Editor: Tim Halliday
Hurt yourself to hurt your enemy: new insights on the function of the bizarre antipredator mechanism in the salamandrid Pleurodeles waltl
Article first published online: 18 AUG 2009
© 2009 The Authors. Journal compilation © 2009 The Zoological Society of London
Journal of Zoology
Volume 280, Issue 2, pages 156–162, February 2010
How to Cite
Heiss, E., Natchev, N., Salaberger, D., Gumpenberger, M., Rabanser, A. and Weisgram, J. (2010), Hurt yourself to hurt your enemy: new insights on the function of the bizarre antipredator mechanism in the salamandrid Pleurodeles waltl. Journal of Zoology, 280: 156–162. doi: 10.1111/j.1469-7998.2009.00631.x
- Issue published online: 25 JAN 2010
- Article first published online: 18 AUG 2009
- Received 25 May 2009; revised 13 July 2009; accepted 14 July 2009
- antipredator behaviour;
The Spanish ribbed newt Pleurodeles waltl shows a bizzare defensive mechanism against predators. X-ray analysis before and after a simulated threat shows that this newt rotates its ribs anteriorly. The maximum measured angle to which the ribs moved was 65°. This forward movement causes the sharply pointed rib tips to lacerate the body wall and project freely from the sides of the trunk as spines. Light microscopy shows the microanatomy, and computed tomography shows the 3D morphology of these unusual weapons. They are ‘spear-shaped’ and hollow proximally, massive distally and are connected to the corresponding vertebra by a well-developed, two-headed joint. The skin in the penetration areas lacks permanent pores through which the ribs could be projected and is pierced de novo by every antipredator posturing. This investigation provides new insight into the functionality of a highly complex, integrated and unusual defensive strategy.
Amphibians are an essential part of the natural food chain. Being numerous, small to moderate in size and having soft skin, some of them are common prey for a huge variety of predators from all classes of vertebrates, as well as for certain arthropods (Duellman & Trueb, 1994). Amphibians have therefore evolved various morphological, physiological and behavioural features, which, alone or in combination, provide varying degrees of protection from potential predators (Duellman & Trueb, 1994; Heiss et al., 2009).
These features include escape behaviour, cryptic coloration and structure, noxiousness or toxicity and encounter behaviour (Duellman & Trueb, 1994). Among urodeles, the family Salamandridae has the greatest diversity of antipredator mechanisms (Brodie Jr, Nussbaum & DiGiovanni, 1984). In the salamandrid genus Pleurodeles and in the closely related genus Echinotriton, unique strategies to decrease palatability and increase survival rates have been described (Nowak & Brodie Jr, 1978; Brodie Jr, 1983; Brodie Jr et al., 1984). When attacked by a potential predator (or provoked with an adequate artificial stimulus), sharp spines appear on the lateral trunk sides. This phenomenon was first mentioned in Pleurodeles waltl by Leydig (1879). This author examined preserved and living material and rebutted earlier (orally referred) notions that the lateral spines of this animal were horny structures. Leydig suggested that the lateral spines of P. waltl are ribs that lie in a lymphatic sheath immediately beneath the skin. A study performed 99 years later by Nowak & Brodie Jr (1978) yielded similar conclusions.
The present study shows new information on the morphological and functional integration of the body wall and the ribs. It also provides new data on how P. waltl protrudes its ribs and on the mechanism in the framework of the antipredator behaviour. We apply photo- and X-ray imaging along with computed tomography (CT) to examine the (micro-) anatomical features of the ribs and histological techniques to study the emersion point of the ribs. We also discuss possible mechanisms preventing self-intoxication or microbial infection that could result from damaging the integrity of the skin. In this context, it is important to clarify whether the tips of the ribs really penetrate the skin or remain covered by integument. If the rib tips are uncoated, it should be determined whether the skin of P. waltl shows distinct and permanent pores or whether the body wall is penetrated de novo by every single antipredator posturing.
