Isolation of a new class of cysteine–glycine–proline-rich beta-proteins (beta-keratins) and their expression in snake epidermis


Lorenzo Alibardi, Dipartimento di Biologia evoluzionistica sperimentale, University of Bologna, via Selmi 3, 40126 Bologne, Italy. T: + 39 051 2094257; F: + 39 051 251208; E:


Scales of snakes contain hard proteins (beta-keratins), now referred to as keratin-associated beta-proteins. In the present study we report the isolation, sequencing, and expression of a new group of these proteins from snake epidermis, designated cysteine–glycine–proline-rich proteins. One deduced protein from expressed mRNAs contains 128 amino acids (12.5 kDa) with a theoretical pI at 7.95, containing 10.2% cysteine and 15.6% glycine. The sequences of two more snake cysteine–proline-rich proteins have been identified from genomic DNA. In situ hybridization shows that the messengers for these proteins are present in the suprabasal and early differentiating beta-cells of the renewing scale epidermis. The present study shows that snake scales, as previously seen in scales of lizards, contain cysteine-rich beta-proteins in addition to glycine-rich beta-proteins. These keratin-associated beta-proteins mix with intermediate filament keratins (alpha-keratins) to produce the resistant corneous layer of snake scales. The specific proportion of these two subfamilies of proteins in different scales can determine various degrees of hardness in scales.


Snakes renew their epidermis through a complex cyclical process termed the ‘shedding cycle’ where the outer part of the mature epidermis is lost as a molt (Maderson, 1985; Landmann, 1986; Maderson et al. 1998; Alibardi, 2002). The renewing epidermis of snakes has been subdivided in a sequence of epidermal layers – oberhautchen, beta-layers, mesos layers and alpha-layer. The outer epidermal generation that gives rise to the molt is connected with inner epidermal generation by desmosomes of the clear layer of the outer generation initially joined to the oberhautchen of the inner generation. In both snakes and lizards, the junctions between the two generations (the last layer, the clear layer of the outer generation, and the first layer of the inner generation, the oberhautchen) are rapidly degraded during the end of the renewal phase of the epidermis so that shedding takes place (Alibardi, 1998).

Previous studies have indicated that a soft type of keratin (alpha) is synthesized in the alpha-layers (mesos, alpha, lacunar and clear), whereas a harder type of keratin (beta- or phi-) is produced in the oberhautchen and beta-layers (Wyld & Brush, 1979, 1983; Thorpe & Giddings, 1981; Maderson, 1985; Landmann, 1986; Sawyer et al. 2000). In modern terms keratins are considered intermediate filament proteins in which the central portion of the molecule presents numerous alpha-helixes (Steinert & Freedberg, 1991). The proteins referred to as beta-keratins do not actually belong to the keratin type as they present numerous differences in gene structure and primary and secondary protein structure, and possess different chemical-physical properties from alpha-keratins (Alibardi et al. 2007, 2009). Beta-keratins are believed to contain numerous amino acid sequences organized in a beta-pleated conformation, responsible for the appearance of the beta-pattern (Baden & Maderson, 1970; Fraser et al. 1972; Baden et al. 1974; Gregg & Rogers, 1986). The different program of synthesis of alpha-keratin vs. that of beta-keratins is cyclically activated during the renewal phase of the shedding cycle (Maderson, 1985; Maderson et al. 1998).

Recent molecular studies have identified and sequenced numerous transcripts coding for ‘beta-keratins’ and have furthered the understanding of the molecular control of the shedding cycle in lizards and snakes (Dalla Valle et al. 2005, 2007a,b; Alibardi et al. 2007, 2009; Toni et al. 2007). These studies have also helped to explain why in the other reptilian species an evident skin shedding does not take place, namely in the epidermis of crocodilians and turtles (Dalla Valle et al. 2009a,b).

The sequenced transcripts and the deduced 12–20-kDa proteins, classically called beta-keratins, are small proteins that associate with (alpha-)keratins of the intermediate filaments. Therefore we have re-named the proteins ‘keratin-associated beta-proteins’ (KAbetaPs), as they associate to alpha-keratins and contain beta-pleated sheets (Alibardi et al. 2007, 2009). In fact, it appears that the role of these small proteins in the process of cornification in reptilian scales is to produce a hard glue material between alpha-keratins, a role that corresponds to that of mammalian keratin-associated proteins (KAPs) in hairs, claw, horns and nails. These small proteins are produced within fusiform cells (beta-cells) of the epidermis of lizards and snakes, where they form the hard corneous material of the beta-layer. Because the term ‘beta-keratins’ has been used for about 60 years in the literature, in the present study we will use this term as synonymous with KAbetaPs. Recent studies have suggested that the timing of the production of KAbetaPs determines the formation of an oberhautchen and beta-layer, leading to skin shedding.

