Augmented CO2 tolerance by expressing a single H+-pump enables microalgal valorization of industrial flue gas

Transcriptome landscaping during high CO₂ conditions

We first performed an RNA sequencing (RNA-seq)-based transcriptome analysis of C. reinhardtii under extremely high CO₂ conditions (i.e. a 20% CO₂-enriched condition) to investigate its cellular response to a high CO₂ environment and identify a gene candidate that could augment CO₂ tolerance¹¹. In addition, transcriptome profiling of algae grown in a CO₂-independent acidic environment (i.e. pH value of 5.0) was also carried out. The stress-imposing conditions were determined according to the results from preliminary condition screening experiments to determine the exposure levels of CO₂ and H⁺ that reduce the algal growth rate by 50% (Supplementary Fig. 1). The RNA-seq analyses were then performed using six biological replicates of each condition (Supplementary Fig. 2). Because H⁺ is known to play a key role in CO₂ poisoning under high CO₂ conditions and decrease the viability of algal cells (Fig. 1a), the latter analysis was conducted as a benchmark to potentially elaborate the H⁺-related CO₂ tolerance mechanisms even though they are unlikely to share the exact same genetic response mechanism (Fig. 1b, Supplementary Fig. 3, and Supplementary Data 1)⁷. The assay of the specific growth rate under several high CO₂-mimicking conditions (i.e. conditions with CO₂-independent low pH and a high concentration of dissolved inorganic carbon (DIC); details are provided in Supplementary Note 1) indeed suggested the existence of subtle differences in the cellular responses to acidic conditions depending on the presence and absence of CO₂. The results showed that the decrease in the pH of the extracellular environment caused by high CO₂ levels could be a minor toxic factor, although it was not the only direct cause of the reduced cell viability. This finding strongly implies that the interaction between intracellularly absorbed CO₂ and cellular activity would aggravate intracellular pH (pH_i) acidification and thereby amplify the toxicity (Table 1 and Supplementary Fig. 4). Namely, CO₂ intoxication seemed to be synergistically caused by an influx of H⁺ from the acidic extracellular space as well as the intracellular generation of H⁺ via CO₂ hydration (facilitated by the activities of cellular enzymes, such as carbonic anhydrase; Supplementary Fig. 5). Notably, the toxicity of high [DIC], which could be accompanied by high CO₂ conditions and trigger dose-dependent stresses, such as osmotic pressure, was also evaluated, but the analysis revealed that it was not a toxic factor.

a Growth of the WT strain under high CO₂ and low pH conditions. The conditions for EC₅₀ were selected (Supplementary Fig. 1). The arrow indicates the time point of biomass sampling for RNA-seq (n = 3). Data are the mean ± SD of three biological replicates. b Scatter plot of the expression levels of each gene (in RPKM values). A pseudocount was added to all RPKM values to enable plotting. The transcriptional changes in the PMA2 and PMA3 genes are highlighted by black circles and triangles, respectively. c Venn diagrams of the number of genes that were commonly upregulated (log₂Fold change (FC) > 1, p < 0.05) or downregulated (log₂FC < 1, p < 0.05) under excess proton-derived stress conditions compared with their expression in cells grown under ambient conditions (as a control). The areas are to scale. d The left panel shows the predicted cellular responses to excess proton-derived stresses, and the right boxes present a heatmap of the log₂FC values of the cell’s knowledge-based selected genes. The detailed roles and descriptions of the genes under CO₂-enriched conditions are described in Supplementary Data 3. An asterisk in the boxes indicates a statistically significant difference (p < 0.05) compared with the expression level under ambient conditions. An asterisk is shown only for the transcriptomes that showed over 2-fold changes in expression. In c and d, the statistically significant differences were determined by the generalized linear model (GLM) likelihood ratio (LR) test. Source data are provided as a Source Data file.

