Ribulose-1,5-Bisphosphate Carboxylase/Oxygenase (RubisCO) Is Essential for Growth of the Methanotroph Methylococcus capsulatus Strain Bath

ABSTRACT The ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO) enzyme found in plants, algae, and an array of autotrophic bacteria is also encoded by a subset of methanotrophs, but its role in these microbes has largely remained elusive. In this study, we showed that CO2 was requisite for RubisCO-encoding Methylococcus capsulatus strain Bath growth in a bioreactor with continuous influent and effluent gas flow. RNA sequencing identified active transcription of several carboxylating enzymes, including key enzymes of the Calvin and serine cycles, that could mediate CO2 assimilation during cultivation with both CH4 and CO2 as carbon sources. Marker exchange mutagenesis of M. capsulatus Bath genes encoding key enzymes of potential CO2-assimilating metabolic pathways indicated that a complete serine cycle is not required, whereas RubisCO is essential for growth of this bacterium. 13CO2 tracer analysis showed that CH4 and CO2 enter overlapping anaplerotic pathways and implicated RubisCO as the primary enzyme mediating CO2 assimilation in M. capsulatus Bath. Notably, we quantified the relative abundance of 3-phosphoglycerate and ribulose-1,5-bisphosphate 13C isotopes, which supported that RubisCO-produced 3-phosphoglycerate is primarily converted to ribulose-1-5-bisphosphate via the oxidative pentose phosphate pathway in M. capsulatus Bath. Collectively, our data establish that RubisCO and CO2 play essential roles in M. capsulatus Bath metabolism. This study expands the known capacity of methanotrophs to fix CO2 via RubisCO, which may play a more pivotal role in the Earth’s biogeochemical carbon cycling and greenhouse gas regulation than previously recognized. Further, M. capsulatus Bath and other CO2-assimilating methanotrophs represent excellent candidates for use in the bioconversion of biogas waste streams that consist of both CH4 and CO2. IMPORTANCE The importance of RubisCO and CO2 in M. capsulatus Bath metabolism is unclear. In this study, we demonstrated that both CO2 and RubisCO are essential for M. capsulatus Bath growth. 13CO2 tracing experiments supported that RubisCO mediates CO2 fixation and that a noncanonical Calvin cycle is active in this organism. Our study provides insights into the expanding knowledge of methanotroph metabolism and implicates dually CH4/CO2-utilizing bacteria as more important players in the biogeochemical carbon cycle than previously appreciated. In addition, M. capsulatus and other methanotrophs with CO2 assimilation capacity represent candidate organisms for the development of biotechnologies to mitigate the two most abundant greenhouse gases, CH4 and CO2.

M ethanotrophs are metabolically unique bacteria that are capable of utilizing CH 4 as a carbon and/or energy source (1). These microbes occupy an array of ecological niches across the globe and are vital in regulating atmospheric CH 4 by either preventing its release into the atmosphere or directly sequestering it from the air (2). Three metabolic modes of CH 4 assimilation have been described in phylogenetically diverse methanotrophic bacteria (3)(4)(5). These include (i) the ribulose monophosphate (RuMP) cycle, primarily utilized by gammaproteobacterial methanotrophs; (ii) the serine cycle, primarily utilized by alphaproteobacterial methanotrophs; and (iii) the Calvin-Basham-Benson (CBB) cycle, utilized by the verrucomicrobial and candidate phylum NC10 methanotrophs. The oxidation of CH 4 is essential for energy generation in all methanotrophs, but various CH 4 oxidation products, including formaldehyde, formate, and/or CO 2 , serve as carbon sources in these bacteria, depending on their single-carbon assimilation pathway.
