Spatial and temporal patterns of genetic variation in the widespread antitropical deep-sea coral Paragorgia arborea

Molecular Ecology (2012)

doi: 10.1111/mec.12074

Spatial and temporal patterns of genetic variation in the widespread antitropical deep-sea coral Paragorgia arborea ´ NCHEZ† S . H E R R E R A , * † T . M . S H A N K ‡ and J . A . S A *Massachusetts Institute of Technology, Woods Hole Oceanographic Institution, Joint Program in Oceanography, 266 Woods Hole Road, Woods Hole, MA 02543, USA, †Laboratorio de Biologia Molecular Marina (BIOMMAR), Departamento Ciencias Biologicas, Universidad de los Andes, Carrera 1E No 18A – 10, Bogota, Colombia, ‡Biology Department, Woods Hole Oceanographic Institution, 266 Woods Hole Road, Woods Hole, MA 02543, USA

Abstract Numerous deep-sea species have apparent widespread and discontinuous distributions. Many of these are important foundation species, structuring hard-bottom benthic ecosystems. Theoretically, differences in the genetic composition of their populations vary geographically and with depth. Previous studies have examined the genetic diversity of some of these taxa in a regional context, suggesting that genetic differentiation does not occur at scales of discrete features such as seamounts or canyons, but at larger scales (e.g. ocean basins). However, to date, few studies have evaluated such diversity throughout the known distribution of a putative deep-sea species. We utilized sequences from seven mitochondrial gene regions and nuclear genetic variants of the deep-sea coral Paragorgia arborea in a phylogeographic context to examine the global patterns of genetic variation and their possible correlation with the spatial variables of geographic position and depth. We also examined the compatibility of this morphospecies with the genealogical-phylospecies concept by examining specimens collected worldwide. We show that the morphospecies P. arborea can be defined as a genealogical-phylospecies, in contrast to the hypothesis that P. arborea represents a cryptic species complex. Genetic variation is correlated with geographic location at the basinscale level, but not with depth. Additionally, we present a phylogeographic hypothesis in which P. arborea originates from the North Pacific, followed by colonization of the Southern Hemisphere prior to migration to the North Atlantic. This hypothesis is consistent with the latest ocean circulation model for the Miocene. Keywords: coral, deep sea, DNA barcoding, phylogeography, species, widespread Received 1 June 2012; revision received 28 August 2012; accepted 1 September 2012

Introduction Several marine species, particularly from deep-sea environments, have apparent widespread yet discontinuous distributions (e.g. review by Roberts et al. 2009; Bik et al. 2012). Various mechanisms have been suggested to explain the apparent existence of such species, including recent connectivity among populations mediated by long-distance dispersal (Bucklin et al. 1987; France & Correspondence: Juan A. Sa´nchez, Fax: +57 1 3394949*2817; E-mail: [email protected] © 2012 Blackwell Publishing Ltd

Kocher 1996; Darling et al. 2000; Won et al. 2003; Pawlowski et al. 2007; Lecroq et al. 2009; Etter et al. 2011), large population sizes and similar selective pressures in a stable environment (Bisol et al. 1984; Brinkmeyer et al. 2003; Etter et al. 2011), relatively recent events of colonization mediated by jump dispersal over barriers (Darling et al. 2000; Etter et al. 2011), and cryptic speciation (France & Kocher 1996; Howell et al. 2004). A number of deep-sea coral morphospecies are among these widespread species with discontinuous distributions, for example Lophelia pertusa, Solenosmilia variabilis and Madrepora oculata (Roberts et al. 2009).

´ NCHEZ 2 S . H E R R E R A , T . M . S H A N K and J . A . S A Deep-sea corals are some of the most conspicuous invertebrate inhabitants of hard-bottom benthic environments worldwide. They are not only more diverse, in terms of number of species, than their shallow counterparts (Cairns 2007), but also they play a fundamental role as foundation species and ecosystem engineers, creating three-dimensional habitats that are occupied by a high diversity of associate species (BuhlMortensen & Mortensen 2005; Costello et al. 2005; Etnoyer & Morgan 2007; Buhl-Mortensen et al. 2010; Shank 2010). Coral ecosystems also support fisheries (D’Onghia et al. 2011; Soeffker et al. 2011) and have been identified as important sources of marine natural products (Leal et al. 2012). Deep-sea corals have evolved in a relatively stable and energy-poor environment; they tend to have slow growth rates (Roberts et al. 2009; Sun et al. 2010), great longevity (Roark et al. 2009) and size-dependent fecundity (Cordes et al. 2001). These characteristics make deep-sea coral ecosystems highly susceptible to disturbance events, especially those generated by human activities, that is, bottom-trawling, deep-sea mining, hydrocarbon extraction, waste disposal, climate change and ocean acidification (reviewed in Ramirez-Llodra et al. 2011). The characterization of spatial distribution patterns of genetic types is of fundamental importance to identify the factors that shape the ranges of deep-sea taxa, and that ultimately drive biodiversity patterns in the ocean (McClain & Mincks 2010). Widespread taxa thus can be used as models to understand how the effects of these factors operate at a global scale. Such information provides critical baseline data with which the potential effects of disturbances on populations inhabiting earth’s largest biome can be assessed. A handful of studies have examined the genetic diversity of deep-sea coral taxa in a regional context (Le Goff-Vitry et al. 2004; Smith et al. 2004; Thoma et al. 2009; Morrison et al. 2011). These have suggested that genetic differentiation does not seem to occur at small geographic scales often associated with discrete features such as individual seamounts or canyons, but presumably at larger scales, that is, broader oceanic regions. However, no studies to date have evaluated such hypotheses throughout the entire known distribution of a putative deep-sea coral species (see Pante & Watling 2012; for a comparison between two distant regions). In this study, we examined the spatial patterns of genetic variation in the widespread bubblegum coral Paragorgia arborea (Linnaeus, 1758) (Octocorallia: Paragorgiidae), which is one of the most prominent coral morphospecies in cold-water sublittoral and bathyal hard-substrate habitats. Paragorgia arborea plays an important ecological role generating microhabitats for numerous species; they are

the structural analog of large trees in a rain forest (Buhl-Mortensen & Mortensen 2005; Metaxas & Davis 2005; Watanabe et al. 2009; Buhl-Mortensen et al. 2010). Single colonies of P. arborea can harbour hundreds of individuals from dozens of associated species (e.g. ophiuroids, copepods, shrimp, anemones, polychaetes, ostracods, barnacles, amphipods, hydroids and foraminiferans) (Buhl-Mortensen et al. 2010). The fauna associated with this coral can be two to three times richer than the fauna associated with equivalent shallow-water tropical gorgonians (Buhl-Mortensen & Mortensen 2004, 2005). Paragorgia arborea has been reported from polar, subpolar and subtropical regions of all of the world’s oceans. This conspicuous and locally abundant species can grow massive colonies, which can reach up to 8 m in height (Sa´nchez 2005). Paragorgia arborea lives in regions of high productivity (Sarmiento & Gruber 2006 depth-integrated primary production > 10 mol/C/m2/ year) and high export fluxes (Sarmiento & Gruber 2006 particle export at 100 m > 2 mol/C/m2/year), water temperatures lower than 12 °C and relatively high local current velocities of 5–30 cm/s (Mortensen & Buhl-Mortensen 2004; Bryan & Metaxas 2006; Etnoyer & Morgan 2007; Roberts et al. 2009; Watanabe et al. 2009). The known distribution of P. arborea in the Northern Hemisphere includes numerous observations in both eastern and western North Atlantic waters and also in the eastern and western North Pacific (WNP), from Japan to the Aleutian Islands and to the Californian seamounts. In the Southern Hemisphere, it has been reported around the Crozet Islands, the Patagonian Shelf and the western South Pacific off New Zealand (Grasshoff 1979; Tendal 1992). Since the publication of these records, both fishing pressure and scientific research in the deep sea have increased significantly, and the number of new records for this species has increased in tandem. Some of these records can now be found in biodiversity databases such as the Ocean Biogeographic Information System (OBIS, http://www. iobis.org) and Global Biodiversity Information Facility (GBIF, http://data.gbif.org); however, many others remain unconsolidated in scattered publications and local databases. Thus, an updated picture of the global distribution of this species is in order. In this study, we provide an up-to-date summary of the global distribution of P. arborea and genetic insights into the global phylogeography of this species. By examining the genealogy of mitochondrial and nuclear genetic variants from specimens collected over nearly its entire known distribution, we tested the compatibility of the morphospecies P. arborea with the genealogical-phylospecies concept. We evaluated the hypothesis that the morphospecies P. arborea is a complex of © 2012 Blackwell Publishing Ltd

PHYLOGEOGRAPHY OF DEEP-SEA CORAL PARAGORGIA ARBOREA 3 cryptic species in a barcoding framework. We also examined the global patterns of genetic variation and their possible correlation with the spatial variables of geographic position and depth. We propose a scenario that could explain the observed evolutionary and present-day patterns in this and other species.

Methods Global distribution To illustrate the currently known global distribution of Paragorgia arborea, we plotted on a gridded geographic map all the unique records available to date from the databases of the OBIS, the GBIF, the Smithsonian Institution National Museum of Natural History (http://www.mnh.si.edu/rc/), the Yale University Peabody Museum of Natural History (www.peabody. yale.edu), the Harvard University Museum of Comparative Zoology (www.mcz.harvard.edu), the Muse´um National d’Histoire Naturelle France (www.mnhn.fr), the National Institute of Water & Atmospheric Research (www.niwa.cri.nz) and several local databases and publications (Bruntse & Tendal 2000; Wareham & Edinger 2007; Mortensen et al. 2008; Roberts et al. 2008; Hibberd & Moore 2009; Laptikhovsky 2011). Geographic coordinates were reconstructed using Google Earth for records with known collection locality, but no latitude and longitude information. Similarly, missing depth data were reconstructed using data from the global GEBCO_08 30 arc-second grid.

Molecular methods We analysed a total of 130 specimens of P. arborea available from various museum and laboratory collections (see Table S1, Supporting information). The examined material, collected since 1878, covers most to the known geographic distribution as well as the entire depth distribution of P. arborea (see Table S1, and Figs S1–S4, Supporting information for comparison). For more details on the sequencing of old specimens, see the Appendix S1 (Supporting information). Additional material from other paragorgiid morphospecies was included for comparisons. Total DNA was extracted from dry or ethanol-preserved (70–96%) samples using a CTAB–proteinase K–PCI protocol (Coffroth et al. 1992) or using an automated extraction system (AutoGenprep 965; AutoGen Inc.) as described in the study by Herrera et al. (2010). DNA was eluted in TE buffer and stored at 70 °C. Mutation rates in octocoral mitochondria are significantly lower than in most other organisms (Bilewitch & Degnan 2011). The implication of this lower mutation © 2012 Blackwell Publishing Ltd

rate is that mitochondrial markers in octocorals are useful to infer phylogeographic patterns and connectivity at broader spatial and temporal scales. Thus, to maximize the amount of variability captured from the genome of this organelle, we obtained sequences from seven gene regions, amplified by five primer pairs (Herrera et al. 2010), adding up to approximately 3000 base pairs (bp). These regions include the 3′-end of the NADH dehydrogenase subunit 6 (nad6), the nad6-nad3 intergenic spacer (int), the 5′-end of the NADH dehydrogenase subunit 3 (nad3), the 3′-end of the cytochrome c oxidase subunit I (cox1), the 5′-end of the DNA mismatch repair protein – mutS – homolog (mtMutS), two different regions of the large subunit ribosomal RNA (16S) and the 5′-end of the NADH dehydrogenase subunit 2 (nad2). We also sequenced the nuclear ribosomal internal transcribed spacer 2 (ITS2). In octocorals, ITS has been assessed in a number of groups providing enough resolution for diverse phylogenetic inferences (Alcyoniidae McFadden et al. 2001; McFadden & Hutchinson 2004; Nephtheidae van Ofwegen & Groenenberg 2007). ITS2 has also provided enough resolution for intraspecific and phylogeographic studies in Caribbean shallowwater octocorals (Sa´nchez et al. 2007; Gutie´rrez-Rodriguez et al. 2009). Furthermore, ITS2 has also provided valuable information for the analysis of genetic structure of deep-sea corals (Le Goff-Vitry et al. 2004; Miller et al. 2011). Polymerase chain reactions (PCR) and sequencing reactions for mitochondrial gene regions were performed following the protocols used by Herrera et al. (2010). ITS2 PCR amplicons, from a subset of 19 geographically representative individuals, were examined to assess the possibility of intragenomic variants through denaturing gradient gel electrophoresis (DGGE). Gels contained 8% polyacrylamide, 19 TAE buffer and a linear urea–formamide denaturing gradient from 45% to 80%. The gels were pre-ran at 60 °C and 90 V for 30 min, followed by 13 h at 60 °C and 90 V. Gels were stained with ethidium bromide for 15 min and visualized using a Bio-Rad Chemidoc system. PCR products from DGGE-excised bands were subsequently cleaned and sequenced. Complementary chromatograms were assembled and edited using the SEQUENTM CHER 4.8 software (Gene Codes Corp.). Sequences of each region were aligned independently using MAFFT 6.8 (Katoh et al. 2002). The G-INS-i and Q-INS-i algorithms (gap opening penalty = 1.53, offset value = 0.07) were employed for the protein coding and ribosomal regions, respectively. Secondary structures of ribosomal regions were inferred to improve the alignments, following the protocols used in the study by Herrera et al. (2010). To correct possible mistakes, all

´ NCHEZ 4 S . H E R R E R A , T . M . S H A N K and J . A . S A alignments of protein coding sequences were visually inspected and translated to amino acids in GENEIOUS 5.3 (Drummond et al. 2010), using the genetic code of Hydra attenuata (Pont-Kindon et al. 2000). No unusual stop codons or suspicious substitutions were identified, suggesting that no nuclear pseudogenes were sequenced (Lopez et al. 1994; Bensasson et al. 2001). Mitochondrial sequences were concatenated for each individual and treated as one single locus in most subsequent analyses, given that the mitochondrial genome is assumed to be nonrecombining. Mitochondrial and ITS2 genetic variants, with alignment gaps included as an informative state (Giribet & Wheeler 1999), were identified using DNASP 5.0 (Librado & Rozas 2009) and will be referred hereafter as haplotypes.

