Sea ice protects the embryos of the Antarctic sea urchin Sterechinus neumayeri from oxidative damage due to naturally enhanced levels of UV-B radiation

In the past, the Antarctic marine environment has been exposed to relatively low levels of ultraviolet-B radiation (UV-B) due to its polar location, seasonal sea ice coverage and high atmospheric ozone (O₃) concentrations (McKenzie et al., 2003; Karentz et al., 2004; Lesser et al., 2004; Lamare et al., 2007). However, in recent years two relatively rapid changes to the global climate have occurred that are causing biologically important changes to the Antarctic environment.

Firstly, a reduction in stratospheric ozone has resulted in increased levels of UV-B reaching the Earth's surface (Madronich et al., 1998). The greatest loss of stratospheric O₃ has occurred over the Antarctic continent and the surrounding Southern Ocean, where mean austral springtime O₃ concentrations are now 40–50% lower than they were in the late 1970s (Malloy et al., 1997; Balis et al., 2009). This reduction in springtime O₃ concentrations, to less than 220 Dobson units, has resulted in the formation of the Antarctic ‘ozone hole’ (Madronich et al., 1998), which results in large transient increases in the levels of UV-B as the ozone hole moves across the Antarctic continent (McKenzie et al., 2003; Karentz et al., 2004; Lesser et al., 2004; Lamare et al., 2007). Although research suggests that global ozone levels are stabilizing (Häder et al., 2003), full recovery of the ozone layer is not expected for many decades (Newman et al., 2006; McKenzie et al., 2007) and hence UV-B levels are expected to remain high for the immediate future. For example, in 2008 the fifth largest ever ozone hole was recorded over Antarctica (http://www.nasa.gov/topics/earth/features/ozonemax_2008.html).

Secondly, increased greenhouse gas emissions have resulted in an increase in global temperatures and as a consequence a rapid reduction in Arctic sea ice coverage (Turner and Overland, 2009). Although, at present, the impact of global warming on Antarctic sea ice has not been as significant as that seen in the Arctic, researchers have shown accelerated melting of the West Antarctic ice sheet (Velicogna and Wahr, 2006; Shepherd and Wingham, 2007) and it is likely that increasing temperatures will eventually lead to an even greater loss of Antarctic sea ice.

Although it has been demonstrated that UV-B can, to a limited extent, penetrate sea ice and cause damage to animals living under the ice (Lesser et al., 2004), studies have also shown that sea ice, which is relatively opaque and highly reflective especially when covered with snow, reduces the UV-B exposure of aquatic organisms living beneath it to a few percent of that above the ice (Trodahl and Buckley, 1989; Perovich, 1993; Karentz et al., 2004). Therefore sea ice can act as a protective filter greatly reducing the potential for severe UV-B-induced damage in organisms living under it. Most Antarctic marine organisms have highly specialised adaptations to survive at low temperatures and have evolved under relatively low UV-B levels (Karentz, 1991; Karentz et al., 2004; Lesser et al., 2004). Studies have already shown that increased UV-B can reduce primary productivity (Bidigare, 1989; Smith et al., 1992; Arrigo, 1994; Cullen and Neale, 1997) and affect reproduction and development in Antarctic marine organisms (Karentz et al., 2004; Lesser et al., 2004) and hence the combination of elevated UV-B and a reduction in sea ice could have significant ecological consequences (Häder et al., 2007).

At the level of the individual organism UV-B induces cellular damage directly, as it is absorbed by macromolecules such as DNA and proteins, and indirectly by increasing the formation of reactive oxygen species (ROS) such as the superoxide anion (O₂•), hydrogen peroxide (H₂O₂), and the extremely reactive hydroxyl radical (•OH) (Lesser et al., 2001; Lesser and Barry, 2003). Direct absorption of UV-B by DNA results in the formation of cyclobutane-pyrimidine dimers (CPDs) and 6-4 photoproducts, and causes increased mutation rates in phytoplankton, macroalgae, and in the eggs and larval stages of fish and other aquatic animals (Arrigo, 1994; Malloy et al., 1997; Lesser et al., 2001; Lesser et al., 2003; Lamare et al., 2006; Tedetti and Sempéré, 2006; Lamare et al., 2007). ROS are produced continuously in living cells, largely as a normal by-product of respiration in animals, however, changes in environmental conditions that result in stress can lead to the over production of ROS (Halliwell and Gutteridge, 1999). When ROS are produced at high enough levels to overcome the antioxidant defences that normally keep an organism's ROS levels in check, oxidation of DNA, proteins and membrane fatty acids occurs, the later resulting in lipid peroxidation and a loss of membrane function (Halliwell and Gutteridge, 1999). Such damage is commonly referred to as oxidative stress and is considered to be a very sensitive biomarker of many important environmental stressors, including UV-B (Lesser, 2006; Burritt and MacKenzie, 2003; Burritt, 2008). Although CPD formation has been extensively studied, comparatively little is known about the impact of UV-B exposure on oxidative damage and prevention in Antarctic marine invertebrates, especially under field conditions.

