International Conference on Heterosis in Plants · The vigor and stature of maize plants has been investigated in a comparative study at multiple ploidy levels. The monoploid is less

International Conference on Heterosis in Plants

University of Hohenheim, Stuttgart, Germany September 7 - 9, 2009

Page Program............................................................................................2-5 Abstracts for oral presentations........................................................6-50 Posters…………………………………………………………………..51-67 List of participants………………………………………………………68-73 Organizers and Organizing Committee……………………………… 74

1

2

International Conference on Heterosis in Plants University of Hohenheim, Stuttgart, Germany

September 7 - 9, 2009

07.09.09 Euroforum; Conference office opens at 7:30 Opening 9:00 Liebig, Hans-Peter; Rector of the University of Hohenheim, Stuttgart, Germany Melchinger, Albrecht E.; Conference Chair; University of Hohenheim, Stuttgart,

Germany Weber, Gerd; Research Center ‘Biotechnology and Plant Breeding’; University of

Hohenheim, Stuttgart, Germany Session 1: Exploring genetic hypotheses of heterosis; chair: Alain Charcosset 9:20 Zamir, Dani; Hebrew University of Jerusalem, Rehovot, Israel:

Genes that drive heterosis

9:50 Birchler, James; University of Missouri, Columbia, USA: Studies on heterosis in maize

10:20 – 10:50 Coffee Break

10:50 Rafalski, Antoni; DuPont, Willmington, USA: Copy number variation in the maize genome and heterosis

11:20 Abel, Stefan; Limagrain GmbH, Rosenthal, Germany: The effect of fixed heterosis in allopolyploid plants and implications for rape seed hybrid breeding

11:40 Becker, Heiko C.; University of Göttingen, Göttingen, Germany: Analysing fixed heterosis by comparative mapping of QTL for early biomass in Brassica napus, B. rapa and B. oleracea

12:00 Geiger, Hartwig H.; University of Hohenheim, Stuttgart, Germany: Does "marginal overdominance" contribute to heterosis in maize? First inferences based on transcript abundancy and enzyme activity studies

12:20 – 14:00

Lunch break and poster viewing

Session 2: Quantitative genetic analysis of heterosis in maize; chair: Qifa Zhang 14:00 Melchinger, Albrecht E.; University of Hohenheim, Stuttgart, Germany:

Quantitative genetic models and experimental designs for heterosis research

14:30 Jiansheng, Li; China Agricultural University, Beijing, China: Dissection of the genetic basis of heterosis in an elite maize hybrid by analyzing an immortalized F2 population with molecular markers

14:50 Schön, Chris-Carolin; TU München, Freising, Germany: High congruency of QTL positions for heterosis of grain yield in three crosses of maize

15:20 Frascaroli, Elisabetta; University of Bologna, Bologna, Italy: Heterosis in maize: from QTL analysis to development and evaluation of near isogenic lines for heterotic QTL

15:40 – 16:10 Coffee break

Session 3: Quantitative genetic analysis of heterosis in other species; chair: Mei Guo 16:10 Zeng, Zhao-Bang; North Carolina State University, Raleigh, USA:

QTL mapping and genetic basis of heterosis in maize and rice

16:40 Zhang, Qifa; Huazhong Agricultural University, Wuhan, China: Genetic and molecular bases of heterosis in rice

17:10 Sun, Qixin; China Agricultural University, Beijing, China: Genetic and molecular bases of heterosis in wheat (Triticum aestivum L.)

17:40 Ecke, Wolfgang; University of Göttingen, Göttingen, Germany: Mapping of QTL for heterosis in rapeseed (B. napus L.)

18:15 Reception at the castle of Hohenheim Welcome address by Albrecht E. Melchinger, University of Hohenheim and Patricia Schmitz-Möller, DFG, Bonn, Germany

19:00 Group photo

19:30 Dinner at Mensa

08.09.09 Euroforum

Session 4: New tools in heterosis research; chair:Thomas Altmann 9:00 Schnable, Patrick S.; Iowa State University, Ames, USA:

Heterosis in maize: new tools, complexities and insights

9:30 Guo, Mei; Pioneer Hi-Bred, Johnston, USA: The maize cell number regulator genes: implications for plant and organ size control, crop yield enhancement and heterosis

10:00 Mette, Michael Florian; IPK Gatersleben, Gatersleben, Germany: Relationship between intraspecies hybridisation and DNA/histone modifications in Arabidopsis thaliana L.

10:20 van Eeuwijk, Fred; Wageningen UR, Wageningen, The Netherlands: Detection of dominance at the QTL level in hybrid breeding programs


Session 5: Plant physiology and heterosis; chair: Günter Strittmatter 11:10 Altmann, Thomas; IPK Gatersleben, Gatersleben, Germany:

Molecular and genetic analysis of biomass-heterosis in Arabidopsis thaliana

11:40 Hochholdinger, Frank; University of Tübingen, Tübingen, Germany: Molecular dissection of heterosis manifestation in maize roots

12:00 Scholten, Stefan; University of Hamburg, Hamburg, Germany: Heterosis and associated gene expression patterns in early embryo and endosperm development

12:20 Weber, Hans; IPK Gatersleben, Gatersleben, Germany: Molecular physiology and genetics of seed heterosis in the model "Vicia faba L."

12:40 – 14:20 Lunch break and poster viewing

3

4

Session 6: From structure to function; chair: Roberto Tuberosa 14:20 Springer, Nathan; University of Minnesota, St. Paul, USA:

Potential implications for genome and transcriptome content variation on heterosis in maize

14:50 Piepho, Hans-Peter; University of Hohenheim, Stuttgart, Germany: The use of geostatistical methods in the analysis of cDNA microarrays for heterosis research

15.10 Snowdon, Rod; University of Gießen, Gießen, Germany: Comparative mapping of QTL and eQTL related to seedling and yield heterosis in oilseed rape

15:30 -16:00 Coffee break

16:00 Willmitzer, Lothar; MPI Golm, Potsdam, Germany:

Metabolomics in heterosis research

16:30 Gierl, Alfons; TU München, Freising, Germany: Metabolic phenotype of heterosis in maize

16:50 Andorf, Sandra; FBN Dummerstorf, Dummerstorf, Germany: Molecular network structures in heterozygotes: A systems-biology approach to heterosis

17:10 de Vienne, Dominique; INRA, Gif-sur-Yvette, France: Systemic properties of metabolic networks lead to an epistasis-based model for heterosis

18:00 Departure from Hohenheim to Bad Cannstatt

19:00 Kurhaus Bad Cannstatt: Conference Dinner

22:15 Departure back to Hohenheim

09.09.09 Euroforum Session 7: Prediction of heterosis and hybrid performance;

chair: Albrecht E. Melchinger 9:00 Maenhout, Steven; University College Ghent, Gent, Belgium:

Hybrid prediction through machine learning on commercial maize breeding data

9:30 Schrag, Tobias; University of Hohenheim, Stuttgart, Germany: Prediction of heterosis and hybrid performance in maize by marker-based approaches

9:50 Frisch, Matthias; University of Gießen, Gießen, Germany: Transcriptome-based prediction of heterosis in experimental hybrids of maize

10:10 Gärtner, Tanja; University of Potsdam, Potsdam, Germany: Improved heterosis prediction in Arabidopsis by combining information on DNA-markers and metabolites


Session 8: Establishing and broadening the genetic basis of heterotic groups; chair: Chris-Carolin Schön

11:00 Graham, Geoff; Pioneer, Johnston, Iowa, USA: Exploitation of heterosis for 80 years in a commercial hybrid breeding program of maize

11:30 Meng, Jinling; Huazhong Agricultural University, Wuhan, China: Broadening the avenue of intersubgenomic heterosis for oilseed Brassica

12:00 Reif, Jochen C.; University of Hohenheim, Stuttgart, Germany: Broadening the Central European heterotic groups in rye

12:20 Knapp, Steven; University of Georgia, Athens, USA: Mining allelic diversity for heterotic traits in desert-adapted native American Landraces and wild species in sunflower

12:50 Closing remarks

13:00 Lunch

14:30 – 17:00 Post congress tours: 1. Garden of the University of Hohenheim 2. German Museum of Agriculture 3. Experimental Research Station of Plant Breeding, University of Hohenheim, (Maize and Sunflower)

5

HETEROSIS IN PLANTS

Session 1:

Exploring genetic hypotheses of heterosis

Chair: Alain Charcosset

Sponsored by

6

Genes that drive heterosis

Dani Zamir¹, Uri Krieger¹, Zachary Lippman²

The Hebrew University of Jerusalem¹, Israel, Spring Harbor Laboratory², NY, USA

E-Mail: [email protected]

In 1955 Mather wrote ”Appreciation of the practical value of hybrid vigour is as old as the mule, but its scientific investigation began only relatively recently” – this sentence is still valid more than 50 years later. Our research employs tomato as a yield-heterosis model due to its ‘selfing’ mode of reproduction, which is accompanied by only modest effects of ‘inbreeding depression’ compared to maize, thereby allowing a better resolution of true overdominant genes and their phenotypic effects. We employ a diversity of phenotyped tomato populations to statistically characterize the mode of inheritance of individual component traits that drive yield and heterosis. We are focusing on heterotic chromosomal segments that improve yield, beyond the best parent, and are derived from the wild species S. pennellii introgression lines ILs: IL1-2 (40% over best parent), IL2-5 (50%) and IL8-3 (50%). The gene(s) responsible for IL heterosis are now being sought through high-resolution mapping followed by map-based cloning. We have also developed a novel approach to identify heterotic genes by producing hybrids that are parented by isogenic mutants 2,3. Such a screen has led to the identification of individual genes that drive heterosis when a single locus is heterozygous in an otherwise homozygous background (ODO-1 - 50% & ODO-2 - 38%). These observations provide strong support for the model that true-overdominant genes exist and can drive yield heterosis

7

8

Studies on heterosis in maize

James A. Birchler, Hong Yao, Nicole Riddle, Siva Chudalayandi, Akio Kato

University of Missouri, Columbia, Missouri, USA

E-Mail:[email protected]

The vigor and stature of maize plants has been investigated in a comparative study at multiple ploidy levels. The monoploid is less vigorous than the diploid. With increasing ploidy above the diploid level for inbred derivatives, the plant vigor and stature decline. However, duplex hybrids producing by crossing tetraploid inbred derivatives are vigorous. A further increase in vigor is achieved by producing double cross quadruplex hybrids that have the potential for four different alleles at any one locus, a phenomenon referred to as progressive heterosis. Thus, with maximum homozygosity vigor declines with increasing ploidy but with maximized heterozygosity vigor increases with ploidy level. With inbreeding of matched diploid and tetraploid hybrids, vigor decline occurs at similar rates in contrast to theoretical prédictions. Triploids derived from inbred lines have reduced vigor compared to the diploid progenitors. Reciprocal triploid hybrids that carry different numbers of alleles from the two parents exhibit consistently different magnitudes of heterosis indicating an impact of allelic dosage on hybrid vigor. Global patterns of gene expression in a 1-4x inbred ploidy series vary for many genes compared to the diploid but the differences are not of great magnitude. Nonadditive expression for many genes is observed in duplex tetraploid hybrids compared to the parents. A greater number of genes show nonadditive gene expression in the quadruplex hybrids and to a greater magnitude, but there is little overlap of nonadditive genes common to all hybrid states.

9

Copy number variation in the maize genome and heterosis

Andre Belo1,2, Mary K Beatty3, Bailin Li1, Antoni Rafalski1,2

¹DuPont Crop Genetics Research, Wilmington, DE, USA, ²Department of Plant and Soil Sciences, University of Delaware, Newark, DE, USA, ³Pioneer Hi-Bred, Johnston, IA, USA


Structural variations, including copy-number variations (CNVs), have been recently studied in mammals by using array-comparative genome hybridization (aCGH). They represent a class of DNA polymorphisms that can contribute to the determination of phenotypes, including gene expression and association with diseases. We investigated the presence of structural variations in the maize genome by aCGH. Digested genomic DNA of B73, Mo17 and other 12 maize inbred lines were individually labeled and hybridized to a gene-expression chip containing more than 100,000 60-mer probes. The position of each probe in relation to the B73 genome was estimated by BLAST of all probes against the B73 BAC sequences. The analysis of the ratio of the probes and their positioning against the reference genome (B73) allowed us to identify cluster of probes that deviated from the 1:1 ratio, which is expected in case of perfect colinearity. Such differences in the number of copies of a genomic segment (CNV) or the presence in one genome and absence in the other are possible structural variations. Some of these variants are large and comprise multiple genes. PCR and real-time qPCR experiments of selected regions confirmed the presence of such structural variations. Moreover, the comparison of different maize inbred lines against B73 revealed that many of these structural variants have moderate allele frequency in maize germplasm. Such structural variations comprise yet another type of genetic polymorphism potentially affecting maize phenotypes. We hypothesized that a subset of such structural variant differing between two parents of a hybrid may contribute to dominance heterosis for hybrid yield. Using a bootstrapping approach we were able to identify the variants that predict a significant fraction of hybrid yield.

