Mechanisms of Resorption and Deposition of Calcium in the Carapace of the Crab Carcinus Maenas

Among the most conspicuous events concerned with moulting in the calcified crustaceans is the resorption of calcium and carbonate from the exoskeleton during the premoult period, and the subsequent calcification of the cuticle during postmoult. Our understanding of these processes rests largely on morphological data, and the visible events associated with changes in the calcium content of the cuticle are now well known. The mechanism of calcium movement, on the other hand, has not been elucidated.

The first sign of preparation for the moult is apolysis, the separation of the hypodermis from the overlying cuticle through the enzymic solation of the membranous layer (Jeuniaux, 1959a, b; Bade & Stinson, 1978). The outer epithelial cells of the hypodermis change from squamous to columnar and begin to elaborate the new epi- and exocuticular matrices (Drach, 1939; Travis, 1965; Green & Neff, 1972). At the same time as these pre-exuvial layers are deposited, old cuticle is resorbed. Resorptive activity reaches a maximum during late premoult (stage D₂) (Drach, 1939; Green & Neff, 1972) and results in the elevation of haemolymph Ca2+ and concentrations (Robertson, 1960; Dejours & Beekenkamp, 1978).

Following the moult, the pre-exuvial layers are calcified beginning in the most external regions and proceeding proximally (Travis & Friberg, 1963; Bouligand,1970). Mineral apparently reaches the outer portions of the cuticle via the pore canals, which are protoplasmic extensions of the epithelial cells that penetrate the cuticle up to or into the level of the epicuticle (Hegdahl, Gustavsen & Silness, 1977). Calcium has been localized histochemically in the distal portions of the epithelial cells and appears to be extruded in vertical rows corresponding in position to the pore canals of the new cuticle (Travis, 1957, 1963, 1965; Travis & Friberg, 1963; Chockalingham, 1971). Endocuticular deposition begins after the moult and its calcification is concomitant with the formation of the organic lamellae (Drach, 1939; Travis,1957,1963,1965; Travis & Friberg, 1963).The epithelial cells are also apparently involved in the movement of Ca into the endocuticle, for mineral granules are seen tobe extruded along the apical cell borders in horizontal rows during endocuticle formation (Travis, 1957, 1963, 1965). The end of postmoult is marked by the deposition of the membranous layer and the cessation of net calcium deposition (Passano, 1960).

In addition to these morphological studies, electron-probe X-ray microanalysis of the hypodermis of Crangon crangon has demonstrated Ca peaks in the epithelial cells during the periods of maximal resorption (stage D) and maximal deposition (stage B) (Hubert & Chassard-Bouchaud, 1978). It seems, therefore, that calcium moves through the hypodermis both during the process of resorption of mineral from the cuticle in premoult and during the calcification of the cuticle in postmoult.

The mechanism of calcium translocation is not known. The present study was conducted to determine the mechanism of calcium transport across the hypodermis in premoult and in postmoult. Toward this end, unidirectional and net fluxes of calcium were measured in the isolated integument of the crab Carcinus maenas.

Green crabs, Carcinus maenas, having a carapace width of 3–6 cm were obtained from Northeast Marine Specimens Co., Woods Hole, Mass., and shipped by air in moist seaweed. Animals were held in a 25 gal recirculating aquarium containing artificial sea water (Instant Ocean, Aquarium Systems, Inc., Eastlake, Ohio). Animals were induced to autotomize six pereiopods by firmly pinching the carpus just distal to the merus with a haemostat. Autotomization was employed to accelerate the moult cycle so as to have a ready supply of animals in different moult stages (Bliss, 1960; Skinner & Graham, 1970, 1972). Crabs were fed on raw fish twice weekly.

A Carcinus Ringer solution (CRS) was formulated by modifying those of Pantin (1934) and Davenport (1941) to conform more closely with the ionic composition of Carcinus haemolymph as described by Webb (1940) and Greenaway (1976). CRS contained (in mm): NaCl, 491·44; KC1, 12·23; CaCl₂, 11·10; MgCl₂, 5·71, and NaHCO₃, 1·79. The resultant pH was 7·0 and was not adjusted further. In some experiments dextrose was added to a concentration of 5·55.mm CRS. Na-free CRS was prepared by substituting 491·44 mm choline chloride for NaCl and 1·79 mm KHCO₃ for NaHCO₃.