Methylococcus capsulatus Bath is a gammaproteobacterial methanotroph that has served as a model methanotrophic bacterium; much of what has been learned about biological methane conversion, including the metabolic pathways and enzymes involved in CH 4 oxidation and assimilation, are based on seminal research utilizing this bacterium (6)(7)(8)(9)(10). Several independent laboratories have observed that M. capsulatus Bath exhibits inconsistent growth if exogenous CO 2 is not included in the gas phase (11,12). This phenotype can be circumvented in continuous culture if the gas flow rate is significantly decreased (13,14). Many hypotheses related to this phenotype have been proposed, including that supplying exogenous CO 2 supports the serine cycle in this organism under certain conditions, which is based on observations that CO 2 supplementation increases growth of alphaproteobacterial methanotrophs (15). Another hypothesis is that, in contrast to the majority of culturable alphaproteobacterial and gammaproteobacterial methanotrophs, M. capsulatus Bath requires exogenous CO 2 for growth. This hypothesis is supported by the presence of RubisCO and phosphoribulokinase in this organism (16,17), which would enable a complete CBB cycle commonly used by chemoautotrophic bacteria and phototrophs for CO 2 assimilation. Notably, CO 2 assimilation by M. capsulatus Bath has been observed in the presence of an energy source such as CH 4 or H 2 (11,18). Furthermore, M. capsulatus Bath autotrophic growth with H 2 as an energy source has been demonstrated on solid medium in a sealed growth vessel, but attempts to cultivate it autotrophically in liquid culture have failed (19). Thus, the roles of RubisCO and the CBB cycle in M. capsulatus Bath central metabolism and physiology remain elusive. In this study, we revisited the capacity of M. capsulatus Bath to utilize CO 2 and the potential role of RubisCO in CO 2 fixation and the central metabolism of this organism. Using reverse genetics approaches, we show that RubisCO-mediated CO 2 assimilation is essential for the growth of M. capsulatus Bath. Furthermore, 13 CO 2 isotopic tracing analyses indicated that a CBB cycle variant is active in this bacterium and highlight extensive overlap between CH 4 and CO 2 utilization pathways. These results establish that RubisCO and CO 2 are central to M. capsulatus Bath metabolism and provide insight into the CO 2 -dependent methanotrophy occurring in this bacterium.

RESULTS
CO 2 assimilation improves growth of M. capsulatus Bath. Initially, we tested the effect of CO 2 addition on M. capsulatus Bath growth with CH 4 as the primary carbon source. CO 2 supplementation to the gas phase of sealed serum vials slightly improved M. capsulatus Bath growth compared to that with CH 4 alone (Fig. 1a). Isotopic elemental analysis of M. capsulatus Bath cultured in serum vials with 8% 13 CO 2 in the gas phase mixture containing 20% CH 4 in air showed that a significant percentage (.10%) of M. capsulatus Bath biomass and excreted compounds in the culture medium were derived from exogenous CO 2 (Fig. 1b), which was 4 to 10Â higher than CO 2 assimilated via basal carboxylation reactions in Escherichia coli or in the related gammaproteobacterial methanotroph Methylotuvimicrobium alcaliphilum 20Z R , which does not encode RubisCO (see Fig. S1a in the supplemental material). Titrating 13 CO 2 in the gas phase indicated that maximum CO 2 assimilation by M. capsulatus Bath was limited below 8% CO 2 (vol/vol) (Fig. 1c), which is stoichiometrically consistent with the total biomass generated and the maximum percentage derived from CO 2 under these growth conditions. 13 CO 2 was assimilated immediately upon introduction to the culture (Fig. 1d), and a positive correlation between growth and 13 CO 2 incorporation into biomass was observed during active growth (Fig. 1d); thus, M. capsulatus Bath displays concurrent CH 4 and CO 2 metabolisms. We note the possibility that the percentage of total biomass derived from CO 2 is potentially even higher than that determined by elemental analysis, as unlabeled CH 4 -derived CO 2 evolved into the headspace is likely assimilated by the bacteria under these growth conditions. CO 2 is required for M. capsulatus Bath growth in an unsealed bioreactor with continuous gas supply. M. capsulatus Bath can be cultivated in sealed serum vials when CH 4 is the only carbon source supplied in the vial gas phase (Fig. 2a). However, we observed that M. capsulatus Bath did not grow in an unsealed bioreactor with con- capsulatus Bath in a sealed serum vial or a continuous gas reactor (CGR) with 20% CH 4 