The diatom T. pseudonana (CCMP 1335) was obtained from the Yellow Sea Fisheries Research Institute Microalgae Culture Center of the National Marine Genetic Resource Center (Xu et al., 2022)¹. In the laboratory, cells were grown in semi-continuous cultures in sterile seawater enriched with modified f/2 medium containing 100 μM N, 6 μM P and 100 μM Si and maintained at 20 ± 1°C under an irradiance of 120 μmol photons m⁻² s⁻¹ and a 12 h/12 h light–dark cycle (light on at 8:00 am and off at 8:00 pm).

The growth inhibition of T. pseudonana under different concentrations of copper (Cu), was quantified before the start of the experiment. T. pseudonana was grown in 400 ml cultures at cell densities of 8 × 10⁴ cells ml⁻¹ and supplemented with a series of copper (prepared with CuSO₄·5H₂O) ranging from 0 μmol L⁻¹ (using the background concentration in natural seawater, < 0.03 μmol L⁻¹) to 35 μmol L⁻¹ for 96 h (during exponential phase). At the beginning and end of the experiment (96 h), 0.5 ml of T. pseudonana culture was collected and preserved in Lugol’s solution to estimate microalgal growth by directly counting cell numbers using a hemocytometer and optical microscope (Nikon, Tokyo, Japan). The growth rate (μ, day⁻¹) was calculated using Equation (1):

where N₁ and N₀ represent cell concentrations at t₁ and t₀, respectively. The percentage of growth inhibition was calculated with reference to the control (no additional copper) using Equation (2):

where μ₂ and μ₁ represent the growth rate at a certain copper concentration and in the absence of copper, respectively. The concentrations of copper causing a 50% reduction in growth (IC₅₀, _96 h, 20 μmol L⁻¹, Supplementary Figure S1) was determined by linear interpolation. Although Cu was presented in the base medium, the group without extra copper addition was set as the control group. In the copper toxicity experiment, the compound EDTA (ethylenediaminetetraacetic acid, disodium salt, dehydrate; Na₂EDTA·2H₂O) and FeCl₃·6H₂O as well as other micronutrients in f/2 medium were not added into sterile seawater medium, due to the chelating properties of EDTA, which could decrease copper toxicity.

Four experiments were set up in this study (Supplementary Figure S2). Firstly, to assess the effect of elevated pCO₂ on the physiological performance of T. pseudonana, a selection experiment was set up in laboratory, where algae were cultured under ambient and elevated pCO₂ for 720 days. Secondly, to assess the evolutionary response of T. pseudonana to elevated pCO₂, shift experiments were set up in the laboratory, where the selected lineages at ambient and elevated pCO₂ were transferred into five different concentrations of pCO₂. Thirdly, to compare the differences in response between the indoor and outdoor experiments, an outdoor culture system was set up, where T. pseudonana was cultured under ambient and elevated pCO₂ using natural temperature and light. Fourthly, we conducted transcriptome sequencing of T. pseudonana using the long-term selected population (after 720 days’ selection in laboratory), to gain a mechanistic understanding of how elevated pCO₂ influences copper metabolism of T. pseudonana. For all four experiments, algae were acclimated under ambient or elevated pCO₂ for more than 30 generations, and then exposed to Cu without acclimation. In all experiments, the respective pCO₂ levels of 420, 1,000, and 2,000 μatm were established by bubbling the liquid medium with air or air/CO₂ premixed gas using a CO₂ chamber (HP1000G-D, China, for the laboratory experiment) or CO₂ Enricher (CE-100B; Wuhan Ruihua Instrument & 25 Equipment Ltd., for the outdoor culture system).

To study the effect of elevated CO₂ on the physiological performance of T. pseudonana, three different pCO₂ levels (420, 1,000, and 2,000 μatm; mimicking current and future pCO₂ rises up to the year 2,300 under IPCC scenario RCP 8.5) were set up using nutrient-replete f/2 medium. T. pseudonana (acclimated at 20 ± 1°C under an irradiance of 120 μmol photons m⁻² s⁻¹) was transferred during the exponential phase of growth to an inoculum of the relevant CO₂-modified f/2 medium at identical densities of 8 × 10⁴ cells ml⁻¹ and the cultures diluted every 4–6 days to maintain a stable carbonate system. In the selection experiment, algal cells were grown under semi-continuous culture conditions under the above pCO₂ levels for 720 days. Three biological replicates were set up using 1,000 ml flasks with 800 ml medium for each pCO₂ treatment. To study whether long-term selection under elevated pCO₂ has an impact on how T. pseudonana deal with exposure to copper stress, cells from each selected sample were transferred at a density of 8 × 10⁴ cells ml⁻¹ into CO₂-modified f/2 medium supplemented with or without copper for another 96 h at time intervals of 60 days (Supplementary Figure S2). Three concentrations of Cu exposure including a control (without additional Cu), low Cu (1 μmol L⁻¹, as usually occurs in polluted coastal areas), and high Cu (20 μmol L⁻¹, according to the IC₅₀, _96 h) were set up in the copper exposure experiment.

