Selenium enrichment in the marine microalga Nannochloropsis oceanica

Se-enriched ingredients have recently gained interest in the aquaculture industry as feed supplements due to their positive effects on fish health, growth, and potential effects on animal welfare. This study aims to assess which inorganic selenium (Se) species is suitable to produce Se-enriched Nannochloropsis oceanica ( N. oceanica ) biomass for aquafeed applications. The effective concentration for 50% growth inhibition (EC 50 ) and Se bio-accumulation of the two inorganic forms of Se, sodium selenite (Na 2 SeO 3 ), and sodium selenate (Na 2 SeO 4 ), were assessed at different concentrations after twelve days of cultivation. Toxicity results showed that selenate, EC 50 = 32.93 μ M, had a greater negative effect on cell growth than selenite, EC 50 = 163.82 μ M. Total intracellular Se was analysed by inductively coupled plasma - optical emission spectrometry (ICP-OES) and high resolution inductively coupled plasma mass spectrometry (HR-ICP-MS), which revealed that selenite was better accumulated by N. oceanica . Further investigation at 30 μ M of selenite in the growth medium resulted in Se bio-accumulation with a minor effect on cell growth and reached a Se intracellular content of 0.131 g Se /kg biomass after 12 days. Thus, 30 μ M of selenite was selected for batch pilot-scale cultivation in a 1500 L tubular photo-bioreactor. Total Se accumulated in the biomass at pilot-scale was in the same order of magnitude compared with flasks (0.104 – 0.159 g Se /kg biomass ). The results from this pilot-scale study are fundamental for a proof of concept from laboratory to pilot-scale production and they represent a critical bridging step for the potential use of Se-enriched N. oceanica in aquafeed.


Introduction
Selenium (Se) deficiency in fish can interfere with growth, cause greater fish mortality and hinder the fish immune response [1]. Se is a trace element naturally present in aquatic environments (both fresh and seawater) in two inorganic forms: selenite (SeIV) and selenate (SeVI) in the range of nanomolars (nM) to micromolars (μM) [2]. However, it is challenging to naturally meet Se dietary requirements in aquaculture, and efforts to mitigate this issue include adding mineral additives to aquafeed such as mineral mixes containing a higher supply of inorganic Se [3]. However, inorganic Se is not an effective aquafeed supplement since it is not easily absorbed by fish [3,4]. Thus, studies have focused on understanding the bioavailability of different Se sources used in aquafeed, including inorganic Se [5], organic Se (in the form of selenomethionine) [5], Se-enriched yeast [3], and Se-enriched microalgae [6,7]. Microalgae are considered an important aquafeed ingredient since they are the primary producers of (polyunsaturated fatty acids) PUFAs in the ocean and contain, carbohydrates, proteins, vitamins, essential amino acids, minerals and pigments which are essential to meet fish nutritional demands [8][9][10]. Studies have shown that microalgae play a crucial nutritional role in early development stages of finfish and for molluscs (such as oysters) during all stages of development [11,12].
The effect of Se on microalgae growth varies greatly and can be either beneficial or toxic [13]. Several studies on Se exposure in microalgae have shown that Se toxicological effects are mostly due to four main factors: the inorganic Se chemical species (e.g. selenite, selenate) [14,15]; the concentrations of Se in the media [16][17][18]; a reflection of a microalgae species-specific response [19], and the concentration of sulphur (S) present in the media [7]; since S and Se share chemical similarity, they are incorporated inside the cells through the same transporters, leading to an antagonistic uptake of S and Se [20][21][22][23].
Although several toxicological studies mention the effect of Se on growth, few studies have looked into the effect of the two inorganic species of Se on cell growth and accumulation under the same experimental conditions. In addition, no research has been conducted on pilotscale production of Se-enriched marine microalgae of commercial interest such as Nannochloropsis oceanica (N. oceanica). N. oceanica is a key species in the aquafeed sector due to its nutritional value in pigments, fatty acids (up to 30.7%) [24], and eicosapentaenoic acid (EPA) content (2.3 to 5.6% DW) [24][25][26][27][28]. The aim of this paper is to determine for the first time which inorganic Se source is less toxic and more accumulated in N. oceanica, thus being suitable to produce Se-enriched biomass at pilot-scale.

