Mixotrophic growth of the extremophile Galdieria sulphuraria reveals the flexibility of its carbon assimilation metabolism

Summary Galdieria sulphuraria is a cosmopolitan microalga found in volcanic hot springs and calderas. It grows at low pH in photoautotrophic (use of light as a source of energy) or heterotrophic (respiration as a source of energy) conditions, using an unusually broad range of organic carbon sources. Previous data suggested that G. sulphuraria cannot grow mixotrophically (simultaneously exploiting light and organic carbon as energy sources), its photosynthetic machinery being repressed by organic carbon. Here, we show that G. sulphuraria SAG21.92 thrives in photoautotrophy, heterotrophy and mixotrophy. By comparing growth, biomass production, photosynthetic and respiratory performances in these three trophic modes, we show that addition of organic carbon to cultures (mixotrophy) relieves inorganic carbon limitation of photosynthesis thanks to increased CO2 supply through respiration. This synergistic effect is lost when inorganic carbon limitation is artificially overcome by saturating photosynthesis with added external CO2. Proteomic and metabolic profiling corroborates this conclusion suggesting that mixotrophy is an opportunistic mechanism to increase intracellular CO2 concentration under physiological conditions, boosting photosynthesis by enhancing the carboxylation activity of Ribulose‐1,5‐bisphosphate carboxylase‐oxygenase (Rubisco) and decreasing photorespiration. We discuss possible implications of these findings for the ecological success of Galdieria in extreme environments and for biotechnological applications.


Introduction
The unicellular red alga Galdieria sulphuraria belongs to the Cyanidiophyceae, a class that includes five species often flourishing in different extreme environments (Merola et al., 1981;Gross et al., 1998;Gross & Oesterhelt, 1999;Oesterhelt et al., 2007). From a phylogenetic perspective, plastids of the red algae gave rise to the complex plastids of, e.g. diatoms via secondary endosymbiosis (Yoon et al., 2002Bhattacharya et al., 2003). Like other members of this class (Doemel & Brock, 1971;Reeb & Bhattacharya, 2010) G. sulphuraria has an extremophile lifestyle, withstanding low pH (pH optimum at 2) and elevated temperatures (up to 56°C). It thrives in soils and forms biomats on rocks surrounding hot springs, fumaroles, or acid mining sites and even on burning coal spoil heaps (Moreira et al., 1994;Gross et al., 1998;Castenholz & McDermott, 2010;Barcyt e et al., 2018). Some mesophilic species have been isolated from environments with moderate temperatures and/or a neutral pH (Gross et al., 2002;Yoon et al., 2006;Az ua-Bustos et al., 2009;Iovinella et al., 2018). Genome analysis (Barbier et al., 2005;Schonknecht et al., 2013;Rossoni et al., 2019a) has pinpointed a very high metabolic flexibility of this alga, which is confirmed by its ability to grow in photoautotrophy (exclusive use of light as an energy source) and heterotrophy (organic carbon respiration of more than 50 different substrates (Gross & Schnarrenberger, 1995)). This capacity, along with the peculiar pH optimum for growth, allow G. sulphuraria to be cultivated in open ponds containing organic matter, overcoming other microorganisms, considered as contaminants in this case. Given these advantages for large-scale cultivation, G. sulphuraria is considered an emerging system for biotechnology applications (Schmidt et al., 2005;Henkanatte-Gedera et al., 2017;Cizkova et al., 2019).
The relationship between photosynthesis and glycolysis/respiration in higher plants and microalgae is complex (Avelange et al., 1988;Kromer et al., 1988;Gemel & Randall, 1992;P€ arnik & Keerberg, 1995;Hoefnagel et al., 1998;Tcherkez et al., 2008). Photosynthesis/glycolysis/respiration interactions are prone to perturbation by mixotrophy, in which external organic carbon often interferes with carbon flow between chloroplasts, the cytosol and mitochondria. Several green microalgae are capable of mixotrophic growth (Combres et al., 1994;Wan et al., 2011;Johnson & Alric, 2012;Cecchin et al., 2018). While mixotrophy has always beneficial consequences on respiration, its effects on photosynthesis differ depending on the microalga considered: enhancement of photosynthesis was reported in one case (Ettlia oleoabundans (Ferroni et al., 2018)), while in other algae, including the diatoms Phaeodactylum tricornutum (Liu et al., 2009;Villanova et al., 2017) and Nannochloropsis (Fang et al., 2004;Xu et al., 2004) photosynthesis was unaffected. Decreased photosynthetic activity in mixotrophy has been reported in Chlorella vulgaris (Martinez & Orus, 1991;Cecchin et al., 2018) and Chlamydomonas reinhardtii where the carbon concentrating mechanism (CCM (Bogaert et al., 2019)) and the light harvesting capacity (Perrineau et al., 2014) is decreased by acetate along with the enhancement of respiration. While it has been reported that G. sulphuraria 074G could grow in the simultaneous presence of light and a carbon source, heterotrophy seemed to prevail in these conditions, as no photosynthetic oxygen (O 2 ) production could be measured in the presence of glucose (Oesterhelt et al., 2007).
Here, we show instead that photosynthesis and carbon metabolism (glycolysis and respiration) operate simultaneously in Galdieria sulphuraria SAG21.92 (a close relative of the 074G strain) under mixotrophic conditions, provided that the temperature conditions are kept close to the ones experienced by this alga in its natural environment. We show that performances in mixotrophy, exemplified by photosynthetic activity and biomass production, actually exceed the sum of the heterotrophic and photoautotrophic yields under limiting inorganic carbon. This synergistic effect stems from a stimulation of photosynthesis by CO 2 of respiratory origin, which overcomes the limitation to the Calvin-Benson-Bassham cycle and suppresses photorespiration.
Limitation originates from the very low inorganic carbon concentration available in acidic conditions (pH 2) that constitute the alga's natural growth environment. Notably, under these conditions, inorganic carbon is almost exclusively available as dissolved CO 2 (around 10 µM) while soluble bicarbonate is virtually absent. Consistent with this hypothesis, the synergistic effect of 'light' and 'dark' energetic metabolisms is sensitive to respiration inhibitors and is lost upon addition of exogenous CO 2 , which outcompetes endogenous CO 2 of respiratory origin in relieving inorganic carbon limitation of the Calvin-Benson-Bassham cycle. We conclude that mixotrophy constitutes an efficient mechanism to increase intracellular CO 2 concentration under physiological conditions, allowing G. sulphuraria to successfully exploit all the energy resources available for growth in its rather challenging environment.

