cDepartamento de Genética, IB/Universidade Federal do Rio Grande do Sul–UFRGS, P.O. Box 15053, 91501-970 Porto Alegre-RS, Brazil
dDepartamento de Ecologia, Campus de Rio Claro, Universidade Estadual de São Paulo, 13506-900 Rio Claro, SP, Brazil
eUnit of Ecology and Evolution, Department of Biology, University of Fribourg, Chemin du Musée 10, CH-1700 Fribourg, Switzerland,
Received 18 September 2014. Revised 5 December 2014. Accepted 4 January 2015. Available online 7 January 2015. Edited by Bohumil Mandák.
Highlights
•
Vriesea gigantea is a cryptically self-incompatible species.
•
Vrieseagigantea has a mixed mating system according to all experiments performed.
•
High selfing and low inbreeding depression from center to periphery in V. gigantea.
•
Studying shifts in mating systems provides key information on Bromeliads evolution.
Abstract
Bromeliaceae is a Neotropical family that evolved ecological key innovations in association with extensive adaptive radiation. Its species present a variety of different mating system strategies varying within and among species, within genera and subfamilies. Also, species with a wide geographical range can display large variation in mating system, reproductive success and genetic diversity. Here we combined data from hand pollinations and genetic analysis to assess outcomes of contemporary gene flow and mating system variation at the range edge of Vriesea gigantea. Results from pollen germination rates showed that this species is cryptically self-incompatible. Hand-pollination experiments and genetic analysis of progeny arrays revealed that V. gigantea has a mixed mating system, with high selfing rates (s= 0.612), and high inbreeding coefficient (F = 0.372). Inbreeding in V. gigantea at southern edge of its distribution range was caused by high levels of selfing rather than by mating among relatives. Moreover, strong pollen pool genetic structure was observed (Φ’FT = 0.671), with an increase from north to south. The parameters observed help us to understand historical and ecological conditions under which V. gigantea has experienced moderate to high levels of selfing in the face of reduced pollen flow from central to peripheral populations due to recent southward range expansion.
Keywords
Bromeliaceae
Cryptic self-incompatibility
Microsatellites
Mixed mating system
Outcrossing rates
Introduction
Neotropics comprise the most species-rich region on Earth (Antonelli and Sanmartín, 2011). Nevertheless, relatively fewer genetic studies have examined Neotropical plant species, for example those species from the Atlantic Forest (AF), compared to temperate species (Hewitt, 2004, Jaramillo et al., 2006 and Turchetto-Zolet et al., 2013). Bromeliaceae is an example of a Neotropical group that has evolved “key innovations” in association with its extensive adaptive radiation, like tank habit and crassulacean acid metabolism (CAM) (Givnish, 2010, Silvestro et al., 2014 and Givnish et al., 2014). Moreover, it comprises subfamilies with different and relatively recent evolutionary histories. In particular, tank bromeliads from the AF have the highest diversification rate among bromeliad species, which can be associated with ecological trends such as evolution of differences in growth form, floral morphology, and mating systems (Benzing, 2000, Wendt et al., 2008, Givnish, 2010, Matallana et al., 2010, Givnish et al., 2011, Silvestro et al., 2014 and Givnish et al., 2014). Studies on variation of reproductive traits within species can provide important insights in evolution and radiation of recent groups from Neotropics, especially in widespread species adapted to different ecological circumstances (Barrett, 2008 and references therein; Levin, 2012), and should be an important aspect for populations occurring at the species’ distribution margins, where adaptations to specific conditions are decisive for their persistence (Darling et al., 2008, Geber, 2008, Sexton et al., 2009and Moeller et al., 2012).
Factors which determine species’ range limits include abiotic and biotic factors, which are fundamentally related to the ecological niche of species (Sexton et al., 2009). A recent evolutionary model includes the notion that a fundamental parameter of mating system evolution, the inbreeding depression, may shape the association of traits along geographic distributions and create stable limits of edge dispersion if inbreeding depression decreases from center to periphery. This parameter is thus worth considering when analyzing empirical data (Pujol and Pannell, 2008, Cheptou and Massol, 2009, Pujol et al., 2009 and Sun and Cheptou, 2012). Moreover, according to Baker’s law, plants colonizing populations located at the margins of a distribution area tend to be autogamous as a consequence of pollinator or mate scarcity (Baker, 1955, Baker, 1959 and Sun and Cheptou, 2012). Indeed, Kalisz et al. (2004), based on inbreeding depression and marker-based estimates of selfing for wild populations of Collinsia verna (Scrophulariaceae), demonstrated that when pollinator visits decrease, populations shift toward intermediate selfing rates through an increase in the proportion of autonomously selfed seeds. As widely known, selfing may allow reproduction when circumstances reduce opportunities for a union of gametes produced by different individuals, a phenomenon known as reproductive assurance (Holsinger, 2000).
