A novel E-Subgroup Pentatricopeptide Repeat Protein DEK55 is Responsible for RNA Editing at 15 Sites and Splicing of nad1 and nad4 in Maize

Background Pentatricopeptide repeat (PPR) proteins is a large protein family, which participate in RNA processing in organelles and plant growth. Previous reports have generally considered E-subgroup PPR proteins as an editing factors for RNA editing. However, the underlying mechanisms and effects of E-subgroup PPR proteins remain to be investigated. Results In this study, we recognized and identied a new maize kernel mutant with arrested embryo and endosperm development, defective kernel 55 (dek55). Genetic and molecular evidences suggest that the defective kernels resulted from a mononucleotide alteration (C to T) at + 449 in the open reading frame (ORF) of Zm00001d014471 (hereafter referred to as DEK55). DEK55 encodes an E-subgroup PPR protein within mitochondria. Molecular analyses suggest that DEK55 plays crucial roles in RNA editing at multiple sites of ribosomal protein S13, ATP synthase subunit1, NADH dehydrogenase-6 (nad6), and nad9 transcripts as well as in splicing of nad1 and nad4. The mutation of DEK55 lead to the dysfunction of mitochondrial complex I. Conclusions Our results demonstrate that the DEK55 mutation is responsible for the dek55 mutant phenotypes, as it affects mitochondrial function that is essential for maize kernel development. This study also provides novel insight into the molecular function of E-subgroup PPR proteins in plant organellar RNA metabolism. DEK55 for editing at rps13-56, atp1-1490, nad6-159, and 12 nad9 editing sites. Meanwhile, DEK55 is responsible for the trans-splicing of two nad1 introns (intron 1 and intron 4) and cis-splicing of the nad4 intron 1 in mitochondria. Our results suggest that the E-subgroup PPR protein DEK55 plays an important role in RNA editing and splicing of introns of maize mitochondrial transcripts. These results provide novel view for understanding the molecular function of E-subgroup PPR proteins in RNA processing in plant organelles.

Here, we identi ed the maize mutant dek55 with an embryo-lethal phenotype with arrested endosperm development, which is caused by the mutation of the mitochondria-localized E-subgroup PPR protein DEK55. In the dek55 mutant, the splicing e ciency of nad1 intron 1 and intron 4 trans-splicing and nad4 intron 1 cis-splicing were decreased. Moreover, the editing ratio of multiple editing sites in ribosomal protein S13 (rps13), ATP synthase subunit1 (atp1), NADH dehydrogenase-6 (nad6), and nad9 transcripts was also signi cantly reduced. Our results suggest that the E-subgroup PPR protein DEK55 participates in both RNA editing and group II intron splicing in maize mitochondria.