10

The effect of fixed heterosis in allopolyploid plants and implications for rape seed hybrid breeding

S. Abel¹, H. C. Becker²

1Limagrain GmbH, Rosenthal, Germany ²University Göttingen, Göttingen, Germany


The spontaneous hybridisation of related species by combining their genomes (allopolyploidy) has played a prominent role in plant evolution. Many wild species as well as major field crops like wheat, oats, cotton, tobacco and rapeseed are allopolyploids. One explanation for the success of allopolyploids are favorable interactions between homoeologous genes of the parental genomes. These favorable interactions may also occur in homozygous allopolyploids and can therefore be called fixed heterosis. However, experimental data on fixed heterosis is not available, since the high performance of allopolyploids can also be attributed to an effect of polyploidy which is difficult to separate from the effect of fixed heterosis. In the present study Brassica genotypes have been created and analysed which are completely balanced in the allelic contributions but differ in the amount of heterozygosity (homozygous vs. heterozygous), ploidy (diploid vs. tetraploid) and genomic diversity (autotetraploid vs. allotetraploid). Based on a set of homozygous Brassica rapa (AA) and Brassica oleracea (CC) genotypes synthetic allopolyploid Brassica napus (AACC) as well as the corresponding autotetraploids (AAAA and CCCC) have been produced in order to quantify the effects of fixed heterosis and polyploidy. Moreover heterozygous genotypes have been analyzed to study the interrelation of fixed and classical heterosis.The magnitude of fixed heterosis averaged 50% relative to the parental mean. In contrast the effect of polyploidy was negative. The mean midparent classical heterosis in the diploid species was 62% (B. rapa) and 90% (B. oleracea). There is also a clear correlation between genetic distance and the magnitude of classical heterosis in the diploid species. In the allopolyploid B. napus the mean midparent classical heterosis was considerable lower (20%) and there is no correlation between genetic distance and the magnitude of classical heterosis.The relative small magnitude of classical heterosis is also a problem for rape seed hybrid breeding. In addition correlations between genetic distance and heterosis are rarely found. One explanation for these phenomena could be that the allopolyploid B. napus benefits already from heterotic interactions via fixed heterosis and therefore shows less classical heterosis than its diploid progenitors.

11

Analysing fixed heterosis by comparative mapping of QTL for early biomass in Brassica napus, B. rapa and B. oleracea

Franziska Wespel, Stefan Abel, and Heiko C. Becker

Georg-August Universität Göttingen, Department of Crop Sciences


The spontaneous hybridization of related species by combining their genomes (allopolyploidy) has played a prominent role in plant evolution. A main reason for the success of allopolyploids are the favourable interactions between genes on their homoeologous chromosomes which is similar to the positive interactions between different alleles at one locus in a heterozygote plant causing classical heterosis. Those favourable interactions between homoeologous loci should result in an increased performance of allopolyploids compared to their parental species, even in homozygous genotypes. Therefore, such positive interactions can be called “fixed heterosis”. Brassica napus (genome constitution AACC) is a very suitable model system to analyze “fixed” heterosis because artificial “resynthesized” lines can easily be developed from diploid parental species B. rapa (AA) and B. oleracea (CC). The aim of this project was to estimate the amount of fixed heterosis in a large number of different combinations and to perform a comparative mapping of QTL involved in fixed heterosis between the parental species B. rapa and B. oleracea and the allopolyploid. All possible resynthezised genotypes between 8 B. rapa and 8 B. oleracea lines were produced. The average amount of fixed heterosis for total biomass of young plants was about 50 %. For QTL mapping, segregating populations were established for the two parental species B. rapa and B. napus and for genetically balanced alloploid B. napus. The three populations were analyzed with 28 amplified fragment length polymorphism (AFLP) primer combinations resulting in 276 up to 297 marker points. In the allopolyploid population also 20 single sequence repeat (SSR) primer pairs were used to allow an alignment with genetic maps of former studies. In total, 29 QTL for early biomass were detected in the allopolyploid. Eighteen QTL occurred in B rapa, eight corresponding with QTL found in the allopolyploid. In B. oleracea 30 QTL could be observed of which eight correspond with QTL detected in the allopolyploid. Some of the detected QTL are located near already known “hot spots” for seed yield and heterosis in B. napus on the linkage groups N6, N7 and N16.

12

Does “marginal overdominance” contribute to heterosis in maize? First inferences based on transcript abundancy and

enzyme activity studies

Vera Kühr, Hartwig H. Geiger,

University of Hohenheim, Institute for Plant Breeding, Seed Science, and Population Genetics, Stuttgart, Germany


Heterosis, maize, transcript abundance, enzyme activity, enzyme profilingThis project refers to a genetic hypothesis proposed by Wallace (1968, Topics in Population Genetics, p.213) for explaining the phenomenon of heterosis. It assumes that a given allele may be superior to a second one under certain “conditions” but inferior under others. Such contrasting “conditions” could be different environments, organs, or developmental stages of an organism. If the heterozygote would under most conditions resemble the superior homozygote, the overall performance of the heterozygote may well surpass that of both parents even if no overdominance occurs under individual “conditions”. For this type of heterozygote superiority Wallace coined the term marginal overdominance. In our study we investigated Wallace’ hypothesis at the molecular level.Four unrelated inter-pool hybrids of maize were compared with their parent lines. Materials were raised in hydroponics (glasshouse) and soil (foilhouse and field). Tissue samples of different plant organs and leaf development stages were taken for RNA and protein analysis. Transcript abundance was assessed for gln4, d8, me1, me2, got1, got2, and got3. Enzyme activity was measured for SPP, SPS, ME and GOT, and enzyme profiling was performed for PEPC, NADP-MDH, PK, pPGI (plastidic PGI) and PGM.Transcript abundance and enzyme activity of the hybrids were mostly in the range of their parents. Overdominance was observed for various hybrid/gene(enzyme) combinations under individual “conditions”. Though the extent of allelic interaction varied over “conditions”, the type and sign of interaction generally remained unchanged. In a few cases the ranking of the two parent lines changed without altering the expression/activity level of the hybrid. Averaged across “conditions”, all cases of overdominance also showed overdominance under one or more individual “conditions”. In few of these cases, the parents changed their ranking under certain conditions. In conclusion, our gene expression and enzyme activity studies did not validate Wallace’ marginal overdominance hypothesis as an important explanation of heterosis. This does not mean that his ideas are wrong if applied to more complex, polygenic traits, i.e. as we walk from the simple molecular traits to those resulting from highly complex biological networks.

HETEROSIS IN PLANTS

Session 2:

Quantitative genetic analysis of heterosis in maize

Chair: Qifa Zhang

Sponsored by

13

14

Quantitative genetic models and experimental designs for heterosis research

Albrecht E. Melchinger

University of Hohenheim,350 Institute of Plant Breeding, Seed Science and Population Genetics, 70593 Stuttgart, Germany


Heterosis is widely exploited in plant and animal breeding, but its genetic basis is not well understood. Using an extended concept of system-level epistatic interactions, we developed a general theoretical framework to determine the contributions of different types of genetic effects to midparent heterosis (MPH) and identified additive-by-additive [aai] epistasis of individual locus i with the entire genetic background to be a component of MPH. On this basis we defined an augmented dominance effect di*, that comprises the dominance effect (di) at each QTL minus half the sum of [aai] interactions. With the design III involving recombinant inbred lines (RILs) and using composite-interval mapping (CIM), we demonstrated that genotypic expectations of heterotic QTL effects precisely equal genotypic expectations of MPH, thereby identifying genomic regions relevant for heterotic expression. Further, the theory for QTL mapping of multiple traits was applied to the simultaneous mapping of newly defined genetic effects to improve the power of QTL detection and distinguish between dominance and overdominance. Thus, it was shown that in the analysis of the genetic causes of heterosis we need to identify genomic regions that harbor effects di*. We also demonstrated that in the analysis of heterosis with triple testcross (TTC) design, dominance-by-additive interactions of individual QTL with the genetic background can be estimated with a one-dimensional genome scan using CIM. Further, quantitative genetic expectations of pairwise interactions between marker loci using two-way ANOVA were also given. Genetic resolution of QTL mapping is increased, when near-isogenic lines (NILs) are employed. We provided the theoretical basis for estimating different types of genetic effects with NILs and their TTC progenies, and showed that NIL populations have advantages over RILs for the power of QTL detection of genetic main effects and QTL-by-background interactions. Further, we extended the approach and proposed a two-step procedure to efficiently map additive-by-additive digenic epistasis: (i) an analysis of means of various generations of TTC design involving one-segment NILs to identify in one-dimensional scans loci showing QTL-by-background interactions, and (ii) use of one-segment NILs with significant additive-by-additive background interactions to produce two-segment NILs to test for additive-by-additive epistatic interactions between these segments.

15

Dissection of the genetic basis of heterosis in an elite maize hybrid by analyzing an immortalized F2 population with molecular

markers

Jihua Tang¹, Jianbing Yan¹,², Xiqin Ma¹, Wentao Teng¹, Weiren Wu³, Jingrui Dai¹, Albrecht E. Melchinger4 and Jiansheng Li¹

¹National Maize Improvement Center of China, China Agricultural University, Beijing, China; ²International Maize and Wheat Improvement Center (CIMMYT), Mexico, ³Institute of Bioinformatics, Zhejiang University, Hangzhou, China; 4Institute of Plant Breeding, Seed

Science and Population Genetics, University of Hohenheim, Stuttgart, Germany.


Heterosis has contributed greatly to the production of high grain yielding in some crop crosses during past century. The developing molecular markers and saturated linkage maps have provided new useful tools to analyze the genetic basis of heterosis. In this study, the genetic basis of heterosis for grain yield and its components was investigated at single- and two-locus levels using molecular markers with immortalized F2 (IF2) populations, which were constructed by a set of 441 pair crosses among recombinant inbred lines (RILs) derived from an elite maize hybrid, Yuyu22. Mid-parent heterosis for each cross in the set of IF2 population was used to valuate heterotic effects. The grain yield and its component traits were measured including shelled grain yield in metric tons/hectare (t/ha); row number on each ear; ear length (cm), and 100-kernel weight (g). In total, 283 SSR markers were used to constructed a genetic linkege map. A total of 13 heterotic loci (HL), which is defined as a locus expressing a significant difference in mid-parent heterosis between the heterozygote and the mean of the two homozygotes, were detected. They included three HL for grain yield, seven for ear length, one for row numbers of ears, and two for 100-kernel weight. All HL detected for the measured traits expressed over-dominant effects. A total of 143 digenic interactions contributing to heterosis were detected at the two-locus level. All three forms of digenic interactions, additive-by-additive (AA), additive-by-dominance (AD/DA), and dominance-by-dominance (DD), contributed to heterosis in IF2 population. There were 25 digenic interactions for grain yield, 36 for ear length, 31 for row numbers of ears, and 51 for 100-kernel weight. Totally HL at single-locus together with DD interactions at pairs of loci only accounted for 31.41% of heterotic effects for ear length, while the AD/DA interaction was speculated to contribute the residual heterosis to ear length. The effects of HL at single-locus, as well as DD and AD/DA interactions in two-locus may play important roles in the genetic basis of heterosis for grain yield and its components in Yuyu22.

16

High congruency of QTL positions for heterosis of grain yield in three crosses of maize

C.C. Schön, ¹B.S. Dhillon,² H.F. Utz² and A.E. Melchinger²

. ¹Plant Breeding, Center of Life and Food Sciences Weihenstephan, Technische Universität

München, Freising, Germany, ²Institute of Plant Breeding, Seed Science and Population Genetics, University of Hohenheim, Stuttgart, Germany


The genetic basis of heterosis in maize has been investigated in a number of studies but results have not been conclusive. Here, we compare QTL mapping results for grain yield, grain moisture and plant height from three studies conducted with the design III, all derived from the heterotic pattern Iowa Stiff Stalk Synthetic x Lancaster Sure Crop and all analyzed with advanced statistical methods specifically developed to examine the genetic basis of midparent heterosis (MPH). In two populations, QTL analyses were conducted with a joint fit of linear transformations Z1 (mean across pairs of backcross progenies) and Z2 (half the difference between pairs of backcross progenies) to estimate augmented additive and augmented dominance effects of each QTL as well as their ratio. QTL positions for the third population were obtained from the literature. For Z2 of grain yield, congruency of QTL positions was high across the three populations and a large proportion of the genetic variance (~ 70%) could be accounted for by QTL. This was neither the case for the other two traits nor for Z1. We conclude that different alleles have been fixed in each heterotic group which in combination with the allele(s) from the opposite heterotic group lead to high MPH for grain yield. Their positive interactions very likely form the base line for the superior performance of the heterotic pattern under study.

17

Heterosis in maize: from QTL analysis to development and evaluation of near isogenic lines for heterotic QTL

Elisabetta Frascaroli1, Giorgio Pea2, Pierangelo Landi1, Maria Angela Canè1,

Michele Morgante3, M. Enrico Pè2

¹University of Bologna, Bologna, Italy; ²Sant’Anna School of Advanced Studies, Pisa, Italy; ³University of Udine, Udine, Italy


Despite a century of investigations, the genetic bases of heterosis are not yet fully understood. We undertook a long term research in maize, aimed at providing a framework of comprehensive quantitative trait loci (QTL) phenotyping, to be integrated with map based cloning of genes contributing to heterosis. As a first step, classical genetic and QTL analyses were applied to genetic materials derived from the heterotic single cross B73 × H99. Recombinant inbred lines (RILs) were crossed to the testers B73 and H99, following a North Carolina III (NCIII) mating design and testcrosses were evaluated, together with the RILs, in three environments. Several QTL with heterotic effects on agronomic traits were detected and most of them were characterized by dominant or overdominant gene action, whereas non-allelic interaction proved to be of minor importance. On the basis of these results, genetic materials suitable for validation and precise estimate of the effects of heterotic QTL were developed. For this purpose, a residual heterozygous lines (RHL)-based introgression program was followed to produce pairs of near-isogenic lines (NILs), homozygous either for one or the other parental inbred allele at the selected heterotic QTL regions. Nine pairs of NILs, corresponding to six QTL, three of which in two genetic backgrounds, were obtained. NIL pairs were then investigated (i) per se and as cross between the two lines of the pair, to verify additive and dominance effects in inbred genetic background; (ii) in combination with related H99 and B73 inbred lines, to characterize QTL effects and dissect components of hybrid vigour in semi-inbred genetic background, at different stress levels; (iii) in combination with unrelated inbred lines (representing different breeding groups), to investigate if QTL effects were stable across maize genetic backgrounds. Results confirmed additive and dominance effects of heterotic QTL for traits that had shown high heterosis in our previous studies. For yield per plant and other agronomic traits, the contribution of additive effects declined from non-stress to stress condition, while dominance effects increased. QTL effects proved to vary to some extent according to unrelated tester's breeding groups. This study allowed us to detect heterotic QTL, to validate them in different genetic backgrounds and to verify their role in coping with stress environments due to high plant density.

