| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Systematics and Phytogeography |
2Departamento de Ciencias Biológicas, Facultad de Ciencias Exactas y Naturales, Universidad de Buenos Aires, 1428 Buenos Aires, Argentina; 3Instituto de Botánica Darwinion, Labardén 200, 1642 San Isidro, Buenos Aires, Argentina
Received for publication June 12, 2001. Accepted for publication December 20, 2001.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Acacia • Argentina • differentiation • Fabaceae • genetic variability • isozymes • morphometry • RAPD
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Karlin et al., 1997
). Their wood is hard, heavy, and durable. It is used in carpentry, chassis, parquet flooring, tie beams, and the like (Cialdella, 1984
; Karlin et al., 1997
). Gum Arabic is obtained from some Acacia species (Martínez et al., 1996
; León de Pinto et al., 1998
), and Acacia flowers are important for the perfume industry (Dimitri and Biloni, 1973
; Allen and Allen, 1981
). Some species are grown as ornamental and/or shade trees, while armed shrubs are suitable for fences.
The genus Acacia (Fabaceae, Mimosoideae) includes over 1200 species of pantropical distribution (Guinet and Vassal, 1978
; Ross, 1981
). They inhabit tropical and subtropical regions of the Americas, Australia, Africa, and southern Asia. In northern and central Argentina Acacia is represented by 21 species in three series: Filicinae, Gummiferae, and Vulgares (Cialdella, 1984, 1997
). All the Argentinean species are woody, and most are trees or shrubs 2–6 m high, although some species reach 20 m. The branches are generally armed—the Gummiferae species have thorny stipules and the Vulgares species have stings. The Filicinae species, however, lack such structures.
Bentham (1842, 1875)
subdivided the genus into six series: Botrycephalae, Filicinae, Gummiferae, Phyllodineae, Pulchellae, and Vulgares. Later, Vassal (1972)
, on the basis of Bentham's system, proposed three subgenera: Acacia (Gummiferae), Aculeiferum (Vulgares), and Heterophyllum (Botrycephalae, Phyllodineae, and Pulchellae). This classification is incomplete for Argentina, because one Argentinean species belonging to Filicinae Benth. cannot be included in Vassal's proposed system. For that reason, Cialdella (1984)
, following Bentham's subdivision, proposed the placement of Argentinean species in three series: Filicinae, Gummiferae, and Vulgares.
Although several authors have studied the taxonomy of Acacia using morphological characters (Bentham, 1842, 1875
; Vassal, 1972
; Guinet and Vassal, 1978
; Cialdella, 1984, 1997
; Pedley, 1986
), in the last ten years some have used biochemical and molecular markers instead (Playford, Appels, and Baum, 1992
; Bukhari, 1997a, b
; Clarke, Downi, and Seigler, 2000
). Biochemical and molecular studies have been conducted on African and Australian Acacia species to provide markers useful for plant breeding and conservation programs (Moran, Muona, and Bell, 1989a, b
; Muona, Moran, and Bell, 1991
; Joly et al., 1992
; Playford, Appels, and Baum, 1992
; Sedgley et al., 1992
; Playford, Bell, and Moran, 1993
; Fagg, Barnes, and Marunda, 1997
; McGranahan et al., 1997
; Butcher, Moran, and Perkins, 1998
). However, no population genetic studies have been carried out so far on Argentinean species of Acacia.
Isozyme electrophoresis and random amplified polymorphic DNA (RAPD) analysis are broadly used in plant population genetic studies (Soltis and Soltis, 1990
; Avise, 1994
; Soltis, Soltis, and Doyle, 1998
; Hollingsworth, Bateman and Gornall, 1999
). Mainly, RAPD has allowed the resolution of complex taxonomic relationships (Voigt, Schleiler, and Brückner, 1995
; Comincini et al., 1996
; Cottrell, Forrest, and White, 1997
; Wolff and Richards, 1999
). Likewise, morphological characters constitute basic information for plant systematics. Therefore, in the present study we used morphometric, RAPD, and isozyme electrophoretic approaches to characterize four Argentinean species of Acacia so that we could evaluate their phenetic relationships.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
and Saidman and Vilardi (1993)
. Approximately 50–80 seed pods were collected from at least ten mother trees that were separated from each other by more than 50 m.
|
|
). It blooms from September to November, and fruits ripen from March to June. We studied two populations of this species from Campo Quijano and Huilla Catina (Fig. 1, Table 1).
Acacia caven (Molina) Molina
This species comprises trees and shrubs 2–5 m high. Their branches have small, thin, conical thorny stipules, and their flowers are arranged in globose inflorescences. Commonly known as "aromita," "cavén," "churqui," "espinillo," or "espino," this species is distributed in northern and central Argentina, Chile, Paraguay, Bolivia, Brazil, and Uruguay (Cialdella 1984, 1997
). It blooms in August and September, and fruits ripen between January and April. Cialdella (1984, 1997)
described four different varieties for this species on the basis of fruit morphology. We studied two populations of A. caven var. caven, a subspecies characterized by ellipsoid or spherical indehiscent fruits (Cialdella, 1984
), from Campo Quijano and La Caldera (Fig. 1, Table 1).
