A multilocus molecular phylogeny of boxfishes (Aracanidae, Ostraciidae; Tetraodontiformes)

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: http://www.elsevier.com/copyright

Author's personal copy

Molecular Phylogenetics and Evolution 66 (2013) 153–160

Contents lists available at SciVerse ScienceDirect

Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev

A multilocus molecular phylogeny of boxfishes (Aracanidae, Ostraciidae; Tetraodontiformes) Francesco Santini a,⇑, Laurie Sorenson a, Tina Marcroft a, Alex Dornburg b, Michael E. Alfaro a,⇑ a b

University of California Los Angeles, Department of Ecology and Evolutionary Biology, 610 Young Drive South, Los Angeles, CA 90095, USA Department of Ecology and Evolutionary Biology, Yale University, New Haven, CT 06520, USA

a r t i c l e

i n f o

Article history: Received 29 July 2012 Revised 10 September 2012 Accepted 17 September 2012 Available online 2 October 2012 Keywords: Aracanidae Ostraciidae Tetraodontiformes Phylogeny Fossils Molecular clock

a b s t r a c t Boxfishes (superfamily Ostracioidea, order Tetraodontiformes) are comprised of 37 species within the families Aracanidae (13 sp.) and Ostracidae (24 sp.). These species are characterized by several dramatic reductive trends in their axial and appendicular skeleton, and by the presence of a carapace formed by enlarged and thickened scale plates. While strong support exists for the monophyly of both families, interspecific relationships remain unclear as no species-level molecular phylogeny currently exists for either of these two clades, and the only hypotheses of relationships are based on morphological studies that were mostly restricted to generic-level relationships. Here we present the results of a new phylogenetic study of a dataset composed of 9 loci for 26 species of boxfishes using both likelihood and Bayesian methods. Our topology strongly supports the monophyly of both groups, and additionally provides strongly supported resolution for the vast majority of species-level interrelationships. Based on this new phylogeny, we suggest changing the taxonomic status of the species Lactoria fornasini to Tetrasomus fornasini, and Rhynchostracion nasus to Ostracion nasus. Using a Bayesian approach to divergence time estimation we inferred a Paleocene origin of the Ostracioidea, with an estimated origin of the reefassociated ostraciids spanning the Eocene and Oligocene, and a Miocene/Pliocene origin of the aracanids. Ó 2012 Elsevier Inc. All rights reserved.

1. Introduction Ostracioids, from here on also referred to as boxfishes, represent one of the most distinct groups of teleost fishes, and are largely characterized by extreme body armor modifications. Body armor has originated multiple times within actinopterygian fishes, with independent origins in lineages as divergent as Siluriformes, Gasterosteiformes, Syngnathiformes, Scorpaeniformes (Nelson, 2006) and fossil tetraodontiforms that include the Cretaceous Plectocretacicoidei, the Eocene Spinacanthus and Protobalistum from the Ypresian of Monte Bolca (Italy), and likely Eospinus from the Eocene of Caucasus (Tyler 1973; Tyler and Gregorova, 1991; Tyler and Bannikov 1992; Tyler and Sorbini, 1996; Tyler and Santini, 2002, 2005; Santini and Tyler, 2003, 2004). Yet, despite the multiple origins of body armor across actinopterygians, the degree of carapace shape disparity demonstrated in boxfishes reflects a complex evolutionary history of armor restructuring and modification not common in other armored fish lineages. The armored carapace in boxfishes is composed of thickened and enlarged scale plates that abut one another (Tyler, 1980;

⇑ Corresponding authors. Fax: +1 310 206 3987. E-mail addresses: [email protected] (F. Santini), michaelalfaro@ ucla.edu (M.E. Alfaro). 1055-7903/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ympev.2012.09.022

Besseau and Bouligand, 1998), and the evolution of the structure has been facilitated by reductions in both the axial and post-cranial skeleton. Boxfishes lack the first spiny dorsal fin, including all associated pterygiophores, the pelvic fin complex, and pleural ribs (Tyler, 1980). Additionally boxfishes have experienced fusion of several abdominal vertebrae to each other and to the occipital region of the skull, as well as the fusion of several elements of the caudal plate to each other (Tyler, 1962, 1980; Klassen, 1995, 1996; Santini and Tyler, 2003; Britz and Johnson, 2005). The boxfish carapace has long been thought to represent an anti-predatory defense mechanism (Brainerd and Patek, 1998), and the shape of the carapace varies widely. Boxfishes range from having nearly square transverse sections (e.g., Ostracion), to more triangular (e.g., Tetrasomus) or oblong (e.g., Anoplocapros) shapes. In addition to shape, the surface structure of the carapace is also very variable, with species of boxfishes having either a relatively smooth, uniform surface (e.g., Ostracion, Anoplocapros), or a surface covered with protruding irregularities such as keels, spikes and spines (e.g., Lactoria, Kentrocapros). A number of recent studies have shown that the carapace shape and structure likely plays a role in swimming by minimizing vortices and drag, as boxfishes have evolved a peculiar style of swimming (ostraciiform locomotion) to offset the restrictions to locomotor function imposed by the presence of rigid full body armor (Hove et al., 2001; Bartol et al., 2005; Gordon et al., 2001).

