Nitrogen sufficiency enhances thermal tolerance in habitat-forming kelp: implications for acclimation under thermal stress

Local and global changes associated with anthropogenic activities are impacting marine and terrestrial ecosystems. Macroalgae, especially habitat-forming species like kelp, play critical roles in temperate coastal ecosystems. However, their abundance and distribution patterns have been negatively affected by warming in many regions around the globe. Along with global change, coastal ecosystems are also impacted by local drivers such as eutrophication. The interaction between global and local drivers might modulate kelp responses to environmental change. This study examines the regulatory effect of NO3− on the thermal plasticity of the giant kelp Macrocystis pyrifera. To do this, thermal performance curves (TPCs) of key temperature-dependant traits–growth, photosynthesis, NO3− assimilation and chlorophyll a fluorescence–were examined under nitrate replete and deplete conditions in a short-term incubation. We found that thermal plasticity was modulated by NO3− but different thermal responses were observed among traits. Our study reveals that nitrogen, a local driver, modulates kelp responses to high seawater temperatures, ameliorating the negative impacts on physiological performance (i.e. growth and photosynthesis). However, this effect might be species-specific and vary among biogeographic regions – thus, further work is needed to determine the generality of our findings to other key temperate macroalgae that are experiencing temperatures close to their thermal tolerance due to climate change.

Thermal performance curves (TPCs). Most of the physiological traits of Macrocystis showed the typical non-linear relationship between temperature and performance (Figs. 1-4; electronic supplementary material, Table S2). However, the effect of the internal N status on the TPCs' shape and position varied among the physiological traits (Table 1). RGR and photosynthesis showed significant differences in the shapes and fits of TPCs between N-replete and N-deplete blades (Figs. 1 and 2) ( Table 2). RGR only evidenced vertical shifts (µ max ) whilst photosynthesis exhibited both vertical and horizontal shifts (µ max and T opt ) in the TPCs. In all of these cases, differences were explained by the considerably higher maximum rates (µ max ) in N-replete blades than in N-deplete blades. Note that at the highest temperature (27 °C) seaweeds died in all experimental treatments and the data were not included in TPC analyses. Therefore, the TPCs are plotted for experimental treatments from 6-24 °C (Figs. 1 and 2). In contrast to the photosynthetic and growth rates, the shape and fit of Fv/Fm and NR activity TPCs were not significantly affected by the internal N status of the alga (Figs. 3 and 4, Table 2).
Pigment concentrations (i.e. Chl a, c and Fx) did not fit any of the non-linear models, suggesting a linear relationship between pigments and temperature that was not influenced by the internal N status of the alga (Table 2 and electronic supplementary material, Fig. S1).
Chlorophyll a fluorescence of PSII. The initial slope (α) of the RLCs did not vary between N-replete and N-deplete blades, but decreased significantly with increasing temperature (ANOVA, P < 0.05; electronic supplementary material, Fig. S3). The E k for the RLCs curves did not vary significantly between N-replete and N-deplete blades across all temperature treatments (ANOVA, P > 0.05), and ranged from 23.95 µmol m −2 s −1 at 10 °C to 76.87 µmol m −2 s −1 at 24 °C for N-replete blades, and from 32.95 µmol m −2 s −1 at 14 °C to 78.05 µmol m −2 s −1 at 24 °C for N-deplete (electronic supplementary material, Fig. S3). The ETR max varied significantly among temperatures, however, neither the internal N status nor the corresponding interaction were significant (ANOVA two-ways, p = <0.001, 0.612, 0.157).

Discussion
Thermal plasticity is a mechanism by which populations rapidly acclimate to warming and to extreme high temperature-related events such as MHWs 40 . Our study is the first to describe the plasticity of temperature-dependant traits of the ecologically and economically important kelp Macrocystis. We show that NO 3 − availability and hence the internal N status of the alga modulated the thermal plasticity of Macrocystis, buffering the negative impact of high temperature on its physiological performance, at least over a short-term incubation (Fig. 5). This may be of great importance with short-term extreme events, such as MHWs, increasing in frequency and intensity 68 . Rapid physiological acclimation to short-term variability in temperature and/or other environmental drivers (e.g., nutrient availability) can play an important role in seaweed responses to ongoing climate change. Our results suggest that populations of Macrocystis which are naturally exposed to moderate inputs of NO 3 − due to e.g., upwelling events or anthropogenic activities, may have a greater tolerance to high temperatures. However, populations that are exposed to limiting nutrient concentrations, like those from Tasmania, might be strongly negatively affected by OW and MHWs. We suggest that local drivers will play an important role in driving kelp responses to global change, and hence region-specific responses can be expected -as suggested by Krumhansl et al. (2016).