Results
Genetic and phenotypic analysis of the dek55-1 mutant A mutant with a defective kernel phenotype was isolated from an ethyl methane sulfonate-induced maize B73 background population, named defective kernel 55 − 1 (dek55-1). The dek55-1 kernels were segregated from self-pollinated progenies of dek55-1/+ heterozygotes in a 1:3 ratio (Fig. 1a, Additional le 1: Table S1). This mutant was con rmed in other populations generated from dek55-1/+ heterozygotes crossed with the inbred lines C733 or S162 (Additional le 1: Table S1). These results suggest that dek55-1 as a recessive phenotype is caused by a monogenic mutation.
The dek55-1 kernels could be distinguished from wild type (WT) kernels at 15 days after pollination (DAP) (Fig. 1a). The dek55 mutant kernels exhibited a whitened pericarp and were smaller than WT kernels, which exhibited a yellow color (Fig. 1a). At the maturity stage, dek55-1 kernels became smaller and more shriveled (Fig. 1b, c). To further dissect the mutant phenotype, both WT and dek55-1 kernels were longitudinally sliced at different developmental stages. At 15 DAP, the pericarp of WT kernels, but not dek55-1 mutant kernels, was lled with endosperm cells (Fig. 1d, e). Furthermore, dek55-1 exhibited a smaller mature embryo and a decreased proportion of hard endosperm than that in WT (Fig. 1f, g). dek55-1 kernels could not germinate in the experimental eld (0/100), implying that the arrested embryo was lethal in mutants. In addition, the kernel weight of dek55-1 was reduced by approximately 70% compared to that of WT kernels (Fig. 1h).
To further investigate the developmental structure of dek55-1 kernels, we examined the tissue structure of WT and dek55-1 kernels at 12 and 18 DAP (Fig. 1i − l). At 12 DAP, WT embryos contained visible coleoptiles, shoot apical meristems, scutella, and two leaf primordia (Fig. 1i). In contrast, dek55-1 embryos only had a small scutellum that was arrested at the coleoptile stage (Fig. 1j). Moreover, WT kernels were lled with endosperm cells, whereas a large interspace between endosperm and seed coat in dek55-1 was observed (Fig. 1i, j). At 18 DAP, WT embryos had developed into relatively complete structures containing four leaf primordia, shoot apical meristems, and a clearly seen root apical meristem (Fig. 1k), while dek55-1 embryos only generated one leaf primordium (Fig. 1l). Less starch grains were accumulated in dek55-1 than in WT endosperm cells at this stage (Fig. 1k, l). In addition, a cavity was observed in dek55-1 endosperm (Fig. 1l). These results indicate that developmental defects in embryo and endosperm are present in dek55-1 mutants.

Map-based cloning of DEK55
To identify the DEK55 gene, we performed the classical map-based cloning strategy to detect F 2 mutant kernels, which were segregated from self-pollinated lial 1 (F 1 ) hybrid ear. Four genomic DNA pools (10 mutant kernels per pool) and both parents were used for correlation analysis with polymorphic simple sequence repeats. The six simple sequence repeats at chromosome 5 were highly correlated with defective kernel phenotypes, implying that the candidate gene may be at chromosome 5. Further analysis showed that the DEK55 gene is located between umc1705 and umc2302 on chromosome 5 (Fig. 2a). Six polymorphic molecular markers in this region were used to analyze 1868 mutant kernels from the F 2 population. Finally, the DEK55 gene was located on an approximately 1.29 Mb region between molecular label M3 and M4 (Fig. 2a). There are 25 putative protein-coding genes in this region (http://ensembl.gramene.org/Zea_mays/Info/Index). To identify the mutated genes, genomic DNA of 25 candidate genes were ampli ed and sequenced. This revealed that the E-subgroup PPR protein gene (Zm00001d014471) has a single nucleotide change (C to T) at + 449 in dek55-1, which might result in an amino acid replacement (Ser to Phe) in the protein sequence but not in expression level of DEK55 change ( Fig. 2a − d). To validate this result, we obtained a new mutant, dek55-2, from the maize ethyl methane sulfonate-induced mutant database [34]. The dek55-2 mutant had a single nucleotide mutation (G to A) at + 729 bp (Fig. 2b), which leads to protein truncation (Fig. 2d). The mutant dek55-2 also exhibited defective kernels with small and white pericarps (Fig. 2e). The allelic test between dek55-1 and dek55-2 heterozygotes revealed that normal and mutant kernels were segregated with the expected 3:1 ratio (normal/mutant; 450/143; P = 0.62) in the F 1 ear (Fig. 2e). As a control, all the kernels from the ear that were crossed between dek55-2 heterozygote and WT were normal (Fig. 2e). These results indicate that the PPR gene Zm00001d014471 mutation is responsible for defective kernel phenotype, and the annotated gene was designated DEK55.

