| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. Crop Sci., North Carolina State Univ., Raleigh, NC, 27695-7620 USA
b USDA-ARS and Dep. Crop Science, North Carolina State Univ., Raleigh, NC, 27695-7631 USA
c Dep. Statistics, North Carolina State Univ., Raleigh, NC, 27695-7803 USA
d Pioneer Hi-Bred Int., Inc., 7300 NW 62nd Ave., P.O. Box 1004, Johnston, IA, 50131 USA
e Dep. of Crop and Soil Sciences, and R.S. Hussey, Dep. of Plant Pathology, Univ. of Georgia, Athens, GA, 30602-7272 USA
tommy_carter{at}ncsu.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
Abbreviations: NOAL, no aluminum treatment (0 µM Al3+ activity) • HIAL, high aluminum treatment (2 µM Al3+ activity) • PC, percentage of control (HIAL/NOAL x 100) • RFLP, restriction fragment length polymorphism • QTL, quantitative trait loci • cM, centimorgan • LOD, Log of the Odds • rp = phenotypic correlation of genotypic means • rg = genetic correlation between two traits • POP-LG, population-specific linkage group • USDA/ISU, the USDA/Iowa State Univ. soybean genetic map
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
Although soil amelioration and irrigation can minimize harmful Al effects on crop performance, they are often economically unfeasible (Spehar et al., 1993; Alva et al., 1986). Genetic adaptation of plants to Al-toxic soils is an attractive and potentially less expensive alternative. In that regard, soybean is grown on Al-rich soils in many parts of the world (Smith and Huyser, 1987), and the identification of Al-tolerant soybean cultivars has been pursued for many years (Hanson and Kamprath, 1979; Sartain and Kamprath, 1978). However, genetic manipulation of Al tolerance has proven difficult. From the breeder's perspective, the problem can be characterized as a dearth of genetic diversity among improved cultivars and a wealth of discrepancies among screening methods (Campbell and Carter, 1990; Hanson, 1991; Foy et al., 1993; Bennet and Breen, 1991; Foy et al., 1992; Dall'Agnol et al., 1996).
Advances in DNA marker technology have added a new dimension to the study of genetic traits in the past decade, offering hope that marker technology can begin to clarify the genetics of Al tolerance as an aid to practical breeding. In soybean, DNA marker analyses have identified chromosomal regions with genes for pest resistance (Diers et al., 1992a; Weisemann et al., 1992), nitrogen fixation (Landau-Ellis et al., 1991), and important agricultural crop characteristics such as plant height, lodging, and maturity (Diers et al., 1992b; Lee et al., 1996a, c), seed composition (Lee et al., 1996b) and seed weight (Mian et al., 1996b). Reide and Anderson (1996) identified a single major gene for Al tolerance using RFLP marker technology in wheat (Triticum aestivum L.). Association between molecular markers and Al tolerance has not been reported in soybean.
In the 1980s, a soybean plant introduction from Japan, PI 416937, was identified as drought tolerant (Sloane et al., 1990). Since its discovery, PI 416937 has been included in numerous studies related to drought stress tolerance in soybean (Mian et al., 1996a; Hudak and Patterson, 1996). Further investigation revealed that the PI 416937 was also Al tolerant (Campbell and Carter, 1990). Heritability of Al tolerance derived from the PI is reported (Bianchi-Hall et al., 1998). The objective of our research was to identify QTLs that control Al tolerance in a F4-derived soybean population from the cross of Young and PI 416937. Knowledge of the genetic basis for Al tolerance should facilitate transfer of this trait from exotic germplasm to economically important cultivars.
|
Materials and methods |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
The hydroponic system consisted of eight individual tanks located in a controlled environment chamber in the Biological Resource Center at NCSU. Details of the hydroponic system and experimental set up were provided elsewhere (Bianchi-Hall et al., 1998). Each tank held 84 L, and continual water movement between the upper and lower reservoirs assured aeration of the system. The growth medium was deionized water with 800 µM Ca2SO4 maintained at a pH of 4.3 ± 0.1 which was automatically monitored and adjusted. The aluminum treatments consisted of 0 and 2 µM Al3+ activity (NOAL and HIAL, respectively), achieved through the addition of Al Cl3. Following germination in the dark at 25°C, seedlings corresponding to stage G3 (Muthiah et al., 1994) were transferred into the hydroponics system and roots were submerged in the medium. After 24 h, individual root lengths were measured and then Al treatments were imposed. After 72 h, root lengths were again measured. Linear tap root extension was defined as final minus initial root length for the 3-d period. Experiments were completed during November 1995 and February 1996.
