| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Crop and Soil Sciences, Univ. of Georgia, Athens, GA 30602
b Plant Science Dep., North Dakota State Univ., Fargo, ND 58105-5051
c Dep. of Agronomy and Plant Genetics, 411 Borlaug Hall, Univ. of Minnesota, St. Paul, MN 55108-6026
Corresponding author (elias{at}prairie.nodak.edu)
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: cM, centimorgan • GPC, grain protein concentration • LDN(DIC-6B), Langdon-T. turgidum L. var. dicoccoides chromosome 6B substitution line • RFLP, restriction fragment length polymorphism • QTL, quantitative trait locus • RI, recombinant inbred
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Breeding efforts to increase GPC in durum wheat have been marginally successful (Blanco and De Giovanni, 1995). Protein concentration is a complexly inherited trait and is strongly influenced by environmental conditions. A negative correlation between grain yield and protein concentration has hampered breeding progresses (McNeal et al., 1972; Steiger et al., 1996). However, in general, the lack of breeding progress may mostly be due to the unavailability of high protein genotypes within the adapted durum wheat breeding gene pool (E.M. Elias, personal communication). Production of durum wheat with high protein concentration is typically obtained through agronomic practices using high rates and late application of nitrogen fertilizers (Feillet, 1988).
Many studies have indicated that emmer wheat (T. turgidum L. var. dicoccoides), a wild relative of cultivated durum wheat, contains a large reservoir of high GPC genotypes (Avivi, 1978; Nevo et al., 1986). T. turgidum L. var. dicoccoides accessions with GPCs of 200 to 240 g kg-1 have been reported (Ephrat et al., 1968). The T. turgidum L. var. dicoccoides gene pool represents a new source of genetic variability for GPC and may provide potentially useful sources of high protein genes for introgression into durum and bread wheat breeding germplasms.
Joppa and Cantrell (1990) substituted individually all 14 chromosomes of a high protein T. turgidum L. var. dicoccoides accession FA-15-3 into the genetic background of the durum cultivar Langdon. This and the subsequent studies showed that five of the 14 T. turgidum L. var. dicoccoides substitution lines (2A, 3A, 6A, 5B, and 6B) have a higher GPC than the Langdon parent, indicating the presence of one or more genes controlling GPC on these chromosomes (Joppa and Cantrell, 1990; Cantrell and Joppa, 1991; Joppa et al., 1991). Joppa et al. (1997) later mapped a QTL for protein concentration in the LDN(DIC-6B) substitution line. In that study, a recombinant inbred chromosome line population was used where the 6B chromosome of T. turgidum L. var. dicoccoides and Langdon were recombined in an otherwise Langdon background. The high protein locus (QGpc.ndsu.6Bb) was located on the short arm and near the centromere of chromosome 6B, flanked by RFLP loci Xmwg79 and Xabg387. The QGpc.ndsu.6Bb locus was inherited as a single genetic factor (i.e., segregating in a 1:1 ratio in a recombinant inbred population), and explained 66% of the total phenotypic variation in protein concentration.
Mesfin et al. (1999) also mapped a high GPC QTL in hexaploid wheat (Triticum aestivum L.) introgressed from the same T. turgidum L. var. dicoccoides accession used by Joppa to develop the Langdon T. turgidum L. var. dicoccoides substitution lines. This high GPC QTL was likely the same locus reported by Joppa et al. (1997) as it was introgressed from the same high GPC T. turgidum L. var. dicoccoides accession and also mapped near Xmwg79. However, in the three mapping populations, Xmwg79 explained only 23.1 to 34.6% of the phenotypic variation (Mesfin et al., 1999), suggesting that the QTL behaves differently in different genetic backgrounds. These results suggest further evaluation to verify the presence and effects of this high GPC QTL in an adapted genetic background and over different environments is necessary before the implementation of marker-£assisted breeding. Also, since this high GPC QTL is from an unadapted germplasm, its evaluation in a high yielding, good quality genetic background should be done to detect undesirable linkage drag or pleiotropic effects on other agronomic and quality traits.
We have crossed the LDN(DIC) substitution lines (Joppa and Cantrell, 1990) to the durum cultivar ‘Vic’ (Steiger et al., 1996; Elias et al., 1996). Field testing of these populations indicated that the LDN(DIC-6B)/Vic population had higher mean GPC than the recurrent parent Vic. A majority of these lines was found to have rather poor agronomic and quality characteristics such as excessive plant height, susceptibility to lodging, and low yield (Elias et al., 1996; Steiger et al., 1996). Nevertheless, several lines with acceptable agronomic performance were selected for further improvement.
