ORFB is the product of a gene that is conserved in plant mitochondrial genomes, and which, on the basis of sequence motif and structural similarity, is predicted to be the homologue of yeast and mammalian ATP8, part of the FO component of the F1 FO‐ATP synthase. We have shown that, in sunflower, orfB transcripts are edited, increasing the similarity of the predicted protein to ATP8 proteins from non‐plant species. Blue‐native polyacrylamide gel electrophoresis and peptide sequencing confirm that ORFB localizes to the ATP synthase complex. The predicted amino‐terminal 19 amino acids of ORFB are identical to those in the chimeric mitochondrial ORF522 protein, which is associated with cytoplasmic male sterility (CMS) in sunflower. Assays comparing respiratory complexes from a male‐sterile line expressing ORF522 with those from a male‐fertile line show a specific decrease in ATP hydrolysis by the ATP synthase. These observations allow us to propose a mechanism underlying CMS that is associated with the expression of chimeric open reading frames containing part of the orfB gene.
Plant mitochondrial genomes contain approximately 60 open reading frames (ORFs), most of which have been identified (Unseld et al., 1997). Only a few, such as orfB, orf25 and orfx, which are conserved across all plant mitochondrial genomes sequenced so far, have no known function. Gray et al. (1998), on the basis of a comparison of the genome sequences of a range of protists (for example, Reclinomonas americana), predicted that the ORFB protein in plants is the homologue of the yeast and mammalian ATP8. The atp8 gene encodes a core proteolipid subunit of the FO component of the F1 FO‐ATP synthase (complex V), which is thought to be involved in the assembly of FO (Hadikusumo et al., 1988; Mueller, 2000). However, the sequence conservation between ORFB and ATP8 is low and ORFB has not been shown to be a component of the plant F1 FO‐ATP synthase.
We, and others, have shown that chimeric mitochondrial genes that contain part, or all, of the orfB gene are involved in cytoplasmic male sterility (CMS) in several plant species (Laver et al., 1991; Schnable & Wise, 1998; Horn & Friedt, 1999; Nakajima et al., 2001). CMS is a maternally inherited trait that has been described in more than 150 plant species, and is characterized by the inability to produce functional pollen, although vegetative development is unaffected. The CMS phenotype, in conjunction with the nuclear restorer of fertility genes (Rf genes), is used in the production of F1 hybrid crops. In the best‐characterized systems, CMS is associated with aberrant recombination events in the mitochondrial genome that result in the generation of chimeric ORFs, which are expressed as novel polypeptides (Schnable & Wise, 1998). In sunflower, the PET1–CMS protein is associated with the expression of a novel mitochondrial gene, orf522, which is located downstream of the atp1 gene. orf522 is co‐transcribed with the atp1 gene, and is expressed as a mitochondrial polypeptide of ∼15 kDa in all tissues (Horn et al., 1991; Köhler et al., 1991; Laver et al., 1991). The male‐fertile phenotype can be restored by the introduction, in a cross, of nuclear Rf genes, which lead to a specific reduction in the levels of the atp1–orf522 co‐transcript and of the ORF522 protein in the male florets (Monéger et al., 1994; Smart et al., 1994). In addition, we have shown that increases in the level of polyadenylation of the atp1–orf522 transcript are correlated with the tissue‐specific instability of the transcript in the male florets of the hybrid plants with restored fertility (Gagliardi & Leaver, 1999).
In this study, we show that, in sunflower, ORFB is localized to the F1 FO‐ATP synthase complex, and that editing of the orfB transcript creates a leucine in the amino terminus that increases the level of sequence similarity with the N‐terminal amino acids of ATP8 from non‐plant species. This provides further evidence that ORFB is the plant equivalent of ATP8. In‐gel assays of the enzyme activities associated with the main enzyme complexes of the mitochondrial inner membrane revealed a specific reduction in the ATPase activity of the F1 FO complex in mitochondria from seedlings of the cytoplasmic male‐sterile line compared with that in fertile lines. In contrast, the activities associated with other respiratory complexes were unaffected. These observations allow us to propose a mechanism underlying CMS in higher plants in which part of the orfB gene contributes to the generation of novel chimeric genes.
