Uncovering the role of the PPR protein PHOTOSYSTEM ONE BIOGENESIS FACTOR6 in splicing chloroplast group II introns

ABSTRACT

Nucleus‐encoded proteins are involved in intron splicing in plant chloroplasts. Although many splicing factors were discovered, the mechanisms for chloroplast intron splicing remain unknown. Here, we identified a P‐class pentatricopeptide repeat (PPR) protein, PHOTOSYSTEM ONE BIOGENESIS FACTOR6 (PBF6), that is essential for the accumulation of photosystem I complex. PBF6 bound to ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts, and was required for their splicing. Truncated PBF6 containing the 10 N‐terminal PPR motifs bound to specific sequences to ycf3 intron 1, clpP1 intron 2, and to the introns of petB, ndhA, and ndhB. PBF6 formed two multi‐subunit splicing complexes with other known splicing factors from the CRM, peptidyl‐tRNA hydrolase, RNase III, PORR, APO, DEAD‐box RNA helicase, and mTERF families. PBF6‐containing complex I was about 600 kDa, comprising nine known splicing factors: CAF1, CAF2, CRS2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2. PBF6‐containing complex II was about 300 kDa and contained three known splicing factors: CAF1, CAF2, and CFM2. Furthermore, two known PPR‐type splicing factors, PBF2 and ECD2, both of which are required for splicing the ycf3 intron 1, also formed two multi‐subunit splicing complexes with other known splicing factors. Importantly, these three PPR‐type splicing factors formed their splicing complexes independently. Our data suggest that a PPR‐type splicing factor forms splicing complexes with other known splicing factors under transient expression conditions to facilitate intron splicing and that several PPR‐type splicing factors work together to promote the splicing of the same intron through forming respective splicing complexes.

The P‐class pentatricopeptide repeat (PPR) protein PHOTOSYSTEM ONE BIOGENESIS FACTOR (PBF6) forms splicing complexes with other known splicing factors to facilitate chloroplast intron splicing. PBF6 cooperates with other PPR splicing factors to promote the splicing of the same intron through forming respective splicing complexes.

INTRODUCTION

Chloroplasts in eukaryotic cells originated from a cyanobacterial progenitor through endosymbiosis. During evolution, most genes from the genome of this cyanobacterial progenitor were either lost or transferred to the host nuclear genome, resulting in the retention of only a small number of genes in the chloroplast genomes of land plants. Thus, the chloroplast gene expression is highly regulated by many host nucleus‐encoded proteins, in particular at the post‐transcriptional level, including proteins modulating RNA stabilization, editing, and intron splicing in the green organelle (Small et al., 2023; Zhang et al., 2023a; Li et al., 2025; Qi et al., 2026).

Intron splicing is crucial for the chloroplast gene expression. Introns are categorized into two families, group I and group II, based on distinct structures and splicing mechanisms (de Longevialle et al., 2010). Group II introns predominate in land plant organelles, whereas group I introns are widespread in algae, exemplified by Chlamydomonas reinhardtii. For instance, the chloroplast genome of Arabidopsis thaliana comprises one group I intron and 20 group II introns. Group I intron splicing involves two consecutive trans‐esterification reactions, using an external free guanosine as a cofactor. While group II introns in bacteria are capable of self‐splicing, a process facilitated by their cognate intron‐encoded maturase (Mat), group II introns present in plant organelle have lost this self‐splicing capacity. Thus, nucleus‐encoded, chloroplast‐translocated splicing factors have evolved to splice group II introns (de Longevialle et al., 2010; Wang et al., 2022; Zeng et al., 2022; Wang and Tan, 2025).

In plant chloroplasts, a variety of protein factors encoded by the nuclear genome have been found to participate in the splicing process of introns. These splicing factors mainly include chloroplast RNA splicing and ribosome maturation (CRM) proteins, accumulation of photosystem one (APO) domain proteins, plant organellar RNA recognition (PORR)‐domain proteins, mitochondrial transcription termination factor (mTERF) proteins, DEAD‐box RNA helicases, a protein with an RNase III domain, a peptidyl‐tRNA hydrolase, and pentatricopeptide repeat (PPR) proteins (Wang et al., 2022; Hu et al., 2025; Wang and Tan, 2025; Zeng et al., 2022, 2026). A total of approximately 30 splicing factors have been characterized in Arabidopsis, and 15 members of these belong to the PPR protein family. However, the detailed molecular mechanisms underlying their functional actions remain poorly elucidated (Small et al., 2023).

The PPR family is characterized by a tandem array of degenerate motifs consisting of ~35 amino acids, and each PPR protein contains two to 30 such motifs (Barkan and Small, 2014; Zhang and Lu, 2019; Wang and Tan, 2025). The PPR family is greatly expanded in land plants, and the Arabidopsis genome contains more than 450 PPR genes. PPR proteins are ubiquitous in all eukaryotes; most are located in chloroplasts, mitochondria, or both (Wang and Tan, 2025). In plant mitochondria and chloroplasts, PPR proteins participate in almost all post‐transcriptional processes, and they play crucial roles in plant growth and development, cytoplasmic male sterility restoration, and responses to biotic and abiotic stresses (Schmitz‐Linneweber and Small, 2008; Rovira and Smith, 2019; Gong et al., 2025; Wang and Tan, 2025). They show sequence‐specific RNA‐binding activity and bind to the RNA targets in accordance with the one‐repeat‐one‐nucleotide binding rule, with the amino acids at positions 5 and 35 of each PPR motif determining the recognition of specific nucleotides (Yin et al., 2013; Shen et al., 2016). Based on their domain structure, PPR proteins can be categorized into two subfamilies, namely, the P and PLS subfamilies. P‐class PPR proteins function in RNA stabilization and intron splicing, and most PLS‐class PPR proteins participate in RNA editing (Barkan and Small, 2014; Wang and Tan, 2025).

A variety of P‐class PPR proteins were reported to be involved in the splicing of one to several introns in chloroplasts. About 15 P‐class PPR splicing factors have been identified in Arabidopsis, such as ALBINO EMBRYO AND SEEDLING (AES), EARLY CHLOROPLAST DEVELOPMENT 2 (ECD2), EMBRYO DEFECTIVE 1270 (EMB1270), EMB1417, PHOTOSYSTEM ONE BIOGENESIS FACTOR2 (PBF2), PIGMENT‐DEFECTIVE MUTANT 4 (PDM4), and SUPPRESSOR OF THF1 5 (SOT5, also reported as EMB2279) (Zeng et al., 2022; Huang et al., 2018; Wang et al., 2020a, 2020b, 2021; Zhang et al., 2021; An et al., 2023; Small et al., 2023; Huo et al., 2024). Recently, studies on the functions of P‐class PPR proteins in chloroplast intron splicing have expanded to rice (Oryza sativa), in which 21 PPR splicing factors have been identified, e.g., ALBINO SEEDLING‐LETHAL (OsASL), PALE GREEN LEAVES 3A (OsPGL3A), OsPPR11, YELLOW‐GREEN SEEDLING 1 (OsYS1), OsYS2, RICE ALBINO SEEDLING‐LETHAL (RAS), WHITE PANICLE RACHIS (WPR), and YOUNG LEAF WHITE STRIPE (YLWS) (Zeng et al., 2022; Lan et al., 2023; Small et al., 2023; Zhang et al., 2023b; Li et al., 2024; Xu et al., 2024; Qiu et al., 2025; Sun, Nie, et al., 2025; Zhao et al., 2025). Although numerous studies have demonstrated that P‐class PPR proteins participate in chloroplast intron splicing, their precise roles are still not fully elucidated (Small et al., 2023; Wang and Tan, 2025). Therefore, it is important to characterize the function of each PPR protein individually to gain a comprehensive, system‐level understanding of this family, when taking into account the scale and complexity of the PPR family.

As one of the two large pigment–protein complexes, photosystem I (PSI) is responsible for the conversion of solar energy into chemical energy across all oxygenic photosynthetic organisms. PSI is among the largest and most complex macromolecular assemblies (Amunts and Nelson, 2009). Indeed, plant PSI comprises a core complex with 14 subunits, four PSI light‐harvesting complex (LHCI) subunits, and hundreds of cofactors (Qin et al., 2015; Pan et al., 2018). PSI is recognized as one of the most efficient devices known in nature for the capture and conversion of light energy (Croce and van Amerongen, 2013; Nelson and Junge, 2015; Feng et al., 2025). Thus, elucidating how PSI is produced and assembled is crucial for engineering photosynthesis. One powerful method to reveal the mechanism behind PSI biogenesis is forward genetic screens, looking for mutants with impairments in PSI biogenesis. Many factors involved in PSI biogenesis were reported in land plants, including the hypothetical chloroplast open reading frames Ycf3 and Ycf4, PALE YELLOW GREEN 7 (PYG7), Ycf3‐INTERACTING PROTEIN 1 (Y3IP1), PSBP‐DOMAIN PROTEIN 1 (PPD1), PBF2, PBF8, PSI ASSEMBLY 2 (PSA2), and PSA3 (Ruf et al., 1997; Albus et al., 2010; Krech et al., 2012; Liu et al., 2012; Fristedt et al., 2014; Shen et al., 2017; Yang et al., 2017; Wang et al., 2020a; Zhang et al., 2024). However, the molecular mechanism of PSI biogenesis remains incompletely understood (Rolo et al., 2024).

