Differential ecophysiological responses to inorganic nitrogen sources (ammonium versus nitrate) and light levels in the seagrass Zostera noltei

: Seagrasses can use both ammonium (NH 4+ ) and nitrate (NO 3− ) as inorganic nitrogen (N) sources. However, NO 3− uptake and assimilation are energetically more expensive and tightly regulated than NH 4+ uptake. The objective of this study was to test the complex interactive effects between different forms of N enrichment (NH 4+ and NO 3− ) and light levels on the morphological and physiological traits in the intertidal seagrass Zostera noltei . Plants were cultured over 40 d under 2 levels of light (low and high) with 2 inorganic N concentrations supplied at the same dose, NO 3− (25 μM) and NH 4+ (25 μM), and a control, following a 2-factorial design. Results showed a differential response in Z. noltei depending on the inorganic N source and light dose. NH 4+ enrichment negatively affected almost all morphometric and dynamic variables analyzed, both in isolation and combined with low light conditions. In contrast, NO 3− enrichment had a positive effect on Z. noltei survival compared with the control treatment in terms of net growth rate and rhizomatic growth, mainly under high light conditions. Therefore, our study demonstrated that the effects promoted by nutrient enrichment largely depend on the source of N used. Light levels play a crucial role in this response by potentially shifting the effects from toxic (under low light) to beneficial (under high light) when NO 3− is the main N source. Our findings highlight that N form in eutrophication events should be considered when evaluating the potential impacts of nutrient enrichment and light reduction on seagrass communities.


INTRODUCTION
Seagrasses are coastal foundation species that are among the most productive coastal habitats. They provide a wide range of ecosystem services such as carbon (C) burial, amelioration of natural hazards and habitat and nursery functions (Nordlund et al. 2018). They sit between the land and the sea; therefore, the increase in human population density in coastal zones favors an increase in nutrient loads de -rived from watersheds and sewage and agricultural runoff, thereby driving eutrophication processes (Vitousek et al. 1997, Verhoeven et al. 2006. Eutrophi cation negatively affects seagrass ecosystems (Waycott et al. 2009) by both decreasing light levels and increasing the concentrations of dissolved inorganic nitrogen (DIN) (Touchette & Burkholder 2000). DIN usually reaches estuaries in the form of nitrate (NO 3 − ) (Weller & Jordan 2020), with ammonium (NH 4 + ) making up less than 10% of the DIN in these ABSTRACT: Seagrasses can use both ammonium (NH 4 + ) and nitrate (NO 3 − ) as inorganic nitrogen (N) sources. However, NO 3 − uptake and assimilation are energetically more expensive and tightly regulated than NH 4 + uptake. The objective of this study was to test the complex interactive effects between different forms of N enrichment  ) and light levels on the morphological and physiological traits in the intertidal seagrass Zostera noltei. Plants were cultured over 40 d under 2 levels of light (low and high) with 2 inorganic N concentrations supplied at the same dose, NO 3 − (25 μM) and NH 4 + (25 μM), and a control, following a 2-factorial design. Results showed a differential response in Z. noltei depending on the inorganic N source and light dose. NH 4 + enrichment negatively affected almost all morphometric and dynamic variables analyzed, both in isolation and combined with low light conditions. In contrast, NO 3 − enrichment had a positive effect on Z. noltei survival compared with the control treatment in terms of net growth rate and rhizomatic growth, mainly under high light conditions. Therefore, our study demonstrated that the effects promoted by nutrient enrichment largely depend on the source of N used. Light levels play a crucial role in this response by potentially shifting the effects from toxic (under low light) to beneficial (under high light) when NO 3 − is the main N source. Our findings highlight that N form in eutrophication events should be considered when evaluating the potential impacts of nutrient enrichment and light reduction on seagrass communities.
