ARF3-Mediated Regulation of SPL in Early Anther Morphogenesis: Maintaining Precise Spatial Distribution and Expression Level

Abstract

Early anther morphogenesis is a crucial process for male fertility in plants, governed by the transcription factor SPL. While the involvement of AGAMOUS (AG) in SPL activation and microsporogenesis initiation is well established, our understanding of the mechanisms governing the spatial distribution and precise expression of SPL during anther cell fate determination remains limited. Here, we present novel findings on the abnormal phenotypes of two previously unreported SPL mutants, spl-4 and spl-5, during anther morphogenesis. Through comprehensive analysis, we identified ARF3 as a key upstream regulator of SPL. Our cytological experiments demonstrated that ARF3 plays a critical role in restricting SPL expression specifically in microsporocytes. Moreover, we revealed that ARF3 directly binds to two specific auxin response elements on the SPL promoter, effectively suppressing AG-mediated activation of SPL. Notably, the arf3 loss-of-function mutant exhibits phenotypic similarities to the SPL overexpression mutant (spl-5), characterized by defective adaxial anther lobes. Transcriptomic analysis revealed differential expression of the genes involved in the morphogenesis pathway in both arf3 and spl mutants, with ARF3 and SPL exhibited opposing regulatory effects on this pathway. Taken together, our study unveils the precise role of ARF3 in restricting the spatial expression and preventing aberrant SPL levels during early anther morphogenesis, thereby ensuring the fidelity of the critical developmental process in plants.

1. Introduction

The anther is a vital male reproductive organ in plants responsible for the production of male gametophytes. Proper development of anthers is crucial for successful plant sexual reproduction and ensuring crop yield [1,2]. Anther morphology in angiosperms is highly conserved, characterized by four lobes and a well-defined development process [3,4,5,6,7]. In Arabidopsis thaliana, the stamen primordium initially forms as a spherical structure composed of three layers of somatic cells, with archesporial cells giving rise to primary sporogenous cells (PSC) internally and primary parietal cells (PPC) externally. Subsequently, PPCs differentiate into secondary parietal cells (SPC), and the SPCs further differentiate into the middle layer (ML), endothecium (En), and tapetum (T), while sporogenous cells (SSC) become microsporocytes [1,2]. The establishment of adaxial and adaxial polarity, as well as the correct differentiation of early cells, are crucial for anther morphogenesis.

Significant progress has been made in understanding anther morphogenesis and discovering key genes and regulatory mechanisms involved [8,9,10,11,12]. The MADS-BOX transcription factor SPL is considered a core regulator during early anther morphogenesis. The loss-of-function mutant spl-1, obtained from the Ac/Ds mutant library, exhibits anthers composed of highly vacuolated parenchyma cells, leading to the absence of anther lobes and the microsporocytes, resulting in male sterility [13]. mRNA in situ hybridization experiments have demonstrated specific expression of SPL in developing microsporocytes [13]. AG directly binds to the CArG-box element downstream of the SPL coding region, activating SPL and thereby initiating microsporogenesis. The ag-1 mutant fails to form microspores, while SPL can induce microspore production in the absence of AG, indicating the importance of SPL in the microsporocytes [14]. SPL also plays a significant role in anther organogenesis. Ectopic expression of SPL leads to the transformation of petals into stamens in the weak mutant of the APETALA2 (AP2), ap2-1. Ectopic expression of SPL overexpression also affects the expression of AG, SEP3, and AP2, suggesting that SPL is involved in the regulation of stamen organ identity determination [15]. BAM1 and BAM2 encode two functionally redundant receptor-like kinases that are essential for normal cell division and differentiation during early anther development. The bam1 bam2 mutant lacks endothecium, middle layer, and tapetum and exhibits morphological characteristics of microsporocytes within the epidermis, indicating the involvement of BAM1 and BAM2 in cell fate determination. Further research has revealed a negative feedback regulation between BAM1/2 and SPL, with SPL promoting microsporocytes preferential expression of BAM1/2 and BAM1/2 restricting SPL expression to the microsporocytes [16]. SPL is not only involved in the division and differentiation of archesporial cells but also directly participates in early anther polarity establishment.

