Perfluorocarbon Nanoemulsions Create a Beneficial O₂ Microenvironment in N₂-fixing Biological | Inorganic Hybrid

SUMMARY

Powered by renewable electricity, biological | inorganic hybrids employ water-splitting electrocatalysis and generate H₂ as reducing equivalents for microbial catalysis. The approach integrates the beauty of biocatalysis with the energy efficiency of inorganic materials for sustainable chemical production. Yet a successful integration requires delicate control of the hybrid’s extracellular chemical environment. Such an argument is evident in the exemplary case of O₂ because biocatalysis has a stringent requirement of O₂ but the electrocatalysis may inadvertently perturb the oxidative pressure of biological moieties. Here we report the addition of perfluorocarbon (PFC) nanoemulsions promote a biocompatible O₂ microenvironment in a O₂-sensitive N₂-fixing biological | inorganic hybrid. Langmuir-type nonspecific binding between bacteria and nanoemulsions facilitates O₂ transport in bacterial microenvironment and leads to a 250% increase in efficiency for organic fertilizers within 120 hours. Controlling the biological microenvironment with nanomaterials heralds a general approach accommodating the compatibility in biological | inorganic hybrids.

eTOC BLURB

Hybrids of materials and microorganisms is versatile and efficient towards a sustainable chemical production. Yet a biocompatible interface that mimics biocatalysts’ native environment is required despite varying application scenarios. Here the addition of nanoemulsion is reported to promote a biocompatible microenvironment for O₂-sensitive microorganisms in air for electricity-driven N₂ fixation.

INTRODUCTION

Natural biological systems possess time-tested catalytic capability of chemical synthesis. Therefore, one approach to tackle the challenge of distributed chemical production with renewable energy is to interface microbial catalysts with inorganic electrochemical materials powered by solar panels, namely the biological | inorganic hybrid, which combines the energy efficiency of synthetic materials with the selectivity and specificity of biochemical synthesis^1–3. Significant development has been observed in this fledgling direction for solar-driven reduction of CO₂ and N₂ with high efficiency and reaction throughput^4–10.Water-splitting inorganic electrocatalysts generate H₂ as the reducing equivalents for microbial catalysts, which enables a selective production of multi-carbon commodity chemicals^4,8,10–16 pharmaceutical precursors^10,11 and nitrogen fertilizers from N₂ even at low partial pressures^{5,6,9,17–20} Fundamentally, the key towards a successful hybrid system is the compatibility between the abiotic and biological components. For example, the design and application of an inorganic water-splitting electrocatalytic system that yields minimal reactive oxygen species achieves high biocompatibility with bacteria of CO₂/N₂ fixation, leading to energy efficiency of solar-driven CO₂ fixation up to 10%^8,9. Addressing biocompatibility with ingenious materials design offers a general route viable for effective production of chemicals with a biological | inorganic hybrid.

Reconciling the O₂ demand of biochemical synthesis with water-splitting electrocatalysis in a biological | inorganic hybrid is essential for efficient production of a wide range of chemicals. O₂ serves as the terminal electron acceptor in many biochemical pathways and is ubiquitous in extracellular medium²¹. Nonetheless, many of the upstream catalysts in biochemical pathways are sensitive if not incompatible with even trace amounts of O₂. Take biological N₂ fixation as one example. N₂-fixing nitrogenase enzymes reduce N₂ with the supply of reducing equivalents and adenosine triphosphate (ATP), and are sensitive to O₂ due to its low-valence Fe- and FeMo-based sulfur clusters^22–24. Meanwhile, many N₂-fixing diazotrophs demand O₂ as the terminal electron acceptor for the production of ATP and reducing equivalents, the ubiquitous energy source for all biochemical reactions and microbial survival. Therefore, there is a stringent requirement of microbial O₂-intake flux, represented as an extracellular O₂ concentration of roughly 1~5%, for typical N₂-fixing microbes that utilize O₂ as the terminal electron acceptors in the electron transport chain^9,25–27. In the biological | inorganic hybrid, the water-splitting electrocatalysis and its generation of O₂ introduce perturbations to the extracellular environment that may adversely impact the delicate O₂ balance for N₂-fixing microbes. Previously reported biological | inorganic hybrids with N₂/CO₂-fixing bacterium Xanthobacter autotrophicus^26–27 addresses such an O₂ requirement by feeding a constant gas stream of O₂/CO₂/N₂ gas with 2% O₂ and forcing a biocompatible microaerobic atmosphere⁹⁹. Such an enforced microaerobic environment is not ideal for practical application nor energy efficient at the system level due to the energy cost of gas mixing for a precise O₂ partial pressure. O₂-reactive redox mediators such as viologen ^28–30 or O₂-reducing wire array electrodes¹⁹ were employed previously to create microenvironments of reduced O₂ concentrations commensurate with the enzymatic or microbial O₂ demands. Yet those designs based on redox reactions with O₂ consume electricity and pose additional overhead towards the hybrids’ efficiency. Alternatively, we aim to design a strategy that does not consume electrons while still creates a desirable extracellular O₂ microenvironment. We hypothesize that the oxidative generation of O₂ from water-splitting electrocatalysis, previously untapped and perceived as an unwanted perturbation to biological O₂ compatibility, should be sufficient to support the O₂ demand of the N₂-fixing X. autotrophicus in an O₂-free external environment, if the electrochemically generated O₂ can be efficiently transported for a desirable bacterial extracellular microenvironment (Figure 1A). Such consideration leads to the search of a design from a materials perspective that creates a controllable biological gas microenvironment.

Figure 1. Microbial microenvironment created by perfluorocarbon (PFC) nanoemulsion.

A, schematic of the bacterial microenvironment created by PFC nanoemulsions and its benefits for O₂ mass transport. CoP, cobalt-phosphorous alloy for hydrogen evolution reaction; CoPi, cobalt phosphate for water oxidation reaction. B, Schematic of colocalization experiment with fluorescent confocal microscopy. C, Fluorescent images of bacteria Xanthabacter autotrophicus (X. a.) stained with SYTO™ 9 (pseudo-colored green), PFC nanoemulsions tagged with fluorous rhodamine (pseudo-colored red), and the overlay images of both fluorescent emissions. PFC nanoemulsion loading, 2.5 volume percentage (v/v). D, Zoom-in images in the highlighted area in C. Scale bars = 10 and 1 μm in C and D, respectively.