KEY WORDS: Seagrass · Nitrogen metabolism · Ammonium · Nitrate · Eutrophication · Toxicity · Dissolved inorganic nitrogen · Light intensity OPEN PEN ACCESS CCESS discharges (Ward et al. 2011). However, the load of NH 4 + in coastal areas has increased worldwide in the last decade (Glibert et al. 2010, Malone & Newton 2020, Peñuelas & Sardans 2022. Seagrasses typically exhibit higher uptake rates of NH 4 + than NO 3 − (Lee & Dunton 1999, Dudley et al. 2001, primarily because NH 4 + assimilation is energetically less costly (Turpin 1991). Previous studies have indicated that a moderate increase in NH 4 + availability (<10 μM) stimulates seagrass growth and biomass when seagrasses grow under nutrient-limited conditions (e.g. Orth 1977, Alcoverro et al. 1997, Peralta et al. 2003, Invers et al. 2004). However, several studies have also demonstrated that high concentrations of NH 4 + (~25 μM) can be toxic to some seagrass species in the presence of low light (LL) levels, phosphate deficiency, alkaline pH, high temperature and/or high salinity, among other factors (e.g. Burkholder et al. 1992, van Katwijk et al. 1997, Brun et al. 2002, van der Heide et al. 2008, Christianen et al. 2011, Villazán et al. 2013. The negative effects of high NH 4 + concentrations on seagrasses have traditionally been explained by intracellular accumulation of NH 4 + , which can affect internal pH and enzyme kinetics, uncouple photosynthetic ATP production, increase respiration and decrease the up take of other cations (e.g. Marschner 1995). In addition, continued uptake and assimilation of NH 4 + can deplete C reserves and thus compete with other C-demanding or energy-consuming metabolic pathways. For example, carbohydrate reserves (mainly in the form of sucrose and starch) in the seagrass Zos te ra noltei have been reported to be crucial for avoiding NH 4 + toxicity. If internal carbohydrate reserves (mainly sucrose) fall below a critical level, NH 4 + can become toxic because NH 4 + is not assimilated into amino acids as a result of the limited available C skeletons (Brun et al. 2002). In the case of NO 3 − , toxicity effects have rarely been described (but see Burkholder et al. 1992Burkholder et al. , 1994, possibly because of the tight interdependence between nitrogen (N) and C metabolism within plants, which require a continual supply of energy and C skeletons for NO 3 − assimilation and a partitioning of photosynthetic products among carbohydrate synthesis, amino acid synthesis and other plant functions (Huppe & Turpin 1994, Foyer et al. 2001, Stitt et al. 2002. Therefore, different effects have been recorded in seagrasses ex posed to high levels of NO 3 − , such as decreases in their C reserves (Jiang et al. 2013), an increased rate of Labyrinthula zosterae in fection in the presence of the herbicide Diuron (Hughes et al. 2018) or de creased shoot survival under diminished light (Burkholder 2000).
Beyond the likely direct toxic effects of nutrients on seagrasses, a common indirect phenomenon during eutrophication events is diminished light resulting from the proliferation of epiphytes and macroalgae in seagrass communities. Many studies have examined the responses of seagrasses to diminished light (Brun et al. 2008, Collier et al. 2009, Christianen et al. 2011, Serrano et al. 2011 and their subsequent recovery dynamics , Bité et al. 2007, Biber et al. 2009, Collier et al. 2009). During periods of depressed photosynthesis caused by light limitation, seagrasses mobilize stored non-structural carbohydrates (NSCs) to maintain metabolic processes (Alcoverro et al. 1999). Shading-induced NSC depletion may modify the responses of seagrasses to other environmental stresses, such as high levels of N, because NSC reserves play an important role in determining seagrass growth ).
To our knowledge, a direct comparison of the 2 common sources of DIN (oxidized vs. reduced) applied at the same dose to a seagrass community is lacking in the literature. This study aimed to fill this research gap by exploring the complex interactive effects between DIN supply in different forms (NH 4 + and NO 3 − ), both alone and combined with light availability (2 contrasting light intensities), on the ecophysiological responses of the intertidal seagrass Z. noltei. We examined the effects of light levels and DIN sources on morphometric (i.e. above-[AG] and belowground [BG] biomass and leaf and root lengths [L L and L R ]), dynamic (i.e. survival, net growth rate [NGR], shoot and internode appearance rates [SAR and IAR], rhizomatic growth rate [RGR]) and physiological traits (i.e. internal N, C and NSC reserves). The fast-growing Z. noltei was used as a model species because it is widely distributed along the coasts of the Atlantic Ocean (Green & Short 2003), in areas usually subjected to high nutrient levels that are exhibiting declining seagrass population trends (Short et al. 2011), necessitating protection and monitoring. Although the effects of nutrient enrichment and diminished light on this species have been investigated (e.g. Brun et al. 2002, Cabaço et al. 2013, Villazán et al. 2013, most studies have focused on NH 4 + enrichment, while less attention has been paid to NO 3 − (e.g. uptake processes; Alexandre et al. 2011). We hypothesized that both forms of DIN would have negative effects on this species under LL conditions by increasing the demand on C reserves, whereas we expected to observe positive effects for both DIN forms in plants under high light (HL) conditions.

