What Does the Molecular Genetics of Different Types of Restorer-of-Fertility Genes Imply?

Abstract

Cytoplasmic male sterility (CMS) is a widely used trait for hybrid seed production. Although male sterility is caused by S cytoplasm (male-sterility inducing mitochondria), the action of S cytoplasm is suppressed by restorer-of-fertility (Rf), a nuclear gene. Hence, the genetics of Rf has attained particular interest among plant breeders. The genetic model posits Rf diversity in which an Rf specifically suppresses the cognate S cytoplasm. Molecular analysis of Rf loci in plants has identified various genes; however, pentatricopeptide repeat (PPR) protein (a specific type of RNA-binding protein) is so prominent as the Rf-gene product that Rfs have been categorized into two classes, PPR and non-PPR. In contrast, several shared features between PPR- and some non-PPR Rfs are apparent, suggesting the possibility of another grouping. Our present focus is to group Rfs by molecular genetic classes other than the presence of PPRs. We propose three categories that define partially overlapping groups of Rfs: association with post-transcriptional regulation of mitochondrial gene expression, resistance gene-like copy number variation at the locus, and lack of a direct link to S-orf (a mitochondrial ORF associated with CMS). These groups appear to reflect their own evolutionary background and their mechanism of conferring S cytoplasm specificity.

1. Introduction

Cytoplasmic male sterility (CMS) is a genetic character that prevents plants from producing functional pollen [1]. In most cases, CMS affects nothing but the male reproductive organ; however, some pleiotropic effects are known such as disease susceptibility and flower morphology (e.g., narrow petals) [2,3]. Due to its maternal inheritance, CMS is widely used in plant breeding to prepare seed parents for hybrid-seed production [1]. In other cases, CMS is a potential obstacle for crossing between a breeding line and genetic resources [4,5,6]. Controlling CMS expression is one of the major challenges in plant breeding.

In the widely held genetic model, the male-sterility inducing cytoplasm is referred to as S cytoplasm, whereas cytoplasm without the male-sterility inducing factor is termed N cytoplasm. As was shown in some initial genetic analyses (e.g., [7]), CMS is governed not only by S cytoplasm but also is affected by nuclear genes because offspring from a CMS plant do not always express male sterility depending on the pollen parental line. Differences among such nuclear genotypes cannot be seen in the presence of N cytoplasm, hence a genetic interaction between S cytoplasm and a nuclear gene is postulated. In many cases, a nuclear suppressor gene is proposed as a restorer-of-fertility in the genetic model where a dominant allele (Rf) suppresses S. An Rf affects nothing when expressed in a different source of S cytoplasm, indicating that the Rf specifically suppresses the cognate S. This observation means that differences in S cytoplasm can be defined by differences in Rf [8]. Note that Rf genes have been named in each plant species; consequently, maize Rf2 (Rf2a) and rice Rf2 are not orthologous genes. Unraveling the genetics of Rf is a prerequisite for the efficient use of CMS in hybrid seed production, otherwise discerning the Rf genotype for plant breeding would be laborious [9].

Molecular analysis of CMS has associated the S cytoplasm with mitochondrial open reading frames (ORFs) composed of pieces of duplicated genes and/or origin-unknown sequences [10,11,12], whereas other studies show associations with truncated mitochondrial genes, mitochondrial non-coding RNAs, and over-expression of mitochondrial genes [13,14,15]. Plants that have lost S cytoplasm-associated ORFs (hereafter S-orfs) from the mitochondrial genome have been obtained by mutation, somaclonal segregation or mitochondrial genome editing [10,16,17]. The phenotypes of these plants indicate that S-orf is responsible for male sterility but plays no other role in plant development.

The identified S-orfs differ in their primary sequences [11], hence knowing the nucleotide sequence of a mitochondrial genome is insufficient for identifying an S-orf. Nonetheless, alteration of S-orf expression in a fertility-restored plant by Rf is frequently observed [10]. Based on this finding, a common approach for identifying S-orfs is to find an alteration in gene expression of a mitochondrial gene/ORF by Rf. Note that, although this approach is broadly applicable for CMS study, there are some exceptions as discussed below.

Since the first molecular cloning of Rf [18], a number of Rf loci have been analyzed to identify Rfs or Rf candidates. As reports increase, Rf-gene products are now known to be numerous and variable (Table 1). We raise the question as to whether each of these Rfs is an idiosyncratic example that episodically emerged during plant-species diversification or is there a tendency or trend underlying Rf evolution. To approach this question, we will briefly review plant Rfs. Recent reviews on CMS, Rf and other types of male sterility will help to understand the progress of this research area [19,20,21,22,23,24,25,26,27,28,29,30,31,32]. This question has also been addressed from the viewpoint of evolutionary genetics (reviewed ib [33]).

2. Pentatricopeptide Repeat Protein is the Most Abundant Type of Rf Gene Product

Pentatricopeptide repeat (PPR) proteins have a succession of degenerate ~35 amino acid sequence motifs with the potential for binding to single-stranded RNA in a nucleotide-sequence specific manner [45]. PPR protein genes are known to form one of the largest gene families among land plant genomes [45]. In most cases, PPR protein gene products are thought to be imported into mitochondria, plastids, or both [45]. PPR protein genes play pivotal roles mainly in post-transcriptional mechanisms and are subdivided into several classes based on the length of the PPR motifs and the presence/absence of additional domains [45]. According to this classification, most of the Rfs encoding PPR proteins (hereafter PPR Rf) belong to the P class (reviewed in [34,35]), in which canonical repeats occupy almost the entire coding region [45]. To date, the P-class PPR protein is the most frequently reported class of Rf-gene product.

The molecular action of most PPR Rfs involves a reduction in S-ORF-protein accumulation in the cognate S cytoplasm [34]; however, details of the action differ among plants. For example, altering the transcript profile of the S-orf is the case for rice Rf1a but not for radish Rfo/Rfk [34]. Additionally, in spite of the notion that PPR protein has the potential to bind RNA, gene products of PPR Rf do not necessarily bind with the cognate S-orf RNA even when accumulation of the S-ORF protein is reduced [34]. Another factor has been reported to mediate between PPR RF protein and S-orf mRNA [46].

