Identification of the CYCD and CDK gene families in Populus tomentosa

To identify CYCD and CDK genes in P. tomentosa, hidden Markov models (HMMs) and Blastp were used to query the whole genome. After the elimination of redundant sequences and examination of domains, we finally identified 43 CYCD and 27 CDK family members (Table S1). A phylogenetic tree was constructed with 24 CYCD and 18 CDK in Populus trichocarpa (PtrCYCD and PtrCDK) to classify and name the successfully identified members12 (Fig.1). Results showed that 47 PotomCYCDs were divided into six subclasses, of which 12, 4, 10, 6, 9 and 2 members belonged to D1, D2/4, D3, D5, D6 and D7 subclasses, respectively. A total of 27 PotomCDKs were divided into seven subclasses, which belonged to 2 members of the CDKA subclass, 1 member of the CDKB subclass, 5 members of the CDKC subclass, 4 members of the CDKD subclass, 4 members of the CDKE subclass, 2 members of the CDKF subclass, and 9 members of the CDKG subclass. The identified members of the P. tomentosa gene family were named in accordance with the gene family members of P. trichocarpa. Considering that P. tomentosa is an allodiploid, a was added to the name of the gene searched from subgenome A (PtA), and b was added to the name of the gene searched from subgenome D (PtD). In phylogenetic tree analysis, most of genes had their alleles with close relationships. However, we found PotomCYCD6;3a was closer to the orthologs gene PtrCYCD6;3 rather than its allele PotomCYCD6;3b. This phenomenon was also found between PotomCYCD3;5, PotomCYCD5;1, PotomCYCD7;1, PotomCDKA;1, PotomCDKE;1 and PotomCDKF;1 and there corresponding orthologs, which revealed that these alleles differentiated during the evolutionary process. An interesting finding was that a small number genes PotomCYCD6;2b, PotomCDKG;2a, PotomCDKC;2a, PotomCDKC3;b, PotomCDKC4;b and PotomCDKB1;1a were lack of their alleles. We speculated that it was the chromosomal variation and transposon insertion that resulted in the loss of these alleles. All these evidences indicated that PotomCYCDs and PotomCDKs were both conserved and differentiated.

Phylogenetic tree analysis of D-type cyclin (CYCD) and cyclin-dependent kinases (CDK) gene family in P. tomentosa and P. trichocarpa. A PotomCYCA gene from P. tomentosa genome was selected as an outgroup gene. (a) Phylogenetic tree analysis of 43 PotomCYCDs and 22 PtrCYCDs. All CYCDs were classified into six distinct groups on the basis of the subfamily of P. trichocarpa CYCDs (from D1 to D7) and were distinguished by different colours. (b) Phylogenetic tree analysis of 27 PotomCDKs and 18 PtrCDKs. All CDKs were classified into seven distinct groups on the basis of the subfamily of P. trichocarpa CDKs (from CDKA to CDKG) and were distinguished by different colours.

To analyse the sequence differences of alleles from different subgenome, we analysed the protein sequence identity of all members of the two gene families. Results showed that in the PotomCYCD gene family, PotomCYCD1;3a and PotomCYCD1;3b (99.0%) had the highest sequence similarity, and PotomCYCD1;2a and PotomCYCD1;2b (80.6%) had the lowest sequence similarity (Table S2). In the PotomCDK gene family, the highest sequence similarity was PotomCDKE;2a and PotomCDKE;2b (99.5%), and the lowest sequence similarity was PotomCDKA;1a and PotomCDKA;1b (82.4%, Table S3). At the same time, we also investigated the basic characteristics of the two family members, such as their AA length, isoelectric point (PI), molecular weight (MW), and subcellular localization (Table S4). Results showed that the AA length of PotomCYCD gene family varied from 256 to 408 AA. The largest protein was PotomCYCD2;1b (45.24kDa), and the smallest protein was PotomCYCD6;2b (29.51kDa). The PI of most CYCD proteins varied around 57. Subcellular localisation prediction results indicated that all PotomCYCD proteins were located in the nucleus. The difference was that the basic characteristics of different subclasses of proteins in the PotomCDK gene family varied remarkably, but the characteristics of different members of the same subclass were relatively similar. The CDKG subclass (except PotomCDKG;2a) had the largest protein (78.6589.63kDa), and the CDKA and CDKB subclasses had the smallest protein (33.7936.53kDa). Although PotomCDKG;2a is only 117aa in length, it contains part of the conserved domains required by CDK, and it is speculated that it might not be functional due to its short length or mutated during evolution. The protein PIs of different CDK subclasses different but were very similar in the same subclass. Subcellular localisation prediction results showed that all PotomCDK proteins were located in the nucleus and that PotomCDKA;1b might also be located in the cytoplasm. PotomCDKG;4b might also be located in the cell membrane and cytoplasm. PotomCDKG;5a might also be located in the cell membrane.