At the end of each batch-culture experiment, a 10 ml aliquot from each replicate was collected and centrifuged at 10,000 g for 10 min at room temperature. The algal precipitate was rinsed twice with Milli-Q water and an ice-cold phosphate buffer to remove extracellular copper, and then harvested and stored at −20°C to estimate intracellular copper content (Angel et al., 2017).

During the selection experiment the pH in each culture medium was measured before and after each dilution with a pH meter (Orion ROSS, Fisher Scientific Instruments), which was calibrated before use (NBS, National Bureau of Standards; variation range ± 0.05). Temperature, salinity, and total alkalinity (TA) were also measured periodically during the selection experiment at time intervals of 60 days. The average value was used to calculate carbonate system parameters using the CO2SYS Package (Pierrot et al., 2006; Supplementary Table S1).

Following 720-day selection in the respective CO₂ levels (corresponding to 833 generations, 920 generations and 887 generations in ambient pCO₂ of 420 μatm and high pCO₂ of 1,000 and 2,000 μatm, respectively), cells were transferred to another five increasing pCO₂ concentrations, ranging from 420 to 2,000 μatm and maintained in semi-continuous batch culture and transferred to fresh media every 4–6 days (Supplementary Figure S2). After acclimation for 30 days, each lineage under the respective pCO₂ was cultured with or without added exposure to Cu for another 96 h. The pH, temperature, salinity, and total alkalinity (TA) were measured before and after each dilution and the average was used to calculate the parameters of the carbonate system (Supplementary Table S2). Cells without Cu exposure were set as the control. In order to test acute Cu toxicity, high Cu (20 μmol L⁻¹, according to the IC₅₀, _96 h) was used in the Cu exposure experiment. Cell growth and intracellular copper accumulation were determined after 96 h, using the methods described above. The direct response to the selection (S) was measured by comparing the growth rate (or intracellular copper concentration) of elevated pCO₂ (1,000, 2,000 μatm) selected cells to that of cells from the ambient pCO₂ culture using Equation (3) (Brennan et al., 2017)

where E is the growth rate (or intracellular copper concentration) of elevated pCO₂ selected cells, and A is the growth rate (or intracellular copper concentration) of ambient pCO₂ selected cells.

To investigate the effect of increasing pCO₂ on copper metabolism of T. pseudonana under more natural environmental conditions, cells selected at pCO₂ values of 420, 1,000, and 2,000 μatm for 720 days in laboratory were transferred outdoors and cultured at the respective pCO₂ on the sea (37°06 ‘N, 122°33 ‘E), using 10 L tanks with 8 L cultures, which were set as long-term selection treatments and deployed on the sea surface (LT; Supplementary Figure S3). Additionally, stock cells without any selection treatment in the laboratory were also transferred to outdoors and cultured under the same series of pCO₂ of 420, 1,000, and 2,000 μatm, and set as short-term acute treatments and cultures also deployed on the sea surface (ST). After acclimation in filtered natural seawater under the respective pCO₂ for 30 days, each lineage was cultured with or without additional Cu exposure for another 96 h. Cell growth and intracellular copper accumulation were determined after 96 h, using the methods described above. The variations of pH, temperature, salinity, and total alkalinity (TA) were measured every day, and were used to calculate the parameters of the carbonate system (Supplementary Figure S4).

To investigate the molecular responses of T. pseudonana to increasing pCO₂ and copper exposure, cultures evolved under 420 and 1,000 μatm pCO₂ for 720 days were transferred in triplicate to fresh media under the same CO₂ conditions, with or without copper exposure (20 μmol L⁻¹) for a further 24 h. At the end of the experiment, 250 ml of culture from each replicate were harvested, centrifuged at 4°C and frozen in liquid nitrogen and stored at −80°C for RNA extraction, transcriptional sequencing, and RT-qPCR. The primers designed for reference and target genes were designed and are listed in Supplementary Table S3. The raw sequence data reported in this paper have been deposited in the Genome Sequence Archive in BIG Data Center, Beijing Institute of Genomics, Chinese Academy of Sciences, under accession numbers CRA001653, that are publicly accessible at http://bigd.big.ac.cn/gsa. A further 100 ml of culture from each replicate was harvested, centrifuged at 4°C and frozen in liquid nitrogen and stored at −80°C for enzyme activity determination. The antioxidant enzyme activities of glutathione peroxidase (GPX), catalase (CAT) and ascorbate peroxidase (APX) were determined using commercial reagent kits (Comin Biotechnology Co., Ltd., Suzhou, China) following the manufacturer’s instructions.

