The Effect of Nutrient Source and Beneﬁcial Bacteria on Growth of Pythium -Exposed Lettuce at High Salt Stress

: High salinity, nutrient imbalance, and pathogens are some of the challenges of closed soilless cultivation systems, e.g., those combining hydroponics (HP) with aquaculture efﬂuents (AE). Plant growth-promoting microorganisms (PGPM) can support plants to cope with stressing agents. To address these topics, lettuces were grown in soilless systems (20 boxes) at an electrical conductivity of around 4.2–5 mS/cm, following a full factorial design with two nutrient sources and ﬁve bacterial treatments. The nutrient sources were either organic (AE) or inorganic (HP); the treatments were either commercial PGPM or sludges of an aquaculture farm or of an urban wastewater treatment plant. Finally, half the plants were exposed to pathogen Pythium sp. After 61 days of culture, most of the differences between HP- and AE-plants could be attributed to the composition of the nutrient solutions. Nutrient imbalances, salinity, and the pathogen exposition did not cause severe damage, except for tip burn. Fresh weight was signiﬁcantly higher in HP (177.8 g) than in AE (107.0 g), while the chlorophyll and ﬂavonoid levels tended to be higher in AE. The leaf sodium and chlorine concentrations were higher than the values found in similar studies; however, AE plants contained a lower content of sodium and chlorine (35.0 and 21.5 mg/g dry weight) than the HP ones (44.6 and 28.6 mg/g dry weight). Many macro- and micronutrients in the AE-grown plants tended to be higher when the commercial PGPM or the sludges were administered, supporting the idea that those treatments contain a ﬂora that helps to extract nutrients from organic sources. The study demonstrated that lettuce can be successfully cultured at relatively high salt concentration. To further investigate beneﬁcial services such as nutrient extraction, salinity mitigation, and pathogen protection, we suggest administering bacterial communities of known composition, or single microbial strains. The study also showed that PGPM can be found in sludges of different origins; isolating beneﬁcial strains from sludge would additionally transform its management from a burdensome cost to a source of beneﬁcial services.


Introduction
Global food production will need to increase if human population growth is to be sustained, and one way to expand production is through sustainable intensification [1]. Soil-less agriculture technologies, such as hydroponics (HP), have many features that enable future challenges to be met, such as improved efficiency of water and fertiliser use, optimised growing conditions and consequently improved yield, decreased incidence of pathogens and pests, the potential for cultivation on degraded urban or arid soils, and avoidance of the negative aspects of monoculture, such as soil fatigue [2]. However, since HP normally relies on the use of inorganic fertilisers, research needs to address not only considering the usual concentrations in aquaculture farms, while the use of sludge can be seen as a way of harnessing the potentially beneficial flora against stressors, such as salt, nutrient scarcity, and pathogens; a previous investigation [21] showed that inoculates from aquaculture sludge can manifest biocontrol activity. In addition, the effect of these additives against the exposition to pathogen Pythium sp. was assessed to test the lettuces under a common disease in soilless systems.

Experimental Design
The experiment took place in summer 2021 at the Institute of Natural Resource Sciences (Zurich University of Applied Sciences, Wädenswil, Switzerland) in a climate chamber (Kälte 3000 AG, Landquart, Switzerland). Air temperature, humidity, and carbon dioxide were controlled, and full-spectrum light provided. Six lettuces (L. sativa L., variety Salanova; Rijk Zwaan Welver GmbH, Welver, Germany) per box (20 cm between plants, 25 plants/m 2 ) seeded in rockwool cubes were positioned on styrofoam rafts floating in 10 plastic boxes (40 × 60 × 12 cm). A full factorial experimental design (Figure 1) was applied with two different nutrient sources (S; levels HP and AE) and five different treatments (T; levels C, T1, T2, T3, T4), resulting in ten combined nutrient solutions corresponding to ten groups (HP-C, HP-T1, HP-T2, HP-T3, HP-T4, AE-C, AE-T1, AE-T2, AE-T3, AE-T4). The stock nutrient solutions and the treatments were: -AE: predominantly organic nutrient source functioning as a proxy for decoupled AP consisting of water from an aquaculture farm (Steinibach-Flühli, Switzerland) producing pikeperch (Sander lucioperca) ( Table S1). The AE was collected at the beginning of the experiment and stored at +7.5 ± 0.5 • C; -HP: inorganic nutrient source consisting of an aqueous solution of inorganic fertilisers (Table 1) emulating the nutrient levels of AE with pH adjusted to that of AE with HCl. HP was formulated at the beginning of the experiment and stored at +7.5 ± 0.5 • C; -T1: RhizoVital42 ® (Andermatt Biocontrol AG, Grossdietwil, Switzerland), which contains Bacillus velezensis (synonym B. amyloliquefaciens ssp. Plantarum), administered at a concentration of 0.04% v/v; -T2: sludge from the drum filter of the aquaculture farm (Steinibach-Flühli, Switzerland), administered at a concentration of 1% v/v, collected at the experiment start, and stored at +7.5 ± 0.5 • C. The dry matter content was 0.42%; -T3: sludge from a municipal wastewater treatment plant (Abwasserreinigungsanlagen, Au, Switzerland), administered at a concentration of 1% v/v, collected at the experiment start, and stored at +7.5 ± 0.5 • C. The dry matter content was 0.35%; -T4: activated effective microorganism BodenFIT ® (EM Schweiz AG, Arni, Switzerland), administered at a concentration of 0.1% v/v.
The application concentrations of T1 and T4 followed the manufacturers' recommendations, while the concentrations of T2 and T3 were selected according to the expertise of the Institute where the experiment took place.

