Sunflower genetic, genomic and ecological resources

Abstract

Long a major focus of genetic research and breeding, sunflowers (Helianthus) are emerging as an increasingly important experimental system for ecological and evolutionary studies. Here, we review the various attributes of wild and domesticated sunflowers that make them valuable for ecological experimentation and describe the numerous publicly available resources that have enabled rapid advances in ecological and evolutionary genetics. Resources include seed collections available from germplasm centres at the USDA and INRA, genomic and EST sequences, mapping populations, genetic markers, genetic and physical maps and other forward- and reverse-genetic tools. We also discuss some of the key evolutionary, genetic and ecological questions being addressed in sunflowers, as well as gaps in our knowledge and promising areas for future research.

Introduction

Iconic symbols in myth, art, politics and religion, sunflowers represent solar deities, power, nuclear nonproliferation, longevity and mortality. Over the past several decades, Helianthus has also emerged as an excellent experimental system for studying the ecological genetics of speciation, species boundaries, hybridization and domestication. With growing genomic resources, extensive publicly available seed collections, a rapidly developing genetic tool kit, important economic impacts and fascinating ecology, it is an ideal taxon for many ecological and evolutionary questions.

One of the core strengths of the system is the tremendous variation found within the genus. The diversity of speciation mechanisms and barriers to gene flow are truly remarkable, making it ideal for understanding speciation and divergence from many angles. The 49 named sunflower species, native to diverse habitats throughout most of North America (Seiler & Rieseberg 1997), include examples of allo- and autopolyploids (Timme et al. 2007), ecologically isolated sympatric and allopatric species (Heiser et al. 1969), karyotypically divergent species (Chandler et al. 1986; Burke et al. 2004; Lai et al. 2005b), allopatric species with weak barriers to gene flow other than geography (Heiser et al. 1969) and several homoploid hybrid species (Rieseberg 1991). This variation has made Helianthus a model system for studying speciation (e.g. Rieseberg et al. 1995).

Key aspects of within- and among-species variation have been harnessed during domestication and improvement of several sunflower species, most prominently the Jerusalem artichoke (H. tuberosus), which was domesticated for its tuber, and the common sunflower H. annuus, cultivated worldwide for edible oil, edible seeds and the cut flower industry. Sunflower-breeding programmes have benefited from the introgression of wild germplasm from numerous annual species as well as from perennial species such as H. tuberosus and H. giganteus (Škorić 1992). Wild germplasm contains numerous ecologically important traits that can be useful in cultivation, including disease resistance (Liu et al. 2010), drought tolerance (Seiler et al. 2006) and cytoplasmic male sterility (CMS) and restoration (Jan 2000). New breeding efforts, of which many involve introgression of wild material, promise to substantially improve drought and salt tolerance in cultivated sunflowers and also may lead to cultivars that produce substantial cellulosic biofuel as a byproduct of cultivation for food. Also of economic importance are several weedy and invasive sunflowers (Muller et al. 2011), again including varieties of H. annuus as well as several other annual and perennial species that have escaped cultivation, been accidentally introduced to new habitats or evolved into noxious weeds from wild populations. Not only has weediness evolved in numerous sunflower species, novel weedy traits have also evolved within species multiple times (Kane & Rieseberg 2008; Muller et al. 2011). Herbicide tolerance is one such trait of Helianthus weeds, which makes them difficult to manage agricultural pest (Burton et al. 2004). Ironically, alleles conferring herbicide tolerance in wild sunflowers can be used to our advantage by introgressing them into the domesticated sunflower for weed control purposes. Clearly, for this and many other traits, a better understanding of basic ecology, evolution and genetics goes hand in hand with agricultural and other applied purposes, making sunflower an interesting system on numerous levels and broadening funding opportunities beyond those available for many ecologically and evolutionarily important wild species.

Here, we briefly review what is currently known about the phylogeny, natural history and genome evolution in Helianthus; we also describe the genetic and genomic resources available for the system and discuss several recent findings from analyses of the new genomic data sets.

Geographical range

Helianthus is indigenous to North America. Although the native ranges of most species are restricted to the continental United States, several extend into northern Mexico or southern Canada (Heiser et al. 1969). In addition to their indigenous distributions, numerous Helianthus species have become naturalized elsewhere in the world as a consequence of both intentional and inadvertent introductions by humans. Indeed, 22 Helianthus taxa are considered naturalized or invasive in Europe (Rehorek 1997), and sunflowers are abundant in parts of southern South America (Cantamutto et al. 2010) and southern and western Australia (Seiler et al. 2008).

