Molecular Phylogenetics and Evolution 55 (2010) 123–135

Contents lists available at ScienceDirect

Molecular Phylogenetics and Evolution

journal homepage: www.elsevier .com/ locate /ympev

Molecular systematics of the bubblegum coral genera (Paragorgiidae, Octocorallia)and description of a new deep-sea species

Santiago Herrera a,b,1, Amy Baco c, Juan A. Sánchez a,*

a Departamento de Ciencias Biológicas-Facultad de Ciencias, Laboratorio de Biología Molecular Marina (BIOMMAR), Universidad de los Andes,Carrera 1E No 18A – 10, (J 409, J309 Lab), Bogotá, Colombiab Department of Invertebrate Zoology, National Museum of Natural History, MRC-163, Smithsonian Institution, P.O. Box 37012, Washington, DC 20013-7012, USAc Department of Oceanography, Florida State University, 117 N. Woodward Avenue, Tallahassee, FL 32306-4320, USA

a r t i c l e i n f o

Article history:Received 25 June 2009Revised 30 November 2009Accepted 4 December 2009Available online 16 December 2009

Keywords:Bubblegum octocoralParagorgiaSibogagorgiaDeep-sea coralSeamountPhylogeny

1055-7903/$ - see front matter � 2009 Elsevier Inc. Adoi:10.1016/j.ympev.2009.12.007

* Corresponding author.E-mail address: juansanc@uniandes.edu.co (J.A. Sá

1 Present address: Massachusetts Institute of Technoographic Institution Joint Program, 2–40 Redfield Labora02543, USA. sherrera@mit.edu

a b s t r a c t

Bubblegum octocorals (Paragorgia and Sibogagorgia) play an important ecological role in many deep-seaecosystems. However, these organisms are currently threatened by destructive fishing methods such asbottom trawling. Taxonomic knowledge of conservation targets is necessary for the creation and imple-mentation of efficient conservation strategies. However, for most deep-sea coral groups this knowledgeremains incomplete. For instance, despite its similarities with Paragorgia, Sibogagorgia is particular inlacking polyp sclerites, which are present in groups like Paragorgia and the Coralliidae. Although twokinds of sclerites are very similar between Paragorgia and Sibogagorgia, other characters challenge themonophyly of these genera. Here we help to clarify the taxonomy and evolutionary relationships ofthe bubblegum octocorals and related taxa by examining molecular data. We employed nucleotidesequences of mitochondrial (ND6, ND6-ND3 intergenic spacer, ND3, ND2, COI, msh1 and 16S) and nuclear(28S and ITS2) genomic regions from several taxa to infer molecular phylogenetics and to examine thecorrespondence of morphological features with the underlying genetic information. Our data stronglysupported the monophyly of the genus Paragorgia, the family Coralliidae (precious corals), and a groupof undescribed specimens resembling Sibogagorgia. Further morphological observations were congruentregarding the uniqueness of the undescribed specimens, here defined as a new species, Sibogagorgia cau-liflora sp. nov., which occurs in both sides of the North American landmass at depths below 1700 m. Thisnew species resembles S. dennisgordoni with branching in one plane but has fairly different radiate scle-rites and significantly divergent DNA sequences. The existence of several diagnostic characters of Sibogag-orgia in S. cauliflora indicates that they indeed belong to this genus. It is however remarkable that a smallnumber of medullar canals are also found in this species; medullar canals have been considered as themain diagnostic character of Paragorgia. Thus, the evidence generated here indicates that the presenceor absence of these canals per se is not a conclusively diagnostic character for either genus. The lack ofinternal-node resolution in the inferred phylogenetic hypotheses of these genera does not allow us topropose a clear scenario regarding the evolution of these traits.

� 2009 Elsevier Inc. All rights reserved.

1. Introduction

The bubblegum octocorals (Paragorgiidae, Octocorallia) areamong the most abundant and widely distributed sessile benthicinvertebrates in deep-water ecosystems, including seamounts,lithoherms, canyons and continental shelves (DeVogelaere et al.,2005; Leverette and Metaxas, 2005; Messing et al., 1990; Morten-sen and Buhl-Mortensen, 2005). Paragorgiids play an important

ll rights reserved.

nchez).logy and Woods Hole Ocean-tory MS#33, Woods Hole, MA

ecological role in many deep-sea ecosystems, equivalent to the roleof large trees in a rain forest, by generating three-dimensional hab-itats for a great number of micro and macro organisms (Austeret al., 2005; Buhl-Mortensen and Mortensen, 2004, 2005; DeVogel-aere et al., 2005; Metaxas and Davis, 2005; Nedashkovskaya et al.,2005). In a study that examined the diversity and abundance ofinvertebrates associated with bubblegum octocorals, 1264 ani-mals, representing 47 recognized species, were found in just 13colonies (Buhl-Mortensen and Mortensen, 2005). This observationindicates that the fauna associated with bubblegum octocorals isconsiderably richer than the fauna associated with shallow-watertropical gorgonians. Unfortunately, bubblegum octocorals anddeep-sea fauna in general are rapidly becoming threatened by hu-man activities. Due to the depletion of mid-water fisheries around

124 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

the world, alternative destructive fishing techniques (e.g. bottomtrawling) are being implemented in deeper waters (Crowderet al., 2008). These practices not only have a direct negative impacton commercial fish populations, but also on all the slow-growingbenthic fauna (e.g. habitat-building octocorals) and therefore onthe ecosystems as a whole (Koslow et al., 2001; Roberts, 2002;Waller et al., 2007; Watling and Norse, 1998).

Biodiversity conservation efforts in the deep-sea (e.g., MarineProtected Areas) are crucial to ensuring that the exploitation offood resources becomes a sustainable practice (Danovaro et al.,2008; Davies et al., 2007; Morgan et al., 2005). Knowledge of con-servation targets is necessary for the creation and implementationof efficient conservation strategies. Such knowledge must includewell-founded taxonomic inventories that allow us to identify thenumber of species and their distribution patterns (Dubois, 2003).Although bubblegum octocorals have been studied since the 18thcentury, their phylogenetic relationships and taxonomic status re-main controversial.

Unlike many octocorals, bubblegum corals lack a calcified orcorneous skeleton, and thus have been included within the appar-ently polyphyletic group scleraxonia, i.e., branching alcyonaceanswith a soft skeleton composed of densely packed unfused sclerites(Bayer et al., 1983; McFadden et al., 2006; Sánchez et al., 2003).The presence or absence of the boundary canals network, a systemof reticulated canals separating the cortex of the branches from themedulla (Bayer et al., 1983), has been utilized as one of the maincriteria to split the bubblegum octocorals in two genera, ParagorgiaMilne Edwards and Haime, 1857 and Sibogagorgia Stiasny, 1937(Sánchez, 2005; Verseveldt, 1940, 1942). Unlike Sibogagorgia, Para-gorgia does not have boundary canals, but rather a few larger ca-nals perforating the sclerital medulla. Other characters, such asthe surface and medulla sclerites in Sibogagorgia, are very similarto those seen in Paragorgia. Both genera have a thin outer surfacelayer composed of radiate sclerites, whose radial ornaments canbe remarkably similar in both genera (Sánchez, 2005). Likewise,the medulla in both groups contains spindle sclerites with variousdegrees of ornamentation, reaching an extreme in the nearly bareSibogagorgia spindles (Sánchez, 2005). These two kinds of scleritesare homologous to those of other polymorphic octocorals (i.e., twokinds of polyps: autozooids and siphonozooids) such as Coralliumand Anthomastus, which according to molecular phylogeneticsstudies are closely related to Paragorgia (Berntson et al., 2001;France et al., 1996; McFadden et al., 2006; Sánchez et al., 2003;Strychar et al., 2005).

Despite its similarities with Paragorgia, Sibogagorgia is particu-lar in lacking polyp sclerites, which are present in groups like Para-gorgia and Corallium as homologous ornate stubby rods. Thus,although two kinds of sclerites are very similar between Paragorgiaand Sibogagorgia, the internal canals and polyp sclerites are indic-ative of their evolutionary divergence. A number of taxonomicstudies examining these morphological features have placed thesegenera in either a unique family, Paragorgiidae (Bayer, 1956; Sán-chez, 2005), or in two separate families, Paragorgiidae and Sibo-gagorgiidae (Verseveldt, 1942). This discrepancy is attributable tothe low number of identifiable anatomical apomorphic characters.Consequently, the aim of this study is to help clarify the taxonomyand evolutionary relationships of the bubblegum octocorals andrelated taxa by examining molecular data. We employ nucleotidesequences of mitochondrial and nuclear genomic regions from sev-eral taxa to: (1) infer historical patterns of cladogenesis, and (2)examine the correspondence of morphological features with theunderlying genetic information. Another major goal of this studyis to (3) define a new species of bubblegum octocoral from a setof undescribed specimens. These specimens share particular char-acteristics not found in any of the currently described bubblegumoctocoral species.

2. Materials and methods

Octocoral specimens (dry and ethanol-preserved) were ob-tained from the collections of the National Museum of Natural His-tory of the Smithsonian Institution (Washington DC, USA) and theNational Institute of Water and Atmospheric Research (Wellington,New Zealand).

