Skip to main content
NCBI home page
Journal of Anatomy logoLink to Journal of Anatomy
. 2021 Aug 3;240(1):1–10. doi: 10.1111/joa.13531

Sharp and fully loaded: 3D tissue reconstruction reveals how snake fangs stay deadly during fang replacement

Silke G C Cleuren 1,✉, William M G Parker 1, Hazel L Richards 1,2, David P Hocking 1,3, Alistair R Evans 1,2
PMCID: PMC8655161  PMID: 34346066

Abstract

Snake venom is produced, transported and delivered by the sophisticated venom delivery system (VDS). When snakes bite, the venom travels from the venom gland through the venom duct into needle‐like fangs that inject it into their prey. To counteract breakages, fangs are continuously replaced throughout life. Currently, the anatomy of the connection between the duct and the fang has not been described, and the mechanism by which the duct is reconnected to the replacement fang has not been identified. We examined the VDS in 3D in representative species from two families and one subfamily (Elapidae, Viperidae, Atractaspidinae) using contrast‐enhanced microCT (diceCT), followed by dissection and histology. We observed that the venom duct bifurcates immediately anterior to the fangs so that both the original and replacement fangs are separately connected and functional in delivering venom. When a fang is absent, the canal leading to the empty position is temporarily closed. We found that elapid snakes have a crescent‐shaped venom reservoir where venom likely pools before it enters the fang. These findings form the final piece of the puzzle of VDS anatomy in front‐fanged venomous snakes. Additionally, they provide further evidence for independent evolution of the VDS in these three snake taxa.

Keywords: DiceCT, fangs, tooth replacement, venom delivery system, venom duct, venomous snakes

The connection between snake fang and venom gland is presented alongside the maintenance of these connections during the process of fang replacement, using histological imaging and diceCT. A bifurcating connection with both the original and replacement fang was observed, allowing both fangs to be functional at delivering venom. Differences across front‐fanged families was observed, where a venom reservoir anterior to the fang was only observed in Elapidae.

graphic file with name JOA-240-1-g005.jpg

1. INTRODUCTION

Many reptiles replace their dentition throughout or during a long portion of their lives (Edmund, 1969; Osborn, 1974; Richman & Handrigan, 2011). Unlike most juvenile mammals, which replace their deciduous teeth as a set, reptiles continuously replace teeth along their toothrows, with teeth at different stages of replacement occurring along the row at any given time (Edmund, 1969). This pattern is called polyphyodonty and is broadly present in reptiles because their bodies and heads continue to increase in size during life and thus experience indeterminate growth (Edmund, 1969). Additionally, it ensures that teeth remain in good condition, preserving tooth shape and sharpness. In acrodont lizards, the mature size is reached, and tooth replacement ceases at a certain age (oligophyodonty), whereas in crocodilians, replacement eventually slows down or becomes more and more irregular over time. In most snakes and lizards, however, the tooth replacement persists and does not slow down throughout life (Edmund, 1969).

Fangs of venomous snakes are also continuously replaced throughout their lives. Compared with the other teeth in snake jaws, these fangs take longer to develop as a result of their larger size and more complex structure, especially the tubular design seen in the front‐fanged taxa Viperidae, Elapidae and Atractaspidinae. The solution for ensuring a quick turn‐over of fangs is the production of up to eight replacement fangs at the same time, all in varying stages of development (Jackson, 2007; Zahradnicek, Horacek & Tucker, 2008). In the maxilla there are two fang sockets, one for the functional fang and one for the replacement fang (Nagy et al., 2013) (Figure S1). One of the sockets is thought to be empty most of the time and will only be occupied by a replacement fang right before the functional fang is shed (Nagy et al., 2013; Schaefer, 1975), making the socket of the previously functional fang empty for the next replacement round. This means that the functional fang will alternate in position after each replacement. Venomous snakes, thus, avoid having a period where there is no fang, by ankylosing the replacement fang in the second socket before the previous one is lost (Nagy et al., 2013).

Venomous snakes are unusual among vertebrates in that their teeth function not just in penetrating food but also in delivering venom into their prey. This adds an additional complication to the replacement of these teeth, for as the teeth are replaced they must also maintain their connection to the venom gland. This requires a mechanism by which the venom duct is reconnected as fangs are lost and replaced. In the past, it has been difficult to determine exactly how this mechanism works due to the small size and complex arrangement of these structures. This has led to three plausible hypotheses for how these teeth stay connected to a duct as they are replaced (Figure 1):

  1. One opening that terminates just before reaching the fangs (Schaefer, 1976): When two fangs are present, this one opening spans across both sockets. For this scenario to work, Schaefer (1976) explained that the sheath surrounding the fang together with the central partition between the fangs would help locate the duct opening above the ankylosed fang(s). When one fang is present the central partition would lay against the fang sheath of the empty position, blocking the venom so that it does not flow into the mouth cavity. When two fangs are present, the sheath, as well as the duct opening, would be stretched so that it spans over both fang orifices, while the partition would be in the middle allowing the venom to go both ways. The anatomy proposed by Schaefer allows both fangs to be functional in delivering venom.

  2. Alternating duct that changes its connection from one fang to the other (Burnett, 1854): When one fang is present, the duct is directed to the entrance orifice of the fang. When a replacement fang ankyloses, the duct disconnects from the original fang and connects to the replacement. Changing the tissue connection likely takes some time, resulting in a short period where none of the fangs are capable of delivering venom. This ‘disconnected’ period makes this hypothesis less likely given the presence of two sockets that can allow venomous snakes to always have a fang in position.

