Molecular phylogeny, character evolution and historical biogeography of Cryptanthus Otto & A. Dietr. (Bromeliaceae)

Accepted Manuscript Molecular phylogeny, character evolution and historical biogeography of Cryptanthus Otto & A. Dietr. (Bromeliaceae) Geyner A.S. Cruz, Georg Zizka, Daniele Silvestro, Elton M.C. Leme, Katharina Schulte, Ana M. Benko-Iseppon PII: DOI: Reference:

S1055-7903(16)30281-0 http://dx.doi.org/10.1016/j.ympev.2016.10.019 YMPEV 5659

To appear in:

Molecular Phylogenetics and Evolution

Received Date: Revised Date: Accepted Date:

9 March 2016 24 October 2016 25 October 2016

Please cite this article as: Cruz, G.A.S., Zizka, G., Silvestro, D., Leme, E.M.C., Schulte, K., Benko-Iseppon, A.M., Molecular phylogeny, character evolution and historical biogeography of Cryptanthus Otto & A. Dietr. (Bromeliaceae), Molecular Phylogenetics and Evolution (2016), doi: http://dx.doi.org/10.1016/j.ympev. 2016.10.019

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Molecular phylogeny, character evolution and historical biogeography of Cryptanthus Otto & A. Dietr. (Bromeliaceae) Geyner A. S. Cruza, Georg Zizkab,c, Daniele Silvestro d, Elton M. C. Lemee, Katharina Schultef, Ana M. Benko-Iseppona a

Department of Genetics, CB, Universidade Federal de Pernambuco (UFPE), Av. Prof. Moraes Rego, 1235, CEP 50670-420, Recife, PE, Brazil

b

Department of Botany and Molecular Evolution, Senckenberg Research Institute Frankfurt and Goethe University, Senckenberganlage 25, 60325 Frankfurt/Main, Germany

c

Department of Diversity and Evolution of Higher Plants, Institute of Ecology, Evolution and Diversity, Goethe University Frankfurt, Senckenberganlage 25, 60325 Frankfurt/Main, Germany

d

Department of Plant and Environmental Sciences, University of Gothenburg, Carl Skottsbergs Gata 22B, 413 19 Gothenburg, Sweden;

e

Herbarium Bradeanum – HB, 20031-970, Rio de Janeiro, RJ, Brazil

f

Australian Tropical Herbarium (CNS) & Centre for Tropical Biodiversity and Climate Change (CTBCC), James Cook University, Cairns, Australia

*For correspondence: [email protected] / [email protected]

Abstract Cryptanthus comprises 72 species endemic to eastern Brazil with a center of diversity in the Atlantic Forest. The majority of the species are threatened due to habitat loss. We reconstructed phylogenetic relationships in Cryptanthus based on amplified fragment length polymorphisms (AFLP) including 48 species and 109 accessions. The Bayesian phylogenetic analysis revealed four major lineages in Cryptanthus and provided further evidence for the paraphyly of subgen. Hoplocryptanthus, while subgenus Cryptanthus was resolved as monophyletic. Monophyly of previously recognized morphological species groups at sectional level could not be confirmed. Based on the phylogenetic reconstruction we inferred the evolution of the sex system in Cryptanthus via maximum likelihood (ML) ancestral character reconstruction. Homoecy, the possession of hermaphrodite flowers only, was reconstructed as the ancestral state in the genus and characterizes three of the four main lineages within Cryptanthus. Andromonoecy, the possession of male and hermaphrodite flowers on the same plant, evolved only once and represents a synapomorphy of the fourth main lineage, C. subgen. Cryptanthus. The ancestral biome analysis reconstructed Cerrado (semiarid scrublands and forests) and campos rupestres (rock fields) as the most likely 1

ancestral biomes for the genus. A shift to the Atlantic Forest biome was reconstructed to have occurred twice, in the ancestor of the first diverging lineage within the genus and in the ancestor of the C. subgen. Cryptanthus clade. A shift to the Caatinga (tropical dryland savanna) and one reversal to Cerrado (campos rupestres - rock fields) was reconstructed to have occurred once, in C. bahianus and C. arelii, respectively. The ancestral biome reconstruction indicates a high degree of niche conservatism within Cryptanthus with rare biome shifts throughout the evolution of the genus. Further, our results imply that the current infrageneric taxonomy of Cryptanthus is problematic and requires revision.

Keywords: Bromelioideae, Cryptanthus, AFLP, Hoplocryptanthus, biome, niche conservatism.

1. Introduction The monocot family Bromeliaceae Juss. (58 genera, ca. 3,472 species, Gouda et al., 2015) is almost exclusively Neotropical, with a single disjunct species [Pitcairnia feliciana (A. Chev.) Harms and Mildbraed] occurring in West Africa (Porembski and Barthlott 1999; Jaques-Félix, 2000). The bromeliad family constitutes a noteworthy case of adaptive radiation and displays a high ecological versatility, occupying a wide range of terrestrial, lithophytic and epiphytic habitats including arid coastal plains, humid forests and dry environments on rock soils (Benzing, 2000; Rex et al., 2007; Schulte and Zizka, 2008). Based on the results of previous molecular phylogenetic studies eight subfamilies are recognized in the family: Brocchinioideae, Bromelioideae,

Hechtioideae,

Lindmanioideae,

Navioideae,

Pitcairnioideae,

Puyoideae and Tillandsioideae (Givnish et al., 2007; 2011). The subfamily Bromelioideae comprises 33 genera (Luther, 2012) distributed throughout Central and South America with a centre of diversity in eastern Brazil (Smith and Downs, 1979). Within Bromelioideae the genus Cryptanthus Otto & A. Dietr. comprises 72 terrestrial and rupicolous species (Luther, 2012) endemic to Brazil. The genus occupies a wide variety of habitats, such as the moist Atlantic Forest and semi-arid habitats like Cerrado, including campos rupestres (rock fields) and Caatinga (semiarid scrublands and forests). Cryptanthus grows in shady places in lowland forests as well as in exposed habitats at high altitudes. Centres of diversity

2

are the Atlantic Forest and campos rupestres in the states of Espírito Santo and Minas Gerais (Ramírez Morillo, 1996; Ramírez Morillo and Brown, 2001). Cryptanthus species (Fig. 1) have in general narrow geographic distributions. The majority of the species occur within the Atlantic Forest domain with 48 species being endemic to this biome. Due to the high extent of deforestation of the Atlantic Forest, 26 species of Cryptanthus are already included in the Brazilian list of threatened species (Martinelli et al., 2008). Given the high degree of endemism and narrow distributional ranges of most Cryptanthus species, habitat loss and fragmentation constitute major threats. Additionally, forest fragmentation leads to habitat dissection and invasion by weeds, which renders fragments prone to forest fires, and increases plant collecting by local human populations (Clark et al., 1995; Tabarelli et al., 2004). Additionally, some endemic species of Cryptanthus of the campos rupestres in Minas Gerais and Bahia are especially threatened, since these habitats are increasingly under anthropogenic pressure due to activities such as mining, stock farming, burning, and wood extraction from gallery forests (Versieux et al., 2008). In previous taxonomic treatments, Cryptanthus has been divided into two subgenera: (1) Cryptanthus, comprising andromonoecious plants, i.e. plants which harbour both hermaphroditic and male flowers on the same plant, and (2) Hoplocryptanthus Mez with homoecious plants (i.e. with hermaphroditic flowers only) (Mez, 1896, Ramírez Morillo, 1996; 1998). Cryptanthus has been regarded as a derived genus, mainly because of the occurrence of andromonoecy, fragrant flowers, CAM and unique chromosome numbers within Bromeliaceae, mainly n = 17 and n = 18, as compared to n = 25 for most remaining members of the family (Ramírez Morillo, 1996; Ramírez Morillo and Brown, 2001; Gitaí et al., 2014). The last monograph of the genus recognized five species groups (proposed as sections) for subgenus Cryptanthus and four for subgenus Hoplocryptanthus (Ramírez Morillo 1996). In previous molecular phylogenetic studies including Cryptanthus the genus was represented by only few taxa and formed a well-defined clade together with Orthophytum Beer and Lapanthus as early divergent lineages within the eubromelioids, yet relationships among the three genera remained unclear due to a lack of statistical support (Silvestro et al., 2014, Louzada et al., 2014). Due to limited sampling of Cryptanthus in previous phylogenetic studies, infrageneric relationships 3

in Cryptanthus are still unclear. Moreover, the taxonomic value of characters used in previous treatments of the genus has not been assessed in a phylogenetic framework yet. Cryptanthus subgenus Cryptanthus includes andromonoecious species, which constitutes a unique trait within the family. Previous molecular phylogenetic studies in Bromeliaceae based on DNA sequencing suffered from low sequence divergence of standard molecular markers, resulting in low resolution and support values among closely related species (e.g. (Aguirre-Santoro et al., 2015; de Sousa et al., 2007; Krapp et al. 2014). Amplified fragment length polymorphism (AFLP; Vos et al., 1995) is applicable to all organisms without a priori sequence information. It combines high reproducibility with high variability and enables a cost-efficient genome-wide sampling. AFLPs have proven to be a powerful tool to assess species relationships in evolutionary complex groups, such as those undergoing hybridization, introgression and polyploidization (Koopman et al., 2008; Perrie and Shepherd, 2009; Jakob and Blattner, 2010; Rebernig et al., 2010). The AFLP technique is well suited for analyzing phylogenetic relationships between closely related taxa (Hodkinson et al., 2000; Després et al., 2003), and has been applied successfully at inter- and intraspecific level within Bromeliaceae (Horres et al., 2007; Rex et al., 2009, Schulte et al., 2010). In this study we used AFLP data to reconstruct the infrageneric relationships of Cryptanthus. Based on the AFLP results we assessed whether taxa distinguished by morphological characters also represent evolutionary distinct, monophyletic units, and reconstructed the evolution of the sex distribution (andromonoecy versus homoecy) and the historical biogeography of the genus.

