were very low in above five organs.
Forests 2018, 9, x FOR PEER REVIEW 3 of 18 The expression level of PmCYCDs was examined in five different types of organ tissue: Root, stem, leaf, bud, and fruit (Table S4) with our previous transcriptome data (No.GSE40162).As shown in Figure 4, we found that PmCYCDs exhibited specific expressions in five different organs.Most of the genes presented relatively higher transcript level in the stem, except PmCYCD6;1 and PmCYCD2;2, PmCYCD2;1, PmCYCD3;2 and PmCYCD3;3 had high expression levels in five samples.PmCYCD1;1 and PmCYCD5;1 had higher expression levels in both the buds and stems, suggesting simultaneous involvement in the regulation of bud and stem development.The expression levels of the PmCYCD6;1 and PmCYCD7;1 genes were very low in above five organs.To verify the accuracy of the transcript levels of PmCYCDs in transcriptome data, specific primers were designed for each cyclin, and their expression patterns in flowers, young buds, shoots, stems, young leaves, and roots were investigated using qRT-PCR with Pmactin and PmPP2A as housekeeping genes (Figure 5).The expression patterns were observed as follows: All PmCYCD subfamily genes were highly expressed in shoot, stem, and young buds except PmCYCD7;1 and PmCYCD5;2.PmCYCD7;1 appeared to be highly expressed in flowers, and PmCYCD5;2 was more highly expressed in root tissues.To verify the accuracy of the transcript levels of PmCYCDs in transcriptome data, specific primers were designed for each cyclin, and their expression patterns in flowers, young buds, shoots, stems, young leaves, and roots were investigated using qRT-PCR with Pmactin and PmPP2A as housekeeping genes (Figure 5).The expression patterns were observed as follows: All PmCYCD subfamily genes were highly expressed in shoot, stem, and young buds except PmCYCD7;1 and PmCYCD5;2.PmCYCD7;1 appeared to be highly expressed in flowers, and PmCYCD5;2 was more highly expressed in root tissues.

The Transcript Level of PmCYCDs under Sucrose and Phytohormone Treatments
As essential regulators of the cell cycle, CYCD genes are important for understanding their differential responses to mitogenic signals.As shown in Figure 6, the expression pattern of PmCYCD genes was observably induced by sucrose and phyto-hormones.The transcript level of PmCYCD6;2 was significantly up-regulated 4.2-fold with sucrose treatment.In the presence of sucrose, PmCYCD1;2, PmCYCD3;1, PmCYCD5;1, and PmCYCD7;1 can induce 2-3 fold increase in expression level, while the transcript level of PmCYCD2;2 was less than 2-fold.Auxin and cytokinin could strongly stimulate the expression of PmCYCD3;1, PmCYCD6;2, and PmCYCD1;2.When GA3 was applied, the transcript levels of PmCYCD1;2, PmCYCD2;2, PmCYCD3;1, PmCYCD6;2, and PmCYCD7;1 were markedly upregulated, especially, expression level of PmCYCD3;1, and PmCYCD6;2, was more than three times higher than that before treatment.The presence of ABA resulted in a distinct down-regulation of most PmCYCDs, except PmCYCD7;1 and PmCYCD1;2.With the combinations of auxin, cytokinin, gibberellin, and sucrose, the expression level of all PmCYCDs was significantly up-regulated by more than two times than in the control, suggesting that there was a synergistic effect between different hormones, or between hormones and sucrose on induced expression of PmCYCDs.Interestingly, the promoting effects on PmCYCD1;2, PmCYCD2;2, and PmCYCD5;1 transcription were not activated when ABA and sucrose were both applied.