Phylogeny

Because Helianthus is a recently evolved and species-rich group with a history of hybridization and polyploidy, phylogenetic reconstruction has proven to be challenging, especially for the perennial species (Schilling & Heiser 1981; Schilling 1997; Schilling et al. 1998; Timme et al. 2007). Combining extensive crossability information (Heiser et al. 1969) with morphological characters for 49 species led to important insights, but crossability information is of limited utility in the perennials because of polyploidy (Schilling & Heiser 1981). The addition of genetic data from ribosomal genes clarified the relationships among most of the annual species, including the reticulation events forming the homoploid hybrid species (Rieseberg 1991). Chloroplast restriction fragments (Schilling 1997) and nuclear ribosomal internal transcribed spacer (ITS) sequence data (Schilling et al. 1998) shed light on relationships among the major lineages within the genus and also clarified the relationships between Helianthus and related genera (Schilling 2001), but failed to resolve finer-scale relationships among species due to paucity of informative characters. The phylogeny for Helianthus with the best resolution and support was generated by analysing sequence data from the external transcribed spacer of the 18S–25S nuclear ribosomal DNA region (Timme et al. 2007). This phylogeny successfully resolved relationships among the perennial species for the first time, identified the parentage of several hybrid and polyploid lineages and showed that an annual life history has evolved multiple times within the genus. Nonetheless, relationships among the basal lineages in the genus were poorly supported and discordant with phylogenetic analyses based on restriction site data from chloroplast DNA (Schilling 2001) and sequence data from the ITS of nuclear ribosomal DNA (Schilling et al. 1998).

To visualize the most well-understood relationships among species in the genus, we show relationships among five sections within the genus (Fig. 1a), as well as more detailed species relationships within the more widely studied annual clade section Helianthus (Fig. 1b). In Fig. 1a, we present a hypothetical phylogeny summarizing broad relationships of clades within the genus supported by multiple studies (Schilling & Heiser 1981; Rieseberg 1991; Schilling 1997, 2001; Schilling et al. 1998; Timme et al. 2007). The phylogeny presented in Fig. 1a differs from the traditional sectional classification of the genus, with sections Agrestis and Helianthus apparently monophyletic but sections Ciliares and Divaricati polyphyletic. Within section Ciliares, however, Series Ciliares (here Sect. Ciliares clade A) and Pumili (here Sect. Ciliares clade B) appear to be monophyletic. The series within Divaricati are more difficult to delineate, and the clades identified by molecular phylogenetic analyses do not correspond to any previous treatment. Species within each section are listed in Table 1, which also presents information on genome size, ploidy, life history and publicly available resources for each species as described in the next section.

Figure 1. Phylogenetic trees for Helianthus. (a) Phylogenetic tree for sections of the genus based on sequence analysis of the external transcribed spacer of nuclear ribosomal DNA (simplified from Timme et al. 2007). Numbers of species in each clade are given in parentheses following the section name. Note that sections Ciliares and Divaricatus are polyphyletic. (b) Phylogenetic network for section Helianthus based on inferences from nuclear ribosomal DNA analyses (Rieseberg 1991; Timme et al. 2007) and 11 single-copy nuclear genes (Moody & Rieseberg 2012). Putative hybrid speciation events are indicated by dashed lines.

Table 1. Publicly available lines from the USDA and INRA, EST and genomic sequence data available from NCBI, characterized wild populations with information at the USDA, ploidy, chromosome numbers, genome size, life history and section for wild and domesticated sunflower species and hybrids. The ‘habitat data’ column lists the number of historical and current accessions for which habitat location and other data exist, regardless of whether the accession is still maintained. Letters following sections refer to the clades in Fig. 1

