PERSPECTIVES

Genome Evolution Due to Allopolyploidizationin Wheat

Moshe Feldman1 and Avraham A. LevyDepartment of Plant Sciences, The Weizmann Institute of Science, Rehovot 76100, Israel

ABSTRACT The wheat group has evolved through allopolyploidization, namely, through hybridization among species from the plantgenera Aegilops and Triticum followed by genome doubling. This speciation process has been associated with ecogeographicalexpansion and with domestication. In the past few decades, we have searched for explanations for this impressive success. Our studiesattempted to probe the bases for the wide genetic variation characterizing these species, which accounts for their great adaptabilityand colonizing ability. Central to our work was the investigation of how allopolyploidization alters genome structure and expression.We found in wheat that allopolyploidy accelerated genome evolution in two ways: (1) it triggered rapid genome alterations throughthe instantaneous generation of a variety of cardinal genetic and epigenetic changes (which we termed “revolutionary” changes), and(2) it facilitated sporadic genomic changes throughout the species’ evolution (i.e., evolutionary changes), which are not attainable atthe diploid level. Our major findings in natural and synthetic allopolyploid wheat indicate that these alterations have led to thecytological and genetic diploidization of the allopolyploids. These genetic and epigenetic changes reflect the dynamic structuraland functional plasticity of the allopolyploid wheat genome. The significance of this plasticity for the successful establishment ofwheat allopolyploids, in nature and under domestication, is discussed.

HYBRIDIZATION and polyploidization are ubiquitousmodes of evolution in plants and in other eukaryotes

(reviewed by Van de Peer et al. 2009). The wheat group(genera Aegilops and Triticum) emphasizes the impact ofhybridization and polyploidization on species evolution innature and under domestication. For example, bread wheathas a complex genome consisting of three related genomesthat derived from three different diploid species; it is calledan allohexaploid (allo, from Greek, meaning “different”).Pasta wheat is an allotetraploid. Overall (Table 1), the wheatgroup contains 13 diploid species and 18 allopolyploid spe-cies (Kihara 1924, 1954; Sax 1927; Sears 1948, 1969; Kiharaet al. 1959; Morris and Sears1967; Van Slageren 1994; andreference therein). Attempts to determine the timing ofwheat speciation, using DNA data, suggest that the diploidprogenitors of allopolyploid wheat have diverged froma common progenitor some 2.5–4.5 MYA (Huang et al.2002). Because of this fairly recent divergence, most ofthe diploid wheat species have relatively limited morpholog-ical and molecular variation, occupy only a few well-defined

ecological habitats, and are distributed throughout relativelysmall geographical areas (Eig 1929; Zohary and Feldman1962; Kimber and Feldman 1987; Van Slageren 1994). In his-torical terms, allotetraploid wheat developed about 300,000to 500,000 years ago (Huang et al. 2002), while allohexa-ploid wheat was formed only about 10,000 years ago (Feldmanet al. 1995; Feldman 2001). According to Stebbins (1950),newly formed allopolyploids are often characterized by lim-ited genetic variation, a phenomenon he referred to as the“polyploidy diversity bottleneck.” This bottleneck arises be-cause only a few diploid genotypes were involved in theallopolyploid speciation events, because the newly formedallopolyploid is immediately isolated reproductively from itstwo parental species, and because time was not sufficient forthe accumulation of mutations.

Despite the diversity bottleneck in newly formed allopoly-ploids and despite the fact that all Aegilops and Triticumallopolyploids formed much later than their ancestral dip-loids, they differ radically from their diploid progenitors.The allopolyploids show wider morphological variation, occupya greater diversity of ecological habitats, and are distributedover larger geographical area than their diploid progenitors(Eig 1929; Zohary and Feldman 1962; Kimber and Feldman1987; Van Slageren 1994). They are also dynamic colonizers

Copyright © 2012 by the Genetics Society of Americadoi: 10.1534/genetics.112.1463161Corresponding author: Plant Sciences, The Weizmann Institute foe Science, 234 HerzlSt., Rehovot 76100, Israel. E-mail: moshe.feldman@weizmann.ac.il

Genetics, Vol. 192, 763–774 November 2012 763

compared to their diploid progenitors (Zohary and Feldman1962). Therefore, students of wheat evolution have been fasci-nated by this paradox and have dedicated themselves to unrav-eling the processes and mechanisms that contributed to thebuild up of genetic and morphological diversity of allopolyploidsand to their great evolutionary success in term of proliferationand adaptation to new habitats, including under domestication.During the past several decades, methods and materials havebeen developed that facilitate studies of genomic alterationstriggered by allopolyploidization. Studies of natural and syn-thetic wheat and related allopolyploids, as well as genomesequencing data, indicate that a broad range of DNA rear-rangements occur during, immediately, or within a few gen-erations following allopolyploidization. These rapid changescontribute to increased diversity at the intra-specific level.

What triggers these changes is a fascinating and still un-answered question, but it is clear that these developments

are rapid and extensive, including the loss of coding andnoncoding DNA sequences, transposon activations, and genesilencing or duplication and pseudogenization (Levy andFeldman 2004; Feldman and Levy 2005, 2009; and refer-ences therein). We therefore have distinguished betweenrevolutionary changes—which occur rapidly, within a fewgenerations—and evolutionary changes, which take placegradually in the polyploid lineage, on an evolutionary timescale of hundreds or thousands of generations (Feldman andLevy 2005, 2009). Note that these evolutionary changes mayalso be relatively faster than run-of-the-mill evolution due tothe buffering of mutations in the polyploid genome (Mac Key1954, 1958; Sears 1972; Thompson et al. 2006), whichspeeds neo- or subfunctionalization of genes and brings ondiploidization and divergence from the diploid progenitorgenomes. In this manuscript, we focus on the contributionof wheat and related species to our understanding of evolu-tionary processes that shape allopolyploid genomes.

Early Studies on Interspecific Hybridization

A newly formed allopolyploid wheat species results from aninterspecific or intergeneric hybridization, followed by spon-taneous somatic or, more likely, meiotic chromosome doublingin the sterile F1 or, alternatively, derived from crosses betweenparental species with unreduced gametes (Ramsey andSchemske 1998). Indeed, early studies of synthetic interge-neric hybrids of wheat showed that the frequency of unre-duced gametes in the F1 hybrids could be as high as 50%(Kihara and Lilienfeld 1949) and identified the occurrenceof spontaneous chromosome doubling, thus demonstratingthe possibility of species formation via allopolyploidy in thewheat group (Von Tschermak and Bleier 1926). Thus, a newlyformed allopolyploid is a hybrid that contains two or moredifferent genomes enveloped within a single nucleus. Conse-quently, the evolutionary process of allopolyploidization exertsconsiderable stress on the young species whose genomes arenot always compatible, as seen with the hybrid necrosissyndrome (Caldwell and Compton 1943; Hermsen 1963;Tsunewaki 1970). This stress was referred to by BarbaraMcClintock as a genomic shock that could activate transpo-sons and further reduces the fitness of the hybrid (McClintock1984). This raised the question, discussed below, on the ex-tent and modes of hybridization that actually occur in nature.

Zohary and Feldman (1962) studied the extent and pat-tern of hybridization and allopolyploidization that occur innature in Aegilops and Triticum species. Evidence that manywheat allopolyploids were genetically interconnected throughhybridizations between various allopolyploid species andtherefore did not evolve independently was presented. Thesewheats were divided into three groups, each containing severalallopolyploids that share a single, common genome but differin their other genome(s) (Zohary and Feldman 1962). Theallopolyploids within each group grow in mixed populationsin many sites, where they can easily hybridize (Feldman1965a). In laboratory hybridization studies of allopolyploids,

Table 1 The species of the wheat group (the genera Aegilops,Amblyopyrum, and Triticum)

Ploidy level Species Genomea

Diploids(2n = 2x = 14)

Amblyopyrum muticum(=Ae. mutica)

Aegilops speltoidesAe. bicornisAg. LongissimaAe. sharonensisAe. searsiiAe. tauschii(=Ae. squarrosa)

Ae. caudataAe. umbellulataAe comosaAe. uniaristataTriticum monococcumT. urartu

TT

SSSbSb

SlSl

SlSl

SsSs

DD

CCUUMMNNAmAm

AA

Tetraploids(2n = 4x = 28)

Ae. biuncialisAe. geniculata(=Ae. ovata)

Ae. neglecta(=Ae. triaristata 4x)

Ae, columnarisAe. triuncialisAe. kotschyiAe. peregrina(=Ae. variabilis)

Ae. cylindricaAe. crassa 4xAe. ventricosaT. turgidumT. timopheevii

UUMMMMUU

UUMM

UUMMUUCC; CCUUSSUUSSUU

DDCCDDMMDDNNBBAAGGAA

Hexaploids(2n = 6x = 42)

Ae. recta(=Ae. triaristata 6x)

Ae. vaviloviiAe. crassa 6xAe. juvenalisT. aestivumT, zhukovskyi

UUMMNN

DDMMSSDDDDMMDDMMUUBBAADDGGAAAmAm

a Genome designations according to Kimber and Tsunewaki (1988); underlineddesignation indicates a modified genome.

