Two new species of the gorgonian inhabiting barnacle, Conopea (Crustacea, Cirripedia, Thoracica), from the Gulf of Guinea

Abstract Two new species of Conopea (Say 1822) are described from the Gulf of Guinea: Conopea saotomensis sp. n.and Conopea fidelis sp. n. These two new species were collected from the historically isolated volcanic islands of São Tomé and Príncipe. The relationship between Conopea saotomensis sp. n., Conopea fidelis sp. n.and two other Atlantic barnacle species, Conopea calceola (Ellis 1758) and Conopea galeata (Linnaeus 1771), is examined. The methods employed are the construction of a molecular phylogeny using mitochondrial COI and nuclear H3 gene sequence data along with morphological comparisons of calcareous and cuticular body parts. It is found that Conopea saotomensis sp. n., Conopea fidelis sp. n.and Conopea calceola are most closely related to each other but the relationship among them is unresolved. Gorgonian hosts are identified. Preliminary observations show species level host specificity for Conopea fidelis sp. n.


introduction
The Gulf of Guinea island chain consists of Bioko, São Tomé, Príncipe, and Annobón. This study focuses on São Tomé and Príncipe, which are approximately 140 km apart and 274 km west of northern Gabon. They are the products of large shield volcanoes originating 3,000 m below the ocean's surface along the Cameroon line. São Tomé and Príncipe are old islands, 13 and 30 myo, respectively, and have never been connected to the African mainland.

Genus Conopea
Say (1822) designated a new genus Conopea to accommodate a new species, C. elongata, he described from eastern Florida. He included a previously known barnacle Balanus galeatus (=Lepas galeata Linnaeus 1771) in Conopea. Conopea elongata is later listed, by Darwin (1854), as a junior synonym of C. galeata. Say describes Conopea as 'Shell sessile, fixed, composed of two cones joined by their bases, the lines of junction carinate each side: inferior cone entire, attached by its anterior side and tip to marine bodies; with an aperture at the summit, closed by a quadrivalved operculum. ' Conopea is a widespread genus that is found in temperate and tropical oceans around the world. Currently, there are 21 described species of Conopea. In general, Conopea is not a well documented group. There is very little data on host associations, species ranges are not well defined, published descriptions are often incomplete and occasionally contain questionable information. Darwin (1854) thought Conopea to be closely related to the genera Megabalanus and Acasta whereas Hoek (1913) thought Conopea to be closely related to Balanus. All species of Conopea live in an obligate commensal symbiotic relationship with either a gorgonian or an antipatharian. The barnacle lives almost completely covered by host tissue, the basis of its shell clasps the axis of the host, with only the opercular opening exposed.

