Biomineralisation is a phenomenon, widespread in the five realms, which appeared firstly in Bacteria [81] and [114]. Among the animals, numerous invertebrates are able to secrete biominerals, the first major function of which is the hardening of a skeleton, a structure that provides support and/or protection against environment [81], [114] and [147]. The fossil species, discovered so far, revealed that the first calcified metazoan exoskeletons were probably elaborated at the end of the Precambrian (Proterozoic) [77] and [147]. The reason for the appearance of such a process remains speculative, but several hypotheses have been evoked, among which protection against predators or against the ionic composition (notably in calcium) of the primitive ocean evolving with time. In the same time, the ocean became rich in phosphates and carbonates, promoting the precipitation of calcium as calcium phosphates and/or carbonates. Thus some biomineralisations may have been elaborated as a consequence of a neutralization process against external toxic ions and/or molecules. The calcifications externally elaborated could have been favourable to the survival of species, which were then protected against chemical and physical environmental pressures.

The biomineralisation process is especially well developed in arthropods, currently the largest group of animals on Earth. As a consequence of the presence of a rigid exoskeleton, also called cuticle, the growth and the whole physiology of these animals are tightly linked to moulting cycles, each of which is characterized by the complete replacement of the exoskeleton. Moreover, arthropods harden their new cuticle by a process called sclerotization (proteins-polysaccharides cross-linking), whereas most of the crustaceans proceed also by calcification.

The major source of calcium is exogenous: the water in which most crustaceans live. In seawater, calcium concentration is very high, but some groups also live in freshwater or on land, where the availability of calcium at ecdysis (moment where the animal leaves its old cuticle) ranges from great to complete absence. These animals have developed different strategies to solve the problem of calcification, notably by storing calcium during the pre-moult period [48] and [56], a phenomenon particularly well developed in the terrestrial species. Thus, the source of calcium used to mineralise the cuticle depends on the way of life of the considered crustacean. The food, a possible source of calcium for each crustacean, represents a minor contribution to the calcification.

Crustaceans are notably interesting because of their particularly active calcium metabolism and their ability to form cyclically not only an external calcified structure, but also, for several species, calcium storage forms.

In crustaceans, a cellular hypodermis of at least three layers underlies the carapace. The outer layer is responsible for the elaboration of the exoskeleton. This cuticle comprises four layers, from the external to the most internal: the epicuticle, the exocuticle, the endocuticle, and the membranous layer [108] and [128]. Except for the innermost layer, the three other layers are mineralised, essentially by precipitation of calcium carbonate and, to a lesser extent, of amorphous calcium phosphate into the twisted lamellar structure of the chitin–protein cholesteric matrix [19], [40], [108] and [128]. Calcification of the carapace has been particularly well studied in Decapoda [21], [39], [75], [108], [109] and [128]. It occurs at different sites within the cuticle: in the chitin–protein fibres, at the level of interprismatic septa and pore-canals. The latter are vertically extensions of the outer epithelial cells, which pervade the whole cuticle in very numerous sites. They are responsible for the transport of components necessary for the sclerotization as well as for the calcification of the cuticle in post-moult (quinone, phenoloxidase, calcium and carbonic anhydrase, for example). This route can also be taken by molecules responsible for the partial resorption of the old cuticle in pre-moult (protease, chitinase, and also carbonic anhydrase).

The carbonic anhydrase (CA) is a particularly important enzyme involved in the calcification processes in crustaceans, as in all the invertebrates elaborating calcified biomineralisations [62] and [114] by catalysing the reversible reaction: CO₂ + H₂O ⇔ HCO₃ ^– + H⁺. This enzymatic activity has been first evidenced in crustacean in the barnacle mantle by Costlow in 1959 [28] and in the carapace by Giraud in 1977 [37], where it has been localized at the calcification sites described above [38] and [39]. Implication of CA in cuticle calcification has not been extensively studied so far [63] and [64]. Except for aragonite, the various polymorphic forms of calcium carbonate (amorphous, calcite, vaterite) have been found in the crustacean cuticle. Nevertheless, the crystalline calcite seems to be the major one. The part of the new cuticle synthesized by the hypodermis during the pre-moult period is not mineralised before the old cuticle has been discarded. A portion of the minerals that made up the old cuticle are resorbed during the pre-moult period and lost into the environment or stored as calcium reserves in different organs. It is worth noting that to avoid toxification of the epidermal cells, calcium ions resorption does not occur via the cytoplasm: transport mechanisms used are pinocytosis vesicles or paracellular channels (extracellular dilated network delineated by cell membranes). Moreover, calcium is in general in bound or precipitated form. The same anti-toxification strategy is used by the most part of epithelial cells comprising the storage organs. Nevertheless, it is worth noting that at this level, calcium transport may also occur via a transcellular pathway using Ca²⁺-ATPase and Na⁺/Ca²⁺ exchanger mechanisms [4], [5] and [108].

