n-Butanol production by Rhodopseudomonas palustris TIE-1

Bacterial strains, media, and growth conditions

All strains used in this study are listed in Supplementary Table 8E. coli strains were grown in lysogeny broth (LB; pH 7.0) at 37 °C. For aerobic growth, Rhodopseudomonas palustris TIE-1 was grown at 30 °C in YP medium (3 g/L yeast extract, 3 g/L peptone) supplemented with 10 mM MOPS [3-N (morpholino) propanesulphonic acid] (pH 7.0) and 10 mM succinate (YPSMOPS) under the illumination of an infrared LED (880 nm). For growth on a solid medium, YPSMOPS or LB was supplemented with 15 g/L agar. For anaerobic phototrophic growth, TIE-1 was grown in anoxic bicarbonate buffered freshwater (FW) medium²⁷. All FW media was prepared under a flow of 34.5 kPa N₂ + CO₂ (80%, 20%) and dispensed into sterile anaerobic Balch tubes. The cultures were incubated at 30 °C in an environmental chamber fitted with an infrared LED (880 nm). For photoheterotrophic growth, the FW medium was supplemented with 50 mM MOPS at pH 7.0 and sodium 3-hydroxybutyrate or sodium acetate at pH 7.0, to a final concentration of 50 mM. For photoautotrophic growth on iron, anoxic sterile stocks of FeCl₂ and nitrilotriacetic acid (NTA) were added to reach final concentrations of 5 mM and 10 mM, respectively. For photoautotrophic growth on H₂, TIE-1 was grown in FW medium at pH 7.0 and 12 psi of 80% H₂/20% CO₂²⁷. For all carbon and electron sources, either ammonium chloride (5.61 mM) or dinitrogen gas (8 psi) was supplied as nitrogen source²⁷. All sample manipulations were performed inside an anaerobic chamber with a mixed gas environment of 5% H₂/75% N₂/20% CO₂ (Coy Laboratory, Grass Lake, MI). When needed, 400 μg/mL kanamycin was added for TIE-1, and 50 μg/mL kanamycin was added for E. coli.

R. palustris TIE-1 deletion mutant construction

We constructed three mutants, two of which were double mutants using the method described in a previous study³⁰. Respectively, Glycogen synthase knockout was created by deleting Rpal_0386 (glgA), nitrogenase knockout was created by deleting Rpal_5113 (nifA1), and Rpal_1624 (nifA2), and hydroxybutyrate polymerase knockout was created by deleting and Rpal_2780 (phaC1) and Rpal_4722 (phaC2) were deleted resulting in hydroxybutyrate polymerase knockout. Briefly, the 1 kb upstream and 1 kb downstream regions of the gene were PCR amplified from the R. palustris TIE-1 genome, then the two homology arms of the same gene were cloned into pJQ200KS plasmid. The resulting vector was then electroporated into E. coli and then conjugated to R. palustris TIE-1, using the mating strain E. coli S17-1/λ. After two sequential homologous recombination events, mutants were screened by PCR, as shown in Supplementary Fig. 9. The primers used for mutant construction and verification are listed in Supplementary Tables 9 and 10.

