A new type of carboxysomal carbonic anhydrase in sulfur chemolithoautotrophs from alkaline environments

ABSTRACT

Autotrophic bacteria are able to fix CO₂ in a great diversity of habitats, even though this dissolved gas is relatively scarce at neutral pH and above. As many of these bacteria rely on CO₂ fixation by ribulose 1,5-bisphospate carboxylase/oxygenase (RubisCO) for biomass generation, they must compensate for the catalytical constraints of this enzyme with CO₂-concentrating mechanisms (CCMs). CCMs consist of CO₂ and HCO₃⁻ transporters and carboxysomes. Carboxysomes encapsulate RubisCO and carbonic anhydrase (CA) within a protein shell and are essential for the operation of a CCM in autotrophic Bacteria that use the Calvin-Benson-Basham cycle. Members of the genus Thiomicrospira lack genes homologous to those encoding previously described CA, and prior to this work, the mechanism of function for their carboxysomes was unclear. In this paper, we provide evidence that a member of the recently discovered iota family of carbonic anhydrase enzymes (ιCA) plays a role in CO₂ fixation by carboxysomes from members of Thiomicrospira and potentially other Bacteria. Carboxysome enrichments from Thiomicrospira pelophila and Thiomicrospira aerophila were found to have CA activity and contain ιCA, which is encoded in their carboxysome loci. When the gene encoding ιCA was interrupted in T. pelophila, cells could no longer grow under low-CO₂ conditions, and CA activity was no longer detectable in their carboxysomes. When T. pelophila ιCA was expressed in a strain of Escherichia coli lacking native CA activity, this strain recovered an ability to grow under low CO₂ conditions, and CA activity was present in crude cell extracts prepared from this strain.

IMPORTANCE

Here, we provide evidence that iota carbonic anhydrase (ιCA) plays a role in CO₂ fixation by some organisms with CO₂-concentrating mechanisms; this is the first time that ιCA has been detected in carboxysomes. While ιCA genes have been previously described in other members of bacteria, this is the first description of a physiological role for this type of carbonic anhydrase in this domain. Given its distribution in alkaliphilic autotrophic bacteria, ιCA may provide an advantage to organisms growing at high pH values and could be helpful for engineering autotrophic organisms to synthesize compounds of industrial interest under alkaline conditions.

INTRODUCTION

A significant portion of global CO₂ fixation is catalyzed within carboxysomes (1). Carboxysomes are proteinaceous microcompartments that facilitate CO₂ fixation by many autotrophic bacteria that use the Calvin-Benson Bassham cycle [CBB (2)]. Carbon fixation by the CBB cycle is catalyzed by RubisCO (ribulose 1,5-bisphosphate carboxylase/oxygenase), which has a low affinity for CO₂, and can also use O₂ as a substrate for a wasteful side reaction (3). In order to compensate for the catalytic constraints of RubisCO, many autotrophic bacteria have a CO₂-concentrating mechanism (CCM), which consists of two components: dissolved inorganic carbon (DIC; CO₂, HCO₃⁻, and CO₃²⁻) transporters, which transport DIC across the cell membrane to generate elevated levels of HCO₃⁻ within the cytoplasm, and carboxysomes (4, 5). Carboxysomes contain two key enzymes: (i) carbonic anhydrase (CA), which interconverts HCO₃⁻ and CO₂ and (ii) RubisCO, which uses CO₂ as its substrate to carboxylate ribulose 1,5-bisphosphate (6–9). When cytoplasmic HCO₃⁻ enters carboxysomes, CA converts it to CO₂, and it is fixed by RubisCO (10).

Two types of carboxysomes are known: α-carboxysomes, which are present in autotrophic Proteobacteria and some Cyanobacteria, including marine globally abundant Prochlorococcus and Synechococcus spp. (1), and β-carboxysomes, which are found exclusively in Cyanobacteria (11–13). While the two types have the same function in CCMs, they arose independently and can be distinguished by the forms of RubisCO and CA they encapsulate (9, 12). α-Carboxysomes contain form IA RubisCO and a β-class CA [CsoSCA (8)], while β-carboxysomes contain form IB RubisCO and either β-class or γ-class CA or both (14–17). Form IA and form IB RubisCO share a common ancestry, while the different forms of carboxysomal CA (β and γ) are evolutionarily independent (18). Notably, these are not the only classes of CA that exist. CA consists of a superfamily of diverse enzymes and is present in all domains of life. Currently, there are eight (α, β, γ, δ, ζ, η, Ɵ, and ι) classes of CA known, and at least six of these classes are evolutionary independent forms [reviewed in reference (5)]. All classes facilitate the reversible hydration of CO₂ to HCO₃⁻, and most require metals for activity [commonly Zn⁺² (18)].

Given the uniformity of CA presence in biochemically characterized α- and β-carboxysomes, and the critical role CA plays in carboxysome function, it was surprising to find that sulfur-oxidizing autotrophic members of genus Thiomicrospira, from phylum Gammaproteobacteria, had carboxysome loci lacking genes encoding the CA typically associated with α-carboxysomes [CsoSCA; Fig. 1 (19, 20)]. In its place, they have a smaller gene, csoSX, which does not encode residues necessary for CsoSCA activity. These differences are interesting since other evidence suggests the presence of a typical CCM in members of Thiomicrospira. Thiomicrospira pelophila and Thiomicrospira thyasirae, like many other organisms with CCMs (19, 21), have genes encoding both carboxysomal RubisCO (form I; cbbL and cbbS), as well as a presumably non-carboxysomal form II RubisCO (cbbM). Other organisms with multiple RubisCO-encoding genes upregulate cbbM when CO₂ is abundant and upregulate carboxysomal cbbL and cbbS when CO₂ is scarce as part of their CCM (22, 23). Carboxysomes have previously been noted in electron micrographs of members of Thiomicrospira (24–26), and it was puzzling that these carboxysomes could function in an autotrophic cell in the absence of CA. Recently, an additional form of CA was discovered in phytoplankton [iota carbonic anhydrase (ιCA) (27)], and a homolog to ιCA is present in carboxysome loci in all members of genus Thiomicrospira (Fig. 1). So far, ιCA has only been described and biochemically characterized in a handful of organisms [Thalassiosira pseudonana, Burkholderia territorii, Bigelowiella natans, and Anabaena sp. PCC7120 (27–30)], and homologs are common in organisms from all three domains. In Bacteria, two ιCA have been biochemically characterized (28, 29), but their roles in the physiologies of these organisms have not been studied. Given the presence of genes encoding ιCA homologs in carboxysome loci from multiple organisms, our hypothesis is that ιCA functions as the carboxysomal CA in these organisms, the first indication that ιCA may play a role in these microcompartments.

