4-CPA (4-Chlorophenoxyacetic Acid) Induces the Formation and Development of Defective “Fenghou” (Vitis vinifera × V. labrusca) Grape Seeds

For some horticultural plants, auxins can not only induce normal fruit setting but also form fake seeds in the induced fruits. This phenomenon is relatively rare, and, so far, the underlying mechanism remains unclear. In this study, “Fenghou” (Vitis vinifera × V. labrusca) grapes were artificially emasculated before flowering and then sprayed with 4-CPA (4-chlorophenoxyacetic acid) to analyze its effect on seed formation. The results show that 4-CPA can induce normal fruit setting in “Fenghou” grapes. Although more seeds were detected in the fruits of the 4-CPA-treated grapevine, most seeds were immature. There was no significant difference in the seed shape; namely, both fruit seeds of the grapevines with and without 4-CPA treatment contained a hard seed coat. However, the immature seeds lacked embryo and endosperm tissue and could not germinate successfully; these were considered defective seeds. Tissue structure observation of defective seeds revealed that a lot of tissue redifferentiation occurred at the top of the ovule, which increased the number of cell layers of the outer integument; some even differentiated into new ovule primordia. The qRT-PCR results demonstrated that 4-CPA application regulated the expression of the genes VvARF2 and VvAP2, which are associated with integument development in “Fenghou” grape ovules. Together, this study evokes the regulatory role of 4-CPA in the division and continuous redifferentiation of integument cells, which eventually develop into defective seeds with thick seed coats in grapes.


Introduction
One of the most obvious phenomena of parthenocarpy is the formation of completely seedless fruits; for instance, an auxin application can induce aborted seeds in tomatoes, cucumbers, and pears [1,2]. Auxin was the first plant hormone to be discovered, and it has important regulatory effects on plant cell division and differentiation, lateral root formation, tropism, flowering, and embryonic development [3][4][5]. In particular, auxin accompanies another plant hormone, gibberellin (GA; both synthesized and accumulated in the fertilized ovaries after pollination/fertilization), both of which modulate the plant fruit set [6,7]. The current model suggests that GA biosynthesis is under the control of auxin in the early stages of fruit development; for instance, exogenous auxin application induces GA biosynthesis, whereas GA treatment does not activate the auxin response in Arabidopsis and tomato [8,9]. Furthermore, the exogenous application of those hormones on ovaries or genetic mutations in genes involving hormone synthesis (such as overexpression of genes of auxin biosynthesis or silencing of the gene DELLA, which have been considered master negative regulators of GA signaling) can induce pollination-independent fruit set [6,10,11].

Plant Material
The hybrid grape cultivar "Fenghou" (V. vinifera × V. labrusca) was planted in a nursery at Wenquan Town, Beijing, the People's Republic of China. All the treatments were conducted over three consecutive years and started with 15-year-old grapevines.

Pharmacological Treatments
The synthetic pesticide 4-chlorophenoxyacetic acid (4-CPA, Sigma-Aldrich, St. Louis, MS, USA), which can modulate grape growth, was employed to induce defective seed generation in this study, with a concentration of 0.08 mM [14]; 4-CPA was dissolved in a solution of 94.9% water, 5% ethanol, and 0.1% Tween 80. Based on different treatments, these inflorescences were mainly divided into the emasculated group (30 inflorescences) and the nonemasculated group (10 inflorescences). In the emasculated group, 20 inflorescences were sprayed with 0.08 mM 4-CPA, and the remaining 10 inflorescences were treated with solvent alone. The 10 inflorescences of the nonemasculated group were used as the controls.
The inflorescence samples were respectively collected 4,8,12,16, and 21 days after treatment (DAT) and divided into 3 groups (each group contained the same number of inflorescences): Group I was fixed with FAA (formaldehyde-acetic acid-ethanol fixative; 50% ethanol, 5% glacial acetic acid, and 5% formaldehyde), and used to make paraffin sections for microscopic observation; Group II was immediately frozen in liquid nitrogen after collection and stored at −80 • C for RNA extraction; Group III was prepared for in situ hybridization and fixed with PBS that contained 4% (w/v) formaldehyde at 4 • C.

