Scale keratin in lizard epidermis reveals amino acid regions homologous with avian and mammalian epidermal proteins



Small proteins termed β-keratins constitute the hard corneous material of reptilian scales. In order to study the cell site of synthesis of β-keratin, an antiserum against a lizard β-keratin of 15–16 kDa has been produced. The antiserum recognizes β-cells of lizard epidermis and labels β-keratin filaments using immunocytochemistry and immunoblotting. In situ hybridization using a cDNA-probe for a lizard β-keratin mRNA labels β-cells of the regenerating and embryonic epidermis of lizard. The mRNA is localized free in the cytoplasm or is associated with keratin filaments of β-cells. The immunolabeling and in situ labeling suggest that synthesis and accumulation of β-keratin are closely associated. Nuclear localization of the cDNA probe suggests that the primary transcript is similar to the cytoplasmic mRNA coding for the protein. The latter comprises a glycine-proline-rich protein of 15.5 kDa that contains 163 amino acids, in which the central amino acid region is similar to that of chick claw/feather while the head and tail regions resemble glycine-tyrosine-rich proteins of mammalian hairs. This is also confirmed by phylogenetic analysis comparing reptilian glycine-rich proteins with cytokeratins, hair keratin-associated proteins, and claw/feather keratins. It is suggested that different small glycine-rich proteins evolved from progenitor proteins present in basic (reptilian) amniotes. The evolution of these proteins originated glycine-rich proteins in scales, claws, beak of reptiles and birds, and in feathers. Some evidence suggests that at least some proteins contained within β-keratin filaments are rich in glycine and resemble some keratin-associated proteins present in mammalian corneous derivatives. It is suggested that glycine-rich proteins with the chemical composition, immunological characteristics, and molecular weight of β-keratins may represent the reptilian counterpart of keratin-associated proteins present in hairs, nails, hooves, and horns of mammals. These small proteins produce a hard type of corneous material due to their dense packing among cytokeratin filaments. Anat Rec Part A, 288A:734–752, 2006. © 2006 Wiley-Liss, Inc.

Reptilian epidermis produces two major groups of keratins, indicated as α- and β (ϕ)-keratins. These proteins have different solubility, molecular weight, amino acid composition, secondary conformation, degree of elasticity properties, packing modality, X-ray diffraction pattern, and biosynthesis (Baden and Maderson,1970; Fraser et al.,1972; Wyld and Brush,1979,1983; Maderson,1985; Landmann,1986; Fraser and Parry,1996; Sawyer et al.,2000; Alibardi and Sawyer,2002; Alibardi et al.,2004a).

α-keratins represent a scaffold for the formation of the soft and pliable stratum corneum, termed α-layer. The latter allows scale mobility and houses the barrier against water loss. Instead, β-keratins form a hard and inflexible layer that protects scales from mechanical stress.

As opposed to α-keratins, β-keratins of birds and reptiles are not extensible and are small proteins with a β-pleated secondary conformation. These proteins aggregate into a densely packed lattice that produces very resistant microfibrils with a typical X-ray and ultrastructural pattern of 3–4 nm electron pale filaments in a denser matrix. The mechanism of aggregation of β-keratins to form the very resistant cables is poorly known (Gregg and Rogers,1986; Fraser and Parry,1996). In reptilian epidermis, the β-layer mainly determines mechanical resistance but is also the depositary of most epidermal pigmentation: both resilience and pattern ornamentation are the first positive characteristics exploited in the leather industry using reptilian skin. The above studies have shown that β-keratins in reptiles, as those of birds, have a smaller molecular mass (10–22 kDa) than α-keratins (40–68 kDa), but very little is known about their amino acid sequences and spatial conformation.

β-keratin is believed to be phylogenetically more recent than α-keratin but no information is presently available on the evolution of β-keratin from the likely more primitive α-keratin (Sawyer and Knapp,2003). Previous studies have shown that in lizard epidermis, acidic and basic cytokeratins similar to those of mammalian epidermis are present (Fuchs and Marchuk,1983; Alibardi et al.,2000,2001; Alibardi and Toni,2006a).

α- (soft) keratins are composed of 8–10 nm thick keratin filaments (intermediate filaments) and in epidermal cells are referred to as cytokeratins. The filaments require the association of one acidic and one basic keratin chain to form a dimeric coil-coiled molecule that has a central α-helical rod and two variable N- and C-heads (Fuchs and Marchuk,1983; Klinge et al.,1987; O'Guin et al.,1987; Steinert and Freedberg,1991; Fuchs and Weber,1994; Coulombe and Omary,2002).

In the mammalian epidermis, the keratin pair is K5/K14 in basal and suprabasal layers and K1/K10 pair in upper spinosus and cornifying layers. β-keratin filaments are thinner (3–4 nm) than cytokeratin filaments.

The epidermis of lepidosaurian reptiles (snakes and lizards) is produced from cyclically alternating epidermal generations (shedding cycle) (Maderson,1985; Landmann,1986; Maderson et al.,1998). β-keratin is produced in specific layers, termed oberhautchen and β-layer, but is replaced by α-keratin in the following layers (mesos, α-, lacunar, and clear layer). These keratins are synthesized during the renewal phase of the shedding cycle of lizard epidermis, which terminates with the molt of the external (outer) epidermal generation. After shedding, the epidermis enters in a resting phase in which proliferation is halted almost completely: this phase generally occupies most of the time of the shedding cycle, until the next renewal phase.

The renewal phase of the shedding cycle in the epidermis of lizards can be induced after tail amputation, a process that stimulates the regeneration of the tail. The regenerating tail epidermis forms a proximodistal sequence of scale neogenesis that reproduces the renewal phase of both embryonic and normal epidermis (Maderson et al.,1998; Alibardi,2000,2001,2003). Thus, the differentiation of a new β-layer can be induced for studying the expression of β-keratins. The differentiation of the first β-layer can also be studied in embryonic scales in late embryonic stages of lizards (stages 38–39) (Alibardi,1996).