764 M. Feldman and A. A. Levy

it was found that the common genome acts as a buffer en-suring some seed fertility following pollination by either oneof the parents, while the chromosomes of the dissimilar ge-nomes, brought together from different parents, may pair tosome extent and exchange genetic material (Feldman 1965b,c). Consequently, the dissimilar genomes of these allopoly-ploids are recombined, or modified genomes (Zohary andFeldman 1962), which contain chromosomal segments thatoriginated from two or more diploid genomes. These speciesoverlap in their genetic variation ranges and are intercon-nected by a series of morphological intermediate forms.Thus, despite the genetic barriers and the genomic shockinvolved, interspecific hybridization, which is rare at thediploid level, has played a decisive role in the establishmentof a wide range of genetic variation in the allopolyploidspecies (Zohary and Feldman 1962). This has probably sig-nificantly contributed to their evolutionary success.

In addition, the newly formed allopolyploids—especiallyannual, predominantly self-pollinated species, like those ofthe wheat group—must overcome several immediate chal-lenges to survive at the cytological, genetic, and epigeneticlevels (Levy and Feldman 2002, 2004; Feldman and Levy2005, 2009, 2011). We argue, in the following sections, thatovercoming these challenges, and increasing the nascentallopolyploid fitness, is achieved through genomic plasticity,namely through the immediate triggering of a variety ofcardinal genetic and epigenetic changes that affect genomestructure and gene expression.

Cytological Diploidization

Bread wheat is an allohexaploid species (2n = 6x = 42,genomes BBAADD) that originated from hybridizationevents involving three different diploid progenitors classi-fied in the genera Triticum and Aegilops: (i) T. urartu, thedonor of the A genome (Dvorak 1976; Chapman et al. 1976),(ii) a yet-undiscovered extant or extinct Aegilops speciesclosely related to Ae. speltoides, the donor of the B genome,and (iii) Ae. tauschii, the donor of D genome (McFaddenand Sears 1944, 1946; Kihara 1944). On the basis of geneticsimilarities, the 21 pairs of homologous chromosomes ofbread wheat (seven pairs in each genome) fall into sevenhomeologous groups, each containing one pair of chromo-somes from the A, B, and D genomes, respectively (Sears1954; Figure 1). Hence, homeologous group 1, for example,contains the pair 1A, 1B, and 1D. In each group, homeologouschromosomes, which are derived from a relatively recent,common ancestral chromosome, i.e., only 2.5–4.5 MYA (Huanget al. 2002), still share a high degree of gene synteny andDNA sequence homology. However, they differ from oneanother by a number of noncoding and highly repetitiousDNA sequences (Flavell 1982), and many functional genecomplexes (Wicker et al. 2011 and references therein). Inspite of this genetic relatedness, the homeologues do not pairwith each other at meiosis, a phenomenon that still requiresclarification.

In allopolyploid wheat, the restriction of pairing to homo-logous chromosomes, i.e., cytological diploidization, has de-veloped through two independent but complementarysystems. The first system to be studied is based on the ge-netic control of pairing. The second depends on the physicaldivergence of the homeologous chromosomes.

In 1958, a gene was discovered on the long arm ofchromosome 5B of bread wheat that restricts meiotic pairingto completely homologous chromosomes (Riley and Chapman1958; Sears and Okamoto 1958). Gene(s) with similar effecthave not been found in the homeologous chromosomes 5Aand 5D of common wheat (Sears 1976). Discovery of thisgene has had a great impact on wheat cytogenetics: it becamea subject of intensive cytogenetic studies that attempted tounderstand its mode of action, evolution, and breeding signif-icance. The gene, called Ph1 (pairing homeologues; Wall et al.1971), was further positioned some 1.0 cM from the centro-mere of the long arm of chromosome 5B (5BL) (Okamoto1957; Sears 1984). It is a dominant gene that suppressespairing of the homeologues (intergenomic pairing) whileallowing that of homologs (intragenomic pairing) (Riley1960; Sears 1976; and reference therein). Consequently, onlybivalents are formed during allopolyploid wheat meiosis.

The mechanism controlling the Ph1 mode of action is stillunclear. Six extra doses of the Ph1-containing arm of chro-mosome 5B caused partial asynapsis of homologs at meiosis,concomitantly allowing some pairing of homeologous chro-mosomes and inducing a high frequency of interlocking biva-lents (Feldman 1966). Similar effects were observed bypremeiotic treatment with colchicine (Driscoll et al. 1967;Feldman and Avivi 1988). It was therefore assumed thatthe hexaploid nucleus still maintains some organizationalaspects of the individual ancestral genomes; i.e., each genomeoccupies a separate region in the nucleus, which in turn, isrecognized by Ph1 (Feldman 1993 and references therein).

A model was thus proposed whereby Ph1 exerts its effectat premeiotic stages, before the commencement of synapsis,and affects the premeiotic alignment of homologous andhomeologous chromosomes. It therefore controls the regu-larity and pattern of pairing (Feldman 1993 and referencestherein). In euploid bread wheat, two doses of Ph1, whilescarcely affecting homologous chromosomes, keep thehomeologues apart, thereby leading to exclusive homolo-gous pairing in meiosis. In the absence of Ph1, the threegenomes are mingled together in the nucleus and conse-quently, the homeologues can also pair with each other.Six doses of Ph1 or premeiotic treatment with colchicineinduces separation of chromosomal sets and the randomdistribution of chromosomes in the premeiotic nucleus, lead-ing to an increased distance between homologs. This resultsin partial asynapsis of homologs that are relatively distantand some pairing of homeologues that happen to lie close toone another. Thus, an interlocking of bivalents can occurwith their chromosomal constituents coming to pair froma relative distance and catching other chromosomes be-tween them (Feldman 1993 and references therein).

Perspectives 765

Studying the effects of Ph1 on centromere behavior inmeiotic cells, Vega and Feldman (1998) concluded thatthe Ph1 gene affects the interaction between the centro-meres and microtubules, possibly through a microtubule-associated protein (MAP) whose activity is located nearthe centromere. G. Moore and co-workers have localizedPh1 to a 2.5-Mb interstitial region of wheat chromosome5B (Griffiths et al. 2006; Yousafzai et al. 2010). This regioncontains a structure consisting of a segment of subtelomericheterochromatin that inserted into a cluster of cdc2-relatedgenes after polyploidization (Griffiths et al. 2006). The cor-relation of the presence of this structure with Ph1 activity inrelated species, and the involvement of heterochromatinwith Ph1 and cdc2 genes with meiosis, makes the structurea good candidate for the Ph1 locus. These mapping datasuggest that the mode of action of this complex locus isdetermined by the effect of cyclin-dependent kinases oncell-cycle progression (Griffiths et al. 2006; Al-Kaff et al.2008; Yousafzai et al. 2010). However, direct evidence asto the identity of the coding gene and its mode of action isstill missing. Recently, K. Gill and associates (personal com-munication) identified a candidate gene in the Ph1 regionthat is different from the one that was proposed by the labof G. Moore. Silencing of the newly identified gene disruptsthe alignment of chromosomes on the metaphase plate inearly stages of meiosis, suggesting a role of microtubules inits function (K. Gill, personal communication).

It was suggested by several authors that Ph-like genesexist in diploid wheat species (Okamoto and Inomata1974; Avivi 1976; Waines 1976; Maan 1977) and that theybecame more effective at the polyploid level as a result ofduplication. This dosage-dependent effect might have beenselected to yield an allopolyploid with improved fertility.

The suppressive effect of Ph1 on homeologous pairing ininterspecific and intergeneric wheat hybrids is absolute.However, its effect on homeologous pairing in bread wheatitself might not be indispensible as plants deficient for thisgene exhibit relatively little homeologous pairing. This isevidenced from the small number of multivalents (less thanone per cell), resulting from intergenomic pairing in theseplants (Sears 1976). Interestingly, and in accord with theabove, Ph-like gene(s) have not been found in any of theallopolyploid species of the closely related Aegilops genus(Sears 1976). Nevertheless, these species also exhibit exclu-sive bivalent pairing of fully homologous chromosomes.Hence, additional mechanisms that ensure homologous pair-ing must be called for and are discussed below.