Author's personal copy

154

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

In spite of their peculiar morphology, boxfishes have received relatively scant attention from fish systematists. Based primarily on carapace armoring, boxfishes are currently classified into two families, the Aracanidae (deepwater boxfishes), and the Ostraciidae (boxfishes, cowfishes and trunkfishes) (Tyler, 1980; Santini and Tyler, 2003; Froese and Pauly, 2011). However, only a handful of morphological studies have attempted to resolve the relationships within these families. Tyler (1980) investigated the osteology of representative species within both families, and suggested a hypothesis of relationships among the then valid genera based on an evolutionary taxonomic approach in which the extinct genus Proaracana and the extant genus Strophiurichthys (that was later folded within the genus Anoplocapros), represented the ancestors to all other extant aracanids, while the Eocene Eolactoria represented the ancestor to all ostraciids, which were then divided ino two groups: one including Acanthostracion, Rhinesomus and Lactophrys, and the other including Lactoria, Tetrasomus, Ostracion and Rhynchostracion. Subsequently, Winterbottom and Tyler (1983) investigated the relationships among the genera of aracanids with a cladistic analysis (based on Hennigian argumentation) of a combined myological and osteological dataset (including osteological data from Tyler, 1980), and concluded that the then recently described Polyplacapros together with Kentrocapros formed the sister group to all other aracanids, while the monotypic Capropygia and Caprichthys appeared to be nested deeply within a clade formed by the remaining genera, Aracana and Anoplocapros. These results agreed with Tyler (1980) in the placement of Capropygia and Caprichthys appearing deeply nested within the aracanids, but disagreed in the placement of Anoplocapros. Klassen (1995) analyzed a large osteological dataset for 19 species of ostraciid species, and found a clade formed by the genera Acanthostracion and Lactophrys, both characterized by a triangular transverse plane shape, representing the sister group to all other ostraciids. A second clade, formed by Ostracion and Rhyncostracion, appears as sister to the Lactoria + Tetrosomus clade. Although the monophyly of all genera, except the paraphyletic Lactoria, was highly supported, the relationships within genera appeared to be rather uncertain, with several potential topologies being equally parsimonious (e.g., his Fig. 32). In this study we present the first molecular phylogeny of the boxfishes providing species level resolution for close to 80% of extant boxfish species. We employ two mitochondrial genes and seven single copy nuclear genes in a suite of Maximum Likelihood and Bayesian analyses and also a framework that allows speciestree inference to accommodate the potential for gene tree heterogeneity. Based on our results we reclassify the taxonomic affinities of two species of boxfish, formally renaming Lactoria fornasini into Tetrasomus fornasini, and Rhynchostracion nasus into Ostracion nasus. Using a Bayesian approach to divergence time estimation, we present the first species level timescale of boxfish evolution, providing the critical foundation for subsequent investigations into the evolutionary history of this charismatic group of marine fishes.

ships, we included both a balistid (Rhinecanthus aculeatus) and a triacanthodid (Triacanthodes ethiops) as outgroups. Higher-level relationships among the major tetraodontiform groups are currently highly contentious, and characterized by major disagreement between datasets. Cladistic analyses of osteological and myological data from adult specimens (Winterbottom, 1974; Santini and Tyler, 2003, 2004) agree with evolutionary taxonomic studies (Tyler, 1980) in strongly supporting a sister group relationships between ostracioids and a clade formed by Balistidae (triggerfishes) and Monacanthidae (filefishes), while larval characters (Leis, 1984; Britz and Johnson, 2005) support a close relationships between Ostraciidae, Molidae (ocean sunfishes) and Diodontidae (porcupine fishes). Molecular analyses, however, have so far failed to support either of these topologies. Analyses of the nuclear locus Rag1 and the ribosomal genes 12S and 16S (Holcroft, 2005; Alfaro et al., 2007) support a clade with ostracioids and Molidae + Triodontidae, while analyses of mitochondrial genomes support a clade with boxfishes, Triacanthodidae (deep sea spikefishes) and Triodontidae (threetooth puffer) (Yamanoue et al., 2008). 2.2. DNA extraction, PCR amplification, and sequencing DNA was extracted from muscle tissue samples or fin clips previously stored in 70% ethanol using the Qiagen DNAeasy kit (Qiagen, Valencia, CA, USA), following the protocol suggested by the manufacturer. Two mitochondrial genes, cytochrome oxidase subunit I (cox1) and cytochrome b (Cytb), and seven nuclear genes, early growth response gene 1 (EGR1), interphotoreceptor retinoidbinding protein (IRBP); mixed-linked Leukemia-like gene (MLL); cardiac muscle myosin heavy chain 6 alpha (myh6), recombination activating gene 1 (Rag1), rhodopsin (Rh), and zic family member 1 (zic1) were amplified using the polymerase chain reaction (PCR). One to two microliters of genomic template was used per 25-lL reaction, containing 5 lL of 5  Go-Taq Flexi PCR buffer (Promega), 2 lL MgCl 2 (25 mM), 0.5 lL dNTPs (2.5 lM), 1.25 lL of each primer (10 lM) (Table 2), and 0.125 lL of Promega GoTaq Flexi DNA polymerase (5 U/L). Primers and PCR conditions were obtained from the literature: (Ward et al., 2005) for cox1; (Sevilla et al., 2007) for Cytb; (Chen et al., 2008) for EGR1; (Li et al., 2009) for IRBP and MLL; (Li et al., 2007) for myh6 and zic1; (López et al., 2004) for Rag1; (Chen et al., 2003) for Rh. PCRs were performed on a MJ Research PTC-200 Peltier or Eppendorf Mastercycler ProS thermal cyclers. All products were stored at 20 °C after amplification. We used ExoSap (Amersham Biosciences) to remove the excess dNTPs and unincorporated primers from the PCR products; purified products were then cycle-sequenced using the BigDye Terminator v.3.1 cycle sequencing kit (1/8th reaction) (Applied BioScience) with each gene’s original or additional internal primers (Table 2) used for amplification. The cycle sequencing protocol consisted of 25 cycles with a 10-s 94 °C denaturation, 5-s of 50 °C annealing, and a 4-min 60 °C extension stage. Sequencing was conducted at the Yale University DNA Analysis Facility using an ABI 3730xl DNA Genetic Analyzer (Applied Biosystems).

