Molecular Phylogenetics and Evolution 77 (2014) 11–22
Contents lists available at ScienceDirect
Molecular Phylogenetics and Evolution
journal homepage: www.elsevier.com/locate/ympev
Morphological and host specificity evolution in coral symbiont barnacles (Balanomorpha: Pyrgomatidae) inferred from a multi-locus phylogeny ⇑ Ling Ming Tsang a, Ka Hou Chu b, Yoko Nozawa c, Benny Kwok Kan Chan c, a Institute of Marine Biology, National Taiwan Ocean University, Keelung, Taiwan b Simon F.S. Li Marine Science Laboratory, School of Life Sciences, The Chinese University of Hong Kong, Shatin, Hong Kong c Biodiversity Research Center, Academia Sinica, Taipei 115, Taiwan article info abstract
Article history: Coral-inhabiting barnacles (Thoracica: Pyrgomatidae) are obligatory symbionts of scleractinian and fire Received 27 July 2013 corals. We attempted to reconstruct the phylogeny of coral-inhabiting barnacles using a multi-locus Revised 15 February 2014 approach (mitochondrial 12S and 16S rRNA, and nuclear EF1, H3 and RP gene sequences, total Accepted 3 March 2014 3532 bp), which recovered a paraphyletic pattern. The fire-coral inhabiting barnacle Wanella milleporae Available online 15 March 2014 occupied a basal position with respect to the other coral inhabiting barnacles. Pyrgomatids along with the coral-inhabiting archaeobalanid Armatobalanus nested within the same clade and this clade was sub- Keywords: divided into two major lineages: Armatobalanus + Cantellius with species proposed to be the ancestral Cirripedia stock of extant coral barnacles, and the other comprising the remaining genera studied. Ancestral state Ancestral state reconstruction Host shift reconstruction (ASR) suggested multiple independent fusions and separations of shell plates and opercu- Host switching lar valves in coral barnacle evolution, which counters the traditional hypothesis founded on a scheme of Symbiosis morphological similarities. Most of the coral barnacles are restricted to one or two coral host families Coral reef only, suggesting a trend toward narrow host range and more specific adaptation. Furthermore, there is a close linkage between coral host usage and phylogenetic relationships with sister taxa usually being found on the same coral host family. This suggests that symbiotic relationships in coral-inhabiting bar- nacles are phylogenetically conserved and that host associated specialization plays an important role in their diversification. Ó 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/).
1. Introduction evolution of symbiotic relationships not only in barnacles, but also in marine fauna in general (Sotka, 2005). Coral-inhabiting barnacles (Thoracica: Balanomorpha: The pyrgomatid barnacles have cup-shaped or tube-shaped Pyrgomatidae) are obligatory symbionts of scleractinian and fire bases that are embedded deeply in the skeleton of host corals. corals (Ross and Newman, 1973). Stable isotope ratios of C12 and Other than pyrgomatid barnacles, balanomorph species of the C13 on corals and coral barnacles (Achituv et al., 1997) suggest that genus Armatobalanus (family Archaeobalanidae, e.g. Armatobalanus corals contribute one of the carbon sources to the barnacles as the allium) and Megabalanus (family Balanidae, e.g. Megabalanus ajax barnacles feed on the coral organic matter and zooxanthellae and M. stultus) also live symbiotically with corals (Liu and Ren, expelled by the corals. On the other hand, the ammonium excreted 2007; Ross, 1999). These species bear six-plated shells that are dis- by the coral barnacles is absorbed by the coral zooxanthellae con- tinguished from the four-plated shells or fused shells found in tributing to coral growth (Achituv and Mizrahi, 1996). Evolution pyrgomatids (Fig. 1; Ross and Newman, 1973; Anderson, 1992). and adaptation to an obligate symbiotic relationship with other The fused shell is unique to pyrgomatids and believed to be an marine fauna including corals and sponges have contributed to adaptation to a symbiotic lifestyle (Anderson, 1992), and in an increased diversity of marine fauna (e.g. Duffy, 1996; Munday response to overgrowth of corals over their surface (Barnes et al., et al., 2004; Sotka, 2005; Macdonald et al., 2006; Faucci et al., 1970). In pyrgomatid barnacles, adaptations include wall plates 2007). Therefore elucidating the evolutionary history of this obli- that have flattened to form a discoidal-shaped apex and a base that gate symbiotic life style would provide valuable insights into the has become cup-shaped. Moreover, their shell plates are reduced from six (a plesiomorphic condition in most of the balanomorph barnacles) to four (carina, rostrum and paired latera) or even fused ⇑ Corresponding author. Fax: +886 02 27899624. into a single shell (Fig. 1A; Anderson, 1992) to provide improved E-mail address: [email protected] (B.K.K. Chan). http://dx.doi.org/10.1016/j.ympev.2014.03.002 1055-7903/Ó 2014 The Authors. Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/3.0/). 12 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Fig. 1. (A) Shell structure and opercular valve morphology of various Pyrgomatidae genera (Carina [C], Rostrum [R], and two Laterals [L]). Hypotheses of phylogenetic relationship among pyrgomatid genera based on morphological similarities as proposed by (B) Ross and Newman (1973); and (C) Anderson (1992). strength between apex and base. In addition to strengthened considered key characters in traditional phylogenies given their shells, pyrgomatid barnacles erect an aperture frill over the orifice important role in adaptation to a symbiotic lifestyle. Ross and which is believed to be a glandular fold that inhibits the growth of Newman (1973) presented a non-cladistic phylogenetic tree of corals. Secretions from the frill act to chemically inhibit coral over- Pyrgomatidae (Fig. 1B). They suggested Pyrgomatidae evolved growth (Anderson, 1992). from an Armatobalanus-like (family Archaeobalanidae) ancestor The morphology of the shell walls, opercular plates and aper- that had a six-plated shell and separate opercular valves. The ture frills of coral-inhabiting barnacles have, therefore been four-plated shell and separated opercular valves observed in L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22 13