Furthermore, some previous reports on mating systems of bromeliads have been based on inferences from population genetic data (Wright’s F-statistics) for a single plant generation. They have indicated mixed system for different species, Alcantarea imperialis (Carrière) Harms, A. geniculata (Wawra) J.R. Grant (Barbará et al., 2007and Barbará et al., 2009), A. glaziouana (Lemaire) Leme and A. regina (Vell.) Harms (Barbará et al., 2008 and Barbará et al., 2009), V. gigantea (Palma-Silva et al., 2009), V. minarum (Baker) L.B.Sm. (Lavor et al., 2014) and outcrossing system for Bromelia antiacantha Bertol. (Zanella et al., 2011 see also Zanella et al., 2012 for a review), Encholirium horridum L.B.Sm. (Hmeljevski, 2013), and Pitcairnia encholirioidesL.B.Sm., which is a self-incompatible species (Hmeljevski et al., 2014). Determination of mating system through a more comprehensive approach, using combinations of direct experiments and genetic analysis (e.g., progeny arrays), has been reported only for three bromeliad species, Guzmania monostachia (L.) Rusby ex Mez. and Tillandsia fasciculata Sw., both species showed high selfing rates (Cascante-Marín et al., 2006), and Dyckia ibiramensis Reitz, which presented mixed mating system (Hmeljevski et al., 2011).
In plants, mating systems have a profound effect on the genetic composition of natural populations (Hamrick, 1982). Likewise, variation in gene flow via pollen transfer and seed dispersal also contributes to the shape of genetic diversity within populations (Charlesworth, 2003 and Duminil et al., 2007). Disruption of gene flow via pollen transfer and seed dispersal also contributes to loss of genetic diversity within populations isolated by habitat fragmentation (Hamrick, 1982), which may contribute to increase inbreeding coefficient. In a previous study, populations throughout entire V. gigantea (Bromeliaceae) distribution range showed latitudinal trends of decreasing genetic diversity and gene flow from north to south, which is consistent with historical forest expansion from northern half of the present distribution range (Palma-Silva et al., 2009). Further expansion toward the south appears to be inhibited by a lack of gene flow at edges of the current range (Palma-Silva et al., 2009 and Paggi et al., 2010). Accordantly, Palma-Silva et al. (2009) found variation in the inbreeding coefficient (FIS) among populations, with higher values occurring in V. giganteapopulations from southern edge of range. In addition, Sampaio et al. (2012) also found geographic variation in inbreeding depression estimates, with a moderate value (δ = 0.31), for the same central and southern populations of V. gigantea; those populations are the focus of the present study (Table 1). Herein, we investigate the mating system and contemporary gene flow of Vriesea gigantea at southern edge of the species range in Brazil, by combining data from hand-pollination experiments and genetic analysis of nuclear molecular markers. Specifically, we aimed to: a) characterize the mating system of V. gigantea at its southern range edge by estimating different parameters using progeny array and pollen pool genetic structure approaches, and by hand pollination experiments; and b) identify the role of two components, population subdivision or high selfing rate, that may be responsible for the high inbreeding coefficient (FIS) observed previously. Results will help us to understand the relative roles of pollen flow and selfing in shaping southern range limit in V. gigantea.
Table 1. Natural populations of V. gigantea in Atlantic Forest, Rio Grande do Sul Brazil. Populations sampled with their geographical parameters, analysis performed, and plant habit.
Population
Conservation unit
City
Geographical coordinates
Habitat
Mean of temperature
Annual rainfall mm/year
V. gigantea habit
Mating system and pollen germination
Molecular analysis
Maquiné
Serra Geral ecological reserve
Maquiné
29°48′S, 50°16′W
Subtropical ombrophilous forest
21.5 °C
1650
Epiphytic/saxicolous
–
X
Itapuã
Itapuã state park
Viamão
30°22′S, 51°01′W
Sandy soil and shrubs with high light incidence
17.5 °C
1200
Epiphytic/saxicolous
X
X
Taim
Taim ecological station
Rio grande and Santa vitória do palmar
31°56′S, 52°25′W
Sandy coastal shrubs
18 °C
1100
Epiphytic exclusively
–
X
Materials and methods
Study species
Vriesea gigantea (Tillandsioideae) presents an inflorescence with a central axis and several branches on each side, and flowers on each side of the lateral axis (Reitz, 1983; Fig. 1A). The flowers have three petals and their yellowish color and tube-shaped corolla are in accordance with the chiropterophilous syndrome (Vogel, 1969and Krömer et al., 2008). Besides, V. gigantea flowers lasted just one night (c.a. 12 h), and petals, stamens, and the style abscised during the following days. Also, the flowers present protogyny, in which style is receptive before anther dehiscence, in the beginning of anthesis (Paggi, 2009). The studied species is a tank bromeliad, frequently epiphytic or rupicolous, occurring in AF, across a wide geographic distribution, from Espírito Santo to Rio Grande do Sul states, Brazil (20°07′S, 40°30′W to 31°56′S, 52°25′W) (Fig. 1C). Vriesea gigantea is the only Vriesea species occurring in the southernmost part of AF (Smith and Downs, 1977, Reitz, 1983, Zimmermann et al., 2007, Martinelli et al., 2008 and Palma-Silva et al., 2009; Fig. 1C).
