Overexpression of TCP5 or Its Dominant Repressor Form, TCP5-SRDX, Causes Male Infertility in Arabidopsis

1. Introduction

TEOSINTE BRANCHED1, CYCLOIDEA, and PROLIFERATING CELL FACTORS (TCP) constitute a family of plant-specific transcription factors that exist from green algae to higher plants [1]. All known TCP members share a conserved TCP domain of 59 amino acids, forming a non-canonical basic helix-loop-helix structure responsible for DNA binding, protein-protein interaction, and dimerization [1,2,3]. TCP family members are classified into two classes, I and II, based on the features of the sequences within and outside the TCP domain. The class I members lack four amino acids in the basic region but include 6-7 more amino acids in the second helix than class II [1]. In addition, class I TCPs have short sequences flanking the TCP domain, but most class II TCPs possess an arginine-rich region or R domain that may be crucial for protein-protein interaction [4]. Phylogenetically, the class I TCPs are closely related, whereas the class II TCPs can be further divided into the CYC/TB1 and CIN (CINCINNATA-like) clades [3]. However, it is uncertain which class of TCP genes is more ancient, as even basal plant species have two classes of TCPs even though these species only have a few TCPs [1]. Thirteen of the 24 TCP members in the Arabidopsis genome are members of class I, while 11 are members of class II, which is composed of the CYC/TB1 (TCP1, TCP12/BRC2, and TCP18/BRC1) and CIN (TCP2/3/4/10/24 and TCP5/13/17) clades. In the CIN clade, miR319 regulates the TCP2/3/4/10/24 genes [5].

TCPs act as transcriptional regulators, binding to the promoters of their target genes and regulating their expression. Both class I and II TCPs bind to GC-rich sequences, but their binding preferences differ. Class I TCPs prefer the sequence GTGGGNCC, while class II prefers GTGGNCCC [3,5]. The two TCP classes might share target genes by binding to the same and different cis-elements or an overlapping promoter sequence [1,3]. TCPs can form homodimers or heterodimers from the same or other classes. However, heterodimers between specific members of the same class are more predominant. These TCP dimers reportedly showed higher binding efficiencies [1].

Class I TCP genes generally promote cell division and plant growth, whereas class II genes inhibit cell division and regulate differentiation [6,7]. Genetic analysis of TCP single and multiple mutants or transgenic plants with overexpression of TCP and its dominant repressor form (TCP-SRDX) revealed that TCP genes exert a crucial role in various developmental processes, including seed germination [8,9,10], hypocotyl growth [11,12,13], trichome initiation and development [14,15], root growth [16], leaf patterning and development [17,18,19,20,21,22,23,24,25], axillary bud outgrowth [26], axillary branching [27], flowering [28,29,30], circadian clock [31], plant hormone biosynthesis, and signaling pathways [3,32].

Several studies reported that the TCP genes also regulate the development of floral organs, including petals, stamens, filaments, gynoecium, and ovules [33,34,35,36,37,38,39]. TCP15, a class I gene, is expressed in stamen filaments. Down-regulation of TCP15-related class I TCPs, or expression of TCP15 dominant repressor form, resulted in short stamens. TCP15 and its related class I TCPs modulate GA-dependent stamen filament elongation by directly targeting SAUR63 family genes, while KNAT1 directly regulates the expression of TCP15 [36,40]. TCP4, a miR319-targeted CIN-like gene, inhibits petal and stamen development when expressed in a miR319a mutant background, indicating that a functional TCP4 is critical for developing these organs [35]. However, how TCP4 influences stamen development needs to be determined. Up-regulation of OsTCP1/PCF5, rice’s TCP4 orthologue, resulted in a premature anther opening phenotype via the JA-dependent pathway [41]. TCP24, another miR319-targeted CIN-like gene, was found to inhibit secondary cell wall thickening in anther endothecium. TCP24 overexpression in Arabidopsis resulted in non-dehiscent anthers, which led to a male sterile phenotype [42].

