Morphological, Transcriptome, and Hormone Analysis of Dwarfism in Tetraploids of Populus alba × P. glandulosa

Breeding for dwarfism is an important approach to improve lodging resistance. Here, we performed comparative analysis of the phenotype, transcriptome, and hormone contents between diploids and tetraploids of poplar 84K (Populus alba × P. glandulosa). Compared with diploids, the indole-3-acetic acid (IAA) and gibberellin (GA3) contents were increased, whereas the jasmonic acid (JA) and abscisic acid (ABA) contents were decreased in tetraploids. RNA-sequencing revealed that differentially expressed genes (DEGs) in leaves of tetraploids were mainly involved in plant hormone pathways. Most DEGs associated with IAA and GA promotion of plant growth and development were downregulated, whereas most DEGs associated with ABA and JA promotion of plant senescence were upregulated. Weighted gene co-expression network analysis indicated that certain transcription factors may be involved in the regulation of genes involved in plant hormone pathways. Thus, the altered expression of some genes in the plant hormone pathways may lead to a reduction in IAA and GA contents, as well as an elevation in ABA and JA contents, resulting in the dwarfing of tetraploids. The results show that polyploidization is a complex biological process affected by multiple plant hormone signals, and it provides a foundation for further exploration of the mechanism of tetraploids dwarfing in forest trees.

Recent studies have extensively examined the effects of polyploidization on plant development at the transcriptome level. With a change in ploidy, gene expression may be

Comparative Analysis of Phenotypic and Anatomical Characteristics between Tetraploids and Diploids
To understand the growth patterns of diploid and tetraploid poplar 84K at different stages of development, the plant height at 1, 3, and 5 months after transplantation was measured. Compared with the diploids, the tetraploids grew slowly and showed dwarfism. The average height of the tetraploids was 60.73 ± 3.65 cm at 3 months, whereas the average height of the diploids was 87.30 ± 3.94 cm (Table S1). The height of the tetraploids was approximately 30.44% lower than that of the diploids ( Figure 1A,B).
To observe anatomical differences between the diploids and tetraploids, paraffinembedded sections of stem internodes and leaves were examined. In general, in longitudinal sections of the stem internodes, the cells of the tetraploids were shorter and wider than those of the diploids to some extent, whereas the cell area of the leaf in transverse section was larger in the tetraploids. The average length and width of the stem internode cells of diploids were 70.47 ± 5.20 µm and 26.54 ± 4.18 µm, whereas those of the tetraploids stem internode cells were 54.82 ± 11.61 µm and 45.63 ± 4.94 µm, respectively ( Figure 1C).

Transcriptome Sequencing Data Analysis
To analyze the gene expression differences between triploid and diploid leaves of poplar 84K during continuous growth, development, and senescence, the first, fifth, and 11th leaves (numbered from the shoot tip) were selected as experimental materials, which were labeled as D1, D5, D11, T1, T5, and T11, respectively. Three biological replicates were analyzed for each leaf node sampled, and the genome sequence for poplar 84K was selected as the reference genome. The transcriptome of the 18 leaves was compared and analyzed. A total of 137 G of valid data were obtained from the 18 cDNA libraries. The Q20 and Q30 quality scores attained at least 98.38%, and the overall GC content was more than 44% (Table S2). On the basis of the comparison of the sequencing reads and reference genome, more than 80.37% mapped reads could be found in all libraries (Table S3). These results showed that the selection of the reference genome was appropriate and the sequencing results were acceptable; thus, they were suitable for subsequent analysis.

Transcriptome Sequencing Data Analysis
To analyze the gene expression differences between triploid and diploid leaves of poplar 84K during continuous growth, development, and senescence, the first, fifth, and 11th leaves (numbered from the shoot tip) were selected as experimental materials, which were labeled as D1, D5, D11, T1, T5, and T11, respectively. Three biological replicates were analyzed for each leaf node sampled, and the genome sequence for poplar 84K was selected as the reference genome. The transcriptome of the 18 leaves was compared and analyzed. A total of 137 G of valid data were obtained from the 18 cDNA libraries. The Q20 and Q30 quality scores attained at least 98.38%, and the overall GC content was more than 44% (Table S2). On the basis of the comparison of the sequencing reads and reference genome, more than 80.37% mapped reads could be found in all libraries (Table S3). These results showed that the selection of the reference genome was appropriate and the sequencing results were acceptable; thus, they were suitable for subsequent analysis.
Using fold change (FC) ≥ 2 or FC ≤ 0.5 (log 2 FC ≥ 1) and p-value < 0.05 as the screening threshold, 7917 differentially expressed genes (DEGs) were identified in the leaves at the three nodes. In total, 1622 DEGs were detected in the first leaf, comprising 1080 upregulated genes (66.6%) and 542 downregulated genes (33.4%), 3747 DEGs were detected in the fifth leaf, consisting of 1840 upregulated genes (49.1%) and 1907 downregulated genes (50.9%), and 2548 DEGs were identified in the 11th leaf, including 1417 upregulated genes (55.6%) and 1131 downregulated genes (44.4%) (Figure 2A). A Venn diagram showed that 447 DEGs (7.3%) were differentially expressed between diploids and tetraploids at all three leaf nodes ( Figure 2B). A cluster analysis heatmap clearly divided the 18 samples into two groups based on the expression level of DEGs, which separated the diploids and tetraploids ( Figure 2C). These results indicated that the biological replicates of each sample were highly similar, and that the ploidy and leaf positions may lead to differences in gene expression patterns.