Acacia macracantha Willd
The shrubs and trees of this species are 2–4 m high. Their branches have compressed thorny stipules with one or several longitudinal ribs. The globose inflorescences and fruits are very similar to those of A. aroma. Its local name is "tusca." The species inhabits western South America, from Ecuador to Bolivia, Paraguay, and northern Argentina. It blooms in October, and fruits ripen in February. We sampled one population of this species at Campo Quijano (Fig. 1, Table 1).
Acacia furcatispina Burkart
This species comprises shrubs and trees 1.5–4 m high. The branches have small branchlets (5–18 mm long) that end in two divergent stings. The species can be easily identified by this trait; flowers and fruits don't have to be present. The inflorescences are globose, and the fruit is a typical legume. The local names are "garabato" and "teatín." The species inhabits sandy soils in northern and central Argentina, and they can also be found in Paraguay and Bolivia. It blooms in October and their fruits ripen in January and February. We studied one population of this species from Cabra Corral (Fig. 1, Table 1).
Isozyme methods and data analysis
Seven isozyme systems were assayed: alcohol dehydrogenase (ADH), (Enzyme Commission) E.C. 1.1.1.1; glutamate oxalacetate transaminase (GOT), E.C. 2.6.1.1; isocitric dehydrogenase (IDH), E.C. 1.1.1.42; peroxidases (PRX), E.C. 1.11.1.7; shikimic dehydrogenase (SKD), E.C. 1.1.1.25; 6-phosphogluconic dehydrogenase (6-PGD), E.C. 1.1.1.43; and superoxide dismutase (SOD), E.C. 1.15.1.1 (IUPAC-IUB, 1979
). We used horizontal electrophoresis on 5% (GOT) or 7% (remaining systems) polyacrylamide gels. Buffers and electrophoretic conditions are described elsewhere (Saidman, 1985, 1986
; Saidman and Vilardi, 1987
; Casiva, 2001
). Homogenates were made from 6–7-d-old cotyledons for all systems except ADH. For that system, we analyzed 12–18-h-old seedlings (the age of cotyledons is measured from the moment the seeds were placed into a wet chamber to stimulate germination). Isozyme loci were named according to decreasing mobility, relative to bromophenol, of their corresponding bands. Superscript numbers were used to name allelic variants according to their relative mobility.
Because seeds from controlled crosses are not available, we used open-pollinated progenies to infer the genetic control of each isozyme system. Bands that did not conform with Mendelian segregation patterns were omitted from the analysis. Allelic and genotypic frequencies were estimated for each locus in each population. Departures of observed frequencies from Hardy-Weinberg expectations were evaluated with a chi-square test.
Genetic variability was estimated with the following parameters: percentage of polymorphic loci (Pp, 0.95 criterion), unbiased expected heterozygosity (He; Nei, 1978
), and mean number of alleles per locus (A). Population structure was analyzed by F statistics (Wright, 1951
), and indirect estimations of gene flow (Nm) were obtained from the differentiation among populations (FST) according to the relationship, Nm = (1 – FST)/4FST. Nei's (1978)
genetic distances among populations were represented in a phenogram constructed with the unweighted pair group method with arithmetic mean (UPGMA) (Sneath and Sokal, 1973
). Distortion between distance matrix and phenogram was evaluated by the cophenetic correlation (Sokal and Rohlf, 1962
). All previously mentioned analyses were performed using the computational program Biosys-1, version 1.7 (Swofford and Selander, 1981
).
By the bootstrap method 100 derived matrices were obtained with the program Seqboot, in the PHYLIP package (Felsenstein, 1993
). The matrices were used to create a consensus tree (using the programs Gendist, Neighbour, and Consense in the same package) to test the reliability of the phenogram branches. With the STATISTICA program (StatSoft, 1995
), we conducted a principal component analysis (PCA) to identify the most important isozyme systems for differentiating populations and species (Crisci and Armengol, 1983
).
RAPD methods and data analysis
Genomic DNA was extracted from 7-d-old cotyledons using a modification of the method of Dellaporta, Wood, and Hicks (1983)
. Both cotyledons from each individual were ground to powder in porcelain mortars with liquid air and suspended in 500 µL buffer (100 mmol/L Tris, 50 mmol/L EDTA, 500 mmol/L NaCl) and 1.4% SDS. The suspension was incubated in a water bath for 10 min at 65°C. After incubation 175 µL potassium acetate (5 mol/L, pH 5.2) was added, followed by ice cooling for 20 min. Homogenates were centrifuged at 16 000 x g for 15 min, and supernatant was recovered in fresh tubes. One volume of 25 : 24 : 1 phenol : chloroform : isoamyl alcohol was added, the mixture was centrifuged at 16 000 x g for 5 min, and the supernatant was carefully removed. Nucleic acids were precipitated by the addition of 50 µL sodium acetate (3 mol/L, pH 5.2) and 1 mL ethanol and were stored overnight at –20°C. A new centrifugation at 16 000 x g was done for 10 min. The pellet was washed with 70% ethanol, airdried, and redissolved in 100 µL sterile distilled water. RNA was removed using 20 µg RNAse, followed by incubation at 37°C for 1 h. DNA was reprecipitated with 50 µL sodium acetate (3 mol/L, pH 5.2) and 1 mL ethanol, followed by centrifugation at 16 000 x g for 10 min. The pellet was washed with 70% ethanol, air dried, redissolved in 100 µL sterile distilled water, and stored at 4°C.
