The Loom

Bryan%20full.jpgBack in February I discovered the remarkable work of Australian biologist Bryan Grieg Fry, who has been tracing the evolution of venom. As I wrote in the New York Times, he searched the genomes of snakes for venom genes. He discovered that even non-venomous snakes produce venom. By drawing an evolutionary tree of the venom genes, Fry showed that the common ancestor of living snakes had several kinds of venom, which had evolved through accidental “borrowing” of proteins produced in other parts of the body. Later, these genes duplicated to create a sophisticated cocktail of venoms–a cocktail that varied from one lineage of snakes to another.

As I report tomorrow in the Times, Fry has taken this research the next logical step. He set out to find out when that ancestral venom evolve. In his search, Fry made an astonishing discovery: snakes are 100 million years old, but snake venom is 200 million years old. This conclusion arose from the fact that some lizards produce the same sorts of venom as snakes–including this desert spotted monitor that Fry is posing next to on one of his venom expeditions in the Outback.

I find that these stories make the most sense if I can map them onto an evolutionary tree. Fortunately this study (published online at Nature) comes with a particularly nice one, which I’ve reprinted here. It shows how snakes are related to lizards, based on a new large-scale study of DNA. Snakes descend from a close common ancestor with monitor lizards, gila monsters, iguanas, and other related species. They’re more distantly related to skinks, gekkos, tuataras, and other less well known lizards.

Fry%20lizard%20tree.jpg

The venomous species did not turn up on random branches of the tree. Instead, they all belong to the same branch, marked on the tree with the name “venom clade.” (Clade just means a group of species sharing a common ancestor.) The venoms shared by all of them are marked in red and brown. Red-marked venoms are produced in glands on the top and bottom of the mouth, and brown in the top. Snakes then added on 17 other kinds of venoms. Other lizard clades have venoms of their own, in other marked in yellow. The tree also shows how mouth glands evolved at the same time, beginning as a way to make prey slimey enough to go down the throat to a way to pour venom into a wound.

The story revealed in this tree reveals how a 200-million year old lizard species evolved venom that helped it disable prey long enough to kill them. Over time, new kinds of venom emerged before the common ancestor of living snakes and their close relatives. Snakes later became venom specialists, in some cases evolving a bite that was even fatal to humans and other big animals. At the same time, though, other lizards were acquiring some new venoms of their own, probably as an adaptation to new kinds of prey. It’s a process that’s going on today, as reflected in different cocktails of venom found in different populations of the same species of lizard.

It will be interesting to see what happens as Fry and others add more branches to this tree. For example, pythons are not known to be venomous. Did they lose venom as they became adept at constrictors? And how did the function of the venoms change with time? A Komodo dragon and a coral snake occupy very different ecological niches, so they need to manipulate their prey in different ways.

Incidentally, I would stress that this discovery does not mean that your pet iguana is going to strike you down tomorrow morning. Lizard venoms are sophisticated, but they typically come in such small doses that they won’t cause you any significant harm. And if you do happen to get bit by a Komodo dragon, you’ll be able to be distracted from the effects of its venom by the fact that your arm is missing.

Comments

  1. #1 Henry
    November 21, 2005

    A few details here and there on this:

    First, the animal in the picture isn’t a lace monitor (V. varius), but is actually a Sand Monitor / Goana (possibly V. panoptes or gouldi or flaviurus; the taxonomy of this group is utterly insane). Lacies are much prettier, IMHO (google images and see what I mean).

    Secondly, they’re interesting results, but I’m somewhat uncomfortable with the phylogeny, since if you actually read the footnotes, they only test 5 lizard species and 4 of those were from the same genus, Varanus. The only non-varanid was a lone Iguanian (the super-family, part of family agamidae). As such, it’s both too early to write off most of those other branches as not having venom, and too soon to start claiming clades without knowing the extent to which convergence of these protiens has occured over the 3000 odd lizard species (in 20-some-odd families).

    Another thing that leapt to mind for me when reading was “so what?”, since they don’t address the relevance to the ecology/behavior of the species. It’s clearly not for prey capture; the lone iguanian eats insects and plants (and just crunches them and gulps them down), and varanids are among the last things needing help dispatching prey (all species analyzed mostly eat small mammals, and they can dispatch those easily; I’ve personally cleaned rabbit intestines off the *ceiling* of a room-sized cage).

