Atmospheric CO2 availability induces varying responses in net photosynthesis, toxin production and N2 fixation rates in heterocystous filamentous Cyanobacteria (Nostoc and Nodularia)

Heterocystous Cyanobacteria of the genus Nodularia form major blooms in brackish waters, while terrestrial Nostoc species occur worldwide, often associated in biological soil crusts. Both genera, by virtue of their ability to fix N2 and conduct oxygenic photosynthesis, contribute significantly to global primary productivity. Select Nostoc and Nodularia species produce the hepatotoxin nodularin and whether its production will change under climate change conditions needs to be assessed. In light of this, the effects of elevated atmospheric CO2 availability on growth, carbon and N2 fixation as well as nodularin production were investigated in toxin and non-toxin producing species of both genera. Results highlighted the following: Biomass and volume specific biological nitrogen fixation (BNF) rates were respectively almost six and 17 fold higher in the aquatic Nodularia species compared to the terrestrial Nostoc species tested, under elevated CO2 conditions. There was a direct correlation between elevated CO2 and decreased dry weight specific cellular nodularin content in a diazotrophically grown terrestrial Nostoc species, and the aquatic Nodularia species, regardless of nitrogen availability. Elevated atmospheric CO2 levels were correlated to a reduction in biomass specific BNF rates in non-toxic Nodularia species. Nodularin producers exhibited stronger stimulation of net photosynthesis rates (NP) and growth (more positive Cohen’s d) and less stimulation of dark respiration and BNF per volume compared to non-nodularin producers under elevated CO2 levels. Biomass and volume specific biological nitrogen fixation (BNF) rates were respectively almost six and 17 fold higher in the aquatic Nodularia species compared to the terrestrial Nostoc species tested, under elevated CO2 conditions. There was a direct correlation between elevated CO2 and decreased dry weight specific cellular nodularin content in a diazotrophically grown terrestrial Nostoc species, and the aquatic Nodularia species, regardless of nitrogen availability. Elevated atmospheric CO2 levels were correlated to a reduction in biomass specific BNF rates in non-toxic Nodularia species. Nodularin producers exhibited stronger stimulation of net photosynthesis rates (NP) and growth (more positive Cohen’s d) and less stimulation of dark respiration and BNF per volume compared to non-nodularin producers under elevated CO2 levels. This study is the first to provide information on NP and nodularin production under elevated atmospheric CO2 levels for Nodularia and Nostoc species under nitrogen replete and diazotrophic conditions.

• Biomass and volume specific biological nitrogen fixation (BNF) rates were respectively almost six and 17 fold higher in the aquatic Nodularia species compared to the terrestrial Nostoc species tested, under elevated CO 2 conditions. • There was a direct correlation between elevated CO 2 and decreased dry weight specific cellular nodularin content in a diazotrophically grown terrestrial Nostoc species, and the aquatic Nodularia species, regardless of nitrogen availability.
• Elevated atmospheric CO 2 levels were correlated to a reduction in biomass specific BNF rates in non-toxic Nodularia species. • Nodularin producers exhibited stronger stimulation of net photosynthesis rates (NP) and growth (more positive Cohen's d) and less stimulation of dark respiration and BNF per volume compared to non-nodularin producers under elevated CO 2 levels.