Gene structural diversity and conserved motif divergence are possible mechanisms for the evolution of multigene families43. To further study the gene and protein structure of the CYCD and CDK gene family, we analysed the number and distribution of exons. The results of gene structure analysis showed that the structures of PotomCYCD genes were similar and that the number of exons varied from 4 to 7. The number of exons in the D3 subclass was 4, and the number of exons in other subclasses except PotomCYCD2;1b and PotomCYCD2;2a was 5 or 6 (Fig.2a). The results of the gene structure analysis of PotomCDK genes showed that the number of exons in CDK ranged from 1 to 13. The gene structure amongst members of the same subclass was relatively conserved. For example, none of the four members of CDKE were introns. Amongst the nine members of CDKG, eight members consisted of 1 long exon and 5/6 short exons. A high number of exons were found in CDKA and CDKC, whereas only 3 exons were observed in the CDKF subclass (Fig.2c).

Gene structure and conserved motif compositions of PotomCYCDs and PotomCDKs. (a) Exon/intron structures of PotomCYCDs. (b) Architecture of conserved protein motifs of PotomCYCDs. (c) Exon/intron structures of PotomCDKs. Yellow boxes and black lines indicate exons and introns, respectively, at each CYCD and CDK gene. (d) Architectures of conserved protein motifs of PotomCDKs. These coloured boxes indicate distinct motifs and their corresponding positions in each CYCD and CDK protein sequence. The detailed characteristics of each motif are shown in Table S5.

To elucidate the distribution of the motifs in CYCD and CDK proteins and their function, ten types of motifs and their distributions of CYCDs and CDKs were predicted using the MEME program (Fig.2). Our results indicated that CYCDs contained similar motif types. However, some differences existed amongst different subclasses. Motifs 15 were contained in all subclasses. Motif 6 was unique to D1 and D2/4 subclasses. Motifs 7 and 8 were contained in most subclasses except D5 and D7. Motif 9 was only D7 subclass Class missing. Motif 10 was included in most subclasses except D1 and D2/4 (Fig.2b). In the CDK gene family, motif 1, 2 and 68 were included in all subclasses. Motif 3 was included in most subclasses except CDKA and CDKB. Motif 4 was not found in CDKA subclasses. Motif 5 was not found in CDKF. Motif 9 was only found in CDKE and CDKG subclasses. Motif 10 was unique to CDKG (Fig.2d). Some interesting findings appeared in some alleles. The PotomCYCD5;1a has one more exon than its allele PotomCYCD5;1b and motif 3 presented in the former but not the latter. PotomCYCD6;2b without allele was lack of motif 7 when comparing to other PotomCYCD6. PotomCDKB1;1a, the only one member in PotomCDKB, had almost the same motifs but an extra motif 4 than the PotomCDKA;1. PotomCDKC;4b was lack of the motif 3 and motif 5 when comparing to other PotomCDKC. These findings revealed that some special alleles might have different gene structures and conserved motifs leading to different functions.

At the same time, we analysed the domains and conserved motifs of CYCDs. LxCxE is a key motif for CYCD binding to RBR33. The PEST sequence, a region full of P(Pro), E(Glu), S(Ser) and T(Thr), might result in itself degradation and were often found in D-type cyclins32,33,34. Results showed that all PotomCYCD proteins had Cyclin_N and Cyclin_C domains. No LxCxE motif was observed in the CYCD6 subclass, and the LxCxE motif existed in other subclass proteins. Most CYCD proteins had the PEST motif, but the position of the PEST motif was not fixed (Fig. S1). Multiple sequence alignment with P. trichocarpa CDK gene family proteins showed that their conserved domains were highly consistent with their orthologous proteins in P. trichocarpa (File S1). Amongst them, CDKA had PSTAIRE, CDKB had PPTALRE or PPTTLRE, CDKC had PITAIRE, CDKE had SPTAIRE, and CDKG had PLTSLRE9. However, PotomCDKC;3b and PotomCDKC;4b were not observed with this characteristic motif.