HETEROSIS IN PLANTS

Session 3:

Quantitative genetic analysis of heterosis in other species

Chair: Mei Guo

Sponsored by

18

19

QTL mapping and genetic basis of heterosis in maize and rice

Zhao-Bang Zeng¹, Antonio Augusto Franco Garcia²

¹Departments of Genetics, Statistics & Bioinformatics Research Center, North Carolina State University, Raleigh, USA;

²Departamento de Genetica, Escola Superior de AgriculturaLuiz de Queiroz, Universidade de Sao Paulo, Brazil


Heterosis (or hybrid vigor) is a phenomenon in which an F1 hybrid has superior performance over its parents. It has been observed in many plant and animal species. The utilization of heterosis is responsible for the commercial success of plant breeding in many species and leads to the widespread use of hybrids in several crops and horticultural species. In maize {the most notable example, heterosis is the primary reason for the success of commercial industry. In China, hybrid rice varieties showed bout 20% yield advantage over inbred varieties and made a tremendous impact on rice production around the world. Despite its importance, the genetic basis of heterosis has been debated for almost one century and is still not explained satisfactorily. The dominance hypothesis suggests that the alleles from one parent are dominant over the alleles from the other parent, and due to the cancelation of deleterious effects at multiple loci the F1 hybrid is superior to the parents. The overdominance hypothesis assumes that the loci with heterozygous genotypes are superior to both homozygous parents. Epistasis is also frequently mentioned as a possible cause of heterosis. In a recent study (Garcia et al. 2008), we extended the multiple interval mapping method to Design III populations and applied the method to the data of Stuber et al. (1992) and Xiao et al. (1995). The re-analysis results strongly support the interpretation that dominance of favorable alleles is the main cause of heterosis in maize, whereas in rice the negative additive by additive epistatic effects may be the main cause for heterosis. Maize and rice evolved from a common ancestor but have different reproductive biology. This may explain the distinctively different genetic basis for heterosis observed in this and other studies. With open-pollination maize tend to have more deleterious recessive alleles than rice, masked by their corresponding dominant counterparts. When inbred, unfavorable alleles are expressed as homozygous, causing inbreeding depression. In self-pollinating species such as rice, deleterious alleles tend to be eliminated by natural (and artificial) selection since the individuals are homozygous. Therefore, outcrossing species could be selected for true dominant loci to avoid the expression of deleterious alleles, thus causing outbreeding advantage; whereas in self-pollinating species the selection for dominance is less important and, when a F1 cross shows heterosis, it is more likely due to epistatic interactions (aa) among loci.

20

Genetic and molecular bases of heterosis in rice

Qifa Zhang

National Key Laboratory of Crop Genetic Improvement and National Center of Plant Gene Research (Wuhan), Huazhong Agricultural University, Wuhan, China


Utilization of heterosis has greatly increased the productivity of many crops including maize, rapeseeds and rice. However, the biological mechanism underlying heterosis is still a debating issue. Our group has been working on various aspects of genetics and molecular bases of heterosis in rice using a combination of approaches, with the ultimate goal to achieve a comprehensive understanding of the biology of heterosis in rice, from the molecular mechanism to phenotypic performance. To fully characterize the genetic basis of heterosis, we focused our effort in one of the best hybrids, Shanyou 63, a cross between Zhenshan 97 and Minghui 63. This hybrid has been planted to the largest area in China in the last more than two decades because of high yield and wide adaptability. A range of experimental materials were developed from this cross, including F2, F2:3, recombinant inbred line (RIL) and “immortalized F2” populations. Molecular map-based analyses of these populations detected a large number of QTLs for agronomically important traits, and also identified many loci producing heterosis. In addition, the analyses also detected widespread digenic interactions in the rice genome. A general conclusion based on the genetic analyses is that accumulation of small advantages of the heterozygotes over the mid-parent across a large number of loci in the genome can adequately account for the genetic basis of heterosis in the F1 hybrid. Several lines of follow-up work have been undertaken to characterize the molecular basis underlying the genetic components of heterosis: (1) complete dissection and full recovery of the genetic effects of heterosis in this population with a SNP bin map making use of the next generation sequencing technology; (2) construction of chromosomal segmental substitution lines to evaluate the heterotic effects under near isogenic backgrounds; (3) transcriptomes of the parents and RILs to characterize the regulatory network and to associate the genotypes with phenotypes; and (4) molecular cloning of the major loci individually conferring large heterotic effects. I will present our current progresses in each of the areas in the symposium.

21

Genetic and molecular bases of heterosis in wheat (Triticum aestivum L.)

Qixin Sun Zhongfu Ni Yingyin Yao Huiru Peng Jinkun Du

China Agricultural University, Beijing 100183, China.


Whole genome expression analysis in hybrid and its parental inbreds provides a platform to understand the molecular mechanisms of heterosis. In this study, two wheat (Triticum aesticum L.) hybrid F1 derived from same female parent but displaying contrasting heterosis in primary root are used for expression analysis by using wheat genome array. The expression polymorphism analysis between the parental inbreds indicates that up to 4% genes display expression difference, but more than 3 times more present-absent genes between the two parental inbreds are detected in highly heterotic Hybrid A than in nonheterotic Hybrid B. Differential expression (DE) analysis in hybrids and their parental inbreds identify 1019 (4.94%) and 698 (3.23%) DE genes in Hybrid A and B, respectively. It is interesting to note that heterotic Hybrid A tends to have more DE genes of dominance and partial dominance expression modes than nonheterotic Hybrid B which, however, tends to have more DE genes of negative partial dominance expression mode. We also find that a substantial number of stress-related genes as well as retrotransposon-like and transposon-like genes are also included in the DE genes. We propose that as compared to the interspecific hybridization which can be a source of genomic shock as described by Barbara McClintock, hybrids derived from less distantly-related two inbreds can be a source of “mild genomic shock” or “intrinsic stress” in the hybrid genome, which, in turn, could cause expression changes of genes, especially stress-related genes and retrotransposon. Heterosis in internode elongation and plant height are commonly observed in hybrid plants, and higher GAs contents were found to be correlated with the heterosis in plant height. Our results indicated that among the 18 genes analyzed, genes encoding enzymes that promote synthesis of bioactive GAs, and genes that act as positive components in the GA response pathways were up-regulated in hybrid, whereas genes encoding enzymes that deactivate bioactive GAs, and genes that act as negative components of GA response pathways were down-regulated in hybrid. This correlation is also validated by using 16 wheat hybrids with different level of heterosis in plant height.

22

Mapping of QTL for heterosis in rapeseed (B. napus L.)

Wolfgang Ecke , Mladen Radoev, Heiko C. Becker

Department of Crop Sciences, Georg-August-University Göttingen, Germany


Heterosis is the superior performance of F1 hybrids relative to the midparent value or to the better parent. Three genetic hypotheses have been proposed which explain heterosis by the action of dominance, overdominance and epistatic effects. To elucidate the genetic basis of heterosis in rapeseed QTL were mapped for nine heterotic traits including early plant biomass, yield and yield related traits as well as plant height and phenological traits. Dominance, overdominance and epistatic effects of the QTL were estimated and compared. A doubled haploid population was developed from a cross between the winter rapeseed variety ‘Express’ and a resynthesized rapeseed genotype, ‘R53’. A total of 250 lines of the doubled haploid population were evaluated in field trials at four locations together with testcrosses of the lines with the parent ‘Express’. In addition, a linkage map comprising 235 SSR and 144 AFLP markers covering 2045 cM of the rapeseed genome was constructed in the doubled haploid population. Using this map QTL were mapped independently in the phenotypic data from the doubled haploid population, the testcrosses and a third dataset representing the midparent heterosis values of the doubled haploid lines, giving estimates of the additive, additive ± dominance, and dominance effects of the mapped QTL, respectively. For 27 QTL additive as well as dominance effects could be estimated. Of these, 13 showed partial to full dominance while 14 exhibited overdominance, indicating that both, dominance and overdominance effects are involved in the expression of heterosis in rapeseed. In addition, analysis of digenic epistatic interactions identified a large number of interacting locus pairs in each of the three data sets, clearly showing that epistasis also contributes to heterosis in rapeseed. The results indicate that, at least in rapeseed, all three of the major hypotheses on the genetic basis of heterosis are valid, but that none provides the sole explanation for the observed heterosis.

HETEROSIS IN PLANTS

Session 4:

New tools in heterosis research

Chair: Thomas Altmann

Sponsored by

23

24

Heterosis in maize: new tools, complexities and insights

Patrick S. Schnable1 Ruth A. Swanson-Wagner1, Rhonda DeCook2, Yi Jia1, Tieming Ji1, Xuefeng Zhao1, Li Li1, Dan Nettleton1, Kai Ying1, A. Leonardo Iniguez3, Heidi

Rosenbaum3, Eddy Yeh1, Jacob Kitzman3, Todd Richmond3, Tieming Ji1, Wei Wu1, Li Li1, Brad Barbazuk4, Dan Nettleton1, Jeffrey Jeddeloh3, Nathan Springer5, Yan

Fu1and The Maize Genome Sequencing Project1,6,7,8 1Iowa State University; 2University of Iowa State University; 3NimbleGen; 4University of

Florida; 5University of Minnesota; 6Washington University; 7Cold Spring Harbor Laboratory; 8University of Arizona, USA


A project to generate a near-complete sequence of the genome of maize inbred B73 was initiated in November 2005. This effort is utilizing a minimal tiling path of approximately 16,000 mapped BAC clones. The project is focusing on producing high-quality sequence coverage of the genic regions of the maize genome. These regions are being ordered, oriented and, along with all of the intergenic sequences, anchored to the extant physical and genetic maps of the maize genome. Following its domestication ~10,000 years ago, breeders have exploited the extensive genetic diversity of maize. The roles of structural variation, including insertions, deletions and copy number variation on the phenotypic diversity and plasticity of this important crop have not been elucidated. Whole-genome array-based comparative genomic hybridizations (aCGH) revealed that structural variation between the B73 and Mo17 inbreds is not evenly distributed across the genome. Analysis of altered segments of DNA identified hundreds of sequences that exhibit copy number variation among the two genotypes, as well as thousands of sequences that are present in B73 but not Mo17. Sequences present in B73, but not Mo17, genome include full-length genes, gene fragments and transposons. Although widely exploited in agriculture, the mechanisms responsible for heterosis are not well understood. Being monoecious it is possible to use a given maize plant as both male and female parents of crosses. Regardless of cross direction, the inbred lines B73 and Mo17 produce heterotic hybrids. These reciprocal hybrids differ phenotypically from each other despite having identical nuclear genomes. Consistent with these phenotypic observations, thousands of genes are differentially expressed between the reciprocal hybrids. An eQTL experiment conducted to better understand the regulation of gene expression in inbred and hybrid lines detected over ~4,000 eQTL associations. The bulk of eQTL act in trans. Surprisingly, for many of the trans-eQTL heterozygous eQTL differentially regulate transcript accumulation in a manner consistent with gene expression in the hybrid being regulated exclusively by the paternally transmitted allele. Because the designs of these experiments control for cytoplasmic and maternal effects, these findings suggest that widespread parent-of-origin effects may contribute to the phenotypic differences between reciprocal hybrids.

25

The maize cell number regulator genes: implications for plant and organ size control, crop yield enhancement and heterosis

Mei Guo1, Mary A. Rupe1, Jijun Zou1, Daniel Spielbauer1, Keith, Duncan2, Richard J. Howard2, Zhenglin Hou1 and Carl R. Simmons1

1Pioneer Hi-Bred, A DuPont Company, Johnston, USA2DuPont Crop Genetics Research, Wilmington, USA


Genes involved in cell number regulation may affect plant growth and organ size and ultimately yield. The tomato fruit weight gene fw2.2 for instance governs a QTL controlling 30% of fruit size variation, with increased fruit size chiefly due to increased carpel ovary cell number. To expand investigation of how related genes may impact other crop plant or organ sizes, we identified the maize gene family of putative fw2.2 orthologs, naming them ZmCNR (Cell Number Regulators). This family represents an ancient eukaryotic family of cysteine-rich proteins containing the PLAC8 or DUF614 conserved motif. We focused on native expression and transgene analysis on the two maize members closest to Le-fw2.2, namely ZmCNR1 and ZmCNR2. ZmCNR2 expression was found to be negatively correlated with growth activity in multiple tissue series analyzed, including maize leaf, embryo, endosperm and silk; that is, as these tissues matured, ZmCNR2 expression increased. Transgenic maize plants ectopically over expressing the ZmCNR1 transcript had reduced overall plant size, and this reduced size phenotype correlated at the molecular level to the transgene expression level. Six of seven of these transgenic events showed this size reduction, but one event which proved to co-suppress the ZmCNR1 expression resulted instead, in increased plant and organ size. This indicated that under- or over-expression of this single gene ZmCNR1 gene can push maize plant size either up or down respectively. Leaf epidermal cell counts showed that the reduced transgenic organ size was due to reduced cell number not cell size. These data are consistent with ZmCNR1 being a negative cell number regulator akin to LeFW2.2, but here operating in a different species and in more diverse plant organs. The effects of the ZmCNR gene on plant size and cell number call to mind heterosis, which increases plant size primarily through increased cell number. The exact mechanism(s) of heterosis is unknown, but it is commonly viewed as a complex multi-dimensional complementation of diverse gene allelic variation, but sometimes a smaller set of key growth-regulating genes dominating heterosis is envisioned. We are exploring how ZmCNR and related genes may contribute to native heterosis, or could mimic heterosis by increasing cell number to generate more vigorous and productive crop plants.