Molecular analyses were performed on 17 coral specimens thatwere previously identified by expert taxonomists (Bayer, unpub-lished data; Sánchez, 2005). These included 10 ingroup (Paragorgiaand Sibogagorgia) and seven outgroup individuals, the latter repre-senting two different families (Coralliidae and Plexauridae), whichreflect different degrees of relatedness to the ingroup (Table 1)(Bayer, 1992; France and Hoover, 2002; France et al., 1996; McFad-den et al., 2006; Sánchez et al., 2003). Only one specimen of a de-scribed species of Sibogagorgia, S. dennisgordoni, had beenpreserved suitably for genetic analyses, but morphological com-parisons with the other described species were available from aprevious study by Sánchez (2005). New nucleotide sequences ofseven mitochondrial and two nuclear regions were generated foreach individual. Mitochondrial genomic regions include: the 30

end of the NADH dehydrogenase-6 (ND6) gene, the ND6–ND3 inter-genic spacer, the 50 end of the NADH dehydrogenase-3 (ND3) gene,the 30 end of the cytochrome oxidase-I (COI) gene, the 50 end of theDNA mismatch repair protein- mutS – homolog (msh1) gene, twodifferent regions of the ribosomal large sub-unit (16S), and the 50

end of the NADH dehydrogenase-2 (ND2) gene. Nuclear genomicregions are the complete internal transcribed spacer-2 (ITS2) anda short region of the 50 end of the ribosomal large sub-unit (28S).Additional contributed outgroup sequences from Anthomastus,Eleutherobia (16S, msh1, ND2: C. McFadden, unpub. data; McFad-den et al., 2006) and Corallium (ITS2: N. Ardila, unpub. data) wereincluded in the analyses. The latter were needed because we werenot able to obtain sequences from the same coralliid individualsfrom which we obtained the mitochondrial data.

Among the 10 ingroup specimens included in the molecularanalyses, there were four individuals (Table 1) that did not haveprior species-level identification. These specimens were intention-ally incorporated in the analyses because our preliminary morpho-logical and molecular observations were not congruent with thefeatures found in any of the currently described species. Additionaldetailed sclerite observations of these specimens were conductedusing Scanning Electron Microscopy (SEM), following the prepara-tions and procedures described by Sánchez and Cairns (2004) andSánchez (2005).

2.1. Molecular laboratory methods

Total genomic DNA was extracted from tissue samples (1–2 pol-yps) using a phenol/chloroform extraction (Mouse Tail Protocol)performed in an automated DNA isolation system (AutoGenprep965, AutoGen Inc.). The standard protocol was modified by addingan extra DNA washing step, increasing the centrifugation time inthe debris-removal step to 15 min., and diluting the purified DNAto final volume of 0.1 ml. Polymerase chain reactions were con-ducted employing five mitochondrial and two nuclear pairs of con-ventional octocoral primers (Table 2). The only exception was theMSH1 forward primer, which needed to be designed de novo to ob-tain successful amplifications in Paragorgia, Sibogagorgia, Corallium,and Paracorallium. Such abnormality is likely to be caused by a spe-cial case of mitochondrial gene rearrangement in these taxa (Bru-gler and France, 2008). The PCR reaction mixes were prepared to afinal volume of 10 l (1 ll of template) resulting in the followingconcentrations of reagents and enzymes: 1 � NH4 Buffer,2.5 � BSA, 0.5 mM dNTPs (0.125 mM each), 2.0 mM MgCl2, 0.5 u

Table 1Sequences and specimen information of the outgroup and ingroup specimens used in this study. Acronyms as follows: National Museum of Natural History of the Smithsonian Institution, USA (USNM); The National Institute of Waterand Atmospheric Research, New Zealand (NIWA); Museum and Art Gallery of the Northern Territory, Australia (NTM; Collection of C.S. McFadden (CSM). Taxonomic classification following McFadden et al. (2006) and Sanchez (2005).Sequences not generated in this study are marked with an asterisk (*). Type specimens are indicated with a cross (�). Specimens marked with (§) are the four undescribed specimens here described as a new species; their names in thetable correspond to their former identifications in the USNM collection.

Taxa Collection Locality GenBank Accession Numbers

16S ND2–16S COI MSH1 ND6-int-ND3

ITS2 28S

O. Alcyonacea

Family PlexauridaeAlaskagorgia sp. USNM 1115602 (51.52, �177.95): Off Aleutian Islands: Bering Sea: USA, 485 m GQ293240 GQ293318 GQ293270 GQ293299 GQ293337 N/A N/AMuricea purpurea USNM 1016584 Off Taboguilla Island: Panama Bay: Panama, 6 m, 1976 GQ293246 GQ293323 GQ293275 GQ293304 GQ293342 N/A N/A

Family AlcyoniidaeAnthomasttus ritteri CSM-ANRI (36.58, �122.1): Off Pebble Beach: Californa: USA, 300 m, 1998 GQ377455 DQ302893 N/A DQ302816 N/A N/A N/AEleutherobia aureum (�30.33, 30.19): Park Rynie: Kwazulu-Natal: South Africa, 22–28 m, 2008 GQ377454 N/A N/A N/A N/A N/A GQ377456Eleutherobia sp. NTM C014902 West Channel: Palau, 2005 N/A DQ302883 N/A DQ302809 N/A N/A N/A

Family ParagorgiidaeParagorgia aotearoa NIWA 3325� (�42.83, 176.92): Off east coast: New Zealand, 700 m, 1996 GQ293247 GQ293324 GQ293276 GQ293305 GQ293343 GQ293295 GQ293261Paragorgia arborea NIWA 3310 (�44.75, 174.82): Off east coast: New Zealand, 687 m, 1999 GQ293252 GQ293330 GQ293281 GQ293311 GQ293349 GQ293294 GQ293264Paragorgia arborea USNM 80937 (40.38, �67.66): Lydonia Canyon: USA, 613–430 m, 1979 GQ293253 GQ293331 GQ293282 GQ293312 GQ293350 GQ293293 GQ293260Paragorgia kaupeka NIWA 3320� (�36.16, 176.81): Off east coast: New Zealand, 820 m, 1989 GQ293254 GQ293332 GQ293283 GQ293313 GQ293351 GQ293292 N/AParagorgia regalis USNM 1014743 (19.74, �158.3): Cross Seamount: Hawaii: USA, 452 m, 2003 GQ293248 GQ293326 GQ293278 GQ293307 GQ293345 GQ293298 N/AParagorgia wahine NIWA 3326� (�42.79, �179.99): Off east coast: New Zealand, 900 m, 2001 GQ293255 GQ293333 GQ293284 GQ293314 GQ293352 GQ293296 GQ293263Paragorgia yutlinux USNM 1073480� (50.23, �128.58): Off Vancouver Isl.: British Columbia: Canada, 846–861 m, 2003 GQ293256 GQ293334 GQ293285 GQ293315 GQ293353 GQ293297 GQ293262Paragorgia sp. USNM 54831§ (23.55, �82.78): Straits of Florida: Havana: Cuba, 1638–1757 m, 1968 GQ293250 GQ293328 N/A GQ293309 GQ293347 N/A N/AParagorgia sp. USNM 1081143§ (52.98, �161.25): Derickson Seamount: Alaska: USA, 2766 m, 2004 GQ293249 GQ293327 GQ293279 GQ293308 GQ293346 GQ293289 N/ASibogagorgia dennisgordoni NIWA 3329� (�36.69, 176.46): Off east coast: New Zealand, 820 m, 1998 GQ293257 GQ293335 GQ293286 GQ293316 GQ293354 GQ293291 GQ293259Sibogagorgia sp. USNM 1122229§ (35.83, �122.61): Davidson Seamount: California: USA, 2502 m, 2006 GQ293258 GQ293336 GQ293287 GQ293317 GQ293355 GQ293290 GQ293266Sibogagorgia sp. USNM 1122230§ (35.63, �122.83): Davidson Seamount: California: USA, 3042 m, 2006 GQ293251 GQ293329 GQ293280 GQ293310 GQ293348 GQ293288 GQ293267

Family CoralliidaeCorallium kishinouyei USNM 1072441 (25.7, �171.45):: Off Laysan Island: Hawaii: USA, 1490 m, 2003 GQ293242 GQ293319 GQ293271 GQ293300 GQ293338 N/A GQ293268Corallium laauense USNM 1071433 (19.8, �156.13):: Off Keahole Point: Hawaii Island: Hawaii: USA, 867 m, 2004 GQ293243 GQ293320 GQ293272 GQ293301 GQ293339 N/A GQ293265Corallium secundum USNM 1010758 (20.88, �156.73): Off Maui: Hawaii: USA, 240 m, 2001 GQ293245 GQ293322 GQ293274 GQ293303 GQ293341 N/A N/ACorallium sp. USNM 1075800 (56.32, �142.44): Pratt Seamount: Alaska: USA, 1627 m, 2004 GQ293244 GQ293321 GQ293273 GQ293302 GQ293340 N/A N/ACorallium sp. CL8 NIWA 28233 (�37.03, 177.34): New Zealand, 910–1048 m, 1997 N/A N/A N/A N/A N/A GQ358527 N/ACorallium sp. CL54 NIWA 15662 (�36.95, �177.34): New Zealand, 1105–1113 m, 2004 N/A N/A N/A N/A N/A GQ358526 N/ACorallium sp. CL59 NIWA 41840 New Zealand, 910 m, 2007 N/A N/A N/A N/A N/A GQ358528 N/AParacorallium sp USNM 1089600 (�23.71, 168.257) New Caledonia, 470–621 m, 2003 GQ293241 GQ293325 GQ293277 GQ293306 GQ293344 N/A GQ293269Paracorallium thrinax CL30 NIWA 28215 (�30.56, �178.51), New Zealand, 165 m, 1974 N/A N/A N/A N/A N/A GQ358529 N/A

S.Herrera

etal./M

olecularPhylogenetics

andEvolution

55(2010)

123–135

125

Table 2Primers list and PCR thermal profiles.

