OsCCRL1 is Essential for Phenylpropanoid Metabolism in Rice Anthers

Abstract

Phenylpropanoid metabolism and timely tapetal degradation are essential for anther and pollen development, but the underlying mechanisms are unclear. In the current study, to investigate this, we identified and analyzed the male-sterile mutant, osccrl1 (cinnamoyl coA reductase-like 1), which exhibited delayed tapetal programmed cell death (PCD) and defective mature pollen. Map-based cloning, genetic complementation, and gene knockout revealed that OsCCRL1 corresponds to the gene LOC_Os09g32020.2, a member of SDR (short-chain dehydrogenase/reductase) family enzyme. OsCCRL1 was preferentially expressed in the tapetal cells and microspores, and localized to the nucleus and cytoplasm in both rice protoplasts and Nicotiana benthamiana leaves. The osccrl1 mutant exhibited reduced CCRs enzyme activity, less lignin accumulation, delayed tapetum degradation, and disrupted phenylpropanoid metabolism. Furthermore, an R2R3 MYB transcription factor OsMYB103/OsMYB80/OsMS188/BM1, involved in tapetum and pollen development, regulates the expression of OsCCRL1. Finally, the osmyb103 osccrl1 double mutants, exhibited the same phenotype as the osmyb103 single mutant, further indicating that OsMYB103/OsMYB80/OsMS188/BM1 functions upstream of OsCCRL1. These findings help to clarify the role of phenylpropanoid metabolism in male sterility and the regulatory network underlying the tapetum degradation.

Background

The phenylpropanoid metabolism produces many important secondary metabolites, such as lignin, flavonoids, lignans, phenylpropanoid esters, hydroxycinnamic acid amides, coumarins, and sporopollenin, which are vital for plant development and survival (Knaggs 2003; Dong and Lin 2021). This pathway begins with three general enzymes: phenylalanine ammonialyase (PAL), cinnamic acid 4-hydroxylase (C4H), and 4-coumarate-CoA ligase (4CL) (Fraser and Chapple 2011). Subsequently, it splits into different branches to synthesize different products. Increasing numbers of studies have shown that phenylpropanoid metabolism is closely related to fertility in plants.

Sporopollenin, the main constituent of pollen exine, is a highly resistant biopolymer composed of many substances, such as aromatics, phenolics, long-chain aliphatic acids, and some phenylpropane metabolites, like ferulate, p-hydroxybenzoate, p-coumarate and lignin units (Domínguez et al. 1999; Bubert et al. 2002; Ahlers et al. 2003), and interfering with the phenylpropanoid pathway could affect the biosynthesis of sporopollenin (Xue et al. 2020). Besides, anther cuticle is also important to protect pollen from external environment, and it is mainly composed of cutin and wax. The generation of cutin and wax also requires the phenylpropane metabolites such as triterpenoids, flavonoids, phenylpropanoids esters. Therefore, when the anther cuticle was destroyed, a whitish anther epidermis often could result (Jetter et al. 2006).

Lignin, an important end-product in phenylpropanoid pathway, is also indispensable for anther development. The lignification and thickening of anther endothecium layer mediate anther dehiscence, thereby affecting fertility (Zhang and Wilson 2009; Gui et al. 2011). Suppressing Os4CL3 (4-Coumarate: coenzyme A Ligase 3) expression causes a wrinkled exterior appearing on the anther surface and the failure of the endothecium layer lignification, resulting in anther dehiscences and ultimately infertility (Schilmiller et al. 2009; Gui et al. 2011). Cinnamoyl CoA reductase (CCR) and cinnamyl alcohol dehydrogenase (CAD), two important enzymes in the biosynthesis of lignin units, also play important roles in plant fertility (Baucher et al. 1995; Lacombe et al. 1997; Barros et al. 2019). In Arabidopsis, the simultaneous repression of CCR and CAD severely damages the lignin structure and reduces the lignin content. The Arabidopsis cad c cad d ccr1 triple mutant displays a severe dwarf phenotype and male sterility (Thévenin et al. 2011). Among them, CCR is the first committed enzyme in the monolignol pathway for lignin biosynthesis and catalyzes the conversion of hydroxycinnamoyl-CoAs to hydroxycinnamaldehydes. And it is accepted that CCRs involved in lignification using feruloyl-CoA as substrates, mainly in grasses (Ma et al. 2007; Escamilla-Treviño et al. 2010). Besides, CCRs are considered members of SDR (short-chain dehydrogenase/reductase) superfamily, whose N-terminal region binds the coenzymes NAD(P)(H) and the C-terminal region constitutes the substrate binding part. SDR family is a large family that contains more than 3000 members, with a substrate spectrum ranging from alcohols, sugars, steroids and aromatic compounds to xenobiotics (Kallberg et al. 2002).

Flavonoid, another major phenylpropanoid metabolites, also takes part in male fertility. Except for sporopollenin, the pollen exine is also covered by flavonoid glycosides, and the biosynthesis of pollen intine also requires the flavonoid metabolites. Therefore, the flavonoid metabolites are necessary for the integrity of pollen exine and intine (Grunewald et al. 2020; Battat et al. 2019). Chalcone synthase (CHS) catalyzes the initiation of flavonoid biosynthesis and the rice oschs1 mutant produces flavonoid-depleted pollen and shows a loss of male fertility (Wang et al. 2020).

In general, when the phenylpropanoid metabolic pathway in anthers is affected, processes such as the development of the anther epidermis, sporopollenin, and the pollen wall, as well as anther dehiscence, are abnormal. Therefore, phenylpropanoid metabolism is vital for anther and pollen development.

In this study, we isolated the male-sterilemutant osccrl1 from indica rice and determined that it has abnormal anthers and defective mature pollen. OsCCRL1 was localized to the nucleus and cytoplasm and expressed in the tapetum and microspores. The osccrl1 mutant exhibited reduced CCRs enzyme activity, less lignin accumulation, disturbed phenylpropanoid metabolism, and delayed tapetum degradation. We also determined that the transcription factor OsMYB103/OsMYB80/OsMS188/BM1 regulates OsCCRL1 expression. At the genetic level, the osmyb103 osccrl1 double mutants showed the same phenotype as osmyb103 single mutant, indicating that OsCCRL1 functions the down stream of OsMYB103/OsMYB80/OsMS188/BM1. These findings provide evidence for the importance of phenylpropanoid metabolism in rice male sterility and enrich the network underlying the tapetal PCD.

Materials and Methods

Plant Material

The rice (Oryza sativa L. ssp. indica) osccrl1 mutant was derived from ‘Xinong 1B’ population treated by 1% (v/v) ethyl methanesulfonate (EMS) and inherited stably after more than three years. In phenotypic characterization and cytological analyse experiments, ‘Xinong 1B’ was used as the wild-type (WT) control. Male-sterile plants in the F₂ population generated from the cross between ‘Jinhui 10’ and the osccrl1 mutant were selected to OsCCRL1. All plants including transgenic plants were bred in Southwest University, Chongqing, China, under natural conditions.