Preparation of the Stock Solution
Before experiment start, the AE stock solution was analysed with Hach cuvette tests (LCK Küvetten-Test-System; Hach Lange GmbH, Berlin, Germany) for Ca 2+ , Mg 2+ , K + , PO 4 3− -P, NO 3 − -N, Cl − and water hardness (LCK 327, 327, 328, 350, 340, 311 and 327, respectively). As the aquaculture farm used NaCl to balance peaks of toxic N forms, the stock solution was diluted with demineralised water to make the electrical conductivity (EC) suitable for lettuce culture, reaching an EC of 4.2 mS/cm. Potassium (K) was added in AE as potassium hydroxide (KOH; Sigma-Aldrich Chemie GmbH, Buchs, Switzerland; 85%) to meet lettuce requirements. During the experiment, when the lettuce showed signs of chlorosis (Day 32), iron (Fe) was added as Fe EDTA (Ökohum GmbH, Herrenhof TG, Switzerland; 6.7%), in order to reach the same Fe EDTA concentration in both stock solutions (60 mg/L). Table 2 contains the nutrient concentrations of the stock solutions, sampled at seeding, and analysed by ion chromatography, as detailed in Section 2.4.  to grow Salanova lettuce (Lactuca sativa), 6 lettuce per box, 10 boxes for unexposed plants, and 10 boxes for plants exposed to the pathogen Pythium sp., 120 lettuces in total. Elements of the scheme were collected from disegnidacoloraregratis.it. Abbreviations: HP, stock solution consisting of demineralised freshwater mixed with inorganic fertilisers; AE, water from an aquaculture farm functioning as a proxy for decoupled aquaponics.

Plant Growth Conditions
The conditions in the climate chamber were set to 23/15 • C day/night, 65% relative humidity (RH), 16-8 h day-night cycle, 0.12% CO 2 concentration, full-spectrum light with 150 photosynthetic photon flux density (PPFD) during 26 days (corresponding to seedling culture and one additional day to ensure that plants could adapt to the new conditions after seedling culture), and with 250 PPFD after day 27, measured at plant level.
At the start of the experiment (day 1), two lettuce seeds were inserted into each rockwool cube in a tray with~1.5 cm demineralised water. On day 12, the weaker plantlet was removed, and 60 cubes were transferred to the 10 plastic boxes containing 15 L of a mixture of demineralised water and either one of the two nutrient solutions. To gradually adapt plants to the high EC of the stock nutrient solutions (Table 2), the ratio between nutrient solution and freshwater in the boxes was progressively increased until day 25 (Spreadsheet S1). On day 25 after seeding, the treatments were administered. On day 30 after seeding, a challenge test was started by placing 0.25 g millet, previously infected with Pythium sp., on each rockwool cube of one replicate of each treatment (60 rockwool cubes in total, 6 rockwool cubes per treatment). The actions carried out during lettuce culture are summarised in Table S4.
The solutions were aerated with an air stone to ensure that dissolved oxygen (DO) was above 7.14 mg/L; the water temperature fluctuated within 2 • C between days at the same time. A black-white plastic cover was used to shade the water in order to prevent algae formation.

Water Quality Monitoring
Temperature, DO, EC, pH, nutrient solution use, and nutrient concentration were monitored weekly (Table 3). Samples for nutrient concentration were collected weekly and determined at experiment end with a 930 Compact IC flex ion (Metrohm Schweiz AG, Zofingen, Switzerland). For anions, a Metrosep A Supp 5-250/4.0 column (Metrohm Schweiz AG, Switzerland) was injected with a sample volume ranging 20-250 µL and i.d.R. 100 µL, eluting with 3.2 mmol Na 2 CO 3 and 2 mmol NaHO 3 in H 2 O; for cations, a Metrohm C 6-150/4.0 column (Metrohm Schweiz AG, Switzerland) was injected with 20 µL of the sample, eluting with 7 mmol/L HNO 3 in H 2 O. The flow and temperature were set at 0.6 mL/min and at 30 • C, respectively, for both anions and cations. The weekly evapotranspired water was recorded and was replaced by demineralised water. In addition, half of the nutrient solution (7.5 L) from each box was replaced weekly with the stock nutrient solutions and the corresponding treatment.