2. Material and Methods 2.1 Taxon sampling A data set of 109 accessions was analysed including 48 Cryptanthus and five Orthophytum species (Tab. 1 and Fig. 2), the latter chosen as outgroup based on previous phylogenetic analyses (Schulte et al., 2005; 2009; Schulte and Zizka, 2008; Givnish et al., 2011, Louzada et al., 2014). In cases where accessions could not be assigned unambiguously to a species, they are listed as C. aff. spec. or as C. spec., if no morphologically similar species could be identified. In at least some of these cases we suppose that these may constitute yet undescribed species. 4

Plant material was collected during field expeditions as well as from scientific living collections of Dr. Elton Leme (Refúgio dos Gravatás) and the Instituto de Botânica de São Paulo (Ibt), with vouchers deposited at Herbarium Bradeanum (HB) at the Universidade Estadual do Rio de Janeiro and the Herbarium of the Instituto de Botânica (SP), respectively. Leaf material collected was transferred into a sodium chloride saturated aqueous solution of CTAB (cetyl-trimethylammonium-bromide, 20 g/L) according to Rogstad (1992), and stored at 7°C until processing.

2.2 DNA Isolation Total genomic DNA was isolated according to Doyle and Doyle (1987) with modifications described by Weising et al. (2005). Moreover, DNA purification was performed including precipitation of polysaccharides (Michaels et al., 1994) and RNAse (10 μg/mL) treatment for 2 hours at 37 ºC. DNA quantification and quality control were carried out with a NanoDrop® 2000c (NanoDrop Technologies, Wilmington, DE), measuring its absorbance at 260 and 280 nm.

2.3 AFLP assays AFLP analyses were accomplished following Vos et al. (1995), with modifications by Bänfer et al. (2004), Rex et al. (2007) and Carmen Jung (personal communication, May 15, 2011). The first step consisted of genomic DNA digestion in a final volume of 25 L at 37°C with restriction endonucleases HindIII and MseI for 12 h and ligation to HindIII and MseI adapters in the same reaction. Two consecutive PCR amplifications were performed as pre-selective and selective amplification using an Eppendorf Mastercycler Pro thermal cycler using primers with one (+1), or three (+3) selective nucleotides at their 3’ ends. The reaction mix of pre-selective PCR (total volume: 10 L) contained 2 L of the 1:10 diluted restriction–ligation product, 0.5 M of unlabelled HindIII (+1) primer, 0.5 M of unlabelled MseI (+1) primer, 1 x PCR buffer (Peqlab, Germany), 2 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate (dNTP), and 0.025 U Taq DNA polymerase (Peqlab). Amplifications were subjected to an initial denaturation of 94ºC for 2 min followed by 30 cycles of amplification, each consisting of 94°C for 20 s, 56°C for 30 s, and 72°C for 2 min. Final extension was at 72°C for 2 min, followed by 60°C for 30 min.

5

An initial screening of selective primers using 12 primer combinations with three selective nucleotides each was performed with seven species of Cryptanthus. From this initial test nine primer combinations were chosen for the final analyses since they produced well scorable polymorphic patterns (Supplementary Table S1). Selective PCRs were carried out with 2.5 L of the 1:20 diluted preselective PCR product and different combinations of the unlabelled MseI (+3) primer (0.25 M) (Carl Roth, Karlsruhe, Germany) and the fluorescence-labelled HindIII primer (NED, VIC and FAM, Sigma-Aldrich, Munich, Germany) (0.05 M) with three selective bases. Furthermore, the PCR reaction contained 1 x PCR buffer (Peqlab), 2 mM MgCl2, 0.2 mM of each deoxynucleoside triphosphate (dNTP), and 0.025 U Taq DNA polymerase (Peqlab). The protocol included an initial denaturation at 94ºC for 2 min, followed by 35 cycles with the initial 15 cycles consisting of 94°C for 20 s, 66°C for 30 s and 72°C for 2 min, reducing the annealing temperature by 0.7°C at each step, followed by 20 cycles of 94°C for 20 s, 56°C for 30 s and 72°C for 2 min. Final extension was at 60°C for 30 min. Final products of the selective PCR were run on an automated sequencer (ABI Applied Biosystems) as a multiplex of three primer combinations labelled with distinct fluorescent dyes (NED, VIC and FAM) and an internal DNA size standard (ABI Applied Biosystems).

2.4 Data Analysis AFLP banding patterns were scored manually for presence or absence of a band at a particular position using the software GeneMarker, version 1.7 (SoftGenetics, State College, PA, USA). Two independent AFLP analyses of 10 accessions were run per primer pair, and the error rate was calculated according to the Jaccard distance (Holland et al. 2008).

2.5 Tree inferences Phylogenetic reconstructions based on the binary AFLP matrix were carried out using Neighbor-Joining (NJ) and Maximum Parsimony (MP) performed in PAUP 4.0b (Swofford, 2002) and Bayesian inference (BI) in MrBayes 3.2 (Ronquist et al., 2012). The neighbor-joining analyses were based on the Nei-Li distance measure (Nei and Li, 1979) with statistical support values for nodes estimated by bootstrap analyses (BS) with 1,000 replicates. For the parsimony analysis a strict consensus tree was 6

generated from heuristic searches with 10,000 random addition sequence (RAS) replicates, branch swapping via tree bisection reconnection (TBR), and the MULTREES option in effect. Statistical support of the tree topology was assessed with bootstrap analysis performing 1,000 pseudo-replicates, with 10 RAS replicates and TBR branch swapping. The extent of homoplasy was estimated using the consistency (CI) and retention indices (RI). A Bayesian analysis was conducted using the Metropolis-coupled Markov Chain Monte Carlo (MCMC) algorithm as implemented in the program MrBayes v. 3.2 (Ronquist and Huelsenbeck, 2003). Two evolutionary models, with and without the gamma distributed across site rate heterogeneity, were tested and compared by the respective marginal likelihoods. These values were calculated using the steppingstone algorithm (Xie et al., 2011) assuming 10 steps, each sampled for 1,000,000 generations after an initial burn-in. The substitution model with rate heterogeneity obtained a strong support (marginal likelihood = -20,016.76) over the simpler model without gamma variation (marginal likelihood = -20,755.86). Therefore, the Bayesian analysis was performed based on the gamma model with four independent MCMC runs and heated chains, for 10,000,000 generations, sampling trees every 1,000 iterations. The burn-in phase and the efficiency of the MCMC sampling were assessed by examining the log files with the program Tracer (Rambaut and Drummond, 2007). After excluding the burn-in fraction (i.e. the initial two million generations), the sampled trees from four independent runs were combined to generate a consensus tree with posterior probability values (PP) quantifying the statistical support for the nodes.

2.6 Ancestral character state and ancestral biome reconstruction We explored the evolution of the sex distribution in the genus by ancestral character mapping using the program R 3.2.1 (http://cran.r-project.org). The ancestral characters were calculated using the function ‘ace’ with the maximum likelihood optimization as implemented in the package Ape (Paradis et al., 2004). The phylogenetic signal was tested through the re-rooting method described by Yang et al. (1995), using the ER model, as implemented in the package Phytools (Revell, 2012). The character states (andromonoecy or homoecy) were coded for each species represented in the phylogeny and plotted onto the Bayesian consensus tree.

7

An ancestral biome reconstruction was conducted by tracing the occurrence of Cryptanthus species in Brazil’s biomes along the Bayesian consensus tree. The coding of the biomes were 1) Atlantic Forest, 2) Caatinga, and 3) Cerrado including campos rupestres. The reconstruction of the ancestral habitats was inferred using the ER rerooting model described by Yang et al. (1995) as implemented in the R-package Phytools.

3. Results 3.1 AFLP data Distinct AFLP profiles with nine primer sets (Supplementary Table S1) from 109 accessions, including the outgroup, produced 489 characters. The scored characters per primer combination varied between 44 and 66 with the fragment sizes ranging from 90 to 480 bp. Of 489 scored characters, 98% were variable in Cryptanthus. The estimated error rate (Jaccard distance) varied between 27.1 and 34.9% with a mean of 32.6%. The pairwise Nei-Li distance values within the total data set ranged from 0.05 to 0.48. The lowest value was detected at the intraspecific level within C. sergipensis, whereas the highest occurred at interspecific level between C. beuckeri and C. lavrasensis.