2. Material and methods 2.3. Phylogenetic analysis 2.1. Sampling Tissue samples for 26 species of boxfishes, plus two outgroups, were obtained through tissue loans from other university or museum collections or purchases through the pet trade (Table 1). Additional sequence data for two species was downloaded from Genbank (Table 1). Our sampling includes 17 ostraciids and 9 aracanids, including representatives from every described genus in the superfamily, with the exception of the aracanid Polyplacapros. Due to uncertainty in higher-level tetraodontiform relation-

The chromatograms were checked and assembled into contigs using Geneious 5.3 (Drummond et al., 2010). The consensus sequences for each individual gene were aligned in Geneious using the MUSCLE software (Edgar, 2004), and the alignments subsequently checked by eye for accuracy. The sequences were trimmed to minimize missing characters, and our final data matrix consisted of 680 bp for cox1, 1088 bp for Cytb, 854 bp for EGR1, 782 bp for IRBP, 704 bp for MLL, 787 bp for myh6, 1353 bp for Rag1, 770 bp for Rh and 726 for zic1, for a total of 7844 nucleotides used in

Author's personal copy

155

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

Table 1 List of Taxa included in this study with tissue voucher and GenBank numbers. Tissues were from the personal collections of Michael E. Alfaro (MEA), Brian Victor Alfaro (VRA) and Francesco Santini (FS) at UCLA, and Peter Wainwright (PW) at UC Davis, as well as from the Natural History Museum in Victoria (NMV), South African Institute of Aquatic Biology (SAIAB), Natural History Museum of the University of Kansas (KU), and Australian Museum (EBU). Scientific Name

Common name

Tissue #

CO1

Cytb

EGR1

irbp

mll

myh6

rag1

rhod

Zic1

Triacanthodes ethiops Balistapus undulatus Anoplocapros amygdaloides Anoplocapros inermis Anoplocapros lenticularis Aracana aurita Aracana ornata

Spikefish

EBU 82004 MEA 144

JQ861002

JQ861145

JQ861169

JQ861121

JQ861195

JQ861070

JQ861052

JQ861028

JQ861096

JQ861003

JQ861146

JQ861170

JQ861122

JQ861196

JQ861071

EU108869

JQ861029

JQ861097

NMV Z 10897 NMV Z 10901 MEA 140

JQ861004

JQ861147

JQ861171

JQ861197

JQ861072

JQ861005

JQ861148

JQ861172

JQ861123

JQ861198

JQ861073

AY700346

JQ861030

JQ861098

JQ861006

JQ861149

JQ861173

JQ861124

JQ861199

JQ861074

JQ861053

JQ861031

JQ861099

JQ861007 JQ861008

JQ861150 JQ861151

JQ861174 JQ861175

JQ861125

JQ861200 JQ861201

JQ861075 JQ861076

JQ861054 AY700348

JQ861032

JQ861100 JQ861101

Caprichthys gymnura Capropygia unistriata Kentrocapros aculeatus Kentrocapros rosapinto Acanthostracion polygonius Acanthostracion quadricornis Lactophrys bicaudalis Lactophrys trigonus Lactophrys triqueter Lactoria cornuta

Rigid boxfish

MEA 171 PW 1286 NMV Z 8002 NMV Z 10837

JQ861009

JQ861152

JQ861176

JQ861126

JQ861202

JQ861077

JQ861055

JQ861033

JQ861102

JQ861010

JQ861153

JQ861177

JQ861127

JQ861203

JQ861078

JQ861056

JQ861034

JQ861103

NC009864

NC009864

SAIAB 820891 KU 1545

JQ861012

JQ861155

JQ861179

JQ861129

JQ861205

JQ861080

JQ861057

JQ861036

JQ861105

JQ861011

JQ861154

JQ861178

JQ861128

JQ861204

JQ861079

AY700343

JQ861035

JQ861104

MEA169

JQ861013

JQ861156

JQ861180

JQ861130

JQ861206

JQ861081

AY700345

JQ861037

JQ861106

Lactoria diaphana Lactoria fornasini Ostracion cubicus Ostracion immaculatus Ostracion meleagris Ostracion rhinorhynchos Ostracion solorensis Ostracion whitleyi Rhynchostracion nasus Tetrosomus concatenatus Tetrosomus gibbosus Tetrosomus reipublicae

Orange-lined triggerfish Western smooth boxfish Eastern smooth boxfish White-barred boxfish Striped cowfish Ornate cowfish

Black-banded pigmy boxfish Basketfish Basketfish Honeycomb cowfish Scrawled cowfish Spotted trunkfish

FJ583605

AY700342

Buffalo trunkfish

MEA 245

JQ861017

JQ861160

JQ861184

JQ861134

JQ861210

JQ861085

JQ861061

JQ861041

JQ861110

Smooth trunkfish

MEA 258

JQ861018

JQ861161

JQ861185

JQ861135

JQ861211

JQ861086

AY700344

JQ861042

JQ861111

Longhorn cowfish

SAIAB 806982 NMV Z 6676 SAIAB 83906 FS005

JQ861014

JQ861157

JQ861181

JQ861131

JQ861207

JQ861082

JQ861058

JQ861038

JQ861107

JQ861015

JQ861158

JQ861182

JQ861132

JQ861208

JQ861083

JQ861059

JQ861039

JQ861108

JQ861016

JQ861159

JQ861183

JQ861133

JQ861209

JQ861084

JQ861060

JQ861040

JQ861109

JQ861019 NC009865

JQ861162 NC009865

JQ861186

JQ861136

AY362207

JQ861087

JQ861062

JQ861043

JQ861112

PW 1272 VRA 014

JQ861020

JQ861163

JQ861187

JQ861137

JQ861212

JQ861088

JQ861063

JQ861044

JQ861113

JQ861021

JQ861164

JQ861188

JQ861138

JQ861213

JQ861089

JQ861064

JQ861045

JQ861114

PW 1273 MEA uncat NMV Z 6670 NMV Z 10856 MEA 175

JQ861022

JQ861165

JQ861189

JQ861139

JQ861214

JQ861090

JQ861065

JQ861046

JQ861115

JQ861190

JQ861144

JQ861215

JQ861091

JQ861066

JQ861047

JQ861120

JQ861191

JQ861140

JQ861216

JQ861092

JQ861067

JQ861048

JQ861116

JQ861192

JQ861141

JQ861217

JQ861093

AY308794

JQ861049

JQ861117

JQ861026

JQ861167

JQ861193

JQ861142

JQ861218

JQ861094

JQ861068

JQ861050

JQ861118

EBU 40868

JQ861027

JQ861168

JQ861194

JQ861143

JQ861219

JQ861095

JQ861069

JQ861051

JQ861119

Roundbelly cowfish Thornback cowfish Yellow boxfish Boxfish Whitespotted boxfish Horn-nosed boxfish Reticulate boxfish Whitley’s boxfish Shortnose boxfish Triangular boxfish Humpback turretfish Smallspine turretfish

