Academia.eduAcademia.edu
ARTICLE IN PRESS Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) based on four chloroplast DNA regions Martina Rex a, Katharina Schulte b, Georg Zizka b, Jule Peters a, Roberto Vásquez c, Pierre L. Ibisch d,e, Kurt Weising a,* a Plant Molecular Systematics, Institute of Biology, Dept. of Sciences, University of Kassel, Heinrich-Plett-Str. 40, D-34132 Kassel, Germany Research Institute Senckenberg and Goethe-University, Frankfurt am Main, Germany c Sociedad Boliviana de Botánica, Casilla 3822, Santa Cruz, Bolivia d Faculty of Forest and Environment, University of Applied Sciences, Eberswalde, Germany e Fundación Amigos de la Naturaleza (F.A.N.), Dept. de Ciencias, Santa Cruz de la Sierra, Bolivia b a r t i c l e i n f o Article history: Received 11 September 2008 Revised 30 December 2008 Accepted 5 January 2009 Available online xxxx Keywords: Fosterella Bromeliaceae Pitcairnioideae s.str. Multilocus chloroplast DNA phylogeny Biogeography Character evolution a b s t r a c t The about 31 species of Fosterella L.B. Sm. (Bromeliaceae) are terrestrial herbs with a centre of diversity in the central South American Andes. To resolve infra- and intergeneric relationships among Fosterella and their putative allies, we conducted a phylogenetic analysis based on sequence data from four chloroplast DNA regions (matK gene, rps16 intron, atpB-rbcL and psbB-psbH intergenic spacers). Sequences were generated for 96 accessions corresponding to 60 species from 18 genera. Among these, 57 accessions represented 22 of the 31 recognized Fosterella species and one undescribed morphospecies. Maximum parsimony and Bayesian inference methods yielded well-resolved phylogenies. The monophyly of Fosterella was strongly supported, as was its sister relationship with a clade comprising Deuterocohnia, Dyckia and Encholirium. Six distinct evolutionary lineages were distinguished within Fosterella. Character mapping indicated that parallel evolution of identical character states is common in the genus. Relationships between species and lineages are discussed in the context of morphological, ecological and biogeographical data as well as the results of a previous amplified fragment length polymorphism (AFLP) study. Ó 2009 Elsevier Inc. All rights reserved. 1. Introduction The genus Fosterella L.B. Sm. (Bromeliaceae) comprises a group of meso- to xerophytic, usually stemless, terrestrial species with rosette leaves and inconspicuous, mostly whitish flowers. Almost all species are entomophilous. Fruits are mainly septicidal capsules that release minute, appendaged seeds. The genus is distributed across central South America, with a centre of diversity in arid and semi-humid habitats of the northeastern Andean slopes of Bolivia. Fosterella micrantha has a disjunct occurrence in Central America. Many Fosterella species are rare, endemic, and (or) restricted to certain habitats. In the most recent monograph of the genus, Smith and Downs (1974) recognized 13 Fosterella species. Numerous new taxa have been described since then, raising the species number to 31 (Rauh, 1979, 1987; Luther, 1981, 1997; Smith and Read, 1992; Kessler et al., 1999; Ibisch et al., 1997, 1999, 2002, 2008; Peters et al., 2008a,b). The exact taxonomic position of Fosterella is not yet satisfactorily settled. Traditionally, the genus has been placed in the subfam* Corresponding author. Fax: +49 561 8044200. E-mail address: weising@uni-kassel.de (K. Weising). ily Pitcairnioideae of Bromeliaceae, mainly because of its fruit characteristics (Smith and Downs, 1974). Based on a morphological cladistic analysis, Varadarajan and Gilmartin (1988a,b) divided Pitcairnioideae into three tribes, i.e., Pitcairnieae, Brocchinieae and Puyeae. In their system, Fosterella was assigned to Pitcairnieae, with Connellia and Navia being its closest relatives. The classification of Varadarajan and Gilmartin (1988b) was also adopted by Smith and Till (1998), but received no support by subsequent molecular data. Instead, several studies based on chloroplast DNA sequence variation provided unequivocal evidence that Pitcairnioideae, in their traditional circumscription, are paraphyletic (Terry et al., 1997; Horres et al., 2000, 2007; Crayn et al., 2004; Givnish et al., 2004, 2007; Schulte et al., 2005). Based on a reasonably well resolved ndhF tree, Givnish et al. (2007) proposed the division of the former Pitcairnioideae into six new subfamilies, i.e., Brocchinioideae, Lindmanioideae, Hechtioideae, Navioideae, Puyoideae, and Pitcairnioideae s.str. In their system, which we also follow here, Fosterella remains within Pitcairnioideae s.str. Whereas the monophyly of Fosterella has never been questioned, relatively little is known about its infrageneric phylogeny. In previous studies, we used random amplified polymorphic DNA (RAPD; Ibisch et al., 2002) and amplified fragment length polymor- 1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.01.001 Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 2 M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx phisms (AFLP; Rex et al., 2007) to unravel the evolutionary history of the genus. Morphologically defined species boundaries were mostly confirmed by the molecular data, and several groups of closely related species were identified that also share similar leaf anatomy patterns (Rex et al., 2007). However, bootstrap support of individual groups was generally low, phylogenetic relationships among groups remained mostly ambiguous, and basal nodes were poorly resolved. Moreover, the high extent of RAPD and AFLP variation prevented us from extending our investigations to the level of genera. Up to now, the vast majority of molecular phylogenetic studies dealing with Bromeliaceae relied on chloroplast genes and introns (e.g., ndhF gene: Terry et al., 1997; Givnish et al., 2007; trnL intron: Horres et al., 2000; matK gene plus rps16 intron: Crayn et al., 2004). Our initial attempts to infer the infrageneric phylogeny of Fosterella from sequence analysis of single chloroplast DNA loci such as the trnL intron were hampered by low levels of sequence divergence (Rex, unpublished data). This is most likely a consequence of the young age of the crown group of Bromeliaceae, which is assumed to be less than 20 Mya (Givnish et al., 2004, 2007). Higher resolution can often be obtained by combining DNA sequences from several chloroplast DNA loci, as has been demonstrated for the subfamilies Tillandsioideae (Barfuss et al., 2005) and Bromelioideae (Schulte and Zizka, 2008). Here we present a multilocus chloroplast DNA phylogeny of Fosterella and related genera based on the matK gene, the rps16 intron, and the atpB-rbcL and psbB-psbH intergenic spacers. The objectives of our study were (1) to determine the position and sister group relationships of Fosterella within Pitcairnioideae s.str., (2) to reconstruct an infrageneric chloroplast DNA phylogeny, (3) to compare the topology of the cpDNA trees with that of the (presumably nuclear) AFLP tree of Rex et al. (2007), and (4) to investigate the evolution and taxonomic significance of morphological characters in Fosterella. 2. Materials and methods 2.1. Plant material Ninety-six accessions were included in the analysis, corresponding to 60 species from 18 genera of Bromeliaceae (Table 1). The subfamily Brocchinioideae sensu Givnish et al. (2007) has consistently been shown to be the sister group of the remainder of Bromeliaceae (e.g., Terry et al., 1997; Horres et al., 2000; Givnish et al., 2007). Consequentially, two members of this subfamily, Brocchinia acuminata and B. uaipanensis, were used as outgroup. The ingroup consisted of five species of Tillandsioideae, two Hechtioideae, four Puyoideae, 10 Bromelioideae, and 37 Pitcairnioideae s.str. As our study was focussed on Fosterella and its position within Pitcairnioideae s.str., members from the other new subfamilies proposed by Givnish et al. (2007) were not included. Fifty-seven accessions represented 22 of the 31 recognized Fosterella species. We also analyzed one additional Fosterella accession that does not correspond morphologically to any described species (F. spec. in Table 1; previously referred to as F. spec. 4 by Rex et al., 2007). Plant material from the remaining Fosterella species was not available for this study. Voucher specimens and duplicates have been deposited in various herbaria and living collections. 