Chemoecology (2011) 21:131–141 DOI 10.1007/s00049-011-0075-5 CHEMOECOLOGY RESEARCH PAPER Chemical defense in pelagic octopus paralarvae: Tetrodotoxin alone does not protect individual paralarvae of the greater blue-ringed octopus (Hapalochlaena lunulata) from common reef predators Becky L. Williams • Vanessa Lovenburg Christine L. Huffard • Roy L. Caldwell • Received: 24 February 2011 / Accepted: 1 April 2011 / Published online: 14 April 2011 Ó Springer Basel AG 2011 Abstract Some pelagic marine larvae possess anti-predator chemical defenses. Occasionally, toxic adults imbue their young with their own defensive cocktails. We examined paralarvae of the greater blue-ringed octopus (Hapalochlaena lunulata) for the deadly neurotoxin tetrodotoxin (TTX), and if present, whether TTX conferred protection to individual paralarvae. Paralarvae of H. lunulata possessed 150 ± 17 ng TTX each. These paralarvae appeared distasteful to a variety of fish and stomatopod predators, yet food items spiked with 200 ng TTX were readily consumed by predators. We conclude that TTX alone does not confer individual protection to paralarvae of H. lunulata, and that they possess an alternative defense. In larger doses, tetrodotoxin is a deterrent to the predatory stomatopod Haptosquilla trispinosa (mean dose = 3.97 lg/g). This corresponds to 12–13 paralarvae per predator based on the TTX levels of the clutch we examined. Thus, the basic assumption that individual paralarvae of H. lunulata are defended by TTX alone was disproved. Instead, functionality of TTX levels in paralarvae may arise through alternative selective pathways, such as deterrence to B. L. Williams (&) Department of Biology, MSC 3AF, New Mexico State University, P.O. Box 30001, Las Cruces, NM 88003-8001, USA e-mail: toxwilliams@gmail.com V. Lovenburg
R. L. Caldwell Department of Integrative Biology, University of California, Berkeley, 3060 Valley Life Science Bldg #3140, Berkeley, CA 94720-3140, USA C. L. Huffard Conservation International Indonesia, 17 Jalan Dr. Muwardi, Denpasar, Bali, Indonesia parasites, through kin selection, or against predator species not tested here. Keywords Chemical defense
Cephalopod blue-ringed octopus
Hapalochlaena lunulata
Palatability
Tetrodotoxin
Paralarvae
Pelagic larvae Introduction Planktivorous predators in the nekton have shaped the evolution of anti-predator defenses including morphologies that are difficult to ingest (e.g. spines in crab zoeae; Morgan 1989), transparency that serves as crypsis (e.g. Johnson 2001), and chemical defenses including unpalatability (e.g. Lindquist and Hay 1996; Lindquist et al. 1992; McClintock and Vernon 1990; Young and Bingham 1987). Cephalopods comprise a diverse lineage of benthic and midwater adults and paralarvae alike, with a diverse repertoire of anti-predator defenses. Adults are integral prey to seabirds, sharks, marine mammals, and teleosts (e.g. Jereb et al. 2005), and this demonstrated widespread palatability suggests that planktonic paralarvae may also be at high risk of predation. Because of morphological constraints in cephalopods, paralarvae especially may benefit from chemical defenses. Inking represents one such chemical defense by mollusks such as cephalopods (e.g. Derby 2007). For example, tyrosinase, a cytotoxin present in the ink of several cephalopods, may irritate chemosensory receptors and thus deter predators (Prota et al. 1981; Russo et al. 2003). Paralarvae of many octopuses can expel ink upon hatching and during their tenure in the water column (reviewed by Villanueva and Norman 2008). Immature common blanket octopuses (Tremoctopus violaceus), mid-water animals, 123 132 possess a different and unusual example of chemical defense in octopods; they harvest man-o-war (Physalia) tentacles and present these armaments to predators in defense (Jones 1963). While inking has received some attention, and the co-opting of cnidarian defenses serves as a unique example, little else is known about chemical defenses in pelagic phase cephalopods. Blue-ringed octopuses (Hapalochlaena spp.) represent a well-known example of chemically defended benthic cephalopods, some of which produce pelagic paralarvae. Noxious larvae coincide with noxious adults among other marine taxa surveyed for chemical defenses (Lindquist and Hay 1996). Adult blue-ringed octopuses harbor tetrodotoxin (TTX; Sheumack et al. 1978), which appears to be produced by symbiotic bacteria (Hwang et al. 1989). This deadly neurotoxin shuts down action potentials in nerve and muscle tissue by blocking sodium channel pores (e.g. Narahasi 2001). This toxicity hypothetically supplants inking as a chemical defense in this genus. Ink sacs in adult blue-rings atrophy (Robson 1929), though adults of H. lunulata and two undescribed congeners can still expel ink (Huffard and Caldwell 2002; Norman 2000). The benthic hatchlings of H. fasciata ink only at ages 1.5–4.5 wks (Tranter and Augustine 1973); newly hatched pelagic paralarvae of H. lunulata also expel minute amounts of ink (pers. obs.). However, these young octopuses ink infrequently and may depend instead on TTX as a chemical defense. Sheumack et al. (1984) identified TTX in the eggs of one species of blue-ringed octopus (probably H. fasciata based on the authors’ locality). Whether TTX defends early life stages from predators is unknown. Additionally, chemical defenses can be costly as they are in aspens and wild parsnip (Zangerl and Berenbaum 1997; Häikiö et al. 2009). In these taxa tradeoffs exist between defense and growth and reproduction. Bullard et al. (1999) empirically observed that many pelagic larvae are chemically undefended. No young of any blue-ringed octopus have yet been examined for the presence of TTX or other evidence of chemical protection. Despite the efficacy against vertebrate neurosystems (e.g. Narahasi 2001) and the presence of the toxin in many unrelated and widespread