26

Relationship between intraspecies hybridisation and DNA/ histone modifications in Arabidopsis thaliana L.

Ali Mohammad Banaei Moghaddam1, Jörg Fuchs1, Michael Seifert1, Francois Roudier2, Tobias Czauderna1, Vincent Colot2, Marc Strickert1, Andreas

Houben1 and Michael Florian Mette1

1 Leibniz Institute of Plant Genetics and Crop Plant Research (IPK), Gatersleben, Germany; 2 CNRS UMR 8186, Ecole Normale Supérieure, Paris, France;


Arabidopsis thaliana accessions Col-0, C24, Cvi, and their reciprocal hybrids were employed as a model system to investigate the potential relationship between changes in DNA methylation, histone modifications, chromatin structure, endopolyploidisation and expression of heterosis. Nucleolus size, endopolyploidisation level and distribution of DNA and histone H3 methylation at the microscopic level do not differ between Col-0, C24, and their hybrids. Methylation sensitive amplified polymorphism (MSAP) revealed a largely constant pattern of DNA methylation (97% of signals analysed) Col-0/C24 crosses. The parental expression profile of selected genes was maintained in the hybrid offspring. No correlation was found between expression pattern and DNA methylation levels at restriction sites within 5’regulatory regions. Genome-wide ChIP on chip analysis revealed naive variations of histone H3K4 and H3K27 methylation among the tested accessions Col-0, C24 and Cvi. Parental histone H3K4 methylation patterns were additively inherited after hybridization in the offspring of Col-0/C24 and Col-0/Cvi crosses. Histone H3K27 methylation showed some dynamics in response to intra-species hybridization. The results are being scanned for chromatin modifications that potentially influence the expression of heterosis, which will enable us to draw conclusions on the importance of histone or DNA modifications on gene regulation in general and intra-species hybrid formation in particular. The work was supported by Deutsche Forschungsgemeinschaft in the frame of SPP 1149 “Heterosis in Plants”.

27

Detection of dominance at the QTL level in hybrid breeding programs

Fred van Eeuwijk¹, Radu Totir²,Marco Bink¹,Deanne Wright²,Martin Boer¹, Chris Winkler², Dean Podlich², Keith Boldman², Andy Baumgarten²,

Matt Smalley², Martin Arbelbide², Cajo ter Braak¹, Mark Cooper²

1Biometris, Wageningen UR, Wageningen, The Netherlands, 2Pioneer Hi-Bred International, Johnston, Iowa, USA


Although increasingly epistasis is thought to form the basis for heterosis, dominance remains a major mode of gene action to think of as a possible explanation. To test for relevant types of dominance at the QTL level one would be strongly interested to use the advanced stages of hybrid prediction programs. The germplasm in those stages provides a highly relevant genetic background within which to search for dominance. Recently suggested mixed model methodologies for QTL analyses allowing dominance did not target the rather special conditions of advanced hybrid prediction programs. We will discuss mixed model and Bayesian approaches for the detection of dominance at the QTL level in the very special conditions occurring in advanced hybrid prediction programs.

HETEROSIS IN PLANTS

Session 5:

Plant physiology and heterosis

Chair: Günter Strittmatter

Sponsored by

28

29

Molecular and genetic analysis of biomass-heterosis in Arabidopsis thaliana

Thomas Altmann¹, Berit Ebert², Barbara Kusterer³, Jan Lisec², David Riewe¹, Romy Schmidt4, Matthias Steinfath4, Kathleen Weigelt¹, Albrecht E.

Melchinger³, Joachim Selbig²,4, Lothar Willmitzer², Rhonda C. Meyer¹

¹Leibniz-Institute of Plant Genetics and Crop Plant Research (IPK), Gatersleben, Germany, ²Max-Planck-Institute of Molecular Plant Physiology, Potsdam-Golm, Germany, ³University

of Hohenheim, Germany, 4University of Potsdam, Germany


The molecular basis of heterosis is analysed in the C24 x Col-0 cross of Arabidopsis thaliana accessions that exhibits strong mid-parent heterosis of vegetative growth / biomass accumulation (up to 161% at high light). Using QTL-mapping in combination with metabolite profiling and gene expression profiling, loci responsible for the heterosis have been identified and heterosis-associated gene expression / metabolite composition has been analysed. Differences in mRNA and metabolite levels of parents and hybrids indicate additive, dominant, and overdominant effects with different prevalence at early and late developmental stages. In order to identify genomic regions involved in biomass heterosis QTL, generation means, and mode-of-inheritance classification analyses were performed using a modified North Carolina Design III: 429 recombinant inbred lines (RILs) and 140 introgression lines (ILs) were backcrossed to the two parental accessions and analysed for biomass accumulation at 15 days after sawing. Mid-parent heterosis of shoot dry weight in the RILs ranged from -31% to 99%. Treating mid-parent-heterosis and augmented dominance effect as quantitative traits, ten genomic positions involved in heterosis were detected, that explain between 2.4% and 15.7% of the phenotypic variation. While overdominant gene action was prevalent in heterotic QTL, the results suggest that a combination of dominance, overdominance and epistasis is involved in biomass heterosis in this Arabidopsis cross. A strong heterotic biomass QTL region of chromosome 4 that co-locates with a cluster of 23 metabolic heterotic QTL was narrowed down to 14 genes by using subILs and segregating IL families. This genomic region is further investigated by comparative sequence analysis (C24 vs. Col-0) and the candidate genes are studied using k.o. mutants subjected to crosses and by transformation of alleles into the opposite genotype.

30

Molecular dissection of heterosis manifestation in maize roots

Anya Paschold, Caroline Marcon, Nadine Hoecker, Frank Hochholdinger

University of Tuebingen, Center for Plant Molecular Biology, Department of General Genetics, Tuebingen, Germany


Heterosis describes the superior performance of heterozygous F1 hybrids compared to their genetically distinct parental inbred lines and is of paramount agronomic importance (Hoecker and Hochholdinger, 2007). Although, heterosis has been successfully exploited in maize (Zea mays L.) hybrid breeding for decades, the molecular basis of heterosis remains enigmatic. Heterosis is not only observed during adult development, but is already manifested during seedling development. The maize primary root, which is the first organ that emerges after germination, is a suitable model to study the early stages of heterosis manifestation. Various seedling root traits including primary root length, lateral root density, or the number of seminal roots display heterosis (Hoecker et al., 2006). Microarray studies suggest organ specific patterns of nonadditive gene expression in maize hybrids (Hoecker et al., 2008a). Moreover, such experiments support the notion that global expression trends in maize primary roots are conserved between different hybrids. Furthermore, nonadditive expression patterns of specific genes such as a SUPEROXIDE DISMUTASE 2 might contribute to the early manifestation of heterosis (Hoecker et al., 2008a). Proteome profiling experiments of maize hybrid primary roots revealed nonadditive accumulation patterns that were distinct from the corresponding RNA profiles underscoring the importance of posttranscriptional processes such as protein modifications that might be related to heterosis (Hoecker et al., 2008b). Finally, metabolite profiling experiments imply that a subtle regulation of particular biochemical pathways such as the phenylpropanoid pathway in hybrids might contribute to the manifestation of heterosis in maize primary roots (Hoecker et al., 2008b). In the future, recently developed molecular tools will facilitate the analysis of the molecular principles underlying heterosis in maize roots.

31

Heterosis and associated gene expression patterns in early embryo and endosperm development

Stephanie Meyer, Barbara Sarholz, Heike Pospisil, José F. Gutiérrez-Marcos, Hans-Peter Piepho and Stefan Scholten

University of Hamburg, Hamburg, Germany


Hybrid maize embryos exhibit a strong crossbreeding advantage as early as six days after fertilization. By allele-specific expression analyses we confirmed that maize, in contrast to the situation in other plants and in animals, has evolved a strategy to activate the paternal genome immediately following fertilization in zygotes before the first cell division. These immediate equivalent parental contributions provide an explanation for the heterotic proliferation rates of early hybrid embryo development. Hybrid endosperm also exhibits heterosis shortly after fertilization associated with differential gene expression of similar pattern as in embryos but to varying extent. The gene expression patterns in both tissues indicate gene regulatory interactions among the parental alleles and genes with consistent expression pattern among both tissues and the genotypes we analysed imply metabolic and epigenetic regulatory processes to be involved in specifying heterosis in the plant life cycle.

32

Molecular physiology and genetics of seed heterosis in the model “Vicia faba L.”

Tobias Meitzel, Ruslana Radchuk, Wolfgang Link, Hans Weber

Leibniz-Institut für Pflanzengenetik und Kulturpflanzenforschung (IPK), Gatersleben, Germany (T.M, R.R, H.W), Universität Göttingen, Department für

Nutzpflanzenwissenschaften, Göttingen, Germany


Heterosis describes the phenomenon of enhanced vigour and performance of hybrids relative to the means of their homozygote parents. Heterosis is expressed very early already on the level of growing seeds. Thus, when grown on the same mother plant, hybrid embryos become larger than inbred embryos indicating that the seed genotype displays a marked, significant heterotic impact on its own final seed weight. Using unique Vicia faba breeding material the influence of the genotype of growing seeds on their growth and final seed weight has been studied with the aim to comprehensively characterise and compare the development of inbred vs. pertinent crossed seeds on phenotypic, physiological and molecular level. Selfing and reciprocal crossing has been carried out with the Vicia faba inbred lines Hedin and Troy and both generations of seeds are balanced in the sense of showing the same average position in the pod and the same average node position on the plant stem, thus allowing valid, orthogonal comparisons. The heterotic impact on mature seed weight was determined to be about 9 %. Change of global gene expression was analysed using cDNA-arrays, encoding more than 5,500 mainly seed-specific genes from Pisum sativum cv. Fribo. Three stages of seed development were analysed, early, mid and late maturation, with three biological replications. Differential gene expression was analysed between selfed and crossed embryos; stage-specific mRNA populations from the two selfed seed genotypes and from the two crossed seed genotypes were combined, respectively. Identified genes upregulated in crossed versus selfed embryos were allocated into the following functional groups: C:N metabolism, cell proliferation, hormonal functions, mitochondrial activity and stress relation. The results indicate specific developmental shifts in crossed seeds compared to selfed seed.

HETEROSIS IN PLANTS

Session 6:

From structure to function

Chair: Roberto Tuberosa

Sponsored by

33

34

Potential implications for genome and transcriptome content variation on heterosis in maize

Nathan M. Springer

University of Minnesota, Saint Paul, Minnesota, USA


Heterosis arises when intraspecific variation is combined in a heterozygous individual. The recent availability of detailed information on population structure and genetic diversity in maize have provided new opportunities to study relationships of heterosis for different traits and the relationship of genetic distance and heterosis in a set of 300 maize hybrids. In general, heterosis is quite prevalent for nearly any quantitative phenotype measured but the amount of heterosis for different traits often shows low correlation. In order to further understand heterosis it will be important to characterize the molecular variation within a species and to study how this variation combines in hybrids. There is evidence for both quantitative and qualitative variation in the transcriptomes of different maize inbreds. Data from allele-specific expression assays and comparative genomic hybridization suggests that many of the major differences in gene expression levels can be explained by cis-acting regulatory variation. This cis-acting regulatory variation is expected to combine in an additive fashion in maize hybrids and this is frequently observed. In general, the majority of differentially expressed genes display transcript levels that could be described as additive, partially dominant or dominant. There are very few examples of hybrid transcript levels that are outside the parental range in the tissues and hybrids that have been studied in my lab. I will discuss the potential for both genome wide expression and CGH data to further illuminate certain aspects of heterosis in maize.

35

The use of geostatistical methods in the analysis of cDNA microarrays for heterosis research

Hans-Peter Piepho

University of Hohenheim, Stuttgart, Germany


The analysis of two-colour cDNA microarray data usually includes subtracting background values from foreground values prior to normalization and further analysis. This approach has the advantage of reducing bias, and the disadvantage of blowing up the variance of lower abundant spots. Whenever background subtraction is considered, it implicitly assumes locally constant background values. In practice, this assumption is often not met which casts doubts on the usefulness of simple background subtraction. In order to improve background correction we propose local background smoothing within the pre-processing pipeline of cDNA microarray data prior to background subtraction. For this purpose, we use a geostatistical framework with ordinary kriging using both isotropic and anisotropic models of spatial correlation. Using real data from a study on heterosis in maize as well as a self-versus-self experiment in Arabidopsis, we show that application of local background smoothing may be beneficial in comparison to using raw background estimates. Using locally smoothed background values increases the number of genotypic contrasts that are significant as well as the accuracy of contrast estimates.