The Transcript Level of PmCYCDs under Sucrose and Phytohormone Treatments
As essential regulators of the cell cycle, CYCD genes are important for understanding their differential responses to mitogenic signals.As shown in Figure 6, the expression pattern of PmCYCD genes was observably induced by sucrose and phyto-hormones.The transcript level of PmCYCD6;2 was significantly up-regulated 4.2-fold with sucrose treatment.In the presence of sucrose, PmCYCD1;2, PmCYCD3;1, PmCYCD5;1, and PmCYCD7;1 can induce 2-3 fold increase in expression level, while the transcript level of PmCYCD2;2 was less than 2-fold.Auxin and cytokinin could strongly stimulate the expression of PmCYCD3;1, PmCYCD6;2, and PmCYCD1;2.When GA3 was applied, the transcript levels of PmCYCD1;2, PmCYCD2;2, PmCYCD3;1, PmCYCD6;2, and PmCYCD7;1 were markedly upregulated, especially, expression level of PmCYCD3;1, and PmCYCD6;2, was more than three times higher than that before treatment.The presence of ABA resulted in a distinct down-regulation of most PmCYCDs, except PmCYCD7;1 and PmCYCD1;2.With the combinations of auxin, cytokinin, gibberellin, and sucrose, the expression level of all PmCYCDs was significantly up-regulated by more than two times than in the control, suggesting that there was a synergistic effect between different hormones, or between hormones and sucrose on induced expression of PmCYCDs.Interestingly, the promoting effects on PmCYCD1;2, PmCYCD2;2, and PmCYCD5;1 transcription were not activated when ABA and sucrose were both applied.

Discussion
The organs of higher plants are constantly evolving throughout their life cycle, and this process relies on the flexible control of cell division and cell proliferation.In the model plant Arabidopsis, there are 32 cyclin genes that can regulate the cell cycle, including 10 A-type, 11 B-type, 10 D-type, and 1 H-type [47].Each cyclin gene exhibits a specific expression pattern in the cell cycle.In general, A-type cyclins (CYCAs) not only control the progression of S-M phase, but also regulate the transformation of G2/M phase and the progression of M phase with CYCBs [48,49].CYCDs are expressed continuously in the cell cycle to control G1/S transition and cell proliferation in response to corresponding external signals, including plant hormones and carbon sources [8,15].In this study, 12 PmCYCD genes in P. mume were identified and named in the light of homologs in A. thaliana.The results of the ML phylogenetic tree illustrated that six subgroups of PmCYCDs were similar to those in A. thaliana and O. sativa.Each subgroup of CYCD was composed of at least one gene: 2 in CYCD1 subgroup, 2 in CYCD2/CYCD4 subgroup, 3 in CYCD3 subgroup, 2 in CYCD5 subgroup, 2 in CYCD6 subgroup and 1 in CYCD7 subgroup (Figure 1 and Table S2).
The genomic organizations of D type cyclin genes and three subgroups (CYCD1, CYCD2/4, and CYCD6) are highly conserved in angiosperms [50].In our study, PmCYCD1;2, PmCYCD2;1, PmCYCD2;2, PmCYCD6;1, PmCYCD6;2, and PmCYCD7;1 all had six exons and seven introns (Figure 2b), which is similar to those in Arabidopsis and maize [12,50].Three AtCYCD3 and six PtrCYCD3 members from Arabidopsis and P. trichocarpa all have four exons, with a conserved length in the first exon and central exon, suggesting that plant CYCDs are derived from the structure of ancestral chromosomes [20].In P. mume, three PmCYCD3 subgroup genes and PmCYCD1;1 genes contain four exons, suggesting the species in dicotyledons shared similar evolutionary events.In addition, other PmCYCD genes have five exons.The third and fourth ancestral exons fused into one exon in all CYCD5 genes of angiosperms.In contrast, CYCD7 genes may have lost and/or gained different introns among different plant species [20].Highly conserved motifs in the PmCYCD were further identified by MEME tool (Figure 2c).Most of PmCYCDs in the same subgroup have similar motifs, suggesting these motifs are indispensable in subgroup-specific functions.However, the structure difference of PmCYCD6 subgroups is very obvious.Among them, only seven motifs were found in PmCYCD6;1, with only motif one in its cyclin-N region, whereas PmCYCD6;1 has 12 motifs.Many motifs were on both the N-and C-terminals, indicating the complex nature of the function.Interestingly, the N-terminal in PmCYCD2 contains a specific motif (motif 12) that was not found in