3.2 Phylogenetic relationships Tree inferences from the binary character matrix were made using Bayesian inference, maximum parsimony and neighbor-joining (Fig. 3, Supplementary material Fig. S1 and S2). The parsimony analysis yielded a total of 28 most parsimonious trees with length of 5,793 steps. The consistency index (CI) for these trees was 0.08 and the retention index (RI) was 0.47. From 489 characters in total, 470 were parsimonyinformative and 15 characters constant. In general, the tree topologies from the three analyses were largely congruent with only minor incongruences at lower taxonomic levels. In the following the results from the Bayesian inference are presented (Fig. 3). The Bayesian consensus tree shows four major Cryptanthus clades (Fig. 3). Clades I to III comprise the species from subgen. Hoplocryptanthus, which is shown as paraphyletic with clades I to III forming a basal ladderized grade. Clade III is sister to the clade IV, which comprises all species of subgenus Cryptanthus. Clade I which constitutes the first diverging group is weakly supported (PP 0.70, BS 42). It comprises 8 species and 14 accessions. C. aff. scaposus branches off 8

first, followed by C. pseudoscaposus sister to a clade with two subclades, A and B. Subclade A comprises C. aff. leuzingerae, C. pseudoglazioui, C. scaposus and C. whitmanii. Within subclade B C. odoratissimus (E34E) and C. latifolius are shown in sister group position to a clade, comprising C. odoratissimus (21E), C. pseudoglazioui, C. santaluciae, and C. microglazioui. The two samples of C. santalucieae form a highly supported monophyletic group (BS 100), and the two samples of C. microglazioui are also depicted as monophyletic, whereas the C. odoratissimus and C. pseudoglazioui are shown as non-monophyletic. Clade II receives maximum support values (PP 1, BS 100) and comprises eight species and 16 accessions. Cryptanthus caracensis, C. lavrasensis and C. glazioui form a clade sister to the remainder, which split into a dichotomy with subclade C comprising C. aff. glazioui (24E), C. aff. schwackeanus (43E), C. regius and C. tiradentensis. Subclade D harbours C. schwackeanus, C. aff. regius (11E), C. ferrarius and C. aff. glazioui (18E). The monophyly of C. tiradentensis is moderately supported (PP 97, BS 65), and three taxa are shown as non-monophyletic (C. aff. glazioui, C. ferrarius, C. schwackeanus). Clade III receives moderate support (PP 0.98, BS 66) and comprises the two species C. micrus and C. leopoldo-horstii. Cryptanthus micrus is nested within accessions of C. leopoldo-horstii, a relationship which is moderately supported (PP 0.87, BS 72). Within the Clade IV, the subclade E received moderate statistical support (PP 0.85 and BS 65) and harbours 19 species, one putatively new taxon (C. spec. nov. 88E), and 34 accessions. The two samples of C. capitellatus form a clade, which branches off first, followed by C. viridipetalus, which is found in sister group position to a larger clade. Within the latter, high support was found for the monophyly of C. bahianus (PP 1, BS 100). Four taxa are depicted as non-monophyletic: C. beuckeri, C. aff. burle-marxii, C. colnagoi, and C. dianae. Cryptanthus coriaceus is found in sister group position to the remainder of clade IV. In turn, subclade F comprises 15 species and 35 accessions. Cryptanthus sergipensis forms a highly supported monophyletic group (PP 1, BS 100). One well supported clade (PP 0.96, BS 67) was formed by all accessions of C. alagoanus, with C. pickelii nested within, indicating C. alagoanus being not monophyletic. Three further taxa were depicted as non-monophyletic: C. beuckeri, C. sinuosus, and C. aff. bivittatus.

9

The species shown as non-monophyletic in the phylogenetic reconstructions were each confined to one of the four main clades. For example, C. beuckeri was found in both subclades E (with 2 accessions) and F (with 4 accessions).

3.4 Ancestral character and ancestral biome reconstructions The reconstruction of the evolution of the sex system (homoecious vs. andromonoecius) showed homoecy as the ancestral character state in Cryptanthus. A single origin of andromonoecy was reconstructed at the base of the Cryptanthus subgen. Crypanthus clade. Within the latter, no reversal of character states was observed. Thus, andromonoecy can be considered a synapomorphy of subgenus Cryptanthus (Fig. 4). The ancestral biome analysis reconstructed Cerrado and campos rupestres as most likely ancestral biomes for Cryptanthus (Fig. 5). According to the ancestral area reconstruction, a shift to the Atlantic Forest occurred twice, in the ancestor of the first diverging clade (clade I) and in the ancestor of the C. subgen. Cryptanthus clade. Within the latter subgenus, one reversal to Cerrado and campos rupestres was observed (in C. arellii), and one shift to the Caatinga biome (in C. bahianus). The ancestral biome for clade II and III (both subgen. Hoplocryptanthus) was inferred as Cerrado and campos rupestres.

4. Discussion Within Bromelioideae, Cryptanthus possesses two distinct features: the occurrence of andromonoecy in subgenus Cryptanthus (Ramírez Morillo, 1996; 1998) and chromosome numbers unique within the family (2n = 32-34; Brown and Gilmartin 1989b; Gitaí et al., 2014). Previous molecular studies showed that Cryptanthus constitutes an early diverging lineage within the eu-bromelioid clade (Schulte and Zizka, 2008, Schulte et al. 2009, Silvestro et al., 2014). The studies also confirmed the close phylogenetic relationships of Cryptanthus, Orthophytum and Lapanthus, which together form a monophyletic group (Louzada et al., 2014, Silvestro et al., 2014). These three genera are terrestrial and lack the tank habit, whereas the tank habit characterizes the species rich eu-bromelioid clade. The genus Cryptanthus occurs in different Brazilian biomes, such as the seasonally dry Caatinga and Cerrado (campos rupestres) and displays its highest species richness in the humid Atlantic Forest biome. 10

So far, phylogenetic relationships in Cryptanthus are poorly understood, and infrageneric concepts, such as the subgeneric classification of the genus (Mez, 1896, Ramírez Morillo, 1996) or the sectional classification proposed by Ramírez Morillo (1996), have not been tested in a phylogenetic framework yet. Further, species delimitation in Cryptanthus is often problematic, which is in part due to poorly typified species (e.g. C. bromelioides, C. praetextus, C. sinuosus) and to limited insights into the plasticity of morphological characters and their taxonomic value. This explains the high number of accessions without identification or assigned to general affinity (affinis) to a putative closely related species or assignment to a species complex (e.g. C. acaulis complex, C. beuckeri complex). This AFLP study provides insights into infrageneric relationships in Cryptanthus, the taxonomic value of morphological characters within the genus, and the diversification of the group within Brazil’s biomes.

Phylogenetic relationships Our phylogenetic study showed four main lineages within Cryptanthus, with the

first

three

diverging

clades

formed

by

representatives

of

subgen.

Hoplocryptanthus, thus implying its paraphyly, and the fourth lineage composed of all representatives of subgen. Cryptanthus, thus supporting its monophyly. The latter is morphologically well characterized by the presence of andromonoecy, which is a unique trait within the family, sublinear-lanceolate petals, pollen with broadly reticulate surface and usually scentless flowers (Ramírez Morillo, 1996; Leme et al., 2010). The species of subgen. Hoplocryptanthus are characterized by hermaphroditic, usually fragrant flowers, broadly spathulate or obovate petals, and pollen with a smooth or minutely reticulate surface (Ramírez Morillo, 1996, Leme et al., 2010). Subgenus Cryptanthus occurs from the State of Rio de Janeiro in the south, through Minas Gerais and Espírito Santo to Rio Grande do Norte in the north, growing from sea level to app. 700 m a.s.l. In contrast, subgenus Hoplocryptanthus is distributed in the mountainous areas of the Atlantic Forest in Espírito Santo and in Cerrado specifically in campos rupestres of the Espinhaço range in Minas Gerais, mainly at elevations above 600 m a.s.l. (Ramírez Morillo, 1996; Leme and Siqueira-Filho, 2006; Leme et al. 2010). The differences in floral and pollen characters point towards different pollinators in both groups (Ramírez Morillo, 2001).

11

In her monograph of Cryptanthus, Ramírez Morillo (1996) recognized 38 species (later 45 spp., Ramírez Morillo, 1998) and confirmed the two major infrageneric groups treated as subgenera established by Mez (1896) based on a cladistic analysis of mainly morphological characters. Furthermore, she recognized five informal sections in subgen. Cryptanthus (Cryptanthus, Bahianae, Beuckeriae, Lacerdae, Zonatae), and four informal sections in subgen. Hoplocryptanthus (Hoplocryptanthus, Mesophyticae, Schwackeanae, Xerophyticae). However, these taxa were not validly published. Thus, the proposed sections will be referred here as groups without precise taxonomic rank and not printed in italics. Despite several more recent, relevant publications on Cryptanthus, the work of Ramírez Morillo (1996, 1998) remains a very detailed and comprehensive treatment of the genus, thus an important data source and therefore the recognized species groups are discussed here in the light of our molecular data.