Gene trees and molecular clock To evaluate the compatibility of P. arborea with the functional definition of genealogical-phylospecies sensu De Queiroz (2007), that is, all alleles of a given locus in individuals of P. arborea being ‘descended from a common ancestral allele not shared with those of other species’ (Avise & Ball 1990; Baum & Shaw 1995), we performed independent phylogenetic analyses of the mitochondrial and ITS2 haplotypes. Homologous sequences from eight other paragorgiid morphospecies were included as outgroups (see Table S1, Supporting information). Phylogenetic estimation was performed using Bayesian inference (BI) in MRBAYES 3.12 (Huelsenbeck & Ronquist 2001; Ronquist & Huelsenbeck 2003) as implemented in the CIPRES portal (http://www.phylo. org). Most likely nucleotide substitution models were selected for each region based on the Akaike Information Criteria (AIC) as implemented in JMODELTEST 2.0. Models for the mitochondrial regions are shown in Table S2 (Supporting information). The general time reversible model with a gamma-distributed rate variation across sites (GTR+G) was selected for the ITS2. Default prior distribution settings were assumed for all parameters. Four independent analyses of 10 000 000 Monte Carlo Markov chain (MCMC) generations (94 chains) were run with a sampling frequency of 1000 generations (burn-in = 25%). Combined BI analysis of the mitochondrial locus was performed with explicit character partitions for each concatenated region, along with their independently selected models of evolution. To account for the rate variation among partitions (Marshall et al. 2006), we allowed the rates to vary under a flat Dirichlet prior distribution (ratepr = variable). The parameters of nucleotide frequencies, substitution rates, gamma shape and invariant site proportion were unlinked across partitions. MCMC runs were analysed in the program TRACER 1.5 (Rambaut & Drummond

2007). Convergence was indicated by the ‘straight hairy caterpillar’ (Drummond et al. 2007) shape of the stationary posterior-distribution trace (generations vs. log-likelihood) of each parameter. Other examined convergence and mixing diagnostics included the standard deviation of partition frequencies (200) and the similitude of posterior probabilities of specific nodes between different runs in the program AWTY (http://ceb.csit.fsu.edu/awty) (Nylander et al. 2008). High correlations between runs and no obvious trends in the split frequency plots were observed. Tree files for each run were combined, after burn-in, using the program LOGCOMBINER v1.7.1 (Drummond et al. 2012). The most probable trees were summarized into a maximum clade credibility tree using TREEANNOTATOR v1.7.1 (Drummond et al. 2012). A Bayesian-MCMC joint estimation of gene genealogy and divergence times was performed in BEAST 1.7.1 (Drummond et al. 2012) for the mitochondrial marker assuming the same substitution model mentioned above. We assumed an uncorrelated relaxed lognormal molecular clock model, which allows for the variation in mutation rates among branches, with the Yule model of constant speciation rate (Yule 1925; Gernhard 2008) and the coalescent model of constant population size (Kingman 1982), as the tree priors. Additional sequences from specimens of the sister family, Coralliidae, were added to estimate divergence time within the phylogeny as this family contains some of the few fossils available for Octocorallia. The coralliid node was calibrated implementing a normal prior distribution for the time to the most recent common ancestor (TMRCA) with a mean of 83.5 million years before present (Myr BP) and a standard deviation of 0.7, corresponding to Campanian age stratum, in which the oldest known fossil in this family has been found (Schlagintweit & Gawlick 2009). Three MCMC independent analyses were run for 30 000 000 generations with a sampling frequency of 3000 (burn-in = 25%). Convergence diagnostics (generations plot and EES) were also examined for the combined runs in TRACER 1.5 (Rambaut & Drummond 2007) as mentioned above. The most probable trees were summarized into a maximum clade credibility tree with median node heights using TREEANNOTATOR v1.7.1 (Drummond et al. 2012). To infer the historical patterns of dispersal in P. arborea, we used the Bayesian phylogeography framework proposed by Lemey et al. (2009), as implemented in BEAST 1.7.1 (Drummond et al. 2012). We mapped the geographic ocean region where each haplotype was sampled to the time-scaled mitochondrial gene genealogy, which was inferred with the assumption of an uncorrelated relaxed lognormal molecular © 2012 Blackwell Publishing Ltd

PHYLOGEOGRAPHY OF DEEP-SEA CORAL PARAGORGIA ARBOREA 5 clock and the coalescent constant-population-size tree prior, as explained above. This framework allows the reconstruction of discrete states of geographic location for ancestral nodes by posterior probability estimation.

Barcoding and species delimitation To test for the possibility of cryptic species in the morphospecies P. arborea, we calculated pairwise uncorrected distances among individuals of P. arborea and other paragorgiid morphospecies for the mtMutS and cox1 sequences, as proposed by McFadden et al. (2011), in PAUP* 4.0b10 (Swofford 2002). Neighbour-joining trees were built using the calculated distances. We also examined the ITS2 secondary structures for the presence of compensatory base changes (CBCs) using the visualization program 4SALE (Seibel et al. 2006); CBCs are altered pairings in a helix of the secondary structure of the ITS2 RNA transcript, and empirical work has suggested that they could be used as indicators of species boundaries in most metazoans (Muller et al. 2007; Coleman 2009). We also used the coalescent-based species delimitation method described by Pons et al. (2006) and Monaghan et al. (2009), as implemented in the SPLITS R-package (available from http://r-forge.r-project.org/ projects/splits/). This likelihood method is based on a general mixed Yule-coalescent (GMYC) model, which estimates phylospecies boundaries in a clock-constrained calibrated tree by identifying increases in branching rates (looking forward in time). Such increases are assumed to be characteristic of transition points between interspecific speciation–extinction processes and intraspecific coalescent processes, that is, populations (Pons et al. 2006; Monaghan et al. 2009). Single- and multiple-threshold models with explicit and upper and lower limits for the estimation of scaling parameters (0 and 10, respectively) were used in the analysis of the time-calibrated trees obtained with the Yule and coalescent models tree priors.

Genetic variability Genetic variability among individuals and populations was measured for each locus according to the haplotype diversity (h) and genetic diversity (average number of pairwise differences hp) indices (Tajima 1983) using ARLEQUIN 3.5 (Excoffier & Lischer 2010). Fu’s Fs statistic was calculated to determine whether the observed pattern of polymorphism was consistent with a neutral model of evolution (Tajima 1989; Fu 1997). Global FST statistics were calculated to evaluate for possible differentiation in the genetic composition among populations worldwide. Pairwise comparisons of population differ© 2012 Blackwell Publishing Ltd

entiation were made in Arlequin and significance values estimated after 1000 permutations. To visualize the spatial patterns of genetic variation for each marker, the specimens were colour-coded according to haplotype, their geographic collection coordinates were plotted using IMAP v3.5 (Biovolution), and their collection depth was plotted on an X-Y scatter plot. To assess the amount of variability in the populations of P. arborea that was represented in our samples, we generated haplotype accumulation curves (Gotelli & Colwell 2001) by calculating estimates of the mean and variance for the number of accumulated haplotypes through 1000 random permutations, using the program R-package SPIDER v1.1 (Brown et al. 2012).

Results Global distribution A total of 341 high-confidence geographic location records of Paragorgia arborea were gathered (see Fig. S3, Supporting information). Paragorgia arborea is an antitropical taxon, occupying a band between 30° and 70° degrees of latitude in both hemispheres. These bands are, in general, areas of high surface primary productivity and export (Sarmiento & Gruber 2006). Most records of P. arborea are from depths shallower than 1000 m, indicating a preference for upper-bathyal environments. Despite the fact that other coral species that share part of their ranges with P. arborea have been commonly observed in tropical and subtropical regions (e.g. Lophelia pertusa and Madrepora oculata), P. arborea has never been found in these areas. This suggests that the currently known distribution of P. arborea is not a result of undersampling at lower latitudes.

Molecular data A total of 92 specimens were positively screened for the mitochondrial marker. The concatenated mitochondrial alignment for the morphospecies P. arborea had a length of 2922 bp, of which 2881 were invariable sites (pairwise identity of 99.7%); eight sites were parsimonyinformative. The mean ungapped sequence length was 2917.9 bp (SD = 2.8 bp), with a range of 2910 and 2921 bp. The G-C content was 39.5%. The only noncoding region in the mitochondrial locus data set, the nad6nad3 intergenic spacer (int), contained one indel, but no nucleotide substitutions. The ITS2 was successfully sequenced for 48 specimens, of which 83% overlapped with the mitochondrial set. The ITS2 alignment had a length of 312 bp, of which 301 were invariable (pairwise identity of 99.2%); 35 sites were parsimony-informative. The mean ungapped sequence length for ITS2 was

´ NCHEZ 6 S . H E R R E R A , T . M . S H A N K and J . A . S A 282 bp (SD 0.9 bp), ranging between 280 and 284 bp. The G-C content was 41.1%. No intragenomic variability was revealed, using DGGE, in the ITS2. The predicted secondary structure of ITS2 showed the characteristic shape of a helicoidal ring with four helixes (Coleman 2007); stems III and IV were particularly long in this species (Fig. S5, Supporting information). The number of haplotypes and genetic diversity estimates for each population with both loci are shown in Figs 1 and 2. The nad6-int-nad3 region (hereafter referred as nad6 for simplicity) contained most of the variable sites (21) and the greatest number of haplotypes (11) found in the individuals with complete mitochondrial data sets. Haplotype differences were also located in the ITS2: one at helix I, three at helix III and five at helix IV; the remaining two were free nucleotides at the structure main ring (Fig. S5, Supporting information).

Gene genealogies and phylogeographic history Individual mitochondrial gene trees were largely congruent, although resolution was generally low (Fig. S6, Supporting information). The inferred phylogeny based on each independent loci (i.e. concatenated mitochondrial and ITS2) highly supported the monophyly of P. arborea (Fig. 3) and had much greater clade resolution. Both Bayesian and neighbour-joining analyses inferred the same evolutionary relationships. Branch lengths were appreciably shorter within the clade of P. arborea, when compared to the ones among morphospecies. Relationships within Paragorgia were not fully resolved, particularly among Paragorgia wahine, Paragor-

gia yutlinux and Paragorgia sp. 1. The systematic relationships of Paragorgia spp. are outside of the scope of this study and will not be further discussed. The time-scaled trees estimated assuming the coalescent model of constant population size had, in general, shorter shallower and longer deeper branches than the tree estimated assuming the Yule model of constant speciation rate (Figs S7 and S8, Supporting information). Consequently, the time to the TMRCA of the genus Paragorgia was estimated to be 61 Myr BP (95% CI: 31–101) under the coalescent model and 54 Myr BP (95% CI: 41–94) under the Yule model. The TMRCA of P. arborea based on the coalescent model and the Yule model was 10.1 Myr BP (95% CI: 4.4–18.8) and 14.1 Myr BP (95% CI: 6.7–26.3), respectively. The Bayesian phylogeographic analysis indicates that the lineage of P. arborea likely originated in the North Pacific (posterior probability 0.38, see Fig. 4). Dispersal to the South Pacific and subsequent colonization of the North Atlantic likely occurred between the midMiocene and early Pliocene.

Genetic distances, CBCs and GMYC Maximum uncorrected genetic distances among conspecifics and minimum uncorrected genetic distances among congeners were used to measure the intraspecific and interspecific variation, respectively, of mtMutS and cox1 sequences as in the study by McFadden et al. (2011). The maximum distances within P. arborea, 0.3% for mtMutS and 0.9% for cox1, were in general smaller than the distances among morphospecies of Paragorgia, Fig. 1 The global geographic distribution of mitochondrial haplotypes in Paragorgia arborea. The gene tree in the centre of the figure shows the inferred relationships among haplotypes. Each haplotype is indicated by a different colour. Framed circles represent individuals. Pie charts indicate the frequency of haplotypes in each population (global region): North Atlantic (NA), South Atlantic (SA), South Pacific (SP), western North Pacific (WNP) and eastern North Pacific (ENP). The size of each pie is proportional to the number of samples from each population (n). The number of haplotypes (H), haplotype diversity (h), genetic diversity (hp) and the Fu’s Fs statistic are also indicated.

© 2012 Blackwell Publishing Ltd

PHYLOGEOGRAPHY OF DEEP-SEA CORAL PARAGORGIA ARBOREA 7 Fig. 2 The global geographic distribution of nuclear internal transcribed spacer 2 haplotypes in Paragorgia arborea. The gene tree in the centre of the figure shows the inferred relationships among haplotypes. Each haplotype is indicated by a different colour. Framed circles represent individuals. Pie charts indicate the frequency of haplotypes in each population: North Atlantic (NA), South Indian Ocean (SI), South Pacific (SP), western North Pacific (WNP) and eastern North Pacific (ENP). The size of each pie is proportional to the number of samples from each population (n). The number of haplotypes (H), haplotype diversity (h), genetic diversity (hp) and the Fu’s Fs statistic are also indicated.

NP NP NP

NP NP

NP

NP NP SP SP

NP SP SA NA

SP SP NA NA NA

Time (million years)

Fig. 3 Unrooted gene tree hypotheses for the mitochondrial (top) and nuclear internal transcribed spacer 2 (bottom) markers in Paragorgia.

which ranged between 0.5–6.1% and 0–2.3% for mtMutS and for cox1, respectively (Table S3, Supporting information). The minimum distances between P. arborea © 2012 Blackwell Publishing Ltd

Fig. 4 Maximum clade credibility ultrametric time-scaled mitochondrial gene tree for Paragorgia arborea. Branch colours show the most probable location states: North Atlantic (NA) in blue, South Pacific (SP) in green, South Atlantic (SA) in violet and North Pacific (NP) in orange. Pie charts show the posterior probabilities of location states for each ancestral node (total pie area = 1). The most probable location state of each node is also indicated.

and the other morphospecies of Paragorgia ranged between 1.8–5.1% for mtMutS and 0.6–1.8% for cox1. A similar pattern was observed for ITS2 distances, 0.6% in

´ NCHEZ 8 S . H E R R E R A , T . M . S H A N K and J . A . S A intraspecific comparisons within P. arborea (maximum distances) and 3.3–7.4% in interspecific comparisons among morphospecies of Paragorgia. The minimum distances between P. arborea and the other morphospecies ranged between 3.3 and 5.6%. The predicted ITS2 secondary structure was essentially the same for all haplotypes of P. arborea and no CBCs or hemi-CBCs were observed. A transition point between species- and populationlevel branching patterns was identified, by the singlethreshold GMYC method, at ca. 21 Myr BP for the time-scaled mitochondrial gene genealogy estimated with the Yule model tree prior (Fig. S7, Supporting information). For the time-scaled gene genealogy estimated with the coalescent model tree prior, this transition point was inferred to be at ca. 16 Myr BP (Fig. S8, Supporting information). In both cases, the GMYC model showed a marginally significant (i.e. a = 0.05) better fit to the data than the null model of uniform coalescent branching rates (LR = 4.84, d.f. = 3, P = 0.18, compared to LR = 5.84, d.f. = 3, P = 0.12, respectively). The implementation of a multiple-threshold GMYC model did not yield a significantly (i.e. a = 0.05) better fit than the single-threshold GMYC for either case (v2 = 2.55, d.f. = 3, P = 0.47 and v2 = 0.83, d.f. = 3, P = 0.84, respectively).