  3. Bifurcating duct that connects to both fangs: The duct bifurcates, allowing each branch to attach to the entrance orifice of a fang. This would allow both fangs to be functional at delivering venom at the same time, ensuring a good physical duct connection. However, between replacement events only one fang is present, so the duct towards the empty socket will need a mechanism to temporarily close off from the mouth cavity, so venom does not flow into it.

FIGURE 1.

FIGURE 1

Open in a new tab

Hypotheses proposed for the venom duct connection during fang replacement

Previous studies have hypothesised that both fangs are capable of administering venom, making scenarios A and C most likely; however, the data on this are not conclusive (Edmund, 1969; Nagy et al., 2013; Schaefer, 1975).

Additionally, it is possible that the connection between the duct and the tubular fang is not the same across the three front‐fanged venomous snake taxa. Physiological variation in glands (Fry et al., 2003, 2008; Gans, 1978; Kerkkamp, Casewell & Vonk, 2017; Kılıç, Akat & Arıkan, 2016; Kochva, 1978; Weinstein, Smith & Kardong, 2009), compressor muscles originating from a different muscle in all taxa (Kerkkamp, Casewell & Vonk, 2017), and different fang designs (protero‐ or solenoglyphous) as well as the presence of a connection furrow or a completely smooth anterior fang surface (Berkovitz & Shellis, 2016; Chipman, 2009; Jackson, 2007; Jackson & Fritts, 1995; Kardong, 1979; Pyron et al., 2014; Vonk et al., 2008; Zahradnicek, Horacek & Tucker, 2008) indicate that the front‐fang venom delivery system (VDS) evolved independently in the two families and subfamily within the Colubroidea (Fry et al., 2006; Iwanaga & Suzuki, 1979; Vidal, 2002; Zeller, 1948).

To examine the connectivity of the duct to the fangs and determine whether this system is consistent among the families, we used diffusible iodine‐based contrast‐enhanced Computed Tomography (diceCT). This method enables us to visualise and differentiate soft tissues in CT scans by increasing the radiodensity of the muscles, epithelia and connective tissues, while keeping the head intact. Because of challenges of penetrating large specimens, this approach has largely been restricted to use on smaller or partly dissected specimens (Gignac & Kley, 2014; Jeffery et al., 2011; Klinkhamer et al., 2017). We will, therefore, provide methodology for staining snake heads keeping in mind full penetration of soft tissues and minimising shrinkage. We further verified our diceCT findings in one specimen via gross dissection and histological imaging after CT scanning.

This study aims to provide a detailed description of the fang replacement mechanism in the three taxa of front‐fanged venomous snakes (Elapidae, Viperidae and Atractaspidinae) and to determine how replacement fangs achieve their connection to the venom gland via the venom duct. By identifying features that vary among these families, we add further evidence for the independent evolution of the VDS within venomous snakes. Understanding this mechanism will help us better grasp how these animals are able to maintain their unique method of feeding, with seemingly delicate fangs staying both sharp and loaded throughout a snake's life.

2. METHODS

2.1. Specimens

Eight frozen snake specimens were loaned from the collections of Museums Victoria (NMV) in Melbourne (Australia) and Steinhardt Museum of Natural History (SMNH) in Tel Aviv (Israel). Three of the Tel Aviv specimens were donated to the Evans EvoMorph Lab and have been accessioned to the lab collection (EEM). We examined five elapid species: Notechis scutatus (NMV D77186), Pseudonaja textilis (NMV D77144), Hydrophis platurus (NMV D75974), Austrelaps superbus (NMV D77128) and Walterinnesia aegyptia (SMNH R14060); one atractaspidin species: Atractaspis engaddensis (EEM‐R1); and two viperid species: Echis coloratus (EEM‐R2) and Daboia palaestinae (EEM‐R3).

2.2. DiceCT and scanning protocol

Museums Victoria specimens were thawed, decapitated one head length behind their head and fixed in 10% neutral buffered formalin for 3 days. Specimens were stained for 27 days in 1.5% w/v Lugol's solution (diluted from 15% stock solution (10 g KI, 5 g I2 in 100 mL H2O)) in a foil‐covered jar on a gentle 2D rocking platform at room temperature (Metscher, 2009a, 2009b; Nasrullah, Renfree & Evans, 2018) (Figure 2a). This is an approximately isotonic concentration of Lugol's solution to minimise shrinkage of the specimen. The staining duration was confirmed first on Notechis scutatus (NMV D77186), where a progress scan taken at 10 days of staining showed incomplete penetration of Lugol's iodine, and staining was continued to 27 days.

FIGURE 2.

FIGURE 2

Open in a new tab

Methodology for diceCT staining and scanning. (a) Snake heads were stained in 1.5% w/v Lugol's solution and (b) microCT scanned. (c) CT scans were segmented in Avizo software. (d) Segmented regions were 3D reconstructed and exported. (e) The anatomy determined from the segmentation and reconstruction was illustrated to portray the anatomy observed in each of the front‐fanged venomous snake taxa

After staining, the specimens were rinsed with water to remove any Lugol's solution in the mouth cavity, patted dry with paper towel and placed in a zip‐lock bag to prevent drying out and refrigerated until scanning. Each of the eight specimens was placed in a Falcon tube, then mounted onto the scanning platform and scanned in a Zeiss Xradia 520 Versa microCT at the Monash University X‐ray Microscope Facility for Imaging Geomaterials (XMFIG) (Figure 2b). For each specimen two scans were taken, one of the entire head to include the complete extent of the venom gland and one zoomed in on the snout region to achieve a higher resolution to visualise the venom duct connection to the fangs. All zoomed scans were performed using source settings of 80 kV and 87–88 μA with 3–5 s of exposure. The cubic voxel size varied based on the size of the snake head, ranging from 12.49 to 21.07 μm (Table S1).

