Poison Hunters

Tim Paine

For years explorers have searched the Neotropics for extraordinary finds, lost cities of gold, lands of warrior princesses, mythical beasts, and most recently the source of deadly frog poisons. It was said that frog poisons were used by Indian tribes for hunting and were powerful enough to kill a man several times. The poisons gave the group of frogs from which they were derived the name of poison-dart frogs, poison-arrow frogs, and dart poison frogs. All of these names could however be considered an exaggeration. 

This family, Dendrobatidae, currently consists of less than 200 species with well over 100 of these species being in the non-toxic* genera, Colostethus¹. The brightly coloured frogs in the genus Dendrobates are assuredly toxic but are not considered lethal, at least to humans. In fact, only three species of frogs in the five species genera, Phyllobates, were used to poison darts². This use was also limited to the western Colombian tribes of Noanama and Embera Choco³ whereas most arrow and dart poisons used by other tribes are plant-based curares from the Strychnos genus². 

Over the last few decades nearly 500 chemicals have been isolated from Dendrobatids⁴. These chemicals are in a group called alkaloids, generally plant-derived chemicals, which contain nitrogen rings. 

Alkaloids frequently have very strong effects when administered to animals. Nicotine, cocaine, morphine, and strychnine are all examples of alkaloids⁵. The hunt for these poisons began in earnest in the early 1960's by Dr. Bernhard Witkop and has been joined by a number of researchers, primarily Dr. John W. Daly and colleagues of the National Institutes of Health. The work began with the discovery of batrachotoxin in Phyllobates bicolor and P. aurotaenia and quickly expanded to include several other major classes of Dendrobatid alkaloids². Currently there are 22 structural classes4. So what are these classes, what do they do to make them toxic, and where do they come from?

These structural classes are based on the number of rings, the placement of the nitrogen atom, and other chemical components such as presence and number of hydroxyl (OH-), methyl (CH3-), or other side chains (R-, R'-). Five of these 22 classes are considered major classes and they are batrachotoxins, pumiliotoxins, histrionicotoxins, indolizidines, and decahydroquinolines. Some classes have many constituent alkaloids and may be widespread in Dendrobatids⁶. Others may be limited in variety e.g. stereochemistry, or frequency. The batrachotoxins are highly toxic and found only in frogs in the genus Phyllobates and strangely enough in feathers of New Guinean birds in the genus Pitohui⁷. Pumiliotoxin and its subclasses are also very toxic and capable of causing death in mammals in high enough doses⁷. Indolizidines and decahydroquinolines cause locomotor difficulties in mice, while histrionicotoxins require much higher doses to elicit similar responses⁷.

The efficacy and importance of these chemicals is based on vertebrate physiology. Animal muscle and nerve cells act as miniature batteries. The cells utilize ions such as potassium (K+), sodium (Na+), chlorine (Cl-), and calcium (Ca2+) that, when not in a state of equilibrium within the cell, cause a slight electrical current. 

Embedded within the surface and interior membranes of these cells are ion channels and 'pumps'. The channels can be opened and closed based on chemical imbalances, gradients, and receptor sites. Furthermore, the pumps can move these ions against concentration gradients. This complex arrangement is what allows nerve cells to 'fire' and muscle cells to contract and relax. Alkaloids, and Dendrobatid toxins in particular, interfere with the channels and pumps in cell walls causing loss of function or hyperfunction of cells. For example muscle cells can twitch excitedly to the point of spasm and permanent contraction.

Batrachotoxin is one of the most toxic natural compounds known³. It has an extremely high affinity for the sodium channels in cell walls. The presence of batrachotoxin in muscle cells causes the sodium channels to remain open allowing an influx of sodium ions into the cell. The muscle cells, especially heart muscle, rapidly contract and remain contracted causing cardiac arrest within minutes⁷.  Batrachotoxin is so toxic that it is estimated that somewhere between 100-300 micrograms (µg) will cause death in humans³. The average adult P. terribilis contains over 1100 µg of batrachotoxins⁸. 

Pumiliotoxin causes the release of calcium ions from within the sarcoplasmic reticulum, another membrane system within muscle cells. Pumiliotoxin also prevents the return of calcium ions to storage sites⁹. The effects of pumiliotoxin are muscle contractions, locomotor difficulties, salivation, uncontrolled chewing, and other such physiological responses⁷. 

Histrionicotoxins interfere with the sodium and potassium channels in muscle cells as well as acetylcholine receptors in nerve cells⁹. The decahydroquinolines block acetylcholine receptors in muscle cells. The effect, though very weak, prevents muscle contraction⁷. The effects of the indolizidines have been little studied but this large and varied class appears to effect, among other potential sites, nicotinic receptors in muscle and nerve cells⁷.

To fully utilise this cornucopia of chemistry each frog could contain dozens of toxins². It was initially assumed that the frogs synthesized these alkaloids completely or from precursor enzymes⁸. Granular glands and their structure were discovered in a number of den-drobatid species and it was hypothesised that these glands could hold the enzymes to synthesise the toxins². 