Introduction
Cyanobacteria, in their role as primary producers, form an essential part of the global C and N cycles, both in terrestrial and aquatic environments (Visser et al. 2016;Elbert et al. 2012). The process of oxygenic photosynthesis, whereby energy from the sun is used to reduce inorganic carbon with the accompanying oxidation of water, is thought to have evolved during the Archean era when there was no free oxygen in the Earth's atmosphere (Lyons et al. 2014). The enzyme catalysing CO 2 fixation in Cyanobacteria and modern-day C3 plants is ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), thought to be the most abundant enzyme on Earth. Rubisco binds CO 2 and generates 2 molecules of 3-phosphoglycerate (3PGA) which is further processed in the Calvin-Benson-Bassham (CBB) cycle to produce ribulose-1,5-biphosphate and glutamate. In order to reduce undesirable oxygenase activity, Cyanobacteria have evolved the carbon concentrating mechanism (CCM) to increase the effective concentration of CO 2 around the Rubisco active site by up to 1000-fold (Price 2011). CO 2 diffuses freely into the cell and is converted to bicarbonate in an NADPH-dependent reaction. Most Cyanobacteria sequenced to date carry the high flux, low affinity CO 2 converting enzyme, NDH-I 4 , as well as the low flux, high affinity NDH-I 3 variant. Uptake of bicarbonate from the surrounding liquid requires an investment in energy and the synthesis of specific transporters. Two sodium -dependent symporters, BicA (high flux, low affinity) and SbtA (low flux, high affinity) bicarbonate transporters occur occasionally together with the BCT1 high affinity low flux transporter, found in almost all Cyanobacteria investigated to date on the cell membrane (Burnap et al. 2015;Visser et al. 2016). The presence of BicA provides aquatic Cyanobacterial species a growth advantage under elevated levels of HCO 3 − availability (Sandrini et al. 2014). Oxygenic photosynthetic organisms that rely on the construction of a carbon concentrating mechanism (CCM) are thought to be sensitive to changes in pCO 2 (e.g. Raven et al. 1991;Rost et al. 2003;Price 2011;Shi et al. 2012;Raven et al. 2017). The plasticity of the CCM to elevated levels of atmospheric CO 2 was found to be high in Cyanobacteria when compared to the more recently evolved haplophytes and diatoms (Van de Waal et al. 2019). This phenotypic plasticity in carbon fixation was demonstrated on Microcystis grown under conditions of elevated CO 2 (Ji et al. 2020). The maximum CO 2 uptake rate of Microcystis grown at 1000 ppm CO 2 was increased 1.5-1.8 times compared to the low CO 2 control cultures, suggesting that elevated CO 2 conditions may stimulate Cyanobacterial bloom growth (Ji et al. 2020). Furthermore, by reducing the levels of dissolved CO 2 and increasing the pH in dense blooms, Cyanobacterial species succession is thought to be driven towards strains with a more efficient carbon concentrating mechanisms (Lines and Beardall 2018).
Globally, an increase in phytoplankton blooms, including Cyanobacterial harmful algal blooms, has been recorded since the 1980's (Ho et al. 2019). While the reasons for the observed increase is unclear, temperature, elevated atmospheric CO 2 levels and eutrophication especially of the freshwater lakes are potential drivers of this phenomenon. Approximately a third of all anthropogenic CO 2 released dissolves in the oceans, reducing the pH by increasing the partial pressure of CO 2 , accompanied by a smaller relative increase in HCO 3 − and a decrease in CO 3 2− (Sabine et al. 2004;Raven et al. 2017). Although the speciation of dissolved inorganic carbon is directly linked to pH, how changes in their balance affects Cyanobacterial bloom occurrence and toxicity is unclear (Raven et al. 2020). The increase in growth rate observed for the marine, nonheterocystous, filamentous diazotrophic Cyanobacterium, Trichodesmium, grown at 900 ppm CO 2 was ascribed to down regulation of the CCM, thereby reducing the energy demands on the cell (Kranz et al. 2011). Under Fe-limiting conditions, decreasing the medium pH reduced N 2 fixation rates in Trichodesmium, with the reduced N 2 fixation rates corresponding to reduced nitrogenase efficiency at lower pH (Kranz et al. 2011). Exposing cultures of the freshwater diazotroph, Nostoc muscorum, to raised HCO 3 − concentrations under diazotrophic conditions resulted in enhanced growth, O 2 and pigment production and nitrogenase activities (Bhargava et al. 2013). The brackish diazotroph, Nodularia spumigena sp. KAC12, when grown at elevated CO 2 of 960 ppm, demonstrated increased photochemical yield after 5 days exposure (Karlberg and Wulff 2013), suggesting higher potential net primary productivity rates. Nodularia spumigena CCY9414, grown under elevated CO 2 conditions (548 ppm), exhibited increased C fixation rates compared to control cultures, with increased carbon to nitrogen (POC:PON) and nitrogen to phosphate ratios recorded (Wannicke et al. 2012). Only a slight increase was observed in the C:N ratios in three Cyanobacterial cultures grown at elevated CO 2 (~ 900 ppm) in continuous culture in bubble reactors, namely Cyanothece sp. ATCC51142, Nodularia spumigena IOW-2000/1 and Calothrix rhizosoleniae sp. SC01 (Eichner et al. 2014). This study emphasised the need to generate more data on the effects of elevated CO 2 levels on Cyanobacterial BNF, and highlighted the diversity in observed responses of marine Cyanobacterial species to elevated atmospheric CO 2 . Wannicke et al. (2018b), in their metadata study, found indications that ocean acidification would benefit BNF in the future ocean. They also drew attention to the fact that these studies were mostly conducted on only two species, the filamentous Trichodesmium and unicellular Crocosphaera. Very few studies were published on filamentous heterocystous Nodularia, Calothrix and Anabaena (alias: Dichlospermum) species (reviewed by Wannicke et al. 2018b). A more recent study suggested that growth of the diazotrophic Dolichospermum circinale might benefit from increased CO 2 levels of 1700 ppm (Symes and van Ogtrop 2019).
Studies investigating the effect of climate change on filamentous diazotrophic Cyanobacteria in terrestrial habitats are rare too. Terrestrial surfaces are often inhabited by cryptogrammic covers, including Cyanobacteria that contribute a significant amount to global net primary productivity (Elbert et al. 2012). Specifically, it is estimated that N 2 fixation by cryptogrammic covers may account for almost half of biological nitrogen fixation on land, ~ 49 Tg per year (Elbert et al. 2012). Biological soil crusts showed a decrease in Cyanobacterial abundance when grown under elevated atmospheric CO 2 for 10 years, suggesting a negative impact of climate change on arid soil crusts (Steven et al. 2012). The ability of Cyanobacterial soil crusts to increase net primary production under high CO 2 (HC) exposure was shown to be dependent on water availability (Lane et al. 2013). This was in agreement with previous research demonstrating that terrestrial Nostoc flagelliforme exhibited its highest relative growth rate under conditions of high CO 2 (1500 ppm) in moist conditions when compared to mats grown at 350 ppm CO 2 (Gao and Yu 2000). Rodriguez-Caballero et al. (2018) suggested that dryland soil crusts are under threat due to anthropogenically induced climate change. Their projected loss of 25-40% cryptogrammic coverage will result in reduced microbial contributions to nitrogen cycling and soil surface stabilisation. Additionally, little information exists as to how Cyanobacterial biocrusts specifically, will respond to increasing atmospheric CO 2 levels (Reed et al. 2016).
Given the evolutionary history of Cyanobacteria already existing under raised CO 2 levels, researchers (Gehringer and Wannicke 2014;Sandrini et al. 2016;Visser et al. 2016;Buratti et al. 2017) have voiced concerns for increased Cyanobacterial bloom occurrence and toxin production under the elevated levels of CO 2 proposed by current climate change scenarios, particularly in eutrophic waters (Ma et al. 2019). Especially toxin producing Cyanobacteria capable of fixing atmospheric N 2 would offer a potential threat to human safety, as they could thrive in otherwise nitrogenlimited habitats (O'Neil et al. 2012). The most commonly occurring Cyanobacterial toxins are microcystin and nodularin, both strong protein phosphatase inhibitors, capable of inducing extensive hepatocellular bleeding and collapse in exposed individuals and animals (Gehringer 2004;Ibelings et al. 2015;Buratti et al. 2017). Microcystin and nodularin are synthesized by non-ribosomal peptide synthetases (Dittmann et. al. 2001;Moffit and Neilan 2004) for which the control mechanisms remain largely unknown. The levels of toxin production within Cyanobacterial blooms is largely determined by several abiotic factors such as light intensity and quality, pH and nutrient availability (Reviewed by Gehringer and Wannicke 2014;Visser et al. 2016;Buratti et al. 2017). Raised temperatures and elevated CO 2 levels in the range of those proposed under climate change, are linked to increased primary production (Paerl and Huisman 2009) and toxin production by Cyanobacteria (El-Shehawy et al. 2012;Kleinteich et al. 2012). Increased production of the secondary metabolite, microcystin, is linked to maintaining the C:N balance in the cell in the non-diazotrophic Microcystis aeruginosa (Downing et al. 2005), particularly when N uptake exceeds the relative growth rate. Elevated CO 2 levels have the capacity to affect the community composition and toxicity of Microcystis blooms significantly (Liu et al. 2016;Van De Waal et al. 2011;Sandrini et al. 2016;Buratti et al. 2017). Microcystin synthesis requires active photosynthesis (Sevilla et al. 2012) and, like nodularin synthesis, is regulated by the global N uptake regulator, NtcA, supporting the proposed importance of the C:N balance on toxin production (Neilan et al. 2013). This agrees with observed anthropogenically induced alterations in environmental N:P ratios, resulting in the appearance of Cyanobacterial blooms (Beversdorf et al. 2013) and increased toxin production (Horst et al. 2014). Inorganic nitrogen limitation was thought to induce a shift to N 2 fixing, diazotrophic Cyanobacteria, thereby increasing organic N availability and a subsequent increase in toxin production (Posch et al. 2012;Gehringer and Wannicke 2014). Recent investigations of bloom dynamics in Lake Müggelsee suggest that the predominant Cyanobacterial diazotrophs, Aphanizomenon sp. and Anabaena sp (Dichlospermum sp.), do not proportionally increase in numbers relative to non-nitrogen producers under conditions of reduced N availability (Shatwell and Köhler 2019). The changes in Ci availability, nitrogen fixation rates and potential cyanotoxin production levels were not reported. Transcription of the nda cluster in Nodularia spumigena AV1 was found to be altered in response to changes in ammonia and phosphate availability, however, the levels of intra-and extracellular nodularin were not significantly altered (Jonasson et al. 2008). Production of cylindrospermopsin and microcystin is thought to be constitutive, with cell cyanotoxin quotas being relatively fixed (Orr et al. 2018;Pierangelini et al. 2015). Orr et al. (2018) furthermore argued that toxicity is not affected through any stimulatory or trigger effect on the toxin production pathway itself, but via changes in rates of cell division and growth of different strains with genetically different cyanotoxin cell quotas. If the Cyanobacterial specific cyanotoxin rate matches the specific cell division rate, the overall cell cyanotoxin remains fixed. Dense Cyanobacterial blooms require excessive CO 2 to support their continued growth (Paerl and Huisman 2009) with CO 2 availability often limiting bloom growth, a restriction that could be removed under increased atmospheric CO 2 levels. Only aquatic Cyanobacteria carrying the high flux, low affinity BicA HCO 3 − receptor were able to benefit from elevated CO 2 levels and increase their growth rates (Sandrini et al. 2015;Visser et al. 2016).
Most studies on elevated CO 2 effects reported for toxin producing Cyanobacteria have focused on the production of the heptapeptide toxin, microcystin, in freshwater unicellular non-diazotrophic Microcystis aeruginosa species. The microcystin content of Microcystis aeruginosa HUB 5-2-4 grown at elevated CO 2 was raised, while growth rates were kept constant ( Van de Waal et al. 2009). Sandrini et al. (2015) reported that the shift of Microcystis aeruginosa PCC 7806 from 200 ppm pCO 2 to 1450 ppm pCO 2 in a continuous culture, resulted in a ∼ 2.7-fold increase of Cyanobacterial biomass and ∼ 2.5-fold elevation in microcystin per cell. Moreover, at high pCO2, gene expression of the high flux, low affinity BicA HCO 3 − receptor was down-regulated and cells shifted to CO 2 and low-affinity, high flux receptors for bicarbonate uptake. Interestingly, the expression of the mcy genes involved in microcystin synthesis remained constant, suggesting additional regulatory steps are involved in toxin synthesis under elevated CO 2 conditions. Studies investigating the competition of microcystin and non-microcystin producing strains of Microcystis at low and elevated pCO 2 levels found that non-toxic strains outcompete toxic strains under conditions of low light and high CO 2 availability (Van De Waal et al. 2011;Yu et al. 2015). On the other hand, toxin-producing strains display a better fitness under growthlimiting conditions suggesting that the benefit of producing the toxin outweighs its costs under unfavourable conditions (Briand et al. 2008;Van De Waal et al. 2011).
To our knowledge, there is no peer-reviewed publication concerning the effect of elevated CO 2 on nodularin production in Cyanobacteria. This study is directed at studying diazotrophic Cyanobacterial species from both terrestrial and aquatic environments to investigate the effect of elevated CO 2 levels on net photosynthesis, toxin production, growth and N 2 fixation rates in a multiple matrix approach. To do so, seven different species were chosen of which three are able to produce the toxin nodularin and four are non-nodularin producers. Finally, we analysed the data set gained in this study by applying weighted mean effect sizes to test the hypothesis that Cyanobacteria react differently towards elevated CO 2 depending on the whether they produce nodularin or not.