In the current study, we identified an Arabidopsis mutant with impaired PSI biogenesis, which was designated pbf6. PBF6 encodes a P‐class PPR protein that is localized to chloroplasts. Knockout mutants of PBF6 show an albino phenotype and a complete loss of the PSI complex, resulting in seedling lethality. We show that PBF6 is essential for splicing ycf3 intron 1 and clpP1 intron 2, as well as the introns of petB, ndhA, and ndhB transcripts in chloroplasts, through its direct binding to these introns in the corresponding transcripts. Our data suggest that PBF6 forms two splicing complexes comprising nine or three known splicing factors. In addition, the two known PPR‐type splicing factors PBF2 and ECD2, together with PBF6, are required for splicing ycf3 intron 1, while they assemble with different known splicing factors to form two splicing complexes independently. Our findings reveal the detailed mechanism of the PPR protein PBF6 in regulating chloroplast group II intron splicing.

RESULTS

PBF6 is essential for the PSI function

To explore the molecular mechanism underlying PSI biogenesis, we performed screening on an Arabidopsis T‐DNA insertion mutant collection. Specifically, we screened for mutant plants with impaired PSI function as described in our previous studies (Wang et al., 2020a; Zhang et al., 2024). We thus identified and characterized two alleles for the photosystem I biogenesis factor 6 (pbf6) mutant: pbf6‐1 (SALK_150731) and pbf6‐2 (SALK_034168). Compared to wild‐type (WT) seedlings, the pbf6‐1 mutant showed retarded growth, with yellowish leaves. The pbf6‐2 mutant showed an albino phenotype and died early, being unable to grow photoautotrophically when in soil. However, the pbf6‐2 mutant could grow on Murashige and Skoog (MS) medium supplemented with sucrose, despite showing the albino phenotype and developing much more slowly than WT seedlings (Figure 1A). We observed a ca. 20% decrease in the leaf chlorophyll content in pbf6‐1 in comparison with WT seedlings, but detected no chlorophyll accumulation in the cotyledons of the pbf6‐2 mutant (Figure S1A). The light‐induced changes in P700 absorbance at 820 nm, an index representing PSI function, were about 50% lower in pbf6‐1 and were completely absent in pbf6‐2 (Figure 1B), suggesting that PSI function is considerably impaired in pbf6‐1 and totally abolished in pbf6‐2.

Figure 1.

Spectroscopic characterization of pbf6 mutants

(A) Representative photographs of wild‐type (WT), pbf6‐1, pbf6‐2, pbf6‐1 com, and pbf6‐2 com seedlings grown on Murashige and Skoog (MS) medium for 10 days. Scale bars, 0.1 cm. (B) P700 redox kinetics of leaves from WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings. The P700 redox state was investigated as P700 absorbance at 820 nm induced by exposure to far‐red light (720 nm). AL, actinic light (100 μmol m⁻² s⁻¹); FR, far‐red light; SL, saturating light pulse; ox., P700 oxidation; red., P700 reduction; and ΔAmax, maximum photo‐oxidizable P700. (C) 77 K fluorescence emission spectra of thylakoid membranes isolated from WT, pbf6‐1, and pbf6‐1 com seedlings. After excitation at 436 nm, fluorescence emission signals were acquired and then normalized to the PSII emission maximum at 688 nm. (D) Photochemical quantum yield of PSI, Y(I) in WT, pbf6‐1, and pbf6‐1 com seedlings. Data are means ± standard deviation (SD; n = 3 biological replicates). (E) Chlorophyll a fluorescence induction kinetics in WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings. AL, actinic light (100 μmol m⁻² s⁻¹); F_m, maximal fluorescence; F_o, minimal fluorescence; FR, far‐red light; SL, saturating light pulse. In (A–C, E), three independent biological replicates were performed, and a representative image is shown.

To investigate the changes in PSI function in the pbf6 mutants in more detail, we determined their 77K fluorescence emission spectra and effective photochemical yield of PSI, Y(I). We observed a maximum fluorescence emission of PSI at 731 nm in WT seedlings, but this peak shifted to 729 nm in pbf6‐1 (Figure 1C); we detected no fluorescence signal in pbf6‐2. This 2‐nm blue shift in the maximum fluorescence emission for PSI in pbf6‐1 is indicative of the dissociation of LHCI from PSI. This characteristic was previously described in PSI‐deficient mutants (Albus et al., 2010; Liu et al., 2012; Wang et al., 2017; Yang et al., 2017; Zhang et al., 2024). Y(I) was also lower in pbf6‐1 than in WT seedlings; again, we detected no values above the background for Y(I) in pbf6‐2 (Figure 1D).

To comprehensively explore the characteristics of the pbf6 mutants, we examined the fluorescence kinetics of chlorophyll a to assess PSII function (Figure 1E). There was a slight decrease in the maximal efficiency of PSII photochemistry (F _v /F _m ) in pbf6‐1 (0.696 ± 0.002 vs. WT 0.798 ± 0.002), but a greater decrease in F _v /F _m in pbf6‐2 (0.122 ± 0.02).

The T‐DNA insertions in pbf6‐1 and pbf6‐2 are located in the 5′ untranslated region (5’ UTR) and the first exon of At5g27270, respectively. We named this gene PBF6. The transcript levels of PBF6 in pbf6‐1 were about 50% of those in the WT and were almost undetectable in pbf6‐2 (Figure S1B–D). To confirm that the phenotypic defects of the pbf6 mutants result from PBF6 gene disruption, we assembled a complementation construct that is composed of the presumptive PBF6 coding sequence fused to the 3×FLAG tag sequence driven by the 35S promoter and introduced this construct into the pbf6‐1 mutant and heterozygous pbf6‐2/PBF6 plants. The complemented lines with each pbf6 mutant background were obtained and named pbf6‐1 com and pbf6‐2 com, respectively. PBF6 transcript levels in the pbf6‐1 com and pbf6‐2 com lines were comparable to those in the WT (Figure S1D). Moreover, both pbf6‐1 com and pbf6‐2 com displayed WT phenotypes (Figures 1, S1). These data confirm that the phenotypic defects of the pbf6 mutants are caused by the disruption of PBF6 (At5g27270). Moreover, pbf6‐1 is a knockdown mutant with a pale green phenotype, whereas pbf6‐2 is likely a knockout mutant with an albino phenotype that leads to seedling death.

PBF6 is essential for the accumulation of the PSI complex and the Cyt b ₆ f complex

To investigate how PSI is affected in pbf6 mutants, we performed immunoblotting to assess the abundance of PSI subunits in the WT and pbf6 (Figure 2A). In pbf6‐1, the abundance of PSI subunits PsaA, PsaD, PsaF, PsaG, and Lhca1 was ~50% or less than that in the WT plants. Cyt b ₆ and Cyt IV, two subunits of the cytochrome b ₆ f complex in pbf6‐1, accumulated to about 50% of WT levels, based on a dilution series. Nevertheless, the abundance of the PSII subunits (D1, D2, CP43, PsbO, and light‐harvesting complex II (LHCII)) was comparable in the WT and pbf6‐1, as was the diagnostic subunit of ATP synthase (CF1α). In pbf6‐2 seedlings, the subunits related to PSII, the cytochrome b ₆ f complex, PSI, and ATP synthase (CF1α) were either undetectable or present in trace amounts.

Figure 2.

Levels of thylakoid membrane proteins and complexes in WT, pbf6‐1 , pbf6‐2 , and pbf6‐2 com seedlings

(A) Immunoblot analysis of thylakoid membrane proteins using equal amounts of total leaf proteins (10 μg). Thylakoid membrane protein complexes and their diagnostic components are indicated on the left. (B) Blue native PAGE (BN‐PAGE) analysis of thylakoid membrane complexes (equivalent to 8 μg of total leaf proteins). Thylakoid membranes from the leaves of WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings were solubilized and separated by BN‐PAGE. The designations of thylakoid membrane protein complexes are indicated on the left. (C) Two‐dimensional (2D) BN/SDS‐PAGE separation of thylakoid membrane complexes (12 μg of total leaf proteins equivalent). Identities of the relevant proteins are indicated. The PSI reaction center proteins PsaA/B are indicated by arrows and PSI minor proteins are highlighted by the black ellipses. In (A–C), three independent biological replicates were performed, and a representative image is shown.

We next examined the accumulation of the PSI complex in the WT and pbf6 using blue native polyacrylamide gel electrophoresis (BN‐PAGE) (Figure 2B). Seven distinct chlorophyll‐containing protein bands were resolved. These bands represent the NDH–PSI complex (band I), PSII supercomplexes (band II), PSI–LHCI and dimeric PSII (band III), monomeric PSII (band IV), CP43‐free PSII (band V), trimeric LHCII (band VI), and monomeric LHCII (band VII) (Wu et al., 2025). The levels of chlorophyll–protein complexes in band III were lower in pbf6‐1 than in the WT, whereas the levels of chlorophyll–protein complexes in bands I, II, IV, and V were comparable in pbf6‐1 and the WT. Since band III is mainly composed of the PSI complex (Liu et al., 2012; Shen et al., 2017; Zhang et al., 2024), and as the levels of PSII subunits were unchanged largely in pbf6‐1 (Figure 2A), we conclude that the decrease in the levels of chlorophyll–protein complexes in band III observed in pbf6‐1 indicates a compromised accumulation of the PSI complex. In pbf6‐2, the chlorophyll–protein complexes in the seven bands were undetectable. The analysis of two‐dimensional SDS‐urea‐PAGE confirmed the lower abundance of PSI core proteins in pbf6‐1, while the levels of the PSII subunits (D1, D2, CP43, and CP47) and ATP synthase subunits (CF1α and CF1β) were comparable in the WT and pbf6‐1. In pbf6‐2, we detected trace amounts of PSI core subunits, PSII subunits (D1, D2, CP43, and CP47), and ATP synthase subunits (CF1α and CF1β) (Figure 2C). Collectively, these results indicate that the lack of PBF6 primarily leads to a major decrease in the accumulation of the PSI complex and the Cyt b ₆ f complex. The absence of the NDH–PSI complex and PSII‐related complexes in pbf6‐2 may be a secondary effect due to the lack of the PSI complex and the Cyt b ₆ f complex.

