New Species of Clade B Symbiodinium (Dinophyceae) from the Greater Caribbean Belong to Different Functional Guilds: S

Home , Antillogorgia bipinnata, Symbiodinium

NEW SPECIES OF CLADE B SYMBIODINIUM (DINOPHYCEAE) FROM THE GREATER CARIBBEAN BELONG TO DIFFERENT FUNCTIONAL GUILDS: S. AENIGMATICUM SP. NOV., S. ANTILLOGORGIUM SP. NOV., S. ENDOMADRACIS SP. NOV., AND S. PSEUDOMINUTUM SP. NOV.1

John Everett Parkinson2 Department of Biology, The Pennsylvania State University, University Park, Pennsylvania 16802, USA Mary Alice Coffroth Graduate Program in Evolution, Ecology and Behavior and Department of Geological Sciences, State University of New York at Buffalo, Buffalo, New York 14260, USA and Todd C. LaJeunesse Department of Biology, The Pennsylvania State University, University Park, Pennsylvania 16802, USA

Molecular approaches have begun to supersede as corals (Trench 1993). Despite their ecological traditional morphometrics in the species delineation importance, the formal taxonomic description of of micro-eukaryotes. In addition to fixed differences Symbiodinium species has lagged far behind the in DNA sequences, recent genetics-based recognition of molecular diversity within the system, descriptions within the dinoflagellate genus impeding research efforts (LaJeunesse et al. 2012). Symbiodinium have incorporated confirmatory One major reason for this delay is that most mem- morphological, physiological, and ecological bers of the genus bear striking morphological simi- evidence when possible. However, morphological larities to one another. Of the handful of species and physiological data are difficult to collect from descriptions that exist, earlier work emphasized species that have not been cultured, while the natural morphometrics (Freudenthal 1962, Trench and ecologies of many cultured species remain unknown. Blank 1987, Blank and Huss 1989, Trench and Here, we rely on genetic evidence—the only data Thinh 1995, Hansen and Daugbjerg 2009), while consistently available among all taxa investigated—to recent work has shifted toward greater reliance on describe four new Clade B Symbiodinium species. The genetic evidence (LaJeunesse et al. 2012, 2014, ‘host-specialized’ species (S. antillogorgium sp. nov. 2015, Jeong et al. 2014, Hume et al. 2015, Lee et al. and S. endomadracis sp. nov.) engage in mutualisms 2015). with specific cnidarian hosts, but exhibit differences Traditionally, only dinoflagellate species that in our ability to culture them in vitro. The could be cultured were formally described. How- ecologically ‘cryptic’ species (S. aenigmaticum sp. nov. ever, culturing is a highly selective process (von and S. pseudominutum sp. nov.) thrive in culture, but Wintzingerode et al. 1997), particularly for symbi- their roles or functions in the ecosystem (i.e., niches) otic dinoflagellates (Rowan 1998, Carlos et al. 1999, are yet to be documented. These new species call Santos et al. 2001). Efforts to culture the mutualists further attention to the spectrum of ecological guilds that are numerically dominant in a given host have among Symbiodinium. often failed (Santos et al. 2001, LaJeunesse 2002, Goulet and Coffroth 2003). Although most of these Key index words: Clade B; culture; ecological guilds; evolutionarily derived, ‘host-specialized’ lineages do free-living; species recognition; Symbiodinium; not persist in artificial culture (LaJeunesse et al. symbiosis; taxonomy 2005, Krueger and Gates 2012), they are some of Abbreviations: Cob, cytochrome b, cp23S, chloro- the most abundant and ecologically important Sym- plast 23S rDNA; ITS, internal transcribed spacer biodinium. At the same time, those Symbiodinium that perform well in culture are often low abundance types distinct from the dominant symbiont in hos- The genus Symbiodinium encompasses a diverse pite (Santos et al. 2001, LaJeunesse 2002). They array of unicellular dinoflagellates, many of which might be background endosymbionts, opportunistic participate in mutualisms with cnidarian hosts such parasites, ectosymbionts, or even free-living hetero- trophs temporarily captured in the gastrovascular cavity or surface mucus (LaJeunesse 2002, Jeong 1Received 1 April 2015. Accepted 24 July 2015. 2Author for correspondence: e-mail [email protected]. et al. 2014, LaJeunesse et al. 2015). As these Editorial Responsibility: S. Lin (Associate Editor) lineages are difficult to characterize ecologically, but

