The Evolution of Polyploidy
|
by: Sebastian Molnar Abstract Polyploidy is widespread in the plant kingdom. Therefore, some explanation is required as to how polyploids originate and evolve. Frequency-dependent processes must be overcome before rare polyploids can become established in populations of their diploid progenitors. Mechanisms and processes that are involved in the origins, establishment, and evolution of polyploids are reviewed. In particular, the minority cytotype exclusion principle, mechanisms of unreduced gamete formation, and factors involved in plant genome evolution, are discussed. Introduction Polyploidy is a characteristic feature in plants, and yet it is relatively rare in animals and fungi (Stebbins, 1950 p 366; Grant, 1971, p 224-6). There is, however, some evidence that the genomes of animals, fungi, and some plants are the products of polyploidy, but had later become diploidized (Soltis and Soltis, 1999, and references therein; Gaut et al., 2000). In the angiosperms, estimations of polyploid abundance range from 47% to 70% (Masterson, 1994). Polyploidy is also found in most pteridophytes, some bryophytes and algae, and few gymnosperms (Stebbins, 1950; Grant, 1971). Although the phenomenon is widespread across the plant kingdom, some plant groups apparently lack polyploidy altogether. In the gymnosperms, for example, no polyploids have yet been found in ginkgo or in the cycads (Grant, 1971, p 225). As such there is variation in polyploid abundance between plant families. Classification and the characteristic features of different types of polyploids have been described by Stebbins (1950, 1971) and by Grant (1971), and will therefore not be dealt with in much detail here. In general, an autopolyploid is defined as "an organism containing three or more sets of homologous chromosomes" (i.e. it has three or more chromosome sets derived from the same species), while an allopolyploid is one "containing separate sets of non-homologous chromosomes" due to hybridization between different species (Grant, 1971, p 234). Autopolyploids typically have multivalent chromosome pairing at meiosis and polysomic inheritance patterns, while allopolyploids typically exhibit bivalent pairing and disomic inheritance (Stebbins, 1971). Segmental allopolyploids are essentially intermediate forms between auto- and allo-polyploids (Stebbins, 1950). Examples of segmental allopolyploids, although apparently few in number, have been described by Stebbins (1950) and Grant (1971). Allopolyploids are generally considered to be much more common than autopolyploids, however, estimates of autopolyploid abundances may be greatly underestimated (Soltis and Soltis, 2000). Some examples of autopolyploids include Andropogon gerardii (Keeler et al., 1987; Keeler and Davis, 1999), Tolmeia menziesii (Soltis, 1984; Soltis and Riesberg, 1986), and Dactylis glomerata (Zohary and Nur, 1959; Bretagnolle and Thompson, 1996). Numerous examples of allopolyploids have been reported, such as Tragopogon (Ownbey, 1950) and Asplenium (Werth et al., 1985). Many crop species (e.g. Brassica, Coffea, Glycine, Oryza, Saccharum, Triticum, Zea) are polyploids, therefore understanding polyploid evolution could be useful in agricultural settings (Mok and Peloquin, 1975; Qu and Vorsa, 1999). An important feature of polyploidization is that entire genomes are duplicated in the process -- not just particular genes. This would potentially allow for a higher genetic diversity in a polyploid than in its diploid progenitor, since more than two alleles would be present per gene loci in the polyploid (while the diploid would have only two alleles per loci). It had been thought that chromosome doubling (especially in autopolyploids) would predominantly have detrimental effects on organismal survival and therefore must be compensated for by other processes (Stebbins, 1971, p 126). In an extensive review, however, Levin (1983) argued that although chromosome doubling does have consequences at many levels (e.g. cytological, biochemical, physiological, developmental), it can bring about adaptive changes which may cause ecological differentiation between cytotypes. Therefore, chromosome doubling by itself is not necessarily harmful. As suggested by Roose and Gottlieb (1976), polyploids have the capacity for greater physiological buffering than their diploid progenitors due to enzyme multiplicity. Increased enzyme activity, novel enzymes and metabolites, and increased metabolic regulation may enable polyploids to invade new habitats that are not occupied by their diploid progenitors. In a review, Thompson and Lumaret (1992) identified three major issues in the evolution of polyploids. These are 1) the origin of polyploids, 2) the establishment and coexistence of polyploids in diploid populations, and 3) the spread of polyploids into habitats not occupied by their diploid progenitors. Polyploids originate through some means of chromosome doubling. There are two major types of such mechanisms: somatic doubling and gamete nonreduction. How tetraploids can arise through these mechanisms from diploid progenitors have been described in some detail by Grant (1971, p 248-9). Somatic doubling is considered to be more rare than the production of unreduced gametes (Harlan and deWet, 1975), and therefore, factors that regulate or influence unreduced gamete formation may play critical roles in polyploid establishment and evolution. Although important, unreduced gamete formation is not the only factor involved in polyploid establishment. The processes that are involved in the establishment and