The embryonic and early larval stages of many marine invertebrates are ecologically extremely important and have been shown to be especially sensitive to increased levels of UV-B as they have less effective avoidance strategies than adults, because of their small size, rapid cell division, low metabolic activity, potentially limited energy reserves and distribution near the ocean surface (Hoegh-Guldberg et al., 1991; Johnsen and Widder, 2001; Karentz et al., 2004). Although many embryonic and larval marine invertebrates do contain UV-B screening molecules in the form of mycosporine-like amino acids (MAAs), provided by the maternal parent, the levels of these molecules can vary greatly depending on the diet of the parent and are often not present at sufficiently high enough levels to provide complete protection from UV-B induced damage (Dunlap et al., 2000; Adams and Shick, 2001; Shick and Dunlap, 2002; Lesser and Barry, 2003; Lamare et al., 2007). This lack of complete protection by screening molecules is compounded by the fact that springtime phytoplankton blooms and many invertebrate spawning times coincide with the development of the ozone hole, making it highly likely that they will be exposed to maximal UV-B levels (Smith et al., 1992).

Sea urchins (Echinodermata: Echinoidea) are a diverse class of benthic invertebrates that are present in relatively high abundance across most marine biomes. Most sea urchins have a free-swimming larval stage that can be in the plankton for days to months depending on the species. Several species of sea urchins are important broadcast spawners in the benthic communities of Antarctica, with their embryos and larvae found throughout the water column for up to four months in spring and summer (Pearse et al., 1991; Lamare et al., 2006). The Antarctic sea urchin, Sterechinus neumayeri Meissner is one of the most common animals in the shallow subtidal benthos surrounding Antarctica (Brey, 1991; Brey et al., 1995) and so any climatic changes that negatively influence S. neumayeri populations could have a detrimental impact on the whole Antarctic marine ecosystem. S. neumayeri embryos and larvae when compared with those of other echinoids have relatively low metabolic rates and require 20 days of development after fertilisation before they are capably of feeding, during which time they are dependent on the reserves deposited in what is a relatively small egg (Bosch et al., 1987; Shilling and Manahan, 1994). Previous work has shown that despite the presence of UV-B-absorbing MAAs, the embryos and larvae of S. neumayeri are very sensitive to changes in UV-B levels, possibly as a result of their evolution in a low UV-B environment and their slow physiological adaptation rates in the cold, food-poor Antarctic waters (Lesser et al., 2004; Lamare et al., 2007).

The present study was designed to test the potential impact of naturally increased UV-B doses, due to seasonal ozone depletion, combined with an artificial reduction in sea ice cover on S. neumayeri embryos. It has been suggested, but not experimentally demonstrated in the field, that the embryos and larvae of Antarctic invertebrates such as S. neumayeri could be predisposed to oxidative damage because of adaptations required for life at low temperatures (Lesser, 2006). For this reason we chose to use oxidative damage to proteins and lipids as two sensitive biomarkers of oxidative stress, in addition to monitoring abnormal embryo development and aspects of enzymatic antioxidant defence. We discuss our results in the context of life history trade-offs and the potential impact of climate change on Antarctic marine communities.

In situ experiments were conducted at Cape Armitage in McMurdo Sound, Antarctica (77.06°S, 164.42°E) on the embryos of Sterechinus neumayeri Meissner (Echinidae) over two consecutive four-day periods in the spring of 2008 (26–30 October and 1–5 November 2008). Ambient sea temperature during the experimental period was −1.9°C. Adult urchins were collected by SCUBA divers from the seafloor in the immediate vicinity of the study site, and were placed in indoor flowing-seawater tanks plumbed directly with seawater from McMurdo Sound.