Strains, growth and media composition
Galdieria sulphuraria SAG21.92 and 074G were obtained from the Culture Collection of Algae at G€ ottingen University (SAG), Germany, and were grown in sterile 2 9 GS modified Allen medium, pH 2.0, containing 20 mM of sodium nitrate (NaNO 3 ), and 5 mM of inorganic phosphate (K 2 HPO 4 and KH 2 PO 4 in a 2 : 1 ratio (Allen, 1959)) at 42°C without or with organic substrates as indicated in the text. The concentration of organic substrates was selected on the basis of data reported in the literature (Oesterhelt et al., 2007). More precisely, glucose was employed at the concentration of 25 mM, as in Oesterhelt et al. (2007). The concentration of all the other organic compounds was adjusted to reach the same carbon atom concentration (150 mM). Galdieria sulphuraria was grown either in 250 ml flasks (50 ml culture volume), in an incubator (Infors, Bottmingen-Basel, Switzerland, continuous light, 30 µmol photons m À2 s À1 , 42°C, 100 rpm) or in a photobioreactor (Multicultivator; Photon System Instruments, Dr asov, Czech Republic). Inside the multicultivator, cells were provided with air or CO 2enriched air by active bubbling (see Supporting Information Methods S1). Moreover, the incident light intensity was adjusted daily to maintain constant transmitted light through the culture (see Results section). This 'luminostat' regime ensures maximal absorption of light without allowing a dark zone to develop inside the photobioreactor (Cuaresma et al., 2011). Growth was monitored daily by cell counting with a LUNA cell counter (Logos Biosystems Inc., Annandale, VA, USA). Sorbitol consumption was measured using the D-sorbitol/xylitol assay kit (Megazyme, Bray, Ireland).

Cell fresh weight and dry weight quantification
Cell pellet was resuspended in a small volume of water and centrifuged in pre-weighed Eppendorf tubes and the pellet was weighed. For dry weight determination fresh cell pellets were dried for three days at 60°C, weighed and expressed as g l À1 .

Clark electrode oxygen measurements
Oxygen exchanges in solution were measured with a Clark-type electrode (Hansatech Instruments, King's Lynn, UK) at 42°C. Respiration and gross photosynthesis were quantified by measuring the slope of O 2 changes in the dark and under light exposure, respectively. Net photosynthesis was calculated assuming O 2 consumption by the mitochondrion in the light is identical to that in the dark (Net photosynthesis = v O2light + |v O2dark |).

Photophysiology measurements
Photosynthetic parameters were derived from quantification of chlorophyll fluorescence emission by cultures within the multicultivator. To this aim, we employed a custom-made fluorescence imaging system based on a previously published setup (Johnson et al., 2009) modified as described in Methods S1. The photosynthetic electron transfer rate (ETR) (a proxy of the carbon assimilation capacity (Maxwell & Johnson, 2000)) was calculated as (Fm 0 À Fs)/Fm 0 9 PFD, where Fm 0 and Fs are the fluorescence intensities measured after exposure to a saturating pulse and in steady state, respectively, in light-acclimated cells and PFD (photosynthetic flux density) is the incident light intensity, measured in µmol photons m À2 s À1 . The cells were allowed to reach steady-state fluorescence emissions at each intensity (5-10 min of light exposure depending on the intensity) before increasing the photon flux.

