Phenotypic variation in space and time of toxin levels in a chemically defended amphibian
Within a population or species, natural selection may push a phenotype to an adaptive peak; a process whereby genetic variation leads to evolutionary change. When interacting phenotypes evolve reciprocally in two species, a coevolutionary relationship develops and, in rare occurrences, the trait affecting the coevolutionary process can be both the agent and target of selection. One such example is the potent neurotoxin, tetrodotoxin (TTX), which sits at the interface of the ecological interaction between tetrodotoxic newts that are preyed upon by garter snakes. This relationship is a widely known, generally undisputed example of a coevolutionary parallel arms race. However, there is relatively little understanding of how newts control TTX, either at the proximate or ultimate level, whether newts can biosynthesize TTX or if it is the result of endosymbionts, and if TTX is genetically controlled and able to evolve in response to predation pressure. In light of these complexities, it seems pivotal to describe and more fully understand newts and the TTX phenotype.
Using multiple research approaches, we have assessed if TTX can affect evolutionary and ecological processes by (i) longitudinally quantifying TTX concentrations of newts within and between breeding localities and testing whether TTX levels vary spatially and temporally, (ii) performing field and laboratory bioassays to test if TTX has community-wide effects, and (iii) testing whether the trait responds to environmental changes. Five years of regular sampling of wild individuals and populations of newts across more than 30 sites spanning California indicate that individual and group mean TTX concentrations can fluctuate tremendously within a breeding season and between years. The results of field and laboratory experiments indicate that newt chemical cues and TTX do affect macroinvertebrate behavior, that adults can be induced to produce greater concentrations of TTX, and that environmental conditions affect TTX levels. Overall, individual and group mean TTX values indicate that TTX is not a static phenotypic trait, and in conjunction with experimental results, collectively provide a novel dimension to our understanding of TTX in nature, potentially requiring a reevaluation of the lockstep predator resistance model generally employed to explain the evolution of this potent toxin.