Gene expression profiling of cuticular proteins across the moult cycle of the crab Portunus pelagicus

Arthropods, such as arachnids, crustaceans and insects, represent the most abundant phylum in the animal kingdom. They are common throughout marine, freshwater, terrestrial, and aerial environments. The success of arthropods can partly be attributed to the protection offered by their characteristic tough exoskeleton. With the advantage of this protective armour, however, comes the problem of growth restriction; arthropods overcome this through periodic moulting, or shedding of the exoskeleton.

The crustacean moult cycle is divided into four discrete stages; pre-moult, ecdysis, post-moult, and intermoult [1], based on the morphology of the exoskeleton. During pre-moult the underlying epidermis separates from the old cuticle (apolysis), the old cuticle is partially digested and reabsorbed, and new epi- and exo-cuticle are secreted. The old exoskeleton is shed during ecdysis. After ecdysis, during the post-moult stage, expansion of the partially formed new exoskeleton occurs followed by the tanning and mineralisation of the pre-ecdysial layers, and the deposition and hardening of the endocuticle. At the intermoult stage a mature, fully developed exoskeleton is formed [2].

The exoskeleton of crustaceans and insects is formed by cells of the hypodermis, an epithelial layer located beneath the cuticle, during both the pre- and post-moult stages [5, 6]. At the beginning of pre-moult, the hypodermis secretes a moulting fluid containing enzymes for digestion of the inner layers of the old cuticle. Secretion of a crustacean's new exoskeleton by the hypodermal cells begins during pre-moult, before the old exoskeleton is shed [2, 7]. Formation of the new exoskeleton occurs in layers. The thin epicuticle is formed first, secreted into the extracellular space between the epidermis and old exoskeleton, and a thicker exocuticle layer then appears beneath the new epicuticle during late pre-moult. At pre-moult the developing integument has invaginations over the entire surface of the animal. This increases the surface area of the new exoskeleton within the restricted space of the old exoskeleton. Ecdysis enables the expansion of the new exoskeleton [8]. Immediately after ecdysis, while the new exoskeleton is still soft, unfolding and expanding to a larger size than the old exoskeleton, the crab continues secreting the endocuticle layer beneath the new exocuticle. Mineralisation of the new exoskeleton is initiated after moulting [9]. The endocuticle layer continues to increase in thickness, while calcification and sclerotization of the new exoskeleton progresses.

The hypodermis synthesizes almost all of the proteins in the cuticle, however, hemocytes and several hemolymph proteins also contribute to the synthesis of the new exoskeleton [10]. Crustacean cryptocyanin, a member of the hemocyanin gene family, has been implicated in the transport of hormones, phenols, and some cuticular proteins to the hypodermis. Cryptocyanin itself may also be used directly as a structural component of the new exoskeleton [11–14].

The organic matrix of the crustacean cuticle is a complex structure composed mainly of α-chitin microfibrils embedded in a protein matrix [8]. The protein matrices contain a large number of proteins with different properties. These proteins can be grouped according to the type of domain that they possess. One domain type, the chitin_bind_4 domain (Pfam nomenclature), containing the Rebers-Riddiford (RR) consensus sequence, has chitin binding properties [15]. While another, the cuticle_1 domain (Pfam nomenclature), is thought to be associated with calcification as it has been found in proteins isolated from the calcified regions of the crustacean cuticle [16, 17].

Many genes that are involved in the formation of the new exoskeleton have been isolated both from insects and crustaceans. However, despite extensive research, the molecular events associated with cuticle formation in crustaceans still remain poorly understood. To investigate and gain a better understanding of this phenomenon, we set out to identify genes associated with cuticle formation, and to study their expression patterns throughout the moult cycle of the blue swimmer crab Portunus pelagicus. We used microarray technology, which provides a powerful, holistic approach to study gene expression in relation to changing physiological states.