The polymerase chain reaction (PCR) was carried out in 50 µL final volume using 25–30 ng DNA, 5 µL dNTP mix (0.1 mmol/L dATP, dCTP, dGTP, dTTP), 3 ng primer, 3 mmol/L MgCl2, and 2.5 units of Taq Polymerase in 5 µL thermophilic DNA buffer (Promega, Madison, Wisconsin, USA). The mixture was placed in a thermocycler with the following program: 1 cycle of 94°C for 1 min, 40 cycles of 94°C for 1 min, and 36°C for 1 min, and 72°C for 2 min. A final extension step at 72°C for 6 min was performed to ensure complete extension. Ten arbitrary primers (Biodynamics oligonucleotidos, Buenos Aires, Argentina) were screened in a small number of individuals. Two of these (B07, 5'-AGATCGAGCC-3'; B09, 5'-ATGGCTCAGC-3') were chosen because they showed clearly reproducible banding patterns. Amplification products were resolved on 1.4% agarose gels stained with ethidium bromide. To test the reliability of PCR products, we routinely used three controls: one without primer, a second with no DNA polymerase, and a third with no genomic DNA. No amplification occurred in any of these controls.
The RAPD data were scored as band present (1) or absent (0). Using the program RAPDBIO in the package RAPD (Black, 1996
), we converted these data into allelic frequencies according to the method proposed by Lynch and Milligan (1994)
. From the allelic frequency matrix, genetic variability was quantified with the same parameters described above for the isozyme data, Nei's (1978)
genetic distances were also estimated and represented in a UPGMA phenogram (Sneath and Sokal, 1973
). The reliability of this phenogram was evaluated by bootstrapping the data matrix. Relationships between populations and species were also evaluated by a PCA analysis of allelic frequencies using the program STATISTICA (StatSoft, 1995
).
Because Lynch and Milligan's (1994)
method assumes allelic frequencies are in Hardy-Weinberg equilibrium, which cannot be proved, we used an alternative method (one without any assumptions) to evaluate phenetic relationships. Similarities between individuals were estimated on the basis only of the number of shared bands, using Nei and Li's (1985)
index. From similarities among individuals, a matrix of similarities among populations was obtained using the expression of Bardakci and Skibinsi (1994)
. These similarities were converted into distances by substracting the corresponding value from 1. Then a UPGMA phenogram was created with the program STATISTICA and compared with that obtained from allelic frequency estimates.
Morphometric methods and data analysis
We studied 15 individuals of A. aroma, 12 of A. macracantha, 15 of A. caven, and 15 of A. furcatispina from the herbarium of the Instituto de Botánica Darwinion. (Voucher details: Arenas 2268(SI), Burkart 13122(SI), Burkart 21774(SI), Burkart 29343(SI), Burkart 7656(SI), Cabrera 21065(SI), Cabrera 24041(SI), Cabrera 25535(SI), Cabrera 27285(SI), Cabrera 29710(SI), Cabrera 27983(SI), Cabrera 31361(SI), Cabrera 31899(SI), Cabrera 32087(SI), Cabrera 34351(SI), Cabrera 3789(SI), Cabrera 4550(SI), Calabretto 83(SI), Castellanos 394(SI), Castellanos 1427(SI), De Huajardo 4630(SI), Deginani 39(SI), Deginani 42610(SI), Hunziker 6138(SI), Kiesling 1252(SI), Kiesling 3021(SI), Kiesling 3022(SI), Kiesling 4899(SI), Kiesling 5843(SI), Krapovickar 1132(SI), Krapovickar 13090(SI), Lautre 5381(SI), Leuenberger 3592(SI), Martinez 9583(SI), Melillo 2358(SI), Meyer 2041(SI), Nicora 3151(SI), Nicora 8454(SI), NN 1120(SI), NN 16968(SI), Pedersen 3957, Pedersen 10785(SI), Pissinini 2020(SI), Rodrigo 2909(SI), Rodriguez 7(SI), Rodriguez 1195(SI), Rotman 146(SI), Rotman 4181(SI), Ruiz Leal 9199(SI), Schreiter 68587(SI), Urtubey 13(SI), Venturi 1087(SI), Venturi 144(SI), Venturi 5376(SI), Venturi 908(SI), Zuloaga 2647(SI).) The available material covers most of the geographic range of each species. To evaluate morphological differentiation of the four species, we measured nine morphometric traits: (1) leaf length, (2) pinna length, (3) number of pinnae per leaf, (4) thorny stipule length, (5) inflorescence diameter, (6) peduncle length, (7) seed pod shape, (8) fruit length, and (9) fruit width. For quantitative traits the statistical analysis was based on the actual measure (in centimeters). Seed pod shape was coded as follows: loment = 1, globose = 2, typical legume = 3.
