Downstream of GA4, PbCYP78A6 Regulates Parthenogenesis by Mediating Cell Cycle-Related Genes in Pear (Pyrus Bretshneider Rehd.)

Haiqi Zhang Northwest A&F University: Northwest Agriculture and Forestry University Wei Han Northwest Agriculture University: Northwest Agriculture and Forestry University Huibin Wang Northwest Agriculture University: Northwest Agriculture and Forestry University Liu Cong Northwest Agriculture University: Northwest Agriculture and Forestry University Rui Zhai Northwest Agriculture University: Northwest Agriculture and Forestry University Chengquan Yang Northwest Agriculture University: Northwest Agriculture and Forestry University Zhigang Wang Northwest Agriculture University: Northwest Agriculture and Forestry University Lingfei Xu (  lingfxu2013@sina.com ) Northwest Agriculture University: Northwest Agriculture and Forestry University

exposure. The ectopic overexpression of PbCYP78A6 contributed to parthenocarpic fruit production in tomato. The PbCYP78A6 expression coincided with fertilized and parthenocarpic fruitlet development and the expression of fruit development-related genes as assessed by cytological observations and RT-qPCR, respectively. PbCYP78A6 RNA interference and overexpression revealed that the gene is an upstream regulator of fruit development-related genes in pear.
Conclusions: Our ndings indicate that PbCYP78A6 plays a critical role in cell proliferation and provide insights into controlling parthenocarpy.

Background
Angiosperms have evolved double-fertilization processes, which require coordinated communication between gametophytic and sporophytic tissues, and fruit development as pivotal steps of their survival and dispersal strategies [1][2]. Fruit initiation requires successful pollination and fertilization, but parthenocarpy uncouples this synchronized association and triggers fruit development [3].
Phytohormones are thought to be triggers induced by fertilization. Many strategies to produced virgin fruit include the exogenous application or overproduction of plant hormones, particularly auxins and gibberellins (GAs) [4][5][6], as well as the mutation of speci c genes in these two plant hormone signaling pathways [7][8]. Auxins partially act upstream of GAs in inducing fruit set [9]. Complex mechanisms involving different hormones have been revealed, but there is limited knowledge regarding the mechanisms that underlie parthenocarpy.
Parthenocarpy results in traits attractive to both consumers and breeders. It also overcomes the obstacle of self-incompatibility in the fruit set of horticultural crops, including pear (Pyrus bretshneider) [3]. The eshy fruit of pear is derived from the hypanthium and is known as an accessory fruit. Owing to the traits of perennial fruit trees, knowledge regarding the mechanism behind parthenocarpy in pear is still limited.
Overexpressing CYP78A9, a member of cytochrome P450 78A subfamily, identi ed by a transfer DNA activation-tagging screen, produces the parthenocarpic phenotype in Arabidopsis thaliana [15]. CYP78A6 acts redundantly with CYP78A9 to control reproductive organ development [16]. To date, the catalytic function of CYP78A enzymes remain largely unknown, but the expression patterns and effects of related genes have been widely elucidated [17][18][19]. Overexpressing members of the CYP78A family increases non-autonomous cell proliferation and the sequential formatting of large organs [20][21]. The growth and development of multicellular organisms is controlled by cell-cycle progression, which is mediated by the periodic activation of complexes containing cyclins (CYCs) and cyclin-dependent kinases (CDKs) [22].
The functions of CYP78A family members and their roles in parthenocarpy have not been reported, nor have correlations between their expression levels and cell proliferation-related genes in pear.
In this work, we provide a detailed description of PbCYP78A6's functions in pear parthenogenesis. We produced matured parthenocarpic fruit from pollination-dependent and GA 4 -induced 'Dangshansu' pear.
Using comparative transcriptome and qRT-PCR analyses, we determined that PbCYP78A6 expression was intimately correlated with fruit set and development. Stable transgenes in tomato demonstrated the contribution of PbCYP78A6 to parthenogenesis. The overexpression and RNA interference of PbCYP78A6 in pear calli revealed that the gene is an upstream activator of cell proliferation. Thus, PbCYP78A6, induced by GA 4 , regulates parthenogenesis by mediating cell proliferation-related genes.

