Molecular process involved in keratinization of mammalian epidermis are relatively well understood. In contrast, only little information information is available on genes and proteins involved in the process of keratinization in the epidermis of non-mammalian amniotes. In birds, most information is available on feather and scale beta-keratinization. In comparison with alpha-keratinization in avian soft apteric and interfollicular epidermis, very little is known on specific proteins involved in the process of keratinization in these tissues.

In reptiles, which are ancestors of both mammals and birds, little information is available on alpha- and beta-keratins. In particular, hard (beta)keratins can be fractionated into two main groups, those that are glycine–cysteine rich (resembling those of avian beak and scales) and those that are glycine–tyrosine rich.

Among amniotes, the soft epidermis of turtles (neck, limbs and tail) resemble more than any other type of epidermis that of basic amniotes of the Carboniferous, living approximately 340 millions years ago. In the soft epidermis of turtles, processes of keratinization are simple and show a specific adaptation to the semi-aquatic environment. Chelonians are a very ancient group of reptiles with an origin dating back nearly 300 millions of years, and it is possible that mechanisms of cornification in their soft epidermis (limbs, neck and tail) are conserved and resemble those in basic amniotes that were also living in amphibious conditions in the Carboniferous. Whether these mechanisms have remained unchanged and/or are still similar to mechanisms in basic amniotes is a matter of speculation. The histological structure of soft continuously renewed epidermis of turtles is relatively simple: a pluristratified keratinized epithelium without a granular layer.

During periods of intense growth, stratification of supra basal cells in the soft epidermis varies, but structures equivalent to mammalian keratohyalin are not formed in the pre-corneous (transitional) layer before cells enter the stratum corneum. The lack of keratohyalin in turtles has been confirmed by previous histoautoradiographic studies after administration of tritiated histidine. It is known that most turtle alpha-keratins have a molecular weight ranging between 40 and 61 kDa.

Another type of epidermis, hard and resistant but inflexible, is present in the carapace and plastron of turtles. The corneous layer consists prevalently of beta-keratins which are deposited over a scaffold of alpha-keratin.

In non-growing zones of the epidermis of carapace and plastron, the epithelial stratification is simple: few fusiform suprabasal keratinocytes and a dense beta-keratin network are produced, and transitional cells slowly enter the stratum corneum. In growing zones of the carapace or plastron, the epidermis becomes pluristratified and numerous layers of fusiform keratinocytes progressively form a new compact stratum corneum. Beta-keratin in these cells is probably rapidly accumulated, although kinetic of this process are not available.

The present study is the first to address the identification of specific proteins during soft and hard keratinization in turtle epidermis, and is part of a continuing program to analyze proteins involved in epidermal differentiation in reptiles. The study correlates cytological and biochemical data supporting the presence of epitopes that are shared with mammalian epidermis. The distribution patterns of protein markers of the cornified cell envelope (loricrin, sciellin and transglutaminase) and of the interkeratin matrix (histidin-rich proteins and filaggrin) have been analyzed as well in the present study. Our data suggest a phylogenetic relationship between molecular processes involved in cornification of amniote epidermis.

The present study was performed on young specimens (1–3 years of age) of the turtle Chrysemys picta, which were purchased in authorized pet shops. Animals were maintained in a water tank before the experiments, and in small tanks with a moist bottom during the experiments.

The specimens were injected with 4–5 ?Ci/g body weight of tritiated histidine (L-2,5-³H-histidine, specific activity 53.0 Ci/mmol; Amersham, Bukinghamshire, UK) between 10.00 and 12.00 to avoid circadian variations. Animals were killed by decapitation, and the soft skin (tail, limb and neck) and the hard skin from the carapace-plastron were collected at 1 h (n=5), 2 h (n=3), 4 h (n=5), and 24 h (n=5), 2 (n=3), and 4 (n=3) days post-injection. The skin was utilized for both histological and biochemical studies.

For autoradiography, small pieces of tissues were immediately fixed in 2.5% glutaraldehyde in phosphate buffer, osmicated or non-osmicated, and embedded in Durcupan resin. Sections, 1–4 ?m thick, were collected using an ultramicrotome, coated with K5 (Ilford, Mobberley, Cheshire, UK) nuclear emulsion for light microscopical autoradiography, exposed for 1–2 months, and developed with D19 developer (Kodak, Rochester, NY, USA). For ultrastructural autoradiography, soft and hard skin were collected at 4 h post-injection on collodium-coated slides, covered with K4 Nuclear emulsion (Ilford, Mobberley, Cheshire, UK), exposed 2–4 months, and developed with Kodak D19. Sections were picked up on nickel or copper grids, and stained for electron microscopical purposes.

For immunocytochemistry, small pieces of skin were fixed in Carnoy or 3% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4, for 4–6 h, dehydrated and embedded in Bioacryl resin, that was prepared by mixing organic chemical components as reported in Scala et al. (1992). The pan-keratin antibody was purchased from Sigma (St. Louis, MO, USA). The antibody against rat filaggrin was generously supplied by Dr. B.A. Dale (Department of Oral Biology, University of Washington, Seattle, WA, USA). The anti-loricrin antibody was a gift of Dr. E. Fuchs (Howard Hughes Medical Institute, University of Chicago, IL, USA, rabbit polyclonal anti-mouse loricrin), or purchased from BabCo-Covance (AF 62, Richmond, CA, USA). The rabbit polyclonal anti-chick b-1 keratin was supplied by Dr. RH Sawyer, University of South Carolina, Columbia, SC, USA.

The antibody against transglutaminase ab421 against guinea pig transglutaminase was a rabbit polyclonal antibody, and was purchased from Abcam Limited (Cambridge, UK). Finally, the rabbit polyclonal anti-human sciellin was a generous gift by Dr. H. Baden (Massachussetts General Hospital, Harvard Medical School, Boston, MA, USA).

Light microscopical immunocytochemistry was performed using fluorescein thioacyanate-conjugated antibodies (Sigma) as previously described. As negative controls, primary antibodies were omitted. Positive controls were used as well.