Pentatricopeptide repeat proteins in plants: Cellular functions, action mechanisms, and potential applications

Abstract

Pentatricopeptide repeat (PPR) proteins are involved in nearly all aspects of post-transcriptional processing in plant mitochondria and plastids, playing vital roles in plant growth, development, cytoplasmic male sterility restoration, and responses to biotic and abiotic stresses. Over the last three decades, significant advances have been made in understanding the functions of PPR proteins and the primary mechanisms through which they mediate post-transcriptional processing. This review aims to summarize these advancements, highlighting the mechanisms by which PPR proteins facilitate RNA editing, intron splicing, and RNA maturation in the context of organellar gene expression. We also present the latest progress in PPR engineering and discuss its potential as a biotechnological tool. Additionally, we discuss key challenges and questions that remain in PPR research.

PPR proteins play critical roles in post-transcriptional processing in plant mitochondria and plastids. This review focuses on elucidating the mechanisms of PPR proteins in mediating these processes in plant organelles and explores potential applications of engineered PPR proteins as biotechnological tools.

Introduction

Pentatricopeptide repeat proteins (PPRs) belong to a family of proteins characterized by tandem degenerate repeats of approximately 35 amino acid (aa) residues. These proteins are prevalent in land plants (Small and Peeters, 2000; O'Toole et al., 2008; Gutmann et al., 2020). Since their initial identification (Small and Peeters, 2000), PPR proteins have been extensively studied. Most reported PPR proteins target mitochondria and/or plastids (Colcombet et al., 2013), where they participate in post-transcriptional processing events, including RNA C-to-U editing, intron splicing, RNA maturation, stabilization, and either the activation or repression of specific mRNA translation (Barkan and Small, 2014). Loss of function of a PPR protein often leads to dysfunction in mitochondria and/or plastids, thereby affecting plant growth, development, cytoplasmic male sterility (CMS) restoration, and responses to abiotic and biotic stresses.

In this review, we summarize the latest advances in research on PPR proteins. We focus on discussing the mechanisms by which PPR proteins recognize RNA fragments, mediate organelle gene expression, and regulate plant growth, development, and stress responses. We explore potential mechanisms by which PPR proteins are involved in organelle gene expression and address unresolved fundamental questions. Furthermore, we review progress toward the application of artificial PPR (aPPR) proteins and provide an outlook on their use as biotechnological tools.

Roles of PPR proteins in plant growth and development

The diverse functions of PPR proteins in plant growth and development, including, plant growth, seed development, photosynthesis, cytoplasmic male sterility (CMS), and responses to biotic and abiotic stresses, have been substantiated by extensive research over the past 30 years (Barkan and Small, 2014; Meng et al., 2024). Mutations in PPR genes often lead to a range of defective phenotypes in plants, including defective embryogenesis and endosperm development, restricted growth, photosynthetic defects, CMS, aberrant leaf development, changes in leaf pigmentation, hypersensitivity to abiotic stress or abscisic acid (ABA), and enhanced resistance to biotic and abiotic stresses (Figure 3) (Barkan and Small, 2014; Meng et al., 2024). Most mitochondria-localized PPR proteins are associated with defective seed development and plant growth. The loss of function of these PPR proteins generally results in one or several deficient transcripts of mitochondrial genes, disrupting the oxidative phosphorylation respiratory pathway or protein translation, leading to dysfunctional mitochondria. Mutations in PPR genes often lead to embryo lethality in monocotyledons (i.e., Zea mays and Oryza sativa), whereas mutants in dicots such as Arabidopsis thaliana are often viable but exhibit a slow growth phenotype. These defects in seed and/or plant development are likely due to deficits in energy supply. It is also possible that signals from the mitochondria, such as reactive oxygen species (ROS), activate the expression of certain nuclear genes, thereby inhibiting embryogenesis and endosperm development in Z. mays and O. sativa. On the other hand, plastid-localized PPR proteins are primarily associated with photosynthesis. Mutations in these proteins lead to defects in photosynthesis and manifest as albino, pale, or yellow seedlings or leaves, indicating that photosynthesis is dispensable for seed development. However, loss of plastid translation due to mutations in some plastid-localized PPR proteins is embryonically lethal in Arabidopsis, whereas similar mutations result in albino seedlings in cereals (Barkan and Small, 2014). Intriguingly, the loss of plastid translation can result in either embryonic lethality or albino seedling phenotypes, depending on the genetic background in maize (Zhang et al., 2013; Xu et al., 2021). Therefore, the loss of plastid translation may send signals to the nucleus that subsequently activate or inhibit the expression of nuclear genes, arresting embryogenesis in Arabidopsis and some maize inbreds. In addition, plants with mutations in plastid-localized PPR genes exhibit aberrant leaf development.

Figure 3.

Localization and functions of PPR proteins in plant mitochondria and plastids, and their regulation in plant growth and response to stresses.

CMS, which harbors fertile eggs and sterile pollen, is maternally inherited in flowering plants and jointly regulated by cytoplasmic genes and their corresponding restorer of fertility (Rf) genes (Chen and Liu, 2014). The CMS/Rf system represents gynodioecy—the coexistence of females and hermaphrodites in natural populations—and has been widely utilized for hybrid seed production. This system also provides ideal genetic material for studying heterosis and communication between mitochondria and the nucleus (Chen and Liu, 2014). Male sterility factors typically consist of unusual genes or open reading frames (ORFs) encoded by mitochondrial genes, which translate into deleterious mitochondrial proteins that prevent the production of functional pollen (Gaborieau et al., 2016). Multiple PPR genes have been identified as Rf genes in different species, which restore male fertility through various mechanisms, as detailed in Supplemental Table 1. Thus, PPR proteins are crucial for restoring CMS in plants and hold significant value in hybridization breeding.

Accumulating evidence suggests that PPR proteins are important regulators in response to biotic and abiotic stress. These proteins are upregulated or downregulated under conditions such as salinity, osmotic, oxidative, pathogenic, salicylic acid (SA), methyl jasmonate (MeJA), and abscisic acid (ABA) stresses in different species (Chen et al., 2018; Xing et al., 2018; Su et al., 2019; Che et al., 2022; Luo et al., 2022). To date, several PPR proteins have been reported to play critical roles in responding to different abiotic and biotic stresses (Liu et al., 2010, 2016a; Laluk et al., 2011; Murayama et al., 2012; Yuan and Liu, 2012; Tan et al., 2014; Qiu et al., 2021; Xiao et al., 2021; Yang et al., 2022b; Luo et al., 2022; Zu et al., 2023). Thus, PPR proteins likely play important roles in plant responses to biotic and abiotic stresses. ROS are detrimental in plant cells, capable of oxidizing and damaging cellular structures, macromolecules, nucleic acids, proteins, and lipids. Mutations in PPR proteins can stimulate ROS production due to defects in the post-transcriptional processing of mitochondrial or plastid genes. Abiotic and biotic stresses can also induce ROS production. These coupled ROS likely constitute the major reason for sensitivity to abiotic and biotic stresses (Zsigmond et al., 2008; Liu et al., 2010; Laluk et al., 2011; Murayama et al., 2012; Tan et al., 2014; Xiao et al., 2021). However, mutations in two PPR genes, RTP7 and OsNBL3, which decrease intron splicing at nad7 and nad5, respectively, result in reduced complex I activity and enhanced resistance to pathogens and salt stress in Arabidopsis and rice, respectively (Qiu et al., 2021; Yang et al., 2022b). It is speculated that the enhanced resistance of rtp7 to Phytophthora parasitica is associated with an elevated mitochondrial reactive oxygen species (mROS) burst but might not rely on the ROS burst associated with plasma membrane-localized NADPH oxidases (Yang et al., 2022b).

