Species Selection

To predict the size of the olfactory repertoire in turtles across diverse habitats and feeding strategies, 7 species native to the southeastern United States were selected. Three species of sea turtles were chosen for this study, leatherback, loggerhead, and green turtles. These 3 species represent the array of feeding strategies used by sea turtles; primarily carnivorous forager (loggerhead), herbivorous grazer (green), and specialist (leatherback) (Limpus 1973; Balazs 1979; Mortimer 1981; Ernst et al. 1994; Davenport 1998; Tomas et al. 2001; Constantino and Salmon 2003). Painted turtles (Chrysemys picta bellii) are freshwater omnivorous generalists that forage for prey in aquatic vegetation relying heavily on visual stimuli (Ernst et al. 1994; Rowe and Parsons 2000). Musk turtles (Sternotherus odoratus) are freshwater omnivores and scavengers that forage for prey both underwater and on land (Mahmoud 1968; Ernst et al. 1994; Ford and Moll 2004). Box turtles (Terrapene carolina) are primarily terrestrial omnivores that forage for food in water and on land (Klimstra and Newsome 1960; Strang 1983). Finally, gopher tortoises (Gopherus polyphemus) are fully terrestrial herbivores that appear to eat only select plant genera (MacDonald and Mushinsky 1988; Ernst et al. 1994).

Genomic DNA Isolation

DNA isolation for this study was noninvasive. Blood samples from all species were obtained from several research groups with stored samples for ongoing population genetics studies. Genomic DNA was isolated from 100 µL whole blood diluted 1∶1 with lysis buffer (100 mM Tris, 100 mM EDTA and 1% SDS) by using a QIAgen QIAmp Tissue Kit, which yielded approximately 1 µg DNA per 1-µL product.

Primer Design and Validation

Degenerate polymerase chain reaction (PCR) primers were designed against an alignment of conserved regions from known OR gene sequences of rat, mouse, chicken, and human to conduct a broad survey of OR genes present in turtles. The sens primer was designed from a conserved sequence in the second transmembrane domain and reads (5′-CCYATGTAYTTBTTBCT-3′). The antisens primer was designed from conserved sequence in the sixth transmembrane domain and reads (5′–GSHRCADGTNKARAADGCYT-3′). To test the inclusiveness of these primers, a virtual PCR was performed by using a linux script that allowed the primers to search the full repertoire of mouse OR gene sequences. Of the 267 mouse OR gene families identified by Zhang and Firestein (2002), only 11 large gene families (those with 10 or more OR gene sequences) were excluded in this search. Although these primers are unlikely to sample the entire OR gene repertoire of turtles, they were accepted as reasonable for a survey of most OR gene families.

PCR Amplification of OR Genes

PCR was performed by using 1 µg (1 µL) of genomic DNA as template. The reaction mix included 5 µL 10× Promega buffer, 1 µL 10 mM dNTPs, 3 µL 25 mM MgCl², 0.5 µL Promega taq polymerase, 5 µL 25 mM sens and antisens, and 29.5 µL sterile distilled water for a final volume of 50 µL. PCR was performed on a Perkin Elmer Cetus Thermocycler by using the following touchdown PCR protocol: 2 minutes at 94° ×1, 10 seconds at 94°, 30 seconds at 39°, 1 minute at 74° ×2, 10 seconds at 94°, 30 seconds at 38°, 1 minute at 74° ×2, 10 seconds at 94°, 30 seconds at 37°, 1 minute at 74° ×2, 10 seconds at 94°, 30 seconds at 36°, 1 minute at 74° ×30, and 3 minutes at 74° Celcius. PCR products were evaluated for size distribution by Ethidium Bromide staining on a 1.5% agarose gel and acceptably sized bands (approximately 550 base pairs) were excised. DNA was isolated by using a Bio 101 Geneclean Turbo Kit.

Cloning and Sequencing

Products were end polished then ligated into a cloning vector by using a Stratagene PCR Script Amp Cloning Kit. Ligated products were then electroporated into supplied bacterial cells and plated on Xgal/IPTG prepared agar plates. Approximately 200 of the resulting colonies were selected based on blue and white screening and were transferred to a master plate for short-term storage and analysis. A QIAgen Miniprep Kit was used to purify the vector and insert, and 1 µL of this product was used in a sequencing PCR with 8 µL 0.25 ABI Big Dye solution, 2 µL T3 vector primer, and 9 µL water for a final volume of 20 µL. PCR products were precipitated then sequenced by using an ABI Prism 377 Sequencer. Clones later identified as ORs by BLAST search comparisons were sequenced again by using the T7 and M13 Forward vector primers to ensure sequence accuracy.

Sequence Analysis

Sequences were initially characterized by using the tblastx protocol on the NCBI Network Blast Server to test for significant (< 0.05 E-value) similarity to known OR genes. Sequences then were translated, and the appropriate reading frame was characterized based on known conserved regions. Pseudogenes were identified by the presence of a stop codon within the reading frame.

Total Number of OR Genes Identified and Possible Species-Specific Expansion of Certain OR Gene Families

OR genes were successfully amplified in all 7 turtle species. The number of unique OR genes identified ranged from 14 in the green sea turtle to 63 in the painted turtle (Table 1). Notably, fewer (at least half as many) OR genes were identified in the 3 sea turtle species and musk turtle compared with painted turtle, box turtle, and gopher tortoise. Although the number of OR genes identified in this study likely represents only a sampling of the total number of OR genes in each species, a greater number of OR genes amplified by using a given primer set may support an expansion of certain OR gene families in some turtle lineages. Expansion and conservation of genes within a given gene family reflects the functional importance of that gene family for an organism. Niimura and Nei (2006) contend that the number of OR genes in an animal's genome correlates with the number of distinct odors it can detect. Steiger et al. (2008) estimated the number of OR genes in birds to range from 100 to well over 600, depending on the species. It would be reasonable to expect a similar species-specific expansion of particular OR gene families in turtles with more chemically complex habitats. The larger number of OR genes identified in the aquatic painted turtle and terrestrial box turtle and gopher tortoise supports the expansion of certain OR gene families in these species.

