This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Molecular phylogeny and character evolution in terete-stemmed Andean opuntias
, J. Reiker
, G. Charles
, P. Hoxey
, D. Hunt
, M. Lowry
, W. Stuppy
, N. Taylor
Senckenberg Museum of Natural History Görlitz, Am Museum 1, D-02826 Görlitz, Germany
Justus-Liebig-University Gießen, Institute of Botany, Department of Systematic Botany, Heinrich-Buff-Ring 38, D-35392 Gießen, Germany
International Organization for Succulent Plant Study, c/o David Hunt, Hon. Secretary, 83 Church Street, Milborne Port DT9 5DJ, United Kingdom
Millennium Seed Bank, Royal Botanic Gardens, Kew & Wakehurst Place, Ardingly, West Sussex RH17 6TN, United Kingdom
Singapore Botanic Gardens, 1 Cluny Road, Singapore 259569, Singapore
Received 22 November 2011
Revised 23 April 2012
Accepted 23 July 2012
Available online 2 August 2012
Nuclear ribosomal DNA
The cacti of tribe Tephrocacteae (Cactaceae–Opuntioideae) are adapted to diverse climatic conditions
over a wide area of the southern Andes and adjacent lowlands. They exhibit a range of life forms from
geophytes and cushion-plants to dwarf shrubs, shrubs or small trees. To conﬁrm or challenge previous
morphology-based classiﬁcations and molecular phylogenies, we sampled DNA sequences from the chlo-
roplast trnK/matK region and the nuclear low copy gene phyC and compared the resulting phylogenies
with previous data gathered from nuclear ribosomal DNA sequences. The here presented chloroplast
and nuclear low copy gene phylogenies were mutually congruent and broadly coincident with the clas-
siﬁcation based on gross morphology and seed micro-morphology and anatomy. Reconstruction of hypo-
thetical ancestral character states suggested that geophytes and cushion-forming species probably
evolved several times from dwarf shrubby precursors. We also traced an increase of embryo size at
the expense of the nucellus-derived storage tissue during the evolution of the Tephrocacteae, which is
thought to be an evolutionary advantage because nutrients are then more rapidly accessible for the ger-
minating embryo. In contrast to these highly concordant phylogenies, nuclear ribosomal DNA data sam-
pled by a previous study yielded conﬂicting phylogenetic signals. Secondary structure predictions of
ribosomal transcribed spacers suggested that this phylogeny is strongly inﬂuenced by the inclusion of
paralogous sequence probably arisen by genome duplication during the evolution of this plant group.
Ó2012 Elsevier Inc. All rights reserved.
Cacti have fascinated botanists since the discovery of the Amer-
icas in the 15th century. Botanists and gardeners have paid much
attention to their remarkable evolution of succulence and their
large colourful ﬂowers. But whereas cactus enthusiasts have gener-
ally focused their interest on the smaller-growing members of the
very diverse subfamily Cactoideae (containing approximately 80%
of cactus species), the subfamily Opuntioideae has not enjoyed
the same attention. The subfamily Opuntioideae is widespread
throughout the Americas from Canada to southern Patagonia. It
has traditionally been recognised as a monophyletic taxonomic en-
tity (Anderson, 2001; Backeberg, 1966; Britton and Rose, 1919;
Hunt et al., 2006; Schumann, 1897–1898; Stuppy, 2002). It is char-
acterised by a number of synapomorphies: (1) presence of glochids:
small, deciduous barbed spines (Robinson, 1974); (2) woody funic-
ular tissue surrounding the seed (‘funicular envelope’, Stuppy,
2002); (3) high amounts of calcium oxalate monohydrate druses
and monoclinic cluster crystals in the outer hypodermis of stems
(Gibson and Nobel, 1986; Hartl et al., 2007), and (4) polyporate pol-
len grains with peculiar exine structures (Leuenberger, 1976).
Molecular phylogenetic investigations support the monophyly
of the Opuntioideae (Bárcenas et al., 2011; Edwards et al., 2005;
Grifﬁth and Porter, 2009; Hernández-Hernández et al., 2011; Nyff-
eler, 2002; Wallace and Dickie, 2002) but the sister group relation-
ship to one of the other subfamilies of the Cactaceae remains
unclear (Bárcenas et al., 2011; Hernández-Hernández et al., 2011;
Nyffeler, 2002). Traditional classiﬁcations of the Opuntioideae
based on gross morphology have treated Opuntia (L.) Mill. itself
as a large genus of up to 200 species, subdivided into infrageneric
units (Barthlott and Hunt, 1993; Britton and Rose, 1919; Endler
and Buxbaum, 1974; Rowley, 1958; Schumann, 1897–1898), or
independent genera (Backeberg, 1958–1962). Current classiﬁca-
tions recognise about 15 genera (Anderson, 2001, 2011; Hunt
1055-7903/$ - see front matter Ó2012 Elsevier Inc. All rights reserved.
Abbreviation: nrITS, nuclear ribosomal transcribed spacer.
Corresponding author. Fax: +49 (0)3581 47605102.
E-mail address: Christiane.firstname.lastname@example.org (C.M. Ritz).
These authors contributed equally to the study.
Molecular Phylogenetics and Evolution 65 (2012) 668–681
Contents lists available at SciVerse ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
Author's personal copy
et al., 2006; Nyffeler and Eggli, 2010; Stuppy, 2002) arranged in up
to ﬁve tribes (Doweld, 1999; Wallace and Dickie, 2002).
In the present study we focus on the South American spherical
to terete-stemmed Opuntioideae of the tribe Tephrocacteae sensu
Hunt (2011), which consists of the genera Austrocylindropuntia
Backeb.,Cumulopuntia F.Ritter, Maihueniopsis Speg., Punotia
D.R.Hunt, Pterocactus K.Schum. and Tephrocactus Lem. (Fig. 1).
Thus, Tephrocacteae in its broader circumscription sensu Hunt
(2011) include the tribes Austrocylindropuntieae Wallace & Dickie
and Pterocacteae Doweld. These genera develop many different life
forms ranging from small geophytes, hemispherical cushion-
plants, dwarf shrubs, shrubs and columnar cacti consisting of
either indeterminate branches (in Austrocylindropuntia) or deter-
minate terete or spherical segments. Although some of the genera
are closely similar morphologically, seed anatomical structures
provide diagnostic characters to differentiate them (Stuppy, 2002).
The phylogenetic relationships within the Tephrocacteae and its
position within the Opuntiodeae are not yet ﬁrmly understood. In
the molecular phylogenetic studies of Bárcenas et al. (2011) and
Hernández-Hernández et al. (2011), the Tephrocacteae formed a
polytomy with the two major clades of Opuntioideae: the terete
stemmed Opuntioideae (Cylindropuntieae sensu Hunt et al.,
Fig. 1. Distribution area of the genera of the tribe Tephrocacteae in Southern South America. Localities known to the authors are presented as dots, localities of specimens
analysed during this study are presented as triangles.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 669
Author's personal copy
2006) and the mainly ﬂat-stemmed Opuntioideae (Opuntieae sen-
su Hunt et al., 2006). However, the phylogeny presented by Grifﬁth
and Porter (2009) indicated extensive polyphyly of the genus Mai-
hueniopsis because some species clustered with the ﬂat- or terete-
stemmed Opuntioideae or were sister to all other Opuntiodeae.
The phylogeny from Grifﬁth and Porter (2009) was based on a
combined data set of the chloroplast trnL-trnF intergenic spacer
and the nuclear ribosomal internal transcribed spacer sequences
(nrITS). In this study the chloroplast DNA sequences were much
less variable than the nrITS sequences (72% of variable sites in
the combined alignment). Thus, the nrITS data obviously strongly
inﬂuenced the topology of the resulting tree. Nuclear rITS data have
been widely used for phylogenetic reconstructions, especially at
infrageneric level e.g. (Baldwin et al., 1995; Bruyns et al., 2006;
Martins, 2006; von Hagen and Kadereit, 2001), because the high
copy numbers of the tandemly repeated ribosomal DNA arrays in
the nucleolar organiser region (NOR) facilitates its ampliﬁcation.
