Molecular Ecology Resources (2011) 11, 271–278 doi: 10.1111/j.1755-0998.2010.02946.x

DNA BARCODING Genetic barcoding of marine (Ozobranchus spp.) from Florida sea turtles and their divergence in host specificity

AUDREY E. McGOWIN,* TRIET M. TRUONG,* ADRIAN M. CORBETT,† DEAN A. BAGLEY,‡§ LLEWELLYN M. EHRHART,‡§ MICHAEL J. BRESETTE,§ STEVEN T. WEEGE§ and DAVE CLARK§ *Department of Chemistry, Wright State University, Dayton, OH 45435, USA, †Department of Neuroscience, Cell Biology and Physiology, Wright State University, Dayton, OH 45435, USA, ‡Department of Biology, University of Central Florida, Orlando, FL 32816, USA, §Inwater Research Group, Inc., 4160 NE Hyline Dr., Jensen Beach, FL 34957, USA

Abstract Ozobranchus margoi and Ozobranchus branchiatus are the only two of marine turtle leeches (Ozobranchus spp.) known to inhabit the Atlantic coast of the United States and theGulfofMexico.Inearlyreports of fibropapillomatosis (FP) in green turtles (Chelonia mydas), O. branchiatus was implicated as a vector in the transmission of Fibropapilloma- associated turtle herpesvirus (FPTHV). It is imperative that the species be identified to elucidate the role Ozobranchus spp. may play in disease transmission. In this study, Ozobranchus branchiatus has been identified for the first time on a log- gerhead (Caretta caretta) turtle, and the molecular data for this species is now available for the first time in GenBank. Both species of leeches were also found infecting a single C. mydas. Using morphological combined with distance- and character-based genetic sequence analyses, this study has established a DNA barcode for both species of Ozobranchus spp. leech and has shown it can be applied successfully to the identification of leeches at earlier stages of development when morphological taxonomy cannot be employed. The results suggest a different haplotype may exist for O. branchiatus leeches found on C. caretta versus C. mydas. Leech cocoon residue collected from a C. mydas was identified using the new method.

Keywords: cocoon, DNA barcoding, Fibropapillomatosis-associated turtle herpesvirus, new host record, Ozobranchus branchiatus, Ozobranchus margoi, sea turtle leech, two species infection Received 1 September 2010; revision received 11 October 2010; accepted 21 October 2010

1975). A distinguishing morphological characteristic of Introduction the Ozobranchidae family is finger-like terminally There are only two known species of marine turtle branching lateral protuberances (gills or branchidae) that leeches (Ozobranchus spp.) known to inhabit the Atlantic decrease in complexity and size from the anterior to the coast of the United States and the Gulf of Mexico, Ozo- posterior and begin on segment XIII of the abdomen branchus margoi and Ozobranchus branchiatus. The primary (Sawyer 1986). The primary difference in these two spe- host for O. margoi is the loggerhead turtle (Caretta caretta) cies of leeches is in the number of gills with one pair per but O. margoi has been reported on green turtles (Chelonia segment; O. margoi has five pairs (Davies 1978) and mydas), hawksbill turtles (Eretmochelys imbricata) and O. branchiatus sports seven pairs (Raj & Penner 1962). Kemp’s Ridley turtles (Lepidochelys kempi) as well (Bunk- Both species have a short, segmented body with two eye ley-Williams et al. 2008). Only rarely is a host other than spots, a proboscis and a large posterior sucker. These C. mydas reported for O. branchiatus, although its occur- leeches attach to the mouth, neck, cloaca and under the rence on black turtles (Chelonia agassizi a.k.a. Chelonia flippers of turtles and deposit eggs in cocoons on the car- mydas of the Eastern Pacific) (McDonald & Dutton 1990), apace. Almost nothing is known about the life cycle of hawksbill turtles (Bunkley-Williams et al. 2008), and olive sea turtle leeches, including whether they can exist in the ridley turtles (Lepidochelys olivacea) in Mexico has been absence of a turtle host or if they have alternate hosts. documented (Hernandez-Vazquez & Valadez-Gonzalez Ozobranchus spp. can be very small (millimetres in 1998). These warm water marine leeches belong to the length) with various life stages so differentiation amongst family , Ozobranchidae (Sawyer et al. species is difficult. Only a few descriptions of Ozobran- Correspondence: Audrey E. McGowin, Fax: +1 937 775 2717; chus spp. exist in the scientific literature (Sawyer et al. E-mail: [email protected] 1975) primarily because of the challenges of studying

