| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Inst. of Plant Breeding, Seed Science, and Population Genetics, Univ. of Hohenheim, 70593 Stuttgart, Germany
b CIMMYT, Mexico D.F., Mexico
c Present address: Nordic Gene Bank, P.O. Box 41, SE-230 53 Alnarp, Sweden
* Corresponding author (melchinger{at}uni-hohenheim.de)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: AMOVA, analysis of molecular variance • FST, pairwise fixation indices • He, gene diversity • LC, landrace cultivar • MB, multiple bands • MRD, modified Rogers' distance • PC, principal coordinate • PCoA, principal coordinate analysis • SSR, simple sequence repeat
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Landrace cultivars undoubtedly represent an important source of genetic variation in wheat. One of the prime examples is the use of Rht dwarfing genes that became available through the Japanese wheat ‘Norin 10’, derived from the LC Shiro Daruma (Kihara, 1982). Two important genes, Rht1 and Rht2, were observed to directly effect yield because of reduced lodging. Moreover, a considerable LC diversity was found for resistance to pests such as stem rust (caused by Puccinia graminis Pers.:Pers. f. sp. tritici Eriks. & E. Henn.), leaf rust [P. recondita Roberge ex Desmaz. f. sp. tritici (Eriks. & E. Henn.) D.M. Henderson], or Russian wheat aphid (Diuraphis noxia Mordv.) (Skovmand and Rajaram, 1990; Skovmand et al., 1994), and for tolerance to abiotic stresses like heat (Hede et al., 1999; Skovmand et al., 2001).
With a few exceptions, all evaluations for desired traits in wheat LCs were done in ex situ collections. Examinations included either a random bulk of LC genotypes or the collections of LC sublines. Preliminary evaluation data were usually recorded during the first seed multiplication and consisted of observations that were highly heritable, easily detectable, and expressed in different environments (DeLacy et al., 2000). However, little information about the genetic variation within LCs and associations among LC accessions is available. It is also still questionable which strategy is the best to ensure an appropriate maintenance of this variation for future generations.
Molecular markers can support a more detailed characterization of genetic resources. A vast potential lies in their ability to identify the structure of genetic diversity within and among accessions, which can be of great importance for the optimization of collections, the planning of seed regeneration, and the successful implementation of prebreeding approaches. Molecular markers provide a direct measure of genetic diversity and go beyond indirect diversity measures based on agronomic traits or geographic origin. Simple sequence repeats are highly polymorphic in wheat and, therefore, suitable for the discrimination of genotypes. They are generally genome specific, abundant, codominant, and cover all 21 wheat chromosomes. They have been successfully employed to characterize genetic diversity in seed bank collections of improved wheat germplasm (Börner et al., 2000; Huang et al., 2002) and wild relatives (Li et al., 2000; Hammer, 2000).
The objectives of our study were to (i) determine SSR-based genetic diversity among and within two sets of hexaploid wheat LCs stored in the plant genetic resources center of CIMMYT, (ii) compare the form of conservation in bulks and individual plant collections, which were applied to maintain these LCs, and (iii) evaluate the use of SSRs as a tool to improve the management of wheat genetic resources.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Two sets of germplasm were used to analyze the genetic variation of hexaploid wheat LCs stored in CIMMYT's plant genetic resource center. Set 1 included 36 LCs accessions, either collected as a bulk, composed of a single LC subline, or of an unknown collection method (Table 1). Set 2 consisted of supplementary individual plant collections of five Mexican and four Turkish LCs already included in Set 1 (refer to Table 2 for the names and available number of sublines per LC).
|
|
SSR Analyses
Genomic DNA of each LC accession in Set 1 was extracted from fresh leaves of 10 to 12 randomly selected seedlings by a modified CTAB (cetyltrimethylammonium bromide) method (Hoisington et al., 1994). Quality and quantity of the isolated DNA was determined on 1% (w/v) agarose gels by comparing bands to known concentrations of lambda DNA. Equal quantities of eight DNA samples per LC accession were bulked together. For Set 2, genomic DNA was extracted from each LC subline, applying the same method.