    Furthermore, the Komodo strongly indicates against the potency of these chemicals, since, if it had something even approximating useful venom, why didn’t it simply enhance that (as Gilas did) to subdue prey rather than evolving a totally distinct, bacteria-based mechanism? The fact that these two took totally separate routes to the same effective end from the same starting points indicates to me that the situation is substantially more complex that it first appears.

    Another question I have concerns his previous work as applied to this: if venom protiens are recruited from protiens elsewhere, shouldn’t we expect a high degree of convergence, if this recruitment occurs often (which I’d argue that the diversity of snake venoms that have evolved since the Miocene indicates it does)?

    Personally, it’s neat that he’s found these protiens in lizard species, but I think a lot more work needs to be done before forming clades and assessing the role of these protiens in squamate evolution.

    Oh and as for constriction vs venom, *usually* they’re not found together, but in some species, notably Boiga irregularis (the famous Brown Tree Snake of Guam), both constriction and venom are used, with the former for mammals such as mice and birds, and the latter for skinks and geckos. But this species is only weakly venomous (as I’ve directly found out), and is what is commonly called “rear-fanged”.

    Sorry, I just can’t resist talking about reptiles.

  2. #2 John Wilkins
    November 21, 2005

    Is there any further discrimination within the Serpentes clade of how the various kinds of venoms evolved? All those toxins must be more or less monophyletic.

  3. #3 Dr. Bryan Fry
    November 22, 2005

    G’day Henry,

    A couple corrections:

    Carl correctly pointed out that that the animal in the picture is a desert spotted monitor (Varanus panoptes rubidus). He didn’t say it was a lace monitor.

    >evolving a totally distinct, bacteria-based mechanism

    Actually that is the entire point. They have venom, not bacteria. The whole bacteria thing has been dogma but a complete red herring. The effects produced by goanna bites (such as rapid swelling, pain and prolonged bleeding, persisting for several hours) are totally inconsistent with bacterial effects. In contrast, we did bioactivity testing of crude varanid venom as well as purified toxin type and demonstrated bioactive effects.

    In regards to the glands, the histology has been done previously by one of my co-authors (Elazar Kochva) on all the various lineages. Only the iguanians and anguimorpha of all the lizards have protein secreting glands. The protein secreting glands are not saliva glands, they are newly created structures. There is no homologous gland found elsewhere.

    As far as convergence of proteins, convergent selection of the same protein type (e.g. PLA2) would result in toxin clades that are not reciprocally monophyletic in relation to related non-toxin proteins. In contrast, a single shared origin would result in a single toxin clade that is monophyletic in relationship to related non-toxin proteins but not monophyletic for the taxa from which the toxins were sequenced. For more information on this approach, have a read of our 2004 ‘Assembling an Arsenal’ paper published in Molecular Biology and Evolution. In this paper we used this test to show the single origin of venom in snakes. The same methodology was applied here to show the single origin of venom shared between snakes and lizards.

    In regards to toxicity of colubrid venoms, pull a couple of our colubrid venom papers from pubmed. Neurotoxic effects of Boiga venom is actually equipotent to highly toxic elapids such as Acanthophis. The key is that the venom yield is smaller and the delivery less efficient. For animals such as Boiga that are feeding mostly on soft-skinned, non-dangerous prey items this is sufficient for them.

    We are now investigating the relationships to prey items and actually finding the same sorts of regional variation in venom composition that we have previously with snakes.

    Cheers
    Bryan

  4. #4 coturnix
    November 22, 2005

    Just a sidenote – Tuataras are not lizards and are used here as an outgroup.

  5. #5 Dave
    November 22, 2005

    Further to Henry’s comments above on the phylogeny, it should also be noted that this is based on DNA evidence alone. Phylogeneticists will happily tell you of the common problem that afflicts DNA-based phylogenies in that they often look VERY different to those based on morphological characteristics. This is not to say that either is right or wrong, per se, but snakes are notorious for being very difficult to trace their origins.

