Research Paper Natural Hybridization in Saxifraga callosa Sm. F. Grassi1, M. Labra2, L. Minuto3, G. Casazza3, and F. Sala1, 4 1 Botanical Garden, Department of Biology, University of Milan, Via Celoria 26, Milan, Italy 2 DISAT, University of Milano-Bicocca, Piazza della Scienza 1, 20126 Milan, Italy 3 DIP.TE.RIS, University of Genova, Corso Dogali 1M, 16136 Genova, Italy 4 CNR, IBF, Via Celoria 26, Milan, Italy Received: June 22, 2005; Accepted: November 10, 2005 Abstract: Saxifraga callosa Sm. is an evergreen perennial species distributed from Eastern Spain, through the Western Alps and the Apennines, to southern Italy. The existence of high morphological variation within different subspecies indicates that phenotypic characters are useful but not sufficient taxonomic tools. Indeed, available morphological data already suggested that S. callosa subentity lantoscana may be an outcross between S. callosa and S. cochlearis. In this work, by analyzing ITS (Internal Transcribed Sequences), AFLP (Amplified Fragment Length Polymorphisms), and cpDNA (chloroplast DNA) markers, a comprehensive study of the genomic relationships among S. callosa and related species has been carried out. The sequence of the ITS region of S. callosa subentity lantoscana gave no conclusive results on the taxonomy status of S. callosa subentity lantoscana. On the other hand, the use of the “NewHybrids” software to analyze an AFLP data-set (208 polymorphic amplified fragments) supported a significant posterior probability that S. callosa subentity lantoscana individuals are natural hybrids between S. callosa and S. cochlearis. The level of introgression of genes from alien genomes was confirmed by a simpler and quick methodology that analyze length variation in cpDNA sequences. Key words: AFLP, introgression, ITS, hybridization, morphology, Saxifraga callosa, Saxifraga cochlearis, TrnL. Introduction Natural hybridization and hybrid zones in plants have been widely studied (Arnold, 1996; Rieseberg and Wendel, 1993; Rieseberg et al., 1999; Barton, 2001). Genetic and ecological consequences have been proposed for introgression of genetic material through hybridization, including: increased genetic diversity, origin or transfer of adaptations, origin of ecotypes or species, breakdown or reinforcement of isolating barriers, and promotion of dispersion and colonization (Rieseberg and Wendel, 1993; Potts and Reid, 1988). Spontaneous hybridization is relatively common in the plant kingdom and played an important positive role in their evolution (Barton and Hewitt, Plant Biol. 8 (2006): 243 – 252 © Georg Thieme Verlag KG Stuttgart · New York DOI 10.1055/s-2005-873047 ISSN 1435-8603 1985, 1989; Barton, 2001). Hybridization may also negatively affect the preservation of species and the ability of plants to reproduce, compete, and deter pathogens and predators (Levin et al., 1996; Abbot, 2003). In some cases, it also negatively affects the growth rate of parental species, due to higher competition of hybrids for habitat or resources. This may be critical in the case of rare species (Soltis and Gitzendanner, 1999). The aim of this paper was the characterization of the genetic structure of the species Saxifraga callosa, with emphasis on the possibility of hybridization with related species. The genus Saxifraga includes about 440 species (Kohlein, 1984; Webb and Gornall, 1989). Most of these occur in the Arctic and temperate zone. The areas with the greatest species diversity include the Himalayan-Tibetan region, both sides of the Northern Pacific (Western North America and Eastern Asia), and the mountains of Europe. Flower morphology, as well as shape, colour, and leaf dimension, number of carpels and petal colour, allowed inclusion of Saxifraga species in the Ligulatae Haworth section. The latter are the ‘’silver‘’ or ‘’encrusted‘’ saxifrages, characterized by conspicuous calcareous encrustations on the leaf margins (Conti et al., 1999). Based on morphological variation, section Ligulatae has been subdivided into three subsections (Webb and Gornall, 1989): Aizoonia (Tausch) Schott, Mutatae (Engler and Irmscher) Gornall, and Florulentae (Engler and Irmscher) Gornall. More recent molecular data suggested that section Ligulatae does not constitute a monophyletic group (Conti et al., 1999): actually, only the eight Aizoonia species (Webb and Gornall, 1989) can be defined as the “true” silver saxifrages, characterized by the combination of glaucous leaves, conspicuous calcareous encrustation, white or pale pink petals usually rounded at the tip and bicarpellate ovary. Saxifraga callosa Sm. is an evergreen perennial, member of the Aizoonia group present in Europe and with specific habitat requirements. Plants occur in more or less vertical limestone rocks between 1000 and 2000 m altitude and range from eastern Spain through the south-western Alps, the Apuan Alps, and the Apennines down to southern Italy. Based on morphology, Webb (1993) distinguished S. callosa into four subtaxa: a) subsp. callosa with an inner subdivision into b) var. australis (Moricand) D. A. Webb, c) subsp. catalunica