Effect of carbonic anhydrase on silicate weathering and carbonate formation at present day CO2 concentrations compared to primordial values

It is widely recognized that carbonic anhydrase (CA) participates in silicate weathering and carbonate formation. Nevertheless, it is still not known if the magnitude of the effect produced by CA on surface rock evolution changes or not. In this work, CA gene expression from Bacillus mucilaginosus and the effects of recombination protein on wollastonite dissolution and carbonate formation under different conditions are explored. Real-time fluorescent quantitative PCR was used to explore the correlation between CA gene expression and sufficiency or deficiency in calcium and CO2 concentration. The results show that the expression of CA genes is negatively correlated with both CO2 concentration and ease of obtaining soluble calcium. A pure form of the protein of interest (CA) is obtained by cloning, heterologous expression, and purification. The results from tests of the recombination protein on wollastonite dissolution and carbonate formation at different levels of CO2 concentration show that the magnitudes of the effects of CA and CO2 concentration are negatively correlated. These results suggest that the effects of microbial CA in relation to silicate weathering and carbonate formation may have increased importance at the modern atmospheric CO2 concentration compared to 3 billion years ago.

It is widely recognized that carbonic anhydrase (CA) participates in silicate weathering and carbonate formation. Nevertheless, it is still not known if the magnitude of the effect produced by CA on surface rock evolution changes or not. In this work, CA gene expression from Bacillus mucilaginosus and the effects of recombination protein on wollastonite dissolution and carbonate formation under different conditions are explored. Real-time fluorescent quantitative PCR was used to explore the correlation between CA gene expression and sufficiency or deficiency in calcium and CO 2 concentration. The results show that the expression of CA genes is negatively correlated with both CO 2 concentration and ease of obtaining soluble calcium. A pure form of the protein of interest (CA) is obtained by cloning, heterologous expression, and purification. The results from tests of the recombination protein on wollastonite dissolution and carbonate formation at different levels of CO 2 concentration show that the magnitudes of the effects of CA and CO 2 concentration are negatively correlated. These results suggest that the effects of microbial CA in relation to silicate weathering and carbonate formation may have increased importance at the modern atmospheric CO 2 concentration compared to 3 billion years ago.
T he evolution of the Earth has been a complex, long-term process 1 . The overall trend in the composition of its surface minerals has involved a constant decrease in silicate and an increase in carbonate minerals. Physical 2,3 and chemical 4 weathering processes are the main forces driving silicate weathering. In recent decades, the fact that living creatures, especially microorganisms, are involved in mineral weathering has been recognized by a growing number of researchers [5][6][7] . Microbial weathering results from a combination of many factors 8 including: bio-mechanical action, the secretion of organic acids, chelation effects, redox reactions, and others. Participation of some active substances during biological weathering makes mineral weathering and enzymatic action more closely linked. Thus, it is worth exploring whether organisms can secrete enzymes to accelerate the weathering of silicate minerals or not and, if they do, how big a role the biological enzymes play in different habitats.
The free metal ions that arise from silicate weathering are involved in the precipitation of carbonates, and this process is accompanied by the fixing of atmospheric CO 2 [9][10][11] . An important constraint on the formation of carbonates is the concentration of carbonate (CO 3 22 ) in the metallogenic environment 12 . The acceleration of carbonate formation due to the action of biological enzymes is thus attributed to the increased formation of HCO 3 2 and CO 3 22 in the carbonate deposition process. Carbonic anhydrase (CA) was first found in human erythrocytes 13 and is widely present in animals, plants, and microorganisms. CA shows appreciable CO 2 hydrase activity (catalytic constants k cat lie in the range 3.9-8.0 3 10 5 s 21 and kinetic efficiencies k cat /K m are in the range 4.3-9.7 3 10 7 M 21 s 21 ) 14 . Thus, CA is capable of catalyzing the reversible hydration reaction, CO 2 , of atmospheric and self-generated CO 2 15,16 . It has been found that when the environmental CO 2 concentration changes, organisms may be able to regulate the expression level of the CA gene to adapt to those changes 17,18 . For example, the CA gene expression level in mature leaves of young legumes changes following the diversification of CO 2 concentration: the CA expression level is reduced if the CO 2 concentration is elevated 19 . Chlamydomonas reinhardtii will also increase its CA expression level to take full advantage of CO 2 when the CO 2 concentration decreases from 5 to 0.04% 20 . The expression level of their CA genes is increased at lower CO 2 concentrations. The above research shows that CA may not work at higher CO 2 concentrations or, perhaps, that it has a more important role in the face of a CO 2 deficiency. Consequently, it seems more meaningful to express this kind of gene to capture CO 2 when available levels are low.
