Some species of butterflyfish have had preyed upon corals for millions of years, yet the mechanism of butterflyfish specialized coral feeding strategy remains poorly understood. Certain butterflyfish have the ability to feed on allelochemically rich soft corals, e.g. Sinularia maxima. Cytochrome P450 (CYP) is the predominant enzyme system responsible for the detoxification of dietary allelochemicals. CYP2-like and CYP3A-like content have been associated with butterflyfish that preferentially consumes allelochemically rich soft corals. To investigate the role of butterflyfish CYP2 and CYP3A enzymes in dietary preference, we conducted oral feeding experiments using homogenates of S. maxima and a toxin isolated from the coral in four species of butterflyfish with different feeding strategies. After oral exposure to the S. maxima toxin 5-episinulaptolide (5ESL), which is not normally encountered in the Hawaiian butterflyfish diet, an endemic specialist, Chaetodon multicinctus experienced 100% mortality compared to a generalist, Chaetodon auriga, which had significantly more (3–6 fold higher) CYP3A-like basal content and catalytic activity. The specialist, Chaetodon unimaculatus, which preferentially feed on S. maxima in Guam, but not in Hawaii, had 100% survival, a significant induction of 8–12 fold CYP3A-like content, and an increased ability (2-fold) to metabolize 5ESL over other species. Computer modeling data of CYP3A4 with 5ESL were consistent with microsomal transformation of 5ESL to a C15-16 epoxide from livers of C. unimaculatus. Epoxide formation correlated with CYP3A-like content, catalytic activity, induction, and NADPH-dependent metabolism of 5ESL. These results suggest a potentially important role for the CYP3A family in butterflyfish-coral diet selection through allelochemical detoxification.
Citation: Maldonado A, Lavado R, Knuston S, Slattery M, Ankisetty S, Goldstone JV, et al. (2016) Biochemical Mechanisms for Geographical Adaptations to Novel Toxin Exposures in Butterflyfish. PLoS ONE 11(5): e0154208. https://doi.org/10.1371/journal.pone.0154208
Editor: Ilaria Corsi, University of Siena, ITALY
Received: September 17, 2015; Accepted: April 11, 2016; Published: May 3, 2016
This is an open access article, free of all copyright, and may be freely reproduced, distributed, transmitted, modified, built upon, or otherwise used by anyone for any lawful purpose. The work is made available under the Creative Commons CC0 public domain dedication.
Data Availability: All relevant data are within the paper and its Supporting Information files.
Funding: This work received support from the following sources: Resource Allocation Program of the Agricultural Research Station for UCR to DS; Summer funding by Hilda and George Liebig Environmental Sciences Summer Fellowship and travel grant Albert March Environmental Sciences Scholarship from the University of California, Riverside; and the Chemistry and DMPK CORE of COBRE, P20GM104932 from the National Institute of General Medical Sciences (NIGMS), a component of the National Institutes of Health (NIH) supported the chemistry studies. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interest exist.
Charles Darwin, in the Origin of Species, was the first to emphasize the importance of interspecific interactions in driving the diversification of life. In spite of the long-recognized importance of such interactions, little is known about the diverse system that governs interspecific interactions. The pairwise reciprocal coevolutionary scenario can predict the development of a chemical “arms race”, in which reciprocity between chemical defenses and counter-adaptive detoxification takes place. In marine systems, however, knowledge of genes conferring consumer tolerance to dietary allelochemicals is in its infancy, when compared to that of terrestrial systems [1–3]. Butterflyfish-coral interactions are an ideal model to study consumer tolerance in marine ecosystems because they share many of the same characteristics as plant-insect interactions. Plants and corals are sessile creatures producing allelochemicals to prevent predation [4,5]. Just as insects have preyed upon plant hosts for 400 million years, butterflyfish have likely preyed upon corals over the past 50 million years [6,7]. Coral consumption (corallivory) is a unique adaptation as only 128 fish species eat corals, out of the 5000 or more fish species recorded from coral reefs, and 61% belong to a single-family, the butterflyfish (f. Chaetodontidae) [6,8]. Despite being one of the most intensively studied families of reef fishes, the evolution of coral feeding remains poorly understood .
Among corallivorous butterflyfish, sympatric species often exhibit highly contrasting levels of dietary specialization (S1 Table) . Specialist species tend to have a narrow dietary range, while generalist species are able to thrive on a widely varied diet . Generalists, such as Chaetodon auriga, consume species from polychaete worms to hard corals , are hypothesized to have liver enzymes that can act on a broad range of substrates to facilitate the biotransformation of a wide variety of toxins. In contrast, generalists consume small amounts of a variety of toxins that are processed through a diverse set of detoxification pathways without overloading any one pathway. This hypothesis has become firmly entrenched in the ecological literature to the extent that it is accepted as the predominant factor regulating the foraging ecology of generalist herbivores e.g. . However, there have been relatively few empirical tests to confirm this hypothesis [13–15]. Specialists, such as Chaetodon multicinctus, consume three species of corals , limiting their range of available food and potentially increasing their capacity to detoxify specific toxins in high concentrations.
