Kleptoplast photoacclimation state modulates the photobehaviour of the solar-powered sea slug Elysia viridis

The incorporation and maintenance of intracellular functional chloroplasts by host cells (kleptoplasty) occurs in a single taxon of metazoans – sacoglossan sea slugs (Rumpho et al., 2011; Serôdio et al., 2014). ‘Stolen’ plastids (kleptoplasts) may remain functional in cells of the digestive gland of these sea slugs for periods ranging from a few days to several months (Händeler et al., 2009). In accordance, different levels of kleptoplasty have been proposed, including long-term, medium-term, short-term and non-functional retention (Clark et al., 1990; Händeler et al., 2009). In sea slugs with higher levels of functional retention, often termed ‘solar-powered sea slugs’ (Rumpho et al., 2000), kleptoplast photosynthesis has been shown to be nutritionally relevant (Pierce et al., 2015; Cartaxana et al., 2017).

Positive phototaxis has been shown for several sea slugs displaying kleptoplasty and it has been suggested that this behaviour promotes light harvesting by stolen plastids (Gallop et al., 1980; Weaver and Clark, 1981; Schmitt and Wägele, 2011; Miyamoto et al., 2015). However, these animals seem to avoid high light intensities, possibly as a strategy to prevent the premature loss of kleptoplast photosynthetic function (Weaver and Clark, 1981; Cruz et al., 2013). A distinguishable behaviour in response to light has been recorded in Elysia timida: a change in the position of its lateral folds (parapodia) from a closed position to a spread, opened leaf-like posture under lower irradiance and the opposite behaviour under high light levels (Rahat and Monselise, 1979; Jesus et al., 2010; Schmitt and Wägele, 2011). This behaviour in response to light changes was only recorded for this species and has been assumed to be linked to long-term retention of kleptoplasts (Schmitt and Wägele, 2011).

In this study, we determine whether the photobehaviour of the solar-powered sea slug Elysia viridis is linked to the photophysiological state of kleptoplasts (i.e. photoacclimation). In order to photoacclimate the kleptoplasts, E. viridis and its algal food source, Codium tomentosum, were maintained under low and high light levels. We describe a correlation between the position of E. viridis parapodia and irradiance and present evidence that light preference and parapodia closure are dependent on the photoacclimation state of the kleptoplasts.

The sea slugs Elysia viridis (Montagu 1804) and Placida dendritica (Alder and Hancock 1843), as well as the macroalga Codium tomentosum Stackhouse 1797, were collected during low tide in the intertidal rocky area of Aguda beach, Vila Nova de Gaia, Portugal (41°02′50.2″N, 8°39′15.2″W). Both sea slugs were sampled from dense stands of C. tomentosum. Animals were kept in aerated water collected at the sampling site and transported to the laboratory within 2 h. Sea slugs and their macroalgal food were maintained in recirculated life-support systems operated with artificial seawater (ASW) at 18°C and a salinity of 32. Photoperiod was 14 h light:10 h dark at an irradiance of 40 µmol photons m^–2 s^–1 (low light acclimated, LL_ac) provided by T5 fluorescent lamps. Elysia viridis and C. tomentosum were also maintained in the same conditions as described above but under an irradiance of 200 µmol photons m^–2 s^–1 (high light acclimated, HL_ac). Chosen irradiance levels are within the range measured in the shaded intertidal rock pools where the macroalgae and animals were collected. The duration of the light-acclimation period was 2 weeks. Animals used for pigment analysis showed no significant differences in dry mass (t-test, P=0.129) between LL_ac and HL_ac treatments (2.64±0.63 and 3.20±0.38 mg, respectively; means±s.d.).