Moulting is an important biological process in arthropods as it is essential for growth, metamorphosis and reproduction. The formation of the new exoskeleton is integral to the moulting process. Although many proteins involved in cuticle synthesis and structural integrity have been previously isolated, little is known about the expression profiles of genes related to cuticle formation across the entire moult cycle. The objective of the current study was to use a P. pelagicus cDNA microarray developed in our laboratory to identify genes (both previously identified and new) that are involved in the formation of the crustacean exoskeleton. The microarrays were produced from 5000 cDNAs expressed in both the entire animal and in specific organs such as the brain, eyestalk, MO and Y-organ from all moult cycle stages. Thus the arrays were designed to study global gene expression profiles of genes relevant to the moulting process, across all moult stages. Microarray technology offers the potential to examine the expression patterns of many genes simultaneously, thus gaining a better understanding of gene function, interaction, and regulation.

Crustacean cuticles are composed of chitin rods embedded in a protein matrix [4], the physical properties of which depend, among other things, on the sequence of the constituent proteins and the extent of mineralisation [19]. Many of the proteins associated specifically with the cuticular matrix of crustaceans can be divided into groups, based on the type of domain that they contain. One group contains a cuticle_1 domain (Pfam); proteins with this domain have previously been isolated from the hard, calcified cuticle of crustaceans [16, 17]. The restricted occurrence of this domain in calcified crustacean cuticle led to the suggestion that it could be involved in the calcification process, either as a nucleation factor for crystal formation or in regulating the growth and size of the calcium carbonate crystals once they have been formed [16]. We have isolated and profiled the differential expression of 13 unique transcripts containing such a domain across the moult cycle of P. pelagicus. Another group contains the RR consensus sequence (chitin_bind_4 in Pfam), demonstrated to be involved in chitin binding [15]. Proteins with chitin_bind_4 domains have previously been isolated from both calcified and un-calcified crustacean cuticle [16, 20], and also from the cuticles of insects [21, 22]. In the present study, four transcripts with the chitin_bind_4 domain were found to be differentially expressed across the P. pelagicus moult cycle. We have also isolated four unique, differentially expressed transcripts containing a domain that as yet remains unannotated, but has previously been isolated from the cuticle of C. pagurus [16], tentatively termed PfamB_109992 (Pfam). It is likely that the different domain types found in cuticle proteins are functionally relevant to the cuticles in which they occur. The expression of cryptocyanin was also found to be moult-cycle related. Unlike the other cuticle proteins, cryptocyanin is a hemolymph protein secreted by the hepatopancreas that has also been implicated in the formation of post-ecdysial cuticle [14].

Proteins with the cuticle_1 domain have previously been isolated from the calcified cuticle of two decapod crustaceans; eight from the crab Cancer pagurus [16] and 10 from the lobster Homarus americanus [17]. All of the proteins previously isolated contain either 1 or 2 copies of cuticle_1 domains. In our study, transcripts with up to four cuticle_1 domains (see Figure 2) were isolated. Six transcripts containing the cuticle_1 domain isolated in this study have a deduced signal peptide, and four of these also contain a deduced transmembrane region, suggesting that they are secreted across the membrane (Figure 2). The lack of a signal peptide in the other transcripts may be due to an incomplete cDNA sequence, therefore the presence of a signal peptide in all cuticle_1 domain-containing transcripts cannot be ruled out. Each cuticle_1 domain consists of two repeated 15 amino acid motifs with a spacer sequence typically 9 to 12 amino acids in length. The alignment in Figure 3 has been created from the cuticle_1 domain-containing proteins, isolated from C. pagurus and H. americanus, and from 13 unique differentially expressed cuticle_1 domain-containing transcripts (conceptually translated) isolated in this study. Figure 3 depicts the location of the cuticle_1 domain (denoted by shading). While the cuticle_1 domain is common to all the sequences, many regions outside of this domain do not share homology. Figure 4 is a phylogram of the proteins described above for Figure 3. A phylogram is a branching diagram (tree) assumed to be an estimate of a phylogeny, branch lengths are proportional to the amount of inferred evolutionary change [23]. Figure 4 depicts the hypothesised branching order of the cuticle_1 domain-containing sequences where the branch lengths are proportional to the amount of inferred evolutionary change. The tree suggests that the cuticle_1 domain-containing proteins do not group according to species, implying that these sequences have evolved to have distinct functional roles that are conserved between species. The sequence variability seen in Figure 3 together with the apparent functional grouping viewed in the phylogram, points to a functional variation of the proteins despite their shared domain.

