Hybridization, speciation and novelty

Hybridization has been alternatively viewed as evolutionary noise (e.g. Wagner, 1970) and an engine of biodiversity (e.g. Arnold, 1997), and debate continues over the role of hybridization in evolution. In their report from last year's workshop on Hybridization and Speciation, Abbott et al. (2013) offer a fresh perspective on the evolutionary significance of hybridization by focusing on the many contexts in which hybridization may occur. Their synthesis identifies several key outcomes of hybridization and raises issues in need of further study. Here, I highlight and build upon an underlying, but often implicit, theme of Abbott et al.'s (2013) article: the role of hybridization in generating novelty. As the authors note, hybridization can affect speciational processes in multiple ways. Hybridization between diverged or diverging lineages can retard or reverse differentiation through the homogenizing effects of gene flow and recombination. Conversely, hybridization may accelerate speciation either by transferring adaptive traits via introgression, through establishment of recombinant forms (homoploid hybrid speciation) or via allopolyploidization, which may lead to instantaneous (or nearly instantaneous) speciation. Common to all of these outcomes is the origin of new genetic combinations via hybridization and their establishment in new spatio-temporal contexts; thus, genetic novelty resulting from hybridization, with or without introgression, may play a potentially pivotal role in speciation. However, the extent and form of genetic novelty will vary among these scenarios and may determine whether the outcome is homogenization, local introgression or hybrid-mediated speciation. At this point, we lack the context, theory and data to make predictions about the likely outcomes of hybridization under a set of conditions, but ongoing analysis of well-characterized hybrid systems, coupled with development of new theory, should lead to progress in this area. Evolutionary novelty in hybrids takes many forms. Most obvious, and perhaps best studied, are those associated with allopolyploidization (see Soltis & Soltis, 2009). Some changes, such as chromosome doubling per se, are instantaneous and carry nearly instantaneous consequences, some deriving from polyploidization and others from hybridization. Allopolyploids are typically immediately reproductively isolated (or largely so) from their diploid progenitors, thus leading to new forms of biodiversity. Furthermore, the immediate union of parental genomes in a single nucleus generates both novelty and conflict on an intracellular scale. Such conflict manifests itself as ‘genomic’ (McClintock, 1984) and ‘transcriptomic shock’ (e.g. Hegarty et al., 2006; Buggs et al., 2011) and by extension, perhaps proteomic shock as well. These phenomena have been examined in synthetic F1 hybrids and synthetic allopolyploids. For example, massive genome restructuring and shifts in gene expression occur in synthetic wheat (reviewed in Feldman et al., 2012; but see alternative results for synthetic wheat from different parental sources, Mestiri et al., 2010). Furthermore, in the evolution of bread wheat, which is an allohexaploid that arose within the past 10 000 years, genetic changes that prevent pairing of homeologous chromosomes (those derived from different diploid species) became fixed relatively quickly and contributed to the stabilization and establishment of the new allohexaploid. New genetic variation may arise via hybridization, introgression and allopolyploidization. Not only are alleles brought together in new combinations by these processes, but copy number variation, resulting from genomic union, may also contribute to genetic novelty. Furthermore, new alleles, generated via recombination of parental alleles (cf. Golding & Strobeck, 1983), may have unique properties that confer novel phenotypes at the biochemical, physiological or even morphological level. The origin of genetic novelty in hybrid derivatives may not be restricted to the nuclear genome. The many examples of chloroplast transfer between species (e.g. Rieseberg & Soltis, 1991) may in fact represent the outcomes of positive selection in which new combinations of chloroplast and nuclear genes have a selective advantage. The effect of alternative cyto–nuclear combinations is almost unstudied. The extent of neo- and subfunctionalization of genes duplicated by allopolyploidization is unclear, but studies are beginning to demonstrate examples of tissue-specific expression of duplicate genes in polyploid plants (e.g. Adams et al., 2003; Buggs et al., 2011). However, in some species, gene copies are lost at higher rates than they are silenced or subfunctionalized (e.g. Buggs et al., 2012). Thus, although theory predicts resolution of duplicate genes via silencing and neo- and subfunctionalization, empirical data demonstrate that polyploid genomes physically lose duplicates from the genome (e.g. Buggs et al., 2012; Feldman et al., 2012), perhaps providing a mechanism for observations of ‘genome downsizing’ (Leitch & Bennett, 2004). Furthermore, alterations at the chromosomal level may explain at least some of the observed gene losses. Intergenomic translocations, coupled with possible homeologous recombination, may yield novel nonadditive combinations of parental genes. Extensive karyotypic variation in Brassica (Xiong et al., 2011) and Tragopogon (Asteraceae; Chester et al., 2012) allopolyploids may mediate genetic changes, such as extensive gene loss, in these plants. Such karyotypic variation may ultimately lead to reproductive isolation among populations of recently formed polyploids. This hypothesis, particularly when coupled with hypotheses of reproductive isolation arising through silencing of alternative homeologs (Werth & Windham, 1991; Lynch & Conery, 2000; Lynch & Force, 2000; McGrath & Lynch, 2012), supports the concept of Abbott et al. (2013) that hybridization may yield outcomes that are not only novel themselves but that accelerate the formation of additional biodiversity (see also McGrath & Lynch, 2012). The discovery of extensive genetic, chromosomal and genomic novelty in hybrids and allopolyploids has been relatively recent. However, the concept of ecological novelty of hybrid derivatives is long standing. Classic models of hybridization and introgression propose that a hybrid or introgressant requires a novel ‘hybridized’ habitat – it cannot compete with its established parents in either of their habitats. Products of hybridization and introgression are therefore often predicted to occur (and often do; Stebbins, 1985; Bayer et al., 1991) in habitats that are intermediate between those of their parents. In some cases, however, they occur in unique habitats that extend beyond the bounds of those occupied by the parental species (e.g. arctic and alpine plants; see Brochmann et al., 2004). Hybridization and polyploidization generate novelty that may be manifested as ecological adaptations, such as the ability to occupy niches separate from those of the parental species. However, the ecological implications of hybridization on biodiversity extend beyond single species. Interspecific hybridization of plants contributes to biological diversity in fundamental ways. Beyond generating novelty through introgression, homoploid hybrid speciation and polyploidization, hybridization in plants increases overall biological diversity through its influence on the abundance and diversity of herbivores and their natural enemies, and their pathogens (Fritz et al., 1999; Whitham et al., 1999; reviewed in Soltis, 2013). As the authors note, distinguishing between those changes that accompany allopolyploidization and those that have accumulated post-polyploidization is typically very difficult. However, there may be broad categories of changes that correspond to allopolyploid formation and post-formation processes. Feldman et al. (2012) have described these as ‘revolutionary’ and ‘evolutionary’ changes respectively. ‘Revolutionary changes’ are triggered by allopolyploidization and include (i) both genetic and epigenetic changes, particularly those that improve ‘harmonic functioning’ of the divergent genomes newly combined in a single nucleus, (ii) those that lead to cytological diploidization and (iii) those that stabilize the allopolyploid and facilitate its establishment and isolation. In contrast, ‘evolutionary changes’ are facilitated by allopolyploidy and comprise mostly genetic changes that contribute to the genetic diversity of the new species and lead to adaptation. These categories are a first step towards developing a predictive framework for investigating and understanding the types of changes that arise in allopolyploids, when they occur and what their impacts may be. Abbott et al. (2013) rightly emphasize the many processes affected by and resulting from hybridization. However, one outcome of hybridization – the formation of evolutionary novelty – is emerging as ripe for investigation from many perspectives, ranging from genetic to ecosystem, yet received less explicit attention in this study than it perhaps deserves. A failure to focus on this impact of hybridization will preclude potential discoveries of the patterns and processes of diversification. A combination of new theory and data will certainly yield exciting results on the topic of hybridization and speciation in general, and of the origin of evolutionary novelty in particular.