Variation between individual nrITS copies is usually rapidly
homogenised by unequal cross overs and gene conversion de-
scribed as concerted evolution (Bailey et al., 2003; Buckler and Hol-
tsford, 1996; Eickbush and Eickbush, 2007). However, several
studies, including some within Cactaceae, have demonstrated that
intra-individual nrITS polymorphisms originating by gene or gen-
ome duplication (paralogous genes) persist because mechanisms
of sequence homogenisation are retarded or lacking (Harpke and
Peterson, 2006, 2007, 2008b; Hartmann et al., 2001, 2002). Some
of the paralogous copies might become non-functional (pseudo-
genes) and often evolve at higher mutation rates, which can be re-
ﬂected by a lower GC-content of the sequences and less stable RNA
secondary structures (Harpke and Peterson, 2006, 2008b). The evo-
lution and persistence of paralogous loci may thus result in errone-
ous species trees, if incompletely sampled, but may also provide
the opportunity to identify ancient paralogs or to unravel the
hybridogenic origin of a taxon when homeologs can be traced in
parental species (Alvarez and Wendel, 2003; Bailey et al., 2003).
The aim of our study was to disentangle the phylogenetic rela-
tionships between the Tephrocacteae sensu Hunt (2011).We
therefore sequenced the chloroplast trnK/matK region and the nu-
clear low copy gene phyC to reconstruct molecular phylogenies.
Using ancestral character state reconstructions of morphological
traits, we traced the evolution of different life-forms and anatom-
ical structures of seeds within the Tephrocacteae. We closely
examined nrITS sequences published by Grifﬁth and Porter
(2009) by analysing their GC-content and their secondary structure
to assess whether the nrITS based phylogeny represents or fails to
represent a species tree because of paralogous nrITS sequences.
2. Material and methods
2.1. Plant material
We analysed 45 taxa of the Tephrocacteae sensu Hunt (2011):
we sampled ﬁve species out of six species of Austrocylindropuntia,
10 species out of 10 species of Cumulopuntia, 11 species out of 12
species of Maihueniopsis, three species out of nine species of Ptero-
cactus, the monotypic genus Punotia and nine species out of nine
species of Tephrocactus. We followed the nomenclature of Hunt
et al. (2006) and Hunt (2011). Taxonomic adjustments resulting
from this study have separately published by Hunt (2011). The
plant material used was taken from vegetatively propagated spec-
imens in documented collections held in Europe including the Roy-
al Botanic Gardens Kew (Table S1). Specimens of investigated
plants were deposited in the spirit collection of the Royal Botanic
Gardens Kew (K). In accordance with previous studies (Bárcenas
et al., 2011; Grifﬁth, 2002; Hernández-Hernández et al., 2011;
Nyffeler, 2002) we used sequences of Portulaca oleracea L., Maihue-
nia patagonica Britton & Rose, Pereskia grandifolia Pfeiff., Pereskiop-
sis spp. Britton & Rose, Opuntia quimilo K.Schum.,O. sulphurea
G.Don., Brasiliopuntia brasiliensis (Willd.) A.Berger and Tunilla spp.
D.R.Hunt & Iliff as outgroups. Sampling details and GenBank acces-
sion numbers are presented in Table S1 in the Supplementary
2.2. DNA Extraction, sequence isolation
DNA was extracted using Qiagen Plant Mini Kit (Qiagen, Hilden,
Germany) according to the manufacturer’s instructions but all cen-
trifugation steps were performed at 14,000 rpm.
The trnK intron including the matK gene (trnK/matK region) was
ampliﬁed in 25
l reactions using the primers trnK-3914F and
trnK-2R (Johnson and Soltis, 1994). The reaction mixture con-
l 10-fold concentrated polymerase buffer (100 mM
Tris-HCl pH 8.8, 500 mM KCl, and 0.8% (v/v) Nonidet (P40), 2.5
(25 mM), 2.5
l dNTPs (2 mM), 1
l of each primer
M), 1 unit of Taq polymerase (MBI Fermentas, St. Leon-Rot,
Germany), and 1
l DNA template from diluted extracts (app.
50 ng DNA). The ampliﬁcation protocol started with an initial
denaturation of 120 s at 94 °C, after which 35 cycles were per-
formed each consisting of 30 s of denaturation at 94 °C, 45 s of
annealing of 49.6 °C, and 180 s of extension at 72 °C and ended
with a ﬁnal extension of 180 s at 72 °C.
Ampliﬁcation of the exon 1 region of phyC followed the same
protocol as for the trnK/matK region but primers were taken from
Helsen et al. (2009) and annealing temperature was set to 55 °C.
Puriﬁed PCR products (Qiaquick Gel extraction Kit, Qiagen, Hilden,
Germany) were cloned into vector pJet1 (CloneJet PCR cloning Kit,
Thermo Fisher Scientiﬁc, Schwerte, Germany). Ligation products
were electroporated into E. coli DH5
PCR products puriﬁed with ExoSAP-IT Kit (Affymetrix, Santa
Clara, CA, USA) or plasmids of ﬁve positive clones per taxon were
sequenced by the company Macrogen (South Korea, Seoul) using
the same primers as for ampliﬁcation and in case of the trnK/matK
region with additional internal primers trnk-23F and trnK-71R
2.3. Phylogenetic analyses
Sequences of the trnK/matK region and phyC gene were aligned
manually. Alignment of the nrITS data set (nrITS-1,2and 5.8S
rDNA) was taken from Grifﬁth and Porter (2009). The best ﬁtting
nucleotide substitution models for the phyC and the trnK/matK re-
gion (the matK gene within the trnK intron was analysed as sepa-
rate partition) were estimated with MrModeltest v. 2.3 according
to the corrected Akaike Information Criterion (Table 1,Nylander,
2004). The resulting model parameters were employed to recon-
struct four phylogenies based on different data sets: (1) trnK/matK
region, (2) phyC, (3) nrITS and (4) sequences combined trnK/matK
region and phyC sequences.
The alignment of combined markers was pruned to taxa se-
quenced for both markers; we randomly chose one of the phyC se-
quences but excluded sequences which were apparently not
orthologous (C. boliviana ssp. ignescens O-08/1-3, 5; C. chichensis
O-13/2-3; Maihueniopsis hickenii O-22/2, 8 and Tunilla orurenis O-
05/2). For Pereskiopsis we combined the chloroplast sequence of
P. diguetii and P. aquosa. We controlled for combinability of trnK
and phyC data sets using incongruence length difference (ILD) test
(Farris et al., 1995) implemented in Paup 4.0b10 (Swofford, 2002)
employing a heuristic search with TBR branch swapping, 1,000
partition replicates, each with 10 random sequence addition repli-
cates and a maximum of 500 saved trees.
670 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
We ran all analyses with MrBayes v.3.1.2 (Huelsenbeck and
Ronquist, 2001) with two simultaneous runs over 10,000,000 gen-
erations, sampling every 100th generation and discarding the ﬁrst
25,000 trees as burn-in resulting in a 50% majority rule consensus
tree showing all compatible partitions supported by posterior
probabilities (PP) for each node. Additionally, we calculated Maxi-
mum Likelihood Bootstrap percentages using RaxML (Stamatakis,
2006). We ran 1000 replicates of standard bootstrap using the
model GTR + G.