Results
A transcription analysis identi ed a cytochrome P450 gene, PbCYP78A6, potentially responsible for parthenocarpy The GA 4 treatments resulted in a relatively high fruit set rate of 93.2% and induced seedless fruits with decreased weights and increased fruit indices and hardness levels compared with seeded fruits, whereas GA 3 treatments induced a relatively low fruit set rate of 56.8% and did not induce mature fruit formation (Table 1; Additional le 1: Fig. S1). To further analyze the molecular mechanisms underlying parthenocarpy, an RNA-seq analysis was used to identify potential related genes. PbCYP78A6 (LOC103964254) was commonly up-regulated in pollinated and GA 4 -treatment groups (Additional le 2: Fig. S2; Additional le 3: Supplementary Table S1).
A phylogenetic analysis demonstrated that PbCYP78A6 in P. bretshneider displayed very high similarity levels to AtCYP78A9 and AtCYP78A6 (Fig. 1). PbCYP78A6 was highly expressed in the sepal, which is an important component of pear fruit (Fig. 2A). During the early process of fruit set, PbCYP78A6's expression level signi cantly increased in the effective treatment groups, including hand-pollinated and GA 4 -treated ( Fig. 2B). We separated the developing fruit into hypanthium, ovary wall and ovule (Fig. 2C). A relatively high PbCYP78A6 expression level was detected in the hypanthium compared with the ovary wall and ovule (Fig. 2D). We speculated that PbCYP78A6 expression correlated with fruit development and parthenogenesis in pear.
PbCYP78A6 overexpression contributed to parthenocarpic fruit development in tomato To determine the potential roles of PbCYP78A6 in fruit development and parthenogenesis, transgenic tomato lines overexpressing the PbCYP78A6 gene were obtained (Additional le 4: Fig. S3). Under natural pollination conditions, transgenic tomato overexpressing PbCYP78A6 had signi cantly reduced numbers of seeds per fruit, and most of the fruits produced were seedless ( Fig. 3A). At the mature stage, there were a few large seeds along with tracks of underdeveloped seeds in the transgenic fruits (Fig. 3B).
Observations and a statistical analysis of seeds determined that overexpressing PbCYP78A6 reduced the number of seeds per fruit (Fig. 3C). Seeded fruits were produced by pollinated ovaries in wild-type (WT) lines (Fig. 3D). The ovaries of PbCYP78A6-overexpression (OE) lines continued to develop and produce parthenocarpic fruits after being emasculated (Fig. 3D, E), whereas WT lines did not produce fruits without pollination. Thus, overexpressed PbCYP78A6 was involved in the regulation of fruit development and parthenocarpy.
Cell division and expansion were induced as PbCYP78A6 expression increased CYP78A6 overexpression results in the production of large fruit owing to increased cell proliferation [16,18]. To explore the cellular changes after PbCYP78A6 expression was highly induced, fruitlets were embedded in para n and sectioned. Phenotypic observations of the fruitlet sections were recorded (Fig. 4). The ovaries of the pollinated and GA 4 -treated groups were larger than those of the un-pollinated and GA 3 -treated groups ( Fig. 4A-D). The thicknesses of the calyxes in GA 4 -treated and pollinated samples were signi cantly greater than those of the un-pollinated and GA 3 -treated groups ( Fig. 4E-K).
Compared with the un-pollinated treatment, hand pollination and GA 4 exposure induced increased celllayer production and cell-area enlargement (Fig. 4J, K). Thus, the increased PbCYP78A6 expression activated cell division and expansion.
The expression of fruit development-related genes was activated during parthenogenesis To further understand the molecular mechanisms underlying parthenogenesis, the dynamic growth and correlated transcript levels of cell expansion-and division-related genes were observed in 'Dangshansu' fruitlets at an early stage. Notably, GA 4 -treated and pollinated fruitlets underwent rapidly growth, whereas GA 3 -treated and un-pollinated fruitlets did not show signi cant growth-related changes (Fig. 5A).
Intriguingly, the PbCYP78A6 expression pattern was almost consistent with the those of fruitlet development-related genes (Figs. 2 and 5). We speculated that PbCYP78A6 acts upstream of cell proliferation-related genes to regulate fertilized and parthenocarpic fruit growth.
PbCYP78A6 was involved in the regulation of fruit development-related genes To further characterize the correlations between PbCYP78A6 and parthenogenesis-related genes, transgenic PbCYP78A6-OE and -RNA interference (RNAi) calli were generated. Green uorescent protein (GFP) signals were detected in PbCYP78A6-OE and -RNAi lines (Fig. 6A). RT-qPCR was used to identify changes in PbCYP78A6 expression (Fig. 6B, C). In PbCYP78A6-RNAi calli, PbExpA4 and PbCDKI6L expression levels were extremely signi cantly increased (Fig. 6B), and the PbCDKB22, PbCDKB22L, PbCCNB2L and PbCyclinA24 transcript levels were also increased ( Fig. 6B; Additional le 5: Fig. S4), compared with in WT. However, in PbCYP78A6-OE calli, the positive regulators of cell proliferation, including PbExpA4, PbCDKB22, PbCDKB22L and PbCyclinA24, were signi cantly up-regulated, while a negative regulator of cell proliferation, PbCDKI6L, was down-regulated compared with in WT ( Fig. 6C; Additional le 5: Fig. S4). Thus, PbCYP78A6 regulated cell proliferation-related genes involved in fruit growth.

