The role of solar radiation and tidal emersion on oxidative stress and glutathione synthesis in mussels exposed to air

Fieldwork

Fieldwork was conducted in a rocky shore in the city of Penha (Santa Catarina State, Brazil) at 26°47′12.2″S, 48°36′18.2″W. On November 2nd (2014), adult mussels (Brachidontes solisianus, Bivalvia: Mytilidae) weighting 0.2–0.4 g were collected from a single rock at several time-points from 7:30 a.m. to 9:25 p.m. During this period the animals were exposed to air for at least 2 h (>2 h-AE group), submerged for 7 h (1 h-W, 3 h-W, and 7 h-W groups), and then exposed to air again for 4 h (0.5 h-AE and 4 h-AE groups). In summary, mussels were exposed to air twice: first in the early morning (with very little exposure to sunlight) and then in the evening (after a whole day of sunlight exposure). Animals were taken from the rock by hand, quickly cleaned and flash frozen in liquid N₂ on site. At the end of the day, the frozen specimens were transported to the laboratory and stored at −80 °C. This study was conducted under the license issued by ICMBio and SISBIO (permit number 20030–9). The experimental groups that were collected on November 2nd, 2014, were exactly the same experimental groups collected on the day before (November 1st, Fig. 1), whose biochemical data have already been published (Moreira et al., 2021b). The site of collection was exactly the same in both days and each individual mussel was an experimental unit, with an N of 3–11 per experimental group/variable (all non-enzymatic redox variables presented N ≥7).

Sampling was done on the same beach-rock with a height of approximately 2 ft. During the emersion periods, mussels were completely exposed to air and presented closed valves. In contrast, their shell valves were clearly open when underwater. Collected bivalves when underwater down to 1 to 2 ft from the surface. During both collection days, seawater was transparent down to the depth and location where mussels were collected. Water transparency in the site, evaluated by the Secci disk, ranges from 5 to 12 ft along all year long (Resgalla et al., 2007). Furthermore, B. solisianus are vertically oriented on the substrate, perpendicularly to the surface, which favors solar radiation reaching the shells’ interior despite the narrow aperture. The shells for adults of this species are fully opaque. Thus, solar radiation cannot reach the internal organs when shut (Abbott, 1974; Ruppert, Fox & Barnes, 2004; Monteiro-Ribas et al., 2006).

In addition to the recording of seawater and air temperatures on site using a mercury thermometer, global solar radiation (GSR) was determined with a pyranometer (CM6B, Kipp & Zonen, Delft, Netherlands) located at the Itajaí-A868 weather station from the National Meteorological Institute (INMET), 25 km from the mussel collection site. GSR strongly relates to solar UVR, and cloudy periods present significantly smaller GSR, with proportionally less UVR when compared with sunny periods (Porfirio et al., 2012; Wang et al., 2013). We calculated the cumulative global solar radiation by summing hourly GSR up to the respective hour.

Mussel collection happened on a clear sky day on November 1st, 2014, and a cloudy day on November 2nd. Thus, days 1 and 2 of collection are also called “sunny” and “cloudy”. Thus, the following groups were formed for each day: >2 h-AE group, 1 h-W, 3 h-W, 7 h-W, 0.5 h-AE, and 4 h-AE; a total of 12 experimental groups (six collections points ×2 days). For comparison purposes, the group of mussels underwater for 7 h was appointed as the “control” group.

Antioxidant and metabolic enzymes

Mussels were individually weighted and then pulverized in liquid nitrogen. The pulverized sample was transferred to a glass tissue homogenizer (Tenbroeck type) containing 0.1 M Tris (pH 7.6), 5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride and 1:1,000 (v/v) protease inhibitor cocktail (P8340; Sigma Aldrich, St. Louis, MO, USA). After homogenization, samples were centrifuged at 10, 000 × g for 15 min at 4 °C and the supernatants were collected for the assessment of enzymatic activities. The activities of catalase, glutathione transferase (GST), glutathione reductase (GR), glutathione peroxidase (GPX), glucose 6-phosphate dehydrogenase (G6PDH), pyruvate kinase (PK), malate dehydrogenase (MDH) and isocitrate dehydrogenase (IDH) were measured using spectrophotometric kinetic assays as described in Moreira et al. (2021b). The activities of opine dehydrogenase (ODH), strombine dehydrogenase (SDH), and alanopine dehydrogenase (ADH) were also measured using spectrophotometric kinetic assays, as described in Schiedek (Schiedek, 1997).