2.2. DNA isolation and PCR amplification Total genomic DNA was extracted from 100 to 150 mg of fresh, frozen or lyophilized leaves of individual plants, using either a DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), or a cetyl trimethylammonium bromide (CTAB) procedure (Weising et al., 2005). Final DNA concentrations were determined electrophoretically versus known amounts of k-DNA as standards. Primer pairs used for PCR amplification of the atpB-rbcL spacer were designed by Manen et al. (1994) and of the psbB-psbH spacer by Xu et al. (2000). For the amplification of the matK gene, the primers matK5 F (Crayn et al., 2000) and trnK2 R (Johnson and Soltis, 1995) were used. Sequences of forward and reverse primers for amplifying the rps16 intron were adopted from Wallander and Albert (2000) and Oxelman et al. (1997), respectively. Except for the matK region, primers carried an M13-tail at their 50 ends to facilitate subsequent sequencing (see below). Primers were either purchased from Metabion GmbH (Martinsried, Germany) or from MWG Biotech (Ebersberg, Germany). All PCRs were performed in 25 lL volumes using a Biometra T-Gradient cycler (Biometra GmbH, Göttingen, Germany). Each assay contained 2–5 ng of template DNA, 2.5 mM MgCl2, 1 lM each of forward and reverse primer, 0.2 mM of each dNTP, 20 mM Tris–HCl pH 8.0, 50 mM KCl, 5 lg BSA and 0.02 U of Taq DNA polymerase (Invitrogen GmbH, Karlsruhe, Germany). After an initial denaturation at 94 °C for 5 min, PCR was performed for 30 cycles, each consisting of 94 °C for 1 min, 2 min at annealing temperature, and 72 °C for 2 min. Annealing temperatures were 50 °C for the matK gene, 58 °C for the atpB-rbcL spacer, 56 °C for the psbB-psbH spacer, and 60 °C for the rps16 intron. Final extension was at 72 °C for 10 min. To check for the presence of distinct, single bands, aliquots of PCR products were electrophoresed on agarose gels and stained with ethidium bromide. The remainder of the reaction was subjected to sequencing. 2.3. DNA sequencing Double-stranded PCR products of the atpB-rbcL and psbB-psbH spacer and the rps16 intron were cycle-sequenced by the dideoxynucleotide chain termination method without further purification. Both strands were sequenced bidirectionally in the same reaction, using a ThermoSequenase kit (Amersham Biosciences Europe, Freiburg, Germany) and 3 pmoles of IRDye700- and 5 pmoles of IRDye800-labeled M13 universal primers for the forward and reverse reaction, respectively (M13 universal: 50 -TGT AAA ACG ACG GCC AGT-30 , M13 reverse 50 -CAG GAA ACA GCT ATG ACC-30 ). Sequencing primers were purchased from MWG Biotech (Martinsried, Germany). Sequencing assays followed the protocol of the kit manufacturer. After initial denaturation at 95 °C for 5 min, cycle sequencing was performed for 25 cycles, each consisting of 95 °C for 15 s, 57 °C for 30 s, and 72 °C for 45 s. Final extension was at 72 °C for 7 min. Sequencing products were mixed with one volume of formamide buffer, denatured at 85 °C for 5 min, and separated on 6% denaturing polyacrylamide gels (Sequagel XR, Biozym Scientific, Hessisch-Oldendorf, Germany) in an automated Li-Cor L4200L sequencer (Li-Cor Bioscience GmbH, Bad Homburg, Germany). Double-stranded PCR products of the matK gene were purified with the NucleoSpin extract kit (Macherey & Nagel, Düren, Germany) according to the manufacturer’s protocol. Cycle sequencing was performed for 25 cycles, each consisting of 5 s at 96 °C, 1 min at 50 °C, and 4 min at 60 °C, using the BigDye Terminator Premix V 3.0 (Applied Biosystems, Darmstadt, Germany). In addition to the amplification primers, the following internal primers were employed for sequencing: TOmatK480 F (Hilu et al., 2003), BROmatK860 F and BROM1 R (Schulte et al., 2005), and an alternative primer to TOmatK480 F that has specifically been designed for this study (BROM650F: 50 -GCG ATT CTT TCT CCA CGA AT-30 ). The matK sequencing products were purified by ethanol precipitation and analyzed on an ABI 377 automated sequencer (Applied Biosystems, Darmstadt, Germany), according to the manufacturer’s protocols. For a few difficult templates, PCR products were sent to a commer- Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 SF Species Collector (Herbarium)/living plant Collection location DNA-No. State, Dpto., Prov. PI Fosterella albicans (Griseb.) L.B. Sm. Fosterella batistana Ibisch, Leme & J. Peters Fosterella caulescens Rauh Fosterella christophii Ibisch, R. Vásquez & J. Peters Fosterella cotacajensis M. Kessler, Ibisch & E. Gross floridensis Ibisch, R. Váquez & E. Gross gracilis (Rusby) L.B. Sm. graminea (L.B. Sm.) L.B. Sm. heterophylla Rauh kroemeri Ibisch, R. Vásquez & J. Peters micrantha (Lindl.) L.B. Sm. Fosterella penduliflora (C.H. Wright) L.B. Sm. Fosterella rexiae Ibisch, R. Vásquez & E. Gross Fosterella robertreadii Ibisch & J. Peters Fosterella rusbyi (Mez) L.B. Sm. Fosterella spectabilis H. Luther Fosterella vasquezii E. Gross & Ibisch Fosterella villosula (Harms) L.B. Sm. Fosterella weberbaueri (Mez) L.B. Sm. atpB-rbcL psbB-psbH rps16-Intron BO, La Paz, Inquisivi BO, Cochabamba, Chapare BO, Santa Cruz, Florida BR, Pará, Itaituba BO, La Paz, Caranavi BO, La Paz, Caranavi BO, La Paz, Caranavi BO, Santa Cruz, A. Ibañez 62a 94c 64a 129a 142b 80a 3a 25f EU681843 EU681838 EU681846 EU681859 EU681839 EU681845 EU681840 EU681863 EF639772 EF639783 EF639774 EF639736 EF639748 EF639779 EF639763 EF639756 EF643061 EF643072 EF643063 EF643025 EF643037 EF643068 EF643052 EF643045 EF643162 EF643173 EF643164 EF643126 EF643138 EF643169 EF643153 EF643146 BO, Cochabamba, Ayopaya BO, La Paz, Inquisivi BO, Santa Cruz, Florida BO, Beni, Ballivian BO, La Paz, Larecaja BO, La Paz, Caranavi BO, La Paz, Caranavi GT, Suchitepequez — BO, Tarija, Gran Chaco BO, Santa Cruz, Cordillera BO, La Paz, Inquisivi BO, Tarija, ÓConnor BO, Santa Cruz, Cordillera BO, Santa Cruz, Florida BO, Santa Cruz, Florida BO, Santa Cruz, Ñ. de Chávez BO, Santa Cruz, Guarayos BO, Cochabamba BO, Santa Cruz, Guarayos — BO, La Paz, Caranavi BO, La Paz, Caranavi — — PE, Cuzco, Vilcanota PE, Cuzco, Vilcanota — BO, La Paz, Sud Yungas BO, La Paz, Nor Yungas BO, Santa Cruz, Florida — BO, Santa Cruz, Velasco BO, Santa Cruz, Velasco BO, Cochabamba, Chapare BO, Cochabamba, Chapare BO, Cochabamba, Chapare BO, La Paz, Caranavi BO, La Paz, Larecaja 76d 13a 67d 117a 71c 26a 28b 133a 132a 45c 22c 55a 46d 50b 18c 35a 34a 118a 136b 120a 93a 9b 10a 86a 143a 135b 137b 141b 60a 107b 87a 144a 63b 23b 104a 138b 48d 121a 95c EU681876 EU681877 EU681878 EU681923 EU681917 EU681849 EU681918 EU681860 EU681862 EU681868 EU681866 EU681874 EU681867 EU681871 EU681873 EU681870 EU681919 EU681864 EU681865 EU681869 EU681914 EU681844 EU681841 EU681912 EU681848 EU681847 EU681922 EU681853 EU681854 EU681842 EU681913 EU681852 EU681856 EU681855 EU681861 EU681857 EU681858 EU681911 EU681910 EF639778 EF639745 EF639775 EF639732 EF639776 EF639757 EF639758 EF639739 EF639738 EF639764 EF639754 EF639768 EF639765 EF639767 EF639752 EF639760 EF639759 EF639733 EF639741 EF639734 EF639782 EF639785 EF639731 EF639779 EF639780 EF639749 EF639742 EF639747 EF639771 EF639730 EF639781 EF639750 EF639773 EF639755 EF639729 EF639743 EF639766 EF639735 EF639784 EF643067 EF643034 EF643064 EF643021 EF643065 EF643046 EF643047 EF643028 EF643027 EF643053 EF643043 EF643057 EF643054 EF643056 EF643041 EF643049 EF643048 EF643022 EF643030 EF643023 EF643071 EF643074 EF643020 EF643068 EF643069 EF643038 EF643031 EF643036 EF643060 EF643019 EF643070 EF643039 EF643062 EF643044 EF643018 EF643032 EF643055 EF643024 EF643073 (continued EF643168 EF643135 EF643165 EF643122 EF643166 EF643147 EF643148 EF643129 EF643128 EF643154 EF643144 EF643158 EF643155 EF643157 EF643142 EF643150 EF643149 EF643123 EF643131 EF643124 EF643172 EF643175 EF643121 EF643169 EF643170 EF643139 EF643132 EF643137 EF643161 EF643120 EF643171 EF643140 EF643163 EF643145 EF643119 EF643133 EF643156 EF643125 EF643174 on next page) ARTICLE IN PRESS Fosterella Fosterella Fosterella Fosterella Fosterella Fosterella Vásquez 3617/FAN RV 3617 Vásquez 4626b/FAN RV 4626b Ibisch 98.0204 (LPB)/FAN PI 98.0204 Fernandes da Silva s.n. (HB, SEL)/LEME 5078 Rauh 40579 a (HEID)/BGHD 103532 Rauh 40579 a (SEL)/MSBG 1989-0220 Rauh 40579 a (U, WU)/FRP 99-18434-3 Ibisch 98.0173 (FR, LPB, SEL, USZ, WU)/FAN PI 98.0173 Kessler 9620b (LPB, SEL)/FAN MK 9620b Vásquez 3612 (SEL)/FAN RV 3612 Ibisch 02.0001 (SEL)/FAN PI 02.0001 Rex & Schulte 281002-6/FAN RS 281002-6 Müller 216 (SEL)/FAN RM 216 Vásquez 3661 (LPB, SEL)/FAN RV 3661 Krömer 1398b (LPB)/FAN TK 1398b Welz 3124 (B, FR, HBG, U)/BGHD 103726 s.n. (B)/BGB 079-02-92-34 Vásquez 4003 (SEL)/FAN RV 4003 Vásquez 3817 (LPB, SEL, USZ)/FAN RV 3817 Vásquez 3624 (FR)/FAN RV 3624 Vásquez 4051b (FR)/FAN RV 4051b Ibisch 00.0036 (FR)/FAN PI 00.0036 Ibisch 98.0098 (FR, LPB, SEL, USZ)/FAN PI 98.0098 Vásquez 3406 (LPB, SEL)/BGHD 125586 Vásquez 3762 (LPB)/FAN RV 3762 Rex & Schulte 301002-3 (FR)/FAN RS 301002-3 Friesen 18605 (OSBU)/BGOS 94-05-0016-20 Vásquez 4685 (VASQ)/FAN RV 4685 Müller 150/FAN RM 150 Vásquez 3673 (LPB)/FAN RV 3673 Vásquez 3666 (USZ)/FAN RV 3666 s.n. (SEL)/MSBG 1978-0905 s.n. (B)/BGB 115-19-83-80 Friesen 18606 (OSBU)/BGOS 94-17-0050-80 Rauh 20866 (WU, HEID)/BGHD 103751 Friesen 18604 (OSBU)/BGOS 94-17-0049-80 Nowicki 2061/FAN CN 2061 Rex & Schulte 251002-3 (SEL)/FAN RS 251002-3 Cathcart B-17 (SEL, HB)/MSBG 1995-0415 s.n. (B)/BGB 290-08-00-84 Ibisch 98.0116 (FR)/FAN PI 98.0116 Ibisch 98.0117/FAN PI 98.0117 Vásquez 4623 (FR)/FAN RV 4623 Krömer 7286 (FR, LPB)/BGHD 105332 Vásquez 3570 (FR, LPB, SEL, USZ)/FAN RV 3570 Vásquez 4656 (FR)/FAN RV 4656 Müller 217 (SEL)/FAN RM 217 GenBank Accession Nos. matK-Gene M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 Table 1 Plant material, collectors, collection sites, and GenBank accession numbers. Laboratory codes for DNA samples are included to facilitate the assignment of individual accessions to positions in the trees. Species are arranged according to their subfamily (SF) assignments sensu Givnish et al. (2007): PI, Pitcairnioideae s.str.; BC, Brocchinioideae; BM, Bromelioideae; HE, Hechtioideae; PU, Puyoideae; TI, Tillandsioideae. Countries of origin: AR, Argentina; BO, Bolivia; BR, Brazil; CL, Chile; CR, Costa Rica; GT, Guatemala; MX, Mexico; PE, Peru; VE, Venezuela. 