Spatial patterns of genetic variation Overall, in P. arborea, 16 haplotypes were defined based on the mitochondrial locus and 11 based on the ITS2. The genealogical relationships among haplotypes inferred by the Bayesian inference did not reveal reciprocal monophyly of the specimens from Northern and Southern Hemispheres, or from different oceans, for example Pacific vs. Atlantic (Figs 1 and 2). In fact, a number of haplotypes were shared across large geographic spans. Genetic diversity, as measured by the haplotype diversity (h, the probability that two randomly chosen haplotypes are different in a population sample) and the average number of nucleotide differences between all pairs of haplotypes in the sample (hp), was highest in the western North Pacific Ocean (mitochondrial h = 0.74, SD = 0.11, and hp = 7.44, SD = 0.11; nuclear h = 0.75, SD = 0.14, and hp = 2.68, SD = 1.81) and South Pacific Ocean (SP; mitochondrial h = 0.79, SD = 0.03, and hp = 6.37, SD = 3.10; nuclear h = 0.73, SD = 0.09, and hp = 1.64, SD = 1.15) regions, intermediate in the North Atlantic Ocean (NA; mitochondrial h = 0.26, SD = 0.12, and hp = 0.46, SD = 0.42; nuclear h = 0.71, SD = 0.11, and hp = 1.02, SD = 0.82) and lowest in the eastern North Pacific Ocean (ENP; mitochondrial hp = 0.00; nuclear h = 0.49, SD = 0.18, and hp = 1.16,

SD = 0.91) (Figs 1 and 2). Overall, the relative levels of genetic diversity for both loci were highly similar (see Figs 1 and 2). No significant deviations from neutrality were found in either locus (i.e. a = 0.05). Mitochondrial haplotypes m15 and m12 were shared between the NA and SP regions, although their frequencies were dissimilar (Fig. 1). Haplotype m15 was the dominant form in the NA, with a frequency of 0.86, whereas only one specimen was found having the m12 variant. In the SP, these haplotypes represented two of the three most common ones, with frequencies of 0.26 for m15 and 0.31 for m12. The rest of the haplotypes in these regions represented private alleles, that is, variants exclusive to a particular area. The haplotype from the single specimen from the South Atlantic Ocean (SA) region was a private allele. No haplotypes from the North Pacific Ocean (NP) were shared with other regions. Within the NP region, there was a clear break between western and eastern subregions, separated by the Alaska Peninsula. All haplotypes within these subregions represented private alleles. Two dominant haplotypes were found in the WNP, m2 and m7, with frequencies of 0.5 and 0.21, respectively. In the ENP, there was a single haplotype. In the ITS2 data set, regional differences were less pronounced than in the mitochondrial data set but, similarly, the haplotype frequencies varied greatly among regions (Fig. 2). Haplotypes i1 and i6 had near-cosmopolitan distributions. Haplotype i1 was found within all regions with frequencies of 0.23 (NA), 0.67 [South Indian Ocean (SI)], 0.07 (SP), 0.25 (WNP) and 0.09 (ENP). Haplotype i6 was found in specimens from the NA, SP and ENP, with frequencies of 0.15, 0.07 and 0.73, respectively. Haplotype i2 was shared between NA and SP, being dominant in both regions, with frequencies of 0.54 and 0.43, respectively. Lastly, haplotype i5 was found in both the NA (frequency = 0.09) and the SI (frequency = 0.33). Private alleles were found in the Pacific, with high frequencies in the SP (i7, frequency = 0.36) and the NWP (i8, frequency = 0.5), but none were found in the NA or SI. No detectable differentiation in the haplotype distributions could be explained by depth differences (Fig. 5). Haplotype accumulation curves revealed that the global mitochondrial and nuclear ITS2 diversities have not been fully sampled, as indicated by the steep slopes of both lines (Fig. S9, Supporting information). Both markers showed similar levels of diversity at given sampling efforts, ITS2 being slightly lower than the mitochondrial. When the individual genes in the mitochondrial data set were examined individually, it was clear that there are significant differences in their contributions to overall diversity estimates and that a single one does not capture the diversity found in the combined mitochondrial marker. By far, the gene that © 2012 Blackwell Publishing Ltd

0

0

500

500

Depth (m)

Depth (m)

PHYLOGEOGRAPHY OF DEEP-SEA CORAL PARAGORGIA ARBOREA 9

1000

1500

1000

1500 1 2 3

5 6 7 8 9 10 11 12 13 14 15 16 Mitochondrial haplotype

1

2

3

4

5 6 7 8 ITS2 haplotype

9

10 11

Fig. 5 The distribution of mitochondrial (left) and nuclear internal transcribed spacer 2 ITS2 (right) haplotypes of Paragorgia arborea with depth. Individuals are represented by dots. Each haplotype is indicated by a different colour, as in Figs 1 and 2. The prefix m denotes mitochondrial haplotypes, and the prefix i denotes nuclear ITS2 haplotypes (haplotype numbers are equivalent to the ones in Table S1, Supporting information).

captures the largest mitochondrial diversity in terms of haplotypes in P. arborea is nad6 (11), followed by nad2 (6), mtMutS (5), 16S (4) and cox1 (3). The combination of nad6 and 16S captures 14 haplotypes, and the addition of nad2 to these two regions captures all 16 haplotypes found in the combined mitochondrial marker.

Discussion The data and analyses generated in this study showed that the morphospecies Paragorgia arborea can be defined as a genealogical-phylospecies, in contrast to the hypothesis that P. arborea represents a cryptic species complex. Genetic variation in this lineage is correlated with geographic location at the basin-scale level, but not with depth. We present a phylogeographic hypothesis for P. arborea in which this independently evolving lineage originates from the North Pacific, followed by colonization of the Southern Hemisphere prior to migration to the North Atlantic. We argue that this hypothesis is consistent with the latest ocean circulation model for the Miocene.

A globally distinct evolving lineage? The distinction among species, incipient species and structured populations in many deep-sea invertebrates remains contentious due to the difficulties in defining the species boundaries for certain groups and to the paucity of the genetic, ecological and taxonomic data available to date (Vrijenhoek 2009; McFadden et al. 2011). Commonly used species concepts (e.g. biological species concept) have been traditionally developed in terrestrial models. However, the biological and ecological information required to apply such concepts to deepsea organisms (e.g. reproductive success, behaviour) is, at this time, impracticable to obtain. It is now recognized that a combination of morphologic and phylogenetic criterions is most practical to discern among deep-sea coral species (e.g. Herrera et al. 2010; Pante & Watling 2012). © 2012 Blackwell Publishing Ltd

Here, we examined, for the first time, the compatibility of traditional taxonomical identifications and molecular information in a putative deep-sea coral species at a global scale. The mitochondrial and nuclear gene trees have congruent topologies, showing that alleles of P. arborea have a common ancestor not shared with other paragorgiid morphospecies. Thus, the morphospecies P. arborea is compatible with the genealogicalphylospecies concept. Furthermore, the branch lengths among haplotypes of P. arborea are much shorter than the branches among other putative species, which is to be expected for genetic variability within a phylospecies. The only consistent morphological variant corresponds to the populations found in the NP, which were previously referred as Paragorgia pacifica Verrill, 1922 but synonymized with P. arborea by Grasshoff (1979). Individuals from these populations seem to have reduced sclerite size and ornamentation when compared to the characteristics that defined P. arborea prior to Grasshoff synonymizing the two (Sa´nchez 2005). Our data do not support P. pacifica as a valid species (see discussion below); we rather suggest that it may represent a subspecies. Taken together, this evidence indicates that P. arborea is a globally distinct lineage, implying that identifications based on morphology can accurately distinguish this taxon.

A complex of cryptic species? The presence of cryptic species complexes has been detected in various presumed widespread marine morphospecies (e.g. clams Goffredi et al. 2003; isopods Raupach et al. 2007; limpets Johnson et al. 2008; gastropods Duda et al. 2009; Vrijenhoek 2009). Despite being morphological indistinguishable, cryptic species have been detected through molecular data, on the basis of genetic dissimilarity. Here, we tested for the possibility of cryptic species within the specimens of P. arborea, by analysing the pairwise uncorrected genetic distances among haplotypes using a DNA barcoding framework

´ NCHEZ 10 S . H E R R E R A , T . M . S H A N K and J . A . S A based on the cox1 and mtMutS gene regions (see McFadden et al. 2011). Under this framework, pairwise uncorrected distances greater than 1% for mtMutS or cox1 genes can be confidently used to indicate cryptic species (McFadden et al. 2011). Based on this threshold, the maximum intraclade distances (0.3% for mtMutS and 0.9% for cox1) among specimens of P. arborea do not suggest the presence of cryptic species. These distances are consistent with the intraspecific distances found within other paragorgiid species, for example 0.8% for mtMutS in Sibogagorgia cauliflora (Herrera et al. 2010). However, the suggested threshold is unidirectional, meaning that distances smaller than 1% do not imply the absence of species boundaries. Additional genetic and biological data are needed to test for this possibility. The uniformity of predicted ITS2 secondary structures and the absence of CBCs or hemi-CBCs are also consistent with intraspecific levels of variation (Muller et al. 2007; Coleman 2009; Ruhl et al. 2010). However, similar to the mitochondrial barcoding threshold discussed above, this criterion is also unidirectional; thus, the absence of cryptic species is not implied (Coleman 2009). Lastly, the branching transition points inferred by the GMYC likelihood method indicate that the lineage of P. arborea is independently evolving with a branching pattern characteristic of a population-level coalescent process (Figs S7 and S8, Supporting information). In summary, the levels and patterns of genetic variability in mitochondrial and ITS2 loci do not provide actual evidence for cryptic species boundaries within P. arborea.

Global patterns of genetic variation Genetic diversity in P. arborea is not randomly distributed, as it would be expected under a scenario of global panmixia. It is highest at the ENP and SP populations and lowest at WNP and NA populations, as indicated by the haplotype diversity and the average pairwise differences among alleles (Figs 1 and 2). The significantly high global FST value (0.61 for the mitochondrial locus and 0.39 for the ITS2, see Table 1) indicates that there are significant differences in the genetic composition among worldwide populations, when defined at the basin/regional scale. This is consistent with the results from other studies of deep-sea corals (e.g. Smith et al. 2004; Thoma et al. 2009; Miller et al. 2011; Morrison et al. 2011). However, to test over scales smaller than regional for genetic structuring, it will be necessary to examine larger numbers of independent, highly variable markers (e.g. Le Goff-Vitry et al. 2004 and Morrison et al. 2011). Regional geographic differences were sorted out by comparing the FST values of genetic differentiation among populations (Table 1). The pairwise differences

Table 1 Global and pairwise FST values for the mitochondrial (top) and nuclear ITS2 (bottom) markers among populations of Paragorgia arborea Mitochondrial FST global

NA SP WNP ENP Nuclear ITS2 FST global

NA SP WNP ENP SI

0.61 NA

SP

WNP

0.27 0.67 0.98

0.39 0.67

0.74

ENP

0.39 NA

SP

WNP

ENP

0.16 0.47 0.27 0.13

0.51 0.43 0.31

0.48 0.24

0.26

SI

NA, North Atlantic; SA, South Atlantic; SP, South Pacific; WNP, western North Pacific; ENP, eastern North Pacific; SI, South Indian Ocean; ITS2, internal transcribed spacer 2. All values are significant (i.e. a = 0.05).

among populations for both markers suggest strong differentiation between the EPN population and all the other populations, including the neighbouring WNP. There is also a significant break between North and South Pacific populations. South Pacific and North Atlantic populations are the less dissimilar, which suggests a more recent connection between them. Gene genealogies of P. arborea showed no reciprocal monophyly of alleles among populations. Two nonexclusive and equally plausible mechanisms could have lead to this observed pattern: (i) gene flow between populations for which recent connectivity could be conceived given a temporal continuity of favourable environmental conditions, and (ii) incomplete lineage sorting caused by a rapid succession of divergence events among populations, combined with large ancestral effective population sizes (Maddison 1997; Edwards 2009). The earliest, divergent lineage of alleles as well as the highest genetic diversity was found in the WNP, which lends support to the idea that P. arborea originated in this region. The nuclear ITS2 showed signs of lower genetic differentiation among populations than the mitochondrial locus. The effects of differing effective population sizes on processes such as genetic drift and genetic sweeps could explain this difference given that the mitochondrial genome has one quarter the effective population size of the nuclear genome (given that it is haploid and assuming maternal inheritance only). Similarly, the nuclear gene tree had much lower resolution compared © 2012 Blackwell Publishing Ltd

P H Y L O G E O G R A P H Y O F D E E P - S E A C O R A L P A R A G O R G I A A R B O R E A 11 to the mitochondrial one, which is likely due to the smaller number of phylogenetically informative sites present in the short ITS2 sequence. In contrast to patterns observed in other deep-sea organisms (Cho & Shank 2010; Etter et al. 2011; Miller et al. 2011), depth does not appear to be an important large-scale structuring factor in populations of P. arborea. This is perhaps not surprising given the widespread distribution of this organism, which suggests that it is capable of living under a relatively broad range of conditions. Alternatively, as mentioned above, small-scale genetic structuring related to depth could be revealed with higher-resolution markers. The mitochondrial nad6 gene contained the greatest amount diversity in terms of haplotypes in this data set, that is, was the most variable mitochondrial marker. This result contrasts with previous studies, in which the mtMutS gene has been found to be significantly more variable than any other mitochondrial gene region (France & Hoover 2001; McFadden et al. 2004, 2010; Herrera et al. 2010). We suggest that the levels of variation among different mitochondrial gene regions in octocorals vary among taxa (see McFadden et al. 2010, 2011; Bilewitch & Degnan 2011), and thus, there is not a single universal region that provides the largest amount of variability. For the samples of P. arborea examined here, we found that the combination of nad6 + 16S +nad2 is the most informative. The nuclear ITS2 still seems to be a good cost-effective alternative to detect genetic variation among individuals, in the absence of intragenomic variants.