In a previous study (Dalla Valle et al.,2005), we have selected and sequenced by RT-PCR/rapid amplification of cDNA ends (RACE) analysis a lizard scale keratin cDNA and deduced the amino acid sequence (Genebank accession number AJ890445).

In the present study, we have utilized the β-keratin cDNA probe, generated in the above study, to analyze further the expression of the mRNAs by light microscopy and ultrastructural in situ hybridization in both regenerating and embryonic epidermis. This was done in conjunction with an immunolocalization study with an antiserum directed against a 15–16 kDa protein (β-keratin) isolated from lizard epidermis.

Moreover, a comparative analysis of the amino acid sequence of the lizard protein with mammalian and avian keratins or with keratin-associated proteins (KAPs) provided some clues on the molecular evolution of hard proteins in amniote skin appendages such as scales, hairs, and feathers.


Experimental Procedures

Tail amputation was made as previously indicated (Alibardi,2001) by inducing tail autotomy, a natural process of tail loss obtained by pulling and twisting the two-third proximal of the tail in adult males and females of the lizard Podarcis sicula and P. muralis. Tail regeneration took place at a temperature ranging from 24 to 28°C. Six specimens of P. sicula and five specimens of P. muralis with regenerating tails of 3–5 weeks postamputation were used. The specimens of P. muralis were studied only for the histological study on the process of tail regeneration. After this period, new scales were forming in proximal regions near the tail stump of the 0.8–2.0 cm long regenerating tail. The regenerating tail was collected by induced autotomy in the proximal tail stump and skin was sectioned off and immediately fixed. Eight specimens of P. sicula were used for the microscopic study.

Nine lizard embryos of P. sicula at embryonic stage 37 (n = 3), 38 (n = 3), and 39 (n = 3) were collected and fixed for the histological study. In the last two stages, a β-layer is differentiating in embryonic scales (Alibardi,1996). Sections from embryos at stage 39 were selected for the in situ hybridization study for β-keratin messenger expression and for immunocytochemistry.

β-Keratin Isolation, Antibody Production, and Immunoblotting

Normal and regenerating skin from the lizard Podarcis sicula were collected, immediately frozen in liquid nitrogen, and stored at −80°C. The skin was incubated in 5 mM EDTA in phosphate-buffered saline (PBS) for 3–5 min at 50°C and 2–4 min in cold PBS. The epidermis was separated from the dermis by dissection under the stereomicroscope and the dermis discarded. Then, the epidermis was homogenized in lysis buffer (8 M urea, 50 mM Tris-HCl, pH 7.6, 0.1 M 2-mercaptoethanol, 1 mM dithiothreithol, 1 mM phenylmethylsulphonyl fluoride) to extract keratins according to the method of Sybert et al. (1985). The protein concentration was determined by the Lowry methods.

For antiserum production, lizard epidermis proteins were extracted (as described above) and separated by SDS-PAGE. Following electrophoresis, the gel was fixed and stained with Coomassie blue and two close protein bands with a molecular weight of 15–16 kDa, corresponding to a β-keratin (as confirmed by immunocytochemistry), were detected. The bands were sliced off with a razor blade and directly emulsified in 2 ml of PBS-CFA (complete Freund's adjuvant; Sigma) solution (1:1 vol/vol). The adjuvant/antigen solution containing approximately 150 μg of the selected proteins was injected subcutaneously in a rat that was previously bled to collect blood sample (preserum). One booster immunization, prepared as above, was performed 2.5 week after the first injection. The rat was sacrificed after about 3 weeks after the boost injection and blood was collected. Serum was then separated by centrifugation from clot and debris and stored at −20°C. Following the above protocol, two different antisera (A612 and A68B) against lizard β-keratin were produced.

For Western blot analysis, epidermis proteins separated by SDS-PAGE (50 μg per lane) were transferred to nitrocellulose paper (Hybond C Extra; Amersham, U.K.) and incubated with rat antisera (A612 or A68B; dilution 1:25–100) as primary antibodies. The filters were then washed and further incubated for 1 hr at room temperature with antirat HRP-conjugated secondary antibody (dilution 1:1,000) in TBS-Tween + 5% nonfat milk powder. Detection was performed by using the enhanced chemiluminescence's procedure developed by Amersham (ECL).

Sequence Analysis

The amino acid sequences of lizard scale (Dalla Valle et al.,2005) and lizard claw (Inglis et al.,1987) were compared with some avian and mammalian keratin-like proteins. These proteins represented a keratin-associated protein, an α-keratin, a scale β-keratin, and a feather β-keratin. Furthermore, a broader phylogenetic analysis was also performed to correlate the amino acid sequences of β-keratins of scales, feathers, and claws, the sequences of cytokeratins of different vertebrates, and the sequences of keratin-associated proteins deposited in the NCBI protein database (protein accession numbers are reported in Table 1). These sequences were aligned using ClustalX (1.81) software (Thompson et al.,1997). The phylogenetic tree was constructed by the neighbor-joining method (Saitou and Nei,1987) and was visualized using the NJPLOT software. ClustalX (1.81) and NJPLOT software are available for free download (

Table 1. List of compared sequences with their NCBI accession number
ProteinAccession numberProteinAccession number
Beta[Varanus](Inglis et al.,1987)KAP 5-1[Homo]Q6L8H4
Beta-scale[Podarcis](Dalla Valle et al.,2005)KAP6-1[Oryctolgus]423336
hypothetical protein [Dictyostelium]XP_647551K-D[Gallus]CAA35555
K-10[Lampetra]CAC87096K-g3 [Lampetra]CAC87092
K-13b[Homo]NP_002265K-gamma2 [Lampetra]CAC87101
K-8[Lampetra]CAC87094Lfl-K 1 [Lampetra]CAC24702


Two to 5 mm long skin pieces collected from three regenerating tails and from three embryos at stage 39 were fixed for 4–5 hr in 4% paraformaldehyde in 0.1 M phosphate buffer at pH 7.4. Tissues were dehydrated and embedded in the hydrophilic resin bioacryl (Scala et al.,1992). Thick (2–4 μm) and thin (40–70 nm) sections were collected respectively for light and ultrastructural cytochemistry. For histological study, sections were stained in 0.5% toluidine blue or by hematoxylin and eosin stain.

