The seeds of S. europaea were purchased from Rühlemann’s Kräuter & Duftpflanzen (Germany) and the seeds of C. officinalis and A. hortensis were from Magic Garden Seeds (Germany). The plants were germinated in soil [substrate type P; pH = 5.9; 120 mg L⁻¹ of N; 120 mg L⁻¹ of PO 4 2 − ; 170 mg L⁻¹ of K; 120 mg L⁻¹ of Mg; density, 430 kg/m³; 70% raised bog peat (degree of decomposition: H2-H5), 30% clay; Einheitserdewerke Werkverband e.V., Germany]. When two leaves had fully developed, the plants were transferred to pots (diameter, 8 cm) containing one-third of soil [substrate type T; pH 5.9; 180 mg L⁻¹ of N; 180 mg L⁻¹ of PO 4 2 − ; 260 mg L⁻¹ of K; 130 mg L⁻¹ of Mg; density, 430 kg/m³; 70% raised bog peat (degree of decomposition: H2-H5), 30% clay; Einheitserdewerke Werkverband e.V., Germany], one-third of fine quartz sand (grain size 0.5–1 mm), and one-third of coarse quartz sand (grain size 2–3 mm) (Euroquarz GmbH, Germany). The water content of the soil with respect to salt treatment can be found in Table S2 . The plants were irrigated with a modified Hoagland solution ( Table S1 ).

To determine the fresh mass, the 12 plants were cut at the root and then the aboveground part of the plants was weighed as whole plants; then, three plants were pooled and the pooled leaves were weighed separately. For further analysis of the metabolite content, the main leaves of A. hortensis and C. officinalis were harvested, and the green aboveground part (which is later on referred to as leaves) of S. europaea was harvested and pooled into four technical replicates per salt treatment at each experiment, resulting in eight replicates per light regime from two independent experiments. After harvesting, the plants were immediately frozen in liquid nitrogen and then freeze-dried for 1 week until completely dry. Dry mass was determined by weighing the pooled leaves before (FM) and after (DM) freeze drying. Percent dry mass was calculated as DM/FM*100. For further analysis, plant samples were homogenized using a Retsch mill (Retsch MM 400; Retsch GmbH, Germany) [three to five times for 50 s with three to five metal beads (diameter, 9 mm) at 25 Hz].

The chloride content of the sampled leaves was determined by ion chromatography. For this purpose, 10 mg of dried plant material was dissolved in 1 ml of ultrapure water. As an internal standard, 0.5 ml of sodium bromide solution (0.6 mg ml⁻¹) was added. The samples were sonicated on ice for 10 min and then centrifuged for 5 min (4,500×g, 4°C). Next, the samples were diluted with ultrapure water, according to the expected salt concentration of the samples, to fit into the calibration range of chloride. Chloride determination was carried out using a 930 Compact IC Flex ion chromatograph (Metrohm AG, Switzerland) equipped with a conductivity detector and suppression system. A Metrosep A Supp 5-250/4.0 column was used with a flow rate of 0.7 ml min⁻¹ and an injection volume of 20 µl. Gradient elution was performed using Na₂CO₃ (3.2 mM) and NaHCO₃ (1 mM). The final chloride concentration was calculated with external calibration using a chloride standard (>99%; Carl Roth GmbH, Germany).