Microscopy

The phenotypes of the osccrl1 mutant and the wild type were compared over the entire growth period under natural conditions. During the flowering period, wild type and osccrl1 mutant spikelets and anthers were randomly selected and observed using a SMZ1500 stereomicroscope (Nikon, Tokyo, Japan). For scanning electron microscopy (SEM) observations, the anther epidermis and inner surface were observed using a SU3500 scanning electron microscope (Hitachi, Tokyo, Japan) under an accelerating voltage of 5.0 kV at − 20 °C. To analyze pollen fertility, mature pollen grains were stained with 1% (w/v) I₂/KI solution and photographed using an Eclipse E600 microscope (Nikon). To observe anther development in the wild type and osccrl1 mutant, anthers at different stages were collected and embedded for semi-section as described previously (Li et al. 2006). Briefly, anthers were pre-fixed in 2.5% (v/v) glutaraldehyde and 2% (w/v) paraformaldehyde at 4 °C, and then post-fixed with 1% (w/v) osmium tetroxide at room temperature. Subsequently, samples were dehydrated with a graded ethanol series [50% (v/v), 70% (v/v), 85% (v/v), 95% (v/v), 100% (v/v), 100% (v/v), and 100% (v/v)] and embedded in SPI-PON 812 resin (Shanghai Physion Instruments, Shanghai, China). 2 μm sections were prepared using an EM-UC6 microtome (Leica, Wetzal, Germany), and then stained with 0.5% (w/v) toluidine blue. For TUNEL analyses of the wild type and osccrl1 mutant anther, anthers at different stages were collected and embedded in paraffin (Sigma-Aldrich, St Louis, MO, USA) in accordance with a previously described method (Fu et al. 2014). In situ analyses of nick-end labeled nuclear DNA fragments in wild-type and osccrl1 anthers was performed with the DeadEnd™ Fluorometric TUNEL Kit (Promega, Madison, WI, USA) in accordance with the manufacturer’s recommendations. TUNEL signals were analyzed using a LSM 800 confocal laser scanning microscopy (Zeiss, Jena, Germany). The pretreatment and embedding of wild-type and osccrl1 anthers for transmission electron microscopy (TEM) analyse were similar to thatin semi-thin section. Using an EM-UC6 microtome, 60 nm ultrathin sections were cut, and then stained with uranyl acetate and citrate aqueous. After staining, the sections were observed and photographed using H-7500 TEM microscope (Hitachi).

Genetic Analysis and Fine-Mapping of OsCCRL1

The osccrl1 mutant was crossed with ‘Jinhui 10’ to generate the F₁ population and the F₂ population was derived by the self-fertilization of F₁ individuals. The F₂ individuals that exhibited male sterility were selected to map OsCCRL1 gene. Gene mapping was conducted using simple sequence repeat markers obtained from the publicly available rice databases Gramene (http://www.gramene.org) and the Rice Genomic Research Program (http://rgp.dna.affrc.go.jp/E/publicdata/caps/index.html). Insertionand deletion markers were developed from comparison of genomic sequences for ‘Xinong 1B’ and ‘Jinhui 10’.

Complementary Verification of OsCCRL1

To complement the male-sterile phenotype of the osccrl1 mutant, a 6383 bp genomic fragment from the wild-type genomic DNA was amplified and cloned into the binary vector pCAMBIA1301. The complementary vector was introduced into the osccrl1 mutant using the Agrobacterium tumefaciens-mediated method (Ma et al. 2017). The primer sequences used for vector construction are listed in Additional file 1: Table S1.

CRISPR-CAS9 Mediated Gene Editing

In regularly interspaced short palindromic repeats (CRISPR)-CAS9 (CRISPR-associated protein 9) mediated gene editing, the specific and low mop rate target sequences were designed according ‘CRISPR direct’ and ‘RNA-fold’ web server. The designed knockout site was constructed into the knockout vector pCAR-CAS9 according to the system (Ma and Liu 2016). The vectors were introduced into the corresponding plants using Agrobacterium tumefaciens-mediated method. The primer sequences used for vector construction are listed in Additional file 1: Table S1.

Reverse Transcription Real-Time Quantitative PCR and In Situ Hybridization

Total RNAs from isolated tissue were extracted and purified using the RNAprep Pure Plant Kit (Tiangen, Beijing, China). 1 µg RNA was used to synthesis cDNA with the SuperScript® III Reverse Transcriptase Kit (Invitrogen, Shanghai, China). After reverse transcription, all samples were diluted with sterilized water. The reverse transcription real-time quantitative PCR (RT-qPCR) analysis was performed on the ABI 7500 Sequence Detection System (Thermo Fisher Scientific, Waltham, MA, USA) using the SYBR® Premix Ex Taq™ GC Kit (Takara, Dalian, China). Using OsACTIN1 as an endogenous control, three replicates were performed and the normalization of relative expression was calculated using the ∆∆C_t method (Livak et al. 2001). For mRNA in situ hybridization of OsCCRL1, the 331 bp specific probe of OsCCRL1 was amplified and labeled using the DIG RNA Labeling Kit (Roche, Basel, Switzerland). Pretreatment of sections, hybridization, and immunological detection were performed as described previously (Ma et al. 2017). The primer sequences used for RT-qPCR and in situ hybridization are listed in Additional file 1: Table S1.

Subcellular Localization

To generate fusion proteins with GFP (green fluorescent protein), the full coding region of OsCCRL1 and LOC_Os09g32020.1 without stop codon was amplified and cloned into pAN580 (2 × 35S-GFP-Nos) vector. The generated plasmids, 2 × 35S::OsCCRL1-GFP and 2 × 35S::LOC_Os09g32020.1-GFP, were introduced into rice protoplasts following a previously described method (Ma et al. 2017). The nuclear-localized marker OsH2B-mCherry and endoplasmic reticulum (ER)-localized marker OsHDEL-mCherry were used. After incubation at 28 °C overnight, fluorescence signals were observed using a LSM 800 confocal laser scanning microscope (Zeiss). For the subcellular localization of Arabidopsis protein, the full-length coding sequence of AtTKPR1 and AtTKPR2 was also cloned into the pAN580 vector and introduced into Arabidopsis protoplasts as a previously described method (Wang et al. 2015). For the subcellular localization in Nicotiana benthamiana, the full-length coding sequence of OsCCRL1, AtTKPR1 and AtTKPR2 was cloned into the pCAMBIA1300 (35S-GFP-Nos) vector driven by one CaMV 35S promoter. Constructs were introduced into Nicotiana benthamiana leaf cells by Agrobacterium-mediated infiltration. The nuclear was indicated by 4',6-diamidino-2-phenylindole (DAPI). The primer sequences used for vector construction are listed in Additional file 1: Table S1.

Western-Blot Analysis of Extracted Nuclear and Cytoplasmic Protein

The cytoplasmic and nuclear proteins from tobacco leaf cells expressing 35S::OsCCRL1-GFP was extracted with previous methods and the western-blot detection was performed (Pandey et al. 2006; Xu et al. 2006; Chen et al. 2014). Histone antibodies was used as the internal reference antibody for detecting nuclear protein. Tubulin antibodies was used as the internal reference antibody for detecting cytoplasmic protein.

Phylogenetic Analysis

To identify the putative homologs of OsCCRL1, its protein sequence was used as a BLAST query on the ‘Gramene’ and ‘NCBI’ database. Phylogenetic tree construction was conducted by Maximum Likelihood method. The protein sequences used for phylogenetic tree are listed in Additional file 1: Table S2.

Total CCRs Enzyme Activity, Lignin Staining and Content of Anther

Under the protection of liquid nitrogen, total CCRs enzyme solutions were extracted from equal amounts anthers in WT and osccrl1 mutant according to the previous method, and were measured the absorbance of A366 by adding the feruloyl-coA as the substrate (Goffner et al. 1994). Briefly, The reaction mixture consisted of 0.1 mM NADPH, 70 μM feruloyl-coA, and different volumes of extracted CCRs from WT and osccrl1 mutant in 100 mM sodium/potassium phosphate buffer (pH 6.25) in a total volume of 500 μL. For lignin staining, the anthers of stage 12 were dissected and stained with phloroglucinol-HCl (Hao et al. 2014), then photographed using an Eclipse E600 microscope (Nikon). To quantify the lignin content, the extraction and quantification from wild-type and osccrl1 mutant spikelet using a previous method with three biological replicates (Moreira-Vilar et al. 2014; Borah and Khurana 2018). Briefly, 0.2 mg sample was rinsed in H-buffer (50 mM Tris–HCL, 10 g/L Triton, 1 M NaCl, PH 8.3) and acetone several times. Then grind them into powder under the protection of liquid nitrogen after drying. All samples were placed into a screw-cap centrifuge tube containing the reaction mixture (1.2 mL of thioglycolic acid and 6 mL of 2 M HCl) and heated (98 °C, 4 h). A standard curve for lignin (alkali, Aldrich) by measuring the absorbance of A280 was generated. The lignin content in wild-type and osccrl1 mutant spikelet was calculated by referring to the standard curve.