Experimental setup
A 2-factorial experiment was conducted at an indoor mesocosm system at the Faculty of Marine and Environmental Sciences of the University of Cádiz in the spring (from March to April). Healthy appearing shoots with intact rhizomes of Zosters noltei were collected from an intertidal seagrass meadow at Santibáñez (Cádiz Bay Natural Park; 36.47°N, 6.25°W, Cádiz, Southern Spain), transported to the laboratory and kept in aerated seawater under saturating light (~231 μmol photons m −2 s −1 ; Peralta et al. 2002) in a 16 h light:8 h dark cycle at 15°C for 2 d before the experiment. Apical shoots formed by 2 rhizome internodes with one apical shoot and one lateral shoot, joined to the associated roots, were selected as the ex perimental plant unit (EPU). Before transplantation, epiphytes were wiped away from shoots with a soft tissue paper.
Aquaria were illuminated by lamps with cool fluorescent tubes (T5 High Output Blau Aquaristic aquarium color extreme fluorescent bulbs) in a 16 h light: 8 h dark cycle. The water temperature was kept constant at 17°C to achieve optimal growth (Nejrup & Pedersen 2008). Next, 36 EPUs were planted in each of the 18 aquaria (n = 648 EPUs), which were then provided with either 25 μM NH 4 + , 25 μM NO 3 − or no inorganic N (as a control) and exposed to 2 contrasting light levels, corresponding to sub-saturating (LL: 52 ± 5.01 mol photons m −2 s −1 ) and saturating (HL: 262 ± 13.5 mol photons m −2 s −1 ) light conditions for this spe-  . NH 4 + and NO 3 − were added from a stock solution to each aquarium (25 μM treatments) as a daily pulse (375 μmol of both N forms). The NH 4 + concentration was chosen because concentrations above 25 μM are known to be harmful to Z. noltei (Brun et al. 2002(Brun et al. , 2008. The same concentration of 25 μM of NO 3 − was also selected so that the N load throughout the experiment was equal to the treatments with NH 4 + . Seawater samples were collected from each aquarium and filtered through Whatman GF/F filters (0.7 μm) before and 10 min after NH 4 + / NO 3 − addition and then were immediately frozen at −20°C for further analysis. Water sampling for analyses was re peated 3 times wk −1 , and physico-chemical parameters (i.e. light, temperature, salinity and pH) were monitored on Days 0, 2, 5 and 7 each week during the experiment (40 d). Water in all aquaria was renewed weekly (approximately every 6−7 d) to prevent any excessive accumulation of NH 4 + or NO 3 − . During water renewal, the aquarium walls were cleaned with soft tissues to remove salt and epiphytes and floating (detached) seagrass leaves. Before and after water renewal, water samples were collected to calculate the nutrient accumulation rate throughout the incubation period. The mean net N up take rates of NO 3 − and NH 4 + (μmol N gWW −1 d −1 ) ( Fig. S1 in the Supplement at www.int-res.com/articles/suppl/m702 p057_supp.pdf) were estimated on DIN supply treatments (i.e. LL + NH 4 + , LL + NO 3 − , HL + NH 4 + , HL + NO 3 − ) among periods of seawater renewal over the course of the experiment (40 d), based on the total DIN added during the time interval (25 μM DIN multiplied by the number of times DIN was supplied, n), the aquarium volume (V = 15 l) and the amount of DIN before water renewal (DIN ren concentration, μM multiplied by V ), divided by the seagrass biomass in the aquarium and by the elapsed time (t) between water renewals. Moreover, aquaria positions were randomly interchanged at each renewal period (i.e. weekly) to minimize the effects of any slight differences in experimental conditions among the treatments (e.g. light or aeration). The analytical methods used to determine NH 4 + and NO 3 − were essentially an adaptation of the spectro photometric methods described by Hansen & Koroleff (1999) using a UNICAM UV-1700 Pharma Spec spectrophotometer. NH 4 + was determined on the basis of the reaction of 120 μl of salicytate-catalyst (reagent A), and 200 μl of nitratehypochlorite as a catalyzer (mixing alkaline-citrate and Na-hypo chlorite 10%), measured at 64 nm (Bower & Holm-Hansen 1980). NO 3 − measurements were based on the colorimetric measurement of nitrite formed after the re duc tion of NO 3 − by nitrate reduc-tase. The NO 3 − produced was measured spectrophotometrically (540 nm) after the addition of 600 μl of vanadium chloride plus 150 μl of reagent mix (sulfanilamide acid and N-1-naphthylethylenediamine dihydrochloride) to 750 μl of samples (Schnetger & Lehners 2014). In both cases, standard curves were constructed according to the same procedures with known concentrations of NH 4 + and NO 3 .