Determination of ABA content was performed as previously described (Errard et al., 2015) with modifications. In brief, 10 mg of the dried plant material was extracted with 0.2 ml of methanol/water (60:40, v/v), and an internal standard [(+)-abscisic acid-d6, Toronto Research Chemicals, Canada] was added. First, the solution was sonicated on ice for 15 min and then centrifuged for 10 min (12,298×g, 4°C). Next, the supernatant was collected in a micro reaction vessel and the extraction steps were repeated twice. Then, the collected supernatant was filtered through a PTFE filter tube (0.2 µm, Thermo Fisher Scientific Inc., USA) and transferred to HPLC vials. Finally, the filtrate was diluted 1:2 with MS water (Supelco, VWR, Germany) + 0.1% acetic acid. The measurement was performed using an Agilent Technologies 1260 Infinity HPLC (Agilent Technologies Sales and Services GmbH & Co. KG, Germany) in combination with a Triple Quadrupole Q-Trap^® 6500-MS/MS system (AB Sciex LLC, USA). Chromatographic separation was performed using a Zorbax Eclipse Plus C18 column (1.8 µm, 2.1 mm × 50 mm; Agilent Technologies, Germany), a column temperature of 30°C, a flow rate of 650 µl min⁻¹, and a mobile phase consisting of solvent A: MS water + 0.1% acetic acid and solvent B: acetonitrile + 0.1% ultrapure water. The injection volume was 10 µl. The initial gradient was 90% solvent A for 1 min, reduced to 15% solvent A for 4 min, and then reduced to 0% solvent A for 4 min. The mass spectrometer was operated in negative ionization mode and an electron spray ionization source was used. The MS parameters were set as follows: ion source temperature, 500°C; ion spray voltage, −4,500 V; curtain gas pressure, 50 psi; drying gas pressure, 50 psi; nebulizer gas pressure, 50 psi; auxiliary gas pressure, 65 psi; and multireaction monitoring (MRM) at a dwell time of 0.3781 s. Identification was based on the retention time and MRM transitions of the following: ABA 263 → 153 [quantifier; collision energy (CE), −15 V], 263 → 203 (qualifier; CE, −40 V), and 263 → 122 (qualifier; CE, −48 V) and ABA-d6 269 → 159 (quantifier; CE, −15 V), 269 → 209 (qualifier; CE, −40 V), and 269 → 128 (qualifier; CE, −48 V). The final ABA concentration was calculated from a calibration curve of the quantifier ratios between an ABA standard (≥98.5%, Sigma Aldrich) and the internal standard. Data analysis was performed with the Analyst 1.6.2 software (AB Sciex LLC, USA).

The extraction of pigments was performed according to Frede et al. (2019). In brief, 5 mg of plant material was dissolved in 0.5 ml of methanol/tetrahydrofuran (1:1, v/v) and incubated for 5 min in a shaker (1,400 rpm, 20°C) followed by centrifugation for 5 min (4,500×g, 20°C). The supernatant was collected in a vial, and the extraction was performed five times. The solution was evaporated to dryness under a nitrogen stream, dissolved in dichloromethane/isopropanol (1:5, v/v), sonicated (3 min, 20°C), filtered (PTFE filter tubes), and transferred to an HPLC vial. The analysis was performed using Agilent Technologies 6530 QToF-DAD-UHPLC-MS (Agilent Technologies Sales and Services GmbH & Co. KG, Germany) according to Frede et al. (2017). Identification was achieved using mass spectra and UV/VIS spectra ( Figure S1 ), and quantification was achieved using external calibration with carotenoid standards (CaroteNature GmbH, Switzerland) of all-trans-isomers from β-carotene, lutein, and zeaxanthin and 9-cis-neoxanthin as well as chlorophyll standards (Sigma Aldrich Chemie GmbH, Germany) of chlorophyll a and b at a wavelength of 450 nm. Data analysis was performed using the TOF Quantitative Analysis (Quant-My-Way) 10.2 (MassHunter, USA).

Statistical differences between the light regime and salt treatments were tested using SigmaPlot (14.0) with a two-way ANOVA followed by a Bonferroni post-hoc test (p ≤ 0.05) (dF, F, and p-values represented in Table S3 ). Data are presented as means ± SEM of two individual experiments per light regime. Twenty-four plants per light regime were used for the determination of fresh mass. For the analysis of the selected metabolites, eight replicates per light regime were used, pooled from three individual plants and two independent experiments.

To evaluate the variation in both light regimes, LR1 and LR2, the light spectra, light intensity, and daily light hours were measured ( Table 1 ; Figures S2, S3 ). The major differences between both light regimes were detected in the daily light hours and light intensity. To encounter both, the DLI was calculated ( Table 1 ). Light regime 1 showed an average DLI of 3.18 ± 0.26 µmol m⁻² day⁻¹ that was only 18% of the DLI of light regime 2, which was on average 17.98 ± 0.24 µmol m⁻² day⁻¹. This is due to 3 h less day light and a 290 µmol m⁻² s⁻¹ lower light intensity in light regime 1 (LR1, greenhouse) compared with light regime 2 (LR2, indoor farming). Light regime 1 showed, due to variations in natural light, variations in DLI ( Figure S2 ).

To estimate the effect of the light regime on salt treatment on photosynthesis, we analyzed the pigment (carotenoids and chlorophylls) content by UHPLC-DAD-QToF-MS. Chlorophyll a and b, as well as all-trans-isomers of lutein, β-carotene, zeaxanthin, and violaxanthin and the 9-cis-isomer of neoxanthin, were detected in all plant species ( Figure S1 ). Both carotenoids and chlorophylls were affected by salt treatment and light regime, and an interaction between these two factors was observed, with different responses for facultative halophytes (C. officinalis and A. hortensis) and the obligate halophyte (S. europaea). Due to the high impact on water content, the metabolites are presented on a dry mass basis ( Tables 2 , 3 ), while the results based on a fresh mass basis can be found in Tables S5, S6 .