Recently, miR319 and its target TCPs (TCP2/3/4/10/24) were shown to control early anther development [43], but there is no information on other TCPs. TCP5 is a non-miR319-targeted member of the same clade (CIN-like) that regulates petal growth and differentiation [33,34]. Our study discovered that TCP5 is expressed in the early anthers. Overexpression of TCP5 and its dominant repressor form, TCP5-SRDX, resulted in smaller anthers with fewer pollen sacs and pollen. In addition, TCP5 overexpression, like TCP24, resulted in a non-dehiscent anther phenotype, implying that TCP5 may have a redundant function with other CIN-like TCPs during anther development.

2. Results

2.1. Overexpression of TCP5 Resulted in Male Sterility

We first acquired a T-DNA mutant (SM_3_29639) from Nottingham Arabidopsis Stock Center to examine the putative function of TCP5 in the reproductive phase. The T-DNA mutant was null, as validated by the RT-PCR data (Figure S1), but neither male nor female organs displayed any apparent abnormalities. We then overexpressed TCP5 in Arabidopsis using the UBQUITIN 10 promoter [44]. The transgenic plants developed longer and slightly narrower rosette leaves than the wild-type (Figure 1A, B). During the flowering stage, the transgenic plants exhibited a dwarf and slightly bushy phenotype (Figure 1C), with the apex of their inflorescence usually twisted and the flowers arranged in clusters (Figure 1C and Figure S2). These flowers produced few to no seeds. However, they produced more seeds when pollinated with wild-type pollen, suggesting that the transgenic plants’ sterility is primarily due to the male side.

We dissected the flowers of transgenic plants with severe phenotypes using a stereomicroscope. The transgenic flowers were slightly smaller than the wild-type (Figure 1D, H). The sepals and petals were morphologically normal, but the stamens, which consisted of short filaments and small anthers, surrounded a gynoecium with fewer stigmatic papillae (Figure 1E, I). Notably, these anthers did not shed pollen during the anthesis stage (Figure 1F, J). However, they produced living pollen inside, as evidenced by Alexander staining (Figure 1G, K), indicating that these anthers could not dehisce.

We examined TCP5 expression levels in several transgenic plants with varying degrees of sterility. We discovered that plants with a more severe phenotype had more TCP5 transcripts, revealing a positive correlation between male sterility and TCP5 expression levels (Figure S3).

2.2. TCP5 Overexpression Reduced Lignin Accumulation in the Anther Endothecium

We used phloroglucinol staining and investigated the lignin content in the anthers of wild-type and TCP5-overexpressing plants with severe phenotype to determine the cellular basis of the anther non-dehiscence (Figure 1C, L15 in Figure S3). After staining, a dark red hue appeared in the wild-type anther loci. However, this color was much reduced in the anthers of TCP5 overexpression plants (Figure 2A, B), indicating a lower lignin level in these transgenic anthers than in the wild-type. Sufficient lignin accumulation on the anther endothecium is essential for anther dehiscence during anthesis. We thus focused on this layer after anther clearing. The continuous fiber band formation on the endothecium in the wild-type anthers was indicative of constant thickening of the secondary cell wall (Figure 2B, C). However, the TCP5 overexpression anthers exhibited discontinuous fiber band formation, meaning no fiber bands developed on specific endothecium layer cells, particularly in the upper half of the anther (Figure 2F, G). A cross-section of the anther at anthesis further supported the difference in fiber band distribution on the endothecium between the wild-type and the TCP5 overexpression plant (Figure 2D, H). The wild-type developed dense fiber bands on the endothecium layer. In contrast, the TCP5 overexpression plant had a few fiber bands on the counterpart region, though we occasionally found dense fiber bands near the connective tissue.

These findings demonstrate that TCP5 overexpression reduced lignin accumulation on the endothecium, resulting in a non-dehisced anther phenotype.

2.3. TCP5 Overexpression Resulted in Fewer Pollen Sacs and Pollen

TCP5 overexpression transgenics had fewer pollen sacs in their anthers and a dehiscence defect (Figure 3). The wild-type anther had four pollen sacs. However, transgenic plants with a severe phenotype developed two or three pollen sac anthers, with an average of 2.5 pollen sacs per anther (Figure 4A). Consequently, the transgenic plants also exhibited reduced pollen numbers. On average, the wild-type anther contained 321 pollen grains, but TCP5 overexpression plants generated only 107-274 pollen grains (Figure 4B).