Specialists are thought to have evolved novel liver enzymes with greater specificity [17,18,19–22]. For example Chaetodon capistratus preferentially feeds on allelochemically-rich gorgonian corals, while C. striatus and C. ocellatus do not feed on gorgonians. C. capistratus possessed 2- to 3-fold more total CYP and significantly more CYP2B-like and CYP3A4-like proteins when compared to C. striatus and C. ocellatus . These results suggest that CYP2 and CYP3 may be involved in allelochemical biotransformation conferring potential feeding advantages . Similarly, phylogenetic evaluations of hepatic CYP1A in Australian butterflyfish indicated grouping according to feeding strategies . Although chemical arms race dogma would suggest consistent benefit to animals with elevated detoxification enzymes, some studies have shown that organisms that detoxify toxins associated with their unique diet may be susceptible to novel toxins. For example, when the woodrats Neotoma stephensi (specialists) and Neotoma albigula (generalists) were exposed to a novel dietary toxin from Larrea tridentate, fitness metrics of the specialists were more negatively impacted than the generalists .
Pre-exposure to allelochemicals has been shown to induce biochemical defenses in insects, and specialists have also been shown to have unique CYPs that allow detoxification [19–22]. However, no studies have directly investigated how butterflyfish specialists survive on diets containing novel coral toxins compared to that of generalists fed the same coral toxin. This study investigated the biochemical ability of butterflyfish with different feeding strategies to consume a novel Alcyoniidae soft coral (Sinularia maxima) prey, which has been shown to deter feeding in several species of fish. The following questions were asked: (i) whether S. maxima or the isolated allelochemical from S. maxima 5-episinuleptolide (5ESL) can deter feeding in butterflyfish with different feeding strategies, (ii) whether S. maxima or 5ESL can induce CYP in butterflyfish with different feeding strategies, and (iii) whether butterflyfish liver microsomes could biotransform 5ESL into metabolites predicted from CYP(s) that were uniquely induced by 5ESL or S. maxima extracts.
Compliance with Ethical Standards
The author and co-authors of this paper have acted ethically in conducting the described research, having undertaken careful analysis of data and the submitted manuscript to avoid errors. Animals in this study were treated humanely, with protocols overseen and approved under the University of Hawaii at Manoa, Institutional Animal Care & Use Committee (IACUC), protocol # 13–1701.
Analytical grade methanol, ethanol and acetonitrile were purchased from Fisher (Pittsburg, PA). Glycerol, Tris(hydroxymethyl)aminomethane and potassium chloride were also purchased from Fisher (Pittsburg, PA). 14C-Testosterone (150 μCi/μmol; 97.6% purity) was purchased from Perkin-Elmer (Waltham, MA). MS222, EDTA, gelatin and NAPDH were purchased from Sigma-Aldrich (St. Louis, MO). Tween was purchased from EMD Millipore (Billerica, MA).
5,8-epoxy-11-hydroxy-18-nor-3,6-dioxo-12-cembradien-15-methyloxirane-20,10-olide (5ESLO) was synthesized directly from 5-episinuleptilide (5ESL). Sodium bicarbonate was added to a solution of 5ESL (10 mg; 0.003 mmol) in 2 mL of CH2Cl2, 3.0 mg (0.032 mmol) and the mixture was cooled to 0°C. A solution of meta-chloroperbenzoic acid (7.5 mg; 0.032 mmol; 70%) in 1.0 mL of CH2Cl2 was added slowly over a period of 5 minutes. The resulting slurry was stirred for 30 minutes at 0°C, and then stirred at ambient temperature for 3 hrs, as the progress of the reaction was monitored by TLC (5% CH3OH/ 95% CH2Cl2; ninhydrin staining) and LC-MS (MNa+ = 387). The reaction mixture was diluted with an additional 5.0 mL of CH2Cl2 and was washed twice with saturated sodium carbonate solution (10 mL x2). The organic layer was washed twice with saturated Na2S2O3 solution, dried over anhydrous sodium sulfate, filtered, and concentrated to give 11 mg of crude epoxide product. Purification was accomplished using silica gel and a mobile phase consisting of 5% CH3OH:95% CH2Cl2. Fraction 1 contained unreacted starting material (2.0 mg) and fraction 2 (5.0 mg) corresponded to the epoxide product. 1H NMR (400 MHz, CDCl3): Δ 6.64 (dd, J = 13.1, 5.6 Hz, 1H), 4.69 (s, 1H), 4.63 (d, J = 7.9 Hz, 1H), 3.89 (d, J = 10.1 Hz, 1H), 3.45 (d, J = 6.8 Hz, 1H), 3.26 (d, J = 6.6 Hz, 1H), 2.49 (dt, J = 17.3, 8.7 Hz, 1H), 2.38 (s, 1H), 2.20 (s, 2H), 2.04 (d, J = 12.7 Hz, 1H), 1.91 (d, J = 3.2 Hz, 1H), 1.51 (s, 1H), 1.39 (s, 1H), 1.25–1.13 (m, 7H), 0.90–0.72 (m, 3H). LC-MS (ESI+) m/z (MNa+) 387.3.
Butterflyfish C. auriga (generalist), C. unimaculatus (Montipora spp. hard coral specialist in Hawaii and S. maxima soft coral specialist in Guam), C. multicinctus (Porites spp hard coral. specialist and Hawaiian endemic) and C. kleinii (facultative coral and plankton feeder) were chosen for this experiment because of their diverse feeding preferences and abundance in Hawaii. Butterflyfish were collected from Kaneohe Bay (C. auriga and C. unimaculatus) and Yokahama Bay (C. multicinctus and C. kleinii) reef systems, surrounding the island of Oahu during June 2012 and 2013. Both locations required no specific permissions, as it is legal in the state of Hawaii to collect fish in Kaneohe Bay and Yokahama Bay. Butterflyfish species collected in this experiment were not endangered or protected. Professional fish collectors caught butterflyfish with nets. C. multicinctus (11 ±3 g; 8 ± 0.2 mm), C. kleinii (23 ± 7 g; 9 ± 1 mm), C. auriga (43 ± 8 g; 10 ± 1 mm) and C. unimaculatus (10.4 ±6 g; 7.2 ±1.7 mm) were acclimated for two weeks before treatments. Fish were juveniles and sexually immature by visual determination of the gonads after dissection. Fish were held at the Waikiki Aquarium in a large (15 ft diameter and 3 ft height) flow-through tank, with floating cages for separation of doses. S. maxima and S. polydactyla were collected from Piti Bomb holes, Guam and identified by M. Slattery .