taxa (reviewed by Noguchi and Arakawa 2008), the ecological function of TTX has rarely been tested empirically (Williams 2010). In newts, TTX appears to be an effective defense against avian, fish, and snake predators (Brodie 1968; Mobley and Stidham 2000; Mochida 2009; Williams et al. 2010). Although TTX is common amongst pufferfish (e.g. Noguchi and Arakawa 2008), and some freshwater fish gustatory neurons respond to TTX (Yamamori et al. 1987), empirical data on the realized anitpredator efficacy is lacking. Conversely, although originally thought to be a defense (Miyazawa et al. 1987), TTX is used as venom in a polyclad worm 123 B. L. Williams et al. (Ritson-Williams et al. 2006). Here we test whether the pelagic paralarvae of H. lunulata contain TTX and, if so, whether TTX protects the paralarvae from predation by common reef predators. Materials and methods Sampling We obtained 30 adult Hapalochlaena lunulata, shipped from Bali, Indonesia, through the tropical aquarium industry between June 2004 and November 2006. Octopuses were wild-caught at our request and, according to the importer, were in transit no more than 1 week. We also obtained five male and two female adult Octopus cf. mercatoris via the pet trade from Florida, USA; these small octopuses occupy similar habitat but provided a non-TTX bearing control. Adult octopuses were housed in 38 l aquaria at 25°C. Hatchlings and paralarvae were housed individually in 0.5 l cups with approximately 150 ml artificial seawater (ASW) mixed at 34 ppt (Aqua Craft, Inc.). Every third day adult octopuses were fed live grass shrimp, ghost shrimp, snails, brine shrimp, or thawed prawn or fish pieces as availability permitted. Males were introduced into home aquaria of females for mating (see Cheng and Caldwell 2000 for a detailed description of mating). Out of 14 introductions of H. lunulata, seven matings were completed. Two female H. lunulata (one unmated at our facilities) deposited eggs. These females were offered food approximately twice weekly, but stopped feeding during brooding and expired soon after eggs hatched as previously observed in semelparous octopuses (i.e. H. lunulata and H. maculosa, Overath and von Boletzky 1974; Tranter and Augustine 1973). The eggs of the unmated female did not develop. The other female (mated three times) yielded a fertile clutch of *260 paralarvae, which hatched after 32–34 days. Approximately 100 paralarvae hatched synchronously in the early morning on the second and third days. Newly hatched paralarvae were *2–3 mm in length. Both control O. cf. mercatoris females laid fertile eggs within one week of arrival and hatchlings emerged 23–24 days hence at 5–6 mm in length. Twelve H. lunulata paralarvae and two O. cf. mercatoris hatchlings were immediately frozen at -80°C for later toxicity analyses. The remaining animals were reserved for behavioral trials (below). Representative stomatopod predators (Haptosquilla trispinosa) were collected from Lizard Island, Queensland Australia in 2005 and 2007. The stomatopods were housed at UC Berkeley, California, USA in plastic cups (*150 ml of 32 ppt ASW in 0.5 l cups), and kept under 14:10 light:dark conditions at 25°C. The water was changed twice Chemical defense in pelagic octopus paralarvae 133 weekly. Various species of larger stomatopods, which were wild-caught or obtained via the pet trade for other projects during the previous 5 years, were housed at UC Berkeley in larger cups (*300 ml ASW in 1 l cups) or individual aquaria (38 l). Smaller stomatopods were fed twice weekly with 10–20 live brine shrimp or an equivalent amount of frozen brine shrimp or frozen Mysis shrimp as supply permitted. Larger stomatopods were fed thawed shrimp or fish, or freeze-dried krill. Fish predators were housed at the Octopus’ Garden pet store in Berkeley, California, USA. This retail marine aquarium store provided a high diversity of tropical reef fish, many of which occur sympatrically with Hapalochlaena and may be natural predators of octopus paralarvae. Fish were housed in 5 l aquaria. after mixing. A water jacket cooled the derivatization products before detection by a Jasco FP-1512 (or Dynamax model FL-2) fluorescence detector with excitation wavelength set at 365 nm and emission wavelength at 510 nm. The Beckman System Gold Software (or a Hewlitt Packard 3390A Integrator) recorded the chromatographs and calculated peak areas. Tetrodotoxin peaks in tissue samples were identified and quantified by comparison with a standard of 1.0 lg/ml TTX in 0.10 N acetic acid (or 2.5 lg/ml TTX in 0.05 N acetic acid depending on the sensitivity of the HPLC system used). Paralarvae of H. lunulata were compared with the hatchlings of O. cf. mercatoris (which do not contain TTX) as a control. We used a peak skimming formula to conservatively calculate peak areas where the TTX peak eluted on top of a broad band. Tetrodotoxin extraction and quantification Antipredator efficacy of TTX Tetrodotoxin was extracted from samples by modifying the methods of Hanifin et al. (1999). We extracted TTX from paralarvae by grinding in the microcentrifuge tube with a plastic pestle or via sonication at level 3 in 100 ll of 0.05 N acetic acid. Food samples spiked with TTX (see Experiments 2–4 below) and O. cf. mercatoris hatchlings were extracted individually in the same manner. Extracts were heated to 100°C in boiling water for 5 min before cooling in an ice bath. The extracts were then centrifuged at 13,000 rpm for 15 min, the supernatant collected, and the supernatant was filtered through 5,000 nominal molecular weight limit (NMWL) 0.5 ml Ultrafree-MC Millipore tubes by centrifugation at 13,000 rpm for 20 min. Filtrates were stored at -80°C until quantification. We quantified TTX in sample extracts using 20 ll aliquots on two different reverse phase high performance liquid chromatography (HPLC) systems by modifying the procedures of