A total of 76 SSRs was applied to fingerprint the LC accessions in Set 1. On the basis of these results the 44 most polymorphic SSRs, equally distributed over the entire genome, were selected for the analyses of Set 2. Simple sequence repeat information was provided by the Institute of Plant Genetics and Crop Plant Research, Gatersleben, Germany (Röder et al., 1998; Röder, unpublished data, 2000) and DuPont (Wilmington, DE) (Eujayl et al., 2002; DuPont, unpublished data, 2001). In addition, the marker WMC56 developed by the Wheat Microsatellite Consortium (Agrogene, France) was used. Information on map location, repeat type, annealing temperature, fragment sizes, number of alleles, as well as polymorphic information content for each SSR is available at http://www.cimmyt.org/english/webp/support/publications/support_materials/ssr_mw1.htm (verified 14 Nov. 2004). PCR amplification and allele detection were performed with an ABI-Prism Sequencer 377 in combination with the computer software GeneScan 3.1 and Genotyper 2.1 (PerkinElmer Biotechnologies, Foster City, CA), as described in detail by Dreisigacker et al. (2004).
Statistical Analyses
The proportion of SSRs showing multiple bands (MB) was determined to estimate the genetic variation of each accession in Set 1. The presence of MB indicates that for a given SSR more than one allele was observed, which may reflect residual heterozygosity and/or segregation at the respective SSR marker. Ordinary t tests were calculated to compare the observed genetic variation of LC accessions composed of bulks or single LC sublines (SAS Institute, 1990).
For the comparison of the LCs in Set 2, which was based on different numbers of LC sublines, standardized average numbers of observed alleles per locus and standardized numbers of unique alleles were calculated. Standardized values were computed by resampling 10 sublines per Mexican and two per Turkish LC and taking means over 5000 repetitions. Gene diversity of each Mexican and Turkish LC was calculated according to Nei (1973).
Analyses of molecular variance (AMOVA) were conducted on the basis of SSR data to divide the genetic variation in Set 2 into components attributable to variance components among and within LCs. Pairwise fixation indices were determined to estimate the extent of LC isolation by distance within the two countries, Mexico and Turkey. Significance levels were computed by permuting sublines between LCs.
Modified Rogers' distance was calculated for each pairwise combination in Set 1 and Set 2 according to the following equation (Wright, 1978):
 |
The AMOVA and pairwise FST values were calculated by the software package Arlequin (Schneider et al., 2000). All other analyses were performed by applying the Plabsim software (Frisch et al., 2000), which is implemented as an extension of the statistical software R (Ihaka and Gentleman, 1996).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
In the AMOVA, 18.4% of the total variance was found between the combined populations of Mexican vs. Turkish LCs in Set 2. Considering exclusively Mexican LCs, the variance within the populations accounted for 52.3%. All Mexican LCs were significantly (P < 0.05) different from each other, whereas corresponding pairwise FST values ranged from 0.37 to 0.68 (Table 4). Variance within was twice as large (67.6%) than between the Turkish LCs. The highest FST value (0.62) was found between 84TK523.001.01 and 84TK567.002 by far the smallest value (0.07) between 84TK.567.001 and 84TK.567.002. A considerably higher range of FST values was observed for the Mexican LCs.
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In our study, a considerable genetic diversity was revealed within rather than between Mexican and Turkish LCs. A surprisingly high intrapopulation diversity seemed to be in contrast to the high selfing rate of wheat but was consistent with previous results found in Pakistani wheat LCs analyzed with protein markers (Tahir et al., 1996) and Italian LCs of emmer [Triticum turgidum L. subsp. dicoccum (Schrank ex Schübl.) Thell.] analyzed with RAPDs (Barcaccia et al., 2001). The higher diversity observed within the Turkish than within the Mexican LCs can be explained by a much longer evolutionary history of wheat in Turkey. Furthermore, wheat LCs or varieties mainly transferred from Spain to the New World were presumably limited in population size, thus resulting in a founder effect.