Biological measurements
At the beginning of the experiment, morphometric measurements (L L , number of leaves and AG and BG biomasses) were conducted on 10 EPUs randomly selected from the pool of collected plants. Before transplantation into aquaria, each EPU was weighed (initial wet weight, WW) and each rhizome was individually tagged with a label. Then each EPU was weighted (g WW), and the initial number of leaves per plant was recorded. At the end of the experiment, all surviving plants were carefully harvested and weighed (mg WW) to estimate the net growth production per EPU (mg WW EPU −1 d −1 ) from the net change in individual plant weight during the experiment. At the end of the experiment, morphometric measurements were collected from all plants (L L , L R and internode abundance). In addition, each harvested EPU was split into leaves (AG) and rhizomes/ roots (BG), freeze-dried and weighed to determine the AG:BG ratio. According to procedures described in Peralta et al. (2006) and de los Santos et al. (2010) ( Table 1), morphometric information was used to calculate plant dynamic properties (survival, NGR, SAR and IAR, RGR), to estimate the growth of the plants over the duration of the experiment (40 d).

Physiological traits
The concentrations of NSCs (i.e. sucrose and starch) were measured in leaf and rhizome samples (n = 6) from each aquarium at the end of the experimental period. Samples were freeze-dried and ground before analysis. Total NSCs were measured according to Brun et al. (2002). Sugars (sucrose and starch) were first solubilized by 4 sequential extractions in 96% (v/v) ethanol at 80°C for 15 min. The ethanol extracts were evaporated under a stream of air at 40°C, and the residues were then dissolved in 10 ml of deionized water for analysis. Starch was extracted from the ethanol-insoluble residue by incubation for 24 h in 1 N NaOH. The sucrose and starch content were determined spectrophotometrically with a resorcinol and anthrone assay with absorbances of 486 and 640 nm, respectively, with sucrose as the standard. NSC concentration was calculated as the sum of AG and BG sucrose and starch in each plant (Alcoverro et al. 1999). Total C and N content was determined in duplicate freeze-dried, ground samples of leaves and roots/rhizomes from each aquarium with a Perkin-Elmer 2400 elemental analyzer.

Statistical analyses
Before any statistical analysis, data were verified for normality (Shapiro-Wilk normality test) and homoscedasticity (Bartlett test for homogeneity of variance test). Repeated measures ANOVA was used to test whether light, salinity and pH at the end of each week varied over the course of the experiment and between treatments. We used 2-factorial permutational multivariate analysis of variance (PERM-ANOVA) to test the overall effects of N (control, NH 4 + enrichment and NO 3 − enrichment) and light conditions (LL vs. HL) on morphometric and dynamic variables (i.e. AG:BG ratio, L L , L R , internode abundance, survival, NGR, SAR, IAR and rhizome growth rate). The multivariate approach was chosen because some of the measured response variables were likely to be correlated. To test the effects of the treatment factors on each response variable more specifically, after the multivariate analyses we performed univariate PERM ANOVA (2-or 3-factorial), as suggested by Quinn & Keough (2002). A 2-way ANOVA was used for sucrose, starch and total N and C content in rhizomes and roots. When ANOVA assumptions were not satisfied (i.e. N and C in leaves), a non-parametric comparison (Kruskal-Wallis matched pairs test) was applied to assess statistically significant differences. When significant differences were found, the Tukey post hoc test was applied to compare both the levels and interaction factors. Data are presented as means ± SE. The significance level (α) was set at 0.05.