When exposed to high CO₂ levels, microalgae attempt to maintain their cytosolic pH, namely, pH homeostasis, against the environment by (i) upregulating proton extrusion extracellularly in response to the inward diffusion of H⁺; (ii) inactivating the carbon concentrating mechanism (CCM) to minimize the collateral intracellular acidification caused by enriched CO₂ and facilitating CO₂ hydration intracellularly; (iii) modifying the composition of the cellular membrane to strengthen its role as a proton barrier; and iv) increasing ATP synthesis to provide biological energy for the operation of ATP-driven transporters and protein recovery processes^7,8,12,13. Therefore, we probed the expression levels of several genes of interest related to the four abovementioned cellular responses based on a transcriptome analysis to examine whether these tolerance mechanisms function properly. In addition, the expression patterns of molecular chaperones (e.g. heat shock proteins; HSPs) were also inspected because these are ubiquitously expressed proteins under most stress conditions and play a crucial role in the recovery of proteins impaired by various stress factors¹⁴. In this context, the modulation of HSP expression is also an important response during acclimation to high CO₂-derived acidic conditions since the denaturation and aggregation of key proteins (i.e. enzymes) by excessively generated H⁺ negatively contribute to cell proliferation¹⁵. Overall, 3867 and 1487 differentially expressed genes (DEGs; log₂fold change (FC) > 1, p < 0.05) were identified from the high CO₂ and low pH sets, respectively. Among these DEGs, 15.9% of the significantly upregulated and 18.1% of the downregulated genes in the high CO₂-treated cells were commonly found in the low pH-treated cells (Fig. 1c and Supplementary Data 1). These overlaps indicate that the algae may manage high CO₂-derived stress in a similar but not equal manner to counteract the stress attributed to low pH condition, and this conclusion is likely because the two inhibitory conditions may not exert identical toxic effects on the cell, as was experimentally revealed in this study (Table 1 and Supplementary Fig. 4).

Prior to the detailed inspection of individual gene expression levels, a Gene Ontology (GO) enrichment analysis was performed to extract and identify the biological processes (BPs) that are sensitively regulated in response to high CO₂ conditions. From the 55 enriched GO terms, diverse modes of cellular responses for enduring the high CO₂ condition were determined (Supplementary Data 2). Among the upregulated gene clusters under the high CO₂ condition, GO-BP terms associated with the maintenance of cellular homeostasis (GO:0019725, GO:0006873, GO:0030003, and GO:0055082) were enriched, which may be attributed to heightened cellular efforts for minimizing the radical change in the intracellular condition caused by the extracellular condition. The enrichment of GO-BPs relevant to carbohydrate metabolism (GO:0044262, GO:0016051, and GO:0034637) was likely a response to the increased cellular activities against the high concentration of CO₂ because carbohydrates are known to not only help adjust the intracellular osmotic pressure but also act as protective agents for maintaining cellular homeostasis¹⁶. Additionally, under the high CO₂ treatment, protein and amino acid metabolism-related GO-BP terms (GO:0006520, GO:1901605, GO:0009072, GO:0006551, GO:0009100, GO:0006528, GO:0030163, GO:0006511, GO:0019941, GO:0044257, and GO:0051603) were also found to be overrepresented, which was likely due to the activated cellular responses related to catabolizing proteins and amino acids that may lead to the production of NH₃ and the reduction of intracellular H⁺ through deamination and decarboxylation, respectively, both of which are widely known as pivotal cellular mechanisms for microbial resistance to intracellular acidification¹⁷.

Subsequently, the expression levels of the tolerance-related essential genes were assessed to verify the response behavior of the genes individually and to predict the detailed genetic cause of the low tolerance of C. reinhardtii to high concentrations of CO₂ by comparing the purported cell responses with the actual gene expression patterns (Fig. 1d and Supplementary Data 3). As expected, a majority of essential genes that constitute the CCM were significantly downregulated (16 out of 21 investigated genes) under extremely high CO₂ conditions as a counteraction to the increased environmental DIC concentration. Because an increased extracellular DIC concentration can trigger a series of reactions that ultimately results in excessive endomembrane acidification by the accelerated inward flux of DIC species and their intracellular hydration, which promotes additional generation and accumulation of H⁺ in the cytoplasm, this gene regulation seems to be an example of a cellular endeavor against high CO₂ conditions (Supplementary Fig. 5)⁸. Countertrends were found for CAH2, RHP1, and RHP2, and the plausibility of this hypothesis has been previously validated¹⁸. Interestingly, although the expression of CCM genes has been known to be mainly influenced by the external CO₂ concentration, our findings showed that many CCM components (CAH1, CAH4, CAH5, HLA3, LCIA, CCP1, CCP2, LCID, and LCIE) expressed in response to low pH conditions showed similar patterns (i.e. downregulated) to those found under the high CO₂ condition, which likely indicates that several CCM genes are also related to the response to CO₂-independent low pH conditions. Regarding the determinants of cell components, the expression of key enzymes involved in controlling the composition of the cellular membrane appears to be regulated under extremely high CO₂ conditions. For example, upregulated expression was observed for SQD3, which contributes to a decrease in the membrane fluidity of the cell envelope by increasing the composition ratio of saturated membrane lipids. In addition, the results showed increases in the expression levels of Cre08.g365950, Cre08.377300, Cre12.g521650, and DGTT1, which are responsible for converting unsaturated lipids in intracellular and plasma membrane (PM) regions into triacylglycerol (TAG), a typical energy storage sink synthesized under various stress conditions. Consequently, these responses may inhibit the entry of H⁺ into the cell, which decreases passive proton influx and eventually leads to stress evasion^19,20,21,22. Additionally, as expected, under high CO₂ conditions, various HSPs from a wide range of HSP subfamilies were highly expressed. These increases in the expression levels of stress-inducible molecular chaperones may contribute to the turnover and refolding of proteins that are denatured and aggregated by excessive H⁺-derived oxidative stress¹³. Meanwhile, a higher expression of a number of building blocks for ATP synthase was also observed under harsh conditions. In particular, mitochondrial ATP synthase-associated (ASA) proteins, such as ASA1, ASA3, and ASA4, which are involved in the dimerization of ATP synthase, showed 1.93-, 1.59-, and 1.63-fold upregulation (all p values were less than 0.01), which in turn can result in increased ATP synthesis activity of the ATP synthase complex^23,24. These changes may correspond to cellular activity to increase ATP synthesis because acidic conditions lead to the consumption of ATP for the operation of H⁺-pumps and the activation of ATP-dependent HSPs, among other processes^7,25.

