Expression and characterization of Pantoea CO dehydrogenase to utilize CO-containing industrial waste gas for expanding the versatility of CO dehydrogenase

Although aerobic CO dehydrogenases (CODHs) might be applicable in various fields, their practical applications have been hampered by low activity and no heterologous expression. We, for the first time, could functionally express recombinant PsCODH in E. coli and obtained a highly concentrated recombinant enzyme using an easy and convenient method. Its electron acceptor spectra, optimum conditions (pH 6.5 and 30 °C), and kinetic parameters (kcat of 12.97 s−1, Km of 0.065 mM, and specific activity of 0.86 Umg−1) were examined. Blast furnace gas (BFG) containing 20% CO, which is a waste gas from the steel-making process, was tested as a substrate for PsCODH. Even with BFG, the recombinant PsCODH retained 88.2% and 108.4% activity compared with those of pure CO and 20% CO, respectively. The results provide not only a promising strategy to utilize CO-containing industrial waste gases as cheap, abundant, and renewable resources but also significant information for further studies about cascade reactions producing value-added chemicals via CO2 as an intermediate produced by a CODH-based CO-utilization system, which would ultimately expand the versatility of CODH.

enzymatic CO-utilization for higher specificity and productivity. Given that (i) an aerobic CODH is able to catalyze CO in the industrial waste gases to CO 2 and (ii) CO 2 is often used as a building block for producing useful chemicals (e.g., urea, methanol, formate, and acetone) through chemical and biological routes with environmental and economic benefits 11 , constructing a CODH-based CO-utilization system via CO 2 as an intermediate was our challenge. Moreover, aerobic CODHs could be practically used for treating hazardous CO in daily life. For example, aerobic CODHs immobilized on the filter of a cigarette seem to effectively convert hazardous CO generated during smoking into non-toxic CO 2 .
Herein, we, for the first time, reported the heterologous expression of an aerobic CODH in E. coli for producing a highly concentrated recombinant enzyme using an easy and convenient process for practical applications. In particular, anaerobic CODH obtained from Pantoea species YR343 (PsCODH) was newly discovered through genome-mining and phylogenetic analysis. To characterize the recombinant PsCODH, electron acceptor spectra, optimum conditions, and kinetic parameters were determined. Furthermore, the feasibility of PsCODH was evaluated for constructing the CO-utilization system with a CO-containing waste gas from the steel-making process as cheap, abundant, and sustainable feedstock.

Results
A newly discovered aerobic CODH from the Pantoea species YR343 (PsCODH) was overexpressed in E. coli. Although aerobic CODHs are significant for developing enzymatic CO-utilization systems, the heterologous expression of an aerobic CODH has not been reported up to date. In order to discover aerobic CODHs that could be expressed in E. coli, we explored proteins from diverse microorganisms listed in the UniProt Knowledgebase (UniProtKB, http://www.uniprot.org/uniprot/), which is the most widely used database for functional information on proteins and includes accurate and detailed annotations. To search for potential CODHs from UniProtKB, we used the aerobic CODH from Oligotropha carboxidovorans (OcCODH) as a reference, since OcCODH has been one of the most intensively researched aerobic CODHs and has structural information is available (PDB ID: 1N5W). OcCODH is a dimer of heterotrimers consisting of CoxS with two Fe 2 S 2 clusters, CoxM with non-covalently bound FAD, and CoxL containing a molybdopterin active site. Therefore, several keywords including aerobic CODH, aerobic coxS, aerobic coxM, and aerobic coxL were used, and several hundreds of hits were resulted (data not shown).