Novel molecular tools for investigating wheat genomicshas advanced our understanding of the exclusive intra-genomic pairing in bread wheat. Using micromanipulation,Vega et al. (1994, 1997) isolated more than 30 isochromo-somes of the long arm of chromosome 5B from first meioticmetaphase spreads of a monoisosomic 5BL line of breadwheat. The dissected isochromosomes were amplified bydegenerate oligonucleotide-primed PCR and a large numberof DNA sequences were produced from this arm (Vega et al.1994, 1997). These sequences were classified by us (Feldmanet al. 1997) into the following four classes (Figure 1): (i) non-specific sequences (mainly repetitious sequences) that arefound in all or many of the wheat chromosomes; (ii) group-specific sequences (mainly coding sequences) that occur inthe chromosomes of one homeologous group, e.g., 1A, 1B,and 1D; (iii) genome-specific sequences that occur in severalchromosomes of one genome, e.g., 1A, 2A, 3A, etc.; and (iv)chromosome-specific sequences (CSSs) that occur in onlyone homologous chromosome pair, e.g., 1A and 1A. Most ofthe CSSs are noncoding sequences (Feldman et al. 1997) andare present in all the diploid species of Aegilops and Triticum,but occur in only one pair of chromosomes of allopolyploidwheat, suggesting that they were lost during or after allopoly-ploidization (Feldman et al. 1997)

A most surprising discovery is that allopolyploidization inthe wheat group causes immediate nonrandom eliminationof specific noncoding, low-copy, and high-copy DNA se-quences (Feldman et al. 1997; Liu et al. 1998a,b; Ozkanet al. 2001; Shaked et al. 2001; Han et al. 2003, 2005; Salinaet al. 2004). The extent of DNA elimination was estimatedby determining the amounts of nuclear DNA in natural allo-polyploids and in their diploid progenitors, as well as innewly synthesized allopolyploids and their parental plants(Ozkan et al. 2003; Eilam et al. 2008, 2010). Natural wheatand related allopolyploids contain 2–10% less DNA than thesum of their diploid parents, and synthetic allopolyploids

Figure 1 Schematic of the wheat karyotype. The wheat karyotype isarranged into genomes A, B, and D and into seven homeologous groups(e.g., group 1 consists of chromosomes 1A, 1B, and 1D). This arrange-ment is after Sears (1954), who classified homeologous chromosomesbased on their ability to compensate for each other’s absence. Examplesof the different types of sequences are drawn on top of the chromo-somes, namely: group-specific sequences that are present in only onehomeologous group, chromosome-specific sequences (CSSs) that arepresent in only one chromosome pair, genome-specific sequences (GSSs)that can be on more than one chromosome pair but only in one of thegenomes, and nonspecific sequences that are present on both homeol-ogous and nonhomeologous chromosomes. The nonspecific sequencesare mainly dispersed repeats (adapted from Levy and Feldman 2004).

766 M. Feldman and A. A. Levy

exhibit a similar loss, indicating that DNA elimination occurssoon after allopolyploidization (Nishikawa and Furuta 1969;Furuta et al. 1974; Eilam et al. 2008, 2010). Also, the nar-row intraspecific variation in DNA content of the allopoly-ploids supports that the loss of DNA occurred immediatelyafter the allopolyploid formation and that there was almostno subsequent change in DNA content during the allopoly-ploid species evolution (Eilam et al. 2008). In triticale (a syn-thetic allopolyploid between wheat and rye, Secale cereale),Boyko et al. (1984, 1988) and Ma and Gustafson (2005)found that there was a major reduction in DNA content inthe course of triticale formation, amounting to �9% forthe octoploid and 28–30% for the hexaploid triticale. In thissynthetic allopolyploid, the various genomes were not af-fected equally: the wheat genomic sequences were relativelyconserved, whereas the rye genomic sequences underwenta high level of variation and elimination (Ma et al. 2004; Maand Gustafson 2005, 2006). Similarly, in hexaploid wheat,genome D underwent a considerable reduction in DNA,while the A and B genomes were not reduced in size (Eilamet al. 2008, 2010).

Bento et al. (2011) reanalyzed data concerning genomicanalysis of octoploid and hexaploid triticale and differentsynthetic wheat hybrids, in comparison with other polyploidspecies. This analysis revealed high levels of genomicrestructuring events in triticale and wheat hybrids, namelymajor parental band disappearance and the appearance ofnovel bands. Furthermore, the data showed that restructuringdepends on parental genomes, ploidy level, and sequencetype (repetitive, low copy, and/or coding); is markedly differ-ent after wide hybridization or genome doubling; and affectspreferentially the larger parental genome (Bento et al. 2011).

Screening the parental wheat diploid species for a numberof CSSs, each located on different chromosome arms incommon wheat, it was found that many of them occur in allthe diploids (Feldman et al. 1997). However, in the allopoly-ploids these sequences occur in only one chromosome pair andare absent from the homeologous chromosomes (Feldmanet al. 1997). By analyzing newly formed Triticum and Aegilopsallopolyploids with genomic combinations similar to thoseof the natural allopolyploids, it was shown that the CSSswere eliminated from one genome immediately or a fewgenerations after the formation of the allopolyploid (Feldmanet al. 1997; Ozkan et al. 2001). Thus, chromosome-specificsequences are found in each homologous pair, but differenthomeologous pairs have their own unique signatures.

Since CSSs are the only sequences that determine chro-mosomal homology, it is assumed that they are implicatedin recognition and initiation of homologous pairing atmeiosis. Therefore, if upon allopolyploid formation thesesequences are eliminated from one pair of homeologouschromosomes in tetraploids or from two pairs in hexaploids,a cytological diploidization process that strongly augments thephysical divergence of the homeologous chromosomes takesplace. It is then extremely difficult or even impossible for themto pair at meiosis. Thus, cytological diploidization leads

to exclusive intragenomic pairing, i.e., diploid-like meioticbehavior.

DNA elimination seems not to be random at the intra-chromosomal level as well (Liu et al. 1997). For example,these authors found that the CSSs on chromosome arm 5BLin allohexaploid wheat are not distributed randomly butcluster in terminal (subtelomeric), subterminal, and intersti-tial regions of this arm. Such structures make these regionsextremely chromosome specific—or homologous. Hence, itwas tempting to suggest that these chromosome-specificregions are equivalents to the classical “pairing-initiationsites” that play a critical role in homology search and initi-ation of pairing at meiosis (Feldman et al. 1997). Computermodeling also shows that homeolog divergence in associa-tion with pairing stringency drives disomic inheritance (LeComber et al. 2010),

To sum up, cytological diploidization in allopolyploidwheat was engendered by two independent, complementarysystems. One is based on the physical divergence of chromo-somes and the second, on the genetic control of pairing. ThePh-gene system superimposes itself on and takes advantageof—and thereby reinforces—the above-described system ofthe physical differentiation of homeologous chromosomes.In addition, stringent selection for fertility might well favorthe development of two systems to effect the suppression ofmultivalent formation and the promotion of bivalent pairingin nature and, more so, in domesticated material.

The process of cytological diploidization in wheat groupallopolyploids has been critical for their establishment innature. The restriction of pairing to completely homologouschromosomes ensures regular segregation of genetic mate-rial, high fertility, genetic stability, and disomic inheritancethat prevents the independent segregation of chromosomesof the different genomes. Homologous pairing also allows forthe maintenance of favorable intergenomic genetic interac-tions. On the other hand, disomic inheritance sustains theasymmetry in the control of many traits by the differentgenomes (Feldman et al. 2012). In addition, since cytologicaldiploidization facilitates genetic diploidization, existing genesin double and triple doses can be diverted to new functionsthrough mutations, thereby favoring the creation of favor-able, new intergenomic combinations.

Structural Changes That Occur in the Polyploid Lineagethrough Time

Several structural changes are known to have sporadicallyoccurred during the history of allopolyploid wheat genomes.These generate an additional source of variation that, someof which, could not have taken place in the diploid parentalgenomes and occur almost exclusively in an allopolyploidbackground. Such intergenomic changes include the horizon-tal transfer of chromosomal segments, repetitive sequences,transposons, or genes via intergenomic translocations resultingfrom interchange between nonhomologous or homeologouschromosomes. Intergenomic translocations that characterize

Perspectives 767

specific populations or biotypes are widespread in allohex-aploid wheat (Maestra and Naranjo 1999 and referencestherein). Invasion of the A genome by sequences from B—most probably transposons—was detected in wild allotetra-ploid wheat using genomic in situ hybridization (GISH)(Belyayev et al. 2000). The possibility of intergenomic trans-fer adds to allopolyploid genomic plasticity and enables thecreation of new genetic combinations beyond those possiblethrough the intragenomic mechanisms of the individualgenomes.

Moreover, in contrast to the diploids, which are geneticallyisolated from each other and have undergone divergentevolution, wheat group allopolyploids exhibit convergentevolution because they contain genetic material from twoor more unique diploid genomes that can be exchanged viahybridization and introgression, resulting in new genomiccombinations. Examples for introgression between allote-traploid Aegilops species were provided by Zohary and Feld-man (1962) and Feldman (1965a,b,c). Additional evidencefor the existence of introgressed genomes in allopolyploidAegilops was obtained by C-banding analysis (Badaeva et al.2004). Introgression of a DNA sequence from allopolyploidwheat to lines of the allotetraploid Aegilops species, Ae. per-egrine, was also described (Weissmann et al. 2005).