Sorghum Rf1 and barley Rfm1 are associated with a PLS-DYW-class PPR protein, where long- and short versions of PPR motifs are included in the repeat array and an extra domain, termed DYW, is added at the carboxyl terminus [36,37]. As PLS-DYW-class PPR proteins almost exclusively play a role in RNA editing, a post transcriptional process that converts specific cytidine residues into uridine [47], it is possible that sorghum Rf1 and barley Rfm1 restore pollen fertility by editing S-orf transcripts or other target RNAs that are currently unknown. Molecular analysis of the cognate S cytoplasm is necessary.

3. What Do Rfs Encode Other than PPR Proteins?

The interaction between nuclei and mitochondria is very complex in plants, as it is in other eukaryotes [48]. Hence, it may be not surprising that some Rfs encode proteins other than PPR proteins. We refer to such Rfs as non-PPR Rfs.

Maize Rf2a is the first Rf whose gene product was identified and shown to encode a mitochondrial aldehyde dehydrogenase (ALDH) [18,42]. Its recessive alleles (i.e., male-sterility inducing alleles) have a transposon insertion or a missense mutation that reduces the level of gene expression or abolishes enzymatic activity, respectively [42]. Maize Rf2a, in conjunction with Rf1, restores male fertility to plants with CMS-T, one of the three S cytoplasms in maize [10]. The S-orf of CMS-T is identified as a chimeric ORF consisting of different parts of rrn26, but Rf2a does not affect transcripts or protein products of this S-orf [10]. The target molecule of Rf2a in anthers is unknown, but recombinant RF2A protein expressed in E. coli can oxidize a broad range of aldehydes [49].

Maize Rf4 is a restorer of another maize S cytoplasm called CMS-C [43]. Its gene product is a basic helix-loop-helix (bHLH) transcription factor [43]. Surprisingly, the RF4 protein has a nuclear-localization signal, suggesting that the protein may localize outside of mitochondria [43]. This finding appears to discount the prevailing notion that Rf gene products are all imported into mitochondria. The identity of the S-orf in CMS-C mitochondrial genomes is unknown, although three candidates have been found (discussed by Allen et al. [50]), making the mechanism for how Rf4 restores pollen fertility more elusive. Maize Rf4 is an allele of Ms23 that was originally identified as playing an essential role in the differentiation of tapetal tissue in anthers [51]. Null alleles of ms23 cause male sterility even in N cytoplasm plants, whereas the rf4 allele (inducing male sterility in CMS-C but male fertile in N cytoplasm) has an amino acid substitution that is potentially harmful to the stability of a heterodimer composed of RF4 and another protein bHLH51 [43]. The male-sterility inducing rf4 allele appears to be leaky.

Rice Rf17 encodes a protein that is partially homologous to an acyl-carrier protein synthase that is likely imported into mitochondria [44]. Although a candidate S-orf was identified, no alteration in the transcript profile was observed [52]. Rf17 restores pollen fertility in a gametophytic manner (i.e., those pollen grains receiving a restoring allele are fertile but otherwise sterile). Molecular analysis associated a reduction in the expression of the restoring allele with fertility restoration [44]. This notion was confirmed by an observation that ectopic expression of RF17 protein canceled fertility restoration [44]. A working hypothesis is that RF17 transmits a retrograde signal that is emitted from S mitochondria to exert male sterility, but the reduced expression of Rf17 results in the blockage of this transmission [44]. The allelic difference is attributed to a single nucleotide substitution in the region upstream of the start codon [44].

In contrast to the abovementioned Rfs, some non-PPR Rfs are known to or potentially affect S-orf expression in different ways. Rice Rf2 reduces the amount of dicistronic mitochondrial transcripts encoding atp6 and orf79 (an S-orf of the cognate S cytoplasm), whereas monocistronic atp6 mRNA remains unaffected, resulting in the reduction of ORF79 protein accumulation [53]. Positional cloning of rice Rf2 revealed that this gene encodes a protein having a characteristic glycine-rich region [38], but whether there is molecular interaction between RF2 protein and orf79 transcript is unclear [53]. The male-sterility inducing allele is expressed to a level comparable to that of the restoring allele, but it has a single amino acid substitution [38].

Barley Rfm3 is genetically associated with two linked genes encoding mitochondrial transcription termination factor family (mTERF) proteins [39]. Interestingly, rye Rfp1 shows tight linkage with an mTERF protein gene [40]. mTERF protein, as well as PPR protein, belongs to a large family of helical-repeat proteins capable of binding to nucleic acids [54]. The copy number of mTERF protein genes in plants exceeds that of vertebrates; at least 33 and 28 mTERF protein genes have been identified in the Arabidopsis and rice genomes, respectively, whereas only four mTERF protein genes are found in human and mouse [55]. A few of the plant mTERF protein genes have been characterized; their gene products are imported into mitochondria or plastids to bind DNA or RNA and are involved in transcription, splicing, or tRNA maturation [54]. Further study is necessary to elucidate the identity and targets of Rfm3 and Rfp1.

Sugar beet Rf1 is a duplicated gene of Oma1 [41,56]. Oma1 (named after overlapping activity with m-AAA protease) is known to be involved in quality control of mitochondria and mitochondrial dynamics in yeast and mammals [57,58,59]. Sugar beet S cytoplasm is associated with an S-orf termed preSatp6 whose origin is unknown [60]. There is no observable effect of Rf1 on the transcript size or the amount of preSatp6 translation product [60,61]; however, whereas preSATP6 protein is detected from a 250-kDa complex (presumably a homo-oligomer form) in CMS anthers, the 250-kDa complex disappears in fertility restored anthers and, instead, novel complexes appear [60]. RF1 protein expressed from a restoring allele can bind to preSATP6 protein, but the gene product from a non-restoring allele does not show such activity [60]. The Rf1 appears to alter the higher order structure of preSATP6 protein without degradation, as do some molecular chaperones. These findings also imply that a higher order structure of preSATP6 protein is an important factor for CMS expression, whereas the preSATP6 protein per se may be less harmful for pollen production [60].