The differences among the studied species were tested by Kruskal-Wallis's (1952)
method, using the program STATISTICA. Phenetic relationships among species were evaluated by cluster analysis from the matrix of average values of morphometric traits. Each species was considered an operational taxonomic unit (OTU). The relationships among the species were estimated by the Pearson correlation coefficient (r). Distance values (D) were defined as follows: D = 1 – r. The distance matrix was represented in a phenogram by the UPGMA clustering method (Sneath and Sokal, 1973
), and the cophenetic correlation (Sokal and Rohlf, 1962
) was also calculated. To identify those morphological characters that were most important in the differentiation of species, we analyzed the principal components with the STATISTICA program.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Five monomorphic loci in the peroxidases (PRX) system were identified in A. furcatispina. Only two (Prx-3 and Prx-4) were present in A. aroma, A. caven, and A. macracantha. The superoxide dismutase (SOD) patterns can be explained if one assumes that four loci occur. Three of them (Sod-1, Sod-3, and Sod-4) were present in the species belonging to the series Gummiferae and proved to be polymorphic. The observed patterns allow one to infer a dimmer structure for this enzyme. In A. furcatispina all four loci were monomorphic. Sod-1 and Sod-2 showed nonallelic interaction, producing a nonsegregating pattern of three bands with intensities 1 : 2 : 1.
The shikimic dehydrogenase (SKD) system was coded by one locus. Its allelic variants allowed for differentiation of A. furcatispina from the other three species. Two monomorphic loci for 6-phosphogluconic dehydrogenase (6-PGD) were found in A. furcatispina. The Gummiferae species showed only one polymorphic locus, which apparently is not homologous to those of A. furcatispina. For the isocitric dehydrogenase (IDH) system two different monomorphic loci were found, one in the Gummiferae species and the other in A. furcatispina. The zymograms of the alcohol dehydrogenase (ADH) system exhibited two polymorphic zones in the Gummiferae species that seemed to be the expression of two loci (Adh-1 and Adh-2). The enzyme would have a dimmer structure, and the products of interaction between and within loci were observed. A. furcatispina exhibited only one monomorphic locus (Adh-2). Glutamate oxalacetate transaminase (GOT) was coded by four monomorphic loci in the Gummiferae species. Only two, Got-1 and Got-2, were present in A. furcatispina, and Got-2 was polymorphic.
The studied populations showed differences in allelic frequencies (Table 2). Estimates of genetic variability in different populations of species of Gummiferae were similar (i.e., their confidence intervals overlapped). Acacia furcatispina showed significantly lower values for all variability estimates (Table 3).
|
|
Acacia furcatispina showed important qualitative differences with respect to the Gummiferae species and had several diagnostic loci. For this reason, calculation of FST was restricted to the Gummiferae populations. When every population in this series was included, all FST estimates were highly significant (Table 4). In intraspecific comparisons FST estimates were significant for only some of the loci. Surprisingly, the comparison between A. aroma and A. macracantha yielded an average FST value similar to those resulting from intraspecific comparison and even lower than that obtained from the comparison between populations of A. caven.
|
From the Nei's (1978)
genetic distance matrix (Table 5), we constructed a UPGMA phenogram representative of phenetic relationships among populations (Fig. 2A). Conspecific populations are clustered, and A. aroma is tightly linked to A. macracantha. Acacia furcatispina is far away from the group of Gummiferae species (Fig. 2A). The high cophenetic correlation (0.99) indicates no significant distortion between data matrix and the phenogram.
|
|
Random amplified polymorphic DNA analysis
The two studied primers showed a total of 34 RAPD loci considering all six analyzed populations and allowed us to distinguish most species. Acacia aroma and A. macracantha did not have diagnostic loci that distinguish one species from the other, but both species differed clearly from A. furcatispina and A. caven. No diagnostic bands among conspecific populations were found for the analyzed primers.
Fourteen bands of diverse molecular mass (229–1353 bp) were analyzed for primer B07 (Fig. 3). Acacia macracantha and A. aroma showed seven bands that clearly differed from A. caven; A. furcatispina, however, did not show any amplification product for this primer. The absence of any band was confirmed with several amplification repetitions; we also varied the concentration of DNA, MgCl2, buffer, enzyme, and/or primer.
|
Following the methodology proposed by Lynch and Milligan (1994)
, we used the data matrix of presence (1) and absence (0) of bands to estimate allelic frequencies (data not shown) and genetic variability parameters (Table 6). The mean number of alleles per locus (A) varied from 1.1 to 1.3, and no significant differences occurred among populations or among species (Table 6). The percentage of polymorphic loci (Pp) varied between 8.8 and 29.4, with an average of 19.1 over all populations. Estimates of Pp and A were lower than those calculated from isozyme data. The heterozygosity was lower, too, and A. furcatispina presented the lowest values of genetic variability (Table 6).
|
|
|
similarities were converted into distance values among populations (Bardakci and Skibinsi, 1994
) and represented in a UPGMA phenogram (Fig. 5), A. furcatispina was separated from the Gummiferae species, in agreement with taxonomic criteria. Acacia aroma and A. macracantha showed a very high similarity; in fact, A. aroma from Campo Quijano was closer to A. macracantha from the same location than to the other A. aroma population (Fig. 5).
