Potential Association of Reactive Oxygen Species With Male Sterility in Peach

Male sterility is an important agronomic trait for hybrid vigor utilization and hybrid seed production, but its underlying mechanisms remain to be uncovered. Here, we investigated the mechanisms of male sterility in peach using a combined cytology, physiology, and molecular approach. Cytological features of male sterility include deformed microspores and tapetum cells along with absence of pollen grains. Microspores had smaller nucleus at the mononuclear stage and were compressed into belts and subsequently disappeared in the anther cavity, whereas tapetum cells were swollen and vacuolated, with a delayed degradation to flowering time. Male sterile anthers had an ROS burst and lower levels of major antioxidants, which may cause abnormal development of microspores and tapetum, leading to male sterility in peach. In addition, the male sterility appears to be cytoplasmic in peach, which could be due to sequence variation in the mitochondrial genome. Our results are helpful for further investigation of the genetic mechanisms underlying male sterility in peach.


INTRODUCTION
A typical flower has four types of floral organs: sepal, petal, stamen, and pistil. The function of the flower is to make seeds for plant propagation. Seed development is initiated by fertilization in which male and female gametes fuse, with the former produced in the anther, a part of stamen, whereas the latter is produced in the ovary, an interior part of the pistil. Hence, the development of male and female gametes is crucial for reproduction of flowering plants. However, some plants lack male gametes, also called pollen grains, despite having complete floral organs. The development of pollen grains is a complex and delicate process. Initially, the pollen mother cell in the pollen sac undergoes meiotic division to produce four daughter cells that are called a tetrad. Later, the tetrad dissociates and develops into mononuclear microspores with the nucleus in the center of the cell. Tapetum cells nourish the microspores, increasing in the size and gradually forming a large central vacuole in the cell. As a result, the nucleus gradually moves from the center to the side. Subsequently, the microspores undergo a mitotic division, producing a larger vegetative cell and a smaller generative cell. Most angiosperms, such as lily, cotton, peach, and orange, have bicellular pollen grains, containing vegetative and generative cells, when the pollen is mature. However, some plants, such as cereals and rapeseed, have trinuclear pollen grains as their generative cells undergo additional mitotic division to generate two sperm cells before the pollen is released. Pollen grains at the bicellular and tricellular stages are also referred to as male gametophytes. Previous studies have shown that numerous genes are involved in pollen development (Wilson and Zhang, 2009;Zhang et al., 2011;Gómez et al., 2015). Abnormal structure or expression of pollen development-related genes may cause failure of the anther to produce mature pollen grains, resulting in male sterility (McCormick, 2004;Borg et al., 2009;Cai et al., 2015).
To date, numerous genes that are related to male sterility have been discovered in plants (Singh et al., 2010;Lukaszewski, 2017). Particularly, functional disorder of genes related to tapetum development is one of the main causes of pollen abortion. Tapetum is located in the innermost part of anther tissue, and its main function is to provide nutrition for the development of pollen grains. During the late stages of pollen development, the tapetum undergoes cellular degradation, and anthers subsequently crack, causing mature pollen grains to be released (Mariani et al., 1990;Parish and Li, 2010;Zhu et al., 2011;Xie et al., 2014). Therefore, normal tapetum is crucial for the development of pollen grains (Li et al., 2017). For example, mutations of OsMS1 and OsTDR genes that regulate tapetum development can block the degradation of tapetum, leading to male sterility in rice (Li et al., 2006;Yang et al., 2019).
In addition to the dysfunction of genes involved in the regulation of tapetum development, other factors such as the excessive production of reactive oxygen species (ROS) in anthers can also cause male sterility. For example, reducing the expression of MT-1-4b gene encoding a type 1 small Cys-rich and metal binding protein increases the level of superoxide anion (O 2 − ), which affects tapetum development and thus decreases pollen infertility (Hu et al., 2011). ROS can act as signaling molecules and maintaining their basal levels is essential for plant development (Gechev et al., 2006). Mitochondrion is a main source of ROS production as it is the site of oxidative phosphorylation. Mitochondria structural variation can lead to the ROS burst (Shadel and Horvath, 2015). In addition, the ROS burst can be induced by environmental stress and dysfunction of genes associated with ROS elimination (Gechev et al., 2006;Suzuki et al., 2012;Miller et al., 2010). Plants contain a variety of antioxidants, such as carotenoids, glutathione (GSH), ascorbate, and flavonoids, which serve as ROS scavengers (Mittler et al., 2004). In plants, the production and elimination of ROS are a balanced process, which is called ROS homeostasis. Many studies have shown that loss of ROS homeostasis can lead to male sterility (Hu et al., 2011;Qu et al., 2014;Zhou et al., 2019;Wang et al., 2020;Zhu et al., 2020).
Peach (Prunus persica) is a fruit tree widely cultivated in temperate regions of the world, with ornamental and edible properties. Although male sterile frequently occurs in peach, its molecular mechanism is yet to be fully uncovered. In this study, we investigated the mechanism of male sterility in "Jinxiang" ("JX"), a popular yellow flesh peach cultivar in China. Our results indicated that abnormal development of tapetum and microspores are caused by disruption of ROS homeostasis, resulting in male sterility. This finding provides an insight into mechanisms underlying male sterility in peach.