Functional Enrichment Analysis of DEGs in the Transcriptomes
To explore the functions of the DEGs, GO annotation and KEGG pathw ment analysis were performed on the DEGs of tetraploids and diploids. Al DEGs were classified into three functional categories: biological process, ce ponent, and molecular function. The most highly enriched GO terms are sho ure 3A-C (p < 0.05). Among the 20 most enriched GO terms in the biological p egory, 2-187 DEGs in the first leaf were involved in 'hormone transport', 's thesis', and 'response to auxin, brassinolide, JA, and other hormones' (Figure fifth leaf, 4-267 DEGs were involved in the 'signaling pathways of cytokinin, and brassinolide', 'starch and sugar degradation', and 'redox reaction' (Figure  11th leaf, 5-262 DEGs were involved in the 'stress response to ABA, ethylene signaling pathways', 'photosynthesis', 'starch and sugar metabolism', 'cell and 'leaf senescence' ( Figure 3F). These processes may be crucial for the varia raploid traits.
To further examine the potential biological processes affected by polyplo KEGG pathway enrichment analysis was conducted on the DEGs. All signi

Functional Enrichment Analysis of DEGs in the Transcriptomes
To explore the functions of the DEGs, GO annotation and KEGG pathway enrichment analysis were performed on the DEGs of tetraploids and diploids. All annotated DEGs were classified into three functional categories: biological process, cellular component, and molecular function. The most highly enriched GO terms are shown in Figure 3A-C (p < 0.05). Among the 20 most enriched GO terms in the biological process category, 2-187 DEGs in the first leaf were involved in 'hormone transport', 'sucrose synthesis', and 'response to auxin, brassinolide, JA, and other hormones' ( Figure 3D). In the fifth leaf, 4-267 DEGs were involved in the 'signaling pathways of cytokinin, gibberellin, and brassinolide', 'starch and sugar degradation', and 'redox reaction' ( Figure 3E). In the 11th leaf, 5-262 DEGs were involved in the 'stress response to ABA, ethylene, and auxin signaling pathways', 'photosynthesis', 'starch and sugar metabolism', 'cell apoptosis', and 'leaf senescence' ( Figure 3F). These processes may be crucial for the variation in tetraploid traits.
To further examine the potential biological processes affected by polyploidization, a KEGG pathway enrichment analysis was conducted on the DEGs. All significantly enriched KEGG pathways are shown in Figure 3G-I (p < 0.05). The DEGs were significantly enriched in the 'flavonoid biosynthesis', 'plant hormone signal transduction', and 'circadian rhythm' pathways in the first leaf ( Figure 3G). In the fifth leaf, the DEGs were mainly enriched in the 'galactose metabolism', 'starch and sucrose degradation', and 'plant hormone signal transduction' pathways ( Figure 3H). In the 11th leaf, the DEGs were mainly significantly enriched in the 'plant hormone signal transduction', 'MAPK signal', 'plant-pathogen interaction', and 'photosynthesis' pathways ( Figure 3I). It was notable that, in the GO and KEGG enrichment analyses, a large number of DEGs in the leaves at all three nodes were significantly enriched in the process of plant hormone synthesis and signal transduction.