2010 Blackwell Publishing Ltd 272 DNA BARCODING their hosts, a formidable task because of the limited aligned and unique substitutions at specific positions can knowledge of sea turtle life history patterns (Bolten 2003). be identified as characteristic attributes (CAs). A set of Interest in Ozbranchus spp. has increased in recent these CAs for any species becomes the character-based years because of its possible relationship to a panzootic ‘DNA barcode’ for species identification. The identifica- that has primarily afflicted C. mydas but has now spread tion of a new species with DNA barcoding and those not to other sea turtles (Herbst 1994; Williams et al. 1994). included in a genomic repository must be accompanied The disease, fibropapillomatosis (FP), is a condition char- by standard taxonomy (DeSalle et al. 2005). Still care acterized by the growth of multiple visceral and cutane- must be exercised when using a single gene such as COI ous fibrovascular tumours on the eyes, neck, cloaca and alone because more than one gene may be necessary to flippers of sea turtles (the same places where leeches delineate phylogenetic relationships between species attach themselves), which can interfere with the turtle’s (DeSalle et al. 2005). ability to swim, see and feed. Fibropapilloma-associated For leeches, DNA barcoding could be particularly use- turtle herpesvirus (FPTHV) has been identified as the ful, as their identification using standard taxonomic tech- causative agent of FP (Quackenbush et al. 1998), although niques can sometimes be ambiguous (Bely & Weisblat the primary vector triggering this chronic tumour-form- 2006). In addition, morphological characteristics are only ing disease is still unknown. In early reports of FP in useful when dealing with more mature specimens, not C. mydas, O. branchiatus was implicated as a vector specimens in the larval or cocoon stages, whereas DNA organism (Smith & Coates 1938; Nigrelli & Smith 1943). barcoding would be useful for all stages of development. Four separate viral variants were detected in an analysis DNA barcoding of marine leeches could facilitate marine of FPTHV in sea turtles in Florida waters (Ene et al. 2005). leech life cycle elucidation because the nucleotide The viral variant specific to Indian River Lagoon, a green sequence will be identical at every life stage. Although turtle foraging area along the Atlantic coast of Florida, the phylogeny of leeches has been studied extensively appears to be endemic to the habitat, leading to the sup- through the lens of COI sequences (Siddall & Burreson position that FP is transmitted during the juvenile phase 1998; Light & Siddall 1999; Utevsky et al. 2007) and when turtles move to nearshore feeding grounds after includes O. margoi (GenBank accession number spending several years at sea (Ene et al. 2005). Therefore, AF003268; Siddall & Burreson 1998), O. branchiatus has a vector organism could be responsible for the epizootic previously been excluded as no COI sequence had been occurring at Indian River Lagoon where as high as 72% determined. In this study, COI sequences of both species of C. mydas have been affected (Hirama & Ehrhart 2007). of leeches were obtained for specimens collected from An analysis of FPTHV concentration in five types of tur- sea turtles captured by net in the Indian River Lagoon tle ectoparasites collected in Hawaii showed only Ozo- and the St. Lucie Power Plant on Hutchinson Island in branchus spp. leeches contained sufficient viral load for Florida, resulting in the addition of O. branchiatus to the transmission of the virus between turtles (Greenblatt GenBank database. Using COI sequences of closely et al. 2004). Unfortunately, the leech species was not spec- related species already in GenBank (Siddall & Burreson ified in this publication. Knowledge of the species 1998; Light & Siddall 1999; Utevsky et al. 2007), DNA bar- removed from turtles with FP could shed light on the role codes were generated for both of these species of leeches of marine leeches as a possible vector organism in disease in the Ozobranchidae family and morphological data transmission (Williams & Bunkley-Williams 2006; Bunk- were used to corroborate DNA barcode identification. ley-Williams et al. 2008). These simple character-based attributes (sCAs) were Sequencing of particular mitochondrial genes in ani- applied to a sample of leech cocoon residue (a sample for mals, such as the cytochrome c oxidase I (COI) gene, can which taxonomic identification would be impossible) yield phylogenetic information as well as aid in the iden- taken from a C. mydas during a cold-stunning event that tification of species that have molecular data in genetic occurred in Florida in January 2010. databases such as BOLD, NCBI, GBIF, DDBJ or EMBL. The Universal Folmer primers (Folmer et al. 1994) have Materials and methods been used successfully for the amplification of COI genes in other species of leeches (Siddall & Burreson 1998; Light Collection sites & Siddall 1999; Utevsky et al. 2007). Unfortunately, spe- cies for which a gene sequence has not been recorded Leeches were collected from turtles captured by net in cannot be identified by a database search. DNA barcod- the Indian River Lagoon, Florida about 3 miles south of ing is a genomic method used to distinguish between Sebastian Inlet (27 49¢ N, 80 27¢ W). The Indian River various species of organisms (Hebert et al. 2003; Ratnas- Lagoon is a 150-mile long estuary along the Florida ingham & Hebert 2007). Once a DNA sequence is Atlantic Coast with five ocean inlets. It serves as a forag- obtained, closely related species’ sequences can be ing habitat for turtles (Hirama & Ehrhart 2007) and has

