The alternative oxidase (AOX) increases sulphide tolerance in the highly invasive marine invertebrate Ciona intestinalis

Ecological communities and biodiversity are shaped by two main factors, climate change and invasive species (Mainka and Howard, 2010). With the currently ongoing change in climate, many environments have already experienced alterations in their characteristics, and will continue to do so, while extremes increase in abundance and severity, consequently affecting ecological communities (Bellard et al., 2012; Mainka and Howard, 2010; Worm and Lotze, 2021). Invasive species play another important role in shaping ecological communities and biodiversity (Bax et al., 2003; Molnar et al., 2008). By nature, invasive species have to be well prepared to cope with diverse stressors, and this in turn may be one key feature that enables them to compete in a new environment. Therefore, understanding the underlying mechanisms of how organisms in general and invasive species in particular respond to extreme environments is crucial to improve predictions for ecological communities and biodiversity, especially in the context of climate change.

The oceans, covering approximately 70% of the Earth's surface, play a central role in climate change. With increased carbon dioxide levels in the atmosphere, the oceans are greatly affected by changing temperature (becoming warmer) and pH (becoming more acidic). The consequences of those changes are manifold, ranging from less oxygenated oceans to rises in sea levels, changes in ocean currents and increasing weather extremes (Diaz and Rosenberg, 2008). One additional consequence with considerable impact on the survival of multicellular organisms has rarely been mentioned in the climate change debate: the increase in hydrogen sulphide (H₂S) in the oceans. In general, high sulphide concentrations are due to geothermal and biological processes (Bagarinao, 1992). Interestingly, biological processes relate to oxygen levels: with depleted oxygen levels, bacterial sulphate reduction takes place and H₂S is produced (Jørgensen et al., 1982; Schunck et al., 2013), and eventually moves up the water column (Bakun and Weeks, 2004; Brüchert et al., 2006; Stewart, 2011). This sulphide–oxygen link becomes rather important because the formation, size and abundance of oxygen minimum zones and anoxic bottom waters are increasing (Breitburg et al., 2018; Diaz and Rosenberg, 2008; Stramma et al., 2010), consequently causing bacterial-associated sulphide production.

The reason for the detrimental potential of H₂S is the inhibition of the aerobic energy production pathway (Bouillaud and Blachier, 2011; Cooper and Brown, 2008). Most eukaryotic organisms produce their energy in the form of adenosine triphosphate (ATP), the majority of which is produced in mitochondria via oxidative phosphorylation (OXPHOS). In this pathway, H₂S inhibits cytochrome c oxidase (COX), which limits or blocks ATP production, ultimately leading to the death of the organism. However, diverse ecological communities are found around in environments with high H₂S concentrations that are lethal for most species, including deep-sea hydrothermal vents, cold seeps, marine sediments, freshwater springs and caves (Riesch et al., 2015; Tobler et al., 2016). Those species have manifold behavioural, physiological and biochemical mechanisms to avoid or eliminate H₂S (Abel et al., 1987; Greenway et al., 2014; Jahn et al., 1997; Laudien et al., 2002; Miron and Kristensen, 1993; Pfenninger et al., 2014). Interestingly, many species among almost all kingdoms of life possess a sulphide-resistant, non-protonmotive alternative oxidase (AOX). AOX can substitute for COX and thus ensures the maintenance of energy production, albeit at a decreased level, as well as preserving mitochondrial redox homeostasis, both of which are crucial for mitochondrial function and, ultimately, survival. In protists, fungi, plants and also some animal species, AOX has been shown to be involved in coping with a multitude of abiotic and biotic stressors impairing OXPHOS, such as extreme temperatures, nutrient limitations, metals and pathogen infections (McDonald, 2008; Rogov et al., 2014; Saari et al., 2019; Tward et al., 2019).

In this context, the tunicate Ciona intestinalis is of great interest for two reasons. First, C. intestinalis is a highly successful invasive marine invertebrate that imposes considerable ecological impacts on newly invaded communities (Blum et al., 2007; Therriault and Herborg, 2008). Second, C. intestinalis possesses AOX (McDonald and Vanlerberghe, 2004), which may be part of the reason for its tolerance to sulphide and, hence, its invasive potential.