5-ESL was isolated in the same manner as Kamel et al. . In summary, S. maxima was initially frozen on dry ice, thoroughly extracted with methanol: dichloromethane (1:1), concentrated under reduced pressure, and subjected to silica gel vacuum liquid chromatography. The column was eluted with hexane, hexane–ethyl acetate, ethyl acetate–methanol to methanol to yield 11 fractions, which were concentrated under reduced pressure. 5-ESL was eluted with 80% ethyl acetate:hexane, recrysallized and washed successively with chloroform and methanol.
To examine the effect of S. maxima tissue homogenate and 5ESL on survival and hepatic CYP expression, 4–6 fish of each species and each dose were gavaged with either high doses (250mg/kg) or low doses (50mg/kg) of S. maxima tissue homogenate (Summer 2010), or high dose (3.0mg/kg), or low dose (1.0mg/kg) of 5ESL (Summer 2011). 1M Tris-buffer (pH 8.2) served as a negative control. Doses of the compound which is found in Sinularia were based on average daily consumption rates of other prey items by C. multicinctus, C. auriga, C. unimaculatus and C. kleinii e.g. [12,16]. Fish were anesthetized with waterborne MS-222 (0.1 g/L) for 10 min during gavage to reduce stress. Treatments occurred with gavage quickly and efficiently to reduce stress on days 1, 3 and 5. Prior and during treatments fish were monitored bidaily for survival, health and parasites. Fish were euthanized with MS-222 during the experiment if they were immobile or severely floundering. At the end of the treatment period (day 7) fish were given an anesthetic waterborne overdose of MS-222 (10 g/L). After approximately 10–20 minutes of fish gills not moving, fish were removed and dissected. There were no unexpected deaths during this experiment. Livers were removed, frozen and maintained in -80°C until microsomes were prepared for immunoblot and catalytic activity analyses. All experimental manipulations were reviewed and approved by IACUC, protocol # 13–1701.
Preparation of Microsomal Fractions
Microsomal fractions were prepared as described in Maldonado et al. . Proteins were measured by the Coomassie Blue method using a commercial kit (Pierce Inc., Rockford, IL) with bovine serum albumin as a standard.
CYP3A-like and CYP2-like protein levels were determined by Western blot as described in Maldonado et al. . The membrane was probed with either 1:500 dilution (v/v) of primary rabbit anti-rainbow trout polyclonal CYP2K1 and/or CYP2M1 or 1:1000 dilution (v/v) of primary rabbit anti-rainbow trout polyclonal CYP3A27 antibodies provided by Dr. Malin Celander University of Gothenburg and Dr. Don Buhler, Oregon State University. Imager ChemiDoc XRS+ Imaging System (BioRad) image analyzer. The data were presented in optical density units per mg protein.
14C Testosterone Hydroxylase Activity
14C Testosterone hydroxylase activity was measured as described by Maldonado et al. . The metabolites were quantified by integrating the area under the radioactive peaks (recovery 98.9%-99.7%); the detection limit was 0.2 pmol/min/mg protein.
5-Episinuleptolide Metabolism and Metabolite Identification
Incubations contained 100 μg of hepatic microsomal protein with 5ESL from 1 to 30 μM and included 7.5 μM NADPH in a final volume of 0.204 mL of 50 mM Tris–HCl pH 7.4. Samples were incubated for 45 min at 30°C. Incubation reactions were stopped with 100 μl of acetonitrile, followed by centrifugation at 10,000 g for 10 min. Then 1 μM of the internal standard caffeine was added, and finally 40 μl of supernatant was injected onto a reverse-phase HPLC column. HPLC analyses were performed on an SCL-10AVP Shimadzu HPLC system equipped with a 250Å~4.6 mm Atlantis C18 (5 μm) reverse-phase column (Waters, Milford, MA). Separation of 5ESL metabolites was attained using mobile phase systems of (A) 95% water and 5% acetonitrile and (B) 100% acetonitrile, at a rate of 1 ml/min. The gradient started with 100% A to 50% B in 40–45 min. Chromatographic peaks were monitored with an SPD-10A VP SHIMADZU UV-VS detector (SHIMADZU Co., Carlsbad, CA) at wavelength 249 nm. 5ESL 15,16 oxide and the internal standard were quantified by integrating the area under the peaks using a 6 point standard curve, with a limit of detection of 5 pmol/min/mg protein.
C. unimaculatus samples were selected for further metabolite characterization by LCMS/MS in positive electrospray ionization mode using a Agilent 6460 triple quadrupole mass spectrometer at gas temperature 300°C and capillary voltage of 2500V. A 1 μL aliquot of metabolite was loaded onto an Agilent Poroshell 120 SB- C18 column (2.1mm x 50mm, 2.7 μm particle size, Santa Clara, CA) with a mobile phase flow rate of 0.2 mLmin−1. Samples were eluted with a gradient composed of 5 mM ammonium acetate, water containing 0.1% formic acid (A) and acetonitrile containing 0.1% formic acid (B). The linear gradient transitioned from a 5% solvent B to 50% for the first 15min, to a gradient of 50% B over 0.2min. Then the initial solvent (5% B) was restored over a 1min linear gradient and re-equilibrated. The ion trap was programmed to collect a MS2 spectrum from 100 to 1000 m/z, a fragment voltage of 20V and a cell acceleration voltage of 7V from collision- induced dissociation of the most intense target ion from the appropriate product ion spectrum. MS data was analyzed with Agilent Mass Hunter (Agilent Technologies, Santa Clara, CA).