Over the years the hunt for the frog poisons posed a number of questions that eventually led away from this theory. Wild caught P. terribilis still had very high levels of batrachotoxin in their system over 6 years after their capture². Captive born P. terribilis from the wild parents showed no batrachotoxin in their skins. The long time lapse after capture would certainly lead credence to the theory of frogs producing their own toxins. But why did the offspring not contain batrachotoxin? 

It had been known that different populations of frogs, even of the same species, had varying profiles of toxins². Daly studied three populations of Dendrobates auratus from Panamá and Costa Rica and found that although each population had its own fingerprint profile variance could even occur on an individual level¹⁰. Additionally, it appeared the further apart the populations the fewer the shared toxins². Daly then analysed the toxins in a population of D. auratus introduced into Hawaii in 1932 and found a large profile divergence between the supposed founder stock in Panamá. In fact, the Hawaiian population was completely lacking in some of the major alkaloids from the Panamanian stock¹⁰. More extensive studies of captive raised animals of a number of Dendrobatid species failed to detect any alkaloids¹⁰. When some of these captive animals were raised in outdoor terraria they did produce profiles similar to the wild stock of the same locale but at reduced levels.

By the early 1990's the problems began to point to a dietary component to the poisons. Daly fed fruit flies dusted with isolated alkaloids extracted from the skins of wild frogs and found that these chemicals were taken up into the skins of the captive animals in the genera Dendrobates, Epipedobates and Phyllobates¹¹. The alkaloids were not detected in internal organs and tissues, however¹¹. Furthermore, Colostethus frogs did not take up these alkaloids, a mirror of their wild life lack of toxins.

These results provided some answers to results seen years previously. This discovered uptake system could show how wild caught animals could maintain their alkaloids for many years. It also showed that the frogs were immune to chemicals that are toxic to other animals, an expansion of an earlier observation that tissues from frogs in the genus Phyllobates showed no reaction to the presence of batrachotoxin .

So now the thought was shifting from internal synthesis to dietary origins and those origins needed to be discovered. Some of the lesser Dendrobatid alkaloids such as pyrrolizidines, indolizidines, pyrrolidines, and piperidines had been isolated from ants and other lesser alkaloids were found in certain beetles and millipedes but none of the major classes had been detected in arthropod prey¹². Daly still considered the possibility that the frogs were ingesting building block chemicals and that further synthesis was occurring¹². 

Daly and colleagues investigated further by capturing Dendrobates auratus tadpoles from a population in Panamá and raising them. The metamorphosed frogs were fed leaf-litter arthropods collected from a second locale in Panamá while control frogs were fed flightless fruitflies. When they were later sacrificed it was found that the frogs raised on leaf-litter insects contained a significant amount of alkaloids while the control frogs contained none¹². Interestingly the frogs showed alkaloids profiles similar to the frogs from the region from where the leaf-litter was accumulated and not from the parent stock of their collected location¹². 

In time a shortlist of minor alkaloids shared between Dendrobatid frogs and arthropods was generated¹³. Unfortunately, this list, as of a 2000 published study, only consisted of 22 of the near 500 Dendrobatid alkaloids. These studies are certainly pointing to a dietary source for the vast array of chemicals, but where are they coming from? Hindering the search is the fact that arthropod fauna are extremely diverse and tiny or microscopic. The most recent published progress by Daly and colleagues examined arthropod collections from a number of sites in Panamá and extracted the alkaloids. While the arthropod taxa were too diverse to isolate the exact source of the extracted chemicals the researchers were able to detect two of the pumiliotoxins and an indolizidine, both major classes of Dendrobatid toxins⁴. These data are showing promise that bioprospecting of arthropod taxa will reveal the sources of many, if not all, of the Dendrobatid toxins.

Interestingly, as Dendrobatid toxins have been found in frogs and toads from Madagascar, South and Central America and Australia^{14, 15 & 16}  as well as the previously mentioned Pitohui bird from New Guinea, the arthropod source puzzle should shed some light on the broader puzzle of evolutionary theory.

* Colostethus inguinalis was shown to contain the potent toxin tetrodotoxin in its skin¹⁷.

References :-
I. Grant, T. and M.C. Ardila-Robayo. 2002. A new species of Colostethus (Anura: Dendrobatidae) from the eastern slopes of the Cordillera Oriental of Columbia. Herpetológica 58(2): 252-260.

2. Daly, J.W., G.B. Brown, M. Mensah-Dwumah, and C.W. Myers. 1978. Classification of skin alkaloids from Neotropical poison-dart frogs (Dendrobatidae). Toxicon 16: 163-188.

3 Myers, C.W., J.W. Daly, and B. Malldn. 1978. A dangerously toxic new frog (Phyllobates) used by Embera Indians of western Colombia, with discussion of blowgun fabrication and dart poisoning. Bull. Am. Mus. Nat. Hist. 161(2):309-365.

4. Daly, J.W. et al. 2002. Bioactive alkaloids of frog skin: Combinatorial bioprospecting reveals that pumiliotoxins have an arthropod source. Proc. Nat. AcaD. Sci. 99(22): 13996-14001.