2.4. Morphological traits and character state reconstruction
For this we used the phylogeny based on the alignment of com-
bined sequences because statistical support was highest in this
tree and the ILD test did not reject homogeneity between parti-
tions (P= 0.12). The analyses were done using the Markov-k-state
1 parameter model (Maddison and Maddison, 2006) implemented
in the Mesquite package (Maddison and Maddison, 2010). We
reconstructed ancestral character states of seven morphological
characters observed in living material or (seed characters) taken
from Stuppy (2002). We assigned only unambiguous character
states; in case several character states were observed in one spe-
cies, we treated them as unknown. The following traits were
scored: (1) roots (ﬁbrous; tuberous (including taproots formed by
root tissue only or from root and hypocotyls tissue); (2) growth
mode (indeterminate, determinate); (3) life form (tree; columnar;
shrub; dwarf shrub; cushion; geophytes); (4) leaf rudiments (per-
sistent; caducous); (5) pericarp (juicy, indehiscent; dry dehiscent);
(6) shape of embryo and amount of perisperm (circular embryo
curved around large perisperm; hook-shaped embryo curved
around reduced perisperm; hook-shaped embryo curved around
very strongly reduced perisperm; spirally enrolled embryo bent
around strongly reduced perisperm); (7) anatomy of the funicular
envelope (all cells of the funicular envelope orientated parallel to
the funicular girdle; cells subtending the funicular girdle trans-
versely orientated and continuing on the insides of the ﬂanks; cells
subtending the funicular girdle transversely orientated and contin-
uing on the outsides of the ﬂanks; inner layers of the funicular
envelope sclerenchymatous, outer layers and funicular girdle
aerenchymatous; cells subtending the funicular girdle without dis-
tinct orientation). We analysed characters for species of the in-
group only, and for Brasiliopuntia brasiliensis,Tunilla orurenis and
2.5. GC-content and secondary structure models of nrITS
In order to assess the presence of pseudogenes within the nrITS
data set taken from Grifﬁth and Porter (2009) we determined the
GC-content of nrITS sequenced using Paup v. 4.0b10 (Swofford,
2002). Minimum free energy secondary structures of nrITS-1 and
nrITS2 were predicted with the program Mfold (Zuker, 2003) using
default parameters. Resulting nrITS-2 structures were compared to
conserved structures within green algae and ﬂowering plants (Mai
and Coleman, 1997). Nuclear ribosomal ITS regions were checked
for the occurrence of conserved sequence motifs: one motif within
nrITS-1(Liu and Schardl, 1994), six motifs in nrITS-2 (Hershkovitz
and Zimmer, 1996), the consensus sequence from (Harpke and Pet-
erson, 2006), three motifs in 5.8S RNA (Harpke and Peterson,
2008a; Jobes and Thien, 1997).
3.1. Phylogenetic reconstructions
Sequences of the trnK/matK region were 2374–2501 bp long,
those of the phyC gene were 815–938 bp long and the variability
of both markers was high within the ingroup (Table 1). Intraindi-
vidual sequence polymorphism among phyC copies obtained from
one species ranged from 2 to 25 bp (mean = 10 bp). In four taxa we
found apparently paralogous alleles, which differed by 56 bp in
Maihueniopsis hickenii (O-22), 26 bp in Cumulopuntia boliviana
ssp. ignescens,18bpinC. chichensis and 12 bp in Tunilla orurensis.
The phylogenies based on the trnK/matK region and the phyC
gene indicated the same major clades (Figs. 2 and 3) but relation-
ships between them were only supported by the combined analy-
sis (Fig. 4). The topology of these three trees differed largely to that
of the nrITS based tree (Fig. 5, see below). The ingroup was strongly
supported by the trnK/matK region and the combined data set (1.00
PP, 96% BS; 1.00 PP, 96% BS, respectively) and weakly supported by
phyC phylogeny (0.92 PP, 62% BS). The sister group relationship be-
tween Maihueniopsis and Pterocactus was only supported by the
analyses based on combined sequences (1.00 PP, 92% BS; Fig. 4).
Maihueniopsis clavarioides, which is the nomenclatural type of the
small genus Puna R.Kiesling, was sister to the remaining species
of Maihueniopsis in all phylogenies, except in the nrITS based one.
Three species formerly assigned to different genera were nested
within Tephrocactus: Tephrocactus verschaffeltii, formerly treated
as Austrocylindropuntia;T. recurvatus formerly referred to Cumulo-
puntia, and T. bonnieae formerly assigned to Puna.(Figs. 2–4).The
genera Austrocylindropuntia and Cumulopuntia including both sub-
species of C. subterranea (formerly assigned to Puna), and Punotia
lagopus (formerly treated as Austrocylindropuntia lagopus) were clo-
sely related (Figs. 2–4). In the phyC based tree these genera did not
form monophyletic groups because some phyC copies of Cumulo-
puntia boliviana ssp. ignescens and C. chichensis appeared as unsup-
ported sister group to the remaining species (Fig. 3). We also
sequenced two phyC copies of Maihueniopsis hickenii, which did
not cluster within the Maihueniopsis clade but with Austrocylindro-
puntia shaferi and A. vestita (O-22/4) and with Tephrocactus alexan-
deri (O-22/8). We also detected one phyC copy of Tunilla orurensis
(O-05/2) within the Maihueniopsis clade (Fig. 3).
In accordance with the tree based on combined nrITS and
trnL-trnF intergenic spacer sequences published by Grifﬁth and
Porter (2009) the phylogeny based on these nrITS sequences only
Sequence information for the different sequence data sets.
Alignment characteristics Genes
trnK/matK region phyC nrITS region Combined trnK/matK and phyC
No. of taxa 54 40 29 40
No. of sequences 62 155 41 41
Range of sequence lengths 2374–2501 815–938 573–589 –
Alignment length 2530 957 622 3487
Informative positions within the ingroup 122 116 42 171
Constant positions within the ingroup 2326 674 518 3160
Nucleotide substitution model
HKY + G (GTR + G) HKY + G GTR + G
Best-ﬁt model according to Akaike information Criterion as implemented in MrModeltest (Nylander, 2004).
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 671
Author's personal copy
Fig. 2. Bayesian phylogeny based on the trnK/matK region of the chloroplast DNA. Posterior probabilities (PP) higher than 0.9 and Maximum Likelihood Bootstrap percentages
(BS) higher than 70% are given above branches (PP/BS). The branch leading to Portulaca oleracea is not presented in the original scale of length. Seeds drawn after Stuppy
(2002) are presented right to each clade. Seed structures were observed in species marked with an asterisk (Stuppy, 2002). The left scheme represents a longitudinal section
of an opuntioid seed (principal structures are exempliﬁed by the seed structures of Cumulopuntia and Austrocylindropuntia). The right scheme illustrates a cross section
through the funicular envelope. Circles represent longitudinally elongated cells, lines represent crosswise elongated cells. Tissues marked with an asterisks are typical for
Tephrocactus but were not observed in T. nigrispinus,T. verschaffeltii and T. recurvatus (Stuppy, 2002; Gilmer and Thomas, 2000; Stuppy pers. comm.).
672 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
Fig. 3. Bayesian phylogeny based on exon 1 region of the phyC gene encoded by nuclear DNA. Posterior probabilities (PP) higher than 0.9 and Maximum Likelihood Bootstrap
percentages (BS) higher than 70% are given above branches (PP/BS). Statistical support for copies of the same species within each clade is not shown.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 673
Author's personal copy
Fig. 4. Bayesian phylogeny based on combined sequences of the trnK/matK region and phyC exon 1 region. Posterior probabilities (PP) higher than 0.9 and Maximum
Likelihood Bootstrap percentages (BS) higher than 70% are given below branches (PP/BS). Character states of seven morphological characters are present in boxes right of the
taxa. Estimates of ancestral character states for internal nodes of the tree are presented in pie charts. Characters marked with an asterisk are taken from Stuppy (2002).
Abbreviations of species names: C. bol. = Cumulopuntia boliviana; C. sub. = Cumulopuntia subterranea.
674 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
contained three major clades (A–C), that were not congruent with
the clades in Figs. 2–4, because Pereskiopsis was sister to Cumulo-
puntia and the genus Maihueniopsis was polyphyletic (Fig. 5).