Discussion
In higher plants, an ovary can either be fertilized and subsequently grow into a fruit, or, owing to the lack of successful fertilization, it can enter the abscission process. Exogenous applications of multiple hormones mimic fertilization functions and fruit set, independently of fertilization, resulting in parthenocarpy [9,23]. In 'Dangshansu' pear, GA 4 is effective in inducing parthenocarpy [10][11][12]24], whereas GA 3 does not induce parthenocarpy. Here, the difference between GA 4 -and GA 3 -induction in pear was assessed to investigate the mechanisms of parthenocarpy-related downstream hormones.
In a comparison of transcriptomes between pollination-dependent and parthenocarpic fruit set, PbCYP78A6 was identi ed as a candidate gene involed in the regulation of parthenogenesis. PbCYP78A6 is similar to AtCYP78A6 and AtCYP78A9 ( Fig. 1), but its expression pattern differed from those of AtCYP78A6 and AtCYP78A9 [16,21]. PbCYP78A6 was determined to be widely expressed in reproductive organs, but it was highly expressed in the sepal ( Fig. 2A), which partly develops into an accessory fruit in pear. PbCYP78A6 expression was signi cantly activated by GA 4 and pollination, and the expression level increased as the fruit set and developmental processes continued (Fig. 2B), which is consistent with AtCYP78A9 activation being detected in developing seeds, septa, funiculi and placental tissues following fertilization [21]. Overexpressing AtCYP78A6 or AtCYP78A9 promotes the growth of not only seeds, but also the reproductive tissues, including sepals and siliques [16,21], which indicates that AtCYP78A6 and AtCYP78A9 control their development. In pear, the cells of the exocarp are capable of cell division, in which new cell layers arise owing to periclinal cell division. In our study, PbCYP78A6 expression was rarely detected in ovules, but occurred at a high level in the pericarp and hypanthium, in which cell division occurs vigorously (Fig. 2C, D). This indicated that PbCYP78A6 is involved in pear fruit development. Thus, the different structures between fruit and silique hint at an important role for PbCYP78A6 in parthenogenesis.
Under normal pollination conditions, the owers of PbCYP78A6-OE transgenic plants produced fruits with reduced seed numbers and even some seedless fruits (Fig. 3). This reproductive organ phenotype was similar to that seen in Arabidopsis overexpressing AtCYP78A6 [16]. The auxin production and GAmediated responses of ovules are dependent on pollination events that trigger fruit development [25]. Fruit development occurred normally in transgenic lines having low numbers of, and even no, seeds  (Fig. 3D, E). In Arabidopsis, large and seedless fruit have also been induced by the overexpression of CYP78A9 [18]. Here, we report that like AtCYP78A9, PbCYP78A6 was capable of inducing parthenocarpic fruit set and development in tomato.
Silencing the Arabidopsis CYP78A6/EOD3 homolog PaCYP78A6 decreases fruit size by affecting cell proliferation [19], indicating that PaCYP78A6 acts upstream of cell division and expansion. Thus, the parthenocarpy produced by overexpressing PbCYP78A6 may also arise from the activation of cell division and expansion. The differential expression of PbCYP78A6 was consistent with the signi cantly increased cell division and expansion that occurred in the tissue-containing calyx and mature ovary compared with those of the un-pollinated group (Fig. 2). The process of fruit set is accompanied by cell division and expansion [26], indicating that PbCYP78A6 plays an essential role in fruit formation.