RFLP Linkage Map
The population was used to establish an RFLP linkage map. The DNA isolation and RFLP protocols were conducted at the Univ. of Georgia and have been described elsewhere (Lee et al., 1996a). The linkage map was constructed with marker data using the Kosambi (1944) map function of GMendel (Holloway and Knapp, 1993) taking into account the F4-derived population structure. For combining markers into population-specific linkage groups (POP-LG), a minimum LOD of 3.0 and maximum distance of 50 cM between linkage markers were used. The genetic map consisted of 155 RFLP markers in 33 linkage groups and covered approximately 973 cM or about half of the soybean genome (Mian et al., 1996a). To facilitate the use of our results, the map was reconciled with the more highly saturated USDA/ISU soybean genetic map using anchor probes and markers (Shoemaker and Specht, 1995). Anchor probes had the same restriction enzyme and an identical banding pattern with the USDA/ISU map.
|
Experimental Design and Data Analysis |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
Analyses of variance, heritability, and genetic correlation estimates were reported previously (Bianchi-Hall et al., 1998). Neither heterogeneity of error variance nor deviations of residuals from normality were a factor in this study. Thus, results based only on untransformed data were employed here. Progeny means from the four sets were combined into a single data set for comparison with RFLP data. To standardize across the four sets, progeny means were adjusted from triplicate entries of the parents as a reference by the following formula: Adjusted set mean for progeny = Unadjusted set mean for progeny - (Parental mean of set - Parental mean over all sets). This adjustment minimized bias that may have arisen from small weekly differences in experimental conditions (Table 1) .
|
for each linkage group as 0.05. Subsequently, the Bonferroni method was employed to determine significance for each individual test within a linkage group (Senn, 1997). On average, approximately five markers were represented per linkage group, and, thus, significance of each marker was tested at probability level /5 or 0.01. Markers were tested for association with rooting phenotype based upon type III mean squares obtained from the GLM procedure of SAS (SAS, 1990; Fig. 1
and Table 2)
.
|
|
Results from the single-factor and interaction analyses were used to construct a multiple regression analysis. Those markers which interacted significantly (P < 0.01) were included in a multiple regression analysis along with the best markers (i.e., those most strongly associated with phenotype) from single factor analyses within each linkage group with the restriction that the best markers were also significant at P < 0.01. In this multiple analysis, exclusion of heterozygous loci was not practical because this would have forced the deletion of all breeding lines which were heterozygous for one or more markers in the regression model. In a multiple analysis of five markers, for example, exclusion of heterozygotes would forfeit as much as one-half of the data set (i.e. probability that a F4-derived progeny is homozygous at five independent markers is 0.8755 = 0.51) and weaken inferences from the analysis. As an alternative, we retained heterozygotes in the regression and then partitioned the two degrees of freedom associated with a marker into linear and quadratic effects. Linear corresponded to the contrast of the homozygote effects and quadratic corresponded to the contrast of the heterozygote class to the average of the homozygote classes. The linear x linear interaction corresponded to the two-way interactive effects of homozygote classes for the markers. Linear effects and linear x linear interactions were employed to assess marker effects. The heterozygote vs. homozygote contrast main effects and interactions were generally small and pooled with error for testing.
|
Results |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
|
|
RFLP Markers Associated with PC
Progeny differed significantly (P < 0.01) for Al tolerance expressed as PC. The PC of the 120 lines showed continuous variation and slight transgressive segregation for this trait, although the highest ranking F4-derived lines were not significantly different from PI 416937 (Table 1 and Bianchi-Hall et al., 1998). Single factor and pairwise analysis of two-way interactions indicated that RFLP markers on five linkage groups were potentially associated with PC (P < 0.01, Tables 2 and 3). The best marker from each of these five linkage groups (i.e., the marker which had the strongest association with phenotype as a main effect or interaction) was incorporated into a multiple regression analysis. This analysis confirmed a positive association (P < 0.01) of four independent RFLP markers to PC (Table 4), with three alleles derived from PI 416937 and one from Young (for marker EV2-1) (Table 2). The general location of one putative QTL on the linkage map (B122-1) was confirmed by three additional linked markers which were also associated with PC but to a lesser degree (Fig. 1). The heritability of PC, which was based on genotypic means of five replicates, was 0.76 (Bianchi-Hall et al., 1998) while the multiple regression analysis accounted for 24% of the phenotypic variation.