Then, we assumed that microorganisms phylogenetically close to E. coli would result in good functional expression of their genes in E. coli. In particular, we focused on microorganisms between E. coli and Bacillus species in the phylogenetic trees. The Bacillus species are obligate aerobes or facultative anaerobes, and their proteins also usually express well in E. coli [12][13][14][15] . Thus, CODHs from species between E. coli and Bacillus were investigated from the CODHs found in our UniProtKB search. As a result, a CODH from the Pantoea species YR343 (PsCODH) was chosen because it was (i) closer to the CODH from E. coli, a commonly used host strain for overexpression, in the phylogenetic tree (Fig. 1); (ii) it was classified as an Enterobacteriaceae similar to E. coli; and (iii) it was phylogenetically located on the border between aerobic and anaerobic CODHs (Fig. 1), thereby expecting that PsCODH may partially share features of anaerobic CODH that are heterologously expressed in E. coli 16,17 .
For the overexpression of PsCODH in E. coli, genes encoding the L, M, and S chains of PsCODH were cloned into the pQE80L vector. A hexahistidine-tag was added to the C-terminus of the L chain for purification as shown in Fig. 2(A). After overexpression and purification using Ni-NTA resin, the homogeneity of the recombinant PsCODH was investigated by SDS-PAGE as shown in Fig. 2(B). The molecular weights of the L, M, and S chains of the recombinant PsCODH were estimated to be 78.0, 33.7, and 19.6 kDa, respectively. To calculate the molecular mass of the gross recombinant PsCODH, size exclusion chromatography (SEC) was performed. As shown in Fig. 2(C) and Fig. S1, the molecular weight of the gross recombinant PsCODH was estimated to be 127.4 kDa. Additionally, heterologous expression enabled us to obtain highly concentrated recombinant CODH with a yield of up to 4.8 mg from a 300 mL culture.
Furthermore, in vivo and in vitro FAD reconstitution were performed to obtain functional recombinant PsCODH, and the reconstitution was evaluated by determining the ratio of A 450 /A 280 18 . As a result, the A 450 /A 280 of the recombinant PsCODH without reconstitution was 0.71 ± 0.11, whereas in vivo reconstitution increased the A 450 /A 280 to 1.81 ± 0.50. For in vitro reconstitution, the A 450 /A 280 was about 0.63 ± 0.23. Consequently, in vivo reconstitution was the most effective for FAD-reconstitution and thus additional experiments were performed with the in vivo reconstituted recombinant PsCODH in this study. Biochemical characterization and kinetic analysis. For validating the functionality of the recombinant PsCODH, the extent of CO-driven and DT-driven PsCODH reduction were compared 19 . As shown in Fig. 3, CO-purging for 2 min reduced the absorbance at 450 nm (ε FAD, 450 nm = 11.3 mM −1 cm −1 ) by approximately 71.8 % compared to the DT-reduced PsCODH, which demonstrated that 71.8 % of the recombinant PsCODH was functional and was able to catalyze CO to CO 2 . Further incubation of the recombinant PsCODH with CO for 30 min resulted in complete reduction, which was similar to the reduction of the DT-reduced form. This result indicated  that a certain amount of non-functional PsCODH was slowly reduced, probably due to intermolecular electron transfer from the reduced functional PsCODH 19 .
We already evaluated the functionality of the recombinant PsCODH by monitoring the reduced FAD ( Fig. 3), but the CO-oxidizing activity of CODH was typically assayed with artificial electron acceptors and often depended on the type of electron acceptor used 20 . Accordingly, the electron acceptor spectrum of the recombinant PsCODH was examined. As shown in Table 1, the recombinant PsCODH used methylene blue (MB), phenazine methosulfate -nitroblue tetrazolium chloride (PMS-NBT), thionine acetate (TA), benzyl viologen (BV), and methyl viologen (MV) as electron acceptors to varying extents. Even if MV was the most widely used electron acceptor for assaying anaerobic CODH, it was less effective for aerobic recombinant PsCODH and exhibited only 22.0 % and 8.6 % activity relative to MB and PMS-NBT, respectively. PMS-NBT and TA have not been used for assaying anaerobic CODH, but these compounds were effective for aerobic recombinant PsCODH, probably due to the different redox center in the active site. The aerobic CODH usually contains a [Mo-Fe-S] cluster, whereas anaerobic CODH contains an [Ni-Fe-S] cluster in its activecenter 17,21 . Among the tested electron acceptors, PMS-NBT was the most effective for accepting electrons from recombinant PsCODH.