Classification of PPR proteins

PPR proteins belong to the α-solenoid superfamily and are characterized by tandem arrays of PPR motifs, ranging from 2 to 30 repeats (Small and Peeters, 2000; Lurin et al., 2004). Initially defined as a 35-amino acid repeat, the PPR motif folds into a pair of antiparallel α-helices (Ke et al., 2013; Yin et al., 2013; Gully et al., 2015; Yan et al., 2017; Small and Peeters, 2000; Barkan and Small, 2014). Subsequent analyses in Arabidopsis revealed new motifs consisting of 31 aa (PPR-like S for short) and 35–36 aa (PPR-like L for long) (Lurin et al., 2004). Based on motif composition, PPR proteins are classified into P and PLS classes. P-class PPR proteins, composed of bona fide “canonical” P motifs, are evolutionarily ancient and found in almost all eukaryotes (Figure 1A) (Schallenberg-Rüdinger et al., 2014; Gutmann et al., 2020). Several P-class PPR proteins, such as AtNUWA, ZmNUWA, and bCCP1, feature additional domains like a basic region/nuclear localization signal leucine zipper motif (bZIP) domain at the N terminus and a coiled-coil at the C terminus (Andrés-Colás et al., 2017; Guillaumot et al., 2017; Wang et al., 2023b, 2024). GRP23, for example, contains a bZIP domain at the N terminus and a WQQ motif at the C terminus (Ding et al., 2006). ATP4, PPR-SMR1, SOT1, and PPR53 harbor a small MutS-related (SMR) domain at the C terminus (Figure 1A) (Zoschke et al., 2012b, 2016; Wu et al., 2016; Chen et al., 2019). PLS-class proteins, assumed to be more evolutionarily recent and largely restricted to land plants, consist of P, L, and S motifs (O'Toole et al., 2008; Gutmann et al., 2020). Most PLS-class PPR proteins often contain an E, E+, and/or DYW domain at their C termini (Figure 1B), leading to further subdivision into PLS, PPR-E, PPR-E+, and PPR–DYW subclasses (Lurin et al., 2004; Barkan and Small, 2014). Structural modeling has identified 10 different variants of the PPR motif, including P, P1, P2, S1, S2, SS, L1, L2, E1, and E2 (Cheng et al., 2016). The canonical P motifs are exclusively found in P-class PPR proteins, whereas P1 and P2 motifs, which mainly differ in the first helix, are present only in PLS-class PPR proteins (Cheng et al., 2016). S1 and S2 motifs consist of 31 and 32 aa, respectively, and L1 and L2 motifs consist of 35 and 36 aa. The arrangement of PPR motifs in PLS-class PPR proteins typically follows the pattern (P1–L1–S1)_n–P2–L2–S2 (Figure 1B) (Rivals et al., 2006; Cheng et al., 2016). In addition, the S-like motif (SS, 31 aa) is always found juxtaposed with other S motifs, with the first and second helices resembling those of the P1 and S1 motifs, respectively. SS motifs frequently occur between S1 and P1 motifs and can appear singly or as several adjacent SS motifs (Figure 1B). The E domains have been redefined into two PPR-like motifs, E1 and E2, each comprising 34 aa. This redefinition raises questions regarding the contributions of E1 and E2 to RNA binding and base recognition (Cheng et al., 2016). Evidence suggests that E1 and E2 may provide additional selectivity for target recognition beyond the capabilities of traditional PPR motifs (Okuda et al., 2014; Ruwe et al., 2018; Yang et al., 2023a). Furthermore, the E2 domain and the “PG box” of the DYW1-interacting PPR protein are thought to complement the formation of the substrate-binding pocket of the DYW1 protein (Toma-Fukai et al., 2023). The PPR–DYW protein was originally characterized to contain E, E+, and DYW (approximately 106 aa) domains at their C termini (Lurin et al., 2004). Based on crystal results and sequence analysis, the C-terminal DYW domain (approximately 136 aa) starting from the conserved PG box (Hayes et al., 2013) was previously considered part of the E domain, including all the sequence beyond the E2 motif (Figure 1D) (Cheng et al., 2016; Takenaka et al., 2021). This domain provides deaminase activity essential for C-to-U conversion (Oldenkott et al., 2019; Hayes and Santibanez, 2020; Takenaka et al., 2021), and may also exert further selectivity for target recognition (Ichinose and Sugita, 2018; Maeda et al., 2022; Yang et al., 2023a). In addition, the E+ domain is considered a degenerate or truncated DYW domain, supported by E+ domains within numerous PPR-E+ proteins from various species containing a PG box and the N terminus of the gating domain of the DYW deaminase (Takenaka et al., 2021; Wang et al., 2023b). Some atypical PPR–DYW proteins consist of only a few low-conservative PPR motifs and a C-terminal DYW domain (i.e., PCW1 and MEF8/8S) or a truncated DYW domain (i.e., AtDYW2 and ZmDYW2A/2B) but lack the E1 and E2 domains (Figure 1C) (Andrés-Colás et al., 2017; Guillaumot et al., 2017; Wang et al., 2023b).

Figure 1.

Architecture of PPR proteins in plants.

(A–C) Schematic structures of P-class PPR proteins (A), PLS-class PPR proteins (B), and atypical PPR–DYW proteins (C). E+ represents a degenerate or truncated DYW domain. Truncated DYW indicates the domain loss in the N-terminal region, including the PG box and the N terminus of the gating domain.

(D) Schematic comparison of the DYW and DYW:KP domains. “PG” in the DYW domain indicates the conserved PG box (Hayes et al., 2013), while HSE and CxDC indicate the zinc-binding deaminase signature sequence (Salone et al., 2007; Hayes et al., 2013, 2015; Boussardon et al., 2014; Wagoner et al., 2015). DYW are the C-terminal residues Asp-Tyr-Trp. “KP” in the DYW:KP domain represents the conserved Lys-Pro dipeptide. x represents aa with lower conservation (Gutmann et al., 2020).

In addition, the DYW domain includes a variant known as DYW:KP. Unlike the canonical DYW consensus, which contains a functionally important PG-box at the start of the DYW domain (Figure 1D), the DYW:KP is distinguished by a conserved Lys-Pro dipeptide at the beginning of this domain (Figure 1D). The DYW:KP variant exhibits a limited phylogenetic distribution, being found only in hornworts, lycophytes, and monilophytes (Gutmann et al., 2020), and within the latter two groups, its presence is inconsistent, with some families lacking the variant altogether (Gutmann et al., 2020). Research has shown that the DYW:KP domain is associated exclusively with U-to-C editing clades and demonstrates a quantitative correlation with U-to-C editing within those clades (Kugita, 2003; Grewe et al., 2011; Oldenkott et al., 2014; Knie et al., 2016; Gutmann et al., 2020). Based on its phylogenetic distribution and sequence characteristics, the DYW:KP variant is speculated to be a candidate for RNA-editing factors that catalyze U-to-C conversion (Gutmann et al., 2020). This hypothesis is further supported by experiments demonstrating that a designer PPR (dPPR)–DYW:KP protein can target and convert uridine to cytidine in bacterial and human cells (Ichinose et al., 2022).

Distribution of PPR proteins

The PPR protein family is one of the largest gene families in plants. By contrast, the genomes of non-plant organisms harbor very few PPR genes: only six, five, and two predicted PPR-encoding genes in the genomes of Homo sapiens, Drosophila melanogaster, and Saccharomyces cerevisiae, respectively, with no PPR genes found in prokaryotes (Pusnik et al., 2007). The parasitic protozoan Trypanosoma brucei contains 28 PPR genes, the highest number among non-plant organisms (Pusnik et al., 2007). The number of PPR genes varies widely across plants, ranging from hundreds to thousands. Algae, by contrast, harbor only a few dozen PPR genes, especially PLS-class PPR genes (Gutmann et al., 2020). Most red and green algae species lack PLS-class PPR genes, with only a few species containing one or two. However, numerous putative PPR-editing factors (PLS-class PPRs) account for nearly 10% of the expressed protein-coding genes in some non-seed plant groups, notably hornworts, some lycophytes, and most monilophytes (Gutmann et al., 2020). This phenomenon indicates a significant expansion of PPR genes during the evolution of land plants (O'Toole et al., 2008; Gutmann et al., 2020). The number of PLS-class PPR proteins varies greatly across plants and correlates with the number of RNA-editing sites in organelles, potentially driving these dramatic changes in their numbers—from none or only a few to thousands within the same genus of non-seed plants. In contrast, the distribution of P-class PPR proteins is relatively conserved, with small numbers in red and green algae and a steady increase observed from the earliest to the most recent terrestrial plant clades (Gutmann et al., 2020). Gene structure and evolutionary relationships of PPR proteins in Physcomitrella patens, A. thaliana, and O. Sativa suggest that in seed plants, one or more waves of retrotransposition may be responsible for the expansion of PPR genes, rather than genome duplication (O'Toole et al., 2008). In addition, several restorer loci in CMS lines contain a cluster of PPR genes that share a high level of sequence homology, suggesting a pattern of PPR evolution through local sequence duplication at specific loci (Gaborieau et al., 2016).

