SEARCH

SEARCH BY CITATION

Keywords:

  • allopolyploidy;
  • Brassica;
  • chromosome rearrangement;
  • gene conversion;
  • homoeologous recombination;
  • Muller’s ratchet;
  • polyploidy;
  • recombination

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

Polyploidization and recombination are two important processes driving evolution through the building and reshaping of genomes. Allopolyploids arise from hybridization and chromosome doubling among distinct, yet related species. Polyploids may display novel variation relative to their progenitors, and the sources of this variation lie not only in the acquisition of extra gene dosages, but also in the genomic changes that occur after divergent genomes unite. Genomic changes (deletions, duplications, and translocations) have been detected in both recently formed natural polyploids and resynthesized polyploids. In resynthesized Brassica napus allopolyploids, there is evidence that many genetic changes are the consequence of homoeologous recombination. Homoeologous recombination can generate novel gene combinations and phenotypes, but may also destabilize the karyotype and lead to aberrant meiotic behavior and reduced fertility. Thus, natural selection plays a role in the establishment and maintenance of fertile natural allopolyploids that have stabilized chromosome inheritance and a few advantageous chromosomal rearrangements. We discuss the evidence for genome rearrangements that result from homoeologous recombination in resynthesized B. napus and how these observations may inform phenomena such as chromosome replacement, aneuploidy, non-reciprocal translocations and gene conversion seen in other polyploids.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

Cytogenetic studies provided the earliest observations for genomic shock and meiotic instabilities inherent to newly formed hybrids and polyploids (Navashin, 1934; Stebbins, 1971; Grant, 1975; Sybenga, 1975; McClintock, 1984; Levin, 2002). In the last decade, molecular data from resynthesized and natural allopolyploids indicate that genetic and epigenetic changes are common consequences of polyploidization across a wide range of species (Wendel, 2000; Osborn et al., 2003b; Chen, 2007; Doyle et al., 2008; Hegarty & Hiscock, 2008; Leitch & Leitch, 2008; Soltis & Soltis, 2009). We understand little of the mechanisms that lead to these changes and know even less about their directed or random nature. However, data from some species suggest that genetic changes may result from homoeologous recombination. Recombination involves some of the most important mechanisms contributing to genetic variation and genome structural diversity in plants (Gaut et al., 2007). It is a process that generates novel combinations of genetic material, eliminates deleterious mutations and plays a role in DNA repair. During meiosis in plants, recombination predominantly occurs among allelic sequences on homologous chromosomes; however, in allopolyploids which lack diploid pairing fidelity recombination may occur ectopically among paralogous or homoeologous sequences. Homoeologous recombination may lead to reciprocal exchange and gene conversion, explaining many of the small and large genetic changes detected in newly formed allopolyploids.

Our goal in this paper is to give an overview of mechanisms of recombination in plants, chromosome pairing, recombination and segregation in allopolyploids, and how these mechanisms might contribute to genetic changes in natural and resynthesized polyploids. We review the evidence for homoeologous recombination and chromosome rearrangements in Brassica napus polyploids. We introduce the idea of a ‘polyploid ratchet’ (inspired by ‘Muller’s ratchet’, Muller, 1964; Felsenstein, 1974) to account for the high frequency of genetic changes in early generations of resynthesized B. napus when compared with natural, domesticated B. napus. Finally, we suggest that B. napus offers a window into allopolyploid genome evolution involving deletions, chromosome rearrangements, chromosome replacement and aneuploidy – all phenomena now being reported in natural allopolyploids such as Tragopogon (Lim et al., 2008; Z. Xiong & J. C. Pires, unpublished). We hypothesize that homoeologous recombination, multivalent pairing, and segregation of translocation heterozygotes are plausible mechanisms for dynamic structural changes in allopolyploids, included gene loss and conversion.

