The Mitochondrial Pentatricopeptide Repeat Protein PPR18 Is Required for the cis-Splicing of nad4 Intron 1 and Essential to Seed Development in Maize

Pentatricopeptide repeat (PPR) protein comprises a large family, participating in various aspects of organellar RNA metabolism in land plants. There are approximately 600 PPR proteins in maize, but the functions of many PPR proteins remain unknown. In this study, we defined the function of PPR18 in the cis-splicing of nad4 intron 1 in mitochondria and seed development in maize. Loss function of PPR18 seriously impairs embryo and endosperm development, resulting in the empty pericarp (emp) phenotype in maize. PPR18 encodes a mitochondrion-targeted P-type PPR protein with 18 PPR motifs. Transcripts analysis indicated that the splicing of nad4 intron 1 is impaired in the ppr18 mutant, resulting in the absence of nad4 transcript, leading to severely reduced assembly and activity of mitochondrial complex I and dramatically reduced respiration rate. These results demonstrate that PPR18 is required for the cis-splicing of nad4 intron 1 in mitochondria, and critical to complex I assembly and seed development in maize.

Mature mitochondrial transcripts undergo extensive post-transcriptional processing events, among which the most reported are RNA editing and RNA splicing [5][6][7]. In flowering plants, RNA editing usually alters cytidine to uridine through a deamination reaction in mitochondria and plastids and RNA splicing is a processing event in which noncoding segments (intron) of precursor RNA are removed and coding sequences are joined. Based on the distinctive structures, introns are divided into two families, group I and group II [4,8]. In flowering plants, most of organellar transcripts contain group II introns with conserved secondary structure, consisting of six domains extending from a central hub [9]. In bacteria, the splicing of a group II intron is self-facilitated by its cognate maturase encoded in the intron domain IV [10], but in plants, nearly all introns lost the maturase gene with only one intron maturase gene (matK) remains in the plastid genome and one matR gene in the mitochondrial genome [11,12]. Instead, four maturase genes (nMat1 to 4) are found in the nuclear genome in Arabidopsis [13][14][15]. Besides, numerous additional nucleus-encoded splicing co-factors have been reported to be involved in the splicing of organellar introns, such as the chloroplast RNA splicing and ribosome maturation (CRM) domain-containing proteins [16,17], RNA helicase [18,19], mitochondrial transcription termination factors (mTERF) [20,21], plant organellar RNA recognition (PORR) domain proteins [22,23], regulator of chromosome condensation (RCC1) domain proteins [24], and the pentatricopeptide repeat (PPR) proteins [5,25,26].
PPR proteins are a large family of RNA binding proteins, with more than 400 members in angiosperms [5,27]. PPR proteins contain multiple 35-amino-acid tandem repeats and each repeat forms a helix-loop-helix structure. Based on domain constitution, PPR proteins are divided into PLS (repeat P-L-S motif)-class proteins and P-class proteins [27]. The PLS-subclass PPR proteins contain characteristic triplets of P, L, and S motifs with additional E, E+, DYW, or other domains at the C-terminus, whereas the P-subclass PPR proteins contain arrays of only P motifs [5]. The PLS-class PPR proteins are implicated in the C-to-U RNA editing that in most cases is to restore the evolutionary conserved amino acids [28]. Functions of the P-subclass PPR proteins are diverse, which includes RNA cleavage, RNA splicing, RNA stabilization and maturation, and translation initiation [5]. Most PPR proteins are localized in mitochondria or chloroplasts. They bind RNA in a sequence specific manner that one PPR motif binds to one nucleotide of the target RNA. The recognition nucleotides were determined by the different combinations of the amino acid residues at position 5th and 35th of each PPR repeat, which is known as the PPR codes [29][30][31]. In plant mitochondria, most of group II introns are present in genes that code for subunits of mitochondrial complex I. In maize mitochondria, out of 22 identified group II introns, 19 resides in nad1, nad2, nad4, nad5, and nad7 transcripts, while 3 in rps3, cox2, and ccmFc transcripts [32,33]. Accurate splicing of these group II introns is critical to mitochondrial function and biogenesis, which is important for plant growth and development. For instance, EMP11 and DEK2 are involved in the splicing of nad1 introns, and the loss of function mutation of Emp11 and Dek2 affects the assembly of complex I with severely arrested embryo and endosperm development [34,35]. EMP10, EMP12, EMP16, DEK37, and PPR20 are responsible for the splicing of nad2 introns in maize. These mutations result in a loss of mitochondrial complex I assembly and activity, impairing the mitochondrial function and embryogenesis and endosperm development [25,[36][37][38][39].
In this study, we characterized a maize seed mutant ppr18, which exhibits arrested embryo and endosperm development phenotype. PPR18 encodes a mitochondrion-targeted P-type PPR protein with 18 PPR motifs. The loss of PPR18 function leads to the splicing deficiency of nad4 intron 1, severely reduced assembly and activity of mitochondrial complex I, resulting in the impairment of mitochondrial function and seed development in maize.