Culture conditions and experimental design
The experimental design contained a multiple matrix approach with different factorial designs for the Cyanobacteria tested, using a combination of different atmospheric CO 2 treatments of high CO 2 (HC), low CO 2 (LC), culture medium containing nitrogen in the form of NaNO 3 or not (N+ and N−) and the ability to produce the hepatotoxin, nodularin (+ and −) (Fig. 1a).
In total, six species of heterocystous filamentous Cyanobacteria belonging to two families, Nostocaceae and Aphanizomninaceae, within the order Nostocales, were selected for investigation. Four representative species of the genus Nostoc were analysed, with two being nodularin producers, ie Nostoc punctiforme sp. 73.1 and Nostoc muscorum sp. 65.1 and two non-nodularin producing species: Nostoc punctiforme sp. 40.5 and Nostoc entophytum sp. C1.8 (Gehringer et al. 2010(Gehringer et al. , 2012. Two representative species of the genus Nodularia, were investigated, with one nodularin producing species, Nodularia spumigena CCY9414 (Voss et al. 2013) and its related non-toxic mutant strain Nodularia spumigena NSBL06 [analogous to N. spumigena NSBL05 (Bolch et al. 1999;Moffit et al. 2001)] and the non-toxic benthic species of Nodularia harveyana SAG 44.85 (Lyra et al. 2005;Řeháková et al. 2014). The combination of the three variable factors, atmospheric CO 2 levels, nitrogen content and toxin production, generated four different factorial designs for the species tested (Fig. 1b).
The terrestrially isolated Nostoc species were maintained since their isolation on the nitrogen free medium, BG11 0 , medium with ferric ammonium citrate replaced with ferric citrate (Gehringer et al. 2010), thereby ensuring their ability to fix nitrogen was not lost. Three months prior to this study, the Nostoc species were also subcultured into BG11 medium containing NaNO 3 (17.6 mM). The aquatic diazotrophic species Nodularia spumigena CCY9414 (Culture Collection Yerseke), Nodularia spumigena NSBL 06 (kindly provided by Hanna Mazur-Marzec, University of Gdansk) and the benthic Nodularia harveyana SAG 44.85 (Culture Collection of Algae, SAG, Georg August University, Göttingen) were cultivated in nitrogen free brackish sea water medium (F/2) with a salinity of 10 containing vitamins (UTEX, Austin). The benthic species, Nodularia harveyana SAG 44.85, required the addition of 5 ml l −1 of soil extract.
Fifty ml of stationary phase cultures were inoculated into 150 ml of the appropriate media in a Fernbach flask (Duran, d = 45 mm) for maximal volume to surface area ratio, and placed at the control or experimental conditions for 14 days to allow them to adjust to their new conditions (Eichner et al. 2014). The inoculum cultures were then diluted 1:1 with fresh medium and divided into two ventilated T 175 polystyrene cell culture flasks (Greiner). In this manner three toxin producing species (n = 3), namely Nodularia spumigena CCY9414, Nostoc punctiforme sp. 73.1 and Nostoc muscorum sp. 65.1 and four control, non-toxin producing species (n = 4), namely Nodularia spumigena NSBL06, Nodularia harveyana SAG 44.85, Nostoc punctiforme sp. 40.5 and Nostoc entophytum sp. C1.8, were studied (Fig. 1). The flasks were laid flat to minimise shading effects, resulting in a culture depth of 1 cm that maximised gas exchange at the culture surface (Herrmann and Gehringer 2019). Experimental cultures were exposed to elevated CO 2 of 2000 ppm (defined here as "High CO 2 "-HC), 10:14 h light:dark cycle, 22 °C, 60% humidity and 130 µmol photons m −2 s −1 (Plant growth chamber E-22L, Percival, USA). Control cultures were exposed to CO 2 at present day level, ~ 440 ppm in Kaiserslautern, Germany (defined here as "Low CO 2 "-LC), with the same culture conditions as above. Different parameters were sampled on days 7-and 14-post inoculation. The numbers of replicates for each sampling day and total numbers are provided in Table 1. In our probe study, a trade off was made between using a variety of species (toxic and non-toxic) from aquatic and terrestrial origin versus increasing the number of replicated incubation bottles per sampling time point. We chose to use a repeated measure approach for most of the parameters with two replicate incubation bottles and two sampling time points. In the case of N 2 fixation, we ended up with technical replicated sampling from each bottle at one sampling time point, applying pseudo-replication in this case.
Determination of growth curves based on optical density were set up separately from the experimental bottles after the adjustment phase with a start inoculum of OD 650 of ~ 0.1 and 200 µl pipetted into 6 wells of a 96 well microtiter plate for each Cyanobacterial culture. Individual Cyanobacteria were grown in their appropriate medium, both diazotrophically and in nitrogen replete medium, at LC and HC. Cultures were resuspended by pipetting and shaking just before the OD 650 was read in a Multiscan Microtitre plate reader (Thermo Scientific) over 14 days post inoculation at experimental conditions identical to the cell culture flasks. To ensure growth rates based on optical density measurements in the 96 well microtiter plate were not biased compared to Fig. 1 Species characteristics and three factorial design chosen for the seven Nostocaceae species tested (a). Factors include the ability to produce nodularin (Factor A); the CO 2 treatment (Factor B) with present day levels (440 ppm-"Low CO 2 " LC) and elevated CO 2 (2000 ppm-"High CO 2 "-HC), and nitrogen availability (Factor C) with cultures grown either diazotrophically or in N-replete medium (N + an N−). The different factorial designs in a) result from the combination of factors A-C for the different species illustrated in (b) the actual experimental set-up in culture flasks, we reputed growth rates determination in culture flasks for 14 days of incubation. Growth curves were therefore set up in culture flasks, in triplicate for terrestrial Nostoc species Nostoc punctiforme sp. 73.1, Nostoc muscorum sp. 65.1, Nostoc punctiforme sp. 40.5 and Nostoc entophytum sp. C1.8 in freshwater growth medium under both diazotrophic (N−) and non-diazotrophic (N+) conditions at LC and HC atmospheric conditions. Similarly, growth curves were established in triplicate for Nodularia spumigena CCY9414 and Nodularia spumigena NSBL06 in brackish sea water growth medium, in both N-replete and N-free medium. T 75 ventilated suspension culture flasks (Sarstedt, Germany) containing 75 ml of the appropriate medium, were inoculated with stationary phase cultures from the respective atmospheres to give a starting Chl a content of 0.1 µg ml −1 . **Readings were repeated on days listed on cultures resuspended in 96 well-microtiter plates every over a period of 14 days. Calculation was done for the period of day 3-14 (10 absorbance readings) Two ml samples were collected from agitated cultures on day 3, 5, 7, 10, 12 and 14 for Chl. a determinations (below). Two ml of culture material was harvested for nodularin analysis on day 14 (below) and a 20 ml volume was centrifuged, drained and dried in a 60 °C oven to obtain the biomass per volume.

