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