Ambient nitrogen reduction cycle using a hybrid inorganic–biological system

Significance

The nitrogen cycle and the fixation of atmospheric N₂ into ammonium are crucial to global food production. The industrial Haber–Bosch process facilitates half the global nitrogen fixation in the form of ammonia but it is energy- and resource-intensive, using natural gas as the source of energy and hydrogen at elevated temperature and pressure. Our alternative approach synthesizes ammonium from N₂ and H₂O at ambient conditions powered by water splitting, which may be driven renewably. The inorganic–biological hybrid system fixes atmospheric nitrogen into NH₃ or soluble biomass with high fluxes and energy efficiency. Simultaneously, this system cultivates a living soil bacterium that acts as a potent biofertilizer amenable to boosting crop yields.

Abstract

We demonstrate the synthesis of NH₃ from N₂ and H₂O at ambient conditions in a single reactor by coupling hydrogen generation from catalytic water splitting to a H₂-oxidizing bacterium Xanthobacter autotrophicus, which performs N₂ and CO₂ reduction to solid biomass. Living cells of X. autotrophicus may be directly applied as a biofertilizer to improve growth of radishes, a model crop plant, by up to ∼1,440% in terms of storage root mass. The NH₃ generated from nitrogenase (N₂ase) in X. autotrophicus can be diverted from biomass formation to an extracellular ammonia production with the addition of a glutamate synthetase inhibitor. The N₂ reduction reaction proceeds at a low driving force with a turnover number of 9 × 10⁹ cell^–1 and turnover frequency of 1.9 × 10⁴ s^–1⋅cell^–1 without the use of sacrificial chemical reagents or carbon feedstocks other than CO₂. This approach can be powered by renewable electricity, enabling the sustainable and selective production of ammonia and biofertilizers in a distributed manner.

The reduction of N₂ into NH₃ is essential for maintaining the global biogeochemical nitrogen (N) cycle (1). Fixed, organic N in food, biomass, and waste is eventually returned to the atmosphere as N₂ through biological denitrification. As a ubiquitous, synthetic nitrogenous fertilizer, NH₃ synthesized from atmospheric N₂ via the Haber–Bosch process has been added to agricultural soils to drive global increases in crop yields (2). Despite its high efficiency and scalability, the Haber–Bosch process unsustainably employs natural gas as a H₂ feedstock, operates at high temperatures and pressures, and relies on a significant infrastructure for NH₃ distribution (1). A distributed approach toward NH₃ synthesis from renewable energy sources at ambient conditions would enable on-site deployment and reduce CO₂ emissions. To this end, significant effort has been devoted to promoting the reduction of nitrogen to NH₃ with the use of transition metal catalysts (3–5), electrocatalysts (6), photocatalysts (7–11), purified nitrogenases (N₂ases) (11, 12), and heterotrophic diazotrophs (13, 14), potentially powered by renewable energy and operating at ambient conditions. Such approaches, however, typically use sacrificial reductants to drive conversion at low turnover or suffer poor selectivity.

More broadly, the limitations of synthetic NH₃ as a fertilizer have become apparent in recent years as decreasing efficiency of fertilizer use, coupled to environmental damage, has provided an imperative for the development of sustainable biofertilizers (15, 16). Soil microorganisms facilitate efficient nutrient uptake and recycling (17), pathogen resistance (18), environmental adaptation (19), and long-term soil productivity (15). However, the diminished yields of organic/sustainable agriculture have demonstrated that nutrient cycling alone, accentuated by natural variabilities in the soil microbiome, is insufficient to meet an increasing worldwide food demand (20). Attempts to establish robust, productive soil communities through microbial inocula have shown promise (21), but the limited natural flow of organic carbon into these soils results in a bottleneck in the biological activity of these largely heterotrophic biomes (22). An alternative solution would leverage the increasing abundance of renewable energy to cultivate and feed such soil microbiomes, effectively supplementing the natural process of microbial N₂ fixation and plant beneficial interactions.

