The energetic cost of exposure to UV radiation for tadpoles is greater when they live with predators

Authors


Correspondence author. E-mail: l.alton@uq.edu.au

Summary

1. Global increases in ultraviolet-B radiation (UVBR) associated with stratospheric ozone depletion are thought to be contributing to the rapid disappearance of amphibian populations from pristine habitats around the world. Much research has been dedicated to understanding the effects of ultraviolet radiation (UVR) alone and in combination with other environmental stressors on the early life stages of amphibians. Little is known, however, of how UVR affects amphibian metabolism, or how amphibian metabolism may change in response to exposure to other stressors in addition to UVR.

2. Using a controlled laboratory experiment, we examined the independent and interactive effects of UVR and non-lethal predatory chemical cues (PCC) on the tissue and whole-animal metabolic rate (MR) and activity of striped marsh frog Limnodynastes peronii tadpoles. PCC signal risk of predation, which is a natural stressor that can cause tadpoles to alter their behaviour, morphology, and life history, and for which the metabolic cost remains little explored.

3. We found that exposure to UVR caused tissue MR to increase by 36%, but whole-animal MR to decrease by 14%, which is most likely due to tadpoles reducing their activity levels by 56%. Exposure to PCC had no significant effect on tissue or whole-animal MR, but caused tadpoles to reduce their activity levels by 36%, indicating that the whole-animal MR of tadpoles exposed to PCC is elevated relative to their activity levels. Compared with tadpoles exposed to neither stressor, tadpoles exposed simultaneously to UVR and PCC showed no change in whole-animal MR despite reducing their activity levels by 62%.

4. This research shows that, for tadpoles, there is an energetic cost associated with being exposed to UVR and PCC independently, and that this cost is greater when they are exposed to both stressors simultaneously. Our previous research has shown that exposure of tadpoles to PCC enhances the lethal effects of UVR, and the present study suggests that this synergistic interaction may arise as a consequence of the effect of these combined stressors on MR. Global increases in UVBR may therefore be contributing to amphibian population declines by compromising energy allocation towards growth and development as well as energy allocation towards coping with additional environmental stressors.

Introduction

The world is arguably experiencing a sixth mass extinction event and amphibians, being the most threatened of all vertebrate taxa, are at the forefront of this biodiversity crisis (Lawton & May 1995; Wake & Vredenburg 2008; IUCN 2010). Although amphibians are being lost globally, it is the rapid disappearance of populations from seemingly pristine habitats that poses the most concern because habitat loss and over-exploitation do not apply in these cases (Stuart et al. 2004). Consequently, these enigmatic declines have prompted consideration of more subtle environmental stressors, particularly those arising from atmospheric disturbances such as increased ultraviolet-B radiation (UVBR) associated with stratospheric ozone depletion (Stuart et al. 2004; WMO 2007). The role of increased UVBR in amphibian declines has received considerable attention with research indicating that exposure to ultraviolet radiation (UVR) may lead to population declines in some species through its lethal and sublethal effects on embryonic and larval life stages (see reviews Blaustein et al. 1998, 2001; Crump 2001; Blaustein et al. 2003; Bancroft, Baker & Blaustein 2008; Croteau et al. 2008). While research has predominantly focused on the effect of UVR on survival, growth, development, morphology and behaviour of early life stage amphibians, comparatively little consideration has been given to its effect on metabolism, a fundamental physiological process that dictates energy allocation to the fitness-enhancing processes of survival, growth and reproduction (Brown et al. 2004).

Alteration of metabolism in response to UVR exposure may occur as a consequence of a number of different factors. For instance, exposure to UVR has the potential to be energetically costly as it inflicts damage upon a number of cellular components that then require repair. Within cells, UVR triggers damaging photochemical reactions in DNA, proteins and lipids through either direct absorption or the production of reactive oxygen species (ROS) (Tyrrell 1991; Tevini 1993). The damage caused by UVR stems primarily from its absorption by DNA where it causes the formation of pyrimidine dimers, which are covalent linkages between adjacent pyrimidine nucleotide bases that compromise DNA replication and transcription (Jagger 1985; Friedberg et al. 2006). In several taxa, including amphibians, this form of damage can be repaired by the energetically efficient and accurate process of enzymatic photoreactivation (EPR), which is catalysed by enzymes known as photolyases that utilize 350–450 nm light as an energy source, not ATP, to reverse the DNA damage in a single-step reaction (i.e. break the bond between the pyrimidines) (Blaustein et al. 1994; Friedberg et al. 2006). ATP is consumed, however, by an alternative DNA repair pathway known as nucleotide excision repair (NER), which rather than reversing the damage, removes and replaces the damaged section of DNA (Sancar & Tang 1993). Similarly, the processes of degradation and resynthesis of proteins and lipids are also likely to promote energy expenditure, as is the up-regulation of antioxidant proteins needed to mitigate the detrimental effects of ROS (Lesser et al. 2001; Fischer et al. 2006). As such, exposure to UVR might cause the metabolism of an organism to increase if the damage to DNA, proteins and lipids is sufficient to activate these ATP-consuming repair pathways. Alternatively, if UVR-induced damage is severe and accumulates because of insufficient repair, a reduction in metabolism may occur, followed by apoptosis, and possibly mortality (Fischer et al. 2006). The effect of UVR on metabolism is further complicated by its effects on organismal behaviour and activity. For example, exposure to UVR can cause the tadpoles of some amphibian species to swim erratically (Grant & Licht 1995; Nagl & Hofer 1997), while others reduce their swimming activities (Zaga et al. 1998; Garcia, Paoletti & Blaustein 2009). Such alterations in activity level may then either exacerbate, or compensate for, any change in metabolism that arises as a consequence of UVR-induced cellular damage. Alternatively, changes in activity levels may be a product of UVR-induced changes in metabolism, though the cause for changes in activity as a result of UVR exposure has not been examined.