Plant Materials
Peach varieties used in this study are maintained in the orchard at Shijiazhuang Pomology Institute, Hebei Academy of FIGURE 1 | Comparison of anther phenotype between "ZH" and "JX". Panels Frontiers in Plant Science | www.frontiersin.org Agriculture and Forestry Sciences, Shijiazhuang, China. Anthers were collected from a male sterile cultivar, JX, and a fertile cultivar, "Zaohong" ("ZH"), as a control. These two cultivars have the same flowering time. Samples were collected at 5-day intervals from March 09, 2019, to April 03, 2019. All samples were divided into six stages based on sampling time, S0, S1, S2, S3, S4, and S5, which were confirmed by paraffin section according to previously reported protocol (Fang et al., 2016). S0-S3 corresponded to 20, 15, 10, and 5 days before anthesis, respectively. S4 corresponded to the anthesis, and S5 represented 5 days after anthesis.

Anatomical Analysis of the Anther Structure
Anthers from the unopened and fully opened flower buds were taken out, put in culture dish, and left to dry. The dried anthers were fixed with glutaraldehyde (pH 6.8) at 4 • C for 1.5 h and then washed three times with 0.1 mol/L phosphoric acid buffer (pH 6.8) for 10 min each. Subsequently, the treated samples were dehydrated with 50, 70, 80, and 90% ethanol for 10-15 min each time and finally dehydrated three times with 100% ethanol for 10-15 min each time. Then, the anthers were immersed in 100% ethanol:tert-butanol (vol/vol = 1) for 15 min and transferred to pure tert-butanol for 15 min. The treated samples were dried for 4 h using LABCONCO Freeze Vacuum Dryer (Labconco, MO, United States). Finally, the observation surface of the sample was put in an upward position, fixed with conductive tape, and sprayed with gold powder using Leica EM ACE Ion sputtering coating instrument (Leica, Wetzlar, Germany). Anther morphology was observed using scanning electron microscope (SEM) (Hitachi, Tokyo, Japan). In addition, transmission electron  microscopy (TEM) was performed according to a previous report (Li et al., 2006).

Detection of Pollen Viability
Pollen viability was analyzed by both 1% iodine/potassium iodide (I 2 /KI) and fluorescein diacetate (FDA) staining. For the I 2 /KI assay, the samples were stained for 5 min and scanned with microscope (Nikon, Tokyo, Japan). For the FDA staining assay, the pollen grains were immersed in a concave slide containing 2 µg/mL FDA solution for 30 min and observed under the fluorescence microscope (Nikon, Tokyo, Japan) after washing out FDA.

Measurement of ROS and Antioxidant Contents in Anthers
The hydrogen peroxide (H 2 O 2 ) content in anthers was measured using H 2 O 2 assay kit (Beyotime, Shanghai, China) following the manufacturer's instruction. The O 2 − radical content in anthers at each developmental stage was assayed using nitroblue tetrazolium (NBT) staining according to a previous report (Xie et al., 2014).
The content of GSH was measured using GSH assay kit (Beyotime, Shanghai, China) according to the manufacturer's instruction. Carotenoid content was measured using plant carotenoid detection kit (Solarbio, Beijing, China) following the manufacturer's instruction. The content of ascorbate peroxidase (APX) and GSH S-transferase (GST) was measured by APX and GST detection kit (GeRuiSi, Nanjing, China), respectively, according to the manufacturer's instruction.