The Enzyme Activity of OsCCRL1 In Vitro

The full-length coding region of OsCCRL1 and osccrl1 was amplified and recombined into the pGEX-4 T-1 vector, and then transformed into Escherichia coli BL21 (DE3) to obtain the OsCCRL1-GST and osccrl1-GST fusion protein. The CCR enzyme activity of OsCCRL1 and osccrl1 was measured according to the method (Lüderitz and Grisebach 1981). The reaction mixture consisted of 0.1 mM NADPH, 100 μM feruloyl-coA, and 5 μg of purified recombinant protein in 100 mM sodium/potassium phosphate buffer (pH 6.25) in a total volume of 500 μL. The enzyme reactions were carried out at 30 °C and the absorbance of A366 was monitored. The primer sequences used for vector construction are listed in Additional file 1: Table S1.

Transient Expression Assay

The promoter of OsCCRL1 (1576 bp) were amplified and cloned into the pGreenII0800-LUC double-reporter vector (firefly luciferase and renilla luciferase). The full-length coding region of OsMYB103 was amplified and cloned to pAN580-no-GFP (2 × 35S-Nos) vector. The constructs were then transformed into rice protoplasts. After overnight incubation at 28 °C, LUC and Renilla (REN) luciferase activities were measured using the Dual Luciferase Assay Kit (Promega, USA), and analysed using the Luminoskan Ascent Microplate Luminometer (Thermo Fisher Scientific, USA) according to the manufacturer’s instructions. The pAN580-no-GFP was used as an control. At least six transient measurements were conducted for each assay. The primer sequences used for vector construction are listed in Additional file 1: Table S1.

Chromatin Immunoprecipitation Analysis

The ChIP assays were performed in Ubi::OsMYB103-GFP transgenic anthers using the EpiQuik™ Plant ChIP kit (EpiGentek, Farmingdale, NY, USA) in accordance with the manufacturer’s recommendations. Briefly, fresh tissues were crosslinked with 1% (v/v) formaldehyde solution under vacuum and decrosslinked by adding 2 M glycine solution (final concentration 0.125 M). The protein–DNA complexes were sheared into 200–500 bp fragments by sonication from the protease inhibitor cocktail protection. After sonication, a portion of chromatin was reverse-crosslinked and used as the input DNA control for ChIP-qPCR analysis. A dilution of the resulting chromatin was incubated with anti-GFP antibody ab290 (Abcam, Cambridge, UK) or anti-IgG antibody at room temperature with orbital shaking. The amount of immunoprecipitated genomic DNA was performed on an ABI 7500 Sequence Detection System (Thermo Fisher) with three biological repetitions. The calculation of relative enrichment followed a previously described method (Li et al. 2011). Relative enrichment was measured by comparing the input and ChIP groups. Normal anti-IgG antibody was used as a negative control. The primer sequences used for vector construction and ChIP-qPCR analysis are listed in Additional file 1: Table S1.

Electrophoretic Mobility Shift Assay

The full coding sequence of OsMYB103 was cloned into the pET-28a vector (Novagen) and then transformed into Escherichia coli strain BL21(DE3) to obtain the OsMYB103-HIS fusion protein. The electrophoretic mobility shift assay (EMSA) was performed using the LightShift™ Chemiluminescent EMSA Kit (BersinBio, Guangzhou, China) in accordance with the manufacturer’s instructions. The enriched regulatory sequences ‘P2’ and ‘P7’ in ChIP-qPCR analysis were synthesized to shorter sequences and annealed to biotin-labeled double-stranded DNA probe (Novagen), named ‘P2s’ and ‘P7s’, respectively. Binding reactions were then separated in 6% acrylamide negative gel and transferred to nylon membranes. Through blocking and washing, membranes were placed in a film cassette and exposed to X-ray film for 5 min. The primer sequences used for vector construction and DNA probes are listed in Additional file 1: Table S1.

The Callose Assays

For callose assays, anthers at stage 8b were collected, dehydrated, and embedded in paraffin (Sigma, St Louis, MO, USA). 10 μm thickness paraffin sections were cut and stained with 0.25% (w/v) aniline blue (Sigma-Aldrich) following a previous method (Wan et al. 2011). The stained sections were observed with a Nikon-80 fluorescence microscope (Nikon) under ultraviolet light condition.

Sudan Red 7B Staining

For lipidic staining of anther, anthers at the mature pollen stage were soaked in a solution containing 0.1% (w/v) sudan red 7B, 50% (v/v) polyethylene glycol-400, 45% (v/v) glycerol. The processing and staining were performed according to the method reported previously (Xu et al. 2017).

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBL libraries under the following accession numbers: OsMYB103 (LOC_Os04g39470), OsCCRL1 (LOC_Os09g32020.2), AtTKPR1 (AT4G35420), AtTKPR2 (AT1G68540), Os4CL3 (LOC_Os02g08100), OsCHS1 (LOC_Os11g32650), OsAP25 (LOC_Os03g08790), OsAP37 (LOC_Os04g37570), C6 (LOC_Os11g37280), OsPKS2 (LOC_Os07g22850).

Results

Isolation and Phenotypic Characterization of the osccrl1 Mutant

To identify additional genes that function in anther and pollen development, we screened a rice mutant library generated by ethyl methanesulfonate (EMS) mutagenesis of Oryza sativa L. ssp. indica ‘Xinong 1B’. We selected one mutant with complete male sterility, osccrl1 (cinnamoyl coA reductase-like 1), for further analysis. At the vegetative stages, there were no obvious phenotypic differences between the wild type and osccrl1 mutant (data not shown). At the flowering stage, both wild-type and osccrl1 spikelets developed normally, and their floral organs were visible (Fig. 1A–D). However, compared to the wild type, osccrl1 mutant anthers dehisce abnormally, without any cracking of the apexes or bases (Fig. 1C, D). In addition, pollen iodine potassium iodide (I₂/KI) staining revealed no viable pollen in osccrl1 anther, which showed osccrl1 was completely male-sterile (Fig. 1E, F).

Fig. 1

Phenotypic comparison between the WT and the osccrl1 mutant. A, B Spikelet of the wild type and osccrl1 mutant. Scale bars, 5 mm. C, D Spikelet of the wild type and osccrl1 mutant after removal of the palea and lemma. Scale bars, 5 mm. E, F Pollen grains of the wild type and osccrl1 mutant stained by I₂/KI solution. Scale bars, 200 μm. G, H The scanning electron microscopy of anther in wild type and osccrl1 mutant. Scale bars, 200 μm. I, J The scanning electron microscopy of anther epidermis surface in wild type and osccrl1 mutant. Scale bars, 50 μm. K, L The scanning electron microscopy of anther inner surface in wild type and osccrl1 mutant. Scale bars, 50 μm

We performed scanning electron microscopy of the anthers epidermis and inner surface to further investigate the morphological differences between the wild type and osccrl1 mutant. The wild-type anthers clearly dehisced, whereas osccrl1 mutant anthers did not dehisce normally (Fig. 1G, H). In addition, a well-ordered nano-ridged structure formed by anther cuticle appeared on the wild-type anthers epidermis, whereas the osccrl1 anther epidermis at the same stage appeared smoother, with reduced deposition of the cutin and wax polymers (Fig. 1I, J). On the anther inner surface, many dense Ubisch bodies with sharp protrusions were uniformly distributed in the wild type, whereas almost no Ubisch bodies were observed in the osccrl1 mutant (Fig. 1K, L).

