EMPTY PERICARP11 serves as a factor for splicing of mitochondrial nad1 intron and is required to ensure proper seed development in maize.

Group II introns are common in the mitochondrial genome of higher plant species. The splicing of these introns is a complex process involving the synergistic action of multiple factors. However, few of these factors have been characterized in maize. In this study, we found that the Empty pericarp11 (Emp11) gene, which encodes a P-type pentatricopeptide repeat (PPR) protein, is required for the development of maize seeds. The loss of Emp11 function seriously impairs embryo and endosperm development, resulting in empty pericarp seeds in maize, and alteration in Emp11 expression leads to quantitative variation in kernel size and weight. We found that the emp11 mutants showed a failure in nad1 intron splicing, exhibited a severe reduction in complex I assembly and activity, mitochondrial structure disturbances, and an increase in alternative oxidase AOX2 and AOX3 levels. Interestingly, the emp11 phenotype was very severe in the W22 inbred line but could be partially recovered in B73 BC2F2 segregating ears. These results suggest that EMP11 serves as a factor for the splicing of mitochondrial nad1 introns and is required for mitochondrial function to ensure proper seed development in maize.

Introduction

Mitochondria are key players in plant development, fitness and reproduction. They perform a variety of fundamental functions, for example in pyruvate oxidation, the Krebs cycle, and the metabolism of amino acids, fatty acids, and steroids, with the most crucial being the production of ATP by oxidative phosphorylation (Mackenzie and McIntosh, 1999; Gualberto ). Although most mitochondrial proteins are encoded by nuclear loci, the mitochondrion is a semi-autonomous organelle with its own genome. The maize NB (Normal type in nuclear background B73) mitochondrial genome encodes 58 genes incorporating 22 introns, which all belong to group II introns based on their distinctive structures (Clifton ; Bonen, 2008). Group II intron splicing depends on various protein factors including maturases (MATs), which are encoded by group II introns in bacteria and in yeast (Singh ; Mohr and Lambowitz, 2003). The MATs can also be encoded by nuclear genomes of plant species, such as the four maturases, nMAT1–4, in Arabidopsis. (Mohr and Lambowitz, 2003; Keren ; Keren ; Cohen ).

Pentatricopeptide repeat (PPR) proteins, a large protein family in plant species, also play constitutive and essential roles in the diverse processes of organelle RNA processing, including RNA editing, cleavage, splicing, and stability (Lurin ; Schmitz-Linneweber and Small, 2008; Fujii and Small, 2011; Barkan and Small, 2014). Mutations in PPR genes result in distinct physiological and morphological defects, indicating the importance of PPR proteins for plant growth and development, especially for embryo and endosperm development (Barkan and Small, 2014). For example, OTP43, a P-type PPR protein, is required for trans-splicing of the mitochondrial nad1 intron 1 in Arabidopsis. otp43 mutants show severe defects in seed development, germination, and plant growth (Falcon de Longevialle ). Four P-type PPR proteins have been reported to be involved in the splicing of mitochondrial genes in maize. EMPTY PERICARP16 is required for mitochondrial nad2 intron 4 cis-splicing in maize (Xiu ). Plants with mutations in EMP16 show an empty pericarp phenotype, reduced complex I assembly and activity, increased accumulation of complex III, and increased expression of alternative oxidase AOX2 (Xiu ). DEK35 is required for the cis-splicing of mitochondrial nad4 intron 1. The dek35 mutant exhibits impaired mitochondrial structure and delayed seed development (Chen ). A third PPR protein, Dek2 targets mitochondria as well; the dek2 mutation reduces the splicing efficiency of mitochondrial nad1 intron 1 and also leads to small kernels (Qi ). Recently, EMP10 was found to be involved in the splicing of mitochondrial genes, which affects the cis-splicing of nad2 intron 1 and seed development in maize (Cai ). Although the PPR family has many members, there seems to be little redundancy between different family members (Barkan and Small, 2014). The maize genome encodes hundreds of PPR proteins, however, the fundamental molecular functions of the majority of PPR proteins are still unknown. Here, we report the biological functions of the P-type PPR gene Emp11 in maize. Emp11 is required for the splicing of all nad1 introns. The loss of Emp11 function leads to a severe reduction in complex I assembly and activity and an increase in alternative oxidase AOX2 and AOX3 levels. The emp11 mutants are arrested in both embryo and endosperm development, resulting in empty preicarp seeds, but the emp phenotype can be partially recovered in B73 BC2F2 segregating ears. In summary, we present a new PPR protein that differs from previously reported genes by being required for splicing of all nad1 introns in maize.

Materials and methods

Plant materials

The emp11-1 and emp11-2 mutants were isolated from two UniformMu lines requested from the UniformMu collection stocks (McCarty ): UFMu-08197 and UFMu-04323, respectively. The two UniformMu lines were backcrossed to the W22 inbred line into BC2. Mutator insertion loci were screened using gene-specific primers and Mutator-specific primers (McCarty ). Normally developed kernels in a segregating ear were randomly planted and the kernel phenotype of each progeny individual was then evaluated. Meanwhile, total DNA was extracted from a leaf of each individual and each kernel with the pericarp removed in the segregating ear, using a modified CTAB method (Shahzadi ), and was then used for genotype analysis and co-segregation analysis. In addition, emp11-1 heterozygous (+/emp11-1) individuals on the W22 background were crossed to the B73 line to develop a BC2F2 population where the +/emp11-1 and emp11-1/emp11-1 genotypes were selected and referred to as +/emp11-1 (B73) and emp11-1/emp11-1 (B73).