4. How Have S-orf and Rf Evolved? An organizational comparison between S- and N mitochondrial genomes and between Rf and rf alleles has been conducted to identify associated genes and to identify the molecular basis of relevant DNA markers. The results have invoked studies on the evolutionary aspects of S-orf and Rf. 4.1. Evolution of S-orf

The size of plant mitochondrial genomes ranges from 66 kbp to 11.3 Mbp but is usually 200–700 kbp or more than 12 times larger than mammalian counterparts [25,62]. Despite their large size, the number of plant mitochondrial genes is 50–60, less than twice that of mammals [25]. Hence, the plant mitochondrial genome is referred to as a gene sparse genome, a term that means the existence of large intergenic regions [25]. Most S-orfs occur in the vicinity of a mitochondrial gene (in many cases, genes encoding subunits of ATP synthase) [11]. S-orfs often exhibit a chimeric nature in which pieces of mitochondrial gene fragments are joined together [10]. The chimeric nature of S-orf reflects the principal mechanism for plant mitochondrial genome diversification in which frequent inter- and intra-molecular recombination plays a very important role [25]. In fact, duplication of DNA segments and rearrangement of the mitochondrial genome between closely related species or even within species can be explained by this mechanism [25]. It is likely that S-orfs independently evolved in each plant lineage as by-products of plant mitochondrial evolution.

4.2. Evolution of PPR Rfs

The PPR Rf locus frequently contains arrayed gene copies that are highly similar to PPR Rf [35], suggesting the occurrence of local gene duplication. Genes having similarity with PPR Rf are designated restorer-of-fertility like (RFL) genes [63]. Organizational differences between restoring and non-restoring alleles of a PPR Rf locus often involve copy number variation (CNV) of RFL genes; detailed comparison between the alleles revealed traces of interallelic recombination with unequal crossing over [34]. Such an evolutionary mechanism could result in the production of multiple, different molecular variants with various numbers of RFL copies within an Rf locus. In fact, analysis of the rice Rf1 (a typical PPR Rf) locus using 59 accessions of Oryza genetic resources revealed the existence of at least six molecular variants with CNV [64].

RFL and PPR Rf (except for those encoding PLS-DYW class proteins) constitute a unique subgroup within the P-class of PPR genes. RFL is almost ubiquitous in plant genomes even when the plant appears to have no relationship to CMS, such as poplar (a dioecious plant species) [63,65,66,67]. RFL copy number in a genome varies among plant species (sometimes close to 40 copies) [68]. RFL copies tend to cluster with each other in several chromosomal regions, but the location of the cluster is less preserved even between closely related species. For this reason, RFL is said to be “nomadic” in evolutionary scale [69]. According to a phylogenetic analysis, RFL already existed before the split between monocot- and dicot lineages. The current RFL members form species-specific clusters that are paralogous to each other, a similar evolutionary pattern observed in pathogen resistance (R) genes [34].

RFL is thought to participate in the quality control of the transcript pool in plant mitochondria [34]. In plant mitochondrial genomes, many ORFs that are obviously distantly related or unrelated to genuine mitochondrial genes occur in the large intergenic regions [70]. As transcriptional control in plant mitochondria is relaxed [34], potentially harmful RNAs including those encoding aberrant proteins can be generated [71]. A post-transcriptional control system to cope with such harmful RNA is employed by plant mitochondria, and RFL is an important component of this system. The emergence of PPR Rf from RFL is a very likely scenario.

4.3. Evolution of Non-PPR Rfs

The organizational diversity of non-PPR Rfs in genetic resources is less investigated. Several haplotypes based on single nucleotide polymorphisms and small insertion/deletions were identified in two linked mTERF protein genes at the barley Rfm3 locus, and one haplotype is exclusively found from a non-restoring genotype [39].

CNV within a non-PPR Rf locus has rarely been reported to date. An exceptional case is the sugar beet Rf1 locus, in which Rf1-like genes are tandemly clustered [56], reminiscent of the PPR Rf locus. In fact, CNV is seen in the Rf1 locus of sugar beet and other Beta vulgaris genetic resources (e.g., Arakawa et al. [72]), suggesting that a similar evolutionary mechanism to that of PPR Rf may be involved (i.e., interallelic recombination with unequal crossing over).

The genetic function of multiple Rf1 molecular variants is largely unknown. A handful of these variants have been subjected to genetic analysis, and semi-dominant- (fertility restoration is insufficient when heterozygous) and hypomorphic (the protein product has the potential to bind with preSATP6 protein but barely restores fertility due perhaps to an insufficient amount of mRNA) alleles were identified, in addition to dominant and recessive alleles [72,73]. This observation indicates multiple allelism of the Rf1 locus. The number of beet Rf1 molecular variants is increasing as more B. vulgaris genetic resources are investigated [74]. One could raise the question of the significance of a series of different strength Rfs.

The clustered genes in the sugar beet Rf1 locus were designated RF1-Oma1 [56]. From the sugar beet genome, another Oma1 homologue that is more similar to Arabidopsis Oma1 (atOma1) was found and named bvOma1. In yeast, a Zn²⁺-binding motif in the peptidase M48 domain was shown to be critical for Oma1 function [57]. The peptidase M48 domain with a functional Zn²⁺- binding motif was predicted from the translation products of atOma1 and bvOma1 but not from RF1-Oma1 due to a critical amino acid substitution [56]. The occurrence of multiple Oma1 homologues in the sugar beet genome contrasts with Arabidopsis and rice, both of which have a single Oma1 copy as is also the case for yeast and humans [41]. An Arabidopsis plant with a defect in atOma1 is viable but has malfunctioning oxidative phosphorylation [75]. A comparison of the proteins’ ability to bind with preSATP6 and the expression pattern among RF1-Oma1, bvOma1 and atOma1 is summarized in Table 2; neither bvOma1 nor atOma1 had binding activity with preSATP6 protein [56]. The expression pattern during anther development was different among the three genes, of which only RF1-Oma1 was expressed at the meiosis stage [56]. Our unpublished data indicated that RF1-Oma1 is likely a paralogue of Oma1 [76]. Taken together, these results suggest that neofunctionalization may be favored to explain beet Rf1 evolution in which an ancestral Oma1 (possibly represented by atOma1) was duplicated, then the duplicated copy acquired several novel functions to evolve into RF1-Oma1.