3.2. Reconstruction of morphological traits
Assignment of the morphological traits of each species and
results from the reconstruction of ancestral character states are
presented in Fig. 4. Thus, according to our analyses, the ancestor
of the Tephrocacteae was probably a dwarf shrub (0.84 PP) with
determinate shoot segments (1.00 PP), caducous leaves (1.00 PP)
and ﬁbrous roots (0.98 PP). The ancestors of Tephrocacteae had
probably a curved embryo with large amounts of perisperm (0.68
PP), and the cells of the funicular envelope were orientated parallel
to the funicular girdle (0.69 PP; i.e. parallel to the funicular
vascular bundle that runs inside the girdle; see also sketches of
seed structures of Maihueniopsis and Pterocactus in Fig. 2). Recon-
structions also imply that tuberous roots originated independently
in the Maihueniopsis–Pterocactus clade, in Tephrocactus and in
3.3. Structural analysis of the nrITS region
Sequence lengths of nrITS-1 and nrITS-2 taken from Grifﬁth and
Porter (2009) were not substantially variable and ranged within
the ingroup from 188 to 191 bp and from 221 to 227 bp, respec-
tively. Sequence variability within the ingroup was low (app. 3%,
Table 1); but sequences of Tephrocactus molinensis and Pereskiopsis
aquosa deviated app. 7% from the other nrITS sequences. Mean and
standard deviation of the GC-content was 0.68 ± 0.01 for both
Fig. 5. Bayesian phylogeny based on nrITS region (nrITS-1, 2, 5.8SrDNA) re-analysed from the study of Grifﬁth and Porter (2009). Posteriori probabilities (PP) higher than 0.9
and Maximum Likelihood Bootstrap percentages (BS) higher than 70% are shown above branches (PP/BS). Genera were abbreviated by ﬁrst letter: A. = Austrocylindropuntia,
C. = Cumulopuntia,M.=Maihueniopsis,P.=Pterocactus,T.=Tephrocactus. Original voucher information for each sequence is presented in rectangular brackets. Sequences that
contained mismatches to conserved motifs within nrITS-2 and 5.8S rRNA were marked with an asterisk (⁄) or circle (°), respectively. Columns right of the phylogeny present
data from secondary structure predictions of nrITS-1 and nrITS-2. Within nrITS-1 structures we found two major types, one containing helix II and the other containing helix Ia
(Table S2,Fig. 6). Within nrITS-2 we also observed two main structures: one type consists of four helices and the second type of ﬁve helices (Table S2,Fig. 6). A question mark
is assigned to structures, which could not be conﬁned to one of these types.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 675
Author's personal copy
nrITS-1 and nrITS-2. The lowest GC-content was observed in Tep-
hrocactus molinensis (0.66 for nrITS-1 and nrITS-2;Table S2 in the
We found no deviation from the conserved nrITS-1 motif
GCCRY-(4-7n)-GYGYCAGGAA published by Liu and Schardl
(1994). We detected mismatches in four of six conserved motifs
within the nrITS-2 region (Harpke and Peterson, 2008b; Hershko-
vitz and Zimmer, 1996;Table S3 in the Supplementary material).
One of three conserved motifs within the 5.8S region differed from
published consensus sequences (Harpke and Peterson, 2008a;
Jobes and Thien, 1997,Table S3).
Secondary structure models resulted in mean free energy values
of 78.78 ± 3.52 for nrITS-1 and 109.47 ± 4.79 for nrITS-2
(Table S3). The highest values for nrITS-1 were observed in T. artic-
ulatus (papyracanthus) and T. articulatus (strobiliformis;
70.33 kcal/mol) and for nrITS-2 in T. molinensis (95.65 kcal/
mol). Within secondary structures of nrITS-1, we detected two ma-
jor types: one consisting of the helices I, II, III and the other of heli-
ces Ia, II, III, with the substructure Ia consisting of two smaller
helices (Table S3,Fig. 6). Sequences forming helix II are conﬁned
to clade A and to the genus Austrocylindropuntia in the nrITS based
phylogeny (Fig. 5). Secondary structures containing helix Ia were
found in clade C, and in the genera Cumulopuntia and Tephrocactus
(Fig. 5). The secondary structure of Tephrocactus molinensis was
very dissimilar to all other investigated structures and did not con-
tain one of the detected helices (data not shown). Nuclear rITS-2
sequences formed structures with four or ﬁve helices, which were
concordant with common angiosperm structures (Fig. 6,Mai and
Coleman 1997). Except for Brasiliopuntia brasiliensis and Maihueni-
opsis ovata 2, nrITS-2 sequences forming ﬁve helices were found in
clade A in the nrITS-based phylogeny (Fig. 5). Nuclear rITS-2 struc-
tures of Tephrocactus molinensis and Pereskiopsis aquosa could not
be conﬁned to one of the found nrITS-2 types (data not shown).
4.1. Relationships and evolution of morphological traits within
Our results and those of others studies clearly imply that the
Pterocactus and Maihueniopsis, now circumscribed as tribe Tephro-
cacteae (Hunt, 2011) represent a distinct lineage from the remain-
ing terete-stemmed opuntias (tribe Cylindropuntieae sensu Hunt
et al., 2006), and the mainly ﬂat-stemmed genera (tribe Opuntieae
sensu Hunt et al., 2006; Bárcenas et al., 2011; Edwards et al., 2005;
Hernández-Hernández et al., 2011; Wallace and Dickie, 2002).
Within the Tephrocacteae, we detected four strongly supported
clades representing the genera (1) Pterocactus, (2) Maihueniopsis,
(3) Tephrocactus and (4) Punotia,Cumulopuntia and Austroclindro-
puntia (Figs. 2–4).
Reconstruction of morphological character states implies that
the morphological traits that have been traditionally considered
as plesiomorphic in the Opuntioideae, are not ancestral within
the Tephrocacteae. Wallace and Dickie (2002) assumed that pre-
cursors of Opuntioideae shared many characters with Austrocylin-
dropuntia because they interpreted its indeterminate growth and
the persistent leaves as plesiomorphic analogous to the ancestral
position of the leafy trees of the genus Pereskia (Britton and Rose,
1919; Edwards et al., 2005). Our results suggest that Austrocylin-
dropuntia and Cumulopuntia are sister to Tephrocactus and the
ancestors of them probably developed determinate segments and
caducous leaves (Fig. 4). Within Punotia,Austrocylindropuntia and
Cumulopuntia a stepwise reduction of the size of leaf rudiments
is observed. Punotia and Austrocylindropuntia have conspicuous
leaves (reaching 7 cm or more in length in A. subulata), which per-
sist for at least one growing season whereas leaves of Cumulopuntia
are vestigial and soon caducous. The loss and regain of leaﬁness oc-
curred evidently several times independently in Opuntioideae (Ed-
wards and Donoghue, 2006; Grifﬁth, 2004). These ﬁndings are
supported by the observation that rudimentary leaves are present
at early developmental stages in many species of subfamily Cactoi-
deae (Mauseth, 2007).
Different life forms are very homoplastic within the Tephrocac-
teae (Fig. 4). Geophytes evolved independently in all major clades
(Fig. 4) from ancestors that were probably dwarf shrubs. The geo-
phytes Maihueniopsis clavarioides,Tephrocactus bonnieae and
Cumulopuntia subterranea are not closely related (Figs. 2–4) and
so cannot be grouped together as the separate genus Puna (Kie-
sling, 1982). Though Maihueniopsis and Cumulopuntia are not very
closely related, their morphological and ecological similarities are
striking; both form dwarf shrubs or dense cushions with determi-
nate stem-segments and mesotonic to sub-acrotonic branching
and they occur often in sympatry. Their seeds do, however, provide
useful diagnostic characters, both morphological and anatomical
(Fig. 2;Iliff, 2002; Stuppy, 2002). The so-called funicular envelope,
one of the features unique to the Opuntioideae, is of particular
interest. Effectively taking over the function of the seed coat (i.e.
mechanical protection), this aril-like structure encasing the seed
does not originate from the integument but from the funiculus,
which completely surrounds and covers the ovule already during
early development. Its central vascular bundle together with a
sheath of longitudinally orientated sclerenchymatic cells forms a
protruding ridge named funicular girdle (Stuppy, 2002). The tissue
of the funicular envelope subtending the funicular girdle usually
consists of transversely orientated ﬁbres but in Maihueniopsis,
these and all other sclerenchymatic cells of the funicular envelope
are orientated in parallel to the funicular girdle (Fig. 2;Stuppy,
The ancestral character state reconstructions within the Tep-
hrocacteae corroborate hypotheses on the evolution of embryo
shape and volume of perisperm within opuntioid seeds (Fig. 4;Stu-
ppy, 2002). Seeds of Maihueniopsis and Pterocactus contain a com-
parably large amount of perisperm tissue surrounded by a ring-
shaped embryo, which is also most likely the character state for
the ancestor of Tephrocacteae (Figs. 2 and 4). The enlargement of
the embryo at the expense of the perisperm derived from nucellar
tissue found in Austrocylindropuntia,Cumulopuntia and Tephrocac-
tus (Fig. 2) are probably advantageous because the storage tissue
is directly incorporated in the embryo thus saving energy for trans-
porting nutrients at the time of germination (Stuppy, 2002).