RNA-Sequencing Analysis
Three development stages of JX and ZH anthers, S0, S1, and S3, were selected for RNA-sequencing (RNA-seq) analysis. RNA extraction and purity, library construction, and RNA-seq were conducted according our previous report (Wang L. et al., 2013). After removing adapters and low-quality reads, clean reads were mapped to the peach reference genome (International Peach Genome Initiative, 2013). Gene expression levels were estimated based on the value of expected number of fragments per kilobase of transcript sequence per million base pairs sequenced (FPKM). Differentially expressed genes (DEGs) were identified using the following criteria: fold change ≥2.0 and a false discovery rate <0.01.
Frontiers in Plant Science | www.frontiersin.org chain reaction (qRT-PCR) was performed using TAKARA SYBR R Premix EX Taq TM II (Tli RNaseH Plus), and the amplification program was as follows: one cycle of 30 s at 95 • C, followed by 40 cycles of 5 s at 95 • C and 34 s at 60 • C. Relative expression levels were normalized against the reference gene GADPH (Tong et al., 2009

JX Has Abnormal Anther Morphology and Lacks Pollen Grains
Flower buds showed no obvious difference in appearance between the male sterile cv. JX and the fertile cv. ZH before the full blooming stage (Supplementary Figure 1), but their inner anthers were different in size and color (Figures 1A,B). The anthers of ZH were yellow, swollen, and significantly larger than the shrunken anthers of JX that were pale in color. The anthers of ZH cracked to release pollen grains, and the color changed from yellow to dark orange at the anthesis, whereas the anthers of JX were indehiscent with no obvious change in color (Figures 1C,D). Five days after anthesis, the anthers of ZH wilted, whereas dehiscence was observed in some anthers of JX, with filament color both changing to purple (Figures 1E,F). Both I 2 /KI and FDA staining showed that the anthers of ZH contained viable pollen grains, whereas the anthers of JX had no pollen grains (Figures 1G-J).
Scanning electron microscope assay showed that the surface of ZH anthers had opening cracks, whereas abnormal opening cracks were observed in the surface of JX anthers (Figures 2A,B). The anthers of ZH released numerous pollen grains during dehiscence, whereas the disrupted anthers of JX had no pollen grains, with few pollen-like spheres ( Figure 2D) that had no typical wrinkles on their surface as shown in Figure 2C.
In summary, the above results suggested that the anthers of JX underwent abnormal development, with delayed dehiscence in some anthers that had no pollen grains.

Microspores of JX Were Compressed Into Bands That Disappeared in Anther Cavity
To investigate the cytological mechanism of pollen abortion in JX, we examined paraffin-embedded anther samples at five stages (S1-S5) using microscope. Microspores were abundant in the anthers of both ZH and JX at S1; however, the anthers JX showed irregular morphology (Figures 3A,B). Microspores in each pollen sac were compressed into a single band in JX, whereas no morphological change was detected for microspores of ZH (Figures 3C,D). The anthers of ZH cracked normally during the S3 stage, but not for the anthers of JX that had empty pollen sacs, with no pollen grains (Figures 3E,F). Interestingly, anther cracking was delayed to the anthesis in JX (Figure 3H), in which mature pollen grains were released in ZH (Figure 3G). At S5, anthers showed no difference in structure between ZH and JX (Figures 3I,J). These results suggested that microspore development was disrupted, leading to pollen abortion in JX.

Ultrastructural Feature of Microspores and Tapetum in ZH and JX
To gain deep insights into abnormal anther development in JX, the ultrastructures of microspores and tapetum were examined using TEM. A single nucleus was present in microspores at S1 and S2 in JX and ZH (Figure 4). The microspores of ZH contained large nucleus, whereas the microspores of JX were severely vacuolated, with smaller nucleus (Figures 4A,B). Moreover, the exine of microspores in ZH had regular rod-like protrusions, whereas the exine structure of microspores in JX was irregular and incomplete (Figures 4A,B). At S2, the nuclear in microspores of ZH was pushed to the side because of vacuole enlargement, whereas microspores of JX were compressed into bands (Figures 4C,D).
Tapetum degradation occurred at S1 in ZH, whereas the development of tapetum was abnormal in JX (Figures 5A,B). The tapetum cells of JX had loose cytoplasm and heavy vacuolization in contrast to those of ZH at S1. The tapetum was almost completely degraded at S2 in ZH (Figure 5C), whereas the tapetum cells of JX were swollen and contained large vacuoles (Figure 5D). At S4, the tapetum of ZH was completely degraded (Figure 5E), whereas tapetum degradation was just initiated in JX ( Figure 5F). Altogether, these results suggested that tapetum development was abnormal and could not be timely degraded to nourish microspores, leading to pollen abortion in JX. In  *B and C represent "ZH" and "JX", respectively. S0, S1, and S3 indicate three developmental stages of anthers. Each sample contains three biological replicates.
Frontiers in Plant Science | www.frontiersin.org addition, we observed that tapetal cells at S2 were filled with round, electron dense, and granular organelles, which are typical characteristics of peroxisomes ( Figure 5D). This suggested the presence of oxidative stress in microspores of JX (Hu et al., 2011).