Cantellius represent a plesiomorphic state in the Pyrgomatidae 2. Materials and methods (Fig. 1B). Two major lineages have evolved from Cantellius; the first lineage began with Hiroa (four-plated shells and separated, modi- 2.1. Taxon sampling fied opercular valves; Fig. 1A) while the other originated with Savignium (including today’s Trevathana, Neotrevathana and Wanel- Coral symbiont barnacles were collected in outlying islands of la with fused shell and separated opercular valves; Fig. 1A). The Taiwan by SCUBA diving at depth of 5–25 m (for site locations, other genera were subsequently derived from these two stem please see Chan et al., 2013b and Table 1). Pieces of corals groups through further modifications in shell and opercular valves ðv7cm2Þ bearing symbiotic barnacles were photographed in situ morphologies (Fig. 1A and B). Anderson (1992) proposed an alter- underwater and then collected using a hammer and chisel. Coral native phylogenetic hypothesis for Pyrgomatinae based on the barnacles were dissected and preserved in 95% ethanol upon col- opercular valve structure, functional morphology and cirral activ- lection. The host corals were identified from the preserved coral ity. Likewise, he also considered Cantellius, specifically Cantellius pieces and in situ photographs. We also included seven species of euspinulosum, as the ancestral stock of extant coral-inhabiting bar- Archaeobalanidae and Balanidae, which belong to the same super- nacles derived from a balanoid-like ancestor (six-plated shell) family (Balanoidea) as the Pyrgomatidae, for outgroup comparison. (Fig. 1C). Cantellius euspinulosum further diverged into three lin- eages, namely the ‘‘septimus’’ group (type I aperture frill mecha- nism defined in Anderson, 1992), the ‘‘pallidus’’ group (type II 2.2. Laboratory protocol and phylogenetic analyses aperture frill mechanism) and the ‘‘secundus’’ group (type III aper- ture frill mechanism). Other coral barnacle genera evolved from Total genomic DNA was extracted from adductor or abdominal these three stem groups following a shell plate reduction gradient muscle tissue using the commercial QIAamp Tissue Kit (QIAGEN). from ancestors (four-plated shell) to descendants (fused shell) We first sequenced the mitochondrial 12S rRNA gene sequences (Fig. 1C). from all coral barnacle samples collected to confirm the species A recent molecular phylogenetic study, however, does not sup- identity and detect the presence of any cryptic species that may port these morphology-based hypotheses and proposes a highly exhibit differential host specificity (see Tsang et al., 2009; Chan contrasting scenario (Simon-Blecher et al., 2007). Their phyloge- et al., 2013a). Based on the 12S data, we selected a subset of repre- netic tree inferred from mitochondrial 12S and 16S rDNA, and sentative individuals for further multi-locus phylogenetic infer- the nuclear 18S rDNA, has placed Armatobalanus within Pyrgomat- ence to achieve more robust phylogeny (Table 1) and idae, whilst the fire-coral inhabiting barnacle Wanella is more clo- morphological examination (including shell structure and opercu- sely related to the other free-living balanids. Archaeobalanidae is lar valves). The coral host usage data for all coral barnacle species thus polyphyletic and Armatobalanus does not represent the evolu- were also recorded (see Table 1) and these provided baseline infor- tionary stem of the Pyrgomatidae. The tree also suggests the num- mation for ancestral state reconstruction (ASR) in subsequent anal- ber of shell plates and the modification of opercular valves are yses (see below). homoplasious features which arose multiple times in coral-barna- We sequenced four additional molecular markers, the mito- cle evolutionary history (Simon-Blecher et al., 2007). However, this chondrial 16S rRNA gene and the nuclear elongation factor 1a sub- study includes only a limited number of taxa and molecular mark- unit (EF1), RNA polymerase subunit II (RPII) and histone 3 (H3) ers, resulting in many internal nodes with poor support requiring gene segments from the selected individuals. Information on the further verification. primers adopted and the annealing temperature in PCR profiles Previous phylogenies have focused almost exclusively on the for the five loci are listed in Table 2. Amplicons were then purified evolution of morphological characters (e.g. Ross and Newman, using either the QIAquick gel purification kit (QIAGEN) or QIAquick 1973; Newman and Ross, 1976; Anderson, 1992; Simon-Blecher PCR purification kit according to the manufacturer’s instructions. et al., 2007). The evolutionary history of host specificity and Sequencing reactions were carried out using the same sets of prim- switching has, however, rarely been explored in the coral barna- ers and the ABI Big-dye Ready-Reaction mix kit following the stan- cles. This is partly attributed to the lack of a robust phylogeny for dard cycle sequencing protocol on the ABI3700 automated this group. Furthermore, information on the range of coral hosts sequencer. utilized by individual barnacle species is far from complete. Many Sequences were aligned with MUSCLE (Edgar, 2004) using the coral barnacle species that were once considered generalists that default parameter settings and then checked by eye. The concate- inhabit a number of coral hosts (e.g. Wanella milliporae, Treva- nated dataset was analyzed using maximum likelihood (ML) and thana dentatum, Galkinia indica) were recently found, using Bayesian inference (BI) approaches. We used PartitionFinder molecular tools, to be a collection of morphologically similar or v1.0.1 (Lanfear et al., 2012) to determine the best partitioning even indistinguishable cryptic species, with high host specificity strategy for the concatenated dataset according to the Bayesian (Mokady et al., 1999; Tsang et al., 2009; Brickner et al., 2010; information criterion (BIC) and ‘‘greedy’’ search strategy. The anal- Chan et al., 2013a). To understand the evolution of the symbiotic ysis suggested splitting the dataset into three partitions: (1) the lifestyle in coral barnacles, it is essential to collect the barnacles mitochondrial 16S + 12S, (2) the first + second codon position for with their species identity and host data recorded accurately as the three nuclear protein-coding genes, and (3) the third codon baseline data for mapping on a molecular phylogeny. position for the three nuclear protein-coding genes, as the best par- In the present study, we attempted to reconstruct a robust titioning scheme. PartitionFinder further suggested GTR + I + G as molecular phylogeny of the Pyrgomatidae and its allies using mul- the best-fit nucleotide substitution model for partitions (1) and tiple molecular markers from mitochondrial and nuclear genomes (3), and GTR + I for partition (2). The ML analysis was implemented to test the various hypotheses concerning coral-inhabiting barna- with RAxML 7.0.3 (Stamatakis, 2006) and the model GTRGAMMAI cle evolution. In addition, we quantitatively investigated the host was used for all the three partitions. The gamma individual-shape usage pattern by different barnacle species based on molecular parameters, GTR-rates and base frequencies were estimated and data. Combination of the molecular phylogeny and host usage data optimized for each partition during analyses. We conducted 1000 allows us to explore the evolution of host switching and host bootstrap (BP) runs and searched for the best-scoring ML tree. specificity in individual barnacle species by ancestral state recon- Bayesian analysis was conducted using MrBayes v.3.2.1 (Ronquist struction that represent the most comprehensive effort for the et al., 2012) with two independent runs carried out with four dif- group. ferentially heated Metropolis coupled Monte Carlo Markov Chains 14 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Table 1 Sampling localities, coral host identities of the barnacle samples used for phylogenetic studies and GenBank accession numbers of gene sequences obtained in the study. ‘‘sp. nov.’’ denotes tentative cryptic species identified based on 12S analysis. ‘‘n.a.’’ (not available) indicates missing sequence data.
