| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, U.S. Dairy Forage Research Center, 1925 Linden Dr. West, Madison, WI 53706-1108
b Dep. of Agronomy, Univ. of Wisconsin, Madison, WI 53706-1597
* Corresponding author (mdcasler{at}wisc.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: DM, dry matter • NDF, neutral detergent fiber • RAPD, random amplified polymorphic DNA
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cell wall concentration, estimated as NDF, is a measure of fibrous bulk, which is negatively associated with voluntary intake (Van Soest, 1994). The concentration of NDF is generally considered to be the single laboratory variable most closely associated with voluntary intake potential (Van Soest, 1994). The concentration of NDF is heritable in several forage species, and rates of gain from selection for reduced NDF concentration of herbage are reported to be as high as 12 g NDF kg–1 DM year–1 (Casler and Vogel, 1999).
In perennial forages, selection for reduced NDF concentration generally leads to reduced forage yield (Casler, 1999; Surprenant et al., 1988). As an approach to ameliorate this positive correlation, combined selection for low NDF and high forage yield reduces the correlated responses for low forage yield, but also reduces the genetic gains for reduced NDF concentration (Surprenant et al., 1988). Observed selection responses for forage yield, following selection for NDF concentration, suggest that linkage and/or pleiotropy are the cause of genetic correlations between these two traits (Casler, 1999). The distinction between linkage and pleiotropy is important because linkages can be broken by selection and recombination, whereas trait–trait associations caused by pleiotropy cannot be broken.
Segregating progeny are required to separate the effects of linkage from pleiotropy. If pleiotropy governs the genetic correlation between forage yield and NDF concentration, the relationship should not change across generations, such as F1 or S1 progeny. However, linkage between loci controlling forage yield and NDF concentration would result in changes to the genetic correlation between the two traits (Lande, 1984).
In addition, random drift can also explain some of the reduction in forage yield associated with selection for low NDF concentration (Casler, 1999). Drift is a potentially reversible process when selection for low NDF is conducted in multiple and diverse germplasm pools. Hybridization of diverse low-NDF populations could restore forage yield potential via heterosis, possibly retaining the low-NDF trait unless the heterotic response of NDF is similar to that of forage yield (Casler, 1999).
The objectives of this study were (i) to determine the relationship between forage yield and NDF concentration in parents and segregating progeny of smooth bromegrass and (ii) to determine the heterotic responses of forage yield and NDF in smooth bromegrass F1 hybrids.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
A diallel with bulked reciprocals was made among the seven clones by planting paired-clone crossing blocks at Arlington, WI, USA, in April 1997. Each crossing block consisted of two adjacent squares of 18 plants each, on a 60-cm spacing, and planted in a 3 x 3 configuration. The plants on the periphery of the square were from one clone but the plant in the center was from the opposite clone and was the only clone from which seed was harvested. Adjacent squares represented reciprocals of a cross. The crossing blocks were isolated by at least 100 m from other smooth bromegrass to avoid pollen contamination. Seed was harvested from female plants in July 1997 and bulked for both members of a cross.
All seven clones were also self-pollinated in the greenhouse during the winter of 1997–1998 (Delgado et al., 2000). Smooth bromegrass is a self-incompatible species with different degrees of incompatibility, so 12 to 18 plants per clone were used for selfing. The clones were transplanted from 30-cm3 plastic pots to 225-cm3 plastic pots in November 1997, allowed to grow for 2 to 3 weeks and then taken to cold frames during the first week of December. The temperature in the cold frames ranged from –6 to 4°C during unusually warm winter days and the plants were kept in the cold frames for 4 wk. After the vernalization period, the plants were taken into the greenhouse and were isolated from one another by cheesecloth cages. Every morning, during flowering time, the plants were shaken to allow the pollen to flow among panicles within the cages. Flowering occurred from mid-February to mid-March. Seeds were harvested between April and May 1998 and were stored in the refrigerator until the end of May when they were planted for their crown rust evaluation. All seven clones produced self-pollinated seed to varying degrees.
Field Evaluation
Seeds from all crosses and self-pollinated families were germinated in the greenhouse in June 1998. Three self-pollinated families were excluded because of low seed numbers, low germination, and low vigor and survival in the winter 1998–1999 greenhouse. Clonal ramets of each parent were established in the greenhouse along with the progeny. Parents and progeny were transplanted to the field at Arlington, WI, in May 1999. The soil was a Plano silt loam (fine-silty, mixed, mesic, Typic Argiudoll). The experimental design was a randomized complete block with four replicates. Each cross was represented by two plots within each block. Plots consisted of either 12 different progeny seedlings or 12 identical (parental) clonal ramets in a 2 x 6 arrangement. Plants were spaced 0.3 m apart within plots and 0.9 m apart between plots. Plants were clipped twice during the establishment year and fertilized with 56 kg N ha–1 in late June and late August.