Sampling and Chemical Analysis of Plants
A day before harvest, i.e., day 60 after seeding, a Dualex ® scientific leaf clip (ForceA; Orsay, France) [32] was used on 4 middle-aged leaves of 6 lettuces per treatment (120 measurements on 30 plants) to estimate chlorophyll, flavonol, and nitrogen contents. On day 61, the lettuces were harvested and shoots and roots were separated by cutting right above and right below the rockwool cube, respectively. Fresh shoots and roots of two plants per treatment were pooled and weighed, dried at 60 • C for 5 days, and weighed again to determine shoot and root dry matter contents. To analyse the chemical composition, root samples from each treatment were pooled. Each dried sample was ball-milled to a fine powder (3 pools of shoot samples per treatment; 1 pool of root samples per treatment), and 0.1 g of each sample was folded in aluminium foil for C, H, and N analysis in a TruSpec CHN Macro Analyser (LECO GmbH, Mönchengladbach, Germany). In parallel, aliquots of each powdered sample were pressed into tablets with Cereox wax (Fluxana ® GmbH&Co. KG, Bedburg-Hau, Germany) for X-ray fluorescence spectrometry with a SPECTRO XEPOS (Xepos GmbH, Kleve, Germany) [33].

Statistical Analysis
Data of unexposed and exposed plants were analysed separately using R software [34] to perform two-way ANOVAs with system (S; levels: HP, AE) and treatment (T; levels: C, T1, T2, T3) as factors with the car package. The estimated marginal means were calculated with the emmeans package. The superscripts, displayed only when interaction was significant, were calculated with the multcomp packages. Statistics was not computed on nutrient solution characteristics nor on root element content. The full model (type 3 ANOVA) function was: Y ij = µ + S i + T j + S i × T j + ε ij . When interaction was not significant, a reduced model (type 2 ANOVA) disregarding the interaction was applied as follows:

Results
The lettuce culture lasted 61 days. EC and pH of the nutrient solutions were higher at the end of the experiment than at the beginning. Nonetheless, the parameters of the nutrient solution did not cause severe stress on the plants, except for visible internal tip burn damages, which appeared around day 37 in lettuces grown in HP conditions and around day 48 in lettuces grown in AE conditions. No other major differences at the harvest were noticed, although fresh weight and the leaf nutrient concentration of some elements were partly modulated. On the other hand, growing lettuce in AE nutrient solution also led to increased chlorophyll and flavonoid levels according to Dualex results. The administration of potential bacteria showed intriguing effects that will be described and discussed in detail in the following sections. No severe Pythium sp. infection was noticed in the lettuces exposed to the pathogen.

Development of Nutrient Solution Characteristics during the Experiment
The nutrient solutions showed increased EC and pH from start to end ( Figure S1). For instance, in boxes of unexposed plants, pH slightly increased on average from 7.8 and 7.7 in HP and AE, respectively, to 7.9 and 7.9 in HP and AE. AE nutrient solutions tended to have lower values of both EC and pH. Single treatments followed the same trend, i.e., they increased from start to end. The exposition to the pathogen did not change the patterns observed.
On the other hand, the measured ions tended to decrease from the beginning to the end of the experiment. The concentration of each detected ion is reported in detail in Spreadsheet S1, while the present paragraph highlights the most important information related to the nutrient solutions of unexposed plants. Nitrate (NO 3 − ), phosphate (PO 4 2− ), and K + were consistently lower in both systems at the end of the experiment, irrespective of the treatment, to the extent of being below the detection limits of the chosen analytical method. In contrast, Na + , Ca 2+ , and Mg 2+ concentrations in the nutrient solutions of unexposed plants were generally lower in HP-C, HP-T4, and in all AE treatments except for T3, but higher in the remaining treatments, i.e., HP-T1, HP-T2, HP-T3, and AE-T3. Cl − and sulphate (SO 4 2− ) generally increased in HP groups and diminished in AE groups at the end when compared to the beginning of the experiment.
The nutrient solutions of the plants exposed to the pathogen showed a more consistent decrease in all ions.