Reproductively mature sea urchins were induced to spawn within days of collection by intracoelomic injection of potassium chloride (0.5 mol l⁻¹). Injected animals were inverted over 200 ml beakers containing filtered seawater in order to collect gametes. After the gamete flow had stopped, animals were removed and the gametes washed by serial partial water changes. In order to gain sufficient material for each experiment, eggs from 13 females were combined and fertilised by adding several drops of dilute sperm from three ripe males. Fertilisation was determined by the appearance of a fertilisation envelope, and only batches of eggs with a fertilisation rate >95% were used in experiments. After 5 min freshly fertilised eggs were washed, to remove excess sperm, and stored at a density of approximately 25 individuals per ml, in 3 litre sterile plastic containers, until they were used for experiments. Temperature was maintained at the environmental ambient sea temperature of −1.9°C. Embryos were used within 12 h of fertilisation and remained in a non-feeding stage throughout the experiments.

In situ exposures of embryos to ambient solar radiation were undertaken using an adaptation of the method of Lesser et al. (Lesser et al., 2004). Briefly, embryos were sealed in one litre UV-transparent polyethelene bags (Whirlpak, Nasco–USA) and suspended in the sea at depths of 1 and 4 m by attaching them to polyvinyl chloride racks (31 cm×26 cm horizontal dimensions), with Vexar netting, moored to a rope with appropriate anchors and flotation devices.

At each depth, the bags of embryos were subjected to one of three radiation treatments: (1) photosynthetically active radiation (PAR) but no UV (‘UV-O’ treatment); (2) PAR and UV-A, but no UV-B (‘UV-A’ treatment); and (3) PAR and both UV-A and UV-B (‘UV-T’ treatment). The three UV treatments were achieved using Plexiglas filters (50 cm×50 cm) attached by cable ties directly above the bags of embryos, thus exposing the embryos to electromagnetic radiation passing through the filters. The UV-T treatment consisted of UV transparent Plexiglas that transmitted PAR (84.5%), UV-A (84.6%) and UV-B (80.6%); the UV-A treatment consisted of UV transparent Plexiglas that transmitted PAR (77.9%) and UV-A (46.5%) but minimal UV-B (0.1%) and the controls (UV-O treatment) consisted of UV opaque Plexiglas that transmitted PAR (81.0%) but minimal UV-A (5.2%) and UV-B (0.0%; Fig. 1). For each treatment at each depth, four replicate bags of embryos were used. Nine moorings (three per treatment) were deployed under sea ice (approximately 1.8 m deep) and three moorings (one per treatment) were deployed in an ice-free area (approximately 2 m in diameter) for 90 h. This entire set-up was repeated a second time with different larvae and conducted for 92 h.

At the termination of each experiment the embryos were removed from the racks and immediately placed in a dark, thermally insulated container containing ambient temperature seawater until processing. Care was taken to avoid exposure to direct sunlight during removal of the bags. Processing of each replicate involved taking three, 1 ml random sub-samples from each bag for later analysis of abnormality, with the remaining embryos concentrated and stored frozen at −80°C, in 1.5 ml Eppendorf tubes, for later biochemical analyses.

Daily images of column ozone concentrations during the study period were obtained from the US National Aeronautics and Space Administration (NASA) web site and were accessed routinely during the field season (see http://jwocky.gsfc.nasa.gov). Daily values for stratospheric ozone concentrations over the study area were provided from the Total Ozone Mapping Spectrometer (TOMS) mounted on the Earth Probe Satellite. Hourly data on ambient UV-B (280–320 nm), UV-A (320–400 nm) and the 400–600 nm subset of visible radiation (VIS) were obtained from the National Science Foundation UVR Monitoring Program using a SUV-100 spectroradiometer (Biospherical Instruments Inc.). Data are available at the Biospherical Instruments, Inc. website (http://www.biospherical.com).

The number of dead or abnormal embryos was determined within a few hours of the termination of an experiment by direct microscopic observation at magnifications of ×100 and ×400. Three replicate sub-samples containing 100 embryos from each sample under each light treatment and depth were counted. Embryos were identified as abnormal if they showed one or more of the following developmental aberrations: pronounced thickening of the blastodermis in combination with reduction of the blastocoel, abnormal development of primary mesenchyme cells, occlusion of the blastocoel by cellular debris, exogastrulation or other forms of aberrant archenteron development (Lamare et al., 2007). Abnormality was cross checked by comparison with control embryos kept in the laboratory and not exposed to field conditions of UV radiation.