Plots were harvested with a flail-type harvester in June and September 2000 and 2001 at a cutting height of 9 cm. Individual transplants had tillered sufficiently so that they were indistinguishable by April 2000. Plots were fertilized with 112 kg N ha–1 in early spring and after the first harvest. A 500-g sample of fresh forage from each plot was dried at 60°C and used for dry matter determination.
Forage samples were ground to pass a 1-mm screen in a Wiley-type mill. Each sample was scanned on a near-infrared reflectance spectrophotometer and a calibration subset of 72 samples was chosen by cluster analysis of spectral reflectance values (Shenk and Westerhaus, 1991). The calibration samples were analyzed for NDF by the procedure of Van Soest et al. (1991) with the exceptions that sodium sulfite and 
-amylase were excluded. Values of NDF were predicted for all samples by a single calibration equation: SEP (standard error of prediction) = 9.1 g kg–1 and R2 = 0.93.
Molecular Markers
Genomic DNA was extracted from the seven parental clones. Fresh leaves (0.1–0.2 g) were macerated in potassium ethyl xanthogenate (PEX) DNA extraction buffer with a ceramic bead using a FastPrep FP120 machine from BIO 101 Inc. (Carlsbad, CA). The remainder of the DNA extraction procedure followed Johns et al. (1997) with minor modification. The samples were ground in 450 µL of DNA extraction buffer in 2.0-mL microcentrifuge tubes. All remaining DNA extraction procedures were performed in 1.5-mL microcentrifuge tubes.
Reactions for RAPD analysis were performed in 10-µL volumes in 96-well plates in an MJ PTC-100 incubator (MJ Research, Watertown, MA) following the methods of Johns et al. (1997). Sixteen 10-mer primers (A9, A18, AE4, AE10, AE12, AE14, AE16, AE17, AE20, AF6, AF14, AF20, AG10, AG20 from Operon Technologies, Inc. and UBC204 and UBC318 from Univ. of British Colombia) were selected for this study because of the consistent clarity and reproducibility of polymorphic bands and polymorphisms on a wide array of unrelated DNA samples of B. inermis. All RAPD reaction products were electrophoresed on agarose gels as described by Johns et al. (1997). Gels were run for 2 h at 300 V, stained with ethidium bromide, illuminated by UV light, photographed, and manually scored for presence/absence of clear bands. Comigrating polymorphic fragments, possessing unambiguous differences among the clonal DNA samples and ranging from 0.2 to 2.1 kb, were visually scored for presence (1) or absence (0) of the band. A total of 329 polymorphic bands were scored and used in the RAPD data analyses.
Statistical Analyses
Genetic distances among the seven clones in all pairwise combinations were estimated as the complement to Jaccard's similarity coefficient (Gower, 1972). The 7 x 7 genetic distance matrix was characterized by two orthogonal coordinates using a multidimensional scaling (MDS) procedure (PROC MDS; SAS Institute, 1999). The MDS procedures have been used to represent genetic relationships among genotypes using molecular-marker-derived genetic distances (Beebe et al., 1995; Skroch et al., 1998; Rodriguez et al., 1999). Cluster analysis, using the unweighted pair-group method of arithmetic averages (UPGMA), was used to group the seven clones on the basis of 329 RAPD bands scores.
Three autocorrelations, representing three distance classes, were computed from the 7 x 7 genetic distance matrix. The three distance classes were within populations, among steppe populations, and between steppe and meadow populations. Autocorrelation coefficients were computed by the method of Smouse and Peakall (1999). First, the distance matrix was converted to a covariance matrix. Second, three 7 x 7 incidence matrices were defined, one for each of the three distance classes. Incidence matrices consisted of 0 (absent from distance class) or 1 (present in distance class) for the off-diagonal elements and number of "hits" for each clone on the diagonal (Smouse and Peakall, 1999). Third, the autocorrelation for distance Class h, the combined correlation between all pairwise members of a distance class, was computed as
 |
Phenotypic data were analyzed by analysis of variance using the split-plot-in-time model (Steel et al., 1996). Mean squares for crosses were partitioned into sum of squares for general and specific combining ability according to Griffing's method 4 for fixed effects (Griffing, 1956). Interactions of GCA and SCA with harvests or years were computed by the following two-step procedure: (i) compute GCA and SCA sums of squares for each year and for totals over years, (ii) compute the sums of squares for interaction as SSGCAxY = SSGCA(y1) + SSGCA(y2) – [SSGCA(T)/y], where y1 and y2 = Years 1 and 2, T = totals, and y = number of years. Effects for GCA and SCA, and their standard errors, were computed according to Griffing (1956). All effects were assumed to be fixed, except replicates, which were assumed to be random.