3 Species Collector (Herbarium)/living plant Collection location Fosterella weddelliana (Brongn.) L.B. Sm. Nowicki 2034 (FR)/FAN CN 2034 Vásquez 3636a2 (FR, LPB, SEL)/FAN RV 3636a2 Vásquez 3620/FAN RV 3620 Nowicki 2076b (LPB, SEL, USZ)/FAN 2076 Vásquez 3627 (FR)/FAN 3627 Mijagawa s.n. (HEID)/BGHD 104866 Ibisch 03.0016 (LPB)/FAN PI 03.0016 Vásquez 4510 (LPB)/FAN RV 4510 S. Reichle SR1 (LPB)/FAN SR1 Vásquez 3729a (LPB)/FAN RV 3729a Vásquez 3729 (LPB)/FAN RV 3729 BO, BO, BO, BO, BO, BO, BO, BO, BO, BO, BO, DNA-No. GenBank Accession Nos. matK-Gene atpB-rbcL psbB-psbH rps16-Intron 37a 12a 56a 58c 36a 145a 139b 140b 72a 128a 19g EU681875 EU681915 EU681882 EU681880 EU681881 EU681879 EU681851 EU681916 EU681850 EF639762 EF639737 EF639769 EF639770 EF639761 EF639751 EF639744 EF639746 EF639777 EF639753 EF643051 EF643026 EF643058 EF643059 EF643050 EF643040 EF643033 EF643035 EF643066 EF643042 EF643152 EF643127 EF643159 EF643160 EF643151 EF643141 EF643134 EF643136 EF643167 EF643143 EU681872 State, Dpto., Prov. Fosterella windischii L.B. Sm. & Read Fosterella yuvinkae Ibisch, R. Vásquez, E. Gross & Reichle Fosterella spec. La Paz, Inquisivi La Paz, Sud Yungas La Paz, Inquisivi La Paz, Sud Yungas La Paz, Inquisivi La Paz Santa Cruz, Velasco Santa Cruz, Chiquitos Santa Cruz, Chiquitos Chuquisaca, Luis Calvo Chuquisaca, Luis Calvo Brocchinia uaipanensis (Maguire) Givnish Brocchinia acuminata L.B. Sm. Horres 011 (FR)/FRP 92-9510-2 Horres 001 (FRP)/FRP 95-15043-3 — VE, Bolivar F7a F19a EU681909 EU681908 EF639829 EF639828 EF643118 EF643117 EF643219 EF643218 BM Aechmea gamosepala Wittm. Aechmea kertesziae Reitz Aechmea orlandiana L.B. Sm. Bromelia pinguin L. Bromelia serra Griseb. Deinacanthon urbanianum Mez BR BR, Santa Catarina BR PE, San Martin — BR — BR, Santa Catarina — CL CL F17a F9a, H270 F18a F38a F12a, H029 F20a H018 F49a F29a, H080 F22a F36a EU681835 AY950039 EU681836 EU681899 AY950019 EF639806 EF639805 EF639804 EF639810 EF639809 EF639813 EF643097 EF643096 EF643093 EF643099 EF643098 EF643102 EF643198 EF643197 EF643194 EF643200 EF643199 EF643203 Edmundoa lindenii (Regel) Leme var. rosaea (E. Morren) Leme Neoregelia laevis (Mez) L.B. Sm. Ochagavia litoralis (Phil.) Zizka, Trumpler & Zoellner Ochagavia carnea (Beer) L.B. Sm. & Looser Schulte 130105-6 (FR)/FRP 89-16960-3 Schulte 290104-3 (FR)/FRP 98-16935-3 Schulte 190203-3 (FR)/FRP 90-1666-2 Rauh 53676 (HEID)/BGHD 103787 Horres 029 (FR)/FRP 98-17751-0 Schulte 110403-3 (FR)/FRP 98-17786-2 Horres 018 (FRP)/FRP 98-17786-0 Buckup s.n./BGHD 107435 Schulte 170305-3 (FR)/FRP 98-16962-3 Horres 015a (FR)/FRP 98-16853-2 Horres 115 (FR)/FRP 94-14614-3 AY950017 EU681903 AY950008 EU681904 EU681905 EF639807 EF639808 EF639811 EF639812 EF643094 EF643095 EF643100 EF643101 EF643195 EF643196 EF643201 EF643202 HE Hechtia argentea Baker Hechtia stenopetala Klotzsch Schulte 280408-2/FRP 88-19332-3 Schulte 280408-4/FRP 0-19454-2 — — F11a F21a EU681833 EU681834 EF639826 EF639827 EF643115 EF643116 EF643216 EF643217 PI Deuterocohnia brevifolia (Rauh) M.A. Spencer & L.B. Sm. Deuterocohnia brevispicata Rauh & L. Hrom. Deuterocohnia glandulosa E. Gross Deuterocohnia lotteae (Rauh) M.A. Spencer & L.B. Sm. Deuterocohnia scapigera (Rauh & L. Hrom.) M.A. Spencer & L.B. Sm. Dyckia encholirioides (Gaudich.) Mez Dyckia estevesii Rauh Balfanz 075/BGHD 107170 Hromadnik 5213 (HEID)/BGHD 102379 Hromadnik 5167 (HEID)/BGHD 103854 Hromadnik 5131 (HEID)/BGHD 103817 Hromadnik 5275 (HEID)/BGHD 130020 Schulte 280408-3/FRP 94-19369-3 HEID 602151-602159 (HEID)/BGHD 105188 Esteves-Pereira s.n. (HEID)/BGHD 105012 Rauh 67622 (HEID)/BGHD 105013 Schindhelm s.n./BGHD 108213 FRP s.n./FRP s.n. Schulte 280408-1/FRP 89-16095-2 Graf 6468 (HEID)/BGHD 102579 Schmidt s.n. (HEID)/BGHD 104044 Rauh 53676 (HEID)/BGHD 103787 Rauh 52598/FRP 1-19497-3 BO, Tarija BO, Chuquisaca BO, Tarija BO, Tarija BO, Potosi BR BR BR, Goias BR, Minas Gerais BR, Minas Gerais — CR VE, Tachira MX, Jalisco PE, San Martin MX, Oaxaca F42a F51a F43a F44a F45a F13a F50a F40a F39a F46a F15a F34a F16a F47a F48a F14a EU681895 EU681889 EU681893 EU681894 EU681888 EU681883 EU681886 EU681885 EU681884 EU681887 EU681896 EU681837 EU681891 EU681897 EU681892 EU681890 EF639794 EF639791 EF639792 EF639793 EF639790 EF639786 EF639787 EF639789 EF639788 EF639795 EF639802 EF639801 EF639799 EF639803 EF639797 EF639796 EF643084 EF643080 EF643082 EF643083 EF643079 EF643075 EF643076 EF643078 EF643077 EF643081 EF643091 EF643090 EF643088 EF643092 EF643086 EF643085 EF643185 EF643181 EF643183 EF643184 EF643180 EF643176 EF643177 EF643179 EF643178 EF643182 EF643192 EF643191 EF643189 EF643193 EF643187 EF643186 Schulte 280408-6/FRP 91-18506-3 Krömer 6581/BGHD 105240 Gouda s.n./BGHD 130080 Leuenberger 4490b/BGHD 103912 CL BO BO AR, Cordoba, Colón F23a F53a F41a F52a EU681920 EU681901 EU681900 EU681902 EF639817 EF639815 EF639814 EF639816 EF643106 EF643104 EF643103 EF643105 EF643207 EF643205 EF643204 EF643206 Horres 05.03.2001/FRP 90-2846-2 Rauh 66031 (HEID)/BGHD 102386 /FRP 92-10645-0 B194/96 (WU)/ Zizka 0293 (FRP)/FRP 90-9689-4-2 Zizka 1582 (FRP)/FRP 90-1326-2 — PE, Pasco — — PE, Piura — F10a F8a F33a MB-14 F25a F26a EU681898 EU681907 EF639824 EF639822 EF639823 EF643113 EF643111 EF643112 EF643214 EF643212 EF643213 EF639819 EF639820 EF643108 EF643109 EF643209 EF643210 Dyckia goehringii E. Gross & Rauh Encholirium horridum L.B. Sm. Pitcairnia albiflora Spreng. Pitcairnia atrorubens (Beer) Baker Pitcairnia grafii Rauh Pitcairnia loki-schmidtiae Rauh & Barthlott Pitcairnia rubro-nigriflora Rauh Pitcairnia breedlovei L.B. Sm. PU Puya Puya Puya Puya coerulea Miers var. violacea (Brongn.) L.B. Sm. & Looser herzogii Wittm. mirabilis (Mez) L.B. Sm. spathacea Mez TI Catopsis floribunda (Brongn.) L.B. Sm. Guzmania glaucophylla Rauh Guzmania musaica (Linden & André) Mez Racinaea pugiformis (L.B. Sm.) M.A. Spencer & L.B. Sm. Tillandsia fraseri Baker AY614058 EU681921 EU681906 ARTICLE IN PRESS BC M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 SF 4 Table 1 (continued) ARTICLE IN PRESS M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx 5 cial sequencing facility (Science Research and Development, Oberursel, Germany). indels, using the software MacClade, V4.06 (Maddison and Maddison, 2003). 2.4. Data analysis 3. Results Forward and reverse sequences were compared using the DNA sequencing software e-seq, V2.0 (Li-Cor Bioscience GmbH, Bad Homburg, Germany). Consensus sequences were aligned and edited using the Align-IRTM Assembly and Alignment software, V1.2 or V2.0 (Li-Cor), with default settings. Automated alignments were adjusted manually where necessary. Ambiguously aligned nucleotide positions were excluded from the analysis. Indels were automatically coded, applying the simple indel coding method of Simmons and Ochoterena (2000) and the GapCoder computer program (Young and Healy, 2003). The resulting presence/absence matrix of indels was appended to the alignment. All sequences obtained in the present study have been deposited in Genbank (accession numbers listed in Table 1). Phylogenetic analyses were performed separately for each locus, as well as for all loci combined, with and without the respective indel matrices. Maximum parsimony ratchet analyses (Nixon, 1999) were conducted in PAUP* 4.0b10 (Swofford, 2002) with command files generated with PRAP (Müller, 2003). Gaps were treated as missing data. Character state optimization was conducted under the ancillary of accelerated transformation (ACCTRAN). Trees were rooted with Brocchinia acuminata and B. uaipanensis. For each of the 1000 random replicates, 200 ratchet iterations were performed. Each iteration comprised ten rounds of TBR swapping, saving one shortest tree. Multiple parsimonious trees were combined to form a strict consensus tree. The extent of homoplasy was estimated using the consistency (CI) and retention indices (RI). Statistical support values for nodes and clades were estimated by bootstrap analyses with 1000 replications (Felsenstein, 1985), each with one random addition replicate, followed by TBR swapping, the MULTREES option activated, saving no more than 500 trees per pseudo replicate. For Bayesian analysis, the best-fit model of evolution was inferred using ModelTest, V3.6 (Posada and Crandall, 1998) based on the Akaike Information Criterion AIC (Akaike, 1974). Bayesian analyses were performed with MrBayes, V2.01 (Huelsenbeck and Ronquist, 2001), using the GTR + G + I evolutionary model, as proposed by AIC. Indels were treated as missing data. One cold chain and three incrementally heated Markov chain Monte Carlo (mcmc) chains were run for 100,000 or 4 million cycles, with trees sampled every 100th or 1000th generation, using random trees as starting point and a temperature parameter value of 0.1. For each data set, mcmc runs were repeated three times as a safeguard against spurious