Of the 13 P. pelagicus transcripts containing the cuticle_1 domain, described above, two transcripts, PpCUT12 and PpCUT13, were up-regulated (eightfold) in the post-moult stage when compared to intermoult. The 11 remaining transcripts, PpCUT1-11, displayed a combined average up-regulation of 9.6-fold in the intermoult stage when compared to crabs in pre-moult. Four of these transcripts, PpCUT7-10, were also down-regulated in post-moult when compared to intermoult by an average of 6.2-fold. Additionally, six transcripts (PpCUT1, 4, 5, 9, 11 and 12) were down-regulated in pre-moult when compared to crabs at ecdysis. No differentially expressed transcripts were detected between the early and late pre-moult stages or between ecdysis and post-moult. For a graphical representation of these data see Table 8 and Figure 5. From these data we can deduce that the up-regulation of many transcripts containing the cuticle_1 domain begins at ecdysis; expression then differentiates into two groups, those transcripts up-regulated in post-moult compared to intermoult and those down-regulated in post-moult compared to intermoult (Figure 5a and Table 8). This converse expression profile of transcripts with the cuticle_1 domain further suggests that a functional and perhaps regulatory difference exists, even between transcripts containing the same domain. The proteins represented by these cuticle_1 domain-containing transcripts may have different mechanisms or modes of action that facilitate their respective roles in exoskeleton formation. High levels of expression of all cuticle_1 domain transcripts (except PpCUT12 and 13) in the intermoult stage compared to early pre-moult (Figure 5b) indicate that formation and/or repair of the exoskeleton continue well into intermoult while expression of these transcripts then decreases dramatically during pre-moult. Multiple proteins containing the cuticle_1 domain have previously been identified in crabs. However, until now, the expression patterns of their corresponding genes across the moult cycle have not been traced. Although the functional role of these proteins, or specifically the cuticle_1 domain, has not been postulated, the domain is thought to be associated with calcification as it has only been identified in proteins isolated from the calcified cuticle of crustaceans [16]. The high level of expression of PpCUT12 and PpCUT13 in post-moult (Figure 5a), and the up-regulation of the other transcripts containing cuticle_1 domains in the intermoult stage (Figure 5b), further support the proposal that proteins with this domain confer moult cycle related changes to the calcified cuticle of crustaceans.

Many proteins containing the chitin_bind_4 domain have been isolated from the cuticle of both insects and crustaceans. In addition to their role in chitin binding [15], several crustacean proteins containing this domain also appear to participate in the calcification of the exoskeleton. Calcification associated peptide (CAP) 1 and 2, isolated from the crayfish Procambarus clarkii, are multifunctional peptides with anti-calcification, calcium binding and chitin-binding properties [24–26]. Crustocalcin from the prawn Penaeus japonicus on the other hand, promotes calcification in addition to having calcium binding properties [27, 28]. Despite the commonality of the chitin_bind_4 domain in these proteins, the promotory role of crustocalcin in the calcification process contrasts with the inhibitory properties described for CAP 1 and 2. To investigate the levels of homology between proteins containing the chitin_bind_4 domain, a phylogram (Figure 6) was constructed from the seven (conceptually translated) transcripts isolated in this study (four of these were differentially expressed), CAP 1 and 2, crustocalcin, and two cuticle proteins isolated from the crab Callinectes sapidus, all containing the chitin_bind_4 domain. The sequences of the chitin_bind_4 domain-containing proteins appear not to group according to species. Variable regions to either side of the conserved chitin_bind_4 domain are apparent in Figure 7 (the chitin_bind_4 domain is shaded). This points to functional specificity of the sequence outside the domain that is likely to influence the role of the chitin_bind_4 domain in the calcification process. Three of the four differentially expressed transcripts containing chitin_bind_4 domains isolated in this study contained a deduced signal peptide, and two of these also contained a deduced transmembrane region (Figure 8), indicating that they are secreted across the membrane.

