TY - JOUR
AU - Benko-Iseppon, Ana, M
AB - Abstract In the Atlantic Rainforest located north of the São Francisco River (northeast Brazil), the humid enclaves called brejos de altitude play a significant role in the diversity dynamics of local flora and fauna. The related species Dyckia pernambucana and D. limae (Bromeliaceae) are characterized by their narrow endemic occurrence in such brejos, and their species status remains unclear. In order to understand the species delimitation in those assigned taxonomic entities, patterns of gene flow and genetic variability were calculated using nuclear and plastid microsatellites and AFLP markers. In this regard, we collected populations of the Pernambuco complex (D. limae and D. pernambucana, from the Borborema Plateau) and the closest relative D. dissitiflora (from the northern Espinhaço Range). Genetic diversity was moderate, despite the possible influence of genetic drift and selfing rates. Dyckia limae could not be undoubtedly discriminated from the remaining populations of Pernambuco, and we propose the synonymization of these species. Thus, the conservation of D. limae as a formerly single species would not reflect the conservation of the minimal gene pool of the studied lineage (D. pernambucana). We also propose the revalidation of the conservation status of this species, endemic to anthropomorphic island-like mountains environments. Finally, the associations found here were consistent with the historical patterns of colonization and fragmentation of the Atlantic Rainforest. AFLPs, gene flow, microsatellites, Pitcairnioideae, species delimitation, taxonomy INTRODUCTION Among several biogeographical units that comprise the Atlantic Forest domain, the north-eastern Atlantic Forest is considered as one of the most important centres of endemism due to its historical association with other Brazilian vegetation types, such as the Amazon Forest and semi-arid Caatinga (Prance, 1982; Tabarelli & Santos, 2004). In this part of Atlantic Forest, located to the north of the São Francisco River, on the Borborema Plateau, the humid enclaves called brejos de altitude (or continental islands, granitic rocky outcrops or inselbergs) play an important role in the diversity of the local flora and fauna composition, because of their unique ecological conditions, spatial isolation and demarcation (Tabarelli & Santos, 2004; Porembski, 2007). Thus, this vegetation type found in the states of Pernambuco, Paraíba, Alagoas, Rio Grande do Norte and Ceará, is thought of as ‘exceptional areas’ in the semi-arid region, also comprising the Caatinga. The brejos de altitude occur at elevations from 600 to 1200 m and higher levels of precipitations are found in these inselbergs than in the surrounding Caatinga. Therefore, they are considered as ecological intermediate areas between Caatinga and Atlantic Forest, with specific vegetation, humidity and microclimates, and are recognized as biodiversity refuges (Lins, 1989; Tabarelli & Santos, 2004). Bromeliaceae have undergone remarkable adaptive radiation in the Neotropics, displaying key innovations, including crassulacean acid metabolism (CAM), leaf succulence, tank habit and presence of trichomes. In Dyckia Schult & Schult.f. (subfamily Pitcairnioideae), CAM and succulence are present. The spiny leaves often have an indumentum that helps to reduce transpiration. All these adaptations allow the plants to succeed in rocky semi-arid environments (Givnish et al., 2014; Silvestro, Zizka & Schulte, 2014; Palma-Silva et al., 2016; Mota et al., 2018). Also, it is well known that events of hybridization and reticulate evolution have played an important role in the process of diversification in Bromeliaceae; these often present a challenge to phylogenetic approaches and species delimitation (Krapp et al., 2014; Barfuss et al., 2016; Palma-Silva et al., 2016; Pinangé et al., 2016; Goetze et al., 2017b). Dyckia (Bromeliaceae) is one of the largest groups of subfamily Pitcairnioideae (173 described species; Butcher & Gouda, 2019) and is distributed across eastern South America; c. 80% of the species are endemic to Brazil. Dyckia spp. generally inhabit xerophytic, terrestrial or rupicolous habitats where their specimens thrive in campos rupestres of Cerrado and Caatinga and in the Atlantic Forest domain (Smith & Downs, 1974). Dyckia is known for its high degree of intraspecific morphological plasticity (Leme, Ribeiro & Miranda, 2012), making species delimitation and phylogenetic reconstruction difficult. Phylogenetic studies based on distribution, plastid and nuclear data have revealed geographically associated clades rather than taxonomic assignments, since Dyckia is known for its continuous morphological variation and geographical distribution pattern, characterized by several endemic and micro-endemic species, as well as a recent origin and diversification (c. 4.0 Mya), especially for the species from north-eastern Brazil (c. 2.5 Mya) (Krapp et al., 2014; Pinangé et al., 2016). The endangered species Dyckia pernambucana L.B.Smith and D. limae L.B.Smith are characterized by their narrow endemic occurrence in the brejos de altitude of Pernambuco (Siqueira-Filho, 2004). Fabricante et al. (2014), in their work regarding the categorization of distribution of D. limae individuals, reported an extremely restricted range of occurrence of 15 km2, and the area of occupancy was estimated at 5 km2, restricted to the rocky outcrops of the Catimbau National Park (Caatinga domain). Further, according to Siqueira-Filho & Leme (2006), and based on morphological features, these species are closely related, with small differences in the inflorescences (shorter in D. limae), leaves (arcuate vs. straight and unilaterally recurved) and spines on the leaf blades (longer and partially retrorse in D. limae). Therefore, there are overlapping characters used in species diagnosis between those two formerly described species, also determined by continuous rather than discrete morphological variation (Siqueira-Filho & Leme, 2006; pers. obs.). These morphological features allow us to raise some questions about the correct taxonomic assignment of populations and/or species. Ornithophily has been reported for both Dyckia spp., with records for the same hummingbird species (Chlorostilbon aureoventris and Chrysolampis mosquitos) as the main pollinators for both species (Machado & Lopes, 2004; A. M. Wanderley, personal communication; D. S. Pinangé, personal observation). The winged seeds do not seem to be effectively dispersed by gravity and wind among populations, as indicated for other Dyckia spp. (Rogalski & Reis, 2009; Rogalski et al., 2009). Dyckia dissitiflora Schult. & Schult.f. is a rupicolous or terrestrial species widely distributed in the rocky outcrops of the Diamantina Plateau, north-eastern Brazil, in the Caatinga Domain (BFG, 2018). This species shows a morphological relationship with the species that make up the Pernambuco complex (PC) (hereafter PC populations), sharing similarities in the main descriptive morphological features, such as habit, leaves, inflorescence, bracts, form and colour of the petals (Smith & Downs, 1974; M. G. Wanderley, pers. comm.). Indeed, Pinangé et al. (2016) in their phylogenetic reconstruction, confirmed the association, with these species appearing in the same clade as D. dissitiflora. The understanding of the mechanisms that have shaped evolutionary and biogeographic histories of natural populations in such particular environments can be achieved through studies of genetic variability. Hence, molecular markers can provide an unbiased approach to quantify genetic diversity levels within and between species (Hamrick, 1994; Hamrick & Godt, 1996; Frankham, Ballou & Briscoe, 2010). In plants, the evolutionary dynamics between and among populations is a result of the intrinsic effects of pollen and seed dispersal, which are always related to the colonization history of both populations and species (Chapman, Parh & Oraguzie, 2000). Genetic structure analyses among populations have been reported for some bromeliads (Cavallari et al., 2006; Barbará et al., 2007, 2008a, b; Palma-Silva et al., 2009, 2011; Hmeljevski et al., 2011; Hmeljevski, Reis & Forzza, 2015; Souza-Sobreira et al., 2015; Goetze et al., 2016, 2017a, b; Gonçalves-Oliveira et al., 2017; Rogalski et al., 2017), and these previous data provided significant insights into genetic patterns of species with varying breeding systems and life histories and a notable genetic differentiation between individual inselbergs, suggesting low levels of gene flow as a result of an isolation effect provided by those insular habitats. Despite the existence of some previous population analyses in Bromeliaceae, little is known about infrageneric relationships, genetic structure, gene flow patterns and mechanisms of speciation in the large genus Dyckia. Using allozyme markers, Hmeljevski et al. (2011) and Rogalski et al. (2017) conducted the most comprehensive population genetic study in two rare Dyckia spp. (D. ibiramensis Reitz and D. brevifolia Baker, respectively) so far. The authors revealed significant population structuring, founder effects and the occurrence of genetic drift as critical events in genetic structuring of the analysed populations. In addition, Gonçalves-Oliveira et al. (2017) performed the only available population genetic study so far on sympatric inselberg populations of Encholirium spectabile Mart. ex Schult.f., closely related to Dyckia (Forzza & Wanderley, 1998; Krapp et al., 2014; Pinangé et al., 2016). This species displays a much higher population distribution when compared with the narrow endemic Dyckia spp. studied here. The authors revealed a considerable level of genetic diversity and a significant genetic structuring among populations especially for the plastid markers employed, indicating a higher efficiency of gene flow through pollination as compared with anemochorous seed dispersal. Here, we provide the first population genetic survey on species cohesion involving the populations of the PC (D. pernambucana and D. limae), and the closely related D. dissitiflora, analysing patterns of structure and genetic diversity and comparing indices from multilocus fingerprinting DNA, co-dominant nuclear and plastid markers to address three main hypotheses. (1) The genetic structure in the PC corresponds primarily to geographical association, not to the taxonomic concept, pattern observed in the only phylogenetic studies available to date (Krapp et al., 2014; Pinangé et al., 2016) involving Dyckia (including the aforementioned species). (2) Although there is a morphological and phylogenetic association with the species of the PC, D. dissitiflora is a distinct lineage. (3) Gene flow among species populations is hindered by island-like distributions, and anemochorous seed dispersal is less efficient than hummingbird pollination for PC populations in this respect. Thus, the recently diversified related PC populations form one biological entity with historical gene flow. In view of the above, our main premise is to verify whether morphologically identified species in PC are genetically differentiated, providing some insights into species delimitation. Finally, evolutionary implications in the life history of the populations the quantification of genetic diversity across the populations in the light of propositions for priority conservation areas are also discussed. MATERIAL AND METHODS Population sampling and DNA extraction Populations of the three morphologically related saxicolous Dyckia spp. that are endemic to outcrops (D. pernambucana, D. limae and D. dissitiflora) in north-eastern Brazil were sampled (Supporting information, Appendix S1; Fig. 1). Many Dyckia spp. are rare and narrow endemics, with low population densities (Smith & Downs, 1974), and we attempted to cover the area of occurrence as completely as possible. Thus, individuals of PC populations were randomly sampled on five inselberg; we collected 50 samples in total (one and four populations of D. limae and D. pernambucana, respectively), with ten individuals per population (Supporting Information, Appendix S1). According to Leme et al. (2012), most Dyckia spp. are prone to vigorous clonal propagation. After an intensive search at the perimeter of each population, we observed that the majority of the populations had limited distributions. Thus, the sampling strategy involved the collection of a single sample from each clump, to avoid the collection of several ramets from the same genet (reflecting the limited population size). As mentioned above, D. limae is restricted to a sedimentary rock formation in the Catimbau National Park in the São Francisco River basin (Fabricante et al., 2014). Thus, the effort of sample collection comprised the already mentioned narrowly distributed species, in terms of areas of occurrence and occupancy (Fabricante et al., 2014). On the other hand, D. pernambucana has somewhat a larger distribution range on granitic rock outcrops on the remaining inselbergs in Pernambuco. Thus, the populations of these assigned species do not occur in sympatry (Supporting Information, Appendix S1; Fig. 1). Figure 1. Open in new tabDownload slide A, Elevational map showing the collections sites of Dyckia dissitiflora populations, located in the rocky outcrops of the Chapada Diamantina, and representatives of the Pernambuco complex (PC) populations with their plastid DNA haplotype frequencies. Pie charts are shown