|
Discussion |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
QTL and Al Tolerance
Genotypic variation for root extension was larger under Al stress than for control conditions, suggesting that genes may exist which are specific for Al tolerance in the population (Bianchi-Hall et al., 1998). The QTL analysis revealed that this is likely the case. Multiple regression analyses identified a total of six independent putative QTL for Al tolerance (expressed as either growth under HIAL stress or as percentage of control), but only two (on POP-LG13 and POP-LG18) were common to linkage groups associated with growth in the non-stress treatment (Table 4). The other four appeared to have a unique role in Al tolerance. Of the two linkage groups affecting growth both in the presence and absence of Al, one (POP-LG18), was likely a carrier of an additional fifth gene specific for Al tolerance as indicated by the following allelic effects. For this linkage group, an allele from PI 416937 (marker B122-1 on linkage group 18) improved growth only in the presence of Al, while an allele from Young (marker B074-1) improved growth only in the absence of Al. It is unlikely that both markers identify a single gene locus responsible for such dissimilar effects. The two-gene hypothesis for this linkage group is further supported by the fact that four widely dispersed markers on POP-LG18 are all associated with PC, while only the one marker (B074-1) was associated with NOAL conditions (Fig. 1).
Because the genetic correlation between HIAL and NOAL treatments was significantly greater than zero (rg = 0.51**) (Bianchi-Hall et al., 1998), we inferred that, in addition to QTL specific for HIAL conditions, there must be other genes in the population which contribute to general plant vigor, thus enhancing root growth under both HIAL and NOAL conditions. Bouton and Parrott (1997) have suggested that this may also occur in alfalfa (Medicago sativa L.). Despite the plausibility of this idea, the theory could not be confirmed by marker results. There was no instance in which an allele from a single marker was positively (P < 0.01) associated with growth under both HIAL and NOAL conditions. The only potential example of a single QTL affecting growth under both conditions was provided by POP-LG13. For this linkage group, the PI appeared to possess two linked alleles for growth—one detected only under HIAL and the other only under NOAL conditions (Tables 2 and 4). However, the alleles were from two markers approximately 40 cM apart, indicating a strong possibility that the two marker alleles were linked to two distinct QTL rather than one (Fig. 1).
Al tolerance expressed as growth in Al per se and as PC were highly correlated (rp = 0.70** based on genotypic means; rg was not employed here because of inherent correlated errors between Al per se and PC), suggesting that many of the same genes control these traits. In fact, this was the case. Three of the four independent markers associated with PC were also associated with root extension under HIAL stress. Surprisingly, despite the high correlation and similar QTL for the two Al tolerance measures, the distributions of progeny means for HIAL conditions and PC were clearly different (Bianchi-Hall et al., 1998). For HIAL conditions, the distribution fit a normal curve, while for PC the distribution was skewed and bimodal, implying a more qualitative inheritance than for HIAL (Bianchi-Hall et al., 1998). Given the contrasting progeny distributions but similar RFLP marker results, one may suspect that undiscovered QTLs remain present in the population which affected PC more than growth per se under HIAL stress.
Relation of Al Tolerance to Agronomic Traits
A breeder may question whether or not selection for Al tolerance will lead to deleterious effects on other traits. In other work, this population has been evaluated for maturity date, 100-seed weight, lodging, plant height, and seed protein and oil content (Mian et al. 1996b; Lee et al., 1996a, b, c). Aluminum tolerance, whether expressed as growth under HIAL conditions or as PC, was not correlated phenotypically to any of these traits. The most important agronomic trait associated with Al tolerance was pod dehiscence. An allele from PI 416937 for marker B122-1 was associated with both Al tolerance and pod dehiscence (Bailey et al., 1997). Water use efficiency (WUE) may also be associated with Al tolerance. Mian et al. (1996a) identified four QTL associated with water use efficiency (WUE) in the same population, two of which were on the same POP-LG as the B122-1 marker. Both alleles for poor WUE derived from the PI, indicating that selection for Al tolerance has the potential to decrease WUE as well as increase pod dehiscence.
Implications to Breeding
Our results provide the first clear basis for the underlying genetics of Al tolerance in soybean. We detected six putative QTL for Al tolerance under HIAL conditions in hydroponics assay. This estimate of gene number agrees roughly with an earlier estimate of three to five genes for Al tolerance for this same population (Bianchi-Hall et al., 1998). This earlier estimate was based upon an algebraic analysis of parental phenotypes in the progeny and probably underestimated the true number of genes because of a crucial assumption that all positive alleles derive from one parent. We found that while most RFLP alleles for Al tolerance were derived from the PI 416937, an allele from susceptible Young (for marker EV2-1) was associated with Al tolerance expressed as PC. Fulton et al. (1997) reported a similar phenomenon in which a QTL for large fruit was identified in a small fruited relative of the cultivated tomato (Lycopersicum esculatum Mill.). In soybean, Hnetkovsky et al. (1996), also found a beneficial QTL derived from a susceptible parent when mapping markers associated with Sudden Death Syndrome disease. In the present study, the Al stress treatments were designed to impose the maximum effect on susceptible cultivar Young without inhibiting root extension of tolerant PI 416937. Thus, it was not possible to detect levels of Al tolerance greater than that of the PI in the progeny. Although this is a tantalizing possibility, the ability of the allele from Young to raise Al tolerance beyond that observed in the PI 416937 could not be ascertained in this study.