The activity of all enzymes was expressed as U per milligram of protein in the supernatant, which was measured using Coomassie brilliant blue G-250 and a standard curve built with bovine serum albumin as standard (Bradford, 1976).

Glutathione and oxidative stress markers

For the measurement of reduced (GSH), disulfide (GSSG) and total glutathione (tGSH = GSH + 2 ×GSSG) concentrations, as well as thiobarbituric acid reactive substances (TBARS) and protein carbonyl levels, mussels were individually weighted and then pulverized in liquid nitrogen. The pulverized sample was transferred to a glass tissue homogenizer (Tenbroeck type) containing trichloroacetic acid (TCA) and homogenized. The crude homogenate was used to measure TBARS levels as described by Buege & Aust (1978). Another set of homogenates was centrifuged at 10,000 × g for 6 min at 4 °C; the supernatants were used to measure GSH, GSSG, and tGSH levels in the enzymatic recycling method (Tietze, 1969; Griffith, 1980) with modifications (Welker, Moreira & Hermes-Lima, 2016), and the pellets were used to measure protein carbonyl content as described by Fields & Dixon (1971). Additional details can be found in Moreira et al. (2021b).

Statistics

Before any statistical analysis, we compiled the original data generated in the present study (day 2) with those already published in a similar study (day 1, Moreira et al., 2021b). The reasons are: the experimental groups (time of air exposure and time underwater) in the two datasets are the same; the animals were collected from the very same rock; small variability in the moment of animal collection (∼10 min; Fig. 1). First, data were tested for normality by group. Secondly, UVR and temperature values on days 1 and 2 were compared using t-tests. In order to compare within animals’ groups, a two-way ANOVA was used for each variable using two factors: day and time under condition (underwater or exposed to air). For glutathione’ parameters, we compared air-exposed groups (“>2 h-AE”, “0.5 h-AE” and “4 h-AE”) and underwater animals (“1 h-W”, “3 h-W” and “7 h-W” groups combined, yielding the large group “1,3,7 h-W”). Considering that glutathione results from the three underwater groups were unchanged, the clustering simplifies comparisons, increases the number of underwater replicates, and is more representative of underwater “control” animals. When necessary, we performed a Tukey’s test. Two-way ANOVAs were also used for data on enzymatic activity (catalase, GST, GR, and G6PDH), comparing air-exposed (“0.5 h-AE” and “4 h-AE”) with underwater mussels (“3 h-W” and “7 h-W” combined as “3,7h-W”) in sunny and cloudy days. Two-way ANOVA tests were also employed for metabolic enzymes, TBARS, and protein carbonyl levels, comparing air-exposed with underwater animals on both days.

In addition to the biochemical data of animals collected on November 2nd (day 2), we are also presenting novel data on global solar radiation on both days and enzymatic activity of enzymes involved in anaerobic metabolism in animals collected on November 1st (day 1) and November 2nd (day 2).

Environmental data

Global solar radiation (GSR) was higher in day 1 (sunny) (mean: 1,183 kJ/m²/h) than in day 2 (cloudy) (mean: 422 kJ/m²/h) (paired t-test = 3.0, p = 0.0102) with variations along the day (Fig. 2). Total cumulative GSR was 15,381 kJ/m² for day 1 and 5,489 kJ/m² for day 2. Air and water temperature values ranged narrowly between 23 °C and 26 °C and were similar in the two days between 6:00 and 21:30 (air temperature t-test = 0.42, p = 0.67 and water temperature t-test = 0.04, p = 0.96).