36

Comparative mapping of QTL and eQTL related to seedling and yield heterosis in oilseed rape

Rod Snowdon¹, Panjisakti Basunanda¹, Bertha Salazar¹, Mladen Radoev², Wolfgang Ecke², Heiko Becker² , Wolfgang Friedt¹

¹Department of Plant Breeding, Justus Liebig University Giessen, Germany ²Department of Crop Sciences, Georg-August-University Göttingen, Germany


Little is known about the genetic control of heterosis in the complex polyploid crop species oilseed rape (Brassica napus L.). In this study, two large doubled haploid (DH) mapping populations and two corresponding sets of backcrossed test hybrids were analysed in controlled greenhouse experiments and extensive field trials for seedling biomass and yield performance traits, respectively. Genetic maps from the two populations, aligned with the help of common simple sequence repeat (SSR) markers, were used to localise and compare quantitative trait loci (QTL) related to the expression of heterosis for seedling developmental traits, plant height at flowering, thousand seed mass, seeds per silique, siliques per unit area and seed yield. QTL were mapped using data from the respective DH populations, their corresponding test hybrid populations and from mid-parent heterosis data, allowing additive and dominance effects along with digenic epistatic interactions to be estimated. A number of genome regions containing numerous heterosis-related QTL involved in different traits and at different developmental stages were identified at corresponding map positions in the two populations. The co-localisation of per se QTL from the DH population datasets with heterosis-related QTL from the mid-parent heterosis data could indicate regulatory loci that may also contribute to fixed heterosis in the highly duplicated B. napus genome. Given the key role of epistatic interactions in the expression of heterosis in oilseed rape, these QTL hotspots might harbour genes involved in regulation of heterosis (including fixed heterosis) for different traits throughout the plant lifecycle, including a significant overall influence on heterosis for seed yield. We are currently performing expression QTL analysis using Illumina EST tag profiling data in seedlings of our DH mapping population. The intention is to identify key trans-eQTL caused by transcription factors that might be responsible for the heterosis-related QTL clusters revealed by the classical QTL analysis.

37

Metabolomics in heterosis research

Lothar Willmitzer, Jan Lisec, Zoran Nikoloski, Joachim Selbig³, A. Melchinger², Thomas Altmann¹

¹Max-Planck-Institut für Molekulare Pflanzenphysiologie,Potsdam-Golm, Germany ¹Leibniz Institute of Plant genetics and Crop Plant Research (IPK), Gatersleben, Germany

²Institute of Plant Breeding, Seed Science and Population Genetics, University of Hohenheim, Stuttgart, Germany, ³Institute of Biochemistry and Biology,

University of Potsdam, Potsdam-Golm Germany


Metabolites are the last level of the realization of genetic information and the one molecular level most near to the final phenotype. In comparison to genetic markers, transcripts or proteins, by their very nature metabolites are well positioned to reflect complex genetic interactions such as the ones underlying heterosis. We thus have started a number of different approaches to see whether or not metabolomics could be a useful tool in describing, predicting or even contributing to an understanding of heterosis. To this end various populations of A. thaliana and to a lesser extent maize displaying different levels of heterosis have been analyzed using metabolomics technologies. These data have been mined by a variety of statistical tools as well as graph based methods. Results linking metabolic phenotypes to biomass and heterosis will be discussed.

Metabolic phenotype of heterosis in maize

Ulrich Genschel, Lilla Römisch-Margl, Alfons Gierl

Lehrstuhl für Genetik, Technische Universität München, Freising, Germany


Background Heterosis results in the phenotypic superiority of a hybrid over its parents with respect to traits such as growth rate, reproductive success and yield. Heterosis has been extensively exploited in plant breeding, particularly in maize. Growth and grain yield are the end results of a series of biochemical reactions. We studied the heterosis phenomenon with respect to metabolites to answer the following questions: i) Is the carbohydrate flux pattern associated with heterosis? ii) Is the metabolite pool size dependent from the hybrid state? iii) Is the metabolite pattern a potential marker for heterosis? Methods In order to quantify the contributions of the major pathways of central carbohydrate metabolism we used stable isotope labeling of developing maize kernels and quantitative 13C NMR. This methodology was applied to the B73 and Mo17 inbred lines, their F1-hybrids, and F2-progeny at three developmental stages ranging from 11 days post pollination (DPP) to 32 DPP. The levels of carbohydrates and amino acids, a total of 26 metabolites, were monitored throughout the grain filling at 8 to 30 DPP in the B73, Mo17 and B73�Mo17 lines using HPLC-PAD and GC-MS methods. Metabolite patters (112 metabolites) were determined in seedling roots of two European flint (UH002 and UH005) and two European dent (UH250 and UH301) lines and their six reciprocal hybrid pairs using GC-MS. Results i) The hybrid state revealed no significant reprogramming of relative carbohydrate fluxes in heterotic hybrids compared with inbreeds. Similarly, among several kernel mutants with strongly reduced yield and low starch content, only the mutants which lack AGPase, had a significantly altered flux pattern that was characterized by increased hexose cycling. Our results indicate that the carbohydrate flux pattern in maize kernels is very robust to genetic, developmental, and environmental perturbations. ii) The B73, Mo17, and B73�Mo17 genotypes differ significantly with respect to their carbohydrate and amino acid compositions. Approximately one third of the analyzed metabolites exhibited nonadditive heterotic effects. Among them high- and low-parent dominance, underdominance and overdominance were observed. iii) Hybrids and inbreeds differ in their metabolite pattern. There are significant differences between mid-parent heterosis values for metabolites grouped according to chemical composition.Conclusions Relative carbohydrate fluxes have little value as markers for heterosis in the maize kernel. The metabolite pool sizes and metabolite patterns are potential markers for heterosis.

38

39

Molecular network structures in heterozygotes: A systems-biology approach to heterosis

Sandra Andorf¹, Matthias Steinfath², Hanna Witucka-Wall³, Tanja Gärtner²,Thomas Altmann4, Joachim Selbig², Dirk Repsilber¹

²University of Potsdam, Potsdam, Germany, ³TeS-Lab/ Go:Incubator, Potsdam, Germany, 4IPK Gatersleben, Germany, ¹FBN Dummerstorf, Germany


Background: Heterosis is a well-known phenomenon, but the molecular mechanisms are not yet established. Early in heterosis research it was suggested that heterozygotes are likely to posses a greater biochemical versatility by carrying a greater diversity of alleles. Additional alleles at heterozygous loci may lead to additional regulatory interactions in the molecular network. Equipped with an enlarged repertoire of regulatory possibilities, hybrids may possibly be able to correctly respond to a higher number of environmental challenges leading to higher adaptability and, thus, the heterosis phenomenon. In our work, we took a systems biological approach to model this hypothesis, measured appropriate experimental data and tested the predictions of our model. Methods:Our systems biological model proposes differences in regulatory network structures between homozygous and heterozygous genotypes. It is based on a two-layer feed forward network also known as association matrix. We analyzed metabolite and gene expression profiles of Arabidopsis thaliana Columbia and C24 homozygous parents and their reciprocal crossings. Partial correlations were used to characterize the global interaction structure of the belonging regulatory networks. Results:Interpreting heterosis as increased adaptability, our model predicts that the biological networks involved show increasing connectivity of regulatory interactions. Furthermore, the model predicts a limit for increasing hybrid vigour with increasing heterozygosity. Our experimental metabolite as well as gene expression data show a system-wide increase in partial correlations in the examined heterozygous genotypes compared to their homozygous parents. These results support our systems biological hypothesis. Moreover, enrichment analyses revealed that the pathway of photosynthesis contains particularly enriched partial correlation heterosis effects. The predicted limit for increase of hybrid vigour is a known phenomenon in the literature. Conclusion: The early hypothesis of a greater biochemical versatility in heterozygotes, leading to additional regulatory interactions, was modelled taking a systems biological approach. Experimental data support the predictions of our model.

40

Systemic properties of metabolic networks lead to an epistasis-based model for heterosis

D. de Vienne¹, C. Dillmann¹, J. B. Fiévet²

¹Univ Paris-Sud, UMR 0320/8120 Génétique Végétale, Gif-sur-Yvette, France ²AgroParisTech, UMR 0320/8120 Génétique Végétale, Gif-sur-Yvette, France


The genetic and molecular approaches to heterosis usually do not rely on any model of the genotype-phenotype relationship. We chose the flux through a metabolic network as model quantitative trait, and used a generalization of the Kacser & Burns’ biochemical model for dominance and epistasis to analyse the genetic consequences of the joint variation of several enzymes of the network. We hypothesized that metabolic heterosis should be observed because the response of the flux towards enzyme activities and/or concentrations follows a multi-dimensional hyperbolic-like relationship. To corroborate this, we simulated the genetic variability of four enzymes of the upstream part of glycolysis by varying in silico enzyme concentrations. From each distribution of enzyme concentrations (i.e. each “parent”), and using previously known values of systemic parameters accounting for the kinetic behaviour of the enzymes, we computed the flux of each virtual parent. Then we “crossed” these parents to get 1000 hybrids, and compared the hybrid and parental flux values. Due to the convexity of the flux-enzyme relationship, mid-parent heterosis was observed. More interestingly, best-parent heterosis was observed whenever the distributions of enzyme concentrations were contrasted. The decomposition of the flux value into genetic effects, with the help of a novel multilocus epistasis index, revealed that antagonistic additive-by-additive epistasis effects play the major role in this framework of the genotype-phenotype relationship. This result is consistent with various observations in quantitative and evolutionary genetics, and provides a model unifying the genetic effects underlying heterosis.

HETEROSIS IN PLANTS

Session 7:

Prediction of heterosis and hybrid performance

Chair: Albrecht E. Melchinger

Sponsored by

41

42

Hybrid prediction through machine learning on commercial maize breeding data

Steven Maenhout¹, Geert Haesaert¹, Bernard De Baets²,

¹University College Ghent, Belgium, ²Ghent University, Belgium


Commercial maize breeding programs rely on extensive field trials to assess the genetic potential of new hybrids. A vast amount of phenotypical information on the offspring of new parental crossings is collected on a yearly basis. The basic idea of this research is to combine this “free” phenotypical data with molecular marker information, in an attempt to construct a prediction model that should allow to assess the potential of new parental combinations prior to field evaluation. All individual plot measurements that are gathered during the course of a commercial breeding program represent an inherently unbalanced data set of enormous proportion. We developed a heuristic which identifies the optimal data subset that contains a predefined number of parental genotypes. The optimization criterion makes a trade-off between the data structure (level of balance) and the information content (number of measurements) such that the prediction error variance on estimated breeding values is minimised. A linear mixed model analysis allows to estimate breeding values from unbalanced phenotypical data. Usually, genotypical effects are assumed to be random random variables and some estimate of the coefficient of coancestry is used to model their covariance. Taking advantage of the specific heterotic structure of hybrid breeding pools, we developed the Weighted Alikeness In State (WAIS) coancestry estimator. This new measure of genetic similarity is ideally suited for use in variance structures of linear mixed models, as it guarantees, contrary to most other estimators, a positive semi-definite coancestry matrix. We show that WAIS produces a low root mean squared error under different breeding circumstances and provides a fit to the data that is comparable to that of several other marker-based alternatives. The actual prediction model is constructed by means of ε-insensitive support vector machine regression (ε-SVR). The prediction accuracy is determined by means of cross-validation routines and a balanced validation field trial. We show that the results of both validation methods can differ substantially. In general, the obtained prediction accuracy of ε-SVR is comparable to that of a Best Linear Prediction-based method and at an acceptable level for the traits grain moisture content and days until flowering. Predictions on grain yield, however, remain problematic.

Prediction of heterosis and hybrid performance in maize by marker-based approaches

Tobias A. Schrag1, AE Melchinger1, M Frisch2

1 Institute of Plant Breeding, Seed Science and Population Genetics, University of Hohenheim, Germany

2 Institute of Agronomy and Plant Breeding II, Justus-Liebig-University Giessen, Germany


In maize hybrid breeding, the majority of the large number of possible inter-pool cross combinations cannot be produced and tested in field trials due to economical constraints. However, the performance of such untested hybrids can be predicted using molecular markers. Marker-based approaches for prediction of heterosis and hybrid performance in factorial crosses were developed and evaluated. Quantitative trait loci were identified for grain yield and grain dry matter content, using field data of up to 400 hybrids as well as field and marker data of their 79 inbred parents. Prediction based on estimated marker effects was compared with prediction from general combining ability (GCA). Cross-validation procedures were designed such that the situation in applied breeding programs was mimicked. The efficiencies for marker-based predictions of hybrids between two tested parental lines were obtained by cross-validation and were nearly as high as with GCA-based predictions. In cross-validation procedures for grain yield prediction of hybrids having one or none tested parents and a very unbalanced data structure, marker-based approaches showed equal or superior efficiencies for prediction of untested hybrids if pedigree information was absent. Application of marker-based prediction is possible at several stages during hybrid maize breeding, e.g., after production of new doubled haploid lines, after testcross evaluation of new lines, or after field testing of experimental hybrids. In the near future, the use of marker-based prediction will benefit from recent developments of high throughput platforms for molecular markers, ongoing increase of available computing power, and the implementation of the doubled haploid technique in plant breeding programs.

43

Transcriptome-based prediction of heterosis in experimental hybrids of maize

Matthias Frisch, Alexander Thiemann, Junjie Fu, Tobias A. Schrag, Stefan Scholten, Albrecht E. Melchinger

Institute of Agronomy and Plant Breeding II, Justus-Liebig-University Giessen, Germany


Grouping of germplasm and prediction of hybrid performance and heterosis are important applications in hybrid breeding programs. Gene expression analysis is a promising tool to achieve both tasks efficiently. Our objectives were to (a) investigate distance measures based on transcription profiles, (b) compare these with genetic distances based on AFLP markers, and (c) assess the suitability of transcriptome-based distances for grouping of germplasm and prediction of hybrid performance and heterosis in maize. We analyzed transcription profiles of RNA extracted from seedlings of the 21 parental maize lines of a 7$\times$14 factorial with a 46k oligonucleotide array. The hybrid performance and heterosis of the 98 hybrids was assessed in field trials. In cluster and principal coordinate analyses for germplasm grouping, the transcriptome-based distances were as powerful as the genetic distances for separating flint from dent inbreds. Cross validation showed that prediction of hybrid performance with transcriptome-based distances using selected markers was more precise than earlier prediction models using DNA markers or general combining ability estimates using field data. Our results suggest that transcriptome-based prediction of hybrid performance and heterosis has a great potential to improve the efficiency of maize hybrid breeding programs.