Atlantic species of Conopea
There are three known species of Conopea found in the Atlantic Ocean and Caribbean Sea: C. calceola, C. galeata, and C. merrilli. Conopea calceola was originally described from the Strait of Gibraltar, by Ellis (1758). Conopea calceola has subsequently been recorded from the Mediterranean to South Africa, the Persian Gulf to western Australia, the Indian Ocean, and Japan (see Newman and Ross 1976 for literature summary). Conopea merrilli was described from South Carolina (Zullo 1966) and has since been recorded from the west coast of Florida and Puerto Rico (see Newman and Ross 1976). Conopea galeata was described by Linnaeus in 1771 but no type locality was given and the type specimen is lost (Pilsbry 1916). Darwin (1854) gave a description of C. galeata (as Balanus galeatus) and listed its known localities as South Carolina, Florida, West Indies, and Central America. Pilsbry (1916) also gave a description of C. galeata (as Balanus galeatus) and listed the distribution as South Carolina to the West Indies and Central America, and southern California. The current distribution range of Conopea galeata is North Carolina through the West Indies, the Gulf of Mexico to Venezuela, southern California to Panama, and the Galápagos Islands (see Newman and Ross 1976).
Morphologically C. merrilli and C. galeata are clearly different and easily distinguishable from C. calceola, C. saotomensis sp. n., and C. fidelis sp. n.. Conopea calceola is morphologically similar to the Gulf of Guinea species and is therefore compared in detail to aid in future identifications. Conopea galeata was chosen over C. merrilli as an outgroup for molecular analysis because of its larger distribution range and greater availability of specimens.
Conopea galeata from St. Catherine Is., Georgia (USA) were borrowed from the California Academy of Sciences Invertebrate Zoology Department (CASIZ). Conopea galeata from South Padre Is., Texas and Mexico Beach, Florida were collected by Mary Wicksten. Conopea galeata from Port Aransas, Texas were collected by Carol Cox.
Barnacle cirri, mouthparts and opercular plates from São Tomé, Príncipe, and Portugal specimens were dissected for morphological comparisons. These physical traits, along with shell shape, in particular basis shape and presence/absence of longitudinal tubes in shell wall plates, are traditionally used for identification. The cirri and mouthparts were mounted on microslides and photographed at 100x with a Leitz microscope imaging system. Images of the opercular plates were taken with a scanning electron microscope (SEM, LEO/Zeiss 1450VP).
Identification of host gorgonians was based on external and sclerite morphology. Branching patterns, polyp shape, color and sclerite types were examined. Sclerites were isolated by dissolving small amounts of gorgonian tissue in sodium hypochlorite solution, followed by rinsing with water and then 75% ethanol. Images of the sclerites were taken with SEM and Leitz optical microscope imaging systems. All gorgonians harboring barnacles were identified using Grasshoff (1988Grasshoff ( , 1992. However some bar- nacle specimens lacked host tissue and were found attached to only the gorgonian axis. Therefore, identification of those hosts was impossible. Genomic DNA was extracted from adductor muscle tissue using the Qiagen DNeasy Blood and Tissue kit (Valencia, CA). The cytochrome c oxidase subunit I (COI) primers COI-N: TGAGAAATTATTCCGAAGGCTGG (Van Syoc 1994, 1995 and LCO 1490: GGTCAACAAATCATAAAGATATTGG (Folmer et al. 1994) were used to amplify approximately 700 base pairs of the mitochondrial genome (mtDNA). Additionally, the Histone 3 primers H3F: ATGGCTCGTACCAAGCA-GAC VGC and H3R: ATATCCTTRGGCATRATRGTGAC (Colgan et al. 1998) were used to amplify approximately 350 base pairs of the nuclear protein coding gene (nDNA). The COI thermal profile was an initial step of 94°C for 3 min, then 35 cycles of: 94°C for 30sec, 47°C for 30sec, and 72°C for 1min. H3 thermal profile was 3 initial steps of 94°C for 3 min, 50°C for 2 min, 72°C for 2 min, then 35 cycles of: 94°C for 35 sec, 50°C for 30 sec, and 72°C for 40 sec. The resulting sequence data were edited in Sequencher 4.7 (Gene Codes) and BioEdit 7.0.9 (Hall 1997). Alignments were initially performed with ClustalW 1.8 and then edited by hand.
Molecular phylogeny was determined by Bayesian and likelihood analyses. Semibalanus balanoides (GenBank accession AF242660.1), another archaeobalanid, was used as an outgroup. Bayesian analyses were run in Mr. Bayes (Huelsenbeck and Ronquist 2001) for 50 million generations with a sample frequency of 1000 using the CAS CCG PhyloCluster (a 280-core Apple Xserve High Performance Computing Cluster with 8-12 GB RAM/node (232 GB total)). The concatenated dataset was partitioned into 1 st , 2 nd , and 3 rd codon positions so that models of substitution could be estimated for each site with Mr.Modeltest 2.3 (Nylander 2004). Burn-in and convergence values were determined using Tracer v1.5 (Drummond and Rambaut 2007). Likelihood analyses were run in PAUP* 4.0b10 (Swofford 2003). Heuristic searches were performed along with bootstrap analyses; 10,000 bootstrap replicates with 10 random sequence additions to each bootstrap. The best-fit DNA substitution models were determined with Mr.Modeltest. All analyses were performed on the gene datasets separately as well as concatenated. Uncorrected nucleotide pairwise-distance matrices among and within groups were determined in MEGA 5.05 (Tamura et al. 2007).