Plasmid construction

All plasmids used in this study are listed in Supplementary Table 11. There are five genes involved in the n-butanol biosynthesis: phaJ, ter, adhE2, phaA, and phaB (Fig. 1a). Among these five genes, TIE-1 has homologs of the first two (phaA and phaB). Hence, we designed two different cassettes, namely, a 3-gene cassette (3-gene), which has phaJ, ter, adhE2, and a 5-gene cassette (5-gene), which has the 3-gene plus a copy of the phaA-phaB operon from TIE-1. phaJ, ter, and adhE2 sequences were obtained from published studies³⁴. The phaJ gene, isolated from Aeromonas caviae, was chosen because it codes for an enzyme that has a higher specificity for its substrate^34,70. The ter gene isolated from Euflena gracilis was selected because it is unable to catalyze the reverse oxidation of butyryl-CoA³⁴. The adhE2 gene isolated from C. acetobutylicum is chosen because the enzyme encodes for specifically catalyzes the reduction of the butyryl-CoA³⁴. All three foreign genes (phaJ, ter, and adhE2) were codon-optimized by Integrated DNA Technology (IDT) for TIE-1. The cassette was synthesized as G-blocks by IDT, which we then stitched together by overlap extension and restriction cloning. The phaJ-ter-adhE2 cassette was then inserted into plasmid pRhokS-2, resulting in pAB675. PhaA and phaB were amplified as an operon from the R. palustris TIE-1 genome. The phaA-phaB cassette was then cloned into pAB675 to obtain pAB744. Upon obtaining mutants and plasmids, either the 3-gene or the 5-gene was conjugated into WT TIE-1 or the mutants, using mating the strain E. coli S17-1/ λ. All conjugations were successful, except for the 5-gene into the ΔphaC1ΔphaC2. The primers used for cassette construction are listed in Supplementary Table 12. The primers used for cassette sequencing are listed in Supplementary Table 13.

Substrate measurement

Substrate concentrations at the beginning (T₀) and the end (T_f) were measured to calculate carbon and electron conversion efficiency to n-butanol. The incubation time of each experiment can be found in Supplementary Table 3.

CO₂ and H₂ analysis by gas chromatography

CO₂ and H₂ were analyzed using a method described in a previous study²⁷. Gas samples were analyzed using gas chromatography (Shimadzu BID 2010-plus, equipped with Rt^®-Silica BOND PLOT Column, 30 m × 0.32 mm; Restek, USA) with helium as a carrier gas. To measure the CO₂ content of the liquid phase, 1 mL of the cell-free liquid phase was added to 15 mL helium-flushed septum-capped glass vials (Exetainer, Labco, Houston) containing 1 mL 85% phosphoric acid. Then 40 μL of the resulting gas from the Balch tube was injected into the Shimazu GC-BID, using a Hamilton^TM gas-tight syringe. To measure the CO₂ and H₂ contents of the gas phase, either 40 μL of the gas phase was directly injected into the Shimadzu GC-BID, or 5 mL of the gas phase was injected into a 15 mL helium-flushed septum-capped glass vial (Exetainer, Labco, Houston), using a Hamilton^TM gas-tight syringe. Then 50 μL of the diluted gas sample was injected into the Shimazu GC-BID, using a Hamilton^TM gas-tight syringe. A standard curve was generated by the injection of 10 μL, 25 μL, and 50 μL of H₂ + CO₂ (80%, 20%). The total moles of CO₂ in the reactors were calculated using the ideal gas law (PV = nRT)⁷¹.

Organic acid analysis by ion chromatography

For measuring organic acid concentration, after 1:50 dilution, the acetate and 3-hydroxybutyrate concentrations at the starting and endpoint of culture for each sample were quantified using an Ion Chromatography Metrohm 881 Compact Pro with a Metrosep organic acid column (250 mm length). Eluent (0.5 mM H₂SO₄ with 15% acetone) was used at a flow rate of 0.4 mL min⁻¹ with suppression (10 mM LiCl regenerant)²⁷.

Ferrous iron [Fe(II)] analysis by ferrozine assay

The Fe(II) concentration measurement was done using 10 μL of culture mixed with 90 μL 1 M HCl in a 96-well plate inside the anaerobic chamber (filled with 5% H₂/75% N₂/20% CO₂, Coy Laboratory, Grass Lake, MI). After the plate was removed from the anaerobic chamber, 100 μL of ferrozine (0.1% (w/v) ferrozine in 50% ammonium acetate) was added to the sample. Then the 96-well plate was covered with foil and incubated at room temperature for 10 min. before the absorbance was measured at 562 nm. The absorbance was then converted to Fe(II) concentration based on a standard curve generated by measuring the absorbance from 0 mM, 1 mM, 2.5 mM and, 5 mM Fe(II).