The amounts of excreted or stored calcium vary considerably from one species to another. The stored calcium represents 4% to 75% of the calcium needed for the complete calcification of the exoskeleton [48], [56] and [114]. This calcium is accumulated during pre-moult and thus available at ecdysis. Whatever its contribution for the cuticle mineralisation, this stored calcium is used firstly before any other exogenous sources, sometimes restricted to the cuticle of some primordial organs such as the masticatory parts. In addition to this endogenous calcium, water, and food, which includes the rejected old cuticle (also called the exuviae), are exogenous sources of calcium used in post-moult.

For aquatic species, the loss of calcium at ecdysis is related to their ability to remove quickly a great amount of calcium from their environment. However, losses are limited in supralittoral or terrestrial crustaceans. They do not excrete, but store the calcium originating from their old cuticle. Storage sites are diverse and may occur in common or peculiar organs, such as haemolymph, hepatopancreas, cardiac stomach wall, sternites, or posterior ceca [48] and [114].

Calcium, from an endogenous or exogenous source, reaches the cuticle via the haemolymph. Similarly to the outer layer of the hypodermis, responsible for the synthesis of the cuticle, the organs storing calcium are generally formed by an unstratified epithelium across which calcium is translocated. These calcium-transporting epithelia are the subject of extensive studies [4], [5], [97], [139], [141] and [142] because they represent interesting models as compared to similar vertebrate ones such as the intestine, kidney, placenta, and chick chorioallantoic membranes [101]. Moreover, at this level, calcium transport involves, as in vertebrates, activation of calcium pumps, such as Ca²⁺-ATPases pumps [23], [108] and [140] and other common ion pumps such as H⁺-ATPase and Na⁺/K⁺-ATPase [114].

Except for the cirripeds (among which barnacles) [114] and copepods, where calcium storage has not been reported so far, all the other crustaceans have developed mineral storage strategies related to their moulting cycle.

Carapace and storage structures are biomineralisations, which means that they contain an organic matrix. The knowledge of the features of the matrix components, the study of the interactions occurring between these molecules to construct the molecular framework within which calcium is precipitated, and the discovery of the chemical interactions occurring at the interface between the matrix and the mineral are of first importance. That could lead us to understand the remarkable properties (elasticity, mechanical resistance...) of these minerals when comparing their ‘biological’ form to their inorganic form, as well as to better understand in a general way why and how biomineralisations are elaborated.

Molluscs and echinoderms represent two invertebrate groups where the characterization of proteins involved in biomineralisations is considerably increasing since a decade [16], [83], [145] and [146]. In crustaceans, few molecules have been characterized to date (Table 1) [10], [30], [61], [68], [69], [70], [71], [72], [73], [74], [77], [82], [98], [121], [122], [123], [129] and [137].

Previous investigations have shown that the crustacean cuticle contains numerous proteinaceous components [9], [116], [136] and [144]. At this time, 33 proteins from calcified parts of the cuticle (also called hard or solid cuticle contrarily to the membranous arthrodial cuticle called the soft cuticle) have been completely sequenced. Twenty-six cuticular matrix proteins have been sequenced by the team of Andersen from Copenhagen: one from the cuticle of the shrimp Pandalus borealis [74] (another one from the same origin has been characterized by Suzuki et al. [121]), 13 from the carapace of the American lobster Homarus americanus [77] and [98] and 12 from the exoskeleton of the crab Cancer pagurus [10]. Nevertheless, nothing is clearly known about the involvement of these cuticular proteins in the sclerotization or the calcification process. The chemical features of some of them only suggest that they could participate to the calcium precipitation and/or bind chitin.