Under extremely high external CO₂, the CO₂-intolerant species C. reinhardtii struggles to maintain its cytosolic neutrality by dynamically modulating genes associated with CCM, cell composition modification, protein recovery, and ATP synthesis in addition to previously elucidated CO₂ tolerance mechanisms. In contrast, the gene expression of H⁺-pumps (e.g. PMAs, vacuolar H⁺-ATPase – V-ATPase, and H⁺-pyrophosphatase)²⁶ profiled under high CO₂ conditions reveals that H⁺ sequestration from the cytoplasm to the extracellular region or contractile vacuole does not seem to be adequately regulated (i.e. not significantly upregulated) in the microalga, which is clearly different from previously reported expression trends observed with acid- and CO₂-tolerant microalgal species^7,11,13. Rather intriguingly, a decrease in the expression of the PMAs – PMA2 (Cre10.g459200; slightly downregulated) and PMA3 (Cre03.g164600; highly downregulated with statistical significance) – was observed under high CO₂ conditions, which is an important finding because PMAs play a crucial role in life support during exposure to various acidic conditions^13,19. The reproducibility of the gene expression results at the transcriptomic and translational levels under high CO₂ conditions was verified by qRT-PCR and immunoblotting (Supplementary Figs. 6a–c and Supplementary Table 1), respectively, which implies the mRNA-protein expression correspondence and, at the same time, suggests that the downregulation can directly influence the congenital low tolerance. Although their distinctive functions have not yet been explicated, PMA2 and PMA3 are two PMA proteins identified in C. reinhardtii with a sequence similarity of 47.4%, and both of these proteins have highly conserved regions that are generally found in well-characterized PMAs (Supplementary Fig. 7)^27,28,29.

Since PMA is a key and primary molecular workhorse located at the forefront of the cell and responsible for adjusting the intracellular H⁺ concentration by pumping out H⁺ into the extracellular region, which results in controlling the pH_i and generating an electrochemical gradient to offer an H⁺-mediated driving force for various secondary transport proteins^19,26, the atypical expression of PMAs under high CO₂ conditions may limit the discharge of excessive H⁺ out of the cell, which would lead to undesired H⁺ accumulation, even more acidification of the intracellular region, and disruption of the proton gradient for nutrient and ion transport. All of these changes can inhibit normal cell propagation (Fig. 1a and Table 1). Therefore, we conjectured that the CO₂ intolerance of the alga can be attributed to insufficient expression of PMA genes and the concomitant low ability of the cell to extrude the excessive H⁺ generated under external CO₂ conditions. We thus hypothesized that the overexpression of PMAs would improve the CO₂ tolerance of the alga by increasing proton efflux and improving the ability of the alga to maintain its endomembrane pH at near neutrality even under high CO₂ environments.

Construction of microalgal strains with high CO₂ tolerance

To augment the CO₂ tolerance of C. reinhardtii by compensating for the low expression of H⁺-ATPase, the overexpression target was a typical PMA, i.e. plasma membrane H⁺-ATPase isoform 4 (hereafter referred to as PMA4) from Nicotiana plumbaginifolia, and its features have been well characterized and successfully expressed in various bioplatforms, such as other plants and yeasts^30,31. If this pump can be activated in algal systems, the exogenous gene expression strategy can provide several merits. Because of its host-independent property, the expression of PMA4 holds the potential to be a versatile approach for various microalgal species suffering from stagnant H⁺ extrusion and consequent CO₂ intolerance, which is a general but critical issue that interferes with the practical application of microalgae-based CO₂ conversion technology^3,8. In addition, PMA4 could help prevent cosuppression, which often occurs in homologous overexpression cases³². Given these findings, the gene insert was designed for mutagenesis. Because deletion of the autoinhibitory domain (located in the C-terminus; Supplementary Fig. 7), which is regulated by interacting with the regulatory 14-3-3 proteins that exist in every eukaryotic cell (including the microalga, C. reinhardtii), increases the activity of the H⁺-pump^31,33, we decided to use the truncated form of PMA4 (referred to as PMA4∆Cter) instead of the full-length protein to confer the greatest acid-derived tolerance to the microalgal cell.