In addition to evolutionary changes that are almostunique among allopolyploids, other types of mutations (e.g.,point mutations, microsatellite instability, transposition, etc.)may contribute, perhaps in an accelerated manner, to evo-lutionarily relevant structural or functional changes in allo-polyploids. For example, presence of duplicate or triplicategenetic material in wheat allopolyploids might have relaxedconstraints on gene structure and function. Thus, the accu-mulation of genetic variation through mutation or hybrid-ization might be more readily tolerated in an allopolyploidthan in a diploid species. While there is no direct evidencefor this assumption, there is indirect support from experi-mental data showing a higher resistance of allohexaploidwheat to irradiation compared to their diploid progenitors(Mac Key 1954, 1958; Sears 1972). Such an increase in mu-tation resistance with increased ploidy was shown in yeastto be correlated with higher evolutionary ability and fitness(Thompson et al. 2006).

Genetic or Functional Diploidization

In some cases, an extra genetic dose can be beneficial, whilein others it may be deleterious. Most duplicated genes thatcode for enzymes are active in allopolyploid wheats (Hart1987). The extra gene itself might provide a favorable effector it could lead to the buildup of positive intergenomic inter-actions if genes or regulation factors on homeologous chro-mosomes are divergent. In some duplicated genes, however,increased dosage has led to redundancy or to a deleteriouseffect. In these cases, functional diploidization processes bringthe redundant or unbalanced gene systems into a diploid-likemode of expression. Functional or genetic diploidization can

be achieved through elimination, inactivation, or diversionof the redundant, duplicated genes to new functions.

In studies on the genetic control of high-molecular-weight(HMW) glutenin subunits, an important component of wheatstorage proteins, allopolyploidy was found to affect genefunction through a variety of genetic and epigenetic mecha-nisms (Galili and Feldman 1984; Forde et al. 1985; Galiliet al. 1986; Waines and Payne 1987). The levels of HMWglutenin subunits in the endospermwas determined in a seriesof wheats containing different doses of chromosomes 1A, 1B,or 1D, which carry the genes controlling the production ofthese proteins. It was found that HMW glutenin subunit pro-duction was nonlinear in response to gene dosage, indicatingthe operation of a mechanism of dosage compensation (Galiliet al. 1986). Such dosage compensation is commonly ob-served as a way to instantaneously reduce the negative effectof overproduction and inefficiency of genes that exist in superoptimal doses (Birchler et al. 2001).

Another aspect regulating gene action in newly formed al-lopolyploids is intergenomic suppression (Galili and Feldman1984; Galili et al. 1986). By crossing hexaploid with tetra-ploid wheat, backcrossing the pentaploid offspring of eachgeneration back to the hexaploid parent, and finally selfingthe pentaploid to obtain tetraplid plants, Kerber (1964)extracted the A and B genomes of allohexaploid wheat (ge-nome BBAADD) to produced an extracted allotetraploidwheat line lacking the D genome. This line facilitated thestudy of intergenomic relationships between the D genomegenes and those of A and B. When compared with the motherallohexaploid, polyacrylamide gels of storage proteins of theextracted tetraploid line exhibited several bands with increasedstaining intensity as well as new bands (Galili and Feldman1984). These novel bands are assumed to have resultedfrom a novel activity of genes located on the A or B genomes,which are no longer repressed by the removed D genome. Re-addition of the D genome to the extracted allotetraploid in-stantaneously resuppressed these genes. Galili and Feldman(1984) suggested that these endosperm-protein genes wererepressed immediately following the formation of allohexa-ploid wheat, about 10,000 years ago, but they still have re-tained their potential for being activated.

Likewise, using microarray analysis, a group of geneslocated on chromosomes of genomes A and B were found tobe strictly regulated by the D genome. They are not expressedin allohexaploid wheat; they are expressed in an extractedallotetraploid (genome BBAA) and they are silenced againupon re-adding the D genome (B. Liu, personal communi-cation). Similarly, Kerber and Green (1980) described anintergenomic suppression of a rust-resistance gene in ge-nome D by gene(s) in genome A or B. Intergenomic suppres-sion of disease-resistance genes is a common phenomenonin wheat and related allopolyploids, as was noted in severalnatural and newly formed allopolyploids (Y. Anikster, J.Manisterski, and M. Feldman, unpublished data). Compara-ble results were obtained by Dhaliwal and co-workers(Aghaee-Sarbarzeh et al. 2001) in Triticum durum–Aegilops

768 M. Feldman and A. A. Levy

amphiploids. They found that dominant leaf-rust- and stripe-rust-resistance genes from the Aegilops parents were sup-pressed by genes in the AB genomes of the wheat parent.Another well-studied example of intergenomic suppressionis the silencing in triticale of the ribosomal RNA genes of ryein the presence of the wheat genome (Appels et al. 1986 andreferences therein). Cytosine methylation is involved in thissilencing as suggested by reactivation of the rye ribosomalRNA genes by the demethylation agent 5-azacytidine or meth-ylation sensitive/insensitive isoschizomers (Houchins et al.1997). Intergenomic suppression is a common control mecha-nism for instantaneously reducing the negative effect ofoverproduction and the inefficiency of genes that exist insuper-optimal doses.

In bread wheat, HMW glutenin genes are located onchromosomes of homeologous group 1, i.e., 1A, 1B, and 1D.These chromosomes carry two genes: Glu1-1, coding for theslow-migrating subunit, and Glu1-2, coding for the fast mi-grating subunit (Payne et al. 1981). Galili and Feldman(1983b) analyzed 109 different lines of allohexaploid wheatrepresenting a wide spectrum of genetic backgrounds andfound that 22 lines (20.2%) had no HMW glutenin subunitscontrolled by chromosome 1A, 44 lines (40.4%) had only onesubunit controlled by 1A, and 43 lines (39.4%) had two suchsubunits. Moreover, in all lines having one subunit controlledby 1A, this subunit was always the slow-migrating variety;i.e., only Glu-1A-1 was active. Hence, in 60% of the hexa-ploid lines studied, Glu-A1-2 was inactive despite the factthat this gene is regularly active in diploid wheat (Wainesand Payne 1987).

The HMW glutenin subunits in 456 accessions of the wildallotetraploid wheat T. turgidum subsp. dicoccoides, originat-ing from 21 different populations in Israel, were studied(Levy and Feldman 1988; Levy et al. 1988). In 82% of theaccessions, the fast-migrating subunit of genome A was ab-sent, and in 17% of the accessions the slow-migrating sub-unit of this genome was also absent. In all 11 studied lines ofthe primitive domesticated allotetraploid wheat, T. turgidumsubsp. dicoccum, the fast-migrating subunit of genome A,was absent, i.e., Glu-A1-2 was inactive, and in all 19 linesof modern allotetraploid wheat, subsp. durum, Glu-A1-1 wasalso inactive, while only Glu-B1-1 and Glu-B1-2 were active(Feldman et al. 1986). Moreover, the HMW glutenin loci ofgenome B were much more polymorphic than those of ge-nome A (Felsenburg et al. 1991). The reduced polymorphismof the A genome loci apparently reflects the nonrandom in-activation of the HMW glutenin genes, as well.

Thus, in both allotetraploid and allohexaploid wheat,inactivation of HMW glutenin genes was massive andnonrandom and occurred in the glutenin genes of genomeA (Galili and Feldman 1983a,b; Feldman et al. 1986; Levyet al. 1988; and reference therein). In hexaploid wheat, thistendency has been found for HMW gliadin genes, as well(Galili and Feldman 1983a,b). The nonrandom nature of theprocess is shown by the fact that not only were the HMWglutenin genes of the A genome specifically affected but their

order of inactivation was also nonrandom, starting with therapidly migrating subunits and continuing with the slowlymigrating ones. The fast-migrating glutenin gene of genomeA was not inactivated by elimination, but from the position-ing of a terminating sequence inside the transcribed portionof the gene (Forde et al. 1985).

The development of molecular genetic techniques for thestudy of plants in general and wheat in particular hasprovided tools for further investigating genome changes dueto allopolyploidization that are involved in the developmentof functional diploidization. Examining newly formed allo-polyploid wheat by cDNA-AFLP, Shaked et al. (2001) foundthat genes were either eliminated or silenced via cytosinemethylation of DNA sequences. They found that alterationsin cytosine methylation (demethylation or new methylation)affected about 13% of a random set of genomic loci. SeveralcDNAs were expressed in the allopolyploids but not in thediploid progenitors (Shaked et al. 2001; Kashkush et al.2002). Kashkush et al. (2002) studied the response of thetranscriptome to allopolyploidization in the first generationof newly formed allopolyploid wheat and in its two diploidprogenitors. They found that transcript disappearance wasthe result of gene loss or silencing, the latter being associatedwith cytosine methylation. The silenced/lost genes includedrRNA genes as well as genes involved in metabolism, diseaseresistance, and cell-cycle regulation. The activated geneswith known functions were all retroelements. Similarly, Heet al. (2003) found that the expression of a significant frac-tion of the genome (7.7%) is altered in a synthetic hexaploidwheat.