Organizational Structure Observation and Statistical Analysis
The ovaries or young fruits of Group I, which have been fixed by FAA, were firstly dehydrated via passing through an ethanol series and then embedded in paraffin. These samples were cut into 8-µm sections and stained with Safranin and Fast green. An Olympus CX31 microscope was used to observe and photograph the sample's organizational structure.
ImageJ software was used to measure the ovule integument cell diameter and thickness in the ovary of young fruit at 4, 8, 12, and 16 DAT.

RNA Extraction and Quantitative Real-Time PCR Analysis
The total RNA was extracted from "Fenghou" grape ovaries/fruits of Group II using a modified CTAB method, and they were purified with RNase-free DNase I (Takara, Kyoto, Japan). An atomic ultraviolet spectrophotometer was used to determine RNA concentration. First-strand cDNA synthesis was initiated from 2 µg total RNA using reverse transcriptase (Promega, Madison, WI, USA) and oligo dT primers according to the manufacturer's instructions. We performed quantitative real-time PCR (qRT-PCR) with the cDNAs on an ABIPRISM 7500 based on the manufacturer's instructions. The reaction conditions were described previously [14]. All experiments were performed in three replicates and two technical replicates. We used the Primer software 5.0 (PREMIER Biosoft International, Palo Alto, CA, USA) to design primers for VvAGL11, VvTTG2, VvAP2, and VvARF2. The relative transcription levels were calculated using the 2 −∆∆Ct method [33] and normalized to the relative transcript levels of the housekeeping gene VvUBQ as an internal standard. All quantitative primers are listed in Table S1.

In Situ Hybridization
Probes for VvAP2 and VvARF2 were PCR-amplified from cDNA templates with gene-specific primers. We sequenced the PCR product and found it was composed of 500 bp of VvAP2 and 500 bp of VvARF2. The DNA fragment showed 99% similarity to the published sequence. The PCR product was ligated into the pMD-18 vector (Takara Bio, Beijing, China). According to the manufacturer's instructions, antisense probes were synthesized using the DIG RNA labeling kit (Sp6/T7; Roche, Basel, Switzerland). As de-scribed by Drews et al. (1991), digoxigenin labeling was used for RNA probe labeling, tissue preparation, and in situ hybridization [34].

Statistical Analysis
Kruskal-Wallis tests and Student's t-tests were conducted to assess statistical significance among different treatments. Significant differences between the means were based on the least significant method when the F-values were significant (p < 0.05).

4-CPA Inhibits Fruit Growth But Not Fruit Setting
To explore the regulatory role of 4-CPA on "Fenghou" grape fruit setting and fruit growth, we used exogenous 4-PCA sprayed on the grape ovary at 8 DBA. At 28 DAT, the ovary diameter of 4-CPA-treated grapevines and control grapevines were 8 and 15 mm, respectively. When grapevines were emasculated but not sprayed with any 4-PCA, the development of the ovaries ceased (Figure 1a). A similar fruit setting rate was detected in 4-CPA-treated grapevines and control grapevines: 23.87% and 26.33%, respectively. In agreement with our observations on ovary development, the untreated fruit setting rate after emasculation was 0 ( Figure 1b). On the side of fruit growth and development, the control fruits started to be colored at 60 DAT and were fully colored and mature at 80 DAT. Compared with control fruits, the 4-CPA-treated fruits were smaller after 20 DAT, and the fruit shape was oblate; they were not colored normally until 80 DAT (Figure 1c). The fruit growth curve also revealed that the horizontal and vertical diameters of the control fruits were both larger than 4-CPA-treated fruits from 20 DAT. Before 16 DAT, the horizontal diameter of the 4-CPA-treated fruit was larger than that of the control fruit, but the vertical diameter of the fruit was not significantly different between both (Figure 1d,e). Furthermore, the development of 4-CPA-treated fruits was almost stagnated after 44 DAT. Additionally, there were no significant differences in fruit size at 80 DAT. The average single-fruit weight of 4-CPA-treated fruits was 1.83 g, which was much lower than the control fruit (10.1g) (Figure 1f). The fruit shape index after the 4-CPA treatment was lower than that of the control fruit ( Figure 1g). Thus, our results revealed that the 4-CPA was able to reduce fruit growth after 20 DAT but had no impact on fruit set.  Significance analysis was conducted with two-tailed Student's t-tests (*** p < 0.001; **** p < 0.0001).