MPK3 and MPK6 encode two functionally redundant protein kinases. The mpk3/+mpk6/-mutants exhibited a deficiency in forming the two adaxial anther lobes. Immunofluorescence analysis revealed the co-expression of MPK3, MPK6, and SPL in the microsporocytes. Subsequent investigations demonstrated the direct interaction between MPK3, MPK6, and SPL, leading to the phosphorylation of SPL. This phosphorylation event plays a regulatory role in the early establishment of polarity during anther development [17]. PHB, encoding a transcription factor, specifically governs the adaxial development of lateral organs in Arabidopsis. Studies have shown that microRNA165/6 negatively regulates the expression of PHB. In situ hybridization results revealed opposite expression patterns between microRNA165/6 and PHB during early anther development while exhibiting similar expression patterns with SPL. Further investigations have demonstrated that PHB inhibits SPL expression by directly binding to the SPL promoter, thereby influencing the formation of anther dehiscence zones [8]. In addition, microRNA165/6 also regulates the expression of UCL1 to control anther dehiscence [18].

SPL also exhibits responsiveness to various hormonal processes within anthers. BZR1, a key transcription factor in the brassinosteroid signaling pathway, and its homologous gene mutant, exhibit similar defective anther lobe phenotypes to the spl mutant. RT-PCR and in situ hybridization experiments indicated that BZR1 and its homologue transcription factors regulate early anther morphogenesis through its effects on SPL [19]. The expression of the auxin transport protein gene, PIN1, is reduced in spl mutant [20]. In addition, spl-D, an endogenous overexpression mutant of SPL, shows a leaf polarity defective phenotype with severe leaf curling. Further investigations found significantly downregulated accumulation of auxin and expression of auxin synthesis genes, YUCCA2 and YUCCA6, in spl-D [21].

The establishment of adaxial–abaxial polarity identity, a crucial aspect of proper anther morphogenesis, is regulated by two classes of antagonistic genes [22]. ARF3, also known as ETTIN, is essential for abaxial polarity establishment [3,10,11,23]. ARF3 physically interacts with an abaxial regulator, ABERRANT TESTA SHAPE (ATS), which encodes a KANADI (KAN) transcription factor. This interaction forms a complex that mediates the specification of abaxial cell fate [24]. In rice, an adaxial identity gene OsPHB3, encoding an HD-ZIP III family transcription factor, is expressed in the adaxial domain where stomium differentiation occurs, while OsETT1 is expressed in the abaxial domain formed between thecae [3]. ARF3 and its homologous gene ARF4 redundantly contribute to organ asymmetry by modulating KANADI activity [23]. Both ARF3 and ARF4 are suppressed by a class of ta-siRNAs (tasiR-ARF) derived from TAS3 [25,26,27]. TasiR-ARFs move from the adaxial interface to the abaxial interface, creating a gradient of small RNAs that patterns the abaxial expression of ARF3 in leaf primordia [28,29]. Through ChIP-seq and RNA-seq data analysis, 663 putative direct targets of ARF3 have been found [30]. However, the direct target of ARF3 during early anther development remains unknown.

Taken together, SPL is crucial for anther morphogenesis and AG has been established as a positive regulator of SPL during the initial stages of stamen initiation, while the negative regulator determinants acting upstream of SPL remain elusive. In this investigation, we identified ARF3 as a negative regulatory factor governing SPL expression based on a series of comprehensive analysis. ARF3 exerts its regulatory effects by directly interacting with the SPL promoter, thereby restricting the expression of SPL in the microsporocytes, curbing excessive and deleterious SPL expression levels, and safeguarding the accurate progression of early anther morphogenesis.

2. Results

2.1. The Critical Role of SPL Expression Level in Anther Morphogenesis

Previous investigations revealed that SPL is required for early sporogenesis and germ cell initiation [31]. In this study, we identified and characterized two novel T-DNA insertion mutants of SPL (Figure 1A), namely spl-4 and spl-5. RT-PCR results showed that spl-4 represents a knockout mutant, while spl-5 exhibited an overexpressed phenotype (Figure 1B). To assess the impact of SPL knockout and overexpression on anther morphogenesis and the establishment of four anther lobes, we examined the morphology of mature anthers in these mutants. Scanning electron microscopy results showed that the mature anthers of the wild-type exhibited the typical four-lobe structure (Figure 1C), with mature fertile pollen within the sporangia (Figure 1F). In contrast, spl-4 displayed significant abnormalities in anther polarity establishment (Figure 1D), resulting in the absence of pollen and complete infertility (Figure 1G). Notably, spl-5 displayed dysplasia phenotype in the two adaxial anther lobes, leading to abnormal anther polarity (Figure 1E), although fertile pollen remained unaffected (Figure 1H). These findings underscore the fact that both SPL functional deficiency and overexpression can cause varying degrees of abnormal anther morphogenesis, suggesting the critical importance of maintaining the appropriate expression level of SPL.