when a population possesses more than one haplotype (except for ‘Srita’ population), and different colours represent different haplotypes. B, Median-joining network among plastid DNA haplotypes. Circle sizes correspond to the relative frequency of a particular haplotype in the total. Line size between haplotypes represents mutation events, according to the scale. PC populations comprise D. pernambucana and D. limae from the Borborema Plateau, Pernambuco, Brazil. Figure 1. Open in new tabDownload slide A, Elevational map showing the collections sites of Dyckia dissitiflora populations, located in the rocky outcrops of the Chapada Diamantina, and representatives of the Pernambuco complex (PC) populations with their plastid DNA haplotype frequencies. Pie charts are shown when a population possesses more than one haplotype (except for ‘Srita’ population), and different colours represent different haplotypes. B, Median-joining network among plastid DNA haplotypes. Circle sizes correspond to the relative frequency of a particular haplotype in the total. Line size between haplotypes represents mutation events, according to the scale. PC populations comprise D. pernambucana and D. limae from the Borborema Plateau, Pernambuco, Brazil. As mentioned before, D. dissitiflora was included in the analysis due to its taxonomic association with the species of the PC populations and for a better hypothesis test regarding species delimitation. Therefore, we have not included all the known distribution (from the Diamantina Plateau, in the northern Espinhaço Range); instead we collected samples from three different populations/sites comprising 37 individuals (Supporting Information, Appendix S1). Considering the entire sampling, young leaves of 87 individuals were stored in sodium chloride saturated aqueous solution of cetyl–trimethylammonium-bromide (20 g CTAB/L) following Tel-Zur et al. (1999), until DNA extraction. Herbarium vouchers were collected and deposited at the Herbarium da Universidade Federal de Pernambuco (UFP) (Supporting Information, Appendix S1). Total genomic DNA was isolated, using the CTAB procedure of Doyle & Doyle (1987), with modifications as described by Weising et al. (2005). AFLP analysis The AFLP fingerprinting was performed according to the original protocol of Vos et al. (1995) and Debener & Mattiesh (1999) with minor changes. In brief, 30 ng genomic DNA was digested in a final volume of 25 µL at 37 °C with simultaneous restriction using endonucleases HindIII and MseI and ligation of adaptors for 12 h. The pre-selective and selective amplifications were carried out using primers with one (+1), two (+2) or three (+3) selective nucleotides at their 3′ ends. The pre-selective PCR reactions contained 2 µL 1:10 diluted restriction-ligation product, 0.5 µM unlabelled MseI +1 primers, 1 µL 10× PCR buffer (Peqlab blue), 25 mM MgCl2, 0.2 mM each deoxynucleoside triphosphate (dNTP) and 0.025 U Taq polymerase (SawadyTaq, Peqlab, Germany). The reaction mixtures were subjected to 20 cycles of PCR amplification, each consisting of 94 °C for 20 s, 56 °C for 30 s and 72 °C for 2 min. Finally, the final extension was at 72 °C followed by 60 °C for 30 min. Six primer combinations were used in the selective PCR (ACA/ACC, AAC/ACA, AAC/ATC, AGC/ATC, AAC/CAG, AGC/CTA) as follows: 2.5 µL of the 1:20 diluted pre-amplification product and different combinations of the MseI (+3) primer (Carl Roth, Karlsruhe, Germany) (0.25 µM) and the fluorescence-labelled HindIII (+3) primer (WellRED- D2, -D3, -D4, Sigma Aldrich, Munich, Germany) (0.05 µM), were performed. Final products of the selective PCR were run on an automated sequencer (CEQ8800, Beckman Coulter, Krefeld, Germany) as a multiplex of three primer combinations labelled and an internal size standard (GenomeLab DNA Size Standard Kit 600, Beckman Coulter, Krefeld, Germany). The AFLP amplification failed for three samples, even after repetitions of the procedure. Thus, these samples were excluded and the final sampling comprised 87 individuals. The AFLP banding patterns were scored semi-automatically as presence or absence using the software Genemarker 1.9 (SoftGenetics, State College, PA, USA). The reproducibility test was also performed with c. 15% of the total data set selected randomly for this purpose. The intensity of each individual peak was normalized on the basis of a fixed threshold, with the cut-off at < 10% of the second highest signal intensity. Microsatellite analysis Fifteen nuclear microsatellite loci (Table 3), previously characterized for D. marnier-lapostollei L.B.Sm (Wöhrmann et al., 2012b), were selected for the population analysis, using either unlabelled or labelled fluorescent primer (forward or reverse), with an IRDye700 or IRDye800 label as detailed by Wöhrmann et al. (2012a). To analyse maternal inheritance, all individuals were screened for variation at eight (SSRL01, SSRL04, SSRL06, SSRN04, SSRN05, SSRN10, SSRN11 and SSRN18) plastid SSR markers, previously developed for D. marnier-lapostollei (Krapp et al., 2012). The PCR amplifications were carried out using an indirect labelling fluorescence procedure, with the labelled forward primer IRDye700-M13 (Schuelke, 2000), as also described by Krapp et al. (2012). The PCR products were visualized on an automated sequencer (Li-Cor, 4200 IR2, Li-Cor Biosciences, Bad Homburg, Germany), as detailed in Wöhrmann et al. (2012a). The scoring procedure was determined by visual examination of the fragment sizes, using an external size standard from sequences of the AT-rich psbL-trnS region of Macaranga indistincta Whitmore that provides only the T reaction from the sequencing procedure (for details, see Guicking et al., 2008). Thus, the identification and assignment of alleles were achieved from the resulting sets of T fragments with known sizes that were run in every sixth or seventh lane of the gel, as described by Wöhrmann et al. (2012a). Data analysis: genetic diversity and F-statistics To avoid putative scoring errors, the AFLP, nuclear and plastid SSR raw data matrices were double-checked. For the AFLP assays, the percentage of polymorphic loci, unbiased estimates of genetic diversity (Hj, analogous to He) and the genetic differentiation with statistical significance (FST) among species/populations, were calculated using the software AFLP-SURV v.1.0 (Vekemans et al., 2002) and obtained by means of 5000 random permutations. Therefore, the allelic frequencies of AFLP fragments were evaluated using the Bayesian approach for diploid species proposed by Zhivotovsky et al. (1999). This program was designed to use the approach of Lynch & Milligan (1994) to calculate population genetic parameters on the basis of the expected heterozygosity of dominant marker loci. The analysis of molecular variance (AMOVA) was performed, using the program Arlequin v.3.5 (Excoffier, Laval & Schneider, 2005), to estimate variance components, partitioning of the variation of each species/individual among populations with significance tests of 10 000 permutations. Regarding the nuclear SSR loci (nSSRs), the number of polymorphic loci (P), number of alleles, heterozygosity measurements (HO and HE) and the variance in allele length (Var) were measured by using the software Microsatellite Analyser (MSA; Dieringer & Schlötterer, 2003). The software Genepop (Raymond & Rousset, 1995; Rousset, 2008) was used to estimate the inbreeding coefficient FIS (Weir & Cockerham, 1984) and the deviation from Hardy–Weinberg equilibrium (HWE). The global and pairwise index of genetic differentiation (FST) was estimated to infer the degree of population subdivision using MSA software, and resampling with 10 000 permutations to test its significance. The software Arlequin was also employed in the nSSR data set in order to evaluate the molecular variance (AMOVA). Furthermore, we performed a test using Bottleneck v.1.2.02 software (Cornuet & Luikart, 1997) to verify whether the populations have undergone a putative loss of genetic variation, looking for evidence of excess heterozygosity relative to allele numbers, as a result of recent reductions in effective population size. Infinite alleles (IAM), stepwise mutation (SMM) and two-phase mutation (TPM) models were used for these tests. The significance of the genetic diversity excess (observed heterozygosity, Ho > expected heterozygosity under the model, He) was calculated using the Wilcoxon test, since it is more suitable for analysis containing less than 20 loci (Piry, Luikart & Cornuet, 1999), based on 10 000 replications. For the plastid data (plastid SSRs), length variants at each locus were combined into haplotypes (Fig. 1). We calculated the effective number of haplotypes (NE) and the number of private haplotypes (Hpr). Additionally, the haplotypic richness (R_h) was also estimated using the rarefaction method (El Mousadik & Petit, 1996) and the unbiased Nei’s index of gene diversity (HE) (Nei, 1973). Furthermore, the average genetic distance among individuals within datasets was estimated using the value of Dsh2 (Goldstein et al., 1995) based on the averaged square sum of all length differences at microsatellites. All these indices were calculated by using the program Haplotype-Analysis v.1.05 (Eliades & Eliades, 2009). In addition, a median-joining (MJ) (Bandelt, Forster & Röhl, 1999) network demonstrating the genetic relationships among the haplotypes were estimated with Network v.5.0.1.1 (http://www.fluxus-engineering.com). Arlequin was employed in order to estimate the patterns of nuclear and plastid DNA differentiation in hierarchical models and to evaluate the global and pairwise index of genetic differentiation (FST), for plastid SSR markers. Isolation by distance (Wright, 1965) was verified by calculating the correlation between geographical and genetic differentiation matrices, with the application of Mantel test (Sokal & Rohlf, 1995), using 10 000 randomizations to determine significance. With the aim to distinguish the contributions between pollen and seed dispersal, we conducted a pollen/seed migration ratio, calculated from nuclear and plastid data, as proposed by Ennos (1994). Genetic distances The AFLP data matrix was used to estimate genetic distances among individuals and populations. Jaccard’s dissimilarity coefficient was calculated for the binary using the software DARwin v.5.0 (Perrier & Jacquemoud-Collet, 2006). Dissimilarity coefficients were used for clustering analysis based on the weighted neighbor-joining (WNJ) method, with bootstrap (Felsenstein, 1985) of 1000 replications. Regarding the nSSR data, we performed a genetic distance-based analysis using Poptreew (Takezaki, Nei & Tamura, 2014). Thus, the Dsw distance coefficient (Shriver et al., 1995), only applicable to microsatellite data, was employed (for each individual) and the neighbor-joining method (Saitou & Nei, 1987) was used for cluster analysis with 1000 bootstrap replications. One problem of haplotype analysis followed by the network approach for plastid SSR data is that the stepwise mutation model (SMM) is employed only to the point of the haplotype’s designation. On the other hand, an evolutionary population tree construction employed by Poptreew would fully take into account the SSM. Thus, we performed an evolutionary analysis of plastid allele frequency data (for each individual) focusing on the PC populations, in the mentioned software. Therefore, the Dsh2 distance coefficient, suitable only for microsatellite DNA data (Goldstein et al., 1995), was calculated, whereas the cluster analysis was performed using neighbor-joining methods. Bayesian admixture analysis A Bayesian approach was used with both nuclear markers (AFLP and nSSR) to assign each individual to one of the most appropriate genetic clusters (K-value) and to estimate admixture proportions (Q) for each individual, without considering sampling locations, using the software Structure v.2.2 (Pritchard, Stephens & Donnelly, 2000; Falush, Stephens & Pritchard, 2003). The analyses were performed under the admixture model, assuming correlated allele frequencies. Each run had 800 000 iterations with a burn-in period of 250 000, and ten repetitions for each K. To determine the most likely number of clusters (K), we used the method proposed by Evanno, Regnaut & Goudet (2005) based on second-order rate of change regarding to K of the likelihood function (∆K), with the online program Structure Harvester v.0.6.94 (Earl & von Holdt, 2012). For nSSRs, Structure analysis was performed, including all samples (N = 87). RESULTS Genetic diversity The six AFLP primer combinations resulted in 340 non-monomorphic markers, of which 307 markers were present in PC populations and 323 in D. dissitiflora populations. The proportion of polymorphic loci ranged from 69.7% (‘Brejo’) to 84.5% (‘Morrao’) with the populations of D. dissitiflora exhibiting higher average values of polymorphic loci than the PC populations (Table 1). Table 1. Comparison of genetic diversity and global differentiation among populations from the Pernambuco complex (D. pernambucana and D. limae) and D. dissitiflora based on 340 AFLP markers Populations . N . TNL . PL . PPL . HJ . SE (HJ) . HT . HW . Hb . Global FST . ‘Cat’ 09 307 227 73.9 0.22091 0.00938 0.277 0.228 0.054 0.196 ‘Brejo’ 08 307 214 69.7 0.19118 0.00939 ‘Pesq’ 10 307 218 71.0 0.23320 0.00930 ‘Pico’ 10 307 239 77.9 0.23921 0.01083 ‘Srita’ 10 307 237 77.2 0.22887 0.00977 ‘Morrao’ 09 323 273 84.5 0.30315 0.00915 0.304 0.285 0.018 0.060 ‘Cach’ 17 323 