Faculty: William K. Hayes

Laboratory of behavioral ecology and conservation

William Hayes homepage


Research on Biological Roles and Variation or Snake Venoms (with photographs)

Snake venoms vary remarkably in composition. They are generally comprised of multiple proteins serving a wide range of functions. Venom differences are often dramatic within the range of a species. Several rattlesnake taxa, for example, produce a dangerous neurotoxin (Mojave toxin) in one portion of their range and not in other parts. Variation also occurs between individuals within a local population and even within a single snake as it ages. What explains this high amount variation? Is the variation random, as seems to be the case for many biochemical traits (e.g., human blood types), or is it adaptive? Though questions such as these pose a real challenge for us to answer, they fascinate me. I wish I had more time to explore them further, but here I will share what few insights I have learned from several interesting studies.

Adaptive variation of venoms  |  Distribution of Mojave Toxin

If variation among snake venoms is adaptive, then the venom must serve specific functions-or biological roles-that are subject to selective forces in the environment. What are the biological roles of venom? For snakes, we can view the functions of their venom from two broad contexts:

Predation - Snake venom can serve the following roles:

  • Rapid immobilization of prey
  • Rapid death of prey
  • Facilitation of prey relocation (for prey released after envenomation)
  • Accelerated digestion of prey (when proteolytic toxins are present)

Defense - Snake venom can serve the following roles:

  • Predator recognition that the snake is dangerous
  • Termination of encounter after painful envenomation

In other animals, venoms can serve these and even additional roles. For example, ants can rely on their venom for communication and the male platypus can use its venom for sexual competition (male-male fighting). The behavioral ecology of venom has largely been neglected!

The changes in venom composition associated with rattlesnake ontogeny (growth) illustrate nicely the adaptive roles that venom serves in snakes. Young rattlers, which possess only a small amount venom, produce a more toxic venom for killing relatively small, easily digested prey. Adults, in contrast, have much more venom available to kill, but because they feed on large, bulky prey that are more difficult to digest, they produce a venom that is less toxic and more proteolytic. Thus, from an ontogenetic perspective, rattlesnake venoms seem especially well-suited to meet the particular needs of the organism at any age. (Curiously, for some species that produce highly neurotoxic venom, toxicity may remain unchanged as the snake grows.)

I am interested in the extent to which snake venoms vary because of selection acting on specific functions of venom. Venom, for example, should be designed to optimally subdue a snake's preferred prey. If the prey species is difficult to kill (e.g., ectotherms, such as frogs and lizards), the venom should be more toxic. If the prey presents more of a challenge to digest (e.g., bulky endotherms, such as rodents and birds), the venom should have greater digestive capacity. As another example, venoms that serve a more defensive role should elicit much stronger pain than venoms used primarily for predation. Spitting cobras illustrate this hypothesis well, as their venoms have toxins that induce substantial pain when sprayed defensively into the eyes of a predator (that attacks to consume the snake) or antagonist (that does not intend to consume the snake). Non-spitting cobras lack these toxins, producing a venom with more toxic and less pain-inducing proteins that better serve a predatory capacity.

One could readily imagine that geographic variation in venom composition is related to diet. Several studies support this view, though they are based on an association of venom constituents or properties with preferred prey types. A more appropriate test of this hypothesis-that venom variation is largely adaptive-would be to examine whether venoms function most effectively when administered against preferred prey.

Relying on tests of function, I have examined the relative killing effectiveness of venoms from different taxa. Clearly, one major role of venom is to kill prey rapidly. However, different venom components may be required to effectively dispatch prey of different species. Several experiments offer the following conclusions:

The venom of the Midget-faded Rattlesnake (Crotalus concolor), when injected in biologically relevant doses (i.e. quantities that snakes actually use when biting), kills both lizards and mice more quickly than venom from the Prairie Rattlesnake (C. viridis). I hypothesized that the highly toxic venom of C. concolor, which contains a neurotoxin homologous to Mojave toxin, would be more effective killing lizards than C. viridis venom. I also hypothesized that C. viridis venom, being more proteolytic (digestive) in composition, would be more effective killing mice. However, C. concolor venom was more toxic for both prey types.  Source: unpublished study.

To test adaptive function of venom in a more comprehensive design, I compared the killing effectiveness of three venoms against their preferred prey types. The venoms were from the Cottonmouth (Agkistrodon piscivorus, a fish and amphibian specialist, though it also eats rodents), Rock Rattlesnake (C. lepidus, a lizard specialist that eats other ectotherms and some rodents), and Western Diamondback (C. atrox, a rodent specialist). The three prey types were Frogs (Hyla chrysocoelus), Lizards (Anolis sagrei), and Mice (Mus musculus). For time to prey death following injection of biologically relevant doses, there was a highly signficant interaction between venom type and prey species. Frogs were killed most rapidly by Cottonmouth venom, lizards were killed most quickly by Cottonmouth and Rock Rattlesnake venoms (Western Diamondback venom was least effective), and mice were most effectively killed by Cottonmouth and Western Diamondback venoms (Rock Rattlesnake venom was least effective). The statistical outcome supports the hypothesis that venom differences correspond to preferred prey types. Source: unpublished study.