Fruit development is largely dependent on cell division and expansion, and cell division is governed by two gene families, CDKs and CYCs [27]. Two other families, EXPA and EXPB, have the ability to regulate cell expansion by extending cell walls [28]. CDKBs, CYCAs and EXPAs participate in regulating fruit development [29][30][31]. In accordance with our previous study [10,12], the expression levels of selected cell division-and expansion-related genes were analyzed during the early fruit set stage (Fig. 5). Among them, PbExpA4, PbCyclinA2-4, PbCCNB2L, PbCDKB22 and PbCDKB22-like had up-regulated expression pattrens, which agreed with the histological observations (Fig. 4). Moreover, a CDK inhibitor gene, CDKI6L, was identi ed, and its expression decreased in GA 4 -treated fruitlets. However, the GA 3 treatment failed to repress the expression of CDKI6L. ICK1 and ICK2 expression reduce the CDK activity and affect cell division in A. thaliana [32], indicating that the inhibition of CDKI6L also plays an important role in fruit set. Thus, these genes are important for pear fruit development. Intriguingly, PbCYP78A6 expression was almost consistent with the expression patterns of PbEXPA4, PbCyclinA2-4, PbCCNB2L, PbCDKB22 and PbCDKB22-like, and it was negatively correlated to the expression of the repressor CDKI6L (Figs. 2B and 4B).
PaCYP78A9's effect on plant organ size is regulated by cell cycle-related genes [19], whereas PbCYP78A6's effect on fruit development was mediated by speci c fruit development-related genes in pear. In PbCYP78A6-RNAi pear calli, silencing PbCYP78A6 did not halt cell proliferation ( Fig. 5B; Additional le 5: Fig S4A). Overexpressing PbCYP78A6 promoted the expression of fruit growth-related genes and particularly repressed PbCDKI6L expression in pear calli ( Fig. 5C; Additional le 5: Fig S4B). The PbCDKB22, PbCDKB22-like, PbCyclinA2-4 and PbCCNB2L expression levels were negatively correlated with the PbCYP78A6 expression level. Similarly, silencing PaCYP78A6 or PaCYP78A9 does not completely repress the expression of all the cell proliferation-related genes [19,33]. The evidence indicates that there must be other factors involved in controlling cell proliferation. We also identi ed a CDK repressor, PbCDKI6L, which was increasingly induced as PbCYP78A6 expression decreased (Fig. 5).
The reduced CDK activity was attributed to the increased ICK1 expression, which represses cell division [32]. Thus, PbCDKI6L plays a key role in the regulation of PbCYP78A6 expression in fruit development.
Cell-wall loosening is critical for rapid cell division, and it is most often controlled by EXPs [34]. Both silencing and overexpressing PbCYP78A6 signi cantly promoted PbExpA4 expression, perhaps owing to the consequences of cell proliferation. The effects of PbCYP78A6 on cell cycle-related genes and parthenogenesis were con rmed, but the catalytic function of the PbCYP78A6 enzyme remains largely unknown. Thus, our ndings indicate that PbCYP78A6 acts as an upstream regulator of fruit development during parthenogenesis (Fig. 7).