Plant Growth and Composition
Shoot dry matter and fresh weight, root dry matter, root:shoot ratio, the chlorophyll and flavonoid levels, and nitrogen balance index (NBI) differed when taking the system factor into consideration (Tables 4 and 5). In detail, fresh weight was significantly higher in HP, while the dry matter for both shoots and roots was significantly higher in AE. The chlorophyll and flavonoid leaf levels tended to be higher in AE. No significant difference was observed on the NBI. The root:shoot ratio, calculated on fresh weight, was significantly higher in plants grown in the HP systems ( Figure 2). Furthermore, the five different treatments had different effects on the root:shoot ratio (Table 4), with the T1 and T2 treatments expressing the highest and lowest ratios, respectively. Notes: Different superscripts indicate significant differences for the system factor: ns, no statistically significant difference; *, p < 0.05; ***, p < 0.001. When the treatment factor is significant, please refer to Table S3 for the means of each single treatment. Abbreviations: SEM, standard error of the mean; HP, nutrient solution from inorganic nutrient source; AE, nutrient solution from aquaculture effluent from an aquaculture farm; C, control; T1, addition of RhizoVital42 ® (Andermatt Biocontrol AG, Switzerland); T2, addition of sludge from the aquaculture farm; T3, sludge from a wastewater treatment plant (Abwasserreinigungsanlagen, Au, Switzerland); T4, activated effective microorganism BodenFIT ® (EM Schweiz AG, Switzerland). DM, dry matter; FW, fresh weight; NBI, nitrogen balance index; T, treatment factor; S: system factor, <LOD: below level of detection. Note: It was not possible to use the full model function to perform ANOVA on the NBI parameter because the instrument was not able to take measurements on HP-T3 lettuces. The application of the reduced model showed no statistically significant differences for neither S nor T factors. Means of the instrument output are shown in the table in place of the estimated marginal means.  Treatment  C  T1  T2  T3  T4  C  T1  T2  T3  T4 Shoot DM, % 6.6 6.5 6.1 6. Notes: ns, no statistically significant difference; *, p < 0.05; **, p < 0.01; ***, p < 0.001. Abbreviations: see Table 4. The plants exposed to Pythium sp. (Table 5) showed the same trends as the une plants, while the treatment factor did not show any significant effect on the e plants. Nitrogen balance index was higher in the AE plants, with a different trend the unexposed plants.   Table 4. Bars indicate standard deviation. The plants exposed to Pythium sp. (Table 5) showed the same trends as the unexposed plants, while the treatment factor did not show any significant effect on the exposed plants. Nitrogen balance index was higher in the AE plants, with a different trend than in the unexposed plants.

23, 15, x FOR PEER REVIEW
When comparing the element composition, the system factor (HP vs. AE) influenced the content of most elements of the lettuces (Tables 6 and 7); the whole set of detected elements is shown in the Spreadsheet S2. The carbon content was consistently higher in the lettuces grown under AE conditions, irrespective of pathogen infection. The opposite was found for phosphorus, which was lower in AE, while the C/P ratio was higher in AE-grown plants (Figure 2). On the other hand, nitrogen content was not statistically different between HP and AE in unexposed plants, while in plants that underwent pathogen infection it was higher in AE than in HP. The C/N ratio was similar between the different groups, irrespective of pathogen infection. P, Ca, Mg, Na, Cl, Mn, Cu, and Zn were significantly more abundant in the shoots of plants grown in HP than in AE, whether exposed to the pathogen or not. K content was higher in the AE-grown plants, exposed or not. S content was significantly higher in the HP-grown unexposed plants than in the AE ones, while it was not statistically different between the HP and AE-grown exposed plants. Finally, no significant effect due to the system was noticed in Al and Fe contents of the plants exposed and not, although Fe contents tended to be higher in the AE-grown plants. Notes: Different superscripts indicate significant differences for the system factor: ns, no statistically significant difference; *, p < 0.05; **, p < 0.01; ***, p < 0.001. When the treatment factor is significant, please refer to Table S3 for the means of each single treatment. Abbreviations: see Table 4.
By inoculating the nutrient solutions with either the commercial products containing beneficial bacteria (T1 and T4) or the sludges (T2 and T3), the only elemental contents showing statistical differences were the ones of Mg, S, and Zn in the unexposed plants (Table S3). Specifically, T1 and T4 resulted in the lowest values for each of the three nutrients, while plants of the C and T2 showed the highest Mg and S contents and the T3 had the highest Zn content. In the plants exposed to the pathogen, the treatments had a significant effect on Cu shoot contents, with the highest values found in the T3 and the lowest found in the C and T4 (Table S4). Table 7. Average lettuce element concentrations (in mg per g of dry weight) in the shoots of exposed lettuces at harvest (day 61); values are means from three pools of two lettuces per treatment. HP  AE  ANOVA  SEM   Treatment  C  T1  T2  T3  T4  C  T1  T2  T3  Notes: Different superscripts indicate significant differences for the system factor: ns, no statistically significant difference; *, p < 0.05; **, p < 0.01; ***, p < 0.001. When the treatment factor is significant, please refer to Table S4 for the means of each single treatment. Abbreviations: see Table 4.