To further the development of distributed fertilization and natural N cycling, we demonstrate the reduction of N₂ coupled to H₂O oxidation by interfacing biocompatible water-splitting catalysts with the growth of N₂-reducing, autotrophic, biofertilizing microorganisms in a single reactor (Fig. 1). This nitrogen cycle builds on our previous efforts that use H₂ from water splitting to power a bioengineered microorganism to fix CO₂ to biomass and liquid fuels at energy efficiencies far exceeding natural photosynthesis (23, 24). The biocompatible catalysts, a cobalt–phosphorus (Co–P) alloy for the hydrogen evolution reaction (HER) and an oxidic cobalt phosphate (CoP_i) for the oxygen evolution reaction (OER), permit low driving voltages (E_appl) under mild conditions (pH 7, 30 °C). The combination of these electrocatalysts with H₂-oxidizing microbes yields CO₂ reduction efficiencies (η_elec,CO2) up to ∼50% (24). The modular design of this renewable synthesis platform may be leveraged beyond fuel production, toward more complex reactions such as the nitrogen reduction reaction (NRR), as well as cultivation of living whole-cell biofertilizers depending on the specific synthetic capabilities of the microorganism. We exploit this design flexibility to perform the efficient synthesis of NH₃ from N₂ and H₂O by driving the NRR within the H₂-oxidizing, autotrophic microorganism Xanthobacter autotrophicus. This Gram-negative diazotrophic bacterium can use H₂ under microaerobic conditions (<5% O₂) as its sole energy source to fix CO₂ and N₂ into biomass (25). We further demonstrate that X. autotrophicus functions as a potent electrogenerated biofertilizer, increasing yields of radishes (Raphanus sativus L. var. “Cherry Belle”), a fast-growing model food crop.

Fig. 1.

Schematic of the electroaugmented nitrogen cycle. A constant voltage (E_appl) is applied between CoP_i OER and Co–P HER electrodes for water splitting. The H₂ases of X. autotrophicus oxidize the generated H₂, fueling CO₂ reduction in the Calvin cycle and N₂ fixation by N₂ases. The generated NH₃ is typically incorporated into biomass (pathway 1), but can also diffuse extracellularly by inhibiting biomass formation (pathway 2). This process is powered by renewable, solar-derived electricity taking N₂ and CO₂ from the environment. Cells of X. autotrophicus form an electrogenerated biofertilizer which may be added to soils to improve plant growth. The pathway of natural N cycling/N₂ fixation is indicated, with line width denoting relative flux of these pathways. Red pathways indicate carbon cycling; blue pathways indicate N cycling.

Results and Discussion

When interfaced with CoP_i | Co–P water-splitting catalysts (Fig. 1), X. autotrophicus accumulates fixed N derived from the NRR. An O₂/CO₂/N₂ gas mixture (2/20/78) was maintained in the single-chamber reactor housing the Co–P HER cathode and the CoP_i OER anode (SI Appendix, Fig. S1). At the beginning of each experiment, X. autotrophicus was inoculated into the organic-free, N-free minimal medium. A constant driving voltage (E_appl = 3.0 V) was applied to the CoP_i | Co–P catalyst system, and aliquots from the reactor were periodically sampled for the quantification of biomass (optical density at 600 nm, OD₆₀₀) as well as fixed N (detected by two colorimetric assays, SI Appendix, Fig. S2A). The H₂ generated from water splitting provides the biological energy supply for X. autotrophicus to perform the NRR, as well as CO₂ reduction, into biomass without the need for sacrificial reagents (Fig. 2). The amount of faradaic charge passed into water splitting was proportional to biomass accumulation (OD₆₀₀) as well as the total N content in the medium (N_total) during 5-d experiments (Fig. 2A). No biofilm formation was observed on either electrode. The fixed N was assimilated into biomass as evidenced by no change in the extracellular soluble N content (N_soluble). Over the course of the experiment, 72 ± 5 mg L^–1 of N_total, as well as 553 ± 51 mg L^–1 of dry cell weight accumulated (n = 3, entry 1 in Fig. 2B and SI Appendix, Table S1). In contrast, no accumulation of N_total is observed in controls that omit one of the following elements in our experiment: H₂ from water splitting, X. autotrophicus, a single-chamber reactor, or a microaerobic environment (entry 2–5 in Fig. 2B and SI Appendix, Table S1). The small increases in OD₆₀₀ for entries 4 and 5 in Fig. 2B are likely due to light scattering from the accumulation of poly(3-hydroxybutyrate) (PHB). In the dual-chamber experiment where cathode and anode are separated by an anion-exchange membrane (AEM) (entry 4 in Fig. 2B), the absence of N_total accumulation is concurrent with an increase of soluble Co²⁺ (as determined by inductively coupled plasma mass spectroscopy, ICP-MS) in the medium from 0.9 ± 0.2 μM to 40 ± 6 μM over the course of 24 h, which approaches the ∼50-μM half-maximum inhibitory concentration (IC₅₀) of X. autotrophicus (SI Appendix, Fig. S2B). As the AEM is not permeable to cations including Co²⁺, its use prevents the redeposition of adventitious Co²⁺ onto the CoP_i anode via self-healing, as we have previously described (24, 26), emphasizing the importance of biocompatibility in our system.

