Azotobacter vinelandii N₂ fixation increases in co-culture with the PGPR Bacillus subtilis in a nitrogen concentration-dependent manner

ABSTRACT

Biological nitrogen fixation (BNF) is an essential source of new nitrogen (N) for terrestrial ecosystems. The abiotic factors regulating BNF have been extensively studied in various ecosystems and laboratory settings. Despite this, our understanding of the impact of neighboring bacteria on N₂ fixer activity remains limited. Here, we explored this question using a co-culture of the two model species: the free-living diazotroph Azotobacter vinelandii and the non-fixing plant growth-promoting rhizobacteria Bacillus subtilis. We observed that the interaction between the two bacteria was modulated by N availability. Under N-replete conditions, B. subtilis outcompeted A. vinelandii in the co-culture. Under N-limiting conditions, BNF activity by A. vinelandii was enhanced in the presence of B. subtilis. Reciprocally, the presence of A. vinelandii repressed sporulation by B. subtilis and supported its growth likely through N transfer. N inputs by A. vinelandii were doubled in the presence of B. subtilis compared to the monoculture, primarily due to the retention of a robust N₂ fixation activity in the stationary phase. A proteomic analysis revealed that A. vinelandii N metabolism, particularly the molybdenum nitrogenase isoform protein levels (NifK and NifD), was upregulated during the stationary growth phase in the presence of B. subtilis. This study revealed that N stress drives bacterial interactions and activity in a two-species community, especially in the stationary phase.

IMPORTANCE

Reducing inputs of chemical N fertilizers is essential to develop a more sustainable agriculture. The stimulation of biological nitrogen fixation by N2 fixers in multispecies cultures, here the plant growth-promoting rhizobacteria Azotobacter vinelandii and Bacillus subtilis, opens opportunities for the formulation of biofertilizers consortia. While most research on N2 fixation historically focussed on the exponential growth phase of microorganisms, we observed that Bacillus subtilis stimulated Azotobacter vinelandii N2 fixation mostly during the stationary phase. This result highlights that more research on the factors controlling N2 fixation repression during the stationary growth phase, especially bacteria-bacteria interactions, is eagerly needed.

INTRODUCTION

In soil, bacteria co-exist in complex and diverse communities that are essential to the function of terrestrial ecosystems. Community members usually have complementary traits important for the community overall function (1). For instance, some bacteria rely on cross-feeding for their survival (2). It is increasingly clear that deciphering soil function requires considering the soil microbial community as a whole instead of independent entities.

The acquisition of macronutrients [i.e., carbon, nitrogen (N), phosphorus, and sulfur] and micronutrients (e.g., metals) is one of the primary drivers of microbial interactions in soil (3). N is, with phosphorus, the nutrient most often reported to limit primary production in terrestrial ecosystems (4). To support crop productivity, modern agriculture relies on massive use of chemical N fertilizers that are energetically costly to produce and harmful to the environment (5). They can reduce soil fertility and alter microbial community structure, diversity, and function (6).

Some members of the microbial soil community, the diazotrophs, can reduce the atmospheric N₂ into bioavailable ammonium. This reaction, catalyzed by the enzyme nitrogenase, is energetically costly and consequently most active in the carbon-rich matrices of the soil, such as the rhizosphere. Several edaphic (e.g., nutrient availability and pH) and climatic (e.g., temperature and humidity) factors are known to affect biological nitrogen fixation (BNF) (7–9). The repression of BNF by N availability in the environment is well documented (10) and could constitute a bottleneck for using diazotrophs as biofertilizers in agriculture. Still, interest in diazotrophs as an alternative or a complement to chemical N fertilizers is quickly rising (10–13).

Our understanding of the role of multispecies interactions on BNF remains limited but could hold the key to unlocking the potential of diazotrophs in agricultural settings. N₂ fixers are a source of biogenic N input in soil that can shape microbial community structure and function (14, 15). N₂ fixers can promote the metabolism of surrounding microorganisms through N transfer (16). For instance, the free-living N₂-fixing bacterium Azotobacter vinelandii was shown to restore microalga growth in an N-depleted medium through the secretion of N-rich metallophores (17). Conversely, recent evidence suggested that neighboring non-fixing bacteria can influence N₂ fixer activity. Studies on Rhodopseudomonas palustris, Azospirillum sp., and N₂ fixers isolated from the xylem of maize reported that co-cultures of N₂ fixers with non-fixers lead to higher nitrogenase activity than in monocultures (18). Multispecies bacterial consortia, combining N₂ fixers and non-fixers, capable of alleviating or easing the need for N-fertilizers while promoting plant growth, would be valuable tools for developing a more sustainable agriculture.

