The two genomes (A and C) of the allopolyploid Brassica napus have been clearly distinguished using genomic in situ hybridization (GISH) despite the fact that the two extant diploids, B. rapa (A, n = 10) and B. oleracea (C, n = 9), representing the progenitor genomes, are closely related. Using DNA from B. oleracea as the probe, with B. rapa DNA and the intergenic spacer of the B. oleracea 45S rDNA as the block, hybridization occurred on 9 of the 19 chromosome pairs along the majority of their length. The pattern of hybridization confirms that the two genomes have remained distinct in B. napus line DH12075, with no significant genome homogenization and no large-scale translocations between the genomes. Fluorescence in situ hybridization (FISH)—with 45S rDNA and a BAC that hybridizes to the pericentromeric heterochromatin of several chromosomes—followed by GISH allowed identification of six chromosomes and also three chromosome groups. Our procedure was used on the B. napus cultivar Westar, which has an interstitial reciprocal translocation. Two translocated segments were detected in pollen mother cells at the pachytene stage of meiosis. Using B. oleracea chromosome-specific BACs as FISH probes followed by GISH, the chromosomes involved were confirmed to be A7 and C6.
IT is estimated that polyploidy has occurred in up to 70% of angiosperm species and it is widely accepted to have been a major force in the evolution of plant genomes (reviewed by Wendel 2000). Many agricultural crops are polyploids, including Brassica napus (oilseed rape, Canola, swede).
The relationship between six agriculturally important Brassica species, three diploid and three allopolyploid, was demonstrated in a classical cytogenetic study by U (1935); this relationship is commonly referred to as the “U triangle.” That study showed that B. oleracea (genome CC, 2n = 18) has produced allopolyploids both with B. rapa (AA, 2n = 20) to form B. napus (AACC, 2n = 38) and with B. nigra (BB, 2n = 16) to form B. carinata (BBCC, 2n = 34). The third amphidiploid, B. juncea (AABB, 2n=36), was formed from B. rapa and B. nigra. Comparisons among the genomes of B. rapa, B. oleracea, synthetic B. napus, and natural B. napus through molecular marker analyses of various crosses concluded that the A and C genomes of natural B. napus have remained essentially unaltered since the formation of the species and that they are similar to those of modern-day B. rapa and B. oleracea (Parkin and Lydiate 1997; Udall et al. 2005). However, it is known that B. napus cultivars differ from each other with regard to the presence of homeologous reciprocal or nonreciprocal translocations caused by recombination between homeologous regions of the A and C genomes (Parkin et al. 1995; Sharpe et al. 1995; Osborn et al. 2003; Udall et al. 2005).
Research into the genome constitution of allopolyploids, the pairing of homeologous chromosomes, and the identification of translocations between chromosomes of different genomes has been aided by recent advances in molecular cytogenetics, particularly fluorescence in situ hybridization (FISH) and a related technique, genomic in situ hybridization (GISH) (reviewed by Raina and Rani 2001). With GISH, the total genomic DNA from one parental species is labeled and used as a probe. Usually, an excess of unlabeled DNA (blocking DNA) from the other parent is included to increase specificity. Because sequences common to both species are blocked, labeled DNA will hybridize only to genome-specific sequences on the chromosomes of its own genome.
Identification of the diploid genomes in the allopolyploid brassicas, using GISH, has been successful for those containing the B genome. Snowdon et al. (1997) demonstrated that in B. juncea and B. carinata the diploid genomes could be distinguished. Maluszynska and Hasterok (2005) combined a GISH analysis of B. juncea with FISH using labeled 5S rDNA and 45S rDNA probes to discriminate several chromosomes within each genome. GISH was also used in the analysis of the behavior of B genome chromosomes in the trigenomic hybrids ABC (Ge and Li 2007). However, an attempt to clearly discriminate the A genome from the C genome in B. napus by GISH was unsuccessful (Snowdon et al. 1997).