In vivo production of n-butanol

The plasmids with the n-butanol pathway were unstable when adapting the strain to the nitrogen-fixing or photoautotrophic conditions. To avoid this problem, a twice-washed heavy inoculum from YPSMOPS was used under all conditions. All strains were inoculated in 50 mL of YPSMOPS with kanamycin with a 1:50 dilution from a pre-grown culture. When the OD₆₆₀ reached 0.6–0.8, the culture was inoculated into 300 mL of YPSMOPS with kanamycin. When the OD₆₆₀ reached 0.8~1, 10 mL of culture was saved for a PCR check (Supplementary Fig. 10). The rest of the culture was washed twice with ammonium-free FW medium and resuspended using anoxic ammonium-free FW medium inside the anaerobic chamber. Finally, the culture was inoculated into the medium containing different carbon sources and electron donors (acetate, 3-hydroxybutyrate, H₂, Fe(II), or electrode) in either a sealed Balch tube (initial OD₆₆₀ ~1) or a BEC (initial OD₆₆₀ ~0.7). The tubes and the reactors were sealed throughout the process, and samples were taken after the cultures reached the stationary phase (incubation time listed in Supplementary Table 3), using sterile syringes.

Extraction and quantification of n-butanol and acetone

After the culture entered the late stationary phase, 1 mL of culture was removed from the culture tube using a syringe and centrifuged at 21,100 × g for 3 min. The supernatant was then filtered using a syringe filter, and the filtrate or the standard was extracted with an equal volume of toluene (containing 8.1 mg/L iso-butanol as an internal standard) and mixed using a Digital Vortex Mixer (Fisher) for 5 min. followed by centrifugation at 21,100 × g for 5 min. After centrifugation, 250 μL of the organic layer was added to an autosampler vial with an insert. The organic layer was then quantified with GC-MS (Shimazu GCMS-QP2010 Ultra), using the Rxi^®-1ms column. The oven was held at 40 °C for 3 min, ramped to 165 °C at 20 °C/min, then held at 165 °C for 1 min. Samples were quantified relative to a standard curve for 0, 0.2025, 0.405, 0.81, 2.025, 4.05, and 8.1 mg/L of n-butanol and 0, 0.784, 3.92, 7.84, 39.2, 78.4, and 392 mg/L of acetone. An autosampler was used to reduce the variance of injection volumes.

Bioelectrochemical platforms and growth conditions

A three-electrode sealed-type bioelectrochemical cell (BEC, C001 Seal Electrolytic Cell, Xi’an Yima Opto-electrical Technology Com., Ltd, China)^30,68 containing 80 mL of FW medium was used for testing n-butanol production. The three electrodes were configured as a working electrode (a graphite rod, 3.2 cm²), a reference electrode (Ag/AgCl in 3.5 M KCl), and a counter electrode (Pt foil, 5 cm²). FW medium (76 mL) was dispensed into sterile, sealed, three-electrode BECs, which were bubbled for 60 min. with N₂ + CO₂ (80%/20%) to remove oxygen and pressurized to ~7 psi. Four BECs were operated simultaneously (n = 3 biological replicates) with one no-cell control. All photoelectroautotrophic experiments were performed at 26 °C under continuous infrared light (880 nm) or halogen light. The electrical potential of 0.5 V (E_appl= 0.5 V) was constantly applied (240 h) to the working electrode with respect to the reference electrode (Ag/AgCl in 3.5 M KCl) and counter electrode using a grid powered potentiostat (Interface 1000E, Gamry Multichannel potentiostat, USA) or The solar panel (Uxcell 0.5 V 100 mA Poly Mini Solar Cell Panel Module) with the output voltage 0.5 V (E_appl= 0.5 V) was directly connected to the bioreactors for 240 h and the resulting current uptake/electron uptake to the bioreactor was measured with the resistor using ohm’s law of electrical current. Electron uptake was collected every 1 min using the Gamry Echem Analyst™ (Gamry Instruments, Warmister, PA) software package. At the end of the bioelectrochemical experiment, the samples were immediately collected from the BEC reactors. n-butanol, acetone, and substrates were measured as described above.