Yotsu et al. (1989), and Hanifin et al. (1999) to optimize peak separation and quantification for each system. Data sets used for statistical analyses were collected on one system or the other; samples were not split between both. Samples were eluted with either an isocratic gradient of 1.0% (or 3.0%) by volume acetonitrile, 1.28% (or 0.97%) by weight heptafluorobutyric acid, and 0.29% by volume acetic acid. Each mobile phase was adjusted to pH 5.0 with 50% NH4OH. We separated analytes with a Synergi 4 lm Hydro-RP 80A C18 column (250 9 4.6 mm, Phenomenex, USA) on both systems. A Beckman Coulter System Gold 126 Solvent Module provided a primary flow rate of 0.5 ml per min for the mobile phase and a secondary flow rate of 1.0 ml per min of 5 N sodium hydroxide for mixing post-column (or a Shimadzu LC-10ADvp module and a Beckman 110A for the primary [0.5 ml/min] and secondary [0.9 ml/min] flow rates, respectively). A Pickering CRX 400 reactor heated elutants to 115°C (or 130°C) Fish and stomatopod predators were offered treatment and control food items in each of five experiments. Experiments were not conducted concurrently; individual predators are not necessarily the same between experiments. No individual predator was used more than once per experiment. ‘‘Injections’’ of treatment or control solution were carefully made under a dissecting microscope on a paper towel to ensure no solution leaked from the food items before presentation to the predators. Dry food items absorbed the TTX or control solutions coated on the outside and were rehydrated. Mysis were injected in between the carapace and first body segment; any applied pressure (i.e., handling by the predator) consequently released the body cavity contents. As an additional control, food items injected with treatment solution were dipped in water, to simulate the time a food item was in the water during behavioral trials, and then extracted to quantify TTX via HPLC. The ‘‘injection’’ procedure and HPLC assay verified that injected solutions did not diffuse from food items before predators accepted them and that TTX was accessible to the predator as the food item was sampled. Food items were introduced into home aquaria or cups that were occupied by a single predator. We randomized the order of control or treatment food items. We recorded whether treatment and control items were taken into the predator maw (sampled), and whether they were eaten if sampled. Any organism possesses a suite of defenses, including behavioral, morphological, and chemical. Because we were interested in the ‘‘last line of defense’’ chemical defenses, we focused on the interaction of predators and prey beginning when the predator physically contacted the prey item (sampling). If the item was sampled the two possible outcomes were ‘Eaten’ or ‘Rejected.’ ‘Eaten’ literally meant that the food item was consumed and was not ejected at any later time. ‘Rejected’ was recorded when the item was sampled but then ejected/ 123 134 B. L. Williams et al. discarded. Trials were discarded in the following instances: on the uncommon occurrence (29 in all experiments) that items were repeatedly mouthed and rejected prior to ingestion because TTX may have been dispelled from the food item; if the predator did not appear to see the food item within 5 s to control for possible diffusion of TTX away from the food item; and if the predator did not sample both treatment food items and controls we assumed they were not hungry (if they sampled both we assumed they were hungry even if both items were rejected and scored this data). If the food item was investigated but not sampled by stomatopods, we recorded ‘Rejected.’ For example, stomatopods may approach, investigate with antennae or antennules, and may or may not strike at the food item without taking it in toward the mouth parts before disregarding the item. This is akin to sampling in other predators and is a typical mode of ‘‘smelling’’ for stomatopods. Regardless of whether stomatopods rejected Table 1 Results of Experiment 1: blue-ringed octopus paralarvae palatability experiment paralarvae by ‘‘smelling’’ or ‘‘tasting,’’ the paralarvae can still be considered ‘‘rejected.’’ Predators that ingested a food item containing TTX were observed for intoxication or any ill effects for approximately 60 min and then again 24 h after the experiment. Experiment 1 We tested palatability of paralarvae in the first experiment with potential octopus predators both sympatric and allopatric to H. lunulata. Paralarval H. lunulata (2.5 mm length 9 1.5 mm diameter) between one and two days after emerging from their egg cases were offered to fish (13 species, N = 13) and stomatopod (5 species, N = 17) predators (Table 1). Live brine shrimp were offered to predators as controls because of their similar size and movement compared to paralarvae and because they were available at the time paralarvae hatched. Scientific name Common name PS Paralarvae Brine Shrimp Acanthurus triostegus Amblygobius rainfordi Convict tang Old glory Y Y R R E E Amphiprion ocellaris Clown anenomefish Y R E Centropyge aurantonotus Flameback anglefish N R E Centropyge flavicauda Whitetail pygmy anglefish Y R E Cirrhilabrus rubriventralis Social wrasse N R E Dendrochirus brachypterus? Dwarf lionfish Y R E Gonodactylus chiragra Mantis shrimp N R E Gonodactyaceus falcatus Mantis shrimp Y R E Gonodactyaceus falcatus Mantis shrimp Y E E Gonodactyaceus falcatus Mantis shrimp Y R E Gonodactyaceus falcatus Mantis shrimp Y R R Gonodactyaceus falcatus Mantis shrimp Y R E Gonodactyaceus falcatus Mantis shrimp Y R E Gonodactyaceus falcatus Mantis shrimp Y R E Gonodactyaceus falcatus Lienardella fasciata Mantis shrimp Harlequin tusk wrasse Y Y R E E E Odonodactylus havanensis Mantis shrimp N E E Odonodactylus havanensis Mantis shrimp N R R Odonodactylus havanensis Mantis shrimp N R E Odonodactylus havanensis Mantis