2.3. 3D reconstruction

MicroCT scans were reconstructed using Automatic Reconstructor (Zeiss). Interpretation of the anatomy and creation of high‐resolution, three‐dimensional surface models of the fangs, venom ducts and glands were achieved using Avizo Lite 9.0.7 (FEI) (Figure 2). Every structure of the VDS was segmented in Avizo (Figure 2c). Fang models were made using a combination of magic wand with thresholding and brush strokes with limited range. Where possible, soft tissues were partly selected using the magic wand with thresholding and manual painting using the brush. When large portions needed to be selected using the brush, interpolation was used. For each specimen, the four resulting models (fangs, duct, accessory and venom glands) were exported together as one model and separately as.ply files (Figure 2d). Lateral views of the 3D models of the VDS of each individual were captured (Figure S2). These were used to make simplified illustrations of the anatomy observed in the three front‐fanged venomous snake taxa (Figure 2e).

2.4. Dissection

After diceCT scanning, P. textilis was placed in 70% ethanol for 11 days to remove the colouration of the iodine staining, refreshing the solution after two and four days of destaining. Using standard tools under a dissection microscope, the skin layer on the right side of the snake's head was removed starting posterior to the venom gland and gradually moving anterior to the location of the fang. This revealed the complete VDS, which was captured using a Zeiss SteREO Discovery V20 microscope equipped with an Axiocam 506 colour camera. The workflow of the dissection is visualised in the SI along with a close‐up of the removed VDS (Figure S3 and S4, respectively).

2.5. Histological imaging

To confirm the presence and identity of specific tissue and cell types that were identified in the diceCT scan, the extracted VDS of P. textilis was histologically sectioned. It was embedded in paraffin and sectioned in the coronal plane in 4 μm slices using a Leica Histocore Biocut microtome (2019) at the Monash Anatomy and Developmental Biology Histology Platform. In total, eight levels consisting of five slides with between 4 and 6 slices were made. On the first two levels, only a few slices were stained with haematoxylin and eosin (H&E), from level 3 onwards multiple slices were H&E stained to pinpoint to location of the duct connection to the fangs. Two slices on level 4 slide 2 were stained used Masson's trichrome to identify the nature of the tissue (connective or muscular). Resulting stained sections were photographed using a BX51 Olympus microscope with a Leica MC 190 HD microscope camera at Monash University School of Earth Atmosphere and Environment (EAE) microscope facility. Scale bars were calculated using Leica LAS EZ imaging software.

3. RESULTS

3.1. DiceCT

All scans resulted in clear differentiation of the soft tissues while avoiding extensive shrinkage of the specimens. Only the zoomed scans were of a high enough resolution to visualise the venom duct from gland to fang. Connection between the fang (either one or two fangs) and the venom duct was observed for all eight snake species, where the duct is found to bifurcate before reaching the ankylosed fang(s) (Figure 3). All three families, thus, possess a bifurcating venom duct to enable functionality of the replacement fang as soon as it is ankylosed. When only one fang was present, remnants of a canal leading to the empty pocket were visible (Figure S5a,c,e). This indicates the presence of a mechanism to temporarily close the venom canal that leads towards an empty socket in between replacement events, to prevent venom from flowing into the mouth cavity. However, some differences in the VDS across the three front‐fanged taxa were observed.

FIGURE 3.

FIGURE 3

Open in a new tab

Venom duct connection to the fangs of front‐fanged venomous snakes. MicroCT slices showing the venom duct connection to the fang's venom entrance orifice (a, e, i) and schematic representation of the anatomy of the entire venom delivery system (VDS) for each of the venomous snake families in anterior (b, f, j) and lateral views (c, g, k). Insets show lateral views of the 3D models generated from the diceCT scans (d, h, l). Interactive 3D pdf files of (d, h, l) showing the VDS and the skull can be found in the supplementary files, as well as Figure S3 showing the lateral view of all eight species. VDS comprises all or some of the following components: ankylosed fang (light grey); replacement fang (dark grey); separate ducts connecting to the entrance orifice (purple); venom chamber (light purple); venom duct (pink); accessory gland (light blue) and venom gland (turquoise)

In Elapidae, the venom duct originates from the venom gland and is surrounded by the accessory gland most of the way to the fangs (Figure 3c). The venom duct and accessory gland are positioned lateral to the ectopterygoid and maxillary bones. Before reaching the entrance orifices of the ankylosed fang(s), the duct expands and forms a venom reservoir where the venom likely pools before being pushed into the fang's venom tube. From the centre of this venom reservoir, a canal originates that splits into two canals that connect to the entrance orifice of each ankylosed fang (Figure 3a–d). The presence of this crescent‐shaped reservoir is only observed in the Elapidae specimens included in this study.

In both Atractaspidinae and Viperidae, the duct connects directly to both entrance orifices, when present, by splitting the duct into two (Figure 3e–l). These two families do, however, differ in the arrangement of their venom duct and glands, which corresponds to earlier descriptions of their anatomy.

The location of the duct also seems to differ slightly; in atractaspidins and viperids the fangs are approached by the duct more dorsally, whereas a more lateral approach is seen in elapids (Figure 3). In viperids, the dorsally located duct immediately bends ventrally when it reaches the fang, whereas in atractaspidins, the duct first continues more anteriorly and then takes a U‐turn towards the entrance orifice. In addition, we note that the fangs of Viperidae and Atractaspidinae are still positioned parallel to the upper jaws and venom duct, but when they are used they would rotate ventrally.