Two-thirds of PPR proteins in Arabidopsis were initially predicted to target either mitochondria or chloroplasts (Small and Peeters, 2000). This prediction was confirmed by a systematic study on the subcellular localization of PPR proteins in Arabidopsis (Colcombet et al., 2013). To date, the localization of PPR proteins has been predicted in multiple species, with most targeting mitochondria or plastids (Lurin et al., 2004; Liu et al., 2016b; Wei and Han, 2016; Chen et al., 2018; Li et al., 2018; Subburaj et al., 2020; Che et al., 2022). Up to now, over 200 PPR proteins have been characterized across various species, nearly all of which are specifically localized to mitochondria, plastids, or both (Supplemental Table 1. Characterized CMS/PPR Rf gene systems in plants, Supplemental Table 2. PPR proteins in RNA editing in plants, Supplemental Table 3. PPR proteins in intron splicing in plants, Supplemental Table 4. PPR proteins in RNA maturation, stabilization, and translation in plants). This localization is consistent with the primary role of PPR proteins in the post-transcriptional processing of organellar genes. In addition, a few PPR proteins are also found to localize to the nucleus. For example, the P-class PPR protein bCCP1, which is involved in RNA editing at numerous sites, is dual-localized to both the mitochondria and nucleus in maize, though its nuclear function remains unknown (Wang et al., 2023b). Another P-class PPR protein, GRP23, is triple-targeted to mitochondria, plastids, and nuclei in Arabidopsis (Yang et al., 2022c), where it likely functions as a transcriptional regulator through interactions with RNA polymerase II (Ding et al., 2006). In rice, OsNPPR1 and FLO14 have been reported to localize to the nucleus (Hao et al., 2019; Xue et al., 2019). As the post-transcriptional processing of nuclear genes is also mediated by numerous factors, PPR proteins with RNA-binding characteristics may serve as mediators in this process.

Crystal structure of PPR proteins and “PPR codes”

The PPR motif is predicted to form antiparallel α helices, with tandem PPR motifs assembling into a superhelix that encloses a groove or tunnel (Figure 2A–2C) (Small and Peeters, 2000). This configuration has been confirmed by the crystal and solution structures of natural PPR proteins, such as PPR10, THA8, and THA8L, and artificial PPR proteins, aPPRs (Ban et al., 2013; Ke et al., 2013; Yin et al., 2013; Li et al., 2014; Gully et al., 2015; Shen et al., 2016; Yan et al., 2017). The Z. mays PPR10 specifically targets chloroplasts, binding two similar RNA sequences in the intergenic regions between Atp1-AtpH and PsaJ-Rpl33, referred to as AtpH and PsaJ, respectively. This binding protects the target RNA from exoribonuclease degradation and activates mRNA translation by remodeling the ribosomal binding sites (Pfalz et al., 2009). The crystal structure of an RNA-free PPR10 fragment (residues 61–786) with quadruple cysteine mutations (C256S/C279S/C430S/C449S) reveals that the 19 PPR10 repeats form a right-handed superhelical spiral, and the two PPR10 molecules intertwine in an antiparallel manner (Yin et al., 2013). This configuration results in an antiparallel, intertwined homodimer that exhibits considerable conformational changes upon binding to 18 nt PsaJ (Yin et al., 2013). Subsequent research reveals that the antiparallel homodimer formation of PsaJ-PPR10 is likely due to the over-short N-terminus of the PPR10 fragment combined with the cysteine mutations (Li et al., 2014). An RNA-free PPR10 with a longer N terminus and no mutations exists in a dimeric state, whereas atpH-PPR10 and PsaJ-PPR10 exist as monomers (Li et al., 2014; Gully et al., 2015). THA8, a small PPR protein with only five P motifs, binds within the ycf3 intron. In the absence of RNA, the apostructure of THA8 is a monomer. RNA binding induces the assembly of two THA8 monomers into an asymmetric dimer. In the complex structure of RNA-THA8, the RNA is bound at the dimeric interface, formed by the C-terminal part of one THA8 monomer and the N-terminal part of another. Further analysis shows that a conserved G nucleotide of the bound RNAs makes extensive contact with both THA8 monomers (Ke et al., 2013). PLS-class PPR proteins and the multiple organellar RNA-editing factor (MORF) are critical editing factors. MORF9 significantly enhances the RNA-binding activity of the designer PLS–PPR protein (PLS)₃PPR. The crystal structure of (PLS)₃PPR and the (PLS)₃PPR–MORF9 complex reveal that (PLS)₃PPR undergoes a notable conformational change upon MORF9 binding, significantly increasing its RNA-binding activity. This conformational change mainly results from pronounced conformational change of L motifs. The enhanced RNA-binding activity of PLS–PPRs mediated by MORF9 has also been observed in natural PLS-class PPR proteins (Ban et al., 2013).

Figure 2.

Structure of the PPR motifs and alignment of PPR protein on RNA target.

(A) Predicted models of P motifs with conserved residues as described by Cheng et al. (2016), using https://swissmodel.expasy.org/interactive. The fifth and the last (35th) aa residues in the P motif are marked by green balls. The predicted distance between the fifth and the last aa residues is 8.58 Å. N, N terminus; C, C terminus.

(B and C) Predicted models of a tract of 11 P motifs using https://swissmodel.expasy.org/interactive.

(D) Residues at the fifth position (black) in the front helix and the last position (gray) in the loop connecting adjacent helices specify which nucleotide is bound (dotted lines). The PPR codes are as described by Barkan et al. (2014) and Yan et al. (2019).

Tandem PPR motifs in PPR proteins specifically recognize RNA sequences through two key combinatorial residues (residues 6 and 1′ according to Barkan et al., 2012; residues 5 and 35 according to Yin et al., 2013; and residues 4 and ii according to Yagi et al., 2013) within the PPR motifs. Barkan et al. used computational methods to deduce a nucleotide recognition code involving these combinatorial residues at residue 6 of one PPR repeat and residue 1′ of the next (Barkan et al., 2012). This model was validated by engineering a PPR protein to bind a novel RNA sequence (Barkan et al., 2012). The crystal structure of PsaJ-PPR10 reveals that nucleotide recognition primarily involves the combinatorial residues at positions 5 and 35. A polar aa at residue 5, which forms direct hydrogen bonds with its recognized nucleotide in each repeat, appears to be the most critical determinant of RNA base specificity. In addition, it was proposed that H₂O molecules may mediate hydrogen bonding between residue 35, the second major determinant, and the nucleotide base (Yin et al., 2013). This hypothesis was confirmed by the crystal structure of RNA-bound dPPRs (Shen et al., 2016), which also indicated that the polar aa at the fifth position in each repeat is the primary determinant of RNA base specificity (Shen et al., 2016). Due to the variation in the number of residues in PPR motifs, the code involving two combinatorial residues—at the fifth residue of the first helix and the last residue of the loop connecting adjacent motifs within the same PPR motif—may offer practical advantages (Cheng et al., 2016). This code has been widely used for predicting nucleotide recognition by PPR proteins (Yan et al., 2019). The model of PPR protein alignment on an RNA target is shown in Figure 2D.

Molecular functions of PPR proteins

To date, over 200 PPR proteins have been characterized, the majority of which are involved in the post-transcriptional processing of mitochondrial and plastid transcripts. These processes include RNA C-to-U editing, intron splicing, RNA maturation, and stabilization (Figure 3; Supplemental Table 1. Characterized CMS/PPR Rf gene systems in plants, Supplemental Table 2. PPR proteins in RNA editing in plants, Supplemental Table 3. PPR proteins in intron splicing in plants, Supplemental Table 4. PPR proteins in RNA maturation, stabilization, and translation in plants). Additionally, some PPR proteins either activate or repress the translation of organelle mRNAs or function as components of ribosomes in mitochondria and plastids (Figure 3). These diverse functions underscore the critical role of PPR proteins in regulating gene expression within organelles.

RNA C-to-U editing

RNA C-to-U editing is a post-transcriptional modification process that converts specific cytidine residues to uridine (C-to-U) in both mitochondrial and plastidial transcripts of flowering plants. This editing alters the coding sequence of DNA at the RNA level, often restoring nucleotide codes for conserved aa residues in proteins (Edera et al., 2018; Brenner et al., 2019), generating translation start or stop codons (Kotera et al., 2005), facilitating tRNA maturation (Fey et al., 2002), or maintaining the structural conformation required for intron splicing (Xu et al., 2020). Editing events are prevalent in organelles from lower to higher plant species (Giegé and Brennicke, 2001; Notsu et al., 2002; Kugita, 2003; Wolf et al., 2004; Mower and Palmer, 2006; Hecht et al., 2011; Bentolila et al., 2013; Oldenkott et al., 2014; Maas et al., 2015; Wang et al., 2019; Shen et al., 2024), and the number of editing sites varies dramatically across species. C-to-U editing is primarily mediated by the PLS-class PPR proteins (Supplemental Table 2) (Small et al., 2020). Notably, PLS-class PPR proteins have expanded significantly in land plants, correlating with an increased number of RNA-editing sites in these species (Giegé and Brennicke, 2001; Notsu et al., 2002; Kugita, 2003; Wolf et al., 2004; Mower and Palmer, 2006; Rudinger et al., 2009; Hecht et al., 2011; Bentolila et al., 2013; Oldenkott et al., 2014; Maas et al., 2015; Wang et al., 2019; Gutmann et al., 2020; Shen et al., 2024). This correlation suggests a co-evolution between the presence of editing sites and the expansion of PLS-class PPR proteins.