PPR18 Is a Mitochondrion-Localized P-Type PPR Protein
PPR18 (GRMZM2G438456) is an intronless gene, encoding an 85 kDa protein with 768 amino acid residues ( Figure 1A). Motif prediction analysis by algorithm TPRpred (http://tprpred.tebingen. mpg.de/tprpred) revealed that PPR18 contained 18 tandemly repeated PPR motifs without any other domains, suggesting that PPR18 is a canonical P-type PPR protein ( Figure 1A,B). A phylogenetic analysis based on the maize PPR18 and its homologous proteins revealed extensive conservation in the sequences in both monocots and dicots ( Figure S1). Most of PPRs are localized in organelles, either chloroplasts or/and mitochondria, except GRP23 and PNM1, which both have nucleus localized signals [5,40,41]. To determine the subcellular localization of PPR18, the 550 amino acid residues of the N-terminal PPR18 were fused to the green fluorescent protein (GFP) in the binary vector pGWB5, then transiently expressed in the tobacco leaves via Agrobacterium EHA105 infiltration. Confocal laser-scanning microscopy revealed that the strong green fluorescence signals of PPR18 N550 -GFP are merged with the red signals of MitoTracker ( Figure 1C), indicating that PPR18 is localized in mitochondria. signals of PPR18 N550 -GFP are merged with the red signals of MitoTracker ( Figure 1C), indicating that PPR18 is localized in mitochondria.

Embryo and Endosperm Development Are Arrested in ppr18
To characterize the function of PPR18, we isolated two independent Mutator (Mu) insertional mutants containing Mu insertions at 880 bp and 939 bp downstream from the start codon of PPR18 from the UniformMu population in the inbred W22 genetic background, named ppr18-1 and ppr18-2, respectively [42]. The selfed progeny of both ppr18-1/+ and ppr18-2/+ heterozygotes segregated at a 3:1 ratio of wild-type (WT) and empty pericarp (emp) kernels, suggesting that both mutants are monogenic nuclear recessive mutations and homozygous lethal (Figures 2A and S2A). Cosegregation analysis of a small isolated population with 72 individuals was performed to test the linkage of ppr18-1 by genomic PCR using PPR18-R1 and Mu TIR8 primers [43]. The results showed that all the self-pollinated progenies of ppr18-1/+ plants produced emp kernels, indicating that the Mu insertion is tightly linked to the mutation ( Figure S3). Crosses between ppr18-1/+ and ppr18-2/+ heterozygotes produced heteroallelic progeny ppr18-1/ppr18-2 with approximately 25% emp kernels ( Figure S2B), confirming that the ppr18 phenotype results from the disruption of GRMZM2G438456.

Embryo and Endosperm Development Are Arrested in ppr18
To characterize the function of PPR18, we isolated two independent Mutator (Mu) insertional mutants containing Mu insertions at 880 bp and 939 bp downstream from the start codon of PPR18 from the UniformMu population in the inbred W22 genetic background, named ppr18-1 and ppr18-2, respectively [42]. The selfed progeny of both ppr18-1/+ and ppr18-2/+ heterozygotes segregated at a 3:1 ratio of wild-type (WT) and empty pericarp (emp) kernels, suggesting that both mutants are monogenic nuclear recessive mutations and homozygous lethal (Figure 2A and Figure S2A). Co-segregation analysis of a small isolated population with 72 individuals was performed to test the linkage of ppr18-1 by genomic PCR using PPR18-R1 and Mu TIR8 primers [43]. The results showed that all the self-pollinated progenies of ppr18-1/+ plants produced emp kernels, indicating that the Mu insertion is tightly linked to the mutation ( Figure S3). Crosses between ppr18-1/+ and ppr18-2/+ heterozygotes produced heteroallelic progeny ppr18-1/ppr18-2 with approximately 25% emp kernels ( Figure S2B), confirming that the ppr18 phenotype results from the disruption of GRMZM2G438456. The developing kernels phenotype of WT and ppr18-1 in the same segregating ear are compared in Figure 2B,C. The ppr18-1 mutant kernels are remarkably smaller than the wild type, which exhibited pale, half-translucent, and collapsed appearance at 35 days after pollination (DAP; Figure  2B-E). Compared with the WT siblings in the same segregating ear, the embryo and endosperm development are arrested in the ppr18-1 kernels ( Figure 2B-E). To examine the developmental arrest of embryogenesis in ppr18-1, we examined the embryo and endosperm development process between ppr18-1 and WT siblings in a segregating ear by light microscopy. At 9 DAP, the WT embryos reached the coleoptilar stage, whereas the ppr18-1 embryos remained at the pre-embryo stage ( Figure 2F,G). At 14 DAP, the WT embryos reached late embryogenesis stage, while the ppr18-1 embryos stayed at the transition stage without any discernable differentiation ( Figure 2H,I). These results indicate that the loss of PPR18 severely arrests both embryo and endosperm development.
To assess the impact of the Mu insertion on the PPR18 expression, we analyzed the transcript level of PPR18 in two ppr18 alleles by reverse transcription PCR (RT-PCR). Results showed that no transcript of PPR18 was detected in both alleles ( Figure S4A), indicating that both alleles are probably null mutations. In wild type, PPR18 transcripts can be detected in all vegetative and reproductive tissues by quantitative real-time PCR (qRT-PCR; Figure S4B). Relative high mRNA expression of PPR18 was in bract and low expression in root, flower and kernel at developmental stages, indicating that PPR18 is a constitutively expressed gene throughout growth and development in maize, rather than a seed specific gene. As the mutants are embryo lethal, impacts on other tissues and during development cannot be determined.