Fig 1.

Carboxysome loci of members of Thiomicrospira and others that include ιCA homologs. The carboxysome locus from Halothiobacillus neapolitanus, which is a model organism for carboxysomes in Proteobacteria, is included for comparison. For genera Thioalkalivibrio and Thiothrix, many strains have carboxysome loci with ιCA homologs nearby, and representative strains are included in this figure. Carboxysome loci with ιCA homologs were found via cassette cluster search of the Integrated Microbial Genomes and Microbiomes database (https://img.jgi.doe.gov/), to find members of Pfams 08332 (ιCA), 12288 (CsoS2), and 00101 (CbbS, RubisCO small subunit) within 20 kbp of each other on a genome sequence. Abbreviations: bfr, bacterioferritin; cbbLS, large and small subunits of RubisCO; cbbO and cbbQ, RubisCO activases (31); cbbR, LysR-type regulatory protein (32); csoS1, S1D, and S4, carboxysome shell proteins (33–35); csoS2, carboxysome assembly protein (36); csoSCA, carboxysomal carbonic anhydrase (8); csoSX, homologous to the amino terminus of csoSCA, lacking CA active-site residues encoded in csoSCA; ham1, Ham1 family protein (Pfam01725); ιCA, iota carbonic anhydrase homologs; pII, signal transduction protein (37); parA, plasmid partitioning protein homolog (Pfam01656); raf, RubisCO assembly factor (38); sbtA, bicA/sulP, and DAC, DIC transporters (19).

RESULTS

Purified carboxysomes from T. pelophila and Thiomicrospira aerophila

Carboxysomes were successfully enriched from T. pelophila and T. aerophila grown under low-DIC conditions (Table 1). Suspensions of carboxysomes were dominated by carboxysome proteins, as determined by LC-MS/MS (Table S1). A small number of major bands are visible in Coomassie blue-stained SDS-PAGE gels, and these bands were excised and identified as carboxysome proteins via LC-MS/MS (Fig. 2A). For T. aerophila carboxysome preparations, porin was present, indicating some contamination from the outer membrane. For both species, αCA was detected at low abundance in some carboxysome preparations (≤0.025% of signal intensity; Table 1) and was not enriched in carboxysome preparations. Intact and broken carboxysomes were visible via transmission electron microscopy of negatively stained carboxysome preparations (Fig. 2B).

TABLE 1.

Relative abundance of carboxysome-associated proteins in carboxysome preparations

	% of total signal intensitya
Proteins	T. pelophila carboxysomes (n = 2)	T. pelophila cells (n = 3)	T. aerophila carboxysomes (n = 2)	T. aerophila cells (n = 3)
Cso + CbbL + CbbS	82.9, 37.6	5.2 ± 0.2	41.5, 42.4	6.6 ± 0.3
CbbL	32.8, 17.2	1.90 ± 0.06	17.5, 15.8	2.43 ± 0.13
CbbS	11.2, 3.6	0.44 ± 0.07	6.77, 7.01	0.76 ± 0.22
CsoS2	10.7, 8.6	1.19 ± 0.08	7.3, 10.3	1.49 ± 0.04
CsoSX	0.99, 0.20	0.020 ± 0.001	0.10, 0.13	0.0054 ± 0.0014
CsoS4A	1.23, 0.17	0.027 ± 0.005	0.17, 0.19	0.0024 ± 0.0008
CsoS4B	0.17, 0.03	0.003 ± 0.002	0.04, 0.10	0.0040 ± 0.0023
CsoS1 (A-C)	23.6, 7.6	1.63 ± 0.08	9.31, 8.40	0.086 ± 0.079
CsoS1D	2.18, 0.29	0.044 ± 0.003	0.35, 0.45	0.040 ± 0.003
Bfrb	6.91, 1.58	0.021 ± 0.03	1.31, 1.07	0.11 ± 0.01
Ham1c	0.59, 0.10	0.016 ± 0.002	0.15, 0.18	0.035 ± 0.013
ιCA	3.53, 0.40	0.063 ± 0.007	–	–
ιCA 112d	–	–	0.07, 0.05	0.0049 ± 0.0007
ιCA 115e	–	–	0.95, 1.23	0.14 ± 0.01
αCAf	0.025, NDg	0.027 ± 0.002	ND, 0.0049	0.0075 ± 0.0002

^a % of signal intensity for all proteins detected in carboxysome samples or cells by LC-MS/MS.

^b Bacterioferritin; member of Pfam00210.

^c Ham1 family protein, XTP/dITP diphosphohydrolase; member of Pfam01725.

^d ιCA 112 is encoded in T. aerophila by IMG gene object identifier 2507134112.

^e ιCA 115 is encoded in T. aerophila by IMG gene object identifier 2507134115.

^f αCA is encoded by IMG gene object identifier 2568511177 in T. pelophila and 2507133695 in T. aerophila.

^g ND, none detected.

Fig 2.

Carboxysome preparations from wild-type T. pelophila and T. aerophila. (A) SDS-PAGE analysis of carboxysome preparations from T. pelophila and T. aerophila (100 µg protein was loaded for each; protein ladder is labeled in kDa). Proteins from the major bands were identified via LC-MS/MS, including CsoS2A and B (Fig. S1). Two bands are apparent for CbbS; the one labeled CbbS* may be post-translationally modified (please see Fig. S2 and S3, and text for further information on bands A–C). Abbreviations: Bfr, bacterioferritin; CbbL and CbbS, large and small subunits of RubisCO; CsoS1, S1D, and S4, carboxysome shell proteins (33–35); CsoS2, carboxysome assembly protein (36); CsoSX, homologous to the amino terminus of CsoSCA, lacking CA active-site residues found in CsoSCA; Ham1, Ham1 family protein (Pfam01725); ιCA, iota carbonic anhydrase homologs. (B) Transmission electron micrographs of carboxysomes negatively stained with ammonium molybdate. Carboxysomes from T. pelophila were magnified to 140,000×, and carboxysomes from T. aerophila were magnified to 100,000×.

