Changes in cuticle structure during growth in two terrestrial isopods (Crustacea: Isopoda: Oniscidea)

sequential layers change their orientation by a smaller angle. We found no significant differences in morphometric parameters between the cuticles of A. vulgare and P. scaber when controlling for cuticle thickness. This suggests that known differences in cuticle structure and composition between these two species may result from differences in cuticle thickness.

The cuticle covers the bodies of arthropods and forms their exoskeleton. In crustaceans, it is often not only sclerotized but also mineralized with calcium carbonates and calcium phosphates. This makes it a natural composite material with chitin-protein fibers embedded in a mineralized matrix.
Several layers are distinguishable in the crustacean cuticle (Roer et al., 2015). The epicuticle is a thin surface layer consisting predominantly of proteins and forming various scales and other surface ornaments (Schmalfuss, 1978;Price and Holdich, 1980) and containing only minute amounts of mineral . The remainder of the cuticle is divided into the exocuticle, located immediately beneath the epicuticle, and the underlying endocuticle. An additional, non-mineralized innermost layerthe membranous layer -can be identified in some crustaceans (Hild et al., 2008). Its thickness is variable and it may be completely absent in some individuals, even in species in which it is recognizable (Waugh et al., 2009;Wood et al., 2017).
In the crustacean cuticle, chitin-protein fibers change their direction in a helicoidal sequencethe Bouligand pattern -as the cuticle is deposited (Bouligand, 1972;Raabe et al., 2005;Roer et al., 2015). As a result, the cuticle has a lamellar appearance in cross-section, with each lamella corresponding to the distance in which the orientation of fibers rotates by 180°.
As crustaceans grow and develop, they require an ever-larger exoskeleton. Just as in other arthropods, the crustacean exoskeleton does not grow continuously. Instead, crustaceans periodically secrete a new cuticle and shed the old one, growing, developing and regenerating damaged structures in a series of molts.
Terrestrial isopods or woodlice (Oniscidea) have successfully established themselves in terrestrial environments. Different woodlice are found in a wide range of habitats, including amphibious as well as desert species (Hornung, 2011). As proposed by Schmalfuss (1984), woodlice with similar body types and life-styles can be classified into morphotypes which evidently evolved convergently in several lineages. Some of these morphotypes reflect different strategies of defense against predators. The rollers, which roll into a ball when threatened, are slow and have thick tergites that protect them from predators. The runners have elongated bodies and long legs and rely on speed to escape predators. The clingers have flat, oval bodies and cling to the substrate when threatened, preventing predators from reaching their vulnerable undersides.
Differences in cuticle thickness, structure and composition have been found between animals with different defense strategies and habitat preferences (Neus et al., 2007;Hild et al., 2008;Ayari et al., 2016;Wood et al., 2017;Csonka et al., 2018). While cuticle thickness in large adults in a given population is fairly consistent between specimens (Csonka et al., 2018;Khemaissia et al., 2018), little is known about the changes in the thickness and structure of the exoskeletal cuticle during the growth of terrestrial isopods. Furthermore, there is currently no information on how the ultrastructural characteristics of the cuticle, such as the relative thickness of its layers or chitin-protein fiber stacking, change as it becomes thicker.
The best studied species of woodlice are undoubtedly the roller Armadillidium vulgare (Latreille, 1804) and the clinger Porcellio scaber Latreille, 1804. Among other aspects of their biology, the structure and composition of their cuticles have been studied in more detail than in any other woodlouse (Neus et al., 2007;Hild et al., 2008;. The tergal cuticle in A. vulgare is thicker than in P. scaber when adults of similar body lengths are compared, which may reflect their different defense strategies (Hild et al., 2008;. However, it is not known if the cuticles in these species become thicker or how the structure of their cuticles changes as these animals grow.
In this study, we determined the principles of upscaling of the exoskeletal cuticle in A. vulgare and P. scaber. We performed a morphometric analysis of cuticles from individuals of different sizes. Our goals were to determine whether the cuticle thickens with growth and how the relative contributions of different cuticular layers change if this occurs. We also aimed to show whether the thickness of cuticular lamellae changes with growth. One possible mechanism by which the cuticle could become thicker would be stacking more layers of chitin-protein fibers with Cuticle structure during growth in isopods Nauplius, 28: e2020041 similar changes in fiber orientation, resulting in more numerous yet similarly thick lamellae in thicker cuticles. Alternatively, sequential planes of chitinprotein fibers might change their orientation by a different angle in thicker cuticles, resulting in a change of lamellar thickness. Our aim was to establish which of these potential mechanisms characterizes changes in cuticles thickness in the studied species.