PBF6 is localized to chloroplasts

We predict that PBF6 is a P‐class PPR protein and has 22 PPR motifs according to the online tool PPRCODE (https://biolib.com/YaoYinYing/pprcode/) (Yan et al., 2019). According to the subcellular localization prediction program TargetP (http://www.cbs.dtu.dk/services/TargetP), a chloroplast transit peptide (TP) at the N terminus of PBF6 is predicted, indicating that PBF6 is a chloroplast‐localized protein. To test this possibility, we investigated the subcellular and suborganellar localization of PBF6 (Figure 3). To this end, we cloned the sequence encoding the first 187 aa of PBF6, including the TP, in‐frame and upstream of the green fluorescent protein gene (GFP); we then transfected Arabidopsis protoplasts with the resulting construct. Confocal laser‐scanning microscopy detected GFP fluorescence evenly within chloroplasts (Figure 3A), revealing that the PBF6–GFP fusion is localized to chloroplasts. We further determined the subcellular location of PBF6 in chloroplasts. We thus isolated the fractions of the stroma and membranes from chloroplasts using pbf6‐2 com seedlings and performed an immunoblot analysis with anti‐FLAG antibodies, which indicated that PBF6 mainly accumulates in the chloroplast stroma (Figure 3B).

Figure 3.

PBF6 localizes to chloroplasts

(A) Subcellular localization test of PBF6–GFP in Arabidopsis protoplasts. PBF6–GFP, fusion between GFP and the first 187 amino acids of PBF6; RbcS 1A‐GFP, ribulose 1, 5‐bisphosphate carboxylase small subunit 1‐GFP, chloroplast localization control; FRO1‐GFP, FERRIC REDUCTION OXIDASE 1‐GFP, mitochondrion control; Fibrillarin‐GFP, nuclear control; and pBI221‐GFP, control with free GFP. Chlorophyll autofluorescence is shown in magenta. Scale bars, 5 μm. Representative images were selected on the basis of the observation of about 150 protoplasts in 2 independent experiments. (B) Subcellular fractionation assay confirming that PBF6 is located in the chloroplast stroma. D2, a subunit of PSII and integral thylakoid membrane protein; RbcL, Rubisco large subunit, an abundant stromal protein. Intact chloroplasts isolated from PBF6‐3×FLAG transgenic seedlings were separated into the membrane and stromal fractions, and their proteins were probed with antibodies against FLAG, RbcL, or D2. Three independent biological replicates were performed, and a representative replicate is shown.

We also identified PBF6 homologs in land plants by using the NCBI database and reconstructed a phylogenetic tree. Arabidopsis PBF6 is highly similar to its homologs in dicotyledonous plant species (e.g., wild cabbage [Brassica oleracea], soybean [Glycine max], and grapevine [Vitis vinifera]), and monocots (e.g., maize [Zea mays], rice, and wheat [Triticum aestivum]). Arabidopsis PBF6 is more distantly related to its homologs in bryophytes (Physcomitrium patens) and ferns (Selaginella moellendorffii) (Figure S2). We conclude from this phylogenetic analysis that PBF6 is evolutionarily conserved in land plants (Figure S3).

PBF6 is essential for the accumulation of ycf3, clpP1, petB, ndhA, and ndhB transcripts

As shown above, PBF6 is essential for the accumulation of the PSI complex, the Cyt b ₆ f complex, and possibly other thylakoid complexes. As the PPR protein is crucial to organellar RNA metabolism (Barkan and Small, 2014), the transcript levels of all 114 chloroplast‐encoded genes were determined in the WT and pbf6 (Figure S4). Among 114 genes, the transcript abundance of 109 genes was similar in the WT and the two pbf6 mutants. Notably, the transcript levels of ycf3, petB, ndhA, ndhB, and clpP1 were much lower in pbf6‐1 and pbf6‐2. These data suggest that PBF6 is essential for the accumulation of ycf3, petB, ndhA, ndhB, and clpP1 transcripts.

PBF6 is essential for splicing ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts

The P‐class PPR protein is required for organellar RNA splicing in plants (Stern et al., 2010; Barkan and Small, 2014). To investigate the possible role of PBF6 in chloroplast intron splicing, we performed rRNA‐depleted total RNA‐seq on total RNA isolated from the WT and pbf6‐2. We aligned the sequencing data to the chloroplast genome (TAIR10 version) using TopHat 2, followed by data processing, filtering, and normalization, to obtain the relative sequencing depth at each chloroplast genomic locus. Full‐scale coverage of the chloroplast genome enables comprehensive profiling of the organellar transcriptional landscape (Figure S5). The abundance of exons and introns in the transcripts from chloroplast genes was compared between the WT and pbf6‐2 (Figure 4A). There was no remarkable decrease in the abundance of chloroplast exons in pbf6‐2. Importantly, the abundance of several introns, including ycf3 intron 1 and clpP1 intron 2, was higher in pbf6‐2 than in the WT (Figure 4A), implying that PBF6 is required for the splicing of some chloroplast introns. When we calculated the splicing efficiency of chloroplast introns (excluding six transfer RNA [tRNA] introns) using the rRNA‐depleted total RNA‐seq data according to the method by Castandet et al. (2016), we observed a markedly lower splicing efficiency for ycf3 intron 1, clpP1 intron 2, petB, ndhA, and ndhB introns in pbf6‐2 relative to that in the WT (Figure 4B).

Figure 4.

Analysis of RNA splicing efficiency in WT and pbf6‐2 seedlings based on RNA‐seq

(A) Differential abundance of exons and introns from chloroplast transcripts between WT and pbf6‐2 seedlings. The average RPKM values from n = 3 biological replicates were calculated and subjected to Log₁₀ normalization. Exons in the large inverted repeat were included only once. The arrows indicate ycf3 intron 1 and clpP1 intron 2. RPKM, reads per kilobase per million mapped reads. (B) Splicing efficiency of the indicated introns in WT and pbf6‐2 seedlings. The splicing efficiency was calculated according to Castandet et al. (2016). Please note that the splicing efficiency for introns in tRNA could not be calculated due to their poor coverage in the RNA‐seq data set. Data are means ± SD (n = 3 biological replicates). Statistical significance was determined using a two‐tailed Student's t‐test. *P < 0.05; **P < 0.01; and ***P < 0.001.

To verify that PBF6 is required for the splicing of these chloroplast introns, we analyzed the splicing pattern of 21 chloroplast introns using RT‐PCR (Figures 5A, S6). We observed a substantial accumulation of unspliced RNA molecules containing ycf3 intron 1, clpP1 intron 2, or the introns from petB, ndhA, and ndhB transcripts in the two pbf6 mutants, with a greater apparent abundance in the null allele pbf6‐2. The splicing pattern of another 16 chloroplast introns was normal in the pbf6 mutants. We then calculated the ratio of spliced to unspliced transcripts (splicing efficiency) in WT, pbf6‐1, and pbf6‐2 seedlings by performing RT‐qPCR. In agreement with the RT‐PCR analysis, the splicing efficiency for ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts was considerably lower in pbf6 than in the WT, while the splicing efficiency for other chloroplast introns was comparable among the WT, pbf6‐1, and pbf6‐2 (Figure 5B).

Figure 5.

PBF6 is required for the splicing of ycf3 intron 1, clpP1 intron 2, and introns from petB , ndhA , and ndhB transcripts

(A) RT‐PCR analysis of the splicing pattern in transcripts from the chloroplast intron‐containing genes ycf3, clpP1, petB, ndhA, and ndhB in WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings. un, unspliced; s, spliced. (B) Quantification of the splicing efficiency of chloroplast introns in WT, pbf6‐1, and pbf6‐2 seedlings by RT‐qPCR. The ratio between the WT and each mutant was subjected to Log₂ normalization. Data are means ± SD (n = 3 biological replicates). Statistical significance was determined using a two‐tailed Student's t‐test. *P < 0.05; **P < 0.01; and ***P < 0.001. (C) Northern blot analysis of ycf3, clpP1, petB, ndhA, and ndhB splicing in WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings. In gene structure schematic diagrams, exons and introns are represented by black and gray boxes, respectively. The numbers below exons and introns indicate the size of each exon and intron. The probes for exons and introns are represented by black lines. The numbers below probes indicate the size of each probe. The size of RNA markers is shown in kb to the left of each panel. Each lane was loaded with 15 μg of total RNA. Ethidium bromide staining of the 25S rRNA band was used as a loading control. Ex, exon; In, intron; M, mature spliced RNA; P, precursor. In (A, C), three independent biological replicates were performed, and a representative image is shown.

We then explored the defect in the splicing of ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts in pbf6‐1 and pbf6‐2 (Figure 5C). We determined that the precursors of ycf3, clpP1, petB, ndhA, and ndhB transcripts overaccumulate in the pbf6 mutants, with mature transcripts for ycf3, clpP1, petB, ndhA, and ndhB being nearly undetected in pbf6‐2. Together, these data indicate that PBF6 is involved in the splicing of ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts.

PBF6 binds to ycf3 intron 1 and clpP1 intron 2, and to the introns of petB, ndhA, and ndhB transcripts

We wondered whether PBF6 binds to ycf3 intron 1 and clpP1 intron 2, and to the introns of petB, ndhA, and ndhB transcripts in vivo. We therefore performed RNA immunoprecipitation (RIP) using pbf6‐2 com seedlings. We observed a clear 4–5‐fold enrichment of PBF6 at these introns (Figure 6A, B). These data indicate that PBF6 functions in the splicing of introns from ycf3, clpP1, petB, ndhA, and ndhB transcripts by binding to their respective introns in vivo.

Figure 6.