CA activity in T. pelophila and T. aerophila carboxysomes

CA activity was present in purified carboxysomes that had been freeze-thawed to release the enzymes they contain [Fig. 3; Table 2 (39)]. For carboxysomes from either T. pelophila or T. aerophila, boiled samples had significantly lower rates of CO₂ hydration due to chemical (non-enzymatic) CO₂ hydration remaining after heat treatment (P < 0.005; Table 2). CA activities, calculated as the difference between boiled and unboiled rates, were similar for carboxysomes from both organisms (Table 2). T. pelophila carboxysome-associated CA was inhibited by ethoxzolamide (EZA) dissolved in dimethyl sulfoxide (DMSO) relative to the solvent control (EZA vs DMSO solvent control, P < 0.015; the addition of DMSO itself did not inhibit CA activity, P > 0.96). CA activity from T. aerophila carboxysomes was lower in the presence of EZA, though not to statistical significance (P < 0.1, EZA vs DMSO, P > 0.7, DMSO vs no additions). CA activity from T. aerophila carboxysomes was sensitive to dithiothreitol (P < 0.005), while CA from T. pelophila was not (P < 0.56).

Fig 3.

CO₂ hydration by carboxysomes purified from T. pelophila (A) and T. aerophila (B). CO₂ hydration was measured indirectly via the response of phenol red to assay acidification by protons released by carbonic acid formed from CO₂. CO₂ hydration was measured in three samples of saline (sal), carboxysomes (csomes), and boiled carboxysomes (boiled). Absorbances were calculated by subtracting the initial absorbance value (A₀) from each timepoint absorbance value (A_t; A₅₅₇ = A_t - A₀) to compensate for turbidity due to the presence of carboxysomes.

TABLE 2.

Activity and inhibition of carboxysome-associated carbonic anhydrase

		Carbonic anhydrase activity (µmoles CO2 s−1 mg protein−1) ±SD (n = 3)
Organism	mg protein mL−1	No inhibitors	10 mM DTTa	1% (vol/vol) DMSOb	1 mM EZAc, 1% (vol/vol) DMSO
T. pelophila	0.7	3.1 ± 0.9	3.9 ± 0.3	3.5 ± 1.0	1.5 ± 0.4
T. aerophila	2.3	2.1 ± 0.3	0 ± 0.2	2.0 ± 0.2	1.6 ± 0.3

^a DTT, dithiothreitol.

^b DMSO, dimethyl sulfoxide.

^c EZA, ethoxzolamide; 6-ethoxy-2-benzothiazolesulfonamide.

Characterization of CO₂-sensitivite T. pelophila mutants

From the library of 10,000 mutant strains of T. pelophila, 19 were unable to grow under low-CO₂ conditions (Fig. S4). Because the initial screen of the library was a growth/no growth assay under low-CO₂ conditions, it is possible that other strains in the library, beyond the 19 studied further here, were capable of lower rates of growth under low-CO₂ conditions. In these 19 strains, Tn5-RL27 was inserted within the carboxysome operon (Fig. 4A). Some of these insertions fell within the genes encoding carboxysomal RubisCO (cbbL; Fig. 4A). Presumably, these mutant strains were still capable of growth under high-CO₂ conditions via expression of cbbM, which encodes non-carboxysomal form II RubisCO encoded elsewhere in the genome. Carboxysomes were not apparent in mutants whose Tn5-RL27 insertion sites fell within genes encoding shell-associated proteins (csoS2 and csoSX; Fig. 4C and D). However, mutants with an interruption within the gene encoding the ιCA homolog were able to generate intact carboxysomes (Fig. 4E and F).

Fig 4.

CO₂-sensitive T. pelophila strains generated by transposon-mediated random mutagenesis. (A) Transposon insertion sites (red arrows) present in the CO₂-sensitive strains (one per strain); exact locations of insertion sites are indicated in Fig. S4. Transmission electron micrographs (14,000× magnification) of wild-type T. pelophila (B) and CO₂-sensitive mutants (C–F) with arrows indicating carboxysomes, and numbers after gene names indicating the location of the transposon insertion site in nucleotides from the first nucleotide of the start codon of the targeted gene. In (C), a cell has inclusions marked with (?), which appear to be slightly smaller than wild-type carboxysomes. Abbreviations: bfr, bacterioferritin; cbbL and cbbS, large and small subunits of RubisCO; cbbR, LysR-type regulatory protein (32); csoS1, S1D, and S4, carboxysome shell proteins (33–35); csoS2, carboxysome assembly protein (36); cbbO and cbbQ, RubisCO activases (31); csoSX, homologous to the amino terminus of csoSCA, lacking CA active site residues encoded in csoSCA; ham1, Ham1 family protein (Pfam01725); ιCA, iota carbonic anhydrase homologs; pII, signal transduction protein (37); parA, plasmid partitioning protein homolog (Pfam01656); raf, RubisCO assembly factor (38); sbtA, bicA/sulP, and DAC, DIC transporters (19).

Growth and CA activities of Escherichia coli constructs expressing genes from T. pelophila

E. coli Lemo21(DE3) ΔyadF ΔcynT successfully expressed CsoS2, CsoSX, and ιCA from T. pelophila, based on the presence of these proteins in cell pellets (Table S2). Cells expressing CsoS2 or CsoSX were unable to grow under low-CO₂ conditions despite growing robustly under a 5% CO₂ headspace (Fig. 5). When expressing the ιCA homolog from T. pelophila, E. coli Lemo21(DE3) ΔyadF ΔcynT was able to grow under ambient air (0.04% CO₂; Fig. 5). This construct had measurable CA activity 6 hours after inducer was added (6 hours: 0.53 ± 0.02 µmoles CO₂ s⁻¹ mg protein⁻¹, P < 0.002; 8 hours: 0.36 ± 0.01 µmoles CO₂ s⁻¹ mg protein⁻¹, P < 0.05; 24 hours: 0.49 ± 0.03 µmoles CO₂ s⁻¹ mg protein⁻¹, P < 0.05). CA activity was not detected in E. coli Lemo21(DE3) ΔyadF ΔcynT in the absence of genes from T. pelophila (P > 0.05 at all timepoints).

Fig 5.

Growth curves of E. coli Lemo21(DE3) ΔyadF ΔcynT (double carbonic anhydrase knockout) expressing genes required for low CO₂ growth by T. pelophila. Duplicate cultures (1, 2) expressing genes controlled by the P_BAD promoter were grown. Top: csoSX (homologous to the amino terminus from csoSCA); middle: csoS2 (shell assembly protein); bottom: homolog of ιCA. Induced cells were cultivated in a medium supplemented with 6 mM arabinose, while the growth medium for repressed cells was supplemented with 11 mM glucose.