2010 Blackwell Publishing Ltd DNA BARCODING 273 significant input from agricultural runoff and sewage opmental stage and size. The cocoon residue was tan and treatment plants. Leeches were also collected from sea green in colour and contained green algae-like flecks. turtles rescued from the cooling water intake canal of the Microscopic examination revealed small circular open- St. Lucie Nuclear Power Plant on Hutchinson Island, ings through which most of the larvae had apparently Florida (27 21¢ N, 80 14¢ W). Turtles entering the intake already hatched. See Supporting Information for photos canal come from the Atlantic Ocean on the east side of of Ozobranchus branchiatus and Ozobranchus margoi from the island and are removed from the canal and returned the St. Lucie Power Plant on Hutchinson Island, FL. to the ocean within a day. The two sites are at approxi- mately the same latitude. DNA extraction, PCR amplification and sequencing Genomic DNA was extracted from individual and com- Sample collection bined specimens. The number of pooled specimens anal- The specimens were analysed and their locations are ysed for each individual turtle species is listed in Table 1. listed in Table 1, along with GenBank accession numbers Prior to DNA extraction, all tissues were washed with for their COI DNA sequences. Specimens were collected double deionized water. Only tissues obtained from the in the field directly from C. caretta and C. mydas that were branchidae, anterior sucker, and posterior (caudal) released soon after capture. The host species was sucker of the leech were analysed (13–25 mg) to avoid recorded at the time of collection. All leech specimens contamination by host DNA in the gut. The leech cocoon were postlarval with branchidae in various stages of residue sample had a mass of 30.6 mg. DNA extraction maturity. Leech cocoon residue was collected from a and purification were performed using the procedures C. mydas found in the Indian River Lagoon in or near Bre- and materials provided in a Qiagen DNeasy Blood and vard County, Florida during a cold-stunning event on Tissue kit (Qiagen, Inc., Valencia, CA, USA). Materials January 12, 2010. Leeches and cocoon residue were pre- not provided in the kit were proteinase K (Qiagen, Inc.), served in 99% denatured ethanol at ambient tempera- RNase (Qiagen, Inc.) and ethanol. Purified genomic DNA ture. was eluted by adding 200 lL of Buffer AE to the same spin column in a new collection tube and centrifuged at 5200 g for 1 min. The centrifuge step was repeated once Morphological identification again using the same spin column and collection tube for Each leech was examined with a microscope. Leeches a total of 400-lL sample volume. DNA samples were were assigned an identity based upon the number and stored at )20 C and concentration measured with a ND- shape of branchidae (gills) on each side of the abdomen; 1000 NanoDrop Spectrophotometer (Thermo Scientific, O. branchiatus has seven equally sized pairs, while Wilmington, DE, USA). O. margoi has only five pairs (Table 1). The complexity Mitochondrial COI sequences (658 bp) were amplified and branching of the gills and the progressive size reduc- from purified genomic DNA using the Universal Folmer tion towards the posterior are much more pronounced in primers LCO 1490 (5¢-GGT CAA CAA ATC ATA AAG O. margoi than in O. branchiatus. Leeches were transpar- ATA TTG G-3¢) as forward primer and HCO2198 ent to whitish in colour and varied greatly in their devel- (5¢-TAA ACT TCA GGG TGA CCA AAA AAT CA-3¢)as

Table 1 Sampling locations in Florida for two species of sea turtle leeches with host species and GenBank accession numbers. Also listed are the haplotype designations. The symbol n is the number of pooled specimens

St. Lucie Power Plant on Hutchinson Island (27 21¢ N, 80 14¢ W) Indian River Lagoon (27 49¢ N, 80 27¢ W)

GenBank GenBank Number of Host Haplotype Collection accession Host Haplotype Collection accession Taxon gill pairs species designation date(s) (n) number species designation date (n) number

Ozobranchus 7 Chelonia mydas OB-CM 3.21.2010 (2); GU985465 Chelonia mydas OB-CM 12.7.2009 (12) GU985465 branchiatus 3.23.2010 (2)* Caretta caretta OB-CC 8.25.2009 (2) GU985466 Ozobranchus 5 Caretta caretta 8.13.2009 (2) GU985467† margoi Chelonia mydas 3.23.2010 (1)* HM590711†

*Collection dates are identical because leech specimens were collected from the same individual sea turtle. †Accession number is different because leeches were found on different hosts and not based upon any genetic variance or haplotype difference.

