Genetic variation of RAPD markers for North American white clover collections and cultivars. (Crop Breeding, Genetics & Cytology).
Despite the fact that naturalized forms of white clover appear voluntarily in old or newly seeded pastures, their productivity is not sufficient to ensure a productive stand (Papadopoulos et al., 1997). White clover persistence can be improved by utilizing naturalized local ecotypes (Cope and Taylor, 1985), as they appear to have significant phenotypic variability (Fraser, 1988 and 1989; Caradus and Christie, 1998). Evaluated ecotypes of white clover collected from old pastures in Atlantic Canada, for example, have shown considerable variability for useful characteristics (Fraser, 1989; Papadopoulos et al., 1996; B.R. Christie, 2000, personal communication). Characterizing the genetic diversity and the degree of association between and within white clover varieties and naturalized ecotype collections is a first step toward developing and improving white clover germplasm and cultivars.
Random amplified polymorphic DNA technology is a reliable method for characterizing variation within and among species and populations (Excoffier et al., 1992; Gustine and Huff, 1999; Huff et al., 1993), although use of RAPDs is not appropriate for some applications, such as parentage analysis (Karlovsky, 1990; Riedy et al., 1992; Perez et al., 1998; Quiros et al., 1995; Scott et al., 1992). Genetic variation in naturalized populations of white clover was evaluated in sampled trifoliate leaves from rotationally stocked, intensively grazed pastures on 18 farms in New York, Pennsylvania, and Vermont (Gustine and Huff, 1999). Analysis of molecular variance on RAPD marker profiles showed that although these populations were highly variable across the northeast, they were genetically similar to each other. Genetic variation within and among white clover populations from a broader area of the USA and Canada, and the effects of breeding on variability in developed cultivars, has not been examined.
There has been relatively little white clover improvement in the upper midwestern or northeastern USA or in Canada, during the past 50 yr (Pederson, 1995). Most of the breeding effort in the USA, centered in the southeast and west, has focused on Ladino or large-leaf types of white clover, which tend to be less persistent than the small- or medium-leaf genotypes (Pederson, 1995). Because white clover is adapted to grazing systems, the development of new cultivars should be a priority. Naturalized plants throughout most temperate pastures provide a potential source for the germplasm needed to initiate breeding programs. The objective of this study was to evaluate, quantify, and compare genetic variation in populations of five recently collected germplasms and three released cultivars.
MATERIALS AND METHODS
Eight populations of white clover were compared using RAPD profiling. These populations consisted of: (i) five collections of GA, IA, PA, WV, and PEI; and (ii) three commercial large-leaf white clover cultivars. Alsike clover, a pasture and hay legume adapted to and naturalized in parts of northeastern USA and Canada, was included for comparison. We determined the RAPD profiles for 40 individual plants from each of the nine clover populations. The genetic distances among individual plants were measured using Euclidean metric distances calculated among all possible pairwise combinations (Gustine and Huff, 1999; Huff et al., 1993). The matrix was used with AMOVA (Excoffier et al., 1992) to estimate variance components for RAPD phenotypes. The interpopulation matrix [[PHI].sub.st] for the nine populations was used with NTSYS-pc (Rohlf, 1996) to calculate an unweighted paired group method of arithmetic averages (UPGMA) dendrogram.
Aurora is a winter-hardy cultivar developed in western Canada. Under some management and environmental conditions, alsike clover leaves resemble those of white clover, but alsike clover is not stoloniferous.
White Clover Cultivars
Regal is a large-leaf white clover cultivar developed in Alabama from five carefully selected germplasm collections (Johnson et al., 1970). Two of these were based on collections from Alabama pastures, while the other three came from a Pennsylvania hybrid and from seed produced in Iowa and Georgia. Because only five parental germplasms were used, Regal may have a narrower genetic base than many other white clover cultivars. Seeds of this old cultivar are produced in California (Caradus and Woodfield, 1997).
Sacramento is a large-leaf cultivar with an unclear pasture history, and was developed in California from California Ladino (Caradus and Woodfield, 1997).
Will, also a large-leaf cultivar, originated from 46 plants collected from North Carolina pastures (Caradus and Woodfield, 1997). After collection, the base germplasm was subjected to three cycles of selection for plant appearance and vigorous growth while exposed to pathogenic fungi and viruses. Selection pressure was relatively intense; [approximately equal to] 3% of the plants were saved per cycle of selection.