Species Host Sampling location 12S 16S EF1 H3 RP Cantellius Acropora General Rock, Green Island, Taitung County, KF776142 KF776191 KF776240 na KF776341 acutum lutkeni Taiwan Cantellius Porites Nanwan, Kenting, Pingtung County, Taiwan KF776143 KF776192 KF776241 KF776291 KF776342 arcuatus Cantellius Porites Nanwan, Kenting, Pingtung County, Taiwan KF776144 KF776193 KF776242 KF776292 KF776343 arcuatus Cantellius Porites Nanwan, Kenting, Pingtung County, Taiwan KF776145 KF776194 KF776243 KF776293 KF776344 euspinulosum Cantellius Porites Nanwan, Kenting, Pingtung County, Taiwan KF776146 KF776195 KF776244 KF776294 KF776345 euspinulosum Cantellius Porites Nanwan, Kenting, Pingtung County, Taiwan KF776147 KF776196 KF776245 KF776295 KF776346 euspinulosum Cantellius Porites sp. Nanwan, Kenting, Pingtung County, Taiwan KF776148 KF776197 KF776246 KF776296 KF776347 euspinulosum Cantellius sp. Unknown Nanwan, Kenting, Pingtung County, Taiwan KF776150 KF776199 KF776248 KF776298 KF776349 nov. 1 Cantellius Pocillopora sp. General Rock, Green Island, Taitung County, KF776151 KF776200 KF776249 KF776299 KF776350 pallidus Taiwan Cantellius Pocillopora Neipi Sea Shore, Suao, Yilan County, Taiwan KF776152 KF776201 KF776250 KF776300 KF776351 pallidus damicornis Cantellius Acropora Turtle Tail, Turtle Island, Yilan County, KF776153 KF776202 KF776251 KF776301 KF776352 secundus muricata Taiwan Cantellius Acropora elseyi Neipi Sea Shore, Suao, Yilan County, Taiwan KF776154 KF776203 KF776252 KF776302 KF776353 transversalis Cionophorus Astreopora Nanwan, Kenting, Pingtung County, Taiwan KF776149 KF776198 KF776247 KF776297 KF776348 soongi randalli Darwiniella Cyphastrea Turtle Tail, Turtle Island, Yilan County, KF776155 KF776204 KF776253 KF776303 KF776354 angularis chalcidicum Taiwan Darwiniella Cyphastrea Turtle Tail, Turtle Island, Yilan County, KF776156 KF776205 KF776254 KF776304 KF776355 angularis serailia Taiwan Darwiniella Cyphastrea Nanwan, Kenting, Pingtung County, Taiwan KF776157 KF776206 KF776255 KF776305 KF776356 conjugatum serailia Galkinia Platygyra pini Nanwan, Kenting, Pingtung County, Taiwan KF776158 KF776207 KF776256 KF776306 KF776357 altiapiculus Galkinia Favites Turtle Tail, Turtle Island, Yilan County, KF776159 KF776208 KF776257 KF776307 KF776358 depressa pentagona Taiwan Galkinia equus Favites abdita Neipi Sea Shore, Suao, Yilan County, Taiwan KF776160 KF776209 KF776258 KF776308 KF776359 Galkinia indica Hynophora Nanwan, Kenting, Pingtung County, Taiwan KF776161 KF776210 KF776259 KF776309 KF776360 microconos Hiroa stubbingsi Astreoporasp. Shanfu Fishing Harbor, Siaoliouciou Island, KF776162 KF776211 KF776260 KF776310 KF776361 Pingtung County, Taiwan Neotrevathana Favites abdita Nanwan, Kenting, Pingtung County, Taiwan KF776163 KF776212 KF776261 KF776311 KF776362 elongatum Neotrevathana Favites abdita Shanfu Fishing Harbor, Siaoliouciou Island, KF776164 KF776213 KF776262 KF776312 KF776363 elongatum Pingtung County, Taiwan Nobia grandis Galaxea Nanwan, Kenting, Pingtung County, Taiwan KF776165 KF776214 KF776263 KF776313 KF776364 fascicularis Nobia grandis Galaxea Nanwan, Kenting, Pingtung County, Taiwan KF776166 KF776215 KF776264 KF776314 KF776365 fascicularis Pyrgoma Turbinaria Nanwan, Kenting, Pingtung County, Taiwan KF776167 KF776216 KF776265 KF776315 KF776366 cancellatum frondens Pyrgoma Turbinaria Nanwan, Kenting, Pingtung County, Taiwan KF776168 KF776217 KF776266 KF776316 KF776367 cancellatum frondens Savignium sp. Platygyra General Rock, Green Island, Taiwan KF776169 KF776218 KF776267 KF776317 KF776368 nov. 1 lamillena Savignium sp. Platygyra Nanwan, Kenting, Pingtung County, Taiwan KF776170 KF776219 KF776268 KF776318 KF776369 nov. 2 ryukyuensis Savignium sp. Merulina Shanfu Fishing Harbor, Siaoliouciou Island, KF776171 KF776220 KF776269 KF776319 KF776370 nov. 4 ampliata Pingtung County, Taiwan Savignium sp. Merulina Shanfu Fishing Harbor, Siaoliouciou Island, KF776172 KF776221 KF776270 KF776320 KF776371 nov. 4 ampliata Pingtung County, Taiwan Savignium sp. Goniastrea Nanwan, Kenting, Pingtung County, Taiwan KF776173 KF776222 KF776271 KF776321 KF776372 nov. 5 edwardsi Savignium sp. Goniastrea Shanfu Fishing Harbor, Siaoliouciou Island, KF776174 KF776223 KF776272 KF776322 KF776373 nov. 5 pectinata Taiwan Trevathana sp. Echinopora Nanwan, Kenting, Pingtung County, Taiwan KF776175 KF776224 KF776273 KF776323 KF776374 nov. 1 lamellose Trevathana sp. Echinopora Nanwan, Kenting, Pingtung County, Taiwan KF776176 KF776225 KF776274 KF776324 KF776375 nov. 1 lamellose Trevathana Platygyra Shanfu Fishing Harbor, Siaoliouciou Island, KF776177 KF776226 KF776275 KF776325 KF776376 dentata lamellina Pingtung County, Taiwan Trevathana Platygyra Shanfu Fishing Harbor, Siaoliouciou Island, KF776178 KF776227 KF776276 KF776326 KF776377 dentata lamellina Pingtung County, Taiwan L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22 15
Table 1 (continued)
Species Host Sampling location 12S 16S EF1 H3 RP Trevathana sp. Favites abdita Lobster Cave, Siaoliouciou Island, Pingtung KF776179 KF776228 KF776277 KF776327 KF776378 nov. 3 County, Taiwan Trevathana sp. Favites abdita Lobster Cave, Siaoliouciou Island, Pingtung KF776180 KF776229 KF776278 na KF776379 nov. 3 County, Taiwan Wanella Millipora Lobster Cave, Siaoliouciou Island, Pingtung EU854816 EU854795 KF776286 KF776335 KF776387 milliporae dichotomus County, Taiwan Wanella Millipora Red Sea EU854817 EU854796 KF776287 KF776336 KF776388 milliporae dichotomus
