2020 Volume 70 Issue 5 Pages 637-641
In addition to Ogura cytoplasmic male sterility (CMS), which is used extensively for F1 hybrid seed production in Brassicaceae crops, two other CMS systems, NWB CMS and DCGMS, have also been identified. The causal gene for the latter two CMS systems has been identified as a novel chimeric gene, orf463. We previously reported that orf463 is specific to black radish cultivars and that it is present in line ‘RS-5’ of Raphanus raphanistrum; however, the orf463 sequence in ‘RS-5’ differed from that of black radish cultivars. Though, R. raphanistrum with an orf463 sequence identical to that found in black radish cultivars was recently identified. We therefore sought to determine whether the orf463 gene in line ‘RS-5’ induces CMS in radishes. We crossed ‘RS-5’ as a female parent with a cultivated radish, ‘Uchiki-Gensuke’, as a male parent, and examined the gross plant morphology and pollen fertility of the resulting progeny. The F2 population contained both male sterile plants and plants with black roots. The findings showed that R. raphanistrum contains two types of orf463 genes that induce CMS, and that the origin of black radishes could be attributed to R. raphanistrum having orf463 gene.
Two mitochondrial genes causing cytoplasmic male sterility (CMS) have been identified in radish to date. The first gene is orf138, which causes Ogura CMS (Bonhomme et al. 1991). Ogura CMS was first discovered in Japanese radish (Ogura 1968), and the cytoplasm was introduced to Brassica crops. Given that the sterility associated with Ogura CMS is stable and Brassica crops do not have a fertility restorer gene (Rf gene), the Ogura CMS system was widely used in seed production for Brassica crops worldwide. It has been reported that orf138 is widely distributed in Japanese wild radishes, and wild species such as Raphanus raphanistrum (Yamagishi and Terachi 1996, 1997).
However, most of the Korean radish cultivars possess Rf gene(s) for Ogura CMS. Consequently, since Ogura CMS cannot be applied to practical F1 breeding in Korea, Korean scientists searched for other CMS systems. As a result, the NWB CMS (Nahm et al. 2005) and DCGMS (Lee et al. 2008) systems were discovered. Park et al. (2013) sequenced the mitochondrial genome that induced DCGMS, and identified a novel chimeric gene, orf463, as the causal gene; this is the second CMS gene. We also obtained the complete mitochondrial sequences of the ‘Black radish’ variety, and found that the genome contained orf463 (Yamagishi et al. 2019). Interestingly, while orf463 was concentrated uniquely in the black radish cultivars belonging to the ‘Niger’ group, one line of the wild species, Raphanus raphanistrum, also possessed orf463 (Yamagishi et al. 2019). Recently, Wang et al. (2020) demonstrated that all six varieties of Raphanus sativus var. niger (Black Spanish radish) possessed orf463, and one line each of R. raphanistrum and Raphanus maritimus also possessed orf463. Furthermore, Wang et al. (2020) demonstrated that NWB cytoplasm has a gene that is identical to orf463, indicating that the same gene causes NWB CMS and DCGMS.
The cytoplasm of the four cultivars of the ‘Niger’ group induced male sterility in their progeny (Yamagishi et al. 2019). The cultivars had sequences that were identical to the orf463 sequence determined by Park et al. (2013) (Yamagishi et al. 2019). Wang et al. (2020) also observed that cultivars of black Spanish radish and a line of R. raphanistrum have sequences that are identical orf463 (Wang et al. 2020). However, the orf463 found in R. raphanistrum (Yamagishi et al. 2019) and in R. maritimus (Wang et al. 2020) had nine nucleotide substitutions, of which seven were not synonymous.
We therefore sought to determine whether the orf463 that we found in R. raphanistrum also causes male sterility. We crossed R. raphanistrum bearing orf463 and the cultivar, ‘Uchiki-Gensuke’, that we used in our previous experiment (Yamagishi et al. 2019) to see if we could induce male sterility in F2 progeny. The results showed that male sterile plants were produced in the F2 populations. In addition, black radishes also appeared in the F2 populations. The findings indicate that the wild species, R. raphanistrum, contains the genetic resources required for producing ‘Niger group’ cultivars and for inducing CMS due to orf463.