3.2. Histological analysis

The comparison of the diceCT scan slices of P. textilis shows extremely close alignment with the histological sections (Figure 4 and Figure S6). The crescent‐shaped venom reservoir is clearly visible with the budding off of the venom duct at the lateral end. The canal that points posteriorly and bifurcates was identified using the corresponding diceCT slice (Figure 4b). The complete bifurcation is, however, not visible in the histological sections: It appears that the tissue portion with the majority of the tissue where the bifurcation would be located was not excised from the specimen during dissection (Figure S7).

FIGURE 4.

FIGURE 4

Open in a new tab

H&E‐stained histological slice through the venom reservoir, duct and part of the canal that connects to the fang orifice of Pseudonaja textilis (NMV D77144) (a). The corresponding CT slice of that region shows identical anatomy (b). Masson's trichrome stained slice (c). Structure labels are added in each panel (CtO, canal to orifice; VD, venom duct; VR, venom reservoir)

From the Masson's trichrome staining no smooth muscle is found that could influence the duct or reservoir.

4. DISCUSSION

By visualising the VDS in front‐fanged venomous snakes, we identified that the venom duct bifurcates when two fangs are present so that both are functional at delivering venom. We also described key differences of anatomy between the taxa, with elapids possessing a venom reservoir anterior to the duct bifurcation. The description of the connection between duct and fang completes the puzzle of the anatomy of the VDS. With this study, we include the revised and complete VDS anatomy of front‐fanged venomous snakes (Figure 5).

FIGURE 5.

FIGURE 5

Open in a new tab

Summary of the current knowledge on the venom delivery system (VDS) of front‐fanged venomous snake taxa (Elapidae, Atractaspidinae and Viperidae). Overview of the complete VDS, type and anatomy of the venom gland (turquoise) and accessory gland (light blue), origin of the compressor muscle, our new findings on the anatomy of the venom duct connection to the fang (pink and purple), morphology of the maxillary bone (orange) and number of teeth present, and the fang design (grey). Information on gland, muscle and fang anatomy is gathered from (Berkovitz & Shellis, 2016; Kerkkamp, Casewell & Vonk, 2017; Savitzky, 1980; Weinstein, Smith & Kardong, 2009)

4.1. Bifurcating duct

In all three taxa, both the original and the new replacement fang, when ankylosed, are separately connected to the venom duct via two smaller venom canals. These canals, however, do not seem to be permanently connected to the tissue pocket reserved for the new fang. Instead, when the original fang is shed, the canal seems to have a dead end and does not run all the way to the empty fang pocket (Figure 3a, c, e). When a replacement fang ankyloses in that empty position, the canal reconnects to the entrance orifice of the replacement fang. Because of the ability to temporarily close off the canal leading towards the empty fang position, the venom is only directed into the venom tube of the present fang and is not injected into the mouth cavity.

Given these findings, we reject hypotheses A and B (Figure 1). The existence of two separate canals in all of the investigated species supports hypothesis C. Hypothesis A, put forward by Schaefer (Schaefer, 1976), also allowed the functionality of both the existing and new fang in delivering venom. Schaefer's proposed mechanism is anatomically simpler because making a new connection every time the fang is replaced is more complex and likely more energetically costly. Our results, however, indicate that all three taxa do create a new connection to each replacement fang, thereby directly connecting the venom duct to the entrance orifice of the fang, ensuring perfect transition of the venom from the duct to the fang's venom tube.

Slivers of a closed canal can be seen in the diceCT slices leading towards the empty fang pocket. Only the radiopaque epithelial lining of the canal is present, whereas there is no evidence of a lumen, which appears darker in the active, patent duct. This might indicate that the canal feeding into an empty socket temporarily collapses until a new fang is fully developed and ankyloses in that position. The mechanism by which this canal transitions between closure and functional patency is unknown. It could be a passive process related to shifting pressure gradients, analogous to the developmental changes in blood flow via shunts around the heart, lungs and GI tract in tetrapod foetuses (Dzialowski, 2018). In these snakes, the collapse of the venom duct might occur due to pressure loss, and then as a new fang matures the canal reconnects and pressure builds, resulting in the canal regaining patency as it receives venom through its lumen. Alternatively, the canal closure could be an active process whereby a myoepithelium lining contracts, constricting the canal in a sphincter‐like manner until the new fang is fully ankylosed (although myoepithelial cells are normally only present in the walls of ducts that lie within glandular tissue, such as in exocrine structures like salivary glands (McMillan & Harris, 2018)). Looking at the Masson trichrome sections, we can rule out closure by smooth muscle activity. Lastly, ankylosis of a fang to the maxilla might trigger a signal for the regeneration of the canal. Identifying these tissues and understanding the mechanistic details behind this system, however, requires further histological and functional analysis of the bifurcated canal in multiple species.

Because of the presence of two fang sockets in each maxilla, up to eight fangs at different developmental stages on each side of their jaw, and the ability to connect both the original and the replacement fangs equally well to the venom duct at the same time, front‐fanged venomous snakes will always have a functional fang. This finding provides a clear answer to the hypotheses made in previous literature on whether both the replacement and ankylosed fang are functional in delivering venom.

4.2. Verifying diceCT with histology

The small size of the ducts and canals made dissection and correct interpretation of the anatomy very difficult. However, using the histology and linking it with the corresponding slices of the diceCT scan confirmed the anatomy of the proximal portion of the VDS described based on the diceCT scan.