JQ861023 JQ861024

JQ861166

JQ861025

the concatenated analyses. All sequences generated for this study were deposited in GenBank (Table 1). We used jModelTest (Posada, 2008) to select the best fitting model of sequence evolution from the candidate pool of models that can be utilized in MrBayes 3.1.2 (Ronquist and Huelsenbeck, 2003) using AIC (Akaike, 1973). We did not include the proportion of invariant sites parameter in the candidate pool, as this parameter is already taken into consideration by the gamma parameter (Yang, 2006). The GTR + G model was selected as the most appropriate for cox1, myh6 and Rag1, while HKY + G was selected as the best model for Cytb, EGR1, IRBP, MLL, and Rh. The individual gene datasets were subject to Maximum Likelihood analyses using RAxML (Stamatakis, 2006), in order to test

for incongruence between gene histories of different loci and to identify potentially contaminated or mislabeled sequences. We assigned a GTR + G model to each individual gene partition, implementing the RAxML model closest to the JModeltest results, and 500 fast bootstrap replicates using the GTR + CAT model were run. The eight individual gene datasets were then concatenated in Mesquite 2.75 (Maddison and Maddison, 2011), and the full dataset was partitioned by gene and subject to Maximum Likelihood analyses with RAxML (Stamatakis, 2006). Each partition was again assigned its own GTR + G model, and 1000 fast bootstrap replicates were generated using the GTR + CAT model. Analyses of phylogenetic relationships were conducted in a Bayesian framework using the software package MrBayes 3.1.2

Author's personal copy

156

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

(Ronquist and Huelsenbeck, 2003). The concatenated dataset was partitioned by locus and each locus was assigned the model selected by JModeltest. Each Bayesian analysis was run for 10 million generations, with 4 chains (one cold, three heated) and sampling every 1000 generations. The trace files were checked in Tracer 1.5 (Drummond and Rambaut, 2007) to ensure that the chains had reached convergence, and the first 25% of trees was discarded as burn in. Post-burn in trees were combined to obtain a 50% majority rule consensus tree. We also utilized a Bayesian speciestree gene-tree approach using BEST (Liu, 2008); the mtDNA was treated as single partition and assigned a HKY + G model selected by jModeltest, all nuclear loci were treated as individual partitions and assigned the same models employed for the MrBayes analyses. BEST was run for 20 million generations, with sampling every 1000 generations. Convergence was assessed by checking the trace files in Tracer 1.5 (Drummond and Rambaut, 2007), and the first 20% of the trees were discarded as burnin.

2.4. Divergence time estimation The concatenated alignment was analyzed, both as a single partition or as three unlinked partitions (Mitochondrial loci with HKY + G model of sequence evolution, nuclear genes with HKY + G model, nuclear genes with GTR + G model), with uncorrelated exponential priors in BEAST 1.6.2 (Drummond and Rambaut, 2007). Analyses in which all eight loci were treated as individual partitions, each with the model selected by JModeltest were also attempted, but did not reach convergence, so their results were not further considered. A birth–death prior was assigned to rates of cladogenesis. For each dataset two analyses of 20 million each, with sampling every 10000 generations, were conducted. We used Tracer 1.5 (Rambaut and Drummond, 2009) to inspect the trace files and insure that the chains had reached convergence and the ESS for all parameters was greater than 200. We removed the first 10% of the trees as burnin. We then used LogCombiner to merge the files with the remaining trees, and TreeAnnotator (Drummond and Rambaut, 2007) to obtain a timetree. Due to the potential of clade-specific rate heterogeneity potentially misleading divergence time estimates (Dornburg et al., 2012) and observations that boxfish possess much slower rates of molecular evolution relative to most other tetraodontiforms, as seen by branch lengths in both previously published phylogenetic studies (e.g., Alfaro et al., 2007) and this study, the outgroups were removed before performing the dating analyses in BEAST. Boxfish have a relatively rich fossil record that spans back to the middle Eocene (Ypresian) deposits of Monte Bolca (50 Ma), enabling us to use two fossil placed calibration points for this study. We dated the split between Ostraciidae and Aracanidae with the fossils of the Eocene Proaracana dubia, a stem aracanid, and Eocalctoria sorbinii, a stem ostraciid (Tyler and Santini, 2002; Santini and Tyler, 2003, 2004). Both of these fossils date from the Ypresian of Monte Bolca (50 Ma), and mark the minimum age for the crown ostracioid clade. We used the age of the Santonian Protriacanthus gortani (Tyler and Sorbini, 1996) to put a soft upper boundary of 85 ma on the prior age for this node. We note that while the oldest taxon currently assigned to the tetraodontiforms is Plectocretacicus clarae, from the Cenomanian of Lebanon (96 Ma), its phylogenetic affinity has been questioned by current morphological work (Francesco Santini, unpublished). For this reason we prefer to use Protriacanthus. We also used the Paleocene Moclaybalistes danekrus, (Santini and Tyler, 2003) a stem balistoid from the late Paleogene of Denmark (58–59 Ma) to put a minimum age prior on the root of the tree. Protriacanthus gortani was again used to set the soft upper bound for the stem age (crown ostracioid: offset = 50, mean = 11; root: offset = 59, mean = 8).