Two F2 populations were produced by self-fertilizing two F1 plants obtained from a cross between ‘RS-5’ (R. raphanistrum) and ‘Uchiki-Gensuke’ (hereafter, ‘UC-G’). ‘RS-5’ was originally collected in Spain, and the seeds were stocked in the Gene Bank of Tohoku University. The seeds were kindly gifted to us by Tohoku University. Although the interspecific cross-fertility was much lower when R. raphanistrum was the female parent than the reciprocal cross (Yamagishi and Terachi 1997), we obtained the F1 hybrids. Both the two F1 plants showed normal seed fertility, producing 2.2 seeds per silique by the self-fertilization. One of the self-fertilized F1 plants, (RS-5-5 × UC-G)-3, had red-purple root, and another one, (R-5-5 × UC-G)-4, showed green root color. All eight F1 plants, including the two used for the self-fertilization, were male fertile, as described in our previous paper (Yamagishi et al. 2019). Among them, the two self-fertilized plants, (RS-5-5 × UC-G)-3 and (RS-5-5 × UC-G)-4, possessed the pollen fertility of 74.0% and 82.3%.
‘UC-G’ is a maintainer of Ogura CMS, and the progenies between ‘UC-G’ as a pollen parent and the ‘Niger’ group cultivars showed segregation of male sterile plants (Yamagishi et al. 2019). In addition, we cultivated the F1 plants produced by a cross between ‘RS-5’ and ‘UC-G’, and the reciprocal progeny of ‘UC-G’ × (‘RS-5’ × ‘UC-G’). An additional ten ‘RS-5’ plants were also newly cultivated and observed in this study (Table 1). All of the materials were cultivated in a glass house.
| Line or Progeny | Number of plants | ||||
|---|---|---|---|---|---|
| Total | White | Green | Red-purple | Black | |
| RS-5 | 10 | 0 | 7 | 3 | 0 |
| UC-G | 10 | 7 | 3 | 0 | 0 |
| F1 (RS-5-5 × UC-G) | 6 | 2 | 1 | 3 | 0 |
| F2 ((RS-5-5 × UC-G)-3) | 35 | 4 | 2 | 24 | 5 |
| F2 ((RS-5-5 × UC-G)-4) | 18 | 5 | 10 | 0 | 3 |
The plants were cultivated in pots. Approximately two months after sowing when vegetative growth was observed, the color of the root surface was recorded in the parts above the soil (hypocotyls). Root color was classified as ‘White’, ‘Green’, ‘Red-purple’, or ‘Black’.
Observation of pollen fertilityAfter each plant bloomed, we observed anther morphology and repeatedly touched the anthers with our fingers to check for pollen adhesion. As some of the progeny died before flowering, the number of plants in which pollen fertility could be observed was decreased. Plants were judged to be fertile if their anthers produced pollen, and male sterile if the pollen never adhered to the fingers. In addition, individual plants that were male sterile but which produced fertile flowers were classified as having unstable fertility. Microscopic observation of the pollen was conducted for the plants judged to be male sterile by staining with acetocarmine.
DNA sequencing of orf463 in R. raphanistrumBased on the finding that the ‘RS-5’ line of R. raphanistrum possessed orf463 (Yamagishi et al. 2019), we newly cultivated the line. DNA was isolated from the leaves of each plant using a DNeasy Mini kit (Qiagen, Valencia, CA, USA) and the isolated DNA was used as a template for PCR. The PCR and the sequencing of orf463 were performed as described in Yamagishi et al. (2019). We then compared the deduced protein sequences of ORF463 obtained from the sequences using the TMHMM server version 2.0 (Krogh et al. 2001).
