Functional and morphological plasticity of crocodile (Crocodylus porosus) salt glands

The estuarine crocodile, Crocodylus porosus, displays broad euryhaline capabilities, with animals observed in waters ranging in salinity from 0‰ to over 60‰ (Grigg et al., 1986). Data from field surveys demonstrate that animals can maintain plasma osmolarity within a narrow range (298–309 mOsm)along such salinity gradients (Grigg,1981). In order to successfully osmoregulate in saline environs, C. porosus utilise extra-renal salt-secreting glands(Taplin and Grigg, 1981), much like marine reptiles and pelagic sea birds. Marine birds have paired orbital salt glands, which secrete excess salt(Fänge et al., 1958; Schmidt-Nielsen et al., 1958),whereas in marine turtles, lachrymal salt glands are responsible for salt secretion (Schmidt-Nielsen and Fänge,1958). Sea snakes, however, have sublingual salt glands, which open into the oral cavity (Dunson and Taub, 1967). In crocodiles, the salt-excreting glands are located on the tongue immediately below the lingual epithelium and are responsible for the excretion of sodium and chloride ions(Taplin and Grigg, 1981; Taplin et al., 1982).

Lingual salt glands, unique to crocodilians, were first identified and described in C. porosus by Taplin and Grigg(Taplin and Grigg, 1981). The salt glands appear as 20–40 distinct pores on the posterior half of the lingual surface. Classified as compound, branched, tubular structures, each gland is composed of 10–14 distinct lobes. Lobes consist of a number of short, blind-ending secretory tubules which join together to form multiple interlobular ducts. Secretions exit the gland via these ducts. Numerous arterioles and venules punctuate the connective tissue around and between gland lobes, and a dense network of capillaries can be found in the connective tissues both surrounding and between the secretory tubules(Franklin and Grigg, 1993). Nerve fibres are also present in large numbers in the connective tissues between and throughout the salt gland lobes and tubules(Cramp et al., 2007; Franklin et al., 2005). Methacholine chloride, an acetylcholine agonist, stimulates lingual salt gland secretion in C. porosus (Taplin and Grigg, 1981) suggesting that active secretion may be under the control of the cholinergic nervous system(Franklin et al., 2005). In marine birds and turtles, salt loading by intravenous or intraperitoneal infusion of hypertonic salt solutions also stimulates secretions from the salt glands (e.g. Fänge et al.,1958; Schmidt-Nielsen and Fänge, 1958; Schmidt-Nielsen et al., 1958). Salt loading of restrained crocodiles does not stimulate secretion from the salt glands; an increase in the rate of secretion from the salt glands following salt loading only occurs in those that are freely moving and unstressed, implying a significant involvement of the adrenergic nervous system in regulating (suppressing) crocodilian salt gland secretion(Franklin et al., 2005; Taylor et al., 1995).

Avian salt glands show remarkable levels of phenotypic plasticity –both morphological and physiological – as a consequence of drinking saline water (for a review, see Shuttleworth and Hildebrandt,1999). In birds that have never been exposed to saline conditions,salt glands secrete little; initial exposure to an osmotic stressor initiates a cascade of morphological and physiological responses, which results in a marked increase in salt gland secretion rate. The most striking change in avian salt glands upon initial exposure to salt water is the increase in gland size (Ellis et al., 1963)which is brought about by changes in secretory cell volume, cell number and blood flow to and within the gland(Ballantyne and Wood, 1969; Hanwell et al., 1971; Hanwell and Peaker, 1975; Hossler, 1982; Knight and Peaker, 1979). These changes in morphology accompany increases in mitochondrial and glycolytic enzymes (McFarland et al.,1965; Spannhof and Jurss,1967; Stainer et al.,1970) and increases in the abundance and activity of Na⁺,K⁺-ATPase (e.g. Ernst et al., 1967; Ernst and Mills, 1977; Fletcher et al., 1967; Stewart et al., 1976). In addition, most of these adaptive responses appear to be reversible during de-adaptation (e.g. Hossler et al.,1978; Mcarthur and Gorman,1978).