In mammalian skin derivatives (claws, horns, nails, hooves and hairs), three general types of protein have been distinguished in KAPs: high glycine–tyrosine proteins, high sulphur proteins, and ultra-high sulphur proteins (Gillespie, 1991; Powell & Rogers, 1994; Rogers et al. 2006).

When the induction of the synthesis of KAbetaPs is repressed in the reptilian epidermis, these proteins rapidly decrease in amount and the remaining bundles of alpha-keratin form a softer type of cornification (alpha-layers). The regulatory elements in the gene encoding for KAbetaPs are still unknown, but the turning off and on of the gene(s) eventually determines the differentiation of facing layers of different consistencies (the shedding complex, Maderson, 1985) that favour shedding.

The initial molecular studies conducted on beta-keratins of lepidosaurian reptiles (lizards and snakes) isolated and sequenced proteins of 13–16 kDa, rich in glycine, proline and serine but relatively poor (< 5%) in cysteine (Dalla Valle et al. 2005, 2007a,b). These high glycine–proline–serine proteins are expressed in the normal and regenerating scales of the body, scales that possess a good pliability and stretchability in conjunction with hydrophobic properties. However, the richness in sulphur and cysteine in reptilian scales (Thorpe & Giddings, 1981; Alibardi, 2001; Alibardi & Toni, 2006) also suggests that cysteine-rich proteins are likely present in reptilian scales. This expectation has recently been confirmed by the discovery and sequencing of high cysteine–glycine–proline proteins in digital scale epidermis of the tokay gecko, a lizard (Hallahan et al. 2009) and from the analysis of the genome of Anolis carolinensis (Dalla Valle et al. 2009c). The latter studies have definitely indicated that transcripts for the high cysteine–glycine–proline proteins are present in scales of different toughness in adhesive pad lamellae and claws in lizards. In the present study we have extended our analysis of KAbetaP variation to a representative of the other large group of lepidosaurians, the snakes. We here report on the first case of deduced high cysteine–glycine–proline proteins isolated from mRNAs expressed in normal scale epidermis of a snake, confirming the presence of these proteins in snake scales.

Materials and methods

Microscopic methods and immunocytochemistry

Three juveniles of the American corn snake (Elaphe guttata, recently re-named Pantherophis guttatus, colubrids) were utilized for the histological and molecular analysis of the epidermis (living plus corneous layers; see Dalla Valle et al. 2007b). In one case the epidermis was in resting stage, but in two cases, the epidermis was in renewal phase, as later confirmed by observations on the histology of the three individuals.

Pieces of skin 2 × 4 mm in size were fixed in 4% paraformaldehyde in 0.1 m phosphate buffer pH 7.4 for 5 h, dehydrated in 80% ethanol, and embedded in the hydrophilic resin Bioacryl under UV light at 0–4 °C (Scala et al. 1992). After sectioning with a microtome, 2–4-μm-thick sections were serially collected over microscope slides, and some sections were stained using 0.5% toluidine blue. Most sections were collected over gelatin-coated slides for immunocytochemistry (see below). The beta-1 antibody, produced in rabbit against a chick scale beta-keratin isolated on two-dimensional gel electrophoresis, was generously donated by Dr R. H. Sawyer (Biological Science Department, University of South Carolina, Columbia, SC; see information on the antibody in Sawyer et al. 2000). The antibody has shown positive reactivity on beta-layers in previous studies on snake epidermis (Alibardi, 2002; Alibardi & Sawyer, 2002; Toni & Alibardi, 2007).

Sections were pre-incubated for 30 min in 5% normal goat serum and 2% bovine serum albumin (BSA) in 0.05 m Tris/HCl buffer at pH 7.6 to neutralize non-specific antigenic sites in the sections. The sections were incubated overnight at 4 °C in BSA-Tris buffer containing the primary beta-1 antibody (dilutions 1 : 200). In control sections the primary antibody was omitted, whereas in positive controls both chick and zebrafinch scales were reacted with the beta-1 antibody, producing positive results. After rinsing in the BSA-Tris buffer, the sections were incubated for 1 h at room temperature in the same buffer containing 1 : 70 of anti-rabbit-IgG FITC-conjugated secondary antibodies. After rinsing, the sections were mounted in Fluoromount (EM Sciences, USA), and observed under a Zeiss epifluorescence microscope equipped with a fluorescein filter.

From two animals, 2–4-mm pieces of ventral and lateral skin were collected and fixed in 4% formaldehyde as above. Tissues were dehydrated in ethanol at increasing concentrations, infiltrated in xylene and embedded in wax for the following in situ hybridization (see below).

Tissue preparation and nucleic acid extraction

Tissues were immediately frozen in liquid nitrogen and stored at −80 °C until analysed. Total RNA was extracted using the commercial product Trizol (Invitrogen, Milan, Italy) according to the manufacturer’s instructions. The RNA samples were kept at −80 °C until use. Genomic DNA was extracted from the liver with the Genomic DNA Purification kit (Fermentas, M-Medical, Milan, Italy) according to the manufacturer’s instructions.