Cultivation medium, pre-cultivation
Nannochloropsis oceanica (N. oceanica) (CCAP 849/10) was cultivated in chloride medium (where all elements containing sulphate forms (SO 4 2− ) were substituted by chloride (Cl − ) forms) as previously described [29]. The medium was filter sterilised (0. 22  μM) to assess the optimal concentration for pilot-scale production. Experiments were performed over 12 days in biological triplicates. A control with no Se (0 μM) was included in each screening series. An overview of the concentrations used can be found in Table 1. All experimental cultures had an initial cell concentration of ~2.4 × 10 7 cells/mL (OD 750 of 0.5).

Pilot-scale production of selenium (Se) Se-enriched N. oceanica
Se accumulation was studied during batch cultivation in a 1500 L LGem tubular photobioreactor (PBR) (GemTube MK-1 1500s, LGem, The Netherlands), with a liquid volume of 1300 L, at AlgaePARC greenhouse facility (Wageningen University & Research) (Fig. 1). The pH was controlled at 7.5 by sparging CO 2 on demand, and the culture was mixed by using a combination of a liquid and air pump (approx. 55 L/min). Temperature in the culture ranged between 18 and 25 • C. N. oceanica biomass was sampled daily (up to 5 L a day) and the total sampling volume was maintained below 10% of the initial experimental volume (<130 L). N. oceanica was cultivated in natural seawater enriched with the same nitrate, phosphate, trace elements and Se stocks used for shake flask experiments. The pre-cultures were incrementally scaled up from Erlenmeyer flasks, to a 20 L flat panel PBR, to a 300 L tubular PBR, and finally transferred to the 1500 L tubular PBR. Microalgal cultures were inoculated on average with a starting OD 750 between 0.4 and 0.5, to avoid photoinhibition. The PBRs were sterilised by adding 100 ppm of sodium hypochlorite for at least 24 h, followed by flushing the reactor twice with filtered tap water (Millipore Opticap® XL10 Durapore® 0.22 μM, Merck). At the end of each batch cultivation, approximately 1000 L was harvested from the PBR and the remaining 300 L in the PBR was used as inoculum for the next batch cultivation. The five consecutive batch reactor runs (16-26 days) described in this experiment occurred during Autumn and Winter (September 2019 -March 2020) (Fig. 1). The remaining microalgal biomass was firstly dewatered by a spiral plate technology centrifuge (EVODOS 25 SPT) and then lyophilised to 30% solids by freeze-drying (EKS 30-3, Zirbus). Freeze-drying was preferred for further processing to maintain the high nutritional composition of the microalgae and to avoid any heat effect on sensitive cellular components, such as fatty acids [24].

Biomass samples for cell growth determination
During all experiments, microalgae growth was monitored daily. Cell growth was determined by measuring N. oceanica optical density (OD) in triplicate with a UV-VIS-spectrophotometer (Hach Lange, DR 6000) (750 nM) and cell concentration was measured in triplicate with a cell counter (Beckman Multisizer™ 3 Coulter Counter) using a 50 μM aperture tube. Particles between 2 μM and 7 μM were considered to be N. oceanica cells. Dry weight was determined using the method previously described by Guimarães et al., [29]. Volumetric productivity (g/L/ day) of each batch in the pilot-scale studies was calculated by subtracting the final dry weight from the initial dry weight and dividing it by the duration of the batch (days). Areal productivity (g/m 2 /day) of each batch was calculated by multiplying the volumetric productivity by the culture volume in the PBR (1300 L) and dividing it by the aerial footprint of the LGem PBR (19 m 2 ).