Isopod culture
Cultures were established from 2 ovigerous females of A. vulgare and 12 adults of P. scaber of both sexes. The isopods were sampled in Središče ob Dravi in northeastern Slovenia in July 2016. They were maintained in culture at room temperature (21 ± 1 °C) and a natural light and day rhythm in Ljubljana, Slovenia. The cultures were kept in containers with soil from the sampling site and hazel leaf litter as food. Occasionally, fresh potatoes, zucchini and carrots were added to the culture as additional food items.
The offspring of collected isopods was sampled from the culture at different body sizes over the course of several months. In the case of A. vulgare, manca larvae less than 3 days after release from the brood pouch were the youngest specimens measured. Mancae were not obtained from the culture of P. scaber and the smallest individuals analyzed in this species were juveniles. Animals that did not possess sternal CaCO 3 deposits that indicate the premolt stage of the molt cycle (Zidar et al., 1998) were fixed in absolute ethanol and used for microscopic analyses. In total, 18 specimens of A. vulgare and 25 specimens of P. scaber were analyzed.

Imaging
For measurements of body lengths and cephalothorax widths of animals, fixed specimens were imaged in ethanol with an MZ FLIII stereomicroscope equipped with a DFC425 C camera (both from Leica).
For scanning electron microscopy (SEM), tergites from pereonites 2-4 were dissected, further dehydrated with two changes of absolute ethanol and air-dried. Dry tergites were attached to aluminum holders with adhesive carbon disks (SPI supplies), fractured sagitally to obtain cross-sectional surfaces and sputter-coated with platinum using an SCD 050 sputter coater (Bal-Tec). Images were obtained with a JSM-7500F field emission scanning electron microscope (JEOL).

Measurements and statistical analyses
Body lengths and cephalothorax widths of the woodlice were measured in stereomicroscope images. Cephalothorax width is an established measure for intraspecific comparisons (Sutton, 1968), as it is not inf luenced by differences in the overlap of sequential body segments. Nevertheless, the shape of the cephalothorax is a species-specific character. To facilitate interspecific comparisons, body length was chosen as a measure of body size in this study. It has been used reliably for both intra-and interspecific comparisons before (Montesanto et al., 2012;Karagkouni et al., 2016).
The thickness of the entire cuticle, the endocuticle and the exocuticle were measured for each individual on SEM images. In the endocuticle, the thickness of lamellae were measured and the measurements averaged to obtain a single value, the average lamellar thickness for that individual. The membranous layer, when present, was not included in measurements concerning the endocuticle. In the exocuticle, the thickness of the distal exocuticle (the distal smooth layer) was measured for each individual. In addition, the thickness of visible lamellae in the proximal exocuticle were determined and averaged to obtain a single average value for each individual. For each parameter, one measurement was performed per individual on images of sagittally fractured tergites in their central regions, approximately halfway along their length. All measurements were performed on pereonites 2-4 using Fiji software (Schindelin et al., 2012).
Coefficients of correlation and the corresponding significance values were determined for selected pairs of morphometric parameters. Differences between species were evaluated with ANCOVA with species as independent variables, morphometric features as dependent variables and cuticle thickness as the covariate. These analyses were performed in Past3 (Hammer et al., 2001). To compare the relationship Cuticle structure during growth in isopods Nauplius, 28: e2020041 between body length and cuticle thickness, the difference of regression slopes between species was tested with the test of regression slope differences for independent samples (Blejec, 1973).