Loss of PPR18 Affects Mitochondrial Respiratory Activity
Previous reports showed the mutation of mitochondrion-localized PPR proteins with arrested seed development often suffers defects in mitochondrial respiration [44][45][46]. Thus, we investigated whether the loss of PPR18 affects mitochondrial respiratory activity in maize by determining the respiratory activity, as shown by three mitochondrial respiratory rates, including total respiratory The developing kernels phenotype of WT and ppr18-1 in the same segregating ear are compared in Figure 2B,C. The ppr18-1 mutant kernels are remarkably smaller than the wild type, which exhibited pale, half-translucent, and collapsed appearance at 35 days after pollination (DAP; Figure 2B-E). Compared with the WT siblings in the same segregating ear, the embryo and endosperm development are arrested in the ppr18-1 kernels ( Figure 2B-E). To examine the developmental arrest of embryogenesis in ppr18-1, we examined the embryo and endosperm development process between ppr18-1 and WT siblings in a segregating ear by light microscopy. At 9 DAP, the WT embryos reached the coleoptilar stage, whereas the ppr18-1 embryos remained at the pre-embryo stage ( Figure 2F,G). At 14 DAP, the WT embryos reached late embryogenesis stage, while the ppr18-1 embryos stayed at the transition stage without any discernable differentiation ( Figure 2H,I). These results indicate that the loss of PPR18 severely arrests both embryo and endosperm development.
To assess the impact of the Mu insertion on the PPR18 expression, we analyzed the transcript level of PPR18 in two ppr18 alleles by reverse transcription PCR (RT-PCR). Results showed that no transcript of PPR18 was detected in both alleles ( Figure S4A), indicating that both alleles are probably null mutations. In wild type, PPR18 transcripts can be detected in all vegetative and reproductive tissues by quantitative real-time PCR (qRT-PCR; Figure S4B). Relative high mRNA expression of PPR18 was in bract and low expression in root, flower and kernel at developmental stages, indicating that PPR18 is a constitutively expressed gene throughout growth and development in maize, rather than a seed specific gene. As the mutants are embryo lethal, impacts on other tissues and during development cannot be determined.

Loss of PPR18 Affects Mitochondrial Respiratory Activity
Previous reports showed the mutation of mitochondrion-localized PPR proteins with arrested seed development often suffers defects in mitochondrial respiration [44][45][46]. Thus, we investigated whether the loss of PPR18 affects mitochondrial respiratory activity in maize by determining the respiratory activity, as shown by three mitochondrial respiratory rates, including total respiratory (V t ), cytochrome respiratory capacity (V cyt ), and alternative respiratory capacity (V alt ). The ratio of V cyt /V t was significantly reduced in the ppr18-1 mutant compared with WT (Table 1), indicating that loss of PPR18 resulted in a severe reduction of the cytochrome pathway and impaired mitochondrial respiration. Meanwhile, the ratio of V alt /V t was markedly increased in ppr18-1 (Table 1), supporting that alternative respiratory pathway was enhanced in the ppr18-1 mutant. Moreover, we detected the expression of alternative oxidase (AOX) protein by Western blot assay using the specific antibody of AOX. Results showed that the abundance of AOX was increased drastically in the ppr18-1 mutant compared to WT ( Figure 3C), confirming that mutation of PPR18 enhances the expression of alternative oxidases. The maize genome contains three AOX genes, AOX1, AOX2, and AOX3. Both RT-PCR and qRT-PCR assays showed that the expression of AOX2 and AOX3 was dramatically increased in the two ppr18 alleles ( Figure 3A,B), indicating that excessive accumulation of AOX is caused by upregulation of AOX2 and AOX3 expression in the ppr18 alleles. Table 1. Alteration of the respiration rate of the wild type (WT) and ppr18-1 kernels. Mitochondrial total respiration rate (V t ), the alternative pathway (V alt ), and the capacity of the cytochrome pathway (V cyt ) were indicated by the oxygen consumption of nmol O 2 min −1 g −1 fresh weight of the maize kernels at 11 DAP using a Clark-type oxygen electrode. Data are mean values ± SEs from three independent biological samples. (Vt), cytochrome respiratory capacity (Vcyt), and alternative respiratory capacity (Valt). The ratio of Vcyt/Vt was significantly reduced in the ppr18-1 mutant compared with WT (Table 1), indicating that loss of PPR18 resulted in a severe reduction of the cytochrome pathway and impaired mitochondrial respiration. Meanwhile, the ratio of Valt/Vt was markedly increased in ppr18-1 (Table 1), supporting that alternative respiratory pathway was enhanced in the ppr18-1 mutant. Moreover, we detected the expression of alternative oxidase (AOX) protein by Western blot assay using the specific antibody of AOX. Results showed that the abundance of AOX was increased drastically in the ppr18-1 mutant compared to WT ( Figure 3C), confirming that mutation of PPR18 enhances the expression of alternative oxidases. The maize genome contains three AOX genes, AOX1, AOX2, and AOX3. Both RT-PCR and qRT-PCR assays showed that the expression of AOX2 and AOX3 was dramatically increased in the two ppr18 alleles ( Figure 3A,B), indicating that excessive accumulation of AOX is caused by upregulation of AOX2 and AOX3 expression in the ppr18 alleles. Mitochondrial total respiration rate (Vt), the alternative pathway (Valt), and the capacity of the cytochrome pathway (Vcyt) were indicated by the oxygen consumption of nmol O2 min −1 g −1 fresh weight of the maize kernels at 11 DAP using a Clark-type oxygen electrode. Data are mean values ± SEs from three independent biological samples.