PBF6 binds to target RNAs in vivo and in vitro

(A) Diagrams of the ycf3, clpP1, ndhA, ndhB, petB, and petD genes, and positions of the RT‐qPCR primers used for detecting their unspliced transcripts. (B) RT‐qPCR analysis of the levels of ycf3, clpP1, ndhA, ndhB, and petB transcripts immunoprecipitated from PBF6‐3×FLAG transgenic seedlings with an anti‐FLAG antibody. petD was used as the negative control. Data are means ± SD (n = 3 biological replicates) and were normalized to the IgG negative control (set to 1). Statistical significance was determined using a two‐tailed Student's t‐test. *P < 0.05; **P < 0.01; and ***P < 0.001. (C) Diagram of PBF6 showing the PPR motifs and their predicted binding sites according to the PPR code. “Residue #5” and “Residue #35” refer to amino acids 5 and 35 in each PPR motif, respectively. “X” indicates positions where no specific RNA base was identified in biochemical assays. The truncated PBF6 variant PBF6‐10N, consisting of the first 10 PPR motifs and used for RNA electrophoretic mobility shift assays (REMSAs), is indicated. (D–H) REMSAs showing the binding of recombinant‐purified PBF6‐10N‐His to the introns of ycf3 (D), clpP1 (E), ndhA (F), ndhB (G), and petB (H), as well as to sequences ‘GCCGAGG’ (I) and ‘GCCGAUG’ (J). PBF6‐10N‐His, His, biotin‐labeled probes, and cold probe were present (+) or absent (−) in each reaction. The cold probe was used as a competitor at increasing amounts of 50× and 100×; the biotin‐labeled ycf3‐exon3 probe was used as a noncompetitor. In (D–J), three independent biological replicates were performed, and representative images are shown.

According to the PPR code, we further identified the potential binding sites of PBF6 using the MEME online tool (https://meme-suite.org/meme/index.html) (Yan et al., 2019). The specific RNA sequence recognized by PBF6 was predicted to be GX(C/U)GAGGXXXG(U/C)XXGCUAUXGX, where X is any RNA base. Although this predicted binding sequence is not present in the ycf3, clpP1, petB, ndhA, or ndhB genes, we did detect the first seven nucleotides (GCCGAGG) of this predicted sequence in ycf3 intron 1 and clpP1 intron 2, and in the introns of ndhA and ndhB transcripts, and a highly similar sequence ‘GCCGAUG’ in the petB intron (Figure 6C).

Previous studies have shown that the C‐terminal and N‐terminal regions of several PPR proteins with very long PPR motifs may recognize their respective RNA‐binding sequences (Fujii et al., 2013). For instance, the PPR protein PROTON GRADIENT REGULATION 3 (PGR3) harbors 27 PPR repeats. Its 16 PPR repeats at the N‐terminal region stabilize petL transcripts through sequence‐specific binding, whereas its 11 PPR repeats at the C‐terminal region participate in regulating the translation of petL and ndhA transcripts (Fujii et al., 2013). As the target introns of PBF6 are recognized by its first seven PPR motifs, we hypothesized that PBF6 recognizes its specific target sequences in ycf3 intron 1 and clpP1 intron 2, and in the introns of petB, ndhA, and ndhB transcripts via its 7 PPR repeats at the N‐terminal region, whereas its 15 PPR repeats at the C‐terminal region may be more flexible in terms of binding specificity. To test this hypothesis, we conducted RNA electrophoretic mobility shift assays (REMSAs) using purified recombinant variants of PBF6. Given the difficulty in generating full‐length mature PBF6, we therefore generated its first 10 N‐terminal PPR motifs fused to a His tag, yielding the recombinant protein PBF6‐10N. We mixed recombinant‐purified PBF6‐10N with probes corresponding to the predicted PBF6‐binding sequences in ycf3 intron 1 and clpP1 intron 2, and in the introns of petB, ndhA and ndhB transcripts (Figure 6D–H). We established that PBF6‐10N binds to the probes derived from all introns tested, indicating that PBF6 directly binds to the predicted binding sequences of these introns. The predicted binding sites for PBF6 in the introns of ycf3, clpP1, ndhA, and ndhB all start with ‘GCCGAGG’, and recombinant‐purified PBF6‐10N bound to a probe with this consensus sequence (Figure 6I). In addition, recombinant‐purified PBF6‐10N also bound to a probe with a highly similar sequence “GCCGAUG” in the petB intron (Figure 6J). These results indicate that the consensus sequence “GCCGAGG” in ycf3 intron 1 and clpP1 intron 2, and the introns of ndhA and ndhB transcripts and a highly similar sequence ‘GCCGAUG’ in the petB intron are the specific binding sequences for PBF6.

PBF6 forms two splicing complexes to facilitate the splicing of ycf3 intron 1, clpP1 intron 2, petB, ndhA, and ndhB introns

Our above results demonstrated that PBF6 is required for the splicing of ycf3 intron 1, clpP1 intron 2, petB, ndhA, and ndhB introns. Earlier studies have demonstrated that these introns are also bound by 13 known chloroplast splicing factors in Arabidopsis, that is, CAF1, CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, LEFKO, APO1, APO2, RH3, ISE2, and mTERF2, belonging to CRM, peptidyl‐tRNA hydrolase, RNase III, PORR, APO, DEAD‐box RNA helicase, and mTERF families (Table S1). We speculated that PBF6 might cooperate with some or all 13 splicing factors to participate in the splicing of these introns.

We asked whether PBF6 associates with 13 splicing factors to form splicing complexes. Thus, we performed co‐immunoprecipitation (Co‐IP) assays with anti‐FLAG antibodies using total protein extracted from Arabidopsis protoplasts co‐transfected with constructs encoding PBF6‐3×HA and each splicing factor tagged with FLAG (Figure 7). We found that PBF6 coimmunoprecipitates with 10 known chloroplast splicing factors, namely, CAF1, CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2, but not with APO1, LEFKO, or ISE2. To confirm whether PBF6 coimmunoprecipitates specifically with these ten splicing factors, we used a non‐splicing factor PPR protein ECB2 (EARLY CHLOROPLAST BIOGENESIS2) containing 18 PPR motifs as a negative control, since ECB2 is involved in chloroplast RNA editing but not in RNA splicing (Yu et al., 2009). We showed that ECB2 does not coimmunoprecipitate with these ten splicing factors, confirming that PBF6 is specifically associated with these ten splicing factors. In addition, our results demonstrated that the association of PBF6 with these ten splicing factors is independent of RNA, implying that this co‐association is shared through protein–protein interactions rather than shared RNA substrates. Collectively, these results suggest that PBF6 associates with these ten splicing factors via protein–protein interactions and may form one or more complexes with some or all of them.

Figure 7.

PBF6 associates with ten known splicing factors

Co‐immunoprecipitation assays testing the potential interaction of PBF6 with 13 known chloroplast splicing factors: CAF1, CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, LEFKO, APO1, APO2, RH3, ISE2, and mTERF2. Arabidopsis protoplasts were co‐transfected with constructs encoding 3×HA‐tagged PBF6 and 3×FLAG‐tagged CAF1 (or CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, LEFKO, APO1, APO2, RH3, ISE2, and mTERF2), with each gene driven by the 35S promoter. The resulting protoplasts were lysed, treated with or without RNase, and incubated with anti‐FLAG Magnetic Agarose. The immunoprecipitates were probed by immunoblotting with antibodies against HA and FLAG. A non‐splicing factor PPR protein ECB2 was used as a negative control. Three independent biological replicates were performed, and representative images are shown.

To investigate whether PBF6 forms complexes with these ten splicing factors, we performed affinity purification of PBF6‐3×FLAG with anti‐FLAG antibodies to purify PBF6‐containing complex(es) using Arabidopsis protoplasts co‐expressing PBF6‐3×FLAG and CAF1‐3×HA, CAF2‐3×HA, CRS2‐3×HA, CFM2‐3×HA, CFM3a‐3×HA, RNC1‐3×HA, WTF1‐3×HA, LEFKO‐3×HA, APO1‐3×HA, APO2‐3×HA, RH3‐3×HA, ISE2‐3×HA, or mTERF2‐3×HA (Figure 8A). We analyzed the affinity‐purified products co‐purifying with PBF6‐3×FLAG by 2D BN‐PAGE/SDS‐PAGE with an anti‐FLAG antibody, which revealed two PBF6‐containing complexes, PBF6‐containing complexes I and II, with molecular weights of about 600 and 300 kDa, respectively. We then explored the polypeptide composition of these PBF6‐containing complexes by immunoblot analysis with anti‐HA antibodies. We found that nine splicing factors are present in the PBF6‐containing complex I, namely, CAF1, CAF2, CRS2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2. The cumulative molecular weight of these nine factors, plus PBF6, was about 580 kDa, which is close to the observed molecular weight of PBF6‐containing complex I. PBF6‐containing complex II comprised the three known splicing factors CAF1, CAF2, and CFM2. The cumulative molecular weight of these three factors, plus PBF6, was about 280 kDa, which is close to the observed molecular weight of PBF6‐containing complex II. These results suggest that PBF6 forms two splicing complexes with three or nine known splicing factors as components under transient expression conditions.

Figure 8.