Ligands and crystal structures were prepared for docking using AutoDockTools v1.5.4; , with the addition of polar hydrogens and the assignment of Gasteiger charges. Ligand structure and charge minimization was performed with semi-empirical methods (PM6 Hamiltonian in Mopac2009) . Gasteiger-Marsili partial charges were used in the final docking runs. High throughput docking was performed using Autodock Vina (v1.1.2). Flexible ligands were docked into models with rigid protein backbones and rigid side chains. 100 replicate dockings were performed, retaining a broad range of calculated energies (6 kcal/mol). Docking visualization was performed using PyMol (v. 22.214.171.124, Schrödinger LLC, Portland, OR).
Statistical analyses were conducted using Prism5 v5.0a software. For the feeding deterrence assay, since the control pellets were typically all consumed (i.e., zero pellets rejected), Fisher’s Exact Test was the most appropriate statistical analysis for this dataset. Prior to statistical analysis, all data was analyzed to meet the normality and variance assumptions of parametric tests. For normally distributed data, an initial one-way ANOVA analysis was carried out to evaluate the differences between doses, not between treatments. If a P-value less than 0.05 was observed, it was considered statistically significant. If significance was determined, the Tukey’s multiple range test was performed to determine differences between groups. If data did not meet assumptions of the parametric test, a Kruskal–Wallis test and two-tailed multiple comparisons Dunn's test was used.
S. maxima Oral Exposure
Survival, hepatic CYP3A-like content, CYPA catalytic activity, CYP2-like content, and CYP2 catalytic activity was measured in 4 species of butterflyfish after oral gavage treatments with of S. maxima. The four species of butterflyfish included C. auriga (generalist), C. kleinii (Facultative coral feeder), C. multicinctus (specialist), and C. unimaculatus (specialist). Oral exposure to soft coral S. maxima tissue homogenate had no significant effect on survival. In C. kleinii, oral treatment with the S. maxima tissue homogenate caused a statistically significant dose dependent reduction in CYP3A-like content and catalytic activity (16ß hydroxytestosterone formation) (P ≤ 0.05) (Fig 1A and 1C), and a statistically significant dose dependent reduction in CYP2 catalytic activity (16a-hydroxytestosterone formation) (P ≤ 0.05) (Fig 2C). The formation of 6ß-hydroxytestosterone in C. auriga was also induced by S. maxima at the high dose by a factor of 2.3 (Fig 1B). C. auriga’s CYP2K-like content and 16a-hydroxytestosterone formation in C. auiria were significantly induced by S. maxima tissue homogenate at the high dose with a 2.33-fold change in content and a 3.2-fold change in catalytic activity (P ≤ 0.05) (Fig 2A and 2C). Oral exposure to S. maxima caused a significant 1.5-fold increase in C. multicinctus CYP2-like activity (16a-hydroxytestosterone) at the high dose (Fig 2C). C. multicinctus 6ß-hydroxytestosterone formation (CYP3A activity) was significantly decreased by the high dose 2.5-fold (Fig 1B). Oral treatment of fish with S. maxima tissue homogenate caused a significant induction in the formation of 6ß, 16ß and 16a-hydroxytestosterone (10–20 fold) (Figs 1B,1C and 2C) and CYP3A-like and CYP2-like content (5–11 fold) in C. unimaculatus (Fig 1A), but caused no significant changes in C. auriga. In C. unimaculatus, CYP3A catalytic activity (16ß and 6ß hydroxytestosterone formation) was induced 4-fold in protein and 2-fold in activity (P ≤ 0.05) (Fig 2A–2C).
Different letters indicate significant (P ≤ 0.05) differences between untreated and treated fish (N = 6–8).
5ESL Oral Exposure
Survival, hepatic CYP3A-like content, CYPA catalytic activity, CYP2-like content, and CYP2 catalytic activity was also measured in the same 4 species of butterflyfish after gavage treatment of 5ESL.During the oral 5ESL treatments, C. multicinctus experienced 100% mortality in both the high and the low doses, with 100% survival in the control (Fig 3). In contrast, 80% survival was observed in C. auriga, with 45–60% survival observed for C. kleinii after treatment (Fig 4). C. unimaculatus had 100% survival in all treatments (Fig 3). 5ESL caused 1.3-fold induction of CYP3A-like content in C. auriga, (P ≤ 0.05) but a reduction by a factor of 7.2 in CYP3A catalytic activity (6ß hydroxytestosterone) (P ≤ 0.05) (Fig 4A and 4C). 5ESL also caused a reduction in CYP2 catalytic activity by 0.5-fold at the high dose (P ≤ 0.05) (Fig 5B). In C. kleinii, no statistically significant changes in content or catalytic activity were observed after exposure to 5ESL. Due to excessive mortality, liver microsomes from only the C. multicinctus control group were evaluated for CYP activity and expression. Hepatic CYP3A-like content from C. unimaculatus had the most significant dose dependent induction following 5ESL treatment compared to the other species, with 8–12 fold times larger values observed after treatment (P ≤ 0.05) (Fig 4A). The induction of CYP content in C. unimaculatus correlated with catalytic activity of CYP3A, increasing 20–25 fold (R2 = 0.66728) (Fig 4B and 4C).
