Crop Science
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 20 May 2008
Published in Crop Sci 48:1080-1089 (2008)
© 2008 Crop Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Agricola
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Related Collections
Right arrow Other Forage Crops
Right arrow Plant Genetic Resources
Right arrow Crop Genetics

Diversity for AFLP and SSR in Natural Populations of Lotus corniculatus L. from Italy

Maria Luisa Savo Sardaroa,*, Maroun Atallaha, Elahe Tavakola, Luigi Russib,* and Enrico Porcedduc

a Programma Internazionale di Dottorato in Agrobiodiversità, Scuola Superiore Sant'Anna di Pisa, c/o ENEA CR Casaccia, Santa Maria di Galeria 00060 Rome, Italy
b Dipartimento di Biologia Applicata, Univ. degli Studi di Perugia, Borgo XX Giugno 74-06121 Perugia, Italy
c DABAC, Univ. degli Studi della Tuscia, Via S.C. de Lellis, 01100 Viterbo, Italy

* Corresponding authors (mlsavo{at}libero.it, lrussi{at}unipg.it).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Birdsfoot trefoil (Lotus corniculatus L.) is a species native to the Mediterranean basin and one of the most widely distributed perennial forage legumes. It is found in cultivated fields as well as in natural pastures of the Mediterranean and temperate regions of Europe, Asia Minor, North Africa, South and North America, Australia, and New Zealand. It is a potential species for marginal, salty, and degraded areas. The genetic variation present in 11 populations of L. corniculatus collected in natural pastures in Italy was assessed by using four amplified fragment length polymorphisms and five simple sequence repeat (SSR) markers, which were previously identified in L. japonicum. The amount of within-population variation was high, but the among-populations variation was higher, allowing discrimination among accessions. Amplified fragment length polymorphisms and SSRs markers provided an almost equal measure of the variation, although the latter provided a better characterization in terms of F-statistic FST and {rho} (estimate of population differentiation in autotetraploids) and indicated that the selfing rate in the species was higher than expected. The matrices of SSR genetic distances and the geographic distances of the collection sites were highly correlated. The information on the genetic structure of the populations is briefly discussed in terms of breeding perspectives.

Abbreviations: AFLP, amplified fragment length polymorphism • AMOVA, analysis of molecular variance • GS, genetic similarity • HW, Hardy–Weinberg • MGS, mean genetic similarity • PCR, polymerase chain reaction • SSR, simple sequence repeat • UPGMA, unweighted pair-group method with arithmetic mean


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
THE GENUS LOTUS is a large, polymorphic and widely distributed taxonomic group comprising approximately 200 annual and perennial species (Grant, 1965). Lotus corniculatus L. (birdsfoot trefoil) is cross-pollinated, tetraploid (2n = 4x = 24) species, although diploid populations have also been reported (Grant and Small, 1996). Birdsfoot trefoil is the most agronomically important species in the genus used for pasture, hay, and silage production in Europe, Asia, Africa, South and North America, Australia, and New Zealand. Its importance derives from its wide adaptability to marginal areas (Bennett, 2003), tolerance to acidic and poorly drained soils (Seaney and Henson, 1970; Marten et al., 1987), high nutritive value, and nonbloating features (Seaney and Henson, 1970; Beuselinck and Grant, 1995). A great diversity is present in the Mediterranean basin (Grant, 1991), although its center of diversification has yet to be identified (Grant and Small, 1996).

Birdsfoot trefoil populations have been assessed for cyanogenic reaction (Blaise et al., 1991), seed globulin polypeptides (Steiner and Poklemba, 1994), random amplified polymorphic DNA sequences (RAPD) (Campos et al., 1994; Steiner and Garcia de los Santos, 2001), and rDNA internal transcribed spacer sequence polymorphisms (Steiner, 1999). High variation was also found for specific agronomic traits, including herbage regrowth and quality, flowering habit, insect resistance (Chrtková-Zertová, 1967, 1973; Seaney and Henson, 1970; Duke, 1981; McGraw et al., 1989; Steiner et al., 2001), and reproductive compatibility (Garcia de los Santos et al., 2001). This wide range of variation and adaptation to different environments is the result of the high intraspecific hybridization between populations (Chrtková-Zertová, 1973).