Subsequently, to label a reporter for probing the subcellular localization of the expressed acid-secreting nanomachine, the PMA4∆Cter coding gene was translationally fused at its 3´-coding-end to the monomeric Venus fluorescent protein (mVenus) coding gene with a flexible linker sequence (GGSGGGSG) (referred to as PMA4∆Cter-V) and placed under the control of a strong constitutive promoter (Hsp70A-Rbc S2) (Fig. 2a and Supplementary Fig. 8a). Afterwards, the codon-optimized insert cassette was delivered into the cell by electroporation.

a Simplified schematic representation of the codon-optimized PMA4∆Cter-V (PMA4 protein whose autoinhibitory domain in the C-terminus was deleted and fused with a fluorescent Venus tag) insert (total 6835 bp). The gene scheme is not to scale. b Genotyping of two Zeocin-resistant colonies by PCR to confirm insertion. A and P indicate the actin (an endogenous control) gene (expected amplicon length: 501 bp) and the PMA insert gene (expected amplicon length: 4691 bp), respectively. c Identification of the insertion site revealed by NGS. In the transgenic algal cells, a single insert was confirmed. In the cartoon, Chr, P, Z, L, and V represent chromosome, promoter, Zeocin resistance, linker, and Venus tag, respectively. The scheme is not to scale. d mRNA expression verification by qRT-PCR. Data are the mean ± SD of two biological replicates. The relative expression levels were normalized to the expression levels of a housekeeping gene (IDA5). N.D. stands for not detected. e Western blotting analysis for confirming protein expression. The expression of the ATP synthase β-subunit was simultaneously examined as an endogenous control. f Fluorescence analysis of the expressed protein by flow cytometry. Details of gating strategy used for the analysis is described in Supplementary Method 5 and Supplementary Fig. 9a–d. g Representative confocal images (bar scale of 5 μm) to confirm the proper localization of the exogenously expressed protein. In contrast to the results obtained for the WT strain, the fluorescence signals were observed exclusively in the plasma membrane of the transgenic strains. Source data are provided as a Source Data file.

After transformation, two independent mutants were screened — whose genetic and phenotypic stabilities are maintained over time – from selective medium plates containing Zeocin via colony PCR (Fig. 2b) using a designed screening primer set (Supplementary Table 1 and Supplementary Method 1). The transgenic cells were named PMA4ΔCter-V4 and PMA4ΔCter-V10 according to the order of their isolation. Due to the lack of an efficient homologous recombination method for C. reinhardtii³⁴, the inserted gene was randomly intercalated into the genome. Gene insertion leads to disruption of the original genome sequence, which can impact transgenic cell phenotypes. To clarify the cause of the phenotypic changes (vide infra), insertion site analysis was performed on the isolated strains by next-generation sequencing (NGS)-based whole-genome sequencing using HISAT2 software (Fig. 2c; see Supplementary Method 2 for details). In the PMA4ΔCter-V4 strain, only one insertion site was identified in chromosome 7, and this mutation deleted 23 base pairs (bps) from the genome and disrupted a predicted protein-encoding gene (Cre07.g331114). In PMA4ΔCter-V10, one inserted gene was also found on chromosome 6. During the insertion process, an approximately 5.5-kbp truncation of the genome occurred, and three protein-encoding genes, tail-specific protease 2 (TSP2; Cre06.g265850), a predicted protein (Cre06.g265900), and dynein heavy chain 3 (DHC3; Cre06.g265950), were unintentionally lost. The insertion site revealed by NGS was visually confirmed by PCR using appropriate primers (Supplementary Fig. 8b, c and Supplementary Table 1). Given the annotations of the disrupted genes (Supplementary Table 2) and previous reports on their functions^35,36,37, it is unlikely that the gene deletion events directly affected the ability of the transgenic strains to maintain their pH_i against acid-derived CO₂ stress, which implies that the phenotypic modifications described below are generally not associated with the lost genes but rather caused by exogenous PMA expression. Following this analysis, the expression of the inserted gene was confirmed by quantitative real-time RT-PCR (Fig. 2d) using a designed primer set (Supplementary Table 1 and Supplementary Method 3), and the successful transcription of mRNA encoding PMA4ΔCter-V in the transgenic strains was confirmed by cDNA amplification. Moreover, no expression of PMA4ΔCter-V was detected in the wild-type (WT) strain. The expression of the protein of interest was then verified by Western blotting, and a band at approximately 130 kDa (Fig. 2e), which is located near the expected molecular weight of the PMA4ΔCter-V protein (120.81 kDa; estimated from its sequence), was visualized only with the transgenic strains when the same amount of protein is loaded (i.e. 20 μg), which reveals PMA overexpression properly occurred in the mutants (see Supplementary Method 4 for details)³¹.

