Euglena gracilis cells were grown in a medium containing 0.1 g L⁻¹ from CH₃CO₂K, 0.1 g L⁻¹ beef extract, 0.2 g L⁻¹ bacto-tryptone, 0.2 g L⁻¹ yeast extract, 0.02 g L⁻¹ KNO₃, 0.002 g L⁻¹ (NH₄)₂HPO₄, 0.001 g L⁻¹ MgSO₄ × 7H₂O, and 0.0052 g L⁻¹ CaSO₄, supplemented with Hutner's trace elements (Hutner et al., 1966). The medium was supplemented with 50, 100, and 150 mM NaCl during salt stress treatment. The strains were grown on a rotary shaker with gentle shaking at 110 rpm at 24°C under 16 h of illumination (40 μmol photons m⁻² s⁻¹) and 8 h of darkness. The subcultures were started by setting the OD at 750 nm ~0.2. The experiments were performed for 4 days with daily sampling. The culture usually reached the stationary phase within 4–5 days.

The pigments were extracted from the cell suspensions in 90% methanol as described previously by Devars et al. (1992) and the chlorophyll (Chl) contents were determined spectrophotometrically using the molar absorption coefficients described by Lichtenthaler and Wellburn (1983).

For SEM, the cells were fixed in 2.5% glutaraldehyde for 3 h in the suspension, then filtered and washed on poly-L-lysine-coated polycarbonate filters. After post-fixation in 1% OsO₄, the samples were dehydrated in increasing ethanol concentrations, critical-point dried, covered with 15 nm gold by a Quorum Q150T ES sputter, and visualized in a JEOL JSM-7100F/LV scanning electron microscope. For TEM, cells were fixed in 2.5% (v/v) glutaraldehyde for 2 days and suspended in 2.5% tepid agar. After solidification, small cubes were cut from the samples and post-fixed in 1% (w/v) OsO₄ for 1 h. The fixatives were buffered with 0.07 M Na₂HPO₄-KH₂PO₄ (pH 7.2). Following dehydration in aqueous solutions of increasing ethanol concentrations, the samples were embedded in Durcupan ACM resin (Fluka, Switzerland). Ultra-thin sections were cut by a Reichert Jung Ultracut M microtome (Reichert-Jung Ltd., Austria), mounted on the copper grids, then contrasted with 5% uranyl acetate and Reynold's lead citrate solution and observed by a Hitachi 7100 (Hitachi Ltd., Japan) and JEOL JEM 1011 (Jeol Ltd., Japan) TEM. To determine the repeat distance of thylakoid membranes from TEM images, we used ImageJ software. On the micrographs made with ×80 magnification, the sharp regions were selected where the thylakoid membranes appeared to be sliced perpendicularly to their membrane planes (Ünnep et al., 2014). We applied fast Fourier transformation (FFT) analysis on the selected region, which provided values on the periodicity (repeat distances—RD) of the thylakoids. This RD incorporates the widths of the two aqueous phases (the lumen and the interthylakoidal aqueous phase), and two times the width of the membrane. Each average RD was calculated from more than 160 data of stacked thylakoids, which were measured on the images of chloroplasts from 18–19 cells.

The experiments were performed on the SANS II instrument of the Paul Scherrer Institute (PSI, Villigen, Switzerland). The applied settings for the measurements of the samples were: sample-to-detector distance (SD), 3 m; collimation, 3 m; neutron wavelength (λ), 5.52 Å. As a subtractable sample background, we measured the D₂O-based culture medium, which was used as a suspension buffer for the algal cells during the measurements. The instrumental background was recorded with the beam blocked by a cadmium plate. All experimental data were corrected for detector efficiency obtained from the measurements performed on Cd plate, cuvette, and H₂O in a quartz cuvette with 1 mm path length, with instrument setting of SD, 1.5 m; collimation, 2 m; and λ, 5.52 Å. For data treatment, the “Graphical Reduction and Analysis SANS Program” package (GRASP) (developed by C. Dewhurst, Institut Laue–Langevin [ILL]) was used. The observed 2D scattering patterns did not reveal significant anisotropy; therefore, further analyses were performed on 360° radially-averaged scattering curves. The scattering curves exhibited two characteristic peaks, ~0.035 and 0.06 Å⁻¹. To quantify any structural changes observed with SANS, the 1D scattering curves were fitted with a sum of constant, a power function, and two Gaussian functions, similar to the method described earlier (Nagy et al., 2013). The fitting range was 0.0204–0.0856 Å⁻¹. Calculated repeat distance values for the thylakoid membranes in the different samples are represented in Table 1.