Together with Trifolium pratense and T. repens, birdsfoot trefoil was one of the most common perennial legume species identified in an ecogeographic survey undertaken in more than 79 sites scattered in the central and southern areas of Italy (Russi et al., 2003). The assessment of the characteristics of natural populations may provide useful information on the environmental conditions of the growing sites and identify the most suitable material for breeding programs (Steiner and Greene, 1996; Steiner, 1999).

The aim of the present work was to assess the genetic variation present in 11 Italian natural populations of L. corniculatus using two molecular marker system (amplified fragment length polymorphism [AFLP] and simple sequence repeat [SSR]) and to compare the amount of information provided by each marker system.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Eleven accessions of L. corniculatus (Table 1 ) were sampled from a wide collection of perennial legumes from central Italy (Russi et al., 2003; Pagnotta et al., 2003). Fifteen seeds from each accession were planted in 20-cm pots and grown in a greenhouse. Five to 10 plants per population were analyzed, with a total of 88 plants. Genomic DNA was extracted following the CTAB method (Chen and Ronald, 1999).


View this table:
[in this window]
[in a new window]

  		Table 1. Collection sites in central Italy of the 11 accessions of Lotus corniculatus analyzed.

 
Five SSR primer pairs, previously identified in L. japonicum (Table 2 ), were used in the analysis. Polymerase chain reaction (PCR) was performed in a total volume of 20 µL containing 50 ng of DNA template, 1x PCR Buffer, 0.2 mM dNTP, 10 pmol of forward primer, 10 pmol of reverse primer, and 0.1 unit Taq polymerase. Polymerase chain reaction products were analyzed by electrophoresis on high-resolution agarose gels (Sigma-Aldrich, St. Louis, MO) (Fig. 1 ).


View this table:
[in this window]
[in a new window]

  		Table 2. Primers used in SSR analysis.

 

Figure 1
View larger version (106K):
[in this window]
[in a new window]

  		Figure 1. Example of simple sequence repeat banding patterns in Lotus corniculatus using the primer BM1397. The first lane is the DNA ladder.

 
We performed AFLP analyses according to Vos et al. (1995) with minor modifications. The primary template was prepared by simultaneous digestion of genomic DNA with EcoRI or PstI and MseI and subsequent ligation of enzyme specific adaptors (Table 3 ). Preamplification of digested–ligated DNA was performed with adaptor specific primers. Selective amplification was conducted with two or three selective bases at the 3' end of both primers. The EcoRI/PstI selective primer was end-labeled with fluorescein. Selectively amplified products were resolved on 6% denaturing polyacrylamide gels. Fluorescence labeled DNA fragments were acquired by Genomix DNA sequencer (Beckman Coulter, Fullerton, CA) (Fig. 2 ). The reproducibility of the AFLP fingerprints was assessed on three DNA samples by replicating the entire procedure starting from the original DNA for all the primer combinations.


View this table:
[in this window]
[in a new window]

  		Table 3. Adaptors and primers used in the amplified fragment length polymorphism analysis.

 

Figure 2
View larger version (143K):
[in this window]
[in a new window]

  		Figure 2. Representative acrylamide gel showing the amplified fragment length polymorphism pattern polymorphism in accessions of Lotus corniculatus using the E-AAT and M-GAG primer combination. Preitoni individuals, as indicated on the gel, were collected from south Italy.

 
Statistical Analysis
SSR Data Matrix
We considered SSR fragments as actual alleles. The scoring and determination of amplicon size were performed by Quantity One v. 4.6.0 Software (Bio-Rad, Hercules, CA), assisted by visual determination of the pattern in each lane. Lanes with four fragments were scored as four alleles in single doses (ABCD); those with one fragment were scored as one allele in a quadruple dose (AAAA). With three fragments in a lane, the most intense one was scored as a double dose, and the other two as a single dose (AABC). Two fragments of the same intensity were scored as two alleles, each in a double dose (AABB). When one fragment of two was more intense, it was considered as a triple dose (AAAB). Allele doses were scored assuming absence of null alleles.

The within-population genetic diversity was assessed by estimating the number of alleles (A), the effective number of alleles (Ae). and the genotype richness. Differences among populations for these parameters were tested by the nonparametric Kruskall–Wallis test. The percentage of polymorphic loci (a locus was defined polymorphic when the most common allele had a frequency <0.95) and mean gene diversity by Nei's index (Nei, 1973) were also calculated. Observed and expected heterozygosity was estimated by AUTOTET, a software specifically designed for autotetraploids (Thrall and Young, 2000). Chromosomes were assumed to segregate randomly because, according to Fjellstrom et al. (2001), the first to detected random chromatid segregation in L. corniculatus, a chromosomal-type segregation predominates. A goodness-of-fit test was used to assess the departure from the Hardy–Weinberg (HW) equilibrium of the observed heterozygosity, by weighting each genotype according to the proportion of potential heterozygous gametes (Bever and Felber, 1992).