Outgroup Archaeobalanidae Armatobalanus Porties sp. General Rock, Green Island, Taitung County, KF776138 KF776187 KF776236 KF776288 KF776337 allium Taiwan Armatobalanus Pavona cactus Nanwan, Kenting, Pingtung County, Taiwan KF776139 KF776188 KF776237 KF776289 KF776338 allium Armatobalanus Pavona venosa Nanwan, Kenting, Pingtung County, Taiwan KF776140 KF776189 KF776238 KF776290 KF776339 allium Armatobalanus Pavona venosa Nanwan, Kenting, Pingtung County, Taiwan KF776141 KF776190 KF776239 n.a. KF776340 allium Striatobalanus Port Dickson, Malaysia KF776186 KF776235 KF776285 KF776334 KF776386 amaryllis Balanidae Amphibalanus Hong Kong KF776181 KF776230 KF776279 KF776328 KF776380 amphitrite Balanus balanus Fish Market, Athens, Greece KF776182 KF776231 KF776280 KF776329 KF776381 Balanus On oysters imported from Canada to Taiwan KF776183 KF776232 KF776281 KF776330 KF776382 graduala Megabalanus Green Island, Taiwan KF776185 KF776234 KF776283 KF776332 KF776384 ajax Megabalanus Hong Kong NC006293 NC006293 KF776284 KF776333 KF776385 volcano Austrobalanidae Austrominus Southampton, UK KF776184 KF776233 KF776282 KF776331 KF776383 modestus
Table 2 Primer sequences used for PCR amplification, annealing temperature used and their sources. Refer to Table 3 for details of individual gene information.
Primer Direction Sequence (50–30) Annealing temperature (°C) Source 12S FB Forward GTGCCAGCAGCTGCGGTTA 50 Tsang et al. (2009) R2 Reverse CCTACTTTGTTACGACTTATCTC Tsang et al. (2009) 16S Val-F Forward CTGTTTTAGCATTTCATTTACACTG 50–55 Tsang et al. (2009) 16S-CR Forward TTACGGTACCTTTTGTATTAG This study 16S-SR Reverse CCGGTCTGAACTCAAATCGTG Tsang et al. (2009) 1472 Reverse AGATAGAAACCAACCTGG Crandall and Fitzpatrick (1996) EF1 EF1-for Forward GATTTCATCAAGAACATGATCAC 60 This study EF1-rev Reverse AGCGGGGGGAAGTCGGTGAA This study H3 AF Forward ATGGCTCGTACCAAGCAGACVGC 50 Colgan et al. (1998) AR Reverse ATATCCTTRGGCATRATRGTGAC Colgan et al. (1998) RP RP-for1 Forward CACAAGATGAGTATGATGGG 57–60 This study RP-for4 Forward GAYTTTGACGGCGAYGAGATGAA This study RP-rev1 Reverse CGTGCCGTCGTAGTTGACCAT This study RP-rev4 Reverse GAGACCCTCRCGRCCWCCCAT This study
for 10 million generations started from a random tree. Model Alternative a priori phylogenetic hypotheses from previous parameters were estimated during the analysis and chains were morphological analyses were statistically tested using the likeli- sampled every 1000 generations. Convergence of the analyses hood-based approximately unbiased (AU) test (Shimodaira, was validated by the standard deviation of split frequencies reach- 2002). The null hypothesis for all topology testing was that there ing <0.01 and by graphically monitoring the likelihood values over was no difference between trees in the AU test. Alternative tree time using Tracer v1.5 (Rambaut and Drummond, 2007). The trees topologies were constructed using RAxML by setting constraints prior to the achievement of stationary of the log likelihood values on taxa monophyly according to the a priori hypotheses. The per- (5000 trees) were discarded as burn-in. A 50% majority-rule site log likelihood values of individual sites for the trees were esti- consensus tree was constructed from the remaining trees to esti- mated using the same program and then the confidence values of mate posterior probabilities (PP). the tree topology were calculated by CONSEL (Shimodaira and 16 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Hasegawa, 2001) with 1000 BP replicates to access the p values of topologies concerning inter-generic relationships with only minor the testing topology. differences in arrangements among individuals from the same species. Fig. 2 presents the best ML topology with nodal supports obtained from the two analyses. The Pyrgomatidae was revealed 2.3. Ancestral state reconstruction of host usage pattern and to be paraphyletic with Armatobalanus allium nested within (ML morphological characters BP = 91; BI PP = 1.00). The fire coral inhabiting barnacle Wanella milliporae occupied a basal position in the Pyrgomati- We adopted a Bayesian inference approach to reconstruct the dae + Armatobalanus clade but this relationship was only strongly evolutionary history of morphological traits and host transitions supported by ML analysis (BP = 79). The AU test, however, could in the coral barnacles. The Bayesian application was preferred over not reject the alternative hypothesis of a monophyletic Pyrgomat- traditional parsimony or ML frameworks because the latter two idae (p = 0.094). rely on implicit or explicit mapping of the characters of interest The barnacles inhabiting scleractinian corals were divided into onto a single topology, and therefore alternative reconstructions two strongly supported lineages, the first containing Cantellius on the same character (mapping uncertainty) and phylogenetic and Armatobalanus (ML BP = 99, BI PP = 1.00) with the remaining uncertainty including alternative topologies, model parameters genera (Cionophorus, Darwiniella, Hiroa, Galkinia, Neotrevathana, are ignored (Ronquist, 2004; Drummond et al., 2006), leading to Nobia, Pyrgoma, Savignium and Trevathana) constituting the second potentially erroneous results. Mapping and phylogenetic uncer- major clade (ML BP = 100, BI PP = 1.00). Species having fused shell tainties can be reflected through estimation of the posterior prob- walls, four-plated walls, or fused opercular valves were dispersed abilities of different ancestral states on the phylogenies under in our topology and AU tests strongly rejected the monophyly of Bayesian inference (Ronquist, 2004). We employed MrBayes for any of these groupings (all p < 0.001). On the other