Up until now, attempts to visualise this final region of the VDS by dissection or histology have been unsuccessful (Schaefer, 1976; Young et al., 2001). Both these studies likely missed the reservoir and bifurcating canal because of the very small size of these structures, which we found ranged between 0.03 and 0.08 mm in diameter. Where possible, combining diceCT and histological imaging is best practise in these types of studies because it provides the ability to cross‐validate and reinforce the observations made using both methods.

During the dissection of the VDS, recognising whether the portion of interest, the location of the bifurcating canal, was successfully removed was impossible to determine. Based on the histological sections, we can conclude that some parts of this tissue were not collected for histological analysis. To visualise the complete bifurcation and connection to the fang, we suggest exploring the use of demineralisation of the head so that the complete region can be histologically sectioned without the need for dissection.

4.3. Presence of a venom reservoir

Our results show that viperids and atractaspidins have the simplest duct anatomy, where the duct exits the gland (accessory gland and venom gland, respectively) and leads directly to the entrance orifices of the fangs. There, it bifurcates just before reaching the fangs to connect to both separately. Elapids have a slightly more complicated anatomy where the duct does not connect directly to the fangs’ entrance orifices, but rather expands into a large venom reservoir, where the venom may accumulate. From this reservoir, two terminal venom canals lead to the fangs’ orifices.

Presence of a venom reservoir anterior to the fang's entrance orifice in elapid snakes has been mentioned before, where a crescent‐shaped cavity was observed within the dental sheath anterior to the fang (Schaefer, 1976). When viperids were investigated, no evidence was found of a reservoir structure (Schaefer, 1976). The function of the cavity was hypothesised to help seal the area around the base of the fang and add additional pressure to the venom (Schaefer, 1976). The proposed function of forming a buffer before the fangs is a likely explanation because in contrast to atractaspidins and viperids, the fangs in elapids are always erect and therefore always maintain an open connection with the venom duct. If small amounts of venom were exuded from the venom gland unintentionally, a reservoir might hold this venom and prevent it from leaking into the snake's mouth. For viperids and atractaspidins, it has been hypothesised that the folding of their maxilla obstructs the venom flow (Young et al., 2001), and they would, therefore, not need a venom reservoir.

In contrast to our findings and those of Schaefer (1976), Mead (1702) and more recently Young et al., (2001) proposed the existence of a venom reservoir anterior to the fang of viperid snakes. Mead (1702: 43–44) described the path of the venom in viperids as follows: “’Tis a Conglomerated Gland, composed of many smaller ones contained in a common Membrane; each of These send off an Excretory Vessel, all which do afterwards Unite and Form one Duct, which running towards the Roots of the Fangs, discharges the Yellow Liquor into a Bag. This bag … [covers] the Fangs near the Root … This consists of Muscular Fibres, both Longitudinal and Circular, by Means of which it can Contract it self when the Fangs are erected; and by this Contraction the Venom is press'd into the Hole at the Root of the Tooth, and forced out at the Fissure near the Point” (Mead, 1702). Young and colleagues verified Mead's findings by injecting the distal part of the venom duct of Crotalus atrox (viperid) with blue India ink. They found that venom pooled at the base of the ankylosed fang and confirmed the presence of a tissue chamber (Young et al., 2001).

From the microCT scans and histological imaging, the venom reservoir of an elapid snake seems to be a passive structure, with no smooth muscle surrounding it that could contract to add additional pressure and push venom into the fang venom tube. In diceCT scans, muscle tissue elsewhere in the head is strongly stained by iodine and is clearly visible in the scans. The lack of strong staining associated with this reservoir indicates there is no obvious muscular layer in the wall of the reservoir. To further establish this Masson's trichrome staining was used where only connective tissue was identified around the reservoir and duct. In the two viperid species included in this study, no reservoir was observed in the diceCT scans, whereas Mead (1702) described this structure in Crotalus. It is, thus, possible that either viperids do not have a reservoir, contradicting Mead's (1702), or the reservoir is not homologous across viperids. From our findings, we can confidently say that there is no muscular tissue associated with the reservoir, duct or terminal canals, contradicting Mead's (1702) description.

To conclude, our results found a crescent‐shaped reservoir in elapids as proposed by Schaefer (1976) and found no evidence of such a structure in the examined viperids and atractaspidins. All five Elapidae species examined here display a venom reservoir. It is reasonable to conclude that a venom reservoir probably is present in most Elapidae since we observed it in two closely related Australian species (A. superbus and N. scutatus), two more distantly related species (P. textilis (Australia) and W. aegyptia (Middle East)), as well as in a sea snake from the subfamily Hydrophiinae (H. platurus), thus spanning a considerable portion of the Elapidae. The two viperid species examined in this study (D. palaestinae and E. coloratus) had no venom reservoir, although it has been previously observed in a viperid species (Mead, 1702; Young et al., 2001). We are aware that the two species examined by us represent a very small part of viperid diversity and, therefore, encourage others to further investigate the presence of a venom reservoir in many other Viperidae species as well as in Atractaspidinae. This, in turn, indicates that the simplified VDS model presented below might not be the rule for all Viperidae and Atractaspidinae species.

4.4. Overview of the complete venom delivery system

The present study describes the complete duct anatomy as well as the presence of a bifurcating duct that connects to the fangs for the first time and thus forms the final piece of the puzzle of the VDS. Below we include the revised and now complete anatomy of the VDS of front‐fanged venomous snakes (Figure 5).