The PPR tract of PLS-class PPR proteins forms a binding surface that recognizes and binds to sequences of single-stranded RNA. Current models and experimental evidence suggest that the terminal S motif, typically the S2 motif, contacts the nucleotide at the −4 position upstream of the editing site, allowing the DYW domain to specifically catalyze the deamination of target Cs (Barkan et al., 2012; Takenaka et al., 2013; Yagi et al., 2013; Cheng et al., 2016). MORF9 enhances the RNA-binding activity of PLS-class PPR proteins by binding to them and altering the conformation of the L motifs, positioning the last residue closer to the fifth residue at L motifs (Yan et al., 2017). However, this binding raises questions about how a PPR protein dissociates from the target RNA once editing is completed. High-affinity binding of PPR to RNA can block RNA translation, a topic that will be discussed later.

The DYW domain at the C terminus of PPR–DYW proteins contains the conserved CDAs-like zinc-binding signature residues (HxE(x)nCxxC), which are essential for deaminases (Salone et al., 2007; Boussardon et al., 2014). The deaminase activity of the DYW domain was recently confirmed. In Escherichia coli, PPR–DYW proteins PpPPR56 and PpPPR65 from P. patens efficiently perform C-to-U editing at targeted sites, while mutations in the DYW domain abolish this activity, thus validating the deaminase function of the DYW domain (Oldenkott et al., 2019; Hayes and Santibanez, 2020). Recent structural analyses have resolved the crystal structures of both the inactive and active states of the DYW domain (Takenaka et al., 2021; Toma-Fukai et al., 2023). These structures and functional data suggest that the DYW domain of OTP86 transitions between an inactive ground state and an activated state. A conserved gating domain at the N terminus of the DYW domain regulates the active site both sterically and mechanistically (Takenaka et al., 2021). Amino acid replacements in the gating domain either abolish or dramatically decrease the deaminase activity of the DYW domain (Takenaka et al., 2021). The difference between typical and atypical DYW domains became apparent when the crystal structure of DYW1 was analyzed (Toma-Fukai et al., 2023). Unlike typical DYW domains, DYW1 lacks any PPR motifs and harbors only an N-terminally truncated DYW domain (Boussardon et al., 2012). A notable structural difference is the presence of an α-helical fold in DYW1, as opposed to the β-finger observed in the gating domain of OTP86. Moreover, the substrate-binding pocket in DYW1 is incompletely formed and is complemented by the E2 domain and the PG box of the interacting PPR protein (Toma-Fukai et al., 2023).

The PLS tract of PPR–DYW proteins typically binds to the RNA sequence upstream of the editing sites, and its DYW domain catalyzes the editing process (Figure 4A and 4B) (Oldenkott et al., 2019; Hayes and Santibanez, 2020; Small et al., 2020). However, over half of the PLS-class PPR proteins in plants lack the DYW domain (Lurin et al., 2004; Cheng et al., 2016; Wei and Han, 2016), and approximately half of the known PPR-editing factors do not belong to the PPR–DYW subclass (Supplemental Table 2). Genetic and biochemical experiments suggest that the absence of deaminase domains in PPR-E and PPR-E+ proteins can be compensated for in trans by atypical PPR–DYW proteins (Figure 4A and 4B). For example, the atypical PPR-E+ protein CRR4, which comprises 14 PPR motifs and E and partially E+ domains, interacts directly with DYW1, which only contains a truncated DYW domain lacking the PG box, to carry out the editing at the ndhD-1 site in Arabidopsis (Kotera et al., 2005; Boussardon et al., 2012). A CRR4-DYW1 fusion can restore the editing at the ndhD-1 site in the crr4::dyw1 double mutant (Figure 4B; Boussardon et al., 2012), providing convincing evidence that deaminase activity can be provided in trans by DYW domain-containing proteins for PPR-E+ proteins. Two research groups independently discovered that the known PPR-E+ editing factors SLO2 in mitochondria and CLB19 in plastids recruit the atypical PPR–DYW protein AtDYW2 (Figure 1C) as the deaminase to carry out editing at specific sites (Chateigner-Boutin et al., 2008; Zhu et al., 2012; Andrés-Colás et al., 2017; Guillaumot et al., 2017; Takenaka et al., 2021; Wang et al., 2023b).

Figure 4.

Models of PPR proteins in RNA C-to-U editing in plant mitochondria and plastids.

The potential mechanisms by which PPR-E, PPR-E+, PPR–DYW, and P-class PPR proteins are involved in RNA C-to-U editing within mitochondria and plastids.

(A) In mitochondria: P: P-class PPR proteins (AtNUWA, ZmNUWA, bCCP1, GRP23, and ZmGRP23). M: MORF1 and/or MORF8.

(B) In plastids: P-PPR: P-class PPR proteins (AtNUWA and ZmNUWA). MORF: MORF2, MORF8, and/or MORF9.

Moreover, it was found that the unique P-class PPR protein AtNUWA bridges and enhances the recruitment of PPR-E+ proteins to AtDYW2 in Arabidopsis mitochondria and plastids (Figure 4A and 4B) (Andrés-Colás et al., 2017; Guillaumot et al., 2017). MORF/RIP proteins, which serve as editing factors, are required for the editing of nearly all sites in both mitochondrial and plastid genomes (Bentolila et al., 2012, 2013; Takenaka et al., 2012). MORFs can form homo- and heterodimers and selectively interact with PPR-E, PPR-E+, and PPR–DYW proteins (Bentolila et al., 2012; Takenaka et al., 2012; Zehrmann et al., 2015; Bayer-Csaszar et al., 2017). In Arabidopsis, MORF1 pulled down AtNUWA and AtDYW2 (Bayer-Csaszar et al., 2017), suggesting that MORF proteins may be components of PPR-E+ editosomes. A similar mechanism operates in maize, where PPR-E+ proteins specifically recruit ZmDYW2A/2B as the deaminase for C-to-U editing, and ZmNUWA assists in this recruitment (Figure 4A and 4B). ZmMORF1/8 may also support the recruitment of PPR-E+ to ZmDYW2A/2B in maize mitochondria (Figure 4A), and the PPR-E+ editing complex in plastids might include MORFs, ORRM1, and OZ1, in addition to PPR-E+, ZmDYW2A/2B, and ZmNUWA (Figure 4B) (Wang et al., 2024). Atypical PPR–DYW proteins PCW1 and MEF8/8S, which are deaminases recruited by PPR-E proteins in maize and Arabidopsis, respectively, and the P-class PPR proteins bCCP1 and GRP23 assist these recruitments in maize and Arabidopsis mitochondria (Figure 4A) (Yang et al., 2022c; Wang et al., 2023b). Moreover, MORF proteins also bridge and enhance the recruitment of atypical PPR–DYW deaminases by PPR-E proteins (Figure 4A) (Yang et al., 2022c; Wang et al., 2023b). The specific recruitment of PCW1 and MEF8/8S by PPR-E proteins, and DYW2 by PPR-E+ proteins, is dictated by their structural requirements. PCW1 and MEF8/8S contain the complete domains of the DYW deaminase (Figure 1C), whereas PPR-E proteins do not, together forming a complete substrate binding and catalysis unit. In contrast, DYW2 possesses only a degenerated N terminus of the gating domain (Figure 1C). An analysis of the E+ domains of 5107 PPR-E+ proteins from various species shows a strictly conserved N terminus of the gating domain. The missing N terminus of the gating domain in DYW2 could be provided in trans by the E+ domain in PPR-E+ proteins (Takenaka et al., 2021; Wang et al., 2023b). Supporting evidence for this structural requirement was found in the CRR4-DYW1 complex, where the E2 domain and the PG box of CRR4 complement the substrate-binding pocket in DYW1 deaminase (Toma-Fukai et al., 2023). PPR-E protein DEK56 facilitates RNA editing by recruiting PCW1, a process supported by ZmMORFs and ZmGRP23 in maize (Zang et al., 2024).