Materials and methods
Five male and four female adult (3–5 years old) P. waltl were used in the present study. The animals were obtained commercially and kept in a 300 L tank with a 12-h dark/12 h light cycle and fed with larval chironomids, earthworms and fish pieces. For behavioural experiments, the reactions to ‘predator-like stimulations’ were documented using a Canon EOS 350D digital camera (Canon Inc., Tokyo, Japan). To simulate a predator attack, the animals were touched repeatedly – but gently – with a cotton bud until they showed defensive behaviour. The animals recovered rapidly after the experiments and all showed natural behaviour such as feeding or mating immediately after the experiments.
For radiographic analyses, dorsoventral radiographs were made with a Siemens Polydoros 80 S machine (Siemens AG, Munich, Germany). Animals were examined with an analogue (Kodak InSight-System, Eastman Kodak Company, Rochester, NY, USA) or digital system (DirectView CR 850 with Kodak EHR M screens for mammography, Kodak GmbH, Health Group, Stuttgart, Germany). Two male and two female newts were positioned in a commercial plastic box and recorded in a relaxed position, and after stimulation. The animals were exposed for 0.56–0.8 ms at 50–65 kV with a film focus distance of 60 cm.
The animals showed two types of antipredator posturings: a flattened body or an arched body (see Figs 1b and 2b). For rib angle measurements, only the flattened antipredator posturing was considered because the dorsal flexion of the vertebral column in the arched posturing could distort the results.
The changes in the angle between the longitudinal axis of the rib and the longitudinal axis of the vertebral body before and after the ‘predator-like stimulus’ were measured. As a reference for the measurements, the imaginary line connecting the distal tip of the rib to the dorsal costo-vertebral joint, and the neural spine of the associated vertebra were used.
Ten angles on both sides were measured twice (measures A and B) before and after the stimulus on three different individuals; another unrepeated measure (only measure A) was taken on a fourth individual. The resulting dataset consisted of 280 angles and involved four variables: stimulus (relaxed vs. stimulated), side (right vs. left), measure (A vs. B) and individual (1–4). To test which variables influenced the mean angles, nonparametric tests were performed for each of the four variables. Regarding the three dichotomous variables stimulus, side and measure, exact Wilcoxon's paired rank tests were performed to compare the dependent (paired) data groups. Regarding the four individuals, a Kruskal–Wallis test for comparison of more than two independent groups of data was performed.
The trunk of one preserved male newt was scanned using CT. For CT, the Sub-μm-device Nanotom (Phönix|x-ray, Wunstorf, Germany) was used. During measurement, projection images were grabbed using an amorphous Silicon matrix detector at several angular positions. After a full 360° rotation, 1500 images were generated. A mathematical algorithm calculated a 3D dataset using the projection images. Grey values corresponding to the tissue density were assigned to each spatial element (voxel). The size of each voxel was 6 μm3. Surface and volume reconstructions were performed using Amira 4.1 (Mercury Computer Systems, Chelmsford, MA, USA) software.
For histological investigations, two adult newts (one male and one female) were stimulated as described above until they pierced their ribs during immobile posturing. While remaining in the immobile position, the animals were anaesthetized by immersing them into dissolved MS222 (0.01%), after which they were decapitated and immediately immersed in buffered (pH=7.2) 4% formalin for 10 days. After washing out the formalin with tap water, tissue samples (1-cm-thick slices from the trunk) were removed, dehydrated in a graded ethanol series and embedded in light curing resin (Technovit 7200 VLC+BPO; Kulzer and Co., Wehrheim, Germany). For further processing, the Exakt cutting and grinding equipment (Exakt Apparatebau, Norderstedt, Germany) was used to reduce the thickness of the plastic-slide-mounted sections to c. 30 μm. The undecalcified sections were stained with Levai-Leczko stain (Donath, 1988) and documented on a Nikon Eclipse 800 light microscope (Nikon Co., Tokyo, Japan).