Glutathione

Following the same approach on day 1 (Moreira et al., 2021b), we first measured glutathione levels, as total glutathione (tGSH), reduced glutathione (GSH), disulfide glutathione (GSSG), and disulfide/total ratio (GSSG/tGSH). Two-way ANOVA tests revealed that air exposure had a significant effect on tGSH and GSH levels, whereas the day of animal collection had no significant effect (File S1). Moreover, a significant interaction between the effects of air exposure and day of collection on both tGSH and GSH was detected (File S1), indicating that the effect of air exposure on these variables depended on the day of collection. In other words, since the two days differed mainly in terms of solar radiation incidence, the effect of air exposure on tGSH and GSH depended on the degree of radiation exposure. Multiple comparisons analyses identified that the concentrations of tGSH (Fig. 3A) and GSH (Fig. 3B) in mussels exposed to air for 4 h (4 h-AE) in day 1 were significantly different from those in mussels underwater (1, 3, 7 h-W) in either day 1 or day 2 (File S1). Levels of tGSH were also higher in animals exposed to air for 4 h in day 1 versus those exposed to air for 4 h in day 2 (File S1) The average increase after 4 h of air-exposure for the two glutathione variables (GSH and tGSH) in relation to the underwater group was ∼1.6 fold.

A two-way ANOVA showed no statistically significant interaction between the effects of air exposure and day of collection, as well as no simple main effect of these factors, on GSSG levels (Fig. 3C; File S1). In the case of GSSG/tGSH (Fig. 3D), a significant effect of air exposure was detected (File S1). However, multiple comparisons tests did not detect significant differences between experimental groups. No significant effect of day of collection or the interaction between day and air exposure on GSSG/tGSH was detected (File S1).

Antioxidant enzymes and oxidative stress markers

Next, we focused on fewer groups to investigate any changes in the activity of antioxidant enzymes (Table 1) or oxidative stress markers (Fig. 4). Two-way ANOVA tests did not reveal any significant effect of day of collection, air exposure or interaction between them on catalase, GST, and G6PDH activities (File S1). A significant interaction between the effects of day of collection and air exposure on GR activity was detected, but multiple comparison tests did not reveal any differences between groups (File S1).

A significant interaction between the effects of day of collection and air exposure on TBARS content was detected (File S1) and multiple comparison tests indicated that animals exposed to air for 4 h on day 1 had higher TBARS levels than those in mussels underwater for 7 h on day 1 (Fig. 4A). Similarly, the effects of day of collection and air exposure on carbonyl protein levels had a significant interaction (File S1). Multiple comparisons tests revealed that carbonyl protein levels in mussels exposed to air for 4 h were higher than those in mussels underwater for 7 h on day 1 and mussels exposed to air for 4 h on day 2 (Fig. 4B).

Metabolic enzymes

There was no significant effect of day of collection, air exposure or interaction between them on the activities of intermediary metabolism enzymes (Table 2) and anaerobic metabolism enzymes (Table 3; File S1).

The transition from day 1 to day 2

It remains to be elucidated how increased levels of GSH and oxidatively damaged proteins and lipids from the evening of the sunny day decreased to basal levels in the morning of the next day? Several mechanisms might have contributed to this outcome, including repair and replacement (indirect repair) processes. Oxidized proteins, for example, are degraded by proteases to their amino acids, some oxidized others undamaged. Those undamaged amino acids are then used for a new synthesis of proteins, closing a recycling process. The lack of a known pathway that directly reverses protein-carbonyl modifications (Moskovitz & Oien, 2010) further supports the idea that oxidized proteins were repaired (directly or indirectly) from one day to another. A similar process also happens for oxidatively damaged lipids. Phospholipases, some with an augmented affinity for oxidized lipids (van Kuijk et al., 1987), hydrolyze these compounds, decreasing their levels. Other enzymes, such as glutathione peroxidase and glutathione transferases, act on lipid hydroperoxides and lipid peroxidation products, respectively. In the case malondialdehyde, aldehyde dehydrogenase and decarboxylase initiate their degradation to CO₂ in a series of enzymatic steps (Ayala, Muñoz & Argüelles, 2014). Lastly, since oxidized lipids and proteins returned to basal levels in the early morning of day 2 in “our mussels”, the degree of oxidative damage to proteins, possibly caused by UVR exposure plus aerial exposure, did not overcome mussels’ repair and replacement capacity.