Phylogeographic hypothesis Here, we suggest a phylogeographic scenario in which P. arborea originated in the North Pacific, possibly in the WNP followed by colonization of the South Pacific and spreading eastward around the Southern Hemisphere in a stepping stone fashion (possibly via the Antarctic Circumpolar Current). The colonization of the North Atlantic seems to have occurred through a more recent dispersal event from the South Pacific, via the Central American Seaway, or from the SA. Similarities between other deep-sea coral taxa from the South Pacific and the North Atlantic have been independently observed (Thoma et al. 2009; Pante & Watling 2012), which gives support to the idea of a more recent connection between South Pacific and North Atlantic deep-sea communities. This scenario is an alternative to the trans-Arctic interchange hypothesis (Vermeij 1991), which suggests a recent North Pacific and North Atlantic connection as indicated by the distributions of several shallow-water taxa (e.g. red algae Vanoppen et al. 1995; asteroids, bivalves, gastropods, barnacles © 2012 Blackwell Publishing Ltd

Wares & Cunningham 2001; seagrass Olsen et al. 2004; cnidarians Govindarajan et al. 2005). Paragorgia arborea shares a similar phylogeographic history and genetic diversity patterns with the spiny dogfish Squalus acanthias (Verissimo et al. 2010) and the bryozoan Membranipora membranacea (Schwaninger 2008), both of which have modern antitropical distributions. The time of divergence between the WNP and South Pacific populations of the spiny dogfish has been estimated to be around 7.8 Myr BP and approximately 13.3 Myr BP (9.9–21.9) for the bryozoan, which is comparable to our estimates of 4.5 Myr BP (95% CI = 2.0– 8.3 using the coalescent model) and 8.1 Myr BP (95% CI = 3.6–15.3 using the Yule model) for P. arborea (Fig. 4, Figs S7 and S8, Supporting information). The timing of colonization of the North Atlantic has been estimated to be between 3.6 and 5.3 Myr BP for the dogfish, 6.2 Myr BP (95% CI = 4.6–10.2) for the bryozoan and between 1.7 Myr BP (95% CI = 0.4–3.6) and 4.0 Myr BP (95% CI = 1.3–8.3) for P. arborea. The similar and independently estimated times for these events give support to the idea that a common set of oceanographic conditions in the Miocene and early Pliocene lead to the current distributions of these species. The latest Miocene ocean circulation models (Butzin et al. 2011) indicate that there was a dominant southward horizontal flow that carried deep waters from the WNP to the South Pacific, passing along the eastern side of the New Zealand landmass, during the mid- to late Miocene (~5–15 Myr BP). This flow decreased during the late Miocene. The Antarctic Circumpolar Current started to develop during the mid- to late Eocene (ca. 37 –40 Myr BP) (Scher 2006) and thus was already well established as the dominant feature of ocean circulation during the Miocene, transporting massive amounts of water eastward. At the same time, during the mid-Miocene, the deep-water formation in the North Atlantic and its southward flow were absent or weak, likely due to the dominant barotropic water flux from the Pacific to the Atlantic. The formation of deep water in this time period mainly took place in the Southern Ocean. Deepwater formation in the North Atlantic and the dominant southward flow, as we know them today, were later established during the late Miocene as the Central American Seaway closed (Butzin et al. 2011). Evolutionary migrations inferred from genetic diversity patterns presented here for P. arborea are consistent with this history of ocean circulation. Historical changes in the global patterns of ocean circulation and climate may have caused shifts in the habitat and thus the distribution of P. arborea. Widespread ocean cooling during glacial periods in the late Miocene–early Pliocene (Mercer & Sutter 1982) and throughout the Quaternary (Ehlers et al. 2011) could have aided the trans-equatorial

´ NCHEZ 12 S . H E R R E R A , T . M . S H A N K and J . A . S A exchange by increasing the area of suitable habitat for stepping stone populations towards the tropics (McIntyre et al. 1989). Isolated relict low-latitude populations might still exist. We hypothesize that the described set of conditions could explain the current distribution patterns of many other marine taxa (e.g. deep-sea coral associates, such as ophiuroids and chirostylid crabs) and thus might have played an important role shaping extant deep-sea faunal diversity.

Acknowledgements Support for this study was generously provided by a mini-grant from 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 los Andes to J.A.S, the National Systematics Laboratory of NOAA’s National Marine Fisheries Service, a Smithsonian Graduate Student Fellowship to S.H., an award from the Systematics Research Fund of the Systematics Association and the Linnean Society of London to S.H., and a Grantin-Aid of Research from the Sigma Xi Research Society to S.H. We are especially thankful to S.D. Cairns, A.G. Collins, C.L. Agudelo, N. Ardila, L. Duen˜as, A. Ormos, J. Hunt, L. Weigt, L. Monroy, M. Herrera and M. Sangrey for their generous support, assistance and advise. Laboratory work was performed at the Laboratories of Analytical Biology NMNH, Smithsonian Institution and BIOMMAR, Universidad de los Andes. Samples were generously provided by P. Alderslade (CSIRO), A. Andouche (MNHN), A. Andrews (MLML), A. Baco (FSU), A. Baldinger (MCZ), J. A. Boutillier (DFO), S.D. Cairns (USNM), S. Davies (DFO), M. Eriksson (UUZM), Y. Imahara (WPMNH), D. Janussen (SMF), E. Lazo-Wasem (YPM), P. Lozouet (MNHN), L. Lundsten (MBARI), S. Mills (NIWA), K. Schnabel (NIWA), and B. Stone (NOAA), D. Tracey (NIWA), and R. Weber (Te Papa Tongarewa). We also thank J. McDermott, N. Roterman and C. Munro for their comments on earlier versions of the manuscript. We are grateful for the helpful input from the editor and two anonymous reviewers.

References Avise JC, Ball RM (1990) Principles of genealogical concordance in species concepts and biological taxonomy. Oxford Surveys in Evolutionary Biology, 7, 45–67. Baum DA, Shaw KL (1995) Genealogical perspectives on the species problem. In: Experimental and Molecular Approaches to Plant Biosystematics (eds Hoch PC, Stephenson AG), pp. 289– 303. Missouri Botanical Gardens, St. Louis. Bensasson D, Zhang D, Hartl D, Hewitt G (2001) Mitochondrial pseudogenes: evolution’s misplaced witnesses. Trends in Ecology & Evolution, 16, 314–321. Bik HM, Sung W, De Ley P et al. (2012) Metagenetic community analysis of microbial eukaryotes illuminates biogeographic patterns in deep-sea and shallow water sediments. Molecular Ecology, 21, 1048–1059. Bilewitch JP, Degnan SM (2011) A unique horizontal gene transfer event has provided the octocoral mitochondrial genome with

an active mismatch repair gene that has potential for an unusual self-contained function. BMC Evolutionary Biology, 11, 228. Bisol PM, Costa R, Sibuet M (1984) Ecological and genetical survey on two deep-sea holothurians: Benthogone rosea and Benthodytes typical. Marine Ecology Progress Series, 15, 275–281. Brinkmeyer R, Knittel K, Jurgens J, Weyland H, Amann R, Helmke E (2003) Diversity and structure of bacterial communities in Arctic versus Antarctic pack ice. Applied Environmental Microbiology, 69, 6610–6619. Brown SDJ, Collins RA, Boyer S et al. (2012) Spider: an R package for the analysis of species identity and evolution, with particular reference to DNA barcoding. Molecular Ecology Resources, 12, 562–565. Bruntse G, Tendal OS (2000) Marine Biological Investigations and Assemblages of Benthic Invertebrates from the Faroe Islands. Kaldbak Marine Biological Laboratory, Faroe Islands. Bryan TL, Metaxas A (2006) Distribution of deep-water coral along the North American, continental margins: relationships with environmental factors. Deep-Sea Research Part I Oceanographic Research Papers, 53, 1865–1879. Bucklin A, Wilson RR Jr, Smith KL Jr (1987) Genetic differentiation of seamount and basin populations of the deep-sea amphipod Eurythenes gryllus. Deep Sea Research Part A. Oceanographic Research Papers, 34, 1795–1810. Buhl-Mortensen L, Mortensen PB (2004) Crustaceans associated with the deep-water gorgonian corals Paragorgia arborea (L., 1758) and Primnoa resedaeformis (Gunn., 1763). Journal of Natural History, 38, 1233–1247. Buhl-Mortensen L, Mortensen PB (2005) Distribution and diversity of species associated with deep-sea gorgonian corals off Atlantic Canada. In: Cold-Water Corals and Ecosystems (eds Freiwald AR, Roberts JM), pp. 849–879. SpringerVerlag, Berlin Heidelberg. Buhl-Mortensen L, Vanreusel A, Gooday AJ et al. (2010) Biological structures as a source of habitat heterogeneity and biodiversity on the deep ocean margins. Marine Ecology, 31, 21–50. Butzin M, Lohmann G, Bickert T (2011) Miocene ocean circulation inferred from marine carbon cycle modeling combined with benthic isotope records. Paleoceanography, 26, PA1203. Cairns S (2007) Deep-water corals: an overview with special reference to diversity and distribution of deep-water scleractinian corals. Bulletin of Marine Science, 81, 311–322. Cho W, Shank TM (2010) Incongruent patterns of genetic connectivity among four ophiuroid species with differing coral host specificity on North Atlantic seamounts. Marine Ecology, 31, 121–143. Coffroth MA, Lasker HR, Diamond ME, Bruenn JA, Bermingham E (1992) DNA fingerprints of a gorgonian coral: a method for detecting clonal structure in a vegetative species. Marine Biology, 114, 317–325. Coleman AW (2007) Pan-eukaryote ITS2 homologies revealed by RNA secondary structure. Nucleic Acids Research, 35, 3322–3329. Coleman AW (2009) Is there a molecular key to the level of “biological species” in eukaryotes? A DNA guide. Molecular Phylogenetics and Evolution, 50, 197–203. Cordes EE, Nybakken JW, VanDykhuizen G (2001) Reproduction and growth of Anthomastus ritteri (Octocorallia: Alcyonacea) from Monterey Bay, California, USA. Marine Biology, 138, 491–501. Costello M, McCrea M, Freiwald A et al. (2005) Role of coldwater Lophelia pertusa coral reefs as fish habitat in the NE

© 2012 Blackwell Publishing Ltd

P H Y L O G E O G R A P H Y O F D E E P - S E A C O R A L P A R A G O R G I A A R B O R E A 13 Atlantic. In: Cold-Water Corals and Ecosystems (eds Freiwald A, Roberts JM), pp. 771–805. Springer, Berlin Heidelberg. Darling KF, Wade CM, Stewart IA, Kroon D, Dingle R, Brown AJL (2000) Molecular evidence for genetic mixing of Arctic and Antarctic subpolar populations of planktonic foraminifers. Nature, 405, 43–47. De Queiroz K (2007) Species concepts and species delimitation. Systematic Biology, 56, 879–886. D’Onghia G, Indennidate A, Giove A et al. (2011) Distribution and behaviour of deep-sea benthopelagic fauna observed using towed cameras in the Santa Maria di Leuca cold-water coral province. Marine Ecology Progress Series, 443, 95–110. Drummond AJ, Ho SYW, Rawlence N, Rambaut A (2007) A rough guide to BEAST 1.4. Drummond A, Ashton B, Buxton S et al. (2010) Geneious v5.3. Available from http://www.geneious.com. Drummond AJ, Suchard MA, Xie D, Rambaut A (2012) Bayesian phylogenetics with BEAUti and the BEAST 1.7. Molecular Biology and Evolution, 29, 1969–1973. Duda TFJ, Kohn AJ, Matheny AM (2009) Cryptic species differentiated in Conus ebraeus, a widespread tropical marine gastropod. The Biological Bulletin, 217, 292–305. Edwards SV (2009) Is a new and general theory of molecular systematics emerging? Evolution, 63, 1–19. Ehlers J, Gibbard PL, Hughes PD (2011) Quaternary Glaciations – Extent and Chronology: A Closer Look. Elsevier, Amsterdam. Etnoyer P, Morgan LE (2007) Predictive habitat model for deep gorgonians needs better resolution: comment on Bryan & Metaxas (2007). Marine Ecology Progress Series, 339, 311–312. Etter RJ, Boyle EE, Glazier A, Jennings RM, Dutra E, Chase MR (2011) Phylogeography of a pan-Atlantic abyssal protobranch bivalve: implications for evolution in the Deep Atlantic. Molecular Ecology, 20, 829–843. Excoffier L, Lischer HEL (2010) Arlequin suite ver 3.5: a new series of programs to perform population genetics analyses under Linux and Windows. Molecular Ecology Resources, 10, 564–567. France SC, Hoover LL (2001) Analysis of variation in mitochondrial DNA sequences (ND3, ND4L, MSH) among Octocorallia (=Alcyonaria) (Cnidaria: Anthozoa). Bulletin of the Biological Society of Washington, 10, 110–118. France SC, Kocher TD (1996) Geographic and bathymetric patterns of mitochondrial 16S rRNA sequence divergence among deep-sea amphipods, Eurythenes gryllus. Marine Biology, 126, 633–643. Fu Y (1997) Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics, 147, 915–925. Gernhard T (2008) The conditioned reconstructed process. Journal of Theoretical Biology, 253, 769–778. Giribet G, Wheeler WC (1999) On gaps. Molecular Phylogenetics and Evolution, 13, 132–143. Goffredi S, Hurtado L, Hallam S, Vrijenhoek R (2003) Evolutionary relationships of deep-sea vent and cold seep clams (Mollusca: Vesicomyidae) of the “pacifica/lepta” species complex. Marine Biology, 142, 311–320. Gotelli NJ, Colwell RK (2001) Quantifying biodiversity: procedures and pitfalls in the measurement and comparison of species richness. Ecology Letters, 4, 379–391. Govindarajan AF, Halanych KM, Cunningham CW (2005) Mitochondrial evolution and phylogeography in the hydrozoan Obelia geniculata (Cnidaria). Marine Biology, 146, 213–222.