Cytological observation

Kernels at 6, 8, 12 and 15 d after pollination (DAP) were collected and then cut along the longitudinal axis. The cut kernels were fixed overnight in 4% paraformaldehyde (Sigma, Santa Clara, CA, USA), dehydrated in an ethanol gradient series (30, 50, 70, 85, 95, and 100% ethanol), and embedded in Paraplast Plus (Sigma, Santa Clara, CA, USA). The sample blocks were sectioned into 8 µm slices using a Leica RM2265 microtome (Leica Microsystems, Wetzlar, Hesse-Darmstadt, Germany) and stained with 0.5% toluidine blue O. Images were captured using a Leica MZFLIII microscope (Leica Microsystems, Wetzlar, Hesse-Darmstadt, Germany). Additionally, for TEM analysis, endosperms at 10 DAP were cut into 1 mm3 pieces. Fresh tissues were fixed overnight in 2.5% (w/v) glutaraldehyde in 0.1 M phosphate buffer at pH 7.4, fixed in 2% OsO4 in the same buffer, and then dehydrated and embedded in epoxy resin and SPI-812 (Structure Probe, Inc., West Chester, PA, USA). Ultra-thin sections obtained using a Leica UC6 ultra microtome (Leica Microsystems, Wetzlar, Hesse-Darmstadt, Germany) were stained with uranyl acetate and subsequently with lead citrate. The observations and recording of images were performed using a Hitachi H-7650 transmission electron microscope (HITACHI, Chiyoda-Ku, Tokyo, Japan) at 80 kV and a Gatan 832 CCD camera (Gatan, Pleasanton, CA, USA). The procedures were performed as described by Yi .

RNA extraction and gene expression analysis

To analyze the expression of Emp11, the developing tissues, including the root, stem, leaf, tassel, ear, silk, ovary, endosperm, and embryo were collected. Total RNA was extracted from plant tissues using Ambion Pure Link Plant RNA Reagent (Life Technologies, Invitrogen, Carlsbad, CA, USA) and was then reverse transcribed using M-MLV reverse transcriptase (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. Reverse transcription PCR (RT-PCR) was performed for expression pattern analysis of Emp11 with primer pairs RT-Emp11-3F and RT-Emp11-3F. Quantitative real time PCR (qRT-PCR) was completed using SYBR Select Master kit (Life Technologies, Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions with three biological replicates, and the maize actin gene (GRMZM2G126010) was used as the internal control. The relative expression levels were calculated using the comparative Ct method (Livak and Schmittgen, 2001). For mitochondrial and AOX gene expression, total RNAs were extracted from 15 DAP kernels of wild-type and two emp11 mutant alleles with the pericarp removed. The RNAs were treated with RNase-free DNase and then normalized against both total RNA and ZmActin (GRMZM2G126010). The primers used for mitochondrial gene RT-PCR (Fig. 4A and Supplementary Fig. S6 at JXB online) were reported by Xiu . All the primers were anchored to the 5ʹ-UTR and 3ʹ-UTR, near the ATG and stop codon, of every gene in order to test the full-length coding region.

Fig. 4.

Deficiency of mature mitochondrial nad1 transcripts in emp11 mutants. (A) RT-PCR analysis of transcript levels of 35 mitochondrial-encoded genes in wild-type and emp11-1 mutant plants. In each of the gels, the left lane is wild-type and the right lane is the emp11-1 mutant. The RNA was isolated from the same ear segregating for wild-type and emp11-1 mutants with the pericarp removed. (B) Structure of the maize mitochondrial nad1 gene. Introns 1, 3, and 4 are trans-spliced introns and intron 2 is a cis-spliced intron. The emp11 mutants were impaired in the splicing of nad1 intron 1, 2, 3, and 4. The expected amplification products using different primer pairs are indicated. The RNA was isolated from the same ear segregating for wild-type and emp11-1 mutants. (C) qRT-PCR analysis of splicing efficiency of all the 22 group II introns in emp11-1, emp11-2, and wild-type plants. The RNA was isolated from two mutant alleles and wild-type kernels at 15 DAP. Values represent the mean and standard deviation of three biological replicates. The expression levels were normalized to ZmActin (GRMZM2G126010). (D) RT-PCR analysis of nad1 intron splicing-related genes in wild-type and emp11-1.

Rapid Amplification of cDNA Ends (RACE)

RACE was performed with the SMART RACE cDNA amplification kit (BD Biosciences Clontech, Franklin Lakes, NJ, USA) according to the recommended protocol. Reactions for 3ʹ- RACE were conducted using primer F69 and nested primer F206. For the 5ʹ- RACE reactions, the universal primer provided with the kit was used in combination with primer R468 and nested primer R565. Primer sequences are listed in Supplementary Table S1.