The root color of the wild line, ‘RS-5’, was green or red-purple, and the color of ‘UC-G’ was white or green, as shown in Table 1 and Fig. 1. The F1 plants showed variation in root color, ranging from white to red-purple (Table 1). In comparison, the plants of the F2 generation varied from white to black (Table 1, Fig. 1A), and the radishes with black roots appeared in both of the F2 populations (Table 1, Fig. 1B). The dominant root color of the F2 populations differed depending on the F1 plants. In one population, F2 ((RS-5-5 × UC-G)-3), most of the plants had roots with a red-purple color, but individuals with green roots were dominant in another population, F2 ((RS-5-5 × UC-G)-4) (Table 1). The major color in each of the F2 populations corresponded to the F1 plant from which the F2 generation derived, respectively.
Appearance of black radishes in the F2 generation between ‘RS-5’ and ‘UC-G’. A: Color of hypocotyl surface, (1): ‘RS-5’, (2) ~ (5): F2 plants, (6): ‘UC-G’ (Bars indicate 3 cm). B: Black radish plants in the two F2 populations, F2 ((RS-5-5 × UC-G)-3) (left) and F2 ((RS-5-5 × UC-G)-4) (right) (Bars indicate 10 cm).
As in our previous study (Yamagishi et al. 2019), all of the F1 plants produced from a cross between ‘RS-5’ and ‘UC-G’ were male fertile (Table 2). The progeny between crosses of ‘UC-G’ as a female parent and one of the F1 plants were fertile (Table 2). On the other hand, although most of the plants in the F2 populations were fertile, male sterile plants were also observed (Table 2). The male sterile plants had short filaments compared to ‘RS-5’ and ‘UC-G’, and the anthers were rudimentary (Fig. 2A). The microscopic observations of the pollen revealed that the anthers of the sterile plants only produced deformed pollen, whereas the fertile plants produced normal pollen (Fig. 2B). The pollen fertility of the fertile plant shown in Fig. 2B was 85%, and the value was almost equal to that of ‘UC-G’ (88%) observed on the same day. The results showed that the fertility was restored to the level of the plants with normal cytoplasm. In addition to the fertile and sterile plants, the plants that showed unstable fertility were observed in both of the F2 populations (Table 2).
| Line or Progeny | Number of plants | |||
|---|---|---|---|---|
| Total | Fertile | Sterile | Unstable | |
| F1 (RS-5-5 × UC-G) | 5 | 5 | 0 | 0 |
| UC-G × ((RS-5-5 × UC-G)-3) | 18 | 18 | 0 | 0 |
| F2 ((RS-5-5 × UC-G)-3) | 21 | 16 | 2 | 3 |
| F2 ((RS-5-5 × UC-G)-4) | 14 | 12 | 1 | 1 |
| RS-5 | 10 | 10 | 0 | 0 |
| UC-G | 6 | 6 | 0 | 0 |
Stamen and pollen of the male sterile F2 plant between ‘RS-5’ and ‘UC-G’. A: Stamens of ‘RS-5’ (left), the male sterile F2 (center), and ‘UC-G’ (right) (Bar indicates 1 cm). B: Pollen grains stained with acetocarmine (left: sterile F2 plant, right: fertile F2 plant) (Bars indicate 100 μm).
Because the number of plants was not large in the two populations, the inheritance mode of pollen fertility could not be estimated. However, the results indicated that male sterility was caused by the cytoplasm of ‘RS-5’. The results also demonstrated that ‘RS-5’ possesses Rf gene(s) against orf463. Further, the facts that the two self-fertilized F1 plants had pollen fertility higher than 70%, and that the male sterile plants were segregated in the F2 populations suggested that the CMS was sporophytic type.
Sequence of orf463 in ‘RS-5’The DNA sequence of orf463 in the 10 ‘RS-5’ plants that were newly cultivated in this experiment was identical to that observed in our previous report (Yamagishi et al. 2019). Compared with the orf463 associated with DCGMS, the gene sequenced in this study contained nine nucleotide substitutions. The findings therefore indicated that all of the plants in ‘RS-5’ had this mitochondrial type. By the prediction of products of orf463, it was found that the two proteins had a similar structure and possessed 12 transmembrane domains as shown by Park et al. (2013) (Fig. 3).
Locations and probability of transmembrane domains of ORF463. A: RS-5, B: DCGMS, Arrows: Positions of amino acid substitutions.