Carbonate chemistry
The pH was determined on day 14 from sample filtrates using an electrode (Radiometer analytical PHM210, France) calibrated with a three-point calibration using NBS (National Bureau of Standards) buffers giving values of pH relative to the NBS scale. Total alkalinity (A T ) was determined using the colorimetric SOMMA system according to Johnson et al. (1993). The system was calibrated with carbon reference material provided by A. Dickson (University of California, San Diego) and yielded a precision of about ± 2 μmol kg −1 . Total carbon (C T ) and pCO 2 in the growth media were calculated using CO2SYS (Lewis et al. 1998). Media control carbonate chemistry was similarly assessed with experimentally obtained values for pH and total alkalinity (TA) determined by manual titration (Dickson et al. 2007) and calculated using the Seacarb package in RStudio version 1.0.153, as input data (Hermann and Gehringer 2019).

Net CO 2 uptake/net photosynthesis (NP)
Culture material was removed by pipetting under sterile conditions in a clean bench in LC and HC conditions on days 7 and 14 representing mid to late exponential growth phases. Harvested Cyanobacteria were filtered onto a 3 µM SSWP (Millipore) glass fibre filter (Ritchie 2008), placed onto an appropriate, moist agar plate and incubated under experimental conditions until CO 2 uptake determinations (between 1 and 4 h after sampling) by means of CO 2 gas exchange measurements (GFS 3000, WALZ, Effeltrich, Germany). Bacterial covered filters of LC acclimated cultures were placed in the sample cuvette and CO 2 uptake determined at 80% humidity at 440 ppm (Herrmann and Gehringer, 2019), while for HC acclimated cultures measurements was done at 1500 ppm CO 2 (readings at 2000 ppm were too unstable at 80% humidity in the sample cuvette of the GFS 3000). The respiration rate was determined for each filter after 5 min dark incubation at the start and end of the measuring period to ensure the cultures were not stressed. CO 2 assimilation rates were determined at 500 µmol photons m −2 s −1 (approximate light saturation point for all Cyanobacteria used in this study, determined from light curves), and expressed per µg chlorophyll a. Net photosynthesis (NP) rates, representing the total assimilation of CO 2 minus the CO 2 released during respiration, were calculated.

Chlorophyll a determination
Chlorophyll a was extracted from the bacterial filter discs used for the gas exchange experiments above. After CO 2 uptake measurements, each filter was placed in a 2 ml centrifuge tube containing 100 mg of 0.1 mm zirconia silica beads (BioSpec) to which 1.5 ml 90% HPLC grade methanol was added (Meeks and Castenholz 1971). The samples were bead-beated (Retch, Germany) for 1 min at 30 beats per min and incubated at 4 °C in the dark overnight. Samples were subsequently centrifuged at 10 000 rcf for 5 min at 20 °C and the OD 665 was determined (Lambda 35 UV/VIS spectrometer, Perkin-Elmer). Chlorophyll a content was calculated using the equation: Chl a µg ml −1 = OD 665 × 12.7 (Meeks and Castenholz 1971). Cell pellets obtained from centrifuging two ml of culture were extracted in the same manner to generate the Chl a based growth rates.

