Incorporation of Functional Rubisco Activases into Engineered Carboxysomes to Enhance Carbon Fixation

The carboxysome is a versatile paradigm of prokaryotic organelles and is a proteinaceous self-assembling microcompartment that plays essential roles in carbon fixation in all cyanobacteria and some chemoautotrophs. The carboxysome encapsulates the central CO2-fixing enzyme, ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco), using a polyhedral protein shell that is selectively permeable to specific metabolites in favor of Rubisco carboxylation. There is tremendous interest in repurposing carboxysomes to boost carbon fixation in heterologous organisms. Here, we develop the design and engineering of α-carboxysomes by coexpressing the Rubisco activase components CbbQ and CbbO with α-carboxysomes in Escherichia coli. Our results show that CbbQ and CbbO could assemble into the reconstituted α-carboxysome as intrinsic components. Incorporation of both CbbQ and CbbO within the carboxysome promotes activation of Rubisco and enhances the CO2-fixation activities of recombinant carboxysomes. We also show that the structural composition of these carboxysomes could be modified in different expression systems, representing the plasticity of the carboxysome architecture. In translational terms, our study informs strategies for engineering and modulating carboxysomes in diverse biotechnological applications.


■ INTRODUCTION
Cells exploit the physical and chemical nature of molecules to generate self-assembling supramolecular complexes, membrane domains, and organelles, which provides a means for segregating specific functions into different subcellular regions to modulate metabolic reactions in space and in time. 1,2 While the emergence of compartmentalization and confinement in the cell is widely accepted as a key event in the evolution of eukaryotic cells, more recent work has documented that compartmentalization is also ubiquitous in prokaryotes. A versatile paradigm is the bacterial microcompartment (BMC) that encapsulates diverse metabolic enzymes within the nanoscale compartments using a polyhedral protein shell. 3−9 BMCs are widespread in the bacterial phyla and are of paramount importance for CO 2 fixation, pathogenesis, and microbial ecology. 10−12 Carboxysomes are the canonical BMCs found in all cyanobacteria and some chemoautotrophs. Carboxysomes encapsulate the key CO 2 -fixing enzymes ribulose-1,5-bisphosphate carboxylase oxygenase (Rubisco) and carbonic anhydrase (CA), using a protein shell made of numerous protein paralogs (Figure 1a). 8,13,14 Rubisco is the central enzyme in the Calvin−Benson−Bassham cycle of photosynthesis, mediating CO 2 fixation by catalyzing the carboxylation of its substrate ribulose-1,5-bisphosphate (RuBP). Although Rubisco is highly productive on a global scale, collectively fixing about 10 11 tons of carbon annually, 15 this enzyme is somewhat inefficient given its distinct substrate specificity for both CO 2 and O 2 and relatively slow catalytic rate. These features make the catalytical reaction of Rubisco the limiting step in photosynthetic CO 2 fixation. 16 To overcome this, in the carboxysome-containing organisms, Rubisco is encased by a protein shell that is selectively permeable to HCO 3 − , permitting substantial accumulation of HCO 3 − within the organelle. 17 The coencapsulated CA then dehydrates HCO 3 − to CO 2 and supplies a high concentration of CO 2 around Rubisco. 18,19 The exquisite carboxysome architecture and the semipermeability of the protein shell ensure enhanced CO 2 assimilation capacity of carboxysomes that are estimated to contribute to approximately 25% of global carbon fixation. 8 Introducing functional carboxysomes into heterologous organ-isms via synthetic biology approaches has proven to be a promising strategy to supercharge CO 2 fixation and enhance agricultural productivity. 20−26 Based on the types of the enclosed Rubisco, carboxysomes can be categorized into α-carboxysomes that contain Form 1A Rubisco and β-carboxysomes that encase plantlike Form 1B Rubisco. 