We present a strategy to distinguish between the A and C genomes in B. napus and to identify individual chromosomes within a genome. It is based on a cytological technique using material from anthers, which provide many meiotic cells as well as mitotic cells from the tapetal layer. We demonstrate that our modified GISH technique, which includes a repetitive probe in the block, is effective on both mitotic and meiotic cells. We have used it predominantly on meiotic pachytene chromosomes and on chromosomes at late diakinesis. Further, we have developed a sequential procedure with FISH and GISH that allows us to identify several specific chromosomes. Using a B. napus cultivar carrying a known reciprocal translocation, we also demonstrate that it is possible to visualize intergenomic translocations and to identify the chromosomes involved by the sequential use of FISH with BACs from B. oleracea, which are linked to a genetic map (Howell et al. 2005) and GISH.
MATERIALS AND METHODS
Anthers with pollen mother cells at meiotic stages between prophase I and metaphase I were collected from B. napus doubled haploid line DH12075 (provided by D. Lydiate, AAFC, Saskatoon, Saskatchewan, Canada) and from B. napus cv. Westar and N-o-1, a DH line derived from Westar (provided by John Innes Centre, Norwich, UK). Methods for staging and fixing anthers have been described previously (Armstrong et al. 1998; Howell et al. 2002). Slides were prepared from these anthers following enzyme digestion (Armstrong et al. 1998; Howell et al. 2002) but cytohelicase was omitted from the enzyme mixture and the digestion time was extended to 2 hr. Slides were screened for well-spread chromosomes that were free of cytoplasm.
Probes and blocking DNA:
For GISH, total genomic DNA was isolated from young leaf tissue of B. oleracea var. alboglabra doubled haploid line A12DHd (C genome) and B. rapa cultivar Golden Ball (A genome) using a DNA extraction kit (Tepnel Life Sciences). After treatment with RNase and quantification against λ-phage DNA standards (Sigma) by gel electrophoresis, some of the DNA was mechanically sheared and then labeled with biotin-16-dUTP by nick translation (Roche). The remainder was autoclaved for 5 min twice to obtain DNA fragments of 100–500 bp for use as blocking DNA. The intergenic spacer (IGS) region of B. oleracea 45S rDNA was amplified from A12DHd genomic DNA by PCR using primers based on EMBL X60324. The primers were Fwd:CAGCCCTTTGTCGCTAAG (within 25S gene) and Rev:GGCAGGATCAACCAGGTA (within 18S gene) and the annealing temperature was 61° (Howell et al. 2002). The PCR product was cleaned and concentrated through a PCR purification kit (Qiagen) and used for blocking.
The probes used for FISH were (1) 45S rDNA from clone pTa71 (Gerlach and Bedbrook 1979), EMBL X07841, labeled with digoxigenin-11-dUTP; (2) BoB061G14, a BAC clone that hybridizes to the pericentromeric heterochromatin of six B. oleracea chromosomes, labeled with biotin-16-dUTP; and (3) BoB028L01, BoB057M06, and BoB028I05, three BAC clones assigned to B. oleracea chromosome C6 (BolC6) (Howell et al. 2002, 2005). The first two BACs were labeled with digoxigenin-11-dUTP and the third was labeled with biotin-16-dUTP. All probes were labeled by nick translation (Roche).
In situ hybridization:
Slide pretreatment, chromosome and probe denaturation, and hybridization and post-hybridization treatments were the same for GISH and FISH and followed previous methods (Howell et al. 2002) except that denaturation at 75° was reduced to 3 min 30 sec. Probes were applied to a slide in 20 μl of a probe mixture that contained 50% deionized formamide, 2× SSC, 10% dextran sulfate, labeled probes, and blocking DNA if required.
For GISH, 50–100 ng labeled C genome DNA and ∼500 ng or 1100 ng blocking A genome DNA were used, with or without 100 ng IGS DNA. On two slides, labeled A genome DNA (50 ng) was used with 5000 ng blocking C genome DNA.