RNA in situ hybridization

mRNA in situ hybridization was performed on developing kernels at 10 DAP; kernels were trimmed along the medial-lateral axis and immediately fixed in 4% paraformaldehyde, following the procedure for cytological observation. Slides were deparaffinized and treated with 10 mg m-1 proteinase K. In vitro transcription of the digoxigenin-UTP-labeled (Roche, Basel, Basel-Stadt, Switzerland) probe was completed. RNA sense and antisense probes were obtained using T7 and SP6 polymerases. The probe used to detect the Emp11 transcript corresponds to a 467 bp fragment from +656 to +1123bp of Emp11 cDNA and was constructed using the RT-Emp11-3F/3R primers. The primer sequences are listed in Supplementary Table S1. The hybridization was performed at 42°C overnight in 50% formamide buffer containing 0.5 ng ml-1 kb-1 probe, 0.5 mg ml-1 yeast tRNA, 10% dextran sulfate, 20 mM NaCl, 10 mM NaH2PO4, and 1X Denhardt’s solution. The slide was washed with 50% formamide in 2X saline sodium citrate (SSC) buffer solution. After RNase A treatment, a high stringency wash was performed twice using 2X SSC and twice using 0.2X SSC, as described in Shen . Digoxigenin detection and signal visualization were completed using nitroblue tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate (Roche, Basel, Basel-Stadt, Switzerland) according to the manufacturer’s instructions. Images were captured using a Nikon Eclipse 80i differential interference contrast microscope (Nikon, Chiyoda-Ku, Tokyo, Japan).

Subcellular localization of EMP11

Subcellular localization was performed as described by Liu . The full-length coding sequence of Emp11without the termination codon was amplified using the Emp11-Kpn1-F and Emp11-Xba1-R primers, which are listed in Supplementary Table S1. The PCR product was purified and sequenced. The purified product with the correct sequence was then inserted into the pM999-GFP vector to generate a GFP fusion construct. The fusion construct was introduced into maize protoplasts from seedling leaves with polyethylene glycol (PEG)/calcium-mediated transformation (Yoo ). MitoTracker Red (Thermo Fisher Scientific, Waltham, MA, USA) was used to label the mitochondria.

Mitochondrial complex activity assay

Crude mitochondria were isolated from the emp11 mutants and the wild-type kernels at 12 DAP. The isolation of crude and intact mitochondria, blue native PAGE, and the in-gel complex I activity assay were performed as described (Meyer ; Keren ). Kernels with the pericarp removed at 12 DAP were ground at 4°C in extraction buffer comprising 0.3 M sucrose, 5 mM tetrasodium pyrophosphate (Sigma, Santa Clara, CA, USA), 10 mM KH2PO4 at pH 7.5, 2 mM EDTA, 1% [w/v] polyvinylpyrrolidone 40, 1% [w/v] BSA, 5 mM cysteine, and 20 mM ascorbic acid, using a Polytron or porcelain mortar and pestle. The extract was filtered through two layers of cheesecloth and two layers of Miracloth (Calbiochem Co., La Jolla, CA, USA). The homogenate was centrifuged for 5 min at 3 000 g. Mitochondria were obtained by centrifugation of the clear supernatant at 20 000 g for 10 min. The pellet was resuspended in wash buffer comprising 0.3 M sucrose, 1 mM EGTA, and 10 mM MOPS/KOH at pH 7.2, and subjected to the same low speed (3 000 g) and high speed (20 000 g) centrifugations. Mitochondrial proteins weighing 200–500 µg, were solubilized with dodecyl maltoside (Sigma, Santa Clara, CA, USA), 1–2% [w/v] final or 4% digitonin (Sigma, Santa Clara, CA, USA) in amino caproic acid (ACA) buffer comprising 750 mM amino caproic acid (Sigma, Santa Clara, CA, USA), 0.5 mM EDTA, and 50 mM Tris-HCl at pH 7.0, and incubated for 30 min at 4°C. The samples were centrifuged for 10 min at 20 000 g, and Serva Blue G (0.2% [v/v] final) was added to the supernatant as described in the manual. The samples were loaded onto a 3% to 12% gradient gel (Invitrogen, Carlsbad, CA, USA) with cathode buffer containing 0.02% dodecyl maltoside. For the in-gel activity assay, complex I was optionally analyzed on separate gel strips. Gel strips loaded with extracts from 200 µg of maize mitochondria were assayed for complex I activity in an assay buffer comprising 25 mg of nitrotetrazolium blue (Sigma, Santa Clara, CA, USA) and 100 µl of 10 mg/ml NADH (Sigma, Santa Clara, CA, USA) added to 10 ml of 5 mM Tris/HCl at pH 7.4.

Maize transformation and progeny analysis

Antisense Emp11 (GRMZM2G353301) was amplified with primers GM301-1F and GM301-1R (see Supplementary Table S1) and ligated into the binary vector pCAMIBA3300. The resulting RNAi vector, pZZ-GM301, was transformed into the maize ZZC01 line using Agrobacterium-mediated transformation. Genotypes were tested with the TransDNA-F/R primers, with the TransDNA-F primer designed for the Emp11 sequence and the TransDNA-R primer designed for the vector sequence. Expression levels were tested with TransRNA-F/R primers designed for the 3ʹ-UTR of Emp11. The primers are listed in Supplementary Table S1.