Pieces of hypodermis removed from beneath the dorso-branchial carapace were mounted in duplicate between halves of plexiglass Ussing chambers (Ussing & Zerahn, 1951) equipped with Sylgard (Dow Corning Corp.) washers to reduce edge damage. A circle of polyethylene mesh (111 μm openings, Spectramesh, Spectrum Medical Industries) was applied to one side of the tissue to lend support. The chambers were connected via Luer fittings to a glass bubble lift, and via 2 % agar-KCl bridges and calomel electrodes to a Keithley 160 Digital Multimeter. Ten ml CRS with dextrose were then added to each side of the bubble lift, and the chambers were mixed and aerated with 95%: 5% O₂/CO₂ (the resultant pH was 6·8). The transepithelial potential (TEP) was monitored and recorded throughout the course of the experiment.

The experiment was begun by the addition of 100 μCi of 45Ca (New England Nuclear, Boston, Mass.) in 50–100 μl CRS to one side of the chamber. A 10 μl sample of this bathing solution was taken at the beginning and end of the experiment. At 30 min intervals after the addition of 45Ca, 100 μl samples of the bathing medium from the opposite side of the chamber were removed and replaced with non-radioactive medium. Samples were placed on aluminium planchets, dried at 60 °C, and counted in a low background gas flow counter.

Rates of Ca incorporation into the intermoult and postmoult integument were also measured. Paired pieces of the dorsal-branchial carapace were removed from crabs. The hypodermis was not stripped from the carapace, but the integument was mounted in Ussing chambers intact between Sylgard washers. The Ussing chambers used here were of simpler construction and were not attached to bubble lifts. One piece of the paired tissues served as the control in each experiment.

Experiments were begun by the addition of an experimental medium (4 ml) containing 10 μCi45Ca per ml to the hypodermal side, and 4 ml of CRS to the carapace side. The medium on the hypodermal side was aerated with 95 % :5 % O₂/ CO₂ via a No. 23 needle. Two 10 μl samples of the experimental medium were taken at the beginning and end of the experiment, and one 100 μl sample was taken from the carapace side of the chamber at the beginning and end.

The preparations were incubated for 3 h, at the end of which the media from both sides were drawn off. The pieces of integument were removed from the chambers and washed vigorously in non-radioactive CRS for 30 s. The hypodermis was then stripped from the carapace with a cotton swab and the carapace was once again rinsed for 30 s in CRS. Calcium was extracted from the carapace by placing it in 10 ml of 0·1 mm EDTA in distilled water (pH 4·5) for 2 days. Following extraction, four 100 μl samples of the extraction medium were taken, placed on planchets, dried and counted as described above.

The counts were corrected for extraction volume and, using the specific activity of the experimental media, were converted to values expressing mmol total Ca corporated into the carapace per cm2 per h.

Animals were staged according to the criteria of Drach & Tchernigovtzeff (1967) based on the hardness of the postmoult cuticle and the presence and condition of the pre-exuvial cuticle in premoult. Premoult animals were further categorized as _eD₀, D_o, D₁′, D₁″, D₁‴ and D₂ by microscopic examination of the first maxilla to determine the state of apolysis and setagenesis according to the criteria of Stevenson (1972) and Reaka (1975). Carapace width and sex were recorded along with the moult stage of each crab prior to use.

NaCN (Baker and Adamson) was prepared as a 100 mm stock solution in distilled water and added to experimental media to give a final concentration of 0·5 mm.

2,4-Dinitrophenol (DNP; Matheson, Coleman and Bell) was prepared as a 100 mm stock in ethanol and was added to experimental media to give a final concentration of 0·5 mm (86 mm ethanol). Paired controls received the vehicle alone to a concentration of 86 mm ethanol.

Quercitin (Sigma) was prepared as a 25 mm solution in dimethylsulfoxide (DMSO) and added to experimental media to give a final concentration of 100 μM (56 mm DMSO) (Fewtrell & Gomperts, 1977a, b). Paired controls received the vehicle alone to a concentration of 56 mm DMSO.

Ouabain (ouabain octahydrate, Strophanthin-G ; Sigma) was prepared as a 1 mm stock solution in CRS or Na-free CRS. Stock solutions were stored at 0 °C for no longer than 1 week. Ouabain was used at a concentration of 10 μM.

Quinine monohydrochloride dihydrate and quinidine sulphate were the gift of Dr John Parker, Department of Medicine, University of North Carolina, Chapel Hill. Both were prepared and used as 1 mm solutions in CRS or Na-free CRS with dextrose.