Dek55 Is A Mitochondrial E-subgroup Ppr Protein
Sequence alignment demonstrated that the DEK55 gene has one exon containing an 1893 bp ORF, which encodes a 630 amino acid residue protein containing 13 PPR motifs and an E domain on the carboxyterminal end (Fig. 2b − d and Additional le 1: Fig. S1). Mutated sites in dek55-1 and dek55-2 were located in the third and fth PPR motifs, respectively (Fig. 2d). The mutation in dek55-2 resulted in a truncated DEK55 protein without the last eight PPR motifs or E domain.
To examine the subcellular localization of DEK55, the p35S:DEK55-EGFP vector was constructed and transformed into maize protoplasts. The uorescence signal of DEK55-EGFP overlapped with Mito Tracker, which is a mitochondria-speci c dye (Fig. 3a), suggesting that the DEK55 protein is a mitochondrial PPR protein in maize (Fig. 3a). In addition, DEK55 expression analysis in various maize tissues demonstrated that DEK55 is highly expressed in root, anther, and ear, but lowly expressed in stem, leaf, silk, tassel, and kernel (Fig. 3b).
DEK55 is essential for the trans-splicing of nad1 introns 1 and 4 and for the cis-splicing of nad4 intron 1 The transcript levels of 35 maize mitochondrial genes were examined, and nad1 and nad4 were signi cantly downregulated in the dek55 mutant (Fig. 5a). The genomic DNA of nad1 contains four group II introns; intron 2 is a cis-splicing intron and the others are trans-splicing introns (Fig. 5c). The genomic DNA of nad4 has three cis-splicing introns (Fig. 5d) [13,35]. The full maturation of nad1 and nad4 transcripts requires complete intron splicing. We further detected the intron splicing e ciency of nad1, nad4, and other genes in WT and dek55-1 by qRT-PCR. Compared with that in WT, the splicing e ciency of the rst and fourth introns of nad1 and the rst intron of nad4 in dek55-1 mutant were decreased (Fig. 5b). Furthermore, we ampli ed each intron and full transcripts of nad1 and nad4 by RT-PCR (Fig. 5c, d). The transcriptional abundance of nad1 exon 1-2, exon 4-5, and full-length DNA fragments were signi cantly decreased (Fig. 5c). The unspliced DNA fragments (1F + 2R, 3F + 4R, 4F + 5R) were not ampli ed by RT-PCR in WT and dek55, as nad1 introns 1, 3, and 4 are too long (Fig. 5c). The unspliced intron 2 fragments of nad1 in dek55 mutants were similar to those in WT (Fig. 5c). The abundance of nad4 spliced exon 1-2 and full-length DNA transcript fragments were signi cantly decreased, and the abundance of nad4 unspliced intron 1 transcript was signi cantly increased (Fig. 5d). This suggests that the signi cant decrease in the nad4 and nad1 transcript abundance in dek55 mutants was caused by the abnormal splicing of nad4 intron 1, nad1 intron 1, and intron 4, respectively ( Fig. 5a − d). Therefore, DEK55 is necessary for the trans-splicing of two nad1 introns (1 and 4) and cis-splicing of the rst nad4 intron in maize.

dek55-1 mutant exhibits reduced complex I activity and increased alternative respiratory pathway activity
The four genes nad1, nad4, nad6, and nad9 encode the subunits of complex I NAD1, NAD4, NAD6, and NAD9, respectively [35]. The rps13 gene encodes ribosomal protein, and atp1 encodes the ATP1 subunit of ATP synthase F1 [35]. Defects in post-transcriptional processing of these genes may impair the biosynthesis of mitochondrial complexes [17,[36][37][38]. We performed blue native polyacrylamide gel electrophoresis (BN-PAGE) and the in-gel NADH dehydrogenase activity assay to investigate the accumulation level and activity of mitochondrial complexes in WT and dek55-1 endosperm. BN-PAGE showed that the abundance of complex I and super-complex I + III 2 in dek55-1 mutants signi cantly decreased (Fig. 6a). However, no signi cant differences were observed in the abundance of complex V between WT and dek55-1 (Fig. 6a). Furthermore, dek55-1 de ciency the activities of complex I and I + III 2 (Fig. 6b). These results indicate that defects in mitochondrial transcript splicing and editing might affect the abundance and activity of mitochondrial complex I.
The mitochondrial respiratory chain in plants contains the cytochrome c and alternative oxidase (AOX) pathways [39]. When the main cytochrome c pathway is blocked, AOX activity can be increased to compensate respiration pathways [40]. In dek55-1, the functions of complex I were abolished (Fig. 6a, b). Thus, we performed qRT-PCR to detect the expression levels of Aox genes in WT and dek55-1. The expression of the Aox2 gene was increased approximately 512-fold in the dek55-1 mutant (Fig. 6c). Taken together, our results indicate that the respiration pathway is severely blocked in dek55-1 mitochondria.