When microorganisms grow in environments that have no limits on the availability of elements, many metabolic pathways become very slow (or even stop) to avoid unnecessary material and energy use. A more efficient, economical way is always chosen if they grow in relatively harsh conditions. Expression levels of one, or several, genes will be different according to the difficulty in obtaining nutrition. An anaplerotic role for CA has been proposed which, for example, accounts for the unusual behaviour observed in terrestrial cyanobacteria such as Nostoc flagelliforme during hydration-dehydration cycles 21 . The present authors recently showed that the level of CA gene expression in Aspergillus fumigatus 22 and Aspergillus niger 23 is enhanced if the only potassium source available is potassium feldspar (to allow the organisms to obtain potassium more effectively). As it accelerates CO 2 hydration, CA can promote the generation of H 2 CO 3 , thus promoting weathering of silicate minerals and facilitating the release of K 1 . Moreover, the increased expression of CA by Bacillus mucilaginosus favours its survival when the growth environment lacks Ca 21 but is rich in calcite 24 . Therefore, an enhanced expression level of the CA gene has a positive impact on microbial growth in environments in which soluble mineral elements are lacking but mineral particles are abundant. The microorganisms not only acquire mineral nutrition but also, at the same time, accelerate the weathering of silicate or calcite. Thus, biological adaptation, with the aid of CA, makes carbon, calcium, and silicon circulation more active.
Carbonate formation is not only an important part of the evolution of surface minerals but also a significant method of fixing atmospheric CO 2 11,25 . Microbial lithification may be the by-product of metabolism 26,27 . Some organisms can actively capture CO 2 and convert it into solid carbonate through CA catalysis 28 . When the CO 2 concentration is reduced by several orders of magnitude, biomineralization behaviour (in which CA takes part) may affect the growth and even survival of the organism. Previous studies have confirmed that many organisms, such as microbes 29,30 , coral 31,32 and animals 33,34 , can take advantage of CA's role in CaCO 3 formation at atmospheric levels of CO 2 . Miyamoto et al., for example, showed that CA from the nacreous layer in oyster pearls is conducive to the formation of CaCO 3 crystals 35 . Moreover, CA accelerates deposition of minerals and shows greater activity at low CO 2 concentrations 36 . It has also been reported that CA can contribute to carbonate precipitation at high concentrations 37 . Thus, there is no definitive conclusion as to whether the role of CA is more obvious with a reduction of CO 2 concentration during CaCO 3 deposition, or not.
In the work presented here, we use real-time quantitative PCR (RT-qPCR) to study the effect of sufficiency or deficiency in calcium and CO 2 concentrations on CA gene expression. Inversely, the function of CA in wollastonite dissolution and CaCO 3 formation, at different CO 2 concentrations, was investigated using heterologous expression and protein purification. The object of the study is to explore whether the magnitude of the silicate weathering and carbonate formation produced by CA is different at the modern atmospheric CO 2 concentration compared to that 3 billion years ago.

Results
The involvement of CA in wollastonite weathering at the atmospheric CO 2 level. The results from Experiment 1 are shown in Fig. 1 (see the Methods section for details on the different experiments performed). The trends in the pH variation for the two treatments (i.e. with and without CaCl 2 ) are similar (Fig. 1a). There was a sharp initial decrease in pH from day 0 (the primary culture) to day 2. In the days which followed, the pH rose slightly. A significant difference was that a moderate reduction in pH occurred with wollastonite as the only calcium source (compared with that containing CaCl 2 ) from day 4 to day 6.
The soluble silicon content (SSiC) of the group with added CaCl 2 was significantly greater than that in the other group on day 2 (Fig. 1b). However, the differences were not statistically significant (p 5 0.09 and 0.37, respectively) when the two conditions were compared on days 4 and 6 ( Fig. 1b). From day 4 to 6, the SSiC did not increase appreciably (p 5 0.066) when the medium contained CaCl 2 . However, there was a statistical difference (p 5 0.033) when the medium only contained wollastonite.