Mechanisms of recombination in plants

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

Meiosis and the genes responsible for carrying out recombination are highly conserved (Zickler & Kleckner, 1999; Bhatt et al., 2001; Anderson and Stack, 2002). Many plant homologs for genes involved in meiotic recombination have been identified (for review see Schwarzacher, 2003; Schuermann et al., 2005). During meiosis, homologous chromosomes undergo reciprocal exchange (crossing over; CO) and gene conversion (noncrossover; NCO), events leading to novel haplotypes. Recombination is initiated by double-strand DNA breaks (DSBs) (Fig. 1). Homologous recombination is the major route for DSB repair during meiosis, which involves homology search, DNA synthesis and DNA repair. Two models have been used to explain CO and NCO: the double-strand break-repair (DSBR) model, which suggests a mechanism for reciprocal exchange and gene conversion during meiotic recombination (Szostak et al., 1983; Fig. 1a–e), and the synthesis-dependent strand annealing (SDSA) model, which can explain gene conversions (Nassif et al., 1994;Rubin & Levy, 1997; Puchta, 2005; Cromie & Smith, 2007; Fig. 1f–g). Evidence from several model organisms suggests that while the DSBR model provides a likely explanation for many COs, it cannot explain all NCOs (Cromie & Smith, 2007). Most DSBs are repaired in a nonreciprocal manner, and few involve COs (Mezard et al., 2007). While reciprocal exchanges can be quite large, the size of gene conversion varies (generally a few hundred to a few thousand base pairs) (Hilliker et al., 1994; Xu et al., 1995; Haubold et al., 2002). For example, conversion tracts in maize may range from 1 to 3 kb (Dooner & Martinez-Ferez, 1997).

image

Figure 1.  Meiotic recombination may lead to crossover and/or gene conversion. The double strand break-repair (DSBR; b–e) and the synthesis-dependent strand annealing (SDSA; f–g) models for recombination are shown. Homologous chromosomes (grey and black) are shown with both chromatids, each of which is shown with both strands. (a) Under the DSBR and SDSA models, recombination is initiated by a double-strand break (DSB) in one chromatid. Double-stranded DNAs are resected (degraded by nucleases) at 5′ ends, exposing 3′ single strands, which invade homologous chromatids (b,f). (b) Under the DSBR model, the displaced strand forms a D-loop that invades the broken chromatid and hybridizes with the other free 3′ single-stranded DNA, forming recombination joints. (c) 5′ to 3′ synthesis (arrow heads) extends invading DNA, and branch migration leads to heteroduplexes (checkered) near the point of exchange. Heteroduplexes are stretches of DNA composed of hybridized homologous single-stranded DNAs that have mismatches, which are recognized by mismatch-repair machinery. Two Holliday junctions are formed. (d) If one Holliday junction is cut in the strands involved in exchange, and the other is cut in the strands not involved in exchange (arrows), reciprocal recombination of flanking regions will occur. Heteroduplexes will be repaired, and if an invading strand is used as the repair template, gene conversion will also occur (not shown). (e) If both Holliday junctions are cut in the strands involved in exchange (arrows) or both are cut in the strands not involved in exchange, no crossover will occur; however, if the heteroduplexes are repaired using the invading strand as a template, gene conversion results. (f) Under the SDSA model, a single-strand DNA invades a homologous sequence and copies the invaded sequence via DNA synthesis. (g) The reaction concludes with strand ‘pull out’, and mismatch repair of heteroduplexes, which can lead to gene conversion.

Download figure to PowerPoint

The main factor driving recombination between any two sequences is homology (for a review see Naranjo & Corredor, 2008). During meiosis in plants, DSBs are induced by Spo11 homologs (a topoisomerase-related enzyme) during leptotene (Keeney et al., 1997; Keeney, 2001). These breaks lead to exposed single-stranded DNAs which act as targets for homology search, a process involving homologs of Rad51 and DMC1 (Franklin & Cande, 1999; Paques & Haber, 1999; Pawlowski et al., 2003). In maize, RAD51 foci have been observed along synapsed chromosomes, which are believed to be the locations where homologous strand invasion has led to the formation of recombination joints (Franklin et al., 1999). Telomere and centromere clustering may also play a role in chromosome association, however, in plants, homology is essential to homologous chromosome pairing (Dernburg et al., 1995; Dawe, 1998; Stewart & Dawson, 2008).