Midparent heterosis effects for forage yield and NDF concentration were computed for each cross within each replicate and year and expressed as a percentage of the midparent mean. Midparent heterosis effects were analyzed by analyses of variance as described above. Inbreeding effects were computed as the reduction of S1 progeny performance relative to parental performance. Inbreeding effects were tested by contrasts within the analyses of variance for forage yield and NDF concentration. The phenotypic correlation coefficient between midparent heterosis effects for forage yield and NDF was tested by permutation test for matrix correlations (Smouse et al., 1986).
Genetic distances of each parental pair were used as a descriptor of the genetic dissimilarity of the parents. Patterns of midparent heterosis effects for forage yield and NDF were investigated by (i) linear regression of heterosis effects on parental-pair genetic distance and (ii) orthogonal contrasts of Steppe x Steppe within populations vs. Steppe x Steppe between populations and Steppe x Steppe vs. Steppe x Meadow.
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
|
|
The phenotypic correlation coefficient between parent clone means and GCA effects was r = 0.83 (P < 0.05) for forage yield and r = 0.92 (P < 0.01) for NDF concentration (Table 3). The phenotypic correlation between forage yield and NDF concentration of parent means was r = 0.86 (P < 0.05), but was reduced to r = 0.63 (P > 0.05) for GCA effects and to r = 0.49 (P < 0.05) for the 21 cross means. The genotypic correlation between forage yield and NDF was reduced from rg = 0.99 ± 0.01 for parents to 0.71 ± 0.14 for the 21 crosses. Four GCA effects were significantly different from zero for both forage yield and NDF, only two of which had a common sign, the two clones with the most extreme values (B15 and S28). Crosses with high mean forage yield and low mean NDF were identified (Fig. 3) .
|
|
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Clone S28 was the most distant (genetically unique) of the seven clones. This clone, derived from a wild population collected in the Altai Mountains of southern Siberia, is classified as the meadow climatype of smooth bromegrass (Casler and Carlson, 1995; Vogel et al., 1996). Meadow climatypes (more northern adapted in North America) derive largely from Russia, while steppe climatypes (more southern adapted in North America) derive largely from central Europe (Vogel et al., 1996). Meadow climatype germplasm is phenotypically distinct from steppe climatype germplasm, generally having a less extensive root system, less lateral spread by rhizomes, and a lower height ratio of reproductive to vegetative tillers (Vogel et al., 1996). Meadow and steppe climatypes can also be distinguished on the basis of agronomic and forage nutritive value traits (Casler et al., 2000).
The low autocorrelation between clones within populations was a reflection of the large amount of genetic variation that exists within smooth bromegrass populations, a pattern similar to other allogamous species. For outcrossing grasses, within-population genetic variation is generally greater than 90% (Ferdinandez et al., 2001; Huff et al., 1993; Huff, 1997; Peakall et al., 1995). For these four smooth bromegrass populations, within-population variation ranged from 94 to 95% of the total variation for a different set of 97 RAPD markers (Diaby and Casler, 2005). The autocorrelation coefficients between clones in different populations and climatypes indicated greater divergence for RAPD markers between clones derived from different populations compared with clones derived from a common population. Thus, despite the wealth of genetic variability that resides within each of these smooth bromegrass populations, the populations themselves are considerably divergent from each other.
The large correlations between parent clone means and GCA effects for both forage yield and NDF suggested that much of the variation among these clones can be attributed to differences in breeding value for both traits. The large positive phenotypic correlation between forage yield and NDF of the parent clones was expected on the basis of the results of other studies in smooth bromegrass (Casler, 1999; Han et al., 2001). A positive genetic correlation between forage yield and NDF occurs across diverse smooth bromegrass germplasm sources, although the absolute magnitude of this correlation is probably not constant (Casler, 2005).