White Clover Collections
The GA population, known as GA-EAT or GA-N, was produced by intermating plants from a collection of 100 ecotypes originating from grasslands in Putnam County, GA, near Eatonton (33 [degrees] N lat). It has an intermediate leaf size, high stolon density, and profuse flowering across long time periods (Bouton et al., 1998; Brink et al., 1999).
The IA population was developed from plants collected in 1996 from 18 pastures in IA ([approximately equal to] 42 [degrees] N lat) that had not been seeded for at least 10 yr and had been heavily grazed by beef or dairy cattle. The collected plants were potted and placed in a greenhouse for the winter. In the spring of 1997, 40 plants were selected based on large leaf size and high stolon density, and intercrossed by bees in the greenhouse. Seed was harvested from the 30 seed-bearing plants (Brummer and Barnhart, 1998).
The PA population was developed from 28 plants collected in 1996 on a farm in Juniata County, PA (40 [degrees] N lat) from a 10-yr-old alfalfa--orchard grass pasture that had not been seeded with white clover and had been heavily grazed by sheep. The collected plants were potted, placed in a greenhouse, and intercrossed by bees.
The PEI population was developed from a collection of white clover seed from grassland that had not been seeded for 50 yr on a farm in Hazel Grove, PEI, Canada (46 [degrees] N lat), and grown as CRS-102 in 1990 and 1993. Remnant seed was used to establish a seed block undercage to produce Syn 1 seeds during the 1994 and 1995 growing seasons. Syn 2 seeds, used for this study, were produced in Idaho in 1997 (Caradus and Christie, 1998).
The WV population originated from plants collected in 1997 from a lawn planted in 1994 with grass seed free of clover seed. The lawn was located in a formerly forested area (37 [degrees] N lat) that had been cleared and graded. About 100 plants were collected, potted, and grown in a greenhouse during the winter. The following spring, plants were placed in an isolation cage and intercrossed.
Random Amplified Polymorphic DNA Markers
At least 40 seeds per clover population were germinated in the greenhouse. After [approximately equal to] 8 wk of growth, one trifoliate leaf was collected from each of the 40 individual plants, the fresh weight was recorded, and the leaf was placed in a plastic 2-mL microcentrifuge tube. The samples were maintained at -70 [degrees] C until they were lyophilized. Genomic DNA was prepared from 25 to 100 mg of plant tissue, as described by Gustine and Huff (1999), using 100 [micro] L of rapid one-step extraction buffer (Steiner et al., 1995) for each 25 mg (fresh weight) of powdered tissue. The RAPD markers were produced with primers OPA08, OPB14, and OPH12 (Operon Technologies, Inc., Alameda, CA (1)). Polymerase chain reactions based on genomic DNA, gel electrophoresis, and ethidium bromide staining were performed according to Gustine and Huff (1999). Gels were documented using a Kodak DC120 digital camera and bands detected with Kodak 1D Image Analysis Software (v 2.1, Eastman Kodak Co., Scientific Imaging Systems, Rochester, NY). Although a single DNA extraction was done for each individual plant, polymerase chain reactions (PCRs) were performed at least three times on each DNA preparation, and only repeatable bands were scored.
Statistical Analysis of RAPD Profiles
Each of nine populations, consisting of 40 individuals grown from randomly selected seed, was arbitrarily divided into two subpopulations of 20 RAPD profiles for AMOVA (Excoffier et al., 1992) to evaluate variance within and among the subpopulations. AMOVA calculates the correlation of random genotypes within subpopulations relative to that of random pairs of genotypes drawn from the whole population. Calculations of Euclidean metric distance matrices, AMOVA, and UP-GMA dendrogram for the nine populations were conducted as described by Gustine and Huff (1999). A single interpopulation distance matrix of nine populations was created from seven overlapping matrices. The assembled interpopulation distance matrix was used for sequential, agglomerative, hierarchical, and nested clustering method (SAHN; Sneath and Sokal, 1973) cluster analysis using the NTSYS-pc UPGMA algorithm (Rohlf, 1996).