Here we report that the addition of perfluorocarbon (PFC) nanoemulsions modulates microbial O₂ microenvironment and satisfies the O₂ demand of X. autotrophicus for electricity-driven fixation of N₂ and CO₂ in a reactor headspace deprived of O₂ (Figure 1A). An exemplary case of biocompatible nanoemulsions with a gas solubility 20 times of aqueous solution^31–33, PFC nanoemulsions have been reported to facilitate the mass transport of H₂ as a redox mediator and increase the overall reaction throughput of CO₂ fixation in a biological | inorganic hybrid¹². We posit that the PFC nanoemulsions will bind to the microbes thus forming a microenvironment surrounding X. autotrophicus for efficient transport of electrochemically generated O₂ in a O₂-limited reactor headspace. Quantitative characterizations indicate that Langmuir-type nonspecific binding between the microbes and PFC nanoemulsions creates a gas microenvironment of enhanced O₂ transport and availability. The addition of PFC nanoemulsions preserves cell viability in a O₂-free gas environment, fully utilizes the electrochemically generated O₂, and leads to a 250% increase in the overall electron efficiency for N₂ fixation within a 120-hour duration. Designing the extracellular microenvironment with nanomaterials offer a viable route that reconciles O₂-dependent microbial energy production of and the O₂ sensitivity of nitrogenase to create synergy and achieve optimal compatibility in biological | inorganic hybrid.

RESULTS AND DISCUSSION

Fluorescent colocalization experiments using confocal microscopy suggest that PFC nanoemulsions bind strongly to CO₂/N₂-fixing X. autotrophicus and creates a unique bacterial microenvironment. X. autotrophicus (American Type Culture Collection, ATCC 35674) was autotrophically cultured in a minimal medium without any nitrogen and organic carbon under a 1-bar gas environment of 2% O₂, 60% N₂, 20% H₂, and 18% CO₂ (See Supplemental Information section 1.1)^9,19,26. PFC nanoemulsions of perfluorodecalin and perfluorohexane were prepared by ultrasonication in the same minimal medium with Pluronic F68 as surfactant (See experimental procedures)³⁴. Experiments of dynamic light scatting determined an averaged diameter of about 240 nm for the prepared PFC nanoemulsions (Figure S1), consistent with the diameter of a stable emulsion formulated in our previous report¹². In order to tag the microbes and nanoemulsions with different fluorescent emitters, X. autotrophicus was incubated with SYTO™ 9, a green fluorescent nucleic acid stain, while red-emitting fluorous rhodamine reported previously³⁵ was incorporated into nanoemulsion during the ultrasonication stage of nanoemulsion preparation (See experimental procedures). When mixtures of fluorescently tagged X. autotrophicus (OD₆₀₀ = 0.1) and PFC nanoemulsions (2.5% volume percentage, v/v) were deposited on a coverslip (Figure 1B), in vitro confocal microscopy images in Figure 1C, 1D and Figure S2 suggest strong colocalization of the green emission from microbes (490~520 nm, named “STYO™ 9” in the figure) and the red emission from nanoemulsions (590~650 nm, named “Rhodamine” in the figure). Control experiment of X. autotrophicus in the absence of nanoemulsions displays minimal red emission, excluding any significant contribution of red microbial autofluorescence in the results. A closer examination of the overlay from both emissions (Figure 1C and 1D) display a corona of red emissions surrounding the green ones. This indicates that the PFC nanoemulsions surround and bind strongly to X. autotrophicus, creating an extracellular microenvironment that could dictate bacterial metabolism.

The binding between bacteria and PFC nanoemulsions can be quantitatively described by a Langmuir-type adsorption model with high binding affinity and low binding specificity towards microbial surface properties (Figure 2A). Non-labelled X. autotrophicus cultures were mixed with fluorescently tagged PFC nanoemulsions (10 % v/v) at different dilution ratios. The resultant mixture, whose concentrations C_PFC = C_PFC,0 / dilution ratio (C_PFC.0 = 10 % v/v), were studied by flow cytometry. Here the employment of flow cytometry allows us to quantitatively distinguish the free and microbe-bound nanoemulsions (Figure 2B and 2C) and determine the average number of PFC nanoemulsions per microbe (N) as a function of nanoemulsion dilution ratios. We observed a linear relationship between 1/N and the dilution ratio, we subsequently concluded that 1/N is linearly dependent on 1/C_PFC (Figure 2D). This observation suggests that a Langmuir monolayer adsorption model, illustrated in Figure 2A, is sufficient to describe the binding between microbes and nanoemulsions:³⁶

$$\frac{1}{N}=\frac{1}{N_{\mathit{max}}}+\frac{1}{N_{\mathit{max}}{K}_{eq}\left(C_{PFC}/C_{PFC,0}\right)}=\frac{1}{N_{\mathit{max}}}+\frac{dilutionratio}{N_{\mathit{max}}{K}_{eq}}$$ (1)

Here N_max denotes the maximum number of adsorbed nanoemulsions for one bacterial cell, K_eq the dimensionless equilibrium constant of the binding event, and C_PFC,0 = 10 % (v/v) as the PFC concentration in the starting nanoemulsion. Linear regression was conducted to extract the values of K_eq and N_max for X. autotrophicus: N_max = 13.8±5.9 and K_eq = 4.8±2.1×10⁴ (Figure 2D; n = 3, same below unless noted specifically). Because the PFC nanoemulsions are about 240 nm in diameter (Figure S1) and X. autotrophicus exhibits a typical ellipsoidal morphology of about 1 μm in size (Figure 1C, 1D and Figure S2)^19,27, the measured N_max value suggests a near-complete surrounding of nanoemulsion on the bacterial surface at maximum binding, i.e. a compartmentalized microenvironment created by the nanoemulsion for the N₂/CO₂-fixing bacterium. The large value of measured K_eq between PFC nanoemulsion and X. autotrophicus ensures the stability and prevalence of the created microenvironment under typical working conditions for the biological | inorganic hybrid (vide infra). We also determined the values of K_eq and N_max for 2-μm polystyrene microspheres of positive or negative charges (–NH₂ or –COOH functionalized, respectively), as well as acetogenic Sporomusa ovata that were previously employed with PFC nanoemulsions for boosted throughput in electricity-driven CO₂ fixation (Figure S3)¹². The values of K_eq and N_max persist at about 10⁴~10⁵ and 10⁰~10¹ (Figure 2E), respectively. In comparison with microbes, the mildly higher value of K_eq for polystyrene microspheres hints that the binding strength could be enhanced with enriched surface charges, yet other factors are also at play given the slightly lower values of N_max despite microspheres’ slightly larger size. Nonetheless, the microbial binding of PFC nanoemulsion seems strong, nonspecific, and can serve as a general venue of creating a customized microenvironment for microbial catalysts.

