Interactions between seagrasses and seaweeds during surge nitrogen acquisition determine interspecific competition

Seagrasses dominate shallow coastal environments where nitrogen (N) availability in the water column is often sporadic and mainly in the form of pulses. We investigated the N uptake competition between seagrasses and seaweeds through a series of 15N surge uptake experiments combining single-species and mixed incubations across ammonium concentrations. N surge uptake rates of seagrasses were 2 to 14-fold higher than those of seaweeds in the majority of combinations, showing that seagrasses are generally in a competitive advantage over seaweeds in N-poor environments with N-pulses. No threshold concentration of ammonium was found beyond which seaweeds performed better than seagrasses. Mixed incubations revealed interspecific interactions that affected rates positively and negatively. Uptake rates obtained in single-species incubations, therefore, cannot always be used to predict the outcome of uptake competition. Only two (Zostera marina vs. Ulva rotundata and Zostera marina vs. Codium decorticatum) of the nine combinations tested (Z. marina, Z. noltei and Cymodocea nodosa vs. U. rotundata, C. decorticatum and Dictyota dichotoma) were found to enhance macroalgal uptake. Our results showed that the surge uptake capacity of seagrasses represents an important mechanism in their N acquisition strategy that justifies their dominance in shallow oligotrophic environments.

that often exceeds the required level for growth by several-fold 18 , is an important physiological mechanism that allows species to take advantage of transient peaks of nitrogen. The existence and characterisation of the surge uptake phase has been well described in macroalgae e.g. [19][20][21][22][23] , and in a few seagrass species, like Z. marina (for ammonium and nitrate), Z. noltei (for phosphate) and Amphibolis antarctica (for ammonium) [24][25][26] . In Z. noltei, the ammonium uptake rates by the leaves were 3 to 4 fold higher within the first 30 min of incubation, and within 120 min in the case of nitrate 27 . Surge uptake often occurs in areas where nutrients are limited and may be crucial to sustain growth under nutrient poor conditions 18,19,28 . We hypothesise that in N-poor environments with N-pulses seagrasses have a competitive advantage due to a greater surge uptake capacity relative to seaweeds, allowing them quicker to capture nitrogen and therefore become better N competitors.
We assessed the nitrogen competition dynamics between seagrasses and seaweeds using a variety of seagrasses and co-occurring seaweed species through a series of 15 N surge uptake experiments combining single-species and mixed incubations, i.e. species were incubated individually and under direct competition. Specifically, we aimed to provide answers to the following questions: i) are seagrasses better than seaweeds in surge uptake, and up to what threshold N concentration do seagrasses perform better, ii) are there any significant interspecific interactions between seagrasses and seaweeds that affect their nitrogen uptake rates and iii) which seagrass vs. seaweed combinations are most prone to macroalgal development, i.e., in which combinations do seaweeds perform better at taking up N? Nitrogen is a fundamental nutrient for seagrass and seaweed growth and one of the most limiting in the marine environment. Ammonium was chosen as the inorganic nitrogen source in these experiments because, in Ria Formosa lagoon, where the study was carried out, N pulses from the sediment to the water column with the incoming tide are mostly in the form of ammonium 15 , and because a compilation of more than thirty published studies comprising eight seagrasses and thirty-four seaweed species showed that ammonium is consistently preferred over nitrate 29,30 .

Results
The surge ammonium uptake rates of both seagrasses and seaweeds increased with nutrient concentration both when incubated in isolation or combination with other species (Fig. 1). In general, the effects of species (S), treatment (T) and nitrogen concentration (C) on the ammonium surge uptake of macrophytes were highly significant (Tables 1 and 2). The relative uptake rates of species when incubated in isolation or combination were maintained along the gradient of ammonium concentration in the majority of cases. The uptake rates of the seagrass Z. noltei were 3 to 14-fold higher than the seaweeds Ulva and Dictyota irrespective of treatment and N concentration ( Fig. 1a,b), and higher than Codium only in monospecific incubations at 100 µM N (Fig. 1c). The surge uptake rates of the seagrass Z. marina exceeded those of seaweeds only in the combination Z. marina vs. Dictyota, where seagrass uptake rates were 4 to 17-fold higher (Fig. 1d). Z. marina uptake rates were 3 and 6-fold lower than those of Ulva and Codium, respectively (Fig. 1e,f). N surge uptake rates of C. nodosa were 2 to 14-fold higher than all seaweed species, except in competition with Codium ( Fig. 1g) and Ulva at 3 µM N (Fig. 1h), where rates were similar.