For FISH, ∼20–50 ng of each labeled probe was included in the probe mixture. Either both 45S rDNA and BoB061G14 were used or the three chromosome BolC6 BACs were used simultaneously. Approximately 1 μg C0t -1 DNA, prepared from A12DHd, was added with the BolC6 BACs to block repetitive sequences.
Biotin-labeled DNA and dioxigenin-labeled DNA were detected with Cy3 streptavidin (Cambio) and antidigoxigenin-fluorescein (Roche), respectively. Slides were counterstained with 4′,6-diamidino-2-phenylindole (DAPI; 1 μg/ml) in Vectashield (Vector).
After image capture, FISH slides selected for sequential probing with GISH were soaked in 2× SSC for 10 min to remove the coverslip, dehydrated through an ethanol series (70 and 85% for 30 sec and 100% for 1 min) and air dried. A GISH probe mixture was applied and the chromosomes and probes were denatured together at 75° for 3 min. The procedures following this step were the same as those for the first probing.
Images were captured and analyzed using SmartCapture 2 software (Digital Scientific UK) and a Photometrics Sensys CCD camera attached to a Nikon E600 fluorescence microscope. When slides were probed sequentially, recapture of the same chromosome spreads was facilitated by an automated stage. In all images, the hybridization sites of Cy 3 streptavidin-detected probes are red, the sites of antidigoxigenin–fluorescein-detected probes are green, and DAPI staining is blue.
When labeled total genomic DNA from B. oleracea (C genome) was hybridized to mitotic or meiotic chromosomes from anthers of B. napus line DH 12075 in the presence of blocking DNA from B. rapa (A genome), nine pairs of chromosomes were strongly labeled. This is expected if the chromosomes of B. oleracea (n = 9) are homologous to the C genome chromosomes of B. napus (n = 19). With the higher concentration of blocking DNA (1100 ng/slide) the differentiation between the genomes was clear. With the lower concentration of blocking DNA (500 ng/slide), the nine C chromosome pairs could still be distinguished but the differentiation was reduced due to increased labeling of the pericentromeric regions of the other chromosomes (not shown). Repeatability was confirmed using DNA samples that had been extracted, labeled, and autoclaved on a different occasion.
Even with the higher concentration of blocking DNA, the signal strength was not uniform along the chromosomes. In general, the intensity decreased from the pericentromeric regions toward the telomeres but very strong signals occurred at one end of each chromosome of two labeled pairs. It was likely that these signals were from C genome 45S rDNA sites and therefore an unlabeled PCR product of the IGS region of B. oleracea 45S rDNA was added to the probe mixture. This resulted in a reduction in the intensity of the strongest signals and therefore produced a more even distribution of the hybridization signal intensity in both mitotic (Figure 1A) and meiotic (Figure 1, B and C) preparations. On pachytene spreads, the coalesced centromeric regions (synizetic knot) fluoresced brightly and the hybridization signals became fainter and less frequent toward the telomeres, with the telomeric and subtelomeric regions being unlabeled. Nevertheless, with this combination of probe, IGS, and A genome block, the free chromosome arms could be categorized as labeled or unlabeled (not shown).
In contrast, when A genome DNA was used as the labeled probe with C genome DNA as the block (ratio of 1 probe to 100 block) on preparations at metaphase I, hybridization was restricted to limited regions on 10 bivalents. The signals differed in strength and the bivalents with the weakest signals were not clearly differentiated from some C genome bivalents, which fluoresced slightly at pericentromeric regions (not shown). This combination was not investigated further.
FISH with repeat probes followed by GISH:
We then used FISH (Figure 1D) with 45S rDNA and BoB061G14, which hybridizes to pericentromeric regions of many chromosomes, followed by GISH (Figure 1E) on preparations of DH12075 at diakinesis. The distribution of the signals from these probes among the chromosomes of the A and C genomes is shown in Table 1. The signals differed in intensity; for example, the strongest 45S rDNA signal was located distally on an A genome bivalent that lacked BoB061G14 signals and the weakest BoB061G14 signal was on an A genome bivalent. Although the chromosome preparations deteriorated slightly with this sequential probing, the two genomes were clearly differentiated. Without the addition of unlabeled IGS DNA, strong GISH signals were seen at the 45S rDNA sites on two C genome bivalents (Figure 1, D and E). In contrast, none of the A genome 45S rDNA sites showed labeling with GISH. Associations between 45S rDNA sites were seen frequently within and between genomes at diakinesis.