The ROS Content Was Elevated in the Anthers of JX Compared to Those of ZH
As oxidative stress is associated with excess production of ROS that causes damage to cell structure (Gechev et al., 2006), we measured the contents of H 2 O 2 and O 2 − , two main components of ROS (Apel and Hirt, 2004), in anthers of JX and ZH at stages 1-5. The concentration of H 2 O 2 was significantly higher in JX than in ZH at S1, S2, and S4 ( Figure 6A). However, the concentration of H 2 O 2 in anthers at S3 and S5 was similar between JX and ZH. Moreover, the NBT staining assay demonstrated that the content of O 2 − at stages 1-4 was higher in JX than in ZH ( Figure 6B). These results suggested an excessive accumulation of ROS in the anthers of JX.

Comparative Transcriptome Analysis Between the Anthers of JX and ZH
As mentioned previously, abnormal development of microspores was initially detected at S1. Thus, S0 and S1 represent transitional stages that are crucial for comparative transcriptome analysis according to a previous study (Liao et al., 2020). Moreover, microspores gradually degraded and finally disappeared in S3. Based on these findings, three stages, S0, S1, and S3, were selected to conduct comparative transcriptome analysis to further reveal molecular mechanisms of pollen abortion in JX. A total of 18 libraries were sequenced, generating approximately 5.77-Gb raw reads for each library, with an average Q30 value of 94.77% ( Table 1). The clean reads of each sample were mapped against the peach reference genome 1 , with unique mapping rates ranging from 81.64 to 92.57%. Pearson correlation coefficients were greater than 0.9 among three biological replications of each sample (Supplementary Figure 2), and 3D principal component analysis (PCA) plot displayed that three biological repeats of each sample clustered together (Supplementary Figure 3). This indicated high consistency between biological replicates, which is suitable for conducting comparative transcriptome analysis.

The Content of Major Antioxidants in the Anthers Was Lower in JX Than in ZH
To further confirm the association of pollen abortion with the ROS metabolism in JX, we measured the content of four major ROS scavengers in the anthers at S1-S5. Overall, the concentrations of APX and GST were significantly lower in anthers of JX than in those of ZH throughout the development (Figure 8). Similarly, the contents of carotenoids and GSH were significantly lower in the anthers of JX than in those of ZH. These results indicated there is a disruption of ROS homeostasis in anthers of JX.
FIGURE 8 | Content of antioxidants in the anthers of "ZH" and "JX" at stages 1-5. Error bars indicate ±SD of three biological replicates. Statically significant differences at P ≤ 0.05 and P ≤ 0.01 (Student t test) are indicated by * and * * , respectively.

Pollen Abortion Is Associated With Abnormal Development of Microspores and Tapetum in Peach
To our knowledge, this study reports for the first time the cytological and physiological traits associated with male sterility in peach. Our results show that microspores were severely vacuolated, with deformed and incomplete exine structure, and compressed into belts that disappeared at mononuclear stage in male sterile cv. JX. Moreover, the tapetum cells were swollen, vacuolated, with a delayed degradation to flowering time. Thus, male sterility in JX is characterized by abnormal development of microspores and tapetum at mononuclear stage of pollen development. As mentioned previously, tapetum provides nutrition for the development of pollen grains. Abnormal development of microspores could be partially attributed to the disrupted development of tapetum in peach.
Similar phenotype of male sterility has also been reported in model plants such as rice and Arabidopsis, in which pollen abortion is frequently found to be associated with genes that are involved in tapetum development and degradation. For example, mutations of tapetum development-related genes in Arabidopsis, such as MYB103, MS1, TDF, DYT1, PRX9, and PRX10, can lead to comparable pollen abortion phenotypes to those observed in this study (Ito et al., 2007;Zhang et al., 2007;Zhu et al., 2008Zhu et al., , 2015Jacobowitz et al., 2019). In rice, functional mutations of TDR, API5, DEX1, GPAT3, and DPW3 cause delayed degradation of tapetum, leading to male sterility with features similar to those in this study (Li et al., 2006(Li et al., , 2011Yu et al., 2016;Men et al., 2017;Mondol et al., 2020). Therefore, degradation of tapetum in mononuclear stage is crucial for normal development of microspores into mature pollen grains. The delayed degradation of tapetume is likely responsible for male sterility in peach cv. JX.

