Populations of introduced species in their new environments are expected to differ from native populations, due to processes such as genetic drift, founder effects and local adaptation, which can often result in rapid phenotypic change. Such processes can also lead to changes in the genetic structure of these populations. This study investigated the populations of Rhododendron ponticum in its introduced range in Ireland, where it is severely invasive, to determine both genetic and flower width diversity and differentiation. We compared six introduced Irish populations with two populations from R. ponticum’s native range in Spain, using amplified fragment length polymorphism and simple sequence repeat genetic markers.We measured flower width, a trait that may affect pollinator visitation, from four Irish and four Spanish populations by measuring both the width at the corolla tip and tube base (nectar holder width). With both genetic markers, populations were differentiated between Ireland and Spain and from each other in both countries. However, populations displayed low genetic diversity (mean Nei’s genetic diversity = 0.22), with the largest proportion (76-93 %) of genetic variation contained within, rather than between, populations. Although corolla width was highly variable between individuals within populations, tube width was significantly wider (>0.5 mm) in introduced, compared with native, populations. Our results show that the same species can have genetically distinct populations in both invasive and native regions, and that differences in floral width may occur, possibly in response to ecological sorting processes or local adaptation to pollinator communities.
Sampling for genetic analysis was carried out in six Irish populations (Table 1), chosen to cover the geographic range of R. ponticum within the country, including the west coast (County Galway), the south-west (County Kerry) and the east (County Dublin). Irish populations were relatively large (>100 adult plants, Table 1). In addition, two Spanish populations were sampled within the Parque Natural Los Alcornocales (∼5 km inland from the Strait of Gibraltar). These populations were sampled in 2002; they were the largest populations in the Los Alcornacales region, but were still comparatively small (18 and 27 adult plants per population). All of the Spanish populations occur within an ∼50 × 30 km area, and are mostly confined to the Aljibe Mountains, where they are restricted to riparian forest habitats (Mejías et al. 2007). Nine to 12 individual plants within both introduced and native populations were randomly selected from each population (Table 1). To avoid sampling clones, distinct individuals, separated by >5 m, were selected. We used this sampling procedure as previous work has shown that vast majority of pollinator visits occur within-plant and that the majority of seeds land close to maternal plants (Stephenson et al. 2007; Stout 2007b). In addition, this sampling procedure ensured that replicate samples were taken in the native range to compare with invasive populations. Rhododendron ponticum populations used for genetic analysis and genetic diversity estimates within populations using (i) AFLP markers and (ii) SSR markers. Size, approximate number of mature, flowering plants in a population; N, number of individuals analysed; Tb, total number of bands; Pb, number of private bands; P, percentage of polymorphic loci at the 5 % level; Hj, Nei’s genetic diversity; Na, observed allele number; Ne, effective allele number; HO, observed heterozygosity; HE, expected heterozygosity; H, average heterozygosity. Leaf material was collected and stored in silica gel (Chase and Hills 1991). DNA was extracted from ∼0.1 g of dried material using a modified 2× hexadecyltryltrimethyl-ammonium bromide procedure (Doyle and Doyle 1987; Hodkinson et al. 2007), and was purified with JetQuick columns (GENOMED Gmbh) according to the manufacturer’s protocol. Two polymerase chain reaction (PCR)-based methods were employed to assess genetic diversity: AFLPs (Vos et al. 1995) and nuclear microsatellites—simple-sequence repeats (SSRs). Sampled DNA was restricted with the endonucleases EcoRI and MseI and ligated to appropriate double-stranded adapters according to the manufacturer’s protocols. Amplified fragment length polymorphism analysis was performed according to the AFLP plant mapping protocol of Applied Biosystems, Inc. Two steps of amplification followed: a pre-selective amplification using primer pairs with one selective base was followed by a selective amplification to further reduce the number of fragments. For the second amplification, the following three selective primer pairs were selected sequentially: EcoRI-ACA/MseI-CAG, EcoRI-AAG/MseI-CTC and EcoRI-AGC/MseI-CAG. The products were sized using an Applied Biosystems 310 Genetic Analyzer with GeneScan version 3.1 and Genotyper version 3.7 software. Amplified fragment length polymorphism profiles were manually scored with the presence of each peak recorded as ‘1’ and the absence of a peak as ‘0’. Only peaks ranging from 50 to 500 bp were scored. A peak was scored as present if it was separated by at least 1 bp and has a relatively high peak height threshold (Meudt and Clarke 2007). In order to reduce genotyping error, AFLP