Transcriptome Analysis Reveals the Regulatory Networks of Cytokinin in Promoting Floral Feminization in Castanea henryi

Castanea henryi is a monoecious plant with a low female-to-male ratio, which limits its yield. The phytohormone cytokinin (CK) plays a crucial role in flower development, especially gynoecium development. Here, the feminizing effect of CK on the development of C. henryi was confirmed by the exogenous spraying of N-(2-chloro-4-pyridyl)-N’-phenylurea (CPPU). Spraying CPPU at 125 mg·L−1 thrice changed the male catkin into a pure female catkin, whereas at 5 mg·L−1 and 25 mg·L−1, only a part of the male catkin was transformed into a female catkin. A comparative transcriptome analysis of male catkins subjected to CPPU was performed to study the mechanism of the role of CKs in sex differentiation. Using Pearson’s correlation analysis between hormone content and hormone synthesis gene expression, four key genes, LOG1, LOG3, LOG7 and KO, were identified in the CK and GA synthesis pathways. Moreover, a hub gene in the crosstalk between JA and the other hormone signaling pathways, MYC2, was identified, and 15 flowering-related genes were significantly differentially expressed after CPPU treatment. These results suggest that CK interacts with other phytohormones to determine the sex of C. henryi, and CK may directly target floral organ recognition genes to control flower sex.


Introduction
Castanea henryi (Skam) Rehd. et Wils. is an important woody food plant of the Fagaceae family. Its nuts are renowned for their sweet and glutinous taste and high nutritional value [1]. However, the low female-to-male flower ratio of C. henryi, with over two thousand male flowers arranged in the male catkins, only one bisexual catkin for every five male catkins and only two female flowers growing at the base of the bisexual catkins [2], has significantly limited its seed yield and industrial potential. Therefore, revealing the molecular mechanism of C. henryi sex determination is of great significance for improving its fruit yield.
Various stages of plant growth and development are controlled by a variety of phytohormones. Phytohormones, such as auxin, ethylene (ETH), cytokinin (CK) and gibberellin (GA), regulate sex differentiation. It is generally believed that CK, auxin and ETH are positive regulators of pistil development, whereas GA promotes the development of androecium organs [3][4][5][6]. Spraying CK on Mercurialis annua increases the number of female flowers [7]. Moreover, the exogenous spraying of 6-benzylaminopurine  in Jatropha curcas can increase the total number of flowers on the catkin, induce bisexual flowers and improve the ratio of males to females [8]. Likewise, the fine-tuning of endogenous CKs can increase flower number and male-to-female ratios in J. curcas [9]. In C. henryi, CK is accumulated more in female flowers than in male flowers, and the exogenous application henryi, CK is accumulated more in female flowers than in male flowers, and the exogenous application of CK increases the number of female flowers. However, the mechanisms through which CK regulates sex differentiation remain unknown.
In the present study, we proved that CK is a crucial regulator of sex determination in C. henryi. The feminization effect of N-(2-chloro-4-pyridyl)-N'-phenylurea (CPPU) treatment, which has been reported to be 10-100 times more effective than 6-BA [10] and has a considerable feminization effect on Vitis amurensis [11] and Litchi chinensis [12], was detected at different spray times (once, twice and thrice) and spray concentrations (5 mg·L −1 , 25 mg·L −1 , and 125 mg·L −1 ). Thus, the levels of nine endogenous hormones were determined, and a comparative transcriptomic analysis was performed of male catkins under different treatments in C. henryi to identify the regulation networks that participate in promoting the floral feminization in C. henryi. The present study provides a basic framework for the regulatory networks of sex determination in C. henryi, and transcriptome data provide a theoretical basis for C. henryi genetic diversity analysis and molecular-assisted breeding.

Feminization Effects of CPPU Treatment on Floral Development in C. henryi
The catkins of C. henryi are divided into two types: male and bisexual.

Effect of CPPU Treatment on the Content of Endogenous Hormones in C. henryi Male Catkins
The levels of the four measured cytokinins, isopentenyladenine (iP), isopentenyladenosine (iPR), zeatin (tZ) and trans-zeatin riboside (tZR), were significantly lower in the treatment group than in the control group (Figure 3a). The levels of the two gibberellins observed, GA 1 and GA 4 , were significantly lower in the treated group than in the control group ( Figure 3b). Auxin (IAA) levels did not differ significantly between the control and treatment groups. Abscisic acid (ABA) levels showed no clear trend among the treatment groups. The level of jasmonate (JA) in the CH_25 group was significantly lower than it was in the other groups ( Figure 3c).