GISH on Westar:
Genetic linkage mapping with RFLPs established that a reciprocal translocation involving the lower portions of linkage groups A7 and C6 is present in several annual oilseed B. napus genotypes, including the cultivar Westar (Osborn et al. 2003). The translocation appears to be interstitial on the basis of segregation data for one or two markers at the end of these linkage groups in different populations and the presence of two synapsis exchange points in synaptonemal complex spreads of an F1 hybrid between DH line N-o-1, which was derived from Westar, and a line not carrying the translocation (Sharpe et al. 1995; Osborn et al. 2003).
When preparations of Westar undergoing mitosis or diakinesis (Figure 1F) were subjected to GISH, we identified nine labeled pairs, nine unlabeled pairs, and one pair with labeling confined to one end. We conclude that one A genome pair had a translocation from a C genome pair. Although a reciprocal translocation could not be identified with certainty at these stages, on good pachytene spreads, where long lengths of bivalents were free from the synizetic knot, one predominantly unlabeled bivalent had a long, labeled, distal section and, conversely, a labeled bivalent had an unlabeled section of similar size at the end (not shown). This unlabeled section showed no hybridization signals close to the telomere. Hybridization signals on the translocated labeled section decreased toward the telomere, with the telomeric and subtelomeric regions being unlabeled, similar to other labeled bivalents. Although our technique was sensitive enough to detect these translocations, we wanted to confirm that they were the result of a reciprocal translocation and that they represented the reciprocal translocation identified previously with molecular markers.
FISH with B. oleracea BACs followed by GISH on DH12075, Westar, and DH line N-o-1:
First we used three B. oleracea BACs, previously assigned near the bottom of chromosome BolC6 (Howell et al. 2005), as FISH probes to determine whether we could identify the corresponding region in the C genome of B. napus. Second, we investigated whether we could establish that the translocations visualized by GISH in Westar involved the regions identified by linkage mapping, since BolC6 corresponds to linkage group B. napus C6 (BnaC6) (Parkin et al. 1995).
The BAC probes, tested in pairs on pachytene spreads of DH12075, hybridized to two bivalents in positions similar to those on chromosome BolC6 (Howell et al. 2005), with BoB028I05 (red) nearest the telomere, BoB028L01 (green) nearby, and BoB057M06 (green) farther away from the telomere. Signals were often stronger on one bivalent. Subsequently, FISH, with all three BACs applied simultaneously (Figure 2A), was followed by GISH (Figure 2C). One hybridization site of BoB057M06 (green) is obscured in Figure 2A but was seen in other images. The bivalent with the strongest BAC signals was labeled by GISH whereas the second bivalent was not. This indicates that the BACs preferentially hybridized with the B. napus C genome and that there was sufficient sequence homology between the BACs and the homeologous region in the A genome for some hybridization to occur there as well.
On Westar, the three BACs gave the same pattern of hybridization signals as on DH12075 (Figure 2B). However, on reprobing with GISH and comparing the sequential images, a distinct change from labeled to unlabeled chromatin could be seen between the two green signals (Figure 2, B, D, and F). Therefore, on one bivalent, the BoB028I05 and BoB028L01 sites were on a GISH-labeled section while the BoB057M06 site was on an unlabeled section whereas on the other bivalent the reverse situation occurred. The sequential probing of DH12075 and Westar is shown diagrammatically in Figure 2, E and F, respectively. This confirmed that the translocation observed in Westar is reciprocal and involves regions of A and C genome chromosomes that have homology to part of BolC6.