Discussion
The organs of higher plants are constantly evolving throughout their life cycle, and this process relies on the flexible control of cell division and cell proliferation.In the model plant Arabidopsis, there are 32 cyclin genes that can regulate the cell cycle, including 10 A-type, 11 B-type, 10 D-type, and 1 H-type [47].Each cyclin gene exhibits a specific expression pattern in the cell cycle.In general, A-type cyclins (CYCAs) not only control the progression of S-M phase, but also regulate the transformation of G2/M phase and the progression of M phase with CYCBs [48,49].CYCDs are expressed continuously in the cell cycle to control G1/S transition and cell proliferation in response to corresponding external signals, including plant hormones and carbon sources [8,15].In this study, 12 PmCYCD genes in P. mume were identified and named in the light of homologs in A. thaliana.The results of the ML phylogenetic tree illustrated that six subgroups of PmCYCDs were similar to those in A. thaliana and O. sativa.Each subgroup of CYCD was composed of at least one gene: 2 in CYCD1 subgroup, 2 in CYCD2/CYCD4 subgroup, 3 in CYCD3 subgroup, 2 in CYCD5 subgroup, 2 in CYCD6 subgroup and 1 in CYCD7 subgroup (Figure 1 and Table S2).
The genomic organizations of D type cyclin genes and three subgroups (CYCD1, CYCD2/4, and CYCD6) are highly conserved in angiosperms [50].In our study, PmCYCD1;2, PmCYCD2;1, PmCYCD2;2, PmCYCD6;1, PmCYCD6;2, and PmCYCD7;1 all had six exons and seven introns (Figure 2b), which is similar to those in Arabidopsis and maize [12,50].Three AtCYCD3 and six PtrCYCD3 members from Arabidopsis and P. trichocarpa all have four exons, with a conserved length in the first exon and central exon, suggesting that plant CYCDs are derived from the structure of ancestral chromosomes [20].In P. mume, three PmCYCD3 subgroup genes and PmCYCD1;1 genes contain four exons, suggesting the species in dicotyledons shared similar evolutionary events.In addition, other PmCYCD genes have five exons.The third and fourth ancestral exons fused into one exon in all CYCD5 genes of angiosperms.In contrast, CYCD7 genes may have lost and/or gained different introns among different plant species [20].Highly conserved motifs in the PmCYCD were further identified by MEME tool (Figure 2c).Most of PmCYCDs in the same subgroup have similar motifs, suggesting these motifs are indispensable in subgroup-specific functions.However, the structure difference of PmCYCD6 subgroups is very obvious.Among them, only seven motifs were found in PmCYCD6;1, with only motif one in its cyclin-N region, whereas PmCYCD6;1 has 12 motifs.Many motifs were on both the N-and C-terminals, indicating the complex nature of the function.Interestingly, the N-terminal in PmCYCD2 contains a specific motif (motif 12) that was not found in other subgroups of PmCYCDs (Figure 2c), suggesting that motif 12 was involved in a specific function to play specific roles in cell cycle progression.
Cyclin core, as the core composition of a typical cyclin protein, consists of a highly conserved cyclin-N and less conserved domains [51].Specific cyclins, without cyclin-C domain, which also play an important role in biological functions, indicate that the cyclin-C domain may not be critical for its function [52].In Figure 3, twelve PmCYCDs all contain cyclin-N and cyclin-C, with the exception of PmCYCD5;1, which is similar to maize [50] and cucumber [53].The CYCDs in higher plants and animals generally contain an N-terminal LxCxE (x stands for any amino acid) domain, which binds to, and phosphorylates, the RB protein to release E2Fs, thus activating cell transition from G1 phase to S phase [15,54,55].In A. thaliana, LxCxE domain in AtCYCD5 was replaced by FxCxE motif, whereas CYCD4;2 and CYCD6;1 were absent of LxCxE domain interacting with RB protein [12].In P. mume, all 12 PmCYCDs studied contained the LxCxE motif except for PmCYCD1;1, PmCYCD6;1, and PmCYCD6;2.Interestingly, PtrCYCD1;4, PtrCYCD5;3, and PtrCYCD6 in poplar also have no LxCxE motifs [20].In addition, CYCDs also contain a conserved cyclin box, which is the binding domain of CDK or inhibitors of CDK (ICK) and forms a CDK/ICK-Cyclin complex [56,57].It has also been reported that some CYCD cyclins can be rapidly degraded, which depends mainly on the presence of PEST sequences [2,58].
After comparing three AtCYCDs in Arabidopsis and four NtCYCDs in tobacco, it was found that PmCYCD1;1 and PmCYCD1;2 both have one potential PEST region at the N-terminus, which is similar to the predicted results of NtCYCD1;1 and NtCYCD3;1 (Table S5).In addition, PmCYCD1;2 also has a potential PEST region at the C-terminus, which is similar in Arabidopsis.AtCYCD1;1, AtCYCD2;1, and AtCYCD3;1 also have one PEST region at the C-terminus (Table S5).There are various types of protein post-translational modifications, such as methylation, acetylation, bio-tinylation, glycosylation, and phosphorylation.Among these modifications, phosphorylation and glycosylation may result in an increase in protein molecular weight.O-Glycosylation and N-glycosylation of the PmCYCD protein were separately predicted, using glycosylation prediction software NetOGlyc 4.0 Server and NetNGlyc 1.0 Server.The results showed that all PmCYCD proteins contained O-glycosylation sites (T or S), and 7 of 12 PmCYCDs contained N-glycosylation sites (N in NLS) (Figure 3 and Table S4).The molecular weight of PmCYCDs may change significantly in the natural state in plants, which may be due to the increase in the quantity or length of carbohydrate chains during the glycosylation process.Previous studies showed that the PtFLA6 gene in poplar was modified by glycosylation, leading to a significant increase in the molecular weight [59].