Loss of PPR18 Affects the Assembly and Activity of Complex I
The limited cytochrome pathway is closely relevant to the defective transfer electrons from mitochondrial respiratory complexes, complex I to IV [47,48]. To determine the assembly and abundance of mitochondrial respiratory complexes, we performed blue native (BN)-PAGE using crude mitochondria from ppr18 alleles and WT maize kernels ( Figure 4). As indicated by Coomassie Brilliant Blue (CBB) staining, complex III and V was substantially accumulated, whereas complex I and supercomplex I + III 2 were depleted in both ppr18 alleles compared to WT ( Figure 4A), indicating that loss of PPR18 affects the assembly of mitochondrial complex I. Furthermore, we analyzed the NADH dehydrogenase activity of complex I by in-gel NADH activity assay, which showed a consistent result with the CBB staining. As shown in Figure 4B, the dehydrogenase activity of complex I and supercomplex I + III 2 were completely deficient in both ppr18 alleles. Besides, we detected the assembly of complex III, IV, and V by Western blot analysis using the anti-Cyt C1 , anti-Cox2, and anti-ATP synthase α-subunit antibody, respectively. Results showed that complex III, IV, and V were increased in the ppr18-1 mutant ( Figure 4C-E). Collectively, these results imply that PPR18 is important for the assembly and activity of mitochondrial complex I in maize.
CBB (Coomassie Brilliant Blue) staining demonstrates that equal amounts of mitochondrial proteins were loaded.

Loss of PPR18 Affects the Assembly and Activity of Complex I
The limited cytochrome pathway is closely relevant to the defective transfer electrons from mitochondrial respiratory complexes, complex I to IV [47,48]. To determine the assembly and abundance of mitochondrial respiratory complexes, we performed blue native (BN)-PAGE using crude mitochondria from ppr18 alleles and WT maize kernels ( Figure 4). As indicated by Coomassie Brilliant Blue (CBB) staining, complex III and V was substantially accumulated, whereas complex I and supercomplex I + III2 were depleted in both ppr18 alleles compared to WT ( Figure 4A), indicating that loss of PPR18 affects the assembly of mitochondrial complex I. Furthermore, we analyzed the NADH dehydrogenase activity of complex I by in-gel NADH activity assay, which showed a consistent result with the CBB staining. As shown in Figure 4B, the dehydrogenase activity of complex I and supercomplex I + III2 were completely deficient in both ppr18 alleles. Besides, we detected the assembly of complex III, IV, and V by Western blot analysis using the anti-CytC1, anti-Cox2, and anti-ATP synthase α-subunit antibody, respectively. Results showed that complex III, IV, and V were increased in the ppr18-1 mutant ( Figure 4C-E). Collectively, these results imply that PPR18 is important for the assembly and activity of mitochondrial complex I in maize. In addition, we determined the abundance of the mitochondrial complex proteins in ppr18-1 and WT maize kernels by Western blot analysis using antibodies against Nad9 (complex I), CytC1 (complex III), Cox2 (complex IV), and ATPase (complex V). As shown in Figure 3C, the protein abundance of Nad9, CytC1, Cox2, and ATPase was increased in ppr18-1, speculating that the lack of PPR18 may enhance the expression of subunit from complex I, III, IV, and V in a feedback mechanism. In addition, we determined the abundance of the mitochondrial complex proteins in ppr18-1 and WT maize kernels by Western blot analysis using antibodies against Nad9 (complex I), Cyt C1 (complex III), Cox2 (complex IV), and ATPase (complex V). As shown in Figure 3C, the protein abundance of Nad9, Cyt C1 , Cox2, and ATPase was increased in ppr18-1, speculating that the lack of PPR18 may enhance the expression of subunit from complex I, III, IV, and V in a feedback mechanism.