Determination of Endogenous Hormones and Expression of Hormone-Signaling Genes in Diploids and Tetraploids
To examine the relationship between the phenotype and hormones after polyploidization, we measured the contents of endogenous hormones (auxin, gibberellin, JA, and ABA) in leaves of diploids and tetraploids at three nodes, and the hormone signal-related genes that encode receptors and response factors were identified by KEGG pathway enrichment analysis.
Auxin (indole-3-acetic acid, IAA) and gibberellin (GA 3 ) can promote plant growth and internode elongation [25]. In the present study, the IAA and GA 3 contents of tetraploids were significantly lower than those of diploids ( Figure 4A). In the auxin signaling pathway, two auxin-responsive SMALL AUXIN-UP RNA 36 (SAUR36) genes (Pop_G03G078252 and Pop_A03G050447), which promote leaf senescence, were upregulated at the three leaf nodes of tetraploids. The AUX/IAA transcriptional regulator family of proteins inhibits auxin signal transduction. Compared with the diploids, one AUX/IAA gene was upregulated in the fifth leaf, and five AUX/IAA genes were upregulated in the 11th leaf in the tetraploids. GH3 proteins are negative regulators in response to auxin. In tetraploids, most GH3 genes were upregulated in the first and 11th leaves ( Figure 4B; Table S4). Moreover, in the gibberellin signaling pathway, six genes (Pop_A01G017618, Pop_A06G062163, Pop_A09G014464, Pop_G06G082035, Pop_G09G014084, and Pop_G16G025336) encoding DELLA protein were upregulated in all leaves. Two gibberellin receptor GID1 genes (Pop_A14G000831 and Pop_G14G045139) were downregulated in the first and fifth leaves of the tetraploids ( Figure 4B; Table S4). These results suggested that the upregulation of most auxin signaling-related DEGs and a large number of DELLA proteins may contribute to the decrease in auxin and gibberellin contents in leaves of tetraploids.
MYC2 gene (Pop_A14G044750) was downregulated in the fifth leaf ( Figure 4B; Table S4). In addition, in the ABA signaling pathway, the ABA response factors ABF (Pop_A09G026716 and Pop_G09G077307), bZIP (Pop_G09G011691), and SnPK2 (Pop_A02G088618 and Pop_G02G065896) were upregulated in tetraploids. The ABAencoding receptors PYL4 (Pop_A06G089383, Pop_A16G090069, Pop_G06G051642, and Pop_G16G068642) and PYL6 (Pop_A10G069584 and Pop_G10G048195) were not significantly differentially expressed in the first leaf, but two PYL4 genes (Pop_A16G090069 and Pop_G06G051642) were downregulated in the fifth leaf, and all were downregulated in the 11th leaf ( Figure 4B; Table S4).  JA and ABA inhibit growth and development, as well as promote plant senescence [26]. Compared with diploids, the JA and ABA contents of tetraploids were significantly increased ( Figure 4A). JASMONATE ZIM-DOMAIN (JAZ) protein is an inhibitor of the JA signaling pathway. Here, four genes that encode JASMONATE ZIM-DOMAIN (JAZ) proteins were differentially expressed. Among them, one JAZ gene (Pop_G01G035869) was downregulated in the first leaf of tetraploids compared with that of diploids. Three JAZ genes (Pop_A08G086419, Pop_G08G046460, and Pop_G15G074364) were downregulated in the fifth leaf. MYC2 can regulate the biosynthesis of sesquiterpenoids and is a core regulator of the JA signal transduction pathway. In tetraploids, two MYC2 genes (Pop_A09G015372 and Pop_G01G089277) were upregulated in the three leaves, and one MYC2 gene (Pop_A14G044750) was downregulated in the fifth leaf ( Figure 4B; Table S4). In addition, in the ABA signaling pathway, the ABA response factors ABF (Pop_A09G026716 and Pop_G09G077307), bZIP (Pop_G09G011691), and SnPK2 (Pop_A02G088618 and Pop_G02G065896) were upregulated in tetraploids. The ABAencoding receptors PYL4 (Pop_A06G089383, Pop_A16G090069, Pop_G06G051642, and Pop_G16G068642) and PYL6 (Pop_A10G069584 and Pop_G10G048195) were not signifi-cantly differentially expressed in the first leaf, but two PYL4 genes (Pop_A16G090069 and Pop_G06G051642) were downregulated in the fifth leaf, and all were down-regulated in the 11th leaf ( Figure 4B; Table S4).

Co-Expression Modules Associated with Hormone Content
The WGCNA method was used to identify co-expressed gene modules associated with phenotypic traits. A module is a cluster of genes whose expression is highly correlated. Genes in the same module have extremely high correlation coefficients with each other [27]. In the present study, the transcriptome sequencing data and four hormone indicators (GA, IAA, JA, and ABA) were used to construct the co-expression network. The genes were divided into 32 gene modules, each represented by a different color in the network ( Figure 5A). The correlations between individual gene modules of each phenotype are shown in Figure 5B. The largest module (turquoise) contained 14,072 genes, whereas the smallest module (midnight blue) contained 36 genes. The gray module represented genes that could not be classified into any one module and included a total of 488 genes.