Region Primer Source Sequence (50-30) PCR profile

16S Octo1-L France et al. (1996) AGACCCTATCGAGCTTTACTGG (94 �C:30 s; 45 �C:60 s; 72 �C:180 s) � 30Octo2-H France et al. (1996) CGATAAGAACTCTCCGACAATA

COI COIOCTf France and Hoover (2002) GGTATGGTCTATGCTATGAT (94 �C:30 s; 55 �C:30 s; 72 �C:60 s) � 35COI-BB6 France and Hoover (2002) GGTCACCCTGAGGTCTAYAT

ND2 16S544F McFadden et al. (2006) CGACCTCGATGTTGAGTTGCGG (94 �C:60 s; 56 �C:60 s; 72 �C:90 s) � 32ND21418R McFadden et al. (2004) ACATCGGGAGCCCACATA

ND6 ND61487F McFadden et al. (2004) TTTGGTTAGTTATTGCCTTT (94 �C:45 s; 49 �C:45 s; 72 �C:45s) � 35ND32126R McFadden et al. (2004) CACATTCATAGACCGACACTT

MSH1 AnthoCorMSH This study AGGAGAATTATTCTAAGTATGG (94 �C:45 s; 50 �C:45 s; 72 �C:60 s) � 32Mut-3458R Sánchez et al. (2003) TSGAGCAAAAGCCACTCC

ITS2 5.8S-436 Aguilar and Sanchez (2007) AGCATGTCTGTCTGAGTGTTGG (94 �C:30 s; 59 �C:30 s; 72 �C:45 s) � 3228S-663 Aguilar and Sanchez (2007) GGGTAATCTTGCCTGATCTGAG

28S F635sq Medina et al. (2001) CCGTCTTGAAACACGGACC (94 �C:30 s; 55 �C:60 s; 72 �C:60 s) � 31R1411sq Medina et al. (2001) GTTGTTACACACTCCTTAGCGG

126 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

Taq polymerase (Biolase™, Bioline), and 1 lM–1.33 lM primers.Negative controls were run in every experiment to test for contam-ination. The reactions were carried out in MJ Research Thermocy-clers PTC-225 (GMI, Inc.), with an initial denaturation step of5 min at 94 �C and a final elongation of 10 min at 72 �C. PCR prod-ucts were cleaned using the Exonuclease-I/Shrimp Alkaline Phos-phatase (ExoSAP-IT™, USB Corp.) method. Cycle sequencingreactions were performed using aABI BigDye Terminator v3.1 kit(Applied Biosystems Inc.), following manufacturer’s protocols. Sub-sequent purification was done through gel filtration with SephadexG-50 (Sigma–Aldrich Corp.). Automated sequencing was com-pleted using a 3730xl DNA analyzer (Applied Biosystems Inc.).Complementary chromatograms were assembled and edited usingthe Sequencher™ 4.8 software (Gene Codes Corp.).

2.2. Alignments and secondary structure inference

Each set of sequences was aligned independently using MAFFT(Katoh et al., 2002), employing the G-INS-iand Q-INS-i algorithms(gap opening penalty = 1.53, offset value = 0.07) for the proteincoding and ribosomal regions, respectively. To correct possiblemistakes, all alignments of protein coding sequences were visuallyinspected and translated to amino acids in MacClade 4.08 (Madd-ison and Maddison, 2005), based on the genetic code of Hydraattenuata (Pont-Kindon et al., 2000). No unusual stop codons, mis-placed reading frames or suspicious substitutions were identified,indicating that no nuclear pseudogenes were amplified (Bensassonet al., 2001; Lopez et al., 1994). The program 4SALE (Seibel et al.,2006; Seibel et al., 2008), which allows the incorporation ofmolecular secondary structure information, was used to improvethe alignment of the ITS2 rRNA sequences. The secondary struc-tures were inferred in the MFOLD server (http://frontend.bioinfo.rpi.edu/applications/mfold/cgi-bin/rna-form1.cgi) based on struc-tural homology with the typical eukaryotic 4-ring model (Coleman,2003), and minimization of the folding’s Gibbs free energy (Zuker,2003).

2.3. Phylogenetic Inferences

Each data set of each individual region (excluding the ND3–ND6intergenic spacer) was analyzed independently to infer its evolu-tionary history using maximum parsimony (MP), maximum likeli-hood (ML) and Bayesian inference (BI) methods. Branch-and-bound semi-exhaustive searches under the MP optimality criterionwere conducted in PAUP* 4.0b10 (Swofford, 2002). The number oftrees saved during the searches was not restricted. Gaps were trea-ted as missing data. Statistical confidence on nodes was estimatedvia 1000 non-parametric bootstrap pseudo-replicates (50 repli-cates for each heuristic search with random addition sequence).

Nucleotide substitution models and their correspondent param-eter values were selected for every region based on the AkaikeInformation Criteria (AIC) as implemented in Modeltest 3.7 (Posada,1998; Posada and Buckley, 2004) (Table 3). Gene trees were esti-mated under the ML optimality criterion using the program GARLI0.951 (Zwickl, 2006). The analyses were run applying the selectedsubstitution models, base frequencies, substitution rates andparameters of variation among sites. All other general and geneticalgorithm settings were left by default. Non-parametric bootstrap(1000 pseudo-replicates) was performed in RAxML 7.0.4 (Stamata-kis et al., 2008) at the CIPRES portal (http://www.phylo.org).

Bayesian Inference (BI) of gene phylogenies was carried out inMrBayes 3.1.2 (Huelsenbeck and Ronquist, 2001; Ronquist andHuelsenbeck, 2003) using the same substitution models. Defaultprior distribution settings were assumed for all parameters. Threeindependent analyses of 1 � 107 Monte Carlo Markov Chain(MCMC) generations (�4 chains) were run for every independent re-gion, with a sampling frequency of 100 generations (burn-in = 2500). MCMC runs were analyzed in the program Tracer 1.4.1(Rambaut and Drummond, 2007). Convergence was indicated bythe ‘‘straight hairy caterpillar” (Drummond et al., 2007) shape ofthe stationary posterior-distribution trace (generations vs. LnL) ofeach parameter. Other examined convergence and mixing diagnos-tics included the standard deviation of partition frequencies (<0.01),the potential scale reduction factor PSRF (�1.00), the effective sam-ple sizes EES (>200), and the similitude of posterior probabilities ofspecific nodes between different runs in the program AWTY (http://ceb.csit.fsu.edu/awty) (Nylander et al., 2008). The obtained treeswere summarized into majority-rule (50%) consensus tree.

Several combined analyses were conducted for different typesof concatenated datasets: (1) all mitochondrial (mt) regions, (2)all mt protein-coding regions, and (3) a set of three mt regions(16S, msh1 and ND2) created to examine the effect of the inclusionof different outgroups in the analyses (see Table 3). The later set in-cluded only these three regions in order to minimize the possibleeffects of missing data in the inferences. Maximum parsimonyand maximum likelihood phylogenetic inferences were carriedout as mentioned above. For the ML analyses the general time-reversible model, with a proportion of invariant sites and a gammadistributed rate variation across sites (GTR + I + G), was assumed.Combined Bayesian Inference analyses were performed with expli-cit character partitions for each concatenated region, along withtheir independently-selected models of evolution. Monte CarloMarkov chains were run for 1 � 107 generations. To account forthe rate variation among partitions (Marshall et al., 2006) we al-lowed the rates to vary under a flat Dirichlet prior distribution(ratepr = variable). The parameters of nucleotide frequencies,substitution rates, gamma shape, and invariant-sites proportionwere unlinked across partitions.

Table 3Summary of the sister relationships found for the group of undescribed specimens under all employed data sets and inference methods. The relationships are expressed in New notation (Felsenstein et al., 1986). Only node supportvalues above 50% are shown. Nodes with support values lower than 50% are indicated with a cross (+). The ‘‘(x of n)” format represents the number of most parsimonious trees (ou f the total) that contained a given node. The asterisk (*)symbolizes extremely short branch lengths for the supporting branch. The information in the last column ‘‘Characters and Model” indicates (in this order): total number of ch acters in the aligned matrix, number of variable sites,number of parsimony informative sites, percentage of variable sites, and assumed model of nucleotide evolution. The dash (–) indicates absence of support.