Understanding gene expression patterns is the first step in confirming gene biological function.Plant CYCDs are predominantly expressed in tissues or organs involved in cell division and proliferation, such as meristems in root or stems and differentiating tissues.Low level expression of major PmCYCDs were detected in leaf tissue (Figures 4 and 5), suggesting they are either by-products of previous processes or involved in differentiation-related processes, with an unknown role.Genes with similar expression profiles in different tissues or developmental stages may participate in similar functions or processes.A transcriptional level of 12 PmCYCDs can be divided into three groups.The first group contains a majority of PmCYCDs, which shows a high transcript level in young bud, shoot, and stem organs (Figure 5).The specific expression of second-group CYCDs genes were detected in roots.The third-group CYCDs genes were majorly detected in flowers.Interestingly, CYCDs in the same subgroups performed similar expression patterns, such as PmCYCD3;1, PmCYCD3;2 and PmCYCD3;3.The similar expression patterns of PmCYCD2;1 and PmCYCD2;2, suggest that the same subgroup genes may share functional redundancy and conservation.Surprisingly, some CYCDs of P. mume and Arabidopsis, within the same clade in phylogenetic analysis showed a distinct expression pattern.AtCYCD7;1 was not detected in leaves, flowers, and roots [11], whereas PmCYCD7;1 was detected in all of the P. mume tissues examined and showed the highest expression level in flowers (Figure 5).AtCYCD5 was highly expressed in inflorescences [60], while PmCYCD5;2 showed a higher expression level in roots.The aforementioned results demonstrate the homologous CYCDs from different species may perform similar biological function during the cell cycle, whereas the specific roles may exist in different plants.In summary, the specific expression pattern lays a foundation for further study of the function of PmCYCDs.
Plant cyclin genes respond to different exogenous stimulus signals, such as mitogenic signals, as seen with the addition of different phyto-hormones and sucrose [61][62][63].CYCDs participate in the cell division at the G1 to S phase, which can be regulated by various phyto-hormones [64].Gibberellins (GAs) are one type of phyto-hormone that regulate various developmental processes, such as growth, stem elongation, flower development dormancy, and senescence [65].Evidence that GA3 regulates cell division and proliferation has been declared [65,66].ABA (Abscisic acid) could maintain tobacco cells in G1 phase by inhibiting DNA synthesis [67].Moreover, the addition of exogenous auxin and cytokinin results in CYCD3 transcript accumulation and increases cell division in Arabidopsis [17,68].Using cell cultures, AtCYCD3 was verified in response to several hormones at differential transcript level, but no response was observed for AtCYCD2, suggesting the differential roles of CYCDs [69].In our study, the transcript levels of PmCYCDs were modulated by phyto-hormones and various combinations of phyto-hormones with sucrose (Figure 6).PmCYCD1;2, PmCYCD2;2, PmCYCD3;1, PmCYCD5;1, PmCYCD6;2, and PmCYCD7;1 were upregulated by sucrose treatment.PmCYCD6;2 was strongly upregulated by sucrose, as it underwent the largest fold change in expression.Moreover, PmCYCD3;1 may be more sensitive to sucrose than PmCYCD2;1 and PmCYCD1;1.The results of sucrose and phyto-hormone applications showed that CYCDs may play an important role in responding to carbon source signals.Previous work has reported that AtCYCD3 can rapidly respond to a variety of hormones [19,70,71].Conversely, At CYCD2 was not involved in induction of hormones [72].The response relationship between phyto-hormones and CYCD is not clear.In this paper, the 2000 bp upstream promoter sequences of 12 PmCYCDs were predicted and hormone-related motifs were screened (Table S6).PmCYCD1;2 was induced by NAA (1-Naphthylacetic acid) and the auxin-responsive element was only found in the PmCYCD1;2 promoter.PmCYCD6;2 may have important roles in responding to external stimulus because it was upregulated under all sucrose and hormone treatments except for the abscisic acid treatment.In PmCYCD6;2, gibberellin-, abscisic acidand methyl jasmonate-responsive elements were identified but found no auxin-or cytokinin-related elements, which is similar to the results of the hormone response.The complex expression patterns of PmCYCD genes under sucrose and phyto-hormone treatments indicated that the six separate PmCYCD groups possess different expression characteristics.As above, our study lay the foundation for systematic analysis of PmCYCDs gene family function in plant stem development and provides a basis for selecting members of the cyclin gene family as candidate genes for further study.