Molecular analyses
Two major clades resulted from molecular analysis. One clade contains C. calceola, C. saotomensis sp. n. and C. fidelis sp. n, and the other contains C. galeata. Unfortunately, the gene data used was not sufficient to completely resolve the relationship among the three eastern Atlantic species. We do know that they are each other's closest relatives but we do not know which two of the three are most closely related. Bayesian (Fig. 1) and likelihood phylogenies, based on concatenated COI and H3 datasets, between C. saotomensis sp. n., C. fidelis sp. n., and C. calceola are unresolved. Bayesian phylogeny generated solely on COI data shows the two Gulf of Guinea species as being more closely related to each other than to C. calceola. However the likelihood phylogeny generated with solely COI data again showed an unresolved relationship among C. saotomensis sp. n., C. fidelis sp. n., and C. calceola. Separate Bayesian and likelihood analyses of H3 sequence data showed similar unresolved topologies among the three eastern Atlantic species.
Pairwise uncorrected p-distances (Table 2) of COI and H3 also could not resolve the relationship. Distances for COI indicate that C. saotomensis sp. n. and Conopea fidelis sp. n. are more closely related to each other (8.2%) than to C. calceola (8.8% and 10.4%, respectively) whereas H3 distances indicate that C. saotomensis sp. n. and Conopea fidelis sp. n. are more closely related to C. calceola (1.4% and 1.3%, respectively) than to each other (2.2%). Pairwise uncorrected p-distances within groups is as follows: C. saotomensis sp. n. = 0.8%/0.0%, C. fidelis sp. n. = 0.7%/0.2%, C. calceola = 0.7%/0.0%, C. galeata = 0.3%/0.1% (COI/H3 respectively).  (7 specimens Description. Exterior of shell with minute bumps, most prominent on parieties. Color variable, white with varying shades of purple concentrated on parietes and basis often at carina side of shell. Radii usually white but can be colored, basis lighter shade of purple to light purplish-red ( Fig. 2A-B). Opercular opening round to diamond shaped, small in comparison to shell. Mantle tissue purple near opercular opening. Basis boat shaped ( Fig. 2A-B) highly variable depth and length. Basis length of the paratypes 9-21mm. Basis elongated in rostro-carina axis, often deeply indented and/ or warped from growing around axis of gorgonian. Carina convex. Rostrum often elongate. Basis and parieties with longitudinal tubes, alae and radii solid. Tubes of basis hollow near wall plate suture where outgrowths from wall plates articulate, otherwise secondarily filled. Wall plates with small, hollow tubes close to external plate surface. Sutural margins denticulated. Shell strong, not disarticulating in sodium hypochlorite solution.
Scutum (Fig. 3A-D) with fairly straight tergal and occludent margins, occludent margin may be concave. Basal margin curved. Apex acute. Articular ridge about ⅔ length of tergal margin. Articular furrow present. Adductor ridge absent. Depressor muscle pit deep, medium to large in diameter. Adductor muscle pit shallow. Interior surface of articular ridge and above adductor muscle pit rough with small flat ridges, remainder of interior surface smooth. Interior and exterior of tergum white with varying degrees of purple coloration, most often dark purple, concentrated at apex. Tergum ( Fig. 3E-H) with concave scutal and convex carinal margins, basal margin slightly convex or straight. Apex acute. Basiscutal angle shallow upper corner recessed. Spur smooth, broad, corners rounded approximately ½ to ⅓ width of tergum. Spur margin bearing 3-5 small teeth. Length of spur teeth variable. Spur furrow open. Articular ridge low ⅓ to ½ length of scutal margin. Articular furrow shallow. Depressor muscle crests faint. Interior surface rough with multiple low longitudinal ridges. Coloration matches that of scutum. Labrum (Fig. 4A) with deep medial notch, 0-3 teeth on both or one side of notch. Mandibular palp (Fig. 4B) elongate; superior margin convex, partially covered with long setae; apex with long setae; inferior margin with many shorter setae (Fig. 3).
Maxilla I (Fig. 4D-E) with 7-10 large thick spines, either evenly distributed or concentrated on ⅔ of the cutting margin near superior margin, remaining cutting margin covered in short fine setae. Many fine short setae along inferior margin near cutting margin and a few fine setae along superior margin near cutting margin.
Etymology. Conopea saotomensis sp. n. is named after the island from which it was first collected, São Tomé.
Distribution. Conopea saotomensis sp. n. is known from São Tomé and Príncipe at depths ranging from 5-34 m living on species of Leptogorgia and Eunicella.
Remarks. Conopea saotomensis sp. n. differs from Conopea calceola by the following: distance between scutal depressor muscle pit and articular furrow is wider in C. saotomensis sp. n. than in C. calceola; angle between tergal spur and basal margin is smaller in C. saotomensis sp. n. than in C. calceola; in C. saotomensis sp. n. large spines on cutting edge of maxilla span ⅔ or entire length, span entire length in C. calceola. Description. Exterior of shell covered in very small bumps; color variable, white with pink or light purple on parietes and basis, radii usually white or lighter in color, rostrum often white (Fig. 6A-B). Opercular opening round to diamond shaped, small compared to shell size. Mantle tissue purple near opercular opening. Basis boat shaped (Fig. 6A-B), highly variable depth and length. Basis length of paratypes 14-32 mm. Basis elongated in rostro-carina axis, often deeply indented and/or warped from growing around axis of gorgonian. Carina convex. Rostrum often elongate. Basis with radiating longitudinal tubes, secondarily filled, hollow near wall plate suture. Wall plates with small longitudinal tubes near external surface of shell. Alae and radii lacking tubes. Sutural margins denticulated. Shell wall compartments strongly fused, not disarticulating in sodium hypochlorite solution.