Taxon	USDA accessions	Habitat data	INRA accessions	NCBI nucleotide sequences	NCBI ESTs	Ploidy	Chromosome number	Genome size (Mb)	Life history	Section
Helianthus agrestis	9	11	1	5	0	2	34	12 691	Annual	Agrestis
H. angustifolius	22	28	1	5	0	2	34	5978	Perennial	Divaricati C
H. annuus cultivars	1867	NA	421	8189	10 5684	2	34	3528	Annual	Helianthus
H. annuus wild	930	1061	267	5908	28 014	2	34	3528	Annual	Helianthus
H. anomalus	6	14	3	194	0	2	34	5488	Annual	Helianthus
H. argophyllus	49	57	26	663	35 720	2	34	4336.5	Annual	Helianthus
H. arizonensis	2	5	2	5	0	2	34		Perennial	Ciliares B
H. atrorubens	14	21	2	7	0	2	34		Perennial	Divaricati D
H. bolanderi	7	14	5	4	0	2	34	4312	Annual	Helianthus
H. californicus	21	22	3	1	0	6	102		Perennial	Divaricati A
H. carnosus	2	3	0	3	0	2	34		Perennial	Divaricati C
H. ciliaris	26	27	2	7	21 590	4	68		Perennial	Ciliares B
H. cusickii	20	23	1	2	0	2	34		Perennial	Ciliares A
H. debilis	53	71	18	29	0	2	34	3577	Annual	Helianthus
H. decapetalus	30	32	4	6	0	2	34		Perennial	Divaricati D
H. deserticola	21	26	0	264	0	2	34	5635	Annual	Helianthus
H. divaricatus	26	45	3	5	0	2	34	8281	Perennial	Divaricati D
H. eggertii	12	15	2	7	0	6	102		Perennial	Divaricati C
H. exilis	30	35	2	4	33 961	2	34	4704	Annual	Helianthus
H. floridanus	8	13	1	2	0	2	34		Perennial	Divaricati C
H. giganteus	25	34	5	12	0	2	34	4728.5	Perennial	Divaricati D
H. glaucophyllus	11	13	2	3	0	2	34		Perennial	Divaricati D
H. gracilentus	6	11	1	2	0	2	34		Perennial	Ciliares A
H. grosseserratus	44	57	6	4	0	2	34		Perennial	Divaricati D
H. heterophyllus	17	26	0	8	0	2	34	4802	Perennial	Divaricati B
H. hirsutus	12	26	3	4	0	4	68		Perennial	Divaricati C
H. laciniatus	7	9	1	3	0	2	34		Perennial	Ciliares B
H. laevigatus	7	11	4	3	0	4	68		Perennial	Divaricati C
H. longifolius	3	4	0	3	0	2	34		Perennial	Divaricati D
H. maximilianii	64	91	23	7	0	2	34		Perennial	Divaricati D
H. microcephalus	13	21	2	4	0	2	34	4998	Perennial	Divaricati D
H. mollis	27	39	6	3	0	2	34		Perennial	Divaricati D
H. neglectus	28	29	3	132	0	2	34	3136	Annual	Helianthus
H. niveus	30	40	3	5	0	2	34	3577	Annual	Helianthus
H. nuttallii	42	46	20	4	0	2	34		Perennial	Divaricati D
H. occidentalis	15	21	2	4	0	2	34		Perennial	Divaricati D
H. paradoxus	2	10	2	268	30 517	2	34	5243	Annual	Helianthus
H. pauciflorus	46	59	3	3	0	6	102		Perennial	Divaricati D
H. petiolaris	139	185	23	1896	27 484	2	34	3332	Annual	Helianthus
H. porteri	8	9	1	7	0	2	34		Annual	Divaricati A
H. praecox	41	43	15	11	0	2	34	3454.5	Annual	Helianthus
H. pumilus	52	55	2	2	0	2	34		Perennial	Ciliares A
H. radula	37	47	1	3	0	2	34	5757.5	Perennial	Divaricati B
H. resinosus	23	31	4	3	0	6	102		Perennial	Divaricati D
H. salicifolius	19	24	3	5	0	2	34		Perennial	Divaricati D
H. schweinitzii	1	2	1	3	0	4	68		Perennial	Divaricati C
H. silphioides	15	21	2	2	0	2	34		Perennial	Divaricati D
H. simulans	4	4	1	7	0	2	34		Perennial	Divaricati C
H. smithii	6	10	2	1	0	2	34		Perennial	Divaricati D
H. strumosus	33	45	14	2	0	6	102		Perennial	Divaricati D
H. tuberosus	90	112	21	205	40 362	6	102	12 299	Perennial	Divaricati D
H. verticillatus	2	2	0	5	0	2	34		Perennial	Divaricati D
H. x laetiflorus	11	11	2	0	0	6	102		Perennial	Divaricati D
Helianthus hybrid	17	13	0	11	0	2	34		Annual	Various
All Helianthus	4052	2684	942	17 945	32 3332