E. Bourgeau, and d) subentity lantoscana. The first three taxa are geographically well separated: subsp. callosa colonized southern France and the northern Apennines, while its var. australis is present in the central and southern Apennines; subsp. catalaunica colo- 243 244 Plant Biology 8 (2006) Table 1 F. Grassi et al. Summary of the species, populations, and number of individuals used for each molecular analysis Saxifraga Population Name Code Location ITS AFLP cpDNA S. callosa subsp. callosa M. Toraggio M. S. Bernardo M. Saccarello M. Pizzacuto To1 SB Sac Piz 43858′N, 07839′E 44804′N, 07851′E 44803′N, 07843′E 44804′N, 07846′E 1 1 1 1 6 6 – – 12 12 8 8 S. callosa subentity lantoscana M. Toraggio M. Pietravechia To2 Pie 43858′N, 07839′E 43859′N, 07840′E 1 1 6 6 8 8 S. callosa subsp. catalunica Monserrat S. callosa subsp. australis M. Mucchia Cat 43858′N, 07839′E 1 6 8 Aus 44805′N, 07836′E – 6 6 S. cochlearis M. Grai M. Toraggio Gra To3 43859′N, 07840′E 43858′N, 07839′E 1 1 3 2 6 6 S. paniculata S. longifolia S. aizoides S. hostii M. Toraggio M. Toraggio M. Toraggio M. Toraggio – – – – Botanical Garden of Genova Botanical Garden of Genova Botanical Garden of Genova Botanical Garden of Genova 1 1 1 1 1 – 1 – – – – – nized the Montserrat massif near Barcelona (Spain). In the past, different authors described the morphological variation of subsp. callosa in the Maritime Alps (Boissier, 1956; Arcangeli, 1882; Fiori, 1929; Pignatti, 1982). Based on this, the subentity lantoscana was proposed, the taxon being defined either as subspecies or variety (S. lantoscana Boiss. et Reut.; S. callosa subsp. lantoscana [Boiss. et Reut.] Arcang.; S. callosa var. lantoscana [Boiss. et Reut.] Fiori). In fact, the almost continuous morphological variation among different Saxifraga subtaxa makes their identification problematic, as phenotypic variation is important but not sufficient to produce a reasonable taxonomic classification (Webb, 1993). The occurrence of hybridization among related Saxifraga species is well documented (Kohlein, 1984). This contributes to complicate their taxonomic definition. Within the genus, several inter-specific hybrids are cited in the literature (Gugerli, 1997) and often used in horticultural activities. S. callosa, outcrosses with S. cotyledon, S. longifolia, S. paniculata, and a possible hybrid with S. cochlearis (S. × farreri Druce) has been described (Webb and Gornall, 1989). However, in general it is difficult to detect genome hybridization merely on the observation of morphological characters, since backcross hybridization frequently leads to resemblance to one of the parental species. Molecular markers of nuclear and plastid DNA (cpDNA), in combination with morphological characters, are widely used today to study the effect of recombination and of genetic barriers between natural populations (Rieseberg and Soltis, 1991; Rieseberg et al., 1999). The use of cpDNA uniparental markers is essential since they allow detection of backcross hybrids (Rieseberg and Ellstrand, 1993; Rieseberg and Brunsfeld, 1992; Rieseberg and Wendel, 1993). In this paper, independent molecular data sets, based on ITS (Baldwin et al., 1995), AFLP (Vos et al., 1995), and cpDNA, have been adopted, in combination with morphological traits, to resolve relationships within the S. callosa taxon. Materials and Methods Plant material and DNA extraction Individuals of each subspecies of S. callosa (subsp. callosa, subsp. australis, subsp. catalaunica, and subentity lantoscana) (Table 1) were sampled in their distribution area (Fig. 1). In addition, individuals belonging to other Saxifraga species (S. cochlearis Reichenb., S. paniculata Miller, S. aizoides L., and S. hostii Tausch) were collected in the same areas where S. callosa was sampled. On the basis of a previous study (Conti et al., 1999), other taxa completing the section Ligulatae (S. valdensis DC., S. cotyledon L., S. crustata Vest, S. florulenta Moretti, S. longifolia Lapeyr., and S. mutata L.) (Table 1) were obtained from the Botanical Garden of the University of Genova. Reference codes, geographical locations, and sampling data are given in Table 1 and Fig. 1. In particularly, S. valdensis confined to a small region of the south-western Alps, mostly near the Franco-Italian frontier. One/two leaves were collected from each individual, washed to remove calcareous deposits and total genomic DNA extracted by using the Qiagen Kit method. Morphological characters Morphological characters, listed in Table 2, were coded taxon by taxon because discontinuities could not be unambiguously coded within taxa. If the quantitative variation in one taxon included two or more “states”, the taxon was scored as polymorphic. The morphological data matrix was run using PAUP 4.0 (beta version 4.0b7, Swofford, 2001) with maximum parsimony methods. Bootstrap values were calculated with branch and bound algorithm based on 10 000 replicates. ITS region strategies The ITS1 and ITS2 spacers and the 5.8S gene (within the ITS region) were amplified and sequenced as a continuous region. Forward and reverse primers were TCC GTA GGT GAA CCT GCG