Fig. 1. (A) Individual showing large characteristic inflorescence of this species in the fruiting season. Photo: courtesy of CM Zanella. (B) Ovules showing pollen germination on a flower fixed 48 h after manual self-pollination, arrow showing a penetrated ovule. Photo: courtesy of N.S. Jr. Bittencourt. (C) Map showing the current geographic distribution of V. gigantea (Reitz, 1983, Zimmermann et al., 2007 and Palma-Silva et al., 2009) and the three studied sites from Rio Grande do Sul State, Brazil.
Study sites
This study was conducted in three populations of V. gigantea (Maquiné, Itapuã and Taim; Table 1), located in AF of Rio Grande do Sul state, Brazil. Vriesea gigantea is restricted to AF, which currently occurs in a narrow range in Brazilian coastal lowland, therefore we found single populations at different latitudes. Previous genetic analysis showed more central populations of Maquiné and Itapuã to have low gene flow with marginal population of Taim, the southernmost population (Palma-Silva et al., 2009). Climate in study region is classified as humid subtropical (Cfa) by the Köppen Climate Classification System. Fieldwork was carried out throughout the species’ flowering and fruiting seasons (January–September 2006/2008). Itapuã population is included in a protected area of Itapuã State Park, created in the 1970’s, which was previously affected by deforestation and stone extraction. However, there is a large number of specimens with more than 1000 adult plants distributed in clustered subpopulations in this population. Maquiné and Taim populations are located near the Ecological Reserve of Serra Geral and Taim Ecological Station, respectively, and have experienced some disturbance due to farming. In Maquiné population plants are distributed in clustered subpopulations, comprising more than 1000 adult individuals per subpopulation. In the southernmost population of Taim, specimens are distributed mainly on trees of Ficus organensis (Miq.) Miq., and in a lower density compared to other populations (personal observation).
Breeding system and pollen germination experiments
Hand-pollination experiments were performed in situ in Itapuã State Park population, where the experiment could be carried out without interference (Table 1). Controlled hand-pollination experiments were designed to test for spontaneous selfing (bagging) and outcrossing (in emasculated open flowers). Flowers under natural pollination conditions were used as controls. In emasculation process, anthers were removed gently before anthesis; in bagging process, mesh bags were used (maximum mesh size 0.9 mm) to cover the flowers. Three to five flowers were used per plant, in a total of 16 plants, totaling 47 flowers for the control, 48 for the spontaneous selfing, and 35 for the outcrossing treatments (Table 2). All treatments were applied in all sampled individual plants. Fruit set (i.e., the fraction of flowers that developed into mature fruit) was estimated as % fruits = number of fruits per plant/number of flowers per plant × 100, for sampled flowers per plant in each treatment. Differences in fruit set among treatments was detected using ANOVA followed by Tukey test (5%) (SAS, version 8).
Table 2. Hand pollination experiments in V. gigantea: number of fruits and flowers used in each treatment. Fruit setting (%) obtained in each pollination treatment. Standard errors are in parentheses.
Estimates with the same letter are not significant different, ANOVA followed by the Tukey test (5%), P = 0.0209.
In the same population, manual self- and cross-pollination treatments were performed to compare differences in pollen tube growth and number of penetrated and non-penetrated ovules. Experiments were performed at beginning of anthesis, when stigmas were receptive, and before anther dehiscence (Paggi, 2009). After 48, 72, or 96 h, flowers from hand self- and cross-pollination were fixed in 3:1 ethanol: glacial acetic acid, to allow further observation using fluorescence microscopy and aniline blue staining (Martin, 1959). Five to seven pistils per individual per treatment were prepared for examination, for a total of 33 pistils. Proportion of penetrated ovules in the self- and cross-pollinated pistils was assessed by scraping out ovules and scoring for presence of pollen tube “tails” at micropyle. Differences in the mean proportion of penetrated ovules in the self- and cross-pollinated pistils were detected using t-test (SAS, version 8).
Mating system analysis
Naturally pollinated maternal seeds were randomly collected from all available reproductive individuals at all studied populations, five mother plants (MPs) in Taim, 10 in Maquiné, and 11 in Itapuã, for a total of 26 MPs. Location of each MP was recorded using a global positioning system (GPS) receiver for TwoGener analysis. On average, 17 seedlings from each MP were used for genetic analysis, for a total of 501 progenies. The seedlings were obtained by in vitro seed germination, as described by Paggi et al. (2007). Fresh leaves (∼5 cm2) from MPs and seedlings were stored in liquid nitrogen until DNA extraction. Total genomic DNA was extracted using the protocol described by Doyle and Doyle (1990).