For immunocytochemistry, the above-described rat immune serum against a lizard β-keratin of 15–16 kDa was utilized on sections of lizard scales and embryos.

Nonspecific binding components were blocked with 5% normal goat serum and 1% BSA in PBS containing 0.3% Triton X-100 for 30 min at room temperature. The primary antiserum (1:50–100 in 0.5 M Tris buffer at pH 7.6 containing 2% BSA) was overlaid on each section for 14–16 hr at 4°C. Sections were subsequently incubated with a goat antirat FITC-conjugated secondary antibody diluted 1:50 (Sigma) for 30 min at room temperature. Immunoreactivity was detected under a UV epifluorescence microscope. The specificity of the immunolabeling was tested by omitting the primary antibody.

Thin sections of skin (40–90 nm thick) were collected on nickel grids using an ultramicrotome. Sections were incubated overnight with the primary rat antibody as for light microscopy immunocytochemistry, rinsed in buffer, and incubated for 1 hr at room temperature with goat antirat IgG conjugated with 10 nm gold particles (Sigma; diluted 1: 50). Detection was done using a secondary antirat antibody conjugated with 10 nm gold particles (Sigma; diluted 1: 50). After rinsing, sections were stained for 7 min with 2% aqueous uranyl acetate, and grids were observed under the electron microscope.

In Situ Hybridization

Tails were collected from six lizards with regenerating tail showing scale neogenesis (3–5 weeks) and fixed in 4% paraformaldehyde as above for immunocytochemistry. The other three embryos (stage 39) were fixed in 4% paraformaldehyde. Tissues were dehydrated in ethanol at increasing concentration, infiltrated in xylene, and embedded in wax.

Sections of 6–8 μm thicknesses were collected using a rotary microtome, dewaxed with xylene, and hybridized using incomplete digoxigenin-labeled antisense probes (corresponding to the last part of the coding region and the 3′-UTR sequence of lizard keratin, nucleotides +382/+835), or complete digoxigenin-labeled antisense cDNA probes (corresponding to the complete coding region). The plasmids containing these fragments [prepared in previous work; see details in Dalla Valle et al. (2005)] were amplified by PCR with the M13 forward and reverse primers. The two fragments obtained were then used to produce, by asymmetric PCR with digoxigenin-labeled nucleotides (DIG, Roche), digoxigenin-labeled sense and digoxigenin-labeled antisense incomplete and complete cDNA probes.

For the hybridization, the probe concentration was 0.5–1.5 ng/μl of digoxigenin-labeled antisense probe in hybridization solution (50% formammide, 4 × SSC, 0.1% Tween-20, 50 μg/ml tRNA, 10 mM EDTA, 50 μm/ml heparin, 0.5% blocking reagent; Roche, Applied Science, Milan, Italy). Hybridization was carried out overnight at 40–42°C. Posthybridization treatment included a washing procedure with 2 × SSC (3 × 10 min, at 50°C), 0.5 × SSC (1 × 15 min, at 60°C), 0.2 × SSC (1 × 15 min, at 60°C), and 0.1 × SSC (1 × 15 min, at 60°C). Afterward, the incubation with the anti-DIG-fluorescein Fab fragment antibodies (Roche) diluted 1:15 in Tris buffer was performed for 2 hr. The labeling was detected under fluorescence microscopy. Other sections were incubated with anti-DIG alkaline phosphatase-conjugated antibody (Roche) diluted 1:500 in PBS and color reactions were started with NBT/BCIP substrate.

For ultrastructural in situ hybridization, 40–80 nm thick sections of regenerating scales containing differentiating β-cells were collected on nickel grids and incubated with the same hybridization solution as described above for 4–5 hr. Rinsing steps were the same as described above. Afterward, the sections were incubated at room temperature for 20 min in Tris 0.05 M, pH 7.6, with 2% bovine serum albumin and for 3 hr with the mouse anti-DIG antibody (Roche) diluted at 1:100. After repeated rinsing in the buffer, grids were reacted for 1 hr at room temperature with the secondary antimouse antibody (conjugated with gold particles of 5 nm in diameter). After rinsing, grids were stained for 5–7 min in 2% uranyl acetate, dried, and observed under a CM-100 Philips electron microscope. In situ hybridization controls were performed on sections consecutive to positive reactions and included omission of the probes and hybridization with a labeled sense probe.



The injection of epidermal protein bands at 15–16 kDa of P. sicula epidermis (Fig. 1, lane A) into rats produced immune sera (Fig. 1, lanes B and C for serum A612, and D and E for serum A68B). Immune sera generally reacted with lizard epidermal proteins, producing a specific immunolabeled band at 15–16 kDa, sometimes with a weak band at 30–32 kDa. These bands were absent in presera. In negative controls, omitting the primary antiserum, no immunolabeled bands appeared.

Figure 1.

Gel electrophoretic separation of epidermal proteins in P. sicula with the chosen bands (square) used for immunization (lane A). The preimmune sera (lanes B and D) show no labeled bands in the 14–16 kDa range, which instead appear in their respective immune sera (A612 serum, lane C; A68B serum, lane E).