Neofunctionalization might be, however, an over-simplification considering the difference in expression patterns between bvOma1 and atOma1; the difference is characterized by the lack of expression at the tetrad stage in bvOma1 (Table 2). Interestingly, RF1-Oma1 is expressed at the tetrad stage, which appears to complement this missing expression. Another molecular evolutionary mechanism posits that two duplicated genes can escape from evolutionary decay when each has a unique expression pattern and they complement each other to fulfil the expression pattern of their parental gene [77]. Whether the apparent degenerative expression pattern in bvOma1 is involved in the evolution of RF1-Oma1 is unknown. The important question is whether RF1-Oma1 has functions other than altering the higher order structure of preSATP6 protein. Concerning this question, the organization of rf1 alleles from different origins was investigated by analyzing sugar beet lines selected for the non-restoring genotype [78]. Interestingly, all genotypes had preserved, intact RF1-Oma1 copies despite lacking the ability to bind with preSATP6 protein [60,78], strongly suggesting an unknown function of RF1-Oma1. A more detailed study will be necessary.

5. Conclusions and Perspectives The molecular genetics of Rf has been deepened by the studies on PPR Rf. Conversely, studies on non-PPR Rf have shed light on aspects that cannot be investigated by examining PPR Rf. Clearly, both types of studies are complementary to each other and are essential for understanding Rf and S cytoplasm function.

Non-PPR Rf now becomes a group of miscellaneous gene products; however, it seems possible to find some shared features between different types of Rfs or with PPR Rf. Our present investigation is to introduce new criteria into Rf grouping instead of that solely based on PPR. As shown below, we propose three groups that are not necessarily mutually exclusive (Table 1). Some remaining questions are also posed.

Group 1: Association with a Post-Transcriptional Mechanism for Regulating Mitochondrial Gene Expression.

No doubt, the number of Rfs associated with a post-transcriptional mechanism is highest due to the large number of PPR Rfs [34,35]. Some non-PPR Rfs are also associated with post-transcriptional mechanisms such as in rice Rf2 [38]. It is possible that barley Rfm3 is associated with this mechanism by analogy with mTERF protein function (this might also be the case for rye Rfp1) [39,40,54], but identifying the mitochondrial gene responsible for CMS and studying the effect of Rf at the molecular level is necessary.

The prevalence of this group in Rf is reminiscent of genomic and proteomic studies in which a large portion of the plant mitochondrial proteome is composed of proteins capable of RNA binding or RNA processing, suggesting the importance of post-transcriptional mechanisms in plant mitochondria [47]. The abundance of genes for such proteins could be a rich resource for Rf evolution.

Group 2: R Gene-Like CNV at the Locus.

Loci of PPR Rf and sugar beet Rf1 (encoding OMA1-like protein) exhibit CNV in genetic resources [34,35,64,72]. This pattern of diversification is analogous to the R gene that evolves to convey pathogen resistance, whereas the pathogen evolves to overcome the R gene, a situation referred to as the “arms race” [63,79]. In terms of evolutionary genetics, mitochondrial and nuclear genomes are two conflicting parties because the former is inherited maternally, meaning that the male gamete is useless for mitochondria, whereas the male gamete is an important vehicle for the bi-parentally inherited nuclear genome. Hence, whereas mitochondria welcome CMS, the nuclear genome responds by implementing Rf [80].

R gene products usually detect pathogen elicitors in either a direct or indirect way to initiate a defense response [81]. Note that specific physical interaction between an RF protein and an S-orf transcript or the protein is the key to the fertility restoration of Group 2 members, such as some PPR Rf and sugar beet Rf1 [60,82,83]. Possibly, the role of Group 2 members is to detect male-sterility inducing factors, analogous to the relationship between the R gene and an elicitor. However, reports on rice Rf5 (identical to rice Rf1) and Rf6 (encoding a P-class PPR protein) in which the two Rfs showed no binding activity with the cognate S-orf [46,84] are the caveats of this proposed function. It is currently unknown what shapes the CNV at an Rf locus.

Group 3: Lack of a Direct Link with S-orf. This group includes maize Rf2a, maize Rf4 and rice Rf17, all of which lack evidence demonstrating that they affect S-orf transcripts or proteins. Note that the S-orf is uncharacterized in maize CMS-C (the cognate S cytoplasm of Rf4) and rice CMS-CW (that of Rf17).

Interestingly, the null alleles of maize rf2a generated by transposon mutagenesis greatly reduce male fertility when combined with an N cytoplasm [42]. Another recessive allele of maize rf2a recovered from a breeding line preserved an intact ORF that was expressed to the level of the dominant allele but had amino acid substitutions [42]. These findings parallel the relationship between ms23 and rf4, the former is a null allele causing tapetum abnormality and the latter is a missense allele in the same locus [43]. These results suggest that, in this group, recessive alleles that can be used for CMS expression but should secure pollen production with an N cytoplasm might be weak alleles at the molecular level. The role of such weak alleles possibly may be increasing anther susceptibility to the S cytoplasm for CMS expression. If so, the Rfs of this group would be a useful tool for determining how the S cytoplasm induces male sterility. The link between S-orfs and the Rfs of this group will be more obvious following a detailed physiological and developmental study.

A puzzling question remains. In the genetic model, Rf specifically suppresses the cognate S cytoplasm (hence the Rf can be used to diagnose S cytoplasm) [8]. Given that the molecular basis of S cytoplasm is associated with S-orf, the differences in S cytoplasms are attributable to the differences in S-orfs. In the cases where specific binding of RF protein with an S-orf transcript or the protein is critical for fertility restoration (e.g., some Rfs in Group 2), the genetic specificity of an Rf to the cognate S cytoplasm is easily explained. However, such data are missing in many cases, and the question of why an Rf specifically restores the cognate S cytoplasm remains unanswered. This issue is conspicuous in the cases of Group 3 Rfs. Possibly, each S cytoplasm has a unique mechanism to induce male sterility, and the Rf arbitrarily targets one of the steps within the entire cascade. Alternatively, the mechanism for male-sterility induction is generally common among S cytoplasms [85], but one (or some) of the steps within the entire cascade may be uniquely vulnerable in each S cytoplasm and is, thus, exposed as a target of Rf.