Within Toxicofera, protein secreting glands evolved along the jaws, with snakes only maintaining the glands along the upper jaw (Chipman, 2009; Fry et al., 2006). Throughout evolution these gland have undergone major modifications in each of the venomous snake families independently. Front‐fanged snakes all possess venom glands with appreciable lumens, different from Duvernoy's glands present in rear‐fanged snakes (Fry et al., 2003; Gans, 1978; Kochva, 1978). In Elapidae, the venom gland consists of clumped tubular cisternae lined with secretory cells that have long secretory tubules; these all empty into a central lumen from where the venom flows into the venom duct that runs through the accessory gland (Berkovitz & Shellis, 2016; Kerkkamp, Casewell & Vonk, 2017; Weinstein, Smith & Kardong, 2009) (Figure 5). Atractaspidinae have a venom gland comprised of spoke‐like tubules that also empty into a central lumen connected to the venom duct (Berkovitz & Shellis, 2016; Kerkkamp, Casewell & Vonk, 2017) (Figure 5). The glands present in Viperidae consist of clumped tubular cisternae lined with secretory cells that empty in a large basal lumen which leads to a venom duct, which in turn leads to an accessory gland followed by another venom duct (Berkovitz & Shellis, 2016; Kerkkamp, Casewell & Vonk, 2017; Weinstein, Smith & Kardong, 2009).

The compressor muscle (compressor glandulae) associated with the VDS lies proximally and is responsible for squeezing the venom gland and pushing the venom through the rest of the system. The origin of this muscle varies in all three snake taxa discussed here: in Elapidae, this compressor muscle evolved from the adductor externus superficialis; in Viperidae, it evolved from the adductor externus profundus and in the Atractaspidinae from the adductor externus medialis (Kerkkamp, Casewell & Vonk, 2017). Beyond the front‐fanged snakes, venom gland compressor muscles have evolved on three additional occasions as they are present in Brachyophis, Dispholidus and Gonionotophis (formerly Mahelya) (Taub, 1967). This results in the evolution of a compressor muscle associated with the venom gland on at least six occasions in total.

All taxa possess a venom duct that carries the venom from a glandular structure to the base of the fang. In Elapidae and Viperidae this duct is associated with an accessory gland. In Elapidae the accessory gland buds off the venom gland, while in Viperidae the primary venom duct forms the connection between venom and accessory glands and the secondary duct connects the accessory gland to the entrance orifice of the fang (Kerkkamp, Casewell & Vonk, 2017).

From this study, we now also know the differences and similarities in the connection between duct and fang. From the five elapid species examined, we find that all of them possess a venom reservoir positioned at the base of the fangs; from there, two canals make their way to the fangs’ entrance orifices. In the species of Viperidae and Atractaspidinae examined here, no such reservoir was observed and the venom duct simply bifurcated into two smaller canals leading to the fangs’ orifices. In all three taxa the physical connection between the venom canal and the orifice is made anew every time a new replacement fang ankyloses to the maxilla. With these findings we provide an additional line of evidence to support the hypothesis that the VDS evolved independently within each (sub)family of front‐fanged snakes.

AUTHOR CONTRIBUTIONS

The study was designed by SGCC and supervised by ARE and DPH. Scanning of museum specimens on the Zeiss Xradia microCT scanner at Monash Engineering was performed by ARE. Specimen preparation and staining was performed by SGCC and HLR. 3D computer modelling and interpretation of the data were conducted by SGCC. Snake head dissection was overseen by SGCC and executed by HLR and WMGP. WMGP and ARE arranged the histological imaging, photographing and aided in the interpretation. SGCC wrote the manuscript on which all authors provided feedback.

Supporting information

Supplementary Material

Click here for additional data file. (9.4MB, pdf)

Atractaspis‐engaddensis

Click here for additional data file. (2.7MB, pdf)

Daboia‐palaestina

Click here for additional data file. (3.3MB, pdf)

Pseudonaja‐text

Click here for additional data file. (5.4MB, pdf)

ACKNOWLEDGEMENTS

Thanks to Karen Roberts, Jane Melville and Ricky‐Lee Erickson (Museums Victoria, Australia) for giving us access to specimens and approving destructive sampling by means of Lugol’s staining for the snakes on loan. Thanks to Shai Meiri and Erez Maza (Steinhardt Museum of Natural History, Tel Aviv) for providing fixed specimens. We would like to thank Camilla Cohen (Histology Platform Manager) from the Department of Anatomy and Developmental Biology at Monash University for performing the histological preparation, sectioning and staining. Also thanks to the Monash Earth Atmosphere and Environment (EAE) microscope facility for granting access to equipment. Thanks to Dr. Amanda Meyer (University of Western Australia) who corroborated the histological findings.

Cleuren, S.G.C. , Parker, W.M.G. , Richards, H.L. , Hocking, D.P. & Evans, A.R. (2022) Sharp and fully loaded: 3D tissue reconstruction reveals how snake fangs stay deadly during fang replacement. Journal of Anatomy, 240, 1–10. 10.1111/joa.13531

Funding information

This paper is part of the PhD project of SGCC funded by the Monash Graduate Scholarship and the Monash International Tuition Scholarship. This project was also partly funded by the Holsworth Wildlife Research Endowment—Equity Trustees Charitable Foundation & the Ecological Society of Australia. Support was given by Australian Government Research Training Program Scholarships (WMGP and HLR) and Robert Blackwood Partnership PhD Top‐Up Scholarships (WMGP and HLR). ARE was supported by an Australian Research Council Discovery Project (DP180101797).