Characterization of carboxysomes purified from T. pelophila with a mutation in the gene encoding ιCA

Carboxysomes from mutant T. pelophila strain ιCA-135 were successfully purified. When subject to SDS-PAGE analysis, a small number of major bands were visible in Coomassie blue-stained gels (Fig. 6A), and LC-MS/MS analysis confirmed that carboxysome proteins comprised 92% of all proteins in the sample (Table S1). Intact and broken carboxysomes were visible in transmission electron micrographs (Fig. 6B). Relative amounts of many carboxysome proteins were similar in mutant and wild-type carboxysomes. However, amounts of ιCA and CsoS1D were lower in two independent preparations of mutant carboxysomes (Fig. 6C).

Fig 6.

Carboxysomes from T. pelophila ιCA-135. (A) SDS-PAGE of a carboxysome preparation from T. pelophila ιCA-135 (100 µg protein; protein ladder labeled in kDa). (B) Transmission electron micrographs of carboxysomes from T. pelophila ιCA-135 negatively stained with ammonium molybdate. Carboxysomes were magnified 56,000×. (C) Relative amounts of carboxysome proteins in two independent replicate carboxysome preparations from wild-type T. pelophila (wild-type rep 1 and wild-type rep 2) and T. pelophila ιCA-135 (ιCA-135 rep 1 and ιCA-135 rep 2). % of csome proteins is calculated from the intensity of LC-MS/MS signal from a particular protein (e.g., CbbL) divided by the total intensity of LC-MS/MS signal from all carboxysome-associated proteins (CbbL + CbbS + CsoSX + CsoS4A + CsoS4B + CsoS1.1 + CsoS1.2 + ιCA + CsoS1D). Abbreviations: CbbL, large subunit of RubisCO; CsoS1.1, CsoS1.2, CsoS1D, CsoS4A, and CsoS4B, carboxysomal shell proteins; CsoS2, carboxysomal shell assembly protein; CsoS4a CsoSX, homologous to the amino terminus of csoSCA, lacking CA active-site residues encoded in csoSCA; ιCA, homolog of ιCA. (D) Signal intensities from ιCA-derived peptides from two independent replicate carboxysome preparations from wild-type T. pelophila (wild-type rep 1 and wild-type rep 2) and T. pelophila ιCA-135 (ιCA-135 rep 1 and ιCA-135 rep 2). Signal intensities from peptides resulting from ιCA protein fragmentation for LC-MS/MS analysis were summed and used as the denominator to calculate the percent of signal from individual ιCA peptides. Numbers on the x-axis correspond to the location of the peptide in the complete ιCA sequence. (E) CO₂ hydration by carboxysomes isolated from T. pelophila ιCA-135. CO₂ hydration was measured indirectly via the response of phenol red to assay acidification by protons released by carbonic acid formed from CO₂. CO₂ hydration was measured in three samples of saline (sal), carboxysomes (csomes), and boiled carboxysomes (boiled). Absorbances were calculated by subtracting the initial absorbance value (A₀) from each timepoint absorbance value (A_t; A₅₅₇ = A_t - A₀) to compensate for turbidity due to the presence of carboxysomes.

The ιCA present in mutant carboxysomes consists of the amino terminus of the protein (Fig. 6D), which reflects the insertion site of the transposon. The diminishment in CsoS1D abundance likely reflects the polar effects of transposon insertion in the gene encoding ιCA, which precedes the gene encoding CsoS1D (Fig. 1). Additionally, one of the lower molecular weight bands from wild-type preparations (Fig. 2; labeled A–C) is absent in mutant carboxyomes (Fig. 6A; labeled B and C). The identities of the proteins in these bands were confirmed by LC-MS/MS analysis of bands excised from SDS-PAGE gels. The composition of bands B and C from mutant carboxysomes resembles the compositions of their counterparts from wild-type cells (Fig. S2), suggesting the loss of the proteins comprising band A from wild-type carboxysomes in mutant carboxysomes. Given that band A in wild-type carboxysome preparations is similar in composition to band B and consists primarily of CbbS (Fig. S2), it is possible that CbbS (or CsoS1.1 and CsoS1.2, which are also present in band A) is post-translationally modified in wild-type cells. The abundance of peptide fragments generated during LC-MS/MS from CbbS, CsoS1.1, and CsoS1.2 was compared (Fig. S3), figuring that post-translational modification of the intact protein at a particular amino acid residue could change the molecular weight and/or charge of the fragment, resulting loss of detection. Indeed, protein fragments corresponding to CbbS (peptide corresponding to residues 70–83), CsoS1.1 (peptides corresponding to residues 1–15 and 39–51), and CsoS1.2 (same as CsoS1.1) are absent in wild-type band A, which is consistent with potential post-translational modification of these proteins (Fig. S3). It is possible that the gene necessary for post-translational modification of these proteins is encoded downstream from iCA, and its expression was diminished similarly to CsoS1D.

CA activity was not detected in the mutant carboxysomes (Fig. 6E). There was no significant difference in CA activity between the carboxysome suspension and the boiled control (P > 0.12). The concentration of protein in the assay was 0.4 mg mL⁻¹, lower than the assay of wild-type carboxysomes (0.7 mg mL⁻¹). However, carboxysome proteins were more enriched in preparations of mutant carboxysomes (92%) than for the preparation from wild-type T. pelophila used for this assay (38%), suggesting that the negative result is not a dilution effect. These carboxysomes had RubisCO activities (0.3 µmol CO₂/min mg protein) similar to wild-type carboxysomes (two independent preparations of wild-type carboxysomes: 0.8 and 0.4 µmol min⁻¹ mg⁻¹ protein, respectively), suggesting that the absence of CA activity was not due to carboxysome breakage and loss of contents (e.g., RubisCO, CA) during purification.

Phylogenetic analysis of ιCA homologs from genus Thiomicrospira

Genes encoding ιCA from Thiomicrospira spp. fall among those from other members of Gammaproteobacteria, divergent from those previously described and biochemically characterized (Fig. 7). For those Thiomicrospira species with two ιCA homologs, these genes fall into two separate clusters.

Fig 7.