System
For the unexposed plants, the interaction between system and treatment factors showed statistically different values in P, K, Ca, S, Cl, Mn, and Cu contents. The HP-C plants showed the highest value of each element except Mn, which was highest in HP-T4 and second highest in HP-C. The lowest values of P, Ca, Cl, and Mn were found in the AE-C. The results of the K content of the unexposed plants showed that HP-C was the highest, which was statistically different from HP-T3 (the lowest value); the second highest K content was found in plants of the AE-T3. Mimicking the P content trend in the unexposed lettuces, an interaction between system and treatment was seen in the C/P ratio, in ascending order: HP-C, HP-T2, HP-T3, HP-T4, HP-T1, AE-T2, AE-T3, AE-T4,  AE-T1, AE-C. P, K, Ca, Mg, Mn, Fe, and Ni in the roots globally reflected the trend in the shoots, with P, Ca, Mg, and Mn higher in the HP plants, and K, Fe, and Ni higher in the AE plants. Na and Cl contents in the roots of unexposed plants were not significantly different between groups, while the roots of exposed lettuces globally showed higher Na and Cl in plants grown in AE in comparison to the ones grown in HP. Different patterns were observed concerning the S, Al, Si, and Cu root contents, which were generally lower in HP-grown than in AE-grown lettuces, irrespective of the Pythium infection. Considerably higher values in AE-grown plants were found in Al and Si contents, irrespective of the pathogen inoculation. The full list with elements detected in the roots is shown in Spreadsheet S3.

Discussion
The present study aimed to investigate whether nutrient origin and/or the administration of bacterial inocula of different origins could play a role in soil-less lettuce growth and element content. Plants grown in either HP or AE solutions were compared, where the HP solution was created by mimicking the pH, nutrient concentrations, and salinity of the AE solution. Obtaining the same nutrient concentration in both HP or AP is often difficult [8], and the authors of the present paper are aware of the limits of discussing the results within the context of the existing literature. Most of the differences in results between the plants grown in the HP and AE systems can be attributed to the intrinsic nutrient composition, as detailed in the following paragraphs. In addition, nutrients taken up by the plants disappear over time, while the ones that are not used accumulate, leading to increased EC of the nutrient solution. The data from this study followed this pattern: the nutrient solutions showed higher EC values at the end of the growth cycle, but with a decrease in the main macronutrients.

Impact of Inorganic or Organic Nutrient Sources on Lettuce Harvest
Lettuce growth was most likely affected by the compositions of the nutrient solutions, thus confirming the results of Yang and Kim (2020) [16], who stated that among the concurrent causes leading to differing growth results in HP-and AP-grown lettuces is the unbalanced nutrient composition and the chemical properties of aquaponic water.
The chlorophyll level was lower in HP-than in AE-grown lettuces, similar to what found by comparing tomatoes grown using a nutrient-film-technique in AP conditions to HP-grown tomatoes [35]. As chlorophyll is closely connected to photosynthesis rates [36], the lower chlorophyll level suggests a lower sugar content in lettuces grown in HP in the present study. However, despite the different chlorophyll levels between systems, the leaf nitrogen content and NBI were not significantly different between the differently grown unexposed plants, indicating that the nitrogen status of leaves was not affected. Further investigations should explain these contrasting results.
The leaf flavonoid level generally tended to be higher in AE plants in comparison to HP. Belonging to the polyphenol family, flavonoids have antioxidant activity and play a role in stress response; they protect from harmful radiations, bind phytotoxins, and contribute to regulating the stress response by controlling auxin transport [37]. The literature shows that higher flavonoid values could be a result of higher stress [38,39]. Further detailed analysis in this direction could explain if this was the case in the AE plants of the present study. On the other hand, flavonoids in the human diet are proved to be beneficial by in vitro, in vivo, and clinical studies, thanks to their potential anti-inflammatory, cholesterol-lowering, antidiabetic, and anti-cancer properties [14]. Besides, polyphenol consumption enables the settling of beneficial bacteria in the human intestine, with the production of beneficial metabolites [40].
Aquaponic systems commonly tend to accumulate zinc, though they do not accumulate iron, manganese, copper, and other micronutrients supplied by the fish feed. Nonetheless, deficiency symptoms are mainly detected only for iron [41]. The pH levels of the nutrient solutions of the present study, ranging 7.8-8.0, lowered the availability of phosphorus, manganese, aluminium, iron, copper, and zinc [31,41], and the results should be considered in this framework.
The higher percentage of shoot and root dry matter in AE in comparison to HP hints at a phosphate deficiency in AE. It is well known [42] that deficiencies of essential macronutrients (for instance nitrogen and phosphorus) generally modify the carbohydrate metabolism and increase the root:shoot ratio in plants. In lettuce, phosphate deficiency increases dry matter contents and non-structural carbohydrates while decreasing the nitrate content in the shoot [43]. On the other hand, the significantly lower root:shoot ratio in the AE system of the present study could be a response to potassium deficiency. As potassium is very mobile, increased root growth (expanding the explored volume of the growing medium) would not be beneficial [42]. The values of the nutrient solutions back up the hypothesised P and K deficiencies in AE-grown plants.
Magnesium and calcium leaf content showed an opposite trend to that of the chlorophyll, with lower values in plants grown in the AE system. Our finding agrees with Yang and Kim (2020) [16], who found that, in contrast to tomato and basil, the magnesium and calcium contents in lettuce leaves did not correlate with SPAD value, an index of chlorophyll content. In the present study, the inconsistency between high magnesium and calcium storage and low chlorophyll content in the HP groups suggests that magnesium in HP was probably diverted to other cellular components and vacuoles rather than used to build chlorophyll [44]. On the other hand, the lower values of magnesium and calcium in AE plants can be explained by the lower magnesium content in the AE nutrient solutions, aggravated by the high cation concentration (high salinity) that competes with magnesium and calcium uptake [45].
Leaf iron concentration did not differ between HP-and AE-grown plants, though iron is often the limiting nutrient in AP production [41,46]. Iron concentration in the AE used in the present study was well below the lettuce requirement and the element was therefore added in form of Fe EDTA to obtain a more balanced nutrient solution.
Similar to an earlier study comparing lettuces grown in HP and AP [16], plants grown in the AE system of the present study accumulated less phosphorus than the HP-grown plants, possibly as a result of the lower phosphorus concentration in AE nutrient solutions coupled with the fact that calcium and phosphorus easily form precipitates unavailable for plant uptake [46]. However, clear phosphorus deficiencies were not observed in our trials, confirming earlier findings on basil grown in HP saline (up to 1750 mg/L NaCl), phosphorus-poor (down to 3 mg/L) nutrient solution, positing that nutrient solution phosphorus can be managed independently from salinity [47]. Hence, the potential for using poor quality water (saline) with low phosphorus concentration, resulting in a high nutrient use efficiency for basil [47], could be investigated in lettuce with further in-depth analyses.
As it is common in AE [48], the sulphur concentration in the nutrient solutions of the present study was low in the AE system. Nonetheless, the shoot sulphur content was not drastically reduced by growing lettuces in the AE setting. Other studies grew lettuce in AP-derived water where sulphur concentration was much higher than the HP control water, and they showed that although the shoot of AP-grown lettuces retained more sulphur than the HP-grown ones, the difference in shoot sulphur content was small (0.02% on dry matter) [9]. In a study on AP-grown Cannabis sativa, leaf sulphur content increased following the addition of micronutrients (in mg/L: 1.4 Fe 3+ , 0.02 Cu 2+ , 0.26 B 3+ , 0.01 Mo 3+ , 0.4 Mn 2+ , and 0.08 Zn 2+ ) and of potassium (in mg/L: 75, 113, or 150), although the concentration of sulphur in the nutrient solution was unvaried [49]. It seems that shoot sulphur content is not highly susceptible to its own concentration in the nutrient solution, but rather to the interactions or synergies with potassium and micronutrients.

