Immunolocalization and characterization of cornification proteins in snake epidermis



Little is known about specific proteins involved in keratinization of the epidermis of snakes, which is composed of alternating β- and α-keratin layers. Using immunological techniques (immunocytochemistry and immunoblotting), the present study reports the presence in snake epidermis of proteins with epitopes that cross-react with certain mammalian cornification proteins (loricrin, filaggrin, sciellin, transglutaminase) and chick β-keratin. α-keratins were found in all epidermal layers except in the hard β- and α-layers. β-keratins were exclusively present in the oberhautchen and β-layer. After extraction and electrophoresis, α-keratins of 40–67 kDa in molecular weights were found. Loricrin-like proteins recorded molecular weights of 33, 50, and 58 kDa; sciellin, 55 and 62 kDa; filaggrin-like, 52 and 65 kDa; and transglutaminase, 45, 50, and 56 kDa. These results suggest that α-layers of snake epidermis utilize proteins with common epitopes to those present during cornification of mammalian epidermis. The β-keratin antibody on extracts from whole snake epidermis showed a strong cross-reactive band at 13–16 kDa. No cross-reactivity was seen using an antibody against feather β-keratin, indicating absence of a common epitope between snake and feather keratins. © 2005 Wiley-Liss, Inc.

Snakes renew their epidermis through a cyclical loss of the superficial portion of the epidermis that is eliminated as a single piece (molt). The epidermis of snakes is made of six epidermal layers termed oberhautchen, β, mesos, α, lacunar, and clear, which are produced during the renewal phase of the epidermis (Maderson, 1985; Landmann, 1986; Maderson et al., 1998). A new sequence of epidermal layers is cyclically produced beneath the previous one. The oberhautchen layer of the new (inner) epidermal generation is connected to the clear layer of the old (outer) generation before shedding. During their differentiation, clear and oberhautchen cells produce an interface where complementary spinulae form a zip-fastened structure. The latter rapidly opens up at maturation so that shedding takes place (Roth and Jones, 1967, 1970; Roth and Maderson, 1968; Landmann, 1979; Maderson et al., 1998; Alibardi, 2002a; Alibardi and Thompson, 2003).

A soft type of keratin (α) is synthesized in the α-layers (mesos, α, lacunar, and clear), while a harder type of keratin (β) is produced in the oberhautchen and β-layer (Gregg and Rogers, 1986; Sawyer et al., 2000; Alibardi and Sawyer, 2002). α-keratin accumulated in the narrow cells of the mature α-layer is pliable and houses lipids for the formation of a barrier against water loss. β-keratin forms a sincytial corneous layer that gives mechanical resistance to snake scales. The above studies have shown that during the process of epidermal maturation in snakes (and lizards), the oberhautchen layer loses its plasma membrane and merges with the multistratified β-layer.

Previous light microscopic autoradiographic studies have shown that the shedding complex in lepidosaurian reptiles (lizards and snakes) takes up most of tritiated histidine, which, apart from the intense metabolism of its cells, might be incorporated into specific proteins for the cornification of clear and oberhautchen cells (Alibardi, 2001, 2002a, 2002b; Alibardi et al., 2003). Biochemical and ultrastructural autoradiographic studies on histidine uptake in lizard epidermis have shown that, apart from keratins, nonkeratin histidine-rich proteins may also be present. It is unknown whether the neosynthesized keratin and nonkeratin material may contain histidine-rich proteins also in snake epidermis. Very little is presently known about the specific proteins involved in the cornification of snake epidermis.

In order to improve the knowledge of the process of differentiation and cornification of the epidermis of snakes, we have conducted a study by immunocytochemistry, autoradiography, and immunoblotting after electrophoresis of snake epidermal proteins. The study shows that proteins with some cross-reactivity with antibodies directed against proteins involved in mammalian cornification are present in the epidermis of snakes, together with unique β-keratins of low molecular weight.