Figure 1b presents a more detailed phylogeny for the annual section Helianthus, including the common sunflower H. annuus, the plains sunflower H. petiolaris and their most closely related wild relatives. The two main lineages of annual sunflowers, those species related to H. petiolaris and those related to H. annuus, are well supported and distinct, although hybridization between the two lineages has given rise to at least three homoploid hybrid species, H. paradoxus, H. anomalus and H. deserticola. Relationships within the two lineages, however, are not uncontroversial, and some species delineations are not entirely clear. In particular, H. debilis is probably not monophyletic and may have to be split or combined with H. praecox (Timme et al. 2007), and the sister species pair H. bolanderi and H. exilis is often combined into a single species despite the strong differences in ecology and morphology (e.g. Heiser et al. 1969). A future goal should be to use the extensive EST data generated for Helianthus to develop a more accurate and highly supported phylogeny for the genus. Such a phylogeny could provide a framework for documenting the extent and timing of reticulate evolution in the genus, determining the genomic composition of hybrid and introgressed lineages, identifying sister taxa for studies of speciation and estimating the number and independence of ecological and evolutionary transitions within the genus.

Ecology and breeding system

Sunflowers occupy a broad range of habitats throughout their native and introduced ranges. The majority of species occupy open habitats such as disturbed areas, grasslands and deserts, but a few species do well in forest edges (H. divaricatus and H. decapetalus) or even grow as woodland understorey species (H. microcephalus and H. radula) and in seasonal bogs or marshes (H. heterophyllus and H. paradoxus). Several species are known for extreme abiotic stress tolerance, particularly the halophyte H. paradoxus, xeric species H. deserticola, dune-adapted species such as H. anomalus and H. neglectus and the serpentine soil specialist H. exilis. Because multiple lineages of sunflowers have been introduced into non-native habitats, and because some species have been introduced into numerous locations independently, sunflowers are also emerging as a key model system for understanding the genomics and ecology of range expansions and the evolution of invasive taxa. Fascinating work has been done to document and explain invasive range expansion across the Argentine landscape (Cantamutto et al. 2010) and to understand the ecological conditions favouring invasion by particular Helianthus species (Cantamutto et al. 2008). Genetic and morphological work indicates that non-native populations have had multiple origins and that source populations probably include wild natives and crop-wild hybrids (Poverene & Cantamutto 2010; Muller et al. 2011). An emerging pattern is that range expansions in Helianthus are frequently associated with introgression from divergent lineages. This has been seen in the historical expansion of H. annuus into Texas, which involved introgression of herbivory tolerance and other traits (Heiser 1951; Whitney et al. 2006, 2010; Scascitelli et al. 2010) and California (Heiser 1949; Rieseberg et al. 1988; Carney et al. 2000), which involved high levels of interspecific hybridization and possibly adaptive introgression. Similarly, invasions onto other continents are known to have high levels of hybridization among species (Gutierrez et al. 2011) and between wild H. annuus and domesticated cultivars (Poverene & Cantamutto 2010; Muller et al. 2011; Lai et al. 2012). Whether this pattern is truly general, and whether this introgression is adaptive or due to demography or other neutral causes, remains a key question to be addressed in this system.

Most sunflowers are obligate outcrossers as a consequence of a sporophytic self-incompatibility system. Exceptions include H. agrestis, an annual species found in Florida, and all widely cultivated varieties of domesticated H. annuus, which are self-compatible (Heiser et al. 1969). Both wild and domesticated sunflowers are pollinated by a diverse array of wild solitary bees, as well as domesticated honey bees (Neff & Simpson 1991; Sapir 2009). CMS segregates in the wild due to cytonuclear interactions (Serieys & Vincourt 1987; Rieseberg et al. 1994) and possibly plays a role in ecological adaptation and selection (Sambatti et al. 2008). In domesticated sunflowers, CMS was obtained for breeding and seed production purposes from hybridization between H. petiolaris and H. annuus by P. Leclercq (Leclercq 1969). Genes giving restoration of male fertility in the presence of this cytoplasm were obtained from both wild H. annuus (Kinman 1970) and the H. petiolaris accession which provided CMS (Leclercq 1971). Although most cultivated lines currently use CMS from this H. petiolaris material, over 60 more sources of CMS and 30 more sources of restorer alleles have been identified (see references in Jan 2000), suggestive of the large amount of diversity and the prevalence of CMS in wild Helianthus. As the specific genetic factors underlying CMS and restoration are identified and characterized, sunflower will probably become an important model for understanding cytonuclear interactions and their role in promoting or inhibiting speciation and interspecific gene flow. The advent of a reference nuclear and mitochondrial genome for this group as well as the many other resources being developed will surely promote this and related work.

