| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
a Dep. of Agronomy and Plant Genetics, University of Minnesota, St. Paul, MN 55108
b USDA-ARS Biosciences Research Lab, Fargo, ND 58105
* Corresponding author (ander319{at}umn.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONConclusions and Future ProspectsREFERENCES |
|---|
Abbreviations: EST, expressed sequence tag • FHB, Fusarium head blight • MAS, marker-assisted selection • NIL, near-isogenic line • PIC, polymorphism information content • QTL, quantitative trait locus • SSR, simple sequence repeat
Received for publication July 9, 2007.
a Dep. of Agronomy and Plant Genetics, University of Minnesota, St. Paul, MN 55108
b USDA-ARS Biosciences Research Lab, Fargo, ND 58105
* Corresponding author (ander319{at}umn.edu).
The difficulties of breeding for Fusarium head blight (caused by Fusarium graminearum Schwabe [teleomorph: Gibberella zeae]) resistance, a quantitatively inherited fungal disease, caused us to initiate a marker-assisted selection (MAS) approach to accelerate our gains from selection. Although MAS for simply inherited traits has become commonplace in many plant breeding programs, there are few examples of its application with quantitatively inherited traits. Several barriers to MAS for a quantitative trait locus (QTL) must be addressed before it can be integrated into a breeding program, including (i) its efficiency or gain compared to phenotypic selection; (ii) the usefulness of markers in breeding-relevant populations; and (iii) the cost, throughput, and expertise required. We identified a major QTL, Fhb1, for Fusarium head blight resistance in wheat (Triticum aestivum L.) and validated its effect in an additional mapping population and near-isogenic lines developed from segregating lines in our breeding program. The effect of this QTL was large and consistent enough to justify complementing our extensive phenotypic screening efforts for this disease with MAS for this major QTL. Fhb1 is located in a highly polymorphic region, and we developed highly diagnostic markers while fine mapping this QTL. The establishment of the USDA-ARS Regional Small Grains Genotyping Centers has dramatically increased our capabilities to apply MAS by providing access to high-throughput DNA extraction and genotyping equipment. Because a limited number of individuals can be subjected to MAS, we use a process of retrospective breeding to identify those populations that are most likely to produce cultivar candidates. More efficient DNA extraction technologies and marker platforms will allow us to fully implement MAS in breeding programs.
Abbreviations: EST, expressed sequence tag • FHB, Fusarium head blight • MAS, marker-assisted selection • NIL, near-isogenic line • PIC, polymorphism information content • QTL, quantitative trait locus • SSR, simple sequence repeat
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONConclusions and Future ProspectsREFERENCES |
|---|
Background on Fusarium Head Blight Resistance
Fusarium head blight is an important disease in the hard red spring wheat producing areas of the United States and Canada (Wilcoxson et al., 1992) as well as other regions of the world. The fungus infects during anthesis or grain filling and causes severe yield reduction and decreases grain quality (Bai and Shaner, 1994). The ability of the pathogen to cause significant damage when appropriate climatic conditions are present makes rapid incorporation of durable resistance into adapted genotypes a priority for the entire wheat industry.
Resistance screening can be effectively accomplished under greenhouse or field conditions (Dill-Macky, 2003), but field screening is both time and resource intensive, and results are often confounded by environmental factors, and needs to be repeated over environments (Campbell and Lipps, 1998; Groth et al., 1999; Fuentes-Granados et al., 2005). We determined an optimum resource allocation to screen breeding materials for FHB resistance that included a single evaluation environment to discard highly susceptible homozygous lines, but three to four discriminating environments are necessary to accurately assess a line's response to this disease. Because of the difficulties in breeding wheat for FHB resistance using phenotypic methods, the identification of DNA markers associated with resistance has been a high priority for wheat breeders and geneticists.