To evaluate the additive, synergistic or antagonistic effects of the significant interactions that arose under 40 d of stressors, we compared the observed re sponses to pairs of stressors with an additive null model (Darling & Côtê 2008). We tested whether the effects of combined stress imposed by LL, NH 4 + enrichment and NO 3 − enrichment were either additive or non-additive (i.e. synergistic or antagonistic) by using the relative response ratios (RR) for each variable in the following equation: where 'stress treatment' is the measured mean response for each stress treatment (i.e. LL, NH 4 + or NO 3 − enrichment, and combinations of these treatments), and 'non-stressed' represents the control con ditions (i.e. HL, N control). We used an additive null model as the expected additive response (  We then compared the observed combined response and the expected additive response. If the observed combined response was less than the expected additive response, the effect was classified as antagonistic. Otherwise, if the observed combined re sponse was greater than the expected additive response, the effect was classified as synergistic. If the observed combined response overlapped with the ex pected additive response, the effect was classified as additive. Statistical analyses were performed in R statistical software v.4.0.2 (R Core Team 2019).

Physico-chemical traits
The temperature in the seawater averaged 16.8 ± 0.06°C, pH averaged 8.04 ± 0.12 and salinity averaged 35.61 ± 0.17 across all treatment combinations and sampling days. Repeated measures ANOVA did not reveal any significant variations in these physicochemical variables over time or across treatments (all p > 0.39). The NH 4 + and NO 3 − concentrations in the water after enrichment (i.e. 10 min after 25 μM addition) differed considerably depending on the treatment (Table S1) and averaged 0 μM in treatments without N addition (data not shown). NH 4 + and NO 3 − accumulated in the seawater, particularly under LL. The LL + NH 4 + treatment, in comparison with the other treatments, showed a continual decrease in NH 4 + uptake capacity ( Fig. S1; ANOVA, p < 0.001). The net uptake rate represented a mean value of 15 ± 5% of the DIN added in the LL + NH 4 + treatment, whereas the rest of the treatments averaged 65−75%, thus indicating accumulation of NH 4 + in the LL + NH 4 + treatment over time (Fig. S1).

Morphometric and dynamic traits
The multivariate response of all morphometric and dynamic variables was affected by both N forms (i.e. NH 4 + and NO 3 − ) and light conditions (LL vs. HL) (Fig. 2, Table S2). However, no significant differences were observed in the interaction between both factors (N and light), except for the maximum leaf length (L Lmax ).
L Lmax was significantly affected by the interaction be tween N and light. Plants growing in LL were longer than those growing in HL under control and NO 3 − treatments (PERMANOVA, F 1, 2 = 5.30, p = 0.002 and F 1, 2 = 3.27, p = 0.001; Fig. 2B). In contrast, this pattern was reversed under NH 4 + loading: L Lmax was lower under the LL + NH 4 + than HL + NH 4 + treatment (PERMANOVA, F 1, 2 = 4.33, p = 0.03). Similar L Lmax values were observed in HL for both N treatments. An inverse pattern was found in the maximum root length (L Rmax ); significant differences were detected in HL, with L Rmax decreasing under NO 3 − and NH 4 + enriched treatments compared with the control (PERMANOVA, F 1, 2 = 2.56, p = 0.002 and F 1, 2 = 3.89, p = 0.001; Fig. 2C). Survival increased significantly from 70% (control treatment) to 80% with NO 3 − enrichment in HL (PERMANOVA, F 1, 2 = 3.03, p = 0.002). However, it was approximately 60% lower in the LL + NH 4 + than the LL + control treatment (PERMANOVA, F 1, 2 = 0.79, p < 0.001; Fig. 2D). This decrease was more pronounced in LL for the control vs. NH 4 + enriched treatment, and survival reached values near 20% in the latter (Fig. 2D). Moreover, in LL, the addition of NH 4 + resulted in a negative NGR (−1.50 mg WW d −1 ) compared to those in the HL treatment (5 mg WW d −1 ). NGR was significantly higher in HL under NO 3 − enrichment than in the control treatment (PERMANOVA, F 1, 2 = 0.98, p = 0.029; Fig. 2E). A similar pattern was found for the SAR (Fig. 2F). However, NH 4 + affected the SAR negatively compared to NO 3 − and the control treatment, and the lowest values were observed in LL. The IAR was lower (no significant difference) under the NO 3 − and NH 4 + treatments than the control treatment in LL. RGR was significantly higher under NO 3 − enrichment than in the control treatment (PERMANOVA, F 1, 2 = 2.29, p = 0.04) but was slightly lower with NH 4 + treatment in HL (PERMANOVA, F 1, 2 = 1.67, p = 0.022; Fig. 2G,H).