In addition to molecular-level expression tests, whether the protein is expressed in the intended subcellular part (i.e. PM) should be determined before evaluating the cellular performance because the protein must be localized appropriately before it can pump H⁺ into extracellular space³¹. The localization of the protein was examined with the aid of a fluorescent tag protein (i.e. mVenus) (see Supplementary Method 5 for details). For precise fluorescence comparison, a transgenic microalgal cell line expressing only mVenus (referred to as V8) was constructed using a separate insert cassette (Supplementary Fig. 8d) as a fluorescence reference strain. The integration of the transgene was genotypically verified (Supplementary Fig. 8e). First, the quantitative fluorescence of the cell lines was estimated by flow cytometry using BD Accuri^TM C6 Plus software and Flowjo software (Fig. 2f). The results showed that mVenus expression induced significant shifts in the fluorescence peaks toward a higher intensity in the transgenic algal strains, including PMA4ΔCter-V4, PMA4ΔCter-V10, and V8 (Supplementary Table 3 and Supplementary Fig. 9e). Subsequently, the localization of the fluorescence source in C. reinhardtii was specified by confocal microscopy. The image-based assessment visually demonstrated that bright fluorescent proteins were preferentially concentrated near the PM of PMA4ΔCter-V4 and PMA4ΔCter-V10, whereas no notable concentration of fluorescence in the specific subcellular region was detected in the WT (Fig. 2g and Supplementary Fig. 9f–i). This result indicated that the chimeric protein was successfully situated at the targeted position, namely, the PM, and ensures that the minimum requirement for proper protein function was met. In the fluorescent reference strain, V8 (Supplementary Fig. 9i), a strong fluorescence signal was observed throughout the cytoplasm, which suggested that the fluorescent proteins were floating freely without showing any specific localization. Compared with the fluorescent reference strain, the PMA-expressing cell lines exhibited relatively weak fluorescence, which can be primarily explained by the C-terminal truncation-derived lower stability and higher turnover of the expressed protein, as reported previously³¹. Another possible reason is that the protein expression level may not be that high because excessive protein expression can be harmful to the host, and in particular, excessive expression of PMA proteins can impose a huge burden on cells and consequently adversely affect their viability³⁸. Either way, the pump protein was deemed to be sufficiently expressed at the posttranslational level to confer tolerance to H⁺-related stresses without having a notable negative influence (vide infra). To further corroborate the localization of the protein, the fluorescent tag protein was probed by enzyme-linked immunosorbent assay (ELISA) using the PM-enriched fraction isolated with the two-phase partitioning method. Detection of the fluorescent tag in the sample also confirmed the proper localization of the protein (Supplementary Fig. 9j).

Characterization of the performance of the transgenic cells

Since the CO₂ tolerance improvement strategy described in this study is in line with the approach of enhancing the tolerance to low pH conditions, the productivity of the low pH-grown cell lines was evaluated first. The evaluation of the pH tolerance of the algal cells was conducted under two independent cultivation modes: mixotrophic and autotrophic conditions. For the mixotrophic culture of the facultative autotroph, acetic acid was applied as an organic carbon source. At a pH value of 5.5, the transformed cell lines showed a clear difference in growth performance compared with the WT cells. In particular, under mixotrophic cultivation at pH 5.5 (Fig. 3a, b), the proliferation rate of the WT cells was significantly stunted, and the productivity of the transgenic strains was high, which indicated that the acid tolerance of the mutants was successfully augmented. Because the growth of the transgenic strains at this low pH was comparable to that at neutral pH (Supplementary Figs. 10a and 11a), this constitutive PMA overexpression strategy did not appear to have a negative effect on the cells. In the photoautotrophic acidic cultivation, the exogenous PMA-expressing mutants also exhibited a relatively improved growth performance under the same acidity (pH 5.5) but did not show a dramatic growth difference compared with the results obtained with the corresponding mixotrophic culture (Supplementary Figs. 10b, c and 11b, c). This trophism-dependent growth performance difference can be ascribed to the fact that the membrane-permeable weak acid protonated acetate (i.e. acetic acid), which was added to create the mixotrophic conditions, may exacerbate intracellular acidification by generating additional H⁺ in the microalgal cells³⁸. In this context, the acetic acid-supplemented pH 5.5 environment can inflict markedly higher levels of acid stress on the cells than a medium with an identical pH but without acetic acid, and this higher stress will eventually lead to the apparently dramatic difference in tolerance observed between the two different carbon regimes (Fig. 3b and Supplementary Fig. 11a–c). These results thus imply that the engineered cells possess high tolerance not only to extracellular acidic conditions but also to factors that aggravate pH_i acidification.