Effect of CPPU Treatment on the Content of Endogenous Hormones in C. henryi Male Catkins
The levels of the four measured cytokinins, isopentenyladenine (iP), isopentenyladenosine (iPR), zeatin (tZ) and trans-zeatin riboside (tZR), were significantly lower in the treatment group than in the control group (Figure 3a). The levels of the two gibberellins observed, GA1 and GA4, were significantly lower in the treated group than in the control group ( Figure 3b). Auxin (IAA) levels did not differ significantly between the control and treatment groups. Abscisic acid (ABA) levels showed no clear trend among the treatment groups. The level of jasmonate (JA) in the CH_25 group was significantly lower than it was in the other groups (Figure 3c).
In plants, tZ and iP are the primary bioactive CKs [13], and GA1 and GA4 are the primary bioactive gibberellins [14]. Hence, the levels of tZ + iP and GA1 + GA4 were determined ( Figure 3d). The results show that, after CPPU treatment, the levels of tZ + iP and GA1 + GA4 were significantly lower in the treatment groups than they were in the control groups. The levels of endogenous tZ + iP and GA1 + GA4 in the male catkins of C. henryi treated with different concentrations of CPPU. The values represent the means ± standard deviations (n = 3). Student's t-test was used for the statistical analyses. Small letters mean a significant difference at the 0.05 level. In plants, tZ and iP are the primary bioactive CKs [13], and GA 1 and GA 4 are the primary bioactive gibberellins [14]. Hence, the levels of tZ + iP and GA 1 + GA 4 were determined (Figure 3d). The results show that, after CPPU treatment, the levels of tZ + iP and GA 1 + GA 4 were significantly lower in the treatment groups than they were in the control groups.

Transcriptome Sequencing of the Male Catkins of C. henryi in Response to CPPU Treatment
The induction of abundant female flowers on the male catkin by the CPPU treatment suggests that CK is a key factor in regulating sex differentiation in C. henryi. Subsequently, a comparative transcriptome analysis of male catkins treated with CPPU was performed to investigate the mechanism of action of CKs in the sex differentiation of C. henryi.
In C. henryi, flower buds are located in the catkin. Two weeks after the third spray, male catkins from different treatments were sampled. The samples from CK, A7, B7 and C7 treatments in Table 5 were named CH_0, CH_5, CH_25 and CH_125, respectively. To identify early response genes involved in CK-regulated sex determination, three replicates of four groups of samples (three different concentrations of CPPU-treated and pure water-treated male catkins) were collected for subsequent transcriptome sequencing and quantitative real-time polymerase chain reaction (RT-qPCR). Twelve cDNA libraries were constructed and sequenced using an Illumina high-throughput sequencing platform. After sequencing, the raw data were obtained, the unqualified sequences were filtered out to obtain clean data and statistics were performed to obtain basic information about the data ( Table 1). The distributions of gene functional elements in all samples and the Pearson correlation coefficient analysis between each sample are shown in Figure 4. The induction of abundant female flowers on the male catkin by the CPPU treatment suggests that CK is a key factor in regulating sex differentiation in C. henryi. Subsequently, a comparative transcriptome analysis of male catkins treated with CPPU was performed to investigate the mechanism of action of CKs in the sex differentiation of C. henryi.
In C. henryi, flower buds are located in the catkin. Two weeks after the third spray, male catkins from different treatments were sampled. The samples from CK, A7, B7 and C7 treatments in Table 5 were named CH_0, CH_5, CH_25 and CH_125, respectively. To identify early response genes involved in CK-regulated sex determination, three replicates of four groups of samples (three different concentrations of CPPU-treated and pure watertreated male catkins) were collected for subsequent transcriptome sequencing and quantitative real-time polymerase chain reaction (RT-qPCR). Twelve cDNA libraries were constructed and sequenced using an Illumina high-throughput sequencing platform. After sequencing, the raw data were obtained, the unqualified sequences were filtered out to obtain clean data and statistics were performed to obtain basic information about the data ( Table 1). The distributions of gene functional elements in all samples and the Pearson correlation coefficient analysis between each sample are shown in Figure 4.