Specific cis-element motifs can be recognised by transcription factors and participate in gene expression regulation. To further study the potential regulatory mechanisms of PotomCYCDs and PotomCDKs in a diversified biological process, particularly in plant hormones and specific expression, 2.0kb upstream sequences from the translation start sites of CYCD and CDK genes were submitted to the PlantCARE database to detect cis-elements. Results showed that the types and numbers of various cis-acting elements of genes in PotomCYCDs and PotomCDKs were similar, thereby implying their functional relatedness. Multiple hormone-responsive elements, such as ABA responsive (DRE1, ABRE, ABRE2, ABRE3a, ABRE4, TCA-element), auxin and/or salicylic acid activation (as-1), auxin responsive (TGA-element), ethylene responsive (ERE), gibberellin responsive (GARE-motif, P-box, TATC-box) and MeJA responsive (CGTCA-motif, TGACG-motif) elements, were found in cis-acting elements in two gene families. Cis-acting element prediction results showed that the two gene families had similar responses to hormones and had the most responsive elements responsive to ABA followed by ethylene responsive. The numbers and positions of various hormone responsive elements on the promoters of each gene are shown in the Fig. S2S3. In the CYCD gene family, the numbers of ABA responsive and ethylene responsive elements in the D7 subclass with only 2 members were 13 and 10, respectively, but no gibberellin responsive and auxin responsive elements were found in the D7 subclass, which might suggest that the D7 The subclass predominantly responded to ABA and ethylene. The number of MeJA responsive elements in the D5 subclass with 6 members was 18, which was the largest amongst all subclasses. This result might suggest that the D5 subclass predominantly responded to MeJA. In the D6 subclass with only 9 members, 16 Gibberellin responsive elements were found, accounting for the largest proportion and suggesting that the D6 subclass predominantly responded to gibberellin. The auxin responsive element was found in all D1D5 subclasses but not in D6 and D7 subclasses (Fig. S2). In the CDK gene family, the number of ABA and ethylene responsive elements in CDKD subclasses with only 4 members was as high as 20 and 17, but gibberellin responsive and auxin responsive elements were not found in CDKD subclasses, which might imply that CDKD subclasses responded to ABA and ethylene. The numbers of MeJA responsive, gibberellin responsive and auxin responsive elements accounted for 8, 3 and 2, respectively, in CDKF subclasses with only 2 members and had the largest proportion. Auxin responsive element was not found in CDKA, CDKB and CDKD subclasses (Fig. S3).

A number of specific expression elements was also found in cis-acting elements in the two gene families, and the proportions of specific expression elements in the two gene family members were also similar. Pollen specific activation elements were the most numerous, with 49 and 41 in CYCD and CDK families, respectively. In addition, two families also included some different cis-acting elements, such as the seed specific regulation element (RY-element) that only existed in the CDK gene family and the cell cycle regulation element (MSA- like) that only existed in the CYCD gene family (Fig. S2S3).

On the basis of the information from the P. tomentosa genomic database, we determined the chromosomal distributions of CYCD and CDK genes (Table S4). Results suggested that CYCD genes were distributed on 26 chromosomes, whereas CDK genes were mapped onto 21 chromosomes (Fig.3). Although 35 chromosomes contained CYCD or CDK genes, the overall distribution was mostly nonuniform. Chromosomes 2A, 2D, 14A and 14D all contained three CYCD genes, whereas only one CYCD gene was distributed on chromosome 4A. Chromosomes 12A and 12D both contained 3 members of CDK genes, whereas chromosome 1A contained only one CDK gene. Interestingly, chromosomes 16D, 17A and 17D did not contain any CYCD or CDK gene.

Circos figure for chromosome distribution with synteny links. Grey and colourful lines represent synteny blocks and duplicated CYCD and CDK gene pairs, respectively, in P. tomentosa. The gene ids in red font are members of the PotomCYCD gene family, and the gene ids in blue font are members of the PotomCDK gene family.