results. The first 60,000 trees were discarded as burn-in, and the remaining trees were used to construct a 50% majority rule consensus tree. 3.1. Alignments and sequence characteristics 2.5. Evaluation of major morphological transitions In the course of our ongoing revision of Fosterella, we are studying specimens from the herbaria B, BA, BM, CGE, CUZ, F, FR, G, GOET, GZU, HB, HBG, HEID, HSB, HUH, K, LI, LIL, LPB, M, MA, MCNS, MO, NY, P, RB, SEL, U, US, USM, USZ, as well as living material from the Palmengarten Frankfurt/M., the Botanical Gardens of Berlin, Göttingen, Hamburg, Heidelberg, Leipzig (all Germany), Vienna (Austria), the Fundación Amigos de la Naturaleza (FAN), Santa Cruz (Bolivia) and the private collection of Elton Leme, Teresópolis, Rio de Janeiro, Brazil (Peters et al., unpublished data). To reconstruct character evolution within the genus, we encoded the states of ten selected characters that have been evaluated during our morphological work (Table 2). Character states were then mapped onto the strict consensus tree resulting from parsimony analysis of the combined data matrix of four chloroplast loci, excluding the coded Sequences of all four chloroplast loci were generated for all 96 investigated accessions of Bromeliaceae (Table 1). The final alignments comprised 1778 aligned positions of the matK gene, 928 of the rps16 intron, 713 of the atpB-rbcL spacer and 729 of the psbB-psbH spacer. Altogether, the concatenated sequence matrix contained 4148 characters. The atpB-rbcL spacer harbored three highly variable microsatellites (polyA, polyT and polyC) that were excluded from all analyses. A total of 128 indels were coded as presence/absence characters, and added to the sequence matrix in some analyses. Fourteen indels were detected in the matK gene, 54 in the rps16 intron, 33 in the atpB-rbcL spacer and 27 in the psbB-psbH spacer. The concatenated data matrix including the indels contained 4276 characters. Of these, 3566 were constant, 244 represented autapomorphies and 466 were parsimony-informative. The highest percentage of synapomorphies was observed in the atpB-rbcL spacer (12.9%; 96 of 746 characters), followed by the rps16 intron (12.1%; 119 of 982), the matK gene (10.3%; 185 of 1792), and the psbB-psbH spacer (8.7%; 66 of 756). 3.2. Phylogenetic analyses Tree topologies resulting from separate analysis of the four individual loci slightly differed from each other, but no well-supported conflicting nodes were observed (data not shown). All data sets were therefore combined, and subjected to maximum parsimony ratchet analysis (MPR; with the appended indel matrix either included or excluded) and Bayesian inference. The combined trees generally showed higher resolution than any of the single-locus trees. A total of 3529 shortest trees of 1142 steps were found in the MPR analysis of the combined data set including the indels, whereas 1459 shortest trees of 875 steps were found when the indels were excluded. The strict consensus tree obtained without coded indels is shown in Fig. 1. Arrows indicate branches that collapse in the tree with the indels included. The 50% majority rule consensus tree resulting from Bayesian analysis of the same data set is shown in Fig. 2. All trees resulting from the combined data set revealed highly similar topologies, with only a few incongruencies at lower taxonomic levels. A grade is formed by the ingroup, with successive branching of several clades that receive moderate to high bootstrap support (BS excluding/including the indels) in the MPR trees and/ or posterior probability (PP) in the Bayesian analysis (Figs. 1 and 2). The first branch comprises all investigated Tillandsioideae and Hechtioideae. The two subfamilies are each monophyletic and sister to each other, but this sister relationship receives only little support (BS 72/62; PP 76). The second branch (BS 89/90, PP 100) splits into a dichotomy with Puyoideae/Bromelioideae forming one clade (BS 94/93; PP 100), and monophyletic Pitcairnioideae s.str. forming the other clade (BS 91/77, PP 100). A grade is also formed within the latter. Two species of Pitcairnia (P. loki-schmidtiae and P. albiflora) are branching off first, followed by a clade comprising four more species of the same genus. Pitcairnia is therefore paraphyletic in both trees, but the paraphyly is only weakly supported (BS 69/54; PP 74). The three genera Dyckia, Encholirium and Deuterocohnia together form a clade which is sister to Fosterella. Strong support was obtained for the monophyly each of Fosterella (BS 100/100, PP 100) and the Dyckia/Encholirium/Deuterocohnia clade (BS 99/99, PP 100), as well as for their sister group relationship (BS 94/85, PP 100). Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 6 M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Table 2 Data matrix used to reconstruct character evolution within Fosterella. (1) Caulescence: 0, acaulecent; 1, subcaulescent; 2, caulescent. (2) Leaves: margin: 0, entire; 1, serrate. (3) Petals: inflection during/after anthesis: 0, straight/straight; 1, slightly recurved/straight; 2, strongly recurved/straight; 3, recoiled like watchsprings/recoiled like watchsprings. (4) Peduncle bracts: margin: 0, entire; 1, serrate. (5) Leaf arrangement: 0, flat rosette; 1, erect rosette; 2, leaves spirally arranged along elongated stem; (6) Inflorescence: density of vestiture: 0, glabrous/glabrescent; 1, sparsely lanate; 2, densely lanate. (7) Leaves, abaxial: trichome types: 0, stellate; 1, peltate with dentate margin; 2, peltate with entire margin; 3, peltate with elongated cells forming a loose fringe; (8) Leaves, abaxial: density of vestiture: 0, glabrescent; 1, sparse; 2, dense, not completely covering the leaf surface; 3, very dense, completely covering the leaf surface. (9) Inflorescence: ramification: 0, raceme/compound raceme; 1, panicle. (10) Petals: color: 0, white; 1, yellow; 2, red. Species Caulescence Leaves: margin Petals: inflection during/after anthesis Peduncle bracts: margin Leaf arrangement Inflorescence: density of vestiture Leaves, abaxial: trichome types Leaves, abaxial: density of vestiture Inflorescence: ramification Petals: color F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. F. 1 0 2 0 2 0 0 0 2 0 0 0 2 0 0 0 0 0 1 2 0 0 1 0 1 0 1 0 0 1 0 0 0 0 1 0 1 0 0 0 0 1 0 0 3 1 3 1 3 0 1 3 3 3 1 1 3 3 3 0 3 1 2 3 3 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 1 0 0 1 0 2 0 2 1 0 1 2 1 0 0 2 0 0 0 0 0 0 2 0 0 2 1 2 2 0 2 0 0 0 0 1 0 1 1 0 0 0 2 0 0 0 0 3 1 2 1 3 2 0 2 2 3 1 1 2 2 3 1 3 1 1 3 3 0 3 0 2 1 3 2 1 2 2 3 1 1 2 2 3 0 3 1 0 3 3 2 1 0 1 1 1 1 1 1 1 1 1 0 1 1 1 0 1 1 0 1 1 1 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 albicans batistana caulescens christophii cotacajensis floridensis gracilis graminea heterophylla kroemeri micrantha penduliflora rexiae robertreadii rusbyi spectabilis vasquezii villosula weberbaueri weddelliana windischii yuvinkae Within Fosterella, six monophyletic lineages are resolved which we informally refer to as rusbyi, albicans, weberbaueri, micrantha, weddelliana, and penduliflora group (Figs. 1 and 2). Each of these receives good support, with posterior probabilities of 100% throughout, and bootstrap values (BS) ranging from 85% to 100%. The rusbyi group consists of F. rusbyi, F. vasquezii, F. windischii, F. yuvinkae, F. floridensis and F. spectabilis. As most of the species of this group are rare endemics and only represented by one or two accessions in our sampling, monophyly of species and boundaries between species are difficult to assess beyond the clearly existing morphological differences. The two accessions of F. spectabilis form a subclade (BS 95/95, PP 100), which is sister to the remainder of the group. This relationship is moderately supported in the bootstrap analysis (BS 66/74), but receives a high a posteriori probability in the Bayesian analysis (PP 100). The albicans group comprises all accessions of F. albicans, F. rexiae, F. caulescens, F. kroemeri, F. graminea, F. heterophylla and F. robertreadii. Species boundaries within this group remain largely unresolved, mainly due to limited sequence variation. The four accessions of F. robertreadii form a well-supported subclade (BS 100/100, PP 100). One of the three accessions of F. albicans included in the sampling takes a separate position from the others, suggesting a hidden, morphologically still not recognized species. The rusbyi and the albicans groups are sister to each other in all trees, with moderate support (BS 78/73 PP 96; Figs. 1 and 2). The small weberbaueri group only comprises F. weberbaueri and F. batistana. It is sister to the highly supported micrantha group (BW 100/100, PP 100), which combines F. micrantha, F. villosula and F. christophii. A sister relationship between the micrantha plus weberbaueri group and the rusbyi plus albicans group receives good support in all trees (BS 98/99; PP 100). The weddelliana and penduliflora groups also harbor two species each. The former comprises all accessions of F. weddelliana and F. cotacajensis, but these two species are only poorly delimited from each other. Within the penduliflora group, the only available accession of the yellow-flowering F. gracilis is sister to the widely distributed F. penduliflora, a relationship that receives high levels of support. The undescribed morphospecies (F. spec. in Table 1) is nested within F. penduliflora, but differs from the latter in some morphological characters. A sister group relationship of the weddelliana and penduliflora groups is only weakly supported (BS 70; PP 91), and collapses in the strict consensus tree when indels are included in the analysis (Fig. 1, arrow). 