For a better understanding of the nasal cavity of alligators, cryosections of the nasal cavity were stained with Kernechtrot-Lichtgrün-Orange (KLO) (nuclear red-light green-orange G) and semithin sections were stained with toluidin blue 33 . To describe the olfactory epithelium macroscopically, the criteria commonly applied to olfactory epithelium were used. The receptor neurons are categorized according to their height and structure within the epithelium: tall cells with their nuclei in the lower portion of the epithelium and long slender dendrites are considered ciliated OSNs. Intermediate cells with their nuclei in the middle portion of the epithelium and short thick dendrites are considered microvillous OSNs (this latter cell type did not occur in the alligator epithelium) 34 . For convenience of description the nasal cavity was divided into 5 anterior-posterior zones, numbered 1 to 5 from anterior to posterior (Fig. 1C). The epithelium in each zone varies in thickness between 50 and 150 μm (Fig. 2). As seen in other vertebrates, the epithelium is pseudostratified with long, slender cells (Figs. 2E,F, 3A) that stretch from the basal lamina to the lumen of the nasal cavity. Bowman glands lie beneath the epithelium (Fig. 2B,D,E, 5B) and olfactory axons aggregate above and under the basal lamina (Fig. 4B). The occurrence of OSNs, however, is not limited to either the thin or the thick type of epithelium (see also Immunocytochemistry). As described before, almost the entire nasal cavity is lined with olfactory epithelium 19 . A few areas contain only nonsensory epithelium, mainly limited to the most ventral portion of the cavity and small ventral areas of the turbinates. These cells can be easily distinguished from the OSNs that have larger, more deeply placed nuclei. The vestibule (Fig. 1B), the duct leading from the naris opening down to the nasal cavity proper, is lined with short, flattened epithelial cells similar to those of the adjacent skin. OSNs are not present in the vestibule.

The description of the ultrastructure of the OSNs follows the criteria commonly used for olfactory epithelium. As mentioned above, receptor cells are categorized macroscopically by their height within the epithelium. At the ultrastructural level, cell organelles are used to categorize OSNs, e.g. structural features of the cilia, presence or absence of rootlets, "checker-board" pattern of chromatin in nuclei, longitudinally arranged mitochondria, presence or absence of centrioles 35 36 . Electron micrographs (TEM) show that in the pseudostratified epithelium, the supporting cells are the most prominent (Fig. 3A). Their nuclei lie above the level of nuclei of OSNs. Abundant clear vesicles ca. 0.6 to 1.0 μm in diameter, fill the area above the nucleus and the most apical area leaving a small portion in between for mitochondria and electron-dense cytoplasm (Fig. 3A,B). Occasionally centrioles (usually two) are present in the upper portion of the cell which is devoid of vesicles (not shown). At the surface bordering the lumen, the width of the supporting cells ranges from 2.8 to 4.1 μm. This is only a rough estimate since these cells often envelop the OSNs. Consequently, at the ultrastructural level two supporting cell profiles, one on each side of an OSN, can be the same cell seen twice. The surface of the supporting cells is covered with numerous microvilli, about 4 to 5 μm long and 0.08 to 0.17 μm wide (Figs. 3B, 4C, 5A–C). OSNs are dispersed between the supporting cells (Figs. 3, 4A, 5C,D); the width of their apical dendrites (1 to 1.4 μm) is even smaller than the upper portion of the supporting cells. As seen in the scanning electron microscope (SEM), the pronounced olfactory knob bears about 6 to 7 cilia (10 to 15 μm long, approx. 0.3 μm wide) that do not taper at their end. In addition, the olfactory knobs bear several thin microvilli (about 1 μm long), and thus shorter and smaller than those of the supporting cells (Figs. 3B–D, 4A). Some TEM sections suggest the presence of OSNs with only microvilli. However, serial sections reveal that this is misleading (Figs. 3C, 4A), since the original plane of section may not include all of the cilia. All OSNs seen in serial ultrathin sections were so-called mixed receptor cells (bearing both cilia and microvilli) as described for birds by Graziadei 37 . Nevertheless, based on dendritic morphology, two different types of OSNs do exist. The distribution of these two types of OSNs within the nasal cavity has no obvious pattern. The olfactory knobs of both OSN types contain small light vesicles. Rootlets are absent from cilia of both cell types (Fig. 3B–D). Although both types bear cilia and microvilli on their olfactory knobs, their dendrites show different features. Type I OSNs are relatively electron-lucent and contain small vesicles (70–100 nm) and few microtubules. A few mitochondria are arranged longitudinally (Fig. 3B). Beneath the olfactory knob – often even deep in the dendrite – several centrioles are present (Fig. 4A), a feature that never occurred in the second OSN type. The cytoplasm of the type II OSN is electron-dense and filled with abundant mitochondria that sometimes reach into the olfactory knob. Their cytoplasm is electron-dense (Fig. 3B,C,D). Centrioles are lacking.