Before determining the kinetic parameters, the optimum pH and temperature for CO-oxidizing activity were investigated. Figure 4 describes that the optimum pH and temperature were 6.5 (200 mM, potassium phosphate buffer) and 30 °C, respectively. The optimum parameters for CO-oxidation were not far from the growth conditions for the Pantoea species, i.e., typically 25-30 °C and pH 7 22,23 .
As shown in Fig. 5, the initial TA reduction velocity depended on the CO concentration under optimum conditions. The initial velocity exhibited a rectangular-hyperbolic increase, and the data was fitted with a Michaelis-Menten equation with a high regression coefficient (R 2 = 0.9972), thereby indicating a k cat of 9.31 s −1 and a K m of 0.065 mM.
PsCODH utilized CO-containing waste gas from the steelmaking process. In order to evaluate the feasibility of recombinant PsCODH for constructing an enzymatic CO-conversion system with industrial waste gas as the feedstock, the specific activity was compared with pure CO, 20 % CO balanced with N 2 , and BFG containing 22 % CO and 20 % CO 2 as the substrate. Although we aimed to compare all waste gases in Table S1, COG and LDG were associated with a risk of explosion and thus were not suitable for testing in the laboratory despite the fact that LDG contained the highest amount of CO, up to 65 %, which was a promising feedstock for enzymatic CO utilization. Thus, only BFG was evaluated in this study. As a result, BFG retained 88.2 ± 7.4 % activity relative to that of pure CO (Fig. 6). In addition, BFG maintained 108.4 ± 7.6 % of relative activity compared to that of 20 % CO. The results implied that (i) H 2 and N 2 in BFG did not negatively influence CODH activity and (ii) the recombinant PsCODH was not inhibited by CO 2 as a product from PsCODH catalysis. Consequently, the recombinant PsCODH was found to be a promising candidate for developing an enzymatic CO-conversion system using industrial waste gases as a cheap and sustainable feedstock.

Discussion
Although aerobic CODHs are essential for developing enzymatic CO-utilization systems for producing useful chemicals and fuels from CO, the heterologous expression of an aerobic CODH has not been reported up to date. To the best of our knowledge, this is the first report about the heterologous functional expression of an aerobic CODH in E. coli. Through genome-mining and phylogenetic analysis, we newly identified an aerobic CODH from the Pantoea species YR343 (PsCODH) and characterized the electron acceptor spectra, optimum pH and temperature, kinetic parameters, and specific activity of the PsCODH. SDS-PAGE and SEC in Fig. 2 represented that PsCODH may exist as a heterotrimer (LMS) that is distinct from OcCODH, which exists as an (LMS) 2 hexamer 24,25 and is used as the reference CODH for genome-mining in this study. In the near future, we are planning to study the 3-dimensional structure of PsCODH to provide structural information regarding aerobic CODHs that can be expressed in E. coli and to provide insight into the catalytic mechanisms of the electron transfer pathway. As shown in Table 1, PsCODH exhibited broad electron acceptor spectra to varying extents. Comparing CODHs from Carboxydothermus hydrogenoformans 17 and O.carboxydothermus 21 , the metal cluster in the redox center might affect the preference of the electron acceptor and thus the structural validation of PsCODH could explain not only the electron acceptor preference but also the role of the metal cluster in the catalytic mechanism. The kinetic parameters of PsCODH were compared to the parameters of aerobic CODHs that were previously reported ( Table 2). CODHs from Bradyrhizobium japonicum 24 and O.carboxydothermus 19 , which are capable of chemolithoautotrophic growth using CO as a sole carbon and energy source, oxidized CO to CO 2 to varying extents with various electron acceptors 19,24 . Even though it is unclear whether P. species YR 343 utilizes CO as the sole carbon and energy source, the