Mechanical defensive adaptations of three Mediterranean sea urchin species

Abstract In the Mediterranean, Paracentrotus lividus and Sphaerechinus granularis are important drivers of benthic ecosystems, often coexisting in sublittoral communities. However, the introduction of the invasive diadematoid Diadema setosum, which utilizes venomous spines, may affect these communities. To describe the mechanical properties of the test and spines of these three species, specimens were collected in winter of 2019 from the sublittoral zone of the Dodecanese island complex, southeastern Aegean Sea. This region serves as a gateway for invasive species to the Mediterranean Sea. Crushing test was conducted on live individuals, while 3‐point bending test was used to estimate spine stiffness. Porosity and mineralogy of the test and spine, thickness of the test, and breaking length of the spine were measured and compared, while the microstructural architecture was also determined. The test of S. granularis was the most robust (194.35 ± 59.59 N), while the spines of D. setosum (4.76 ± 2.13 GPa) exhibited highest flexibility. Increased porosity and thickness of the test were related to increased robustness, whereas increased flexibility of the spine was attributed to high porosity, indicating that porosity in the skeleton plays a key role in preventing fracture. The spines of S. granularis exhibited highest length after fracture % (71.54 ± 5.5%). D. setosum exhibited higher values of Mg concentration in the test (10%) compared with the spines (4%). For the first time, the mineralogy of an invasive species is compared with its native counterpart, while a comparison of the mechanical properties of different species of the same ecosystem also takes place. This study highlights different ways, in which sea urchins utilize their skeleton and showcases the ecological significance of these adaptations, one of which is the different ways of utilization of the skeleton for defensive purposes, while the other is the ability of D. setosum to decrease the Mg % of its skeleton degrading its mechanical properties, without compromising its defense, by depending on venomous bearing spines. This enables this species to occupy not only tropical habitats, where it is indigenous, but also temperate like the eastern Mediterranean, which it has recently invaded.


| INTRODUC TI ON
Echinoids are a group of organisms belonging in the phylum echinodermata. They are important grazers in most marine benthic sublittoral communities (Sala et al., 1998), which can change the type of the ecosystem they live in. Predation is one of the main factors, which control their density and structure population (Shears & Babcock, 2002). To defend against predators, sea urchins utilize their calcareous skeleton, which consists mainly of the test covered by epidermis and the spines, in various ways.
Among all echinoids, diadematoids and echinothurioids are the only ones acquired with venomous spines. However, it is not yet known whether this evolutionary trait is related to decreased skeletal development, leading to decreased robustness (Koch et al., 2018).
Diadematoids have also distinct skeletal morphological characteristics both for the test and for the spines (Coppard & Cambell, 2004. These may also provide them with distinct mechanical properties (Burkhardt et al., 1983).
Diadema setosum (Leske, 1778) diverged first from the other extant Diadema in the Miocene. This clade then split into two clades, one around the Arabian peninsula and the other in the Indo-West Pacific (Lessios et al., 2001). Both environments are considered tropical, characterized by high annual temperatures. However, the recent increase in temperature of the Mediterranean Sea seems to accelerate the introduction of invasive tropical species (Bianchi, 2007). The first report about the occurrence of D. setosum in the Mediterranean dates in 2006 (Yokes & Galil, 2006). Established populations of this species have been observed in the Aegean Seas (Vafidis et al., 2021). Paracentrotus lividus (Lamarck, 1816) and Sphaerechinus granularis (Lamarck, 1816) are two species belonging to the order camarodonta, which play a key role in the sublittoral Mediterranean ecosystems. These species often coexist, with S. granularis occurring mainly in soft substrates and P. lividus on rocks and boulders . Seagrass meadows of Zostera marina and Posidonia oceanica appear to be their main habitats (Antoniadou & Vafidis, 2014).
The sea urchin skeleton is composed of ossicles made up of a trabecular meshwork, named the stereom, which in turn consists of high-magnesium calcite (Weiner, 1985). The concentration of Mg in the calcite may differ in test and spine (Smith et al., 2016).
Differences in Mg concentration seem to grant the skeletal elements different mechanical properties. Specifically, it is reported that higher magnesium concentration seems to make the structure stronger and more rigid (Weber et al., 1969). The Mg/Ca ratio is affected by temperature and salinity, meaning there might be intraspecific differences in Mg concentration in different environments (Borremans et al., 2009). Variations of the porosity of the material also seem to play a key role in its mechanical properties (elasticity, hardness, etc.; other factors like the stereom architecture may further affect these properties; Lauer et al., 2018).
The three studied species play a key role in the mid-littoral and upper sublittoral zone. There are a limited number of studies related to the functional role of the skeleton of sea urchins, regarding the Mediterranean Sea, where attachment tenacity, spine length and thickness, and robustness of the test of P. lividus, A. lixula, or S. granularis in western Mediterranean are discussed (Collard et al., 2016;Di Giglio et al., 2020;Guidetti & Mori, 2005;Santos & Flammang, 2007). However, there are no studies regarding the eastern Mediterranean, which is characterized by peculiar oceanographic conditions. Furthermore, the study of an invasive sublittoral species, which heavily depends on its venomous spines for protection, is very important not only because it may further explain differences between mechanical and chemical adaptations but also because it is possible that this species may alter the dynamics of the Mediterranean sublittoral ecosystems, beginning from the eastern Mediterranean. This study aims to assess and explain the mechanical differences among two native camarodonts and an invasive diadematoid from an ecological standpoint, hypothesizing that D. setosum will exhibit inferior mechanical properties, as it mainly depends on its venom for protection. This might enable it to better adapt to environmental changes, as a degeneration of its skeleton due to environmental factors might not be a delimiting factor regarding its adaptive capability.  Byrne et al. (2014) was followed. Briefly, to determine the force needed to crash the test, ten live individuals from each species were crushed, using an INSTRON 3382 Universal Testing Machine of a load capability of 100 kN (22400 lbf). The device is equipped with a data acquisition system controlled from Instron Bluehill ® Lite Software to record load and displacement. A circular plate (15 cm diameter) was attached to the motorized tester so that an even force was applied to the test. The speed of the tester was set at 100 mm min −1 . At the end of the experiment, the thickness of each test was measured by haphazardly selecting pieces of the test from the ambital region of each individual (n = 10 per species). The thickness of three pieces per individual was measured using an electronic caliper to the nearest 0.01 mm, and the average value was used in the analysis.