Eight microsatellite loci previously characterized for bromeliad species Tillandsia fasciculata, Guzmania monostachia (loci: e6b and CT5, respectively; Boneh et al., 2003), and V. gigantea (loci: VgA04, VgA06, VgB06, VgF01, VgF02, and VgG02; Palma-Silva et al., 2007) were selected for this study considering their appropriated amplification in study species and polymorphism observed in each sampled population. We performed PCR reactions in a PE Applied Biosystems 9700 thermocycler (Applied Biosystems, Foster City, CA, USA) as described by Palma-Silva et al. (2007). Microsatellite alleles were resolved on a 3100 DNA Analyzer (Applied Biosystems) and precisely sized against a ROX molecular size standard (Applied Biosystems) using Genescan 3.7 and Genotyper 3.7 (Applied Biosystems).
Multilocus outcrossing rate (tm) was estimated under mixed mating model described by Ritland (2002), with Mltr 3.4 (Ritland, 2009). This procedure allowed us to estimate multilocus outcrossing rate (tm), mean single-locus outcrossing rate (ts), difference between estimates (tm–ts), which represents the outcrossing rate between related individuals (=biparental inbreeding), and inbreeding coefficient of maternal parents (F). Significance of tm and F were determined using one-tailed Student’s t-tests based on null hypotheses that tm is significantly less than 1 and F is significantly greater than zero (SAS, version 8). Mltr software also estimates correlated mating parameters: correlation of selfing (rs), which reflects variation in selfing rates among families, and the correlation of paternity (rp), which is the probability that two siblings are outcrossed full-sibs (Ritland, 2002). Significance of rs was also determined using one-tailed Student’s t-tests based on null hypotheses that rs is significantly greater than zero (SAS, version 8). Standard errors were estimated based on 1000 bootstraps between individuals within a progeny array. From these estimates, we calculated the proportions of progeny that are full sibs (tm·rp) and half sibs [(tm (1 − rp)] (Mantovani et al., 2006).
TwoGener procedure was used to analyze genetic structure of pollen pools (ΦFT) of individual plants relative to global pollen pool (Smouse et al., 2001) based on AMOVA approach of Excoffier et al. (1992). We computed 99% confidence interval of ΦFT by bootstrapping among loci with 1000 replicates, using Poldisp 1.0 software (Robledo-Arnuncio et al., 2007). Significance of ΦFT was determined using one-tailed Student’s t-tests based on null hypotheses that ΦFT is significantly less than 1 (SAS, version 8).
Inbreeding coefficient was high and significant in adult populations of V. gigantea (FIS= 0.269; Palma-Silva et al., 2009), which may bias estimates of pollen pool genetic structure, increasing this estimate. Hence, the estimate of ΦFT was corrected using the formula described by Austerlitz and Smouse (2001), Φ’FT = ΦFT (1 + F), where ΦFTis pollen structure parameter due only to limited pollen dispersal, and F is adult inbreeding coefficient. Estimate of Φ’FT was used to calculate variance in pollen flow, computed as σ2 = (ΦFT8πd)−1, average distance of pollen dispersal (δ = √σ2π/2), effective neighborhood pollination area (Aep = 4π σ2), and effective number of pollen donors (Nep = 4π σ2d) (Austerlitz and Smouse, 2001). ANOVA followed by Tukey test were performed to determine whether genetic parameters differed among localities using families as grouping factors. All statistical analyses were performed using SAS (version 8).
Results
Breeding system and pollen germination experiments
Results from hand-pollination experiments showed that proportion of fruit set was 47.9% in control treatment, 27.1% in spontaneous selfing treatment, and 17.7% in emasculated treatment, which estimates the amount of pollen from other flower (from same or other individual) in the emasculated flowers (Table 2). Fluorescence microscopy analysis showed both self- and cross-pollinated grains germinated on stigma and produced tubes that grew toward ovary (Figs. 1B, 2). A higher proportion of penetrated ovules were observed in cross-pollinated flowers than self-pollinated at every time points tested. At 48 h after self-pollination and cross-pollination, 9.5% and 22.8% of ovules, respectively, were penetrated (F = 5.63, P = 0.0451). After 72 h, around 90% of ovules were penetrated in the cross-pollination treatment, while 76% of ovules were penetrated in self-pollination treatment (F = 33.47, P = 0.0001). Difference was even greater after 96 h, with 92% of cross-pollinated ovules penetrated and 68% of self-pollinated ovules penetrated (F = 10.69, P = 0.0114; Fig. 2). Pollen tubes from self-fertilization grew more slowly than pollen tubes from cross-fertilization; as a result, cross-pollination occurs foremost in flowers of V. gigantea.
Fig. 2. Proportion of penetrated ovules in 48, 72 and 96 h after self (triangles) and cross-pollination (circles).