Figure 2. Quantification of microenvironment formation with flow cytometry.

A, The binding between the bacteria and PFC nanoemulsions follows the Langmuir adsorption model. N and N_max, the experimentally determined average and theoretical maximal number of nanoemulsions per microbe; K_eq, the dimensionless binding constant; C_PFC,0 & C_PFC, the concentrations of PFC nanoemulsion before and after dilution under a certain “dilution ratio”, respectively. B, Representative two-parameter density plot in flow cytometry based on forward scattering (FSC) and side scattering (SSC) which gates the populations of X. autotrophicus potentially with nanoemulsion adsorption (red circle) and free PFC nanoemulsion (blue circle). C, Histograms of fluorescent intensities for X. autotrophicus (red circle in B) in PFC-free microbial suspension (blue, dilution ratio = ∞), emulsion-microbe mixture with a dilution ratio = 10⁵ (yellow) and 10⁴ (red). D, Plot of 1/N vs. dilution ratio for X. autotrophicus. E, Experimentally determined K_eq and N_max values for X. autotrophicus (X. a.), Sporomusa ovata (S. o.), and 2-μm polystyrene microspheres with –NH₂ and –COOH surface functionalization groups (“PS-NH₂” and “PS-COOH, respectively), n = 3. Data are represented as mean ± SEM.

The observed bacterial microenvironment created by PFC nanoemulsions led us to probe the possible modulation of local O₂ supply due to PFC’s high gas solubilities^31–33. As the binding of PFC nanoemulsion appears nonspecific and persistently strong (vide supra), we posit that a platinum-based metallic electrode could serve as a surrogate of the bacterial surface for O₂-respiring X. autotrophicus, thus the electrochemical activity of the 4-electron O₂ reduction on Pt electrode^37,38 in the presence of nanoemulsions will yield information of the created microenvironment, similar to our previous study that employed PFC nanoemulsions for electricity-driven microbial CO₂ reduction¹². A Pt-based rotating disk electrode³⁹ (RDE) was deployed to create a well-defined O₂ mass transport for quantitative study (Figure 3A). A pH = 7.0 phosphate buffer solution was applied in lieu of the microbial minimal medium solution to avoid the interference from trace metals and other competing biochemical reactions. In this phosphate buffer, a PFC nanoemulsion concentration of 0.75% (v/v) was used for the RDE experiments, as the emulsion’s surfactant interferes with the catalytic activity of the Pt electrode at higher concentrations¹². Linear scan voltammograms at different rotation rates at 1-bar O₂ condition were obtained with the absence and presence of PFC nanoemulsions in the microbial minimal medium (Figure 3B and 3C, respectively), under a three-electrode electrochemical setup (See experimental procedures). A larger magnitude of reduction current densities (|i|) were observed with the 888nanoemulsion’s presence, suggesting a facilitated mass transport induced by nanoemulsions. Quantitatively, RDE’s well-defined profile of mass transport introduces the Koutecký-Levich equation³⁹:

$$\frac{1}{\left|i\right|}=\frac{1}{i_{\text{k}}}+\frac{1}{0.62nF{D}^{2/3}{v}^{-1/6}{C}_0}{\omega }^{-1/2}$$ (2)

Here i_k denotes the intrinsic current density of O₂ reduction on Pt surface, n = 4 for the presumed 4-electron reduction of O₂ on Pt³⁸, F the Faradaic constant, D the diffusion coefficient of O₂ in water, v the kinematic viscosity of liquid, C₀ the O₂ solubility at 1-bar O₂ condition, and ω the RDE’s angular rotation rate. Figure 3D plots 1/|i| versus ω^−1/2 at 0.3 V vs. Reversible Hydrogen Electrode (RUE) based on the data in Figure 3B and 3C. The similar slope between the two curves in Figure 3D suggest that the local O₂ solubility near the Pt electrode is not significantly perturbed with the presence of PFC nanoemulsion. Indeed, at 0.3 V vs. RHE, C₀ = 1.05 and 0.82 mM with and without nanoemulsion, respectively, when treating the values of D and v in the phosphate buffer the same as the ones in pure water (D = 2.5×10⁻⁹ m²•sec⁻¹, v = 0.801×10⁻⁶ m²•sec⁻¹)⁴⁰. Despite the similar values of C₀, at 0.3 V vs. RHE i_k = 91.0 and 4.44 mA/cm² in the presence and absence of PFC nanoemulsions, respectively. The 20-times difference in the value of i_k indicates that the observed bacterial microenvironment with PFC nanoemulsions facilitates the O₂ transport in the extracellular space proximate to the microbes without perturbing the local O₂ solubility. Since it is the O₂ transport into the microbes that dictates the intracellular O₂ concentration, the O₂ transport also dictates the delicate balance between biological N₂ fixation and O₂-dependent respiration,^25–27 the enhanced O₂ transport kinetic introduced by PFC nanoemulsion under steady state should lead to an increased O₂ availability for microbial metabolism in a O₂-limited gas environment in which the only O₂ supply comes from electrochemical water oxidation.

Figure 3. Electrochemical probing of O₂ kinetics in a nanoemulsion-based microenvironment.