Following allopolyploidization events in wheat, thesteady-state level of expression of long terminal repeats(LTR) in retrotransposons was massively elevated (Kashkushet al. 2002, 2003). In addition, the transcriptional activity ofthe Wis2-1A LTR element was shown to be associated withthe production of read-out transcripts flanking toward hostDNA sequences, a process that occurred in a genome-widemanner (Kashkush et al. 2003). In many cases, these read-out transcripts were associated with the expression of adja-cent genes, depending on their orientation. They knockeddown or knocked out the gene product if the read-out tran-script was in the antisense orientation relative to the orien-tation of the gene transcript (such as the iojap-like gene), orthey overexpressed the gene if the read-out transcript was inthe sense orientation (such as the puroindoline-b gene)(Kashkush et al. 2003). The mechanisms by which transcrip-tional activation of TEs influences the expression of neighbor-ing genes are poorly understood. Recent studies (Kraitshteinet al. 2010; Yaakov and Kashkush 2011) on tracking methyl-ation changes around transposons in the first generations ofa newly formed wheat allopolyploid showed massive changesin methylation and that read-out activity was restricted to thefirst generations of the nascent polyploid species.

Transcriptional activation of transposons in allopolyploidsis part of the “genomic shock” syndrome that results frominterspecific hybridization and polyploidization as proposed

Perspectives 769

by McClintock (1984). The mechanism that leads to thisactivation is not well understood.

Recently, Kenan-Eichler et al. (2011) found that smallRNAs (siRNAs) that are thought to repress transposons un-der normal conditions can be disregulated in allopolyploid.The repression of these repressors is correlated with hypo-methylation of the transposons, which in turn may enabletheir transcriptional activation (Kenan-Eichler et al. 2011).An additional pathway of small RNAs-mediated epigeneticcontrol of genome stability and expression might be achievedthrough the activity of microRNAs. Changes in microRNAsexpression in newly synthesized wheat were observed, suchas that of miR168, which targets the Argonaute1 gene (Kenan-Eichler et al. 2011). Changes in microRNAs expression werealso found in Arabidopsis allopolyploids that presumably haveled to changes in gene expression, growth vigor, and adapta-tion (Ha et al. 2009).

Inactivation of homeoalleles may be a nonrandom effect.The data of Koebner and co-workers (Bottley et al. 2006)suggested that for leaf transcripts, there is a modest biastoward silencing of the D genome copies, but this patterndoes not extend to root transcripts. Interestingly, He et al.(2003) found that Ae. tauschii genes (genome DD) wereaffected much more frequently than T. turgidum genes (ge-nome BBAA) in a BBAADD synthetic allopolyploid, and Dgenome homoeoalleles were silenced twice as frequently asthose from the A or B genomes in a newly synthesized hexa-ploid wheat line. Silencing of the same genes was also foundin the bread wheat cultivar Chinese Spring (He et al. 2003).

Thus, allopolyploidization triggers gene silencing, geneelimination, or gene activation and transposon activation viagenetic and epigenetic alterations immediately upon allo-polyploid formation.

Concluding Remarks

Formation of an allopolyploid species is accomplished ina small number of steps. However, its establishment andsurvival in nature probably depend on its ability to self, thenumber of allopolyploid individuals that are formed, andperhaps also exchange genes with its progenitors, as well ason a high level of genomic plasticity that enables it to overcomepotential incompatibilities in its genome and to gain new traits.The studies reported here suggest that allopolyploid wheat canachieve genomic plasticity through the induction of a seriesof cardinal nonadditive genomic changes. Some of them,genetic and epigenetic, are rapid and non-Mendelian, occur-ring during or immediately after the formation of theallopolyploid (revolutionary changes; Table 2). Other changesoccur sporadically over a long period of time during theevolution of the allopolyploid (evolutionary changes; Table2). From a population point of view, the chances of severalindividuals of a nascent allopolyploid being established asa new species is almost null, unless they exhibit increasedfitness compared to their parents or new traits that can en-able them to colonize new niches. This must occur within a

few generations, otherwise the new species will rapidly be-come extinct.

The revolutionary changes described here may contributeto the establishment of new allopolyploid species. Instanta-neous elimination of sequences from one genome in the newplant increases the divergence of the homeologous chromo-somes, leading to exclusive intragenomic pairing and improv-ing fertility. Mechanisms enabling the loss of deleteriousgenes (e.g., the removal of genetic incompatibilities), positivegene dosage effects, and new intergenomic heterotic interac-tions may all rapidly increase fitness of the nascent species.

Evolutionary changes, however, contribute to the buildupof genetic variability and thereby increase adaptability,fitness, competitiveness, and colonizing ability. In nature,most hybridization events do not lead to the formation ofa new species. However, the wheat group is remarkablyequipped with a battery of molecular mechanisms that enablethe appearance of phenotypic novelty and successful specia-tion through allopolyploidy. Future work should help toclarify (i) the role of specific genes and DNA sequences inallopolyploid speciation; (ii) the mechanisms that conferrobustness to the genome under the shock of allopolyploidy;(iii) the activation of transposable elements; (iv) the mech-anisms that enable the orchestration of chromosome division;and (v) the control of bivalent pairing during meiosis.

The rapid processes of cytological and genetic diploid-ization allow for the development of two contrasting andhighly important genetic phenomena in allopolyploid wheatthat presumably contribute to their evolutionary success: (i)the buildup and maintenance of enduring intergenomicfavorable genetic combinations and (ii) genome asymmetryin the control of a variety of morphological, physiological,and molecular traits, namely, the total or predominant controlof certain traits by a single constituent genome. While thefirst phenomenon was taken for granted by plant geneticists,genomic asymmetry in interspecific hybrids and allopoly-ploids was mainly recognized in ribosomal RNA genes(reviewed in Pikaard 2000). Only recently has this beendocumented in allopolyploid wheat, where genome A wasfound to control morphological traits while genome B inallotetraploid wheat and genomes B and D in allohexaploidwheat were shown to control reactions to biotic and abioticfactors (Feldman et al. 2012 and references therein). Inter-genomic pairing might have led to disruption of the linkageof the homeoalleles, which contribute to positive intergeno-mic interactions and might have also led to the segregationof genes that participate in the control of certain traits bya single genome. Intergenomic recombination could there-fore produce many intermediate phenotypes that may neg-atively affect the functionality, adaptability, and stability ofthe allopolyploids.

The revolutionary and evolutionary genomic changes re-ported for wheat allopolyploids emphasize the dynamic plas-ticity of their genomes with regard to structure and functionalike. These changes might have improved and currentlyimprove the adaptability of the newly formed allopolyploids

770 M. Feldman and A. A. Levy

and facilitated their rapid colonization of new ecologicalniches. No wonder, therefore, that domesticated allohex-aploid wheat exhibits a wider range of genetic flexibilitythan diploid species and has been able to adapt itself to animpressive variety of environments over a very short evolu-tionary time scale.

Ohno’s (1970) proposal that evolution moves forwardthrough polyploidy (whole-genome duplication) is presentlygenerally accepted. The current use of accurate and sensi-tive molecular methods, e.g., sequence analysis, shows thatpolyploidy (recent and ancient) is a widespread phenome-non occurring in up to 80% of angiosperms (Soltis and Soltis1993; Masterson 1994; Otto and Whitton 2000), in 95% ofpteridophytes (Grant 1971), and in the lineage of all verte-brates (Van de Peer et al. 2009 and references therein). Spe-cies that were considered typical diploids (e.g., maize, rice,Arabidopsis, yeast, and humans) are in fact ancient poly-ploids, i.e., paleopolyploids, that underwent one or morerounds of chromosome doubling during their evolution(Van de Peer et al. 2009 and references therein). The pro-genitors of paleopolyploids cannot be identified due to bothcytological and genetic diploidization. The fact that all poly-ploids, including paleopolyploids, recent allopolyploids, anddiloidized autopolyploids, underwent cytological and ge-netic diploidization supports the concept that polyploidywithout diploidization is an evolutionary dead end. Theworks reported here, showing rapid diploidization, suggestthat polyploids that did not diploidize could not surviverather than an alternative model of gradual decay of redun-

dant loci. Hence, the processes described above and similarprocesses described in other plant polyploids (e.g., Chen andNi 2006; Chaudhary et al. 2009; Tate et al. 2009; Buggset al. 2011; and references therein) suggest that diploidiza-tion was essential for the successful establishment of poly-ploid plant and animal species.

Acknowledgments

The constructive comments of four anonymous reviewersand of the editor are acknowledged with appreciation.

Literature Cited

Aghaee-Sarbarzeh, M., H. S. Dhaliwal, and H. Singh, 2001 Suppressionof rust resistance genes from distantly related species in Triti-cum durum-Aegilops amphiploids, pp. 195–201 in Proceedings ofthe First Regional Conference on Yellow Rust in the Central andWest Asia and North Africa Region, Karaj, Iran, edited by R. John-son, A. Yahyaoui, C. Wellings, A. Saidi, and H. Ketata, May 8–14.

Al-Kaff, N., E. Knight, I. Bertin, T. Foote, N. Hart et al.,2008 Detailed dissection of the chromosomal region containingthe Ph1 locus in wheat Triticum aestivum: with deletion mutantsand expression profiling. Ann. Bot. (Lond.) 101: 863–872.