Mating system and contemporary gene flow
Outcrossing rate (tm) was variable among populations, ranging from 0.176 to 0.477; outcrossing rate across all populations was 0.388 (Table 3). A significant self-fertilization rate (s = 1 − tm) was detected (s = 0.612) since the overall outcrossing rate was significantly lower than one (t= − 22.778, P < 0.0001). Outcrossing rate between relatives (biparental inbreeding; tm–ts) ranged from 0.069 to 0.141, south to center, respectively. Estimates for individual MPs showed significant correlation in selfing (rs, among-plant variation in the selfing rate, t = − 17.129, P < 0.0001) within the Maquiné population alone. Distribution of individual outcrossing rates for this population indicated correlation probably arose from few families exhibiting significant selfing (s > 0.20; data not shown).
Table 3. Estimates of mating system parameters (SE in parentheses) for three V. giganteapopulations from Rio Grande do Sul State, Brazil. * Mixed mating system parameters: F, (estimated) parental inbreeding; tm, multilocus outcrossing rate; ts, single-locus average outcrossing rate; (tm–ts), biparental inbreeding; rp, correlation of paternity; rs, correlation of selfing; and Nep, effective number of pollen donor.
Estimates with the same letter are not significant different, ANOVA followed by the Tukey test (5%). Correlation of selfing rs, (F = 8.92; P = 0.0024); multilocus outcrossing rate tm, (F = 5.69; P = 0.0106); and single-locus average outcrossing rate ts, (F = 7.06; P = 0.0036).
**
One-tailed Student’s t-tests: F > 0, t = − 12.524, P < 0.0001; tm < 1, t = − 22.778, P < 0.0001; rs > 0, t = − 17.129, P < 0.0001.
Inbreeding coefficient (F), which is based on genotypic frequencies in maternal plants, was positive and significantly greater than zero (F = 0.372; t= − 12.524, P < 0.0001; Table 3), considering all populations. Maquiné and Itapuã populations had higher outcrossing rates (0.477 and 0.336, respectively) compared to Taim population, which showed the highest selfing rate (0.824). Mating system parameters exhibited conspicuous trends toward southern margin of species’ range, with some parameters increasing and others decreasing (Table 3 and Table 4). Multi- and single-locus outcrossing rates and correlation of paternity showed significant differences among populations, being lower in southernmost population of Taim (Table 3).
Table 4. Estimates of pollen pool genetic structure parameters based on TwoGener Model for three V. gigantea populations from Rio Grande do Sul State, Brazil. TwoGener model parameters: Φ’FT, corrected pollen structure parameter; ΦFT, observed (uncorrected for inbreeding) pollen structure parameter; Nep, effective number of pollen donors; δ, average distance of pollination; and Aep, effective neighborhood pollination area.
Parameters not significantly different among populations. ANOVA followed by the Tukey test (5%). (F= 1.12; P= 0.3450).
***
One-tailed Student’s t-tests: Ф’FT < 1, t = − 24.164, P < 0.0001.
Genetic structure of pollen pool in V. gigantea populations was ΦFT = 0.532, indicating significant differentiation in pollen pools among populations (Table 4). Considering significant inbreeding coefficients previously observed in a “single-generation” population genetic study (FIS = 0.269; Palma-Silva et al., 2009), we corrected pollen pool structure parameter and found a slightly greater value of Ф’FT = 0.671 (t = − 24.164, P < 0.0001; Table 4). From this estimate (Ф’FT = 0.671), effective number of pollen donors (Nep) was calculated as 0.75, effective neighborhood pollination area (Aep) was 173 m2, and average distance of pollination (δ) was 147 m, considering all populations. In agreement with Ritland’s mating system parameters which decreased from north to south, Ф’FT values increased toward species’ range edge (Table 4).
Discussion
Combining results from hand-pollinations experiments, pollen germination and progeny array analysis, we showed that V. gigantea exhibits mixed mating system and cryptic self-incompatibility, consistent with levels of genetic diversity data reported by Palma-Silva et al. (2009). Population in extreme southern edge of the species' distribution (Taim) exhibited the highest selfing rates and pollen pool genetic structure. Results presented here are important for ongoing debates regarding limits to adaptation in marginal plant populations, particularly considering variation and evolution of plant mating systems. Ecological and genetic patterns play important roles in persistence of plant populations in marginal portions of their geographic distribution (Pannell and Dorken, 2006, Eckert et al., 2008, Geber, 2008, Pujol and Pannell, 2008, Foxe et al., 2009, Pujol et al., 2009 and Barringer et al., 2012).