The defect in the pollen sac formation occurred before the meiosis stage. In the stage 4 anthers [45], sporogenous cells appeared in all four corners of the wild-type but only two or three in the TCP5 overexpression plants (Figure 3A, B). While the anthers from the TCP5 overexpression plants only had two or three loci at stages 6–8, the wild-type anther had four. Meiosis appeared to proceed normally in the anther loci of both wild-type and TCP5 overexpression plants (Figure 3C-F). At stage 9, microspores developed in the anther loci, and the tapetum gradually degenerated (Figure 3G, H). At stages 11-12, the fiber band develops on the endothecium of the anthers of both wild-type and TCP5 overexpression plants, but less in the latter (Figure 3I, J). During anthesis, stage 13, the wild-type anthers with dense fiber bands on the endothecium opened to release pollen (Figure 3K). The anthers of TCP5 overexpression plants with fewer fiber bands on the endothecium could not open despite the septum disappearing and the stomium partially breaking down, resulting in no pollen release from these anther loci (Figure 3L). This was consistent with our observations in Figure 2.

2.4. TCP5 Overexpression Affected the Expressions of Genes Involved in Secondary Cell Wall Formation and Other Key Anther Genes

We investigated the expression level of genes involved in secondary cell wall formation and several vital anther genes to comprehend the molecular mechanism underlying the defective anther phenotype in the TCP5 overexpression plants.

The secondary cell wall predominantly comprises cellulose, hemicellulose, and lignin. We used qRT-PCR to determine the expression levels of genes involved in cellulose biosynthesis (IRX1, IRX3, IRX5, CESA1, IRX8, and IRX9) [46,47,48,49] and lignin biosynthesis (PAL4, C4H, C3H, 4CL1, COMT1, CCOAMT, HCT, CAD4, CAD5, CAD6, F5H1, and CCR1) [50]. Overexpression of TCP5 in plants resulted in down-regulation of all genes except IRX1 and IRX5, which remained unchanged, and CCOAMT, which increased expression (Figure 5A). In addition, three transcription factor genes, MYB26 and its downstream targets NST1 and NST2, which influence secondary cell wall thickness [51,52,53,54], were down-regulated. AHP4, a negative regulator of secondary cell wall thickening [55], was significantly up-regulated in these TCP5 overexpression plants compared to the wild-type (Figure 5A).

Several genes are known to be activated during anther development. Twenty genes were selected to compare their expression levels in wild-type and TCP5 overexpression plants. Except for TPD1, a tapetum determinant gene [56], which showed no significant expression change, all nineteen genes were down-regulated in TCP5 overexpression transgenics (Figure 5B). The most down-regulated genes were BAM1 and BAM2, two receptor kinase-like genes involved in cell fate specification [57], and AMS and DYT1, two bHLH transcription factors that regulate tapetum development [58,59]. The second group consisted of two glutaredoxin genes, ROXY1 and ROXY2, and two TF genes, SPL9 and TGA9 [60,61,62]. The third group included SPL8 and SPL2 as well as SERK1 and EMS1, which control anther sporogenesis, SPL/NZZ, a well-studied early anther gene, and BIM1, which is involved in brassinosteroid signaling [61,63,64,65,66,67]. Next came MYB33 and TGA10 [62,68], whose expression levels were lowered by one-third compared to the wild-type. SPL11, MYB65, and SERK2, the final group of genes [61,65,68], had their expression levels reduced by two-thirds, slightly more than half, and one-half, respectively, compared to the wild-type (Figure 5B).

These findings suggest that overexpressing TCP5 can influence the expression of numerous genes involved in secondary cell wall thickening and early anther development.

2.5. TCP5 Was Expressed in the Early Anthers

We conducted RT-PCR analysis in various organs and flower buds or flowers at different developmental stages to investigate the TCP5 expression profile. The highest expression signal was detected in the cauline leaf, followed by a moderate expression level in the seedling, rosette leaf, and inflorescence, a low expression signal in the young silique, a feeble signal in the stem and older silique, and no signal in the root (Figure S4A). Furthermore, the TCP5 expression signal was observed in all flower buds and open flowers, with the expression level slightly lower in open flowers (Figure S4B). Contrarily, a robust GUS signal was discovered in the open flowers, including the sepal, petal, stamen, and upper part of the gynoecium, a weak signal in the young flower buds, and no signal in the early anthers (Figure S4C, D).