Available resistance to FHB in wheat is quantitatively inherited with a continuous distribution among progeny (Snijders and van Eeuwijk, 1991; Bai and Shaner, 1994; Snijders, 1994; Waldron et al., 1999; Anderson et al., 2001, Kolb et al., 2001; Buerstmayr et al., 2002). The Chinese cultivar Sumai 3 and its derivatives are the most popular sources of resistance genes. Cultivar x isolate interactions have been tested, but none found for the predominant Fusarium species that cause head blight (Snijders and van Eeuwijk, 1991). Therefore, genetic vulnerability with the known resistance genes should not be a concern, but with the large genetic variability known to exist in Fusarium spp. (Bowden and Leslie, 1999), the introduction of at least a few different genes would be a wise approach.
Fhb1
We first reported the location of a major QTL, Fhb1 (syn. Qfhs.ndsu-3BS), for FHB resistance (Waldron et al., 1999) and verified it in a second mapping population (Anderson et al., 2001). The best simple sequence repeat (SSR) markers explained 24.8 and 41.6% of the phenotypic variation in FHB resistance in the two mapping populations (Anderson et al., 2001). The large effect of this QTL was confirmed in other populations by other researchers (Bai et al.,1999; Buerstmayr et al., 2002; Zhou et al., 2002, 2003; Yang et al., 2003). We also have developed and tested more than 19 pairs of near-isogenic lines (NILs) for this QTL in diverse genetic backgrounds and have found a consistent effect of about 27% reduction in FHB-infected grains (Pumphrey et al., 2007). This is the largest, most consistent QTL effect discovered to date for FHB resistance in wheat.
Criteria for Marker-Assisted Selection
We have considered three broad, practical criteria that must be satisfied for MAS to be effectively implemented in a breeding program: (i) efficiency or gain compared to phenotypic selection; (ii) usefulness of markers in breeding-relevant populations; and (iii) the cost, throughput, and expertise required.
Efficiency or Gain Compared to Phenotypic Selection
Theoretical considerations aside, it should be noted that any MAS activity is competing against or complementing, and in most cases, not replacing a well-established evaluation system based on phenotype. The proportion of the trait variance explained by individual QTLs is probably the most significant barrier to efficient implementation of MAS. This is not an issue with qualitative traits expressed as distinct phenotypes (e.g., resistant vs. susceptible), except to the extent that the genotype does not equate to the marker phenotype due to recombination. Theoretically, MAS for QTLs is most advantageous when heritability is low and the proportion of variance explained by markers is high. However, these situations are not common. A preponderance of QTL studies reveal moderate effects (R2 of 10–20%) at one or a few loci, and several other loci explaining less than 10% of the variance. Population sizes used to establish QTL linkages (100–200) are not large enough to provide precise estimates of QTL effects (Beavis, 1998; Bernardo, 2004). Instead, marker effects in such small populations are usually overestimated and there is evidence that MAS using smaller effect QTLs can actually diminish gains from selection (Bernardo, 2001; Bernardo and Charcosset, 2006). Inconsistent QTL effects across breeding populations are the main reason that there are few examples of their use in MAS programs in plants.
Parents for developing mapping populations are chosen, in part, due to their disparate values for the trait or traits to be mapped. If the QTL is subject to epistatic effects, its effect could be quite different when present in other (adapted) genetic backgrounds. Two common approaches to validate QTL position and effect are to (i) measure QTL effects in additional mapping or validation populations; or (ii) backcross the QTL into one or few genotypes to create sets of NILs to estimate QTL effect. Unfortunately, both of these approaches are very time consuming and their results still may not adequately represent QTL effects in breeders' populations. We have chosen to investigate QTL effects over multiple genetic backgrounds through the development of QTL-NILs (Pumphrey et al., 2007). The QTL-NILs are developed by identification of lines heterozygous for the QTL (via markers) at the F4 stage (headrow selection in our breeding program [Table 1 ]) followed by selfing to extract homozygous lines that differ for marker alleles flanking the QTL. The resulting homozygous QTL-NIL lines are phenotypically evaluated in multiple environments to estimate the QTL effect on the trait of interest and detect linkage drag and pleiotropic effects. Testing the QTL effect in 10 to 20 genetic backgrounds should give the breeder an accurate estimate of the magnitude and consistency of its effect in elite breeding germplasm. Developing NILs over a representative cross-section of the breeding germplasm in this manner fulfills two important functions by assessing: (i) the informativeness of markers per se for the QTL; and (ii) the effect of the QTL in different genetic backgrounds, thus taking into account the presence of other, unselected genes. The resulting QTL-NILs are directly useable in the breeding program, as opposed to other QTL validation methods that may, at best, produce new parental material for the breeding program.
