Maintenance of Methyl-Esterified Pectin Level in Pollen Mother-Cell Stages Is Required for Microspore Development

Pectin modification and degradation are vital for plant development, although the underlying mechanisms are still not well understood. Furthermore, reports on the function of pectin in early pollen development are limited. We generated OsPME-FOX rice lines with little methyl-esterified pectin even in the early-pollen mother-cell stage due to overexpression of the gene encoding pectin-methylesterase. Overexpression of OsPME1 in rice increased the activity of PME, which decreased the degree of pectin methyl esterification in the cell wall. OsPME1-FOX grew normally and showed abnormal phenotypes in anther and pollen development, especially in terms of the pollen mother-cell stage. In addition, we examined modifications of cell-wall polysaccharides at the cellular level using antibodies against polysaccharides. Immunohistochemical staining using LM19 and LM20 showed that methyl-esterified pectin distribution and the pectin contents in pollen mother-cell wall decreased in OsPME1-FOX compared with the wild type. Thus, the maintenance of methyl-esterified pectin plays a role in degrading and maintaining the pollen mother-cell wall during microspore development.


Introduction
Pollen development is accompanied by a characteristic series of complex changes in the cell wall [1,2]. Immediately before meiosis, pollen mother cells produce callose within their cell walls, union by plasmodesmata between adjacent pollen mother cells, and the tapetum disappears. The cell walls between neighboring pollen mother cells are separated by degradation, and after cytoplasmic division, microspores are surrounded by callose, a cell wall called the prime exine. Later, in the tetrad stage, the cell wall is further degraded, and the microspores are released into the locule. At this stage, the tapetum disintegrates [1]. Thus, cell-wall degradation and synthesis, as well as modification, are highly regulated in microspore development ( Figure S1).
The pollen wall is the outer structure of the pollen grain that protects the male gametophyte from physical and abiotic stresses. The pollen wall plays a vital role in pollen development, a key process in the life cycle of angiosperms [3]. The pollen wall consists of exine and intine and is elaborately organized. Intine is composed of pectin, structural proteins [4], and microfibril cellulose [5].
There are few reports on factors that specifically regulate changes in the cell-wall structure during pollen formation. The dex1 mutant of Arabidopsis thaliana exhibits abnormal pollen formation and male sterility [6]. Another male sterility mutant, ms33, also exhibits abnormal intine formation [7]. There are many reports that the genes encoding cell-wallmodifying enzymes are expressed specifically in the pollen and anthers [8][9][10][11][12]. A recent results, we show that proper maintenance of methyl esterification level of pectin via PME is important for normal microspore development.

OsPME1-FOX and OsPME34-FOX Overexpressed PME
For a systematic approach to analyze a gain-of-function phenotype, Hakata et al. [29] developed the FOX hunting system by using expression libraries for full-length cDNAs (fl-cDNAs) from rice at a maximum of 28,000 fl-cDNA clones in total, and individually overexpressed the fl-cDNAs in rice driven by the maize ubiquitin-1 gene promoter. Among approximately 14,500 FOX rice lines, we identified two lines overexpressing fl-cDNAs for OsPME1 and OsPME34 and named them OsPME1-FOX and OsPME34-FOX. The OsPME1 and OsPME34 transcript levels were elevated in OsPME1-FOX and OsPME34-FOX, respectively ( Figure 1A). PME activity in the flowers was also increased 5.2-and 4.1-fold in OsPME1-FOX and OsPME34-FOX, respectively, compared with that of the WT ( Figure 1B). the pollen mother-cell wall decreased in OsPME1-FOX compared with WT. From these results, we show that proper maintenance of methyl esterification level of pectin via PME is important for normal microspore development.

