One of the more intriguing features of snakes is their lack of arms or legs to provide support and locomotion. Their sinusoidal, wave-like movements have long fascinated our kind, perhaps even impacting the evolution of primate visual systems (a discussion for another time, though see 1 on the intriguing concept of Snake Detection Theory). While it may seem simplistic, the body plan of snakes is actually rather utilitarian. They can ball up and minimize surface area exposure to conserve heat, or alternatively stretch out to increase their surface area and reach. Snakes are phenomenal climbers, swimmers, crawlers, burrowers, sprinters, even side-winders or aerial gliders, all with a rather similar design; long and legless.
It initially seems somewhat counter-intuitive to to suggest that snakes lost their limbs, so useful are our own arms and legs to our lives. Why would an animal with perfectly functional legs loose them? Is there some “anti-evolutionary” forces going on here? Surely loosing such a feature is a huge detriment and would be punished by natural selection, no?
‘No’ indeed! Evolution doesn’t necessarily proceed in a way that increases complexity. Natural selection has no obligation to punish the loss of limb function if there are gains and benefits in another area. For example, the energy and resources used in building legs might be better put to producing more young, perhaps increased sperm production or more ovarian follicles. Aside from this resource partitioning benefit, there may be many ecological scenarios (e.g. living in as dense, grassy habitats and requiring speed) in which arms or legs are little more than a hindrance. In fact there seems to be two main cases where limblessness is an advantage in reptiles; in the rather shorter, stumpier burrowers, digging their way through the substrate with external limbs just getting in the way and causing friction, and the more elongate terrestrial sliders/grass swimmers we mentioned above (2).
Putting aside the many limbless invertebrates, let us focus on the reptiles for now for they will prove instructive. It’s a common perception the reptiles can be divided into the following groups; there are the turtles and their like, crocodiles and their relatives such as alligators, the lizards, and finally, the legless snakes. This picture is not only incomplete, for example missing the birds which modern science now includes with the Archosauria (3,4), but assumes leglessness belongs only to the snakes. In reality, a elongated, limb-reduced form appears to have evolved independently some 25 times in reptiles, while true limblessness may have arisen in the Squamates, the snakes and lizards, a dozen times or so in seven different families (2, 5).
It seems that leglessness is a fairly common strategy among vertebrates and reptiles in particular. While initially implausible, the idea of doing away with cumbersome and obstructive limbs and hips seems to work perfectly well, with many lineages exploiting available niches in dense vegetation or underground, adapting to these lifestyles over time (5). Perhaps it’s the apparent major rearrangement that is necessary to rid one of limbs that makes such ideas so counter intuitive. One can certainly understand why we upright hominds, bipedal walkers and incessant tool users that we are, might have a negative responses to the concept of limb loss. However most legged reptiles drag their body somewhat anyway, spending plenty of time with their ventral surfaces in contact with the ground. Their inability to even use the simplest telephones further demonstrates the limited utility of reptile limbs for purposes we might ourselves consider important but have little impact on the lives of reptiles. What hubris!
Aside from considering our own teleological tendencies on the matter, remembering that what works for one lineage doesn’t necessarily work for all, we can approach the evolution of leglessness from a developmental angle. In considering the frequent progressive loss of limbs in reptiles, how easily can leglessness evolve? Through understanding how such a drastic change in structure came about biologically, we might better understand it’s repeated evolution, and also come to a greater appreciation of development, the building of a complex organism from a single celled zygote.
The seemingly intractable task of contemplating the growth and organization of some 37 trillion cells to make up your own meaty, mobile, thinking “gene-vehicle” called a body (6) is more manageable if we break things down into stages. First, fertilization by two haploid cells leads to the formation of a zygote. The zygote cell begins to divide in a process known as cleavage, until 16 or so cells form a tight ball. These cells now begin to differ as they divide, forming a mass of cells known as the blastocyst which enters the womb and implants itself in the uterine wall. We shall not dwell on the details of how the blastomeres, the blastocyst cells, differentiate into trophoblasts, embryoblasts, and so on, except to say it is fascinating and you can read more here (an excellent resource, 7). We will satisfy ourselves for the moment with the blastocyst implanting on the uterine wall around week five in humans, becoming an embryo, beginning as a two layered embryonic disc with an upper “epiblast” and the lower “hypoblast” layer, respectively formed of primitive ectoderm and endoderm cells.