RESULTS AND DISCUSSION
The three primers produced 28 polymorphic markers in replicate amplifications that were used for assessing genetic variation within and among the nine populations. No pairs of RAPD profiles matched, indicating that no clonal individuals were present in the populations.
To partition the total variance among individuals for a population, RAPD profiles of two subpopulations of each clover population were analyzed by AMOVA to determine the intrapopulation genetic variances. The populations had 0 to 12% and 88 to 100% of their genetic variance, respectively, among and within respective pairs of subpopulations (Table 1). These results indicate that all clover populations had high genetic variance within subpopulations and low genetic variance among subpopulation pairs. On the basis of these results, we assumed that each group of 40 plants was representative of its source population. However, the genetic variance for the Regal, Will, GA, and PA subpopulation pairs was significant (P < 0.05). Results in Table 1 show that Regal, despite its original base of only five germplasms, did not have a narrow genetic base.
Assessment of the interpopulation distance matrix in Table 2 through SAHN cluster analysis produced the dendrogram shown in Fig. 1. In the dendrogram, T. hybridum cultivar Aurora was separated as expected from all tested T. repens populations at 0.34 [[PHI].sub.st] distance units. Within the white clover populations, two distinct groups were formed at 0.25 [[PHI].sub.st] distance units. One group consisted of GA, PEI, PA, and IA, and the second group contained Regal, Sacramento, WV, and Will. Lower level groupings occurred at 0.15 or less [[PHI].sub.st] distance units.
[FIGURE 1 OMITTED]
Two surprising and striking comparisons can be found in Fig. 1. First, a population bred for improved winter hardiness in Atlantic Canada (Caradus and Christie, 1998), based on PEI (Christie, 1990), and one bred for improved grazing persistence (Bouton et al., 1998) from collections in Georgia, were closely related. Secondly, a population from West Virginia was closely related to the commercial cultivar Sacramento.
Unpublished data from related studies suggest that the experimental white clover collections evaluated in this study vary widely in leaf size, and have small to medium leaf type compared with the three cultivars. In research conducted by one of the authors (P. Voigt, 1999, unpublished data), Will white clover had a mean leaf size of 8.1 [cm.sup.2], compared with mean leaf sizes ranging from 4.7 [cm.sup.2] for GA to 6.0 [cm.sup.2] for PA, with the IA and PEI populations intermediate in leaf area. In a second study at a different location (P. Voigt, 2000, unpublished data), mean leaf sizes of Regal and Sacramento were 10.3 and 9.2 [cm.sup.2], respectively, while that of WV was only 3.8 [cm.sup.2]. Although comparisons across experiments are not possible, these data indicate that the five experimental populations used in this study had small to intermediate leaves and they varied widely in leaf size.
Small-leaf white clover tends to have short petioles, and produce dense networks of well-rooted stolons (Hay and Hunt, 1989). The networks of stolons are dense because the internodes are short, and in addition to having a high leaf-appearance rate, small-leaf types tend to have higher branching rates (Grant and Barthram, 1991). In contrast, the large-leaf types have longer petioles, longer internodes, a less dense and more poorly rooted network of stolons, and lower leaf appearance and branching rates. All these characteristics make small-leaf white clover less susceptible to grazing pressure; they are more persistent, though lower yielding, than larger leaf white clover.
The leaf size differences are in partial agreement with the relationships determined from the distance matrix. All of the larger leaf germplasm are in one of the two white clover groups (Fig. 1). However, the relatively small-leaf WV germplasm was also placed in that group. Thus, an association between leaf characteristics and the RAPD marker polymorphisms on which the dendrogram is based does not provide a good explanation of our results. Similarly, recentness of collection cannot explain these results.
Our results suggest that leaf area is not a good predictor of genetic similarity or dissimilarity. For example, Regal, Sacramento, and Will are large-leaf cultivars, while WV is a small-leaf germplasm, yet they are grouped together in Fig. 1. This is probably a function of the RAPD markers used in analyzing genetic distances. The 28 markers selected amplified portions of genomic DNA sequences of unknown function, and are based only on the randomly produced sequence of each oligonucleotide. The marker sequences thus may originate from any part of the genome, including genes, promoters, repeat stretches of DNA, and noncoding regions. Additional RAPD markers linked to genes controlling leaf size are needed to enable better discrimination of this phenotypic character, and to possibly predict white clover productivity.