PPR18 Is Required for the Splicing of nad4 Intron 1
Previous study showed that most P-type PPR proteins function on intron splicing, RNA maturation, RNA stabilization or RNA cleavage in organelles [5]. To reveal the molecular function of PPR18, we analyzed the transcript levels of 35 mitochondrion-encoded genes between WT and ppr18 alleles by RT-PCR and qRT-PCR. The results showed that the expression level of most mitochondrion-encoded genes was indistinguishable between the WT and ppr18 alleles, only the expression of nad4 was obviously different between WT and ppr18 alleles ( Figure 5 and Figure S5). The mature nad4 transcript was not detectable in both ppr18 alleles. Instead, a band larger than the mature nad4 transcript was dramatically increased in both ppr18 mutants ( Figure 5). The sequencing results of the larger fragments showed that these fragments contain unspliced nad4 intron 1. These results indicate that the loss-of-function in PPR18 abolishes the splicing of nad4 intron 1 in mitochondria. maturation, RNA stabilization or RNA cleavage in organelles [5]. To reveal the molecular function of PPR18, we analyzed the transcript levels of 35 mitochondrion-encoded genes between WT and ppr18 alleles by RT-PCR and qRT-PCR. The results showed that the expression level of most mitochondrion-encoded genes was indistinguishable between the WT and ppr18 alleles, only the expression of nad4 was obviously different between WT and ppr18 alleles (Figures 5 and S5). The mature nad4 transcript was not detectable in both ppr18 alleles. Instead, a band larger than the mature nad4 transcript was dramatically increased in both ppr18 mutants ( Figure 5). The sequencing results of the larger fragments showed that these fragments contain unspliced nad4 intron 1. These results indicate that the loss-of-function in PPR18 abolishes the splicing of nad4 intron 1 in mitochondria. The nad4 precursor RNA contains three cis-splicing introns ( Figure 6A). To confirm the function of PPR18 on the nad4 intron 1 splicing, we amplified fragments containing each of the three introns in the nad4 transcript by RT-PCR using specific primers. As shown in Figure 6B, only the splicing of nad4 intron 1 was impaired in both ppr18 alleles. In addition, we analyzed the splicing efficiency of 22 group II introns in mitochondria by qRT-PCR. Results showed that the splicing efficiency of nad4 intron 1 was dramatically decreased in the ppr18 alleles ( Figure 7). These data suggest that PPR18 is indeed required for the splicing of mitochondrial nad4 intron 1. Previous report showed that PPR proteins bind specific RNA via a modular recognition code in which the nucleotide specificity primarily relies on combination at the 5th and 35th amino acid residues of each PPR motif [29][30][31]. Based on PPR recognition code, potential binding sites of PPR18 in mitochondrial nad4 intron 1 were The nad4 precursor RNA contains three cis-splicing introns ( Figure 6A). To confirm the function of PPR18 on the nad4 intron 1 splicing, we amplified fragments containing each of the three introns in the nad4 transcript by RT-PCR using specific primers. As shown in Figure 6B, only the splicing of nad4 intron 1 was impaired in both ppr18 alleles. In addition, we analyzed the splicing efficiency of 22 group II introns in mitochondria by qRT-PCR. Results showed that the splicing efficiency of nad4 intron 1 was dramatically decreased in the ppr18 alleles ( Figure 7). These data suggest that PPR18 is indeed required for the splicing of mitochondrial nad4 intron 1. Previous report showed that PPR proteins bind specific RNA via a modular recognition code in which the nucleotide specificity primarily relies on combination at the 5th and 35th amino acid residues of each PPR motif [29][30][31]. Based on PPR recognition code, potential binding sites of PPR18 in mitochondrial nad4 intron 1 were predicted ( Figure 6C). Results showed that the nucleotides of nad4 intron 1 are well aligned to the combinatorial codes. We predicted the secondary structure of nad4 intron 1 and mapped the putative binding site of PPR18 in domain I of nad4 intron 1 ( Figure S6). A phylogenetic analysis based on genomic DNAs in the GenBank, the putative binding site of PPR18 in nad4 intron 1 appeared to be highly conserved in both monocots and dicots ( Figure S7). predicted ( Figure 6C). Results showed that the nucleotides of nad4 intron 1 are well aligned to the combinatorial codes. We predicted the secondary structure of nad4 intron 1 and mapped the putative binding site of PPR18 in domain I of nad4 intron 1 ( Figure S6). A phylogenetic analysis based on genomic DNAs in the GenBank, the putative binding site of PPR18 in nad4 intron 1 appeared to be highly conserved in both monocots and dicots ( Figure S7).