Here, we investigated how N stress impacts the interaction between an N₂ fixing and a non-fixing plant growth-promoting rhizobacteria (PGPR) using a simplified bipartite model consisting of A. vinelandii ATCC 12837 and B. subtilis NCIB 3610. Although not co-isolated, these species are ubiquitous soil bacteria found in the rhizosphere of the two most important crops in agriculture, i.e., rice (19, 20) and wheat (21, 22).

A. vinelandii is a PGPR that can provide bioavailable N to plants in various chemical forms (23) through BNF. While A. vinelandii is an obligate aerobe, it has developed various mechanisms to prevent nitrogenase inhibition by O₂ (24), such as alginate encapsulation (25) and enhanced respiration (26). These mechanisms add to the already costly N₂ fixation process (27). B. subtilis is a non-fixing PGPR that synthesizes growth-promoting phytohormones such as cytokinins (28) and antimicrobial compounds such as surfactin (29) and iturin (30). It can protect plants against pathogens by activating the induced systemic resistance (31). Under N deficiency, B. subtilis forms endospores (32, 33) that are highly resistant to various stresses but mostly metabolically inactive. Thus, B. subtilis cannot exert most of its PGPR traits in N-deficiency.

Since nutrient availability can polarize bacterial interactions, we hypothesized that the N concentration in the growth medium would modulate the relation between A. vinelandii and B. subtilis. We predicted that under N limiting conditions, A. vinelandii will relieve N stress for B. subtilis supporting the growth of both organisms. Indeed, we observed that at low N, the two bacteria co-existed in the co-culture, and that community led to an increase in N₂-fixation. At high N, competitive interactions characterized the co-culture. Mechanisms underlying B. subtilis and A. vinelandii mono and co-culture response to N were further investigated through N mass balance and proteomic analyses.

RESULTS

Growth of mono and co-cultures in response to nitrogen availability

We assessed the effect of N availability on A. vinelandii and B. subtilis interaction by monitoring their growth in mono and co-culture along an N gradient. As predicted, in monoculture, biomass production after one week was positively influenced by N availability for both species. B. subtilis growth was proportional to N availability between 0.075 mM to 3 mM, showing that in this N range, growth is N-limited. As expected, for A. vinelandii, NH₄Cl content had little impact on growth, illustrating that N-fixation allows alleviating N limitation (Fig. 1A and B; Fig. S1). Nitrogen limitation constraining B. subtilis growth in the absence of NH₄Cl is further illustrated by strong sporulation observed in the absence of A. vinelandii (Fig. 1C). In the co-culture, the presence of B. subtilis had no measurable impact on A. vinelandii up to 7.5 mM NH₄⁺, but significantly impeded growth above this threshold proportionally to the N concentration (Fig. 1A). Conversely, B. subtilis growth was not affected by the presence of A. vinelandii above 7.5 mM NH₄ but was significantly enhanced (up to 10 times) below this threshold (Fig. 1B). B. subtilis is a well-known producer of various antimicrobial compounds (34). The reduced growth of A. vinelandii in the presence of B. subtilis under high N concentration could result from the production of such compounds. A triple mutant deleted for the genes encoding three main antimicrobial compounds [i.e., surfactin, bacillomycin, and plipastatine (34)] showed strong antagonism against A. vinelandii (Fig. 2). In the absence of N, the growth of B. subtilis in the presence of A. vinelandii was maintained after 3 days, while it collapsed in the monoculture (Fig. 1D). Still, the maximal growth of B. subtilis in these conditions did not exceed 10% of the maximal growth observed at high N (Fig. 1D), suggesting that its growth remained restricted by N availability.

Fig 1.

