The restoration fertility complex (RFC) was previously identified in Honglian (HL)-cytoplasmic male sterility (CMS) rice (Oryza sativa), and glycine-rich protein 162 (GRP162) is responsible for binding to the CMS-associated transcript atp6-orfH79.
Here, we engineered a recombinant GRP162 containing the mitochondrial transit peptide, termed Mt-GRP162, as an artificial restorer of fertility (Rf) gene. Mt-GRP162 was confirmed to bind to CMS-associated RNA and to localize to the mitochondria. The transgenic plants showed restored fertility with partially functional pollen.
We found that the expression of ORFH79 decreased in transgenic plants, while the expression of atp6-orfH79 was not changed. These findings indicate that Mt-GRP162 restores fertility by suppressing the expression of the cytotoxic protein ORFH79 at the post-transcriptional level rather than via the cleavage of atp6-orfH79 in the presence of RFC.
These findings contribute to our understanding of the mechanisms of restoration through diverse pathways.
Cytoplasmic male sterility (CMS), a maternally inherited inability to produce functional pollen, has been observed in c. 200 species of higher plants (Hanson, 2004; Chase, 2007). Several CMS-associated genes have been found to be encoded by the mitochondrial genome. The CMS phenotype can be rescued by a class of nuclear-encoded genes termed ‘restorer of fertility’ (Rf) genes, highlighting the significance of the interactions of mitochondrial and nuclear gene products. Rf2 in T-CMS maize encodes an aldehyde dehydrogenase that accumulates in the mitochondria (Cui et al., 1996). Rf17, which has a partial acyl-carrier protein synthase (ACPS) domain, restores the fertility of CW-CMS rice via a retrograde regulation signal pathway (Fujii & Toriyama, 2009). The Rf2 gene in lead type-CMS (LD-CMS) in rice is a glycine-rich protein (GRP) (Itabashi et al., 2011). The remaining known Rf genes belong to the pentatricopeptide (PPR) family, a large family of genes that encode organelle-targeted proteins. These include PPR592 in petunia, Rfk1 in Kosena rapeseed, Rfo in radish, PPR13 in sorghum, Rf1a/Rf1b in Boro II type CMS (BT-CMS) rice and Rf5 in Honglian type CMS (HL-CMS) rice (Bentolila et al., 2002; Brown et al., 2003; Desloire et al., 2003; Kazama & Toriyama, 2003; Koizuka et al., 2003; Akagi et al., 2004; Komori et al., 2004; Klein et al., 2005; Wang et al., 2006; Hu et al., 2012).
The Rf genes act to reduce the accumulation of CMS-associated RNAs and/or proteins through different mechanisms, acting at the DNA, RNA, or protein level, or even by metabolic complementation. We consider three levels of regulation based on the changes in CMS genes in F1 hybrids, even though some Rf genes remain to be characterized. The first level is the DNA level. For example, Fr can restore the pollen fertility of a CMS common bean by decreasing the copy number of orf239 (Mackenzie & Chase, 1990; Arrieta-Montiel et al., 2001). The second level is the post-transcriptional level; most restoration processes can be categorized as this type. In T-CMS maize, for example, Rf1, Rf 8, and Rf * process the CMS RNA, T-urf13, to achieve restoration, while in S-CMS maize, Rf3 can alter the transcription pattern of orf355 and orf 77 (Wise et al., 1999). In rapeseed, the CMS RNAs orf224/atp6 in Polima CMS are processed in the presence of Rfp, and orf222 in napus CMS is processed in the presence of Rfn (Yuan et al., 2003; Geddy et al., 2005). In sunflower, PET1 might destabilize the CMS-associated transcript orf522 in a tissue-specific manner to restore male fertility (Moneger et al., 1994). The Petunia PPR592 not only reduces the transcript level of pcf but also affects its protein level (Hanson et al., 1999). The CMS RNA atp6-orf79 not only can be cleaved by PPR791 (RF1A) but also can be degraded by RF1B in Boro II CMS rice, while atp6-orfH79 in HL-CMS is cleaved by an RFC (Wang et al., 2006; Kazama et al., 2008; Hu et al., 2012). The destabilization and/or translational control of CMS proteins are additional potential mechanisms for restoration. The levels of the CMS proteins ORF125 and ORF138 are dramatically reduced in the presence of PPRB/ORF687 (Koizuka et al., 2003; Uyttewaal et al., 2008). In LD-CMS rice, RF2 might interact directly with the CMS-causing protein or form a multimolecular complex to achieve fertility restoration (Itabashi et al., 2011). The third level of regulation is metabolic complementation of the damage caused by CMS-associated proteins. For instance, URF13 expression in T-CMS maize is not reduced in the presence of Rf2, which encodes an ALDH protein that accumulates in the mitochondria (Chase, 2007). The pollen must eliminate the impairments caused by CMS genes by any means necessary in order to survive. Therefore, many types of Rfs have developed throughout evolution.