a Spot culture with serial dilution on agar medium (supplemented with acetic acid) at pH 5.5. b The pH tolerance during liquid cultivation (pH 5.5; supplemented with acetic acid) is markedly enhanced via PMA4∆Cter expression. Shown are data from duplicate cell culture. The cell states at Day 2 of WT, PMA4∆Cter-V4, and PMA4∆Cter-V10 are displayed from the far left of the top panel. c The photoautotrophic biomass production (upper panel) and CO₂ fixation rate (R_CO2; lower panel) are improved under extremely high CO₂ conditions (20% CO₂). The two graphs share the same timescale. The arrows indicate the time points (distinguished by the colors) that the cell lines reached their maximum concentrations. Data from duplicate cell cultures are shown. Each dot in the lower panel depicts two biological replicates. d Per-cell gross photosynthetic rate of the strains exposed to high CO₂. The parameter was estimated as light-dependent O₂ evolution plus dark respiration. The double asterisks indicate statistically significant differences in the gross photosynthetic rate of each strain under the high CO₂ environment compared with those under the ambient conditions (p < 0.05). Each bar represents the mean ± SD of three independent experiments. e Box and whisker plots demonstrating the pH_i of the microalgal strains exposed to high CO₂ conditions derived from four biological replicates. The median (the center line) ± whiskers (1.5× the interquartile range from the lower and upper quartiles) with the interquartile range (the boundaries) are provided. The double asterisks indicate statistically significant differences compared with the pH_i of the WT strain after 2 h of exposure to CO₂ (p < 0.05). N.S. represents not significant. f In vitro ATPase activity assay using isolated PM-enriched fraction samples. The double (p < 0.05) and triple (p < 0.01) asterisks denote statistically significant differences compared with the ATPase activity of the WT strain, respectively. Each bar represents the mean ± SD of four independent experiments. In d, e, and f, the statistically significant differences were determined by a two-tailed Student’s t test. g Representative image of bromocresol purple-mediated in vivo ATPase activity test. Source data are provided as a Source Data file.

Based on the previous pH tolerance evaluation, an autotrophic CO₂ tolerance test (without the organic carbon source) was performed with the cell lines after exposure to an extremely high CO₂ concentration (20%) to exclude the effect of other influencing factors on acid tolerance. Although the CO₂ concentration was somewhat higher than that found in typical industrial exhaust gas⁶, it was selected to test whether the strains possessed sufficient acid tolerance to be used in practice given the additional acidification effect that may be induced by the various acidic gaseous components (e.g. SO_X and NO_X) contained in general industrial exhaust gas streams. In this assay, the algal biomass production rates estimated from the turbidities of the cultures were compared to accurately assess CO₂ sequestration-related performance metrics because the microalgal CO₂ fixation rate (R_CO2) is directly proportional to biomass productivity². For the estimation, an OD₈₀₀-biomass correlation curve (Supplementary Fig. 12) and Eqs. (2) and (3) were used. According to the tolerance test, the WT strain showed stagnant biomass productivity (especially in the initial culture phase; Supplementary Fig. 11d) when exposed to high CO₂ concentrations compared with its growth under ambient conditions, which clearly reveals the vulnerability of these algal cells to elevated CO₂ levels. According to the entire batch culture result, eventually, the WT showed 0.66-fold decreased maximum biomass production while exhibiting 1.09-fold increased productivity under high CO₂ conditions (by Day 6) compared to those under ambient CO₂ conditions (by Day 17) during reaching their maximum cell growths (Fig. 3c and Supplementary Fig. 11c). The higher productivity is mainly attributed to the relatively shorter cultivation period because of the rapid cell death by high-CO₂ stresses. However, the slight increase would be hardly meaningful in terms of the CO₂ removal because the autotrophic growth rate that is not correspondingly increased against the elevated supply of CO₂ to the algal culture, inevitably causes the generation of unfixed CO₂, which significantly lowers the overall CO₂ reduction efficiency³⁹. Moreover, the dramatically reduced maximum cell density would also be unfavourable from the same point of view, because the final biomass concentration is one of the most influential factors affecting the economic feasibility and energetic efficiency — which is intimately related to the net CO₂ removal ability of CO₂ conversion processes — from the life-cycle perspective^2,40.