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
) and is similar to the average value (0.211) for other tropical angiosperms, such as Eucalyptus (0.174) and Casuarina (0.2119) (Moran, 1992
).
All the estimated genetic variability parameters were higher in the species of Gummiferae than in A. furcatispina. If we compare the estimates obtained for the Gummiferae species with those of other tropical angiosperms, the differences are even higher. Playford, Bell, and Moran (1993)
also found high levels of genetic diversity (0.208) in A. melanoxylon populations in association with a great genetic differentiation among geographic areas. High values of heterozygosity (H) were also found in African populations of A. albida (Joly et al., 1992
), with values ranging from 0.260 to 0.442. However, estimated heterozygosity values were lower in Australian species of Acacia—for example, A. auriculiformis and A. crassicarpa (0.146 and 0.141, respectively; Moran, Muona, and Bell, 1989b
), A. aulacocarpa (0.112; McGranahan et al., 1997
), and A. mangium, which had the lowest value (0.017; Moran, Muona, and Bell, 1989b
).
High levels of genetic variability, such as those observed in the present study, may be related to mating system and geographic distribution. Species with crossed fecundation and a wide geographic range have higher levels of genetic diversity than do selfer and endemic species. Likewise, species whose seeds are dispersed by animal ingestion or by wind maintain high levels of within-population genetic variability (Hamrick and Godt, 1990
). Both populations of A. caven studied here showed high levels of genetic variability, and animals are known to play an important role in the dispersal of their seeds (Gutiérrez and Armesto, 1981
). The fruits of A. aroma and A. macracantha are loments (Cialdella, 1984
), and although the seed-dispersional mechanism is unknown, animals likely play an important role. In A. furcatispina, which had the lowest genetic variability values in our study, fruits are typical legumes (Cialdella, 1984
) that open to liberate the seeds and, although the means of seed dispersal is not known, H values are close to those recorded by Hamrick and Godt (1990)
for species with explosive or gravity distribution (H = 0.062 and 0.101, respectively).
Studies on the reproductive biology of several species of Acacia indicate that acacias are open-pollinated by insects and are generally self-incompatible (Bernhardt, Kenrick, and Knox, 1984
; Kenrick, Kaul, and Williams, 1986
; Moran, Muona, and Bell, 1989b
; Muona, Moran, and Bell, 1991
). In some cases acacias form hybrid complexes (Ali and Qaiser, 1980
; Sedgley et al., 1992
).
The low level of genetic variability observed in A. furcatispina coincides with that observed in selfer species (H = 0.074) by Hamrick and Godt (1990)
and is close to that of mixed-animal breeding systems (0.090). In Gummiferae species genetic variability is higher than in outcrossing animal breeding systems (0.124) or outcrossing wind breeding systems (0.148).
The reproductive biology of South American Acacia species is not well known. The low variability of A. furcatispina might be attributed to selfing. However, Bernhardt, Kenrick, and Knox (1984)
suggested that this species would be self-incompatible. An alternative hypothesis for low variability would be limited seed dispersal. In fact, A. furcatispina legumes are dehiscent, and seeds apparently fall down and germinate close to the mother plants. Such behavior favors biparental inbreeding.
Acacia aroma and A. macracantha are frequently visited by insects, mainly bees, which could be important in pollen dispersal (Zapata and Arroyo, 1978
; Bernhardt, Kenrick, and Knox, 1984
). Therefore, these species would have crossed fecundation, with high levels of genetic variability.
To analyze population structure, we calculated F statistics (Wright, 1951
). In most cases FIS was not significant, suggesting that populations are panmictic. The fixation index (FIS) was significant only for some of the analyzed isozyme loci, but both positive and negative values were recorded. Moreover, the observed bias from Hardy-Weinberg expectations seem not to follow any defined pattern, because the same loci showed different behavior in different populations.
The FST index represents the correlation of two uniting gametes within a subpopulation with respect to the whole population (Wright, 1951
), and it may be used as a measure of the degree of differentiation among related populations. Because A. furcatispina showed low homology at isozyme loci relative to the other three species, this index was used only to estimate the differentiation among species in the series Gummiferae. The differentiation observed among all populations was highly significant for all polymorphic loci. Adh-1 was the locus with the highest contribution to FST, showing different allelic variants in A. caven than in A. aroma and A. macracantha. Skd-1 was the polymorphic locus with the lowest differentiation, showing the same allelic variants in all populations, with differences limited to allelic frequencies. The differentiation among conspecific populations was significant for only some of the polymorphic loci analyzed. The differentiation between populations of A. aroma and A. macracantha was lower than that expected for distinct species, with FST estimates similar to those observed between conspecific populations.
Indirect estimates of gene flow (Nm) among conspecific populations were higher than one migrant per generation. By contrast, Nm among all populations was lower. This effect is the consequence of the inclusion of A. caven populations, because Nm among populations of A. aroma and A. macracantha is similar to the values observed for conspecific populations. The latter result is compatible with genetic exchange among these species, either in the present or in the relatively recent past.