3. Results 3.1. Phylogenetic analyses Both Maximum Likelihood and Bayesian analyses of the concatenated dataset strongly support the same topology with very high support values present for the inter-specific and generic relationships. Nineteen out of 25 nodes within the ostracioids, including these supporting the monophyly of Ostracioidea, Aracanidae and Ostraciidae have a posterior probability (pp) and bootstrap (bsp) support of >0.99 and 100% (Fig. 1 and 2). With the exception of the ostraciid Lactoria, all genera are resolved as monophyletic with high support (pp. > 0.99, bsp > 92%). Within Aracanidae, a clade formed by the monotypic Caprichthys (rigid boxfish) and Capropygia (black banded pigmy boxfish) appear as sister to all remaining species with the two species of Kentrocapros (basketfishes) being the next group to split from the remaining deepwater boxfishes, however these relationships are poorly supported (pp = 0.77; bsp = 44%). The remaining deepwater boxfishes in our tree are split between the reciprocally monophyletic Aracana (ornate boxfishes) and Anoplocapros (smooth boxhfishes). Within Ostraciidae, we recover three highly supported (pp > 0.99; bsp = 100%) clades: (1) Acanthostracion (scrawled and honeycomb cowfishes) + Lactophrys (trunkfishes), (2) Lactoria (cowfishes) + Tetrasomus (triangular boxfishes and turretfishes) and (3) Rhynchostracion (shortnose boxfish) + Ostracion (yellow, whitespotted and horn-nosed boxfish). Acanthostracion + Lactophrys appear to be the sister taxon to all remaining ostraciids, with Lactoria + Tetrasomus sister to Rhynchostracion + Ostracion. We found strong support for the paraphyly of Lactoria, with L. fornasini appearing as the sister taxon to Tetrasomus. The monophyly of all remaining genera is strongly supported in both Bayesian (pp. > 0.99) and Likelihood (bsp = 91% or greater) analyses. The gene-tree species-tree analysis performed with BEST (not shown) recovers an identical topology for the aracanids, however relationships within the otraciids differ slightly. When allowing for gene tree heterogeneity, a clade composed of Lactoria cornuta and L. diaphana is sister to a clade with (L. fornasini + Tetrasomus sister to Rhynchostracion + Ostracion. Although the monophyly of most subclades is strongly supported (pp. = 0.96 or greater), that of the clade formed by L. diaphana + Tetrasomus sister to Rhynchostracion + Ostracion is not (pp. = 0.4).

3.2. Divergence time estimation The topology of the BEAST analysis (Fig. 3) is highly congruent with that the Maximum Likelihood and Mrbayes trees, with the exception of Kentrocapros appearing as sister taxon to all remaining aracanids, though this placement was not highly supported. We estimate the origin of crown ostracioids at 63 Ma. However, aracanids are estimated to be far more recent in origin, with a crown age of 26 Ma and a 95% highest posterior density (HPD) interval between 18 and 36 Ma. Within aracanids, we estimate the split between the Caprichthys + Capropygia and the remaining aracanids at approximately 24 Ma (95% HPD interval 16–33 Ma), and the split between Aracana and Anoplocapros is 13 Ma (95% HPD interval 9–19 Ma). Although our analyses place the initial radiation of crown aracanids in the Miocene, the bulk of the extant species-level diversity likely radiated during the Pliocene, with the three Anoplocapros species originating likely in the last 6 Ma (95% HPD interval 3–9 Ma), Aracana aurita and A. ornata being only 1.8 Ma old (95% HPD interval 0.7–3 Ma) and the Caprichthys + Capropygia clade being only 5 Ma old (95% HPD interval 2–9 Ma).

Author's personal copy

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

157

Fig. 1. Maximum likelihood phylogenetic hypothesis of boxfish relationships based on analysis of the concatenated dataset using RAxML. Black circles indicate nodes with bootstrap support values equal or greater than 90%; gray circles indicate bootstrap support between 70% and 90%, white circles indicate support lower than 70%. Abbreviations above branches leading to Ostraciidae (Ost.) and Aracanidae (Ara.).

Fig. 2. Bayesian phylogenetic hypothesis of boxfish relationships based on analysis of the concatenated dataset using Mrbayes 3.1.2. Black circles indicate nodes with posterior probability values equal or greater than 0.95; gray circles indicate posterior probability between 0.94 and 0.75, white circles indicate support lower than 0.75.

In contrast to aracanids, we estimate a substantially older origin for ostraciids, with a mean crown age of 56 Ma (95% HPD interval 45–68 Ma), and all extant genera originating by at least the Oligocene. The Acanthostracion + Lactophrys divergence estimated to have occurred approximately 32 Ma (95% HPD interval 21– 45 Ma); the Lactoria + Tetrasomus and Rhynchostracion + Ostracion

divergence is 39 Ma (95% HPD interval 29–49 Ma); Lactoria cornuta and L. diaphana separated from Tetrasomus (+ L. fornasini) 35 Ma (95% HPD interval 26–45 Ma),and the Rhynchostracion + Ostracion divergence occurring approximately 15 Ma (95% HPD interval 10–21 Ma). In contrast to aracanids, most ostraciid lineages originiated by the end of the Miocene (5.3 Ma), with the exceptions

Author's personal copy

158

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

Fig. 3. Boxfish timetree based on a Bayesian relaxed clock approach as implemented in BEAST 1.6.2. 95% HPD confidence intervals on each node indicated by gray bars. Green box indicates Cretaceous (up to 65 Ma), orange box shadows Paleogene (65–23 Ma), yellow box shadows Neogene (23 Ma-present). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

of Ostracion cubicus and O. whiteyi + O. solorensis that diverged in the Pliocene age, approximately 4 Ma (95% HPD interval 2–6 Ma), and the O. whiteyi + O. solorensis divergence approximately 1 Ma (95% HPD interval 0.4–2 Ma).