Subgenus Cryptanthus The species of the subgenus Cryptanthus form a monophyletic group (clade IV, Fig. 3), which received moderate statistical support. The AFLP tree shows two major subclades within subgen. Cryptanthus plus C. coriaceus, which was found in sister group position to these. Comparing the molecular phylogeny with the subgeneric classification of Ramírez Morillo (1996) shows that representatives of four of the five groups are found in both major subclades (Bahianae, Beuckeri, Cryptanthus and Zonatae group), the Lacerdae group is confined to subclade E, and none of them form monophyletic groups. Thus, the infrageneric classification of subgenus Cryptanthus proposed by Ramírez Morillo (1996) is not supported by our molecular results (Tab. 1). The five morphological groups for subgenus Cryptanthus were largely defined based on leaf characters, such as shape, indumentum, succulence, and spines. Our study implies that these characters possess a high level of homoplasy within subgen. Cryptanthus and are of only limited taxonomic value. Looking at the subgroups of subclade F and E, the first one comprises four groups. In subclade F a clear geographical group may be recognized, mostly including species occurring north of São Francisco River from the states of Alagoas (AL), Pernambuco (PE), and Sergipe (SE), considered isolated from the remaining ones (Siqueira et al., 2007). Within this complex C. pickelii, C. alagoanus and C. felixii share morphological features, formed in a well-supported group. Moreover the 12

mentioned species have also a similar distribution in common, in the Pernambuco center of endemism (Siqueira et al., 2007). Additionally, C. burle-marxii and C. zonatus are characterized by a unique pattern of peltate trichomes on the foliar blades (Ramírez Morillo, 1996), and are shown as non-monophyletic in the subclades E and F, implying that this morphological character is of only limited taxonomic value. Cryptanthus sergipensis and C. bahianus belong to the Bahianae group and present some morphological similarities such as narrow triangular leaves, serrate leaf margin, and adaxially glabrous and abaxially lepidote leaf surfaces (Ramírez Morillo, 1996). In our phylogenetic reconstruction, C. bahianus formed a highly supported group within subclade E and C. sergipensis was found as a highly supported group in subclade F within the subgenus Cryptanthus clade, thus providing evidence that the Bahianae group is not a natural group (Fig. 3). Cryptanthus beuckeri is known from the southern region of Bahia state and from northern Espírito Santo (Thomas et al., 2003). It is the sole member of the Beuckeriae group, characterized by its petiolate leaves (Ramírez Morillo, 1996). Nevertheless, the monophyly of C. beuckeri remained unsupported by our phylogenetic results with representatives of this species being found in subclades E and F, indicating that petiolate leaves may have arisen several times in the evolution of the genus. Similar situations of non-monophyletic position of some members of the same species in the phylogeny have also been observed in other fast-evolving bromeliad genera (e.g. Tillandsia; Barfuss et al. 2005; Puya; Jabaily and Sytsma 2010; Dyckia; Krapp et al. 2014) and may be the result of hybridization and introgression events. Considering the actual configuration of this group within the phylogeny, further studies are required to clarify the taxonomic status of Cryptanthus plants with petiolate leaves. Likewise, the monophyly of the Cryptanthus group proposed by Ramírez Morillo (1996) remained unsupported in our molecular study. The species C. acaulis, C. bromelioides, C. colnagoi, C. coriaceus, C. correia-araujoi, C. dianae, C. lutherianus, C. marginatus, C. maritimus and C. ubairensis are found in both major subclades of the subgenus Cryptanthus (Fig. 3).

Subgenus Hoplocryptanthus Mez Our phylogenetic study resolved subgen. Hoplocryptanthus as a paraphyletic group, representing three major lineages (I-III) (Fig. 3). The four morphological 13

groups discerned in subgen. Hoplocryptanthus by Ramírez Morillo (1996), i.e. Hoplocryptanthus, Mesophyticae, Schwackeanae and Xerophyticae, did not form monophyletic groups in our molecular study. Instead, representatives of the respective groups were found in more than one of the three major Hoplocryptanthus clades (Hoplocryptanthus, Mesophyticae, and Xerophyticae group), or grouped together with representatives of other morphological groups within one of the major clades (Schwackeanae group, clade II) (Fig. 3, Tab. 1). The Hoplocryptanthus group is primarily defined by a long-caulescent habit, being here represented by C. microglazioui and C. pseudoglazioui which were found in clade I, and C. glaziouii which was part of clade II, (Fig. 3), implying that the character used to circumscribe this group is probably homoplastic. The Mesophyticae group is mainly characterized by leaf morphological characters of the indumentum (Ramírez Morillo, 1996). The representatives of this group (C. latifolius, C. odoratissimus, C. pseudoscaposus, C. scaposus and C. whitmanii) where found in clade I. Similar to the sections defined based mainly on leaf morphological characters in subgen. Cryptanthus, in subgen. Hoplocryptanthus indumentum characters, such as the distribution and density of leave trichomes, were found to be of limited taxonomic value and appear unsuitable for the infrageneric classification of the genus. Clade II is strongly supported (Fig. 3), and includes species that share a suite of morphological similarities with C. schwackeanus, although with distinct synapomorphies among them (Leme, 2007; Leme and de Paula, 2009; Ramírez Morillo, 1996), for example C. regius and C. tiradentesensis, which form a well supported subclade and which share morphological features such as stemless habit, propagation by short basal rhizomes and bipinnate inflorescence (Leme, 2007). Nevertheless, clade II also harbours species with different morphology, and from all four morphological groups discerned by Ramírez Morillo (1996). Previous molecular studies also showed subgen. Hoplocryptanthus as a paraphyletic group (Silvestro et al., 2014, Louzada et al., 2014). Nevertheless, relationships between the major Cryptanthus clades did not receive statistical support, thus further molecular studies are required to clarify relationships between the major clades, as well as between Cryptanthus and the closely related genera Orthophytum and Lapanthus. Next-generation sequencing approaches such as targeted sequence

14

capture of nuclear and plastid markers are promising molecular tools for such future studies.

4.2 Ancestral character reconstruction Andromonoecy is considered a derived character in Bromeliaceae, taking into account that almost all members of the family are homoecious. Only Hechtia Klotzsch (Hechtioideae) includes species with male flowers and hermaphroditic flowers on the same individual (Smith and Downs, 1977; Brown and Gilmartin, 1989a). Thus, the andromonoecy observed in the subgenus Cryptanthus is a unique character state in the family (Ramírez Morillo, 1996). In the andromonoecious Cryptanthus species, staminate flowers open first and are located mainly in the mid to apical sector of the inflorescence, while the perfect flowers are concentrated in the basal racemes (Leme et al., 2010). Within an inflorescence, the staminate flowers are smaller than the hermaphroditic ones (Ramírez Morillo, 1996). The reconstruction of the evolution of staminate flowers within the genus Cryptanthus reveals a clear pattern (Fig. 4). In our study, homoecy was inferred as ancestral within the genus. The ancestral character reconstruction showed a single origin of the andromonoecy, with all species of the subgenus Cryptanthus being characterized by andromonoecy. When acquiring this innovation, the group may have been favored by some advantages such as resource reallocation, in which the production of staminate flowers reduces the resource investment in functionally male flowers, thus enabling the resources saved to be re-allocated to other fitness-enhancing traits (Bertin, 1982; Solomon, 1985; Spalik, 1991; Emms, 1993). Sex allocation is an example of this process, where the resource-depleting factors such as shading or water stress (Solomon, 1985), florivory (Krupnick and Weis, 1998) and fruiting success of earlier flowers (Diggle, 1994; Gibbs et al., 1999) may alter the production of perfect and staminate flowers. The staminate flowers can be more efficient as pollen donor and pollinator attractor for many reasons, such as higher production of pollen. Furthermore, may also reduce the possibility of self-pollination (Podolsky, 1992, 1993; Harder and Barrett, 1996; Elle and Meagher, 2000; Barrett, 2003). These hypotheses have been tested in different andromonoecious species. For instance, Zhang and Tan (2009) observed in the shrub Capparis spinosa L. that male flowers save resources for female function and they primarily serve to attract pollinators as 15

pollen donors. Interestingly, the few genera with other than perfect flowers (Cryptanthus, Hechtia, Catopsis) occur in a variety of habitats (dry to humid), grow terrestrially as well as epiphytic but are similar in having less showy and usually small, often white flowers that are presumably pollinated by insects.

4.3 Ancestral biome analysis The Cerrado biome (including campos rupestres), is a semiarid habitat which shelters a peculiar vegetation due to poor shallow soils on rock underground and high radiation levels (Porembski et al., 1998; Porembski and Barthlott, 2000). In our analysis the Cerrado biome including campos rupestres was inferred as ancestral for the genus, and only few shifts between biomes were reconstructed (Fig. 5). For the Atlantic Forest biome, which is characterized by a strong seasonality and diverse landscape including semi-deciduos and deciduous forests (Tabarelli et al., 2010), two shifts were reconstructed, in the ancestor of the first diverging lineage within the genus, and in the ancestor of subgenus Cryptanthus clade. The first three clades diversified exclusively within only one of the biomes, that is in the biome which was reconstructed as ancestral to them (clade I: Atlantic Forest, clade II and III: Cerrado and campos rupestres). In the fourth clade, the ancestral biome of subgen. Cryptanthus (Atlantic Forest) was retained with only two exceptions: one shift to the tropical savanna biome Caatinga (in C. bahianus), and one reversal to Cerrado and campos rupestres (in C. arelii). Thus, the ancestral biome reconstruction shows a strong degree of niche conservatism in Cryptanthus with rare biome shifts throughout the evolution of the genus. A high degree of niche conservatism has also been found in ancestral habitat reconstructions for other neotropical plants lineages, including Bromeliaceae (Särkinnen et al. 2012, Wagner et al., 2013). Our results have important conservation implications, as our ancestral biome reconstruction implies that bromeliad species adapted to a certain biome may have only a low potential to adapt to different environmental conditions within short evolutionary timeframes. Today, the Atlantic Forest harbours the greatest species diversity not only within Cryptanthus but also for the species rich core bromelioids, with many endemic species. Our ancestral biome reconstruction implies that species adapted to moist forest environments, such as those within subgen. Cryptanthus, have a low potential to adapt and persist into the future if their habitat is altered through deforestation or climate change. 16

Our molecular study also showed another interesting pattern: a high degree of homoplasy is found in morphological characters within Cryptanthus, whereas habitat preferences appear to be more conserved within the genus. Our comparison of groups based on morphology, mainly vegetative characters, showed a high degree of parallel evolution of similar traits in different lineages, which rendered the recognition of natural groups and phylogenetic relationships based on morphology difficult. The results of our ancestral biome reconstruction for the ancestor of the whole genus is of course strongly influenced by the outgroup choice. So far, the relationships among the three closely related genera Cryptanthus, Orthophytum and Lapanthus are not fully understood and phylogenetic reconstructions lack support for the deeper nodes (Silvestro et al., 2014, Louzada et al., 2014, this study). Therefore, further studies are needed to clarify these relationships and to reconstruct the biogeographic history of this larger alliance. We regard the Cryptanthus/Orthophytum/Lapanthus clade, which is the most diverse among the early diverging eu-bromelioids, as a key group for the understanding of the radiation of Bromelioideae in the South American savannah and forest biomes.