850 NEW SYMBIODINIUM SPECIES FROM DIFFERENT GUILDS 851 clearly occupy distinct niches compared to classically recently begun the process of speciation and those mutualistic host-specialized species, we refer to them that are farther along. as ‘cryptic’ (sensu LaJeunesse 2001). Here, we use a combination of genetic concor- Population genetic analyses have shown that sex- dance at multiple phylogenetic markers and some ual recombination occurs within but not between ecological information to test the hypothesis that sympatric Symbiodinium populations with distinct four Symbiodinium Clade B lineages are genetically ecological distributions (LaJeunesse et al. 2010, isolated and represent distinct species. Two of the 2014, Thornhill et al. 2014). Such patterns reflect focal lineages associate with cnidarian hosts and reproductive isolation and serve as a powerful appear host-specialized. The first, which has an ITS2 basis for species delineation (LaJeunesse et al. sequence corresponding to type B1 (sensu LaJe- 2014). However, developing microsatellite markers unesse 2001), is mutualistic with sea plumes in the can be resource-intensive, and their utility may be genus Antillogorgia. The second, corresponding to restricted to a few closely related lineages. In the type B7 (sensu LaJeunesse 2002) is mutualistic with absence of population genetic markers, the com- stony corals in the genus Madracis (Diekmann et al. bined analysis of more conservative, unlinked phy- 2003, Pettay and LaJeunesse 2007). The former logenetic markers spanning nuclear, shows a degree of culturability while the latter has mitochondrial, and plastid genomes can effectively not been successfully brought into unialgal culture distinguish evolutionarily divergent lineages (Sam- using current techniques and media. The other two payo et al. 2009). Recombination within species focal lineages thrive in culture but to date have not results in different genealogies among unlinked been detected in host tissues using conventional loci, while between species, drift removes ancestral genetic approaches, and therefore exhibit cryptic, alleles, producing concordant genealogies. Thus, potentially free-living ecologies. The first is closely species can be delimited if they exhibit congruent related to type B23 (sensu Finney et al. 2010). The reciprocal monophyly or nondiscordance at multi- second is closely related to S. minutum (with which ple unlinked loci (Sampayo et al. 2009, Leliaert it shares ITS2 type B1; LaJeunesse et al. 2012). et al. 2014). We recently used this multi-gene approach to delineate S. minutum and S. psygmophilum, two Clade MATERIALS AND METHODS B Symbiodinium (LaJeunesse et al. 2012). These taxa Specimen collection. Tissues from the host cnidarians Antillo- exhibited clear differentiation and reciprocal mono- gorgia bipinnata, A. elisabethae and Madracis spp. were collected phyly across multiple genomic loci, indicating a lack from the Florida Keys, Barbados, Belize and/or Curacßao to of sexual recombination which is consistent with the target symbionts with host-specialized associations. Some of Biological Species Concept (Mayr 1942, Avise and the samples had been used in an earlier study (Finney et al. Wollenberg 1997). Several additional lines of evi- 2010) and were previously screened for the presence of Clade dence were also available, though not all of them B Symbiodinium using denaturing gradient gel electrophoresis of the ITS region as described by Sampayo et al. (2009). New are needed to describe species. S. minutum and samples were similarly screened. Whole tissues were preserved S. psygmophilum were known to function as common in high-salt, 20% DMSO buffer (Seutin et al. 1991) and host-specialized symbionts, so their ecologies and stored at À20°C until processed for molecular analysis. biogeographies had been well studied and deter- Cultures and cell size analysis. Monoclonal cultures of Clade mined to be distinct (satisfying the Ecological Spe- B Symbiodinium isolated from western Atlantic cnidarian hosts cies Concept; Van Valen 1976). They were readily were acquired from the Robert K. Trench collection and the maintained in culture, allowing for physiological Buffalo Undersea Reef Research Culture Center (BURRCC) collection. Cultures were maintained in liquid media (ASP8- experiments that revealed further functional differ- A; Ahles 1967) and kept in incubators delivering 80– À À ences in thermal tolerance. Cell size disparities sug- 120 lmol photons Á m 2 Á s 1 photosynthetically active radia- gested morphological divergence reflective of tion on a 14:10 L:D photoperiod under Philips fluorescent evolutionary divergence. Indeed, subsequent elec- tubes (Koninklijke Philips Electrons, Amsterdam, the Nether- tron microscopy identified differences in amphies- lands) at a temperature of 26°C. mal plate tabulations between the species (Lee et al. Cultures were photographed during log phase growth 2014). Because ecological and morphological data under bright-field illumination at a magnification of 400–1,0009 using an Olympus BX61 compound microscope have thus far corresponded to concordant phyloge- (Olympus Corp., Tokyo, Japan) with a Jenoptik ProgRes CF netic data from multiple DNA markers (LaJeunesse Scan digital camera (Jenoptik, Jena, Germany). Uncultured et al. 2012, 2014, 2015, Jeong et al. 2014, Hume tissue homogenates were photographed under identical con- et al. 2015, Lee et al. 2015), observing the same ditions. Cell length and width for at least 40 individuals per type of genetic patterns as found between S. minu- culture or tissue sample were calculated with the program ImageJ (Abramoff et al. 2004), and used to calculate ellipsoid tum and S. psygmophilum among other groups is a p 9 strong indication that they represent different spe- volume as (4/3) abc, where a, b, and c are equal to half the length, width, and height, respectively (cell height was cies. Note, however, that the absence of strong assumed to be identical to cell width). The data were neither genetic patterns does not necessarily reflect the normally distributed nor homoscedastic, so size differences absence of isolation, given the spectrum of evolu- between putative species were assessed with a nonparametric tionary divergence between species that have just Kruskal–Wallis test. Significant groupings were