In adult C. intestinalis, elevated H₂S concentrations elicit a strong response in the AOX transcript levels in heart and neural complex tissue (Saari et al., 2019), indicating that AOX may play a role in mitigating H₂S stress in adults. However, a functional verification of the transcript level response is still lacking and the effect of H₂S on the early developmental stages of C. intestinalis remains unknown. This information, however, is of great interest. These early developmental stages are those that invade new habitats and ecological communities. However, early developmental stages have also been shown to be more susceptible to external stressors compared with adult stages in many species (Pineda et al., 2012).

Although AOX has been detected at the sequence and transcript levels in many animal phyla from Placozoa to Chordata (McDonald et al., 2009; Tward et al., 2019), studies on the functional role of AOX will improve our understanding of how animals tolerate and invade potentially harsh environments. Hence, our aim was to understand the role of AOX during early development of the highly invasive ascidian C. intestinalis. Specifically, we tested whether a lack of AOX protein during early C. intestinalis development causes lower developmental success under elevated H₂S concentrations, relative to control groups. With this study, we aimed to shed light on the ecological implication of AOX in C. intestinalis and consequently other AOX-possessing animals, and thereby encourage a new perspective on the alternative mitochondrial pathway in the context of the ecology of aquatic communities and climate change.

We conducted all animal experiments and survival analyses at the Observatoire Océanologique de Villefranche-sur-Mer, France. The mRNA analyses were performed at the BioMediTech, Tampere University, Finland.

Adult Ciona intestinalis (Linnaeus 1767) were collected off the coast of Brest, France, by the Centre de Ressources Biologiques Marines of the Station Biologique de Roscoff. Prior to experiments, animals were kept in an 18°C, temperature-controlled flow-through system with natural seawater at the station. A detailed visualisation and description of C. intestinalis embryonic development is provided by Hotta et al. (2007).

We performed all experiments at 18°C using artificial seawater (ASW, 420 mmol l⁻¹ NaCl, 9 mmol l⁻¹ KCl, 10 mmol l⁻¹ CaCl₂, 24.5 mmol l⁻¹ MgCl₂, 25.5 mmol l⁻¹ MgSO₄, 2.15 mmol l⁻¹ NaHCO₃, 10 mmol l⁻¹ Hepes buffer, pH 8.0, 0.05 g l⁻¹ kanamycin sulphate to prevent potential bacterial growth in embryo cultures, sterilized with a 0.22 µm filter) and 1% agarose-coated Petri dishes (Ø 55 mm, Gosselin, Hazebrouck, France) (Sardet et al., 2011). In all experiments, we used unrelated families as biological replicates, i.e. sperm and eggs for each family were only used once. On the day of the experiment, eggs and sperm for each family were freshly sampled and eggs were dechorionated and fertilized according to Sardet et al. (2011). For all sulphide treatments, we used a sodium sulphide (Na₂S, 99.99% trace metals basis, no. 431648, Sigma Aldrich, St Louis, MO, USA) stock solution (10 mmol l⁻¹) and used specific volumes to obtain the desired final concentration. In solution, Na₂S dissociates into three sulphidic species that equilibrate quickly into sulphide (S²⁻), sulfanide (HS⁻) and hydrogen sulphide (H₂S). At the average physiological pH of 7.4, H₂S is present at approximately 20–30% and HS⁻ at 70–80%, while S²⁻ is only present in negligible quantity (Szabo et al., 2014). For simplicity and comparable reasons with previous studies that use this approach, we hereafter use ‘sulphide’ to refer to all three species collectively.

To determine the effect of sulphide on the developmental success of C. intestinalis and the concentration that resulted in 50% embryos being dead before completing development to tailbud stage (LC₅₀), we tested five different sulphide concentrations (0, 10, 20, 25 and 50 μmol l⁻¹) on five families with technical duplicates (two Petri dishes per family) per concentration for three of the five families. After fertilization, we distributed zygotes of each family evenly across treatment dishes filled with 10 ml of ASW and directly added sulphide at concentration-specific amounts. We terminated the experiment after 18.5 h of post-fertilization development by adding paraformaldehyde (PFA) to each dish. We then analysed the successful development to tailbud stage by counting the number of embryos that did and did not fully develop to that stage using a stereo microscope and calculating the proportion of those embryos that fully developed to tailbud stage.