Given the potential for UVR to disrupt the homeostasis of an organism through its effects on metabolism, it is important to consider how this may then affect the ability of the organism to cope with additional environmental stressors. It has long been recognized that many amphibian population declines are unlikely to be a consequence of a single stressor, but rather a result of multiple stressors interacting to have compounding negative effects (Blaustein & Wake 1990; Blaustein & Kiesecker 2002; Linder, Krest & Sparling 2003). Multi-factorial studies that examine the interactive effects of UVR combined with an additional stressor generally consider chemical contaminants, such as pesticides and fertilizers (Bancroft, Baker & Blaustein 2008; Croteau et al. 2008). Such research has shown that UVR and contaminants can interact synergistically, enhancing embryonic and larval mortality above the additive effects of the independent stressors (Bancroft, Baker & Blaustein 2008; Croteau et al. 2008). If, however, contaminants are less likely to be present in the undisturbed habitats where enigmatic amphibian declines have occurred, then consideration of how UVR interacts with natural stressors is particularly important.

One natural stressor that can impact greatly upon the early life stages of amphibians is predation. Aside from causing direct mortality through consumption, predators invoke defensive strategies in tadpoles in the form of changes in behaviour, morphology and life history (e.g. Skelly & Werner 1990; Relyea 2001, 2007). In a previous study, we examined the independent and interactive effects of UVR and non-lethal predatory chemical cues (PCC), which are chemical cues that signal risk of predation, on the survival and morphology of striped marsh frog Limnodynastes peronii (Duméril and Bibron, 1841) tadpoles (Alton, Wilson & Franklin 2010). As has been shown in other amphibian species, we found that exposure to UVR alone caused tadpoles to be smaller (Croteau et al. 2008; Alton, Wilson & Franklin 2010), as did exposure to PCC alone (Relyea & Werner 1999; Alton, Wilson & Franklin 2010). Exposure to PCC alone also caused tadpoles to change their morphology (Alton, Wilson & Franklin 2010), which is known to increase their chance of survival during encounters with predators (Kraft, Franklin & Blows 2006). The reduced investment of resources into growth suggests there is an energetic cost associated with being exposed to UVR and PCC independently, with resources possibly being allocated towards the repair of UVR-induced damage, and the production and maintenance of morphological defences, respectively. As with other multi-factorial studies concerned with the interactive effects of UVR (Bancroft, Baker & Blaustein 2008), we found that simultaneous exposure to UVR and PCC caused greater tadpole mortality than the additive mortality caused by the two stressors independently, indicating a synergistic interaction (Alton, Wilson & Franklin 2010). We also found that tadpoles exposed simultaneously to UVR and PCC were smaller than those exposed to these stressors independently, and did not exhibit the predator-induced morphological defence that was evident in tadpoles exposed to only PCC (Alton, Wilson & Franklin 2010). If exposure to UVR and PCC each requires an energetic investment as suggested by their independent effects on growth, the synergistic interaction between UVR and PCC suggests that the combined energetic costs of UVR and PCC exceeds the total capacity of tadpoles to expend energy. These findings also suggest that the division of resources between the response mechanisms for UVR and PCC potentially compromises the ability of tadpoles to cope adequately with either stressor. Put another way, exposure of tadpoles to UVR compromises their energetic investment in coping with PCC, and exposure to PCC compromises their energetic investment in coping with UVR. This then may ultimately lead to an accumulation of UVR-induced damage as well as reduced investment into the production and maintenance of predator-induced defensive structures.

In the present study, we examine the independent and interactive effects of UVR and PCC on the whole-animal and tissue metabolic rate (MR), and activity of L. peronii (Fig. 1) tadpoles in an attempt to understand the physiological basis for the previously observed effects of these stressors, alone and in combination, on tadpole survival and morphology. We hypothesized that independent exposure to both UVR and PCC would lead to an increase in MR, and that exposure to both stressors simultaneously would be more energetically costly than exposure to either stressor independently.

Figure 1.

 Adult (top) and tadpole (bottom) of striped marsh frog Limnodynastes peronii (Duméril and Bibron, 1841). Photographs: Lesley Alton.