F-statistics were calculated by Gene4X software (Ronfort et al., 2000), and the level of population differentiation was estimated through FST and {rho} parameters. The parameter {rho} is an analog of the correlation between truly outcrossed mates and is an accurate estimate of population differentiation, independent of {alpha} (proportion of double reduction) and s, the selfing rate (Ronfort et al., 1998). Significance was checked by Fisher's test expanded to autotetraploids (Raymond and Rousset, 1995).

The genetic distances among the 11 populations were estimated (i) by allele frequency at the five loci (Nei, 1972) and (ii) by an index specific for codominant markers (Kosman and Leonard, 2005). Distance matrices were then clustered by the unweighted pair-group method with arithmetic mean (UPGMA) (Sneath and Sokal, 1973). To estimate the goodness-of-fit of the clustering to the data matrix, cophenetic matrices were derived from dendrograms and compared with the original similarity matrix by the test of Mantel (1967) (Rohlf and Sokal, 1981). The results from cluster analysis were also validated by principal coordinate analysis (Gower, 1966), by double-centering the similarity matrix, extracting the eigenvalues and eigenvectors, and displaying the relationship among individuals in three dimensions. Genetic distances, cluster analysis, Mantel test, and principal coordinate analysis were performed with NTSYS-pc software (Rohlf, 1993).

AFLP Data Matrix
Amplified fragment length polymorphism fragments were scored as dominant markers. Fragments that could unequivocally be scored across all individuals were included in the analysis and were scored 1 for presence and 0 for absence, using Cross checker software v. 2.91 (Buntjer, 1999). Scores were used to prepare a data matrix based on the 88 individuals belonging to the 11 populations of L. corniculatus, and 254 markers. The AFLP data were inspected for the presence of unique fragments and among- and within-population polymorphisms. The AFLP binary data matrix was used to derive a matrix of genetic similarity (GS) by using the coefficient of Jaccard (1908). The among- and within-population mean genetic similarity (MGS) estimates were obtained by importing into a spreadsheet the similarity matrix and averaging the GS estimates over the individuals belonging to the 11 populations. Genetic similarity and MGS matrices were analyzed by UPGMA cluster analysis, and the goodness-of-fit was validated through bootstrap analysis (Felsenstein, 1985) and cophenetic procedure (Rohlf and Sokal, 1981), respectively. Amplified fragment length polymorphism data were also analyzed by AMOVA (Genealex software) to estimate the among- and within-population variation. Similarity estimates, cluster analysis, Mantel test, and principal coordinate analysis were performed by NTSYS-pc software (Rohlf, 1993); bootstrap analysis was performed by WinBoot software (Yap and Nelson, 1996).

Amplified fragment length polymorphisms were also used in a classification procedure: (i) a stepwise discriminant procedure to look for the most important predictor variables; (ii) a canonical discriminant procedure to find all possible discriminant functions and the loadings of these with the original variables; and (iii) a discriminant analysis, where the group-specific density was estimated using the kernel and k-nearest-neighbor nonparametric methods because binary-type data does not support a multivariate normal assumption (Rosenblatt, 1956; Parzen, 1962). Discriminant analysis was performed with SAS software (SAS Institute, 1999).

Geographic distance among sites of collection were computed as the shortest distance between any two points on the surface of a sphere, measured through the geographic coordinates by using the formula

Formula
where r = 6373 km (the Earth's radius), and {phi} and {lambda} are the latitudes and longitudes of point 1 and 2, respectively. The matrix of geographic distances among collection sites was compared with the matrices of genetic similarity (AFLP), genetic distances and population differentiation (SSR) by using the Mantel test (Mantel, 1967).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Diversity for Microsatellite
The gels of the five microsatellite loci belonging to the 11 natural populations of L. corniculatus, scored according to the allele doses as described above, showed a total of 68 alleles. The average number of alleles per locus was 13.6, ranging from 10 (locus TM0023) to 17 (locus TM0083). The 11 populations did not exhibit any allele that was specific to a single population. Populations did not differ in terms of number of alleles (A) and effective number of alleles (Ae) (Kruskal–Wallis test, {chi}2 = 14.7, P = 0.14, df = 10 and {chi}2 = 4.6, P = 0.92, df = 10, respectively) (Table 4 ). Low values of Ae for each population were likely to be the effect of a few alleles at high frequency. No significant differences were detected for loci polymorphism and gene diversity. Populations differed for genotype richness ({chi}2 = 19.0, P < 0.05, df = 10), although the significance was basically due to the low value of material from Preitoni, which is geographically apart from other sites.