© 2012 Blackwell Publishing Ltd

Grasshoff M (1979) Zur bipolaren verbreitung der Oktokoralle Paragorgia arborea (Cnidaria: Anthozoa: Scleraxonia). Senckenbergiana maritima, 11, 115–137. Gutie´rrez-Rodriguez C, Barbeitos MS, Sa´nchez JA, Lasker HR (2009) Phylogeography and morphological variation of the branching octocoral Pseudopterogorgia elisabethae. Molecular Phylogenetics and Evolution, 50, 1–15. Herrera S, Baco A, Sa´nchez JA (2010) Molecular systematics of the bubblegum coral genera (Paragorgiidae, Octocorallia) and description of a new deep-sea species. Molecular Phylogenetics and Evolution, 55, 123–135. Hibberd T, Moore K (2009) Field identification guide to Heard Island and McDonald Islands benthic invertebrates: a guide for scientific observers aboard fishing vessels. The Department of Environment, Water, Heritage and the Arts, Australian Antarctic Division and the Fisheries Research and Development Corporation. Howell KL, Rogers AD, Tyler PA, Billett DSM (2004) Reproductive isolation among morphotypes of the Atlantic seastar species Zoroaster fulgens (Asteroidea: Echinodermata). Marine Biology, 144, 977–984. Huelsenbeck JP, Ronquist F (2001) MRBAYES: Bayesian inference of phylogenetic trees. Bioinformatics (Oxford), 17, 754–755. Johnson S, Ware´n A, Vrijenhoek R (2008) DNA barcoding of Lepetodrilus limpets reveals cryptic species. Journal of Shellfish Research, 27, 43–51. Katoh K, Misawa K, Kuma K, Miyata T (2002) MAFFT: a novel method for rapid multiple sequence alignment based on fast Fourier transform. Nucleic Acids Research, 30, 3059–3066. Kingman JFC (1982) The coalescent. Stochastic Processes and their Applications, 13, 235–248. Laptikhovsky V (2011) Migrations and structure of the species range in ridge-scaled rattail Macrourus carinatus (Southwest Atlantic) and their application to fisheries management. ICES Journal of Marine Science: Journal du Conseil, 68, 309–318. Le Goff-Vitry MC, Pybus OG, Rogers AD (2004) Genetic structure of the deep-sea coral Lophelia pertusa in the northeast Atlantic revealed by microsatellites and internal transcribed spacer sequences. Molecular Ecology, 13, 537–549. Leal MC, Puga J, Serodio J, Gomes NCM, Calado R (2012) Trends in the discovery of new marine natural products from invertebrates over the last two decades – where and what are we bioprospecting? PLoS ONE, 7, e30580. Lecroq B, Gooday AJ, Pawlowski J (2009) Global genetic homogeneity in the deep-sea foraminiferan Epistominella exigua (Rotaliida: Pseudoparrellidae). Zootaxa, 2096, 23–32. Lemey P, Rambaut A, Drummond AJ, Suchard MA (2009) Bayesian phylogeography finds its roots. Plos Computational Biology, 5, e1000520. Librado P, Rozas J (2009) DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics, 25, 1451–1452. Lopez JV, Yuhki N, Masuda R, Modi W, O’Brien SJ (1994) Numt, a recent transfer and tandem amplification of mitochondrial DNA to the nuclear genome of the domestic cat. Journal of Molecular Evolution, 39, 174–190. Maddison WP (1997) Gene trees in species trees. Systematic Biology, 46, 523–536. Marshall DC, Simon C, Buckley TR (2006) Accurate branch lengths estimation in partitioned Bayesian analyses requires accommodation of among-partition rate variation and attention to branch length priors. Systematic Biology, 55, 993–1003.

´ NCHEZ 14 S . H E R R E R A , T . M . S H A N K and J . A . S A McClain C, Mincks SL (2010) The dynamics of biogeographic ranges in the deep sea. Proceedings of the Royal Society of London Series B, Biological Sciences, 277, 3533–3546. McFadden CS, Hutchinson MB (2004) Molecular evidence for the hybrid origin of species in the soft coral genus Alcyonium (Cnidaria: Anthozoa: Octocorallia). Molecular Ecology, 13, 1495–1505. McFadden CS, Donahue R, Hadland BK, Weston R (2001) A molecular phylogenetic analysis of reproductive trait evolution in the soft coral genus Alcyonium. Evolution, 55, 54–67. McFadden CS, Tullis ID, Breton Hutchinson M, Winner K, Sohm JA (2004) Variation in coding (NADH dehydrogenase subunits 2, 3, and 6) and noncoding intergenic spacer regions of the mitochondrial genome in Octocorallia (Cnidaria: Anthozoa). Marine Biotechnology, 6, 516–526. McFadden CS, Sa´nchez JA, France SC (2010) Molecular phylogenetic insights into the evolution of Octocorallia: a review. Integrative and Comparative Biology, 50, 389–410. McFadden CS, Benayahu Y, Pante E, Thoma JN, Nevarez PA, France SC (2011) Limitations of mitochondrial gene barcoding in Octocorallia. Molecular Ecology Resources, 11, 19–31. McIntyre A, Ruddiman WF, Karlin K, Mix AC (1989) Surface water response of the equatorial Atlantic Ocean to orbital forcing. Paleoceanography, 4, 19–55. Mercer JH, Sutter JF (1982) Late miocene-earliest pliocene glaciation in southern Argentina: implications for global ice-sheet history. Palaeogeography, Palaeoclimatology, Palaeoecology, 38, 185–206. Metaxas A, Davis J (2005) Megafauna associated with assemblages of deep-water gorgonian corals in Northeast Channel, off Nova Scotia, Canada. Journal of the Marine Biological Association of the United Kingdom, 85, 1381–1390. Miller KJ, Rowden AA, Williams A, Haussermann V (2011) Out of their depth? Isolated deep populations of the cosmopolitan coral Desmophyllum dianthus may be highly vulnerable to environmental change. PLoS ONE, 6, e19004. Monaghan MT, Wild R, Elliot M et al. (2009) Accelerated species inventory on Madagascar using coalescent-based models of species delineation. Systematic Biology, 58, 298–311. Morrison CL, Ross SW, Nizinski MS et al. (2011) Genetic discontinuity among regional populations of Lophelia pertusa in the North Atlantic Ocean. Conservation Genetics, 12, 713–729. Mortensen PB, Buhl-Mortensen L (2004) Distribution of deepwater gorgonian corals in relation to benthic habitat features in the Northeast Channel (Atlantic Canada). Marine Biology, 144, 1223–1238. Mortensen PB, Buhl-Mortensen L, Gebruk AV, Krylova EM (2008) Occurrence of deep-water corals on the Mid-Atlantic Ridge based on MAR-ECO data. Deep Sea Research Part II: Topical Studies in Oceanography, 55, 142–152. Muller T, Philippi N, Dandekar T, Schultz J, Wolf M (2007) Distinguishing species. RNA, 13, 1469–1472. Nylander JA, Wilgenbusch JC, Warren DL, Swofford DL (2008) AWTY (are we there yet?): a system for graphical exploration of MCMC convergence in Bayesian phylogenetics. Bioinformatics (Oxford), 24, 581–583. van Ofwegen LP, Groenenberg DSJ (2007) A centuries old problem in nephtheid taxonomy approached using DNA data (Coelenterata: Alcyonacea). Contributions to Zoology, 76, 153–178. Olsen JL, Stam WT, Coyer JA et al. (2004) North Atlantic phylogeography and large-scale population differentiation of the seagrass Zostera marina L. Molecular Ecology, 13, 1923–1941.

Pante E, Watling L (2012) Chrysogorgia from the New England and Corner Seamounts: Atlantic-Pacific connections. Journal of the Marine Biological Association of the United Kingdom, 92, 911–927. Pawlowski J, Fahrni J, Lecroq B et al. (2007) Bipolar gene flow in deep-sea benthic foraminifera. Molecular Ecology, 16, 4089–4096. Pons J, Barraclough TG, Gomez-Zurita J et al. (2006) Sequencebased species delimitation for the DNA taxonomy of undescribed insects. Systematic Biology, 55, 595–609. Pont-Kindon G, Vassort CG, Rahul W, Okimoto R, Beagley T, Wolstenholme DR (2000) Mitochondrial DNA of Hydra attenuata (Cnidaria): a sequence that includes an end of one linear molecule and the genes for l-rRNA, tRNA f-Met, tRNATrp, COII, and ATPase8. Journal of Molecular Evolution, 51, 404–415. Rambaut A, Drummond AJ (2007) Tracer v1.4. Available from http://beast.bio.ed.ac.uk/Tracer. Ramirez-Llodra E, Tyler PA, Baker MC et al. (2011) Man and the last great wilderness: human impact on the deep sea. PLoS ONE, 6, e22588. Raupach MJ, Malyutina M, Brandt A, Wagele J-W (2007) Molecular data reveal a highly diverse species flock within the munnopsoid deep-sea isopod Betamorpha fusiformis (Barnard, 1920) (Crustacea: Isopoda: Asellota) in the Southern Ocean. Deep Sea Research Part II: Topical Studies in Oceanography, 54, 1820–1830. Roark EB, Guilderson TP, Dunbar RB, Fallon SJ, Mucciarone DA (2009) Extreme longevity in proteinaceous deep-sea corals. Proceedings of the National Academy of Sciences of the United States of America, 106, 5204–5208. Roberts JM, Henry LA, Long D, Hartley JP (2008) Cold-water coral reef frameworks, megafaunal communities and evidence for coral carbonate mounds on the Hatton Bank, north east Atlantic. Facies, 54, 297–316. Roberts JM, Wheeler A, Freiwald AR, Cairns SD (2009) ColdWater Corals: The Biology and Geology of Deep-Sea Coral Habitats. Cambridge University Press, Cambridge. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics (Oxford), 19, 1572–1574. Ruhl MW, Wolf M, Jenkins TM (2010) Compensatory base changes illuminate morphologically difficult taxonomy. Molecular Phylogenetics and Evolution, 54, 664–669. Sa´nchez JA (2005) Systematics of the bubblegum corals (Cnidaria: Octocorallia: Paragorgiidae) with description of new species from New Zealand and the Eastern Pacific. Zootaxa, 1014, 1–72. Sa´nchez JA, Aguilar C, Dorado D, Manrique N (2007) Phenotypic plasticity and morphological integration in a marine modular invertebrate. BMC Evolutionary Biology, 7, 9. Sarmiento JL, Gruber N (2006) Ocean Biogeochemical Dynamics. Princeton University Press, Princeton, New Jersey. Scher HD (2006) Timing and climatic consequences of the opening of Drake Passage. Science, 312, 428–430. Schlagintweit F, Gawlick HJ (2009) The incertae sedis Carpathoporella Dragastan, 1995, from the Lower Cretaceous of Albania: skeletal elements (sclerites, internodes/branches, holdfasts) of colonial octocorals. Facies, 55, 553–573. Schwaninger HR (2008) Global mitochondrial DNA phylogeography and biogeographic history of the antitropically and longitudinally disjunct marine bryozoan Membranipora membranacea L. (Cheilostomata): another cryptic marine sibling species complex? Molecular Phylogenetics and Evolution, 49, 893–908.

© 2012 Blackwell Publishing Ltd

P H Y L O G E O G R A P H Y O F D E E P - S E A C O R A L P A R A G O R G I A A R B O R E A 15 Seibel PN, Muller T, Dandekar T, Schultz J, Wolf M (2006) 4SALE–a tool for synchronous RNA sequence and secondary structure alignment and editing. BMC Bioinformatics, 7, 498. Shank T (2010) Seamounts deep-ocean laboratories of faunal connectivity, evolution, and endemism. Oceanography, 23, 108–122. Smith P, McVeagh S, Mingoia J, France S (2004) Mitochondrial DNA sequence variation in deep-sea bamboo coral (Keratoisidinae) species in the southwest and northwest Pacific Ocean. Marine Biology, 144, 253–261. Soeffker M, Sloman KA, Hall-Spencer JM (2011) In situ observations of fish associated with coral reefs off Ireland. Deep Sea Research Part I: Oceanographic Research Papers, 58, 818–825. Sun Z, Hamel JF, Mercier A (2010) Planulation periodicity, settlement preferences and growth of two deep-sea octocorals from the northwest Atlantic. Marine Ecology Progress Series, 410, 71–87. Swofford DL (2002) PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Tajima F (1983) Evolutionary relationship of DNA sequences in finite populations. Genetics, 105, 437–460. Tajima F (1989) Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics, 123, 585–595. Tendal OS (1992) The North Atlantic distribution of the octocoral P. arborea (L., 1758) (Cnidaria, Anthozoa). Sarsia, 77, 213–217. Thoma J, Pante E, Brugler M, France S (2009) Deep-sea octocorals and antipatharians show no evidence of seamount-scale endemism in the NW Atlantic. Marine Ecology Progress Series, 397, 25–35. Vanoppen MJH, Draisma SGA, Olsen JL, Stam WT (1995) Multiple trans-arctic passages in the red alga Phycodrys rubens: evidence from nuclear rDNA ITS sequences. Marine Biology, 123, 179–188. Verissimo A, McDowell JR, Graves JE (2010) Global population structure of the spiny dogfish Squalus acanthias, a temperate shark with an antitropical distribution. Molecular Ecology, 19, 1651–1662. Vermeij GJ (1991) Anatomy of an invasion – the trans-Arctic interchange. Paleobiology, 17, 281–307. Vrijenhoek RC (2009) Cryptic species, phenotypic plasticity, and complex life histories: assessing deep-sea faunal diversity with molecular markers. Deep-Sea Research Part II, Topical studies in oceanography, 56, 1713–1723. Wareham VE, Edinger EN (2007) Distribution of deep-sea corals in the Newfoundland and Labrador region, Northwest Atlantic Ocean. Bulletin of Marine Science, 81, 289–313. Wares JP, Cunningham CW (2001) Phylogeography and historical ecology of the North Atlantic intertidal. Evolution, 55, 2455–2469. Watanabe S, Metaxas A, Sameoto J, Lawton P (2009) Patterns in abundance and size of two deep-water gorgonian octocorals, in relation to depth and substrate features off Nova Scotia. Deep-Sea Research Part I Oceanographic Research Papers, 56, 2235–2248. Won Y, Young CR, Lutz RA, Vrijenhoek RC (2003) Dispersal barriers and isolation among deep-sea mussel populations (Mytilidae: Bathymodiolus) from eastern Pacific hydrothermal vents. Molecular Ecology, 12, 169–184. Yule GA (1925) A mathematical theory of evolution, based on the conclusions of Dr. J.C. Willis. Philosophical Transactions of the Royal Society of London Series B, Biological Sciences, 213, 21–87.

S.H. and J.A.S. conceived and designed the study. S.H. performed the research, analysed the data and wrote the article with contributions from T.M.S. and J.A.S.

© 2012 Blackwell Publishing Ltd

Data accessibility Sample information and locations are provided as supporting information in Table S1 (Supporting information). DNA sequences are available in GenBank and accession numbers appear in Table S1 (Supporting information). DNA sequence alignments are available in DRYAD doi:10.5061/dryad.ns23j.