4. Discussion 4.1. Boxfish relationships Our study provides strong support for the monophyly of both the Ostraciidae and Aracanidae, agreeing with the previous studies of tetraodontiform relationships (Tyler, 1980; Santini and Tyler, 2003). Within the aracanids both likelihood and Bayesian analyses suggest a topology with the two monotypic genera Caprichthys and Capropygia representing the sister taxon to all remaining aracanids. Although not highly supported, this finding differs from that of Winterbottom and Tyler (1983), who recovered these two genera deeply nested within the Aracana + Anoplocapros clade. Winterbottom and Tyler (1983) indicate six characters to support their topology, however two of these characters (peduncular scale plates and the degree of separation of fibers within the sternobranchialis muscle) are also used to support two subsequent nodes each in their tree (their Fig. 12). Winterbottom and Tyler (1983) used Hennigian argumentation, and not a numerical cladistic approach for data analysis, and thereby did not produce a character matrix. This prevents us from understanding what characters or character states correspond to nodes within their tree. Besides the Caprichthys + Capropygia clade, our hypothesis of genus-level relationships in the Aracanidae match Winterbottom and Tyler (1983). Our topological inference of ostraciid relationships were mostly concordant with Klassen (1995), who analyzed a large osteological dataset of 108 characters for 19 extant species. While the relationships among the genera of ostraciids in Klassen (1995), which were all supported by five or more morphological characters, are fully congruent with both the Bayesian and likelihood based analyses of our molecular dataset, we differ from Klassen in two major regards. First we obtain a strongly supported relationship of

Rhyncostracion nasus as sister to Ostracion, while Klassen (1995) placed R. nasus deeply nested within Ostracion. Second, we find strong support for a clade of Lactoria diaphana and L. cornuta, which appears to be the sister taxon to L. fornasini + Tetrasomus, while Klassen (1995) inferred the existence of a clade formed by (L. cornuta (L. fornasini (Tetrasomus))). While Klassen (1995) observed several characters that supported the monophyly of Tetrasomus, and of a Lactoria fornasini + Tetrasomus clade, the only character that supports Lactoria cornuta as the sister taxon to L. fornasini + Tetrasomus is the shape of a pair of ventrolateral processes of the haemal spine of caudal vertebra #6, which join to the base of the last anal basal pterygiophore (character 88–2 in Klassen, 1985). Figs. 28–30 in Klassen (1995), as well as additional illustrations of boxfish osteology in Tyler (1980), reveal a broad variation in this feature, and we could not appreciate the difference between states 88–1 and 88–2 in Lactoria and Tetrasomus illustrated in Fig. 30. We thus assume that this character state may not be a reliable indicator of relationships within this clade. 4.2. The timing of boxfish evolution Our timetree suggests an early Paleocene origin for the boxfish, with the two families splitting from one another 63 Ma, and most of the extant boxfish species-richness accumulating in the Miocene and Pliocene. This results broadly agrees with the ages inferred by Alfaro et al. (2007) in their analyses of tetraodontiform higherlevel relationships, as well as the pattern observed in other tetraodontiform lineages (Dornburg et al., 2008, 2011; Santini et al 2009). However, we estimate a much older crown age of the Aracanidae (Ma) compared to Alfaro et al. (2007) (7 Ma) due to the inclusion in this study of Caprichthys, Capropygia and Kentrocapros. This late Oligocene/early Miocene (considering the 95% HPD interval) crown age of Aracanidae provides new insights into the radiation of these fishes into cooler temperate waters. With the exception of Kentrocapros, almost all aracanids are restricted to the rocky temperate seas surrounding Australia and New Zealand (Froese and Pauly, 2011). The enigmatic placement of Kentrocapros

Author's personal copy

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

(Figs. 1–4) clouds a specific inference of whether temperate aracanids represents a single transition between temperate rocky coasts and tropical coral reef habitats. As dispersal out of the temperate waters of Australia and New Zealand has been invoked for several marine taxa (e.g., Burridge and Smolenski, 2004; Briggs and Bowen, 2011), placing Kentrocapros will be critical for future biogeographic studies of the Aracanidae. However, even when allowing for topological uncertainty, our timetree strongly suggests an origin of the temperate lineages between the Oligocene and early Miocene, an interval of time that marks the onset of cooler oceanic conditions globally (e.g., Zachos et al., 2001). Despite the onset of temperate conditions, we find no evidence of an Oligocene origin of deepwater boxfish species diversity and instead find extant lineages to have all originated subsequent to the Late Miocene. Geologic evidence suggests a second period of surface water warming during the middle Miocene around Southern Australia and New Zealand (e.g., Gallagher et al., 2001) that corresponds to a global period of oceanic warming (e.g., McGowran and Li, 1993; Flower and Kennett, 1994; Zachos et al., 2001; Shevenell et al., 2004) suggesting that aracanids either diversified very slowly since the Oligocene or that lineages originating during this time went extinct as a result of warming waters and changes in upwelling, a hypothesis recently proposed for Antarctic fishes during this same time period (Near et al., 2012). 4.3. Taxonomic reclassification All of our phylogenetic analyses reveal Lactoria fornasini (Bianconi, 1846) to be much more closely related to Tetrasomus than to the other species of the genus Lactoria. For this reason we recommend removing fornasini from the genus Lactoria and including it in Tetrasomus. We also recommend changing the generic status of Rhynchostracion nasus, to Ostracion nasus due to the fact that the taxonomy of the Ostracion + Rhynchostracion clade has been in flux during the past few decades, with species that are alternatively recognized as member of one genus or the other (e.g., O. rhinorhynchus, recognized as R. rhinorhynchus in Klassen, 1995), and that we could not verify any significant morphological difference that can set the two genera apart. 5. Conclusion Our study provides the first molecular analysis of boxfish interrelationships, providing taxonomic revisions that are consistent with the evolutionary history of these fishes. We find that monophyly of both families, and that of most genera, is strongly supported (pp > 0.99; bsp > 91%). We revise the taxonomy of ostraciids to eliminate the non-monophyly of the genus Lactoria. By placing the phylogeny of boxfishes into a time calibrated framework, we infer that despite an origination of crown boxfishes by the early Paleocene (63 Ma) and the formation of the two boxfish families by the Eocene or Oligocene, most of the extant diversity in both families is relatively young, having originated during the Miocene or Pliocene (23 Ma or younger). We hypothesize that the diversification of the aracanids might have been driven by changes in marine habitats during subsequent Miocene cycles of warming and cooling; while the diversification of the coastal, tropical ostraciids might be connected to the expansion of coral reef ecosystems during the Oligocene and Miocene (Wood, 1999). Acknowledgments Funding for this project was provided by NSF Grant DEB 0842397 ‘‘Systematics and Influence of Coral Reefs on Diversification in Tetraodontiform Fishes.’’ to MEA and FS. This research Pro-

159

ject was made possible by the generous loan or gift of tissues from a number of colleagues and institutions: Victor Brian Alfaro (UCLA), P. Wainwright (UC Davis), Andrew Bentley and Ed Wiley (University of Kansas), Unathi Lwana (South African Institute for Aquatic Biodiversity), Dianne Bray (Museum Victoria), Mark McGrouther (Australian Museum). This manuscript was improved by the helpful comments of two anonymous reviewers.