In order to investigate whether AOX plays a role in the tolerance to sulphide in C. intestinalis embryos, we conducted knock-down experiments using a morpholino (MO) anti-sense oligomer targeted to the 5'-untranslated region (UTR) of AOX transcripts to block their translation. The set-up for the MO injections is described in Yasuo and McDougall (2018). We sampled eggs and sperm following the same protocol as in the dose–response experiment and tested seven C. intestinalis families on three ‘MO groups’ and four of those families in a rescue group: (1) uninjected eggs (uninjected), (2) control-MO injected eggs (MO_Ctrl), (3) AOX-MO injected eggs (MO_AOX) and (4) AOX-MO and MO-immune AOX mRNA-injected eggs (MO_AOX+AOX mRNA). The rescue group addresses the specificity of AOX-MO; the MO-immune AOX mRNA consists of the AOX open reading frame [obtained from pWPI-AOX (HindIII); courtesy of Dr Eric Dufour, Faculty of Medicine and Health Technology, Tampere University, Finland] and re-cloned into the pBluescript RN3 vector (Lemaire et al., 1995). The standard control-MO (CCTCTTACCTCAGTTACAATTTATA) (Gene Tools, LLC) and AOX-MO (TTCCGGTAGACAACATATTTGTTGC) (Gene Tools, LLC) were injected at 1 mmol l⁻¹, while AOX mRNA was injected at 0.5 μg μl⁻¹. The control-MO targets a human β-globin intron mutation that causes beta-thalassemia and is broadly used as a negative control and it causes no visible phenotype in C. intestinalis. AOX mRNA was synthesized from linearized pRN3-AOX-3HA using the mMESSAGE mMACHINE T3 Kit (Ambion, Thermo Fisher, AM1348). Approximately 1 h after injection and no injection for the uninjected control group, eggs were fertilized and selected for correctly developed embryos when they reached the 8-cell stage. We then distributed embryos of each MO group equally into two Petri dishes (∼50 eggs per dish) and started the experiment by adding sulphide to one of the two dishes to obtain a final concentration of 15 μmol l⁻¹ in a final volume of 10 ml ASW (LC₅₀; Fig. 1). As with the dose–response experiment, we terminated this experiment after 18.5 h of development by adding PFA to each dish and determined the proportion of embryos showing successful development to tailbud stage.

To test the effect of sulphide on AOX transcript levels along developmental stages, we exposed developing C. intestinalis embryos to either LC₅₀ sulphide (15 μmol l⁻¹) or control (no sulphide) conditions. Embryos of four families were collected at egg (control condition only), 32-cell, mid-gastrula, mid-neurula and mid-tailbud stages and immediately frozen and stored at −80°C. We extracted RNA (Arcturus PicoPure RNA Isolation Kit, ABI, Foster City, CA, USA) and removed residual genomic DNA by treatment with RNase-free DNase I (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's protocol, followed by reverse transcription (High-Capacity cDNA Reverse Transcription Kit, ABI, Foster City, CA). Extracted RNAs were quantified spectrophotometrically (NanoDrop 2000, Thermo Fisher Scientific). We determined transcript levels based on qPCR (ABI StepOne Plus instrument and Fast SYBR Green Master Mix, ABI) with gene-specific primer sets [AOX: efficiency=1.81, (Saari et al., 2019); β-actin: efficiency=1.90, (Fujikawa et al., 2010); TCEB3 (transcription elongation factor B polypeptide 3): efficiency=1.87, forward primer (5′–3′) GTATGCCTGGTACGATTCCCAACT, reverse primer (5′–3′) AAACTGCCTATCTTCCAAAGATGCTC (present study)]. On each of four 96-well plates, we included sample duplicates for AOX and two housekeeping genes (β-actin, TCEB3) and no-template controls per gene (no contaminations detected). We used the online tool Real-time PCR Miner (Zhao and Fernald, 2005) to estimate quantification cycle (C_q) values and amplification efficiencies for each qPCR replicate (E; subsequently averaged per gene and plate) from the raw qPCR fluorescence data. To allow analysis of C_q values via linear models, we efficiency-standardized C_q values as E^C_q per gene and plate and then standardized to a common log₂C_q scale across genes and plates as log₂(E^C_q). Accordingly, all linear model contrasts represent the log₂ of traditionally used C_q ratios.