The distribution of both the antiserum and the probe for the β-keratin mRNA was studied in lizard epidermis. No difference in the histology of regenerating tail was observed between the two species used (P. muralis and P. sicula). Normal epidermis consisted of a corneous layer where an outer β-layer was followed by the α-layer and the remaining living epidermal layers (Fig. 2A). After 2–3 weeks postamputation of the tail in both species, a wound epidermis made of spinosus-like keratinocytes covered the regenerating tail (data not shown). In proximal regions, near the skin of the tail stump, the wound epidermis formed pegs penetrating the dermis (Fig. 2B). The pegs progressively elongated in distal-proximal direction, and in the following 1–-2 weeks their axial part started to differentiate fusiform cells of the β-layer (Fig. 2C and D). The following maturation and compaction of β-cells allowed the new formation of a β-layer (Fig. 2E). The detachment from the upper cornified wound epidermis occurred by the differentiation of a shedding complex, comprising a granulated layer (termed clear layer, with down-pointing serration) and an oberhautchen layer (with up-pointing serration). Therefore, beneath the cornified wound epidermis, the sequence of layers was clear layer with its granules, serrated oberhautchen, and β-layer. The α-keratogenic cells differentiated underneath the β-layer to reconstitute later a new α-keratin layer (data not shown).

Figure 2.

Light microscopic view of normal (A), regenerating (BE), and embryonic (FH) epidermis of lizard. A: Beneath the α-layer, precorneous cells (arrows) are differentiating. Scale bar = 10 μm. B: Epidermal peg penetrating the dermis in regenerating tail at 3 weeks postamputation. Beneath the lacunar epidermis, a granulated (clear) layer is forming (arrows), followed by early differentiating β-cells above a stratified epidermis. Scale bar = 10 μm. C: Cross-sections of proximal regenerating scale at 3 weeks postamputation showing the formation of large granules in the clear layer (arrows) with the thick differentiating β-layer. Scale bar = 10 μm. D: Detail of epidermal differentiation in elongating pegs at 4 weeks of tail regeneration. Beneath the granulated layer (arrows, the future shedding line), 3–4 layers of fusiform β-cells are forming the β-layer. Scale bar = 10 μm. E: Regenerating scales of 4 weeks postamputation. The figure shows the passage from a cellularized β-layer near the tip of the regenerating scale to the dark compact β-layer. Cells of the granulated layer (arrow) also become dense and form the compact clear layer (arrowhead). The latter produces denticles that interdigitate with the specular oberhautchen serration. Scale bar = 20 μm. F: Symmetric ventral scale at stage 37. Scale bar = 10 μm. G: Tail scale at stage 39 with a 4–5 dark layers of flat β-cells in the outer surface (arrow) beneath the embryonic epidermis (arrowhead). The dark layer remains monocellular in the inner surface and hinge region. Scale bar = 20 μm. H: Part of mature dorsal scale at stage 39 showing the packed β-layer (arrow) and dense detaching embryonic epidermis (arrowhead). Scale bar = 20 μm. β-, β-layer; βC, compacted β-layer; D, dermis; E, living epidermis; H, hinge region; I, inner scale surface; L, lacunar layer; W, wound epidermis.

In the embryo, at stages 37 in some areas of the body (e.g., midtrunk lateral and limbs), the formation of bumps represented the symmetric stage of scale development (Fig. 2F). Bumps in other regions of the body (e.g., lumbar and tail) at stage 37 and in the remaining body areas at stage 38 became asymmetric and rapidly formed the outer scale surface (Fig. 2G). At this stage, the epidermis was made of 3–4 layers of cells above the germinal layer.

During elongation of the outer scale surface, fusiform β-cells started to accumulate beneath the embryonic layers and the epidermis became thicker (Fig. 2G). In the inner side and hinge region of the developing scales, no fusiform β-cells were present and the epidermis remained thin. The three more external, very flat layers formed the outer and inner periderm and the third layer represented the embryonic oberhautchen (data not shown). Beneath the latter, fusiform and dark cells started to form the embryonic β-layer that was very reduced in the inner scale surface and hinge regions (Fig. 2G). In the outer scale surface, the β-layer was formed by 3–5 layers of fusiform β-cells and eventually formed a compact β-layer at stage 39 (Fig. 2H). In some scales (e.g., in the proximal tail and ventral scale), both the embryonic epidermis and the β-layer packed into two dense layers, and the embryonic epidermis started to detach from the β-layer (Fig. 2H). However, scales in different parts of the body were at slightly different stages, and the β-layer was still immature (unpacked).

Light Microscopy Immunocytochemistry and In Situ Hybridization

The β-layer of normal, regenerating, and embryonic scales was immunofluorescent using our rat antiserum A68B against the 15–16 kDa lizard β-keratin (Fig. 3A–C). In normal epidermis, the β-layer corresponded to the external compact corneous layer (Fig. 3A). In the regenerating epidermis, only the forming β-layer (Fig. 3B) and the mature compact β-layer of regenerated (neogenic) scales were immunofluorescent (Fig. 3C). Negative controls were not immunolabeled (Fig. 3D).

Figure 3.

Immunofluorescence of lizard scales after application of the lizard β-antiserum (A68B) in normal scale (A), regenerating scales (B and C), and negative control (D). Immunostaining with the lizard keratin antiserum on embryonic scales (E and F). In situ hybridization immunofluorescence with antisense DNA incomplete probe (aDi) on regenerating scales (G and H), sense DNA incomplete (sDi) probe (I), and negative control (co) probe (J). H, hinge region; T, scale tip. Dashes underline the basal layer; arrows indicate the more or less compact β-layer. Scale bars = 10 μm.

In the embryonic epidermis at stage 39, the β-layer was also immunostained and the labeling tended to disappear in the hinge region (Fig. 3E and F). Other layers of the epidermis or other tissues of the skin were immunonegative. Negative controls were not immunolabeled (data not shown).