Given the wide range of salinities across which the estuarine crocodile Crocodylus porosus can successfully osmoregulate in the wild, we speculate that crocodile salt glands should demonstrate a broad functional flexibility. Moreover, this functional flexibility should also be manifested in the cellular morphology of the glands. In previous studies, it has been shown that the gross morphology of the salt glands changes significantly with prolonged acclimation to saline environments(Cramp et al., 2007; Franklin and Grigg, 1993). Although we know that C. porosus transferred from fresh to salt water are immediately capable of maintaining the plasma ion homeostasis(Taplin, 1985), suggesting that salt gland function can be upregulated rapidly, little is known about long-term acclimatory changes in the structure and function of the glands following prolonged exposure to hypersaline environs.

In the present study, we have examined whole animal osmoregulatory capacities, maximal secretory rates, and tissue metabolic rate in C. porosus salt glands following prolonged acclimation to both fresh and hypersaline environments. Additionally, we performed a detailed morphological study of salt gland secretory cells from acclimated animals to determine if morphological changes to cells occur with prolonged saltwater acclimation which may correlate with functional changes in the tissue.

Estuarine crocodiles Crocodylus porosus Schneider 1801 were obtained from Fleays Wildlife Park (Queensland, Australia), as eggs. The eggs were incubated in a Clayson (Brendale, Queensland, Australia) IM 1000 incubator at 30°C in plastic containers covered in moist vermiculite. Upon hatching, all animals were transferred to 1000 l Nally bins containing 20 l of heated (30±2°C) freshwater. A basking area was supplied with a standard 40 W globe, turned on daily from 09:00 h until 16:00 h. The crocodiles were fed ad libitum on a mixture of live crickets and dead baby mice. Water changes were carried out twice weekly. Once all animals were feeding well (∼2 weeks), animals were assigned to salinity treatment groups.

Hatchling crocodiles were randomly assigned to one of two treatment groups:freshwater (FW) or saltwater (SW) acclimation. SW-acclimated animals were placed into 70% seawater made using commercially available artificial sea salt(Aquasonic, Wauchape, New South Wales, Australia). Animals in both treatments were fed and had their water changed twice weekly. Animals remained in these treatments for 6 months.

Animals (N=6 per treatment group) were fasted for 1 week prior to experimentation. Animals were removed from holding tanks, weighed and restrained on a foam-lined board and their mouths propped open with a cork and secured with fabric tape. The tongue surface was rinsed clean with water and then blotted dry. Animals then received 2 mg kg^–1methacholine chloride in sterile saline via intraperitoneal injection and were left undisturbed for 10 min. Whatman filter papers pieces were cut to the size and shape of the animal's tongue. Secretions from the salt gland pores were collected by placing the filter papers over the tongue and pores and allowing the secretion to be absorbed into the filter paper. Filter paper pieces were large enough that they did not become completely saturated with secretion during the sampling period and that ensured secretions were continually absorbed by capillary action into the paper. Filter paper pieces were left on the tongue for 10 min. At the end of 10 min, the filter paper was removed and placed into a glass vial containing 1 ml of ultrapure water(Millipore, 18.2 MΩ cm). The process of secretion collection was repeated three more times. Vials containing salt secretions were allowed to sit overnight for the salt to diffuse into the surrounding fluid. Na⁺ concentrations were measured with an EEL flame photometer(Corning Glassworks, Corning, NY, USA), taking into account the additional Na⁺ present in the filter paper prior to the addition of the secretions. Na⁺ secretion rate data were scaled to body mass and are presented as μmol 100 g^–0.7 body mass h^–1. The highest rate of secretion achieved over the 40 min period was deemed the maximal secretory rate. Mass and maximal secretory rates were compared with Student's t-tests using the statistical program SigmaStat™. Data are presented as means ± s.e.m.