Mixed mating system and cryptic self-incompatibility
Based on results from hand-pollination experiments and pollen germination, V. gigantea exhibits a mixed mating system, since it can set fruits through both self-fertilization and outcrossing (Table 2; Paggi et al., 2007). Mixed mating systems are particularly common in self-compatible species pollinated by animals (Barrett and Eckert, 1990, Barrett and Charlesworth, 1991, Vogler and Kalisz, 2001, Barrett, 2002, Barrett, 2003, Goodwillie et al., 2005, Futuyma, 2009 and Goodwillie et al., 2010). Moreover, selfing component of a mixed mating system could arise through geitonogamy (G.M. Paggi and C. Palma-Silva, personal observations) and might be a nonadaptive cost associated with the large floral display of V. gigantea (Reitz, 1983, Sazima et al., 1999 and Paggi et al., 2007), which is typically required to attract animal pollinators. Sazima et al. (1999), based on floral features and pollinator observations in populations from southeastern Brazil, concluded V. gigantea is an outcrossing and bat-pollinated species; nevertheless, that study was performed only at the center of the species' distribution (southeastern Brazil, 23°50′S, 46°08′W), and the present study was conducted at the central-southern edge of the species' range (Table 1). The geographic distribution of the bat species described as pollinator of V. gigantea in southeastern Brazil, Anoura caudifer É. Geoffroy, only reaches Maquiné population in southern Brazil; this bat species does not occur in Itapuã or Taim populations (Fábian et al., 1999 and Oprea et al., 2009). In addition, we observed populations from southern Brazil have diurnal anthesis and are preferentially pollinated by bees and hummingbirds rather than bats (Paggi, 2009). These facts indicate a shift in reproductive biology of V. gigantea across its geographic distribution. Populations at margin of a species’ distribution may experience poorer environmental conditions than populations in center of range (Geber, 2008). Therefore, marginal populations may have low fitness (Tsaliki and Diekmann, 2009and Barringer et al., 2012) or may face reductions in pollinator services resulting in higher cross-pollination limitation, which would consequently lead to high selfing rates through autonomous selfing, the reproductive assurance (Holsinger, 2000, Barrett, 2002, Darling et al., 2008, Zhang and Li, 2008, Cheptou, 2012 and Moeller et al., 2012). Surprisingly, based on flower, fruit and seed production (Paggi et al., 2007 and Sampaio et al., 2012) and adult morphology aspects, seed germination and seedling survival, high levels of plant fitness for V. gigantea populations at the range margin were reported (Sampaio et al., 2012), although pollen limitation was observed (Paggi et al., 2007 and Palma-Silva et al., 2008).
Pollen germination experiments revealed that V. gigantea is self-compatible as described by (Paggi et al., 2007), but presents cryptic self-incompatibility, which is a form of physiological SI found in self-compatible plants in which outcrossing is promoted through differential pollen-tube growth between outcross- and self-pollen tubes (Franklin-Tong, 2008). Several plant species exhibit cryptic forms of self-incompatibility in which seed set following cross- and self-pollination is similar but differences exist in germination speed and quality, suggesting that some form of competition exists between self- and cross-pollination (Good-Avila et al., 2008). Self-compatibility appears to be a widespread trait in Bromeliaceae, mainly in the subfamily Tillandsioideae, including species from Alcantarea, Guzmania, Racinea, Tillandsia, Vriesea, and Werauhia (Lasso and Ackerman, 2004, Ramírez-Morillo et al., 2004, Cascante-Marín et al., 2005, Cascante-Marín et al., 2006, Matallana et al., 2010and Schmid et al., 2011). Growing evidence indicates many self-compatible plants control the level of self-fertilization through post-pollination processes that give a siring advantage to cross-pollination over self-pollination through cryptic SI (Eckert and Allen, 1997). This mechanism has also been reported in several species of Lythraceae, Melastomataceae and Spigeliaceae as CSI and may function to reduce inbreeding (Eckert and Allen, 1997, Erbar and Leins, 1999, Erbar, 2003 and Santos et al., 2010). Paggi et al. (2013) reported self-sterility due to late-acting SI in a congener species, V. friburgensis Mez, another study conducted in a species’ range edge. In the present work, selection for reduced CSI may have occurred at periphery of species’ range, but this has not been tested nonetheless.
High selfing rates and inbreeding coefficient
Results from the progeny array showed moderate to low outcrossing rates in populations of V. gigantea (Table 3), which agrees with mixed mating system. High inbreeding coefficient observed (F = 0.212–0.718) was similar to previous results from same populations in AF (FIS = 0.225–0.489; Palma-Silva et al., 2009). This result can be explained by high levels of selfing (s = 0.612) rather than mating among relatives (tm–ts = 0.144) in studied populations (Table 3). Since in V. gigantea each plant opens many concurrently flowers per day (Sazima et al., 1999 G.M. Paggi personal observation), observed low outcrossing, high inbreeding coefficient and high levels of selfing are probably due to pollinator-mediated autogamy and geitonogamy, apparently contributing to high selfing rates in V. gigantea. Also, a positive value of rswas observed in Maquiné population, which means that two selfed progeny sampled from same progeny array are not result of two independent selfing events. For instance, such correlated selfing may occur when movement of a pollinator causes within-flower or -individual selfing (G.M. Paggi personal observation). Therefore such correlations are expected to be high among seeds within a single fruit (Takebayashi et al., 2006). Likewise, correlation of paternity (the probability of two siblings being outcrossed full-sibs) varied among populations, and was high in Maquiné, where outcrossing rate was also high, indicating that even with pollinator service, gene flow persist low, with few pollen donors per population and among populations (Table 3).