In a previous study, we showed that GRP162 is involved in the process of fertility restoration as a key component of the restoration fertility complex (RFC) in HL-CMS rice. Results established that GRP162 specifically bound to CMS RNAs directly. The expression of GRP162 was increased, and the localization of GRP162 was also markedly altered, in the presence of RF5 (Hu et al., 2012). Hence, we proposed that GRP162 could be rebuilt as a restorer of fertility gene with the addition of a mitochondrial transit peptide.
In this study, an artificial protein with CMS-associated transcript-binding activity was introduced into the mitochondria. The pollen was rescued in the presence of Mt-GRP162, as shown in a transgenic complementation assay. The results confirmed the mechanism of ORFH79 suppression by the artificial restorer fertility gene Mt-GRP162 and suggested a new pathway for fertility restoration. These findings also imply that GRP162 may not be the only essential factor for RNA cleavage in RFC.
Materials and Methods
Construction of Mt-GRP162
The mitochondrial transit peptide of RF5 was amplified with specific primers (Rf5-atg-BglII-F: CagatctATGGCGCGCCGCGCCGCTTCC, Rf5-MT-BamHI-R: CggatccGTAGGTGCACAAGTCGGGAG) (restriction enzyme sites indicated by lowercase letters) and inserted into the HBT-sGFP vector to produce a HBT-Mt-sGFP construct for the validation of the subcellular localization of the mitochondrial transit peptide. The mitochondrial transit peptide was first inserted into pCAMBIA1301-UBI to produce the donor vector pCAMBIA1301-UBI:Mt. The GRP162 whole cDNA was amplified with specific primers (GRP162-BamHI-F: CCggatccATGGCGGCGCCGGATGT, GRP162-BamHI-R: CggatccGTTCCTCCAGTTCCCGTC) (restriction enzyme sites indicated by lowercase letters) and inserted into the pCAMBIA1301-UBI:Mt vector to produce pCAMBIA1301-UBI:Mt-GRP162, in which the artificial Mt-GRP162 containing a mitochondrial transit peptide was driven by the ubiquitin promoter.
Subcellular localization of Mt-GFP
The constructs were transformed into rice protoplasts for transient expression to validate their subcellular localization. The rice protoplasts were isolated from four-wk-old etiolated seedling leaves incubated in a dark room and prepared according to the standard protocol (Faraco et al., 2011). Approximately 20 μg of HBT-sGFP and HBT-Mt-sGFP plasmid DNA were transformed into 5 × 105 protoplasts, followed by culture for 14–20 h in a dark room. Mitochondria were labeled with MitoTracker Red (Invitrogen) and observed by confocal microscopy (Olympus, Tokyo, Japan).
Expression of Mt-GRP162
To confirm the RNA recognition activity of Mt-GRP162, the Mt-GRP162 fragment was ligated into pET32a for expression in BL21. The recombinant proteins were purified by affinity chromatography on Ni-NTA agarose (GE Healthcare, NJ, USA), washed with buffer A (20 mM sodium phosphate, pH 7.4, 2 mM imidazole) and eluted with buffer B (20 mM sodium phosphate, pH 7.4, 180 mM imidazole). The His-Mt-GRP162 was further dialyzed against DEPC-treated buffer C (40 mM HEPES-KOH, pH 7.0, 1 M KCl) and quantified using a BCA protein assay (Pierce, Rockford, IL, USA).