Gastrulation, the folding of the epiblast inwards to form a cavity, the whole structure now with differentiated endoderm, ectoderm, and mesoderm layers, is followed by organogenesis where the organs of the body begin to develop (8). A simple example; among many other roles the mesoderm of vertebrates gives rise to the heart, blood cells, and blood vessels, as wells as performing myogenesis (formation of muscle fibres). As organogenesis continues, the human fetus develops arm and leg buds at around week 6, toes appearing around week 9 when all the embryo’s major organs have begun to grow in size. This all occurs in the first trimester of human pregnancy, the first 12 weeks after conception. The next two trimesters proceed as mother and child develop in step, further developing the now formed organ systems, totaling just over 280 days, or 40 weeks, until birth (9).
Let’s imagine, unethical as it may seem, that we wish to design a snake-like bodyplan from an extant tetrapod lizard, perhaps as a new kind of pet for unscrupulous billionaires. We have the power to insert mutations as we please into the germline, but with some cost to us (perhaps we lose a toe per each mutation) so a minimum number of mutations would be ideal. Where to begin in de-limbing our lizard linneage? For my money, limb budding seems to be a reasonable place to start, as we might switch off the development of limbs right at the start, or before they form at all. How then do these limb buds form? Something must initiate the formation of the limb buds at week six. Perhaps we can pinpoint this trigger and interfere with it in some manner. This initiation occurs through genes, or more specifically, regulation of gene expression. The chemical reactions of fertilization caused by a variety of proteins trigger other reactions, with more transcription factors expressed and binding to upstream promoters and so on, the beginnings of an ongoing cascade of chemical activity in the developing organism which will continue until it is born, grows, and one day expires. At each stage, different sets of genes are activated producing an array of interactions such as flow-on effects and feedback loops of gene expression to create specific proteins and cell structures, orchestrating the development of the zygote, blastula, gastrula, embryo, fetus, child, and onward.
Genes, the sequences of DNA in cell nuclei which the cell translates into proteins (putting aside introns/exons and splice variation for now), aren’t constantly active. Upstream of a gene’s “start” sequence one will find various promoter regions: genetic switches which, at least in us complicated eukaryotic organisms, may work alongside other elements (enhancers, silencers, insulators) to control a gene’s transcription rate. The presence or concentration of specific regulatory proteins inside cells direct transcription factors, proteins containing a DNA-binding domain, to bind to a certain gene’s upstream promoter and/or enhancer regions. Transcription factors can either promote (activators) or inhibit (repressors) the recruitment of local RNA polymerase enzymes to transcribe of the gene to mRNA, which is itself translated to a protein following transport to one the cell’s ribosmes. These proteins might be structural, such as the common mammalian collagen, or enzymatic, catalyzing some chemical reaction, perhaps kinases or phosphotases in signal transduction for cellular communication, Pepsin or Amylase for digesting food, or various regulatory proteins including transcription factors themselves. With all this in mind, the development of limb buds in the early embryo, the control of gene expression by transcription factors and upstream regulatory sites, and the cascade of interacting functional or regulatory proteins that is biology and cellular development, let’s examine how these steps really create a four limbed embryo.
Cranking the human developmental clock up to week five, we have our wonderful gastrula, recently invaginated (the process of inward folding of the blastocyst) with differentiated layers of meso-, ecto-, and endoderm cells ready to begin organogenesis (8). Something causes the mesoderm cells to grow outward into a paddle like protrusion covered by the ectoderm, the limb bud of the embryo. We can rather safely assume that signaling molecules within the cells activate transcription factors to bind to promoters/enhancers of certain genes to control this process. If we knew these genes, or their transcription factors and regulatory mechanisms, our legless creature may be available at the local freaky-science-petstore sooner than expected!
What we are interested in is a specific set of regulatory genes important for development, known as HOX genes. These genes are common throughout many complex organisms, essential for directing the formation of the bodyplan (10, 11). HOX genes control the development of an embryo along a head to tail axis, with genes ranged along the chromosome in order of bodyplan. Quite literally, HOX genes for head development are at one end, genes for tail end on the other, with various body segments in the middle (a phenomenon known as colinearity).
These important regulatory genes are part of a related family of genes, known as the homeotic genes (10,11,12). A variety of homeotic genes exist, including the ParaHox & Mads-BOX genes, all which regulate the development of anatomical structures. The protein products of homeotic genes are transcription factors working upstream to activate a cascade for genes (including other TFs) producing certain organs or body parts. Homeotic genes do not produce the overall body segments themselves. Rather, once the segments are formed, homeotic genes (particularly the closely related HOX family) determine the identity of the segments (head vs. abdomen etc.) and which structures will form where.