Fig. 2.

N₂ reduction on the CoP_i | Co–P | X. autotrophicus hybrid system. (A) OD₆₀₀, the amount of charge passed through, the concentration of total N content (N_total), and soluble N content (N_soluble) are plotted versus the duration of experiments. n ≥ 3; error bars denote SEM. (B) Change of N_total and OD₆₀₀ under different experimental conditions in 5-d experiments (SI Appendix, Table S1). “No AEM” indicates a single-chamber reaction without an anion-exchange membrane. *, not applicable because no bacteria were introduced. n ≥3; error bars denote SEM. (C) A qualitative gas chromatography comparison of the whole-cell acetylene reduction with 100-ppm standard sample. t, incubation time after C₂H₂ injection. (D) Linear scan voltammetry (line, 10 mV s^–1) and chronoamperometry (circle, 30-min average) of Co–P HER cathode in X. autotrophicus medium, iR corrected. The thermodynamic values of HER and NRR (E_HER, E_NRR) are displayed. (E) Contributions of voltage drops within the applied E_appl = 3.0 V, as calculated SI Appendix. η_HER and η_OER, overpotentials of HER and OER. a.u., arbitrary units.

Whole-cell acetylene (C₂H₂) reduction (Fig. 2C) demonstrated high N₂ase activity in X. autotrophicus grown in our hybrid system. This assay provides a measure of the NRR turnover number (TON) by using C₂H₂ as an isoelectronic surrogate of N₂ (25, 27). Aliquots of X. autotrophicus sampled directly from our hybrid system reduced C₂H₂ exclusively to C₂H₄ at a rate of 127 ± 33 µM h^–1 OD₆₀₀^–1 (n = 3), when the whole-cell aliquots were incubated under a C₂H₂/O₂/H₂/CO₂/Ar gas environment (10/2/9/9/70). The measured C₂H₂-reduction activity corresponds to a rate of N₂ fixation at ∼12 mg N_total L^–1⋅d^–1 for cultures of OD₆₀₀ = 1.0, consistent with the measured N_total accumulation during the 5-d experiments (entry 1 in Fig. 2B and SI Appendix, Table S1). Combining a measured cell density of 3.8 ± 0.7 × 10⁸ cells mL^–1, this C₂H₂-reduction activity translates to an NRR turnover frequency (TOF) of 1.9 × 10⁴ s^–1 per bacterial cell, or roughly ∼4 s^–1⋅N₂ase^–1, assuming a N₂ase copy number of about 5,000 (28). The similarity between the measured NRR TOF per N₂ase and the values reported in previous studies (12, 27, 29) indicated that the N₂ases are fully functional in our experiments, highlighting the biocompatibility of the water-splitting catalysts in the CoP_i | Co–P | X. autotrophicus hybrid system. The equivalent TON is ∼9 × 10⁹ bacterial cell^–1 and 2 × 10⁶ N₂ase^–1, at least two orders of magnitude higher than that previously reported for synthetic (3–5, 7) and biological (11, 29) catalysts (SI Appendix, Table S2).