Tests of lethality-primarily the LD50 test in mice-are widely represented as the definitive means for comparing relative lethality of various venoms. However, do LD50 values, obtained from injection of very minute venom quantities (micrograms) usually directly into a vein, correspond to biological reality? Snakes generally inject milligram amounts of venom into a wide range of tissues and/or organs. Previous studies by Harold Heatwole using sea snake venoms injected into various eel species suggested that relative lethality-and even mechanism of prey death-will vary depending on dose. Recent case studies of human envenomation (Sean Bush et al., unpubl. data) suggest that the neurotoxic venom of the Mojave Rattlesnake (C. scutulatus), universally considered to be highly lethal, may actually be less dangerous than the venom of Southern Pacific Rattlesnake (C. helleri), which normally has less-toxic LD50 values. In a simple study that needs to be repeated, I injected biologically relevant doses (mg) of three venom types into the right mid-dorsal region of mice (i.e., where the fangs often penetrate). To my surprise, the venom of C. helleri specimens lacking neurotoxicity actually killed mice significantly faster than the venoms of C. helleri and C. scutulatus specimens having Mojave neurotoxins (see Distribution of Mojave Toxin below for evidence that some C. helleri specimens possess neurotoxicity). This unexpected outcome turns conventional wisdom on its head, suggesting that venoms with Mojave toxin do not necessarily kill prey more rapidly when an appropriate predatory dose is injected. Source: unpublished study.

These studies illustrate the need to take a more ecological approach to the study of venom. Studies based exclusively on mice tell us next to nothing about the functional roles that venom serves. We need instead experiments that test function in actual prey or predators, using venom quantities that approximate the doses that snakes normally inject when feeding or defending themselves. LD50 values (derived from ug venom per g mouse) may model reasonably well the relative lethality of venoms in humans (mg venom per kg human), as the amount of venom injected in each case is very small relative to the victim's mass. However, don't be surprised to learn one day that some venoms may be much more lethal in humans than mouse-derived LD50 values suggest!   

Distribution of Mojave Toxin  |  Adaptive Variation of Venoms

Why do some rattlesnakes have more toxic venoms than others? Why do some have neurotoxins and others lack them? These questions are exceedingly difficult to answer. The highly varied effects of venom on human snakebite victims has generated considerable interest in the biochemical, pharmocological, and toxicological properties of venom. Because we like to simplify things, herpetologists have dichotomized snake venoms into two broad types, despite the fact that individual snakes can have components of both in their venom. The two venom types can be summarized as follows:

Neurotoxic venoms - These venoms consist primarily of neurotoxins that generally cause death by muscle paralysis. Their primary role appears to be producing rapid prey death, especially in prey that are highly resistant to venom (including ectotherms such as fishes, amphibians, and reptiles). In mammals, these proteins appear to be highly toxic (having low LD50 values) but do not inflict much pain.

Proteolytic venoms - These venoms consist of various protein-degrading enzymes that cause a wide range of toxic effects, including those resulting from cytotoxic (cell-destroying), hemotoxic (blood-destroying), myotoxic (muscle-destroying), and hemorrhagic (bleeding) activities. The term “proteolytic” encompasses a broader and more descriptive range of activities than the more widely and not-so-appropriately used terms “hemotoxic” and “cytotoxic.” Although toxic in their own right, these proteins contribute significantly to prey digestion and create tremendous pain and tissue damage when injected into a potential predator or antagonist.

In rattlesnakes, most taxa produce venoms that are strongly proteolytic with little to no neurotoxicity. However, there are some species that possess venoms with neurotoxins, and these are widely regarded to be the most dangerous (though some unpublished data may challenge this idea).  In general, rattlesnake venoms with neurotoxins have reduced amounts of proteolytic enzymes, though there are individuals with venoms that possess substantial neurotoxic and proteolytic activity.

Recently, I have become interested in the distribution of Mojave (or Mojave-like) toxin in several rattlesnake species. Mojave toxin is a large, basic protein with two subunits. Whereas the basic subunit is a highly neurotoxic phospholipase enzyme, the acidic subunit lacks lethality and appears to chaperone the basic subunit to specific binding sites. Both subunits must be present for the toxin to be lethal. Immunologically homologous toxins are found in a variety of rattlesnake taxa, and these are often given different names despite their virtually identical structure. We now know that a Mojave-like toxin is present in the following taxa:

Species Neurotoxin Distribution of toxin
Mexican West Coast Rattlesnake
Crotalus basiliscus basiliscus
Mojave-like toxin Present in at least some individuals examined
Midget-faded Rattlesnake 
Crotalus concolor
"Concolor toxin" Possibly entire range (Utah, 
Wyoming, Colorado)
South American Rattlesnake 
Crotalus durissus
"Crotoxin" Some but not all of range 
(Central and South America)
Southern Pacific Rattlesnake 
Crotalus helleri
Mojave-like toxin Only in Mt. San Jacinto, CA, 
area in US portion of range
Timber Rattlesnake 
Crotalus horridus
"Canebrake toxin"
Only in southern portion of its broad eastern US range 
Rock Rattlesnake 
Crotalus lepidus
Mojave-like toxin Only in western portion of its 
southwestern US range
Baja Speckled Rattlesnake 
Crotalus mitchelli mitchelli
Mojave-like toxin Distribution of toxin not known; absent from C. m. pyrrhus in U.S. 
Mojave Rattlesnake 
Crotalus scutulatus
"Mojave toxin" Much or all of U.S. range except for southcentral Arizona
Uracoan Rattlesnake 
Crotalus vegrandis
"Vegrandis toxin" I don't have details; please 
send to me if you know!
Sistrurus catenatus
"Sistruxin" Present in at least some eastern and western forms

This table illustrates the remarkably disjunct distribution of Mojave-like neurotoxins in rattlesnakes. Even within a single taxon, some populations produce the toxin and others do not. In time, one can expect more species to be added to this table. The presence of neurotoxin, however, does not necessarily mean that the snake's bite is especially dangerous. Some taxa express only small amounts of neurotoxin in their venom; hence, their venom and bite may be less toxic than those that produce large amounts in their venom.

Could there be other (non-phospholipase) neurotoxins in rattlesnake venoms? A smaller basic polypeptide found in some snakes, such as the Eastern Diamondback (C. adamanteus), has been implicated as a neurotoxin, but I believe the jury is still out on this. However, Sean Bush, MD, of Loma Loma University Medical Center has treated cases of Southern Pacific Rattlesnake (C. helleri) envenomation involving unambiguous neurotoxicity, yet the offending snake in one widely-publicized case clearly lacked Mojave-like toxin. Cases such as these suggest that an unidentified neurotoxin exists in some individuals of this species. We hope to eventually isolate and characterize the hypothesized toxin.

The first case of C. helleri-induced neurotoxicity that Sean treated triggered an idea. Investigators had previously screened a handful of C. helleri venom samples and reported the absence of neurotoxin. However, because this one particular specimen originated in the foothills ecotone where the ranges of C. helleri and C. scutulatus meet, perhaps this snake had Mojave toxin in its venom because of historic or recent gene exchange between the two species (an idea that Steve Grenard popularized in his provocative July-August 2000 article in Natural History). Thus, we sought to learn whether a Mojave-like neurotoxin exists in some specimens of C. helleri from southern California.

With help from colleagues (Sean Bush and Mike Cardwell), students, my very tolerant wife, and some cooperative snake owners/collectors, I collected a number of venom and blood samples from snakes throughout southern California and shipped them off to Eppie Rael and Wendy French at the University of Texas, El Paso. The venom was tested using monoclonal antibodies that bind specifically to Mojave toxin. The blood samples yielded DNA that could be tested to determine whether genes were present for both the acidic and basic subunits of Mojave toxin. Initially, we focused on C. helleri, but we later collected samples from Southwestern Speckled Rattlesnakes (C. mitchelli pyrrhus) and Tiger Rattlesnakes (C. tigris).

The results for C. helleri gave us a surprise. Some specimens indeed possessed a Mojave-like toxin, but none were from populations in close proximity to the range of the Mojave rattlesnake. They were all from a relatively small area (near Lake Hemet) on Mt. San Jacinto, in western Riverside County. All five specimens tested from this area possessed the toxin.

Although we are still examining samples from the Speckled Rattlesnakes, all of the Tiger Rattlesnake samples tested positive for Mojave venom (see Powell et al., 2004, J. Herpetol. 38:149-152.)

The phylogeographic distribution of Mojave-like neurotoxins supports the view that these toxins have either evolved independently multiple times in various rattlesnake lineages or were present in an ancestral form but have since become been lost from most lineages. Frankly, Grenard's idea that rattlesnake venoms are rapidly evolving to become more dangerous, aided by rampant hybridization, finds little support from what we know about the distribution of Mojave-like toxins. The adaptive value of having a highly toxic venom seems questionable, as most individual snakes actually produce less toxic venom as they grow. Nonetheless, much remains to be learned.

Hayes and snake
So what's in your venom? A Southwestern Speckled Rattlesnake (Crotalus mitchelli pyrrhus) yields a venom sample for science. I was relieved to see this phase of the project-venom-collecting-completed. The venom and a blood sample (containing DNA) will be sent to the laboratory of Eppie Rael at University of Texas, El Paso, to be screened for Mojave toxin (applying monoclonal antibodies to the venom) and the presence of Mojave toxin genes (using primers to the acidic and basic units). Photograph: Shelton S. Herbert.

Cardwell in lab
Mike Cardwell calmly prepares for data collection in the lab. His years of experience seem obvious to even the casual observer. The snake will give us venom and blood samples, as well as data on body size and distance between fangs. Photograph: Shelton S. Herbert.

Stick with the seabirds, Tony! Obviously, this US Virgin Islander (a.k.a. Tony Trimm) had no experience whatsoever with snakes during his deprived childhood. We're working on him... Photograph: Shelton S. Herbert.

William Hayes homepage

[Top of page] [Biology program]

[Department of Earth and Biological Sciences homepage]