As part of our ongoing breeding efforts to increase protein concentration in durum wheat, we have further crossed one line selected from the LDN(DIC-6B)/Vic population to the durum cultivar Renville, in an attempt to move the high protein trait from T. turgidum L. var. dicoccoides into a more adapted genetic background. Our first objective was to evaluate the effect of this quantitative trait locus (QTL) on GPC, grain yield, and other agronomic traits in an adapted genetic background. Our second objective was to determine the chromosomal location of the high GPC gene(s) in the LDN(DIC-6B)/Vic//Renville population and to determine its relationship to the high GPC locus previously identified. Finally, we intended to identify genotypes carrying a minimum amount of donor chromatin retained near the protein locus that could be used for durum wheat improvement.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Plant height was measured from the ground to the tip of spike, excluding awns. Days to heading was the number of days from seeding to the date when approximately 50% of the plants had heads completely emerged from the boot. At maturity, individual plots were hand-harvested and threshed. For GPC determination in 1995 and 1996, 10 g of clean kernels from each plot was ground to whole-meal flour using a Udy Cyclone Mill (UDY Corp., Ft. Collins, CO) with a 1-mm mesh sieve. Grain protein concentration was determined by near-infrared reflectance measured by a Technicon InfraAlyzer model 400 (Technician Industrial Systems, Tarrytown, NY). In each year, the bias of the protein measurements was adjusted using the high and low protein parents from all tested locations analyzed in duplicate by standard Kjeldahl procedures (AACC method 46-10, 1983). Protein concentration measurements were adjusted to 140 g kg-1 moisture basis (AACC method 39-11, 1983).
Experimental Design and Statistical Analyses
In 1995, the RI lines, the two parents and two check cultivars (Vic and Medora) were planted in an augmented randomized block design with repeated checks (Federer, 1961) in the Prosper and Langdon locations. Each block contained 23 RI lines, with one of the parents or checks repeated every five entries in all blocks. Because of limited seed availability, RI lines were not replicated, and only 101 lines were planted in Prosper. Analysis of variance for protein data was performed in each location separately by the General Linear Model procedure (SAS Institute, 1995). Because of significant block effects, least square means adjusted for blocks effects were used for further analysis. Analysis of variance combined over the two locations was conducted by least square means as replication.
A randomized complete block design with two replications was used at all three locations in 1996. Each replicate consisted of 110 RI lines, two parents, and three checks [Medora, LDN(DIC-6B) and Vic]. Analysis of variance was computed for each trait for each location with SAS (SAS Inst., 1995). Homogeneity of error variances for the three locations was tested by Bartlett's 
2 test (Steel and Torrie, 1980). The error mean squares were homogeneous, thus analysis were combined across locations. In the combined analyses, locations, replications within locations, and genotypes were considered random effects. Genotypes x locations mean square was used to calculate Fisher's least significant differences (F-LSD) to compare genotype means. This F-LSD was used to separate the high versus low protein genotypes similar to procedures used by Riede and Anderson (1996) and Joppa et al. (1997). A high GPC line was defined as one LSD greater than low protein parent, and a low protein line as one LSD less than the high parent. On the basis of this classification, a slight overlap occurred between 150.5 and 151.5 g kg-1. One RI line had GPC within the overlap region and it was labeled inconclusive. The segregation pattern for GPC was determined by the Chi-square goodness-of-fit test based on the GPC classes separated by the F-LSD (Steel and Torrie, 1980). A RI line was considered a transgressive segregant if it had a statistically higher GPC than the high parent. Correlation of GPC with agronomic traits was computed for each location in 1996 with entry means. The correlation coefficients among locations were homogeneous (Gomez and Gomez, 1994) and therefore were pooled.
Genetic Mapping
Although Renville or Vic may have contributed minor genes for increased GPC in the LDN(DIC-6B)/Vic//Renville population, studies by Joppa et al. (1997) and Steiger et al. (1996) indicated that the high GPC trait was introgressed from a T. turgidum L. var. dicoccoides ancestor through the Langdon-T. turgidum L. var. dicoccoides 6B substitution line. Therefore, we choose RFLP markers that mapped to chromosome 6B for this study. Eighty-five low-copy DNA clones were selected based on their chromosomal location (Chen et al., 1994; Van Deynze et al., 1995; Marino et al., 1996) or on a priori information regarding the probable location of the high GPClocus (Joppa et al., 1997). The sources of these DNA clones were from barley genomic and cDNA (bcd, abg, abc), oat cDNA (cdo), rice genomic (rz), and wheat genomic and cDNA (fba, fbb, ksu, wg) libraries. Techniques for DNA isolation, and Southern hybridization have been described (Riede and Anderson, 1996). These clones were surveyed for parental polymorphisms using five restriction endonucleases (EcoRI, EcoRV, DraI, HindIII, and XbaI). Chromosome location of the hybridized fragments was determined by probing the appropriate clones to aneuploid ‘Chinese Spring’ stocks as described by Anderson et al. (1992). The polymorphic clones were first mapped on a set of 87 randomly chosen RI lines. Clones that mapped to the chromosome 6B linkage group were later probed on the remaining LDN(DIC-6B)/Vic//Renville population.