Meiotic recombination predominantly occurs among homologous chromosomes (for a review see Naranjo & Corredor, 2008) but it does not occur randomly along chromosomes and hotspots vary among species (Anderson et al., 2001; Anderson & Stack, 2002; Jensen-Seaman et al., 2004). Hotspots have been found within genes in maize (Civardi et al., 1994; Xu et al., 1995) and in intergenic regions in Arabidopsis thaliana (Kim et al., 2007). While many plants display a propensity for recombination at the distal ends of chromosomes near the telomeric or subtelomeric regions and lack recombination near the centromere (Drouaud et al., 2006), the reverse is true in some species of Allium (Khrustaleva et al., 2005). Meiotic recombination predominantly occurs among allelic regions of homologous chromosomes (for a reviewed see Naranjo & Corredor, 2008). Recombination can also occur among repeats within the same chromosome (intrachromosomal exchange; Lysak et al., 2006), unequally among repeats on homologous chromosomes (see Jelesko et al., 2004) or promiscuously among nonhomologous chromosomes that share some degree of homology (for a review see Gaut et al., 2007; Mezard et al., 2007). The consequence of these events includes genetic change (Fig. 2).

image

Figure 2.  Recombination and chromosome rearrangement. (a) Deletion by intrachromatid recombination among duplicated sequences in diploids and polyploids. (b) Deletions and duplications caused by unequal pairing and crossover (CO) among repeat sequences on homologous chromosomes in diploids and polyploids. (c) Intergenomic deletions and duplications can result from homoeologous recombination in an allotetraploid.

Download figure to PowerPoint

DNA repair may also cause mutations. During DNA synthesis, DSBs can initiate breakage-induced replication during which stalled replication forks undergo strand breakage and single-strand invasion of sister or non-sister chromatids, which are used as templates for DNA repair (reviewed by Sung & Klein, 2006). The transfer of genetic material is nonreciprocal, and can result in gene conversion. Double strand breaks in somatic cells can also be repaired by nonhomologous end joining (NHEJ), which has no requirement for homology (Puchta, 1999; reviewed by Puchta, 2005). Errors during NHEJ can result in insertions and deletions at the site of ligation, and the ligation of unlinked DNA may generate translocation (Pipiras et al., 1998; Puchta, 2005). For example, in Arabidopsis unlinked DSBs can act as substrates for reciprocal exchange by both homologous recombination and non-homologous end joining (Pacher et al., 2007).

In summary, recombination during meiosis and DNA repair can cause chromosomal rearrangements (deletions, duplications, gene conversions and translocations) in somatic or meiotic cells. While genetic variation may arise in the genomes of somatic cells of a multicellular organism during development, most will not contribute to gametophytic tissues or be transmitted to the next generation. Meanwhile, genetic changes that arise during meiosis as a consequence of recombination have a better chance of contributing novel variation to the next generation. In the following sections we explore the consequences of homoeologous recombination in allopolyploids.

Chromosome pairing, recombination, and segregation in allopolyploids

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

In this review, we distinguish allopolyploidy from autopolyploidy by parentage (Ramsey & Schemske, 1998, 2002) and not chromosome pairing behavior. Allopolyploids may arise from the fusion of unreduced gametes or by hybridization followed by genome doubling (for a review see Ramsey & Schemske, 1998; Comai, 2005). The consequence is that two or more divergent genomes (subgenomes) reside within one nucleus, each of which contains complete sets of homologous chromosomes. Sets of homologous chromosomes in a polyploid are considered homoeologous to one or more sets from the other subgenomes if the chromosomes share a common ancestral type (and thus genomic synteny). As a result, a somewhat challenging scenario is posed during meiosis in allopolyploids compared with diploids. Homologous chromosomes must pair faithfully and nonhomologous associations avoided, lest the genome be subject to a breakdown in disomic inheritance, the consequences of which may include chromosome rearrangements and aneuploidy (Figs 2c, 3).

image

Figure 3.  Meiosis I in an allopolyploid carrying a homoeologous translocation. In this example, homoeologous chromosomes A (black bars) and C (grey bars) are shown. (a) The C-homoeologs carry a terminal translocation from the distal end of the A chromosome (CA) resulting from homoeologous recombination in a previous generation. Lines carrying such translocations pair in a cross-like configuration. (b) In this example, if we assume no new recombination, chromosome segregation that is alternate (Alt) or adjacent 1 (Adj1) will generate daughter cells of the parental type. However, if adjacent 2 (Adj2) segregation occurs, daughter cells will be produced that lack one homoeolog or the other. In this way, homoeologous chromosome loss and replacement can occur, as well as the loss of homoeologous centromeres. (c) Depending on the location and number of crossovers, meiosis I chromosome segregation could also lead to nondisjunction (3 : 1 or 4 : 0 chromosome segregations) and aneuploidy.