persistence of polyploids is the primary focus of this paper. The Minority Cytotype Disadvantage A problematic and critical step in polyploid evolution is the establishment, and subsequent persistence, of the polyploid (Fowler and Levin, 1984). The predominance of one cytotype excludes other cytotypes from reaching high frequencies in a randomly mating population due to the ineffective matings of the rare cytotype (Husband, 1999). This is known as the minority cytotype exclusion principle. A new, and therefore rare, polyploid in a diploid population would be at a major fertility disadvantage, since most pollinations of the polyploids will involve pollen from diploids (i.e. rare polyploids will be less fit, since mostly sterile or inviable triploid progeny would be produced) (Fowler and Levin, 1984). A number of models have been developed to determine how polyploids may become established in parental diploid populations (e.g. Fowler and Levin, 1984; Felber, 1991; Rodriguez, 1996). Fowler and Levin (1984) compared two models, which were based on cytotype competition. Both models had essentially three possible outcomes: 1) stable equilibrium or coexistence of the two cytotypes, 2) replacement of one cytotype by the other, regardless of initial population size and cytotype frequencies, and 3) replacement of one cytotype by the other, but the outcome (i.e. whichever cytotype that is excluded) is dependent on initial population size and cytotype frequencies. Fowler and Levin (1984) suggest that niche separation, caused by a change in ploidy level, allows for the coexistence of diploids and tetraploids. Polyploids are known, in general, to have higher stress tolerances and therefore may occupy separate habitats from their diploid ancestors (Levin, 1983). Small population size would also play an important role in polyploid evolution, since chance events could allow the minority cytotype (i.e. the rare polyploid) to gain a frequency advantage necessary to replace the parental diploid cytotype (Fowler and Levin, 1984). Stochastic events, however, could also work against rare cytotypes, possibly leading to their extinction. Unlike Fowler and Levin (1984), Felber (1991) developed a model to include the production of unreduced (2n) gametes by the diploid cytotype, so that some frequency of tetraploids form with each generation. Felber (1991) found that tetraploids can be maintained at low frequencies in a diploid population if 2n gamete production by the diploids is below some threshold value (e.g. below 17.16%, in his model). Once 2n gamete production exceeds this threshold, the tetraploids are able to replace diploids. By modifying the fertility and viability of the cytotypes, this threshold value required for the spread of rare polyploids varies. If either viability or fertility of the polyploid is twice that of the diploid, the required threshold value dropped to 10% in Felber’s model. If both viability and fertility of the polyploid are twice that of the diploid, the threshold dropped even lower to 6% (Felber, 1991). The opposite effect is observed when Felber (1991) noted that a number of factors, such as genetic and environmental changes, as well as stochastic events in small populations, may influence the frequency 2n gamete production. These factors will be described in more detail in the next section. Rodriguez (1996a) has argued that the above models are too restrictive and are probably not likely to occur in reality. In particular, Rodriguez (1996a) demonstrated that conditions for the successful invasion of a rare cytotype into a diploid population is much less restricted than that found by Fowler and Levin (1984). Rodriguez (1996b) improved on an earlier model to test for the effects of hybrid viability, iteroparity, and demographic stochasticity. In the earlier model (Rodriguez, 1996a), hybrids were assumed to be infertile and inviable. In the newer model (Rodriguez, 1996b), hybrids were assumed to be infertile, but viable, and it was found that the presence of such hybrids do not change the conditions required for polyploid establishment. Also, iteroparity (where an organism can reproduce more than once in a life cycle) may enhance polyploid establishment over semelparity (where an organism reproduces only once) (Rodriguez, 1996b). This would support the finding that polyploidy is more common in perennials than in annuals (Stebbins, 1950). Finally, whether demographic stochasticity hinders polyploid establishment or not, really depends on the conditions of the model (Rodriguez, 1996b). This raises the issue as to how useful such models are in reality. Models may contain underlying assumptions and conditions that are not necessarily present in nature. On the other hand, they can be used to devise theoretical principles which may be tested experimentally in greenhouse or field populations; and vice versa, experimental observations may be better clarified by testing models. One such experimental test was done by Husband (2000) using populations of diploid and tetraploid plants of Chamerion angustifolium at varying proportions. The conditions for frequency-dependent effects on fitness (i.e. seed set for one season) in this species was found to hold true for diploids, but not for tetraploids (Husband, 2000). Husband (2000) concluded that the frequency-independent effects in tetraploids (i.e. seed set was not dependent on tetraploid frequency) was due to assortative mating patterns caused by non-random pollinator behaviour. Differences in diploid and tetraploid flowering times was excluded as a possible reason for the