Biochemistry and proteomic analysis
Western blot analysis was performed on cells grown for 7 d in the indicated conditions. Cells (10 9 cells) were broken with a Precellys homogenizer (Bertin, Beaumont-Village, France), through three cycles of 30 s at 10 000 rpm separated by a 30-s interval. Total protein extracts were analyzed by immunoblotting with an anti-Ribulose-1,5-bisphosphate carboxylase-oxygenase (anti-Rubisco) antibody (Agrisera, V€ ann€ as, Sweden). An antibody against the b subunit of the ATPsynthase complex (Agrisera) was used as a loading control. 10 µg of protein was loaded per well.
For proteomic analysis, algae were cultivated under the three conditions photoautotrophy, mixotrophy and heterotrophy in parallel in the same cultivator. Three multicultivator experiments were carried out one week apart and constituted the three independent biological replicates. Cells were collected on day 7 (i.e. 4 d after addition of 25 mM D-sorbitol to mixotrophic and heterotrophic cultures). Proteins from whole cell extracts (40 µg each) were solubilized in Laemmli buffer before being stacked in the top of a 4-12% NuPAGE gel (Life Technologies, Waltham, MA, USA), stained with R-250 Coomassie blue (Bio-Rad, Hercules, CA, USA) and in-gel digested using modified trypsin (sequencing grade; Promega, Madison, WI, USA) as previously described (Bouchnak et al., 2019). Resulting peptides were analyzed by online nanoLC-MS/MS (Ultimate 3000 RSLCnano coupled to Q-Exactive HF; ThermoFisher Scientific, Waltham, MA, USA) using a 200-min gradient. Peptides and proteins were identified using MASCOT (v.2.6.0, Matrix Science). Spectra were searched against Uniprot (G. sulphuraria taxonomy, July 2019 version, 7347 sequences) concomitantly with a home-made list of contaminants frequently observed in proteomics analyses (trypsins and keratins, 250 sequences).
The PROLINE software (Bouyssie et al., 2020) was used to filter the results (conservation of rank 1 peptides, peptide identification false discovery rate (FDR) < 1% as calculated on peptide scores by employing the reverse database strategy, minimum peptide score of 25, and minimum of one specific peptide per identified protein group). Proline was then used to extract the MS1-based intensities values of protein groups from unique peptides. Proteins identified in the reverse and contaminant databases (i.e. trypsin or keratin), and proteins identified with only one peptide with a score < 40 were further discarded from the list. Proteins identified in only one or two conditions were kept for analysis without statistical treatment. Proteins identified in all three conditions were submitted to statistical differential analysis using PROSTAR (Wieczorek et al., 2017;Wieczorek et al., 2019). Detailed procedures are described in Methods S1.

Metabolite extraction and analysis
Cells grown in a multicultivator were harvested at day 5 by centrifugation (4°C, 5 min and 3000 rcf), washed with ice-cold 0.9 % (w/v) sodium chloride (NaCl), snap frozen in liquid nitrogen, and lyophilized overnight. Cells (5 9 10 8 ) were disrupted in a Mixer Mill (MM 400; Retsch GmbH, Haan, Germany) for 60 s at 30 Hz using metal beads and 500 µl of ice-cold chloroform and methanol (1 : 2.3 ratio; containing 5 µM Ribitol and 2,4dimethylphenylalanine (both from Sigma-Aldrich, Munich, Germany) as internal standards) with another round of shaking in the Mixer Mill. After a 2 h incubation at À20°C, 400 µl of icecold deionized water (Milli Q; Merck Chemicals GmbH, Darmstadt, Germany) were added to induce phase separation. The samples were vortexed and centrifuged for 5 min at 4°C and 16 000 rcf. The aqueous phase was transferred to a new reaction tube and the organic phase was re-extracted with 400 µl of icecold deionized water. Aqueous phases of each sample were combined and lyophilized overnight. After resuspension in 500 µl of 50% methanol 50 µl were dried in a glass inlet for analysis by gas chromatography-mass spectrometry (GC-MS) and ion chromatography-mass spectrometry (IC-MS).
For GC-MS analysis the samples were prepared and analyzed as described by Gu et al. (2012) and Shim et al. (2020). Identification of metabolites was performed with MASSHUNTER Qualitative (v.b08.00; Agilent Technologies, Santa Clara, CA, USA) by comparing spectra to the NIST14 Mass Spectral Library (https://www. nist.gov/srd/nist-standard-reference-database-1a-v14) and to a quality control sample containing all target compounds. Peaks were integrated using MASSHUNTER Quantitative (v.b08.00; Agilent Technologies). For relative quantification, all metabolite peak areas were normalized to the peak area of the internal standard ribitol.
For IC-MS a combination of a Dionex ICS-6000 HPIC and a high field Thermo Scientific Q Exactive Plus quadrupole-Orbitrap mass spectrometer following the method described in Schwaiger et al. (2017) and in Methods S1. Data analysis was conducted using COMPOUND DISCOVERER (v.3.1, ThermoFisher Scientific). Setting parameters for the untargeted metabolomics workflow and peak annotation criteria are described in Methods S1.