4.2. Relationships within major clades of Tephrocacteae
Contrary to the results of Grifﬁth and Porter (2009) and to the
nrITS based phylogeny based on sequences taken from the study
of these authors (Fig. 5), the genus Maihueniopsis appears to be a
strongly supported monophyletic group (Figs. 2–4). Possible rea-
sons for the incongruence of genetic markers are discussed in the
section ‘nuclear ITS evolution’.
The geophytic species Maihueniopsis clavarioides is sister to the
remaining species of Maihueniopsis and morphologically very sim-
ilar to Pterocactus, which is distributed in the Southern Andes in
Northern Argentina. In contrast to the wide distribution of the rest
of the Maihueniopsis clade, which is found from the eastern Andes
to the Atlantic coast (Fig. 1), M. clavarioides is also distributed in
the southern Andes in northern Argentina. Interestingly, the next
species which branches off after M. clavarioides is M. domeykoensis.
This species is found in the western Andes of northern Chile, and
thus we assume that the ancestors of Maihueniopsis and Pterocactus
676 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
originated in the Southern Andes, followed by an early dispersal of
Maihueniopsis to the west, and a further colonisation of Maihueni-
opsis over its entire range.
Specimens formerly treated as M. minuta (G. Charles, pers.
comm.) were not closely related with each other (O-11, O-12, O-
20 in Fig. 2). Thus, the plant material was re-evaluated and a
new species, M. glochidiata G.Charles, was described (Hunt,
2011). This taxon is part of a clade (0.98 PP) consisting of species
with very small segments (Fig. 2). The samples O-22, O-48, O-49
did not cluster together but were treated as M. darwinii in the
Fig. 6. Examples for secondary structure predictions of the nrITS region. Within nrITS-1 sequences, we observed two major types of secondary structures. The ﬁrst type
(represented by the nrITS-1 structure of Pterocactus tuberosus (decipiens) contains the helices I to III. The second type (represented by the nrITS-1 structure of Tephrocactus
articulatus consists of the helices Ia, II and III. The conserved nrITS-1 motif (Liu and Schardl, 1994) is presented in bold letters and marked with ‘‘m1’’. Within nrITS-2 sequences
we observed either structures consisting of ﬁve helices, e.g. Austrocylindropuntia subulata or of four helices, e.g. Pterocactus tuberosus (decipiens). All helices found in nrITS-2
are concordant to structures found in ﬂowering plants by Mai and Coleman (1997) Six conserved nrITS-2 motifs (Hershkovitz and Zimmer, 1996) are presented in bold letters
and marked with ‘‘m1-m6’’.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 677
Author's personal copy
collection of G. Charles and the Royal Botanical Gardens Kew,
respectively (Fig. 2). A close examination of the plants revealed
that none of them matched the description of the type of M. dar-
winii, thus the species M. hickenii and M. platyacantha were re-
established (Hunt, 2011).
Within Maihueniopsis hickenii and Tunilla orurenis we detected
phyC copies which did not cluster within the clade of the respective
genus (O-05/2; O-22/4, O-22/8 in Fig. 3). Present hybridisation
events seem to be an improbable reason for this peculiar pattern
because all above mentioned species are spatially isolated from
each other. A likely scenario for the presence of paralogous alleles
in these species is ancient hybridization between the ancestors of
the genera combined with a loss of alleles in some taxa because
nuclear ribosomal DNA data resulted also in a conﬂicting phylog-
eny (see section ‘nuclear rITS evolution’). Although the intraindi-
vidual sequence polymorphism among phyC clones was high in
some taxa (2–25 bp) we do not expect that the paralogous phyC
copies originated by Taq polymerase and cloning errors because se-
quences did not contain unprecedented mutations like premature
stop codons as similarly shown by Thornhill et al. (2007). Never-
theless, more detailed research including the determination of
the ploidy level of investigated plants, the exact copy number of
the phyC gene in the genome and more sequence information from
other nuclear markers are needed to disentangle possible explana-
tions for paralogy ranging from ancient hybridisation, gene dupli-
cation and incomplete lineage sorting.
The incorporation of Tephrocactus nigrispinus,T. verschaffeltii
and T. recurvatus in this genus, suggested by molecular data
(Figs. 2–4) and also assumed by Backeberg and Knuth (1936),
Backeberg (1963) and Nyffeler and Eggli (2010) for the latter two
species, complicates the morphological description of Tephrocactus
because the otherwise characteristic aerenchymatic funicular gir-
dle of the seeds is not developed in these species. This suggests
that these species are not dispersed by wind and water but proba-
bly by endozoochory (Fig. 2;Stuppy, 2002). However, species of
Tephrocactus are characterised by acrotonic branching resulting
in typical chains of segments, and their glochids are sunken into
cavities with small openings (‘encrypted glochids’; Stuppy, 2002).
The species with tiny joints, Tephrocactus bonnieae,T. recurvatus
and T. molinensis are closely related (Figs. 2–4). Tephrocactus
recurvatus was originally described by Backeberg as T. curvispinus
Backeb. (Backeberg, 1963), but this turned out to be an invalid
name because it was based on living material. Later Gilmer and
Thomas described this species validly and transferred it to Cumulo-
puntia because of similar seed characters (Gilmer and Thomas,
2000). However, they also mentioned the intermediate morphol-
ogy between Tephrocactus and Cumulopuntia and suspected a hy-
brid origin of the species (Gilmer and Thomas, 2000). Our results
do not support a hybrid origin, but only copy of the phyC gene
was sampled (Fig. 3). Interestingly, T. recurvatus was found to be
tetraploid but contained only two nucleolar organiser regions
(Las Peñas et al., 2009), thus it could be an allopolyploid species
which has lost one of the parental ribosomal loci.
Tephrocactus nigrispinus is sister to T. verschaffeltii (Figs. 2–4).
The latter species was treated as Austrocylindropuntia (Backeberg,
1966). Stuppy’s detailed analysis of seed anatomy already revealed
that seed structures of T. verschaffeltii are exceptional within
Austrocylindropuntia, but nevertheless he considered them to be
typical for Austrocylindropuntia because of their rudimentarily
developed funicular girdle (Stuppy, 2002). Tephrocactus nigrispinus
and T. verschaffeltii share a number of morphological characters
that are unusual for the genus. Both species have widely opened
orange to red ﬂowers, whereas the remaining species have mostly
white or pink ﬂowers. Both species are distributed in high altitudes
in the eastern Andes in contrast to the other Tephrocactus species
which are usually found at medium to low elevations in the east-
ern Andes (Fig. 1).
The close relationship between Tephrocactusaoracanthus,T. alex-
anderi and T. articulatus (Figs. 2–4) is supported by morphology;
these plants form apically branching shrubs up to 30 cm in height
consisting of rather large moniliform segments, which easily de-
tach from the plants.
Tephrocactus weberi is not included in any of the above men-
tioned clades (Figs. 2–4) and is characterised by a variety of ﬂower
colours including yellow ﬂowers which are not found in Tephrocac-
tus elsewhere (Eggli and Leuenberger, 1998).
4.2.3. Punotia and Austrocylindropuntia
Austrocylindropuntia as treated in Hunt et al. (2006) is not
monophyletic. Austrocylindropuntia lagopus, which has now been
referred as a monotypic genus Punotia (Hunt, 2011), is sister to
the remaining species of Austrocylindropuntia and Cumulopuntia
(Figs. 2–4). Austrocylindropuntia and Punotia are characterised by
indeterminate growth, persistent leaves and fruits containing pulp,
whereas Cumulopuntia develops determinate segments with soon
caducous leaf rudiments and pulp-free fruits. Thus, leaﬁness,
which was regained by the ancestors of the Austrocylindropuntia–
Cumulopuntia–Punotia clade, was again lost during the evolution
of Cumulopuntia (Fig. 4). Besides leaﬁness and indeterminate
growth, Austrocylindropuntia is well characterised by seed charac-
ters because the vascular bundle of the funiculus is not covered
by the tissue of the funicular girdle (‘naked vascular bundle’; Stu-
The close relationship of Austrocylindropuntia subulata and A.
pachypus (Figs. 2–4) was also assumed by Iliff (2002). Both species
have cylindrical upright stems but the columnar A. pachypus has
smaller leaves than the shrubby A. subulata. These two sister spe-
cies have a vicariant distribution: A. pachypus is distributed in
low altitudes of western Andes, whereas A. subulata is naturally
found at high elevations east of the Andean continental divide,
but is also cultivated in large parts of South America.