OsPME1-FOX and OsPME34-FOX Overexpressed PME
For a systematic approach to analyze a gain-of-function phenotype, Hakata et al. [29] developed the FOX hunting system by using expression libraries for full-length cDNAs (fl-cDNAs) from rice at a maximum of 28,000 fl-cDNA clones in total, and individually overexpressed the fl-cDNAs in rice driven by the maize ubiquitin-1 gene promoter. Among approximately 14,500 FOX rice lines, we identified two lines overexpressing fl-cDNAs for OsPME1 and OsPME34 and named them OsPME1-FOX and OsPME34-FOX. The OsPME1 and OsPME34 transcript levels were elevated in OsPME1-FOX and OsPME34-FOX, respectively ( Figure 1A). PME activity in the flowers was also increased 5.2-and 4.1-fold in OsPME1-FOX and OsPME34-FOX, respectively, compared with that of the WT ( Figure  1B).

Effect of Overexpression of PME1 and PME34 on the Pectin of the Cell Wall
The level of degree of pectin methyl esterification in the early stage of the flower was decreased to 27.2% and 69.7% in OsPME1-FOX and OsPME34-FOX, respectively, compared with the WT (Figure 2A). We also determined the uronic acid contents from polygalacturonase-soluble fraction of the cell wall of OsPME1-FOX and OsPME34-FOX. The pectin content in OsPME1-FOX and OsPME34-FOX was slightly decreased to 63.2% and 78.9%, respectively, compared to that of the WT ( Figure 2B). The decrease in the level of degree of pectin methyl esterification was more significant in OsPME1-FOX than in OsPME34-FOX. Figure 1. Characteristics of the OsPME1-FOX and OsPME34-FOX lines. (A) RT-PCR analysis of inflorescence transcripts from WT, OsPME1-FOX and OsPME34-FOX lines. The expression levels of OsPME1-FOX and OsPME34-FOX were higher than in the WT. The numbers in parentheses show the numbers of PCR cycles. These analyses were performed at least three times with similar results. (B) Pectin methylesterase activities in OsPME1-FOX and OsPME34-FOX. Pectin methylesterase activity is shown as the ratio of the activity in each FOX line inflorescence to that in the control. Error bars shows the SD (n = 3). Letters in each panel show significant differences at p < 0.01 (Tukey's test).

Effect of Overexpression of PME1 and PME34 on the Pectin of the Cell Wall
The level of degree of pectin methyl esterification in the early stage of the flower was decreased to 27.2% and 69.7% in OsPME1-FOX and OsPME34-FOX, respectively, compared with the WT (Figure 2A). We also determined the uronic acid contents from polygalacturonase-soluble fraction of the cell wall of OsPME1-FOX and OsPME34-FOX. The pectin content in OsPME1-FOX and OsPME34-FOX was slightly decreased to 63.2% and 78.9%, respectively, compared to that of the WT ( Figure 2B). The decrease in the level of degree of pectin methyl esterification was more significant in OsPME1-FOX than in OsPME34-FOX.

OsPME1-FOX and OsPME34-FOX Showed Abnormal Anther Development
OsPME1-FOX and OsPME34-FOX and control plants showed no differences in vegetative tissue ( Figure 3). OsPME-FOX1 and OsPME34-FOX plants showed various abnormal phenotypes in the male reproductive organs, but no differences from WT were observed in the female reproductive organs ( Figure 4A,F). In male reproductive organs, anthers were curled and short with a white color in OsPME1-FOX plants compared to WT ( Figure 4B,C). Anthers in OsPME34-FOX plants were slightly shorter than in WT ( Figure  4G, H). Furthermore, in OsPME-FOX1, it was observed that few pollen grains were present in the anthers ( Figure 4E) compared to WT ( Figure 4D). However, in OsPME34-FOX, pollen grains were similar to WT ( Figure 4I,J). During reproductive development, the

