Observations on the Histochemistry and Ultrastructure of ...lemastm/Publications/PDF/Snake Skin/Alibardi_et_al-2003-Journal_of... · dermis of Sphenodon punctatus conﬁrms previous

Observations on the Histochemistry and Ultrastructureof the Epidermis of the Tuatara, Sphenodon punctatus(Sphenodontida, Lepidosauria, Reptilia): A Contributionto an Understanding of the Lepidosaurian EpidermalGeneration and the Evolutionary Origin of theSquamate Shedding ComplexLorenzo Alibardi1 and Paul F.A. Maderson2*

1Dipartimento di Biologia evoluzionistica sperimentale, University of Bologna, Bologna 40126, Italy2Biology Department, Brooklyn College of CUNY, Brooklyn, New York 11210, USA

ABSTRACT Histochemical and TEM analysis of the epi-dermis of Sphenodon punctatus confirms previous histologi-cal studies showing that skin-shedding in this relic speciesinvolves the periodic production and loss of epidermal gen-erations, as has been well documented in the related Squa-mata. The generations are basically similar to those thathave been described in the latter, and their formation in-volves a cyclic alternation between �- and �-keratogenesis.The six differences from the previously described squamatecondition revealed by this study include: 1) the absence of awell-defined shedding complex; 2) the persistence of plasmamembranes throughout the mature �-layer, including theoberhautchen; 3) the concomitant presence of lipogenic la-mellar bodies and PAS-positive mucous granules in mostpresumptive �-keratinizing cells; 4) the presence of the se-creted contents of these organelles in the intercellular do-mains of the three derived tissues, the homologues of thesquamate mesos, �-, and lacunar cells; 5) the paucity oflamellated lipid deposits in such domains; 6) the presence ofkeratohyalin-like granules (KHLG) in the presumptive lacu-nar, clear, and oberhautchen cells. In toto, the absence ofmany of the precisely definable, different pathways of cyto-genesis discernible during squamate epidermal generationproduction might be interpreted as primitive for lepidosaurs.However, when the evolutionary significance of each of thesix differences listed is evaluated separately, it becomesclear that the epidermis of S. punctatus possesses primitiveamniote, shared and derived lepidosaurian, and someunique characters. This evaluation further elucidates theconcept of a lepidosaurian epidermal generation as a derivedmanifestation of the sauropsid synapomorphy of vertical al-ternation of keratin synthesis and shows that further studyof keratinocyte differentiation in the tuatara may contributeto our understanding of the origin and evolution of�-keratinization in sauropsid amniotes. J. Morphol. 256:111–133, 2003. © 2003 Wiley-Liss, Inc.

KEY WORDS:tuatara; Sphenodon; Lepidosauria; squa-mate; epidermis; keratin; evolution

Skin shedding in lizards and snakes involves theperiodic formation and loss of epidermal genera-

tions. Over the past 40 years the underlying cycliccellular activities have been well documented bylight (LM) and electron microscopy (EM) (reviewedby Maderson, 1985; Landmann, 1986; Maderson etal., 1998). Fundamentally, squamate epidermis is astratified squamous epithelium wherein populationsof daughter cells derived from a morphologicallyhomogeneous stratum germinativum (Maderson etal., 1972) manifest a conspicuous vertical alterna-tion between two distinct keratogenic pathways, em-phasizing either �- or �-proteins. Although mecha-nisms directing this alternation remain uncertain(Maderson, 1985, pp. 555–558), many studies haveconfirmed the presence of morphologically and func-tionally distinct epithelia associated with the tran-sitions from one pathway to the other.

During the transition from �- 3 �-keratogenesis,the flattened cells of the mesos layer are laid down.The mature tissue houses the barrier to cutaneouswater loss (CWL) (Lillywhite and Maderson, 1982),a function that depends on mesos granules, a type oflipogenic lamellar body (LB). Transmission electronmicroscopic (TEM) study shows that, in mammals,LBs exocytose their contents into the extracellulardomain, where they form lipid-rich, bilayered struc-tures (Menon et al., 19921996; Menon and Ghadi-ally, 1997; Menon and Menon, 2000).

During the transition from �- 3 �-keratogenesis,a shedding complex is laid down. During the pre-shed renewal phase, membranes of cells of the clear

Contract grant sponsor: Italian MURST 60% fund.

*Correspondence to: Dr. P. Maderson, 210 Axehandle Road, Quak-ertown, PA 18951-5904. E-mail: [email protected]

DOI: 10.1002/jmor.10079

JOURNAL OF MORPHOLOGY 256:111–133 (2003)

© 2003 WILEY-LISS, INC.

layer (the innermost component of the outer gener-ation, the next to be shed) develop interdigitationswith those of the subjacent oberhautchen (the out-ermost component of the inner generation that willbe exposed next) whose eventual separation permitsshedding. These events are barely discernible byLM. Details of membrane interdigitation and theresultant microornamentation demand, respec-tively, TEM and scanning electron microscopic(SEM) analysis (Maderson et al., 1998).

Vertical alternation of �- and �-keratogenesis inthe epidermis is a sauropsid synapomorphy andlepidosaurian skin shedding is its best-known man-ifestation (Maderson and Alibardi, 2000). The exis-tence of two unique squamate tissues (mesos layerand shedding complex) associated with the alternat-ing transitions raises questions regarding their evo-lutionary origin, answers to which must be sought inthe closest living relative of squamates, the relicsphenodontid species, the New Zealand tuatara(Sphenodon punctatus).

An LM study of tuatara epidermis (Maderson,1968) confirmed earlier reports of the presence ofepidermal generations. It was found that, in con-trast to its syncytial form and chromophobia insquamates, the mature �-layer retains its stratifiedsquamous morphology so that its tinctorial proper-ties grade insensibly into those of the �-layer. Pres-ence of a mesos layer was suggested by artifactualsplitting of mature corneous tissues during tissueprocessing. Identification of a shedding complex wasuncertain due to lack of renewal phase material (seebelow). By LM, an oberhautchen was discernibleonly as a refractive line on the surface of the �-layer:the paucity of microornamentation was later estab-lished in an SEM study (Peterson, 1984).

The present TEM study extends identification ofepidermal components in Sphenodon punctatus.New data 1) permit more precise comparison withthose of squamates and extend understanding of thelepidosaurian epidermal generation, 2) elucidate theevolutionary origin of the squamate barrier to CWLand shedding complex, and 3) offer insights into theorigins of amniote keratinization.

MATERIALS AND METHODS

Pieces of caudal skin (1–3 mm long for TEM, 5 mm for histol-ogy) were sampled from six specimens of Sphenodon punctatus.Three were 3-month-old juveniles hatched in the laboratory fromeggs collected on Stephens Island, Cook Strait, New Zealand, byauthority of the NZ Department of Conservation (Thompson,1990). The other three were adults captured during the summeron Stephens Island (Alibardi, 1992). Status of individuals withinshedding cycles was unknown.

Samples for TEM were fixed in 2.5% glutaraldehyde in Ringer’s(pH 7.4–7.6), stored in Ringer;s (12–16 h), postfixed in 1% os-mium tetroxide (2 h), dehydrated in graded ethanols, embeddedin Epon, and sectioned on an ultramicrotome.

Semithick sections of normal skin from different regions of thetail were stained with 0.5% toluidine blue. Some sections weredouble-stained with eosin for 3–10 sec over a hotplate, washed,

and then stained with 0.5% toluidine blue. Others were deplas-ticized with sodium-ethoxide prior to staining with PAS and 1.0%alcian blue to detect mucosubstances (Troyer, 1980). Thin sec-tions collected on 200–100 mesh copper or nickel grids werestained with uranyl acetate and lead citrate and examined with aPhillips CM 100 electron microscope.

Larger pieces of tail (5–10 mm) were fixed in 10% formalde-hyde, ethanol dehydrated, cleared in xylene, and wax embedded.Sections 6–10 �m were stained with hematoxylin and eosin.

RESULTSGross Morphology

Tail scales vary in size and shape depending onthe region examined (Fig. 1). Those on the charac-teristic mid-dorsal ridge are heavily keeled. Whilemost dorsolateral scales overlap little or not at all,the larger medioventral scales are slanted and par-tially overlapped, permitting identification of outerand inner surfaces and hinge regions (Maderson,1964).

Light Microscopic ObservationsData presentation — definitions of three epi-

dermal cytologies. With material from only sixanimals available for study, some aspects of cyclicepidermal activity were not represented. Addition-ally, a description of semithick sections (Figs. 2–6) isconfounded by 1) great variability in tinctorial prop-erties of keratinized tissues within and betweenspecimens, and 2) unique cytologies. These factorscomplicate correlation with previous LM studies(Maderson, 1968) and comparison with squamatecycle stages (Maderson, 1985; Maderson et al., 1998,fig. 1).

To facilitate communication, we define six differ-ent epidermal cytologies in our material as Condi-tions One, Two, and Three. Each definition is accom-panied by a brief comment on the general relation ofthe condition to the squamate cycle. The definitionsare followed seriatim by LM descriptions of 1) as-pects of keratinized tissues of the outer generationcommon to all specimens, 2) the inner generation inone specimen, and 3) the stratum germinativum andsuprabasal cells in all six. These descriptions permittentative assignment of each of the six specimens toa stage of the squamate cycle. In accounts of TEMand histochemical data for each condition that com-plete the results, the assignments permit specificidentifications of cell types that are justified in theDiscussion section.

Condition One, seen in samples from two juvenilesand one adult, shows an incomplete outer genera-tion above flattened basal cells (Fig. 2). It reflects aresting stage of squamates.

Condition Two, seen in samples from two adults,shows a complete, but immature, outer generationabove cuboidal/columnar basal cells (Fig. 3). This isidentified as a Stage 2 of squamates.

112 L. ALIBARDI AND P.F.A. MADERSON

Condition Three, seen in one juvenile, shows acomplete, but immature, outer generation, an in-complete inner generation, and a germinal layer(Figs. 4, 5). This reflects a renewal phase but uniquefeatures of the inner generation prohibit directequation with any single squamate stage.

Histology of the corneous tissues of the outergeneration (Figs. 2–6). The mature �-layer’s sur-face is slightly waved: some toluidine blue-stained

sections show a dark line with occasional serrations.The entire tissue is intensely toluidinophilic in ma-terial from juveniles (Fig. 2), but stains lighter, ornot at all, in adults (Fig. 3). In wax sections, treatedwith eosin, it stains only slightly, or not at all, what-ever the animal’s age. A distinctive lamellate, tile-like pattern is visible in lightly stained sections (Fig.6). On the outer scale surface, the �-layer is thickerin adults (�5 cell layers) (Figs. 3–5) than in juve-

Fig. 1. Sphenodon punctatus. Gross view of normal adult tail showing different forms of scales dorsally (D), laterally (L) andventrally (V). Bar � 2.5 mm.

Fig. 2. Sphenodon punctatus. Juvenile. Light micrograph Condition One epidermis. Above very flattened basal cells (arrowheads)is a single layer of living cells. Keratinized tissues of the outer generation comprise a densely stained �-layer (�) artifactually separated(arrow) from a thin �-layer (�). Numbered squares or rectangles, here and in other figures, refer to comparable cells or tissues shownin subsequent figures. Toluidine blue. Bar � 10 �m.

Fig. 3. Sphenodon punctatus. Adult. Light micrograph Condition Two epidermis. Basal cells (B) are cuboidal and there are strikingdifferences in tinctorial properties of keratinized tissues (contra Fig. 2). �-Layer (�) is artifactually separated (arrowheads) from�-layer (�). Upward and downward pointing open arrowheads associated with regions indicated by “8, 9” here, and in Figure 4, indicateextent of artifactual splitting through mesos region. Pigment granules scattered among suprabasal cells (S) derive from arborizedmelanocytes. D, dermis. Toluidine blue. Bar � 20 �m.