observed frequency-independent effects, since the entire tetraploid flowering time period occurred within the diploid period (Husband, 2000). Other potential factors causing non-random mating patterns had not been tested for, however, this study demonstrates that frequency-dependent processes can be overcome (Husband, 2000). By reducing inter-cytotype matings, rare cytotypes could become established despite the minority cytotype disadvantage. Assortative mating (either by differences in flowering time or by non-random pollinator visitation) is one means through which rare polyploids can contribute seeds to the next generation. Selfing, apomixis, and vegetative propagation could also enable rare polyploids to become more frequent. Inbreeding depression would be one potential difficulty faced by rare polyploids with high selfing rates. It has, however, been demonstrated in certain species that inbreeding depression can be less severe in tetraploids than in diploids (Husband and Schemske, 1997). Thus, there are several possible ways in which polyploids may become established. The next section will deal with the factors that influence unreduced gamete production, as this could be an important mechanism for generating polyploids. Factors Influencing Unreduced Gamete Formation Although the formation of unreduced (2n) gametes is considered to be rare in general (McCoy, 1982), 2n gamete production is likely to play a major role in polyploid origins (Harlan and deWet, 1975; Vorsa and Bingham, 1979). A number of factors -- genetic and environmental -- have been shown to influence the frequency of 2n gamete formation (Sax, 1937; Thompson and Lumaret, 1991). For example, temperature has been demonstrated to increase 2n pollen formation (Sax, 1937; McHale, 1983). In Tradescantia, plants that had been grown at a relatively low temperature (18 C) and then transferred to an environment with a higher temperature (38 C), showed meiotic irregularities (e.g. spindle disruption and failure of segregation) in pollen mother cells (Sax, 1937). This sudden change in temperature somehow prevents the nuclear division of a single pollen mother cell and can lead to the formation of two diploid pollen grains (e.g. dyads), each of which are larger than normal haploid pollen grains (Sax, 1937). Low temperature has been shown to have an effect on the formation diploid pollen grains in other species (McHale, 1983). In Solanum phureja, plants that had been grown in cool environments consistently produced a higher frequency of dyads and large pollen grains than plants grown in a warmer environment (McHale, 1983). The effects of temperature in producing diploid pollen grains has been demonstrated in a number of different plant species (Sax, 1937). Similarly, dehydration has been shown to inhibit spindle formation, therefore preventing normal segregation during meiosis, and it can also inhibit cell division, generating bi- and quadri-nucleate cells (Giles, 1939). Other factors that can induce chromosome doubling, besides temperature and the degree of hydration, include X-rays, ultraviolet radiation, mechanical injury, certain chemicals, infection by certain viruses or mites, and genetics (Sax, 1937). Many of these factors appear to have similar effects, as they disrupt the synchronization of nuclear and cellular division (Giles, 1939). Under natural conditions, sudden changes in temperature, water abundance, and (now, with the ever decreasing ozone layer) UV radiation, can be expected to be the major environmental factors which could influence the population dynamics of polyploids through unreduced gamete formation. Genetics is another major factor in unreduced gamete formation. Genetic control of 2n gamete production has been demonstrated in a few species (Mok and Peloquin, 1975; Veilleux et al., 1982; McCoy, 1982; Qu and Vorsa, 1999). In Solanum, three mechanisms (one for parallel spindle formation and two for premature cytokinesis) of 2n pollen (or “diplandroid”) formation were found to be each controlled by simple recessive genes, although the frequency of diplandroid production had wide-ranging variability (Mok and Peloquin, 1975). Certain genotypic clones of potato (Veilleux et al., 1982), alfalfa (McCoy, 1982), and blueberry (Qu and Vorsa, 1999), have been observed to consistently produce high frequencies of 2n pollen, while others are more variable. Genetic controls of unreduced gamete production is important to polyploid evolution, since such genes could become fixed in populations (especially, in small populations) which would enable rare polyploids to become more frequent and possibly overcome the minority cytotype disadvantages in diploid populations. The production of unreduced gametes is only one factor in the establishment of a new cytotype. In order to establish and persist, a polyploid must also be competitive (Harlan and deWet, 1975). For example, superior vegetative growth would allow a rare cytotype to outcompete and even replace its diploid progenitor (Harlan and deWet, 1975). On the other hand, it could work the other way around with the diploid progenitor having the higher competitive ability and excluding the rare cytotype from becoming more frequent. As mentioned above, niche separation, assortative mating, stochastic variation, and fitness differences between cytotypes, also play roles in polyploid evolution. Many of these factors need not be mutually exclusive, and several of them probably work in combination at any one time. Therefore, broad generalizations may not apply to specific cases. As such, different species need to be