5. Conclusions Our AFLP study offers the first phylogeny, ancestral character reconstruction and historical biogeography analysis of Cryptanthus. Our phylogenetic reconstruction supported the monophyly of subgenus Cryptanthus, which is well characterized by the presence of andromonoecy, while subgen. Hoplocryptanthus was shown to be paraphyletic and requires revision. Further molecular studies are needed to clarify phylogenetic relationships within the Cryptanthus/Orthophytum/Lapanthus clade, which will form the basis for a taxonomic revision of these three genera.

Acknowledgments This study is part of a doctoral dissertation by GAS Cruz undertaken at the Federal University of Pernambuco. We acknowledge support from the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) and the German Academic Exchange Service (DAAD) for a fellowship at the Goethe University Frankfurt, Germany, awarded to GAS Cruz. This work was supported by 17

CAPES/DAAD (PROBRAL project), by the Programa de Apoio e Desenvolvimento da Botânica (PNADB/CAPES), by the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq), and by the Deutsche Forschungsgemeinschaft (DFG, ZI 557/7-1). We thank Dr. Rafael Batista Louzada, Dr. Diego de Barros Pinangé and Dr. Santelmo Selmo de Vasconcelos Junior for valuable discussions and assistance during the field work. The authors also acknowledge the comments received by the anonymous reviewers. References Aguirre-Santoro, J., Betancur, J., Brown, G.K., Evans, T.M., Salgueiro, F., AlvesFerreira, M., Wendt, T., 2015. Is Ronnbergia (Bromeliaceae, Bromelioideae) a geographically disjunct genus? Evidence from morphology and chloroplast DNA sequence data. Phytotaxa 3, 261-275. Bänfer, G., Fiala, B., Weising, K., 2004. AFLP analysis of phylogenetic relationships among myrmecophytic species of Macaranga (Euphorbiaceae) and their allies. Plant Syst. Evol. 249, 213-231. Barrett SCH., 2003. Mating strategies in flowering plants: the outcrossing-selfing paradigm and beyond. Philos. T. Roy. Soc. B. 358, 991–1004. Barffus, M.H.J., Samuel, R., Till, W., Stuessy, T.F., 2005. Phylogenetic relationships in subfamily Tillandsioideae (Bromeliaceae) based on DNA sequence data from seven plastid regions. Am. J. Bot. 92, 337-351. Benzing, D.H., 2000. Bromeliaceae: profile of an adaptive radiation, Cambridge University Press, Cambridge, UK. 690 p. Bertin, R.I., 1982. The evolution and maintenance of andromonoecy. Evol. Theory 6, 25–32. Brown, G.K., Gilmartin, A.J., 1989a. Stima types in Bromeliaceae-A Sistematic survey. Syst. Bot. 14, 110-132. Brown, G.K., Gilmartin, A.J., 1989b. Chromosome numbers in Bromeliaceae. Amer. J. Bot. 76, 657–665. Clark, D.A., Clark, D.B., Sandoval, M.R., Castro, C.M.V., 1995. Edaphic and human effects on landscape-scale distributions of tropical rain forest palms. Ecology. 76, 2581–2595. Després, L., Gielly, L., Redoutet, B., Taberlet, P., 2003. Using AFLP to resolve phylogenetic relationships in morphologically diversified plant species 18

complexes when nuclear and chloroplast sequences fail to reveal variability. Mol. Phylogenet. Evol. 27, 185–196. de Sousa, L.O.F., Wendt, T., Brown, G.K., Tuthill, D.E., Evans. T.M., 2007. Monophyly and phylogenetic relationships in Lymania (Bromeliaceae : Bromelioideae) based on morphology and chloroplast DNA sequences. Syst. Bot. 32, 264-270. Diggle,

P.K.,

1994.

The

expression

of

andromonoecy

in Solanum

hirtum (Solanaceae): phenotypic plasticity and ontogenetic contingency. Am. J. Bot. 81, 1354–1365. Doyle, J.A., Doyle, J.L., 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochem. Bull. 19, 11–15. Elle, E., Meagher, T.R., 2000. Sex allocation and reproductive success in the andromonoecious perennial Solanum carolinense (Solanaceae). II. Paternity and functional gender. Amer. Nat. 156, 622– 636. Emms, S.K., 1993. Andromonoecy in Zigadenus paniculatus (Liliaceae): spatial and temporal patterns of sex allocation. Am. J. Bot. 80, 914–923. Gibbs, P.E., Lewis, G.P., Lughadha, E.N., 1999. Fruit-set induced changes in the sex of flowers in Caesalpinia calycina (Leguminosae). Plant Biol. 1, 665–669. Gitaí, J., Paule, J., Zizka, G., Schulte, K., Benko-Iseppon, A.M., 2014. Chromosome Numbers and DNA Content in Bromeliaceae: New Data and Critical Review. Bot. J. Linn. Soc. 176, 3: 349-368. Givnish, T.J., Milliam, K.C., Berry, P.E., Sytsma, K.J., 2007. Phylogeny, adaptive radiation, and historical biogeography of Bromeliaceae inferred from ndhF sequence data. Aliso 23, 3–26. Givnish, T.J., Barfuss, M.H., Van Ee, B., Riina, R., Schulte, K., Horres, R., Gonsiska, P., Jabaily, R., Crayn, D., Smith, J., Winter, K., Brown, G., Evans, T., Holst, B., Luther, H., Till, W., Zizka, G., Berry, P., Sytsma, K., 2011. Phylogeny, adaptive radiation, and historical biogeography in Bromeliaceae: Insights from an eight-locus plastid phylogeny. Am. J. Bot. 98, 872–895. Gouda, E.J., Butcher, D., Gouda, K., 2015. Encyclopaedia of Bromeliads Version 3.1. Available from: http://botu07.bio.uu.nl/bcg/encyclopedia/brome/ (accessed 30 January 2015)

19

Harder, L.D., Barrett, S.C.H., 1996. Pollen dispersal and mating patterns in animalpollinated plants. In: Lloyd, D.G., Barrett, S.C.H., eds. Floral biology: studies on floral evolution in animal-pollinated plants. New York, pp. 140–190. Hodkinson, T.R., Renvoize, S.A., Chonghaile, G.N., Stapleton, C.M.A., Chase, M.W., 2000. A comparison of ITS nuclear rDNA sequence data and AFLP markers for phylogenetic studies in Phyllostachys (Bambusoideae, Poaceae). J. Plant Res. 113, 259–269. Holland, B.R., Clarke, A.C., Meudt., H.M., 2008. Optimizing automated AFLP scoring parameters to improve phylogenetic resolution. Syst. Biol. 57, 347– 366. Horres, R., Schulte, K., Weising, K., Zizka, G., 2007. Systematics of Bromelioideae (Bromeliaceae) – evidence from molecular and anatomical studies. Aliso. 23, 27– 43. abail

R. .

tsma K.J., 2010. h logenetics of Puya (Bromeliaceae): placement

ma or lineages and evolution of hilean species. Am. J. Bot. 97, 337–356. Jakob, S.S., Blattner, F.R., 2010. Two extinct progenitors were involved in allopolyploid formation in the Hordeum murinum (Poaceae: Triticeae) taxon complex. Mol. Phylogen. Evol. 55, 650–659. Jacques-Félix H., 2000. The discovery of a bromeliad in Africa: Pitcairnia feliciana. Selbyana 21, 118-124. Koopman, W.J.M., Wissemann, V., De Cock, K., Van Huylenbroeck, J., De Riek, J., Sabatino, G.J.H., Visser, D., Vosman, B., Ritz, C.M., Maes, B., Werlemark, G., Nybom, H., Debener, T., Linde, M., Smulders, J.M., 2008. AFLP markers as tool to reconstruct complex relationships: a case study in Rosa (Rosaceae). Am. J. Bot. 95, 353–366. Krapp, F., de Barros Pinange, D.S., Benko-Iseppon, A.M., Leme, E.M.C., Weising, K., 2014. Phylogeny and evolution of Dyckia (Bromeliaceae) inferred from chloroplast and nuclear sequences. Plant Syst. Evol. 300, 1591-1614. Krupnick, G.A., Weis, A.E., 1998. Floral herbivore effect on the sex expression of an andromonoecious plant, Isomeris arborea (Capparaceae). Plant. Ecol. 134, 151–162. Leme, E.M.C., Siqueira-Filho, J.A., 2006. Taxonomia das bromélias dos fragmentos de Mata Atlântica de Pernambuco e Alagoas. In: Siqueira Filho, J.A. & Leme, E.M.C. (eds.). Fragmentos de Mata Atlântica do Nordeste, Biodiversidade, 20