Fig. 4. Sphenodon punctatus. Juvenile. Low-power light micrograph Condition Three epidermis. Outer generation is representedby a darkly stained �-layer (�o) artifactually separated (open arrowheads) from chromophobic �-layer (�o) on whose inner aspect aresome living cells not resolvable at this magnification (see Fig. 5). The complex form of tissues of immature, inner generation lying abovethe stratum germinativum (B) is shown in Figure 5. Toluidine blue. Bar � 40 �m.

Fig. 5. Sphenodon punctatus. Juvenile. Light micrograph Condition Three epidermis. �-Layer of outer generation lost duringsectioning. Note strands of darkly stained material on outer aspect of chromophobic �-layer (�o) and some extremely flattened cells(arrowhead) on its inner aspect. A differentiating inner generation lying between the latter and the columnar basal cells presents fourdifferent images, unique in toto to the tuatara, identified in vertical sequence as regions A–D. In regions A (chromophobic) and B(darkly stained) no cell outlines can be discerned. In region C there are 2–4 layers of swollen, polygonal cells, some of which aretoluidinophilic and have a floccular cytoplasm (arrows). In region D there are 5–8 layers of cells whose degree of flattening increasesfurther away from the basal layer. See text for further explanation. Toluidine blue. Bar � 15 �m.

Fig. 6. Sphenodon punctatus. Light micrograph of mature �-layer showing characteristic lamellar/tile-like stratified squamousmorphology. Arrowheads point to external surface. Hematoxylin and eosin. Bar � 5 �m.

113TUATARA EPIDERMIS

niles (3–5 cell layers). The �-layer is always thinnerin the inner scale surface and hinge region, some-times only one cell thick.

An artifactual split traversed by thin strands, var-iously stained, always occurs at the base of the�-layer (Figs. 2–4). Such were interpreted as LMindications of a mesos layer (Maderson, 1968) butTEM data demand reevaluation of direct homology.

The �-layer is thinner in the three specimensshowing Condition One (Fig. 2) than in the otherspecimens (Figs. 3–5). The flattened, toluidinopho-bic (but eosinophilic) cells have a typical stratifiedsquamous epithelial form at high magnification.

Some extremely flattened, apparently keratin-ized, cells visible at the base of the �-layer in Con-dition Three (Fig. 5) are the innermost componentsof the outer generation: their specific identities arediscussed after presentation of their ultrastructuralfeatures.

Histology of the inner generation (Figs. 4, 5).The inner generation and basal cells in ConditionThree vary greatly. This might reflect a centrifu-

gal gradient of cytodifferentiation across the outerscale surface during the renewal phase (cf. squa-mates, Maderson et al., 1998, p. 16), but addi-tional unique features in Sphenodon punctatuswould defy interpretation without TEM data.While it would be inappropriate to offer specificidentifications of the various cell populations inFigure 5 before such data are presented, any at-tempt to present them without reference to theentire tissue would be confusing. Here, we identifyregions A–D in Figure 5 and offer brief commentson the LM appearance of each.

Region A has a stratified squamous form resem-bling that shown in Figure 6. Apart from its mostsuperficial portion it is very lightly toluidinophilic.

Region B is so strongly toluidinophilic that itssquamous form is only discernible at high magnifi-cation.

Region C comprises 2–4 layers of slightly swollen,polygonal cells: toluidinophilia increases away fromthe basal layer. Healthy nuclei can be seen in thefloccular cytoplasm.

Fig. 7. Sphenodon punctatus. Low-power TEM micrograph of innermost tissues in Condition One epidermis showing part of a flatgerminal cell (B) with basement membrane (arrowheads). A single suprabasal cell with organelles typical of an immature keratinocytelies between the stratum germinativum and immature �-layer (�). The latter comprises only four cells that become flatter towards themesos region (MR). Bar � 1 �m.

Fig. 8. Sphenodon punctatus. Medium- to high-power TEM micrograph of �-layer (�) and mesos region (MR) in Condition Oneepidermis. The gradually increasing degree of flattening of �-cells (Fig. 7) is seen again. In some extracellular spaces desmosomalremnants (arrows) and occasional lipid lamellae can be discerned. Bar � 200 nm.

Fig. 9. Sphenodon punctatus. High-power TEM micrograph of indistinct transition between �-layer (bottom) and mesos region (top)in Condition One epidermis. All cells have filamentous contents and similarly thickened marginal regions (arrows). Occasional,scattered images of extracellular lipid lamellae can be seen in these tissues. Bar � 200 nm.


Region D comprises 5–8 layers of living cells. Thedeepest are polygonal: the others are increasinglyelongate and flattened away from the basal layer.

A shedding stage assignment for Condition Threeis deferred until the germinal and suprabasal cellshave been described.

Histology of the stratum germinativum andsuprabasal cells. The deepest living cells in Con-dition One (Fig. 2) differ from those in the other twoconditions (Figs. 3–5). Germinal cells are so flat thatovoid nuclei lie with long axes paralleling the skinsurface and suprabasal nuclei are sparse. This formcharacterizes the squamate “perfect resting stage”(PRS), but the ultrastructure of superjacent keratin-ized tissues defines a post-shed resting stage (seeDiscussion).

The cuboidal/columnar basal cells and polygonalsuprabasal cells in Condition Two (Fig. 3) haverounded nuclei, often with prominent nucleoli. Epi-dermal melanocytes are arborized, with widely dis-persed pigment granules. Some pycnotic, flattenednuclei occur at the base of the �-layer. All thesefeatures suggest direct homology with very earlyrenewal phase (Stage 2) of the squamate cycle.

Chromophobia of columnar basal cells in Condi-tion Three (Figs. 4, 5) resemble that seen in squa-mate epidermis during mesos maturation (lateStage 5 to early 6), but features of regions B and Csuggest an earlier stage in the renewal phase. Ul-trastructural data do not resolve this anachronisticaspect of tuatara epidermis, but emphasize that cy-todifferentiation here is less precise than in squa-mates.

Ultrastructural and HistochemicalObservations

Condition One (Figs. 7–9). Resting on a base-ment membrane, flattened germinal cells havemany desmosomes where they contact suprabasalcells (Fig. 7) and other organelles typical of undif-ferentiated amniote keratinocytes. The suprabasalcells resemble presumptive �-cells. The superjacent10–16 layers of keratinized cells become progres-sively thinner in ascending vertical sequence (Figs.7–9). The first 5–6 are moderately electron-dense.Their irregular profiles show marginal layers 10–17nm wide, desmosomal remnants, and intercellularlipid deposits (Figs. 7, 8). A pattern of electron-lucent, 6–8 nm-wide filaments against a darkerbackground matrix (lowermost cells) changes to onewith 15–20 nm filaments — all features resemblingmammalian �-keratinization. Sequential flatteningof the remaining 5–10 layers produces cells �0.1 �deep (Figs. 7–9) that resemble those in mature squa-mate mesos layer, but there is no clear distinctionbetween “�-” and “mesos” cells in the tuatara.

Ultrastructural features of the mature �-layercommon to all three conditions are described below.

Condition Two (Figs. 10–17). Figure 10 con-firms LM data (Fig. 3) showing differences in generalform of living cells from those in Condition One (Fig.2). While bundles of 6–8 nm filaments in suprabasalcells (Figs. 11, 13, 15) imply an �-keratogenic capacitycomparable to that of presumptive �-cells in ConditionOne (Fig. 7), several features indicate additional syn-thetic capacities.

The electron-lucent, amorphous contents of vesi-cles blebbing from the maturing (trans) face of thewidely distributed Golgi complexes coalesce to formtwo types of electron-dense bodies: small ovoids, andlarger, irregular or roundish units (Figs. 11–14). Thesmall ovoid vesicles (0.2–0.3 � 0.09–0.15 �m) showan internal lamellar pattern on a finely granularbackground (Figs. 11–13). Lamellar periodicity,measured as the distance between adjacent denselines, was usually 6.0–6.5 nm but ranged up to 8.0nm. These organelles are identified as lamellar bod-ies (LBs) (Menon and Ghadially, 1997). Coarse gran-ulation in the larger (0.2–1.0 �m) bodies resemblesmucous granules (MGs) of amphibian skin (Parrakaland Matoltsy, 1964; Lavker, 1974). Contents of bothLBs and MGs appear to be discharged into the ex-tracellular domain at the base of the �-layer (Fig.15) in regions where occasional cells with numerousMGs occur (Fig. 16).

Cells on the �-layer’s inner aspect (Figs. 10, 13),whose recent addition to it is suggested by con-densed chromatin in the electron-dense nuclei, arenot as flat as more superficial cells: they resembletransitional cells of mammalian stratum corneum.

The ultrastructure of the other keratinized tissuesresembles that seen in Condition One.

PAS staining (Fig. 17) is weak or negative in themature �-layer, but positivity increases steadily to-wards the base of the �-layer. Occasional living cellswith large PAS-positive granules are apparentlythose that contain numerous MGs (Figs. 16, 17).

Condition Three (Figs. 18–26). The manyunique features are best approached by noting theorientation within the entire epidermis (Figs. 4, 5) ofa low to medium power TEM (Fig. 18) that straddlesthe tissues of region A through the uppermost tis-sues of region C (Fig. 5). There is extreme variationin the form of contacts between the electron-lucentcells, ranging from flat to a complex interdigitation(Fig. 18). Plasma membranes are thickened withdeposits of electron-dense material 8–12 nm wide.Desmosomal remnants are visible throughout, butmany cells remain separated by spaces 10–15 nmwide (Fig. 19). The cytoplasm of a single cell orientedobliquely across Figure 18 has a “basket-weave” pat-tern similar to that seen in squamate presumptive�-cells (Maderson et al., 1972, 1998; Maderson,1985). The round or ellipsoid inclusions forming thispattern are similar in electron-lucency to the morehomogeneous cytoplasm of the adjacent, more flat-tened cells. They may be associated with bundles offilaments converging on desmosomes. They merge


with one another in cells as nuclear heterochroma-tin increases, ribosomes diminish, and the cyto-plasm becomes swollen and vacuolated. At higherresolution (Fig. 19) it is sometimes possible to mea-sure a granular pattern of 3–5 nm electron-lucentfilaments. A similar pattern is also discernible in theoutermost cells of the mature �-layer (Figs. 6, 20). Inmature cells (Figs. 19, 20) occasional melanosomesand irregular electron-dense areas are seen, but10–20 nm-thick filaments are absent. All these fea-tures suggest that Figure 18 shows cells in variousstages of �-keratogenesis and that the inclusions arepackets of �-keratin filaments (Alibardi, 1998a).During final maturation the cell extremities (Figs.18, 19) fit together producing the characteristic formof the mature �-layer (Fig. 6).

The variability in membrane form (see above) canbe seen on a single cell at the surface of the �-layerof the OG (Fig. 20), where the almost continuouselectron-dense thickening of the inner leaflet, 8–12nm wide, highlights occasional short (0.15–0.20 �m)projections. By definition, this is the oberhautchen,

and the projections are the sparse microornamenta-tion reported by Peterson (1984).

Interpreting regions A–C (Fig. 5) as sequentialsteps in �-layer differentiation permits identifica-tion of the extremely flattened cells in region D.Because sparse intracellular LBs and their extracel-lular products occur throughout most �-keratogenictissues, cells in region D must be precursors of thekeratinized tissues in Figures 7–9.