Semi-thin transverse sections were performed to further examine the anther morphological defects in wild type and osccrl1 mutant. At the stages of 8a and 8b, both wild-type and osccrl1 mutant anthers contained four layers: the epidermis, endothecium, middle layer, and tapetum. At the stage 8a, the tapetal cells appeared vacuolated, and the meiocytes formed ellipsoidal shaped dyads. At stage 8b, the tapetal cell continued the process of vacuolation, and the dyads had undergone meiosis to form a haploid tetrad (Fig. 2A–D). At stage 9, no obvious morphological difference was visible between wild-type and osccrl1 mutant anthers, as they all contained three cellular layers and free globular microspores (Fig. 2E, F). The morphological difference between osccrl1 mutant and wild-type anthers began to appear at stage 10. At this stage, the tapetum appeared thicker in osccrl1 mutant than that in the wild type, and the osccrl1 microspores showed smaller spherical and uneven distribution in anther chamber comparing with that in wild type (Fig. 2G, H). At stage 11, wild-type microspore had undergone mitotic division and appeared the falcate shape, and the tapetum had undergone normal PCD and become denser. In the osccrl1 mutant, the microspores also appeared to have undergone mitosis and become sickle-shaped, but the tapetal cells showed a delay in PCD, as they contained a thicker tapetum layer than that in the wild type (Fig. 2I, J). At stage 12, wild-type microspores were darkly stained and were filled with starch. By contrast, osccrl1 microspores were abnormally shrunken, showing an infertile phenotype (Fig. 2K, L). And at stage 13, the wild-type anther normally dehisced, whereas osccrl1 mutant anthers did not dehisce and had none pollen (Fig. 2M, N).

Fig. 2

Semi-thin section and TUNEL assay of the wild type and osccrl1 mutant. A − N Semi-thin sections of anthers at stages 8a to 13 in the wild type and osccrl1 mutant. The images show a single locule of an anther. E, epidermis; En, endothecium; M, middle layer; T, tapetum; Dy, dyad cell; Tds, tetrads; Msp, microspore; BP, bicellular pollen; MP, mature pollen; dMP, mature pollen. Scale bars, 25 μm. O − X TUNEL analyses of wild-type and osccrl1 anthers at stages 8a to 11. Propidium iodide staining is indicated by red fluorescence and TUNEL positive staining is indicated by yellow to green fluorescence (arrow indication). Scale bars, 100 μm

The PCD of tapetal cells is characterized by the random cleavage of nuclear DNA, which can be detected through a terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling (TUNEL) assay. We therefore performed a TUNEL assay of wild-type and osccrl1 anthers to test whether the PCD of tapetal cells was defective in the osccrl1 mutant. At stage 8a, 8b, and 9, a positive TUNEL signal was observed in both wild-type and osccrl1 tapetal cells (Fig. 2O–T), and the TUNEL fluorescence signal reached the maximum at the stage 8b and 9. At stage 10, the DNA fragment signal become very weak in wild type but the osccrl1 tapetum still has some stronger PCD signal than that in wild type (Fig. 2U, V). At stage 11, no any DNA fragment signal was observed in wild-type tapetal cells, but an obvious TUNEL signal was detected in osccrl1 tapetal cells (Fig. 2W, X). These results confirm that the tapetum degradation is delayed in the osccrl1 mutant.

Map-Based Cloning and Functional Verification of OsCCRL1

To determine whether the infertile phenotype of osccrl1 mutant is controlled by a single gene, we crossed osccrl1 mutants with the maintainer line ‘Jinhui 10’. All F₁ individuals exhibited a normal phenotype, and the F₂ generation showed a segregation ratio of approximately 3:1 (normal:sterile = 387:128, χ2 = 0.0026 < χ2_0.05 = 3.841), indicating that a single recessive nuclear gene controls this male-sterile phenotype. The OsCCRL1 locus was mapped by map-based cloning to a 75-kb region flanked by markers LZY17 and LZY23 on chromosome 9. Sequencing analysis identified a single-nucleotide substitution from ‘T’ to ‘A’ in LOC_Os09g32020, which was described as OsTKPR1 (Tetraketide α-Pyrone Reductase 1) gene before (Xu et al. 2019). The LOC_Os09g32020 gene generates two transcripts whose coding sequences do not overlap: LOC_Os09g32020.1 with one exon (993 bp) and LOC_Os09g32020.2 with 6 exons (1101 bp) (Additional file 1: Fig. S1A, B). The ‘T’ to ‘A’ substitution occurred in the fourth exon of LOC_Os09g32020.2 in the osccrl1 mutant, causing an amino acid conversion from valine (V; a hydrophobic amino acid) to glutamic acid (E; a negatively charged amino acid) at amino acid 231 (Fig. 3A). What needs to be mentioned was that there was one base difference, ‘T’ to ‘G’, in the LOC_Os09g32020.2 genome between the reference sequence given online and the sequence we sequenced from wild-type DNA (Additional file 1: Fig. S1C), which caused an amino acid change (D to E, aspartic acid to glutamic acid) occurring at the site 80 of OsCCRL1 protein (data not shown).

Fig. 3

Molecular cloning, complementary validation of OsCCRL1, and CRISPR-Cas9-mediated targeted mutagenesis of OsCCRL1. A Map-based cloning of OsCCRL1 on chromosome 9. The thick black line represents the chromosome, the markers used for gene mapping are indicated above the black line, numbers in brackets below the black line represent recombinants, and the value below the line represents the genetic distance (cM) between two markers. The osccrl1 mutant harbors a nucleotide ‘T’ to ‘A’ transition in exon 4, leading to a Val-to-Glu amino acid conversion at site 231. B Anther morphology and pollen grains stained by I₂/KI solution in the wild type, osccrl1 mutant and OsCCRL1-complementary transgenic plant. Scale bars, 2 mm (b-1, b-3 and b-5); 200 μm (b-2, b-4 and b-6). C Reference sequence and sequencing of OsCCRL1^cas9 mutant lines. D Spikelet, anther, anther dehiscence and pollen grains stained by I₂/KI solution in wild type and OsCCRL1^cas9 mutant lines. Scale bars, 2 mm (d-1 to d-6); 500 μm (d-7 to d-12); 400 μm (d-13 to d-18); 200 μm (d-19 to d-24)

To verify that the phenotype of the osccrl1 mutant is caused by the mutation of LOC_Os09g32020.2, we introduced a complementation vector containing the normal genomic sequence of the LOC_Os09g32020.2 into the osccrl1 mutant to generate the corresponding transformant OsCCRL1-C. Among the 35 transgenic plants generated, 7 complementation-positive plants were obtained. Compared to osccrl1 mutant and the wild type, the anthers and pollen were restored to the wild-type phenotype in the OsCCRL1-C transformants (Fig. 3B).

We then used CRISPR-CAS9 gene editing to knock out LOC_Os09g32020.2 in the wild-type cultivar ‘Xinong 1B’ background. Among the 21 transgenic plants generated, 16 knockout-positive plants were obtained. Sequencing revealed five types of mutations among these transgenic plants (Fig. 3C). We observed the 16 transformants throughout the growth period, finding that they were all male sterile. All CRISPR/CAS9 mutant lines grew normal spikelets. When removing the palea and lemma, the anthers of different mutant lines were a bit whiter and shorter than wild-type anthers, but the other floral organs were normal in mutant lines. During the later period of development, when anthers emerge from the glume, wild-type anther filaments grew longer and the anthers protruded from the upper part of the glume. However, the anther filaments of the mutant lines did not seem to grow, and the anthers emerged directly from the dehiscent glumes. Although there were some slight differences in the color of anther (deeper and lighter) and the length of filament (longer and shorter) between different CRISPR/CAS9 mutant lines, the I₂/KI staining revealed that all of those different mutant lines lacked mature pollen or accumulated little pollen. Therefore, they were infertile (Fig. 3D). These result, coupled with the effect of functional complementation, support the conclusion that OsCCRL1 is identical to LOC_Os09g32020.2.

Expression Analysis and Subcellular Localization of OsCCRL1

To examine the expression pattern of OsCCRL1, we quantified OsCCRL1 transcripts in wild-type mature tissues by reverse transcription quantitative PCR (RT-qPCR). The OsCCRL1 (LOC_Os09g32020.2) transcripts were abundant in spikelets, with small amounts in leaves, implied that OsCCRL1 mainly functions in spikelets (Fig. 4A). At the same time, we found LOC_Os09g32020.1 transcript was highly expressed in roots, sheaths, and spikelets, but not culms, indicating that LOC_Os09g32020.1 does not function specifically in spikelets (Additional file 1: Fig. S1D). When we performed RT-qPCR in anthers at different developmental stages, OsCCRL1 transcripts were abundant at stages 9 and 10 relative to the other periods (Fig. 4B) and LOC_Os09g32020.1 transcripts were also abundant at stages 9 and 10 (Additional file 1: Fig. S1E). For more detailed analysis of OsCCRL1 expression, we performed mRNA in situ hybridization of fixed wild-type anthers using digoxygenin (DIG)-labeled antisense and sense probes. Strong DIG signals were observed in tapetal cells and microspores (Fig. 4C). These results confirm that OsCCRL1 is closely associated with anther and pollen development.