Three juveniles (about 2 months posthatching) of the corn snake (Elaphe guttata, colubrids) were utilized for the histological and biochemical analysis of the whole epidermis (living plus corneous layers). Additionally, three juveniles of the grass snake (Natrix natrix, colubrids) were used as previously reported (Alibardi, 2002a, 2002b). Their epidermis was in the renewal phase. They were injected intraperitoneally with a saline solution containing tritiated histidine (L 2,5-3H-histidine, specific activity 40–60 Ci/mmole; Amersham) at a concentration of 5–8 μCi/g body weight.

The skin of three water python snakes (Liasis fuscus, pythonids) of about 2.5 months posthatching and in the renewal phase was collected and fixed for immunocytochemistry without any experimental treatment (Alibardi and Thompson, 2003). Since the epidermis of N. natrix and L. fuscus was in the renewal stage, it was possible to study with precision the formation of the shedding complex.

Four hours after the injection of tritiated histidine, the snakes (N. natrix) were killed by decapitation and skin from various areas of the body was collected. Pieces of tissues between 1 and 4 mm in length were immediately fixed for 5–8 hr in cold (0–4°C) phosphate buffer solution (0.1 M; pH 7.2–7.4) containing 2.5% glutaraldehyde. After 30-min rinsing in the same buffer, the tissues were immersed for 90–120 min in 2% OsO4, dehydrated by ethanol, immersed in propylene oxide, and embedded in the resin Durcupan.

Microscopic Techniques

After sectioning with an ultramicrotome, 1–4 μm thick sections were serially collected over microscope slides. Thin sections (40–90 nm thick) were collected over collodion-coated slides for the ultrastructural autoradiographic analysis. Thick sections were coated with K5 (for light microscopic autoradiography) or with L4 (for ultrastructural autoradiography) Nuclear Emulsion (Ilford, Cheshire, U.K.) in a darkroom. Thick sections were left for 1–2 months to expose for tracer localization, while thin sections were exposed for 3–4 months. After developing with D19 (Kodak, Rochester, NY) and fixing in Ilford fixer, sections were slightly stained in 0.5% toluidine blue or were left unstained. Thin sections supported by the collodium film were then collected on copper or nickel grids, stained with uranyl acetate and lead citrate, and observed under a Philips CM-100 electron microscope.

Other pieces of skin of N. natrix were fixed in 4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, for 5 hr, dehydrated in 80% ethanol, and embedded in the hydrophilic resin Bioacryl under UV light at 0–4°C (Scala et al., 1992). Skin from L. fuscus was fixed in 4% paraformaldehyde as above and embedded in the hydrophilic resin Lowicryl K4M under UV polymerization at 0–4°C. Sections of both N. natrix and L. fuscus skin were obtained with an ultramicrotome at 1–3 μm thickness for immunocytochemical analysis.

The β1 antibody, produced in rabbit against a chick scale β-keratin, and an FBK antibody, produced against chick feather β-keratin, were generously donated by Dr. R.H. Sawyer, Biological Science Department, University of South Carolina (Sawyer et al., 2000). The antiloricrin antibody, produced in rabbit against a 15 amino acidic sequence toward the C-terminal of mouse loricrin (Mehrel et al., 1990), was purchased from BaBco. An AE3 antibody, a broadly characterized antibody against basic cyto(α-)keratins (Sun et al., 1983), was purchased from Progen (Heidelberg, Germany). The antibody against guinea pig transglutaminase 1 (ab421, rabbit polyclonal) was purchased from Abcam (Cambridge, U.K.). The antibody against human sciellin was generously donated by Dr. H. Baden, Massachusetts General Hospital (Kvedar et al., 1992). Finally, the antirat filaggrin antibody was generously donated by Dr. B.A. Dale, Department of Oral Biology, University of Washington (Presland et al., 1997).