We constructed a synteny analysis between CYCD and CDK genes in P. tomentosa (Table S6). The synteny blocks and the duplicated CYCD and CDK gene pairs were showen by the grey and colourful lines (Fig.3). All CYCD genes and 24 of 27 CDK genes were identified as collinear genes involving WGD or segmental duplication, whereas 2 CDK genes (i.e. PotomCDKC;4b and PotomCDKG;2a) were considered as dispersed genes, and 1 CDK gene (i.e. PotomCDKB1;1a) was considered as a singleton gene (Table S6). Remarkably, some CYCD and CDK genes were associated with at least five syntenic gene pairs, which implied that these genes might be involved in some critical roles during the evolutionary process. Interestingly, tandem duplication events were not found between CYCD and CDK genes in P. tomentosa. Evidence suggested that WGD or segmental duplication led the expansion of CYCD and CDK genes in P. tomentosa.

To further explore the evolutionary relationships of the CYCD and CDK gene families, we performed syntenic analyses amongst Arabidopsis, P. trichocarpa and P. tomentosa (Table S7). Between Arabidopsis and P. trichocarpa, 7 CYCD and 11 CDK gene pairs were found. A total of 81 CYCD and 38 CDK gene pairs were identified between P. trichocarpa and P. tomentosa (Fig.4). We found that 4 of 7 CYCD gene pairs and 9 of 11 CDK gene pairs between Arabidopsis and P. trichocarpa, respectively, were distributed on chromosome 1 and 4 in Arabidopsis. In P. trichocarpa, CYCD gene pairs were predominantly distributed on chromosomes 1, 2, 7, 9, 14 and 19, whereas CDK gene pairs were predominantly located on chromosomes 1, 3, 8, 10 and 12. The situations of CYCD and CDK gene pair distributions on chromosomes in P. tomentosa were similar to P. trichocarpa. Between P. trichocarpa and P. tomentosa, although abundant gene pairs were identified, 1 CYCD (i.e. PotomCYCD6;3b) and 2 CDK (i.e. PotomCDKC;4b and PotomCDKG;2a) genes did not find any syntenic gene.

Synteny of CYCD and CDK genes in Arabidopsis, P. trichocarpa and P. tomentosa. Red and blue lines represent duplicated CYCD and CDK gene pairs, respectively.

To investigate the possible roles of the PotomCYCDs and PotomCDKs, the expression levels of 43 CYCD and 27 CDK genes were determined by transcriptome results in three various tissues, i.e. leaf, stem and root41. Our results indicated that CYCD and CDK genes showed similar expression profiles in different tissues. Most genes were expressed in all three tissues, and alleles from different subgenomes had similar or even the same expression pattern (Figs. S4S5). In CYCD and CDK gene families, most genes showed this expression trend, with the highest expression in stems, followed by roots and leaves. In the CYCD gene family, most D3 subclass gene expression levels conformed to this trend. However, CYCD3;1a, CYCD3;2a and CYCD3;2b gene expression levels were highest in roots, and the expression of CYCD3;2b in leaves was highest amongst all genes. In terms of the overall expression levels of genes in different subclasses, D3 subclass genes had the highest expression, followed by the D1 and D2 subclasses, whereas the D7 subclass was not expressed in all tissues. In the CDK gene family, the expression trends of CDKA and CDKB subclass genes were the same as those mentioned before. The genes of other subclasses are also highly expressed in leaves and roots, such as CDKC;1a/b and CDKG;1a/b had the highest expression in leaves and CDKC;2a/b and CDKG;4a/b had the highest expression in roots. In terms of the overall expression levels of genes in different subclasses, the CDKA subclass had the highest gene expression followed by the CDKG subclass. These results suggested that CYCD and CDK genes had similar expression patterns, implying their functional relevance, and the overall expression level of CYCD genes was lower than that of CDK genes. These results might also indicate that alleles with the same function as in an allodiploid species co-regulated the growth and development of the organism.