ORFB localizes to the F1 FO‐ATP synthase complex
To determine whether ORFB is a subunit of the mitochondrial F1 FO‐ATP synthase (complex V), highly purified mitochondria were isolated from four‐day‐old, dark‐grown sunflower seedlings or florets, and the mitochondrial respiratory complexes were separated in their native forms by blue‐native polyacrylamide gel electrophoresis (BN‐PAGE). The component subunits in each complex were then resolved using SDS–PAGE in the second dimension (Fig. 1A). The complexes were identified by their order of separation in the gel on the basis of their known molecular weight, the patterns of subunit composition resolved in the second dimension (Jänsch et al., 1996) and by the use of histochemical staining to detect the characteristic enzyme activities associated with each complex (see below). The identity of the F1 FO‐ATP synthase (550–600 kDa) was confirmed further by peptide mass fingerprinting of the F1 α‐subunit (data not shown), and by immunological staining of the α‐ and β‐subunits (Fig. 1B) with an antibody against the whole F1 complex from spinach mitochondria (Hamasur et al., 1990). In sunflower mitochondria, this antibody recognizes the α‐ and β‐subunits and, to a lesser extent, the γ‐subunit. Combining with a 3:3:1 stoichiometry of α:β:γ only the α‐ and β‐subunits are usually detected. In addition, these two subunits are similar in terms of molecular weight (approximately 54 kDa each), which prevents good resolution of them, especially when tricine is used instead of glycine in the SDS–PAGE. This results in the detection of only one spot on immunoblots. A band from the F1 FO‐ATP synthase complex containing a polypeptide of approximately 15 kDa (indicated by an arrowhead in Fig. 1A) was excised from the gel. In‐gel trypsin digestion was then carried out on the excised band, and peptide mass fingerprinting was performed using matrix‐assisted laser desorption ionization/time‐of‐flight (MALDI–TOF) mass spectrometry. The spectrum obtained contained peptides matching those predicted for the mitochondrially encoded ORFB from sunflower (Monéger et al., 1994) and Arabidopsi s (GenBank accession number NP_085508). Mass spectrometry was carried out on two peptides that exactly matched the predicted molecular mass for ORFB in order to determine their amino‐acid sequences. The sequences of these peptides were identical to sequences found in the sunflower ORFB protein, as predicted from the genomic sequence (Fig. 1C; Quagliariello et al., 1990). These results provide the first biochemical evidence that ORFB is a subunit of the F1 FO‐ATP synthase.
The sunflower orfB transcript is edited
Editing of mitochondrial transcripts is a common phenomenon in plants, often leading to codon changes (Maier et al., 1996). To obtain a more accurate prediction of the amino‐acid sequence of sunflower ORFB, seven complementary DNAs corresponding to the orfB transcript were sequenced in both the forward and reverse directions, and were compared with the orfB genomic sequence. Three C to T conversions were found, in all of the cDNAs sequenced, at nucleotide positions 47, 58 and 452 of the coding sequence. Each of these results in the change of an amino‐acid residue as compared with the sequence predicted from the genomic DNA as follows: amino‐acid residues S16 to L, L21 to F and P151 to L (Fig. 2). Residues L16 and F21 are conserved in ORFB proteins from higher plants, and L16 is also conserved in ATP8 of the protist Reclinomonas americana and in the mammalian ATP8 orthologue. The editing site at nucleotide 47 (amino acid S16 to L) is also present in the related chimeric orf522 gene of sunflower (Monéger et al., 1994), confirming that the predicted N‐terminal 19 amino acids of ORFB and ORF522 are identical.