Download figure to PowerPoint

Strict homologous chromosome pairing is displayed by both diploids and diploidized allopolyploids (Sybenga, 1975). Disomic inheritance requires paired centromeres that are aligned for equal segregation of homologs in Meiosis I and equational segregation of chromatids during Meiosis II. At the other extreme, inheritance in an allopolyploid may resemble that of an autopolyploid, in which there is no discrimination between the multiple sets of homologous chromosomes (see Cifuentes et al. (2100) in this volume). Most allopolyploids fall somewhere between these two extremes, and in some cases some chromosomes may behave in a disomic manner, while others may pair as multivalents. Stebbins (1971) refers to this situation as ‘segmental allopolyploidy.’ Multivalents can result in the interlocking of homoeologous and homologous chromosomes, and during metaphase pairing configurations may include cross-structures, rings and chains of chromosomes (Sybenga, 1975; Grant, 1975: Fig. 3). The segregation of chromosomes from these structures often leads to duplication and deficiency gametes, and aneuploidy, all of which lead to reduced fertility (Gillies, 1989). Segregation for any given gene is a consequence of centromere proximity and the frequency and distribution of crossovers among paired chromosomes.

There is significant evidence that homologous (and thus homoeologous) pairing is under genetic control. In the well-studied allohexaploid bread wheat, three distinct, yet related genomes coordinate meiotic pairing such that all three sets of chromosomes (A, B and D genomes) pair faithfully with their homologs and segregate disomically. Genetic control of chromosome pairing is mediated by the PH1 locus (Okamoto, 1957; Riley & Chapman, 1958; Sears, 1977; Griffiths et al., 2006). Mutations at this locus lead to homoeologous recombination and gross chromosomal rearrangement (Qi et al., 2007). The PH1 locus includes 70 Mb region of chromosome 5B; however, although it is 50 yr since its first discovery, we are only starting to understand the mechanism through which it acts. Similarly, in B. napus polyploids the PrBn locus regulates chromosome pairing, although its effect is only observed strongly at the allohaploid and allotriploid levels (Jenczewski et al., 2003; Nicolas et al., 2009).

While mispairing may involve homoeologous associations, it may also include paralogous associations, both of which lead to complicated modes of inheritance, chromosomal rearrangement, and aneuploidy. For example, the genome of B. napus contains approximately six copies of many loci because the progenitors species each contain approximately three loci resulting from ancient duplications (defined as paralogs). In B. napus allohaploids, most bivalents observed during meiosis show allosyndesis (involving pairing of homoeologous chromosomes), while up to 30% show autosyndesis (involving chromosomes within the same parental genome that presumably contain duplicated regions) (Nicolas et al., 2009). If pairing and recombination occur among homoeologous or paralogous sequences, disomic inheritance is disrupted and genome rearrangements may result; however, there is limited data on the rate of genetic changes resulting from paralogous associations in allotetraploids. Since pairing control and fertility are often related, it is assumed that fertile allopolyploids must have either had some level of pre-existing control over pairing, or in some way acquired genetic control for this during their evolution. It is also possible that structural changes that occur in newly formed polyploids (e.g. expansion or contraction of repeat elements or other genomic rearrangements) contribute to divergence among homoeologous chromosomes and facilitate proper homolog pairing.