Calculations of CCE, electron conversion efficiency, electrical energy conversion efficiency, and electron consumption of n-butanol, acetone, CO₂, H₂, and biomass biosynthesis

CCE, electron conversion efficiency, and EECE were calculated by dividing the total carbon/electrons/electrical-energy consumption by the final carbon/electrons/energy content in n-butanol, respectively.

To determine carbon consumption, acetate, 3-hydroxybutyrate, or CO₂ consumption was calculated by subtracting the amount in the sample at the end of the experiment from the amount at the beginning of the experiment. Then all the carbon substrate consumptions were converted to moles of carbon, using Eq. (1). The amount of carbon converted to n-butanol was calculated based on the n-butanol production, using Eq. (2). The CCE was calculated using Eqs. (1), (2), and (3) below:

The theoretical total number of electrons available from each consumed electron donor was calculated as described below (Eq. (4)). The total available electrons from the complete oxidation of each organic acid were calculated with the assumption that the final oxidation product was CO₂. The inorganic electron donors such as Fe(II) and H₂ release 1 mole e⁻ and 2 moles e⁻ per mole, respectively. Electrons supplied for the photoelectroautotrophy condition were calculated directly from BEC-based experiments wherein the total current uptake was integrated over the operational time. The total electron uptake was used to calculate the electron conversion efficiency to n-butanol because the electrode is the direct electron donor under this growth condition. The number of electrons required for n-butanol production was calculated from the oxidation state of the carbon in each carbon source and n-butanol. Supplementary Table 14 lists the specific oxidation state, and the number of electrons required per mole of n-butanol is listed for all studied sources and n-butanol.

To calculate the total available electrons from each substrate, the amount of consumed substrate (in moles) was multiplied by the theoretical total available electrons per mole of the substrate when fully oxidized to CO₂ (Eq. (4)). For photoelectroautotrophy, the total available electron was calculated based on data collected from a data acquisition system (DAQ, Picolog Datalogger). To obtain the electrons required for n-butanol production, the n-butanol production (in moles) was multiplied by the theoretical number of electrons required per mole (Eq. (5)). The conversion efficiency was calculated by dividing the moles of electrons required for n-butanol production by the theoretical total available electrons (Eq. (6)).

Calculation of the EECE to n-butanol was adapted from a previous study¹⁹. The EECE was calculated by Eq. (7). The charge supplied to the bioelectrochemical platforms was calculated from data collected by DAQ.

The Gibbs free energy gains (({Delta }_{{{{{{rm{r}}}}}}}{{{{{{rm{G}}}}}}}^{0})) for n-butanol was calculated similarly with a previous study¹⁹ by reaction 8 and Eq. (9)⁷².

Electron consumption of n-butanol, acetone, CO₂, H₂, and biomass biosynthesis, were calculated by the mole of production multiplied by the electrons required for each mole of product. The molar production of n-butanol and acetone was determined by the titer divided by molecular weight. The molar production of CO₂, H₂, was measured by GC-BID. The molar production of biomass was calculated by the OD₆₆₀ change between T₀ and T_f by Eq. (10).

Cell number/ml: 8 × 10⁸ cell/ml/OD, cell weight: 10⁻¹²g/cell⁷³, Molecular weight of biomass: 22.426 g/mol⁴².