View this table:
[in this window]
[in a new window]

  		Table 4. Mean number of alleles (A) over the five simple sequence repeat loci and effective number of alleles (Ae), genotype richness (G), percentage of polymorphic loci (Pl), and gene diversity (H) in 11 populations of Lotus corniculatus from central Italy.

 
Significant departures from the HW equilibrium were observed for almost all populations and at nearly all loci (Table 5 ). Out of a total of 55 pair comparisons, there was a significant deficiency of heterozygotes in as many as 40 (P ranging from 0.05 to 0.001); seven comparisons showed HW equilibrium, in three cases (Moggio at locus TM0083 and Preitoni at loci TM0023 and TM0083), heterozygotes were significantly higher than expected, while in five cases (Sigillo, Trecine, and Moggio at locus TM0023, and Preitoni at loci TM0212 and TM0023), there was only one fixed allele (Table 5). The overall departure from the HW equilibrium was very consistent. The high FIS values for each population, an estimate of the level of inbreeding, indicate that in almost all populations the selfing rate and the consequent low level of gene flow was high and could have promoted population diversification. Only the material from Preitoni had an excess of heterozygotes. These observations are supported by the overall high values of FST (0.274) and {rho} (0.390, P < 0.001), the latter being a parameter independent of the selfing rate and of double reduction, events that tend to increase the level of homozygotes and explain the high level of population differentiation. However, since events of double reductions would be rare in L. corniculatus (Fjellstrom et al., 2001), it is the selfing rate that may play a major role in population differentiation. This conclusion is also supported by the fact that the effect of the selfing rate was shown to be greater in autotetraploid than in diploid species (Ronfort et al., 1998).


View this table:
[in this window]
[in a new window]

  		Table 5. Observed (Ho) and expected (He) heterozygosis for the five simple sequence repeat loci and F-statistics{dagger} in 11 populations of Lotus corniculatus from central Italy.

 
The {rho} values, obtained by population pair comparisons (Table 6 ), were highly significant (P < 0.001) according to the exact tests (Raymond and Rousset, 1995), including the comparisons between material from Tuoro and Allerona ({rho} = 0.115, P = 0.0083) and from Moggio and Monte Fausola ({rho} = 0.068, P = 0.0114), which are within very short distances of one another. Analysis of variance (Weir and Cockerham, 1984), adapted to autotetraploids according to Ronfort et al. (1998), indicated that the source of variation among and within populations accounted for 45 and 55%, respectively, of the total variation, values in the range for outcrossing species (Allard, 1999).


View this table:
[in this window]
[in a new window]

  		Table 6. Estimated {rho} pairs among 11 populations of Lotus corniculatus collected in Italy. All values were statistically significant (Moggio–Monte Fausola at P < 0.05, Tuoro–Allerona and Trecine–Passignano at P < 0.01, and all the others at P < 0.001).

 
Diversity for AFLP
The AFLP analysis with four primer combinations produced 254 polymorphic products. The primer combination E-ACC/M-GCA generated the highest number of polymorphic loci (31%), whereas the primer combination E-AAG/M-GAC generated the lowest polymorphisms (11%). The mean number of polymorphic loci for all 11 populations was 56%, with Preitoni having the lowest percentage (39%) and Trecine having the highest (75%). A representative AFLP autoradiogram is shown in Fig. 2.