Discussion
DEK55 is required for maize kernel development Previous reports have shown that PPR proteins play important roles in maize kernel development and that the loss of function of some PPR proteins results in empty pericarp as well as small and defective kernel phenotypes of different genetic backgrounds [13, 14, 17-21, 30, 31, 38, 41, 42]. These ppr mutants exhibit developmentally arrested embryos and endosperm. The embryos usually reached the coleoptilar stage or the leaf stage 1 (L1), and endosperm exhibited signi cantly reduced starch and protein levels [14,33,42]. The dek55 mutant kernels exhibited a shriveled pericarp and small size (Fig. 1a − c). The dek55-1 mutant kernels also exhibited smaller embryos and a decreased proportion of hard endosperm compared with WT ( Fig. 1d − l). In particular, dek55-1 embryos were severely arrested and only had one leaf primordium. Thus, the mutant kernel could not germinate in the eld. Allelic tests indicated a nonsense mutant dek55-2, an allelic mutant with dek55-1 and dek55-1/dek55-2 heterozygous kernels, which displayed a similar phenotype to dek55-1. This suggests that dek55 dysfunction is responsible for defective kernel phenotype and that DEK55 is required for kernel development in maize.
The E-subgroup PPR proteins are characterized by an E domain on the carboxy-terminal end that might be responsible for interactions between proteins [43,44]. In the dek55-1 mutant, there is a single nucleotide change (C to T) at + 449 in dek55, which caused phenylalanine (Phe) to replace serine (Ser) on the third PPR motif of DEK55 at 150 amino acid sites (Fig. 2b, d; Additional le 1: Fig. S1). This mutation altered the a nity of amino acids to water, from hydrophilic to hydrophobic amino acids. Our evidence suggests that the amino acid change (Ser→Phe) is responsible for defective kernels in dek55 mutants. Therefore, the amino acid change at this site in DEK55 might cause altered conformation and function loss. In dek55-2 mutants, the mutation resulted in a loss of the last eight PPR motifs and E domain on the carboxy-terminal end of the DEK55 protein (Fig. 2b, d), which might prevent it from forming complexes with other proteins and from binding to targets.
DEK55 is necessary for C-to-U editing of multiple sites in the mitochondrial transcripts of maize PPR proteins, including DYW2, EMP21, NUWA, and MEF8, are involved in C-to-U editing at multiple sites [45][46][47][48]. In this study, DEK55 participated in RNA editing at 15 sites, suggesting it might be a novel Esubgroup PPR protein. However, PPR proteins do not share uniform protein features. Among them, DYW2 and MEF8 harbor only ve PPR repeats and belong to an atypical DYW subgroup [45,46]. NUWA belongs to the P-class of PPR proteins [45,47]. EMP21 contains 11 PPR-motifs in addition to E and DYW domains and belongs to PPR-DYW proteins [48]. DEK55 is considered as an E-subgroup PPR protein that contains the canonical E domain. Therefore, PPR proteins that target multiple sites for editing might not share similar structures.