ATPase activity in mitochondria expressing ORF522
To determine whether there is a correlation between the presence of ORF522 and any change in the activity of the F1 FO‐ATP synthase that might be linked to pollen abortion in the sterile PET1–CMS line, we analysed the activity of the native ATP synthase in mitochondria isolated from seedlings of fertile, sterile and hybrid sunflower lines. Mitochondrial membranes were solubilized using N‐dodecyl β‐D‐maltoside, and the electron transport complexes and F1 FO‐ATP synthase were separated by BN‐PAGE. The ATPase activity of the F1 FO‐ATP synthase complex was visualized by histochemical staining on the basis of the precipitation of lead nitrate by the inorganic phosphate produced during ATP hydrolysis (Yoshida et al., 1975). The in‐gel ATPase activity of the isolated complex was reproducibly and significantly lower in mitochondria from the sterile compared with those from the fertile line, but was higher in the hybrid line (Fig. 3A, B). To investigate whether this pattern of activity was unique to the ATP synthase complex, the enzyme activities associated with complex I (NADH dehydrogenase), complex II (succinate dehydrogenase) and complex IV (cytochrome c oxidase) were visualized in parallel gels (Fig. 3A). Although slight differences were observed, they were not statistically significant (Fig. 3B), nor did they correlate with the consistently reproducible lower level of ATPase activity observed in the sterile line.
To investigate whether the decrease in ATPase activity was linked to a decreased protein level on the gel in the sterile line, the bands containing the F1 FO‐ATP synthase complex in the gel stained with Coomassie blue were scanned and quantified using Multi‐Analyst software (Biorad) (Fig. 3C). The variations in the intensities of total areas of the bands in all three lines were not significantly different, except for complex V. The apparent slight decrease in band intensity in complex I of the sterile line is not correlated with any significant change in the activity of the complex (Fig. 3B).
The existence of orfB in plant mitochondrial genomes was first reported in Oenothera by Hiesel et al. (1987), identified in sunflower by Quagliariello et al. (1990) and subsequently shown to be conserved in the mitochondrial genome of diverse plant species. Gualberto et al. (1991) reported that the ORFB protein is expressed extensively in wheat, suggesting that it has biological significance. However, it was not until the genome sequences of a range of protists became available that it was suggested, on the basis of structural and sequence motif similarities, that ORFB is the homologue of yeast and mammalian ATP8, a subunit of the FO component of F1 FO‐ATP synthase (Gray et al., 1998). However, the overall sequence similarity is weak, and the predicted molecular weight of ORFB (18.240 kDa) is much higher than that predicted for yeast ATP8 (5.822 kDa) and the human homologue (7.992 kDa). Our demonstration that ORFB is a subunit of the F1 FO‐ATP synthase complex (Fig. 1), together with our demonstration that editing of the orfB transcript increases the similarity of the N‐terminal 19 amino acids to ATP8 from non‐plant species, provides further support for the above suggestion (Fig. 2).
Numerous studies have shown that cytoplasmic male sterility in plants is associated with aberrant recombination events in the mitochondrial genome, resulting in the generation of chimeric ORFs that are expressed as novel polypeptides (Schnable & Wise, 1998). The sequences that contribute to the generation of these chimeric orfs are usually derived from coding and non‐coding regions of existing genes, but are occasionally from unknown origins. ORF522, the chimeric protein responsible for the CMS phenotype in sunflower, shares similarities with ORFB in its N‐terminal amino‐acid sequence, polypeptide domain structure and molecular weight (Fig. 2). The sequence identity of the N‐terminal 19 amino acids of ORFB and ORF522, coupled to a novel carboxy‐terminal domain, could result in competition between the two proteins, leading to impaired biogenesis and uncoupling or decreased phosphorylation activity of the F1 FO‐ATP synthase complex (Balk & Leaver, 2001). In support of this, we have shown, by comparing the in‐gel enzyme activities associated with several mitochondrial respiratory complexes between fertile and sterile lines, that the ATPase activity of the F1 FO‐ATP synthase is specifically and significantly reduced in the sterile line as compared with the fertile line. The reduction in ATPase activity in the male‐sterile line correlates with the expression of ORF522, and is restored in the presence of nuclear restorer of fertility genes in the fertile hybrid line (Fig. 3).