Compared to the olfactory epithelium of other groups of animals, the OSNs in the alligator's olfactory epithelium are rather scarce. In alligators, the density of OSNs varies in the different zones of the olfactory epithelium. The highest density occurs on the turbinates [4.6 (+/- 0.34 SE) apical knobs of OSNs per 25 μm length of epithelium, as assessed in ultrathin sections]. By way of comparison: goldfish or mouse olfactory epithelia contain approx. 8 – 10 OSNs per 25 μm length of olfactory epithelium in ultrathin sections (A.H., personal observation). Towards the rostral end of the nasal cavity of alligators, the density declines with the lowest numbers close to the vestibule [1.5 (+/- 0.22 SE) OSNs per 25 μm length of epithelium] (Fig. 6). ANOVA/Tukey tests of these cell countings revealed that zones 1, 2, and 3 were significantly different from zones 4 and 5. The difference between zone 4 and 5 was not significant (df between groups = 4, total = 87, F = 22.029, p < 0.05).

Beneath the layer of OSNs a layer of basal cells occurs. Basal cells rest on the basal lamina, their shape often following the curves of the basal lamina (Fig. 4B). Within the olfactory epithelium a few cells occur that contain cell debris enveloped in large vesicles (Fig. 4D). The nature of these presumably phagocytotic cells is unknown.

Strictly non-olfactory areas of epithelium are confined to the most ventral areas of the nasal cavity and small areas in the folding of the turbinates. These non-olfactory areas are mostly lined with columnar epithelial cells. They bear either tiny microvillous-like protrusions (Fig. 4E) or numerous cilia (Fig. 5E) which unlike cilia present on sensory neurons have pronounced rootlets extending into the cell (arrowheads Fig. 4D). Some of the microvilli-bearing cells are of mucous character (compare 19 ) containing dark, electron-dense vesicles of various sizes. This part of the epithelium is loosely-packed and the wide interstitium is filled with dense extracellular matrix (Fig. 4E).

The antisera that worked in the alligator tissue corroborated the results obtained at the electron microscopic level.

(1) PGP9.5 showed labelled OSNs throughout the nasal cavity (Fig. 7A–B). The highest density of labelled cells was present in the turbinate areas (Fig. 7B), a lesser amount in the more anterior areas (Fig. 7A). Evidently, the height of the nasal epithelium did not correlate with sensory versus non sensory areas. PGP-positive OSNs occurred in areas with thick as well as in areas with thin epithelium. In nonsensory areas, PGP9.5 unspecific label was present in the nonsensory kinocilia (Fig. 7C). Occasionally, an apical ending that appeared to be dome-shaped showed PGP9.5-label (Fig. 7B). These cells may be SCCs (see below).

(2) Gα_olf, a marker for the canonical transduction cascade in olfaction, was present in the apical endings of OSNs (Fig. 8A–B). The label occurred throughout the nasal cavity proving the presence of OSNs in wide areas of the nasal epithelium. The degree of label, however, varied from area to area with the strongest label in the turbinate areas (Fig. 8A) and the weakest label in the anterior areas of the cavity (Fig. 8B) corresponding to PGP 9.5 results showing fewer OSNs in the anterior areas (Fig. 7A). Also as seen in PGP 9.5 experiments, Gα_olf occurred both in thick and in thin epithelia.