hand, all of ASR analyses using the same settings as in previous phylogenetic the genera/species with multiple species/individual analyzed were inference except one node of interest was constrained as mono- confirmed to be monophyletic with the exception of paraphyletic phyletic in each MCMC run and the PP of alternative states were Trevathana which was interposed by Neotrevathana. estimated for the constrained node. States of the characters were unordered and equally weighted. The analysis was repeated for 3.2. Ancestral state reconstruction of morphological traits and host all major nodes in our molecular phylogeny. usage The shell structure (six-plate, four-plate or fused) and opercular valve morphology (separate or fused) are two important characters Ancestral state reconstructions revealed multiple independent in traditional phylogenetic inferences of the Pyrgomatidae and circumstances of fusion and separation in both shell plates and hence were included in the ASR analysis. We determined the coral opercular valves (Figs. 3 and 4). The most recent common ancestor host range of barnacle species based on the mitochondrial 12S (MRCA) of Pyrgomatidae + Armatobalanus most likely had a six- analysis (see above and Table 1). Although most of the coral hosts plated shell with separate opercular valves (Fig. 3). The first shell could be identified down to species level, we restricted our analysis fusion event occurred in Wanella when it branched out as an early to infer only the transition in coral host at the family level in ASR. offshoot among the coral inhabiting barnacles. Cantellius reduced This was justified as a more appropriate and conservative approach the number of plates from six in its MRCA with Armatobalanus to due to two reasons. First, our host data were obtained only from four as observed in the extant species. The MRCA of the remaining Taiwanese populations and geographical variation in host coral genera was inferred to exhibit a completely fused shell already species usage may occur. Furthermore, the host specificity is not when it diverged from the Armatobalanus + Cantellius clade merely a function of how many host species a symbiont can (Fig. 3). The fused shell was separated into four plates indepen- exploit, but also how closely related the host species are (Poulin dently in Hiroa and Galkinia. Therefore, at least three shell fusion and Mouillot, 2003). Hence, classifying host according to family events followed by two shell separation events were deduced dur- constitutes a more reliable and conservative approach. Secondly, ing coral barnacle evolution in our molecular phylogeny. On the we found that individual coral barnacle species demonstrated high other hand, the basal coral-inhabiting barnacle taxa (Wanella, dissimilarities in coral host at the genus or species level. There Armatobalanus and Cantellius) retained separated opercular valves would be little phylogenetic information if we considered coral whilst the fused opercular valves first evolved in the MRCA of host at the species level given the large number of coral host spe- the remaining genera (Fig. 4). Subsequently, there were two shifts cies versus the relatively small number of coral barnacle species. from fused to separated opercular valves in Hiroa and the MRCA of Pyrgoma + Savignium + Trevathana + Neotrevathana clade (Fig. 4). 3. Results The opercular valves became fused again in Neotrevathana. The coral hosts colonized by the MRCA of coral-inhabiting bar- 3.1. Sequence characteristics and phylogenetic inference nacles could not be inferred with confidence in our ASR as the posterior probabilities for different candidate families were simi- We sequenced a total 1300 barnacle individuals collected from lar (Fig. 5). The patterns became clearer in the more recent nodes. 553 individual pieces of corals belonging to 58 coral species for the The Armatobalanus and Cantellius clade contained both specialist mitochondrial 12S rRNA gene. A subset of 41 samples from 26 and generalist species and the ancestral state for this lineage pyrgomatid species in 11 genera was chosen for subsequent was not well-resolved. The more basal Armatobalanus can be multi-locus phylogenetic analyses. Other coral-inhabiting barna- found on all coral families whereas Cantellius species comprised cles, including Armatobalanus allium (four individuals) and Mega- of specialists of various coral families (e.g. Acroporidae, Poritidae balanus ajax (one individual) belonging to the Archaeobalanidae and Agariciidae) and species inhabiting two to three coral host and Balanidae, respectively, were analyzed as well. The final con- families (Fig. 5). On the contrary, the MRCA of the remaining gen- catenated dataset of the five molecular markers consisted of era exhibited much higher host specificity with >50% probability 3532 base pairs (see Table 3 for details of individual genes and that would only recruit on single host family, Faviidae. Alterna- Supplementary figures for individual gene trees) with no indels tively, there were 17% and 11% probabilities that it could be observed in the three nuclear protein-coding genes. found on Acroporidae or both Acroporidae and Faviidae (Fig. 5). Analyses under maximum likelihood (ML) and Bayesian infer- In either case, a trend of toward a smaller number of hosts is ence (BI) of the concatenated dataset resulted in highly congruent apparent. This MRCA diverged among two lineages. The first L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22 17
Table 3 Summary of information of individual gene markers.