The phenogram based on Nei's (1978)
genetic distances (D) clustered conspecific populations. The distances observed between populations of the same species ranged from 0.02 to 0.01. In other species of Acacia intraspecific genetic distances showed great variation, from 0.009 in A. mangium (Moran, Muona, and Bell, 1989a
) to extremely high values among widely distributed populations of A. melanoxylon (0.34; Playford, Bell, and Moran, 1993
), A. aulacocarpa (0.26; McGranahan et al., 1997
), and A. albida (0.27; Joly et al., 1992
).
In the present study A. furcatispina was the most differentiated species, as would be expected given that it belongs to a different series. Among Gummiferae species A. aroma and A. macracantha are more similar to each other, while A. caven is more differentiated. The PCA agreed with the cluster analysis because the same groups of populations and species were formed. The genetic distances between A. macracantha and A. aroma populations were similar to those observed between conspecific A. caven populations. The average genetic distance among all Gummiferae populations was low and was comparable to those for local populations of the same species (Ayala et al., 1974
) and for populations of species of section Algarobia of genus Prosopis (Fabaceae) that belong to a syngameon (Saidman and Vilardi, 1987
).
Random amplified poymorphic DNA analysis
One of the most significant disadvantages of RAPD is its dominance, which results in less accurate estimates of allelic frequency than those obtained by codominant markers, such as isozymes. Therefore, more RAPD loci should be studied in comparison with isozymes (Lynch and Milligan, 1994
). In this study, the number of RAPD loci was about 1.5 times higher than the number of isozyme loci (34 and 21, respectively). Genetic variability estimates based on RAPD markers were lower than those based on isozymes. However, in both cases the lowest variability was observed in A. furcatispina, and the variability among Gummiferae species did not differ significantly.
In some cases genetic variability estimated by both techniques are similar (Saidman et al., 1998
; Mamuris, Stamatis, and Triantaphyllidis, 1999
; Bessega, Saidman, and Vilardi, 2000
), while in others there is discrepancy. RAPD genetic variability is usually lower than isozyme variability (Jenczenwski, Prosperi, and Ronfort, 1999
; Sun et al., 1999
). These differences could be due to the characteristics of each marker. The isozyme technique is able to detect variation only in genic functional products, while RAPD can detect variation in the whole genome, in both coding and noncoding regions. Therefore, isozyme loci could be more affected by natural selection than RAPD loci, and the evolutionary rate of these markers would be different.
The RAPD differentiation among populations is usually higher than isozymal differentiation. Consequently, RAPD is useful for studying relationships among isolated geographic populations (Li, 2000
) as well as among highly related ones that cannot be accurately differentiated with isozyme methods (Sun et al., 1999
; Bessega, Saidman, and Vilardi, 2000
). The PCA from RAPD data in this work allowed us to differentiate three groups: (1) A. macracantha and A. aroma populations, (2) A. caven populations, and (3) A. furcatispina, which is far from the other groups. This result is roughly consistent with that from the PCA based on isozymes. However, the UPGMA phenograms from RAPD and isozyme Nei's (1978)
distance matrices are different. The isozyme data agree with taxonomic criteria and separate the species of Gummiferae from A. furcatispina. This pattern was not observed in the RAPD phenogram, in which A. furcatispina (Vulgares) is associated with A. aroma and A. macracantha (Gummiferae). In both phenograms, however, A. aroma and A. macracantha are clustered together. Discrepancies among the topologies of phenograms obtained from different markers have also been observed in other plant genera (Saidman et al., 1998
; Sun et al., 1999
; Bessega, Saidman, and Vilardi, 2000
; Li, 2000
). In the present study the cause might be the assumptions of Lynch and Milligan's (1994)
method for estimating allelic frequencies from RAPD data. Indeed, when the RAPD-based phenogram was obtained from band presence/absence, instead of estimated allelic frequencies data, the result is completely consistent with that from isozyme data. The main difference between the two methods for evaluating similarities from RAPD data is that Lynch and Milligan's (1994)
method assumes Hardy-Weinberg equilibrium and gives the null and the active allelic frequencies the same loadings in the estimation of final distance. Because the absence of an RAPD band may be the consequence of different gene mutations, genetic similarities between populations in which the same band is absent are likely to be overestimated. Nei and Li's (1985)
index is based on shared bands and appears to be more accurate in estimating relative similarities, especially in cases like A. furcatispina in which many bands are absent.
Morphometric analysis
A morphometric characterization of these four Acacia species has been previously done (Cialdella, 1984, 1997
), although it was not used to evaluate phenetic relationships among them. In the present study we used the correlations between species for morphometric measures to obtain a phenogram representative of phenetic similarities among species. The results of cluster and PCA analyses agree with those obtained from molecular data. Acacia aroma and A. macracantha show the highest similarity, while A. furcatispina is the most differentiated species.
The most important characters in the differentiation of these species were thorny stipule length, seed pod shape, and number of pinnae per leaf. Most Acacia species are armed (Cialdella, 1984
). In Gummiferae species stipules develop and then become thorny, whereas A. furcatispina has divergent stings on a small branchlet. The importance of this trait is evidenced by its high loading in all three principal components.