4. Discussion The comparative sequencing of coding and noncoding regions of the chloroplast genome of Bromeliaceae has proven valuable and informative for phylogenetic reconstruction at the level of genera and subfamilies (Terry et al., 1997; Horres et al., 2000, 2007; Givnish et al., 2004, 2007; Crayn et al., 2000, 2004; Barfuss et al., 2005; Schulte et al., 2005). However, the relatively low level of sequence variation within the family renders any meaningful investigation at the infrageneric level difficult, and only few studies succeeded in reconstructing species trees from chloroplast data in Bromeliaceae (Givnish et al., 1997; Duval et al., 2003; Barfuss et al., 2005; De Sousa et al., 2007). The nuclear ribosomal internal transcribed spacer (ITS) region, which is a standard marker at low taxonomic levels in other systems, proved to be recalcitrant to DNA sequencing in Bromeliaceae (Tuthill and Brown, 2004; Barfuss, pers. comm.), and multilocus markers like RAPDs (Zizka et al., 1999; Ibisch et al., 2002) and AFLPs (Rex et al., 2007; Horres et al., 2007) have only rarely been used. As a consequence, species-level phylogenies only exist for a few genera in Bromeliaceae. In the present study, the combination of four moderately polymorphic cpDNA loci yielded a relatively well-resolved species tree in Fosterella. A similar increase in phylogenetic resolution by adding more loci has been observed by Barfuss et al. (2005) in their seven-locus chloroplast analysis of subfamily Tillandsioideae. 4.1. Monophyly and taxonomic position of Fosterella The gross topology of our trees and the arrangement of subfamilies is in general agreement with other studies. With the exception Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx 7 Fig. 1. Strict consensus tree of a parsimony ratchet analysis based on four chloroplast loci (matK gene, rps16 intron, atpB-rbcl and psbB-psBH intergenic spacer), coded indels excluded. The analysis yielded 1459 most parsimonious trees of 875 steps length (consistency index CI = 0.737, retention index RI = 0.922). When indels were included in the data matrix, 3529 most parsimonious trees of 1142 steps length were obtained (consistency index CI = 0.677, retention index RI = 0.899). Arrows indicate branches that collapse in the strict consensus tree when indels were included. Numbers below and above branches indicate bootstrap support values obtained with and without indels, respectively. of Puyoideae/Bromelioideae, all subfamilies suggested by Givnish et al. (2007)—as far as they are included in our investigation—are monophyletic. Puya is a large genus, and limited sampling may therefore account for its apparent non-monophyly in the Puyoi- deae/Bromelioideae clade. When sampling is more extensive, Puya comes out as the sister group of Bromelioideae (Schulte et al., 2005; Horres et al., 2007). The molecular systematics of Bromelioideae, Puyoideae, Tillandsioideae, Brocchinoideae, Lindmanioideae, Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 8 Fosterella rusbyi group micrantha weberbaueri group group penduliflora group Tilliandsioideae Puyoideae Bromelioideae Pitcairnia clade Dyckia clade weddellliana grou up F. rexiae 9d F. rexiae 10a F. kroemeri 28d 81 F. albicans 62a F. albicans 94c F. caulescens 142b 99 F. caulescens 3a F. caulescens 80a 79 F. graminea 71c F. robertreadii 86a 100 F. robertreadii 135b 89 F. robertreadii 143a 100 F. robertreadii 137b F. albicans 64a F. heterophylla 26a 99 F. vasquezii 63a F. vasquezii 23a 96 96 F. windischii 139b F. yuvinkae 72a 98 F. yuvinkae 140b 100 F. rusbyi 60a 100 100 F. rusbyi 141b F. rusbyi 107b 100 F. floridensis 67a 100 100 F. spectabilis 87a F. spectabilis 144a 100 F. weberbaueri 138b 99 F. weberbaueri 48d 100 F. weberbaueri 95c 100 F. weberbaueri 121a F. batistana 129a 100 100 F. micrantha 132a F. villosula 104a 100 F. micrantha 133a F. christophii 25f F. penduliflora 45c 100 F. penduliflora 46d 100 F. penduliflora 55a F. penduliflora 22c F. penduliflora 50b 98 F. penduliflora 93a F. spec. 4 19g F. penduliflora 35a 100 F. penduliflora 18c F. penduliflora 34a 100 F. penduliflora 136b F. penduliflora 118a F. penduliflora 120a F. gracilis 117a 100 F. cotacajensis 76d 91 100 F. cotacajensis 13a F. weddelliana 56a 100 F. weddelliana 37a F. weddelliana 36a 100 F. weddelliana 12a F. weddelliana 58c F. weddelliana 145a Dyckia encholiroides F13a 100 Dyckia estevesii F50a 74 Dyckia goehringii F39a 100 Dyckia estevesii F40a 100 Encholirium horridum F46a 100 Deuterocohnia scapigera F45a 100 Deuterocohnia brevispicata F51a 100 Deuterocohnia glandulosa F43a 100 100 Deuterocohnia lotteae F44a Deuterocohnia brevifolia F42a 100 Pitcairnia rubro-nigriflora F48a Pitcairnia breedlovei F14a 100 Pitcairnia grafii F16a Pitcairnia atrorubens F34a 100 Pitcairnia loki-schmidtiae F47a Pitcairnia albiflora F15a Aechmea kertesziae F9a 100 Aechmea gamosepala F17a 100 96 Neoregelia laevis F29a 100 Edmundoa lindenii F49a 85 Aechmea orlandiana F18a 100 Bromelia pinguin F38a 100 Bromelia serra F12a 100 Ochagavia litoralis F22a Ochagavia carnea F36a 98 Deinacanthon urbanianum F20a 100 90 Puya herzogii F53a 100 100 Puya mirabilis F41a Puya spathacea F52a Puya coerulea F23a 100 Racinea pugiformis F25a 100 Tillandsia fraseri F26a 100 Guzmania musaica F33a 100 Guzmania glaucophylla F8a 76 Catopsis floribunda F10a 100 Hechtia argentea F11a Hechtia stenopetala F21a Brocchinia acuminata F19a Brocchinia uaipanensis F7a albicans group M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx 0.1 Fig. 2. The 50% majority rule consensus tree of 3401 trees obtained from four rounds of Bayesian analysis of the combined data set (four chloroplast loci, coded indels excluded), implementing the GTR + G + I model. Branch lengths reflect changes per site. Posterior probabilities are given above branches. Navioideae and Hechtioideae have been analyzed in more detail elsewhere (Schulte et al., 2005; Barfuss et al., 2005; Horres et al., 2007; Givnish et al., 1997, 2007; Schulte and Zizka, 2008) and will not be discussed further here. Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Morphologically, Fosterella is distinguished from related genera of Pitcairnioideae s.str. by a number of flower characters, including naked petals, basifixed anthers (which are coiled at anthesis), and inner filaments adnate to the petals. Monophyly of the genus was supported by all molecular investigations where more than one Fosterella species was included (e.g., Horres et al., 2000; Crayn et al., 2004; Schulte et al., 2005; this study). Importantly, our four-locus tree also lends high support to the sister group relationship of Fosterella and a clade comprising Dyckia, Deuterocohnia (including Abromeitiella) and Encholirium. The same relationship has been detected earlier, but with low levels of support (Crayn et al., 2004; Givnish et al., 2007). It is interesting in at least two respects. First, all Fosterella species investigated so far conduct C3 photosynthesis only, whereas Dyckia, Deuterocohnia and Encholirium species (coined the ‘‘Dyckia” clade by Crayn et al., 2004) perform CAM (Martin, 1994; Crayn et al., 2004; Givnish et al., 2007). Second, given that sister groups by definition have the same age, the Dyckia clade with about 170 species (Luther, 2006) has diversified much more than Fosterella with only about 30 species. Presently, one can only speculate about the reasons behind the higher extant species richness in the Dyckia clade, which could be the result of a higher extinction rate in Fosterella, or a higher speciation rate in the Dyckia clade, or both. A possible scenario is that the common ancestor of Dyckia, Deuterocohnia and Encholirium has acquired the ability to conduct CAM photosynthesis. This could have served as a key innovation that fostered diversification in discontinuous arid habitats of the southern Andes, Argentina, and southern and eastern Brazil. Fosterella species lack this innovation, and remained mostly associated with more or less mesic habitats. 