In conclusion, some of the different types of Rfs can be grouped by molecular genetic features other than PPR. The reason why these features work is not well understood, but the involvement of their evolutionary background is possible. Cross-disciplinary studies including molecular genetics, evolution, physiology and developmental studies will be necessary to refine the details of Rf and CMS. The outcome will be beneficial to plant breeding where controlling CMS is highly desired.

Gene Product 1		Post-transcriptional Mechanism in Mitochondria	R Gene-like CNV at the Locus	Direct Link with S-orf	References
PPR protein	P-class	Yes	Yes	Yes	[34,35]
	SPL-DYW	Undetermined	Undetermined	Undetermined	[36,37]
GR protein		Yes	Undetermined	Undetermined	[38]
mTERF protein		Undetermined	Undetermined	Undetermined	[39,40]
OMA1-like protein		No	Yes	Yes	[41]
ALDH		No	No	No	[18,42]
bHLH		No	No	No	[43]
ACP-like protein		No	No	No	[44]

1 PPR, pentatricopeptide repeat protein; GR, glycine-rich; mTERF, mitochondrial transcription termination factor family; OMA1, a gene symbol named after overlapping activity with m-AAA protease; ALDH, aldehyde dehydrogenase; bHLH, basic helix-loop-helix transcription factor; and ACP, acyl-carrier protein synthase.

Gene Name 2	Protein Product to Bind with PreSATP6	mRNA Detection
		Tissue	Anther developmental stage
			Meiosis	Tetrad	Microspore
RF1-Oma1	Yes (dominant allele)/ No (recessive allele)	Tapetum	+	+	-
		Meiocyte/Tetrads/ Microspore	+	+/-	-
bvOma1	No	Tapetum	-	-	+
		Meiocyte/Tetrads/ Microspore	-	-	+
atOma1	No	Tapetum	-	+	+
		Meiocyte/Tetrads/ Microspore	-	+	+

1 Based on [55]. 2 RF1-Oma1, Oma1-like genes clustered in the sugar beet Rf1 locus; bvOma1 and atOma1, probable Oma1 orthologues of sugar beet and Arabidopsis, respectively.

Author Contributions

Conceptualization, T.K., T.A. and K.K.; Writing-Original Draft, T.A., Y.H. and K.K.; Writing-Review & Editing, T.K.; Visualization, T.K. and T.A.; Funding Acquisition, T.K. All authors have read and agreed to the published version of the manuscript.

Funding

This work is partly supported by JSPS KAKENHI [Grant Number 18K05564] and NARO Bio-oriented Technology Research Advancement Institution (BRAIN) (Research program on development of innovative technology [Grant Number 30001A]. TA was a recipient of a JSPS Research Fellowship for Young Scientists [Grant Number 16J01146].

Acknowledgments

Part of this work is based on our experiments at the Field Science Center for Northern Biosphere, Hokkaido University.

Conflicts of Interest

The authors declare no conflict of interest.

References

1. Kaul, M.H.L. Male Sterility in Higher Plants; Springer: Berlin/Heidelberg, Germany, 1988; pp. 3-14.

2. Levings, C.S., III. Thoughts on cytoplasmic male sterility in cms-T maize. Plant Cell 1993, 5, 1285-1290.

3. McVetty, P.B.E.; Pinnisch, R.; Scarth, R. The significance of floral characteristics in seed production of four summer rape cultivar A-lines with pol cytoplasm. Can. J. Plant Sci. 1989, 69, 915-918.

4. Goto, S.; Yoshioka, T.; Ohta, S.; Kita, M.; Hamada, H.; Shimizu, T. QTL mapping of male sterility and transmission pattern in progeny of Satsuma mandarin. PLoS ONE 2018, 13, e0200844.

5. Hosaka, K.; Sanetomo, R. Development of a rapid identification method for potato cytoplasm and its use for evaluating Japanese collections. Theor. Appl. Genet. 2012, 125, 1237-1251.

6. Donoso, J.M.; Eduardo, I.; Picañol, R.; Batlle, I.; Howad, W.; Aranzana, M.J.; Arús, P. High-density mapping suggests cytoplasmic male sterility with two restorer genes in almond X peach progenies. Hortic. Res. 2015, 2, 15016.

7. Jones, H.A.; Clarke, A.E. Inheritance of male sterility in the onion and the production of hybrid seed. Proc. Am. Soc. Hortic. Sci. 1943, 43, 189-194.

8. Duvick, D.N. Cytoplasmic pollen sterility in corn. Adv. Genet. 1965, 13, 1-56.

9. Bosemark, N.O. Genetics and breeding. In Sugar Beet; Draycott, A.P., Ed.; Oxford: Blackwell, UK, 2006; pp. 50-88.

10. Schnable, P.S.; Wise, R.P. The molecular basis of cytoplasmic male sterility and fertility restoration. Trends Plant Sci. 1998, 3, 175-180.

11. Hanson, M.R.; Bentolila, S. Interactions of mitochondrial and nuclear genes that affect male gametophyte development. Plant Cell 2004, 16, S154-S169.

12. Chase, C.D. Cytoplasmic male sterility: A window to the world of plant mitochondrial-nuclear interactions. Trends Genet. 2007, 23, 81-90.

13. Ducos, E.; Touzet, P.; Boutry, M. The male sterile G cytoplasm of wild beet displays modified mitochondrial respiratory complexes. Plant J. 2001, 26, 171-180.

14. Stone, J.D.; Koloušková, P.; Sloan, D.; Štorchová, H. Non-coding RNA may be associated with cytoplasmic male sterility in Silene vulgaris. J. Exp. Bot. 2017, 68, 1599-1612.