The effects of elevated CO₂ and Cu exposure on algal growth rate or intracellular copper concentration ([Cu]_intra) over the selection period of 720 days were analyzed with a mixed effects model in R,² using the packages lme4 and lmerTest. Algal growth rate or intracellular copper concentration ([Cu]_intra) were used as the experimental variable to be explained. Replicates were used as random effects. To test the individual and combined effects of increasing pCO₂ and Cu exposure on algal growth rate and intracellular copper concentration [Cu]_intra across different culture duration in selection experiment, different pCO₂ levels or different Cu exposure scenario or culture duration was set as one fixed effect. Different pCO₂ levels, different Cu exposure scenario, and culture duration were then tested separately and each in interaction with pCO₂, and the most parsimonious model was chosen for reporting based on the smallest AICc score (Akaike Information Criterion for small sample sizes); To test the algal adaptive response to increasing pCO₂ in shift experiment, the assay pCO₂ was set as a fixed factor and the different selection pCO₂ was used as random effects; To test algal physiological responses to increasing pCO₂ in outdoor culture experiment, different pCO₂ levels were used as fixed factors. The different Cu exposure scenarios and selection length (short-term, STS and long-term, LTS) were used as random effects.

Elevated pCO₂ significantly increased growth rates of T. pseudonana in all Cu exposure scenarios (Figure 1A; F_4,88 = 4.00, p < 0.05). Under elevated pCO₂ conditions without Cu, maximal growth rates were found at 1000 μatm pCO₂ until month 14 (Figure 1A; F_2,92 = 159.71, p < 0.001). Trajectories of algal growth rate over the course of the selection experiment differed substantially between the ambient and elevated pCO₂. After acclimation for 24 months (approximately 1,000 generations), populations selected under 2,000 μatm pCO₂ showed higher growth rates than populations selected at 1,000 and 420 μatm pCO₂. In addition, the variation in growth rates between the ambient and elevated pCO₂ reduced as the experiment proceeded (F_2,100 = 56.03, p < 0.001).

In contrast to the variation in growth rate, elevated pCO₂ significantly reduced copper accumulation within algal cells in all Cu exposure scenarios (Figure 1B; F = 79.632, p < 0.001). Populations selected under 1,000 and 2,000 μatm pCO₂ showed opposite trends in copper accumulation across the selection periods. Over 24 months of selection, copper accumulation gradually increased in populations selected under 1,000 μatm pCO₂, whereas copper accumulation gradually decreased in populations selected under 2,000 μatm pCO₂, in each of the control, the low Cu or high Cu treatments.

We found evidence of adaptation to elevated CO₂ conditions by measuring the direct response to selection (growth rate of evolved populations relative to that of control assayed in the same selection environment). Populations evolved under 1,000 μatm and 2,000 μatm pCO₂ showed a positive response to selection in their respective selection environments, indicating that evolution under elevated pCO₂ conditions increases growth rates beyond the plastic response (of the evolved control populations). While populations evolved under elevated pCO₂ conditions converge on a similar evolved growth rates (Figure 1), populations evolved under the highest pCO₂ conditions (2,000 μatm) evolved more than populations evolved under pCO₂ 1,000 μatm to arrive at the same endpoint. However, the strength of selection in each environment is similar (indicated by the initial population growth rate of evolved control populations in the same environments; Figure 2). Population growth rates generally increased with increasing pCO₂ across all selected lines, including treatments without (Figure 2A; F_4,38 = 6.441, p < 0.001) or with Cu exposure (Figure 2B; F_4,38 = 2.924, p < 0.05). However, growth rates in elevated pCO₂-selected samples decreased and showed lower levels than the ambient-selected ones, when they were shifted back to ambient pCO₂ at 420 μatm or a slightly higher pCO₂ at 700 μatm.