Results

Phenotypic characterization of the emp11-1 mutant

The emp11-1 mutant was isolated from a UniformMu stock (No. UFMu-08197) with active Mutator (Mu) in the W22 genetic background (McCarty ). The emp11-1 mutant exhibited empty pericarp seeds in mature maize ears (Fig. 1A). The mutant kernels could be clearly distinguished from their wild-type siblings at 6 DAP on the basis of their reduced size. At 12 DAP, the mutant kernels were dramatically different from wild-type kernels in a segregating ear, namely they were small, white, and shrunken. At maturity, the mutant kernels had almost no starch in the pericarp. During sectioning of the mature kernels, normal embryo or endosperm structures could not be found in the mutant kernels and they were non-viable (Fig. 1B). Homozygous mutant seedlings could therefore not be obtained.

Fig. 1.

Embryo and endosperm development is arrested in emp11-1 mutants. (A) The ears segregated 3:1 for wild-type and emp11-1 mutant kernels (arrows). The emp11-1 mutants were defective with an empty, collapsed, and wrinkled pericarp. (B) The embryo or endosperm structures of emp11-1 kernels were abnormal at maturity. (C–J) Sections of developmental kernels at 12 DAP (C,D, G, H) and 15 DAP (E, F, I, J). Wild-type (C, E) and emp11-1 kernels (D, F). (G–J) are magnified images of the micrographs above. Arrows indicate the BETL. em, embryo; en, endosperm. Scale bar, 1 cm in (A), 2 mm in (B), 1 mm in (C–F) and 500 µm in (G–J).

To observe the development of the embryo and endosperm, emp11-1 and wild-type kernels from a segregating ear were sectioned and examined (Fig. 1 and Supplementary Fig. S1). At 6 DAP, the embryos of emp11-1 kernels were at the proembryo stage, similar to the wildtype, but the endosperm volume was dramatically smaller than wild-type kernels (Supplementary Fig. S1A, B). At 8 DAP, wild-type embryos had progressed from the proembyro to the coleoptile stage and endosperm cells filled the space inside the pericarp (Supplementary Fig. S1C), while emp11-1 embryos remained at the proembyro stage, resembling 6 DAP embryos (Supplementary Fig. S1D), and emp11-1 kernels were not filled by endosperm cells, with an obvious space between the pericarp and endosperm that persisted in later stages. At 12 DAP, wild-type embryos had a visible shoot apex (Fig. 1C) but emp11-1 embryos were arrested at the transition stage (Fig. 1D). At 15 DAP, wild-type kernels had developed a nearly mature embryo and a starch-filled endosperm (Fig. 1E and Supplementary Fig. S1I) but emp11-1 embryos remained arrested and the endosperm showed limited starch accumulation (Fig. 1F and Supplementary Fig. S1J). In summary, embryo and endosperm development were arrested in emp11-1 mutants.

Moreover, an observation of the basal endosperm transfer layer (BETL) found that emp11-1 kernels had a similar cellular morphology to wild-type kernels at early stages (Fig. 1G, H and Supplementary Fig. S1E–H). However, at 15 DAP, the BETL in emp11-1 kernels was invisible (Fig. 1J). These results suggest that the emp11-1 kernel defects may result from severe defects in development of the basal endosperm transfer cells.

Characterization of the Emp11 gene

The emp11-1 mutation behaved as a monogenic recessive trait since the mutant kernels on ears of self-pollinated heterozygous plants segregated in a ratio of 2235:800, wild-type: emp11, corresponding to a 3:1 segregation ratio (χ2-test, P>0.05; Fig. 1A). To identify the causal gene, the Mu-flanking sequences were identified using a method described by Settles . We found that a Mu7 element was inserted into GRMZM2G353301 at +531 bp downstream of the predicted translation start site (ATG) in UFMu-08197 (Fig. 2A). A segregating F2 population was developed by self-crossing a heterozygous +/emp11-1 plant. Genomic DNA was isolated from the individual F2 plants, and the phenotype was identified by selfed plants and checked for emp11-1 segregation. As emp11-1 homozygotes were not viable, only heterozygotes (segregating) or the wild-type (non-segregating) were available for this analysis. Importantly, the insertion site co-segregated with the phenotype in all the 901 individuals tested (+/+: +/emp11-1, 308:593, corresponding to 1:2 ratio, χ2-test, P>0.05) (Supplementary Fig. S2A, B).

Fig. 2.

Characterization of the Emp11 gene. (A) Emp11 encodes a P-type PPR protein that contains 16 PPR repeats. Gene structure of Emp11 and locations of the Mu insertions in two independent alleles. The Mu insertion sites in emp11 alleles are indicated with triangles. (B) The 16 PPR motifs of the EMP11 protein are loosely conserved in alignment.