Circular dichroism spectra from the control and salt-stressed cells were recorded at room temperature in the spectral range of 400–800 nm with 3 nm spectral resolution using a J815 spectropolarimeter (Jasco, Japan). The scan speed was set to 100 nm min⁻¹ with an integration time of 2 s. The cells were diluted to a Chl concentration of 20 μg ml⁻¹ for each sample. The measurements were carried out in standard glass cuvettes with an optical path length of 1 cm. For baseline correction, the culture media were used. The CD spectra were normalized to the Chl Qy absorption band with a reference wavelength at 750 nm. Differences of the amplitudes of (+)690 and (–)674 nm psi-type CD bands were calculated to compare the effect of salt treatment on the psi-type CD in the red spectral region.

Chl a fluorescence transients from the control and salt-stressed cells were measured using Aqua Pen AP-C 100 fluorometer (PSI, Czech Republic). Cultures were dark adapted for 20 min before the measurements. A cell suspension containing a total of 1 μg Chl was pipetted into a cuvette. Illumination (650 nm) was provided by an LED array which was focused on the sample to provide uniform irradiance. The fluorescence measurements were carried out with a 5 s flash of 3,000 μmol photons m⁻² s⁻¹. Fluorescence transients were recorded from the three biological repetitions.

The Chl a fluorescence and light-induced absorbance changes at 820 nm of intact cells were measured using a Dual-PAM-100 chlorophyll fluorometer. (Heinz Walz GmbH, Germany). Cell suspension equivalent to 40 μg Chl was filtered onto a Whatman glass microfibre filter (GF/C) of 25 mm diameter, which was placed in between the two microscope coverslips separated by a spacer. The basal fluorescence level (Fo) when all reaction centers open in PSII was determined first after 20 min of dark adaptation. Fm in the dark-adapted state was obtained by triggering a saturating light pulse at a light intensity of 3,000 μmol photons m⁻² s⁻¹. Actinic red light with increasing intensity was applied consecutively in the range of 10–664 μmol photons m⁻² s⁻¹ for 2 min which was enough to achieve the steady-state Fo. The Fm' level was obtained at each actinic light intensity by a saturating light pulse. Calculated parameters are represented in Table 2. ETR and redox changes of P700 were determined from the light response curves. The PSI yield (φ_I) and PSII yield (φ_II) were calculated according to Kramer et al. (2004) and Klughammer and Schreiber (1994), respectively.

The protocol of Genty et al. (1989) was used to measure the non-photochemical quenching (NPQ) of Chl fluorescence. The measurements were done using a Dual-PAM-100 chlorophyll fluorometer. The dark-adapted cells were first exposed to a weak modulated beam of 2 μmol photons m⁻² s⁻¹ followed by a saturating flash of 8,000 μmol photons m⁻² s⁻¹ for a duration of 800 ms. After 20 s, the cells were exposed for 15 min to a continuous actinic light of 660 μmol photons m⁻² s⁻¹. Thereafter, the cells were exposed to a saturating pulse of 8,000 μmol photons m⁻² s⁻¹ for 800 ms. Continuous actinic light was turned off after 10 s, followed by 5 min of dark adaptation (to determine the recovery of fluorescence).

Fluorescence emission spectra at 77 K were recorded with a Fluorolog 3 double-monochromator spectrofluorometer (Horiba Jobin-Yvon, IL, USA). The cell suspension equivalent to 5 μg Chl/ml was filtered and deposited on Whatman glass fiber (GF/B) discs. The filters were flash frozen in liquid nitrogen and transferred to a Dewar vessel filled with liquid nitrogen that was placed into the measurement chamber of the spectrofluorometer. The emission spectra from 650 to 800 nm were recorded with excitation wavelengths of 436 and 480 nm. The excitation and emission bandwidths were set to 3 and 5 nm, respectively. The measurements were performed with 1 nm resolution and 1 s integration time. Three independent repetitions were made for each treatment and averaged.