Materials and methods

Animal Collection and Maintenance

Eight freshly laid Limnodynastes peronii foam egg masses were collected from an ephemeral creek near The University of Queensland, Brisbane, Australia (27°30′22·81″S, 152°59′22·99″E). Egg masses were collected at 9 am and transported to The University of Queensland where an equal number of eggs were randomly selected from each mass and divided among 20 experimental tanks, such that there were 32 eggs within each tank. Embryos and larvae were reared inside larval fish nets that sat 20 mm below the water surface inside the tanks. Tanks were filled with 5 L of water that was purified by reverse osmosis (RO water) and supplemented with ocean salt (Aquasonic, Wauchope, NSW, Australia) to give a salinity of approximately 0·3 ppt. Ammonia levels within the tanks were monitored regularly using ammonia test kits (Aquasonic) and regular water changes were performed to maintain water quality. Eggs hatched 3 days after collection, and tadpoles reached Gosner stage 25 (Gosner 1960) the following day and were fed with boiled spinach ad libitum. To account for deaths that occurred during embryonic development and to minimize density effects, 5 days after collection, randomly selected tadpoles were removed such that there were 20 tadpoles per tank remaining, except for two tanks that had only 16 and 17 tadpoles remaining (N.B. we chose not to reduce the number of individuals to 16 to ensure that there would be enough tadpoles within the majority of tanks until the end of the experimental period). Temperature (25 ± 1 °C) and photoperiod (12L : 12D) were kept constant for the duration of the experimental period.

Experimental Treatments

Tadpoles were exposed to a factorial combination of two UVR treatments [UVR absent (UV); UVR present (UV+)] and two predator treatments [PCC absent (PCC); PCC present (PCC+)], such that there were four experimental treatments in total (i.e. UV PCC, UV+ PCC, UV PCC+, and UV+ PCC+) with five replicate tanks in each.

Embyros and larvae within the UV+ treatment were exposed to UVBR, UVAR and visible light emitted from two 40 W linear fluorescent light bulbs (Repti Glo 8.0, Exo Terra®, Montreal, QC, Canada) that were on for 12 h each day, and an additional four 40 W linear fluorescent light bulbs (two Repti Glo 8.0, Exo Terra®; and two T-40, Vilber Lourmat, Marne-la-Vallée, France) that were on for 4 h each day at the photoperiod midpoint. Embyros and larvae within the UV treatment were exposed to visible light and a negligible amount of UVBR and UVAR emitted from two 40 W linear fluorescent light bulbs (Repti Glo 2.0, Exo Terra®) that were on for 12 h each day.

The absolute irradiance of UVBR in ambient midday sunlight during the peak breeding season of L. peronii (i.e. summer) in Brisbane, Australia, has been measured previously as 5 W m−2 (van Uitregt, Wilson & Franklin 2007), which corresponds to a UV index (UVI) of 11 (WHO 2002; ARPANSA 2010b). The peak absolute irradiance of UVBR generated by our UV+ lighting regime was 3·5% of 5 W m−2, and the UVI within the UV+ treatment was <1 (see Appendix S1 and Table S1, Supporting information). Over the course of a cloudless day in Brisbane during summer (December to February), the UVI is below 1 only in the early morning (before 7 am) and in the late afternoon (after 5 pm), with the average summer daily maximum UVI between December 2004 and February 2010 being 11 (ARPANSA 2010a).

Freshwater shrimp Macrobrachium australiense (Holthius, 1950) were used as predators and were obtained from a local aquarium supplier (Westside Pets & Aquarium, Taringa, Qld, Australia). Three shrimp were placed outside of the larval fish net in each of the tanks that were assigned to the PCC+ treatment. In all 20 experimental tanks, a piece of black plastic was placed under the larval fish nets to ensure that shrimp received minimal exposure to the lighting treatments. Shrimp were fed L. peronii tadpoles (Gosner stage 24–26; Gosner 1960) ad libitum; thus, PCC+ tadpoles were exposed to chemical cues from both the predators and damaged conspecifics (Chivers & Smith 1998). To ensure adequate mixing of PCC through the larval fish nets, all 20 experimental tanks were aerated with an air stone.

Activity and Whole-Animal Respirometry Measurements

After 9 days of exposure to experimental treatments, the activity and oxygen consumption rate (i.e. MR) of 30 tadpoles from each treatment (i.e. six tadpoles from each experimental tank) was measured. Food was removed from experimental tanks a minimum of 18 h prior to respirometry measurements. For logistical reasons, activity and MR measurements were performed across six batches over the course of 1 day (10·5 h) with one randomly selected tadpole from each tank measured in each batch. As measurements were conducted over such a long period, all experimental tanks were removed from their respective UVR treatments and maintained under laboratory fluorescent lighting containing no UVR so that tadpoles across batches were not exposed to varying levels of UVR prior to being measured for activity and MR. To standardize exposure of tadpoles across batches to experimental treatments immediately prior to being measured for activity and MR, individual tadpoles were placed in a container filled with 12 mL of water from their respective PCC treatments (water depth of 20 mm), and placed within their respective UVR treatments for 45 min. The containers holding PCC+ tadpoles were filled with aerated water pooled from all the PCC+ treatment tanks, while containers holding PCC tadpoles were filled with aerated RO water supplemented with ocean salt (Aquasonic) to give a salinity of approximately 0·3 ppt. Containers holding UV tadpoles were placed under UV lighting, and UV+ tadpoles were placed under UV+ lighting (N.B. UV+ tadpoles were exposed to peak photoperiod midpoint lighting).

Following the standardized 45 min exposure to UVR treatments, tadpoles were filmed from above within their containers with a camcorder for 15 min to assess activity levels. For logistical reasons, activity could not be measured simultaneously with respirometry. During filming, tadpoles were not exposed to direct artificial lighting, but to ensure that tadpoles were in focus in the video footage natural light was allowed to enter the room through a window. Video footage was played back in real time and the total time each individual spent being active was recorded using a stopwatch with any movement considered as activity.

