Diss Caroline

8/13/2019 Diss Caroline

1/91

Diss. ETHZ No.17544

Molecular and phenotypic diversity of wheat(Triticum aestivumL.) for winter hardiness

A dissertation submitted to the

ETHZURICH

for the degree of

DOCTOR OF SCIENCES

presented by

CAROLINE PLASSE

M.Sc. in Agronomical Sciences, INPL Nancy, France

born 12.07.1978

Citizen of France

accepted on the recommendation of

Prof. Dr. P.Stamp, examiner

Prof. Dr. B. Keller, co-examiner

Dr. J. Leipner, co-examiner

Zurich, 2007


2/91


3/91

The tools of genome research may finally

unleash the genetic potential of our wild

and cultivated germplasm resources

for the benefit of the society

(Tanksley and McCouch 1997)


4/91


5/91

I

Table of Contents

Summary III

Rsum V

List of abbreviations VII

1 General Introduction 1

1.1 Frost-induced damage and cellular response to low temperature ................................ 1

1.2 Cold acclimation and winter hardiness of wheat ......................................................... 2

1.3 Determination of frost tolerance .................................................................................. 3

1.4 Genetic aspects of frost tolerance in wheat.................................................................. 4

1.5 Use of genetic resources for wheat improvement........................................................ 5

1.6 Contribution of genomic research and molecular markers to crop improvement........ 6

1.7 Aim of the study........................................................................................................... 7

2 Phylogenetic analysis of wheat (Triticum aestivumL.) accessions by microsatellitemarkers and parametric genetic mixture analysis 8

2.1 Abstract ........................................................................................................................ 8

2.2 Introduction..................................................................................................................8

2.3 Material and Methods .................................................................................................. 92.3.1 Plant material...................................................................................................... 9

2.3.2 Microsatellite analysis ......................................................................................10

2.3.3 Genetic diversity assessment ............................................................................ 12

2.4 Results........................................................................................................................ 13

2.4.1 Microsatellite polymorphism............................................................................ 13

2.4.2 Estimation of the number of genetic clusters and assignment of individuals topopulations........................................................................................................ 14

2.4.3 Within and between population diversity......................................................... 16

2.4.4 Causes of the observed population structure .................................................... 172.5 Discussion ..................................................................................................................19

3 Analysis of winter hardiness of divergent wheat material under field conditionsusing chlorophyll fluorescence analysis 22

3.1 Abstract ...................................................................................................................... 22

3.2 Introduction................................................................................................................22

3.3 Material and Methods ................................................................................................ 24

3.3.1 Plant material and growth conditions ............................................................... 24

3.3.2 Phenotype assessment....................................................................................... 243.4 Results........................................................................................................................ 26


6/91

II

3.4.1 Temperature conditions .................................................................................... 26

3.4.2 Comparison of tested accessions with standard lines....................................... 27

3.4.3 Effect of experimental season on winter survival ............................................27

3.4.4 Time course of Fv/Fmand leaf greenness throughout the winter season .......... 28

3.5 Discussion ..................................................................................................................32

4 Analysis of the promoter region of the transcription factor TaCBF12 in wheataccessions of contrasting winter hardiness 35

4.1 Abstract ...................................................................................................................... 35

4.2 Introduction................................................................................................................35

4.3 Material and Methods ................................................................................................ 37

4.3.1 Primer design.................................................................................................... 37

4.3.2 DNA analysis.................................................................................................... 37

4.3.3 Data analysis.....................................................................................................384.4 Results........................................................................................................................ 38

4.4.1 Sequence variability .........................................................................................38

4.4.2 Identification of transcription regulatory elements ..........................................40

4.5 Discussion ..................................................................................................................41

5 Identification of sequence variant associated with freezing tolerance in wheat bycandidate gene-based and genome-wide association analyses 45

5.1 Abstract ...................................................................................................................... 45

5.2 Introduction................................................................................................................455.3 Material and Methods ................................................................................................ 46

5.3.1 Plant material and growth conditions ............................................................... 46

5.3.2 Sequence analysis .............................................................................................46

5.3.3 Population structure and association mapping ................................................. 47

5.3.4 False discovery rate control.............................................................................. 47

5.4 Results........................................................................................................................ 47

5.4.1 Comparison of haplotypes................................................................................ 47

5.4.2 Population structure..........................................................................................48

5.4.3 Candidate gene based association mapping...................................................... 50

5.4.4 Genome wide association mapping ..................................................................51

5.5 Discussion ..................................................................................................................52

6 General Conclusions 55

7 References 58

8 Appendix 69

Acknowledgements 77

Curriculum vitae 79


7/91

III

Summary

Wheat (Triticum aestivumL.) is cultivated on more than 240 million hectares, a larger area

than for any other crop. Furthermore, the trade in wheat worldwide is greater than for all the

crops combined. The intraspecific reduction of wheat genetic diversity caused by the

development of high-yielding cultivars enhances the risk of loss of adaptation to abiotic stress

such as frost, which periodically accounts for significant losses in the production of winter

wheat. The use of genetic resources for improving wheat is considered to be the most

sustainable way conserving valuable genetic material for the future. In the recent past,

genomics emerged as the result of technical advances in molecular biological techniques and

in bioinformatics. Genomic research can contribute to the use of genetic resources for wheatadaptation to winter hardiness as it is possible to identify alleles useful for wheat

improvement. This project aimed to survey genetic resources for wheat winter hardiness

improvement by investigating genetic and phenotypic diversity and by determining the

genetic basis of frost tolerance in wheat by association mapping.

The plant material was a set of 320 wheat accessions provided by the genebank of Agroscope

Changins-Wdenswil (Switzerland). This material originated from 40 countries and each

accession was genotyped at 32 loci with microsatellite (SSR) markers. The genetic diversitystudy detected a high genetic diversity. A parametric genetic mixture analysis was conducted

to analyze the stratification. This analysis identified twelve subpopulations. Many of the

populations consisted mainly of accessions from a certain geographic area or of a particular

growth type.

A subset of 103 accessions was grown under field conditions without snow cover in Jura in

two consecutive winters. Maximum quantum efficiency of PSII primary photochemistry

(F

v/F

m) and leaf greenness were determined throughout the season, and winter survival andshoot dry weight were assessed at the end of the winter. Due to low temperatures, as low as

-29 C in both seasons, frost damage occurred in the less winter hardy accessions. The

determination of Fv/Fm and leaf greenness enabled a classification of the accessions with

respect to winter hardiness and revealed large phenotypic variation within the studied wheat

accessions.

In order to investigate the genetic basis of wheat adaptation to frost, a candidate gene-based

and a genome-wide association study was performed in a collection of 95 wheat accessions.

The gene TaCBF12, which codes for a transcription factor of the AP2 type, was used as


8/91

IV

candidate gene due to its positional association with the Fr-A2 locus, which is located on

chromosome 5A and is known to be the major locus for frost tolerance in wheat. The

sequencing of the promoter region of TaCBF12 revealed nine completely associated single

nucleotide polymorphisms (SNPs) resulting in two haplotypes. The promoter region was

found to contain several cis-acting elements of low temperature regulation. The SNPs caused

that the two haplotypes differed in the presence, respectively absence, of two cis-acting

elements, which are probably involved in the acclimation to low temperature. Therefore, the

two haplotypes were tested for an association with the variation in winter hardiness. No

association was found, suggesting a minor effect of TaCBF12on phenotypic variation in frost

tolerance. However, a highly significant association was found between the markerXcfa2173

on chromosome 4D and variation in the fluorescence parameterFv/Fmat the beginning of the

winter. The identified alleles may be useful in the selection for frost tolerance during winter

hardening of wheat.


9/91

V

Rsum

Avec plus de 240 millions dhectares, le bl est la culture occupant la plus vaste surface dans

le monde. Ces changes sur le march mondial sont plus importants que ceux des autres

cultures runies. Le dveloppement de cultivars hauts rendements a entran une baisse de la

variabilit gntique des varits cultives ce qui augmente le risque dun manque

dadaptation aux stress abiotiques. Parmi ces stress, le gel est rgulirement responsable

dimportantes pertes de rendement chez le bl. La caractrisation des ressources gntiques

est un pr-requis pour leur intgration future en amlioration vgtale et la restauration de la

diversit gntique de cette espce. Ces dernires annes, la gnomique a merg grce aux

avances conjointes des techniques molculaires et bioinformatiques. La recherchegnomique qui permet lidentification dallles utiles lamlioration du bl, dtient le

potentiel de contribuer lutilisation des ressources gntiques pour ladaptation du bl au

gel. Lobjectif de ce projet est dvaluer, chez le bl, les ressources gntiques pour la

tolrance au gel par ltude de la diversit phnotypique et par la dtermination des bases

gntiques de la tolrance au gel au moyen de la gntique dassociation.

Un panel de 320 accessions de bl dorigines diverses fourni par la banque de gnes de

lAgroscope de Changins-Wdenswil (Suisse) a t utilis. Pour valuer la diversit gntiquede ce panel, chaque accession a t value au moyen de marqueurs microsatellites au niveau

de 32 loci. Une grande diversit gntique a t mise en vidence. Une analyse dhybridation

gntique a t mene afin de mettre en vidence la stratification. Cette analyse a permis

lidentification de douze sous-populations. Un certain nombre de sous-populations

rassemblaient des individus dune origine gographique donne ou dun type particulier de

bl (hiver ou printemps).