Local adaptations in thermal physiology have been recorded for macroalgal species across the latitudinal distribution of the species and ecotypes 58,69-72 but we know little about the influence of nitrogen on the thermal performance of macroalgae. For the annual kelp Undaria pinnatifida, Gao et al. 55 found differences in the thermal tolerance of geographically separated populations, where individuals with a higher thermal tolerance had  the greatest capacity to store N. These results support our hypothesis that populations of Macrocystis that are naturally exposed to greater N supply will have a greater thermal tolerance than that those exposed to limiting nutrient concentrations. For microalgae, it is well documented that they are more vulnerable to high temperatures under N limited conditions compared to N-sufficient ones 73,74 . However, further studies comparing the effects of www.nature.com/scientificreports www.nature.com/scientificreports/ nitrogen and/or other local drivers on the thermal plasticity of populations separated geographically are urgently required to more precisely predict species' responses to climate change.
We found that the internal N status modulated the thermal plasticity of Macrocystis, but its effects on TPC shapes varied among traits. Maximum photosynthetic and growth rates (µ max ) were enhanced in N-replete blades. The positive effects of higher nitrogen availability on metabolic rates has been described in other brown kelp e.g., Saccharina japonica 57 . However, both the photosynthetic and growth rates of Macrocystis showed differences in the degree of plasticity. Only optimum temperatures (T opt ) for photosynthesis were increased in N-replete blades while T opt for growth did not change between N-replete and N-deplete blades. Also, T opt for photosynthesis was higher than that for growth, which agrees with previous studies on macroalgae 41,75 . The differences observed in the thermal plasticity of growth and photosynthesis might be associated with the regulatory effect of temperature and nitrogen on each trait. For example, the photosynthetic machinery can rapidly acclimate to increases in temperature 76 , while growth is an integrated parameter that is regulated by many metabolic processes, including dark respiration, efflux of organic carbon, nitrogen uptake and assimilation 41,75 . This suggests that growth acclimation to changes in temperatures might be slower than for net photosynthesis.
Many aspects of thermal acclimation on key physiological traits are poorly studied in seaweeds compared to other marine organisms (e.g., invertebrates and corals) [77][78][79][80] and there are currently no standardized protocols for performing thermal tolerance experiments in seaweeds. Ours is a physiological study looking into the rapid acclimation responses to temperature stress; in order to acclimate seaweeds to the experimental temperatures we used a ramping of 2 °C/hour from the acclimation temperature of 17 °C to a maximum of 27 °C and a minimum of 6 °C. We recognise that changes of 10 °C in a span of 5 h are unlikely to occur in the shallow subtidal system from which we collected Macrocystis, and although this may have contributed to increased physiological stress, the experiments were designed to develop the thermal performance curves (TPCs) of key physiological traits. Other studies have performed thermal ramps of 5 °C per day 81 , 3 °C per three days 82 and some have not performed any thermal acclimation before the start of experiments 56 . In order to make more realistic long-term predictions to climate change (i.e. OW and MHWs), development of appropriate protocols for thermal stress experiments (i.e. thermal ramping and acclimation procedure) are required, understanding that these conditions might vary among regions, depending on daily and seasonally local variability (i.e. in temperature).