Nodularin analysis
Samples for toxin determinations were obtained on day 14 of the growth curves for nodularin producing Cyanobacterial cultures. A 2 ml volume of culture material was centrifuged and the cell pellet drained. One hundred mg of 0.1 mm zirconia silica beads (BioSpec) were added to the pellet with 1.5 ml 70% HPLC grade methanol (Gehringer et al., 2012). The samples were lysed by bead beating as above and incubated at room temperature in the dark overnight. The following morning the samples were vortexed, the lysed cell material removed by centrifugation as above, and the supernatant fluid used in a competitive ELISA assay (Abraxis #522015, Eurofins, Luxembourg) following the manufacturer's instructions. The amount of toxin extracted for each nodularin producing Cyanobacterium under diazotrophic and non-diazotrophic conditions was calculated from the standard curve (R 2 = 0.9937) and expressed as total soluble cellular nodularin content per dry biomass [ng nodularin. µg dry weight −1 ].

Particulate organic matter and N 2 fixation
The PON and POC content were measured for Nodularia cultures and Nostoc punctiforme sp. 73.1 grown in N-free medium on day 14. Due to budget constraints, the remaining Nostoc cultures were not studied. Filters containing culture samples were trimmed, sectioned, then loaded into tin capsules and palletised for isotopic analysis. Measurement was done by means of flash combustion in a Carlo Erba EA 1108 at 1020 °C in a Thermo Finnigan Delta S mass-spectrometer. Calibration material for N and C analysis was acetanilide (Merck). N 2 fixation activity was determined by incubating cultures in two replicates per treatment with bubble addition of 15 N-N 2 enriched gas (99% 15 N 2 ) for 24 h, guaranteeing sufficient dissolution of the 15 N gas in the incubation bottle (Wannicke et al. 2018a). Tracer incubations were terminated by gentle vacuum filtration (< 200 mbar) of the culture material through pre-combusted GF/F filters (Whatman) that were then dried at 60 °C, analysed and the N 2 fixation rates calculated (Montoya et al. 1996). Two technical replicates were conducted per bottle.

Statistical analysis
Statistical analyses were done either by using the Student´s t test or Mann-Whitney Rank Sum Test comparing the mean effect sizes and mean values or by using one-way and repeated measures ANOVA to determine the CO 2 treatment effect. It has to be noted that in the case of Nostoc punctiforme sp. 73.1, only two samples for N 2 fixation were successfully measured. Two replicates were lost due to an autosampler error during processing. No statistical analysis comparing the treatment groups was applied in this case.
Prior to statistical analysis, data were tested for normality and homogeneity of variances using Wilk-Shapiro and Levene's tests. All analyses were performed using the software SigmaPlot 13 (Systat Software Inc., San Jose, CA, USA).

Response ratios and weighted mean effect sizes
To investigate a possible modulating effect of toxin production in response to elevated CO 2 and of elevated CO 2 on volume and dry weight specific nodularin production, we determined the response ratio, i.e. lnRR = x T x C for net photosynthesis, dark respiration, growth and N 2 fixation in selected Cyanobacteria. Here lnRR is the natural-log proportional change in the means ( x ) of the CO 2 treatment (T, i.e. HC) and control group (C, i.e. LC). Negative values of lnRR denote lower rates/ growth at elevated CO 2 compared to control, and vice versa.
To examine the modulating effect of nodularin production over all species tested, pooled lnRR values were combined to give a mean effect size (i.e. Cohen's d). A weight was assigned to each lnRR obtained from individual species which was inversely proportional to its sampling variance (DerSimonian and Laird 1986) as represented by the following equation: d = x T −x C SampleSDpooled . Sub-group calculations were done for the groups "nodularin producer" and "non-nodularin producer" (see Fig. 1 for the toxin status of each Cyanobacterium investigated) and for the nodularin production per volume and per dry weight. To calculate the weighted mean effect sizes, their significance and 95% confidence intervals, a random effect model was applied (DL = DerSimonian-Laird estimator) using Meta-Essential (Suurmond et al. 2017a, b).

Carbonate chemistry
Carbonate chemistry of the media confirmed that experimental application of a continuous atmospheric gas supply ensured enrichment with CO 2 in cultures grown at elevated CO 2 levels. The pCO 2 in the growth media of cultures incubated at HC in the Nostoc cultures was determined to be 1987 ± 42 µatm for N-free media, while the pCO 2 in the LC treatment was 293 ± 38 µatm. Nodularia cultures displayed a mean pCO 2 of 1701 ± 83 µatm in N-replete brackish seawater medium (Suppl. Table 1). Control cultures at LC conditions revealed significantly reduced pCO 2 availability at 232 ± 26 µatm. Control medium at HC was 2728 ± 323 µatm and 2203 ± 405 µatm for fresh and brackish N-free media respectively. Also, carbonate chemistry determined in the experimental bottles showed significant differences between LC and HC treatments (Suppl. Table 1). The N-free media TA values determined agreed with previously published data (Wannicke et al. 2012) in low nutrient media. The TA values for experimental cultures grown at HC in media containing N were also exceedingly high, suggesting interference of biologically synthesised compounds interfering with the TA assessment. These values were therefore not reported.

Effect of elevated CO 2 on Cyanobacterial growth
The growth curves for each species grown under LC and HC conditions highlight the different responses between Cyanobacterial cultures (Suppl. Fig. 1, 2, 3) to atmospheric CO 2 and / or nitrogen availability. The data for Nostoc species 65.1, 40.5 and C1.8, under N limitation, are not presented as the NP readings fell below the level of detection in the gas exchange measurements.
Nostoc punctiforme sp. 73.1 exhibited higher growth rates at HC conditions than at LC, with higher growth rates observed under N replete conditions. Nostoc punctiforme sp. 40.5 did not show a significant response towards HC. On the other hand, both Nostoc muscorum sp. 65.1 and Nostoc entophytum sp. C1.8 displayed lower growth rates at HC (Suppl. Growth rates determined in 96-well microtitre plates and culture flasks are mostly in agreement (Suppl. Fig. 4). In Nostoc species, growth rates calculated from optical density in the 96-well microtiter plate were mostly lower than those derived from tracking Chl a content in the culture flasks, with the largest deviation of ~ 35% in Nostoc entophytum sp. C1.8 and Nostoc punctiforme sp. 73.1 grown at N-replete conditions, at LC and HC respectively. In Nodularia species, the largest deviations in recorded growth rates were 28% for N. spumigena NSBL206 and 17% for N. spumigena CCY9414, grown diazotrophically, at LC and HC respectively, with higher growth rates determined when optical density was used to track growth.