Mixed incubations of seagrasses and seaweeds revealed the existence of both negative and positive interactions between macrophytes that affected their individual ammonium uptake rates (Table 3). Negative effects on uptake rates were observed on both competitors (Z. noltei and Ulva), or only on one competitor (on Dictyota in the presence of Z. noltei or Z. marina). Positive effects on uptake rates of the seagrasses Z. marina and C. nodosa were found in the presence of Ulva, on Z. noltei and C. nodosa in the presence of Dictyota and of the seaweeds Ulva in the presence of Z. marina (only at 100 µM) and Codium in the presence of C. nodosa ( Table 3). The absence of any interaction between macrophytes was found in the combinations Z. noltei vs. Codium and Z. marina vs. Codium at all ammonium concentrations. Thus, in most cases seagrasses were winners over seaweeds when competing for ammonium surges.

Discussion
Surge uptake is an important component of the uptake process as it may determine the competitive ability of a species to obtain the necessary nutrients in environments where nutrient concentrations generally are low. We showed here that seagrasses exhibit a remarkable uptake capacity of ammonium surges, which exceeded that of co-occurring seaweeds by several-fold in the majority of combinations. All seagrass species studied were able to take up ammonium more rapidly than seaweeds, except Z. marina when combined with Ulva and Codium. No threshold concentration of ammonium was found beyond which seaweeds performed better than seagrasses, suggesting that competition between seagrasses and seaweeds for ammonium surges is determined by the species-specific surge uptake rate rather than by the surge concentration.
Mixed species incubations revealed the existence of interactions between seagrasses and seaweeds. Both positive and negative effects on uptake rates were observed relatively to the rates of monospecific incubations, determining the competitive uptake winner at specific combinations. This important finding shows that the uptake rates of macrophytes, when incubated individually, cannot always be used to predict the outcome of uptake competition between seaweeds and seagrasses. Similar findings were also reported in a competition study of macroalgae vs. phytoplankton 31  be predicted using individual nutrient uptake parameters. To our knowledge, our work is the first that directly measures competitive dynamics of nutrient uptake rates between seagrasses and macroalgae. The ability of seagrasses to quickly remove nitrogen as it becomes available reflects their adaptation to environments where nutrient concentrations are typically very low but where pulses of nutrients normally occur 32,33 . In N-limited environments, where competition for the nutrient is high, seagrasses may be in a competitive advantage over seaweeds since they can explore short-lived pulses of nitrogen from the water column more efficiently, thus increasing their ability to maintain growth in environments with fluctuating N concentrations. Our results suggest that the surge uptake capacity of seagrasses represents an important mechanism in their N acquisition strategy that favors their survival and dominance in shallow oligotrophic environments.
In this study, the leaves and roots of seagrasses were incubated in the same compartment at the same concentrations, and the uptake by both plant parts was integrated as a whole-plant uptake, which we compared with the seaweed uptake. Even though ammonium-rich sediments are often considered the primary source of nitrogen for seagrasses, previous studies showed that the uptake of ammonium through the roots does not contribute significantly to the overall seagrass N acquisition because root uptake is typically much lower compared to those by the leaves in several seagrass species 34 and references therein. Thus, the whole-plant uptake rates of ammonium obtained in this study by incubating leaves and roots in the same concentration should not vary much from uptake rates obtained from incubating both plant parts separately at different concentrations. To completely unveil the hypothesis currently formulated that seagrasses are more efficient than seaweeds in nutrient uptake under low nutrient concentrations 4 , and may thus prevent macroalgal development when some nutrient threshold is exceeded, the long-term uptake rates of seagrasses versus seaweeds must also be analysed. However, care should be taken to address this hypothesis as nutrient uptake studies have been mostly done in single-species incubations. As we showed here, in some specific cases the uptake rates can be significantly altered in the presence of other species.