    Still a fascinating piece of research and worthy of its publication in Nature.

  6. #6 RyanJW
    November 22, 2005
  7. #7 Carl Zimmer
    November 22, 2005

    For some reason I’m not getting comments sent to me by email, and so I’m a bit slow these days on the responses. The lace monitor/spotted monitor confusion is my fault–I put lace in the original post, then changed it to the correct name and then realized that I should have used the strikethrough function to show that it had been updated. Arg. Anyway, now it’s right. I’ll fix the tuatara language too.

    Thanks to Bryan for getting into the gorey details.

    For those who want more on the lizard/snake phylogeny, check out this paper by two of the co-authors that’s in press. PDF here: http://evo.bio.psu.edu/hedgeslab/Publications/PDF-files/171.pdf

  8. #8 Cameron Peters
    November 22, 2005

    Just as added interest, Nicolas Vidal and colleagues have a new paper that really shakes up the genealogical tree of reptiles (including snakes). It is in the current issue of C.R. Biologies (easily found on PubMed). Very interesting.

  9. #9 David B
    November 23, 2005

    Are there any verified cases of mammals with venom? And if not, why not?

    I think I’ve read somewhere that a few shrew species have venomous saliva. And the male platypus has a venomous spike on its feet, used in male-male combat.

    Perhaps venom in snakes, insects, etc, has evolved mainly to enable predators to subdue prey larger or more mobile than themselves. Maybe this is seldom the case with mammalian predators. But one might expect e.g. weasels to find venom useful in subduing larger prey such as rabbits.

    An alternative possibility is that venom has evolved from salivary enzymes originally used to aid digestion. Mammalian predators tend to tear their prey into chunks, so maybe they did not need strong salivary enzymes.

    This is all just hasty speculation.

  10. #10 Mrs Tilton
    November 23, 2005

    Fascinating stuff. I’m glad to learn from Dave at No. 5 that the phylogeny’s unsettled, otherwise I suppose that, based on the tree up top, we’d really have to stop talking about ‘snakes’ altogether!

  11. #11 jorzo
    November 23, 2005

    The venomous species did not turn up on random branches of the tree. Instead, they all belong to the same branch, marked on the tree with the name “venom clade.” (Clade just means a group of species sharing a common ancestor.)

    Wouldn’t that be how the tree was designed? Oops, I don’t mean THAT designed. Isn’t it like saying “all the A’s weren’t in random sections of the alphabetical list.” Aren’t animals grouped by common characteristics and naturally would show up in the same branch of the tree? Just a nit pick because I’m a nit picker. :)

  12. #12 Thanh
    November 23, 2005

    Out of curiosity, does the platypus belong to the same “venom clade”?

  13. #13 Jon H
    November 26, 2005

    David B: “Are there any verified cases of mammals with venom?”

    The “slow loris” is a primate which produces a venomlike substance.

    “On the inside of their elbows, sebaceous tissue secretes a toxin (like sweat pores, which is rather fitting since the toxic mixture smells remarkably like sweaty socks). The lorises take it into their mouth and deliver it in the bite. “

  14. #14 Dr. Bryan Grieg Fry
    November 27, 2005

    >Are there any verified cases of mammals with venom? And if not, why not?

    Yes, shrews, platypus (the closely related echidna has secondarily lost the venom after it evolved defensive spines) and even a primate group (the lorises). Some of the extinct mammaliforma also had deeply grooved venom delivering teeth.

    >Out of curiosity, does the platypus belong to the same “venom clade”?

    No, the venom clade in this context is the clade within the squamate reptiles (lizards and snakes) that have venom. Incidently, this clade has now been given the name ‘Toxicofera’ in Vidal & Hedges’ (two of my co-authors on the Nature paper) followup paper to our Nature paper. The followup paper can be downloaded from http://evo.bio.psu.edu/hedgeslab/Publications/PDF-files/171.pdf The platypus is rather unrelated ;-) and represents an independent evolution of venom.