Genetic changes in natural and resynthesized polyploids: from pattern to mechanism

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

Evidence for genetic changes in polyploids has been found in resynthesized polyploids (e.g. Arabidopsis suecica, B. napus, Triticum aestivum, Gossypium hirsutum, Nicotiana tabacum and Triticale) and natural polyploid species (e.g. Tragopogon, Senecio, Spartina and Glycine). Studies have reported on sequence deletion (Ozkan et al., 2001; Shaked et al., 2001; Ma & Gustafson, 2006, 2008; Tate et al., 2006, 2009), gene conversion (Wendel et al., 1995; Kovarik et al., 2004, 2005, 2008; Salmon et al., 2010), rDNA loci changes (Lim et al., 2000, 2008; Joly et al., 2004; Pontes et al., 2004), transposon activation (Kashkush et al., 2002, 2003; Madlung et al., 2005; Parisod et al., 2010), chromosomal rearrangements (Kenton et al., 1993; Parkin et al., 1995; Lim et al., 2004, 2006; Pires et al., 2004; Udall et al., 2005; Gaeta et al., 2007), and aneuploidy (Lim et al., 2008). However, despite the renaissance in observations of genetic changes, we have made less progress on identifying the underlying mechanisms of these changes.

Genetic changes in polyploids are frequently observed using molecular markers; however, the problem with marker analysis in polyploids is that genetic changes are often detected after the initial event has occurred, leaving the details of their origins a mystery. Indeed, many studies resort to ‘bandology’, counting missing parental markers as gene losses. In some cases parental bands share the same molecular weight, making it difficult identify some deletions unless stringent methods for resolution are used (e.g. single-strand conformation analysis). In addition, markers that are not sensitive to duplications cannot distinguish gene losses or gene conversion from loss/duplication events as discussed below. This limitation is further exacerbated by a general difficulty in distinguishing between homologous and homoeologous markers. It is useful to have mapped molecular markers to distinguish linked events, as a single deletion or rearrangement could account for the loss of many genes. Strong evidence for the mechanisms of genetic change comes from the use of genome-wide mapped molecular markers, combined with physical analysis of chromosomes by fluorescent in situ hybridization (FISH). Recently, there has been renewed interest in using cytogenetic analysis with FISH in polyploids (Lim et al., 2007, 2008; Pires & Hertweck, 2008; Z. Xiong & J. C. Pires, unpublished).

Studies in B. napus have provided strong evidence that homoeologous recombination contributes to genetic changes (Jenczewski et al., 2003; Pires et al., 2004; Udall et al., 2005; Leflon et al., 2006; Liu et al., 2006; Osborn et al., 2003a; Gaeta et al., 2007; Nicolas et al., 2007, 2008, 2009; Cifuentes et al., (2010); Szadkowski et al., (2010)). In B. napus mapping populations and resynthesized polyploids, restriction fragment length polymorphism markers have revealed both the loss and the duplication of homoeologous loci (Parkin et al., 1995; Sharpe et al., 1995; Osborn et al., 2003a; Pires et al., 2004; Udall et al., 2005; Gaeta et al., 2007). These changes have been described as homoeologous nonreciprocal transpositions (HNRTs) because they are detectable when the loss of one homoeologous marker occurs together with a duplication of the other (loss/duplications). However, as rightly pointed out by Nicolas et al. (2008); the term ‘nonreciprocal’ is a bit misleading because the initial event is probably an intergenomic crossover (CO), and may therefore have been reciprocal. Reciprocal exchanges are difficult to detect because individuals carrying them will demonstrate marker additivity. However, reciprocal exchanges segregate during self-pollinations and their products are detectable when they are homozygous, thus appearing nonreciprocal (Figs 2c, 4). When loss–duplication dosage changes are observed among linked homoeologous markers, HNRTs are visible as cytological translocations (Z. Xiong & J. C. Pires, unpublished).