8,27 Rubisco of the two forms is a hexadecameric complex composed of eight large subunits and eight small subunits, denoted as CbbL 8 S 8 in α-carboxysomes or RbcL 8 S 8 in β-carboxysomes. The biogenesis of Rubisco requires a series of chaperones, such as GroELS, 28 Rubisco assembly factor 1 (Raf1), 29−31 and RbcX 32,33 for Form 1B Rubisco. Rubisco also requires conformational repair by Rubisco activases (Rca) to be catalytically active. To fulfill the functionality, the active site of Rubisco must be carbamylated by nonsubstrate CO 2 molecules. However, binding of RuBP prior to carbamylation or other misfire sugar bisphosphates, such as xylulose-1,5bisphosphate, 2,3-pentodiulose-1,5-bisphosphate, and 2-carboxy-D-arabinitol-1-phosphate, can inhibit Rubisco by closing the catalytic site and impeding reactions with either CO 2 or O 2 . 34 Rca is required to remove these inhibitors from Rubisco to restore its carboxylation activity, 35 through binding with Rubisco over one of the catalytic sites of red-type Rubisco 36 or the RbcL N-terminus of Form 1B Rubisco. 37 In the chemoautotroph Halothiobacillus neapolitanus (H. neapolitanus), Rca comprises a prokaryotic AAA+ protein CbbQ (∼30 kDa) and a Rubisco adaptor CbbO (∼82 to 88 kDa) (Figure 1b). 38,39 CbbQ appears as a hexameric ring of the typical AAA + -ATPase domain and was indicated to be associated with the α-carboxysome by interacting with the shell protein CsoS1. 39 CbbO has a C-terminal VWA domain with a metal ion-dependent adhesion site, which is vital for interacting with Rubisco. 35,38 Both cbbQ and cbbO genes are often present concurrently downstream of the Rubisco genes in the carboxysome-encoding operons. 40 It has been shown that one CbbQ hexamer can bind one CbbO monomer in vitro to form a bipartite complex, and the binding of CbbO was presumed to be key for the Rca activity. 38 While evidence indicates that CbbQ is associated with the H. neapolitanus carboxysome shell, 39 how the CbbQO complex promotes activation of Rubisco in α-carboxysomes remains enigmatic.
Here, we develop genetic constructs to coexpress the CbbQO Rca complex with the H. neapolitanus α-carboxysomes in E. coli, and characterize the incorporation of CbbQO within the recombinant carboxysomes and their roles in promoting CO 2 fixation of the carboxysomes. Our study provides insight into the significance of Rca in mediating the structure and functionality of α-carboxysomes. It has implications for synthetically engineering carboxysome structures with the capacity of modulating their composition and functionality.

■ RESULTS AND DISCUSSION
Integration of CbbO and CbbQ in the α-Carboxysome. Previous studies have shown that expressing the H. neapolitanus α-carboxysome cso operon could lead to the formation of catalytically functional α-carboxysome structures in E. coli 20,21,26,41 and a Gram-positive bacterium. 22 To coexpress CbbO and CbbQ with recombinant α-carboxysomes and investigate their functions in carboxysome activities, we generated a series of constructs using a pAM2991 vector (Figure 1c). The pS1D plasmid consists of the α-carboxysome cso operon from H. neapolitanus, including the genes encoding Rubisco large and small subunit proteins (CbbL and CbbS), the shell proteins CsoS1A/B/C and CsoS4A/B, the shellassociated protein CsoS2, the CA protein CsoSCA, and the csoS1D gene. The pS1DQ, pS1DO, and pS1DQO plasmids integrate the cbbQ, cbbO, and cbbQO genes, respectively, into the α-carboxysome expression operon, downstream of csoS1D neapolitanus and the synthetic operons for producing α-carboxysome structures in E. coli. pS1D, pS1DQ, pS1DO, and pS1DQO were generated using a pAM2991 vector. His-tags are fused to the 3′' end of cbbQ and cbbO genes. (d) PCR (polymerase chain reaction) verification of the carboxysome-expressing vectors using the primers shown in (c) ( Table 3). The sizes of the PCR products were 1.1, 3.5, 1.9, and 4.4 kb for pS1D, pS1DO, pS1DQ, and pS1DQO, respectively.
( Figure 1c,d). Polyhistidine tags were fused to the C-termini of CbbQ and CbbO for immunoblot assays.