Hybridization

Where species co-occur, hybridization leads to gene flow in many (but not all) cases, with widely varying outcomes. In some cases, gene flow is so low as to be virtually nonexistent, but in others it is quite high. Helianthus annuus ssp. texanus appears to have arisen through introgression from H. debilis, which apparently facilitated the southward expansion of H. annuus in Texas (Heiser 1951; Whitney et al. 2006, 2010). Rates of gene flow between these species are between four and seven migrants per generation (Scascitelli et al. 2010), but H. debilis remains strongly isolated and morphologically distinct from H. annuus and other sympatric species. In contrast, rampant hybridization has occurred between H. annuus and H. bolanderi since the introduction of H. annuus into California (Heiser 1949; Rieseberg et al. 1988), with the result that many populations of H. bolanderi are predominantly hybrids. Interestingly, the closely related H. exilis, sister to H. bolanderi and also native to California, has not shared this fate, perhaps due to its strong ecological isolation (Carney et al. 2000): H. exilis thrives on serpentine soil, rarely coming into contact with H. annuus or other congeners. This suggests that the unique ecology of H. exilis may be a barrier to gene exchange, underlining the importance of ecology in promoting and maintaining isolation among diverging sunflower lineages.

Perhaps the best-studied hybridizing species are H. annuus and H. petiolaris, which co-occur over much of their broad ranges (Heiser 1947). Both species span the continental United States, from the Atlantic to Pacific and from Mexico to Canada. The primary ecological difference between the species appears to be that H. annuus prefers mesic clay soils while H. petiolaris thrives in drier sandy soils. These soil types often occur in close proximity, where the species can hybridize, resulting in low but measurable gene flow as a result, both at chloroplast (Dorado et al. 1992) and nuclear markers (Yatabe et al. 2007; Strasburg & Rieseberg 2008; Kane et al. 2009). These hybrid zones tend to be quite restricted, because the hybrids have extremely low levels of fertility (<1% viable seed), but because of the high effective population sizes of the two species, effective long-term levels of migration have been estimated as approximately 0.5 migrants per generation in each direction (Strasburg & Rieseberg 2008). Because of this low level of gene flow, divergence due to drift may be impeded, and advantageous or neutral alleles may move between the two species, but divergence will be high near loci where genetic background, ecology or other factors select against introgression.

Speciation

Evolutionary dead ends and low levels of gene flow are not the only results of hybridization. In several cases, hybridization between H. annuus and H. petiolaris has resulted in the origins of novel lineages of extremophiles (Rieseberg et al. 1990; Rieseberg 1991; Schwarzbach & Rieseberg 2002). Interestingly, these hybrid species (H. anomalus, H. deserticola and H. paradoxus) all evolved without polyploidization, a process known as homoploid hybrid speciation. Because the origins of these species can in some ways be mimicked in the laboratory and the field by crossing the parental species, these speciation events have a particular fascination. Indeed, even without imposing ecological selection, crosses derived from H. annuus × petiolaris hybrids rapidly converge on a genetic make-up similar enough to the three natural homoploid hybrid sunflower species that they are largely interfertile with all three hybrid species and highly infertile with the parental species (Rieseberg 2000; Rieseberg et al. 2003; Lai et al. 2005b). Re-playing of evolutionary history is repeatable in only some respects: while hybrids from the laboratory and nature are interfertile and appear to have similar genetic make-up, their ecological niches are entirely different (Welch & Rieseberg 2002a,b; Rosenthal et al. 2005). Field experiments examining performance of synthetic hybrids in the homoploid hybrids' native environments have shown that the range of phenotypes in the hybrids exceeds that of the parents, to the extent that transgressive segregation may lead to some individuals with superior stress tolerance compared to either parent (Gross et al. 2004; Ludwig et al. 2004). However, because some aspects of the homoploids' genetic make-up and ecology were not re-created in these synthetic hybrids, it has been suggested that ecologically mediated selection throughout the process of homoploid hybrid speciation may be important (Karrenberg et al. 2007). Clearly, these homoploid hybrids are a system with many fascinating secrets yet to be revealed.