A comparable scenario was observed in Perna perna mussels exposed to air under laboratory conditions. After animals were exposed to air for 24 h, the levels of oxidative damage to DNA increased in the gills and digestive gland, and the degree of lipid peroxidation increased in the digestive gland (Almeida et al., 2005). Three hours after re-submersion, the levels of these oxidative stress markers returned to baseline levels in both tissues (Almeida et al., 2005). The occurrence of repair mechanisms is supported by the upregulation of proteins involved in repairing oxidative damage in the salt marsh mussel Geukensia demissa during recovery from air exposure (Fields et al., 2014). Other evidence of the upregulation of damage repair is the increase in the activity of GST in the digestive gland of P. perna mussels during recovery from air exposure (Almeida et al., 2005).

At least two processes might explain the decrease in GSH concentration from ∼0.2 mM to ∼0.1 mM in mussels during night-time, between the evening of the sunny day to the early morning of the cloudy day. The first is the consumption of glutathione in the repair mechanisms presented above. The conjugation of glutathione with exogenous and endogenous electrophiles, such as lipid peroxidation products, would remove it from the tGSH pool (i.e., it does not simply oxidize it to GSSG). The second is the degradation of GSH to its corresponding amino acids or dipeptides. Two recently discovered enzymes are involved in cytosolic glutathione degradation in animals. They are the ChaC1 and ChaC2 enzymes, members of the gamma-glutamylcyclotransferases family. These enzymes degrade glutathione into 5-oxoproline and Cys-Gly (Kaur et al., 2017; Bachhawat & Yadav, 2018). The amino acids resulting from GSH degradation could, in turn, be used for new synthesis of proteins and peptides. Nevertheless, studying how enzymes involved in GSH biosynthesis and degradation regulate glutathione homeostasis and the POS-phenotype in mussels is a new challenge.

POS mechanism induced by UVR

In our system, the peak of solar UVR (from about 10 AM to 3 PM) does not match the moment of increased variables of redox metabolism (glutathione, TBARS and carbonyl protein). When these biochemical variables increased the solar radiation had already decreased. However, the critical factor is the cumulative exposure to UVR (lasting 9 h, during which animals were underwater with valves open), which is maximal when the low tide starts. UVR triggers several cellular/biochemical events that increase RONS formation. In our hypothesis, the level of cellular stress able to induce oxidative damages and the POS response (increased GSH synthesis) is attained only by a further increase in RONS formation in aerial exposure. Such increase was recorded only after 4 h of low tide –it could have happened earlier than 4 h, but later than 30 min of aerial exposure (it is relevant to note the high variability in tGSH levels at 30 min of air exposure in the sunny day; 1/3 of the specimens had over 0.3 mM tGSH). Thus, in our setting neither of the abiotic events alone (UVR and aerial exposure) could induce significant changes in biochemical variables related to redox metabolism. They had to happen in synergism –one right after the other –to induce the POS phenotype.

Temperature effect

One environmental variable that could have affected the redox metabolism in B. solisianus is temperature. The ability of high temperatures to influence the redox metabolism in tissues of aquatic animals is well-known. However, there is not a clear pattern of antioxidant responses to temperature, and higher temperatures do not necessarily induce oxidative stress nor activation of endogenous antioxidants in tissues of various aquatic invertebrate species, being a species-specific response (Matozzo et al., 2013; Vinagre et al., 2014; Jimenez et al., 2016; Kamyab et al., 2017; Moreira et al., 2017a; Grilo et al., 2018; Feidantsis et al., 2020). In our case, temperature oscillated minimally along the sampling hours of both days, varying from 23 °C to 26 °C, either in air or seawater. There was also no relevant difference in temperatures between sunny and cloudy days. In addition, temperatures peaked in mid-afternoon, when mussels were underwater with unchanged redox biochemical markers. Thus, the observed increased in redox markers in the evening if day 1 was probably unrelated to temperature.