The results of unidirectional Ca2+ fluxes across isolated pieces of hypodermis from intermoult and premoult crabs are summarized in Table 1. The values for the unidirectional influx of Ca2+ (carapace to haemolymph) for intermoult (C₄), early premoult (_eD₀, D_o, D₁′, and D₁″)> and mid-premoult (D₁‴) tissues were all approximately 4 × 10−8 mol/cm2.h, with no significant differences distinguishable between the mean values for the various stages. There was, however, a significant increase in the unidirectional influx (J_i) in late premoult (D₂) tissues to a value of 22·6 ± 5·9 × 10−8 mol/cm2.h. The unidirectional effluxes (J_o, haemolymph to carapace) across both intermoult and premoult tissues were approximately 3 × 10−8 mol/cm2.h. There were no significant differences between effluxes measured in different staged animals, including late premoult (D₂). There was, however, considerable individual variation in the flux values, either influx or efflux, within a given stage. The problem of individual variation in comparing influx values to efflux values was alleviated by comparing those flux values obtained from paired tissue pieces from individual crabs. The results of one such experiment in which the influx of Ca2+ was measured on one piece and the efflux measured on another piece removed from a late premoult (D₂) animal is shown in Fig. 1. The flux ratios (J_i/J₀) from all such experiments performed on paired pieces of tissue from intermoult and premoult crabs are shown in Fig. 2. The flux ratios clearly demonstrate the differences between the influx and efflux values of crabs in a given moult stage. Intermoult tissues (C₄) showed a wide range of flux ratios, while premoult (stage D) flux ratios were, with one exception, greater than one. Additions of DNP (0·5 mm) during the course of the unidirectional flux preparations gave inconsistent results. DNP showed a tendency to decrease the unidirectional influx in intermoult and premoult tissues by an average of 23%, but the difference was not significant.

Intermoult and premoult tissues showed a small transepithelial potential (TEP) when first mounted in the Ussing chambers. The potentials invariably declined rapidly over the first 15 min of incubation, approaching 0 mV by the time sampling was initiated (1 h after the addition of 45Ca). Intermoult tissues had a slightly positive initial TEP (haemolymph side w.r.t. carapace side) of +0· 51 mV. In early premoult (_eD₀) tissues, the polarity was reversed and was − 0· 56 mV. The later premoult stages had more negative TEPs, −3· 17, − 1·10 and −2·71 mV for stage D₁″, D₁‴ and D₂ tissues respectively.

Although the small initial potentials of premoult tissues would favour the inward movement of Ca2+, the magnitude of these potentials in the absence of a chemical gradient was insufficient to account for the flux ratios which were obtained, according to the Ussing-Teorell flux ratio equation. The flux ratios predicted by this equation are represented by + signs in Fig. 2. Moreover, during the entire course of sampling, the potentials across all tissues were essentially zero.

The values for the net uptake of calcium from CRS with dextrose by isolated postmoult and intermoult integument (carapace plus hypodermis) are summarized in Table 2. The net uptake rate was highest soon after the moult (stages A₁ and A₂) and decreased as the animals advanced toward intermoult. Within a given moult stage, there was considerable individual variability in net rates of calcium uptake into the carapace which is reflected by the high standard errors of the mean values (Table 2). Consequently, it was necessary to use paired pieces of integument from individual crabs to elucidate the mechanism of calcium uptake through the use of inhibitors.

The effects of the various inhibitors and manipulations of the incubation media are summarized in Table 3. The addition of 0·5 mm NaCN to the medium significantly reduced the net Ca uptake to an average of 45 % that of control tissue incubated in CRS with dextrose alone (Table 3). Quercitin (100 μM), a plant flavone with reported general transport ATPase inhibitory properties (Fewtrell & Gomperts, 1977a, b), also significantly inhibited net uptake by 70% with respect to controls in CRS with dextrose (Table 3). This inhibitory effect on Ca uptake was also seen with the addition of 0·1 mm ouabain, an inhibitor of the Na, K-ATPase. Ouabain inhibited uptake by an average of 86% (Table 3).

The replacement of Na+ with choline in the incubation medium resulted in a marked and significant increase in net Ca uptake, an average of 318% above control values (Table 3). Quinine (and quinidine) also caused a highly significant increase in net calcium uptake into the carapace of Carcinus (Table 3, with those tissues incubated in the presence of the alkaloid accumulating 177% more Ca than controls. With no Na in the incubation medium, however, ouabain had no effect on the uptake of Ca into the carapace. Similarly, the stimulatory effect of quinine was abolished in the absence of Na, compared to controls in Na-free CRS with dextrose and no quinine (Table 3).

To see whether or not quercitin exhibited its inhibitory properties independent of its possible effects on Na, K-ATPase, quercitin was added to preparations in the presence of quinine. Quercitin inhibited Ca uptake by 77% in the presence of quinine compared to paired controls incubated in CRS with dextrose and quinine, but without quercitin (Table 3). Ouabain, on the other hand, had no effect in the presence of quinine (not shown).