In order to analyze the interaction between different CYCD and CDK proteins, the STRING website was used to predict the interaction between different CYCD subclasses and CDK proteins. First, all PotomCYCD and PotomCDK genes were compared with the Arabidopsis database of the STRING website. The comparison results and annotation information are shown in Table S8. The interaction between different subclasses of CYCD and CDK was predicted and the line thickness indicates the strength of data support (Fig.5). Results showed that CDKA (CDC2) was at the core of the interaction relationship. The protein interaction prediction results showed that the D1 subclass could interact with CDKA, CDKD1;1 and CDKE;1 and had the strongest interaction with CDKA. The D2/4 subclass only could have a strong interaction with CDKA. In the D3 subclass, the proteins CYCD3;1 and CYCD3;3 were obtained from the alignment and could interact with CDC2 and CDKE;1, and the interaction with CDKA was stronger. The D5 subclass could interact with CDKA, CDKB1;2, CDKD1;1, CDKD1;3, CDKE;1 and CAK1AT (CDKF). The D6 subclass could interact with CDKA, CDKB1;2, CDKD1;1 and CDKD1;3 has an interaction relationship. The D7 subclass only had a weak interaction relationship with CDKA (Fig.5). These results suggested that different subclasses of CYCD proteins might differ in gene sequence and protein properties and in their interaction with CDKs.

Prediction of the interaction between different subfamilies of PotomCYCDs and PotomCDKs gene family proteins by using the STRING website. (a) Interaction between D1 subfamily and CDKs. (b) Interaction between D2/4 subfamily and CDKs. (c) Interaction between D3 subfamily and CDKs. (d) Interaction between D5 subclass and CDKs. (e) Interaction between D6 subfamily and CDKs. (f) Interaction between D7 subfamily and CDKs. Line thickness indicates the strength of data support.

Previous in vitro yeast two-hybrid (Y2H) experiments and molecular docking experiments showed that PtoCYCD3;3 protein interacts with 12 PtoCDK proteins, of which the strongest interaction is PtoCDKE;212. To verify the reliability of the in vitro Y2H results, in vivo validation in plants was performed using the Bimolecular Fluorescent Complimentary (BIFC) assay. The 12 PtoCDKs screened by Y2H and PtoCYCD3;3 were fused to the N- and C-termini of YFP to construct fusion vectors (YFPNPtoCDKs and YFPCPtoCYCD3;3), and transient infection mediated by Agrobacterium in two fusion proteins were co-expressed in the lower epidermal cells of tobacco. If the two proteins interacted, the two fragments of the fluorescent protein will be close to each other in space, complementary to each other and reconstructed into an active and complete fluorescent protein molecule, thereby generating fluorescence. At the same time, GUS was fused to the N-terminus of YFP (YFPNGUS) and co-expressed with YFPCPtoCYCD3;3 as a negative control. The 12 PtoCDKs (PtoCDKA;1, PtoCDKB1;1, PtoCDKB1;2, PtoCDKB2;1, PtoCDKB2;2, PtoCDKC;2, PtoCDKD;1, PtoCDKD;2, PtoCDKE;2, PtoCDKF;1, PtoCDKG;3 and PtoCDKG;4) and PtoCYCD3;3, showed fluorescence in the nuclei of tobacco epidermal cells (Fig.6). The results indicated that PtoCYCD3;3 also interacted with these 12 PtoCDKs proteins in plants.

BiFC validation in tobacco epidermis. PtoCYCD3;3 connects pSPYCE(MR) vector and PtoCDKs connects pSPYNE(R)173 vector. After co-expression in tobacco leaf epidermal cells, the fluorescence signal was observed under a laser confocal microscope. YFPC-PtoCYCD3;3+YFPN-GUS were used as negative control.YFP: yellow fluorescent signal; NLS-mCherry: nuclear signal localisation label; Bright field: brightfield vision; Merge: superposition of fluorescence signals, bar=50m.

Our previous research found that transgenic PtoCYCD2;1 and transgenic PtoCYCD3;3 poplars have completely opposite phenotypes. Transgenic PtoCYCD2;1 plants have reduced plant height, curled leaves and thin stems42, whereas transgenic PtoCYCD3;3 poplar plants have increased height, curled leaves, thickened stem and branched in advance12. Cloned PtoCYCD2;1 and PtoCYCD3;3 genes were combined with the CYCD gene families of P. tomentosa and P. trichocarpa to construct a phylogenetic tree. Results showed that PtoCYCD2;1 and PtoCYCD3;3 belonged to the D2 and D3 subclasses, respectively, and closely related to the PotomCYCD gene (Fig. S6).