On the contrary, the transgenic microalgal strains exhibited noticeable improvements in any photosynthetic production metrics, including maximum biomass productions and rates, even in a culture system to which an extremely high level of CO₂ was provided. Their maximum biomass productions were considerably above that found for the WT strains under identical CO₂ conditions, and 3.22- and 3.03-fold improvements were found for PMA4ΔCter-V4 and PMA4ΔCter-V10, respectively (upper panel of Fig. 3c). The accumulative productions and the rates of the transgenic strains under high CO₂ conditions (by Day 7) were also much higher (rather 1.12- and 1.09-fold improvements in the accumulative biomass productions and 2.77- and 2.62-fold improvements in the productivities for PMA4ΔCter-V4 and PMA4ΔCter-V10, respectively) than those under ambient CO₂ conditions (i.e. autotrophic pH 7.0) (by Day 17) (Fig. 3c and Supplementary Fig. 11c), which illustrates that the PMA-expressing mutants can even beneficially utilize excessive CO₂ while attenuating the intracellular acidification effect. This growth improvement also indicates that the amount of CO₂ fixation obtained with each mutant during the culture period was substantially improved compared with that found for WT under conditions with high CO₂ provision (lower panel of Fig. 3c). In terms of the rate of CO₂ fixation, PMA4ΔCter-V4 and PMA4ΔCter-V10 showed 3.23- and 2.99-fold increased R_CO2 values compared with the WT strain, respectively, during their exponential growth phases when their effective CO₂ fixations occur (between Day 2 and 6). Taken together, this tolerance improvement strategy raises the mutants’ maximum biomass concentrations during a shorter culture period under high CO₂ supply conditions and improves their CO₂ removal rates and efficiencies, all of which are requisites of microalgal strains for efficient CO₂ sequestration and economically feasible CO₂-derived product productions^39,40.

In parallel to the growth assessment, the per-cell basis photosynthetic O₂ evolution of the cell lines exposed to high CO₂ conditions was evaluated (see Supplementary Method 6 for details). Consistent with the productivity data, the high CO₂-grown WT strain exhibited a lower gross photosynthetic rate (i.e. O₂ evolution plus dark respiration) than the WT strain cultivated under ambient conditions (approximately 50.4% decrease), which implies that the algal photosynthetic apparatus was severely impaired by the high CO₂ condition. In contrast, the PMA4ΔCter-V4 and PMA4ΔCter-V10 strains grown under high CO₂ conditions showed increased photosynthetic O₂ evolution rates in comparison to those found during growth under atmospheric conditions (approximately 58.6% and 37.3%, respectively) (Fig. 3d). Based on observations from previous reports and the results of this study, this excess CO₂-related photosystem impairment is likely the primary reason for the growth retardation observed in the presence of high levels of CO₂^8,41. Moreover, the increased O₂ evolution observed with the transgenic cells is attributable to the activation of oxygenic photosynthesis-related enzymes under tolerably high CO₂ conditions to facilitate biological CO₂ assimilation, as has been well documented⁴². Thus, C. reinhardtii, which is a congenitally CO₂-intolerant microalga, became resistant to high CO₂ conditions via PMA expression and its activated core machinery for photosynthesis was thereby able to generate an increased amount of oxygen.

To probe whether PMA expression indeed improves photosynthetic O₂ evolution and biomass productivity by providing cells with the ability to maintain their pH_i near neutrality, we next monitored the time-lapse cytoplasmic pH changes after high CO₂ treatment using a membrane-permeable pH-sensitive fluorescent dye, BCECF, AM (2′,7′-bis(2-carboxyethyl)-5 (6)-carboxyfluorescein, acetoxymethyl ester) (see Supplementary Method 7 for details)²⁶. To eliminate the background signal, including the autofluorescence of the cells (for all strains) and the fluorescence of the fluorescent probe (i.e. mVenus for the transgenic algal lines), from the fluorescence of interest emitted by the pH_i-sensitive dye, the pH_i of each algal strain was calibrated independently (Supplementary Fig. 13). Consequently, the WT strain showed a dramatic decline in their pH_i during 2 h of exposure to high CO₂ (a median pH_i value of 6.56), whereas the transgenic cell lines showed a better ability to robustly maintain a near-neutral pH_i under acidic conditions (median pH_i values of 6.65 and 6.78 were obtained for PMA4ΔCter-V4 and PMA4ΔCter-V10, respectively) (Fig. 3e), which demonstrates that high CO₂ conditions largely affect the viability of microalgae by lowering the cytosolic pH and that PMA expression in algae results in a positive response to CO₂-derived intracellular acidification. Given that the photosynthetic CO₂ fixation system and O₂ evolution machinery are highly sensitive to the pH and CO₂ conditions of the surroundings as well as to the cytosolic and subcellular pH^43,44, the differences in the ability of the cells to maintain their pH_i under high CO₂ conditions appear to be the fundamental cause of the tolerance gap between the WT and transgenic strains.

