Electroenzymatic Nitrogen Fixation Using a MoFe Protein System Immobilized in an Organic Redox Polymer

Abstract We report an organic redox‐polymer‐based electroenzymatic nitrogen fixation system using a metal‐free redox polymer, namely neutral‐red‐modified poly(glycidyl methacrylate‐co‐methylmethacrylate‐co‐poly(ethyleneglycol)methacrylate) with a low redox potential of −0.58 V vs. SCE. The stable and efficient electric wiring of nitrogenase within the redox polymer matrix enables mediated bioelectrocatalysis of N3 −, NO2 − and N2 to NH3 catalyzed by the MoFe protein via the polymer‐bound redox moieties distributed in the polymer matrix in the absence of the Fe protein. Bulk bioelectrosynthetic experiments produced 209±30 nmol NH3 nmol MoFe−1 h−1 from N2 reduction. 15N2 labeling experiments and NMR analysis were performed to confirm biosynthetic N2 reduction to NH3.


Experimental Procedures Materials
All chemicals and materials were purchased from Alfa-Aesar, Sigma Aldrich, VWR, and Acros Organics and were used as received except otherwise noted. The monomers glycidyl methacrylate (GMA), methyl methacrylate (MMA) and poly(ethylene glycol methacrylate) (PEGMA, Mn = 500, dissolved in iso-propyl alcohol, 0.052 g mL -1 ) were passed to a short column filled with an inhibitor remover and stored at -20 °C (GMA, MMA) or room temperature (PEGMA). The initiator AIBN (azobisisobutyronitrile) was recrystallized from hot toluene or hot methanol. [1] CD 3 CN for NMR spectroscopy was purchased from Deutero and stored at 4° C. For the modification of the polymer backbone P(GMA-MMA-PEGMA) with neutral red, the free base was used which was precipitated from an aqueous solution by adding a base. The free base (red-orange powder) was dried in an oven and stored at room temperature.

Methods
All NMR experiments were conducted with a DPX250 spectrometer from Bruker with a 1 H resonance frequency of 250.13 MHz. All measurements were conducted at room temperature. Samples were filtered to remove particles. All UV-vis measurements were conducted with a Cray60 spectrophotometer from Agilent in quartz cuvettes with an optical path length of 1 cm. Electrochemical characterization of the redox polymer was performed with a PalmSens2 v4.4 (PalmSens) potentiostat with a three-electrode configuration in standard electrochemical glass cells. Polymer modified electrodes were prepared by a standard drop cast process and dried for several our prior to measurement. Size exclusion chromatography (SEC) was performed against polystyrene (PS) standards in THF at 30 °C. Data from the RI detector were analyzed with the PSS WinGPC Unity software. Sample concentration was 15 mg mL -1 .

Synthesis of the polymer backbone poly(glycidyl methacrylate-co-methyl methacrylate-co-poly(ethylene glycol) methacrylate)), P(GMA-MMA-PEGMA)
The synthesis of the polymer backbone P(GMA-MMA-PEGMA) was conducted following procedures as described in ref. [2] The monomers GMA (500.16 mg, 3.52 mmol), MMA (280.8 mg, 2.81 mmol) and PEGMA (0.052 g mL -1 , 6.73 mL, 350 mg, 0.7 mmol) were dissolved in THF (5 mL) under argon atmosphere. The radical initiator AIBN (4.9 mg, 0.03 mmol) was added and the reaction mixture was heated to 70 °C under an argon atmosphere (closed Schlenk tube, connected to a pressure release valve). The transparent and colorless reaction mixture was stirred for 285 min. The solution was cooled down to room temperature and the product was precipitated by adding 120 mL of n-pentane to yield the crude polymer as a colorless solid. The turbid solution was decanted off and centrifuged and the precipitate was combined with the colorless crude polymer. The product was re-dissolved in THF (≈15 mL) and precipitated again by adding n-pentane. The solution was again decanted off and centrifuged and the purification step was again repeated. The combined solids were air dried and the remaining solvent was removed by shortly applying a vacuum. The solid was immediately dissolved in MeCN (20 mL) and stored as a stock solution with a concentration of 19 mg mL -1 at room temperature. Yield: 380 mg (33 %). 1

Synthesis of the redox-active polymer neutral red modified poly(glycidyl methacrylate-co-methyl methacrylate-co-poly (ethylene glycol) methacrylate), P(GMA-MMA-PEGMA)-NR
Under an argon atmosphere, neutral red (free base, 36.25 mg, 0.13 mmol, 0.46 eq with respect to the GMA amount in the backbone) was suspended in 5 mL of DMSO, and the organic base NEt 3 (21.9 mg, 0.22 mmol) was added. The slurry was stirred until the base was fully dissolved. Then, P(GMA-MMA-PEGMA) solution (19 mg mL -1 in MeCN, 5.25 mL, 99.75 mg, 0.28 mmol GMA units assuming 49 wt% of GMA within the backbone) was slowly added with a syringe. The resulting mixture was heated to 60 °C and stirred for two days in a closed vessel equipped with a pressure release valve. The brown reaction mixture was cooled to room temperature, and most of the organic solvent was removed under vacuum. The remaining solution was quenched with water and acidified with a few drops of diluted aqueous HCl (protonation of the neutral red moieties to increase water solubility, color changes from brown to red). The crude mixture was purified by membrane filtration using membrane filters with a molecular weight cutoff of 10 kDa (VivaSpin, Turbo 15, Sartorius) and centrifugation. The aqueous workup was stopped when the washing solution became colorless. The polymer was obtained as a red-brown aqueous solution with a concentration of 17 mg mL -1 . Note that the secondary amine that is formed after binding of NR to the backbone shows a higher nucleophilicity than the primary amine in the free base and might thus react with a second epoxide function (intra-and/or interpolymeric reaction). Hence, the exact degree of binding of the NR moiety is not known It should be noted that for clarity reasons only the product of the mono-addition of the NR modifier to the epoxide modified polymer is shown in Figure 2, indicated by the dashed line. The secondary amine formed at the polymer may react with an additional epoxide function within the backbone. The resulting P(GMA-MMA-PEGMA)-NR shows high electrochemical and physical stability. After 50 continuous CV scans of P(GMA-MMA-PEGMA)-NR at 100 mV s -1 the reductive peak currents decreased only by <7%.