Identification of differentiating �-cells of the in-ner generation permits further comment on ober-hautchen fine structure (Fig. 20). Because most ofthe remaining keratinized components of theouter generation in Condition Three are un-changed from Condition Two, attention can focuson flattened cells at the base of its �-layer (Fig. 5).These show diverse TEM images (Figs. 21–24).The morphology of cell– cell contact varies fromregion to region (Figs. 21, 22) and across singlecells (Fig. 24). The cytoplasm is generally moreelectron-dense than that of subjacent �-cells (moreobvious in Figs. 21 and 24 than in Fig. 22), and a

Fig. 10. Sphenodon punctatus. Low power TEM micrograph of Condition Two epidermis showing a small portion of a keratinizedcell at base of mature �-layer (arrow, top left). Above cuboidal cells of stratum germinativum (B) lie five layers of suprabasal cells (S).A melanocyte dendrite (double arrowhead) can be seen between the basal cells. Bar � 2.5 �m.

Fig. 11. Sphenodon punctatus. Medium- to high-power TEM micrograph of cytoplasm of a suprabasal — presumptive lacunar —cell in Condition Two epidermis (cf. Figs. 12–17). Some blebbing Golgi vesicles (arrowheads) contain finely granulated material; othersshow lamellated material (arrow). Larger, membrane-bounded bodies (M) store coarsely granulated mucus-like material. Bar � 250nm.

Fig. 12. Sphenodon punctatus. Medium- to high-power TEM micrograph Condition Two epidermis. In this presumptive lacunarcell, the beginning of lamellation within a finely granulated organelle identifies it as a lamellar body. Bar � 250 nm.


lighter 3–5 nm filament pattern is just discernibleagainst a finely granulated matrix. Usually, scat-tered, round or irregular, electron-dense granules(Fig. 23) distinguish them from “typical” �-cells inSphenodon punctatus. Desmosomal remnants oc-cur frequently and in regions of greatest mem-brane complexity (Fig. 22) conspicuous intercellu-lar gaps (perhaps magnified by sectioning) occur.

A similar mixture of features suggesting both �-and �-keratogenic capabilities in the cells describedabove characterizes the most superficial cells of the�-layer of the outer generation (Fig. 25).

DISCUSSIONLepidosaurian Epidermal Generation:Introductory Comments

Squamate epidermal generations are complexunits with outer �-, and inner �-keratogenic, re-gions. When mature, the former (including the su-perficial oberhautchen) is syncytial but its multicel-lular origin is evident during differentiation. Mature�-keratogenic tissues retain cellular form and derivefrom four definable cell types (Maderson, 1985, tableI). This picture of the basic squamate epidermalgeneration as comprising six distinct cell types hasbeen well substantiated. Its application to Spheno-don punctatus in LM studies (Maderson, 1968)seemed justified and, until recently, it seemed that

epidermal generations were basically similar in alllepidosaurs (Maderson et al., 1998).

This study shows that the epidermal form inSphenodon punctatus differs from previous TEM ac-counts of squamate EGs in two locations, both in-volved with transitions from one pattern of kerato-genesis to another. The �- 3 �-, and �- 3�-transitions concern, respectively, the locations ofthe squamate mesos layer and shedding complex,the two specialized tissues whose evolutionary ori-gin we sought to elucidate. Interpretation of differ-ences revealed by TEM is hindered by the unknownstatus of individuals within their shedding cyclesand the limited material. We encountered other un-expected difficulties whose resolution necessitatesreview of the general pattern of cyclic epidermalactivity in lepidosaurs before we address the evolu-tionary issues.

Cyclic Events in Tuatara and SquamateEpidermis: Equating ConditionsWith Stages

Recent research on lepidosaurian skin has fol-lowed premises established over a century ago (re-viewed by Lange, 1931). Earlier workers definedwhat was lost from the body as an outer generationand wrote of its replacement by a newly formedinner generation. They recognized the most superfi-

Fig. 13. Sphenodon punctatus. Low-power TEM micrograph of Condition Two epidermis. Nuclei (N) of mature �-keratogenic cells(top) contain clumped heterochromatin. In cytoplasm of subjacent presumptive lacunar, numerous dense and lamellated LBs (arrows)(also see insert for high-power TEM micrograph, bar � 200 nm) stand out against lighter pre-keratinized cytoplasm. Bar � 1 �m.

Fig. 14. Sphenodon punctatus. Medium- to high-power TEM micrograph Condition Two epidermis. In this presumptive lacunarcell, condensed mucous granules (M) lie next to a Golgi complex (G) with blebbing vesicles. Bar � 200 nm.


Figure 15–20


cial generation component as the oberhautchen andthe innermost as the clear/granular layer. The lat-ter, and indeed several other components of a gen-eration, were variously named. A nomenclaturebased on LM study (Maderson, 1965) avoided termslikely to cause confusion with mammalian epider-mis, e.g., stratum intermedium, s. spinosum, or s.granulosum, but was necessarily modified as EMdata appeared (Maderson et al., 1972, 1998). How-ever, because interpretation of cellular identity de-pends on how the system is conceptualized, newdata raise new problems. The first report on squa-mate epidermis, based on longitudinal biopsies(Maderson, 1965), conceptualized its cyclic histogen-esis as resembling the mammalian hair cycle with“resting phases” alternating with “renewal phases.”On the assumption that an entire generation wasproduced during a renewal phase, suprabasal cellspresent at shedding were interpreted as presump-tive lacunar and clear cells that matured concur-rently with inner generation differentiation duringthe next renewal phase (loc. cit.).

The original conceptualization was revised whenlater studies permitted more precise analysis of celllineage and elucidated cycle dynamics. Although thecapacity for cyclic activity is intrinsic to the epider-mis (Flaxman et al., 1968), frequency of its expres-sion is controlled by hormones that alter the dura-tion of the perfect resting stage (PRS) (Maderson etal., 1970a; Maderson, 1985). Because longitudinalbiopsying of lizards such as Gekko gecko or Anoliscarolinensis that shed every 14–16 days reveals noPRS morphologies, epidermal form in infrequentshedders, e.g., most snakes, can be predicted bynoting the animal’s status within a cycle. Such datahave been incorporated into a model of the still

incompletely understood mechanisms involved inswitching between two keratogenic patterns (Mad-erson, 1985, pp. 557–558, fig. 10). The model pro-poses that a late resting stage morphology reflects adownregulated “�-synthesizing signal” and an up-regulated “�-synthesizing signal,” while a Stage 4morphology (Maderson et al., 1998, fig. 1) reflectsthe reverse.

The squamate PRS is definable morphologically.An incomplete outer generation, comprising mature�-, mesos, and �-layers, and one layer of suprabasalcells lies above a stratum germinativum, whose cellsare conspicuously flattened. It is uncertain whetherthe system is absolutely quiescent when a PRS oc-curs (Maderson, 1985, pp. 552–557). While extensivelongitudinal biopsies from Elaphe g. guttata re-vealed occasional “proliferating histologies” (Zucker,1977), 87% of 328 field-collected specimens of twolizard species showed PRS (Maderson et al., 1970b).Differences in epidermal structure between Spheno-don punctatus and squamates could reflect differ-ences in patterns of cyclic activity underlying gen-eration production. In the absence of dataconcerning renewal phase duration in S. punctatus,we can approach questions pertaining to the evolu-tionary origin of the squamate shedding complexand mesos layer by comparing epidermal morpholo-gies reflecting their formation in the two taxa. Thisnecessitates equating the three conditions describedhere for S. punctatus with squamate stages (Mader-son et al., 1998, fig. 1).

Only Condition Two equates directly with a des-ignated squamate stage, Stage 2 of early renewal, asindicated by a mature �-layer and subjacent, ar-borized, epidermal melanocytes. The cytology (Fig.3) is identical to that of two snakes, Constrictor

Fig. 15. Sphenodon punctatus. Medium- to high-power TEM micrograph Condition Two epidermis shows mucus-like materialbeing discharged into intercellular space (arrowheads) at base of mature �-layer (�). Bar � 200 nm.

Fig. 16. Sphenodon punctatus. Low-power TEM micrograph Condition Two epidermis showing cytoplasm of a presumptive lacunarcell with a flattened nucleus (N) lying above numerous large mucous granules (M). Bar � 1 �m.

Fig. 17. Sphenodon punctatus. High-power LM Condition Two epidermis showing a mucogenic lacunar cell containing PAS-positivegranules (arrows). K, innermost aspect of keratinized �-layer. PAS technique. Bar � 10 �m.

Fig. 18. Sphenodon punctatus. Low-power TEM micrograph Condition Three epidermis showing �-layer of inner generation. Somemature �-cells have electron-dense, flat, or complexly interdigitated cell membranes (arrows) and electron-dense areas (arrowheads).Numerous �-packets (P) are merging in immature cell oriented top left to bottom right. Image (bottom left) of a mature cell apparentlylying “beneath” immature cell is an artifact reflecting extremely complex interdigitation of �-cells seen in the tuatara (cf. Figs. 20, 22and 24). Bar � 1 �m.

Fig. 19. Sphenodon punctatus. Medium-power TEM micrograph Condition Three epidermis showing melanosomes (arrowheads)embedded in homogeneous �-keratin matrix. Some very narrow intercellular spaces are visible (arrows). Cell boundaries areelectron-dense and vary in form from complex (right) to flat (left). Bar � 500 nm.

Fig. 20. Sphenodon punctatus. Medium- to high-power TEM micrograph Condition Three epidermis where outer and innergenerations are artifactually separated. The cell can possibly be identified as the oberhautchen by the few spinulae on �-layer’s outersurface that are highlighted by electron-dense thickening on inner aspect of the cell membrane (arrowheads) (Maderson et al., 1998).The identity of the large, external, electron-dense object (arrow) is unknown, but material lying more laterally is probably remnantinter-cellular cement representing a broken-down desmosome. Bar � 200 nm.


constrictor (Szabo et al., 1973, fig. 6) and Elaphe o.quadrivitta (Maderson, 1985, fig. 6G). Unique fea-tures of living cells beneath the �-layer (Figs. 10–17) distinguish them from suprabasal cells in Con-dition One (Fig. 7).

Neither histological (Figs. 4, 5), nor ultrastruc-tural (Figs. 18–24), criteria permit equation of Con-dition Three of Sphenodon punctatus with a specificsquamate stage. The tripartite tinctorial properties(regions A–C, Fig. 5) of the immature �-layer reflectcytogenetic patterns never before observed in juxta-posed vertical sequence. Condition Three (whichshows cytologies seen from early Stage 4 to lateStage 5 in squamates; Maderson et al., 1998, fig. 1)is a “mid-renewal phase” that would precede thepre-sloughing condition described previously (Mad-erson, 1968, figs. 4, 5).

Condition One (Fig. 2), a morphology never previ-ously recorded in Sphenodon punctatus, is difficult

to interpret. The keratinized outer generation tis-sues are unlike those in Conditions Two (Fig. 3) andThree (Fig. 4). The apparent absence of an �-layer inLM preparations, in contrast to the three layers ofkeratinized �-cells revealed by TEM (Fig. 7), reflecta tissue immaturity that defines a squamate post-shed RS, not a PRS. While ultrastructural identifi-cation of the single suprabasal cell as a presumptive�-cell confirms that the �-layer is still immature, theextreme basal cell flattening characterizes a PRS.These mutually exclusive data imply that �-layerhistogenesis in the tuatara must differ from thatdescribed for squamates, a possibility that requiresreevaluation of previous studies.

On the basis of autopsy samples from two ani-mals, Gabe and St. Girons (1964, p. 12) give aschema of epidermal form in Sphenodon punctatus.In it, the keratinized tissues of the outer generationare divided into three squamous strata in vertical

Fig. 21. Sphenodon punctatus. Medium-power TEM micrograph Condition Three epidermis showing the junction between inner-most aspect of outer generation (top 2/3) and outermost aspect of inner generation (bottom 1/3). Granulation size is similar inelectron-dense and electron-lucent areas of cells of outer generation: their membranes are linear. Although cytoplasm of cells of mature�-layer (�I) of inner generation exactly resembles that in Figures 18–20, here their electron-dense, thickened membranes emphasizelinearity of cell surfaces (cf. Fig. 22). Bar � 500 nm.