Datasets (Coralliidae, undescribedspecimens)

(Paragorgia, undescribedspecimens)

(S. dennisgordoni/outgroup, undescribedspecimens)

Unresolved Characters and model

Individual regionsParagorgia, Sibogagorgia, Corallium and Paracorallium

28S MP (3 of 6) – – (3 of 6) 597, 16, 5, 2.7%, TIM + IML 52 – – –BI - – – +

ITS2 MP - (3 of 3), 59 – – 367, 89, 53, 24.2%, HKY+GML - 92 – –BI - 83 – –

ND6 MP (1 of 2) – (1 of 2) – 468, 75, 42, 16.0%, GTR+GML - – +* –BI - – 52 –

ND3 MP - – – + 126, 19, 11, 15.1%, HKY+IML - – +* –BI - – – +

COI MP (3 of 12) (1 of 12) (8 of 12) – 459, 48, 25, 10.5%, GTR+GML + – – –BI - – – +

MSH1 MP (3 of 3), 55 – – – 729, 152, 82, 20.8%,HKY+GML + – – –

BI 78 – – –16S MP (1 of 3) (1 of 3) (1 of 3) – 547, 63, 33, 11.5%, TVM+I

ML - – +* –BI - – – +

ND2 MP (2 of 2), 89 – – – 564, 95, 50, 16.8%, HKY+GML 92 – – –BI 83 – – –

All mtDNA coding regionsParagorgia, Sibogagorgia, Corallium and Paracorallium

ND6-intND63-ND3-COI- MSH1–16S-ND2

MP (2 of 2), 90 – – –ML 100 – – – 2958, 460, 248BI 90 – – –

All mtDNA protein-coding regionsParagorgia, Sibogagorgia, Corallium and Paracorallium

ND6-ND3-COI- MSH1–16S-ND2 MP (2 of 2), 91 – – – 2346, 389, 210ML 100 – – –BI 0.91 – – –

Additional outgroupsEleutherobia + Paragorgia, Sibogagorgia, Corallium and Paracorallium

28S MP (3 of 3) – – –ML - – + – 597, 16, 5BI - – – +

16S, MSH1, ND2 MP (1 of 2) – (1 of 2) –ML 98 – – – 1470, 312, 156BI 90 – – –

(continued on next page)

S.Herrera

etal./M

olecularPhylogenetics

andEvolution

55(2010)

123–135

127

ickt oar

Tabl

e3

(con

tinu

ed)

Dat

aset

s(C

oral

liid

ae,u

nde

scri

bed

spec

imen

s)(P

arag

orgi

a,u

nde

scri

bed

spec

imen

s)(S

.den

nisg

ordo

ni/o

utgr

oup,

un

desc

ribe

dsp

ecim

ens)

Un

reso

lved

Ch

arac

ters

and

mod

el

Ant

hom

astu

s+

Para

gorg

ia,S

ibog

agor

gia,

Cora

llium

and

Para

cora

llium

16S,

MSH

1,N

D2

MP

(1of

2)–

(1of

2)–

ML

-–

+*–

1463

,285

,145

BI

-–

52–

Eleu

ther

obia

and

Ant

hom

astu

s+

Para

gorg

ia,S

ibog

agor

gia,

Cora

llium

and

Para

cora

llium

16S,

MSH

1,N

D2

MP

(1of

2)–

(1of

2)–

ML

-–

+*–

1463

,323

,161

BI

-–

98–

Eleu

ther

obia

,Ant

hom

astu

s,M

uric

eaan

dA

lask

agor

gia

+Pa

rago

rgia

,Sib

ogag

orgi

a,Co

ralli

uman

dPa

raco

ralli

um16

S,M

SH1,

ND

2M

P(3

of3)

––

–M

L-

–+*

–14

92,4

23,2

59B

I-

––

+

128 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

A ‘‘total evidence” combined analysis was not performed be-cause of two main reasons: (1) the specimens, and likely the spe-cies, of the selected outgroup Coralliidae from which weobtained the nuclear ITS2 and the mitochondrial sequences werenot the same (please refer to Section 1 and Table 1); (2) it has beenshown that the combined analysis of genes or gen-complexes (i.e.,linked genes such as the mitochondrial ones) with incongruentevolutionary histories can produce positively misleading resultsby increasing the support of wrong phylogenetic tree (Edwardset al., 2007; Hedtke et al., 2006; Kubatko and Degnan, 2007)

2.4. Divergence estimates

Genetic divergence, measured as genetic distances amongspecies and genera, was estimated using the mitochondrial pro-tein-coding dataset. This data set was chosen because it was morecomplete than any of the nuclear data sets and also excluded theinsertion–deletion (INDELS) uncertainty present in non-coding re-gions; therefore it minimized the noise that can arise from missingdata. Maximum likelihood distances were calculated in PAUP*employing a Neighbor-Joining tree and the GTR + I + G model.

3. Results

Mostly complete sequence sets were obtained, except in thefew cases where nucleotide sequences from direct PCR productshad low quality (see Table 1). Sequence length variations, for theindividual regions of Paragorgia, Sibogagorgia, Corallium, Paracoral-lium and the undescribed specimens, were observed in the 28S(587–592 bp), ITS2 (280–301 bp), ND6 (462–465 bp), ND6–ND3intergenic spacer (22–45 bp), msh1 (720–723) and 16S (338–353,163–192 bp). The number of nucleotide residues in the ND3(127 bp) ND2, (564 bp) and COI (461 bp) was constant among allexamined taxa. The percentages of nucleotide variation withineach aligned region ranged between 2.7% (28S) and 24.2% (ITS2)for the nuclear and 10.5% (COI) and 20.8% (msh1) for the mitochon-drial (Table 3).

3.1. Molecular phylogenetic inferences

Phylogenetic hypotheses obtained for all combined and individ-ual datasets, under all inference methods (MP, ML and BI), sharedtwo overall patterns:

(1) The well-supported monophyly of the clades Paragorgia,Coralliidae (Corallium + Paracorallium) and the group of unde-scribed specimens, (2) the failure of the undescribed specimensto group with any described species.

The analyses of each independent genomic region yielded, asmentioned above, gene trees with highly supported deep nodes,but very poorly resolved relationships among major clades(Fig. 1A). This was true for all regions, with the exception of theITS2, msh1 and ND2. These three regions produced fully resolved,although dissimilar, tree topologies; mitochondrial and nucleartrees differed appreciatively in arrangement of the internal nodes.The ND2 and msh1 data grouped the undescribed specimens withthe Coralliidae, and Paragorgia with S. dennisgordoni (not shown).In contrast, the nuclear ITS2 recovered the undescribed specimensgrouping together with Paragorgia, and the Coralliidae with S. den-nisgordoni (Fig. 1B). It is noteworthy that these datasets showedthe highest percentage of nucleotide variation (P16.8%) of all indi-vidual genomic regions, and trees based on them were the only oneswith relatively high support values for the internal nodes (see Table 3for an example of the positioning of the Sibogagorgia sp. group).

Combined phylogenetic analyses of the all-mt-regions and pro-tein-coding datasets (Table 3) produced fully resolved trees. The

Coralliidaeundescribed

specimensS. dennisgordoni

Paragorgia

Und. Spec.3

Und. Spec.2Und. Spec.4

A

CB

Und.

Spe

c.1

Und. Spec.2

Und. Spec.3

Und. Spec.4

Fig. 1. Unrooted evolutionary-tree hypotheses of the relationships among genera and species of the families Coralliidae (Corallium and Paracorallium), Paragorgiidae(Paragorgia and Sibogagorgia) and the unidentified specimens, generated through BI analyses. The unidentified specimens where enumerated; number 1 corresponds to thespecimen with catalog number USNM 54831, number 2 to USNM 1081143, number 3 to USNM 1122229 and number 4 to USNM 1122230 (see Table 1). (A) Dendrogramsummarizing the gene-tree hypotheses; this dendrogram was obtained through the loose consensus method as implemented in the program Dendrogram 2.4 (Huson, 2007).(B) Phylogram obtained with the ITS2 sequences. (C) Phylogram obtained with the all-mt-regions concatenated dataset. Support values P 90% are shown as circular pies.Each color represents support found with each one of the inference methods: magenta for MP, cyan for ML and green for BI. The scale bar represents the number ofsubstitutions per site. (For interpretation of references to color in this figure legend, the reader is referred to the web version of this article.).

S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135 129

topologies were identical and all nodes (except for the one linkingP. regalis and P. yutlinux) were well supported by high MP and MLbootstrap values, and BI posterior probabilities (>90%). Thesehypotheses reflected the same relationships recovered by msh1and ND2 data sets (Fig. 1C).

In general, the establishment of a rooting point in the trees wasproblematic due to the discordance in the positioning of the se-lected outgroup (Coralliidae) with respect to the ingroup (Paragor-giidae). The addition of more distantly related outgroups in theanalyses of the 16S-msh1-ND2 dataset indicated that S. dennisgor-doni is the extant member of the earliest divergent lineage in theCoralliidae + Paragorgiidae group (Fig. 2), although this interpreta-tion may be inaccurate due a possible effect of long branch attrac-tion (LBA). The inclusion of Eleutherobia derived into the samewell-supported and resolved tree topology found with the all-mt-regions and protein-coding datasets (Fig. 2C). However, this re-sult was not consistent when the sequences of Muricea, Alaskagor-gia or Anthomastus were incorporated in the analyses. Suchincorporation reduced the deep-nodes resolution. Incongruentrelationships were obtained among different out group-subsetsand phylogeny estimation methods. Trees were characterized byvery low values of node support, and either unresolved topologies(Table 3), or extremely short inner branches (Fig. 2).

