Eugene Metkovsky,Viktor Melnik*,Mrt Rodriguez-Quijno,Vldimir UpelniekJose Mri Crrillo

aUnidad de Genetica,Universidad Politecnica de Madrid,Madrid,Spain

bVavilov Institute of General Genetics Russian Academy of Sciences,Moscow,Russia

Keywords:Gliadin alleles Genetic variation Gene geography Intra-varietal non-uniformity Allopolyploid Triticum aestivum

ABSTRACT A new,improved version of the catalog of 182 alleles at the six Gli loci of common wheat(T.aestivum L.)shown in electrophoregrams of 128 standard genotypes was used for analysis of 1060 cultivars and lines bred in the 20th century.The most frequent alleles in the studied germplasm occurred with frequencies of 18%–40%,with 30 unique alleles,one in each cultivar.Extremely high genetic diversity was found(average H for the six main Gli loci was 0.870±0.046),nearly identical in winter(H=0.831)and spring(H=0.856)wheats but differing among 28 groups of cultivars released in 22 countries.Each country or region was characterized by its own specific set of the most frequent Gli alleles,and the 28 cultivar groups formed five main relationship clusters if polymorphism at the six Gli loci was considered.However,different levels of similarity between groups of cultivars were found if polymorphism of the A,B,or D genomes of common wheat was tested separately.In general,the 20th century germplasm of common wheat was differentiated and structured by country or region and cultivar type(spring or winter).Each elemental genome(in particular,A and D)contributed to the structure of the polymorphism studied.We propose that the structure of the wheat germplasm was a result of natural selection under the ecoclimatic conditions of cultivation specific to each country or region.As many as 27.4%of cultivars studied violated the requirement of the DUS rules for uniformity,being represented by mixtures of two or more closely related genotypes.However,the composition of a cultivar as a set of related but different genotypes may contribute to its adaptivity,and thereby to the known high plasticity of common wheat.

1.Introduction

Gliadin is an alcohol-soluble storage protein that accounts for about 40%,by weight,of all proteins of wheat flour.Comprising a considerable part of wheat-based products,gliadin is one of the major protein components of the human diet.Moreover,as one of the most polymorphic proteins in nature,gliadin may serve in genetic studies to resolve some questions of great importance to science and to agricultural practice.

The complex genome of allohexaploid wheat,Triticum aestivum L.,consists of three related elemental genomes,A,B,and D,each with seven chromosomes[1].The availability of nulli-tetrasomic lines[2]allowed the identification of chromosomes 1A,1B,1D,6A,6B,and 6D as those encoding gliadins.There are six main gliadin-coding(Gli)loci situated distally on the short arms of these chromosomes(loci Gli-A1,Gli-B1,Gli-D1,Gli-A2,Gli-B2 and Gli-D2,respectively)[3,4].

In one wheat kernel,it is possible to identify 20–25 electrophoretic components (bands)using a standard method,one-dimensional acid polyacrylamide gel electrophoresis(APAGE)[5],and up to several tens of gliadin polypeptides(gliadins)using variants of two-dimensional electrophoresis [6].It has been shown that gliadin electrophoregrams are cultivar-specific and do not depend on plant growing conditions(eco-climatic conditions,level of fertilization and/orwatering),grain protein content,or incomplete kernel maturation[6].

Conventionally,the gliadin electrophoregram of a wheat genotype is divided into four zones depending on electrophoretic mobility(EM)of bands,α (the fastest bands),β,γ,and ω,using nomenclature suggested by Woychik et al.[6].Three main structure types of gliadin polypeptides(and their encoding genes)are known:α-,γ-,and ω-gliadins[3,7].

By analysis of segregating progenies of crosses between common wheat cultivars,it was first discovered by Sozinov,Poperelya,and theirco-workers (AgriculturalInstitute,Odessa,Ukraine)that each Gli locus encodes several different gliadin polypeptides(a “block”)inherited together as one Mendelian trait.Moreover,multiple allelism at each Gli locus has been described,such that alleles differ in number and EM of gliadin bands composing a block[8].Catalogs of gliadin alleles identified using one-dimensional electrophoresis in acid starch gel have been published[8,9].

There are two important peculiarities of gliadin polymorphism.First,a block may consist of bands located in different zones of an electrophoregram(e.g.,in β- γ-,and ω-zones).Second,there are blocks encoded by alleles at the same Gli locus which lack a single electrophoretic band in common.Indeed,the DNA sequence of a Gli locus may include expressed genes of different types(for example, γ-and ωgliadin-encoding genes)as well as gliadin pseudogenes and non-gliadin DNA[10–12].

Using APAGE,a total of 111 alleles at the six Gli loci have been identified and shown using 49 standard wheat genotypes[13].More alleles were discovered later[14–21]and are included in a list of 939 common wheat cultivars studied(www.aaccnet.org)[22].Thus,the need has emerged for a new catalog that includes all Gli alleles identified to date and shows their encoding blocks.