Changes in thermal tolerance involve various adjustments in algal metabolism, in which nitrogen plays a critical regulating role 83 . These adjustments include increased production of heat shock proteins (HSPs), which are important to tolerate temperature-stress conditions [84][85][86] and changes in fatty acid composition and lipids of the thylakoid membrane 41,87 . We did not measure HSPs across the experimental treatments, but further investigation of the fatty acid and lipid composition in selected experimental treatments (6, 17 and 24 °C) (M. Schmid unpublished data) showed that under high nitrate concentrations, Macrocystis maintains a high proportion of polar lipids (PL) and shows no increase in free fatty acids (FFAs). Under low nitrate concentrations, however, PL decreased markedly with increasing temperatures, with a concomitant increase in FFAs. PL are key component of cellular membranes 88 , and hence a decrease in their proportion can negatively affect membrane stability at high temperatures. These results showed that nitrogen can rapidly influence lipid metabolism, and is a likely mechanisms by which Macrocystis acclimates to short-term thermal stress.
The thermal plasticity of Fv/Fm, which is often used as an indicator of photosynthetic stress and photoinhibition 89 , was not influenced by the internal N status of the algae, suggesting that the initial stage of the light reaction of photosynthesis (i.e. the transport of electrons through PSII) was stable under both low and enriched NO 3 − concentrations. In phytoplankton 90,91 and the green macroalga Ulva sp. 92,93 , N limitation negatively affects PSII efficiency, indicated by a decline in Fv/Fm, likely due to decreases in protein synthesis and reaction centres 90 . However, we observed no effect of the internal N status of the alga on PSII efficiency (i.e., Fv/Fm, ETR max , E k and α) and nor on the photosynthetic pigments, at least over the short term incubation. Our results, along with those of Mabin et al. 61 who found no effect of NO 3 − (0.5-3.0 µM) on the PSII parameters Fv/Fm and rETR max of Macrocystis, suggest that seawater NO 3 − concentrations were above those required for the efficient functioning of PSII. In south-eastern Australia, this species is naturally exposed to low NO 3 − concentrations and hence the functioning of the PSII might be locally adapted to low nutrient concentrations.
Availability of NO 3 − can directly affect NR activity by regulating the synthesis and degradation of the protein 94 , and it can be strongly regulated by internal NO 3 − pools 95 . In macro-and microalgae, NR activity is also responsive to changes in temperature 96 , with thermal plasticity (T opt ) varying between phytoplankton species [97][98][99] . However, there are no studies describing the regulating role of NO 3 − on NR thermal plasticity in macroalgae or phytoplankton species. The NR thermal plasticity of Macrocystis was not influenced by the internal N status of the algae. T opt (15 °C) for NR in N-replete and N-deplete blades was similar to those described for phytoplankton species, ranging between 10 to 20 °C, which are typically close to the T opt for growth 97 . This agrees with our study where T opt for NR was similar to that for growth (13-14 °C). The lack of effect of NO 3 − in regulating the NR thermal plasticity of Macrocystis may be due to the high intraspecific variability observed among individuals (activities ranged from 0.77-4.0 nmol NO 3 − g −1 FW min −1 ), or that NR activity was not limited by the NO 3 − concentration in the treatments. Similar to higher plants, two NR forms, one inducible and one constitutive, may occur in macroalgae 100 . It is possible that the inducible NR might be regulated by external environmental parameters, and the constitutive form maintains a constant activity rate 101 , which could explain our results. Thus, even when external NO 3 − concentrations are low, NR activity remains active. However, the regulation of the constitutive NR form has not been studied in macroalgae. Although we did not observe an increase in NR activity in N-replete blades, higher growth rates were observed across most of the temperature treatments, suggesting that more NO 3 − was assimilated and converted to N to support growth. Moreover, Macrocystis from New Zealand can rapidly respond to changes in N availability, up-regulating its N metabolism 95 . However, similar to PSII, NR activity for Macrocystis from Tasmania might be adapted to the local low ambient nutrient concentrations, and thus responses to environmental variability (physiological plasticity) might be distinct from other populations across the world.