image

Figure 4.  Simplified model of segregating homoeologous exchange in self-pollinated resynthesized allotetraploid Brassica napus. Grey chromosomes represent A1 from the Brassica rapa subgenome and black chromosomes represent C1 from Brassica oleracea subgenome. (a) Recombination among homoeologous chromatids occurs, and segregation of exchange products assuming a disomic chromosome model of inheritance (homologs move to opposite poles and recombined homoeologs move to opposite poles). Each form of the reciprocal exchange (or in the case of gene conversion, the nonreciprocal change) would segregate 3 : 1 in the gametes. (b) Self-pollination would fix opposite forms of HNRTs in c. 1/16 progenies, parental types in c. 4/16 progenies, rearranged but balanced (reciprocal translocations) individuals in c. 2/16 progenies, and most progenies (c. 8/16) would be heterozygous for one form of the rearrangement or its reciprocal and contain a 3 : 1 dosage of parental homoeoalleles. The two lines fixed for homoeologous rearrangements (upper left and lower right corners) are detectable by molecular marker analysis. A 1/16 frequency is similar to what has been observed in segregants derived from the first round of meiosis in resynthesized allopolyploids (Gaeta et al., 2007, supplementary data). This model is an oversimplification, and when pairing occurs among homoeologs the resulting chromosome segregation patters would be dependent on the location, size, and frequency of chiasmata, and segregation may fall between disomic and tetrasomic models.

Download figure to PowerPoint

In recently formed Tragopogon allopolyploids, the loss of genetic markers has also been found to correlate with karyotype changes observable by FISH, possibly caused by homoeologous recombination in some instances (Tate et al., 2006, 2009; Lim et al., 2008). Research with Nicotiana and Arabidopsis allopolyploids has found that rDNA loci may be lost (or homogenized) or translocated after polyploidization, and in some cases unit amplification has been observed (Lim et al., 2000; Pontes et al., 2004; Kovarik et al., 2008). In these species rDNA loss and homogenization may have resulted from homoeologous recombination and gene conversion, and random segregation, or from chromosome breakage and NHEJ. Pontes et al. (2004) suggested that transposon-mediated breakage and recombination might also play an important role in rDNA restructuring in A. suecica; this is corroborated by the observation that transposons are found in high densities near the nucleolus organizer region that is most unstable in the allopolyploid. Unit amplification reported in Nicotiana may have resulted from unequal crossing over (see Fig. 2b; Kovarik et al., 2008). In some resynthesized species, deletions have been reported to occur specifically in the paternal genome (Song et al., 1995; Skalicka et al., 2005). These events could also be caused by intrachromatid exchanges (see Fig. 2a) or transposon excision; however, the reason for their paternally directed nature remains unknown. In resynthesized wheat, low-copy, coding and non-coding sequences are lost from specific chromosomes and subgenomes (Feldman et al., 1997; Ozkan et al., 2001; Shaked et al., 2001; Kashkush et al., 2002). Because homoeologous pairing is strongly suppressed in wheat by the PH1 locus, some of these deletions could be caused by intrachromosomal recombination (Fig. 2a) or DNA transposon excision. In contrast to most resynthesized allopolyploids, few genetic changes have been detected in resynthesized Gossypium and mapping studies suggest that intergenomic translocations are rare in natural allopolyploids, suggesting that homoeologous pairing and recombination are rare events in this species (Rong et al., 2004; Liu et al., 2001).

Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References

Lukens et al. (2006) and Gaeta et al. (2007) studied a population of 50 independently resynthesized B. napus allopolyploids in the S0 : 1 (S0) and S5 : 6 (S5) generations. These self-pollinated allopolyploids were created using doubled-haploid parents, allowing for unambiguous identification of mapped homoeologous molecular markers. Homoeologous translocations were detected at a very low frequency in the S0 generation of these resynthesized allopolyploid lines when bulked DNA samples of 8–12 individuals per line were analysed. In the S5 generation nearly 5% of markers analysed were deleted, most of which resulted from HNRTs. Rearrangements were not randomly distributed across the genome and occurred frequently among highly similar homoeologous linkage groups A1–C1, A2–C2 and A3–C3. Reciprocal forms of most rearrangements were found segregating in independent lineages. Some of the markers that detected rearrangement in this population of resynthesized allopolyploids had previously detected preexisting and de novo homoeologous rearrangements in synthetic × natural mapping populations (Udall et al., 2005). Similarly, these same chromosomes underwent frequent homoeologous pairing and recombination in other studies of resynthesized B. napus allohaploids and allotetraploids (Jenczewski et al., 2003; Szadkowski et al., (2010)).