Fig. 4

Expression analysis and subcellular localization of OsCCRL1. A, B Abundance of OsCCRL1 transcripts in wild-type different tissues and anthers at different developmental stages. Data are means ± SD (n = 3). C mRNA in situ hybridization of OsCCRL1 transcripts in the wild-type anther. Scale bars, 50 μm. D Subcellular localization of 2 × 35S::GFP and 2 × 35S::OsCCRL1-GFP in rice protoplasts. OsH2B-mCherry was used to indicate the nucleus. Scale bars, 5 μm. E Subcellular localization of 35S::GFP and 35S::OsCCRL1-GFP with DAPI in Nicotiana benthamiana. Scale bar: 5 μm. F Immunoblotting of extracted nuclear proteins from tobacco leaves expressing OsCCRL1-GFP using anti-GFP antibodies and Histone antibodies. Histone antibodies was used as the internal reference antibody. G Immunoblotting of extracted cytoplasm proteins from tobacco leaves expressing OsCCRL1-GFP using anti-GFP antibodies and Tubulin antibodies. Tubulin antibodies was used as the internal reference antibody.

We then transiently expressed GFP driven by two CaMV 35S promoters and 2 × 35S::OsCCRL1-GFP fusion protein in rice protoplasts, along with OsH2B-mCherry as a nuclear marker. Using laser confocal microscopy, we detected GFP from the empty vector throughout the protoplast, except the vacuole. By contrast, the GFP signal from 2 × 35S::OsCCRL1-GFP was located to the cytoplasm and nucleus, partially overlapping the signal with OsH2B-mCherry (Fig. 4D), which was different from the previous reporter that OsTKPR1 was localized in the endoplasmic reticulum (ER) (Xu et al. 2019). Based on this confusion, we transiently expressed the 35S::OsCCRL1-GFP fusion protein driven by single CaMV35S promoter in Nicotiana benthamiana leaves and 4',6-diamidino-2-phenylindole (DAPI) was performed to stain the nucleus. In N. benthamiana leaves, 35S::OsCCRL1-GFP localized to the cytoplasm and nucleus, as also observed in protoplasts (Fig. 4E). Finally, to confirm the localization of OsCCRL1, we extracted nuclear and cytoplasmic proteins from N. benthamiana leaves expressing 35S::OsCCRL1-GFP and subjected them to immunoblot analysis with anti-GFP antibody. And OsCCRL1 was identified in nuclear and cytoplasmic extracts from N. benthamiana leaves (Fig. 4F, G). These results indicate that OsCCRL1 is expressed in the nucleus and cytoplasm. At the same, we transiently expressed 2 × 35S::LOC_Os09g32020.1-GFP fusion protein in rice protoplasts. OsH2B-mCherry and OsHDEL-mCherry were used as nuclear and ER markers, respectively. The LOC_Os09g32020.1 protein localized to the ER and the nucleus (Additional file 1: Fig. S1F).

OsCCRL1 Affects the Lignin Accumulation and Tapetal PCD

LOC_Os09g32020.2 encodes a SDR family enzyme that uses NAD(P)(H) as a co-factor for catalysis. As the plant SDR family is a large family (Moummou et al. 2012), we constructed a phylogenetic tree mainly derived from rice. This phylogenetic analysis suggested that OsCCRL1 might belong to two enzyme classes, dihydroflavonol-4-reductase (DFR) and cinnamoyl coA reductase (CCR) (Additional file 1: Fig. S2A). DFR is a pivotal oxidoreductase that catalyzes the conversion of dihydroflavonols to generate leucoanthocyanindins during flavonoid biosynthesis (Halbwirth et al. 2006; Petit et al. 2007); CCR catalyzes the hydroxycinnamaldehydes (p-coumaraldehyde, coniferaldehyde, and sinapaldehyde) synthesis from hydroxycinnamoyl-CoA thioesters (p-coumaroyl-CoA, feruloyl-CoA, sinapoyl-CoA) during lignin biosynthesis (Lacombe et al. 1997; Pan et al. 2014). DFR and CCR may originate from a common ancestor, as indicated by sequence alignment. Plant CCR and DFR, bacterial uridine diphosphate-galactose-4-epimerase (UDP-galactose-4-epimerase) and mammalian 3β-hydroxysteroid dehydrogenase show significant similarities, suggesting that UDP-galactose-4-epimerase might be the common ancestor of CCR and DFR proteins (Lacombe et al. 1997).

The ostkpr1 mutants, concluding carrying a T-DNA insertion in the second exon of LOC_Os09g32020.2 and a 15-base deletion in the fourth exon of LOC_Os09g32020.2 induced by ⁶⁰Co γ-radiation, lack mature pollen grains (Wang et al. 2013; Xu et al. 2019). Therefore, LOC_Os09g32020.2 was considered to encode tetraketide α-pyrone reductase 1 (TKPR1), a rice ortholog of Arabidopsis TKPR1, which plays an important role in the fatty acids and sporopollenin synthesis with acyl-CoA synthetase (ACOS) and polyketide synthase (PKS). TKPR enzymes were initially considered to be DFR-like proteins (Yau et al. 2005; Tang et al. 2009). We therefore constructed a phylogenetic tree of OsCCRL1 based primarily on rice and two Arabidopsis TKPR proteins, AtTKPR1 and AtTKPR2. Based on this analysis, OsCCRL1, AtTKPR1, and AtTKPR2 are evolutionarily closer to DFR (Additional file 1: Fig. S2B). We then transiently expressed GFP, 35S::AtTKPR1-GFP, and 35S::AtTKPR2-GFP fusion proteins in N. benthamiana leaves and performed DAPI staining for the nucleus, and found that AtTKPR1 and AtTKPR2 localized to the cytoplasm and nucleus (Additional file 1: Fig. S3A). We transiently expressed 2 × 35S::AtTKPR1-GFP and 2 × 35S::AtTKPR2-GFP in Arabidopsis protoplasts. AtVirD2NLS-mCherry was used as a nuclear marker. The GFP signals from 2 × 35S::AtTKPR1-GFP and 2 × 35S::AtTKPR2-GFP were located in the cytoplasm and nucleus (Additional file 1: Fig. S3B).

In previous study, the Arabidopsis DFR (tt3) null mutation could not be complemented by 35S::AtDRL1/AtTKPR1 gene, and standard DFR substrates (dihydrokaempferol and dihydroquercetin) could not be metabolized by N. tabacum (Nt) TKPR1 or OsTKPR1 in vitro (Tang et al. 2009; Wang et al. 2013). For this, a conclusion that TKPR1 were unlikely to process the activity of DFR enzymes participating in flavonoid biosynthesis was decided (Petit et al. 2007). In our study, we quantified the transcript level of OsPKS2 (Polyketide Synthase 2), which is involved in fatty acid biosynthesis with OsTKPR1, in wild-type and osccrl1 anthers by RT-qPCR, and observed no significant differences of OsPKS2 expression between wild-type and osccrl1 mutant (Additional file 1: Fig. S4A). Given this and previous findings, we investigated whether LOC_Os09g32020.2 was invovled in lignin synthesis. First, we extracted total CCRs from wild-type and osccrl1 anthers and performed an assay using ferulyl-CoA in vitro according to a published procedure (Goffner et al. 1994). When we measured the UV absorption of the reaction mixture at 366 nm after incubation, the absorption of wild-type CCRs decreased more rapidly than that from the osccrl1 mutant (Fig. 5A), implying that the osccrl1 mutant has reduced CCRs activity. CCR is a limiting enzyme with a key role in lignin biosynthesis. We therefore examined lignin levels in wild-type and osccrl1 anthers upon phloroglucinol-HCl staining (Hao et al. 2014). Wild-type anthers had darker staining than that of osccrl1 mutant, indicative of higher lignin contents than osccrl1 anthers (Fig. 5B). The lignin of wild-type and osccrl1 anthers was extracted by thioglycolic acid-HCL and dissolved in NaOH, and it was found that the lignin content of wild-type anthers was relatively higher than that of osccrl1 mutant (Fig. 5C). Next, the enzyme activity of recombinant protein OsCCRL1 and osccrl1 was assayed in vitro using with ferulyl-CoA substrate. The UV absorption of the reaction mixture at 366 nm after incubation of OsCCRL1 decreased more rapidly than that of osccrl1 (Fig. 5D). These indicated that OsCCRL1 participated in the synthesis of lignin.