Labeling experiments with 13 C-glucose
Cells were cultivated in 250 ml Erlenmeyer flasks (50 ml culture volume) for 4 d under continuous light at 60 µmol m À2 s À1 , 40°C and ambient air (0.04% CO 2 ). U-13 C 6 -glucose (Cambridge Isotope Laboratories Inc., Tewksbury, MA, USA) was added at day 4 in a final concentration of 25 mM and the irradiance was increased to 100 µmol m À2 s À1 either under ambient or elevated (2%) CO 2 conditions. Next, 1-2.5 9 10 8 cells were harvested 1, 4, 12, 24, 36, 48 and 60 h after glucose addition as described earlier. Metabolites were extracted and measured by IC-MS as described earlier.
Data analysis was conducted with COMPOUND DISCOVERER (v.3.1, ThermoFisher Scientific) and the standard workflow for stable isotope labelling from COMPOUND DISCOVERER was chosen. The default settings, which are 5 ppm mass tolerance, 30% intensity tolerance and 0.1% intensity threshold for isotope pattern matching were used and the maximum exchange rate was set to 95%.

Galdieria sulphuraria metabolizes several organic substrates
Previous work (Oesterhelt et al., 2007) has suggested that G. sulphuraria is unable to grow mixotrophically in the presence of light plus organic carbon but rather alternates between heterotrophy (in the presence of an external source of organic carbon) and phototrophy (upon inorganic carbon consumption). This conclusion was based on experiments carried out at 25°C, i.e. a temperature that is far from the physiological optimum of this alga (above 40°C), and therefore decreases photosynthetic performances (Doemel & Brock, 1971;Ford, 1979;Rossoni & Weber, 2019;Rossoni et al., 2019b). Thus, we decided to reinvestigate the possible occurrence of mixotrophy under conditions that resemble natural growth conditions (42°C, pH 2.0) in which G. sulphuraria displays maximum photosynthetic capacity.
First, we sought compounds that could improve algal growth in presence of light. We found that several hexoses, disaccharides and pentoses, but also some polyols and amino acids (Lalanine, L-glutamate) were able to boost G. sulphuraria's growth when compared to strict photoautotrophic conditions (Fig. S1). Conversely, some organic acids (malic and citric acid) and amino acids (L-aspartate, L-leucine, L-valine, L-isoleucine, L-asparagine) had either no effect or led to a growth inhibition when compared to photoautotrophic conditions. Acetic acid was lethal to the cells. Compound concentrations were 25 mM, i.e. the same value employed in the study by Oesterhelt et al., 2007. In the case of disaccharides, we reduced the concentration by a factor of two, to keep the overall carbon atoms concentration constant.
Based on these results, we focused on sorbitol, a compound that boosts growth in the dark without inducing a significant loss of photosynthetic pigments (Gross & Schnarrenberger, 1995). We confirmed that sorbitol was able to sustain cell division in the dark (Fig. 1), but its effect on growth was largely enhanced by a concomitant exposure to light. The final dry weight and number of cells collected at the end of the exponential phase in the light plus carbon condition (orange symbols, Fig. 1) exceeded the sum of cells obtained in the heterotrophic (black symbols, Fig. 1) plus the photoautotrophic (green symbols, Fig. 1) conditions. This finding indicates that G. sulphuraria is not only able to perform true mixotrophy under high temperature, dim light and presence of external organic carbon sources, but that this trophic mode is highly beneficial for its growth capacity. We obtained similar effects replacing sorbitol with a monosaccharide (glucose) and a disaccharide (saccharose) (Fig. S2).