Sections were preincubated for 30 min in 5% normal goat serum and 2% bovine serum albumine (BSA) in 0.05 M Tris/HCl buffer, pH 7.6, in order to neutralize nonspecific antigenic sites on the sections, then incubated overnight at 4°C in the BSA-Tris buffer containing the primary antibody (dilutions 1:100–200 for β1 and 1:200 for loricrin). In control sections, the primary antibody was omitted. After rinsing in the BSA-Tris buffer, the sections were incubated in the same buffer for 1 hr at room temperature containing 1:50 of antirabbit-IgG FITC-conjugated secondary antibodies. After extensive rinsing, sections were mounted in Fluoromount (EM Sciences) and observed under a Zeiss epifluorescence microscope equipped with a fluorescein filter. Positive controls for each of the antibodies utilized produced immunostaining (Alibardi and Maderson, 2003).

For immunoelectron microscopy, 40–90 nm thick sections were collected on nickel grids and immunostained with the primary antibody as above using 1% cold-water fish gelatin in 0.05 M Tris-HCl buffer to block nonspecific binding sites. An antirabbit IgG conjugated to 10 nm large gold particles (Sigma or Biocell) was used as secondary antibody. Sections were studied, unstained or lightly stained with 2% uranyl acetate, under a CM-100 Philips electron microscope.

Electrophoresis and Immunoblotting

Samples of the epidermis from Elaphe guttata (colubrid snake) were extracted and analyzed by electrophoresis according to the method by Sybert et al. (1985). Briefly, the skin was incubated in 5 mM EDTA in phosphate-buffered saline for 5 min at 50°C and 2–4 min in cold buffer. The epidermis was separated from the dermis by dissection under the stereomicroscope. The epidermis and the molts were homogenized in 8 M urea, 50 mM Tris-HCl, pH 7.6, 0.1 M 2-mercaptoethanol, 1 mM dithiothreithol, 1 mM phenylmethylsulphonyl fluoride, and the particulate matter was removed by centrifugation at 10,000 rpm for 10 min. Protein concentration was assayed by the Bio-Rad method and mini Protean III cells; immunoblotting and detection systems were used to fractionate the extracted proteins. Proteins, denaturated in the sample buffer, were separated in 12% or 15% SDS-polyacrylamide gels (SDS-PAGE) according to Laemmli (1970). For Western blotting, the proteins were separated by SDS-PAGE and transferred to nitrocellulose paper to be incubated with primary antibodies (dilution 1:500 for loricrin and sciellin; 1:2,000 for filaggrin and transglutaminase; 1:3,000 for β1) in TBS-Tween + 5% nonfat milk powder. Detection was performed by using the enhanced chemiluminescence procedure developed by Amersham (ECL). Controls were achieved by omitting the primary antibody in the incubation solution.


Microscopic Observations

The renewing epidermis of the outer scale surface of N. natrix showed a superficial compact β-layer followed by a dark mature α-layer (Fig. 1). Beneath the latter, a pale and immature lacunar (α) layer were present. All the living layers (but not the cornified outer β-layer) showed some immunoreactivity for the AE3 antibody, including in the forming inner β-layer (Fig. 2). This was due to the cross-reactivity of the AE3 antibody with dense bundles of α-keratin in immature keratinocytes (Fig. 3).

Figure 1.

Histological structure of renewal epidermis of N. natrix showing the outer β-layer (arrow), the outer α-layer (arrowhead), the immature α-layer above the forming inner β-layer (double arrowhead). Magnification = 700×. Figures 1–11 are light and electron microscopic views of snake epidermis. a, α-layer; AE3, AE3 immunolabeling for α-keratin; b, β-layer; BE, immunoreactivity for β-keratin; bi, inner β-layer; bo, outer β-layer; bp, β-keratin packets; c, clear layer; g, germinal (basal) layer; HIS, tritiated histidine autoradiographic labeling; l, lacunar layer; m, mesos layer; o, oberhautchen layer; s, oberhautchen spinulae.

Figure 2.

AE3-immunolabeled living epidermal layers of N. natrix (the outer β-layer has been dislocated by sectioning). Magnification = 460×.

Figure 3.

Ultrastructural detail of AE3-labeled (arrowheads) keratin bundles in immature cell of the α-layer of N. natrix. Magnification = 53,500×.