Three markers at the end of A1–C1 homoeologous linkage groups were used for genotyping several lines across generations, as well as individual plants (segregants) within generations (Gaeta et al., 2007; supplemental data). As stated above, when S1 segregants within each line were bulked for analysis of the S0 generation polyploids, genetic changes were extremely rare; however, when individual S1 plants derived from eight independent S0 lines were genotyped, rearrangements were segregating from the first meiosis at a low frequency (3/118 S1 segregants carried a HNRT), indicating that some HNRTs were missed in the bulk analysis. A reciprocal exchange among a pair of homoeologs during meiosis would generate reciprocal forms at a frequency of c. 1/16 segregants following self-pollination under disomic inheritance (Fig. 4). The three HNRTs detected among segregants occurred in three different lines at frequencies of 1/14, 1/15, and 1/16, respectively (see Gaeta et al., 2007; supplemental data). This frequency is very similar to the expected frequency (1/16); however, owing to limited sample size, further analysis is warranted. Within a self-pollinated lineage, a single homoeologous exchange would lead to an array of progeny genotypes carrying various dosages and arrangements of homoeoalleles (Fig. 4).

Genetic marker analysis on bulked DNA across generations revealed that some HNRTs progressed across generations (Fig. 5; see Gaeta et al., 2007; supplemental data). In one example, an interstitial rearrangement was extended toward the telomere in a later generation (Fig. 5a). In another case, we detected a terminal rearrangement that preceded the detection of a linked, interstitial rearrangement in a later generation (Fig. 5b). In this second example, the linked, interstitial transposition detected in the later generations was nonreciprocal for the other parental fragment. Several rounds of homoeologous pairing and genetic exchange could explain these results. Initial exchanges may have destabilized subsequent pairing events, leading to more extensive chromosome rearrangements in subsequent meioses. In each case, individual segregants were genotyped from the generation before the detection of a HNRT in the bulk analysis, and almost invariably the HNRTs were segregating 3 : 1 (see Fig. 5). Rearrangements of variable length (determined by the number of linked markers involved) were observed in some segregating individuals, and in one case reciprocal rearrangements were detected (1/16 carried one form and 1/16 carried the reciprocal form). Together, these data suggest that homoeologous translocations in resynthesized B. napus allopolyploids originated from intergenomic recombination events that segregated in a somewhat disomic pattern and fixed during the process of selfing. As intergenomic homogenization occurred through the fixation of translocations, sequence similarity increased between homoeologs and may have increased homoeologous pairing and exchange. Over many generations translocation heterozygotes would increase, leading to distorted segregation, aneuploidy, chromosome loss-and replacement and centromere homogenization (see Fig. 3). The consequence is somewhat ratchet-like (polyploid ratchet), leading to increased instability in the absence of acquired pairing control. These types of chromosome anomalies have recently been reported in recently formed natural Tragopogon allopolyploids and may arise as a consequence of homoeologous recombination (Lim et al., 2008).

image

Figure 5.  The case of the polyploid ratchet. After an initial exchange among homoeologs, further pairing and exchange in future meioses can lead to additional translocation. The consequence is ‘ratchet-like’, leading to increased instability. Gaeta et al. (2007) analysed several self-pollinated B. napus allopolyploids from generations S0 to S5. Examples from two resynthesized polyploid lines (A and B) are shown. Grey chromosomes represent A1 from the Brassica rapa subgenome and black chromosomes represent C1 from the Brassica oleracea subgenome. Three markers at the end of A1–C1 homoeologous groups were used to screen bulked tissue across generations, and individual plants (segregants) within generations (see Gaeta et al., 2007, supplemental data). (a) In line A, an interstitial nonreciprocal transposition was detected in the S4 generation (loss of B. rapa markers, duplication of B. oleracea markers). A terminalized translocation was detected in the S5 generation and was segregating in exactly 25% of 16 S4 plants when individually genotyped. Because the terminalized translocation was not yet fixed it went undetected in the bulk analysis of the S4 generation. Thus, the rearrangement detected in the S5 generation had arisen as a consequence of several rounds of meiosis and genetic exchange during self-pollination. (b) In line B, a distal/terminal nonreciprocal translocation was detected in the S2 generation by the two most terminal markers screened. Again, exactly 25% of 16 S1 plants were segregating for this rearrangement. In the S5, the most proximal marker detected a linked, interstitial rearrangement; however, it was nonreciprocal for the other parent relative to the distal rearrangement. This proximal rearrangement was segregating in three of 16 individual S4 plants. These observations are consistent with several rounds of genetic exchange, which resulted in larger, more complex rearrangements.