Maximum likelihood analysis of ιCA amino acid sequences. Taxon names of biochemically characterized ιCA are underlined (28, 29, 40). ●, carboxysome locus-associated, with csoSX in the carboxysome locus; ○, carboxysome locus-associated, with csoSCA in the carboxysome locus; green-thickened branches, amino acid sequences with multiple cysteine residues near the amino terminus. Numbers preceding the taxon names are gene object ID numbers from IMG [https://img.jgi.doe.gov/ (41)] and accession numbers from the nonredundant protein sequences database (https://www.ncbi.nlm.nih.gov/). Sixty-nine sequences with 291 positions were used to generate the tree using the Le_Gascuel_2008 model (42) and a gamma distribution with five categories. The scale indicates the numbers of substitutions per site. Bootstrap values >70 are shown and are based on 100 resamplings of the alignment.

DISCUSSION

ιCA is likely to function as the carboxysomal CA in the genus Thiomicrospira

Multiple lines of evidence presented here are consistent with the displacement of CsoSCA, formerly the only CA associated with α-carboxysomes, with ιCA in carboxysomes from members of Thiomicrospira. Genes encoding ιCA homologs are present in Thiomicrospira carboxysome loci, which lack genes encoding CsoSCA for all members of this genus sequenced thus far (Fig. 1). Though CsoSCA is absent, Thiomicrospira carboxysomes have CA activity (Table 2; Fig. 3), ιCA proteins are present in preparations of purified carboxysomes (Table 1; Table S1), interruption of the gene encoding ιCA results in loss of CA activity in carboxysomes (Fig. 6), and heterologous expression of ιCA genes from T. pelophila in E. coli Lemo21(DE3) ΔyadF ΔcynT (double CA knockout) rescues the strain’s ability to grow under low CO₂ conditions (Fig. 5) and results in measurable CA activity. To our knowledge, this is the first study to show the replacement of CsoSCA in α-carboxysomes, as well as a physiological function for ιCA in Bacteria.

ιCA enzymes from members of the genus Thiomicrospira share many similarities to those previously described in other Bacteria. Residues corresponding to those present in the active site of ιCA from Nostoc sp. PCC7120 [T106, Y127, H197, and S199; HHSS motif at the carboxy terminus (29)] are present in ιCA from Thiomicrospiras (Fig. S5). Similarities extend to predicted tertiary structure. Structures predicted via Alphafold2 (43) are similar to the structure empirically determined for ιCA from Nostoc sp. PCC7120 [Fig. 8 (29)].

Fig 8.

Structures predicted for ιCA homologs from members of genus Thiomicrospira. (A and B) Predicted structure of ιCA from T. pelophila via Alphafold2 (43), implemented via Colabfold (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb) using default parameters, without comparison to Protein Data Bank (PDB) structures. (A) Comparison of ιCA sequence from T. pelophila to homologs used to predict its structure. (B) Predicted structure of T. pelophila ιCA, shaded by confidence (see legend). (C) Superposition of structures predicted for ιCA homologs from T. pelophila and T. aerophila onto the empirically determined structure of ιCA from Nostoc sp. PCC7120 [PDB ID 7C5V (29)], and a table of results from pairwise structural alignments of Thiomicrospira ιCA proteins to ιCA from Nostoc sp. PCC7120. Predicted structures for the Thiomicrospira ιCA homologs were compared to 7C5V via TM-Align (44) implemented at RSCB Protein Data Bank (https://www.rcsb.org/alignment). Abbreviations: pLDDT, predicted local distance difference test, a measure of confidence in local structure (45); RMSD, root mean square deviation of distances in angstroms between alpha carbons in compared structures; TM-score, template modeling score, proportional to the agreement in topology between compared structures.

CsoSX is not a carboxysomal CA

T. pelophila CsoSX presence does not result in the growth of E. coli Lemo21(DE3) ΔyadF ΔcynT under low CO₂ conditions (Fig. 5), consistent with an absence of CA activity. CsoSX sequences from Thiomicrospiras (250–332 aa) are substantially shorter than CsoSCA sequences (471–531 aa) and lack all of the residues necessary for CA activity by CsoSCA (Fig. S6; residues corresponding to C173, H242, and C253 in H. neapolitanus) (18, 46). Predicted structures for CsoSX proteins (Fig. 9A) are low-confidence and do not resemble the empirically determined structure for CsoSCA from H. neapolitanus [PDB 2FGY (46); root mean square deviation of distances is high, >6 Å, and template modeling scores are low, ~0.2].

Fig 9.

Comparison of CsoSX proteins from members of genus Thiomicrospira. (A) Structures predicted for CsoSX from T. pelophila and T. aerophila via Alphafold2 (43), implemented via Colabfold (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb) using default parameters, without comparison to PDB structures. Abbreviations: pLDDT, predicted local distance difference test, a measure of confidence in local structure (45). (B) Sequence alignment [via MUSCLE, implemented in MEGA11 (47, 48)], and (C) sequence logo created from the alignment of CsoSX amino acid sequences from members of genus Thiomicrospira with CsoSCA from H. neapolitanus. The complete logo is depicted in Fig. S6.

Despite this lack of predicted structural similarity to CsoSCA, it is likely that CsoSX is a modified version of the amino terminal region of CsoSCA. All CsoSX proteins share a short amino terminal sequence motif with CsoSCA proteins (RXPXGXXXV, corresponding to H. neapolitanus CsoSCA residues 11–20; Fig. 9B and C). Additionally, the amino terminal region of CsoSX from T. thyasirae (residues 63–131) matches Interpro family IPR043065 [N-terminal domain of CsoSCA, e < 2E-5; https://www.ebi.ac.uk/interpro/ (49)]. The remainder of the sequence (residues 132–332) has no match, which is unlike CsoSCA sequences, which also match IPR043066 (carboxysome shell carbonic anhydrase and C-terminal domain superfamily).

Conservation of the amino terminal residues of CsoSX suggests that the presence of the protein has been selected for, and CsoSX is indeed present in carboxysome preparations (Table 1; Table S1). However, CsoSX proteins appear to be unique to Thiomicrospira; none of the other CsoSX sequences from Thiomicrospira match Interpro families nor do they match sequences in the nonredundant protein database or structures available to searches with Foldseek [https://search.foldseek.com/search (50)]. The amino termini of CsoSCA proteins are disordered and interact with CbbL and CbbS [Fig. 10 (51)]. The amino termini of CsoSX proteins are predicted to lack much secondary structure; perhaps they too are disordered (Fig. 9). It is possible that CsoSX interacts with CbbL and CbbS and facilitates packing them into Thiomicrospira carboxysomes.