236 73.1 0.27105 0.00987 ‘Lajes’ 09 323 264 81.7 0.28217 0.00950 Populations . N . TNL . PL . PPL . HJ . SE (HJ) . HT . HW . Hb . Global FST . ‘Cat’ 09 307 227 73.9 0.22091 0.00938 0.277 0.228 0.054 0.196 ‘Brejo’ 08 307 214 69.7 0.19118 0.00939 ‘Pesq’ 10 307 218 71.0 0.23320 0.00930 ‘Pico’ 10 307 239 77.9 0.23921 0.01083 ‘Srita’ 10 307 237 77.2 0.22887 0.00977 ‘Morrao’ 09 323 273 84.5 0.30315 0.00915 0.304 0.285 0.018 0.060 ‘Cach’ 17 323 236 73.1 0.27105 0.00987 ‘Lajes’ 09 323 264 81.7 0.28217 0.00950 N, number of individuals per population; PL, polymorphic loci; TLN, total number of loci; PPL, percentage of polymorphic loci; HJ, analogous expected heterozygosity (Zhivotvski, 1999); HT, total diversity; HW, average diversity within-population; Hb, average diversity between populations, FST, global fixation index. Open in new tab Table 1. Comparison of genetic diversity and global differentiation among populations from the Pernambuco complex (D. pernambucana and D. limae) and D. dissitiflora based on 340 AFLP markers Populations . N . TNL . PL . PPL . HJ . SE (HJ) . HT . HW . Hb . Global FST . ‘Cat’ 09 307 227 73.9 0.22091 0.00938 0.277 0.228 0.054 0.196 ‘Brejo’ 08 307 214 69.7 0.19118 0.00939 ‘Pesq’ 10 307 218 71.0 0.23320 0.00930 ‘Pico’ 10 307 239 77.9 0.23921 0.01083 ‘Srita’ 10 307 237 77.2 0.22887 0.00977 ‘Morrao’ 09 323 273 84.5 0.30315 0.00915 0.304 0.285 0.018 0.060 ‘Cach’ 17 323 236 73.1 0.27105 0.00987 ‘Lajes’ 09 323 264 81.7 0.28217 0.00950 Populations . N . TNL . PL . PPL . HJ . SE (HJ) . HT . HW . Hb . Global FST . ‘Cat’ 09 307 227 73.9 0.22091 0.00938 0.277 0.228 0.054 0.196 ‘Brejo’ 08 307 214 69.7 0.19118 0.00939 ‘Pesq’ 10 307 218 71.0 0.23320 0.00930 ‘Pico’ 10 307 239 77.9 0.23921 0.01083 ‘Srita’ 10 307 237 77.2 0.22887 0.00977 ‘Morrao’ 09 323 273 84.5 0.30315 0.00915 0.304 0.285 0.018 0.060 ‘Cach’ 17 323 236 73.1 0.27105 0.00987 ‘Lajes’ 09 323 264 81.7 0.28217 0.00950 N, number of individuals per population; PL, polymorphic loci; TLN, total number of loci; PPL, percentage of polymorphic loci; HJ, analogous expected heterozygosity (Zhivotvski, 1999); HT, total diversity; HW, average diversity within-population; Hb, average diversity between populations, FST, global fixation index. Open in new tab Similarly, the Hj value within Pernambuco populations ranged from 0.191 (‘Brejo’) to 0.239 (‘Pico’), with the populations of D. dissitiflora also displaying higher diversity values (Table 1). The global values of diversity were also observed for D. dissitiflora populations, for which the diversity within populations (HW) was higher than in the PC, whereas this latter revealed higher values of diversity between populations (HB) (Table 1). The fraction of polymorphic nuclear SSR loci among species ranged from 66% to 93% and was slightly higher in D. dissitiflora (Table 2). The allelic richness (RS) ranged from 2.183 (‘Brejo’) to 5.333 (‘Morrao’), and observed and expected heterozygosities ranged from 0.127 to 0.481 and 0.260 to 0.653, respectively, also with D. dissitiflora populations exhibiting the highest values (Table 2). Conversely, the highest variation in allele length (Var) was obtained for a population from the PC (‘Pico’) with 206.547, and the smallest value (35.098) was also observed for a population of this group (‘Cat’). Finally, most loci deviated from the HWE through the observation of their FIS estimation (Table 2). Table 2. Indices of genetic diversity in PC populations of Dyckia pernambucana, D. limae and D. dissitiflora for microsatellite markers Population ID . Nuclear microsatellites . . . . . . Plastid DNA . . . . . . P . Rs . Var . Ho . He . FIS (W&C) . Ne . Hpr . Rh . He . Dsh2 . ‘Cat’ a 73 3.350 35.098 0.380 0.391 0.010 1.852 3 1.900 0.511 1.114 ‘Brejo’ b 93 2.183 68.790 0.127 0.260 0.367* 1.724 2 1.000 0.467 0.058 ‘Pesq’ b 66 2.412 74.723 0.253 0.281 0.033* 1.724 2 1.000 0.467 0.058 ‘Pico’ b 93 2.832 206.547 0.253 0.481 0.372* 2.941 4 2.800 0.733 0.167 ‘Srita’ b 66 2.798 193.826 0.180 0.403 0.338* 1.000 1 0.000 0.000 0.000 PC/Mean 78.200 2.715 115.797 0.239 0.363 0.010 2.060 2.750 1.675 0.545 0.349 ‘Morrao’ c 93 5.333 200.044 0.481 0.653 0.091* 3.857 5 5.000 0.833 3.174 ‘Cach’ c 80 3.788 51.540 0.460 0.482 0.142* 1.695 1 1.420 0.433 3.000 ‘Lajes’ c 93 4.467 46.964 0.415 0.525 0.253* 5.400 5 5.000 0.917 2.569 Overall/mean 82.125 3.395 109.692 0.319 0.4345 0.010 2.524 2.875 2.265 0.545 1.268 Population ID . Nuclear microsatellites . . . . . . Plastid DNA . . . . . . P . Rs . Var . Ho . He . FIS (W&C) . Ne . Hpr . Rh . He . Dsh2 . ‘Cat’ a 73 3.350 35.098 0.380 0.391 0.010 1.852 3 1.900 0.511 1.114 ‘Brejo’ b 93 2.183 68.790 0.127 0.260 0.367* 1.724 2 1.000 0.467 0.058 ‘Pesq’ b 66 2.412 74.723 0.253 0.281 0.033* 1.724 2 1.000 0.467 0.058 ‘Pico’ b 93 2.832 206.547 0.253 0.481 0.372* 2.941 4 2.800 0.733 0.167 ‘Srita’ b 66 2.798 193.826 0.180 0.403 0.338* 1.000 1 0.000 0.000 0.000 PC/Mean 78.200 2.715 115.797 0.239 0.363 0.010 2.060 2.750 1.675 0.545 0.349 ‘Morrao’ c 93 5.333 200.044 0.481 0.653 0.091* 3.857 5 5.000 0.833 3.174 ‘Cach’ c 80 3.788 51.540 0.460 0.482 0.142* 1.695 1 1.420 0.433 3.000 ‘Lajes’ c 93 4.467 46.964 0.415 0.525 0.253* 5.400 5 5.000 0.917 2.569 Overall/mean 82.125 3.395 109.692 0.319 0.4345 0.010 2.524 2.875 2.265 0.545 1.268 Abbreviations: P, percentage of polymorphic loci; Rs, allelic richness; Var, variance in allele size; HO, observed heterozygosity; HE, expected heterozygosity; FIS, inbreeding coefficient (Weir & Cockerham, 1984); Ne, effective number of haplotypes; Hpr, number of private haplotypes; Rh, haplotypic richness; HE, genetic diversity; Dsh2, mean genetic distance between individuals. aDyckia limae;bD. pernambucana; cD. dissitiflora. PC, Pernambuco complex. Deviations of intra-population inbreeding coefficients from Hardy–Weinberg Equilibrium (HWE) are denoted by asterisks. Open in new tab Table 2. Indices of genetic diversity in PC populations of Dyckia pernambucana, D. limae and D. dissitiflora for microsatellite markers Population ID . Nuclear microsatellites . . . . . . Plastid DNA . . . . . . P . Rs . Var . Ho . He . FIS (W&C) . Ne . Hpr . Rh . He . Dsh2 . ‘Cat’ a 73 3.350 35.098 0.380 0.391 0.010 1.852 3 1.900 0.511 1.114 ‘Brejo’ b 93 2.183 68.790 0.127 0.260 0.367* 1.724 2 1.000 0.467 0.058 ‘Pesq’ b 66 2.412 74.723 0.253 0.281 0.033* 1.724 2 1.000 0.467 0.058 ‘Pico’ b 93 2.832 206.547 0.253 0.481 0.372* 2.941 4 2.800 0.733 0.167 ‘Srita’ b 66 2.798 193.826 0.180 0.403 0.338* 1.000 1 0.000 0.000 0.000 PC/Mean 78.200 2.715 115.797 0.239 0.363 0.010 2.060 2.750 1.675 0.545 0.349 ‘Morrao’ c 93 5.333 200.044 0.481 0.653 0.091* 3.857 5 5.000 0.833 3.174 ‘Cach’ c 80 3.788 51.540 0.460 0.482 0.142* 1.695 1 1.420 0.433 3.000 ‘Lajes’ c 93 4.467 46.964 0.415 0.525 0.253* 5.400 5 5.000 0.917 2.569 Overall/mean 82.125 3.395 109.692 0.319 0.4345 0.010 2.524 2.875 2.265 0.545 1.268 Population ID . Nuclear microsatellites . . . . . . Plastid DNA . . . . . . P . Rs . Var . Ho . He . FIS (W&C) . Ne . Hpr . Rh . He . Dsh2 . ‘Cat’ a 73 3.350 35.098 0.380 0.391 0.010 1.852 3 1.900 0.511 1.114 ‘Brejo’ b 93 2.183 68.790 0.127 0.260 0.367* 1.724 2 1.000 0.467 0.058 ‘Pesq’ b 66 2.412 74.723 0.253 0.281 0.033* 1.724 2 1.000 0.467 0.058 ‘Pico’ b 93 2.832 206.547 0.253 0.481 0.372* 2.941 4 2.800 0.733 0.167 ‘Srita’ b 66 2.798 193.826 0.180 0.403 0.338* 1.000 1 0.000 0.000 0.000 PC/Mean 78.200 2.715 115.797 0.239 0.363 0.010 2.060 2.750 1.675 0.545 0.349 ‘Morrao’ c 93 5.333 200.044 0.481 0.653 0.091* 3.857 5 5.000 0.833 3.174 ‘Cach’ c 80 3.788 51.540 0.460 0.482 0.142* 1.695 1 1.420 0.433 3.000 ‘Lajes’ c 93 4.467 46.964 0.415 0.525 0.253* 5.400 5 5.000 0.917 2.569 Overall/mean 82.125 3.395 109.692 0.319 0.4345 0.010 2.524 2.875 2.265 0.545 1.268 Abbreviations: P, percentage of polymorphic loci; Rs, allelic richness; Var, variance in allele size; HO, observed heterozygosity; HE, expected heterozygosity; FIS, inbreeding coefficient (Weir & Cockerham, 1984); Ne, effective number of haplotypes; Hpr, number of private haplotypes; Rh, haplotypic richness; HE, genetic diversity; Dsh2, mean genetic distance between individuals. aDyckia limae;bD. pernambucana; cD. dissitiflora. PC, Pernambuco complex. Deviations of intra-population inbreeding coefficients from Hardy–Weinberg Equilibrium (HWE) are denoted by asterisks. Open in new tab The BOTTLENECK analyses were carried out on our dataset and estimated with different models (IAM, SMM and TPM; Supporting Information, Appendix S2), where the relationship (Ho > He) means that there is an excess of heterozygosity in relation to the heterozygosity in the equilibrium. In this regard, the populations ‘Pico’ and ‘Srita’ from the PC displayed substantial excess of heterozygosity, detected by the Wilcoxon test, with a probability < 0.01. Thus, these data revealed a deviation from mutation-drift equilibrium under two models (IAM and TPM), but not under the SMM, implying a recent bottleneck over the past few generations (Supporting Information, Appendix S2). Likewise, for the D. dissitiflora population ‘Cach’, the same significant deviation (< 0.05) could be notified under the model IAM, but not under SMM and TPM. Conversely, we observed no significant deviation from mutation-drift equilibrium in the remaining populations, suggesting an absence of a recent event of bottleneck (Supporting Information, Appendix S2). Twenty-eight plastid SSR alleles across the whole sampling were found in the present work (ranging from 62 to 100 bp). Briefly, the plastid DNA markers showed that all populations displayed polymorphic sites in their plastid microsatellite markers, except for the ‘Srita’ population that showed a single fixed haplotype; the remaining populations had two to four haplotypes (Table 2; Supporting Information, Appendix S4). The populations ‘Morrao’ and ‘Lajes’ exhibited the highest haplotypic richness (5.000 in each population) and values of genetic diversity (0.833 and 0.917, respectively; Table 2). Variability across loci All 15 nuclear microsatellite loci were variable, with three to 36 alleles, giving a total of 181 alleles (12.0 alleles per locus; N = 87; Table 3). The average for observed and expected heterozygosity ranged from 0.032 to 0.669 and from 0.032 to 0.834, respectively (Table 3). In relation to the F-statistics values, most of the nuclear SSR loci departed from the HWE in the intra-population level (FIS estimates) for the PC populations (D. limae and D. pernambucana). Table 3. Levels of genetic diversity and indices of F-statistics for 15 microsatellite loci (Wörhmann et al., 2012b) in populations of Dyckia limae, D. pernambucana and D. dissitiflora 1  Locus . 2   A . 3  HO . 4  HE . 5  FIT* . 6  FST* . 7  FIS* . ngDy_1 8   09 9  0.294 10  0.401 11  0.156 12  0.142 13  0.016 ngDy_3 14  09 15  0.331 16  0.385 17  0.252 18  0.247 19  0.006 ngDy_8 20  03 21  0.032 22  0.032 −0.013 −0.013 0.000 ngDy_10 05 0.317 0.269 −0.017 −0.017 0.000 ngDy_16 11 0.387 0.485 0.564 0.387 0.289 ngDy_17 14 0.298 0.520 0.628 0.184 0.544 ngDy_22 20 0.417 0.730 0.568 0.325 0.359 ngDy_24 13 0.243 0.375 0.591 0.374 0.346 ngDy_25 11 0.299 0.588 0.693 0.363 0.519 ngDy_27 36 0.669 0.834 0.378 0.147 0.271 ngDy_30 06 0.159 0.357 0.905 0.532 0.797 ngDy_31 11 0.185 0.370 0.732 0.168 0.679 ngDy_32 12 0.259 0.421 0.814 0.471 0.648 ngDy_45 13 0.508 0.520 0.397 0.319 0.114 ngDy_49 08 0.296 0.245 0.080 0.270 −0.259 Mean 12.067 0.313 0.435 0.449 0.260 0.289 1  Locus . 2   A . 3  HO . 4  HE . 5  FIT* . 6  FST* . 7  FIS* . ngDy_1 8   09 9  0.294 10  0.401 11  0.156 12  0.142 13  0.016 ngDy_3 14  09 15  0.331 16  0.385 17  0.252 18  0.247 19  0.006 ngDy_8 20  03 21  0.032 22  0.032 −0.013 −0.013 0.000 ngDy_10 05 0.317 0.269 −0.017 −0.017 0.000 ngDy_16 11 0.387 0.485 0.564 0.387 0.289 ngDy_17 14 0.298 0.520 0.628 0.184 0.544 ngDy_22 20 0.417 0.730 0.568 0.325 0.359 ngDy_24 13 0.243 0.375 0.591 0.374 0.346 ngDy_25 11 0.299 0.588 0.693 0.363 0.519 ngDy_27 36 0.669 0.834 0.378 0.147 0.271 ngDy_30 06 0.159 0.357 0.905 0.532 0.797 ngDy_31 11 0.185 0.370 0.732 0.168 0.679 ngDy_32 12 0.259 0.421 0.814 0.471 0.648 ngDy_45 13 0.508 0.520 0.397 0.319 0.114 ngDy_49 08 0.296 0.245 0.080 0.270 −0.259 Mean 12.067 0.313 0.435 0.449 0.260 0.289 Abbreviations: A, number of alleles per locus; Ho, observed heterozygosity; HE, expected heterozygosity; FIT, overall inbreeding coefficient; FST, fixation index; FIS, inbreeding coefficient. *Values calculated only for the focused PC populations of D. pernambucana and D. limae. Open in new tab Table 3. Levels of genetic diversity and indices of F-statistics for 15 microsatellite loci (Wörhmann et al., 2012b) in populations of Dyckia limae, D. pernambucana and D. dissitiflora 1  Locus . 2   A . 3  HO . 4  HE . 5  FIT* . 6  FST* . 7  FIS* . ngDy_1 8   09 9  0.294 10  0.401 11  0.156 12  0.142 13  0.016 ngDy_3 14  09 15  0.331 16  0.385 17  0.252 18  0.247 19  0.006 ngDy_8 20  03 21  0.032 22  0.032 −0.013 −0.013 0.000 ngDy_10 05 0.317 0.269 −0.017 −0.017 0.000 ngDy_16 11 0.387 0.485 0.564 0.387 0.289 ngDy_17 14 0.298 0.520 0.628 0.184 0.544 ngDy_22 20 0.417 0.730 0.568 0.325 0.359 ngDy_24 13 0.243 0.375 0.591 0.374 0.346 ngDy_25 11 0.299 0.588 0.693 0.363 0.519 ngDy_27 36 0.669 0.834 0.378 0.147 0.271 ngDy_30 06 0.159 0.357 0.905 0.532 0.797 ngDy_31 11 0.185 0.370 0.732 0.168 0.679 ngDy_32 12 0.259 0.421 0.814 0.471 0.648 ngDy_45 13 0.508 0.520 0.397 0.319 0.114 ngDy_49 08 0.296 0.245 0.080 0.270 −0.259 Mean 12.067 0.313 0.435 0.449 0.260 0.289 1  Locus . 2   A . 3  HO . 4  HE . 5  FIT* . 6  FST* . 