|
Usefulness of Markers in Breeding-Relevant Populations
The marker systems of the late 1980s and early 1990s were plagued by low polymorphism and detection of relatively few alleles (Anderson et al., 1993). The polymorphism information content (PIC) of SSR markers, the mostly widely used marker type in wheat today, is satisfactory for most regions (Harker et al., 2001). The five SSR markers near Fhb1 have PIC values of 0.70 to 0.84 in a diverse set of 74 genotypes (Liu and Anderson, 2003a). Moreover, the allele associated with Fhb1 is unique enough that these markers could be used to predict the presence of this QTL in germplasm without prior information on pedigree or FHB resistance. The synteny between rice (Oryza sativa L.) and wheat and the huge amount of rice genomics data and wheat expressed sequence tags (ESTs) were exploited to develop DNA markers closely linked to Fhb1 (Liu and Anderson, 2003b). We fine-mapped the Fhb1 gene, using a population developed by crossing two NILs in a susceptible background, into a 1.2-cM interval (Liu and Anderson, 2003b; Liu et al., 2006). The sequencing of BAC clones in this region revealed a number of polymorphisms that were exploited to design highly diagnostic markers that are less than 0.1 cM from this QTL and >99% effective in revealing polymorphism in the hundreds of cultivars and breeding lines that we have evaluated to date (data not shown).
Facilities and Expertise
Efficient implementation of MAS demands the use of high-throughput equipment and trained personnel. Although MAS is becoming a new capability in many wheat breeding programs, its implementation is limited by the cost to support trained personnel and purchase equipment and reagents. Backcrossing using markers is a cost effective way of utilizing markers. A recently completed USDA-funded project, MASwheat, utilized backcrossing with DNA markers to introgress 43 genes into 75 genetic backgrounds (Dubcovsky, 2004; http://maswheat.ucdavis.edu/ifafs.htm). This germplasm has been valuable to breeding programs and regions and also laid the groundwork for a current USDA-CSREES–funded Coordinated Agricultural Project (CAP) (http://maswheat.ucdavis.edu/Index.htm). The WheatCAP project is utilizing the four new USDA-ARS Regional Small Grains Genotyping Centers for processing MAS samples. The genotyping centers were proposed by Van Sanford et al. (2001) and have greatly enhanced our ability to use markers in selection, effectively removing the barriers to efficient use of this technology by individual breeding programs. Analogous to the USDA-ARS regional wheat quality laboratories, the regional genotyping centers provide support to small grain breeding programs in the United States through use of automated DNA extraction and high-throughput marker screening procedures. They provide a bioinformatic interface between molecular genetic data and breeding programs. Concentrating the genotyping activities in regional labs will ensure that the expertise and equipment will keep pace with improvements in technology. This allows breeders to focus their attention on choosing populations and markers, and utilization of the resulting information.
Although SSRs are the most widely used marker in wheat at this time, single nucleotide polymorphisms are being developed (http://wheat.pw.usda.gov/SNP/new/index.shtml) with the expectation that they will replace SSRs because of their greater abundance compared to other marker types and potential for use in lower cost, higher-throughput systems (Hernandez, 2005; Huang and Röder, 2005).