5. Solomons, T.W. 1988. Inorganic Chemistry. 4th ED. (Wiley and Sons, New York)

6. Daly, J.W., C.W. Myers, and N. Whittaker. 1987. Further classification of skin alkaloids from the Neotropical poison frogs (Dendrobatidae), with a general survey of toxic/ noxious substances in the amphibia. Toxicon 25(10): 1023-1095.

7. Heatwole, H. eD. 1994. Amphibian Biology: Vol. 1 The Integument (Surrey Beatty & Sons Pty Ltd., NSW Australia).

8. Daly, J.W., C.W. Myers, J.E. Wamick, and E.X. Albuquerque. 1980. Levels of batrachotoxin and lack of sensitivity to its action in poison-dart frogs (Dendrobatidae). Science 208: 1383-1385.

9. Myers, C.W, and J.W. Daly. 1983. Dart-Poison Frogs. Sci. Am. 248(2): 120-133.

!0. Daly, J.W. et al. 1992. Variability in alkaloid profiles in Neotropical poison frogs (Dendrobatidae): Genetic versus environmental determinants. Toxicon 30(8): 887-898.

II. Daly, J.W. et al. 1994. An uptake system for dietary alkaloids in poison frogs (Dendrobatidae). Toxicon 32(6): 657-663.

12. Daly, J.W. et al. 1994. Dietary source for skin alkaloids of poison frogs (Dendrobatidae)? J. of Chem. Ecol. 20(4): 943-955.

13. Daly, J.W, et al. 2000. Arthropod-Frog connection: Decahydroquinoline and pyrrolizidine alkaloids common to microsympatric myrmicine ants and Dendrobatid frogs. J. of Chem. Ecol. 26(1): 73-85.

14. Daly, J.W., R.J. Highet, and C.W. Myers. 1984. Occurrence of skin alkaloids in non-Dendrobatid frogs from Brazil (Bufonidae), Australia (Myobtachidae) and Madagascar (Mantellinae). Toxicon 22(6): 905-919.

Daly, J.W. et al. 1986, Alkaloids from Dendrobatid frogs: structures of two O-hydoxy congeners of 3-butyl-5-propylindolizidine and occurrence of 2,5-disubstituted pyrrolidines and a 2,6-disubstituted piperidine. J. of Nat. ProD. 49(2): 265-280.

15. Garraffo, H.M., et al. 1993. Alkaloids from bufonid toads (Melanophyriniscus): Decahydroquinoli-nes, pumiliotoxins and homopumiliotoxins, indolizidines, pyrrolizidines, and quinolizidines. J. of Nat. ProD. 56(3): 357-373.

16. Garraffo, H.M., et al. 1993. Alkaloids in Madagascan frogs (Mantella): Pumiliotoxins, indolizidines, quinolizidines, and pyrrotizidines. J. of Nat. ProD. 56 (7): 1016-1038.

17. Daly, J.W. et al. 1994. First Occurrence of tetrodotoxin in a Dendrobatid frog (Colostethus inguinalis), with further reports for the bufonid genus Atelopus. Toxicon 32(3): 279-285.

Other References
Myers, C.W. and J. W. Daly. 1976. Preliminary evaluation of skin toxins and vocalizations in taxonomic and evolutionary studies of poison-dart frogs (Dendrobatidae). Bull. Am. Mus. Nat. Hist. 157(3): 173-262.

Mensah-Dwumah, M. and J.W. Daly. 1978. Pharmacological activity of alkaloids from poison-dart frogs (Dendrobatidae). Toxicon 16: 189-194. 

Neuwirth, M., J. W. Daly, C.W. Myers and L.W. Tice. 1979. Morphology of the granular secretory glands in skin of Poison-dart frogs (Dendrobatidae). Tissue & Cell 11(4): 755-771.

Myers, C.W. and J.W. Daly. 1979. A name for the poison frog of Cordillera Azul, Eastern Perú, with notes on its biology and skin toxins (Dendrobatidae). Am. Mus. Novitates 2674: 1-24

Myers, C.W. and J.W. Daly. 1980. Taxonomy and ecology of Dendrobates bombetes, a new Andean poison frog with new skin toxins. Am. Mus. Novitates 2692: 1-23

Edwards, M.W., J.W. Daly, and C.W. Myers. 1988. Alkaloids from a Panamanian frog Dendrobates speciosus: Identification of pumiliotoxin-A and allopumiliotoxin class alkaloids, 3,5-disubstituted indolizidines, 5-substituted 8-methylindolizidines, and a 2-methyl-6-nonyl-4-hydroxypiperidine. J. of Nat. ProD. 51(6): 1188-1197.

Jain, P. et al. 1995. A new subclass of alkaloids from a Dendrobatid poison frog: a structure for de-oxypumiliotoxin 251H.   J. of Nat. ProD. 58(1): J GO-104.

Garraffo, H.M.P. Jain, T.F. Spande, and J.W, Daly. 1997. Alkaloid 223A: The first tri-substituted indolizidine from Dendrobatid frogs. J. of Nat. ProD. 60: 2-5