Fig. 22. Sphenodon punctatus. Medium-power TEM micrograph Condition Three showing a quite different image of interfacebetween outer (top) and inner (bottom) generations than that shown in Figure 21. Cells closest to upper margin of figure (�) areidentical to those at base of �-layer of outer generation in Conditions One and Two (Figs. 7, 8, 15). Cells at bottom of figure (�) areidentical to those in Figures 18–21 and are clearly part of newly formed �-layer of inner generation, as evidenced by arrowedelectron-dense areas. Cells at center of micrograph (SP) show a mixture of features mentioned, with complex membrane interdigita-tions and incorporated dense granules (double arrowheads). Large arrowheads indicate possible splitting regions. The striking,regional differences between membrane form as shown here and in Figure 21 can be understood more readily by reference to Figure26. Bar � 1 �m.


sequence: such are discernible in accompanying pho-tomicrographs. However, the forms of the 5–6 layersof suprabasal and germinal cells in the schema donot reflect the photomicrographs (loc. cit., Pl. I, figs.1, 2). The latter resemble our Condition Two. Gabeand St. Girons (1964, p. 14) argue that their animalshad shed “recently.” We disagree for three reasons.First: Their photographs show arborized melano-cytes (cf. Fig. 3). Second: Suprabasal cells (their“stratum intermedium [granulosum]”) are said tobe: “…assez pauvres en grains keratohyaline …comparables a celles que signale Goslar chez Na-trix…” (“…somewhat sparsely provided with kerato-hyalin granules … comparable to the situationGoslar has demonstrated in Natrix…”). Goslar(1958, 1964) noted minute, intracellular, dark-stained “keratohyalin granules (KHG)” in a tissuelater identified by TEM as the mesos layer in thissnake (Landmann, 1979). Barely discernible by LM,these organelles are in fact LBs and/or MGs (Figs.10–16). Third: Gabe and St. Girons’ (1964) allusionsto basal mitoses do not necessarily imply a post-shedRS when proliferative activity is diminishing — theycould equally describe its resumption at the onset ofa renewal phase (Maderson, 1985, fig. 10). The pos-sibility that Gabe and St. Girons’ specimens showedour Condition Two is supported by other data.

Neither Gabe and St. Girons (1964) nor Maderson(1968) evaluated allusions to classical reports of tu-atara epidermis in Lange’s (1931) review because ofuncertainties concerning the number of animalsstudied, their unknown cycle status, quality of fixa-tion, lack of photographic record, and nomenclaturalconfusion. All these problems are confounded by athird translation from the original German!

It can now be seen that Lange (1931), whose treat-ment of each topic evaluates data and/or interpreta-tions within and between living reptilian taxa,makes important points. Emphasizing that layeringof mature keratinized tissues characterizes lepido-saurs, he notes that staining of the squamate com-ponents now termed the �-, mesos, and �-layers ismore distinct than that of their homologs in Sphe-nodon punctatus. His remarks in several contextsimply that data known to him for this species werelimited. When describing the squamate oberhaut-chen, he does not mention a tuatara shedding com-plex. His description of undifferentiated, reptiliankeratinocytes implies a familiarity with the formdefined here as the squamate PRS, but he does notallude to such in the tuatara. Rather, he emphasizesthat its complex dermo–epidermal junction, germi-nal, and suprabasal cells more resemble those ofcrocodilians and chelonians than those of squa-

Fig. 23. Sphenodon punctatus. Very high-power TEM micrograph Condition Three showing irregular dense granules embeddedwithin cytoplasm of cell lying between “typical” �- and �-cells in Figure 22. Bar � 100 nm.

Fig. 24. Sphenodon punctatus. Medium- to high-power TEM micrograph Condition Three showing complex membrane of cell lyingbetween �- and �-cells in Figure 22. Electron-dense thickenings along membranes (arrowheads). Desmosomal remnants (arrows) areaccompanied by intercellular lipid lamellae. Bar � 250 nm.

Fig. 25. Sphenodon punctatus. Medium-power TEM micrograph outermost surface of �-layer (“oberhautchen”) in Condition One.Intracellular, electron-dense areas are visible (arrows). Double arrowhead points to a melanosome. Microbial bodies (arrowheads) areoften seen together with electron-dense debris. Bar � 500 nm.


mates. In describing nonsquamate epidermis heuses the term stratum spinosum and, in compara-tive allusions to that of mammals, he speaks of“tonofibrils.” In classical LM the latter denoted des-mosomes, intercellular junctions not fully character-ized until TEM became a research tool 35 yearslater. Because TEM reveals extensive desmosomesamong immature keratinocytes in mammalian stra-tum spinosum (Matoltsy, 1986) we conclude thatLange (1931) knew only of post-shed, or late RSmaterial, in S. punctatus.

We reexamined LM samples from dorsal, ventral,and gular regions of 14 specimens reported previ-ously (Maderson, 1968). Of 12 whose fixation per-mitted analysis of unkeratinized tissues, four werepre-shed, and eight showed “resting stages.” Of thelatter, four post-shed RS morphologies were uniformin all regions sampled and the other four showedslight regional variation. Although LM alone pre-cludes quantitation of �-layer development (cf. Figs.2, 7), living cells lacked PRS morphologies. Supra-basal cells, identifiable by TEM as presumptive�-cells (Figs. 2, 7), were discernible in all specimensand were illustrated in both “post-shedding” (restingstage) and “pre-sloughing” (Stage 6) categories inMaderson (1968).

A summary of “squamate equivalent” cycle stagesknown for Sphenodon punctatus (Table 1) showsthat nine of 20� reported individuals showed re-newal phase histologies. Of six reported here, 50%were in renewal and 50% showed the paradoxicalCondition One. The tuatara is an infrequent shed-der: In nature it sheds once (Hazley, 1982; M.B.Thompson, pers. obs.), and in captivity twice peryear (Mertens, 1971). Allowing for the uncertainnumber of individuals known to Lange (1931), bycomparison with squamates it is surprising that ap-proximately 45% of individuals known in toto shouldshow renewal phase histologies. Three mutuallycompatible possible explanations could be resolvedby studies involving longitudinal skin biopsies ofliving animals of known cycle status. 1) Sphenodonpunctatus has a renewal phase longer than the 14

days documented for several squamates (Maderson,1985). The fact that 33% of Maderson’s (1968) spec-imens were “pre-shed” does not necessarily supportthis possibility. Because samples from different bodyregions were similar, we may conclude that the tu-atara sheds piecemeal following a pan-body, syn-chronized renewal phase, as do many lizards (Mad-erson et al., 1998). 2) The early renewal histologyseen in 20% of 20� specimens might reflect stress.Field-captured squamates kept for a few days beforebiopsy show a disproportionate percentage of earlyrenewal stages (Maderson, pers. obs.). No data areavailable concerning the pre-sacrifice history of an-imals reported by Gabe and St. Girons (1964) andMaderson (1968). 3) The site of origin of the skinsamples reported here might have influenced nor-mal cell dynamics. For the six animals, 50% were inrenewal and the other 50% showed the anomalousCondition One. We describe normal skin just proxi-mal to neogenic scales on regenerating tails. Noth-ing is known of the possible effects of such proximityon normal skin (Alibardi and Maderson, 2003).

�-Keratogenic Cells in Tuatara Epidermis:Lipogenesis, Mucogenesis, andKeratohyalin-Like Granules (KHLGs)

Condition Three shows approximately 20 layersof keratinized cells between the base of the �-layerof the outer generation and the inner generation.This increase over the number present in Condi-tions One and Two shows that the outer genera-tion is complete and mature in Condition Three.The steady increase in thickness from withoutinwards revealed by TEM (Figs. 7–10, 13, 18, 21,22, 24) is one of several “gradients” discernible inthe �-keratogenic cells. In Condition One (Figs.7–9), intra- and extracellular deposits of lipidsand mucus are heaviest in more superficial cells.Although organelles responsible for their deposi-tion (lamellar bodies [LBs] and mucous granules[MGs]), have only been seen in Condition Two(Figs. 10 –17), late maturing, keratinized, outergeneration tissues in Condition Three (Figs. 5,21–24) show less deposited lipid and mucus thando “typical” �-cells (Figs. 7–13). Because lipogenicLBs occur throughout �-keratogenic tissues in thetuatara, a “mesos layer” is not definable by “mesosgranules” (Roth and Jones, 19671970; Landmann,19791986). Hereafter, we use the term “mesos re-gion” to distinguish the most superficial, flattenedcells (Figs. 7–9) from subjacent �-cells.

Most of the late-maturing �-keratogenic tissuescomprise typical �-cells (cf. Figs. 7–13, 21), but oc-casional cells (possibly in the innermost layer [Figs.22, 24]) show tortuous membranes typical of �-cells(Fig. 18), and other ubiquitous features resemble thelatter. The uniformly granular cytoplasm seen inCondition Three (Figs. 21–24) derives from aggrega-tion of keratohyalin-like granules (KHLGs, Alibardi,

TABLE 1. Summary of works available documenting theepidermal morphology of Sphenodon punctatus and the stages

of the squamate shedding cycle described in them

PublicationNumber of

animalsNumber of specimens (n) andshedding stage assignments

Lange (1931) ?????? (?) Resting StageGabe and St.

Girons (1964)Two (2) Stage 2*

Maderson (1968) Twelve (8) Resting Stage; (4) Stage 6Present work Six (3) Resting Stage **;

(2) Stage 2;(1) Stage “4” ***

Symbols: *Stage reassigned in present work (see Discussion);**Unique PRS never before observed in this species; ***A provi-sional assignment for this specimen (see Discussion).


1998b) around tonofilament bundles and �-keratinpackets. Such mixed �- and �-keratogenic capacitiesalso occur in the most superficial of the subjacent“�-keratogenic tissues.”

�-Keratogenic Cells in Tuatara Epidermis:�-Layer and the Shedding Complex

Persistent cell membranes distinguish the mature�-layer from that of squamates, and absence of syn-cytiality may explain the unique cytology in Condi-tion Three (Figs. 5, 18). In contrast to squamates,the boundaries of all �-keratogenic cells, not justthat of the outer-facing oberhautchen membrane(Maderson et al., 1998), are thickened (Figs. 18, 19,21–23). The deposits of electron-dense materials, ofunknown chemical nature, decrease as cells mature(Fig. 18), implying involvement in the reduction andloss of membrane tortuosity during flattening.

The basis for the toluidinophilia and basophilia ofdifferentiating �-cells, properties mostly lost at ma-turity, is uncertain, although copious ribosomes maybe partly responsible. Certainly, the mature tissue isnever as chromophobic as is that in squamates. Sim-ilarly, the basis for the replacement of toluidino-philia by eosinophilia during �-layer maturation isunclear, although in this respect Sphenodon punc-tatus resembles squamates.

Less discrete than in squamates, the distinctionbetween tissues of the innermost outer generationand outermost inner generation varies regionally(cf. Figs. 21, 22). Because cells show mixed �- and�-keratogenic features, the �-3 �-“switch” is clearlyless precise than in squamates. The absence of therectilinear, i.e., square, columnar, or rectangular insection, cellular morphologies that characterize thecells of the clear layer and oberhautchen of squa-mates precludes identification of a shedding complex(Maderson, 1985, figs. 6G–K, 7A–E; Maderson et al.,1998). The sparse, irregular microstructure seenwith TEM at the �-layer surface in Sphenodon punc-tatus supports Peterson’s (1984) SEM observationsand raises important functional questions.