Appels, R., L. B. Moran, and J. P. Gustafson, 1986 The structure ofDNA from the rye (Secale cereale) NOR R1 locus and its behav-iour in wheat backgrounds. Can. J. Genet. Cytol. 28: 673–685.

Avivi, L., 1976 The effect of genes controlling different degrees ofhomoeologous pairing on quadrivalent frequency in inducedautotetraploid lines of Triticum longissimum. Can. J. Genet.Cytol. 18: 357–364.

Table 2 Genomic changes in allopolyploid wheat thought to promote and facilitate speciation (adapted from Feldman et al. 2012)

Structural Functional

1. Revolutionary changes occur during or soon after allopolyploidization,lead to diploidization and are often reproducible

Genetic• Elimination of low-copy DNA sequences• Elimination, reduction or amplification of high-copysequences

• Intergenomic invasion of DNA sequences• Elimination of rRNA and 5S RNA genes

Genetic• Gene loss/loss of function• Rewiring of gene expression through novel intergenomicinteractions

• Novel dosage responses (positive, negative, dosage compensation)• Gene suppression or activation• Transcriptional activation of transposons that may affect nearbygenes

• New transposon insertion/excisionEpigenetic

• Chromatin remodeling• Chromatin modification• Heterochromatinization• DNA methylation• Small RNA activation or repression

Epigenetic• Gene silencing or activation through changes in methylation orsmall RNAs or through chromatin modifications

• Transposon silencing or activation through demethylationor changes in small RNAs or through chromatin modifications

2. Evolutionary changes facilitated by allopolyploidy in wheat occurduring the evolution of the species by promoting speciesbiodiversity (current knowledge is limited to the geneticrather than epigenetic changes)• Chromosomal re-patterning (intra- and intergenomic translocations)• Introgression of chromosomal segments from alien genomesand production of recombinant genomes

• Intergenomic transposons invasion

• Subfunctionalizations• Neofunctionalizations• New dosage effects through copy number variation• New allelic variations

Perspectives 771

Badaeva, E. D., A. V. Amosova, T. E. Samatadze, S. A. Zoshchuk, N.G. Shostak et al., 2004 Genome differentiation in Aegilops. 4.Evolution of the U-genome cluster. Plant Syst. Evol. 246: 45–76.

Belyayev, A., O. Raskina, A. Korol, and E. Nevo, 2000 Coevolutionof A and B genomes in allotetraploid Triticum dicoccoides. Ge-nome 43: 1021–1026.

Bento, M., J. P. Gustafson, W. Viegas, and M. Silva, 2011 Sizematters in Triticeae polyploids: larger genomes have higher re-modeling. Genome 54: 175–183.

Birchler, J. A., U. Bhadra, M. P. Bhadra, and D. L. Auger,2001 Dosage-dependent gene regulation in multicellular eu-karyotes: implications for dosage compensation, aneuploid syn-dromes, and quantitative traits. Dev. Biol. 234: 275–288.

Bottley, A., G. M. Xia, and R. M. D. Koebner, 2006 Homoeologousgene silencing in hexaploid wheat. Plant J. 47: 897–906.

Boyko, E. V., N. S. Badaev, N. G. Maximov, and A. V. Zelenin,1984 Does DNA content change in the course of Triticalebreeding? Cereal Res. Commun. 12: 99–100.

Boyko, E. V., N. S. Badaev, N. G. Maximov, and A. V. Zelenin,1988 Regularities of genome formation and organization incereals. I. DNA quantitative changes in the process of allopoly-ploidization. Genetika 24: 89–97.

Buggs, R. J. A., L. Zhang, N. Miles, J. A. Tate, L. Gao et al.,2011 Transcriptomic shock generates evolutionary noveltyin a newly formed, natural allopolyploid plant. Curr. Biol. 21:551–556.

Caldwell, R. M., and L. E. Compton, 1943 Complementary lethalgenes in wheat causing a progressive lethal necrosis of seed-lings. J. Hered. 34: 67–70.

Chapman, V., T. E. Miller, and R. Riley, 1976 Equivalence of the Agenome of bread wheat and that of Triticum urattu. Genet. Res.27: 69–76.

Chaudhary, B., L. Flagel, R. M. Stupar, J. A. Udall, N. Verma et al.,2009 Reciprocal silencing, transcriptional bias and functionaldivergence of homeologs in polyploid cotton (Gossypium). Ge-netics 182: 1–15.

Chen, Z. J., and Z. Ni, 2006 Mechanisms of genomic rearrange-ments and gene expression changes in plant polyploids. Bioes-says 28: 240–252.

Driscoll, C. J., N. L. Darvey, and H. N. Barber, 1967 Effect ofcolchicine on meiosis of hexaploid wheat. Nature 216: 687–688.

Dvorak, J., 1976 The relationship between the genome of Triti-cum urattu and the A and B genomes of Triticum aestivum. Can.J. Genet. Cytol. 18: 371–377.

Eig, A., 1929 Monographisch-kritische Übersicht der Gattung Ae-gilops. Feddes Repert. 55(Beih.): 1–228.

Eilam, T., Y. Anikster, E. Millet, J. Manisterski, and M. Feldman,2008 Nuclear DNA amount and genome downsizing in naturaland synthetic allopolyploids of the genera Aegilops and Triticum.Genome 51: 616–627.

Eilam, T., Y. Anikster, E. Millet, J. Manisterski, and M. Feldman,2010 Genome size in diploids, allopolyploids, and autopolyploidsof mediterranean triticeae. J. Bot. 210: 10.1155/2010/341380.

Feldman, M., 1965a Further evidence for natural hybridizationbetween tetraploid species of Aegilops Section Pleionathera. Evo-lution 19: 162–174.

Feldman, M., 1965b Fertility of interspecific F1 hybrids and hy-brid derivatives involving tetraploid species of Aegilops SectionPleionathera. Evolution 19: 556–562.

Feldman, M., 1965c Chromosome pairing between differentialgenomes in hybrids of tetraploid Aegilops species. Evolution19: 563–568.

Feldman, M., 1966 The effect of chromosomes 5B, 5D and 5A onchromosomal pairing in Triticum aestivum. Proc. Natl. Acad. Sci.USA 55: 1447–1453.

Feldman, M., 1993 Cytogenetic activity and mode of action of thepairing homoeologous (Phl) gene of wheat. Crop Sci. 33: 894–897.

Feldman, M., 2001 The origin of cultivated wheat, pp. 1–56. InThe Wheat Book, edited by A. Bonjean, and W. Angus. LavoisierTech & Doc, Paris.

Feldman, M., and L. Avivi, 1988 Genetic control of bivalent pairingin common wheat: the mode of Ph1 action, pp. 269–279 in TheThird Kew Chromosome Conference, edited by P. E. Brandham.Kew Publishing, Richmond, Surrey, UK.

Feldman, M., and A. A. Levy, 2005 Allopolyploidy—a shapingforce in the evolution of wheat genomes. Cytogenet. GenomeRes. 109: 250–258.

Feldman, M., and A. A. Levy, 2009 Genome evolution in allopoly-ploid wheat—a revolutionary reprogramming followed by grad-ual changes. J. Genet. Genomics 36: 511–518.

Feldman, M., and A. A. Levy, 2011 Instantaneous genetic andepigenetic alterations in the wheat genome caused by allopoly-ploidization, pp. 261–270 in Transformations of Lamarckism,From Subtle Fluids to Molecular Biology, edited by S. B. Gissis,and E. Jablonka. The MIT press, Cambridge, MA.

Feldman, M., G. Galili, and A. A. Levy, 1986 Genetic and evolu-tionary aspects of allopolyploidy in wheat, pp. 83–100 in TheOrigin and Domestication of Cultivated Plants, edited by C. Bar-igozzi Elsevier, Amsterdam.

Feldman, M., F. G. H. Lupton, and T. E. Miller, 1995 Wheats, pp.184–192 in Evolution of Crop Plants, Ed. 2nd, edited by J. Smarttand N. W. Simmonds. Longman Scientific, London.

Feldman, M., B. Liu, G. Segal, S. Abbo, A. A. Levy et al.,1997 Rapid elimination of low copy DNA sequences in poly-ploid wheat: A possible mechanism for differentiation of homoe-ologous chromosomes. Genetics 147: 1381–1387.

Feldman, M., A. A. Levy, T. Fahima, and A. Korol, 2012 Genomeasymmetry in allopolyploid plants - wheat as a model. J. Exp.Bot. 63: 10.1093/jxb/errs311932.

Feldman, M., A. A. Levy, B. Chalhoub, and K. Kashkush,2012 Genome Plasticity in Polyploid Wheat, Polyploidy andGenome Evolution, edited by Soltis, P. S., and D. Soltis.Springer-Verlag, Berlin/Heidelberg (in press).