A, Electrochemical reduction of O₂ at a Pt rotating disk electrode (RDE) as a surrogate detects the kinetics of O₂ mass transport at the microbial microenvironment created by PFC nanoemulsions. B & C, linear scan voltammograms depicting the current density (i) vs. electrode potential (E_Disk) on a Pt RDE of different spinning rates for oxygen reduction reaction in the absence (B) and presence (C) of 0.75% (v/v) PFC nanoemulsion. 1-bar O₂; 0.1 M sodium phosphate buffer (pH 7.0); 50 mv/dec; iR corrected; RHE, reversible hydrogen electrode. D, the Koutecký-Levich plot of |i_Disk|⁻¹ vs. ω ^−1/2 at 0.3 V vs. RDE that extracts the kinetic of O₂ delivery expressed as i_k. ω, angular rotating frequency of RDE.

The persistency of the bacterial microenvironment created by biocompatible PFC nanoemulsions and the associated enhancement of O₂ availability presents PFC nanoemulsion a suitable candidate to tackle the issue of O₂ biocompatibility without consuming additional electrons as the cases reported previously^10,19. In a single-chamber hybrid system without an external O₂ supply, the introduction of PFC nanoemulsions presumably enable the O₂ generated from electrochemical water oxidation alone to be sufficient to satisfy the O₂-dependent N₂/CO₂ fixation in X. autotrophicus, by facilitating the mass transport and availability of O₂ in the bacterial microenvironment (Figure 1A). We strive to test this working hypothesis and one preparatory step in our investigation is to ensure that the introduction of PFC nanoemulsions does not adversely affect the electrochemical water-splitting process in the hybrid. Biocompatible cobalt phosphate (CoP_i) loaded on carbon cloth and cobalt-phosphorous alloy (CoP) loaded on stainless steel mesh, whose water-splitting mechanisms have been studied before,^41,42 were prepared following previous literature for electrochemical generation of O₂ and H₂, respectively (See experimental procedures)^8,9,12. PFC nanoemulsions of 0%, 1.25%, 2.5% and 3.75% (v/v) were added into minimal medium without any nitrogen and organic carbon along with X. autotrophicus inoculations (See experimental procedures) (Figure 4A). Multi-step choronoamperometry yields near-identical i-V curves among all four conditions (Figure 4B). It shows that the constitution of PFC nanoemulsions used in this work does not have a significant influence on the catalytic activity of the electrodes over the range of applied potentials. It also hints that the accelerated mass transport enabled by nanoemulsion for gas-consuming reaction does not alter the kinetics of gas-generating water-splitting reaction. Electrochemical impedance spectroscopy was also conducted in the hybrid system and the Nyquist plots at all four conditions are similar to each other as displayed in Figure 4C. The derived values of series resistance R form such measurements, indicative of solution’s electrical conductivity, remain unchanged (inset in Figure 4C). Those electrochemical measurements reveal that PFC nanoemulsions, within the percentage range used in this work, do not have significant impacts on the electrochemical water splitting reaction in our hybrid system. We further tested whether the introduction of PFC nanoemulsions affects the growth of X. autotrophicus. X. autotrophicus were cultured in minimal medium with 0 %, 1.25 %, 2.5 % and 3.75 % (v/v) PFC nanoemulsions under a 1-bar gas mixture of 2% O₂, 60% N₂, 20% H₂ and 18% CO₂ (See experimental procedures). Similar growth rates of X. autotrophicus, determined from the changes of total nitrogen values in a 5-day period, were observed among 0 %, 1.25% and 2.5% PFC, while a slightly lower growth rate was observed for microbes in 3.75% PFC nanoemulsions (Figure 4D). The absence of any significant growth enhancement under a O₂-containing gas environment, different from the observed enhancement for S. ovata in H₂,¹² indirectly suggests that the similarly enhanced mass transport of H₂ should not contribute significantly in our current system. The similar growth rates of X. autotrophicus in the presence of PFC nanoemulsions suggest the biocompatibility of PFC nanoemulsions up to 3.75% (v/v).

Figure 4. O₂-dependent microbial N₂/CO₂ fixation in a O₂-free headspace driven by electricity.

A, Reactor schematic of the biological | inorganic hybrid in a O₂-free headspace for N₂/CO₂ fixation. B & C, The relationships between current density (i) and applied potential (E_appl) (B), and representative Nyquist plots from electrochemical impedance spectroscopy (C), between CoP and CoPi electrodes for reactors of varying concentrations of PFC nanoemulsions. Inset in C, series resistance (R) of the reactor determined from the Nyquist plots (n = 8). D, The relative growth rate of X. autotrophicus in mineral medium during a 5-day gas fermentation under varying concentrations of PFC nanoemulsions. n = 4. E to G, the amount of fixed nitrogen (E), the corresponding overall electron efficiency of nitrogen reduction reaction (NRR) (F), and the percentage of dead X. autotrophicus (X. a.) (G), after a 5-day reaction under varying concentrations of PFC nanoemulsions. 1-bar O₂-free atmosphere of 20% CO₂ and 80% N₂; E_appl = 3.0 V; 30 °C; n = 4. Data are represented as mean ± SEM.

