Engineered microbial biofuel production and recovery under supercritical carbon dioxide

Abstract

Culture contamination, end-product toxicity, and energy efficient product recovery are long-standing bioprocess challenges. To solve these problems, we propose a high-pressure fermentation strategy, coupled with in situ extraction using the abundant and renewable solvent supercritical carbon dioxide (scCO₂), which is also known for its broad microbial lethality. Towards this goal, we report the domestication and engineering of a scCO₂-tolerant strain of Bacillus megaterium, previously isolated from formation waters from the McElmo Dome CO₂ field, to produce branched alcohols that have potential use as biofuels. After establishing induced-expression under scCO₂, isobutanol production from 2-ketoisovalerate is observed with greater than 40% yield with co-produced isopentanol. Finally, we present a process model to compare the energy required for our process to other in situ extraction methods, such as gas stripping, finding scCO₂ extraction to be potentially competitive, if not superior.

Fig. 1

Schematic of integrated fermentation and extraction under supercritical CO₂. A scCO₂-tolerant microbe can be engineered to produce compounds, such as medium-chain alcohols (C4-C5), that may serve as biofuels and may be preferentially extracted from aqueous media into the scCO₂ phase. Collection of scCO₂ followed by partial de-pressurization will facilitate high-purity biofuel extraction. The presented biphasic scCO₂ separation strategy is expected to simultaneously provide a contaminant-free environment for the engineered organism due to the broad microbial lethality of scCO₂, and continuously strip microbially produced biofuels to eliminate end-product toxicity. scCO₂ is only weakly soluble in the aqueous media phase ( < 1%) and vice versa, resulting in a biphasic system

Fig. 2

Heterologous protein expression and promoter evaluation in SR7. a Schematic of the promoters tested in SR7. pXyl is a xylose-inducible promoter and contains the xylR gene and xylA promoter from B. megaterium strain DSM 319. pHysp contains the IPTG-inducible hyperspank promoter and lacI, both from plasmid pDR111. The p43 promoter is taken from B. subtilis strain KS438 and is growth-associated. Expected inducers are indicated in parentheses. b Bulk fluorescence measurements of SR7 populations containing gfp-encoding plasmids, with (circles, solid lines) or without (diamonds, dashed lines) inducers. Error bars represent the standard deviation of biological triplicate samples. c Fluorescence populations of SR7 cells expressing gfp at 4 h post induction measured by FACS (Alexa Fluor 488). A GFP positive gate was established using the autofluorescence of SR7 cells containing an empty plasmid control. d Phase-contrast and fluorescence microscopy of SR7 pXyl gfp cells with and without xylose induction. Images were taken 4 h post induction. Source data are provided as a Source Data file

Fig. 3

β-Galactosidase activity for SR7 cultures grown under 0.1 MPa CO₂ and 10 MPa scCO₂ at 37 °C. Specific β-galactosidase activity measured in lysates of SR7 cultures that contain an empty pXyl plasmid or pXyl lacZ. Activity was normalized to total protein concentration in the lysate to control for growth differences and differential cell lysis. Errors represent standard deviation of biological replicate cultures (n > 4). For cultures grown under 0.1 MPa CO₂, overnight anaerobic cultures grown from spores were subcultured into CO₂-evacuated vials to a uniform optical density. Induction with xylose occurred 2 h after subculture, and the time point for LacZ activity measurement was taken 4 h post induction. For cultures grown under 10 MPa scCO₂, SR7 pXyl lacZ spores were loaded into stainless steel culture vessels with and without xylose inducer. Cultures were grown for 21 days and microscopy was used to determine which samples showed at least a tenfold increase in cell number prior to measuring LacZ activity. Source data are provided as a Source Data file

Fig. 4

Isobutanol and isopentanol production in SR7. a Two-step enzyme pathway to convert the valine intermediate 2-ketoisovalerate (αKIV) to isobutanol using the decarboxylase enzyme KivD from L. lactis and an alcohol dehydrogenase. b Evaluation of isobutanol production at 24 h post induction in SR7 with an empty pXyl plasmid, pXyl kivD, pXyl kivD ADH6_Sc, or pXyl kivD yqhD_Ec, where ADH6 and YqhD are alcohol dehydrogenases from S. cerevisiae and E. coli, respectively. Cultures were grown aerobically in LB medium in the presence and absence of 5 g l^–1 xylose as an inducer and 5 mM αKIV as substrate for the pathway. For all aerobic experiments, error bars represent the standard deviation of biological triplicate cultures. c Co-production of isopentanol measured in SR7 samples shown in b. d Accumulation of the isobutyraldehyde intermediate at short culture times, 4 h post induction. Alcohol dehydrogenases ADH6 from S. cerevisiae, AdhA from L. lactis, AdhP from E. coli, YqhD from E. coli, and AdhA from B. megaterium SR7 were evaluated to decrease the buildup of isobutyraldehyde intermediate for aerobic cultures fed 5 mM αKIV and induced with xylose. e Production of isobutanol at 24 h post induction for the cultures shown in d. f Production of isobutanol for SR7 cultures grown under 0.1 MPa CO₂ (biological triplicate cultures) and 10 MPa scCO₂ at 37 °C. Under scCO₂, cultures with at least a tenfold increase in cell number, as enumerated by microscopy, were analyzed for alcohol production, sugar consumption and fermentation product generation. Average isobutanol titers from the aqueous phase of cultures showing at least 1 mM glucose consumption (classified as high activity; Supplementary Fig. 11) are provided. Source data are provided as a Source Data file

Fig. 5

Isobutanol energy recovery requirements for integrated fermentation-extraction under scCO₂. a Simplified process schematic of the envisioned scCO₂ bioprocess used to develop an Aspen Plus model, including compression of CO₂ to scCO₂, fermentation/extraction, de-pressurization/biofuel recovery, and CO₂ recycle. Further detail is provided in Supplementary Fig. 15a. b Energy requirement for isobutanol recovery as MJ kg⁻¹ for several scenarios. Separate fermentation and scCO₂ recovery processes represent the base case (scenario A). Scenario B is coupled fermentation and extraction without use of partial extraction stream de-pressurization or scCO₂ recycle. Scenario C de-pressurizes the extraction stream to 6.5 MPa to induce formation of an isobutanol-rich stream, which permits recycling of compressed CO₂, set at 50% (by mass). Scenario D again de-pressurizes to 6.5 MPa and recycles the maximum possible amount of CO₂ (87%) that results in no accumulation of CO₂ in the system. Values for b can be found in Supplementary Table 6. c Comparison of the energy requirements for isobutanol production found for the scCO₂ process relative to published literature values for alternative in situ recovery methods. The scCO₂ extraction energy requirement in c is for coupled fermentation and extraction with partial de-pressurization and full recycle (i.e., Scenario D in b). Source data are provided as a Source Data file