Our results showed that thermal plasticity, tolerances and sensitivities, can vary markedly among physiological traits, with some traits responding faster than others to short-term variability in environmental temperature. Similarly, Wernberg et al. 69 showed that in three habitat-forming seaweeds (Sargassum fallax, Ecklonia radiata, Scytothalia dorycarpa) optimum temperature for photosynthesis ranged from 23 °C-25 °C, depending on the species; however, no optimum temperatures were detected for respiration. Previous studies on ectothermic animals have shown similar variability among traits (e.g., respiration, growth) 27 . These results highlight the importance of selecting the right traits for predicting the effects of warming on the whole organism 27 . For macroalgae, photosynthesis can rapidly respond to changes in temperature, which agrees with our study, and it is considered a good proxy to compare thermal performance between species 75 . However, for long term predictions, photosynthesis might not be the best parameter to use because it over estimates the upper thermal tolerances for long term growth 75 . Studies describing thermal plasticity of different traits in macroalgae are urgently needed to identify the most relevant traits to precisely predict and compare the effects of ongoing warming at the organism and population level.
Recent studies have highlighted the importance of plasticity and local adaptation in macroalgal responses to warming, suggesting that some ecotypes might be more resilient or vulnerable to high temperatures than others, depending on the conditions to which they are usually exposed 40,86,102 . Previous studies have illustrated the physiological plasticity of Macrocystis across its wide geographical distribution. For example, marked differences in thermal tolerance and N storage capacities have been observed in populations that are geographically isolated 66,[103][104][105] . Similarly, distinct thermal tolerances and survival abilities have been observed in microscopic life stages from populations locally exposed to a different gradient of temperature (warmer vs. cool temperate sites) 106 . These results suggest that populations that are naturally exposed to highly variable thermal and nutrient regimes can have different responses to future oceanic conditions compared to populations that are exposed to more stable environmental conditions. Further studies linking molecular (e.g., expression of heat-shock-proteins) with physiological responses will provide a better understanding of macroalgal plastic and adaptive capacities to respond to climate change.

Materials and Methods
Seaweed collection. Sampling was performed at Bruny Island (45°47′S, 170°43′E), Tasmania, Australia, in March 2016. At the time of collection, temperature and NO 3 − concentrations in surface waters ranged from 18.7-19.9 °C and from 0.14-0.53 µM NO 3 − , respectively. Also, temperatures were 3-4 °C above the average of previous summers due to the most intense and longest MHW recorded in the Tasman Sea 107 . A total of 80 young blades (the 2 nd and 3 rd blades below apical scimitar) were collected from different individuals of Macrocystis (3-4 blades from each of 26 sporophytes). Blades were transported to the laboratory in an insulated container with ambient seawater. At the laboratory, blades were gently rinsed and cleaned with 0.5 µm filtered natural seawater (NSW) of any visible epibionts by gently brushing. Each of the 80 blades were cut to a similar size of 11 cm × 3.5 cm (fresh initial weight 1.0 ± 0.2 g), at 2 cm from the neumatocyst/blade junction (meristematic zone). Initial physiological measurements on field collected samples (electronic supplementary material, Table S1) showed that seaweeds were healthy at the start of the experiment. Then, blade sections were incubated for 12 h to allow marginal wounds to heal, in transparent 2 L-jars (0.5 µm filtered NSW at 17 °C). Mixing in the culture tanks was provided by pumping air. Eight blade sections were used to assess their initial physiological status (i.e. photosynthetic parameters, growth, nitrate reductase (NR) activity and nitrogen and carbon content) as described below. − concentrations in each culture tank, and 100 mM PO 4 − was used to avoid P limitation through the experiment (5:1 N:P). Seawater mixing was provided by pumping air. A saturating photon flux density of 120-130 µmol m −2 s −1 was provided overhead by florescent white tubes (Envirolux CE F28T5/4100K-120477 240V) set on a 12L:12D photoperiod. Incident light was measured using a Li-Cor LI-1400 data logger equipped with a flat underwater radiation sensor LI-192. SW samples (10 mL) were taken every day before and after renewing the medium to monitor NO 3 − concentrations (electronic supplementary material, Table S1). Prior to the experiments and during the pre-experimental conditions, the cultures were maintained in a temperature-controlled room at 17 °C.

Temperature and nitrate incubations, and experimental design.