After isopropyl β-D-1-thiogalactopyranoside (IPTG) induction to ensure the expression of α-carboxysome proteins, the recombinant α-carboxysomes were purified by sucrose gradient centrifugation. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and immunoblot analysis confirmed the presence of the carboxysome protein components in the carboxysome preparations from pS1D, pS1DQ, pS1DO, and pS1DQO cells (Figure 2a,b), consistent with previous results. 41 In addition, we verified the presence of CbbQ in the pS1DQ and pS1DQO carboxysomes and the incorporation of CbbO into the pS1DO or pS1DQO carboxysomes using an anti-His antibody (Figure 2a,b), demonstrating that CbbQ and CbbO can be structurally integrated into recombinant αcarboxysomes as intrinsic components. Consistently, CbbQ has been identified in the H. neapolitanus α-carboxysomes. 39 Quantitative analysis of immunoblots indicated that the ratio of CbbQ and CbbO in the pS1DQO carboxysomes is ∼6:1 (data were calculated from immunoblot results in Figure 2b),   Table 2. Error bars represent SD of at least three independent replicates. supporting the functional forms of CbbQ as a hexamer and CbbO as a monomer. 38 Our results suggest that the expressed CbbQ or CbbO alone can be integrated into recombinant αcarboxysomes (Figure 2a,b). In support of our observation, CbbQ was proposed to integrate into the carboxysome via interacting with the shell protein. 39 It has been suggested that the Rca could be packed into β-carboxysomes and binds with Rubisco via its Rubisco small subunit-like domains and AAA+ core. 37 As indicated by SDS-PAGE and immunoblot analysis, the Rubisco contents were similar among the four samples, suggesting that the integration of CbbO and CbbQ did not affect the Rubisco content (Figure 2a,b).
Negative-staining electron microscopy (EM) showed that the recombinant carboxysomes produced in the pS1D, pS1DO, pS1DQ, and pS1DQO constructs exhibited a polyhedral shape with defined edges and vertices (Figure 2c). The average diameters of the recombinant α-carboxysomes are 122 ± 19 nm for pS1D (mean ± SD, n = 43), 129 ± 14 nm for pS1DQ (n = 39), 124 ± 17 nm for pS1DO (n = 35), and 128 ± 18 nm for pS1DQO (n = 50) (Figure 2d). No significant difference in diameter was observed among the four types of recombinant carboxysomes, suggesting that integration of CbbO and CbbQ has no notable effects on the carboxysome structure. The sizes were comparable with those of the native carboxysome purified from H. neapolitanus 42 and the cyanobacterium Synechococcus WH8102, 43 as well as recombinant H. neapolitanus carboxysomes 20 and empty α-carboxysome shells produced in E. coli. 41 The Activase Activity of CbbO and CbbQ within the α-Carboxysome. While a functional CO 2 -concentrating mechanism (CCM) pathway has been reconstructed in E. coli, 26 building on evidence that multiple proteins including CbbQO are required for CCM function, no study has yet examined the roles of CbbQO in isolated carboxysomes. Observing the successful incorporation of potentially functional activase proteins in recombinant carboxysomes, we examined the activase activity of CbbQO in these structures at different concentrations of carboxyarabinitol-1,5-bisphosphate (CABP), which is a tight-binding inhibitor of Rubisco. 44 As expected, 0.1 μM CABP could inhibit up to 95% of Rubisco activity (Figure 3a; Table 1); Rubisco activity appeared to be linear in the absence of CABP, and there is no significant difference in the Rubisco activity between recombinant carboxysome types under these conditions. In contrast, remarkable differences were observed when assaying Rubisco activities at 0.05 μM CABP. The Rubisco activity of the pS1DQO carboxysomes was higher than that of pS1DQ (∼1.2 fold) and pS1DO (∼1.4 fold), and the pS1D carboxysomes that lack CbbQ and CbbO had the lowest Rubisco activity among these recombinant carboxysomes (Figure 3a, Table 1). Meanwhile, supplementing isolated carboxysomes with ATP could diminish Rubisco inhibition by CABP and enhance Rubisco activity (Figure 3b), consistent with the ATP requirement for CbbQ. 38 Taken together, our results indicate that integration of both CbbQ and CbbO could improve the Rubisco carboxylation activities of recombinant carboxysomes, confirming their roles as Rca in dissociating the tightly bound CABP from the inhibited Rubisco holoenzymes and thereby enhancing the carboxylation of Rubisco. 