After filming, tadpoles were sealed individually in 5 mL glass vials with an integrated oxygen sensor spot (W-In-SP-PSt3-NAU-D5-YOP; Presens, Regensburg, Germany). As before, vials containing PCC+ tadpoles were filled with aerated water pooled from all the PCC+ treatment tanks, and vials containing PCC tadpoles were filled with aerated RO water supplemented with ocean salt (Aquasonic) to give a salinity of approximately 0·3 ppt. After all 20 vials were sealed, along with four control vials without tadpoles (two vials with PCC+ water and two vials with PCC water), they were placed in random order on a 24-channel oxygen meter (SDR SensorDish® Reader; Presens) that measured the oxygen concentration (% air-saturation) at 20 s intervals for 1 h in the 24 vials simultaneously using optical fluorescence-based oxygen respirometry (see Köster, Krause & Paffenhöfer 2008 for details). Respirometry measurements were performed in a dark temperature-controlled room that maintained temperature at 24 ± 0·5 °C. Following respirometry measurements, each tadpole was blotted on paper towel and weighed to the nearest 0·0001 g (tadpole mass range of 0·0070–0·0167 g with mean ± SE tadpole mass of 0·0101 ± 0·0002 g) using an analytical balance (XS204; Mettler-Toledo Ltd, Port Melbourne, Vic., Australia). All tadpoles were then euthanized by an overdose of MS-222.

To ensure that the effects of PCC on MR and activity were not because of the presence of conspecific cues in the water, we performed an additional experiment that examined the effect of conspecific cues (CC) on the whole-animal MR and activity of L. peronii tadpoles. We randomly allocated 160 L. peronii tadpoles across 16 tanks such that there were 10 tadpoles per tank. Tanks were filled with 1 L of RO water supplemented with ocean salt to give a salinity of approximately 0·3 ppt, and tadpoles were fed with boiled spinach ad libitum. After 3 days of being maintained in these tanks, the activity and MR of four randomly selected tadpoles from each tank was measured using the same methodology as described earlier. Two of the four tadpoles had their activity and MR measured in water pooled from the 16 holding tanks (i.e. exposed to CC, CC+), while the other two had their activity and MR measured in RO water supplemented with ocean salt to give a salinity of approximately 0·3 ppt (i.e. not exposed to CC, CC). Whole-animal MR data were analysed using a mixed-model ancova with CC (present or absent) as a fixed factor, mass of tadpole as a covariate, and trial and tank as random effects. Activity data were arcsine square root transformed to satisfy assumptions of normality and homogeneity of variance, and were analysed using an anova with CC (present or absent) as a fixed factor. The results from this experiment show that CC have no effect on either whole-animal MR (F1,44 = 0·20, P = 0·66, n = 31–32) or activity (F1,62 = 0·36, P = 0·55, n = 32), and thus, the effects observed as a result of the PCC+ treatment are attributable to PCC, not CC.

In Vitro Tissue Respirometry

Following whole-animal respirometry measurements, the remaining tadpoles continued to be exposed to experimental treatments for 15 extra days to allow tadpoles to grow to a size that would generate enough tissue for respirometry measurements. Whole-animal respirometry could not be performed at the time tissue respirometry was performed because tadpoles were too large to fit into the 5 mL respirometry vials. At the time tissue respirometry was performed, 10–14 tadpoles were remaining in all but two tanks that had only 3 and 8 tadpoles remaining. The difference in tadpole numbers across tanks is attributable to some accidental deaths that occurred during water changes, and other deaths, the cause of which were unknown. For logistical reasons, tissue respirometry was performed in two batches over 2 days, and as with activity and whole-animal respirometry measurements, all experimental tanks were removed from their respective UVR treatments and maintained under laboratory fluorescent lighting containing no UVR during this time. Food was removed from experimental tanks a minimum of 18 h prior to respirometry measurements. On the first day of tissue respirometry measurements (batch number 1), 2–6 randomly selected tadpoles from each tank were placed in a container filled with 60 mL of water from their respective PCC treatments (water depth of 20 mm), and placed within their respective UVR treatments for 45 min. The containers holding PCC+ tadpoles were filled with aerated water pooled from all the PCC+ treatment tanks, while containers holding PCC tadpoles were filled with aerated RO water supplemented with ocean salt (Aquasonic) to give a salinity of approximately 0·3 ppt. Containers holding UV tadpoles were placed under UV lighting, and UV+ tadpoles were placed under UV+ lighting (N.B. UV+ tadpoles were exposed to peak photoperiod midpoint lighting).