Fig. 5

OsCCRL1 process the CCR enzyme activity and is involved in phenylpropane metabolism and tapetum apoptosis. A The comparison of extracted CCRs enzyme activity in the wild type and osccrl1 mutant anthers. Data are means ± SD (n = 3). B The lignin staining of wild-type and osccrl1 anthers by phloroglucinol-HCl. Bars, 20 μm (left) and 80 μm (right). C The lignin content of wild-type and osccrl1 spikelets extracted by thioglycolic acid-HCl. Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). D In vitro enzyme activity of OsCCRL1 and osccrl1 proteins. Data are means ± SD (n = 3). E − G The relative expression comparison of Os4CL3, OsCHS1 and OsCCRL1 in the wild type and osccrl1 anthers. Data are means ± SD (n = 3). Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). H − J The relative expression comparison of OsAP25, OsAP37 and C6 in the wild-type and osccrl1 anthers. Data are means ± SD (n = 3). Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). K − P Transmission electron micrograph of pollen wall and anther cuticle in the wild type and osccrl1 mutant. Po, pollen; Te, tectum; Ba, bacula; Ne, nexine; In, intine; dPo, defective pollen; dIn, defective intine; dEx, defective exine; E, epidermis; C, cuticle; dC, defective cuticle. Scale bars, 1 μm in (k, m), 0.2 μm in (l, n), and 2 μm in (o, p)

Based on this finding, we named the LOC_Os09g32020.2 gene as cinnamoyl coA reductase-like 1 (OsCCRL1). Because CCR also plays an important role in the shunting of components in the phenylpropanoid pathway, we examined the expression of Os4CL3 (4-Coumarate: coenzyme A ligase 3), OsCHS1 (chalcone synthase 1), and OsCCRL1 in wild-type and osccrl1 anthers, which are regarded to associate with phenylpropanoid pathway. The levels of all these transcripts were reduced in osccrl1 mutant (Fig. 5E–G), indicating that the mutation of OsCCRL1 also affects the phenylpropanoid metabolism. As normal tapetal PCD was also delayed in osccrl1 anther revealed by TUNEL assays above, we measured the transcript levels of some key genes involved in tapetal PCD, finding that OsAP25 (Aspartic Protease 25), OsAP37 (Aspartic Protease 37), and OsC6 transcript levels were also strongly reduced in osccrl1 mutant (Fig. 5H–J). Tapetum cells could secrete sporopollenin precursors for the formation of pollen exine, and the sporopollenin synthesis and pollen intine formation requires the participation of phenylpropanoid metabolites. In osccrl1 mutant, tapetum development was delayed and phenylpropane metabolism was disordered, which suggested that the pollen wall formation of osccrl1 mutants might be affected. Therefore, the wild-type and osccrl1 mutant pollen wall were observed by transmission electron microscopy (TEM), and it was found that the wild-type pollen had a complete pollen wall structure, including tectum, bacula, nexine and intine (Fig. 5K, L). However, the pollen wall structure of osccrl1 mutant was abnormal. In osccrl1 mutant, the bacula was missing, causing tectum and nexine fused together to form a thinner and defective exine. In addition, the pollen intine of osccrl1 mutant was also thinned (Fig. 5M, N). The synthesis of cutin and wax also requires the phenylpropane metabolites. Scanning electron microscopy found that the anthers epidermis of osccrl1 mutant was smoother and defective. When observing the pollen wall by TEM, we also found obvious differences between the surface of wild-type and osccrl1 mutant anther epidermis. A protruding structure, anther cuticle, composed of cutin and wax was found in the wild type, but the osccrl1 mutant lacked this (Fig. 5O, P).

OsCCRL1 is Directly Regulated by OsMYB103

Given that OsCCRL1 is expressed in a relatively spatiotemporally specific pattern, we then investigated the precise regulation of it. Various MYB transcription factors, the MBW ternary complex, and additional transcription factors are key transcriptional regulators of phenylpropanoid metabolism, especially lignin and flavonoid biosynthesis (Zhong et al. 2010; Feller et al. 2011; Hichri et al. 2011; Liu et al. 2015; Xu et al. 2015; Yang et al. 2017; Ma and Constabel 2019). In Arabidopsis, many MYB transcription factors function in the biosynthesis of secondary cell wall components, including cellulose, xylan, and lignin, such as MYB15, MYB46, MYB52, MYB54, MYB103, MYB85, MYB43, and MYB20 (Zhong et al. 2008; Chezem et al. 2017; An et al. 2019). In rice, OsMYB30 binds to and activates the OsPAL6 (Phenylalanine Ammonia-Lyase 6), OsPAL8 (Phenylalanine Ammonia-Lyase 8), Os4CL3 (4-Coumarate: coenzyme A Ligase 3), and Os4CL5 (4-Coumarate: coenzyme A Ligase 5) promoters to regulate phenylpropanoid metabolism (He et al. 2020; Li et al. 2020). OsMYB103 is an important R2R3 MYB transcription factor that positively regulates tapetum degradation (Zhang et al. 2010; Pan et al. 2020; Xiang et al. 2020; Han et al. 2021; Lei et al. 2022), and different type mutation of OsMYB103 in several different rice were male sterile. We previously identified the osmyb103 mutant (Additional file 1: Fig. S4B). In addition to delayed PCD in the tapetum, the osmyb103 mutant shows whitish anthers, abnormal sporopollenin synthesis, lack of Ubisch bodies, smoother anther epidermis, and defective pollen (Lei et al. 2022), suggesting that it may regulate phenylpropanoid metabolism. To investigate whether OsCCRL1 is regulated by OsMYB103, OsCCRL1 expression was detected by RT-qPCR in the wild type and osmyb103 anther at first [This osmyb103 mutant was isolated and identified in our previous study (Lei et al. 2022)]. The expression of OsCCRL1 almost none was detected in osmyb103 mutant (Fig. 6A), indicating that mutation of OsMYB103 affected the expression of OsCCRL1. We also examined the expression of LOC_Os09g32020.1 in the wild type and osmyb103 anther, finding no significant difference between them (Fig. 6B). These results, combined with the results of the subcellular localization analysis, suggest that the LOC_Os09g32020.1 and LOC_Os09g32020.2 mRNAs are two different transcripts.

Fig. 6

OsMYB103 regulates the expression of OsCCRL1. A, B The relative expression comparison of OsCCRL1 and LOC_Os09g32020.1 between the wild-type and osmyb103 mutant anthers. Data are means ± SD (n = 3). Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). C Transient expression assay of the LUC reporter gene with the multiple cloning site (MCS) effector plasmid using as a control. Data are means ± SD (n = 3). Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). D Schematic diagrams of the distribution of ‘AACC’ motifs (cis-elements recognized by OsMYB103 transcription factors) in the OsCCRL1 promoter regions. E A ChIP-qPCR analysis of OsMYB103 binding the ‘AACC’ motifs to OsCCRL1 promoter regions in Ubi::OsMYB103-GFP plants with anti-GFP antibody. Data are means ± SD (n = 3). F An EMSA assay of OsMYB103 protein directly binding to the probe ‘P2s’ (93 bp) and ‘P7s’ (40 bp) of OsCCRL1 promoter. ‘Biotin-P2s mutant’ and ‘Biotin-P7s mutant’ mean the mutant probes at ‘AACC’ motifs

Then, OsMYB103 could activate the OsCCRL1 promoter in a dual luciferase assay in rice protoplasts, with about three-fold higher LUC/REN ratio compared to the control (Additional file 1: Fig. S4C; Fig. 6C), suggesting that the OsCCRL1 promoter is regulated by OsMYB103. To confirm this notion, we performed chromatin immunoprecipitation and quantitative-PCR (ChIP-qPCR) analysis in Ubi::OsMYB103-GFP transgenic anthers. We detected 16 ‘AACC’ cis-elements recognized by OsMYB103 in the OsCCRL1 promoter within a 1700-bp region upstream of the transcription start site (Fig. 6D). Using primers containing the ‘AACC’ cis-element designed based on the OsCCRL1 promoter, all regions except ‘P5’ were enriched in Ubi::OsMYB103-GFP transgenic anthers with anti-GFP antibody (Fig. 6E). To further confirm that OsMYB103 directly binds the OsCCRL1 promoter, we performed an electrophoretic mobility-shift assay (EMSA) and found that OsMYB103 bound directly to the ‘P2s’ and ‘P7s’ probes of OsCCRL1 promoter (Fig. 6F). These results confirm that OsMYB103 directly regulates OsCCRL1 expression.