The in situ hybridization study by immunofluorescence was done using the antisense DNA incomplete probe (aDi; digoxigenin-labeled corresponding to the last part of the coding region and the 3′-UTR sequence of lizard keratin; nucleotides +382/+835). The antisense DNA complete probe (aDc; digoxigenin-labeled whole 485 nucleotidic coding region probe; see sequence in Fig. 9A) was also used. The probes were added to sections of regenerating tail (4 weeks postamputation) at a concentration of 0.5–1.5 ng/μl (see details in Dalla Valle et al.,2005), and the formed hybrid produced a specific immunofluoresecence in differentiating fusiform β-cells present in the regenerating outer scale surface (Fig. 3G and H). Both the sense DNA incomplete probes (sDi) and the negative control showed a very weak or absent labeling (Fig. 3I and J).

The in situ hybridization using the detection system based on alkaline phosphatase produced a reddish-purple positive reaction exclusively in the forming β-layer of regenerating scales (4 weeks postamputation) or embryonic scales (stage 39; Fig. 4A–J). In apical regions of the regenerating tail, where the epidermis was still undifferentiated (stratified wound epidermal cells), no detection of labeled cells was seen in epidermal pegs (Fig. 4A and B). In more proximal regions where scales were differentiating new epidermal layers underneath the wound epidermis, the first fusiform β-cells appeared labeled with the probes. Differentiating fusiform cells of the β-layer appeared distinctively reactive for the probes and produced a red-purple stain. The stratification of the forming β-layer increased to 3–5 cell layers in fully differentiating scales (Fig. 4C–E). No other skin tissues were stained. Negative and sense controls were not stained, and only a brown pigmentation was sometimes seen (Fig. 4F). In embryonic scales, 2–3 layers of β-cells appeared specifically stained with the probe, while negative and sense controls showed no labeling or very low, mainly due to the pigmentation (Fig. 4K and L).

Figure 4.

In situ hybridization with alkaline phosphatase detection of antisense DNA complete probe of early regenerating stages (epidermal pegs; A and B), incomplete probe (aDi; C) on elongated pegs, complete probe (aDc) of later stages of elongated pegs (D and E), sense DNA incomplete probe (sDi; F) in regenerating scales. aDc in dorsal scales of embryo at stage 39 (G) and its control (co; H). Detail of labeled (aDc) fusiform β-cells in dorsal (I) and tail (J) scales of embryo at stage 39. Negative control (K) and sense control (L) of tail scales of embryo at stage 39. E, living epidermis; H, hinge region; T, scale tips; W, wound epidermis. Arrows indicate the differentiating the β-layer; arrowheads indicate pigments; dashes underline the basal layer; asterisks indicate non specific reaction. Scale bars = 10 μm.

Ultrastructural Immunocytochemistry and In Situ Hybridization

The ultrastructural examination, using immunogold of 10 nm, showed that the lizard β-antibody (A68B, directed against a 15–16 kDa lizard epidermal protein) decorated only electron-pale bundles of keratin among ribosomes of β-cells (Fig. 5A). The pale keratin bundles of the oberhautchen layer were less labeled or unlabeled with this antibody. High-magnification observation on early differentiating β-cells showed that small β-packets were labeled with the antibody (Fig. 5B). The small keratin packets formed a typical reticulate pattern or a network of matrix material of low electrondensity, mainly around more densely packed filaments of β-keratin material. The latter appeared at a following stage of compaction of the small β-packets forming the reticulate pattern.

Figure 5.

Ultrastructural immunolabeling of β-keratin filaments with the lizard keratin antiserum. A: Region of passage from the granulated (clear) layer to the underlying oberhautchen and β-layer. While no labeling is present over keratohyalin-like granules and keratin bundles (arrowheads) of clear cells, gold particles cover some oberhautchen β-keratin packets and intensely the pale keratin filaments of β-cells (arrow). Scale bar = 500 nm. B: Detail of labeling (arrowheads) of β-keratin packets in early differentiating β-cell. Scale bar = 100 nm. G, dense granule; KH, keratohyalin-like granules; O, oberhautchen serration.

Dense granules in oberhautchen and β-keratin cells were less labeled or completely unlabeled. Keratohyalin granules of the clear (granulated) layers and α-keratin bundles were unlabeled. The large β-keratin filaments were merging into a denser mass of corneous material in more mature compacting β-cells. These tangled filaments were intensely labeled with the lizard β-keratin antibody. The labeling disappeared in the differentiating mesos and α-cells that were formed underneath the β-layer (data not shown).

Ultrastructural in situ analysis over regenerating scales using the antisense complete DNA probe (aDc) showed clusters of gold particles sparse in the cytoplasm of differentiating β-keratin cells (Fig. 6A). The clusters were localized among ribosomes or associated with small β-packets: the latter formed the pale β-keratin reticulate pattern previously described (in Fig. 6B) among the larger β-keratin filaments (Fig. 6). This finer material was associated with the periphery of larger and compact β-keratin filaments, while their central part was generally devoid of clusters of gold particles (Fig. 6B). Gold particles often formed clusters among the reticulate network of material of low electrondensity: the latter represented early stages of β-keratin accumulation (Fig. 6C).

Figure 6.

Ultrastructural in situ hybridization in the cytoplasm of differentiating β-cells. A: Gold clusters (arrowheads) in the cytoplasm among β-packets. B: Detail of clusters of gold particles associated with the periphery of larger β-keratin filaments (arrowhead). C: Details of cluster of gold particles among alveolate β-packets (arrowhead) near a compact mass of β-keratin filament (arrow). Scale bars = 100 nm.