When copper is added to the environments, simulating environmental pollution of the aquatic environment, the benefits of evolution under elevated pCO₂ are reversed in all but the highest pCO₂ environments (Figures 2B,D; copper accumulation is greater in selected lines than the plastic response indicated by the dashed line). Overall, copper accumulation decreased significantly with increasing pCO₂ in the shift experiments under treatments both without (Figure 2C; F_4,40 = 11.23, p < 0.001) and with Cu exposure (Figure 2D; F_4,40 = 5.12, p < 0.001), indicating that copper accumulation may be reduced in photosynthetic algae under future pCO₂ conditions.

In outdoor culture experiments, population growth rates increased under elevated pCO₂ (1,000 and 2,000 μatm), in both the short-term (ST) or long-term (LT) acclimation experiments and in all copper treatments (Figures 3A,B; F_1,69 = 28.40, p < 0.001). By comparison between ST and LT samples, the former showed the highest growth at 1,000 μatm pCO₂, whereas the latter showed a maximum at 2,000 μatm pCO₂.

Elevated pCO₂ (1,000 and 2,000 μatm) also reduced copper accumulation across all experimental treatments, even using the natural seawater where the background Cu concentration was less than 0.03 μmol L⁻¹ (Figures 3C,D; F_1,68 = 15.18, p < 0.001). Copper accumulation in the short-and long-term mirrored the pattern found for growth rates. Under the control treatment without Cu exposure, intracellular copper concentration ([Cu]_intra) in ST samples was reduced by 38.0 and 30.5% under 1,000 and 2,000 μatm pCO₂ respectively, and that in LT samples was reduced by 31.7 and 46.3% under 1,000 and 2,000 μatm pCO₂, respectively. Under Cu exposure, [Cu]_intra in ST samples was reduced by 33.4 and 29.5% under 1,000 and 2,000 μatm, respectively, and that in LT samples was reduced by 37.0 and 43.7% under 1,000 and 2,000 μatm pCO₂, respectively_.

Transcriptome analysis as well as real-time quantitative polymerase chain reaction (RT-qPCR; Figure 4A; Supplementary Figure S5) revealed 32 differentially expressed genes (Supplementary Dataset 1) associated with copper metabolism among the different treatments. Although the Cu⁺ transporter CTR was up-regulated under elevated pCO₂, the reduction reaction from Cu²⁺ to Cu⁺ by FRE was significantly down-regulated. As a result, Cu²⁺ uptake rate was reduced (Supplementary Figure S6). Within the cell, glutathione synthetase (GS), glutathione reductase (GR), and phytochelatin synthetase (PCS) were up-regulated to synthesize phytochelatins (PC_n), which can be used to chelate with Cu⁺ and sequester it in the vacuole. Elevated pCO₂ also resulted in down-regulation of the gene expression of Cu chaperones (COX17 and SCO1) to decrease Cu toxicity to mitochondrion. Cu-transporting P1B-type ATPases (CTP) were up-regulated under elevated pCO₂ to enhance the efflux of intracellular Cu⁺ from the cell, this was also verified by a measured increase in Cu efflux rate under elevated pCO₂ (Supplementary Figure S7). Additionally, to eliminate reactive oxygen species (ROS) induced by Cu stress under elevated pCO₂, glutathione peroxidase (GPX) and catalase (CAT) were up-regulated both in gene expression and enzyme activity (Figure 4C). However, ascorbate peroxidase (APX) showed decreased activity under higher pCO₂ (1,000 and 2,000 μatm), compared with ambient pCO₂ (420 μatm), when exposed to copper.

Many studies have found different responses between short-term acclimation and long-term adaptation of phytoplankton to elevated pCO₂ (Schlüter et al., 2016; Li et al., 2017; Tong et al., 2018). Our findings are in line with previous studies indicating selection under elevated pCO₂ promoted algal growth rates in diatom species (Wu et al., 2014; Zhu et al., 2017; Qu et al., 2018). The positive effect of elevated pCO₂ may result from the reduced energy requirement of carbon concentration mechanisms (CCMs) with the saved energy being used to support carbon fixation and growth (Raven et al., 2011). Hopkinson et al. (2011) reported that doubling of ambient pCO₂ reduced CCM-related energy expenditure by ~20% and decreased the total energy demand on carbon fixation by up to 6%. As the selection experiment continued, the response differences between ambient and elevated pCO₂ decreased among different treatments after 24 months selection, which may be due to an effect induced by adaptation to the laboratory setting (Collins, 2016). Shift experiments further confirmed the evolutionary responses of T. pseudonana to high pCO₂, in that the long-term selected samples under elevated pCO₂ showed a lower growth rate than ambient selected ones when they were transplanted back into ambient pCO₂ conditions. Collins and Bell (2004) found that the long-term selected cells under high pCO₂ showed less efficient carbon concentration mechanisms, or a higher per-cell requirement for inorganic carbon, which could result in decreased growth rate when they were assayed under ambient pCO₂.