| 3-point bending test
To describe the stiffness of the spines, the bending modulus of elasticity (Young's modulus) was determined. The arrangement of the samples was parallel to the central axis of the spine. Ten ambital primary spines per species (one per individual) with no trace of regeneration were used for the determination of the elastic module.
Force was built up with a non-cutting blade. It was impossible to use a pin, due to the small thickness of the examined spines. The bending tests were conducted on a FTC Mechanical Testing Device 25 of a load capability of 25 KN equipped with appropriate three-point bending grips. Feed motion was set at 5 mm s −1 . Force (N) and deflection (mm) were monitored. The second moment of inertia (I, mm 4 ) was determined by the software ImageJ (1.53f51), using the BoneJ plugin, taking the porous nature of the spines into account (Doube et al., 2010). The equations for the calculation of stress, strain, and Young's modulus are shown below: with σ f the stress (MPa), F the force at fracture (N), L the active length (mm), R the radius of the spine at fracture (mm), ε f the strain, D the deflection at fracture (mm), d the thickness of the spine (mm), and E the Young's modulus (MPa). The active length for D. setosum was set at 33.75 mm, while for P. lividus and S. granularis at 10.085 mm, due to their smaller length. The length before and after fracture was measured to determine the percentage of fracture.

| Scanning electron microscopy-Porosity
Scanning electron microscopy (SEM) was carried out on a JEOL JSM 6510. After the mechanical tests, fragments from the fracture site of the interambulacral region of the test were retrieved and together with the lower section of the fractured spines were prepared for SEM. At first, the samples were bleached to remove the soft tissues, following the method of Collard et al. (2016), NaOCl 2.5% for 1.5 h, then NaOCl 5.25% for a further 2.5 h, and air-dried for 24-48 h. The specimens were then mounted on metal stabs with carbon-based tape and coated with carbon. The inner surface of 10 plates and the cross and longitudinal sections of each spine (n = 10) were examined, using a 10 kV acceleration voltage with a 26-31 mm working distance, and secondary electron images were taken. SEM micrographs were analyzed for porosity (%) by calculating the ratio of pore area to total area using ImageJ software, as proposed by Schneider et al. (2012). The mean porosity for each species was then determined and compared. Morphological descriptions for the stereom of the tests and spines were carried out using the terminology of Smith (1980).