recombinant PsCODH exhibited a specific activity of 0.86 Umg −1 . The purified anaerobic CO-oxidizing:H 2 -evolving enzyme complex from C. hydrogenoformans was slightly inhibited by the substrate CO 9 , whereas aerobic recombinant PsCODH did not exhibit substrate inhibition at all. Given that the recombinant PsCODH used in this assay was approximately 71.8 % active (Fig. 3), the corrected k cat value for the fully functional PsCODH seemed to be maximized up to 12.97 s −1 with TA as the electron acceptor 19 . Various industrial processes discharge considerable amounts of CO into the atmosphere. For example, POSCO, which is the most competitive steel-maker in the world, emits huge amounts of CO-containing waste gases with different compositions (Table S1). Currently, most waste gases are combusted as fuels at POSCO, but the inert gases in the waste gas are impediments to fuel efficiency. Accordingly, it would be helpful to establish methods for using waste gases for other purposes. Therefore, we focused on evaluating whether waste gases may be cheap and sustainable resources for CO-utilization for producing value-added chemicals. Figure 6 implies that (i) H 2 and N 2 in BFG did not negatively influence CODH activity and (ii) the recombinant PsCODH was not inhibited by CO 2 that was a product from PsCODH catalysis. Consequently, the recombinant PsCODH was found to be a promising candidate for developing an enzymatic CO-conversion system using industrial waste gases as cheap, abundant, and sustainable feedstock for producing value-added chemicals. Since CODH is the key enzyme for CO-metabolism in CO-utilizing microorganisms, research on CODH has been focused on the catalytic mechanisms and structural information. However, practical applications of CODH to produce value-added chemicals have not been thoroughly explored to date. The results described in this study might contribute to expanding the versatility of CODH towards practical uses even under aerobic conditions. For example, aerobic CODH immobilized on a cigarette filter could convert toxic CO generated from smoking to non-toxic CO 2 . Furthermore, the results might open a new door to cascade reactions consisting of CO conversion to CO 2 and CO 2 conversion to useful fuels and chemicals via CO 2 as an intermediate. Accordingly, we plan on performing cascade reactions that produce formate, which is the most beneficial bulk chemical obtained from CO 2 from an environmental and economic perspective 11 , by utilizing BFG as a feedstock in the near future.   MA, USA) and Qiagen (Valencia, CA, USA), respectively. Pfu polymerase, EcoRI, HindIII, and DpnI were purchased from Enzymonics (Daejeon, Korea). As a CO-containing waste gas from the steelmaking process, BFG was prepared from Seongkwang (Seoul, Korea). Unless otherwise stated, all chemicals were used at the highest grade available without further purification.

Cloning and overexpression of PsCODH in E. coli.
Genes encoding the large (coxL; GenBank accession: EJM96790), medium (coxM; GenBank accession: EJM96789), and small (coxS; GenBank accession: EJM96790) chains of PsCODH were synthesized by GenScript (Piscataway, NJ, USA). The gene sequences were codon-optimized for expression in E. coli and were then cloned into the pQE80 vector between the EcoRI and HindIII sites, which generated pQE80-Ps_L, pQE80-Ps_M, and pQE80-Ps_S, including coxL, coxM, and coxS, respectively. For the construction of pQE80-Ps_SML containing the large, medium, and small chains of PsCODH, we performed restriction-free (RF) cloning 26,27 . To amplify the medium and small chains, a polymerase chain reaction (PCR) was performed using pQE80-Ps_M and pQE80-Ps_S as templates with the primers listed in Table S2, respectively. To construct pQE80-Ps_ML containing the large and medium chains of PsCODH, RF cloning was performed with the PCR products from pQE80-Ps_M and pQE80-Ps_L as a primer and a template, respectively. To construct pQE80-Ps_SML containing the