How is this achieved? Thanks to DNA sequencing efforts, we know that these HOX genes all contain a similar sequence known as the “homeobox” (10,11,12). This 180 base-pair long sequence encodes a 60 amino acid “homeodomain”, a helix-turn-helix folded protein of three alpha-helices joined by short amino acid loops, forming a structure capable of binding to a strand of DNA given certain complementary sequences are present. These complementary sequences are found in the upstream promoter regions of certain genes involved in the development the appropriate characteristics for each body segment.
Let’s consider the HOX genes of the fruit fly, Drosophila melanogaster, for a moment. This tiny organism has been extensively used in biological studies, thus we know practically all we can about it’s genome, development, physiology, behaviour and ecology (12). Like other insects it possesses 8 different but closely related HOX genes on it’s third chromosome, clustered into two groups. The Antennapedia complex contains five genes: labial (lab), proboscipedia (pb), deformed (Dfd), sex combs reduced (Scr), and Antennapedia (Antp). The remaining three form the Bithorax complex, which includes Ultrabithorax (Ubx), abdominal-A (abd-A) and abdominal-B (abd-B) genes. The order in which these genes are presented above is also the order of their arrangement along chromosome 3 in insects, as well as the the order they’re expressed in along the bodyplan, with lab expression producing the most anterior head parts while abd-B expression produces the tail segment. The action and colinearity of HOX genes in bodyplans can be demonstrated by some rather gruesome mutations in D. melanogaster. During development, various mouth and head structures initially develop outside the head, then fold inwards in a process called involution (10). Loss of function in lab means head structures including the salivary glands, pharynx, and the labial appendage, fail to form. A chromosomal inversion in the Antp gene forward has the unfortunate consequence being that a leg, rather than an antenna, grows out of the fly’s heads. It seems Antp is an important upstream regulator of, among other things, the development of legs in fruit flies and many other insects.
What about other groups of organisms? Do all legged taxa share similar pathways for leg development? Unfortunately it’s not that simple. The Arachnid Antp gene actually represses leg development in the first abdominal segment. Loss of function or down regulation of Antp in spider embryos in fact leads to a ten-legged phenotype, rather than no legs as one might suspect (13). HOX genes are still important in leg development just, just not the ones we might suspect if we look at insects alone. It seems that the ancestors of spiders and scorpions stumbled upon an alternative way to use their HOX genes to develop limbs.
As we have mentioned, insects have eight HOX genes in two clusters, however mice, humans, and many other vertebrates have 39 closely related HOX genes in four clusters (A to D), likely the result of multiple gene duplication events (4,10,14). Traditional theories on the subject suggest that the snake bodyplan is “deregionalised”, with HOX genes for thoracic and cervical being expressed concurrently rather than separately in clearly defined neck, torso, and tail regions etc. Snake vertebrae and ribs display little obvious differentiation from tail to neck, and it was presumed that HoxC6 and HoxC8 expression was restricted to the first cervical vertebrae, eliminating both the torso and leg development (15). However recent research shows that despite the superficial similarity, snake vertebra are actually differentiated by region. By statistically examining the minor variations in vertebra and HOX expression across a wide range of limbed, limb-reduced, and limbless reptiles, Head & Polly (16) found that cervical, thoracic, and caudal regions do in fact exists in snakes, the boundaries of which are defined by HOX expression. Why then are there no limbs to go along with these regions?
Another fruitful avenue of research has been to look at the gene expression in more ancient snake lineages (5,10,15). Modern snakes are entirely limbless, however they all evolved from a legged ancestor, with various intermediate fossils showing gradual loss of the front legs followed by the back legs. In fact, modern pythons and boas still retain a hip girdle and even elements of the femur. Scientist have long been interested in these species for their value as comparative subjects for studying leg development. While the details of cellular communication, division, differentiation, function, and controlled death known as apoptosis are all intricately controlled and orchestrated by the expression of different genes and enzymes, and far too complex to fully cover in a simple essay, let’s quickly cover the basics of what we know from python legs.
During development, python embryos initially develop the typical vertebrate hindlimb bud. Unlike human limb buds which continue to develop from week 6 onwards, python buds bloom into little more than tiny vestigial limbs with a basic, incomplete femur and a single claw (17,18). Developmental studies have shown that this is due to two regions of the limb bud which fail to form properly; at the posterior base of the bud the zone of polarizing activity (ZPA) controls outgrowth and polar patterning of tetrapod limbs, while at the tip of the bud the apical ectodermal ridge (AER) further communicates with the mesenchyme (formed from the mesoderm) and ZPA to determine axial patterning of the limb cells and the distal outgrowths (fingers/toes and the rest of the supporting hand/foot bones).