Adaptation and domestication

A new focus of research is on understanding phenotypic variation within and among sunflower species. This includes the study of local adaptation and the evolution of ecotypes within species, which may sometimes be the beginnings of a more long-term divergence (Kane & Rieseberg 2007). For instance, substantial variation in flowering phenology exists within H. annuus (Blackman et al. 2011a,b). Unlike other well-studied species, control of flowering time varies substantially in sunflowers, including short-day, long-day and day-neutral photoperiod responses (Blackman et al. 2011a,b). In the case of dune-adapted H. petiolaris in Colorado, differences in flowering time may be a key cause of reproductive isolation from nearby nondune populations (R. Andrew, personal communication).

Some of the most interesting within-species variation is in cultivated sunflower species, particularly H. annuus. The domesticated common sunflower has been cultivated for a variety of purposes and is highly diverse in terms of phenology and morphology. The genetics of sunflower domestication are becoming increasingly well understood (Burke et al. 2002, 2005; Baack et al. 2008), with a single cultivated lineage clearly originating in what is now the east–central United States (Harter et al. 2004; Wills & Burke 2006; Blackman et al. 2011a,b) and domestication having involved many genetic changes of small effect (Burke et al. 2002; Gandhi et al. 2005; Wills & Burke 2007). Some of the genes underlying important differences in phenotypes and phenology are also beginning to be understood, which has further supported our understanding of domestication and improvement (Blackman et al. 2010, 2011a,b). Additionally, characterization of the genes underlying iconic traits in mutated ornamental varieties has given insights into the genetics of floral patterning and inflorescence architecture and development (Chapman et al. 2012). Gene duplication and specialization of paralogous copies has been an important part of the evolution of inflorescence architecture (Chapman et al. 2008a, 2012; Thtiharju et al. 2012). As the genes underlying the regulation of development become increasingly well characterized, we will gain a better understanding of evolution of the compound inflorescence that defines the sunflower family (Asteraceae).

Genome evolution

Analysis of the age distribution of duplicate genes indicates that, while genetically diploid, Helianthus has experienced at least three episodes of polyploidy over the past 60 Myr (Barker et al. 2008; J. Bowers, personal communication). Two of these episodes appear to have occurred near the origin of the Compositae family, whereas the third duplication is coincident with an approximate doubling of chromosome numbers at the base of the sunflower tribe, Heliantheae. However, most of the genes that were duplicated during these ancient whole-genome duplications have been silenced or lost. As a consequence, the 3.6 Gb H. annuus genome is dominated by repetitive elements; a recent analysis of more than 100 Gb of whole-genome shotgun sequence indicates that 78.5% of the genome consists of repeats, mainly LTR retrotransposons (Kane et al. 2011).

There is interesting variation in genome size among Helianthus species with a fourfold size difference across the diploids. In general, the perennial species have larger genomes than most of the annuals (Sims & Price 1985; and see Table 1). Among the annual species of section Helianthus, there is a roughly 50% increase in genome size in the three homoploid hybrid species (H. anomalus, H. deserticola, H. paradoxus). Ungerer et al. (2006) showed that large-scale expansions of LTR retrotransposons are responsible for the genome size increases in the three hybrid lineages. Because this genome expansion occurred at least three times independently from hybridization from the same parental species, these lineages represent an ideal system to study what causes the activation and inactivation of retrotransposons, and what effects such expansions in retrotransposon copy number have on fitness, phenotypes and genome architecture. A baseline for this work has been set by recent efforts to characterize in detail the repetitive content in H. annuus (Staton et al. 2012).

Helianthus is also well known for its high rate of karyotypic evolution, and closely related species typically differ by multiple chromosomal translocations and inversions (Heiser 1947; Chandler et al. 1986; Burke et al. 2004; Heesacker et al. 2009). The most striking example of rapid karyotypic change comes from comparative mapping studies of the three homoploid hybrid species and their parental species, H. annuus and H. petiolaris (Rieseberg et al. 1995; Lai et al. 2005b). Not only are the parental species highly divergent in karyotype, differing by at least eight translocations and three inversions, but the genomes of the three hybrid species are comprehensively restructured relative to their parents. Interestingly, most QTLs underlying hybrid pollen sterility in sunflowers map to chromosomal breakpoints (Lai et al. 2005a), and analyses of patterns of introgression across hybrid zones in nature suggest that interspecific gene flow is reduced in these genomic regions (Yatabe et al. 2007; Strasburg et al. 2009). However, marker densities on these maps are too low to detect micro-chromosomal rearrangements, and the location of chromosomal breakpoints cannot be inferred precisely. Thus, a more complete understanding of karyotypic evolution in Helianthus and its role in speciation will require high-density maps and genome-wide analyses of sequence divergence, both of which are underway. Because the annual sunflowers, in particular, co-occur in numerous distinct hybrid zones, and because these hybrid zones involve different species pairs in different locations, Helianthus represents an ideal model system for understanding the relationship between chromosomal evolution and speciation.