In order to find out the interaction between PtoCYCD2;1 and PtoCDKs, in vitro Y2H and in vivo BIFC experiments were used to detect the interaction. Y2H vectors (pGBKT7PtoCYCD2;1 and pGADT7PtoCDKs) were constructed and co-transformed into yeast AH109 competent cells, and the successfully identified positive yeast strains were spread on different AA-deficient media for Y2H experiment. As a competitive inhibitor of HIS3, 3-AT can inhibit the expression of HIS3 to a certain extent by adding this substance to the culture medium. The growth rates on SD-Trp-Leu-His+10/20mM 3-AT and SD-Trp-Leu-His-Ade culture media were observed to detect the interaction strength between PtoCYCD2;1 and different PtoCDKs proteins. During the 6day observation period, PtoCDKD;3, PtoCDKF;1 and PtoCDKG;5 could grow on SD-Trp-Leu, SD-Trp-Leu-His+10/20mM 3-AT and SD-Trp-Leu-His-Ade culture media, suggesting the strongest interaction. PtoCDKA;1, PtoCDKB1;1, PtoCDKB2;1, PtoCDKB2;2, PtoCDKD;2, PtoCDKE;1, PtoCDKE;2 and PtoCDKG;1 could grow on SD-Trp-Leu, SD-Trp-Leu-His+10/20mM 3-AT culture media, suggesting a strong interaction. PtoCDKC;1, PtoCDKD;1, PtoCDKG;3 and PtoCDKG;4 could only grow on SD-Trp-Leu culture medium, but not on SD-Trp-Leu-His+10/20mM 3-AT and SD-Trp-Leu-His-Ade culture media, indicating that no direct interaction between these proteins (Fig.7). The growths on the SD-Trp-Leu-His+10/20mM 3-AT and SD-Trp-Leu-His-Ade culture media for 3 and 6days were photographed and observed, and the interaction strength of the PtoCDKs gene family members with PtoCYCD2;1 was observed to have the following relationships: PtoCDKD;3=PtoCDKF;1=PtoCDKG;5>PtoCDKA;1>PtoCDKG;1>PtoCDKE;2>PtoCDKE;1>PtoCDKB2;1>PtoCDKB2;2>PtoCDKD;2>PtoCDKB1;1.

Proteinprotein interactions of PtoCYCD2;1 with PtoCDKs. Yeast cells were co-transformed with pGBKT7 and pGADT7 constructs carrying the corresponding genes and grown on SD-Trp-Leu, SD-Leu-Trp-His+10mM 3-AT, SD-Leu-Trp-His+20mM 3-AT, SD-Leu-Trp-His-Ade. AD+BD, AD+BD-PtoCYCD2;1 and AD-PtoCDKs+BD, were used as negative control. Day represents the number of days of growth on the corresponding medium. The triangle represents dilution in a 0.1-fold gradient (1,0.1 and 0.01).

To verify the interaction between PtoCYCD2;1 and PtoCDKD;3, PtoCDKF;1, PtoCDKG;5 and the fastest-growing PtoCDKA;1 on SD-Trp-Leu-His+10/20mM 3-AT culture medium in plants, we successfully constructed BIFC vectors (YFPNPtoCDKs and YFPCPtoCYCD2;1), which were transiently expressed in the lower epidermis of tobacco through Agrobacterium-mediated transient infection. Results showed that only PtoCDKA;1 and PtoCYCD2;1 showed fluorescence in the nuclei of tobacco cells and that PtoCDKD;3, PtoCDKF;1, PtoCDKG;5 and PtoCYCD2;1 co-expressed no fluorescence signal (Fig.8). These results indicated that PtoCYCD2;1 only interacted with PtoCDKA;1 in plants.

BiFC validation in tobacco epidermis. PtoCYCD2;1 connect pSPYCE(MR) vector and PtoCDKs connect pSPYNE(R)173 vector. After co-expression in tobacco leaf epidermal cells, the fluorescence signal was observed under laser confocal microscopy. YFP: yellow fluorescent signal; NLS-mCherry: nuclear signal localisation label; Bright field: bright field vision; Merge: superposition of fluorescence signals. Scale bars forYFPC-PtoCYCD2;1+YFPN-PtoCDKD;3, 50m; for others, 25m.

Continued here:
Study on the interaction preference between CYCD subclass and CDK family members at the poplar genome level | Scientific Reports - Nature.com

Comments are closed.