44

Improved heterosis prediction in Arabidopsis by combining information on DNA-markers and metabolites

Tanja Gärtner¹, Matthias Steinfath¹, Sandra Andorf², Jan Lisec³, Rhonda C. Meyer4, Thomas Altmann4, Lothar Willmitzer³, Joachim Selbig¹

¹University of Potsdam, Potsdam, Germany, ²FBN Dummerstorf, Dummerstorf, Germany, ³MPIMP, Potsdam-Golm, Germany, 4IPK Gatersleben, Gatersleben, Germany


Hybrids represent a cornerstone in the success story of breeding programs. The fundamental principle underlying this success is the phenomenon of heterosis. It describes an advantage of the offspring as compared to the two parental lines with respect to physiological traits such as growth. The heterosis level clearly depends on the combination of the parents used for offspring production. This results in a major challenge for plant breeders, as usually several thousand combinations of parents have to be tested for identifying the best combinations. Thus, any approach to reliably predict heterosis levels based on properties of the parental lines would be highly beneficial for plant breeding. In addition, the identification of markers essential for heterosis prediction should contribute to the understanding of the molecular basis for heterosis. We compared the predictive power of parental metabolites and DNA-markers for biomass heterosis. Therefore we analyzed experimental data on a recombinant inbred line population of Arabidopsis thaliana derived from the accessions C24 and Col-0. We used linear regression methods in combination with a variable selection algorithm and a minimum description length based strategy. Recently, genetic markers have been used to predict heterosis. Here we show that more integrated markers such as parental metabolite levels are similarly predictive for heterosis. Interestingly, a linear combination of parental genetic and metabolic markers significantly improves the prediction of biomass heterosis as compared to pure genetic linear models. A possible explanation is, that metabolite levels reflect parts of the complex gene interactions involved in the manifestation of heterosis. Heterosis prediction has been improved by combining genetic and metabolic markers. Our findings will help furthering the understanding of the molecular basis of heterosis, revealing, for instance, the presence of nonlinear genotype-phenotype relationships. In addition, we get hints how to further improve models for accelerated selection in plant breeding.

45

HETEROSIS IN PLANTS

Session 8:

Establishing and broadening the genetic basis of heterotic groups

Chair: Chris-Carolin Schön

Sponsored by

46

47

Exploitation of heterosis for 80 years in a commercial hybrid breeding program of maize

Graham, Geoff, Dean Podlich, and Andy Baumgarten

Pioneer Hi-Bred International, Inc., Johnston, Iowa, USA


Pioneer Hi-Bred International, Inc., has had an active corn breeding program since the early 1920’s. The company was founded based on the concept that hybrid vigor was a better solution to improving corn yields compared to open pollinated varieties. Since the inception of the company, average corn yields in the United States have increased over 5-fold. While not all of this can be contributed to genetic gain, there is no question the effect of a long-term reciprocal recurrent selection program has significantly improved modern maize hybrids. In this talk we will describe phenotypic and genetic changes to our germplasm pool as a result of 80 years of selection.

48

Broadening the avenue of intersubgenomic heterosis for oilseed Brassica

Jun Zou and Jinling Meng

National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, Wuhan, China


Accumulated evidence shows that the long history of intensive artificial selection in different Brassica species has resulted in profound genomic changes in Brassica crops. We distinguish three sets of basic genomes and subgenomes in a modified U’s triangle, building on the cytogenetic relationships of Brassica species and their genomes identified last century. For example, the A subgenome from B. rapa is identified as Ar, whereas the A subgenome from B. napus is identified as An. To utilize the genetic difference between subgenomes, we partially replaced the An and Cn genome of B. napus by the Ar genome of B. rapa and Cc genome of B. carinata, resulting in the first generation of new type B. napus. Heterosis between “normal” B. napus and the new type B. napus can therefore be defined as “intersubgenomic heterosis”. Significant intersubgenomic heterosis was observed in hybrids between the new type B. napus and normal B. napus over the past ten years of field trials. By selecting the plants with higher proportion of exotic-subgenomic components from segregated populations of first generation of new type B. napus, a second generation of new type B. napus was developed with stronger intersubgenomic heterosis. To increase the content of the exotic-subgenomic components further and to enable a sustainable breeding for novel lines of new type B. napus, we proposed an idea of developing a gene pool for new type B. napus with plenty of genetic variation at both Ar/Cc genome. New evidence has shown that, after interspecific crossing, great genomic alterations would happen in new type B. napus far beyond the introgression of exotic genomic components, and the genepool would enable us to carry out a recurrent selection for the alleles with positive contribution to the intersubgenomic heterosis. New approaches in broadening the avenue of intersubgenomic heterosis for oilseed Brassica are discussed.

49

Broadening the Central European heterotic groups in rye

S. Fischer1, A. E. Melchinger¹, V. Korzun4, P. Wilde4, B. Schmiedchen4, J. Möhring², H.-P. Piepho², B.S. Dhillon¹, T. Würschum³, and J.C. Reif³

1 Institute of Plant Breeding, Seed Science and Population Genetics,Univ. of Hohenheim, Stuttgart, Germany; ²Bioinformatics Unit, Institute for Crop Production and Grassland

Research, ³State Plant Breeding Institute, Univ. of Hohenheim, Stuttgart, Germany; 4KWS LOCHOW GMBH, Bergen, Germany


Broadening the genetic base of heterotic pools is a key to ensure continued genetic gains in hybrid breeding and extend hybrid cultivation to new areas. In the present study two Central European heterotic pools (Carsten and Petkus) and five Eastern European open-pollinated varieties (OPVs, Pop-1 to Pop-5) were studied with the objectives to: (1) investigate the genetic diversity in OPVs and the heterotic pools using molecular and field data, (2) evaluate the molecular diversity among OPVs, (3) examine the combining ability for grain yield of the OPVs when crossed with testers in field trials, (4) and develop a strategy for targeted introgression of OPV germplasm into the heterotic pools. In total, 610 S0 clones, 347 from OPVs and 263 from heterotic pools, were developed. The clones of OPVs were crossed with two testers belonging to each heterotic pool, while clones of heterotic pools were crossed with only the opposite tester. Testcrosses were evaluated for grain yield in multi-location trials. In addition, 589 S0 clones were fingerprinted with 30 SSR markers. The data revealed that the Carsten pool has a narrow genetic base and should be the primary target for broadening the established heterotic pattern. Mean and genetic variance suggested that Pop-2 and Pop-4 are good candidates for introgression in Petkus pool and Pop-5 in Carsten pool. Nevertheless, introgression of Pop-5 in Carsten could reduce the genetic diversity between heterotic pools. Therefore, we suggest that either selected clones of Pop-5 should be introgressed or more Eastern European germplasm should be fingerprinted and field evaluated to identify promising germplasm for broadening the established heterotic pattern.

50

Mining allelic diversity for heterotic traits in desert-adapted native American landraces and wild species in sunflower

Eleni Bachlava, Jessica Barb, Sukhpreet Sandhu, Chris Taylor, Dong Zhang, Juan Rey, Sameer Khanal, Jennifer Wood, Jason Prothro, Suzie Fatkin,

Rebecca Okahshah, Nelly Khalilian, and Steven J. Knapp

Institute of Plant Breeding, Genetics, and Genomics, University of Georgia, Athens, USA


Sunflower (Helianthus annuus L.), a heterotic, outbreeding diploid (x = 17), carries a tremendous genetic load. Since the discovery of cytoplasmic-genic male-sterility, development of first-generation elite inbred lines, and emergence of the hybrid sunflower seed industry in the 1970s, seed yield gains have been modest in the US, Europe, and elsewhere (historical records are lacking for the pre-hybrid era). This poses a serious problem in an increasingly competitive marketplace where the pressures of population growth and global climate change and increased demand for food, feed, and fuel have steadily driven seed yields upward in other species. The yield gains in sunflower starkly contrast with those in maize (Zea mays L.), for example, where significant yield gains have been achieved year over year for several decades. Why? First, several non-genetic factors have negatively impacted sunflower production around the globe, e.g., production on marginal lands and slow adoption of modern hybrids. Second, selection for increased seed yield has not been aggressively and systematically applied on a large scale. Lastly, the hybrid sunflower seed industry was built around and grew out of a single public hybrid (HA89 x RHA274)__still widely grown in many permutations__heterotic groups and patterns are not sharply delineated, and the discovery and transfer of novel favorable alleles for seed yield from elite and exotic sources into the parents of elite single-cross hybrids has been limited. We describe several research activities in our laboratory geared towards significantly increasing seed yields in sunflower, primarily by genome-wide selection among progeny from elite x elite crosses within heterotic groups and by identifying and transferring favorable alleles from Native American land races and desert-dwelling wild species currently not found in the parents of elite single-cross hybrids. Significant resources have been developed for genomics-assisted discovery and marker-assisted breeding (MAB) in sunflower. Next-generation DNA sequencing and single nucleotide polymorphism (SNP) genotyping technologies are enabling further, rapid expansion of genomic resources__high-density genetic mapping and sequencing of the sunflower genome (3,500 Mbp) are underway. Phenotyping continues to be a serious bottleneck for many traits and will be critical for increasing seed yield and the global competitiveness of sunflower. Genome-wide selection and analogous approaches have become a practical reality with the emergence of highly parallel SNP genotyping and have the potential to partially alleviate the problem.

HETEROSIS IN PLANTS

Posters

51

Models for predicting hybrid performance among a subset of parental accessions from the USDA ALFALFA core collection

A.H. Al Lawati1, I.M. Ray2

1Directorate General of Agriculture and Livestock Research, Ministry of Agriculture, Oman 2Department of Plant & Environmental Sciences, New Mexico State University,

Las Cruces, U.S.A.


Utilization of nuclear genetic distance (GD) information based on molecular markers has been proposed to help reduce the time required, and resources needed, for selecting proper parents that produce heterosis among their F1 progenies. In alfalfa, this process usually takes at least four years. In the first year, hybrids are generated between several parents. In the second, third, and fourth years, forage yield general combining ability (GCA) of the parents and specific combining ability (SCA) of the individual hybrids are estimated. The purpose of this study was to evaluate if regression models based on GD alone, or in various combinations with the variables of GCA, variety effects (VE), fall regrowth (FR), winter-hardiness (WH), and chloroplast genetic distance (Chlor. GD) of nine parental populations were able to predict the performance and heterosis of their 36-diallel hybrids. These variables were also tested alone, or in various other combinations, to predict hybrid performance and heterosis. The results revealed that AFLP genetic distance was able to predict yield performance for 63% of 19 superior hybrids. The best single model to predict hybrid performance was GCA. It was able to identify 79% of superior hybrids. Combining GD with GCA did not improve the prediction ability to identify superior hybrids. WH and GCA+ Chlor. GD models were able to identify 74% of the superior hybrids. Chloroplast genetic distance was able to predict 53% of the 19 superior hybrids. High parent heterosis was significantly predicted by FR (r=0.47) and WH (r=0.33). Mid-parent heterosis was predicted by GCA (r=0.44), VE (r=0.71), FR (r=0.51) and WH (r=0.46), but not by GD (r=0.18). The implication of these results is that evaluation of VE, and FR or WH, could offer a practical approach to predict hybrid performance and heterosis among the parents used in this study. Thus, resources that would be normally used to estimate the GCA of “n” parents could be used to evaluate VE of [n(n-3)/2] populations.

52

A molecular mechanism for gene expression changes and growth vigor in hybrids and allopolyploids

Z. Jeffrey Chen, Enu-Deok Kim, Misook Ha, Erika Lackey, Danny Ng, Changqing Zhang

The University of Texas at Austin, Austin, USA


Hybrids produced within and between species often grow bigger and more vigorously than the parents. Some crops including corn and rice are grown mainly as hybrids, and many other crops such as wheat, cotton, and canola are grown as allopolyploids. An allotetraploid is a “doubled interspecific hybrid”. We use the model plant Arabidopsis to study the mechanisms for gene expression changes in hybrids and allotetraploids. Arabidopsis suecica is a natural allotetraploid, and new allotetraploids are readily resynthesized by hybridizing the related species Arabidopsis thaliana and Arabidopsis arenosa. We found genome-wide nonadditive expression of homoeologous genes in the allotetraploids and expression and phenotypic dominance of A. arenosa over A. thaliana. The nonadditively expressed genes include microRNA targets and the genes encoding transcription factors. One set of transcription factors controls circadian clock regulation in Arabidopsis. Circadian clocks mediate metabolic pathways and increase fitness in plants and animals. We showed that epigenetic modifications of the circadian clock genes CIRCADIAN CLOCK ASSOCIATED 1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY) and their reciprocal regulators TIMING OF CAB EXPRESSION 1 (TOC1) and GIGANTEA (GI) mediate expression changes in downstream genes and pathways. During the day, epigenetic repression of CCA1 and LHY induced expression of TOC1, GI and downstream genes that contain CCA1 binding site (CBS) in chlorophyll and starch metabolic pathways in allotetraploids and F1 hybrids, which produced more chlorophyll and starch than the parents in the same environment. Daily repression of cca1 in TOC1:cca1-RNAi transgenic plants increased expression of downstream genes and chlorophyll and starch content, whereas constitutively expressing CCA1 or ectopically expressing TOC1:CCA1 had the opposite effects. The causal effects of CCA1 on output traits suggest that hybrids and allopolyploids gain advantages from the control of circadian-mediated physiological and metabolic pathways, leading to growth vigor and increased biomass.