determined 852 JOHN EVERETT PARKINSON ET AL. with a KW post hoc multiple comparison test implemented in S. minutum and S. psygmophilum (Fig. 1). Crucially, the R package ‘pgirmess’ at P < 0.05. Each species was repre- although not all species were resolved at every mar- sented by four strains (either unique cultures or unique host ker (though all were resolved at Sym15), no sepa- colonies). Previously published measurements from the two other described Clade B species (S. minutum and S. psyg- rate gene genealogies were discordant, such that mophilum; LaJeunesse et al. 2012) were included for compari- species monophyly became apparent using the com- son. bined approach (Fig. 2). Reciprocal monophyly DNA extraction, PCR amplification, sequencing, and phylogenetic indicated that genetic recombination has been con- analysis. Whole tissue DNA extractions were performed as strained to members within each phylogenetic described by LaJeunesse (2001); culture DNA extractions grouping, validating that each lineage adhered to were performed as described by Coffroth et al. (1992) with the precepts of the Biological Species Concept. The the addition of a bead-beating step (0.4–0.6 mm glass beads, Sigma) to help rupture the algal cell wall. In addition to the Ecological Species Concept was supported in all nuclear (ITS2, Sym15 flanking region), mitochondrial (cob), instances by the specific host associations among and plastid (cp23S) DNA sequences used for species delin- host-specialized Symbiodinium and the rarity with eation by LaJeunesse et al. (2012), we also sequenced the which cryptic Symbiodinium known primarily from nuclear eukaryotic 28S large ribosomal subunit (LSU). The culture could be detected in the wild. primers and thermal cycler conditions for all markers are Taxonomic descriptions. Symbiodinium aenigmaticum listed in Table S1 in the Supporting Information. J. E. Parkinson, M. A. Coffroth & T. C. LaJeunesse To amplify DNA, reactions were performed in 25 lL vol- umes containing 2.5 lL of 2.5 mM dNTPs, 2.5 lLof25mM sp. nov. MgCl , 2.5 lL standard Taq Buffer (New England Biolabs, Diagnosis. Coccoid cells range in average size from 2 À Ipswich, MA, USA), 0.13 lLof5UÁ lL 1 Taq DNA Poly- 6.7 to 8.4 lm at maximum diameter during log merase (New England Biolabs), 1 lL of each forward and phase growth (Figs. 2 and 3). The combined reverse primer at 10 lM, and 1 lLof5–50 ng DNA template. nucleotide sequences of cp23S (GenBank accession: Products were cleaned with ExoSAP-IT (Affymetrix, Santa KT149335), Sym15 microsatellite flanking region Clara, CA, USA) and directly sequenced on an Applied Bio- (KT149352), nuclear ribosomal ITS2 (KT149340) sciences sequencer (Applied Biosciences, Foster City, CA, USA) at the Pennsylvania State University Genomics Core and LSU (KT149345), and mitochondrial cob Facility. Electropherograms were checked and sequences (KT149327) are characteristic of this species. aligned using CodonCode Aligner software (CodonCode, Holotype designation. United States Natural History Dedham, MD, USA) and MEGA v.6 (Tamura et al. 2013). Museum (USNM) sample US 222922 deposited in Previously published sequence information for cultures of the Type Specimen Collection, USNHM, Smithso- the Clade B species S. minutum and S. psygmophilum (LaJe- nian Institution, Washington, DC, USA. unesse et al. 2012) were included in the primary alignment Authentic strain. CCMP3448 (=mac04-180) depos- for reference. A complete list of tissue sample/culture identi- ties, host origins, and collection locations is provided in ited at the Provasoli-Guillard National Center for Table S2 in the Supporting Information. Raw sequences and Marine Algae and Microbiota, East Boothbay, ME, alignments for each gene can be found in the Penn State USA. ScholarSphere database (https://scholarsphere.psu.edu/col- Type locality. Cultured from tissue of the sclerac- lections/x346dq69q). Phylogenetic analyses were performed * tinian mustard hill coral, Porites astreoides, Florida on aligned data sets in PAUP v.4.a136 (Swofford 2000) Keys, USA, though it is not one of the dominant under maximum parsimony with indels included as a 5th character state. Bootstrap support was calculated with 1,000 symbionts in this host. Appears most closely related replicates. Bayesian posterior probabilities were calculated to ITS2 type B23 found in hydrozoan fire corals of with the software MrBayes v.3.2.3 (Ronquist and Huelsenbeck the genus Millepora (Finney et al. 2010), but differs 2003), using the optimal nucleotide substitution model for by two insertions and five substitutions at this mar- each gene (or a partitioned model for the concatenated ker. sequence) based on corrected Akaike Information Criterion Etymology. The Latin ‘aenigma’ (mystery) refers to as calculated with ModelTest v.3.7 (Posada and Crandall 1998). the obscure ecology of this species, which thrives in culture but has only been detected in the wild once, likely as a background or contaminant symbiont. RESULTS Other notes. The authentic strain was originally iso- Phylogenetic delineation of species. Complementary lated in 2004 by Mary Alice Coffroth; it is also evidence from multiple genetic markers with differ- known as culture mac04-180 from the Buffalo ent evolutionary rates spanning mitochondrial (cob: Undersea Reef Research Culture Center collection. 915 bp), nuclear (ITS2: 253 bp; and LSU: 597 bp), Symbiodinium antillogorgium J.E. Parkinson, M.A. and plastid genomes (cp23S: 600 bp), along with Coffroth & T.C. LaJeunesse sp. nov. the flanking region of a single copy microsatellite Diagnosis. Coccoid cells range in average size from (Sym15: 223 bp), supported the designation of four 7.1 to 8.1 lm at maximum diameter during log phase new Symbiodinium species in Clade B (Figs. 1 and 2). growth (Figs. 2 and 3). The combined nucleotide To aid in visualization, we concatenated all sequences of cp23S (GenBank accession: KT149336), sequences (2,588 bp from five markers), producing Sym15 microsatellite flanking region (KT149353), a phylogeny composed of monophyletic lineages nuclear ribosomal ITS2 (KT149341) and LSU with high bootstrap and Bayesian support, and dis- (KT149346), and mitochondrial cob (KT149328 tinguished these lineages from previously described -KT149331) are characteristic of this species. NEW SYMBIODINIUM SPECIES FROM DIFFERENT GUILDS 853