Respiration boosts photosynthesis in mixotrophic Galdieria cells
To further characterize the consequences of mixotrophy in G. sulphuraria, we transferred half of the cells grown in mixotrophy ** ** 0.18 ± 0.04 1.40 ± 0.06 2.00 ± 0.18 1.69 ± 0.04 Growth enhancement of Galdieriasulphuraria SAG21.92 by reduced carbon sources is light dependent. (a) Growth curves. Data from three biological replicates AE SD. Error bars are shown when larger than the symbol size. Galdieriasulphuraria was grown in flasks at ambient CO 2 in photoautotrophic (light only, 30 µmol photons m À2 s À1 , green symbol), mixotrophic (30 µmol photons m À2 s À1 plus D-sorbitol 25 mM, orange symbol) and heterotrophic (absence of light, presence of D-sorbitol 25 mM, black symbol). The initial cell concentration was 1.5 9 10 6 cells per milliliter. At day 5, the mixotrophic culture was split in two parts, and light was switched off in one culture (gray symbol). Growth was carried out at 42°C with shaking at 100 rpm, pH 2. (b) Dry weight estimated at day 9. Mixotrophic biomass (orange bar) exceeds the sum of photoautotrophic (green bar) and heterotrophic (black bar) biomass, highlighting the existence of a synergy under mixotrophic conditions. Data from three biological replicates AE SD. **Indicates that at the 0.01 level the means of the two populations (mixotrophy on one side; heterotrophy + photoautotrophy on the other one) means are statistically different (ANOVA test). The concentration of inorganic nitrogen was 20 mM, while that of inorganic phosphate was 5 mM.  Galdieriasulphuraria SAG21.92 cells and biomass production. Cells were inoculated at 3.5 9 10 6 cell per milliliter and grown in a photobioreactor in the light (transmitted light 10 µmol photons m À2 s À1 ) and air before D-sorbitol was added in the absence (heterotrophy, black) and in the presence of light (mixotrophy, orange). Light was increased every day to keep the transmitted light to a constant value of 10 µmol photons m À2 s À1 . Growth was followed at 42°C and pH 2. (a, d, g) After 5 d of growth (i.e. 2 d after the addition of D-sorbitol), ETR was measured directly on cultures within the photobioreactor, to avoid possible temperature stress. Measurements were done in air, in the absence (a, b) and presence (d, e) of respiratory inhibitors (SHAM (1 mM) and myxothiazol (10 µM), added 24 h before measurements), or in a CO 2 -enriched (0.5%) atmosphere (g, h). (a, d, g) Photosynthetic electron transfer: data from three biological replicates AE SD. (b, e, h) Biomass production in photoautotrophic (green, data from 12 biological replicates AE SD), heterotrophic (black, data from eight biological replicates AE SD) and mixotrophic (orange, data from eight biological replicates AE SD) conditions, respectively. Cells were collected after 7 d of growth (i.e. 4 d after addition of D-sorbitol). (c, f, i) Sketches representing possible CO 2 sources for photosynthesis in the three examined conditions. **Indicates that at the 0.01 level the means of the two populations (mixotrophy on one side; heterotrophy + photoautotrophy on the other one) means are statistically different (ANOVA test).

Research
New Phytologist for 5 d (Fig. 1) to heterotrophic condition and monitored growth of the four samples (photoautotrophic: green bar; heterotrophic: black bar; mixotrophic: orange bar; mixotrophic transferred to heterotrophy: gray bar) for another 5 d. While the mixotrophic cells continued to display a higher growth capacity than the heterotrophic plus photoautotrophic ones, the mixotrophic cells transferred to the dark (gray symbols) slowly reached the same cell density (Fig. 1a) and dry weight (Fig. 1b)  culture, i.e. they lost the benefits provided by the simultaneous exposure to light and organic carbon within a few days. We conclude therefore that mixotrophy promotes a synergistic interaction between light and dark energy metabolisms, which slowly disappears when the light supply is halted.