38 CbbQO has also been suggested to function as the Rca in both Form I and Form II Rubisco. 38,45 To further evaluate the functions of CbbQ and CbbO in Rubisco activities of recombinant α-carboxysomes, we carried out 14 C radiometric Rubisco assays as a function of the RuBP concentration (normalized by the total protein abundance) and then calculated V max and K m for RuBP using a Michaelis− Menten kinetic model. The pS1DQO carboxysomes possessed a higher V max than the pS1D, pS1DQ, and pS1DO carboxysomes, indicating that the overall carbon-fixation activity of carboxysomes was stimulated in the presence of CbbQO (Figure 3c; Table 2). Moreover, immunoblot analysis indicated the equal quantities of Rubisco in these recombinant carboxysomes (Figure 2a,b), suggesting that the Rubisco functionality per active site was enhanced in the pS1DQO carboxysomes. K m (RuBP) of these recombinant carboxysomes was relatively similar ( Table 2), suggesting that the CbbQO hetero-oligomer may specifically release tight-binding inhibitory sugar phosphates during Rubisco activation. Since incorporation of CbbQO could mediate activation of inhibited Rubisco and improve the CO 2 -fixation activities of carboxysomes (Figure 3), coexpressing the catalytically active CbbQO Rca and carboxysomes could be an effective approach to stimulate carboxysome function in heterologous hosts. 26 Variability of the α-Carboxysome Architecture. To elucidate whether different expression systems can affect the formation and structure of carboxysomes, we also generated the carboxysome-expression vectors using a pBAD33 vector that is induced by arabinose. The carboxysome structure with a polyhedral shape could be formed by expressing the created pBAD33-S1D, pBAD33-S1DQ, pBAD33-S1DO, and pBAD33-S1DQO vectors (Figure 4a). No visible difference in size was observed between the recombinant carboxysomes expressed using different vectors. SDS-PAGE and immunoblot analysis showed the typical distribution pattern of α-carboxysome proteins (Figure 4b). However, the protein levels of CbbQ and CbbO were significantly lower in the purified recombinant carboxysomes expressed from the pBAD33-based vectors (Figure 4b) than those from the pAM2991 vector ( Figure  2b). SDS-PAGE profile analysis further confirmed that the protein content of some components within the carboxysome structures differs among the carboxysomes generated by pAM2991 and pBAD33 vectors (Figure 4c). For example, the pBAD33-S1DQO carboxysome contains a relatively high content of Rubisco and CsoS2B in comparison with the pS1DQO carboxysomes, specifying the stoichiometric and organizational variations of the α-carboxysome architecture. Stoichiometric plasticity has been recently assessed as a general feature of natural and recombinant BMCs, including βcarboxysomes, 13,14,46 the propanediol utilization metabolosome, 47 and several recombinant shell structures. 46,48,49 This structural variation may have important implications on the flexible protein−protein interactions and the modulation of shell permeability for the regulation of BMC assembly and function in response to a varying environment. It also implies the requirement of tuning expression of carboxysome operons for functionality. 26

■ CONCLUSIONS
In this study, we experimentally verify that the Rca proteins CbbQ and CbbO could serve as structural components of reconstituted α-carboxysomes, without detectable effects on the carboxysome structure. Incorporation of both CbbQ and CbbO into the recombinant carboxysomes could promote catalytic activation of inhibited Rubisco in the presence of 0.05 μM CABP and enhance the CO 2 -fixation activities of recombinant carboxysomes in the presence of ATP. Moreover, we show that the assembly and organizational composition of recombinant carboxysomes could be modified by using different expression systems, highlighting the plasticity of the carboxysome architecture, which may be physiologically vital for carboxysome self-assembly, repair, and permeability regulation. Our study may offer new strategies for rational design, engineering, and modulation of carboxysome structure Homologous sequences for Gibson assembly, restriction enzyme sites, and His-tag coding sequences are shown in bold, italic, and underlined, respectively. and function in synthetic biology, emphasizing the requirement for carboxysomal Rca for correct functions.