To gain insight into the influence of light regime and salt treatment on the dynamic metabolic variation, a PCA analysis was performed ( Figure 4 ). The greatest influence was due to the difference in salt treatment. For all three halophytes (C. officinalis, A. hortensis, and S. europaea), distinct clusters were found for the treatments with and without salt based on PC1 and PC2 ( Figures 4A–C ). With respect to the light regime, the effects on the metabolite profiles were less pronounced. However, interactions on individual metabolite levels have been demonstrated, e.g., for carotenoids and chlorophylls ( Tables 2 , 3 ).

In the evaluation of salt-tolerant crops, the yield loss in response to the salt concentrations in soil or water is a key aspect to be considered. Considering the salt treatments, a reduction in fresh mass was found at 50/200 mM for C. officinalis, at 200 mM for A. hortensis, and at 600 mM and no salt for S. europaea. Evaluating the effect of the light regime in the non-salt-stressed conditions (A. hortensis and C. officinalis no salt and for S. europaea 50 and 200 mM of salt), the fresh mass is higher in indoor farming (LR2). This suggests that the DLI in the greenhouse (LR1) was too low for optimal growth. However, relative fresh mass still differs in the light regimes at salt treatments 50 and 200 mM for S. europaea and C. officinalis ( Figure S7 ). Additionally, we observed differences in the influence of salt treatment on the water content, between both the facultative halophytes (A. hortensis and C. officinalis) and the obligate halophyte (S. europaea). Water content was less affected by salt treatment for the obligate halophyte. Succulent halophytes (salt-accumulating), like Suaeda maritima, show a different osmotic adjustment and, thus, a different water content under salinity (Flowers and Colmer, 2008). To evaluate the influence of light regime in relation to salt treatment, further insights were obtained by studying changes in the ABA and chloride contents. ABA serves as an indicator of salt stress response in halophytes and glycophytes and mediates the stomatal movement of guard cells (Zhang et al., 2006; Karimi et al., 2021). According to previous research, we observed a correlation between increased salt stress (increased ABA content) and yield loss (Breckle, 2002; Metselaar, 2013). However, the response of ABA differs between halophytes and glycophytes, at least with respect to the salt level. A study conducted by Karimi et al. (2021) showed only a short-term response of ABA in Thellungiella salsuginea (a halophyte) at 200 mM, while Arabidopsis thaliana (a glycophyte) showed a long-term response. Ben Hassine et al. (2009), on the other hand, showed an increased ABA content in the seedlings of the halophyte Atriplex halimus in a short- and long-term response to 160 mM of NaCl treatment. This suggests that ABA regulation not only differs among glycophytes and halophytes but also among halophyte species. Aside from sodium content, chloride content also changes with salinity treatment, and accumulation varies between halophyte species. For instance, the ratio of sodium and potassium cations to chloride anions varies among halophyte species, which could influence the external chloride uptake (Flowers and Colmer, 2008) and should be considered in future studies.

In this study, C. officinalis showed an increased ABA content correlating with increased chloride and reduced growth already from 50/200 mM. This is in accordance with a lower salinity tolerance (de Vos and Broekman, 2013). In contrast, de Vos and Broekman (2013) observed a higher percent dry weight at 200 mM and a leaf succulence, whereas, in our study, it was only significantly increased at 600 mM. The lowest chloride accumulation in the leaves of C. officinalis compared with the two other plant species (A. hortensis and S. europaea) would support the salt exclusion mechanism if the salt is translocated into the xylem and root (Chen et al., 2018). Atriplex hortensis showed an increased ABA content (only in LR1), paired with a reduction in fresh mass and chloride accumulation at 200 mM of salt. This is in accordance with the literature, where a salt tolerance of up to 250 mM of salt was shown for another variety of A. hortensis (red orache) (Wilson et al., 2000). Furthermore, we observed excreted salt crystals on the leaf surface, which is typical for halophytes with salt bladders (Schirmer and Breckle, 1982). Therefore, when considering chloride content in A. hortensis, it is important to consider the salt deposition on the leaf surface. One possibility is to wash off the salt from the leaves before measurement, but this may not reveal the transport of salt into the leaves, making it more difficult to compare salt tolerance mechanisms. Nevertheless, it would be interesting to distinguish between the salt excreted and the salt accumulated from and in the leaf. Since S. europaea is an obligate halophyte, it showed an increase in ABA and a decrease in fresh mass not only at 600 mM but also at no salt, unlike the other plant species, suggesting that this salt concentration and very low salt lead to stress. This can be explained by the fact that in obligate halophytes, salt uptake is essential for maintaining turgor and for optimal growth and is also reflected in the water content, which changes only slightly with salt treatment (Glenn and O’Leary, 1984). Furthermore, when considering the differences between chloride content in shoots based on dry and fresh mass, it can be clearly observed that S. europaea has a lower chloride/fresh mass ratio at higher salt treatments compared with the other plant species. This is due to lower salt-induced water loss, indicating better osmotic regulation in S. europaea at higher salinity levels.