The permanent use in breeding programs of a limited set of“successful”parental genotypes may cause,theoretically,a decrease in the level of genetic variation(genetic erosion)of wheat germplasm and thereby complicate wheat improvement[23,24].To evaluate genetic variation in common wheat,one needs reliable genetic markers for description of cultivar genotypes.It has been proposed that Gli alleles might be used for this purpose[6,8].

The aim of the present work was to develop a new and enlarged version of the catalog of Gli alleles and to use it for description and analysis of 20th-century common wheat germplasm.This effort revealed that wheat genetic polymorphism was,in general,extremely high and,besides,was highly structured with respect to countries and regions,winter and spring wheats,and elemental genomes A,B,and D.It is suggested that the main contributors to the known plasticity of common wheat have been the allopolyploidy of the species and the non-uniformity of cultivars.

2.Materials and methods

2.1.Grain materials

Grain samples(one sample,from 10 kernels up to 0.5 kg of grain)of common wheat were obtained mainly from genetic and/or breeding labs in the country of origin of cultivars:Agricultural Research Institute(Martonvásár,Hungary),CSIRO Plant Industry (Sydney, Australia), INRA Station d'Amelioration des Plantes(Clermont-Ferrand,France),Institute of Agriculture(Omsk,Russia),Institute of Agriculture(Sadovo,Bulgaria),Institute of Agriculture of South-East(Saratov,Russia),Institute of Cytology and Genetics(Novosibirsk,USSR),Institute of Selection and Genetics(Odessa,Ukraine),InstituteofSmallCrops(Kragujevac,Serbija),Istituto Sperimentale per la Cerealicoltura (S.Angelo Lodigiano,Italy),Plant Breeding Institute(Cambridge,UK),Research Applied Technology Department,NIAB(Cambridge,England),Shortandy Agricultural Station(Kazakhstan),Scientific and Breeding Centre “Don”(Rostov,Russia),University of Ljubljana (Slovenia),University ofManitoba (Canada),Universidad Politecnica de Madrid(Spain),and some others.Grain samples of some cultivars were obtained from two or more sources,but a few of them came from large international germplasm collections.

Names of cultivars follow http://www.wheatpedigree.net/(Table S1,Fig.S1).This website was also used to know the pedigree and winter/spring habit of a cultivar.

2.2.Methods

One-dimensional acid electrophoresis(APAGE)of gliadin from single kernels was performed by the authors in several wheat genetics laboratories of Europe and Australia under the conditions usually used in these labs,but always following a standard procedure[5]with modifications(vertical gel,higher voltage)(discussed in[6]).The results of identification of Gli alleles in cultivars bred in Canada[14],France[15],Italy[16],Russia[17],Spain[18,19],UK[17],Ukraine[20],and the former Yugoslavia[21],and at http://www.aaccnet.org/[22]were all revised in the present study in accordance with the new catalog of Gli alleles.The results of our analysis of Gli alleles in 120 further cultivars performed at the Universidad Politecnica,Madrid and in the Institute of General Genetics,Moscow have not previously been published.

Blocks of gliadin electrophoretic bands inherited together(Gli alleles)in the set of standard genotypes(Fig.S1)were identified by analysis of segregating progenies of crosses between cultivars and by comparison of gliadin electrophoregrams of natural biotypes of non-uniform cultivars with those of cultivars closely related in their pedigree[6,22].Gli alleles in each new cultivar were routinely identified by comparison of electrophoregrams(side-by-side in the same APAGE slab run)of a cultivar under study and of the cultivar Bezostaya-1 and appropriate standard cultivars.In fact,thisapproach useseach band ofthe electrophoregram of Bezostaya-1(the genetic control of each electrophoretic band of this cultivar is well studied[6,22])as an individual marker of EM for a given APAGE run.

Genetic variation(H)at a locus was calculated as H=1-Σp2,where p is the frequency of an allele at this locus in a group of cultivars[25].Cluster analysis was performed as described[26].More details are presented in Box 1.

3.Results and discussion

3.1.Reproducibility of gliadin APAGE electrophoregrams

APAGEelectrophoregramsofgliadin of128selected(standard)genotypes are shown in Fig.S1.Comparison of pairs of electrophoregrams of some of these genotypes(Asiago, Caprock, Centauro, Cluj-650, De-Carolis,Dneprovskaya-521, Gallo, Ghurka, Hereward, Kanzler,Krasnodonka,Lesostepka-75,Longbow,Marquis,Mercia,Odesskaya-16,Salmone,Silvana,Skala,Tovstik-line-awnless)obtained in different runs confirmed the satisfactory reproducibility of the APAGE procedure used.

The independence of a genotype-specific electrophoregram of gliadin from plant growing conditions was confirmed.In the present study,52 cultivars were represented by more than one grain sample obtained from different countries,and 40 cultivars by two independent grain samples from the same country. Identical gliadin electrophoregrams(and thus the same Gli allelic compositions)were shown by,for example,two grain samples of the Canadian cultivar Katepwa(alleles at the six Gli loci are m,d,j,m,c,h,respectively)obtained from Australia and Canada,and by three grain samples of the French cultivar Pistou(f,f,b,o,o,n)received from France,Italy,and Spain.See Box 2 for more details.