shrimp N R E Odonodactylus havanensis Mantis shrimp N R E Paralarvae of Hapalochlaena lunulata and brine shrimp controls were offered to predators. Predators rejected H. lunulata paralarvae more often than brine shrimp controls Odonodactylus scyllarus Peacock mantis shrimp Y E E Paraglyphidodon nigroris Black-and-gold chromi Y R E Pseudochromis diadema Diadem dottyback Y R E Pseudochromis paccagnellae Royal dottyback Y R E Raoulserenea pygmaea Mantis shrimp N R E PS possible sympatry, E eaten, R rejected Raoulserenea pygmaea Mantis shrimp N E E v2 = 16.03, df = 1, N = 30, P \ 0.001 Salarius sp. Blenny ? R E Unknown Wrasse (yellow) ? R E 123 Chemical defense in pelagic octopus paralarvae Experiment 2 with either 200 ng TTX (2 ll of a 0.1 mg/ml solution in ASW with 0.01 N acetic acid) or control ASW (Table 3). In Experiment 2 we tested the antipredator efficacy of TTX against a large diversity of predators. Fish and stomatopod predators were offered dry food items (11 mm length 9 2 mm diameter) injected with either 200 ng TTX (5 ll of a 0.04 mg/ml solution in ringer solution) based on the average amount of TTX measured by HPLC in H. lunulata paralarvae (see ‘‘Results’’ section) or a control injection of 5 ll Ringer’s solution (Table 2). Fish (8 species, N = 9) were given floating tropical food pellets while stomatopods (9 species, N = 15) were offered pieces of freeze dried krill, typical food items for these captive animals. However, unknown constituents in these food items interfered with accurate quantification of TTX in samples by HPLC (see ‘‘Results’’ section). Experiment 3 Experiment 3 confirmed the results of Experiment 2 with a subset of predators offered another food item that did not interfere with TTX quantification. In this experiment, stomatopods (8 species, N = 23) were offered Mysis shrimp (6 mm length 9 1 mm diam.) that were injected Table 2 Results of Experiment 2: dried food palatability experiment 135 Experiment 4 In Experiment 4 we opportunistically tested another nonTTX bearing cephalopod as the food item. During Experiment 4 we offered predators a small set of benthic Octopus cf. mercatoris hatchlings. Sixteen hatchlings were injected with either 200 ng TTX (5 ll of a 0.04 mg/ml solution in Ringer’s solution) or a control injection of 5 ll Ringer’s solution. We offered one spiked and one control hatchling to each of four species of fish and four species of stomatopods (N = 8). Experiment 5 In the fifth experiment we offered multiple food items spiked with TTX to in a single population sample of stomatopods (Haptosquilla trispinosa from Queensland, Australia) to test the possibility that greater quantities of TTX than found in paralarvae may be deterrent. Tetrodotoxin (1 mg) was dissolved in 1 ml of 0.05 N acetic acid in ASW, then diluted by half in ASW. Mysis shrimp were Scientific name Common name PS Spiked food Control Balistapus undulatus Undulated triggerfish Y E E Balistoides conspicillum Clown triggerfish Y E E Ctenochaetus tominiensis Tomini tang Y E E Gonodactylus chiragra Mantis shrimp N R E Gonodactylus chiragra Mantis shrimp N E E Gonodactylus chiragra Mantis shrimp N R R Gonodactyaceus falcatus Mantis shrimp Y R R Haptosquilla stoliura Mantis shrimp Y R E Haptosquilla stoliura Mantis shrimp Y R E Lysiosquillina maculata Odonodactylus brevirostris Mantis shrimp Mantis shrimp Y N E E E E Odonodactylus havenensis Mantis shrimp N E E Odonodactylus havenensis Mantis shrimp N E E Odonodactylus havenensis Mantis shrimp N R R Odonodactylus latirostris Mantis shrimp Y R E Odonodactylus scyllarus Peacock mantis shrimp Y E E Odonodactylus scyllarus Peacock mantis shrimp Y E E Plepidogenys? Damsel ? E E Pseudochromis splendens Splendid dottyback Y E E Pseudosquilla ciliata False mantis/rainbow mantis Y E E Rhinecanthus aculeatus Picasso triggerfish Y E E PS possible sympatry, E eaten, R rejected Siganus puellus Blue line rabbitfish Y E E v2 = 0.421, df = 1, N = 24, P = 0.516 Siganus puellus Blue line rabbitfish Y E E Yongeichthys nebulosus Shadow goby Y E E We injected commercial food items (dried krill and tropical fish food pellets) with either 200 ng TTX or ringer solution as a control. Predators did not discriminate between innocuous food items and those spiked with TTX 123 136 Table 3 Results of Experiment 3: Mysis shrimp palatability experiment B. L. Williams et al. Scientific name Common name PS Spiked Mysis Control Acanthosquilla multifasciata Mantis shrimp Y R R Acanthosquilla multifasciata Mantis shrimp Y E E Acanthosquilla multifasciata Mantis shrimp Y E E Gonodactylellus caldwelli Blue-lined mantis shrimp Y R R Gonodactylus chiragra Mantis shrimp N E E Gonodactylus chiragra Mantis shrimp N E R Gonodactylus smithii Purple spotted mantis shrimp Y E E Gonodactylus smithii Purple spotted mantis shrimp Y E E Haptosquilla stoliura Mantis shrimp Y E E Haptosquilla stoliura Mantis shrimp Y E E Haptosquilla stoliura Mantis shrimp Y R R Haptosquilla stoliura Mantis shrimp Y E E Haptosquilla trispinosa Mantis shrimp Y E R Haptosquilla trispinosa Mantis shrimp Y E E Haptosquilla trispinosa Neogonodactylus curacaoensis Mantis shrimp Mantis shrimp Y N R R E R Odontodactylus havenensis Mantis shrimp N R E Odontodactylus havenensis Mantis shrimp N E E Odontodactylus havenensis Mantis shrimp N E E Odontodactylus scyllarus Peacock mantis shrimp Y E E PS possible sympatry, E eaten, R rejected Odontodactylus scyllarus Peacock mantis shrimp Y E E v2 = 0.000, df = 1, N = 23, P = 1.00 Odontodactylus scyllarus Peacock mantis shrimp Y E E Pseudosquilla ciliata False mantis/rainbow mantis Y R R We injected Mysis shrimp with either 200 ng TTX or artificial seawater as a control. No difference existed in the number of spiked versus control Mysis shrimp consumed spiked with either 2 ll of 500 lg/ml TTX solution (1 lg TTX injected in total) for the treatment group, or 2 ll of 0.025 N acetic acid in ASW as a