Whereas the influence of the light regime on fresh mass was clearly visible, the effect on metabolite profiles was less pronounced. However, the influence of the light regime on metabolite profiles also differed within plant species ( Figure 4 ). Again, different patterns were observed with respect to salinity treatment, suggesting a different adaptation to salinity and a different influence of the interaction between salinity treatment and light regime in different halophyte species. This interaction is particularly interesting for photosynthetic pigments.

Carotenoids and chlorophylls have multiple functions in plants; for example, carotenoids are accessory pigments, and also they have essential photoprotective properties, while chlorophylls are the main pigments of photosystems. Carotenoid and chlorophyll biosynthesis and metabolism are affected by light, e.g., light quality or light intensity, as well as salinity (Pizarro and Stange, 2009; Tanaka and Tanaka, 2011; Soltabayeva et al., 2021). All the pigments studied are part of the photochemical apparatus but have different functions according to which they can be divided into two groups. First, chlorophyll a and b, lutein, and neoxanthin, in a simplified way, function as absorbers and converters for the incoming light energy (Choudhury and Behera, 2001). Second, violaxanthin, β-carotene, and zeaxanthin, on the other hand, function as dissipators of excessive light energy (Choudhury and Behera, 2001). This should be taken into account as we observed a different pattern in pigment accumulation between the two light regimes in the salt stress and salt tolerance range.

Salt stress leads to limited activity in several parts of the photosynthetic apparatus (e.g., RuBisCO activity, NADPH oxidase activity) and, thus, increased formation of reactive oxygen species (ROS) (Hasanuzzaman et al., 2020). In high light stress, the high photon flux density leads to excessive light energy that exceeds the capacity of the photosynthetic apparatus, resulting in the formation of ROS, which can cause subcellular damage and photooxidation of pigments (Gilmore, 2001). A study by Simkin et al. (2003) showed that in pepper (Capsicum annuum L. cv. Yolo Wonder), the photooxidation of carotenoids already occurs at the transition of light intensity from 150 to 280 µmol m⁻² s⁻¹. Since we observed a combined effect of light regime and salt treatment and found a strong difference in DLI between the two light regimes, this difference must be taken into account. Higher light energy combined with the salt stress-induced limited activity of the photosynthetic apparatus results in overexcitation of the photosynthetic apparatus and increased ROS formation that exceeds antioxidant capacity (Carillo, 2018). Accordingly, we observed a salt stress-induced reduction of all pigments in indoor farming (LR2) with a higher DLI. In contrast, within salt tolerance ranges, we observed an accumulation of the carotenoids violaxanthin, β-carotene, and zeaxanthin, which act as dissipators of excess light energy and thus scavenge ROS, and a decrease in lutein and neoxanthin and chlorophylls, which act as absorbers and converters of incoming light energy and thus maintain photosynthetic activity (Choudhury and Behera, 2001). These effects are in accordance with the changes in our study in indoor farming (LR2) and resulted in higher levels of violaxanthin, β-carotene, and zeaxanthin during salt stress at lower DLI in greenhouse (LR1) and suggest that DLI affects the carotenoid profiles as a function of salt concentration with respect to their different functions in the photosynthetic process.

Interestingly, the response of the obligate halophyte (S. europaea) in salt stress (no salt) is different from the response of the facultative halophytes (C. officinalis and A. hortensis). For S. europaea, we observed particularly low contents of pigments in both light regimes at no salt. An explanation could be a different adaptation of the photosynthetic apparatus to salt. It is suggested that halophytes have the ability to regulate steady chloride concentrations by a different ion (Na⁺, Cl⁻, and K⁺) transport compared with glycophytes (Bose et al., 2017). Since salt is essential for maintaining intracellular pH, altered pH in the thylakoid interior could affect the function of crucial enzymes for photosynthesis, e.g., RuBisCO or NADPH oxidase. This could influence photosynthetic activity, e.g., photooxidation of pigments and biosynthesis of carotenoids and chlorophylls (Glenn and O’Leary, 1984).