Functional Annotation and Classification of Differentially Expressed Genes (DEGs)
To compare the differences in gene expression between different samples, the expression levels of significantly different genes in all samples were extracted into a list, which was then used to draw a volcano plot of differentially expressed genes. The differences of gene expression levels were plotted with colored dots, and there were separately 2129, 5275 and 2786 genes differentially expressed after three different concentrations (5 mg·L −1 , 25 mg·L −1 and 125 mg·L −1 ) of CPPU treatment ( Figure 5). To visualize the similarity and overlap in the number of DEGs in different samples, Venn diagrams are used to show the number of shared and unique genes in different samples ( Figure 6). To verify the accuracy of the transcriptome sequencing results, RT-qPCR was used to detect the expression levels of nine DEGs. The primer sequences for the reference gene (GAPDH) and the nine selected genes are listed in Table       For an overall description of their biological functions, the DEGs were analyzed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and Gene Ontology (GO) analyses. The KEGG analysis divided the DEGs into 22 categories, among which categories of "Metabolism" (66%) and "Genetic information processing" (14%) were the most prominent ( Figure 8). In addition, GO analysis divided the DEGs into "cellular process", "metabolic process", "cell", "cell part" and "catalytic activity", which were related to the basic biological function of CK to promote cell growth and metabolism.

Screening of Key Genes in the Hormone Biosynthesis Pathway
GAs and CKs play a leading role in the flower sex differentiation of C. henryi [19]. The genes related to the biosynthetic pathways of these two hormones were investigated, and the results show that the concentration of bioactive CKs is significantly positively correlated with the genes LOG1 (Che016788), LOG3 (Che010959) and LOG7 (Che008994) in their metabolic pathways (Table 3), and these three genes were significantly down-regulated after CPPU treatment compared with the control group. LOG is a key gene family that can regulate the synthesis of CKs, and changes in its expression positively affect the The CK signal transduction pathway is also known as the two-component signaling system [18]. In this study, some genes involved in CK signal transduction were found to be significantly differentially expressed. The results show that HP5 (Che022632, Che027964 and Che022630) and ARR9 (Che022324) were significantly down-regulated, whereas HP4 (Che000841), HK4 (Che018412) and ARR17 (Che019431) were significantly up-regulated after CPPU treatment.

Screening of Key Genes in the Hormone Biosynthesis Pathway
GAs and CKs play a leading role in the flower sex differentiation of C. henryi [19]. The genes related to the biosynthetic pathways of these two hormones were investigated, and the results show that the concentration of bioactive CKs is significantly positively correlated with the genes LOG1 (Che016788), LOG3 (Che010959) and LOG7 (Che008994) in their metabolic pathways (Table 3), and these three genes were significantly downregulated after CPPU treatment compared with the control group. LOG is a key gene family that can regulate the synthesis of CKs, and changes in its expression positively affect the concentration of bioactive CKs in C. henryi. The bioactive GA concentration is significantly positively correlated with KO (Che010129) in its biosynthetic pathway (Table 4). Table 3. Correlations between the concentrations of cytokinin and its metabolism-related genes.

Identification of the DEGs Related to Floral Development
The development of floral organs of typical higher plants can be divided into four rounds, with different genes regulating the floral organs in each round. Most ABC floral organ regulator genes belong to the MADS-box gene family [20,21]. Transcriptome data show that 16 genes related to flower development responded to CPPU treatment ( Figure 10

Identification of the DEGs Related to Floral Development
The development of floral organs of typical higher plants can be divided into four rounds, with different genes regulating the floral organs in each round. Most ABC floral organ regulator genes belong to the MADS-box gene family [20,21]. Transcriptome data show that 16 genes related to flower development responded to CPPU treatment ( Figure  10