4.2. Infrageneric relationships in Fosterella The combined chloroplast tree resolves six distinct evolutionary lineages, which we refer to as rusbyi, albicans, weberbaueri, micrantha, weddelliana, and penduliflora groups (Figs. 1 and 2). Furthermore, we found evidence for a close association between (1) the rusbyi and the albicans group; (2) the weberbaueri and the micrantha group, and (3) the weddelliana and the penduliflora group. Finally, a moderately supported clade is formed by the rusbyi + albicans and the weberbaueri + micrantha groups. One drawback of our trees is the low resolution within each group, which is mainly due to limited sequence variation. In the following, the six species groups are discussed in some detail, and relationships suggested by the chloroplast trees are compared to those observed in our previous AFLP study (Rex et al., 2007). For better comparison, both trees are shown opposite to each other in Fig. 3. The names of several taxa have changed in the course of our ongoing revision (Peters et al., 2008a). To facilitate the comparison between the AFLP tree of Rex et al. (2007) and the present chloroplast trees, old and new names of the respective taxa are compiled in Table 3. The rusbyi group corresponds to groups A, B, C + F. spectabilis in the AFLP tree of Rex et al., 2007; see Fig. 3. It comprises a set of species from rather diverse environments, which intuitively would not have been grouped together. Apparently, the Andean cloud forest species F. rusbyi is related to the lowland species from the Brazilian shield, F. vasquezii, F. windischii and F. yuvinkae, which are morphologically clearly distinct from each other. Both F. floridensis and F. spectabilis are taxa from semi-humid areas south of the so-called Andean knee, a region in Bolivia where the main Andean cordillera takes a sharp bend towards south. Contrasting to their sister group, the latter two species are characterized by straight, non-recoiled petals. Fosterella spectabilis is the only representative of the genus with rather large, red (and possibly ornithophilous) flowers. The albicans group (corresponding to groups D, F, G, H + F. graminea in the AFLP tree of Rex et al., 2007; see Fig. 3) is entirely 9 Andean with a distribution from northern Argentina to southern Peru with a focus in the montane Yungas rain forests of La Paz in northern Bolivia. Morphologically, the group is highly heterogeneous, but petals that are recoiled like watchsprings characterize all of its species. Caulescence has developed in F. caulescens, F. rexiae, F. heterophylla and to some extent in F. albicans. The genetic heterogeneity of the widely distributed F. albicans requires further investigation. It is possible that the albicans-like populations north of the Andean knee, which represents an important climatic and biogeographical boundary, belong to another species yet to be described. With respect to the high variability concerning the key morphological characters illustrated in Fig. 4, the albicans group certainly takes a prominent position among the six lineages that we discern within the genus. It is currently unknown what evolutionary processes and factors could have fueled the extraordinary divergent development in this group, especially since distribution patterns and sequence variability are not significantly different from those of the other groups. Further investigations are needed, including field studies that aim to assess the spectrum of pollinators and their relevance for diversification. Unfortunately, pollination in Fosterella has only rarely been documented so far (see Ibisch et al., 2002). The micrantha group consists of three species and corresponds to group J in the AFLP tree of Rex et al. (2007; Fig. 3). Fosterella christophii and F. villosula are both characterized by relatively small distribution ranges in the Bolivian Andes, whereas F. micrantha represents an obvious case of long-distance dispersal from the tropical humid Andes to Central America (see Rex et al., 2007 for a discussion). Fosterella christophii occurs at lower altitudes in the semi-humid sub-Andean belt, whereas F. villosula is found in very humid Andean forests of the Chapare region. The weberbaueri group is closely related with the micrantha group in the chloroplast tree, but not in the AFLP tree of Rex et al. (2007), where it corresponds to group L (Fig. 3). It only contains two species, i.e., F. weberbaueri that also occurs at lower altitudes of the sub-Andean belt, and F. batistana, which is found in the Amazon lowlands of Brazil. The species of these two groups are rather broad-leaved, morphologically related to each other, and share the same flower morphology, with lily-like, slightly to strongly recurved petals. Fosterella cotacajensis and F. weddelliana together make up the weddelliana group (corresponding to group E in the AFLP tree of Rex et al., 2007). Both form a caulescent stem and carry serrate leaves. The position in the tree suggests that F. cotacajensis has been derived from a F. weddelliana progenitor conquering higher and more arid habitats. The penduliflora group (group K in the AFLP tree of Rex et al., 2007) comprises F. gracilis from northern Bolivia and the widely distributed and morphologically variable F. penduliflora from the southern tropical Andes (south of the Andean knee) and the Chiquitano lowlands. All plants of this group are acaulescent herbs with low rosettes and ‘lily-like’ flowers with slightly recurved petals. Previously, two additional taxa, F. chiquitana and F. latifolia, had been separated from F. penduliflora (Ibisch et al., 1999). Taking into account the results of our recent AFLP study, the high morphological variability of F. penduliflora and the lack of diagnostic characters, we decided to synonymize F. chiquitana and F. latifolia with F. penduliflora (Peters et al., 2008a). Nonetheless, the chloroplast DNA results suggest that the lowland plants from the Pre-Cambrian Brazilian shield are somewhat distinct, and that F. penduliflora may have begun to split up rather recently. Also a member of this group is the so far undetermined F. spec. 4. This specimen could well represent a hybrid involving F. penduliflora as one parent. The AFLP and cpDNA trees are somewhat difficult to compare, mainly because of the slightly uneven taxon representation in both samples (Fig. 3). Some species worked well in the sequencing Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 10 M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Fig. 3. Comparison of the infrageneric phylogenies of Fosterella deduced from AFLP analysis (neighbor joining tree, modified from Rex et al., 2007; left panel) and from DNA sequence data at four chloroplast loci (strict consensus tree of a parsimony ratchet analysis, modified from Fig. 1 of the present study; right panel). New names of Fosterella taxa (Table 3) have been used throughout. Table 3 New names (used in the present study) versus old names of Fosterella taxa (used in the AFLP tree of Rex et al., 2007). Collector Name in AFLP tree (Rex et al., 2007) Name in chloroplast tree (present study) RS 251002-3 CN 2061 CN 2076b RV 3636 RV 3612 TK 1398b PI 98.0173 PI 98.0098 RV 3406 RV 3762 RS 301002-3 RV 4685 Leme 5078 F. elata H. Luther F. elata H. Luther F. nowickii Ibisch, R. Vásquez & E. Gross F. nowickii Ibisch, R. Vásquez & E. Gross F. weddelliana (Brongn.) L.B. Sm. F. spec. 3 F. villosula (Harms) L.B. Sm. F. latifolia Ibisch, R. Vásquez & E. Gross F. latifolia Ibisch, R. Vásquez & E. Gross F. chiquitana Ibisch, R. Vásquez & E. Gross F. chiquitana Ibisch, R. Vásquez & E. Gross F. chiquitana Ibisch, R. Vásquez & E. Gross F. spec. 8 F. F. F. F. F. F. F. F. F. F. F. F. F. study, but were recalcitrant to AFLP analysis (most likely due to poor DNA quality, especially in the case of herbarium specimens). Nevertheless, the topologies of the cpDNA trees were generally congruent with those of the AFLP trees, and the six clades defined in the cpDNA tree roughly correspond to combinations of groups defined by AFLPs. A few taxa take contrasting positions in the two trees (e.g., F. spectabilis, F. graminea and F. gracilis) and need further study. In general, the monophyly of species groups and relationships between these are better resolved by the chloroplast rusbyi (Mez) L.B. Sm. rusbyi (Mez) L.B. Sm. weddelliana (Brongn.) L.B. Sm. weddelliana (Brongn.) L.B. Sm. cotacajensis M. Kessler, Ibisch, E. Gross kroemeri Ibisch, R. Vásquez & J. Peters christophii Ibisch, R. Vásquez & J. Peters penduliflora (C.H. Wright) L.B. Sm. penduliflora (C.H. Wright) L.B. Sm. penduliflora (C.H. Wright) L.B. Sm. penduliflora (C.H. Wright) L.B. Sm. penduliflora (C.H. Wright) L.B. Sm. batistana Ibisch, Leme & J. Peters syn. nov. syn. nov. syn. nov. syn. nov. rev. sp. nov. sp. nov. syn. nov. syn. nov. syn. nov. syn. nov. syn. nov. sp. nov. trees, whereas resolution within groups was higher in the AFLP trees, which also showed relatively long terminal branches. 4.3. Character evolution in Fosterella The states of ten selected morphological characters were coded for all species under study (Table 2), and then mapped onto the four-locus chloroplast DNA phylogeny. Character-state transformations for six of these characters are summarized in Fig. 4. Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx 11 Fig. 4. Inference of character evolution in Fosterella using parsimony. Morphological transitions for six selected characters (a–f) were mapped on the strict consensus tree of a parsimony ratchet analysis of four chloroplast loci, coded indels excluded, using MacClade. 