Germplasm and genomic resources

Supporting this extensive historical and ongoing research are numerous well-maintained, publicly available lines that can be obtained from germplasm centres in the United States (Marek et al. 2008, 2012) and elsewhere. A total of 2797 accessions of H. annuus and 1344 accessions of other species (Table 1), collected from native, cultivated and invasive populations from around the globe, are catalogued at the USDA's sunflower genebank in Ames, Iowa (http://www.ars-grin.gov/cgi-bin/npgs/html/site_holding.pl?NC7). A partially overlapping set of 521 wild lines and 421 cultivated accessions (including 331 old open-pollinated varieties and a core collection of 96 inbred lines) are also available from INRA's seed centres in France (Table 1, Table S1). In addition to making seed available, the USDA's website has an extensive database of measured traits, including morphology, phenology, chemistry (particularly seed oil characters) and genotypic information. These lines are thus both well maintained and well characterized, making it possible to identify subsets of lines representing the majority of phenotypic and genetic variation within a species (Brothers & Miller 1999; Mandel et al. 2011).

In addition to inbred and open-pollinated accessions, several well-characterized mapping populations have been developed (e.g. Burke et al. 2002; Tang et al. 2002; Baack et al. 2008; Bowers et al. 2012). Crosses have been made between wild and domesticated lines (Burke et al. 2002; Wills & Burke 2007; Baack et al. 2008; Wills et al. 2010), elite cultivars and Native American landraces not subjected to modern breeding efforts (J.G. Barb et al., personal communication), between elite modern breeding lines (Berry et al. 1995; Tang et al. 2002) and between species (Rieseberg et al. 1995; Burke et al. 2004; Lai et al. 2005b; Heesacker et al. 2009). The existing sunflower genetic maps can be viewed and compared using CMap at the Sunflower CMap database (http://www.sunflower.uga.edu/cmap/). A consensus map based on two RIL populations including 453 RIL has been also made available at http://www.heliagene.org/Web/public/mapping_downy_mildew_resistance_genes.html. As of this writing, there are thus multiple genetic maps that contain many thousands of markers with known positions in the genomes of H. annuus (Bowers et al. 2012) and related species, including one of the highest-density genetic maps ever made, with 2.6 million mapped SNPs, for H. annuus (C. Grassa, personal communication). Currently being developed, a physical map containing unique sequence-based markers every 5–6 kb will be a major addition, greatly facilitating the ongoing assembly of the domesticated sunflower genome (Kane et al. 2011).

The current version of the H. annuus genome sequence represents over 80% of the sunflower genome and is comprised of tens of thousands of scaffolds (Kane et al. 2011). Incorporation of the physical and genetic maps, as well as substantial new long-read Roche 454 FLX Plus and GS FLX Titanium sequence data, is certain to improve these measures substantially. Substantial Illumina GAII whole-genome shotgun sequence data are publicly available (GenomeProject ID #64989) for H. annuus. In addition to the domesticated H. annuus genome, which has been the primary target of the sequencing effort, substantial genomic sequence data are being generated for numerous wild species, and high-quality annotated genomes will be assembled for several, including the annuals H. argophyllus, H. petiolaris and the perennial H. maximiliani.

Over 2000 microsatellites and other markers have been developed and genetically mapped (Berry et al. 1995; Gentzbittel et al. 1995, 1999; Jan et al. 1998; Gedil et al. 2001; Burke et al. 2002, 2004; Tang et al. 2002; Yu et al. 2003; Lai et al. 2005a,b; Chapman et al. 2008b; Heesacker et al. 2008, 2009), some of which occur in expressed genes and some in anonymous markers located largely in intergenic regions. Several large EST databases in sunflowers have provided important additional resources for developing SSRs quickly and cost-effectively (Pashley et al. 2006; Ellis & Burke 2007) as have user-friendly bioinformatic pipelines for identifying potentially useful microsatellite markers (e.g. find SSR, Kane & Rieseberg 2007; Barker et al. 2010). These resources have greatly facilitated population genetic work characterizing diversity and gene flow within species (Liu & Burke 2005; Kane & Rieseberg 2007; Chapman et al. 2008b; Kane et al. 2009) and across species boundaries (Yatabe et al. 2007; Scascitelli et al. 2010). The origin of several of the homoploid species has also been investigated using these microsatellites (Schwarzbach & Rieseberg 2002; Welch & Rieseberg 2002b; Gross et al. 2003, 2007; Sapir et al. 2007).