The germplasms we studied had a relatively wide range of origin: from 33 [degrees] N lat for GA to 46 [degrees] N lat for PEI, with PA and IA falling in between. The point of origin of the WV population is fairly close to that of Will, the only cultivar with a clearly defined geographical point of origin (North Carolina). In this context, the separation of Will and WV from IA, PA, and PEI would appear appropriate. However, the inclusion of GA with germplasms from more northern latitudes suggests that latitude of origin does not explain the observed results.
One factor that unifies the GA, IA, PA, and PEI populations, and separates them from the WV populace, is that they have a pasture origin. Although Will, grouped with WV, also has a pasture origin, its collection from North Carolina pastures was followed by three cycles of relatively intensive selection (3.3% of plants were saved per cycle) for tolerance to pathogens and plant vigor. This could have resulted in changes that affected molecular markers. Also, the large-leaf size of Will probably has a different origin, perhaps a natural introgression of genes from large-leaf germplasm planted either intentionally or unintentionally (through hay, animals, or other means). The selection of large-leaf plants during the collection of Will's base germplasm could have increased the relationship to large-leaf cultivars such as Regal and Sacramento.
Of particular interest is the great divergence between the West Virginia population and the other recent collections. The dissimilarity of the WV to the other experimental populations argues against a common origin of all small-leaf collections in this region. Reasons for the close relationship of the WV population to the large-leaf cultivars are not known. Ostensibly, WV derived from a similar initial pool of germplasm and has a comparatively similar morphology. Its genetic divergence may make it a useful germplasm for breeding purposes. Perhaps this population, despite its smaller leaf size, derived in part from past hybridization with ladino type germplasm that had been planted in the area.
The selection of superior plants should be possible within any of the recent collections evaluated. These five experimental populations appear to represent different germplasm compared with major large-leaf cultivars. The fact that four of the experimental populations cluster together indicates that they may share similar germplasm. Whether or not they bring substantially different genetics to a breeding program needs to be further investigated. Certainly, the environmental conditions under which they were grown differ considerably, and the mild selection pressure used in their development may have selected them for different traits. If they all came from a more-or-less similar genetic base, originating with the initial settlers of North America, then the slight genetic differences among them may represent differential adaptations.
Abbreviations: [[PHI].sub.st] an analog of F as described by Excoffier et al. (1992); AMOVA, analysis of molecular variance; IA, germplasms from Iowa; GA, germplasms from Georgia; PA, germplasms from Pennsylvania; PCR, polymerase chain reaction; PEI, germplasms from Prince Edward Island, Canada; RAPD, random amplified polymorphic DNA; SAHN, sequential, agglomerative, hierarchical, and nested clustering method; UPGMA, unweighted paired group method of arithmetic averages; WV, germplasms from West Virginia.
Table 1. Analysis of molecular variance ([dagger]) of RAPD profiles for eight white clover populations and one alsike clover population. Pairs of subpopulations consisted of two randomly selected groups of 20 individuals from the same population. Variance Among Within subpopulation subpopulation P-Value among Populations variance variance subpopulations % `Aurora' 0.3 99.7 0.52 `Regal' 6.3 93.7 0.001 `Sacramento' 0.0 100.0 0.43 `Will' 7.4 92.6 0.001 GA 11.6 88.4 0.001 PEI 0.3 99.7 0.29 PA 6.2 93.8 0.01 IA 0.5 99.5 0.26 WV 0.2 99.8 0.42 ([dagger]) Analysis of molecular variance is the correlation of random genotypes within subpopulations relative to that of random pairs of genotypes drawn from the whole population. ([double dagger]) For the GA population, n = 41 samples; for the remaining populations, n = 40 samples. Table 2. Interpopulation distance matrix for eight white clover populations and one alsike clover population. Population `Aurora' GA PEI IA PA [[PHI].sub.st] ([dagger]) `Aurora' ([double 0.0000 dagger]) GA 0.2793 0.0000 PEI 0.2818 0.0722 0.0000 IA 0.3463 0.1390 0.1750 0.0000 PA 0.3470 0.1458 0.1076 0.1366 0.0000 `Regal' 0.3380 0.2508 0.2782 0.3379 0.2548 `Sacramento' 0.3900 0.2540 0.2759 0.2712 0.2415 WV 0.3635 0.1906 0.1997 0.2120 0.1986 `Will' 0.3613 0.2295 0.2734 0.2370 0.2876 Population `Regal' `Sacramento' WV `Will' [[PHI].sub.st] ([dagger]) `Aurora' ([double dagger]) GA PEI IA PA `Regal' 0.0000 `Sacramento' 0.1351 0.0000 WV 0.1369 0.0691 0.0000 `Will' 0.1474 0.1405 0.0957 0.0000 ([dagger]) [[PHI].sub.st] value, which is equivalent to the proportion of the total variance that is partitioned between two populations. All blank cells in the upper diagonal have a value of zero, indicating all corresponding values in the lower matrix diagonal are highly significant. ([double dagger]) For the GA population, n = 41 samples; for the remaining populations, n = 40 samples.