4.3.1. Caulescence The majority of Fosterella species are acaulescent plants, forming a basal rosette. Pronounced caulescence (with stems >5 cm long in adult plants) has developed in only five species (F. cotacajensis, F. weddelliana, F. heterophylla, F. rexiae and F. caulescens). In two additional species (F. weberbaueri and F. albicans), adult plants are subcaulescent, i.e., they develop a shorter, rather inconspicuous stem less than 5 cm long. Mapping this character onto the phylogeny (Fig. 4a) clearly indicates that acaulescence is the ancestral state within the genus. Apparently, Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 12 M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx caulescence has evolved independently at least twice: once in the weddelliana group and once in the albicans group. Subcaulescence is also found in two distinct lineages, and might be regarded as an intermediate character state in the albicans group. twice) within the albicans group, and once in F. floridensis. A spiral arrangement of leaves along an elongated stem independently evolved within two lineages (the weddelliana and the albicans group), together with the caulescent habit. 4.3.2. Leaf blade margins Most Fosterella species possess entire leaves, but seven species (F. cotacajensis, F. weddelliana, F. rusbyi, F. rexiae, F. caulescens, F. albicans, and F. graminea) are characterized by serrate leaf margins. Serration is generally more pronounced above the leaf sheath, and becomes less distinct to absent towards the leaf apex. Parsimony reconstruction of the character suggests that entire leaves are ancestral in the genus, with serrate leaves being derived (Fig. 4b). The reconstruction of character evolution based on the molecular phylogeny indicates that serrate leaves evolved at least three times independently, i.e., within the weddelliana, the rusbyi and the albicans group. 4.3.6. Vestiture of the inflorescence axis Most Fosterella species exhibit a glabrous or glabrescent inflorescence axis. Several species possess an inflorescence axis that is conspicuously densely lanate, as in F. christophii, F. villosula, F. floridensis, F. caulescens and F. albicans, whereas others are only sparsely lanate (F. micrantha, F. batistana, F. robertreadii, F. rexiae, F. pearcei, F. petiolata, and F. aletroides). Tracing the character on the molecular phylogeny indicates that glabrous/glabrescent inflorescences are ancestral within the genus and lanate inflorescences are the derived condition (Fig. 4f). The molecular phylogeny implies several independent origins of densely lanate inflorescences, i.e., once each in the micrantha and the albicans group as well as once in F. floridensis. With the exception of F. batistana, sparsely lanate inflorescences are only found in groups that have also developed densely lanate inflorescences (the micrantha and the albicans group). The former might therefore represent an intermediate character state. 4.3.3. Inflection of petals during and after anthesis Several years ago, the late Robert W. Read (pers. comm.) discovered a morphological difference that divides Fosterella into two subgroups: one with petals that are recoiled like watchsprings and stay so after anthesis; and the other one with straight or ‘‘lily-like”, more or less recurved petals that become straight again after anthesis. Read felt that two subgenera could be described based upon this difference, but never published his observations. In order to reconstruct the evolution of the character, we retained the ‘‘recoiled like watchsprings” category, but further subdivided the ‘‘straight or lily-like” category of Read into (1) petals that remain straight during and after anthesis, (2) petals that are slightly recurved (‘‘lily-like‘‘) during anthesis and straight afterwards and (3) petals that are strongly recurved during anthesis and become straight afterwards. Mapping the four different character states onto the molecular phylogeny revealed that a principal subdivision of the genus based upon this character would not reflect natural groups (Fig. 4c). Nevertheless, several groups are indeed characterized by the presence of a single, peculiar type of petal inflection. For example, the species of the penduliflora and the micrantha group all possess slightly recurved petals during and after anthesis, whereas all species of the weddelliana and the albicans group have petals that are recoiled like watchsprings during anthesis and remain so afterwards. The rusbyi group is quite heterogeneous in this respect, with three different types of petal inflection, and the weberbaueri group exhibits two types (Fig. 4c). Parsimony reconstruction did not answer the question of which type of petal inflection might be regarded as ancestral. Nevertheless, it becomes obvious that the different types must have been gained or lost several times independently within the genus. 4.3.4. Peduncle bract margins The majority of Fosterella species possess inflorescences with entire peduncle bracts. Serrate peduncle bracts are only found in F. weddelliana and F. rexiae. Mapping of the character on the phylogeny (Fig. 4d) suggests that entire peduncle bracts are the ancestral state and that serrate peduncle bracts evolved twice independently. 4.3.5. Leaf arrangement Leaves of mature plants are arranged in a flat rosette in most Fosterella species. In F. floridensis, F. albicans, F. kroemeri, and F. petiolata, mature plants possess an erect, funnel-shaped rosette. In the caulescent species, leaves are spirally arranged along the elongated stem. The inferred evolution of leaf arrangement indicates that flat rosettes are the ancestral condition within the genus (Fig. 4e). As the reconstruction of character evolution based on the molecular phylogeny implies, erect rosettes evolved at least once (and maybe 4.3.7. Trichomes The majority of Fosterella species carry peltate trichomes on the abaxial leaf. Stellate trichomes are only found in F. gracilis and F. yuvinkae. The fine structure of peltate trichomes varies within the genus. Thus, the margin of the trichome shield can either be entire, dentate or consist of elongated cells that form a loose fringe. Mapping the character on the molecular phylogeny suggests that peltate trichomes with entire margins represent the ancestral condition within the genus (not shown). The other three types each evolved several times independently, and trichome types also vary within most groups. Only three of the six Fosterella lineages defined by our chloroplast trees are characterized by a single trichome type. These are the micrantha and the weberbaueri group (both with peltate trichomes with entire margins) and the weddeliana group (with peltate trichomes and elongated rim cells that form a loose fringe). The character appears to be highly variable within the genus and therefore of limited taxonomic utility. 4.3.8. Vestiture on the abaxial leaf surface Within Fosterella, the density of vestiture on the abaxial leaf surface ranges from glabrescent to very dense, and then completely covering the leaf surface. Character state reconstruction suggests that a sparse vestiture on the abaxial leaf is the ancestral condition (not shown). A very dense vestiture appears to have evolved in three lineages independently, i.e., within the weddelliana, the rusbyi and the albicans group. Glabrescent leaves are typical for the weberbaueri group and for F. spectabilis, and apparently evolved in both lineages independently. The character shows a high variability, especially within the rusbyi group. 4.3.9. Inflorescence: ramification Most Fosterella species are characterized by a panicle. Only few species develop a raceme or a compound raceme. The reconstruction of the evolution of the inflorescence type implies that the panicle is ancestral and the raceme/compound raceme represents the derived condition (not shown). Racemes/compound racemes appear to have evolved independently within three separate lineages: in the penduliflora and the weberbaueri group as well as in F. spectabilis. 4.3.10. Petal color All but two Fosterella species have white petals. Only F. gracilis and F. spectabilis exhibit a different petal color, which is yellow Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx and red, respectively. As the molecular phylogeny implies (not shown), white flowers can be regarded as the ancestral condition within the genus. 