TwoGener analysis indicated significant differentiation in pollen pool genetic structure among populations, with pollen-mediated gene movement being severely limited (Φ’FT = 0.671; Table 4). Effective pollination neighborhood (Aep) around a MP was only 173 m2, suggesting that MPs are preferentially pollinated by pollen drawn from neighboring plants (average distance of pollination δ = 147 m; Table 4). This was confirmed by moderate level of biparental inbreeding (tm–ts = 0.144). Likewise, V. gigantea has low gene flow via seeds, since direct observations showed seeds disperse over short distances (<3 meters, Paggi et al., 2010). Moreover, low effective numbers of pollen donors (Nep; Table 3 and Table 4) indicate that the majority of seeds within each fruit were sired only by a few pollen donors, with few progeny having different fathers, either due to pollen carryover or multiple pollinator visits.
Considering populations of V. gigantea sampled throughout its geographic distribution, Palma-Silva et al. (2009) reported a strong negative correlation between genetic diversity and population latitude, consistent with historical Atlantic Forest expansion from northern half of present distribution range (Rambo, 1951, Rambo, 1960, Smith, 1962 and Turchetto-Zolet et al., 2013). Generally, populations at edges of species’ ranges may face higher risks of extinction if they are less able to respond to selection under a changing environment (Pannell and Dorken, 2006, Eckert et al., 2008, Geber, 2008, Pujol and Pannell, 2008 and Levin, 2012). Also, two major factors, the timing and a severe founder event (bottleneck), support hypothesis that selfing may have been favored during colonization as new habitats emerged after the last glaciation (Pannell and Dorken, 2006, Foxe et al., 2009 and Pujol et al., 2009). Studying populations of V. gigantea from the range margin, Sampaio et al. (2012)reported low to moderate levels of inbreeding depression (δ = 0.02 to 0.39), which was consistent with our previous results on fertility, genetic diversity and gene flow (Paggi et al., 2007 and Palma-Silva et al., 2009). Besides, Paggi et al. (2010)described short distances of seed dispersal for this species, in agreement with reported aggregate distribution of bromeliad seedlings (Zotz, 1997, van Dunné, 2001and Cascante-Marín et al., 2005), and with previous molecular studies, revealing V. gigantea populations are genetically structured, with low levels of gene flow (Palma-Silva et al., 2009 this study). All analyzed parameters indicates that historical and ecological conditions following the Pleistocene climatic oscillations are reflected in the contemporary genetic structure of V. gigantea, which has experienced reduction in gene flow (Palma-Silva et al., 2009 this study) and low to moderate levels of inbreeding toward the southern margin of its distribution (Sampaio et al., 2012).
CSI of V. gigantea favors germination and fertilization of outcross pollen grains over selfing ones, which results in a mixed mating system in this self-compatible species. Those features, combined with protogyny and floral display, may contribute to mixed mating system, since they promote delayed selfing and represent a trade-off between pollinator attraction and geitonogamous selfing. CSI may also contribute to the maintenance of high levels of genetic diversity and fitness and moderate levels of inbreeding depression (Paggi et al., 2007, Palma-Silva et al., 2008, Palma-Silva et al., 2009 and Sampaio et al., 2012 this study), which could facilitate the persistence of this species at its range edge. High selfing rates observed in the studied populations were the main processes that created high FIS values in the southern range margin. However, population subdivision due to biparental inbreeding (Table 3) could also contribute to elevated FIS values in populations at southern edge, since low levels of pollen flow (high estimates of pollen pool genetic structure) were observed in all populations.
Although number of investigated families and populations is limited, our results support the idea that selfing ability associated with a decrease in inbreeding depression (Sampaio et al., 2012) from center to periphery can have great importance for plant population establishment at stable range limits (Sun and Cheptou, 2012). This is in line with results for other plant groups and helps to explain the evolution of selfing in plant species (Darling et al., 2008, Geber, 2008, Pujol et al., 2009, Barringer et al., 2012 and de Groot et al., 2012). A solid knowledge base of genetic variation and reproductive biology in bromeliad species with different mating systems and ecological adaptations would provide key information for better understanding the importance of shifts in mating systems during diversification of Bromeliaceae. This information would also benefit studies that assess impacts of habitat fragmentation on biological diversity in this large ecologically important plant family that has undergone continent-wide adaptive radiation.