To gain further insights into the role of SPL during anther morphogenesis, we conducted semi-thin section experiments to examine the early morphology of spl-4 anthers. In the wild-type anthers, the primary parietal cells and primary sporogenous cells are formed at stage 3 (Figure 1I,R). Subsequently, at stage 5–6, the primary parietal cells differentiate into four layers of somatic cells, while the secondary sporogenous cells develop into microsporocytes (Figure 1J,K,S,T). In comparison to the previously reported phenotype of the SPL loss-of-function mutant spl-1 with no lobe formed, we observed two distinct phenotypes in spl-4. Type I spl-4 anthers exhibited evident vacuolization and nuclear shrinkage in the primary parietal cells and primary sporogenous cells at stage 3 (Figure 1L,U). At stages 5 to 6, these anthers displayed pronounced vacuolization and failed to form anther lobes (Figure 1M,N,V,W), similar to the spl-1 phenotype. Intriguingly, type II spl-4 anthers exhibited no apparent abnormalities in the primary parietal cells and primary sporogenous cells at stage 3 (Figure 1O). At stage 5 to 6, these anthers formed a single anther lobe consisting of four layers of somatic cell cells and one layer of germ cells (Figure 1P,Q).

2.2. ARF3 Suppresses SPL Expression in Arabidopsis Anthers

Previous studies demonstrated that AG positively regulates SPL, thereby initiating microsporogenesis [14]. Given the crucial requirement of SPL for maintaining the correct expression, as deviations in either direction lead to an abnormal anther polarity development, it becomes imperative to identify negative regulatory factors acting upstream of SPL. We first explored the upregulation of SPL expression resulting from gene mutations using available published data. Transcriptome analysis revealed that elevated SPL expression in arf3 mutant, and ChIP data showed that ARF3 can bind to the SPL promoter [30,32]. To validate the regulatory role of ARF3 in SPL expression, we assessed the expression level of SPL in the arf3-29 mutant. The results unequivocally demonstrated significant up-regulation of SPL expression in arf3-29 compared to the wild-type (Figure 2A). To further ascertain the direct regulation of SPL by ARF3, we detected SPL expression after treatment with dexamethasone (DEX) in ARF3::ARF3-GR arf3-29 plants. DEX treatment induces the translocation of ARF3-GR fusion protein into the nucleus, thereby restoring the phenotype of arf3-29 [33]. RT-PCR results showed a slight downregulation of SPL expression after 4 h of DEX treatment (Figure 2B), with a statistically significant downregulation observed after 6 h of DEX treatment (Figure 2B), providing compelling evidence for the direct regulatory role of ARF3 in moduling SPL expression.

To investigate the spatial regulation of SPL by ARF3, we examined the expression pattern of SPL in wild-type and arf3-29 mutants. In the wild-type, SPL exhibited preferential expression in the archesporial cells, secondary sporogenous cells, and microsporocytes at stage 2 to 5 (Figure 2C–E). However, in arf3-29, the expression of SPL was observed to be preferentially located to one side of the archesporial cells at stage 2 (Figure 2G). At stage 4, SPL expression was concentrated in the secondary parietal cells, secondary sporogenous cells, and the middle region surrounding the adaxial and abaxial lobes on one side (Figure 2H). By stage 5, SPL expression was detected in the microsporocytes, middle layer, the endothecium, and the tapetum (Figure 2I), indicating an expanded expression range of SPL in arf3-29. Subsequently, we examined the expression pattern of ARF3 in early anthers. mRNA in situ hybridization revealed preferential expression of ARF3 in the vascular tissues during stage 2 to 5 (Figure 2K–M), with slight expression observed in the microsporocytes at stage 5 (Figure 2M). These findings indicated an inverse expression pattern between ARF3 and SPL in anthers. The aforementioned results together suggested that ARF3 exerts negative regulation on SPL at both the transcriptional and spatial levels.