In the experiments described earlier, photoautotrophic growth was most likely limited by the low light intensity (30 µmol photons m À2 s À1 ), which was kept constant during growth, and therefore rapidly became limiting for photosynthesis when the cell concentration was increased in the flasks. Gas diffusion could also be limiting in flasks. Therefore, we repeated experiments in a photobioreactor (Fig. S3), where air bubbling ensured a more efficient gas delivery to the algae. Moreover, the intensity of the incident light was progressively increased in this setup to maintain a linear relationship between the cell number and the absorption. We reproduced the improvement of biomass productivity in mixotrophy with the photobioreactor (Fig. 2), where we could monitor biomass production and photosynthetic performances on actively growing cells using a custom-built fluorescence imaging setup to monitor photosynthesis via the ETR parameter (Maxwell & Johnson, 2000) (Fig. S3).
The ETR in the photoautotrophic cultures was lower than in mixotrophic ones (Fig. 2a). The difference was larger during the first days of culturing (Fig. S4, days 4 and 5, i.e. the first and second day after sorbitol addition), and then photosynthesis progressively diminished in the mixotrophic cells (Fig. S4, days 6 and 7, i.e. the third and fourth day of mixotrophy), where we also observed a large variability in the photosynthetic capacity (Fig. S4, day 7). We ascribe this variability to a differential consumption of sorbitol, and therefore of the mixotrophic synergistic effect, in the various samples. Consistent with this hypothesis, enhancement of photosynthesis was re-established in mixotrophic cultures, provided that sorbitol, nearly exhausted after 4 d (Fig. S5, day 7) was added again to the growth medium (Fig. S6).
To explain the synergistic effect of mixotrophy, we reasoned that photosynthesis could be limited by the availability of inorganic carbon in photoautotrophic cells. The latter is present only as CO 2 and at very low concentration (10 µM (Gross et al., 1998)) at ambient air in G. sulphuraria's cultures, due to the acidic pH (pH 2). Mixotrophy could alleviate this limitation by providing extra CO 2 of mitochondrial origin via enhanced respiration (Figs 2c, S7). We tested this hypothesis by two approaches. First, we poisoned mixotrophic and photoautotrophic cultures with myxothiazol and SHAM, known inhibitors of the cyanide sensitive and insensitive respiratory pathways, respectively, and measured consequences on photosynthetic activity. Addition of these inhibitors completely abolished the enhancement of photosynthetic activity and of biomass productivity by mixotrophy (Fig. 2d-f). Next, we increased the CO 2 availability to the cells in mixotrophic cultures, to outcompete endogenous CO 2 of respiratory origin with an excess of exogenous inorganic carbon. As a prerequisite for this experiment, we calibrated the CO 2 requirement for optimum photosynthesis in our growth conditions (transmitted light of 10 µmol photons m À2 s À1 ). We found that photoautotrophic growth was increased by CO 2 up to a concentration of 2% CO 2 (Fig. S8a), the apparent affinity for CO 2 being about 0.5%. Upon addition of external CO 2 , the photosynthetic capacity in photoautotrophic conditions increased and the biomass produced in heterotrophic plus photoautotrophic conditions became equal to that observed in mixotrophy (Fig. 2g-i). Overall, these data indicate that the synergy between respiration and photosynthesis is lost when the respiration is inhibited or when the photosynthesis becomes saturated with externally supplied CO 2 . Moreover, we hypothesize that photorespiration should be decreased to very low rates at saturating CO 2 , possibly contributing to the gain in biomass productivity at high CO 2 concentrations. In these experiments, we also observed that the amount of biomass produced by phototrophic cells supplemented with air was the same irrespective of the light intensity employed (Fig. S8b). Conversely, biomass production could be increased by increasing the light intensity in both phototrophic cells supplemented with 0.5% CO 2 (Fig. S8c) or in mixotrophic cells (Fig. S8d). These findings are fully consistent with the hypothesis that photosynthesis is CO 2 limited in air, and that this limitation can be alleviated by endogenous (mixotrophy) or exogenous CO 2 .