control. Stomatopods of approximately the same size class were chosen for the experiment. Immediately after injections the Mysis shrimp were presented to the stomatopods in their cups via small forceps. The stomatopods were offered control or TTXspiked Mysis shrimp until rejection or until they accepted a total of five. Half of the stomatopods, chosen at random, were tested with control Mysis and half with TTX-spiked Mysis on the first day. On a different day, those that first received controls received TTX-spiked Mysis and vice versa. Stomatopods were typically fed twice a week. Trials were conducted on a regular feeding day; all experimental subjects were fed *15 Mysis shrimp as per their normal feeding schedule at the conclusion of the experiment. Analyses Tetrodotoxin was extracted and quantified in spiked food samples and compared with unspiked samples to confirm the validity of the spiking procedure. We tested whether food items were rejected more often than controls in Experiments 1–3 with v2 goodness-of-fit tests (Tables 1, 2, 3). In order to 123 test whether dose (amount of TTX/size predator) varied between experiments we compared size of predators between experiments. We estimated fish to the nearest cm and stomatopods to the nearest 0.5 cm after feeding trials as we did not anesthetize predators and wished to reduce handling time. We then assigned size classes to predators based on 2 cm gradations for the analyses to account for error of our estimates. We used t-tests to evaluate whether predator size class varied between the blue-ring paralarvae palatability (Experiment 1) and each of the two spiked food palatability experiments (dry food and Mysis shrimp; Experiments 2 and 3). Data from Experiment 4 (O. cf. mercatoris hatchlings) were not analyzed because of the small sample size. In Experiment 5 we compared the number of TTX-injected Mysis shrimp eaten to the number of controls eaten with the non-parametric Wilcoxon Sign-Rank test. Although we chose predators of similar size for this experiment, we also evaluated whether the number of TTX-spiked Mysis eaten varied with predator size with ANOVA. For this analysis, the number of TTX-spiked Mysis was transformed [log (# eaten ? 1)] to approximate a normal distribution as indicated by the Shapiro–Wilks Test. Untransformed stomatopod size did not differ from a normal distribution. Analyses were performed with JMP version 7.0, copyright 2007 SAS Institute, Inc. Chemical defense in pelagic octopus paralarvae Results The paralarvae of H. lunulata contained mean ± SE = 150 ± 17 ng TTX per parlarava (N = 11); no TTX was found in the O. cf. mercatoris hatchlings examined (Fig. 1). The lower detection limit of this HPLC was *10 ng TTX. Antipredator efficacy of TTX 137 dipped in water. Mysis shrimp presented no interfering compounds to hamper the quantification of TTX in the spiked samples. Estimated TTX quantity in two samples of Mysis shrimp spiked with 200 ng TTX were 263 and 321 ng of TTX respectively. Because the amount we injected into food items was near the lower detection limit, variability of estimates was high. However, given that we injected 200 ng TTX—approximately 33% above the mean amount of TTX in paralarvae (150 ± 17 ng)—and anything under 100 ng was undetectable on this machine, we Spiked food items such as dried krill, fish food pellets, or Mysis shrimp, were compared to un-spiked samples of the same food via HPLC. Tetrodotoxin peaks were visualized in chromatographs of dried food items (krill and fish food pellets); however, the presence of a co-eluting, unidentified constituent interfered with highly accurate quantification of the TTX peak (Fig. 2). Given that the limit of detection of this HPLC was 100 ng and that the injected TTX was visualized in spiked food samples, we could ascertain that more than 100 ng of the 200 ng of TTX injected was recovered from spiked food samples even after having been Fig. 1 Chromatographs illustrating a recovery of tetrodotoxin (TTX) in a paralarvae of Hapalochlaena lunulata, and b absence of TTX in an Octopus cf. mercatoris hatchling Fig. 2 Chromtographs illustrating recovery of tetrodotoxin (TTX) in food items spiked with 200 ng of TTX, and absence of TTX in control food items. A commercial standard of TTX is given for comparison. The dotted line is an aid to visualize the retention time of TTX. a Extract of dried krill. b Extract of dried krill spiked with TTX. c Extract of Mysis shrimp. d Extract of Mysis shrimp spiked with TTX 123 138 are confident that the amount of TTX in all spiked food items equaled or exceeded the amount of TTX in paralarvae. Predators rejected H. lunulata paralarvae more often than brine shrimp controls (Experiment 1; v2 = 16.03, df = 1, N = 30, P \ 0.001; Table 1). Most paralarvae sampled were rejected instantly. Only 5 of 30 predators ate paralarvae. Neither predators that ate paralarvae nor those that rejected paralarvae displayed any obvious signs of TTX intoxication such as unsteadiness, paralysis, or reduced ventilation rate (e.g. Kao 1966; Noguchi and Ebesu 2001). All rejected paralarvae survived attempted ingestion by predators. Predators did not discriminate between spiked and innocuous dry food items (Experiment 2; v2 = 0.421, df = 1, N = 24, P = 0.516; Table 2). Similarly, no difference existed in the number of spiked (16/23) versus control (16/23) Mysis shrimp consumed by predators (Experiment 3; v2 = 0.000, df = 1, N = 23, P = 1.00; Table 3). Predators also found O. cf. mercatoris hatchlings palatable whether they were injected with TTX or Ringer’s solution (Experiment 4; 7/8 hatchlings consumed in both treatments). As in Experiment 1, no predators exhibited any signs of TTX intoxication or ill effects after Experiments 2–4. The predator size classes for Experiment 1 involving the blue-ring paralarvae (mean ± SD = 3.47 ± 1.07) and Experiment 2 involving dry food (3.66 ± 1.27) were