Ecological biochemistry field studies

Our main motivation to obtain data from animals collected directly from the field environment, with no intervention at all, is to take a glimpse at the biochemical response of animals in a real-world situation (Spicer, 2014). Data from neuroendocrine studies indicate that the results from animals kept in captivity might differ, in some cases drastically, from those obtained from field animals (Calisi & Bentley, 2009). Although completely natural field experiments, as done here, do not strictly control experimental conditions (e.g., temperature, solar radiation, etc.) that might influence results, they are as close to reality one can get, providing valuable insights into the POS-mechanism in the actual environment (Spicer, 2014). On the other hand, it can be very hard to accurately reproduce the conditions of one experiment to another, given the large natural variations in the environment. Despite obtaining biochemical data for only two days of field work, this study managed to collected animals in a set of six different natural conditions with relevant characteristics and differences in terms of environmental variables. N ≥ 7 animals were collected for each of these groups to measure glutathione and oxidative stress markers. Six major natural conditions were covered: (i) early morning low tide, day 1; (ii) early morning low tide, day 2; (iii) high tide, day 1 (late morning plus afternoon); (iv) high tide, day 2 (late morning plus afternoon); (v) late afternoon plus evening low tide, day 1; and (vi) late afternoon plus evening low tide, day 2.

Although we, purposely, did not control every environmental variable as in lab research, we could compare animals exposed to a very similar tidal cycle (i.e., temperature, time of exposure, duration) but with a tremendous (i.e., 2.8-fold) difference in terms of global solar radiation. This is why we attribute the observed biochemical changes in oxidative stress markers and glutathione to a possible effect of UVR (POS mechanism induced by UVR). However, we cannot rule out the effects of other environmental factors whose data we do not have (e.g., salinity, microbiota, and contaminants). But, for example, salinity is unlikely to have changed significantly from one day to another, since natural variations in salinity occur seasonally rather than on a daily basis in this region (Resgalla et al., 2007). As mentioned above, air and seawater temperatures were remarkably stable in the period of animal collection, and is unlikely to have played any role. Nevertheless, our hypothesis needs to be further tested under strictly controlled settings in the lab. Another effect we cannot ignore is any inter-day influence on the biochemical parameters. We do not know what happened the day before day 1 and we cannot exclude that the events occurring in day 1 affected the response on day 2. New field expeditions for the study the solar radiation in mussel redox variables would be desirable to strengthen our hypothetical explanations.

Finally, we calculated the possibility that all three redox indicators (glutathione, TBARS and carbonyl protein) become increased only during the evening low tide of day 1 (i.e., in only one condition, out of six). This is a situation where air exposure happens after mussels had been exposed underwater to intense solar radiation for hours. To simplify calculations, we just tabulated the variables as “not increased” (summing “no significant change” and “decreased” in this category) or “increased”. So, the possibility that redox biomarkers are all increased in this particular condition (evening low tide of day 1) would be just 1 in 816, or 0.12%. Therefore, it is highly improbable that the increase in all three redox indicators happened by chance at “condition 3” (see Table 4).

Controlled experiments versus observational biology

It is fundamental to differentiate the present observational results, also called natural experiments (Diamond, 1983), obtained from wild animal sampling, to those of strictly controlled laboratory studies, where researchers run a well-defined control group versus experimental ones. As stated in the Introduction section, almost all animal studies on the POS response were done in the lab, except for a small list, including a mussel work from Moreira et al. (2021b). When we performed the bivalve captures in 2014, we did not anticipate the possibility that solar UVR could play any role in the redox metabolism of mussels. This only became clear when we analyzed samples from another expedition-day when no changes were observed in any of the metabolic and redox variables. The sole difference between the two days of animal sampling was solar radiation. So, the present study is an attempt to rationalize this unpredicted observation. According to Jared Diamond, “the obvious weakness of natural experiments is that the observer does not create a known difference between two situations, but must instead decide what difference between two existing situations is the salient one.” (Diamond, 1983). In our case, it is solar radiation. It would have been better if this had started as a lab-controlled experiment. However, the trade-off for the inability to manipulate the wilderness is more realism (Diamond, 1986). Still, this proposed effect of UVR is actually a “side effect” of our search for the POS-mechanism in nature. Therefore, it is a must that such mechanism of “POS-UVR” be analyzed for other intertidal bivalve species.