The CMS gene, orf463, was first found in a collection from Uzbekistan (Lee et al. 2008). The observation that, other than male sterility, the morphology of the collection was not significantly different from that of normal radish (Lee et al. 2008) suggests that the accession belongs to cultivated radish. We previously found orf463 specifically in black radishes (Yamagishi et al. 2019), and Wang et al. (2020) corroborated this finding recently. They observed that all six accessions of black Spanish radish possessed the orf463 gene (Wang et al. 2020). In addition to the cultivars mentioned above, orf463 has also been identified in R. raphanistrum (Wang et al. 2020, Yamagishi et al. 2019) and R. maritimus (Wang et al. 2020).
The cytoplasm of the black radishes was also shown to induce male sterility (Yamagishi et al. 2019). In this study, the cytoplasm of ‘RS-5’ was shown to cause male sterility in a similar manner to DCGMS as demonstrated by Lee et al. (2008). These observations indicate that the orf463 of black radish was derived from the wild species, R. raphanistrum. Although the DNA sequence of ‘RS-5’ differs from that of black radish, Wang et al. (2020) found a wild radish with an orf463 sequence that was identical to black radish.
Black roots were observed in the F2 generation produced from a cross between ‘RS-5’ and ‘UC-G’, even though the both parents did not have black roots (Table 1). As reported in Nishio (2017), red and purple roots have been studied in radish cultivars; however, this is the first report to demonstrate the segregation of the black root color in progeny resulting from cross hybridization. This result also indicates the origin of black radish. Further, haplotype analysis using the chloroplast genome showed that ‘RS-5’ is closely related to the black radish variety (Yamagishi et al. 2009). Thus, based on the sharing of orf463, the appearance of black roots in progeny, and the similarities in the chloroplast genome, the results strongly suggest that the black radish originated from a wild species with orf463 in their mitochondria.
Sequence variation of orf463The DNA sequence of orf463 in ‘RS-5’ was fixed to one type that had nine substitutions compared to DCGMS. Despite these differences, both types of orf463 caused CMS. Although only two types have been found in orf463, nine types of orf138 are known in the Ogura CMS gene (Yamagishi and Terachi 2001). In the case of orf138, the average number of DNA variations among the types was 2.28, and average number of non-synonymous nucleotide substitutions was 1.56. When we compare the size of the coding sequences of orf138 and orf463, the number of nucleotide substitutions is comparable between the two CMS genes. Further, as shown in Fig. 3, the deduced ORF463 proteins exhibited similar characteristics despite the non-synonymous substitutions. Thus, the products of both orf463 sequences share the same function of inducing male sterility. A more extensive search would clarify the sequence variations that exist in orf463 in wild species, like orf138.
Rf genes for orf463Lee et al. (2008) reported finding DCGMS, and that the inheritance pattern associated with this male sterility system varied according to parental lines. Specifically, they found that in one case a single locus was involved, and in another cross combination at least three genes were involved. Wang et al. (2020) also examined three crosses between a male sterile line and radish lines having Rf gene(s). One F2 population exhibited the segregation of 15 fertile: 1 sterile plants, fitting an inheritance mode controlled by two loci. However, in the other two F2 populations, no male sterile plants were observed in approximately 100 F2 plants. In that case, at least three genes were involved. Compared to these two studies, we found the male sterile plants in the F2 populations of smaller size, but it was difficult to estimate the inheritance mode.
In addition, we found the plants in which male fertility was unstable, the fertile or sterile phenotype differing between the flowers in an individual. This phenomenon was also observed when black radishes were used as the female parent (Yamagishi et al. 2019), which made it difficult to estimate the number of Rf genes. If the observed instability is a characteristic of the male sterility induced by orf463, then it would affect the practical usefulness in breeding. However, such instability was not reported in two previous articles (Lee et al. 2008, Wang et al. 2020). To better understand fertility restoration, cloning of the Rf genes is needed, and the DNA markers developed by Cho et al. (2012) would be useful for such a purpose.
Conceptualization, H.Y. and T.T.; Methodology, H.Y.; Investigation, H.Y., A.H. and A.F.; Writing, H.Y. and T.T.; Supervision, H.Y.