After the pre-experimental incubation, N-deplete (blades coming from the 5 µM NO 3 − treatment) and N-replete blades (blades coming from the 80 µM NO 3 − treatment) were haphazardly selected and placed into each of 64 Erlenmeyer 250 mL flasks, containing either low (n = 32) or enriched-NO 3 − SW (n = 32). After that, each 250 mL culture flask was randomly assigned to one of the seven temperatures treatments, 6-10-14-17-20-24-27 °C, with four replicates for each experimental treatment. Blades were gradually acclimated from the pre-incubation temperature (17 °C) to the experimental temperatures. Thermal ramps were performed with linear temperatures changes of 2 °C per hour over a span of 5 h. Although the temperature range, 6 °C to 27 °C, and the thermal ramp does not fully coincide with the conditions experienced by the species in the field, the temperature range and ramping were selected to estimate the short-term thermal acclimation and precisely develop the TPCs for the specie. The culture flasks under each temperature treatment were maintained in a controlled temperature water bath for three days and subjected to a 12L:12D photoperiod under a saturating light intensity of 120-130 µmol m −2 s −1 provided and measured as described in the pre-experimental incubations. SW was changed daily and mixed by pumping air. Temperature and light conditions within each temperature treatment were monitored continuously using HOBO pendant temperature/light data loggers (64K-UA-002-64). SW samples (10 mL) were taken every day before and after renewing the medium to monitor NO 3 − concentrations in each treatment (electronic supplementary material, Table S1). After the 3-day incubation, Macrocystis blades were harvested to determine their physiological and biochemical responses (i.e. photosynthesis, growth, nitrate assimilation, and carbon and nitrogen content).
Physiological and biochemical parameters. Photosynthetic rates expressed as oxygen evolution were measured on the last day of the temperature/nitrate experiment, for each of the 56 experimental Macrocystis blades (n = 4 for each treatment combination). To do this, each blade was incubated separately in a biochemical oxygen demand (BOD) bottle, containing 266 mL of filtered 0.5 µm SW. A control BOD bottle without seaweed was also carried out. BOD bottles were placed on the top of an orbit shaker table set at 100 rpm to provide water movement. Dissolved oxygen (DO) was measured at the start and after 1 h of incubation using an optical dissolved oxygen sensor (Hach LBOD10101) and a DO meter (HQ40d Hach). Photosynthetic rates were determined under a saturating light of 120-130 µmol m −2 s −1 that was provided overhead by florescent white tubes (Envirolux CE F28T5/4100K-120477 240 V). For each temperature treatment, BOD bottles were put inside a transparent plastic box, where the temperature was controlled by an aquarium heater set up at each temperature. Photosynthetic rates were estimated from the initial and final oxygen concentrations (mg/l), and standardized by the fresh weight (g) of each blade and the incubation volume (l).
After measuring photosynthetic rates, chlorophyll a fluorescence of photosystem II was measured using a Pulse Amplitude Modulation fluorometer (diving-PAM, Walz, Germany). Macrocystis blades from the different treatment combinations were dark-adapted for 20 min before exposure to the PAM's photosynthetic active radiation (PAR, 0-422 μmol photons m −2 s −1 ). Rapid light curves (RLCs), relative ETR (rETR) versus irradiance, were conducted right after dark adaptation. Calculations of rETR of algal samples were estimated using the equation: where Y(II) is the quantum yield of photochemical quenching, E the incident irradiance of PAR, A the average ratio of light absorbed by algal tissue (0.8 for kelps) and the factor 0.5 assumes that 50% of the all absorbed energy has been utilized by PSII 108 . The rETR-RLCs were fitted according to Eilers and Peeters (1988), so that the light saturation (E k ), initial slope (α) and the maximum ETR (ETR max ) parameters were calculated from each curve. The optimum quantum yield (Fv/Fm), which represents a good indicator of maximal algal photosynthetic efficiency 109 , was calculated right after the dark adaptation. Relative growth rates (RGR, % days −1 ) were estimated after measuring photosynthetic rates and Chlorophyll a fluorescence by the difference in fresh weight (FW) after 3 days of incubation, using the formula: where W 0 is the initial FW and W t is the final FW after t days of incubation. The FW was estimated after blotting the blade section gently with tissue. After the growth measurement, blade