This hybrid system displays high faradaic efficiency for the NRR. A current of 10–12 mA for 100 mL of X. autotrophicus culture was needed to maintain sufficient microbial growth. This current was achieved with a cell potential of E_appl = 3.0 V (SI Appendix, Fig. S2C); this high driving voltage is due to the dilute medium salinity as X. autotrophicus is sensitive to high salt concentrations (25). The contributions of the HER and the OER overpotentials were determined by examining these half-reactions in the low salinity culture medium. Fig. 2D shows the I–V characteristics of the Co–P cathode for the HER vs. the normal hydrogen electrode (NHE). A 12-mA current is achieved for a driving voltage of –0.72 V vs. NHE or at a 0.3-V HER overpotential (η_HER) (E_HER = –0.41 V at pH = 7). Following a similar procedure for the OER half-reaction, we determined that the OER overpotential (η_OER) is 0.62 V. As summarized in Fig. 2E, with the measured overpotentials, the contribution of ohmic resistance due to low ionic conductivity is ∼28% of E_appl (∼0.85 V). Given the standard potential difference between the HER and the NRR (E_HER = −0.28 V), the driving force for the NRR by X. autotrophicus is ∼0.43 V, which is lower than the previously reported values that are summarized in SI Appendix, Table S2 (3–5, 29, 30). Based on the overall cell potential of 3.0 V, the energy efficiency of NRR (η_elec,NRR) is 1.8 ± 0.3% (n = 3) for a 5-d experiment (SI Appendix, Table S1). Parallel to NRR, X. autotrophicus fixes CO₂ into cellular biomass (Fig. 2A) at a measured η_elec,CO2 = 11.6 ± 1.9% (n = 3). The theoretical value of η_elec,NRR at E_appl = 3.0 V is 7.5 ∼11.7% based on the reaction stoichiometry of N₂ase and upstream biochemical pathways. Thus, our observation of η_elec,NRR = 1.8 ± 0.3% is 15 ∼23% of the theoretical η_elec,NRR. The calculated NRR faradaic efficiency is 4.5%, higher than the faradaic efficiencies or quantum yields of other reported NRR systems (6–11) at ambient conditions (SI Appendix, Table S2).

Modeling shows that linear microbial growth may be achieved by controlling the H₂ concentration relative to the Michaelis constant of H₂ase (24). Fig. 2A shows that linear growth conditions may be achieved for X. autotrophicus by balancing the H₂ produced from water splitting and microbial H₂ oxidation. Importantly, the H₂ inhibition constant of N₂ase is K_i(D₂) ∼11 kPa (31); the low H₂ partial pressure generated by water splitting (roughly 0.3% or 0.3 kPa H₂, depending on gas flow rate) does not impede N₂ fixation and/or reduce the NRR energy efficiency. When the reactor is pressurized with an external H₂ source, microbial growth is attenuated (SI Appendix, Fig. S3A). This contention is illustrated by numerical simulations (SI Appendix, Fig. S3B), which show slower biomass accumulation in the case of microbes under high H₂ concentration but linear growth in CoP_i | Co–P water splitting. In this regard, the direct hybrid device provides potential benefit, as the generated H₂ from water splitting not only provides cellular energy but also controls metabolic pathways, thus allowing growth conditions to be controlled with fidelity.

X. autotrophicus cells can be applied directly to promote plant growth and in this regard is, in effect, a living biofertilizer. Cultures of X. autotrophicus were collected, washed, resuspended in 50 mM NaCl saline, and applied to greenhouse radish growth experiments to assess their ability to improve harvest yields (Fig. 3). Increasing amounts of X. autotrophicus, applied weekly, increased edible radish storage root mass by up to ∼130% compared with unfertilized controls from 3.4 ± 1.2 g plant^–1 to 8.2 ± 1.4 g plant^–1 (n = 12 radishes for both) (Fig. 3 A and B and SI Appendix, Fig. S4A), and total mass by up to ∼100% (SI Appendix, Fig. S4 B and C). To evaluate a potential mechanism of improving plant growth, in vitro fertilization assays were conducted by suspending X. autotrophicus cells in a 50 mM NaCl solution and measuring extracellular secretion of N and P. Only living cells of X. autotrophicus released NH₄⁺ and PO₄^3– over the course of the 7-d experiment, with ethanol-killed controls displaying no or negative nutrient release behavior (Fig. 3C and SI Appendix, Fig. S5 A and B). Under these saline “starvation” conditions, these bacteria maintained ∼50% cell viability, likely metabolizing stored PHB carbon reserves (SI Appendix, Fig. S5A) (32, 33), demonstrating their capacity to be self-sustaining even in the absence of external energy inputs. Whereas actively growing X. autotrophicus cells do not secrete N_soluble (Fig. 2A), under these starvation conditions, leaky reutilization pathways potentially slowly release N and P from polyglutaminyl-rich cell membranes (34) and polyphosphate granules (32). From this, we contend that living X. autotrophicus cells improve plant growth as a slow-release source of bioavailable N and P, demonstrating the capability of this hybrid inorganic–biological NRR cycle to effectively bridge the gap between atmospheric N₂ and plant biomass.

Fig. 3.