Gene No. of sites No. of variable sites No. of parsimony informative sites %AT 12S 486 223 186 69.8 16S 931 460 407 77.4 EF1 852 289 235 37.4 H3 261 93 78 37.2 RPII 1002 299 243 41.5
Fig. 2. Maximum likelihood topology inferred from a concatenated dataset (3532 bp from five loci) analysis of phylogenetic relationships among coral barnacle species. Numbers in parentheses denote individuals from different species sequenced (see Table 1 for detailed information). The Bayesian posterior probabilities and maximum likelihood bootstrap support measures are indicated at the corresponding nodes for all values P0.95 for posterior probability and P70 for bootstrap values.
comprised Darwiniella that inhabits Acroporidae and Faviidae, 4. Discussion and Cionophorus + Hiroa that became specialists on Acroporidae. The second lineage consists almost exclusively of species that 4.1. Morphological evolution and taxonomy of Pyrgomatidae lived only on Faviidae, with one or two species that could be occasionally found on other hosts. Furthermore, Nobia and Pyrg- Our molecular phylogenetic analyses and ancestral state recon- oma which were derived from this lineage are specialized for structions revealed that the previously proposed phylogenetic Euphyllidae and Dendrophyllidae, respectively. hypotheses based on morphological characters, primarily shell 18 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Fig. 3. Ancestral state reconstruction of the shell morphology of coral barnacles based on Bayesian inference implemented in MrBayes. The estimated posterior probabilities of alternative ancestral states are indicated by pie charts on the corresponding nodes. The circles next to the species names indicate the traits of the species with multiple individuals from the same species collapsed into a single representative for ease of visualization. structure and opercular valves, do not reflect the evolutionary his- fusions shared a common ancestry, at least two reversals and even tory of the coral-inhabiting barnacles. Our topology is largely con- ‘‘re-fusion’’ were detected as well. Recent molecular phylogenetic gruent with the molecular phylogeny presented by Simon-Blecher analyses have falsified the phylogenetic hypothesis of a gradual et al. (2007) based on 12S, 16S and 18S rRNA sequences but we reduction of shell plate number in multiple barnacle groups (e.g. have achieved improved resolution at most of the internal nodes stalked barnacles, Pérez-Losada et al., 2004, 2008; Chthamalidae, with expanded taxon and gene sampling. These results strongly Fisher et al., 2004; Pérez-Losada et al., 2012). These results indicate suggest that the shell and opercular valve evolution is a highly that the barnacle shell is evolutionarily a rather plastic feature that dynamic process and substantial taxonomic revision is needed in is highly versatile and responsive to environmental selection pres- coral-symbiotic barnacles. sure. Accordingly, phylogenetic inferences based on this character Fusion of shell and opercular valves is a feature unique to the may be misleading or erroneous. Pyrgomatidae and is believed to be an important adaptation to Wanella is the only pyrgomatid living in association with the withstand the pressure from growth of the coral skeleton (Ross fire coral Millipora (Hydrozoa). It was once believed to be affiliated and Newman, 1973; Anderson, 1992). Three lineages (Wanella, with the Savignium group due to morphological similarities Cantellius, and the remaining pyrgomatids) were identified as (Anderson, 1992). The molecular phylogeny of Simon-Blecher evolving fused shells independently, such that the fused shell is a et al. (2007) placed Wanella among the other free-living balanoids result of parallel evolution in adapting to a symbiotic lifestyle. but our statistical tests could not reject the alternative hypothesis More importantly, the fused shell did not represent an evolution- of classifying Wanella in Pyrgomatidae (with Armatobalanus trea- ary dead-end and two reversals were observed in Hiroa and Galki- ted as a pyrgomatid). Likewise, we also demonstrated here that nia respectively. Similarly, while most of observed opercular valve Wanella is highly divergent from the other pyrgomatids, as well L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22 19
Fig. 4. Ancestral state reconstruction of the opercular valve morphology of coral barnacles based on Bayesian inference implemented in MrBayes. The estimated posterior probabilities of alternative ancestral states are indicated by pie charts on the corresponding nodes. The circles next to the species names indicate the traits of the species with multiple individuals from the same species collapsed into a single representative for ease of visualization. as the other free-living balanoids analyzed. Given its high genetic lineage in the coral-inhabiting barnacles. It exhibits a generalist divergence from the Pyrgomatidae (including Armatobalanus) and lifestyle, living on any corals forming symbiotic interactions with its very different host specificity (fire coral), a separate family barnacles. The preservation of its plesiomorphic features may have placement for Wanella milleporae may be justified. enabled it to survive with a broad range of hosts instead of evolv- On the other hand, molecular phylogenetic inferences have con- ing specific adaptations for a particular host. Moreover, investiga- sistently placed Armatobalanus within the Pyrgomatidae, as the sis- tions on sperm ultrastructure have shown that Armatobalanus ter group of Cantellius (Simon-Blecher et al., 2007). However, sperm bears features distinct from those of other archaeobalanids, Armatobalanus retains many ancestral balanid features, including yet similar to those of pyrgomatids (Healy and Anderson, 1990). six-plated shells without any aperture frills (a glandular fold that Armatobalanus species associated with corals (e.g. A. allium and A. inhibits the growth of corals, see introduction) and a conical shell arcuatus) have some morphological adaptations that allow them that is elevated above the coral surface. This contrasts with pyrg- to live in coral tissues, including the development