Although less frequent than in the cytoplasm, some gold clusters were also seen inside the nucleus of β-cells, especially among the finer euchromatine (Fig. 7). Clusters were sometimes seen associated with the nuclear membrane or localized in the cytoplasm near the nucleus: in this way, the sparse probe labeling extended from the internal part of the nucleus to the cytoplasm. The immunogold labeling was completely absent or randomly diffused and did not form clusters in other cells of the epidermis and dermis (data not shown) or in sense controls (Fig. 8).

Figure 7.

Ultrastructural in situ hybridization in nuclei of differentiating β-cells. A: Clusters of gold particles (arrow) within the nucleus (N). Arrowheads indicate unlabeled β-keratin filaments in the cytoplasm. B: Details of cluster of gold particles among reticulate euchromatin material. C: Details of a cluster of gold particle (arrow) near the nuclear membrane. N, nucleous; Cy, cytoplasm. Scale bars = 100 nm.

Figure 8.

Ultrastructural view of a sense control section of a differentiating β-cell. Both the cytoplasm and β-keratin filaments (double arrowheads) show completely absent or very sparse gold particles. Scale bar = 100 nm.

Comparative Analysis of Protein Sequences

The comparison between amino acid sequences of lizard scale (Dalla Valle et al.,2005) and claw (Inglis et al.,1987) with a keratin-associated protein, an α-keratin, a scale β-keratin, and a feather β-keratin is presented (Figs. 9–12). They showed numerous region homologies as it is reported in the Discussion section. A phylogram of different cytokeratins, β-keratins, and KAPs is reported in Figure 13.

Figure 9.

Sequence comparison of the lizard scale protein (B), deduced from its complete cDNA (A) (Dalla Valle et al.,2005), with KAP 16-1 of mouse (C), and α-2 keratin of lamprey (D). Stars indicate identities; colon indicates conserved substitutions; period indicates semiconserved substitutions detected by the ClustalW software. Glycine residues are evidenced in gray shadows. Underlines indicate the glycine-rich regions (mammalian-like) in N- and C-terminal ends of the lizard protein, while the boxed sequence indicates the central feather-claw-like sequence. Accession numbers of sequences analyzed are reported in Table 1.

Figure 10.

Sequence comparison of the amino acids of the lizard scale protein (A), with chicken scale keratin (B) and chicken feather keratin (C). Stars indicate identities; colon indicates conserved substitutions; period indicates semiconserved substitutions detected by ClustalW software. Glycine residues are evidenced in gray shadows. Underlines indicate the glycine-rich regions (mammalian-like) in N- and C-terminal ends of the lizard protein, while the boxed sequence indicates the central feather-claw-like, sequence Accession numbers of sequences analyzed are reported in Table 1.

Figure 11.

Sequence comparison of the lizard claw (A), with human KAP 5-1 (B) and a human cytokeratin type II (C). Stars indicate identities; colon indicates conserved substitutions; period indicates semiconserved substitutions detected by ClustalW software. Glycine residues are evidenced in gray shadows. Underlines indicate the glycine-rich regions (mammalian-like) in N- and C-terminal ends of the lizard protein, while the boxed sequence indicates the central feather-claw-like sequence. Accession numbers of sequences analyzed are reported in Table 1.

Figure 12.

Sequence comparison of the lizard claw sequence (A), with chicken scale β-keratin (B) and a chicken feather β-keratin (C). Stars indicate identities; colon indicates conserved substitutions; period indicates semiconserved substitutions detected by ClustalW software. Glycine residues are evidenced in gray shadows. Underlines indicate the glycine-rich regions (mammalian-like) in N- and C-terminal ends of the lizard protein, while the boxed sequence indicates the central feather-claw-like sequence. Accession numbers of sequences analyzed are reported in Table 1.

Figure 13.

Phylogram among different cytokeratins derived from some representative of different groups of vertebrates, from fish to mammals (bottom group in parenthesis), β-keratins from avian skin (intermediate group in parenthesis), and keratin-associated proteins from hairs (KAPs; upper group in parenthesis) that includes the scale and claw proteins of lizard (boxed). Numbers indicate distances (percent divergence) between all pairs of sequence. Accession numbers of sequences analyzed are reported in Table 1.

In Figure 9, we have compared the lizard scale protein with a keratin-associated protein (16-1) of mouse hair and with a cytokeratin (α-2) of lamprey epidermis. The homology between the lizard glycine-rich protein and the keratin-associated protein is high, while glycine-rich regions of the lamprey cytokeratin toward the N- or the C-terminal present relatively high homology with the lizard scale protein. In Figure 10, we have compared the lizard scale protein with a scale keratin of chick and with a feather keratin of chick. The homology is also high in most region of the scale keratin and in some regions of feather keratin. In Figure 11, we have compared the lizard claw protein with a human keratin-associated protein (5-1) and with a human keratin of type II. There is some homology in various glycine-rich regions between the lizard claw and a human KAP and less homology with the cytokeratins. In Figure 12, we have compared the lizard claw protein with scale and feather keratin of the chick, where the homology is high in various regions.

Finally, a phylogenetic analysis using the ClustalX (1.81) software has shown that lizard scale and claw glycine-rich proteins are more related to mammalian KAPs than avian β-keratins and less related to several vertebrate cytokeratins (Fig. 13). As a result of this analysis on the only two known primary sequences of reptilian keratins (lizard scale and claw), it appears that lizard keratins are a mosaic of sequences present in KAPs and in β-keratins of mammals and birds.

The conformational prevision, using the program “consensus secondary structure prediction,” for the lizard scale protein shows that no β-pleated sheets are present. In particular, 104 amino acids (64.2%) have a random coil conformation, 50 amino acids (30.7%) an extended strand conformation, and only 9 amino acids (5.5%) an α-helical conformation. The specific X-ray pattern and tertiary structure of this protein remain to be analyzed. It is possible that the strand conformation or interchain interactions are responsible for the β X-ray pattern ascribed to these proteins (Fraser et al.,1972; Fraser and Parry,1996).