Following the standardized 45 min exposure to UVR and PCC treatments, tadpoles were euthanized by decerebration with a surgical blade and their tails were cut away from the body and immediately placed in aerated McKenzie’s amphibian Ringer solution (pH 7·4; 111 mm NaCl, 2·5 mm KCl, 1·8 mm CaCl2, 1 mm MgCl2, 5 mm HEPES, 10 mm glucose). To ensure that the oxygen consumption of the tail tissue was detectable, all the tails (2–6) from each tank were pooled together and sealed in the same 5 mL glass vials used for whole-animal respirometry. Whole tails were chosen to measure tissue MR because they are thin and thus permit adequate diffusion of oxygen while maintaining organizational integrity (Umbreit, Burris & Stauffer 1972). After all 20 vials were sealed, along with four control vials containing only Ringer solution and no tissue, they were placed in random order on the 24-channel oxygen meter that took an initial measurement of the oxygen concentration (% air-saturation) in each vial. After the first measurement of oxygen concentration, all 24 vials were placed on a rotator (Roto-Shake Genie®; Scientific Industries, Inc., Bohemia, NY, USA) that continuously rotated the vials through 360° to ensure adequate mixing within the vials. Initially, vials were transferred back to the oxygen meter every 15 min to record oxygen concentration, but the recording rate was reduced to once every hour after 1 h. The oxygen consumption of the tissues was recorded for 7 h, after which the tails from each vial were blotted on paper towel and weighed together to the nearest 0·0001 g (tissue mass range of 0·0126–0·0309 g with mean ± SE tissue mass of 0·0199 ± 0·0009 g) using the analytical balance. As with the whole-animal respirometry measurements, tissue respirometry measurements were performed in a dark temperature-controlled room at 22·5 ± 0·5 °C. On the second day of tissue respirometry measurements (batch number 2), this process was repeated with an hourly sampling rate, 3–7 tails pooled together in each vial, and five control vials.

Metabolic Rate Calculations and Statistical Analyses

The oxygen consumption rate (inline image, mL O2 h−1) of tadpoles and tissues was calculated by fitting a linear regression to the data obtained from the oxygen meter (% air-saturation on time), and using the equation:

image

where: mt is the slope derived from the trial with the tadpole or tissue (% air-saturation h−1); mc the mean slope derived from the controls (% air-saturation h−1) (note that the difference between mt and mc is divided by 100 to convert the percentage of oxygen in the media to a fraction); Vv the volume of the respirometry vial (L) = 0·005 L; Vt the volume of tadpole or tissue (L) = tadpole or tissue mass (kg)/density of muscle (1060 kg m−3; Alexander 1982); βO2 is the O2 capacitance of the media (mL L−1) = 5·89 mL L−1 for whole-animal O2 consumption rate (air-saturated freshwater at 24 °C) (Riley & Chester 1971) and = 5·80 mL L−1 for tissue O2 consumption rate (Ringer solution with a salinity of 8 ppt at 22·5 °C) (Riley & Chester 1971).

In the whole-animal MR trials, we observed that the slope of the first half hour of data was consistently steeper than that of the remainder. We therefore excluded these data from inline image calculations given that this observation is likely to be attributed to animals being more active at the beginning of the trial having recently been handled, and also to equilibration in temperature between the vials and oxygen meter. Similarly, in tissue MR trials performed on the first day, we observed that oxygen levels increased during the first half hour after which they levelled off and began to decline after one hour. We therefore excluded the first hour of data from trials performed on both days given that this observation is likely to be attributed to equilibration in temperature between the vials and oxygen meter.

Whole-animal MR data were analysed using a two-way mixed-model ancova with UVR (present or absent) and PCC (present or absent) as fixed factors, an interaction term between UVR and PCC, mass of tadpole as a covariate, and tank, batch and vial number as random effects. Of the 120 whole-animal MR trials, 92 showed linear decreases in oxygen saturation with time. Data for the remaining 28 trails (11, 4, 7 and 6 trials from the UV PCC, UV PCC+, UV+ PCC and UV+ PCC+ treatments, respectively) showed generally linear declines in oxygen saturation but were interspersed with occasional transient (∼5 min) downward peaks. As the cause of this malfunction is unknown, and the appearance of the transient downward peaks is apparently unrelated to treatment conditions, data for these 28 trials were discarded. Tissue MR data were analysed using a two-way mixed-model ancova with UVR (present or absent) and PCC (present or absent) as fixed factors, an interaction term between UVR and PCC, mass of tissue as a covariate, and tank and vial number as random effects. Of the 39 tissue MR trials, one trial from the UV PCC+ treatment was excluded from the data set because it was identified as an outlier. We present the results with this data point excluded; however, retaining this data point in the analysis does not change the conclusions. Activity data were arcsine square root transformed to satisfy assumptions of normality and homogeneity of variance, and were analysed using a two-way mixed model anova with UVR (present or absent) and PCC (present or absent) as fixed factors, an interaction term between UVR and PCC, and tank and batch number as random effects.

Results

Activity

Tadpoles that were not exposed to UVR or PCC spent nearly one-third of the time being active (Fig. 2a). Exposure to UVR caused tadpoles to significantly reduce their activity (F1,16 = 36·1, P < 0·001), as did exposure to PCC (F1,16 = 6·8, = 0·02), with no interaction between these two fixed factors (F1,16 = 2·2, P = 0·16) (Fig. 2a). Exposure to UVR caused PCC and PCC+ tadpoles to reduce their activity by 56% and 40%, respectively (Fig. 2a). Exposure to PCC caused UV and UV+ tadpoles to reduce their activity by 36% and 12%, respectively (Fig. 2a).

Figure 2.