CPPU Has Strong Feminization Effects on the Floral Development in C. henryi
To increase the yield of monoecious plants, it is important to increase the ratio of male to female flowers. Phytohormones such as CK [22], GA [23], auxin [24] and ETH [25] promote the conversion of male flowers into female flowers in different plants. Exogenous CK spraying has been reported to alter the ratio of male to female flowers in Actinidia chinensis [26], Jatropha curcas [6] and Mercurialis annua [7]. The results of the present study on C. henryi are consistent with these findings, as CK treatment showed strong feminization effects on floral development. This suggests that an astonishing 100% conversion rate can be achieved with appropriate treatment.
These results suggest that CK is a key regulator controlling female flower development as well as a potential determinant of sex differentiation in C. henryi. In addition, C. henryi was observed to be less sensitive to low concentrations of CPPU, with the conversion of only some of the male flowers into female flowers, which was significantly lower than the promotional effect of high concentrations. Therefore, it is suggested that low concentrations of CPPU are insufficient for the complete conversion of male flowers to female flowers. However, exogenous CPPU treatment was still effective in inducing the conversion of male flowers to female flowers. Transcriptome data analysis of different treatment groups can provide precious information for studying the gene network of CKs that regulate sex differentiation in C. henryi.

Exogenous Cytokinin Treatment Alters Endogenous Cytokinin Levels and the Expression of Cytokinin Biosynthesis and Signaling-Related Genes
In plants, CKs exist in various forms, and the supply of exogenous CKs affects endogenous cytokinin content [27]. In this study, exogenous cytokinin (CPPU) treatment was observed to significantly reduce the level of endogenous cytokinin, which may be due to the presence of bioactive cytokinin in other bound forms, reducing endogenous cytokinin levels. Furthermore, genes related to the CK metabolic pathway were identified, and the results show that the expression levels of three genes (CKX3, CKX7 and CKX5) of the CKX gene family, which are involved in CK degradation and the regulation of CK homeostasis in plants [28,29], are negatively correlated with cytokinin levels. The expression levels of three genes (LOG7, LOG1 and LOG3) of the LOG gene family, which encode a 5 -ribose monophosphate hydrolase that directly converts iPRMP and tZRMP to iP and tZ [13], are significantly positively correlated with CK levels. The results show that the exogenous CPPU treatment of C. henryi altered the expression of genes related to the CK metabolic pathway, which in turn affected the levels of endogenous CK. In LOG mutants, rice panicles were severely reduced in size, and branching patterns were abnormal. This is accompanied by the abnormal development of floral organs, with flowers having only one stamen and no pistil [15]. Hence, it was inferred that the decrease in the relative expression of LOG in the treatment group may affect the normal formation of stamens and promote female flower differentiation.
In the signaling model of CK, it binds to the receptor of histidine kinases (HKs) and transfers the phosphate group to downstream histidine transfer proteins (HPs), which in turn transfers the phosphate group to type-A response regulators (type-A RRs) or type-B response regulators (type-B RRs) [30] (Figure 11). Type-A and type-B RRs regulate the transcription of CK-responsive genes, thereby regulating plant growth and development [30,31]. In this study, five genes were differentially expressed in the signal transduction pathway. Among them, the relative expression of HK4, ARR17 and HP4 in the treatment group was significantly higher than it was in the control group, indicating that the activities of genes related to the CK signaling pathway in female flowers were higher than those in male flowers.

Exogenous CK Treatment Alters Endogenous GA Levels and the Expression of GA Biosynthesis and Signaling-Related Genes
GA has been demonstrated to be a key hormone that promotes male flower formation in C. henryi [19]. In this study, the GA content in the treatment group was significantly lower than that in the control group. Moreover, the GA content showed a decreasing trend with increasing CPPU concentrations, which was consistent with the increase in the proportion of female flowers. The significantly decreased expression levels of CPS, KO and GA3ox and the increased expression levels of GA2ox may be the reason for the decreased endogenous GA levels. Ent-kaurene oxidase KO is a single-gene regulatory enzyme that functions upstream of GA synthesis. It is a cytochrome P450 and NADPH-dependent monooxygenase belonging to the CYP701 subfamily that catalyzes the final formation of ent-kaurene acid by catalyzing ent-kaurene C19 [33]. The down-regulation of AtKO1 in transgenic Arabidopsis results in a decrease in GA4 content [34]. Tissue-specific studies have shown that KO gene expression exhibits distinct spatial differences [35]. The gene AtKO in Arabidopsis is expressed in all tissues, with the highest expression observed in the inflorescence. In this study, the gene KO showed the highest relative expression in male catkins of the control group and is significantly positively correlated with GA levels, implying that it affects the formation of catkins.
The gene GASA is one of the few identified target genes downstream of the GA signaling pathway [36,37], and it is mostly expressed strongly in young plant tissues, organs and parts of vigorous growth [38,39]. In Arabidopsis, GASA participates in the regulation of floral meristem decisions, promotes flowering [40] and regulates GA by promoting GA signaling and inhibiting redox activity to promote seed germination [36]. In this study, three members of the GASA gene family were identified as DEGs, indicating that they affect plant hormone signal transduction.