Phenotypic Analysis of Osmyb103 osccrl1 Double Mutant

Taking into account the relationship between OsMYB103 and OsCCRL1, we generated the osmyb103 osccrl1 double mutant by introducing the OsCCRL1 knockout vector into the osmyb103 mutant. We identified osmyb103 osccrl1 double mutant transgenic plants by sequencing (Additional file 1: Fig. S4D, E) and measured OsMYB103 and OsCCRL1 transcript levels in wild-type and osmyb103 osccrl1 double mutant anthers by RT-qPCR. OsMYB103 transcript levels did not significantly differ between two lines (Fig. 7A), but OsCCRL1 was not expressed at all in osmyb103 osccrl1 double mutant anthers (Fig. 7B), implying that OsMYB103 functions upstream of OsCCRL1. When we compared the phenotypes of the osmyb103 mutant, osccrl1 mutant and osmyb103 osccrl1 double mutants, the phenotype of osmyb103 osccrl1 double mutant was closer overall to that of osmyb103 single mutant, although in some respects the phenotypes of osmyb103 and osccrl1 were not easily distinguishable. For example, all three mutants lacked pollen, none showed anther dehiscence, and all had severely defective anther epidermis and inner surface (Fig. 7C, D).

Fig. 7

Phenotypic analysis of osmyb103 osccrl1 double mutant transgenic plants. A, B The relative expression comparison of OsMYB103 and OsCCRL1 between the wild-type and osmyb103 osccrl1 double mutant anthers. Data are means ± SD (n = 3). Asterisks indicate the significance differences determined by Student’s t-test (*, 0.01 ≤ P < 0.05; **, P < 0.01). C Spikelet, anther and pollen stained by I2/KI solution of wild type, osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant. Scale bars, 2 mm (c-1 to c-4); 500 μm (c-5 to c-8); 200 μm (c-9 to c-12). D The scanning electron microscopy of anther, anther epidermis surface, anther inner surface in wild type, osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant. Scale bars, 200 μm (d-1 to d-4); 50 μm (d-5 to d-12)

Therefore, semi-thin sections were performed and it showed that the abnormality of osmyb103 osccrl1 double mutant anthers began to be detectable at stage 9. During this stage, the tapetum of the double mutant began to swell, resembling the phenotypes of the osmyb103 mutant. At stage 11, the abnormality of the microspores appeared to be more serious in the double mutant than in osccrl1, such as the abortion phenotype of microspores, resembling the phenotype of the osmyb103 mutant (Fig. 8A–P). To identify the phenotype of osmyb103 osccrl1 double mutant was closer to osmyb103 mutant, we performed aniline blue staining to observe anther callose at stage 8b, observing a clear callose signal in the wild type and osccrl1 mutant, but not in osmyb103 mutant or osmyb103 osccrl1 double mutant (Fig. 8Q–T). Then, we stained anthers with sudan red 7B to identify the fatty acid differences. The anthers of the osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant showed less sudan red 7B staining than those of the wild type and the staining pattern of the osmyb103 osccrl1 double mutant was closer to that of osmyb103 mutant (Fig. 8U–X) again suggesting that OsCCRL1 functions downstream of OsMYB103.

Fig. 8

Semithin section, aniline staining and sudan red staining of the wild type, osmyb103, osccrl1, and osmyb103 osccrl1 double mutants. A–P Semi-thin sections of anthers at stages 9, 10, 11 and 12 in the wild type, osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant. The images show a single locule of an anther. E, epidermis; En, endothecium; T, tapetum; Msp, microspore; BP, bicellular pollen; MP, mature pollen; dMP, mature pollen. Scale bars, 25 μm. Q–T Callose staining of the wild type, osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant anther at stage 8b. The images show a single locule of an anther. Scale bars, 25 μm. U–X Sudan red of the wild type, osmyb103 mutant, osccrl1 mutant, and osmyb103 osccrl1 double mutant anther. Scale bars, 200 μm

Discussion

The Potential Role of OsCCRL1 in Phenylpropane Metabolism

In flowering plants, environmental changes easily affect the developing male gametophytes. To cope with this issue, plants have evolved two protective structures, the anther cuticle and pollen wall, composed of lipidic and phenolic compounds. Therefore, anther and pollen development requires the accumulation of large amounts of phenylpropanoid metabolites, especially lignin, hydroxycinnamic acid amides, and flavonol glycosides. Changes in the composition and levels of phenylpropanoids can affect anther development and pollen wall patterning (Herdt et al. 1978; Fellenberg and Vogt 2015; Shi et al. 2015). Arabidopsis REF3 gene encodes C4H, which transforms para-hydroxylates cinnamic acid to p-coumaric acid in the phenylpropanoid pathway. P-coumaric acid ester was considered to be the precursor of sporopollenin and the ref3-2 mutant produces little mature pollen and indehiscent anthers. The sterile phenotype of ref3-2 is due to the decrease in C4H enzyme activitywhich leads to reduced p-coumaric acid ester and ultimately decreased sporopollenin production (Schilmiller et al. 2009). In the current study, we identified the male-sterile rice mutant osccrl1 from indica, with abnormal anther cuticles and Ubisch bodies, as revealed by cytological observation, along with seriously defective pollen wall and tapetum development. Our phenotypic analyses indicated that OsCCRL1 has a comprehensive effect on anther development, especially on anther cuticle, pollen wall, Ubisch bodies and anther dehiscence, which are somehow related to phenylpropanoid metabolism.

Previous studies had identified LOC_Os09g32020.2 as OsTKPR1, which was considered to function in the fatty acid metabolic pathway for sporopollenin synthesis and pollen exine formation with ACOS and PKSA/B (Xu et al. 2019; Wang et al. 2013). ostkpr1-2, which was completely male sterile with smaller and paler anthers containing no viable pollen grains, was screened from a radiated mutant library by ⁶⁰Co γ-ray of 9522 (O. sativa ssp japonica) and showed a 15 nucleotide deletion in the fourth exon of LOC_Os09g32020.2 (Xu et al. 2019). Our mutant osccrl1 from indica was also completely male sterile without any viable pollen grains and showed one base substitution in the fourth exon of LOC_Os09g32020.2. Combining ostkpr1-2 and osccrl1 mutants, it showed that LOC_Os09g32020.2 controlled the anthers fertility both in japonica and indica. The difference between ostkpr1-2 and osccrl1 mutants was that the anthers of ostkpr1-2 were smaller and paler, and the abnormality of anthers in ostkpr1-2 began at the 9 stage, but the anther of osccrl1 mutant did not show smaller and whiter, and the abnormality of anthers in osccrl1 began at the 10 stage, which may be caused by their mutation sites, causing ostkpr1-2 mutant more serious.