This work was partly supported by funding from the Cooperative State Research Education and Extension Service, USDA, to the NE-144 Cooperative Regional Project, Forage crop genetics and breeding to improve yield and quality.
(1) Mention of a trademark, vendor, or proprietary product does not constitute a guarantee or warranty of the product by the USDA and does not imply its approval to the exclusion of other products that may also be suitable.
Bouton, J.H., C.S. Hoveland, D.R. Woodfield, and J.R. Caradus. 1998. Selection and testing of white clover germplasm derived from naturalized ecotypes, p. 33. In C.R. Grau (ed.) Proc. Trifolium Conference, 15th, Madison, WI. 10-12 June 1998. Univ. Wisconsin, Madison, WI.
Brink, G.E., G.A. Pederson, M.W. Alison, D.M. Ball, J.H. Bouton, R.C. Rawls, J.A. Steudemann, and B.C. Venuto. 1999. Growth of white clover ecotypes, cultivars, and germplasms in the southeastern USA. Crop Sci. 39:1809-1814.
Brummer, E.C., and S.K. Barnhart. 1998. Collection of forage crop germplasm throughout Iowa. Leopold Center Progress Rep. 7:3-5.
Caradus, J.R. and B.R. Christie. 1998. Winter hardiness and artificial frost tolerance of white clover ecotypes and selected breeding lines. Can. J. Plant Sci. 78:251-255.
Caradus, J., and D.R. Woodfield. 1997. World checklist of white clover varieties II. N.J. Agric. Res. 40:115-206.
Christie, B.R. 1990. Ecotypes of white clover, p. 12. In R.R. Smith (ed.) Clovers and special purpose legumes research. Vol. 23. USDAARS, Madison, WI.
Cope, W.A., and N.L. Taylor. 1985. Breeding and genetics, p. 383-404. In N.L. Taylor (ed.) Clover Science and Technology. Agron. Monogr. 25. ASA, CSSA, and SSSA, Madison, WI.
Excoffier, L., P.E. Smouse, and J.M. Quattro. 1992. Analysis of molecular variance inferred from metric distances among DNA haplotypes: Application to human mitochondrial DNA restriction data. Genetics 131:479-491.
Fraser, J. 1988. Seasonal yield changes in white clover varieties and ecotypes in Nova Scotia. J. Agric. Sci. 110:609-617.
Fraser, J. 1989. Characteristics of naturalized populations of white clover (Trifolium repens) in Atlantic Canada. Can. J. Bot. 67:2297-2301.
Gibson, P.B., and W.A. Cope. 1985. White clover, p. 471-490. In N.L. Taylor (ed.) Clover science and technology. Agron. Monogr. 25. ASA, CSSA, and SSSA, Madison, WI.
Grant, S.A., and G.T. Barthram. 1991. The effects of contrasting cutting regimes on the components of clover and grass growth in micro swards. Grass Forage Sci. 46:1-13.
Gustine, D.L., and D.R. Huff. 1999. Genetic variation within and among white clover populations from managed permanent pastures of the northeastern U.S. Crop Sci. 39:524-530.
Hay, R.J.M., and W.F. Hunt. 1989. Competition from associated species on white and red clover in grazed swards, p. 311-326. In G.C. Marten et al. (ed.) Persistence of Forage Legumes. ASA, CSSA, and SSSA. Madison, WI.