Exogenous CK Treatment Alters the Levels of Other Endogenous Phytohormones
CK coordinates plant growth and development with phytohormones, such as auxin, ABA, ETH and JA, in a synergistic or antagonistic manner [41]. Previous studies have shown that CKs and auxins regulate their biosynthesis and signal transduction during plant growth and development [42]. In this study, although 15 genes related to auxin biosynthesis and signal transduction pathways were differentially expressed, there was no significant difference in auxin content in the treatment groups, possibly owing to the time delay between gene expression and protein synthesis [43]. The role of ABA in sex differentiation differs among plants. In this study, ABA levels showed no clear trend among the treatment groups, suggesting that endogenous ABA levels were not significantly affected by the exogenous CK treatment. CK also interacts with ETH and JA to regulate flowering, and genes related to the two hormones' signal transduction pathways were found to be differentially expressed in this study. It is worth mentioning that the gene

Exogenous CK Treatment Alters Endogenous GA Levels and the Expression of GA Biosynthesis and Signaling-Related Genes
GA has been demonstrated to be a key hormone that promotes male flower formation in C. henryi [19]. In this study, the GA content in the treatment group was significantly lower than that in the control group. Moreover, the GA content showed a decreasing trend with increasing CPPU concentrations, which was consistent with the increase in the proportion of female flowers. The significantly decreased expression levels of CPS, KO and GA3ox and the increased expression levels of GA2ox may be the reason for the decreased endogenous GA levels. Ent-kaurene oxidase KO is a single-gene regulatory enzyme that functions upstream of GA synthesis. It is a cytochrome P450 and NADPH-dependent monooxygenase belonging to the CYP701 subfamily that catalyzes the final formation of ent-kaurene acid by catalyzing ent-kaurene C19 [33]. The down-regulation of AtKO1 in transgenic Arabidopsis results in a decrease in GA 4 content [34]. Tissue-specific studies have shown that KO gene expression exhibits distinct spatial differences [35]. The gene AtKO in Arabidopsis is expressed in all tissues, with the highest expression observed in the inflorescence. In this study, the gene KO showed the highest relative expression in male catkins of the control group and is significantly positively correlated with GA levels, implying that it affects the formation of catkins.
The gene GASA is one of the few identified target genes downstream of the GA signaling pathway [36,37], and it is mostly expressed strongly in young plant tissues, organs and parts of vigorous growth [38,39]. In Arabidopsis, GASA participates in the regulation of floral meristem decisions, promotes flowering [40] and regulates GA by promoting GA signaling and inhibiting redox activity to promote seed germination [36]. In this study, three members of the GASA gene family were identified as DEGs, indicating that they affect plant hormone signal transduction.

Exogenous CK Treatment Alters the Levels of Other Endogenous Phytohormones
CK coordinates plant growth and development with phytohormones, such as auxin, ABA, ETH and JA, in a synergistic or antagonistic manner [41]. Previous studies have shown that CKs and auxins regulate their biosynthesis and signal transduction during plant growth and development [42]. In this study, although 15 genes related to auxin biosynthesis and signal transduction pathways were differentially expressed, there was no significant difference in auxin content in the treatment groups, possibly owing to the time delay between gene expression and protein synthesis [43]. The role of ABA in sex differentiation differs among plants. In this study, ABA levels showed no clear trend among the treatment groups, suggesting that endogenous ABA levels were not significantly affected by the exogenous CK treatment. CK also interacts with ETH and JA to regulate flowering, and genes related to the two hormones' signal transduction pathways were found to be differentially expressed in this study. It is worth mentioning that the gene MYC2, identified in the JA signaling pathway, is a "master switch" in the crosstalk between JA and the other hormone signaling pathways, and it was significantly up-regulated after CPPU treatment. These results indicate that CKs interact with auxin, ABA, GA, ETH and JA to determine the sex of C. henryi. Moreover, these findings suggest that CK determines endogenous hormone levels in C. henryi by altering the expression levels of genes involved in the synthesis and transduction of endogenous CK, GA, auxin, ABA, ethylene and JA, ultimately altering the sex of C. henryi flowers.