Supporting information Additional Supporting Information may be found in the online version of this article. Appendix S1 Sequencing of old samples. Table S1 Collection and sequence information for the specimens used in this study. Table S2 Nucleotide substitution models for mitochondrial gene partitions, as selected by the AIC criterion in JMODELTEST. Table S3 Interspecific and intraspecific (i.e. coalescent depths) uncorrected pairwise distances (%) among haplotypes of species of Paragorgia and Sibogagorgia. Fig. S1 Sampling location of specimens of Paragorgia arborea examined in this study. Fig. S2 Depth distribution of samples shown in Fig. S1 (Supporting information). Fig. S3 The revised geographic distribution of Paragorgia arborea. Fig. S4 Depth distribution of records shown in Fig. S3 (Supporting information). Fig. S5 Predicted ITS2 secondary structure of Paragorgia arborea. Fig. S6 Individual mitochondrial gene tree hypotheses in Paragorgia. Fig. S7 Fit of the GMYC single-threshold model to the mitochondrial time-calibrated gene tree generated with the Yule model tree prior. Fig. S8 Fit of the GMYC single-threshold model to the mitochondrial time-calibrated gene tree generated with the coalescent model tree prior. Fig. S9 Haplotype accumulation curves in Paragorgia arborea.

Appendix S1. Sequencing of old samples - We arbitrarily define samples as ‘old’ if they predate the year of 1979, as this is the year of collection of the oldest specimens of P. arborea in a 40-year time window before present (as suggested by the reviewer). - All the samples at the Smithsonian Institution in Washington DC (main source of the old NA samples, i.e. >90%) are stored in separate glass jars or individual cabinets (in the case of dry samples) and bags, there's no 'common pool' storage. The same is true for the SP samples, which came from NIWA, Wellington, New Zealand. - During the sub-sampling process of specimens for molecular work, all dissecting instruments and working surfaces were carefully sterilized in between samplings. Furthermore, this subsampling took place in a different building than where the DNA extractions and PCR took place (buildings are separated by several kilometers). - DNA was extracted with an automated robotic system (Autogen), which minimizes human error (the only step that requires hands-on work is the insertion of the actual samples into the 96-well extraction plate). - This extraction system is a state-of-the-art instrument that is routinely used by the barcoding facility at the Smithsonian (thousands of samples per week), with extremely rare reports of malfunctioning. There was no reported malfunctioning of the system during the period in which the DNA from these specimens was extracted. - No previous work on this species (or even family) had been performed at the laboratories where molecular work took place, before this study. - DNA extraction was performed in a different room (in a different floor) than the room used for PCRs. - All pipette tips utilized for this study had filters and aerosol barriers. Extreme caution was used while setting up PCR reactions. - All extraction plates and PCR reactions contained negatives. None of these amplified in any reaction. - Given that the old specimens represent only 40% of all the NA specimens, excluding their sequences from our analysis would not change the interpretations, i.e. definition of P. arborea as genealogical-phylospecies, absence of cryptic species, phylogeographic history, shared haplotypes with the South Pacific, low genetic diversity in the NA, genetic differentiation in the species at the oceanic basin-scale. - The following figure shows the plan of the 96-well extraction/PCR/sequencing plate containing the great majority of the old specimens (98%). There are 40 old samples in this plate (36/40 of these came from the NA): 13 dry-preserved and 27 wet-preserved (most likely originally fixed in formalin before preservation in ethanol). 8/13 of the dry specimens were successfully sequenced (all from the NA), but only 1/27 of there wet specimens 1/27 of wet specimens amplified.

A B C D E F G H

A B C D E F G H

1 Parborea9 8080 Parborea1 016643 Parborea4 089 Parborea5 116 Parborea1 6937 Parborea3 3733 Parborea1 00758 Parboreap 5107

2 Parborea4 569 Parborea1 7264 Parborea4 910 ParboreaN ew568 ParboreaN ew545 Parborea1 092766 Parborea1 7263 Parborea4 238

3 Parborea3 0238 Parborea1 7262 Parborea1 00843 ParboreaN ew100 Parborea8 0936 Parborea1 011079 Parborea3 3562 Parboreap 17266

4 Parborea8 0937 Parborea1 00818 Parborea1 002420 Parborea1 092765 Parborea1 010787 Parboreap 33560 Parborea1 7261 Parboreap 4573

5 Parboreap 17265 Parborea4 911 Parborea5 4497 Parborea1 00846 Parborea3 3559 Parborea1 011014 Parborea1 5880 Parboreap 4143

6 Ppacifica5 2432 Parborea1 5874 Parborea3 0239 Parborea1 7260 Parborea1 011097 Parborea5 0890 Parborea5 4298 Parborea5 0028

7 Parborea1 00817 Parborea8 0828 Parborea1 011360 Parborea8 0838 Parborea1 011098 Parborea3 0240 Parborea3 3561 Parborea4 599

8 Parboreap 30241 Parborea1 092764 Parborea4 242 Parborea4 091 Parborea2 1855 Parborea1 011360 Parborea3 308 Parborea3 309

9 Parborea3 310 Parborea3 311 Palisonae 3312 Parborea1 7969 Palisonae 3312 Palisonae 3313 Parborea1 7970 Parborea2 5527

10 Parborea2 8123 Parborea2 8154 Parborea2 8155 Parborea2 8156 Parborea2 8157 Parborea2 8158 Parborea2 8159 Parborea2 8161

11 12 Parborea2 ParboreaS 8160 MF4588 Parborea2 ParboreaS 8392 MF4605 Parborea2 ParboreaS 8422 MF4600 Parborea2 ParboreaS 8423 MF9554 Parborea2 ParboreaS 8425 MF9559 ParboreaZ1 PcfarboreJ 1166 apan ParboreaZ PcfarboreJ C0706 apan ParboreaS Blank MF1246

1 Parborea9 8080 Parborea1 016643 Parborea4 089 Parborea5 116 Parborea1 6937 Parborea3 3733 Parborea1 00758 Parboreap 5107

2 Parborea4 569 Parborea1 7264 Parborea4 910 ParboreaN ew568 ParboreaN ew545 Parborea1 092766 Parborea1 7263 Parborea4 238

3 Parborea3 0238 Parborea1 7262 Parborea1 00843 ParboreaN ew100 Parborea8 0936 Parborea1 011079 Parborea3 3562 Parboreap 17266

4 Parborea8 0937 Parborea1 00818 Parborea1 002420 Parborea1 092765 Parborea1 010787 Parboreap 33560 Parborea1 7261 Parboreap 4573

5 Parboreap 17265 Parborea4 911 Parborea5 4497 Parborea1 00846 Parborea3 3559 Parborea1 011014 Parborea1 5880 Parboreap 4143

6 Ppacifica5 2432 Parborea1 5874 Parborea3 0239 Parborea1 7260 Parborea1 011097 Parborea5 0890 Parborea5 4298 Parborea5 0028

7 Parborea1 00817 Parborea8 0828 Parborea1 011360 Parborea8 0838 Parborea1 011098 Parborea3 0240 Parborea3 3561 Parborea4 599

8 Parboreap 30241 Parborea1 092764 Parborea4 242 Parborea4 091 Parborea2 1855 Parborea1 011360 Parborea3 308 Parborea3 309

9 Parborea3 310 Parborea3 311 Palisonae 3312 Parborea1 7969 Palisonae 3312 Palisonae 3313 Parborea1 7970 Parborea2 5527

10 Parborea2 8123 Parborea2 8154 Parborea2 8155 Parborea2 8156 Parborea2 8157 Parborea2 8158 Parborea2 8159 Parborea2 8161

11 12 Parborea2 ParboreaS 8160 MF4588 Parborea2 ParboreaS 8392 MF4605 Parborea2 ParboreaS 8422 MF4600 Parborea2 ParboreaS 8423 MF9554 Parborea2 ParboreaS 8425 MF9559 ParboreaZ1 PcfarboreJ 1166 apan ParboreaZ PcfarboreJ C0706 apan ParboreaS Blank MF1246

Figure showing the plan of the 96-well plate configuration used to extract, amplify and sequence 98% of the old specimens. TOP shows the distribution of old samples in the plate (underlined names with colored backgrounds). Blue background indicates that the specimen was wet-preserved and red indicates that the specimen was dry-preserved. BOTTOM shows the distribution of all sequenced samples in the plate. Colored backgrounds indicate the mitochondrial haplotype of the specimens of Paragorgia arborea (as in figures 1, 3 and 5 and table S1 of the manuscript): Orange for haplotype m15, light blue for haplotype m3, light green for haplotype m14, light orange for haplotype m16, aquamarine for haplotype m2, violet for haplotype m4, dark green for haplotype m12, brown for haplotype m5, fuchsia for haplotype m13, yellow-green for haplotype m10, yellow for haplotype m11, and pink for haplotype m6. Cells with white background and black font indicate non-NA specimens of Paragorgia cf. arborea that did not amplify. Cells with white background and maroon font indicate modern samples (post-1979) from other species that were sequenced for all genes. Cells with white font indicate NA specimens of Paragorgia cf. arborea, and those underlined indicate old specimens. Cells with black background indicate old NA specimens that did not sequence.

- The ‘old’ specimens (predating 1979) from the NA that were successfully sequenced represented 3 mitochondrial haplotypes: m15 (orange) Globally 28 individuals had this haplotype. 9 from the South Pacific (SP) and 19 from the North Atlantic (NA). All the old samples (collected before 1979) with this haplotype were collected in the NA, and add up to 7 specimens (37% of the total m15 NA samples). m12 (dark green) Globally 10 individuals had this haplotype. 9 from NZ and 1 from the NA, the later is an old sample. m16 (light orange)

This is a unique haplotype from an old NA sample (therefore no contamination is possible in this case). - This plate also contained modern samples from P. arborea specimens representing 12 mt haplotypes, plus representatives from at least 5 other species. - It is thus extremely unlikely that if contamination occurred it was restricted to 2 particular haplotypes, and that it was strongly biased towards dry specimens from the NA - We are aware of two publications that have successfully sequenced old octocoral specimens: - The Aguilar and Sanchez 2008 Bull Mar Sci paper has an ITS2 sequences from the octocoral Calyptrophora japonica collected in 1900s from the USNM collection. - The Aguilar and Sanchez 2007 MPE has several octocoral specimens collected from the Atlantide expedition 1945-1946. Their ITS2 sequences and RNA secondary structures are unique compared to any other octocoral.  In conclusion, all these lines of evidence strongly indicate that our sequences from old specimens are legitimate and are not products of contamination. It is plausible that DNA in old specimens of Paragorgia can be still viable for the sequencing of mitochondrial and nuclear ribosomal DNA if the tissue material is fixed and preserved correctly. Wet specimens were traditional fixed in formalin, which is well known to damage DNA, thus these are mostly unusable for genetic purposes. Dry preservation of specimens of Paragorgia and other octocorals is an acceptable, although not ideal, way to preserve mitochondrial and nuclear ribosomal DNA.

Coll. Number 3312 DAVI1 DAVI3 20070178B01 J2104-6-2 ZC0706ROV01 15721 28057 28697 51244 412 422 425 430 439 3308 3309 3310 3311 17969 17970 17971 25527 28123 28154 28156 28157 28158 28160 28161 28392 28422 28425 41829 41854 41870 41999 42000 42001 42003 44156 44608 44609 46314 46315 46316 46317 46318 46319 46320 46377 Z11166 4089 4091 4238 4242 4569 33559 33560 33561 33562 50890 80838 80936 80937 100758 100817 100818 100843 100846 1007340 1010787 1011097 1011360 1014919 1016320 1027060 1075738 1075744 1075745 1075746 1075753 1075754 1075760

1927 1979 1979 1979 1994 1994 1994 1994 1994 2001 2000 2002 2001 2003 2002 2003 2004 2004 2004 2004 2004 2004 2004

1879 1879 1879 1879 1878

2002 2002 1999 1996 2002 2001 2004 2004 2000 1981 2002 1999 2001 2001 2001 2001 2001 1997 2006 2007 2008 2007 2008 2007 2007 2007 2007 2007 2008 2008 2007 2008 2008 2006 2008 1927 2007

2002

2002 2001 1878 1979

Date 1998 2002 2004 2004 2004

Locality New Zeland Davidson Seamount, California, USA Davidson Seamount, California, USA Gulf of Alaska, USA Amchitka Pass, Andreanof Islands, Aleutian Islands Zhemchung Canyon, Bering Sea Atlantic Ocean Nantucket Island, Massachussetts, USA off Maine, USA Georges Bank, Massachusetts, USA off Norway Trondhjems Fjord, Norway Indian Ocean, Crozet Islands Indian Ocean, Crozet Islands Indian Ocean, Crozet Islands New Zeland New Zeland New Zeland Southern Havre trough, New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland Coral Hill, New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland New Zeland Sable Island, 50 Mile E Of E Light, Nova Scotia, Canada Banquereau Bank, Nova Scotia, Canada Banquereau Bank, South Of, Nova Scotia, Canada Grand Banks, W Part Of, Newfoundland, Canada Sable Island Bank, Nova Scotia, Canada Fishing Banks, North Carolina, USA Fishing Banks, North Carolina, USA Off NE North America, USA Off NE North America, USA Burdwood Bank, S Of Falkland Islands, Scotia Sea Baltimore Canyon, Off Eastern Shore, Maryland, USA Lydonia Canyon, Massachusetts, USA Lydonia Canyon, Massachusetts, USA Aleutian Islands Atka Island, Andreanof Islands, Aleutian Islands Semisopochnoi Island, Rat Islands, Aleutian Islands Tanaga Island, Andreanof Islands, Aleutian Islands Yunaska Island, Islands of Four Mountains, Aleutian Islands Vancouver Island, Canada Norfolk Canyon,Virginia, USA Buldir Reef, Rat Islands, Aleutian Islands off Umnak Island, Fox Islands, Aleutian Islands Davidson Seamount, California, USA British Columbia, Canada Pioneer Seamount, South of farallon Islands, California, USA Dickins Seamount, Gulf of Alaska, USA Dickins Seamount, Gulf of Alaska, USA Dickins Seamount, Gulf of Alaska, USA Welker Seamount, Gulf of Alaska, USA Welker Seamount, Gulf of Alaska, USA Welker Seamount. Gulf of Alaska, USA Pratt Seamount, Gulf of Alaska, USA 1168 375-489 160 102 1313 1152-1192 1712 760 851 849 780 1112 1084 959