Growth of Azotobacter vinelandii and purification of nitrogenase
Mo-dependent nitrogenase from A. vinelandii was expressed as described previously. [3,4] Briefly, the wild-type cells were cultivated in 18 L modified Burk medium supplemented with 10 mM ammonium (NH 4 + ). Then, the cells were collected via centrifugation and resuspended into NH 4 + free growth media to enhance nitrogenase expression. The cells were regrown until the optical density at 600 nm reached 1.5. All of the following steps were carried out under strictly anaerobic conditions. The cells were lysed by sonication in the presence of 2 mM sodium dithionite (DT). The crude extract was separated by anion exchange chromatography (Q-Sepharose, GE Healthcare) over a linear gradient of sodium chloride (NaCl) from 200 mM to 640 mM to separate the MoFe and FeP proteins. Histagged MoFe was purified using HisTrap HP column (GE Healthcare) and were desalted using a Hitrap desalting column equilibrated with 100 mM MOPS and 2 mM DT at pH = 7.0. The purified protein was concentrated to ≥ 20 mg mL -1 and was shock-frozen in liquid N 2 until further use. The purity of the MoFe proteins was greater than 95% based on the SDS-PAGE analysis using Coomassie blue staining. The purity of the protein was verified by SDS-PAGE gel ( Figure S3). Pure protein was concentrated to ~20 mg/mL.

Electrochemical analysis
All electrochemical analyses were performed anaerobically (<1 ppm O 2 ) to avoid competitive O 2 reduction at the working electrode and to preserve the activity of the MoFe nitrogenase. pH 6.0 100mM MOPS buffer was degassed and used as the supporting electrolyte to avoid migration issues associated with N 3and NO 2 -. Electrochemical measurements were performed using a CH Instruments potentiostat model 1230b, with a saturated calomel reference electrode (SCE), a platinum mesh counter electrode, and the carbon paper working electrode (0.5 cm x 0.5 cm) in a standard 3-electrode configuration. Electrolysis was carried out at a constant potential of either −0.3, −0.4, −0.5, −0.6, −0.7 or −0.8 V vs SCE with rapid stirring (400 rpm) using 100 mM MOPS buffer, pH 6.0 at 25 ºC. All voltammograms are plotted in IUPAC notation with cathodic currents being negative.

Ammonia/ammonium fluorescence detection assay
The produced NH 3 was quantified as previously reported. [3,5,6] A reagent to detect NH 3 was prepared by mixing 100 mL phosphate buffer (pH 7.3, 200 mM), 25 μL of 2-mercaptoethanol, and ortho-phthalaldehyde (270 mg dissolved in 5 mL of ethanol, Figure S4). Note: this reagent is light-sensitive and it has to be used and stored in the absence of light. Following bioelectrosynthesis experiments, 250 μL of the electrolyte solution containing the formed NH 3 was mixed with 1000 μL of the premade reagent for NH 3 detection. The mixture was incubated for 30 min. The produced NH 3 was quantified by measuring the fluorescence emission at 472 nm when excited at a wavelength of 410 nm. All samples were triplicated and corrected against background samples prepared with BSA or denatured MoFe protein bioelectrodes.

15
N labeled experiments were conducted to verify that NH 3 is produced electrochemically. Bulk electrolysis cells were assembled under an Ar/H 2 (3.2%) atmosphere. Experiments performed under 15 N 2 were carried out in a sealed three-neck flask. Electrolysis cells were degassed after assembly under vacuum at room temperature for 1 h to ensure all H 2 was removed and then filled with 15 N 2 . All electrolysis experiments were carried out for 24 h. Electrolysis experiments were performed under 15 N 2 from Sigma Aldrich (MDL # MFCD00084232, >98% isotopic purity purity 15 N; 99% (CP) assay). Gases were purged through sequential solutions of 0.1 M HCl and 10 mM sodium dithionite to trap any contaminant NH 3 and O 2 , respectively. After electrolysis experiments, samples were first spun through a 10 kDa MWCO membrane to remove any leached protein. The NMR spectra were collected at room temperature on a 500 MHz ( 1 H) Varian Inova spectrometer equipped with a cryoprobe. 1 H 1-D spectra were collected using a selective excitation pulse for water suppression and a spectral width of 8012.8 Hz (16.0 ppm). Spectra were collected both with and without 15 N decoupling during acquisition to demonstrate the absence or presence of 15 N-enriched molecules. 1 H-NMR signals were averaged over 24 h in all cases, and baseline subtraction was utilized to account for the overlapping residual peak corresponding to water (after solvent suppression). Figure S1. The nitrogenase catalytic cycle illustrating N2 reduction involving the transient binding of the Fe protein with the MoFe protein and ATP hydrolysis..           Table S1. Calculated Faradaic efficiencies for N3and NO2reduction by MoFe/P(GMA-MMA-PEGMA)-NR bioelectrodes. The theoretical maximum quantity of NH3 anticipated to be generated as a function of the electron transferred during the electrosynthesis was also calculated based on the reaction equations.