Determination of glycogen content

TIE-1 cells were grown in freshwater medium with NH₄Cl or under nitrogen-fixing condition (with N₂) supplemented with 10 mM 3-hydroxybutyrate to and OD₆₆₀ of 1.8 mL of bacterial culture was pelleted and washed three times with ultrapure water and resuspended in 30% (w/v) KOH with for glycogen extraction. Samples were then incubated at 95 °C for 2 h. Glycogen was precipitated by the addition of ice-cold ethanol to a final concentration of 75%. Samples were put on ice for 2 h followed by 10 min. centrifugation 10,000 × g at 4 °C. The precipitated glycogen was then washed twice with pure ethanol and dried for 20 min. at 60 °C. Glycogen samples were resuspended in 250 µL of 100 mM sodium acetate (pH 4.5) and digested with 2 mg/ml amyloglucosidase (Sigma Aldrich A7420) for 2 h at 60 °C. Samples were added with infinity glucose hexokinase liquid reagent (Thermo scientific TR1542) at a ratio of 1:150 according to the manufacturer’s recommendation and absorbance reading was done at 340 nm⁷⁴.

RNA extraction, cDNA synthesis, and RT-qPCR

To extract RNA for cDNA synthesis and eventually perform RT-qPCR for analyzing the expression level of the individual genes, culture samples (2.5 ml to 15 ml depending on OD₆₆₀) were taken at the late exponential (T_m) or stationary phase (T_f). Samples were immediately stabilized with an equal volume of RNAlater (Qiagen, USA). After incubation at room temperature for 10 min, samples were centrifuged at 21,100 × g for 3 min. After the supernatant was removed, the pellet was stored at −80 °C before RNA extraction using the Qiagen RNeasy Mini kit (Qiagen, USA), following the manufacturer’s protocol. DNA was removed using a Turbo DNA-free Treatment and Removal Kit (Ambion, USA). DNA contamination was ruled out by PCR using the primers listed in Supplementary Tables 12 and 13.

Purified RNA samples were then used for cDNA synthesis by an iScript^TM cDNA Synthesis Kit (Biorad, USA). The same mass of RNA was added to each cDNA synthesis reaction. The synthesized cDNA was used for RT-qPCR. RT-qPCR was performed using the Biorad CFX connect Real-Time System Model # Optics Module A with the following thermal cycling conditions: 95 °C for 3 min, then 30 three-step cycles of 95 °C for 3 s, 60 °C for 3 min, and 65 °C for 5 s, according to the manufacturer’s manual. The reaction buffer was iTaq SYBR Green Supermix with ROX (Bio-Rad). The primers used for RT-qPCR (listed in Supplementary Table 15) were designed using primer3 software (http://bioinfo.ut.ee/primer3/). The primer efficiencies were determined by performing RT-qPCR using different DNA template concentrations. The genes clpX and recA, which have been previously validated as internal standards, were used^29,30. The gene code for kanamycin resistance was also used as an internal standard for the plasmid. After RT-qPCR, the data were analyzed using the ΔΔC_T method.

NADH/NAD⁺ measurement

Wild-type TIE-1 and Nif mutants were grown with either 3-hydroxybutyrate or H₂ and using NH₄Cl as a nitrogen source. Briefly, 1.8 mL cell cultures from the stationary phase (same phase at which samples were taken for measuring n-butanol) was spined at 21,000 × g for 1 min. inside an anaerobic chamber. Then the pellet was resuspended in either 300 μL 0.2 M sodium hydroxide (for NADH extraction) or 300 μL 0.2 M hydrochloric acid (for NAD⁺ extraction). The resuspension was then incubated at 50 °C for 10 min. and cooled to below 20 °C on an ice block. While vortexing on medium speed, Equal volume 0.1 M acid or base was added to neutralize the sample. After spinning at 21,000 × g for 5 min, the supernatant was stored in freezer for the following assays. The enzyme cycling assays were then performed on a BioTek Synergy^TM HTX 96-well plate reader measuring absorbance at 570. The amount of NADH/NAD⁺ was quantified relative to a standard curve ranging from 0 to 5 μM.