Cell growth along the N gradient. (A) Final growth was obtained after 6 days post-inoculation of Azotobacter vinelandii (Av) mono- or co-culture with Bacillus subtilis (Bs) grown on Subtiburk containing different concentrations of NH₄Cl. (B) Final growth obtained after 6 days post-inoculation of B. subtilis in mono- or co-culture with A. vinelandii grown in Subtiburk containing different concentrations of NH₄Cl. (C) Sporulation dynamic of B. subtilis in Subtiburk without fixed N, in presence or absence of A. vinelandii. (D) Log transformed growth curve of mono- or co-culture of Av and Bs grown in Subtiburk without added N. Species were inoculated at an OD₆₀₀ ratio of 1:1 (OD_i ~10⁶ CFU/mL). Values represent the mean of three technical replicates, and the experiments are representative of at least three independent biological replicates. For all panels, error bars represent standard deviation. The number of stars translates a significant change compared to the control (analysis of variance; *P < 0.05, **P < 0.01, ***P < 0.001; ).

Fig 2.

B. subtilis antagonism at high N (15 mM) against A. vinelandii is non-antimicrobial dependent. Azotobacter vinelandii (Av) final cell density in CFU.mL⁻¹ after 6 days with and without B. subtilis (Bs) WT and (A) Bs single NRPS deletion mutants [srfAA (surfactin), pksX (bacillaene) and sfp (plipastatine and surfactin)] or (B) the triple deletion mutant of B. subtilis srfAA, plsX, ppsA. The number of stars translates to a significant change in Av cell density in co-culture with the different Bs compared to the monoculture (one-way analysis of variance, ***P < 0.001, ****P < 0.0001).

The beneficial effect of A. vinelandii on B. subtilis in the absence of N is further illustrated by the substantial reduction of B. subtilis sporulation in co-cultures (11%), as compared to monocultures (100%), suggesting that B. subtilis retrieved a significant amount of N from A. vinelandii (Fig. 1C). As B. subtilis is a well-known producer of various antimicrobial compounds, this bioavailable N could result from A. vinelandii cell lysis. However, a LIVE/DEAD staining revealed no significant difference in cell death between the mono- and the co-culture (Fig. 3).

Fig 3.

Cell death in mono- and co-culture in the absence of N (0 mM) in the medium. Azotobacter vinelandii (Av) cell death (in %) during time (days) and with and without Bacillus subtilis (Bs). Cell death was assessed with LIVE/DEAD cell staining and quantified by flow cytometry using a BD Accuri C6 Plus. The number of stars translates to a significant change in Av cell death in co-culture compared to the monoculture (two-way analysis of variance, **P < 0.01).

Nitrogen fixation rates

Acetylene reduction assays (ARAs) revealed that BNF N input by A. vinelandii was inversely correlated to bioavailable N in both the mono and co-culture (Fig. 4A), which is consistent with the known regulation of BNF by N (32). However, at low N (<7.5-mM N), we observed that the N input in the co-culture after 7 days was up to three times higher (i.e., at 0.15 mM) in the co-culture than the monoculture (Fig. 4A), suggesting a stimulation of BNF in the presence of B. subtilis. This important enhancement of N input in the co-culture was driven by higher N₂ fixation rates in the late exponential and stationary phases of the co-culture compared to the monoculture (Fig. 4B; Fig. S2).

Fig 4.

B. subtilis influences N₂ fixation efficiency along the N gradient. (A) Cumulative nitrogen fixed after 1 week of A. vinelandii grown in mono- or co-culture with B. subtilis (molNred·mL⁻¹) under different concentrations of NH₄Cl. The total Nred was calculated as the sum of the Cell density (CFU/mL) × nitrogen fixation rate (molN·CFU⁻¹·mL⁻¹·h⁻¹) over 6 days. (B) Nitrogen fixation rates during 6 days in mono- and co-culture reported as the nmol·Nred.CFU⁻¹. Values represent the mean of at least three technical replicates, and the experiments are representative of at least three independent biological replicates. Error bars represent standard deviation. The number of stars translates a significant change compared to the control (one way analysis of variance; *P < 0.05, **P < 0.01, ***P < 0.001, ).