Role of Potentially Beneficial Bacteria in the Nutrient Solutions
Different sources of potentially beneficial microorganisms were administered as treatments in this experiment with the intent of screening for beneficial effects such as nutrient extraction [30], pathogen protection [50], and salinity mitigation [29]. Taking HP into consideration, it was noticed that most of the control shoots showed the highest values of mineral concentration. On the other hand, the control of the AE-grown plants often showed the lowest values. This asymmetry supports the idea that the addition of commercial PGPM or sludge, containing potentially beneficial flora, did not influence promptly available nutrients of inorganic fertilisers; however, it helped to extract nutrients from organic sources. More specifically, the AE control plants showed the lowest phosphorus, potassium, and manganese contents. Calcium, aluminium, and copper shoot contents showed the lowest values in control and T1 of the AE group; considering that T1 only contained a single bacterial strain (i.e., B. amyloliquefaciens ssp. plantarum), it is possible that the bacterium was not able to extract calcium, aluminium, and copper from the organic nutrients present in the AE nutrient solution, as the study of its genome did not find genes fulfilling such functions [51]. Finally, magnesium, sulphur, and zinc showed a similar asymmetrical trend, that is with plants belonging to AE of either T1 or T4 (i.e., containing fungi including yeasts and bacteria including actinomycetes, lactic acid and photosynthetic bacteria; [52,53]) showing the lowest contents, C plants of AE in the third lowest position, and C plants of HP with mostly the highest values. In contrast to our results, lettuce grown in deep-water culture in an AP system rearing red Mozambique tilapia (Oreochromis mossambicus) contained a higher amount of zinc when the water was administered with the commercial Bacillus mixture Sanolife ® PRO-W (INVE Technologies, Dendermonde, Belgium) [54]. In contraposition with our study, the genome of B. amyloliquefaciens contains a gene for zinc transport [51], and Bacillus spp. were able to solubilise zinc oxide [55]. Results similar to ours were reported by Iriti and colleagues (2019) [56], who showed that the zinc content of seeds from bean plants (Phaseolus vulgaris L.) grown in a greenhouse substrate was lower when treated with effective microorganisms, i.e., the commercial product administered in T4 treatment.
Although not statistically significant, growing under AE conditions seemed to impart a higher aluminium content than growing under HP conditions, although the high pH should exert a mitigating effect on the toxicity of aluminium. The higher aluminium content in AE plants can be explained by the possible presence of Myroides xuanwuensis, previously isolated from tilapia-rearing water [22], which was proved to release aluminium, silicon, and iron from biotite [57]. In the present study, the aluminium and silicon contents in the roots were more than two times higher in the AE-grown than in the HP-grown plants. In HP-cultured barley (Hordeum vulgare), silicon was shown to ameliorate the condition of the plants [58]; also, silicon and root exudate were shown to ameliorate aluminium toxicity in conventionally grown maize (Zea mays L.) [59]. This knowledge helps to explain why the higher aluminium content in AE did not cause plant toxicity in the present study.
It cannot be excluded that the nutrients dissolved in the relatively nutrient-rich HP solutions were partly chelated by the organic matter found in T2 and T3 (sludges), which accordingly diminished the shoot element content. In the relatively nutrient-poor AE system, the formation of organo-metallic complexes serves as a sink, slowly supplying nutrients when the chemical equilibrium is less favourable to sequestration [41,60]. Finally, within system, the minerals dissolved in the solutions varied in a treatment-specific manner; therefore, the treatments are thought to have brought nutrients to the solutions. However, the same trends were not found in shoots; for example, in HP, calcium in the nutrient solutions was highest in T3, followed by T2, T1, T4, and finally C, while the calcium content in the shoots went from C to T2, T1, T4, and finally T3. Such disparities between solution's nutrients and shoots' nutrients were also seen for other minerals and can be either explained by different chelation properties of the treatments, or by the different bacterial flora introduced with the treatment administration. The role of bacteria in modulating the element contents in the shoots deserves further investigation, for instance by inoculating microbial communities with a known composition or by performing simpler experiments where systems are treated with single bacterial strains. Also, the isolation of beneficial strains from sludge should be encouraged as to enlarge the range of available PGBM.