3.2. Genetic divergence

Genetic divergence within clades (Paragorgia, Coralliidae andthe group of undescribed specimens) was, on average, lower than

when compared to the divergence between them (including S. den-nisgordoni) (Fig. 3). The clades with larger and smaller ranges of ge-netic variation were the Paragorgia (0.43–4.87%) and theundescribed specimens group (0–0.75%), respectively. Distanceswere the greatest when measured between S. dennisgordoni andeach one of the clades. On the other hand, the genetic divergenceswithin Coralliidae and the undescribed specimens were the small-est of all the inter-clade comparisons.

3.3. Posterior morphological examination of unidentified specimens

The examined undescribed specimens, which were previouslyidentified as Paragorgia sp. (2) and Sibogagorgia sp. (2), showedvery similar morphological features, both macroscopic and micro-scopic. Surface radiate sclerites were almost identical among allthe four specimens (Figs. 4 and 5), but did not resemble the onesfound in any of the currently accepted species in either genus (Sán-chez, 2005). These specimens also presented both boundary andmedullar canals (Fig. 7). Branch coloring was either beige or red.Pink polyps were observed in a couple of beige specimens.

4. Discussion

4.1. Correspondence of morphologically-base taxonomy and moleculardata

The molecular data generated in this study indicated heteroge-neous levels of correspondence with the currently accepted

D

BA

C

Und. Spec.1Und. Spec.2

Und. Spec.3

Und. S

pec.4

Und.

Spe

c.1

Und. Spec.2

Und. Spec.3

Und. Spec.4

Und. Spec.1Und. Spec.2

Und. Spec.3

Und. S

pec.4

Und.

Spe

c.1

Und. Spec.2

Und. Spec.3

Und. Spec.4

Fig. 2. Unrooted phylogenetic hypotheses of the genera and species of the families Coralliidae, Paragorgiidae and the unidentified specimens, including additional outgroups.All trees were obtained through ML analyses of the concatenated 16S-msh1-ND2 datasets. The unidentified specimens where enumerated; number 1 corresponds to thespecimen with catalog number USNM 54831, number 2 to USNM 1081143, number 3 to USNM 1122229 and number 4 to USNM 1122230 (see Table 1) (A) Outgroups:Anthomastus, Eleutherobia, Muricea and Alaskagorgia (B) Outgroups: Anthomastus and Eleutherobia. (C) Outgroup: Eleutherobia. (D) Outgroup: Anthomastus. Supportvalues P 90% are shown as circular pies. Each color represents support found with each one of the inference methods: magenta for MP, cyan for ML and green for BI. The scalebar represents the number of substitutions per site. (For interpretation of references to color in this figure legend, the reader is referred to the web version of this article.).

130 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

morphologically-based taxonomy. Our results strongly supportmonophyly of the genus Paragorgia, the family Coralliidae (pre-cious corals), and the group of undescribed specimens. Each ofthese clades exhibited low intra-clade and high inter-clade geneticdistances (Fig. 3). Paragorgia presents a special case of very low ge-netic divergence, similar to the one found within P. arborea (<0.5%),between the pairs of Pacific morphospecies P. aotearoa and P. wa-hine, and P. yutlinux and P. regalis (Sánchez, 2005). It would notbe surprising if the morphological differences that had been foundbetween these pairs of morphospecies, such as color, scleritalstructure, and branch shape (see Sánchez, 2005), were productsof phenotypic plasticity within a single species rather than from

genetic differentiation (e.g., Forsman et al., 2009; Kim et al.,2004; Prada et al., 2008; Sánchez et al., 2007; West et al., 1993).It is also possible that this pattern could be explained by a recentdivergence process or even by the low rate of mitochondrial evolu-tion found in octocorals (France and Hoover, 2002). However, amore extensive interspecific and intraspecific sampling is neededto discern between these alternatives. Despite the inferred mono-phyly of the Coralliidae, the results from the analyses of our mito-chondrial data are against the monophyly of the precious coralgenus Corallium. In the mitochondrial phylogeny (Fig. 1C) Paracor-allium appears nested within Corallium, which suggests paraphylyof the later. It is important to emphasize that these results need

0

1

2

3

4

5

6

7

8

9

I II III IV V VI VII VIII IX

Gen

etic

Dis

tanc

e (%

)

Fig. 3. Genetic variation within and between Paragorgia, Coralliidae, the unde-scribed specimens and S. dennisgordoni ‘‘clades”, based on the protein-codingdataset. The amount of genetic variation was calculated as maximum likelihoodcorrected distances and is expressed in percentage ranges. (I) Within Coralliidae. (II)Within Paragorgia. (III) Within the group of undescribed specimens. (IV) BetweenCoralliidae and S. dennisgordoni. (V) Between Paragorgia and S. dennisgordoni. (VI)Between group of undescribed specimens and S. dennisgordoni. (VII) BetweenCoralliidae and Paragorgia. (VIII) Between Paragorgia and group of undescribedspecimens. (IX) Between Coralliidae and the group of undescribed specimens. Meandistance values represented with a circle. Bars show the range of variation amongspecimens. The intraspecific distance calculations in P. arborea and in the C. laauenseand Corallium sp. clade were averaged for graphical representation.

Fig. 4. Sibogagorgia cauliflora sp. nov. Holotype (USNM 1122229). (A) Radiates fromthe surface cortex (scale 0.01 mm). (B) Ornamentation detail of radiates from thesurface cortex. (C) Sclerites from the medulla (scale 0.1 mm). (D) Intermediateradiate-spindle forms from the inner cortex.

Fig. 5. Sibogagorgia cauliflora sp. nov. Paratype (USNM 1122230). (A) Radiates fromthe surface cortex (scale 0.01 mm). (B) Sclerites from the medulla (scale 0.01 mm).

Fig. 6. Sibogagorgia cauliflora sp. nov. (A) In situ photograph during the collection ofthe Holotype (USNM 1122229) using the Tiburon remotely operated vehicle. (B) Insitu photograph during the collection of the Paratype (USNM 1122230). (C)Examined fragment of the Holotype. (D) Examined fragment of the Paratype. ScaleC–D in centimeters with millimeter marks. Arrows indicate siphonozooids.

S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135 131

to be taken with prudence given that the number of specimens ofprecious corals sampled for the nuclear (ITS2) is smaller than the

one sampled for the mitochondrial sequences. Futhermore, theITS2 sequences were obtained from different specimens than the

Fig. 7. Transversal cross-section of a terminal branch of S. cauliflora sp. nov. (a)Boundary canals. (b) Medullar canals. Scale 1.0 mm.

132 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

mitochondrial ones. Besides, the precious corals group comprises arelatively high number of described species, ca. 32, (Bayer andCairns, 2003; Castro et al., 2003) thus the inclusion of more specieswill likely be necessary to elucidate its internal relationships(Graybeal, 1998).

All our morphological observations and molecular-data analy-ses are conclusive regarding the monophyly and uniqueness ofthe group of undescribed specimens. The existence of two diagnos-tic characters of Sibogagorgia in the undescribed specimens (i.e.presence of boundary canals and lack of tentacular sclerites inthe polyps) indicate that they indeed belong to this genus (Sán-chez, 2005). It is however remarkable that a small number ofmedullar canals are also found in these specimens (Fig. 7); medul-lar canals are a diagnostic character of Paragorgia (Sánchez, 2005)and thus indicates that the presence or absence of medullar andboundary canals per se might not be a conclusively diagnostic char-acter of neither Paragorgia or Sibogagorgia. Nevertheless, the lack ofinternal-node resolution in the inferred phylogenetic hypothesesdoes not allow us to propose a clear scenario regarding the evolu-tionary history of this character. Another noteworthy particularityobserved in the group of undescribed specimens was the pattern ofornamentation of their surface radiate sclerites (Figs. 4A, B and 5A).Such pattern is strikingly different from the one found in any othercurrently described Sibogagorgia species (Sánchez, 2005). All thisleads us to believe that the group of undescribed specimens is anindependent taxonomic unit, i.e. species. Furthermore, the rela-tively long branch that supports this group in all phylogenetichypotheses, which indicates large genetic differentiation reinforcethis idea (Figs. 1–3). Thus, we propose the establishment of a newspecies: Sibogagorgia cauliflora sp. nov.

4.2. Species description: Sibogagorgia cauliflora sp. nov

Family Paragorgiidae Kukenthal, 1916Genus Sibogagorgia Stiasny, 1937Sibogagorgia cauliflora sp. nov.

Holotype: USNM1122229, T947-A9, (35.83, �122.61), DavidsonSeamount, ROV Tiburon Dive 947, 3042.4 m, 1 February 2006(Figs. 6A and C). Paratype: USNM 1122230, T945-A12, (35.63,�122.83), Davidson Seamount, ROV Tiburon Dive 945, 2502.1 m,30 January 2006 (Figs. 6B and D). Type specimens are depositedat the invertebrate zoology collection of the National Museum ofNatural History, Smithsonian Institution, Washington, DC, USA.Additional examined material: USNM 1081143, (52.98, �161.25),Derickson Seamount, Alaska, USA, ROV Jason Dive 93, Spec 31,2766 m, 2004. USNM 54831, (23.55, �82.78), Straits of Florida,northwest of Havana, Cuba, 1638–1757 m, 25 March 1968, cruise6802 (Pillsbury R/V), station 634.