Un sous-ensemble de 103 accessions a t cultiv en conditions de champ, sans couvertureneigeuse, dans le Jura pendant deux hivers conscutifs. Le rendement quantique maximum de

la photochimie de PSII (Fv/Fm) et la vert des feuilles ont t mesurs durant la saison

hivernale ; la survie et le poids sec de la partie foliaire ont t valus la fin de lhiver. Des

tempratures atteignant -29C pendant les deux hivers ont occasionns des dgts lis au gel

sur les bls les moins tolrants. La dtermination du rapportFv/Fmet de la vert ont permis

une classification des accessions en fonction de leur degr de tolrance et ont rvl une

considrable variation phnotypique parmi les gnotypes de bl tudis.


10/91

VI

Afin dtudier les dterminants gntiques de ladaptation du bl au gel, des tudes

dassociation par les approches gne-candidat et gnome-scan ont t menes chez une

population de 95 accessions de bl. Le gne TaCBF12, qui code un facteur potentiel de

transcription de type AP2, a t choisi comme gne-candidat cause de son association

positionnelle avec le locusFr-A2, cartographi sur le chromosome 5A et connu comme tant

le locus majeur de tolrance au gel chez le bl. Le squenage de la rgion promotrice de

TaCBF12a mis en vidence neuf polymorphismes simples nuclotides (SNPs) compltement

associs entre eux, gnrant deux haplotypes. Des lments cis-rgulateurs de rgulation en

basses tenpratures taient visibles dans la rgion promotrice. Les SNPs ont engendr la

prsence de deux lments cis-rgulateurs, probablement impliqus dans lacclimatation au

froid, diffrenciant les deux haplotypes. Ces deux haplotypes ont ensuite t tests pour leur

association avec la variation phnotypique. Aucune association na t trouve suggrant un

effet mineur de TaCBF12sur la variation phnotypique de la tolrance au gel. Toutefois, une

association fortement significative a t trouve entre le marqueur Xcfa2173 sur le

chromosome 4D et la variation du paramtre de fluorescence Fv/Fmau dbut de lhiver. Les

allles identifis pourraient tre utiles en slection pour la tolrance au gel pendant la phase

dendurcissement du bl.


11/91

VII

List of abbreviations

bp Base pair

CBF C-repeat binding factor

cM Centimorgan

COR Cold-regulated

DRE/CRT Drought-responsive element/C repeat

DREB Drought-responsive element binding protein

Fr-A2 Frost tolerance locus on chromosome 5A

Fv/Fm Maximum quantum efficiency of Photosystem II primary photochemistry

GEVES French group of varieties and seeds study and controlGLM General linear model

ICE Inducer of CBF expression

INRA French national institute of agronomical research

LD Linkage disequilibrium

MLM Mixed linear model

PAM Pulse amplitude modulation fluorometer

PCR Polymerase chain reactionPSII Photosystem II

QTL Quantitative trait locus

RAC Federal station of agronomical research of Changins

Rpm Rotation per minute

SNP Single nucleotide polymorphism

SPAD Soil plant analyses development

SSR Simple sequence repeats


12/91


13/91

Chapter 1

1

1 General Introduction

Wheat (Triticum aestivumL.) was one of the first domesticated food crops and has been the

major staple in Europe, North Africa and central Asia for 8000 years. It is still one of the most

important food crops. Wheat is cultivated on more than 240 million hectares, which is the

largest area of any crop. The most suitable latitudes for wheat cultivation are between 30 and

60 N and between 27 and 40 S (Nuttonson 1955). Wheat grows from 0 to 3,000 m a.s.l.

The optimum growing temperature is 25 C with a minimum of 3 to 4 C and a maximum of

30 to 32 C (Briggle 1980). Precipitations for wheat cultivation range between 250 and 1750

mm (Leonard and Martin 1963). World trade in wheat is greater than for all the other crops

combined. In 2005, worldwide production of wheat reached 625 million tons (FAOSTAT, seehttp://faostat.fao.org/). The world population is predicted to increase to 7.9 billion by 2025

(United States Census Bureau, see http://www.census.gov/ipc/www/idb/worldpopinfo.html).

Assuming that wheat consumption remains stable, 786 million tons of wheat will be required

by 2025. The only way to meet this demand is to develop marginal lands for wheat

production. Frost is a major factor that limits the expansion of wheat and it periodically

accounts for significant reductions in plant productivity. Making use of the genetic diversity

in genetic resources of wheat, improving frost tolerance is the most sustainable way ofsupporting food security. In the last decade, tremendous progress has been made in wheat

genomic research, which has the potential to solve the major problems associated with the

genetic basis of frost tolerance.

1.1 Frost-induced damage and cellular response to low temperature

Low temperature reduces the velocity of many enzymatic reactions and, in particular, inhibits

certain enzymes in the dark reaction of photosynthesis. In contrast, photosynthetic electron

transport is affected only slightly by temperature (Fracheboud and Leipner 2003). Under these

conditions, chlorophyll still absorbs light energy but can not use this energy for

photosynthesis. This excess of excitation has to be dissipated as heat, because pigments that

remain in an excited state pose the risk of generating free radicals, which lead to oxidative

stress (Hurry et al. 2000). This production of reactive oxygen species contributes to

membrane damage (Mc Kersie and Bowley 1997). When temperature drops below 0 C, ice

formation is initiated in the intercellular spaces; the extracellular fluid has a higher freezingpoint than the intracellular fluid due to its lower solute concentration. The formation of


14/91

General Introduction

2

extracellular ice results in the movement of water from inside the cell to the intercellular

spaces, which result in severe dehydration of the plant cell. The accumulation of ice in the

intercellular spaces can cause physical disruption of the cell (Levitt 1980). Moreover, low

temperature triggers protein denaturation, which could potentially harm the cell (Guy and Li

1998), and induces lipid transformations (Steponkus et al. 1993; Uemura and Steponkus

1997).

The strategy of plants to avoid frost-induced injury seems to depend mainly on achieving: (i)

membrane stability, (ii) inhibition of ice formation, (iii) prevention of dehydration and (iv)

macromolecular stability. Membrane stability seems to be the result of the action of lipids and

proteins. In particular, an asymmetry of membrane lipids contributes to stability (Sung et al.

2003). Similarly, amphipathic -helical regions which form by late embryogenesis abundant

(LEA) polypeptides stabilize membranes against damage caused by freezing (Thomashow

1999). Furthermore, proteins encoded by cold-regulated (COR) genes seem to play a role in

the stabilization of the chloroplast membrane (Crosatti et al. 1999). The inhibition of

intracellular ice formation is achieved by antifreeze proteins (AFP) (Smallwood et al. 1999;

Griffith et al. 2005; Tremblay et al. 2005). Dehydrins prevent dehydration (Fu et al. 2000).

Moreover, sugars such as sucrose, glucose, fructose, raffinose and stachyose have a

cryoprotectant effect (Castonguay et al. 1995), and proline, glycine, betaine and sorbitol act as

osmoprotectants by increasing the osmotic pressure in the cytoplasm (Rudolph et al. 1986).

Protection from protein denaturation is achieved by molecular chaperones (Wang et al. 2003).

1.2 Cold acclimation and winter hardiness of wheat

Many plant species have the potential to cold acclimate, acquiring an enhanced tolerance to

negative temperature when exposed to a period of low but positive temperature. Cold

acclimation is a reversible process, which is suppressed when temperatures exceed 0 C, for

example during a spell of mild weather or in spring. In a second phase of hardening, exposure

to negative but non injurious temperatures can confer to plants an increased tolerance to

severe frost (Trunova 1965; Olien 1984; Castonguay et al. 1993; Castonguay et al. 1995;

Livingston 1996). Winter wheat is less frost tolerant than rye but is more tolerant than barley

(Hmm 1994). The acquired frost tolerance results from a combination of developmental,

genotypic and environmental factors. According to Lecomte et al. (2003), the main factors

that influence the acquired frost tolerance of wheat are growth stage, hardening conditions,

the length of cold spells, sequences of frost and defrost, snow cover, maximum tolerance,


15/91

Chapter 1

3

speed of hardening, and the ability to maintain tolerance. The growth stage is important

because at early developmental stages (e.g. first leaf stage), wheat seedlings can withstand

temperatures only as low as -12 C. To achieve a good frost tolerance, a growth period of

three to five weeks at temperatures below 15 C is necessary; the plants can accumulate

reserves. Negative temperatures favor a high and sustainable frost tolerance, but severe frost

near the tolerance threshold of the variety causes a loss of tolerance. Furthermore a succession

of frost and warm periods leads to a progressive decrease in frost tolerance. Snow cover has

an insulating and, therefore a protective effect even when the snow layer is less than 5 cm.

In wheat, the maximum frost tolerance is around -12 C for sensitive varieties and about

-32 C for the most tolerant varieties. However, as mentioned above, the maximum resistance

can be expressed only when the plant has reached a sufficient developmental stage and when

the conditions for hardening have been met. The time required to reach the maximum

tolerance is between three and five weeks, depending on the variety. This difference in

acclimation speed can play a role during early cold spells, as fast acclimating varieties are

better able to tolerate these cold spells. Furthermore, the ability of a variety to maintain cold

hardiness contributes to its general winter hardiness when late cold spells occur after a warm

period at the end of winter.