480 680-370 613-430

457 457

750-855 952-1118 891 512 366 457

-54.50 38.17 40.38 40.38 52.00 53.00 52.17 52.00 53.00 48.44 37.07 51.96 53.68 35.70 53.70 37.40 54.55 54.51 54.51 55.05 55.07 55.07 56.17

-44.74 -44.74 -44.74 -44.45 -47.25 -47.53 -44.52 -44.50 -44.71 -44.50 -34.82 -47.47 -46.91 -47.53 -50.05 -47.55 -44.52 -47.58 -44.73 -49.50 -46.48 -50.00 -44.44 43.90 44.58 43.90 45.00 43.42 36.00 45.00

753 753 753 858 1044 867-986 600 1203-1288 669 1283-1393 780 720-741 1106-1357 870-967 843-998 888-1015 600 931-1025 794-987

140-200 325 290-305 1225 955 687 1525 900 753 1235 826 872 427 920 959 753

Latitude -48.02 35.75 35.75 55.91 51.68 -47.54 37.67 41.97 42.60 41.00 66.70 63.50 -46.08 -46.67 -46.13 -33.93 -33.93 -44.75 -35.84 -44.74 -44.74 -44.80 -42.83 -47.31 -43.35 -33.92 -44.58 -44.74

Depth (m) 980 1313 1313 867 857 171 480 156 245

-59.10 -73.84 -67.66 -67.66 -170.00 -174.00 179.72 -178.00 -171.00 -126.38 -74.66 176.83 -169.11 -122.70 -133.42 -123.44 -136.84 -136.91 -136.91 -140.31 -140.41 -140.41 -142.70

-176.81 -177.19 -176.81 -179.96 178.33 177.87 175.78 -174.79 -177.04 -174.82 169.86 177.02 171.88 177.92 174.73 177.86 175.77 177.78 -177.04 176.00 170.60 176.06 175.54 -58.80 -57.68 -58.67 -54.00 -60.00 -74.00 -53.50

Longitude 166.08 -122.70 -122.70 -154.02 -179.58 177.93 -74.65 -65.87 -65.73 -67.00 11.60 10.50 50.62 51.67 50.73 167.92 167.91 174.82 177.91 -177.19 -177.19 -177.12 177.42 165.83 178.66 167.92 -177.88 -177.18 J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez B. Stone J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez R. Rathbun A.Verrill R. Rathbun A.Verrill R. Rathbun A.Verrill A.Verrill A.Verrill A.Verrill F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer F. Bayer R. Stone F. Bayer F. Bayer S. Cairns S. Cairns S. Cairns S. Cairns S. Cairns S. Cairns S. Cairns S. Cairns S. Cairns

Tixier-Durivault

Identified By J. Sánchez A. Andrews A. Andrews B. Stone J. Sánchez J. Sánchez A. Johnston A. Johnston A. Johnston

ID Palisonae3312 ParagospDAVI1 ParagospDAVI3 Paragosp200701 ParboreaJ210462 ParboreaZC0706 ParboreaMCZ15721 ParboreaMCZ28057 ParboreaMCZ28697 ParboreaMCZ51244 ParboreaMNHN0412 ParboreaMNHN0422 ParboreaMNHN0425 ParboreaMNHN0430 ParboreaMNHN0439 Parborea3308 Parborea3309 Parborea3310 Parborea3311 Parborea17969 Parborea17970 Paragosp17971 Parborea25527 Parborea28123 Parborea28154 Parborea28156 Parborea28157 Parborea28158 Parborea28160 Parborea28161 Parborea28392 Parborea28422 Parborea28425 Parborea41829 Parborea41854 Parborea41870 Parborea41999 Parborea42000 Parborea42001 Parborea42003 Paragosp44156 Parborea44608 Parborea44609 Paragosp46314 Paragosp46315 Paragosp46316 Paragosp46317 Paragosp46318 Paragosp46319 Parborea46320 Paragosp46377 ParboreaZ11166 Parborea4089 Parborea4091 Parborea4238 Parborea4242 Parborea4569 Parborea33559 Parboreap33560 Parborea33561 Parborea33562 Parborea50890 Parborea80838 Parborea80936 Parborea80937 Parborea100758 Parborea100817 Parborea100818 Parborea100843 Parborea100846 Paragosp1007340 Parborea1010787 Parborea1011097 Parborea1011360 Paragosp1014919 Ppacifica1016320 Paragosp1027060 Paragosp1075738 Paragosp1075744 Paragosp1075745 Paragosp1075746 Paragosp1075753 Paragosp1075754 Paragosp1075760

16S JX128380 JX128430 JX128367 JX128397 JX128425 JX128372 JX128389 JX128427 JX128351 JX128376 JX128370 JX128429 N/A N/A N/A JX128440 JX128398 JX128388 JX128426 JX128381 JX128419 JX128444 JX128423 JX128415 JX128441 JX128420 JX128424 JX128437 JX128384 JX128386 JX128438 JX128434 JX128418 JX128414 JX128435 JX128421 JX128387 N/A JX128354 N/A JX128383 JX128392 JX128406 JX128394 JX128369 JX128408 JX128396 JX128371 JX128413 JX128439 JX128379 JX128375 JX128431 JX128391 JX128363 JX128352 JX128359 JX128436 JX128410 JX128358 N/A JX128412 JX128416 JX128362 JX128385 JX128355 JX128404 JX128433 JX128382 JX128411 JX128364 JX128368 JX128357 JX128378 JX128373 JX128361 JX128446 JX128365 JX128432 JX128366 JX128445 JX128356 JX128405 JX128393

nad2-16S JX128457 JX128464 JX128542 JX128465 JX128524 JX128449 JX128521 JX128473 JX128483 JX128520 JX128540 JX128507 N/A N/A N/A JX128539 JX128455 JX128475 JX128476 JX128515 JX128513 JX128480 JX128512 JX128496 JX128448 JX128472 JX128510 JX128538 JX128537 JX128531 JX128498 JX128514 JX128456 JX128523 JX128506 JX128478 JX128470 N/A JX128474 N/A JX128509 JX128497 JX128492 JX128454 JX128450 JX128505 JX128458 JX128482 JX128532 JX128522 JX128461 JX128487 JX128493 JX128501 JX128451 JX128485 JX128491 JX128466 JX128499 JX128508 N/A JX128467 JX128517 JX128518 JX128469 JX128534 JX128494 JX128536 JX128459 JX128489 JX128500 JX128488 JX128528 JX128462 JX128495 JX128519 JX128463 JX128502 JX128453 JX128535 JX128471 JX128526 JX128468 JX128447

GenBank Accesion Numbers coxI mtMutS JX128345 JX128349 JX124636 JX124577 JX124652 JX124542 JX124684 JX124587 JX124642 JX124555 JX124687 JX124598 JX124657 JX124606 JX124646 JX124551 JX124702 JX124561 JX124638 JX124600 JX124681 JX124609 JX124667 JX124564 N/A N/A N/A N/A N/A N/A JX124645 JX124552 JX124656 JX124535 JX124641 JX124571 JX124686 JX124525 JX124624 JX124607 JX124637 JX124605 JX124677 JX124592 JX124689 JX124548 JX124682 JX124546 JX124625 JX124553 JX124635 JX124610 JX124671 JX124531 JX124661 JX124566 JX124694 JX124523 JX124614 JX124568 JX124616 JX124575 JX124697 JX124601 JX124619 JX124550 JX124660 JX124585 JX124683 JX124611 JX124643 JX124558 JX124658 JX124588 N/A N/A JX124678 JX124574 N/A N/A JX124692 JX124563 JX124673 JX124537 JX124672 JX124534 JX124666 JX124594 JX124621 JX124569 JX124617 JX124573 JX124695 JX124603 JX124698 JX124583 JX124612 JX124595 JX124703 JX124570 JX124648 JX124549 JX124664 JX124533 JX124615 JX124591 JX124627 JX124579 JX124629 JX124608 JX124632 JX124567 JX124701 JX124596 JX124670 JX124559 JX124626 JX124572 JX124633 JX124562 N/A N/A JX124651 JX124521 JX124676 JX124544 JX124622 JX124543 JX124613 JX124599 JX124634 JX124584 JX124669 JX124539 JX124700 JX124597 JX124674 JX124529 JX124679 JX124538 JX124650 JX124556 JX124639 JX124528 JX124659 JX124545 JX124631 JX124582 JX124668 JX124578 JX124680 JX124581 JX124620 JX124560 JX124649 JX124526 JX124690 JX124565 JX124675 JX124602 JX124630 JX124540 JX124685 JX124527 JX124665 JX124593 JX124699 JX124530 nad6-int-nad3 JX128568 JX128629 JX128606 JX128579 JX128591 JX128551 JX128610 JX128578 JX128550 JX128631 JX128633 JX128595 N/A N/A N/A JX128620 JX128585 JX128615 JX128574 JX128619 JX128621 JX128618 JX128566 JX128571 JX128549 JX128548 JX128632 JX128625 JX128562 JX128544 JX128581 JX128588 JX128587 JX128558 JX128555 JX128592 JX128590 N/A JX128613 N/A JX128601 JX128582 JX128567 JX128598 JX128603 JX128611 JX128597 JX128575 JX128609 JX128573 JX128616 JX128563 JX128636 JX128589 JX128577 JX128596 JX128634 JX128561 JX128635 JX128554 N/A JX128630 JX128553 JX128556 JX128557 JX128547 JX128626 JX128546 JX128584 JX128570 JX128572 JX128617 JX128560 JX128543 JX128622 JX128612 JX128594 JX128623 JX128607 JX128599 JX128580 JX128605 JX128624 JX128545

ITS2 N/A JX128683 JX128680 N/A N/A JX128666 N/A N/A N/A N/A JX128641 JX128658 JX128662 JX128663 JX128673 JX128642 JX128664 JX128656 JX128648 JX128652 N/A JX128669 N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A N/A JX128644 N/A JX128672 JX128668 JX128649 N/A JX128681 JX128655 N/A JX128654 N/A N/A N/A N/A N/A N/A JX128651 N/A N/A JX128685 N/A JX128679 JX128653 JX128657 N/A N/A N/A JX128665 N/A JX128661 N/A N/A N/A N/A JX128675 JX128660 N/A JX128678 JX128671 JX128674 N/A N/A N/A JX128643 JX128676 N/A N/A JX128646 JX128682 N/A JX128645 m4 m15 m15 m15 m3 m3 m3 m3 m3 m9 m15 m2 m3 m9 m9 m9 m9 m9 m9 m9 m9 m9 m9

m12 m12 m11 m12 m11 m12 m12 m11 m11 m12 m11 m11 m15 m12 m15 m15 m15 m16 m15 m15

m11

m15 m15 m5 m5 m15 m12 m12 m15 m15 m15 m15 m15 m12 m11 m13 m13 m13 m10 m12 m13 m12 m15

m9 m9 m9 m7 m3 m15 m15 m15 m15 m15 m15

i6

i6 i11

i6 i6

i6 i1 i8

i8 i8

i2

i2

i1 i2 i6

i6

i6

i2

i4 i2

i2 i2 i2

i2

i1

i2 i2 i1 i1 i5 i7 i7 i7 i7 i7

i8

i6 i6

Haplotypes Mitochondrial ITS2

Acronyms as follows: National Museum of Natural History, Smithsonian Institution, USA (USNM); The National Institute of Water and Atmospheric Research, New Zealand (NIWA); Museum of Comparative Zoology, Harvard University, USA (MCZ); Muséum National d'Histoire Naturelle, Paris, France (MNHN); Senckenberg Research Institute And Natural History Museum Frankfurt, Germany (SMF); Uppsala University Evolutionsmuseet, Sweden (UUZM); Wakayama Prefectural Museum of Natural History, Japan (WPMNH); Yale Peabody Museum of Natural History, USA (YPM).

Collection NIWA A. Andrews A. Andrews B. Stone B. Stone B. Stone MCZ MCZ MCZ MCZ MNHN MNHN MNHN MNHN MNHN NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA NIWA USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM

Table S1. Collection and sequence information for the specimens used in this study.

Species Paragorgia alisonae Paragorgia arborea

1075761 1075766 1092764 1092765 1092766 1120444 1122233 1122237 1122240 1123932 1123935 1123936 1123937 1123938 WPMNH 27002A 98785 73767 3320 1122305 3326 1073480 1122229

2004 2004 2000 2000 2000 2008 2002 2002 2002 2002 2004 2004 2004 2004 2005 2000 1995 1984 1989 2004 2001 2003 2006

Pratt Seamount, Gulf of Alaska, USA Welker Seamount, Gulf of Alaska, USA East of Virginia Beach,Virginia, USA East of Virginia Beach,Virginia, USA East of Virginia Beach,Virginia, USA off Maryland, USA San Juan Seamount, California, USA Rodriguez Seamount, California, USA San Juan Seamount, California, USA South of Trinity Islands, Aleutian Islands Amlia Island, Andreanof Islands, Aleutian Islands Amlia Island, Andreanof Islands, Aleutian Islands Adak Canyon, Andreanof Islands, Aleutian Islands Amchitka Pass, Andreanof Islands, Aleutian Islands Off Yaizu-shi, Shijuoka Prof., Japan Atlantic Ocean, Bear Seamount, near continental shelf East Pacific Rise, off Mexico Little Bahama Bank, Bahamas Off east coast, New Zealand British Columbia, Canada Off east coast, New Zealand Off Vancouver Isl., British Columbia, Canada Davidson Seamount, California, USA 900 846–861 3042.4

941 1114 375-489 375-489 375-489 400 1360.8 894.5 1362.9 746 843 843 1269 747 760-800 1439-1460 1950 608 820

56.17 55.07 37.07 37.07 37.07 37.06 32.97 34.06 32.97 55.87 51.81 51.81 51.51 51.72 33.00 39.88 12.73 27.10 -36.16 51.20 -42.79 50.23 35.63

-142.70 -140.41 -74.66 -74.66 -74.66 -74.62 -121.04 -121.08 -121.04 -154.06 -173.83 -173.83 -177.04 -179.58 138.40 -67.44 -102.60 -79.70 176.81 -130.14 179.99 -128.58 -122.83 J. A. Moore F. Bayer F. Bayer J. Sánchez J. Boutillier J. Sánchez J. Sánchez S. Herrera

S. Cairns S. Cairns J. Sánchez J. Sánchez J. Sánchez S. Cairns J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez J. Sánchez