DEK55 is involved in group II intron splicing in maize mitochondria
The E-subgroup PPR proteins are usually considered to be editing factors for RNA editing in organelles [10], whereas few of these proteins are considered to play a role in splicing [28,29,52]. SLO4 is necessary for RNA editing of nad4 and the e cient splicing of nad2 intron 1 in Arabidopsis mitochondria [29]. AEF1/MPR25 not only participates in RNA editing of atpF and nad5, but also modulates atpF splicing in both Arabidopsis and rice [28]. Furthermore, the plastid PPR protein OTP70 only participates in the intron splicing of the rpoC1 transcript [52]. Interestingly, in this study, DEK55 (an E-subgroup PPR protein) participated in both RNA editing of 15 sites and group II intron splicing in maize mitochondrial transcripts (Figs. 4, 5a − d). However, the RNA editing of nad1 and nad4 transcripts was not affected. It has been reported that intron splicing is usually mediated by RNA editing events, in which the key sites of introns are edited [53,54].
Some proteins have been identi ed that participate in the splicing of nad1 and nad4 pre-RNAs. The nuclear maturases 1 [55], DEK2 [42], and EMP11 [41] participate in trans-splicing of nad1 intron 1. EMP11, EMP8, and ZmSMK3 are required for nad1 intron 4 trans-splicing [41]. The proteins include NMS1 [56], DEK35 [19], EMP8 [13], DEK43 [20], EMP602 [57], and ZmSMK3 [58] and are implicated in cis-splicing of nad4 intron 1. In our study, we have demonstrated that DEK55 is involved in both transand cisdual splicing. It appears that the splicing of one intron possibly requires the involvement of multiple factors to constitute a putative spliceosome. This is supported by the nding that PPR-small MutS-related-1 can interact with the Zm-mCSF1 formation protein complex. The protein complex participates in the intron splicing of multiple transcripts within mitochondria [59]. Therefore, DEK55 might interact with P-type PPR proteins or other splicing factors involved in group II intron splicing.

Conclusions
In this study, we have demonstrated that DEK55 is a mitochondria-localized E-subgroup PPR protein.
Mutation of DEK55 lead to embryo-lethal and arrested endosperm development phenotype in maize. DEK55 is required for editing at rps13-56, atp1-1490, nad6-159, and 12 nad9 editing sites. Meanwhile, DEK55 is responsible for the trans-splicing of two nad1 introns (intron 1 and intron 4) and cis-splicing of the nad4 intron 1 in mitochondria. Our results suggest that the E-subgroup PPR protein DEK55 plays an important role in RNA editing and splicing of introns of maize mitochondrial transcripts. These results provide novel view for understanding the molecular function of E-subgroup PPR proteins in RNA processing in plant organelles.

Plant materials
Maize dek55-1 was identi ed from ethyl methane sulfonate populations from a B73 background, which was provided by Prof. Xiaoduo Lu of Qilu Normal University. The allele mutant dek55-2,which original material name is EMS4-073342. The EMS4-073342 was purchased from the maize ethyl methane sulfonate-induced mutant database (http://www.elabcaas.cn/memd/) [34] which can be found by searching for gene ID (Zm00001d014471). To purify the genetic background of the dek55-1 mutant, dek55-1 was crossed into the B73 inbred line twice to harvest BC 2 F 1 . BC 2 F 2 kernel was used for further experiments. The dek55-1 heterozygote as the male parent was crossed with our laboratory inbred lines C733 and S162, then F 1 progenies were self-pollinated to generate the F 2 population that was used for gene mapping. Ru Chang Ren and Xu Wei Yan undertook the formal identi cation of the plant materials. All plant materials were planted in the experimental station of Shandong Agricultural University (Taian, Shandong province).

Histological Analysis
The WT and defective kernels were obtained from the self-pollinated heterozygous plant at 12 and 18 DAP, respectively. The middle part of the kernel along the longitudinal axis was selected and placed in formalin-acetic acid-alcohol solution for at least 12 h on ice. This was followed by treatment with 50%, 70%, 85%, 95%, and 100% ethanol as well as 100% xylene for 2-4 h. After dehydration, materials were treated in para n for 72 h at 60 °C and then embedded in para n. The para n-embedded samples were cut into 12 µm thin slices using a microtome (Leica RM2235, Germany). Section staining was performed based on the methods of Ren et al. (2020). Finally, the sections were photographed with a light microscope (OLYMPUS DP72).