All of the atypical PPR–DYW proteins (AtDYW2, ZmDYW2A/2B, PCW1, and MEF8/8S) and P-class PPR proteins (AtNUWA, ZmNUWA, bCCP1, and GRP23) are involved in RNA editing at numerous sites within plant organelles (Andrés-Colás et al., 2017; Diaz et al., 2017; Guillaumot et al., 2017; Yang et al., 2022c; Wang et al., 2023b, 2024). The proposed mechanism of PPR proteins in RNA editing is illustrated in Figure 4A and 4B. (1) The PPR motif array of a PPR–DYW protein specifically binds to the RNA sequence upstream of the editing sites, and its DYW domain catalyzes the editing at these sites in mitochondria and plastids (Figure 4A and 4B). (2) The PPR tract within the PPR-E+ protein recognizes the target Cs and recruits the atypical PPR–DYW protein DYW2 to catalyze the C-to-U editing. NUWA and MORFs assist this recruitment in both mitochondria and plastids (Figure 4A and 4B). (3) The PPR tract of a PPR-E protein specifies the editing site and recruits the atypical PPR–DYW protein PCW1 or MEF8/8S to catalyze the C-to-U editing. P-class PPR proteins (bCCP1 or GRP23) and MORFs support the recruitment of deaminase in mitochondria (Figure 4A).

Only three atypical PPR–DYW proteins—PCW1, ZmDYW2A, and ZmDYW2B—have been identified in maize, and they are involved in editing at approximately 200 sites (Wang et al., 2023b, 2024). At some of these sites, the editing activity is reduced rather than completely abolished (Wang et al., 2023b, 2024). Moreover, ZmDYW2A and ZmDYW2B do not edit some of the sites targeted by PPR-E+ proteins in maize (Wang et al., 2024). Given that there are over 600 editing sites in maize mitochondria, additional deaminases may be responsible for these editing events. Although typical PPR–DYW proteins are reported to be involved in editing at many sites (Wang et al., 2019; Liu et al., 2020; Yang et al., 2022a), this does not exclude the possibility that PPR-E/E+ proteins may also recruit other deaminases.

Intron splicing

Canonical group II introns, characterized by their highly complex structures, are widely present in the genes of bacteria and yeast, where they exhibit self-splicing activity (Lehmann and Schmidt, 2003; Lambowitz and Zimmerly, 2004, 2011; Lambowitz et al., 2015). In contrast, plant mitochondria and plastid genomes encode group II introns that have lost this self-splicing capability and thus require the assistance of numerous nucleus-encoded factors for splicing (Brown et al., 2014). A variety of PPR proteins, predominantly from the P-class, are involved in the intron splicing of mitochondrial and plastid transcripts (Supplemental Table 3). However, the mechanism by which PPR proteins mediate intron splicing remains elusive. RNA-binding protein immunoprecipitation (RIP-chip) and RNA EMSA (REMSA) assays have demonstrated that several PPR proteins specifically bind to introns involved in their splicing activities (Supplemental Table 3) (Aryamanesh et al., 2017; Liu et al., 2021; Zhang et al., 2021). It is proposed that most of these PPRs assist intron splicing by occupying RNA fragments that block the formation of proper intron folding. P-class PPR proteins PPR4 and EMB2654 are essential for the trans-splicing at rps12 intron1, binding to rps12 intron1a and intron1b, respectively (Schmitz-Linneweber et al., 2006; Lee et al., 2019). These two PPRs are required to juxtapose the two intron halves and maintain RNA structures competent for the efficient splicing of intron1 within rps12 pre-RNA (Lee et al., 2019). In addition, PPR protein binding at termini can protect precursor transcripts containing introns from degradation by endonucleases and/or exonucleases (Figure 5A). For instance, the P-class PPR protein EMS1, which binds to the 3′ end of the nad2 exons 1–2 precursor transcript, is essential for both the production and stabilization of its mature form, corresponding to the 5′ half of nad2 trans-intron2 (Wang et al., 2023a).

Figure 5.

Models of PPR proteins mediating the trimming and stabilization of the 5′ and 3′ terminus or cleavage of the intercistronic region between two ORFs.

(A) The PPR protein likely acts as a barrier when bound to the RNA sequence at the 5′ or 3′ terminus or the trans-intron, blocking degradation by exonuclease.

(B) PPR protein binding to the RNA sequence unmasks the cleavage site of endonuclease in the intercistronic region between two ORFs, and the PPR proteins binding at the 5′ and 3′ termini of the resulting cleaved bicistronic mRNA block degradation by exonuclease.

(C) PPR protein binding to the RNA sequence unmasks the cleavage site of endonuclease within the 5′ terminus of the pre-RNA, mediating the trimming of the 5′ terminus.

Although most reported PPR proteins are required for the splicing of only a few introns (Supplemental Table 3), two unique PPR proteins, PPR-SMR1 (with 12 PPR motifs and a C-terminal SMR domain) and small PPR protein 2 (SPR2, with only four PPR motifs), mediate the splicing of multiple mitochondrial introns in Z. mays. PPR-SMR1 is essential for the splicing of 16 introns (nad1 intron1–4; nad2 intron1–4; nad4 intron1–3; nad5 intron1, 3, and 4; nad7 intron2; and rps3), while SPR2 is crucial for splicing 15 introns (nad1 intron1–4; nad2 intron1–4; nad4 intron1 and 3; nad5 intron1, 2, and 4; and nad7 intron1 and 2) of the 22 introns in maize mitochondrial transcripts, respectively. Multiple lines of evidence indicate that PPR-SMR1 interacts with SPR2, and both proteins interact with other P-class PPR proteins such as Zm-mCSF1 (a CRM/CRS1-YhbY domain protein) and ZmRH48 (DEAD-box RNA helicase 48) (Chen et al., 2019; Wang et al., 2020a; Cao et al., 2022; Yang et al., 2023b). In addition, several other P-class proteins have been found to interact with CRM domain-containing proteins, RNA helicases, and/or RAD52-like proteins in mitochondria (Wang et al., 2020a; Fan et al., 2021). Similarly, various PPR proteins have been reported to interact with other splicing factors, such as CRM domain proteins and/or WTF1 to mediate the splicing of specific introns in plastids (Wang et al., 2020b; Zhang et al., 2021, 2022; Huo et al., 2024). THA8, a small P-class PPR protein with only four PPR motifs, is required for the splicing of ycf3 intron2 and trnA. The crystal structure suggests that THA8 likely forms oligomers, which may organize and condense the ycf3 intron for splicing due to its multiple THA8-binding sites and the G nucleotide, as described above (Ke et al., 2013). In addition, THA8 forms large complexes with ycf3-intron2 and trnA-intron and co-immunoprecipitates with known trnA-splicing factors WTF1 and RNC1 (Khrouchtchova et al., 2012). Several studies have detected intron-splicing factors in large ribonucleoprotein complexes alongside known splicing factors, which are required for the splicing of identical introns in chloroplast stroma (Jenkins and Barkan, 2001; Asakura and Barkan, 2007; Hammani and Barkan, 2014). Additionally, interactions among CRM domain proteins, PORR domain proteins, RNA helicases, WTF1, and RNC1 have been observed in mitochondria and plastids (Germain et al., 2013; Brown et al., 2014). These findings suggest that interactions among PPR proteins and other splicing factors likely help introns form and maintain a splicing-competent structure, facilitating splicing in a manner similar to the canonical group II introns found in bacteria and yeast.

RNA maturation and stabilization

Many studies have reported that PPR proteins, especially P-class PPRs, are essential for the processing and stabilization of the 5′ or 3′ termini of transcription units as well as termini resulting from processing between ORFs in bicistronic or polycistronic transcripts in chloroplasts and mitochondria (Supplemental Table 4). PPR proteins mediate processing and stabilization through multiple mechanisms, as reviewed by Barkan and Small (2014). However, all these mechanisms are based on the characteristic ability of PPR proteins to specifically recognize and tightly bind to target RNA sequences. When a PPR protein binds tightly to a target RNA, it likely acts as a roadblock to hinder exoribonuclease activity (Figure 5A and 5B). This roadblock mechanism is supported by the identification of many small chloroplast RNAs that display PPR footprints, as identified through transcriptome data mining in species such as Chlamydomonas, Arabidopsis, maize, rice, and barley (Ruwe and Schmitz-Linneweber, 2012; Zhelyazkova et al., 2012; Loizeau et al., 2014). These small RNAs are typically mapped to approximately 20 nucleotides at the 5′ termini of almost all processed mRNAs and at many 3′ termini, indicating that PPR proteins or PPR-like proteins act as barriers to protect these termini from exoribonucleolytic degradation. This roadblock mechanism is further supported by multiple experimental findings. For example, in the green alga Chlamydomonas reinhardtii, the P-class PPR protein MCA1 protects the entire transcript from 5′→3′ degradation by specifically acting on the first 21 nucleotides of the petA 5′ UTR in chloroplasts (Loiselay et al., 2008). PPR10, one of the best-studied PPR genes, encodes a plastid-localized P-class PPR protein that binds to similar sequences within the atpI-atpH and psaJ-rpl33 intercistronic regions, thereby protecting these target RNAs from 5′ or 3′ exonucleases and resulting in processed RNAs with PPR10 bound at the 5′ or 3′ terminus (Pfalz et al., 2009; Prikryl et al., 2011). Similarly, the P-class PPR protein HCF152 is required for the accumulation of processed RNA at the 5′ or 3′ termini in the psbH-petB intergenic regions in Arabidopsis chloroplasts (Meierhoff et al., 2003). Further analysis revealed that HCF152, acting analogously to PPR10, binds to the psbH-petB intergenic region, impeding 5′ and 3′ exonucleases, and then defines the position of the processed RNA termini in this region (Pfalz et al., 2009; Zhelyazkova et al., 2012). Several other PPR proteins have also been reported to stabilize specific mRNA termini in plant chloroplasts through this roadblock mechanism (Supplemental Table 4).