Conopea fidelis
Scutum (Fig. 7A-D) with straight to mildly convex tergal margin, occludent margin usually straight , occasionally with curve above basioccludent angle. Basal margin  variable, sinuous. Apex subacute. Articular ridge prominent, extending ⅔-¾ length of scutum. Articular furrow present. Adductor ridge absent. Adductor muscle pit shallow, fairly large. Depressor muscle pit large, deep, broad, may converge with basil margin. Majority of interior surface of scutum smooth, articular ridge and apex with low, flat ridges. Interior and exterior white with varying shades of purple coloration concentrated at apex. Tergum (Fig. 7E-G) scutal and carinal margins curved. Basal margin straight or slightly curved. Apex acute. Basicutal angle shallow, upper corner recessed. Spur broad, bears no teeth, about ½ to ⅓ width of tergum, spur furrow open. Articular ridge ⅓ to ½ length of tergum. Articular furrow shallow. Depressor muscle crests faint. Interior rough with multiple small ridges. Coloration matches that of scutum. Labrum (Fig. 8A) with deep notch, 0-3 teeth on both or one side of notch. Mandibular palp (Fig. 8B) slightly convex oval shape, superior margin with curved ridge and sparse fine long setae, inferior margin with dense shorter setae.
Cirrus IV (Fig. 9D) rami long, tapered, anterior side with long dense setae and small spines at base of setae (Fig. 9E) extending from first to twentieth segment (end segment variable), posterior side with short sparse setae at segment divisions.
Cirrus V (Fig. 9F) rami long with long dense setae on anterior side and short sparse setae at segment divisions on posterior side, about equal in length.
Cirrus VI (Fig. 9G) rami long with long dense setae on anterior side, short sparse setae only at segment divisions of posterior side, similar length.
Etymology. Conopea fidelis sp. n. is named so because it is found to be faithful to one host species of gorgonian, Muriceopsis tuberculata. From the Latin fidelis: faithful or true.
Distribution. Conopea fidelis sp. n. is known from São Tomé and Príncipe at depths ranging from 5-34 m and is found living on the gorgonian Muriceopsis tuberculata.
Remarks. Morphological differences between Conopea fidelis sp. n. and Conopea calceola are as follows: C. fidelis sp. n. does not have tergal spur teeth, C. calceola does; scutal depressor muscle pit may converge with basal margin in C. fidelis sp. n., it does not in C. calceola; C. fidelis sp. n. maximum basis length is longer than that of C. calceola.
Morphological differences between Conopea saotomensis sp. n. and Conopea fidelis sp. n. are as follows: C. saotomensis sp. n. shell color ranges from dark purple to light purplish-red, C. fidelis sp. n. shell color ranges from light purple to pink; C. fidelis sp. n . basis length maximum is longer than that of C. saotomensis sp. n.; length of scutal articular furrow in C. saotomensis sp. n. is shorter than C. fidelis sp. n.; scutal depressor pit may converge with basal margin in C. fidelis sp. n. but not in C. saotomensis sp. n.; angle between tergal spur and basal margin is smaller in C. saotomensis sp. n. than C. fidelis sp. n.; tergal spur teeth present in C. saotomensis sp. n., absent in C. fidelis sp. n.; C. saotomensis sp. n. length of tergal articular ridge is equal or longer to that of C. fidelis sp. n.; cutting edge spines of maxilla I span entire margin or just ¾ in C. saotomensis sp. n, span entire margin in C. fidelis sp. n.; C. fidelis sp. n. maxilla I may have a notch, C. saotomensis sp. n. does not.