Each value represents the total survival of 6–8 individuals.
Different letters indicate significant (P ≤ 0.05) differences between untreated and treated fish (N = 6–8).
5ESL Docking Experiment
5ESL was docked into a crystal structure of CYP3A4, (PDB:3NXU) with the ligand removed, to assess the ability of CYP3A4 to metabolize 5ESL. Several low energy (high affinity) positions were found well within the oxidizable distance to a computed Fe-O position. In general, two main docking areas were found—one with an oxidizable carbon within 2–3 A of the Fe-bound oxygen, and the other at about a 7–8 A distant. Both could be occupied at the same time, as the Hill coefficient for CYP3A4 is usually 2 (that is, the reaction rate of CYP3A4 with substrate is dependent on the square of the substrate concentration [S] in Michaelis-Menten type kinetics, implying double occupancy of the active site). There were four different carbons (C16, C12, C10, C19) that could be oxidized depending on preferred ligand orientation in the active site, one of which was an epoxide formed on the C15-16 (Fig 6B).
5ESL Microsomal Incubations
To investigate the metabolism of 5ESL, liver microsomes of C. auriga, C. unimaculatus, C. kleinii, C. multicinctus and human CYP3A4 supersomes were incubated with 5ESL and the metabolites were identified. Microsomal in vitro incubations with 5ESL (Fig 6A) showed NADPH-catalyzed clearance in C. auriga, C. unimaculatus, C. multicinctus and human CYP3A4 supersomes (Table 1). Enzyme saturation as determined by Michaels-Menten kinetics occurred at 11 μM. An increased trend for biotransformation of 5ESL with NADPH was observed in the following species: in increasing order—C. multicinctus < C. kleinii < C. auriga < C. unimaculatus; mimicking the survival trend. Human CYP3A4 completely metabolized 5ESL when incubated with NADPH after 45 min (Table 1). There were three significant peaks at retention times of 2, 3 and 4.5 minutes in the HPLC chromatogram. We further analyzed the peak at retention time 3 min using LCMS/MS and compared the chromatogram to the synthesized 5ESL 15–16 epoxide (5ESLO) (Fig 6B) based on docking data. A 365 m/z molecular ion was detected with the metabolite at 3 min and a purified standard of 5ESLO. Loss of 114 and 247 m/z was observed in both compounds from the corresponding pseudomolecular ions. The structures of the remaining metabolites remain unknown.
5ESL Epoxide Formation
After identifying 5ESL15-16 Oxide (5ESLO) as a metabolite, the rates of 5ESLO formation were calculated in liver microsomes of C. auriga, C. kleinii and C. unimaculatus. 5ESLO formation correlated with CYP3A activity, with an R2 = 0.862. The highest 5ESLO formation was seen in microsomes obtained from C. unimaculatus exposed to the high dose of 5ESL relative to untreated C. unimaculatus, C. auriga and C. kleinii (P ≤ 0.05) (Table 1). Formation of 5ESLO was below the limit of detection in microsomes from C. multicinctus.
Previous studies have found that generalist feeding is deterred by allelochemicals, whereas specialists are either unaffected or attracted to the allelochemicals of their preferred diet [29,30]. The capability of S. maxima to deter predators under laboratory and field conditions has been well described e.g. [23,31]. However, few studies have looked at the effect of different feeding strategies within the same genus or species on deterrence. Homogenates of S.maxima and the individual compound, 5ESL have been shown to be deterrent to generalists but not specialists (Slattery et al., submitted). Deterrence in the generalists may indicate possible toxicity by learned avoidance in the wild or distastefulness during exposure [32–35]. In contrast, the lack of deterrence in specialist species of butterflyfish may indicate an evolved adaptation to the toxins in the soft coral S. maxima. C. unimaculatus preferentially feed on the soft coral S. maximus in Guam  but not in Hawaii where the coral does not occur .
In Hawaii, exposure to homogenates of the soft coral S. maxima had no significant effect on survival or noteworthy effects on CYP2 or CYP3A-like proteins or activities on all 4 species. In contrast, 5ESL treatment caused significant mortality in C. multicinctus and C. kleinii. However, survival in C. auriga following exposure to 5ESL was unaffected. Since survival correlated with basal CYP3A-like and CYP2-like content and activities, CYP3A and CYP2 of C. auriga may function as multipurpose detoxification enzymes of dietary chemicals. If the fish is capable of detoxifying a broad range of lipophilic compounds, this species of fish may exploit a larger range of prey as a generalist feeder [13–15,37]. Conversely, C. multicinctus, with a limited range of available food may have CYP’s with greater specificity [21,22,38]. A potential negative effect of this trait is that the specialists are less energetically and/or mechanistically efficient at eliminating novel toxins that are not associated with their specific prey . Consistent with this hypothesis, C. multicinctus had the lowest survival of all four species and the lowest CYP2-like and CYP3A-like basal content and activities. An example of specialist herbivores being negatively impacted by novel defense compounds is the woodrat, Neotoma stephensi that had reduced locomotor activity when exposed to novel plant toxins from Larrea tridentata [14,19].