DH line N-o-1, probed sequentially with the BACs and GISH, gave results that were indistinguishable from those of Westar. The images of whole bivalents (Figure 3, A–H) illustrate the proportion of the chromosomes involved in this reciprocal translocation. Although every BAC hybridization signal is not present on these images, with BoB057M06 being the least reliable, the sites were observed on other pachytene spreads. The occasional presence of a faint green site near the BoB028L01 site (Figure 3, C and D) suggests that this BAC hybridizes weakly to a duplicate region that was identified in BolC6 (Howell et al. 2005) and likely to be present in BnaC6. Some cross-hybridization to the A genome pericentromeric heterochromatin can be seen. The translocated section of a C genome chromosome on the A genome pair seen at diakinesis could also be identified as part of BnaC6 by using these BACs prior to GISH on spreads at diakinesis (Figure 3, I and J).
The diploid species B. rapa (A genome), B. nigra (B genome), and B. oleracea (C genome), which represent the progenitors of the three allotetraploids of U's triangle (U 1935), are closely related. Molecular phylogenetic analyses of the tribe Brassiceae within the family Brassicaceae place B. nigra in a different lineage to B. oleracea and B. rapa (Warwick and Black 1991). These lineages split ∼7.9 million years ago (MYA) (Lysak et al. 2005) whereas B. oleracea and B. rapa diverged more recently, ∼3.75 MYA (Inaba and Nishio 2002). This close relationship is also indicated from FISH experiments because only a few probes appear to be species specific. Some sequences used as probes, such as the “centromeric retrotransposon of Brassica” and a tandem repeat (TR805), hybridized to chromosomes of all three species (Lim et al. 2007) whereas others, for example, the tandem repeat sequences pBcKB4 (CentBr1), pBoKB1, and CentBr2, hybridized to several chromosomes of both the A and C genomes but not to the B genome (Harrison and Heslop-Harrison 1995; Lim et al. 2005, 2007).
It has been suggested that this similarity would prevent the successful application of GISH to B. napus. Excessive cross-hybridization in pericentromeric regions and a low level of hybridization in chromosome arms was observed in one GISH experiment regardless of which genome was used as the labeled probe (Snowdon et al. 1997). We have shown that the A and C genomes can be distinguished clearly, provided that B. oleracea (C genome) DNA is used as the probe and B. rapa (A genome) DNA as a block, with the addition of the C genome IGS sequence to the block improving the differentiation.
When we used labeled A genome DNA with C genome DNA as a block, the pattern of labeling suggested that A-genome-specific sequences are concentrated in the pericentromeric regions and rDNA arrays. “Pericentromeric retrotransposon of Brassica rapa” (a group of four families of Ty3-Gypsy elements) and a degenerate tandem repeat (TR238) are likely to be among these sequences because they hybridized to pericentromeric heterochromatin of several A genome but no C genome chromosomes in a FISH experiment (Lim et al. 2007).
In contrast, we conclude that C-genome-specific sequences must be widely distributed across the chromosomes because the GISH labeling was widespread when the C genome was the probe. It is likely that several different types of sequence are involved and, because every C-genome-specific sequence is labeled, the coverage that we obtain with GISH is better than that achievable with an individual transposable element or repetitive sequence.
A telomere-like repetitive sequence, pBo1.6, appeared to be C genome specific with FISH. It was localized at interstitial and/or telomeric/subtelomeric regions of all B. oleracea chromosomes and detected on 18–24 chromosomes in B. napus. However, although it was rarely detected on B. rapa chromosomes by FISH, it hybridized to B. rapa DNA on a Southern blot (Galvao Bezerra Dos Santos et al. 2007). In our GISH experiments, we used a high concentration of A genome DNA in the block and it is likely that there were enough copies of A genome pBo1.6 to hybridize with the more abundant C genome sites. As the C genome telomeric and subtelomeric regions were not labeled in our GISH experiments, we suggest that the pBo1.6 sites were blocked. The minimum level of sequence similarity required for a sequence from the A genome to effectively block a C genome site is not known but is probably >85% because, despite there being 82–85% sequence similarity between the CentBr1 and CentBr2 repeat sequences, a CentBr1 probe did not hybridize to CentBr2 sites in B. rapa FISH experiments (Lim et al. 2005).