PBF6 mediates the formation of two different splicing complexes

(A) Separation and identification of PBF6‐containing complexes from the affinity‐purified PBF6‐3×FLAG immunoprecipitates on BN‐PAGE gels. PBF6‐3×FLAG was co‐transfected in Arabidopsis protoplasts with CAF1‐3×HA, CAF2‐3×HA, CRS2‐3×HA, CFM2‐3×HA, CFM3a‐3×HA, RNC1‐3×HA, WTF1‐3×HA, LEFKO‐3×HA, APO1‐3×HA, APO2‐3×HA, RH3‐3×HA, ISE2‐3×HA, or mTERF2‐3×HA. The polypeptide composition of PBF6‐containing complexes was analyzed using 2D BN‐PAGE/SDS‐PAGE. FLAG‐trap beads were used for affinity purification of the PBF6‐containing complexes; anti‐FLAG was used for the detection of PBF6‐3×FLAG and anti‐HA antibodies were used for the detection of other splicing factors fused with the 3×HA tag. A non‐splicing factor PPR protein ECB2 was used as a negative control. Three independent biological replicates were performed, and a representative replicate is shown. (B) Luciferase complementation imaging assays testing the interaction of PBF6 with CAF1, CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, LEFKO, APO1, APO2, RH3, ISE2, and mTERF2. PBF6 was fused to the N‐terminal half of firefly luciferase (nLUC), while each chloroplast splicing factor (CAF1, CAF2, CRS2, CFM2, CFM3a, RNC1, WTF1, LEFKO, APO1, APO2, RH3, ISE2, and mTERF2) was individually fused to the C terminal half of LUC (cLUC). The PBF6‐nLUC and cLUC constructs were co‐infiltrated in Nicotiana benthamiana leaves via Agrobacterium‐mediated infiltration. LUC activity was measured 48 h post‐infiltration. Data are means ± SD (n = 8 biological replicates).

To understand how PBF6 forms these two splicing complexes, we performed luciferase complementation imaging analyses to identify the splicing factors that physically interact with PBF6 (Figure 8B). To this end, we fused PBF6 with the N‐terminal half of firefly luciferase (nLUC), while each splicing factor was individually fused with the C‐terminal half of LUC (cLUC). We co‐infiltrated the leaves of Nicotiana benthamiana plants with PBF6‐nLUC and each cLUC‐based construct, and measured LUC activity 48 h later. We detected high LUC activity only when PBF6‐nLUC was co‐expressed with CAF1‐cLUC or CAF2‐cLUC (Figure 8B), suggesting that PBF6 interacts with CAF1 and CAF2, but not with the other splicing factors, suggesting that PBF6 forms two splicing complexes by interacting with CAF1 and CAF2.

The interaction interface of PBF6 for CAF1 and CAF2 was delineated using yeast two‐hybrid assays (Figure S7). We divided the 22 PPR motifs into three nonoverlapping segments: motifs 1–8, 9–15, and 16–22. We found that motifs 1–22 and 16–22 show self‐activation activity when fused to the DNA‐binding domain of yeast GAL4, precluding us from testing their interaction with CAF1 and CAF2. Motifs 1–8 interacted with CAF1 and CAF2, while motifs 9–15 did not. Further dissection of motifs 1–8 indicated that PBF6 can interact with both CAF1 and CAF2 through its PPR motifs 4–7.

PBF6 and other PPR‐type splicing factors PBF2 and ECD2 independently form splicing complexes that separately facilitate the splicing of ycf3 intron 1

We found that the splicing of over half of the chloroplast introns requires the participation of at least two PPR‐type splicing factors (Table S2). The ycf3 intron 1 splicing requires the participation of four PPR splicing factors PBF2, ECD2, EMB1270, and PBF6 as shown in this study (Wang et al., 2020a; Wang et al., 2021; Zhang et al., 2021). We sought to investigate how these four PPR proteins cooperate in ycf3 intron 1 splicing. As we could not obtain an expression vector for EMB1270, we focused on PBF2 and ECD2 in this study.

We investigated whether ECD2 and PBF2 also form splicing complexes by performing co‐IP with anti‐FLAG antibodies on total protein extracted from Arabidopsis protoplasts transfected with an ECD2‐3×FLAG or PBF2‐3×FLAG construct. Indeed, we detected ECD2 in two complexes: ECD2‐containing complexes I and II (Figure 9A). As with the PBF6‐containing complexes, ECD2‐containing complex I was about 600 kDa and comprised nine known splicing factors (CAF1, CAF2, CRS2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2); ECD2‐containing complex II was about 280 kDa and contained three known splicing factors (CAF1, CAF2, and CFM2). We performed a similar analysis for PBF2, which suggested that PBF2 was also part of two complexes: PBF2‐containing complexes I and II (Figure 9B). PBF2‐containing complex I was about 460 kDa and contained the three known splicing factors CAF2, CRS2, and APO1. PBF2‐containing complex II was about 260 kDa and comprised the three known splicing factors CAF1, CAF2, and CFM2. The composition of PBF2‐containing complex II was similar to those of PBF6‐containing complex II and ECD2‐containing complex II. These results indicate that various PPR‐type splicing factors may be part of similar or different splicing complexes.

Figure 9.

Different PPR‐type splicing factors required for splicing of a single intron independently mediate the formation of splicing complexes with other known splicing factors

(A, B) Separation and identification of the components in the ECD2‐ or PBF2‐containing complexes from affinity‐purified ECD2‐3×FLAG (A) and PBF2‐3×FLAG (B) immunoprecipitates on BN‐PAGE gels. The ECD2‐3×FLAG or PBF2‐3×FLAG construct was co‐transfected in Arabidopsis protoplasts with CAF1‐3×HA, CAF2‐3×HA, CRS2‐3×HA, CFM2‐3×HA, CFM3a‐3×HA, RNC1‐3×HA, WTF1‐3×HA, LEFKO‐3×HA, APO1‐3×HA, APO2‐3×HA, RH3‐3×HA, ISE2‐3×HA, or mTERF2‐3×HA. The polypeptide composition of ECD2‐containing complexes or PBF2‐containing complexes was analyzed using 2D BN‐PAGE/SDS‐PAGE. FLAG‐trap beads were used for affinity purification of the ECD2‐ or PBF2‐containing complexes; anti‐FLAG was used for the detection of ECD2‐3×FLAG or PBF2‐3×FLAG and anti‐HA antibodies were used for the detection of other splicing factors fused with the 3×HA tag. (C) PBF6, ECD2, and PBF2 do not co‐immunoprecipitate with each other. (i) Co‐IP assay of PBF6‐3×FLAG in Arabidopsis protoplasts co‐transfected with PBF6‐3×FLAG and PBF2‐3×HA or ECD2‐3×HA. (ii) Co‐IP assay of PBF2‐3×FLAG in Arabidopsis protoplasts co‐transfected with PBF2‐3×FLAG and PBF6‐3×HA or ECD2‐3×HA. (iii) Co‐IP assay of ECD2‐3×FLAG in Arabidopsis protoplasts co‐transfected with ECD2‐3×FLAG and PBF6‐3×HA or PBF2‐3×HA. Total proteins were incubated with magnetic agarose beads conjugated to the HA‐tag antibody. The immunoprecipitates were probed with antibodies against FLAG or HA. For all panels, three independent biological replicates were performed, and a representative replicate is shown.

To validate whether all the detected splicing factors in PBF6‐, ECD2‐, or PBF2‐containing complexes through affinity purification are integral components of these complexes, we used the non‐splicing factor PPR protein ECB2 as a negative control again to examine whether ECB2 co‐purifies with all 13 above‐mentioned splicing factors by 2D BN‐PAGE/SDS‐PAGE (Figure S8). Our results showed that none of the 13 splicing factors was detected in the affinity‐purified products of ECB2‐3×FLAG, although an ECB2‐containing complex with a molecular weight of about 500 kDa was observed. These results indicate that ECB2 does not co‐purify with any of 13 splicing factors examined and that all the detected splicing factors are integral components of the affinity‐purified complexes obtained from PBF6‐3×FLAG, ECD2‐3×FLAG, or PBF2‐3×FLAG.

Next, we asked whether the formation of the above splicing complexes requires other PPR‐type splicing factors. Accordingly, we performed co‐IP assays on total protein extracted from Arabidopsis protoplasts co‐transfected with constructs encoding tagged versions of PBF2, PBF6, and ECD2. We observed that PBF6 does not co‐immunoprecipitate with PBF2 or ECD2; similarly, PBF2 did not co‐immunoprecipitate with PBF6 or ECD2, and ECD2 did not co‐immunoprecipitate with PBF6 or PBF2 (Figure 9C). These results suggest that PBF6, PBF2, and ECD2 are part of independent splicing complexes.

DISCUSSION

PBF6 is essential for the splicing of specific group II introns, chloroplast biogenesis, and plant development

In this study, PBF6, a chloroplast‐localized P‐class PPR protein with 22 PPR motifs, was characterized (Figures 3, S2). Loss‐of‐function mutants of PBF6 showed severe impairments in the splicing of five chloroplast introns: ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts (4, 5, S6). RIP‐qPCR assays demonstrated that PBF6 specifically associates with these introns (Figure 6A, B). In addition, REMSA indicated that PBF6 binds to these introns by recognizing the specific sequences, which matches the predicted sequences of the PPR motifs of PBF6 (Figure 6C–J). Moreover, we observed defects in the splicing of these introns in two pbf6 alleles, which were both rescued to WT levels through the introduction of a 35S:PBF6‐3×FLAG transgene in each pbf6 mutant (1, 2). Taken together, these data demonstrate that PBF6 is required for the splicing of five chloroplast introns.

The five introns reside in the genes ycf3, clpP1, petB, ndhA, and ndhB, which encode Ycf3, ClpP1, Cyt b6, NdhA, and NdhB, respectively. Ycf3 is essential for PSI assembly (Naver et al., 2001). The knockout of ycf3 results in a block in PSI biogenesis and its knockout mutants cannot grow photoautotrophically. Loss‐of‐function mutants in clpP1 develop white cotyledons and lose the ability to produce shoots in tobacco (Kuroda and Maliga, 2003). Deletion mutants for petB failed to accumulate the Cyt b ₆ f complex and developed into yellowish seedlings that eventually died in tobacco (Nicotiana tabacum) (Monde et al., 2000). NdhA and NdhB are subunits of the NDH complex, a crucial component of the photosynthetic cyclic electron flow mediating electron transport between PSI and Cyt b ₆ f. However, mutants lacking the NDH complex show no obvious phenotype when grown under normal conditions (Peltier et al., 2016). Thus, defects in the splicing of ycf3, petB, and clpP1 introns result in the loss of the PSI complex, the Cyt b ₆ f complex, and the Clp protease complex, respectively, which may be responsible for the albino phenotype and seedling death of the pbf6‐2 mutants.