Physiological traits
AG and BG N content was affected by both factors (N and light) and by their interactions. Foliar N was significantly higher under NH 4 + load than under control and NO 3 − treatments (Kruskal-Wallis test, χ 2 1, 5 = 3.056, p = 0.031 and χ 2 1, 5 = 1.089, p = 0.023). The N content in rhizomes and roots differed between light treatments and was higher under LL, except in the NH 4 + treatments, which showed an inverse pattern. The highest N content in rhizomes and roots was found for the HL + NH 4 + treatment compared with the control and NO 3 − treatments (2-way ANOVA in HL, F 1, 2 = 11.16, p = 0.001 and F 1, 2 = 9.43, p = 0.03; Fig. 3B, Table S3). The foliar C content was influenced by light treatments and was significantly lower in LL under control and nutrient-enrichment treatments (Fig. 3C). However, the C content in rhizomes and roots was not influenced by light or N supply (Fig. 3D).
The sucrose/starch concentrations responded negatively to NH 4 + enrichment and LL (Fig. 4, Table S3) compared with the controls, but no differences were observed under NO 3 − enrichment (except for a starch increase in rhizomes/roots). Sucrose content (in leaves and rhizome/root parts) was substantially lower under NH 4 + load and showed the largest decrease in LL, with a foliar sucrose content 40% lower than that in the LL + NO 3 − treatment (2-way ANOVA, F 1, 2 = 1.04, p = 0.0012; Fig. 4A). The content of starch was lower in rhizomes and roots than leaves, and NH 4 + enrichment resulted in a significantly lower overall starch content than did control and NO 3 − treatments. Plants cultivated under NO 3 − enrichment had significantly higher content of starch in rhizomes and roots than those under the control treatment (2-way ANOVA, F 1, 2 = 0.89, p = 0.010; Fig. 4D) and NH 4 + treatments (2-way ANOVA, F 1, 2 = 1.45, p = 0.031; Fig. 4D) under HL conditions.
The combined effects of light and NH 4 + enrichment on the morphometric and physiological responses did not generally differ from the expected additive effects, except in the case of the sucrose content of leaves (Table 2,  at physiological levels were also additive, but the starch content in rhizome/root parts showed an antagonistic effect because the combined effect was lower than the expected additive effect (Table 2, Fig. 3D).

DISCUSSION
As expected, our experiments indicated that diminished light decreased biomass and negatively affected most dynamic parameters in Zostera noltei. The ecophysiological responses of Z noltei were conditioned by the nature of each DIN form supplied: negative effects were observed with NH 4 + and neutral and/or slightly positive effects were observed with NO 3 − . In addition, light levels boosted these responses. Therefore, our initial assumption that NO 3 − might have a negative effect on this species was not supported. Instead, a dual behavior of NH 4 + (i.e. as a nutrient and as a toxic element) was demonstrated and was found to be highly dependent on light levels. The large decrease in AG biomass and SAR observed under LL conditions indicated the high sensitivity of Z. noltei. Decreases in AG biomass and shoot density under limited light conditions have often been reported for Z. noltei and other seagrass species ); it has been described as a plasticity mechanism to maximize available understory light by decreasing self-shading in the population (Collier et al. 2012). Meanwhile, leaves were longer under LL than HL con ditions. This morphological plasticity is also a well-described mechanism in seagrasses and land plants that substantially enhances light harvesting under LL conditions (Erftemeijer & Stapel 1999,  Peralta et al. 2002). The diminished RGR and branching frequency observed in this work and in other studies (Bulthuis 1983, Abal et al. 1994, Gordon et al. 1994, Vermaat & Verhagen 1996, Krause-Jensen et al. 2000 under LL conditions may also explain the observed decrease in SAR observed in our light treatments, as shoot appearance is mainly affected by the growth of the apical shoot in this species (Peralta et al. 2006). Regarding internal composition, the N content in leaves was higher than that in BG tissues (rhizomes and roots) regardless of light level, whereas the N content in both tissues decreased with increasing light levels. Similar responses have been observed in other studies with Z. noltei (Pérez-Lloréns & Niell 1993, Vermaat & Verhagen 1996. This observation may be explained by dilution processes (Stocker 1980): when N utilization is faster than uptake, stored N resources are gradually diluted during growth.