We analysed the dose–response data using a generalized linear mixed model with logit-link function and binomial error distribution under Laplace approximation to the likelihood as implemented in the R function glmer of the R package lme4 (Bates et al., 2015). The proportions of individuals that survived were modelled with a fixed covariate for sulphide concentration (in μmol l⁻¹; representing a logit regression), and random terms for family, family×concentration and family×concentration×replicate (accounting for detected overdispersion). We determined the LC₅₀ relative to the survival at sulphide absence based on the estimated model intercept=1.689±0.163 (estimate±s.e.) and regression slope=−0.132±0.009 as 15.2 μmol l⁻¹ sulphide, which we rounded to 15 μmol l⁻¹ for use in the AOX knock-down experiment.

We analysed the AOX knock-down experiment data using a generalized linear mixed model with logit link function like for the dose–response data. The developmental success binaries (embryos that developed to tailbud stage relative to all embryos) were modelled with fixed terms for the sulphide treatment (15 μmol l⁻¹ or no sulphide), the MO groups and their interaction, and random terms for family, family×treatment and family×MO. We performed pairwise comparisons of predicted means for each MO group within each treatment and between treatments within each MO group using t-tests and adjusted the P-values for the false discovery rate (Benjamini and Hochberg, 1995).

We analysed the qPCR data using linear mixed models. We first fitted a model to test for a constant log₂C_q across development for each of the two housekeeping genes and a constant difference between them, which held for the latter but not the former. Next, we averaged the C_q across housekeeping genes per family, treatment and developmental stage. We then standardized AOX C_q values by subtracting the average housekeeping C_q values. These efficiency- and housekeeping-gene-standardized AOX C_q values were modelled with fixed terms for the sulphide treatment (treatment: 15 μmol l⁻¹, no sulphide), the five stages sampled [stage: egg (control condition only), 32-cell, gastrula, neurula, tailbud] and their interaction, and random terms for family, family×treatment, family×stage and family×stage×treatment. We tested fixed terms using F-tests. We performed multiple comparisons of treatment contrasts within each stage similarly as reported for the AOX knock-down experiment.

In our study, we used an MO-based knock-down approach to test whether AOX plays a role in successful embryonic development of C. intestinalis under elevated sulphide concentrations. First, we determined the sulphide concentration at which 50% of developmental success to tailbud stage was compromised (sulphide LC₅₀). Then, we tested whether AOX knock-down affected developmental success at those concentrations. Lastly, we gained insight into AOX mRNA responses during development in the presence and absence of LC₅₀ sulphide concentrations.

The dose–response experiment indicated that C. intestinalis larval development success depended strongly on the sulphide concentration (Fig. 1). The average maximum developmental success in the absence of sulphide was 84%, whereas at sulphide concentrations around 50 µmol l⁻¹, absolute developmental success was approaching zero. Based on the logistic mixed model, we determined a sulphide LC₅₀ of ∼15 µmol l⁻¹ compared with embryos not exposed to sulphide. We used this concentration in the following experiments.

In the absence of sulphide, we did not detect any significant difference in developmental success among the four MO groups (Fig. 2, Table 1), indicating that injection procedures had little impact on developmental success. In contrast, when embryos were exposed to sulphide LC₅₀, the developmental success decreased within each of the four MO groups compared with control conditions without sulphide (Fig. 2, Table 2).