Transmission electron microscopy (TEM)

Wild-type TIE-1 and Phb mutants grown with 3-hydroxybutyrate with either N₂ or NH₄Cl as a nitrogen source was used as representative samples for TEM. Briefly, 5 mL planktonic cell suspensions were centrifuged at 6000×g for 5 min. followed by primary fixation by resuspending the cell pellets in 2% formaldehyde and 2.5% glutaraldehyde in 0.05 M sodium cacodylate buffer (pH 7.2) for ~45 min. at room temperature. Cell pellets were agar encapsulated followed by primary fixation for ~20 min. Polymerized agar was cut into small cubes and were subjected to secondary fixation for ~5 h followed by acetone dehydration and resin infiltration. Ultrathin sections (~70 nm) were cut on a Reichert Ultracut UCT ultramicrotome (Leica, Buffalo Grove, IL, USA), mounted on copper grids (FCFT300-CU-50, Electron Microscopy Sciences, Hatfield, PA, USA), and counterstained with lead citrate for 8 min⁷⁵. The sample was imaged with a LEO 912 AB Energy Filter Transmission Electron Microscope (Zeiss, Oberkochen, Germany). Images were acquired with iTEM software (ver. 5.2) (Olympus Soft Imaging Solutions GmbH, Germany) with a TRS 2048 × 2048k slow-scan charge-coupled device (CCD) camera (TRÖNDLE Restlichtverstärkersysteme, Germany). Each TEM image was acquired at ×10,000 magnification and 1.37 nm pixel resolution.

Viability analysis of TIE-1 under photoelectroautotrophy

WT TIE-1 was inoculated into the bioelectrochemical reactors described above, with a starting OD of ~0.3. After 72 h of incubation, the viability of the biofilm attached to the electrode was characterized by imaging the electrode after staining with the LIVE/DEAD^® (L7012, Life Technologies) kit. The attached cells were quantified using NIS-Elements AR Analysis 5.11.01 64-bit software. For imaging of the electrode, prior to cutting a piece of the spent electrode, the electrode from the reactor was washed three times with 1× phosphate-buffered saline (PBS) to remove unattached cells. A piece of the spent electrode was then submerged in 1× PBS in a sterile microfuge tube. Prior to imaging, the electrode piece was immersed in LIVE/DEAD^® stain (10 μM SYTO9 and 60 μM propidium iodide) kit and incubated for 30 min. in the dark. The electrode sample was then placed in a glass-bottom Petri dish (MatTek Corporation, Ashland, MA) containing enough PBS to submerge the sample. Further, it was imaged on a confocal microscope (Nikon A1 inverted confocal microscope), using 555 and 488 nm lasers and a ×100 objective lens (Washington University in St. Louis Biology Department Imaging Facility). Electrode attached cells were quantified by Elements Analysis software using the protocol described below: Briefly, for each reactor, three images were processed. Z-stacks of each image were split into two channels (one for live cells, one for dead cells), the MaxIP was acquired for the combined z-stacks. After GaussLaplace, local contrast and smoothing, and thresholding, and Object Count was performed for each channel based on a defined radius (0.8–5 μm). Then the percentage of live (or dead) cells was calculated by

Toxicity study

WT TIE-1 with an empty vector (pRhokS-2) was used to test the tolerance of TIE-1 for acetone and n-butanol. To test the tolerance, 0, 0.25, 0.5, 1, or 2% n-butanol (v/v), or 0, 0.1, 0.25, 0.5, 1, or 2% acetone (v/v), was added to FW media with acetate (10 mM). Growth was monitored by recording OD₆₆₀ over time.

Statistics

All statistical analyses (two tails Student’s t-test) were performed with Python. p-value < 0.05 was considered to be significant. p-values are presented in supplementary data 2, and estimated effect size (Cohen’s d) are presented in supplementary data 3. For most of the experiments data are from n = 3 of biologically independent samples, from each biologically independent samples n = 3 technical replications were performed, For photoelectroautotrophy measurements, data are from n = 2 of biologically independent samples, from each biologically independent samples n = 3 technical replications were performed. For RT-qPCR for photoelectroautotrophy which data are from n = 2 of biologically independent samples, from each biologically independent sample n = 2 technical replications were performed.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.