Cloning of snake cysteine-rich beta-keratin cDNA

The snake cysteine–glycine–proline-rich beta-keratin cDNA (Sn-cgprp) was cloned using a two-step strategy based on 5′- and 3′-RACE analyses that were performed with the FirstChoice RLM-RACE kit (Ambion, Milan, Italy) following the manufacturer’s instructions and as previously described (Dalla Valle et al. 2007b). The cloning progression and the primers used are described under Results.

cRNA and cDNA probe synthesis

To prepare anti-sense cRNA probes, a recombinant plasmid containing the whole coding region of the Sn-cgprp-1 (obtained with the pair of primers Sn-cys-1 and Sn-cys-4) was linearized by restriction cleavage and used as template. The cRNA transcripts were digoxigenin-labelled by in vitro transcription using a DIG RNA Labeling kit (Roche Diagnostics, Milan, Italy) and Sp6 polymerase.

Northern blotting analysis

Total RNAs extracted from snake epidermis were separated by electrophoresis using a 1.1% formaldehyde-denaturing gel, blotted onto a positively charged nylon membrane (Roche), and baked at 80 °C for 2 h. The 28S and 18S rRNA genes present in the samples were visualized by methylene blue staining to check for RNA loading and integrity. The membranes were hybridized overnight at 68 °C with the DIG-labelled cRNA anti-sense probe in 5 × SSC (saline sodium citrate buffer; 1 × SSC = 0.15 m sodium chloride and 0.015 m sodium citrate) containing 50% formamide, 0.02% SDS, 0.1% lauroylsarcosine, 1% blocking reagent and 100 μg mL−1 of transfer RNA. After incubation with an anti-DIG antibody, the signals were detected using the CPD-Star DIG Luminescent reagent according to the manufacturer’s instructions. The signals were revealed after exposing the membranes to an X-ray film for different time intervals (1–30 min).

Nucleotide sequencing and protein analysis

Sequencing was performed on double-stranded DNA using the abi prism Dye Terminator Cycle Sequencing Core kit (Applied Biosystems, Monza, Italy). Electrophoresis of sequencing reactions was completed on the abi prism model 377, version 2.1.1 automated sequencer. The homology searches were carried out using the basic blast program version 2.0 at, and the alignment was performed using the ClustalW program at For the secondary structure prediction of proteins we used the psipred Protein Structure Prediction Server at (McGuffin et al. 2000). The latter program produces an image of the secondary structure of the proteins, showing the regions forming strands or beta-sheets, those organized as random coils, and those in the alpha-helix conformation (see Results).

In situ hybridization

Tissue sections of 6–8-μm thickness were cut using a microtome, dewaxed with xylene, and hybridized using the complete anti-sense-cRNA probes. A negative control was performed omitting the probes, following a hybridization protocol as previously indicated (Dalla Valle et al. 2005). The hybridization medium contained a mixture of 50% formamide, 4 × SSC, 0.1% Tween-20, 50 μg mL−1 tRNA, 10 mm EDTA, 50 μm mL−1 heparin, and 0.5% Blocking reagent (Roche).

For hybridization we have utilized an overnight incubation at 70 °C in the hybridization buffer containing 0.5–1.5 ng μL−1 of digoxigenin-labelled probe. Hybridization was followed by rinsings at decreasing concentrations of standard saline citrate (2 × SSC, 0.5 SSC, 0.2 SSC, 0.1 SSC, increasing stringency) until mixing with the phosphate-saline-Tween buffer (PBT buffer). Sections were incubated for 2 h at room temperature with anti-digoxigenin-fluorescein Fab fragment antibodies diluted 1 : 30 in Tris Buffer (Roche, Mannheim, Germany). The labelling was detected under fluorescence microscopy. Other sections were incubated with anti-digoxigenin alkaline phosphatase-conjugated antibody (Roche, Mannheim, Germany) diluted 1 : 500 in PBT buffer. Detection was done with PBT buffer containing 4-nitroblue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolyl phosphate (BCIP) as substrates, as indicated by the manufacturer (Roche, Mannheim, Germany). These sections were mounted in permanent medium, and studied under bright field optical microscope.