2010 Blackwell Publishing Ltd 274 DNA BARCODING reverse primer (Folmer et al. 1994) synthesized by Invi- identify simple characteristic attributes (sCAs) for the trogen Corporation (Carlsbad, CA, USA). The reaction Ozobranchidae family. Nucleotide positions were solutions contained 25 lL AmpliTaq Gold 360 (Applied selected only if unique characters existed pertaining to Biosystems, Foster City, CA, USA), 5 lLofa2-lM solu- O. branchiatus (found on C. mydas or C. caretta) and tion of each primer, 70–100 ng of DNA template and O. margoi. The COI genetic sequence of cocoon residue enough distilled, deionized water for a total volume of obtained from the carapace of the C. mydas was examined 50 lL. DNA and primer blank samples were run with at all these positions to determine whether the cocoon each batch of samples. The PCR thermal regime for residue originated from Ozobranchus spp. leeches. amplification was 10 min at 95 C, followed by 35 cycles of 30 s at 95 C, 30 s at 50 C, 60 s at 72 C and a final Results elongation of 7 min at 72 C using an Applied Biosys- tems PCR System 2700 thermocycler (Singapore). Purifi- Host species divergence: leech-host pairings at sampling cation of PCR products was performed using the sites procedures and materials provided in a QIAquick PCR Purification Kit (Qiagen, Inc.). Amplification products At the Indian River Lagoon site, 12 of 29 Ozobranchus were sequenced in both directions by Retrogen, Inc. (San branchiatus leeches collected from one C. mydas were Diego, CA, USA). Alignment analysis of nucleic acid con- selected for genetic analysis. Nine individual leeches sensus sequences was performed using ClustalW2 were selected from those collected at the St. Lucie Power (EMBL-EBI). The invertebrate mitochondrial genetic code Plant location from four different turtles. Of the nine was used for translating DNA sequences to protein leeches, two were O. margoi from a C. caretta and another sequences (start codon at second position) with EMBOSS two were O. branchiatus from a C. mydas. A second Transeq (EMBL-EBI). DNA sequences were submitted as C. mydas captured at this location was the only turtle to a nucleotide query using BLAST to determine whether a be infected with both species of leech. Interestingly, match existed in the NCBI GenBank. another C. caretta from the site had two O. branchiatus leeches. This is the first report of O. branchiatus on a C. caretta and the first report of both leech species infect- Distance-based tree approach to species identification ing a single C. mydas. The turtles netted at this site were A BLAST search of GenBank was conducted using the COI captured individually or only a few at a time, so there consensus sequences of leeches sampled in this study. was no overcrowding of turtle specimens. Turtles were Eleven species of the order Rhynchobdellida (one Glossi- returned to the ocean within a day of capture. phoniidae and ten piscicolids from three subfamilies) with molecular data most closely matching the ozobran- Genetic diversity chids’ COI sequences were selected for comparison. Using MEGA 5 Beta (Tamura et al. 2007), a distance-based All GenBank accession numbers for new sequences are tree approach to species identification was carried out by listed in Table 1. The molecular data of O. margoi analy- neighbour-joining the COI sequences of recorded speci- sed in this study from different turtle species (C. mydas mens from the BLAST search and those analysed in our and C. caretta) collected at St. Lucie Power Plant matched study. The Tajima & Nei (1984) method served as the 100% but achieved a lower similarity score of 99% (pair- substitution model for the neighbour-joining tree and wise distance value of 0.0015) when compared to the pairwise distance calculations conducted in MEGA 5 Beta. O. margoi sequence by Siddall & Burreson (GenBank Gaps were treated using the pairwise deletion option, [accession number AF003268]) obtained from a single and a gamma parameter of 0.24 was selected. Addition- leech on a C. caretta at Virginia Beach, VA. Our sequence ally, the bootstrap method was employed as a test of phy- chromatogram strongly indicates an additional thymine logeny using 500 bootstrap replications. Pairwise (T) should be present at the beginning of the genetic distances were given for clade with bootstrap percent- sequence (four T’s instead of three T’s). This difference ages above 95% and offer a tentative indication of the suggests primers were attached either at a different loca- similarity cut-off to determine species status in a certain tion in the chromosome or perhaps the O. margoi leeches clade. in Florida belong to a separate haplotype. The O. branchiatus leeches unexpectedly identified on a C. caretta at the St. Lucie Power Plant (designated OB- Character-based diagnostic approach to species CC) and those found on C. mydas in the Indian River delimitation Lagoon (designated OB-CM) had 25 variable sites in their All COI sequences employed in the neighbour-joining COI sequences (each 658 bp, 219 amino acids). These tree were aligned in MEGA 5 Beta and then examined to nucleotide changes represent 3.8% of the data set and