Abipin03 Abipin04 Abipin19 S. antillogorgium 100/1.0 mac08-0689 (host-specialized) AE5 S. aenigmaticum S. endomadracis Pesci166 10 µm 61/1.0 Pesci214 mac703 S. antillogorgium S. pseudominutum 73/1.0 Mf1.05b S. minutum* rt002 (host-specialized) mac04-180 79/1.0 rt064 mac04-267 rt351 S. aenigmaticum mac04-287 100/1.0 rt012 (cryptic) rt013 mac04-487 S. pseudominutum mac04-488 rt074 95/1.0 (cryptic) 100/1.0 HIAp rt146 Mf10.14b.02 rt147 S. psygmophilum* 91/1.0 PurPFlex Bar85 (host-specialized) 100/1.0 rt005 Bar92 rt141 Be057 99/1.0 S. endomadracis Be058 (host-specialized) Cur09-122 Concatenated Tree Cur09-145 1 Base Pair Change Fl02-41

Pliocene (B19) Radiation Pleistocene (B1) Radiation

FIG. 1. Maximum parsimony phylogeny of the four new and two previously described Clade B Symbiodinium based on the concatenated sequences of cp23S, Sym15 microsatellite ﬂanker region, nuclear ribosomal ITS2 and LSU, and mitochondrial cob. Novel LSU sequences for S. minutum (KTKT149349) and S. psygmophilum (KT149351) were also included. Branch support is indicated by bootstrap values followed by Bayesian posterior probabilities. Values for nodes where support was ≤ 60% using either method were omitted. Asterisks indicate previously described species. The tree’s major axis divides the Pleistocene (B1) and Pliocene (B19) radiations. Inset: light micrographs depict- ing representative cells of each new species.