Moult cycle related differential gene expression was observed in four unique transcripts containing the chitin_bind_4 domain, PpCB1-4, in this study. These transcripts were highly up-regulated (12-fold) in the intermoult stage when compared against early pre-moult (Figure 5b). We found that only two of these transcripts (PpCB1 and PpCB3) were up-regulated in ecdysis when compared to pre-moult (Figure 5c). The lack of differential expression of transcripts containing the chitin_bind_4 domain between the post-moult and intermoult stages indicates that both stages display similar expression levels of these genes. The up-regulation of chitin_bind_4 transcripts in the intermoult stage compared to pre-moult suggests that these genes are also expressed in post-moult, and that expression is significantly reduced in pre-moult. The high level of expression in intermoult is perhaps unexpected, certainly in terms of exoskeleton formation, which is traditionally associated with the pre- and post-moult stages. However, this expression pattern reflects the expression of other chitin_bind_4 domain-containing genes, CsAMP8.1 and CsAMP6.0 of C. sapidus, whereby their expression in arthrodial membranes continued for 32 days post-moult before disappearing [20], apparently due to the late deposition of the un-calcified membranous layer within 16 to 32 days post-moult. The down-regulation of chitin_bind_4 transcripts in the pre-moult stage of the moult cycle when compared with ecdysis, strongly suggest that at least some members of the chitin_bind_4 domain family of proteins are not involved in the synthesis of pre-ecdysial cuticle in crustaceans but rather, are "switched on" at the time of moulting in preparation for the generation of the post-ecdysial, or the hardening of the pre- and post-ecdysial layers. This is consistent with findings for other chitin_bind_4 domain-containing proteins such as CAP 1 and 2, and crustocalcin, which show that expression of their genes occurs in the epidermal tissue only during the post-moult stage [24–28].

Four transcripts containing the domain PfamB_109992, which as yet remains unannotated, were found to display moult cycle related differential expression profiles in P. pelagicus. The PfamB_109992 domain was previously found in a protein isolated from the calcified cuticle of C. pagurus (CPCP1876, accession number P81584) [16], and also in the calcified cuticle protein of C. sapidus (CP15.0, ABB91679). An alignment of their amino acid sequences (Figure 9), and the phylogram in Figure 10, indicates that they do not group according to species, suggesting that they represent distinct genes, perhaps clustering according to function. The sequence homology may result from conserved functionality. PpBD1 and 2 both contain a deduced signal peptide, while only PpBD1 contains a deduced transmembrane region (Figure 11), suggesting possible secretion of these proteins across the membrane. PpBD1-4 were up-regulated (combined average of 6.4-fold) in the post-moult stage when compared against intermoult (Figure 5a), suggesting that they are relevant to cuticle formation or to changes in the physical properties of the exoskeleton observed after ecdysis. Two of these transcripts, PpBD1 and 2, were also highly up-regulated (by a factor of 12-fold) in the intermoult stage when compared against crabs in the pre-moult period (Figure 5b), signifying their continued expression well into intermoult followed by a decrease in expression during pre-moult. The difference in expression between transcripts containing the PfamB_109992 domain in the intermoult stage when compared to pre-moult is suggestive of a divergence in functionality or regulatory mechanisms. To date, little work has been carried out on proteins containing this domain. However, moult cycle related differential expression evident from the present study, and the previously implied specificity of this domain to the calcified cuticle of crustaceans, signal the importance of proteins carrying this domain to exoskeletal formation.