GAA GGA TCA TTG (ITS1 primer) and TCC TCC GCT TAT Natural Hybridization in Saxifraga callosa Sm. Plant Biology 8 (2006) Fig. 1 Geographic distributions of S. callosa and S. cochlearis, as described by Webb and Gornall (1989). Taxon abbreviation: cal, S. callosa subsp. callosa; cat, S. callosa subsp. catalunica; aus, S. callosa subsp. australis; lan, S. callosa subentity lantoscana; coc, S. cochlearis. TGA TAT GC (ITS4 primer), respectively (Conti et al., 1999). PCR amplification was performed with “PCR Ready to go” (Amersham, Italy). PCR products were purified using the QIAquick PCR Purification Kit (Qiagen Inc.). DNA sequencing was performed using dRhodamine Chemistry (Applied Biosystems), following the manufacturer’s protocol, and run in an automated DNA Sequencer (Applied Biosystems). Both strands were sequenced with ITS1 as forward sequencing primers and ITS4 as reverse primer. ITS sequences were aligned with the Staden Package. Phylogenetic analyses of sequence data were performed with PAUP 4.0 (beta version 4.0b7, Swofford, 2001) with maximum parsimony. Bootstrap values were calculated with heuristic searches based on 10 000 replicates. AFLP analysis AFLP analysis was performed as described in the European Patent 0 534 858 (Keygene, Belgium). Essentially, genomic DNA (150 ng) was digested (3 h) with the EcoRI (1 U) and MseI (1 U) restriction enzymes. DNA fragments were ligated, with T4 DNA-ligase, to 5 pmol EcoRI adapter and 50 pmol MseI adapter in a final volume of 50 μl. Restriction and ligation were performed at 37 8C for 6 h. This mixture was used as template in the pre-amplification reaction containing the DNA primers E01 and M01, complementary to the core of the EcoRI and MseI adapters, respectively. The 50-μl amplification mixture contained 20 μl of digested/ligated DNA, 50 ng of the selected primers, 200 mM of each dNTP, 0.5 U Dynazyme II (Finnzymes, Finland), and 5 μl Dynazyme buffer. After 2 min at 94 8C, amplification was carried out for 20 cycles of denaturation (45 s at 94 8C), annealing (30 s at 50 8C), and extension (1 min at 72 8C). After a final elongation step (7 min at 72 8C), the pre-amplification product was diluted 1 : 20 with water and used for selective amplification. This was carried out using one of the selective pairs of primers (M34-E49, M35-E46, M36-E37, M36-E38, M39-E34, M40-E38, M42-E35, M35-E39). The amplification mixture (10 μl, final volume) contained 2.5 μl of the pre-amplification mixture, 5 ng of labelled EcoRI primer, 30 ng of MseI primer, 200 μM of each dNTP, 0.5 U Dynazyme II (Finnzymes, Finland), and 5 μl Dynazyme buffer. After 2 min at 94 8C, 36 cycles of amplification were performed under the following conditions: denaturation for 30 s at 94 8C; annealing for 30 s at 65 8C for the first cycle, followed by a decrease of temperature (0.7 8C in each cycle) in the following 12 cycles. The remaining 23 cycles were under the following conditions: 60 s at 92 8C (denaturation), 30 s at 56 8C (annealing), and 60 s at 72 8C (extension). A total of 5 μl of the PCR-amplified mixture was added to an equal volume of loading buffer (80 % formamide, 1 mg/ml xylene cyanol FF, 1 mg/ml bromophenol blue, 10 M EDTA, pH 8.0), denatured for 5 min at 92 8C, loaded onto a 6 % denaturing polyacrylamide gel and electrophoresed in TBE electrophoresis buffer for 3 h at 75 W. Amplified bands, in the range 80 – 500 bp, were scored for their presence (1) or absence (0) and the resulting data matrix was used for different statistical analysis. Reproducibility was verified in duplicate amplifications and electrophoresis. A dissimilarity matrix was computed using the Nei and Li (1979) coefficient. The Neighbour-Joining (NJ) bootstrapped phenogram (2000 replicates) was calculated using the TREECON software. The degree of AFLP polymorphism was quantified using the Shannon index (Sh) (Shannon and Weaver, 1949) to calculate allelic richness and gene diversity (GD) (Nei, 1987). These were used to determine genetic variability with the POPGENE 1.21. software (Yeh et al., 1997). Principal component analysis (PCA) of the AFLP data was carried out using NTSYS-pc (Rohlf and Sokal, 1981). Raw AFLP data were converted into a similarity matrix using the Dice coefficient with the SIMQUAL function. Three eigenvectors were extracted from the correlation matrix using the EIGEN function. Moreover, for each S. callosa subentity lantoscana individual 245 246 Plant Biology 8 (2006) Table 2 F. Grassi et al. Summary of character number (No.), morphological characters, and character state (0, 1, or 2) of the analyzed accessions No. Morphological character 0 1 Fruit production polycarpic monocarpic Basal leaf 2 3 4 Length (mm) Width (mm) Shape 10 – 100 2 – 10 spathulate 5 6 7 8 9 10 11 12 Base/wide (mm) Base/colour Apex Margin Surface Ornamentation Glands Stolons < 10 <2 linear, linearspathulate, lanceolate <3 red obtuse, rounded entire glaucous lime encrusted Y Y >3 green acute, mucronate, apiculate serrate not glaucous not lime encrusted N N 1 Caulinare leaf 13 Length (mm) 14 Shape 15 Margin 16 Deportment 17 No. leaves before lower inflorescence branches 18 