OsPME1-FOX and OsPME34-FOX Showed Abnormal Anther Development
OsPME1-FOX and OsPME34-FOX and control plants showed no differences in vegetative tissue ( Figure 3). OsPME-FOX1 and OsPME34-FOX plants showed various abnormal phenotypes in the male reproductive organs, but no differences from WT were observed in the female reproductive organs ( Figure 4A,F). In male reproductive organs, anthers were curled and short with a white color in OsPME1-FOX plants compared to WT ( Figure 4B,C). Anthers in OsPME34-FOX plants were slightly shorter than in WT ( Figure 4G, H). Furthermore, in OsPME-FOX1, it was observed that few pollen grains were present in the anthers ( Figure 4E) compared to WT ( Figure 4D). However, in OsPME34-FOX, pollen grains were similar to WT ( Figure 4I,J). During reproductive development, the fertility rate decreased by~12% in OsPME1-FOX and~26%OsPME34-FOX. In the OsPME1-FOX phenotype, the decrease in the degree of pectin methyl esterification level was significant (Figure 2A), and abnormal anther development could be observed more stably than that in OsPME34-FOX; thus, detailed microscopy experiments were performed using OsPME1-FOX. fertility rate decreased by ~12% in OsPME1-FOX and ~26%OsPME34-FOX. In the OsPME1-FOX phenotype, the decrease in the degree of pectin methyl esterification level was significant (Figure 2A), and abnormal anther development could be observed more stably than that in OsPME34-FOX; thus, detailed microscopy experiments were performed using OsPME1-FOX.

OsPME1-FOX Showed Abnormal Pollen Development
To determine the stage at which pollen development is disrupted in OsPME1-FOX, thin sections were obtained from the flowers of WT and OsPME1-FOX plants at various stages [14,30] (Figure 5). We divided rice-anther development after the pollen mother-cell stage into five stages. First, in the stage where pollen mother cells are observed in WT ( Figure 5A,B). In OsPME1-FOX, pollen mother cells stopped developing in the early-pollen mother-cell stage ( Figure 5F,G), and aberrant cells enlarged and filled the anthers where pollen mother cells are normally developed ( Figure 5G-I). In WT, tapetum was developed at the tetrad stage and mitosis stage ( Figure 5C,D), whereas tapetum morphogenesis could not be observed in the anther of OsPME1-FOX ( Figure 5H,I). Observation of anthers in the mature stage showed that mature pollen was formed in WT ( Figure 5E), whereas no pollen was observed in OsPME1-FOX ( Figure 5J).

OsPME1-FOX Showed Abnormal Pollen Development
To determine the stage at which pollen development is disrupted in OsPME1-FOX, thin sections were obtained from the flowers of WT and OsPME1-FOX plants at various stages [14,30] (Figure 5). We divided rice-anther development after the pollen mother-cell ( Figure 5A,B). In OsPME1-FOX, pollen mother cells stopped developing in the early-pollen mother-cell stage ( Figure 5F,G), and aberrant cells enlarged and filled the anthers where pollen mother cells are normally developed ( Figure 5G-I). In WT, tapetum was developed at the tetrad stage and mitosis stage ( Figure 5C,D), whereas tapetum morphogenesis could not be observed in the anther of OsPME1-FOX ( Figure 5H,I). Observation of anthers in the mature stage showed that mature pollen was formed in WT ( Figure 5E), whereas no pollen was observed in OsPME1-FOX ( Figure 5J).