Microscopic and immunocytochemical observations

Body scales in resting phase comprise an external beta-layer made of hard corneous material and poorly stained with toluidine blue, followed by a darker incomplete alpha-layer, and a two-stratified epithelium (Fig. 1A). Body scales in renewal phase show a more stratified epidermis. The outer generation in our samples shows an outer beta-layer and an incomplete alpha-layer, followed by an irregularly stratified (approximately three to five layers) lacunar layer containing numerous vesicles, which is an immature form of alpha-layer (Fig. 1B,C). A thin clear layer completes the outer generation and is in contact with the differentiating oberhautchen of the next, inner epidermal generation. In fact, beneath the outer layers (generation) a new epidermal generation is forming, which consists of a new oberhautchen layer and three to five layers of spindle-shaped beta-cells that accumulate intensely stained corneous material (‘beta-keratin’). Beneath the forming beta-layer, a stratified series of layers of thin cells indicate differentiating mesos cells that are localized above the two or three stratified basal parts of the epidermis. The oberhautchen layer extends over the scale tip and down the inner surface and the hinge region of the scale, forming a blue-stained line (Fig. 1B,C: this actually represents the spinulated interface with the clear layer). Mesos cells are also seen in the inner scale surface and in this region of the scale the basal cells are flat.

Figure 1.

 Histologic sections of scale epidermis in resting (A) and renewal (B–D) stages. (A) Two overlapped scales are covered by a pale beta-layer (arrows). The epidermis (arrows) has one to two layers. Bar: 20 μm. (B) Scale in renewal phase (stage 4 of the shedding cycle) with the differentiating, inner beta-layer (arrows) within the multilayered epidermis. The arrowhead points to the outer (mature and compact) beta-layer. The double arrowheads indicate the thin oberhautchen layer present in the inner scale surface and hinge region. Bar: 20 μm. (C) Detail of renewing epidermis showing the tapering of the inner beta-layer (arrows) from the outer scale to the tip and down the inner surface. In the latter, only the dark oberhautchen layer is present (arrowheads). Bar: 10 μm. (D) Beta-1 immunolabelled outer (arrowhead) and inner (arrow) beta-layers. The other layers in between the two beta-layers, and the basal layer (above the dashes), are unlabelled. Bar: 10 μm. a, alpha-layer; ba, basal layer of the epidermis; de, dermis; h, hinge region; i, inner scale surface; la, lacunar layer; o, outer scale layer; t; scale tip; The basal layer is underlined with dashes.

The immunocytochemical study of the renewal epidermis, using the beta-1 antibody at the stage indicated above, shows that only the compact and mature outer and the still differentiating inner beta-layers are immunofluorescent (Fig. 1D). The latter observation confirms previous studies (Alibardi, 2002).

Cloning of the snake cysteine-rich beta-keratin cDNAs

To clone the cysteine-rich beta-keratin sequences from the snake Elaphe guttata, we have utilized total RNAs extracted from the skin of a specimen in renewal phase (Snake-2), as confirmed in histological sections from the same specimen.

Initially we used a series of new primers selected for this study on the conserved region of the cysteine-rich beta-keratins of the Gecko gecko (Hallahan et al. 2009). With the Cys-1 primer (Table 1 and Fig. 2) we obtained a 900 nucleotide (nt)-long fragment that, after cloning and sequencing, was found to correspond to the coding region and to the 3′-untranslated terminal region (3′-UTR) of new snake beta-keratins. The 5′-UTR was obtained by 5′-RACE analysis (rapid amplification of cDNA ends) using the anti-sense primer Sn-Cys-1R (Table 1 and Fig. 2). The complete sequence was obtained by the precise overlap of one 3′-RACE and one 5′-RACE fragments (Fig. 2), and this sequence was deposited in EMBL/GenBank with accession number (FN297847) (Sn-cgprp-1).

Table 1.   Primers used in RT-PCR, PCR and 5′- and 3′-RACE analyses
PrimerSequencesPosition*Accession number
  1. *Nucleotide position in the deposited sequence.

Cys-15′-ATGTCCTGCTGTCCTCC-3′+1 → +17FN297847
Sn-cys-1R5′-CACGTGGCAGCAAGGATG-3′+275 → +257FN297847
Sn-cys-1F5′-CCTCCATCTTTCTCAAGATG-3′−17 → +3FN297847
Sn-cys-2R5′-CTCTGGGAGAAGACATGTC-3′+407 → +388FN297847
Sn-cys-2F5′-ATCAACTTCTCTGCTTCTCC-3′−65 →−45FN297847
Figure 2.

Figure 2.

 Nucleotide sequences of cysteine-rich snake beta-keratin transcript (Sn-cgrpr-1), of the genes (Sn-cgrpr-2 and 3), and the deduced amino acid sequences. One-letter symbols of the encoded amino acids are shown below the DNA sequence. The numbers refer to the nucleotide and amino acid positions at the end of each line. Dashed lines indicate missing nucleotides or amino acid loss. Intron sequences are indicated in lowercase. The in-frame translation start codon (ATG) as well as the stop codon (TAA or TAG) and the putative polyadenylation signal (ATTAAA) are boxed. Splice signals (gt and ag) are shown in bold and are boxed. The specific oligonucleotide primers used for the cloning (see Table 1) are underlined and the names are reported in bold over the sequences. The sequences are available at the EMBL/GenBank/DDBJ database under accession number FN297847, FN297848 and FN297849 for Sn-cgprp-1, 2 and 3, respectively.