Download figure to PowerPoint

Mapping studies in natural B. napus have detected pre-existing and de novo homoeologous rearrangements (Parkin et al., 1995; Sharpe et al., 1995; Udall et al., 2005), but they occur at a much lower frequency than that which been observed in resynthesized lines (Song et al., 1995; Gaeta et al., 2007). Similarly, comparative genome analyses among extant Brassica oleracea, Brassica rapa, and B. napus have found little evidence for extensive rearrangement in genome microstructure (Rana et al., 2004). As genetic variation in homoeologous pairing exists in B. napus (Jenczewski et al., 2003; Liu et al., 2006), the higher rate of change in resynthesized lines could be attributed to the genetic background of the parental lines used for synthesis.

Perhaps a better explanation for fewer observed homoeologous recombination events seen in domesticated vs resynthesized B. napus is that it is a consequence of natural selection. Domesticated B. napus likely underwent natural selection during its early evolution and currently undergoes selection for fertility and other agronomic traits. Thus, extensive genome rearrangements may have been selected against. By contrast, resynthesized genotypes have been selected for extreme flowering time (Pires et al., 2004) or self-pollinated without selection by single-seed descent, despite extensive rearrangements and deleterious phenotypes (Gaeta et al., 2007). Homoeologous rearrangements in the early generations after formation can contribute significant variation to gene expression and phenotypes (Pires et al., 2004; Gaeta et al., 2007), and this may help a new polyploid become established or exploit new niches. Resynthesized lines with extensive rearrangements generally show low fertility (Gaeta et al., 2007), and in nature such lines would probably not contribute to future generations. Any rearrangements that beget further mispairing and rearrangement lead to a ‘polyploid-ratchet’ effect, and will likely be selected against. Future studies can test these hypotheses in resynthesized B. napus lines by allowing selection in field-type analyses in natural environments or selection for fertility.

Conclusions

Molecular analysis in polyploids has provided evidence that the genomic events that follow polyploidization include genetic and epigenetic changes, changes in gene expression and changes in phenotypic variation. However, the mechanisms of, and consequences for, these changes are still largely unknown. Studies in B. napus polyploids have pointed to homoeologous pairing and recombination as a key mechanism for genome restructuring; however, there is still much to be learned regarding the consequences of homoeologous rearrangement on aneuploidy and chromosome dosage changes, and whether it contributes to deletion and gene conversion reported in other polyploids. Recent studies in Tragopogon that included FISH analysis provided evidence that deletions and rDNA changes detected in recent allopolyploids might also be the consequences of homoeologous rearrangements (Lim et al., 2008). While studies of newly resynthesized allopolyploids continue to provide a window on the early events following hybridization and polyploidization, future studies should include large numbers of independently resynthesized or recently formed natural polyploids, ideally with multiple parental genotypes and reciprocal crosses. This would allow for statistically relevant comparisons of the relative propensities of different mutations (e.g. structural vs epigenetic; nonreciprocal vs reciprocal) across genotypes within a species and among different species or higher taxa. In addition, analyses must go beyond simple ‘bandology’ and combine mapped dosage-sensitive markers with cytogenetics to reveal not only patterns of genetic changes but also mechanisms of change. Future studies might address how environmental variations contribute to genome structure through analysis of newly formed allopolyploids over several generations under field-like conditions, or by selection for specific phenotypes (Pires et al., 2004). Resynthesized polyploids provide a unique model for studying the role of recombination on genetic changes, and how novel mutations may lead to new phenotypes and heterosis that may be important for niche exploitation and speciation under natural selection.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Mechanisms of recombination in plants
  5. Chromosome pairing, recombination, and segregation in allopolyploids
  6. Genetic changes in natural and resynthesized polyploids: from pattern to mechanism
  7. Homoeologous recombination and chromosome rearrangements in B. napus allotetraploids; the case of the polyploid ratchet
  8. Acknowledgements
  9. References