Consequently, we performed RNA in situ hybridization for TCP5 in wild-type anthers. At stages 2-3, the TCP5 expression signal was evenly distributed throughout the region, with a strong signal in the corners (Figure 6A, B). Stages 4-5 saw the formation of the four lobes on the anthers. The TCP5 signal was detected in the epidermis, endothecium, middle layer, tapetum, and sporogenous cells (or pollen mother cells/PMCs), as well as vascular bundles in the connective region (Figure 6C, D). At stages 6-8, the anthers underwent meiosis, and a strong TCP5 signal appeared in the tapetum, PMCs, or tetrads, and the vascular bundles (Figure 6E, F, G). Following meiosis, the TCP5 signal disappeared from stage 9 to stage 13 anthers (Figure 6H, I). Thus, we concluded that TCP5 was expressed in the early anthers, contradicting the expression pattern observed in the GUS plants (Figure S4C, D).

2.6. TCP5-SRDX Transgenic Plants also Had Male Fertility Defects

We used the UBQUITIN 10 and SPL8 promoters to create TCP5-SRDX transgenic plants to rule out putative functional redundancy with other TCPs. SRDX is a sequence that encodes an EAR-motif-based artificial repression domain (LDLDLELRLGFA), which provides the fused TF with dominant transcriptional repression activity [69]. Both pUBQ10:TCP5-SRDX and pSPL8:TCP5-SRDX transgenic plants appeared normal, except for the absence of elongated siliques after flowering, at least for the early-arising flowers. Their flower size was comparable to the wild-type (Figure 7A, D, G). However, their anthers were noticeably smaller and produced less pollen than the wild-type (Figure 7B, C, E, F, H, I). This phenotype appeared to be substantially more severe in the pUBQ10:TCP5-SRDX transgenic plants. Interestingly, these anthers could open to release even less pollen at anthesis (Figure 7F, I). Cross-sections and Alexander staining of these anthers revealed fewer pollen sacs and little or no pollen (Figure S5 and Figure 7J, K, L). The phloroglucinol staining showed that the dark-red color of the lignin in the majority of these anthers was comparable to the wild-type (Figure 7M, N, O), indicating the ability of these anthers to dehisce.

2.7. TCP5-SRDX Transgenic Plants Exhibited Altered Expression of Early Anther Genes

Expression of TCP5-SRDX in Arabidopsis resulted in undersized anthers with fewer pollen sacs and less pollen (Figure 7 and Figure S5). Thus, we compared the expression levels of the twenty early anther genes (Figure 5B) in wild-type and TCP5-SRDX transgenic plants. All genes were significantly down-regulated in the pUBQ10:TCP5-SRDX transgenics, except SPL2, which showed no discernible change in expression level (Figure 8A). AMS was the most down-regulated gene, followed by BAM1/2, MYB33, BIM1, TGA9/10, and SPL/NZZ. TPD1, SERK1/2, ROXY2, SPL9, SPL11, and MYB65 were next, while the rest, including EMS1, DYT1, SPL8, and ROXY1, were less down-regulated, with expression levels not lowered by more than 50% when compared to wild-type (Figure 8A). Most of these genes were also down-regulated in the pSPL8:TCP5-SRDX transgenic plants. AMS was similarly the most down-regulated gene, with nearly undetectable expression (Figure 8B). TGA9 and DYT1, the second most down-regulated genes, had their expression levels drop below 10%. However, unlike in pUBQ10:TCP5-SRDX transgenics, ROXY1 and SPL11 expression levels did not alter significantly, although SPL9 expression levels increased dramatically (Figure 8).

Nevertheless, these results demonstrated that most early anther genes were down-regulated in TCP5-SRDX transgenic plants.

3. Discussion

Plant fertility refers to the male and female activities in the flower and is affected by various factors in the reproductive process. The male reproductive organ, stamen, located in the flower’s third whorl, consists of an anther and a filament. The anther, a butterfly-like structure that produces, stores, and releases pollen, is located at the tip of the filament, a slender stalk that connects the anther to the floral central axis [70]. Any developmental changes or defects in these structures may impact male fertility, which involves numerous genes or regulatory networks.