Salinity
The lower content of sodium and chloride in the AE plants in comparison to the HP plants contrasts with previous results comparing lettuce grown in HP and AP [16]. In absolute terms, the leaf sodium and chloride concentrations of the present study were much higher than the usual values [14] and can be explained by the high sodium and chloride content of the nutrient solutions, irrespective of the system. However, a daily portion of lettuce with the present sodium content would still be lower than the daily recommended intake of sodium (1.2-1.5 g/day, [61]).
So-called tip-burn appeared in all plants irrespective of system and treatment and should be discussed considering the salinity and the calcium concentrations of the nutrient solutions. Tip-burn appears when fast-growing parts of a plant (such as the growing tip in lettuce or the tip of the fruits in tomatoes) have little calcium supply. In addition, in the growing tip in lettuce, calcium can lack because of low transpiration. In the present study, low transpiration could be caused by a moderately high relative humidity coupled with little aeration (due to the conformation of the climate chamber used and to the leaves surrounding the growing tip, thus preventing good air movement), and by high Na and Cl concentration in the nutrient solutions, which exacerbates the situation as the ratios of Na to Ca and K may impair nutrient root uptake activity and translocation within the plant [62]. In summary, the observed tip-burn was likely a consequence of not only saline growing solution, but also of the lack of aeration around the fast-growing tip, and the impaired ratios between calcium and the other cationic macronutrients. In a well-balanced nutrient solution, it was demonstrated that lettuce do not undergo damage with a chloride concentration of up to 2000 mg/L [17], which is higher than the concentration of the nutrient solutions of the present study.
Despite the elemental deficiencies in the AE nutrient solutions, the lettuce of the present study did not show major salinity stress injuries except for the tip-burn, and the overall growth of the plants was not hindered. Besides, neither of the four treatments triggered a marked positive response against saline stress or nutrient unbalance. The hypothesis that beneficial bacteria alleviate saline stress [63,64] remained unanswered in HP and AP settings [65]. However, it was recently shown that silicon application and plant growth promoting rhizobacteria consisting of six pure Bacillus species, singly or in combination, alleviate salinity stress in cucumber (Cucumis sativus L.) grown in a deep-water culture [66]. As previously mentioned, the administration of the four bacterial treatments in the present study brought to the growing systems not only a microbial component but possibly also organic compounds and inorganic nutrients (most likely T2 and T3). For this reason, it was not possible to discern the items leading to specific changes in lettuce harvest outcome.