To determine whether the mutation in GRMZM2G353301 was the cause for the emp11 phenotype, we isolated an independent mutant, UFMu-04323, with a Mu element inserted at +1741 bp in the coding sequence (Fig. 2A) and we named this allele emp11-2. Importantly, the selfed progeny of +/emp11-2 heterozygotes also produced empty pericarp kernels in a recessive pattern in a 3:1 ratio, with a wild-type: emp11 ratio of 1516:497 (χ2-test, P>0.05) (Supplementary Fig. S2C, D). Complementation crosses between +/emp11-1 and +/emp11-2 produced mutant kernels as well with a wild-type: emp11 ratio of 3896:1379 (χ2-test, P>0.05) (Supplementary Fig. S2E), confirming that GRMZM2G353301 is the causative gene for the empty pericarp phenotype.

To further explore the function of Emp11, RNAi transgenic plants were created in the maize inbred line ZZC01. It is likely that severe knockdown lines would be lethal considering the essential role of Emp11 in W22, therefore we assume that only weak knockdown lines were recovered and two representative lines were used for further molecular analysis. qRT-PCR analysis revealed the Emp11 expression level was significantly decreased in transgenic lines TG61 and TG68 compared with a non-transgenic control (NT) (Supplementary Fig. S3A). The T1 progeny of the two transgenic plants segregated defective kernels at different ratios, ~15% in line TG61 and ~22% in line TG68 (Supplementary Fig. S3B). The defective kernels were small, having 61% of the weight of the well-developed kernels in the TG61 line and 30% of the weight in the TG68 line (Supplementary Fig. S3D). These results showed that knockdown of Emp11 expression affected kernel development and led to defective kernels, consistent with the UniformMualleles. Taken together, the results confirm that loss of GRMZM2G353301 is responsible for the emp11 phenotype.

The genomic sequence and cDNA of Emp11 were isolated from the W22 inbred line. Gene-specific primers were designed and used in rapid amplification of cDNA ends (RACE) to generate full-length wild-type Emp11 cDNA (see Materials and Methods and Supplementary Table S1). Emp11 was an intronless gene encoding a PPR protein with 698 amino acids, which were predicted to form 16 PPR motifs (Fig. 2A, B). The emp11-1 allele contained a Mu7 insertion in the fourth PPR motif-encoding region and emp11-2 had a Mu insertion in the coding region between the fourteenth and fifteenth PPR motifs (Fig. 2A). EMP11 was most closely related to Sb05g002210 from Sorghum bicolor and LOC_Os11g04295 from rice (Supplementary Fig. S4).

Expression and localization of Emp11

To examine the expression pattern of Emp11, we performed RT-PCR and found that Emp11 was expressed in all vegetative and reproductive tissues tested, with higher levels in the ear, ovary, and embryo and lower levels in the root, stem, leaf, tassel, silk, and endosperm (Fig. 3A and Supplementary Fig. S5). We next genotyped developing seeds with the pericarp removed. Homozygous emp11 seeds with the pericarp removed were used to analyze expression of Emp11 and mitochondrial genes. Emp11 transcripts could not be detected in the emp11-1 or emp11-2 mutants (Fig. 3B), suggesting that both mutations were likely to be null. To understand the spatial localization of Emp11, mRNA in situ hybridization was performed in the wild-type kernels at 10 DAP. A strong hybridization signal was detected in the aleurone layer (Fig. 3C), the BETL (Fig. 3D) and in the embryo shoot apical meristem (Fig. 3E), with lower levels in other tissues, suggesting that Emp11 was ubiquitously expressed but was higher in more metabolically active tissues. Furthermore, we characterized the subcellular localization of EMP11 using GFP fusion proteins. Transient expression of the EMP11-GFP fusion in maize protoplasts showed that GFP signal overlapped well with MitoTracker Red (Fig. 3I), demonstrating that EMP11 is targeted to mitochondria.

Fig. 3.

Expression and localization of Emp11. (A) RT-PCR analysis of Emp11 expression in the root, stem, leaf, tassel, ear, silk, ovary, endosperm, and embryo 15 DAP. (B) RT-PCR analysis of Emp11 expression in wild-type, emp11-1, and emp11-2 seeds. RNA was isolated from kernels 15 DAP after pericarp removal. (C–F) RNA in situ hybridization was performed on developing kernels at 10 DAP using an Emp11 antisense probe (C, D, E) and sense probe (F, G, H). Scale bar, 50 µm. (G) Subcellular localization of the EMP11 protein in mitochondria. The EMP11-GFP fusion protein was transiently expressed in maize leaf protoplasts and visualized with confocal microscopy. Scale bar, 10 µm.