Holotype designation. USNM sample US 222923 A. elisabethae but as yet not outside the genus Antillo- deposited in the Type Specimen Collection, gorgia. USNHM, Smithsonian Institution, Washington, DC, Symbiodinium endomadracis J.E. Parkinson, M.A. USA. Coffroth & T.C. LaJeunesse sp. nov. Authentic strain. CCMP3449 (=mac08-0689) depos- Diagnosis. Coccoid cells range in average size from ited at the Provasoli-Guillard National Center for 6.1 to 7.2 lm at maximum diameter in hospite Marine Algae and Microbiota, East Boothbay, ME, (Figs. 2 and 3). The combined nucleotide USA. sequences of cp23S (GenBank accession: KT149337), Type locality. Cultured from the gorgonian soft Sym15 microsatellite ﬂanking region (KT149354), coral Antillogorgia bipinnata, Florida Keys, USA. nuclear ribosomal ITS2 (KT149342-KT149343) and Etymology. ‘Antillogorgium’ refers to the genus LSU (KT149347-KT149348), and mitochondrial cob name of the host, Antillogorgia spp., for which this (KT149332-KT149333) are characteristic of this spe- symbiont appears to have a specialized association. cies. Other notes. The authentic strain was originally iso- Holotype designation. USNM sample US 222924 lated in 2008 by Mary Alice Coffroth; it is also deposited in the Type Specimen Collection, known as culture mac08-0689 from the Buffalo USNHM, Smithsonian Institution, Washington, DC, Undersea Reef Research Culture Center collection. USA. The ITS2 type of this species is B1(LaJeunesse Type locality. Collected from the scleractinian hard 2001), which matches S. minutum and other ecologi- coral Madracis spp., Curacßao. cally distinct B1 lineages that possess identical cob Etymology. ‘Endomadracis’ means ‘within Madra- and partial ITS2 sequences but are distinguished at cis,’ referring to the genus name of the host Madra- more rapidly evolving markers (cp23S, Sym15 ﬂank- cis spp., for which this symbiont appears to have a ing region). This species is also detected in the host specialized association. 854 JOHN EVERETT PARKINSON ET AL.

A cob B LSU

S. antillogorgium S. antillogorgium

S. minutum S. minutum S. aenigmaticum S. aenigmaticum

S. endomadracis S. psygmophilum S. endomadracis S. psygmophilum

S. pseudominutum S. pseudominutum

C ITS2 D cp23S

S. antillogorgium S. antillogorgium

S. aenigmaticum S. endomadracis S. endomadracis S. aenigmaticum

S. psygmophilum S. minutum S. psygmophilum S. minutum

S. pseudominutum S. pseudominutum

E Sym15

S. antillogorgium

S. minutum S. aenigmaticum

S. pseudominutum S. psygmophilum

S. endomadracis

FIG. 2. Gene-speciﬁc maximum parsimony trees for all described Clade B Symbiodinium species based on (A) cob, (B) LSU, (C) ITS2, (D) cp23S, or (E) microsatellite Sym15 ﬂanker region sequences (in order of increasing resolution). Line segments below gene names indicate the distance corresponding to one base pair change. To emphasize nondiscordance, species groups are alternately highlighted in light or dark gray for convenience; colors do not necessarily correspond across all phylogenies. Thick lines indicate node support ≥ 60% using both bootstrap and Bayesian posterior probabilities.

Other notes. The holotype tissue is also known as KT149338–KT149339), Sym15 microsatellite ﬂanking sample Cur09-122. This species has yet to be cul- region (KT149355), nuclear ribosomal ITS2 tured successfully for more than a few months. The (KT149344) and LSU (KT149350), and mitochon- symbiont has not yet been detected outside the drial cob (KT149334) are characteristic of this species. genus Madracis. Although species designations Holotype designation. USNM sample US 222925 within Madracis are disputed, this symbiont has been deposited in the Type Specimen Collection, recovered from morphotypes of all currently named USNHM, Smithsonian Institution, Washington, DC, species (Diekmann et al. 2001, 2003, Frade et al. USA. 2008). It can be competitively displaced in certain Authentic strain. CCMP3450 (=rt146) deposited at Madracis hosts by ITS2 types B13 or B15 depending the Provasoli-Guillard National Center for Marine on light or depth conditions (Frade et al. 2008). Algae and Microbiota, East Boothbay, ME, USA. Symbiodinium pseudominutum J.E.Parkinson, Type locality. Cultured from the tissue of the ivory M.A.Coffroth & T.C.LaJeunesse sp. nov. bush coral, Oculina diffusa, Bermuda, though it is Diagnosis. Coccoid cells range in average size from not one of the numerically dominant symbionts in 7.8 to 8.5 lm at maximum diameter during log this host. phase growth (Figs. 2 and 3). The combined nucleo- Etymology. The Latin ‘pseudo’ (false) refers to the tide sequences of cp23S (GenBank accession: morphological similarity but molecular and ecological NEW SYMBIODINIUM SPECIES FROM DIFFERENT GUILDS 855