Total protein was extracted for analysis of protein carbonyls and enzyme activities by mixing a frozen pellet of approximately 25,000 embryos with 300 μl of potassium phosphate buffer (pH 7.0), containing 0.1 mmol l⁻¹ Na₂ EDTA, 1% polyvinylpolypyrrolidone (PVPP), 1 nmol l⁻¹ phenylmethylsulphonyl fluoride (PMSF) and 0.5% Triton X-100, and mixing the sample by vortexing for approx 30 s before centrifugation (13,000 g) for 5 min at 4°C. Protein extracts were stored at −80°C until the assays were conducted. The protein contents were determined as per Fryer et al. (Fryer et al., 1986) with minor modifications.

Lipids were extracted following protein extraction by mixing the pellet containing the remains of 25,000 embryos with 600 μl of methanol:chloroform (2:1) and leaving the suspension to stand for 1 min at room temperature. 400 μl of chloroform was then added and the sample mixed by vortexing for 30 s. 400 μl of deionised water was then added and the sample mixed by vortexing for 30 s, before placing the tubes at 4°C overnight to allow the phases to separate. 50 μl of the chloroform phase was used for the lipid hydroperoxide analysis.

Protein carbonyl levels in the protein extracts were determined via reaction with 2,4-dinitrophenylhydrazine (DNPH) as described by Reznick and Packer (Reznick and Packer, 1994). Lipid hydroperoxide levels in the lipid extractions were quantified using the ferric thiocyanate method described by Mihaljevic et al. (Mihaljevic et al., 1996). Both assays were adapted for measurement using a microplate reader, with glass microplates used for the lipid hydroperoxide analysis. Superoxide dismutase (SOD; EC 1.15.1.1) was assayed using the microplate assay described by Banowetz et al. (Banowetz et al., 2004) with minor modifications. Catalase (CAT; EC 1.11.1.6) was assayed using the chemiluminescent method of Maral et al. (Maral et al., 1977), as modified by Janssens et al. (Janssens et al., 2000). All assays were carried out using a PerkinElmer (Wallac) 1420 multilabel counter (Perkin Elmer, San Jose, California, USA), controlled by a PC, and fitted with a temperature control cell and an auto-dispenser. Data were acquired and processed using the WorkOut 2.0 software package (Perkin Elmer, San Jose, California, USA).

Statistically significant differences in abnormality, lipid hydroperoxides, protein carbonyls, and enzyme activities were tested against depths and treatments using two-way ANOVAs. Any significant treatment or depth effects were then analysed using post-hoc Tukey's HSD tests. Residual values were assessed to ensure that the test assumptions, normality and homogeneity of variances, were met. All analyses were performed using SigmaStat 2.03 (SPSS Inc.) and the level of significance was P≤0.05 for the post-hoc tests.

The ozone hole was located south of the study site in McMurdo Sound during the first experimental period (26–30 October 2008), but was directly over the study site for most of the second experimental period (1–5 November 2008). Ozone concentrations over McMurdo Sound ranged between 233 and 312 DU (with an average of 283 DU) during the first experimental period, decreasing to between 166 and 269 DU (with an average of 210 DU) during the second experimental period (Table 1). Concurrent with changes in ozone concentrations were changes in ambient irradiances, the most noticeable being an increase in incident UV-B during the second experimental period (Table 1). Maximum UV-B irradiance increased 1.7-fold, total UV-B dose increased 3.8-fold and the ratio of UV-B to visible radiation (measured from 400 to 600 nm) doubled (Table 1).

In the sea-ice-free site at a depth of 1 m radiation levels were 24.8% of surface UV-B, 28.4% of surface UV-A and 27% of surface PAR, whereas at 4 m radiation was 2.4% of surface UV-B, 4.7% of surface UV-A and 5% of surface PAR. The extinction coefficients (K_d m⁻¹) for UV-B, UV-A and PAR in the sea ice-free site were 1.14, 0.84 and 0.73 m⁻¹, respectively.

In general, embryos exposed to UV-A but protected from UV-B showed no significant increases in abnormality or oxidative damage, when compared with control embryos protected from both UV-A and UV-B (Figs 2 and 3). The exception was for embryos exposed to UV-A at 4 m under sea ice during the low ambient UV-B (high ozone concentration) experiment which showed increased abnormality (Fig. 3D). In addition, exposure to UV-A did not cause a significant increase in the activities of antioxidant enzymes, when compared with control embryos (Figs 4 and 5).