of a cup-shaped omatids in general which are highly derived and usually deeply base, a feature absent in non-coral inhabiting Armatobalanus or embedded in their scleractinian coral hosts with an aperture frill other archeobalanid species. Although Armatobalanus does not to reduce coral growth over their orifices. Various authors have have any aperture frill to chemically prevent coral overgrowth at therefore, considered Armatobalanus as the stem lineage of extant the orifice, it uses its beak or tergum to mechanically abrade the Pyrgomatidae (Ross and Newman, 1973; Newman and Ross, rim of the orifice to remove overgrowing coral tissue (Anderson, 1976; Anderson, 1992). In our analysis Armatobalanus does not 1992). The tergum of A. allium and A. arcuatus are modified with constitute the ancestral stock but represents a relatively basal a strong, enlarged beak strengthened by a chitinous ridge 20 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Fig. 5. Ancestral state reconstruction of the host usage pattern of coral barnacles based on Bayesian inference implemented in MrBayes. The estimated posterior probabilities of alternative ancestral states are indicated by pie charts on the corresponding nodes. The matrix next to the species names denotes the presence (colored) or absence (empty) of the barnacle species inhabiting the corresponding coral family according to data from the present survey. The numbers of individuals sampled for each barnacle species are indicated at the right hand side. The proportion of individual found on different coral host families is shown by the pie chart at the bottom for the barnacle species that found on more than one coral host family. Multiple individuals from the same species are collapsed into a single representative for ease of visualization. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
(see Anderson, 1992) and spur, in contrast to other archaeobalanid from Savignium. Trevathana differs from Savignium in having an species. However, the fourth cirri of Armatobalanus differ from abbreviated tergal spur (absent from Savignium), small spur tooth those of typical pyrgomatids in having sharp serrated segments (large spur tooth in Savignium), absence of adductor plates (present (hooked segments). The function of these segments is unknown, in Savignium), and a shell sheath not extending to the basal margin but they may be used for removing overgrowing coral tissues on of the shell externally. The division of Savignium and Trevathana the orifice. Based on these multiple lines of evidence, we conclude based on these characters is supported by our molecular phylog- that it is reasonable to remove Armatobalanus from Archaeobalan- eny. Neotrevathana was erected by Ross (1999) and is distinguished idae and treat it as a member of the Pyrgomatidae. from Trevathana in having broad low ridges on the shell surface, Savignium, Trevathana and Neotrevathana are morphologically fused opercular plates and a broad occludent ledge. The tergal spur similar and they form a clade in the present study. Anderson in Neotrevathana is knob like, whilst the tergal tooth in Trevathana (1992) erected Trevathana to accommodate some species differing is an elaboration of the tergal spur. However, Neotrevathana is L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22 21 placed within Trevathana in our topology so its generic status is not a grant from the National Science Council, Taiwan (NSC103- supported. 2621-B-019-004-MY2) to LMT, (NSC-102-2621-B-001-002) to B.K.K.C., and a research grant from the Research Grants Council, 4.2. Evolution and diversification through host shift HKSAR Government (CUHK463509) to KHC.
In contrast to morphology-based phylogenies, we observed that Appendix A. Supplementary material there is a close linkage between coral host usage and phylogenetic relationships. Sister taxa are usually found on the same coral host Supplementary data associated with this article can be found, in family, for instance, Cantellius species living on Acroporidae, Dar- the online version, at http://dx.doi.org/10.1016/j.ympev.2014. winiella/Cionophorus/Hiroa living on Acroporidae, and Savignium/ 03.002. Trevathana/Neotrevathana living on Faviidae. Ecological interac- tions, such as parasitism and mutualism, are evolutionarily con- served across the entire tree of life, with phylogenetically close References species interacting with similar partners (Gómez et al., 2010). Sym- biotic organisms must necessarily evolve morphological and phys- Achituv, Y., Mizrahi, L., 1996. Recycling of Ammonium within a hydrocoral (Millepora dihotoma) – zooxanthellae-cirripede (Saviginium milleporum) iological adaptations to host defense systems. Hence, there are symbiotic association. Bull. Mar. Sci. 58, 856–860. constraints in host switching and invasion, as switching to phylo- Achituv, Y., Brickner, I., Erez, J., 1997. Stable carbon isotope ratios in Red Sea genetically distant hosts likely requires entirely new sets of anti- barnacles (Cirripedia) as an indicator of their food source. Mar. Biol. 130, 243– 247. defense adaptations. The conservation is more pronounced in par- Anderson, D.T., 1992. Structure, function and phylogeny of coral-inhabiting asitic organisms (e.g. viruses) which are under strong pressure to barnacles (Cirripedia: Balanoidea). Zool. J. Linn. Soc. 106, 277–339. overcome host defenses (Gómez et al., 2010). Likewise, coral bar- Barnes, H., Read, R., Topinka, J.A., 1970. The behavior on impaction by solids of some common Cirripedes and relation to their normal habitat. J. Exp. Mar. Biol. Ecol. nacles with shared ancestry should have higher success in coloniz- 5, 70–87. ing closely related coral hosts with similar skeletal structure, Brickner, I., Simon-Blecher, N., Achituv, Y., 2010. Darwin’s Pyrgoma (Cirripedia) growth pattern and defense mechanisms, resulting in correlations revisited: revision of the Savignium group, molecular analysis and description of such as those observed in the present study between barnacle phy- new species. J. Crustac. Biol. 30, 266–291. Chan, B.K.K., Chen, Y.-Y., Lin, H.