The Effect of 4-CPA Treatment on Seed Growth and Development
In addition to inhibiting fruit growth, surprisingly, seed-like products were observed in 4-CPA-treated fruits. These products looked similar to normal seeds, albeit being smaller and in larger quantities (Figure 2a,b). For instance, 92.86% of control fruit only contained 2 seeds, 3.57% had 3 or 4 seeds, and no fruit contained more than 4 seeds. In contrast, more seed-like products were found in 4-PCA-treated fruits: the respective percentage of 2, 3, 4, 5, 6, and 7 seed-like products was 0, 7.69%, 42.31%, 19.23%, 26.92%, and 3.85%, respectively ( Figure 2c). The average number of seeds or seed-like products The scale bar of (a,c) = 1 cm. Abbreviations: 4-CPA, 4-CPA treatment; UP, nonpollination treatment. Significant differences (p < 0.05) between treatments are indicated by different letters according to the Kruskal-Wallis test. Significance analysis was conducted with two-tailed Student's t-tests (*** p < 0.001; **** p < 0.0001).

The Effect of 4-CPA Treatment on Seed Growth and Development
In addition to inhibiting fruit growth, surprisingly, seed-like products were observed in 4-CPA-treated fruits. These products looked similar to normal seeds, albeit being smaller and in larger quantities (Figure 2a,b). For instance, 92.86% of control fruit only contained 2 seeds, 3.57% had 3 or 4 seeds, and no fruit contained more than 4 seeds. In contrast, more seed-like products were found in 4-PCA-treated fruits: the respective percentage of 2, 3, 4, 5, 6, and 7 seed-like products was 0, 7.69%, 42.31%, 19.23%, 26.92%, and 3.85%, respectively ( Figure 2c). The average number of seeds or seed-like products per fruit for the control and 4-CPA treatment was 2.1 and 4.8, respectively (Figure 2d). In terms of the size of seeds or seed-like structures, the transverse diameters of the seeds or seed-like products obtained from the control and 4-CPA treatment were 5.4 and 1.3 mm, and the vertical diameters of the seeds were 9.0 and 2.3 mm, respectively (Figure 2e); the average singleseed weight of the seeds was 839.6 and 253.8 mg, respectively ( Figure 2f). These results indicate that the average number of seed-like products per fruit after 4-CPA treatment is more than that of control fruit and that all fruit after 4-CPA treatment contained three or more seed-like products, whereas the majority of the control fruit contained two seeds. However, the seed-like products obtained after 4-CPA treatment were smaller than those obtained from the control. and the vertical diameters of the seeds were 9.0 and 2.3 mm, respectively (Figure 2e); the average single-seed weight of the seeds was 839.6 and 253.8 mg, respectively ( Figure 2f). These results indicate that the average number of seed-like products per fruit after 4-CPA treatment is more than that of control fruit and that all fruit after 4-CPA treatment contained three or more seed-like products, whereas the majority of the control fruit contained two seeds. However, the seed-like products obtained after 4-CPA treatment were smaller than those obtained from the control. To detect whether there was apomixis under 4-PCA treatment, anatomical observations and germination tests were conducted on these seed-like products. There were no embryo and endosperm in the seed-like products obtained after 4-CPA treatment; only a thicker seed coat was detected (Figure 3a,b,e). Furthermore, these products could not germinate at all. In comparison, the germination rate of normal seeds was 13.3% To detect whether there was apomixis under 4-PCA treatment, anatomical observations and germination tests were conducted on these seed-like products. There were no embryo and endosperm in the seed-like products obtained after 4-CPA treatment; only a thicker seed coat was detected (Figure 3a,b,e). Furthermore, these products could not germinate at all. In comparison, the germination rate of normal seeds was 13.3% (Figure 3c,d). Thus, we named the seed-like products induced by 4-PCA "defective seeds", which were not the result of apomixes.   (c) seed germination; (d) germination rate statistics; (e) seed coat thickness. Abbreviations: sc, seed coat; en, endosperm; em, embryo. Significance analysis was conducted with two-tailed Student's t-tests (** p < 0.01; **** p < 0.0001).