450 d type B23), S. antillogorgium (ITS2 type B1), S. endo- 400 ) madracis (ITS2 type B7), and S. pseudominutum (ITS2 3 350 type B1). Phylogenetically, Clade B is comprised of 300 c c two major radiations that diverged millions of years 250 b bc – 200 ago. The Pleistocene radiation expanded 1.2 2.1 Ma 150 a and its members share ITS2 types derived from an 100 ancestral B1 sequence, while the Pliocene radiation Cell Volume (µm 50 – B1 B19B1 B1 B1 B19 expanded 2.3 4.0 Ma and its members share ITS2 0 types derived from an ancestral B19 sequence (LaJe- unesse 2005). S. antillogorgium, S. endomadracis, S. pseudominutum, and the previously described S. minutum belong to the Pleistocene (B1) radiation; S. aenigmaticum and the previously described S. psygmophilum belong to the Pliocene (B19) radiation FIG. 3. Average cell volume for each of the four new (shaded) and two previously described (unshaded) Clade B Symbiodinium (Fig. 1). The major bifurcation is apparent at all species. Error bars represent 95% confidence intervals. Lower phylogenetic markers used in this study (Fig. 2). case letters above bars indicate membership in statistically signifi- Each radiation includes representatives of both – cant groupings (Kruskal Wallis; KW multiple comparison host-specialized and cryptic ecologies. P < 0.05). Upper case designations within bars indicate radiation membership. For each species, n = 4 cultures or host colonies, Genetic evidence. All four new species were well each with n ≥ 40 measurements. supported by the concatenated gene tree (Fig. 1) and by the nondiscordant genealogies among unlinked loci (Fig. 2). Certain genes lacked resolv- disparity between this species and the closely related ing power when considered in isolation, though S. minutum. none of them conflicted with other gene genealo- Other notes. The authentic strain was originally iso- gies. For example, S. endomadracis, S. minutum, and lated in the mid-1970s by David A. Schoenberg; it is S. pseudominutum shared identical sequences at also known as culture rt146 from the Robert K. cp23S, but clear divergence at the Sym15 flanker Trench collection. The ITS2 type of this species is region (Fig. 2). This situation is similar to S. boreum B1(LaJeunesse 2001), which matches S. minutum and S. eurythalpos—two Clade D species which are and other ecologically distinct B1 lineages that pos- readily distinguished by mitochondrial cob sess identical cob and partial ITS2 sequences but are sequences, rDNA (ITS1-5.8S-ITS2-LSU), and distinguished at the more rapidly evolving Sym15 microsatellite flanker region and allele data but nev- flanking region. This species has also been cultured ertheless share identical cp23S sequences (LaJe- independently from the tissues of diverse hosts unesse et al. 2014). Similarly, a lack of variation at including Cassiopeia xamachana, Aiptasia sp., and An- ITS2 can mask species diversity in some groups tillogorgia sp. from the Caribbean. Another culture (e.g., type B1 shared by S. antillogorgium, S. minutum, of this species, rt013, was informally described as S. pseudominutum, and many other lineages; Santos Symbiodinium ‘bermudense’ nomen nudum (Banas- et al. 2004, Finney et al. 2010, Thornhill et al. 2014, zak et al. 1993, LaJeunesse 2001). This name was this study). On the other hand, variation can exist subsequently used by many researchers to refer to in the dominant or codominant ITS sequences diag- the dominant symbiont of Aiptasia spp., when in fact nostic of members of a single species (e.g., S. eury- they were most likely working with S. minutum (LaJe- thalpos, encompassing types D8, D8-12, D12-13, and unesse et al. 2012). Symbiodinium ‘bermudense’ D-13; LaJeunesse et al. 2014). These scenarios sensu stricto should now be considered synonymous emphasize that a combined approach using multiple with S. pseudominutum. markers is required for effective Symbiodinium spe- Cell morphology. Under simple light microscopy, the cies delineation (LaJeunesse et al. 2012, Stat et al. new species shared similar morphological characteris- 2012). tics (Fig. 1 Inset). S. endomadracis had the smallest Morphological evidence. There are upper and lower average volume (90% confidence interval: 119– cell size limits among Symbiodinium, and previous 133 lm3), significantly smaller than S. antillogorgium studies have found inequalities in cell size that are (203–227 lm3)andS. pseudominutum (222–252 lm3), genetically influenced (Schoenberg and Trench while S. aenigmaticum (179–213 lm3) was of an inter- 1980a,b,c, LaJeunesse et al. 2012, 2014, Lee et al. mediate size (Fig. 3). S. minutum was similar in size to 2014). Differences in cell volume can significantly S. antillogorgium and S. pseudominutum, while S. psyg- affect the physiological differences among species mophilum was larger than all other Clade B species. (Banse 1976). Within the B19 radiation, S. aenigmaticum was smaller than S. psygmophilum. Within the Pleistocene (B1) radiation, only S. endomadracis DISCUSSION was completely distinct. Across radiations, S. aenig- We resolved four new species lineages within Sym- maticum (Pliocene-B19) could not be confidently dis- biodinium Clade B: S. aenigmaticum (related to ITS2 tinguished from S. antillogorgium (Pleistocene-B1), 856 JOHN EVERETT PARKINSON ET AL. showing that highly divergent lineages may never- symbiont may prevent them from reaching larger theless overlap in cell size. This result is consistent population sizes in hospite, and thus they have with previous findings (Trench and Blank 1987), remained undetected using traditional techniques. and reinforces the concept that genetic data are We refer to such Symbiodinium as ecologically ‘cryp- more reliable than certain morphometric data for tic.’ They may be well-known (e.g., S. pilosum; this group. Trench and Blank 1987) or new to science (such as Host-specialized species. Two new Clade B Symbio- those characterized here). The new Clade B species dinium (S. antillogorgium and S. endomadracis) appear S. pseudominutum exemplifies this uncertainty to be specialized and are therefore named after regarding ecological function in the ecosystem their hosts, the octocoral genus Antillogorgia and the where the organism occurs. Despite its shared evolu- scleractinian genus Madracis, respectively. We use tionary history with S. minutum, which is commonly the term ‘host-specialized’ to refer to an ecological detected in hosts collected from nature, S. pseu- guild encompassing mutualistic Symbiodinium that dominutum has rarely been isolated (only four times associate persistently with their hosts as the numeri- total, as a contaminant from Aiptasia, Antillogorgia, cally dominant symbiont. Although capable of sur- Cassiopeia, and Oculina). Compared to S. minutum, viving outside of the host, they give the impression S. pseudominutum exhibits a distinct gene expression of being obligate endosymbionts because few are profile (Parkinson 2014) and features lower half-sat- able to be cultured long-term (see discussion below) uration constants for photosynthesis under thermal and rarely are they detected in the water column. stress (Oakley et al. 2014). Instead, they predictably associate with certain hosts, Another cryptic species, S. aenigmaticum, is geneti- and are almost always recovered in hospite. Host- cally distinct from other members of its Pliocene specialized Symbiodinium are restricted to a small (B19) radiation. It has only been cultured from the number of host species and likely have coexisted tissues of one host population: Porites astreoides from with a particular host species over millions of years the Florida Keys. The coral is never dominated by (Thornhill et al. 2014). S. aenigmaticum, and yet 11 independent cultures Consistent with other dinoflagellate groups, we from multiple host individuals have recovered this recognize that cultures represent only a fraction of symbiont (M.A. Coffroth, personal observation). It is natural diversity (Stern et al. 2010), and only a small therefore incredibly successful at growing in vitro number of host-specialized Symbiodinium can be cul- despite its apparent existence as a background pop- tured with current techniques. Despite multiple ulation associated with P. astreoides at sub-tropical attempts, host-specialized symbionts rarely survive latitudes. The phylogenetically closest lineage to for more than a few months without being replaced S. aenigmaticum, type B23 (currently unnamed), by contaminants or dying out entirely (Santos et al. occurs in the fire coral genus Millepora from the 2001, LaJeunesse 2002, Krueger and Gates 2012). eastern Caribbean (Finney et al. 2010). These symbionts appear to suffer diminished vitality when removed from the host habitat, whereas free- living or generalist species have an advantage CONCLUSION in vitro. The current inability to culture these host- There are likely numerous other species whose specialized Symbiodinium from host tissues is likely a ecology differs from mutualistic species (e.g., S. vo- technical limitation rather than a biological one. ratum, Jeong et al. 2014, S. necroappetans, LaJeunesse Although S. endomadracis fails in culture, S. antillo- et al. 2015). More of these may be cultured in the gorgium is a host-specialized species that nevertheless future or discovered with deep-sequencing tech- cultures well. The previously described Clade B spe- niques (Arif et al. 2014, Thomas et al. 2014). Never- cies S. minutum and S. psygmophilum are both cultur- theless, preliminary genetic data indicate that there able and ecologically common, and frequently are many ecologically distinct species without formal dominate the tissues of compatible hosts (LaJe- description within Clade B (Finney et al. 2010, unesse et al. 2012). S. psygmophilum can associate Prada et al. 2014). Given the likelihood that other with several different host genera, whereas S. antillo- Clades have similar or greater numbers of species gorgium has yet to be identified outside the genus (LaJeunesse et al. 2014, Thornhill et al. 2014, but Antillogorgia. Extensive investigations conducted on see Jeong et al. 2014), it is critical to recognize that other gorgonians will be necessary to assess the resolution at the level of Clades (and in some cases, extent of this Symbiodinium’s specificity. ITS2 types) is insufficient for addressing many ques- Ecologically cryptic species. The ‘contaminants’ that tions about the ecology and evolution of cnidarian- supplant host-specialized symbionts during the cul- Symbiodinium associations. Resolution of species will turing process may occupy one of several separate facilitate a more sophisticated understanding of ecological guilds. They might be completely free-liv- these systems. ing, commensal, or perhaps endosymbiotic but restricted in their capacity to achieve high densities We thank Robert K. Trench for reviewing an early draft of in the host’s tissues (Schoenberg and Trench the manuscript. Thanks also to the many researchers involved 1980c). The presence of a competitively dominant in sample collection throughout the Caribbean, to Allison M. NEW SYMBIODINIUM SPECIES FROM DIFFERENT GUILDS 857