Austrocylindropuntia vestita, A. shaferi and A. ﬂoccosa are closely
related (Figs. 2 and 4). This is supported by morphology because
these species are characterised by hairy, cylindrical stems forming
shrubs, dwarf shrubs or cushions.
Phylogenetic reconstructions revealed two major clades within
Cumulopuntia: the C. boliviana clade and the C. sphaerica clade
(Figs. 2–4). Both clades are well differentiated by morphology
and distribution (Fig. 7). They mostly correspond to the informal
species groups deﬁned by Iliff (2002) and Nyffeler and Eggli
(2010) suggested treating them as separate genera. The species
of the C. boliviana clade form hemispheric cushions or dwarf shrubs
consisting of ovoid joints whose areoles are clustered towards the
apex and their seeds have lateral ridges on the funicular envelope
(Stuppy, 2002). The species of the C. sphaerica clade are character-
ised by non-tuberculate, easily detaching segments with evenly
distributed areoles and their seeds lack lateral ridges on the funic-
ular envelope (Stuppy, 2002). The species of the C. sphaerica clade
and C. corotilla, which is sister to the remaining species of the C.
boliviana clade, grow at low to relatively high elevations on the
western side of the Andes (0–3400 m, Fig. 7). This suggests the
genus Cumulopuntia originated in this region and the C. boliviana
clade spread to higher elevations (2500–4400 m) of the eastern
We sequenced phyC copies from C. chichensis and C. ignescens,
which were sister to each other but did not cluster within the C.
boliviana clade and appeared at the base of the Austrocylindropun-
tia–Cumulopuntia clade (Fig. 3). This indicates a common origin of
678 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
these alleles before speciation within the clade. However, as out-
lined above, more information on ploidy level of species and copy
number of the phyC gene in Opuntiodeae is needed to investigate
whether these results arose from ancient hybridisation combined
with a loss of alleles in related taxa or from incomplete lineage
4.3. Nuclear rITS evolution
The phylogeny based on a subset of nrITS sequences taken from
Grifﬁth and Porter (2009) is congruent to results of these authors
based on a combined data set of nrITS and trnL-trnF intergenic
spacer sequences but contradicts the topology of trees in Figs. 2–
4. The nrITS based phylogeny found Pereskiopsis within the clade
containing Austrocylindropuntia and Cumulopuntia (Fig. 5), whereas
our data assign this genus to the outgroup. The genus Maihueniop-
sis turned out to be monophyletic in Figs. 2–4, but nrITS sequences
of this genus were scattered within all major clades of the nrITS
based phylogeny. This conﬂict between phylogenies is caused
either by the inclusion of misidentiﬁed plant material as assumed
for the study of Grifﬁth and Porter (2009) by Nyffeler and Eggli
(2010) or by the inclusion of paralogous nrITS sequences in the
analysis. It is unlikely that misidentiﬁcation is responsible for all
conﬂicting positions because 12 samples from the Tephrocacteae
Fig. 7. Distribution area of the genus Cumulopuntia. Localities known to the authors are presented as dots, localities of specimens analysed during this study are presented as
triangles. Species of the C. sphaerica clade are found at low to higher elevations at the western side of the Andes, whereas species of the C. boliviana clade are mainly
distributed at higher elevations at the eastern side of the Andes.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 679
Author's personal copy
appeared in the deep lineages of the Opuntioideae and 20 samples
in a derived clade of terete-stemmed Opuntiodeae (Grifﬁth and
Porter, 2009). To investigate whether this topology resulted from
locus-speciﬁc properties of the ribosomal rRNA region we analysed
sequence features of the nrITS region. We did not observe large dif-
ferences between GC-content, minimum free energy of secondary
structures and found no deviations from the nrITS-1 conserved mo-
tif and only few deviations from nrITS-2 conserved motifs (Tables
S2 and S3). Among three conserved motifs within the 5.8S rDNA
sequences, which are suited to test for the presence of pseudo-
genes (Harpke and Peterson, 2008b), we found one deletion within
motif 2 in two samples (A- instead of AA, Table S3), which might be
prone to misinterpretation reading the sequencing ﬁle. These re-
sults imply that the nrITS data set did not comprise a large propor-
tion of highly degraded pseudogenes as found in other Cactaceae,
e.g. Mammillaria,Lophocereus (Harpke and Peterson, 2006, 2007;
Hartmann et al., 2001, 2002). However, we suspect that the nrITS
sequences of Tephrocactus molinensis and Pereskiopsis aquosa are
pseudogenes because they differed approximately 7% from the
other ingroup sequences, which is twice as high as the average se-
quence divergence in the ingroup. Their considerably lower GC-
contents and the substantially deviating secondary structures at
higher minimum free energy values compared to the other sam-
ples (Table S2) also indicate pseudogenization (Mayol and Rossello,
2001). These putative pseudogenes themselves are not found at
unexpected positions in the nrITS phylogeny (Fig. 5). Thus homo-
plasy due to pseudogenization is probably not responsible for
incongruency of phylogenies.
The secondary structure predictions of nrITS-1 and nrITS-2 se-
quences revealed two major types in each region. Nuclear ribo-
somal ITS-1 sequences formed both helix II and III, and helix II
consisted of a conserved motif in angiosperms (Liu and Schardl,
1994;Fig. 6). The structures differed in the assembly of helix
one, which forms two smaller stem structures (Ia) in approxi-
mately half of the sequences (Fig. 6). Nuclear ribosomal ITS-2 se-
quences were either assembled to structures consisting of four or
ﬁve helices, which were found to be conserved in angiosperms
(Mai and Coleman, 1997). Interestingly, the structure types of both
spacer sequences did not vary within major clades of the phylog-
eny (Fig. 5), thus secondary structure predictions contained an
obvious phylogenetic signal as was also shown by other studies
(Grajales et al., 2007; Keller et al., 2010, 2008; Young and Coleman,
2004). This consistency and the incongruence of the tree to other
phylogenies (Figs. 2–4) implies that the nrITS phylogeny might rep-
resent a gene tree, which is not inevitably identical to the species
phylogeny because it is based on a mixture of orthologous and
paralogous nrITS sequences. Paralogous sequences emerge by gene
or genome duplication, the latter is often connected with hybrid-
ization (homeologous genes in allopolyploids). Recent studies on
chromosome numbers revealed that genome duplication occurred
frequently within the evolution of cacti and especially within
Opuntioideae (Arakaki et al., 2007; Las Peñas et al., 2009; Pinkava,
2002). Moreover, Las Peñas et al. (2009) mapped the physical loca-
tion of the nucleolar organiser regions (NOR), which include the
nrITS region, in Cactaceae and observed one NOR-bearing chromo-
some pair in diploid species. Interestingly, different results were
yielded from tetraploid species: tetraploid samples of Maihuenia
poeppigii Speg. contained four NOR-bearing chromosomes as ex-
pected, but in tetraploid samples of Tephrocactus recurvatus only
two NOR regions were found implying that two NOR loci were lost
after duplication. These results strongly support the existence of
paralogous nrITS loci within Opuntioideae. More detailed Fluores-
cent in situ Hybridization (FISH) and cloning experiments are
needed to understand the fates of duplicates ribosomal DNA loci
in these cacti. The investigation of intra-individual nrITS polymor-
phisms can elucidate whether paralogous sequences have been
maintained (Ritz et al., 2005; Sang et al., 1995), were homogenised
to one parental sequence type in certain lineages (Kovar
ˇik et al.,
2005; Wendel et al., 1995) or were recombined to new nrITS types
ˇik et al., 2004). It is generally difﬁcult to assess whether
paralogy is caused by incomplete lineage sorting (ancient sequence
polymorphism predating speciation) or whether it is caused by
hybridization followed by maintenance or loss of homeologs. We
assume that the situation observed here probably emerged from
ancient hybridization events because the presence of multiple
and not closely related alleles of the independent nuclear marker
phyC in e.g. Maihueniopsis and Cumulopuntia (Fig. 3) and the ob-
served polyploidy in the group (Las Peñas et al., 2009) point also
to genome whole doubling coming along with hybridization. How-
ever, the phylogeny based on nuclear low copy gene phyC was lar-
gely congruent to the chloroplast phylogeny, and it was much
better suited to detect conﬂicting phylogenetic signals because
paralogous sequences were simultaneously sampled, which is
accordance to previous suggestions on the use of molecular mark-
ers in plant phylogeny (Alvarez and Wendel, 2003; Small et al.,
We thank British members of the International Organization of
Succulent Plant Study (IOS) for ﬁnancial support of this study. We
thank L. Csiba (Jodrell Laboratory, Royal Botanic Gardens Kew, UK)
for supplying us DNA samples from their collection and F. Katter-
mann (Wantage, USA) for kindly providing locality data for the
maps. We thank H. Krufzik, J. Föller and J. Spies (Justus-Liebig-Uni-
versity Gießen, Germany) for excellent technical support and V.