in air only (white bar) or supplemented with 2% CO 2 (gray bar). (b) M. capsulatus Bath growth in a CGR supplied with 20% CH 4 and 0.2% or 2% CO 2 in air at 1 volume gas mixture/volume medium/min (left). After 2 days of cultivation with 20% CH 4 in air, 2% CO 2 was added (1CO 2 , middle) or removed (2CO 2 , right) from the gas phase. Data represent the mean 6 standard deviation of 4 to 6 biological replicates from two independent experiments. ***, P # 0.001; ** P # 0.01, determined by unpaired Student's t test. tinuous CH 4 supply (20% CH 4 in air; 1 volume gas/volume vessel/min) unless CO 2 was also provided in the gas mixture ( Fig. 2a and b), a phenotype not observed for M. alcaliphilum 20Z R (Fig. S1b). This CO 2 -requiring growth phenotype was confirmed with the additional M. capsulatus strains M. capsulatus Bath (ATCC 33009) and M. capsulatus Texas (ATCC 19069) to rule out potential strain variation artifacts known to exist within methylotrophic bacterial laboratory strains (Fig. S1c) (20). Similar growth kinetics were measured in bioreactors during logarithmic growth between cultures supplied with 0.2% CO 2 and 2% CO 2 in 20% CH 4 and air; however, 0.2% CO 2 became limiting as the culture density increased, restricting the maximum culture density to ;50% of that observed with 2% CO 2 supplementation (Fig. 2b). We successfully cultivated M. capsulatus Bath without CO 2 if the gas flow rate was reduced (20% CH 4 in air; 0.1 volume gas/volume vessel/min), but the bacteria showed significantly slower growth kinetics and reduced culture density (optical density at 600 nm [OD 600 ] of ;4 after 7 days of cultivation; see Fig. S1d) compared to faster gas flow rates when CO 2 was requisite (Fig. 2b). Collectively, these data suggest that CH 4 -derived CO 2 is evolved and can support M. capsulatus Bath growth under some cultivation conditions (sealed serum vials and unsealed vessels with low CH 4 supply rates) but is stripped from an unsealed bioreactor with high flow rates. M. capsulatus Bath is widely used for the production of single-cell protein and is cultivated at industrial scale in proprietary U-loop bioreactors supplied with natural gas and oxygen/air mixtures (21)(22)(23). Notably, the U-loop bioreactor is sealed, such that CO 2 evolved from the gas fermentation is only vented to maintain gas partial pressures that ensure optimal gas solubility. Thus, M. capsulatus Bath likely uses CH 4 -derived CO 2 during U-loop reactor cultivation.
RubisCO is an essential M. capsulatus Bath gene. In M. capsulatus Bath genes encoding RubisCO, phosphoribulokinase, overlapping RuMP and CBB cycle enzymes, and serine cycle enzymes are transcribed at variable levels during active growth with continuous dual CH 4 and CO 2 supply ( Fig. 3a; see also Table S1 in the supplemental material). Relative transcript abundances of carboxylating enzymes that could mediate CO 2 assimilation by M. capsulatus Bath, including the serine cycle's carboxylating enzyme, pyruvate carboxylase (pyc), and the H-protein component of the putative carboxylating glycine cleavage enzyme, were 1.5-fold higher than the RubisCO large subunit transcript (Fig. 3b). Two other potential carboxylating enzymes, the pyruvate::ferredoxin oxidoreductase (pfo) and malic enzyme (sfc), exhibited significantly lower relative transcription compared to the other carboxylating enzyme transcripts (,50 transcripts per million [TPM]).
Pyruvate carboxylase (Dpyc), serine-glyoxylate aminotransferase (Dsga), or glycine cleavage (Dgcv) knockout strains obtained using marker exchange mutagenesis exhibited similar or retarded (Dsga) growth in serum vials compared to that of wild-type M. capsulatus Bath (Fig. 3c). The comparable growth of the Dpyc and Dgcv knockout strains to that of the wild type indicated that these enzymes are not essential for CO 2dependent growth of this organism. Interestingly, we observed a growth defect by the Dsga knockout strain, a similar phenotype to that recently observed in M. alcaliphilum 20Z R , which does not require exogenously supplied CO 2 for growth (24). Collectively, the Dsga and Dpyc knockout strain phenotypes suggest that the serine-glyoxylate aminotransferase is involved in M. capsulatus Bath oxaloacetate and/or glyoxylate conversion, but that a canonical serine cycle that would be dependent on a phosphoenolpyruvate or pyruvate carboxylase is not essential for M. capsulatus Bath CO 2 -dependent growth. We note the possibility that M. capsulatus Bath could encode an alternative enzyme to facilitate a complete serine cycle, although such an enzyme is not readily identified in the available genome.