To get a better understanding of how B. subtilis stimulated A. vinelandii N₂ fixation activity, we added cell-free supernatants from mono and co-cultures and measured the N₂ fixation rates of A. vinelandii after 5 days. The addition of B. subtilis and A. vinelandii mono-supernatants did not affect A. vinelandii N₂ fixation. In contrast, the addition of a co-culture supernatant increased it (Fig. 5A). This result strongly suggests that the stimulation of A. vinelandii N₂ fixation is, at least in part, triggered by soluble compounds produced by B. subtilis when in the presence of A. vinelandii.

Fig 5.

Effect of co-culture and co-culture secretion on B. subtilis growth and A. vinelandii nitrogen fixation activity at low N (0.15 mM). Bacteria were grown in Subtiburk, and after 4 days, their supernatants were collected and concentrated twice and completed at a ratio 1:1 with fresh Subtiburk before inoculation and incubation for 5 days. We then assessed the (A) nitrogen fixation rate of Azotobacter vinelandii (Av) after 5 days of growth with 0.15 mM NH₄Cl, completed with supernatants of Av, Bacillus subtilis (Bs, collected after 4 days post-inoculation) or of the co-culture (AvBs). Also shown is the rate of the co-culture. The stars indicate a significant difference from the control (A. vinelandii alone in new medium). (B) B. subtilis growth after 5 days with 0.15-mM NH₄Cl, completed with supernatants of Bs, Av, or of AvBs. Also shown is the growth in the co-culture. The number of stars translates to a significant increase in B. subtilis growth compared to the control, B. subtilis alone in new medium without supernatant addition (one-way analysis of variance; *P < 0.05, ***P < 0.001).

Cellular N content and N transfer

We further evaluated the fate of this higher amount of N fixed in the co-culture, which could accumulate in A. vinelandii cells, in the medium, or both. Interestingly, supernatants of A. vinelandii monocultures and co-cultures did not significantly increase B. subtilis growth. This result suggests that N concentration in the medium at a given time was not enough to meet B. subtilis need for further growth. Only direct co-culture of the two species was able to increase B. subtilis growth under low N (Fig. 5B). We then measured N concentrations in cells of A. vinelandii grown alone and in culture with B. subtilis in the presence of low (0.15 mM) or high exogenous NH₄Cl (15 mM). At low exogenous N concentration, we approximated that the cellular N concentration measured in co-culture is a good proxy for A. vinelandii cellular content because B. subtilis accounted for only about 11% of the cells. N content in A. vinelandii cells was significantly higher in the co-culture than in the monoculture throughout the growth (up to 35%) (Fig. 6A). This result suggests that A. vinelandii cell material accounted for most of the extra N fixed in the co-culture. In contrast, at high exogenous N (15 mM), the cellular N contents were lower in the co-culture than in the monoculture of A. vinelandii (Fig. 6B). This observation suggests that in N-replete conditions, in the presence of B. subtilis, both species compete for the same N source, resulting in an overall decrease in cellular N content in both species. Further characterization of cellular N in both species would be required to confirm this interpretation, but to do so, a technically challenging separation of both species from the preculture would be needed.

Fig 6.

Intracellular nitrogen concentration of A. vinelandii in mono- or co-culture with B. subtilis at low (A) (0.15-mM NH₄Cl) or high (B) (15-mM NH₄Cl) exogenous nitrogen. Values represent the mean of three technical replicates, and the experiments are representative of three independent biological replicates. Error bars represent standard deviation.

Protein expression in the co-culture

To determine the cellular pathway(s) involved in increased N₂ fixation, we performed a proteomic comparative analysis on mono- and co-cultures of A. vinelandii and B. subtilis in the exponential and stationary phase, with N₂ as the sole source of N. Most proteins differentially expressed in the co-culture compared to the monoculture were observed during the stationary phase (days 5 and 6) (Fig. 7A); of note, both days were very similar in terms of protein composition. Far fewer proteins were differentially present during days 1 and 2. At those time, the overrepresented proteins in the co-culture were mostly conserved hypothetical proteins, proteins involved in catabolism for energy generation such as glycolysis (including GapA) and protein biosynthesis. Proteins underrepresented were in cell stress management and cyst formation. For days 5 and 6, in the co-culture and low N, the proteins upregulated were for N metabolism (including NifK and NifD), chemotaxis (including MotC, FlgN, and FliA), and energy generation pathways (including the electron transport chain with RnfA, the TCA cycle with FumB, and the respiratory chain with the Nuo protein) (Table S1). At the same time, cyst formation, stress response, and metal transport (Fe− and Mo) were downregulated (Fig. 7B). Only 10% of the total pool of proteins identified could be attributed to B. subtilis due to its low abundance in the population under these experimental conditions.