2010 Blackwell Publishing Ltd DNA BARCODING 275 yielded only one amino acid (AA) change. At AA 102, C. caretta and a C. mydas were 0.3139 and 0.2975, respec- only O. branchiatus found on a C. caretta possess isoleu- tively. As the pairwise distance for the two most closely cine, while all other specimens, including O. margoi, had related distinct species, Austrobdella bilobata and Austrobd- valine. Differences at nucleotide positions between the ella translucens, was significantly lower (0.1073), the two distinct Ozobranchus species resulted mainly in syn- leeches taxonomically identified as O. branchiatus must onymous substitutions but four differences were nonsyn- be a separate species from O. margoi. Furthermore, pair- onymous. These included alanine to glycine at AA 8, wise distance values also suggest the O. branchiatus threonine to serine at AA 40, valine to isoleucine at AA leeches found on C. mydas and C. caretta belong to the 41 and isoleucine to valine at AA 47 (O. margoi AA versus same species for they share a score of 0.0438, which is O. branchiatus AA). close in magnitude to the pairwise distance value (0.0109) calculated for Myzobdella lugubris HI and Myzobdella lugubris VA (specimens of the same species but found at Genetic distance and tree-building different locations). However, it should be noted that Prior to this study, O. margoi was the only member of the whether Myzobdella spp. from different locations and Ozobranchidae family to be included in the GenBank environments really belong to the same species is an issue database. As the molecular data for O. branchiatus was open to further discussion (Williams & Burreson 2006). absent from all international databases until now, a BLAST search of the COI sequence for the species in Genbank Character-based diagnosis originally yielded an 84% raw similarity score matching O. margoi. From the neighbour-joining tree (Fig. 1), Characteristic attributes generated from the results O. branchiatus is found to be the sister taxon of O. margoi obtained in these analyses occurred at 51 of 658 nucleo- with a bootstrap value of 100 for this specific clade. Only tide positions (Table 2). Twelve pure diagnostic charac- the topology of branches supported by bootstrap values ters were found for the Ozobranchidae family. above 95% is considered to be reasonably true. The Diagnostic characters distinguishing O. margoi from pairwise distances computed between the COI sequence O. branchiatus were apparent for 31 positions, regardless of O. margoi and the sequences of O. branchiatus on a of turtle host species for O. branchiatus. At certain

Branchellion lobata (DQ414307*)

65 Branchellion parkeri (DQ414308*) 94 Branchellion torpedinis (AF003265§)

53 Austrobdella bilobata (DQ414301*) 0.1073 99 Austrobdella translucens (DQ414306*) 35 Johanssonia arctica (DQ414320*)

30 Myzobdella lugubris HI (DQ414325*) 89 0.0109 100 Myzobdella lugubris VA (DQ414324*)

Calliobdella lophii (DQ414314*)

72 Oxytonostoma typica (EF405596†)

Glossiphonia complanata (AY047321‡)

Ozobranchus margoi (GU985467)

100 Ozobranchus branchiatus CC (GU985466) 0.3139 (CC) and 0.2975 (CM) 0.0438 100 Ozobranchus branchiatus CM (GU985465)

Fig. 1 MEGA 5 Beta generated neighbour-joining tree based upon the COI sequences of specimens analysed in this study and eleven oth- ers closely related to them. Pairwise distances are given for clades with bootstrap percentages above 95% (enclosed by brackets). Boot- strap values (in percentage of 500 bootstrap replicates) are given below the tree branch. Eleven additional leeches not sampled in this study include one Glossiphoniidae and ten Piscicolidaes (comprised of three subfamilies). The thick black branch indicates the specimen belonging to the Glossiphoniidae family, while leeches in the Piscicolidae subfamilies are represented by thick grey branches. GenBank accession numbers are given along with the taxa names and their associated authors. *Williams & Burreson 2006; †Utevsky et al. 2007; ‡Light & Siddall 1999; §Siddall & Burreson 1998.