Sequences within a BAC, BoB014O06, from B. oleracea may be among the GISH C-genome-specific sequences. This BAC specifically hybridized to all C genome chromosomes in B. napus in FISH experiments, although the hybridization signals were not homogeneous along the chromosomes (Leflon et al. 2006; Nicolas et al. 2007). One component of this BAC is Bot1, a recently identified CACTA transposon, which is highly amplified in the C genome with an estimated 1500 copies in the B. oleracea haploid genome and only 13–177 in B. rapa. When one copy was used as a FISH probe on B. napus, all the C genome chromosomes were labeled whereas only minor hybridization signals were detected on A genome chromosomes. A 2.7-kb region of the Bot1 sequence was identified as C genome specific (Alix et al. 2008). Therefore, although some regions of the Bot1 sequence may be blocked by A genome copies, this 2.7-kb region is likely to be involved in the GISH labeling of the C genome.
Other transposable elements are likely to be among the C-genome-specific sequences. Approximately 20% (120 Mb) of the B. oleracea genome is composed of transposable elements with class 1 elements, including both LTR and non-LTR retrotransposons, accounting for 14%. The remaining 6% is composed of class 2 (DNA) elements with the most abundant superfamily being CACTA (Zhang and Wessler 2004). Many subfamilies of retroelements are present in both the A and C genomes but a few that may be C genome specific have been identified through sequence analysis of PCR products from retroelements in the species of U's triangle (Alix and Heslop-Harrison 2004).
We have established that the IGS of the 45S (18S-5.8S-25S) rDNA genes is a component of the C-genome-specific sequences in our GISH experiments. The coding sequences of the 45S and 5S rDNA genes are highly conserved in plants and, consequently, those in the C genome will be blocked. In brassicas, while some sequence variation occurs in the internal transcribed spacer within the set of 45S rDNA genes (Yang et al. 1999), the IGS between the tandemly arrayed sets is much more divergent (Bhatia et al. 1996). In our experiments, the A genome DNA blocked the 45S rDNA sites of the A genome but not those of the C genome. Addition of unlabeled B. oleracea IGS to the block resulted in reduced hybridization signals from these sites, indicating that the IGS in the B. rapa genomic DNA had not hybridized effectively to them. The IGS of B. oleracea A12DHd may not be completely homologous to the IGS of the B. napus C genome since the sites were still not totally blocked. However, total blocking was not required because we wanted to label as much of the C genome chromosomes as possible.
The GISH labeling of the C genome chromosomes indicates that extensive intergenomic sequence homogenization has not occurred in B. napus. Homogenization of the rDNA arrays has occurred in some polyploids (reviewed by Wendel 2000) and it is occurring rapidly in the rDNA of two Tragopogon allotetraploids, which originated within the last 80 years (Kovarik et al. 2005). Although we saw frequent associations between 45S rDNA sites within and between genomes at diakinesis, these were probably remnants of nucleolar associations from the preceding interphase (Hasterok et al. 2005) and it is unlikely that recombination regularly occurs in these regions. The fact that the IGS of the 45S rDNA arrays is genome specific in our GISH experiments indicates that the two parental 45S rDNA array types have been maintained in B. napus. This is supported by evidence from blots of restriction-enzyme-digested DNA probed with an rDNA repeat (Waters and Schaal 1996) and from comparisons of the promoter sequences in the IGS (Chen and Pikaard 1997). This implies that recombination between the 45S rDNA arrays of the two genomes is rare and/or that a mechanism such as gene conversion is acting within but not between genomes.