    >An alternative possibility is that venom has evolved from salivary enzymes originally used to aid digestion

    Proteins in venom are mutated forms of normal body proteins, not salivary proteins. Have a read of my Genome Research paper for more information:

    http://www.venomdoc.com/downloads/2005_BGF_Genome_2_Venome.pdf

    As for Dave’s comment, the historical problem has been that trees based upon morphology have relied upon characters that are ambiguous or even were convergently acquired. DNA is infinitely more reliable. DNA is everything, the rest is just details ;-D

    In regards to the comment “Fascinating stuff. I’m glad to learn from Dave at No. 5 that the phylogeny’s unsettled, otherwise I suppose that, based on the tree up top, we’d really have to stop talking about ‘snakes’ altogether!” this is not a problem unique to this situation, where historical arrangements have been shown to be artificial once more reliable trees are reconstructed. The Whippo clade is a good example (in case you aren’t familiar with this, whales have been shown to be most closely related to hippos). Similarly, reptiles are a concept aren’t monophyletic, the birds inconsiderately split the group. So, in reality we need to come up with new names. That is exactly what Vidal & Hedges have done in their excellent paper.

    Cheers
    Bryan

  15. #15 Jane Shevtsov
    December 1, 2005

    Henry was talking about the Komodo dragon, not goannas, having a bacterial prey-killing mechanism.

  16. #16 Dr. Bryan Fry
    December 1, 2005

    >Henry was talking about the Komodo dragon, not goannas, having a bacterial prey-killing mechanism

    Goannas and komodos are the same thing, all are monitor lizards. The point is that the bacterial prey-killing story has been a red-herring.

    Cheers
    Bryan

  17. #17 Zane JB
    May 17, 2007

    The one thing that has me stubbed is why hasnt there been further study into other reptiles? despite the fact a gecko may not have the venom power to kill a human couldnt there still be an individual evolutionary trait that a gecko may have inhertited to itself? also another point is why do all reptiles seem to have evolved a very small amount of venom that may not have been powerful enopugh to kill anything but why do Blind snakes not possess any venom at all? and is also like to know the potency of the venom from a komodo as due to popular beleif if its not the bacteria that kills you then doesnt this meen we could have an anti-venom made to save people from a serious bite? or will the “bacteria” still kill us later? i found this article interesting and the reply’s people have given in regard’s to it but it think there needs to me further studies done and more studiy done on all the reptile groups.

  18. #18 DDeden
    September 26, 2007
  19. #19 David Marjanovi?
    September 27, 2007

    The link to the Vidal & Hedges paper is wrong — it’s 170, not 171: http://evo.bio.psu.edu/hedgeslab/Publications/PDF-files/170.pdf.

    Fascinating stuff. I’m glad to learn from Dave at No. 5 that the phylogeny’s unsettled, otherwise I suppose that, based on the tree up top, we’d really have to stop talking about ‘snakes’ altogether!

    I don’t understand what you mean. The snakes are all in Serpentes. Not everything that lacks legs is a snake!

    The one thing that has me stubbed is why hasnt there been further study into other reptiles? despite the fact a gecko may not have the venom power to kill a human couldnt there still be an individual evolutionary trait that a gecko may have inhertited to itself? also another point is why do all reptiles seem to have evolved a very small amount of venom that may not have been powerful enopugh to kill anything but why do Blind snakes not possess any venom at all?

    1. The geckos are in the tree (Gekkonidae).
    2. It isn’t “all reptiles”, it’s just the iguanian + anguimorph + snake clade (called Toxicofera by Vidal & Hedges).
    3. The blind snakes must have lost it — or they need more research…

    the historical problem has been that trees based upon morphology have relied upon characters that are ambiguous or even were convergently acquired. DNA is infinitely more reliable. DNA is everything, the rest is just details ;-D

    DNA comes with its own problems — long-branch attraction is much more common in DNA data than in morphological data. However, good modern molecular studies are better than morphological studies done 20 years ago. For a good modern morphological study, you need plenty of taxa (including fossils!) and plenty of characters (at least 3 times as many as taxa). No such study exists for squamate phylogeny. For the whippo question, it has been done (2003 and 2005), and indeed, the whales come out as the sister-group of Anthracotheriidae, which in turn includes the hippos. Given the fact that the venoms and several other morphological characters fit the new molecular squamate tree much better than the traditional morphological tree, I’m optimistic about the former.