When taken from the aquarium, the newts immediately squirmed rather extensively and tried to escape. After a few seconds, the animals calmed down and oriented themselves in the new environment. When ‘predator-like stimulation’ was applied, the animals tried to escape again. As no escape was possible, the newts took on an immobile position and began to actively secrete a milky and viscous secretion onto the body surface (Fig. 1a). This secretion appeared mainly on the neck, the dorsal and lateral trunk and on the tail. Additionally, all tested adult newts stretched the skin of the lateral trunk warts with the sharply pointed rib tips while holding an immobile arched (as shown in Fig. 1b) or flat body position (as shown on the radiograph in Fig. 2b). Pleurodeles waltl typically have eight to 10 such orange warts on the trunk sides. These warts correspond with the position of the directly underlying ribs. The stretched skin of the lateral warts often appeared to be pierced by the rib tips (Fig. 1b).
The X-ray analysis before and after the stimulation clearly showed the changed position of the vertebrae and the ribs (compare Fig. 2a and b and Fig. 2c and d). In the relaxed position, the vertebral column shows its natural, slightly curved configuration and the ribs are posteriorly oriented (Fig. 2a). After stimulation, the vertebral column is held rather straight – relative to the body axis – and the ribs are moved forward (Fig. 2b). The P-values of the rib angles showed significant differences in terms of the stimulus (before and after the stimulus; P-value=6.56e−21), but further significant differences were also recorded regarding the side (right vs. left; P-value=7.35e−3) and the individual (individuals 1–4; P-value=3.2e−4). Thus, not only the stimulus influenced the mean angle but also the side measured. The two measures per individual did not affect the mean angle: there was no significant difference between measures A and B (P-value=0.968).
CT scans showed the rib morphology (Fig. 3a–c). The ribs are connected to the corresponding vertebra by a well-developed, two-headed joint. This joint is composed of the articulations between the tuberculum and diapophysis and between the capitulum and parapophysis (Fig. 3c). The ribs are slightly curved in the transversal and horizontal axis (Fig. 3a, b). While the proximal three-fourths of the ribs run posteriorly and slightly ventrally (Fig. 3a, b), the distal fourth is slightly curved and the distal ends are directed dorso-laterally (Fig. 3a). The rib tips are sharply pointed (Fig. 3a, b).
Histological investigations showed the position of the ribs of a post-stimulated adult male P. waltl. The ribs are extended through and beyond the vertical and horizontal myosepta and are embedded in muscle tissue (M. dorsalis trunci complex dorsally and M. subvertebralis complex ventrally). Even if the animals relaxed their trunk musculature during anaesthesia, some rib tips still penetrated the lateral body wall completely (Fig. 4a, c). A thick layer of loose connective tissue is visible where the body wall is pierced (Fig. 4a). The proximal three-fourths of the ribs are filled with fat tissue composed of inflated fat cells (Fig. 4b). In contrast, the distal one-fourth is composed of massive bony tissue (Fig. 4c). The rib tip, which ruptures the lateral body wall, is coated by a thick and dense periosteum (Fig. 4c).
Amphibians have evolved many antipredator adaptations (for an overview, see Duellman & Trueb, 1994; Stebbins & Cohen, 1997). Among active (e.g. biting) and passive (e.g. mimicry) behaviour, the ability to produce skin secretions seems to be an important factor for amphibians to avoid being eaten. Skin secretions can be used to make the animal's surface slippery to promote escape (Stebbins & Cohen, 1997) or to make it sticky to immobilize a predator (Duellman & Trueb, 1994; Evans & Brodie Jr, 1994). Skin secretions can also be unpleasant tasting or irritating to mucous membranes, making the amphibian unpalatable to predators (Brandon, Labanick & Huheey, 1979; Brodie Jr, 1983; Duellman & Trueb, 1994). Some amphibian skin glands even synthesize noxious or poisonous substances to truly harm or kill would-be predators (e.g. Furlotti, 1910; Vialli, 1934; Nowak & Brodie Jr, 1978; Brodie Jr, 1983; Brodie Jr & Smatresk, 1990; Erspamer, 1994; Daly, 1995; Delfino et al., 1995; Tsuruda et al., 2002; Daly, Spande & Garraffo, 2005). The skin secretion of P. waltl is harsh and irritating to human mucous membranes (E. Heiss, pers. obs.), but lethal even in small amounts if injected intraperitoneally into mice (Nowak & Brodie Jr, 1978). In addition to the poisonous skin secretion, P. waltl uses its protruded, sharply pointed ribs as mechanical weapons to decrease palatability.