Diagnostic characters: Oval surface sclerites, mostly 8-radiatesbut often include fused radial ornaments. Radial ornaments withgrooves and densely packed pyramidal sub-ornamentation(Figs. 4A, B and 5A). Presence of both boundary canals and smallmedullar canals (Fig. 7).

Description: Sibogagorgia (Sánchez, 2005) with 1–2 perforatingcanals in the medulla (Fig. 7). Beige or red branching coloniesspreading in one plane (up to ca. 150 cm in height and ca.120 cm in length). Main branches considerably thicker than termi-nals (Figs. 6A and B). Terminal branches clavate (6–10 mm indiameter: Figs. 6C and D). Autozooids located throughout thebranches in an irregular fashion (Figs. 6C and D). Tiny granularsiphonozooids located uniformly throughout the branches(Figs. 6C and D). Surface of the cortex with oval sclerites from 8-radiate origination (0.05–0.08 mm in length; 0.035–0.4 mm inwidth); characteristic radial ornaments often fused including mul-tiple grooves and pyramidal sub-ornaments. Medulla with eithersmooth to moderately ornate straight spindles of various sizes,up to 0.6 mm in the larger forms (Figs. 4C and 5B). Smaller forms(<0.4 mm) usually ornate (Figs. 4C and 5B). Occasional intermedi-ate radiate-spindle sclerites in the inner cortex (e.g., Fig. 4D).

Morphological variation: The examined specimens have nearlyidentical types of radiates, which is the main diagnostic characterfor this species. The medulla sclerites, on the other hand, vary con-siderably between holotype and paratype. Particularly, the holo-type presents the large, smooth and barely ornate spindles thatare only found in other Sibogagorgia species. Both the holotypeand paratype presented a beige coloration with slightly projectedpink polyp apertures. Contrastingly, the additional observed spec-imens had a uniform red or beige coloration. These observationsare congruent with the results of a previous study, which indicatethat the shape of medullar sclerites and colony coloration arehighly polymorphic characters within the species of both Paragor-gia and Sibogagorgia and their variation seems to be produced byphenotypic plasticity (Sánchez, 2005).

Genetic variation: The measured genetic distances among S.cauliflora specimens ranged between 0% when comparing the Paci-fic specimens and 0.77% when comparing among the Atlantic andPacific specimens. Although such divergence is comparable withthe ones found between some Paragorgia morphospecies, it is pos-sible that these morphospecies are in fact variants within a singlespecies, as discussed in Section 4.1. Another likely possibility isthat the lineage that gave rise to Sibogagorgia has faster mutationrates, which can also explain the relatively long branches that sup-port these taxa. In such case the measured genetic distances withinS. cauliflora are not comparable with the ones within other slower-evolving taxa.

Distribution: Pacific (Northeastern: California and Alaska) andAtlantic (Western: Caribbean) oceans. The discontinuous distribu-tion of S. cauliflora suggests that this might be another cosmopoli-tan bubblegum coral species (Grasshoff, 1979; Sánchez, 2005). TheCaribbean specimen of S. cauliflora constitutes the first record ofthe genus in the Atlantic Ocean. The geographical records of thisnew species enrich the deep-sea diversity inventories of the

S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135 133

regions where they were found, and provide potentially importantinformation for the design conservation strategies such as pro-tected areas.

Species comparisons: All examined specimens, included in mor-phological and molecular analyses were congruent with a singlemonophyletic lineage corresponding to a morphological species.Externally S. cauliflora resembles S. dennisgordoni Sánchez withbranching in one plane. However, S. cauliflora and S. dennisgordonihave appreciatively different radiate sclerites (see for comparisonSánchez, 2005: p. 65–66) and their DNA sequences are also signif-icantly divergent (Figs. 1–3).

Etymology: cauliflora is the Latin word for cauliflower (Brassicaoleracea). It was chosen because the surface radiate sclerites of thisnew species morphologically resemble that vegetable (Figs. 4A and5A).

4.3. Discordance in evolutionary trees

Although the number of molecular markers employed in thisstudy is the largest of any octocoral phylogenetics study to date,we were not able to obtain a strongly supported phylogenetichypothesis for the bubblegum octocorals (Figs. 1 and 2). It has beenwidely suggested that the inclusion of a large number of genomicregions in the phylogenetic analyses would improve the quality ofthe inferences (Cummings et al., 1995). The logical basis of thissuggestion is that ‘‘total evidence” will allow the embracement ofa larger diversity of characters and heterogeneous mutation rates(Cummings et al., 1995; Kluge, 2004). For our particular case, themutation rates of each region, which are somehow reflected bythe amounts of nucleotide variation, were considerably heteroge-neous (see Table 3). The most variable regions in the Corallii-dae + Paragorgiidae group, the ND2, msh1 and ITS2, correspondwith the ones that had been reported for other groups of octocorals(Aguilar and Sánchez, 2007; McFadden et al., 2006). In addition,these same regions were the only ones that yielded totally resolvedand well supported gene trees (Table 3 and Fig. 1). It is then possi-ble that the ND2 and msh1 data sets had a large contribution tophylogenetic signal in the concatenated mitochondrial analyses(i.e., number of informative sites), thus obscuring the phylogeneticinformation provided by other regions (Baker et al., 1998; Edwards,2008). It seems logical that by increasing the number of sampledcharacters per region, and thus increasing the total number of sam-ple variable sites, we will not only be able to obtain gene trees withmore resolved relationships, but to increase the confidence in thehypotheses obtained from concatenated analyses (Joly et al.,2009; Rasmussen and Kellis, 2007). On the other hand, the combi-nation of this strategy with the incorporation of additional inde-pendent markers (preferably nuclear in this case), could lead to abetter estimation of the species-tree in the bubblegum octocorals(Edwards, 2008; Felsenstein, 2006). The main advantage of utiliz-ing novel species-trees inference methods is that they allow theincorporation of information from incongruent gene trees, suchas the ones found between nuclear ITS2 and mitochondrial genes(Carstens and Knowles, 2007; Edwards, 2008; Pollard et al.,2006). Likely sources of incongruence in this group could includeincomplete lineage sorting, old introgression events betweenancestors of today’s Paragorgiids and Coralliids, and horizontalgene transfer (Maddison, 1997).

The results of the mitochondrial phylogenetic analysis thatincorporated different outgroups indicate, although not conclu-sively, that the group Paragorgiidae + Coralliidae is monophyletic(Fig. 2). This observation is congruent with other results from taxo-nomically-broader molecular and morphological studies (Bayer,1992, 1993; Berntson et al., 2001; France et al., 1996; McFaddenet al., 2006; Sánchez et al., 2003; Strychar et al., 2005). In general,the inclusion of such outgroups produced high instability in the

tree topologies. It is plausible that the effects of incomplete taxonsampling, which can lead to long-branch attractions, could haveplayed an important role generating the observed problems intrees resolution, clade stability and statistical support (Graybeal,1998; Hedtke et al., 2006; Hillis et al., 2003; Lecointre et al.,1993). This can be particularly important considering that somespecies of deep-sea bubblegum octocorals are extremely rare –meaning that there could be a fair number of undiscovered (orunidentified) species. The genus Sibogagorgia seems to be a goodexample of that situation. Only one or two individuals per specieshave been collected throughout history (Sánchez, 2005). Conse-quently, notwithstanding the fact that the sister taxa relationshipbetween S. dennisgordoni and S. cauliflora is not supported by ourmolecular data, a more clear alternative is not available, and there-fore the genus should remain valid until a more complete sampleof taxa can be included in the analyses.

Acknowledgements

Support for this study was generously provided by a minigrantfrom the Global Census of Marine Life on Seamounts Project(CenSeam) to J.A.S. and S.H., a grant from the Facultad de Ciencias,Department of Biological Sciences of the Universidad de losAndes to J.A.S, the National Systematics Laboratory of NOAA’sNational Marine Fisheries Service, a Smithsonian Graduate StudentFellowship to S.H., an award from the Systematics Research Fund ofthe Systematics Association and the Linnean Society of London toS.H., and a Grant-in-Aid of Research from the Sigma Xi ResearchSociety to S.H. A.B.T. was supported by grant No. NA07OAR4600292from the NOAA Office of Ocean Exploration. We are especiallythankful to S.D. Cairns, A.G. Collins, M. Consalvey, A. Rowden,M. Clark, D. Tracey, K. Schnabel, S. Mills, C.L. Agudelo, L.F. Dueñas,N. Ardila, C.S. McFadden, M. Taylor, A. Ormos, J. Hunt, L. Weigt,S. Whittaker, L. Monroy, M. Herrera, M. Simpson, C.D. Cadena,S. Alvarado, J. Cortes, M. Sangrey, and B. Stone for their generoussupport and assistance. Samples were provided by the NationalMuseum of Natural History of the Smithsonian Institution,USA (NMNH-SI) and the National Institute of Water and Atmo-spheric Research, New Zealand (NIWA Invertebrate Collection).Coralliid and Paragorgia specimens from Hawaii (except C. secun-dum and P. regalis) were collected by A.B.T. through grants Nos.NA03OAR4600108 and NA04OAR4600071 and from Pratt Sea-mount through grant No. NA04OAR4600051 from the NOAA Officeof Ocean Exploration. Sibogagorgia cauliflora from Derickson Sea-mount was collected by A.B.T. through Grant No, UAF-040118 fromthe NOAA West Coast and Polar Region NURP Center. Specimens ofSibogagorgia cauliflora from Davidson Seamount were provided byLonny Lundsten of MBARI. C.S. McFadden and N. Ardila kindly con-tributed additional sequences for the analyses. Molecular work wasperformed at the Laboratories of Analytical Biology NMNH, Smith-sonian Institution. Image from Fig. 7 was taken by M. Taylor.