Fig 7.

Protein level analysis in low N (0.15 mM) in the co-culture compared to the monoculture growth. (A) Volcano plots representing the significant (−log10 P value) difference in protein expression levels (fold change < or >1.4; P < or >0.05) in the co-culture samples compared to the monoculture ones during the time (at days 1, 2, 5, and 6). (B) Heatmap represented the differentially expressed proteins (log₂FC) in co-culture compared to monoculture on days 1, 2, 5, and 6. Fold change represents means. Positive values indicate that the protein was more abundant in co-culture, and negative values indicate that the protein was less abundant in co-culture. Black lines mean that those proteins were not detected in those conditions. Proteins were then clustered by groups using STRING v.12.0 for functional protein enrichment analysis. Proteins from these functional groups were attributed according to their Gene Ontology term annotation.

DISCUSSION

Nitrogen availability drives B. subtilis and A. vinelandii interaction

The stress-gradient hypothesis (SGH) states that increased nutrient stress drives a shift in interactions between species from predominant negative interactions (e.g., competition and parasitism) in nutrient-replete conditions to dominant positive interactions (e.g., cooperation and facilitation) in limiting conditions (33, 35–40). While the SGH theory remains open to debate, a recent meta-analysis suggests that it could be a relevant and widely applicable ecological concept (41). Our results concur with the concept of the SGH, with a drastic shift from negative to positive interactions between A. vinelandii and B. subtilis along an N availability gradient. We provide a conceptual scheme summarizing our main observations and their interpretations within the frame of the SGH (Fig. 8).

Fig 8.

N availability polarizes A. vinelandii and B. subtilis interaction. At low NH₄Cl concentrations, A. vinelandii fixes at high level and B. subtilis benefits from N release and thus grows in these otherwise N-stressed conditions. At high NH₄Cl concentrations, B. subtilis growth widely benefits from the exogenous N input, which inhibits A. vinelandii growth. Proteins represented are key players of the N nutrition and are overproduced in the co-culture compared to the monoculture (red arrows). This figure was created in BioRender.com (BioRender.com/j06e383).

Under N-replete condition (high N), competition dominated the interaction with the presence of B. subtilis reducing A. vinelandii growth in co-culture. The production of common antibiotics by B. subtilis failed to explain the reduced growth of A. vinelandii in co-culture. This result suggests that other compounds could be involved in B. subtilis antagonism against A. vinelandii at high N or that another mechanism is at play.

Based on our observation, the interaction between A. vinelandii and B. subtilis in low N is best described as facilitation. Facilitation is a positive interaction between organisms in which one species benefits the other by improving access to limiting nutrients (36). The presence of A. vinelandii had clear benefits for the survival of B. subtilis by reducing sporulation and sustaining active growth, likely through the transfer of N, while it had no measurable advantages or disadvantages for A. vinelandii. During nutritional stress (e.g., N limitation), B. subtilis secretes antimicrobial compounds before entering dormancy (42, 43). By secreting N in the medium and lifting B. subtilis N stress, A. vinelandii could prevent the production of antimicrobial compounds detrimental to its survival. From a larger perspective, even if BNF is an energy-costly reaction, it might still be advantageous under N-limiting conditions for the N₂ fixer to share N with starving surrounding species rather than manage and survive multiple negative interactions.

N transfer from A. vinelandii to B. subtilis

Under low N, the sustained growth of B. subtilis in the presence of A. vinelandii demonstrates a transfer of fixed N compound from A. vinelandii to B. subtilis. However, the nature (chemical form) of this transfer is unknown. Indeed, A. vinelandii secretes a variety of N-rich compounds such as ammonium (44), amino acids (45), and siderophores (23), depending on the growth conditions.