1.3 Determination of frost tolerance

There are a number of approaches to studying the frost tolerance of plants. Under field

conditions, winter survival is usually determined or frost damage is visually assessed and

quantified as the percentage of yellow leaf area after strong frost events. In the laboratory,

frost-induced leaf damage is often determined by measuring electrolyte leakage. The

measurement of electrical conductivity permits the assessment of the temperature at which 50

% of the electrolytes had leaked from a leaf or a plantlet (LT50) and, therefore, expresses the

degree of frost hardiness. A similar method is the measurement of the triphenyl tetrazolium

chloride-reduction (TTC-reduction) capacity, which provides the degree of frost hardiness of

the sample expressed as the temperature at which the amount of formazan formed after TTC-

reduction is 50 % of that formed in control leaves. For many years, the chlorophyll

fluorescence has been routinely used for the non-invasive assessment of photosynthetic

performance of plants.

The analysis of chlorophyll fluorescence is a very powerful tool for accessing the plants

tolerance to abiotic stress factors (Fracheboud and Leipner 2003). Many parameters of the


16/91


4

chlorophyll fluorescence are strongly affected by temperature and can be used to determine

frost tolerance. Exposure of plants to low temperature in the light results in an energy

imbalance that can lead to photoinhibition (Osmond 1994). This decrease in the maximum

quantum efficiency of PSII primary photochemistry (Fv/Fm), which can occur due to freezing

stress, indicates a change in the PSII reaction center, from a functional to a down-regulated or

non-functional reaction center (Powles 1984; Krause 1988). The measurement of Fv/Fmhas

an advantage over the other fluorescence parameters under conditions at which frost stress

must be evaluated at fluctuating temperature, Fv/Fmper seis temperature-independent since

the efficiency of energy transfer at the PSII is a quantum tunneling effect. Beside Fv/Fm, the

quantum yield of electron transport at photosystem II (PSII) has been reported to allow to

differentiate Triticumspecies with different cold tolerance (Rekika et al. 1997).

1.4 Genetic aspects of frost tolerance in wheat

Hexaploid wheat presumably originated in Iran, southwest of the Caspian Sea, as a result of

the hybridization between tetraploid wheat (ssp. Dicoccon or ssp. Parvicoccum, AABB

genome) and the diploid Aegilops tauschii(DD genome) followed by chromosome doubling

(Bonjean and Angus 2001). In wheat, at least ten of the 21 chromosome pairs seem to be

involved in a complex inheritance of frost tolerance (Cahalan and Law 1979; Sutka et al.1986). The group 5 chromosomes and the chromosomes 2B and 4B were reported to be

important in determining frost tolerance, and it has often been suggested that chromosome 4D

is involved in frost tolerance. Furthermore, a quantitative trait locus (QTL) for frost tolerance

was detected on chromosome 1D (Bga et al. 2007). The chromosome 5A carries two loci for

frost tolerance Fr-A1 (formerly Fr-1) and Fr-A2. In Triticum monococcum, these loci are

separated by a distance of 30 cM. (Sutka and Snape 1989; Vgjfalvi et al. 2003). Fr-A1, the

gene for frost resistance is closely linked to gene Vrn-A1, which controls vernalization

(Galiba et al. 1995; Sutka et al. 1999). Chromosome 5B carries two loci for frost tolerance

orthologous toFr-A1andFr-A2namedFr-B1andFr-B2(Tth et al. 2003), and chromosome

5D carries the orthologous loci for frost toleranceFr-D1(formerlyFr-2) andFr-D2(Snape et

al. 1997). A cluster of CBFgenes was mapped on theFr-Am2locus of Triticum monococcum

(Miller et al. 2006). The CBFgenes encode cold-responsive transcription factors of the AP2

type. InArabidopsis, the CBF3was shown to play a key regulatory role in freezing tolerance

(Liu et al. 1998). Overexpressing this gene inArabidopsisenhanced freezing tolerance (Jaglo-


17/91

Chapter 1

5

Ottosen. et al. 1998; Kasuga et al. 1999). Therefore, the CBFgenes seem to be a potential

target for dissecting genetic variation in frost tolerance.

1.5 Use of genetic resources for wheat improvement

The intraspecific reduction caused by the development of high-yielding cultivars enhances the

risk of losing adaptation to abiotic stress such as frost, which periodically accounts for

significant reductions in wheat productivity. The use of genetic resources to improve wheat is

considered to be the most sustainable way of conserving valuable genetic resources for the

future. Genetic resources are all the material that is available for improving a cultivated plant

species (Becker 1993) and are divided into four gene pools. The primary gene pool consists of

species (including the target species), which are easily crossed to the target species. Thesecondary gene pool is made up of related species; their crossing with the targeted species

often results in a low percentage of viable seeds and partially sterile progenies. The tertiary

gene pool is defined by species, which can be used only by applying techniques such as

embryo rescue or protoplasm fusion. Today a fourth class exists: isolated genes, which cross

all the boundaries of plants, animals or microorganisms.

Worldwide, more than 1500 genebanks are registered in the WIEWS (World Information and

Early Warning System on PGR) and SIS (Seed Information System) databases. They

conserve more than five million accessions belonging to more than 18000 species. The

WIEWS facilitates the exchange of information among member countries on their gathered

information on plant genetic resources. In Europe, many genebanks exist in which Triticum

aestivumaccessions are held under the following statuses like wild, weedy, traditional cultivar

or landrace, breeders line or advanced cultivar. In Switzerland, three institutes hold 4386

Triticum aestivumaccessions from various countries. Among these institutes, the Agroscope

Changins-Wdenswil, a governmental institute, is engaged in plant genetic resources

activities and holds an ex situ collection of plant genetic resources with 4331 Triticum

aestivum accessions (2000 traditional cultivars/landraces, 1700 breeders lines and 631

undefined accessions) (FAO see http://apps3.fao.org/wiews/). The cereal activities of the

Agroscope Changins-Wdenswil are conservation, characterization, utilization and plant

breeding.


18/91


6

1.6 Contribution of genomic research and molecular markers to cropimprovement

In the last few years, genomicshas emerged from the technical advances made in molecular

biological techniques and bioinformatics. Genomics aims at examining all genes and their

functions of an organism. It has the potential for using genetic resources for the adaptation of

wheat adaptation to frost as genes which can be used to improve wheat are identified.

Genomic regions for a trait of interest can be identified by quantitative trait loci (QTL)

analysis using a segregating population from two inbred lines. However, long stretches of

chromosomes, in linkage disequilibrium (LD), because of a low number of generations after

maximum LD is reached in the F1, lead to large confidence intervals for QTLs. In comparison

with QTL analyses, association studies have the potential to mine directly the allelic diversity

of genetic resources and to identify alleles that are beneficial for the trait of interest. For an

association study, a random population sample of unrelated individuals is required. Because

population structure can lead to false associations, it must be controlled. Such spurious

associations can arise from extensive interchromosomic LD caused by an admixture of

populations with different allele frequencies. Consequently, significant trait-marker

associations can be found with markers, even if they are not in the very close proximity of

trait loci. Two strategies have been developed for association mapping (Oraguzie and Wilcox

2007). One is based on candidate genes: the candidate gene is sequenced within germplasmsshowing phenotypic variation in order to detect polymorphism in the gene, which is

potentially associated with the phenotype. The second strategy consists of genotyping the

genome with evenly distributed markers. A significant association can be detected if the

chosen markers are in LD with genes controlling the phenotypic variation (Haussmann et al.

2004).

In order to avoid spurious associations due to population structure, the genetic stratification of

the sample population is analyzed by means of molecular markers. They are an efficient toolin diversity studies for identifying the degree of genetic similarity. Due to their high rate of

polymorphism, co-dominant character, selective neutrality, distribution across the genome

and cost and labor efficiency, microsatellites or simple sequence repeats (SSRs) are suitable

markers for detecting allele frequency within the population and for assessing population

structure. The SSRs are tandemly repeated motifs of one to six bases. Polymorphism is

created by the loss or gain of repeats. Three mechanisms may be responsible for the

occurrence of new alleles at SSR loci: replication slippage, unequal crossing-over and genetic

recombination. Replication slippage seems to be responsible for affecting the repeat number


19/91

Chapter 1

7

of short tandem arrays (Tachida and Lizuka 1992) whereas unequal crossing-over affects long

tandem arrays (Harding et al. 1992).

In association mapping, in particular candidate gene-based association mapping, single

nucleotide polymorphisms (SNP) are attractive markers because of their high frequency, low

mutation rate and amenability to automation. SNPs are individual differences in nucleotides

between two DNA sequences. SNPs are produced by mutations in part due to errors during

DNA replication. The mutation frequency is not random between two bases; it depends on the

nature of the base, the sequence in the vicinity and the methylation status of the DNA.