OsPME1-FOX Showed Abnormal Distribution of Methyl-Esterified and De-Methyl-Esterified Pectin in Pollen Mother-Cell Stage of Anther
Since the defect responsible for low fertility was expected to be an abnormality in the methyl-esterification condition of pectin at the pollen mother-cell stage, which has abnormal morphogenesis in OsPME1-FOX, we prepared cross-sections of anthers from the early-and late-pollen mother-cell stages of OsPME1-FOX and WT plants by staining with a de-methyl-esterified pectin-specific antibody (LM19) and a methyl-esterified pectin-specific antibody (LM20) (Figures 6 and 7). In WT, signals for both de-methyl-esterified and methyl-esterified pectin were observed in anthers at the early pollen mother-cell stage ( Figure 6B,F), while in anthers at the late pollen mother-cell stage, the methyl-esterified pectin signal was very weak and the signal of the dimethyl-esterified pectin was stronger than in the previous stage ( Figure 7B,F). On the other hand, in anthers at the early pollen mother-cell stage of OsPME1-FOX, the signal of de-methyl-esterified pectin could be observed ( Figure 6J), but the signal of methyl-esterified pectin was not detected ( Figure 6N). Although methyl-esterified pectin is also included in the Golgi vesicles, these experiments were focused on the cell wall, exposure times were the same in WT and OsPME1-FOX, and the amount of pectin itself was much higher in the cell wall than in the Golgi vesicles, and therefore methyl-esterified pectin signals in Golgi vesicle are very difficult to observe. In anthers at the late-pollen mother-cell stage of OsPME1-FOX, and de-methyl-esterified pectin and methyl-esterified pectin were hardly detected ( Figure 7J,N). Green autofluorescence signals for sporopollenin, the main component of exine, were not detected in WT  Since the defect responsible for low fertility was expected to be an abnormality in the methyl-esterification condition of pectin at the pollen mother-cell stage, which has abnormal morphogenesis in OsPME1-FOX, we prepared cross-sections of anthers from the early-and late-pollen mother-cell stages of OsPME1-FOX and WT plants by staining with a de-methyl-esterified pectin-specific antibody (LM19) and a methyl-esterified pectin-specific antibody (LM20) (Figures 6 and 7). In WT, signals for both de-methyl-esterified and methylesterified pectin were observed in anthers at the early pollen mother-cell stage ( Figure 6B,F), while in anthers at the late pollen mother-cell stage, the methyl-esterified pectin signal was very weak and the signal of the dimethyl-esterified pectin was stronger than in the previous stage ( Figure 7B,F). On the other hand, in anthers at the early pollen mother-cell stage of OsPME1-FOX, the signal of de-methyl-esterified pectin could be observed ( Figure 6J), but the signal of methyl-esterified pectin was not detected ( Figure 6N). Although methylesterified pectin is also included in the Golgi vesicles, these experiments were focused on the cell wall, exposure times were the same in WT and OsPME1-FOX, and the amount of pectin itself was much higher in the cell wall than in the Golgi vesicles, and therefore methyl-esterified pectin signals in Golgi vesicle are very difficult to observe. In anthers at the late-pollen mother-cell stage of OsPME1-FOX, and de-methyl-esterified pectin and methyl-esterified pectin were hardly detected ( Figure 7J,N). Green autofluorescence signals for sporopollenin, the main component of exine, were not detected in WT and OsPME1-FOX in the early-and late-pollen mother-cell stages ( Figures 6D,H,L,P and 7D,H,L,P). and OsPME1-FOX in the early-and late-pollen mother-cell stages ( Figures 6D,H,L,P and 7D,H,LP).

Discussion
During pollen development, the cell walls of many tissues changes, including the tapetum, pollen mother cell, vesicle, and transmitting tissue of pistil [31][32][33][34]. A lot of enzymes for pectin degradation and modification have been reported from the anthers of petunia, tobacco, and Arabidopsis [8][9][10][11]32]. These enzymes include polygalacturonases, which directly depolymerize the pectin, as well as PME.
PME plays an important role in plant-cell separation and cell-wall modification. Primary cell-wall pectins are composed of the polysaccharides homogalacturonan, rhamnogalacturonan-I, and rhamnogalacturonan-II. The main component of all three types of pectin is galacturonic acid. When pectin is synthesized in Golgi body, galacturonic acid residues are methyl-esterified by pectin methyltransferase and released into the cell wall as methyl-esterified pectin [33,34]. Pectin chains may need to be depolymerized when cellwall expansion is required for growth or when degradation progresses, as in fruit soften-