Felsenburg, T., A. A. Levy, G. Galili, and M. Feldman,1991 Polymorphism of high molecular weight glutenins inwild tetraploid wheat: Spatial and temporal variation in a nativesite. Isr. J. Bot. 40: 501–508.

Flavell, R. B., 1982 Sequence amplification, deletion and rear-rangement: major sources of variation during species diver-gence, pp. 301–323. In Genome Evolution, edited by G. A.Dover and R. B. Flavell, Academic Press, New York.

Forde, J., J. M. Malpica, N. G. Halford, P. R. Shewry, O. D. Ander-son et al., 1985 The nucleotide sequence of an HMW gluteninsubunit gene located on chromosome 1A of wheat (Triticumaestivum L.). Nucleic Acids Res. 13: 6817–6832.

Furuta, Y., K. Nishikawa, and T. Tanino, 1974 Stability in DNAcontent of AB genome component of common wheat during thepast seven thousand years. Jpn. J. Genet. 49: 179–187. doi:10.1266/jjg.49.179.

Galili, G., and M. Feldman, 1983a Genetic control of endospermproteins in wheat. 2. Variation in high molecular weight glute-nin and gliadin subunits of Triticum aestivum. Theor. Appl.Genet. 66: 77–86.

Galili, G., and M. Feldman, 1983b Diploidization of endospermprotein genes in polyploid wheats, pp. 1119–1123 in Proc. 6th

Int. Wheat Genet. Symp., Kyoto, Japan, edited by S. Sakamoto,Faculty of Agriculture, Kyoto University, Kyoto, Japan.

Galili, G., and M. Feldman, 1984 Inter-genomic suppression ofendosperm protein genes in common wheat. Can. J. Genet. Cy-tol. 26: 651–656.

Galili, G., A. A. Levy, and M. Feldman, 1986 Gene-dosage com-pensation of endosperm proteins in hexaploid wheat Triticumaestivum. Proc. Natl. Acad. Sci. USA 83: 6524–6528.

Grant, V., 1971 Plant Speciation, Columbia University Press, New York.

772 M. Feldman and A. A. Levy

Griffiths, S., R. Sharp, T. N. Foote, I. Bertin, and M. Wanous et al.,2006 Molecular characterization of Ph1 as a major chromo-some pairing locus in polyploid wheat. Nature 439: 749–752.

Ha, M., J. Lu, L. Tian, V. Ramachandran, K. D. Kasschau et al.,2009 Small RNAs serve as a genetic buffer against genomicshock in Arabidopsis interspecific hybrids and allopolyploids.Proc. Natl. Acad. Sci. USA 106: 17835–17840.

Han, F. P., G. Fedak, T. Quellet, and B. Liu, 2003 Rapid genomicexchanges in interspecific and intergeneric hybrids and allopo-lyploids of Triticeae. Genome 46: 716–723.

Han, F. P., G. Fedak, W. L. Guo, and B. Liu, 2005 Rapid andrepeatable elimination of a parental genome-specific DNA re-peat (pGc1R–1a) in newly synthesized wheat allopolyploids.Genetics 170: 1239–1245.

Hart, G. H., 1987 Genetic and biochemical studies of enzymes,pp, Wheat and wheat improvement, Ed. 2nd, edited by E. G.Heyne American Society of Agronomy, Madison, Wisconsin.

He, P., B. R. Friebe, B. S. Gill, and J. M. Zhou, 2003 Allopolyploidyalters gene expression in the highly stable hexaploid wheat. PlantMol. Biol. 52: 401–414.

Hermsen, J. G. T., 1963 Hybrid necrosis as a problem for thewheat breeder. Euphytica 12: 1–16.

Houchins, K., M. O. Dell, R. B. Flavell, and J. P. Gustafson,1997 Cytosine methylation and nucleolar dominance in cerealhybrids. Mol. Gen. Genet. 255: 294–301.

Huang, S., A. Sirikhachornkit, X. Su, J. Faris, B. Gill et al.,2002 Genes encoding plastid acetyl-CoA carboxylase and 3-phosphoglycerate kinase of the Triticum/Aegilops complex andthe evolutionary history of polyploid wheat. Proc. Natl. Acad.Sci. USA 99: 8133–8138.

Kashkush, K., M. Feldman, and A. A. Levy, 2002 Gene loss, silenc-ing and activation in a newly synthesized wheat allotetraploid.Genetics 160: 1651–1659.

Kashkush, K., M. Feldman, and A. A. Levy, 2003 Transcriptionalactivation of retrotransposons alters the expression of adjacentgenes in wheat. Nat. Genet. 33: 102–106.

Kenan-Eichler, M., D. Leshkowitz, L. Tal, E. Noor, C. Melamed-Bes-sudo et al., 2011 Wheat hand polyploidization results in de-regulation of small RNAs. Genetics 188: 263–272.

Kerber, E. R., 1964 Wheat: reconstitution of the tetraploid com-ponent (AABB) of hexaploids. Science 143: 253–255.

Kerber, E. R., and G. J. Green, 1980 Suppression of stem rustresistance in hexaploid wheat cv Canthach by chromosome7DL. Can. J. Bot. 58: 1347–1350.

Kihara, H., 1924 Cytologische und genetische studien bei wichti-gen Getreidearten mit besonderer Rücksicht ouf das Verhaltender Chromosomen und die Sterilitat in den Bastarden, Mem.Cell. Sci., Kyoto Imp. Univ., Ser. B1: 1–200.

Kihara, H., 1944 Die Entdeckung der DD-Analysatoren beim Wei-zen. Agric and Hortic (Tokyo) 19: 889–890.

Kihara, H., 1954 Considerations on the evolution and distributionof Aegilops species based on the analyzer-method. Cytologia(Tokyo) 19: 336–357.

Kihara, H., and F. A. Lilienfeld, 1949 New synthesized 6x-wheat.Proceedings of Eighth International Congress of Genetics, editedby G. B. A. R. Larsson, Stockholm, Sweden, Hereditas (Suppl.),pp 307–319.

Kihara, H., H. Yamashita, and M. Tanaka, 1959 Genomes of 6xspecies of Aegilops. Wheat Inf. Ser. 8: 3–5.

Kimber, G., and M. Feldman, 1987 Wild wheats: an introduction.pp. 1–142, Special Report 353. College of Agriculture, Colum-bia, Missouri.

Kimber, G., and K. Tsunewaki, 1988 Genome symbols and plasmatypes in the wheat group, in pp. 1209-1211, Proceedings of the7th International Wheat Genetics Symposium, edited by T. E.Miller and R.M. D. Koebner, Cambridge, UK.

Kraitshtein, Z., B. Yaakov, V. Khasdan, and K. Kashkush,2010 Genetic and epigenetic dynamics of a retrotransposonafter allopolyploidization of wheat. Genetics 186: 801–889.

Le Comber, S. C., M. L. Ainouche, A. Kovarik, and A. R. Leitch,2010 Making a functional diploid: from polysomic to disomicinheritance. New Phytol. 186: 113–122.

Levy, A. A., and M. Feldman, 1988 Ecogeographical distributionof HMW glutenin alleles in populations of the wild tetraploidwheat Triticum turgidum var. dicoccoides. Theor. Appl. Genet.75: 651–658.

Levy, A. A., and M. Feldman, 2002 The impact of polyploidy ongrass genome evolution. Plant Physiol. 130: 1587–1593.

Levy, A. A., and M. Feldman, 2004 Genetic and epigenetic repro-gramming of the wheat genome upon allopolyploidization. Biol.J. Linn. Soc. Lond. 82: 607–613.

Levy, A. A., G. Galili, and M. Feldman, 1988 Polymorphism andgenetic control of high molecular weight glutenin subunits in wildtetraploid wheat Triticum turgidum var. dicoccoides. Heredity 61:63–72.

Liu, B., G. Segal, J. M. Vega, M. Feldman, and S. Abbo,1997 Isolation and characterization of chromosome-specificsequences from a chromosome arm genomic library of commonwheat. Plant J. 11: 959–965.

Liu, B., J. M. Vega, G. Segal, S. Abbo, M. Rodova et al.,1998a Rapid genomic changes in newly synthesized amphi-ploids of Triticum and Aegilops. I. Changes in low-copy non-coding DNA sequences. Genome 41: 272–277.

Liu, B., J. M. Vega, and M. Feldman, 1998b Rapid genomicchanges in newly synthesized amphiploids of Triticum and Ae-gilops. II. Changes in low-copy coding DNA sequences. Genome41: 535–542.

Ma, X.-F., and J. P. Gustafson, 2005 Genome evolution of allopo-lyploids: a process of cytological and genetic diploidization. Cy-togenet. Genome Res. 109: 236–249.

Ma, X.-F., and J. P. Gustafson, 2006 Timing and rate of genomevariation in triticale following allopolyploidization. Genome 49:950–958.

Ma, X.-F., P. Fang, and J. P. Gustafson, 2004 Polyploidization-in-duced genome variation in triticale. Genome 47: 839–848.

Maan, S. S., 1977 Fertility of amphiploids in Triticinae. J. Hered.68: 87–94.