Paragosp1075761 Paragosp1075766 Parborea1092764 Parborea1092765 Parborea1092766 Paragospnizinski ParboreaT662A29 ParboreaT661A10 ParboreaT662A28 Parborea4178B ParboreaJ2095272 ParboreaJ2095271 ParboreaJ2099211 ParboreaJ210441 PcfarboreJapan Parborea27002A Pcfcorallo98785 Pjohnsoni73767 Pkaupeka3320 Paragosp121 Pwahine3326 Pyutlinux1073480 Sibogasp1T947A9

JX128374 JX128400 JX128395 JX128377 JX128353 JX128422 JX128409 JX128401 JX128402 JX128360 JX128399 JX128407 JX128443 JX128403 JX128442 N/A JX128428 JX128417 GQ293254 JX128390 GQ293255 GQ293256 GQ293258

JX128529 JX128511 JX128525 JX128541 JX128477 JX128479 JX128530 JX128527 JX128452 JX128481 JX128516 JX128460 JX128484 JX128533 JX128486 N/A JX128504 JX128503 GQ293332 JX128490 GQ293333 GQ293334 GQ293336

JX124644 JX124618 JX124688 JX124655 JX124640 JX124663 JX124662 JX124623 JX124691 JX124628 JX124696 JX124693 JX124653 JX124647 JX124654 N/A JX128346 JX128344 GQ293283 JX128343 GQ293284 GQ293285 GQ293287

JX124576 JX124541 JX124589 JX124520 JX124554 JX124547 JX124586 JX124590 JX124522 JX124524 JX124580 JX124532 JX124604 JX124536 JX124557 N/A JX128350 JX128348 GQ293313 JX128347 GQ293314 GQ293315 GQ293317

JX128559 JX128614 JX128569 JX128627 JX128586 JX128552 JX128628 JX128565 JX128564 JX128637 JX128583 JX128602 JX128604 JX128608 JX128593 N/A JX128576 JX128600 GQ293351 JX128638 GQ293352 GQ293353 GQ293355

N/A JX128684 N/A JX128667 JX128640 N/A N/A N/A N/A JX128677 N/A JX128659 JX128639 JX128670 JX128650 JX128647 N/A N/A GQ293292 N/A GQ293296 GQ293297 GQ293290

m9 m9 m15 m14 m15 m15 m9 m9 m9 m9 m7 m6 m7 m1 m8 i1 i9 i10 i1 i1

i1

i5 i2

i6

Acronyms as follows: National Museum of Natural History, Smithsonian Institution, USA (USNM); The National Institute of Water and Atmospheric Research, New Zealand (NIWA); Museum of Comparative Zoology, Harvard University, USA (MCZ); Muséum National d'Histoire Naturelle, Paris, France (MNHN); Senckenberg Research Institute And Natural History Museum Frankfurt, Germany (SMF); Uppsala University Evolutionsmuseet, Sweden (UUZM); Wakayama Prefectural Museum of Natural History, Japan (WPMNH); Yale Peabody Museum of Natural History, USA (YPM).

Paragorgia coralloides Paragorgia johnsoni Paragorgia kaupeka Paragorgia sp. Paragorgia wahine Paragorgia yutlinux Sibogagorgia cauliflora

USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM USNM WPMNH YPM USNM USNM NIWA USNM NIWA USNM USNM

Table S1. Collection and sequence information for the specimens used in this study.

Table S2. Nucleotide substitution models for mitochondrial gene partitions, as selected by the AIC criterion in jModeltest. 16S coxI nad3 nad2 mtMutS nad6 int

Model TPM2+I TPM2+I TPM1 TPM1+I TPM2+I TVMef+G TIM3ef

Partition 10212 10212 12210 12210 10212 12314 12032

-lnL 1086.8644 853.9125 239.1497 1168.0114 1581.1612 1106.6466 68.6581

AIC 2271.7289 1805.8249 574.2994 2434.0228 3260.3225 2315.2932 235.3162

deltaAIC 0 0 0 0 0 0 0

weight 0.1562 0.1728 0.0933 0.0698 0.1328 0.1708 0.1623

cumweight 0.1562 0.1728 0.0933 0.0698 0.1328 0.1708 0.1623

uDelta 115.8078 242.8336 216.9732 319.3048 -

S. cauliflora 2.9 3.2 2.8 3.2 2.8 3.5 3.9 3.2a

S. cauliflora 5.8 5.9 8.0 6.2a

cox1 S. cauliflora P. kaupeka P. coralloides P. johnsoni Paragorgia sp.1 P. yutlinux P. alisonae P. wahine P. arborea

ITS2 S. cauliflora P. kaupeka P. wahine P. yutlinux P. arborea 5.3 7.4 5.6a 6.4 4.7a

P. wahine

0.4 0.9 0.4 0.9 1.5 1.1a

2.0 1.5 2.0 1.5 2.0 2.3 1.8a P. kaupeka

P. coralloides

2.1 2.8 2.2 2.3 2.2 1.8a

4.1 5.4 6.1 5.5 5.6 5.5 5.1a P. kaupeka

P. coralloides

P. kaupeka

3.3a

P. yutlinux

0.4 0.0 0.6 1.1 0.6a

P. johnsoni

1.0 0.5 0.7 0.6 2.1a

P. johnsoni

0.6b

P. arborea

0.4 1.1 1.5 1.1a

Paragorgia sp.1

1.2 1.4 1.3 2.8a

Paragorgia sp.1

0.6 1.1 0.6a

P. yutlinux

0.7 0.6 2.1a

P. yutlinux

a = minimum pairwise distance among congeners, b = maximum pairwise distance among conspecifics.

S. cauliflora 7.1 5.2 6.6 7.3 6.6 6.8 6.6 6.3a

mtMutS S. cauliflora P. kaupeka P. coralloides P. johnsoni Paragorgia sp.1 P. yutlinux P. alisonae P. wahine P. arborea

1.7 1.3a

P. alisonae

0.7 2.2a

P. alisonae

1.3a

P. wahine

2.1a

P. wahine

0.9b

P. arborea

0.3b

P. arborea

Table S3. Interspecific and intraspecific (i.e. coalescent depths) uncorrected pairwise distances (%) among haplotypes of species of Paragorgia and Sibogagorgia.

80° 70° 60° 50° 40° 30° 20° 10° 0° -10° -20° -30° -40° -50° -60° -70° -80° -180°

-160°

-140°

-120°

-100°

-80°

-60°

-40°

-20°

0°

20°

40°

60°

80°

100°

120°

140°

160°

180°

Figure S1. Sampling location of the specimens of Paragorgia arborea examined in this study. Colors indicate depth ranges: yellow from 0 to 500 m, orange from 500 to 1000 m, red from 1000 to 1500 m, and maroon from 1500 to 2000 m. Blue indicates records for which depth data was unavailable.

Number of Samples

50 40 30 20 10 0

NA

0-500 500-1000 1000-15001500-2000 Depth Range (m) Figure S2. Depth distribution of samples shown in Figure S1.

80° 70° 60° 50° 40° 30° 20° 10° 0° -10° -20° -30° -40° -50° ★

-60°

★ ★

-70° -80° -180°

-160°

-140°

-120°

-100°

-80°

-60°

-40°

-20°

0°

20°

40°

60°

80°

100°

120°

140°

160°

180°

Number of Records

Figure S3. The revised geographical distribution of Paragorgia arborea. Circles = high-confidence records, squares = moderate confidence, and stars = low-confidence (confidence qualitatively assessed based on abundance of records in the neighboring area). Colors indicate depth ranges: yellow from 0 to 500 m, orange from 500 to 1000 m, red from 1000 to 1500 m, and maroon from 1500 to 2000 m. Blue indicates records for which depth data was unavailable.

150

100

50

0

NA

0-500

500-1000 1000-1500 1500-2000 Depth Range (m) Figure S4. Depth distribution of records shown in Figure S3.

i1 i2 i3 i4 i5 i6 i7 i8 i9 i10 i11

Figure S5. Predicted ITS2 secondary structure of Paragorgia arborea. The arrows and associated numbers correspond to the 11 haplotypes (as seen in the upper left table) found in the examined specimens. Dashes indicate absence of change relative to a reference 75%-consensus sequence.

16S

coxI

1

m14 m10 m16 m5 m11 m8 m12 m7 m13 m4 m6 m15 m9 m3 m1 m2 P. yutlinux P. wahine P. alisonae Paragorgia sp. P. johnsoni P. coralloides P. kaupeka S. cauliflora

nad3

nad6

nad2 m15 m1 m4 m6 m9 m3 m7 m14 m16 m12 m10 m2 m8 m5 m13 m11 Paragorgia sp. P. johnsoni P. yutlinux P. wahine P. alisonae P. coralloides P. kaupeka S. cauliflora

m14 m15 m16 m12 m3 m13 m5 m6 m7 m8 m9 m11 m10 m4 m2 m1 P. alisonae P. johnsoni P. yutlinux P. wahine Paragorgia sp. P. kaupeka P. coralloides S. cauliflora

m14 m6 m12 m11 m5 m15 m8 m13 m16 m7 m10 m4 m9 m3 m2 m1 P. wahine P. johnsoni P. coralloides P. yutlinux P. alisonae Paragorgia sp. P. kaupeka S. cauliflora

mtMutS m14 m15 m13 m16 m11 m6 m8 m12 m7 m5 m10 m3 m2 m1 m4 m9 P. johnsoni P. alisonae Paragorgia sp. P. wahine P. yutlinux P. coralloides P. kaupeka S. cauliflora

m1 m3 m2 m4 m7 m6 m8 m10 m11 m15 m13 m5 m12 m16 m9 m14 P. johnsoni Paragorgia sp. P. wahine P. yutlinux P. alisonae P. coralloides P. kaupeka S. cauliflora

int Paragorgia sp. m5 m1 m11 m14 m16 P. wahine P. johnsoni m8 m2 m12 m3 m9 m10 m4 m15 P. alisonae m7 m13 P. yutlinux m6 P. coralloides P. kaupeka S. cauliflora

posterior probability > 0.95

Fig. S6. Individual mitochondrial gene-tree hypotheses in Paragorgia. Trees are 50% consensus cladograms of the sampled trees, after burnin, in the Bayesian analyses.

20

-43.0

10

-43.5

Parborea28392

Parborea33559

-44.0

Paragosp46316 Parborea42001

-44.5

Log-likelihood

ParboreaMCZ51244

5

N

Parborea1092765

Parborea28425

ParboreaJ2099211 ParboreaJ2095271

1

-45.5

2

-45.0

PcfarboreJapan

-100

-80

-60

-40

-20

0

Parborea3311 -100

Time (million years)

-80

-60

-40

-20

0

Paragosp1075754

Time (million years)

Parborea50890

Log-likelihood of null model: -45.14365 Maximum log-likelihood of GMYC model: -42.71979 Likelihood ratio: 4.847715 Result of LR test: 0.1832939 n.s. ML number of clusters: 3 95% Confidence interval: 2-5 ML number of entities: 9 95% Confidence interval: 2-23 Threshold time estimate: -21.03428 Ma BP

ParboreaJ210441 Parborea1011097 Parborea1011360 Paragosp121 Pyutlinux1073480 Pwahine3326 Pjohnsoni73767 Palisonae3312 Pcfcorallo98785 Pkaupeka3320 Sibogasp1T947A9 Coralliumsp1075800 Claauense1072452 Claauense1071433 Paracora1089600 Csecundum1010758 Ckishinouyei1072441

140.0

120.0

100.0

80.0

60.0

40.0

20.0

0.0

Time (million years)

Figure S7. Fit of the GMYC single-threshold model to the mitochondrial time-calibrated gene tree generated with the Yule-model tree prior. Ultrametric tree shows the estimated times of divergence under this model. Node bars represent the 95% highest posterior density intervals. Clades with red branches indicate the inferred, independently evolving, lineages. Dotted lines indicate the ML inferred time for the speciation-coalescent threshold for species delimitation. Top boxes show the lineage-through-time plot (left) and the log-likelihood plot (right).

Log likelihood -80

-60

-40

-20

-39.5 -39.0 -38.5 -38.0 -37.5 -37.0 -36.5

20 10 N

5 2 1 -100

ParboreaMCZ51244

0

Parborea1092765 Parborea33559 Parborea28392 Paragosp46316 Parborea28425 Parborea42001 ParboreaJ2099211 PcfarboreJapan ParboreaJ2095271 -100

-80

Time (million years)

-60

-40

-20

Parborea3311

0

Time (million years)

Paragosp1075754 Parborea50890

Log-likelihood of null model: -39.40713 Maximum log-likelihood of GMYC model: -36.48733 Likelihood ratio: 5.839602 Result of LR test: 0.1196801 n.s. ML number of clusters: 3 95% Confidence interval: 2-5 ML Number of entities: 9 95% Confidence interval: 2-19 Threshold time estimate: -16.07062 Ma BP

ParboreaJ210441 Parborea1011097 Parborea1011360 Paragosp121 Pyutlinux1073480 Pwahine3326 Palisonae3312 Pjohnsoni73767 Pcfcorallo98785 Pkaupeka3320 Sibogasp1T947A9 Coralliumsp1075800 Claauense1072452 Claauense1071433 Paracora1089600 Csecundum1010758 Ckishinouyei1072441

160.0

140.0

120.0

100.0

80.0

60.0

40.0

20.0

0.0

Time (million years)

Figure S8. Fit of the GMYC single-threshold model to the mitochondrial time-calibrated gene tree generated with the coalescent-model tree prior. Ultrametric tree shows the estimated times of divergence under this model. Node bars represent the 95% highest posterior density intervals. Clades with red branches indicate the inferred, independently evolving, lineages. Dotted lines indicate the ML inferred time for the speciation-coalescent threshold for species delimitation. Top boxes show the lineage-through-time plot (left) and the log-likelihood plot (right).

8

10

15

4

6

10

Haplotypes

2

5

20

40

60

80

0

20

40 60 Individuals

80

0

20

40

60

80

0

20

40 60 Individuals

80

10

10 8 2

4

5

6

Haplotypes

12

15

14

0

Figure S9. Haplotype accumulation curves in Paragorgia arborea. (Top left) black line indicates mitochondrial haplotypes, and gray line indicates nuclear ITS2 haplotypes. The other graphs show the contribution of individual genes to the mitochondrial diversity. (Top right) ND6 haplotypes (green), ND2 haplotypes (blue), mtMutS haplotypes (red), 16S haplotypes (purple), and COI haplotypes (yellow). (Bottom left) ND6+16S haplotypes (purple), ND6+ND2 (green), ND6+COI (yellow), ND6+mtMutS (red). (Bottom right) ND6+16S+ND2 (green), ND6+16S+COI (yellow), and ND6+16S+mtMutS (red).