DATA AVAILABILITY STATEMENT

DiceCT scans and generated 3D models of the venom delivery system are available on the MorphoSource database under the ‘Sharp and fully loaded: supporting diceCT data’ Project (ID: 366918) (URL: https://www.morphosource.org/projects/000366918).

REFERENCES

  1. Berkovitz, B. & Shellis, P. (2016) Enameloid and enamel. The Teeth of Non‐Mammalian Vertebrates, Amsterdam: Elsevier, pp. 311–330. 10.1016/b978-0-12-802850-6.00012-6. [DOI] [Google Scholar]
  2. Burnett, W.I. (1854) The poison apparatus of the rattlesnake. Proc. Boston Soc. Nat. Hist., 5, 31–34. [Google Scholar]
  3. Chipman, A.D. (2009) On making a snake. Evolution and Development, 5, 3–5. 10.1111/j.1525-142X.2008.00295.x. [DOI] [PubMed] [Google Scholar]
  4. Dzialowski, E.M. (2018) Comparative physiology of the ductus arteriosus among vertebrates. Seminars in Perinatology, 42(4), 203–211. 10.1053/j.semperi.2018.05.002. [DOI] [PubMed] [Google Scholar]
  5. Edmund, A.G. (1969) Dentition. In: Gans, C. & Parsons, T.S. (Eds.) Biology of the Reptilia I. Morphology A. Vol. 1, New York: Academic Press, pp. 117–200. [Google Scholar]
  6. Fry, B.G. et al. (2003) Isolation of a neurotoxin (a ‐colubritoxin) from a nonvenomous colubrid: Evidence for early origin of venom in snakes. Journal of Molecular Evolution, 57, 446–452. 10.1007/s00239-003-2497-3. [DOI] [PubMed] [Google Scholar]
  7. Fry, B.G. , Scheib, H. , van der Weerd, L. , Young, B. , McNaughtan, J. , Ramjan, S.F.R. et al. (2008) Evolution of an Arsenal ‐ Structural and functional diversification of the venom system in the advanced snakes (Caenophidia). Molecular & Cellular Proteomics, 7(2), 215–246. 10.1074/mcp.M700094-MCP200. [DOI] [PubMed] [Google Scholar]
  8. Fry, B.G. , Vidal, N. , Norman, J.A. , Vonk, F.J. , Scheib, H. , Ramjan, S.F.R. et al. (2006) Early evolution of the venom system in lizards and snakes. Nature, 439(7076), 584–588. 10.1038/nature04328. [DOI] [PubMed] [Google Scholar]
  9. Gans, C. (1978) Reptilian venoms: Some evolutionary considerations. In: Gans, C. (Ed.) Biology of the Reptilia, Physiology B. New York: Academic Press INC, pp. 1–42. [Google Scholar]
  10. Gignac, P.M. & Kley, N.J. (2014) Iodine‐enhanced micro‐CT imaging: Methodological refinements for the study of the soft‐tissue anatomy of post‐embryonic vertebrates. Journal of Experimental Zoology Part B: Molecular and Developmental Evolution, 322(3), 166–176. 10.1002/jez.b.22561. [DOI] [PubMed] [Google Scholar]
  11. Iwanaga, S. & Suzuki, T. (1979) Enzymes in snake venom. In: Lee, C.Y. (Ed.) Snake venoms. Handbook of Experimental Pharmacology (Vol. 52). Berlin, Heidelberg: Springer, pp. 61–158. 10.1007/978-3-642-66913-2_4. [DOI] [Google Scholar]
  12. Jackson, K. (2007) The evolution of venom‐conducting fangs: Insights from developmental biology. Toxicon, 49(7), 975–981. 10.1016/j.toxicon.2007.01.007. [DOI] [PubMed] [Google Scholar]
  13. Jackson, K. & Fritts, T.H. (1995) Evidence from tooth surface morphology for a posterior maxillary origin of the proteroglyph fang. Amphibia‐Reptilia, 16, 273–288. [Google Scholar]
  14. Jeffery, N.S. , Stephenson, R.S. , Gallagher, J.A. , Jarvis, J.C. & Cox, P.G. (2011) Micro‐computed tomography with iodine staining resolves the arrangement of muscle fibres. Journal of Biomechanics, 44(1), 189–192. 10.1016/j.jbiomech.2010.08.027. [DOI] [PubMed] [Google Scholar]
  15. Kardong, K.V. (1979) “Protovipers” and the evolution of snake fangs. Evolution, 33, 433–443. [DOI] [PubMed] [Google Scholar]
  16. Kerkkamp, H.M.I. , Casewell, N.R. & Vonk, F.J. (2017) Evolution of the snake venom Delivery System. In: Malhotra, A. (Ed.) Evolution of venomous animals and their toxins. Dordrecht: Springer, pp. 303–316. 10.1007/978-94-007-6458-3_11. [DOI] [Google Scholar]
  17. Kılıç, D.Ç. , Akat, E. & Arıkan, H. (2016) A histological study on venom apparatus of Montivipera xanthina (Gray, 1849) (Serpentes, Viperidae). Brazilian Archives of Biology and Technology, 59, 1–5. 10.1590/1678-4324-2016160040. [DOI] [Google Scholar]
  18. Klinkhamer, A.J. , Wilhite, D.R. , White, M.A. & Wroe, S. (2017) Digital dissection and three‐dimensional interactive models of limb musculature in the Australian estuarine crocodile (Crocodylus porosus). PLoS One, 12(4), 17–22. 10.1371/journal.pone.0175079. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Kochva, E. (1978) Oral glands of the reptilia. In: Gans, C. (Ed.) Biology of the reptilia, physiology B. New York, NY: Academic Press INC, pp. 43–162. [Google Scholar]