not significantly different (t = 0.614, df = 45, P = 0.542), but that for Experiment 3 involving Mysis shrimp (2.65 ± 0.83) was smaller (t = -3.11, df = 51, P = 0.003). Thus, predators presented with TTX-spiked food items received an equal or higher dose of TTX (dose incorporates mass of predator) than predators exposed to TTX in paralarvae, yet still consumed spiked food items with impunity. In Experiment 5 we searched for deterrence by TTX in a sample of stomatopods from one population (Haptosquilla trispinosa). More control Mysis shrimp (mean ± SD = 3.91 ± 1.36) were eaten than TTX-spiked Mysis (1.91 ± 1.45, Z = 21.50, n = 11, df = 11, P = 0.0127). Because each Mysis shrimp was injected with 1 lg TTX these stomatopods consumed on average 1.9 lg of TTX before rejection of the food items. Each paralarvae of H. lunulata averaged *150 ng TTX in the clutch we examined. By extrapolation, we estimated that stomatopods in our sample group could have consumed approximately 12–13 of these paralarvae before reaching a noxious dose of TTX. Although TTX-spiked Mysis were rejected, we observed no obvious signs of intoxication in the stomatopods. However, we did not explicitly test intoxication; because TTX inhibits locomotor function (e.g. Brodie and Brodie 1999), swimming speed or endurance may be a viable assay to test intoxication of predators in future experiments. Because we chose stomatopods of roughly the same size, there was no correlation 123 B. L. Williams et al. between predator mass and amount of TTX consumed in this study (F(1,9) = 0.0507, P = 0.8269). While body size does influence the effect of xenobiotics, there was not enough variation in body size in our sample to discern this effect here. Additionally, individual susceptibility to a toxin will likely play a role but was not assessed in this study. Discussion Paralarvae of Hapalochlaena lunulata were unpalatable to most predators. Although TTX is fast-acting and onset of symptoms in humans and garter snakes occurs in as little as 1–15 min (Kao 1966; Williams et al. 2003), predators rejected paralarvae of H. lunulata upon immediate contact (\1 s)—faster than expected. Additionally, predators did not find TTX-spiked food items distasteful and were not intoxicated after consuming either paralarvae or TTXladen food items. Thus, a heretofore-unknown substance is implicated in paralarval distastefulness. This substance may or may not act in concert with TTX. A distasteful compound need not directly harm a predator. It may facilitate predator learning through multiple cues (e.g. Fisher 1930; Nicolaus et al. 1983). In this scenario, the compound signals the onset of more detrimental symptoms of TTX. While it is theoretically possible the unknown defense is not chemical, we favor this hypothesis because: most young octopus are highly palatable; H. lunulata paralarvae possess no unusual morphologies compared to other cephalopods and do not display the aposematic blue rings of adults; paralarvae were rejected only upon sampling by predators and this limits paralarval options for behavioral defenses; and cephalopods in general are not known for their morphological defenses (the beak is the only hard part; Hanlon and Messenger 1996). It is clear, however, that the quantity of TTX in individual paralarvae was not deterrent to the predators tested here and an alternative defense must exist. Savage and Howden (1977) reported the presence of a second toxic component called hapalotoxin in addition to maculotoxin in H. fasciata adults (maculotoxin = TTX; Sheumack et al. 1978). Hapalotoxin was deemed similar in size to TTX, but exhibited some unique properties and thus may have been an analog of TTX. Additional TTX analogs are common in TTX bearing taxa, though few are as efficacious as TTX itself (Yasumoto and Yotsu-Yamashita 1996). Our analysis revealed no peaks in the HPLC profile of H. lunulata paralarvae that suggested the presence of additional TTX analogs (Fig. 1). Furthermore, to our knowledge, none of the known analogs have been reported as distasteful. Our extraction technique eliminated large molecules over 5000 NMWL by filtration and smaller peptides were likely denatured during the boiling process Chemical defense in pelagic octopus paralarvae and thus likely removed constituents other than TTX and its analogs. Therefore, further work is required to identify any additional defensive compound in paralarvae of H. lunulata. Although predators were not deterred by the amount of TTX found in an individual paralarva, paralarval TTX may provide benefit against certain predators, such as H. trispinosa, through kin selection. TTX in quantities larger than those contained by individual paralarvae did deter predators. Moths (Utetheisa ornatrix) endow their eggs with a protective pyrolizidine alkaloid. Green lacewing larvae (Ceraeochrysa cubana) sample the moths’ eggs before consuming the whole clutch, destroying some eggs in the process, but leaving the siblings intact when the toxicity of the clutch exceeds the lacewing’s tolerance (Eisner et al. 2000). In one sample of stomatopod predators, H. trispinosa, weighing on average 479 ? 