Since there were multiple CYP2-like isoforms observed in western blots, and catalytic activities associated with CYP2 (16α-testosterone hydroxylase) were not correlated with survival, focus was placed upon CYP3A as a potential enzyme for detoxification. 5ESL docking into the crystal structure of CYP3A4 showed several low energy (high affinity) positions where four different carbons could be oxidized, depending on preferred ligand orientation in the active site. CYP3A enzymes are functionally among the most versatile forms of CYPs. Mutation and docking studies have demonstrated that CYP3A proteins have a large substrate-binding pocket in comparison to other members of the CYP superfamily . This large pocket may enable CYP3A enzymes to catalyze biotransformation of the large 5ESL molecule. Further analysis revealed the C15-16 epoxide (5ESLO) (Fig 6B) to be the most likely CYP3A metabolite. The 5ESLO metabolite in microsomal in vitro incubations showed NADPH dependent formation of a C15-16 epoxide confirmed by LCMS with a synthesized standard. In addition, formation of the epoxide correlated with hepatic CYP3A-like content and survival in all species treated with 5ESL. This is consistent with the possibility that epoxide formation of 5ESL by CYP3A is a detoxification pathway.
The corresponding pattern of CYP3A-like content induction and catalytic activity in the liver enzymes of C. unimaculatus dosed with 5ESL is similar to cases in plant-insect interactions, where black swallowtails (Papilio polyxenes) where exposed to the plant alleochemical, xanthotoxin that induced CYP activity in a dose-dependent manner, increasing seven fold . Allelochemicals often act as ligands to induce their own CYP metabolism, [19–22]. Although it is unclear how 5ESL regulates CYP3A, other studies have demonstrated that CYP3A can be induced by allelochemicals that bind the nuclear pregnane X receptor (PXR) [39–41]. Further research is needed to elucidate this mechanism.
C. unimaculatus had the greatest clearance of 5ESL, the highest induction of CYP3A-like content as well as activity, and the highest relative formation of 5ESLO. C. unimaculatus preferentially feeds on S. maxima in Guam and Australia, suggesting the fish has a high tolerance for soft coral toxins regardless of current feeding preferences. Induction of CYP3A after exposure to a dietary item that is geographically removed from Hawaii suggests that C. unimaculatus may have migrated from Guam (where the fish feeds on the soft coral) to Hawaii. The Biodiversity Feedback model indicates that biodiversity flows eastward from the indo-pacific archipelago to peripheral habitats in Hawaii . The Central Pacific area is characterized by low species departure (10–16% Central Pacific), but high dispersal of lineages into the region, from the Indo-Pacific Archipelago . In terms of global diversity for Chaetodontidae, the Indo-Pacific Archipelago stands out as a significant source of diversity in terms of both origination within the region and the expansion of lineages into adjacent regions . Gene flow of two related coral-feeding butterflyfish was found to be high, across the Pacific Ocean over both recent and historical timeframes, in the specialist, Chaetodon trifascialis and more stable in the generalist, C. lunulatus . It is possible that C. unimaculatus has gene flow between Hawaii and the rest of the Pacific or historically migrated from Guam eastward to Hawaii. Similar hypotheses have been proposed for butterflyfish from other areas of the Indo-Pacific  as well as other fish species . Future phylogenetic studies on C. unimaculatus should be conducted to support or refute the evolutionary migration from Guam to Hawaii.
C. multicinctus had the least metabolism of all species tested, and it is endemic to Hawaii, with no evolutionary history with S. maximus. The distinct levels of allelochemical tolerance may explain coral-feeding butterflyfish dietary specialization, since nutritional value of corals has yet to explain these differences . The risks associated with dietary specialization are thought to be offset by increased nutritional benefits (e.g growth) when feeding on preferred prey over generalist feeding on the same prey . However, a study conducted on coral-feeding butterflyfish found that a more specialized species (Chaetodon trifascialis) did not outperform the generalist species (Chaetodon plebeius) when both consumed their preferred prey . In the current study, C. unimaculatus appears to have an innate ability to tolerate toxic dietary allelochemicals allowing them to explore uncompetitive diets of soft corals throughout the Pacific . Losing this advantage may have led C. unimaculatus to consume hard coral Monipora spp. in Hawaii . This may also explain the innate inability of C. multicinctus to tolerate the soft coral toxin 5ESL, and its preference for hard corals. These differences in dietary choices may be explained by their characteristic ability to tolerate dietary allelochemicals.
In conclusion, the use of a toxicological approach has resulted in the identification of butterflyfish cytochrome P450s, specifically CYP3A, as a potential detoxification enzyme. We provide several lines of evidence that implicate inducible butterflyfish CYP3As in the mechanism of adaptation to soft coral allelochemicals. The corresponding pattern of protein content responsiveness and testosterone hydroxylase activity in the liver microsomes of butterflyfish orally exposed to 5ESL coupled with in vitro incubations, computer modeling data, and the formation of 5ESL epoxide metabolite strongly indicates a role for the Chaetodon CYP3As in mediating the metabolism of dietary norcembranoids. The results of our study also demonstrate that functional versatility of counter-defense genes may facilitate generalist (C. auriga), whereas functional specialization of C. multicinctus genes may hinder the ability to detoxify novel toxins. Other genetic and molecular factors may contribute to generalist ability to detoxify novel toxins. Comparative gene sequencing surveys of host use-related genes between pairs of closely related butterflyfish species that diverge in their degree of feeding specialization may resolve the question of whether multiple gene duplication events also aid generalists in coping with their diverse and unpredictable coral defense challenges. This work demonstrates the utility of incorporating molecular approaches to better understand the biochemical novelties that allow marine consumers to detoxify allelochemically defended prey. Evolution of host range has been a central interest in the field of butterflyfish ecology; however, little evidence to support theories proposed to explain macroevolutionary patterns of host use and dietary breadth leave a gap in the science that requires more multidisciplinary involvement . Moreover, identifying the molecular foundations of organismal detoxification has broad implications for understanding the role of species-species interactions on gene function.