Figure 2.

Figure 2.

 Nucleotide sequences of cysteine-rich snake beta-keratin transcript (Sn-cgrpr-1), of the genes (Sn-cgrpr-2 and 3), and the deduced amino acid sequences. One-letter symbols of the encoded amino acids are shown below the DNA sequence. The numbers refer to the nucleotide and amino acid positions at the end of each line. Dashed lines indicate missing nucleotides or amino acid loss. Intron sequences are indicated in lowercase. The in-frame translation start codon (ATG) as well as the stop codon (TAA or TAG) and the putative polyadenylation signal (ATTAAA) are boxed. Splice signals (gt and ag) are shown in bold and are boxed. The specific oligonucleotide primers used for the cloning (see Table 1) are underlined and the names are reported in bold over the sequences. The sequences are available at the EMBL/GenBank/DDBJ database under accession number FN297847, FN297848 and FN297849 for Sn-cgprp-1, 2 and 3, respectively.

The deduced amino acid sequence derived from the cDNA coding the Sn-cgprp-1 is based on an open reading frame (ORF) of 387 bp, which starts from a putative initiation methionine localized 65 bp downstream from the 5′-end and continues until a stop codon TAA (Fig. 2). The 3′-UTR is 487 bp long, excluding the poly-A. The ORF of this sequence encodes a putative protein of 128 amino acids with a calculated molecular weight of 12.45 kDa and a pI at 7.95. This sequence presents < 50% of identity with the glycine-rich beta-keratins previously identified in snake (Dalla Valle et al. 2007b).

Northern blotting analysis

The expression of snake cysteine-rich beta-keratins was analysed by Northern blotting using three samples of total RNAs (8 μg) extracted from the skin of two snake samples collected during two different periods of the epidermal cycle. The first sample, indicated as Snake-1 (line 1 in Fig. 3A), was in resting phase. The second sample, indicated as Snake-2 (lines 2 and 3 in Fig. 3A), was in renewal phase. From Snake-2, total RNA was also extracted from the subcutaneous tissue, including muscle, and used as negative control (line 4 in Fig. 3A). More extracted RNA was loaded with the skin samples of Snake-2, as shown by the methylene blue staining of the 28S and 18S rRNA genes (Fig. 3B).

Figure 3.

 Northern blot analysis. (A) Representative Northern blot analysis of total RNA extracted from the skin of Snake-1 (resting), Snake-2 (renewing), and Snake-2 subcutaneous and muscular tissue, detected with a snake DIG-labelled anti-sense cRNA cysteine-rich beta-keratin probe. Exposition time: 5 min. (1)  RNA samples from the skin of Snake-1; (2,3) RNA samples from the skin of Snake-2; (4) RNA samples from the subcutaneous and muscular tissue of Snake-2. (B) Methylene blue-stained 28S and 18S ribosomal RNA genes present in each mRNA sample.

A transcript, estimated to be approximately 1.3 kb, was detected in the skin samples of Snake-2 (Fig. 3A). The size correlates well with the size of the cDNA of the cloned transcript (950 bp), considering an average addition of 200–300 adenines in the mRNAs (Poly-A tail). After 30 min of exposure, no transcript could be detected in the Snake-1 sample (data not shown), nor could it be found using the RNA extracted from subcutaneous and muscle tissue from Snake-2. The latter analysis confirmed the restricted expression of these genes in the epidermis.

Genomic DNA amplifications

To analyse the occurrence of introns inside the cysteine-rich beta-protein gene, we amplified the genomic DNA extracted from the snake liver, using the pair of primers Sn-cys-2F and Sn-cys-2R, which cover the 5′-UTR and the coding region of the gene (Table 1 and Fig. 2).

The obtained amplified fragment was cloned and five of the clones then sequenced. This analysis led to the discovery of two additional and new cysteine-rich beta-protein sequences, Sn-cprp-2 and Sn-cprp-3. The sequences contain introns of 603 and 792 bp, respectively, which are located 22 nucleotides upstream of the ATG codon (Fig. 2). The introns display splice signals consistent with the GT/AG rule. These sequences were deposited in GenBank with the accession numbers FN297848 and FN297849.

In situ hybridization

Using fluorescein-coupled anti-sense RNA probes, a positive immunofluorescence was observed in suprabasal and spindle-shaped cells of the forming beta-layers in the outer scale surface and in the oberhautchen layer of the inner scale surface (Fig. 4A,B). A positive fluorescent signal was also present in suprabasal cells, which are destined to form spindle-shaped keratinocytes (presumptive beta-cells) (Fig. 4B,C). In negative controls, omitting the probes, no immunofluorescence was seen (Fig. 5D). A confirmation of the cellular sites of expression of mRNAs coding for Sn-cprp-1 was obtained by in situ detection using the alkaline phosphatase colorimetric method. The reddish, positive signal was also localized in spindle-shaped cells of the forming beta-layer of the outer scale surface and in the oberhautchen layer of the inner surface (Fig. 4E–G). In contrast to the fluorescent method, the second method, based on the detection using alkaline phosphatase, mainly produced a positive signal in spindle-shaped cells, and little in the underlying, suprabasal cells. The numerous chromatophores beneath the epidermis non-specifically took up some blue stain, and this staining was also seen in the controls. Negative controls did not show any signal in the epidermis, including in the differentiating beta-layer (Fig. 4H).