Biomass samples for elemental analysis
N. oceanica flask cultures were harvested for elemental analysis on day 12 of the experiment. Cultures were centrifuged (2000 ×g, 15 min at 20 • C) and washed twice with ammonium formate (0.5 M) [29] to remove the excess of salts. After washing, the biomass pellets were lyophilised (Sublimator 2 × 3 × 3-5, Zirbus Technology, Germany). Samples for elemental analysis from the LGem reactor were taken at the end of each batch. For the last batch, samples were taken daily, and these were washed and centrifuged in the same manner.

Microwave-assisted acid digestion
Lyophilised biomass samples were microwave-assisted acid digested as previously described [29]. Each lyophilised microalgal sample (50 mg) was extracted with 10 mL dH 2 O, 7.5 mL of hydrochloric acid (37%) and 2.5 mL of nitric acid (65%) in a microwave oven (milestone S.r.l. ETHOS 1). The total microwave run time was 40 min with a maximum temperature of 175 • C (maintained from 15 to 30 min). The maximum energy was 1400 W.

Sodium selenate-treated cultures.
Samples treated with sodium selenate were measured by high resolution -inductively coupled plasma -mass spectrometry (HR-ICP-MS) (Element2, Thermo Fisher Scientific). HR-ICP-MS was used since Se values were below detection limits of the ICP-OES. The nebulizer used for the ICP-MS was a concentric nebulizer made from PFA (Perfluoroalkoxy alkanes) and the spray chamber was cyclonic made and from quartz. The cones used were made of Nickel.

Statistical analysis
The data generated during this study was subjected to statistical analysis using one-way analysis of variance (ANOVA) to test the effect of Se on cell growth. When significant differences were found, post-hoc Tukey tests were applied. The statistical analysis was performed using SPSS version 25 with a significance level of p < 0.05. In order to perform the dose-response analysis, final obtained cell concentration values from the sodium selenate and selenite toxicity screening were normalised and plotted using Excel. The effective concentration for 50% growth inhibition (EC 50 ) determination was based on Geoffroy et al. [30], using the final cell concentration values adjusted to a four parameter logistic function using a dose-response model and defining the upper and lower limit using R [47] package drc [45]. Briefly, the lower limit was set between 0 and 2.4 × 10 7 cells/mL (the initial cell concentration) and the upper limit was set for the highest cell concentration obtained (3.3 × 10 8 cells/mL).