Sea slugs were placed individually into 150 mm long×10 mm wide glass tubes filled with ASW. The tube was composed of sequential dark and transparent areas as follows: 15 mm dark–15 mm transparent–60 mm dark–60 mm transparent (Fig. S1). In the first light preference experiment, the tube was illuminated from above, resulting in a homogeneous light field of 40 µmol photons m^–2 s^–1 (low light 1, LL₁) in the transparent areas and a gradient of 0–5 µmol photons m^–2 s^–1 (dark) in the dark areas. In the second light preference experiment, the tube was illuminated, resulting in a homogeneous light field of 550 µmol photons m^–2 s^–1 (high light, HL) in the transparent areas and a gradient of 5–80 µmol photons m^–2 s^–1 (low light 2, LL₂) in the dark areas. Irradiance was measured with a submersible spherical micro quantum sensor (US-SQS/L, Heinz Walz GmbH, Pfullingen, Germany). The animals were then monitored for 20 min and the time spent in each part of the tube was registered. Twenty-four specimens of E. viridis were monitored (12 LL_ac and 12 HL_ac). Six P. dendritica were also monitored as a negative control, as this sea slug feeds on the same macroalga but presents non-functional retention of kleptoplasts.

Sea slugs were placed in Petri dishes with ASW for 30 min to 1 h at an irradiance of 10 µmol photons m^–2 s^–1 under a digital microscope (DMS-300, Leica Microsystems, Wetzlar, Germany). When the animals were static and fully relaxed, they were subjected to increasing irradiance at steps of 30 s duration: 10, 30, 120, 350 and 650 µmol photons m^–2 s^–1. Light was delivered from an LED light source (KL-200 LED, Schott, Mainz, Germany). At the end of each light step, a photograph was taken with the digital microscope camera. The exposed dorsal area (EDA), excluding the pericardium, was then calculated from the photographs with open source software ImageJ (version 1.49). For each slug, the percentage EDA was calculated as EDA_E/EDA_max ×100, where EDA_E was the EDA at irradiance E and EDA_max was the maximum EDA. Twenty-eight specimens of E. viridis were monitored (14 LL_ac and 14 HL_ac). As P. dendritica have cylindrical cerata extending from the dorsal surface and lack parapodia, they were not considered in this experiment.

Fluorescence measurements were carried out on LL_ac and HL_ac macroalgal samples using a Junior PAM fluorometer (Heinz Walz GmbH). The distance between the fluorometer fibre optic and the surface of the sample was kept constant at 1 mm during all measurements. Steady-state light curves (SSLCs) were constructed with 11 incremental steps of actinic irradiance (E; 0, 25, 45, 65, 90, 125, 190, 285, 420, 625 and 820 µmol photons m^–2 s^–1). For each step, the effective quantum yield of photosystem II (ΔF/F_m′) was monitored every 2 min and when a steady-state was reached the relative electron transport rate (rETR) was calculated as rETR=E×ΔF/F_m′. The light response was characterized by fitting the model of Eilers and Peeters (1988) to rETR versus E curves and by estimating the parameters α (initial slope of the light curve), ETR_max (maximum rETR) and E_k (light saturation coefficient). The model was fitted iteratively using MS Excel Solver. Curve fit was very good in all cases (r>0.98).

Pigment analysis of LL_ac and HL_ac sea slugs and macroalgae was done as described in detail by Cruz et al. (2014). Briefly, sea slugs and macroalgal samples were frozen in liquid nitrogen, freeze-dried and pigments extracted in 95% cold buffered methanol (2% ammonium acetate). After filtration, the extracts were injected into an HPLC system (Shimadzu, Kyoto, Japan) with a photodiode array detector (SPD-M20A). Pigments were identified from absorbance spectra and retention times and concentrations calculated in comparison with pure crystalline standards from DHI (Hørsolm, Denmark). Pigment concentrations were expressed per dry mass.

Significant differences of photoacclimation state (LL_ac and HL_ac) for measured parameters were tested using independent samples t-tests. Light preference was tested using one-sample t-tests for a test value of 10 min. Assumptions were verified using Levene's (homogeneity of variances) and Shapiro–Wilk (normal distribution) tests. A three-parameter logistic model was applied to EDA versus E data for LL_ac and HL_ac states. Significant differences in E₅₀ (irradiance causing 50% closure of parapodia) between different photoacclimation states were evaluated using a generalized likelihood ratio test. All statistical analyses were carried out using IBM SPSS Statistics 24.