Overview of cuticle structure
Sagittally fractured anterior tergites of A. vulgare and P. scaber (Fig. 1) are presented in Figs. 2 and 3, respectively. The cuticle in both species consists of a thin epicuticle, an exocuticle and an endocuticle, with a membranous layer of variable thickness which may also be absent. The distal smooth layer, corresponding to the distal exocuticle, can be observed just beneath the epicuticle (Figs. 2, 3). Regardless of the size of the individuals, the lamellar appearances of the endocuticle and the proximal exocuticle are discernible. It is evident that larger animals have thicker cuticles with thicker endocuticular lamellae.

Relationship between cephalothorax width and body length
In A. vulgare and P. scaber, cephalothorax width is strongly correlated with body length (r = 0.98 and p < 10 -12 for both species; Fig. 4). This is a good indication that body length is an appropriate measure of body size for both intra-and interspecific comparisons. Specimens of P. scaber grew to a body length of 12.5 mm in culture. In the case of A. vulgare, individuals grew to a maximum size of 11 mm during the study. The animals grown in culture thus reached body sizes characteristic of the species and similar to the animals that were obtained in nature and founded the culture.

Tergal cuticle thickness in relation to body size
In both species, the thickness of the tergal cuticle correlates with body length (r = 0.82 for A. vulgare (p = 6.7 x 10 -5 ) and 0.67 for P. scaber (p = 2 x 10 -4 ) (Fig. 5). The slope of the relationship between body length and cuticle thickness is greater for A. vulgare (p = 0.002).
The thickness of tergites therefore increases with increasing body length at a higher rate in A. vulgare than in P. scaber. Even in animals grown in culture and thus under similar conditions, the variability in cuticle thickness between conspecifics of similar body sizes is fairly large.
Cuticle structure during growth in isopods Nauplius, 28: e2020041  In A. vulgare, the thickness of the tergal cuticles ranged between 3.5 µm (in a 1.4 mm long manca larva less than 3 days after release) and 39.5 µm (in an 11 mm long individual; Fig. 2). In P. scaber, the thickness ranged between 4.4 µm (in a 4.2 mm long juvenile) and 21.2 µm (in a 10 mm long specimen; Fig. 3).

Changes in cuticle structure during growth
As the cuticle grows, the endocuticle thickness to cuticle thickness ratio increases (r = 0.59, p = 0.001 for P. scaber and r = 0.90, p = 8 x 10 -7 for A. vulgare; Fig. 6A). Thus, the relative contribution of the endocuticle in the tergal cuticle becomes larger in both species as they grow. As determined by ANCOVA, the percentage of cuticle thickness represented by the endocuticle does not differ significantly between the two species if adjusted for cuticle thickness as a covariate.
There is a strong correlation between the average thickness of the endocuticular lamellae and the thickness of the cuticle in both species (r = 0.86 for A. vulgare and r = 0.88 for P. scaber; p < 10 -5 for both species ; Fig. 6B). The average thickness of endocuticular lamellae ranged between 360 nm and 2880 nm in A. vulgare and between 430 nm and 1300 nm in P. scaber. The lamellae were thickest in the thickest cuticles -in large specimens of A. vulgare. There was no demonstrable difference in the thickness of the endocuticular lamellae between species when controlling for cuticle thickness as a covariate. Cuticle structure during growth in isopods Nauplius, 28: e2020041 While the correlation between the cuticle thickness and the thickness of the visible lamellae in the proximal exocuticle is significant in A. vulgare (r = 0.62, p = 0.006; Fig. 7A), there is no convincing correlation between these parameters in P. scaber (r = 0.40, p = 0.05). There is also no demonstrable correlation between the cuticle thickness and the thickness of the distal exocuticle in either of the two species (r = 0.56, p = 0.02 and r = 0.11, p = 0.61 for A. vulgare and P. scaber, respectively; Fig. 7B). Once again, no significant differences in these characteristics were demonstrated between A. vulgare and P. scaber.