Discussion
During pollen development, the cell walls of many tissues changes, including the tapetum, pollen mother cell, vesicle, and transmitting tissue of pistil [31][32][33][34]. A lot of enzymes for pectin degradation and modification have been reported from the anthers of petunia, tobacco, and Arabidopsis [8][9][10][11]32]. These enzymes include polygalacturonases, which directly depolymerize the pectin, as well as PME. PME plays an important role in plant-cell separation and cell-wall modification. Primary cell-wall pectins are composed of the polysaccharides homogalacturonan, rhamnogalacturonan-I, and rhamnogalacturonan-II. The main component of all three types of pectin is galacturonic acid. When pectin is synthesized in Golgi body, galacturonic acid residues are methyl-esterified by pectin methyltransferase and released into the cell wall as methyl-esterified pectin [33,34]. Pectin chains may need to be depolymerized when cell-wall expansion is required for growth or when degradation progresses, as in fruit softening [35]. Depolymerization of pectin chains is initiated by de-methyl esterification by PME. The resulting de-methylated pectin is then depolymerized and degraded by polygalacturonase. Hence, pectin, in which methyl groups are removed by PME, is known to be a target of polygalacturonase. Therefore, the regulation of pectin methyl esterification by PME is involved in the control of pectin degradation as well as the control of cell-wall properties by binding to calcium.
The overexpression of OsPME1-FOX and OsPME34-FOX resulted in increased PME activity ( Figure 1) and a reduced degree of methyl esterification of pectin ( Figure 2). On the other hand, the amount of pectin in OsPME1-FOX and OsPME34-FOX was slightly lower than in WT (Figure 2), even though PG activity was not increased compared to WT (data not shown). It is possible that many de-methyl-esterified pectins in OsPME1-FOX and OsPME34-FOX are easy targets for polygalacturonase and are therefore also susceptible to degradation. In the OsPME1-FOX phenotype, the decrease of the degree of pectin methyl esterification level and deficient development in pollen deficient could be observed more stably than in OsPME34-FOX. OsPME1 has a conserved PME domain and PMEI domain (group 2 PME), and OsPME34 only has PME domain (group 1 PME). Group 2 PME may affects pollen development than group 1 PME.
Our results showed that OsPME1-FOX and OsPME34-FOX affect normal pollen development (Figures 4-7). There are many reports of mutants with abnormal pollen development. For example, SPL/NZZ, BAM1, 2, EMS1/EXS, SERK1, 2, TPD1 and OsGRP2 all function at an early stage and determine tapetal formation [4,[36][37][38][39][40][41]. Thus, cell-wall development in tapetum is often required for pollen development. On the other hand, the cell-wall development of the stage of the pollen mother cell is also very important for pollen development. The bam1 and bam2 double mutant displays developmental defects at the early anther stage and lacks somatic cell layers, which suggests that these genes promote cell division and differentiation. In this study, we showed that the ubiquitously high expression of PMEs causes significant changes in pollen development with little pollen development (Figures 4 and 5). This is due to the lack of methyl-esterified pectin distribution of the cell wall in early pollen mother cells (Figure 6), suggesting that overexpression of OsPME1 probably leads to the progressive degradation of pectin in the early stages of pollen mother-cell development ( Figure 5), and thus fails to maintain sufficient pectin levels (Figure 7). PME catalyzes the de-methyl esterification of pectin and can have distinctly different physiological consequences depending on the pattern of de-methyl esterification within the pectin [18]. When de-methyl esterification occurs randomly, protons are released, promoting polygalacturonase activity, and degradation of the cell-wall pectin. On the other hand, de-methyl esterification of pectin residues may occur linearly. As a result, large blocks of negatively charged