profiles were scored at least twice by individuals with no knowledge of the origin of plant material. No SSR markers have been published for R. ponticum, and so nuclear SSR amplification of seven polymorphic loci isolated from R. metternichii var. hondoense was screened according to the methods described in Naito et al. (1998), of which four were informative for R. ponticum (RM3D2, RM2D2, RM9D6 and RM2D5). Polymerase chain reaction amplification followed (Naito et al. 1998), and the amplicons were sized on an Applied Biosystems 310 Genetic Analyzer with GeneScan version 3.1 and Genotyper version 3.7 software. In addition to quantifying genetic variation in native and invasive populations of R. ponticum, in 2011 we quantified floral width in representative plants in both regions to test whether nectar holder width varied between populations and the two regions. To estimate floral width, two measurements were made on each flower in the field using dial callipers (Moore and Wright, CDP150M), with a precision of 0.01 mm: (i) the width of the corolla at the widest point between the upper wing petals and (ii) the width of the corolla tube at the base. These traits were measured as they represent the extent to which R. ponticum flowers are open to insect pollinators in order to access nectar rewards. Due to logistical constraints, and the fact that these data were collected separately from the leaf material for the population genetic study, only relatively few measurements were taken per population and in only one of the populations (El Palancar, Spain) sampled for genetic analysis. Measurements were made in four Irish and four Spanish populations (Table 2). From each population, five completely open flowers (third floral phase, i.e. with corolla wide open, stigma receptive and protruding beyond anthers; Mejías et al. 2002) from each of five individual plants were randomly selected for measurement. Rhododendron ponticum populations used for flower morphology measurements. For the AFLP data set, genetic diversity estimates were calculated with AFLPsurv V.1.0 (Vekemans et al. 2002). To estimate allelic frequencies, the Bayesian method with a non-uniform prior distribution of allele frequencies (Zhivotovsky 1999) was used. Due to the mixed mating system of the species (Stout 2007a), we assumed some deviation (FIS = 0.1) from the Hardy–Weinberg equilibrium. Statistics of gene diversity were calculated according to Lynch and Milligan (1994). For each population, we calculated the proportion of polymorphic loci (P) and Nei’s gene diversity (Hj). For the SSR data set, GenAlEx 6.2 (Peakall and Smouse 2006) was used to test for departures from the Hardy–Weinberg equilibrium. Observed heterozygosity (HO) and Nei’s expected heterozygosity (HE) were calculated with GenAlEx 6.2, and the average heterozygosity was calculated with PopGene 1.32 (Yeh et al. 2000). Euclidean pairwise genetic distances were calculated in GenAlEx 6.2, which allows a common pathway for subsequent statistical analysis for both dominant AFLP markers and codominant SSR markers (Maguire et al. 2002). For both data sets, genetic distances were calculated using Eq. (1), where n is the total number of polymorphic bands and 2nxy is the number of markers shared by two individuals (Peakall et al. 1995; Maguire et al. 2002). Total genetic diversity was partitioned among groups of populations, among populations within groups and within populations using a hierarchical analysis of molecular variance (AMOVA) in GenAlEx 6.2. Genetic structure was tested with AMOVA on the genetic distance matrix (9999 permutations) produced for both sets of markers (Weir 1996). Analysis of molecular variance output nomenclature follows that of Excoffier et al. (1992) in that variation was summarized both as the proportion of the total variance and as φ-statistics (FST analogues). Pairwise genetic distances among populations and their level of significance for both the AFLP and SSR markers were also obtained from the AMOVA (9999 permutations). In addition, a non-hierarchical AMOVA was performed to test population differentiation in Ireland and Spain separately. Unweighted pair group method with arithmetic mean cluster analysis (UPGMA) was performed in PopGene 1.32 using Nei’s genetic distance (Nei 1972) to analyse the patterns of population-level genetic distances across all populations for both the AFLP and SSR data sets. A Mantel test was used to compare pairwise genetic differences from the AFLP and SSR data. Corolla width and tube width were compared between regions (Ireland and Spain), among populations within regions and among plants within populations, using hierarchical (nested) ANOVA (with ‘region’, Ireland or Spain, as a fixed factor, ‘population’ nested within the region as a random factor and ‘plant’ nested within the population as a random factor; n = 5). Analyses were conducted using GMAV5 for Windows (University of Sydney, Australia). Data were tested for heterogeneity of variances using Cochran’s test prior to analysis (P = 0.0976 and 0.0978 for corolla and tube width data, respectively) and were not transformed. Post-hoc Student–Newman–Keuls (SNK) tests were used to determine which means differed from each other (using the standard threshold of significance α = 0.05).
DIMA AI Care