Segments of orfB and atp9 have been shown previously to be involved in the generation of the chimeric proteins associated with CMS in a variety of plant species (Schnable & Wise, 1998). Some of these orfB‐related CMS genes include either part of the sequence of the orfB gene, for example, the CMS‐associated orf224 in Brassica, orf222 and orf138 in radish (Schnable & Wise, 1998), orf324 in sugar beet (Kubo & Mikami 1996), or the entire sequence of the gene, as in carrot (Nakajima et al., 2001). The derivation of these chimeric sequences from progenitor genes that encode subunits of the FO portion of the ATP synthase, suggests that the molecular mechanisms underlying all of these CMS phenotypes may be similar to that for PET1–CMS in sunflower, which is proposed in this work.
It is interesting to speculate as to why mutations in the mitochondrial genome in plants only seem to affect cellular development during male flower development (male gamete production), whereas vegetative development and female flower development are unaffected. We have demonstrated previously that high levels of mitochondrial gene expression, associated with a marked increase in mitochondrial biogenesis (Smart et al., 1994), occur in the meiocyte cells, which give rise to four independent haploid microspores (pollen grains) during anther development in sunflower. The data presented here support the hypothesis that the increased demand for respiratory function and cellular energy in the form of ATP generated by oxidative phosphorylation during anther development may be compromised by the expression of ORF522. The inability to meet the increased energy demand required to sustain anther development may therefore not be met in the sterile line, leading to pollen abortion. Our data also support an alternative hypothesis that we proposed recently, which suggests that impairment of the capacity for mitochondrial oxidative phosphorylation in sterile sunflower lines leads to premature programmed cell death during anther development (Balk & Leaver, 2001). It is of interest to note that the F1 FO‐ATP synthase is thought to have a role in apoptosis in mammals, and possibly yeast, although the precise mechanism remains to be determined (reviewed in Matsuyama & Reed, 2000).
Unexpectedly, we observed that the ATPase activity in hybrid sunflower seedlings was increased as compared with that of fertile and sterile seedlings. It is tempting to speculate that an increased activity or efficiency of the F1 FO‐ATP synthase in the F1 hybrid, by an as yet unknown mechanism, could contribute to the phenomenon known as ‘hybrid vigour’, one of the important goals of plant breeding.
Sunflower (Helianthus annuus L.) lines and growth conditions were as described by Smart et al. (1994).
Analysis of editing sites.
Genomic DNA and orfB cDNAs were amplified, cloned and sequenced using standard molecular biological techniques. Sequences of the primers used in this analysis are available on request to the authors.
Isolation of mitochondria.
Mitochondria from sunflower florets and four‐day‐old etiolated seedlings were isolated by differential centrifugation and gradient purification as described previously (Leaver et al., 1983).
Blue‐native (first dimension) and tricine–SDS (second dimension) polyacrylamide gel electrophoresis.
Sample preparation, electrophoresis conditions, gel fixation and Coomassie blue staining were adapted from Jänsch et al. (1996) and Schägger & von Jagow (1991). A final ratio of 1 g of detergent (N‐dodecyl β‐D‐maltoside) to 1 g of protein gave the best solubilization of respiratory complexes from sunflower mitochondria. Western blot analysis was performed as described in Balk & Leaver (2001).
MALDI–TOF and mass spectrometry for peptide fingerprint analysis.
The in‐gel digestion and mass spectrometry was performed as a service by GARNet (http://www.york.ac.uk/res/garnet/garnet.htm). The results from mass spectrometry data were analysed using the MASCOT search program, available at the website http://www.matrixscience.com/.
In‐gel activity assays of respiratory complexes.
After electrophoresis, gels were rinsed briefly with MilliQ water, equilibrated in reaction buffer without reagents for 10 min and incubated in fresh buffer with reagents (Zerbetto et al., 1997; Jung et al., 2000). Gels were then fixed and destained in 45% (v/v) methanol and 10% (v/v) acetic acid.
We thank E. Glaser for the F1 antibody. The work was supported by a grant from the Biotechnology and Biological Sciences Research Council (C.J.L.), the Gatsby Charitable Foundation and Lincoln College Oxford (J.B.).
- Copyright © 2003 European Molecular Biology Organisation