pectin residues are thought to interact with calcium ions to form pectate-Ca gels and stiffen the cell wall [25]. The mode of de-methyl esterification of a particular PME has been shown to be influenced by many factors, including pH, ion concentration, and the existing state of pectin methyl esterification [42]. In the present study, pectin levels were clearly reduced in OsPME1-FOX ( Figure 7). This strongly suggests that the effect of OsPME was not the formation of the pectin-Ca gel, but rather the progressive degradation of pectin by polygalacturonase. Our results suggest that it is important to maintain a high degree of pectin methyl esterification in the early stages of pollen mother-cell development through appropriate pectin methyl esterification regulation via PME. Additionally, the distribution of methyl-esterified pectin was not detected in WT at the later stages of pollen mother cell development (Figure 7). In QRT3, which encodes a polygalacturonase that degrades de-methyl-esterified pectin in the cell wall of pollen mother cells, loss of function prevents pollen development [3]. This suggests that maintaining a low degree of methyl esterification of pectin in late pollen mother cells and enriching the pollen mother-cell wall with de-methyl-esterified pectin may support QRT3 function.
Overexpression of PME also did not affect the development of the vegetative organ ( Figure 3). Monocotyledonous plants, including rice, have been reported to have very low pectin content in their vegetative organs. Overexpression of polygalacturonase also did not affect development of the vegetative organ [43], suggesting that pectin may not be very important for vegetative organ development in rice.
From our results and these studies, we conclude that maintaining the high degree of pectin methyl esterification in the cell wall during early pollen mother cells is directly and specifically involved in the degradation of pectic polysaccharides in the pollen mother-cell wall at the appropriate timing ( Figure 8). In the Arabidopsis qrt mutant, microspores do not separate and remain attached to the characteristic tetrahedral aggregate of four pollen grains [27]. This separation failure suggests that the pectin component of the cell wall of the pollen mother cell remains around the microspores after callose degradation [28]. It has been suggested that modification and degradation of the pectin component of the microspore primary cell wall within the cell wall of the pollen mother cell are important for the separation between developing microspores. Ectopic overexpression of PME improperly degrades the pectin in the pollen mother-cell wall, presumably causing the developing pollen grains to become mechanically constrained and remain strongly attached to the pollen grains (Figure 8). OsPME1-FOX were mixed and abnormal cells with fused outercell layers were observed. This suggested that this leads to the inter-digitation of the polysaccharides of the developing cell walls of the pollen grains, resulting in the fusion of the exine layers at the point of oppression. Intermixing of cell-wall polysaccharides leading to abnormal cell fusion has previously been invoked to explain the phenotype of the fiddlehead and other mutants [44]. These results provide a partial answer to the question of why plants have numerous enzymes for pectin methyl esterification [45]. This is the first report on the regulation of pectin methyl esterification in the cell wall of pollen mother cells during microspore development. Plants 2023, 12, x FOR PEER REVIEW 12 of 17 Figure 8. A model for Microspore Development in rice. The primary cell wall of the pollen mother cell has pectin. Callose is deposited between the primary wall and the plasma membrane of the pollen mother cell as the secondary cell wall. In OsPME1-FOX, the primary cell wall of the pollen mother cell is de-graded abnormally, which may prevent the normal development of microspores.
In the anther, presumably causing the developing pollen grains to become mechanically constrained and remain strongly attached to the pollen grains.