 The independent and interactive effects of chronic exposure to ultraviolet radiation (UVR) and predatory chemical cues (PCC) on: (a) the proportion time spent being active; and (b) whole-animal metabolic rate (inline image, mL O2 kg−1 h−1) of Limnodynastes peronii tadpoles. (c) The effect of chronic exposure to UVR on the metabolic rate (inline image, mL O2 kg−1 h−1) of Limnodynastes peronii tail tissue. UV and UV+ denote UVR absence and presence, respectively, and PCC and PCC+ denote PCC absence and presence, respectively (a: n = 30 for all treatments; b: UV PCCn = 19, UV PCC+n = 26, UV+ PCCn = 23, and UV+ PCC+n = 24; c: UVn = 19, UV+n = 19). Exposure to UVR caused tadpoles to significantly reduce their activity (< 0·001), as did exposure to PCC (P = 0·02), with no interaction between these two fixed factors (P = 0·16). For whole-animal inline image, there was a significant interaction between UVR and PCC (P = 0·02) with significant and non-significant differences detected between treatment groups connected by stars and ‘NS’, respectively. Exposure to UVR caused tissue inline image to significantly increase, as denoted by the star (< 0·001), but there was no significant effect of exposure to PCC (P = 0·20), with no interaction between these two fixed factors (P = 0·07). Data represent means ± SE.

Whole-Animal Metabolic Rate

Tadpoles that were not exposed to UVR or PCC had a mass-specific oxygen consumption rate of 286 mL O2 kg−1 h−1 (Fig. 2b). There was no significant main effect of UVR (F1,20 = 0·4, P = 0·55) or PCC (F1,23 = 0·6, P = 0·45) on the rate of oxygen consumption of tadpoles, but there was a significant interaction between these two fixed factors (F1,21 = 6·4, P = 0·02) and the effect of mass was significant (F1,84 = 59·4, < 0·001). A Student’s t post-hoc analysis without an adjusted alpha value (Quinn & Keough 2002; Nakagawa 2004) revealed that within the PCC treatment, exposure to UVR significantly reduced the oxygen consumption rate of tadpoles (P = 0·04) with mass-specific oxygen consumption rate being reduced by 14% (Fig. 2b). Within the PCC+ treatment, however, exposure to UVR did not significantly affect the oxygen consumption rate of tadpoles (P = 0·17) (Fig. 2b). Within the UV treatment, exposure to PCC did not significantly affect the oxygen consumption rate of tadpoles (P = 0·23) (Fig. 2b). Within the UV+ treatment, however, exposure to PCC significantly increased the oxygen consumption rate of tadpoles (P = 0·04) with mass-specific oxygen consumption rate being increased by 21% (Fig. 2b). There was also no significant difference in the oxygen consumption rate of tadpoles exposed to both UVR and PCC simultaneously and those exposed to neither stressor (P = 0·90).

In Vitro Tissue Metabolic Rate

There was a significant effect of UVR (F1,16 = 16·6, < 0·001), but not PCC (F1,24 = 1·7, P = 0·20) on the oxygen consumption rate of tadpole tail tissue. There was no interaction between these two fixed factors (F1,16 = 3·9, P = 0·07) and the effect of mass was significant (F1,24 = 56·6, < 0·001). The pooled mass-specific oxygen consumption rate of tail tissue from tadpoles not exposed to UVR was 44 mL O2 kg−1 h−1, while the mass-specific oxygen consumption rate of tail tissue from tadpoles exposed to UVR was significantly higher at 60 mL O2 kg−1 h−1 (Fig. 2c), an increase of 36%.

Discussion

Independent Effects of UVR

This is the first study to show that exposure to UVR causes the MR of animal tissue to increase, demonstrating that exposure to UVR is energetically costly for an animal. Interestingly, however, while exposure to UVR caused the MR of tadpole tail tissue to increase by 36%, it caused tadpoles to reduce their whole-animal MR by 14%. Formicki, Zamachowski & Stawarz (2003) similarly found that exposure to UVBR caused common toad Bufo bufo tadpoles to reduce their rate of oxygen consumption. Formicki, Zamachowski & Stawarz (2003) noted, however, that the movement of tadpoles within respirometers was not completely restricted, as was also the case within the present study, and thus suggested that the changes in oxygen consumption may be attributed to UVR-induced changes in locomotor activity. Unlike Formicki, Zamachowski & Stawarz (2003), we measured activity levels of tadpoles and found that exposure to UVR caused tadpoles to reduce their activity levels by 56%, a response that has been observed in other amphibian species (Zaga et al. 1998; Garcia, Paoletti & Blaustein 2009). Less energetic investment in locomotion is therefore likely to contribute to the observed reduction in whole-animal MR of L.  peronii.

In addition to reduced activity levels, the reduction in L. peronii whole-animal MR may also be attributable to an accumulation of UVR-induced cellular damage. Fischer et al. (2006) suggested that the cause for reduced rates of oxygen consumption in Daphnia catawba following UVBR exposure was excessive damage to various cellular components including those involved in the repair of UVR-induced damage and those with important structural or metabolic roles. Fischer et al. (2006) hypothesized that such widespread damage could depress metabolism, and that a reduction in MR is perhaps a precursor to UVBR-induced mortality. Indeed Fischer et al. (2006) found that the level of UVBR that reduced the MR of D. catawba caused all individuals to die within 2 days of being exposed. The UVR levels used in the present study, however, have been shown not to be lethal to L. peronii tadpoles (Alton, Wilson & Franklin 2011), though survival was not measured in the present study. Thus, while the UVR levels used in the present study may have caused cellular damage that contributed to the observed reduction in whole-animal MR, the observed increase in tissue MR of L. peronii is suggestive of functioning repair mechanisms which may have prevented the accumulation of lethal levels of UVR-induced damage. We therefore suspect that reduced activity levels, not excessive cellular damage, is primarily responsible for reduced whole-animal MR in UVR-exposed L. peronii tadpoles.