Box 1 Analysis of intra-varietal non-uniformity.

3.2.New catalog of Gli alleles and their frequencies in the germplasm of 20th-century common wheat

In the collection of 1060 cultivars,193 alleles were discovered,including nine different new alleles(one differing from known alleles but not catalogued)and two null alleles.Each of the 182 catalogued alleles is shown as its encoded block identified in the electrophoregram of one or more standard cultivars(Fig.S1).

Important changes in the set of alleles at the Gli-A1 locus differentiate the new catalog from the old one.Two years after publication of the catalog[13],so-called “minor”Gli loci were mapped distally on the short arm of chromosome 1A:Gli-A5[27]and Gli-A6(the most distal)[28].Minor Gli loci control two(allele Gli-A5b),and one(allele Gli-A6b)ω-gliadins in the APAGE electrophoregram whereas alleles Gli-A5a and Gli-A6a are null.Severalstudies haveshown thatthegenetic distances between the three loci did not exceed 5%[6,22].

Combinations between alleles at the Gli-A1 and Gli-A5 or Gli-A6 loci produce different wheat genotypes.In the new catalog,we added ω-gliadins controlled by Gli-A5b or Gli-A6b to some Gli-A1-controlled blocks to produce a new variant(“allele”)of the distal part of chromosome 1A.For example,Gli-A1af+Gli-A6a(null)=Gli-A1af(standard cultivars Basalt,Insignia,Fig.S1),and Gli-A1af+Gli-A6b=Gli-A1f(Maris-Freeman,Mironovskaya-808)(more details are shown in Box 3).Owing to involvement of the Gli-A5 and Gli-A6 loci,several new “allelic variants”at the Gli-A1 locus were obtained,broadening for practical purposes the reported polymorphism of gliadin in common wheat.

In total,1060 common wheat cultivars from more than 20 countries of Europe,North America,and Australia and several regions of the former USSR were analyzed(Table S1)and a highly disproportionate distribution of alleles in the germplasm studied was discovered:the most frequent allele,Gli-D1b,occurred in 40% of all cultivars,while 30 unique catalogued alleles were present in one cultivar each(Table1).Moreover,seven unique alleles,Gli-B1u(Fig.S1,cultivar Negrillo),Gli-D1n(Este),Gli-D1p(Skorospelka-Uluchshennaya,bt 2),Gli-A2ad(Rannyaya-73),Gli-A2al(Aquila),Gli-B2x and Gli-D2ad(both in Skorospelka-Uluchshennaya,bt 1),were each discovered only in one of the genotypes of a non-uniform cultivar(Table S1).

Box 2 Reproducibility of gliadin APAGE pattern.

Table 1–Number of all Gli alleles listed in Table S1 including news and nulls(All alleles),number of catalogued Gli alleles(Catalogued),number of catalogued alleles found in only one cultivar or genotype(Uniques),the most frequent alleles(Frequency),and genetic variation(H)in 20th-century wheat germplasm.

Box 3

The average genetic variation H in the collection(Table 1)was extremely high(H=0.870±0.046),being highest at the Gli-A2 and Gli-B2 loci and lowest at Gli-D1(Table 1).

3.3.Spontaneous mutations at Gli loci discovered through analysis of gliadin electrophoregrams

Analysis of individual kernels of a cultivar allows the discovery of single kernels carrying changes in their electrophoregrams caused by mutations at the Gli loci[29].In the set of cultivars studied,three types of mutational changes of electrophoregrams were found.First was a change in EM of one gliadin band(for example,Fig.S1,cultivars Bezostaya-1,Farneto,Omskaya-17,lines M);Second was the disappearance of only one(cultivarAlbidum-3700,lineM)orseveralcomponentsofaknown block(cultivars Leningradka,Omskaya-12,line M),and third was the presence of null mutants.

Obviously,a spontaneous mutation appears firstly in the heterozygous state(Fig.S1,cultivar Rescler,line M).The genotype carrying a mutation at one of its Gli loci may be casually involved in the breeding process.As a result,a new(daughter)cultivar may receive a new(mutant)Gli allele.New(mutant)alleles are normally inherited[30].

Judging from the known structures of different types of gliadin-encoding genes[7,11,31–33]one may suggest that mutation changing the EM of a gliadin polypeptide may be a consequence of change in the number of repeats of repeating DNA present in the γ-gliadin genes,or change in the length of the polyglutamine domain of α-gliadin genes.Indeed,analysis of some pairs of known Gli alleles differing in the EM of one component of their controlled blocks confirmed that increase of EM of the gliadin polypeptide was always accompanied by decrease in its molecular weight(length of the polypeptide)[34].A mutation changing the EM of a gliadin polypeptide(electrophoretic band)generates a new Gli allele and thereby increases the genetic polymorphism of wheat.On the other hand,disappearance of one of the bands of a known block may be plausibly explained as the conversion of the corresponding gene into a pseudogene(such as may occur as a consequence of,for example,transposable element activity).Pseudogenes represent a large fraction of all Gli loci studied[10–12,35].