As far as the effect of sufficiency or deficiency in calcium on CA gene expression is concerned (see Experiment 2 in the Methods), none of the CA genes showed markedly different expressions in the two conditions on days 2 and 4 (Fig. 2). The low expression of CA genes and no difference between expressions in the two culture conditions in the early-and mid-growth stages, demonstrates that CA function may not be essential at these points. All five CA genes showed much higher expression levels on day 6 when only wollastonite was present compared to when wollastonite and CaCl 2 were used ( Fig. 2). Furthermore, there was a sharp increase in expression of all genes from day 4 to day 6 when the culture was deficient in calcium. These results indicate that the participation of CA was urgently needed to accelerate wollastonite dissolution in order to provide Ca 21 under such conditions. If B. mucilaginosus sensed calcium deficiency, significantly increased expression levels of the five CA genes were observed (see Fig. 2). Nevertheless, whether or not a single CA can show any significant effect on the dissolution of wollastonite at the current atmospheric CO 2 concentration remains unanswered. The current authors tried to find an answer to this by testing the effects of recombinant protein (PCA4) from CA4 gene by heterologous expression on wollastonite dissolution (see Experiments 4 and 5 in the Methods section). The effects of PCA4 on wollastonite dissolution can be seen in Fig. 3. The size of the target protein is consistent with the actual calculated value (28.62 kDa), and no contaminating proteins remained after dialysis (Fig. 3a). The ratio of the dialysis and soluble proteins was about 151.225 (gray value), so only a small amount of protein was lost. As can be seen from the dissolution curve ( Fig. 3b), the Ca 21 concentration, after adding PCA4, was higher everywhere compared to that without it. As the reaction continued, the wollastonite dissolution reached equilibrium. There were only trace amounts of Ca 21 released after 8 h. According to the change in the amounts of Ca 21 released over time, a pseudo-second-order kinetics model 38 was constructed to describe the dissolution behaviour of the wollastonite under both conditions (see Table 1). As can be seen from the kinetics equations (Table 1), the value of the dissolution rate k after adding PCA4 was 1.402 3 10 23 mg g 21 min 21 , and 9.24 3 10 24 mg g 21 min 21 without CA.
The effect of CO 2 on CA gene expression and the decreased importance of CA for wollastonite dissolution at high CO 2 concentration. The relative expression levels of the CA genes (displayed in Fig. 4) were significantly different at different sampling times and CO 2 concentrations (see Experiment 3 in the Methods). The expression of CA3, CA4, and CA5 genes showed no obvious differences on days 2 and 4. This indicates that CO 2 does not affect the expression of these three genes at this growth stage. However, the expression of CA1, CA3, CA4, and CA5 genes on day 6 were related to CO 2 concentration. Furthermore, CO 2 concentration and CA gene expression are negatively correlated. The relative level of expression decreased three-to five-fold when the CO 2 concentration increased by two orders of magnitude. Additionally, the difference in expression levels obtained by www.nature.com/scientificreports SCIENTIFIC REPORTS | 5 : 7733 | DOI: 10.1038/srep07733 comparing days 6 and 2 reached two orders of magnitude at 0.039% CO 2 . The CA1 gene demonstrated differential expression on day 4. This suggests that the stress of Ca 21 deficiency was felt by bacteria at that time. In this case, CA1 was preferred to accelerate the dissolution of wollastonite. This selectivity allows the bacteria to not only adapt to vertiginous environments in a timely manner but also prevents a waste of materials and energy due to superfluous gene expression.
To further confirm that the role played by CA in wollastonite dissolution is decreased at higher CO 2 concentrations, the effect of PCA4 on wollastonite demineralization was determined at both CO 2 concentrations (see Experiment 5). It can be seen from the observed trends in the amount of dissolved Ca 21 (see Fig. 5a) that the dissolution of the wollastonite gradually equilibrated at 0.039% CO 2 concentration even though PCA4 was added to the reaction system as well. In contrast, Ca 21 was released continuously under high CO 2 conditions. The difference in Ca 21 concentration emerged as early as the tenth minute. As the reaction proceeded, the difference increased.
Thus, after 8 h, the Ca 21 concentration at 3.9% CO 2 exceeded twice that present at atmospheric CO 2 levels. These results suggest that CA plays a greater role at lower CO 2 concentrations than at higher CO 2 concentrations.