We thank Dr. Andrew Rossiter and the staff of the Waikiki Aquarium for providing laboratory facilities and technical expertise. I would like to thank Dr. Malin Celander (University of Gothenburg) and Dr. Donald R. Buhler (Oregon State University) for the provision of anti-CYP antibodies. We are grateful for the support of undergraduates Bronte Crook, Jenna Harlacher, Danielle Garcia, Stephen Fallon and Chelsea Wolke in fish care.
Conceived and designed the experiments: AM DS. Performed the experiments: AM RL SK MS SA JG RG. Analyzed the data: AM RL SK MS SA JG RG. Contributed reagents/materials/analysis tools: MS KW EH JR GO DS. Wrote the paper: AM.
- 1. DeBusk BC, Slattery M, Ki J, Lee J, Aparicio-Fabre R, Schlenk D. Species differences and effects of soft coral extracts from Sinnularia maxima on the expression of cytochrome P4501A and 2N in butterflyfishes (Chaetodon spp.). Fish Physiol Biochem. 2008;34: 483–492. pmid:18958605
- 2. Sotka EE, Whalen KE. Herbivore offense in the sea: the detoxification and transport of secondary metabolites. Algal chemical ecology. Springer Berlin Heidelberg, 2008: 203–228.
- 3. Sotka EE, Forbey J, Horn M, Poore A, Raubenheimer D, Whalen KE. The emerging role of pharmacology in understanding consumer-prey interactions in marine and freshwater systems. Integrative & Comparative Biology 2009;49: 291–313.
- 4. Lages BG, Fleury BG, Pinto AC, Creed JC. Chemical defenses against generalist fish predators and fouling organisms in two invasive ahermatypic corals in the genus Tubastraea. Marine Ecology. 2010;31: 473–482.
- 5. Fritz RS, Simms EL editors. Plant resistance to herbivores and pathogens: ecology, evolution, and genetics. University of Chicago Press. 1992.
- 6. Cole AJ, Pratchett MS, Jones GP. Diversity and functional importance of coral-feeding fishes on tropical coral reefs. Fish and Fisheries. 2008;9: 286–307.
- 7. Labandeira C. The origin of herbivory on land: Initial patterns of plant tissue consumption by arthropods. Insect Science. 2007;14: 259–275.
- 8. Rotjan RD, Lewis SM. Impact of coral predators on tropical reefs. Mar Ecol Prog Ser. 2008;367: 73–91.
- 9. Bellwood DR, Klanten S, Cowman PF, Pratchett MS, Konow N, Van Herwerden L. Evolutionary history of the butterflyfishes (f: Chaetodontidae) and the rise of coral feeding fishes. J of evolutionary biol. 2010;23: 335–349.
- 10. Berumen ML, Pratchett MS. Trade-offs associated with dietary specialization in corallivorous butterflyfishes (Chaetodontidae: Chaetodon). Behav Ecol Sociobiol. 2008;62: 989–994.
- 11. Krebs JR, Davies NB. Behavioural ecology: an evolutionary approach. John Wiley & Sons. 2009.
- 12. Pratchett MS. Dietary overlap among coral-feeding butterflyfishes (Chaetodontidae) at Lizard Island, northern Great Barrier Reef. Marine Biology. 2005;148: 373–382.
- 13. Sorensen J, Dearing M. Elimination of plant toxins by herbivorous woodrats: revisiting an explanation for dietary specialization in mammalian herbivores. Oecologia. 2003;134: 88–94. pmid:12647185
- 14. Sorensen JS, McLister JD, Dearing MD. Novel plant secondary metabolites impact dietary specialists more than generalists (Neotoma spp). Ecology. 2005;86: 140–154.
- 15. Dearing MD, Cork S. The role of detoxification of plant secondary compounds on diet breadth in a mammalian herbivore, Trichosurus vulpecula. J Chem Ecol. 1999;25: 1205–1219.
- 16. Tricas TC. Determinants of feeding territory size in the corallivorous butterflyfish, Chaetodon multicinctus. Animal Behaviour. 1989;37: 830–841.
- 17. Torregrossa AM, Azzara AV, Dearing MD. Testing the diet-breadth trade-off hypothesis: differential regulation of novel plant secondary compounds by a specialist and a generalist herbivore. Oecologia. 2012;168: 711–718. pmid:21927911
- 18. Vrolijk NH, Targett NM, Woodin BR, Stegman JJ. Comparison of Cytochrome P450 in three butterflyfish species: ecological implications of eleated CYP2 and CYP3A in Chaetodon capistratus. Marine Environmental Research. 1995;39: 11–14.
- 19. Berenbaum MR. North American ethnobotanicals as sources of novel plant-based insecticides. In ACS Symposium series-American Chemical Society (USA). 1989.
- 20. Crawford IP. Toxicity of angular furanocoumarins to swallowtail butterflies: escalation in a coevolutionary arms race? J Clin Invest. 1980;66: 473.
- 21. Cohen MB, Berenbaum MR, Schuler MA. Induction of cytochrome P450-mediated detoxification of xanthotoxin in the black swallowtail. J Chem Ecol. 1989;15: 2347–2355.pmid:24272422
- 22. Li X, Baudry J, Berenbaum MR, Schuler MA. Structural and functional divergence of insect CYP6B proteins: From specialist to generalist cytochrome P450. Proceedings of the National Academy of Sciences of the United States of America. 2004;101: 2939–2944. pmid:14981232
- 23. Kamel HN, Fronczek FR, Khalifa SI, Slattery M. Microbial transformation of 5-episinuleptolide. Chemical and pharmaceutical bulletin. 2007;55: 537–540. pmid:17409543
- 24. Slattery M, Paul VJ. Indirect effects of bleaching on predator deterrence in the tropical Pacific soft coral Sinularia maxima. Marine Ecology-Progress Series. 2008;354: 169.