2010 Blackwell Publishing Ltd 276 DNA BARCODING

Table 2 DNA barcodes for Ozobranchus spp. sea turtle leeches based on pure diagnostic characters

Nucleotide positions were selected only if unique characters exist pertaining to the Ozobranchidae family. Pure diagnostic characters that are found solely in the Ozobranchidae family are shaded in grey. Any nucleic ambiguity or absence of nucleotide at a given posi- tion is represented by an ‘N’. At certain positions, pure characters occur in only some members of a species, also known as private characters.

positions, pure characters occur in only some members of quality of the sample or possibly a third haplotype that a species, also known as private characters. Private char- has yet to be identified. acters occur at 12 positions only for O. branchiatus. For instance, at positions 343 and 628, a ‘C’ diagnoses OB-CC Discussion found on a C. caretta. In addition, OB-CC was separated from all other species of leeches by seven CAs (positions The FP panzootic affecting mainly C. mydas initially has 103, 163, 319, 343, 535, 625, and 628), while OB-CM was now appeared to a lesser degree in other turtle species, separated by four CAs (positions 140, 292, 370, and 518). including C. caretta, that inhabit the same waters. An Position 50 is the only position to contain a distinct pure accurate method for species identification is imperative if CA for each species of leech; a ‘T’ for O. branchiatus and a future studies are performed to determine what species ‘C’ for O. margoi. of Ozobranchus spp. leech is a possible vector organism As the cocoon residue came from posthatched sam- for transmitting FPTHV in sea turtles. Previous studies ples, the molecular data sequenced contained ambigui- have failed to identify what species of Ozobranchus spp. ties. The chromatogram of the reverse sequence was not leeches were collected from turtles, while others relied considered as a result of the low signal-to-noise ratio solely on morphological taxonomy in the absence of that ranged from 45 to 88 (Sequence Scanner v1.0). DNA barcoding. This study is the first to combine mor- Therefore, the forward sequence was used for compari- phological taxonomy with DNA barcoding as an effective son instead of a consensus sequence. Ten of the 12 pure tool to distinguish O. margoi and O. branchiatus at various diagnostic characters for the Ozobranchus spp. family stages of their life cycle. Despite all the information gen- matched (positions 56 and 483 were mismatches). erated, DNA barcoding that only utilizes a tree-based Twenty-one of the 31 characters distinguishing O. mar- approach cannot absolutely confirm species identity nor goi from O. branchiatus matched O. branchiatus, whereas yield substantial hypotheses regarding phylogenetic rela- only one character (position 649) indicated O. margoi. tionships. The neighbour-joining tree generated is based The other positions were mismatched or undefined in solely on a single gene, which raises the possibility it may the chromatogram. Of the 12 characters that distinguish represent the evolutionary divergence of COI amongst the host turtle species, eight indicated the host species the species rather than their phylogenetic relationships was a C. mydas (OB-CM). Two characters (positions 370 (Krane & Raymer 2003). The use of DNA data to distin- and 628) matched OB-CC. One was a mismatch for both guish species must not only be accompanied by morpho- haplotypes (position 571) and the other was undefined logical taxonomy, when possible, but also must be in the chromatogram. This is compelling evidence that consistent with current taxonomic practices (DeSalle et al. the residue found on the carapace of the C. mydas was 2005). Although the molecular data from the COI gene indeed from an O. branchiatus spp. leech. Although it alone may only be enough to produce a gene tree rather more closely matches the OB-CM haplotype, the residue than a species tree, the DNA data can yield significant sequence does contain characteristics of the OB-CC hap- characters for nontree based diagnostic approaches that lotype. The differences could be attributed to the poor are more in line with current taxonomic research. Greater