We next hypothesized that CO 2 assimilation by this organism is mediated by RubisCO, which has been shown to be active during growth on CH 4 (11). To test this hypothesis, we attempted but were unable to obtain a RubisCO knockout strain (DcbbLS) using similar genetic approaches to those used in generating the Dpyc, Dgcv, and Dsga knockout strains, suggesting that RubisCO-encoding genes are essential. However, leveraging custom genetic tools for inducible gene expression (25), we developed a conditional RubisCO knockout strain with inducible, ectopic homologous RubisCO expression (pCBBLS DcbbLS). The pCBBLS DcbbLS strain exhibited similar growth to that of wild-type M. capsulatus Bath in both liquid and solid medium supplemented with the RubisCO inducer anhydrotetracycline (aTc). However, we observed no bacterial growth in liquid medium and significantly reduced growth on solid medium in the absence of RubisCO induction (Fig. 3d), supporting the conclusion that RubisCO is required for M. capsulatus Bath growth. Leaky cbbLS transcription could have enabled the limited growth we observed on solid medium, since we have previously measured leaky expression from the P tet promoter in the absence of aTc induction in M. capsulatus Bath (26). CH 4 and CO 2 enter overlapping M. capsulatus Bath central metabolic pathways. We next cultivated M. capsulatus Bath with 12 CH 4 and 13 CO 2 to identify enzymes mediating CO 2 fixation and metabolites derived from CO 2 . 13 C-labeling patterns of derivatized amino acids indicated that central metabolites derived from CO 2 enter core intermediary metabolic pathways, including the Embden-Meyerhof-Parnas (EMP) glycolytic pathway, the pentose phosphate pathway, and the tricarboxylic acid (TCA) cycle in this organism ( Fig. 4a; see also Fig. S2 and Table S2 in the supplemental material). 3-Phosphoglycerate (3PG)-derived glycine and serine incorporated 13 C exclusively in the C-1 position, which implicated RubisCO as the primary carboxylating enzyme responsible for CO 2 -derived 3PG production (Fig. 4b). In agreement with the genetic analyses, these amino acid isotopomer labeling patterns confirmed that RubisCO is the primary enzyme mediating CO 2 assimilation in M. capsulatus Bath and excluded carboxylation by pyruvate carboxylase, glycine synthase, malate synthase, pyruvate::ferredoxin oxidoreductase, or acetyl-coenzyme A (CoA) carboxylase as primary CO 2 assimilation reactions in this bacterium.
To confirm that 3PG is produced via RubisCO, as the serine and glycine isotopomer labeling patterns suggested, we quantified 3PG isotopomers directly. Consistent with RubisCO-dependent carboxylation of ribulose-1,5-bisphosphate (R1,5P), we measured the M1 3PG isotopomer after 4 h of labeling and found it to be 22% of the total 3PG pool (Fig. 5a). R1,5P isotopomers were also measured to determine if RubisCO-derived 3PG is used for regeneration of the RubisCO substrate (labeled R1,5P) or if RubisCOderived 3PG enters the EMP pathway and is completely converted to pyruvate in this bacterium (unlabeled R1,5P). The M1 R1,5P isotopomer was enriched (;45%), confirming that R1,5P is derived from both CH 4 and CO 2 and that a CBB cycle is active in M. capsulatus Bath (Fig. 5b). Surprisingly, we detected few of the doubly labeled M2 R1,5P isotopes (,5% of total isotopomers) that would be expected if ribulose-5-phosphate (Ru5P) is derived via carbon rearrangements occurring in the nonoxidative branch of the pentose phosphate pathway. The enriched M1 R1,5P isotopomer with limited detection of the M2 isotopomer indicates that R1,5P is regenerated via the oxidative  branch rather than via the nonoxidative branch of the pentose phosphate pathway, representing a noncanonical CBB cycle (Fig. 6). Collectively, isotopomer fingerprints of M. capsulatus central metabolites indicate that core intermediates are derived from both CH 4 and CO 2 carbon sources, supporting a high degree of metabolic plasticity and an essential interplay between CH 4 and CO 2 metabolism that engenders a novel, dual C 1 -fixing RuMP/RuBP pathway in this organism.