(3) YOL 1/34 tubulin (a marker for microtubules) immunoreaction is present in cilia (Fig. 8C) and in fibres coursing below the olfactory epithelium but no positive fibres are seen beneath the nonsensory areas.

Occasionally, cells with the morphology of solitary chemosensory cells (SCCs) occur in the olfactory epithelium. These cells are scarce but stand out in SEM and TEM preparations (Figs. 4C, 5F). Their apical ending is dome-shaped. Supporting cells seem to indent into the upper part of the cell resulting in a "neck" of the presumed SCCs (Fig. 4C). The dome is filled with dense cytoplasm and few tiny vesicles (approx. 0.02 – 0.05 μm). The area below the "neck" on the other hand contains abundant slightly larger vesicles (approx. 0.07 – 0.2 μm).

Immunocytochemical results (synapses in the lower portion of the epithelium, PLCβ2 in apical endings) corroborate the notion that these cells are, indeed, SCCs.

(1) PLCβ2, a marker for the transduction pathway of SCCs and some OSNs 28 38 , labels few apical endings which at the light microscopic level appear to be dome-shaped (Fig. 7E–F).

(2) Substance P labels nerve fibres under the nasal epithelium. Some of the fibres reach into the olfactory and into the nonsensory epithelium (Fig. 8D). Other Substance P-positive fibres are present in the close vicinity of the glands.

The present study of the American alligator confirms Saint Girons' 19 findings insofar as almost the entire nasal cavity is lined with olfactory (= sensory) epithelium. However, the density of OSNs is heterogeneous along the nasal cavity of the American alligator. The highest numbers of OSNs occur in the zones of the two turbinates and the lowest numbers close to the naris. Saint Girons' 19 cell countings in histologically stained sections rendered a proportion of 61% OSNs for Crocodylus and 59% of OSNs for Caiman. These results are very different from what I counted in electron microscopic experiments: In the alligator, the proportion of OSNs versus supporting cells is reversed compared to Saint Girons' countings 19 : the olfactory epithelium contains many more supporting cells than OSNs – even in the turbinate zones where the OSNs have the highest density (Fig. 6). The difficulties in counting supporting cells cannot account for this considerable difference. This could be species-dependent but it is more likely due to the very different techniques. Saint Girons 19 counted nuclei in histologically stained thick sections which is tedious but not very reliable due to the poor resolution and thus difficulties in distinguishing cell types, whereas in the present study cells were counted at the ultrastructural level.

Since alligators hunt above and under water and given the fact that they possess only one nasal compartment the question arises how are volatile and/or water-soluble chemicals detected? A study on the semi-aquatic newt Triturus pyrrhogaster reported a change of morphology and physiological properties of the OSNs depending on the environment. In a terrestrial environment the cilia of the OSNs grew longer than when they were in aquatic surroundings. Slow potentials evoked in the olfactory epithelium changed from responding to odorous fluids in an aqueous medium to responding to odorous vapours in a terrestrial environment 16 . Interestingly, two types of OSNs are present in the olfactory epithelium of the American alligator. Both types bear cilia that do not taper at their ends as seen in rodents. Other than that, the two types of ciliated OSNs show different morphological features that are commonly used to describe OSNs (number of mitochondria, presence/absence of centrioles, electron density of cytoplasm 35 ) so that type I and type II OSNs are easily distinguished at the ultrastructural level. This has not been reported for birds (the second group of the Sauropsida) or other reptiles. What the two different types of OSNs mean in terms of function is yet unknown, but it is tempting to speculate that one type of OSN may be utilized above water for air-borne stimuli while the other is used under water. The results described in this study were collected from juvenile specimens. It remains to be shown whether the findings also apply to fully mature alligators.