The progressive reduction in the magnitude of the forage yield-NDF phenotypic correlation from parents to GCA effects to cross means, and of the genotypic correlation from parents to crosses, suggested that the correlation may be partly a function of linkages between loci controlling forage yield and NDF. Segregation and partial breakup of linkage blocks between generations is the most logical explanation for a change in correlation across generations (Lande, 1984). However, a previous study suggested that less than 5% of the variation in forage yield of smooth bromegrass could be attributed to linkage with loci controlling NDF (Casler, 2005). The previous study was conducted at a population level, where break-up of linkage blocks may be less obvious over a short time period, such as a single generation. Highly heterozygous clones will show obvious and measurable segregation in first-generation hybrids, whereas the breakup of linkage blocks in populations can only be measured across several sexual generations (Bingham, 1994).
Heterosis for forage yield is a common phenomenon in highly heterozygous, cross-pollinated forage crops, occurring on both population-cross and clonal-cross levels (Brummer, 1999). The heterotic effects observed for forage yield were of a similar magnitude to that typically observed in other forage crops (Knowles, 1950; Foster, 1971; Moutray and Frakes, 1973; Waldron, 1920). Crosses involving one clone generally did not result in heterosis, but this was due to the extremely high forage yield of this clone (A05), which was 27% higher than the mean of the other six parent clones. Similarly, the #2 ranked clone for forage yield had a nonsignificant heterotic effect in two crosses.
The lack of a relationship between parental genetic distance and forage-yield heterosis was not surprising. Previous studies suggest that such a relationship may occur in some cases, but it is far from reliable for predicting heterosis and single-cross performance for use in a breeding program (Diers et al., 1996; Godshalk et al., 1990; Lee et al., 1989; Melchinger et al., 1990). Bernardo (1992) itemized several requirements for molecular marker heterozygosity to be predictive of single-cross performance, including high heritability and linkage between markers and quantitative trait loci. The lack of significance of the contrast between Steppe x Steppe vs. Meadow x Steppe crosses suggested that Steppe and Meadow climatypes may not be a priori heterotic groups in this species.
While there was some evidence for midparent heterosis for NDF concentration, it appeared to derive from a different mechanism than for forage yield. The relatively low magnitude and variable sign of NDF heterotic effects was similar to heterotic effects for NDF of alfalfa (Medicago sativa L.) (Riday et al., 2002) and acid detergent fiber (ADF) of maize (Zea mays L.) (Moreno-González et al., 2000), a trait that comprises a large physical component of NDF and is typically highly correlated with NDF. Heterosis can arise by complementation of dominance alleles from different parents or by multiplicative effects (additive x additive epistasis) (Schnell and Cockerham, 1992). Furthermore, multiplicative gene action may be expressed in the form of multiplicative yield components, such as yield per tiller and tiller density (Stuber, 1999).
While any difference in the mechanism of heterosis for forage yield and NDF is purely speculative, the difference in magnitude of heterotic effects and the relatively low correlation between them suggests that low-NDF x low-NDF crosses may be a viable mechanism to combine low fiber and medium-to-high forage yield potential (Casler, 1999). For example, 10 of 21 crosses had a significant positive heterotic effect for forage yield, but no heterosis for NDF. Each of these crosses involved at least one parent that ranked low in NDF, while three crosses were between two clones that ranked low in NDF. Furthermore, the moderate correlation between forage yield and NDF of the crosses resulted from some crosses that ranked high for forage yield and low for NDF. Thus, it appears likely that a moderate to high proportion of crosses between low-NDF populations will have positive heterosis for forage yield but no measurable effect on NDF. Similarly, alfalfa crosses that showed an average of 18% heterosis for forage yield (Riday and Brummer, 2002) had an average of only 3% heterosis for NDF concentration (Riday et al., 2002).
In conclusion, the positive genetic correlation between forage yield and NDF concentration is probably grounded in the physiology of the grass plant, that NDF comprises the majority of plant dry matter and provides the architectural framework for the accumulation of dry matter in stems and leaves, i.e., that much of this correlation is due to pleiotropic effects of genes. However, it appears that this correlation is somewhat malleable: it is probably subject to modification by selection and is dynamic across sexual generations. This suggests that linkage is partially responsible for this correlation. Finally, because drift is an inevitable consequence of recurrent selection, multiple populations should be carried through a recurrent selection program for low NDF. Strain crosses, or semihybrids, produced from the complementary low-NDF populations would restore heterozygosity lost during recurrent selection and would likely retain their low fiber levels.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication March 17, 2004.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in Crop Science:
This article has been cited by other articles:
|
|
 |
|
  M. D. Casler and E. C. Brummer Theoretical Expected Genetic Gains for Among-and-Within-Family Selection Methods in Perennial Forage Crops Crop Sci., May 1, 2008; 48(3): 890 - 902. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