PPR proteins likely employ a similar roadblock mechanism to define the positions of the 3′ termini of processed RNA and stabilize mRNA in mitochondria. Genetic and biochemical analyses have shown that the mitochondrion-localized PPR protein MTSF1 is involved in the 3′-terminal processing and stabilization of nad4 mRNA by binding with high affinity to the last 20 nucleotides of the nad4 mRNA. This protective function of MTSF1 on nad4 mRNA is likely conserved across dicots, as the bound RNA sequence remains strictly conserved across species (Haili et al., 2013). It has also been reported that mitochondria-localized PPR proteins MTSF2, MTSF3, MTSF4, EMS1, and MSP1 bind to their target RNA sequences, thereby protecting the 3′ termini of mature or precursor mitochondrial transcripts (Wang et al., 2017, 2022, 2023a; Best et al., 2023; Jung et al., 2023) (Supplemental Table 4). The evidence strongly supports the notion that PPR protein binding to RNA termini forms barriers that hinder the progression of exoribonucleases in plant mitochondria (Figure 5A). This roadblock mechanism is likely further corroborated by findings that short RNA-binding protein (RBP) footprints coincide with transcript 3′ ends but are largely absent from 5′ ends in Arabidopsis mitochondria (Ruwe et al., 2016). EMS1, another mitochondria-localized PPR protein, is essential for the production and stabilization of the mature form of the nad2 exon 1–2 precursor transcripts. The accumulation of an extended form, rather than a truncated form, of the nad2 exon 1–2 precursor transcripts in the ems1 mutant compared to the wild type suggests that the formation of the 3′ ends of the processed mitochondrial RNA may require the interplay of endonucleolytic and exonucleolytic processing mediated by PPR proteins (Wang et al., 2023a).

Unlike the 3′ ends, the lengths of 5′ UTRs vary significantly within mature mitochondrial transcripts, and no 5′→3′ exonuclease involved in post-transcriptional processing has been identified in mitochondria (Binder et al., 2016). In addition, as mentioned earlier, short RBP footprints are largely absent from 5′ ends in Arabidopsis mitochondria (Ruwe et al., 2016). However, multiple PPR proteins have been reported to shape the mature 5′ termini of mitochondrial transcripts. For example, Rf-like PPR proteins RPF1, RPF2, RPF3, RPF4, RPF6, and RPF8 act on the 5′ UTRs of nad4 and nad4L-atp4, nad9 and cox3, ccmC, ccmB, ccmC, and nad3-rps12, respectively (Jonietz et al., 2010, 2011; Hölzle et al., 2011; Stoll et al., 2017; Schleicher and Binder, 2021). RPF1 is essential for the efficient processing of the nad4 -228 and nad4L-atp4 5′ ends. In the RPF1 inactive ecotype Landsberg erecta (Ler), the major nad4 5′ end is found at position −390, whereas in Columbia (Col) and C24, the major 5′ end occurs at −228 nucleotides upstream of the ATG (Hölzle et al., 2011; Schleicher and Binder, 2021). Similarly, multiple Arabidopsis ecotypes with inactive RPF proteins display extended 5′ UTRs compared to those in ecotypes where RPF proteins are active (Jonietz et al., 2010, 2011; Stoll et al., 2015, 2017; Schleicher and Binder, 2021). Moreover, a similar effect was observed for RPF5, which is not part of the RFL clade. RPF5 stimulates the efficient 5′ processing of three different RNAs; the precursor molecules of these RNAs share conserved sequences that closely match the predicted binding site for RPF5 based on the PPR code (Hauler et al., 2013). Therefore, PPR proteins likely play important roles in promoting RNA cleavage. The proposed mechanism involves the PPR protein binding to the 5′ terminus of precursor RNA, which then recruits the necessary machinery to remodel the RNA structure and expose a cleavage site (Figure 5C). A similar mechanism, where a PPR protein induces endonucleolytic cleavage, was observed with RF1a in rice, which restores fertility in Boro II CMS by blocking ORF79 through endonucleolytic cleavage (Wang et al., 2006).

Translation

Some PPR proteins function as translational activators or crucial components of mitoribosomes in plant organelles. Several PPR proteins are reported to stimulate the translation of specific chloroplast ORFs by stabilizing RNA segments (Supplemental Table 4). These PPR proteins bind to the 5′ UTRs of ORFs to activate their translation. From the study of PPR10 and its role in activating atpH translation, a potential mechanism has been proposed (Barkan and Small, 2014). As illustrated in Figure 6A, a PPR protein binds to an RNA segment near the ribosome's binding sites within the 5′ UTR of the ORF. This binding prevents the formation of a hairpin structure, thereby exposing the ribosome-binding sites for efficient translation. In the absence of the PPR protein, an RNA hairpin forms, which blocks the ribosome-binding site and inhibits translation. It is hypothesized that other proteins may employ similar mechanisms to activate ORF translation (Rojas et al., 2024). HCF173, a member of the short-chain dehydrogenase/reductase superfamily, is associated with the psbA 5′ UTR and is hypothesized to enhance translation by binding to an RNA segment that exposes the ribosome-binding site. An artificial PPR or designed PPR (aPPR or dPPR) protein can substitute for HCF173 when it is bound to the HCF173-binding site, further supporting the mechanism of PPR-mediated translational activation by unmasking the ribosome-binding site (Sane et al., 2005; Hammani et al., 2012; Rojas et al., 2024). In addition, the discovery that many PPR footprints map very close to (but do not overlap with) the anticipated footprints of the initiating ribosome at a start codon further supports this mechanism (Ruwe and Schmitz-Linneweber, 2012; Zhelyazkova et al., 2012). In mitochondria, this mechanism may be conserved. MTL1, a P-class PPR protein, is essential for the translation of mitochondrial nad7 mRNA in Arabidopsis, although it is not involved in the maturation of the 5′ and 3′ extremities of nad7 mRNA (Haili et al., 2013). Mutation of MTL1 partially decreases intron splicing in nad7 intron2. It was also found that ribosome association with nad7 mRNA is specifically disrupted in mtl1 mutants (Haïli et al., 2013).

Figure 6.

Models of PPR proteins mediating the translation of mRNA.

(A) PPR protein specifically binds to the RNA sequence to unmask the ribosome-binding site, activating translation.

(B) PPR protein specifically binds to the ribosome-binding site, inhibiting the initiation or ribosome movement. In addition, rPPRs are components of the ribosome, and mTRAN1 and mTRAN2 may act as universal homing factors to guide the mitoribosome to mitochondrial mRNAs and initiate translation (Tran et al., 2023). rPPR, ribosomal PPR.

Other mechanisms of translation regulation may be adopted by PPR proteins. For example, ATP4 activates the translation of the atpB ORF but binds to an RNA segment hundreds of nucleotides upstream of the atpB start codon (Zoschke et al., 2012a), suggesting that activation may not occur via facilitating ribosome binding. In contrast, PPR proteins can also repress protein translation by masking the ribosome-binding site or impeding ribosome movement along ORFs (Figure 6B). This translational suppression mediated by PPR proteins has been demonstrated in several Rf-type PPR proteins (Supplemental Tables 1 and 4). For example, PPR-B from the Brassica napus Rfo locus inhibits the translation of the sterility-inducing orf138 mRNA without affecting its size or abundance (Uyttewaal et al., 2008b; Wang et al., 2021) by binding within the coding sequence and acting as a ribosome blocker (Wang et al., 2021). Similarly, Rf1a, a PPR-type Rf protein, promotes intercistronic cleavage of the B-atp6-orf79 transcripts by binding to the intercistronic region of this bicistronic mRNA and inhibits the translation of sterility-inducing orf79 mRNA (Wang et al., 2006; Kazama et al., 2008). Analysis has shown that while processed atp6 transcripts are associated with polysomes, the processed orf79 transcripts in the restorer line are not, suggesting that Rf1a may prevent ribosome binding to orf79 transcripts (Kazama et al., 2008).