Immunocytochemical experiments were hampered by the fact that most of the antisera made for rodents do not work in alligators. Nevertheless, it was possible to show that, as seen in fish and mammals, the OSNs in alligators express Gα_olf, the most common G-protein subunit in olfactory transduction 3 6 7 . The neuronal marker PGP9.5 confirmed the presence of OSNs in various areas. This was especially helpful since it corroborated on a larger scale what ultrastructural experiments had shown: (a) that the density of OSNs varies from the anterior end of the snout to the turbinates, and (b) that the height of the epithelium is not correlated to the presence/absence of OSNs. They occur in thicker as well as in thinner areas of epithelium. These two features were also evident in experiments with the YOL 1/34 antibody, a marker for tubulin in cilia, and Gα_olf.

Another interesting feature of the alligator's nasal epithelium is the presence of SCCs. These chemoreceptor cells had been known for fish and amphibians 24 25 . They occur in the skin of the body of fish, in the oral cavity and on the gills 26 39 . In the alligator, SCCs are integrated into the olfactory epithelium, a feature that up to now has been reported only for the goldfish olfactory epithelium 40 . With some exceptions, their function is not well understood. Recently, SCCs were also described for rodents 28 . In this case, the SCCs are located in the respiratory epithelium of the nasal cavity. In rodents, the SCCs express components of the T2R transduction cascade including gustducin and PLCβ2. Electrophysiological and behavioural experiments led to the conclusion that SCCs in rodents serve as "sentinels" in the anterior nasal cavity guarding against potentially harmful or irritant chemicals. SCCs also occur in guinea pigs 41 and cows 30 . In alligators, dome-shaped apical endings were labelled with PLCβ2 as are nasal SCCs in rodents. In addition, Substance P-positive fibres were present in and under the nasal epithelium of the alligator indicating innervation of the SCCs by trigeminal nerve fibres as seen in rodents. The presence of SCCs in alligators as well as in hagfish 23 , elasmobranchs, teleost fish 25 , amphibians 24 42 , rodents 28 , large mammals 30 , and humans 31 support the notion that SCCs are an evolutionarily conserved trait having evolved early in vertebrate phylogeny.

Three juvenile alligator heads (Fig. 1A) fixed in 4% paraformaldehyde (kindly provided by Dres. Catherine Carr and Daphne Soares, University of Maryland) were split at the midline by means of a fretsaw (Fig. 1B). The complete nasal sacs were dissected out and divided into 5 zones for further processing (Fig. 1C). Two animals were used for transmission electron microscopy and immunocytochemistry, the third animal for scanning electron microscopy and immunocytochemistry.

For an overview at the light microscopic level, cryosections (16 μm) were treated with the histological staining Kernechtrot-Lichtgrün-Orange (KLO) (0.1% nuclear red – C.I. 60760, Merck, Darmstadt – in 5% aluminium sulphate followed by 0.2% light green – C.I. 42095, Allied Chemical Corp., New York, 1% orange G – C.I. 16230, Allied Chemical Corp., New York – and 0.5% phosphotungstic acid in 1% acidic acid) 33 . Briefly, the slides were washed in dH₂O, immersed in a mixture of nuclear red in aluminium sulphate for 10 min, washed in dH₂O, and immersed in a mixture of light green and orange G in phosphotungstic acid for 1 – 2 min. After the staining process, the slides were dehydrated in ethanol and xylene. Epon-Araldite-embedded pieces of tissue were cut at 1 μm and stained with 1% toluidin blue (J. T. Baker Chemical Co., Phillipsburg).

Pieces of tissue of all 5 zones of the nasal cavity were postfixed in 4% glutaraldehyde in 0.05 M phosphate buffer (pH 7.2). After rinsing in phosphate buffer, the specimens were dehydrated in a graded series of acetone and isoamyl acetate, critical-point-dried in CO₂, coated with gold, and examined with a LEO 1525 scanning electron microscope (Carl Zeiss, Jena, Germany).