A 1 ml blood sample was taken from the cervical sinus of six crocodiles from each treatment group into individual heparinised syringes. Form these,duplicate 50 μl samples were taken and placed in haematocrit tubes and centrifuged for 2 min at 13 000 g. The remaining blood sample was centrifuged at 2500 g for 3 min and the plasma collected. Osmotic pressure was measured by vapour pressure osmometry, and Na⁺, K⁺ and Cl^– concentrations were determined using an Olympus AU400 automated analyser (Olympus America Inc.,Melville, NY, USA) using the ion-selective electrode (ISE) method. All haematological parameters were compared with Student's t-tests or where data failed tests of normality or equal variance, non-parametric Mann–Whitney rank sum tests were performed. All analyses were performed with the statistical program SigmaStat™. Data are presented as means± s.e.m.

In vitro salt gland oxygen consumption rate was measured using a Strathkelvin Instruments Mitocell S200 micro respirometry system comprising a 782 oxygen meter, 1302 oxygen electrode MT200 respirometer (Strathkelvin Instruments, Motherwell, Scotland). Crocodiles (N=5 per treatment)were killed by overdose of anaesthetic (sodium thiopentone, Jurox, Rutherford,New South Wales, Australia). The tongue was excised whole and placed into ice-cold crocodile ringer (containing in mmol l^–1, NaCl 116,KCl 4, CaCl₂ 2, MgCl₂ 2, glucose 15, sodium pyruvate 5,Hepes-Na 10). The salt glands were dissected free of the underlying connective tissue and overlying epithelia. Small (<0.5 mm thick↔1 mm²)slices of salt gland tissue were placed into beakers containing ringer at room temperature bubbled continuously with air and allowed to incubate for 15 min. One piece was then placed into a MT200 microrespirometer containing a magnetic stirrer bar and 0.5 ml of air-saturated crocodile ringer at 25°C. The chamber was sealed and the rate of oxygen consumption by the tissue in the chamber was recorded instantaneously until the partial pressure of oxygen in the chamber had dropped by 20% (approximately 30 min, depending on the treatment group). At the end of this period the chamber was opened and the ringer was replaced with fresh oxygenated ringer containing 0.1 mmol l^–1 methacholine chloride. The chamber was once again sealed and the rate of oxygen consumption measured until the partial pressure of oxygen in the chamber had dropped by 20%. Prior to the commencement of experimentation, trials were conducted to ensure that oxygen consumption rates did not drop substantially over the course of the experiment. The data (not shown) indicated that, apart from an immediate drop in the metabolic rate of the tissue following excision (first 15 min), the metabolic rate of the tissue was subsequently stable for at least 4 h. Care was taken to ensure that all experimental measures were completed within this 4 h window. At the completion of the experiment, tissue pieces were blotted dry and weighed. Prior to the commencement of, and at the end of each trial, the rate of background oxygen consumption in the chamber (from contaminants and the electrode) in the absence of any tissue was recorded and this mean rate subtracted from the tissue rates. In order to minimize bacterial contamination of the chamber,chambers were cleaned prior to trials with 70% ethanol and the chamber ringer was filtered through 0.2 μm Millipore syringe filters. Rates of oxygen consumption were standardized to tissue wet mass and are reported as μl O₂ g^–1 h^–1. Data were compared statistically using a two-way repeated-measures AVOVA. Post-hoccomparisons were made with Holm–Sidak tests. All statistical comparisons were made using the statistical program SigmaStat™. Data are presented as means ± s.e.m. and treatments were considered as significantly different if P<0.05.

Crocodiles (N=4 per treatment) were euthanased and salt glands dissected out as described previously. Small pieces (circa 2 mm²) containing at least one visible pore were dissected out from the surface layer and were placed into cold 2% glutaraldehyde (ProSciTech,Thuringowa, Australia) in 0.1 mol l^–1 phosphate buffer pH 7.4 for 24 h. Following glutaraldehyde fixation, tissues were washed in phosphate buffer and then post-fixed in 1% osmium tetroxide (ProSciTech, Thuringowa,Australia) in phosphate buffer. Tissue pieces were then dehydrated through an ascending ethanol gradient and infiltrated with Spurrs medium (ProSciTech). Tissues were blocked and allowed to polymerise at 60°C for 48 h. Semi-thin sections were cut, stained with 1% Toluidine Blue in 0.1 mol l^–1 phosphate buffer and viewed under a microscope to determine if the appropriate position in the tissue piece had been reached(i.e. there were secretory cells present). Once the appropriate position within the tissue block was confirmed, ultrathin sections (60 nm) were cut with a diamond knife and collected onto 2 mm copper grids. All sectioning was performed with a Leica Ultracut UC6 Ultramicrotome. Sections were stained with uranyl acetate and lead citrate and viewed at 80 kV in a Jeol 1011 transmission electron microscope. Images were captured with a MegaView III digital camera using the iTEM™ software package.