Cellular Mechanisms of Generation Lossin Lepidosaurs

Reviewing squamate skin shedding, Maderson et al.(1998) noted that loss of mature epidermal tissues inlarge flakes, or sheets, characterizes lissamphibiansand reptiles, in contrast to insensible desquamation inendotherms. Weakening of desmosomal connectionsbetween living cells during maturation characterizesboth modes of loss. The existence of mature, deadtissues at the body surface depends on a stratifiedsquamous form augmented by adhesive properties ofresidual desmosomal proteins and possibly extracellu-lar lipids. Epidermal replacement is needed to: 1) in-crease the surface area of the body to accommodate

somatic growth (all vertebrates); 2) repair abrasivedamage caused by environmental contact (all terres-trial tetrapods); and 3) maintain the barrier to CWL(most amniotes). A pelage (mammals), or plumage(birds), provides the physical protection necessary forboth the organism, and the barrier tissues, but�-keratinized tissues covering reptilian scales serveboth these roles (Maderson et al., 1998). In lepido-saurs, scale overlap also provides integumentary flex-ibility so that, in this clade, the sauropsid synapomor-phy of vertical alternation of �- and �-keratogenesismanifests in periodic, pan-body skin shedding (Mad-erson and Alibardi, 2000).

Protection of immature, squamate �-keratogenic,and mesos tissues relies on the functional outer gen-eration’s being held in place by the three-dimensional complexity of the integument (Mader-son et al., 1998, p. 17). Separation of the twocomponents of the shedding complex involves a pre-cise sequence of behavioral, tissue, cellular, and mo-lecular events. The review noted the absence of ashedding complex in Sphenodon punctatus. How-ever, beyond noting that synchronized sheddingcould occur in the absence of morphological special-ization (Maderson et al., 1998, pp. 20–21), the shed-ding mechanism could not be addressed. New TEMdata, correlated with SEM observations (Peterson,1984; Maderson and Pauwels, pers. obs.) and newdata concerning KHLGs (Alibardi, 1999a, 2000a,b,2001b; Alibardi et al., 2000) permit further com-ment.

Occasionally seen in the innermost �-keratogenictissues (Figs. 22, 24), deposits of electron-dense ma-terials associated with exceptional membrane tortu-osity (Fig. 18) are unique features of immature�-keratogenic cells in Sphenodon punctatus. Bothfeatures are lost during final maturation (Figs. 18,21). The electron-dense deposits resemble the “mar-ginal layer” (“cornified envelope”) of mammaliankeratinocytes (Polakowska and Goldsmith, 1991).This may imply that mechanisms underlying flat-tening of �-keratogenic cells in S. punctatus moreresemble those of more primitive �-keratogenic cells(Maderson and Alibardi, 2000) and could explain thesimilarity between the surfaces of the tuatara’s ma-ture �-layer and mammalian stratum corneum(Maderson et al., 1998, p. 20). Comparison of Peter-son’s (1984) figures with SEMs of squamates (Mad-erson and Pauwels, pers. obs.) reveals that the tu-atara oberhautchen, bearing only occasional pits,lacks impressions of overlying clear layer cells (Irishet al., 1988). Lacking the rectilinear form of juxta-posed, squamate, oberhautchen cells, the exception-ally convoluted margins of homologous cells in thetuatara overlap. They lack consistent orientation:“Where a shift in orientation occurs, the cell marginslie at right angles to each other and interdigitate ir-regularly.” (Peterson, 1984, p. 43, emphasis added).These data facilitate interpretation of our TEMs.


Figure 26 shows juxtaposed, immature genera-tions in Sphenodon punctatus. Tortuosity of �-cellmembranes decreases during maturation so thatTEM shows flat surfaces (Figs. 18, 21) with occa-sional convoluted profiles (Fig. 22). In squamates,TEM reveals vertically oriented membranes inclear layer and oberhautchen (Maderson et al.,1972, figs. 9 –11; Landmann, 1979, figs. 23, 30;Alibardi and Thompson, 1999a, figs. 12, 13, 23,24), but comparable sections in S. punctatus showtheir oblique orientation. In the absence of regu-larly spaced interdigitations (serrations, spines,

setae, etc. forming squamate microornamentation[Maderson et al., 1998]), loss of desmosomal adhe-sion during final cell maturation separates adja-cent units along a splitting zone (Fig. 26). Syn-chronized in some unknown fashion, separationprobably results from a specific sequence of eventsas in squamates (Maderson et al., 1998). While nodescriptions of tuataras removing flakes of outerepidermal generation from the body are available,such could be torn away, so that previously ap-posed oberhautchen cells assume a tiled appear-ance (Peterson, 1984, fig. 1E).

Fig. 26. Schema of interface between outer and inner generations in tuatara towards end of renewal phase. Figures 18, 21, and 22(rectangles) were photocopied to standard magnification and cell profiles therein were drawn as solid lines. Other cell profiles (dottedlines) were reconstructed from other TEMs. Cells in upper half of figure show mixed �- and �-keratogenic features: topographicallythey correspond to the lacunar and clear layer components of the outer generation and the oberhautchen and most superficial cells of�-layer of the inner generation (see Fig. 27). A shedding complex comparable to of squamates is not discernible because cells of clearlayer (clo1, clo2) and oberhautchen (obi1, obi2, obi3) are fusiform and overlap adjacent cells. A potential “splitting zone” within acomplex cannot be recognized in absence of regularly spaced interdigitations between membranes of clear layer and oberhautchencells. Separation of the old, outer generation from the new, underlying, inner generation must be effected by desmosomal breakdownalong line of cell–cell contact (broad arrows). “Irregular tiled” appearance of �-layer surface seen in SEM (Peterson, 1984) may reflectsuch breakdowns along distal interfaces between oberhautchen cells (small arrows).


Fig. 27. Similarities and differences in mature generation structure among three taxa of living lepidosaurs: lack of relevant dataprecludes inclusion of the amphisbaenia. Features unique to Sphenodon punctatus (dark backgrounds); features shared between taxa(medium backgrounds and faint grid-lines between cells); features unique to “lizards” and/or snakes (light backgrounds). Explanatorynotes appearing in cells: 1�-Keratin persists in electron-dense areas of �-layer (Alibardi, 2000a; Alibardi et al., 2000). 2Horizontal levelof cells or tissues being discussed — columns towards right are progressively deeper. 3Phase/stage of squamate cycle (Maderson et al.,1998) when first cells of indicated tissues can be discerned. 4Phase/stage of cycle when they are morphologically and functionallymature. 5“Post-shed �” — this component is revealed as old shed is removed from body surface. 6“Post-shed ��” — final maturationof this tissue, probably involving dehydration of cells, occurs approximately 10–20 h after shedding (Maderson et al., 1998). 7Tuataraloses a unit homologous with squamate generation that has “outer” and “inner” surfaces but, because shedding complex is notdiscernible, the most superficial and the deepest surfaces can only be defined respectively as “oberhautchen” or “clear/granular layer”by their topographic position. 8Many possible cytologies exist in various “lizard” genera ranging from that in desert iguanids(Maderson et al., 1970b) which somewhat resembles that of S. punctatus, and a “typically snake-like” lacunar tissue, e.g., in Anoliscarolinensis (Maderson and Licht, 1967). 9Data for snake microornamentation (MO) (Maderson and Pauwels, pers. obs.). 10Sea snakes(Hydrophidae) have a stratified squamous �-layer, a derived condition (Maderson, 1985) 11Mesos layer, with characteristic mesosgranules, was originally defined in snake epidermis (Roth and Jones, 1967). Previous assumptions that a homologous tissue exists inall lepidosaurs negated by fact that LBs and derived extracellular lipid lamellae may be distributed throughout �-keratogenic tissuesin tuatara and many lizards (Alibardi, 2000b; Dujsebayeva, van Wyk, pers. obs.). In future, the terms “mesos layer” and “mesosgranules” should be used only in descriptions of snake epidermis. For other clades, we recommend that very flattened cells lyingbeneath �-layer and above “typical” �-cells should be termed the “mesos region” with included lipogenic organelles recognized as LBs.12This mucogenicity does not involve distinct granules. Data compiled from various sources in Literature Cited.

Form and Evolution of LepidosaurianEpidermal Generations

Tuatara generation structure differs from that ofsquamates (Fig. 27) in that all the components havea stratified squamous epithelial form reflected in theabsence of �-layer syncytiality and a shedding com-plex, while �-keratin packets, LBs, MGs, andKHLGs are not restricted to specific cell popula-tions. In toto, these differences reflect the lack ofprecision in transitions from �- 3 �- and �- 3�-keratogenesis long thought to characterize lepido-saurs (Baden and Maderson, 1970). They do notmerely suggest a possible primitive condition for theclade; they also elucidate the evolution of amniotekeratinization and its functional correlates.

Discussions of soft-tissue evolution involving com-ment on the adaptive significance of character statesrequire an explicit phylogeny and an understandingof form and function. In these contexts, any attemptto explain the origin and evolution of the squamateshedding complex, and barrier to CWL, faces prob-

lems similar to those pertaining to feathers (Mader-son and Homberger, 2000).

The systematic problem here is that while the statusof the monospecific Sphenodontidae and families of“lizards,” Amphisbaenians, and Ophidians are widelyaccepted, their interrelationships remain uncertain(see references and discussion in Evans, 1995, 1998;Lee, 1997). The database concerning functional mor-phology of lepidosaurian skin is incomplete. Scale di-versity within and between taxa awaits evaluation(Maderson and Alibardi, 2000) and for amphisbae-nians data concern only glands (Jared et al., 1999).This study documents epidermal structure in Spheno-don punctatus but, in contrast to an extensive, albeitdisjunct, literature for several squamates that permitscomparison with other sauropsid amniotes (Lillywhiteand Maderson, 1982,1988; Menon and Menon, 2000),we know nothing of its physiology.

Recent EM studies have extended knowledge ofcytodifferentiation of squamate epidermal genera-tions and biochemical and molecular data concern-ing reptilian keratins are becoming available (Ali-