Pieces of tissue of all 5 zones of the nasal cavity were postfixed in 4% glutaraldehyde in 0.05 M phosphate buffer (pH 7.2) for several hours. After rinsing in phosphate buffer, the tissue samples were postfixed with 1% osmium tetroxide for 2 h. The fixed specimens were dehydrated in a graded series of ethanol and acetone and embedded in Epon-Araldite (Electron Microscopy Sciences, Hatfield, PA). Ultrathin sections (silver to gold) were stained with uranyl acetate and lead citrate and examined with a FEI Tecnai G² electron microscope (Philips, Eindhoven, Netherlands).

Epithelia of all 5 zones were photographed in the transmission electron microscope and printed at a magnification of 4000×, the lowest magnification where it was possible to unequivocally distinguish cell types. Photos of each zone were taken in a single session to ensure that the epithelia were captured only once. Thus, double counting of cells was prevented. On these photos, the apical portion of the epithelium was divided into sections (87 in total) each representing 25 μm of actual tissue. Only OSNs that reached the apical surface were counted. The investigator was blinded as to location of the sample at the time each was counted. Statistical analysis (ANOVA, Tukey test) was done with the program SPSS 15.0 (SPSS Inc., Chicago).

Antisera directed against various markers for OSNs and SCCs were used to test their effectiveness in alligator tissue. These experiments were hindered by the fact that most antisera directed against mammalian proteins did not work in alligators (e.g. NCAM, gustducin, CGRP, Hu-C, calbindin, calretinin, RMRF, OMP, Gα_o). Consequently, it is not possible to tell whether these proteins are present or not by this method. A further difficulty was the high autofluorescence of the tissue. All antisera were used with fluorescence-coupled and biotinylated secondary antibodies. Best results were obtained with antigen retrieval, extensive blocking of endogenous peroxidase and biotinylated secondary antibodies. 16 μm cryosections were processed according to standard methods. Briefly, the sections were subjected to either acidic (30 min in 10 mM citric acid, pH 3.0, at 37 degrees C) or basic (30 min in 0.05 M Tris buffer, pH 8.0, at 80 degrees C) antigen retrieval. After three washes in phosphate-buffered saline (PBS; pH 7.2), the sections were left in blocking solution [1% bovine serum albumin, 5% normal horse serum (NHS), 0.3% Triton X-100] for 2 – 3 hours. Incubation with the primary antisera diluted in blocking solution containing 2% NHS lasted overnight to 3 days. After three washes in PBS, fluorescence-coupled secondary antibodies were used to visualize the antibody labelling. As the autofluorescence of the tissue was usually very high, more immunocytochemical experiments were carried out with biotinylated secondary antibodies. In that case, endogenous peroxidase was quenched with 0.3% H₂O₂ for 30 min. After 3 × 15 min washes in PBS, the blocking solution was applied as described above. Then the slides were incubated in avidin (Blocking Kit, Vector Laboratories Inc., Burlingame, CA) for 15 min, rinsed briefly and incubated in biotin (Blocking Kit, Vector Laboratories Inc., Burlingame, CA) for 15 min. The slides were left in the primary antisera overnight and reacted with the ABC/DAB method according to standard protocols. As a control, various steps (primary antibody, secondary antibody, ABC, DAB) were omitted. The antisera that immunolabeled alligator tissue were: rabbit Protein Gene Product 9.5 (PGP9.5 = Ubiquitin C-terminal Hydrolase, dilution 1:200 – 1:1000, Biogenesis, Kingston, NH), rabbit Gα_olf (dilution 1:1000, St. Cruz Biotechnologies, Santa Cruz, CA), rabbit PLCβ2 (dilution 1:100, St. Cruz, Santa Cruz Biotechnologies, Santa Cruz, CA), rat Substance P (dilution 1:100, Accurate Chemical and Scientific Corp., Westbury, NY), rat YOL 1/34 tubulin (dilution 1:100 – 1:200, Abcam, Inc., Cambridge, MA) (Table 1).