For pigment determination, 1.5 ml of cell suspensions harvested from day 0 to day 4 were pelleted by centrifugation at 3,000 g for 3 min and snap frozen in liquid nitrogen and stored at −80°C until use. The pigments were extracted by resuspending the cells in ice-cold 100% acetone followed by 30 min dark incubation with continuous shaking at 1,000 rpm. The samples were centrifuged at 11,500 g for 10 min at 4°C and the supernatant was passed through a PTFE 0.2 μm pore size syringe filter.

The pigment composition of the cells was determined by high-performance liquid chromatography (HPLC), using a Shimadzu Prominence HPLC (Shimadzu Corp., Japan) system as described by Zsiros et al. (2020). The pigments were identified according to their retention times and absorption spectra and quantified by the integrated chromatographic peak area recorded at the wavelength of maximum absorbance for each kind of pigments using the corresponding molar decadic absorption coefficient (Wright et al., 2005).

The thylakoid membranes were isolated from 4-day old cells in a medium containing 50 mM [4-(2-hydroxymethyl)-1-piperazineethanesulfonicacid] (HEPES)/KOH buffer, pH 7.5, 300 mM Sorbitol, 2 mM Na₂-ethylenediaminetetraaceticacid (EDTA), 1 mM MgCl₂, and 1% w/w bovine serum albumin (BSA) by breaking the cells in a Precellys Evolution (Bertin Technologies, France) homogenizer using the glass beads according to the protocol of Block and Albrieux (2018) with slight modifications. The thylakoid fractions from the disrupted cells were collected on a three-step Percoll gradient, washed with 50 mM HEPES-KOH pH 8.0 and 330 mM Sorbitol containing buffer, and solubilized with 2% (w/v) n-dodecyl-β-D-maltoside (DDM) in 25 mM bis-Tris/HCl pH 7.0, 20% glycerol (w/v) for 30 min at 4°C. The unsolubilized membranes were removed by centrifugation at 18,000 g for 20 min. The solubilized complexes were separated on a 5–13% (w/v) BN-PAGE, according to Schagger et al. (1994).

Paramylon was extracted and purified according to the protocol of Barsanti et al. (2001) with small modifications. The cells (control as well as salt-treated on day 4) were pelleted by centrifugation of 20 ml culture and frozen for a minimum of 4 h. Then, the pelleted cells were resuspended in a medium containing 1% SDS and 5% Na₂EDTA and incubated for 30 min at 37°C. After incubation, the suspension was centrifuged for 10 min at 1,000 g to extract the paramylon granules. The treatment was repeated, and the extract was washed two times with hot distilled water (70°C). After washing, the suspension was deposited on a pre-weighted Whatman glass microfibre filter (GF/C) and dried overnight at 90°C. The dried filters were weighted to quantify the paramylon content. The amount of paramylon per milligram dry weight was calculated for all sample types. The dry weight was determined gravimetrically.

The Origin Pro 8 program (OriginLab Corporation, MA, USA) was used for the statistical analyses. All the measurements were carried out using at least three independent biological experiments. Data are expressed as mean ± SE. The significant levels are tested using one-way ANOVA at p < 0.05. The multiple comparisons of the means to determine the trends and pattern of distribution of the data were done by Tukey's post-hoc test.

The time course of the salt-stress driven changes of the growth rate and total Chl contents of the cells were determined spectrophotometrically. The optical density values at 750 nm for the salt treated cells compared with the control revealed a reduction in the growth rate (Figure 1A). In all cases, the cells appeared to reach the stationary phase on day 4. The total Chl content was decreased with the increasing salt concentrations (Figure 1B). These findings are in agreement with the earlier observation on salt-treated E. gracilis and C. reinhardtii cells (González-Moreno et al., 1997; Subramanyam et al., 2010; Neelam and Subramanyam, 2013).