Map-based Cloning
The DEK55 locus was identi ed using 1868 F 2 defective kernels from the self-pollinated F 1 population (C733 × dek55-1/+). For preliminary mapping, 73 polymorphic simple sequence repeat markers were selected from the entire genome with which the parent F 1 individual plant and F 2 defective kernel DNA pools were analyzed. New molecular markers were selected according to the parent DNA sequences used for ne mapping. The website (http://ensembl.gramene.org/Zea_mays/Info/Index) was used to search for gene annotations between candidate regions in Zea mays (B73_RefGen_v4) [60]. Phanta EVO Super-Fidelity DNA polymerase (Vazyme Code: P503-d1) was used to clone all candidate gene genomic DNA sequences and sequencing. The primers were designed according to candidate gene reference sequences. The primer sequences for cloning of full length DEK55 genomic DNA and map-based cloning are shown in Additional le 1: Table S2.
Rna Extraction, Rt-pcr, And Qrt-pcr Total RNA of WT and dek55 mutant kernels without pericarp and other tissues were extracted with the Ultrapure RNA Kit (CWBIO, China). The residual DNA in the total RNA was removed by DNase. Complementary DNA was obtained by reverse transcription. RT-PCR was performed to amplify mitochondrial transcripts, splicing e ciency of nad1, and nad4 introns. The DNA fragments obtained by RT-PCR were directly sequenced and shifted into the Escherichia coli strain (TOP10) for monoclonal sequencing. The transcripts were ampli ed according to the primers previously reported [61]. Primers used to amplify introns of nad1 and nad4 are shown in Additional le 1: Table S2.
The qRT-PCR equipment and reaction system were used according to a previous report [20]. All qRT-PCR assays were performed with three samples and technical repeats. The primers were designed for group II intron splicing e ciency analysis in mitochondria according to previous reports [17,18,61]. The primers used to analyze DEK55 expression levels are shown in Additional le 1: Table S2.

Subcellular Localization
The termination codon was removed from the whole coding sequence of DEK55 and imported into the pM999-EGFP vector generating a DEK55-EGFP recombinant vector driven by the CaMV 35S promoter. The subcellular localization experiment was performed as previously described [62]. Protoplasts of maize mesophyll were obtained from etiolated leaves by enzymatic hydrolysis as described previously [21]. Recombinant vector (20 µL, 15-20 µg) was added into maize protoplasts (200 µL), 220 µL of 40% (w/v) PEG4000 solution was added and mixed completely, and then the sampled were incubated at 23 °C for 10-15 min. Afterwards, the protoplasts were washed with W5 or WI solution and cultured for 12-16 h in the dark at 23 °C. The mitochondria in the protoplasm were labeled by a speci c probe (MitoTracker Red CMXRos, Thermo Fisher Scienti c), and images were acquired with a laser confocal microscope (LSM 880, Zeiss).

Isolation And Analysis Of Mitochondrial Complexes
The plant mitochondrial isolation kit (Biohao, Wuhan; catalog no. P0045) was used to isolate crude mitochondria from WT and dek55-1 seeds with removed pericarps (on 15 DAP) for analysis of BN-PAGE and complex I activity. The collected mitochondrial precipitate was redissolved in 35 µl of solution buffer and then kept on ice for 30 min. Afterwards, the suspension was centrifuged at 4 °C, the supernatant was collected and loaded on pre-prepared gradient gels (BN1002BOX, Thermo Fisher Scienti c), and electrophoresis was performed according to manufacturer's instructions. Next, the gels were placed in 100 mL of xing solution (methanol/ddH 2 O/acetic acid, 4:5:1) for 30 min and then transferred to 0.02% Coomassie R-250 stain (Sigma-Aldrich) for analysis of mitochondrial complex abundance. The gel strips were incubated in assay buffer for 5 min, and the reaction was terminated with the xing solution for analysis of complex I activity [41].

Consent for publication
Not applicable.

Availability of data and materials
All data generated or analyzed during this study are included in this published article and its supplementary information les.