2.3. ARF3 Is Required for the Establishment of Four Anther Lobes

Anther morphogenesis occurs from the Shoot Apex stage to Flower Stage 9. Analysis of publicly available data revealed high expression levels of ARF3 and SPL during this period (Figure 3A), suggesting their crucial roles in the process of anther morphogenesis. To further investigate the function of ARF3 in anther development, we examined the anther polar developmental phenotype of arf3-29 mutants. Compared to wild-type anthers that display the characteristic four-lobed structure (Figure 3B), the arf3-29 mutants showed defects in anther lobe formation (Figure 3C,D). Cytological experiments further confirmed the presence of anther lobe defects in arf3-29 mutants (Figure 3E–G), highlighting the involvement of ARF3 in regulating anther morphogenesis.

Given the phenotypic similarity between ARF3 loss-of-function mutants and SPL overexpression mutants, we aimed to investigate whether ARF3 and SPL are part of the same signaling pathway. To address this, we generated the arf3-29 spl-1 double mutant through hybridization. Scanning electron microscopy analysis revealed similar phenotypes between spl-1 and arf3-29 spl-1, characterized by significant abnormalities in polarity development (Figure 4A–C) and an inability to form pollen (Figure 4D–F). Furthermore, semi-thin section results demonstrated highly vacuolated anther cells and the absence of any anther lobes in both mutants (Figure 4G–I). These results strongly suggest that ARF3 acts upstream of SPL in the regulation of the pathway governing anther development.

2.4. ARF3 Directly Interacts with the SPL Promoter and Suppresses AG-Mediated SPL Activation

To investigate the direct interaction between ARF3 and SPL promoter, we performed chromatin immunoprecipitation (ChIP) assays in vivo. Floral primordia of ARF3::ARF3-GFP 35S::AP1-GR cal ap1 inflorescences were treated with DEX for 5 days to reach floral stage 9 (anther stage 5). The results demonstrated that ARF3 is preferentially bound to the P1 promoter elements rather than P2, with higher binding observed at anther stage 5 (Figure 5A,B). To further confirm the direct binding of ARF3 to the SPL promoter, we performed electrophoretic mobility shift assays (EMSA). As shown in Figure 5C, ARF3 directly bound to the P1 elements of the SPL promoter, and both TGTCTC boxes were found to be essential for the interaction, while ARF3 did not bind to elements P2 of SPL promoter (Figure 5D). A previous study reported that AG directly activates SPL during microsporogenesis [14]. Therefore, we investigated whether ARF3 affects the transcriptional activation of AG on the SPL promoter using a transient transcription assay. In this assay, we connected 1 kb 3′-untranslated region to a 3.8 kb promoter of SPL (Figure 5E). The expression level of the SPL promoter-LUC construct was approximately 8-fold higher in its presence compared to the control (GFP). However, when ARF3 and AG were co-expressed, ARF3 repressed this increase in expression (Figure 5F).

2.5. ARF3 and SPL Oppositely Regulate Early Anther Morphogenesis

To investigate the functional role of ARF3 and SPL in anther morphogenesis, we conducted a transcriptome analysis. In the arf3-29 mutants, 1312 (73% of a total of 1807 differentially expressed genes) were found to be down-regulated, while 495 (27%) were up-regulated compared to the wild-type. Similarly, in the spl-1 mutants, the expressions of 3479 genes were down-regulated, and 3999 genes were up-regulated (47% and 53% of the 7478 total altered genes, respectively) (Figure 6A). Among these, 739 genes were identified as commonly differentially expressed between arf3-29 and spl-1 (Figure 6B), with 28.7% exhibiting opposite regulatory patterns (11.6% of genes are up-regulated in arf3-29 and down-regulated in spl-1, while 17.1% of genes are down-regulated in arf3-29 and up-regulated in spl-1) (Figure 6C–E). Gene Ontology analysis revealed that the common DEGs which regulate in opposite directions (arf3-29 down spl-1 up and arf3-29 up spl-1 down) are enriched in morphogenesis related pathways (Figure 6F). These findings indicate that ARF3 and SPL exert opposing regulatory effects on anther morphogenesis. Interestingly, the genes specifically regulated in arf3-29 or spl-1 were also enriched in morphogenesis-related pathways (Figure 6G), indicating that ARF3 and SPL each contributed to the regulation of anther morphogenesis through distinct pathways.