Function of Emp11 in mitochondrial nad1 splicing

To investigate the function of Emp11, we analyzed the transcript levels of the maize NB mitochondrial genes in wild-type and emp11-1 mutants by RT-PCR. Total RNAs were extracted from 15 DAP kernels of wild-type and emp11-1 from the same segregating ear with the pericarp removed. The results showed that the expression level of most mitochondrial genes was not different between wild-type and the emp11-1 mutant, but some mitochondrial tRNAs were more highly expressed in the emp11-1 mutant (Fig. 4A and Supplementary Fig. S6A). Intriguingly, the fully spliced nad1 transcript could not be detected in emp11-1 kernels using primers targeted to the 5ʹ- and 3ʹ-UTRs, near the ATG and stop codon, but could be detected in wild-type kernels (Fig. 4A), suggesting that Emp11 is potentially required for the regulation of nad1 expression. The maize mitochondrial nad1 gene contains three very long introns, introns 1, 3, and 4, which can’t be amplified by PCR and one 1393 bp intron, intron 2, which can be amplified using the F2+R2 primer pair shown in Fig. 4B (Brown ). To examine whether an intron-splicing deficiency leads to the absence of mature nad1 transcript, we compared the transcripts for the presence of unspliced introns by RT-PCR. The results showed that intron1-unspliced transcripts were not detected in both wild-type and mutants, due to a failure in PCR amplification of the long intron1 using the F1+R1 primer pair, but levels of intron1-spliced PCR products, 327 bp in size, were lower in emp11 mutants than in wild-type (Fig. 4B). Levels of intron3-spliced PCR products, 170 bp in size, and intron4-spliced PCR products, 315 bp in size, which were amplified using F3+R3 and F4+R4 primer pairs, respectively, were also lower in emp11 mutants than in wild-type (Fig. 4B). Both intron 2-unspliced, 1640 bp in size, and intron 2-spliced, 247 bp in size, transcripts could be amplified using F2+R2 primer pairs, and levels of the intron 2-spliced transcripts were lower in emp11 mutants than in wild-type, whereas levels of unspliced transcripts were higher in emp11 (Fig. 4B). The splicing defects of nad1 introns in emp11 mutants were also revealed by qRT-PCR (Fig. 4C). These results demonstrated that Emp11 was required for the splicing of all nad1 introns. We also examined the splicing efficiency of other mitochondrial group II introns by RT-PCR (Supplementary Fig. S6B). Both the RT-PCR and qRT-PCR results showed that all of the other mitochondrial group II introns were spliced normally in emp11. These results confirm that nad1 intron splicing is specifically defective in the emp11 mutants.

In Arabidopsis, six nuclear-encoded factors are required for the excision of nad1 introns, including OTP43 (Falcon de Longevialle ), nMAT1 (Keren ), and nMAT4 (Cohen ) for intron 1; nMAT2 (Keren ), mCSF1 (Zmudjak ), and PMH2 (Köhler ) for intron 2; and mCSF1, nMAT4, and PMH2 for intron 3. Only nMAT4 is involved in the splicing of nad1 intron 4. The maize orthologs of nMAT1, nMAT2, nMAT4, OTP43, and mCSF1 are GRMZM2G023983, GRMZM2G154119, GRMZM2G375999, GRMZM2G 046058, and GRMZM2G087395, respectively. PMH2 orthologs are GRMZM2G080512, GRMZM2G107984, and GRMZM2G565140. In maize, Dek2 is involved in the splicing of mitochondrial nad1, since the dek2 mutation reduces the splicing efficiency of nad1 intron 1. To investigate the nad1 splicing mechanism in maize, we assayed the expression of dek2 and the eight maize orthologs genes in emp11-1 and wild-type plants by RT-PCR. The results showed that the expression of some genes was significantly higher in emp11-1 plants than in wild-type, for example, dek2, GRMZM2G375999, GRMZM2G107984, and GRMZM2G565140 (Fig. 4D). In Arabidopsis, nad1 has two trans-splicing introns, introns 1 and 3, while in maize, nad1 introns 1, 3, and 4 are trans-spliced. These results show that nad1 splicing in maize likely shares a complex mechanism with Arabidopsis and the absence of Emp11 induces a higher expression of nad1 splicing-related genes.

Mitochondrial structure and respiratory function in emp11 mutants

Nad1 is a central component of complex I and is essential for assembly of the complex in the membrane (Klodmann ). Substantial reduction in nad1 expression is likely to impact complex assembly, stability, and/or activity. To uncover whether the defect in nad1 processing in emp11 mutants affected complex I formation and/or function, we examined complex I assembly with Blue Native polyacrylamide gel electrophoresis (BN-PAGE) and complex I activity with an in-gel NADH dehydrogenase activity assay. The results showed strongly reduced assembly and activity of complex I in emp11 mutants (Fig. 5A, B), indicating that the loss of Emp11 expression led to a reduction in the assembly and activity of complex I.

Fig. 5.

Mitochondrial structure and respiratory function in emp11 mutants. (A) BN gels were stained with Coomassie Brilliant blue. Mitochondrial complexes of the embryo and endosperm of maize kernels at 12 DAP were subjected to a 3% to 12% BN-PAGE. (B) In-gel NADH dehydrogenase activity assay of the mitochondrial protein complexes was assayed using dihydrolipoamide dehydrogenase (DLDH) as a loading control. (C) The transcript levels of AOX1, AOX2, and AOX3 in emp11-1 and emp11-2 kernels. Total RNA was extracted from 15 DAP kernels with the pericarp removed. The expression levels were normalized to ZmActin (GRMZM2G126010). (D) qRT-PCR analysis of AOX1, AOX2, and AOX3 in wild-type, emp11-1, and emp11-2 plants. (E, F) Transmission electron micrographs of ultra-thin sections from the endosperm of wild-type (E) and emp11 mutants (F). Mitochondria are indicated with arrows. Scale bars, 0.5 µm.