Glands < 30 linear entire erect <5 > 30 spath teeth patent 5 – 12 Y N Inflorescence 19 Deportment 20 Branches 21 Inflorescence/caule 22 No. flowers on the basal branch 23 Inflorescence type 24 Ornamentation 25 No. of flowers unilateral < 15 ≤ 1⁄ 2 ≤3 not pyramidal glabrous < 40 all directions 15 – 25 > 1⁄ 2 >3 pyramidal glandular 40 – 100 Flower 26 27 28 29 30 31 32 33 glabrous glabrous lanceolate Y white spath rounded glabre glandular glandular ovate N yellow orbicul acute ciliate Pedicel Hypanthium Sepal shape Glands Petal colour Petal shape Petal apex Petal base analyzed by AFLP, the posterior probability that it belongs to either S. callosa or S. cochlearis or to hybrid classes (F1, F2, or backcross) was estimated using a Bayesian method. With this approach, multilocus genotypic data are used to define a set of populations with distinct allele frequencies, hereafter referred to as clusters, and assign individuals probabilistically to defined clusters without prior knowledge of sampling location. At every iteration of the algorithm, the probability that each individual belongs to any of the different categories, given the current values of the allele frequencies and the mixing proportions, was computed. In the graphic representation, individuals are represented by horizontal bars. The total length of 2 > 100 > 10 oblong > 12 > 25 > 100 triang the bar represents a probability of 1.0, and the length of bar occupied by each colour represents the probability that the individual belongs to the genotype frequency category denoted by that colour. The Bayesian approach uses a Markov-chain Monte Carlo algorithm and is implemented in the “NewHybrids” software (Anderson and Thompson, 2002; Cianchi et al., 2003). Our analyses were performed without assuming prior population information, and runs for 2000 iterations. Natural Hybridization in Saxifraga callosa Sm. Plant Biology 8 (2006) Fig. 2 Most parsimonious trees based on: (A) morphological characters dataset analyzed with the Branch and Bound algorithm, and (B) ITS sequence dataset analyzed with the Heuristic algorithm. Bootstrap support values are given above branches (bootstrap percentages below 50% are not given). * Data obtained from GeneBank (Conti et al., 1999). CpDNA analysis Results Assuming strict maternal inheritance of the plastid genome, cpDNA analysis was used to assess the parentage of hybrid plants. Polymorphism within cpDNA was investigated in the following three sequences: i) the TrnL regions (Borsch et al., 2003), amplified by the Saxcp1R (5′-GAG TGA ATG AGA AAC ATA ACG-3′) and Saxcp1F (5′-GGT TCA AGT CCC TCT ATC CCC3′) primers, ii) the ccmp2, and iii) the ccmp10 microsatellite regions. Regions (ii) and (iii) were amplified with the primers described by Weising and Gardner (1999). The analysis was performed on 8 – 10 individuals for each population. In all cases, the PCR amplification profile was: 3 min at 94 8C; 35 cycles of denaturation (45 s at 94 8C), annealing (30 s at 50 8C), and extension (1 min at 72 8C), then a final step of 7 min at 72 8C. Amplification success was confirmed on 1.4 % agarose gels. Then, amplification products were separated and analyzed by vertical acrylamide (6 %) gel electrophoresis. Morphology The species listed in Table 1 were analyzed with a taxonomic approach based on 33 diagnostic traits (Webb and Gornall, 1989), 32 of which were found to be informative (Table 2). The results show that S. callosa subentity lantoscana shares the same state for 30 characters (90.91 % of total) with S. cochlearis and for 21 characters (63.64 % of total) with subsp. callosa. No exclusive character restricted to S. callosa and S. cochlearis was evident. Therefore, the tree based on morphological data (Fig. 2 A) shows high affinity of S. callosa subentity lantoscana with S. cochlearis, and places other subspecies in a separate clade. Furthermore, the catalaunica subspecies shares fewer character states with other taxa, i.e., 21 with S. callosa, 17 with S. crustata, 19 with S. hostii, and 18 with S. valdensis. The most common traits are typical of subsect. Aizoonia, as is the case for the entire margin of the cauline leaves and the glabrous inflorescence. 247 248 Plant Biology 8 (2006) F. Grassi et al. Fig. 3 Neighbour-joining tree obtained on the basis of the AFLP dataset. Numbers above branches indicate bootstrap support values (boot- strap percentages below 60% are not given). Population codes (in brackets) are those of Table 1. ITS analysis and evolution for the S. callosa subsp. catalunica can be proposed. The two ITS sequences amplified by PCR and sequenced on both strands were analyzed by alignment. Of the 425 aligned nucleotide positions for both combined spacers, 343 characters were constant, 52 were variable, and 30 were informative. The six sequences of S. callosa from the Maritime Alps were nearly identical. Consequently, these accessions were grouped in the same clade, with a bootstrap value of 99 % (Fig. 2 B). The ITS analysis gave significant information in the case of the S. callosa subsp. catalunica, indicating that this taxon and S. valdensis are closely