characterized and systematically analysed to determine what mechanisms are responsible for bringing about the observed polyploid frequencies and subsequent evolutionary patterns. Multiple Origins of Polyploids and Plant Genome Evolution It had been previously thought that polyploids arose through a single origin and that they were evolutionary dead-ends (reviewed in Soltis and Soltis, 1995). There is, however, much evidence for the recurrent formation of polyploids in different geographic locations (Werth et al., 1985; Doyle et al., 1999; reviewed in Soltis and Soltis, 1999, 2000). One well studied example of a recent speciation event (i.e. Tragopogon sp.) may provide some insight into the dynamics of polyploid evolution. Two Tragopogon species of recent origin have been shown to have arisen on several occasions (Soltis and Soltis, 1999). Three Old World parental diploid species (T. dubius, T. porrifolis, and T. pratensis) had been introduced to North America early on in the 1900s, and hybrid plants derived from these diploid progenitors were discovered in Washington and in Idaho by 1949 (Ownbey, 1950). Morphological and cytological evidence (Ownbey, 1950), followed up by studies in isozyme variation (Roose and Gottlieb, 1976) and RFLPs in chloroplast DNA (Soltis and Soltis, 1989), has shown and confirmed that the tetraploid T. mirus was derived from the diploid parents T. dubius and T. porrifolius, while that of T. miscellus was derived from T. dubius and T. pratensis. Early investigations suggested that T. mirus and T. miscellus had two or three independent origins (Ownbey, 1950), but the range and number of the Tragopogon populations in North America have increased since then, and it is now thought T. mirus and T. miscellus have arisen independently at least 12 and 21 times, respectively (Soltis et al., 1995; Soltis and Soltis, 1999). Soltis et al. (1995) sampled various populations of Tragopogon species for isozyme and DNA variation at a number of sites in Washington and Idaho. Tragopon dubius apparently has more genetic diversity, and is more widespread, than either T. porrifolius or T. pratensis in North America, which suggests T. dubius was probably introduced many more times from different source populations than were the other two species (Soltis et al., 1995). This type of variability could give rise to different genotypic races of the derived tetraploids in different populations. Some of this variation may simply be due to the long-distance dispersal of a particular genotype accompained by divergence after colonization (Soltis and Soltis, 1995). However, a number of distinct genotypes were found, and these were used to compare variation between different populations -- it seems likely that the Tragopogon allotetraploids arose on numerous separate occasions (Soltis et al., 1995). Multiple polyploid origins could provide a wide range of genetic diversity across separate geographically isolated populations. Another source of genetic variation is the subsequent divergence after a polyploid originates. There is a controversy over whether or not newly formed polyploids undergo rapid changes in genome structure and in gene expression (Matzke et al., 1999). A recent study, which used synthetic Brassica polyploids to study genome evolution, suggested that rapid genetic change is possible soon after formation (Song et al., 1995). Extensive genomic changes were observed within five generations of these synthetic polyploids (Song et al., 1995). Other studies have demonstrated that, although duplicated genes do evolve independently (Cronn et al., 1999), large and rapid genomic changes do not necessarily occur (Axelsson et al., 2000). In some cases, the rate of gene divergence in polyploids versus their diploid progenitors may not be significantly different (Cronn et al., 1999). The answer probably lies somewhere in the middle: whether subsequent evolutionary change is gradual or rapid may really depend on the particular species and on the particular ecological context. Wendel (2000) has reviewed the mechanisms and processes that are involved in the evolution of polyploid genes and genomes. Evolutionary fates of duplicated genes have been categorized into three major categories: 1) functional divergence, 2) loss or inactivation (e.g. pseudogene formation, epigenetic silencing), and 3) long term maintenance. To this list, Wendel (2000) adds a forth possible fate: novel interactions among duplicated genes. On the whole, several major processes have been recognized that are involved in the evolution of plant genomes and of polyploids, but many aspects remain poorly understood (Thompson and Lumaret, 1992; Soltis and Soltis, 1999; Wendel, 2000). Whether the recent sequencing of the Arabidopsis thaliana genome (Dennis and Surridge, 2000) will provide any insight into plant genome evolution remains to be seen, but it will certainly need to be compared to other plant genomes to be meaningful. In summary, three major, and inter-related, aspects of polyploid evolution were discussed. A number of factors can influence the formation of unreduced gametes, and thereby increase the number of polyploids (of a particular cytotype) in a population. Increasing the number of cytotypes above some threshold level can allow polyploids to overcome frequency dependent processes that would otherwise prevent cytotype establishment. Polyploidy is a rapid speciation mechanism, and polyploids may also originate several times independently due to similar processes working in different populations. Polyploid Evolution: References
|
Get your own Free Home Page