In the final stage of the cell performance characterization, we investigated the H⁺ pumping ability of the algal cells determined by the activity of ATPase on the PM region to determine the CO₂ tolerance conferred to the mutants by exogenously expressed PMA via the H⁺ extrusion function of the cell. This ability was comprehensively evaluated both in vitro and in vivo. In the in vitro ATPase activity test, the PM fraction-enriched protein sample, which is identical to that isolated for assessment of its localization (i.e. ELISA), was used (see Supplementary Method 8 for details)³¹. As expected from the results of the previous performance tests, PMA4ΔCter-V4 and PMA4ΔCter-V10 exhibited approximately 2.5- and 2.6-fold higher ATPase-mediated ATP hydrolysis performance than the WT strain, respectively, which demonstrates that the transgenic strains have substantially higher H⁺ translocation activity in their PM regions than the WT strain (Fig. 3f). The greater ATPase activity in the mutants could cause higher ATP consumption, which may lead to additional energetic costs associated with the exogenous ATPase. However, given algal growth under neutral conditions (Fig. 3b and Supplementary Fig. 11a, c), the viability of the algal cells was not significantly affected. An in vivo assay was performed to confirm the algal H⁺ extrusion performance at the cell level and the direction of the H⁺ flux using a chemical pH indicator, bromocresol purple (see Supplementary Method 9 for details). The colourimetric assay visually demonstrated that the H⁺ flux was appreciably improved in the PMA-expressing mutants, as can be inferred from the color change on the perimeter of the cells caused by the expelled H⁺ and the accompanying extracellular acidification (Fig. 3g and Supplementary Fig. 14), and this result also provides direct evidence of the outward flux direction of H⁺, namely, from the intracellular region to the exofacial space, and hints that the H⁺ pumping protein was properly localized and oriented as desired. On balance, the data clearly show that the expression of the H⁺-pump imbued the algal transformants with markedly improved H⁺ efflux and an augmented ability to maintain their pH_i against factors causing severe intracellular acidification (e.g. membrane-permeable organic acids and CO₂); thus, the algae showed enhanced tolerance to extremely high CO₂ conditions.

Directly utilizing coal-fired flue gas for fuel production

To estimate the real-world practicality of the engineered algal strain as a CO₂-derived biofuel production platform⁴⁵, we grew the PMA4ΔCter-V4 strain, which showed remarkable productivity under the high CO₂ level (Fig. 3c), in the presence of actual coal-fired flue gas discharged from the Taean thermal power station complex unit 5 boiler of Korea Western Power Co., Ltd. The gas stream consisted of approximately 13% CO₂, 20 ppm NO_X, and 32 ppm SO_X on average (vol.%)⁴⁶. In this outdoor cultivation, coal-fired flue gas was chosen as the exemplary industrial flue gas because the discharge of CO₂ from the exhaust is the largest contributor to global emissions, and its reduction is therefore vital⁴⁷. Despite the urgency of mitigating this pollutant, the direct application of microalgae has been considered impractical to date due to its vulnerability to CO₂⁴¹.

The strain was photoautotrophically cultivated with continuously provided CO₂ through the introduction of highly acidic exhaust gas as the sole carbon source in a microalgal CO₂ bioconversion facility constructed next to the coal combustion boiler using a 10 L (in working volume) photobioreactor (PBR) (Fig. 4a). After two weeks of cell culture, we holistically evaluated various performance indices of the strain using Eqs. (2)–(6)⁴⁵. We observed that PMA4ΔCter-V4 exhibited a markedly improved phototrophic growth rate than the WT strain (2.28-fold higher) and beneficially converted the high concentration of CO₂ included in the noxious gas stream. The increased biomass productivity in the transgenic algal strain led to the enhancement of other essential parameters, including the microalgal CO₂ fixation rate (by 2.23-fold), total lipid productivity (by 2.21-fold), total fatty acid methyl ester (FAME) productivity (by 4.68-fold), and calorific productivity (by 2.20-fold) (Fig. 4b, Supplementary Fig. 15, and Supplementary Table 4). The acceleration of CO₂ sequestration, FAME accumulation, and calorific production indicate the practicality of using the engineered strain as a biological platform for producing various types of biofuels (i.e. biodiesel and solid fuel for direct combustion) while reducing industrial CO₂ emissions even when a variety of toxic gaseous compounds are continuously delivered into the culture. Overall, PMA expression allows inherently CO₂-intolerant algae to overcome the cytotoxicity of high CO₂, which leads to the efficient and rapid conversion of greenhouse gases and improves the technological feasibility of using microalgae with augmented tolerance.

a Bioconversion of CO₂ in the flue gas stream emitted from the coal-fired power plant. A house-manufactured bubble column PBR (10 L of working volume) was employed for cultivation in the vicinity of the coal combustion boiler. b Radar plot showing various cellular performances measured in duplicate (n = 2). The covered ranges on the axis indicate the standard deviation of each parameter. Source data are provided as a Source Data file.