The group of Nagasawa, from Tokyo, recently characterized and sequenced, from the exoskeleton of the crayfish Procambarus clarkii, two interesting polypeptides, named CAP-1 and CAP-2, both with anti-calcification (CAP-1 inhibits in vitro calcium precipitation), calcium-binding and chitin-binding properties (presence of Rebers–Riddiford motifs in the sequence and positive in vitro chitin-binding assay) [69], [70] and [71]. These proteins could both be linked to chitin and act as nucleator of the calcification process, as suggested [33] and [71].

Finally, four cDNAs were sequenced, which encode crustacean proteins of the exoskeleton of the prawn Penaeus japonicus (DD4, DD9A, DD9B and DD5) [30], [68] and [137]. Among these, only DD4 [30], encoding an acidic protein is clearly calcium binding and presents similarities with calphotin (a calcium-binding protein from Drosophila photoreceptor cells). DD9A and DD9B correspond to acidic proteins (20% of the amino-acids) with a Rebers–Riddiford motif and DD5 comprises a tandem repeat domain (of 93 to 98 a.a.) also with chitin-binding property.

As regards the storage structures, only two proteins have been evidenced, sequenced and extensively studied. First, an insoluble matrix protein, named GAMP, extracted from the gastroliths of the crayfish Procambarus clarkii, was characterized by Nagasawa’s team. GAMP is rich in acidic amino acids and its molecular weight was estimated to 94 kDa on SDS-PAGE and to 50.5 kDa by mass spectrometry. GAMP is suspected to play a role in the calcification process of the storage structures [72], [73] and [129]. An immunolocalization study revealed that GAMP is present into the gastroliths, as expected, and in the stomach epithelial cells responsible for the elaboration of gastroliths. Moreover, this protein (or a immunologically-related molecule) seems also present as a component of the cuticle organic matrix; except for the epicuticle, all the other layers expressed a GAMP-immunoreactivity [122]. This study also demonstrated that calcium carbonate is present in its amorphous form in gastroliths, whereas only calcite is detected in the cuticle of P. clarkii. The authors suggest that GAMP could have a function in the precipitation of calcium carbonate in the amorphous form in gastroliths.

Second, a soluble organic matrix protein, Orchestin, from the terrestrial crustacean Orchestia cavimana has been well characterized [61], [82] and [123]. Analysis of the proteinaceous components of the organic matrix of the storage forms elaborated by this crustacean revealed that this matrix obtained after dissolution in an EDTA buffer is composed of a soluble fraction (SM) and an insoluble fraction (IM) [82]. The electrophoretic pattern of the SM fraction, which contains most of the constituents, as generally observed in transitory biomineralised structures, showed about 12 polypeptides with acidic isoelectric points. Comparison of this pattern with those corresponding to proteins extracted at different stages of the moulting cycle led to evidence of a SM-specific 23 kDa polypeptide [82]. Biochemical characterization of this polypeptide, called Orchestin, revealed that it is an unglycosylated but phosphorylated calcium-binding acidic protein [123]. Furthermore, its ability to bind calcium is linked to phosphorylations on serines [61]. The corresponding cDNA and gene were sequenced [123]. The molecular mass calculated from the deduced sequence of the protein is 12.4 kDa. The discrepancy of this calculated mass with the SDS-PAGE deduced one is attributed to the richness of the protein in acidic amino acids (30%) and the presence of post-translational modifications [123]. The same kind of explanations can be applied to the other proteins listed in Table 1, with which such a divergence was observed.

The spatio-temporal expression of the gene encoding Orchestin was studied by in situ hybridisation and Northern blotting [61]. The gene is expressed only in the posterior ceca at a rate that increases from negligible during the inter-moult through the pre-moult period to culminate at the end of the pre-moult concomitant to the calcium storage process. After ecdysis, the expression remains not negligible and decreases until the basic rate is attained at the beginning of the next inter-moult. Contrarily to GAMP, no Orchestin-immunoreactivity was observed at the level of the cuticle. Recent ultrahistochemical investigations by electronic microscopy have led to the conclusion that Orchestin is also a component of the organic matrix of the transepithelial post-moult spherules of calcium resorption.