Experimental results support our working hypothesis that the nanoemulsion-induced microenvironment leads to full utilization of electrochemically generated O₂, enhanced cell viability, and a boosted efficiency for electricity-driven microbial N₂/CO₂ fixation. In a single-chamber, two-electrode reactor at 30 °C, a O₂-free gas environment was maintained with 20% CO₂ and 80% N₂ as the sole carbon and nitrogen source for X. autotrophicus. Chronoamperometry at E_appl = 3.0 V was conducted for 5 days and the resultant yields of CO₂ and N₂ fixation were quantified based on appropriate calibrations (Figure S4) following previously published procedure (See experimental procedures)^12,19,43,44. Among the experiments with different PFC nanoemulsion contents (0, 1.25, 2.50, and 3.75 %, v/v), a maximal yield of N₂ fixation was observed with a concentration of 2.50% PFC nanoemulsion (Figure 4E). The amount of fixed total nitrogen tops at 52±13 mg/L, i.e. ~ 3 mM (n = 4). The resultant overall electron efficiency that calculates the percentage of electric charge towards N₂ fixation, in addition to the roughly 21 % overall electron efficiency of CO₂ fixation (See experimental procedures)²¹, reaches 3.07±0.49 % (n = 4) and is about 250 % higher than the control case when no PFC nanoemulsions were introduced (0 % v/v) (Figure 4F). In those experiments, along with free nanoemulsions there are bacterial-bound PFC nanoemulsions of a fully surrounded microenvironment due to the high binding constant K_sp (vide supra). Since previous literature^25–27 and our own experience of culturing X. autotrophicus suggest minimal microbial growth when O₂ is depleted, our observed N₂ fixation of the hybrid system in an O₂-free headspace is in accordance with the hypothesis that the created microenvironment enables an O₂-dependent biocatalytic N₂/CO₂ fixation in an O₂-free gas headspace, by fully utilizing the electrochemically generated O₂ from water oxidation via the enhanced O₂ transport kinetic. Experimental evidence that further supports our argument can be obtained from the percentage of bacterial cells remaining viable during the electricity-driven process of N₂/CO₂ fixation. Aliquots of bacterial cultures were sampled at the day 4 of electrolysis and the viabilities of those bacteria were determined by flow cytometry with a bacterial Live/Dead staining kit (See experimental procedures, Figure S5). About 30 % of X. autotrophicus were dead without the addition of nanoemulsion (0% PFC in Figure 4G), while only 3% of dead bacteria were detected with 2.75% nanoemulsion (2.75% PFC in Figure 4G). The latter value in a presumably unfit O₂-free headspace is comparable with the < 5% observed in healthy cultures grown under a previously reported microaerobic atmosphere of 2% O₂⁹. While the percentage of dead cells did not fully capture the fitness and metabolic status of the microbes due to the presence of stationary cells, the increased cell viability (Figure 4G) and N₂ fixation efficiency (Figure 4F) corroborate that the increased O₂ mass transport induced by nanoemulsion addition enhanced ATP generation, the production of reducing equivalents, and subsequently N₂ fixation with concurrent cell growth. Our results shows that the creation of nanoemulsion-based microenvironment and the subsequently enhanced O₂ transport from electrochemical water oxidation preserves the viability of O₂-demanding X. autotrophicus without external gas supply of O₂.

An excessive amount of PFC nanoemulsion may be detrimental towards the biological | inorganic hybrid despite the created microenvironment. Cell viability assay indicates that excessively high concentration of PFC nanoemulsion (3.75% PFC in Figure 4G) leads to about 35% of dead bacteria in the culture, which is much higher than the 3% of dead microbes observed with 2.75% loading of PFC nanoemulsions. This finding is in agreement with the biocompatibility test results, where X. autotrophicus cultured in 3.75% PFC are showing slower growth than those under conditions with lower PFC percentage (Figure 3D). Such a decrease of cell viability is in accordance with the observed decrease for fixed nitrogen amount and the overall electron efficiency of N₂ fixation (Figure 4E and 4F). One interpretation to such results is that excessively high PFC concentration will lead to a high concentration of intracellular O₂, which benefits cellular ATP and energy production but impedes the activity of bacterial N₂ fixation. Therefore, while a suitable dose of PFC nanoemulsion creates a beneficial microenvironment for bacterial survival, other detrimental factors will overwrite such a benefit under excessively high PFC loading amount. Advanced design of the PFC nanoemulsions with more biocompatible surfactants or PFC composition^34,45 will help to mitigate the toxicity issue at high nanoemulsion concentrations. As the liquid culture of X. autotrophicus themselves were considered as an “organic” version of biofertilizers for crop growth,⁹ our knowledge obtained here will guide future general designs for biological | inorganic hybrid with the production of organic fertilizers.

CONCLUSION

In addition to ensuring biocompatibility, biological | inorganic hybrid offers a great opportunity of expanding the operation conditions for biocatalysis with the advanced designs in materials. In this work, we tackled on one ubiquitous issue: the mismatch of the O₂ demand between the native requirement in biocatalysis and the constrains in practical application. We demonstrate the feasibility of O₂-depenent microbial N₂/CO₂ fixation in a O₂-free headspace, by employing PFC nanoemulsions to maximize the use of electrochemically generated O₂, a typically deemed “waste” in electricity-driven N₂/CO₂ fixation. Given the various application situations and the multitude of intertwining among anaerobic and aerobic pathways in biochemistry⁴⁶, our design is applicable to other biochemical systems whenever the O₂ incompatibility arises in a biological | inorganic hybrid, thanks to PFC’s biocompatibility and the strong, non-specific binding of nanoemulsions. The characterizations presented here offer a microscopic and quantitative picture of the created extracellular microenvironment, which will be beneficial for future development and optimization in other applicable scenarios. The strategy of creating a microbial microenvironment with nanomaterials despite an unwelcoming macroscopic environment offer a viable solution to resolve incompatibility and create synergy at the materials-biology interface.

EXPERIMENTAL PROCEDURES

Resource Availability

Lead Contact

Further information and requests for the resources and reagents should be directed to and will be fulfilled by the lead contact, Chong Liu (chongliu@chem.ucla.edu)

Materials Availability

This study did not generate unique reagents.

Data and Code Availability

Full experimental procedures and experimental data are provided in the Supplementary Information.

Materials and chemicals

All chemicals were purchased from Thermo Fisher, Sigma–Aldrich, or VWR International, unless otherwise stated. Perfluorocarbons (PFCs) were purchased from SynQuest Laboratories. All deionized (DI) water was obtained from a Millipore Millipak® Express 40 system. Stainless steel (SS) mesh was purchased from Alfa Aesar and carbon cloth (CC) was purchased from Fuel Cell Earth. Electrochemical supplies noted here were purchased from CH Instruments, inc.. The LIVE/DEAD™ BacLight™ Bacterial Viability and Counting Kit and the included SYTO™ 9 dye were purchased from ThermoFisher (L34856). 2-μm polystyrene microspheres with surface-functionalized with –NH₂ and –COOH moieties were purchased from Sigma–Aldrich (L0280 and L4530, respectively). The materials and supplies for biomass assay were purchased from Hach Company.