Association Analysis of Differentially Expressed Transcription Factors and Core Genes
Transcription factors play a crucial role in gene expression. In the present study, a total of 42 TFs, representing 19 TF families, were identified in the purple and cyan modules. Most of the TFs belonged to the MYB, GRAS, ARF, and other families ( Figure 6A; Table S6). To identify TFs significantly correlated with hormone signaling pathways, we constructed a co-expression network using the hormone signaling-related DEGs in the module ( Figure 6B; Table S6). The screening threshold for high correlation was set as the absolute value of r > 0.7. Larger nodes indicate stronger associations between genes. In this co-expression network, 165 pairs of corresponding relationships between 18 hormone signaling pathway genes and 28 TFs were identified. Among the hormone signaling pathway genes, BRI1 and SAUR, which were relatively highly associated with TFs, were downregulated in tetraploids compared with diploids ( Figure 6B; Table S6). This result indicated that the hormone signal promoting growth was weakened. Further analysis of TFs directly associated with hormone signaling pathways in the network showed that the TFs such as MYB, G2-like, and BBX had high connectivity with hor- The correlations between gene modules and phenotypes were analyzed. The purple module (comprising 650 genes) showed the strongest positive correlation with IAA (r 2 = 0.86, p-value < 0.001) and GA (r 2 = 0.76, p-value < 0.001) ( Figure 5B). A GO enrichment analysis of genes in the purple module showed that the constituent genes were mainly enriched in hormone synthesis and metabolism (GO:0010423, negative regulation of brassinosteroid biosynthetic process; GO:0090032, negative regulation of steroid hormone biosynthetic process; GO:0032351, negative regulation of hormone metabolic process; GO:0032353, negative regulation of hormone biosynthetic process; GO:0010422, regulation of brassinosteroid biosynthetic process; GO:0090030, regulation of steroid hormone biosynthetic process), cell growth and development (GO:0040007, growth; GO:0009826, unidimensional cell growth; GO:0016049, cell growth; GO:0048589, developmental growth), and other related biological processes (Table S5). A KEGG pathway enrichment analysis of genes in the purple module showed that these genes were mainly involved in plant hormone signal transduction (kegg:04075, plant hormone signal transduction), photosynthesis (kegg:00196, photosynthesis-antenna proteins), brassinosteroid biosynthesis (kegg:00905, brassinosteroid biosynthesis), and other biological processes ( Figure 5C). The cyan module (comprising 145 genes) had the strongest positive correlation with ABA (r 2 = 0.77, p-value = 0.0002) and JA (r 2 = 0.7, p-value = 0.001) ( Figure 5B). The KEGG pathway enrichment analysis showed that these genes were mainly involved in photosynthesis (kegg:00195, photosynthesis) and plant hormone signal transduction (kegg:04075, plant hormone signal transduction) ( Figure 5D). Comparison of the gene annotations in the purple and cyan modules revealed that there were complex signal networks and crosstalk between different hormone signal transduction pathways. These genes jointly regulated the growth and development of cells, and then adjusted the hormone signals, so as to regulate plant growth.

Association Analysis of Differentially Expressed Transcription Factors and Core Genes
Transcription factors play a crucial role in gene expression. In the present study, a total of 42 TFs, representing 19 TF families, were identified in the purple and cyan modules. Most of the TFs belonged to the MYB, GRAS, ARF, and other families ( Figure 6A; Table S6). To identify TFs significantly correlated with hormone signaling pathways, we constructed a co-expression network using the hormone signaling-related DEGs in the module ( Figure 6B; Table S6). The screening threshold for high correlation was set as the absolute value of r > 0.7. Larger nodes indicate stronger associations between genes. In this co-expression network, 165 pairs of corresponding relationships between 18 hormone signaling pathway genes and 28 TFs were identified. Among the hormone signaling pathway genes, BRI1 and SAUR, which were relatively highly associated with TFs, were downregulated in tetraploids compared with diploids ( Figure 6B; Table S6). This result indicated that the hormone signal promoting growth was weakened. Further analysis of TFs directly associated with hormone signaling pathways in the network showed that the TFs such as MYB, G2-like, and BBX had high connectivity with hormone pathway genes in the module (log 2 FC ≥ 1 and p-value < 0.05). Interestingly, these TFs were also downregulated in tetraploids. These results suggested that plant hormone-related TFs were involved in the regulation of dwarfing in tetraploids, and they played important roles in the expression and transcription regulatory network.