Three amino acid regions of the lizard glycine-proline-rich protein are interesting, one in the N-region, one in the C-region, and one in the central region. They are here compared using the BLAST or ClustalX (1.81) software with sequences present in cited publication or deposited in Genbank database. In the present study, only some examples of sequence homology between reptilian and avian proteins are indicated (Figs. 9–12). The homology is indicated by the percentage of identities (= same amino acids) and positivities (= same amino acids or amino acids that preserve the physicochemical properties).

From amino acid residues 19–60 near the N-end, the lizard protein contains glycine-X or glycine-glycine-X sequences with high homology with mouse glycine-tyrosine-rich type I proteins (Genbank accession number AAKO7673.1; 62–78% identities and 65–81% positivities). Other sequence homology is present comparing lizard protein with goat keratin-associated protein (Genbank accession number AAR89458.1; 70% identity, 77% positivity), with human keratin-associated protein (Genbank accession number NP853640.1; 65% identity, 73% positivity), or with glycine-tyrosine rich proteins (Genbank accession number AAP97270.1, 56% identities, 61% positivities).

Near the C-terminal region, the lizard protein also presents homologies with mammalian glycine-tyrosine-rich hair-associated proteins. For example, it resembles a keratin-associated protein of mice hairs (Genbank accession number NP570927.1; 61% identities, 64% positivities), a glycine-tyrosine-rich mouse protein (Genbank accession number AAKO7673.1; 59% identities, 70% positivities), and also a human keratin-associated protein of hairs (Genbank accession number NP853638.1; 62% identities, 66% positivities).

From the above results, it can be seen that the N- and C-termini of the scale lizard glycine-proline-rich protein have mammalian-like homologies that suggest a common ancestry with those of mammals (Gillespie et al.,1982; Marshall and Gillespie,1982).

Finally, the central region between amino acids 77 and 118 resembles (40% identities, 61% positivities) various sequences of avian feather, scale, and claw keratins [see sequences in Table 2 of Sawyer et al. (2003)]. In particular, the analysis of amino acids 77–111 of the lizard protein shows homology with chick claw keratin (Genbank accession number XP 428214.1; 75% identities, 90% positivities), chick feather keratin I (Genbank accession number XP424547.1; 73% identities, 94% positivities), and another feather keratin (Genbank accession number 426556.1; 70% identities, 88% positivities).

The analysis of amino acids in position 26–60 show homology with chicken claw keratin (Genbank accession number P25692; 54% identities, 71% positivities), and analysis of amino acids 23–64 show homology with chick feather keratin (Genbank accession number P04458; 45% identities, 64% positivities). Therefore, a large part of the central region of the lizard glycine-proline-rich protein shows homologies with claw, scale, and feather proteins, which also may indicate some common ancestry with those of archosaurian progenitors. The possibility that these sequences derive from convergence of molecular evolution remains another possibility that requires further analysis of much more reptilian primary structure.

The lizard claw glycine-cystein-proline-rich protein also appears as a chimeric protein and presents homologous regions with both mammalian (high-sulfur- and high-tyrosine-rich proteins) and avian β-keratins (Gillespie et al.,1982; Inglis et al.,1987). Amino acid residues 1–36 have homology with human cytokeratins. Also, residues 1–36, 63–98, and 132–142 have homologies with high-cystein wool keratins. Finally, residue 32–67 shows homologies with feather keratin. Conformational analysis of the 142 amino acid lizard claw β-keratin sequence showed that the amino acid residues 36–67 and 94–125 determine the β-sheet component, as amino acids 24–55 in emu β-keratin of feathers also determines most of the β-sheet conformation (Fraser and Parry,1996). β-keratin sequences of birds, alligator, and lizard claw have a very similar amino acidic sequence from amino acids 39–53 (Fraser and Parry,1996; Sawyer et al.,2000).

Using the “consensus secondary structure prediction” program on the claw protein, it results that 105 amino acids (73.9%) have a random coiled conformation and 37 amino acids (26.1%) an extended strand conformation, while no β-sheet or α-helical conformation is present.


Cellular Synthesis and Distribution of Keratins

The present observations extend a previous study on mRNA isolation and localization for a lizard β-keratin (Dalla Valle et al.,2005). The study showed that the messenger codes for one small protein produced in β-cells of regenerating epidermis, and that β-keratin mRNA was highly expressed in differentiating β-cells. This indication suggested that the utilization of a cDNA probe, which is easier to handle and less susceptible to artifact degradations than cRNA probes, was sensitive enough for our hybridization study. The present study also shows that the mRNA for this protein is expressed also in embryonic epidermis and therefore suggests that the protein is ubiquitous to all β-layers of the epidermis of P. sicula (and P. muralis).

Using our rat antisera against a 15–16 kDa lizard β-keratin, we have for the first time shown that the thick keratin packets and bundles contain this small protein. The results confirm and overlap with the immunolocalization previously seen using antibodies directed toward avian β-keratins (Alibardi and Sawyer,2002; Alibardi et al.,2004a,2004b; Alibardi and Toni,2005). The labeling is uniform over β-keratin filaments and no fibrous structure (reminiscent of cytokeratin filaments) can be detected at high magnification. The study also confirms a previous study that showed the presence of a low immunolabelig for cytokeratin in β-keratin cells (Alibardi,2000). The latter observation indicates that the antiserum labels the amorphous matrix contained in β-filaments that show a thickness of 3–4 nm (the β-keratin pattern) (Maderson,1985; Landmann,1986). Both in situ hybridization and immunological results indicate that these basic proteins (pI 7.4–8.2) (Dalla Valle et al.,2005; Alibardi and Toni,2006a,2006b) belong to β-keratins.