Preparation and characterization of PFC nanoemulsions and electrodes

Nanoemulsions were prepared via previously reported procedure³⁴. Surfactant solution was prepared by combining 2.8 wt.% of surfactant, Pluronic F68, with the relevant buffer and sonicating in a bath sonicator, Branson 3800 ultrasonic cleaner, to thoroughly dissolve the polymer. Perfluorodecalin, PFD, and perfluorohexane, PFH, were combined (450μL each) in a 15-mL centrifuge tube. The relevant surfactant-containing buffer was then added to achieve a total final volume of 10 mL. All-inorganic microbial minimal medium, detailed in Supplemental Information section 1.1, was used as the buffer for all experiments, except rotating disk electrode experiments where a phosphate buffer (pH 7.0, 0.1 M sodium phosphate), was used instead. The PFCs and surfactant-containing buffer were then sonicated at 35% amplitude for 5 mins using a Qsonica point sonicator. Fluorescently tagged PFC nanoemulsion was prepared for microscopic and flow cytometry experiments. Fluorous rhodamine was synthesized as previously reported. Fluorous rhodamine was stabilized in 4 μL methoxyperfluorobutane, and 8 μL of both PFH and PFD (a total of 16 μL). This mixture was then combined with 200 μL of minimal medium, and point-sonicated at 35% amplitude for 90 sec. Procedures were scaled up when larger volumes were needed. Dynamic light scattering (DLS) experiments were conducted with a Malvern Zetasizer Nano ZSP instrument. 20 μL PFC nanomulsions were diluted in 2 mL DI water in a plastic cuvette. A total of 3 measurements, each consisting of 10 scans, were run to determine the average size of the nanoemulsions (~240 nm, see Figure S1).

The CoP_i and Co-P alloy electrodes were both prepared via electrochemical deposition using a Gamry Interface 1000 potentiostat, following previously reported procedures⁸. The electrochemical deposition was run using a three-electrode system consisting of a working electrode of the substrate to be electrodeposited, a Pt counter electrode, and a Ag/AgCl reference electrode (1 M KCl). The Co-P alloy catalyst was prepared via cathodic electrochemical deposition onto SS mesh. The deposition solution consists of 0.15 M H₃BO₃, 0.1 M NaCl, 0.33 M Na₂H₂PO₂, and 0.2 M CoCl₂. The SS mesh was sequentially cleaned with acetone, isopropanol, and then soaked in DI water before electrodeposition. To deposit the Co-P alloy catalyst, the SS mesh as the substrate underwent chronocoulometry for 15 min at an applied voltage of −1.5 V vs. reference. The CoP_i electrode was prepared using a deposition solution containing 10 mM Co(NO₃)₂ and 0.1 M methylphosphonate (MeP_i) buffer of pH 8. CoP_i catalyst loaded on CC substrate was deposited via chronocoulometry at 0.85 V vs. reference until 500 mC cm⁻² of charge was passed.

Protocols of microbial culturing

The freeze-dried samples of aerobic bacterium Xanthobacter autotrophicus (ATCC 35674) and anaerobic bacterium Sporomosa ovata (ATCC 35899) were purchased from American Type Culture Collection (ATCC)⁴⁷. Detailed recipes of culture media listed below are provided in Supplemental Information section 1^12,19. Yellow colonies of X. autotrophicus were selected from succinate agar plates (succinate nutrient broth solidified with 1.5% agar), incubated at 30 °C in succinate nutrient broth, and stored at −80 °C in a mixture of glycerol and succinate nutrient broth (20/80, v/v). Cultures of S. ovata were obtained at 34°C under strict anaerobic condition with DSMZ 311 medium under H₂/CO₂ (80/20) atmosphere, and stored at −80 °C in a mixture of dimethyl sulfoxide and DSMZ 311 medium (20/80, v/v).

X. autotrophicus reported in this work were first grown at 30 °C in succinate nutrient broth for 1 day, collected by centrifugation (6000 rpm, 5 min; Sorvall ST8, Fisher Scientific) after adjusting the culture pH to about 12 with NaOH, and cultured autotrophically at 30 °C in the all-inorganic minimal medium with an anaerobic jar (Vacu-Quick Jar System, Almore) under a 1-bar gas mixture of 2% O₂, 60% N₂, 20% H₂ and 18% CO₂ (200 rpm stirring)^9,19,26. The microbial culture started at OD₆₀₀ ~ 0.2 (optical density at 600 nm) and reached OD₆₀₀ ~ 1 within 5 days under a condition of N₂ and CO₂ fixation. The bacteria were harvested via centrifugation (6000 rpm, 5 min) before experiments. S. ovata were strictly anaerobically cultured at 34°C in DSMZ 311 medium under H₂/CO₂ (80/20) atmosphere for a 3-days autotrophic growth before use.

Colocalization experiment with fluorescent confocal microscopy

Cultures of X. autotrophicus (OD₆₀₀ = 1.0) were harvested and re-suspended with 0.85% NaCl solution with OD₆₀₀ adjusted to 0.1. Each 1 mL of the resulted bacterial suspension was incubated in hard at room temperature for 15 mins with 1.5 μL of microbial-binding SYTO™ 9 dye solution from the LIVE/DEAD™ BacLight™ Bacterial Viability and Counting Kit. The fluorescently tagged X. autotrophicus was separated via centrifugation (6000 rpm, 5 min) and re-suspended in 1 mL minimal medium containing 2.5% fluorescently tagged PFC nanoemulsion (vide supra) ³⁵. Suspension of X. autotrophicus without the addition of nanoemulsion was prepared in parallel as the control sample. The prepared samples incubated in dark for 1 hr for completion of nanoemulsion binding and loaded to a 35-mm glass bottom dish (μ-dish, ibidi), whose bottom glass was coated with a layer of poly-l-lysine (treated with 0.01% poly-l-lysine solution overnight and dried). The mixture was allowed to sit in the dish for 0.5 hr before all liquid was slowly removed by pipetting. The glass-surface of the dish was gently washed 5 times with filtered microbial minimal medium. Last, 1 mL of minimal medium was added to the dish to keep the sample hydrated before imaging.