■ METHODS
Construction of Expressing Vectors. The genetic organization of the operons that express α-carboxysomes is displayed in Figure 1c. For pS1D, the operon was amplified from the pHnCBS1D plasmid (Addgene plasmid # 52065) 20 and then cloned into a modified pAM2991 vector containing a Kanamycin resistance gene by Gibson Assembly (NEB, UK). The cbbQ and cbbO genes were cloned from the genomic DNA (deoxyribonucleic acid) of H. neapolitanus, and a His-tag coding sequence was appended to the 3′-termini of cbbQ and cbbO by PCR. The fragments of cbbQ and cbbO were digested by BsaI and then assembled by T 4 DNA ligase using Golden Gate Assembly 50 to generate the cbbQO expression cassette. Finally, cbbQ, cbbO, and cbbQO were cloned into the pS1D vector at the NotI site to generate pS1DQ, pS1DO, and pS1DQO, respectively. To generate the pBAD33-S1D, pBAD33-S1DQ, pBAD33-S1DO, and pBAD33-S1DQO vectors, the operons in pS1D, pS1DQ, pS1DO, and pS1DQO were cloned into the amplicon of a pBAD33 vector 51 by Gibson Assembly. The positive clones were verified by PCR, and the plasmids were finally confirmed by sequencing. All the primer information used in this research is listed in Table 3.
Protein Expression and Carboxysome Purification. The E. coli BL21(DE3)/TOP10 constructs were cultured overnight at 37°C in 10 mL of LB medium with the corresponding antibiotic, and the cultures were diluted in 800 mL of medium in a 2-L flask. When the optical density (OD) of the culture reaches 0.6, arabinose or IPTG was added to a final concentration of 1 mM or 50 μM to induce protein expression. The cultures were grown at 25°C overnight with a 120-rpm shaking.
Cells were harvested by centrifugation at 5000g for 10 min and washed with 10 mL of TEMB buffer (10 mM Tris-pH 8.0, 10 mM MgCl 2 , 1 mM EDTA (ethylenediamine tetraacetic acid), and 20 mM NaHCO 3 ). The cells were then resuspended in 20 mL of TEMB buffer with the 10% CelLytic B cell lysis reagent (Sigma-Aldrich, USA) and 1% Protease Inhibitor Cocktail (Melford, UK). The cells were broken by sonication and then centrifuged at 10,000g to remove cell debris at 4°C. The supernatant was recentrifuged at 50,000g for 30 min at 4°C to enrich carboxysomes. The pellet was resuspended with 2 mL of TEMB buffer and centrifuged at 10,000g for 1 min, before loading the supernatant onto a 10, 20, 30, 40, and 50% (w/v) sucrose density gradient. Sucrose gradients were subjected to centrifugation at 80,000g for 30 min at 4°C. Carboxysomes were enriched in the 40% sucrose fraction and were collected for further analysis.
Rubisco Activity and Activase Assays. Rubisco activity assays were performed as previously described. 13 Approximately 200 ng μL −1 isolated α-carboxysomes (5 μL) in Rubisco assay buffer (100 mM EPPS (4-(2-hydroxyethyl)-1piperazinepropanesulphonic acid), pH 8.0, 20 mM MgCl 2 , 3.5 mM ATP) were aliquoted into scintillation vials containing NaH 14 CO 3 (1.48−2.22 GBq mmol −1 ) at a final concentration of 25 mM and incubated at 30°C for 2 min. D-Ribulose-1,5bisphosphate sodium salt hydrate (RuBP; Sigma-Aldrich) was then added to the samples with a range of concentrations (0− 0.8 mM) to initiate carbon fixation. The reaction was terminated after 5 min incubation by adding 10% (v/v) formic acid. The samples were then dried on heat blocks at 95°C to remove unfixed NaH 14 CO 3 , and the pellets were resuspended in distilled water in the presence of the scintillation cocktail (Ultima Gold XR; Perkin-Elmer, USA). Radioactivity measurements were carried out using a scintillation counter (Tri-Carb; Perkin-Elmer, USA). Counts per minute were used to calculate the amount of fixed 14 C according to the standard curve and were then converted to the total CO 2 fixation rates. V max was calculated using a Michaelis−Menten kinetic model in Origin Pro 2020b (OriginLab, USA). For each experiment, at least three independently purified carboxysome samples were examined. Results are presented as mean ± SD.
Electron Microscopy. The structures of purified recombinant α-carboxysomes were characterized using negativestaining transmission electron microscopy as described previously. 13,23,31 The sizes of the recombinant carboxysomes were analyzed by ImageJ.