qRT−PCR Validation
To verify the authenticity of the RNA-sequencing (RNA-seq) data, eight DEGs as- (B) Co-expression network between pathway genes (hormone) and TFs. The green circular nodes represent TFs. The red octagonal nodes represent the pathway genes of the plant hormone. Among these nodes, yellow nodes represent significant differences (p-value < 0.05). The node size is positively correlated with the gene degree.

qRT−PCR Validation
To verify the authenticity of the RNA-sequencing (RNA-seq) data, eight DEGs associated with IAA, GA, JA, and ABA signaling were randomly selected for real-time fluorescence quantification (Figure 7; Table S7). The relative expression patterns of all DEGs determined by qRT−PCR were similar to those of the RNA-seq data. Thus, the results obtained by transcriptome sequencing were reliable. circular nodes represent TFs. The red octagonal nodes represent the pathway genes of the plant hormone. Among these nodes, yellow nodes represent significant differences (p-value < 0.05). The node size is positively correlated with the gene degree.

qRT−PCR Validation
To verify the authenticity of the RNA-sequencing (RNA-seq) data, eight DEGs associated with IAA, GA, JA, and ABA signaling were randomly selected for real-time fluorescence quantification (Figure 7; Table S7). The relative expression patterns of all DEGs determined by qRT−PCR were similar to those of the RNA-seq data. Thus, the results obtained by transcriptome sequencing were reliable.

Discussion
With technological progress, the advantages presented by polyploid plants and elucidation of the underlying mechanisms indicate that we will enter an era of artificially induced plant polyploidy [28]. Compared with diploids, polyploidy often leads to phenotypic changes. In the present study, the phenotypic and anatomical characteristics of diploids and tetraploids generated by chromosome doubling were compared. The plant height of the tetraploids was significantly reduced compared with that of the diploids