Observation of Histological Structure of Defective Seed
Since the auxin produced in the endosperm can drive seed coat development in Arabidopsis [19], we assumed that the formation of defective seeds (mainly containing seed coats) was the result of 4-PCA mimicking endosperm auxin. Thus, we further counted the number of cell layers and measured the cell size of the integument in the ovules after 4, 8, 12, and 16 DAT (Table 1). There was almost no difference in the number of cell layers in the inner integument between the 4-CPA treatment and the control. Additionally, the cell diameter and cell area of 4-CPA-treated seeds were slightly larger than that of control. However, the number of cell layers, cell diameter, and cell area in the outer integument were significantly increased under the 4-CPA treatment. Significance analysis was conducted with two-tailed Student's t-tests (** p < 0.01; **** p < 0.0001).

Observation of Histological Structure of Defective Seed
Since the auxin produced in the endosperm can drive seed coat development in Arabidopsis [19], we assumed that the formation of defective seeds (mainly containing seed coats) was the result of 4-PCA mimicking endosperm auxin. Thus, we further counted the number of cell layers and measured the cell size of the integument in the ovules after 4, 8, 12, and 16 DAT (Table 1). There was almost no difference in the number of cell layers in the inner integument between the 4-CPA treatment and the control. Additionally, the cell diameter and cell area of 4-CPA-treated seeds were slightly larger than that of control. However, the number of cell layers, cell diameter, and cell area in the outer integument were significantly increased under the 4-CPA treatment. Observation of the structure of the ovules at different stages of development revealed that the integument had 8-9 layers of cells in the ovary before flowering. At this time, the cells were arranged neatly, the nucellus also had about 8 layers of parenchyma cells, and the embryo sac began to develop (Figure 4a). The embryo development started at the end of pollination and fertilization. During this process, the number of integument cells in the pollinated and fertilized ovary did not change much, but the nucellus tissue continued to grow (Figure 4b-d). At 4 DAT (before flowering), the number of integument cells treated with 4-CPA was not much different from the control (Figure 4e). However, cell redifferentiation began to appear on the top of the ovule at 8 DAT (Figure 4f), and re-differentiated new ovule primordia appeared at 12 DAT (Figure 4g), resulting in abnormally thickened outer cells of the ovule (Figure 4h). Furthermore, some ovule primordia, induced by differentiation at the apex of the ovule, developed into a young ovule-shaped structure (Figure 4i). At 16 DAT, the integument tissue after 4-CPA treatment was abnormally developed, but the nucellus tissue was significantly atrophied (Figure 4j). Until 21 DAT, the developing new ovule-like structure could also be observed in the developed integument tissue after 4-CPA treatment (Figure 4k). Meanwhile, the integument cells were being differentiated at the top of the ovule (Figure 4l), which protruded into a seed shape (Figure 4m). These findings indicate that 4-CPA can stimulate the division and continuous development of integument cells and promote the redifferentiation of cells at the top of the ovule to form a structure similar to a new ovule.