The ZPA functions by secreting Sonic Hedgehog (SHH), an important signaling protein in the Hedgehog signaling pathway (17,18). Produced by activation of the Sonic Hedgehog (shh) gene, SHH helps form the identity of the parts of the limb (femur, tibia/fibia, and feet) depending on both it’s concentration, which decreases with distance from the ZPA, and by suppression by other enzymes (such as GLI3). SHH also induces a gene called Gremlin1 (GREM1) which maintains the activity of the AER. Additionally, the AER stimulates SHH expression in the ZPA through production of fibroblast growth factors (FGFs), creating a feedback loop that causes an increasing expression of these genes and their products, controlling segment identity and cell differentiation, thus leg formation. Interestingly, python leg buds at 1-2 days old showed no evidence of SHH protein production at the ZPA, yet python ZPA cells placed underneath a chicken embryo’s AER will produce SHH. Apparently the SHH production mechanism, the shh gene, is intact in pythons, just not being activated during development.
Like many genes shh expression is controlled by an upstream promoter. The limb-specific enhancer for shh is known as the ZPA Regulatory Sequence (ZRS). Various genes such as HOXD13 and HOXA13 have been shown to bind to the ZRS in mice, activating the shh gene. These ZRS binding genes are also important in the distal patterning of the limb, in finger formation and so on, and show similar expression in the leg buds of both lizards and pythons. If the shh gene is functional, as are the HOX and other genes which regulate it, that leaves only the ZRS itself to blame. In 2016, Leal & Cohn (17) found that pythons have three mutations in their ZRS each which reduce the binding activity of this important enhancer. Together, all 3 mutations reduce ZRS activity to 12.8% of the controls without mutations. More specifically, these mutations occur at the upstream (denoted 5′, the five-prime-end, as opposed to the downstream three-prime-end of the DNA strand, 3′) end of the ZRS, a site for HOXD binding. When comparing the activity of HOXD gene products binding to the enhancer, the activity of the mutant ZRS in pythons is only 2% of that in mice. HOXD enhancers, however, remain intact and in-play.
Considering all this, a picture of leglessness begins to emerge involving progressive loss of shh function while HOX genes remain very much in play. Using molecular clocks (a calibrated mutation rate and observed sequence variation), it appears that one of these mutations arose during the late Upper Cretaceous, initially just interfering with it’s activity as an shh enhancer (17). This “hypofunctionality” led to further mutation over time, eventuating in the loss of hindlimb structures. While this might seem a large jump in morphology, something we generally deride in evolutionary theory, the loss likely occurred through slow, progressive divergence. Not only do the mutations in the ZRS have a cumulative impact, likely evolving one after another and acting in concert, but other species show evidence of a similar pattern. Lizards, pythons & boas, pit vipers, garter snakes, and king cobras (Ophiophagus hannah, an elapid and presumably the most recent lineage of the included taxa) show a degenerative pattern, with the most modern possessing the most mutated, non-functional ZRS, and vice versa.
So, while HOX genes play an important role in leg development and many other roles, they are not, as we suspected at the outset, the key to leglessness in snakes. However, by working from the basics, that is, the body plan and the genes which pattern it, we’ve perhaps come to a better understanding of signaling molecules, transcription factors, and development, which helps us see how HOX gene products bind to the ZRS to increase the transcription rate at shh, promoting SHH protein expression which is essential for leg segment patterning and formation. A mere three mutations inactivate the ZRS promoter, which the authors of the study named ΔA, ΔB, & ΔC, so we need only lose three toes to achieve our de-limbed lizard. Fantastic! It is also of great credit that M.J. Cohn, one of the authors of the 1999 paper suggesting snake skeletons are de-regionalized (15), is the same M.J. Cohn from the 2016 paper identifying these mutations in the ZPR (17), just one fine example of many a scientist who, upon seeing compelling data which might contradict their personal ideas, reviewed and accepted the weight of the evidence and changed their view accordingly, even going on to further the progress on these new theories. Such is the dedication to truth we should all try to achieve, that we may overcome our preconceptions when faced with new, compelling evidence.
Why not just mutate the activating HOX genes themselves, HOXD13 for example? Evolution has no such hindsight, and furthermore, such a drastic changes are more likely to cause problems due to the complexity of living systems. By recruiting promoters into the mix we now have the option to loose function without loosing or even modifying the underlying sequence. Gene expression can be altered without loss of a gene itself, which might be essential elsewhere in development, as one might suggest of such important body patterning genes like the the HOX family. Mutating HOXD genes would do more than just remove legs, for instance it was recently discovered that HOXD genes, but not shh, are important in the development of snake genetalia, and that their expression in both genetalia and digits are under shared genetic regulation. SHH also continues to be important outside of the ZPR, helping pattern brain development and regulate cell division in adults, it’s misfuction potentially having a role in certain types of cancer development (19). Who knows the consequences to snake hemipenes and brains if the entire HOXD cluster was disrupted by mutation. When they fall in specific regulatory elements like the ZRS instead, mutations may accidentally hit upon more targeted genetic switches, the consequences of which, apparently, can define the form and function of a whole lineage of life on earth for many millions of years.