The results of the unidirectional flux experiments demonstrate an increase in the influx of Ca from intermoult and early premoult to late premoult (stage D₂). Efflux values, on the other hand, remained constant throughout premoult. This situation would result in an increase in the net inward movement of Ca in stage D₂ of premoult, which is precisely what is observed in vivo; i.e. the maximum rates of exo-skeletal resorption occur during stage D₂ (Drach, 1939; Green & Neff, 1972). The flux rates observed were measured in the absence of a transmembrane electrochemical gradient. Thus it appears that the premoult resorption of mineral from the carapace is due, at least in part, to the active inward transport of Ca across the hypodermis.

The widespread variation in unidirectional flux values and flux ratios which were obtained for intermoult tissues and for preliminary experiments with postmoult tissues (not shown) probably were the result of tissue damage. Premoult tissue had undergone apolysis prior to the experiments; thus these pieces of hypodermis were naturally separated from the carapace before being removed for study. Intermoult and postmoult hypodermis, on the other hand, had to be peeled from the carapace for unidirectional flux experiments, a procedure of unavoidably rough handling.

The problem was overcome by using a portion of the whole integument (carapace plus hypodermis) to study the mechanisms of Ca movement in postmoult and intermoult crabs. The results of these net uptake experiments showed that the highest rates of Ca incorporation into the carapace in vitro were during stages A₁ and A₂. Again this corresponds to the stages in which maximal rates of Ca incorporation are seen in vivo (Welinder, 1975). In Astacus fluviatilis the rate of Ca uptake into the postmoult cuticle is approximately 4 × 10−7 mol Ca/cm2.h (Welinder, 1975), which is close to that seen in vitro in stage A₁Carcinus in the present study (8·82 × 10−8 mol Ca/cm2.h).

The inhibition of postmoult Ca incorporation into the carapace by cyanide suggests a dependence of outward Ca transport on metabolic energy, either directly or indirectly. The involvement of an ATPase in Ca translocation was indicated by the inhibitory effect of quercitin on Ca uptake. Quercitin has been shown to interfere with the activities of numerous transport ATPases including the mitochondrial ATPase (Lang & Racker, 1974), Mg-ATPase and Na, K-ATPase (Carpenedo et al. 1969), and the Ca-ATPase associated with the extrusion of Ca from the cytosol of most cells (Fewtrell & Gomperts, 1977a).

Ouabain, a specific inhibitor of the Na, K-ATPase, also reduced cuticle uptake of Ca. The large increase in Ca uptake in the absence of Na on the inside of the hypodermis also clearly demonstrated that Na was intimately involved in Ca transport. Moreover, it was also clear that the effects of ouabain on Ca transport were mediated through its effects on Na movements in the hypodermis, since ouabain had no effect in the absence of Na from the incubation medium.

Quinine (and quinidine) inhibits the influx of Ca and the Ca-activated Na efflux in dog red blood cells by blocking a Na/Ca exchange mechanism (Parker, 1978). Quinine increased the uptake of Ca into Carcinus carapace when present in the incubation medium in a concentration of 1 mm. Its effect was dependent upon the presence of Na in the medium as no stimulation of Ca uptake was seen with quinine in Na-free medium. These results suggest that a Na/Ca exchange mechanism present in the hypodermis plays a role in Ca transport. This does not, however, rule out a priori an involvement of a Ca-ATPase type pump operating in conjunction with the exchange mechanism. A Ca pump was indicated by the results of experiments utilizing quinine in conjunction with quercitin. Quercitin retained its inhibitory effect even when the Na/Ca exchange mechanism had been blocked with quinine. This was not the case for ouabain, which demonstrated no effect in the presence of quinine.

It thus appears that the transport of Ca across the hypodermis results from the combination of a Ca-ATPase and a Na/Ca exchange mechanism. A model which fits the data presented is shown in Fig. 3. The components of the model are an outwardly directed Ca pump which pumps Ca from the cytosol to the extracellular space, a Na/Ca exchanger which obligatorily exchanges cytosolic Ca or Na for extracellular Ca or Na with an unknown stoichiometry, and a Na, K-ATPase like that found in most animal cells. It is postulated that these components are located throughout the membranes of the outer epithelial cells of the hypodermis, including those of the pore canals. The model further presumes that the intracellular calcium and sodium activities are maintained at levels below that in the haemolymph, and that the epithelial cells are joined by intercellular junctions which restrict paracellular ionic permeability.