Spiny structures used as weapons to actively repel predators are broadly known among vertebrates: dermal spines (e.g. Pawlowsky, 1927; Bosher, Newton & Fine, 2005), teeth (e.g. Takahashi & Blanchard, 1982; Husak et al., 2006), claws (e.g. Gonyea & Ashworth, 1975; Blackburn, Hanken & Jenkins Jr, 2008) and horns or antlers (e.g. Clutton-Brock, 1982; Price & Allen, 2004). The use of ribs as ‘concealed’ weapons, however, is known only from two phylogenetically closely related salamander genera: Pleurodeles and Echinotriton (Leydig, 1879; Nowak & Brodie Jr, 1978; Brodie Jr, 1983; Brodie Jr et al., 1984). Echinotriton andersoni protrudes its ribs bearing 0–3 dorsally projecting epipleural processes (Nussbaum & Brodie Jr, 1982; Brodie Jr et al., 1984), and a spine projecting laterally from each quadrate bone (Brodie Jr et al., 1984). Pleurodeles waltl only protrudes its (always unbranched) ribs. The ribs of P. waltl are comparatively longer (rib length relative to body length) than those of most other salamandrids (Nowak & Brodie Jr, 1978). While the proximal two-thirds are contacted by fibres of dorsalis trunci musculature, the distal third is surrounded loosely by a connective tissue sheath (according to Nowak & Brodie Jr, 1978). This connective tissue with its loosely arranged collagen fibres possibly advances and accelerates closure of the self-induced wound. We were unable to find a lymphatic sheath in which the distal third of the rib should lie, as observed by Leydig (1879). In response to a threatening stimulus, the ribs are rotated forward by (mainly but not exclusively) dorsalis trunci musculature to a maximum angle of 92° relative to the longitudinal axis of the corresponding vertebra. The rib tips are positioned immediately beneath the lateral orange warts (Nowak & Brodie Jr, 1978), and no pores were found there even if multiple rib penetrations were observed on one wart. The ribs do not pierce the skin passively as suggested by Leydig (1879) due to lateral movements of the animal, but actively during defence. The orange warts provide a potential aposematic signal that would help make the ribs more noticeable.
With regard to forward rotation (e.g. from 27° to 92° relative to vertebrae axis), we propose that the rib mobility is highly significant during the ‘antipredator posturing’. Our results showed that the tested individuals repeatedly showed similar reactions (measured based on rib angle difference before and after stimulus) to the same stimulus. On the other hand, different individuals reacted differently to the same stimulus. This implies that the intensity of the reaction is dependent on the individual: the individual itself seems to react stereotypically depending on the degree of stress. The significant rib angle differences regarding the sides (right vs. left) may be of less importance. However, as only soft stimulations were applied, and predators seldom attack gently, the reported measurements probably do not represent the full behavioural response that has been evolutionarily selected. Further studies including observations of interactions of P. waltl and its predators in the wild would help understand the full range and effectiveness of antipredator responses.
The ability in Pleurodeles to use ribs as ‘spines’ requires certain morphological adaptations. The construction of the two-headed costo-vertebral joint constrains dorso-ventral deflexion but still enables a forward rotation of the rib at over 90° relative to the longitudinal vertebral axis. Two-headed joints are known from most urodeles studied so far (for an overview, see Starck, 1979; Duellman & Trueb, 1994), suggesting that urodeles in general are able to move their ribs in the posterior–anterior direction, possibly to expand the overall size of the body in a defensive posture. However, this ability is of special significance in the Spanish ribbed newt due to the newly gained function in defence. A smooth and stable movement of the ribs, enabled by two-headed costo-vertebral joints, may be advantageous when ribs are rotated forward to stretch the skin – in the case of P. waltl– to the point of piercing it.