When the algae were exposed to copper stress, elevated pCO₂ significantly reduced intracellular copper accumulation and alleviated the negative effect of copper stress (Supplementary Figure S8). Changes to the seawater carbonate chemistry under elevated pCO₂ did not change external copper concentration in the culture medium (Supplementary Figure S9), suggesting that reduced copper accumulation under elevated pCO₂ is biologically mediated. This is supported by the study of de los Santos et al. (2019) who also found that projected OA would ameliorate copper toxicity on photosynthetic performance of Zostera noltei when pH decreased from 8.36 to 8.03. Our findings add to the growing body of evidence that adaptive evolution of the marine diatom community under projected OA would increase their resilience to harsh environments (Domingues et al., 2014; Valenzuela et al., 2018; Dong et al., 2020; Xu et al., 2022).

Previous studies have documented that elevated pCO₂ does not only directly affect primary producers but also changes the distribution, speciation and bioavailability of organic and inorganic trace metals and will therefore modify their interaction with organisms (Millero et al., 2009; Campbell et al., 2014; Stockdale et al., 2016). Decreasing pH could increase the concentration of Cu²⁺ by reducing complex formation with CO32-and OH⁻, thereby altering the bioavailability and increasing copper toxicity to marine organisms (Millero et al., 2009; Roberts et al., 2013). On the other hand, there is competition between H⁺ and Cu to prevent Cu²⁺ from binding at the cell surface (Gao et al., 2017). The decreased inhibitory effect of Cu on Ulva prolifera under elevated pCO₂ (1,000 μatm) suggests that the competition between H⁺ and Cu outcompetes the increased availability of Cu²⁺ in the medium (Gao et al., 2017). Here, although the effect of increasing pCO₂ on Cu speciation in the medium were not determined, the free Cu²⁺ concentration should increase in background seawater medium as Millero et al. (2009) has previously indicated. However, results consistently found reduced Cu bioaccumulation in T. pseudonana under elevated pCO₂, both with and without additional Cu exposure (Figures 1–3). Therefore, it suggested that the alleviation of toxicity could be due to the elevated CO₂ per se but not to the reduced pH.

Transcriptome analysis indicated that algal selected under elevated pCO₂ performed a specific copper detoxification strategy, that includes down-regulation of the reduction reaction from Cu²⁺ to Cu⁺ at the cell membrane to decrease copper uptake, up-regulation of biosynthesis of phytochelatins to transform free toxic Cu⁺ to less toxic organic forms, and up-regulation of a Cu⁺ transporter to enhance Cu⁺ efflux from the cell to decrease the concentration of free ions of Cu in the algal cells. All these processes could mitigate the oxidative stress in cells and enhance its tolerance to Cu exposure (Figure 4). Recent studies have demonstrated that Cu²⁺ is reduced to Cu⁺ extracellularly, which is an obligatory first step in Cu uptake in an oceanic diatom and is mediated by biological processes (Kong and Price, 2020).

Our results indicated that 1 and 20 μΜ Cu exposure inhibited growth and induced significant oxidative stress and damage in T. pseudonana (Figures 1, 4). Although Cu is one of the redox trace elements in biological systems and a basic cofactor of many enzymes (Wang and Ki, 2019), excess copper induces oxidative stress in the diatom and increases accumulation of reactive oxygen species (ROS), which would destroy macromolecules such as proteins, nucleic acids and lipids (Anu et al., 2016). In response to the oxidative stress resulting from copper exposure, Chlamydomonas reinhardtii was found to increase activities of the antioxidant enzymes glutathione S-transferase (GST), glutathione peroxidase (GPX), superoxide dismutase (SOD) and peroxidase (POD) to eliminate ROS (Zheng et al., 2011; Jiang et al., 2016). Bielmyer-Frasera et al. (2018) studied the physiological responses of two coral species, Acropora cervicornis and Pocillopora damicornis to OA and copper exposure. They found that copper exposure increased activities of the antioxidant enzymes CAT, GPX, and GR of these two species. By comparison with copper exposure alone, copper exposure under high pCO₂ (1,000 μatm) further increased activities of CAT, GPX, GR of Acropora cervicornis. This is supported by our results in that the high pCO₂ selected T. pseudonana showed higher activities of GPX and CAT when exposed to external stress than ambient-grown cells did.