Fig 10.

Potential redox sensitivity of amino termini of some carboxysome-associated carbonic anhydrase enzymes. (A) Structure of CsoSCA from H. neapolitanus predicted via Alphafold2 (43), implemented via Colabfold (https://colab.research.google.com/github/sokrypton/ColabFold/blob/main/AlphaFold2.ipynb) using the “pdb100” option for template_mold since an empirically determined structure is available for this protein [PDB 2FGY (46)]. A predicted structure was generated because the published structure does not include the amino terminus. The published structure and this predicted structure are in good agreement (RMSD = 0.66; TM-score = 0.99; TM-Align) (44) and implemented at RSCB Protein Data Bank (https://www.rcsb.org/alignment). Predicted disordered regions of CsoSCA (B), and ιCA from members of genus Thiomicrospira (C), and their potential sensitivity to redox status. “Position” is with respect to the residue number in the amino acid sequence, and “score” is the disorder prediction score, which indicates the probability that a residue is located in a disordered region of the protein. “Minus” and “plus” refer to predicted disorder in the absence (minus; cysteine replaced by alanine) and presence (plus; native sequence) of redox-sensitive amino acids. Disordered regions and impact from redox status were predicted via IUPred2A [(52); https://iupred2a.elte.hu/]. Abbreviations: pLDDT, predicted local distance difference test, a measure of confidence in local structure (45).

Redox sensitivity varies among ιca proteins

The sensitivity of carboxysome CA activity to the reducing agent dithiothreitol (DTT) differs between T. pelophila and aerophila (Table 2). T. pelophila carboxysomal CA activity is insensitive to DTT, while that from T. aerophila is completely inhibited by this compound. Of the two ιCA proteins present in T. aerophila carboxysomes, ιCA115 is present at an order-of-magnitude excess relative to ιCA112 (Table 1), so this sensitivity to DTT likely reflects an impact on ιCA115 activity. Interestingly, CsoSCA from H. neapolitanus is also inhibited by DTT (39). For both T. aerophila ιCA115 and H. neapolitanus CsoSCA, the relative disorder of the structure of the amino terminus is predicted to change due to redox status (e.g., due to disulfide bond formation between cysteine residues); such changes are not predicted for T. pelophila ιCA or T. aerophila ιCA112 (Fig. 10). The amino termini of the T. aerophila ιCA115 and H. neapolitanus CsoSCA have more cysteine residues (two, C54 and C57; and three, C41, C49, and C65, respectively) than the others (one each; C54 for T. pelophila and C53 for T. aerophila ιCA112). Perhaps the sensitivities of some ιCA enzymes to DTT are due to the loss of intra- or intermolecular disulfide bonds in the presence of this reducing agent.

It is tempting to speculate that these α-carboxysomal CA enzymes are redox-regulated, as is the case for the β- and γCA enzymes present in β-carboxysomes from Cyanobacteria, which are also inhibited by DTT (15, 16), and this redox sensitivity has been proposed to be beneficial during carboxysome assembly. Cells with CCMs have elevated concentrations of cytoplasmic HCO₃⁻ generated by active transport of DIC into the cell from the environment by membrane-spanning transporters (53). If carboxysomal CA were active in the cytoplasm during carboxysome assembly, it would convert cytoplasmic HCO₃⁻ into CO₂, which would then be lost from the cell via diffusion through the cell membrane (54). It has been suggested that carboxysomal CA enzymes that are inactive in reducing environments, such as the cytoplasm, will not cause CO₂ leakage from cells during carboxysome assembly (16). Presumably, these enzymes are oxidized and active once packaged within carboxysomes.

Preventing CO₂ leakage during carboxysome assembly would seem to provide a strong selective advantage for carboxysomal inactivation by reducing conditions, which makes the distribution of redox sensitivity among ιCA enzymes present in Thiomicrospira quite puzzling. If multiple cysteine residues at the amino terminus result in redox sensitivity, only three members of Thiomicrospira have redox-sensitive ιCA. Like T. aerophila, both T. cyclica and Thiomicrospira sp. ALE5 have two genes encoding ιCA, one of which (in each species) has two cysteine residues in its amino terminus and predicted redox sensitivity of structural disorder [via IUPred2A (52); https://iupred2a.elte.hu/] similar to ιCA from T. aerophila (Fig. 10). All three of these species also have a gene encoding ιCA without multiple amino terminal cysteine residues. The other four members of the genus only have ιCA with amino termini having single cysteine residues. Based on the observed redox sensitivities of ιCA from T. pelophila and T. aerophila, as well as inferences based on amino terminal sequences, only three members of Thiomicrospira have redox-sensitive ιCA, and all seven species carry genes encoding ιCA that is not sensitive to redox conditions. Perhaps the redox-insensitive version of this enzyme is an advantage for cells growing in highly reduced conditions, where hydrogen sulfide may be abundant enough to inactivate redox-sensitive ιCA (55), or where activating the enzyme by oxidation is less thermodynamically favorable. This may explain why T. microaerophila, which requires low O₂ concentrations for growth (56), only has the redox-insensitive form (Fig. 7). It would be helpful to determine whether the amino terminal cysteines are indeed responsible for redox sensitivity in these ιCA enzymes, and whether this redox sensitivity has a role in preventing CO₂ leakage during carboxysome assembly. It would also be helpful to understand how redox-insensitive carboxysomal CA enzymes are regulated during the assembly of these microcompartments.

ιCA could offer a selective advantage in alkaline habitats

Thus far, Thiomicrospira is the only genus in which CsoSCA appears to have been replaced by ιCA. The other genus in which ιCA is common in the carboxysome locus is Thioalkalivibrio, though for these organisms, a csoSCA gene is also present in the locus (Fig. 1 and Fig. 7). Both of these genera consist entirely (Thioalkalivibrio) or mostly (Thiomicrospira) of alkaliphilic organisms with optimal growth at pH 9–10 (57, 58). Most have been isolated from soda lakes with pH values from 9.2 to 11 (57). Given that the solubility of Zn⁺² and other metals is extremely low under alkaline conditions [e.g., the solubility product of Zn(OH)₂ is ~2 × 10⁻¹⁷ at 15°C–40°C (59)], acquiring sufficient Zn⁺² or other metals necessary for CA activity would be problematic. Some ιCA enzymes are atypical in that they do not require metals for activity (29), and this trait may be common among them (30). A lack of dependence on metals would provide a big advantage for carboxysome function in these habitats and may explain why CsoSCA, which requires Zn⁺² (39), was replaced by ιCA in Thiomicrospira. For members of Thioalkalivibrio, it would be interesting to determine which form of CA is packed into carboxysomes (CsoSCA or ιCA), and whether this depends on cultivation conditions (e.g., pH). It would be interesting to determine whether factors that favor ι- and βCA in α-carboxysomes are similar to those driving the distribution of β- and γCA in β-carboxysomes (12).