Conservação e suas Bromélias. Andrea Jakobsson Estúdio, Rio de Janeiro. pp.190-381. Leme, E.M.C., 2007. "Three Subtle New Cryptanthus Species from Espinhaço Range, Minas Gerais, Brazil." J. Bromeliad Soc. 57, 259-272. Leme, E.M.C., de Paula, C.C., 2009. Cryptanthus ferrarius, a new species from the iron-rich soils of Minas Gerais, Brazil, on the way to extinction. J. Bromeliad Soc. 59, 104-108. Leme, E.M.C., Fraga, C.N., Kollmann, L.J.C., Brown, G., Till, W., Ribeiro, O.B.C., Machado, M.C., Monteiro, F.J.S., Fontana, A.P., 2010. Miscellaneous new species in the Brazilian Bromeliaceae. Rodriguésia. 61, 21-67. Louzada, R.B., Schulte, K., Wanderley, M.G.L., Silvestro, D., Zizka, G., Barfuss, M.H.J., Palma-Silva, C., 2014. Molecular phylogeny of the Brazilian endemic genus Ortophytum (Bromelioideae, Bromeliaceae) and its implications on morphological character evolution. Mol. Phylogenet. Evol. 77, 54-64. Luther, H.E., 2012. An Alphabetical List of Bromeliad Binomials. Holst, B.K. & Rabinowits, L. (eds.), Marie Selby Botanical Gardens, Sarasota & The Sarasota Bromeliad Society, 44 pp. Martinelli, G., Vieira, C.M., Gonzalez, M., Leitman, P., Piratininga, A., Costa, A.F., Forzza, R.C., 2008. Bromeliaceae da Mata Atlântica Brasileira: Lista de Espécies, Distribuição e Conservação. Rodriguésia. 59, 209-258. Mez, C., 1896. Bromeliaceae. In: C. De Candolle. Monographiae Phanerogamarum. Paris. Sumptibus Masson, 9, l-144. Michaels, S.D., John, M.C., Amasino, R.M., 1994. Removal of polysaccharides from plant DNA by ethanol precipitation. BioTechniques. 17, 274–276. Nei, M., Li, W.-H., 1979. Mathematical model for studying genetic variation in terms of restriction endonucleases. Proc. Natl. Acad. Sci. USA 76, 5269-5273. Paradis, E., Claude, J., Strimmer, K., 2004. APE: Analyses of Phylogenetics and Evolution in R language. Bioinformatics. 20, 289-290. Perrie, L.R., Shepherd, L.D., 2009. Reconstructing the species phylogeny of Pseudopanax (Araliaceae), a genus of hybridizing trees. Mol. Phylogenet. Evol. 52, 774–783. Podolsky, R.D., 1992. Strange floral attractors: pollinator attraction and the evolution of plant sexual systems. Science. 258, 791–793.

21

Podolsky, R.D., 1993. Evolution of a flower dimorphism: how effective is pollen dispersal by male flowers. Ecology. 74, 2255–2260. Porembski, S., Martinelli, G., Ohlemüller, R., Barthlott, W., 1998. Diversity and ecology of saxicolous vegetation mats on inselbergs in the Brazilian Atlantic rainforest. Divers. Distrib. 4, 107-119. Porembski, S., Barthlott, W., 1999. Pitcairnia feliciana: the only indigenous African bromeliad. Harvard Pap. Bot. 4, 175-184. Porembski, S., Barthlott, W., 2000. Granitic and gneissic outcrops (inselbergs) as centers of diversity for desiccation-tolerant vascular plants. Plant Ecol. 151, 19-28. Rambaut,

A.,

Drummond,

A.J.,

2007.

Tracer:

Available

from

http://

beast.bio.ed.ac.uk/tracer. Ramírez-Morillo, I., 1996. Systematics, Phylogeny and Chromosome Number Evolution in Cryptanthus (Bromeliaceae). Ph.D. Thesis, University of Missouri–St. Louis, pp. 1–263. Ramírez Morillo, I.M., 1998. A systematic revision of Cryptanthus: major results. J. Bromeliad Soc. 48, 50-59. Ramírez Morillo, I.M., 2001. Pollination in Cryptanthus: a hypotesis. J. Brom. Soc. 52, 65-70. Ramírez-Morillo, I., Brown, G.K., 2001. The origin of the low chromosome number in Cryptanthus (Bromeliaceae). Syst. Bot. 26, 722–726. Rebernig, C.A., Weiss-Schneeweiss, H., Schneeweiss, G.M., Schonswetter, P., Obermayer, R., Villasenor, J.L., Stuessy, T.F., 2010. Quaternary range dynamics and polyploid evolution in an arid bushland plant species (Melampodium cinereum, Asteraceae). Mol. Phylogen. Evol. 54, 594–606. Revell, LJ., 2012. Phytools: an R package for phylogenetic comparative biology (and other things). Methods Ecol Evol. 3, 217-223.
 Rex, M., Patzolt, K., Shulte, K., Zizka, G., Vásquez, R., Ibisch, P.L., Weising, K., 2007. AFLP analysis of genetic relationships in the genus Fosterella L.B. Smith (Pitcairnioideae, Bromeliaceae). Genome. 50, 90-105. Rex, M., Schulte, K., Zizka, G., Peters, J., Vásquez, R., Ibisch, P.L., Weising, K., 2009. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae,

22

Bromeliaceae) based on four chloroplast DNA regions. Mol. Phylogen. Evol. 51, 472–485. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572-1574. Roquist, F., Teslenko, M., Van der Mark, P., Ayres, D.L., Darling, A., Höhna, S., Larget, B., Liu, L., Suchard, M.A., Huelsenbeck, J.P., 2012. MrBayes 3.2: Efficient Bayesian phylogenetic inference and model space. Syst. Biol. 61, 539-542. Rogstad, S.H., 1992. Saturated NaCl-CTAB solution as a means of field preservation of leaves for DNA analyses. Taxon 41, 701–708. Särkinen, T., Pennington, R.T., Lavin, M., Simon, M.F., Hughes, C.E., 2012. Evolutionary islands in the Andes: persistence and isolation explain high endemism in Andean dry tropical forests. Journal of Biogeography, 39, 884– 900. Schulte, K., Horres, R., Zizka, G., 2005. Molecular phylogeny of Bromelioideae and its implications on biogeography and the evolution of CAM in the family (Poales, Bromeliaceae). Senckenbergiana Biol. 85, 113–125. Schulte, K., Silvestro, D., Kiehlmann, E., Vesely, S., Novoa, P., Zizka, G., 2010. Detection of recent hybridization between sympatric Chilean Puya species (Bromeliaceae) using AFLP markers and reconstruction of complex relationships. Mol. Phylogenet. Evol. 57, 1105-1119. Schulte, K., Zizka, G., 2008. Multi locus plastid phylogeny of Bromelioideae (Bromeliaceae) and the taxonomic utility of petal appendages and pollen characters. Candollea 63, 209–255. Schulte, K., Barfuss, M.H.J., Zizka, G., 2009. Phylogeny of Bromelioideae (Bromeliaceae) inferred from nuclear and plastid DNA loci reveals the evolution of the tank habit within the subfamily. Mol. Phylogenet. Evol. 51, 327–339. Silvestro, D., Zizka G., Schulte, K., 2014. Disentangling the effects of key innovations on the diversification of Bromelioideae (Bromeliaceae). Evolution, 68, 163-175. Siqueira, J.A.F., Santos, A.M.M., Leme, E.M.C., 2007. "Atlantic Forest Fragments and Bromeliads in Pernambuco and Alagoas: Distribution, Composition, 23

Richness and Conservation". IN: Filho, J. A. Siqueira & Elton Martinez Carvalho Leme. Fragmentos de Mata Atlântica do Nordeste, Biodiversidade, Conservação e suas Bromélias. Rio de Janeiro, Andrea Jakobsson Estúdio, 1, pp. 100-131. Smith, L.B. and Downs, R.J., 1977. Tillandsioideae (Bromeliaceae). Flora Neotropica Monograph 14: 661–1492. Hafner Press, New York. Smith, L.B. and Downs, R.J., 1979. Bromelioideae (Bromeliaceae). Flora Neotropica Monograph 14: 1493–2141. Hafner Press, New York. Solomon, B. P., 1985. Environmentally influenced changes in sex expression in an andromonoecious plant. Ecology 66, 1321–1332. Spalik, K., 1991. On evolution of andromonoecy and overproduction of flowers: A resource-allocation model. Biol. J. Linnaean Soc. 42, 325–336. Swofford, D.L., 2002. Paup: Phylogenetic analysis using parsimony (and other methods). Sinauer Associate, Sunderland. Tabarelli, M., Silva, J.M.C., Gascon, C., 2004. Forest fragmentation, synergisms and the impoverishment of neotropical forests. Biodiversity and Conservation. 13, 1419–1425. Tabarelli, M., Aguiar, A.V., Ribeiro, M.C., Metzger, G.P., Peres, C.A., 2010. Prospects for biodiversity conservation in the Atlantic Forest: Lessons from aging human-modified landscapes. Biol. Conserv. 143, 2328-2340. Thomas, W.W., Jardim, J.G., Fiaschi, P., Amorim, A.M., 2003. Preliminary list of locally endemic plants Southern Bahia and Northern Espírito Santo, Brazil. Corredor de Biodiversidade da Mata Atlântica do Sul da Bahia. 1-9. Versieux L.M., Wendt T., Louzada R.B., Wanderley M.G.L., 2008. Bromeliaceae da Cadeia do Espinhaço. Megadiversidade. 4, 99-110. 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. Nucl. Acids Res. 23, 4407–4414. Wagner, N., Silvestro, D., Brie, D., Ibisch, P.L., Zizka, G., Weising, K., Schulte, K., 2013. Spatio-temporal evolution of Fosterella (Bromeliaceae) in the Central Andean biodiversity hotspot. J. Biogeogr. 40, 869-880. Weising, K., Nybom, H., Wolff, K., Kahl, G., 2005. DNA fingerprinting in plants, principles, methods, and applications. Second Ed. CRC Press, Taylor & Francis Group, Boca Raton, FA. 24