-C., 2013a. Biodiversity and host-specificity of coral logeny and coral host family usage. barnacles of Galkinia (Cirripedia: Pyrgomatidae) in Taiwan, with descriptions of The coral-inhabiting barnacles further display an evolutionary six new species. J. Crustac. Biol. 33, 392–431. trend towards narrow host range. Armatobalanus and Cantellius at Chan, B.K.K., Chen, Y.-Y., Achituv, Y., 2013b. Crustacean Fauna of Taiwan, Barnacles II: Family Pyrgomatidae. Biodiversity Research Center, Academia Sinica. the basal position of the tree contain many more generalists than Colgan, D.J., McLauchlan, A., Wilson, G.D.F., Livingston, S.P., Edgecombe, G.D., the more derived lineages. This is consistent with the view that Macaranas, J., Cassis, G., Gray, M.R., 1998. Histone H3 and U2 snRNA DNA the ancestral stock of coral barnacles could inhabit a greater num- sequences and arthropod molecular evolution. Aust. J. Zool. 46, 419–437. Crandall, K.A., Fitzpatrick Jr., J.F., 1996. Crayfish molecular systematics: using a ber of host corals and that host ranges narrowed as more derived combination of procedures to estimate phylogeny. Syst. Biol. 45, 1–26. morphological adaptations evolved (Ogawa and Matsuzaki, Drummond, A.J., Ho, S.Y.W., Phillips, M.J., Rambaut, A., 2006. Relaxed phylogenetics 1992). This evolutionary trajectory towards higher host specificity and dating with confidence. PLoS Biol. 4, 699–710. Duffy, J.E., 1996. Species boundaries, specialization, and the radiation of sponge- is commonly observed in other symbiotic animals (Nosil, 2002). dwelling alpheid shrimp. Biol. J. Linn. Soc. 58, 307–324. Although con-generic barnacle species usually share the same Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and coral host family, they are often associated with different coral high throughput. Nucl. Acids Res. 32, 1792–1797. Faucci, A., Toonen, R.J., Hadfield, M.G., 2007. Host shift and speciation in a coral- genera or species, with little overlap in their host range (e.g. Galki- feeding nudibranch. Proc. R. Soc. Lond. B: Biol. Sci. 274, 111–119. nia, Savignium and Trevathana). This implies subsequent host-asso- Fisher, T., Katcoff, D.J., Achituv, Y., 2004. A phylogenetic study of chthamaloids ciated adaptation and speciation are a common process and (Cirripedia; Thoracica; Chthamaloidae) based on 16S rDNA and COI sequence competition for the same host is intensive as has been reported analysis. Biol. J. Linn. Soc. 83, 39–45. Gómez, J.M., Verdú, M., Perfectti, F., 2010. Ecological interactions are evolutionarily in other symbiotic marine fauna (e.g. Mokady et al., 1999; conserved across the entire tree of life. Nature 465, 918–922. Munday et al., 2004; Macdonald et al., 2006; Faucci et al., 2007). Healy, J.M., Anderson, D.T., 1990. Sperm ultrastructure in the Cirripedia and its phylogenetic significance. Rec. Aust. Mus. 42, 1–26. Lanfear, R., Calcott, B., Ho, S.Y.W., Guindon, S., 2012. Partition finder: combined 5. Conclusion selection of partitioning schemes and substitution models for phylogenetic analyses. Mol. Biol. Evol. 29, 1695–1701. Liu, R.-Y., Ren, X.Q., 2007. Fauna Sinica. Invertebrata. Crustacea: Cirripedia Our molecular phylogeny suggests morphological characters Thoracica, vol. 42. Science Press, Beijing, China (in Chinese). including fused shell and fused opercular valves unique to the Macdonald III, K.S., Rios, R., Duffy, E., 2006. Biodiversity, host specificity, and dominance by eusocial species among sponge-dwelling alpheid shrimp on the coral-inhabiting barnacles Pyrgomatidae are a result of conver- Belize Barrier Reef. Divers. Distrib. 12, 165–178. gence and do not reflect the evolutionary history of the group. Mokady, O., Loya, Y., Achituv, Y., Geffen, E., Graur, D., Rozenblatt, S., Brickner, I., The development and evolution of these characters are further 1999. Speciation versus phenotypic plasticity in coral inhabiting barnacles: Darwin’s observation in a phylogenetic context. J. Mol. Evol. 49, 367–375. revealed to be highly dynamic and that reversed separation of Munday, P.L., van Herwerden, L., Dudgeon, C.L., 2004. Evidence for sympatric fused shell/plates has occurred more than once. On the other hand, speciation by host shift in the sea. Curr. Biol. 14, 1498–1504. the host-symbiont interactions appear to be more phylogenetically Newman, W.A., Ross, A., 1976. Revision of the balanomorph barnacles: including a conserved. The symbiotic lifestyle has evolved in the direction of catalog of the species. Mem. San Diego Soc. Nat. Hist. 9, 1–108. Nosil, P., 2002. Transition rates between specialization and generalization in narrower host ranges with more specific adaptation. Host associ- phytophagous insects. Evolution 56, 1701–1706. ated speciation further shaped species diversity and distribution Ogawa, K., Matsuzaki, K., 1992. An essay on host-specificity, systematic taxonomy patterns in these symbiotic fauna thereby reducing competition. and evolution of coral barnacles. Bull. Biogeogr. Soc. Jpn. 47, 87–101. Pérez-Losada, M., Høeg, J.T., Crandall, K.A., 2004. Unraveling the evolutionary radiation of the thoracican barnacles using molecular and morphological Acknowledgments evidence: a comparison of several divergence time estimation approaches. Syst. Biol. 53, 244–264. Pérez-Losada, M., Harp, M., Hoeg, J.T., Achituv, Y., Jones, D., Watanabe, H., Crandall, The authors would like to thank Pei-Chen Tsai, I-Han Chen, Hsi- K.A., 2008. The tempo and mode of barnacle evolution. Mol. Phyl. Evol. 46, 328– Nien Chen, Anteia Feng for (Academia Sinica) supporting the field 346. Pérez-Losada, M., Høeg, J.T., Crandall, K.A., Achituv, Y., 2012. Molecular phylogeny sampling of corals. This work was supported from a Thematic and character evolution of the chthamaloid barnacles (Cirripedia: Thoracica). Grant of Academia Sinica, Taiwan (AS-100-TP-A02-3) to B.K.K.C., Mol. Phyl. Evol. 65, 329–334. 22 L.M. Tsang et al. / Molecular Phylogenetics and Evolution 77 (2014) 11–22
Poulin, R., Mouillot, D., 2003. Parasite specialization from a phylogenetic Shimodaira, H., 2002. An approximately unbiased test of phylogenetic tree perspective: a new index of host specificity. Parasitology 126, 473–480. selection. Syst. Biol. 51, 492–508. Rambaut, A., Drummond, A.J., 2007. Tracer v1.4.