The absence of large-scale translocations between the genomes in DH12075 is evident from the fact that C genome chromosomes were labeled along almost their entire length apart from the telomeres and a subtelomeric region. The only labeling of A genome chromosomes occurred faintly in pericentromeric heterochromatin, most probably due to cross-hybridization rather than translocation. This supports the conclusion drawn from RFLP mapping data that the distinct A and C genomes have been maintained to a large extent in natural B. napus throughout its evolution despite the fact that homeologous recombination occurs both in modern natural cultivars (Sharpe et al. 1995; Udall et al. 2005) and in resynthesized B. napus lines (Parkin et al. 1995; Udall et al. 2005). Individual natural B. napus cultivars possess small numbers of homeologous reciprocal or nonreciprocal translocations (Parkin et al. 1995; Sharpe et al. 1995; Osborn et al. 2003; Udall et al. 2005) but over the course of the evolution of this species there has not been a significant mixing or homogenization of the genomes, which suggests that most nonreciprocal translocations cause a reduction in fitness and that selection has acted against them.
Using GISH we cannot see if rearrangements have occurred within a genome in B. napus compared to the diploids, but no substantial rearrangements were identified when genetic maps were compared (Parkin et al. 1995; Bohuon et al. 1996; Parkin and Lydiate 1997). Therefore, it is likely that the A and C genome chromosomes in B. napus are similar to the chromosomes in the diploids and we can compare them using FISH and GISH. In a study of several species of the Brassicaceae, inter- and/or intravarietal polymorphism in the distribution of rDNA sites was observed (Hasterok et al. 2006). However, comparisons between the A and C diploid species and B. napus (Snowdon et al. 2002) or B. napus haploid lines (Kamisugi et al. 1998) have shown conservation of the number and location of 45S rDNA sites and only minor differences in signal strength. Therefore, the location but not the signal strength of 45S rDNA loci in the diploid genomes may be used to identify homologs in B. napus with some confidence. With 45S rDNA and BoB061G14 as FISH probes followed by GISH, we were able to distinguish several chromosomes of B. napus and to identify potential homologs in the karyotypes of B. oleracea (Howell et al. 2002) and B. rapa (Lim et al. 2005). Since the chromosomes of the diploids have been associated with linkage groups of genetic maps (Howell et al. 2002; http://www.brassica.info) and molecular marker analysis has confirmed homology between these linkage groups and those of B. napus (A1-10 = R1-10 = N1-10 and C1-9 = O1-9 = N11-19) (Parkin and Lydiate 1997), we can tentatively assign linkage groups to these chromosomes. Subsequently, five of these have been confirmed using B. oleracea chromosome-specific FISH probes.
The C genome chromosomes of B. napus line DH12075 showed the same distribution of signals as those of B. oleracea A12DHd except that the interstitial 45S rDNA signal of BolC4 was not present (Table 1). Since this site is absent from other B. oleracea accessions, this was expected. C7 has a BoB061G14 signal and a distal 45S rDNA signal. C8 has a distal 45S rDNA signal but the signal intensity is less than in A12DHd. Neither probe hybridizes to C3 and C9 (confirmed, not shown) and these chromosomes can be differentiated by size and centromere position in mitotic spreads (not shown). The remaining five chromosomes form a group with only BoB061G14 signals. The pattern of hybridization of 45S rDNA and BoB061G14 in the A genome was the same as that described for 45S rDNA and CentBr1 in B. rapa (Lim et al. 2005). Therefore, we can identify two A chromosomes and divide the remainder into two groups (Table 1). A3 has a large distal 45S rDNA signal (confirmed, not shown) and A5 has a 45S rDNA signal near the centromere. A1, A6, and A9 have both 45S rDNA and BAC signals and the remaining five have only a BAC signal. We have also shown that B. oleracea BACs can effectively mark chromosome BnaC6 and, to a lesser extent, the homeologous part of BnaA7. It should be possible to mark any particular homolog in B. napus with BACs from the diploids because chromosome-specific BACs have been identified for all of their chromosomes (http://www.brassica.info).