| X-ray powder diffraction analysis
An X-ray powder diffraction analysis (XRD) of the test and spines of D. setosum was performed on a D8 Advance-BrukerAXS diffractometer using CuKα radiation to determine the crystalline phase both quantitatively and qualitatively. Interambulacral plates and spines from all individuals of D. setosum were triturated in an agate mortar until a fine state (<40 µm) and consequently pooled into one homogenous powder. Afterward, 1 g of powder was placed in a standard sample holder and was mounted in the X-ray diffractometer. Measurements were carried out by a LynxEye detector with Nifilter, operated at the voltage of 35 kV, and the intensity of 35 mA, at a 2θ scanning range of 4-70; analyses were made at a step of 0.02 • and a speed of 0.2 s per step. The evaluation of data was carried out with the Diffracplus EVA-BrukerAXS software. Identification of the experimental data was performed by fitting the diffraction pattern using Crystallography Open Database.

| Statistical analysis
Analysis of variance (one-way ANOVA) was used to examine differences in morphometric characteristics (i.e., test diameter, test height, test thickness, spine % length after fracture), mechanical properties (i.e., crushing force, Young's modulus), and porosity %. Prior to analysis, data were tested for normality with the Anderson-Darling test, while the homogeneity of variances was tested with Cohran's test and, when necessary, data were logtransformed. The Tukey test was used for post hoc comparisons.
ANOVAs were performed using the SPSS software package (IBM

| Morphological characterizations
The three species exhibit differences in their morphological features both in the test and in the spine. Not only the stereom but also their microstructures vary, but there are also similarities worth mentioning.

| Test
Observing the external surface of the interambulacral plates of the three species, morphological differences of their primary tubercles are evident (Figure 1a). D. setosum possesses perforate and crenulate tubercles, while those of the two native species appear to be imperforate and non-crenulate. Regarding the stereom microstructure, the superficial layer seems to be different between P. lividus (dense labyrinthic) and S. granularis and D. setosum (galleried). Notice that the galleried stereom of S. granularis is denser compared to D. setosum. The middle layer of all species seems to be made of galleried stereom. The cross sections highlight the alignment of the galleries with the stereom, which follow a parallel orientation to the test surface ( Figure 1b). Finally, the inner surface of the test appears denser in P. lividus, exhibiting a rather compact, simple perforate stereom.
D. setosum and S. granularis, on the other hand, possess a labyrinthic inner stereom, which is rather coarse in both species but thicker in S. granularis (Figure 1c).

| Spine
Concerning the texture of the external surface, differences in the spines of the three species are easily distinguishable. The verticillations of the spine of D. setosum may play a functional role since the spines of this species are difficult to remove once they pierce through skin (authors' observation). This may enhance the chemical defense of this species through elongated time of exposure to toxin. S. granularis exhibited small, underdeveloped barbs, while the spines of P. lividus were smooth. Regarding the morphology of the spines, two main differences can be detected. First, the external surface of P. lividus is smooth (Figure 2d

| Mechanical tests-Morphometrics
The three species exhibited significant differences for all parameters in both the test and the spines (Appendix S1). Specifically, highest force increase and breakage at maximum input. However, S. granularis showed microfractures before maximum input of load (Figure 4).

| Test and spine morphology
Comparison of SEM micrographs showed differences in the micromorphology of test and spines among the three species. It should be noted that many previous studies have been concerned with the skeletal micromorphology of the examined species (Burkhardt et al., 1983;Régis, 1979Régis, , 1986Regis & Thomassin, 1982;Smith, 1984), although the examined specimens did not originate from the eastern Mediterranean Sea. However, no interspecific differences were observed regarding the type of the trabecular meshwork. External differences regarding the test include the tubercles, which, in D.
setosum are perforate and crenulate, a characteristic feature of the genus Diadema. This perforation serves as a pathway for the central ligament to penetrate the spine and may provide increased mobility of the spine as pointed out by Motokawa (1983). Moreover, since the spines of this species are long in relation to its test, crenulation of the tubercles serves as attachment of a thick and more developed catch connective tissue (Motokawa & Fuchigami, 2015) which also accounts for the rapid movements of the spines. The two native species, one the other hand, have solid non-crenulated tubercles. These may provide the ball socket joint with more stability. In other words, D. setosum presents perforation leading to increased mobility and crenulation to possibly increase the angular range of motion of the spine. On the other hand, the two camarodonts, lacking venomous spines, do not exhibit perforations. Instead, during impact the ball socket joint needs to remail firm, so smooth tubercles may be most suitable. Since the catch connective tissue plays a key role in changing the mechanical properties of the ball socket joint (Wilkie, 1996), the functional morphology of the tubercles is directly related to the activity and mutability of this tissue. More research on this aspect needs to be conducted, as it is evident that different species use spines for different purposes and in different ways.
Regarding the stereom of the test of the three species, that of P. lividus appears to be denser both externally and internally, with indicates that its defense does not involve piercing the predator.
Rather than that this species utilizes its spines to minimize the impact generated by an attack on the test, and this seems to be the main reason why this species prefers maintaining most of the spine during fracture, rather than breaking it. Thus, the spines of S.
granularis function in such a way that the impact is spread among as many spines as possible, resulting in reduced pressure per spine. For that purpose, long, brittle spines would not be suitable.
External barbs seem to play a role in maximizing the distribution of the impact inflicted on the spine, by providing "crack deflecting