To further investigate the consequences of the emp11 mutation, we analyzed the alternative pathway by examining the expression of three alternative oxidases, AOX1 (AY059646.1), AOX2 (AY059647.1), and AOX3 (AY059648.1). We found that AOX2 expression was dramatically increased in emp11 mutants and AOX3 expression was also higher (Fig. 5C, D). These results indicate that the loss of Emp11 function causes a dramatic increase in the expression of AOX genes.

Reduced complex I levels in Arabidopsis nmat1 and nmat4 mutants are considered to contribute to the compromised mitochondrial ultrastructure (Keren ; Sosso ; Cohen ). To observe the ultrastructure of emp11 mitochondria, we used ultra-thin sections of 10 DAP endosperm for TEM and found that wild-type mitochondria showed normal cristae, visible as densely folded inner membranes (Fig. 5E), while approximately 40% of emp11-1 mitochondria had a poorly developed membrane system with a large internal space and abnormal and obscure cristae (Fig. 5F). Based on the mitochondrial ultrastructure, we assume that the respiratory function of the mutant mitochondria may be impaired.

The phenotype of emp11 is modified in B73 BC2F2 segregating ears

In BC2F2 segregating ears in B73, we observed partially filled emp11-1/emp11-1 kernels with weights approximately one fifth of that of wild-type, larger than in W22 (Supplementary Fig. S7A, B). Furthermore, we found that emp11-1 (B73), emp11-1 homozygous kernels in BC2F2 segregating ears in B73, embryos were larger than in W22 with thick scutellum, small coleorhiza, short roots, stunted coleoptile, and abnormal shoot apex lacking clear differentiation (Supplementary Fig. S7C). Moreover, homozygous emp11-1 (B73) seeds had a 55% germination ratio, but these seedlings exhibited low a survival rate of ~23%) and the survived seedlings had much smaller roots and shoots relative to wild-type(B73) (Supplementary Fig. S7D, E). We measured agronomically important traits, including plant height, ear height, kernel row number, leaf width, spikelet density, tassel branches, and tassel length of those surviving seedlings, and found that these traits were severely arrested in emp11-1 (B73) plants (Supplementary Fig. S7F–L).

Discussion

Emp11 affects complex I by modulating the intron splicing of nad1 in maize mitochondria

NAD1 is a central component of complex I, which is located on the mitochondrial membrane (Klodmann ; Meyer ). NAD1 deficiency leads to disturbances in the assembly or stability of complex I (Keren ; Keren ; Cohen ). In maize, Dek2, a mitochondrial P-type PPR protein, is involved in the splicing of nad1 intron 1 (Qi ). In Arabidopsis, there are at least three maturases and various other protein cofactors involved in the maturation of nad1 (Cohen ), including OTP43, a P-type PPR protein, which is specifically required for trans-splicing of mitochondrial nad1 intron 1 and cis-splicing of nad2 intron 1 (Falcon de Longevialle ). In our study, the four introns within nad1 pre-mRNA could not be removed normally because of the loss of Emp11 function, while other mitochondrial introns in emp11 were correctly spliced, suggesting that a single emp11 mutation affects the splicing of all the introns of nad1. It is well known that the mechanism of mitochondrial intron splicing is a very complex procedure involving many splicing factors, which facilitate the splicing of group II introns, including PPR proteins, CRM (chloroplast RNA splicing and ribosome maturation) proteins, RNA DEAD-box helicases, PORR (plant orangellar RNA recognition) domain proteins, RCC (regulator of chromosome condensation) proteins, and others. Among them, PPR, PORR, and RCC may act as the factors that recognize the specific RNA-binding sites (Brown ). To interpret the non-specificity of EMP11 to nad1 introns, we assume that EMP11 may play a role in the recognition of precursor nad1 mRNA and in the maintenance of the nad1 conformation for intron splicing by cooperating with other factors. The absence of mature nad1 transcripts led to the failure of complex I assembly (Fig. 5). This is consistent with early reports in complex I-deficient nmat1 (Keren ), nmat2 (Keren ), and nmat4 (Cohen ) mutants. Nucleus-encoded factors play essential roles in the regulation of mitochondrion development, which requires the coordinated expression of both nucleus-encoded and mitochondrion-encoded genes, such as ppr8522, a plastid mutant that triggers a retrograde signal to the nucleus (Sosso ). These results support the existence of a retrograde signaling pathway coordinating mitochondrial and nuclear gene expression. This retrograde signaling may be associated with the mitochondrial intron splicing machinery to monitor its integrity. In our study, Dek2 and some orthologs of the Arabidopsis nad1 splicing genes, OTP43, nMAT1, nMAT2, nMAT4, mCSF1, and PMH2, had higher transcript levels in emp11 mutants. We propose that nad1 splicing in maize likely shares a similar mechanism as that in Arabidopsis and the absence of Emp11 induces a higher expression of nad1 splicing-related genes in the emp11-1 mutant. The discrepancies in the splicing efficiencies of nad1 introns and up-regulated expression of other nad1 splicing genes in emp11 mutants suggest that Emp11 possibly influences a plethora of metabolic and developmental changes.