large, medium, and small chains of PsCODH, RF cloning was performed with the PCR products from pQE80-Ps_S and pQE80-Ps_ML as a primer and a template, respectively. The RF cloning product was purified, incubated with DpnI at 37 °C for 2 hours, and then transformed into competent E. coli Top10 (DE3) cells generating Top10_pQE80-Ps_SML-expressing cells. The mixture of transformed cells was spread on a Luria-Bertani (LB) agar plate containing 100 μ g mL −1 ampicillin. The gene sequences of PsCODH obtained from several colonies were verified by Macrogen (Seoul, Korea). For overexpression, Top10_pQE80-Ps_SML cells were grown in LB medium (300 mL in 1 L flask) containing 50 μ g mL −1 ampicillin and 1 mM sodium molybdate at 37 °C and 200 rpm. When the optical density of cells at 600 nm reached approximately 1.0, 1.0 mM isopropyl-β -d-thiogalactopyranoside (IPTG) and 100 μ M FAD were added for induction and in vivo reconstitution, respectively. Then, the temperature was lowered to 25 °C. After a 24-hour cultivation, the cells were harvested by centrifugation (12,000 rpm, 30 min, 4 °C), and the cell pellet was stored at − 70 °C before use 18,28 . In vitro reconstitution and purification of recombinant PsCODH. For the cell lysis, the cell pellet was suspended in BugBuster ® (1 g cell pellet in 5 mL BugBuster ® ). After 1 hour of incubation on ice with shaking at 20 rpm, the cell debris were removed by centrifugation (12,000 rpm, 30 min, 4 °C). Then, 100 μ M FAD was added to the crude cell lysate for in vitro reconstitution, and the mixture was incubated on ice for 1 hour with shaking at 20 rpm. Then, the lysate was loaded onto Ni-NTA resin, the column was washed with buffer containing 20 mM of imidazole, and finally 3 mL of the purified recombinant PsCODH was obtained. The purified recombinant PsCODH was quantified by Bradford assay using bovine serum albumin as the standard 18 .
The molecular masses of the L, M, and S chains of PsCODH were identified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) using a 12 % acrylamide gel under denaturing conditions. All protein bands were stained with Coomassie Blue (Bio-Rad, Hercules, CA, USA) for visualization 18 . The molecular mass of the gross recombinant PsCODH was estimated by SEC using a Superdex 200 Increase 10/300 GL column (GE Healthcare Life Sciences, United Kingdom). The purified recombinant CODH was loaded onto the column and eluted at a flow rate of 0.2 mL min −1 with phosphate buffer (50 mM, pH 7.2) containing 150 mM NaCl. Thyroglobulin (669 kDa), ferritin (440 kDa), aldolase (158 kDa), and conalbumin (75 kDa) were used as reference proteins. The molecular mass of the gross recombinant PsCODH was calculated by comparing the migration time with those of the reference proteins 29 .
The optimum pH and temperature were determined using 20 μ M TA as the electron acceptor. For the optimum pH, 200 mM phosphate buffer (pH 6.0-pH 7.5) was used at 30 °C. Then, the optimum temperature was investigated at the optimum pH shown in Fig. 3(A).
For calculating kinetic parameters, 0, 10, 20, 50, and 99.985 % of CO balanced using N 2 were used. By Henry's law: 0, 10, 20, 50, and 99.985 % CO in the headspace corresponded to 0, 0.091, 0.182, 0.455, and 0.909 mM CO in the aqueous phase at 30 °C, respectively. To measure the initial velocity, each concentration of CO was purged into the headspace of a rubber septum-stopped cuvette containing 20 μ M TA in phosphate buffer (200 mM, pH 6.5) for 2 min. For saturating CO into the solution, the cuvette was incubated at 30 °C for 1 hour. Then, the recombinant PsCODH was injected using a syringe to initiate the reaction. The absorbance change at 595 nm was monitored at 30 °C using a UV-visible spectrophotometer. Kinetic parameters (k cat and K m ) were calculated from the Michaelis-Menten equation. To calculate the relative activity with CO-containing waste gas from the steelmaking process, the initial velocity with BFG was measured using the same procedure as described above.