The LCs in Set 1 were not grouping according to their continent or country of origin in the PCoA (Fig. 2). We speculate that many LCs analyzed in our study were relatively late in history transferred from the Near East or Europe to other parts of the world and/or environmental adaptation changed their genetic composition only little. The Turkish LCs, which formed a distinct cluster, were collected in the primary center of diversity of wheat in proximal locations in Hakkari, Turkey. All three LCs show resistance to Russian wheat aphid. As expected, the two LCs Barbela and Barbela 0248 were closely related, the latter being considered as a subrace of Barbela, a very old Portuguese LC showing impressively wide adaptation to different environments, in particular high acid soil and drought tolerance. The Chilean LC Trigo Africano clustered together with the African LC Dikwa 1, which was collected in a small homonymous region in the northeast of Nigeria (Zeven, 1974). The Spanish name Trigo Africano directly refers to the continent of origin, Africa, but not necessarily to the Nigerian region. In view of the lack of historical records, a larger number of accessions per country should be fingerprinted before drawing any firm conclusions about the evolutionary relationships of accessions.
Diversity within wheat LCs rests more on the allelic variation between individual plants than on heterozygous individuals. In our study, this was reflected by a low mean of heterozygosity (2.6%) observed in the sublines of Mexican and Turkish LCs, which was similar to the mean (2.5%) reported for improved lines from CIMMYT's wheat breeding program (Dreisigacker et al., 2004). An exception was the Turkish LC accession 27 (84TK567.001), which was supposedly composed of a single subline, but showed an extremely high percentage of SSRs (20.6%) with MB (Fig. 2). This high variation might be due to outcrossing, seed contamination, or experimental errors. Employing the formula of Crow and Kimura (1970)(p. 93), our estimate of mean heterozygosity corresponds to 1.3% outcrossing rate and is thus slightly higher than reported in the literature (Martin, 1990; Hucl, 1996). This outcrossing rate is sufficient to generate off-types by contamination with foreign pollen. Outcrossing might also explain why some sublines of the Mexican and Turkish LCs were positioned separately from their main clusters in the PCoA (Fig. 3). Furthermore, the Turkish LCs could be intercrossed with wild species, such as goatgrass [Triticum tauschii (Coss.) Schmal.], which are still widely grown in the mountain regions of Hakkari (Braun et al., 2001).
Bulk versus Individual Plant Conservation
In early expeditions of genetic resources acquisition, collections of bulks were preferred since the prime focus was to collect as much material as possible in a short time and to cover widely diverse geographic regions. Collecting individual plants separately was first advocated by Bennett (1970) and later reinforced by Ford-Lloyd and Jackson (1986). On one hand, the conservation in bulks offers the advantage of including seed of many different plants, which prevents a dramatic reduction in the original population size and simplifies the procedure of sampling and conservation (Frankel, 1977; Marshall, 1990). On the other hand, the presence of different genotypes makes a precise characterization of bulks difficult. Bulk accessions must therefore be "de-bulked" or evaluated on a larger scale before the best individuals are identified and used in prebreeding programs.
We observed a low molecular variation in LC accessions conserved as bulks. In general, the variance of genetic diversity measures increases with reduced numbers of examined genotypes (Weir, 1996). The variance of gene diversity in the Turkish LCs of Set 2 was higher than in Mexican LCs, the former being composed of only three to seven LC sublines (Table 2). The regeneration procedure at CIMMYT, where only 100 seeds are sown per accession, could be an additional reason for the low molecular variation observed in the bulk accessions. Small effective population sizes lead to the risk of losing molecular variation during seed regeneration. Major threats are genetic drift and selection as shown in previous studies on barley (Hordeum vulgare L.) (Parzies et al., 2000), rice (Oryza sativa L.) (Gao et al., 2000), and rye (Secale cereale L.) (Chwedorzewska et al., 2002). Therefore, larger samples for seed regeneration are recommended in the literature. Assuming a population with 20000 polymorphic loci and two alleles per locus, Lawrence et al. (1995) concluded that about 172 plants are sufficient to conserve nearly all alleles with frequencies not lower than 0.05. According to Crossa and Vencovsky (1999), for 5 to 100 loci and 2 to 20 alleles per locus, between 105 and 335 plants per population are required to maintain alleles at a 5% frequency. Some of the bulk accessions in our study were probably subsamples received from or shared with cooperators. These samples usually contain only 100 to 200 seeds, which could be another reason for a loss of variation.