The UPGMA clustering based on all individuals is shown in Fig. 3 . Only bootstrap percentages higher than 80% were inserted on the nodes of the dendrogram. Except for two cases with the same AFLP profiles (Allerona04-Allerona05 and Trecine05-Trecine06), the major results are that (i) all individuals of Preitoni always clustered together and far apart from the other populations, and (ii) a somewhat large, consistent group (97% of bootstraps) included individuals of Moggio, Monte Fausola, Panicale and Passignano, thus not providing a clear-cut distinction among the other 10 populations. When the cluster analysis was performed on the MGSs, the dendrogram (not reported) showed a main cluster consisting of eight populations that was joined to individuals branches for Trecine, Canale Monterano, and Preitoni, respectively. The clustering apart of Preitoni confirmed the results of SSRs that indicated that this was the only population in HW equilibrium, characterized by few alleles, low genotype richness, and low levels of polymorphisms. The population from Canale Monterano was also quite distinct relative to the other populations examined, whereas Trecine was not different from Passignano in terms of genotype richness, number of alleles, and diversity. Unexpectedly, the closest populations were Passignano and Monte Fausola, two locations that are quite different with respect to various environmental factors, including rainfall, air temperature, soil characteristics, and altitude. The goodness-of-fit of the UPGMA clustering was confirmed by the cophenetic procedure (r = 0.991, Mantel test = 3.2829, P < 0.0020), and similar results were obtained by principal coordinate analysis (results not shown).


Figure 3
View larger version (20K):
[in this window]
[in a new window]

  		Figure 3. Unweighted pair-group method with arithmetic mean clustering of Jaccard genetic similarity coefficients of 88 individuals belonging to 11 Italian populations of Lotus corniculatus, as based on amplified fragment length polymorphism data.

 
The MGS matrix was highly and negatively correlated to the matrix of geographic distances (r = –0.89376, Mantel test = –2.9300, P = 0.0017). However, by removing Preitoni, the location most distant from the others, the correlation was not significant (r = –0.28707, Mantel test = –1.3278, P = 0.0921).

With the caution that AFLPs polymorphisms of individual plants are considered phenotypes, the AMOVA partitioning of the total variation produced a pattern not very different from that reported for the SSRs: most variation could be ascribed to differences within rather than among populations (70 vs. 30%).

SSR vs. AFLP Comparisons
Both types of markers were informative for detecting variation among and within each of the 11 populations. Simple sequence repeats provided detailed information on the genotype at a single locus and offered the possibility to monitor the variation of alleles. Amplified fragment length polymorphism generated many amplification products in a single reaction but could not provide any information concerning the allele situation. The comparison between the two types of markers was based on how the 11 populations clustered together. This comparison indicated that individuals from Preitoni did not share any allele with those from Moggio, Monte Fausola, and Canale Monterano at any of the five loci monitored, and the estimated genetic distance D was equal to {infty}. A different approach was then adopted since Kosman and Leonard (2005) showed that none of the most popular distance–similarity coefficients (simple matching, Dice, or Jaccard) could adequately assess genetic dissimilarity between homozygous and heterozygous individuals for a given locus, and proposed an index of distance ranging from 0, when two individuals are identical, to 1, for individuals not sharing any allele across all codominant loci tested. The index of Kosman and Leonard was estimated and the distance matrix used in the UPGMA clustering procedure. The resulting dendrogram showed three main clusters (Fig. 4 ): Preitoni, which clustered apart, a large cluster including three pairs of populations, Moggio–Monte Fausola, Passignano–Trecine, and Canale Monterano–Panicale, which are geographically rather close to one another and have similar environmental characteristics, and, unexpectedly, another cluster consisting of the population from Tuoro, a site very close to Passignano and Trecine, along with populations from Allerona, Moggio and Sigillo, three sites with similar environmental conditions. This type of clustering was exactly the same as that based on Nei's genetic distance with 10 populations (dendrogram not shown), the two matrices being highly and significantly correlated (r = 0.81440, Mantel t = 5.2649, P < 0.001). The matrix of genetic distances (according to Kosman and Leonard index), and the among-sites geographic distance matrix were also significantly correlated, even when Preitoni was removed (r = 0.29547, Mantel test = 2.1306, P = 0.017).


Figure 4
View larger version (11K):
[in this window]
[in a new window]

  		Figure 4. Unweighted pair-group method with arithmetic mean clustering of genetic distance coefficients (Kosman and Leonard) of 11 Italian populations of Lotus corniculatus, as based on simple sequence repeat data.

 
Discriminating Populations
The stepwise discriminant procedure reduced the number of predictor variables from 254 to 80; only the top 20 of these were used to find the discriminant functions. As many as 10 discriminant functions were found (Table 7 ), the first nine of which were statistically significant. The Wilks' lambda of the first function was 7.8 x 10–7, which was highly significant and could explain as much as 63% of the among-population variation. However, after removing the first function, a strong association still existed between populations and marker predictors, as well as after removing functions 1 and 2 (Table 7). The first three functions together accounted for a total of 89% of the between-population variation (63, 13, and 12%, respectively).