Fig. 28. Expression, documented or inferred, of 13 epidermal characters (rows 01 3 13, extreme left, boldface) in majorlepidosaurian taxa and antecedent taxa. Lack of relevant data precludes inclusion of amphisbaenia. Shared character states shown incell(s) bounded by heavy margins with taxa (Columns A3 E) separated by light vertical divisions. Darkening of background shadingsto the right across rows (except row 11, Note 23) represents increasingly derived character states relative to basal amniote condition.Explanatory notes and abbreviations appearing in cells: 1Except where indicated, for Column A data see Maderson and Alibardi(2000). 2Except where indicated, for Column B data see Matoltsy (1987), Matoltsy and Huszar (1972), and Matoltsy and Bednarz(1975). 3Amniote keratins reviewed in Sawyer et al. (2000). 4Phyletic distribution amniote keratinization patterns reviewed inMaderson and Alibardi (2000). Vert. Alt., vertical alternation. 5Conceptualization of squamate epidermis as a stratified squamoustissue (e.g., Maderson et al. 1998) emphasizes multicellular origins of its mature components. Here we distinguish actual cytology ofsyncytial squamate �-layer from the more primitive multicellular condition. All Strat.Squam., entire mature epidermis has stratifiedsquamous morphology. 6Epidermal generations, a lepidosaurian synapomorphy (Maderson and Alibardi, 2000). 7N.A., not applicable.8Tissue morphology in locations where transitions between keratogenic pathways occur. 9Cellular basis for skin flaking fromnon-lepidosaurian, reptilian and avian scutate/scutellate scales is poorly known (Maderson et al., 1998: 19). 10S-C, shedding complex.11MO, microornamentation (Maderson et al., 1998). 12It is doubtful whether microarchitecture of tuatara skin surface should bedescribed as MO (Maderson and Pauwels, pers. obs.). 13Whatever the form, size, shape, density, and pattern of constituent units ofsquamate MO, their formation always involves interdigitation between membranes of the clear/granular layer and oberhautchen(Irish et al., 1988; Maderson et al., 1998; Maderson and Pauwels, pers. obs). 14Present in all amniote �-keratogenic cells, absent from�-keratogenic cells (Landmann, 1986) except in Sphenodon punctatus (present study). 15Occurrence in all keratogenic cells in S.punctatus implies retention of a primitive, amniote character state. 16Restriction to outermost facing membrane of oberhautchen cells(Maderson et al., 1998) implies derived squamate character state. 17HRPs are reviewed by Resing and Dale (1991). 18Mammalianstratum granulosum is classically defined by keratohyalin granules (KHGs), whose basophilia in paraffin histological preparationsdiffers from acidophilic granules associated with sauropsid �-keratogenesis (KH-like granules, KHLGs [Alibardi, 1999a, 2001]).Histidine-rich nature of both granules (loc. cit.) implies presence of some “ancestral molecule” in basal amniotes prior to synapsidorigins (Maderson and Alibardi, 2000) but its organelle basis cannot be known. 19HRPs, widely distributed in innermost �-keratogenictissues (Fig. 27), are associated with KHLG aggregation (Alibardi, 1999a,2001, 2002a). 20See Alibardi (2002a). 21Coexpression oforganelles associated with both �- and �-keratogenesis and total keratinization of these tissues unique to S. punctatus. 22Becausemature lacunar tissue (LTO) varies so much in lizards (Fig. 27, Note 8) it is inappropriate to distinguish character states between themand snakes in this context. In both taxa CLO becomes fully keratinized but organelle origin of HRP differs. 23Tetrapod epidermalmucogenicity not separable into discrete character states. Using lissamphibia as living outgroup, a range of size and/or density of MGs(represented here as XXXXX 3 X) is discernible in amniotes. 24Intra- and extracellular lipids, some lamellar, some not (Menon andMenon, 2000), originating from LBs, are unique to amniotes (Landmann, 1986). 25LBs may have evolved prior to amniote origins(Maderson and Alibardi, 2000). 26There is no way of knowing distribution of LBs and/or their products in extinct taxa. Absence of LBsfrom amniote �-keratogenic cells could be a derived character state or a developmental constraint. 27See Figure 27, Note 11.28Restricted distribution of LBs and extracellular lipids in both boids (Roth and Jones, 1967, 1970) and colubrids (Landmann, 1979)implies a derived character state. 29There is no way of knowing whether the first tetrapods to possess a barrier to CWL wereanthracosaurian amphibians or basal amniotes (Maderson and Alibardi, 2000), although some of the former bore keratinized scales(Malakhov and Dujsebayeva, 2001). 30While superficial �-keratinized tissues protect the barrier in living amniotes (Maderson et al.,1998; Maderson and Alibardi, 2000), we have no way of documenting barrier efficacy in basal sauropsids. 31Although morphology oftuatara barrier tissues resembles that of lizards, no data document their efficacy. 32In lizards, number of cell layers comprising tissuenow definable by TEM as mesos region varies greatly. LM data (Maderson et al., 1970b; Ananjeva, Dujsebayeva, pers. obs.) concerningspecies from xeric habitats may imply a form/function relationship. 33No available data explain functional significance of this derivedcharacter state.


Figure 28.


bardi, 2000a; Alibardi et al., 2000; Sawyer et al.,2000). Most data derive from study of three lizards(Gekko gecko [a gekkonine gekkonid], Anolis caroli-nesis [a polychrotid iguanian], and Podarcis muralis[a lacertid]), and two snakes (Constrictor constrictor[a boid], and Natrix natrix [a colubrid]). Althoughthese five species represent major squamate radia-tions, two caveats are important.

First, because the spinulate oberhautchen ingekkonids and polychrotids, independently derived inthe two clades (Irish et al., 1988), is different thanmicroornamentation in other lizards (Alibardi, 1999b,2000b; Alibardi and Thompson, 1999a) and snakes(Alibardi, 2002a; Maderson and Pauwels, pers. obs.),we must be cautious in making generalizations con-cerning factors involved in membrane interdigitationbetween oberhautchen and clear layer cells (Madersonet al., 1998). Second, the shedding complex is so unlikeany other epidermal specialization that it has auniqueness comparable to that of a feather (Madersonand Alibardi, 2000), or a hair (Maderson, 1972a). Sim-ilarities of the clear layer shared by the five best-studied species do not track other features of their�-keratogenic tissues. In all non-eublepharine (Mad-erson, 1972b) gekkonids, but in only some polychrotids(Maderson, pers. obs.), the superjacent lacunar tissueresembles that of snakes. In Anolis carolinensis (Mad-erson and Licht, 1967), but not in gekkonids (Mader-son, 1966), the similarity includes eosinophil invasionof epidermal tissues during the renewal phase. Suchsimilarities (and others regarding microornamenta-tion [Harvey and Maderson, 1999; Maderson and Pau-wels, pers. obs.]) are homoplasies, not reflecting phyl-etic relationships. Lack of understanding of theirfunctional significance constrains identification of evo-lutionary trends.

Although epidermal generations are a lepidosau-rian synapomorphy (Maderson and Alibardi, 2000),there are differences between lizards and snakes(Fig. 27). Given systematic problems, possible biasesin the database, and unresolved questions concern-ing possible adaptive diversity in lizards inter alia,and snakes, and/or amniote keratinization sui gene-ris, their evaluation demands a broad context.

Consideration of 13 epidermal characters (Chars.01–13, Fig. 28) shows that Sphenodon punctatusshares some character states with basal amniotesand/or basal sauropsids (Chars. 01–03, 07), andsome with squamates. Ill-defined generations, lackof a shedding complex and microornamentation, andmixed �- and �-keratogenic capacities (Chars. 04–06, 09) are primitive for lepidosaurs. However, itsmesos tissues resemble those of lizards (Chars. 10,12). Evolutionary patterns can be tentatively iden-tified when this mosaic of character states is consid-ered in relation to 1) other aspects of vertical alter-nation of keratogenesis in sauropsids (Madersonand Alibardi, 2000), and 2) current knowledge ofvertebrate keratinocyte differentiation (Fig. 29).

Landmann’s (1986) skepticism concerning syncy-tial �-keratogenic tissues in nonsquamates has beenvalidated for two chelonians, a crocodilian and a bird(Alibardi and Thompson, 1999bc2000; Alibardi,2002b), while other studies (Alibardi and Thompson,1999a; Alibardi, 2000b,2001a) confirm syncytialityto be a squamate synapomorphy. Found only in tet-rapods (Matoltsy, 1987), a marginal layer (“cornifiedenvelope”) is a primitive feature of amniote�-keratogenic cells (Char. 07, Fig. 28). It is absentfrom the cells of all other sauropsidan �-keratogenictissues except for the squamate oberhautchen (Fig.28). The presence of this envelope in all such cells inSphenodon punctatus implies that the �-layer inthis species is the most primitive �-keratogenic tis-sue known. Mixed organelles in cells in the �- 3�-transition zone (Chars. 05, 09, Fig. 28) tell us thatlack of a shedding complex, and imprecisely con-trolled vertical alternation of keratogenesis, areprimitive for sauropsids. Coexistence of sparseKHLGs with �-packets in lizard oberhautchen cells(Alibardi, 1999b, p. 263) implies that final refine-ment of the squamate �- 3 �-transition occurredafter a well-defined shedding complex appeared inthe evolution of this clade. These conclusions arerelevant to the evolution of amniote keratinization.

Epidermal Keratinization in AmnioteVertebrates

The rapid advances in our understanding of epi-dermis, especially keratinization and lipids, are ev-idenced by treatments of similar topics in Lyne andShort (1965), Sengel (1976), Bereiter-Hahn et al.(1986), Sawyer (1987), Goldsmith (1991), andChuong (1998). Chronologic review of any one topicshows that advances come first from studies of mam-malian (usually human) cells and comparative datafollow. Such a sequence can produce misunderstand-ing when complex functional issues are involved.With our new understanding of amniote phylogeny(Maderson and Alibardi, 2000, fig. 2), an impliedevolutionary “history” of amniote keratinizationbased on chelonian �-keratogenesis (Matoltsy, 1987)is no longer tenable.

Stratified squamous epidermis of basal amniotescomprised �-keratogenic cells that synthesized la-mellar bodies (LBs) while retaining ancestral, ana-mniote mucogenicity (Maderson and Alibardi, 2000,fig. 2) to some extent (Chars. 11, 12, Fig. 28): we canhave no direct knowledge of its barrier efficacy. Ingross form, the tissue may have resembled that ofliving bufonid anurans, although, of course, LBs (loc.cit.), and histidine-rich protein (HRP) and filaggrinare absent from these lissamphibians (Fig. 29, Ali-bardi, pers. obs.).

Filaggrin derives from the histidine-rich profilag-grin of mammalian keratohyalin granules (KHGs)that also contain a cystine-rich protein (Resing andDale, 1991). It forms the matrix within which 70-nm


Fig. 29. Epidermal cell differentiation in living vertebrates compared to that of lepidosaurian reptiles (highlighted for comparisonwith Figs. 27 and 28). Abbreviations and notes appearing in cells: A, absence of pattern of cell differentiation, organelle or molecule;L, organelles or molecules large and/or extensively distributed; M, organelles of medium size; na, not applicable; P, presence of patternof cell differentiation, organelle or molecule; S, organelles or molecules large and/or extensively distributed; ?, data unavailable.1Mucogenicity, a primitive feature of vertebrate epidermal cells, is seen in parakeratotic buccal epithelia and other “mucousmembranes” associated with alimentary and urogenital openings of all amniotes. Otherwise absent from epidermis of mammals andarchosaurians, MGs have been described in �-keratogenic tissues in turtles and lepidosaurs. 2Marginal layer (�cornified cell envelope)is here identified primarily on basis of TEM study. Biochemical data identifying constituent molecules (involucrin, loricin and smallproline-rich proteins) and role of transglutaminases in their assembly are available only for mammalian tissues (Polakowska andGoldsmith, 1991; Akiyama et al., 2000). 3Free lipid droplets have been reported in amphibian epidermal cells, but their origin isunknown. 4In mammalian hair trichohyalin granules, HRP is much reduced by comparison with KH, but both contain arginine. 5HRPpresence in epidermis of nonsquamate, sauropsid amniotes is usually inferred from similarity between KHLGs in these taxa andlizards (Alibardi, 2001). While autoradiography reveals HRP in suprabasal cells beneath �-keratogenic tissues of chelonian carapace(Alibardi, pers. obs.), patterns of cytodifferentiation associated with vertical alternation of keratogenesis therein remain poorlyunderstood (see Fig. 28, Note 8). 6Although biochemical data are lacking for Sphenodon punctatus, the similarity of its KHLGs to thoseof other sauropsids suggests a relation between the HRP component and the transition from �-3 �-keratogenesis. 7“Cornified” tissuesdescribed in gnathostomatous anamniotes are difficult to define: their strongly mucogenic cells contain 70-nm filaments, but lacklamellar bodies (LBs). Those of lissamphibia possess a marginal layer, but lack HRP or filaggrin (Alibardi, pers. obs.). Provisionally,we restrict the term “�-keratinization” to orthokeratotic tissues in amniotes. 8Presence of a marginal layer in all �-keratogenic cellsin S. punctatus is unique among all such tissues. 9Free lipid droplets, of unknown origin but not derived from LBs, have been foundin �-keratogenic tissues in several sauropsid species. 10Positivity to antibodies against HRP-derived filaggrin and/or TEM evidence ofKHLGs has been reported in oberhautchen cells (Alibardi, 2001, 2002a, pers. obs.; this study). Data compiled from various sources inLiterature Cited.


filaments coalesce, while the cystine-rich proteincontributes to the marginal layer (loc. cit.). Respec-tively, the two proteins provide the strength andhydrophilia of mature mammalian keratinocytes,between which lie lipid lamellae derived from la-mellar bodies (Menon and Menon, 2000). Desmo-somal remnants, aided by the adhesive propertiesof those lipids, facilitate the ordered stacking ofmature squames (Odland, 1991) that ensures thatmammalian stratum corneum forms a flexible sur-face to the body, while simultaneously providingphysiological and mechanical protection. The last-named role is augmented by the pelage (Madersonet al., 1998).