The proximal three-fourths of the ribs in Pleurodeles are filled with fat tissue, but the protrusible distal one-fourth is built up by massive bone, possibly to improve mechanical stability and decrease the probability of fractures. The protrusible tip is also coated with a thick periosteum. This tough sheath could also function as a physical barrier against pathogens when the rib is protruded.
Amphibians have an extraordinary ability to repair their skin, whereby antimicrobial peptides provide direct protection against certain bacterial, fungal and protozoan pathogens (for an overview, see Zasloff, 1987, 2002; Schadich, 2009). Pleurodeles waltl not only lives in a wet, microbially contaminated environment but also lacerates its skin during defence. Antimicrobial peptides, released from specialized cutaneous glands (Schadich, 2009), could be of special importance because dangerous infections through the wounds caused by rib protrusions seem to be avoided.
The skin secretion of P. waltl also contains some poisonous components (Nowak & Brodie Jr, 1978; Heiss et al., 2009) that passively may seep into the body through the self-induced wounds, and yet we observed no self-intoxication by the newts. We therefore assume that P. waltl is immune against its own toxins. The high tolerance of urodeles against their own toxins has been demonstrated by Brodie Jr & Gibson (1969). They showed that Ambystoma gracile and Taricha granulosa were tolerant to the intraperitoneal injection of their own skin secretion, but reciprocal injections were lethal even in small amounts for both species.
The clade within the Salamandridae that comprises the three genera Pleurodeles, Echinotriton and Tylototriton is known to be monophyletic – with Pleurodeles as a sister group to the branch that includes Echinotriton and Tylototriton (Weisrock et al., 2006). Interestingly, while Pleurodeles and Echinotriton protrude their ribs, Tylototriton does not (Nowak & Brodie Jr, 1978; Brodie Jr, 1983; Brodie Jr et al., 1984). It seems, therefore, that the use of ribs as concealed weapons within this monophyletic clade is ancestral rather than derived. Only Tylototriton has lost this ability through time. However, to confirm this statement, the detailed mechanisms in Echinotriton and Tylototriton need to be studied similarly.
The authors acknowledge Stefan Tangl and Martina Scheuer (both Medical University Vienna) for kind help in histology, Michael Stachowitsch (University of Vienna) for improving our English, Günter Schultschik (Salamanderland, Kaltenleutgeben) for generously supplying material and Thomas Schwaha and Stefan Handschuh (both University of Vienna) for help in 3D imaging. Furthermore, two anonymous reviewers are acknowledged for their helpful comments. This study was supported by the University of Vienna research grant B-107 and by the Austrian Science Fund FWF P20094-B17.
- 2008). Concealed weapons: erectile claws in African frogs. Biol. Lett. 4, 355–357. , & (
- 2005). The spines of the channel catfish, Ictalurus punctatus, as an anti-predator adaptation: an experimental study. Ethology 112, 188–195. , & (
- 1979). Relative palatability, defensive behavior, and mimetic relationships of red salamanders (Pseudotriton ruber), mud salamanders (Pseudotriton montanus), and red efts (Notophthalmus viridescens). Herpetologica 35, 289–303. , & (
- 1983). Antipredator adaptations of Salamanders: evolution and convergence among terrestrial species. In Plant, animal and microbial adaptations to terrestrial environment: 109–133. Margaris, M.N., Arianoutsou-Faraggitaki, M. & Reiter, R.J. (Eds). New York: Plenum Publishing Corporation. (
- 1969). Defensive behavior and skin glands of the northwestern salamander, Ambystoma gracile. Herpetologica 25, 187–194. & (