1.7 Aim of the study

To achieve a sustainable agriculture under changing climatic conditions, wheat varieties arerequired, which are characterized by a high degree of winter hardiness and a low demand of

vernalization. Therefore, the long-term aim is to better understand the response of wheat to

frost and its genetic background in order to develop methods to improve its winter hardiness.

For this purpose, it is important to use the genetic resources of wheat and to study the genetic

and molecular mechanisms of frost tolerance.

The focus of this thesis is on three major areas in research of the use of genetic resources for

improving wheat with regard to frost tolerance: (i) investigation of genetic and phenotypic

diversity of the wheat genetic resources available at Agroscope Changins-Wdenswil, (ii)

testing the chlorophyll fluorescence analysis for assessing winter hardiness and (iii) the

determination of the genetic basis of frost tolerance in wheat by association mapping.


20/91


21/91

Chapter 2

9

SSRs are multiallelic, chromosome specific and are evenly distributed along the chromosome

(Rder et al. 1998a; Rder et al. 1998b). SSR markers were shown to be highly polymorphic

in diploid species, in the tetraploid wild wheat Triticum dicoccoides, in the D-genome donor

of wheat,Aegilops tauschii, and in hexaploid wheat (Plaschke et al. 1995; Donini et al. 1998;

Hammer et al. 2000; Pestsova et al. 2000; Prasad et al. 2000; Stachel et al. 2000; Fahima et al.

2002). Furthermore, SSRs have been successfully used in wheat for QTL identification

(Parker et al. 1998), resistance gene tagging (Peng et al. 1999; Brner et al. 2000b), marker-

assisted selection (Korzun et al. 1998; Huang et al. 2000) and for investigating the genetic

stability of genebank accessions (Brner et al. 2000a). The use of SSRs has facilitated the

identification of subtle population subdivisions by focusing on the distribution of ancestry

proportions, which can provide additional information about admixture processes (Falush et

al. 2003).

There is a great potential for improving abiotic stress tolerance, e.g. winter hardiness, by

integrating seed material from gene banks in the actual breeding material. However, the

genetic resources stored in gene banks are often poorly analyzed in respect to their

physiological potential as well as their genetic background. Therefore, the investigation of the

genetic diversity in wheat is a prerequisite for identifying genetic resources with high stress

tolerance. Thus, the objective of the present study was to assess the genetic diversity of a

large number of hexaploid wheat germplasms. Furthermore, this study was the basis for

further physiological and molecular biological analyses.

2.3 Material and Methods

2.3.1 Plant material

Two hundred ninety-eight wheat (Triticum aestivum L.) accessions were provided by

Agroscope Changins-Wdenswil (Switzerland). Furthermore, 22 accessions were supplied by

INRA Dijon (France). The accessions originated from 40 countries (Fig. 2.1) and consisted of

released and unreleased cultivars, breeding lines and local variety/landraces (Table Appendix

1). The material consisted of 153 winter, 72 spring and 3 intermediate wheat types; the growth

habit of the remaining 92 accessions was not catalogued in the European Wheat Database

(EWDB) (see http://genbank.vurv.cz/ewdb/).


22/91

Phylogenetic analysis of wheat

10

2

25

24

11

3

2

3

90

3

1

261

3

17

54

25

2

2

24

1

1

14

4

11

13

115+1

3

6

4

33

4

9

14

7

3

4

3

14

15

6

6

1

4

2

25

24

11

3

2

3

90

3

1

261

3

17

54

25

2

2

24

1

1

14

4

11

13

115+1

3

6

4

33

4

9

14

7

3

4

3

14

15

6

6

1

414

4

11

13

115+1

3

6

4

33

4

9

14

7

3

4

3

14

15

6

6

1

4

Figure 2.1: Origin of investigated wheat accessions. Number of accessions from eachcountry, and for four European regions is given above the country. Dark blue, North America;dark red, South America; light blue, northern Europe; dark-green, western Europe; red,southern Europe; green-yellow, Eastern Europe (countries of the Warsaw pact including theformer Yugoslavia but excluding the former Soviet republics Kazakhstan and Armenia); grey,Africa; yellow, Asia; turquoise, Oceania.

2.3.2 Microsatellite analysisFor DNA extraction, wheat plants were grown in growth chambers at 18/16C (day/night) in

0.75-l pots under a 12-h photoperiod (300 mol photons m-2s-1) and at a relative humidity of

60/70 % for 30 days. The plants were watered and fertilized with Wuxal nutrient solution as

required. From five-week-old plants, approximately 5 g of leaf tissue were collected and

frozen in liquid N2. Samples were ground in liquid N2, added to 15 ml of CTAB and

incubated at 65C for 90 minutes. An equal volume of dichloromethan:isoamylalcohol was

added. The mixture was centrifuged at 3500 rpm for 15 minutes (Megafuge 2.0, Heraeus,

Osterode, Germany). The precipitate was added to 15 ml dichloromethan:isoamylalcohol,

mixed well and centrifuged at 3500 rpm for 15 minutes. 75 l RNase A were added to the

supernatant and the mixture was incubated for 30 to 45 minutes. The DNA was then

precipitated with 10 ml of isopropanol. The pellet was hooked up by sterile pipettes and

washed in wash solution 1 (76% ethanol, 200 mM sodium acetate) and wash solution 2 (76%

ethanol, 10 mM ammonium acetate). Finally, the pellet was air-dried and the DNA was

suspended in 1 ml TE-buffer (10 mM Tris/HCl pH 8.0, 1 mM EDTA).

Twenty-three microsatellite markers, detecting 32 loci, were selected for genotyping on the


23/91

Chapter 2

11

basis of their chromosomal location. Markers that representatively cover the genome were

identified (Table 2.1): 12 Xgwmmarkers, developed at IPK Gatersleben (Institute of Plant

Genetics, Germany) (Rder et al. 1998b; Pestsova et al. 2000), 3 Xwmc markers from the

Wheat Microsatellite Consortium (Gupta et al. 2002), 1 Xbarcmarker developed at USDA-

ARS (Beltsville, USA), 5 Xcfa markers developed at INRA Clermont-Ferrand, France

(Guyomarc'h et al. 2002b; Sourdille et al. 2003), 1Xcfdmarker (Guyomarc'h et al. 2002a) and

1 Xgdmmarker (Pestsova et al. 2000; Paillard et al. 2003; Somers et al. 2004). The marker

was also tested on nullisomic-tetrasomic stocks in order to determine the wheat chromosome

which it was assigned. Two methods were used for polymerase chain reactions (PCR):

protocol A and B (Table 2.1). The PCR of protocol A was performed in 10 l in a PTC-200

(MJ Research) Peltier 9600 thermal cycler. The reaction buffer contained 16.25 ng template

DNA, 0.125 mM of each deoxynucleotide, 10 nM Tris-HCl (pH 8.3), 50 mM KCl, 1.5 mM

MgCl2, 0.05 M labeled primer (IRD700 or 800), 0.2 M unlabelled primer and 0.5 U Taq

DNA polymerase. Twenty-four to 35 cycles of 1 minute between 50 to 60C (depending on

the primer combination, see Table 2.1) and 1 minute at 94C were performed, followed by

two-minute elongation step at 72C. Fragment analysis was carried out on a LI-COR 4200

DNA analyzer (Licor, Lincoln, NE, U.S.A.). Fifteen l of formamide tracking dye

(Amersham) was added to each PCR sample and the PCR product was separated on 8%

acrylamide gels. A pattern diversity analysis was performed by visual scoring. For protocol B,

the reaction mixture (10 l) contained 16.25 ng of template DNA, 2.5 mM dNTP, 1 M M13-

tailed forward primer, 10 M reverse primer, 100 mM Tris-HCl, 500 nM KCl, 15 mM MgCl2,

0.75 U Taq polymerase (Sigma) and 10 M one of three fluorophores (6FAM, NED or VIC).

The PCR conditions were as follows: an initial denaturation step of 4 minutes at 94C,

followed by 35 cycles of 1 minute at 94C, 1 minute at Tm temperature (see Table 2.1), 2

minutes at 72C followed by an extension step of 10 minutes at 72C and a final step of 30

minutes at 60C. Amplified products were detected on a DNA sequencer (ABI PRISM 3730

Genetic Analyzer, Applied Biosystems, Foster City, U.S.A.). Samples were prepared by

adding 1 l of PCR product to 11.95 l of HiD formamide and 0.15 l of CIDR size standard.

Pattern diversity analysis was performed using GeneMapper software v4.0 (Applied

Biosystems, Foster City, U.S.A.). When markers produced more than one band, the bands

with clearly separated size ranges were scored independently as a different locus.


24/91


12

Table 2.1: Name, loci, repeat motif, allele number and gene diversity of the 23 testedmicrosatellite markers. The annealing temperature (Tm) and the number of PCR cycles as wellas the protocol are given.