Discussion
With technological progress, the advantages presented by polyploid plants and elucidation of the underlying mechanisms indicate that we will enter an era of artificially induced plant polyploidy [28]. Compared with diploids, polyploidy often leads to phenotypic changes. In the present study, the phenotypic and anatomical characteristics of diploids and tetraploids generated by chromosome doubling were compared. The plant height of the tetraploids was significantly reduced compared with that of the diploids ( Figure 1A), which is similar to results reported for tetraploid Eucalyptus [9] and tetraploid apple [10]. The cross-sectional cell area in the leaf of tetraploid poplar 84K was larger, which was similar to previous findings for polyploid Arabidopsis [29] and tetraploid Medicago sativa [30]. Senescence was accelerated in the tetraploids (Figure 1), which was similar to previous reports for hexaploid Populus [12], tetraploid Populus [21], and tetraploid cucumber [31]. Furthermore, we also measured the endogenous hormone contents of diploids and tetraploids. The contents of IAA and GA, two hormones that promote plant growth and development [25], were significantly increased in tetraploid poplar 84K. In contrast, JA and ABA, which inhibit plant growth [26], were significantly reduced in tetraploids ( Figure 4A). These results are consistent with previous studies of the dwarf shengyin bamboo (Phyllostachys edulis f. tubaeformis) [32], tetraploid apple [10], and dwarf mutants of tartary buckwheat [33].
Previous studies have shown that polyploidy results in changes in gene expression levels [34]. To evaluate the changes at the molecular level after polyploidization of poplar 84K, the current study used the genome sequence of poplar 84K [35] as a reference and RNAseq technology to compare the transcriptome of leaves in tetraploids and diploids. Through comparative analysis, 7917 DEGs were detected (Figure 2A). Among the differentially expressed genes in tetraploids and diploids studied by predecessors, phytohormone-related genes comprised an important assemblage [19,20]. Therefore, GO and KEGG enrichment analyses of the DEGs revealed that a large number of DEGs were involved in hormone signaling pathways (Figure 3G-I; Table S4). To explore the molecular mechanism via which hormones affect plant growth and to determine the relationship between DEGs and the dwarf phenotype of tetraploid poplar 84K, we focused attention on the crucial genes involved in four phytohormone signaling pathways.
Auxin is an important growth regulator and plays an important role in plant growth and development [36]. AUX/IAA, GH3, and SAUR are considered to be early and major auxinresponse genes [37]. Among these genes, the AUX/IAA family plays a crucial role in inhibiting the expression level of genes activated by auxin response factors (ARFs) [38,39]. In the current study, four AUX/IAA4 genes (Pop_A05G001751, Pop_G02G023725, Pop_G05G057440, and Pop_G05G072154) were significantly upregulated in the 11th leaf of tetraploids, and two AUX/IAA13 genes (Pop_A10G047408 and Pop_G10G064704) were significantly upregulated in the fifth and 11th leaves, respectively. The GH3 gene family is involved in hormonal homeostasis by binding to free forms of IAA and JA [40,41]. Overexpression of GH3.1 and GH3.6 can significantly reduce the auxin content and lead to dwarfism in rice [42,43]. Compared with diploids, three GH3.1 genes and one GH3.6 gene were significantly upregulated in the first and 11th leaves of tetraploids. In addition, GH3.5 and LAX3 play positive roles in regulating the auxin signaling pathway [44,45]. In tetraploids, one GH3.5 gene (Pop_A19G055012) was downregulated in the leaves at all three nodes, and one LAX3 gene (Pop_A02G088716) was downregulated in the fifth leaf. SAUR36 is a positive regulator of leaf senescence and mediates auxin-induced leaf senescence in Arabidopsis [46,47]. In the present study, two SAUR36 genes were significantly upregulated in tetraploids ( Figure 4B; Table S4). Thus, the changes in expression of these crucial genes involved in the auxin signaling pathway were integral to the reduction in auxin content of the tetraploids. The decrease in IAA content may be one reason for the slow growth of the tetraploid plants.
Gibberellin plays an important role in regulating plant growth and development [48]. Gibberellin is a vital factor in the control of plant height. For example, dwarfism results if the genes involved in GA synthesis or signal transduction are mutated in sponge gourd (Luffa acutangula) [49]. Increasing evidence indicates that GA promotes nutritional and reproductive growth by triggering the degradation of DELLA proteins, which are the main inhibitors of GA signal transduction [50,51]. In the present study, six genes encoding DELLA proteins were significantly upregulated in all three sampled leaves of tetraploids. The upregulation of DELLA proteins will lead to the reduction in GA content. In addition, the gibberellin receptor GID1 binds to bioactive gibberellins (e.g., GA 1 , GA 3 , and GA 4 ) to degrade DELLA proteins, thereby relieving DELLA-mediated growth inhibition [52,53]. Compared with diploids, two GID1 genes (Pop_A14G000831 and Pop_G14G045139) were downregulated in the first and fifth leaves of tetraploids ( Figure 4A; Table S3). Therefore, GID1 expression was downregulated, leading to DELLA protein accumulation, thus reducing the GA content and resulting in dwarfism of the plants.
Jasmonic acid is a lipid-derived plant hormone that is essential for plant defense and plant development, and it is an important endogenous signal that activates the expression of senescence-related genes and induces leaf senescence [54,55]. A critical element in JA signaling is JAZ proteins, which are inhibitors of JA signaling [56]. In addition, MYC2 acts upstream of the JA signaling pathway and functions as the regulatory center for plant hormone signal transduction by integrating various endogenous and exogenous signals that affect plant growth and development [57]. Previous studies have shown that MYC2 plays an active role in JA-mediated leaf senescence [58]. In the current study, four genes encoding JAZ proteins were downregulated in the first and fifth leaves of tetraploids, and two genes encoding MYC2 were upregulated in the three leaves sampled ( Figure 4A; Table S4). Therefore, the downregulated expression of JAZ proteins in the tetraploids reduced the repression of MYC2 and activated JA to induce leaf senescence.
Abscisic acid is considered to be a growth inhibitor, and an increase in the ABA content inhibits germination and causes growth retardation [59]. SnPK2 is a positive regulator of ABA signaling in Arabidopsis thaliana [60]. The released SNF1-RELATED PROTEIN KINASE 2 (SnRK2) is then activated by autophosphorylation or phosphorylation by other kinases, and the activated SnRK2 can phosphorylate downstream proteins or transcription factors, including ABA response element binding factors (AREB/ABF) and bZIP TFs [61]. In the present investigation, two SnPK2 genes, two ABF3 genes, and one bZIP gene were upregulated in tetraploids, and the number of upregulated genes was greatest in the 11th leaf ( Figure 4A; Table S4). The results showed that the upregulated expression of SnPK2, ABF3, and bZIP genes in the ABA signaling pathway led to accumulation of ABA and, thus, accelerated aging.
The dwarf phenotype of plants is not only affected by the synergistic and antagonistic effects of hormones, but also by the regulatory effects of TFs. The role of TFs is to transform external signals into intracellular signals, thereby stimulating specific hormone signaling pathways and gene expression to maintain growth [62]. Here, the WGCNA co-expression network was used to explore the regulatory relationship between TFs and pathway genes. It was observed that MYB, G2-like, and BBX genes were downregulated in all three sampled leaves, and BRI1, SAUR, and other pathway genes strongly correlated with TFs were downregulated ( Figure 6; Table S6). Plant MYB proteins have been reported to play a role in hormonal responses during seed development and germination [63,64]. MYB33 mediates GA signaling in plant growth and is a positive regulator of GA signal transduction [65]. MYB61 activates the expression of GA3ox1 and GA3ox2 genes, which in turn may lead to increased GA 4 content in Artemisia annua and Arabidopsis thaliana [66]. In the present study, one MYB33 gene and two MYB61 genes were all downregulated in tetraploids, which may have inhibited GA. In addition, G2-like (GLK) TFs play important roles in regulating plant aging and hormone contents [67,68]. Overexpression of GLK in maize and tomato promotes chloroplast development and the expression of a photosynthesis-related gene [69]. In tobacco, GLK inhibits ABA signaling, and the expression of GLK gradually decreases with the aging of leaves [70]. A GLK gene was downregulated in the fifth and 11th leaves of tetraploid poplar 84K, and it might participate in ABA signaling as a negative regulator and accelerate leaf senescence. BBX proteins play an important role in light-regulated signaling pathways and plant hormone-mediated development. BBX negatively regulates the expression of ABI, a positive regulator of ABA signaling [71]. BBX18 is involved in the GA signaling pathway and increases bioactive GA contents to promote growth by enhancing the expression of the GA metabolic genes GA3ox1 and GA20ox1 [72]. A BBX18 gene was downregulated in the 11th leaf of tetraploid poplar 84K, which may have led to the reduction in GA content observed in this study. These changes may have reduced the accumulation of IAA and GA, increased the accumulation of JA and ABA, and inhibited the growth and development of tetraploid poplar 84K.