15. Szklarczyk, M.; Szymański, M.; Wójcik-Jagła, M.; Simon, P.W.; Weihe, A.; Börner, T. Mitochondrial atp9 genes from petaloid male-sterile and male-fertile carrots differ in their status of heteroplasmy, recombination involvement, post-transcriptional processing as well as accumulation of RNA and protein product. Theor. Appl. Genet. 2014, 127, 1689-1701.

16. Budar, F.; Pelletier, G. Male sterility in plants: Occurrence, determinism, significance and use. CR Acad. Sci. Paris Life Sci. 2001, 324, 543-550.

17. Kazama, T.; Okuno, M.; Watari, Y.; Yanase, S.; Koizuka, C.; Tsuruta, Y.; Sugaya, H.; Toyoda, A.; Itoh, T.; Tsutsumi, N.; et al. Curing cytoplasmic male sterility via TALEN-mediated mitochondrial genome editing. Nat. Plants 2019, 5, 722-730.

18. Cui, X.; Wise, R.P.; Schnable, P.S. The rf2 nuclear restorer gene of male-sterile T-cytoplasm maize. Science 1996, 272, 1334-1336.

19. Chen, L.; Liu, Y.G. Male sterility and fertility restoration in crops. Annu. Rev. Plant Biol. 2014, 65, 579-606.

20. Touzet, P.; Meyer, E.H. Cytoplasmic male sterility and mitochondrial metabolism in plants. Mitochondrion 2014, 19, 166-171.

21. Horn, R.; Gupta, K.J.; Colombo, N. Mitochondrion role in molecular basis of cytoplasmic male sterility. Mitochondrion 2014, 19, 198-205.

22. Hu, J.; Huang, W.; Huang, Q.; Qin, X.; Yu, C.; Wang, L.; Li, S.; Zhu, R.; Zhu, Y. Mitochondria and cytoplasmic male sterility in plants. Mitochondrion 2014, 19, 282-288.

23. Yamagishi, H.; Bhat, S.R. Cytoplasmic male sterility in Brassicaceae crops. Breed. Sci. 2014, 64, 38-47.

24. Bohra, A.; Jha, U.C.; Adhimoolam, P.; Bisht, D.; Singh, N.P. Cytoplasmic male sterility (CMS) in hybrid breeding in field crops. Plant Cell Rep. 2016, 35, 967-993.

25. Gualberto, J.M.; Newton, K.J. Plant mitochondrial genomes: Dynamics and mechanisms of mutation. Annu. Rev. Plant Biol. 2017, 68, 225-252.

26. Tang, H.; Xie, Y.; Liu, Y.G.; Chen, L. Advances in understanding the molecular mechanisms of cytoplasmic male sterility and restoration in rice. Plant Reprod. 2017, 30, 179-184.

27. Štorchová, H. The role of non-coding RNAs in cytoplasmic male sterility in flowering plants. Int. J. Mol. Sci. 2017, 18, 2429.

28. Kim, Y.J.; Zhang, D. Molecular control of male fertility for crop hybrid breeding. Trends Plant Sci. 2018, 23, 53-65.

29. Chen, Z.; Zhao, S.; Li, S.; Grover, C.E.; Nie, H.; Wendel, J.F.; Hua, J. Plant mitochondrial genome evolution and cytoplasmic male sterility. Crit. Rev. Plant Sci. 2017, 36, 55-69.

30. Mishra, A.; Bohra, A. Non-coding RNAs and plant male sterility: Current knowledge and future prospects. Plant Cell Rep. 2018, 37, 177-191.

31. Fishman, L.; Sweigart, A.L. When two rights make a wrong: The evolutionary genetics of plant hybrid incompatibilities. Annu. Rev. Plant Biol. 2018, 69, 707-731.

32. Singh, S.; Dey, S.S.; Bhatia, R.; Kumar, R.; Behera, T.K. Current understanding of male sterility systems in vegetable Brassicas and their exploitation in hybrid breeding. Plant Reprod. 2019, 32, 231-256.

33. Rice, W.R. Nothing in genetics makes sense except in light of genomic conflict. Annu. Rev. Ecol. Evol. Syst. 2013, 44, 217-237.

34. Dahan, J.; Mireau, H. The Rf and Rf-like PPR in higher plants, a fast-evolving subclass of PPR genes. RNA Biol. 2013, 10, 1469-1476.

35. Gaborieau, L.; Brown, G.G.; Mireau, H. The propensity of pentatricopeptide repeat genes to evolve into restorers of cytoplasmic male sterility. Front. Plant Sci. 2016, 7, 1816.

36. Klein, R.R.; Klein, P.E.; Mullet, J.E.; Minx, P.; Rooney, W.L.; Schertz, K.F. Fertility restorer locus Rf1 of sorghum (Sorghum bicolor L.) encodes a pentatricopeptide repeat protein not present in the colinear region of rice chromosome 12. Theor. Appl. Genet. 2005, 111, 994-1012.

37. Rizzolatti, C.; Bury, P.; Tatara, E.; Pin, P.A.; Rodde, N.; Bergès, H.; Budar, F.; Mireau, H.; Gielen, J.J.L. Map-based cloning of the fertility restoration locus Rfm1 in cultivated barley (Hordeum vulgare). Euphytica 2017, 213, 276.

38. Itabashi, E.; Iwata, N.; Fujii, S.; Kazama, T.; Toriyama, K. The fertility restorer gene, Rf2, for Lead Rice-type cytoplasmic male sterility of rice encodes a mitochondrial glycine-rich protein. Plant J. 2010, 65, 359-367.

39. Bernhard, T.; Koch, M.; Snowdon, R.J.; Friedt, W.; Wittkop, B. Undesired fertility restoration in msm1 barley associates with two mTERF genes. Theor. Appl. Genet. 2019, 132, 1335-1350.

40. Hackauf, B.; Bauer, E.; Korzun, V.; Miedaner, T. Fine mapping of the restorer gene Rfp3 from an Iranian primitive rye (Secale cereale L.). Theor. Appl. Genet. 2017, 130, 1179-1189.