Name Locus Repeat motifa Allele

number

Gene

diversity

Tm Cycles Protocol

Xbarc133 3B (CT)24 9 0.502 55 29 AXcfa2141 a 5A (GA)18 14 0.885Xcfa2141 b 5D 3 0.196

} 55 30 A

Xcfa2153 1A (CT)27 20 0.920 55 35 BXcfa2163 5A (TC)16 11 0.439 55 35 BXcfa2173 4D (CA)28 16 0.841 55 24 AXcfa2174 a 7A (CT)15(GT)14 15 0.792Xcfa2174 b 7B 6 0.515

} 60 35 B

Xcfd223 3D (TG)18 10 0.733 50 35 BXgdm132 6D (CT)24 15 0.854 60 35 BXgwm190 5D (CT)22 4 0.706 60 35 BXgwm234 a 5A (CT)16(CA)20 3 0.117Xgwm234 b 5B 14 0.866

} 55 28 A

Xgwm311 a 2D (GA)29 13 0.861Xgwm311 b 6B 14 0.274

} 60 35 B

Xgwm335 5B (GA)14(GCGT)3 19 0.865 55 35 BXgwm357 1A (GA)18 12 0.764 55 34 AXgwm37 7D (AG)8GG(AG)21 7 0.509 60 35 BXgwm526 a 2A (CT)16 7 0.624

Xgwm526 b 2B 4 0.148

} 55 35 B

Xgwm538 4B (GT)6T(GT)10 8 0.472 60 29 AXgwm570 6A (CT)14(GT)18 13 0.865 60 35 BXgwm624 4D (GT)26 11 0.816 50 35 BXgwm642 1D (GT)14 8 0.602 60 35 BXgwm779 3A (CA)31(GA)33 25 0.865 60 35 BXwmc150 a 5A (GT)21 5 0.640Xwmc150 b 5D 4 0.506Xwmc150 c 6A 10 0.794Xwmc150 d 7D 6 0.609

} 50 33 A

Xwmc216 a 1B (GT)22 3 0.373

Xwmc216 b 1D 7 0.342} 60 30 A

Xwmc232 4A (CA)11 8 0.454 55 30 A

aRepeat motifs are based on the sequence of the variety Chinese Spring

2.3.3 Genetic diversity assessment

The gene diversity, which corresponds to the polymorphic information content (PIC)

(Anderson et al. 1993), was computed according to Nei (1973) as:

Gene diversity =

=

k

1j

2jp1 , for a single locus and 1-

= =

m

1i

k

1j

2ijp

m

1for several loci,

wherepijis the frequency of the jthof k alleles for the ithof m loci.


25/91

Chapter 2

13

The genetic diversity of the wheat accessions was analyzed by clustering the accessions by

applying a parametric genetic mixture analysis implemented in the softwarestructureversion

2.2 (Pritchard et al. 2000). A number of genetic clusters (K) ranging from 1 to 20 were tested

using the linkage model in conjunction with the correlated allele frequency model (Falush et

al. 2003). The distances of all markers were obtained from the Wheat Composite 2004

(GrainGenes 2.0) with the exception ofXcfa2141at 5D andXwmc150at 6A whose distances

originate from the Wheat Consensus SSR 2004 (Somers et al. 2004). For each K, 20

independent runs were implemented. For each run, 10 000 burn-in periods followed by 10 000

Markov chain Monte Carlo (MCMC) were found to be sufficient. The number of genetic

clusters (K) was estimated using the graphical method of Evanno et al. (2005) with minor

modifications.

2.4 Results

2.4.1 Microsatellite polymorphism

A total of 330 alleles were detected with an average allele number of 10.28 per locus and a

gene diversity of 0.617. The number of alleles ranges from 3 to 25 with the lowest gene

diversity, which is also called polymorphic information content (PIC), for Xgwm234a at

chromosome 5A and the highest for Xcfa2153 at 1A (Table 2.1). The highest number of

alleles per locus was detected in the A genome with 11.92, compared to 9.63 and 9.08 for

Table 2.2:Comparative analysis of genetic diversity among homologues groups and genomesby the average allele number per locus and the gene diversity.

Group Average allele number Gene diversity

All 10.28 0.617

Genome

A 11.92 0.680B 9.63 0.502D 9.08 0.631

Homologues group

1 10.00 0.6002 7.40 0.5443 14.67 0.7004 10.75 0.6465 9.22 0.580

6 13.00 0.6977 8.25 0.606


26/91


14

genomes B and D, respectively (Table 2.2). The corresponding gene diversity ranged from

0.502 for the B genome to 0.680 for the A genome. The highest allele number per locus

among the seven homologous groups was observed in chromosomal group 3 and the lowest in

group 2 which was also reflected by the gene diversity for these two groups which was 0.700

and 0.544, respectively (Table 2.2).

2.4.2 Estimation of the number of genetic clusters and assignment of individuals topopulations

The number of genetic clusters (K) was estimated from the rate of change in the log

probability of data between successive K values (K) obtained from 20 independent runs

under the linkage model (Fig. 2.2). The mean likelihood L(K) over 20 runs decreased slightly

until K = 8 but increased strongly at higher K; at K > 12 the increase in L(K) became

K

2 4 6 8 10 12 14 16 18 20

L

(K)

-14000

-13500

-13000

-12500

-12000

-11500

-11000

K

2 4 6 8 10 12 14 16 18 20

L

'(K)

-300

-200

-100

0

100

200

300

400

K

2 4 6 8 10 12 14 16 18 20

|L

''(K)|

0

100

200

300

400

500

K

2 4 6 8 10 12 14 16 18 20

K

0.0

0.5

1.0

1.5

A

C D

B

Figure 2.2:Determination of the number of genetic clusters (K) using the graphical methodof Evanno et al. (2005). (A) Mean L(K) SD over 20 runs for each K value. (B) Rate ofchange in the mean likelihood distribution calculated as L'(K) = L(K) - L(K-1). (C) Absolutevalues of the second order rate of change in the mean likelihood distribution calculated as|L''(K)| = |L'(K+1) - L'(K)|. (D) Kcalculated as K= |L''(K)|/SD[L(K)].


27/91

Chapter 2

15

considerably less (Fig. 2.2 A). Plotting |L''(K)|, which reflects the second order rate of change

in L(K) with respect toK, againstKresulted in a peak at aroundK= 13 (Fig. 2.2 C). Dividing

|L''(K)| by the standard deviation of L(K) resulted in Kwith a peak value forK= 12 (Fig. 2.2

D). Consequently, the number of genetic clusters (K) of the investigated plant material was

assumed to beK= 12.

The individual analyses of the 20 independent runs under the assumption ofK= 12 revealed

that 7 out of the 12 populations were consistent among all the runs (D to G and I to K). The

populations A and B were joined to one population in 7 runs. For the assignment of the

individuals to one of the 12 populations, the maxp (the maximum proportional membership to

one of the 12 populations of each individual) of the Qmatrices was taken from the run which

showed the highest mean maxp over all individuals. The comparative analysis of the mean of

maxp (corrected with maxp for maximum possible admixture) between successiveKsupported

the estimation ofKby K; the average maxp increased untilK= 12 and remained constant at

higherK.

The value of , which reflects the relative contribution of population kto the genetic material

in the sample, was = 0.377 0.026 (n = 20) at K= 12 suggesting that ancestry of many

accessions originated from admixed populations. Individual proportions of membership in

each group estimated are in agreement with the existence of a genetic stratification

(population structure) in the sample. Furthermore, the estimated value of the -parameter is

consistent with the distribution of the proportional membership of individual accessions (Fig.

2.3).

0.0

0.2

0.4

0.6

0.8

1.0 AB

C

D

E

F

G

H

I

J

K

L

Fig. 2.3:Barplot of the proportional membership of individual accessions within each of thetwelve inferred populations (p ). Each accession is represented by a vertical bar composed of

different colors representing the admixture to the twelve populations.


28/91


16

2.4.3 Within and between population diversity

The twelve populations consisted of 21 (populations C and D) to 31 (populations B, E and J)

individuals. The average allele number was ranged from about 4.1 and 5.3 which

corresponded to a gene diversity of 0.463 to 0.600 and an FSTof 0.858 to 0.480 (Table 2.3).

The least dispersed population was population D. Population J had a higher gene diversity

than the whole set of 320 accessions (see for comparison Table 2.2); consequently, FSTwas

very low for population J (Table 2.3).

Table 2.3: Comparative analysis of genetic diversity among the populations, homologuegroups and genomes by means of the average allele number per locus and the gene diversityas well as the standard measure of genetic distance,FST, for each population.

Population n Average allele number Gene diversity FST

A 24 5.00 0.539 0.594B 31 4.97 0.529 0.467C 21 4.81 0.544 0.743D 21 4.06 0.463 0.858E 31 4.75 0.468 0.763F 28 4.53 0.488 0.651G 25 4.81 0.539 0.619H 25 4.69 0.502 0.708

I 29 5.31 0.600 0.480J 31 7.44 0.662 0.012K 29 5.06 0.553 0.557L 25 4.94 0.559 0.610

The distances among populations were analyzed by the allele frequency divergence among

populations computed by the programstructureusing point estimates of P(Fig. 2.4). There

was a high allele frequency divergence between population J and all the other populations.

Furthermore, this analysis revealed that populations B, D, F and K were distinct from the

other populations.


29/91

Chapter 2

17

A

H

L

EG

C

I

J

B

F

D

K

0.02

F

H

D

E

A

K

L

B

CI

G

J

A

H

L

EG

C

I

J

B

F

D

K

0.02

F

H

D

E

A

K

L

B

CI

G

J

Fig. 2.4: Tree representing the distances among populations. The divergence in allelefrequency among the populations was computed using point estimates of P. The tree wasconstructed using MEGA 3.1.