Acknowledgements
We thank T. Wendt, M.S. dos Reis, T. Souza-Chies, C.M. Zanella, M. Goetze, M.V. Büttow and T. Barbará for helpful discussions. We also thank to C.M. Zanella, J.A.T. Sampaio, L.C.T. da Silveira, M. Bruxel, M. Goetze, M.V. Büttow and G. Chazan for help with field and lab work. We thank to N.S.Jr. Bittencourt for help with pollen germination techniques. Thanks to N.R. Albuquerque for English revision and suggestions. We thank the staff of Itapuã State Park and Taim Ecological Station for helping to collect samples, especially Sr. J. Soares. We thank IBAMA (Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renováveis) and SEMA (Secretaria do Meio Ambiente do Estado) for processing of collection permits. This work was supported by grants from CNPq to C. Palma-Silva and G.M. Paggi; by awards from CAPES to C. Palma-Silva and from Andrew W. Mellon Foundation to G.M. Paggi; by grants from PPGBM to G.M. Paggi; and by grants from IFS to F. Bered.
T. Barbará, C. Lexer, G. Martinelli, S.J. Mayo, M.F. Fay, M. Heuertz
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)
T. Barbará, G. Martinelli, C. Palma-Silva, M.F. Fay, S. Mayo, C. Lexer
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)
T. Barbará, G. Martinelli, M.F. Fay, S.J. Mayo, C. Lexer
Population differentiation and species cohesion in two closely related plants adapted to neotropical high-altitude ‘inselbergs’ Alcantarea imperialis and Alcantarea geniculata (Bromeliaceae)
T.J. Givnish, M.H.J. Barfuss, B.V. Ee, R. Riina, K. Schulte, R. Horre, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H. Luther, W. Till, G. Zizka, P.E. Berry, K.J. Sytsma
Phylogeny adaptive radiation, and historical biogeography in Bromeliaceae: insights from an eight-locus plastid phylogeny
T.J. Givnish, M.H.J. Barfuss, B.V. Ee, R. Riina, K. Schulte, R. Horre, P.A. Gonsiska, R.S. Jabaily, D.M. Crayn, J.A.C. Smith, K. Winter, G.K. Brown, T.M. Evans, B.K. Holst, H. Luther, W. Till, G. Zizka, P.E. Berry, K.J. Sytsma
Adaptive radiation, correlated and contingent evolution: and net species diversification in Bromeliaceae
Mol. Phylogenet. Evol., Volume 71, 2014, pp. 55–78
Genetic and environmental causes and evolutionary consequences of variations in self-fertility in self incompatible species
Self-Incompatibility in Flowering Plants: Evolution, Diversity, and Mechanisms, V.E. Franklin-Tong, 2008, Springer, Verlag Berlin Heidelberg, DE, pp. 33–51
C. Goodwillie, R.D. Sargent, C.G. Eckert, E. Elle, M.A. Geber, M.O. Johnston, S. Kalisz, D.A. Moeller, R.H. Ree, M. Vallejo-Marin, A.A. Winn
Correlated evolution of mating system and floral display traits in flowering plants and its implications for the distribution of mating system variation
Evolution of self-fertilization at geographical range margins? A comparison of demographic, floral, and mating system variables in central vs. Peripheral populations of Aquilegia canadensis (Ranunculaceae)
Am. J. Bot., Volume 92, Issue 4, 2005, pp. 744–751
Biologia Reprodutiva, Genética De Populações E Filogeografia De Encholirium Horridum (bromeliaceae) Em Inselbergues Graníticos Da Mata Atlântica. Doctoral Thesis
2013, Instituto de Pesquisas Jardim Botânico do Rio de Janeiro, Brazil
K.V. Hmeljevski, L. Freitas, R. Domingues, A.R. Pereira, A.S. Cancio, A.C.S. Andrade, M.A. Machado, L.F. Viccini, R.C. Forzza
Conservation assessment of an extremely restricted bromeliad highlights the need for population-based conservation on granitic inselbergs of the Brazilian Atlantic Forest
C. Palma-Silva, M.M. Cavallari, T. Barbará, C. Lexer, M.A. Gimenes, F. Bered, M.H. Bodanese-Zanettini
A set of polymorphic microsatellite loci for Vriesea gigantea and Alcantarea imperialis (Bromeliaceae) and crossamplification in other bromeliad species
C. Palma-Silva, T. Wendt, F. Pinheiro, T. Barbará, M.F. Fay, S. Cozzolino, C. Lexer
Sympatric bromeliad species (Pitcairnia spp:) facilitate tests of mechanisms involved in species cohesion and reproductive isolation in Neotropical inselbergs
Ritland, K., 2009. Multilocus mating system program MLTR. Version 3.4. University of British Columbia, Canada. Free program distributed by the authors http://genetics.forestry.ubc.ca/ritland/programs
Diversity of flower visitors and their role for pollination in the ornithophilous bromeliad Vriesea friburgensis in two different habitats in Southern Brazil
Life-history traits evolution across distribution ranges: how the joint evolution of dispersal and mating system favor the evolutionary stability of range limits?
Corresponding author at: Departamento de Genética, IB/Universidade Federal do Rio Grande do Sul – UFRGS, P.O. Box 15053, 91501-970, Porto Alegre-RS, Brazil. Tel.: +55 51 33086742; fax: +51 51 33087311.
No comments:
Post a Comment