related (bootstrap 94 %) and are more distantly related to S. callosa. Based on this, a different origin AFLP analysis A total of 208 polymorphic AFLP fragments were generated using eight primer pairs across the 40 Saxifraga accessions. The level of polymorphism within different subspecies was low. The Gene diversity value for all populations was: To1 = 0.031, SB = 0.058, To2 = 0.108, Pie = 0.108, Cat = 0.046, and Roc = 0.059 (abbreviations in Table 1). Moreover, the Shannon value for all populations was: To1 = 0.045, SB = 0.085, To2 = 0.161, Pie = 0.164, Cat = 0.068, and Roc = 0.068. Based on both Gene di- Natural Hybridization in Saxifraga callosa Sm. Plant Biology 8 (2006) Fig. 4 Graphic representation produced on the basis of the AFLP data matrix for S. callosa subsp. callosa, S. callosa subentity lantoscana, and S. cochlearis. (A) Plot of S. callosa subentity lantoscana individuals (grey squares) by first- and second-principal components showing high genetic variability and an intermediate position between S. cochlearis (black square) and S. callosa subsp. callosa (white squares) samples. In (B) the length of each bar reflects the Bayesian Posterior Probabilities that the 26 analyzed individuals belong to parental, F1, F2, or BC. Samples 1 – 5 belong to the To1 populations (see Table 1), samples 6 – 10 to the SB populations, samples 11 – 15 to the To2 populations, and samples 16 – 21 to the Pie populations. The last 5 samples (22 – 26) belong to the Gra and To3 populations. Individuals 16, 17, 19, 20, 21 (indicated by arrows) showed a cpDNA fragment deriving from S. cochlearis. versity and Shannon data, the analyzed populations of S. callosa subentity lantoscana (To2 and Pie) revealed the highest polymorphism values. A NJ tree obtained from the AFLP data is shown in Fig. 3. S. callosa subsp. catalunica samples showed a relatively large genetic distance from S. callosa (genetic distance mean value = 16.25 ± 1.24), with at least one unique fragment for amplification with each primer pair. The NJ tree (Fig. 3) revealed a highly structured S. callosa subsp. catalunica taxon. This was also supported by the high bootstrap value. These data agree with the morphological and ITS datasets. Moreover, the S. callosa subsp. australis accessions turned out to be related to S. callosa subspecies (genetic distance mean value = 7.09 ± 1.83). PCA analysis, performed on S. cochlearis, S. callosa subsp. callosa, and S. callosa subentity lantoscana DNA, allowed discrimination of the two species (Fig. 4 A). The first principal component (Axis 1) explained 42 % of the variation, thus showing a clear separation between the two species. The second component (Axis 2) explained 11 % of the variation, pointing to an intermediate position of S. callosa subentity lantoscana together with the occurrence of a large genetic variability. Our results, together with the field observations reported by Webb (1989), imply inter-fertility between S. cochlearis and S. callosa. To verify the hybridization hypothesis (Floate et al., 1994), we estimated the posterior probability that each individual in a sampling belongs to a different hybrid class (Fig. 4 B). The S. callosa accessions represent a single genotype (a pure line) with a probability > 0.97 %. Similar results were observed for all S. cochlearis accessions. However, the histogram of Fig. 4 B showed high values of variability within S. callosa subentity lantoscana. Three samples of this subtaxon (sample 12, 17, and 19 of Fig. 4 B) were assigned to S. callosa subsp. callosa, with a probability of about 50 %, while all other samples showed a lower probability of hybridization. Plastid DNA analysis A clear difference between S. callosa and S. cochlearis was identified in the Saxcp1 plastid sequence. On the other hand, within the Pie population (S. callosa subsp. lantoscana), 75 % of the accessions showed cpDNA fragments characteristic of S. cochlearis (Fig. 5). Thus, cpDNA analysis provided a convenient and efficient tool to search and validate cases of maternal introgression through the analysis of uniparentally inherited genes. Sequencing of the amplification products revealed a deletion of 249 250 Plant Biology 8 (2006) F. Grassi et al. Fig. 5 Electrophoresis of the DNA fragments produced by PCR amplification of locus Saxcp1. Horizontal arrows identify two different alleles. To1, S. callosa subsp. callosa from M. Toraggio; Pie, S. callosa subentity lantoscana from M. Pietravecchia; Cat, S. callosa subsp. catalunica from Monserrat; Aus, S. callosa subsp. australis from Mucchia; Gra, S. cochlearis from M. Gray. 5 base pairs (TGATC) in the Gra (S. cochlearis) and Pie (S. callosa subsp. lantoscana) populations. No nucleotide dissimilarity was found for the other two cpDNA microsatellites (ccmp2 and ccmp10) within and among S. callosa and S. cochlearis species. Discussion The main issue addressed in this paper concerns the occurrence of natural hybridization in S. callosa. Morphological data had already been interpreted as if S. callosa subentity lantoscana was a case of an outcross between S. callosa and S. cochlearis. The Maritime