Our study discovered that TCP5 is involved in anther development. TCP5-like genes (TCP5, TCP13, and TCP17) may have functional redundancy because the single knock-out mutant tcp5 did not exhibit an apparent phenotype in anther development as expected. Thus, the triple mutant tcp5 tcp13 tcp17 requires further investigation. Overexpressing TCP5 in Arabidopsis resulted in a non-dehiscent anther phenotype at anthesis (Figure 1). Further investigation revealed that the endothecium in the anther of the TCP5 overexpressing plant exhibited decreased lignin accumulation and fibrous band formation (Figure 2), which could account for the anther’s non-dehiscence. TCP24, a miR319-targeted gene, demonstrated a similar anther phenotype when overexpressed [42]. However, it appears more severe than TCP5 overexpression plants because of the total absence of lignin accumulation in the anther endothecium [42]. TCP5 and TCP24 were both expressed in early anthers, including the epidermis, endothecium, middle layer, tapetum, and PMCs, but not in late anthers in wild-type plants. The endothecium of late anthers did not exhibit any expression of these two genes during the secondary cell wall thickening. Thus, we conclude that TCP5 and TCP24 act as negative regulators of secondary cell wall thickening, preventing precocious secondary cell wall thickening in the early anther endothecium. Intriguingly, the TCP24 signal was not detected in the endothecium layer of the late anthers in the TCP24 overexpression plants. This indicates that the non-dehiscent anther phenotype resulted from enhanced TCP24 expression rather than ectopic expression [42]. Thus, we need to examine the detailed expression profile of TCP5 in the TCP5 overexpression plants. TCP5 and TCP24 are both members of the CIN-clade, although they belong to separate subclades (TCP5-like and Jaw-D/miR319-targeted), suggesting that other CIN-like genes may perform similar functions. Many genes related to cellulose (IRX1/3/8/9 and CESA1) and lignin biosynthesis (PAL4, C3H, C4H, 4CL1, COMT1, CAD4/5/6, HCT, F5H1, and CCR1) were down-regulated in TCP5 overexpressing plants which supported the existence of their non-dehiscent anther phenotype (Figure 5). A previous study also showed the down-regulation of IRX1/3/5, C4H, 4CL1, CCOAMT, and PAL4 genes in TCP24 overexpression plants [42]. Both TCP5 and TCP24 overexpression plants showed increased levels of AHP4 (Arabidopsis Histidine-containing Phosphotransfer Factor 4), which is consistent with a prior finding that overexpression of AHP4 results in a non-dehiscent anther phenotype [55]. MYB26 and its two targets NST1 and NST2, three transcription factors that regulate secondary cell wall thickening, were down-regulated in TCP5 overexpression plants, but MYB26 exhibited no significant change in TCP24 overexpression plants [42]. Several other discrepancies exist between these two transgenic plants. In the TCP5 overexpression plants, CCOAMT was up-regulated, while IRX5 did not exhibit any significant expression change. In contrast, these two genes were down-regulated in the TCP24 overexpression plants. These expression differences for some of the genes mentioned above indicate that TCP5 and TCP24 somewhat differ in regulating secondary cell wall thickening when overexpressed.

Another distinguishing feature of the TCP5 overexpression plant was a slightly smaller anther with fewer pollen sacs and pollen (Figures 3, 4). The TCP5-SRDX transgenic plants, where TCP5 functioned as a transcriptional repressor, remarkably exhibited a similar anther phenotype, except for anther dehiscence (Figure S5). The lignin intensity in the anthers of TCP5-SRDX plants was comparable to that of wild-type, as indicated by a dark-red hue from phloroglucinol staining (Figure 7). This differs from TCP24, in which TCP24-SRDX plants accumulated more lignin in anthers than wild-type plants [42]. It was unclear whether the TCP24 transgenic plants produced fewer pollen sacs and pollen in the anthers. Nonetheless, a recent study reported that miR319 and its target genes TCP2/3/4/10/24 contribute to early anther development. Multiple mutants of these genes exhibited a similar early anther phenotype as TCP5 overexpression and TCP5-SRDX plants [43]. These findings demonstrated that CIN-like TCP genes are involved in early anther development.