Plant Material and Growth Conditions
Rice plants of the WT (Oryza sativa cv. Nipponbare) and FOX lines, which carry overexpression constructs for OsPME1 (Os01g0788400) and OsPME34 (Os11g0683800), respectively, were used. The plants were grown in soil in a greenhouse during the natural growing season at 28 °C under 1600 µmol s-1 m-2illumination (normal light condition), and in a growth chamber at 28 °C under 115 µmol s-1 m-2 illumination (weak light condition). The primary cell wall of the pollen mother cell has pectin. Callose is deposited between the primary wall and the plasma membrane of the pollen mother cell as the secondary cell wall. In OsPME1-FOX, the primary cell wall of the pollen mother cell is de-graded abnormally, which may prevent the normal development of microspores. In the anther, presumably causing the developing pollen grains to become mechanically constrained and remain strongly attached to the pollen grains.

Plant Material and Growth Conditions
Rice plants of the WT (Oryza sativa cv. Nipponbare) and FOX lines, which carry overexpression constructs for OsPME1 (Os01g0788400) and OsPME34 (Os11g0683800), respectively, were used. The plants were grown in soil in a greenhouse during the natural growing season at 28 • C under 1600 µmol s-1 m-2illumination (normal light condition), and in a growth chamber at 28 • C under 115 µmol s-1 m-2 illumination (weak light condition).
The presence of the constructs in genomic DNA was confirmed by PCR using T3 generation plants. Transgenic lines were selected on hygromycin-containing agar and tested regarding the heritability of the expression pattern and altered sugar traits. To analyze gain-offunction phenotypes, Nakamura et al. (2007) and Hakata et al. (2010). developed the FOX hunting system using fl-cDNA expression libraries from rice containing 28,000 fl-cDNA clones and overexpressed the fl-cDNAs in rice under the control of the maize ubiquitin-1 promoter. Among approximately 14,500 FOX rice lines, two overexpressed fl-cDNAs for OsPME1 (Os01g0788400) and OsPME34 (Os11g0683800); these were named OsPME1-FOX and OsPME34-FOX, respectively.

Monitoring of Anther Development
The developmental stages of the rice pollen were defined in the same methods described in Itoh et al. [30] and Fujita et al. [14], based on anther length. Spikelets of different developmental stages were collected, fixed in PFA, and dehydrated through an ethanol series. The samples were embedded in Technovit 7100 resin (Hereaus Kulzer) and polymerized at room temperature. Transverse sections of 3 µm were cut using a Leica VT1200S (Leica Microsystems, Wetzlar, Germany) and stained with 0.1% (w/v) toluidine blue O in distilled water (DW). The sections were visualized by microscopy (DMRB, Leica Microsystems, Wetzlar, Germany).

Extraction of RNA and Analysis of Gene Expression
Plant material was frozen in liquid nitrogen and ground using a Tissue Lyser II (Qiagen, Hilden, Germany). The following procedure is based on the method described by Hyodo et al. [35]. Flowers with anthers less than 1 mm in length from OsPME1-FOX and OsPME34-FOX were sampled. Total RNA was extracted using an RNeasy Plant Mini Kit (Qiagen) and DNase I (Roche, Basel, Switzerland) according to the manufacturer's instructions. cDNA was synthesized using ReverTra Ace ® (Toyobo, Tokyo, Japan), as specified by the manufacturer. For the OsPME1-FOX line, transcripts were quantified using the primers OsPME1-forward (5 -GAAGCAGTTCCCGACGTT-3 ) and OsPME1-reverse (5 -CGCTCTGGTCCGTGATGA-3 ). For the OsPME34-FOX line, transcripts were quantified using the primers OsPME34-forward (5 -GGCCTCCACTACATCAAGGA-3 ) and OsPME34reverse (5 -TGACGCAGTGGAATTACTCG-3 ). As an endogenous control, 17S rRNA transcript was quantified using the primers 17S rRNA-forward (5 -GCAAATTACCCAATCC TGAC-3 ) and 17S rRNA-reverse (5 -CTATTGGAGCTGGAATTACC-3 ). The products were separated in a 2% agarose gel and stained with ethidium bromide. Quantitative reverse transcription-PCR (qRT-PCR) analysis was performed using SYBR Green I (Qiagen) with cDNA as the template on a Model 7000 Sequence Detection System (Applied Biosystems, Foster City, CA, USA). The ubiquitin gene (EU604080) was used as the normalizing reference gene.

Determination of Pectin Methylesterase Activity
Pectin methylesterase activity was assayed with a continuous spectrophotometric method according to Hyodo et al. [35]. Inflorescence, including anthers less than 1 mm in length from OsPME1-FOX and OsPME34-FOX, were sampled. Activity measurements were conducted at 20 • C and pH 7.5 in a cuvette containing 2 mL pectin, 0.15 mL bromothymol blue (BTB) and 0.55 mL distilled water. An addition of 0.3 mL enzyme extract started the reaction, and the residual enzyme activity was immediately assayed. The change in absorbance at 620 nm was recorded for 10 min in a UV/VIS spectrophotometer (Perkin-Elmer, Waltham, MA, USA). The activity values reported are an average of three independent measurements.