Our phylogenetic analysis classified LOC_Os09g32020.2 into both the DFR family (involved in flavonoid biosynthesis) and the CCR family (involved in lignin biosynthesis). Through phylogenetic analysis, AtTKPR1 and AtTKPR2, two Arabidopsis homologs of LOC_Os09g32020.2, also belong to the DFR family. Therefore, combining previous reporters with our finding, OsTKPR1/OsCCRL1 might be involved fatty acid metabolism, flavonoid biosynthesis, and lignin biosynthesis. However, previous studies have shown that OsTKPR1 might not have a DFR-like function (Tang et al. 2009). And the expression of OsPKS2, thought to be involved in the same metabolic pathway with OsTKPR1/OsCCRL1, was not significantly altered in osccrl1 anther. The expression of OsCCRL1 itself in osccrl1 mutant was reduced, but the expression of OsPKS2 in osccrl1 mutant did not change much, suggesting that OsCCRL1 may be involved in another pathway which differed with OsPKS2. Based on these findings, we focused on the role of LOC_Os09g32020.2 as a CCR enzyme. Then, biochemical analysis revealed that total CCRs activity was higher in wild-type anthers than that in osccrl1 anthers, and lignin accumulation and content was reduced in osccrl1 anther. In addition, the recombinant protein OsCCRL1 may have CCR enzymatic activity in vitro, but the osccrl1 lost it. CCR is considered to bind the coenzymes NAD(P)(H) at the N-terminal region, and bind the substrate at the C-terminal region during the catalytic reaction. Therefore, we guess that the substitution of valine (hydrophobic) to glutamic acid (negatively charged) of OsCCRL1 possibly altered the affinity of OsCCRL1 to the substrate, further affecting the binding to the substrate. And eventually, the osccrl1 losing the CCR enzymatic activity.

CCR is a key enzyme in the phenylpropanoid pathway. Os4CL3 and OsCHS1, two other key enzymes in this pathway, were also greatly reduced in osccrl1 anthers, suggesting that the phenylpropanoid pathway was broadly disrupted. Phenylpropanoid pathway is also altered in ostkpr1-2 mutants (Xu et al. 2019). Therefore, OsTKPR1/OsCCRL1 plays an important role in the phenylpropanoid pathway. And it might be the disturbed phenylpropanoid pathway in the osccrl1 mutant that blocked the several components formation, such as cuticles, sporopollenin and lignin, further resulting in pollen infertility and anther indehiscence.

It is worth noting that the previously reported OsTKPR1 was located in the endoplasmic reticulum in tobacco leaf cells, but OsCCRL1 was localized in cytoplasm and nucleus in rice protoplasts and tobacco leaf cells in our study. We also found that in previous reports, AtTKPR1 was located in endoplasmic reticulum and AtTKPR2 was located in cytoplasm (Grienenberger et al. 2010; Lallemand et al. 2013). But DRL1 (Dihydroflavonol 4-reductase-like 1), that was AtTKPR1, was located in cytoplasm and nucleus (Tang et al. 2009). Therefore, TKPRs may be located in endoplasmic reticulum, nucleus and cytoplasm. What caused the subcellular localization difference of OsTKPR1 and OsCCRL1? We noticed that there was one base difference in the LOC_Os09g32020.2 genome sequence between the reference sequence given online and the sequence we sequenced from wild-type DNA, meaning that there was an amino acid difference at the site 80 between the online sequence and OsCCRL1. At the same time, the amino acid at the site 80 of protein OsTKPR1 was aspartic acid (D), which was consistent with the one online. Therefore, it may be this difference leads to the pattern change of subcellular localization between OsTKPR1 and OsCCRL1.

OsCCRL1 was Directly Regulated by OsMYB103

The MYB transcription factors is highly conserved and could activate or inhibit downstream gene expression. Various MYB transcription factors have been shown to be required for the regulation of phenylpropanoid metabolism. In rice, OsPAL6 and OsPAL8 are directly upregulated by OsMYB30, and OsMYB30 also binds to the promoters of Os4CL3 and Os4CL5, leading to guaiacyl lignin and syringyl lignin accumulation (He et al. 2020; Li et al. 2020). In addition, MYB15 directly binds to the promoters of PAL1, C4H, and COMT to promote guaiacyl lignin biosynthesis (Chezem et al. 2017). LTF1, a MYB transcriptional repressor, represses lignin biosynthesis by binding to the promoter of 4CL, whereas the degradation of LTF1 via the proteasome pathway activates lignin biosynthesis (Gui et al. 2019).

OsMYB103/OsMYB80/OsMS188/BM1 encodes an R2R3 MYB transcription factor that is highly conserved in both dicot and monocot plants. In japonica and indica rice, EMS-treated induced mutants and knockout transgenic plants of OsMYB103/OsMYB80/OsMS188/BM1 are completely male sterile, with small and white anthers. These mutants exhibit a lack of Ubisch bodies, and no mature microspores, and delayed tapetum PCD (Zhang et al. 2010; Xiang et al. 2020; Han et al. 2021; Lei et al. 2022), indicating that OsMYB103/OsMYB80/OsMS188/BM1 plays an important role in tapetum and pollen development, and this gene was easily induced to mutate. Although the osmyb103 mutants obtained by different methods were slightly different in some details, such as anther anthers epidermis, tapetum thickness, and the stage of abnormal anther, these osmyb103 mutants were all male sterile. OsMYB103/OsMYB80/OsMS188/BM1 transcription factor directly binds to the promoter regions of CYP703A3, CYP704B2, OsPKS2, OsPKS1, DPW, ABCG15, EAT1, and PTC1 to regulate their expressions (Xiang et al. 2020; Han et al. 2021; Lei et al. 2022). Comparative transcriptome analysis (RNA-seq) of downstream differentially expressed genes (DEG) in wild type and osmyb80 mutant, it suggested that OsMYB103/OsMYB80/OsMS188/BM1 functions in many processes, such as fatty acid and small-molecule metabolic process, transcriptional regulation process, ubiquitin-dependent protein degradation process, oxidation and reduction process, binding processes with proteins, lipids, and coenzymes, and oxidoreductase and hydrolase activity (Pan et al. 2020). Therefore, OsMYB103/OsMYB80/OsMS188/BM1 is a crucial transcription factor that regulates the expression of many downstream genes and participates in many pathways to function in anther and pollen development. OsMYB103/OsMYB80/OsMS188/BM1 might also be involved in regulating the phenylpropanoid pathway and lignin biosynthesis.

In our study, OsCCRL1 expression was significantly reduced in osmyb103 anthers. A dual luciferase assay in rice protoplasts, a ChIP analysis in Ubi::OsMYB103-GFP transgenic plants, and an EMSA analysis with OsMYB103 protein confirmed that OsMYB103 binds to the OsCCRL1 promoter to regulate its expression. More importantly, compared with the individual mutants osccrl1 and osmyb103, some phenotypes of the osmyb103 osccrl1 double mutant were biased towards those of osmyb103 mutant, such as whiter anthers, abnormal anther morphology at stage 9, lack of callose accumulation in microspores at stage 8b, and reduced sudan red staining of anthers. In the osmyb103 osccrl1 double mutant, OsCCRL1 expression was drastically reduced, but OsMYB103 expression seemed to be unchanged. Therefore, OsMYB103 is located upstream of OsCCRL1 and binds to its promoter to regulate its expression, which showed OsMYB103 also plays an important role in phenylpropanoid metabolism and tapetum development in rice anthers.

Conclusion

In this study, we demonstrated that OsCCRL1 (LOC_Os09g32020.2), one member of SDR, is involved in the phenylpropanoid pathway in rice anthers. OsCCRL1 is specifically expressed in the tapetum and microspores, and a replacement of valine (hydrophobic) by glutamic acid (negatively charged) at position 231 of OsCCRL1 led rice to have abnormal anthers, delayed tapetum development, reduced lignin accumulation, and complete male sterility. Solid genetic evidences, including genetic complementation and CRISPR/CAS9-derived transgenic plants analysis, supported the notion that OsCCRL1 is required for male fertility. Besides, OsCCRL1 was contributed to phenylpropanoid metabolism, which is vital for pollen wall formation, anther cuticle development, and anther dehiscence. In addition, we showed that OsCCRL1 is directly regulated by OsMYB103/OsMYB80/OsMS188/BM1, as confirmed by analysis of the osmyb103 osccrl1 double mutant, revealing a new regulatory pathway of phenylpropanoid metabolism and tapetal development in rice anthers.