The loss of all bands of a block was caused by null mutations(forexample,cultivarsSkorospelka-Uluchshennaya (bt3),Saratovskaya-39,Fig.S1,lines M).Spontaneous null mutants have been described previously at each of the six main Gli loci of the cultivar Saratovskaya-29[29].A double-null mutant(Gli-B1null,Gli-D1null)was discovered in a grain sample of the cultivar Tselinogradka(Fig.S1,lineDM).Asarule,nullmutantslacksome DNA or chromosomal material[36,37].However,even the registered French cultivars Darius[15]and Touzelle(Fig.S1)were found to be a Gli-D1null mutant and a double Gli-B1null,Gli-D1null mutant,respectively.Also,common wheat cultivars(75 of the 1060 studied,Table S1)carried allele Gli-B1l,the result of a substitution of wheat 1BS chromosome arm with alien(rye)genetic material(the 1BL/1RS translocation).

Common wheat,being a polyploid,may survive despite substantial changes in chromosome arrangement,including aneuploidy,because the activity of homoeologous genes may compensate for the loss of genetic material[2,38].The discovery of registered cultivars lacking some genetic material(null mutants at Gli loci)emphasizes once more the high level of plasticity of common wheat.

3.4.Intra-varietal non-uniformity of common wheat of the 20th century

Each wheat cultivar must satisfy three main requirements:to be distinct from other cultivars,uniform,and stable(the socalled DUS rules).Each new cultivar passes a complex series of tests to satisfy the DUS rules and thereby to protect the intellectual property of the breeder[39,40].

However,in performing kernel-by-kernel analysis of grain samplesofcommonwheat,wediscoveredthatmany cultivars did not satisfy one of the requirements of DUS,uniformity.Unexpectedly many cultivars(27.4%of all studied,Table S1)consisted of mixtures(populations)of related genotypes differing by alleles at least one Gli locus.Such cultivars we have called “heterogeneous”.Cases of intravarietal non-uniformity of common wheat cultivars have been previously described[41–43].

Natural biotypes(genotypes)of a cultivar may appear as a consequence of crosses of wheat genotypes performed by a breeder.Thus,genotypes of a non-uniform cultivar would carry alleles at the Gli loci which have come from parental cultivars[22].For example,the grain sample of the heterogeneous Italian cultivar Flaminio(alleles at the six Gli loci,a,g,f,o+g,o,n,respectively)was a mixture of two types of kernel(genotypes)differing at the Gli-A2 locus,a,g,f,o,o,n,and a,g,f,g,o,n.This cultivar had two parents,cultivars Frassineto(f,g,k,o,o,n)and Funo(a,e,f,g,j,a)(Table S1).Obviously,Flaminio inherited alleles Gli-B1g,Gli-A2o,Gli-B2o,Gli-D2n from Frassineto and alleles Gli-A1a,Gli-D1f,Gli-A2g from Funo.

The non-uniform English cultivar Galahad(ab,g,b,l,g+l,g)had three parents,cultivars Maris-Hobbit(o+ab,f,b,l,g,g),Maris-Durin(ab,g,b,ak,g,a)and Joss-Cambier(o,f,b,p,l,n).It consisted of two genotypes differing at the Gli-B2 locus(Table S1).Obviously,the cultivar Galahad inherited its Gli-B2g alleles from Maris-Hobbit or Maris-Durin and its Gli-B2l alleles from Joss-Cambier.

Siberian cultivar Amurskaya-75(m+j,e+d,a,q+m,o,h)was a mixture of eight related genotypes(first,m,e,a,q,o,h,second,j,e,a,q,o,h,third,m,d,a,q,o,h,and so on)and was developed from a cross between cultivars Thatcher(m,d,j,m,c,h)and Lutescens-62(j,e,a,q,o,a)(Table S1).Obviously,Amurskaya-75 inherited its Gli-A1m,Gli-B1d,Gli-A2m,and Gli-D2h alleles from Thatcher and its Gli-A1j,Gli-B1e,Gli-D1a,GliA2q and Gli-B2o alleles from Lutescens-62.It is noteworthy that,in each allelic pair of blocks,one of them could not be transformed into another in one step.For example,one may compare the band compositions of blocks encoded by alleles Gli-A1m(standard cultivar Marquis)and Gli-A1j(Lutescens-62)(Fig.S1).