Xie, W., Lewis, P.O., Fan, Y., Kuo, L., Chen, M.H., 2011. Improving marginal likelihood estimation for Bayesian phylogenetic model selection. Syst. Biol. 60, 150-160. Yang, Z., Kumar, S., Nei, M., 1995. A new method of inference of ancestral nucleotide and amino acid sequences. Genetics. 141, 1641-1650.
 Zhang, T. and Tan, Z.Y., 2009. An examination of the function of male flowers in an andromonoecious Shrub Capparis spinosa. JIPB. 51, 316-324.

25

Figure 1: Examples of plants studied. A. Cryptanthus odoratissimus. B. Cryptanthus bromelioides. C. Cryptanthus sanctaluciae. D. Cryptanthus ferrarius. E. Cryptanthus pseudoscaposus. F. Cryptanthus beuckeri. G. Cryptanthus leopoldo-horstii. H. Cryptanthus schwackeanus. I. Cryptanthus zonatus. J. Cryptanthus capitellatus. Figure 2: Map showing the geographical distribution of studied Cryptanthus species and biomes. The black and red circles on the map show locations of species of subgenera Cryptanthus and Hoplocryptanthus respectively and the blue circles the occurrence of both. Figure 3: Bayesian consensus tree of 104 Cryptanthus accessions based on 489 characters obtained with nine AFLP primer pair combinations, with Orthophytum as outgroup. Values above branches indicate posterior probability values of the Bayesian analysis / the bootstrap support for the Parsimony and Neighbor Joining analyses. Clades I to IV are referred to in the text.

Figure 4: Maximum likelihood reconstruction of the evolution of the sexual system in Cryptanthus, based on relationships revealed by Bayesian consensus tree of nine AFLP primer pair combinations. Orange: homoecy; blue: andromonoecy.

Figure 5: Maximum likelihood ancestral habitat reconstruction in Cryptanthus, based on relationships revealed by Bayesian consensus tree of nine AFLP primer pair combinations. Yellow: Cerrado and campos rupestres; green: Atlantic Forest; red: Caatinga.

Figure S1: Strict consensus tree of a parsimony analysis based on 489 characters obtained with nine AFLP primer pair combinations and Orthophytum as outgroup. The analysis yielded 28 most parsimonious trees of 5793 steps length (consistency index CI = 0.08, retention index RI = 0.47). The squares indicate the type of the subgenus, the references are indicated above in the figure.

Figure S2: Neighbor Joining tree based on Nei-Li distances of 489 characters obtained with nine AFLP primer pair combinations and Orthophytum as outgroup. Bootstrap values are given above branches. 26

27

28

29

30

Table 1. Studied material with voucher numbers, provenance, taxonomic and phylogenetic informations. Abbreviations: ASE, Herbarium of Universidade Federal de Sergipe; BHCB, Herbarium of Universidade Federal de Minas Gerais; HB, Herbarium Bradeanum; IBt, Instituto de Botânica; LC, Living collection, RB, Herbarium of Jardim Botânico do Rio de Janeiro; RG, Refúgio dos Gravatás, living collection, in Teresópolis, Rio de Janeiro; SP, Herbarium of Instituto de Botânica; UFP, Herbarium Geraldo Mariz of Universidade Federal de Pernambuco.

Voucher

DNA No.

Locality

Taxonomic information (Ramirez-Morillo, 1996)

C. acaulis (Lindl.) Beer

Leme 3359 (RG)

17E

RJ, Rio de Janeiro

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade F

C. alagoanus Leme & J.A.Siqueira

Leme 6186 (RG)

62E

PE, Ipojuca

-

Clade IV – subclade F

Leme 6188 (RG)

77E

PE, Ipojuca

-

Clade IV – subclade F

Leme 5193 (RG)

81E

PE, Ipojuca

-

Clade IV – subclade F

14E

BA, Chapada Diamantina

-

Clade IV – subclade E

33E

BA, Bahia

-

Clade IV – subclade E

65E

BA, Santa Terezinha

Cryptanthus Species

C. arelii H.Luther C. argyrophyllus Leme C. bahianus L.B.Sm.

Leme 2830 (RG) Leme 5143 (RG) Leme 4223 (RG) Oliveira et al. 45628 (UFP)

GRV4 PE, Gravatá

Oliveira et al. 45629 (UFP)

GRV6 PE, Gravatá

AFLP clade membership

Clade IV – subclade E Subgenus Cryptanthus, Bahianae group

Clade IV – subclade E Clade IV – subclade E

31

C. beuckeri E.Morren

16E

BA, Arraial da Ajuda

Clade IV – subclade E

Leme 6786 (RG)

2E

ES, Linhares

Clade IV – subclade F

Leme 4020 (RG)

3E

BA, Bahia

Clade IV – subclade F

Leme 7341 (RG)

54E

BA, Camacan

Leme 6994 (RG)

9E

BA, Itapebi

Clade IV – subclade F

Leme 5151 (RG)

58E

BA, Potiraguá

Clade IV – subclade F

C. bromelioides Otto & Dietrich.

Leme 2229 (RG)

4E

RJ, Rio de Janeiro

Leme 3595 (RG)

53E

RJ, Rio de Janeiro

C. capitellatus Leme & L.Kollmann

Leme 7988 (RG)

23E

ES, Santa Teresa

Leme 6921 (RG)

26E

ES, Várzea Alegre

C. caracensis Leme & E.Gross

Leme 1853 (RG)

89E

MG, Santa Bárbara

C. colnagoi Rauh & Leme

Leme 7898 (RG)

32E

MG, Salto da Divisa

Leme 1021 (RG)

66E

BA, Bahia

Leme 5313 (RG)

76E

BA, Itapetinga

Leme et al. 1114

91E

ES, Serra

C. coriaceus Leme

Leme 3891 (RG)

Subgenus Cryptanthus, Beuckeriae group

Subgenus Cryptanthus, Cryptanthus group

Subgenus Hoplocryptanthus, Xerophyticae group

Clade IV – subclade E

Clade IV – subclade F Clade IV – subclade E Clade IV – subclade E Clade IV – subclade E Clade II Clade IV – subclade E

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade E Clade IV – subclade E

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade E

32

(HB) C. correia-araujoi Leme

Leme et al. 2704 (HB)

39E

ES, Domingos Martins

C. delicatus Leme

Leme 2236 (RG)

72E

C. diamantinensis Leme

Leme et al. 3812 (HB)

C. dianae Leme

C. dorothyae Leme C. felixii J.A.Siqueira & Leme

C. ferrarius Leme & C.C.Paula

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade F

RJ, Campos

-

Clade IV – subclade E

64E

BA, Caeté-Açu

-

Clade IV – subclade E

Leme 5038 (RG)

31E

PE, Jaqueira

Leme 5037 (RG)

84E

PE, Jaqueira

Leme 6814 (RG)

90E

PE, Jaqueira

19E

ES, Presidente Kennedy

Leme et al. 6100 (HB)

44E

PE, Bonito

Leme et al. 5540 (HB)

47E

AL, Serra Lisa

Clade IV – subclade F

Leme 6890 (RG)

27E

MG, Catas Altas

Clade II – subclade D

Leme 6540 (RG)

60E

MG, Mariana

Leme 6541 (RG)

67E

MG, Mariana

Leme 2379 (RG)

Clade IV – subclade E Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade E Clade IV – subclade E

-

Clade IV – subclade F Clade IV – subclade F

-

-

Clade II – subclade D Clade II – subclade D

33

C. giganteus Leme & A.P.Fontana C. glazioui Mez

Leme 7988 (RG)

Leme 1856 (RG)

20E

MG, Salto da Divisa

5E

MG, Santa Bárbara

C. lacerdae Antoine

Leme 3400 (RG)

38E

BA, Camaçã

C. latifolius Leme

Leme 5220 (RG)

12E

ES, Guarapari

8E

MG, Santa Bárbara

41E

MG, São Gonçalo R. Pratas

Leme 5565 (RG)

82E

MG, Diamantina

Louzada et al. 384520 (SP)

R20

MG, Minas Gerais

C. lutherianus I.Ramirez

Leme 3083 (RG)

22E

ES, Ibiraçu

C. lyman-smithii Leme

Leme 4344 (RG)

10E

BA, Jaguaripe

Leme et al. 4342 (HB)

57E

BA, Jaguaripe

80E

ES, Domingos Martins

C. lavrasensis Leme C. leopoldo-horstii Rauh

C. marginatus L.B.Sm.

Leme 7620 (RG) Leme 5657 (RG)