Meanwhile, lower C content in leaves under LL conditions is in concordance with the observed decrease in NSCs, because under such conditions, NSCs are mobilized to meet respiratory demands and balance C budgets (Kraemer & Alberte 1993, Zimmerman & Alberte 1996, Lee & Dunton 1997, Brun et al. 2003a. Although the effects of DIN enrichment on seagrasses are well documented, potential differences in ecophysiological responses associated with DIN forms are frequently overlooked. In our study, a significantly higher NGR was observed under NO 3 − enrichment than in the control treatment, independent of light conditions, and survival and SAR of Z. noltei were not compromised in these treatments -a finding opposite from our hypothesis. This result indicates that under our control conditions, experimental plants were nutrient-limited, and even under LL conditions, plants benefit from having this surplus of N (i.e. NO 3 − ). Moreover, NSCs under LL and HL were similar to those in control treatments, and significant differences were found only in BG starch, thus underscoring the positive effects of NO 3 − in our experimental design. However, opposite results have been found by Burkholder et al. (1992) and Burkholder et al. (1994) in Z. marina, in which NO 3 − appeared to damage the plants' meristems and led to leaf loss under pulsed daily additions (approximately 3.5, 7 or up to 10 μM NO 3 − d −1 for 14 wk). In studies with other species (e.g. Thalassia hemprichii), a neutral effect has been found only with use of NO 3 − (Jiang et al. 2013, Ow et al. 2016 Table 2. Relative response ratios (Eq. 1) of significant morphometric and physiological variables (see abbreviations in Table 1) in Zostera noltei plants when exposed to a single factor: low light (LL), high N (NO 3 − vs. NH 4 + ) and when these single factors were combined. The expected additive response is the null model to which the combined response was tested. Values shown are adjusted bootstrap means and 95% confidence interval (in brackets). Add.: additive; Antag.: antagonistic; Synerg: synergistic often been observed (Orth 1977, Peralta et al. 2003, van Lent & Verschnure 1995. The lack of consistent results emphasizes the need for direct comparative studies on NO 3 − enrichment and calls attention to the presence of other factors causing stress to the plants, given that negative effects of NO 3 − enrichment are associated with the tight coupling between C and N metabolism and consequently with decreased Cskeleton availability to respond to such additional stress (Jiang et al. 2013, Hughes et al. 2018. In contrast, NH 4 + enrichment triggered negative ef fects independently of light conditions, but also boosted those found under LL conditions. Unlike NO 3 − enrichment, NH 4 + enrichment led to lower growth, survival and SAR than did control and NO 3 − treatments under both light conditions. Moreover, a remarkable increase in internal N in leaves was observed under NH 4 + enrichment, reaching concentrations greater than 4% under LL treatments. This value is high, given that the N-limitation threshold in seagrasses is approximately 1.2−1.3% of N in AG biomass (Duarte 1990). The uptake of NH 4 + occurs primarily through passive and unregulated processes (Britto & Kronzucker 2002) and has a positive linear relationship with external concentrations (Pedersen et al. 1997, Alexandre et al. 2011, thus leading to high internal concentrations of NH 4 + (Villazán et al. 2015). To limit these toxic effects, plants must assimilate this NH 4 + into amino acids to prevent intracellular storage of NH 4 + (Britto & Kronzucker 2002, Marsch ner 1995, Pedersen et al. 1997, van Katwijk et al. 1997, Villazán et al. 2015, thus increasing the N content in tissues. Similar responses have been observed in laboratory experiments in several species , Egea et al. 2018, Moreno-Marin et al. 2018) as well as in seasonal studies in Z. noltei beds (Pérez-Lloréns & Niell 1993, Vermaat & Verhagen 1996, Brun et al. 2003b). In contrast, the poor ability of Z. noltei to survive under very LL conditions may be explained by the restricted sucrose mobilization throughout the plant under LL levels and the small starch reservoir that this species must use to meet C demands (Brun et al. 2003a). Interestingly, LL and NH 4 + enrichment