41. Matsuhira, H.; Kagami, H.; Kurata, M.; Kitazaki, K.; Matsunaga, M.; Hamaguchi, Y.; Hagihara, E.; Ueda, M.; Harada, M.; Muramatsu, A.; et al. Unusual and typical features of a novel restorer-of-fertility gene of sugar beet (Beta vulgaris L.). Genetics 2012, 192, 1347-1358.

42. Liu, F.; Cui, X.; Horner, H.T.; Weiner, H.; Schnable, P.S. Mitochondrial aldehyde dehydrogenase activity is required for male fertility in maize. Plant Cell 2001, 13, 1063-1078.

43. Jaqueth, J.S.; Hou, Z.; Zheng, P.; Ren, R.; Nagel, B.A.; Cutter, G.; Niu, X.; Vollbrecht, E.; Greene, T.W.; Kumpatla, S.P. Fertility restoration of maize CMS-C altered by a single amino acid substitution within the Rf4 bHLH transcription factor. Plant J. 2020, 101, 101-111.

44. Fujii, S.; Toriyama, K. Suppressed expression of Retrograde-Regulated Male Sterility restores pollen fertility in cytoplasmic male sterile rice plants. Proc. Natl. Acad. Sci. USA 2009, 106, 9513-9518.

45. Barkan, A.; Small, I. Pentatricopeptide repeat proteins in plants. Annu. Rev. Plant Biol. 2014, 65, 415-442.

46. Hu, J.; Wang, K.; Huang, W.; Liu, G.; Gao, Y.; Wang, J.; Huang, Q.; Ji, Y.; Qin, X.; Wan, L.; et al. The rice pentatricopeptide repeat protein RF5 restores fertility in Hong-Lian cytoplasmic male-sterile lines via a complex with the glycine-rich protein GRP162. Plant Cell 2012, 24, 109-122.

47. Small, I.D.; Schallenberg-Rüdinger, M.; Takenaka, M.; Mireau, H.; Ostersetzer-Biran, O. Plant organellar RNA editing: What 30 years of research has revealed. Plant J. 2019.

48. Welchen, E.; García, L.; Mansilla, N.; Gonzalez, D.H. Coordination of plant mitochondrial biogenesis: Keeping pace with cellular requirements. Front. Plant Sci. 2014, 4, 551.

49. Liu, F.; Schnable, P.S. Functional specialization of maize mitochondrial aldehyde dehydrogenase. Plant Physiol. 2002, 130, 1657-1674.

50. Allen, J.O.; Fauron, C.M.; Minx, P.; Roark, L.; Oddiraju, S.; Lin, G.N.; Meyer, L.; Sun, H.; Kim, K.; Wang, C.; et al. Comparisons among two fertile and three male-sterile mitochondrial genomes of maize. Genetics 2007, 177, 1173-1192.

51. Nan, G.L.; Zhai, J.; Arikit, S.; Morrow, D.; Fernandes, J.; Mai, L.; Nguyen, N.; Meyers, B.C.; Walbot, V. MS23, a master basic helix-loop-helix factor, regulates the specification and development of the tapetum in maize. Development 2017, 144, 163-172.

52. Fujii, S.; Kazama, T.; Yamada, M.; Toriyama, K. Discovery of global genomic re-organization based on comparison of two newly sequenced rice mitochondrial genomes with cytoplasmic male sterility-related genes. BMC Genom. 2010, 11, 209.

53. Kazama, T.; Itabashi, E.; Fujii, S.; Nakamura, T.; Toriyama, K. Mitochondrial ORF79 levels determine pollen abortion in cytoplasmic male sterile rice. Plant J. 2016, 85, 707-716.

54. Quesada, V. The roles of mitochondrial transcription termination factors (MTERFs) in plants. Physiol. Plant. 2016, 157, 389-399.

55. Linder, T.; Park, C.B.; Asin-Cayuela, J.; Pellegrini, M.; Larsson, N.G.; Falkenberg, M.; Samuelsson, T.; Gustafsson, C.M. A family of putative transcription termination factors shared amongst metazoans and plants. Curr. Genet. 2005, 48, 265-269.

56. Arakawa, T.; Sugaya, H.; Katsuyama, T.; Honma, Y.; Matsui, K.; Matsuhira, H.; Kuroda, Y.; Kitazaki, K.; Kubo, T. How did a duplicated gene copy evolve into a restorer-of-fertility gene in a plant? The case of Oma1. R. Soc. Open Sci. 2019, 6, 190853.

57. Käser, M.; Kambacheld, M.; Kisters-Woike, B.; Langer, T. Oma1, a novel membrane-bound metallopeptidase in mitochondria with activities overlapping with the m-AAA Protease. J. Biol. Chem. 2003, 278, 46414-46423.

58. Head, B.; Griparic, L.; Amiri, M.; Gandre-Babbe, S.; van der Bliek, A.M. Inducible proteolytic inactivation of OPA1 mediated by the OMA1 protease in mammalian cells. J. Cell Biol. 2009, 187, 959-966.

59. Ehses, S.; Raschke, I.; Mancuso, G.; Bemacchia, A.; Geimer, S.; Tondera, D.; Martinou, J.C.; Westermann, B.; Rugarli, E.I.; Langer, T. Regulation of OPA1 processing and mitochondrial fusion by m-AAA protease isoenzymes and OMA1. J. Cell Biol. 2009, 187, 1023-1036.

60. Kitazaki, K.; Arakawa, T.; Matsunaga, M.; Yui-Kurino, R.; Matsuhira, H.; Mikami, T.; Kubo, T. Post-translational mechanisms are associated with fertility restoration of cytoplasmic male sterility in sugar beet (Beta vulgaris). Plant J. 2015, 83, 290-299.

61. Kubo, T.; Nishizawa, S.; Mikami, T. Alterations in organization and transcription of the mitochondrial genome of cytoplasmic male sterile sugar beet (Beta vulgaris L.). Mol. Gen. Genet. 1999, 262, 283-290.

62. Skippington, E.; Barkman, T.J.; Rice, D.W.; Palmer, J.D. Miniaturized mitogenome of the parasitic plant Viscum scurruloideum is extremely divergent and dynamic and has lost all nad genes. Proc. Natl. Acad. Sci. USA 2015, 112, E3515-E3524.