To further validate the opposing regulation of anther development by ARF3 and SPL, we examined the expression of marker genes for anther development. bHLH010, bHLH089, bHLH091, AMS, MS1, MYB35 serve as markers for anther development, while ROXY2 and ROXY3 are associated with early anther morphogenesis. In arf3-29 mutants, these genes were up-regulated, whereas in spl-1 mutants, they were down-regulated, supporting the opposite regulatory roles of ARF3 and SPL in anther development (Figure 7A–H).

3. Discussion

The process of organogenesis in plants differs from that in animals. While animal organs are typically formed during embryogenesis, plants continuously undergo organogenesis from stem cells throughout their life cycle [34,35,36,37]. Early morphogenesis, which involves the development of plant organs from their initial primordia into polar structures, is a crucial stage. Different from leaf development, which requires once polar establishment, anther early morphogenesis requires once polar establishment and once polar reversal [38], and ultimately forms a four-lobe structure. Therefore, anther development represents an excellent model to study plant organogenesis and morphogenesis, as it requires both polar establishment and polar reversal to form a four-lobed structure consisting of somatic and germ cells.

In this study, we investigated the role of SPL in early anther morphogenesis. We identified two mutants, spl-4 and spl-5, with disrupted SPL expression (Figure 1A–C). The spl-4 anthers exhibited abnormal anther polarity dysplasia and no pollen production, while the spl-5 had significantly reduced size in two adaxial lobes (Figure 1E,F,H). These results indicated that functional deficiency or overexpression of SPL can lead to a different degree of abnormal anther polarity development. Through analysis of the public database, we identified ARF3 as an upstream regulatory element of SPL. We found that ARF3 negatively regulates the expression of SPL at both the transcriptional and spatial levels. In the absence of ARF3, SPL expression was up-regulated, and its expression domain expanded beyond the microsporocytes to include the tapetum, middle layer, and endodermis (Figure 2A–J). Moreover, the lack of ARF3 function resulted in a defect in the number of anther lobes (Figure 3B–G). By generating the arf3-29 spl-1 double mutant (Figure 4A–I), we confirmed that ARF3 acts upstream of SPL in the same signaling pathway. Next, we confirmed that ARF3 can directly bind to the SPL promoter in vivo.

Through chromatin immunoprecipitation (ChIP) assays and electrophoretic mobility shift assays (EMSA) (Figure 5). we demonstrated that ARF3 directly binds to the P1 region of the SPL promoter, specifically to the TGTCTC elements. We also found that ARF3 suppresses the activation effect of AG (AGAMOUS) on SPL in a transient transcription assay. Transcriptome analysis of arf3-29 and spl-1 mutants revealed differentially expressed genes, particularly those with opposite expression patterns, which were significantly enriched in morphogenesis-related pathways. This suggests that ARF3 and SPL regulate anther morphogenesis in opposite directions and contribute to anther morphogenesis through distinct pathways. Previous studies highlighted the positive regulatory roles of AG and BES1 (BRI1-EMS-SUPPRESSOR 1) on SPL during microsporogenesis; negative regulatory factors upstream of SPL were poorly understood. Our identification of ARF3 as a negative regulator provides new insights into the regulatory network of SPL in anther polarity development.

In conclusion, this study enhances the understanding of the regulatory mechanisms involved in early anther morphogenesis. The ARF3-SPL module plays a crucial role in anther polarity development, but additional pathways are also involved. The research sheds light on the complex processes underlying plant organogenesis and provides new evidence and insights into the regulatory network governing early anther morphogenesis.

4. Materials and Methods

4.1. Plant Materials and Growth Conditions

Arabidopsis thaliana plants of Ler or Col-0 background were planted in a greenhouse under conditions of a 16 h light/8 h dark cycle at 22 °C and 65% humidity. All mutants and transgenic A. thaliana lines are in the Ler or Col-0 background. The arf3-29 [39] ARF3::ARF3-GR arf3-29 [33] and spl-1 [13] were previously described. The spl-4 (SAIL_827_A10) and spl-5 (SALK_044645) mutants were obtained from the Nottingham Arabidopsis Stock Centre. The arf3-29 spl-1 double mutants were generated by crossing the related single mutants.

4.2. Phenotypic Analysis of Anthers

To assess pollen viability, stage 12 anthers were collected and subjected to staining with Alexander solution at 65 °C for 1 h [40]. Images of the stained anthers were captured using a digital camera (Nikon, Kyoto, Japan). For the transverse semi-thin sections assay, inflorescences were embedded in Spurr resin, cut into 1 μm thick slices, stained with toluidine blue, and visualized using a Nikon digital camera. Scanning electron microscopy was performed to photograph stage 12 anthers using a TM-3000 (Hitachi, Kyoto, Japan) scanning electron microscope.