Lewis and Drew C. Wham for technical support, to the Aquar- 2014. Genetics and morphology characterize the dinoflagel- ium of Niagara for providing seawater to maintain the late Symbiodinium voratum, n. sp., (Dinophyceae) as the sole BURRCC, and to the Penn State Microscopy and Cytometry representative of Symbiodinium clade E. J. Eukaryot. Microbiol. Facility, University Park, Pennsylvania, USA. Funding was pro- 61:75–94. vided by the U.S. National Science Foundation (NSF DEG- Krueger, T. & Gates, R. D. 2012. Cultivating endosymbionts - host environmental mimics support the survival of Symbiodinium 0750756 to J.E.P., NSF OCE-0926822 to M.A.C., NSF OCE- – 0928764 to T.C.L.). C15 ex hospite. J. Exp. Mar. Biol. Ecol. 413:169 76. LaJeunesse, T. C. 2001. Investigating the biodiversity, ecology, and phylogeny of endosymbiotic dinoflagellates in the genus Abramoff, M. D., Magalhaes, P. J. & Ram, S. J. 2004. Image pro- Symbiodinium using the ITS region: in search of a “species” – cessing with imageJ. Biophotonics Int. 11:36 42. level marker. J. Phycol. 37:866–80. Ahles, M. D. 1967. Some aspects of the morphology and physiol- LaJeunesse, T. C. 2002. Diversity and community structure of sym- ogy of Symbiodinium microadriaticum. PhD dissertation, Ford- biotic dinoflagellates from Caribbean coral reefs. Mar. Biol. ham University, New York, 158 pp. 141:387–400. Arif, C., Daniels, C., Bayer, T., Banguera-Hinestroza, E., Barbrook, LaJeunesse, T. C. 2005. “Species” radiations of symbiotic dinoflag- A. C., Howe, C. J., LaJeunesse, T. C. & Voolstra, C. R. 2014. ellates in the Atlantic and Indo-Pacific since the Miocene- Assessing Symbiodinium diversity in scleractinian corals via Pliocene transition. Mol. Biol. Evol. 22:570–81. next-generation sequencing-based genotyping of the ITS2 LaJeunesse, T. C., Lambert, G., Andersen, R. A., Coffroth, M. A. – rDNA region. Mol. Ecol. 23:4418 33. & Galbraith, D. W. 2005. Symbiodinium (Pyrrhophyta) gen- Avise, J. C. & Wollenberg, K. 1997. Phylogenetics and the origin ome sizes (DNA content) are smallest among dinoflagellates. – of species. Proc. Natl. Acad. Sci. USA 94:7748 55. J. Phycol. 41:880–6. Banaszak, A. T., Iglesias-Prieto, R. & Trench, R. K. 1993. Scripp- LaJeunesse, T. C., Lee, S. Y., Gil-Agudelo, D., Knowlton, N. & siella velellae sp. nov. (Peridiniales) and Gloeodinium viscum sp. Jeong, H. J. 2015. Symbiodinium necroappetens sp. nov. (Dino- nov., dinoflagellate symbionts of two hydrozoans (Cnidaria). phyceae), an opportunistic “symbiont” associated with the – J. Phycol. 29:517 28. bleached and diseased tissues of Caribbean reef corals. Eur. Banse, K. 1976. Rates of growth, respiration and photosynthesis J. Phycol. 50:223–38. – of unicellular algae as related to cell size a review. J. Phycol. LaJeunesse, T. C., Parkinson, J. E. & Reimer, J. D. 2012. A genet- – 12:135 40. ics-based description of Symbiodinium minutum sp. nov. and S. Blank, R. J. & Huss, V. A. R. 1989. DNA divergency and specia- psygmophilum sp. nov. (Dinophyceae), two dinoflagellates sym- tion in Symbiodinium (Dinophyceae). Plant Syst. Evol. biotic with cnidaria. J. Phycol. 48:1380–91. – 163:153 63. LaJeunesse, T. C., Smith, R., Walther, M., Pinzon, J., Pettay, D. T., Carlos, A. A., Baillie, B. K., Kawachi, M. & Maruyama, T. 1999. McGinley, M., Aschaffenburg, M. et al. 2010. Host-symbiont Phylogenetic position of Symbiodinium (Dinophyceae) isolates recombination versus natural selection in the response of from tridacnids (Bivalvia), cardiids (Bivalvia), a sponge (Pori- coral-dinoflagellate symbioses to environmental disturbance. fera), a soft coral (Anthozoa), and a free-living strain. J. Phy- Proc. R. Soc. Lond. B 277:2925–34. – col. 35:1054 62. LaJeunesse, T. C., Wham, D. C., Pettay, D. T., Parkinson, J. E., Ke- Coffroth, M. A., Lasker, H. R., Diamond, M. E., Bruenn, J. A. & shavmurthy, S. & Chen, C. 2014. Ecologically differentiated Bermingham, E. 1992. DNA fingerprints of a gorgonian stress tolerant endosymbionts in the dinoflagellate genus coral - a method for detecting clonal structure in a vegetative Symbiodinium Clade D are different species. Phycologia 53:305– – species. Mar. Biol. 114:317 25. 19. Diekmann, O. E., Bak, R. P. M., Stam, W. T. & Olsen, J. L. 2001. Lee, S. Y., Jeong, H. J., Kang, N. S., Jang, T. Y., Jang, S. H. & Molecular genetic evidence for probable reticulate speciation Lim, A. S. 2014. Morphological characterization of Symbio- in the coral genus Madracis from a Caribbean fringing reef dinium minutum and S. psygmophilum belonging to clade B. Al- – slope. Mar. Biol. 139:221 33. gae 29:299–310. Diekmann, O. E., Olsen, J. L., Stam, W. T. & Bak, R. P. M. 2003. Lee, S. Y., Jeong, H. J., Kang, N. S., Jang, T. Y. & LaJeunesse, T. Genetic variation within Symbiodinium clade B from the coral C. 2015. Symbiodinium tridacnidorum sp. nov., a dinoflagellate genus Madracis in the Caribbean (Netherlands Antilles). Co- common to Indo-Pacific giant clams, and a revised morpho- – ral Reefs 22:29 33. logical description of Symbiodinium microadriaticum, emended Finney, J. C., Pettay, D. T., Sampayo, E. M., Warner, M. E., Oxen- Trench et Blank. Eur. J. Phycol. 50:155–72. ford, H. A. & LaJeunesse, T. C. 2010. The relative signifi- Leliaert, F., Verbruggen, H., Vanormelingen, P., Steen, F., cance of host-habitat, depth, and geography on the ecology, Lopez-Bautista, J. M., Zuccarello, G. C. & De Clerck, O. endemism, and speciation of coral endosymbionts in the 2014. DNA-based species delimitation in algae. Eur. J. Phycol. – genus Symbiodinium. Microb. Ecol. 60:250 63. 49:179–96. Frade, P. R., Englebert, N., Faria, J., Visser, P. M. & Bak, R. P. M. Mayr, E. 1942. Systematics and the Origin of Species, from the Viewpoint 2008. Distribution and photobiology of Symbiodinium types in of a Zoologist. Harvard University Press, Massachusetts, 337 pp. different light environments for three colour morphs of the Oakley, C. A., Schmidt, G. W. & Hopkinson, B. M. 2014. Thermal coral Madracis pharensis: is there more to it than total irradi- responses of Symbiodinium photosynthetic carbon assimila- – ance? Coral Reefs 27:913 25. tion. Coral Reefs 33:501–12. Freudenthal, H. D. 1962. Symbiodinium gen. nov. and Symbiodinium Parkinson, J. E. 2014. The role of intraspecific diversity in coral-al- microadriaticum sp. nov., a zooxanthella: taxonomy, life cycle, gal symbiosis ecology and evolution. PhD dissertation, The – and morphology. J. Protozool. 9:45 52. Pennsylvania State University, Pennsylvania, 221 pp. Goulet, T. L. & Coffroth, M. A. 2003. Stability of an octocoral-al- Pettay, D. T. & LaJeunesse, T. C. 2007. Microsatellites from clade gal symbiosis over time and space. Mar. Ecol. Prog. Ser. B Symbiodinium spp. specialized for Caribbean corals in the – 250:117 24. genus Madracis. Mol. Ecol. Notes 7:1271–4. Hansen, G. & Daugbjerg, N. 2009. Symbiodinium natans sp. nov.: a Posada, D. & Crandall, K. A. 1998. MODELTEST: testing the “free-living” dinoflagellate from Tenerife (Northeast-Atlantic model of DNA substitution. Bioinformatics 14:817–8. – Ocean). J. Phycol. 45:251 63. Prada, C., McIlroy, S. E., Beltran, D. M., Valint, D. J., Santos, S. Hume, B. C. C., D’Angelo, E. G., Smith, E. G., Stevens, J. R., Burt, R., Ford, S. A., Hellberg, M. E. & Coffroth, M. A. 2014. Cryp- J. & Wiedenmann, J. 2015. Symbiodinium thermophilum sp. nov., tic diversity hides host and habitat specialization in a gor- a thermotolerant symbiotic alga prevalent in corals of the gonian-algal symbiosis. Mol. Ecol. 23:3330–40. world’s hottest sea, the Persian/Arabian Gulf. Sci. Rep. 5:8562. Ronquist, F. & Huelsenbeck, J. P. 2003. MrBayes 3: Bayesian phy- Jeong, H. J., Lee, S. Y., Kang, N. S., Yoo, Y. D., Lim, A. S., Lee, logenetic inference under mixed models. Bioinformatics M. J., Kim, H. S., Yih, W., Yamashita, H. & LaJeunesse, T. C. 19:1572–4. 858 JOHN EVERETT PARKINSON ET AL.