The fruit is one of the most important characters for the identification of the species of Gummiferae, but it shows little variation among species within the series Vulgares (Cialdella, 1984
). The characters fruit length and fruit width contribute significantly to PC1 and PC2, and these axes differentiate the Gummiferae species (A. aroma, A. macracantha, and A. caven) from one another and from A. furcatispina. Another character with great contribution to PC3 was number of pinnae per leaf; it allowed for differentiation of A. furcatispina and A. caven from A. aroma. Acacia furcatispina usually has small leaves, jugated 2–6 times, as does A. caven (jugated 3–15 times), whereas the leaves of A. aroma and A. macracantha are bigger (jugated 10–25 and 3–30 times, respectively) (Cialdella, 1984
). The difference between A. aroma and A. macracantha might be overestimated owing to the high variability in A. macracantha. All the morphological characters we analyzed showed highly significant differences among the species. These results suggest a good morphological differentiation. However, A. aroma and A. macracantha are very similar to each other. The only character that differentiates these species is the thorny stipule, which is longer in A. macracantha.
The taxonomic status of these species is matter of discussion. Cialdella (1984)
analyzed material from A. aroma and A. macracantha collected in Argentina and several other South American countries. She found great similarities between these species but noted that the shape of the thorny stipules allowed one to identify each species. In A. macracantha the thorny stipules are laterally compressed on the plant and rhomboidal in cross section, while in A. aroma they are conical on the plant and circular in cross section, although the determination of the stipule shape is sometimes difficult (Cialdella, 1984
). However, Ebinger, Seigler, and Clarke (2000)
state that A. aroma and A. macracantha may be distinguished by several characteristics other than stipule shape, although their distribution map of A. macracantha does not include Argentina. According to these authors, A. aroma has fruits that are constricted between the seeds, peduncles more than 22 mm long, and sessile petiolar glands. In contrast, A. macracantha has nonconstricted fruits, peduncles less than 25 mm long, and stalked petiolar glands. An important disagreement between Cialdella (1984)
and Ebinger, Seigler, and Clarke (2000)
is that the latter described the stipular spines of A. macracantha as terete to oval in cross section and those of A. aroma as terete to angular or rarely flattened in cross section. The characteristics used by Ebinger, Seigler, and Clarke (2000)
cannot be used in Argentinean material because in our study we observed specimens of both species in which fruits on the same branch varied from conspicuously constricted to not constricted. Besides, the range of peduncle length in both species overlaps, and A. aroma and A. macracantha have sessile petiolar glands. The most important difference between A. aroma and A. macracantha was thorny stipule length.
The discrepancies between these results call into question the relationships between A. aroma and A. macracantha and suggest the need for a taxonomic revision, one that compares the materials analyzed by Cialdella (1984)
, Ebinger, Seigler, and Clarke (2000)
, and us.
The results of our morphometric analysis are consistent with the findings of Cialdella (1984)
in showing a high similarity between A. aroma and A. macracantha and indicating that thorny stipules are an important diagnostic character, as much for their shape as for their size. Moreover, genetic similarity between these species is also high. The genetic distances observed from isozyme data correspond to values recorded in other genera for conspecific populations (Ayala et al., 1974
). Besides, RAPD data do not allow one to differentiate A. macracantha from A. aroma populations. One hypothesis for such a high genetic and morphological similarity is that they are varieties of a single species. However, they are sympatric in wide areas of their distribution (Cialdella, 1984
) and thus are not isolated geographically. Moreover, although they seem to be outcrossers (they exhibit high genetic variability and their genotypic frequencies do not depart significantly from Hardy-Weinberg expectations), no intermediate forms have ever been recorded in sympatric areas, suggesting that they are able to hybridize. Alternatively, they might be members of a single species in which thorny stipule shape and length are determined by a few linked loci, as might be true of the traits used by Ebinger, Seigler, and Clarke (2000)
. A third possibility is that they are different species that diverged very recently and have yet to accumulate significant genetic differences. Because controlled crosses in A. aroma and A. macracantha are still impracticable, embryological and genetic studies, as well as indirect analysis of the mating system, may help explain their relationship.