Extraction and Analysis of Cell-Wall Polysaccharides
The cell-wall extraction and analysis were conducted in accordance with Sumiyoshi et al. [46], with slight modifications. Rice inflorescences were frozen in liquid nitrogen and ground using a Tissue Lyser II (Qiagen) at 30 Hz for 2 min. A methanol/chloroform mixture (1 mL, 1:1, v/v) was added to the samples, followed by centrifugation at 15,000 rpm for 5 min and supernatant removal; the process was repeated twice. After the last supernatant removal, the samples were air-dried, and the resulting alcohol-insoluble residue (AIR, dry cell wall) was used as cell-wall material. The dry cell walls were extracted with 50 mM Na 2 CO 3 at room temperature for 2 h as pectic fractions.

Determination of Uronic Acid
Uronic acid was determined using the method of Blumenkrantz and Gustav (1973). Briefly, 1 mL ion-exchange water was added to each sample, and 1 mL iced concentrated sulfuric acid (0.025 M borax) was mixed into 200 µL of each sample. After heating in 100 • C water for 10 min and cooling in ice, 40 µL carbazole solution (125 mg carbazole, 100 mL ethanol) was mixed into each sample. The samples were heated in 100 • C water for 15 min, then cooled in ice; absorbance was measured at 530 nm (GENESIS 10S UV-VIS; Thermo Scientific). The cell-wall material was treated for 4 h at 4 • C with 0.1 M NaOH to saponify the methyl and acetyl esters. The suspensions were adjusted to pH 5.0 with 10% (vol/vol) glacial acetic acid and then treated for 16 h at 30 • C with a homogeneous preparation of endo-polygalacturonase (EPG) from Aspergillus niger [2.5 units, Megazyme, Wicklow, Ireland; 1-unit releases 1 µmol of reducing sugar min −1 from a 1% (weight/volume) solution of polygalacturonate at pH 5.0 and 25 • C. The suspensions were centrifuged, and the insoluble residues were washed with water.

Determination of the Pectin Methyl Ester Content
The methyl ester group was determined quantitatively by an enzymatic method involving an alcohol oxidase/formaldehyde dehydrogenase system. For hydrolysis of methyl esters bound to pectin, 0.1 m KOH (100 µL) was added to the pectin fraction (100 µg/100 µL), followed by standing for 1 h at room temperature. The methanol released was determined. The reaction mixture, composed of 100 mM glutathione (60 µL), 100 mM NAD + (60 µL), alcohol oxidase (1 unit) and FADH (2 units) in 0.2 M potassium phosphate buffer (pH 7.5) in a total volume of 2.9 mL, was placed in screw-cap tubes. Aliquots (100 µL) of methanol standards (0.5-10 µg) or the pectin hydrolysates containing 50 pg of galacturonic acid were added to the tubes. The tubes were incubated at 25 • C for 30 min. The methanol content was calculated using e = 6.2 × 10 3 mol -1 cm -1 for NADH at 340 nm. The degree of methyl esterification was expressed as the molar per cent of methyl ester groups per d-galacturonic acid residues.

Immunohistochemistry
The following procedure is based on the method described by Hasegawa et al. [33]. Samples were fixed in 4% paraformaldehyde, 0.25% glutaraldehyde and 0.05 M phosphate buffer (pH 7.5). Transverse 3 µm sections were cut using a Leica RM2145 microtome (Leica Microsystems) and stained with 0.1% (w/v) toluidine blue O in distilled water. For immunohistochemistry, sections were subjected to immunohistochemical analysis using TSA Kit #12 (Invitrogen, Eugene, OR, USA). Primary antibodies for LM19 and LM20 (PlantProbes, Leeds, UK) were used at a dilution of 1:20. Negative controls lacked the primary antibody. Sections were visualized with fluorescence microscopy (Leica; DMRB).

Statistical Analysis
The data were expressed as the mean values ± SD taken from 4-9 independent biological experiments. The experimental data of the samples were statistically analyzed through one-way analysis of variance (ANOVA) with Tukey's post hoc test using Statistica 13.1 software (StatSoft, Inc., Tulsa, OK, USA). The results with a p-value ≤ 0.05 and a p-value ≤ 0.01 were considered statistically significant.