PPR18 Does Not Show a Direct Interaction with DEK35, EMP8, and EMP602 in Yeast Two-Hybrid Assays
Previous studies reported that three PPR proteins DEK35, EMP8, and EMP602 are involved in the splicing of mitochondrial nad4 intron 1 in maize [46,49,50]. In this study, PPR18 is also required for the splicing of nad4 intron 1. To determine whether PPR18 interacts with DEK35, EMP8, and EMP602, we performed a yeast two-hybrid assay. Results showed that PPR18 has no directly physical interaction with these proteins ( Figure S8).

PPR18 Does Not Show a Direct Interaction with DEK35, EMP8, and EMP602 in Yeast Two-Hybrid Assays
Previous studies reported that three PPR proteins DEK35, EMP8, and EMP602 are involved in the splicing of mitochondrial nad4 intron 1 in maize [46,49,50]. In this study, PPR18 is also required for the splicing of nad4 intron 1. To determine whether PPR18 interacts with DEK35, EMP8, and EMP602, we performed a yeast two-hybrid assay. Results showed that PPR18 has no directly physical interaction with these proteins ( Figure S8).

A Role of PPR18 on nad4 Intron 1 Splicing and the Assembly of Complex I
The maize mitochondria contain a total of 22 group II introns in 8 genes (nad1, nad2, nad4, nad5, nad7, ccmFc, cox2, and rps3) [4]. Some PPR proteins have been reported to be responsible for the splicing of group II introns in maize mitochondria ( Table 2). Loss-of-function of these PPR proteins usually results in empty pericarp (emp) or defective kernel (dek) phenotype in maize. For example, disruption of Dek2 and Dek37 causes small kernels and delayed development, leading to a dek phenotype of maize [35,38]. The mutation of Emp16 and Emp10 severely arrests seed development, resulting in embryo lethality and an emp phenotype of maize [25,39]. Similarly, the loss of PPR18 function severely arrests the embryo and endosperm development (Figure 2), leading to the emp phenotype, suggesting that PPR18 is crucial for maize kernel development. Table 2. List of mitochondrion-localized PPR proteins are involved in intron splicing of mitochondrial genes in maize.

Protein
Target Transcript Reference As shown in Table 2, these PPR proteins (EMP8, EMP10, EMP11, EMP12, EMP16, EMP602, DEK2, DEK35, DEK37, DEK41, DEK43, PPR-SMR1, and PPR20) are reported to participate in the splicing of nad1, nad2, nad4, nad5, nad7, and rps3 introns. These nad genes encode the core subunits of mitochondrial complex I and lack of these Nad proteins causes the disassembly and reduced activity of complex I in maize [51]. For example, EMP11 is responsible for the intron splicing of nad1 and the dysfunction of Emp11 affects the assembly and stability of mitochondrial complex I [34]. EMP12 and PPR20 are essential for the splicing of nad2 introns. The mutation of EMP12 and PPR20 results in disassembly of mitochondrial complex I and a significant reduction in the dehydrogenase activity [36,37]. In this study, loss of PPR18 impaired the cis-splicing of nad4 intron 1 and the accumulation of mature nad4 transcript ( Figure 5), leading to severely defective assembly and activity of mitochondrial complex I in the ppr18 mutants, suggesting the importance of PPR18 in the intron splicing of nad4 transcript and the assembly and activity of mitochondrial complex I. The splicing of nad4 intron 1 was dramatically reduced in dek35, and completely abolished in emp8, emp602, and ppr18 [46,49,50], causing a deficiency in the mature nad4 transcript and severely arrested embryo and endosperm development in maize. Together, these results indicate that expression of nad4 is essential to the mitochondrial function and kernel development in maize. A co-evolution between the mitochondrial nad4 intron 1 and the nuclear PPR18 genes are implicated as indicated by the highly conserved sequences between PPR18 and its putative binding site in nad4 intron 1 in both monocots and dicots (Figures S1 and S7).
In Figure 4, the in-gel NADH activity assay showed that two bands with a smaller size than mature complex I were produced in ppr18 alleles, which are probably partially assembled complex I. Additionally, the two smaller size complexes had the dehydrogenase activity ( Figure 4B), indicating the two partially assembled complex I are stable intermediate of the mitochondrial complex I assembly pathway in maize. Similar cases were also found in the dek35, emp602, dek41, and dek43 mutants, which are defective in the splicing of nad4 introns and [49,50,52,53], supporting that Nad4 is assembled into the complex I at a late stage. As reported in Arabidopsis, Nad4 is located in the P D module of the membrane arm of mitochondrial complex I, which is associated with assembled intermediate to form the mature complex I [51], indicating that Nad4 plays a crucial role during the last phase of the complex I assembly process both in monocots and dicots.
Previous reports showed that impairments in the electron transfer chain (ETC) can enhance AOX pathway in mitochondria [54][55][56]. In the ppr18 alleles, the ratio of V alt /V t , the abundance of AOX protein, and the expression of AOX2 and AOX3 transcripts were notably increased (Figure 3), indicating that AOX pathway is significantly enhanced in ppr18. A retrograde signalling pathway is strongly implicated as the AOXs are nucleus-encoded proteins.