affected the lengths of the leaves, following an opposite pattern from that observed in the other LL treatments because the leaves were shorter. This is a clear ex ample of a tradeoff: to improve light harvesting efficiency under LL conditions, leaves must be longer, but longer leaves could increase passive NH 4 + uptake and consequently exacerbate NH 4 + toxicity, potentially compromising plant survival. This tradeoff may also partially explain why NH 4 + accumulated in seawater during the experimental period, particularly in LL treatments, in agreement with the observed decrease in DIN uptake capacity under the LL + NH 4 + treatment (Fig. S1). Our findings may indicate that NH 4 + toxicity has a negative feedback effect under LL conditions because NH 4 + toxicity is concentration-dependent (van Katwijk et al. 1997, Brun et al. 2002, van der Heide et al. 2008). Therefore, the lower the AG biomass (e.g. because of shoot mortality, shorter leaves, lower biomass, etc.), the lower the NH 4 + uptake from the water, thus increasing the NH 4 + concentration in the water and enhancing its toxicity in a continuous feedback mechanism.
Although the mechanistic processes underlying physiological responses to separate factors (e.g. light and nutrients) can be explored in indoor mesocosms and are well described here, the complexity in nature -where factors interact simultaneously and plants may have opposing responses (e.g. leaf length) to the factors present-makes the final response difficult to predict. Several meta-analyses have indicated important roles for synergistic and antagonistic effects in marine organisms (Crain et al. 2008, Jackson et al. 2016. Some studies on seagrasses have demonstrated that synergistic interactions occur when plants are exposed to a combination of stressors, such as light, salinity, temperature and eutrophication (Collier et al. 2011, Salo & Pedersen 2014, Ontoria et al. 2019. However, as shown by this study, a large fraction of the responses at the physiological level were additive, which is consistent with previous studies of combined multiple stressors on seagrasses (e.g. Egea et al. 2018, Moreno-Marin et al. 2018. Therefore, stressor responses appear to be highly plastic and context-dependent, and de signing ecologically realistic experiments that consider the impact of local stressors (e.g. nutrient in puts) within the context of global stressors (e.g. climate change) will be particularly valuable (Gunderson et al. 2016). In this sense, our results highlight aspects that should be considered in setting up and performing experiments. First, the passive uptake of NH 4 + (in contrast to NO 3 − ) may affect nutrient enrichment, given that some effects were found to depend on the N source used (i.e. NH 4 + vs. NO 3 − ). Furthermore, factors such as hydrodynamics (e.g. narrow boundary layers; La Nafie et al. 2012), temperature (van Katwijk et al. 1997, Brun et al. 2002), pH (van der Heide et al. 2008, Egea et al. 2020, shoot density (van der Heide et al. 2008), salinity (Villazán et al. 2013) and phosphate presence (Brun et al. 2008), among others, can influence NH 4 + effects. Moreover, other plant traits may affect the whole response of the plant to combined stressors. In addition, most of these interrelationship pathways are bidirectional and also affect habitat complexity (van der Heide et al. 2008) and secondarily affect the whole seagrass community (e.g. herbivore and filter-feeder abundance; Jiménez-Ramos et al. 2017).
In summary, our study showed that the form of DIN supplied (reduced vs. oxidized) is of critical importance in seagrass ecosystems because different forms may have opposite ecological consequences. Z. noltei exhibited a positive response under NO 3 − en richment independent of light conditions, but showed diminished growth, survival and NSCs with NH 4 + enrichment, mainly under LL conditions. Although we found positive effects of NO 3 − enrichment, extrapolation of these results to in situ conditions must be performed with caution, as complex relationships in the ecosystem and other indirect ef fects (e.g. increasing photosynthetic growth, de creasing C reserves within the plant, enhancing the settling of organic matter into the sediment, etc.) may blunt this initially beneficial effect.