Intracellular Na and Ca activities are, indeed, maintained at levels below that in the extracellular fluids in the crustaceans as well as in most other animal cells. Intermoult and postmoult hypodermal intracellular Na concentrations in Cancer magister are approximately 50 mm (Guderly, 1977). Conversely, intracellular K concentrations are higher than ambient, and the disequilibrium is maintained by a ouabain-sensitive Na, K-ATPase (Leader & Bedford, 1978). While virtually all intracellular Na is in free ionic form, this is not the case for Ca (Blaustein, 1974). Haemolymph Ca is in the range of 7 –12 mm ; total intracellular Ca in Cancer magister hypodermis is 16 mm (Guderly, 1977). It is highly probably, however, that this value far exceeds the intracellular activity of Ca. This has proved to be the case for jnost animal cells thus far studied ; intracellular Ca activity is in the range of 0 · 01 –1 μ M (Rasmussen, 1970; Blaustein, 1974; Bygrave, 1978).

Junctions, which appear to be continuous septate junctions, are evident in the hypodermal epithelium of Uca (Green & Neff, 1972). The junctions are found at the apical margins of the outer epithelial cells in every stage of the moult cycle. Continuous septate junctions have been implicated in restricting permeability and are impermeable to lanthanum. Moreover, invertebrate septate junctions form ‘blisters’, as do vertebrate tight junctions, under imposed osmotic gradients (Lord & DiBona, 1976).

While the model may explain the transport of calcium during the deposition of the cuticle, a central question remains. How can a net inward movement of Ca during premoult be followed within a matter of days by the net outward movement of calcium during postmoult? If one observes the changes in the shape and the size of the epithelial cells and pore canals during the moult cycle, the proposed model may answer this question also.

The onset of premoult is marked by apolysis, the separation of the epithelial cells from the old cuticle and the concomitant severing of the pore canals from the apical cell surface (Drach, 1939; Jeuniaux, 1959a, b; Bade & Stinson, 1978). The epithelial cells also elongate, changing from squamous to columnar (Drach, 1939; Travis, 1965; Green & Neff, 1972). As the pre-exuvial layers are deposited in late premoult, the pore canals begin once again to elongate. In postmoult, the cells return to the squamous state, and the pore canals are further elongated with the progressive deposition of the new endocuticle (Green & Neff, 1972).

Utilizing light and scanning electron micrographs prepared in conjunction with another study (Dillaman & Roer, 1980), measurements were made of the dimensions of the epithelial cells in the squamous and columnar conditions and the dimensions and numbers of the pore canals (Table 4). Calculations revealed the important contribution of the pore canals to the cell surface area. In Carcinus, there are approximately 950000 pore canals per mm2 of epithelial surface. It is clear that changes in the length of the numerous pore canals will greatly affect the apical surface area (Fig. 4). When the pore canals are severed in premoult (pore canal height = o) and the epithelial cells become columnar, very little (about 3%) of the total cell surface area exists apical to the intercellular junctions. Thus, in premoult, about 97% of the Ca passively entering the cell either from the molting fluid in contact with the cuticle or from the haemolymph will be pumped out, according to the model, on the baso-lateral side of the epithelium. This would effect a net inward movement of Ca during premoult. As premoult continues and the pore canals elongate with the deposition of the pre-exuvial layers, more and more of the total cell surface area comes to lie apical to the junctions and, hence, more and more Ca entering the cell will exit on the apical or outer side of the epithelium. Based upon these calculations, by the time the pore canals have reached a height of only 20 μM, 50 % of the Ca will be pumped out of the cells toward the cuticle (Fig. 4). By the time of the moult, the net movement of Ca will, in large part, be toward the carapace, over 70% by stage A₂. This net outward transport of calcium would thus continue until all the sites for Ca in the cuticle are filled (i.e. the cuticle is fully calcified) and a steady state is reached between inward diffusion and outward pumping of Ca along the apical surface. This steady state would be maintained until apolysis in preparation for the next moult.

An implication of the model is that the hypodermal epithelium of the crab is not a specialized calcium-transporting tissue per se, but that the transepithelial movement of Ca is a consequence of the more general mechanism for cellular calcium extrusion. Net transport is effected by asymmetry of the epithelium; the direction of net transport only depends on the amount of Ca-extruding membrane on one side of the tissue as opposed to the other. This idea was also proposed by Wood & Harvey (1976) for Ca transport across the midgut of Hyalophora cecropia. That the crab hypodermis does not exhibit specialization for net calcium translocation may be a restriction imposed by the requirements for bidirectional net transport at different stages of the moult cycle.

I wish to thank Drs Karl Wilbur and John Parker for their help and advice. This project was supported by grant no. PCM76-81470 by the National Science Foundation.