A major source of novelty in polyploids may come through genome rearrangement and gene silencing after polyploidization, as was mentioned before. Grant (1966) was the first to experimentally demonstrate this possi- bility when he produced a hybrid of tetraploid Gilia malior Gilia mod- ocensis that had seed fertility of 0.007%. He selfed the lines for 11 Table 4.8. Morphological and environmental ranges of diploid and octoploid Fragariain California. A value less than 1.0 indicates that the octoploids had a greater range. (From Hancock and Bringhurst, 1981.)
Range Range
Character 2x/8x Character 2x/8x
Environmental Morphological
January temperature 0.85 Stolon width 0.52
April temperature 0.79 Stolon numbers 1.59
July temperature 0.74 Branch crowns 1.59
Rainfall (mm) 0.95 Petal index 0.82
pH 0.71 Petal area 0.21
Salinity (ppm) 0.52 Peduncle length 1.52
% carbon 0.91 Flower number 0.40
% sand 0.82 Trichome number 0.59
% silt 0.78 Fruit index 1.19
% clay 0.96 First flower 0.86
Last flower 0.88
Flowering period 0.65
Achene weight 0.42
Fruit weight 0.05
Leaf area 1.15
Leaf index 0.55
Petiole length 1.06
Sclerophylly 1.93
generations and was able to isolate a highly self-fertile plant that was repro- ductively isolated from its parents and had a unique combination of parental traits, presumably through chromosomal translocation. More recent work uti- lizing molecular markers has indicated that DNA sequence elimination may be a major, immediate response to allopolyploidization in at least Brassica and Triticum/Aegilops, and in many cases duplicated genes may even be silenced through DNA methylation (Eckardt, 2001).
Song et al. (1995) produced reciprocal hybrids between the diploids Brassica rapaand Brassica nigra, and B. rapaand Brassica oleraceae. The F1 individuals were colchicine doubled and progenies were generated to the F5 generation by selfing. They then conducted a restriction fragment length polymorphism (RFLP) analysis of F2and F5individuals of each line, using 89 nuclear DNA probes, and found substantial genomic alterations in the F5gener- ation, including losses of parental fragments and gains of novel fragments (Fig.
4.9). Almost twice as much change was observed in the combinations involving the two most distant relatives, B. rapa and B. nigra (Table 4.9), and they observed more change in some nuclear/cytoplasmic combinations than others.
7.2 5.0 3.7 2.8 1.7 A
9.0
3.0
0.5
AB BA
A B C F2 F5-1 F5-2 F5-3 F5-4 F5-5 F5-6 F5-7 F5-8 F5-9 F2 F5-1 F5-2 F5-3 F5-4 F5-5 F5-6 F5-7 F5-8 F5-9
BA AC
A B C F2 F5-1 F5-2 F5-3 F5-4 F5-5 F5-6 F5-7 F5-8 F5-9 F2 F5-1 F5-2 F5-3 F5-4 F5-5 F5-6 F5-7 F5-8 F5-9
B
Fig. 4.9. Nuclear RFLP patterns of Brassica rapa-A, Brassica nigra-B, Brassica oleraceae-C, F2hybrids between them and F5populations. (A) HindIII-digested DNAs, probed with EZ3, which show a loss of fragments and a gain of fragments in some F5plants (5.0 kb and 2.8 kb). (B)HpaII-digested DNAs probed with EC3C8 showing a gain of a 0.5 kb fragment in five BA F5plants, which does not exist in either the A or B parental genome, but which is present in the C genome parent and all AC F5plants. (Used with permission from K. Song et al., © 1995, Proceedings of the National Academy of Sciences USA92, 7719–7723.)
In the work on wheat, Feldman et al.(1997) began by examining RFLP patterns in natural diploid and allopolyploid species. They used 16 probes that were from low-copy, non-coded DNA. Nine of these probes were found in all the diploid species, indicating that they were conserved, but, when they examined aneuploid and nullisomic lines, they found that each sequence was only retained in one of the allopolyploid genomes. In follow-up work, Liu et al., (1998a,b) examined RFLP profiles of both coding and non-coding sequences in synthetic tetraploid, hexaploid and octoploids of Triticum and Aegilops that had been selfed for three to five generations. They obtained similar results to Feldman et al. (1997), observing non-random sequence elimination in all the allopolyploids studied, along with the occasional appearance of unique fragments. They also found that some of the changes were brought about by DNA methylation. By comparing crosses with and without the PH1 gene, which regulates bivalent pairing, they were able to deduce that intergenomic recombination did not play a role in the sequence change, as both types of crosses yielded about the same amount of change.
In two further studies, the Feldman group found that the direction of sequence change in wheat followed a different pattern from that observed by Song’s group in Brassica, and confirmed that some sequences were silenced by elimination, while others were silenced through methylation.
Ozkan et al. (2001) analysed diploid parental generations, F1progeny and the first three generations (S1, S2 and S3) of synthetic hybrids of several species of Aegilopsand Triticum. When they followed the rate of elimination of eight low-copy DNA sequences, they found in contrast to Song et al.
(1995) that sequence elimination began earlier in the synthetic allopoly- ploidal whose species composition most closely represented natural occur- Table 4.9. Frequencies and types of genomic changes in F5progenies of synthetic polyploids of Brassicacompared with their parents (modified from Song et al., 1995).
Types of fragment changes
Loss/gain of fragments Fragments gained Fragments found
in F5b in F2c only in F5d
Polyploid linea A B C
AB F5 9/13 25/12 9 19
BA F5 8/12 14/0 5 51
AC F5 7/1 19/4 4 1
CA F5 15/1 16/5 3 4
a A = B. rapa, B = B. nigra, C = B. oleraceae.
bLoss = fragments present in diploid parent and the F2but not present in F5plants; gain = diploid parental fragments absent in F2plants but present in F5; A, B, and C = fragments specific to the various parents.
cFragments found in the F2but not in either parent.
dFragments found in F5plants but not in the diploids or F2s.
ring ones, and sequence elimination was not associated with cytoplasm.
Shaked et al. (2001) used amplified fragment length polymorphism (AFLP) and methylation-sensitive amplification polymorphism (MSAP) fingerprinting to evaluate another set of diploid and tetraploid hybrids within and between genera. They also found considerable sequence elimination after polyploidy and that it occurred most rapidly in allopolyploids of the same species rather than in different species. Further analysis indicated that some of the sequences were eliminated, while others were altered by cytosine methylation.
The Feldman group has suggested that the observed sequence alterations may play a physical role in how chromosomes pair, resulting in the bivalent meiotic behaviour of newly formed allopolyploids. However, sequence losses do not appear to be necessary for bivalent pairing behaviour, as Liu et al.
(1998) found little evidence of change in 22,000 AFLP loci in artificial hybrids of cotton, even though pairing in cotton tetraploids is strictly bivalent.
It is also unknown why the most divergent genomes were altered the most in Brassica, while the opposite was true in wheat. Perhaps transposons play a role, with the direction and degree of perturbations being associated with the unification of genomes with or without unique mobile elements.