Wissemann (Justus-Liebig-University Gießen, Germany) for many
helpful comments during the study and the possibility to work in
his lab. We thank U. Eggli (Sukkulentensammlung, Zürich, Switzer-
land) and R. Nyffeler (University Zürich, Switzerland) for fruitful
discussions during planning this study, N. Korotkova (Freie Univer-
sität Berlin, Germany), D. Harpke (Leibniz Institute of Plant Genet-
ics and Crop Plant Research, Gatersleben, Germany), K. Wesche
(Senckenberg Museum of Natural History Görlitz, Germany) and
the two anonymous reviewers for very helpful comments on the
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.ympev.2012.07.
Alvarez, I., Wendel, J.F., 2003. Ribosomal ITS sequences and plant phylogenetic
inference. Mol. Phylogenet. Evol. 29, 417–434.
Anderson, E.F., 2001. The Cactus Family. Timber Press, Portland, Oregon.
Anderson, E.F., 2011. Das große Kakteenlexikon, second ed. Eugen Ulmer, Suttgart,
Arakaki, M., Soltis, D.E., Speranza, P., 2007. New chromosome counts and evidence
of polyploidy in Haageocereus and related genera in tribe Trichocereeae and
other tribes of Cactaceae. Brittonia 59, 290–297.
Backeberg, C., 1958–62. Die Cactaceae. Gustav Fischer Verlag, Jena.
Backeberg, C., 1963. Descriptiones Cactearum Novarum.
Backeberg, C., 1966. Das Kakteenlexikon. Gustav Fischer Verlag, Jena.
Backeberg, C., Knuth, F.M., 1936. Kaktus-ABC. En haandbog for fagfolk og amatorer.
Nordisk Forlag, Copenhagen.
Bailey, C.D., Carr, T.G., Harris, S.A., Hughes, C.E., 2003. Characterization of
angiosperm nrDNA polymorphism, paralogy, and pseudogenes. Mol.
Phylogenet. Evol. 29, 435–455.
Baldwin, B.G., Sanderson, M.J., Porter, M.J., Wojciechowski, M.F., Campbell, C.S.,
Donoghue, M.J., 1995. The ITS region of nuclear ribosomal DNA: a valuable
source of evidence on angiosperm phylogeny. Ann. Missouri Bot. Gard. 82, 247–
Bárcenas, R.T., Yesson, C., Hawkins, J.A., 2011. Molecular systematics of the
Cactaceae. Cladistics 27, 1–20.
680 C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681
Author's personal copy
Barthlott, W., Hunt, D.R., 1993. Cactaceae. In: Kubitzki, K., Rohwer, J.G., Bittrich, V.
(Eds.), The Families and Genera of Vascular Plants. Springer Verlag, Berlin,
Germany, pp. 161–197.
Britton, N.L., Rose, J.N., 1919. The Cactaceae. Carnegie Institution, Washington DC.
Bruyns, P.V., Mapaya, R.J., Hedderson, T., 2006. A new subgeneric classiﬁcation for
Euphorbia (Euphorbiaceae) in southern Africa based on ITS and psbA-trnH
sequence data. Taxon 55, 397–420.
Buckler, E.S., Holtsford, T.P., 1996. Zea ribosomal repeat evolution and substitution
patterns. Mol. Biol. Evol. 13, 623–632.
Doweld, A.B., 1999. Tribal taxonomy of Pereskioideae and Opuntioideae
(Cactaceae). Sukkulenty 1, 25–26.
Edwards, E.J., Donoghue, M.J., 2006. Pereskia and the origin of the cactus life-form.
Am. Nat. 167, 777–793.
Edwards, E.J., Nyffeler, R., Donoghue, M.J., 2005. Basal cactus phylogeny:
implications of Pereskia (Cactaceae) paraphyly for the transition to the cactus
life form. Am. J. Bot. 92, 1177–1188.
Eickbush, T.H., Eickbush, D.G., 2007. Finely orchestrated movements: evolution of
ribosomal RNA genes. Genetics 175, 477–485.
Eggli, U., Leuenberger, B., 1998. On colour forms of Opuntia weberi (Cactaceae) with
notes on the typiﬁcation of the name. Willdenowia 28, 175–180.
Endler, J., Buxbaum, F., 1974. Die Pﬂanzenfamilie der Kakteen, third ed. Albrecht
Philler Verlag, Minden, Germany.
Farris, J.S., Kallersjo, M., Kluge, A.G., Bult, C., 1995. Constructing a signiﬁcance test
for incongruence. Syst. Biol. 44, 570–572.
Gibson, A.C., Nobel, P.S., 1986. The Cactus Primer. Harvard University Press,
Cambridge, Massachusetts, London, England.
Gilmer, K., Thomas, H.-P., 2000. Alte Bekannte, neu beschrieben: Cumulopuntia
recurvata. Kakteen und andere Sukkulenten 52, 85–92.
Grajales, A., Aguilar, C., Sanchez, J.A., 2007. Phylogenetic reconstruction using
secondary structures of internal transcribed spacer 2 (ITS2, rDNA): ﬁnding the
molecularand morphological gap in Caribbean gorgoniancorals. BMC Evol. Biol. 7.
Grifﬁth, M.P., 2002. Phylogenetic relationships in the Opuntioideae (Cactaceae)
based on nrITS sequences. International Organization for Succulent Plant Study
Bulletin 10, 15–16.
Grifﬁth, M.P., 2004. What did the ﬁrst cactus look like? An attempt to reconcile the
morphological and molecular evidence. Taxon 53, 493–499.
Grifﬁth, M.P., Porter, J.M., 2009. Phylogeny of Opuntioideae (Cactaceae). Int. J. Plant
Sci. 170, 107–116.
Harpke, D., Peterson, A., 2006. Non-concerted ITS evolution in Mammilaria
(Cactaceae). Mol. Phylogenet. Evol. 41, 579–593.
Harpke, D., Peterson, A., 2007. Quantitative PCR revealed a minority of its copies to
be functional in Mammillaria (Cactaceae). Int. J. Plant Sci. 168, 1157–1160.
Harpke, D., Peterson, A., 2008a. 5.8S motifs for the identiﬁcation of pseudogenic ITS
regions. Botany 86, 300–305.
Harpke, D., Peterson, A., 2008b. Extensive 5.8S nrDNA polymorphism in
Mammillaria (Cactaceae) with special reference to the identiﬁcation of
pseudogenic internal transcribed spacer regions. J. Plant Res. 121, 261–270.
Hartl, W.P., Klapper, H., Barbier, B., Ensikat, H.J., Dronskowski, R., Müller, P.,
Ostendorp, G., Tye, A., Bauer, R., Barthlott, W., 2007. Diversity of calcium oxalate
crystals in Cactaceae. Can. J. Bot. 85, 501–517.
Hartmann, S., Nason, J.D., Bhattacharya, D., 2001. Extensive ribosomal DNA genic
variation in the columnar cactus Lophocereus. J. Mol. Evol. 53, 124–134.
Hartmann, S., Nason, J.D., Bhattacharya, D., 2002. Phylogenetic origins of
Lophocereus (Cactaceae) and the senita cactus-senita moth pollination
mutualism. Am. J. Bot. 89, 1085–1092.
Helsen, P., Browne, R., Anderson, D., Verdyck, P., van Dongen, S., 2009. Galapagos
Opuntia (prickly pear) cacti; extensive morphological diversity, low genetic
variability. Biol. J. Linn. Soc. 96, 451–461.