DISCUSSION
It has been 40 years since the initial discovery that M. capsulatus Bath possesses RubisCO activity and derives a portion of its biomass from CO 2 (11,18); however, the importance of CO 2 assimilation and the RubisCO enzyme in M. capsulatus Bath metabolism has remain unclear. In this study, we have provided evidence that both CO 2 and RubisCO are required for optimal M. capsulatus Bath metabolism and growth. Our results indicate that RubisCO assimilates CO 2 to produce 3PG and that a CBB cycle is active in this bacterium.
Our results expand upon the known carbon assimilation routes utilized by diverse methanotrophic bacteria. It is generally accepted that the majority of gammaproteobacterial methanotrophs utilize CH 4 as their sole carbon and energy source, whereas alphaproteobacterial methanotrophs assimilate CH 4 derivatives and CO 2 via the serine cycle (1,3). Recent evidence suggests that verrucomicrobial and candidate phylum NC10 methanotrophs utilize CH 4 as an energy source and CO 2 as a sole carbon source (27)(28)(29)(30)(31). Additionally, the verrucomicrobial Methylacidiphilum spp. have been demonstrated to grow autotrophically when using hydrogen as an energy source (30). Despite this mixotrophic metabolism, neither we nor others have successfully cultivated M. capsulatus Bath autotrophically in liquid medium. Our results show that M. capsulatus Bath is a chemoorganoautotroph that strongly prefers CH 4 as an energy source and requires both CH 4 and CO 2 as carbon sources. Further research is needed to understand why some methanotrophs encode RubisCO and assimilate CO 2 via a CBB cycle while others do not, but the metabolic plasticity afforded by the presence of hydrogenases, methane monooxygenases, and RubisCO may allow these microbes to inhabit and contribute to primary productivity in diverse habitats (2,32,33).
This study demonstrates that CO 2 is essential for M. capsulatus Bath growth. M. capsulatus Bath actively expresses several putative carboxylases that could enable CO 2 assimilation, but our data conclusively show that RubisCO is essential for M. capsulatus Bath under the experimental conditions utilized here. Other carboxylases that are part of the serine cycle and the recently discovered autotrophic reductive glycine pathway (34) that could mediate CO 2 fixation are also encoded by this organism. However, capsulatus Bath CBB cycle variant deduced from 13 CO 2 tracing and metabolite analyses. 3PG, 3phosphoglycerate; G3P, glyceraldehyde 3-phosphate; DHAP, dihydroxyacetone phosphate; F1,6P, fructose 1,6-bisphosphate; F6P, fructose 6-phosphate; E4P, erythrose 4-phosphate; S7P, sedoheptulose 7-phosphate; R5P, ribose 5-phosphate; Xu5P, xylulose 5-phosphate; Ru5P, ribulose 5-phosphate; R1,5P, ribulose 1,5-bisphosphate; G6P, glucose-6-phosphate; 6PG, 6-phosphogluconate. genetic analyses and 13 CO 2 isotopic tracing excluded these pathways as primary routes of CO 2 assimilation in this bacterium. In contrast, ;70% of the biomass of Methylosinus trichosporium OB3b, an alphaproteobacterial methanotroph that uses the serine cycle for CH 4 -derived carbon assimilation, is derived from CO 2 (15). These results underscore that the serine cycle can be a primary route of CO 2 assimilation in both methylotrophs and methanotrophs that use this pathway for methanol or CH 4 assimilation, respectively. These alternative CO 2 fixation pathways may play important metabolic roles for M. capsulatus in specific environments with variable CO 2 availability. Notably, carboxylases and the associated biochemical pathways present in M. capsulatus Bath represent rational metabolic engineering targets for increasing the CO 2 utilization capacity of this methanotroph for complete conversion of biogas or nonphotosynthetic CO 2 capture.