Plant Materials and Growth Conditions
Diploid (2n = 2x = 38) and tetraploid (2n = 4x = 96) individuals of poplar 84K were used as materials in this study. Tetraploids were obtained by colchicine-induced doubling of the somatic chromosomes of diploids [24]. Rooting culture was conducted in half-strength Murashige-Skoog medium supplemented with 0.02 mg/mL 1-naphthaleneacetic acid, 0.05 mg/mL indole-3-butyric acid, 30 g/L sucrose, and 6.5 g/L agar (the culture medium and conditions of all plants were consistent). After culture for approximately 30

Measurement of Phenotypic Traits and Anatomical Observations
The plant height (H) of 15 diploid and 15 tetraploid plants of poplar 84K was measured with a steel tape once per month for a total of 5 months. The fully expanded fifth leaf (including the main leaf vein) and the stem internode between the fourth and fifth leaves (numbered from the shoot tip) were sampled from the tetraploids and diploids. The samples were fixed in FAA fixative (0.25 mL of 38% formaldehyde, 0.25 mL of glacial acetic acid, and 4.5 mL of 50% ethanol) for more than 24 h. The fixed samples were dehydrated sequentially in an ethanol gradient series (50%, 70%, 95%, and 100%), then infiltrated with 100% xylene for 30 min, and embedded in paraffin at 60 • C for 2 h. After solidification, the embedded sample was trimmed to a suitably sized trapezoid block, and sections of thickness~8 µm were cut with a microtome. After dewaxing, the sections were stained with 1% safranin O and 0.1% fast green. All sections were observed and photographed with an Olympus BX51 microscope.

Total RNA Extraction and Transcriptome Sequencing
The materials to be sampled were determined by analyzing the photosynthetic rate of diploids and tetraploids (Table S1). Therefore, the first, fifth, and 11th leaves were sampled simultaneously when the plants were 3 months old. Three replicates were randomly selected for each ploidy. The sampled materials were immediately frozen in liquid nitrogen and stored at −80 • C for subsequent sample sequencing and measurement of physiological indicators. The TRIzol Reagent Kit (TIANGEN Biotech Co., Ltd., Beijing, China) was used to extract RNA from the samples in accordance with the manufacturer's instructions. A NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific Inc., Wilmington, DE, USA) was used to determine the RNA quality. The integrity of the RNA was verified by electrophoresis in 1.5% agarose gel. After the RNA quality was confirmed, the extracted RNA was reverse-transcribed, and cDNA libraries constructed with the FastKing RT Kit (with gDNase). The libraries were sequenced on an Illumina HiSeq platform at the Hangzhou Lianchuan Biological Technology Co., Ltd. (Beijing, China).