Toxicity of selenium species in N. oceanica: selenite and selenate
Selenium (Se) can be either beneficial or toxic to microalgal cultures [13,31], however, the effect of Se in N. oceanica has not yet been described. In this study, various concentration of Se (0-1000 μM) ( Table 1) were tested on N. oceanica. Cell growth was monitored to determine the effective concentration for 50% growth inhibition (EC 50 ) Growth of N. oceanica when exposed to a range of Se concentrations (0, 1, 5, 10, 25, 50, 100, 500, 1000 μM) over 12 days. Two inorganic Se species were studied: a) sodium selenite (Na 2 SeO 3 ) and b) sodium selenate (Na 2 SeO 4 ). Samples were taken daily and each data point represents the average of three biological replicates (n=3).
of Se (Fig. 2). Experiments lasted 12 days to observe the effect over several cell divisions of a culture during exponential and linear growth. Toxic effects that resulted in no cell growth were observed at the highest Se concentrations tested (500 and 1000 μM), for both selenite and selenate. Moreover, for both selenite-and selenate-treated cultures ( Fig. 2, A-B), concentrations ranging from 50 to 100 μM resulted in a significant reduction in cell concentrations (p < 0.001). Differences in toxicity between the two Se species were observed at lower Se concentrations (1-50 μM). Selenate-treated cultures showed the highest toxicity, where the final cell concentration was significantly different from the control (0 μM) at the concentrations of 5 μM (p = 0.034), 10 μM (p = 0.001), and 25 μM (p < 0.001). Consequently, for selenate, non-toxic concentrations were only obtained at the lowest concentration tested, 1 μM (Fig. 2, B). For selenite-treated cultures, nontoxic concentrations were obtained within the concentration range 1-25 μM (Fig. 2, A). Overall, the toxicity screening successfully provided an assessment of toxic and non-toxic effects for both Se ionic species and it showed that selenate is more toxic to N. oceanica than selenite. Similar to our findings, other Se comparative studies showed higher toxicity of selenate compared to selenite in the freshwater microalgae Selenastrum capricornutum (S. capricornutum) [14] and Chlamydomonas reinhardtii (C. reinhardtii) [15].
Selenite concentrations in the range of 25-50 μM seemed promising since no effect on cell growth was observed at 25 μM and only a small effect on cell growth was observed at 50 μM (Fig. 2, A). Therefore, we  Fig. S1). Although some studies report a beneficial effect on microalgal growth when exposed to low concentrations of selenite, for example Arthrospira platensis (0.5-40 mg/L) [16], Chlorella vulgaris (C. vulgaris) (≤75 mg/L) [17], and Chlorella pyrenoidosa (C. pyrenoidosa) (≤40 mg/L) [32], we did not observe any positive effects on growth. This could be due to a microalgal species-specific response to Se [19].
Overall, this broad range of reported levels of Se toxicity (Table 2) emphasises the necessity for species-specific studies. Moreover, other elements in the medium composition may interfere with Se uptake, which also needs to be investigated in more depth. For example, the competitive uptake of sulphur (S) may have an important role in the toxicity and bioaccumulation of Se [15,16].
Overall, the toxicological differences observed in our study confirm that the toxic effect of Se depends on the concentration of Se in the medium, the inorganic Se chemical species, and the N. oceanica speciesspecific response, which has also been observed in other studies [15,16,19,37]. Moreover, it was observed that increasing intracellular Se negatively affected growth [7,38]. Our study emphasises the need for an assessment of Se toxicity and bioaccumulation in the same study.

Selenium bioaccumulation
Se was measured in both selenite-and selenate-treated N. oceanica cultures, after 12 days (Table 3). However, Se analysis was not possible for 100 μM (selenate only), 500 μM, and 1000 μM Se-treated cultures since no growth was observed at the highest Se concentrations tested for both selenite and selenate (cells were bleached) (Fig. 2). Control groups (without Se) had a maximum amount of total Se <0.001 g Se /kg biomass for selenate and <0.01 g Se /kg biomass for selenite cultures.
N. oceanica selenate-treated cultures had a maximum Se accumulation of 0.085 g Se /kg biomass (at 50 μM) (Table 3). Moreover, the increase of selenate concentrations in the media resulted in more Se accumulation in the biomass. Selenate-treated cultures in the range from 5 to 50 μM had significantly (p < 0.05) more Se than control cultures (no added Se), indicating Se accumulation is proportionate to the Se concentration in the media [39].
N. oceanica selenite-treated (100 μM) cultures had a maximum Se accumulation of 2.115 g Se /kg biomass (Table 3). This value is 4-fold higher than what was previously observed in selenite-treated C. pyrenoidosa Overall, Se accumulation was 4 to 8-fold higher when treating N. oceanica cultures with selenite rather than selenate. This was to some extent in line with the published literature where selenite was preferentially taken up by C. vulgaris [40] and selenite was found to be accumulated up to ten times more efficiently than selenate by C. reinhardtii, due to different transport mechanisms [15]. In our study, selenite was the least toxic and the most accumulated Se inorganic form in N. oceanica, which contradicts previous observations where the form of Se that is most accumulated is also the most toxic to microalgae [13]. This emphasises the need for more research in Se uptake and accumulation and its species-specific effect. Additionally to Se, for selenite treated cultures, other major and minor elements (S, P, Ca, Cu, Fe, Mg, Mn, Na, K, Zn) were measured at the end of each cultivation (Supplementary data - Table S1). The elements were present in a decreasing concentration of P [41], where Se is between Na and Zn in all conditions, except the control (Se 0 μM) and the lowest selenite concentrations (1,5 where Se is least accumulated. This suggests that the presence of Se does  that the decrease in growth caused by the toxic effect of Se is also reflected in a decrease in Mg, which is a component of chlorophyll. To produce Se-enriched biomass, a compromise needs to be reached between Se accumulation and the effect on growth during cultivation. Therefore, we consider selenite to be the most suitable Se source for the production of Se-enriched N. oceanica. A selenite concentration of 30 μM, which is just above the NOEC (25 μM), was selected for the pilotscale experiments.