System and Treatment Effects on Lettuces Exposed to Pythium sp.
Microorganisms of aquaculture origin have been found to suppress plant diseases under soil-less circumstances [21,23,[29][30][31] but the extent of the beneficial effects on lettuce have not been thoroughly reported. With an increase of Pythium aphanidermatum density, photosynthesis, transpiration, and dry mass of all components of tomato plants (Lycopersicon lycopersicum L. cv. "Counter") in deep water culture were reduced at 30 • C, but not at lower temperatures [67]. In the present study, the plants did not show clear symptoms of disease, not even in the roots. Similar to the unexposed plants, root dry matter of the exposed plants was statistically lower in the HP group. The root:shoot ratio also aligned with the trend found in the unexposed plants and reinforces the hypothesis that inoculating the rockwool cubes with 0.25 g millet previously exposed to Pythium sp. did not lead to disease.
Subtle variations between the plants cultured in the two systems were observed in leaf nitrogen content, which was higher in the AE plants. In a comparison between tomato plants exposed or not to P. aphanidermatum, leaf nitrogen and potassium concentration decreased in the presence of the pathogen at high temperatures, and the authors blamed the limited nutrient uptake of severely infected roots, suggesting that the decreased nitrogen content and the reduced stomata conductance were the possible mechanisms that could have limited leaf photosynthesis [67]. However, no damage was observed at 20 • C, except for transpiration, leading to the conclusion that tomato plants tolerate infections of P. aphanidermatum below 30 • C at the root zone. The reason why nitrogen was higher in AE plants although nitrogen concentration in AE nutrient solution was lower could be ascribed to a beneficial role played by microorganisms present in the AE.
Although no statistics were computed between exposed and unexposed plants, by looking at the data it seems that no differences were revealed. A possible cause is the relatively low water temperature, in agreement with the study on tomato [67] and according to the general knowledge that maintaining temperatures below 28-30 • C is essential in order to limit the exponential germination of Pythium spores [31]. In addition, the high pH of the present study could have prevented disease development, as it has been demonstrated in soil-systems [68].

Conclusions
The present study presents a first insight into lettuce grown in multiple stressful conditions, including nutrient imbalance, high salinity, and pathogen Pythium sp. exposure. Despite the relatively low number of replicates due to the preliminary nature of the experiment, we successfully showed that the harvest of lettuces grown in either HP-like or AP-like systems was not jeopardised when taking dry matter in consideration, and the electrical conductivity of 4.3-5.0 mS/cm did not cause severe stress to the plants. The fresh weight and the leaf nutrient concentration of some elements were influenced by the different systems. For instance, phosphorus was lower in AE lettuces; however, this did not cause clear deficiencies, suggesting the need to further investigate the potential of using poor quality water (saline) with low phosphorus concentration. Growing lettuces in the AE nutrient solution also led to increased chlorophyll and flavonoid content. The increased nitrogen leaf content of the exposed AE group hints at a possible beneficial effect of microorganisms of aquaculture origin and further studies are encouraged. Finally, future cost analyses would assess the economic feasibility of growing vegetables in nutrient poor/salt-rich nutrient solutions.
The addition of beneficial bacteria to the nutrient solutions is a possible practice to obtain microbiome services such as nutrient extraction, salinity mitigation, and pathogen protection, and was successful in the present study to increase the contents of some nutrients in the AE-grown lettuces, showing its extraction properties from organic sources. In the future, knowing the composition of the administered microbial communities will be important for predicting their effects, with the aim of modulating the microbiota to suit the farm's purpose of producing more yield with less input and less impact on the environment. Finally, future studies on the administration of microbial inocula investigating simpler systems treated with single bacterial strains are to be encouraged. Also, the isolation of beneficial strains from sludge would transform its management from a burdensome cost for aquaculture production into a source of beneficial organisms, with the overarching aim of closing the production loop.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/w15112109/s1, Figure S1: Parameters of the nutrient solutions in boxes of lettuce unexposed (A, B) and exposed (C, D) to pathogen Pythium sp., one day after microbial inoculates addition (day 26) and one day before harvest (day 60); Table S1: Technical details of the pikeperch (Sander lucioperca) aquaculture farm whose effluent water was used as nutrient solution containing organic nutrient sources and functioning as a proxy for decoupled aquaponics; Table S2: Actions carried out during lettuce culture; Table S3: Average element concentrations in the shoots of unexposed lettuces at harvest (Day 61); values are means for each treatment, irrespective of the system; the p-values were calculated with a type 2 ANOVA following the reduced model disregarding the interaction between the two factors; Table S4: Average copper concentrations in the shoots of lettuces exposed to the pathogen Pythium sp. at harvest (Day 61); values are means for each treatment, irrespective of the system; the p-values were calculated with a type 2 ANOVA following the reduced model disregarding the interaction between the two factors; Spreadsheet S1: Nutrient solution values during the adaptation phase; Spreadsheet S2: Average element concentrations in the shoots lettuces at harvest (Day 61); values are means from three pools of two lettuces per treatment; Spreadsheet S3: Average element concentrations in the roots of lettuce at harvest (Day 61); values are means from one pool of six lettuces for each treatment. Funding: This work was supported by the COST Action CA17133 "Implementing nature-based solutions for creating a resourceful circular city" and was conducted during a Short-Term Scientific Mission (STSM). The authors acknowledge the financial support of the Finanziamento di Azioni di Internazionalizzazione issued by University of Florence.
Data Availability Statement: Not applicable.