X. autotrophicus biofertilization of radishes. (A) Yields of radish storage roots from biofertilization with different amounts of X. autotrophicus (X. a) (n = 12 radishes per treatment) in as-supplied potting media. No X. a: OD₆₀₀ = 0, Low X. a: OD₆₀₀ = 0.03, Med. X. a: OD₆₀₀ = 0.3, High X. a: OD₆₀₀ = 3.0, applied at t = 7, 14 d. Corresponding dry masses and shoot masses are given in SI Appendix, Fig. S4 A–C. Significance (P value) calculated by a two-tailed, heteroscedastic Student’s t test. (B) Photographs of radishes from A. (C) Extracellular NH₄⁺ release from live and dead X. autotrophicus after 7 d in 50 mM NaCl starvation conditions. Dead cells were prepared by 70% EtOH sterilization. 0×: OD₆₀₀ = 0, 1×: OD₆₀₀ = 0.5, 10×: OD₆₀₀ = 5 (n = 3 biological replicates). (D) Growth yields of radish seeds with and without seed sterilization by hypochlorite treatment, and preinoculation with and without X. autotrophicus (n = 15). Experiments conducted in sterilized potting medium. (E) Growth yields of radish seeds sterilized and inoculated with X. autotrophicus, B. japonicum, V. paradoxus, or no inoculation, fertilized at t = 7, 14 d with X. autotrophicus biofertilizer in sterilized potting medium. All error bars indicate the SD centered on the arithmetic mean.

X. autotrophicus engages in specific plant–microbe and soil microbe–microbe interactions. To study this, we evaluated the effect of soil microbiome composition on X. autotrophicus biofertilizer performance. As-supplied potting media (Promix HP MYCORRHIZAE) contains a plant-growth-promoting fungal inoculant. Reuse of the potting media or sterilization by autoclaving effectively removes these microbial species (SI Appendix, Fig. S5 C and D), allowing more direct assessment of radish–X. autotrophicus interactions. Interestingly, reused potting media previously fertilized with X. autotrophicus exhibit a carryover effect, demonstrating that this biofertilizer improves plant yields over multiple planting cycles (SI Appendix, Fig. S5E), although comparatively lower yields from cycle 1 to cycle 2 suggest continued biofertilization produces the best radish growth. Radish seeds themselves also carry along native microbial inoculants on their seed coat and can be sterilized by quick treatment with a dilute hypochlorite solution (35). Radish seeds sterilized by this method showed lower growth yields (Fig. 3D) compared with unsterilized controls, attesting to the importance of plant–microbe commensal interactions. Radish seeds preincubated with X. autotrophicus before sowing, in the method of biopriming (35), were able to compensate for the loss of the native radish microbiome, and in fact improve total plant mass by ∼40% compared with unsterilized controls (Fig. 3D). This interaction was found to be somewhat specific to X. autotrophicus as inoculation of sterilized and unsterilized radish seeds with another diazotroph, Bradyrhizobium japonicum [American Type Culture Collection (ATCC) 10324], and a plant-growth-promoting bacterium, Variovorax paradoxus (ATCC 17713) (36), showed similar yields as unsterilized radish seeds (SI Appendix, Fig. S4D). To discern the soil microbe–microbe interactions, radish seeds were inoculated with different plant-growth-promoting bacteria, and fertilized weekly with X. autotrophicus cells (Fig. 3E). Uninoculated radish seeds in sterile soil showed poor growth even with X. autotrophicus fertilization, suggesting that biofertilization is most effective in cooperation with preestablished plant–microbe relationships (from biopriming or soil fungal inoculants) to facilitate nutrient uptake. Growth improved upon inoculation with B. japonicum and no significant change was observed with V. paradoxus, although the greatest growth increase was for X. autotrophicus inoculated and fertilized radishes, at an average storage root mass of 6.2 ± 2.7 g plant^–1 (n = 10) compared with 0.4 ± 0.2 g plant^–1 (n = 14) for uninoculated controls, an improvement of ∼1,440%. These results suggest that X. autotrophicus not only works as a biofertilizer in conjunction with other members of the soil microbiome (as represented by B. japonicum), but that X. autotrophicus itself engages in a unique plant-growth-promoting mechanism with radishes to realize increased yields. Its function as a soil and seed inoculant in addition to acting as a N and P source positions it as a potential strategy to revive biologically degraded and nonarable soils (21). Further studies will be required to unambiguously define the mechanism of X. autotrophicus–plant interactions.