Electrophoretic mobility shift assay (EMSA)
For the electrophoretic mobility shift assay (EMSA), serial dilutions of purified recombinant His-Mt-GRP162 were incubated with 5′-end labeled RNAs probe in a 50 μl reaction mixture including 25 μl of 2× binding buffer (100 mM sodium phosphate, pH 7.5, 10 units RNasin, 0.1 mg ml−1 BSA, 10 mM DTT). In each reaction we applied excess yeast tRNA as competitor RNAs. The mixture was incubated at 37°C for 30 min. Samples were separated by 5% native PAGE in 0.5× TBE buffer, exposed to a phosphorimager screen and analyzed using a Typhoon 9200 instrument (AB, Sunnyvale, CA, USA) or visualized by autoradiography.
Northern blot and western blot
Total RNA was isolated with the TRIzol reagent (Invitrogen). Thus orfH79 was amplified and recovered for use as a probe. Approximately 20 μg of total RNA was transferred to a nylon membrane for northern blotting. A random primer DNA labeling kit 2.0 (Takara, Dalian, Liaoning, China) was used for probe labeling, and the northern blot was performed as described by Hu et al. (2012).
Total protein was isolated from seedling leaves with extraction buffer (66 mM Tris pH 6.8, 2% SDS, 2% β-mercaptoethanol) and incubated on ice for 60 min. After precipitation with acetone, the protein was dissolved in buffer (PBS pH 7.4, 0.1% SDS) and quantified with a BCA kit (Pierce). Approximately 40 μg of total protein was separated by 10% SDS-PAGE for Coomassie Brilliant Blue staining or on a 10% Tricine gel and transferred onto a PVDF membrane (GE Healthcare, Mannheim, Germany) for western blotting. Anti-ORFH79 (1 : 2000) and anti-UGPase (1 : 2000) antibodies were employed as the probe and positive control, respectively. HRP-linked secondary antibodies (mouse anti-rabbit) and the ECL Plus reagent were used to visualize the bands (Thermo Scientific, Rockford, IL, USA).
The mitochondrial transit peptide localizes to the mitochondria
Previous studies indicated that GRP162 interacts with RF5 as a member of the RFC in the HL-CMS restoration pathway and binds to CMS-associated transcripts to restore the fertility in HL-type CMS (Hu et al., 2012). Transgenic lines in which GRP162 was silenced exhibited disturbed restoration of the fertility pathway. Considering the interaction between GRP and CMS-associated transcripts, we proposed that the Mt-GRP162 could capture the CMS RNAs to suppress the expression of the cytotoxic protein ORFH79, consequently alleviating the damage caused by ORFH79 and rescuing the viability of the pollen.
To perform as well as the restorer gene Rf5, the reconstructed protein must be introduced into mitochondria. Because GRP162 is diffusely distributed in the cytoplasm, we engineered the artificial Mt-GRP162 for transformation to ensure its localization to the mitochondria to better understand the function of GRP162 and the RFC. To alter the subcellular localization of Mt-GRP162, an expression vector including a 90 amino acid mitochondrial transit peptide derived from RF5 and driven by the ubiquitin promoter was constructed (Fig. 1a). We first used transient expression in rice protoplasm to validate the function of the mitochondrial transit peptide encoded by the HBT-sGFP vector. Figure 1(b) shows that the localization of Mt-sGFP perfectly overlaps with the mitochondrial signal, suggesting that the mitochondrial transit peptide derived from RF5 is functional and that Mt-GRP should be introduced into the mitochondria by this mitochondrial transit peptide.