sections, from each experimental treatment, were cut along the blade into four pieces in order to assess NR activity, C and N content and pigment concentrations. Tissue samples for NR activity (0.21-0.30 g FW), pigment analysis (0.10-0.15 g FW) were immediately frozen in liquid N 2 and stored at −80 °C until further analyses. Tissue samples for C and N content and stable isotopes (0.008-0.010 g FW) were oven dried for 48 h at 60 °C. NR activity was measured by nitrite production in an in vitro assay 110 . NR extraction methodology is described in detail in Fernandez et al. 95 . Briefly, NR was extracted in a 200 Mm Na-phosphate buffer (pH 7.9), containing 3% w/v BSA, 0.3% w/w polyvinylpyrrolidone (PVP), 2 Mm Na-EDTA and 1% w/v Triton X-100 (all Sigma, St Louis, MO, USA). The content of photosynthetic pigments (chlorophyll a and fucoxanthin) was analysed using methods described in Seely et al. 111 , Wheeler 112 , and Stephens and Hepburn 113 , where dimethyl-sulfoxide (DMSO) was used for the primary extraction and acetone for the secondary. Tissue samples for C and N content and stable isotopes from field collection and pre-experimental incubations, and temperature/nitrate experiments were determined according to Cornwall et al. 114 . Seawater analyses. Nitrate concentrations were analyzed using a QuickChem 8500 series 2 Automated Ion Analyzer (Lachat Instrument, Loveland, CO).

Data analysis.
Prior to analyses, we tested for normality and homoscedasticity for all variables, using the Lilliefors and Levene tests, respectively. Data was transformed either by log 10 or by square root to fulfil the requirements for parametric tests. Comparisons of parameters measured as part of the chlorophyll a fluorescence of PSII, were done via analysis of variance (ANOVA). When differences in the means were significant at the P < 0.05 level, they were also tested with a posteriori Tukey's test (HSD). Statistical analyses were performed with R 3.0.2 software and the package lme4 115 and in the GraphPad Prism software (v.7.03).
The effect of temperature on physiology and performance was described by a continuous nonlinear reaction norm (i.e. TPC) (Huey et al. 1999). Variation in the parameters of the TPCs (i.e. the optimal temperature -T opt ; the thermal breadth -T br ; the maximal performance -µ max ; and the upper and lower limits of temperature at which traits expression decrease -CT min and CT max ) was used to describe mechanistically the variation of thermal sensitivities and tolerances of natural populations (see Gaitan-Espitia et al. 2013, 2014) (Fig. 6). Here, we used www.nature.com/scientificreports www.nature.com/scientificreports/ the TableCurve2D curve-fitting software (version 5.01; Systat Software, Inc.) and the GraphPad Prism software (v.7.03) for model fitting. TPC parameters (µ max , T opt , CT min and CT max ) were extracted from the best models (see below for details). The physiological characteristics of critical thermal maximum (CT max ) and minimum (CT min ) were derived numerically as the intersection points of the resulting thermal performance curve with the temperature axis (μ = 0).
The fit of several linear and non-linear functions (e.g., Gaussian, Quadratic Lorentzian, Weibull) that could describe organismal performance as a function of temperature was analyzed using the Akaike Information Criterion (AIC) (Angilletta 2006). The AIC represents a balance between the likelihood explained by the model and the number of model parameters, with the best model minimizing AIC 116 . Thermal-dependent traits obtained from the TPCs were analyzed using a linear modelling approach. The effects of temperature and NO 3 − (fixed effects) were evaluated through confidence intervals (CI) computed from the likelihood profile 115 . In addition to CI, parameters and shapes of the TPCs were analyzed and compared using AIC and the Extra Sum-of-Squares F test. For linear models, the slopes and intercepts were compared using an analysis of covariance (ANCOVA) F-test. Figure 6. Typical TPCs describing the critical thermal minimum (CT min ) and critical thermal maximum (CT max ), at which physiological responses (e.g., photosynthesis, respiration) are possible. The thermal optimum (T opt ) is the temperature at which the physiological response reaches its maximum performance, and the optimal range represents the temperature range at half the maximum of performance (based on Angilletta et al. 2002, Sinclair et al. 2016). Both curves are representing plasticity in response to some environmental drivers (i.e. temperature), and the shift between curves (red to grey) represent a plastic adjustment to a different condition (e.g., nitrogen).