The developmental success under sulphide LC₅₀ decreased for the uninjected embryos by 49% [from 67% (95% CI 57–76%) to 35% (95% CI 26–45%)], for the MO_Ctrl-injected by 51% [from 58% (95% CI 47–69%) to 28% (95% CI 20–39%)] and for the rescue group by 52% [from 57% (95% CI 43–70%) to 27% (95% CI 17–40%)]. Importantly, however, the AOX-MO group exhibited a stronger effect of the sulphide LC₅₀ than any of the other MO groups; the developmental success decreased by 81% [from 60% (95% CI 49–71%) to 12% (95% CI 7–18%)] (Fig. 2). These sulphide effects can be translated into odds ratios, i.e. the odds for an embryo to survive in control conditions relative to sulphidic conditions within each MO group. For the uninjected group, the odds ratio was 3.9 (95% CI 2.1–7.1), for the MO_Ctrl group, it was 3.5 (95% CI 1.9–6.7), and for the rescue group, it was 3.5 (95% CI 1.7–7.6). For the AOX knock-down group, the odds to survive under control conditions were 11.6 (95% CI 5.9–23.0) times higher than to survive under sulphidic conditions. Further, to evaluate the relative effect of AOX, we contrasted the odds ratio of the AOX knock-down group with that of each of the two control MO groups. In other words, we quantified the ratio of developmental success of control embryos (uninjected and MO_Ctrl-injected) and AOX knock-down embryos under sulphide conditions to control conditions. We found the odds to survive under LC₅₀ sulphide conditions were higher for control embryos than for AOX knock-down embryos. Specifically, we found the odds to survive under LC₅₀ sulphide conditions for the uninjected control group to be three times higher (95% CI 1.7–5.3) and for the MO_Ctrl group 3.3 times higher (95% CI 1.8–6.2) compared with the AOX knock-down group.

We detected differences in the effects of developmental stage on housekeeping gene transcript levels (gene×stage; Table 3), i.e. the housekeeping gene transcript levels change with development asynchronously, which hinders an interpretation of the results as changes across development. However, the developmental stage effect on transcript levels between housekeeping genes did not differ between the 15 and 0 μmol l⁻¹ sulphide treatments (gene×stage×treatment; Table 3). This absence of stage×treatment interaction for the housekeeping genes (used for the AOX C_q standardization) allows for valid AOX transcript level contrasts between sulphide treatments. Therefore, we performed treatment contrasts, i.e. the difference in AOX mRNA levels (relative to housekeeping gene levels) between sulphide and control treatments for each developmental stage, but we did not perform treatment contrasts across development.

Relative AOX transcript levels between the 15 and 0 µmol l⁻¹ sulphide-exposed embryos were very similar at the 32-cell stage, while their difference increased significantly at the neurula and tailbud stages (Fig. 3). Further, the AOX mRNA level differences between the control and sulphide treatment for the 32-cell, gastrula and neurula stages were similar, while the AOX mRNA level difference was significantly higher for the tailbud stage (Tables 4 and 5).

Because of the changing housekeeping gene mRNA levels across developmental stages, the underlying cause of this increase cannot be fully pinpointed (either an increase in AOX transcript levels or a decrease in the average housekeeping gene transcript levels with development).

In this study, we investigated whether AOX is crucial for the development and survival of the highly invasive marine tunicate C. intestinalis under sulphide stress. To do so, we combined an MO-based knock-down approach and studied the transcriptional response of AOX. First, we demonstrated that early developmental success decreases with increasing sulphide concentrations. Second, we provided evidence that AOX is required for sulphide tolerance. Finally, we showed that sulphide elicits an increased relative amount of AOX transcripts. Together, these results provide support for the idea that the presence of AOX in C. intestinalis increases its sulphide tolerance during development and may thereby contribute to its invasive potential.

The determined LC₅₀ of 15.2 μmol l⁻¹ sulphide and the fully stalled development at a concentration of 50 μmol l⁻¹ (Fig. 1) are compatible with known COX-inhibiting sulphide concentrations. Specifically, while the isolated COX enzyme is half-maximally inhibited at ∼0.2 µmol l⁻¹ H₂S, intact cells display a concentration of ∼30 µmol l⁻¹ sulphide at which COX is inhibited (Leschelle et al., 2005; Petersen, 1977; Yong and Searcy, 2001). Because C. intestinalis embryos closely resemble intact cells, our results are consistent with previously estimated limits, suggesting that AOX is recruited when water sulphide levels are above known COX-inhibiting concentrations. Interestingly, the tolerance of C. intestinalis to sulphide is two orders of magnitude higher than that of eggs, fry and juveniles of non-AOX hosting fish species such as walleye, northern pike, sucker and rainbow trout (Adelman and Smith, 1970; Smith and Oseid, 1972, 1974). While the maximum possible safe level of sulphide for fish eggs lies between 0.41 and 0.53 μmol l⁻¹, it is between 0.11 and 0.18 μmol l⁻¹ for yolk-sac fry. Our results also align with the suggestions that developmental stages of many marine invertebrates are more sensitive to stress compared with their adult stages (Pineda et al., 2012; Ringwood, 1992). To our present knowledge, adult C. intestinalis can withstand much higher sulphide concentrations of up to 300 μmol l⁻¹ than determined here for embryos (Saari et al., 2019). The underlying reason may be the limited repertoire of stress responses in embryos, which exclusively rely on cellular response mechanisms, whereas adults may employ additional behavioural and structural coping mechanisms. This illustrates very well the importance of taking all life stages into account, both when studying the ability of a species to survive in response to environmental stressors and predicting environmental effects on species and ecological communities.