The linkage relationships of markers were determined by the program MAPMAKER for Macintosh v 2.0 (Lander et al., 1987). A minimum LOD score of 3.0 was used for two-point analysis to determine linkage relationship among markers and for multipoint analysis to establish the best loci order in the linkage group. Map distance (in centimorgan) among marker loci was calculated by the Kosambi mapping function (Kosambi, 1943).
Chromosomal location and the effects of the high GPC QTL were determined with the MQTL software (Tinker and Mather, 1995). The simple interval mapping (SIM) was performed on the combined least square entry means of the 1995 data and for the entry means of each location in 1996. Significant threshold for type I error at 5% was calculated on the basis of the 1000 permutation of the data. QTL x environment interactions were calculated on the basis of the assumption that environments are fixed (Tinker and Mather, 1995). Percentage of the phenotypic variation associated with the marker loci was calculated on the basis of marker class sum of squares as a fraction of entries sum of squares from simple marker regression with the computer software QGene (Nelson, 1997).
There was an inconsistency between the GPC statistical and MQTL analyses. In the statistical analysis, the locations were considered random, whereas in the MQTL were considered fixed. Since the error mean squares for GPC were homogenous, we could do mapping analyses using the combined data, but we chose not to because we may risk the possibility of not being able to detect environmentally sensitive QTLs. The MQTL software treats environments as fixed effects (a built-in procedure), therefore the results of our analysis applicable only to the environments where the data was gathered. However, the treatment of locations as fixed does not affect inferences that are made about the position of the QTL when using MQTL.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
The top 10% high GPC RI lines had higher GPC than LDN(DIC-6B)/Vic and LDN(DIC-6B) at all locations and years tested (Table 1). Five RI lines had higher GPC than the high parent LDN(DIC-6B)/Vic suggesting transgressive segregation for GPC. The mean grain yield of the top 10% high GPC RI lines was not different from Renville for any locations but was higher than LDN(DIC-6B)/Vic and LDN(DIC-6B) at Prosper and Langdon. The Minot location had moisture stress during the seedling stage, which resulted in lower overall yield and GPC.
Grain protein concentration was not correlated (P > 0.05) with plant height (r = 0.04) in the two locations where measurements were available (Table 2). The pooled correlations of protein concentration with heading date (r = 0.19) and grain yield (r = -0.21) were significant at P < 0.05. However, only one of the three locations showed a significant correlation between GPC and grain yield.
|
|
|
|
Graphical genotyping analysis (Young and Tanksley, 1989) on the eleven RI lines with recombination near the QTL region, and by associating the chromosome breakpoint with GPC phenotypes, further affirmed the protein locus was within the region flanked by Xcdo365 and Xmwg79 (Fig. 5) . For example, the RI genotypes 22-158 and 22-183 which have the T. turgidum L. var. dicoccoides allele for Xmwg79 and Xcdo365 and Renville allele on the two outside flanking markers (Xrz995a and Xfba175a) were in the high protein class. In addition, genotypes with recombination distal to Xmwg79 and Xcdo365 were high in protein concentration (21-155, 23-206, 22-191, and 22-201). However, recombination within the Xmwg79 and Xcdo365 region, which produced Renville alleles, resulted in genotypes with low protein concentration (genotypes 23-220, 23-242, 23-225, and 24-251). Line 24-249 had a T. turgidum L. var. dicoccoides allele at Xcdo365 and a Renville allele at Xmwg79, and retained the high protein allele. These results, along with mapping analysis, reinforce the conclusion that a gene(s) for increased GPC is located within the region flanked by Xmwg79 and Xcdo365 approximately 3.1 and 0.7 cM from each locus, respectively.
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Xcdo365 locus explained up to 72% of the phenotypic variance in GPC in the tested environments, and accounted for a 15 g kg-1 (or 1.5%) average increase in GPC. The magnitude of expression observed in this study and by Joppa et al. (1997) indicates that the GPC locus QGpc.ndsu.6Bb is the major genetic factor segregating in the two mapping populations. This hypothesis is supported by the bimodal segregation pattern observed in this study and by Joppa et al. (1997), and the expression was not diluted despite two subsequent backcrosses into a more agronomically adapted background. Mesfin et al. (1999) observed a much smaller effect of this locus on GPC in hexaploid wheat, which might be due to QTL x environment interactions or to other genes for (increasing or decreasing) GPC not detected in their populations. The possible transgressive segregation for high GPC observed in this study suggests other minor genes for GPC may be segregating in the LDN(DIC-6B)/Vic//Renville population.