PpCBM, a transcript sharing sequence homology with a peritrophic membrane chitin binding protein from the cabbage lopper Trichoplusia ni (AAR06266), was found to be 6.2-fold up-regulated exclusively in the post-moult stage when compared to intermoult. This transcript contains one chitin binding Peritrophin-A domain (CBM 14) and is found in chitin binding proteins, particularly peritrophic matrix proteins of insects [18]. In insects, these proteins are important determinants for the structural formation and function of the peritrophic membrane, which lines the gut of arthropods and protects it from digestive enzymes [18]. In crustaceans, the mucus layer that coats the undigested products in the stomach and assists in their movement through the digestive tract is also referred to as the peritrophic membrane [29]. Within crustaceans, sequences with the CBM 14 domain have to date only been found in penaeid prawns. The specific up-regulation of the peritrophic membrane chitin binding protein transcript in post-moult, demonstrates its increase in expression during a period associated with cuticle synthesis and hardening. A chitinous portion of the stomach of crabs, and other higher crustaceans, is shed along with the exoskeleton at ecdysis [29], which necessitates the synthesis of a new chitinous matrix in the stomach post-ecdysis. Proteins such as those containing the CBM 14 domain, which has been demonstrated to bind chitin [30], may be involved in the formation and/or structural integrity of the digestive tract.

Moult cycle related differential expression profiles were also observed for three transcripts, PpVER1, PpVER2 and PpVER3, which displayed sequence homology to a LDLa domain containing chitin binding protein from Drosophila (Accession number NP_730442), termed vermiform. PpVER1 was up-regulated in both the post-moult stage compared to intermoult, and during intermoult when compared to early pre-moult. Whereas PpVER2 was down-regulated in late pre-moult when compared to ecdysis, while PpVER3 was up-regulated in only in the post-moult stage when compared to intermoult. The specific up-regulation of each of these transcripts during various stages of the moult cycle, points to a functional/regulatory difference between these genes, and implicates their involvement in distinct aspects of cuticle formation. Vermiform, a gene involved in chitin modification, was found to affect the structural properties of the chitinous matrix of the trachean cuticle of Drosophila [31]. The authors demonstrated that this gene is not required for chitin synthesis, secretion, or accumulation but rather for its normal morphology and structure. Vermiform codes for a protein with a LDL-receptor ligand binding motif and chitin binding and deacetylation domains. Chitin deacetylase modulates the physical and chemical properties of chitin by deacetylation, which converts chitin into chitosan [32], and may influence the structure and orientation of chitin fibrils in the arthropod cuticle. Luschnig et al [31] demonstrated that vermiform is expressed in epidermal cells of Drosophila, and that mutations in this gene affect body shape, presumably by altering the structure and rigidity of the epidermal cuticle. Further cuticular abnormalities such as reduced procuticle deposition and aberrant apical membranes were also observed in Drosophila vermiform mutants [33]. Proteins with chitin binding properties are important in the structural modifications occurring in the post-ecdysial cuticle. This is the first report of vermiform like proteins in crustaceans, and their expression patterns indicate that they have a role in the structural modifications occurring in the post-ecdysial cuticle