Pilot-scale production of selenium (Se) Se-enriched N. oceanica
To test the feasibility of pilot-scale production, Se-enriched N. oceanica was cultivated in a 1500 L photobioreactor for 101 days in five consecutive cultivation batches (Table 4). Table 4 gives an overview of growth and the final Se accumulation obtained. Overall, each cultivation batch lasted between 16 and 26 days. The average volumetric productivity of all batches was approximately 0.21 g/L/day, which is in the same order of magnitude as the value achieved in an outdoor study with N. oceanica in tubular PBRs (0.33 g/L/day) during Spring/Summer in Portugal [42]. The average areal productivity reached in this study was 14.07 g/m 2 /day (Autumn/Winter, The Netherlands). Similar results were found in a study with Nannochloropsis sp. in vertical outdoor photobioreactors (10.6-24.4 g/m 2 /day, during Summer in The Netherlands) [43]. Similarly, a study with N. oceanica obtained an average areal productivity of 10.7 g/m 2 /day during Spring/Summer in Portugal [42]. Overall, this emphasises that the cultivation of Seenriched N. oceanica did not affect its volumetric or areal productivity.
The total amount of Se accumulated at the end of each batch was determined. However, Se determination was not possible for batch 1 since the biomass was not pre-washed with ammonium formate and only centrifuged. Overall, at the end of the batch (2-4) Se accumulation did not vary significantly. However, batch 5 had significantly more Se than all other batches (p = 0.001). This could be due to biological variability between the different batches, which occurred between Autumn and Winter.
Selenium accumulation in the biomass was measured daily in batch 5 (Fig. 4, B). In this batch N. oceanica was cultivated for 24 days (Fig. 4, A) to a final biomass concentration of 5 g/L. There was no sampling on day zero (inoculation day) and day 1 since the tubular PBR was still being mixed to a uniform culture. Se inclusion increased during cultivation, reaching 0.110 ± 0.015 g Se /kg biomass on day 12 and 0.159 ± 0.015 g Se / kg biomass on day 24 (Fig. 4, B). Overall, Se accumulation at pilot-scale was similar to the Se accumulation observed in flasks after 12 days of cultivation (0.131 ± 0.017 g Se /kg biomass ), suggesting a stable and growth-dependent Se accumulation (Fig. 4, dotted line). Our results are also in the same order of magnitude as the values achieved with the freshwater microalgae Chlorella (0.251 g Se /kg biomass ) [44].

Conclusion
This study is the first to assess toxicological effects of Se and the production of Se-enriched biomass using N. oceanica. The dose-response effect was evaluated for sodium selenate, EC 50 = 32.93 μM (2.60 mg Se / L) and sodium selenite, EC 50 = 163.82 μM (12.94 mg Se /L), revealing that selenite is less harmful for the growth of N. oceanica. Selenite is more efficiently taken up than selenate, thereby showcasing its potential to be used as an inorganic source for the production of Se-enriched N. oceanica biomass. During pilot-scale experiments, the presence of selenite had a minimal effect on the growth of N. oceanica and resulted in selenium accumulation between 0.104 and 0.159 g Se /kg biomass . These results are in line with what was observed during laboratory scale trials (0.131 ± 0.017 g Se /kg biomass ). Overall, we demonstrated the capability of N. oceanica to accumulate Se, validated this at pilot-scale and provided the proof of concept for using this species as a Se-enriched biomass for aquafeed.

Statement of informed consent, human/animal rights
No conflicts, informed consent, or human or animal rights are applicable to this study.