Disruption of ROS Homeostasis May Hinder Pollen Development, Leading to Male Sterility in Peach
In plants, basal concentration of ROS is important in maintaining normal growth and development. The production and elimination of ROS are a complex network involving participation of multiple factors (Gechev et al., 2006;Miller et al., 2010;Suzuki et al., 2012;Shadel and Horvath, 2015). In this study, the anthers of JX were found to accumulate higher levels of H 2 O 2 and O 2 − at early stages of development. ROS burst in anther has been found to affect tapetum degradation time (Jiang et al., 2007;Huang et al., 2012;Luo et al., 2013;Wang K. et al., 2013;Yan et al., 2014). ROS burst at the mononuclear stage is likely responsible for the delay of tapetum degradation, leading to abortion of microspores in JX. This finding is consistent with previous reports that disruption of ROS homeostasis contributes to male sterility. For example, the BZR1 gene is involved in ROS production, and its mutation can delay tapetal cell degeneration, leading to male sterility in tomato (Yan et al., 2020), whereas increased ROS content negatively impacts pollen development in rice (Qu et al., 2014;Zheng et al., 2019;Guo et al., 2020;Zafar et al., 2020).
Plants have a complex enzymatic and non-enzymatic antioxidant system for maintaining the ROS homeostasis (Mittler et al., 2004). In this study, both major enzymatic antioxidants, APX and GST, and major non-enzymatic antioxidants, GSH and carotenoids, were detected in the anthers of ZH and JX. However, the levels of enzymatic antioxidants and nonenzymatic antioxidants were significantly lower in the anthers of JX than in those of ZH. Thus, ROS homeostasis seems to be crucial for pollen development, and its disruption could be associated with male sterility in peach. In addition, a large number of structural genes related to antioxidant production were differentially expressed in the anthers of ZH and JX. It is worthy of further studies to investigate whether antioxidant accumulation is controlled by regulatory genes in peach.
Mitochondrion is a major site of ROS production that is related to oxidative phosphorylation (Gechev et al., 2006;Miller et al., 2010;Suzuki et al., 2012;Shadel and Horvath, 2015).
Dysfunction of mitochondrial genes has been reported to cause excessive production of ROS (Touzet and Meyer, 2014). Insertion of unknown open reading frames into the mitochondrial genome causes excessive accumulation of ROS, leading to male sterility in rice (Luo et al., 2013;Wang K. et al., 2013;Xie et al., 2018). In this study, cytological and physiological characteristics of male sterility in JX were similar to those of cytoplasmic male sterile rice variety, Honglian. In addition, ROS burst was observed in the anther of JX, along with six mitochondrial genes involved in oxidative phosphorylation that were differentially expressed in the anthers of ZH and JX (Supplementary Table 3). These findings suggest that male sterility could be cytoplasmic due to sequence variation in the mitochondrial genome in peach, which is consistent with previous finding that ROS burst during pollen development causes Honglian type cytoplasmic male sterility (CMS-HL) in rice (Wang K. et al., 2013). Analysis of structural variation of mitochondrial genome will be a practical way to identify potential candidate genes for male sterility in peach. Notably, a previous study has reported a candidate gene on chromosome 6 (Prupe.6G024900) controlling male sterility in peach (Eduardo et al., 2020). However, based on RNA-seq, Prupe.6G024900 showed no expression in the anther at S0, S1, and S3 of both JX and ZH, suggesting it is unlikely responsible for male sterility in JX. More studies are needed to clarify whether diverse mechanisms are associated with male sterility in peach germplasm.
In summary, our study reveals that disruption of ROS homeostasis may cause abnormal development of microspores and tapetum, leading to cytoplasmic-type male sterility in peach. Our results will be useful for further investigation of mechanisms underlying male sterility in peach.

DATA AVAILABILITY STATEMENT
The original contributions generated for this study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

AUTHOR CONTRIBUTIONS
YC conducted most experiments of this study and wrote the manuscript. YC, ZM, and LZ prepared the experimental materials. LL, LW, and BZ participated in the transcriptomic analysis. YH was overall project leader and revised the manuscript. CO and RZ revised the manuscript. All authors read and approved the final manuscript.