The observed interspecific interactions between seagrasses and seaweeds, which affected positively or negatively their individual ammonium uptake rates, may be explained by a specific limitation of other essential elements (e.g. phosphorus and carbon) that interfere with the uptake of ammonium. For example, Z. noltei has been shown to use the dissolved organic carbon excreted by Ulva to enhance growth 35 , something that could have benefit some of the seagrass species in our experiments (Z. marina and C. nodosa). More complex interspecific interactions may be expected if seagrasses are incubated with multiple species of seaweeds, and vice-versa, as in the natural environment. However, the hypothesis that a limitation by essential nutrients may affect the ammonium uptake rates between seagrasses and seaweeds must be experimentally tested. Another possible explanation is allelochemical-mediated interference. Allelopathy, i.e. the release of chemical substances by one plant eliciting positive or deleterious responses on another 36 , is known to be involved in interspecific competition between several aquatic macrophytes [37][38][39][40] . The chemicals released can affect numerous physiological processes in the target species, such as growth, photosynthetic performance, enzymatic activity and nutrient uptake 39,41 . In the present study, the ammonium uptake rates of the seaweeds Ulva and Dictyota were negatively affected by the presence of the seagrasses Z. noltei and Z. marina (except in the combination Ulva vs. Z. marina). We are not aware of any studies reporting allelochemical effects of seagrasses on macroalgae, but water soluble extracts of leaves of Zostera species contain several inhibitory substances, such as zosteric acid, flavonoids and phenolics 42 , which negatively affected the development and photosynthetic carbon uptake of epiphytic diatoms 43,44 , as well as growth of microalgae and marine bacteria 45 . A negative effect of macroalgae on the N uptake rates of seagrasses was found only in the combination Z. noltei vs. Ulva. Ulva species are well known for their strong allelopathic inhibitory effects on micro-and macroalgae 40,46-48 but effects on seagrasses have not been reported. Although by definition beneficial allelopathic effects may also occur, only a few studies reported such effects, and they were mainly observed in crop plants e.g. 49 . Nonetheless, in our study, we found positive seagrass-seaweed interactions in a large number of combinations, where the N uptake rates of at least one species increased relatively to those of monospecific incubations. As a result, the collective nitrogen uptake in mixed incubations was higher than the total N uptake in monospecific incubations. This is an interesting result and suggests that macrophyte diversity  Table 3. Summary of the effects of interspecific interactions on the ammonium surge uptake rates of each species. 0 = no effect; − % = percentage decrease relative to the species uptake in monospecific incubation; + % = percentage increase relative to the species uptake in monospecific incubation. Split cells indicate that more than one effect was observed for a specific combination, depending on the nutrient concentration. Values in brackets indicate the specific ammonium concentration at which the effect occurred.
may increase the total nitrogen uptake capacity of seagrass-dominated ecosystems as reported for seaweeds in tidal rock pools 50 .
The winners for ammonium uptake in different seagrass-seaweed combinations depend not only on the surge uptake capacity of each species but also on their interactions. Consequently, one seagrass species may be a winner in one specific combination but not in others. For example, the seagrass Z. marina was the uptake winner when combined with Dictyota, but not when combined with Ulva or Codium, irrespective of the ammonium concentration. The combinations Z. marina vs. Ulva and Z. marina vs. Codium were the most prone to macroalgal development because the surge uptake rates of the seagrass were always lower than the seaweeds. Z. marina is one of the most threatened seagrass species worldwide 51 , mainly due to the N enrichment of coastal habitats 52 , and is the most endangered of the three existing species in Ria Formosa lagoon 53 . It is possible that the global decline of Z. marina may be related to its lower surge uptake capacity relative to seaweeds. In the specific case of Ria Formosa, Z. marina appears to be close to a light-mediated ecophysiological threshold being less competitive for light than the sympatric seagrass Cymodocea nodosa 54 .