X. autotrophicus interacts favorably with sustainable abiotic fertilizers. One such natural fertilizer is human urine, which is an attractive alternative soluble N and P fertilizer (37), as well as a potential growth substrate for soil bacteria. X. autotrophicus was found to grow autotrophically in a synthetic defined urine medium (38) diluted to an appropriate ionic concentration (SI Appendix, Fig. S5F). This bacterium was also tolerant to the variable loading of N and P associated with their concentration in natural urine sources (SI Appendix, Fig. S5G) (39, 40). These results suggest that this biofertilizer is compatible with existing sustainable fertilizers as a potential medium for X. autotrophicus cultivation and in-soil propagation. As observed for other soil bacteria (15), the ability of X. autotrophicus to transform the labile N and P in urine into stable, slow-release forms is an intriguing prospect for future longitudinal field studies (20).

In addition to fixing N₂ in the form of soluble biomass, the hybrid device can be induced to excrete synthesized NH₃ directly into the extracellular medium. Genome sequencing of the strain of X. autotrophicus used here (SI Appendix, Table S3) indicates that the NH₃ generated from N₂ase is incorporated into biomass via a two-step process mediated by glutamine synthetase (GS) and glutamate synthase (GOGAT) (Fig. 4A), as is consistent with previous biochemical assays (41). If the functionality of this NH₃ assimilation pathway is disrupted, direct production of extracellular NH₃ should occur. Noting that GS inhibitors can induce NH₃ secretion in sugar-fermenting diazotrophs (14), we turned to the specific GS inhibitor phosphinothricin (PPT) (42) to block the NH₃ assimilation pathway and allow NH₃ to passively diffuse out into the extracellular medium (pathway 2 in Figs. 1 and 4A). After adding a sufficiently high concentration of phosphinothricin (50 μM) to inhibit GS (42), the biomass accumulation of X. autotrophicus slows, while N_total and the concentration of free NH₃ in the solution (N_NH3) continues to increase (Fig. 4B), indicating that N₂ fixation after phosphinothricin addition is in the form of extracellular NH₃. ¹⁵N isotope labeling experiments confirm that the generated NH₃ is from N₂ (Fig. 4C). After exposing the microbes to ¹⁵N-enriched N₂, the intensity of the ¹⁵NH₄⁺ doublet peaks (6.91 ppm, J¹_NH = 72.7 Hz) in the ¹H NMR spectrum of the supernatant monotonically increases. This ¹⁵NH₄⁺ signal is not detectable in control samples under N₂ of natural abundance; the ¹H NMR exhibits only the triplet signal of ¹⁴NH₄⁺ (6.95 ppm, J¹_NH = 50.0 Hz). The concentration of N in NH₃ (N_NH3) after terminating the experiment was 11 ± 2 mg L^–1 (∼0.8 mM) and the accumulated N_total reached 47 ± 3 mg L^–1 (n = 3, SI Appendix, Table S1). The rate of N₂ fixation decreases in the latter phase of experiment due to down-regulation at transcriptional and posttranscriptional levels (43).

Fig. 4.

NH₃ production in the extracellular media. (A) Biochemical diagram of NH₃ synthesis from N₂. OG, 2-oxoglutarate; PPT, phosphinothricin. (B) The OD₆₀₀, the amount of charge passed through, the concentration of N_total and N_NH3 are plotted versus the duration of experiments. The addition of PPT as GS inhibitor was displayed. n = 3; error bars denote SEM. (C) The ¹H NMR spectrum evolution of generated NH₄⁺ under ¹⁵N-enriched and naturally abundant N₂. Time counted as the duration after providing ¹⁵N₂. * denotes resonances for the internal standard of H–CON(CH₃)₂. Gln, glutamine; Glu, glutamic acid.

Conclusion

We demonstrate herein the synthesis of solid N biomass and NH₃ from N₂, driven by water splitting at appreciable TON and TOF, under ambient conditions. By driving water splitting with solar energy and renewable electricity, the approach can potentially provide a renewable synthesis platform for the NRR. We demonstrate that the production of NH₃ and X. autotrophicus biofertilizer allows this hybrid inorganic–biological NRR system to effectively connect atmospheric N₂ to plant biomass. The hybrid inorganic–biological approach illustrated for the NRR may be generalized to a renewable biological and chemical synthesis platform, depending on the biomachinery to which water splitting is coupled.