ATP production within mitochondria is linked to electron transfer through complex I (Millar ). Defects in complex I likely lead to loss of ATP synthesis, alter signal transduction pathways, and disturb normal metabolism (Schmitz-Linneweber and Small, 2008). In our study, the mitochondrial structure of emp11 mutants was altered (Fig. 5F). This altered mitochondrial morphology is likely unable to provide enough ATP to maintain normal life activities. Under this low energy condition, the alternative respiratory pathway is activated (Gutierres ; Sabar ; Karpova ; Dutilleul ). Our data showed that AOX2 transcription levels were dramatically increased in emp11 mutants (Fig. 5C, D), suggesting that the AOX pathway was activated by the loss of Emp11 expression. Complex I mutants are anticipated to display retarded phenotypes with empty pericarp seeds, which may result from deficits in energy supply. For example, in Arabidopsis, otp43, nmat1, and nmat4 mutants display similar growth defects as they all fail to process nad1 transcripts correctly and show altered respiratory function as a consequence of complex I defects (Falcon de Longevialle ; Keren ; Cohen ). Taken together, these data suggest that EMP11 serves as a factor involved in the splicing of mitochondrial nad1 introns and affects the respiratory function of complex I, which is the physiological basis for ensuring normal embryo and endosperm development in maize kernels.

Emp11 is required for seed and plant development and the emp phenotype could be partially recovered in B73 BC2F2 segregating ears

The maize embryo and endosperm are two products of double fertilization. Endosperm development includes coenocytic, cellularization, differentiation, and maturation stages. The differentiated endosperm contains four major cell types: the cells of the embryo surrounding region, the BETL, the aleurone layer, and the starchy endosperm (Olsen, 2001). The BETL is responsible for nutrient transfer from maternal tissue to the developing endosperm (Olsen, 2004). mRNA in situ hybridization revealed that Emp11 was highly expressed in BETL cells (Fig. 3D), suggesting that Emp11 is important for BETL development. This inference was then supported by the fact that BETL was absent from emp11 kernels at 15 DAP (Fig. 1J) and was also further supported by the observation that the starchy endosperm was poorly filled in emp11 mutants (Fig. 1F and Supplementary Fig. S1J). Embryo development in maize proceeds in three stages, namely the transition, coleoptilar, and late embryogenesis stages (Olsen, 2001). In this study, we found that Emp11 was highly expressed in embryos (Fig. 3A). In the W22 background, the loss of Emp11 function caused the early arrest of embryo development at the transition stage (Fig. 1 and Supplementary Fig. S1). Homozygous emp11-1 (B73) embryos could sometimes develop and germinate into plants, but they had delayed development relative to wild-type (B73) plants (Fig. 6E), suggesting that EMP11 also acts during post-embryonic development. Background-dependence has also been described for ppr8522 and whirly1, two plastid-targeted PPR genes that confer an embryo-specific (emb) phenotype (Prikrylv ; Sosso ; Zhang ). In the R-scm2 background, the emb8522 mutant causes embryo lethality but causes albino seedlings in the A188 and B73 backgrounds (Sosso ). In addition, the why1 allele in the W22 genetic background causes specific arrest in embryogenesis without a major impact on endosperm development. However, the severe phenotype can be suppressed in maize A188, B73, Mo17 and Oh51a genetic backgrounds, giving rise to seedlings, which show an albino seedling phenotype (Prikrylv ; Zhang ). The retrograde signaling hypothesis is used to explain the background-dependent suppression of why1 in embryogenesis (Zhang ). The emp11 phenotype maybe background-dependent and there appears to be suppressor(s) in B73 that can suppress the emp phenotype of emp11, as we observed that homozygous emp11 seeds in B73 BC2F2 segregating ears have the totipotency to develop a plant. A similar finding was reported by introgresing single emp mutants in different genetic backgrounds and revealed the existence of a cryptic genetic variation (CGV) recognizable as a variable increase in endosperm tissue (Sangiorgio ). The genetic difference between B73 and W22 may provide clues to understanding the background dependence of the emp11 phenotype, which showed embyro lethality under the W22 background but embryo totipotency in B73 BC2F2 segregating ears.

Supplementary data

Supplementary data can be found at JXB online.

Table S1. Primers used in this study.

Fig. S1. Embryo and endosperm development in early stage.

Fig. S2. Genotype and phenotype analysis of emp11 mutants alleles.

Fig. S3. Defective kernels were produced in Emp11 RNAi transgenic lines.

Fig. S4. Alignment of the EMP11 protein with rice ortholog LOC_Os11g04295, sorghum ortholog Sb05g002210 and Arabidopsis ortholog AT1G52620.

Fig. S5. Expression of Emp11 in Multiple Organs.

Fig. S6. Emp11 is specifically involved in nad1 introns splicing.

Fig. S7. The emp phenotype of the emp11-1 mutants can be suppressed in BC2F2 of B73.

Author contributions

Conceived and designed the experiments: ZZ FQ XR ZP. Performed the experiments: XR ZP HZ. Analyzed the data: XR ZP HZ. Contributed reagents/materials/analysis tools: JZ MC. Wrote the manuscript: XR ZP ZZ FQ.

Click here for additional data file.