Discussion
COI has been shown to be useful for delimiting species within the Crustacea (Lefébure et al. 2006) and, in particular, within the Cirripedia (Van Syoc 1995, Wares 2001, Rawson et al. 2003. Costa et al. (2007) found within genus COI divergence levels of crustaceans to range from 4.92% to 31.39%. Van Syoc (1994) found COI divergence levels averaging 1.2% among distantly separated sub-populations of Pollicipes elegans (Lesson 1831) (Crustacea: Cirripedia: Scalpelliformes). Van Syoc (1995) and Van Syoc et al. (2010) also found a range of 15%-28% among species of P. elegans, P. polymerus (Sowerby 1833), and P.pollicipes (Gmelin 1790). Regarding barnacles, histone genes have been shown to be highly conserved and can be used for deep metazoan phylogenies (Pérez-Losada et al. 2004, 2008, Van Syoc et al. 2010) and for phylogenetic analysis of arthropods (Colgan et al. 1998) and thoracic barnacles (Pérez-Losada et al. 2004). Expected divergence levels of H3 among closely related cirriped species is not known but Van Syoc et al. (2010) found low levels, 0-1.3%, of sequence divergence between species of Pollicipes. The overall difference in divergence between genes, higher for mitochondrial COI and lower for nuclear H3, is expected as nuclear genes typically evolve slower than mitochondrial genes for arthropods (Avise et al. 1994, Burton andLee 1994). COI and H3 divergence levels found for C. saotomensis and C. fidelis are both satisfactory for determining a species within Cirripedia.
The barnacles collected from the Gulf of Guinea for this study were originally identified as C. cf. calceola. The initial identifications were tentative because C. calceola is not well studied, has a reportedly large distribution, the original species description (Ellis 1758) contains sparse morphological data, and Darwin's description of C. calceola (used as our reference for morphology) relied on locality for identification. Ellis (1758) designated the type locality of C. calceola as the Strait of Gibraltar, which connects the eastern Atlantic Ocean to the Mediterranean Sea. Darwin (1854) recorded the presence of C. calceola off the west coast of Africa. For his identification he noted that the original description of C. calceola, by Ellis, does not adequately distinguish any morphological characters for positive identification so he relied on locality, the eastern Atlantic. The C. calceola specimens from Portugal used in our phylogenetic and morphological analyses match the morphology of the specimens from the coast of Africa that Darwin described as C. calceola.
Attempts to obtain specimens of C. calceola from other locations/institutions were unsuccessful. Darwin's (1854) description of C. calceola was used as the guideline for the species along with the literature of Hoek (1913), (Hiro 1937), andRen andLiu (1978). Unfortunately, none of these papers state the number of tergal spur teeth and the images are too poor to count them accurately. Therefore, a comparison of number of tergal spur teeth could not be made. But there is a difference between the 7 Portuguese C. calceola specimens of this paper and C. saotomensis sp. n. Conopea calceola was found to have 6-9 tergal spur teeth and C. saotomensis sp. n. 3-5.

Gorgonian host preference
Barnacles are found permanently attached to many different types of living and nonliving substrata. Locating a living substratum, especially one that is mobile or spatially rare, can be challenging for a small marine larva. For example; a gorgonian, a turtle, or a whale is harder to locate than a rock bed. When barnacle larvae locate and settle onto a gorgonian they may be recognizing the substratum, the presence of conspecifics, or both. It has been shown that barnacle larvae can determine where to settle by recognizing pheromone cues from their cohorts (Crisp and Meadows 1962, Knight-Jones 1995, Dreanno et al. 2006 or chemical cues from their host (Pasternak et al. 2004, Nogata andMatsumura 2005). It has also been shown that gorgonians produce barnacle settlement inducers as well as inhibitors (Standing et al. 1983) and prostaglandins that promote hatching (Clare et al. 1985). The inhibitors are water soluble and so found in the water near the gorgonian whereas the inducers are found absorbed in the gorgonian tissue.
Although the details of the settling barnacle larvae and gorgonian interaction are not completely known, it appears, from our observations (specifically that Conopea fidelis sp. n. was found only on Muriceopsis tuberculata) that barnacle larvae may be capable of distinguishing between gorgonian species. Of course, more collections, identifications, and laboratory work testing settlement preference would be needed to answer this question.

Endemism
The possibility that Conopea saotomensis sp. n. and Conopea fidelis sp. n. are endemic to the Gulf of Guinea Islands is likely for the following reasons: the islands' distance (approx. 274 km), age (approx. 13 and 30 myo), and historic isolation from mainland Africa; they are not known from any previous locality; many endemic species, terrestrial and marine, are found on the Gulf of Guinea islands (Jones 1994, Measey et al. 2007). However, further sampling from the west coast of Africa is essential to determine if they are indeed endemic.