Leme 0290 (RG)

Subgenus Hoplocryptanthus, Hoplocryptanthus group

Clade IV – subclade E Clade II

Subgenus Cryptanthus, Lacerdae group

Clade IV – subclade E

Subgenus Hoplocryptanthus, Mesophyticae group

Clade I – subclade B

-

Clade II Clade III

Subgenus Hoplocryptanthus, Xerophyticae group

Clade III Clade III

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade F Clade IV – subclade F

-

Subgenus Cryptanthus, Cryptanthus group

Clade IV – subclade F Clade IV – subclade E

34

C. maritimus L.B.Sm. C. microglazioui I.Ramirez

C. micrus Louzada, Wand. & Versieux C. odoratissimus Leme

Leme et al. 1582 (HB)

92E

ES, Vila Velha

15E

ES, Domingos Martins

Louzada et al. 396794 (SP)

R13

ES, Espírito Santo

Louzada et al. 1474 (BHCB)

R116

MG, Minas Gerais

34E

ES, Santa Leopoldina

21E

ES, Domingos Martins

Louzada sn IBt (LC)

R9

PE, Pernambuco

Leme et al. 1560 (HB)

42E

ES, Santa Leopoldina

Leme et al. 1556 (HB)

73E

ES, Santa Leopoldina

74E

ES, Domingos Martins

Subgenus Hoplocryptanthus, Mesophyticae group

49E

MG, Tiradentes

-

Leme 0152 (RG)

Leme 5207 (RG) Leme 5216 (RG)

C. pickelii L.B.Sm. C. pseudoglazioui Leme

C. pseudoscaposus L.B.Sm. C. regius Leme

Leme 5218 (LC) Leme et al. 6372

Subgenus Cryptanthus, Cryptanthus group

Subgenus Hoplocryptanthus, Hoplocryptanthus group

-

Clade IV – subclade F Clade I – subclade B Clade I – subclade B Clade III Clade I – subclade B

Subgenus Hoplocryptanthus, Mesophyticae group Clade I – subclade B -

Subgenus Hoplocryptanthus, Hoplocryptanthus group

Clade IV – subclade F Clade I – subclade B Clade I – subclade A Clade I Clade II – subclade C

35

(HB) C. reisii Leme

Leme 5019 (RG)

85E

BA, Itapetinga

C. sanctaluciae Leme & L. Kollmann

Louzada 162 IBt (LC)

STL1

ES, Santa Teresa

Louzada 162 IBt (LC)

STL2

ES, Santa Teresa

36E

ES, Domingos Martins

86E

MG, Ouro Preto

C. scaposus E.Pereira C. schwackeanus Mez

C. sergipensis I.Ramirez

Leme 5213 (RG) Leme 6981 (RG)

-

Clade IV – subclade E Clade I – subclade B

-

Louzada et al. 441744 (SP)

SERB MG, Ouro Preto

Louzada et al. 441744 (SP)

SERP MG, Caeté

Clade I – subclade B Subgenus Hoplocryptanthus, Mesophyticae group

Clade I – subclade A Clade II – subclade C

Subgenus Hoplocryptanthus, Schwackeanae group

Clade II – subclade C Clade II – subclade C

Melo et al. 17279 (ASE)

G1

SE, Poço Redondo

Melo et al. 17279 (ASE)

G2

SE, Poço Redondo

Melo et al. 17279 (ASE)

G3

SE, Poço Redondo

Clade IV – subclade F

Melo et al. 17279

G4

SE, Poço

Clade IV – subclade F

Clade IV – subclade F

Subgenus Cryptanthus, Bahianae group

Clade IV – subclade F

36

(ASE) C. sinuosus L.B.Sm.

Leme 2868 (RG) Leme 5229 (RG) Leme 4184 (RG) Leme 5406 (RG) Leme 5084 (RG)

Redondo 68E

RJ, Arraial do Cabo

Clade IV – subclade E

30E

RJ, Búzios

Clade IV – subclade F

7E

RJ, São Pedro D’Aldeia

45E

RJ, São Pedro D’Aldeia

Clade IV – subclade F

63E

RJ, São Pedro D’Aldeia

Clade IV – subclade F

-

Clade IV – subclade F

C. teretifolius Leme

Leme et al. 3073 (HB)

51E

ES, Vitória

C. tiradentesensis Leme

Leme 5240 (RG)

61E

MG, Tiradentes

Leme 5825 (RG)

69E

MG, Tiradentes

Clade II – subclade C

Leme 5225 (RG)

13E

BA, Ubaira

Clade IV – subclade F

Leme et al. 7788 (RB)

59E

BA, Guaratinga

Leme 5228 (RG)

78E

BA, Ubaira

48E

ES, Nova Venecia

C. ubairensis I.Ramirez

C. venecianus Leme & L.Kollmann

Leme 7743 (RG)

-

Subgenus Cryptanthus, MG: Cryptanthus group

Clade IV – subclade E Clade II – subclade C

Clade IV – subclade F Clade V – subclade F

-

Clade IV – subclade E

37

C. viridipetalus Leme C. whitmanii Leme C. zonatus (Visiani) Beer C. aff. bivittatus

Leme 8016 (RG) Leme 5208 (RG) Leme 6518 (RG) Leme 3762 (RG) Leme 3080 (RG)

46E

ES, Boa Esperança

71E

-

Clade IV – subclade E

ES, Domingos Martins

Subgenus Hoplocryptanthus, Mesophyticae group

Clade I – subclade A

55E

PE, Jaqueira

Subgenus Cryptanthus, Zonatae group

Clade IV – subclade F

25E

ES, Domingos Martins

-

Clade IV – subclade F

79E

ES, Domingos Martins

-

Clade IV – subclade F

C. aff. bromelioides

Leme 4244 (RG)

40E

RJ, Rio de Janeiro

-

Clade IV – subclade F

C. aff. burle-marxii

Leme 3860 (RG)

1E

RN, Natal

-

Clade IV – subclade E

Leme 3878 (RG)

28E

RN, Natal

-

Clade IV – subclade E

Leme 6260 (RG)

56E

PE, Pernambuco

-

Clade IV – subclade E

Leme 6325 (RG)

93E

BA, Saude

-

Clade IV – subclade E

18E

MG, Barão de Cocais

-

Clade II – subclade D

24E

MG, Barão de Cocais

-

Clade II – subclade C

29E

ES, Santa Leopoldina

-

Clade I – subclade A

C. aff. diamantinensis C. aff. glazioui

Leme 6877 (RG) Leme 6872 (RG)

C. aff. leuzingerae

Leme 1144 (RG)

38

C. aff. praetextus

Leme 5490 (RG)

52E

ES, Espírito Santo

-

Clade IV – subclade F

C. aff. regius

Leme 6888 (RG)

11E

MG, Catas Altas

-

Clade II – subclade D

C. aff. scaposus

Leme 2725 (RG)

70E

ES, Santa Rita

-

Clade I

C. aff. schwackanus

Leme 6266 (RG)

43E

MG, Sabará

-

Clade II – subclade C

C. aff. sinuosus

Leme 8089 (RG)

50E

RJ, São Fidélis

-

Clade IV – subclade E

Cryptanthus sp. nov.

Leme 6588 (RG)

88E

BA, Rio do Meio

-

Clade IV – subclade E

37E

BA, Vitória da Conquista

-

Clade IV – subclade F

6E

ES, Serra

-

Clade IV – subclade E

83E

BA, Bandeira do Colônia

-

Clade IV – subclade E

Cryptanthus sp. Cryptanthus sp. Cryptanthus sp.

Leme 6479 (RG) Leme 6979 (RG) Leme 6477 (RG)

Orthophytum amoenum (Ule) L.B.Sm.

Louzada, 7106 IBt (LC)

R23

BA, Chapada Diamantina

-

Outgroup

Orthophytum hatschbachii Leme

Louzada, 104 IBt (LC)

R19

BA, Rio de Contas

-

Outgroup

Orthophytum heleniceae Leme

Louzada et al. 2544 (SP)

R38

BA, Bahia

-

Outgroup

Orthophytum ophiuroides Louzada & Wand.

Louzada, 88 IBt (LC)

R5

BA, Chapada Diamantina

-

Outgroup

39

Orthophytum ulei Louzada & Wand.

Louzada, 91 IBt (LC)

R43

BA, Chapada Diamantina

-

BA, Chapada Diamantina

Abbreviations: AL: Alagoas; BA: Bahia; ES: Espírito Santo; MG: Minas Gerais; PE: Pernambuco; RJ: Rio de Janeiro; RN: Rio Grande do Norte.

40

41

Highlights



The presented phylogenetic reconstruction for the genus Cryptanthus makes clear that subgenus Cryptanthus is a monophyletic group while subgenus Hoplocryptanthus forms a paraphyletic group.



The previous morphological groups described in the literature for the genus Cryptanthus were not confirmed as natural groups in our proposed phylogeny.



Regarding the reconstruction of homoecic and andromonoecic species, we found a single origin of the andromonoecy, with all species of the subgenus Cryptanthus being characterized by andromonoecy.



Our molecular findings also revealed an interesting pattern where the habitat preferences appear to be more conserved than morphological characters within Cryptanthus.



Cerrado and campos rupestres where inferred as ancestral biomes for the genus and we identified two shifts to the moister Atlantic forest biome, in the ancestor of the first diverging lineage within the genus, and in the ancestor of subgenus Cryptanthus.

42