It is also worth noting that D. catawba reduced their MR by approximately 80% (Fischer et al. 2006), which is far more severe than what we observed in L. peronii. D. catawba also had no access to photoreactivating light during UVBR exposure which is necessary for DNA repair by photolyase enzymes. Interestingly, Formicki, Zamachowski & Stawarz (2003) found that when B. bufo tadpoles were exposed to UVBR only as tadpoles they also reduced their MR by 80%, but B. bufo tadpoles that had been exposed previously to UVBR during embryonic development reduced their MR by only 22%, which is more similar to our findings in L. peronii. These studies combined indicate that the initial exposure to UVR can be incredibly detrimental to animal health, particularly if the primary DNA repair mechanism of EPR is unable to function because of a lack of photoreactivating light. Consequently, a metabolically depressed state appears to be a symptom of widespread cellular damage and is likely to precede mortality as suggested by Fischer et al. (2006). There does, however, appear to be potential for animals to up-regulate repair mechanisms in response to initial UVR exposure, particularly if initial exposure occurs during early development. This along with behavioural adjustments then seems to enable animals to cope more adequately with the detrimental effects of UVR and prevent them depressing their MR so severely.

Independent Effects of Predatory Chemical Cues

The effect of exposure to predators and PCC on the MR of prey individuals has been the focus of few studies. Researchers interested in the response of animal prey to predators and PCC are primarily focussed on induced defences, such as changes in behaviour and morphology, and their associated fitness costs (e.g. McCollum & Van Buskirk 1996; Van Buskirk 2000; Van Buskirk & Saxer 2001). The physiological response of animal prey to PCC, and hence the potential mechanistic basis for fitness consequences, however, remains relatively unknown. In the present study, tadpoles responded to PCC by reducing their activity, which is a common defensive strategy adopted by prey that reduces their chances of encountering, and being detected by, predators (Lima & Dill 1990; Werner & Anhold 1993; Lima 1998). A fitness cost generally linked to reduced activity is decreased growth because it is thought that those animals that are less active spend less time foraging and therefore consume less food (Lima & Dill 1990; Werner & Anhold 1993; Lima 1998). In a previous study on L. peronii, we found that tadpoles exposed to PCC derived from dragonfly nymphs feeding on L. peronii were smaller in size (and developed slower; L. Alton, unpublished data) compared with those that were not exposed; however, food was always available and tadpoles were reared in individual containers and so did not have to compete for resources (Alton, Wilson & Franklin 2010). This then raises the question: can reduced activity and feeding be solely responsible for reduced growth and development rate? Steiner (2007) addressed this question for pool frog Rana lessonae tadpoles and found that although exposure to predators causes them to reduce their overall activity, and more specifically their feeding activity, they ingest similar amounts of food compared with tadpoles not exposed, and even digest food more efficiently. Despite these apparent advantages, however, R. lessonae tadpoles exposed to predators still suffered from reduced rates of development (Steiner 2007). This therefore implies that there are additional factors that contribute to reduced growth and development rates of tadpoles exposed to predators and PCC.

An elevated MR and increased allocation of resources to morphological defences have both been suggested as potential mechanisms that may explain reduced growth and development rates of animals exposed to predators (Harvell 1990; Stoks 2001; Steiner 2007; Slos & Stoks 2008). An elevated MR may arise as a consequence of the fight-or-flight response which is the release of stress hormones that leads to increased ventilation and heart rates, increased cardiovascular tone and blood pressure, and causes energy stores to be mobilized and made available to the muscles involved in locomotion (Sapolsky 2002). In the present study, exposure to PCC had no significant effect on the whole-animal MR of L. peronii tadpoles, but because tadpoles reduced their activity levels by 36%, no change in MR indicates an elevated MR relative to activity levels. This then suggests that there is an energetic cost associated with being exposed to PCC, and may explain why tadpoles exhibit reduced growth and development rates when reared in the presence of predators.

The effect of PCC on whole-animal MR is not reflected at the tissue level, however, with exposure to PCC having no significant effect on the tissue MR of L. peronii. If an elevated whole-animal MR is a consequence only of the fight-or-flight response, then the lack of an effect on tissue MR may suggest that it is evident only in intact living animals, and not in tail tissue. If, however, the production and maintenance of morphological defences also contributes to the energetic costs associated with exposure to PCC, then the lack of an effect at the tissue level has additional implications. Exposure of tadpoles to PCC generally results in changed tail morphology, which acts to either enhance tadpole escape swimming performance or divert the attention of predators away from the head of the tadpole (Wilson, Kraft & Van Damme 2005; Johnson, Burt & DeWitt 2008). With regard to L. peronii specifically, changes in tail morphology are detectable after 9 days of exposure to PCC (Alton, Wilson & Franklin 2010), but are more strongly expressed after 28 days exposure (Kraft, Wilson & Franklin 2005). Thus, because we measured whole-animal MR after 9 days of exposure to PCC and the MR of tadpole tail tissue after 25 days exposure to PCC, the non-significant effect of PCC on tissue MR may suggest that while the production of morphological defences may be energetically costly, once these morphological defences are developed, their maintenance is not.