Mt-GRP162 interacts with CMS-associated RNA
Another requirement for restoration is that the reconstructed protein should exhibit CMS RNA binding activity. To that end, the Mt-GRP162 was inserted in pET32a for expression (Fig. 2a). Then, the recombinant His-Mt-GRP162 was extensively purified and dialyzed against DEPC-treated buffer for EMSA (Fig. 2b). The recombinant protein was analyzed by western blot with anti-His antibodies, and the results confirmed the high purity of the protein (Fig. 2c). A previous study showed that GRP162 binds to atp6-orfH79 specifically (Hu et al., 2012). Next, EMSA was performed with CMS RNA to confirm the RNA binding capability of the recombinant protein in the presences of yeast tRNA as competitor. The EMSA data suggested that the recombinant Mt-GRP162 retained its RNA binding capabilities, even in the presence of excess yeast tRNA (Fig. 2d). We concluded that the presence of the mitochondrial transit peptide does not impair the RNA binding activity of GRP162, establishing that the artificial Mt-GRP162 is suitable for further transgenic analysis.
Pollen viability is rescued in the presence of Mt-GRP162
To validate our proposal that the artificial Mt-GRP162 can restore the fertility of HL-CMS, the construct pCAMBIA1301-UBI:Mt-GRP162 was introduced into the HL-CMS line YTA for genetic complementation tests. We obtained five positive transgenic plants, as confirmed by PCR assays. The growth of the transgenic reporter lines was indistinguishable from that of wild-type plants, indicating that the fusion protein did not obviously affect plant development. The fertility of independent transformants was assayed by staining pollen with 1% I2-KI (Fig. 3a, Supporting Information Table S1). Three of five T0 plants derived from the pCAMBIA1301-UBI:Mt-GRP162 transformation showed the expected fertile phenotype with average 48.5% fertile pollen, whereas the other two plants showed partially complemented fertility. By contrast, the 12 T0 plants derived from the empty vector were completely male sterile. These results indicated that Mt-GRP162 indeed rescued the viability of pollen in the HL-CMS sterile line.
To genetically confirm the fertility restoration phenotype of transformants, two independent T0 transgenic plants were self-bred and simultaneously test crossed to the HL-CMS line YTA. A 1 : 1 genetic segregation was observed in 34 T1 progeny from T0-2, with 13 plants producing an average of 84.0% fertile pollen and 21 producing an average of 47.3% fertile pollen, reflecting the expected gametophytic genetic segregation (χ2 = 1.44 for 1 : 1, P <0.05) (Fig. 3b, Supporting Information Table S2). The PCR results were consistent with the phenotypes of these progenies, confirming the functional complementation of Mt-GRP162.
Furthermore, 27 F1 progeny of the test crosses presented an average of 45.9% fertile pollen as detected by 1% I2-KI (Supporting Information Table S3). Therefore, we concluded that Mt-GRP162 rescued the pollen viability of the HL-CMS line.
Mt-GRP162 suppresses the expression of the CMS protein rather than the CMS RNA
To obtain insight into the mechanism by which Mt-GRP162 restored the fertility of the HL-CMS line, we next investigated the levels of the CMS-associated transcript and protein. Previous studies have shown that atp6-orfH79 and orfH79(s) encode a cytotoxic protein, ORFH79, that is responsible for the production of sterile pollen in HL-CMS (Li et al., 2008; Peng et al., 2009, 2010). We next examined the expression of the CMS RNAs in the three transgenic T1 plants from T0-2 (T1-3, T1-15 and T1-21), which produced c. 85% fertile pollen, by northern blot assays using orfH79 as probe. The CMS RNAs of atp6-orfH79 and orfH79(s) were still observed in the seedling leaves of the HL-CMS line and transgenic T1 plants, demonstrating that the CMS RNAs are not cleaved or degraded in the presence of Mt-GRP162, unlike in the near isogenic line (Rf5) (NIL(Rf5)) line, in which CMS RNAs are cleaved to suppress the expression of ORFH79 (Fig. 4).
Therefore, we speculated that the process of protein translation might be disrupted. A western blot assay was employed to analyze the expression of ORFH79 in the HL-CMS line, three T1 transgenic plants and the NIL(Rf5) line. ORFH79 protein expression was obviously decreased in the transgenic plants in comparison to the HL-CMS line (Fig. 5), suggesting that the expression of ORFH79 was suppressed in the presence of Mt-GRP162. No signal was observed in the NIL(Rf5) line because the CMS-associated RNAs were cleaved in the presence of RFC, as described in a previous study. These results indicate that the mechanism of fertility restoration via Mt-GRP162 relies on the suppression of the cytotoxic protein ORFH79 rather than the cleavage of atp6-orfH79 in the presence of RFC.