Histidine labeling was diffuse in all living layers of the epidermis (except corneous layers), but the silver grains particularly concentrate along the shedding layer, especially the oberhautchen and clear layers (Fig. 4). The β1 antibody reacted weakly with the oberhautchen layer but intensely with the outer and especially the inner β-layer (Figs. 5 and 6). The oberhautchen layer was interfaced with the upper clear layer along an undulated shedding line (where detachment of the outer from the inner epidermal generation will occur at maturity producing the molt; Fig. 7). The cytological difference between clear and oberhautchen layers was appreciated under the electron microscope. While coarse, fibrous material accumulated along the basal-most membrane of the clear layer; in the underlying oberhautchen layer, many dense β-keratin packets were present, especially within the short cytoplasmic protrusions (spinulae) interdigitating with the clear layer (Fig. 7). The dense β-keratin packets along the shedding line were the sites of maximal uptake of tritiated histidine, especially in the oberhautchen spinulae (Fig. 8). The immunogold labeling of the dense β-keratin packets of the oberhautchen was much lower than the intense labeling of the electron-pale β-keratin packets of the underlying β-cells (Figs. 9 and 10). During the following maturation of β-cells, they merged into a sincytium with the superficial oberhautchen layer. β-packets enlarged to form long β-keratin filaments that eventually merged into a uniformly immunolabeled mass of β-keratin. Spinulae of the oberhautchen at the epidermal surface were still poorly immunolabeled with the β1 antibody (Fig. 11).

Figure 4.

Ultrastructural autoradiography of renewing epidermis of N. natrix after 4 hr from tritiated histidine injection. The spinulae of the outer oberhautchen (arrow) and the cornified β- and α-outer layers are unlabeled. Silver grains increase from the maturing α-layer (arrowheads) to the clear layer and especially over the oberhautchen. Magnification = 6,400×. The inset shows that the higher labeling of the oberhautchen layer forms a labeled line (arrow) as visualized under the light microscope. Magnification = 440×.

Figure 5.

Detail of the β1 immunolabeling of the oberhautchen spinulae (arrowhead) and inner β-layer of L. fuscus epidermis. Magnification = 580×.

Figure 6.

β1-immunolabeled outer and inner β-layer in L. fuscus epidermis. Magnification = 690×.

Figure 7.

Ultrastructural detail of the forming shedding complex of N. natrix (localized as indicated by the arrow in the inset of Fig. 4). Fibrous material (arrowheads) is deposited in the clear layer connected by desmosomes (double arrowheads) to the cross-sectioned oberhautchen spinulae in which dense β-keratin material (arrows) is accumulated. Magnification = 28,700×.

Figure 8.

Autoradiographic ultrastructural detail of the poorly labeled fibrous material (arrowheads) in clear layer while most of the trace grains are localized on the small dense β-keratin (arrows) converging into oberhautchen spinulae. Magnification = 24,500×.

Figure 9.

Low β1 immunogold labeling of dense β-keratin packets (double arrowhead) in oberhautchen layer in comparison to the high labeling of the pale β-keratin packets (arrowhead) in β-cells of L. fuscus. The two layers are still distinct and joined by desmosomes (arrows). Magnification = 44,000×.

Figure 10.

Detail of β1-labeled β-packets among residual α-keratin bundles (arrowheads) in differentiating β-cells of L. fuscus. Magnification = 63,200×.

Figure 11.

Compact outer β-layer of N. natrix with homogeneous β1 immunolabeling, except in the outer spinulae of the oberhautchen. The arrowheads indicated an embedded melanosome. Magnification = 73,000×.


Protein bands extracted from the whole epidermis of E. guttata were generally in the range of 40–80 kDa (Fig. 12, lane A). Some of these bands showed cross-reactivity with the different antibodies. The pan-cytokeratin antibody showed immunopositive bands within a range of 40–67 kDa (Fig. 12, lane B); for loricrin, the bands were at 33, 48–50, and especially 57–58 kDa (Fig. 12, lane C). With sciellin antibody, bands at 40–42, 56–57, and a stronger one at 62 kDa were seen (Fig. 12, lane D). The antifilaggrin antibody showed bands at 38–40, 50, and especially at 62–64 kDa (Fig. 12, lane E), and the antibody against transglutaminase reacted with bands at 42, 45, and 55–57 kDa (Fig. 12, lane F). Finally, a strong band at 13–16 kDa was seen using the β1 keratin antibody, and very weak at 37 and 64–67 kDa (Fig. 12, lane G). The FBK antibody did not produce any labeled band at low molecular weight (the β-keratin range) but only faint bands at 50 and 57 kDa (data not shown). No bands at all were seen using only the secondary antibody (omitting the primary antibody; data not shown).