References
- Van Le, Q., Isbell, L.A., Matsumoto, J., Nguyen, M., Hori, E., Maior, R.S., Tomaz, C., Tran, A.H., Ono, T., Nishijo, H. (2013) Pulvinar neurons reveal neurobiological evidence of past selection for rapid detection of snakes. Proc Natl Acad Sci USA. 110(47):19000-5.
- Wiens, JJ., Brandley, M.C., Reeder., T.W. (2006) Why Does a Trait Evolve Multiple Times within a Clade? Repeated Evolution of Snakelike Body Form in Squamate Reptiles. Evolution 60(1): 123–141
- Nesbitt, S.J. (2011). The early evolution of archosaurs: relationships and the origin of major clades. Bulletin of the American Museum of Natural History. 352: 1–292.
- Böhmer, C., Rauhutm, O.W., Wörheide, G. (2015) New insights into the vertebral Hox code of archosaurs. Evol Dev. 17(5):258-69. doi: 10.1111/ede.12136.
- Brandley, M.C., Huelsenbeck, J.P., Wiens, J.J. (2008) Rates and patterns in the evolution of snake-like body form in squamate reptiles: evidence for repeated re-evolution of lost digits and long-term persistence of intermediate body forms. Evolution. 62(8):2042-64.
- http://www.smithsonianmag.com/smart-news/there-are-372-trillion-cells-in-your-body-4941473/
- https://embryology.med.unsw.edu.au/embryology/index.php/Blastocyst_Development
- https://embryology.med.unsw.edu.au/embryology/index.php/Gastrulation
- https://embryology.med.unsw.edu.au/embryology/index.php/Timeline_human_development
- Gilbert, S.F. (2000) Hox Genes: Descent with Modification. Sunderland (MA): Sinauer Associates.
- McGinnis, W., Garber, R.L., Wirz, J., Kuroiwa, A., Gehring, W.J. (1984) A homologous protein-coding sequence in Drosophila homeotic genes and its conservation in other metazoans. Cell. 37(2):403-8.
- Lemons, D., McGinnis W. (2006) Genomic evolution of Hox gene clusters. Science. 313(5795):1918-22.
- Khadjeh, S., Turetzek, N., Pechmann, M., Schwager, E.E., Wimmer, E.A., Damen, W.G., Prpic, N.M. (2012) Divergent role of the Hox gene Antennapedia in spiders is responsible for the convergent evolution of abdominal limb repression. Proc Natl Acad Sci USA. 109(13):4921-6.
- Pearson, J.C., Lemons, D., McGinnis, W. (2005) Modulating Hox gene functions during animal body patterning. Nat Rev Genet. 6(12):893-904.
- Cohn, M. J., Tickle, C. (1999) Developmental basis of limblessness and axial patterning in snakes. Nature. 399, 474–479.
- Head, J.J., Polly, P.D. (2015) Evolution of the snake body form reveals homoplasy in amniote Hox gene function. Nature. 520 (7545):86-9.
- Leal, F., Cohn, M.J. (2016) Loss and Re-emergence of Legs in Snakes by Modular Evolution of Sonic hedgehog and HOXD Enhancers. Current Biology. 26(21):2966-2973.
- Kvon, E.Z., Kamneva, O.K., Melo, U.S., Barozzi, I., Osterwalder, M., Mannion, B.J., Tissières, V., Pickle, C.S., Plajzer-Frick, I., Lee, E.A., Kato, M., Garvin, T.H., Akiyama, J.A., Afzal, V., Lopez-Rios, J., Rubin, E.M., Dickel, D.E., Pennacchio, L.A., Visel, A. (2016) Progressive Loss of Function in a Limb Enhancer during Snake Evolution. Cell, 167 (3): 633
- Hatton, B.A., Knoepfler, P.S., Kenney, A.M., Rowitch, D.H., de Alborán, I.M., Olson, J.M., Eisenman, R.N. (2006) N-myc is an essential downstream effector of Shh signaling during both normal and neoplastic cerebellar growth. Cancer Research. 66 (17):8655-61.
Snakes On The Brain 