Notably, we observed an almost complete loss of all the thylakoid protein complexes in pbf6‐2 (Figure 2), suggesting a more general chloroplast developmental defect that may be associated with a general defect in chloroplast gene expression or ribosome assembly. In the present study, our results imply that a general defect in plastid gene expression does not occur, as transcript levels showed a significant decrease only in ycf3, petB, ndhA, ndhB, and clpP1, but not in other plastid genes in pbf6‐2 (Figure S4). Instead, we observed almost total loss of three ribosomal subunits (RPS1, RPS2, and RPL15) in pbf6‐2 (Figure S9), indicating a general defect in plastid ribosome assembly in pbf6‐2. What is the causality between the splicing defects and a general defect in ribosome assembly in pbf6‐2? Our results showed that loss of PBF6 function fails to splice clpP1 intron 2, leading to the loss of mature clpP1 transcripts (Figure S4). The plastid Clp protease plays roles in chloroplast biogenesis and plant development (Nishimura and van Wijk, 2015; Rodriguez‐Concepcion et al., 2019). The plastid Clp core complex consists of the asymmetric P‐ring (ClpP3–6 subunits) and the R‐ring (ClpP1 and ClpR1–4 subunits) (Nishimura et al., 2016). Knockout of various Clp genes results in albino embryos or seedling lethality (Kato and Sakamoto, 2010; Olinares et al., 2011). For example, the null mutant for ClpR4 shows albino embryos, with seedling growth arrested at the cotyledon stage, and loss of ClpP1 results in white cotyledons and ablation of tobacco shoots (Kuroda and Maliga, 2003; Kim et al., 2009). It is still unclear how the plastid Clp protease regulates chloroplast development and plant development. Interestingly, it was found that there was a dramatic decrease in the levels of several plastid ribosomal subunits in clpR4 mutants, implying that the plastid Clp protease is indispensable for the plastid ribosome assembly (Wu et al., 2013). A defect in plastid ribosome assembly in pbf6‐2 may be associated with a dramatic decrease of mature clpP1 transcripts, which is due to a severe impairment in the clpP1 intron 2 splicing. Thus, we tentatively propose that a general defect in plastid ribosome assembly in pbf6‐2 mutants may be a secondary consequence of the splicing defects. The exact relationship between the splicing defects and a general defect in plastid ribosome assembly in pbf6‐2 mutants remains to be investigated.

How does PBF6 function in intron splicing?

The chloroplast introns from land plants have lost the capability for self‐splicing during evolution because they are unable to take on a catalytically active conformation. Thus, a variety of nucleus‐encoded splicing factors have stepped in to aid the folding of introns into a conformation with catalytic activity, including members from the CRM, APO, DEAD‐box RNA helicase, mTERF, PORR, and PPR protein families. Although a growing number of splicing factors were uncovered in various land plants (Wang et al., 2022; Zeng et al., 2022; Small et al., 2023; Wang and Tan, 2025), the mechanism of each of these factors is still unknown.

Of the 30 splicing factors identified in Arabidopsis (Table S3), half belong to the PPR family, indicating the importance of PPR proteins in the splicing of chloroplast group II introns. Recently, several PPR proteins were shown to interact physically with other splicing factors in chloroplasts, e.g., CAF1, CAF2, CFM2, CRS2, and WTF1 (Wang et al., 2020a; Zhang et al., 2021; Huo et al., 2024). Through our analysis of splicing factors in Arabidopsis, we have discovered that most splicing factors participate in the splicing of multiple introns. Conversely, the splicing of an individual intron often requires the cooperation of multiple splicing factors (Tables S1–S3). Therefore, it is most likely that a PPR protein may form one or more protein complexes with other splicing factors to splice the targeted introns by recognizing specific sequences in each intron.

Surprisingly, our results suggest that PBF6 forms two splicing complexes with other known splicing factors. These known splicing factors pertain to all possible families of splicing factors and are indispensable for the splicing of all group II introns. PBF6‐containing complex I has a large size, with a molecular weight of about 600 kDa, and comprises at least nine known splicing factors, namely, CAF1, CAF2, CRS2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2. PBF6‐containing complex II is about 300 kDa and contains three known splicing factors: CAF1, CAF2, and CFM2 (Figure 8A). Similarly, ECD2 is also part of two splicing complexes: ECD2‐containing complex I and ECD2‐containing complex II. ECD2‐containing complex I comprises nine known splicing factors, CAF1, CAF2, CRS2, CFM3a, RNC1, WTF1, APO2, RH3, and mTERF2, while ECD2‐containing complex II consists of three known splicing factors: CAF1, CAF2, and CFM2 (Figure 9A). In addition, PBF2 recruits other known splicing factors to form two splicing complexes: PBF2‐containing complex I and PBF2‐containing complex II. PBF2‐containing complex I contains three known splicing factors, CAF2, CRS2, and APO1, while PBF2‐containing complex II consists of at least three known splicing factors: CAF1, CAF2, and CFM2 (Figure 9B). Notably, the components of PBF6‐containing complex I and ECD2‐containing complex I are very similar, as are the components of PBF6‐containing complex II, ECD2‐containing complex II, and PBF2‐containing complex II. Taken together, our findings suggest that PPR‐type splicing factors form different splicing complexes with distinct splicing factors.

PPR proteins predominantly recognize and bind to specific sequences in the introns of transcripts via the one repeat–one nucleotide mechanism (Yin et al., 2013; Barkan and Small, 2014). PBF6 binds to specific sequences in ycf3 intron 1 and clpP1 intron 2, and the introns of petB, ndhA, and ndhB transcripts (Figure 6). ECD2 and PBF2 can also bind to their specific cognate sequences in their targeted introns (Wang et al., 2020a; Wang et al., 2021). Therefore, PPR‐type splicing factors such as PBF6, ECD2, and PBF2 may recognize their specific sequences and guide the corresponding splicing complex to a targeted intron, thus promoting intron splicing.

It should be noted that co‐IP assays and affinity purification were performed in a transient expression system. Thus, the physiological stoichiometry of two splicing complexes formed by PBF6, ECD2, or PBF2 remains to be determined.

How does PBF6 cooperate with other PPR‐type splicing factors to promote the splicing of a single intron?

The splicing of more than half of chloroplast group II introns requires the participation of at least two PPR‐type splicing factors (Table S2). For instance, five PPR‐type splicing factors (PDM1, EMB1270, ECD2, AES, and PPR_66) participate in the splicing of the same ndhA intron. The results of this study and previous studies show that the three PPR proteins PBF6, PBF2, and ECD2 are involved in ycf3 intron 1 splicing (Wang et al., 2020a; Wang et al., 2021), raising a question about their cooperation in ycf3 intron 1 splicing.

The results of this study indicate that PBF6, PBF2, and ECD2 each form two splicing complexes that promote the splicing of their respective target introns. Importantly, these splicing complexes are distinct, as the complexes containing PBF6 (or PBF2, or ECD2) do not contain the other two PPR‐type splicing factors (Figure 9C). These findings indicate that the formation of the splicing complexes harboring each PPR protein is independent of the other PPR‐type splicing factors. PBF6, PBF2, and ECD2 specifically bind to distinct sequences in ycf3 intron 1, with the binding sites of PBF6, PBF2, and ECD2 in ycf3 intron 1 being in domains V, II, and I, respectively (Figure S10). Therefore, we propose that PBF6, PBF2, and ECD2 independently guide their respective associated splicing complexes to their specifically cognate sequences in ycf3 intron 1, after which these splicing complexes cooperate to make the intron into a catalytically active conformation and facilitate ycf3 intron 1 splicing. How these splicing complexes coordinate the ycf3 intron 1 splicing and whether they form a super‐large splicing machinery remain to be investigated.

According to our findings, we put forward a model illustrating the role of PBF6 in the splicing of some group II introns in chloroplasts (Figure 10). PBF6 forms two distinct splicing complexes: One consisting of nine known splicing factors and the other containing three known splicing factors. These known splicing factors belong to the CRM, peptidyl‐tRNA hydrolase, RNase III, PORR, APO, DEAD‐box RNA helicase, and mTERF protein families. PBF6 guides each splicing complex to the targeted intron by recognizing the specific sequence in ycf3 intron 1 and clpP1 intron 2, as well as the introns of petB, ndhA, and ndhB to facilitate the splicing of each individual intron (Figure 10A). PBF6 cooperates with two other PPR‐type splicing factors, PBF2 and ECD2, to splice ycf3 intron 1. As with PBF6, PBF2 and ECD2 each recruit various known splicing factors and form two distinct splicing complexes. These three PPR‐type splicing factors recruit the respective splicing complexes to their specific recognition sequences in ycf3 intron 1, thus facilitating intron splicing (Figure 10B). Our findings suggest how a PPR protein forms distinct splicing complexes to splice multiple introns and how several PPR proteins guide their respective splicing complexes to their specific target sequences to splice a single intron. Unraveling the structures of these splicing complexes will help fully elucidate the chloroplast intron‐splicing mechanism in plants.

Figure 10.