Multiple Splicing Factors Participate in the Splicing of nad4 Intron 1
The defective splicing of nad4 intron 1 was firstly reported in Nicotiana sylvestris nms1 mutant [57]. In maize, defects of splicing of nad4 intron 1 have been found in some ppr mutants, e.g., dek35, emp8, and emp602 [46,49,50]. Our data show that PPR18 also specifically functions on the splicing of nad4 intron 1 in maize. These splicing factors share common intron target with PPR18, implying that splicing of a single intron requires multiple splicing factors. Based on the yeast two-hybrid analyses, however, PPR18 does not directly interact with DEK35, EMP8, and EMP602 ( Figure S8).
It is possible that these PPR proteins function independently by binding to the specific sequences of nad4 intron 1 to maintain a splicing-competent structure, and these proteins do not have protein interactions with each other. Proteins could promote group II intron folding to form the native structure [58]. PPR proteins are RNA binding proteins that could guide intron folding by sequence-specific interactions [59]. PPR18 may play a role in folding of nad4 intron 1 to participate in the intron splicing of nad4 transcript.
As the intron splicing mechanism in mitochondria is not clear yet, all co-factors may have not been identified. The possibility of a ribonucleoprotein complex similar to the nuclear spliceosome exists for some introns. Thus, it is possible that PPR18 may interact with other splicing co-factors. The tested PPR proteins may interact with some key splicing factors exist in splicing complexes to mediate splicing of nad4 intron 1. For example, PPR-SMR1 interacts with Zm-mCSF1 to participate in the splicing of several mitochondrial group II introns, speculating that PPR-SMR1 and Zm-mCSF1 might be the core splicing factors to mediate multiply group II introns splicing [60]. Further studies are necessary to unravel the splicing mechanism of plant organellar group II introns.

Plant Materials
The ppr18 alleles (UFMu-06715 and UFMu-11033) were obtained from the Maize Genetics Cooperation Stock Center (Urbana, IL, USA). For developmental analyses and population generation, the maize plants were cultivated in the experimental filed of Shandong University in Qingdao. Tobacco (Nicotiana tabacum) was grown in climate chambers at 25 • C/22 • C day/night on a 12 h light/12 h dark regime.

Subcellular Localization
To express PPR18 N550 -GFP fusion proteins, the amplified PPR18 N-terminal coding sequence (1650 bp) was cloned into binary vector pGWB5 driven by the 35S promoter. Agrobacterium tumefaciens strain (EHA105) harboring this construct was infiltrated into tobacco (Nicotiana tabacum) leaf epidermis with a syringe, as previously described [61]. After incubation at 24 • C for 24-30 h, the GFP signals were observed and imaged using ZEISS LSM 880 confocal microscope (Carl-Zeiss, Jena, Germany). MitoTracker Red (ThermoFisher Scientific, Waltham, MA, USA) was used as the mitochondrion marker with a working concentration of 100 nM.

Light Microscopy of Cytological Sections
ppr18-1/+ heterozygotes were identified by PCR with Mu primer TIR8 and gene primer PPR18-R1. Immature WT and ppr18-1 kernels were harvested from the self-pollinated heterozygous segregating ear at 9 and 14 days after pollination (DAP). The fixed material and sections were performed as described previously [62]. Paraffin sections were stained with 1% Johansen's Safranin O and imaged with a stereo microscope (Carl-Zeiss, Jena, Germany).

RNA Extraction, RT-PCR, and qRT-PCR
Total RNA was extracted from fresh kernels by removing the pericarp using TRIzol reagent (ThermoFisher Scientific, Waltham, MA, USA). After DNaseI digestion (NEB) to eliminate DNA contamination, reverse transcription was conducted with random primers according to the manufacturer's instructions (TransGen Company, Beijing, China). Quantitative real time PCR (qRT-PCR) was performed with using LightCycler 96 (Roche, Basel, Switzerland) with three biological replicates. The maize actin gene ZmActin (GRMZM2G126010) was used as the reference gene. For functional analysis of PPR18, RT-PCR and qRT-PCR were performed with primers as previously described [45,49].