Metabolic acclimation to mixotrophy in G. sulphuraria
The observation that mixotrophy enhances photosynthesis to a similar level as upon addition of external CO 2 to G. sulphuraria, suggests that mixotrophy behaves as a strategy to traffic CO 2 from the mitochondria to the plastid, allowing this alga to successfully exploit all the energy resources available for growth and to minimize energy loss through photorespiration.
To substantiate this hypothesis, we performed a complete survey of the metabolic changes between the three trophic lifestyles, and relate these changes to modifications in the cell proteome. We found that photosynthetic proteins (complexes of the electron transfer chain, enzymes of the Calvin-Benson-Bassham cycle and transporters involved in triose phosphate export) were downregulated in mixotrophic conditions compared to photoautotrophic conditions (Fig. 3), in agreement with previous suggestions (Gross & Schnarrenberger, 1995;Oesterhelt et al., 2007). Conversely, respiratory protein levels remained relatively constant in the three conditions. This observation (photosynthetic activity is enhanced in mixotrophy despite the downregulation of the photosynthetic machinery) fully supports the conclusion that intracellular increase in CO 2 due to enhanced respiratory activity more than compensates for the decrease in Rubisco (see also Fig. S9).
In line with the hypothesis that mixotrophy channels CO 2 from respiration to photosynthesis, proteomic analysis indicates that all the enzymes involved in photorespiration (phosphoglycolate phosphatase, glycolate oxidase, serine-glyoxylate aminotransferase, glycine decarboxylase, glycine/serine hydroxymethyltransferase, hydroxypyruvate reductase, glycerate kinase) were less abundant in mixotrophy. This is also true for Rubisco activase (Gasu_19410; Datasets S1, S2), an enzyme that was shown to be important under low CO 2 in Chlamydomonas (Pollock et al., 2003). These findings are also corroborated by metabolite analysis (Fig. 4a). The amounts of the oxygenation product of Rubisco, 2-phosphoglycolate are highest under photoautotrophic conditions, lower in mixotrophic conditions, and very low in heterotrophic conditions. Glycine accumulates, likely due to a higher reduction potential inside the mitochondrial matrix under mixotrophic conditions, which will reduce the rate of oxidative decarboxylation of glycine by glycine decarboxylase. Consequently, the glycine to serine ratio was inverted compared to photoautotrophic conditions. The amounts of glycerate and 3-phosphoglycerate under mixotrophic conditions are mimicking those in heterotrophically grown cells (Fig. 4a).
Proteins involved in a putative C4-like carbon concentrating cycle (Rademacher et al., 2017) (Figs 3, 4b) followed the same pattern as the photorespiratory enzymes. Carbonic anhydrase and phosphoenolpyruvate carboxylase were more abundant under photoautotrophic conditions than under mixotrophic or heterotrophic conditions, and, in agreement with this finding, photoautotrophic cells also contained higher amounts of phosphoenolpyruvate (PEP) than mixotrophic or heterotrophic cells (Fig. 4b). The presence of this carbon concentrating cycle in photoautotrophic conditions could be supported by increased steadystate levels of phosphorylated C3 compounds, as suggested by the downregulation of pyruvate kinases in phototrophy (PEP consumption might be decreased) and strong induction of pyruvate ortho-phosphate dikinase (producing PEP). We assume that oxaloacetate (OAA) under photoautotrophic conditions is decarboxylated by PEP carboxykinase (PEPCK), and not after reduction into malate by malate dehydrogenase and decarboxylation by mitochondrial malic enzyme. This hypothesis is supported by the finding that mitochondrial malic enzyme (ME) is strongly decreased in photoautotrophy (Fig. 3) and the malate pool size is smaller under this condition than in mixotrophy or heterotrophy. We note that decarboxylation by PEPCK (as compared to ME) is energy conserving and directly yields PEP for a new round of carboxylation by PEP carboxylase. While part of the increase in biomass production in mixotrophic conditions can be attributed to the repression of photorespiration, other benefits of the mixotrophic lifestyle may come from more carbon units being shuttled into anabolic pathways. Indeed, we observed a general increase of metabolites of the oxidative pentose phosphate pathway (ribulose-5P, sedoheptulose-7P) and nucleoside triphosphates (adenosine triphosphate (ATP), guanosine triphosphate (GTP), uridine triphosphate (UTP), cytidine triphosphate (CTP)) as successor metabolites (Fig. 4c). The amounts do not differ between autotrophic and heterotrophic cultures but are clearly increased under mixotrophic conditions providing evidence for a higher flux of carbon at least into purine and pyrimidine synthesis.
Labeling experiments using carbon-13 ( 13 C)-labeled glucose, a sugar with similar effects on cell growth and biomass production as sorbitol (Fig. S2), further support this notion. We found a rapid incorporation of CO 2 from 13 C glucose into photosynthetic metabolites, such as ribulose 1,5-bisphosphate (RuBP), sedoheptulose 1,7-bisphosphate (SBP), and 2-phosphoglycolate (2-PG) (Fig. 5). While RuBP also occurs in the oxidative pentose phosphate pathway, SBP and 2-PG are metabolites solely formed in the Calvin-Benson-Bassham cycle and can only carry a label when 13 CO 2 released by respiration is fixed. Incorporation of the label by shuttling of carbon backbones from the cytosol into the chloroplast is unlikely since the solute transporters in the G. sulphuraria's chloroplast envelope do not favor the import of glycolytic intermediates under photosynthetic conditions (Linka et al., 2008). The rate of incorporation of 13 C is much reduced when external CO 2 levels are increased to 2%. Interestingly, the tricarboxylic acid cycle intermediate succinate exhibits a similar pattern of delayed labeling under high CO 2 conditions when compared to ambient air, indicating a slowed down glucose usage. Thus, the delay of label incorporation into the Calvin-Benson-Bassham cycle intermediates can be traced back to lower levels of labeled respiratory CO 2 under CO 2 saturated conditions.