5. Conclusions In conclusion, our phylogenetic study has provided us with good evidence that Fosterella is monophyletic and sister to a clade that comprises Dyckia, Deuterocohnia and Encholirium. We have further demonstrated that Fosterella is divided into six well-supported evolutionary lineages that show only limited correspondence with morphological traits or geographical distribution patterns. The majority of character states, which were mapped on our trees, proved to have evolved several times independently. It is therefore not surprising that previous attempts to group Fosterella species according to morphological traits were not successful. Further studies are needed to elucidate the colonization history of the Andes by Fosterella, and the evolutionary origin of narrow endemics. These studies will require the development of markers with improved resolution at and below the species level, such as nuclear and chloroplast microsatellites. Acknowledgments The authors acknowledge support by the Deutsche Forschungsgemeinschaft (DFG grants We 1830/5-1, IB 85/1-1 and ZI 557/6-1), the DAAD (travel grant to Jule Peters), and the University of Kassel (Ph. D fellowship grant to Jule Peters). The authors also acknowledge funding from the Hessian initiative for the development of scientific and economic excellence (LOEWE) at the Biodiversity and Climate Research Centre, Frankfurt/Main. Part of the plant material was kindly supplied by the Palmengarten Frankfurt am Main and the Botanical Gardens of Heidelberg, Berlin and Osnabrück, Marie Selby Botanical Garden, Sarasota, FL and by the living collection of FAN, where Arturo Osinaga provided invaluable help. We also thank Elton Leme, the Deutsche Bromeliengesellschaft and the Möllgard Stiftung for support, and Christoph Nowicki, and Robert Müller for help with collections and data. References Akaike, H., 1974. A new look at the statistical identification model. IEEE Trans. Automat. Control. 19, 716–723. Barfuss, 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. Crayn, D.M., Terry, R.G., Smith, J.A.C., Winter, K., 2000. Molecular systematic investigations in Pitcairnioideae (Bromeliaceae) as a basis for understanding the evolution of Crassulacean Acid Metabolism (CAM). In: Wilson, K.L., Morrison, D.A. (Eds.), Monocots: Systematics and Evolution. CSIRO, Melbourne, pp. 569– 579. Crayn, D.M., Winter, K., Smith, J.A.C., 2004. Multiple origins of crassulacean acid metabolism and the epiphytic habit in the Neotropical family Bromeliaceae. Proc. Natl. Acad. Sci. USA 101, 3703–3708. 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. Duval, M.-F., Buso, G.S.C., Ferreira, F.R., Noyer, J.L., Coppens d́Eeckenbrugge, G., Hamon, P., Ferreira, M.E., 2003. Relationships in Ananas and other related genera using chloroplast DNA restriction site variation. Genome 46, 990–1004. Felsenstein, J., 1985. Confidence limits on phylogenies, an approach using the bootstrap. Evolution 39, 783–791. Givnish, T.J., Sytsma, K.J., Smith, J.F., Hahn, W.J., Benzing, D.H., Burkhardt, E.M., 1997. Molecular evolution and adaptive radiation in Brocchinia (Bromeliaceae, Pitcairnioideae) atop tepuis of the Guayana shield. In: Givnish, T.J., Sytsma, K.J. (Eds.), Molecular Evolution and Adaptive Radiation. Cambridge University Press, UK, pp. 259–311. Givnish, T.J., Millam, K.C., Evans, T.M., Hall, J.C., Pires, J.C., Berry, P.E., Sytsma, K.J., 2004. Ancient vicariance or recent long-distance dispersal? Inferences about phylogeny and South American–African disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int. J. Plant Sci. 165, S35–S54. 13 Givnish, T.J., Millam, K.C., Evans, T.M., Pires, J.C., Berry, P.E., Sytsma, K.J., 2007. Phylogeny, biogeography, and ecological evolution in Bromeliaceae, insights from ndhF sequences. Aliso 23, 3–26. Hilu, K.W., Borsch, T., Müller, K., Soltis, D.E., Soltis, P.S., Savolainen, V., Chase, M.W., Powell, M.P., Alice, L.A., Evans, R., Sauguet, H., Neinhuis, C., Slotta, T.A.B., Rohwer, J.G., Campbell, C.S., Chatrou, L.W., 2003. Angiosperm phylogeny based on matK sequence information. Am. J. Bot. 90, 1758–1776. Horres, R., Zizka, G., Kahl, G., Weising, K., 2000. Molecular phylogenetics of Bromeliaceae, evidence from trnLUAA intron sequences of the chloroplast genome. Plant Biol. 2, 306–315. Horres, R., Schulte, K., Weising, K., Zizka, G., 2007. Systematics of Bromelioideae (Bromeliaceae)—evidence from molecular and anatomical studies. Aliso 23, 27– 43. Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference of phylogenetic trees. Bioinformatics 17, 754–755. Ibisch, P.L., Gross, E., Rauer, G., Rudolph, D., 1997. On the diversity and biogeography of the genus Fosterella L.B. Sm. (Bromeliaceae) with the description of a new species from Eastern Bolivia. J. Bromel. Soc. 47, 211–217. Ibisch, P.L., Vásquez, R., Gross, E., 1999. More novelties of Fosterella L.B. Smith (Bromeliaceae), Pitcairnioideae from Bolivia. Rev. Soc. Boliv. Bot. 2, 117–132. Ibisch, P.L., Vásquez, R., Gross, E., Krömer, T., Rex, M., 2002. Novelties in Bolivian Fosterella (Bromeliaceae). Selbyana 23, 204–219. Ibisch, P.L., Read, R., Peters, J., 2008. Key to the species of the genus Fosterella (Bromeliaceae). Selbyana 29, 195–198. Johnson, L.A., Soltis, D.E., 1995. Phylogenetic inference in Saxifragaceae sensu stricto and Gilia (Polemoniaceae) using matK sequences. Ann. Mo. Bot. Gard. 82, 149– 175. Kessler, M., Ibisch, P.L., Gross, E., 1999. Fosterella cotacajensis, una nueva especie de Bromeliaceae de los valles secos andinos de Bolivia. Rev. Soc. Boliv. Bot. 2, 111– 116. Luther, H.E., 1981. Miscellaneous new taxa of Bromeliaceae I. Selbyana 5, 310–311. Luther, H.E., 1997. A showy new Fosterella from Bolivia. J. Bromel. Soc. 47, 118–119. Luther, H.E., 2006. An Alphabetical List of Bromeliad Binomials, 10th ed. Marie Selby Botanical Gardens, Sarasota, FL. Maddison, D.R., Maddison, W.P., 2003. MacClade 4. Sinauer, Sunderland, MA. Manen, J.F., Natali, A., Ehrendorfer, F., 1994. Phylogeny of Rubiaceae-Rubieae inferred from the sequence of a cpDNA intergene region. Plant Syst. Evol. 190, 195–211. Martin, C.E., 1994. Physiological ecology of the Bromeliaceae. Bot. Rev. 60, 1–82. Müller, K., 2003. PRAP-computation of Bremer support for large data sets. Mol. Phylogenet. Evol. 31, 780–782. Nixon, K.C., 1999. The parsimony ratchet, a new method for rapid parsimony analysis. Cladistics 15, 407–414. Oxelman, B., Lidén, M., Berglund, D., 1997. Chloroplast rps16 intron phylogeny of the tribe Sileneae (Caryophyllaceae). Plant Syst. Evol. 206, 393–410. Peters, J., Vásquez, R., Osinaga, A., Leme, E., Weising, K., Ibisch, P.L., 2008a. Towards a taxonomic revision of Fosterella (Bromeliaceae). Selbyana 29, 182–194. Peters, J., Weising, K., Ibisch, P.L., 2008b. Fosterella nicoliana, eine neue Art des peruanischen Amazonas-Tieflands mit einem bislang unbekannten Samentyp. Die Bromelie 2008(2), 64–69. Posada, D., Crandall, K.A., 1998. Modeltest: testing the model of DNA substitution. Bioinformatics 14, 817–818. Rauh, W., 1979. Bromelienstudien. I. Neue und wenig bekannte Arten aus Peru und anderen Ländern 9. Mitteilung. 4. Fosterella. Trop. Subtrop. Pflanzenwelt 31, 23– 29. Rauh, W., 1987. Bromelienstudien. 4. Fosterella L.B. Sm.. Trop. Subtrop. Pflanzenwelt 60, 24–26. Rex, M., Patzolt, K., Schulte, 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. Schulte, K., Zizka, G., 2008. Multi locus plastid phylogeny of Bromelioideae (Bromeliaceae) gives new insights into the taxomomic utility of petal appendages and pollen characters. Candollea 63, 209–225. Schulte, K., Horres, R., Zizka, G., 2005. Molecular phylogeny of Bromelioideae and its implications on biogeography and the evolution of CAM in the family. Senckenbergiana Biol. 85, 113–125. Simmons, M.P., Ochoterena, H., 2000. Gaps as characters in sequence-based phylogenetic analyses. Syst. Biol. 49, 369–381. Smith, L.B., Downs, R.J., 1974. 13. Fosterella. In: Flora Neotropica, Monograph No. 14, Pitcairnioideae. Hafner Press, New York, pp. 199–209. Smith, L.B., Read, R.W., 1992. Flora Neotropica, Monograph No. 14 1, Suppl. No. 3. Fosterella. Bradea 6, pp. 134–140. Smith, L.B., Till, W., 1998. Bromeliaceae. In: Kubitzki, K. (Ed.), The Families and Genera of Vascular Plants, vol. 4. Springer, Berlin, Heidelberg, New York, pp. 74– 99. Swofford, D.L., 2002. PAUP, Phylogenetic Analysis Using Parsimony,  and other methods, version 4.0b10. Sinauer, Sunderland, MA. Terry, R.G., Brown, G.K., Olmstead, R.G., 1997. Examination of subfamilial phylogeny in Bromeliaceae using comparative sequencing of the plastid locus NDHF. Am. J. Bot. 84, 664–670. Tuthill, D., Brown, G.K., 2004. Bromeliad endophytes and the serendipity of science. J. Bromeliad Soc. 54, 28–32. Varadarajan, G.S., Gilmartin, A.J., 1988a. Phylogenetic relationships of genera within the subfamily Pitcairnioideae (Bromeliaceae). Syst. Bot. 13, 283–293. Varadarajan, G.S., Gilmartin, A.J., 1988b. Taxonomic realignments within the subfamily Pitcairnioideae (Bromeliaceae). Syst. Bot. 13, 294–299. Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001 ARTICLE IN PRESS 14 M. Rex et al. / Molecular Phylogenetics and Evolution xxx (2009) xxx–xxx Wallander, E., Albert, V.A., 2000. Phylogeny and classification of Oleaceae based on rps16 and trnL-F sequence data. Am. J. Bot. 87, 1827–1841. 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. Xu, D.H., Abe, J., Sakai, M., Kanazawa, A., Shimamoto, Y., 2000. Sequence variation of non-coding regions of chloroplast DNA of soybean and related wild species and its implications for the evolution of different chloroplast haplotypes. Theor. Appl. Genet. 101, 712–732. Young, N.D., Healy, J., 2003. GapCoder automates the use of indel characters in phylogenetic analyses. BMC Bioinformatics 4, 1–6. Zizka, G., Horres, R., Nelson, E.C., Weising, K., 1999. Revision of the genus Fascicularia Mez (Bromeliaceae). Bot. J. Linn. Soc. 129, 315–332. Please cite this article in press as: Rex, M., et al. Phylogenetic analysis of Fosterella L.B. Sm. (Pitcairnioideae, Bromeliaceae) ... Mol. Phylogenet. Evol. (2009), doi:10.1016/j.ympev.2009.01.001