The two irradiance levels under which E. viridis and C. tomentosum were maintained induced a clear difference in the photoacclimation state of the plastids, as demonstrated by differences between LL and HL treatments in both photophysiological parameters (SSLC parameters ETR_max and E_k) and pigment concentration (Fig. 1). Photosynthetic capacity at saturating irradiance (ETR_max) and light saturation coefficient (E_k) were significantly higher for HL_ac (42.8±5.0 and 86.3±9.7 µmol m^–2 s^–1, respectively; mean±s.d.) than for LL_ac C. tomentosum (23.0±9.8 and 45.8±9.4 µmol m^–2 s^–1, respectively) (t-tests, P=0.036 and 0.006, respectively). These differences are characteristic of HL- versus LL-acclimated algae (Cruz and Serôdio, 2008). The parameter related to photosynthetic efficiency at limiting irradiance, α (0.49±0.11 and 0.50±0.08 µmol^–1 m² s for LL_ac and HL_ac, respectively; mean±s.d.), did not show significant differences between light treatments (t-test, P=0.907) (Fig. 1A).

Concentrations of pigments from the light harvesting complexes (LHCs) of siphonous green algae chloroplasts, chlorophylls a and b (Chl a and Chl b) and siphonoxanthin (Siph) were significantly higher in LL_ac C. tomentosum and E. viridis (t-tests, in all cases P<0.05) (Fig. 1B,C). This is a consequence of adaptation to the available light conditions, with specimens reared under low irradiance requiring higher concentrations of pigments for light harvesting than individuals reared under high light. An opposite trend was observed for pigments all-trans-neoxanthin (t-Neo) and violaxanthin (Viola) with significantly higher concentrations found in HL_ac C. tomentosum and E. viridis (t-tests, in all cases P<0.05). This is yet more evidence of photoacclimation, as these pigments are involved in photoprotection. Indeed, Uragami et al. (2014) observed the accumulation of t-Neo and Viola in Codium intricatum under high irradiance conditions and hypothesized that these pigments promote oligomerization of the LHCs in order to quench the excess amount of excitation energy. Pigment differences due to light acclimation were similar in sea slugs and the algal food source, indicating a fast turnover of kleptoplasts.

When allowed to choose between LL₁ (40 µmol photons m^–2 s^–1) and dark (0–5 µmol photons m^–2 s^–1) conditions, the long-term kleptoplast retention species E. viridis showed a significant preference for LL₁ (t-tests, P<0.001 in both LL_ac and HL_ac) (Fig. 2A). Choice of low light versus dark was independent of the photoacclimation state (t-test, P=0.634). In contrast, P. dendritica, a non-functional kleptoplast retention sea slug, showed an extremely variable behaviour without any preference for dark, low or high light conditions (t-tests, P=0.586 and P=0.996 for dark versus LL₁ and HL versus LL₂, respectively) (Fig. 2A,B). Positive phototaxis has been shown for several sea slugs species hosting functional kleptoplasts (Gallop et al., 1980; Weaver and Clark, 1981; Schmitt and Wägele, 2011; Miyamoto et al., 2015). Costasiella lilianae and Elysia crispata preferred light intensities of up to 600 μmol photons m^–2 s^–1 over dark, while Berthelinia caribbea and Oxynoe antillarum, which do not retain algal chloroplasts, showed a preference for dark conditions (Weaver and Clark, 1981). Likewise, Miyamoto et al. (2015) reported positive phototaxis for photosynthetic sea slugs and neutral or negative phototaxis for species without functional kleptoplasts. However, light-driven behaviour in sacoglossan sea slugs may not be exclusively related to the presence of functional kleptoplasts. Juveniles of E. timida display a clear preference for light prior to the acquisition of kleptoplasts (Schmitt and Wägele, 2011). Indeed, phototaxis is a widespread animal behaviour and factors such as foraging might play an important role in the response of sacoglossan sea slugs to light.

When allowed to choose between LL₂ (5–80 µmol photons m^–2 s^–1) and HL (550 µmol photons m^–2 s^–1), E. viridis showed a significant preference for low light in LL_ac specimens (t-test, P<0.001), but no specific preference in HL_ac individuals (t-test, P=0.441). Light preference was dependent on the photoacclimation state: HL_ac specimens spent significantly more time in high light than LL_ac conspecifics (t-test, P=0.002) (Fig. 2B).