The decreased Chl content might indicate the alterations in the photosynthetic apparatus under salt stress. Therefore, CD spectroscopy was used to monitor the changes in the macro-organization of the thylakoid membranes upon salt treatment. The CD spectra recorded from the control and salt-treated cells on different days are shown in Figure 2. In the Chl Q_y region (red maximum ~ 696 nm), the spectra showed characteristic psi-type CD signatures similar to the higher plant thylakoid membranes, originating from the long-range chiral dipole-dipole interactions of Chls of PSII and LHCII embedded in the membranes (Lambrev and Akhtar, 2019). However, unlike in higher plants, there was no psi-type CD in the blue region. The CD in the blue region was also somewhat different than in the green alga C. reinhardtii (Tóth et al., 2016) but similar to the diatom Cyclotella meneghiniana (Ghazaryan et al., 2016). The CD signal in the Soret region contains contributions from short-range, excitonic interactions between the Chls and carotenoids; therefore, the different carotenoid compositions can give different CD signals. The psi-type CD is associated with the presence of grana stacking and it disappears upon unstacking the membranes by removal of cations from the medium (Lambrev and Akhtar, 2019). However, the E. gracilis cells lack typical granular organization, similar to some other algae like diatoms, but they still have psi-type CD signature (Szabó et al., 2008). Upon salt treatment, the effects were only visible on the main psi-type CD band, the amplitude of the main band was reduced by the increment of salt concentration. The largest change was observed on the third day. No visible change was observed in the blue region of the spectra. The amplitude of psi-type CD band plotted as a function of salt concentration in Figure 2D, showed significant decrease on the third day of treatment.

We studied the ultrastructural changes of the thylakoid membranes of the E. gracilis cells induced by the salt treatments using SANS. The SANS profiles (Figure 3) of the cells show two well-defined peaks at ~0.035 and 0.06 Å⁻¹; the 0.035 Å⁻¹ peak corresponds to a repeat distance of thylakoid membranes of 179 Å; the peak at a higher q value probably originates from stacked membrane pairs (Nagy et al., 2013). Upon salt treatment, both the peaks were shifted toward higher Q-values (~ 5.5%) indicating shrinking of the periodic lamellar order. However, the shrinking effect did not show a correlation with the salt concentration, differences between the effects of different salt concentrations seem insignificant (Figure 3; Table 1).

We used SEM and TEM to further visualize the effects of salt-treatments on the morphology of E. gracilis cells and the ultrastructure of their thylakoid membranes (Figure 4). The cells did not show significant alteration in their shape and external morphology at 150 mM NaCl treatment (Figures 4A,B). The cells contained 4–5 disc-shaped chloroplasts, which were located close to the periphery of the cells, parallel to the plasma membranes (Figures 4C,D). Chloroplast comprised elongated lamellae, each formed by 2–4 (rarely 5) closely appressed thylakoids (Figures 4C,D). Upon salt-treatment, we observed shrinkage of thylakoid membranes and an increased number of paramylon grains (Figure 4D). The RD of the thylakoid membranes calculated from electron micrographs showed a reduction in the values of the salt-treated cells (18.6 ± 0.2 nm) with reference to the control (20.5 ± 0.2 nm); the shrinkage of thylakoid membranes, qualitatively, agrees well with the SANS observations.

Low-temperature fluorescence spectroscopy provides information about the distribution of excitation energy between the different pigment-protein complexes of the two photosystems. Representative fluorescence emission spectra at 77 K from the control and salt-treated cells of E. gracilis upon Chl a (436 nm) excitation are shown in Figure 5. The contributions of the two photosystems to the fluorescence can be estimated based on the emission spectra. The emission spectra are very similar to the previously published data (Doege et al., 2000) on the same species dominated by the emission from PSI only, with the peak at 727 nm. The small shoulder could also be seen at 685 nm indicating negligible fluorescence from PSII. The emission spectra are very different from the higher plant thylakoid membranes (Walters and Horton, 1995) and other green algae (Kramer et al., 1985), where both the photosystems display distinct peaks. Similar results were obtained upon 480 nm excitation (preferentially exciting Chl b) (data not shown). While comparing different salt treatments, no significant changes were observed, indicating no or negligible effect of salt on the photosynthetic energy transfer.