In individual plant collections, alleles are usually fixed in each accession. Because of their uniformity, the accessions can be more precisely characterized and, hence, exploitation by breeders may proceed more rapidly (DeLacy et al., 2000). Its disadvantages are extensive space and labor costs essential for conservation and seed regeneration. The genetic variation within individual plant collections directly depends on the number of collected sublines. The Mexican and Turkish LCs in Set 2 might therefore represent only a part of the variation present in the original LCs. Indigenous knowledge of LCs would be extremely useful for the optimization of the sampling of sublines of each particular LC (for a review see Zeven, 2002).
Both ex situ conservation methods maintain only a part of the original LCs genetic variation and disregard their integrity. For instance, low input agriculture relies on the buffering effect of LCs, which is responsible for their broad adaptation but requires the intact original level of diversity. Thus, a combination of both conservation forms could be a reasonable solution: the storage of (i) a large bulk to preserve the natural state of the LC variation in a simple manner, and (ii) separate LC sublines representing potentially useful variants for breeding programs.
Implications of SSR-Based Genetic Diversity for Seed Bank Management
Currently some of the limiting factors in the use of LC ex situ collections are (i) missing or incomplete passport data, and (ii) the precise characterization of the collections. Passport data were not available for half of the 36 CIMMYT wheat LC accessions used in our study. Most collecting expeditions were of such a short duration that it was difficult to locate and interview all relevant landowners at the collection sites. Additionally, personnel, management, and political changes in seed banks may have contributed to the incompleteness of the records. Molecular markers may provide new and reliable information for the description and optimization of LC collections in seed banks.
The increasing costs to efficiently manage large ex situ collections encourage curators to identify redundant germplasm accessions. Verifying duplications is complex, because their definition can vary from "accessions with similar passport data" to "identical genotypes" (Hintum, 2000). Suspected duplicates were identified in collections of sorghum [Sorghum bicolor (L.) Moench] and barley by means of 15 and 35 SSRs in combination with passport data and AMOVA as a biometrical tool (Dean et al., 1999; Lund et al., 2003). In our study, 84TK567.001 and 84TK567.002 were assumed to be closely related, because of adjacent collection sites. Applying AMOVA, these LCs showed nonsignificant differences, further strengthening the notion to manage these two individual collections as one LC. However, in wheat, which is particularly suitable for seed storage, conserving either new or existing accessions in perpetuity (including regeneration in 25-yr intervals, germination tests, etc.) is currently still more cost-effective than DNA fingerprinting even with a relatively small number of SSRs (Dreher et al., 2000; Pardey et al., 2001). Thus, the identification and removal of suspected duplicates should not be considered as the main role of molecular screenings in seed bank collections.
The assessment of LC variability in seed banks demands large-scale screenings of collections. According to Zhang et al. (2002), 300 to 400 alleles are required to reflect stable relationships between wheat accessions and effectively establish core collections. In our study, 256 alleles detected with 44 SSRs (two SSRs per chromosome) were sufficient to differentiate individual genotypes of Mexican and Turkish LCs. Moreover, half of the SSRs applied were developed from expressed sequence tags, which are generally less polymorphic but might reflect functional diversity more accurately. Future opportunities to combine these markers and phenotypic data in association studies may narrow the search for new alleles at loci of interest (Thornsberry et al., 2001).
In conclusion, SSRs provide important information about the genetic variation of wheat LCs and demonstrate a powerful tool for the future tasks of seed bank management. However, the high costs warrant the further optimization of SSR application. The standardization of molecular methods, for instance, would allow to coordinate collections of different seed banks and the incorporation of new technologies like GPS, to relate the molecular diversity with their geographic dispersion.
|
ACKNOWLEDGMENTS |
|---|
Received for publication November 24, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. M. Bonman, H. E. Bockelman, B. J. Goates, D. E. Obert, P. E. McGuire, C. O. Qualset, and R. J. Hijmans Geographic Distribution of Common and Dwarf Bunt Resistance in Landraces of Triticum aestivum subsp. aestivum Crop Sci., May 18, 2006; 46(4): 1622 - 1629. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