Mt-GRP can restore the fertility of HL-CMS via the suppression of ORFH79
Several reports have established that the CMS-associated proteins are cytotoxic and that the expression of these aberrant mitochondrial proteins leads directly to cytoplasmic male sterility (Duroc et al., 2005; Peng et al., 2010). A decrease in the expression of CMS-associated RNAs and/or proteins in the presence of restorer genes is the key to fertility restoration. Generally, CMS-associated RNAs are expressed ubiquitously, without spatial and temporal restrictions. It has been reported that the ORF79 protein accumulates specifically in the microspores of a BT-type CMS line but is absent from young seedling leaves (Wang et al., 2006). Meanwhile, another group found that the ORF79 protein accumulated in the calli of the BT-type CMS line (Kazama et al., 2008). We conclude that the CMS-associated protein ORFH79 is expressed in all tissues. It is unclear why the aberrant ORF impairs only pollen maturation even though the protein is also expressed in somatic cells. Recent results from our group suggest that ORFH79 is expressed throughout entire tissues and might be involved in the assembly of Complex V (Liu et al., 2012). In any case, the decrease in ORFH79 expression is doubtless the key for fertility restoration in HL-CMS rice.
Here, an artificial Mt-GRP162 driven under the ubiquitin promoter was introduced into the mitochondria. Mt-GRP162 bound to CMS RNAs, rather than reducing CMS RNA content, and consequently decreased the amount of ORFH79 protein in transgenic plants. The accumulation of CMS RNAs was not impaired, consistent with the view that Mt-GRP162 only binds to CMS-associated transcripts. These findings suggested that GRP162 might not cleave or degrade CMS RNAs directly and also implied the existence of an additional important factor in the RFC that performs CMS cleavage to achieve fertility restoration. ORFH79 was suppressed in the presence of Mt-GRP162, suggesting that Mt-GRP162 blocked the translation of ORFH79 at the post-transcriptional level, possibly coincident with the action of PPR-B/ORF687. These data strongly indicate that GRP162 can function as a novel restorer gene with the addition of a mitochondrial transit peptide.
The divergence of the Rf genes in development
To date, although several types of Rf genes have been identified, little is known about the mechanism of restoration. Rf genes eliminate the impairments caused by CMS RNAs and/or proteins through various mechanisms of action on different developmental processes. PPR proteins such as RF1A and RF1B decrease the accumulation of CMS RNA directly. Other PPR proteins, such as PPR-B/ORF687, can also reduce the amount of CMS protein rather than the amount of CMS RNA (Uyttewaal et al., 2008). Many types of Rf genes have developed, mainly due to the variety of developmental processes that affect fertility.
Previous studies showed that CMS RNAs are cleaved in the presence of Rf5, which assembles a RFC to cleave the CMS RNAs. GRP162 is involved in this fertility restoration process as a key CMS RNA-binding factor in the RFC. In this study, we successfully built a novel restorer gene based on our knowledge of the RFC. The Mt-GRP162 protein restored the fertility of the HL-CMS line by suppressing ORFH79 through another pathway. It also can be imagined that several natural pathways for restoration have arisen during evolution. The CMS line will produce completely non-functional pollen when the candidate Rf genes involved in the fertility restoration process are removed. This might be one reason why there are so many types of divergent CMS lines and Rf genes. The future discovery of additional novel Rf genes will clarify the nature of the interaction between the mitochondria and nuclei and allow heterosis to be used more efficiently to promote human food security.
The authors thank Professor Francesca M. Quattrocchio for technical assistance for rice protoplast. This work is supported by funds from the National High Tech Research and Development Project (2011AA101101), National Key Project for Research on Transgenic Plant (2011ZX08010), and Postdoctoral Science Foundation (2012M511658).