Conclusions
PbCYP78A6 was induced by pollination and GA 4 treatments, and its overexpression resulted in parthenocarpic tomato. The effects of PbCYP78A6 on fruit development were mediated by cell cycledrelated genes. The utilization of CYP78A6 by fruit trees and its altered expression in fruits might provide a method for producing seedless fruits and enhancing crop yields (Fig. 7). The further elucidation of unknown plant growth-related substances may contribute to their utilization in regulating fruit development.

Experimental treatments and samples
Two days before anthesis, all of the treatments and the control were bagged to avoid pollination. To avoid the disturbing fruit set caused by natural pollination, the opened owers and the weak buds were removed to make sure every single pear ower keeping consistent and unopened when the exogenous hormones were applied. GA 3 solutions 50 mg L − 1 and GA 4 solutions of 50 mg L − 1 were independently sprayed on individual un-pollinated owers of 'Dangshansu' pears at anthesis. 1:1 water: ethanol mixture was sprayed on un-pollinated owers as the un-pollinated treatment. Three branches of each treatment were used as three replicates. Fruitlets with complete structure per treatment were randomly sampled. Fruits at 0, 3, 4, 6, 9, 30 and 145 DAA were harvested for morphological observation and other experiments, respectively. After removing the stalks, sepal, stamens an gynoecium, fruitlets left sampled at 4 and 6 days after anthesis (DAA) were used for RNA sequencing and immediately xed in formaldehyde-acetic acid-alcohol for para n section, respectively. Fruitlets at 0, 3, 6, 9 DAA were used for the analysis of gene expression patterns with the same dispose as above. Fruitlets with hand pollination at 3 DAA were subdivided into fruit stem, sepal, petal, stamen, stigma, ovary, and ovule for tissue-speci c expression. Fruits at 30 DAA were subdivided into pericarp, hypanthium, ovary wall and ovule for the expression position of PbCYP78A6 in fruits. Each sample was harvested from 12 fruitlets, pooled, directly frozen in liquid nitrogen and stored at − 80°C.

Determination of fruit set rate
A total of 30 blooms on each three branches was labeled and bagged immediately after receiving treatments. At 20 DAA, the bags were removed. The formula used to calculate the fruit-set rate was as follows: Fruit set (%) = (number of fruitlets remaining/number of total owers counting) × 100.

Para n sectioning and data statistics
To conduct the histological observations, fruit samples of four treatments in 'Dangshansu' pear were collected at 6 DAA when the signi cant changes were observed, immediately xed in formaldehydeacetic acid-alcohol xative [35] and stored at 4 ℃. The ovaries were dehydrated in an ethanol/xylene series and embedded in para n. They were then cut into 8-µm-thick slices, dried and stained with safranin and fast green. The anatomical images were observed using a microscopic imaging system (BX51 + PD72 + IX71, OLYMPUS, Japan). Cell area and calyx thickness were calculated using Image J software (https://imagej.net/Welcome). Cell area calculation method was as follows: rst, a certain area was randomly circled and the total cell area was calculated; second, the total number of cells was counted; and third, the representative cell area of a single cell was calculated as the total cell area divided into the total number of cells. All calculations were performed on more than three replicates.

Transcription analysis
To further understand the potential molecular mechanism underlying parthenocarpy, Samples from unpollinated, hand pollinated, GA 3 -treated (without pollination) and GA 4   cDNA was synthesized by the reversely transcription of 1µg total RNA using a PrimeScript RT reagent kit with gDNA Eraser (Takara, Dalian, China). Primers of target genes were designed by Primer Premier 5.0 software (PREMIER Biosoft) and NCBI Primer-BLAST tools (https://www.ncbi.nlm.nih.gov/tools/primerblast/index.cgi). The Actin 7 gene was used as the internal reference for the gene expression analysis.