The nuclear localization and the structure of the gene for this proteins (no introns appear to be present inside the coding region; Dalla Valle, personal communication) suggest that the probe can directly label the primary transcript inside the nucleus (Dalla Valle et al.,2005). The presence of sparse-labeled clusters from the inner part of the nucleus to the nuclear membrane and then mainly in the cytoplasm may be associated with the movement of the messengers toward the cytoplasm. The mRNAs are present not only in the “free” cytoplasm, but often also among the small β-keratin packets or associated in the periphery of larger β-keratin filaments. This confirms previous studies using tritiated proline that indicated that synthesis and polymerization of this protein occurs on the surface of β-keratins filaments (Alibardi,2004). This process is also typical for the synthesis of feather (β-)keratin in barb and barbule cells (Kemp et al.,1974).

Some Proteins of β-Keratin Filaments Resemble Mammalian Keratin-Associated Proteins

The amino acid sequence of the scale epidermal protein of P. sicula is very different from a proteins isolated from the claw of a varanus lizard (Inglis et al.,1987). The latter is an acidic protein of 13.1 kDa proteins (142 amino acids, pI 5.6), richer in sulfur and glycine, and with an amino acid composition similar to type II glycine-tyrosine-rich proteins of mammals.

The in situ hybridization study indicates that the sequenced glycine-proline-rich protein (Fig. 10A) is present in the β-layer of P. sicula scale. It is possible that our primer selection has isolated mainly the messengers for one of the minor fractions of hard keratins, relatively poor in cystein and resembling type I tyrosine-rich proteins (Marshall and Gillespie,1982). If this may be the case, other keratins richer in cystein should be present in β-cells, as indicated from biochemical (Gillespie et al.,1982) and histochemical (Alibardi,2001,2003) studies. These putative cystein-rich proteins remain to be isolated and sequenced in lizard epidermis.

The lizard scale protein has a molecular weight, amino acid composition, and predicted secondary conformation very different from that of α- or cytokeratins (Fuchs et al.,1987; Steinert and Freedberg,1991). The scale proteins localize in the same sites (β-keratin filaments) where a prevalent incorporation of tritiated proline was previously shown (Alibardi,2001; Alibardi et al.,2004a). The scale glycine-proline-rich protein enters in the composition of β-keratin filaments, which rapidly replaces cytokeratin bundles present in early differentiating β-cells (Alibardi,2000). This process resembles the localization of keratin-associated proteins among trichocytic keratins of mammalian hairs (Gillespie,1991; Powell and Rogers,1994; Rogers,2004). In comparison to mammalian glycine-tyrosine-rich proteins of 7–12 kDa, the glycine-cystein-proline-rich (varanus claw) and glycine-proline-rich (lizard scale) proteins are larger proteins with a molecular weight of 13–16 kDa and with a lower content in tyrosine.

The ultrastructural and molecular data suggest that both the lizard scale and claw glycine-proline-rich proteins have the composition of KAP isolated in mammalian corneous appendages (hair, nail, etc.) (Gillespie,1991; Powell and Roger,1994). Ongoing studies from this laboratory indicate that at least two more members of this protein type are present in the epidermis of P. sicula, and that similar glycine-rich proteins are present in the epidermis of other lizards, snakes, and turtles (data not shown).

Previous studies (Wyld and Brush,1979,1983) indicated that different β (ϕ)-keratin types were present in low amounts in reptilian α-layers and in higher amount in β-layers, as defined by histological and ultrastructural studies (Baden and Maderson,1970; Maderson,1985; Landmann,1986). It is likely that some ϕ-keratins of reptilian skin represent keratin-associated proteins localized in β-keratin filaments: the latter are present in high amount in β-layers and in very low amount in α-layers of reptilian epidermis. Therefore, aside from filamentous proteins, termed β-keratin, at least some proteins composing β-keratin filaments are represented by glycine-rich proteins. The filamentous nature of β-keratins present in β-keratin filaments remains uncertain and only the complete protein analysis of these filaments will clarify this problem (Gillespie et al.,1982; Marshall and Gillespie,1982; Wyld and Brush,1983).

It is likely that α-layers of reptilian epidermis contain a very small fraction amount of β (ϕ)-keratins among the prevalent bundles of cytokeratin. This has been detected by immunological and autoradiographic methods in turtle and lizard soft epidermis (Alibardi et al.,2004b; Alibardi and Toni,2006a,2006b). Differently from the α-layer, β-layers contain higher levels of β (ϕ)-keratins than α-layers (Wyld and Brush,1979,1983). Therefore, it is possible that α- and β-layers of reptilian epidermis simply reflect the massive activation of genes synthesizing keratin-associated proteins and β-keratins in differentiating β-cells or their nearly complete repression in α-keratogenic cells.

Reptilian ancestors of both mammals and birds might have produced glycine-rich proteins. The presence of more glycine-proline-rich and glycine-cystein-rich proteins in lizard epidermis indicates that also in reptiles, like in mammals, more members of KAPs are present. The latter conclusion has important implications for the general evolution of KAPs in amniote skin.

The two proteins so far sequenced in extant lizards (with others unpublished but similar sequences; data not shown) suggest that these proteins represent the prototype of a class of small proteins that might have been present in basal amniotes (reptiles by definition). Some of the genes coding for these proteins were later selected in the diapsid/anapsids (sauropsids) that evolved in modern reptiles and birds, while others proteins were selected in synapsid and therapsids that evolved into mammals. The genes coding for these proteins were inherited and largely modified to produce avian β-keratins, including scale, claw, beak, and possibly also feather keratins, the smallest (Gregg and Rogers,1986; Sawyer et al.,2000). In mammals, KAPs evolved instead to serve as hard material for keratinized appendages such as hairs, claws, and horns (Gillespie,1991; Marshall et al.,1991; Powell and Rogers,1994).