Several CBB cycle variants and RubisCO-dependent pathways function in nature (35,36). The isotopomer labeling patterns of 3PG and R1,5P (both singly 13 C-labeled) suggest that a CBB cycle variant that overlaps with the known primary CH 4 flux through the oxidative pentose phosphate pathway (oxPPP)/Entner-Doudoroff pathway is functional in M. capsulatus Bath (10). In the canonical transketolase-dependent CBB cycle, ribose-5-phosphate (R5P) and xylulose-5-phosphate (Xu5P) are produced from glyceraldehyde-3-phosphate and sedoheptulose-7-phosphate by transketolase during carbon rearrangement reactions of the nonoxidative pentose phosphate pathway (noPPP; Fig. 6a). M2 and M3 doubly and triply labeled R1,5P isotopomers would be expected if Ru5P were derived from the noPPP R5P and Xu5P metabolites due to carbon rearrangements, similarly to what we observed in the isotopomer distributions of the aromatic amino acids phenylalanine and tyrosine, which are derived from the noPPP metabolite erythrose-4-phosphate (Fig. 4a). However, R1,5P isotopomers primarily consisted of singly labeled M1 isotopes (Fig. 5b), which supports that the phosphoribulokinase substrate Ru5P is primarily derived from 6-phosphogluconate (6PG) formed in the oxPPP (Fig. 6b). This in an intriguing observation, given that decarboxylation of 6PG to Ru5P would represent a futile CBB cycle. It is possible that flux through oxPPP to Ru5P supports cellular reducing power, since two NADPH reducing equivalents are generated in this pathway. The reductive power generated by oxPPP is associated with boosting antioxidant defenses in many bacteria and eukaryotes (37)(38)(39)(40); thus, increased flux through this pathway could be essential for M. capsulatus Bath to maintain redox homeostasis and/or to support the increased NADPH required for CO 2 assimilation. Furthermore, the increased growth kinetics and overall biomass yield we observed (Fig. 1) support that RubisCO and CO 2 also serve a biosynthetic role, likely providing an alternative route of anaplerotic metabolites that enables metabolic plasticity for M. capsulatus Bath. Additional experimentation is required to validate the putative CBB cycle variant, determine its potential contribution to cellular energetics, and its relationship with CH 4 assimilation, since Ru5P also serves as the substrate for formaldehyde condensation in this methanotroph.
M. capsulatus Bath is an obligate aerobe, so its RubisCO presumably exhibits both carboxylase and oxygenase activities during growth in the presence of CO 2 and O 2 , generating phosphoglycerate and phosphoglycolate from these substrates, respectively. Indeed, phosphoglycolate phosphatase activity in M. capsulatus Bath cell-free extracts and bacterial phosphoglycolate utilization were previously demonstrated (18). We observed significant transcription of a putative phosphoglycolate phosphatase gene (MCA_RS12655) that likely encodes the enzyme mediating conversion and utilization of RubisCO-generated phosphoglycolate during active growth with dual CH 4 /CO 2 supply (see Table S1 in the supplemental material). M. capsulatus Bath also encodes a putative glycolate oxidase (MCA_RS07375 to MCA_RS07380) that catalyzes the conversion of glycolate to glyoxylate. Glyoxylate entry into the serine cycle would complete a unique metabolic pathway in this methanotroph, a serine cycle variant of the phosphoglycolate salvage pathway recently described in the chemolithotroph Cupriavidus necator (41). In support of this hypothesis, the growth defect observed by the Dsga mutant strain that cannot convert glyoxylate to glycine could be due to a deficiency in the capacity to metabolize phosphoglycolate. Thus, the production of phosphoglycolate and its downstream metabolism may contribute, in part, to the essentiality of RubisCO that is independent of the observed M. capsulatus Bath requirement for CO 2 .