The impact of CO 2 on the value of CA in carbonate formation. The results on the impact of CO 2 on the role of CA in carbonate formation are shown in Fig. 5b. Regardless of whether the reaction system contains PCA4 or not, the CO 2 concentration is positively correlated with CaCO 3 production. At any CO 2 concentration, the CaCO 3 content (w/w) is significantly different due to the participation of PCA4 (p 5 5 3 10 26 , 1.2 3 10 25 , and 3.3 3 10 25 at day 2, 4, and 6, respectively). At 10% CO 2 concentration, the mass of CaCO 3 was approximately 0.065 g without PCA4 and more than 0.070 g with PCA4. The proportion of CaCO 3 that formed due to the participation of PCA4 was about 15%. At low concentrations of CO 2 (0.4%), the masses were approximately 0.005 g without PCA4 and  www.nature.com/scientificreports SCIENTIFIC REPORTS | 5 : 7733 | DOI: 10.1038/srep07733 more than 0.020 g with PCA4. The proportion of the CaCO 3 formed as a result of the recombinant protein was up by 419%, largely due to the behaviour of the CA. Thus, PCA4 causes a much greater difference in the amount of CaCO 3 at lower CO 2 concentrations. The fact that the effect of CA is more remarkable at low CO 2 concentrations, rather than the opposite, is notable.

Discussion
Organisms growing on the surfaces of rocks, and thereby causing weathering to occur, are largely there to obtain nutrition [39][40][41] . Some of the most important inorganic nutrients required for proper cell function are obtained from rocks 42 . In the experiments testing whether wollastonite can induce the expression of CA and whether the weathering behaviour caused by the participation of CA from B. mucilaginosus contributes to a certain proportion of the overall mineral weathering effect at atmospheric CO 2 levels or not, wollastonite was the only available calcium resource when the B. mucilaginosus was cultured in media lacking soluble calcium but containing wollastonite. As one group had artificially added CaCl 2 , it is illogical to describe wollastonite dissolution using Ca 21 concentration. In view of this, SSiC was used to represent wollastonite dissolution. The number of bacteria on day 2 were about (2.78 6 0.48) 3 10 7 ml 21 and (2.70 6 0.69) 3 10 7 ml 21 with and without CaCl 2 , respectively. We observed a sharp decrease in pH, which may be due to organic acids being secreted by B. mucilaginosus in both treatments. Liu et al. showed that B. mucilaginosus produces organic acids to decompose silicate minerals during its growth, e.g. oxalic acid and citric acid 43 . The overall effect of the bacteria on wollastonite weathering with added CaCl 2 was stronger, and soon afterwards more soluble silicon was released. Despite the weaker effect without added CaCl 2 , enough Ca 21 was released to meet the amount needed for bacterial growth on day 2. Consequently, CA protein may not play an obvious role in wollastonite dissolution and the CA gene expression levels showed no significant differences. As culturing continued, the consumption of organic acids may result in a slight increase in pH. In the culture condition with wollastonite as the only calcium resource, the pressure of calcium deficiency may have been felt on day 6. The wollastonite-only group had a relatively larger bacterial population, (1.21 6 0.11) 3 10 9 ml 21 , and less Ca 21 than the group containing CaCl 2 (117. 35 6 10.62 mg/L). A single unit of Ca 21 , which bacteria were able to gain, was much lower in the group without CaCl 2 . In this case (in the medium without CaCl 2 ), the demand of B. mucilaginosus for soluble calcium may be stronger. The RT-qPCR results show that the expression of all the CA genes were up-regulated by two orders of magnitude from day 4 to day 6 when wollastonite was the only calcium resource. The consistency between this increase in CA gene expression and wollastonite dissolution (Fig. 1b, from day 4 to 6) suggests that CA plays a role in the dissolution of wollastonite at atmospheric CO 2 concentrations. The dissolution of wollastonite consumed part of the H 1 produced by CO 2 hydration and generated a certain amount of HCO 3 2 : The overall change can be written: Therefore, the extent of the pH decrease was lower than that with CaCl 2 and wollastonite as calcium resources (Fig. 1a). This is due to the production of HCO 3 2 and consumption of CO 2 in the medium. There are two aspects to the facilitation of wollastonite dissolution by increasing the expression level of CA: (i) The amount of H 1 is an important factor if the bacteria is to obtain adequate Ca 21 from wollastonite dissolution;(ii) CO 2 hydration can produce HCO 3 2 , which is an important substrate for many fundamental biological pathways such as: gluconeogenesis, lipogenesis, ureagenesis, pyrimi- dine synthesis, and synthesis of several amino acids 44 . CA can participate in the formation of malonyl-CoA, which is catalyzed by acetyl-CoA carboxylase, with bicarbonate and acetyl-CoA as the substrate 45 . Therefore, CA is an important regulator of fatty acid metabolism. Synthesis of fatty acids helps to improve membrane fluidity, which has a certain effect on the efficiency of nutrient acquisition 23 .