In contrast to the cuticle proteins discussed above, cryptocyanin was highly up-regulated in the late pre-moult period (9.2-fold) when compared to crabs in ecdysis, but was down-regulated in post-moult (9.6-fold) when compared to intermoult. This indicates that levels of cryptocyanin transcript decrease at the time of ecdysis and continue to remain low throughout the post-moult period. Cryptocyanin expression then increases in intermoult and remains stable during the pre-moult period. These results concur with those of another study in C. magister, in which mRNA levels of cryptocyanin were low immediately post-moult and gradually increased to a peak two thirds into the moult cycle then decreased prior to moulting [14]. The authors demonstrated that similar patterns were exhibited by cryptocyanin protein levels in the hemolymph, where levels remained low in intermoult, increased in pre-moult and dropped dramatically in post-moult [14]. Immunohistochemical and mRNA expression studies in C. magister, demonstrated that cryptocyanin is synthesised in the hepatopancreas, secreted into the hemolymph, transferred across the epidermis and incorporated into the extracellular matrix of the new exoskeleton [14]. The converse expression profiles of cryptocyanin and cuticle protein transcripts across the P. pelagicus moult cycle observed in this study suggest that they are involved in chronologically different stages of cuticle formation. However this temporal delay is likely due to the different sites of synthesis of each protein, as both protein types play a role in post-ecdysial cuticle formation. The time required between cryptocyanin synthesis and secretion from the hepatopancreas, transportation through the hemolymph and final incorporation into the cuticle, may be responsible for the earlier expression of cryptocyanin mRNA compared with that observed for the other cuticle proteins in this study.

Our study describes the temporal expression patterns of cuticle protein transcripts containing the cuticle_1, chitin_bind_4, PfamB_109992, and CBM14 domains, as well as the hemolymph protein cryptocyanin, across the moult cycle of P. pelagicus. Most of the differentially expressed cuticle protein transcripts identified in this study were up-regulated in the intermoult period when compared to the pre-moult. This suggests that synthesis and/or repair of the cuticle occurs well into the intermoult stage of the moult cycle, and that these transcripts are down-regulated in the pre-moult stage and hence not involved in the synthesis of the pre-ecdysial layers of the exoskeleton. Instead these genes may be involved in the synthesis of the post-ecdysial cuticle, or in the exoskeletal hardening process associated with post-moult. None of the cuticle transcripts identified in this study were up-regulated in the pre-moult period, reinforcing the supposition that they are not involved in pre-ecdysial cuticle formation. Only the hemolymph protein cryptocyanin, displayed up-regulation in pre-moult, its involvement in cuticle formation however, also occurs post ecdysis [14].

Temporal variation in individual transcript expression, both within those transcripts that contain the same domain and between transcripts containing different cuticle domains, was observed in the present study. Wynn and Shafer [20] also demonstrated differential temporal and spatial expression between four genes containing the chitin_bind_4 domain in both calcified and un-calcified cuticle, in C. sapidus. This fluctuation in expression levels between cuticle proteins containing the same domain type and the presence of various domains in cuticular proteins indicates that numerous genes are involved in the formation of the crustacean exoskeleton, that they are not up-regulated simultaneously, and that each may have a distinct role in cuticle formation. The moult cycle-related differential expression observed for transcripts containing the same domain type indicates a difference in functionality and perhaps regulation of the corresponding proteins in the crustacean cuticle. We therefore propose that the identification of cuticle proteins be based on the type of domain contained within the protein, rather than the type of cuticle (be it calcified or un-calcified) from which the protein was isolated. As more information becomes available about the role of these domains in the arthropod exoskeleton, the features present within a cuticle protein will become pivotal in describing its function.

Tracing the temporal expression patterns of genes involved in cuticle formation assists in elucidating the mechanisms of cuticle synthesis. Expression information about such genes aids in their functional annotation, as differential gene action underlies the regulation of differential protein accumulation in the cuticle [19]. The expression data presented here provide a chronological depiction of moult cycle related changes to the synthesis of proteins involved in cuticle formation. We have identified a number of candidate genes, by virtue of domain annotation and differential expression data, that could play an important role in the formation and hardening of the crustacean exoskeleton throughout the moult cycle. The expression pattern of these genes, together with conceptual domain annotation, has enabled the discovery of new genes likely to be important to cuticle formation that contain similar motifs to those identified previously. It is evident that cuticle formation and hardening are complex processes involving many components and requiring strict regulatory mechanisms.