Measurements of Respiration Rate
The respiration rates were determined according to the previous report with some modifications [63]. The respiratory activities were measured in a reaction medium (50 mM phosphate buffer, pH 6.8) at 25 • C in the dark using a Chlorolab II liquid oxygen electrode (Hansatech, King's Lynn, UK, http://hansatech-instruments.com). The alternative pathway capacity (V alt ) and cytochrome pathway capacity (V cyt ) are defined as O 2 uptake rate in the presence of 2 mM potassium cyanide (KCN; Sigma-Aldrich, St. Louis, MO, USA, catalog number: 207810) and 2 mM salicylhydroxamic acid (SHAM; Sigma-Aldrich, St. Louis, MO, USA, catalog number: S607), respectively. All the respiration rates were indicated by the oxygen consumption of nmol O 2 min −1 g −1 fresh weight of the maize kernels.

Blue Native (BN)-PAGE and Complex I Activity Assay
Crude mitochondrial proteins were extracted from immature maize kernels with the pericarp removed at 11 DAP. Blue native (BN)-PAGE and measurement of NADH dehydrogenase activity were performed as previously described [64]. Of maize mitochondrial proteins 130 µg was separated by 3%-12.5% gradient gel (ThermoFisher Scientific, Waltham, MA, USA). The gel strips were stained by Coomassie Blue R-250 and assayed for complex I activity in nitroblue tetrazolium (NBT)-NADH buffer (25 mg of NBT, 100 µL of NADH (10 mg mL −1 ), and 10 mL of 5 mM Tris-HCl, pH 7.4). The gel strips were transferred to the nitrocellulose membrane and Western blotting with specific antibodies cytochrome c 1 (Cyt C1 , a gift from G. Schatz, University of Basel, Switzerland), Arabidopsis Cox2 (Agrisera, Vännäs, Sweden, http://www.agrisera.com), and ATPase (ATPase α-subunit, MBL Beijing Biotech, China) for detection of complex III, VI, and V, respectively [64].

Yeast Two-Hybrid Analysis
Yeast two-hybrid analysis was performed according to the manual of Matchmaker™ Gold Yeast Two-Hybrid System (Clontech, Mountain View, CA, USA). The coding sequence (CDS) without signal peptide sequences of PPR18, DEK35, EMP8, and EMP602 were cloned pGADT7 vector and pGBKT7 vector, respectively. The primer sequences are listed in Table S1. Combinations of plasmids were co-transformed into the yeast strain Y2H Gold (Clontech, Mountain View, CA, USA) and placed on SD/-Leu/-Trp (Minimal Media Double Dropouts, DDO) mediums and growth of diploid yeast colonies on SD/-Ade/-His/-Leu/-Trp (Minimal Media Quadruple Dropouts, QDO) mediums for 4 days at 30 • C to reveal protein-protein interactions.

Prediction of PPR18 Binding Site
PPR motif prediction alignment analysis of PPR18 protein was used by algorithm TPRpred (http://tprpred.tebingen.mpg.de/tprpred). The alignment of PPR18 to its nad4 intron 1 binding site was generated as follows the PPR codes [29][30][31]. The recognition nucleotides for the PPR18 protein were predicted by the arrangements of the amino acids at position 5th and 35th of each PPR repeat from PPR18. The respective recognition nucleotides were listed and aligned with nad4 intron 1 to find the best matched nucleotide sites as potential binding sites of PPR18 in mitochondrial nad4 intron 1.

Phylogenetic Analysis
The full-length amino acid sequences of PPR18, the putative binding site of PPR18 in nad4 intron 1, and their orthologs in plant species were downloaded from NCBI database (https://blast.ncbi.nlm. nih.gov). The phylogenetic tree was constructed using MEGA7 software by the maximum likelihood method [66].

Conclusions
In this study, we characterized a maize seed mutant ppr18, which exhibits an arrested embryo and endosperm development phenotype. Through a molecular characterization of the PPR18 gene, we elucidated its function in the cis-splicing of nad4 intron 1 in mitochondria and seed development in maize. The lack of splicing of nad4 intron 1 results in the absence of nad4 transcript, leading to severely reduced assembly and activity of mitochondrial complex I. The profiles of complex I assembly, activity, and component accumulation in the ppr18 mutants shed lights to the assembly process of complex I in maize. Despite PPR proteins have been reported in intron splicing, our study provides additional information on a new PPR protein in intron splicing, complex I assembly, and its essential role in maize seed development.
Supplementary Materials: Supplementary materials can be found at http://www.mdpi.com/1422-0067/21/11/ 4047/s1, Figure S1: Phylogenetic analysis of PPR18 homologs, Figure S2: Phenotypes of the ppr18-2 mutant and the ear of ppr18-1 × ppr18-2, Figure S3: The linkage analysis of the Mu insertion in PPR18-1 and the phenotype in the selfed progeny, Figure S4: Expression of PPR18, Figure S5: Transcript levels of the 35 mitochondrial genes in ppr18 alleles in developing kernels, Figure S6: The predicted secondary structure of maize nad4 intron 1 and the location of the putative binding site, Figure S7: Alignment of the putative binding site of PPR18 in nad4 intron 1 in different plant species, Figure S8: Interaction assay of PPR18 and the related splicing factors, Table S1: Primers used in this study.