The essentiality of M. capsulatus Bath RubisCO could also be partially attributed to a role in controlling intracellular R1,5P levels. Accumulation of R1,5P significantly inhibits the bacterial growth of E. coli expressing heterologous phosphoribulokinase (42). Furthermore, RubisCO has been shown to play important roles in the regulation of both redox and R1,5P levels in purple nonsulfur bacteria (43,44). Notably, recent evidence indicates that one-carbon assimilation in the methylotroph Methylorubrum extorquens is regulated by an R1,5P-responsive allosteric transcriptional activator, QscR, that induces expression of serine cycle enzymes (45). M. capsulatus Bath encodes a QscR homolog (MCA_RS14905) with 41% amino acid identity to the M. extorquens protein. If R1,5P plays a similar regulatory role or causes toxicity in M. capsulatus Bath, the accumulation of R1,5P in the absence of RubisCO could preclude isolation of a RubisCO-null mutant. Thus, additional inquiry into the role of R1,5P in M. capsulatus Bath physiology is warranted.
The M. capsulatus Bath dual CH 4 /CO 2 metabolism described here provides additional insight into one-carbon metabolism and the potential evolutionary relationship between methanotrophy and RubisCO-mediated autotrophy. The extensive overlap in metabolites produced and converting enzymes required for sole fixation of each carbon source in related methanotrophic and autotrophic bacteria supports this evolutionary relationship. Notably, this dual CH 4 /CO 2 metabolism designates M. capsulatus Bath as a promising biocatalyst for simultaneous mitigation and valorization of the two most abundant and harmful atmospheric greenhouse gases (46,47).

MATERIALS AND METHODS
Methanotroph cultivation. Bacterial strains used in this study are shown in Table 1. The primary Methylococcus capsulatus Bath strain used during the course of these investigations was obtained from the Mary Lidstrom laboratory at the University of Washington. Additional M. capsulatus strains were  4 in the gas phase, as previously described (26). Methylotuvimicrobium alcaliphilum 20Z R was cultured in modified NMS medium containing 3% NaCl and carbonate buffer as described previously (48). Strains were grown in 150-ml vials containing 30 ml of growth medium. After inoculation with plate-derived biomass, vials were crimped with gray butyl stoppers to create gas-tight seals. CH 4 was added to the headspace to reach a final CH 4 of 20% in air (vol/ vol), and cultures were incubated at 37°C (M. capsulatus) or 30°C (M. alcaliphilum) at 200 rpm orbital shaking. Continuous gas cultivation was performed using a custom midthroughput gas fermentation reactor (MGFR) in 150-ml Kimax cultivation tubes fitted with stainless steel sparge stones. Culture aliquots (100 ml) were inoculated with plate-derived biomass and supplied with 20% CH 4 in air (vol/vol) or 20% CH 4 /0.2 to 2% CO 2 in air (vol/vol) at a flow rate of 1 volume gas/volume culture/min (vvm) premixed with gas-specific mass flow controllers. For the CH 4 only cultivation of M. alcaliphilum, carbonate buffer was removed from the medium formulation, after which the pH of the medium was stabilized at 9.5 with KOH. Knockout strain construction. Primers used in this study are shown in Table 2. The M. capsulatus Bath cbbLS (MCA_RS13440 to MCA_RS13445) genes encoding the RubisCO large and small subunits, respectively, were amplified using oCAH644 and oCAH655 primers and cloned into pCAH01 via Gibson assembly to generate pRUB with inducible RubisCO expression under the control of the tetracycline promoter/operator. pRUB was transferred to M. capsulatus Bath via biparental conjugation using E. coli S17 as previously described (26). Genomic fragments (1,000 bp) flanking the cbbLS, pyruvate carboxylase (pyc [MCA_RS12165 to MCA_RS12170]), glycine cleavage aminomethyltransferase and glycine dehydrogenase components (gcv, [MCA_RS01715 to MCA_RS01725]), or serineglyoxylate aminotransferase (sga [MCA_RS06920]) genes, an FLP recombination target (FRT)flanked gentamicin resistance cassette from pPS856 (49), and a pCM184 (50) or pCM433kanT (51) fragment containing an origin of transfer (oriT) were amplified by PCR and assembled independently via Gibson assembly. The resulting marker exchange mutagenesis plasmids for generating pyc, gcv, or sga knockout strains were introduced into wild-type M. capsulatus Bath via biparental conjugation, and transformants were selected on solid NMS medium containing 30 mg/ml gentamicin. The cbbLS marker exchange mutagenesis plasmid was introduced to the pCbbLS-containing strain by biparental conjugation in the presence of the anhydrotetracycline (aTc) inducer. After mating, gentamicin-resistant clones were selected and maintained on NMS medium supplemented with aTc (0.5 mg/ml), gentamicin (30 mg/ml), and kanamycin (50 mg/ml). Positive transformants were PCR genotyped for the absence of wild-type pyc, gcv, or sga loci using the primers oCAH1233/ oCAH1234, oCAH151/oCAH152, or oCAH1215/oCAH1216, respectively. The absence of the cbbLS genes and the presence of pCbbLS were confirmed via PCR using the primers oCAH964/oCAH965 and oCAH172/oCAH173, respectively, followed by sequence verification. Conditional growth of the pCbbLS DcbbL::Gm r strain was performed on NMS solid medium with or without 0.5 mg/ml aTc (RubisCO induction) supplementation.