discussion
As woodlice grow, their tergal cuticles grow thicker with successive molts. Among terrestrial isopods, larger individuals have been demonstrated to have thicker tergal cuticles in two species of trichoniscids as well (Vittori and Štrus, 2014). In P. scaber, comparisons have been made between the cuticles of embryos, marsupial mancae -the larval stages of peracarids not yet released from the brood pouch -and adults (Mrak et al., 2014). The conclusions were that the cuticle of marsupial mancae is structurally similar to adult cuticles and is already mineralized, but is much thinner. A similar trend of cuticle thickening with increasing body size has previously been demonstrated in benthic marine malacostracans, including a benthic isopod, whereas cuticle thickness was shown to remain constant during growth in pelagic species (Pütz and Buchholz, 1991;Waugh et al., 2009). Changes in the thickness of the integument during ontogenetic development may have a profound impact on woodlice biology. If the thickness of the cuticle determines desiccation resistance, which is likely the case (Csonka et al., 2018), a thinner exoskeletal cuticle in smaller individuals would make them more vulnerable to water loss and could influence their microhabitat preferences.
The cuticle of the roller A. vulgare is not only thicker at comparable body sizes; its thickness also increases at a higher rate than in P. scaber. The rate of cuticle thickening may be linked to the isopod's defense strategy. As a roller, A. vulgare potentially requires a greater increase of cuticle thickness at a given increase of body size if the cuticle is to perform its defensive function in larger animals.
With increasing cuticle thickness, the proportions of different layers in the cuticle change considerably. The relative contribution of the endocuticle increases, whereas the exocuticle represents an ever-smaller share of the cuticle. This is even more pronounced in the distal exocuticle, which hardly thickens at all. A similar trend has been recognized in the carapace cuticle of the blue crab, Callinectes sapidus Rathbun, 1896, in which the proportion of the exocuticle in the carapace cuticle diminishes as the cuticle becomes thicker (Waugh et al., 2009).
There are many possible consequences of these changes in layer proportions. In A. vulgare as well as in P. scaber, the exocuticle is mineralized predominantly with calcite, with amorphous calcium carbonate (ACC) present predominantly in pore canals, whereas the endocuticle is mineralized with ACC and calcium phosphate (Hild et al., 2008;Seidl et al., Figure 7. Changes in the exocuticle during growth. A, Relationship between the cuticle thickness and the average thickness of the lamellae in the proximal exocuticle in Armadillidium vulgare (circles) and Porcellio scaber (diamonds). B, Relationship between the cuticle thickness and the thickness of the distal exocuticle (smooth layer) in A. vulgare (circles) and P. scaber (diamonds).