63. Fujii, S.; Bond, C.S.; Small, I.D. Selection patterns on restorer-like genes reveal a conflict between nuclear and mitochondrial genomes throughout angiosperm evolution. Proc. Natl. Acad. Sci. USA 2011, 108, 1723-1728.

64. Kato, H.; Tezuka, K.; Feng, Y.Y.; Kawamoto, T.; Takahashi, H.; Mori, K.; Akagi, H. Structural diversity and evolution of the Rf-1 locus in the genus Oryza. Heredity 2007, 99, 516-524.

65. Melonek, J.; Stone, J.D.; Small, I. Evolutionary plasticity of restorer-of-fertility-like proteins in rice. Sci. Rep. 2016, 6, 35152.

66. Sykes, T.; Yates, S.; Nagy, I.; Asp, T.; Small, I.; Studer, B. In Silico Identification of candidate genes for fertility restoration in cytoplasmic male sterile perennial ryegrass (Lolium perenne L.). Genome Biol. Evol. 2017, 9, 351-362.

67. Melonek, J.; Zhou, R.; Bayer, P.E.; Edwards, D.; Stein, N.; Small, I. High intraspecific diversity of Restorer-of-fertility-like genes in barley. Plant J. 2019, 97, 281-295.

68. Anisimova, I.N.; Alpatieva, N.V.; Karabitsina, Y.I.; Gavrilenko, T.A. Nucleotide sequence polymorphism in the RFL-PPR genes of potato. J. Genet. 2019, 98, 87.

69. Geddy, R.; Brown, G.G. Genes encoding pentatricopeptide repeat (PPR) proteins are not conserved in location in plant genomes and may be subject to diversifying selection. BMC Genom. 2007, 8, 130.

70. Marienfeld, J.R.; Unseld, M.; Brandt, P.; Brennicke, A. Mosaic open reading frames in the Arabidopsis thaliana mitochondrial genome. Biol. Chem. 1997, 378, 859-862.

71. Hammani, K.; Giegé, P. RNA metabolism in plant mitochondria. Trends Plant Sci. 2014, 19, 380-389.

72. Arakawa, T.; Ue, S.; Sano, C.; Matsunaga, M.; Kagami, H.; Yoshida, Y.; Kuroda, Y.; Taguchi, K.; Kitazaki, K.; Kubo, T. Identification and characterization of a semi-dominant restorer-of-fertility 1 allele in sugar beet (Beta vulgaris). Theor. Appl. Genet. 2019, 132, 227-240.

73. Arakawa, T.; Uchiyama, D.; Ohgami, T.; Ohgami, R.; Murata, T.; Honma, Y.; Hamada, H.; Kuroda, Y.; Taguchi, K.; Kitazaki, K.; et al. A fertility-restoring genotype of beet (Beta vulgaris L.) is composed of a weak restorer-of-fertility gene and a modifier gene tightly linked to the Rf1 locus. PLoS ONE 2018, 13, e0198409.

74. Kanomata, Y.; Kubo, T. Unpublished data; Hokkaido University: Sapporo, Japan, 2020.

75. Migdal, I.; Skibior-Blaszczyk, R.; Heidorn-Czarna, M.; Kolodziejczak, M.; Garbiec, A.; Janska, H. AtOMA1 affects the OXPHOS system and plant growth in contrast to other newly identified ATP-independent proteases in Arabidopsis mitochondria. Front. Plant Sci. 2017, 8, 1543.

76. Arakawa, T.; Kubo, T. Manuscript in preparation; Hokkaido University: Sapporo, Japan, 2020.

77. Force, A.; Lynch, M.; Pickett, F.B.; Amores, A.; Yan, Y.-L.; Postlethwait, J. Preservation of duplicated genes by complementary, degenerative mutations. Genetics 1999, 151, 1531-1545.

78. Ohgami, T.; Uchiyama, D.; Ue, S.; Yui-Kurino, R.; Yoshida, Y.; Kamei, Y.; Kuroda, Y.; Taguchi, K.; Kubo, T. Identification of molecular variants of the nonrestoring restorer-of-fertility 1 allele in sugar beet (Beta vulgaris L.). Theor. Appl. Genet. 2016, 129, 675-688.

79. Touzet, P.; Budar, F. Unveiling the molecular arms race between two conflicting genomes in cytoplasmic male sterility? Trends Plant Sci. 2004, 9, 568-570.

80. Touzet, P. Mitochondrial genome evolution and gynodioecy. In Mitochondrial Genome Evolution; Marechal-Drouard, L., Ed.; Academic Press: Oxford, UK, 2012; pp. 71-98.

81. Tiffin, P.; Moeller, D.A. Molecular evolution of plant immune system genes. Trends Genet. 2006, 22, 662-670.

82. Kazama, T.; Nakamura, T.; Watanabe, M.; Sugita, M.; Toriyama, K. Suppression mechanism of mitochondrial ORF79 accumulation by Rf1 protein in BT-type cytoplasmic male sterile rice. Plant J. 2008, 55, 619-628.

83. Uyttewaal, M.; Arnal, N.; Quadrado, M.; Martin-Canadell, A.; Vrielynck, N.; Hiard, S.; Gherbi, H.; Bendahmane, A.; Budar, F.; Mireau, H. Characterization of Raphanus sativus pentatricopeptide repeat proteins encoded by the fertility restorer locus for Ogura cytoplasmic male sterility. Plant Cell 2008, 20, 3331-3345.

84. Huang, W.; Yu, C.; Hu, J.; Wang, L.; Dan, Z.; Zhou, W.; He, C.; Zeng, Y.; Yao, G.; Qi, J.; et al. Pentatricopeptide-repeat family protein RF6 functions with hexokinase 6 to rescue rice cytoplasmic male sterility. Proc. Natl. Acad. Sci. USA 2015, 112, 14984-14989.

85. Møller, I.M. A more general mechanism of cytoplasmic male sterility? Trends Plant Sci. 2001, 6, 560.