2.4.4 Causes of the observed population structure

Among the 12 populations which captured the relevant subdivisions of the samples, origin

seemed to play a role in the population structure observed. This is exemplified by the fact that

42.3 % of the southern European accessions were classified in population G (Fig. 2.5 A). In

this population a considerable number of the South American accessions was present. The

largest part of the South American material was classified in population H and made up one

third of the accessions of this population. The material from western Europe was classified in

all the populations, with population B consisting of 58.1 % of western European accessions,

for the most part from France. Population E also had a high percentage of western European

material (32.3 %). However, the special characteristic of this population was the high

percentage of northern European material; 48 % of the northern European accessions were

classified in population E.


30/91


18

Num

bero

faccess

ions

0

10

20

30

40

0

10

20

30

40

Population

A B C D E F G H I J K L

0

10

20

30

40

A

B

C

sub-meridionale

albidum

alborubrum

barbarossa

erythroleucon

erythrosperum

sub-erythrospermum

ferrugineum

sub-ferrugineumgraecum

hostianum

lutescens

meridionale

milturum

turcicum

velutinum

nd

N-America

S-America

N-Europe

W-Europe

E-Europe

S-Europe

Africa

Asia

Oceania

Winter wheat

Spring wheat

Intermediate

nd

Fig. 2.5: Origin (A), type (B) and variety (C) of the 320 accessions comprised in the 12populations (A-L).

It was also possible to identify type-related causes of the populations. Indeed, wheat of the

winter type made up the predominant part of the populations D, E, F and, in particular B (Fig

2.5 B). The spring wheat type was distributed among the remaining populations, in particular

in K and L. A similar distribution was observed in respect to the varieties. The var. lutescens

was found mainly in population E (Fig 2.5 C). In contrast, populations H and K consisted of

52.0 and 44.8 % of var. erythrospermum, respectively. A considerable part of var.

ferrugineumwas found in populations A, I and L.


31/91

Chapter 2

19

A comparative analysis of microsatellite diversity among the nine geographical regions

revealed that the accessions from South America exhibited more genetic diversity than those

from the other regions. In contrast, the northern European material showed the lowest

diversity (not considering Africa because of the low number of accessions from that

continent) (Table 2.4).

Table 2.4:Comparative analysis of genetic diversity among regions analyzed by the averageallele number and gene diversity.

Region n Average allele number Gene diversity

S-America 55 6.66 0.611Asia 13 4.38 0.603E-Europe 53 6.91 0.593

N-America 27 5.38 0.590S-Europe 26 5.25 0.580W-Europe 90 7.72 0.571Oceania 27 5.19 0.568

N-Europe 25 4.53 0.509Africa 3 1.84 0.324

2.5 Discussion

A set of 23 microsatellite (SSR) markers detecting 32 loci was used to characterize 320

accessions of hexaploid wheat. The SSR markers displayed a high level of polymorphism.

According to Huang et al. (2002), the number of alleles can be used to evaluate genetic

diversity because of the significant correlation between gene diversity and the number of

alleles. There was an average allele number of 10.3 per SSR locus in the investigated wheat

material whereas RFLP analyses revealed 4.7 polymorphisms per probe/enzyme combination

of the 81 European wheat cultivars (Siedler et al. 1994). Even fewer polymorphisms were

found with RAPDs primers; they generated only 1.8 polymorphisms per primer among the 15

investigated wheat cultivars (Joshi and Nguyen 1993). These findings support the finding of

Huang et al. (2002) who showed that SSR markers in wheat are more informative than RFLPs

and RAPDs. Similarly to the present results, Rder et al. (2002) found an average of 10.5

alleles per microsatellite locus by genotyping 500 European wheat varieties with a set of 20

SSR markers. Breseghello and Sorrells (2006) found an average number of 4.8 alleles per

SSR locus in 149 elite soft wheat cultivars. A similar number of alleles was found for 60

wheat cultivars from eastern Europe, genotyped at 42 SSR loci (Stachel et al. 2000). Other


32/91


20

estimates of meanallele number per locus in studies using SSR markerson diverse germplasm

were: 5.6 alleles at 70 loci on 58 Triticum durumcultivars of diverse geographical origins,

including oldcultivars (Maccaferri et al. 2003) and 18.1 alleles at 26 loci on 998 accessions of

wheat from the genebank of IPK Gatersleben, Germany (Huang et al. 2002). The latter study

suggests a much larger genetic diversity than that found in the studied 320 accessions.

However, in respect to gene diversity (PIC), the difference between the gene diversity found

in the present study (0.62) and that found in the 998 accessions of the genebank of IPK

Gatersleben (0.77) (Huang et al. 2002) is smaller than the difference in the average number of

alleles. The gene diversity of the B genome was found to be considerably lower than the gene

diversity of the genomes A and D. This is consistent with the fact that the B genome is quite

different from A and D genomes whereas the A and D genomes are more similar (Bonjean

and Angus 2001). Huang et al. (2002) compared the gene diversity between wheat accessions

from northern and southern Europe and found lower gene diversity in varieties from northern

Europe than in those from southern Europe. The present study led to a similar finding. The

gene diversity of the northern European accession was 0.509, much lower than that of the

accessions from southern Europe (0.580).

In order to analyse the genetic stratification of the material, the genetic diversity of the 320

wheat accessions was analyzed by clustering the accessions by means of a parametric genetic

mixture analysis (Pritchard et al. 2000). The model-based clustering method of this analysis

was chosen to use mutilocus genotype data, to infer populations and to assign individuals to

these populations. It was assumed that SSRs frequencies correlated among populations as it is

often more effective than the independent frequencies model at detecting subtle population

structure (Falush et al. 2003). This method is also known to be more suitable for explanatory

data analysis than distance-based methods, because it is easier to assess the validity of explicit

modeling assumptions than to compare the relative merits of more abstract quantities such as

distance measures and graphical representations (Pritchard et al. 2000). The analysis of the

genetic stratification of the 320 accessions suggested a large admixture history between close

and distant relatives, gene flow between cultivated species and mutations, which contributed

to allele frequency modifications. The clustering method was successful because the gene

diversity in the obtained groups was lower than in the whole population. Furthermore,

clustering reflected the geographic origin of the material. Nevertheless, none of the clusters

was made up of varieties of just one origin. This can be easily explained when exchanges of

wheat material from one region to another is considered, which homogenized the wheat pooland results in similar allele frequencies. For example, a relatedness was found between wheat


33/91

Chapter 2

21

accessions from South America and wheat from southern Europe. This relatedness can be

historically explained by the fact that Spaniards brought wheat to Mexico in 1529 from where

it was distributed to South America (Bonjean and Angus 2001).

In summary, a study of the genetic diversity of 320 wheat accessions based on 32 SSR loci

revealed a high genetic diversity among these wheat accessions. The linked model-based

diversity analysis identified twelve subpopulations. This genetic diversity survey will help to

broaden genetic variation in future wheat breeding.


34/91

Analysis of winter hardiness

22

3 Analysis of winter hardiness of divergent wheatmaterial under field conditions using chlorophyllfluorescence analysis

3.1 Abstract

Tolerance to frost is an important goal in wheat cultivation to expand the growth area and to

reduce pre-harvest losses. The goals of this work were to characterize wheat accessions with

contrasting frost tolerance and to investigate the suitability of chlorophyll fluorescence and

leaf greenness traits as selection criteria for winter hardiness. One hundred three wheat

accessions were grown under field conditions without snow cover at the altitude station of

INRA Dijon in Chaux des Prs (France). Maximum quantum efficiency of PSII primaryphotochemistry (Fv/Fm) and leaf greenness were determined throughout two winter seasons.

Winter survival and shoot dry weight were assessed at the end of the winter. There was a

large variation in winter hardiness within the investigated material. The determination of

Fv/Fm in combination with leaf greenness permitted a differentiation of the accessions in

respect to their winter hardiness already at the beginning of the winter. A distinct grouping of

the investigated plants with regard to their ability to maintain frost tolerance was revealed by

an analysis of the time course ofFv/Fm.

3.2 Introduction

Breeding programs for winter hardiness have been traditionally empirical. Classical selection

for winter hardiness takes account of winter survival and frost damage or focuses on the yield

per seas a final integrator of all the processes occurring during the life of the plant. However,

meeting the challenges to increase winter hardiness will depend on a better understanding of

the underlying physiological processes of winter hardiness. As well as the cold acclimationprocess sensu stricto, which permits an enhanced tolerance to subzero temperatures when

plants are submitted to a period of low but positive temperatures, the second phase of

hardening confers increased tolerance to negative temperatures by exposure to negative but

non injurious temperatures (Livingston 1996). In wheat, temperatures as low as -12 C can be

tolerated by sensitive varieties, while tolerant genotypes can withstand temperatures of

-32 C. However, at early developmental stages, seedlings can hardly tolerate temperatures

below -12 C. Good winter hardiness requires a sufficiently long growth period for the plantto accumulate reserves. Furthermore, it requires a mean daily temperature below 15 C and


35/91

Chapter 3

23

between 3 and 5 weeks of low temperature (Lecomte et al. 2003). As well as the temperature

during the phase of cold acclimation, light intensity also plays an important role in acquiring

maximal frost tolerance (quist and Huner 2003). The freezing tolerance of winter wheat can

be increased when plants develop under a higher light intensity (Pocock et al. 2001). Hereby,

the excitation pressure seems to be of great importance, because a large number of genes,

which are induced at low temperature, can also be induced by light (Ndong et al. 2001).