In conclusion, this study clearly shows that seagrasses can compete with seaweeds during surge uptake and thus prevent opportunistic macroalgae blooms in N-poor environments with high, transient ammonium inputs irrespective of their concentration. Significant species-specific interactions may affect the seagrass-seaweed competitive outcome. Although no interspecific interactions were observed in most of the combinations, positive effects were mostly observed over the uptake rates of seagrasses in the presence of seaweeds.

Site description and plant collection. Ria Formosa is a mesotidal coastal lagoon located in South
Portugal (37°01′N, 7°51′W). In this system, Zostera noltei is the most abundant seagrass species, developing extensive meadows along the intertidal mudflats and major contributor to the lagoon's metabolism 55 . The subtidal areas of the lagoon are occupied by the seagrass species Zostera marina and Cymodocea nodosa. Bloom-forming macroalgae co-occur with seagrasses in the lagoon. Ulva species, mostly U. rotundata cover mostly the intertidal areas occupied by the seagrass Z. noltei, but also settle over Z. marina and C. nodosa meadows. Other seaweeds also thrive, namely species of Dictyota dichotoma and Codium decorticatum, although these species develop more frequently in the subtidal areas of the lagoon. Ammonium and nitrate concentrations in the water column are usually less than 5 µM due to a high water exchange with the adjacent ocean during each tidal cycle 56 , but ammonium pulses (∼10 µM) from the sediment to the water column occur with the incoming tide 15 . Ammonium concentration in the sediment pore water is higher (12-38 µM), whereas nitrate concentration is almost negligible (0.2-0.9 µM) 57 . Macroalgae (U. rotundata, D. dichotoma and C. decorticatum) and seagrass species (Z. noltei, Z. marina and C. nodosa) were collected during the autumn and winter of 2012. In the laboratory, seagrass roots were carefully cleaned of adherent sediment and leaves were cleaned of epiphytes. The species were kept separately in aquaria with seawater from the collection site for two days to acclimate to the experimental conditions (seawater temperature of 14 °C and light intensity of 300 µmol quanta m −2 s −1 ). Experimental procedure. In a first experiment, the competitive dynamics of ammonium uptake between seagrasses and seaweeds was studied using three different combinations of co-occurring species in the lagoon: Z. noltei vs. Ulva, Z. marina vs. Dictyota and C. nodosa vs. Codium. In each combination, the ammonium uptake rate of each species was assessed by incubating them separately and in competition in nitrogen-free artificial seawater (salinity of 35‰, pH of 8.24) enriched with 15 NH 4 Cl (atom % = 98, Sigma) at seven different concentrations (3,6,12,25,50, 100, 200 µM) for 30 min. Incubations were performed in triplicate for each nutrient concentration.
In a second experiment, N competition dynamics was studied in another six different combinations of seaweeds and seagrasses (Z. noltei vs. Dictyota, Z. noltei vs. Codium, Z. marina vs. Ulva, Z. marina vs. Codium, C. nodosa vs. Ulva and C. nodosa vs. Dictyota), so that all possible combinations of the three seagrass and seaweed species were tested. The ammonium uptake rate of each species was assessed at low (3 µM) and high (100 µM) ammonium concentrations by incubating species separately or in competition as described above.