CK Regulated the Expression of Genes Related to Floral Organ Development
Studies have shown that DEFICIENS (DEF), PISTILLATA (PI) and AGAMOUS (AG) jointly control stamen formation [44]. Mandel et al. found that the absence of the AP3 (homologous to DEF) gene product in Arabidopsis causes petals to be replaced by a ring-shaped sepal, and it causes stamens to become carpels [45]. Mutations or the ectopic expression of the AG gene has demonstrated its decisive role in stamen and carpel development [45]. In this study, the expression of DEF and AG genes in each treatment group was significantly down-regulated, which may have affected the normal formation of stamens, resulting in a significant decrease in the number of stamens compared to the control group. LFY plays an important role in activating gene activity in floral organs. In the LFY mutant of Arabidopsis, the floral organ-determining gene AP3 was decreased in both the expression region and expression level [46]. LFY and WUS act together to activate the expression of AG and simultaneously act with UNUSUAL FLORAL ORGANS (UFO) to activate the expression of AP3, thereby forming floral organs [47]. In this study, the LFY gene was significantly up-regulated in the treatment groups, presumably indicating its critical role in processes involved in flower sex expression. In rice OsMADS1 gene transformants, the number of stamens was reduced [48]. In this study, the MADS17 gene, which is a direct downstream gene of MADS1 [49], was significantly differentially expressed in the treatment group. Therefore, it is speculated that the differential expression of the MADS17 gene may be the reason for the decreased number of stamens in the treatment group.
The gene PMADS2 was only expressed in the petals and stamens of Petunia [50]. In this study, the expression of PMADS2 was significantly down-regulated, and the number of male flowers in the treatment group was lower than it was in the control group, indicating that the decrease in its expression may affect the development of flowers. The gene SVP can inhibit normal flowering in plants [51]. With an increase in the number of female flowers in the treatment group, the relative expression of the SVP gene decreased, which gradually reduced the inhibitory effect on C. henryi flowering and may be beneficial to the development of pistils. The gene AGL15 is involved in various in vivo regulatory pathways. The gene downstream of AGL15 in Arabidopsis is a key enzyme in the GA regulatory pathway [52]. In the present study, the expression of AGL15 was significantly downregulated in the treatment group. It is speculated that this affected the related regulatory pathway of GA, which in turn changed the number of pistils. This is consistent with the results of the present study. These results suggest that CKs regulate sex differentiation in C. henryi by directly targeting genes associated with androecium and gynoecium development.

Regulatory Networks Participate in Promoting Floral Feminization in C. henryi
In the present study, many genes involved in CK, GA, auxin, ABA, ETH and JA biosynthesis and signal transduction were identified, suggesting that these hormones may be involved in the process of feminization of the male flowers of C. henryi. Particularly, we speculate that the endogenous cytokinin and gibberellin were the key phytohormones that promoted the transformation of male to female flowers in C. henryi, and the expression of genes LOG1/3/7 and KO were the key factors that altered the levels of both phytohormones. Therefore, it would be interesting to investigate how the interaction of these hormones feminizes the male flower of C. henryi. MYC2, which was identified in the JA signaling pathway, has a broad role in regulating developmental programs and controls crosstalk between JA and nearly all the other hormone signaling pathways [53]. MYC2 is not only implicated in cytokinin signaling by interacting with HP5 [54] and HP6 [55] in the cytokinin signaling pathway, but it also transduces the GA signaling to the biosynthesis of the sesquiterpenes pathway by interacting with DELLA [56]. Based on the results, we speculate that MYC2 crosstalks the transduction of CK and GA signaling, which may lead to mutations in one or more genes associated with floral organ development, ultimately causing the transition of androecium to gynoecium in C. henryi (Figure 12). Whether the genes mentioned in the regulatory network are the key target genes to improve the yield of C. henryi remains to be our further study.
signaling pathway, has a broad role in regulating developmental programs and controls crosstalk between JA and nearly all the other hormone signaling pathways [53]. MYC2 is not only implicated in cytokinin signaling by interacting with HP5 [54] and HP6 [55] in the cytokinin signaling pathway, but it also transduces the GA signaling to the biosynthesis of the sesquiterpenes pathway by interacting with DELLA [56]. Based on the results, we speculate that MYC2 crosstalks the transduction of CK and GA signaling, which may lead to mutations in one or more genes associated with floral organ development, ultimately causing the transition of androecium to gynoecium in C. henryi (Figure 12). Whether the genes mentioned in the regulatory network are the key target genes to improve the yield of C. henryi remains to be our further study.