Enrichment Analysis of DEGs
The transcript abundance was calculated as reads per kilobase per million mapped reads (RPKM). The edgeR software (https://bioconductor.org/packages/release/bioc/ html/edgeR.html) (accessed on 24 October 2021) was used for differential expression analysis between samples [74]. Genes with fold difference multiples of |log 2 (FC)| > 1 and p-value < 0.05 were defined as DEGs. In addition, Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses were performed. Using the GO database (Gene Ontology, 2021), the DEGs were annotated using the GO online tools with terms classified into three categories: molecular function, biological process, and cell composition. Using the KEGG database (http://www.kegg.jp/) (accessed on 17 June 2022), the DEGs were annotated with signaling pathways and metabolic pathways.

Determination of Endogenous Hormone Contens
Briefly, 0.5 g of the first, fifth, and 11th leaves stored at −80 • C were used for the determination of hormone content. Contents of the plant hormones IAA, GA 3 , ABA, and JA were determined using an enzyme linked immunosorbent assay in accordance with a previously described method [75]. Each treatment had three biological repetitions and three technical repetitions.

Division of Co-Expression Modules and Visualization of Gene Expression
The gene co-expression network was constructed using the WGCNA package (version 4.0.2) for R software. To maintain gene connectivity and place greater weight on the strongest correlation, a soft threshold for the correlation matrix (β) was selected. The hierarchical clustering dendrogram of the TOM matrix was constructed using the average distance with a minimum size threshold of 30 and the merge cut height of 0.25. Genes with similar expression patterns were clustered and divided into unified modules, and a co-expression network was constructed. The correlations among gene modules were analyzed using Pearson correlation analysis, and the expression patterns of modules and samples were calculated. The FPKM values of the genes were used to map the network of pathways, genes, and TFs. The screening threshold was |R| ≥ 0.7. Positive values indicated a positive correlation, and negative values indicated a negative correlation. The PlantTFDB database (http://planttfdb.gao-lab.org/) (accessed on 17 June 2022) was used to identify genes encoding TFs among the DEGs. The gene co-expression networks were visualized using Cytoscape 3.8.2, and the same software was used to calculate the gene connectivity. The size of the nodes was positively correlated with the degree of gene connection.

qRT−PCR Verification of Differentially Expressed Genes
RNA was extracted from the leaves of diploid and tetraploid individuals. After determining the RNA concentration and quality, the first-strand cDNA was reverse-transcribed using the FastKing RT Kit (TIANGEN Biotech Co., Ltd., Beijing, China). Nine DEGs were selected for verification by qRT-PCR analysis, among which the constitutively expressed Actin gene was used as the internal reference gene (GenBank accession number: EF145577). Real-time PCR amplifications were performed on the Applied Biosystems 7500 Fast Instrument (AB Ltd., Lincoln, NE, USA) using the 2 × SYBR ® Green qPCR Mix Kit (Aidlab, China). Gene-specific primers for the nine DEGs were designed using the Primer3 Plus online tool (http://www.primer3plus.com/) (accessed on 17 June 2022). The primer sequences are listed in Table S7. The RT-qPCR protocol consisted of 40 cycles, comprising initial denaturation at 95 • C for 5 min, denaturation at 95 • C for 10 s, annealing at 60 • C for 30 s, and extension at 72 • C for 15 s. The relative expression level of genes was calculated using the 2 −∆∆Ct method [76]. All reactions comprised at least three biological replicates and three technical replicates.

Statistical Analysis
The experimental data were statistically analyzed using SPSS 20.0 (IBM Corporation Inc., Armonk, NY, USA). The results were expressed as the mean ± standard deviation. The significance of the difference among means was determined by Student's t-test at p < 0.05, p < 0.01, and p < 0.001. The cell area, length, and width for the leaves and stems were determined with ImageJ. TBtools was used to draw the heatmaps [77].

Conclusions
The molecular mechanism of trait variation in plant polyploids is highly complex. Hormone synthesis, signal transduction, and gene regulation are all associated with phenotypic changes. In this study, compared with the diploids, most DEGs positively correlated with growth in the tetraploids were down-regulated, consistent with IAA and GA contents in the leaves, and most DEGs negatively correlated with growth were upregulated, in accordance with JA and ABA contents in the leaves. The changes in endogenous hormone contents were consistent with transcriptomic changes. In addition, co-expression network analysis showed that TFs may also play important regulatory roles that together lead to tetraploid dwarfing. The present results provide a theoretical basis for future research on the mechanism of plant trait variation at the molecular level in polyploid woody plants.