Conclusions
We identified and characterized twelve CYCD genes (PmCYCDs) in P. mume.Phylogenetic analysis showed that there were six sub-families in the CYCD family.The gene structure and conserved domains of each subgroup CYCDs was similar to that of their orthologues in A. thaliana and O. sativa.The qRT-PCR showed that all PmCYCDs, except PmCYCD5;2 and PmCYCD7;1 were primarily highly expressed in leaf buds, shoots, and stems.The transcript levels of PmCYCD genes were up-regulated by exogenous applications of NAA, 6-BA, GA3, ABA, and sucrose.The expression level of PmCYCD1;2 was induced by hormones depended on the presence of sucrose.PmCYCD3;1 was stimulated by NAA, and induction was strengthened when sugar and hormones were applied together.Our study demonstrates that PmCYCDs are functional in plant stem development and provides a basis for the further study of members of the cyclin gene family as candidate genes for ornamental woody breeding.

Supplementary Materials:
The following are available online at http://www.mdpi.com/1999-4907/10/2/147/s1. Figure S1: Chromosomal distribution of PmCYCD genes on eight P. mume chromosomes, Figure S2: Alignment showing domains and motifs in protein sequences of PmCYCDs, Table S1: Primer sequences used for qRT-PCR in this study, Table S2: CYCDs in Arabidopsis, rice, and P. mume divided in the previously reported subgroups, Table S3: Location of cyclin-specific domains and motifs in the protein sequences of P. mume D-type cyclins, Table S4: Log2 value of RPKM of PmCYCDs in five different tissue organs, Table S5: Distribution of PEST sequence of CYCDs in different species, Table S6: Hormone-related Cis-acting elements predicted in the promoter of PmCYCDs.