2010 Blackwell Publishing Ltd DNA BARCODING 277 confidence can be achieved in species delimitation by conservation, this study provides the information needed employing both morphological taxonomy and DNA bar- to adequately identify these leeches. The criteria for an coding in terms of both distance- and character-based effective DNA barcode require that it incorporate genet- techniques. ics, morphology, species behaviour, geographical infor- The character-based diagnosis of O. margoi and mation and other valid species delimitation attributes O. branchiatus reveal a significant number of CAs unique (DeSalle 2006). Our study has established such a DNA to the Ozobranchidae family and include those vital to barcode for two species of Ozobranchidae leech and has species differentiation within this particular family of shown it can be applied successfully to species identifica- leeches. It is important to note that morphological identi- tion even when morphological taxonomy cannot be fication is not always effective at all stages of species employed. Future research should include the identifica- development. A hypothesis can be drawn that the cocoon tion of all leeches species found on individual turtles, residue found on a C. mydas in this study came from an especially those turtles with FP. O. branchiatus leech based upon knowledge of what para- sitic specimens live on C. mydas. The set of CAs estab- Acknowledgements lished for Ozobranchid leeches (Table 2) sampled in this study offers a means of identifying the cocoon residue Dean Rider and Dan Krane of Wright State University provided when morphological identification becomes virtually valuable bioinformatics tutelage. Appreciation is expressed to impossible. The cocoon residue came from samples that John Stireman, III and Jeremy Heath for assistance with photog- raphy and to Ashley Marshall for assistance with the manu- were already hatched, thus, lacking enough DNA to gen- script. Special thanks go to John A. Case for inspiration and erate a strong genetic sequence defined by limited ambi- assistance with travel. guities. Despite the limitations imposed by the absence of larvae in the sample sequenced, multiple selected nucleo- tide positions in the residue DNA sequence reveal pure References CAs associated with the Ozobranchid leeches. Further- more, at certain positions, the sampled residue more Bely AE, Weisblat DA (2006) Lessons from leeches: a call for DNA barcod- ing in the lab. Evolution & Development, 8, 491–501. closely matched the CAs established for O. branchiatus Bolten AB (2003) Variation in sea turtle life history patterns: neritic vs. than O. margoi. Better quality leech cocoon samples that oceanic developmental stages. In: The Biology of Sea Turtles, Volume II have not hatched would allow for higher quality DNA (eds Lutz PL, Musick JA & Wyneken J), pp. 243–257. CRC Press, Boca sequences to better compare selected nucleotides to the Raton, FL. Bunkley-Williams L, Williams EH, Horrocks JA, Horta HC, Mignucci- established set of CAs. Giannoni AA, Poponi AC (2008) New leeches and diseases for the In general, the significance of studying Ozobranchus hawksbill sea turtle and the West Indies. Comparative Parasitology, 75, spp. leeches centres on their connection to sea turtle con- 263–270. servation. All species of sea turtles are threatened and Davies RW (1978) The morphology of Ozobranchus margoi (Apathy) (Hiru- dinoidea), a parasite of marine turtles. The Journal of Parasitology, 64, evidence collected over the years indicates Ozobranchus 1092–1096. spp. leeches may play a significant role in the spread of DeSalle R (2006) Species discovery versus species identification in DNA FP amongst sea turtles (Herbst 1994). This study is barcoding efforts: response to Rubinoff. Conservation Biology, 20, 1545– 1547. the first to record the appearance of O. branchiatus on a DeSalle R, Egan MG, Siddall M (2005) The unholy trinity: taxonomy, spe- C. caretta; an important finding revealing the gradual cies delimitation and DNA barcoding. Philosophical Transactions of the divergence of O. branchiatus in host specificity. This and Royal Society. Series B, 360, 1905–1916. the other recent reports of divergence in host specificity Ene A, Su M, Lemaire S et al. (2005) Distribution of chelonid fibropapillo- matosis-associated herpesvirus variants in Florida: molecular genetic for O. branchiatus could shed light on the recent spread of evidence for infection of turtles following recruitment to neritic devel- sea turtle FP from C. mydas to other turtle species, further opmental habitats. Journal of Wildlife Diseases, 41, 489–497. implicating O. branchiatus as a possible vector organism Folmer O, Black M, Hoeh W, Lutz R, Vrijenhoek R (1994) DNA primers in FPTHV transmission. If O. branchiatus is a vector for amplification of mitochondrial cytochrome c oxidase subunit I from diverse metazoan invertebrates. Molecular Marine Biology and Biotech- organism triggering FP in sea turtles, the evidence pro- nology, 3, 294–299. vided in this study may help explain why the panzootic GenBank http://www.ncbi.nlm.nih.gov/GenBank/ (for Ozobranchus affecting mainly C. mydas (its primary host) initially has margoi cytochrome c oxidase subunit I (COI) gene, partial cds; mito- now spread to a lesser degree in other turtle species. chondrial [accession number AF003268]; accessed 14 June 2010). Greenblatt RJ, Work TM, Balazs GH, Sutton CA, Casey RN, Casey JW Additionally, the molecular data for O. branchiatus made (2004) The Ozobranchus leach is a candidate mechanical vector for the available for the first time in GenBank and other genetic fibropapilloma-associated turtle herpesvirus found latently infecting databases will provide information needed by research- skin tumors on Hawaiian green turtles (Chelonia mydas). Virology, 321, ers interested in further elucidating the phylogenetic 101–110. Hebert PDN, Cywinska A, Ball SL, deWaard JR (2003) Biological identifi- relationships of leeches in the Rhynchobdellida order. cations through DNA barcodes. Proceedings of the Royal Society of Lon- However, for those primarily concerned with sea turtle don. Series B, Biological Sciences, 270, 313–321.