Moreover, PPR proteins are integral components of the mitoribosome (Figure 6A and 6B) and act as translation initiation factors in plant mitochondria (Waltz et al., 2019, 2020; Tran et al., 2023). Biochemical characterization of Arabidopsis mitochondrial ribosomes identified 19 plant-specific mitoribosome proteins, 10 of which are PPR proteins, termed ribosomal PPR (rPPR) (Waltz et al., 2019). rPPR1, also known as PPR336, is associated with ribosomes in Arabidopsis mitochondria (Uyttewaal et al., 2008a) and serves as a generic translation factor for mitochondrial mRNA (Waltz et al., 2019). Structural analysis reveals that rPPR4 and rPPR5 stabilize single-stranded rRNA segments, whereas rPPR6, rPPR9, and rPPR∗ primarily interact with the backbone of the rRNAs. rPPR4, rPPR9, and rPPR5 are embedded in the large subunit of the mitoribosome (mtLSU), with rPPR5 stabilizing additional rRNA extensions in mtLSU, and rPPR9 and rPPR4 directly contacting and stabilizing the remodeled domain III rRNA that contains two main helices largely extending into the solvent (Waltz et al., 2020). rPPR6 and rPPR∗ are part of the mitoribosome small subunit (mtSSU), with rPPR6 stabilizing the rRNA extension (Waltz et al., 2020). Mutations in each plant-specific rPPR, except for rPPR3a and rPPR3b, lead to various growth defects, highlighting their essential roles. Recently, two noncanonical PPR proteins, rPPR10/mS83, also known as mitochondrial translation factors mTRAN1 and mTRAN2, were discovered as components of the plant mtSSU (Tran et al., 2023). These two PPRs directly bind to A/U-rich motifs (CUUUxU and AAGAAx/AxAAAG), which may act as ribosome-binding sites in the 5′ regions of thousands of plant mitochondrial mRNAs both in vitro and in vivo (Figure 6A and 6B). Ribosome footprinting (Ribo-seq) has shown that mTRANs are essential for binding and translating likely all plant mitochondrial mRNAs, suggesting that mTRANs may function as universal homing factors to guide the mitoribosome to mitochondrial mRNAs and initiate translation (Tran et al., 2023). Thus, PPR proteins regulate mRNA translation in multiple ways in plant mitochondria and chloroplasts (Figure 6A and 6B).

Nuclear gene expression

Most reported PPR proteins are targeted to mitochondria and plastids (Supplemental Table 2. PPR proteins in RNA editing in plants, Supplemental Table 3. PPR proteins in intron splicing in plants, Supplemental Table 4. PPR proteins in RNA maturation, stabilization, and translation in plants). However, several are also found to localize to the nucleus, including GRP23, PNM1, OsNPPR1, FLO14, and bCCP1. Notably, GRP23 is localized to mitochondria, plastids, and nuclei (Ding et al., 2006; Yang et al., 2022a) and its size varies in each compartment (Yang et al., 2022a). GRP23 mediates C-to-U editing at many sites in mitochondria and chloroplasts. In the nucleus, it physically interacts with the nuclear RNA polymerase II subunit III (RBP36B) in both Y2H and bimolecular fluorescence complementation (BiFC) assays, suggesting that GRP23 might function as a transcription co-factor to regulate nuclear gene expression in Arabidopsis (Ding et al., 2006). PNM1 is localized to both the mitochondrion and nucleus in Arabidopsis (Hammani et al., 2011) and is associated with polysomes, potentially playing a role in mitochondrial protein translation. In the nucleus, however, the function of PNM1 is not essential for embryogenesis. TCP8 can bind to the promoter of the PNM1 gene, suggesting that PNM1 may regulate its own gene expression in the nucleus through interaction with TCP8. GFP fusion assays indicate that full-length PNM1-GFP is preferentially localized to mitochondria. Localization of PNM1 to the nucleus occurs only when its mitochondrial targeting signal (MTS) is masked or absent, and the mechanisms behind these localization differences remain to be uncovered (Hammani et al., 2011). OsNPPR1 is specifically localized to nuclei and affects mitochondrial function in rice. In vitro systematic evolution of ligands by exponential enrichment (SELEX) and RNA electrophoretic mobility shift assays suggest that OsNPPR1 may bind to the CUCAC motif. In addition, retention of several nuclear introns was observed in the osnppr1 mutant (Hao et al., 2019). bCCP1, which is involved in RNA editing at many sites in mitochondria, is dual-localized to both mitochondria and the nucleus in maize. However, the function of bCCP1 in the nucleus remains to be determined (Wang et al., 2023b). These results indicate that some PPR proteins are localized to the nucleus where they exert diverse functions. Given their RNA-binding activities, nuclear PPR proteins may influence the post-transcriptional processing of nuclear genes.

Potential applications of PPR proteins

RNA modification and metabolism regulated by RBPs play a crucial role in modulating gene expression in nuclear, mitochondrial, and plastid compartments. Abnormal RNA modification and metabolism can inhibit growth and development in plants and lead to diseases in humans. RBPs are implicated in every aspect of RNA maturation and metabolism (Hammani and Giege, 2014; Lee and Kang, 2020). Synthetic RBPs that specifically bind to RNA sequences have the potential to regulate the modification and expression of transcripts. PPR proteins, which belong to the α-solenoid superfamily characterized by regularly spaced helical repeating units that contact the Watson–Crick surfaces of continuous RNA nucleotides, could serve as effective RNA targeting tools (Filipovska and Rackham, 2013). Each repeat in a PPR protein discriminates among nucleotides based on the aa combination at the 5th and last positions, comprising an aa code for nucleotide recognition (Barkan et al., 2012; Yin et al., 2013; Yan et al., 2019). This modular code for RNA binding by PPR proteins holds great promise for engineering synthetic PPRs to target RNA (Filipovska and Rackham, 2013). Due to their natural diversity, a variety of native functions, a relatively clear and simple PPR-binding code (Barkan et al., 2012; Yin et al., 2013; Yan et al., 2019), and versatility, PPR proteins are excellent candidates for synthetic RBPs. Extensive efforts have been made to modify native PPRs to target other RNA sequences or to design aPPRs to regulate transcript processing and RNA expression (McDowell et al., 2022).

Alterations at positions 5 and 35 within the native maize protein PPR10, in accordance with the PPR code, enable effective binding to a new RNA sequence (Barkan et al., 2012). PPR10 prevents 5′-to-3′ exoribonucleolytic decay and increases the translational efficiency of atpH transcripts by binding to an RNA segment approximately 20 nucleotides upstream of the atpH ribosome-binding site, thereby stimulating atpH expression. A variant of PPR10, with aa modifications that alter its sequence specificity, binds to a new target RNA and significantly enhances the expression of a chloroplast transgene (Rojas et al., 2019; Yu et al., 2019). The expression of this PPR10 variant can be regulated using inducible or tissue-specific promoters, offering a means to modulate and optimize the expression of foreign genes in chloroplasts (Rojas et al., 2019; Yu et al., 2019). Similarly, a modified version of the SOT1 protein, with altered PPR motifs, efficiently recognizes and cleaves a specific RNA substrate (Zhou et al., 2017). Another successfully modified native PPR protein, RNA PROCESSING FACTOR 2 (RPF2), targets a new site on mitochondrial nad6 transcripts and mediates their cleavage, effectively abolishing Nad6 protein expression (Colas des Francs-Small et al., 2018). Shen et al. (2015) utilized the most evolutionarily conserved aa of P motifs from A. thaliana as the scaffold for aPPR motifs to establish the primary structure of the RNA base-recognition units. By using different aa combinations at positions 2, 5, and 35, these motifs discriminate among the RNA bases A, U, and C. The designer P-type PPR repeat array is capped by N- and C-terminal domains, derived from positions 37–208 and 737–786 of PPR10, respectively, at the amino and carboxyl termini of the aPPR protein. In vitro, these aPPR proteins recognize RNA bases A, C, and U with high modular selectivity and specifically target the RNA sequence (Shen et al., 2015). Using the approach described by Shen et al. (2015), two aPPR proteins, SCD11 and SCD14, consisting of either 11 or 14 aPPR motifs, and N- and C-terminal segments of the native chloroplast-localized protein PPR10, were designed. These two aPPR proteins demonstrate high specificity to their intended target, the psbA 3′ UTR, in vivo. Ribonucleoproteins primarily bound to psbA RNA are captured using these aPPRs, adding to the toolkit for characterizing native ribonucleoproteins in vivo (McDermott et al., 2019). Recently, an aPPR that binds to HCF173 binding sites has been shown to partially substitute for HCF173, activating psbA translation in A. thaliana (Rojas et al., 2024).