Given that autotrophic organisms grow in every imaginable habitat on Earth, it is not surprising that they have a diversity of biochemistries and physiological adaptations to facilitate CO₂ fixation at any pH, temperature, or solute composition amenable to life (5, 60). Carboxysome diversity is one such adaptation that could facilitate CO₂ fixation at a range of redox conditions or metal availabilities, and it is likely that more carboxysome diversity remains to be uncovered. Indeed, there are members of Nitrosospira (Phylum Pseudomonadota) and Pseudonocardia (Phylum Actinomycetota) whose carboxysome loci lack genes encoding any of the currently known forms of CA (20). An understanding of the full breadth of autotrophic adaptations to carbon fixation in all sorts of habitats will add to an understanding of life here and elsewhere. Furthermore, carboxysomes are being integrated into synthetic systems for generating compounds of industrial importance (61, 62); increasing the parameter space (e.g., pH) for the synthesis by native and engineered autotrophic microorganisms could be helpful for these efforts.

MATERIALS AND METHODS

Bacterial strains and cultivation conditions

T. pelophila DSM 1534^T was obtained from the German Collection of Microorganisms and Cell Cultures (DSMZ) and cultivated at 32°C under ambient air in thiosulfate supplemented artificial seawater medium [TASW, pH 7.5, 1 µg L⁻¹ vitamin B12 (53, 63)]. T. aerophila DSM 13739^T was obtained from DSMZ and cultivated in DSMZ medium 925 (alkaliphilic sulfur respiring strains medium) and incubated at 32°C under ambient air.

Five liter bioreactors were used for large-scale cultivation (Eppendorf, BioFlo 120). The pH was monitored and maintained via the addition of 10 M KOH (pH 7.5 for T. pelophila and pH 9.5 for T. aerophila). O₂ electrodes monitored the oxygen concentration, and cultures were stirred and sparged with O₂ to maintain O₂ levels at 95% air saturation. For mutant strain T. pelophila ιCA-135, which is unable to grow under ambient air, O₂ was maintained by sparging with 95% O₂, 5% CO₂ (vol/vol).

E. coli Lemo21(DE3) ΔyadF ΔcynT was used to identify genes encoding CA. With both genes encoding native CA (yadF/cynT) deleted, this strain cannot grow with air as the source of CO₂ (64). Heterologous expression of HCO₃⁻ transporters or CA enzymes restores its ability to grow under ambient air (64). E. coli Lemo21(DE3) ΔyadF ΔcynT was cultivated in lysogeny broth [LB (65)], supplemented with the appropriate antibiotics (see below), under either an air or 5% CO₂, 95% air (vol/vol) headspace (see below). Cultures were incubated at 37°C and 100 rpm.

Carboxysome purification

Volumes of 100–300 L of T. pelophila and T. aerophila culture were grown to early stationary phase in 5 or 10 L bioreactors (Eppendorf, BioFlo 120) and harvested via centrifugation (Sorvall SLA-1500 rotor, 3,795 × g, 10 min, 10°C–20°C). Carboxysomes were purified from the cells via differential centrifugation and use of sucrose gradients (66). In order to rupture carboxysomes for enzyme assays, pellets were frozen for 30 min at −20°C and stored at −80°C. Carboxysome preparations were monitored via SDS-PAGE (67). LC-MS/MS was used to determine the identities of the proteins, as in reference (64). Carboxysome preparations were also negatively stained with ammonium molybdate and examined with a FEI Morgani 268D transmission electron microscope (68).

CA assays

CA assays were conducted on purified carboxysomes and crude extracts from E. coli expressing genes from T. pelophila via stopped-flow spectrophotometry. To prepare for the assay, frozen carboxysomes were suspended in CA buffer (50 mM HEPES, 0.2 mM ZnSO₄, and pH 8 with NaOH) via vortexing and incubated at room temperature for 30 min. Frozen cell pellets from 500 mL cultures of E. coli Lemo21(DE3) ΔyadF ΔcynT heterologously expressing proteins from T. pelophila were thoroughly suspended in 2 mL CA buffer supplemented with lysozyme and bovine pancreas deoxyribonuclease I (Thermo Scientific, 0.1 mg each). Two milliliters of BPER-II Bacterial Protein Extraction Reagent (Thermo Scientific) was added to facilitate cell lysis, and samples were sonicated four times for 30 s while incubated in ice water. After sonication, samples were incubated at room temperature for 15 min while gyrated at 100 rpm (C10 platform shaker, New Brunswick Scientific).

Stopped-flow spectrophotometric CA assays were conducted similarly to references (39, 69), with an SX20 spectrophotometer (Applied Photophysics), using phenol red to track the fall in pH caused by CO₂ conversion to H₂CO₃ (and deprotonation to form HCO₃⁻ + H⁺). Absorbance values (A₅₅₇) taken 200 times per second, beginning at 0.1 s and ending at 0.2 s, were used to calculate the initial rate of CO₂ hydration since measurements prior to 0.1 s were erratic due to transients resulting from sample injection into the reaction chamber, and those after 0.2 s demonstrated a slow decrease in the rate of net CO₂ hydration, likely due to back-reaction (H₂CO₃ dehydration). A similar diminishment of apparent rates of CO₂ hydration by bovine carbonic anhydrase (Sigma-Aldrich, C3934) was observed in pilot studies to optimize the CA assay for the study presented here and is also apparent in published CA assays of carboxysomes from H. neapolitanus [as measured by pH indicator protonation (68)]. Absorbance values from these timepoints were stripped from the individual Excel files from each run and combined into a single file using a Python script (70) available at Github (https://github.com/scooterboi85/CSV-Conerter). Changes in A₅₅₇ were converted to changes in [H⁺] by using ΔA₅₅₇ values from titrating 1:1 mixtures of CA buffer and indicator solution with a standardized 6 N HCl solution (Thermo Fisher), which provided a measured change of 15.69 mM H⁺ per absorbance unit at 557 nm. Changes in [H⁺] were converted to changes in CO₂ concentrations by assuming 1 H⁺ produced per CO₂ hydrated (CO₂ + H₂O → H₂CO₃→H⁺ + HCO₃⁻). Each sample was run in triplicate, and the rates of CO₂ hydration were compared to those measured for identical samples that had been denatured by incubating them at 95°C for 30 min (also run in triplicate) to determine whether CO₂ hydration by the sample contents was primarily chemical or potentially catalyzed by CA. For those samples whose rate of CO₂ hydration significantly exceeded those measured in denatured samples (see Statistical analyses, below), CA activities were calculated by subtracting the rates from boiled samples from the rates of the unboiled samples.