Phylogenetic analysis
The full-length CDS of PbCYP78A6 (LOC103964254) was isolated from 'Dangshansu' pear (http://www.ncbi.nlm.nih.gov/genome/?term = pyrus). Amino acid sequences of PbCYP78A6 and other CYP78A subgroup members of other plants were aligned using ClustalW [38]. MEGA5.10 was applied to construct the phylogenetic tree with the neighbor-joining statistical method. In addition, 1000 bootstrap replications were performed for testing of phylogeny [39].

Construction of plasmids and generation of transgenic pear calli
To generate transgenic pear calli, the coding region of PbCYP78A6 was cloned using primers PbCYP78A6-OE F/R and introduced into the vectors pGWB414 vector based on gateway recombination technology (Invitrogen) to create an overexpression vector. The vectors pHellsgate2 and pK7WIWG2D were used as RNAi-mediated vectors for silencing PbCYP78A6 as described [40,41]. The vectors were transformed into Agrobacterium tumefaciens strain EHA105 for in ltration.
The induction of pear calli and their transformation were described previously [42]. Brie y, the prefabricated suspension of EHA105 was incubated with 15-day-old pear calli for 15 min. After co-

Production of transgenic lines in tomato
The full-lenth PbCYP78A6 (LOC103964254) coding sequence (CDs) was isolated from 'Dangshansu' cDNA using primers with adaptors (Additional le 9: Supplementary Table S4) designed by Snap Gene software 1.1.3, and then cloned into the BamH I and SacI restriction enzyme sites downstream of a Cauli ower mosaic virus (CaMV) 35S promoter in the pBI121 vector to generate a overexpression construct using a ClonExpress One Step Cloning kit (Vazyme, Nanjing, China). The recombinant vect was transformed into Agrobacterium tumefaciens strain LBA4404 by heat-shock method, and the positive monoclonal Agrobacterium cell proliferated at 28°C, 200 rpm, dark in Luria-Bertani (LB) solid medium containing appropriate antibiotics (kanamycin and rifampicin). After incubation for appropriate time, the Agrobacterium cell concentration achieving to OD 600 ≈ 0.5-0.8 were centrifuged at 600 rpm, collected in the tube, and then suspended in MS isopyknic to LB solution. The bacterial suspension was used for tomato transformation.
Transformation assays were carried out as previously described [43]. Brie y, sterilized tomato seeds were grown until its cotyledons full stretched in glass jars. Cotyledons were cut into sections and placed on Petri dishes containing solidi ed MS1 medium for 2 days in the dark, Then immersed them in bacterial suspensions prepared above for 10 min. Explants impregnated with bacterial suspension were blotted with lter paper, then cultured in the dark for 2 days in MS1 medium. Then the explants were transferred to MS2 solidi ed medium. After explants developed resistant calli produced shoots, 1-2 cm shoots were excised and placed on MS3 medium in glass jars. After took root, explants with root were cultured in pots containing vermiculite, watered with Hoagland's solution, and conditioned in a growth chamber before transferring to the greenhouse. Progeny from the transgenic plants were obtained by sel ng under controlled conditions.

Declarations
Ethics approval and consent to participate: Not applicable.

Consent for publication: Not applicable
Availability of data and materials: Permissions for all the materials used in this experience had been obtained. No further permission was therefore needed. All data generated and analyzed during this study are included in this published article. Extra data has been appended as supplementary Tables. All the genes' sequence and information can be accessed on National Center of Biotechnology Information (https://www.ncbi.nlm.nih.gov/).
Competing interests: The authors declare no con ict of interest.
Funding: This research was funded by the National Key R&D Program of China (2019YFD1001400) and Weinan Experimental Station foundation of Northwest A&F University. Apart from providing nancial support, funding agencies played no role in the design of the studies, the collection, analysis and interpretation of the data, and the writing of the manuscripts.  The results are represented as means of three biological replicates (± SDs). Signi cant differences (P < 0.05) among treatments as determined by a one-way ANOVA are indicated using different lowercase letters.  Signi cant differences (P < 0.05) among treatments as determined by a one-way ANOVA are indicated using different lowercase letters.   Model depicting the role of PbCYP78A6 in controlling fruit development-related genes to regulate parthenocarpy.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.