Alternative to the aforesaid hypotheses, the difference between our findings on the whole-animal and tissue MR of L. peronii may indicate that tadpoles are capable of acclimating to predation stress as suggested by Steiner & Van Buskirk (2009). They found that acute exposure to PCC resulted in an increase in MR of common frog Rana temporaria tadpoles, but chronic exposure resulted in a decrease in MR (Steiner & Van Buskirk 2009). They therefore concluded that although the initial metabolic response of tadpoles to PCC represents an energetic cost, changes in MR cannot account for reduced growth and development rates because this metabolic response is reversible and tadpoles actually acclimate to predation risk by lowering their MR (Steiner & Van Buskirk 2009). The data presented by Steiner & Van Buskirk (2009), however, also show that tadpoles reared in the presence of predators and measured in water containing PCC had similar MRs to tadpoles not reared with predators and measured in water without PCC (see Fig. 1 in Steiner & Van Buskirk 2009). Unlike the present study, Steiner & Van Buskirk (2009) do not consider this comparison, nor do they account for changes in tadpole activity. We therefore feel that Steiner & Van Buskirk’s (2009) conclusions should be viewed cautiously, and, based on our findings, there is certainly potential for an increased MR to explain reduced growth and development rates in tadpoles. Further research into the effects of PCC on tadpole MR is clearly needed, however, and should endeavour to determine how both whole-animal and tissue MR changes over the course of tadpole development. Such research could potentially give greater insight into the energetic costs associated with changes in morphology along with the ability of tadpoles to acclimate to predation stress.

Interactive Effects of UVR and Predatory Chemical Cues

In our previous study on L. peronii, we found that exposure to UVR alone reduced tadpole survival by 17%, whereas exposure to PCC alone had no significant effect on tadpole survival (Alton, Wilson & Franklin 2010). Interestingly, when we exposed tadpoles to UVR and PCC simultaneously, tadpole survival was reduced by 37% (Alton, Wilson & Franklin 2010). Here, we show that the cause for this synergistic interaction on tadpole survival may be due to the combined effect of these stressors on tadpole MR, i.e. tadpoles exposed to UVR and PCC simultaneously exhibited no change in whole-animal MR despite reducing their activity levels by 62% compared with tadpoles exposed to neither stressor. Given that tadpoles exposed to UVR alone reduced their activity levels by a similar margin and consequently showed a reduction in whole-animal MR, it might have been expected that the tadpoles exposed to both UVR and PCC would exhibit a similar reduction in MR. The fact that these tadpoles did not reduce their whole-animal MR gives further support to our hypothesis that exposure to PCC is energetically costly, and indicates that exposure to both these stressors has a compounding influence on MR. Unfortunately, however, because of the complicating factor of activity and because we saw no effect of PCC on tissue MR, we cannot know whether this compounding influence on whole-animal MR is additive or synergistic. Despite this, however, it is clear that being exposed simultaneously to UVR and PCC is more energetically costly than being exposed to either stressor independently. Therefore, if the combined cost of these stressors exceeds the metabolic capacity of a tadpole, prolonged exposure to UVR and PCC may result in death. To better establish a link between the energetic cost of simultaneous exposure to UVR and PCC and reduced tadpole survival, future research should measure survival in addition to MR and activity using lethal doses of UVR.

Conclusions

The phenomenon of global amphibian declines is a testament to the profound effects of human-induced global change on natural environments. With many amphibian populations declines likely to be attributable to multiple environmental stressors (Blaustein & Kiesecker 2002; Linder, Krest & Sparling 2003), both anthropogenic and natural, it is important that we examine the mechanisms by which stressors interact if we are to understand the cause for amphibian declines. Here, we demonstrate that exposure to UVR and PCC independently is energetically costly for L. peronii tadpoles and, more importantly, that this cost is greater when they are exposed to both stressors simultaneously. Measurement of tissue MR and activity levels were critical for the interpretation of the effects of UVR and PCC on whole-animal MR, indicating that future research should consider these response variables when evaluating the potential energetic costs associated with exposure to various environmental stressors.

The findings from the present study suggest that global increases in UVBR are potentially contributing to global amphibian declines by compromising energy allocation towards growth and development as well as energy allocation towards coping with additional environmental stressors. With large increases in damaging UVR predicted to occur in coming years because of climate change (Hegglin & Shepher 2009), it is important that research continues to investigate the interactive effects UVR may have with additional environmental stressors to assess the current and future role UVR may have in affecting amphibian populations.

Acknowledgements

This research was funded by The University of Queensland Research Grant to CEF and The Australian Research Council (grant DP0987626 to CRW). All experimental procedures were approved by The University of Queensland Native and Exotic Wildlife and Marine Animals Ethics Committee (SIB/626/08/URG). We thank anonymous reviewers for their comments on earlier versions of this manuscript.

Ancillary