Between 50 and 70 images of different secretory cell were collected by transmission electron microscopy (TEM) from four animals per treatment group. Mitochondrial volume density, plasma membrane surface area, intercellular space volume, and mitochondrial cross-sectional area were measured using the morphometrics software package, SigmaScanPro™.

The volume density of mitochondria within glandular secretory tissue(V_{V mit}) was determined by systematic point counting of mitochondria in ultrathin sections. Using image analysis software, a coherent test lattice (30×24 μm; with lines spaced 2 μm apart) was superimposed over digital transmission electron micrographs of glandular tissue. The number of points of intersection (i.e. where test lines cross one another) over mitochondria was then counted. To calculate V_{V mit}, this number was divided by the total number of points of intersection over secretory tissue.

The volume density of intercellular space within glandular secretory tissue(V_{V ics}) was determined by systematic point counting using the same coherent test lattice and digital TEMs of glandular tissue as used to determine V_{V mit}. To determine intercellular space, the number of points of intersection over areas of intercellular space was counted and divided by the total number overlaying secretory tissue.

The surface density of basolateral cell membrane within secretory tissue(S_{D cm}) was determined using a coherent test lattice(30×24 μm; with lines spaced 4 μm apart) superimposed over the same TEMs used to determine V_{V mit} and V_{V ics}. To determine S_{D cm}, the number of times test lines intersected the basolateral membrane of secretory cells was counted in micrographs from each animal. To calculate S_{D cm}, this number was doubled and then divided by the total length of line overlaying secretory tissue (L_C). L_C was calculated using the equation: L_C=L_T×V_{V st},where L_T is the total length of test line applied and V_{V st} is the volume density of secretory tissue (st) in micrographs.

Differences in V_{V mit} between treatment groups could reflect a difference in the size and/or density of mitochondria. To investigate differences in V_{V mit}, we compared mitochondrion cross-sectional area between treatment groups. We did this by measuring the cross-sectional area of 200 mitochondria from each animal in both treatment groups. Cross-sectional area was measured directly from digital TEMs of glandular tissue using the image analysis program (SigmaPlot™7.0).

S_{D cm} will depend not only on the extent of elaboration of the basolateral cell membrane, but also the number of cells per unit volume of secretory tissue. To investigate the effects of saltwater and freshwater acclimation on basolateral cell membrane area more thoroughly, we compared the number of nucleated cells, per unit volume density of secretory tissue(V_{V st}), in micrographs from each animal. We did this by counting the number of nucleated cells in TEM micrographs and dividing this by the volume density (V_{V st}) of glandular secretory tissue. All morphological parameters were compared with Student's t-tests using the statistical program SigmaStat™. Data are presented as means± s.e.m.

All reagents were sourced from Sigma Aldrich (Sydney, Australia) unless otherwise specified.

There were no differences in mean animal mass between treatment groups(P=0.5193, t=0.6679, d.f.=10l; FW-acclimated range 110–375 g; SW-acclimated range 130–234 g). Maximal Na⁺secretion rates from the lingual salt glands of C. porosus were highly variable (FW-acclimated range: 12–53; SW-acclimated range:33–137 μmol 100 g^–0.7 body mass h^–1). There was a marked increase in the mean maximal salt gland secretory rate of the SW-acclimated animals when compared with their freshwater counterparts (t-test, P=0.002, t=–4.053, d.f.=10; Fig. 1A). In SW-acclimated animals, maximal secretion rates were attained 15 min post methacholine injection and declined steadily over the subsequent 40 min period. In FW-acclimated animals, secretion rates appeared to have attained their maximal level within 15 min, but remained constant over the subsequent collection period (Fig. 1B).