Wollastonite dissolution proceeds according to the reaction: It can be seen from the stoichiometry of this equation that each mole of Ca 21 released consumes one mole of CO 2 , which means the rela-tionship that holds between the quantity of Ca 21 released and CO 2 consumed during the process of wollastonite dissolution is: Here V, [CO 2 ], and A represent the solution volume, CO 2 concentration, and the area of the mineral surface, respectively; R is the flux of Ca 21 from the wollastonite surface. Thus, for a given volume of solution and mineral surface area, release of Ca 21 is proportional to the consumption of CO 2 . From the results of fitting the data to a pseudo-second-order kinetic equation, it is apparent that the k values, after adding PCA4, were higher than those without PCA4. This further confirms that CA has a significant role in promoting dissolution of wollastonite at 0.039% CO 2 concentration. Therefore, enhancement of the CA expression level is an effective way to promote weathering of minerals in order to get the desired inorganic nutrients when the microorganism grows where the CO 2 concentration is low. This behaviour of the bacteria, to a certain extent, also accelerates the weathering of silicate minerals. The involvement of CA in the demineralization of silicate minerals has also become a recognized part of global biogeochemical cycles.
An increase in CA gene expression level is advantageous to the microbe's survival chances in soluble-calcium deficient environments. However, does CA have a significant accelerating effect on mineral dissolution at high CO 2 concentrations? The proportion of CO 2 in the Earth's primordial atmosphere was up to 10% 48 . Previous studies have shown that bacteria can grow in the presence of a CO 2 concentration of 5% 49,50 and even 10% 51 . In our experiment, the wollastonite underwent dissolution to varying extents at two different levels of CO 2 concentration (Fig. 5a). The low saturation level resulted in a reduction in dissolution rate and the process gradually reached dissolution equilibrium at 0.039% CO 2 concentration. In contrast, at 3.9% CO 2 concentration, the solution had a relatively high saturation level and the dissolution rate remained essentially unchanged as CO 2 continuously dissolved in the reaction. Therefore, the CO 2 primarily affected wollastonite dissolution and the function of the CA was not obvious in a sustained high-CO 2 partial pressure environment. During the early appearance of life (3 billion years ago), silicate weathering mainly occurred due to physical and chemical effects -the contribution of CA to silicate weathering at this time may have been minimal. Atmospheric CO 2 concentrations gradually decreased (by more than two orders of magnitude) during the process of terrestrial evolution. This implies that the current expression level of biological CA is far higher now than it was in the period during which life emerged. Consequently, the participation of CA in silicate weathering may be much higher now than it was three billion years ago. This means that CA played an increasingly important role in the evolution of the Earth.
Whether it is physical, chemical, or biological weathering that affects the silicate minerals, the process is always accompanied by a release of metal ions. Some can react with HCO 3 2 or CO 3 22 in aqueous solution to revert to a solid form 25 . This is also the basic process governing both silicate weathering and carbonate formation. Mineralization experiments have shown that the highest amount of CaCO 3 occurs under 10% CO 2 and yet the required HCO 3 2 or CO 3 22 during this mineralization mainly arises from spontaneous CO 2 hydration. The role played by CA in the formation of CaCO 3 crystals is only responsible for a small proportion of them. Although bacterial CA may have been helpful in promoting the formation of carbonate more than 3 billion years ago 25 , the role might have been negligible because of the small amount of biomass and high CO 2 concentration. As the atmospheric CO 2 concentration decreased, non-enzymatic CO 2 hydration reactions would have become relatively weak. However, the involvement of CA, to some extent, compensated for this reduced rate. When applying CA to capture atmospheric CO 2 , the enzyme efficiency required to accelerate CO 2 capture increases as the partial pressure of the CO 2 decreases 52 . The least amount of total mineralization was found at the minimum concentration of CO 2 , but the difference in the amount of CaCO 3 was maximized at that point. This suggests that CA is more significant at lower CO 2 concentrations. To some degree, the participation of CA mitigates the reduced rate of carbon fixation and carbonate formation due to the decrease in CO 2 concentration. In summary, CA gene expression is negatively correlated with the ease of obtaining soluble calcium and CO 2 concentration. Moreover, considering the importance of the effect of purified CA on wollastonite dissolution and CaCO 3 precipitation, the magnitude of the effect of CA is significantly weakened at higher CO 2 concentrations. In view of the value of the effect of CA at the current atmospheric CO 2 concentration and that 3 billion years ago, the results suggest that the role of microbial CA may have become increasingly more apparent and important as terrestrial surface rocks have evolved.