53

Combining ability of Guinea and Guinea-Caudatum Sorghum hybrid parents in contrasting environments in Mali

Abdoulaye Diallo¹, H.F.W. Rattunde², Eva Weltzien²

¹Institut d’Economie Rural, ¹International Crops Research Institute for the Semi-Arid Tropics,

Bamako Mali


The first series of sorghum hybrid parents specifically developed for the Sudanian zone of West Africa are based on landrace- and improved breeding materials of Guinea-race origin as well as Guinea x Caudatum inter-racial breeding products. The combining abilities of these diverse sorghum hybrid parents and the performance of hybrid groups of differing parental-background combinations were assessed. Five females parents (3 Inter-racial and 2 Guinea) and 12 males parent (6 Inter-racial and 6 Guinea) were crossed to produce 18 hybrids with both parents classified as Inter-racial, 12 hybrids of predominantly Guinea background, and 30 hybrids with one parent classified as inter-racial and the other as Guinea. The 60 hybrids were tested together with 6 check varieties over two years in four environments per year. Male parents of Inter-racial and Guinea-bred origins had the most frequent positive general combing ability (GCA) values over environments whereas Landrace-Guinea and the more Caudatum-phenotype breeding lines often had negative GCA values. Female Parents showed important genotype x environment interaction. The ratio of Male-GCA/SCA ranged from 1.5 to 5.0. The mean performance of Inter-racial-, Guinea-, and Inter-racial x Guinea hybrid groups differed significantly for grain yield in all environments. The Inter-racial group produced the highest mean hybrid yields under high fertility conditions at Samanko and Sotuba. Under the Low-P conditions in contrast, the Inter-racial x Guinea group of hybrids had the highest mean yields, followed closely by the Guinea group. Each group of hybrids had some hybrids significantly superior to the mean of all hybrids, although the frequency differed substantially among groups over the test environment.

54

Мolecular confirmation possibility of fixing heterosis effect in rice hybrids by Strunnikov’s method

J.K.Goncharova, V.V.Gronin, D.B. Krutenko, E.V.Litvinova

All Russian Rice Research Institute Krasnodar Russia, 350059 SRI ARRRI, Krasnodar, p/o Belozеorny


Strunnikov offered in 1999 a theory which not only allows explaining the nature of heterosis, but also answering the main questions of heterosis breeding; how this gene complexes are created, and how to maintain heterosis in the following generations. Efficacy of the heterosis maintenance method in the successive generations was tested on the hybrids of silkworm and drosophila. We received the data confirming offered theory in plant on molecular and organism level. This hypothesis allows integrating all proposed earlier theories. Problem of heterosis maintenance can be solved by means of getting hybrids without lethals and half-lethals. For cleaning hybrid genotype from genes decreasing viability, we used anther culture. Back-crossing hybrid to the samples without bad genes in homozygous and heterozygous conditions leads to progeny, in genotype of which there are no lethals in homozygous condition. High viability of back-crossed generations is of a great interest, because these generations are coming from self-propagation, which usually reduces viability. We obtained backcross generation of 13 hybrid combinations of rice. The absence segregation in back cross generation confirm opportunity of fixing heterosis effect by the offered method. Analysis loci by SSR and ISSR molecular marker, connected with adaptivity traits shown selective allele elimination from double haploid line population. That confirms the possibility using anther culture for removing lethal and half –lethal loci from hybrid genotype.

55

Population-hybrid superiority and combining ability patterns among West African pearl millet landraces

BIG Haussmann*, HK Parzies, I Angarawai, J Gondah, R Zangré, MD Sanogo and O. Sy

*ICRISAT, BP 12404, Niamey, Niger;


Pearl millet [Pennisetum glaucum (L.) R. Br.] is a highly out-crossing staple cereal crop that is most important for food security of small-scale farmers in West and Central Africa (WCA). In order to enhance performance of open-pollinated cultivars, and with a vision for future hybrid breeding, efforts are underway to establish heterotic groups among pearl millet landraces in their WCA centre of origin an diversity. This is done through studying agro-morphological differentiation, genetic diversity at the DNA level (using molecular markers), and through evaluation of population hybrids representing putative inter- and intrapool crosses. Here we report results from a multi-location evaluation of 10×10 factorial crosses in 2006 at seven sites in Nigeria, Niger, Burkina Faso, Mali and Senegal. Parental populations crossed for this trial included two female (one early, one late) and two male parents (one early, one late) each from Senegal, Mali, Burkina Faso, Niger and Nigeria. These were crossed in a factorial mating design, thereby producing putative inter- and intra-pool crosses. The population hybrids out-yielded their parental populations on average by 15%, with the mean hybrid superiorities at the individual testing sites ranging from -1% at Samanko (Mali) to +32% at Bambey (Senegal). Average hybrid superiority for panicle yield was not related to the environmental mean. Hybrid superiority for individual entries averaged across locations ranged from -21 to +70%, pointing to the benefits of enhanced heterozygosity in certain population crosses. Nine out of the ten best hybrids were derived from crosses of parents originating in different countries. Among the five highest yielding hybrids, four were derived from crossing Senegalese varieties with parents from Niger or Nigeria (wide crosses). Combined across locations, general combining ability (GCA) effects of female (F) and male (M) parental populations, specific combining ability (SCA) effects (F×M) as well as the two-fold interactions with test locations (F×E, M×E) were highly significant, while the interaction of SCA effects and locations (F×M×E) was only weakly significant. Therefore, both additive and dominant gene effects have contributed to the performance of the hybrids and may be exploited in the WCA breeding programs. The study illustrates complex combining ability patterns among WCA pearl millet landraces and underlines the benefits of germplasm exchange and wide crosses for enhancing performance of pearl millet open-pollinated varieties, in order to enhance food security in WCA.

56

Diversity arrays technology for whole-genome profiling of crops

Eric Huttner, Vanessa Caig, Jason Carling, Michael Duncan, Margaret Evers, Kasia Heller-Uszynska, Colleen Hopper, Damian Jaccoud, Gosia Kilian, Karl

Miller, Grzegorz Uszynski, Peter Wenzl, Ling Xia and Andrzej Kilian

Diversity Arrays Technology, Canberra, Australia PO Box 7141, Yarralumla, ACT 2600, Australia


Diversity Arrays Technology (DArT) is a novel genetic marker system providing high-density and low-cost whole-genome profiles. DArT markers have been developed for over forty crop and wild plant species. DArT profiles represent a powerful tool for the deep characterisation of breeding material. They enable the effective management and use of heterotic groups in plant breeding. As an example, current genetic progress in many grain legumes is limited by the extremely low genetic diversity of the currently cultivated germplasm. The development of new productive varieties will involve many strategies including the introduction of germplasm from wild relatives and landraces. These breeding endeavours will greatly benefit from the availability of affordable whole-genome molecular profiles based on DArT markers in pigeonpea, chickpea, lupin, groundnuts and others crops.

57

Broadening the avenue of intersubgenomic heterosis for oilseed Brassica

Jun Zou and Jinling Meng

National Key Laboratory of Crop Genetic Improvement, Huazhong Agricultural University, Wuhan, China


Accumulated evidence shows that the long history of intensive artificial selection in different Brassica species has resulted in profound genomic changes in Brassica crops. We distinguish three sets of basic genomes and subgenomes in a modified U’s triangle, building on the cytogenetic relationships of Brassica species and their genomes identified last century. For example, the A subgenome from B. rapa is identified as Ar, whereas the A subgenome from B. napus is identified as An. To utilize the genetic difference between subgenomes, we partially replaced the An and Cn genome of B. napus by the Ar genome of B. rapa and Cc genome of B. carinata, resulting in the first generation of new type B. napus. Heterosis between “normal” B. napus and the new type B. napus can therefore be defined as “intersubgenomic heterosis”. Significant intersubgenomic heterosis was observed in hybrids between the new type B. napus and normal B. napus over the past ten years of field trials. By selecting the plants with higher proportion of exotic-subgenomic components from segregated populations of first generation of new type B. napus, a second generation of new type B. napus was developed with stronger intersubgenomic heterosis. To increase the content of the exotic-subgenomic components further and to enable a sustainable breeding for novel lines of new type B. napus, we proposed an idea of developing a gene pool for new type B. napus with plenty of genetic variation at both Ar/Cc genome. New evidence has shown that, after interspecific crossing, great genomic alterations would happen in new type B. napus far beyond the introgression of exotic genomic components, and the genepool would enable us to carry out a recurrent selection for the alleles with positive contribution to the intersubgenomic heterosis. New approaches in broadening the avenue of intersubgenomic heterosis for oilseed Brassica are discussed.

58

Selection of maize plus-hybrids: relevance of genetic distance for predicting the Xenia effect on grain yield

Magali Munsch1,2, C. Weider1, K.-H. Camp2, P. Stamp1

¹ETH-Zurich, Institute of Plant Sciences, group of Agronomy and Plant Breeding, Zurich, Switzerland, ²Delley Seeds and Plants, Delley, Switzerland


Maize (Zea mays L.) Plus-Hybrids are cytoplasmic male-sterile (CMS) hybrids that are grown in a mixture with unrelated male-fertile hybrids to ensure the pollination of the whole stand. In regular maize stands, huge amounts of pollen are produced and the harvested grains consist of an F2 population, i.e. the first inbreeding generation, resulting from the sib-pollination of single-cross hybrids. In contrast, in maize Plus-Hybrids, the yield of CMS plants, accounting for 80 percent of the stand, can be increased both by the effect of male sterility (CMS effect) and the effect of allo-pollination (xenia effect). The latter could stem from a higher level of heterozygosity in allo- than in sib-pollinated kernels, thus, increasing the final kernel weight. The aim of this stud was to investigate the relationship between the genetic distance (GD) between the two Plus-Hybrid components and the extent of the xenia effect on yield and thousand kernel weight (TKW). Eighteen modern European hybrids, 12 of which were available in the male-fertile and the male-sterile forms, were combined into 64 Plus-Hybrids, which were evaluated in small-plot field trials from 2005 to 2007. Both positive and negative effects of xenia on yield and TKW were found. Several hybrid combinations led to gains in yield due to xenia above 6%, on average across all environments, and from up to 20% in specific environments. Based on the genotyping by 20 simple sequence repeats (SSR) markers, different genetic groups were identified. The phylogenic tree corroborated the pedigree information provided by breeders. However, the GD between the CMS and the pollinator hybrid was not correlated with the xenia effect on yield or TKW. Nevertheless, there were indications that only a limited number of genes might be involved in the expression of xenia, especially for the paternal effect. This should be followed up to develop specific molecular markers, which could be used in marker-assisted selection in order to facilitate the identification of good Plus-Hybrid candidates within the existing genetic pool.

59

Modern maize hybrids can improve grain yield as plus-hybrids by the combined effects of cytoplasmic male sterility and

allo-pollination

Magali Munsch1,2, C. Weider1, K.-H. Camp2, A. Hüsken3, X. Foueillassar4, N. Christov5, P. Stamp1

1- ETH-Zurich, Institute of Plant Sciences, group of Agronomy and Plant Breeding, Zurich,Switzerland; 2- Delley Seeds and Plants, Delley, Switzerland; 3-JKI, Institute for

Biosafety of Genetically Modified Plants, Braunschweig, Germany; 4- Arvalis, Institut du Végétal, Montardon, France ; 5-AgroBioInstitute, Sofia, Bulgaria


In a maize (Zea mays L.) Plus-Hybrid, a cytoplasmic male-sterile (CMS) hybrid is combined with an unrelated male-fertile hybrid, the latter acting as the pollen donor for the whole field. In addition to the heterosis effect on both hybrids, two other phenomena can improve the grain yield of the CMS hybrid: the effect of male sterility (CMS effect), and the effect of allo-pollination (xenia). This study aimed at evaluating the potential gain in the yield of modern European Plus-Hybrids. Combinations of Hybrids of different Breeding Companies were evaluated in small-plot trials in three seasons. The pollinating ability of the male-fertile as well as the combining ability of the male-sterile form of 12 hybrids was evaluated with two testers at three Swiss locations in 2005. Five responsive CMS hybrids and three additional good pollinators were identified as being good candidates for Plus-Hybrids. In the ‘European maize ring experiment’, carried out in 2006 and 2007, these candidate hybrids were evaluated in a factorial manner at 12 locations in Switzerland, Germany, France and Bulgaria. In comparison to the yield of the male-fertile form of the sib-pollinated seed-plants, modern Plus-Hybrids increased the maize grain yield by 10% across all environments. In given environments, the best performing Plus-Hybrids led to yield increases above 20%, often due to variation in the two yield components, number of kernels (KN) as well as the thousand kernel weight (TKW). Variations in KN were usually related to the CMS effect, while variations in TKW were due mainly to xenia. The CMS effect seemed to be very dependent on the environment, and CMS hybrids usually performed better than their male-fertile counterpart under sub-optimal growing conditions. Thus, the CMS effects seemed to be less relevant than xenia effects for a significant impact on grain yield; however, they still remain of interest since CMS plants could act as a buffer, favoring more stable yields in uncertain climate change. On the other hand, the potential gain in yield due to xenia was higher and more consistent across environments, although its extent varied according to environmental conditions. For this reason, it is suggested that breeding companies concentrate on the xenia effect, i.e. on identifying hybrid pairs that combine well in the existing genetic pool and on developing further effective CMS and pollinator hybrids.