| Robustness of test-Stiffness of the spine
The test of S. granularis appears to be able to withstand significantly more load than the other two species. Guidetti and Mori (2005) found the test robustness of P. lividus to range from 11.57 (1180 g) to 109.64 N (11180 g), while test thickness ranged from 0.12 to 0.90 mm. Present results exhibit higher values for P. lividus both for the load and thickness, meaning that an increase in thickness results in a more robust test in the case of P. lividus. D. setosum and P. lividus did not exhibit any significant differences in their ability to withstand load, but the test of the former was thinner and more porous. This indicates that this species builds less material but in a more efficient way and that porosity may play a more important role than thickness in the mechanical design of the test of this species. It is important to note that the mechanical properties of the skeleton of the test are not the only factor that determines the strength of the organism against a potential mechanical attack. Ellers et al. (1998) showed that the sutures that hold the skeleton in place play a key role in the mechanical tenacity of the test. These sutures are comprised of soft tissue, specifically collagenous mutable connective tissue, and provide the test with more strength. D. setosum may utilize these sutures to gain robustness, making up for its decreased thickness.
However, this study is meant to deal exclusively with the mechanical properties of the skeleton and how these are utilized by the three examined species for different purposes.    (Bischoff et al., 1987). Numerous studies have been conducted describing the Mg % in the calcite regarding P. lividus, ranging between 11.02 and 12.9% for the test and 2.9 to 3.9% for the spine (Table 1). However, there is only one study regarding the Mg concentration in the skeleton of S. granularis. This species exhibits the highest Mg concentration among the three studied echinoids regarding the spine and lowest regarding the test (

| CON CLUS ION
The species examined in the present study are two common camarodonts of the eastern Mediterranean and one invasive diadematoid.
Micromorphological aspects, morphometrical features, mechanical TA B L E 1 Mg content % in calcite of the test and spine regarding the three studied species properties, and skeletal mineralogy were compared in order to explain not only the defensive adaptations of these species, but also the invasion of D. setosum in the eastern Mediterranean Sea. It is evident that this species depends on its venomous baring spines in order to be able to inhabit environments of lower temperature, where the skeletal Mg concentration is decreased compared to the tropics, leading to degraded mechanical properties. In this way, it does not compromise its defense against predation. S. granularis is acquired with globiferous pedicellariae, which provide a level of chemical defense, but does also depend on its highly robust test to withstand mechanical impacts. Furthermore, its obtuse spines exhibit a tendency to maintain most of their length after fracture, which may have led to this species being able to inhabit soft substrates. P. lividus mainly depends on its mechanical properties to defend against predation. This might be linked with lower energetic costs, as the production of toxins is known to be metabolically expensive. This is the first study to deal with a direct comparison of three different sea urchin species in terms of how differences in their mechanical properties might explain their different defensive adaptations. However, there are certain aspects that need to be further examined, the most important being the energy costs of these defensive mechanisms, and how they affect the population dynamics of this species. Furthermore, more research needs to be conducted as to how soft tissue affects the mechanical properties of sea urchins in an interspecific level.

ACK N OWLED G M ENTS
The authors thank Anastasios Koutsidis for his technical contributions during the mechanical experiments.

CO N FLI C T O F I NTE R E S T
The authors declare no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data used for this manuscript are uploaded to Dryad: https://doi. org/10.5061/dryad.d51c5 b03b.