7  FIS* . ngDy_1 8   09 9  0.294 10  0.401 11  0.156 12  0.142 13  0.016 ngDy_3 14  09 15  0.331 16  0.385 17  0.252 18  0.247 19  0.006 ngDy_8 20  03 21  0.032 22  0.032 −0.013 −0.013 0.000 ngDy_10 05 0.317 0.269 −0.017 −0.017 0.000 ngDy_16 11 0.387 0.485 0.564 0.387 0.289 ngDy_17 14 0.298 0.520 0.628 0.184 0.544 ngDy_22 20 0.417 0.730 0.568 0.325 0.359 ngDy_24 13 0.243 0.375 0.591 0.374 0.346 ngDy_25 11 0.299 0.588 0.693 0.363 0.519 ngDy_27 36 0.669 0.834 0.378 0.147 0.271 ngDy_30 06 0.159 0.357 0.905 0.532 0.797 ngDy_31 11 0.185 0.370 0.732 0.168 0.679 ngDy_32 12 0.259 0.421 0.814 0.471 0.648 ngDy_45 13 0.508 0.520 0.397 0.319 0.114 ngDy_49 08 0.296 0.245 0.080 0.270 −0.259 Mean 12.067 0.313 0.435 0.449 0.260 0.289 Abbreviations: A, number of alleles per locus; Ho, observed heterozygosity; HE, expected heterozygosity; FIT, overall inbreeding coefficient; FST, fixation index; FIS, inbreeding coefficient. *Values calculated only for the focused PC populations of D. pernambucana and D. limae. Open in new tab Regarding the plastid SSRs, the eight loci generated two to six alleles per locus, allowing the identification of 25 individual haplotypes, 23 of which were exclusive to one population (Supporting Information, Appendix S3). For the PC populations alone, the combined genetic information yielded 12 distinct haplotypes, each exclusive to a single population. The populations from ‘Pico’ and ‘Catimbau’ (hereafter ‘Cat’) revealed the highest numbers of haplotypes: four and three, respectively (Fig. 1; Table 2; Supporting Information, Appendix S3; for population information see Supporting Information, Appendix S1). Patterns of genetic differentiation The AFLP markers showed variation among PC populations. However, it was lower than that of SSR analysis, with global FST of 0.196 (Table 1). On the other hand, nSSR loci revealed higher values of divergence among PC populations than AFLP analysis, across all loci with a global FST value of 0.260 (Table 3). Bearing in mind only the PC populations, for nSSR markers a higher value of global FST was found (0.316, Table 5). On the other hand, for plastid SSRs the Arlequin data yielded a remarkable value of genetic structuring, across all loci with a global FST value of 0.838 (considering only the PC populations, Table 5). The haplotype network analysis (also considering only the PC populations) yielded a correlation with a poor resolution, initially with the formation of a loop with parallel step mutation, where haplotypes (H9, H3 and H4) were linked by a median vector (data not shown). Thus, these results suggest a possibly extant of unsampled or extinct haplotypes and can also be related to the influence of amplicon size homoplasy of the plastid SSR loci (Bandelt et al., 1999; Navascués & Emerson, 2005). To resolve the loop, we took into account the topological and geographical criteria proposed by Crandall & Templeton (1993) (Fig. 1). The AMOVA results among the PC populations were highly congruent between the AFLPs and nSSR data sets. Thus, the variation was higher within populations than among populations (69.19% for AFLP and 68.63% for nSSRs, P < 0.001; Table 5). Conversely, the plastid DNA data revealed that the variance relies on the variation among populations, instead of within populations (83.83 and 16.17%, P < 0.001, respectively; Table 5). We used the pollen/seed ratio, proposed by Ennos (1994), to compare the FST values revealed by the bi- and uniparentally inherited from nuclear and plastid SSRs, respectively. The ratio of pollen flow vs. seed flow was 8.867. Genetic distance-based analysis The dissimilarity analysis performed (WNJ method) using the AFLP data resulted in a clear separation among populations of D. dissitiflora from the PC populations (Supporting Information, Appendix S4). The populations of the first species did not show a clear separation, but the populations of D. pernambucana and D. limae showed two major clusters, except for ‘Pico’ and ‘Srita’, inselbergs located at the Triunfo city (Fig. 1), indicating probable gene flow between those localities, as suggested by the pairwise FST values comparisons for the PC populations (Table 4). The ‘Cat’ population (D. limae) was more closely related to the ‘Pico’ and ‘Srita’ populations, with significant statistical support (Supporting Information, Appendix S4). Table 4. Pairwise FST values (Weir & Cockerham, 1984) based on 15 SSR loci (below the diagonal) and 340 AFLP loci (above the diagonal) 23   . Cat . Brejo . Pesq . Pico . Srita . Morrao . Cach . Lajes . ‘Cat’ - 0.245 0.152 0.127 0.125 0.3002 0.3090 0.3028 ‘Brejo’ 0.345 - 0.179 0.312 0.296 0.3581 0.3846 0.3719 ‘Pesq’ 0.305 0.491 - 0.225 0.222 0.2889 0.3274 0.3169 ‘Pico’ 0.201 0.343 0.368 - 0.020 0.2941 0.3270 0.3067 ‘Srita’ 0.237 0.367 0.381 0.034 - 0.3089 0.3333 0.3101 ‘Morrao’ 0.301 0.356 0.355 0.232 0.269 - 0.0697 0.0380 ‘Cach’ 0.391 0.434 0.422 0.353 0.380 0.117 - 0.0778 ‘Lajes’ 0.370 0.451 0.423 0.322 0.366 0.041 0.078 - 23   . Cat . Brejo . Pesq . Pico . Srita . Morrao . Cach . Lajes . ‘Cat’ - 0.245 0.152 0.127 0.125 0.3002 0.3090 0.3028 ‘Brejo’ 0.345 - 0.179 0.312 0.296 0.3581 0.3846 0.3719 ‘Pesq’ 0.305 0.491 - 0.225 0.222 0.2889 0.3274 0.3169 ‘Pico’ 0.201 0.343 0.368 - 0.020 0.2941 0.3270 0.3067 ‘Srita’ 0.237 0.367 0.381 0.034 - 0.3089 0.3333 0.3101 ‘Morrao’ 0.301 0.356 0.355 0.232 0.269 - 0.0697 0.0380 ‘Cach’ 0.391 0.434 0.422 0.353 0.380 0.117 - 0.0778 ‘Lajes’ 0.370 0.451 0.423 0.322 0.366 0.041 0.078 - Pairwise genetic differentiation within the Pernambuco complex (PC; Dyckia limae and D. pernambucana) in bold. The values highlit in red display non-significant differentiation between ‘Pico’ and ‘Srita’ populations). Open in new tab Table 4. Pairwise FST values (Weir & Cockerham, 1984) based on 15 SSR loci (below the diagonal) and 340 AFLP loci (above the diagonal) 23   . Cat . Brejo . Pesq . Pico . Srita . Morrao . Cach . Lajes . ‘Cat’ - 0.245 0.152 0.127 0.125 0.3002 0.3090 0.3028 ‘Brejo’ 0.345 - 0.179 0.312 0.296 0.3581 0.3846 0.3719 ‘Pesq’ 0.305 0.491 - 0.225 0.222 0.2889 0.3274 0.3169 ‘Pico’ 0.201 0.343 0.368 - 0.020 0.2941 0.3270 0.3067 ‘Srita’ 0.237 0.367 0.381 0.034 - 0.3089 0.3333 0.3101 ‘Morrao’ 0.301 0.356 0.355 0.232 0.269 - 0.0697 0.0380 ‘Cach’ 0.391 0.434 0.422 0.353 0.380 0.117 - 0.0778 ‘Lajes’ 0.370 0.451 0.423 0.322 0.366 0.041 0.078 - 23   . Cat . Brejo . Pesq . Pico . Srita . Morrao . Cach . Lajes . ‘Cat’ - 0.245 0.152 0.127 0.125 0.3002 0.3090 0.3028 ‘Brejo’ 0.345 - 0.179 0.312 0.296 0.3581 0.3846 0.3719 ‘Pesq’ 0.305 0.491 - 0.225 0.222 0.2889 0.3274 0.3169 ‘Pico’ 0.201 0.343 0.368 - 0.020 0.2941 0.3270 0.3067 ‘Srita’ 0.237 0.367 0.381 0.034 - 0.3089 0.3333 0.3101 ‘Morrao’ 0.301 0.356 0.355 0.232 0.269 - 0.0697 0.0380 ‘Cach’ 0.391 0.434 0.422 0.353 0.380 0.117 - 0.0778 ‘Lajes’ 0.370 0.451 0.423 0.322 0.366 0.041 0.078 - Pairwise genetic differentiation within the Pernambuco complex (PC; Dyckia limae and D. pernambucana) in bold. The values highlit in red display non-significant differentiation between ‘Pico’ and ‘Srita’ populations). Open in new tab The population tree based on the nSSRs also revealed an overall separation between the individuals of D. dissitiflora and those from the PC populations with a good support, except for three individuals from the ‘Morrao’ population (Supporting Information, Appendix S5). Those individuals displayed a somewhat different position, more related to one of the groups formed by the individuals from the ‘Pico’ and ‘Srita’ populations (Supporting Information, Appendix S5). In general, lower bootstrap support was observed in comparison with the AFLP, and, for the PC populations, four major groups could be identified in nSSR analyses (Supporting Information, Appendix S5). Unlike the AFLP data, the ‘Pico’ and ‘Srita’ populations formed two different groups, although still in closely related clusters, and we found no clear separation among individuals from different localities, as demonstrated by the AFLP analysis (Supporting Information, Appendix S5). The observed admixture among individuals of PC populations suggests historical nuclear allele sharing in the genetic structuring of sampled populations. In an attempt to obtain a better resolution regarding the genetic relationships revealed by the plastid SSR loci across the PC populations, a population tree was performed in Poptreew. As observed in nuclear data, the populations displayed a significant genetic structuring, with five major groups (Fig. 2). At the same time, we observed that populations of ‘Cat’ placed together with the ‘Pico’ and ‘Srita’ populations in a major group (Fig. 2), displaying a genetic similarity, as also observed in AFLP analysis. Therefore, a better resolution regarding the plastid SSR data could be achieved, as well, a convergence across the molecular markers employed in this work (Fig. 4). Figure 2. Open in new tabDownload slide Population evolutionary neighbour-joining tree on the genetic distance of eight plastid SSRs in Pernambuco complex (PC) populations sampled. Coloured branches indicate individuals of D. limae (red – ‘Cat’) and populations of D. pernambucana (purple – ‘Pesq’; green – ‘Srita’, dark green – ‘Brejo’ and blue – ‘Pico’). Numbers represent the collected individuals per population (PC populations N = 50). Scale bar represents 4% divergence. Figure 2. Open in new tabDownload slide Population evolutionary neighbour-joining tree on the genetic distance of eight plastid SSRs in Pernambuco complex (PC) populations sampled. Coloured branches indicate individuals of D. limae (red – ‘Cat’) and populations of D. pernambucana (purple – ‘Pesq’; green – ‘Srita’, dark green – ‘Brejo’ and blue – ‘Pico’). Numbers represent the collected individuals per population (PC populations N = 50). Scale bar represents 4% divergence. Bayesian clustering The results of hierarchical Structure analyses for both nuclear markers, considering the whole sampling suggested the indication of the most likely number of clusters after Evanno et al. (2005) of two genetic clusters (K = 2), although the K = 3 (for AFLP and nSSRs) and K = 5, for nSSRs (∆K values; Supporting Information, Appendix S6) also presented a high probability value in the hierarchical Bayesian structure analyses (Fig. 3). Thus, for the AFLP data set, the identification of the genetic clusters revealed one single group for D. dissitiflora populations and another for PC populations (Fig. 3A). Likewise, when K = 3 is considered, two clusters were recognized for the PC populations; for the AFLP data set, one cluster comprised ‘Brejo’ and ‘Pesq’ populations and one cluster included ‘Cat’, ‘Pico’ and ‘Srita’ populations, in agreement with the genetic distance-based analysis (data not shown). Figure 3. Open in new tabDownload slide Bayesian admixture Structure proportions of individuals of Dyckia populations analysed, for K = 2 model (AFLP and nSSRs; A and B, respectively) and K = 3 (for nSSRs; C). The results are based on A, 340 AFLP fragments and B,C, 15 nuclear SSR loci. The Pernambuco complex is formed by D. limae (‘Cat’) and D. pernambucana (‘Brejo’, ‘Pesq’, ‘Pico’ and ‘Srita’ populations). Figure 3. Open in new tabDownload slide Bayesian admixture Structure proportions of individuals of Dyckia populations analysed, for K = 2 model (AFLP and nSSRs; A and B, respectively) and K = 3 (for nSSRs; C). The results are based on A, 340 AFLP fragments and B,C, 15 nuclear SSR loci. The Pernambuco complex is formed by D. limae (‘Cat’) and D. pernambucana (‘Brejo’, ‘Pesq’, ‘Pico’ and ‘Srita’ populations). For the nSSR Bayesian Structure analysis, one single genetic cluster was also found for the populations of D. dissitiflora but, unlike the AFLP data, the two groups recognized for PC populations comprised only the admixture between ‘Pico’ and ‘Srita’ related populations in one cluster and ‘Cat’, ‘Pesq’ and ‘Brejo’ populations in another genetic cluster (Fig. 3B, C). Considering both markers, it is noteworthy that it was not possible to confirm the taxonomic expectations regarding ‘Cat’ population (D. limae) as a formal and isolated genetic cluster, as observed for D. dissitiflora. The admixture proportion (Q) for each individual is shown in Fig. 3. DISCUSSION Population genetic diversity Dyckia spp. are interesting models for population analysis due to their intrinsic biological features, such as the rupicolous/saxicolous nature, a wide range of morphological variability and isolated populations, exhibiting a particular colonization history and several examples of adaptive radiation (Forzza, 2001; Forzza & Silva, 2004; Hmeljevski et al., 2011; Leme et al., 2012; Rogalski et al., 2017) Thus, the population parameters provided here from different molecular markers (multilocus, nuclear and plastid) allow us to interpret some population patterns, in a detailed intraspecific scale. The genetic diversity observed in the present analysis using dominant markers (AFLP fragments; HE = 0.277) was similar with the data provided for the populations of three species of the sister genus Encholirium Mart. ex Schult. & Schult.f. (Cavallari et al., 2006). Despite the known occurrence of clonal habit in Encholirium spp. (also registered in some Dyckia spp.), the dominant marker profiles indicated the maintenance of genetic variation. Additionally, the AFLP patterns shown for the Dyckia populations studied here were similar with the genetic analysis of species of Aechmea Ruiz & Pav. (Zhang et al., 2012) and among populations of the endangered species Vriesea cacuminis L.B.Sm and Pitcairnia flammea Lindl. revealed by ISSR dominant markers (Ribeiro et al., 2013; Souza-Sobreira et al., 2015), also with significant values of genetic diversity. The overall genetic variation in the focused PC populations (D. limae and D. pernambucana) can be considered low to moderate for SSR markers, although higher (nSSR, HE = 0.343 and cpSSR, HE = 0.545) than in previously studied bromeliads (Izquierdo & Piñero, 2000; Sarthou, Samadi & Boisselier-Dubayle, 2001; Sgobarti et al., 2004) and other Dyckia spp. (D. ibiramensis with HE = 0.219; Hmeljevski et al., 2011; D. brevifolia with HE = 0.106; Rogalski et al., 2017). We observed these levels of genetic diversity even considering that these populations are narrowly distributed to the brejos de altitude, the isolated island-like mountains environment that are often related to chronic anthropogenic disturbance (CAD) effects, which may lead to reduction in the genetic diversity (Kruckeberg & Rabinowitz, 1985; Tabarelli & Santos, 2004; Arnan et al., 2018). Soares et al. (2018) found significant levels of genetic diversity in populations of Vriesea reitzii Leme & And.Costa from mixed ombrophilous forest, a particular mountain environment with high anthropogenic influence, the same type of scenario as observed in the present study. Relationships among populations may be affected by different factors, such as sample size, inbreeding, genetic drift and fitness. In general, loss or maintenance of genetic diversity may depend on different characteristics of the plant species (mating system, pollination syndromes, seed dispersal), mostly indicating a positive correlation between heterozygosity and population fitness (Loveless & Hamrick, 1984; Hamrick & Godt, 1996; Kageyama et al., 2003; Leimu et al., 2006; Frankham et al., 2010). In this regard, the above-mentioned life history attributes (reproductive strategies, such as preferential xenogamy and pollination strategies) of the isolated PC populations and the particular features of the brejos de altitude may play a pivotal role in the genetic diversity. Indeed, clonal propagation may occur in PC populations, as observed for other Dyckia spp. (e.g. Rogalski et al., 2017) and a mixed mating system was observed for the ‘Cat’ population (D. limae, D. S. Pinangé, unpubl. data). We assume that this is also the case with the other populations. Therefore, the combination of these reproductive strategies may explain the results found here, since a preferred xenogamy can lead to resilient genetic variation, just as clonal propagation can increase the longevity of genets (Hamrick & Godt, 1996; Goetze et al., 2015; Soares et al., 2018). The values found here were lower when compared with the sympatric and widely distributed populations of Encholirium spectabile (HE = 0.662; Gonçalves-Oliveira et al., 2017), Epidendrum secundum Jacq. (Orchidaceae; HE ranged from 0.442 to 0.608; Pinheiro et al., 2014) and populations of Pilosocereus gounellei F.A.C.Weber ex K.Schum (Cactaceae; HE = 0.523; Monteiro et al., 2015), species that were also evaluated using SSR markers and also widely distributed in the sympatric inselbergs of north-eastern Brazil, in the Caatinga Biome. Thus, these values, together with the observation of significant deviation from HWE, indicate that the rare Dyckia spp. (with their reduced population sizes) are more susceptible to possible genetic drift effects. Polymorphisms in plastid microsatellites have also been used to identify plant genetic diversity at various hierarchical levels (Provan, Powell & Hollingsworth, 2001; Cubas, Pardo & Tahiri, 2005; Terrab et al., 2006; Pardo, Cubas & Tahiri, 2008). The plastid SSR markers employed in the present work indicated that the PC populations appear to maintain a high level of genetic variation (Table 2), as observed in some populations of bromeliad with similar patterns of diversity (Palma-Silva et al., 2009; 2011; Gonçalves-Oliveira et al., 2017). This was also observed in the haplotype richness values in the present work, except for the ‘Srita’ population, which exhibited a single haplotype and, therefore, no evidence of genetic variation (Fig. 1; Table 3; Supporting Information, Appendix S4). This can be correlated with the significant deviations from mutation-drift equilibrium data regarding ‘Srita’ and ‘Pico’ (Supporting Information, Appendix S2) bringing evidence of population bottleneck or recent re-colonization in the here analysed populations, as expected for groups of recent evolution, as is the case of Dyckia (c. 2.5 Mya) (Krapp et al., 2014). The genetic diversity estimated at the population level in both groups (PC and D. dissitiflora populations) showed higher values for D. dissitiflora populations for almost all parameters (e.g. allelic and haplotypic richness, observed and expected heterozygosities) across the three molecular markers used. In the present case, it may also reflect the different geographical distribution ranges of the two groups (narrow endemic vs. endemic), as is the case for the PC populations and D. dissitiflora, respectively. A similar pattern of genetic diversity was also found by Barbará et al. (2007) with the populations of Alcantarea imperialis (Carrière) Harms (higher values; endemic) and A. geniculata (Wawra) J.R.Grant (lower values; narrow endemic) for which the authors argued that the results might reflect the different geographical distribution range of the two species or even differences in pollination effectiveness (discussed below). In addition, we could also argue that narrow endemics or rare species are believed to have reduced genetic variability, which is mostly attributed to the influence of evolutionary forces, such as genetic drift or selection (Hamrick & Godt, 1996; Cole, 2003). Genetic differentiation in fragmented populations of the brejos de altitude (north-eastern Brazil) Our results indicate the evidence of moderated genetic differentiation among PC populations between AFLP (FST = 0.196) and SSR (FST = 0.260 and 0.316 for PC populations) loci, also observed in the pairwise analysis (Table 4) among sampled populations. This scenario of clear genetic structuring supports the patterns of differentiation of bromeliad species living in isolated habitats, also found in other species with comparable ecological features and life histories, such as populations of Alcantarea (É.Morren ex Mez) Harms (Barbará et al., 2007, 2008a), which inhabit rocky inselbergs in the Brazilian Atlantic Forest, species of Pitcairnia L’Hér. (Palma-Silva et al., 2011) and sympatric populations of Encholirion spectabile, also occurring in the inselbergs of the Borborema Plateau and in the Caatinga-Atlantic Rainforest interzones (Gonçalves-Oliveira et al., 2017). Additionally, in Dyckia, a strong differentiation among subpopulations of the rare riparian D. ibiramensis was observed by Hmeljevski et al. (2011). These isolated population patterns possibly suggest that the low gene flow (via pollen and/or seeds) among populations allowed the influence of particular microclimatic conditions on each rocky outcrop (as already reported for the brejos de altitude: Tabarelli & Santos, 2004) and genetic drift (Wright, 1931). Given the above, it seems that genetic isolation is a trend among populations with an island-like elevational distribution such as the brejos de altitude or inselbergs. According to Loveless & Hamrick (1984), the mating system is one of the factors that can play an important role in the spatial distribution of genetic variation within and among populations. In general, Dyckia spp. show a mixed mating system (Hmeljevski et al., 2011; Leme et al., 2012), with selfing in some cases (Rogalski et al., 2009). From PC populations, the ‘Cat’ (D. limae) exhibits the existence of selfing or even mating between related individuals as a reproductive assurance strategy (D. S. Pinangé, unpubl. data), possibly indicating an increase in inbreeding rates. Given that moderate to high FIS values caused by a lack of heterozygous individuals (Table 2) were observed in other population genetic studies on bromeliads, including E. spectabile (Gonçalves-Oliveira et al., 2017), also recorded for inselberg species of the Atlantic Rainforest of south-eastern Brazil (Barbará et al., 2007), Vriesea minarum L.B.Sm. from rocky outcrops of the Espinhaço Range, Minas Gerais (Lavor et al., 2014) and Aechmea calyculata (É.Morren) Baker and A. kertesziae Reitz from the southern Atlantic Rainforest (Goetze et al., 2016, 2017a, respectively). These bromeliad population genetic studies imply a pattern credited to effects of inbreeding and limited allele sharing (therefore gene flow) among populations. Similar results of high inbreeding coefficients of populations (FIT) were observed, indicating, therefore, in the occurrence of the Wahlund effect (Hartl & Clark, 1997), which can be explained mainly by reduction of heterozygosity due to a genetic subdivision. The isolation of the PC populations, in agreement with the inbreeding coefficients, was also supported by the NJ distance analysis across the molecular markers in the present work (Fig. 2; Supporting Information, Appendices S4, S5). The AFLP-based analysis confirmed D. dissitiflora as a relevant outgroup in the clustering analysis (Supporting Information, Appendix S2). On the other hand, D. limae (‘Cat’) could not be undoubtedly discriminated from the populations of D. pernambucana in any of the cluster analyses performed (Figs 2, 3; Supporting Information, Appendices S2, S3). The distance-based analysis for AFLP and the better resolved plastid SSR population tree suggested the existence of two major groups in the brejos de altitude: the ‘Brejo’ and ‘Pesqueira’ populations on one side and the ‘Pico’ and ‘Srita’ populations with D. limae on the other side (Fig. 2; Supporting Information, Appendices S4, S5). Additionally, admixture in some individuals could be noticed among populations, indicating the existence of a remaining gene flow via pollen. Given the above, the genetic structure in the PC does not correspond to the formal taxonomic concept, and D. dissitiflora could be confirmed as a distinct lineage in the present work. Additionally, as observed for widely distributed lithophytic Bromeliaceae (Barbará et al., 2007; Boisselier-Dubayle et al., 2010; Palma-Silva et al., 2011; Gonçalves-Oliveira et al., 2017, respectively) the Mantel test revealed no recognizable linear correlation between genetic and geographical distances. The clustering analysis performed here (PC populations) pointed at this lack of correlation, with populations from more distant geographical areas showing historical genetic associations (e.g. ‘Cat’, ‘Pico’ and ‘Srita’; Fig. 2; Supporting Information, Appendices S4, S5), probably related to historical processes and geographical patterns. Regarding the genetic associations found within PC populations, one possible explanation relies on the biogeographical history of the Atlantic Forest to the north of the São Francisco River. According to the hypothesis proposed by Cavalcanti (2003), a sequence of divergence events would have defined a historical relationship between the Atlantic Forest at this region and the south-eastern Atlantic Forest. Briefly, it is proposed that the first event isolated the lowland forests from the large continuous forest, whereas the second event divided the forest corridor coming from the south and south-east, but some connections still remained with the inselbergs in Pesqueira city (the locality for the studied ‘Pesq’ population). The third significant event probably separated the inselbergs in the city of Brejo da Madre de Deus (‘Brejo’) from the others: cities of Triunfo (‘Pico’ and ‘Srita’), Buíque (‘Cat’) and Floresta. Consequently, the historical fragmentation of Atlantic Forest located to the north of the São Francisco River seems to be a possible explanation of the relationships, found in the present work, between the populations of ‘Cat’, ‘Pico’ and ‘Srita’ populations, also corroborating our first assumption (mentioned above) regarding the PC populations. The Bayesian cluster analysis also showed (and supported) these patterns and provided the first insights into the genetic structure of Dyckia populations from Pernambuco. The AFLP clustering analysis showed the existence of two clusters (K = 2) to be most likely, in accordance with the first node separation between the two major groups, and K = 3 also supported the existence of two genetic clusters within the PC populations (data not shown) ‘Pico’, ‘Srita’ and ‘Cat’ (D. limae) (these shared the same genetic cluster for K = 2; Fig. 3A, Supporting Information, Appendix S5). The SSR clustering analysis also displayed a similar result, with D. limae not being discriminated as an isolated genetic cluster (Fig. 3B, C). In this context, the plastid SSR haplotypes revealed stronger genetic structure when compared to nSSR and AFLP data, as also reported for Vriesea gigantea Gaudich. (Palma-Silva et al., 2009) and Encholirium spectabile (Gonçalves-Oliveira et al., 2017). Nevertheless, in the present work there was no evidence of shared haplotypes in the PC populations (Fig. 1, Supporting Information, Appendix S4). Altogether, the present work revealed that the overview of the evolutionary processes in such populations might be due to biparental inbreeding, selfing, clonal growth and restricted seed dispersal as the main determinants of the historical genetic structure in naturally fragmented populations. As suggested for populations of Vriesea and Alcantarea, the capacity of populations