Mac Key, J., 1954 Mutation breeding in polyploid cereals. ActaAgriculturae Scandinavica 4: 549–557.

Mac Key, J., 1958 Mutagenic response in Triticum at differentlrvels of ploidy, pp. 88–111, in Proc. 1st Inter Wheat GeneticsSymp, edited by C. B. Jenkins, Winnipeg, Manitoba, Canada.

Maestra, B., and T. Naranjo, 1999 Structural chromosome differ-entiation between Triticum timopheevii and T. turgidum and T.aestivum. Theor. Appl. Genet. 98: 744–750.

Masterson, J., 1994 Stomatal size in fossil plants: Evidence forpolyploidy in majority of angiosperms. Science 264: 421–424.

McClintock, B., 1984 The significance of responses of the genometo challenge. Science 226: 792–801.

McFadden, E. S., and E. R. Sears, 1944 The artificial synthesis ofTriticum spelta. (Abstr.). Rec. Genet. Soc. Amer. 13: 26–27.

McFadden, E. S., and E. R. Sears, 1946 The origin of Triticumspelta and its free-threshing hexaploid relatives. J. Hered. 37:81–89, 107–116.

Morris, R., and E. R. Sears, 1967 The cytogenetics of wheat andits relatives, pp. 19–87 in Wheat and Wheat Improvement, editedby K. S. Quisenberry, and L. P. Reitz. American Society of Agron-omy, Madison, WI.

Nishikawa, K., and Y. Furuta, 1969 DNA content per nucleus inrelation to phylogeny of wheat and its relatives. Jpn. J. Genet.44: 23–29 10.1266/jjg.44.23.

Ohno, S., 1970 Evolution by Gene Duplication, Springer, New York.

Perspectives 773

Okamoto, M., 1957 Asynaptic effect of chromosome V. Wheat InfServ. 5: 6.

Okamoto, M., and N. Inomata, 1974 Possibility of 5B-like effect indiploid species. Wheat Inf Serv 38: 15–16.

Otto, S. P., and J. Whitton, 2000 Polyploid incidence and evolu-tion. Annu. Rev. Genet. 34: 401–437.

Ozkan, H., A. A. Levy, and M. Feldman, 2001 Allopolyploidy-in-duced rapid genome evolution in the wheat (Aegilops-Triticum)group. Plant Cell 13: 1735–1747.

Ozkan, H., M. Tuna, and K. Arumuganathan, 2003 Nonadditivechanges in genome size during allopolyploidization in the wheat(Aegilops-Triticum) group. J. Hered. 94: 260–264.

Payne, P. I., L. M. Holt, and C. N. Law, 1981 Structural and ge-netical studies on the high-molecular-weight subunits of wheatglutenin. Theor. Appl. Genet. 60: 229–236.

Pikaard, C. S., 2000 The epigenetics of nucleolar dominance.Trends Genet. 16: 495–500.

Ramsey, J., and D. Schemske, 1998 Pathways, mechanisms, andrates of polyploid formation in flowering plants. Annu. Rev.Ecol. Syst. 29: 467–501.

Riley, R., 1960 The diploidization of polyploid wheat. Heredity15: 407–429.

Riley, R., and V. Chapman, 1958 Genetic control of the cytologi-cally diploid behavior of hexaploid wheat. Nature 182: 713–715.

Salina, E. A., O. M. Numerova, H. Ozkan, and M. Feldman,2004 Alterations in subtelomeric tandem repeats duringearly stages of allopolyploidy in wheat. Genome 47: 860–867.

Sax, K., 1927 Chromosome behavior in Triticum hybrids. pp.1267–1284 in Verhandlungen des V. Intern. Kongresses für Ver-erbungswissenschaft, Berlin, Germany.

Sears, E. R., 1948 The Cytology and Genetics of the Wheats andtheir Relatives. Adv. Genet. 2: 239–270.

Sears, E. R., 1954 The aneuploids of common wheat. MissouriAgric. Exp. Stn. Res. Bull. 572: 1–58.

Sears, E. R., 1969 Wheat cytogenetics. Annu. Rev. Genet. 3: 451–468.Sears, E. R., 1972 The nature of mutation in hexaploid wheat.

Symp. Biol. Hung. 12: 73–82.Sears, E. R., 1976 Genetic control of chromosome pairing in

wheat. Annu. Rev. Genet. 10: 31–51.Sears, E. R., 1984 Mutations in wheat that raise the level of mei-

otic chromosome pairing, pp. 295–300 in Gene Manipulation inPlant Improvement, edited by J. P. Gustavson Stadler Sympoium,Columbia, MO.

Sears, E. R., and M. Okamoto, 1958 Intergenomic chromosome rela-tionships in hexaploid wheat. Proc. X Int. Congr. Genet. 2: 258–59.

Shaked, H., K. Kashkush, H. Ozkan, M. Feldman, and A. A. Levy,2001 Sequence elimination and cytosine methylation arerapid and reproducible responses of the genome to wide hybrid-ization and allopolyploidy in wheat. Plant Cell 13: 1749–1759.

Soltis, D. E., and P. S. Soltis, 1993 Molecular data and the dy-namic nature of polyploidy. Crit. Rev. Plant Sci. 12: 243–273.

Stebbins, G., L., 1950 Variation and Evolution in Plants. ColumbiaUniversity Press, New York.

Tate, J. A., P. Joshi, K. A. Soltis, P. S. Soltis, and D. E. Soltis,2009 On the road to diploidization? Homoeolog loss in inde-

pendently formed populations of the allopolyploid Tragopogonmiscellus (Asteraceae). BMC Plant Biol. 9: 80 10.1186/1471–2229–9-80.

Thompson, D. A., M. M. Desai, and A. W. Murray, 2006 Ploidycontrols the success of mutators and nature of mutations duringbudding yeast evolution. Curr. Biol. 16: 1581–1590.

Tsunewaki, K., 1970 Necrosis and chlorosis genes in commonwheat and its ancestral species. Seiken Ziho 22: 67–75.

Van de Peer, Y., S. Maere, and A. Meyer, 2009 The evolutionarysignificance of ancient genome duplications. Nat. Rev. Genet.10: 725–732.

Van Slageren, M. W., 1994 Wild Wheats: A Monograph of AegilopsL. and Amblyopyrum (Jaub. & Spach) Eig (Poaceae), AgriculturalUniversity, Wageningen, The Netherlands.

Vega, J. M., and M. Feldman, 1998 Effect of the pairing gene Ph1on centromere misdivision in common wheat. Genetics 148:1285–1294.

Vega, J. M., S. Abbo, M. Feldman, and A. A. Levy, 1994 Chromosomepainting in plants: in situ hybridization with a DNA probe froma specific microdissected chromosome arm of common wheat. Proc.Natl. Acad. Sci. USA 91: 12041–12045.

Vega, J. M., S. Abbo, M. Feldman, and A. A. Levy, 1997 Chromosomepainting in wheat, pp. 319–332 in Chromosomes Today, Vol. 12,edited by N. Henriques-Gil, J. S. Parker, and M. J. Puertas. Chap-man and Hall, London.

Von Tschermak, E., and H. Bleier, 1926 Über fruchtbare Aegilops-Weizenbastarde, Der Deutsch. Bot Ges 44: 110–132.

Waines, J. G., 1976 A model for the origin of diploidizing mech-anisms in polyploid species. Am. Nat. 110: 415–430.

Waines, J. G., and P. I. Payne, 1987 Electrophoretic analysis ofthe high-molecular-weight glutenin subunits of Triticum mono-coccum, T. urartu and the A genome of bread wheat (T. aesti-vum). Theor. Appl. Genet. 74: 71–76.

Wall, A. M., R. Riley, and M. D. Gale, 1971 The position of a locuson chromosome 5B of Triticum aestivum affecting homoeolo-gous meiotic pairing. Genet. Res. 18: 329–339.

Weissman, S., M. Feldman, and J. Gressel, 2005 Sequence evi-dence for sporadic intergeneric DNA introgression from wheatinto wild Aegilops species. Mol. Biol. Evol. 22: 2055–2062.

Wicker, T., K. F. X. Mayer, H. Gundlach, M. Martis, B. Steuemagelet al., 2011 Frequent gene movement and pseudogene evolu-tion is common to the large and complex genomes of wheat,barley, and their relatives. Plant Cell 23: 1706–1719.

Yaakov, B., and K. Kashkush, 2011 Massive alterations of themethylation patterns around DNA transposons in the first fourgenerations of a newly formed wheat allohexaploid. Genome54: 42–49.

Yousafzai, F. K., N. Al-Kaff, and G. Moore, 2010 Structural andfunctional relationship between the Ph1 locus protein 5B2 inwheat and CDK2 in mammals. Funct. Integr. Genomics 10:157–166.

Zohary, D., and M. Feldman, 1962 Hybridization between amphi-diploids and the evolution of polyploids in the wheat (Aegilops-Triticum) group. Evolution 16: 44–61.

Communicating editor: A. S. Wilkins

774 M. Feldman and A. A. Levy