Experiments of confocal microcopy (Leica Confocal SP8 MP) was conducted at Advanced Light Microscopy and Spectroscopy Laboratory at California Nanoscience Institute, UCLA. The data was acquired using Leica Application Suite X (LASX) on x-y mode at a scanning resolution of 14.6 nm per pixel, taking x-y cross-sectional images with a 100× oil objective lens (Leica 100× HC PL APO OIL CS2 NA/1.4). Fluorescence from SYTO™ 9 in the microbes was monitored at 490nm~520nm by a 470-nm laser excitation; the fluorescence from fluorous rhodamine in PFC nanoemulsions was monitored at 580nm~650nm by a 550-nm laser excitation. The intensities of fluorescence emissions were collected by photon multiplier tube (PMT) detectors. We note that the contribution of out-of-focus signals for SYTO, much brighter than rhodamine in those experiments, is larger than the one from rhodamine dye in the nanoemulsion. Subsequently, the imaged PFC nanoemulsion is more concentrated within the microscopy’s focal plane as compared to the imaged bacteria. The fluorescence images of microbes and PFC nanoemulsions were taken separately and merged as shown in Figure 1C, 1D and S2.

Experiments of flow cytometry related to microbes and PFC nanoemulsions

Experiments of flow cytometry (BD LSR II cytometer) were conducted at the Janis V. Giorgi Flow Cytometry Core, UCLA. The flow cytometer was operated under a slow flow rate setting. Forward Scattering (FSC) was used as the threshold for events. For each sample and control, 10,000 events were collected and recorded. The recorded events were plotted as 2-D scatter plots of FSC and Side Scattering (SSC) for the gating of microbes, particles, and nanoemulsions. Data were analyzed using FlowJo ver. 10 and the gating represents >95% of the clustered events.

The following procedures were applied to quantify the binding between microbes and nanoemulsion. Cultures of X. autotrophicus and S. ovata (OD₆₀₀ ~ 1.0) were harvested and re-suspended in 200-nm filtered minimal medium (OD₆₀₀ = 0.002). Solution of Fluorescently tagged PFC nanoemulsions (10 %, v/v)³⁴ was serial diluted by dilution ratios ranging between 10³ and 10⁶. The diluted solution was mixed with equal volume of microbes to reach a final microbial OD₆₀₀ = 0.001, incubated for 30 mins in dark for the completion of nanoemulsion binding, and tested directly by a flow cytometer without further dilution. Control samples were prepared and tested similarly by mixing solutions of diluted nanoemulsions and filtered minimal medium. A pre-defined fluorescence measurement channel for PE (phycoerythrin) was used to measure the fluorescence intensity of fluorinated rhodamine. When 2-μm polystyrene microspheres surface-functionalized with –NH₂ and –COOH moieties were tested, filtered microsphere solutions whose particle density was about 2×10⁶ mL⁻¹ were used in lieu of the microbial suspension of OD₆₀₀ = 0.002. The average number of PFC nanoemulsions bound to a bacterial cell or microsphere (N) for a specific PFC concentration was determined based on the mean fluorescent intensity with the following equation:

$$N=\frac{I_{mix}-I_{bg}}{I_{PFC}}$$ (3)

Here I_mix and I_PFC are the mean emission intensities of fluorous rhodamine for microbe-nanoemulsion complex and individual PFC nanoemulsion, respectively, at a specific PFC concentration. I_bg is mean background emission intensity for the microbe-nanoemulsion complex.

The following procedures were applied to quantify viability of X. autotrophicus in different concentrations of PFC nanoemulsions. Samples of X. autotrophicus were harvested via centrifugation (6000 rpm, 5 min), and re-suspended in 1 mL of 0.85% NaCl. The samples were then added with 1.5 μL of SYTO™ 9 dye solution and 1.5 μL of propidium iodide solution from the LIVE/DEAD™ BacLight™ Bacterial Viability and Counting Kit, following the instructions from the manufacturer. After a 15-min incubation in dark, the samples were tested by a flow cytometer after a dilution of about 20~50 times. Two pre-defined fluorescence measurement channels, one for AlexaFluor-488 and the other for PE (Phycoerythrin), were used to measure the fluorescence intensity of SYTO ™ 9 and propidium iodide, respectively. The gating of live and dead cells (>95% events) were determined from sample actively grown under N₂/CO₂-fixing condition in a anaerobic jar (live reference) and the one treated with pure ethanol for 1 hr before staining (dead reference). Results of dead and live reference samples were shown in Figure S5.

Biocompatibility test of X. autotrophicus with nanoemulsions

For each set of tests, four 20-mL glass screw top vials were autoclaved with a small stir bar. Microbial culture, PFC nanoemulsions, and minimal medium were combined to achieve a starting volume of 8 mL total. The volume of culture used was adjusted from OD₆₀₀ after resuspension in the minimal medium to give a final OD₆₀₀ of 0.2 in the experiment. PFC nanoemulsion was added to achieve final 1.25, 2.5 and 3.75 % PFC nanoemulsion in the 8 mL, respectively. The total volume was then adjusted to 8 mL with the minimal medium. All samples were cultured autotrophically at 30 °C in the all-inorganic minimal medium with an anaerobic jar (Vacu-Quick Jar System, Almore) under a 1-bar gas mixture of 2% O₂, 60% N₂, 20% H₂ and 18% CO₂. One set of viability tests, which consisted of 4 20-mL vials, one for each percent PFC nanoemulsion (0, 1.25, 2.5 and 3.75 %) were placed in an aforementioned air-tight jar and incubated under N₂ fixation condition for 5 days. 1 mL of sample was taken from each vial and frozen right before the incubation starts and right after the incubation ends. The total nitrogen contents of the samples were analyzed as described in “quantification of total nitrogen”. The increase of total nitrogen content after 5 days’ incubation compared to the total nitrogen content before incubation was considered as the relative growth of microbes.

Electrochemical characterizations and the setup of biological | inorganic hybrid

The experiments of rotating disk electrode (RDE) consisted of a Gamry Interface 1000E potentiostat, an MSR rotator (Pine Research Instrumentation), and an electrochemical glass cell under a controlled gas environment. The electrochemical cell was a 150 mL, five-neck, flat bottom glass flask. The temperature of the setup was maintained at 30 °C by a circulating water bath. A platinum RDE disc electrode of 5.0 mm disk dimeter (AFE5T050PT, Pine Research Instrumentation) was used as working electrode, a Ag/AgCl (1M KCl) as reference electrode and a Pt wire as counter electrode. Prior to experiments, the working electrode was polished, sonicated in DI water, and air dried. The Pt counter electrode was cleaned with dilute HNO₃ then DI water. For every measurement, 120 mL fresh electrolyte was saturated with a stream of 13 seem N₂ or O₂ for at least 20 min. Linear scanning voltammograms were obtained at a controlled spin rates The interval between each measurement was 1 min. The scanning rate was 50 mV s⁻¹ and the iR-corrected results were displayed in Figure 3B and 3C.