Figure 12.

Electrophoretic patterns for proteins in whole epidermis of E. guttata. Red ponceau (lane A), pan-cytokeratin (lane B), loricrin (lane C), sciellin (lane D), filaggrin (lane E), transglutaminase (lane F), β1 (lane G). Polyacrylamide gel at 12%.


α-Keratin and Associated Proteins

Immunocytochemistry on snake epidermis has shown that α-keratins react weakly with the AE3 antibody for mammalian basic keratins (Alibardi, 2002c). Apart from β-keratins, all the remaining immunoreactivities observed in gels are associated with α-keratin layers, suggesting that loricrin-, sciellin-, filaggrin/AE2-positive keratins and transglutaminase-like proteins are present in the epidermis of snakes. Specific preabsorbing controls using the specific antigen searched for (loricrin, sciellin, filaggrin, and transglutaminase) were unavailable in our immunoblotting study. Despite this shortage, our negative controls (omitting the primary antibody) clearly indicate that specific immunoreactive bands were extracted from snake epidermis.

Previous immunocytochemical studies (Alibardi, 2002c) have shown that, when present, loricrin-like immunoreactivity is associated with α-keratin bundles of lacunar and α-cells, but is not localized along the cornified cell envelope as in corneocytes of mammalian epidermis (Steven et al., 1990; Ishida-Yamamoto et al., 2000; Kalinin et al., 2002). Similarly, in the cornified cell envelope, transglutaminase substrates (loricrin, filaggrin, etc.) may have no antigens available for immunodetection, as is the case for mammalian epidermis (Ishida-Yamamoto et al., 2000; Kalinin et al., 2002). The immunolabeling pattern for loricrin, as observed in Western blots, suggests that proteins with some epitopes with mammalian loricrin are present in lacunar cells of snake epidermis (Alibardi, 2002c). A previous study on snake epidermis indicated that a weak protein band of low molecular weight (below 18 kDa) was present (Hohl et al., 1993). The present study instead suggests that the prevalent band has a molecular weight of 57–58 kDa. A cross-reactivity of the loricrin antibody with other proteins of the corneous cell envelope of snake corneocytes is also possible. For example, loricrin shares with involucrin and small proline-rich proteins (two other proteins of the cornified cell envelope of mammalian epidermis) numerous amino acidic sequences in the N- and C-terminal regions (Backendorf and Hohl, 1992), and it is possible that the employed antibody recognizes these common epitopes in different proteins of cornification in snake epidermis. Only a study after isolation of specific cornified envelope proteins from snake epidermis will further clarify the present preliminary data.

Also, the presence of sciellin-immunoreactive protein bands, especially at 62 kDa, suggests that similar protein may be present among keratin filaments of snake α-cells (Kvedar et al., 1992). The immunolocalization of this protein at the microscopic level has still to be done.

The weak filaggrin-like immunoreactivity observed after immunocytochemistry in α-cells and in the lacunar layer of the hinge region of snakes (Alibardi, 2002c) is probably due to a common epitope(s) associated with K1/K10 keratins recognized by the AE2-keratin antibody. The latter recognizes 67–68 and 56.5 kDa α-keratins in mammalian epidermis (Dale and Sun, 1983) and in lizard (Carver and Sawyer, 1987; Alibardi et al., 2000, 2001). The presence of soft α-keratins in the hinge regions of snakes allows the elasticity of these areas among scales, which are used for the movement of snakes. Among α-keratin and associated proteins, a large production of complex lipids and waxes forms the barrier against water loss (Roberts and Lillywhite, 1983; Tu et al., 2002).