The proportion of abnormal embryos and oxidative damage to both lipids and proteins increased significantly in embryos exposed to UV-B at a depth of 1 m under open-water conditions, when compared with embryos protected from UV-B by artificial filters (Fig. 2A–C). In addition, both the proportion of abnormal embryos and the levels of oxidative damage were dependent on column ozone concentration and thus ambient UV-B dose (Fig. 2A–C). Embryos exposed to the higher ambient UV-B levels during the second experimental period, when the ozone hole was directly over the study site, had a significantly higher proportion of abnormal embryos and greater oxidative damage when compared with embryos sampled during the first experimental period, when ambient UV-B levels were lower (Fig. 2A–C). Specifically, during the first experimental period embryos exposed to UV-B showed 1.8-fold and 2.5-fold increases in oxidized lipids and proteins respectively, and a 1.8-fold increase in the proportion of abnormal embryos, compared with those protected from UV-B by filters (Fig. 2A–C). During the second experimental period, with lower ozone concentrations and higher ambient UV-B levels, embryos exposed to UV-B showed a 2.4-fold increase in oxidized lipids, a 4.4-fold increase in oxidized proteins and a 2.5-fold increase in the proportion of abnormal embryos, compared with those protected from UV-B by filters (Fig. 2A–C). A significant UV treatment × ambient UV-B interaction effect was also observed for both measures of oxidative damage (F_2,12=12.245, P=0.005; F_2,12=7.118, P=0.001).

In contrast to embryos exposed to UV-B at 1 m under open water conditions those held at 1 m under the sea ice showed no consistently significant increases in either oxidative damage or embryo abnormality, when compared with embryos protected from UV-B by filters (Fig. 2D–F). An ambient UV-B effect on embryo abnormality and a UV treatment effect for protein carbonyl levels were observed when the data from the two experimental periods were pooled and subjected to an ANOVA (F_1,12=13.031, P=0.005; F_2,12=4.205, P=0.05). However, multiple comparison testing revealed there were no significant differences in embryo abnormality or protein oxidation due to the UV treatments (Fig. 2D,E).

For embryos held at 4 m, despite the differences being smaller and results more variable, the same general trends with respect to oxidative damage as seen in embryos held at 1 m were observed, with oxidative damage to both proteins and lipids increasing significantly in embryos exposed to UV-B under open water conditions (Fig. 3B,C). The amount of oxidative damage was also dependent on column ozone concentration and thus ambient UV-B dose (Fig. 3B,C), but unlike embryos held at 1 m no increase in the proportion of abnormal embryos was observed (Fig. 3A). Sea ice once again protected embryos held at 4 m from oxidative damage caused by UV-B exposure under both high and low ozone conditions and from UV-B-induced abnormality during the low ozone high ambient UV-B period (Fig. 3E,F). However, a significant increase in embryo abnormality was observed in embryos exposed to both UV-A and UV-B during the low ambient UV-B experimental period (Fig. 3D). As there was no increase in oxidative damage observed under these treatments and as the trend was not repeated under high ambient UV-B conditions, this result was assumed to be due to a significant ‘batch effect’, whereby the batch of embryos used had a greater propensity for abnormality or was influenced by some unknown factor (i.e. poly-spermy), rather than as a result of UV exposure.

Under open water conditions, the activities of SOD and CAT increased significantly in embryos maintained at 1 m and exposed to UV-B, when compared with embryos protected from UV-B (Fig. 4A,B). Furthermore, the activities of both SOD and CAT were significantly higher during the second experimental period with higher ambient UV-B levels (Fig. 4A,B), indicating that the increases in enzyme activities were UV-B dose-dependent. Under sea ice no significant differences in the activities SOD or CAT, due to UV-B treatment, were observed (Fig. 4C,D). Moreover, under the sea ice no significant differences in enzyme activities between the two experimental periods differing in ambient UV-B levels were observed.

Under open water conditions the basal activities of SOD and CAT in embryos held at 4 m were similar to those found in embryos held at 1m and the same trends with respect to the presence or absence of UV-B were observed, but were less clearly defined, with increases in enzyme activity being much smaller than those observed in embryos held at 1 m (Fig. 5A,B). Some variability in enzyme activities between the two experimental periods was observed. Under the sea ice no significant differences in the activities of SOD or CAT, directly attributable to UV-B, were observed (Fig. 5C,D).