References Alfaro, M.E., Santini, F., Brock, C.D., 2007. Do reefs drive diversification in marine teleosts? Evidence from the pufferfish and their allies (Order Tetra odontiformes). Evolution 61, 2104–2126. Akaike, H., 1973. Information theory as an extension of the maximum likelihood principle. In: Petrov, B.N., Caski, F. (Eds.), Second International Symposium on Information Theory. Akademiai Kiado, Budapest, pp. 267–281. Bartol, I.K., Gharib, M., Webb, P.W., Weihs, D., Gordon, M.S., 2005. Body-induced vortical flows: a common mechanism for self-corrective trimming control in boxfishes. J. Exp. Biol. 208, 327–344. Besseau, L., Bouligand, Y., 1998. The twisted collagen network of the box-fish scutes. Tissue Cell 30, 251–260. Brainerd, E.L., Patek, S.N., 1998. Vertebral column morphology, C-start curvature, and the evolution of mechanical defenses in Tetraodontiform fish. Copeia 1998, 971–984. Briggs, J.C., Bowen, B.W., 2011. A realignment of marine biogeographic provinces with particular reference to fish distributions. J. Biogeogr. 39, 12–30. Britz, R., Johnson, G.D., 2005. Occipito-vertebral fusion in ocean sunfishes (Teleostei: Tetraodontiformes: Molidae) and its phylogenetic implications. J. Morphol. 266, 74–79. Burridge, C.P., Smolenski, A.J., 2004. Molecular phylogeny of the Cheilodactylidae and Latridae (Perciformes: Cirrhitoidea) with notes on taxonomy and biogeography. Mol. Phylogenet. Evol. 30, 118–127. Chen, W.-J., Bonillo, C., Lecointre, G., 2003. Repeatability of clades as a criterion of reliability: a case study for molecular phylogeny of Acanthomorpha (Teleostei) with larger number of taxa. Mol. Phylogenet. Evol. 26, 262–288. Chen, W.-J., Miya, M., Saitoh, K., Mayden, R.L., 2008. Phylogenetic utility of two existing and four novel nuclear gene loci in reconstructing Tree of Life of rayfinned fishes: The order Cypriniformes (Ostariophysi) as a case study. Mol. Phylogenet. Evol. 423, 125–134. Dornburg, A., Brandley, M.C., McGowen, M., Near, T.J., 2012. Rate Heterogeneity, Uncorrelated Relaxed-Clock Models, and Patterns of Molecular Evolution in Cetaceans. Mol. Biol. Evol. 29, 721–736. Dornburg, A., Santini, F., Alfaro, M.E., 2008. The influence of model-averaging on clade posteriors: an example using the triggerfishes (Family: Balistidae). Syst. Biol. 5, 905–919. Dornburg, A., Sidlauskas, B., Sorenson, L., Santini, F., Near, T., Alfaro, M.E., 2011. The influence of an innovative locomotor strategy on the phenotypic diversification of triggerfishes (Family: Balistidae). Evolution 65, 1912–1926. Drummond, A.J., Ashton, B., Buxton, S., Cheung, M., Cooper, A., Duran, C., et al., 2010. Geneious v5.3. . Drummond, A.J., Rambaut, A., 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol. Biol. 7, 214. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797. Flower, B.P., Kennett, J.P., 1994. The middle Miocene climatic transition: East Antarctic ice sheet development, deep ocean circulation and global carbon cycling. Palaeogeogr. Palaeocl. 108, 537–555. Froese, R., Pauly, D., 2011. FishBase. . Gallagher, S.J., Smith, A.J., Jonasson, K., Wallace, M.W., Holdgate, G.R., Daniels, J., Taylor, D., 2001. The Miocene palaeoenvironmental and palaeoceanographic evolution of the Gippsland Basin, Southeast Australia; a record of Southern Ocean change. Palaeogeogr. Palaeocl. 172, 53–80. Gordon, M.S., Hove, J.R., Webb, P.W., Weihs, D., 2001. Boxfishes as unusually well controlled autonomous underwater vehicles. Physiol. Biochem. Zool. 73, 663– 671. Holcroft, N.I., 2005. A molecular analysis of the interrelationships of tetraodontiform fishes (Acanthomorpha: Tetraodontiformes). Mol. Phylogenet. Evol. 34, 525–544. Hove, J.R., O’Bryan, L.M., Gordon, M.S., Webb, P.W., Weihs, D., 2001. Boxfishes (Teleostei: Ostraciidae) as a model system for fishes swimming with many fins: kinematics. J. Exp. Biol. 204, 1459–1471. Klassen, G.J., 1995. Phylogeny and biogeography of the Ostraciinae (Tetraodontiformes: Ostraciidae). Bull. Mar. Sci. 57, 393–441. Leis, J.M., 1984. Tetraodontiformes: relationships. In: Moser, H.G., Richards, W.J., Cohen, D.M., Fahay, M.P., Kendall Jr., A.W., Richardson, S.L. (Eds.), Ontogeny and Systematics of Fishes. Am. Soc. Ichthyol. Herp., pp. 459–463 (Spec. Publ. No. 1). Li, B., Dettai, A., Cruaud, C., Couloux, A., Desoutter-Meniger, M., Lecointre, G., 2009. RNF213, a new nuclear marker for acanthomorph phylogeny. Mol. Phylogenet. Evol. 50, 345–363. Li, C., Ortí, G., Zhang, G., Lu, G.Q., 2007. A practical approach to phylogenomics: the phylogeny of ray-finned fish (Actinopterygii) as a case study. BMC Evol. Biol. 7, 44–54.