Therefore, the response of the photosynthetic apparatus to low temperature seems to play a

major role in cold acclimation.

In order to determine the effects of environmental factors on the functioning of the

photosynthetic apparatus, chlorophyll fluorescence analysis has a great potential (Fracheboud

and Leipner 2003). Many authors evaluated frost tolerance by this method (Sundbom et al.

1982; Barnes and Wilson 1984; Lindgren and Hllgren 1993; quist et al. 1993). In

particular, the ratio of variable to maximal fluorescence (Fv/Fm), which reflects the maximum

efficiency of PSII primary photochemistry, appeared to be a reliable parameter to quantify

damage to the photosynthetic machinery (Bjrkman and Demmig 1987). After exposure of

cereals to freezing stress,Fv/Fmper se as well as the frost-induced decrease inFv/Fm(Fv/Fm)

show a high correlation with electrolyte leakage (Clement and Van Hasselt 1996; Pocock et

al. 2001; Rizza et al. 2001). Moreover, long-term cold stress results in a decrease ofFv/Fmin

particular when the cold stress was combined with high light intensity or/and when spring

wheat was exposed to these conditions instead of winter wheat (Hurry et al. 1992). Similar

results were obtained when the Fv/Fmof winter wheat was analyzed throughout the winter

season (Groom and Baker 1992). Furthermore the study of Hurry et al. (1992) revealed that

long-term cold stress under high light intensity leads to a reduction of the chlorophyll content.

To improve winter hardiness, the selection efficiency could be increased if specific attributes

related to frost tolerance could be identified and used as selection criteria for complementing

traditional plant breeding. The objective of this study was to examine the suitability of

chlorophyll fluorescence analysis in combination with the determination of leaf greenness as

selection criterion for winter hardiness within a selection of germplasm grown under field

conditions and to describe this material in terms of winter hardiness.


36/91


24

3.3 Material and Methods

3.3.1 Plant material and growth conditions

One hundred three wheat Triticum aestivum L. accessions from 34 countries distributed

worldwide were provided by Agroscope Changins-Wdenswil (Switzerland) (Table Appendix

1). These genotypes were bred to perform well as crops in contrasting winter climates from

continentally cold to subtropical cool; thus they present contrasting level of winter hardiness.

The varieties consisted of 66 winter, 24 spring, 3 intermediate and 10 non-determined wheat

varieties. The varieties were selected based on their degree of unrelatedness to obtain a

diverse test set of germplasm. Plants were grown at INRA Dijon altitude station at Chaux des

Prs, Dpartement Jura, France (46 30' N, 5 52' E, 875 m a.s.l.). The plants were grown

without snow cover by employing a movable shelter which was controlled by a snow sensor;

during snow fall the plot was covered by the shelter while for the rest of the time plants were

exposed to ambient conditions. Experiments were carried out over two seasons (winter

2004/05 and winter 2005/06). Sowing was implemented on the first week of October (5

October 2004 and 6 October 2005). To ensure regular sowing, seeds were attached to a strip

of paper with 2.5 cm between the seeds. Sowing depth was 2.5 cm. Thirty plants of a given

accession were sown in each row. The rows were 75 cm long with a inter-row distance of 10

cm. A replication was done in each year in the same shelter. Root damage due to freeze-thaw

cycles of the soil were minimized by adding peat to the soil every two years. The pre-winter

growth conditions were chosen to fulfill the requirements for hardening and so that all the

accessions reached their maximum tolerance before the first frost.

3.3.2 Phenotype assessment

Each accession was graded (1 to 9) to express the degree of freezing tolerance based on a set

of nine standard lines (Lecomte et al. 2003). The resistance of the standard lines is well

known and corresponds to a grade of the official scale of the French group of varieties and

seeds study and control (GEVES) (Table 3.1). The standard lines were described, after each

cold spell (two to three weeks after the leaves had defrosted) by damaged foliar surface

percentage and at the end of winter by their survival percentage (comparing the surviving

plants with the number of emerged plants). The survival counting were performed on 7 and 8

April 2005 and on 30 and 31 March 2006. The foliar damage of the tested plants was graded

by comparing them with the closest standard lines after each cold spell. They were annotated

using the closest standard line name, using if necessary "+" or "-" in the case a given variety ismore damaged than a standard line but less than the standard line of inferior tolerance coming


37/91

Chapter 3

25

immediately after it in the range. The survival percentage was determined similarly to the

standard lines. The final mark was attributed by comparison with the closest standard line first

considering the survival percentage and the extent of foliar damage. To determine the shoot

dry matter during re-growth, the shoots of ten plants of each accession in winter 2004/05 and

five plants in winter 2005/06 were collected in March, dried at 60 C for one week and

weighed. The photoinhibition was estimated by the maximum quantum efficiency of PSII

primary photochemistry (Fv/Fm) measured with a pulse amplitude modulation fluorometer(PAM-2000, Walz, Effeltrich, Germany). Themeasuring light modulation frequency was set

at 600 Hz to determine the minimum fluorescence (Fo) and at 20 kHz to measure the

maximum fluorescence (Fm). To induce Fm, a saturating pulse of approximately 8000

mol m-2s-1 PPFD for 0.8 s was applied. The Fv/Fm was calculated as (Fm-Fo)/Fm. The

chlorophyll fluorescence measurements were done monthly throughout the winter season on

the middle part of the second leaf with five measurements for each accession in each plot. The

Fv/Fmmeasurements were carried out at nightfall in order to let the plants adapt to the darkfor at least 20 minutes. The greenness of the leaves was assessed using a SPAD-502

chlorophyll meter (Minolta, Osaka, Japan). Measurements were done on the middle part of

the second leaf. Five measurements were done for each accession and values were averaged

by the SPAD-502 chlorophyll meter.

Table 3.1: Range of frost sensitivity of standard lines. The sensitivity threshold correspondsto temperature at appearance of the first foliar necrosis, performed at the INRA altitudestation of Chaux des Prs at optimally frost hardened plants.

Cultivar Sensitivity threshold Mark

Rex -12 C 1Magali -14 C 2Champlein -15 C 3Cappelle -18 C 4Capitole -21 C 5Moisson -24 C 6Arminda -26 C 7Comtal -28 C 8Cheyenne -32 C 9


38/91


26

3.4 Results

3.4.1 Temperature conditions

The first winter of experiment (2004/05) was characterized by a warm autumn (no frost in

October) and no strong cold observed until the 22 December (Fig. 3.1). January was ratherwarm with a sharp decrease in temperatures for the last week of the month to -20 C for 2

days and -24 C on 30 January. The beginning of February was cold and the end of the month

2004/05

Date

01.1

0.

01.1

1.

01.1

2.

01.0

1.

01.0

2.

01.0

3.

01.0

4.

Temperature(C)

-30

-20

-10

0

10

20

max. air temp.

min. air temp.

min. soil temp.

2005/06

Date

01.1

0.

01.1

1.

01.1

2.

01.0

1.

01.0

2.

01.0

3.

01.0

4.

Temperature(C)

-30

-20

-10

0

10

20

max. air temp.

min. air temp.

min. soil temp.

-28.9 C

-31.1 C

-28.4 C

-32.1 Cx x xx

x x xx

Figure 3.1: Daily maximum and minimum air temperature (at 2 m above ground) andminimum soil temperature (at soil surface) during the growth season in winter 2004/05 and2005/06. The minimum air and soil temperatures in each growth season are given. The dateson which physiological parameters were determined are marked by "x".


39/91

Chapter 3

27

was very cold (-26 C on 25 and 26 February). The first week of March was very cold and the

lowest temperatures of the season, -28.4 C and -32.1 C at soil level, were measured on 1

March. There was an abrupt increase in temperature from 15 March. The second winter

season (2005/06) was characterized by a warm autumn (no severe frosts until 15 November).

December was very cold; -20 C at the beginning and -28.9 C at the end. The soil surface

was -31.1 C on 30 December, the lowest temperature of the winter. Until the last 10 days of

March, there were frequent temperature drops below -20 C and only a few days with no

frost.

3.4.2 Comparison of tested accessions with standard lines

The 103 wheat accessions were classified according to winter hardiness in comparison to nine

standard lines (Fig. 3.2). Accessions from Australia, South America and southern Europe

were characterized by low winter hardiness grades compared to accessions from North

America, eastern Europe and northern Europe. There was a highly significant linear

correlation for winter hardiness grades between the two plots (r2= 0.978 in 2004/05 and r2=

0.954 in 2005/06) and between both winter seasons (r2= 0.906) (see Fig. 3.3).

Accession

Rex

Magali

Champlein

Cappelle

Capitole

Moisson

Arminda

Comatal

Cheyenne

Mark

1

2

3

4

5

6

7

8

9

Africa

Oceania

Asia

East Europe

South Europe

West Europe

North Europe

South America

North America

Figure 3.2: Classification of the 103 wheat accessions graded for winter hardiness, based onthe winter hardiness of nine standard lines. The accessions are colored according to thegeographic origin. Values are means SD of four replications (two plots in each of the twowinter seasons).