Demonstration of HRPs in epidermis of repre-sentative species of all sauropsids implies thatsuch characterized basal amniotes (Char. 08, Fig.28), but their functions are problematic. Soon af-ter amorphous granules lacking a surroundingmembrane were reported in lizard �-keratogenictissues, Maderson et al. (1972, p. 196) noted thattheir acidophilia contrasted with the basophilia ofmammalian KHGs. This distinction was confirmed(Maderson, 1985; Alibardi, 1995) and the acronymKHLGs (keratohyalin-like granules) adopted (Ali-bardi, 1998b). The presence of KHLGs in lizardclear layer suggested they were involved in theformation of microornamentation (MO) by facili-tating interdigitation with oberhautchen cells(Alibardi, 1999a,b,2000b, 2001b). This does notexplain: 1) KHLG presence in normal lacunar cellsin some species (Maderson et al., 1970b), or theirenhanced expression therein in vitro (Flaxman etal., 1968) or posttrauma (Maderson, 1985); 2) arange of sizes of KHLGs that does not correlatewith MO size or shape (Maderson and Pauwels,pers. obs.); 3) the absence of KHLGs from snakesthat have MO (Alibardi, 2002a); 4) the presence ofKHLGs in Sphenodon punctatus that lacks MO.Available data on sauropsid KHLGs derive fromTEM, histochemical, and autoradiographic stud-ies. Although future biochemical analysis mayshow them to be as heterogeneous as are mamma-lian KHGs (Resing and Dale, 1991), for thepresent we can draw some tentative conclusions.

Keratinization and the Evolution ofEpidermal Barrier Function in Amniotes

In the stratified squamous epidermis of basal am-niotes (Column A, Fig. 28), there must have been amolecule ancestral to some HRP component of bothKHGs and KHLGs (? granular, see data for snakes[Alibardi, 2002a]). Because �-keratogenesis in mostliving reptiles involves mucous granules, and oftenonly modest expression of intercellular lipid lamel-lae, the ancestral tissue may have resembled am-niote parakeratotic buccal epithelia (Fig. 29). What-ever its barrier efficacy, its delicate structure wouldhave constrained invasion of an abrasive terrestrial

environment. The theropsidan, and sauropsidan,amniote lineages (Maderson and Alibardi, 2000, fig.2), employed different strategies to protect barriertissues (Maderson et al., 1998; Maderson, 2002).

At some point in theropsidan evolution, a tough,flexible stratum corneum resulted from involvement ofKHGs in �-keratogenesis (Maderson, 1972a,2002).Barrier efficacy was enhanced by continuous renewalof stacked cells with copious extracellular lipid lamel-lae: the barrier tissues were further protected when apelage evolved (Maderson, 2002). There are differ-ences in epidermal HRP metabolism between the “na-ked skin” of Homo sapiens and hirsute mammalianspecies (Resing and Dale, 1991). Because such mightbe explicable in the context of differing modes of pro-tecting barrier tissues, they warrant further investi-gation, although, clearly, the hominid condition is sec-ondarily derived.

The evolutionary origin of the sauropsid strategyto cope with a terrestrial environment (Madersonand Alibardi, 2000) is elucidated by the expressionof vertical alternation of keratogenesis in Spheno-don punctatus, whose �-keratogenic tissues are themost primitive yet described. Because KHLGs wereseemingly associated with �-keratinization in basalsauropsids (Char. 08, Fig. 28), at first sight theircoexpression with �-keratin packets in S. punctatus(Char. 09, Fig. 28) is puzzling. However, both or-ganelles occur in occasional cells (Figs. 21, 22) in theinnermost �-keratogenic generation tissues (Fig.27), a region where fewer mucous granules (MGs)and lamellar bodies (LBs) occur than in “typical”�-cells. While no data suggest a causal relationshipbetween the apparently simultaneous evolutionaryorigin of KHLGs and the new �-keratinaceous pro-tein (Maderson and Alibardi, 2000), the expressionof both precludes the presence of MGs and LBs(Fig. 29).

Does the imprecise vertical alternation of keratogen-esis in tuatara epidermis reflect the basal sauropsidcondition? This question cannot be answered withoutdata concerning temporal aspects of epidermal cyclingin this relic species: relevant information may emergefrom an ongoing study of cytodifferentiation in chelo-nian carapace involving longitudinal biopsies (Ali-bardi, pers. obs.). With the evolutionary origin of thesquamate shedding complex, many as-yet unexplainedchanges occurred in KHLG morphology, presumablyimplying altered patterns of HRP metabolism. VariedKHLG expression in innermost �-keratogenic tissuesin lizards, and absence therefrom in snakes (Figs. 27,28), offer a variety of models for study, the results ofwhich could elucidate the role of “sauropsid” HRP insquamate epidermal generations and perhaps en-hance our understanding of its role in the evolutionaryorigin of �-keratinization.

The sharp morphological transition visible withintissues associated with the �- 3 �-transition in lepi-dosaurs (Char. 10, Fig. 28) is due in part to the absenceof MGs and LBs from cells where 30 nm keratin fila-


ments predominate. While syncytiality of the maturesquamate �-layer emphasizes this discontinuity, TEMhas revealed a similarly well-defined inner boundaryof �-keratogenic tissues in chelonians and crocodilians(Alibardi and Thompson, 1999b,c, 2001), avian scutatescales (Sawyer et al., 2000), even feathers (Madersonand Alibardi, 2000). In all lepidosaurs, especiallysnakes, there is extreme flattening of subjacent�-keratogenic cells and thickened membranes arecharacteristic. Marginal layer retention in mature, tu-atara, �-keratogenic cells is the only cytological fea-ture shared with cells of the mesos region (Sphenodonpunctatus and lizards) or mesos layer (snakes). Be-cause the latter are the �-keratogenic tissues whereboth MGs and LBs occur (Fig. 27), similarities anddifferences in cells and organelles may be related tothe evolution of the barrier to CWL (Chars. 11–13, Fig.28).

In the 25 years since the function of cell/molecular, lipid/protein complexes in amniote epi-dermis as barriers to CWL became known, descrip-tive and experimental studies have revealeddifferences in barrier tissue morphology and rates ofCWL in mammals, birds, and squamate reptiles(Menon and Menon, 2000; Tu et al., 2002). However,we may not state that a particular organization ofany tissue is functionally inferior, or superior, tothat of another — both are the products of naturalselection that equip a particular species for life in itsnative habitat. In the context of lepidosaurian evo-lution, it makes little difference that we can neverknow the form, or efficacy, of barrier tissues in ex-tinct taxa (Fig. 28), or that CWL data are unavail-able for Sphenodon punctatus. The fact that thetuatara, and all lizards for which such data areavailable, have a mesos region, in contrast to a dis-creet mesos layer in both “primitive” (boid [Roth andJones, 1967, 1970]) and “advanced” (colubrid [Land-mann, 1979, 1986]) snakes, might be significant inanother context. Extracellular lipid lamellae inmammalian stratum corneum contribute to cell ad-hesion, so that they perform waterproofing and me-chanical functions simultaneously. Restriction ofLBs and their products to the snake mesos layermay similarly contribute to waterproofing while en-hancing flexibility of individual scales. This possibil-ity could be verified or refuted by examining epider-mal structure in serpentiform lizards.

Understanding the evolutionary origins of soft-tissue components is inherently problematic whereappropriate out-group comparison is limited. Thisstudy elucidates the origin of squamate mesos tis-sues by revealing the similarity in such betweentuatara and lizards. The possibility that tuatara’slack of a shedding complex is causally related to lackof precision in the �-3 �-keratogenic transition canbe explored further. Whether a true oberhautchen— a superficial layer of �-keratogenic cells with mi-croornamentation formed by membrane interdigita-tion — exists in all lacertilians awaits EM confirma-

tion (Maderson and Pauwels, pers. obs.). Madersonet al. (1998) showed that useful cytological informa-tion can derive from TEM study of shed skins andsuch material can yield biochemical data (Thorpeand Giddings, 1983). Such future studies could re-inforce the clues to the origins of sauropsid�-keratogenesis provided here, but information con-cerning temporal aspects of epidermal cell cycling inall such tissues is urgently needed.

ACKNOWLEDGMENTS

Dr. Alibardi’s fellowship to stay in New Zealandcame from the NZ University Grant Committee inexchange with the Italian MAE. He thanks Dr. V.B.Meyer-Rochow (University of Waikato, Hamilton,NZ) for hosting him and for his important role ingetting a permit to work on the tuatara. We thankDrs. M. Thompson (University of Sydney, NSW,Australia), A. Cree (Victoria University, Wellington,NZ), and Mr. J. Govey (NZ Department of Conser-vation), who made possible the collection of speci-mens during a visit to Stephen’s Island. Mrs. Luci-ana Dipietrangelo’s photographic skills are greatlyappreciated. The authors thank Dr. Susan Evans(University College, London) and Dr. George Rogers(University of Adelaide, Australia) for comments onan earlier draft of the manuscript.

LITERATURE CITED

Akiyama M, Smith LT, Shimuzu H. 2000. Expression of trans-glutaminase activity in developing human epidermis. Br J Der-matol 142:223–225.

Alibardi L. 1992. Glial cell composition and ultrastructure of thecaudal spinal cord of young and adult tuataras Sphenodonpunctatus. Acta Zool 73:157–162.

Alibardi L. 1995. Electron microscopic analysis of the regenerat-ing scales in lizard. Boll Zool 62:109–120.

Alibardi L. 1998a. Glycogen distribution in relation to epidermalcell differentiation during embryonic scale morphogenesis inthe lizard Anolis lineatopus. Acta Zool 79:91–100.

Alibardi L. 1998b. Differentiation of the epidermis during scaleformation in embryos of lizard. J Anat 192:173–186.

Alibardi L. 1999a. Keratohyalin-like granules in embryos andregenerating epidermis of lizards and Sphenodon punctatus(Reptilia, Lepidosauria). Amphiba-Reptilia 20:11–23.

Alibardi L. 1999b. Formation of large microornamentations indeveloping scales of agamine lizards. J Morphol 240:251–266.

Alibardi L. 2000a. Ultrastructural localization of alpha keratinsin the regenerating epidermis of the lizard Podarcis muralisduring formation of the shedding layer. Tissue Cell 32:153–162.

Alibardi L. 2000b. Epidermal structure of normal and regenerat-ing skin of the agamine lizard Physignathus lesueurii (McCoy,1878) with emphasis on the formation of the shedding layer.Ann Sci Nat Zool (Paris) 21:51–61.

Alibardi L. 2001. Keratohyalin-like granules in lizard epidermis:evidence from cytochemical, autoradiographic and microanalyt-ical studies. J Morphol 284:64–79.

Alibardi L. 2002a. Ultrastructure of the embryonic snake skinand putative role of histidine in the differentiation of the shed-ding complex. J Morphol 251:149–168.

Alibardi L. 2002b. Keratinization and lipogenesis in the embry-onic epidermis of the zebrafinch Taeniopygia guttata castanotis(Aves, Passeriformes, Ploecidae). J Morphol 251:294–308.


Alibardi L, Maderson PFA. 2003. Observations on the histochem-istry and ultrastructure of the regenerating caudal epidermis ofthe tuatara Sphenodon punctatus (Sphenodontida, Lepidosau-ria, Reptilia). J Morphol .