The haematocrit of blood from FW-acclimated animals was significantly lower than that of SW-acclimated animals (t-test, P=0.042; Fig. 2A). Plasma osmolarity was also significantly lower in FW-acclimated animals relative to SW-acclimated ones (t-test, P<0.05; Fig. 2A). Both plasma Na⁺ and K⁺ concentrations were significant higher in SW-acclimated individuals (Na⁺-SW= 173±2.9,Na⁺-FW=147±9.5 mmol l^–1;K⁺-SW=5.9±0.18, K⁺-FW=4.8±0.35 mmol l^–1) but Cl^– concentrations were not different between the treatment groups (Mann–Whitney rank sum test, P=0.065; Fig. 2B).

There was no statistically significant difference in the control (pre-drug)metabolic rates of tissue slices from FW- and SW-acclimated treatment groups(two-way RM ANOVA P=0.842, t=0.205; Fig. 3). Addition of methacholine chloride to tissue slices from FW-acclimated animals had no statistically significant effect on oxygen consumption (P=0.065, t=2.137) whereas in SW-acclimated animals, methacholine chloride addition resulted in a 33% increase in oxygen consumption (P=0.02, t=4.426).

The secretory tubules of C. porosus consist of a single layer of simple cuboido-columnar epithelial cells. The epithelium is composed primarily of just one cell type, the principal secretory cell, although this cell type appeared to have several morphological states. At the blind end of the secretory tubules, the principal cells were typically conical in cross section. They were, on average approximately 15 μm long and 7 μm wide at their widest point, tapering to approximately 3 μm at the apical tip of the cell (Fig. 4A). Morphologically, the principal cells were characterised by a greatly amplified basolateral membrane (BLM). Lateral, microvilli-like projections of the BLM intermeshed with those of adjacent cells(Fig. 4B). By contrast, the apical membrane surface area was relatively reduced. A few, small microvilli projected from the apical surface of the secretory cells into the secretory tubule lumen (Fig. 4C). Adjacent principal cells were linked at their apexes by zona occludens and zona adherens junctions and along their basolateral membranes by zona adherens junctions. Principal cells contained a large, basally located nucleus composed primarily of euchromatin, with small regions of heterochromatin at the nuclear periphery. The cells generally contained a large number of mitochondria, which were evenly dispersed throughout the cell. Numerous secretory vesicles were readily visible at the apical tip and, to a lesser degree, along the basolateral membrane of the principal cells. Large amounts of smooth endoplasmic reticulum (ER) could be seen throughout the cell. Rough ER was less evident in principal cells. In cross section most cells could be seen to contain at least one Golgi complex, generally located close to the nucleus. Many secretory cells also contained small electron-opaque, non-membrane bound vesicles. Principal cells rested upon a well-developed basement membrane and were in close proximity to capillaries(Fig. 4D).

There was no significant effect of saltwater acclimation on either the surface density of the basolateral cell membrane within secretory tissue(S_{D cm}; Table 1; t-test, P=0.714, d.f.=7) or the volume of intercellular space between secretory cells (V_{V ics}; Table 1; t-test, P=0.546, d.f.=7). There was a significant effect of saltwater acclimation on the volume density of mitochondria within the principal secretory cells; there was an increase of almost 23% in mitochondrial volume per μm³ in SW-acclimated animals relative to FW-acclimated animals (Table 1; t-test, P=0.024, d.f.=7).

There were no differences in mitochondrial cross-sectional areas between FW- and SW-acclimated animals (t-test, P=0.55, d.f.=7),suggesting that the increase in mitochondrial volume density is most probably due to an increase in mitochondrial numbers, rather than mitochondrial size.

There was a significant difference in the numbers of cell nuclei perμm² per micrograph, with SW-acclimated animals having significantly fewer nuclei than FW-acclimated animals(Table 1; t-test, P=0.046, d.f.=7), suggesting that cells from SW-acclimated animals were larger than those of FW-acclimated animals.