3.4.3 Effect of experimental season on winter survival

All 103 investigated wheat accessions behaved very similarly in relation to the nine standardlines in both winter; the grades for winter hardiness were closely related between the two


40/91


28

experimental seasons (Fig. 3.3 A). Looking on the survival rate, however, a considerable

difference was observed between the two experimental seasons. In 2004/05 a larger number

of accessions survived the winter than in 2005/06 (Fig. 3.3 B). In particular, accessions which

showed a moderate winter survival in 2004/05 had a very low survival rate in 2005/06. This

effect was also reflected in the shoot dry weight after re-growth (Fig. 3.3 C).

3.4.4 Time course of Fv/Fmand leaf greenness throughout the winter season

TheFv/Fmof the wheat plants was 0.726 (2004/05) and 0.747 (2005/06) at the beginning of

winter (Table 3.2). There was a considerable effect of year and leaf greenness. While in

Mark

1 2 3 4 5 6 7 8 9

2005

/06

1

2

3

4

5

6

7

8

9

Survival (%)

2004/05

0 20 40 60 80 100

0

20

40

60

80

100

Shoot dry weight (g plant-1

)

0.0 0.1 0.2 0.3 0.4 0.5

0.0

0.2

0.4

0.6

0.8

1.0

A B C

r = 0.906

r = 0.826

r = 0.635

Figure 3.3: Relationship of winter hardiness grades (A), winter survival (B) and shoot dryweight after re-growth (C) between the two experimental seasons (2004/05 and 2005/06) in

the 95 investigated wheat accessions. Values are means of the two plots.

Table 3.2: Mean, minimum and maximum of Fv/Fm and SPAD of the investigated wheataccessions in winter 2004/05 and winter 2005/06. Values are derived from means of tenmeasurements for each genotype.

Fv/Fm SPADmean min max mean min max

2004/05Nov 0.726 0.658 0.766 38.3 29.9 44.6Dec 0.723 0.586 0.775 37.9 26.6 48.2Jan 0.390 0.111 0.630 35.4 11.3 51.3Feb 0.214 0.000 0.683 25.5 4.0 54.7

2005/06Nov 0.747 0.632 0.784 48.6 39.4 56.7Dec 0.523 0.116 0.697 45.7 29.0 54.3

Jan 0.162 0.000 0.556 28.9 1.9 54.4Feb 0.011 0.000 0.391 1.0 0.0 36.0


41/91

Chapter 3

29

November 2004/05 the SPAD value was on average 38.3, it was 48.6 in November 2005/06.

When the temperature was lower than -15 C (see Table 3.3), bothFv/Fmand SPAD started to

decrease. At the end of winter 2005/06, most of the wheat accessions exhibited strong

necroses and, consequently,Fv/Fmand leaf greenness were close to zero.

There was a positive correlation between leaf greenness (SPAD) and Fv/Fm from the time

when starting to discriminate for these traits (Fig. 3.4). When minimum temperature was

frequently below -10 C (see Table 3.3), TheFv/Fmof sensitive genotypes declined and leaves

got slightly paler. The Fv/Fmof all wheat accessions declined further when the temperature

dropped down to a minimum of -20 C. This decline was much more pronounced on wheat

accessions which showed a poor rate of winter survival. The decrease in leaf greenness was

seemingly retarded compared to the decline in Fv/Fm as it can be seen from the data of

January 2005 and December 2005 (Fig. 3.4). At this time, a good discrimination between the

frost-tolerant and frost-sensitive genotypes could be done based on the combined

measurements of leaf greenness and Fv/Fm. The discrimination between accessions was even

more pronounced as the winter progressed and strong frost events were more severe.

Furthermore, there was a high linear correlation for leaf greenness between the two replicates

at this time: r2 = 0.679 (February 2005) and r2 = 0.835 (January 2006). The correlation

between both replicates was noticeably the same when considering SPAD as well as Fv/Fm

(data not shown). However a variance analysis of Fv/Fm and SPAD showed a small but

Table 3.3: Temperature conditions before and during determination ofFv/Fm.

Date min. temp.before meas.

average temp.during meas.

2004/0509.11. -9.0 C 2.1 1.420.12. -10.2 C 3.8 1.121.12. -10.5 C -2.6 1.526.01. -17.3 C -3.2 1.127.01. ditto -3.0 1.224.02. -26.4 C 2.8 1.725.02. ditto 1.9 1.9

2005/0619.11. -10.3 C -0.3 0.8

10.12. -19.7 C -1.5 1.908.01. -28.9 C 2.1 2.004.02. -28.9 C -3.8 0.6


42/91


30

significant difference between the replications in each year withpvalues of 0.0016 (2004/05)

and 0.0467 (2005/06).

The discrimination between the accessions was more marked in the second year of experiment

(2005/06) than in the first year (2004/05) at a given month. The genotypes with a higher level

of tolerance were affected by the cold sooner during the second year of experiment. The

ANOVA test indicated that differences among genotypes were highly significant (p= 2.10-16).

However, there was no genotype year interaction (p= 0.2245) indicating that genotypes

responded to the environment in the same way in each winter season.

The accessions could be assigned by hierarchical cluster analysis in six groups according to

their response of Fv/Fm throughout the two experimental seasons (Fig. 3.5). Two of the six

groups consisted of a few (F) or only one accession (C). Group A consisted of accessions

which showed a low winter survival in both winters. Most of the members of this group were

spring wheat accessions from South America (in particular Brazil) and from Australia. The

standard line "Rex" was a member of this group as well. Genotypes with moderate winter

Nov 04

Fv

/Fm

0.0

0.2

0.4

0.6

0.8

100 % Survival

50 % Survival

0 % Survival

Dec 04 Jan 05 Feb 05

Nov 05

SPAD

0 10 20 30 40 50

Fv/F

m

0.0

0.2

0.4

0.6

Dec 05

SPAD

0 10 20 30 40 50

Jan 06

SPAD

0 10 20 30 40 50

Feb 06

SPAD

0 10 20 30 40 50 60

Figure 3.4: Relationship between leaf greenness (SPAD) and maximum quantum efficiencyof PSII primary photochemistry (Fv/Fm) in the 103 wheat accessions throughout the twowinter seasons. The accessions were color-coded according to their winter survival rate for

the corresponding year. Values are means of 10 measurements (5 from each plot, with theexception of November 2004 with only one plot and 5 replications).


43/91

Chapter 3

31

survival in 2004/05 and low winter survival in 2005/06 were based on the Fv/Fm analysis

classified in group B. The decline inFv/Fmof the genotypes from this group was less strong

than that of the group A accessions. Most of the accessions in group B were from Argentina

and Chile but also from western Europe, in particular from France. Groups D and E consisted

of genotypes with a high winter survival in both winters. Both groups contained accessions

from western and eastern Europe and northern Europe, the latter especially in group E, which

was also characterized by a very high percentage of North American accessions. These two

groups were distinct in the decline of Fv/Fm in winter 2004/05; the Fv/Fm of group E

genotypes was clearly higher in February 2005 than the Fv/Fm of group D (Fig. 3.5). The

Fv

/Fm

0.0

0.2

0.4

0.6

0.8

Fv

/Fm

0.0

0.2

0.4

0.6

0.8

01.1

1.

01.1

2.

01.0

1.

01.0

2.

01.1

1.

01.1

2.

01.0

1.

01.0

2.

01.0

3.

Fv

/Fm

0.0

0.2

0.4

0.6

Fv

/Fm

0.0

0.2

0.4

0.6

01.1

1.

01.1

2.

01.0

1.

01.0

2.

Fv

/Fm

0.0

0.2

0.4

0.6

01.1

1.

01.1

2.

01.0

1.

01.0

2.

Fv

/Fm

0.0

0.2

0.4

0.6

2004/05 2004/052005/06 2005/06

A

DC

B

FE

4.7 10.9

0.2 0.8

49.7 37.9

18.1 25.5

31.0

5.0

93.9 11.2 71.5 26.3

99.5 1.3 86.0 8.9 100.0 0.0 100.0 0.0

Figure 3.5: Time course ofFv/Fmfor the 103 wheat accessions. The accessions were assigned

by hierarchical cluster analysis into six groups (A to F) according to their Fv/Fm response.Values are means of 10 measurements (5 from each plot, with the exception of November2004 with onl one lot and 5 re lications . The avera e winter survival SD for each cluster


44/91


32

highest winter survival and the smallest decline in Fv/Fmthroughout the winters were found

for the accessions in group F. The three accessions were the varieties Ridit (USA) and

Cheyenne (USA) as well as the landrace Wagrein 1371 (Austria).

3.5 Discussion

The climatic conditions of both winter seasons showed some common characteristics. The

warm autumn in 2004 and especially in 2005 allowed for fast growth and early tillering phase

at the beginning of November instead of mid-November when tillering usually occurs at the

experimental site in Chaux des Prs (personal communication Alex Giraud). Frost (about

-10 C) during this period had no effect on the maximu

Documents

Diss Caroline