Table 1. Turtle olfactory receptor genes and pseudogenes identified.

View Fullscreen

Proportion of Pseudogenes Characterized and Species-Specific Loss of Functional OR Genes

Of the OR genes identified, the percentage of those presumed nonfunctional based on the presence of stop codons (pseudogenes) ranges from 2.4% in the box turtle to 53% in the leatherback sea turtle (Table 1). The percentage of OR pseudogenes found in a given species correlates with the functional importance of olfaction as a whole or of particular odors in the life history of that species. As detection of certain odorants in the environment becomes less important, perhaps due to reliance on alternative sensory systems, selective pressures that maintain OR genes that code for the detection of those odors relax, mutations accumulate, and the number of pseudogenes increases (Rouquier et al. 2000; Gilad et al. 2004; Niimura and Nei 2006; Kishida et al. 2007). This process also may occur as a result of the historical movement from a complex terrestrial environment to a less chemically complex aquatic environment with fewer odors to detect (Freitag et al. 1998; Rouquier et al. 2000). This study suggests a positive correlation between the loss of functional OR genes and a turtle's increased association with water (Fig. 1). Sea turtles have the greatest percentage of pseudogenes (range, 25%–53%), freshwater turtles (painted and musk turtles) have an intermediate percentage of pseudogenes (range, 9%–27%), and land turtles (gopher tortoise and box turtles) have the lowest percentage of pseudogenes (range, 2.4%–9%). These results support the prediction that a less chemically complex aquatic environment will lead to a reduction in the number of functional OR genes. Inferred differences in the olfactory repertoire among turtle species that share a similar habitat may be explained by differences in feeding strategies.

View Figure

• Download as Powerpoint
• Download Figure

Leatherback and green sea turtles have the highest percentages of pseudogenes identified in this study (52.9% and 42.9%, respectively), whereas, loggerhead sea turtles have 25% pseudogenes. If sea turtles are using their OR system to detect food odors, then feeding strategies may help explain this difference in OR gene repertoire size across sea turtle species. Green turtles are primarily herbivores, grazing on sea grass or algae while passively consuming invertebrates in the vegetation (Balazs 1979; Mortimer 1981; Ernst et al. 1994). Green turtles also have been observed feeding on jellyfish in the open water (Heithaus 2002). Leatherbacks are specialists that primarily eat jellyfish (Ernst et al. 1994; Davenport 1998; Constantino and Salmon 2003). In comparison, loggerheads actively hunt a wide variety of marine invertebrates (including molluscs, crustaceans, jellyfish, and sponges) and small fish (Limpus 1973; Tomas et al. 2001). Perhaps a more active feeding strategy and the wider range of food items taken by the loggerhead necessitates a larger repertoire of functional OR receptors.

Musk turtles have a relatively low percentage of OR pseudogenes (9.1%), which indicates selection for maintaining a large percentage of their OR genes. Musk turtles have been shown to heavily rely on odor cues to mediate social behaviors (Fadool et al. 2001) and actively forage for a wide range of prey and vegetation outside of the water (Mahmoud 1968; Ernst et al. 1994; Ford and Moll 2004). Painted turtles have a pseudogene percentage similar to that of the loggerhead sea turtle (Table 1). Because painted turtles forage primarily in the water and rely heavily on vision to find prey (Ernst et al. 1994; Rowe and Parsons 2000), they may have a reduced reliance on their olfactory systems compared with the musk turtle. The lowest percentage of OR pseudogenes found in the turtles studied were in the 2 terrestrial species, gopher tortoise (8.9%) and box turtle (2.4%). Gopher tortoises feed on a variety of plants and have been found to be fairly selective in the plants that they will eat (MacDonald and Mushinsky 1988; Ernst et al. 1994). Box turtles are omnivorous, taking a large variety of plants and animals from both the water and land (Klimstra and Newsome 1960; Strang 1983). Finding food, for both species, may involve the detection of plant volatiles not important for carnivorous or aquatic turtles. The wider range of habitats encountered and food items taken by the box turtle may necessitate a greater diversity of odor receptors compared with all other turtles studied.

Through the use of a noninvasive molecular strategy for predicting the size of the olfactory repertoire in several turtle species, one can better understand the relative importance of different sensory systems in the life history of these animals. An increased understanding of chemosensory systems may allow for the development of better strategies for conserving these species, particularly for those that use olfaction to choose nesting sites and locate and/or recognize food sources, which may be of particular importance in protecting sea turtles that ingest longline bait, such as loggerhead and olive ridley (Lepidochelys olivacea; Swimmer et al. 2005; Southwood et al. 2007) and those that imprint on natal beach odors to aid in natal homing (Owens et al. 1982; Grassman and Owens 1987). A future study of the OR genes in the 2 closely related Lepidochelys species (olive ridley and Kemp's ridley) might be of particular interest for inferring the evolution of olfactory repertoires. More study is needed to further characterize the olfactory and vomeronasal systems of turtles, determine the relative importance of each in detecting waterborne and volatile odors, and identify the particular odors that are most important in directing turtle behaviors.