The fast Chl a fluorescence induction (OJIP) transients, measured by exposing the dark-adapted samples to high light, show a complex multistep rise curve which represents a fingerprint of the species and the physiological status of the cells (Srivastava and Strasser, 1995). The shape of transient is sensitive to many photosynthetic processes, excitation energy flow from the antenna to the reaction centers, the structural plasticity of the reaction center complexes, and the electron transfer on the donor and the acceptor side of PSII, and the availability and redox state of intersystem electron carriers and downstream electron transfer to PSI. According to the theory of energy flux, any change in any of these processes will induce a change in the shape of the induction curve. The fluorescence induction curves recorded from E. gracilis cells have an unusual shape, with no well-visible J and I steps (Figure 6). In the higher plants (Stirbet et al., 2018) and green alga C. reinhardtii (Kodru et al., 2015), after 100 ms, there is a fast drop in the transient fluorescence which is usually assigned to the state transitions, NPQ or redox state of P700 (Strasser et al., 2004). In the case of E. gracilis, the induction curve stays flat up to 1000 ms. This could be very well explained by the common antenna system of PSI and PSII or the absence of the classical violaxanthin cycle. When comparing the transients for different salt-treatments, no remarkable difference could be observed (Figure 6). Very similar F_v/F_m values, 0.6 were recorded for both the salt-treated and control cells, thus the functional state of PSII is not affected by the salt treatment.

Furthermore, to understand how salt stress affects photosynthesis in E. gracilis cells, the light curves were recorded. ETR and PSI activity in both the control and salt-treated cells were measured on different days (Figure 7). In good agreement with the previous observations, that upon NaCl treatment there is no significant effect on the functional state of PSII (F_v/F_m) of the cells, we also found that the linear ETR and the activity of PSI were not affected. Upon irradiation with high-light intensity, both the control and salt-treated cells displayed considerable induction of non-photochemical quenching. The NPQ of 0.4 was observed which is similar to the findings of González-Moreno et al. (1997). However, the NPQ of the salt-treated cells was not significantly different from the control cells as represented in Table 2.

The reduction kinetics of P700 in both the control and salt-treated cells are shown in Figure 8. The activity of PSI was monitored by the absorbance change at 830 nm due to the oxidation of P700. The fluorescence decay lifetimes of the 50, 100, and 150 mM NaCl treated cells are 11.05 ± 0.238, 10.12 ± 0.177, and 9.997 ± 0.276, which is similar to the lifetime of control cells, 10.98 ± 0.227 on day 4 of the treatment as shown in Table 3. The salt treatment induced no significant changes in the PSI activity.

The pigment extracts of the control and salt-treated cells were analyzed by HPLC to determine the effects of salt on the pigment composition. Typical HPLC profiles exhibit peaks for Chl a, Chl b, β-carotene, Nx, and Ddx. Ddx is found in the antenna of many diatom species (Lavaud et al., 2003), while the antenna of green algae (Goss and Jakob, 2010) and higher plants are known to contain lutein instead (Liu et al., 2004). However, diatoxanthin which is a constituent of classical Ddx-cycle in the diatoms was not detected. During salt stress, the relative composition of carotenoids did not change, however, every carotenoid alone and so the total carotenoid content increased significantly compared with Chl (Figure 9).

Useful information about the composition and interactions of the protein complexes in the thylakoid membrane can be obtained by solubilizing the membrane by the mild detergent treatment followed by their separation on a polyacrylamide gel. BN-PAGE is one of the methods used for the separation of protein complexes in their native and functional form. In this study, BN-PAGE was used to determine whether the salt treatment induces changes in the interactions between the protein complexes or leads to the formation of larger supercomplexes. The patterns of the thylakoid membrane protein complexes, after solubilization with DDM and separation by BN-PAGE (Supplementary Figure 1), showed similarity to that of the higher plants (Järvi et al., 2011). No visible changes were observed in the pattern of BN-PAGE when comparing different salt treatments.

Our ultrastructural analyses indicated the accumulation of paramylon grains in the cytoplasm of stressed cells (Figure 4D). A protocol has been used to isolate paramylon (Supplementary Figure 2) to quantify its content in the different samples. Table 4 shows the changes in paramylon content at a cellular level in the E. gracilis control and salt-treated cells after 4 days of treatment. The accumulation of paramylon increased with the increasing salt treatment with the highest levels of paramylon observed in the cells grown in the presence of 150 mM NaCl. The increase was 2.5-, 3.5-, and 5-fold in the 50, 100, and 150 mM NaCl treated cells, respectively.