Random mutagenesis of T. pelophila and characterization of mutant phenotypes

Random mutagenesis of T. pelophila was undertaken as in reference (71). In short, a clone library of 10,000 mutant strains of T. pelophila was generated by mating with E. coli BW20767 carrying plasmid pRL-27, which encodes transposon Tn5 (72). Once transformed into T. pelophila, Tn5 inserts randomly into the genome, generating a library of mutant strains, selectable by kanamycin since Tn5 encodes resistance to this antibiotic. The library was maintained on a solid medium containing 25 mg L⁻¹ kanamycin under a high CO₂ headspace [5% CO₂/95% air (vol/vol)]. This library was then screened for strains that had lost their ability to grow with air as the source of CO₂ (~0.04% CO₂) using high throughput 96-well growth assays (71). CO₂-sensitive strains were set aside for further characterization.

Characterization of mutant phenotype began by verifying CO₂ sensitivity by growing each of these strains individually in 20 mLTASW in paired cultures (air, 5% CO₂/95% air) and monitoring growth via optical density at 480 nm. Transposon insertion sites were identified using arbitrary PCR (71) to amplify the gDNA adjacent to Tn5 insertion for sequencing (Psomagen, Inc.). CO₂-sensitive mutants were also examined for carboxysome presence via transmission electron microscopy, as in reference (22).

Heterologous expression of genes of interest in E. coli

To clarify if genes interrupted in CO₂-sensitive T. pelophila mutant strains encode CA, these genes were cloned into the plasmid pBAD, for expression in E. coli Lemo21(DE3) ΔyadF ΔcynT under control of the araBp promotor. Plasmids carrying either csoS2 or csoSX were synthesized by the Department of Energy Joint Genome Institute. Plasmids carrying the gene encoding ιCA were synthesized by amplifying this gene from T. pelophila via PCR and inserting it into pBAD (Thermo-Fisher Scientific). Plasmids were initially propagated in E. coli Top10 in lysogeny medium supplemented with 100 mg L⁻¹ ampicillin and subsequently transformed into chemically competent E. coli Lemo21(DE3) ΔyadF ΔcynT (64), which were cultivated under a high CO₂ headspace on lysogeny media supplemented with antibiotics (30 mg L⁻¹ aprimycin and 30 mg L⁻¹ chloramphenicol for expressing CsoSX, CsoS2; 100 mg L⁻¹ ampicillin and 30 mg L⁻¹ chloramphenicol for expressing ιCA).

Inocula for the growth assay for CA activity were prepared by adding cells to 20 mL of LB supplemented with the appropriate antibiotics and cultivating overnight under 5% CO₂, 95% air (vol/vol). The following morning, inocula of 100 µL were introduced into 100 mL cultures containing 6 mM arabinose (induced), 11 mM glucose (repressed), or no additions (neutral) (73). Three replicates were run per condition. Additionally, E. coli Lemo21(DE3) ΔyadF ΔcynT, which was not carrying a plasmid (no plasmid control), was run alongside each experiment. Flasks were incubated with air headspace at 37°C, 100 rpm, and growth was monitored via spectrophotometry at OD₆₀₀ for 48 h.

Larger-scale cultures were used to prepare biomass for CA assays of E. coli (Lemo21(DE3) ΔyadF ΔcynT expressing ιCA from T. pelophila. Five hundred milliliters LB (supplemented with 100 mg L⁻¹ ampicillin and 30 mg L⁻¹ chloramphenicol) was grown overnight, at 37°C, 100 rpm, under 5% CO₂, 95% air (vol/vol) headspace. The next day, the culture was harvested via centrifugation (10 min, 3,795 × g, Sorvall SLA-1500, 10°C–20°C). The cell pellet was re-suspended in 2.5L of LB containing arabinose to induce expression. The 2.5 L culture was split into 0.5 L portions, which were incubated at 37°C, 100 rpm under ambient air, and cell growth (OD₆₀₀) was monitored. Five hundred milliliter portions were harvested via centrifugation after 0, 4, 6, 8, and 24 h. Cell pellets were stored at −80°C, until assayed for CA activity. The successful expression of T. pelophila genes by E. coli (Lemo21(DE3) ΔyadF ΔcynT was verified by subjecting proteins extracted from cell pellets to LC-MS/MS analysis as described above for carboxysome preparations.

Phylogenetic analyses

Homologs to the Thiomicrospira ιCA genes were collected from the Integrated Microbial Genomes database [IMG, https://img.jgi.doe.gov/ (41)] via BLASTp (74), with ιCA homologs from Thiomicrospira as query sequences. Sequences from reference (30) were gathered from the nonredundant protein sequences database (https://www.ncbi.nlm.nih.gov/). Amino acid sequences were aligned using MUSCLE (47), and phylogenetic trees were constructed using the maximum likelihood method (75). Alignment, model evaluation, and phylogenetic analysis were all implemented in MEGA 11 (48).

Statistical analyses

CO₂ hydration rates by carboxysomes were compared to rates measured for assay buffer, boiled carboxysomes, and carboxysomes in the presence of potential inhibitors or solvent (n = 3 for each condition). For carboxysome preparations from each strain (T. pelophila, T. aerophila, or T. pelophila ιCA-135), rates were subjected to ANOVA, with a post-hoc Tukey’s test for significance of difference among the treatments. CO₂ hydration rates by cell lysate from E. coli Lemo21(DE3) ΔyadF ΔcynT expressing genes from T. pelophila were analyzed by comparing boiled (n = 3) and unboiled (n = 3) rates via Welch t tests. All statistical analyses were implemented in R v4.3.2 using the R Stats Package (76).

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01075-24.

Supplemental material. aem.01075-24-s0001.pdf.

Tables S1 and S2; Figures S1 to S6.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.