60

Heterosis and/or epigenetic effects increase the kernel weight of allo-pollinated cytoplasmic male-sterile maize

Christophe Weider1, M. Munsch1,2, K.-H. Camp2, P. Stamp1

1- ETH-Zurich, Institute of Plant Sciences, group of Agronomy and Plant Breeding, Zurich, Switzerland; 2- Delley Seeds and Plants, Delley, Switzerland


Xenia and cytoplasmic male sterility (CMS) together can often increase grain yield of maize (Zea mays L.); this is referred to as the Plus-Hybrid effect. While the CMS effect influences the kernel number, xenia influences the single kernel weight (SKW). An improved knowledge on the genetic control of a SKW increase by xenia could lead to improved breeding strategies. A field study was carried out in three different environments in Switzerland in 2006 to compare the grain yields and grain quality from the combination of three modern European CMS hybrids with six current pollinators. Significant Plus-Hybrid effects were observed for almost all combinations. The effect of allo-pollination on the grain yield and kernel weight was usually positive and significant. Both main compartments of the kernel, the endosperm and the embryo were affected by the xenia effect and their weights were usually increased. Nevertheless, the relative influence on the endosperm / embryo weight ratio in comparison with kernels of the isogenically pollinated CMS hybrid was negative in most cases of allo-pollination; this reflects a stronger increase in weight of the embryo relative to the endosperm, relating rather to an early heterosis effect than to epigenetic pathway. Still, the results depended on the Plus-Hybrid combination (CMS + pollinator) and heterosis as well as genomic imprinting might act together as a booster of kernel weight after allo-pollination.

61

Development of hybrid cotton (Gossypium spp.) using honey bees as pollinators and roundup ready as an indicator trait

Jimmy X. Zumba1 and G.O. Myers1

1LSU AgCenter, School of Plant, Environmental and Soil Sciences104 M.B. Sturgis Hall, Baton Rouge, LA


The use of hybrid cotton (Gossypium hirsutum) in the US has been limited due to seed cost production. The objective of this study was to investigate a novel method for the production of F2 cotton hybrids using honey bees as pollinators and Roundup Ready® gene as selection trait. This research was conducted from 2005-2007 in Louisiana. Crosses between non-transgenic and transgenic varieties were made in 2005 to obtain F1 cottonseeds using honey bees. In 2006, F2 cottonseed was obtained. In 2007, F1, F2, and parents were field tested using a randomized complete block design with 3 replications in two locations. Data analysis was conducted using the SAS PROC MIXED procedure. Results indicate that all crosses exhibited heterosis in the F1 hybrid populations relative to the best parent. The crosses LA1110023/PHY410R and ARKRM24-12-04/PHY410R exhibited a higher degree of heterosis for yield averaging 33.1% and 20.6%, respectively, across locations. Yield heterosis in the F2 population was of 20.9% and 19.5%, respectively, and statistically different from the best parent. The ARK9506-40-05/PHY410R cross had yield heterosis averaging 15.6% in the F1 population and 13.5% in the F2 population; however, these were not significantly different from the best parent. Fiber quality descriptors from the crosses did not have a significant heterosis in the F2 population relative to the best parent. In summary, the use of herbicide resistant varieties as males and Roundup Ready® gene as selection trait, conventional varieties as females and honey bees as pollinators, was shown to be a viable method for developing F2 hybrid varieties. Further variety testing will be required to determine the best combination of parents.

62

The use of heterotic patterns: thirty years of selection in maize

C. Romay, B. Ordás, P. Revilla, A. Ordás

Misión Biológica de Galicia, CSIC


Modern maize (Zea mays L.) breeding is based on the exploitation of heterotic groups. The most common heterotic pattern used in temperate areas is “Reid × Lancaster”. The Reid group is mainly Iowa Stiff Stalk Synthetic (BSSS). Although Lancaster germplasm is used extensively, it seems that it would be more appropriate to set the pattern as “Reid × other dent germplasm” because other materials different from Lancaster are also used. In Western Europe the pattern “Corn Belt Dent (mainly Reid) × European Flint” is also used. A work to establish new heterotic patterns started in 1974 at the Misión Biológica de Galicia (CSIC) in Spain. Five Spanish populations (two from humid Spain and three from dry Spain) and four Northern Corn Belt populations were crossed in a diallel series in 1974 and 1975. The 9 populations and 36 crosses were evaluated at two locations in Spain in 3 years (1976–1978). The analysis of the diallel revealed the existence of three heterotic groups: US Corn Belt, northern (humid) Spain, and southern (dry) Spain. Two populations were developed: EPS6 (northern Spain) and EPS7 (southern Spain) that formed a new heterotic pattern. Three cycles of S1 recurrent selection were carried out in EPS6 and EPS7. The evaluation of these three cycles of selection in two locations in two years (1994 and 1995) showed that grain yield increased significantly with selection in both populations. Grain yield increased linearly at a rate of 0.5 t/ha per cycle (13.0%/cycle) in EPS6, while in EPS7 the gain was lower (9.8%/cycle) and followed a quadratic trend. We began a reciprocal full-sib (RFS) selection program in 1992 with the populations being renamed: EPS6(S)C3 as EPS13 and EPS7(S)C3 as EPS14. After completing three cycles of selection, the program was evaluated in three locations in Spain in 2006 and 2007. The yield of the cross between the two populations has improved along the selection program. The yield of EPS13 has also improved, but not that of EPS14. The initial populations (EPS6, EPS7, EPS13, and EPS14) and the populations resulting from each cycle of the RFS selection were genotyped with 76 SSRs. We detected a small loss of variability and a subsequent increase of inbreeding as expected. The effective population size of EPS14 is slightly less than expected.

63

An eight-way breeding scheme for generating a large multi-parental RI population in maize

Giorgio Pea*1, Htay-Htay Aung1, Luca Gianfranceschi2, Mario Enrico Pè1

1Scuola Superiore Sant’Anna, Pisa, Italy 2DSBB, Università di Milano, Milano, Italy


Studies on heterosis for maize yield demonstrated the existence of several combined effects for many yield and yield components. The most recent observations suggest that manifestation of heterosis might lay in a complex combination of diverse molecular factors. In such a frame, advanced cross designs might boost the power of detection of the genetic bases of complex traits, including heterosis, thus providing unprecedented and powerful resources of information. Here we describe the program currently ongoing for the development of an innovative advanced RI population, aimed at genetic and molecular dissection of heterosis and other complex traits in maize. Eight maize inbred lines were selected to include a wide genetic variability for the expression of complex phenotypes and crossed according to a half-diallel design. This genetic material was used as the starting point for producing an extended (>2000) maize 8-ways Recombinant Inbred lines population (8W-RI) to be used for high-resolution QTL mapping. According to what estimated by The Complex Trait Consortium, which first proposed this model for mouse, such a material should allow mapping QTL with effect size >5% of the total variance to an interval of 0.5 cM using fewer than 1000 lines. Twenty-eight 2W maize hybrids from the 8x8 half-diallel were crossed so that only crosses between entries with no parents in common (e.g. cross AB x CD) were allowed (but not AB x AD or AB x BF, etc.). Such 4W hybrids (210) were bulked in 70 pools, each composed by all the three 4W hybrids bearing the same alleles in all possible parent-of-origin cis combinations (e.g. “ABCD” pool included ABxCD, ACxBD and ADxBC 4W hybrids). 8W hybrids were then produced by crossing complementary 4W hybrids pools (e.g. ABCD x EFGH, CDFG x ABEH, etc.). The production 8W-RI by single-seed descent is currently at the third selfing generation (8W-RI F3). Performing two generations of selfing per year, we expect to obtain the 8W-RI F6 mapping population by the end of 2010, whereas the molecular characterization of parental lines is scheduled to start in the fall of 2009. All 2W hybrids, 4W hybrids and 8W highly-recombinant hybrids plus the parental inbreds will also be available for phenotypic evaluation of heterotic traits. Trials and experimental fields will allow estimating both general and specific combining ability on a broad genetic base.

64

Pearl Millet hybrid parents development: perspectives from an international research center

KN Rai, SK Gupta, V Vengadessan, T Nepolean and CT Hash

International Crops Research Institute for the Semi-Arid Tropics, Patancheru, Andhra Pradesh, India


Pearl millet [Pennisetum glaucum (L.) R. Br.], primarily grown for grain production on more than 26 million ha in some of the harshest arid and semi-arid tropical environments of Asia and Africa, is a highly nutritious cereal crop with wide agro-ecological adaptation. The International Crops Research Institute for the Semi-Arid Tropics (ICRISAT) is a world leader in genetic improvement of this crop. India, cultivating pearl millet on about 10 million ha (more than 90% of pearl millet area in Asia), is the largest producer of this crop globally. The National Agricultural Research System (NARS) and a large number of private seed companies in India are almost all focused on hybrid development. In alignment with this research priority, ICRISAT’s research emphasis for the Asia region has been to undertake trait-based breeding to develop and disseminate a diverse range of breeding lines with high grain yield potential, resistance to downy mildew (DM) caused by [Sclerospora graminicola (Sacc.) Schroet.], and adaptation to various agro-ecoregions. This strategy has strengthened the genetic base of hybrid programs in India as reflected in more than 80 hybrids currently in the market (compared to no more than six hybrids about 30 years ago) of which at least 60 hybrids are based on parental lines developed at ICRISAT, or on proprietary parental lines developed from ICRISAT-bred germplasm. ICRISAT also conducts strategic research with a view to enhance its breeding efficiency. Thus, ICRISAT assembled and discovered alternative cytoplasmic-nuclear male-sterility (CMS) systems, evaluated these for commercial viability and breeding efficiency, diversified the CMS base of male-sterile lines (A-lines), and developed a breeding scheme and demonstrated its effectiveness in producing restorer lines of these alternative CMS systems. ICRISAT pioneered molecular marker-assisted breeding for drought tolerance and DM resistance to add value to parental lines of some of the commercial hybrids. This resulted in the development, release and adoption of a more DM resistant version of an early-maturing hybrid (HHB 67) that matures in 65 days and is cultivated on more than 0.4 million ha. A molecular marker-based approach is now being tested to characterize heterotic groups among hybrid parental lines. Though initial results, based on 38 SSR markers, do not show any correlation between marker-based genetic diversity and heterosis for grain yield, this research will be continued using a larger set of markers well distributed across the pearl millet genome.

65

Using global diversity of guinea-race sorghums for hybrid development in West and Central Africa

H.F.W. Rattunde¹, Sokona Dagnoko², M. Smith³, Eva Weltzien¹

¹International Crops Research Institute for the Semi-Arid Tropics ²World Vegetable Center (AVRDC), ³Cornell University, Bamako Mali


Hybrids based on Guinea-race sorghum parents could provide yield increases while retaining the required adaptive and grain quality characteristics that make this race the predominant race in West Africa from Senegal to Burkina Faso. The extent and pattern of heterosis in Guinea-race sorghum hybrids were studied to determine possibilities for successful hybrid breeding and heterotic pool identification in this race. Hybrids (n=77) were produced by crossing Guinea-race male parents originating in the Sudanian- (13) and Guinean- (7) zones of West and Central Africa, Eastern and Southern Africa (6) and Asia (2) to three A-lines developed from Malian (2) and Sudanese (1) landrace varieties. Yields of hybrids, parents and check varieties were assessed in eight environments in Mali and Niger over two years. Better parent heterosis (Hbp) ranged from 9.8 to 44.6% for the eight environments averaged over all hybrids tested, and from 12.8 to 63.8% for hybrids on the best female parent, FambeA. Mean Hbp of FambeA hybrids over environments was 30.3%. Absolute yield superiorities over better parents ranged from 169 to 558 kg-ha for all hybrids across the eight environments, and from 228 to 807 kg-ha for FambeA hybrids. The hybrid yield superiorities indicate that Guinea-race hybrids can provide agronomically useful advantages over traditional landrace varieties. Although one or more male parents from each geographic region produced superior hybrids with high heterosis, male-parents from the more humid Guinea-zone of West and Central Africa and from Southern Africa showed greatest promise when crossed with female parents from the drier Sudanian zone of West and Central Africa

66

Role of small RNAs in cultivated and wild tomato phenotypes and their hybrids

Padubidri V. Shivaprasad and David Baulcombe

Department of Plant Sciences, University of Cambridge, CB2 3EA, Cambridge, United Kingdom


Tomato is a model system for fleshy fruits and its accessions are highly variable in phenotypes but with less genetic diversity. We aim to establish a correlation between tomato phenotypes of useful agronomic traits and the presence or effects of small (s)RNA. Small RNAs mediate repression of genes through RNA silencing, and are involved in various aspects of plant development including developmental patterning, genome integrity and stress respons. With next generation solexa sequencing we have profiled Argonaute (AGO) 1 associated sRNAs as well as total small RNAs, from four selected cultivars: MicroTom (all purpose), M82 (processive variety), Solanum pimpinellifolium and S. pennellii (two wild varieties with useful traits) to identify variable sRNAs. SiLoCo (UEA small RNA tool kit) has identified many variable sRNA loci among these cultivars. Northern blotting and solexa cloning identified quantitative differences between several micro (mi)RNAs and transacting siRNAs among the tomato lines. One of the novel miRNA which could target pathogenesis-related protein 5 (PR-5), was downregulated upon various stress treatments and appeared to be a housekeeping miRNA. We have also made profiles of small RNAs from M82 X S. pennellii hybrid and few Introgressed Lines and identified several small RNA loci which influence gene expression in these lines. Together, these differences among small RNAs indicate a strong influence of small RNA accumulation on tomato transcriptome which might explain few aspects of hybrid vigour.

67

International Conference on Heterosis in Plants · The vigor and stature of maize plants has been investigated in a comparative study at multiple ploidy levels. The monoploid is less

Documents