Methods
Minerals. The wollastonite, Ca 3 (Si 3 O 9 ), used in the present study was provided by the State Key Laboratory of Environmental Geochemistry, Institute of Geochemistry, Chinese Academy of Sciences (Guiyang, China). The mineral was crushed and washed according to the method described by Daval et al 53 . Briefly, the crushed wollastonite powder was washed using absolute ethanol and sterilized ultrapure water (18.2 MV cm 21 ) to eliminate the fine dust resulting from the grinding procedure. Analysis using X-ray diffraction showed that the wollastonite powder contained only trace amounts of calcite and quartz.
Experiment 1 -Effects of B. mucilaginosus on wollastonite dissolution. B. mucilaginosus was cultured in a nitrogen-containing medium with different calcium resources to test the effect of B. mucilaginosus on wollastonite dissolution (see Table 2). The composition of one kind of medium per litre was as follows: sucrose 10.0 g, (NH 4 ) 2 SO 4 1.0 g, CaCl 2 0.44 g, wollastonite 1.16 g, MgSO 4 0.5122 g, KCl 0.1 g, and Na 2 HPO 4 ?12H 2 O 2.507 g. The other medium consisted of the same components except for omission of the CaCl 2 . As far as possible, to avoid adsorbed metals and cell exudates being introduced into the medium during inoculation, a seed solution was used according to the description given by Fein et al. 54 (with a little modification). Briefly, the bacteria were cleaned using sterilized ultrapure water (SUW), sterilized HNO 3 (1 M), SUW, sterilized EDTA (0.001 M), and SUW, respectively. Finally, the precipitate was suspended using 20 ml of SUW and used as seed liquid for inoculation. This operation removes as much inherent calcium as possible. Meanwhile, control media were inoculated with deactivated bacteria to eliminate interference from abiotic factors or mineral dissolution only in the medium. The ratio of seed solution to medium was about 1510 (v/v). The culture conditions were set at 30uC and 130 rpm at the current atmospheric CO 2 concentration (0.039%). The pH value of the culture solution was tested at set sampling times (2, 4, and 6 days) using a pH-meter (METTLER-TOLEDO SevenEasy S20). The number of bacteria was counted using a microscope (Zeiss Axio Imager A1, Zeiss, Germany). Moreover, some of the culture solution (15 ml) was then centrifuged (10397 g, 4 uC, 30 min) using a centrifuge (Sigma 3 k30) and 5 ml of the supernatant were collected. The remaining liquid was discarded. The precipitate was re-suspended using 10 ml of 1 M ammonium acetate, broken (a minute at a time for a total of three times) using an ultrasonic cell disrupter (Sanyo Soniprep150), and cleaned ultrasonically for 30 min. Ammonium acetate solution was added again to make the total volume up to 15 ml. The solution was mixed and centrifuged (10397 g, 4 uC, 30 min) and the supernatant collected. The two kinds of supernatant were mixed in equal volumes and the concentrations of Ca 21 and SSiC were detected using ICP-AES (Thermo IRIS Intrepid II XSP). A two-tailed t-test was performed using STATISTICA 6.0 software. The data met the assumptions of the test. The mean and its standard deviation were calculated based on three independent experiments. To test the degree of difficulty (or ease) of acquiring Ca 21 on CA expression, the same culture conditions were used as in Experiment 1. After 2, 4, or 6 d culturing time, the bacteria were centrifuged (11500 g, 4 uC, 1 min). The supernatant was discarded, and the collected cells frozen in liquid nitrogen. Total RNA was then extracted (using an Invitrogen kit in accordance with the manufacturer's instructions) and reverse transcribed into cDNA. The correct RT-qPCR reaction conditions were adopted in accordance with the manufacturer's instructions (SYBRH Premix Ex Taq TM (TliRNaseH Plus), TaKaRa). As an internal reference, 16S rRNA was used (see Table S1). After optimization by testing different primers, a single melting temperature was determined for each of the six pairs of primers, 85.8, 88.2, 87.8, 85.7, 84.8, and 85.5 uC, respectively. The Ct value was recorded for subsequent analysis (when the fluorescent signal of each reaction tube reached a set threshold, the number of reaction cycles involved gives the Ct value). The mean of DD Ct was set to zero on the second day when the bacteria was cultured using wollastonite and CaCl 