The potentially important Al tolerance associated with marker B122-1 in PI 416937 may prove difficult to capitalize upon in applied breeding, because of its association with a detrimental pod dehiscence QTL in the PI (Bailey et al., 1997). The five most Al-tolerant progeny in this study averaged twice the level of pod dehiscence as the five most Al susceptible (Bailey et al., 1997; data not shown). All five Al-tolerant types above carried the PI 416937 marker allele for B122-1 compared with none of the five Al-sensitive types. Fourteen of the 15 progeny most resistant to pod dehiscence carried the Young allele for marker B122-1, further illustrating the importance of this marker. The one progeny resistant to pod dehiscence and carrying the PI allele for the B122-1 marker was sensitive to Al and thus it is not clear that a linkage between pod dehiscence and Al tolerance was broken. Additional markers on POP-LG18 may help to distinguish between linkage and pleiotropic effects.
One may speculate that Al tolerance is rare among North American cultivars because of linkage effects associated with intense selection against pod dehiscence in the adaptation of soybean to mechanical harvest. Horst and Klotz (1990) screened more than 1000 accessions from the USDA soybean collection and found one accession comparable to the PI 416937 in terms of Al tolerance. This accession was also highly susceptible to pod dehiscence (T.E. Carter, 1995, personal communication). The relation of Al tolerance with other traits in the population studied here indicates that selection for Al tolerance will probably have little impact on agronomic traits such as maturity, lodging, and plant height in practical breeding. Water use efficiency may be decreased somewhat, however. The economic impact of WUE in this population is under current study.
The putative QTL for Al tolerance identified in this study are candidates for applied breeding. Validation of results presented here will be an important component of commercial application. The actual utility of DNA markers in comparison to direct phenotypic screening for Al tolerance will be determined by cost effectiveness. With refined DNA tagging of QTL, marker assisted selection achieved through service companies may pose a viable alternative to development of an in house phenotypic assay of Al tolerance, especially where Al tolerance is a secondary rather than primary breeding objective. Presently, no service companies offer assays for Al tolerance. Marker assisted selection may offer the added advantage of dissecting and overcoming breeding bottlenecks such as the detrimental linkage that may exist between Al tolerance and pod dehiscence alleles.
The results provided here may also guide the use of additional Al-tolerant PIs in breeding. As new PIs with Al tolerance are identified, markers may elucidate contrasting genetic mechanisms for tolerance in comparison to the Al-tolerant PI 416937 now in use, leading to strategies for the pyramiding of Al tolerance genes in a practical breeding program.SAS Institute 1990
Received for publication January 27, 1997.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMaterials and methodsExperimental Design and Data...ResultsDiscussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  N. C. Collins, F. Tardieu, and R. Tuberosa Quantitative Trait Loci and Crop Performance under Abiotic Stress: Where Do We Stand? Plant Physiology, June 1, 2008; 147(2): 469 - 486. [Full Text] [PDF]
|
|
|
|
 |
|
 J.M. Narvel, T.E. Carter Jr., L.R. Jakkula, J. Alvernaz, M.A. Bailey, M.A.R. Mian, S.H. Lee, G.J. Lee, and H.R. Boerma Registration of NC113 Soybean Mapping Population Crop Sci., March 1, 2004; 44(2): 704 - 706. [Full Text]
|
|
|
|
 |
|
 J. F. Ma, R. Shen, Z. Zhao, M. Wissuwa, Y. Takeuchi, T. Ebitani, and M. Yano Response of Rice to Al Stress and Identification of Quantitative Trait Loci for Al Tolerance Plant Cell Physiol., June 15, 2002; 43(6): 652 - 659. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. R. Villagarcia, T. E. Carter Jr., T. W. Rufty, A. S. Niewoehner, M. W. Jennette, and C. Arrellano Genotypic Rankings for Aluminum Tolerance of Soybean Roots Grown in Hydroponics and Sand Culture Crop Sci., September 1, 2001; 41(5): 1499 - 1507. [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 | |||