Figure 4.

In situ hybridization of renewing scale epidermis visualized by immunofluorescence (A–D) and colorimetric reaction (red) with the AP-method (E–H). (A) Positive signal (arrows) in beta-keratin cells of the outer (arrows) and in the oberhautchen (arrowheads) of the inner scale surface. Bar: 20 μm. (B) Detail of the labelled beta-cells (arrows) of the outer scale surface and in the oberhautchen of the inner scale surface of the previous scale (arrowhead). Bar: 10 μm. (C) Detail of labelled beta-cells of the stratified inner beta-layer of the outer scale surface (arrow). Bar: 10 μm. (D) Control showing absence of fluorescence in both the beta-layer of the outer (arrow) and the inner (arrowheads) scale surface. Bar: 10 μm. (E) Reactive beta-layer (arrows) and oberhautchen layer of the inner scale surface (arrowhead). Bar: 30 μm. (F) Detail of the reactive cells (arrows) forming a single layer (oberhautchen) in the hinge region (left), which becomes a tetra-stratified beta-layer in the outer scale surface (right). Bar: 15 μm. (G) Reactive oberhautchen (arrowhead) and penta-stratified beta-layer (arrow) located in the outer scale surface. Bar: 15 μm. (H) Control sections showing no reactivity in beta-cells (arrows). The arrowhead indicates the layer of melanocytes located underneath the epidermis. Bar: 15 μm. a, alpha-layer; aR, anti-sense RNA probe detection; c, negative control; de, dermis; the epidermis is underline with dashes.

Figure 5.

 ClustalW comparison of the amino acid sequences among the cysteine-rich protein sequences (Sn-cgprp-1, 2, 3) and the glycine-rich protein sequences (Sn-gprp-1 to 5, Dalla Valle et al. 2007b). The extended 32 amino acids appear in the centre of the figure, including the core-box region, which shows the highest identity among these proteins. Four key amino acids (glycine in red, proline in blue, serine in yellow, and cysteine in green) are indicated. According to the ClustalW convention, an asterisk denotes the identity in all sequences of the alignment; colons denote conserved substitutions; dots denote semi-conserved substitutions.

Analysis of the deduced amino acid sequence of keratin-associated beta-proteins

The newly identified cysteine–proline-rich beta-proteins (Sn-cgpr-1,2,3) show a different amino acid composition in comparison with the previously sequenced proteins found in this snake, and indicated as glycine–proline-rich beta-proteins (e.g. Sn-gprp-1 etc., see Dalla Valle et al. 2007b; Table 2). Among other amino acids the content of cysteine is almost doubled in cysteine–proline-rich proteins, whereas the content of glycine is about 4% less in comparison with glycine-rich proteins, which are generally of higher molecular weight. The more predominant amino acids in cystine-rich proteins are glycine (14.8–15.6%), proline (13.9–14%), serine (13.3–14.8%), alanine (8.2–9.1%), and cysteine (10.2–11.5%) (Table 2). All deduced proteins present a net positive charge with a basic pI.

Table 2.   Amino acid composition [Number (N) and percentage (%) of amino acid residues] of the cysteine-rich snake beta-keratin deduced proteins. Sn-gprp-1 refers to a glycine-rich snake beta-keratin deduced protein isolated in a previous work (Dalla Valle et al. 2007b)
Theoretical pI7.497.957.927.95
Ala (A)1510.9118.6108.2119.1
Arg (R)42.943.132.532.5
Asn (N)
Asp (D)
Cys (C)85.81310.21411.51310.7
Gln (Q)42.921.621.621.7
Glu (E)32.221.621.621.7
Gly (G)2518.22015.61814.81814.9
His (H)00.010.810.800.0
Ile (I)
Leu (L)139.575.575.775.8
Lys (K)
Met (M)10.710.810.810.8
Phe (F)42.910.821.621.7
Pro (P)1712.41713.31713.91714.0
Ser (S)1510.91713.31814.81714.0
Thr (T)64.464.754.165.0
Trp (W)
Tyr (Y)21.553.932.532.5
Val (V)107.386.2129.8108.3
Asp + Glu3222
Arg + Lys4444