of Dyckia spp. to colonize isolated rocky outcrops probably stems from their flexible mating systems and ability to endure inbreeding. In addition, the available data for Dyckia spp. (Hmeljevski et al., 2011) also support this mentioned capacity to maintain genetic diversity and adaptive success (in this case, in a rheophytic environment) with significant inbreeding rates, in which the offspring was not composed exclusively by half-sibs, but a mixture of half-sibs, full-sibs and selfing-sibs. Evolutionary implications of pollination and seed dispersal on genetic structure in the Dyckia pernambucana complex The genetic structure data revealed by the two types of marker (nuclear and plastid) diverged significantly (FST = 0.316 for nSSRs, 0.196 for AFLPs and 0.838 for plastid SSRs; Table 5) indicating a much stronger discrepancy for plastid DNA. Comparing the partitioning of the genetic diversity (AMOVA) between nuclear and plastid markers that revealed the distinct roles of pollen and seed dispersal. The congruent multilocus and nuclear data showed that most of the variation lies within populations, in contrast to the results found for plastid SSRs, which showed higher variation rates among populations (Table 5). The available data for rock outcrop species from Caatinga and/or north-eastern Brazil also support high levels of differentiation at the plastid level (Pinheiro et al., 2014; Gonçalves-Oliveira et al., 2017). Similar patterns have been found between the sympatric and related populations of E. spectabile (closely related to Dyckia: Krapp et al., 2014; Pinangé et al., 2016) and the populations studied here, indicating a correlation of ecological and geographical factors for these bromeliads. Table 5. Analysis of molecular variance (amova) and overall FST values in the Pernambuco complex (PC) populations (Dyckia limae and D. pernambucana) based on 307 AFLP, 15 nuclear SSRs and eight plastid DNA loci Source of variation . d.f. . Variance components . Percentage of variation . P value . AFLP markers Among populations 4 11.020 Va 30.81 P < 0.001 Within populations 45 24.744 Va 69.19 P < 0.001 Overall FST: 0.196 Nuclear microsatellite Among populations 4 1.249 Va 31.64 P < 0.001 Within populations Overall FST: 0.316 95 2.699 Va 68.63 P < 0.001 Plastid DNA Among populations 4 1.314 Va 83.83 P < 0.001 Within populations Overall FST: 0.838 45 0.253 Va 16.17 P < 0.001 Source of variation . d.f. . Variance components . Percentage of variation . P value . AFLP markers Among populations 4 11.020 Va 30.81 P < 0.001 Within populations 45 24.744 Va 69.19 P < 0.001 Overall FST: 0.196 Nuclear microsatellite Among populations 4 1.249 Va 31.64 P < 0.001 Within populations Overall FST: 0.316 95 2.699 Va 68.63 P < 0.001 Plastid DNA Among populations 4 1.314 Va 83.83 P < 0.001 Within populations Overall FST: 0.838 45 0.253 Va 16.17 P < 0.001 d.f.: degrees of freedom Open in new tab Table 5. Analysis of molecular variance (amova) and overall FST values in the Pernambuco complex (PC) populations (Dyckia limae and D. pernambucana) based on 307 AFLP, 15 nuclear SSRs and eight plastid DNA loci Source of variation . d.f. . Variance components . Percentage of variation . P value . AFLP markers Among populations 4 11.020 Va 30.81 P < 0.001 Within populations 45 24.744 Va 69.19 P < 0.001 Overall FST: 0.196 Nuclear microsatellite Among populations 4 1.249 Va 31.64 P < 0.001 Within populations Overall FST: 0.316 95 2.699 Va 68.63 P < 0.001 Plastid DNA Among populations 4 1.314 Va 83.83 P < 0.001 Within populations Overall FST: 0.838 45 0.253 Va 16.17 P < 0.001 Source of variation . d.f. . Variance components . Percentage of variation . P value . AFLP markers Among populations 4 11.020 Va 30.81 P < 0.001 Within populations 45 24.744 Va 69.19 P < 0.001 Overall FST: 0.196 Nuclear microsatellite Among populations 4 1.249 Va 31.64 P < 0.001 Within populations Overall FST: 0.316 95 2.699 Va 68.63 P < 0.001 Plastid DNA Among populations 4 1.314 Va 83.83 P < 0.001 Within populations Overall FST: 0.838 45 0.253 Va 16.17 P < 0.001 d.f.: degrees of freedom Open in new tab The estimated ratio of pollen/seed found in the present work (8.87), although lower than in that found by Gonçalves-Oliveira et al. (2017) (17.12) revealed that that gene flow via seeds (along with their wind-dispersed features) is insignificant in comparison with pollen flow in both analyses. This remarkable trend can be positively associated with the hummingbird-driven pollination (Siqueira-Filho & Leme, 2006) and the wind-based seed dispersal in the inselberg populations of Dyckia. Therefore, assumptions can be made regarding a higher efficiency in gene flow through pollen than seed dispersal in those populations, as shown for other bromeliads, and confirming our third hypothesis (gene flow among species populations is hindered by island-like distribution patterns), particularly in the case of anemochorous seed dispersal (Barbará et al., 2008a; Palma-Silva et al., 2009; Paggi et al., 2010; Palma-Silva et al., 2011). On the other hand, the higher estimated pollen/seed ration found in E. spectabile indicate a higher efficiency of the mixed pollination system of this bromeliad (Queiroz et al., 2016) in the maintenance of gene flow and species cohesion. However, as mentioned above, the PC populations are pollinated by two species of hummingbirds (Chlorostilbon aureoventris and Chrysolampis mosquitos; Machado & Lopes, 2004) and display a mixed mating system (‘Cat’ population; D. S. Pinangé, unpubl. data) There is also an association between self-compatibility and hummingbird-pollinated plants (Wolowsky et al., 2013; Wanderley, Lopes & Machado, 2014). According to Wolowsky et al. (2013), the mixed mating system can be easily adapted with the high efficiency of hummingbirds in promoting allogamy in self-compatible plants, reducing the deleterious effects of an only self-compatible mating system, as seen in the patterns of genetic diversity provided here, and also the data provided by nuclear markers. CONCLUSIONS AND PERSPECTIVES FOR CONSERVATION STRATEGIES The combined results of genetic variation from different markers allowed us to identify diversity hotspots in bromeliads from two important and exclusive types of vegetation: the Diamantina Plateau in the northern Espinhaço Range and the brejos de altitude from the Borborema Plateau and related inselbergs. The data suggested that, despite the possible influence of genetic drift and selfing rates, the populations showed moderate patterns of diversity in the continental island-like rock outcrop, as high as those seen in other bromeliad studies. Considering the results, the application of AFLPs and two types of SSR marker (nuclear and plastid DNA), with particular evolutionary histories in the genome, has proven to be significantly powerful in the evaluated populations of Dyckia. Taking into account the morphological characteristics and the molecular data achieved here, D. limae could not be confirmed as a separate taxonomic unit at the species level, suggesting that synonymization with D. pernambucana might be appropriate. The PARNA Catimbau, the National Park located in the city of Buíque, is the only sampling location in the Borborema Plateau with a sedimentary formation (instead of the remaining granitic formations), possibly explaining the existing adaptive distinctions of this species. In this regard, even though isolation and moderate genetic differentiation data may promote allopatric speciation in the group, the data presented here suggest that this lineage separation has not yet occurred. The biogeographic assumptions raised by Krapp et al. (2014) regarding colonization patterns, dating and clade diversification (c. 2.5 Mya) are in agreement with the results presented here. It is reasonable to assume that since there is a recent diversification, the taxa may have had not enough time to accumulate specific polymorphism or alleles, and diversification is probably still ongoing. Therefore, PC populations (D. limae and D. pernambucana) exhibited moderate genetic diversity (Tables 1, 2) and private haplotypes in all sampled areas (Table 2). Altogether, the reduced and endemic population sizes, being more prone to possible genetic drift effects, evidence of bottlenecks and moderate levels of genetic variation make this lineage ideal for conservation strategies. In addition, the ongoing synonymization would enable the conservation of the gene pool of the lineage analysed here, since the conservation of D. limae as a formerly separate species would not reflect the conservation of the minimal gene pool of D. pernambucana. Hence, we propose the revalidation of the conservation status of this species together with the ongoing synonymization including all five sampled areas of brejos de altitude, not only in the PARNA Catimbau as initially proposed by Fabricante et al. (2014), to ensure the survival and preservation of the evolutionary capacity of the group. Likewise, the isolated island-like mountain environments (brejos de altitude) are known for their high levels of endemism, and they are often related to CAD effects, which may lead to reduction in the genetic diversity (Arnan et al., 2018). Finally, studies concerning the fine-scale structure, phenology, pollination ecology, paternity correlation and population dynamics should be addressed in the populations from ‘Pico’, ‘Srita’ (located in Triunfo city) and ‘Cat’ (Catimbau National Park) to deepen the understanding of their relationship. SUPPORTING INFORMATION Additional Supporting Information may be found in the online version of this article at the publisher’s website. Appendix S1. Population names, localities, geographical coordinates and sample sizes of Dyckia spp. from the inselbergs of the Diamantina Plateau (Bahia) and brejos de altitude (Pernambuco), north-eastern Brazil. Appendix S2. Mean observed/expected heterozygosity ratio and significance from Wilcoxon’s test for putative deviations in the mutation-drift equilibrium (genetic bottlenecks) in the populations of Pernambuco complex (PC) (‘Cat’, ‘Brejo’, ‘Pesq’, ‘Pico’ and ‘Srita’) and D. dissitiflora (‘Morrao’, ‘Cach’ and ‘Lajes’) under three different models for nuclear microsatellite data using Bottleneck software. Appendix S3. Distribution and frequencies of the haplotypes (plastid DNA) in all Dyckia populations sampled. Haplotypes from D. pernambucana and D. limae are in boldface. Appendix S4. Dendrogram based on the genetic distance of 340 AFLP bands using the WNJ method in all populations sampled. Black and coloured branches indicate individuals of D. dissitiflora (black), D. limae (red) and populations of D. pernambucana (purple – ‘Pesq’; green – ‘Srita’ and blue – ‘Pico’). Numbers above branches are bootstrap support values. Scale bar represents 3% divergence. Appendix S5. Dendrogram based on the genetic distance of 15 nSSR loci using the neighbor-joining method in all populations sampled. Black and coloured branches indicate individuals of D. dissitiflora (black), D. limae (‘Cat’– red) and populations of D. pernambucana (purple – ‘Pesq’; green – ‘Srita’, dark green – ‘Brejo’ and blue – ‘Pico’). Numbers above branches are bootstrap support values. Appendix S6. Indication of the most likely number of clusters after Evanno et al. (2005) in the Structure analysis. Results from ten replications for each 1≤ ∆K ≤10 values with A, AFLP and B, nuclear SSR data. C, Indication of the most likely number of clusters after Evanno et al. (2005), considering only the PC populations and based on nuclear SSR loci. Results from ten replicates for each 1 ≤ ∆K ≤ 7 values. ACKNOWLEDGEMENTS We thank FACEPE (Fundação de Amparo à Pesquisa do Estado de Pernambuco, Brazil), DAAD (Deutscher Akademischer Austauschdienst – PROBRAL Program), CAPES (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior) and CNPq (Conselho Nacional de Desenvolvimento Científico e Tecnológico) for financial support. We also thank Dr Artur Wanderley and Dr Geyner Cruz for the help with the fieldwork, Marcos Júnior and Dr Rodrigo César for the support in the DNA isolation and Dr Santelmo Vasconcelos and MSc Mariana Dias for all the support in the revision of this work. REFERENCES Arnan X , Leal IR, Tabarelli M , Andrade JF , Barros MF , Câmara T , Jamelli D , Knoechelmann CM , Menezes TGC , Menezes AGS , Oliveira FMP , de Paula AS , Pereira SC , Rito KF , Sfair JC , Siqueira FFS , Souza DG , Specht MJ , Vieira LA , Arcoverde GB , Andersen AN . 2018 . A framework for deriving measures of chronic anthropogenic disturbance: surrogate, direct, single and multi-metric indices in Brazilian Caatinga . Ecological Indicators 94 : 274 – 282 . Google Scholar Crossref Search ADS WorldCat Bandelt J , Forster P , Röhl A . 1999 . Median-joining networks for inferring intraspecific phylogenies . Molecular Biology and Evolution 16 : 37 – 48 . Google Scholar Crossref Search ADS PubMed WorldCat Barbará T , Lexer C , Martinelli G , Mayo S , Fay MF , Heuertz M . 2008a . Within-population spatial genetic structure in four naturally fragmented species of a Neotropical inselberg radiation, Alcantarea imperialis, A. geniculata, A. glaziouana and A. regina (Bromeliaceae) . Heredity 101 : 285 – 296 . Google Scholar Crossref Search ADS WorldCat Barbará T , Martinelli G , Fay MF , Mayo SJ , Lexer C . 2007 . Population differentiation and species cohesion in two closely related plants adapted to Neotropical high-altitude ‘inselbergs’, Alcantarea imperialis and Alcantarea geniculata (Bromeliaceae) . Molecular Ecology 16 : 1981 – 1992 . Google Scholar Crossref Search ADS PubMed WorldCat Barbará T , Martinelli G , Palma-Silva C , Fay MF , Mayo S , Lexer C . 