Author's personal copy

160

F. Santini et al. / Molecular Phylogenetics and Evolution 66 (2013) 153–160

Liu, L., 2008. BEST: Bayesian estimation of species trees under the coalescent model. Bioinformatics 24, 2542–2543. Lopez, J.A., Chen, W.J., Ortí, G., 2004. Esociform phylogeny. Copeia 2004, 449–464. Maddison, W.P., Maddison, D.R., 2011. Mesquite: A Modular System for Evolutionary Analysis, Version 2.75. . McGowran, B., Li, Q., 1993. Miocene planktonic foraminifera from Lakes Entrance in Gippsland; midlatitude neritic signal from a transforming ocean. Mem. Assoc. Austral. Palaeont. 15, 395–405. Near, T.J., Dornburg, A., Kuhn, K.L., Eastman, J.T., Pennington, J.N., et al., 2012. Ancient climate change, antifreeze, and the evolutionary diversification of Antarctic fishes. Proc. Natl. Sci. U.S.A. 109, 3434–3439. Nelson, J.S., 2006. Fishes of the World, fourth ed. John Wiley and Sons, Inc. Hoboken, New Jersey, USA. Posada, D., 2008. JModelTest: phylogenetic model averaging. Mol. Biol. Evol. 25, 1253–1256. Rambaut, A., Drummond, A.J., 2009. Tracer v1.5. . Ronquist, F., Huelsenbeck, J.P., 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19, 1572–1574. Santini, F., Harmon, L.J., Carnevale, G., Alfaro, M.E., 2009. Did genome duplication drive the origin of teleosts? A comparative study of diversification in ray-finned fishes. BMC Evol. Biol. 9, 164–178. Santini, F., Tyler, J.C., 2003. A phylogeny of the families of fossil and extant tetraodontiform fishes (Acanthomorpha, Tetraodontiformes), Upper Cretaceous to recent. Zool. J. Linn. Soc. 139, 565–617. Santini, F., Tyler, J.C., 2004. The importance of even highly incomplete fossil taxa in reconstructing the phylogenetic relationships of the Tetraodontiformes (Acanthomorpha: Pisces). Integr. Comp. Biol. 44, 349–357. Sevilla, R.G., Diez, A., Noren, M., Mouchel, O., Jerome, M., Verrez-Bagnis, V., et al., 2007. Primers and polymerase chain reaction conditions for DNA barcoding teleost fish based on the mitochondrial cytochrome b and nuclear rhodopsin genes. Mol. Ecol. Resour. 7, 730–734. Shevenell, A.E., Kennett, J.P., Lea, D.W., 2004. Middle Miocene Southern Ocean cooling and Antarctic cryosphere expansion. Science 305, 1766–1770. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688– 2690. Tyler, J.C., 1962. The pelvis and pelvic fin of plectognath fishes; a study in reduction. Proc. Acad. Natl. Sci. Philad. 114, 207–250.

Tyler, J.C., 1973. A new species of boxfish from the Eocene of Monte Bolca, Italy, the first unquestionable fossil record of the Ostraciontidae. Boll. Mus. Civ. Stor. Nat. Verona. Studi Ric. Giac. Terz. Bolca 2, 103–127. Tyler, J.C., 1980. Osteology, phylogeny, and higher classification of the fishes of the order Plectognathi (Tetraodontiformes). NOAA Tech. Rep. NMFS Circ. 434, 1– 422. Tyler, J.C., Bannikov, A.F., 1992. A remarkable new genus of Tetraodontiform fish with features of both Balistids and Ostraciids from the Eocene of Turkmenistan. Smithon. Contrib. Paleobiol. 72, 1–14. Tyler, J.C., Gregorova, R., 1991. A new genus and species of boxfish (Tetraodontiformes: Ostraciidae) from the Oligocene of Moravia, the second fossil representative of the family. Smithon. Contrib. Paleobiol. 71, 1–20. Tyler, J.C., Santini, F., 2002. Review and reconstructions of the tetraodontiform fishes from the Eocene of Monte Bolca, Italy, with comments on related Tertiary taxa. Boll. Mus. Civ. Stor. Nat. Verona. Studi Ric. Giac. Terz. Bolca 9, 47–119. Tyler, J.C., Santini, F., 2005. A phylogeny of the fossil and extant zeiform-like fishes, Upper Cretaceous to Recent, with comments on the putative zeomorph clade (Acanthomorpha). Zool. Scripta 34, 157–175. Tyler, J.C., Sorbini, L., 1996. New superfamily and three new families of tetraodontiform fishes from the Upper Cretaceous: the earliest and most morphologically primitive plectognaths. Smithon. Contrib. Paleobiol. 82, 1–59. Ward, R.D., Zemlak, T.S., Innes, B.H., Last, P.R., Hebert, P.D., 2005. DNA barcoding Australia’s fish species. Philos. Trans. Roy. Soc. B. 360, 1847–1857. Winterbottom, R., 1974. The familial phylogeny of the Tetraodontiformes (Acanthopterygii: Pisces) as evidenced by their comparative myology. Smithon. Contrib. Zool. 155, 1–201. Winterbottom, R., Tyler, J.C., 1983. Phylogenetic relationships of aracanin genera of boxfishes (Ostraciidae: Tetraodontiformes). Copeia 1983, 902–917. Wood, R., 1999. Reef Evolution. Oxford University Press, New York. Yang, Z., 2006. Computational molecular evolution. Oxford University Press, Oxford. Yamanoue, Y., Miya, M., Matsuura, K., Katoh, M., Sakai, H., Nishida, M., 2008. A new perspective on phylogeny and evolution of tetraodontiform fishes (Pisces: Acanthopterygii) based on whole mitochondrial genome sequences: basal ecological diversification? BMC Evol. Biol. 8, 212. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms and aberrations in global climate 65 Ma to present. Science 292, 686–693.