Figure 1 .
Figure 1.Phylogenetic tree analysis of P. mume CYCD (PmCYCDs) genes from P. mume, A. thaliana and rice.Multiple sequence alignments of full length sequences were performed and the ML tree was generated using MEGA 5.0 software.Each subgroup is distinguished in a different color.

Figure 1 .
Figure 1.Phylogenetic tree analysis of P. mume CYCD (PmCYCDs) genes from P. mume, A. thaliana and rice.Multiple sequence alignments of full length sequences were performed and the ML tree was generated using MEGA 5.0 software.Each subgroup is distinguished in a different color.

Figure 2 .
Figure 2. Phylogenetic relationship and motif distribution of PmCYCDs.(a) Mutiple alignments of 12 full-length protein sequences of PmCYCDs were performed by MEGA 5.0.The 12 PmCYCDs proteins were clustered into six phylogenetic sub-families.Each sub-family is marked with different background colors; (b) Exon-intron structures of PmCYCDs.Exons and introns are indicated by green rectangles and horizontal lines, respectively.The scale estimates the lengths of exons and introns.(c) Motif distribution of PmCYCDs.Schematic representation of the conserved motifs in PmCYCD proteins were identified by MEME.Each motif is presented by a different colored block.

Figure 2 .
Figure 2. Phylogenetic relationship and motif distribution of PmCYCDs.(a) Mutiple alignments of 12 full-length protein sequences of PmCYCDs were performed by MEGA 5.0.The 12 PmCYCDs proteins were clustered into six phylogenetic sub-families.Each sub-family is marked with different background colors; (b) Exon-intron structures of PmCYCDs.Exons and introns are indicated by green rectangles and horizontal lines, respectively.The scale estimates the lengths of exons and introns.(c) Motif distribution of PmCYCDs.Schematic representation of the conserved motifs in PmCYCD proteins were identified by MEME.Each motif is presented by a different colored block.
five non-contiguous conserved amino acids,

Figure 4 .
Figure 4. Hierarchical clustering of expression profiles of PmCYCDs in different organs.PmCYCD genes in five organs were collected from transcriptome data and the Log2 value of RPKM (RPKM> 0) was converted.

Figure 4 .
Figure 4. Hierarchical clustering of expression profiles of PmCYCDs in different organs.PmCYCD genes in five organs were collected from transcriptome data and the Log2 value of RPKM (RPKM> 0) was converted.

Figure 5 .
Figure 5. Expression analysis of 12 PmCYCD genes under different tissues by qRT-PCR.The relative quantification method (2 −ΔΔCT ) was used to evaluate quantitative variation under different tissues.Error bars represent standard error for three replicates.Different lowercase letters in column differed significantly based on Tukeyʹs test (p < 0.05) after one-way ANOVA.F, flower; YB, young buds; Sh, shoot; S, stem; YL, young leaf; R, root.

Figure 5 .
Figure 5. Expression analysis of 12 PmCYCD genes under different tissues by qRT-PCR.The relative quantification method (2 −∆∆CT ) was used to evaluate quantitative variation under different tissues.Error bars represent standard error for three replicates.Different lowercase letters in column differed significantly based on Tukey's test (p < 0.05) after one-way ANOVA.F, flower; YB, young buds; Sh, shoot; S, stem; YL, young leaf; R, root.

Figure 6 .
Figure 6.The expression levels of PmCYCDs genes were analyzed in roots under different treatments including NAA, 6-BA, GA3, ABA, and sucrose.

Figure 6 .
Figure 6.The expression levels of PmCYCDs genes were analyzed in roots under different treatments including NAA, 6-BA, GA3, ABA, and sucrose.