Other genes, such as SPL/NZZ, ROXY1/ROXY2, TGA9/TGA10, and SPL8/miR156-SPL, have previously been shown to control the early anther development, affecting pollen sac formation and pollen production [60,61,62,64,67]. Single and/or multiple mutants of these genes resulted in fewer pollen sacs or even pollen sac anthers, such as those from TCP5 overexpression/TCP5-SRDX plants and mir319a/b/c mutants [43]. SPL/NZZ has been shown to recruit TPL (TOPLESS) and interact with CIN-like TCPs, including TCP5, to repress the expression of these TCP genes [71,72]. TCP4, miR319’s target, directly binds to the promoters of TGA9/TGA10 and ROXY2, repressing their expression [43]. Many early anther genes, including SPL/NZZ, BAM1/BAM2, AMS, ROXY1/ROXY2, TGA9/TGA10, SPL8, and others [57,58,60,61,62,64,67], were significantly down-regulated in both TCP5 overexpression and TCP5-SRDX plants (Figure 8), indicating that a bona fide TCP5 expression is critical for early anther development. However, it is unknown whether TCP5 can bind to the cis-elements of these genes to regulate their expression or interact with them to govern early anther development.

4. Materials and Methods

4.1. Plant Materials and Growth Conditions

Arabidopsis thaliana ecotype Columbia-0 (Col-0) was used as the wild-type. All other plants used in this study were developed in the Col-0 background. A tcp5 mutant (SM_3_29639) was acquired from the Nottingham Arabidopsis Stock Centre (NASC). Seeds were sown in a plastic petri dish lined with wet tissue paper and stratified for two days in the dark at 4 °C. The stratified seeds were transferred into pots filled with a mixed substrate of nutrient soil: vermiculite (1:1). They were grown in a greenhouse with light intensity 120 µmol/m²/s at 22–23°C under long-day conditions (16 h light/8 h dark).

4.2. Plasmid Construction and Transgenic Plants

The pUBQ10:TCP5 and pUBQ10:TCP5-SRDX constructs were created using gateway cloning. The full-length coding sequence (CDS) of TCP5 was amplified by PCR from wild-type Arabidopsis inflorescence cDNA, and the TCP5-SRDX sequence (TCP5-CDS-without stop codon fused with CTTGATCTTGATCTTGAACTTAGACTTGGATTTGCTTAA) was synthesized directly by Sangon Biotech, Shanghai. These sequences were added to pDONR207 separately using BP reactions and then transferred into the destination vector pUB-Dest [44] using LR reactions. A pOLE:OLE-TagRFP cassette expressed in seeds from a pDe-CAS9 derivative [73,74,75] was cut off and integrated into pUBQ10:TCP5 and pUBQ10:TCP5-SRDX, respectively, to allow for visual screening of the transgenic plants. The 2.1 kb SPL8 promoter, expressed in early anther [61], was acquired via PCR from wild-type leaf genomic DNA for the synthesis of pSPL8:TCP5-SRDX, and the pUBQ10 promoter in pUBQ10:TCP5-SRDX was replaced through appropriate restriction enzyme digestion and ligation. A 5173 bp TCP5 promoter was amplified on the wild-type genomic DNA for pTCP5:GUS construction, and it was inserted into the pGPTV-BAR [76] just before the β-glucuronidase (uidA) reporter gene. PCR, restriction enzyme digestion, and sequencing validated the plasmids. These plasmids were then transformed into wild-type plants by floral dipping using Agrobacterium tumefaciens strain GV3101 [77]. Basta spraying or screening red seeds with a fluorescence stereomicroscope helped in the selection of transgenic plants (Leica M205C; Leica Microsystems, Mannheim, Germany). Table S1 includes a list of the primers used.

4.3. Histology, Histochemistry, and Microscopy

Unopened anthers were dissected from flowers just before anthesis and submerged in drops of Alexander staining solution [78] on a slide for 30-60 min at room temperature. The slide was then covered with a coverslip and examined under a bright field microscope (Leica DFC450; Leica Microsystems, Mannheim, Germany).

Phloroglucinol staining was used to visualize lignin deposition in anthers, as previously described [42]. We removed the sepals and/or petals from open or unopened flowers, dyed the rest of the flowers for 5-10 min, and examined them using a stereomicroscope with bright field (Leica M205C; Leica Microsystems, Mannheim, Germany).