The only common feature of all these sequenced crustacean matrix proteins are that they are acidic, as generally observed for matrix proteins of calcified structures. Comparison of the sequences of these molecules does not permit to find homologies, identities, nor particular domains, except for the Rebers–Riddiford (RR) motif [105], [106] and [144]. This RR consensus region is suspected to adopt a two-stranded β-sheet conformation, which dictates the helicoidal structure of the cuticle [66], [67] and [106]. The possible presence of such a domain could lead to define two groups of cuticular proteins: those linked to the chitin filament system and the others, involved in the calcification process or simply in the structural elaboration of the matrix framework. Nevertheless, the characteristics of CAP-1 and CAP-2 suggest that a protein could be both calcium-precipitation interfering and chitin binding [70] and [71].

Whatever the origin of the calcium used for the hardening of the carapace, calcium metabolism is of great intensity in crustaceans and results from a very precise and hormonally controlled balance between the cuticle and the environment or the storage sites, coordinated with the moulting cycles. Little is known about the hormonal regulation of such calcium translocations.

In arthropods, most events of a moulting cycle are under ecdysteroid control [25], [26], [34], [87], [88], [115] and [118]. If other active ecdysteroids have been found, 20-hydroxyecdysone (20E) can be considered as the active moulting hormone in most of the crustaceans, as in many insects. The moulting hormone is generally synthesized, from cholesterol, in cephalic organs called Y-organs as an inactive precursor, the ecdysone for the most crustaceans. During inter-moult, ecdysone secretion in Y-organs is inhibited by another hormone, MIH (Moult-Inhibiting Hormone), synthesized in the eyestalks by the X-organs, a group of neurosecretory neurones, and released from an adjacent neurhemal organ, the sinus gland [25], [26], [34], [47], [78] and [115].

One of the target tissues of the moulting hormone is the epidermis. Moreover, it has been shown that this hormone has implications in the general crustacean calcification and/or decalcification processes, but the mechanism of control, probably indirect, is still unknown [46], [47], [56], [81], [88] and [139]. It is worth noting that, in general, ecdysteroid titre becomes very low just before ecdysis, which permits the synthesis of the eclosion hormone responsible for this event. This ecdysteroid level remains very low during the post-moult period. Thus, calcification of the cuticle occurring during this period is surely independent of the moulting hormone. At the molecular level, only some changes of protein expression in the cuticle epidermis have been observed after in vivo injection of the moulting hormone, 20E [80], [100], [119], [124] and [133]. Nevertheless, if some proteins stimulated by the moulting hormone have been characterized, none of these seems to be involved in the cuticle calcification or decalcification process [27] and [29].

The role of the moulting hormone in the elaboration of gastroliths has been suspected for many years [85], [86], [87] and [143], notably by eyestalk removal experiments. Ueno and colleagues [134] have clearly shown the presence of ecdysteroid-binding sites in the stomach and gastrolith epithelium (in the cytoplasm, then the nuclei) during the elaboration and resorption periods of the gastroliths. More recently, Nagasawa et al. [129] have shown, by culture of gastrolith discs with their surrounding tissues, that expression of the gene encoding GAMP, the insoluble protein from the matrix gastrolith described above [72], [73], [122] and [129], is induced by 20E.

Previous experiments have shown a possible involvement of ecdysteroids in the calcium turnover in O. cavimana [46], [47] and [114]. The ecdysteroid titre was measured at different stages of a moulting cycle in both haemolymph and whole body extracts by radioimmunoassay [53]. This titre increases during the pre-moult period concomitant to both calcium storage and expression of the gene coding for Orchestin [53] and [123]. Thus, we tested the involvement of 20-hydroxyecdysone (20E), the moulting hormone of Orchestia, in the regulation of the elaboration of concretions. First, we demonstrated that injection of 20E in animals at the beginning of the pre-moult not only induced an early ecdysis, but also, 6 h later, an electrophoretic pattern of the whole PC proteins similar to a late pre-moult one [82]. Second, study of orchestin expression showed a significant stimulation of the gene by 20E, but this effect is probably indirect as demonstrated by the inhibitory action of cycloheximide (a protein synthesis inhibitor) on this stimulation [123]. In summary, it seems that if ecdysteroids are involved in the cuticle calcification and calcium storage control, this control is certainly indirect.

The involvement of other specific crustacean hormones in calcium metabolism is not clear. Nevertheless, some vertebrate-type hormones [35] and [54] have been suspected to play a role in crustacean calcium metabolism.