  20. McMillan, D.B. & Harris, R.J. (2018) Chapter K. The digestive system. An Atlas of Comparative Vertebrate Histology, 1st, Amsterdam: Elsevier, pp. 291–387. 10.1016/B978-0-12-410424-2.00011-1. [DOI] [Google Scholar]
  21. Mead, R. (1702) A mechanical account of poisons in several essays. London: Ralph Smith. [Google Scholar]
  22. Metscher, B.D. (2009a) Micro CT for comparative morphology: Simple staining methods allow high‐contrast 3D imaging of diverse non‐mineralized animal tissues. BMC Physiology, 9(1), 11. 10.1186/1472-6793-9-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Metscher, B.D. (2009b) MicroCT for developmental biology: A versatile tool for high‐contrast 3D imaging at histological resolutions. Developmental Dynamics, 238(3), 632–640. 10.1002/dvdy.21857. [DOI] [PubMed] [Google Scholar]
  24. Nagy, Z.T. et al. (2013) 3D reconstruction of fang replacement in the venomous snakes Dendroaspis jamesoni (Elapidae) and Bitis arietans (Viperidae). Salamandra, 49(2), 109–113. [Google Scholar]
  25. Nasrullah, Q. , Renfree, M.B. & Evans, A.R. (2018) Three‐dimensional mammalian tooth development using diceCT. Archives of Oral Biology, 85, 183–191. 10.1016/j.archoralbio.2017.10.018. [DOI] [PubMed] [Google Scholar]
  26. Osborn, J.W. (1974) On the control of tooth replacement in reptiles and its relationship to growth. Journal of Theoretical Biology, 46(2), 509–527. 10.1016/0022-5193(74)90012-5. [DOI] [PubMed] [Google Scholar]
  27. Pyron, R.A. et al. (2014) Effectiveness of phylogenomic data and coalescent species‐tree methods for resolving difficult nodes in the phylogeny of advanced snakes (Serpentes: Caenophidia). Molecular Phylogenetics and Evolution, 81, 221–231. 10.1016/j.ympev.2014.08.023. [DOI] [PubMed] [Google Scholar]
  28. Richman, J.M. & Handrigan, G.R. (2011) Reptilian tooth development. Genesis, 49(4), 247–260. 10.1002/dvg.20721. [DOI] [PubMed] [Google Scholar]
  29. Savitzky, A.H. (1980) The role of venom delivery strategies in snake evolution. Evolution, 34(6), 1194–1204. [DOI] [PubMed] [Google Scholar]
  30. Schaefer, N. (1975) Fang replacement in the snakes dispholidus typus and thelotornis kirtlandii. Herpetologica, 31(1), 102–104. [Google Scholar]
  31. Schaefer, N. (1976) The mechanism of venom transfer from the venom duct to the fang in snakes. Herpetologica, 32(1), 71–76. [Google Scholar]
  32. Taub, A.M. (1967) Comparative histological studies on Duvernoy’s gland of colubrid snakes. Bulletin of the American Museum of Natural History, 138(1), 1–50. [Google Scholar]
  33. Vidal, N. (2002) Colubroid systematics: Evidence for an early appearance of the venom apparatus followed by extensive evolutionary tinkering. Journal of Toxicology: Toxin Reviews, 21, 21–41. 10.1081/TXR-120004740. [DOI] [Google Scholar]
  34. Vonk, F.J. , Admiraal, J.F. , Jackson, K. , Reshef, R. , de Bakker, M.A.G. , Vanderschoot, K. et al. (2008) Evolutionary origin and development of snake fangs. Nature, 454(7204), 630–633. 10.1038/nature07178. [DOI] [PubMed] [Google Scholar]
  35. Weinstein, S.A. , Smith, T.L. & Kardong, K.V. (2009) Reptile venom glands: Form, function, and future. In Mackessy, S.P. (Ed.) Handbook of venoms and toxins of reptiles. London: CRC Press Inc, pp. 65–92. [Google Scholar]
  36. Young, B.A. , Blair, M. , Zahn, K. & Marvin, J. (2001) Mechanics of venom expulsion inCrotalus, with special reference to the role of the fang sheath. The Anatomical Record, 264, 415–426. [DOI] [PubMed] [Google Scholar]
  37. Zahradnicek, O. , Horacek, I. & Tucker, A.S. (2008) Viperous fangs: Development and evolution of the venom canal. Mechanisms of Development, 125(9–10), 786–796. 10.1016/j.mod.2008.06.008. [DOI] [PubMed] [Google Scholar]
  38. Zeller, E. (1948) Enzymes of snake venoms and their biological significance. In Nord, F.F. (Ed.) Advances in enzymology and related subjects of Biochemistry. VIII New York, NY: Interscience Publishers Inc, pp. 459–491. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplementary Material

Click here for additional data file. (9.4MB, pdf)

Atractaspis‐engaddensis

Click here for additional data file. (2.7MB, pdf)

Daboia‐palaestina

Click here for additional data file. (3.3MB, pdf)

Pseudonaja‐text

Click here for additional data file. (5.4MB, pdf)

Data Availability Statement

DiceCT scans and generated 3D models of the venom delivery system are available on the MorphoSource database under the ‘Sharp and fully loaded: supporting diceCT data’ Project (ID: 366918) (URL: https://www.morphosource.org/projects/000366918).

Articles from Journal of Anatomy are provided here courtesy of Anatomical Society of Great Britain and Ireland

RESOURCES