Rowan, R. 1998. Diversity and ecology of zooxanthellae on coral Western Australia with 454 amplicon pyrosequencing. Mol. reefs. J. Phycol. 34:407–17. Ecol. 23:3113–26. Sampayo, E. M., Dove, S. & LaJeunesse, T. C. 2009. Cohesive Thornhill, D. J., Lewis, A. M., Wham, D. C. & LaJeunesse, T. C. molecular genetic data delineate species diversity in the 2014. Host-specialist lineages dominate the adaptive radia- dinoflagellate genus Symbiodinium. Mol. Ecol. 18:500–19. tion of reef coral endosymbionts. Evolution 68:352–67. Santos, S. R., Shearer, T. L., Hannes, A. R. & Coffroth, M. A. Trench, R. K. 1993. Microalgal-invertebrate symbioses: a review. 2004. Fine-scale diversity and specificity in the most prevalent Endocyt. Cell. Res. 9:135–75. lineage of symbiotic dinoflagellates (Symbiodinium, Dino- Trench, R. K. & Blank, R. J. 1987. Symbiodinium microadriaticum phyceae) of the Caribbean. Mol. Ecol. 13:459–69. Freudenthal, Symbiodinium goreauii sp. nov., Symbiodinium Santos, S. R., Taylor, D. J. & Coffroth, M. A. 2001. Genetic com- kawagutii sp. nov. and Symbiodinium pilosum sp. nov. - gymno- parisons of freshly isolated versus cultured symbiotic dinioid dinoflagellate symbionts of marine invertebrates. J. dinoflagellates: implications for extrapolating to the intact Phycol. 23:469–81. symbiosis. J. Phycol. 37:900–12. Trench, R. K. & Thinh, L. V. 1995. Gymnodinium linucheae sp. nov. Schoenberg, D. A. & Trench, R. K. 1980a. Genetic variation in - the dinoflagellate symbiont of the jellyfish Linuche unguicu- Symbiodinium (=Gymnodinium) microadriaticum Freudenthal, lata. Eur. J. Phycol. 30:149–54. and specificity in its symbiosis with marine invertebrates. 1. Van Valen, L. 1976. Ecological species, multispecies, and oaks. Isoenzyme and soluble-protein patterns of axenic cultures of Taxon 25:233–9. Symbiodinium microadriaticum. Proc. R. Soc. Lond. B 207:405– von Wintzingerode, F., Gobel, U. B. & Stackebrandt, E. 1997. 27. Determination of microbial diversity in environmental sam- Schoenberg, D. A. & Trench, R. K. 1980b. Genetic variation in ples: pitfalls of PCR-based rRNA analysis. FEMS Microbiol. Rev. Symbiodinium (=Gymnodinium) microadriaticum Freudenthal, 21:213–29. and specificity in its symbiosis with marine invertebrates. 2. Morphological variation in Symbiodinium microadriaticum. Proc. R. Soc. Lond. B 207:429–44. Schoenberg, D. A. & Trench, R. K. 1980c. Genetic variation in Supporting Information Symbiodinium (=Gymnodinium) microadriaticum Freudenthal, and specificity in its symbiosis with marine invertebrates. 3. Additional Supporting Information may be Specificity and infectivity of Symbiodinium microadriaticum. Proc. R. Soc. Lond. B 207:445–60. found in the online version of this article at the Seutin, G., White, B. N. & Boag, P. T. 1991. Preservation of avian publisher’s web site: blood and tissue samples for DNA analyses. Can. J. Zool. 69:82–90. Table S1. Gene regions targeted for analyses, Stat, M., Baker, A. C., Bourne, D. G., Correa, A. M. S., Forsman, gene types, primer pairs used for PCR, primer Z., Huggett, M. J., Pochon, X., Skillings, D., Toonen, R. J., sequences, approximate sizes of amplified DNA van Oppen, M. J. H. & Gates, R. D. 2012. Molecular delin- – fragments, and annealing temperatures used to eation of species in the coral holobiont. Adv. Mar. Biol. 63:1 delineate species in Clade B of the genus Symbio- 65. Stern, R. F., Horak, A., Andrew, R. L., Coffroth, M. A., Andersen, dinium. R. A., Kupper, F. C., Jameson, I. et al. 2010. Environmental barcoding reveals massive dinoflagellate diversity in marine Table S2. Additional details for samples used in environments. PLoS ONE 5:e13991. species delineation, including sample ID, species Swofford, D. 2000. PAUP*, Phylogenetic Analysis Using Parsimony membership, material type (either culture or tis- (*and Other Methods). 4.0b10. Sinauer Associates, Sunderland, sue), host species (note that cultures isolated Massachusetts. from a given “host” may not be representative of Tamura, K., Stecher, G., Peterson, D., Filipski, A. & Kumar, S. 2013. MEGA6: molecular evolutionary genetics analysis ver- the dominant symbiont in that species), and col- sion 6.0. Mol. Biol. Evol. 30:2725–9. lection location. Thomas, L., Kendrick, G. A., Kennington, W. J., Richards, Z. T. & Stat, M. 2014. Exploring Symbiodinium diversity and host specificity in Acropora corals from geographical extremes of