Effect of elevated CO 2 on photosynthesis
A statistically significant increase in net photosynthesis was observed in all species tested at HC compared to the LC control cultures (Fig. 2), except for Nodularia harveyana SAG 44.85 which displayed a significant reduction (p ≤ 0.005) in photosynthesis at HC. This effect was independent of inorganic nitrogen in the growth media being present or absent in Nostoc punctiforme sp. 73.1. Overall, photosynthesis rates pooled for all Nostoc species over all treatments was significantly higher than pooled for Nodularia species (F = 36.3, n = 92/22, p ≤ 0.001). There was a trend towards elevated dark respiration at HC in Nostoc punctiforme sp. 73.1 and Nodularia harveyana SAG 44.85 grown diazotrophically (Suppl. Fig. 5). For Nodularia spumigena NSBL206 this trend was statistically significant (Suppl. Fig. 5).

Effect of elevated atmospheric CO 2 on nodularin production per culture volume
Comparing dry weight volume specific biomass nodularin content revealed a statistically significant impact of HC growth conditions in two of the three toxin producing species, Nostoc punctiforme sp. 73.1 and Nodularia spumigena CCY9414. A significant decrease in nodularin content per dry weight was observed at elevated CO 2 in Nostoc punctiforme sp. 73.1 when grown diazotrophically, while there was no significant impact in nitrogen containing growth media (Fig. 3a). No significant impact of HC on nodularin content was detectable in Nostoc muscorum sp. 65.1, regardless of N availability (Fig. 3b). A significant decrease in dry weight specific nodularin content was apparent for Nodularia spumigena CCY9414 when grown under diazotrophic or N-replete conditions (Fig. 3c) at HC. The effect of elevated CO 2 on nodularin production was strongly influenced by the normalisation of actual toxin concentrations. When comparing volume specific nodularin concentration (ng nodularin per culture volume analysed) a strong increase in nodularin production was observed at elevated CO 2 for Nostoc punctiforme sp. 73.1 under N-replete conditions (N+) when compared to diazotrophically grown Nostoc punctiforme sp. 73.1, while there was only a slight increase for Nodularia spumigena CCY9414 (Suppl. Fig. 6) cultured diazotrophically. Dry weight specific nodularin content was decreased in diazotrophically grown Nostoc punctiforme sp. 73.1 in contrast to the decrease observed in nodularin levels of Nodularia spumigena CCY9414 grown under both nutrient conditions (Fig. 3).

Effect on N 2 fixation
N 2 fixation rates in both Nodularia species were significantly elevated by a factor of 6 for biomass specific rates (t = 3.85, p = 0.001, n = 24 for Nodularia/n = 4 for Nostoc) and 17 for volume specific rates (t = 2.83, p = 0.008, n = 24 for Nodularia/n = 4 for Nostoc, Fig. 4) when compared to rates determined for Nostoc punctiforme sp. 73.1. These rates should be interpreted carefully, especially for Nostoc punctiforme sp. 73.1, due to the low number of replicates (n = 2). Moreover, volume specific N 2 fixation rates were significantly higher for Nodularia spumigena CCY9414 and Nodularia harveyana SAG 44.85 under HC growth conditions when compared to the LC cultures (F = 34.9, p = 0.01, n = 4 and F = 11.7, p = 0.04, n = 4), respectively (Fig. 4a). There was no trend in N 2 fixation in Nodularia spumigena NSBL206.
N 2 fixation rates normalized to particulate organic nitrogen were significantly decreased in the two non-toxin producing Nodularia species, Nodularia harveyana SAG 44.85 and Nodularia spumigena NSBL206 at elevated pCO 2 (HC) (Fig. 4a). This may reflect the significant increase (p ≤ 0.05) in PON levels recorded for all three Nodularia cultures under HC conditions, which were elevated by a factor of 3-6.5 under HC conditions (Suppl. Fig. 7a). The BNF rates determined for the Nostoc punctiforme sp. 73.1 (n = 2) showed no definitive trend at HC growth conditions as seen for Nodularia spumigena CCY9414 and Nodularia harveyana SAG 44.85, at HC.

Modulation of CO 2 response in nodularin and non-nodularin producer
Weighted mean effect sizes for the subgroups, nodularin producer and non-nodularin producer, were positive for NP indicating a positive effect of the high CO 2 treatment on net photosynthesis. Nodularin producing Cyanobacteria displayed a significantly higher positive response towards HC compared to non-nodularin producers (p ≤ 0.006, Student's t test, Fig. 5). Weighted mean effect sizes of dark respiration differed in the two subgroups. Nodularin producers showed no significant effect of HC for dark respiration (overlap of confidence interval with 0-response line, Fig. 5), while non-nodularin producer displayed a positive mean effect size. Weighted mean effect sizes of growth rates showed a significant difference in-between the two subgroups. Mean effect size of non-nodularin producing Cyanobacteria did not deviate from the 0-response line indicating no effect of CO 2 treatment on growth, while the nodularin producers displayed a significant positive response to HC. Contrasting patterns were visible for N 2 fixation, depending on the mode of normalisation of rates, namely volume or PON. Weighted mean effect sizes of biomass specific N 2 fixation showed opposing directions and a significant difference in the subgroups (p ≤ 0.001, Student's t test) with a significant decrease in N 2 fixation at HC in the non-nodularin producer and a no effect on nodularin producer (overlap of confidence interval with 0-response line, Fig. 5). Mean effect sizes of volume specific N 2 fixation rates showed an opposing trend. Non-nodularin producers displayed significantly elevated BNF rates at HC, while nodularin producers displayed a non-significant increase at HC.