Some well-known cultivars from different countries used in enormous numbers of breeding programs around the world during the 20th century were non-uniform and consisted of sets of genotypes differing at one or more Gli loci.Examples include Avrora,Bezostaya-1,Charodeika,Damiano,Ghurka,Holdfast,Insignia,Kavkaz,Marimp-8,Maris-Hobbit,Mercia,Odesskaya-16, Partizanka, Pinnacle, Red-Fife, Rusalka,Sadovo-Super, San-Pastore, Selkirk, Siete-Cerros-66,Skorospelka-3,Super-Zlatna,Vilmorin-27,and VPM-1(Table S1).The international standard cultivar Chinese Spring[2]may also be non-uniform[44].Each biotype(genotype)of a cultivar found to be heterogeneous at a Gli locus may consist of a set of genotypes that could be differentiated by use of DNA markers[45].

ThespringcultivarSaratovskaya-29consisted ofsix genotypes differing by alleles at the Gli-A1,Gli-A2,and Gli-B2 loci[46].The relative frequencies of these genotypes in grain samples of Saratovskaya-29 has been maintained over decades if the cultivar was grown in its native Saratov region of Russia.In contrast,drastic changes in relative frequencies of genotypes occurred when the cultivar was grown in other ecoclimatic regions of cultivation.Moreover,in different regions,different genotypes of Saratovskaya-29 began to dominate[47]changing some important characters of the cultivar as a whole[6,48].Thus,this cultivar,the best spring cultivar of the USSR over three decades of the 20th century[49],may be called,in a sense,unstable,besides being non-uniform,because the general parameters of the cultivar may depend on the relative concentrations of certain genotypes in its populations[48].

Thus,a non-uniformity that violates the severe rules of DUS may be valuable for a commercial common wheat cultivar,helping it to adapt,by changing concentrations of its different genotypes,to different environments and conditions of growing.Polyploidy broadens the opportunities to maintain somelevelofnon-uniformity,owingtothe homoeology among chromosomes and genes.Thus,nonuniformity of cultivars may be considered an important contributor to the high plasticity of common wheat.

On the other hand,the presence of a set of genotypes within a cultivar may severely complicate an interpretation of results obtained in any experimental work because the genotypes composing a non-uniform cultivar may differ in parameters tested in the experiment[6,41–43,48].Unfortunately,the possible heterogeneity of a cultivar under study has been considered in few reports describing polymorphism analyses of common wheat using DNA markers.In one study,the use of microsatellites as DNA markers revealed that 74%of Bulgarian common wheat cultivars studied were heterogeneous[50].Analysis of Gli alleles showed 48%of Bulgarian cultivars in the present study to be non-uniform(calculated from Table S1).

In general,in ouropinion,theobviousdiscrepancy between the DUS requirement of uniformity and the existence of intra-varietal non-uniformity in many cultivars as well as a possible contribution of the internal heterogeneity of cultivars to plasticity of common wheat attracts insufficient attention by scientists and breeders working with this self-pollinated species.

3.5.Differentiation of genotypes and genetic diversity between countries and regions

None of the 182 catalogued alleles occurred at least once in each of the 28 groups of cultivars studied,and the distribution of Gli alleles among countries and regions was highly unequal:each country or region produced cultivars carrying their own specific sets of the most frequent Gli alleles(Table 2).

Some alleles may be claimed as a characteristic of cultivars bred in a country or region,for example,Gli-A2c,Gli-A2a,and Gli-D2w in Australia;Gli-A1m,Gli-D1j,Gli-A2m,and Gli-D2h in Canada;“b”alleles at each of the six Gli loci in Bulgaria and Romania;Gli-A1o,Gli-B1f,and Gli-A2g in France;Gli-A2h,and Gli-B2ae in Germany;Gli-D1k,Gli-A2g,and Gli-B2o in Italy;Gli-A2e and Gli-B2e in Croatia;Gli-A2n in the “Don”breeding center;Gli-A1f,Gli-B1e,Gli-D1a,Gli-A2q,Gli-B2s,and Gli-D2e in the Saratov region of Russia;Gli-A1k and Gli-B2k in eastern

Siberia;Gli-A2ak and Gli-B2d in Kazakhstan;Gli-A1af,Gli-A2l,Gli-B2l,and Gli-D2g in the UK.Allele Gli-A1a was the most frequent in the neighboring countries Croatia,Italy,and Serbia(with frequencies of occurrence 39.1%,57.7%,and 43.0%,respectively)(Table 2).

Table 2–Analysis of groups of cultivars bred in different countries and regions:spring or winter habit(s/w),number of cultivars in a group(N),Gli alleles occurring with relative frequency greater than 30%(in parentheses,a relative frequency of 25%–30%),and genetic variation(H).

Fig.1–Relationships of gliadin genotypes among 28 groups of cultivars of common wheat bred in 22 countries(Table 2).Results of identification of alleles at all six Gli loci(Table S1)were used.About the North group see Table 2,footnotes.