Marker-Assisted Selection for Fhb1
We have approached MAS for Fhb1 with the following principles in mind:
121 if only homozygous undesirable classes for any of the four markers were discarded. Although there is theoretical justification for this population-enriching approach to MAS (Bonnett et al., 2005), the majority of these materials will be discarded in their first generation of field-based selection because they lack the desired phenotype for relatively simply inherited traits such as height, maturity, and resistance to leaf rust and stem rust. The concept of retrospective breeding is that at any given stage in a breeding program (e.g., F2, headrows, preliminary yield trial), we judge the genetic potential of a population or collection of lines with common parentage. Those populations that are most desirable or produce the most promising lines are resampled from remnant seed of an appropriate generation, or reconstituted entirely, and then subjected to MAS. Therefore, we have some assurance that the population being subjected to the time and expense of MAS is likely to produce superior inbred lines at the end of the selection and evaluation process. Prior knowledge of performance gives the breeder confidence to subject a much larger population (and screen additional genes [markers], if appropriate) to MAS. This retrospective approach avoids the time and expense of conducting MAS in populations with a low probability of producing improved lines as a result of poor combining ability or absence of genes for other required traits. While resampling an "old" population may not appeal to all breeders, we believe that marker screening in this more directed manner, using much larger population sizes, should substantially increase the possibility of recovering superior progeny and easily negate the loss of time associated with resampling or reconstructing a cross or population.
Marker-Assisted Selection for Fhb1 in Practice
We have used MAS at several different generations in our breeding program. Table 1 shows the flow of our breeding programs and numbers of populations and progeny at each stage. Initially, we began applying MAS for Fhb1 only at the F4 headrow stage in year 2000, using leaf tissue collected from the field and screening with flanking SSR markers (Liu and Anderson, 2003a) in-house using a Li-Cor DNA sequencer (LI-COR Biosciences, Lincoln, NE). The reasoning for this approach was that the F4 headrow stage represents our first stage of major selection and marker screening resources could be concentrated on those lines that had been previously selected for other key agronomic and disease resistance traits. In addition, because only 
1,200 F4s were selected based on phenotype each year and only about half of them were from populations expected to segregate for Fhb1, this created a manageable ceiling for the number of DNA extractions that needed to be performed. In addition, during 2001 to 2004 Fhb1 was backcrossed into six recently released cultivars and advanced, elite lines as part of the MASwheat project. While none of the finished lines were direct cultivar releases, they have been extensively used as parents in crosses.
The establishment of the USDA-ARS Small Grains Genotyping Centers has allowed us to change our approach to MAS with Fhb1 and other markers, from a low-throughput system capable of analyzing a few hundred to thousands of genotypes per year. Currently we are screening F2 and F3 individuals using a single SSR marker (Xbarc133) or sequence tagged site marker developed from an EST identified in our fine mapping of Fhb1. Markers are available for leaf rust resistance genes Lr21 (Huang et al., 2003), Lr46 (William et al., 2003; Rosewarne et al., 2006), and Lr47 (Helguera et al., 2000), and the high molecular weight glutenins (e.g., Ma et al., 2003), that are segregating in many of our breeding populations. Plants are grown in the greenhouse and small portions of leaf tissue are harvested into 96-well microtiter plates and sent to the genotyping center in Fargo, ND. The genotyping center uses automated DNA extraction and combines four 96-well plates into a single a 384-well plate for polymerase chain reaction and fragment analysis using a semi-automated 16-capillary gel system, ABI3130xl, from Applied Biosystems (Foster City, CA). More than 14,000 marker datapoints, representing 6500 genotypes from 21 populations from our breeding program were provided by the genotyping center during the past year. We analyze populations of 384 individuals when seed supplies allowed. Our marker screening efforts are now concentrated on (i) parental screening for key genes to identify additional MAS opportunities; (ii) high priority F2 populations using enrichment; and (iii) BC1F1 or topcross populations to discard individuals that do not contain desired genes from a parent in the first cross. Ultimately, the success of this approach will be judged based on the frequency and performance of cultivars and parental germplasm that were selected using marker information.
|
Conclusions and Future Prospects |
|---|
|
TOPABSTRACTINTRODUCTIONConclusions and Future ProspectsREFERENCES |
|---|
Received for publication July 9, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONConclusions and Future ProspectsREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