We examined indices of salt gland function both in vivo and in vitro and the ultrastructual morphology of the salt secreting tissue of the estuarine crocodile, Crocodylus porosus to determine its long-term adaptive capacity and phenotypic plasticity following chronic exposure to hypo- and hyperosmotic environmental conditions. As previously established, salt glands play a pivotal role in the maintenance of osmotic homeostasis in C. porosus living in hyperosmotic conditions. However,little was known about long-term effects of prolonged exposure to hyperosmotic environments on the maintenance of osmotic homeostasis. All haematological values measured in this study for SW-acclimated animals were within the ranges reported for wild-caught C. porosus living in water bodies of varying salinities (Grigg, 1981; Wells et al., 1991),suggesting that SW-acclimated crocodiles were able to successfully maintain osmotic homeostasis despite extended exposure to hyperosmotic conditions. Moreover, extended exposure to hypersaline conditions appeared to have little long-term impact on the animal's growth, suggesting that the additional energy required in maintaining homeostasis was either relatively minimal or was compensated for by increased food intake (animals were fed ad libitum, so we cannot readily distinguish between these possibilities).

Surprisingly, blood haematocrit, plasma osmolarity and plasma Na⁺ and K⁺ concentrations were significantly lower in FW-acclimated crocodiles when compared with their SW-acclimated counterparts. In previous studies of plasma Na⁺ concentrations and water flux in C. porosus, Taplin (Taplin,1985) demonstrated that FW-acclimated C. porosus actively drank freshwater when available, but when living in hyperosmotic environments,animals avoided drinking saltwater and obtain all necessary water though ingested food and metabolic water (Taplin,1985). The active consumption of freshwater by C. porosus, when available, is thought to be related to the method of excretion of nitrogenous waste utilised while in freshwater [i.e. predominantly ammonia in freshwater as opposed to insoluble urates when in saltwater (Grigg, 1981)] and the associated high requirement for water. The ingestion of such an electrolyte-depleted medium, whilst important for waste excretion, may secondarily be responsible for the relatively reduced plasma osmolarity and haematocrit, and lower electrolyte levels in FW-acclimated animals observed in this study. Moreover, the integument of the cephalic region (including the buccal cavity, but excluding the salt glands) is also highly permeable to Na⁺ and is a major secondary route for Na⁺ efflux (and possibly other ions) in C. porosus(Taplin, 1985). Hence, in freshwater, maintaining a lower plasma osmolarity, relative to SW-acclimated animals, may aid in reducing the rate of Na⁺ efflux from the buccal integument into the surrounding medium.

To investigate salt gland functional capacity in vivo,methacholine chloride, an acetylcholine (ACh) agonist, was used to stimulate maximal salt gland secretory rates in restrained C. porosus. Previously, Taplin reported that short term exposure (∼160 h) to saline environments in C. porosus hatchlings reared in freshwater was not long enough to induce differences in maximal salt gland secretion rates relative to animals maintained continuously in freshwater(Taplin, 1985). However, in this study we found that chronic exposure to hypersaline environments resulted in adaptive changes to the salt gland, which enabled a significantly elevated rate of Na⁺ secretion. Maximal mean lingual gland secretory rates were more than three times higher in SW-acclimated C. porosusrelative to FW-acclimated animals. In FW-acclimated C. porosus,methacholine stimulation appeared to have relatively little effect on secretion rate. These results possibly suggest that prolonged saltwater acclimation results in changes to the cholinergic pathways responsible for salt secretion.

Although acetylcholine can act on numerous aspects of the tissue to increase secretion rate, we believe that there is at least an underlying direct increase in the response of the secretory tissue to exogenous methacholine, which may increase the secretory capacity of the tissue. In SW-acclimated C. porosus salt gland tissue slices, application of methacholine had a marked effect on the in vitro metabolic rate of the tissue, raising it by over 30%. A comparable metabolic response to methacholine application, in vitro, has been observed in both avian and turtle salt gland slices (Shuttleworth and Thompson, 1987). In contrast to tissue slices from SW-acclimated crocodiles, those from FW-acclimated animals showed relatively little response to methacholine. This result suggests that tissue slices from SW-acclimated crocodiles are more responsive to ACh than those of FW-acclimated crocodiles.