Auxin Regulates the Expression and Location of Grape-Ovule-Related Genes
Our results above indicate that 4-CPA induces cell division and redifferentiation of the integument, resulting in abnormal thickening and cell enlargement. Meanwhile, it also induces the integument to form a seed coat. To investigate the related gene expression changes during this process, we firstly used protein sequences of genes known to be associated with integument cell division and development in Arabidopsis as queries for BLAST searches of the grapevine genome database (http://plants.ensembl.org/Vitis_vinifera/Info/Index) (accessed on 25 June 2020). Genes VvARF2, VvAP2, VvTTG2, and VvAGL11 were retrieved from the grapevine genome, respectively. Thus, the transcriptional level changes of integument-developmentassociated genes after 4-CPA treatment were detected by qRT-PCR. Compared with the

Auxin Regulates the Expression and Location of Grape-Ovule-Related Genes
Our results above indicate that 4-CPA induces cell division and redifferentiation of the integument, resulting in abnormal thickening and cell enlargement. Meanwhile, it also induces the integument to form a seed coat. To investigate the related gene expression changes during this process, we firstly used protein sequences of genes known to be associated with integument cell division and development in Arabidopsis as queries for BLAST searches of the grapevine genome database (http://plants.ensembl.org/Vitis_vinifera/ Info/Index) (accessed on 25 June 2020). Genes VvARF2, VvAP2, VvTTG2, and VvAGL11 were retrieved from the grapevine genome, respectively. Thus, the transcriptional level changes of integument-development-associated genes after 4-CPA treatment were detected by qRT-PCR. Compared with the control, we found the expression of VvARF2 was lower at 4 DAT but higher for 8-16 DAT (Figure 5a). The VvAP2 expression was significantly lower than the control at all tested time points after 4-CPA treatment (Figure 5b). Consistently, VvTTG2 induction levels were also lower than the control at 4 and 8 DAT (Figure 5c). However, VvAGL11 transcription levels did not show any differences between the 4-CPA treatment and the control (Figure 5d). were also lower than the control at 4 and 8 DAT (Figure 5c). However, VvAGL11 transcription levels did not show any differences between the 4-CPA treatment and the control (Figure 5d). Since the integument redifferentiation induced by 4-CPA treatment mainly occurred at 12 DAT, the expression of genes VvARF2 and VvAP2 also showed a significant difference during this time; we assumed that these two genes might be involved in a signaling pathway activated by 4-CPA. Therefore, we used in situ hybridization to locate the expression positions of VvARF2 and VvAP2. A strong VvARF2 mRNA hybridization signal was detected in the nucellus and inner and outer integument tissues 12 days after 4-CPA treatment (Figure 6a,b), while the hybridization signal was much weaker in the same positions in the control ovules (Figure 6c,d). In 4-CPA-treated ovules after 12 days, almost no VvAP2 hybridization signals were detected (Figure 6e,f). However, the VvAP2 hybridization signal was detected in the control nucellus and integument tissue ( Figure  6g). Thus, the qRT-PCR results were in line with in situ hybridization, namely, that 4-CPA treatment increased the expression of the VvARF2 gene and inhibited the expression of the VvAP2 gene in the ovule, which affected the number of integument cell layers and cell size, which, in turn, affected seed coat development. Since the integument redifferentiation induced by 4-CPA treatment mainly occurred at 12 DAT, the expression of genes VvARF2 and VvAP2 also showed a significant difference during this time; we assumed that these two genes might be involved in a signaling pathway activated by 4-CPA. Therefore, we used in situ hybridization to locate the expression positions of VvARF2 and VvAP2. A strong VvARF2 mRNA hybridization signal was detected in the nucellus and inner and outer integument tissues 12 days after 4-CPA treatment (Figure 6a,b), while the hybridization signal was much weaker in the same positions in the control ovules (Figure 6c,d). In 4-CPA-treated ovules after 12 days, almost no VvAP2 hybridization signals were detected (Figure 6e,f). However, the VvAP2 hybridization signal was detected in the control nucellus and integument tissue (Figure 6g). Thus, the qRT-PCR results were in line with in situ hybridization, namely, that 4-CPA treatment increased the expression of the VvARF2 gene and inhibited the expression of the VvAP2 gene in the ovule, which affected the number of integument cell layers and cell size, which, in turn, affected seed coat development.