Discussion
At variance with a previous report (Oesterhelt et al., 2007), G. sulphuraria cells are capable of a true mixotrophy, when transferred from strictly photoautotrophic conditions to a light and

Research
New Phytologist organic carbon regime, provided that the temperature conditions are kept close to the ones experienced by this alga in its natural environment. This is not only true for the G. sulphuraria SAG21.92 strain used here, but also for G. sulphuraria 074G, i.e. the strain used in the previous study by Oesterhelt et al. (2007) when tested in the presence of the same external organic carbon sources employed here (see Fig. S10). Acclimation of the strictly photoautotrophic Cyanidiophyceae Cyanidioschyzon merolae to suboptimal growth temperatures of 25°C, led to massive rearrangements of the photosynthetic apparatus and a lower photonto-oxygen conversion rate when compared to cells grown at 42°C (Nikolova et al., 2017). Also, cultivation of G. sulphuraria at 25°C led to a downregulation of transcripts encoding components of the photosynthetic machinery (Rossoni et al., 2019b). Therefore, cultivation at suboptimal temperatures may not be in favor of high photosynthetic rates and drive Cyanidiophyceae species capable of heterotrophic growth towards this trophic state.
Mixotrophy deeply alters the carbon metabolism of G. sulphuraria cells. Under photoautotrophic conditions, CO 2 is mainly concentrated by the PEPC/PEPCK driven CCM (Rademacher et al., 2017). This process is repressed in mixotrophy. Moreover, enhanced respiration relieves the limitation of photosynthesis by inorganic carbon, which is at a low concentration (µmolar range) in the acid growth medium of this alga. We can quantify the extent of this process using the Clark electrode (as in Fig. S7) in a closed configuration, to avoid gas exchanges with the atmosphere (representative traces in Fig. S11a,c). In this case, the amount of O 2 produced by photosynthesis (32.3 AE 7.5 µM and 76.6 AE 2 µM in photoautrotrophy and mixotrophy, respectively) is commensurate with the amount of CO 2 available to Rubisco, i.e. the sum of respiratory CO 2 (22 AE 10.5 µM and 61 AE 1.5 µM respectively, assuming a 1 : 1 stoichiometry with consumed O 2 ) plus the small CO 2 amount present in the medium (about 10 µM at pH 2) (Fig. S11c,d). These estimates clearly indicate that respiration in mixotrophy sets the rate of photosynthesis under limited CO 2 conditions. The proximity between mitochondria and chloroplasts (highlighted in red and green, respectively, in Fig. S12) may favor this process, facilitating intracellular gas exchange between the two cell organelles, as already shown in the case of other microalgae (Lavergne, 1989).
At the same time, increased intracellular CO 2 concentration is expected to lower the rate of photorespiration in mixotrophy. Although the Rubisco enzymes from Cyanidiophyceae show some of the highest carboxylation specificities reported to date (Sugawara et al., 1999), photorespiration is expected to proceed at high rates at the high temperatures and low CO 2 concentrations under which Galdieria grows. Indeed, knockout of peroxisomal glycolate oxidase in the transformable Cyanidiales alga Cyanidioschyzon merolae demonstrated that a functional photorespiratory pathway is essential for survival of these algae under ambient CO 2 concentrations (Rademacher et al., 2016). The observed increase in the rate of photosynthesis and gain of biomass under high CO 2 and mixotrophic conditions (25 AE 6%, when combining data from Figs 1, 2) is consistent with overcoming the expected loss of biomass gain due to photorespiration under photoautotrophic conditions.
In conclusion, by bypassing the possible metabolic antagonism between respiration and photosynthesis, G. sulphuraria can exploit the plethora of transporters encoded by its genome (Schonknecht et al., 2013;Rossoni et al., 2019a) and import organic carbon available in its environment and to boost CO 2 availability for photosynthesis. While this phenomenon certainly exists in other phototrophs (e.g. Rolland et al., 1997), the capacity to enhance photosynthesis with respiratory CO 2 when the latter process is increased by exogenous carbon sources could be particularly relevant in G. sulphuraria. This alga thrives in an extreme environment, where growth is limited by low pH, high temperature and possibly light availability. Dissolved inorganic carbon is very low in this hot and acidic milieu (in fact its concentration inversely correlates with the pH of different collection sites in Yellowstone National Park) Hamilton et al., 2012) and its light-driven uptake is fairly low when compared to alkaline thermal habitats. Conversely, the concentration of dissolved organic carbon can be relatively high in the acidic hot springs (from 17 µM to 3 mM (Nye et al., 2020)). Values could become much higher when this alga proliferates in mats (Gross et al., 1998), where by-products of every group of microorganisms may serve as 'food' for other groups.
Based on these considerations, it is tempting to speculate that while photosynthesis should allow cells to colonize new environments devoid of any organic carbon source, the peculiar division mode of G. sulphuraria (formation of endospores associated with the release in the media of the mother cell wall remnants), may favor mixotrophy on a longer timescale. In the acidic conditions where G. sulphuraria lives, this material is probably rapidly hydrolyzed, providing an organic carbon source to the algae, along with other external sources for dissolved organic carbon, like e.g. high-temperature acid-digested wood (Nye et al., 2020). Thanks to the abundance of transporters, this alga could outcompete other microorganisms such as fungi, which are also found in these extreme conditions, for growth. Thus, the coexistence of phototrophic, mixotrophic, and heterotrophic lifestyles thanks to the subtle compromise between the activity of the two energy-producing pathways (photosynthesis and respiration) would represent a key element for fitness and explain the success of G. sulphuraria to thrive in its extreme ecological niche. High fluctuations in the availability of dissolved organic and inorganic carbon, light, temperature, etc. in these environments possibly selected for the maintenance of a metabolic flexibility (Gross, 1999;Gross et al., 2002;Ciniglia et al., 2004;Cho et al., 2020), which may have also allowed Cyanidiales to invade more moderate habitats (Yoon et al., 2006;Az ua-Bustos et al., 2009;Castenholz & McDermott, 2010) and to disperse over long distances to geographically isolated extreme habitats (Rossoni et al., 2019a).
The same metabolic flexibility opens interesting perspectives to exploit this alga for biotechnology applications. Galdieria sulphuraria has already been employed in the fields of pigment/ antioxidant production, bioremediation and bioenergy (Cizkova et al., 2019) and the possibility to exploit its lifestyle flexibility should be explored to cultivate this alga in organic matter-rich open ponds without contamination by other microorganisms.

Supporting Information
Additional Supporting Information may be found online in the Supporting Information section at the end of the article.
Dataset S1 Proteins involved in photosynthesis, central metabolism and respiration were selected from the complete proteomic dataset (see Dataset S2) and used to build Fig. 3.
Dataset S2 Compared proteomic analysis between photoautotrophic, mixotrophic and heterotrophic growth conditions. Dataset S3 Compared metabolomic analysis between photoautotrophic, mixotrophic and heterotrophic growth conditions.            Methods S1 Microalgae and media composition, growth, cell fresh weight and dry weight estimates, Clark electrode oxygen and photophysiology measurements, mass spectrometry-based proteomic analyses, metabolic analyses by IC-MS, electron microscopy sample preparation and observation.
Please note: Wiley Blackwell are not responsible for the content or functionality of any Supporting Information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.