It was recently demonstrated that crocodile salt glands are richly innervated by cholinergic neurons, suggesting a large role for ACh in the regulation of salt gland secretion(Franklin et al., 2005). Similarly, avian salt-secreting tissues also appear to be under the control of the cholinergic system (e.g. Ash et al.,1969; Schmidt-Nielsen and Fänge, 1958). Moreover, avian salt gland secretory cells also possess large numbers of muscarinic ACh (mACh) receptors, which when activated, initiate a multitude of secretion-related processes within the cell(e.g. Borut and Schmidt-Nielsen,1963; Hootman and Ernst,1981; Hootman and Ernst,1982; Stewart et al.,1979; van Rossum and Ernst,1978). In addition, saltwater acclimation results in a 3-fold increase in the number of mACh receptors on duck salt gland secretory cells(Hootman and Ernst, 1981),suggesting that ACh plays a large, multifaceted role in regulating the secretory process. We suspect that, as in avian models, ACh is likely to act directly on the secretory cells in the salt glands of C. porosus to regulate the secretory activity of the tissue. Moreover, saltwater acclimation is likely to result in an increase in ACh receptor numbers in C. porosus secretory cells, although this, and the other potential sites of action by ACh, remains to be empirically determined.

Increases in the maximal secretory capacity of SW-acclimated crocodiles were reflected by changes in secretory tissue morphology. Studies on birds and turtles have demonstrated that salt-secreting tissues are morphologically plastic and are capable of responding rapidly to changes in plasma osmolarity(e.g. Abel and Ellis, 1966; Benson and Phillips, 1964; Ernst and Ellis, 1969; Reina, 2000; Schmidt-Nielsen and Kim,1964). Changes include increases in blood flow to the gland(Hanwell et al., 1970) and the rapid increase in salt gland size by both cellular hypertrophy and hyperplasia(Abel and Ellis, 1966; Benson and Phillips, 1964; Reina, 2000; Schmidt-Nielsen and Kim,1964). In addition, changes in secretory cell ultrastructure,particularly marked increases in plasma membrane surface area and increases in the number of mitochondria (Abel and Ellis,1966; Bokenes and Mercer,1998) accompany increases in gland size. In C. porosus we found no increase in the surface density of plasma membrane per unit area in SW-acclimated animals, but did observe an increase in the volume density of mitochondria. As mitochondrial cross-sectional areas were not different across treatments, it suggests that changes in secretory cell mitochondrial volume per unit area arise from an increase in mitochondrial numbers (density per unit volume) in SW-acclimated crocodiles. Mitochondrial density is an important proxy for estimates of the metabolic activity of a cell. Hence cells involved in energetically expensive processes such as secretion and absorption often have large numbers of mitochondria. Salt-secreting cells generally have relatively large numbers of mitochondria. Indeed, actively secreting avian salt gland cells have approximately twice the volume of mitochondria compared with their inactive counterparts (Bokenes and Mercer, 1998). Moreover, the metabolic rate of secretory cells from SW-acclimated ducks is almost double that of naive (FW-acclimated)animals (Hootman and Ernst,1980).

In the present study, we also examined the rate of oxygen consumption by isolated crocodile salt gland tissue and found there was no difference in the mass-specific metabolic rate of tissues from FW- and SW-acclimated crocodiles. There are fewer cells per unit mass in SW-acclimated salt gland relative to FW-acclimated ones (Cramp et al.,2007) and when oxygen consumption rates are expressed in terms of the number of cells per unit mass, the metabolic rate of SW-acclimated crocodile salt gland tissue would be higher than that of FW-acclimated cells. Similar values are seen in duck salt gland models(Hootman and Ernst, 1980). Given that there are almost 23% more mitochondria per unit volume in SW-acclimated crocodiles, and again, since these secretory cells are so much larger than those of FW-acclimated crocodiles, this figure is also probably an underestimate of the total cellular volume of mitochondria.