Deckard et al. (1996) indicated that LDN(DIC-6B) had a higher protein concentration because of higher nitrogen accumulation (or uptake) and partitioning (or remobilization to the grain), suggesting a genetic factor(s) involved in nitrogen regulation was contributed by the T. turgidum L. var. dicoccoides 6B chromosome. This may explain why genotypes with this allele from T. turgidum L. var. dicoccoides have higher protein concentration than those without it.
From a durum wheat production standpoint, this high GPC allele can potentially be very valuable in environments where conditions are not conducive to producing high GPC durum with currently available cultivars and agronomic practices. In those environments, this allele could provide a boost in GPC to make the difference between a crop sold at full price and a crop subject to price discount because of lower-than-required protein concentration. Furthermore, this allele may potentially produce a commercially acceptable crop requiring less fertilizer input than durum cultivars lacking it.
From a breeding standpoint, this high GPC allele is important because the highest GPC durum cultivar currently grown in the Northern Plains, Medora, had average GPC of 148 g kg-1 in 1995 and 145 g kg-1 in 1996. Utilizing the LDN(DIC-6B) high protein allele for cultivar improvement should allow selection of breeding lines with protein concentration above the currently available commercial durum cultivars. This protein locus does not appear to affect plant height (Table 2) and gluten strength (unpublished data). Although the correlation between days to heading and protein concentration was statistically significant, it was less than 0.33. Several authors have reported a small (Steiger et al., 1996; Joppa and Cantrell, 1990) or no significant correlation between GPC and grain yield (Joppa et al., 1997; Deckard et al., 1996) for the LDN(DIC-6B) substitution line or genotypes derived from it. A weak negative correlation between yield and GPC was observed. However, breeders should be able to select lines with high GPC that are also high yielding as the mean yield of top 10% high GPC RI lines was not different from Renville (Table 1; also see Fig. 4).
With the tightly linked Xcdo365 marker, marker-assisted breeding can be used to integrate this high GPC allele into commercial durum cultivars. Selection for Xcdo365 would identify all the high GPC genotypes in the mapping population (Fig. 4), demonstrating the potency of this marker for selection. The lack of association between cdo365 marker and grain yield (Fig. 4) further supports our observation that selection with this marker for high GPC may also recover high yielding genotypes. The Xcdo365 locus was polymorphic between the T. turgidum L. var. dicoccoides parent and 51 of the 63 durum cultivars tested from France, Canada, International Center for Agricultural Research in the Dry Area (ICARDA), and the USA (Chee et al., 1997). Selection for Xcdo365 in conjunction with Xmwg79 would ensure the recovery of the GPC allele (probability of recovering a double recombinant in a 3 cM region is low), especially when used in a backcross breeding strategy. The use of RFLP marker assisted selection in early segregating generations may be particularly valuable in identifying genotypes that are fixed for the desirable allele versus those that are heterozygous. Finally, the conversion of Xcdo365 and Xmwg79 to STS-PCR markers (Talbert et al., 1994) would allow a more rapid and feasible screening process for the high GPC allele.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication February 23, 2000.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  J. D. Sherman, S. P. Lanning, D. Clark, and L. E. Talbert Registration of Near-Isogenic Hard-Textured Wheat Lines Differing for Presence of a High Grain Protein Gene Journal of Plant Registrations, May 1, 2008; 2(2): 162 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Uauy, A. Distelfeld, T. Fahima, A. Blechl, and J. Dubcovsky A NAC Gene regulating senescence improves grain protein, zinc, and iron content in wheat. Science, November 24, 2006; 314(5803): 1298 - 1301. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. Uauy, J. C. Brevis, and J. Dubcovsky The high grain protein content gene Gpc-B1 accelerates senescence and has pleiotropic effects on protein content in wheat J. Exp. Bot., August 1, 2006; 57(11): 2785 - 2794. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Mickelson, D. See, F. D. Meyer, J. P. Garner, C. R. Foster, T. K. Blake, and A. M. Fischer Mapping of QTL associated with nitrogen storage and remobilization in barley (Hordeum vulgare L.) leaves J. Exp. Bot., February 1, 2003; 54(383): 801 - 812. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. See, V. Kanazin, K. Kephart, and T. Blake Mapping Genes Controlling Variation in Barley Grain Protein Concentration Crop Sci., May 1, 2002; 42(3): 680 - 685. [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 | |||