2008b . Genetic relationships and variation in reproductive strategies in four closely related bromeliads adapted to Neotropical 'inselbergs': Alcantarea glaziouana, A. regina, A. geniculata and A. imperialis (Bromeliaceae) . Annals of Botany 103 : 65 – 77 . Google Scholar Crossref Search ADS WorldCat Barfuss MHJ , Till W , Leme EMC , Pinzón JP , Manzanares JM , Halbritter H , Samuel R , Brown GK . 2016 . Taxonomic revision of Bromeliaceae subfam. Tillandsioideae based on a multi-locus DNA sequence phylogeny and morphology . Phytotaxa 279 : 1 – 97 . Google Scholar Crossref Search ADS WorldCat BFG. 2018 . Brazilian Flora 2020: innovation and collaboration to meet Target 1 of the Global Strategy for Plant Conservation (GSPC) . Rodriguésia 69 : 1513 – 1527 . Crossref Search ADS WorldCat Boisselier-Dubayle MC , Leblois R , Samadi S , Lambourdière J , Sarthou C . 2010 . Genetic structure of the xerophilous bromeliad Pitcairnia geyskesii on inselbergs in French Guiana - a test of the forest refuge hypothesis . Ecography 33 : 175 – 184 . Google Scholar Crossref Search ADS WorldCat Butcher D , Gouda EJ . cont. updated . The new bromeliad taxon list. Utrecht : University Botanic Gardens . Available at: http://bromeliad.nl/taxonList/. accessed 18 July 2019 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Cavalcanti DR . 2003 . Distribuição altitudinal de plantas lenhosas e relações históricas entre a Floresta Atlântica do Sul-Sudeste e o Centro de Endemismo Pernambuco. Unpublished M.Sc. Thesis, Federal University of Pernambuco . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Cavallari MM , Forzza RC , Veasey EA , Zucchi MI , Oliveira GCX . 2006 . Genetic variation in three endangered species of Encholirium (Bromeliaceae) from Cadeia do Espinhaço, Brazil, setected using RAPD Markers . Biodiversity and Conservation 15 : 4357 – 4373 . Google Scholar Crossref Search ADS WorldCat Chapman HM , Parh D , Oraguzie N . 2000 . Genetic structure and colonizing success of a clonal, weedy species, Pilosella officinarum (Asteraceae) . Heredity 84 : 401 – 409 . Google Scholar Crossref Search ADS PubMed WorldCat Cole CT . 2003 . Genetic variation in rare and common plants . Annual Review of Ecology and Systematics 34 : 213 – 237 . Google Scholar Crossref Search ADS WorldCat Cornuet JM , Luikart G . 1997 . Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data . Genetics 144 : 2001 – 2014 . OpenURL Placeholder Text WorldCat Crandall KA , Templeton AR . 1993 . Empirical tests of some predictions from coalescent theory with applications to intraspecific phylogeny reconstruction . Genetics 134 : 959 – 969 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Cubas P , Pardo C , Tahiri H . 2005 . Genetic variation and relationships among Ulex (Fabaceae) species in southern Spain and northern Morocco assessed by chloroplast microsatellite (cpSSR) markers . American Journal of Botany 92 : 2031 – 2043 . Google Scholar Crossref Search ADS PubMed WorldCat Debener T , Mattiesch L . 1999 . Construction of a genetic linkage map for roses using RAPD and AFLP markers . Theorical Applied Genetics 99 : 891 – 899 . Google Scholar Crossref Search ADS WorldCat Dieringer D , Schlötterer C . 2003 . Microsatellite analyser (MSA): a platform independent analysis tool for large microsatellite data sets . Molecular Ecology Notes 3 : 167 – 169 . Google Scholar Crossref Search ADS WorldCat Doyle JJ , Doyle JL . 1987 . A rapid DNA isolation procedure for small amounts of fresh leaf tissue . Phytochemical Bulletin, Botanical Society of America 19 : 11 – 15 . OpenURL Placeholder Text WorldCat Earl DA , von Holdt BM . 2012 . Structure Harvester: a website and program for visualizing Structure output and implementing the Evanno method . Conservation Genetics Resources 4 : 359 – 361 . Google Scholar Crossref Search ADS WorldCat El Mousadik A , Petit RJ . 1996 . High level of genetic differentiation for allelic richness among populations of the argan tree Argania spinosa (L) Skeels endemic to Morocco . Theoretical and Applied Genetics 92 : 832 – 839 . Google Scholar Crossref Search ADS PubMed WorldCat Eliades N-G , Eliades DG . 2009 . Haplotype Analysis: software for analysis of haplotype data. Göttingen : Forest Genetics and Forest Tree Breeding, Georg-August University . Available at: http://www.uni-goettingen.de/en/134935.html. Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Ennos R . 1994 . Estimating the relative rates of pollen and seed migration among plant populations . Heredity 72 : 250 – 259 . Google Scholar Crossref Search ADS WorldCat Evanno G , Regnaut S , Goudet J . 2005 . Detecting the number of clusters of individuals using the software structure: a simulation study . Molecular Ecology 14 : 2611 – 2620 . Google Scholar Crossref Search ADS PubMed WorldCat Excoffier L , Laval G , Schneider S . 2005 . Arlequin ver. 3.11: an integrated software package for population genetics data analysis . Evolutionary Bioinformatics Online 1 : 47 – 50 . OpenURL Placeholder Text WorldCat Fabricante JR , Araújo KCT , Ferreira JVA , Castro RA , Silva ACCP , Siqueira-Filho JA . 2014 . Categorização do risco de extinção de Dyckia limae L. B. Sm. e Tillandsia catimbauensis Leme, W. Till & J.A. Siqueira por meio de critérios de distribuição geográfica . Biotemas 27 : 203 – 207 . Google Scholar Crossref Search ADS WorldCat Falush D , Stephens M , Pritchard JK . 2003 . Inference of population structure: extensions to linked loci and correlated allele frequencies . Genetics 164 : 1567 – 1587 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Felsenstein J . 1985 . Confidence limits on phylogenies: an approach using the bootstrap . Evolution 39 : 783 – 791 . Google Scholar Crossref Search ADS PubMed WorldCat Forzza RC . 2001 . Filogenia da tribo Puyeae Wittm. e revisão taxonômica do gênero Encholirium Mart. ex Schult. & Schult.f. (Pitcairnioideae – Bromeliaceae). Unpublished D. Phil. Thesis, Universidade de São Paulo . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Forzza RC , Silva BR . 2004 . A new species of Dyckia (Bromeliaceae) from Rio de Janeiro State, Brazil . Novon 14 : 168 – 170 . OpenURL Placeholder Text WorldCat Forzza RC , Wanderley MGL . 1998 . Flora da Serra do Cipó, Minas Gerais: Bromeliaceae – Pitcairnioideae . Boletim de Botânica da Universidade de São Paulo 17 : 225 – 270 . OpenURL Placeholder Text WorldCat Frankham R , Ballou JD , Briscoe DA . 2010 . Introduction to conservation genetics , 2nd edn. Cambridge : Cambridge University Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Givnish TJM , Barfuss HJB , Ee V , Riina R , Schulte K , Horres R , Gonsiska PA , Jabaily RS , Crayn DM , Smith JAC , Winter K , Brown GK , Evans TM , Holst BK , Luther H , Till W , Zizka G , Berry PE , Sytsma KJ . 2014 . Adaptive radiation, correlated and contingent evolution, and net species diversification in Bromeliaceae . Molecular Phylogenetics and Evolution 71 : 55 – 78 . Google Scholar Crossref Search ADS PubMed WorldCat Goetze M , Büttow MV , Zanella CM , Paggi GM , Bruxel M , Pinheiro FP , Sampaio JAT , Palma-Silva C , Cidade , FW , Bered F . 2015 . Genetic variation in Aechmea winkleri, a bromeliad from an inland Atlantic rainforest fragment in southern Brazil . Biochemical Systematics and Ecology 58 : 204 – 210 . Google Scholar Crossref Search ADS WorldCat Goetze M , Capra F , Büttow MV , Zanella CM , Bered F . 2017a . High genetic diversity and demographic stability in Aechmea kertesziae (Bromeliaceae), a species of sandy coastal plains (restinga habitat) in southern Brazil . Botanical Journal of the Linnean Society 186 : 374 – 388 . Google Scholar Crossref Search ADS WorldCat Goetze M , Palma-Silva C , Zanella CM , Bered F . 2016 . East-to-west genetic structure in populations of Aechmea calyculata (Bromeliaceae) from the southern Atlantic rainforest of Brazil . Botanical Journal of the Linnean Society 181 : 477 – 490 . Google Scholar Crossref Search ADS WorldCat Goetze M , Zanella CM , Palma-Silva C , Büttow MV , Bered F . 2017b . Incomplete lineage sorting and hybridization in the evolutionary history of closely related, endemic yellow-flowered Aechmea species of subgenus Ortgiesia (Bromeliaceae) . American Journal of Botany 104 : 1 – 15 . Google Scholar Crossref Search ADS WorldCat Goldstein DB , Ruiz-Linares A , Cavalli-Sforza LL , Feldman MW . 1995 . An evaluation of genetic distances for use with microsatellite loci . Genetics 139 : 463 – 471 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Gonçalves-Oliveira RC , Wöhrmann T , Benko-Iseppon AM , Krapp F , Alves M , Wanderley M das GL , Weising K . 2017 . Population genetic structure of the rock outcrop species Encholirium spectabile (Bromeliaceae): the role of pollination vs. seed dispersal and evolutionary implications . American Journal of Botany 104 : 868 – 878 . Google Scholar Crossref Search ADS PubMed WorldCat Guicking D , Kröger-Kilian T , Weising K , Blattner FR . 2008 . Single nucleotide sequence analysis: a cost and time effective protocol for the analysis of microsatellite and indel rich chloroplast DNA regions . Molecular Ecology Resources 8 : 62 – 65 . Google Scholar Crossref Search ADS PubMed WorldCat Hamrick JL . 1994 . Genetic diversity and conservation in tropical forests. In: Drysdale RM , John SET , Yopa AC , eds. Proceedings, International Symposium on Genetic Conservation and Production of Tropical Forest Tree Seed. Muak-Lek : ASEAN-Canada Forest Tree Centre Project . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hamrick JL , Godt MJW . 1996 . Conservation genetics of endemic plant species . In: Avise JC , Hamrick JL , eds. Conservation genetics. New York : Chapman and Hall , 281 – 304 . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Hartl DL , Clark AG . 1997 . Principles of population genetics. Sunderland : Sinauer & Associates . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Hmeljevski KV , Reis A , Montagna T , Reis MS . 2011 . Genetic diversity, genetic drift and mixed mating system in small subpopulations of Dyckia ibiramensis, a rare endemic bromeliad from southern Brazil . Conservation Genetics 12 : 761 – 769 . Google Scholar Crossref Search ADS WorldCat Hmeljevski KV , Reis MS , Forzza , RC . 2015 . Patterns of gene flow in Encholirium horridum L.B.Sm., a monocarpic species of Bromeliaceae . Journal of Heredity 106 : 93 – 101 . Google Scholar Crossref Search ADS PubMed WorldCat Izquierdo LY , Piñero D . 2000 . High genetic diversity in the only known population of Aechmea tuitensis (Bromeliaceae) . Australian Journal of Botany 48 : 645 – 650 . Google Scholar Crossref Search ADS WorldCat Kageyama PY , Sebbenn AM , Ribas LA , Gandara FB , Castellen M , Perecim MB , Vencovsky R . 2003 . Diversidade genética em espécies arbóreas tropicais de diferentes estágios sucessionais por marcadores genéticos . Scientia Forestalis 64 : 93 – 107 . OpenURL Placeholder Text WorldCat Krapp F , Pinangé DSB , Benko-Iseppon AM , Leme EMC , Weising K . 2014 . Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences . Plant Systematics and Evolution 300 : 1591 – 1614 . OpenURL Placeholder Text WorldCat Krapp F , Wohrmann T , Pinangé DSB , Benko-Iseppon AM , Huettel B , Weising K . 2012 . A set of plastid microsatellite loci for the genus Dyckia (Bromeliaceae) derived from 454 pyrosequencing . American Journal of Botany 99 : e470 – e473 . Google Scholar Crossref Search ADS PubMed WorldCat Kruckeberg AR , Rabinowitz D . 1985 . Biological aspects of endemism in higher plants . Annual Review of Ecology, Evolution, and Systematics 16 : 447 – 79 . Google Scholar Crossref Search ADS WorldCat Lavor P , van den Berg C , Jacobi CM , Carmo FF , Versieux LM . 2014 . Population genetics of the endemic and endangered Vriesea minarum (Bromeliaceae) in the Iron Quadrangle, Espinhaço Range, Brazil . American Journal of Botany 101 : 1167 – 1175 . Google Scholar Crossref Search ADS PubMed WorldCat Leimu R , Mutikainen P , Koricheva J , Fischer M . 2006 . How general are positive relationships between plant population size, fitness and genetic variation? Journal of Ecology 94 : 942 – 952 . Google Scholar Crossref Search ADS WorldCat Leme EMC , Ribeiro OBC , Miranda ZJG . 2012 . New species of Dyckia (Bromeliaceae) from Brazil . Phytotaxa 67 : 9 – 37 . Google Scholar Crossref Search ADS WorldCat Lins RC . 1989 . As áreas de exceção do agreste de Pernambuco. Recife : Sudene . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Loveless MD , Hamrick JL . 1984 . Ecological determinants of genetic structure in plant populations . Annual Review of Ecology and Systematic Evolution 15 : 69 – 95 . OpenURL Placeholder Text WorldCat Lynch M , Milligan BG . 1994 . Analysis of population genetic structure with RAPD markers . Molecular Ecology 3 : 91 – 99 . Google Scholar Crossref Search ADS PubMed WorldCat Machado IC , Lopes AV . 2004 . Floral traits and pollination systems in the Caatinga, a Brazilian tropical dry forest . Annals of Botany 94 : 365 – 376 . Google Scholar Crossref Search ADS PubMed WorldCat Monteiro ER , Mangolin CA , Neves AF , Orasmo GR , Silva JGM , Machado MFPS . 2015 . Genetic diversity and structure of populations in Pilosocereus gounellei (F.A.C.Weber ex K.Schum.) (Cactaceae) in the Caatinga biome as revealed by heterologous microsatellite primers . Biochemical Systematics and Ecology 58 : 7 – 12 . Google Scholar Crossref Search ADS WorldCat Mota MR , Pinheiro F , Leal BSS , Wendt T , Palma-Silva C . 