For example, hormones that resemble the calcitonin (CT) and the calcitonin gene-related peptide (CGRP) of vertebrates have been detected in some crustaceans [11], [12], [22], [36], [54], [79] and [135]. Some results showing a possible involvement of these hormones in calcium metabolism were obtained.

In Orchestia, a CT-like molecule was detected and its titre calculated at different stages of a moulting cycle [54]. Effect of in vivo injection of salmon CT on calcium haemolymph rate suggests that this hormone could induce hypocalcemia [112]. A strong CT-immunoreactivity detected in the central nervous system of Orchestia, and a weaker one at the level of the posterior ceca led the authors to hypothesize that a CT-like hormone could be synthesized in the SNC as a neurohormone, and then participate to the regulation of calcium translocation in the storage organs [55].

Arlot-Bonnemains and associates have purified from a lobster a cysteine protease, which is immunologically related to salmon CT and human CGRP, which interacts with the CT and CGRP specific receptors, stimulates adenylate cyclase, and induces hypocalcemia and hypophosphatemia when injected in young rats [13 ].

Finally, other molecules that resemble hormones regulating vertebrate calcemia, such as parathyroid hormone (PTH) or vitamin D, could be active in crustaceans [81], [88], [94] and [95]. Nevertheless, the function of such molecules, which are not chemically characterized so far in crustaceans, remains unclear.

Crustaceans are a remarkable group of animals as regards the biomineralisation world. Firstly, these animals are able to elaborate cyclically a mineralised exoskeleton. Furthermore, they can also synthesize cyclically other calcified structures, not only different in their morphology but also in their mineralogical composition and in the adopted polymorph (crystalline and/or amorphous, for example). The biomineralisations elaborated as calcium storage deposits are transient structures, reservoirs of calcium ions quickly available after ecdysis, and, for this reason, the storage structures are essentially in amorphous form. To understand why these amorphous mineralisations are stabilized in time in comparison with the same very unstable purely chemical mineral is also of great interest [6] and [20]. In fact, it seems that amorphous calcium carbonate (ACC) is more widely distributed than previously suspected [2] and [138]. On the one hand, many invertebrates seem to elaborate calcium carbonate structures firstly under the ACC precursor form, which is transformed into one of the crystalline calcium carbonate polymorphs [17], [103] and [104]. On the other hand, it has been shown that ACC can coexist with one of the calcium carbonate crystalline forms: ACC and calcite coexist in well-defined domains in the exoskeleton of ascidians and calcareous sponges, for example [8].

It seems that both of these processes, transformation with time of one polymorph into another and coexistence of two polymorphs, could be involved in the calcification of the crustacean cuticle.

The organic matrices of crustacean biomineralisations are relatively accessible, notably those from the storage structures (because of their mainly amorphous form easier to decalcify than their crystalline counterparts). Unfortunately, few molecules really involved in calcification have been well characterized so far in crustaceans. Moreover, if numerous proteins begin to be sequenced in other invertebrates, as in vertebrates, their actual function is not fully understood. The knowledge of the physical and chemical features of each matrix component is a prerequisite to clarify, at the molecular level, how a biomineralisation is elaborated, how the matrix molecules influence the nature of the polymorph obtained, as suggested [1], [6], [7], [8], [15], [32] and [104], and also to determine, by means of comparative studies, why and how calcification could have emerged on Earth. The sequencing of the genes encoding these constituents could lead us to the understanding of the appearance of calcification processes and of the strategy used by evolution to construct different mineralising systems: convergence of different biological systems, adaptation to similar mineralising roles, or exaptation of an ancestral biomineral system in many separate lineages?

Other interesting characteristic of calcium metabolism in crustaceans is that calcium transport is performed by epithelia very similar to some vertebrate ones, and in this way they represent good models to understand how they function. Similarly, calcium pumps and enzymatic systems associated with the vertebrate calcium-transporting epithelia seem operational in the same kind of crustacean tissues.

Finally, crustaceans represent useful models for the understanding of hormonal regulation of the balance occurring cyclically between two sources of calcium, and more generally of hormonal control of mineralising systems. Notably, the tripartite vertebrate calcium-controlling system (CT/PTH/Vitamin D), which probably has emerged from invertebrates, could regulate crustacean calcium metabolism.