Even without the influence of climate change the Antarctic environment is metabolically demanding, with organisms requiring a range of local adaptations to survive (Guderly, 2004; Pace and Manahan, 2006). Most Antarctic marine organisms are stenothermal and have cellular and metabolic adaptations that allow them to survive and complete their life cycles at a constant seawater temperature of −1.86°C (Pörtner et al., 2007). These adaptations may include increased mitochondrial abundance and/or volume density to help compensate for decreases in O₂ conductance at constantly low temperatures, a higher proportion of unsaturated fatty acids in their membrane phospholipids to help maintain membrane fluidity at low temperatures, the maintenance of greater lipid stores, and changes in both the levels and specific activities of enzymes involved in key metabolic processes including respiration (Hazel, 1995; Guderly, 2004). Although these adaptations allow aerobic organisms to survive and function at constantly low temperatures, such adaptations can greatly increase the susceptibility of these organisms to environmental stresses which cause excess ROS production and hence oxidative damage, when compared with organisms living in warmer and more variable environments (Viarengo et al., 1998; Guderly, 2004; Pörtner et al., 2007). In addition, Antarctic seawater is rich in oxygen, further increasing the likelihood of excessive ROS production under environmental conditions that induce stress (Lesser, 2006).

UV-B can directly induce ROS formation either in water, through UV-B absorption by dissolved organic carbon (Cooper et al., 1994; Richard et al., 2007), or inside animals, where photosensitisers such as flavins and reduced pyridine nucleotides absorb UV energy and are elevated to an excited state (Halliwell and Gutteridge, 1999). When returning to the ground state, excited photosensitisers transfer energy to ground-state oxygen producing ¹O₂, H₂O₂ and O₂•, which can in turn lead to the production of •OH. The •OH radical is potentially very damaging to cells as it can initiate lipid peroxidation, which once started can proceed via a free-radical-mediated chain reaction mechanism which continues unabated unless terminated (Halliwell and Gutteridge, 1999). Polyunsaturated fatty acids, which contain multiple double bonds and highly reactive hydrogens, are most sensitive to lipid peroxidation. Animals adapted to cold water environments, such as the Antarctic oceans, often have a higher proportion of unsaturated fatty acids in their membranes and higher levels of lipid reserves, which are also able to undergo peroxidation, and therefore may be particularly susceptible to lipid peroxidation. Moreover, oxidative damage to membranes and other key cellular macromolecules, such as proteins, can lead to mitochondrial damage. It has been estimated that normal undamaged mitochondria convert between 1 and 3% of their entire oxygen consumption to ROS (Sohal and Weindruch, 1996) and so any disruption of mitochondrial function, especially in animals whose mitochondria are adapted to cold-water environments, could result in a large increase in ROS production and therefore greatly increase the potential for oxidative damage.

In the present study we showed that under open-water conditions, at a depth of 1 m, exposure of S. neumayeri embryos to naturally elevated levels of UV-B as a result of the presence of the Antarctic ozone hole over the study site, greatly increased levels of oxidative damage to proteins and lipids, and also increased the proportion of abnormal embryos observed. At a depth of 4 m increases in oxidative damage from UV-B exposure also occurred, but the increases were small compared with those observed at 1 m and no increase in embryo abnormality, which could be directly attributed to UV-B exposure, was observed. This reduction in oxidative damage at increased depth was most likely due to the attenuation of UV-B by sea water, for at 1 m below the surface UV-B levels were 24.8% of surface levels whereas at 4 m UV-B levels were only 2.4% of surface levels. These results support in part the findings of Karentz et al. (Karentz et al., 2004), who found that at depths greater than 3 m reduced or no UV-B-induced embryo abnormality was evident and that DNA damage in the form of CPDs was minimal. However, unlike Karentz et al. (Karentz et al., 2004) who found that compared with visible light and UV-A, UV-B had a relatively small impact on embryo abnormality at depths less than 3 m, the results of the present study support the findings of numerous other studies that have shown that UV-B can have a considerable impact on the embryos of sea urchins and other marine larvae (Lesser et al., 2001; Browman et al., 2003; Lesser and Barry, 2003; Lesser et al., 2003; Lesser et al., 2004; Lamare et al., 2007). As abnormality can be the end result of exposure of a developing embryo to several stressors, is not surprising that under highly variable field conditions, such as those found in Antarctica, the accumulation of cellular damage to a level sufficient to cause abnormalities could be achieved in different ways depending on the combined impacts of many environmental factors. For example, Karentz et al. (Karentz et al., 2004) showed that abnormal embryo morphology could be caused by exposure to 80 kJ m⁻² of UV-B or by exposure to more than 2000 kJ m⁻² of UV-A, suggesting that under conditions of very high UV-A the effects of UV-B could easily be masked, as abnormal embryos would be observed in the presence or absence of UV-B. The UV-A levels to which the embryos were exposed in the present study were lower than those detailed for most of the experiments presented by Karentz et al. (Karentz et al., 2004), hence the UV-A dose in the present study may have been insufficient to cause enough cellular damage in the absence of UV-B to induce abnormal embryo development. This is supported by the failure of UV-A alone to induce significant amounts of oxidative damage.