2018 . The role of hybridization and introgression in maintaining species integrity and cohesion in naturally isolated inselberg bromeliad populations . Plant Biology 21 : 122 – 132 . Google Scholar Crossref Search ADS PubMed WorldCat Navascués M , Emerson BC . 2005 . Chloroplast microsatellites: measures of genetic diversity and the effect of homoplasy . Molecular Ecology 14 : 1333 – 1341 Google Scholar Crossref Search ADS PubMed WorldCat Nei M . 1973 . Analysis of gene diversity in subdivided populations . Proceedings of the National Academy of Sciences of the United States of America 70 : 3321 – 3323 . Google Scholar Crossref Search ADS PubMed WorldCat Paggi GM , Sampaio JAT , Bruxel M , Zanella CM , Goetze M , Büttow MV , Palma-Silva C , Bered F . 2010 . Seed dispersal and population structure in Vriesea gigantea, a bromeliad from the Brazilian Atlantic Rainforest . Botanical Journal of the Linnean Society 164 : 317 – 325 . Google Scholar Crossref Search ADS WorldCat Palma-Silva C , Leal BSS , Chaves CJN , Fay MF . 2016 . Advances in and perspectives on evolution in Bromeliaceae . Botanical Journal of the Linnean Society 181 : 305 – 322 . Google Scholar Crossref Search ADS WorldCat Palma-Silva C , Lexer C , Paggi GM , Barbará T , Bered F , Bodanese-Zanettini MH . 2009 . Range-wide patterns of nuclear and chloroplast DNA diversity in Vriesea gigantea (Bromeliaceae), a Neotropical forest species . Heredity 103 : 503 – 512 . Google Scholar Crossref Search ADS PubMed WorldCat Palma-Silva C , Wendt T , Pinheiro F , Barbará T , Fay MF , Cozzolino S , Lexer C . 2011 . Sympatric bromeliad species (Pitcairnia spp.) facilitate tests of mechanisms involved in species cohesion and reproductive isolation in Neotropical inselbergs . Molecular Ecology 20 : 3185 – 3201 . Google Scholar Crossref Search ADS PubMed WorldCat Pardo C , Cubas P , Tahiri H . 2008 . Genetic variation and phylogeography of Stauracanthus (Fabaceae, Genistae) from the Iberian Peninsula and northern Morocco assessed by chloroplast microsatellite (cpSSR) markers . American Journal of Botany 95 : 98 – 109 . Google Scholar Crossref Search ADS PubMed WorldCat Perrier X , Jacquemoud-Collet JP . 2006 . Dissimilarity analysis and representation for Windows –DARwin software. Program distributed by the author. Available at: http://darwin.cirad.fr/ Pinangé DSB , Krapp F , Zizka G , Silvestro D , Leme EMC , Weising K , Benko-Iseppon AM . 2016 . Molecular phylogenetics, historical biogeography and character evolution in Dyckia (Bromeliaceae, Pitcairnioideae) . Botanical Journal of the Linnean Society 183 : 39 – 56 . OpenURL Placeholder Text WorldCat Pinheiro F , Cozzolino S , Draper D , de Barros F , Félix LP , Fay MF , Palma-Silva C . 2014 . Rock outcrop orchids reveal the genetic connectivity and diversity of inselbergs of north-eastern Brazil . BMC Evolutionary Biology 14 : 49 . Google Scholar Crossref Search ADS PubMed WorldCat Piry S , Luikart G , Cornuet JM . 1999 . BOTTLENECK: a computer program for detecting recent reductions in the effective population size using allele frequency data . Journal of Heredity 90 : 502 – 503 . Google Scholar Crossref Search ADS WorldCat Porembski S . 2007 . Tropical inselbergs: habitat types, adaptive strategies and diversity patterns . Brazilian Journal of Botany 30 : 579 – 586 . Google Scholar Crossref Search ADS WorldCat Prance GT . 1982 . Forest refuges: evidence from woody angiosperms. In: Prance GT , ed. Biological diversification in the Tropics. New York : Columbia University Press , 137 – 157 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Pritchard JK , Stephens M , Donnelly P . 2000 . Inference of population structure using multilocus genotype data . Genetics 155 : 945 – 959 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Provan J , Powell W , Hollingsworth PM . 2001 . Chloroplast microsatellites: new tools for studies in plant ecology and evolution . Trends in Ecology and Evolution 16 : 142 – 147 . Google Scholar Crossref Search ADS PubMed WorldCat Queiroz JA , Quirino ZGM , Lopes AV , Machado IC . 2016 . Vertebrate mixed pollination system in Encholirium spectabile: a bromeliad pollinated by bats, opossum and hummingbirds in a tropical dry forest . Journal of Arid Environments 125 : 21 – 30 . Google Scholar Crossref Search ADS WorldCat Raymond M , Rousset F . 1995 . Genepop (version 1.2): population genetics software for exact tests and ecumenicism . Journal of Heredity 86 : 248 – 249 Google Scholar Crossref Search ADS WorldCat Ribeiro PC , Pinheiro LC , Domingues R , Forzza RC , Machado MA , Viccini LF . 2013 . Genetic diversity of Vriesea cacuminis (Bromeliaceae): an endangered and endemic Brazilian species . Genetics and Molecular Research 12 : 1934 – 1943 . Google Scholar Crossref Search ADS PubMed WorldCat Rogalski JM , Reis A . 2009 . Conservação de reófitas: o caso da bromélia Dyckia brevifolia Baker, Rio Itajaí-Açu, SC . In: Tres DR , Reis A , eds. Perspectivas sistêmicas para a conservação e restauração ambiental: do pontual ao contexto. Florianópolis : Herbário Barbosa Rodrigues , 335 – 344 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Rogalski JM , Reis A , Reis MS , Hmeljevski KV . 2009 . Biologia reprodutiva da reófita Dyckia brevifolia Baker (Bromeliaceae), no Rio Itajaí-Açu, Santa Catarina, Brasil . Revista Brasileira de Botânica 32 : 691 – 702 . OpenURL Placeholder Text WorldCat Rogalski JM , Reis A , Rogalski M , Montagna T , Dos Reis MS . 2017 . Mating system and genetic structure across all known populations of Dyckia brevifolia: a clonal, endemic, and endangered rheophyte bromeliad . Journal of Heredity 108 : 299 – 307 . Google Scholar Crossref Search ADS WorldCat Rousset F . 2008 . Genepop: a complete reimplementation of the Genepop software for Windows and Linux . Molecular Ecology Resources 8 : 103 – 106 . Google Scholar Crossref Search ADS PubMed WorldCat Saitou N , Nei M . 1987 . The neighbor-joining method: a new method for reconstructing phylogenetic trees . Molecular Biology and Evolution 4 : 406 – 425 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Sarthou C , Samadi S , Boisselier-Dubayle MC . 2001 . Genetic structure of the saxicole Pitcairnia geyskesii (Bromeliaceae) on inselbergs in French Guiana . American Journal of Botany 88 : 861 – 868 . Google Scholar Crossref Search ADS PubMed WorldCat Schuelke M . 2000 . An economic method for the fluorescent labelling of PCR fragments . Nature Biotechnology 18 : 233 – 234 . Google Scholar Crossref Search ADS PubMed WorldCat Shriver MD , Jin L , Boerwinkle E , Deka R , Ferrell RE , Chakraborty R . 1995 . A novel measure of genetic distance for highly polymorphic tandem repeat loci . Molecular Biology and Evolution 12 : 914 – 920 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Silvestro D , Zizka G , Schulte K . 2014 . Disentangling the effects of key innovations on the diversification of Bromelioideae (Bromeliaceae) . Evolution 68 : 163 – 175 . Google Scholar Crossref Search ADS PubMed WorldCat Siqueira-Filho JA , Leme EMC . 2006 . Fragmentos de Mata Atlântica do Nordeste. Biodiversidade, conservação e suas Bromélias. Rio de Janeiro : Andréa Jakobson Estúdio . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Smith LB , Downs RJ . 1974 . Pitcairnioideae (Bromeliaceae) . Flora Neotropica Monograph 14 : 1 – 660 . OpenURL Placeholder Text WorldCat Soares EL , Goetze M , Zanella CM , Bered F . 2018 . Genetic diversity and population structure of Vriesea reitzii (Bromeliaceae), a species from the southern Brazilian Highlands . Genetics and Molecular Biology 41 : 308 – 317 . Google Scholar Crossref Search ADS PubMed WorldCat Sokal RR , Rohlf FI . 1995 . Biometry: the principles and practice of statistics in biological research . New York : Freeman and Company . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Souza-Sobreira FB , Souza GB , Rosado CCG , Miranda FD , Soares TCB , Gontijo ABPL . 2015 . Genetic diversity in three natural populations of Pitcairnia flammea (L.) John (Bromeliaceae) estimated by ISSR markers . Genetics and Molecular Research 14 : 15892 – 15901 . Google Scholar Crossref Search ADS PubMed WorldCat Sgorbati S , Labra M , Grugni E , Barcaccia G , Galasso G , Boni U , Mucciarelli M , Citterio S , Benavides Iramátegui A , Venero Gonzales L , Scannerini S . 2004 . A survey of genetic diversity and reproductive biology of Puya raimondii (Bromeliaceae), the endangered queen of the Andes . Plant Biology 6 : 222 – 230 . Google Scholar Crossref Search ADS PubMed WorldCat Siqueira-Filho JA . 2004 . As bromélias nos brejos de altitude em Pernambuco: riqueza de espécies e status de conservação . In: Pôrto KC , Cabral JJP , Tabarelli M , eds. Brejos de altitude em Pernambuco e Paraíba: história natural, ecologia e conservação . Brasília, DF : MMA , 99 – 110 . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Tabarelli M , Santos AMM . 2004 . Uma breve descrição sobre a história natural dos brejos nordestinos. 99–110. In: Pôrto KC , Cabral JJP , Tabarelli , M , eds. Brejos de altitude em Pernambuco e Paraíba: história natural, ecologia e conservação. Brasília : Ministério do Meio Ambiente . Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Takezaki N , Nei M , Tamura K . 2014 . Poptreew: Web version of Poptree for constructing population trees from allele frequency data and computing other population statistics . Molecular Biology and Evolution 31 : 1622 – 1624 . Google Scholar Crossref Search ADS PubMed WorldCat Tel-Zur N , Abbo S , Myslabodski D , Mizrahi Y . 1999 . Modified CTAB procedure for DNA isolation from epiphytic cacti of the genera Hylocereus and Selenicereus (Cactaceae) . Plant Molecular Biology Reporter 17 : 249 – 254 . Google Scholar Crossref Search ADS WorldCat Terrab A , Paun O , Talavera S , Tremetsberger K , Arista M , Stuessy TF . 2006 . Genetic diversity and population structure in natural populations of Moroccan Atlas cedar (Cedrus atlantica) determined with cpSSR markers . American Journal of Botany 93 : 1274 – 1280 . Google Scholar Crossref Search ADS PubMed WorldCat Vekemans X , Beauwens T , Lemaire M , Roldán-Ruiz I . 2002 . Data from amplified fragment length polymorphism (AFLP) markers show indication of size homoplasy and of a relationship between degree of homoplasy and fragment size . Molecular Ecology 11 : 139 – 151 . Google Scholar Crossref Search ADS PubMed WorldCat Vos P , Hogers R , Bleeker M , Reijans M , Van de Lee T , Hornes M , Frijters A , Pot J , Peleman J , Kuiper M , Zabeau M . 1995 . AFLP: a new technique for DNA fingerprinting . Nucleic Acids Research 23 : 4407 – 4414 . Google Scholar Crossref Search ADS PubMed WorldCat Wanderley AM , Lopes AV , Machado ICS . 2014 . Reproductive ecology of Ameroglossum pernambucense (Scrophulariaceae): is this ornithophilous and threatened shrub highly adapted to a naturally fragmented habitat? Plant Systematics and Evolution 300 : 1099 – 1110 . Google Scholar Crossref Search ADS WorldCat Weir BS , Cockerham CC . 1984 . Estimating F-statistics for the analysis of population structure . Evolution 38 : 1358 – 1370 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Weising K , Nybom H , Wolff K , Kahl G . 2005 . DNA fingerprinting in plants, principles, methods, and applications , 2nd edn. Boca Raton : CRC Press . Google Scholar Crossref Search ADS Google Scholar Google Preview WorldCat COPAC Wöhrmann T , Pinangé DSB , Krapp F , Benko-Iseppon AM , Huettel B , Weising K . 2012b . Development of 15 nuclear microsatellite markers in the genus Dyckia (Pitcairnioideae; Bromeliaceae) using 454 pyrosequencing . Conservation Genetics Resources 5 : 81 – 84 . Google Scholar Crossref Search ADS WorldCat Wöhrmann T , Wagner N , Krapp F , Huettel B , Weising K . 2012a . Development of microsatellite markers in Fosterella rusbyi (Bromeliaceae) using 454 pyrosequencing . American Journal of Botany 99 : e160 – e163 . Google Scholar Crossref Search ADS WorldCat Wolowsky M , Saad SF , Ashman TL , Freitas L . 2013 . Predominance of self-compatibility in hummingbird-pollinated plants in the Neotropics . Naturwissenschaften 100 : 69 – 79 . Google Scholar Crossref Search ADS PubMed WorldCat Wright S . 1931 . Evolution in Mendelian populations . Genetics 16 : 97 – 159 . Google Scholar PubMed OpenURL Placeholder Text WorldCat Wright S . 1965 . The interpretation of population structure by F-statistics with special regards to system of mating . Evolution 19 : 395 – 420 . Google Scholar Crossref Search ADS WorldCat Zhang F , Wang W , Ge Y , Shen X , Tian D , Liu J , Liu X , Yu X , Zhang Z . 2012 . Genetic relatedness among Aechmea species and hybrids inferred from AFLP markers and pedigree data . Scientia Horticulturae 139 : 39 – 45 . Google Scholar Crossref Search ADS WorldCat Zhivotovsky LA . 1999 . Estimating population structure in diploids with multilocus dominant DNA markers . Molecular Ecology 8 : 907 – 913 . Google Scholar Crossref Search ADS PubMed WorldCat © 2019 The Linnean Society of London, Botanical Journal of the Linnean Society This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)
TI - Population genetics shed light on species delimitation and life history of the Dyckia pernambucana complex (Bromeliaceae)
JO - Botanical Journal of the Linnean Society
DO - 10.1093/botlinnean/boz106
DA - 2020-03-27
UR - https://www.deepdyve.com/lp/oxford-university-press/population-genetics-shed-light-on-species-delimitation-and-life-Qsigz2oZNx
DP - DeepDyve
ER -