Alps are an important area in which to investigate hybridization events: the region was a possible refuge for European populations during the Pleistocene ice period and later a contact zone between the western European and Italian taxa. However, in all location of the Maritime Alps the three taxa often live together, and those who studied them always found difficulties in separating the three entities (Pignatti, 1982; Webb and Gornall, 1989; Webb, 1993). The subentity lantoscana was proposed to be an ecotype of S. callosa living in more stressed habitats, like rocky cliffs exposed to strong winds. Our results, based on molecular evidence, now demonstrate the occurrence of hybridization between S. callosa and S. cochlearis. Hybridization was made possible by the overlapping distribution of S. callosa subentity lantoscana and S. cochlearis in the Maritime Alps (Pignatti, 1982); both taxa fall into the wider distribution range of S. callosa (Webb, 1993). It should be emphasized that the appearance of new taxonomic entities within the Maritime Alps during the latest Ice Ages has already been documented (Casazza et al., 2005). In this study, the ITS analysis resolved the phylogeny of the genus (Conti et al., 1999), but was ineffective in clarifying relationships within S. callosa. Indeed, the ITS marker has already proved to be a useful source of information for the resolution of phylogenetic relationships above or at species level (Baldwin, 1992; Alvarez and Wendel, 2003), but not within species complexes. Consequently, the AFLP molecular markers were used to distinguish closely related genotypes (Gabrielsen et al., 1997; Gabrielsen et al., 1998; Loh et al., 2000; Sunnucks, 2000; Després et al., 2003; Grassi et al., 2004). Discontinuous markers, such as AFLP, are powerful tools to assign genetic distance among neighbouring taxa (El-Rabey et al., 2002; Pelser et al., 2003; Labra et al., 2003; Creer et al., 2004). However, their use for phylogenetic studies that combine different markers, characterized by different degrees of polymorphism, may resolve relationships at different taxonomic levels. On the other hand, the exclusive use of AFLP to resolve phylogenetic relationships is controversial, because of the possibility that the technique may reveal a high proportion of homoplasious fragments, particularly at high taxonomic levels. The AFLP analysis revealed the existence of a large variability within subentity lantoscana populations, thus suggesting that inter-species transfer of genes through hybridization is a source of conditional genetic variability. The finding of introgression in S. callosa of genomic AFLP markers originally from S. cochlearis, as revealed by the “NewHybrids” software, agreed with data from cpDNA analysis. In fact, this analysis showed the presence in S. callosa subentity lantoscana of alleles which are not found in S. callosa. A large part of the analyzed Pie accessions (S. callosa subentity lantoscana) showed cpDNA sequences that are typical of S. cochlearis. In conclusion, data produced using molecular markers allowed us to postulate that hybridization events have occurred at least in a part of the considered Saxifraga populations. The experimental data allow us to propose that S. callosa subentity lantoscana should not be considered a subspecies of S. callosa, but rather a hybrid taxon. On the other hand, the To2 population did not show the presence of alien plastid fragments. This is expected since hybridization between two species does not necessarily occur with the same frequency in different natural ecosystems. For example, hybridization between Picea marina and Picea rubens has been shown to be extensive on an analyzed coastal island, but limited in continental mountain populations (Bobola, 1996). Our study shows that the AFLP approach is especially suitable for the identification of interspecific hybrids; thus, we encourage its application to solve similar taxonomic problems in Saxifraga and in other systematic groups. Moreover, we are convinced that molecular phylogenetic studies at low taxonomic level should not merely rely on sequence data from single individuals as representative of a species. A further issue addressed by the present work is related to the phylogenetic position of the S. callosa subsp. catalunica. The AFLP and the ITS data concur to suggest that S. callosa subsp. catalunica accessions are the result of an evolutionary pathway dissimilar to that of S. callosa. This substantiates the recent proposal that this entity should be transferred to a higher range (Catroviejo, 1997). S. catalaunica Boiss. et Reut. should be considered a separated species from S. callosa and placed within the subsection Aizoonia, thus increasing to nine the number of its members. No information has been offered by our molecular data on the evaluation of S. callosa subsp. australis. Here, more markers for ecologically relevant traits are needed, also in order to verify its relationship with S. callosa subspecies (Van Tienderen et al., 2002; Rieseberg et al., 2003). It is conceivable that nuclear microsatellite DNA or scnDNA (Single copy nuclear DNA) loci might have the potential to overcome these difficulties, be- Natural Hybridization in Saxifraga callosa Sm. cause of their high mutation rate and their potentially high information content (Richard and Thorpe, 2001; Gugerli et al., 2001). Acknowledgements This research was supported by FIRB-RNBE01SF project, MIUR, Italy. Thanks are given to Prof. Francesco Salamini and to Prof. Elena Conti for perceptive comments that helped to improve this manuscript. We are also grateful to Roberto Brontini and Fabrizio Celia for technical support and molecular analyses. References Abbott, R. J. (2003) Sex, sunflowers, and speciation. Science 301, 1189 – 1190. Alvarez, I. and Wendel, J. F. (2003) Ribosomal ITS sequences and plant phylogenetic inference. Molecular Phylogenetics and Evolution 29, 417 – 434. Anderson, E. C. and Thompson, E. A. (2002) A model-based method for identifying species hybrids using multilocus genetic data. Genetics 160, 1217 – 1229. Arcangeli, G. (1882) Compendio della Flora Italiana. Torino: Loescher. Arnold, M. (1996) Natural Hybridization and Introgression. Princeton: Princeton University Press. Baldwin, B. G. (1992) Phylogenetic utility of the internal transcribed spacers of nuclear ribosomal DNA in plants: an example from the Compositae. Molecular Phylogenetics and Evolution 1, 3 – 16. Baldwin, B. G., Sanderson, M. J., Porter, J. M., Wojciechowski, M. F., Campbell, C. S., and Donoghue, M. J. (1995) The ITS region of nuclear ribosomal DNA: a valuable source of evidence on angiosperm phylogeny. Annals of the Missouri Botanical Garden 82, 247 – 277. Barton, N. H. (2001) The role of hybridization in evolution. Molecular Ecology 10, 551 – 568. Barton, N. H. and Hewitt, G. M. (1985) Analysis of hybrid zones. Annual Review of Ecology and Systematics 16, 113 – 148. Barton, N. H. and Hewitt, G. M. (1989) Adaptation, speciation and hybrid zones. Nature 341, 497 – 503. Bobola, M., Eckert, R., Klein, A., Stapelfeldt, K., Hillenberg, K., and Gendreau, S. (1996) Hybridisation between Picea rubens and Picea marina: differences observed between mountain and coastal island populations. Canadian Journal of Forest Research 26, 444 – 452. Boissier, E. (1956) Diagnoses Plantarum Orientalium Novarum. Ser. II. Leipzig, Paris, pp. 63 – 64. Borsch, T., Hilu, K. W., Quandt, D., Wilde, V., Neinhuis, C., and Barthlott, W. (2003) Noncoding plastid trnT-trnF sequences reveal a well resolved phylogeny of basal angiosperms. Journal of Evolutionary Biology 16, 558 – 576. Casazza, G., Barberis, G., and Minuto, L. (2005) Ecological characteristics and rarity of endemic plants from Italian Maritime Alps. Biological Conservation 123, 361 – 371. Catroviejo, S. (1997) Flora Ibérica. In Plantas Vasculares de la Península Ibérica e Islas Baleares (Catroviejo, S., ed.), Madrid: Real Jard. Bot. C.S.I.C., pp.162 – 247. Cianchi, R., Ungano, A., Marini, M., and Bullini L. (2003) Differential patterns of hybridization and introgression between the swallowtails Papilio machaon and P. hospiton from Sardinia and Corsica islands (Lepidoptera, Papilionidae). Molecular Ecology 12, 1461 – 1471. Conti, E., Soltis, D. E., Hardig, T. M., and Schneider J. (1999) Phylogenetic relationships of the silver Saxifrages (Saxifraga, sect. Ligulatae Haworth): implications for the evolution of substrate specificity, life histories, and biogeography. Molecular Phylogenetics and Evolution 13, 536 – 555. Plant Biology 8 (2006) Creer, S., Thorpe, R. S., Malhotra, A., Chou, W. H., and Stenson, A. G. (2004) The utility of AFLPs for supporting mitochondrial DNA phylogeographical analyses in the Taiwanese bamboo viper, Trimeresurus stejnegeri. Journal of Evolutionary Biology 17, 100 – 107. Després, L., Gielly, L., Redoutet, B., and Taberlet, P. (2003) Using AFLP to resolve phylogenetic relationships in a morphologically diversified plant species complex when nuclear and chloroplast sequences fail to reveal variability. Molecular Phylogenetics and Evolution 27, 185 – 196. El-Rabey, H. A., Badr, A., Schäfer-Pregl, R., Martin, W., and Salamini F. (2002) Speciation and species separation in Hordeum L. (Poaceae) resolved by discontinuous molecular markers. Plant Biology 4, 1 – 9. Fiori, A., (1929) Nuova Flora Analitica d’Italia. Firenze: Ricci. Floate, K. D., Whitham, T. G., and Keim, P. (1994) Morphological versus genetic markers in classifying hybrid plants. Evolution 48, 929 – 930. Gabrielsen, T. M., Bachmann, K., Jakobsen, K. S., and Brochmann, C. (1997) Glacial survival does not matter: RAPD phylogeography of Nordic Saxifraga oppositifolia. Molecular Ecology 6, 831 – 842. Gabrielsen, T. M. and Brochmann, C. (1998) Sex after all: high levels of diversity detected in the arctic clonal plant Saxifraga cernua using RAPD markers. Molecular Ecology 7, 1701 – 1708. Grassi, F., Cazzaniga, E., Minuto, L., Peccenini, S., Barberis, G., and Basso, B. (2004) Evaluation of biodiversity in Pancratium maritimum L. and conservation strategies for the Northern Tyrrhenian Sea. Biodiversity and Conservation 14, 2159 – 2169. Gugerli, F. (1997) Hybridization of Saxifraga oppositifolia and S. biflora (Saxifragaceae) in a mixed alpine population. Plant Systematics and Evolution 207, 255 – 272. Gugerli, F., Senn, J., Anzidei, M., Madaghiele, A., Büchler, U., Sperisen, C., and Vendramin, G. G. (2001) Chloroplast microsatellites and mitochondrial nad1 intron 2 sequences indicate congruent phylogenetic relationships among Swiss stone pine (Pinus cembra), Siberian stone pine (Pinus sibirica), and Siberian dwarf pine (Pinus pumila). Molecular Ecology 10, 1489 – 1497. Kohlein, F. (1984) Saxifrages and Related Genera. London: Batsford. Labra, M., Grassi, F., Bardini, M., Imazio, S., Guiggi, A., Citterio, S., Banfi, E., and Sgorbati, S. (2003) Genetic relationships in Opuntia Mill. genus (Cactaceae) detected by molecular marker. Plant Science 165, 1129 – 1136. Levin, D. A., Francisco-Ortega, J., and Jansen R. K. (1996) Hybridization and the extinction of rare plant species. Conservation Biology 10, 10 – 16. Loh, J. P., Kiew, R., Set, O., Gan, L. H., and Gan, Y. Y. (2000) Amplified fragment length polymorphism fingerprinting of 16 banana cultivars (Musa cvs.). Molecular Phylogenetics and Evolution 17, 360 – 366. Nei, M. (1987) Molecular Evolutionary Genetics. New York: Columbia University Press. Nei, M. and Li, W. H. (1979) Mathematical model for studying genetic variation in terms of restriction endonucleases. Proceedings of the National Academy of Sciences of the USA 76, 5269 – 5273. Pelser, P. B., Gravendeel, B., and Van der Meijden, R. (2003) Phylogeny reconstruction in the gap between too little and too much divergence: the closest relatives of Senecio jacobaea (Asteraceae) according to DNA sequences and AFLPs. Molecular Phylogenetics and Evolution 29, 613 – 628. Pignatti, S., (1982) Flora d’ Italia. Bologna: Edagricole. Potts, B. M. and Reid, J. B. (1988) Hybridization as a dispersal mechanism. Evolution 42, 1245 – 1255. Richard, M. and Thorpe, R. S. (2001) Can microsatellites be used to infer phylogenies? Evidence from Population Affinities of the Western Canary Island Lizard (Gallotia galloti). Molecular Phylogenetics and Evolution 3, 351 – 360. 251 252 Plant Biology 8 (2006) Rieseberg, L. H. and Brunsfeld, S. J. (1992) Molecular evidence and plant introgression. In Molecular Systematics of Plants (Soltis, P. S., Soltis, D. E., and Doyle, J. J., eds.), New York: Chapman and Hall, pp.151 – 178. Rieseberg, L. H. and Ellstrand, N. C. (1993) What can molecular and morphological markers tell us about plant hybridization? Critical Review in Plant Science 12, 213 – 241. Rieseberg, L. H., Raymond, O., Rosenthal, D. M., Lai, Z., Livingstone, K., Nakazato, T., Durphy J. L., Schwarzbach, A. E., Donovan, L. A., and Lexer, C. (2003) Major ecological transitions in wild sunflowers facilitated by hybridization. Science 301, 1211 – 1216. Rieseberg, L. H. and Soltis D. E. (1991) Phylogenetic consequences of cytoplasmic gene flow in plants. Evolutionary Trends in Plants 5, 65 – 84. Rieseberg, L. H. and Wendel, J. F. (1993) Introgression and its consequences in plants. In Hybrid Zones and the Evolutionary Process (Harrison R. G., ed.), New York: Oxford University Press, pp. 70 – 109. Rieseberg, L. H., Whitton, J., and Gardner, K. (1999) Hybrid zones and the genetic architecture of a barrier to gene flow between two sunflower species. Genetics 152, 713 – 727. Rohlf, F. J. and Sokal, R. R. (1981) Comparing numerical taxonomic studies. Systematic Zoology 30, 459 – 490. Shannon, C. E. and Weaver, W. (1949) The Mathematical Theory of Communication. Urbana: University of Illinois Press. Soltis, P. S. and Gitzendanner, M. A. (1999) Molecular systematics and the conservation of rare species. Conservation Biology 13, 471 – 483. Sunnucks, P. (2000) Efficient genetic markers for population biology. Trends in Ecology and Evolution 15, 199 – 203. Swofford, D. L. (2001) PAUP*: Phylogenetic Analysis Using Parsimony (*and other methods), Version 4.0b2. Sunderland, MA: Sinauer Associates. Van Tienderen, P. H., De Haan, A. A., Van der Linden, C. G., and Vosman, B. (2002) Biodiversity assessment using markers for ecologically important traits. Trends in Ecology and Evolution 17, 577 – 582. Vos, P., Hogers, R., Bleeker, M., Reijans, M., Van de Lee, T., Horenes, M., Fiujters, A., Pot, J., Pelerman, J., Kuiper, M., and Zabeau, M. (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Research 23, 4407 – 4414. Webb, D. A. (1993) Saxifraga. In Flora Europaea, Vol. I (Tutin, T. G., Burges, N. A., Chater, A. O., Edmonson, J. R., Heywood, V. H., Moore, D. M., Valentine, D. H., Walters, S. M., and Webb, D. A., eds.), Cambridge: Cambridge University Press. Webb, D. A. and Gornall, R. J. (1989) “Saxifrages of Europe”. London: Christopher Helm. Weising, K. and Gardner, R. (1999) A set of conserved PCR primers for the analysis of simple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous angiosperms. Genome 42, 9 – 19. Yeh, F. C., Yang, R. C., and Boyle, T. (1997) POPGENE, Version 1.21. CIFOR and University of Alberta, Canada. F. Grassi et al. F. Grassi Botanical Garden Department of Biology University of Milan Via Celoria 26 20133 Milano Italy E-mail: fabrizio.grassi@unimi.it Editor: M. Koornneef