Discussion
The recent report by the Intergovernmental Panel on Climate Change indicates that CO 2 emission rates are not being reduced as rapidly as desired, suggesting levels of CO 2 which will most likely exceed 1000 ppm by the year 2100 for the worst-case-scenario (RCP8.5) (IPCC 2019). d represent single measurements. Significant differences between treatments are indicated by *p ≤ 0.05 according to repeated -measure ANOVA Recent reviews have summarised the literature regarding the response of marine and freshwater bloom forming Cyanobacteria to elevated CO 2 levels (Huisman et al. 2018;Visser et al. 2016;Wannicke et al. 2018b), suggesting an increase in bloom formation, possibly favouring diazotrophs, as they would be less susceptible to N-limitation (Gehringer and Wannicke 2014). While numerous studies have been conducted on toxin and non-toxin producing Microcystis aeruginosa strains under elevated CO 2 conditions, little data has been accumulated as to the effects on diazotrophic Cyanobacteria, especially terrestrial Nostoc species (Gehringer and Wannicke 2014). In this study, the effects of elevated atmospheric CO 2 on various growth parameters of six species of diazotrophic, heterocystous Cyanobacteria of two families of the order Nostocales were assessed. The weighted mean effect sizes (Cohens's d) of three nodularin producing Cyanobacteria were compared to four non-nodularin synthesising Cyanobacteria with respect to NP, growth and BNF rates (Fig. 5). Nodularin producers tended to have higher NP rates under HC conditions than non-toxin producers, accompanied by increased growth rates. Non-toxic Cyanobacteria in contrast showed increased respiration at HC conditions, suggesting increased respiratory or oxidative stress. This effect was largely driven by the negative response of Nostoc punctiforme sp. 73.1. Due to the lack of individual species replication, these investigations need to be repeated to make statistically robust conclusions. The BNF rates of both nodularin producers and non -toxic species were unaffected by HC growth conditions when normalised against PON given that the 95% confidence interval crosses the zero line in the Cohens's d plot (Fig. 5), while toxin producers had slightly lower rates. This trend appeared inversed if BNF were normalised against culture volume, where non-nodularin producers were negatively affected by HC, while nodularin producers showed no effect. These trends suggested that nodularin producing diazotrophs did indeed respond differently to changing elevated CO 2 levels compared to non-nodularin producers. We then proceeded to investigate these results in more detail at the individual genus, and species level.

Response of elevated CO 2 on NP and growth
All aquatic and terrestrial diazotrophic Cyanobacteria investigated in this study fixed atmospheric CO 2 in the range of 0.1-1.5 µmol C. ng Chl a −1 s −1 , whether under N-replete or diazotrophic conditions. Additionally, our study has confirmed that most Nostoc species and Nodularia spumigena, grown at 2000 ppm CO 2 , have the capacity for significantly higher NP at HC conditions (Fig. 2), indicating that they are not functioning at saturation under current atmospheric levels of CO 2 . The benthic species, N. harveyana SAG 44.85, however showed a significant reduction in NP with HC. A similarity search conducted using BLASTn (Altschul et al. 1990) on the genome of Nodularia spumigena CCY9414, found that this strain carries a gene (NSP_RS09630) with 75% identitity to the BicA gene of Microcystis aeruginosa PCC7806, thereby suggesting that it can benefit from increased HCO 3 − in the media and thus, increase its NP rates accordingly (Visser et al. 2016). Gas exchange measurements were used to assess NP (Herrmann and Gehringer 2019) as most Cyanobacteria are known to encode the CO 2 converting enzymes, NDH-I 4 and NDH-I 3 (Visser et al. 2016), necessary for the direct, non-energy demanding conversion of CO 2 to bicarbonate for transport to the carboxysome. As non-sequenced environmental isolates were used, there was no information regarding the status of bicarbonate transporters of all the Cyanobacteria under investigation.
Alkalisation of the culture medium of Nostoc species grown under N-replete conditions at both HC and LC culture conditions was recorded, whereas alkalisation was only seen for Nodularia grown under N-replete conditions at LC conditions (Supp. Table 1). The changes in pH do not follow the observed changes in biomass, suggesting species specific responses to changes in pH and dissolved inorganic carbon availability.
In our study, a CO 2 enriched atmosphere led to an increase in inorganic carbon in the control media flasks, while alkalinity was kept constant and pH decreased. Often discussed is the effect of reduced pH by adding acid to keep inorganic carbon stable while total alkalinity and pH decrease. For example, a study by Berge et al. (2010)  , respiration, growth and N 2 fixation for the two subgroups "Nodularin producer" (black circles, for details on Cyanobacteria see Fig. 1a) and "Non-nodularin producer" (white circles, for details on Cyanobacteria see Fig. 1a). Data is presented as means and 95% confidence interval (CI). The horizontal grey line indicates lack of response to the CO 2 treatment (i.e. mean effect size = 0). If the CI crossed the 0 response line, the effect of elevated CO 2 is considered as non-significant. Mean and CI > 0 indicate a stimulation by elevated CO 2 . Mean and CI < 0 indicate a negative effect of elevated CO 2 showed that phytoplankton of the genera dinoflagellates, cryptophytes, diatom and prymnesiophyte were resistant in terms reduced pH and did not increase or decrease their growth rates according to ecological relevant ranges of pH from 7.0 to 9.0. More recently, the response of Raphidiopsis raciborskii to changes in pH and inorganic carbon in water was assessed (Vilar and Molica 2020). The growth of the Cyanobacterium, R. raciborskii was increased with the addition of sodium carbonate and air bubbling, however, saxitoxin production was reduced. Additionally, the authors observed that pH changes were related to significant changes in cellular saxitoxin levels. In general, the potential effect of pH changes on neither the growth, nor nodularin production, of the Nostocales strains investigated in our study, have been characterised and published in the past. Especially the Baltic Sea is depleted with CO 2 during summer time due to the high draw down of dense phytoplankton blooms with high photosynthetic activity (e.g. Huisman et al. 2018), as well as poor diffusion of CO 2 in water and the slow equilibrium between CO 2 and HCO 3 (e.g. Ibelings and Maberly 1998). Photosynthesis in biocrusts is also often limited by CO 2 availability, especially when flooded or desiccated (Tuba et al. 1998;Jauhiainen and Silvola 1999;Lange 2002;Botting and Fredeen 2006;Toet et al. 2006).
The reduction in NP seen for the benthic Nodularia harveyana SAG 44.85, suggests that it may tightly regulate its Ci uptake mechanisms, as elevated CO 2 levels of up to 3000 µatm can occur in benthic layers due to high organic matter decomposition and remineralisation (Haynert et al. 2012). The capacity to tightly regulate C assimilation is an important prerequisite for reducing respiratory stress in this environment. The increase in dark respiration in N. harveyana SAG44.85 suggests it may indeed be under respiratory stress at HC conditions.

Nodularin production under elevated atmospheric CO 2 exposure
Intracellular nodularin content showed significant variation dependent on the Cyanobacterium and/ or atmospheric CO 2 content (Fig. 3). Nostoc sp. 65.1 appears to constitutively produce nodularin at low levels, independent of medium N content and atmospheric CO 2 levels. This Nostoc species also exhibited the lowest growth rate overall (Supplementary  Table 2), which prevented the generation of sufficient biological biomass under diazotrophic growth conditions for NP and BNF determinations. The nodularin content of Nostoc sp. 73.1 was significantly raised under diazotrophic growth conditions at both atmospheres. NP rates (Fig. 2) and growth rates (Supplementary Table 2) of Nostoc sp. 73.1 exceeded those of Nostoc sp. 65.1 under N-replete conditions, factors that may contribute to its higher levels of intracellular toxin production under diazotrophic conditions. Nodularia spumigena CCY 9414 exhibited a very different nodularin synthesis profile, with cultures grown at LC conditions containing significantly more nodularin per dry weight than the cultures in HC conditions, irrespective of medium N content (Fig. 3). The high NP rates observed under HC conditions (Fig. 2), combined with the increased growth rates at HC under diazotrophic conditions suggest that the cells were N depleted and thus not expending resources to produce nodularin. However, the particulate organic nitrogen levels of the HC grown cultures (Supplementary Fig. 7) suggest that the cells were not N-depleted and that some other regulatory mechanism was suppressing nodularin synthesis, compared to the LC cultures.