To cope with ROS generation and to help prevent oxidative damage animals have a number of antioxidant defences including ROS scavenging molecules such as ascorbate, glutathione and α-tocopherol, and ROS detoxifying enzymes such as SOD and CAT (Halliwell and Gutteridge, 1999). Animals adapted to low temperatures often have increased antioxidant defences to enable them to cope with the potential for increased mitochondrial ROS production. In the present study two commonly used enzymatic markers of ROS scavenging capacity, SOD and CAT, were used to determine if S. neumayeri embryos are capable of increasing their ROS scavenging capacity to protect themselves from UV-B-induced oxidative damage. Our results clearly demonstrate that although S. neumayeri embryos contain moderate to high levels of both SOD and CAT, when compared with the developing eggs and larvae of other marine animals (Buchner et al., 1996; Rudneva, 1999; Korkina et al., 2000; Regoli et al., 2002), that exposure to UV-B induces a small but significant dose-dependent increase in both SOD and CAT activities. Lesser et al. (Lesser et al., 2003) also found an increase in SOD activity when the embryos of the sea urchin Strongylocentrotus droebachiensis were exposed to UV-B. In the present study although significant increases in the activities of antioxidant enzymes were observed, the increases were insufficient to completely protect embryos from oxidative damage to proteins and lipids, especially at a depth of 1 m. It is highly likely that this increase in oxidative damage to these key macromolecules contributed significantly to the increase in abnormality observed in embryos exposed to UV-B at this depth.

Although studies have shown that the UV-A portion of the spectrum can cause oxidative damage in animal cells (McMillan et al., 2008), studies on marine animals have provided mixed results, with some studies demonstrating negative effects (Adams and Schick, 1996; Lesser and Barry, 2003; Lesser et al., 2003; Karentz et al., 2004) whereas others showed little effect at all (Béland et al., 1999; Kouwenberg et al., 1999). In the present study no significant effects due to exposure to UV-A radiation were observed. This could have been due to the fact that the filters used to block UV-B also reduced UV-A doses by over 50% and as UV-A levels were not significantly influenced by the presence or absence of the ozone hole this treatment effectively reduced ambient UV-A levels. Moreover, as S. neumayeri embryos were shown in the present study to have moderate to high basal levels of SOD and CAT activity and sea urchin embryos and larvae have been shown in other studies to have antioxidant capacity (Lesser et al., 2003; Lesser, 2006), it is highly likely that significant levels of oxidative damage only occur in embryos when their basal antioxidant capacity is greatly exceeded by ROS production, for example when exposed to high levels of UV-B.