Understanding the underlying mechanisms for coping with environmental stressors, including elevated H₂S, is the basis for predicting ecological communities in the future (Hofmann and Todgham, 2010). In this context, AOX plays an interesting role. AOX is suggested to have fulfilled an important role in the evolution of animals and shaping ecological communities. The appearance of AOX can be dated back to before the great oxygenation event approximately 2.4 billion years ago, a very stable sulphide-rich era (Weaver, 2019). During this time, AOX is suggested to have allowed for metabolic flexibility and hence the rise of mitochondria and the first eukaryotes, and ultimately the rise of the first metazoans after a second (Neoproterozoic) oxygenation event approximately 800 million years ago. After this event, oxygen levels rapidly increased and the environment changed, yielding stable oxygenated conditions (Olson and Straub, 2016; Weaver, 2019). Current sulphide-rich environments including sediments, water columns and areas around vents are found in marine and freshwater environments around the globe. For example, H₂S concentrations can reach up to 15,000 μmol l⁻¹ in the pore waters of sediments off the coast of California, up to 6500 μmol l⁻¹ around the hot vents of the East Pacific rise, or up to 800 μmol l⁻¹ in freshwater springs in Mexico (Bagarinao, 1992; Oeschger and Vetter, 1992; Rosales Lagarde et al., 2014). More temporary, but recurring, sulphide-rich events have been observed off the coast of Peru with up 6 μmol l⁻¹ sulphide (Schunck et al., 2013) and Namibia (up to 30 μmol l⁻¹ sulphide), which lead to regular mass mortalities of fish (Copenhagen, 1953; Lavik et al., 2009). Furthermore, it is expected that areas of toxic H₂S concentrations will increase in abundance, size and severity in many marine environments owing to the increase in anoxic waters (Schobben et al., 2015), which can be expected to affect ecosystems via large animal kills. Interestingly, and as one of the worst-case scenarios, the end-Permian marine mass extinction has been linked to anoxic and sulphide-rich oceanic conditions (Schobben et al., 2015).

To our current and best knowledge, this is the first functional study of AOX in an animal species. This, however, also impedes a direct comparison of our results with those of other studies. Nonetheless, the importance of AOX in the sulphide response was suggested previously. A study on mitochondria of a polychaete, the lugworm (Arenicola marina), found a strong indication that the alternative pathway via AOX is involved in the oxidation of sulphide (Hildebrandt and Grieshaber, 2008). Specifically, under elevated sulphide levels, oxygen consumption was high while no ATP was produced, and this reaction was completely blocked when AOX was inhibited (Hildebrandt and Grieshaber, 2008). These latter results may also partially explain why A. marina can survive quite high sulphide concentrations of up to 10 mmol l⁻¹ (Groenendaal, 1980). A study on another polychaete, the echiuran worm (Urechis unicinctus), found that AOX mRNA levels in the body wall and hindgut tissue increased with both sulphide concentration and sulphide exposure time (Huang et al., 2013). In a previous study on adult C. intestinalis, AOX mRNA levels were increased in heart and neural complex under elevated sulphide concentrations (100 and 300 μmol l⁻¹; Saari et al., 2019). A few additional studies examined the role of AOX under other environmental conditions, such as copper and cadmium exposure, hypoxia, anoxia and extreme temperatures. Interestingly, copper and cadmium, both COX inhibitors, do not elicit a response in AOX mRNA in C. intestinalis (see appendix of Saari et al., 2019). However, under both hypoxic and anoxic conditions, AOX mRNA levels are increased in the Pacific oyster (Crassostrea gigas), the freshwater mussel (Diplodon chilensis) and also in C. intestinalis adults (Saari et al., 2019; Sussarellu et al., 2012; Yusseppone et al., 2018). Notably, in C. intestinalis adults, the combined effects of hypoxia and sulphide exposure appeared to be additive for AOX mRNA levels (Saari et al., 2019). These findings support the idea that AOX may have played an evolutionarily important role under the sulphide-rich and oxygen-poor conditions during the Proterozoic era and allowed for the appearance of metazoans (Weaver, 2019). Only recently has the effect of temperature on the protein level of AOX been investigated in the copepod Tigriopus californicus (Tward et al., 2019). That study detected increased AOX protein levels under cold and heat stress. Interestingly, both the Pacific oyster and the freshwater mussel are also considered invasive species, which supports our suggestion that AOX may indeed be an additional mechanism in slow or sessile marine organisms to cope with a variety of environmental factors and that it may give those species an advantage under future ocean conditions.