- 25. Maldonado A, Johnson A, Gochfeld D, Slattery M, Ostrander GK, Bingham J, et al. Hard Coral (Porites lobata) extracts and homarine on Cytochrome P450 expression in Hawaiian butterflyfishes with different feeding strategies. Comparative Biochemistry and Physiology: Part C. 2016;179: 57–63.
- 26. Morris GM, Huey R, Lindstrom W, Sanner MF, Belew RK, Goodsell DS, et al. AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility. Journal of computational chemistry. 2009;30: 2785–2791. pmid:19399780
- 27. Stewart JJ. Application of the PM6 method to modeling the solid state. Journal of molecular modeling. 2008;14: 499–535. pmid:18449579
- 28. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. Journal of computational chemistry. 2010;31: 455–461. pmid:19499576
- 29. Roslin T, Salminen JP. Specialization pays off: contrasting effects of two types of tannins on oak specialist and generalist moth species. Oikos. 2008;117: 1560–1568.
- 30. Becerro MA, Paul VJ, Starmer J. Intracolonial variation in chemical defenses of the sponge Cacospongia sp. and its consequences on generalist fish predators and the specialist nudibranch predator Glossodoris pallida. Marine Ecology Progress Series. 1998;168: 187–196.
- 31. Wylie CR, Paul VJ. Chemical defenses in three species of Sinularia (Coelenterata, Alcyonacea): effects against generalist predators and the butterflyfish Chaetodon unimaculatus Bloch. Journal of Experimental Marine Biology and Ecology. 1989;129: 141–160.
- 32. Berenbaum MR, Zangerl AR, Nitao JK. Constraints on chemical coevolution: Wild parsnips and the parsnip webworm. Evolution 1986;40: 1215–1228.
- 33. Bernays EA. Aversion learning and feeding. In: Papaj DR, Lewis AC, editors. Insect learning. Springer US; 1993. pp. 1–17.
- 34. Long JD, Hay ME. Fishes learn aversions to a nudibranch’s chemical defense. Marine Ecology Progress Series. 2006;307: 199–208.
- 35. Lawton RJ, Messmer V, Pratchett MS, Bay LK. High gene flow across large geographic scales reduced extinction risk for a highly specialized coral feeding butterflyfish. Molecular Eco. 2011;20: 3584–3598.
- 36. Li X, Schuler MA, Berenbaum MR. Molecular mechanisms of metabolic resistance to synthetic and natural xenobiotics. Annu Rev Entomol. 2007;52: 231–253. pmid:16925478
- 37. Khan KK, Halpert JR. Structure–function analysis of human cytochrome P450 3A4 using 7-alkoxycoumarins as active-site probes. Arch Biochem Biophys. 2000;373: 335–345. pmid:10620357
- 38. Cohen MB, Schuler MA, Berenbaum MR. A host-inducible cytochrome P-450 from a host-specific caterpillar: molecular cloning and evolution. Proc Nati Acad Sci. 1992;89: 10920–10924.
- 39. Li X, Berenbaum MR, Schuler MA. Molecular cloning and expression of CYP6B8: a xanthotoxin-inducible cytochrome P450 cDNA from Helicoverpa zea. Insect biochemistry and molecular biology. 2000;30: 75–84. pmid:10646973
- 40. Kullak-Ublick GA, Becker MB. Regulation of drug and bile salt transporters in liver and intestine. Drug Metab Rev. 2003;35: 305–317. pmid:14705863
- 41. Moore LB, Goodwin B, Jones SA, Wisely GB, Serabjit-Singh CJ, Wilson TM, et al. St. John's wort induces hepatic drug metabolism through activation of the pregnane X receptor. Proc Natl Acad Sci. 2000;97: 7500. pmid:10852961
- 42. Cowman PF, Bellwood DR. The historical biogeography of coral reef fishes: global patterns of origination and dispersal. Journal of Biogeography. 2013;40: 209–224.
- 43. Gaither MR, Rocha LA. Origins of species richness in the Indo-Malay-Philipine biodiversity hotspot: evidence for the centre of overlap hypothesis. Journal of Biogeography. 2013; 40:1638–1648.
- 44. Craig MT, Eble JA, Bowen BW, Robertson DR. High genetic connectivity across the Indian and Pacific oceans in the reef fish Myripristis berndti (Holocentridae). Marine Ecology Progress Series 2007;334:245–254.
- 45. Schoener TW. Theory of feeding strategies. Ann Rev Ecolog Sys 1971;2: 369–404.
- 46. LaJeunesse TC, Thornhill DJ, Cox EF, Stanton FG, Fitt WK, Schmidt GW. High diversity and host specificity observed among symbiotic dinoflagellates in reef coral communities from Hawaii. Coral Reefs. 2004;23: 596–603.
- 47. Cox EF. Aspects of corallivory by Chaetodon unimaculatus in Kaneohe Bay, Oahu. Masters Thesis. University of Hawaii at Manoa.
- 48. Pratchett MS, Berumen ML, Kapoor BG. Biology of Butterflyfishes. CRC Press Taylor & Francis Group, Boca Raton, FL. 2014.