Discussion
Our study has shown that the 4-CPA-induced defective seeds in parthenocarpic grapes are caused by promoting integument development and seed coat formation. This conclusion was generated by the observation that defective seeds were present in fruit formed from unpollinated ovaries under 4-CPA treatment and that such defective seeds contained no embryo and endosperm, only the seed coat (Figures 2 and 3). The development of the seed coat is triggered by auxin that has been synthesized in the developing endosperm and exported to the integument after fertilization [15,19]. Further, the results of the present study support the idea that the plant hormone auxin is a key factor in triggering seed coat formation in grapes.
The seed coat of plants originates from the integument tissue of the ovule [35]. Generally, the integument tissue is divided into outer integument and inner integument, which enclose the nucellus and female gametophyte. After fertilization, the integument tissue grows cooperatively with the embryo and endosperm [36]. Finally, the seed coat is formed to build a barrier to protect the inner seeds [37]. The thickness of the seed coat, which is determined by the number of integument cell layers, can affect seed germination, dormancy, and death [38,39]. Since we have found the 4-CPA can stimulate the division of grape integument cells, this might explain the occurrence of the thicken seed coat under 4-CPA application. Furthermore, in Arabidopsis, integument formation and development are largely regulated by the plant hormone auxin through the modulation of proximaldistal axis polarity [40,41]. Consistently, 4-CPA promotes the redifferentiation of apical cells of the ovule to form new ovules, which might be the regulatory effect of polarity, leading to the formation of more seed-like products. In a word, our findings of 4-PCAregulated seed development in grapes are largely overlapped with auxin's regulation effects on Arabidopsis. However, the changes of thickness of seed coat and seed numbers under auxin regulation are clear and measurable in grapes.
Furthermore, tissue observation is discussed together with the transcriptional changes of several seed-coat-associated genes in this study. The qRT-PCR results revealed that VvARF2 gene expression was upregulated while VvAP2 gene expression was downregulated (Figure 5a,b), indicating that the two genes play distinct roles under 4-CPA treatment. VvARF2 and VvAP2 are homologs of Arabidopsis ARF2 and AP2,