The comparison of the amino acid sequences of snake cysteine-rich beta-keratins with the glycine-rich beta-keratins from the same species of snake shows that identities of 38–40% are present over the whole protein among these two sub-families. In contrast, the three cysteine-rich beta-keratins present identity values of 88–90% among themselves (Fig. 5). In the 32-amino acid region where beta-sheets are present (Fraser & Parry, 2008), here referred to as the extended core-box region, the two sub-families of snake beta-keratins, cysteine-rich and glycine-rich, present the higher identity (range 50–59%, see details in Table 3). Table 3 is based on the values obtained using ClustalW software at

Table 3.   Evaluation of the percentage of identity of the 32-amino acid sequence in the extended core-box region in all known beta-keratins of Elaphe guttata. Sn-gprp-1 to 5: extended core-box of snake glycine-rich beta-keratins (accession numbers CAL49457, CAL49458, CAL49459, CAL49460, CAL51276); Sn-cgprp-1 to 3: extended core-box of cysteine-rich beta-keratins (accession numbers FN297847, FN297848, FN297849)
Sn-gprp-2 10010093595056
Sn-gprp-3  10093595056
Sn-gprp-4   93595056
Sn-gprp-5    595056
Sn-cgprp-1     8487
Sn-cgprp-2      78

The prediction of the secondary structure for these proteins was done using the psipred Protein Structure Prediction Server at (McGuffin et al. 2000; Fig. 6). As previously indicated, this program gives a visual representation of the predicted secondary conformation (random coil, strand and alpha-helix) of different regions of the proteins. The protein prediction shows that these proteins contain a small region (boxed in Fig. 6) with an extended strand structure, indicating the presence of a prevalent beta-sheet conformation in this region (Fraser & Parry, 2008). Using the prediction program, the central region of 32 amino acids with high identity (Fig. 5) shows the presence of at least two strands regions (beta-pleated sheets), as in other beta-keratins previously sequenced in snakes and in other reptiles (see Discussion). At least one other strand is present downstream of the extended core-box region, further indicating that the beta-sheet conformation is prevalent in the central part of these proteins. Outside the central part of the proteins, a random-coiled conformation is prevalent, whereas alpha-helix regions are almost absent.

Figure 6.

 Computer prediction of the secondary structure (using the psipred Protein Structure Prediction Server at of two Sn-cprp proteins. The central core-box is indicated by squares. The confidence of prediction (%) is indicated.


Genomic consideration and expression

The present study has for the first time demonstrated the presence of a transcript and two genomic sequences encoding proteins of 11.6, 11,8 and 12.5 kDa, which are within the typical molecular weight range for beta-keratins. The new proteins have a relatively high cysteine content (over 10%). These new types of scale proteins are here referred to as high cysteine–glycine–proline proteins, and they represent the first members of this type ever identified in a species of snake. The cloning analysis performed in our study has also shown the expression of more beta-keratins rich in cysteine present in snake epidermis, which characteristics are not shown in the present report. In fact, in the present study we have only shown the amino acid sequences of the proteins on which we have determined the entire coding region. Using genomic DNA, two cysteine-rich beta-protein sequences, complete in the 5′-UTR and in the coding regions, were obtained. These data indicate that, as in lizards, more members of cysteine-rich beta-proteins are present in snake scales, perhaps mainly expressed in the scales containing reinforcing structures such as heels (scales were unavailable in our samples). In fact, cysteine-rich beta-proteins in lizards appear typical of the more cornified scales and claws (Inglis et al. 1987; Alibardi & Toni, 2009).

The deduced proteins in snake appear to be expressed in suprabasal keratinocytes (presumptive beta-cells) as well as in differentiating keratinocytes (fusiform) of the forming beta-layer. Northern analysis only showed the presence of transcripts coding for these proteins with RNA extracted from skin during the renewal phase (Snake-2). The size of the transcripts correlates well with the cDNA size of the transcript cloned in this study, and also with the size of the glycine-rich beta-proteins of the shorter transcripts displayed with the Northern analysis in our previous work (Dalla Valle et al. 2007b). However, the possibility of cross-reactivity seems unlikely as the two types of proteins present < l50% of identity with each other at the nucleotide level. Moreover, the new probe does not recognize the longer transcripts displayed with the previous Northern analysis (Dalla Valle et al. 2007b).

The genes for these proteins, like other beta-keratin genes sequenced in other reptilian and avian species, contain an intron located in 5′-UTR (the lower case nucleotide sequence shown in Fig. 2). Although the length of the two introns is different, stretches with a high identity are present in the 5′- and 3′-regions of the introns. Moreover, the alignment of the intron found in the glycine-rich beta-keratin gene (Dalla Valle et al. 2007b) with the two intron sequences of the new cysteine-rich beta-keratin genes have shown that no conserved regions are present between these sequences (data not shown).