The partitioning of wheat genotypes among countries and regions was further confirmed by analysis of their relationships(Fig.1).There were five clusters of groups of cultivars released in the 20th century:first,winter wheat from northwest Europe(France,Germany,Holland,UK);second,spring wheat from countries not including the USSR(Australia,Canada,Mexico,Portugal);third,cultivars of Italy,Spain(both spring and winter),and Croatia;fourth,winter wheat from southeastern Europe(Bulgaria,Don and Krasnodar regions of Russia;Hungary,Rumania,Serbia,and Ukraine);and,fifth,spring cultivars from the former USSR and a small group,North,of 10 cultivars from northern Europe.The long branches of the trees(Fig.1)of these cultivar groups point to the profound differentiation of polymorphism of 20th-century common wheat germplasm among countries and regions.A similar differentiation of polymorphism of microsatellite DNA markers in a set ofEuropean countries was described earlier[51].This differentiation between countries(the structure of polymorphism)might be caused by the necessity of plantstoadapttoeco-climaticconditionsofcultivationthatdiffer amongcountriesand region;inotherwords,by naturalselection,unavoidablyactingtogetherwithartificialselectionperformedby a breeder.Strong genetic differentiation and association of genotypes with environmental factors has been shown in wheat by Nevo et al.[52,53].

Genetic variation also differed among groups of cultivars,being highest in Spanish wheat and lowest in the group of cultivars bred in the Saratov region of Russia(Table 2).The high level of genetic variation in Spanish wheats may be explained by the high variation among eco-climatic conditions suitable for wheat cultivation in that country[18].In contrast,only a restricted number of genotypes are able to tolerate the rigorous environmental conditions of the Saratov region[49].Owing to the differentiation of genotypes among groups of cultivars,the level of genetic diversity H was lower in each group(Table 2)than in the collection of 1060 cultivars(Table1).Eco-climaticand otherenvironmental factors dramatically affect levels of genetic variation[53].Selection of genotypes better adapted to specific environments leads to a reduction in genetic variation[54].

3.6.Comparison of Gli genotypes of spring and winter wheats

It has been suggested that genotypes of spring and winter wheat might differ[55,56].Indeed,the genotypes of the spring and winter cultivars studied shared a few of the most frequent Gli alleles,Gli-B1b being one of them(Table 2).This allele was present in several important cultivars of the 20th century,including the spring wheats Gabo(Australia),Marquis and Kitchener(Canada),and Skala(USSR),and the winter cultivars Bezostaya-1 and its daughters Avrora and Kavkaz(USSR),Mironovskaya-808(Ukraine),and Partizanka(Serbia)(Table S1,Fig.S1).Allele Gli-B1b was discovered in the landraces Selivanovskii-Rusak(spring)and Krymka-Mestnaya(winter)(Table S1)and in some Spanish landraces of diverse habits[19].

Alleles Gli-B2o and Gli-D2a also occurred as the most frequent in some groups of both spring and winter cultivars(Table 2).The presence of allele Gli-B2o in the genotype of spring wheat is correlated with the adaptability of plants,at least,to conditions of the west Siberia[6].

A large genotypic difference between groups of spring and winter cultivars is displayed in Fig.1:the first and fourth clusters represent winter wheats,whereas the second and fifth represent spring wheats.The characteristic genotypes of the cultivars composing the fifth cluster may be the result of the frequent use in breeding,throughout the 20th century,of Saratovskaya-29 and its descendants and relatives.Thus,alleles Gli-A1f,Gli-B1e,Gli-D1a,Gli-A2q,Gli-B2s,and Gli-D2e are common in spring wheat germplasm of the former USSR and Kazakhstan(Table 2).In contrast,groups of spring wheats in the second cluster showed broad diversity of frequent alleles at five Gli loci and only one allele in common,Gli-B2c(Table 2).The presence of this allele in genotypes of French and Italian cultivars has been correlated with improved dough quality[6].

In contrast to spring wheat,winter cultivars carried alleles Gli-A1a,Gli-A1af,Gli-A1o,Gli-B1b,Gli-B1f,Gli-D1b,Gli-A2g,and Gli-D2a as frequent alleles in three or more of the groups studied(Table 2).The winter cultivar Bezostaya-1 and its daughters Avrora and Kavkaz,which carry the 1BL/1RS translocation,wereusedasparentsinmanybreeding programs,especially in southeast Europe during the second half of the 20th century(Fig.1,fourth cluster).As a result,alleles Gli-A1b, Gli-B2b and Gli-D2b, characteristic of Bezostaya-1,became frequent in winter cultivars from this region(Table 2)whereas for cultivars bred in northwestern Europe(Fig.1,first cluster)no single parental genotype was frequently used in breeding programs.

Thus,a clear difference between Gli genotypes of spring and winter common wheats was confirmed.Moreover,spring and winter cultivars each formed two independent groups of genotypes.One cluster(second,Fig.1)of related spring cultivars united all groups based mainly on genotypes bred in the Saratov breeding center and the small group,North,composed of wheat bred in northern European countries.Four other groups of spring cultivars from countries(Australia,Canada,Mexico,and Portugal)on three continents,which apparently had little in common besides their spring habit also showed similarity among their genotypes(fifth cluster,Fig.1).Surprisingly,this similarity was revealed by the use of Gli genetic markers,which had nothing in common with wheat genes controlling the spring/winter habit of wheat.