Isotopic elemental analysis. Cultures for isotopic and elemental analysis were grown in sealed serum vials as described above with 20% CH 4 in air (vol/vol). 13 CO 2 (Sigma) was added to the headspace via syringe to reach the indicated final concentrations, and cultures were incubated at 37°C with 200 rpm orbital shaking for 48 to 72 h. At indicated time points postinoculation, culture density was measured spectrophotometrically using a NanoDrop spectrophotometer (Thermo). Bacterial cells were then pelleted via centrifugation and freeze-dried prior to analysis. Isotopic and elemental analyses were conducted by combustion in a Flash 2000 elemental analyzer (Thermo) coupled with an isotope ratio mass spectrometer (Delta V; Thermo Scientific). Standards were used to account for linearity, drift, and isotopic discrimination.
RNA sequencing. M. capsulatus Bath transcription was evaluated during logarithmic growth in a continuous gas reactor as described above with 20% CH 4 /2% CO 2 in air. Culture samples (2 ml; OD 600 = 1) were pelleted by centrifugation, resuspended in 500 ml RNALater (Ambion), and stored at 220°C. Frozen samples were shipped to Genewiz (South Plainfield, NJ) for RNA isolation, library preparation, and sequencing. RNA was submitted to quality control analysis before rRNA depletion. Sequencing was performed using the Illumina HiSeq platform. Paired-end 150-bp Illumina reads were analyzed by Genewiz using their standard transcriptome sequencing (RNAseq) analysis pipeline. Reads were mapped to the M. capsulatus Bath reference genome (GenBank accession no. NC_002977.6), and transcript abundance was supplied as relative transcripts per million (TPM). 13 C tracer analyses. Sample preparation and gas chromatography-mass spectrometry (GC-MS) for amino acid isotopomer analysis were performed as previously reported (52). Liquid chromatography-tandem mass spectrometry (LC-MS/MS) for determination of 3-phosphoglycerate and ribulose-1,5-bisphosphate isotopomers was conducted as previously reported (53). For all isotopic tracer analyses, 30-ml cultures in 150 ml serum vials were inoculated at an OD 600 of 0.1 and incubated with 20% CH 4 /8% 13 CO 2 in air (vol/vol). Samples for proteinogenic amino acid derivatization were prepared from 5 ml of logarithmically growing cultures at 24 h postinoculation. For determination of phosphorylated compounds, 13 CO 2 (8% final) was added via a syringe to cultures growing exponentially on 20% CH 4 in air (OD 600 = 0.4) and incubated at 37°C with 200 rpm orbital shaking for 4 h. After incubation, cultures were rapidly harvested by vacuum filtration using a Büchner funnel with a mixed cellulose ester (MCE) membrane filter (0.2 mM, 47 mm). Cells were washed with 5 ml ice-cold NMS medium before the filter was immediately transferred using forceps to a 15-ml centrifuge tube and flash frozen in liquid nitrogen. Frozen cells/membrane were stored at 280°C until metabolite extraction.
Data availability. Plasmids, bacterial strains, and other data will be made available upon reasonable request.