Among species lacking the AOX gene, sulphide tolerance varies enormously but can also be quite high (Bagarinao, 1992; Grieshaber and Völkel, 1998). Cellular mechanisms, other than AOX, that explain sulphide tolerance include the sulphide-quinone oxidoreductase (SQR)-related sulphide oxidation as present in most domains of life (Theissen et al., 2003), the H₂S-resistant COX such as found in the shortfin molly (Poecilia mexicana), and the H₂S-binding haemoglobins of the giant tube worm (Riftia pachyptila) and the California killifish (Fundulus parvipinnis) (Bagarinao and Vetter, 1992; Zal et al., 1998). Using H₂S-binding haemoglobins, the California killifish, for example, has an 8 h LC₅₀ of 300 μmol l⁻¹ H₂S (Bagarinao, 1992).

To better understand the control of the AOX response to sulphide, we also investigated the transcriptional response of AOX to sulphide exposure at different developmental stages. Unfortunately, changes of the two housekeeping gene mRNA levels across developmental stages prevented us from being able to make clearer statements on the change of AOX mRNA across developmental stages. This exemplifies the crucial need of a careful analysis of housekeeping gene responses, particularly during dynamic biological processes, such as development, to allow for more reliable conclusions on a target gene response. However, we were able to draw conclusions about the difference of AOX mRNA levels between control and sulphide conditions at each developmental stage. We found no difference in AOX mRNA levels between sulphide and control conditions at the 32-cell stage, whereas the AOX mRNA levels were increasingly higher under LC₅₀ sulphide exposure compared with no sulphide exposure at the later stages (Fig. 3). Further, the sulphide effect was clearly higher for the last stage studied, the tailbud stage, compared with the three earlier stages. These results agree with our findings that AOX indeed plays a crucial role in C. intestinalis survival under sulphide exposure, and is very likely controlled at the transcriptional level.

In summary, we showed that the AOX gene is not required for embryonic development in normal environmental conditions. However, we demonstrated a clear link between sulphide tolerance and AOX, which appears to be regulated, at a minimum, at the level of gene transcription. On a larger scale, these results provide a valuable insight into how fluctuations in sulphide, which are expected to increase under climate change, may affect marine ecosystems. As sulphide exposure increases in frequency and severity, it becomes more and more important to understand whether and how species are affected by sulphide. Many of the AOX-hosting animal species are sessile or slow-moving marine invertebrates with free floating larval stages, making them particularly susceptible to changes in sulphide as they cannot actively avoid the stressor; however, these species comprise many successful invasive species. Our finding about the role of AOX in survival under sulphide stress in C. intestinalis embryos provides the opportunity to extrapolate to other AOX-hosting species, which we expect to have similar competitive advantages under sulphide stress. Lastly, we hypothesize that AOX is an important protein that mitigates not only sulphide but also other environmental stressors affecting aerobic energy production and under whose presence AOX-possessing species may be more competitive as a result.

We thank Clare Hudson and Cathy Sirour for their technical and logistic support at the Laboratoire de Biologie du Développement de Villefranche-sur-mer.