Plant Materials, Growth Conditions and Treatment
The C. henryi "Huali 4" cultivar was obtained from the Central South University of Forestry and Technology (28°818′′N, 113°015′′E), Hunan Province, China. Catkins were selected from three 11-year-old trees that were planted with 3 m × 2.5 m spacing, pruned and fertilized every year in December.
A certain amount of CPPU (5 mg CPPU for 5 mg·L −1 , 25 mg CPPU for 25 mg·L −1 and 125 mg CPPU for 125 mg·L −1 ) was dissolved in 25 mL absolute ethanol to prepare a plant hormone stock solution. Stock solutions were diluted with 1 L of pure water to prepare working solutions of different concentrations. As shown in Figure 1a, the catkins were treated with working solutions one week after the buds sprouted for the first time, when the length of the bud was approximately 1.5 cm (March 22). Three concentrations of CPPU, A (5 mg·L −1 ), B (25 mg·L −1 ) and C (125 mg·L −1 ), were applied at three different periods (the interval is one week in 2021) with a 500 mL plastic sprayer to the C. henryi catkins. At each concentration, CPPU was applied to the C. henryi catkins in seven different ways, and each treatment is represented as A1, B1, C1, …, C7. Three biological replicates were used for each treatment, and the control group CK was treated with pure water instead of hormones.

Plant Materials, Growth Conditions and Treatment
The C. henryi "Huali 4" cultivar was obtained from the Central South University of Forestry and Technology (28 • 818 N, 113 • 015 E), Hunan Province, China. Catkins were selected from three 11-year-old trees that were planted with 3 m × 2.5 m spacing, pruned and fertilized every year in December.
A certain amount of CPPU (5 mg CPPU for 5 mg·L −1 , 25 mg CPPU for 25 mg·L −1 and 125 mg CPPU for 125 mg·L −1 ) was dissolved in 25 mL absolute ethanol to prepare a plant hormone stock solution. Stock solutions were diluted with 1 L of pure water to prepare working solutions of different concentrations. As shown in Figure 1a, the catkins were treated with working solutions one week after the buds sprouted for the first time, when the length of the bud was approximately 1.5 cm (March 22). Three concentrations of CPPU, A (5 mg·L −1 ), B (25 mg·L −1 ) and C (125 mg·L −1 ), were applied at three different periods (the interval is one week in 2021) with a 500 mL plastic sprayer to the C. henryi catkins. At each concentration, CPPU was applied to the C. henryi catkins in seven different ways, and each treatment is represented as A1, B1, C1, . . . , C7. Three biological replicates were used for each treatment, and the control group CK was treated with pure water instead of hormones.

Sample Collection
Two weeks after the third spraying, when the carpel and stamen primordia began to differentiate [19,57] (April 20), the male catkins of treatment groups CK, A7, B7 and C7 (see Table 5 for details) were collected as samples and were named CH_0, CH_5, CH_25, and CH_125, respectively. They were then snap-frozen in liquid nitrogen and were stored in a −80 • C freezer for endogenous hormone concentration determination, transcriptome sequencing and quantitative real-time polymerase chain reaction (RT-qPCR).  Total RNA was extracted using the Plant Total RNA Isolation Kit (Omega, China), and the concentrations and purities of RNA were detected using a UV spectrophotometer and NanoDrop 2000c spectrophotometer (Thermo Fisher, Waltham, MA, USA). A sequencing library construction for each sample with three biological replicates was completed by Igenebook Biotechnology Co., Ltd. (Wuhan, China).