The experiments of biological | inorganic hybrids were conducted in a 250-mL GL45 glass bottles using modified bottle caps with four 1/4-28 ports (Western Analytics). A two-electrode setup was established in which the working electrode is a CC electrode loaded with CoPi catalyst and the reference/counter end is a SS mesh loaded with Co-P alloy catalyst. The total liquid volume was 100 mL, and the electrolyte is a mixture of microbes, minimal medium, and PFC nanoemulsion, with a starting OD₆₀₀ of 0.2 for microbes and the appropriate PFC percentage for the given experiment. A maximum of 8 electrochemical experiments were run in parallel using a Gamry Interface 1000E potentiostat coupled with a Gamry ECMB multiplexer. Steady-state i-V correlations were established with multi-step chronoamperometry. The applied potential (₁E_appl) ranges from 1.3 to 3.0 V with at an interval of 0.05 V. At each voltage step, the corresponding E_appl was maintained for 2 min. Electrochemical impedance spectroscopy (EIS) were measured at open circuit potential with an AC amplitude of 50 mV. Measurements were performed at frequencies between 100 kHz and 10 Hz with 10 data points per decade. Nyquist plots were plotted and used to extract the approximated series resistances (R) of reactors at day 0 of experiments under different conditions. The hybrid operates at a constant E_appl for 5 days under a O₂-free headspace gas environment that only contains N₂ and CO₂ (80/20) for microbial fixation of N₂ and CO₂. Liquid cultures that include both microbial suspension and soluble chemicals were sampled and frozen right before and after the bulk electrolysis for further analysis. Samples for cell viability determination under different concentrations of PFC nanoemulsion were collected on the 4^th day of the experiment.

Performance of biological | inorganic hybrid

The total nitrogen and carbon in both the biomass and liquid solution was analyzed using commercial assay kit from Hach Company. Total nitrogen was performed using the Hach Company total nitrogen reagent kit, Test ‘N Tube (Product #2714100), following the provided procedural instructions (DOC316.53.01085) by observing the absorbance of 410 nm at the end of protocol. As used previously for electricity-driven microbial N₂ fixation,⁹ this measurement protocol allows the quantification of nitrogen content from all nitrogen-containing compounds notwithstanding N₂, which are either dissolved in the liquid solution or accumulated in the biomass. The total nitrogen content was then quantified using calibration curves correlating the absorbance at 410 nm and total nitrogen concentration, constructed using the kit on different standard nitrogen sample solutions (ammonium p-toluenesulfonate, Hach). For each of the different percentage of PFC nanoemulsion (0%, 1.25%, 2.5% and 3.75% in minimal medium) a calibration curve was made using the corresponding PFC nanoemulsion as solvent to prepare standard solutions (Figure S4). The absorbance was measured using Agilent Cary 60 UV-Vis spectrophotometer.

The overall electron efficiency (E. E.) of N₂ reduction reaction in each experiment was determined using the following formula^9,19:

$$E.E.=\frac{\frac{[\left(\Delta N_{\mathrm{TNT}}\left(\mathrm{mg}\right)\right)}{1000\times 14\text{g}{\text{ mol}}^{-1}}]\times V_{\text{solution}}\left(L\right)\times 3\times F\left(\text{C}{\text{ mol}}^{-1}\right)}{\mathit{overall\; charge}\left(\text{C}\right)}\times 100\%$$ (4)

Here ΔN_TNT (mg L⁻¹) is the change of nitrogen content from day 0 to day 5 in the experiment; V_solution the total volume of solution; F the Faraday constant, and Overall charge is the total electric charges passed through the cathodic chamber. The number 3 is the number of electrons required to reduce one nitrogen atom in N₂ reduction reaction. We note that our calculation did not consider the possible loss of electrochemically generated H₂ along with other possible side reactions. Therefore, the calculated values are a lower-bound estimate for the N₂ fixation capablity in our system.

The overall electron efficiency (E. E.) of CO₂ reduction reaction in each experiment was estimated based on previous C/N ration of X. autotrophicus (1.5 mg N cell⁻¹ vs. 6.4 mg C cell⁻¹, i.e. C/N ~ 4.27)²¹, assuming all fixed carbon and nitrogen forms biomass based on the following formula^9,19:

$$E.E.=\frac{\frac{[\left(\Delta N_{\mathrm{TNT}}\left(\mathrm{mg}\right)\times 4.3\right)}{1000\times 12\text{g}{\text{ mol}}^{-1}}]\times V_{\text{solution}}\left(L\right)\times 4\times F\left(\text{C}{\text{ mol}}^{-1}\right)}{\mathit{overall\; charge}\left(\text{C}\right)}\times 100\%$$ (5)

Here ΔN_TNT (mg L⁻¹) is the change of nitrogen content from day 0 to day 5 of the experiment; V_solution the total volume of solution; F the Faraday constant, and Overall charge is the total electric charges passed through the cathodic chamber. The number 4 is the number of electrons required to reduce one carbon atom in CO₂ reduction reaction.

HIGHLIGHTS.

Electricity-driven N₂ fixation in aerobic microbes without external O₂ supply;

Perfluorocarbon nanoemulsion creates desirable O₂ microenvironment for microbes;

Optimal nanoemulsion concentration offers good efficiency and biocompatibility.

BIGGER PICTURE.

A successful deployment of electrical energy to drive microbial catalysis demands a biocompatible electrochemical system that does not interfere with the desired biochemical transformations. As O₂ is one of the most ubiquitous molecules in biology, balancing the O₂ environment near the microorganisms is critical to ensure proper functioning of biochemical pathways despite the perturbation from electrochemistry and the environmental requirements at operating conditions. Here we report that the strategy of introducing perfluorocarbon nanoemulsions to microbial catalysts creates a beneficial O₂ environment near microorganisms and facilitates electricity-driven N₂ fixation into organic fertilizers. The strategy of fine-tuning the local microbial environment with nanomaterials offer a viable approach to achieve harmony among biotic and abiotic components in chemical transformation.