The inflorescences or open flowers of 4- or 5-week-old plants were fixed in 4% glutaraldehyde for at least 24 h at 4°C to create cross-sections of the anthers, which were then embedded in paraffin following dehydration. A microtome (RM2235, Leica Biosystems Nussloch GmbH) was used to cut 5-µm-thick sections, which were then deparaffinized, hydrated, and stained with 0.05% (w/v) toluidine blue for 1-3 min at room temperature. The stained slides were rinsed three times with water, dried at room temperature, and sealed with neutral balsam.

The anthers from stages 12-13 were dissected and cleared for 2-4 h in a chloral hydrate:glycerol:water (8:1:2/w:v:v) clearing solution to see the distribution of fibrous bands on the endothecium of the intact anthers. The anthers were then observed under a microscope (Leica DFC450) with differential interference contrast optics.

Individual unopened anthers were stained in the Alexander staining solution before being smashed and photographed to measure pollen output. We used Image J to count the pollen.

4.4. In Situ Hybridization

RNA in situ hybridization was performed using the previously reported method [79]. The full-length CDS of TCP5 was amplified by PCR using forward and reverse gene-specific primers containing a T3 sequence (aattaaccctcactaaaggg) and a T7 sequence (gtaatacgactcactatagggc), respectively. The PCR product was gel purified and quantified with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). T7 RNA polymerase (Roche Diagnostics GmbH, Mannheim, Germany) was used to synthesize the TCP5 antisense probe per the manufacturer’s instructions. The synthesized TCP5 probe was hydrolyzed before use.

4.5. RNA Extraction and qRT-PCR Analysis

Total RNA from the inflorescences was extracted using RNAiso Plus reagent (Takara, Japan) and then processed with DNase I (Takara Bio Inc., Beijing) to remove DNA contamination. RNA quality and concentration were determined using a NanoDrop 2000 spectrophotometer. Following the manufacturer's instructions, 1 µg of total RNA was used to synthesize first-strand cDNA using an RNA HiScript® III RT SuperMix for qPCR (+ gDNA wiper) Kit (Vazyme, Nanjing, China). We used Roche-LightCycler-480 II Real-Time System with ChamQTM Universal SYBR qPCR Master Mix (Promega, Beijing, China) to conduct qRT-PCR. The qRT-PCR reaction volume was 20 µL, with 10 µL of 2× SYBR qPCR Master Mix, 0.4 µL of primers (10 µM), 8 µL of cDNA template (50 ng/µL), and 1.2 µL of ddH2O. The reaction was performed at 94°C for 2 min, followed by 40 cycles of 94℃ for 15 s, and 60℃ for 31 s. 18S rRNA served as an internal control. The experiments included three technical and three biological replicates. Gene expression levels were determined using the 2^−∆∆CT method. Table S1 lists the gene-specific primers used for qRT-PCR.

4.6. Data Analysis

The genes’ relative expression levels were estimated using a one-way analysis of variance (ANOVA) and post-hoc Tukey’s test. The rest of the data were evaluated using the student t-test. All data are presented as mean ± SD. Asterisks indicate significant differences between the genes or groups (**, p≤ 0.01; ***, p≤ 0.001; ****, p≤ 0.0001).

5. Conclusions

As plant-specific transcriptional regulators, TCP family members regulate multiple developmental processes during a plant’s life cycle. TCP5 belongs to the CIN clade but is not targeted by miR319. Our study discovered that TCP5 expresses in early anthers. Overexpression of TCP5 or its dominant repressor form TCP5-SRDX resulted in undersized anthers with fewer pollen sacs and pollen, as well as a non-dehiscent anther phenotype in TCP5-overexpressing plants. Consistent with their phenotype, several early anther genes and genes associated with secondary cell wall thickening changed their expression levels in these transgenic plants. Our findings demonstrated that adequate TCP5 expression is essential for early anther development, as miR319-targeted CIN genes have previously been shown to regulate early anther development [43]. Thus, all CIN genes in Arabidopsis might play a role in early anther development. The relationship between these TCPs and other early anther genes, such as SPL8 and miR156-targeted SPLs, should be explored further.