Determination of Endogenous Hormones
The levels of endogenous hormones iP, iPR, tZ, tZR, GA 1 , GA 4 , IAA, ABA and JA in the male catkin of C. henryi in each treatment were determined using high-performance liquid chromatography-mass spectrometry (HPLC-MS) [58]. An Agilent 1290 HPLC system (Agilent) and a Qtrap 6500 mass spectrometer (Sciex) were used for quantification. Standards were purchased from Sigma-Aldrich. A reversed-phase poroshell 120 SB-C18 column (2.1 × 150 mm, 2.7 µm) was used as the stationary phase, and the mobile phase was a solution of A: B = (methanol/0.1% formic acid): (water/0.1% formic acid). The column temperature was 30 • C, and the sample volume was 2 µL. The ionization mode was negative. Other parameters detected by mass spectrometry were: curtain gas pressure, 15 psi; spray voltage, −4000 V; nebulizing gas pressure, 65 psi; auxiliary gas pressure, 70 psi; and nebulizing temperature, 400 • C.

Illumina Transcriptome Sequencing and Assembly of Clean Reads
Base calling was used to convert the raw image data files obtained by high-throughput sequencing into raw sequenced reads, namely raw reads or raw data. The original data were filtered and processed to obtain clean reads as follows [59]: the clean reads data were aligned to the reference genome using HISAT2 software [60] (version: 2.0.1-beta); the transcripts of all samples were reconstructed from the alignment information to obtain total reads using StringTie software [61] (version: 2.0.4); and the number of reads of the gene was normalized by the fragments per kilobase of exon per million reads mapped (FPKM) normalization method [62,63]. The genome was divided into five regions: CDS, 5UTR, 3UTR, intron and intergenic, and the total reads aligned to the genome were counted according to the functional elements. Most RNA-seq reads fall within the CDS region under normal conditions. Based on the expression level of each sample gene, the correlation between samples was calculated using the Pearson correlation coefficient. Genes with FDR < 0.05 and |FoldChange| > 2 were considered significantly different genes. Differential expression analysis was performed using edgeR [64]. The expression levels of significantly different genes in all samples were extracted into a list, and the Pheatmap function was used to draw a heat map of the DEGs.
To quantify the DEGs, the FPKM method was used to calculate read counts as the number of fragments per kilobase of the transcript [70].

Validation of DEGs by RT-qPCR
To confirm the Illumina sequencing results, 10 candidate genes were randomly selected, and the expression of the DEGs in three different samples was verified using RT-qPCR. cDNA for RT-qPCR was synthesized according to the instructions of HiScript II Q RT SuperMix (Vazyme Biotechnology, Nanjing, China). The total volume of the qPCR reaction system was 20 µL, including 10 µL of 2 × ChamQ Universal SYBR qPCR Master Mix (Vazyme Biotechnology, Nanjing, China), each with 0.4 µmol/L of forward and reverse primers, 1 µL of ten-fold diluted cDNA template and 8.2 µL ddH 2 O. The PCR conditions were as follows: denaturation at 95 • C for 30 s; 40 cycles of denaturation at 95 • C for 10 s; and annealing and extension at 60 • C for 30 s. Gene expression was calculated using the 2 −∆∆Ct method [71]. The primer sequences of the reference gene (GAPDH) and 9 selected genes are shown in Table 2.

Statistical Analysis
Data were analyzed using Statistical Product and Service Solution (SPSS) software (version 13.0; SPSS, Chicago, IL, USA). The means were compared using Student's t-test at the 5% significance level. Three replicates were performed for each treatment. Figures were generated using OriginPro (Microcal Software Inc., Northampton, MA, USA).

Conclusions
In conclusion, this study demonstrates that the exogenous administration of CPPU has a strong feminization effect on C. henryi. Transcriptome analysis of the treatment group, in which male flowers were converted to female flowers after CPPU treatment, helped to identify candidate genes that determine sex regulation. As was expected, these candidate genes have only been shown to be related to the sex of flowers in some plants, and whether they have the same effect in C. henryi requires further study. This study provides a basis for further understanding the molecular mechanisms of CK regulation in sex determination in C. henryi.