2010 Blackwell Publishing Ltd 278 DNA BARCODING

Herbst LH (1994) Fibropapillomatosis of marine turtles. Annual Reviews of Tajima F, Nei M (1984) Estimation of evolutionary distance between Fish Diseases, 4, 389–425. nucleotide sequences. Molecular Biology and Evolution, 1, 269–285. Hernandez-Vazquez S, Valadez-Gonzalez C (1998) Observations of the Tamura K, Dudley J, Nei M, Kumar S (2007) MEGA4: molecular evolu- epizoa found on the turtle Lepidochelys olivacea at La Gloria, Jalisco, tionary genetics analysis (MEGA) software version 4.0. Molecular Biol- Mexico. Ciencias Marinas, 24, 119–125. ogy and Evolution, 24, 1596–1599. Hirama S, Ehrhart LM (2007) Description, prevalence and severity of Utevsky SY, Utevsky AY, Schiaparelli S, Trontelj P (2007) Molecular green turtle fibropapillomatosis in three developmental habitats on the phylogeny of pontobdelline leeches and their place in the descent east coast of Florida. Florida Scientist, 70, 435–448. of fish leeches (Hirudinea, Piscicolidae). Zoologica Scripta, 36, 271– Krane DE, Raymer ML (2003) Fundamental Concepts of Bioinformatics. 280. Benjamin Cummings, San Francisco, CA. Williams EH, Bunkley-Williams L (2006) Early fibropapillomas in Hawaii Light JE, Siddall ME (1999) Phylogeny of the leech family Glossiphoniidae and occurrences in all sea turtles species: the panzootic, associated based on mitochondrial gene sequences and morphological data. Jour- leeches wide-ranging on sea turtles, and species of study leeches nal of Parasitology, 85, 815–823. should be identified. Journal of Virology, 80, 4643–4644. McDonald D, Dutton P (1990) Fibropapillomas on sea turtles in San Diego Williams JI, Burreson EM (2006) Phylogeny of the fish leeches (Oligochae- Bay, California. Marine Turtle Newsletter, 51, 9–10. ta, Hirudinida, Piscicolidae) based on nuclear and mitochondrial genes Nigrelli RF, Smith GM (1943) The occurrence of leeches, Ozobranchus and morphology. Zoologica Scripta, 35, 627–639. branchiatus (Menzies), on fibro-epithelial tumors of marine turtles, Williams EH, Bunkley-Williams L, Peters EC et al. (1994) An epizootic of Chelonia mydas (Linnaeus). Zoologica, 28, 107–108. cutaneous fibropapillomas in green turtles Chelonia mydas of the Quackenbush SL, Work TM, Balazs GH et al. (1998) Three closely related Caribbean: part of a panzootic? Journal of Aquatic Health, 6, 70– herpesviruses are associated with fibropapillomatosis in marine turtles. 78. Virology, 246, 392–399. Raj PJS, Penner LR (1962) Concerning Ozobranchus branchiatus (Menzies, 1791) (Piscicolidae: Hirudinea) from Florida and Sarawak. Transactions Supporting Information of the American Microscopical Society, 81, 364–371. Ratnasingham S, Hebert PD (2007) BOLD: the barcode of life data Additional supporting information may be found in the system (http://www.barcodinglife.org). Molecular Ecology Notes, 7, online version of this article. 355–364. Sawyer RT (1986) Leech Biology and Behaviour: Volume I Anatomy, Physiol- Appendix S1 Photos of Ozobranchus branchiatus and Ozo- ogy, and Behaviour. pp. 101–104. Oxford University Press, New York. branchus margoi from the St. Lucie Power Plant on Hutch- Sawyer RT, Lawler AR, Overstreet RM (1975) Marine Leeches of the east- ern United States and the Gulf of Mexico with a key to the species. Jour- inson Island, FL. nal of Natural History, 9, 633–667. Siddall ME, Burreson EM (1998) Phylogeny of leeches (Hirudinea) based Please note: Wiley-Blackwell are not responsible for the on mitochondrial cytochrome c oxidase subunit I. Molecular Phylogenet- content or functionality of any supporting information ics and Evolution, 9, 156–162. supplied by the authors. Any queries (other than missing Smith GM, Coates CW (1938) Fibro-epithelial growths of the skin in large material) should be directed to the corresponding author marine turtles, Chelonia mydas (Linnaeus). Zoologica (New York), 23, 93– 98. for the article.

2010 Blackwell Publishing Ltd