The incubation conditions were identical in the two experiments. The media were constantly stirred to decrease the thickness of the boundary layer and to ensure a homogeneous distribution of the isotopic labels. The biomass to volume relationship of the incubations was previously determined in preliminary experiments to ensure that the ammonium concentrations remained constant throughout the specific incubation period (i.e. no substantial change in the nutrient concentration occurred), preventing any significant ammonium limitation that could interfere in the rate of nutrient uptake. When in competition, one single seagrass module (i.e. shoot with respective rhizome and roots) and its equivalent fresh weight of seaweed were collectively incubated to eliminate the possibility that any existing interspecific interactions that could affect the uptake rates was due to differences in biomass between the two species. In single-species incubations, two seagrass modules and the equivalent fresh weight of seaweed were incubated separately. Species were immersed in 1 L of nitrogen-free artificial seawater (salinity of 35‰, pH of 8.24). Seawater was prepared using MilliQ ultrapure water with a pH of 5. The pH of the seawater solution was adjusted to 8.24 using HCO 3 − , which also provided the media with a source of inorganic carbon. The fresh weight of one seagrass module of Z. noltei was on average 0.12 ± 0.07 g (0.02 ± 0.01 g dry weight). One module of Z. marina averaged 1.06 ± 0.90 g fresh weight (0.19 ± 0.16 g dry weight) while one module of C. nodosa averaged 1.13 ± 0.57 g FW (0.20 ± 0.1 g dry weight). The aboveground: belowground biomass ratio was 1.47 for Z. noltei, 4.83 for Z. marina and 0.76 for C. nodosa. All experiments were performed in a walk-in culture chamber at constant temperature (14 °C) and light intensity (300 µmol quanta m −2 s −1 ). This light intensity has been shown to saturate, or nearly saturate, photosynthesis in virtually all studied species [58][59][60][61] . At the end of the incubation, tissues were removed from the media, seagrass leaves were immediately separated from the rhizomes and roots, and all tissues were briefly rinsed with deionised water to remove adherent salts and label ( 15 N). Tissues were dried at 60 °C for 48 h and reduced to a fine powder. Total nitrogen content and the percentage of 15 N of dried tissues were determined using a Flash EA 1112 Series elemental analyser coupled on line via a Finningan conflo II interface to a Thermo delta V S mass spectrometer. Precision in the overall preparation and analysis was better than 0.2‰. 15 N background levels of seagrass leaf, root and rhizome tissues and seaweed tissue were measured as controls (n = 5).
Though in nature the rhizosphere of seagrasses can be anoxic, in these experiments we incubated whole plants in an oxygenated medium. Previous experiments showed no effects of rhizosphere oxygenation on the ammonium uptake rates of leaves and roots of Z. noltei 34,62 , which was confirmed for Z. marina in preliminary experiments of the present study. Data analysis. 15 N enrichment (%) of tissues after incubations were calculated by subtracting the post-incubation 15 N levels from the initial background levels, which was multiplied by the total nitrogen content of the tissue (g) and then divided by its weight (g dry weight). For seagrasses, the uptake rates were expressed as whole-plant uptake rates and were calculated by the sum of the uptake rates of leaves and roots calculated using the surface area of the respective plant part and then divided by the sum of the surface areas of the plant parts. Because nutrient uptake during the surge phase is diffusive, the extent of the surge uptake rate is expected to be a direct function of the number and activity of sites available for the nutrient transport into the cells at the plant surface 63 . We, therefore, believe that, when evaluating interspecific competition during the surge phase, uptake rates should be expressed per surface area units rather than per biomass units. The latter would be more appropriate when comparing uptake rates during the subsequent internally controlled phase, i.e. when uptake rates are controlled by the rate of nutrient assimilation in the cells 22 . All incubated tissues were photographed, and their surface areas were calculated using the software Image J 64 . Significant differences in the nitrogen uptake rates of seagrass and seaweed species (S) incubated alone or in competition (T) at different nitrogen (N) concentrations were analysed for all combinations of species tested using a permutational analysis of variance 65 , using the PERMANOVA module 66,67 within Primer 5 software 68 , with three fixed factors: 1) species, with two different levels: seagrass and seaweed, 2) treatment, also with two levels: alone and in competition; and 3) N concentration, with seven different levels: 3, 6, 12, 25, 50, 100 and 200 µM in the first experiment, and two levels: 3 and 100 µM in the second experiment. This method does not require implicit assumptions about the underlying distribution (i.e. normality) or spread (i.e. variance) of the data, hence does not assume either normality or homoscedasticity. Permutation of residuals under a reduced model, with 9999 permutations on a data matrix of average distance measures was performed as recommended to test distance based homogeneity of dispersion, main effects and pair-wise tests on significant factors/interactions. In case the number of unique permutations was lower than 100, Monte Carlo permutations(9999) p-values were used.
Data availability statement. The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.