Huff, D.R., R. Peakall, and P.E. Smouse. 1993. RAPD variation within and among natural populations of outcrossing buffalo grass [Buchloe dactyloides (Nutt.) Engelm.]. 1993. Theor. Appl. Genet. 86:927-934.
Johnson, W.C., E.D. Donnelly, and P.B. Gibson. 1970. Registration of Regal white clover. Crop Sci. 10:208.
Karlovsky, P. 1990. Misuse of PCR. Trends Biochem. Sci. 15:419.
Papadopoulos, Y.A., E.A. Butler, S.A.E. Fillmore, J. Read, B. Christie, K.B. McRae, D. Garbary, and W.G. Thomas. 1996. The evaluation of white clover varieties and new experimental synthetics under clipping and intensive pasture management systems on Nova Scotia soils. Final Report, Canada/Nova Scotia Agri-Food Development Agreement. Agriculture & Agri-Food Canada, Nappan, NS, Canada.
Papadopoulos, Y.A., E. Charmley, R.C. Martin, A.H. Fredeen, K.B. McRae, D. Mason, and S.A.E. Filmore. 1997. Productivity of renovated naturalized pastures in Atlantic Canada. Session 22. p. 151152. In Proceedings of XVIII International Grassland Congress, Saskatoon, SK, Canada.
Pederson, G.A. 1995. White clover and other perennial clovers. p. 227-236. In Forages: Vol. 1, An introduction to grassland agriculture. 5th ed. R.F. Barnes, et al. (ed.) Iowa State Univ., Ames, IA.
Perez, T., J. Albornoz, and A. Dominguez. 1998. An evaluation of RAPD fragment reproducibility and nature. Mol. Ecol. 7:1347-1357.
Quiros, C.F., P. This, M. Laudie, A. Benet, A.-M. Chevre, and M. Delsney. 1995. Analysis of a set of RAPD markers by hybridization and sequencing in Brassica: a note of caution. Plant Cell Rep. 14: 630-634.
Riedy, M.F., W.J. Hamilton III, and C.F. Aquadro. 1992. Excess of non-parental bands in offspring from known primate pedigrees assayed using RAPD PCR. Nucleic Acids Res. 20:918.
Rohlf, F.J. 1996. NTSYS-pc. Numerical taxonomy and multivariate analysis system. Ver. 1.80. Exeter Publ., Setauket, NY.
Scott, M.P., K.M. Haymes, and S.M. Williams. 1992. Parentage analysis using RAPD PCR. Nucleic Acids Res. 20:5493.
Sneath, P.H.A., and R.R. Sokal. 1973. Numerical Taxonomy. Freeman, San Fransisco.
Steiner, J.J., C.J. Poklemba, R.G. Fjellstrom, and L.F. Elliott. 1995. A rapid one-tube genomic DNA extraction process for PCR and RAPD analyses. Nucleic Acids Res. 23:2569-2570.
David L. Gustine, * Paul W. Voight, E. Charles Brummer, and Yousef A. Papadopoulos
D.L. Gustine, USDA-ARS, Pasture Systems and Watershed Management Research Unit, Curtin Road, Building 3702, University Park, PA 16802-3702; P.W. Voigt, USDA-ARS, Appalachian Farming Systems Research Center, 1224 Airport Road, Beaver, WV 25813-9423; E.C. Brummer, Dep. of Agronomy, Iowa State Univ., Ames, IA 50011; Y.A. Papadopoulos, Agriculture & Agri-Food Canada Crops and Livestock Research Centre, P.O. Box 1210, 440 University Ave., Charlottetown, PE, C1A 7M8 Canada. Received 17 Jan. 2001. * Corresponding author (firstname.lastname@example.org).
|Printer friendly Cite/link Email Feedback|
|Author:||Gustine, David L.; Voigt, Paul W.; Brummer, E. Charles; Papadopoulos, Yousef A.|
|Article Type:||Statistical Data Included|
|Date:||Mar 1, 2002|
|Previous Article:||Genetic diversity patterns among Phytophthora resistant soybean plant introductions based on SSR markers. (Crop Breeding, Genetics & Cytology).|
|Next Article:||A novel gene ef1-h conferring an extremely long basic vegetative growth period in rice. (Crop Breeding, Genetics & Cytology).|