Discussion
Our study has shown that the 4-CPA-induced defective seeds in parthenocarpic grapes are caused by promoting integument development and seed coat formation. This conclusion was generated by the observation that defective seeds were present in fruit formed from unpollinated ovaries under 4-CPA treatment and that such defective seeds contained no embryo and endosperm, only the seed coat (Figures 2 and 3). The development of the seed coat is triggered by auxin that has been synthesized in the developing endosperm and exported to the integument after fertilization [15,19]. Further, the results of the present study support the idea that the plant hormone auxin is a key factor in triggering seed coat formation in grapes.
The seed coat of plants originates from the integument tissue of the ovule [35]. Generally, the integument tissue is divided into outer integument and inner integument, which enclose the nucellus and female gametophyte. After fertilization, the integument tissue grows cooperatively with the embryo and endosperm [36]. Finally, the seed coat is formed to build a barrier to protect the inner seeds [37]. The thickness of the seed coat, which is determined by the number of integument cell layers, can affect seed germination, dormancy, and death [38,39]. Since we have found the 4-CPA can stimulate the division of grape integument cells, this might explain the occurrence of the thicken seed coat under 4-CPA application. Furthermore, in Arabidopsis, integument formation and development are largely regulated by the plant hormone auxin through the modulation of proximaldistal axis polarity [40,41]. Consistently, 4-CPA promotes the redifferentiation of apical cells of the ovule to form new ovules, which might be the regulatory effect of polarity, leading to the formation of more seed-like products. In a word, our findings of 4-PCA-regulated seed development in grapes are largely overlapped with auxin's regulation effects on Arabidopsis. However, the changes of thickness of seed coat and seed numbers under auxin regulation are clear and measurable in grapes.
Furthermore, tissue observation is discussed together with the transcriptional changes of several seed-coat-associated genes in this study. The qRT-PCR results revealed that VvARF2 gene expression was upregulated while VvAP2 gene expression was downregulated (Figure 5a,b), indicating that the two genes play distinct roles under 4-CPA treatment. VvARF2 and VvAP2 are homologs of Arabidopsis ARF2 and AP2, respectively [42,43]. An increment of the number of integument cell layers and thickened seed coats were detected in Arabidopsis ARF2 mutant [28]. In contrast, when 4-CPA induced an increase in the number and size of integument cell layers, the expression of the VvARF2 gene increased in the grapes. Additionally, its hybridization signal was present on both the integuments and at the nucellus. We hypothesized that the opposite VvARF2 expression in grapes might be caused by the strong accumulation of VvARF2 mRNA signals at the nucellus. The Arabidopsis ap2 mutant had enlarged outer integument cells [25,26,44]. In agreement with this, we found that the outer integument cells of defective seeds were enlarged. The decreased VvAP2 expression was in line with the finding that its mRNA hybridization signal was diminished on the integument. Together, these findings suggest that AP2 is functionally conserved in woody plants.
In this study, we demonstrate that 4-CPA-induced defective seeds were caused by regulating the division and continuous redifferentiation of integument cells in "Fenghou" grapes. Our findings reveal that the 4-CPA treatment could upregulate VvARF2 gene expression, inhibit VvAP2 gene expression, and accelerate cell division and cell differentiation in the integument tissues. These events induced the increment of integument cell layers in the ovules and maintained continuous division and growth, resulting in the formation of defective seeds with only thick seed coats. Meanwhile, the redifferentiation of the integument tissue allowed the formation of new ovules, leading to a larger number of seeds in the ovary after 4-CPA treatment ( Figure 7). Overall, the study elucidated specific functions and mechanisms of 4-CPA on seed development in grapes.
Biomolecules 2021, 11, x 12 of 14 respectively [42,43]. An increment of the number of integument cell layers and thickened seed coats were detected in Arabidopsis ARF2 mutant [28]. In contrast, when 4-CPA induced an increase in the number and size of integument cell layers, the expression of the VvARF2 gene increased in the grapes. Additionally, its hybridization signal was present on both the integuments and at the nucellus. We hypothesized that the opposite VvARF2 expression in grapes might be caused by the strong accumulation of VvARF2 mRNA signals at the nucellus. The Arabidopsis ap2 mutant had enlarged outer integument cells [25,26,44]. In agreement with this, we found that the outer integument cells of defective seeds were enlarged. The decreased VvAP2 expression was in line with the finding that its mRNA hybridization signal was diminished on the integument. Together, these findings suggest that AP2 is functionally conserved in woody plants.
In this study, we demonstrate that 4-CPA-induced defective seeds were caused by regulating the division and continuous redifferentiation of integument cells in "Fenghou" grapes. Our findings reveal that the 4-CPA treatment could upregulate VvARF2 gene expression, inhibit VvAP2 gene expression, and accelerate cell division and cell differentiation in the integument tissues. These events induced the increment of integument cell layers in the ovules and maintained continuous division and growth, resulting in the formation of defective seeds with only thick seed coats. Meanwhile, the redifferentiation of the integument tissue allowed the formation of new ovules, leading to a larger number of seeds in the ovary after 4-CPA treatment ( Figure 7). Overall, the study elucidated specific functions and mechanisms of 4-CPA on seed development in grapes.