Biological nitrogen fixation
All Nodularia investigated exhibited significant increases in their culture volume PON content under HC culture conditions ( Supplementary Fig. 7). However, increased BNF rates per volume were only significantly raised (p ≤ 0.05) for Nodularia spumigena CCY 9414 and Nodularia harveyana SAG 44.85 under HC, while those for N. spumigena NSBL06 remained unaffected. If the BNF rates are expressed per PON content, the LC exposed cultures of Nodularia harveyana SAG 44.85 and N. spumigena NSBL06 are significantly higher (p ≤ 0.05) than HC grown cultures (Fig. 4), suggesting inhibition of BNF under HC conditions. A recent study (Boatman et al. 2019) found that dark respiration rates were up to 5 times higher in Trichodesmium erythraeum IMS101cultures exposed to elevated levels of CO 2 (720 μmol mol -1 ). While increases in dark respiration were observed for the non-toxin producing Nodularia spumigena NSBL06 and Nodularia harveyana SAG 44.85 (Supplementary Fig. 5), no increase in BNF rates per PON were recorded (Fig. 4).
Nodularia spumigena CCY9414 and Nodularia harveyana SAG 44.85 exhibited significantly raised volume specific N 2 fixation rates under HC culture conditions, as observed for N. spumigena CCY9414 at 548 ppm CO 2 by Wannicke et al. (2012). Czerny et al. (2009) reported negative effects of HC on cell specific N 2 fixation rates for Nodularia spumigena IOW-2000/1 at 16 °C, whereas Eichner et al. (2014) found no significant changes in the same nodularin producing species exposed to elevated CO 2 through continual bubbling of the cultures. Whether these inconsistencies reflect differences in species selection, culture conditions or method of monitoring of N 2 fixation rates can only be determined with repeat experiments under the identical conditions, preferably related to the ecological environment being investigated. A review of published data of N 2 fixation rates in relation to CO 2 gave evidence for a global positive but non-significant mean effect size for heterocystous species from marine, brackish, and limnic environments (Wannicke et al. 2018b).
When combined with the C assimilation data previously presented, we propose that both the Nodularia spumigena strains are capable of immediately responding to elevated CO 2 levels and rapidly increasing their NP rates (Fig. 2). Additionally, they are capable of increasing their biomass with respect to particulate to PON (Supplementary Fig. 7) in the system under HC conditions, thereby making a significant contribution to the primary productivity in the system. Nodularia harveyana SAG 44.85 appears to be able to regulate NP under HC conditions (Fig. 2), a trait essential for survival in the organic rich benthic zone (Haynert et al. 2012), but still exhibits a significant contribution to the PON (Supplementary Fig. 7) at the elevated CO 2 conditions investigated in this study.
Significantly, the average N 2 fixation rate of the Nostoc punctiforme sp. 73.1 at both LC and HC was 6-17 times lower than that recorded for the Nodularia species investigated, although the low number of repetitions prevents a generalisation of the observed trend (Fig. 4).
This study supports the observation of phenotypic plasticity of carbon fixation rates observed for aquatic freshwater Microcystis cultures grown under elevated CO 2 conditions of 1000 ppm (Ji et al. 2020). While the Nostoc species responded to HC with increased NP rates (Fig. 2), Nostoc punctiforme sp. 73.1 most likely did not invest in the highly energy demanding process of N 2 fixation (Fig. 4) under N limitation as observed for the aquatic Nodularia species studied. To speculate on a general pattern, however, N 2 fixation measurements have to be repeated for all Nostoc species. The PON results (Supplementary Fig. 7) also suggest that Nostoc sp. sp. 73.1 would not increase its contribution to N availability in its direct vicinity, thereby possibly offering an explanation as to the overall reduction in Cyanobacterial biomass observed in dryland soilcrusts exposed to HC of 550 ppm for 10 years (Steven et al. 2012). This negative effect of exposure to HC highlights the complexity of dryland biocrust systems and their response to climate change (Reed et al. 2016) may have been the result of reduced BNF and supply of PON to the system. In contrast, an increase in N 2 fixation was observed in earlier studies in cultures of the Nostoc punctiforme CPCC41 when grown at elevated CO 2 levels of 940 ppm, about half of the CO 2 used in this investigation (Lindo et al. 2017). Further examination of the survival strategies of these important terrestrial primary producers will offer greater insights into the nutrient partitioning and growth strategies, especially under elevated atmospheric CO 2 levels. Additionally, further research into the phenotypic plasticity of carbon fixation within the complex filamentous diazotrophs studied here is crucial to understand the effects of climate change on Cyanobacterial primary productivity under future climate change scenarios.

Conclusion
Our study demonstrates species and strain specific variations to elevated atmospheric CO 2 levels. Interestingly, our data suggests that nodularin producers have, on average, higher NP rates than non-nodularin producers under HC conditions, with lower respiration rates. HC growth conditions induce increases in BNF rates and PON levels per volume of cultures of Nodularia spumigena CCY9414 and N. harveyana SAG 44.85 species, while Nostoc BNF rates are seemingly unaffected. Unexpectedly, the combined BNF of all Nostoc sp. sp. 73.1 determined for LC and HC are significantly lower than those for all Nodularia species tested.
A correlation was observed between HC growth conditions and a decrease in nodularin production under diazotrophic conditions for Nodularia CCY9414 and Nostoc sp. sp. 73.1 (Fig. 3), with Nostoc sp. 73.1 showing increased nodularin content under diazotrophic conditions and Nodularia spumigena CCY9414 under LC conditions. Future studies using similar toxin and non-toxin producing Cyanobacteria for which genomic sequence data exists, need to be undertaken under identical conditions to further elucidate the effects of elevated CO 2 on Cyanobacterial cellular metabolism, and the role of secondary metabolites, like nodularin, in mediating the cellular responses to future climate change conditions. This study would suggest that toxinproducing diazotrophs may be less advantaged under current climate change predictions in diazotrophic conditions, due to impaired N 2 fixation under elevated CO 2 conditions, when compared with similar non-toxin producing species of Cyanobacteria. On the other hand, a higher positive response in NP may outbalance this effect at elevated CO 2 .