Alibardi L, Thompson MB. 1999a. Epidermal differentiation inthe developing scales of embryos of the Australian scincid lizardLampropholis guichenoti. J Morphol 241:139–152.

Alibardi L, Thompson MB. 1999b. Epidermal differentiation dur-ing carapace and plastron formation in the embryonic turtleEmydura macquarii. J Anat 194:531–545.

Alibardi L, Thompson MB. 1999c. Morphogenesis of shell andscutes in the turtle Emydura macquarii. Aust J Zool 47:245–260.

Alibardi L, Thompson MB. 2001. Fine structure of the developingepidermis in the embryo of the American alligator (Alligatormississippiensis, Crocodilia, Reptilia). J Anat 198:265–282.

Alibardi L, Maurizii MG, Taddei C. 2000. Immunocytochemicaland electrophoretic distribution of cytokeratins in the regener-ating epidermis of the lizard Podarcis muralis. J Morphol 246:179–191.

Baden HP, Maderson PFA. 1970. Morphological and biophysicalidentification of fibrous proteins in the amniote epidermis. JExp Zool 174:225–253.

Bereiter-Hahn J, Matoltsy AG, Richards KS (eds). 1986. Biologyof the integument. 2. Vertebrates. Berlin: Springer.

Chuong CM (ed). 1998. Molecular basis of epithelial appendagemorphogenesis. Austin, TX: RG Landes.

Evans SE. 1995. Lizards: evolution, early radiation and biogeog-raphy. In: Sun A, Wang Y, editors. 6th Symposium MesozoicTerrestrial Ecosystems and Biota. Beijing: China Ocean Press.p 51–55.

Evans SE. 1998. Lepidosaurian faunas from the early Creta-ceous: a clade in transition. In: Lucas SG, Kirkland JI, EstepJW, editors. Lower and Middle Cretaceous terrestrial ecosys-tems. New Mexico Mus Nat Hist Bull 14:195–200.

Flaxman BA, Maderson PFA, Szabo G, Roth SI. 1968. Control ofcell differentiation of lizard epidermis in vitro. Dev Biol 18:354–374.

Gabe M, St. Girons M. 1964. Tegument. Contribution al’Histologie de Sphenodon punctatus Gray. Editions du CentreNationale de la Recherche Scientifique, Paris 1–14.

Goldsmith LA (ed). 1991. Physiology, biochemistry and molecularbiology of the skin, 2nd ed. New York: Oxford University Press.

Goslar HG. 1958. Beitrage zum Hautungsvorgang der Schlangen1. Mitteilung: Histologische und topochemische Untersuchun-gen an der Haut von Natrix natrix L. wahrend der Phasen desnormalen Hautungszyklus. Acta Histochem 5:182–212.

Goslar HG. 1964. Beitrage zum Hautungsvorgang der Schlangen2. Mitteilung: Studien zur Fermentopochemie der Keratogen-ese und Keratolyse am Modell der Reptilienhaut. Acta Histo-chem 17:1–60.

Harvey HB, Maderson PFA. 1999. Homoplasies as inadequatelydescribed character states: a case study involving squamateskin. Am Zool 38:203A.

Hazley LC. 1982. Tuatara. Invercargill, NZ: Craig.Irish FJ, Williams EE, Seling E. 1988. Scanning electron micros-

copy of changes in epidermal structure occurring during theshedding cycle in squamate reptiles. J Morphol 197:105–126.

Jared C, Antoniazzi MM, Silva JRMC, Freymuller E. 1999. Epi-dermal glands in squamate: SEM of precloacal glands in Am-phisbaena alba (Amphisbaenia, Amphisbaenidae). J Morphol241:197–206.

Landmann L. 1979. Keratin formation and barrier mechanismsin the epidermis of Natrix natrix (Reptilia, Sepentes): an ultra-structural study. J Morphol 162:93–126.

Landmann L. l986. Reptilian skin. In: Bereiter-Hahn J, MatoltsyAG, Richards KS, editors. Biology of the integument, vol. 2.Vertebrates. Berlin: Springer. p 150–187.

Lange B. 1931. Integument der Sauropsiden. Handb vergl Anat dWirbelthiere 1:375–488.

Lavker RM. 1974. Horny cell formation in the epidermis of Ranapipiens. J Morphol 142:365–378.

Lee MSY. 1997. The phylogeny of varanoid lizards and the affin-ities of snakes. Philos Trans R Soc 352B:53–92.

Lillywhite HB, Maderson PFA. 1982. Skin structure and perme-ability. In: Gans C, Pough HF, editors. Biology of the Reptilia,vol 12. New York: Academic Press. p 397–442.

Lillywhite HB, Maderson PFA. 1988. The structure and perme-ability of the integument. Am Zool 28:945–962.

Lyne AG, Short BF (eds). 1965. Biology of the skin and hairgrowth. Sydney: Angus and Robertson.

Maderson PFA. 1964. The skin of snakes and lizards. Br J Her-petol 3:151–154.

Maderson PFA. 1965. Histological changes in the epidermis ofsnakes during the sloughing cycle. J Zool (Lond) 146:98–113.

Maderson PFA. 1966. Histological changes in the epidermis of theTokay (Gekko gecko) during the sloughing cycle. J Morphol119:39–50.

Maderson PFA. 1968. Observations on the epidermis of the tua-tara (Sphenodon punctatus). J Anat 103:311–320.

Maderson PFA. 1972a. When? Why? and How? Some specula-tions on the evolution of the vertebrate integument. Am Zool12:159–171.

Maderson PFA. 1972b. The structure and evolution of holocrineepidermal glands in sphaerodactyline and eublepharinegekkonid lizards. Copeia 1972:559–571.

Maderson PFA. 1985. Some developmental problems of the rep-tilian integument. In: Gans C, Billett F, Maderson PFA, editors.Biology of the Reptilia, vol 14. New York: John Wiley & Sons.p 523–598.

Maderson PFA. 2002. Mammalian skin evolution: a reevaluation.Exp Dermatol 12:1–5.

Maderson PFA, Alibardi L. 2000. The development of the saurop-sid integument: a contribution to the problem of the origin andevolution of feathers. Am Zool 40:513–529.

Maderson PFA, Homberger DG (eds). 2000. Evolutionary originof feathers — symposium. Am Zool 40:455–707.

Maderson PFA, Licht P. 1967. Epidermal morphology and slough-ing frequency in normal and prolactin treated Anolis carolinesis(Iguanidae, Lacertilia). J Morphol 123:157–172.

Maderson PFA, Chiu KW, Phillips JG. 1970a. Endocrine-epidermal relationships in squamate reptiles. Mem Soc Endo-crinol 18:259–284.

Maderson PFA, Mayhew WW, Sprague G. 1970b. Observationson the epidermis of desert-living iguanids. J Morphol 130:25–36.

Maderson PFA, Flaxman BA, Szabo G, Roth SI. 1972. Ultrastruc-tural contributions to the identification of cell types in thelizard epidermal generation. J Morphol 136:191–210.

Maderson PFA, Rabinowitz T, Tandler B, Alibardi L. 1998. Ul-trastructural contributions to an understanding of the cellularmechanisms involved in lizard skin shedding with comments onthe function and evolution of a unique lepidosaurian phenom-enon. J Morphol 236:1–24.

Malakhov DV, Dujsebayeva TN. 2001. Cornified scalation of Ute-gania shpinari (Seymouriomorpha, Discosauriscidae) and radi-ation of the family during the Paleozoic. Biota 2:157–162.

Matoltsy AG. 1986. Structure and function of the mammalianepidermis. In: Bereiter-Hahn J, Matoltsy AG, Richards KS,editors. Biology of the integument, vol 2. Vertebrates. Berlin:Springer. p 255–271.

Matoltsy AG. 1987. Concluding remarks and future directions.Curr Top Dev Biol 22:255–264.

Matoltsy AG, Bednarz JA. 1975. Lamellar bodies of the turtleepidermis. J Ultrastr Res 53:128–132.

Matoltsy AG, Huszar T. 1972. Keratinization of the reptilianepidermis: an ultrastructural study of the turtle skin. J Ultra-str Res 38:87–101.

Menon GK, Ghadially R. 1997. Morphology of lipid alterations inthe epidermis: a review. Microsc Res Tech 37:180–192.

Menon GK, Menon J. 2000. Avian epidermal lipids: functionalconsiderations and relation to feathering. Am Zool 40:540–552.

Menon, GK, Ghadially R, Williams ML, Elias PM. 1992. Lamellarbodies as delivery systems of hydrolytic enzymes: implications


for normal and abnormal desquamation. Br J Dermatol 126:337–345.

Menon GK, Maderson PFA, Drewes RC, Baptista LF, Price LF,Elias PM. 1996. Ultrastructural organization of avian stratumcorneum lipids as the basis for facultative cutaneous water-proofing. J Morphol 227:1–13.

Mertens R. 1971. Bemerkungen uber die Hautung der Bruck-enechse (Sphenodon punctatus) in Gefangenschaft. Salaman-dra 7:81–82.

Odland GF. 1991. Structure of the skin. In: Goldsmith LA, editor.Physiology, biochemistry and molecular biology of the skin.Oxford: Oxford University Press. p 3–62.

Parrakal PF, Matoltsy AG. 1964. A study of the fine structure ofthe epidermis of Rana pipiens. J Cell Biol 20:85–94.

Peterson JA. 1984. The scale microarchitecture of Sphenodonpunctatus. J Herpetol 18:40–47.

Polakowska RR, Goldsmith LA. 1991. The cell envelope andtransglutaminases. In: Goldsmith LA, editor. Physiology, bio-chemistry and molecular biology of the skin. Oxford: OxfordUniversity Press. p 168–204.

Resing KA, Dale BA. 1991. Proteins of keratohyalin. In: Gold-smith LA, editor. Physiology, biochemistry and molecular biol-ogy of the skin. Oxford: Oxford University Press. p 148–167.

Roth SI, Jones WA. 1967. The ultrastructure and enzymaticactivity of the boa constrictor (Constrictor constrictor) skin dur-ing the resting phase. J Ultrastr Res 32:304–323.

Roth SI, Jones WA. 1970. The ultrastructure of epidermal mat-uration in the skin of the boa constrictor (Constrictor constric-tor). J Ultrastr Res 32:69–93.

Sawyer RH (ed). 1987. The molecular and developmental biologyof keratins. Curr Top Dev Biol 22:1–275.

Sawyer RH, Glenn T, French JO, Mays B, Shames RB, BarnesGL, Ishikawa Y. 2000. The expression of �-keratins in theepidermal appendages of reptiles and birds. Am Zool. 40:530–539.

Sengel P. 1976. Morphogenesis of skin. New York: CambridgeUniversity Press.

Szabo G, Maderson PFA, Roth SI, Speer RM. 1973. Melanocyteactivity in the epidermis of the boa constrictor Constrictor con-strictor during the sloughing cycle. Anat Rec 176:377–388

Thompson MB. 1990. Incubation of eggs of tuatara, Sphenodonpunctatus. J Zool (Lond) 222:303–318.

Thorpe RS, Goodings MR. 1983. The biochemistry of keratins ingreen geckos (Phelsuma) and its systematic value. Herpeto-logica 39:399–410.

Troyer H. 1980. Principles and techniques of histochemistry.Boston: Little Brown.

Tu MC, Lillywhite HB, Menon JG, Menon GK. 2002. Post natalecdysis establishes permeability barrier in snake skin: newinsights into barrier lipid structures. J Exp Zoo 205:3019–3030.

Zucker AH. 1977. Cutaneous water loss in reptiles. UnpublishedPh.D. Thesis, City University of New York.


Documents

Observations on the Histochemistry and Ultrastructure of ...lemastm/Publications/PDF/Snake Skin/Alibardi_et_al-2003-Journal_of... · dermis of Sphenodon punctatus conﬁrms previous