Numerous PLS-class PPR proteins act as specificity factors in RNA C-to-U editing. It has been reported that a single PPR–DYW protein can carry out RNA editing both in vivo and in vitro. The PPR array in a PPR–DYW protein specifically recognizes the editing sites, while its DYW domain catalyzes the deamination in the C-to-U transition. In addition, the extensive presence of PLS-class PPR proteins in plants implies the necessary sequence information for designing consensus-based RBPs. Based on conserved aas in the P, L, and S motifs of all PLS-class PPR proteins in A. thaliana, and in conjunction with PPR-binding codes, the artificial (PLS)₃PPR was designed. (PLS)₃PPR can specifically target the intended RNA sequence, and MORF9 significantly enhances the RNA-binding activity of (PLS)₃PPR (Yan et al., 2017). Subsequently, a dsn3PLS-DYW protein with the motif arrangement (P1-L1-S1)3-P2-L2-S2-E1-E2-DYW, based on the most representative aa at each position in each motif from 9730 PPR protein sequences across 38 different species, was synthesized. The dsn3PLS-DYW protein can target its desired sequence upstream of the rpoA-78691 site and carry out editing at this site in Arabidopsis chloroplasts and heterogeneous bacteria. MORF proteins enhance both the RNA-binding activity of dsn3PLS-DYW protein in vitro and its editing activity at the rpoA-78691 site in bacteria (Royan et al., 2021). The crystal structure of (PLS)₃PPR–MORF9 revealed that MORF9 induces substantial conformational changes in (PLS)₃PPR, primarily resulting from variations in the conformation of L motifs (Royan et al., 2021). This structural insight supports the potential to design a PPR–DYW protein with enhanced editing activity and specificity. Bernath-Levin et al. (2021) designed a PPR protein that includes bona fide designer S motifs and a DYW domain at its C terminus. This designed PPR protein targets and edits the intended editing site in E. coli and functions independently of cofactors in RNA binding and editing (Bernath-Levin et al., 2021). In addition, natural PPR–DYW proteins PpPPR56 and PpPPR65 have been found to perform targeted C-to-U editing of nuclear transcripts in human cells (Lesch et al., 2022). The dPPR–DYW:KP proteins edit targeted U to C in both bacteria and human cells (Ichinose et al., 2022). These successful designs and applications of programmable RNA-editing factors demonstrate their potential as biotechnological tools for manipulating RNA in vivo.

Based on the characteristics of PPR proteins and their successful applications, aPPR proteins have several potential applications (McDowell et al., 2022). (1) Given that the target sequence of PPR-editing factors can be predicted based on the PPR-binding code (Takenaka et al., 2013; Yagi et al., 2013; Yan et al., 2019), it seems feasible to use designed PPR proteins with enzymatic domains (e.g., DYW, DYW:KP) at their C termini to specifically edit desired sites (Figure 7A). However, several challenges must be addressed when using these designed PPR proteins for targeted editing. Firstly, the off-target effects of synthetic or natural variant PPR proteins can significantly limit their use, especially when editing RNA in the nuclear genome, which contains more genes than organelle genomes (Ichinose et al., 2022; Lesch et al., 2022). Secondly, co-factors are necessary for editing in plant organelles (Sun et al., 2016), adding another layer of complexity. Efforts to address this issue have led to the development of a designed S-type editing factor that can effectively edit an RNA target in E. coli without requiring any additional co-factors (Bernath-Levin et al., 2021). Nonetheless, this new tool poses challenges, such as ensuring that the designed PPR-editing factor can quickly dissociate from the target RNA after editing. If the PPR protein binds too tightly, it could potentially inhibit the translation of the target RNA (Figure 6B). (2) It has been reported that natural PPR proteins containing a nuclease domain are involved in RNA processing (Gobert et al., 2010; Gobert et al., 2013; Zhou et al., 2017). For example, proteinaceous RNase P (PRORP) and SOT1 each contain an NYN nuclease domain and an SMR domain, respectively. Therefore, an artificial P-class PPR protein with a nuclease domain at its C terminus could target specific RNA sequences and cleave these RNAs at designated sites (Figure 7B). This approach is similar to the engineering of site-directed RNA endonucleases, such as those developed by fusing ZF and TALE nucleases or PUF-PIN (pilT amino terminus) (Choudhury et al., 2012; Perez-Pinera et al., 2012). In addition, modified natural PPR proteins can facilitate the cleavage of new target RNAs, effectively abolishing their expression. A notable example includes the modified RPF2, which mediates the cleavage of nad6 in Arabidopsis mitochondria, serving as a successful model (Colas des Francs-Small et al., 2018). (3) Artificial PPR proteins can specifically bind to desired RNA sequences and subsequently remodel the RNA, thereby exposing endonuclease target sites. This alteration induces cleavage in the intercistronic regions between two ORFs (Figure 7C). This tool has the potential to restore CMS. (4) Artificial PPR proteins specifically bind to desired RNAs to remodel their secondary and tertiary structures, thereby unmasking ribosome-binding sites. This remodeling impedes the initiation of translation or occupies target RNA sequences, preventing typical intermolecular interactions with the ribosome that could impair translation initiation (Figure 7D). For example, an aPPR that binds to the HCF173-binding sites can partially substitute for HCF173 to activate psbA translation in A. thaliana, demonstrating a successful application for remodeling and blocking (Rojas et al., 2024). (5) RNA distribution within cells is not uniform but varies according to functional requirements, which is critical for the accurate regulation of gene expression and cellular development. However, tracking RNA distribution is challenging. Artificial PPR proteins tagged with a fluorescent protein, such as GFP or YFP, can specifically bind to target RNAs, and the fluorescence emitted by these proteins may indicate the distribution of these RNAs (Figure 7E). Similarly, RNA has been successfully labeled in live cells using other RBP fusions with GFP, for example, modified PUF–GFP fusion proteins that examine the distribution of viral RNAs in infected plant cells (Tilsner et al., 2009). The potential off-target effects of synthetic PPR proteins or natural variations in PPR proteins may lead to inaccurate measurements of RNA distribution, and an excess of PPR motifs tends to allow for more mismatches (Miranda et al., 2018). An alternative method involves using two PPRs with split-fluorescent proteins designed to target the same specific RNA sites (Figure 7E). Fluorescence is only detected when these two dPPRs simultaneously bind to the same RNA, potentially enabling more accurate tracking of RNA distribution. Thus, aPPR proteins offer a promising tool to regulate RNA processing and translation, as well as to track and modify RNA distribution within cells.

Figure 7.

Potential applications for dPPR proteins.

(A) Editing of specific sites by dPPR–DYW or PPR–DYW:KP proteins.

(B) Cleavage of an RNA by a dPPR-endonuclease fusion protein at a specific site.

(C) Remodeling of RNA structure and unmasking of the cleavage site at an intercistronic region between two ORFs by a dPPR, restoring CMS.

(D) Impairment or activation of the translation of a specific ORF by masking or unmasking the ribosome-binding site.

(E) Detection of transcript localization using a fluorescent or luminescent reporter molecule.

As most fertility-restorer genes are found to be PPRs in various species (Supplemental Table 1), PPR genes can be used to manage crop fertility. CMS lines, specifically the wild-abortive type CMS (CMS-WA), have contributed significantly to rice breeding. The mitochondrial genome of CMS-WA contains an aberrant ORF encoding the WA352 protein, which interacts with the mitochondrial protein COX11. This interaction attenuates the function of COX11, leading to premature tapetal programmed cell death and consequent male sterility (Luo et al., 2013). The abundance of WA352 transcripts is reduced in the restorer lines that carry the fertility-restorer gene Rf4, which encodes a PPR protein (Tang et al., 2014). Based on this finding, an engineered PPR gene could be designed to restore the fertility of CMS-WA rice by reducing the abundance or translation of WA352 transcripts. As described in Figure 7A, a dPPR–DYW protein precisely edits C to U at the WA352-109 site to generate a premature stop codon (CAG to UAG), terminating the translation of WA352 mRNA. Alternatively, a dPPR–endonuclease fusion protein can be engineered to specifically recognize and bind to WA352, inducing the cleavage or degradation of WA352 transcripts (Figure 7B). As shown in Figure 7D, a dPPR protein can tightly bind to the WA352 mRNA, blocking its translation. These dPPR proteins could potentially reduce the WA352 protein level and restore fertility in CMS-WA rice. If dPPRs are regulated by an inducible promoter, the fertility of the CMS-WA rice can be controlled. Therefore, by modifying deleterious mitochondrial RNAs, artificial PPR proteins offer a precise and powerful tool for regulating fertility in CMS lines.

Funding

This work was supported by the National Natural Science Foundation of China (project no. 32101640 and 32230075).

Acknowledgments

No conflict of interest is declared.

Author contributions

Y.W. and B.-C.T. conceptualized and wrote this manuscript.

Published: December 5, 2024