Many B. napus cultivars studied with molecular markers have been shown to have homeologous reciprocal or nonreciprocal translocations (Parkin et al. 1995; Sharpe et al. 1995; Osborn et al. 2003; Udall et al. 2005). The one that has been investigated most thoroughly is an interstitial reciprocal translocation involving A7 and C6, which is present in several Canadian and Australian cultivars. The source of this translocation may be one of three annual forms of B. napus, which are common to all their pedigrees (Osborn et al. 2003). With sequential FISH and GISH, on pachytene spreads we identified this reciprocal translocation in Westar and N-o-1. The proximal end of the translocation was between the signals from BoB057M06 and BoB028L01 and therefore between markers pO10E1 and BoAP1-c (Howell et al. 2005), agreeing with the genetic marker data that places it below pO10eNP (Osborn et al. 2003). However, we were unable to confirm that the translocation was interstitial as there was no sign of GISH labeling near the telomeric end of the unlabeled segment. This suggests that the distal breakpoint was in the small region close to the telomere where labeling in the C genome was absent. The fact that the more distal synapsis exchange point of the two indicated in a synaptonemal complex spread of an F1 hybrid N-o-9 × N-o-1 (Osborn et al. 2003) was not far from the telomere supports this suggestion. The exchanged region was approximately one-third of the length of the A7 and C6 chromosomes in synaptonemal complex spreads (Osborn et al. 2003) and our images of the whole chromosomes at pachytene agree with this. C6 is the smallest chromosome of the C genome complement (Howell et al. 2002).
At diakinesis, the A7 chromosome with the segment from the C genome can be seen clearly but the C6 chromosome with the A genome segment cannot and it is also not easy to identify this chromosome in every spread at mitosis. Using GISH on mitotic spreads of wheat containing recombined wheat–rye chromosomes, a small labeled translocated segment was detected on an unlabeled chromosome. However, when the same segment was not labeled but the chromosome was labeled, the segment could not be detected and it was suggested that the halo effect of the probe obscured the unlabeled segment (Lukaszewski et al. 2005). This implies that, since we can label only the C genome in B. napus, small translocations of C genome chromosomes will be easier to detect than the reciprocal translocations. As noted above, small translocations involving only the unlabeled subtelomeric regions will not be detected by GISH.
We have found that using FISH to assign BACs containing paralogous genes to their respective linkage groups is a useful alternative to genetic mapping in B. oleracea, particularly if polymorphism is difficult to detect. Using pachytene spreads of B. oleracea, we can also assist the sequencing of some regions by confirming the order and relative position of selected BACs (Howell et al. 2005). Now that we have developed GISH and shown that B. oleracea BACs can be used on B. napus to identify chromosomes, it will be possible to allocate B. napus BACs to linkage groups in a similar way.
We are using both GISH and FISH on pollen mother cells at metaphase I to study the genetic control of homeologous pairing and recombination in B. napus. Experiments with B. napus haploids (AC) showed that the amount of pairing between chromosomes varied depending on the varieties from which the haploids originated and that this was controlled by a major locus, PrBn, together with other loci. However, the B. napus varieties themselves (AACC) had normal bivalent pairing (Jenczewski et al. 2003; Liu et al. 2006). From genetic mapping, it is apparent that the frequency of homeologous recombination is much higher in resynthesized than in natural B. napus (Parkin et al. 1995) and it may be possible to exploit this situation to determine whether different loci or different alleles at the PrBn locus are involved in the control of homeologous recombination in B. napus. The techniques will also be useful in evolutionary studies on newly synthesized B. napus allopolyploids because some of the intergenomic translocations resulting from homeologous recombination can be identified.
We are grateful to S. Price and K. Staples (University of Birmingham, Birmingham, UK) for technical assistance and to I. A. Parkin (AAFC, Saskatoon, Saskatchewan, Canada) and C. Morgan (JIC, Norwich, UK) for seed. This work was supported by the Biotechnology and Biological Sciences Research Council (grant no. G18621).
Communicating editor: A. H. Paterson
- Received September 9, 2008.
- Accepted October 3, 2008.
- Copyright © 2008 by the Genetics Society of America