Analogously to spring wheat,genotypes of one of the two independent clusters of winter cultivars(fourth,Fig.1)were also based mainly on one cultivar,Bezostaya-1.However,in contrast to spring wheat,winter cultivars of the other large cluster(first,Fig.1)were placed together,probably owing to the similarity of the eco-climatic conditions of cultivation and growth in northwestern Europe.In winter wheat,higher genetic variation was found in groups of cultivars from the Mediterranean and southern regions of Europe(Bulgaria,Croatia,Italy,Romania,Ukraine)than in those from the central and northwestern parts(France,Germany,Holland,Serbia,UK)(Table2).Using microsatellites as genotype markers revealed the higher genetic variation in the southeastern regions of Europe,which are more favorable to wheat cultivation than are the northern regions[57].

The spring and winter germplasm studied showed similar levels of genetic diversity(average for six main Gli loci):H=0.856±0.050(359 cultivars),and H=0.831±0.061(658 cultivars),respectively.Genetic diversity in a group of spring wheats may be higher[55]or lower[56]than that in a group of winter cultivars.

3.7.Common wheat:relation between allopolyploidy and adaptability to eco-climatic environments

The phylogenetic relationships among wheat species and the pathways of origin of hexaploid wheat through successive crosses between diploid donors of the A and B genomes and later between an AB allotetraploid and a diploid donor of the D genome are well known in detail[1,58–60].

As a result of comparison of the composition of allelic variants of blocks known in common wheat,families of similar(and thus closely related,having originated probably through spontaneous mutation)alleles encoded at one Gli locus were identified.Members of one family could hardly arise from other families during the evolution of an already hexaploid wheat.To explain the existence of families of allelic variants of blocks,it has been suggested that multiple crosses occurred between different genotypes of polymorphic diploid donors of the A and B genomes as well as between donors of the AB and D genomes[61,62].Thus,the origin of each family of Gli alleles of common wheat now existing might be traced to a specific genotype of the polymorphic diploid donor.Indeed,at least five and six different genotypes of diploid donors of genomes A and B,respectively,may have participated in the origin of allotetraploid wheat[63]which later crossed with at least two[63–66]or three[34]genotypes of the D-genome donor.As a result,each donor of an elemental genome(A,B,and D)has supplied common wheat with some part of its own genetic variation[61,62].The polymorphism acquired by common wheat from diploid donors and from related species during its origin and subsequent evolution as a hexaploid[1,58–60]now is manifested as extensive multiple allelism with long sets of alleles(divided into families of alleles)at each Gli locus(Fig.S1).

The three genomes of common wheat are related,but differ,for example,in the proportion of gliadin pseudogene DNA sequences located within the Gli loci[35],in the numbers and types of transposons found in the region of the Glu-1 loci[65],in the composition of allelic variants of blocks(Fig.S1),or in different levels of genetic variation at the Gli loci(Table 1).Thus,the genomes of allopolyploid wheat may include many related but different associations of genes,each of them contributing in its own way to the ability of the species to adapt to a wide spectrum of ecoclimatic conditions.

Indeed,the relationship between groups of genotypes(their distribution into clusters)depends on the elemental genome used for the analysis(Figs.2–4)and differs from that obtained for six loci(Fig.1).The fewest differences were observed between clusters of Fig.1(six loci)Fig.3(polymorphism at the Gli-B1 and Gli-B2 loci was considered).However,even in this case,Italian and Croatian cultivars(third cluster,Fig.1)became a subcluster of a large cluster of winter wheat of southeastern Europe,and Spanish wheat of the same third cluster showed more similarity to winter cultivars of northwestern Europe(Fig.3).

Fig.2–Comparison of relationships of gliadin genotypes among 28 groups of cultivars of common wheat.For cluster analysis,results of identification of alleles at the Gli-A1 and Gli-A2 loci were used.

Fig.3–Comparison of relationships of gliadin genotypes among 28 groups of cultivars of common wheat.For cluster analysis,results of identification of alleles at the Gli-B1 and Gli-B2 loci were used.

More information came from comparison of the composition of clusters obtained using the elemental genomes A(Gli-A1 and Gli-A2 loci were considered)and D(Gli-D1 and Gli-D2).Spring and winter cultivars of Italy as well as those of Spain were similar in each of their A,B,and D genomes.In contrast,the A and D genomes,and,particularly,the B genomes of French spring and winter cultivars were more different(Figs.1–4).

Fig.4–Comparison of relationships of gliadin genotypes among 28 groups of cultivars of common wheat.For cluster analysis,results of identification of alleles at the Gli-D1 and Gli-D2 loci were used.

The A genomes of spring cultivars bred in Australia,Canada,Mexico,and Portugal(second cluster,Fig.1),showed similarities with A genomes of different groups of cultivars:first,Australian wheat grouped with the cluster of winter cultivars of northwestern Europe;second,Canadian cultivars formed a strange sub-cluster together with cultivars of the groups North of Europe and Ural of Russia;and,third,cultivars of Mexico and Portugal clustered with some winter cultivars of southeast Europe(Fig.2).However,if polymorphism at the genome D was considered,cultivars of Australia,Canada and Mexico occurred in one cluster together with Spanish wheat,but cultivars from Portugal went far away to cluster with the group of winter cultivars from northwestern Europe(Fig.4).