Working models for PBF6

(A) Proposed role of PBF6 in enhancing the splicing of ycf3 intron 1 and clpP1 intron 2, and of introns from petB, ndhA, and ndhB transcripts. PBF6 forms splicing complexes with other known splicing factors and guides the corresponding splicing complexes to each of these introns by recognizing specific intronic sequences, thus making the intron in a catalytically active conformation and facilitating the splicing of multiple introns. Boxes, exon 1 and exon 2; black line, intron between exon 1 and exon 2. Different colors represent various known splicing factors from the CRM, peptidyl‐tRNA hydrolase, RNase III, PORR, APO, DEAD‐box RNA helicase, and mTERF families. (B) Proposed role of PBF6 in concert with two other PPR‐type splicing factors ECD2 and PBF2 in promoting splicing of ycf3 intron 1. Association of the splicing complexes containing PBF6, ECD2, or PBF2 in a sequence‐specific binding manner makes ycf3 intron 1 into a catalytically active conformation, leading to splicing. The secondary structure of Arabidopsis ycf3 intron 1 is adapted from the tobacco intron model of Michel et al. (1989). The binding sites of PBF6, ECD2, and PBF2 are highlighted in red.

MATERIALS AND METHODS

Plant materials and growth conditions

Columbia‐0 (Col‐0), an accession of the A. thaliana, was used as the wild‐type (WT) material in the present study. Two T‐DNA insertion mutants, pbf6‐1 (SALK_150731) and pbf6‐2 (SALK_034168), were obtained from the Arabidopsis Biological Resource Center (ABRC). T‐DNA insertion loci were confirmed via genotyping PCR and Sanger sequencing. Primer information used for the analysis of T‐DNA insertion loci is shown in Table S4. After surface sterilization with 70% (v/v) ethanol, seeds were stratified for 2 days in the dark at 4°C and then plated on Murashige and Skoog (MS) medium with 2% (w/v) sucrose added. Seedlings were cultivated in a growth chamber at 22°C and exposed to a 14‐h light/10‐h dark cycle, with light density set to 100 μmol m⁻² s⁻¹. Unless stated otherwise, the plant materials used in the present study were obtained from 12‐d‐old seedlings.

Chlorophyll fluorescence measurements

Chlorophyll fluorescence induction kinetics were determined using a pulse‐amplitude modulated fluorometer (PAM‐2500; Heinz Walz, Germany), following a previously described protocol (Liu et al., 2012; Ding et al., 2019). After the entire plant was dark‐adapted for 20 min, the minimal fluorescence yield (F_o) of the attached leaves was determined with weak red measuring light (630 nm, < 0.2 μmol m⁻² s⁻¹). The maximal fluorescence yield (F_m) was acquired by applying a saturating red pulse (630 nm, 3,300 μmol m⁻² s⁻¹, 0.8 s). Actinic light (630 nm, 100 μmol m⁻² s⁻¹) was subsequently applied for 7–8 min to record steady‐state fluorescence (F_s), followed by a second saturating red pulse to determine F_m′. After turning off the actinic light, a 3 s pulse of far‐red light was applied to measure the minimal fluorescence in the light‐adapted state (F_o′).

77 K fluorescence emission spectra

77 K fluorescence emission spectra were detected from 600 nm to 800 nm with excitation at 436 nm using a fluorescence spectrophotometer (F‐7100; Hitachi, Japan).

P700 absorbance

P700 absorbance in attached intact leaves was measured using a Dual‐PAM‐100 system (Heinz Walz, Germany). Absorbance changes were monitored following illumination with far‐red light at 720 nm. The maximum level of photo‐oxidizable P700 (ΔAmax) was measured by applying a saturating light pulse (720 nm, 5,000 μmol m⁻² s⁻¹).

Measurements of PSI complementary quantum yields Y(I)

The effective quantum yield of PSI [Y(I)] was measured using a Dual‐PAM‐100 system (Heinz Walz, Germany) as described previously (Wang et al., 2017). Leaves were dark‐adapted for 20 min before being exposed to far‐red light (720 nm, 20 μmol m⁻² s⁻¹) for 10 s to record the initial P700 signal (P). A saturating light pulse (720 nm, 10,000 μmol m⁻² s⁻¹, 0.2 s) was delivered to acquire the maximal P700 signal (Pm). Actinic light (625 nm, 100 μmol m⁻² s⁻¹) was then turned on and saturating pulses were imposed at 20 s intervals to measure Pm′. Once the induction curve reached steady state (~6 min), the actinic light was removed. The Y(I) value was calculated using the formula Y(I) = (Pm′ − P)/Pm.

Immunoblot, SDS‐PAGE, BN‐PAGE, and 2D BN‐PAGE/SDS‐PAGE analyses

Total leaf proteins were extracted as described previously and separated by SDS‐PAGE (Ding et al., 2019). Following transfer of proteins onto PVDF membranes (Immobilon‐P, USA), the membranes were probed with specific antibodies, and then secondary antibody signals were detected using an enhanced chemiluminescence system (DW101‐01, TransGen) and chemiluminescence signals were visualized using the Tanon 5200 imaging system. Protein concentrations were measured using a Bio‐Rad protein assay kit. All antibodies used are listed in Table S5. Thylakoid membranes were isolated as described previously (Zhang et al., 2024). BN‐PAGE was performed using 5%–12% (w/v) linear gradient acrylamide gels following Zhang et al. (2024). For 2D BN‐PAGE/SDS‐PAGE, first‐dimension gel strips were incubated in denaturation buffer (100 mM Tris‐HCl, pH 6.8, 2% [w/v] SDS, 15% [v/v] glycerol, and 2.5% [v/v] β‐mercaptoethanol). After 30 min of treatment at 37°C, the gel strips were separated via 16% (w/v) SDS‐PAGE.

Subcellular localization of GFP fusions

For subcellular localization of PBF6, a 561‐bp fragment of the 5′ region of PBF6, starting at the ATG start codon, was amplified from cDNA prepared from total RNA extracted from wild‐type seedlings and inserted into the pBI221‐35S‐GFP vector. Constructs for targeting chloroplast, mitochondrial, and nuclear markers were generated in accordance with our previous work (Zhong et al., 2013). Fusion constructs were transfected into Arabidopsis protoplasts. GFP signals were obtained by confocal laser‐scanning microscopy. The detailed procedures followed the method described by Zhong et al. (2013).

Isolation of chloroplasts, thylakoid membranes, and stroma proteins

Seedlings of pbf6‐2 com (~3 weeks old) were ground in an ice‐cold chloroplast extraction buffer (20 mM HEPES‐KOH, pH 8.0, 10 mM NaHCO₃, 5 mM EDTA‐Na₂, 5 mM EGTA, 330 mM sorbitol, and 5 mM MgCl₂). After filtering the homogenate through three layers of gauze, it was centrifuged at 3,000 g for 5 min at 4°C. The pellet was used as the crude chloroplast fraction. Intact chloroplasts were purified by Percoll density gradient centrifugation (Zhong et al., 2013). The resulting intact chloroplasts were resuspended in hypotonic buffer (5 mM MgCl₂, and 10 mM HEPES‐KOH, pH 8.0), vortexed briefly, and then subjected to incubation on ice for 20 min to rupture chloroplasts. The suspension was centrifuged at 20,000 g for 30 min at 4°C; the supernatant and pellet were collected as the stroma and thylakoid membrane fractions, respectively. Protein concentrations were measured prior to immunoblot analysis.

RNA isolation, RT‐PCR, and RT‐qPCR analyses

Total RNA was extracted from fresh leaves using TRIzol reagent (15596018, Ambion) in accordance with the manufacturer's protocol. For RT‐PCR and RT‐qPCR, first‐strand cDNA synthesis was carried out using 1 μg of total RNA as the template using EasyScript reverse transcriptase (AE301, TransGen) including a mixture of random hexamer and oligo(dT) primers. qPCR was carried out using a LightCycler 96 Real‐Time PCR System (Roche Diagnostics) under the following cycling conditions: 30 s at 95°C, then 40 cycles with 7 s at 95°C, 20 s at 60°C, and 10 s at 72°C. Transcript abundance was normalized to that of ACTIN2 using the Delta Ct method. Primers used for splicing assays and chloroplast transcript analysis are listed in Table S4.

rRNA‐depleted total RNA sequencing

Total RNA was extracted from fresh leaves of 12 d‐old Col‐0 and pbf6‐2 seedlings. Total RNA was assessed for integrity using an Agilent 2100 Bioanalyzer. A Ribo‐Zero kit (RZH1046, Epicentre) was used to remove ribosomal RNA, and the resulting rRNA‐depleted RNA was applied to construct libraries using an NEBNext Single Cell/Low Input Library Prep Kit (E6420, NEB). After pooling the libraries, sequencing was performed on an Illumina NovaSeq. 6000 platform to produce 150‐bp paired‐end reads. Reads with quality scores above 30 and lengths exceeding 60 nt were mapped to the A. thaliana reference genome (TAIR10) using TopHat2 (‐p4‐no‐novel‐juncs). Chloroplast gene transcript abundance was quantified using the ChloroSeq pipeline (Castandet et al., 2016).

RNA gel blot analysis

Fifteen micrograms of total leaf RNA was separated on a 1.2% (w/v) agarose–formaldehyde gel and transferred to a nylon membrane by capillary blotting. Membranes were baked at 80°C for 3 h. Hybridization was performed in a buffer containing 7% (w/v) SDS, 0.5 M Na₂HPO₄ (pH 7.0), and 1 mM EDTA at 65°C for 16 h using [α‐³²P]‐labeled probes. Membranes were washed twice with 1× SSC containing 0.1% (w/v) SDS at 65°C and then exposed to an X‐ray film. The probes were generated by PCR amplification of cDNA fragments and labeled with [α‐³²P]‐dCTP using a Random Primer DNA Labeling Kit (U1100, Promega). The primer sequences for preparing probes are listed in Table S4.