- 1984). Antipredator adaptations of Asian salamanders (Salamandridae). Herpetologica 40, 56–68. , & (
- 1990). The antipredator arsenal of fire salamanders: spraying of secretions from highly pressurized dorsal skin glands. Herpetologica 46, 1–7. & (
- 1982). The function of antlers. Behaviour 79, 108–125. (
- 1995). The chemistry of poisons in amphibian skin. Proc. Natl. Acad. Sci. USA 92, 9–13. (
- 2005). Alkaloids from amphibian skin: a tabulation of over eight-hundred compounds. J. Nat. Prod. 68, 1556–1575. , & (
- 1995). Post-golgian maturative processes during the biosynthesis of poison secretion in cutaneous glands of the European common toad Bufo bufo. J. Nat. Toxins 4, 97–113. , , & (
- 1988). Die Trenn-Dünnschliff-Technik zur Herstellung histologischer Präparate von nicht schneidbaren Geweben und Materialien. Präparator 34, 197–206. (
- 1994). Biology of amphibians. Baltimore, MD: The Johns Hopkins University Press. & (
- 1994). Bioactive secretions of the amphibian integument. In Amphibian biology, volume 1: the integument: 178–350. Heatwole, H. & Barthalmus, G.T. (Eds). Chipping Norton, NSW: Surrey Beatty & Sons. (
- 1994). Adhesive strength of amphibian skin secretions. J. Herpetol. 28, 499–502. & (
- 1910). Ulteriori ricerche sulle ghiandole cutanee di Triton cristatus. Ital. J. Anat. Embryol. 9, 274–297. (
- 1975). The form and function of retractile claws in the Felidae and other representative carnivorans. J. Morphol. 145, 229–238. & (
- 2009). Three types of cutaneous glands in the skin of the salamandrid Pleurodeles waltl. A histological and ultrastructural study. J. Morphol. 270, 892–902. , , , & (
- 2006). Bite-force performance predicts dominance in male venerable collard lizard (Crotaphytus antiquus). Copeia 2006, 301–306. , , & (
- 1879). Die Rippenstacheln des Pleurodeles waltlii. Arch. Naturgesch. 45, 211–234. (
- 1978). Rib penetration and associated antipredator adaptations in the salamander Pleurodeles waltl (Salamandridae). Copeia 1978, 424–429. & (
- 1982). Partitioning of the salamandrid genus Tylototriton Anderson (Amphibia: Caudata) with a description of a new genus. Herpetologica 38, 320–332. & (
- 1927). Mit giftigen Hautdrüsen ausgerüstete Fische, die durch ihren Stich vergiften. In Gifttiere: 108–138. Pawlowsky, E.N. (Ed.). Jena, Germany: Gustaf Fischer Verlag. (
- 2004). Exploring the mechanisms regulating regeneration of deer antlers. Philos. Trans. Roy. Soc. Lond. 359B, 809–822. & (
- 2009). Skin peptides activity against opportunistic bacterial pathogens of the African clawed frog (Xenopus laevis) and three Litoria frogs. J. Herpetol. 43, 173–183. (
- 1979). Skelett des Rumpfes. In Vergleichende Anatomie der Wirbeltiere auf evolutionsbiologischer Grundlage. Band 2. Das Skelettsystem: 34–118. Starck, D. (Ed.). Berlin: Springer-Verlag. (
- 1997). A natural history of amphibians. New Jersey: Princeton University Press. & (
- 1982). Attack and defense in laboratory and wild Norway and black rats. Behav. Process 7, 49–62. & (
- 2002). Secretory glands of tetrodotoxin in the skin of the Japanese newt Cynops pyrrhogaster. Toxicon 40, 131–136. , , , , & (
- 1934). Ricerche istochimiche sul veleno cutaneo degli anfibi. Boll. Soc. Ital. Biol. Sper. 9, 600–602. (
- 2006). A molecular assessment of phylogenetic relationships and linage accumulation rates within the family Salamandridae (Amphibia, Caudata). Mol. Phylogenet. Evol. 41, 368–383. , , , , , , , & (
- 1987). Magainins, a class of antimicrobial peptides from Xenopus skin: isolation, characterisation of two active forms, and partial cDNA sequence of a precursor. Proc. Natl. Acad. Sci. USA 84, 5449–5453. (
- 2002). Antimicrobial peptides of multicellular organisms. Nature 415, 389–395. (