4.3. mRNA In Situ Hybridization

Gene-specific RNA probes were synthesized by in vitro transcription using the DIG RNA Labeling Kit (Roche, Basel, Switzerland). The samples were fixed in FAA (3.7% formaldehyde, 5% acetic acid, 50% ethanol) and embedded in wax. The embedded inflorescences were sectioned into 7 μm slices and subjected to dewaxing, rehydration, and dehydration. The sections were then hybridized with the prepared RNA probes at 55 °C overnight. After washing in SSC buffer, the slices were incubated with an anti-digoxigenin-AP antibody (Roche, Basel, Switzerland) at room temperature for 2.5 h. Hybridization signals were detected using NBT/BCIP color reaction (Roche).

4.4. ChIP-qPCR Analysis

Chromatin was isolated from 40 g of frozen tissue and fragmented by sonication to yield approximately 500 bp fragments. Immunoprecipitation was performed by incubating the chromatin with anti-GFP beads (Chromotek, Munich, Germany) overnight at 4 °C. The protein–chromatin complexes were then washed four times to eliminate nonspecific binding. The precipitated DNA was quantified by qRT-PCR. Primers specific for ChIP-qPCR assays are provided in Supplementary Table S1. The binding level was determined by calculating the ratio between IP and MOCK samples, normalized to the internal control ACT.

4.5. Electrophoretic Mobility Shift Assay

The coding regions of ARF3 were cloned into the pGEX4T-1 vector, and the resulting GST fusion protein was expressed in E. coli Rosetta cells. The empty pGEX4T-1 vector was used as a control. Biotin-labeled and unlabeled DNA probes containing the auxin responsive elements (TGTCTC) in the SPL promoter were designed. In vitro binding experiments were performed using the Light Shift Chemiluminescent EMSA system (Thermo Scientific, Waltham, MA, USA). The reaction mixtures were incubated in binding buffer (10 mM Tris-HCl (pH 7.5), 40 mM KCl, 2.5 mM MgCl₂, 1 mM EDTA, 3 mM DTT, and 10% (v/v) glycerol) on ice for 30 min. The samples were then electrophoresed on 6% native polyacrylamide gel at 4 °C. The primers used for the EMSA probes are listed in Supplementary Table S1.

4.6. Transient Transcription Dual-Luciferase Assay

The coding regions of ARF3 and AG were cloned into p1306-GFP binary vectors. The promoter sequence and 3′-end non-coding sequence of SPL were connected and cloned into pGreenII-0800-LUC vectors. Agrobacterial cell suspensions containing TF protein(s) and a promoter construct were mixed at a ratio of 4:1 and infiltrated into young leaves of Nicotiana benthamiana using a published method [41]. After 48 h of cultivation, the leaf cells were lysed with Passive lysis buffer, and the luciferase activity was measured using the Dual-Luciferase assay kit (Promega, Madison, WI, USA). The firefly luciferase luminescence was detected, followed by the measurement of after quenching. The primers used for all constructs are listed in Supplementary Table S1.

4.7. RNA Sequencing and Data Analysis

Approximately 20 inflorescences containing stage 1–9 anthers from Ler and arf3-29 plants were collected and immediately frozen in liquid nitrogen. Total RNA from each sample was extracted using the ZR plant RNA Miniprep™ kit (Zymo Research, Irvine, CA, USA) [30]. Then 3 μg total RNA of each sample was subjected to deep sequencing using an Illumina™ Hi-seq 2000 system (Illumina Ins., San Diego, CA, USA). The sequencing data were mapped and analyzed following a previously reported method [30]. The transcriptomic data of spl-1 were obtained from a published study [42].

4.8. Quantitative Real-Time PCR Analysis

Arabidopsis inflorescences from various mutants were frozen in liquid nitrogen. RNA was extracted using the TRIzol (Invitrogen, Carlsbad, CA, USA) method and treated with DNase I (Takara, Kyoto, Japan). The cDNA was obtained with a reverse transcription system (Takara). Real-time PCR was performed using SYBR premix Ex Taq II (Takara) on a Bio-Rad Real-Time system (Bio-Rad, Hercules, CA, USA), with primers listed in Supplementary Table S1.