View this table:
[in this window]
[in a new window]

  		Table 7. Discriminant functions, canonical correlation, eigenvalues, cumulative percentage of variation explained by each function, and F test.

 
The first discriminant function separated materials from Preitoni and Canale Monterano from the other nine populations, whereas the second function was able to discriminate between the former two (Fig. 5 ). The loadings between original variables and the first three discriminant functions, as shown in Table 8 , indicated that the best predictors for distinguishing between Preitoni and Canale Monterano from the other populations (Function 1) was the fragment MGCPCT_55, almost absent in the former group (present only in one plant of Canale Monterano) and present in all individuals of the other nine populations. The second discriminant function was positively correlated with MGAC-EAAG_07 and MGCPCT_36, and negatively with MGCPCT_60. The presence or absence of these three fragments clearly distinguished Preitoni from Canale Monterano and some other populations. MGCA-EACC_01 was positively correlated with the third function and allowed for the separation of three populations that clustered closely together (Trecine, Sigillo, and Tuoro).


Figure 5
View larger version (11K):
[in this window]
[in a new window]

  		Figure 5. Centroids of the 11 Italian populations of Lotus corniculatus, plotted according to the first two canonical discriminant functions.

 

View this table:
[in this window]
[in a new window]

  		Table 8. Mahalanobis distances among 11 populations of Lotus corniculatus collected in Italy estimated through amplified fragment length polymorphism molecular markers.

 
The matrix of the Mahalanobis distances among the 11 populations (Table 9 ) confirmed basically the graphical representation of the centroids (Fig. 5). Almost all distances were highly significant (P < 0.001), even the lowest values found between Monte Fausola–Passignano and Moggio–Panicale (P < 0.05). The Mahalanobis distance and the geographic distance matrices were significantly correlated (r = 0.78086, Mantel test = 2.8146, P = 0.0080).


View this table:
[in this window]
[in a new window]

  		Table 9. Correlation between the first three canonical discriminant functions and the amplified fragment length polymorphism fragments for 11 populations of Lotus corniculatus collected in Italy.

 
The nonparametric k-nearest neighbor method (Rosenblatt, 1956; Parzen, 1962) permitted the reclassification into their original population of 87 out of 88 individuals. Only plant no. 2 belonging to Sigillo was erroneously classified into Trecine.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Lotus corniculatus is an important forage species with a wide geographic and climatic distribution, which is the result of a large level of genetic diversity. The significance of the present study is primarily the validation of the diversity already assessed in a field evaluation of natural populations of L. corniculatus collected in Italy in 1999 (Pagnotta et al., 2003) following an ecogeographic survey (Russi et al., 2003). The survey clearly confirmed that birdsfoot trefoil is one of the most widespread perennial legumes present in natural pastures over a range of altitudes, soil pH, and rainfall. Regardless of this, in Italy birdsfoot trefoil has not yet been taken into consideration by breeders (the varieties registered in the National list are few) as well as by farmers (certified seeds in the last 5 yr averaged 54 t per year vs. 4700 t of alfalfa and 15,800 t of all forage legume species).

The 11 populations studied here, although not characterized by unique alleles, showed combinations that were useful for distinctiveness and at the same time provided a picture of great diversification. Population diversification was directly related to geographical distances. In most populations the excess of homozygotes at the five loci examined was the result of a consistent, unexpected, selfing rate, combined with a limited gene flow. However, inbreeding, limited gene flow, and specific selection pressures may lead to genetic diversification that could ultimately confer a successful adaptation strategy to a population and allow a superior exploitation of ecological niches.

This type of germplasm is highly valuable to breeding programs. Short-term programs devised to target traits in specific varieties, such as adaptation to low soil pH conditions, are therefore likely to be successful. Benefits to longer-term programs by breeding new varieties adapted to a wider range of agronomic conditions are also possible by exploiting the heterosis arising from intercrossing genotypes selected among genetically distant populations.


   ACKNOWLEDGMENTS
 
We thank Professor Mario Pagnotta for providing us with the plant material and Paola Donate for the assistance with Cross Checker Software.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.

Received for publication May 29, 2007.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 





This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Agricola
Right arrow Articles by Savo Sardaro, M. L.
Right arrow Articles by Porceddu, E.
Related Collections
Right arrow Other Forage Crops
Right arrow Plant Genetic Resources
Right arrow Crop Genetics

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