34 (SE) mg, the lower level of deterrence was 1.9 lg of TTX. This corresponds to approximately 12–13 paralarvae in the clutch examined here. Blue-ring octopus paralarvae may benefit from kin selection during egg development and even immediately after hatching. Paralarvae of H. lunulata hatch synchronously and diurnally (Overath and von Boletzky 1974; this study) and the group may therefore be subject to predation as they hatch and disperse into the water column. Even in the absence of kin selection, the level of TTX seen in this study may still be functionally defensive. At hatching the entire paralarvae possessed a concentration of 16.14 ± 4.84 lg TTX/g. Parasites of the paralarvae may be negatively affected at this dose. Alternately, smaller zooplankters, such as fish or crustacean larvae that might prey upon octopuses in the water column, might be intoxicated or even killed by the level of TTX in paralarvae (*150 ng). Susceptibility to TTX varies by orders of magnitude between taxa (e.g. Brodie 1968) and TTX may be more effective against predators or parasites that were not tested in this study. Thus, antipredator efficacy of TTX (and any potential unknown distasteful substance, if discovered) should be tested in a variety of contexts. Another reason that predators were not deterred by the TTX levels found in the clutch examined could arguably be that the concentration of TTX in this clutch was lower than wild-hatched conspecifics or that TTX levels vary considerably among clutches. Animals that obtain chemical defenses exogenously, such as poison dart frogs (e.g. Daly et al. 1997), often lose toxicity in captivity. For example, puffer fish exhibit very low or no TTX levels when raised in captivity (e.g. Matsui et al. 1982; Matsumura 1996; Noguchi et al. 2006; Saito et al. 1984). However, the TTX levels found in adult H. lunulata from this area (Bali, Indonesia) are comparable to levels found in other species of the genus (Williams and Caldwell 2009; Yotsu-Yamashita et al. 139 2007). Also, TTX stores increase in developing embryos of H. lunulata in captivity after eggs have been laid but before hatching, suggesting that the mechanism of TTX production remains intact (Williams et al. 2011). Finally, the TTX levels in this clutch are comparable to those of other developing embryos of H. lunulata before hatching (Williams et al. 2011). Thus, we conclude that the TTX levels in the clutch examined here are representative of natural conditions. Tetrodotoxin is present in paralarvae of H. lunulata. This TTX is deterrent to predators, but only at amounts far greater than those possessed by paralarvae. Even so, paralarvae were rejected immediately. Predators rejected 83% of the paralarvae and all rejected paralarvae survived. These results suggest the presence of an additional mechanism that confers protection to individual paralarvae of H. lunulata. Further investigation is needed to clarify the role of TTX, if any, and to discover the chemical identity of any additional defensive compound(s) in paralarvae. Acknowledgments The UC Berkeley ACUC authorized this research (protocol #R302-0407). The Great Barrier Reef Marine Park Authority provided a collecting permit to BLW and RLC for Haptosquilla trispinosa (#G05/13381.1). Dr. Charles Hanifin at Hopkins Marine Station, Stanford University, and Dr. Edmund D. Brodie, Jr. at Utah State University provided assistance with HPLC analyses. Stephanie Bush, Julie Himes, and Calida Martinez assisted in the animal behavior lab and at Octopus’ Garden pet store in Berkeley, CA. We would like to thank the owner of Octopus Gardens, Erin Janoff, and an employee, Laith Shabbas, for their indispensable help identifying fish and for logistical support. The manuscript benefited from comments by Dr. Akira Mori at Kyoto University, Maya DeVries, Stephanie Bush, Joey Pakes, Jean Alupay, Dr. Dustin Rubenstein, and Dr. David Lindberg at UC Berkeley, and Krista Heideman, Dr. Jamie Howard, Anne Jacobs, and Anna Garliss at New Mexico State University. Funding was provided by a Sigma Xi Grantin-Aid-of-Research and the Department of Integrative Biology, at UC Berkeley. References Brodie ED Jr (1968) Investigations on the skin toxin of the adult rough-skinned newt, Taricha granulosa. Copeia 1968:307–313 Brodie ED III, Brodie ED Jr (1999) The cost of exploiting poisonous prey: evolutionary tradeoffs in a predator-prey arms race. Evolution 53:626–631 Bullard SG, Niels L, Hay M (1999) Susceptibility of invertebrate larvae to predators: how common are post-capture larval defenses? Mar Ecol Prog Ser 191:153–161 Cheng MW, Caldwell RL (2000) Sex identification and mating in the blue-ringed octopus, Hapalochlaena lunulata. Anim Behav 60:27–33 Daly JW, Garaffo HM, Hall GSE, Cover JF (1997) Absence of skin alkaloids in captive-raised Madagascan mantelline frogs (Mantella) and sequestration of dietary alkaloids. Toxicon 35: 1131–1135 Derby CD (2007) Escape by inking and secreting: marine mollusks avoid predators through a rich array of chemicals and mechanisms. Biol Bull 213:274–289 123 140 Eisner T, Eisner M, Rossini C, Iyengar VK, Roach BL, Benedikt E, Meinwald J (2000) Chemical defense against predation in an insect egg. Proc Natl Acad Sci USA 15:1634–1639 Fisher RA (1930) The genetical theory of natural selection. Oxford University Press, Oxford Häikiö E, Makkonen M, Julkunen-Tiitto R, Sitte J, Freiwald V, Silfver T, Pandey V, Beuker E, Holopainen T, Oksanen E (2009) Performance and secondary chemistry of two hybrid aspen (Populus tremula L. 9 Populus tremuloides Michx.) clones in long-term elevated ozone exposure. J Chem Ecol 35:664–678 Hanifin CT, Yotsu-Yamashita M, Yasumoto T, Brodie ED III, Brodie ED Jr (1999) Toxicity of dangerous prey: variation of tetrodotoxin levels within and among populations of the newt Taricha granulosa. J Chem Ecol 25:2161–2175 Hanlon RT, Messenger JB (1996) Cephalopod behaviour. Cambridge University Press, Cambridge Huffard CL, Caldwell RL (2002) Inking in a blue-ringed octopus, Hapalochlaena lunulata, with a vestigial ink sac. Pac Sci 56:255–257 Hwang DF, Arakawa O, Saito T, Noguchi T, Simidu U, Tsukamoto K, Shida Y, Hashimoto K (1989) Tetrodotoxin-producing bacteria from the blue-ringed octopus, Octopus maculosus. Mar Biol 100:327–332 Jereb P, Roper CFE, Vecchione M (2005) Introduction. In: Roper CFE (ed) Jereb P. an Annotated and Illustrated Catalog of Species Known to Date. FAO Rome Italy, Cephalopods of the World Johnson S (2001) Hidden in plain sight: the ecology and physiology of organismal transparency. Biol Bull 201:301–318 Jones EC (1963) Tremoctopus violaceus uses Physalia tentacles as weapons. Science 139:764–766 