Hernández-Hernández, T., Hernández, H.M., De-Nova, J.A., Puente, R., Eguiarte, L.E.,
Magallón, S., 2011. Phylogenetic relationships and evolution of growth form in
Cactaceae (Caryophyllales, Eucotyledoneae). Am. J. Bot. 98, 44–61.
Hershkovitz, M.A., Zimmer, E.A., 1996. Conservation patterns in angiosperm rDNA
ITS2 sequences. Nucleic Acids Res. 24, 2857–2867.
Huelsenbeck, J.P., Ronquist, F., 2001. MRBAYES: Bayesian inference of phylogeny.
Bioinformatics 17, 754–755.
Hunt, D., Taylor, N., Charles, G., 2006. The new cactus lexicon. Dh Books, Milborn
Hunt, D., 2011. Classiﬁcation of the ‘cylindroid’ opuntias of South America.
Cactaceae Systematics Initiatives 25, 5–29.
Iliff, J., 2002. The Andean opuntias: an annotated checklist of the indigenous non-
platyopuntioid opuntias of South America. Succ. Plant Res. 6, 133–244.
Jobes, D.V., Thien, L.B., 1997. A conserved motif in the 5.8S ribosomal RNA (rRNA)
gene is a useful diagnostic marker for plant internal transcribed spacer (ITS)
sequences. Plant Mol. Biol. Rep. 15, 326–334.
Johnson, A.J., Soltis, D.E., 1994. MatK DNA sequences and phylogenetic
reconstruction in Saxifragaceae s.str. Syst. Bot. 19, 143–156.
Keller, A., Foerster, F., Mueller, T., Dandekar, T., Schultz, J., Wolf, M., 2010. Including
RNA secondary structures improves accuracy and robustness in reconstruction
of phylogenetic trees. Biology Direct 5.
Keller, A., Schleicher, T., Foerster, F., Ruderisch, B., Dandekar, T., Mueller, T., Wolf, M.,
2008. ITS2 data corroborate a monophyletic chlorophycean DO-group
(Sphaeropleales). BMC Evol. Biol. 8.
Kiesling, R., 1982. Puna, un género nuevo de Opuntiodeae (Cactaceae). Hickenia 55,
ˇik, A., Matyášek, R., Lim, K.Y., Skalická, K., Koulaková, B., Knapp, S., Chase,
M.W., Leitch, A.R., 2004. Concerted evolution of 18–5.8–26S rDNA repeats in
Nicotiana allotetraploids. Biol. J. Linn. Soc. 82, 615–625.
ˇik, A., Pires, J.C., Leitch, A.R., Lim, K.Y., Sherwood, A.M., Matyášek, R., Rocca, J.,
Soltis, D.E., Soltis, P.S., 2005. Rapid concerted evolution of nuclear ribosomal
DNA in two Tragopogon allopolyploids of recent and recurrent origin. Genetics
Las Peñas, M.L., Urdampilleta, J.D., Bernardello, G., Forni-Martins, E.R., 2009.
Karyotypes, heterochromatin, and physical mapping of 18S–26S rDNA in
Cactaceae. Cytogenet. Genome Res. 124, 72–80.
Leuenberger, B.E., 1976. Die Pollenmorphologie der Cactaceae. Dissertationes
Botanicae 31, 1–321.
Liu, J.S., Schardl, C.L., 1994. A conserved sequence in internal transcribed spacer-1 of
plant nuclear ribosomal-RNA genes. Plant Mol. Biol. 26, 775–778.
Maddison, W.P., Maddison, D.R., 2006. StochChar: A package of Mesquite modules
for stochastic models of character evolution.
Maddison, W.P., Maddison, D.R., 2010. Mesquite: A modular system for
evolutionary analysis. http://mesquiteproject.org.
Mai, J.C., Coleman, A.W., 1997. The internal transcribed spacer 2 exhibits a common
secondary structure in green algae and ﬂowering plants. J. Mol. Evol. 44, 258–
Martins, L., 2006. Systematics and biogeography of Klasea (Asteraceae-Cardueae)
and a synopsis of the genus. Bot. J. Linn. Soc. 152, 435–464.
Mauseth, J.D., 2007. Tiny but complex foliage leaves occur in many ‘‘leaﬂess’’ cacti
(Cactaeceae). Int. J. Plant Sci. 168, 845–852.
Mayol, M., Rossello, J.A., 2001. Why nuclear ribosomal DNA spacers (ITS) tell
different stories in Quercus. Mol. Phylogenet. Evol. 19, 167–176.
Nyffeler, R., 2002. Phylogenetic relationships in the cactus family (Cactaceae) based
on evidence from trnK/matK and trnL-trnF sequences. Am. J. Bot. 89, 312–326.
Nyffeler, R., Eggli, U., 2010. A farewell to dated ideas and concepts – molecular
phylogenetics and a revised suprageneric classiﬁcation of the family Cactaceae.
Schumannia 3, 109–149.
Nylander, J.A.A., 2004. MrModeltest. Evolutionary Biology Centre, Uppsala
University, Uppsala, Program distributed by the author.
Pinkava, D.J., 2002. On the evolution of the continental North American
Opuntioideae (Cactaceae). Succ. Plant Res. 6, 59–98.
Ritz, C.M., Schmuths, H., Wissemann, V., 2005. Evolution by reticulation: European
dogroses originated by multiple hybridization across the genus Rosa. J. Hered.
Robinson, H., 1974. Scanning electron microscope studies in the spines and glochids
of Opuntiodeae (Cactaceae). Am. J. Bot. 61, 278–283.
Rowley, G.D., 1958. Reunion of the genus Opuntia Mill. Natl. Cact. Succ. J. 13, 3–6.
Sang, T., Crawford, D.J., Stuessy, T.F., 1995. Documentation of reticulate evolution in
peonies (Paeonia) using internal transcribed spacer sequences of nuclear
ribosomal DNA – implications for biogeography and concerted evolution.
Proc. Natl. Acad. Sci. USA 92, 6813–6817.
Schumann, K., 1897–1898. Gesamtbeschreibung der Cactaceae im Verhältnis zu
ihrer systematischen Gliederung. Verlag J. Neumann, Neudamm.
Small, R.L., Cronn, R.C., Wendel, J.F., 2004. Use of nuclear genes for phylogeny
reconstruction in plants. Aust. Syst. Bot. 17, 145–170.
Stamatakis, A., 2006. RaxML-VI-HPC: Maximum Likelihood-based phylogenetic
analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688–
Stuppy, W., 2002. Seed characters and the classiﬁcation of the Opuntioideae. Succ.
Plant Res. 6, 25–58.
Swofford, D.L., 2002. Paup⁄4.0. Phylogenetic analyses using parsimony (and other
methods). Sinauer Associates, Sunderland, Massachusetts.
Thornhill, D.J., Lajeunesse, T.C., Santos, S.R., 2007. Measuring rDNA diversity in
eucaryotic microbial systems: how intragenomic variation, pseudogenes, and
PCR artifacts confound biodiversity estimates. Molec. Ecol. 16 (5326), 5340.
Von Hagen, K.B., Kadereit, J.W., 2001. The phylogeny of Gentianella (Gentianaceae)
and its colonization of the southern hemisphere as revealed by nuclear and
chloroplast DNA sequence variation. Org. Divers. Evol. 1, 61–79.
Wallace, R.S., Dickie, S.L., 2002. Systematic implication of chloroplast DNA sequence
variation in subfam. Opuntioideae (Cactaceae). Succ. Plant Res. 6, 9–24.
Wendel, J.F., Schnabel, A., Seelanan, T., 1995. Bidirectional interlocus concerted
evolution following allopolyploid speciation in cotton (Gossypium). Proc. Natl.
Acad. Sci. USA 92, 280–284.
Young, I., Coleman, A.W., 2004. The advantages of the ITS2 region of the nuclear
rDNA cistron for analysis of phylogenetic relationships of insects: a Drosophila
example. Mol. Phylogenet. Evol. 30, 236–242.
Zuker, M., 2003. Mfold web server for nucleic acid folding and hybridization
prediction. Nucleic Acids Res. 31, 3406–3415.
C.M. Ritz et al. / Molecular Phylogenetics and Evolution 65 (2012) 668–681 681