Closure of parapodia, with a consequent decrease in EDA, was observed under increasing irradiance for both LL_ac and HL_ac E. viridis (Fig. 3; Fig. S2). However, the pattern of parapodia closure was dependent on the photoacclimation state. E₅₀ was significantly higher (χ²_df=1=62.27, P<0.001) for HL_ac E. viridis than for LL_ac conspecifics (486±28.4 and 138±15.4 μmol photons m^–2 s^–1, respectively; mean±s.e.m.). For LL_ac E. viridis, the relationship between EDA and E was exponential, with substantial closure of parapodia occurring even under low irradiance (from 30 to 120 µmol photons m^–2 s^–1). HL_ac E. viridis closed their parapodia under higher irradiance levels (350 and 650 µmol photons m^–2 s^–1) than LL_ac specimens (Fig. 3; Fig. S2). These results are congruent with light acclimation at 40 µmol photons m^–2 s^–1 (LL_ac) and 200 µmol photons m^–2 s^–1 (HL_ac) and indicate that photobehaviour of E. viridis was modulated by the photoacclimation state of the kleptoplasts. In a light preference experiment, the sea slugs Elysia hamatanii, Elysia trisinuata and Plakobranchus ocellatus were reported to choose light intensities that maximized photosynthesis (Miyamoto et al., 2015). Similarly, a clear avoidance of extreme low or high light levels and a preferential accumulation at irradiance levels coincident with the optimum for photosynthetic activity were reported for the symbiotic acoel flatworm Symsagittifera roscoffensis (Serôdio et al., 2011).

Avoidance of high irradiance through parapodia closure is a possible strategy to prevent photoinhibition and a premature loss of kleptoplast photosynthetic function (Weaver and Clark, 1981; Cruz et al., 2013). Rahat and Monselise (1979) observed that E. timida opened their parapodia wide under light intensities between 3×10³ and 3×10⁵ erg cm^–2 s^–1 (approximately 14–1400 μmol photons m^–2 s^–1) and closed them at lower or higher irradiance levels. We found that E. viridis opened their parapodia at an irradiance of 10 μmol photons m^–2 s^–1 and closed them at a lower irradiance than that reported for E. timida. These differences may be related to a distinct kleptoplast distribution in the two species: exclusively in the inner part of the parapodia in E. timida and in both inner and outer sides in E. viridis. This probably makes the closure of parapodia more efficient in protecting kleptoplasts from high light exposure in E. timida. Alternatively, differences may arise from higher sensitivity to light of Codium-derived chloroplasts hosted by E. viridis as a result of the absence of a photoprotective xanthophyll cycle (Cruz et al., 2015), which is present in Acetabularia- and Vaucheria-derived chloroplasts of E. timida and Elysia chlorotica, respectively (Jesus et al., 2010; Cruz et al., 2015). Changes in the position of the parapodia in response to light variations were not observed in Thuridilla hopei, a short-term (ca. 8 days) kleptoplast retention species (Schmitt and Wägele, 2011).

Plastid acquisition during evolution of eukaryotes imposed an increase in the production of reactive oxygen species (ROS) and the development of the redox-signalling network (Woehle et al., 2017). Hence, it is reasonable to hypothesize that kleptoplasty may be challenging to metazoan cells as a result of ROS produced from chloroplast oxygen-based metabolic processes, particularly under high light conditions (Nishiyama et al., 2006). ROS generated as by-products of photosynthesis, especially singlet oxygen and H₂O₂, are signals involved in responses to excessive light (Li et al., 2009), but whether these molecules trigger light avoidance/closure of parapodia in solar-powered sea slugs remains to be investigated.

In conclusion, we describe a correlation between the position of E. viridis parapodia and irradiance. Our results indicate that the photobehaviour (light preference and parapodia closure) of this sea slug species is modulated by the photoacclimation state of the kleptoplasts.

Thanks are due to the anonymous reviewers for their helpful suggestions and comments on previous versions of the manuscript.