Kao CY (1966) Tetrodotoxin, saxitoxin, and their significance in the study of excitation phenomena. Pharmacol Rev 18:997–1049 Lindquist N, Hay ME (1996) Palatability and chemical defense of marine invertebrate larvae. Ecol Monogr 66:431–450 Lindquist N, Hay ME, Fenical W (1992) Defense of ascidians and their conspicuous larvae: adult versus larval chemical defenses. Ecol Monogr 62:547–568 Matsui T, Sato H, Hamada S, Shimizu C (1982) Comparison of toxicity of the cultured and wild puffer fish Fugu niphobles. Bull Jpn Soc Sci Fish 48:253 Matsumura K (1996) Tetrodotoxin concentration in cultured puffer fish, Fugu rubripes. J Agric Food Chem 44:1–2 McClintock JB, Vernon JD (1990) Chemical defense in the eggs and embryos of Antarctic sea stars (Echinodermata). Mar Biol 105:491–495 Miyazawa K, Jeon JK, Noguchi T, Ito K, Hashimoto K (1987) Distribution of tetrodotoxin in the tissues of the flatworm Planocera multitentaculata (Platyhelminthes). Toxicon 25:975–980 Mobley JA, Stidham TA (2000) Great horned owl death from predation of a toxic California newt. Wilson Bull 112:563–564 Mochida K (2009) A parallel geographical mosaic of morphological and behavioural aposematic traits of the newt, Cynops pyrrhogaster (Urodela: Salamandridae). Biol J Linn Soc 97:613–622 Morgan SG (1989) Adaptive significance of spination in estuarine crab zoeae. Ecol 70:462–482 Narahasi T (2001) Pharmacology of tetrodotoxin. J Toxicol Toxin Rev 20:67–84 Nicolaus LK, Cassel JF, Carlson RB, Gustavson CR (1983) Tasteaversion conditioning of crows to control predation on eggs. Science 220:212–214 Noguchi T, Arakawa O (2008) Tetrodotoxin—distribution and accumulation in aquatic organisms, and cases of human intoxication. Mar Drugs 6:220–242 123 B. L. Williams et al. Noguchi T, Ebesu JSM (2001) Puffer poisoning: epidemiology and treatment. J Toxicol Toxin Rev 20:1–10 Noguchi T, Arakawa O, Takatani T (2006) Toxicity of pufferfish Takifugu rubripes cultured in netcages at sea or aquaria on land. Comp Biochem Physiol D 1:153–157 Norman MD (2000) Cephalopods: a world guide. ConchBooks, Hackenheim Overath H, von Boletzky S (1974) Laboratory observations on spawning and embryonic development of a blue-ringed octopus. Mar Biol 27:333–337 Prota G, Ortonne JP, Voulot C, Khatchadourian GN, Palumbo A (1981) Occurrence and properties of tyrosinase in the ejected ink of cephalopods. Comp Biochem Physiol 68B:415–419 Ritson-Williams RR, Yotsu-Yamashita M, Paul V (2006) Ecological functions of tetrodotoxin in a deadly polyclad flatworm. Proc Natl Acad Sci USA 103:3176–3179 Robson GC (1929) A monograph of the recent Cephalopoda based on the collections in the British Museum (Natural History) Part 1 Octopodinae. London, British Museum 1–236 Russo GL, De Nisco E, Fiore G, Di Donato P, d’Ishcia M, Palumbo A (2003) Toxicity of melanin-free ink of Sepia officinalis to transformed cell lines: identification of the active factor as tyrosinase. Biochem Biophys Res Commun 308:293–299 Saito T, Maruyama J, Kanoh S, Jeon JK, Noguchi T, Harada T, Murata O, Hashimoto K (1984) Toxicity of the cultured pufferfish Fugu rubripes rubripes along with their resistibility against tetrodotoxin. Bull Jpn Soc Sci Fish 50:1573–1575 Savage IVE, Howden MEH (1977) Hapalotoxin: a second lethal toxin from the octopus Hapalochlaena maculosa. Toxicon 15:463–466 Sheumack DD, Howden MEH, Spence I (1978) Maculotoxin: a neurotoxin from the glands of the octopus, Hapalochlaena maculosa identified as tetrodotoxin. Science 199:188–189 Sheumack DD, Howden ME, Spence I (1984) Occurrence of a tetrodotoxin-like compound in the eggs of the venomous blueringed octopus (Hapalochlaena maculosa). Toxicon 22:811–812 Tranter DJ, Augustine O (1973) Observations on the life history of the blue-ringed octopus Hapalochlaena maculosa. Mar Biol (Berl) 18:115–128 Villanueva R, Norman MD (2008) Biology of the planktonic stages of benthic octopuses. Oceanogr Mar Biol Ann Rev 46:105–202 Williams BL (2010) Behavioral and chemical ecology of marine organisms with respect to tetrodotoxin. Marine Drugs 8(3):381–398 Williams BL, Caldwell RL (2009) Intra-organismal distribution of tetrodotoxin in two species of blue-ringed octopuses (Hapalochlaena fasciata and H. lunulata). Toxicon 54(3):345–353 Williams BL, Brodie ED Jr, Brodie ED III (2003) Coevolution of deadly toxins and predator resistance: self-assessment of resistance by garter snakes leads to behavioral rejection of toxic newt prey. Herpetologica 59:155–163 Williams BL, Hanifin CT, Brodie ED Jr, Brodie ED III (2010) Tetrodotoxin (TTX) affects survival probability of rough skinned newts (Taricha granulosa) faced with TTX-resistant garter snake predators (Thamnophis sirtalis). Chemoecology 20:285–290 Williams BL, Hanifin CT, Brodie ED Jr, Caldwell RL (2011) Ontogeny of tetrodotoxin levels in blue-ringed octopuses: maternal investment and apparent independent production in offspring of Hapalochlaena lunulata. J Chem Ecol. doi:10.1007/ s10886-010-9901-4 Yamamori K, Nakamura M, Hara TJ (1987) Gustatory responses to tetrodotoxin and saxitoxin in rainbow trout (Salmo gairdneri) and arctic char (Salvelinus alpinus): a possible biological defense mechanism. Ann NY Acad Sci 510:727–729 Yasumoto T, Yotsu-Yamashita M (1996) Chemical and etiological studies on tetrodotoxin and its analogs. J Toxicol Toxin Rev 15:81–90 Chemical defense in pelagic octopus paralarvae Yotsu M, End A, Yasumoto T (1989) An improved tetrodotoxin analyzer. Agric Biol Chem 53:893–895 Yotsu-Yamashita M, Mebs D, Flachsenberger W (2007) Distribution of tetrodotoxin in the body of the blue-ringed octopus (Hapalochlaena maculosa). Toxicon 49:410–412 Young CM, Bingham BL (1987) Chemical defense and aposematic coloration in larvae of the ascidian Ecteinascidia turbinata. Mar Biol 96:539–544 141 Zangerl AR, Berenbaum MR (1997) Cost of chemically defending seeds: furanocourmarins and Pastinaca sativa. Am Nat 150:491–504 123