Cuticle structure during growth in isopods
Nauplius, 28: e2020041 2012). Mineralization of the endocuticle with ACC may facilitate the resorption and potential reuse of mineral from the exoskeleton during molt, as the highly soluble ACC is predominantly resorbed from the cuticle . Temporary deposition and reutilization of minerals are possible in terrestrial isopods as ACC from the endocuticle may be deposited in sternal CaCO 3 deposits during preparation for molt (Ziegler et al., 2005). A relatively thicker endocuticle thus potentially enables the recycling of a greater percentage of the mineral in the exoskeleton. The resorption and storage of minerals may be of greater importance in thick cuticles which require large amounts of calcium, phosphorus and carbonate for their mineralization . This has previously been suggested as an explanation of the differences in composition between the cuticles of rollers and clingers, but our results show that it may be extended to differences between thicker and thinner cuticles in individuals of the same species.
A second likely consequence of the changes in relative contributions of different cuticular layers is the altering of the mechanical properties of the tergites. As the calcite-containing layers are harder and more brittle than the layers mineralized with ACC , the positioning of the calcite-containing exocuticle over the more pliable and fracture-resistant endocuticle likely increases the strength of the tergites, as the endocuticle is subjected to greater tension upon loading . This idea has not been extended to consider the effects of the relative layer thickness on the mechanical properties of the tergite as a whole, but we can expect a larger part of thicker cuticles to be softer.
At this point, we naturally cannot be certain that any of these potential adaptive advantages actually resulted in selection for the observed changes in proportions of cuticular layers with growth. It is just as possible that the exocuticle, which is deposited before molt, does not thicken much during growth due to the physiological mechanisms underlying cuticle deposition and the need for its expansion after the shedding of the old cuticle.
With the growth of the cuticle, endocuticular (and, at least in A. vulgare, exocuticular) lamellae become thicker. The lamellae of crustacean endocuticles can be interpreted as stacks of planes in which chitin-protein fibers periodically shift their orientation in accordance with the Bouligand pattern (Seidl et al., 2011). Thus, thicker lamellae result from fibers shifting their orientation more gradually, with sequential layers of fibers changing direction by a smaller angle (Vittori and Štrus, 2014). Although the mechanical effects of changing lamellar thickness in cuticles with similar composition remain to be experimentally determined, existing studies indicate that thinner lamellae result in harder cuticles. Cuticular lamellae have been shown to be thicker in benthic than in pelagic malacostracans and this has been interpreted in view of the mechanical properties of the cuticle. The proposed explanation was that thinner lamellae provide more rigidity to the relatively thinner cuticles in pelagic crustaceans (Pütz and Buchholz, 1991). Without considering possible differences in mineral content, layers of the cuticle consisting of thinner lamellae have also been shown to be harder in a decapod (Raabe et al., 2005).
Overall, we can theoretically expect thinner lamellae and a relatively thicker calcite-containing exocuticle to result in an increased hardness of thinner cuticles. Thicker cuticles, by contrast, should overall be less hard with more calcium carbonate and calcium phosphate available for mobilization during molt.
Despite the fact that A. vulgare and P. scaber are species with different defense strategies, one being a roller and the other a clinger, there are no morphometric differences in the ultrastructure of their cuticles if the effects of cuticle thickness are taken into consideration. While the tergal cuticle is consistently thicker in A. vulgare, it is apparently structured precisely as the cuticle of P. scaber would be if it were similarly thick. Previous studies have showed that the cuticle of A. vulgare contains relatively more mineral than the cuticle of P. scaber. Also, more mineral is in the form of amorphous calcium carbonate and there is less calcite (Neues et al., 2007). As the cuticle is thicker in A. vulgare, these differences in mineral composition can be explained as resulting from a consequently smaller contribution of the exocuticle, without an inherent difference in composition between the cuticles of rollers and clingers.
Nevertheless, it is known that cuticles of similar thickness are structured differently in certain species when compared to A. vulgare and P. scaber. Ayari et al. (2016) reported that the relative thickness of Cuticle structure during growth in isopods Nauplius, 28: e2020041 the exocuticle in the tergites of the desert isopod Hemilepistus reamurii (Milne-Edwards, 1840) is greater than that determined in A. vulgare and P. scaber in our study. A species that is morphologically transitional between runners and clingers, Balloniscus glaber Araujo and Zardo, 1996, has much thinner endocuticular lamellae than those measured in A. vulgare and P. scaber (Wood et al., 2017). Troglobitic species from European caves similarly have very numerous and thin endocuticular lamellae in addition to very thin cuticles for their body sizes (Vittori and Štrus, 2014). Furthermore, the pattern of calcite and ACC distribution characteristic of A. vulgare and P. scaber tergites is not universal; only roughly the external half of the exocuticle is mineralized with calcite in the cavernicolous woodlouse Titanethes albus (C. Koch, 1841) with the rest of the cuticle containing ACC (Hild et al., 2009). In the beach isopod Tylos europaeus Arcangeli, 1938, only the distal exocuticle contains calcite (Seidl et al., 2011).
The main conclusion we can draw is that the thickness and the structure of the tergal cuticle change continuously during the growth and development of terrestrial isopods. This likely affects traits such as water loss and the ability to evade predators during different life-cycle stages. The structural and compositional differences between exoskeletons of different terrestrial isopods are likely the result of an interplay of phylogeny, defense strategies and adaptations to different habitats, with animal size being an important factor as well.