The results of the open water experiments detailed above clearly demonstrate that naturally increased UV-B doses, due to seasonal ozone depletion, cause oxidative damage in S. neumayeri embryos under field conditions, despite a limited capacity of the embryos to increase the activities of antioxidant enzymes. However, during the Austral spring McMurdo Sound is normally covered with 2–3 m of annual sea ice and it is generally accepted that sea ice provides a physical barrier to UV-B radiation reaching the underlying water column (Perovich, 1993; Perovich, 2002). The results of the present study suggest a protective role for sea ice by demonstrating that sea urchin embryos held beneath almost 2 m of sea ice were completely protected from oxidative damage and UV-B-induced abnormality. Moreover, no significant increases in SOD or CAT activities were observed in animals protected by sea ice, indicating that no upregulation of ROS scavenging capacity was required, or that ROS levels did not increase sufficiently to trigger increases in SOD and CAT activities. Although the upregulation of other antioxidant mechanisms cannot be ruled out, changes in the activities of SOD or CAT are commonly used as cellular markers of oxidative stress and ROS scavenging capacity in most animals (Halliwell and Gutteridge, 1999) and have been shown to be upregulated in response to oxidative stress in marine animals (Lesser, 2006). Despite the above results it should be noted that Lesser et al. (Lesser et al., 2004) found UV-B wavelengths reached a depth of over 6 m after being transmitted through approximately 3 m of annual sea ice in McMurdo Sound. Furthermore, they observed mortality and DNA damage in sea urchin embryos exposed to UV radiation beneath the ice over two consecutive years and that higher mortality and DNA damage were associated with greater irradiances of UV-B. There are many possible reasons for the difference between the Lesser et al. (Lesser et al., 2004) study and the present study. For example, seasonal variations in the levels of screening pigments in maternal parents, which are dependent on food abundance and quality, could influence the levels of UV-B screening pigments passed on to the embryos, fluctuations in ozone concentrations and cloud cover during experiments, variations in ice thickness and optical properties (reviewed by Perovich, 2002), the thickness of snow cover and water transmission characteristics, which are highly variable and constantly changing on both a daily and seasonal basis, making direct comparisons between studies difficult. Irrespective of the above, both studies did show that ice reduces the negative effects of UV-B on S. neumayeri embryos and therefore plays an important protective function.

As mentioned previously, most studies on the influences of UV-B on S. neumayeri embryos have concentrated on direct DNA damage, in the form of CPDs, as a measure of the influence of UV-B exposure at the cellular level and microscopic observations to determine the levels of embryo abnormality. Although CPDs are a sensitive marker for exposure to UV-B, because DNA dimers such as CPDs are produced at high levels in cells exposed to UV-B, most organisms have evolved very effective mechanisms to repair them (Sinha and Hader, 2002). Unlike DNA dimers, which can be reversed effectively by photoreactivation, the overproduction of ROS caused by exposure of an organism to UV-B can have far more wide ranging consequences. Oxidative damage to DNA, proteins and lipids is both energetically expensive to prevent and potentially very resource intensive to repair (Halliwell and Gutteridge, 1999; Sinha and Hader, 2002).

The cost of prevention and/or repair of oxidative damage could be a very important life history trade-off for organisms living in the Antarctic marine ecosystem. Moreover, S. neumayeri embryos require 20 days of development after fertilisation before the resultant larvae are capably of feeding (Bosch et al., 1987) during which time they are totally dependent on stored reserves for growth and development. Thus any relocation of resources from growth and development to the prevention or repair of oxidative damage could significantly influence larval fitness and/or survivability. In response to increased ROS levels, developing embryos could be required to invest valuable cellular resources to prevent oxidative damage, but this is likely to have consequences for other important traits (Monaghan et al., 2009).

In conclusion, basal levels of oxidative damage and antioxidant enzyme activities in S. neumayeri embryos were within the range of values published for other marine invertebrates. In embryos exposed to UV-B under open water conditions antioxidant enzyme activities increased, but not enough to prevent a significant increase in oxidative damage and developmental abnormalities. The amount of oxidative damage was dependent upon ambient UV-B levels and increased as the ozone hole moved over the study site. Embryos protected from UV-B by sea ice showed little oxidative damage with few developmental abnormalities, directly attributable to UV-B. Our results suggest that normal development of sea urchin embryos could require the presence of sea ice, which acts as a UV-B screening filter, protecting the embryos from potentially damaging increases in UV-B. As embryos represent a key life-history stage, lower survival rates caused by increased UV-B as a result of the ozone hole and significant reductions in protective sea ice coverage brought about by global warming, could reduce the long-term viability of affected populations and potentially have a significant impact on the sensitive Antarctic marine ecosystem. Sea urchin embryos could therefore be an ideal indicator species for the potential impact of climate change on Antarctic marine invertebrates.

We especially thank Antarctica New Zealand for logistical support over two field seasons at Scott Base and Kelly Tarlton's Antarctic Encounter and Underwater World for financial support via a postgraduate scholarship. Thanks are also extended to staff at the Portobello Marine Laboratory for assistance during research. Thank you to the members of K-068: Nickolas Isley, Dana Clark, Michael Gonsior and Michelle Liddy (both in the 07/08 and 08/09 seasons) and K-065: Dr Steve Wing, Rebecca Macleod, Brian Grant, Greg Funnell and Jim) for help with field-work in Antarctica. Financial support to write this manuscript was provided by a University of Otago Publishing Bursary.