Each of the three genomes of winter cultivars released in France,Germany,Holland,and UK,being different enough to be easily distinguished by cluster analysis,showed a high level of similarity among each of their three genomes(Figs.2–4).In contrast,winter cultivars bred in two neighboring countries,France and Italy,differed in each of their three genomes.

There were pairs of geographically neighboring countries or regions of southeastern Europe that produced,in the 20th century,winter cultivars similar in each of their three genomes:Croatia and Italy,Bulgaria and Serbia,Romania and Szeged,Don and Ukraine(Figs.1–4).On the other hand,genotypes of cultivars bred at the same time in the same region of Europe might be similar or different when different elemental genomes were tested.For example,cultivars of Romania differed from cultivars bred in Bulgaria or Serbia by Gli alleles of the A genome,and cultivars of Ukraine from those of Martonvasar or Romania by Gli alleles of the D genome.Cultivars bred in Serbia and Croatia,which not only are neighbors but,in the 20th century,were parts of the same country,Yugoslavia,differed considerably at the Gli loci of the B and D genomes(Figs.1–4).Martonvasar and Szeged,two breeding centers in the same country,also produced cultivars that were similar mainly in their A genomes.Cultivars grown in two countries of southern Europe,Italy and Spain,showed some similarity in their A and D genomes,but had nothing in common in their B genomes(Figs.2–4).

Thus,each of the three elemental genomes,especially A and D,madeitsspecificadditionalcontributiontothestructureofthe polymorphism of common wheat of the 20th century.

Divergence of homologous chromosomes of wheat[67–70]reduces their ability to recombine[67,69]and produces so-called“heterohomologous”chromosomes[67,69,70].We suggest that“heterohomologous”chromosomes of the same genome may maintain and keep,in different genotypes,different gene associations,each of them valuable for adaptation to certain environments,and thereby contribute to the structure of wheat polymorphism.

4.Conclusions

We used an updated,enlarged,and improved catalog of 182 alleles at the Gli-1 and Gli-2 loci to analyze common wheat germplasm of the 20th century and found a disproportionate frequency of occurrence of Gli alleles in the collection of 1060 common wheat cultivars.Polymorphism of the common wheat studied was extremely high and was differentiated and structured with respect to countries and regions and with respect to spring and winter cultivars.We suggest that this differentiation may be caused mainly by the peculiarities of eco-climatic conditions of wheat cultivation in each country or region;in other words,by natural selection acting in different ways in different regions.Frequent use in breeding of some “successful” genotypes(such as Bezostaya-1 or Saratovskaya-29)contributes to differentiation and,to some extent,narrowspolymorphisminacountryorregion.However,the necessity of plants to adapt to very different eco-climatic conditions of growth determines and maintains high levels of genetic variation in common wheat as a species.DiploiddonorsoftheA,B,andDgenomessupplied allohexaploidcommonwheatwithpartsoftheirown polymorphism that were valuable for adaptation to certain environments.As a result,the structure of the polymorphism studied differed if elemental genomes were tested separately.We suggest that distinct associations of genes situated on the homoeologous and,probably, “heterohomologous”chromosomes of the A,B,and D genomes may be valuable for adaptation of the wheat plant to specific(different)ecoclimatic conditions of cultivation and growth,thereby providing a high level of plasticity to common wheat.

Also,we found that grain samples of many cultivars were represented by mixtures of related genotypes.We suggest,that,paradoxically,this non-uniformity may be valuable for a commercial common wheat cultivar,helping it to adapt,by changing concentrations of its genotypes,to different environments.Thus,non-uniformity of cultivars may be considered as an important contributor to the known high plasticity of common wheat.

A supplementary table and figure for this article can be found online at https://doi.org/10.1016/j.cj.2018.02.003.

Acknowledgments

This work could not have been performed without samples of grain of common wheat cultivars,which we received,over time,from our colleagues Dr.R.Gupta and Dr.C.W.Wrigley(Australia),Dr.B.Bochev,Dr.G.Ganeva,Dr.I.Panaiotov(Bulgaria),Dr.W.Bushuk(Canada),G.Branlard(France),Dr.J.Sutka(Hungary),Dr.N.Pogna(Italy),Dr.V.Movchan(Kazakhstan),Dr.V.Chernakov,Dr.A.Galkin,Dr.М.Kopus and Dr.S.Koval(Russia),Dr.D.Knezevic(Serbia),Dr.B.Javornik(Slovenia),Dr.R.CookeandDr.P.Payne(UK),and Acad.A.Sozinov and Dr.F.Poperelya(Ukraine).We appreciate very much their scientific discussion,kind encouragement,and valuable advice.