|
FOOTNOTES |
|---|
4 Author for reprint requests (pcasiva{at}bg.fcen.uba.ar
)
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Allen O. N. E. K. Allen 1981 The Leguminosae. A source book of characteristics, uses and nodulation. University of Wisconsin, Madison, Wisconsin, USA
Avise J. C. 1994 Molecular markers, natural history, and evolution. Chapman & Hall, New York, New York, USA
Ayala F. M. L. Tracey L. G. Barr J. F. McDonald S. Pérez-Salas 1974 Genetic variation in natural populations of five Drosophila species and the hypothesis of the selective neutrality of protein polymorphisms. Genetics 77: 343-384
Bardakci F. O. F. Skibinsi 1994 Application of the RAPD technique in tilapia fish: species and subspecies identification. Heredity 73: 117-123
Bentham G. 1842 Notes on Mimosae, with a synopsis of species. London Journal of Botany 1: 318-392, 494–528
Bentham G. 1875 Revision of the suborder Mimosae. Transactions of the Linnean Society of London 30: 335-664
Bernhardt P. J. Kenrick R. B. Knox 1984 Pollination biology and the breeding system of Acacia retinodes (Leguminosae: Mimosoideae). Annals of the Missouri Botanical Garden 71: 17-29[CrossRef][ISI]
Bessega C. B. O. Saidman J. C. Vilardi 2000 Isozyme and RAPD studies in Prosopis glandulosa and P. velutina (Leguminosae, Mimosoideae). Genetics and Molecular Biology 23: 1-5
Black IV W. C. 1996 FORTRAN programs for the analysis of RAPD-PCR markers in populations. Colorado State University, Ft. Collins, Colorado, USA
Bukhari Y. M. 1997a Nuclear DNA amounts in Acacia and Prosopis (Mimosaceae) and their evolutionary implications. Hereditas 126: 45-51[CrossRef][ISI]
Bukhari Y. M. 1997b Cytoevolution of taxa in Acacia and Prosopis (Mimosaceae). Australian Journal of Botany 45: (5)879-891[CrossRef]
Butcher P. A. G. F. Moran H. D. Perkins 1998 RFLP diversity in the nuclear genome of Acacia mangium. Heredity 81: 205-213[CrossRef][ISI]
Casiva P. 2001 Utilización de marcadores moleculares y morfológicos para el estudio de la diferenciación y variabilidad en poblaciones naturales argentinas del género Acacia (Leguminosae). Licensing thesis, University of Buenos Aires, Buenos Aires, Argentina
Cialdella A. M. 1984 El género Acacia (Leguminosae) en la Argentina. Darwiniana 25: 59-111
Cialdella A. M. 1997 Acacia. In A. T. Hunziker [ed.], Flora fanerogámica Argentina, fascicle 25, 3–21. Proflora CONICET, Córdoba, Argentina
Clarke H. D. S. R. Downi D. S. Seigler 2000 Implications of chloroplast DNA restriction site variation for systematics of Acacia (Fabaceae: Mimosoideae). Systematic Botany 25: 618-632[CrossRef][ISI]
Comincini S. M. Sironi C. Bandi C. Giunta M. Rubuini F. Fantana 1996 RAPD analysis of systematic relationships among the Cervidae. Heredity 76: 215-221
Cottrell J. E. G. I. Forrest I. M. S. White 1997 The use of random amplified polymorphic DNA markers to identify and estimate the relatedness of clones belonging to the genus Populus. Botanical Journal of Scotland 49: 89-102
Crisci J. V. M. F. Armengol 1983 Introducción a la teoría y práctica de la taxonomía numérica. Organización de los Estados Americanos, Washington, D.C., USA
Dellaporta S. L. J. Wood J. B. Hicks 1983 A plant DNA minipreparation: version II. Plant Molecular Biology Report 1: 19-21
Dimitri M. J. J. S. Biloni 1973 El libro del árbol, I. Celulosa Argentina S.A. Buenos Aires, Argentina
Ebinger J. E. D. Seigler H. D. Clarke 2000 Taxonomic revision of South American species of the genus Acacia (Fabaceae: Mimosoideae). Systematic Botany 25: 588-617[CrossRef][ISI]
Fagg C. W. R. D. Barnes C. Marunda 1997 African Acacia Trials Network: a seed collection of six species for provenance-progeny test: held at the Oxford Forestry Institute. FAO-FOREST Genetic Resources 25, Oxford, UK
Felsenstein J. 1993 PHYLIP (phylogenetic inference package) v. 3.5c. Department of Genetics, University of Washington, Seattle, Washington, USA
Guinet P. J. Vassal 1978 Hypotheses on the differentiation of the major groups in the genus Acacia (Leguminosae). Kew Bulletin 32: 509-527[CrossRef]
Gutiérrez J. R. J. J. Armesto 1981 El rol del ganado en la dispersión de las semillas de Acacia caven (Leguminosae). Ciencia e Investigación Agraria 8: 3-8
Hamrick J. C. J. W. Godt 1990 Allozyme diversity in plant species. In H. D. Brown, M. T. Clegg, A. L. Kahler, and B. S. Weir [eds.], Plant population genetics, breeding, and genetic resources, 43–65. Sinauer Associates, Sunderland, Massachusetts, USA
Hollingsworth P. M. R. M. Bateman R. J. Gornall [eds.] 1999 Molecular systematics and plant evolution. Taylor and Francis, London, UK
IUPAC-IUB. 1979 Enzyme nomenclature: recommendations of the nomenclature committee of the International Union of Biochemistry. Academic Press, New York, New York, USA
Jenczenwski E. J. M. Prosperi J. Ronfort 1999 Differentiation between natural and cultivated populations of Mendicago sativa (Leguminosae) from Spain: analysis with random amplified polymorphic DNA (RAPD) markers and comparison to allozymes. Molecular Ecology 8: 1317-1330[CrossRef][Medline]
Joly H. I. M. Zeh-Nlo P. Danthu C. Aygalent 1992 Population genetics of an African acacia, Acacia albida I. Genetic diversity of populations from West Africa. Australian Journal of Botany 40: 59-73[CrossRef][ISI]
Karlin O. U. R. O. Coirini L. Catalán R. Zapata 1997 Acacia caven. In Oficina Regional de la FAO para America Latina y el Caribe [ed.], Especies arbóreas y arbustivas para las zonas áridas y semiáridas de América Latina, 157–167. Zonas áridas y semiáridas, 12. FAO/PNUMA, Santiago, Chile