With 323 332 Sanger ESTs available in GenBank as of April 2011, and millions of ESTs sequenced using next-generation technology (Lai et al. 2012), the genetic diversity at expressed genes is extremely well characterized in Helianthus. Sequence variation within species is quite high for both H. annuus and H. petiolaris (Strasburg et al. 2009). In addition to sequence data available on Genbank, information on the ESTs and other resources for sunflowers and related species can be found on the Compositae Genome Project (CGP) database (http://compgenomics.ucdavis.edu/) and at HeliaGene (http://www.heliagene.org/). Taking advantage of the available variation within H. annuus, two SNP chips have been constructed (Kane et al. 2011; Bachlava et al. 2012): a 10 640-SNP Infinium array and a 2.56 million feature Affymetrix chip, based on Sanger EST sequences from H. annuus, H. petiolaris, H. argophyllus, H. exilis, H. paradoxus, H. ciliaris and H. tuberosus sequenced by the CGP. Both have been used successfully for genetic mapping (Bowers et al. 2012; N. Langlade et al. unpublished; J. Bowers et al. unpublished), and high-density genetic maps are expected to be published using each chip in the near future. To assess expression differences throughout the transcriptome, both 4- and 12-plex NimbleGen expression arrays have been developed for common sunflower (Lai et al. 2012), with 68 400 features from 33 376 unigenes in the 4-plex array and 136 454 features from 48 683 unigenes from the 12-plex array. Additional gene expression work in progress uses a 44 000 feature Agilent array that is populated with genes from H. argophyllus and H. annuus (J. M. Burke, personal communication). Reliable genome size estimates have been made for 20 species, with chromosome numbers characterized for all species (Table 1). Replication for each species, more details on these estimates and the methods used can be found on the Kew c-values database (http://data.kew.org/cvalues/).

Reverse-genetic tools are less well developed than forward-genetic approaches in Helianthus. However, substantial work in this area is in progress, with tilling populations in Italy (Sabetta et al. 2009) and in France (P. Vincourt et al., unpublished) and reports of successful transformation from several groups (Schrammeijer et al. 1990; Knittel et al. 1994; Malone-Shoneberg et al. 1994; Grayburn & Vick 1995; Rao & Rohini 1999; Weber et al. 2003; Mohamed et al. 2006; Manavella & Chan 2009; Zhang 2011). Nested core collections have been built from sunflower inbred lines of different origins, and recent work demonstrates that this is a powerful way to identify genetic variation underlying important segregating traits (Coque et al. 2008; Mandel et al. 2011).

Conclusions and future directions

With the forthcoming full genome sequence, new reverse-genetic tools, and increasingly well-characterized variation in important traits such as flowering time and herbivory, it is clear that sunflowers are emerging as an ecological model system with numerous advantages. Work towards more reliable transformation protocols and further development of mapping populations exploiting the substantial standing variation are also quite promising. As sequencing costs drop and genetic and germplasm resources continue to be improved, sunflower evolutionary geneticists are poised to answer key classic questions in ecology, evolution, development and population genetics. These include the following: What is the role of hybridization in adaptive evolution and speciation? What is the genetic basis of transgressive segregation? How do chromosomal differences get established and what is their role in reproductive isolation? How has the unique, compound inflorescence of the Asteraceae evolved, and what are the genetic factors underlying its development? What is the nature of species? What genetic differences define species boundaries? How does hybridization and gene flow between lineages affect adaptation and divergence? What is the genetic basis of parallel adaptations, such as invasiveness or adaptation to extreme environments? Particularly in the annual sunflowers, all of which are diploid, the wealth of population genetic, genomic and ecological data currently available and in progress means that many of these questions are on the verge of being answered.

Acknowledgements

This research was supported by Genome Canada, Genome BC, the French National Institute for Agricultural Research (INRA), the US National Science Foundation (DEB0421630; DBI-0820451) and the US Department of Agriculture (2008-35300-19263, 2008-35300-04579, 2008-35504-04854, 2010-33522-21668).

All authors contributed to the writing of the paper and preparation of tables and figures.