The slight difference observed in the in situ hybridization experiments (fluorescence vs. alkaline phosphatase detection) may be due to a higher sensitivity of the fluorescence method over the colorimetric method based on the alkaline phosphatase reaction. If this is the case, the fluorescence method indicates that the initial synthesis of mRNAs coding for the high cysteine–glycine proteins already occurs in presumptive beta-cells of the suprabasal layer, before the cytological signs for their differentiation are manifested (fusiform shape and keratin accumulation). This early expression was not observed in previous studies that detect the expression of mRNAs encoding for high glycine beta-proteins (Dalla Valle et al. 2007b). The expression of mRNAs for high cysteine–glycine proteins is, however, less marked in spindle-shaped (beta-)cells than the expression of high glycine protein mRNA, perhaps indicating that after their initial production the levels of these transcripts are reduced in the beta-layer of the normal renewing scale epidermis. This pattern suggests that gene products (Cys-rich beta-keratin mRNAs) are already present in suprabasal cells before they accumulate large amounts of the proteins. The observation may indicate that in the scales analysed in the present study, high cysteine–glycine beta-proteins are initially produced in beta-cells but that later their level decreases or is overcome by the production of the high glycine beta-proteins. The latter family of beta-keratins probably produces most of the proteins of the relatively softer beta-layer of the scales analysed in this snake. Perhaps in some scales present in other body areas there is a more marked expression of high cysteine–glycine beta-proteins, but further studies are needed to determine this. The specific sites of expression of the different beta-proteins in snake (and lizards) await further in situ hybridization and immunocytochemical analysis.

Proteomic considerations

The molecular weight and pI of the deduced proteins are within the range previously determined for extracted hard (beta-)keratins in snake epidermis, 10–15 kDa (Alibardi & Toni, 2006; Toni & Alibardi, 2007; Toni et al. 2007), confirming the validity of the present results. The amount of glycine in snake beta-keratins is generally lower than in the proteins of other reptiles (17–19% in glycine-rich beta-keratins and 14–15% in cysteine-rich beta-keratins compared with 25–35% found in other species). This is mainly due to the disappearance of glycine-rich regions in the head and tail regions of the proteins in comparison with the presence of glycine-rich regions in the proteins in other reptiles.

Despite the differences in the amino acid composition and sequence between high glycine–proline proteins and high cysteine–glycine proteins, both these protein families present a central region of 32 amino acids (the extended core-box) with a high identity (78–87%, see Table 3 and Fig. 5). The prediction of the secondary structure for these proteins using the psiprep program has shown the presence of at least two beta-strands in this region. Another analysis of this central region has indicated the presence of four beta-strands in this region, believed to be involved in the formation of the framework of the filaments of beta-keratin (Fraser et al. 1972; Fraser & Parry, 1996, 2008). The beta-keratin filament is the polymer obtained through the polymerization of the single proteins, high glycine beta-proteins or high cysteine–glycine beta-proteins, which function as monomers (Fraser et al. 1972; Brush, 1978).

Together with the other class of KAbetaPs (beta-keratins) rich in glycine previously found in snake, the new type of cysteine-rich beta-proteins may help reinforce the corneous layers of scales, due to the potential capability of these proteins to form numerous disulphide bonds. Although our analysis is limited to some examples of body scales, it is very likely that other protein members of the high cysteine–glycine beta-protein family are also present in snake epidermis. This was suggested from the previous isolation and amino acid determination of numerous cysteine-rich beta-proteins from lizards (Dalla Valle et al. 2009c; Hallahan et al. 2009). It is also possible that the newly discovered proteins may be particularly expressed in harder scales of snakes, such as the gastrosteges of the ventral side of the snake body, or in more cornified scales of other regions. However, our study is not definitive on this point, as these large scales were not available in our study.

It is known that, despite their surface extension, all snake scales have the same epidermal stratification (dorsal, lateral and ventral, the large gastrostege scales). Previous studies (Wyld & Brush, 1979, 1983) have indicated that the mechanical performance of scales in different body regions, or in claws and ramphoteca, derives from the unique proportion of specific beta-keratin monomeres (corresponding to our KAbetaPs) present in the different skin appendages. Although high glycine beta-proteins present in normal snake scales seem more abundant than high cysteine–glycine beta-proteins (Dalla Valle et al. 2007b), it is probable that the latter are major components of the harder scales.

HCGP proteins are prevalent in the adhesive climbing pad lamellae of the gecko (a specialized type of digit scales, see Hallahan et al. 2009) and in lizard claws (Inglis et al. 1987; Alibardi & Toni, 2009). Claws in particular are made of a harder corneous material than normal scales, and this characteristic may also result from the high expression of cysteine-rich beta-proteins in claw keratinocytes.

Analysis of the distribution of cysteine within these proteins (Fig. 5) shows that most of this amino acid is present toward their N- and C-extremities, which are useful sites for the formation of inter-molecular disulphide bonds.


The present study was supported by 60% funding from the University of Padova and in part was self-supported (L.A.).