RNA and protein co‐immunoprecipitation assays

Chloroplasts were isolated from 3‐week‐old seedlings of pbf6‐2 com and ground in RNA immunoprecipitation (RIP) buffer (20 mM Tris‐HCl, pH 8.0, 1 mM EDTA, 150 mM NaCl, 0.5% [v/v] Nonidet P‐40, 2 mM DTT, 1× protease inhibitors, and 50 U mL⁻¹ RNase inhibitor). Following centrifugation to remove debris, supernatants were incubated with magnetic beads that were conjugated to anti‐FLAG or IgG antibodies. After washing, beads were treated with proteinase K buffer and incubated at 65°C. RNA was extracted using TRIzol, reverse‐transcribed, and analyzed by qPCR. The primers for qPCR are listed in Table S4.

Protein production, purification, and REMSA

A fragment of the PBF6 coding sequence (PBF6‐10N, residues 32–601) was inserted into the vector pET‐28a and transformed into Transetta (DE3) cells (TransGen Biotech, China). Protein production was induced by 0.2 mM IPTG and incubated at 18°C for 16 h. The recombinant protein was purified via Ni‐NTA affinity chromatography (GenScript, China). Biotin‐labeled RNA probes (Sangon Biotech, China) were incubated with 50 nM purified PBF6‐10N in EMSA binding buffer (10 mM HEPES, pH 7.3, 20 mM KCl, 2 mM MgCl₂, 1 mM DTT, and 5% [v/v] glycerol) at 25°C for 30 min (Sun, He, et al., 2025). The binding reaction was resolved on native 6% (w/v) PAGE, then transferred to membranes, and visualized using a Tanon 5200 imaging system.

Co‐immunoprecipitation assay

co‐IP experiments were carried out using the Arabidopsis protoplast transient expression system (Zhang et al., 2024). The full‐length coding sequence of PBF6 was inserted into pUC19 harboring the sequence for a 3×HA tag. The full‐length coding sequences for each splicing factor gene (CAF1, CAF2, CRS2, CFM2, CFM3a, WTF1, RNC1, APO1, APO2, LEFKO, RH3, mTERF2, and ISE2) were individually cloned into pUC19 carrying the sequence for a 3×FLAG tag. Paired plasmids were co‐introduced into protoplasts, followed by incubation at 22°C under dim light conditions for 16 h. The protoplasts were lysed using lysis buffer, followed by treatment with 50 μg mL⁻¹ RNase (B600476, BBI) or without RNase at 4°C for 20 min. Lysates were incubated with anti‐FLAG magnetic agarose beads (M185‐10; MBL) and washed multiple times using co‐IP buffer (150 mM HEPES, pH 7.5, 1 mM EDTA, 150 mM KCl, 1 mM DTT, 1% [w/v] n‐dodecyl‐β‐D‐maltoside [β‐DM], and 1× protease inhibitors) at 4°C for 30 min. Bound proteins were eluted with the FLAG peptide and subsequently analyzed by SDS‐PAGE, followed by immunoblotting.

Luciferase complementation imaging experiments

The luciferase complementation imaging experiment was conducted following Chen et al. (2008). The coding sequence of PBF6 was inserted into pCAMBIA‐NLuc, while the coding sequence for each splicing factor gene (CAF1, CAF2, CRS2, CFM2, CFM3a, WTF1, RNC1, APO1, APO2, LEFKO, RH3, mTERF2, and ISE2) was individually inserted into pCAMBIA‐CLuc. Each of the resulting constructs was introduced into GV3101 and the appropriate pairs of GV3101 cultures were co‐agroinfiltrated into the leaves of 6‐week‐old N. benthamiana. After 2 d, the collected leaf discs were incubated with 1 mM luciferin for 15 min and then measured for luminescence using an EnSpire multimode plate reader (PerkinElmer, USA).

Yeast two‐hybrid assays

Yeast two‐hybrid assays were carried out using the Matchmaker GAL4 two‐hybrid system. The coding sequences encoding various truncations of PBF6 were inserted into pGBKT7 (carrying the GAL4 DNA‐binding domain [BD]) to generate bait constructs, while the coding sequences of CAF1 and CAF2 were individually inserted into pGADT7 (with the GAL4 activation domain [AD]) to generate prey constructs. Pairs of bait and prey plasmids were introduced into yeast (Saccharomyces cerevisiae) strain Y2HGold. Positive transformants were screened on synthetic defined SD/−Trp/−Leu medium. Positive colonies were resuspended in sterile water and serially diluted 10‐fold (10⁻¹, 10⁻², and 10⁻³). An aliquot of 1 μL from each dilution was spotted onto fresh triple‐deficient SD medium lacking Trp, Leu, and His with 4 mg mL⁻¹ X‐α‐Gal (B8050, Solarbio). Bait–prey protein interactions were assessed based on LacZ activation and positive staining.

Affinity purification of FLAG‐tagged proteins

For affinity purification, the Arabidopsis protoplast transient expression system was used as described for the co‐IP assay. Arabidopsis protoplasts were co‐introduced with appropriate pairs of plasmids and incubated under dim light conditions at 22°C for 16 h. Then, the protoplasts were collected at 100 g for 2 min and resuspended in 0.5 mL of lysis buffer (150 mM HEPES, pH 7.5, 1 mM EDTA, 150 mM KCl, 1 mM DTT, 0.3% [v/v] Triton X‐100, and 1× protease inhibitor cocktail). The mixture was incubated on ice for 20 min and subsequently centrifuged at 20,000 g at 4°C for 10 min. After collecting the supernatant, it was gently mixed with 100 μL of anti‐FLAG beads (SA109005, Smart‐Lifesciences) and incubated in the dark at 4°C for 1 h. Non‐bound extracts were removed by centrifugation at 1,000 g at 4°C for 1 min. The beads were washed three times with wash buffer (150 mM HEPES, pH 7.5, 1 mM EDTA, 150 mM KCl,1 mM DTT, 1% [w/v] β‐DM, and 1× protease inhibitor cocktail), and the bound proteins were eluted with elution buffer including 0.5 mg mL⁻¹ FLAG peptide. The purified FLAG‐tag preparation was used for subsequent analysis.

Complementation of the pbf6 mutant

The full‐length coding sequence of PBF6 ligated with 3×FLAG sequence was inserted into the pCAMBIA1301 vector under the control of 35S (primer sequences are listed in Table S4). The resulting vector was introduced into the pbf6‐1 mutant and pbf6‐2/PBF6 heterozygous plants by Agrobacterium‐mediated transformation. Arabidopsis seeds were plated on MS medium containing 40 μg mL⁻¹ hygromycin to screen transgenic lines with a wild‐type phenotype and a homozygous mutant background.

Accession numbers

Sequence data for the genes mentioned in this study can be obtained from the Arabidopsis information resource (TAIR, www.arabidopsis.org) under the following accession numbers APO1 (At1g64810), APO2 (At5g57930), CAF1 (At2g20020), CAF2 (At1g23400), CFM2 (At3g01370), CFM3a (At3g23070), CRS2 (At5g38290), ECB2 (At1g15510), ECD2 (At5g50280), EMB1270 (At3g18110), ISE2 (At1g70070), LEFKO (At5g62990), mTERF2 (At2g21710), PBF2 (At3g42630), PBF6 (At5g27270), RH3 (At5g26742), RNC1 (At4g37510), RPL15 (At3g25920), RPS1 (At5g30510), RPS2 (Atcg00160), and WTF1 (At4g01037).

CONFLICTS OF INTEREST

The authors declare no conflicts of interest.

AUTHOR CONTRIBUTIONS

M.L. and C.L. conceived the original research plan. M.L., M.S., J.L., M.S., S.L., A.Z., and C.L. designed the experiments and analyzed the data. M.L., M.S., M.S., and S.L. performed the experiments. C.L. conceived the project. M.L., M.S., and C.L. wrote the article. All authors read and approved the final manuscript.

Supporting information

Additional Supporting Information may be found online in the supporting information tab for this article: http://onlinelibrary.wiley.com/doi/10.1111/jipb.70289/suppinfo

Figure S1. Phenotypic characterization and identification of pbf6 mutants

Figure S2. Multiple‐sequence alignment of PBF6 and its homologous proteins

Figure S3. Phylogenetic analysis of PBF6 and its homologous proteins

Figure S4. Relative transcript levels of chloroplast genes in pbf6‐1 and pbf6‐2 seedlings

Figure S5. Average coverage from RNA‐seq reads across the chloroplast genome in WT and pbf6‐2 seedlings

Figure S6. Splicing of chloroplast introns, as analyzed by RT‐PCR

Figure S7. Delineating the PPR domains of PBF6 that interact with CAF1 and CAF2

Figure S8. ECB2 does not co‐purify with 13 known chloroplast splicing factors

Figure S9. Levels of plastid ribosomal proteins in WT, pbf6‐1, pbf6‐2, and pbf6‐2 com seedlings

Figure S10. Binding sites of three PPR proteins PBF6, ECD2, and PBF2 in ycf3 intron 1

Table S1. Splicing factors except for the PPR family required in splicing of ycf3 intron 1, clpP1 intron 2, ndhA, ndhB, and petB introns, respectively, in Arabidopsis

Table S2. List of introns that require the participation of more than one PPR splicing factor in Arabidopsis chloroplasts

Table S3. List of splicing factors required in group II intron splicing in Arabidopsis chloroplasts

Table S4. Primers used in this study

Table S5. Antibodies used in this study

Biographies

Li, M. , Sun, M. , Lv, J. , Song, M. , Lu, S. , Zhang, A. , and Lu, C. (2026). Uncovering the role of the PPR protein PHOTOSYSTEM ONE BIOGENESIS FACTOR6 in splicing chloroplast group II introns. J. Integr. Plant Biol. 68: 3809–3829.

Data availability statement

The raw reads for the RNA‐seq data presented in this study were deposited to the Genome Sequence Archive of the National Genomics Data Center, China National Center for Bioinformation/Beijing Institute of Genomics, Chinese Academy of Sciences (GSA: CRA035209).