Finally, the presence of some weak immunolabeling for transglutaminase in Western blots suggests that a small amount of isopeptide bonds (the product of transglutaminase reaction on proteins) is formed in α-cells, in the peripheral cytoplasm or even along the cornified cell envelope [the marginal layer of Landmann (1979)]. Immunocytochemistry on snake epidermis shows a very weak reaction for isopeptide bonds, especially in α-cells of hinge regions (data not shown).

All the above proteins appear absent in β-cells of the hard corneous layer of snakes. In the latter, β-keratin is the prevalent protein, which accumulates into a dense and compact corneous mass. The prevalent transglutaminase band at 57 kDa found in the present Western blot study is within the range of molecular weight reported for mammalian and chick transglutaminases (Polakowska and Goldsmith, 1991).

Further detailed studies on snake epidermis (also using immunogold cytochemistry) are needed to determine where these proteins enter in the formation of the corneous cell envelope or the internal corneous mass of α-keratinocytes. However, the presence of cross-reactive protein bands for the first time suggests that common molecular processes of cornification as those operating in mammalian epidermis are present in the epidermis of snakes.


β-keratin layers mechanically protect the softer α-layers underneath and therefore the integrity of the water barrier (Maderson et al., 1998; Tu et al., 2002). β-keratin is deposited during the transition from the last formed α-layer (the clear) of the outer epidermal generation to the first β-layer of the inner generation (the oberhautchen), with the result that a shedding complex is formed.

The electron-dense histidine-labeled fibrous material in clear cells and in dense β-keratin packets in oberhautchen cells during the transition between α-keratin to β-keratin synthesis indicates that histidine is mainly used for the production of new keratins (Landmann, 1979; Alibardi, 2002a, 2002b, 2002c; Alibardi and Thompson, 2003). A specific biochemical study on specific protein bands isolated from snake epidermis is required to show whether nonkeratin histidine-rich proteins are also produced, as is the case for oberhautchen cells of lizards and geckos (Alibardi, 2001, 2002b; Alibardi et al., 2003). Histidine-incorporating proteins rapidly decrease in β-cells and their β-packets. Taken together, these observations suggest that the dense β-keratin packets in oberhautchen cells, but not the pale β-keratin of β-cells, contain histidine-rich proteins.

β-keratin is deposited over or among α-keratin bundles, which rapidly disappear in maturing β-cells before the formation of the syncitial and mature outer β-layer. Most of β-keratin-labeled bands from whole epidermis in snakes have a lower molecular weight than α-keratins, as in the epidermis of the other reptiles (Wyld and Brush, 1979, 1983; Gillespie et al., 1982; Marshall and Gillespie, 1982; Carver and Sawyer, 1987; Sawyer et al., 2000). Bands of variable molecular weights (low sulfur fractions) have been reported for lizard (12–70 kDa) (Gillespie et al., 1982) or gecko (12–72 kDa) (Thorpe and Giddings, 1981) β-keratins, although the main fractions (high sulfur) are those at 13–16 and 20–22 kDa.

In the epidermis of snakes, the smaller β-keratins with a molecular weight of 13–16 kDa appear constantly formed. The evolutions of these small keratins in reptiles from α-keratins remain to be solved, but the abundance of glycine-glycine-Tyr or glycine-glycine-Cys sequences (Gillespie et al., 1982; Marshall and Gillespie, 1982) suggests that β-keratins may be derived from the evolution of the gly-gly-rich sequences at the N- or C-terminal of the external nonhelical sequences of α-keratins (Klinge et al., 1987). Future molecular biology studies will definitely disclose the trend followed during the evolution of these hard forms of keratins in reptiles.


Grass snakes were collected by Dr. L. Rugiero (FIZV, Roma, Italy) and water python by Drs. Gavin Bedford and Keith Christian (Northern Territory University, Darwin, Australia) with permission of the NT Conservation Commission. Dr. Brendon Dunphy (University of Auckland) read a draft of the manuscript.