Taken together, we observed an interaction of light regime and salt treatment in influencing the performance of the three halophyte species. Therefore, when optimizing the light conditions in indoor farming, the plant species, salt tolerance, and salinity of the cultivation medium must be taken into account. In indoor farming, lighting conditions are not only important for the plants but also for evaluating the profitability and sustainability of a production system. Hence, light efficiency use (LUE) is a factor, considering the consumed electricity of the system, which helps to compare indoor farming and greenhouse cultivation. A study by Jin et al. (2022) pointed out that the average LUE in vertical farming is higher than in greenhouse cultivation. Considering the influence of salt stress and light on yield, assuming LUE is higher under lower DLI (greenhouse) than under higher DLI (indoor farming), therefore, lower light intensity in saline indoor farming could decrease light energy while maintaining yield and, thus, optimize LUE. Nevertheless, the DLI in the greenhouse (LR1) was also too low. Therefore, lower DLI with moderate salinity could lead to optimized resource use and even improved nutritional quality by increasing the amount of PSMs. The implementation of UV-B LEDs or colored LEDs could further enhance the PSM content (Wiesner-Reinhold et al., 2021; Frede and Baldermann, 2022). Further research could aim to study the influence of DLI and salinity on other nutritive compounds, such as polyphenols and vitamins.

The major limitation of this study is that the effects are assumed to be due to the daily light integral and not due to light quality. An altered light quality, in this case mainly light spectra, has also an influence on the plant metabolism and pigment content (Alrifai et al., 2019; Frede et al., 2019). For example, different photoreceptors can be activated through changes in the light spectra (Kami et al., 2010). However, our study design aimed to investigate the differences between greenhouse cultivation and indoor farming, and thus, there are differences not only in the light regime but also in light intensity. Since the daily light impact was highly influenced (72% differences between both light regimes) by the light regimes, we focused on this while explaining the results. Nevertheless, it would be beneficial to investigate further influences of different light parameters on the quality of vegetables in indoor farming systems.

Regarding the response of plants to salt stress, it is important to know whether salt was applied in a single step or gradually. If salt is applied in a single step, there is a possibility that plants will suffer from salt shock (Shavrukov, 2012). In our study, salt was applied in a single step. However, salt was applied in an NFT system where the pots were irrigated from below, which resulted in a slower accumulation of salt in the soil. In addition, plants had a long acclimation period of 17 days, during which they could have recovered from the osmotic shock (Shavrukov, 2012). Nevertheless, this is an important point that should be considered in future studies which may affect the tolerance of plants to ionic stress and, hence, their response to varying light regimes.

It would be interesting to study the modification of light conditions with respect to the adjustable salt tolerance of halophytes and the impact of light in relation to the use of different saline water sources. The salt concentration is not only dependent on the water source, e.g., brackish water, wastewater, or brine water, but also on the location (Atkinson and Bingman, 1997). For example, regional brine waters have different salt concentrations and compositions (Fitzner et al., 2021). One option to adjust the salt concentration to the halophyte salt tolerance range is dilution with freshwater. However, freshwater is an exhaustible resource, and in sustainable agriculture, freshwater consumption should be reduced (Gleick, 1993). If there is a way to regulate light intensity, this would be a potential solution.

Further research also could aim to study other halophyte species to broaden the picture of differences between obligate and facultative halophytes and investigate the interaction of the salt tolerance mechanism and the influence of light.

In conclusion, this study highlights the potential applications of halophytes for indoor farming and also hints at the adaptation of photosynthesis during salt stress under different light regimes in halophytes. Furthermore, optimization of indoor farming lighting conditions, taking into account salinity and plant species, could improve resource efficiency and pigment profile. Given the limited land and water use and the changing climate conditions, we argue that indoor farming has a high potential to become a fundamental contributor to global agriculture. In addition to sustainable crop production, healthy and sustainable nutrition will be a valued aspect of future diets. Halophytes are not only suitable for indoor farming but can also be irrigated with saline water, which conserves freshwater resources, and are additionally rich in PSM. Hence, saline indoor farming with halophytes could contribute to food and nutritional security in the future.