References

Aguilar, C., Sánchez, J.A., 2007. Phylogenetic hypotheses of gorgoniid octocoralsaccording to ITS2 and their predicted RNA secondary structures. Mol.Phylogenet. Evol. 43, 774–786.

Auster, P.J., Moore, J., Heinonen, K.B., Watling, L., 2005. A habitat classificationscheme for seamount landscapes: assessing the functional role of deep-watercorals as fish habitat. In: Freiwald, A.R., Roberts, J.M. (Eds.), Cold-water Coralsand Ecosystems. Springer-Verlag, Berlin Heidelberg, pp. 761–769.

Baker, R.H., Yu, X.B., DeSalle, R., 1998. Assessing the relative contribution ofmolecular and morphological characters in simultaneous analysis trees. Mol.Phylogenet. Evol. 9, 427–436.

Bayer, F.M., 1956. Octocorallia. In: Moore, R.C. (Ed.), Treatise on InvertebratePaleontology. Part F Coelenterata. Geological Society of America and Universityof Kansas Press, Lawrence, Kansas, pp. 163–231.

Bayer, F.M., 1992. The Helioporacean Octocoral Ephiphaxum, recent and fossil: amonographic iconography. Stud. Tropical Oceanogr. 15, 1–76.

134 S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135

Bayer, F.M., 1993. Two new species of the gorgonacean genus Paragorgia(Coelenterata: Octocorallia). Precious Corals Octocoral Res. 2, 1–40.

Bayer, F.M., Cairns, S.D., 2003. A new genus of the Scleraxonian family Coralliidae(Octocorallia: Gorgonacea). Proc. Biol. Soc. Wash. 116, 222–228.

Bayer, F.M., Grasshoff, M., Verseveldt, J., 1983. IIllustrated Trilingual Glossary ofMorphological and Anatomical Terms Applied to Octocorallia. E. J. Brill, Leiden,The Netherlands.

Bensasson, D., Zhang, D., Hartl, D., Hewitt, G., 2001. Mitochondrial pseudogenes:evolution’s misplaced witnesses. Trends Ecol. Evol. 16, 314–321.

Berntson, E.A., Bayer, F.M., McArthur, A.G., France, S.C., 2001. Phylogeneticrelationships within the Octocorallia (Cnidaria: Anthozoa) based on nuclear18S rRNA sequences. Mar. Biol. 138, 235–246.

Brugler, M.R., France, S.C., 2008. The mitochondrial genome of a deep-sea bamboocoral (Cnidaria, Anthozoa, Octocorallia, Isididae): genome structure andputative origins of replication are not conserved among octocorals. J. Mol.Evol. 67, 125–136.

Buhl-Mortensen, L., Mortensen, P.B., 2004. Crustaceans associated with the deep-water gorgonian corals Paragorgia arborea (L., 1758), Primnoa resedaeformis(Gunn., 1763). J. Nat. History 38, 1233–1247.

Buhl-Mortensen, L., Mortensen, P.B., 2005. Distribution and diversity of speciesassociated with deep-sea gorgonian corals off Atlantic Canada. In: Freiwald,A.R., Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag,Berlin Heidelberg, pp. 849–879.

Carstens, B.C., Knowles, L.L., 2007. Estimating species phylogeny form gene-treeprobabilities despite incomplete lineage sorting: and example from melanoplusgrasshopers. Syst. Biol. 56.

Castro, C., Thiago, C.M., Medeiros, M.S., 2003. First record of the family Coralliidae(Cnidaria: Anthozoa: Octocorallia) from the western South Atlantic, with adescription of Corallium medea Bayer, 1964. Zootaxa 323, 1–8.

Coleman, A., 2003. ITS2 is a double-edged tool for eukaryote evolutionarycomparisons. Trends Genet. 19, 370–375.

Crowder, L.B., Hazen, E.L., Avissar, N., Bjorkland, R., Latanich, C., Ogburn, M.B., 2008.The impacts of fisheries on marine ecosystems and the transition to ecosystem-based management. Annu. Rev. Ecol., Evol., Syst. 39, 259–278.

Cummings, M.P., Otto, S.P., Wakeley, J., 1995. Sampling properties of DNA sequencedata in phylogenetic analysis. Mol. Biol. Evol. 12, 814–822.

Danovaro, R., Gambi, C., Dell’Anno, A., Corinaldesi, C., Fraschetti, S., Vanreusel, A.,Vincx, M., Gooday, J., 2008. Exponential decline of deep-sea ecosystemfunctioning linked to benthic biodiversity loss. Curr. Biol. 18, 1–8.

Davies, A.J., Roberts, J.M., Hall-Spencer, J., 2007. Preserving deep-sea naturalheritage: emerging issues in offshore conservation and management. Biol.Conserv. 138, 299–312.

DeVogelaere, A.P., Burton, E.J., Trejo, T., King, C.E., Clague, D.A., Tamburri, M.N.,Cailliet, G.M., Kochevar, R.E., Douros, W.J., 2005. Deep-sea corals and resourceprotection at the Davidson Seamount, California, USA. In: Freiwald, A.R.,Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag,Berlin Heidelberg, pp. 1189–1198.

Drummond, A.J., Ho, S.Y.W., Rawlence, N., Rambaut, A., 2007. A Rough Guide toBEAST 1.4.

Dubois, A., 2003. The relationships between taxonomy and conservation biology inthe century of extinctions. Comptes Rendus Biologies 326, S9–S21.

Edwards, S.V., 2008. Is a new and general theory of molecular systematicsemerging? Evolution 63, 1–19.

Edwards, S.V., Liu, L., Pearl, D.K., 2007. High-resolution species trees withoutconcatenation. Proc. National Acad. Sci. 104, 5936–5941.

Felsenstein, J., 2006. Accuracy of coalescent likelihood estimates: do we need moresites, more sequences, or more loci? Mol. Biol. Evol. 23, 691–700.

Felsenstein, J., Archie, J., Day, W.H.E., Maddison, W.P., Meacham, C., Rohlf, F.J.,Swofford, D.L., 1986. The Newick tree format. http://evolution.genetics.washington.edu/phylip/newicktree.html.

Forsman, Z., Barshis, D., Hunter, C., Toonen, R., 2009. Shape-shifting corals:molecular markers show morphology is evolutionarily plastic in Porites. BMCEvol. Biol. 9, 45.

France, S.C., Hoover, L.L., 2002. DNA sequences of the mitochondrial COI gene havelow levels of divergence among deep-sea octocorals (Cnidaria: Anthozoa).Hydrobiologia 471, 149–155.

France, S.C., Rosel, P.E., Agenbroad, J.E., Mullineaux, L.S., Kocher, T.D., 1996. DNAsequence variation of mitochondrial large-subunit rRNA provides support for atwo-subclass organization of the Anthozoa (Cnidaria). Mol. Mar. Biol.Biotechnol. 5, 15.

Grasshoff, M., 1979. Zur bipolaren verbreitung der oktokoralle Paragorgia arborea(Cnidaria: Anthozoa: Scleraxonia). Senckenb Marit 11, 115–137.

Graybeal, A., 1998. Is it better to add taxa or characters to a difficult problem? Syst.Biol. 47, 9–17.

Hedtke, S.M., Townsend, T.M., Hillis, D.M., 2006. Resolution of phylogenetic conflictin large data sets by increased taxon sampling. Syst. Biol. 55, 522–529.

Hillis, D.M., Pollock, D.D., McGuire, J.A., Zwickl, D.J., 2003. Is sparse taxon sampling aproblem for phylogenetic inference? Syst. Biol. 52, 124–146.

Huelsenbeck, J.P., Ronquist, F., 2001. MrBayes: Bayesian inference of phylogenetictrees. Bioinformatics (Oxford) 17, 754–755.

Huson, D.H., Rausch, D.C.R.C., Dezulian, T., Franz, M., Rupp, R., 2007. Dendroscope –an interactive viewer for large phylogenetic trees. BMC Bioinform. 8, 460.

Joly, S., McLenachan, P.A., Lockhart, P.J., 2009. A statistical approach fordistinguishing hybridization and incomplete lineage sorting. Am. Nat. 174,E54–E70.

Katoh, K., Misawa, K., Kuma, K., Miyata, T., 2002. MAFFT: a novel method for rapidmultiple sequence alignment based on fast Fourier transform. Nucl. Acids Res.30, 3059–3066.