Following prolonged saltwater acclimation, we observed no increase in the density of plasma membrane per unit volume of salt gland tissue relative to that from FW-acclimated crocodiles. As detailed above, this result is likely to be an underestimate of the actual value since there are fewer secretory cells per unit volume of SW-acclimated salt gland(Cramp et al., 2007). We suspect that when the surface area measurements are expressed relative to total cells size, there would be an increase in plasma membrane surface area of salt gland following acclimation to hypersaline environments. However, the extent of the change in plasma membrane area from SW-acclimated crocodile cells is significantly less than is observed in duck salt gland cells following osmotic challenge, in which there can be a fivefold increase in plasma membrane surface area (Merchant et al., 1985). These differences in plasma membrane response to salt loading may simply reflect species differences between crocodiles and ducks,but may reflect greater underlying differences in their abilities to respond to osmotic challenges over long periods of time. Further work is required to determine whether the ultrastructural changes we observed in chronically SW-acclimated salt glands are different to those that occur following acute osmotic challenge.

Relative to the salt glands of birds, the functional and morphological effects of saltwater acclimation in C. porosus were comparatively small. Although it is possible that these differences reflect different time courses over which these responses occurred, it is also possible that differing relative salt loads may also play a role. C. porosus can successfully obtain all necessary water requirements from ingested food and metabolic water production alone, without the need to maintain water balance by drinking seawater (Taplin,1984; Taplin,1985). Indeed, several species of marine reptiles (e.g. sea snake Pelarmis platurus, estuarine turtles Callagur borneoensisand Malaclemys terrapin and Crocodylus acutus) are thought to also maintain water balance without the need to drink seawater(Dunson, 1982; Dunson and Moll, 1980; Dunson and Robinson, 1976; Evans and Ellis, 1977; Robinson and Dunson, 1976). Taplin (Taplin, 1985) provided data on the relative salt loads of saline water food items of C. porosus in the Liverpool/Tomkinson River System, northern Australia,which showed that the Na⁺ content of these items ranged from∼55 to ∼215 mmol Na⁺ kg^–1 body mass (the mean Na⁺ content of food items fed to crocodiles in the present study was 145 mmol kg^–1 body mass). Hence when living in seawater, a crocodile's salt load would be equivalent to that of the food it consumes plus any incidental water ingested (i.e. during swallowing). By contrast, marine birds maintain hydration by consuming seawater and excreting the excess salt through their salt glands. With an average [Na⁺] of around 450 mmol l^–1, the ingestion of seawater would contribute to a much greater relative salt load than that experienced by crocodiles. Hence, differences in acclimatory responses to seawater exposure between marine birds and elasmobranchs and crocodiles may reflect different relative salt loads. We would hypothesise that increases in dietary salt loads may stimulate a great phenotypic response in the salt glands of C. porosus.

Crocodylus porosus has a remarkable ability to successfully inhabit severely hypersaline water bodies (>60‰)(Grigg et al., 1986). Animals maintain their osmotic balance by employing extrarenal lingual salt glands to concentrate and remove excess salt accumulated as a consequence of living in marine habitats. Data from the current and previous studies has shown that the salt glands of C. porosus can respond rapidly to changes in environmental salinity/plasma osmolarity(Cramp et al., 2007; Franklin and Grigg, 1993). The salt glands of C. porosus respond to increasing environmental salinity in much the same way as those of other marine reptiles and birds, in that they become larger, have larger secretory cells, larger blood vessels(Franklin and Grigg, 1993) and show some evidence of neural plasticity upon hypertrophying(Cramp et al., 2007). We have shown that a change in the ultrastructure of salt gland secretory cells (i.e. increased mitochondrial numbers and increase plasma membrane surface area)correlates with an increased functional activity of the glands. Thus, the salt glands of crocodiles exposed to hypersaline conditions have a greater secretory rate and higher tissue metabolic rate than those of freshwater-acclimated animals.

The authors would like to thank Rick Webb and Rob Gould for assistance with electron microscopy and Selena Hobbs, Lyn Beard and the UQ Small Animal Clinic Diagnostic Services Unit, for assistance with analysis of haematological parameters. This research was funded by an ARC Discovery Grant to C.E.F. and a UQ research grant to N.S. All experimental procedures were carried out with permission from QPWS (permit no. WISP00993703) and UQ Animal Welfare (permit nos SIB/257/05/ARC/URG and SIB/186/06/ARC/URG).