2 . The relative expression level (REL) was then calculated using the following formula: . Experiment 3 -Effects of CO 2 concentration on CA gene expression. The CO 2 concentration was set to either 0.039% or the higher CO 2 concentration (3.9%) to determine the effect of CO 2 concentration on CA gene expression (see Table 2). In this experiment, the calcium resource in the medium was only wollastonite. The culture conditions were the same as in Experiment 1. Bacteria collection, RNA extraction, reverse transcription, and the RT-qPCR experiment were carried out as in Experiment 2. The mean of DD Ct was set to zero on the second day at 3.9% concentration. REL was then calculated using Eq. (6).  24 . The engineered E. coli, which over-expresses CA protein from transcription and translation of the CA4 gene referred to as PCA4, was used in the present study. Our recently published research showed that PCA4 had the best solubility and activity compared to four other proteins 24 , and so it was selected for use in this study. Briefly, the CA4 gene was amplified using PCR and then two kinds of restriction endonuclease (Kpn I and Hind III) were introduced using the relevant primers. PCR products and plasmid pET30a were both digested using Kpn I and Hind III and then linked to construct the expression vector. Recombinant plasmids were introduced into E. coli BL21 to form recombinant bacteria. Protein was produced by the induction of a final concentration of 1 mM IPTG. After induction, over-expressed PCA4 was obtained using ultrasonication. As there is impure protein mixed with the PCA4, the mixed proteins were purified using Ni-NTA agarose (QIAGEN) in accordance with published research 37 . Shortly after, the mixed proteins were loaded into the Ni-NTA agarose. Then, washing with a buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 40 mM imidazole, pH 8.0) and elution buffer (50 mM NaH 2 PO 4 , 300 mM NaCl, 250 mM imidazole, pH 8.0) was carried out to remove impure proteins and collect the targeted proteins (PCA4). The eluent containing target proteins was dialyzed twice in dialysate (100 mM tris-sulfate 100, pH 8.0) for 16 h in total. The complete process of protein purification and dialysis was carried out at 4 uC. SDS-PAGE (12.5% polyacrylamide) was used to analyze the target protein as described by Laemmli 55 with a little modification. Proteins were stained using Coomassie brilliant blue R-250 and decolouration was performed until the band appeared clear. The ''gray value'' of the proteins were calculated using Photoshop software and used to represent its content.
Experiment 5 -The effect of CA on wollastonite dissolution. Ultrapure water (49 ml) was added to an Erlenmeyer flask containing 0.116 g of wollastonite; three replicates were tested. Then, 1 ml of ultrapure water and the same amount of PCA4 were rapidly added to the flasks at 35 uC and 130 rpm. Samples were collected at 0, 10, 20, 30, 60, 120, 240, and 480 min (the remaining samples were discarded after each sampling time). The liquid was filtered using a 0.45 mm filter membrane. The concentration of Ca 21 was determined by titration using ethylenediaminetetraacetic acid disodium salt (EDTA-Na 2 ). To explore whether the importance of the role of CA in wollastonite dissolution at high CO 2 concentration was similar, the same operation was carried out at 3.9% CO 2 concentration. To analyze the data, a two-tailed t-test was used. The data presented is the mean (along with the standard deviation) of three independent experiments. Experiment 6 -Mineralization reaction under different CO 2 concentrations. A 10 ml (0.2 M) portion of Tirs-HCl (pH 9.0) was mixed with an equal volume of CaCl 2 (0.2 M) in a clean, sterile Petri dish. Then, 1 ml of ultrapure water and an equal volume of PCA4 were added to the reaction system at 35 uC and rotated at 80 rpm under three different CO 2 concentrations (0.4%, 3.9%, and 10%). There were three independent replications of each treatment. After 20 min, the supernatant was discarded and the sample dried overnight at 65 uC. The residual crystalline CaCO 3 was weighed. The dry weight of 1 ml of PCA4 solution is only a few micrograms, so it is negligible in relation to the weight of the CaCO 3 formed. The statistical approach used to analyze the data was the same as that described above.