Kim, E., Lasker, H.R., Coffroth, M.A., Kim, K., 2004. Morphological and geneticvariation across reef habitats in a broadcast-spawning octocoral. Hydrobiologia530 (531), 423–432.

Kluge, A.G., 2004. On total evidence: for the record. Cladistics 20, 205–207.Koslow, J.A., Gowlett-Holmes, K., Lowry, J.K., O’Hara, T., Poore, G.C.B., Williams, A.,

2001. Seamount benthic macrofauna off southern Tasmania: communitystructure and impacts of trawling. Mar. Ecol. Progr. Ser. 213, 111–125.

Kubatko, L.S., Degnan, J.H., 2007. Inconsistency of phylogenetic estimates fromconcatenated data under coalescence. Syst. Biol. 56, 17–24.

Lecointre, G., Philippe, H., Le, H.L.V., Le Guyader, H., 1993. Species sampling has amajor impact on phylogenetic inference. Mol. Phylogenet. Evol. 2, 205–224.

Leverette, T.L., Metaxas, A., 2005. Predicting habitat for two species of deep-watercoral on the Canadian Atlantic continental shelf and slope. In: Freiwald, A.R.,Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag, BerlinHeidelberg, pp. 467–479.

Lopez, J.V., Yuhki, N., Masuda, R., Modi, W., O’Brien, S.J., 1994. Numt, a recenttransfer and tandem amplification of mitochondrial DNA to the nuclear genomeof the domestic cat. J. Mol. Evol. 39, 174–190.

Maddison, D.R., Maddison, W.P., 2005. McClade: Analysis of Phylogeny andCharacter Evolution. Version 4.08. Sinauer Associates, Sunderland,Massachusetts.

Maddison, W.P., 1997. Gene trees in species trees. Syst. Biol. 46, 523–536.Marshall, D.C., Simon, C., Buckley, T.R., 2006. Accurate branch lengths estimation in

partitioned Bayesian analyses requires accommodation of among-partition ratevariation and attention to branch length priors. Syst. Biol. 55, 993–1003.

McFadden, C.S., France, S.C., Sánchez, J.A., Alderslade, P., 2006. A molecularphylogenetic analysis of the Octocorallia (Cnidaria: Anthozoa) based onmitochondrial protein-coding sequences. Mol. Phylogenet. Evol. 41, 513–527.

Messing, C.G., Neumann, A.C., Lang, J.C., 1990. Biozonation of deep-waterlithoherms and associated hardgrounds in the Northeastern Straits of Florida.Palaios 5, 15–33.

Metaxas, A., Davis, J., 2005. Megafauna associated with assemblages of deep-watergorgonian corals in Northeast Channel, off Nova Scotia, Canada. J. Mar. Biol.Assoc. U.K. 85, 1381–1390.

Morgan, L.E., Etnoyer, P., Scholz, A.J., Mertens, M., Powell, M., 2005. Conservationand management implications of deep-sea coral and fishing effort distributionsin the Northeast Pacific Ocean. In: Freiwald, A.R., Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag, Berlin Heidelberg, pp. 1171–1187.

Mortensen, P.B., Buhl-Mortensen, L., 2005. Deep-water corals and their habitats inThe Gully, a submarine canyon off Atlantic Canada. In: Freiwald, A.R., Roberts,J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag, BerlinHeidelberg, pp. 247–277.

Nedashkovskaya, O.I., Bum Kim, S., Lysenko, A.M., Frolova, G.M., Mikhailov, V.V.,Sook Bae, K., 2005. Bizionia paragorgiae gen. Nov., sp. nov., a novel member ofthe family Flavobacteriaceae isolated from the soft coral Paragorgia arborea. Int.J. Syst Evol. Microbiol. 55, 375–378.

Nylander, J.A., Wilgenbusch, J.C., Warren, D.L., Swofford, D.L., 2008. AWTY (are wethere yet?): a system for graphical exploration of MCMC convergence inBayesian phylogenetics. Bioinfomatics 24, 581–583.

Pollard, D.A., Iyer, V.N., Moses, A.M., Eisen, M.B., 2006. Widespread discordance ofgene trees with species tree in Drosophila: evidence for incomplete lineagesorting. PLoS Genet. 2, 1634–1647.

Pont-Kindon, G., Vassort, C.G., Rahul, W., Okimoto, R., Beagley, T., Wolstenholme,D.R., 2000. Mitochondrial DNA of hydra attenuata (Cnidaria): a sequence thatincludes an end of one linear molecule and the genes for l-rRNA, tRNA f-Met,tRNATrp, COII, and ATPase8. J. Mol. Evol. 51, 404–415.

Posada, D., 1998. Model test: testing the model of DNA substitution. Bioinformatics(Oxford) 14, 817–818.

Posada, D., Buckley, T., 2004. Model selection and model averaging inphylogenetics: advantages of Akaike information criterion and Bayesianapproaches over likelihood ratio tests. Syst. Biol. 793, 808.

Prada, C., Schizas, N., Yoshioka, P., 2008. Phenotypic plasticity or speciation? A casefrom a clonal marine organism. BMC Evol. Biol. 8, 47.

Rambaut, A., Drummond, A.J., 2007. Tracer v1.4, Available from: <http://beast.bio.ed.ac.uk/Tracer>.

Rasmussen, M.D., Kellis, M., 2007. Accurate gene-tree reconstruction by learninggene- and species-specific substitution rates across multiple completegenomes. Genome Res. 17, 1932–1942.

Roberts, C.M., 2002. Deep impact: the rising toll of fishing in the deep sea. TrendsEcol. Evol. 17, 242–245.

Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inferenceunder mixed models. Bioinformatics (Oxford) 19, 1572–1574.

Sánchez, J.A., 2005. Systematics of the bubblegum corals (Paragorgiidae:Octocorallia: Cnidaria) with description of new species from New Zealandand the Eastern Pacific. Zootaxa 1014, 1–72.

Sánchez, J.A., Aguilar, C., Dorado, D., Manrique, N., 2007. Phenotypic plasticity andmorphological integration in a marine modular invertebrate. BMC Evol. Biol. 7,122.

Sánchez, J.A., Cairns, S.D., 2004. An unusual new gorgonian coral (Anthozoa:Octocorallia) from the Aleutian Islands, Alaska. Zoologische Mededelingen 78,265–274.

S. Herrera et al. / Molecular Phylogenetics and Evolution 55 (2010) 123–135 135

Sánchez, J.A., Lasker, H.R., Taylor, D.J., 2003. Phylogenetic analyses among octocorals(Cnidaria): mitochondrial and nuclear DNA sequences (lsu-rRNA, 16S and ssu-rRNA, 18S) support two convergent clades of branching gorgonians. Mol.Phylogenet. Evol. 29, 31–42.

Seibel, P., Muller, T., Dandekar, T., Schultz, J., Wolf, M., 2006. 4SALE – a tool forsynchronous RNA sequence and secondary structure alignment and editing.BMC Bioinform. 7, 498.

Seibel, P., Muller, T., Dandekar, T., Wolf, M., 2008. Synchronous visual analysis andediting of RNA sequence and secondary structure alignments using 4SALE. BMCRes. Notes 1, 91.

Stamatakis, A., Hoover, P., Rougemont, J., 2008. A rapid bootstrap algorithm for theRAxML Web servers. Syst. Biol. 57, 758–771.

Strychar, K.B., Hamilton, L.C., Kenchington, E.L., Scott, D.B., 2005. Geneticcircumscription of deep-water coral species in Canada using 18S rRNA. In:Freiwald, A.R., Roberts, J.M. (Eds.), Cold-Water Corals and Ecosystems. Springer-Verlag, Berlin Heidelberg, pp. 679–690.

Swofford, D.L., 2002. PAUP*. Phylogenetic Analysis Using Parsimony (*and OtherMethods). Version 4. Sinauer Associates, Sunderland, Massachusetts.

Verseveldt, J., 1940. Studies on Octocorallia of the families Briareidae, Paragorgiidaeand Anthothelidae. Temminckia 4, 1–142.

Verseveldt, J., 1942. Further studies on Octocorallia. Zoologische Mededelingen 24,159–186.

Waller, R., Watling, L., Auster, P., Shank, T., 2007. Anthropogenic impacts on thecorner rise seamounts north-west Atlantic Ocean. J. Mar. Biol. Assoc. U.K. 87,1075–1076.

Watling, L., Norse, E.A., 1998. Disturbance of the seabed by mobile fishing gear: acomparison to forest clearcutting. Conserv. Biol. 12, 1180–1197.

West, J.M., Harvell, C.D., Walls, A.M., 1993. Morphological plasticity in a gorgoniancoral (Briareum asbestinum) over a depth cline. Mar. Ecol. Prog. Ser. 94, 61–69.

Zuker, M., 2003. Mfold web server for nucleic acid folding and hybridizationprediction. Nucl. Acids Res. 31, 3406–3415.

Zwickl, D.J., 2006. Genetic algorithm approaches for the phylogenetic analysis oflarge biological sequence datasets under the maximum likelihood criterion.Ph.D. dissertation. The University of Texas at Austin.