Here, the medium had a low Fe concentration (<5 µM), which likely stimulated the production of siderophores, around 10⁻⁵ M, to support Fe acquisition (46, 47). The most common siderophores of A. vinelandii are N-rich compounds containing 2 (azotochelin), 4 (protochelin), and 13 (azotobactin) atoms of N, and these molecules were shown to restore the growth of algae in an N-depleted medium (17). Since B. subtilis encodes for several xenosiderophores transporters (48), siderophore scavenging could be a viable source of N. However, the downregulation of siderophore transporters in co-culture (FhuC, Avin_25350, Avin_37020, Avin_25420, Avin_25580, Avin_25610, Avin_25550; pvdE, Avin_25560) (Table S1) suggests that Fe stress and siderophore production were somehow reduced in the presence of B. subtilis compared to the monoculture. Thus, while the transfer of N from A. vinelandii to B. subtilis through siderophores cannot be ruled out, it is unlikely to be the primary pathway for N transfer.

During N₂ fixation, A. vinelandii releases ammonium (NH₄⁺) in the medium by passive diffusion (49). To counter these losses, A. vinelandii expresses an ammonium transporter to retrieve this extracellular NH₄⁺ (50). The improved BNF activity of A. vinelandii in the stationary phase of co-cultures could enhance this NH₄⁺ loop. Indeed, many proteins involved in nitrogen metabolism and N assimilation into the biomass [i.e., glutamate-glutamine synthase protein GltD and hydrogenases HoxK, HoxO, and HypB, usually overexpressed during diazotrophy (51)] were overexpressed by A. vinelandii in co-cultures during the stationary phase. We hypothesize that this enhanced NH₄⁺ loop provides a means for B. subtilis to benefit from the additional N reduced by A. vinelandii during the stationary phase. This hypothesis is further supported by the strong presence of an NH₄⁺ transporter NrgB in the B. subtilis (Table S2) population in co-culture with A. vinelandii. Further, faster cycling of NH₄⁺, while providing a steady supply of N, would not lead to significant (<0.1 ppm) accumulation of N in the medium at any given time during the growth due to fast consumption by both bacteria, thus explaining why B. subtilis could not grow in the supernatant transfer experiments (Fig. 5B). Further research is needed to test the validity of this hypothesis.

The co-culture stimulates BNF

The stimulation of BNF activity in the presence of another bacteria has been reported in a handful of studies (18, 52–55). The proposed mechanisms underpinning this behavior include oxygen protection and the transfer of small energy-rich compounds. Consequently, we offer two hypotheses about how B. subtilis could impact A. vinelandii BNF. First, the presence of B. subtilis cells could ease oxidative stress. A. vinelandii, an obligatory aerobic bacterium, manages O₂ toxicity to the nitrogenase using respiratory protection (25). The presence of another species consuming O₂ in the medium could reduce the cost of respiratory protection by lowering the medium O₂ concentration. Second, B. subtilis could support the energy demand from the nitrogenase by supplying A. vinelandii with readily available C compounds. A similar mechanism was described for E. coli, which can stimulate Rhodopseudomonas palustris growth and N₂ fixation in co-culture by providing fermentation by-products such as formate, ethanol, and organic acids (52). The presence of a protein involved in lactic acid synthesis (Ldh) (Table S2) suggests that a similar phenomenon could happen in the B. subtilis and A. vinelandii co-culture.

The results of our study stress two important pieces of information. The first is the necessity to include nutrient variation when studying mixed communities to have a dynamic view of the interactions as they might drastically change according to the environment. This knowledge is essential when the endpoint is product development, such as biofertilizers, since the presence or absence of nutrients in the environment might influence their efficacy. Nitrogen biofertilizers are promising strategies to reduce chemical N intrants. However, they are realistically not meant to replace chemical fertilizers as their efficacy can simply not match the N need for intensive cropping (at least in the foreseeable future). Testing the performance of biofertilizers under different contexts of chemical N addition is necessary to evaluate their applicability properly. The second is that it is essential to use an extended time course when examining environmental bacteria since physiologic effects could be revealed only in the stationary phase. This growth stage is rarely investigated in laboratory studies despite being an abundant growth stage in nature.

MATERIALS AND METHODS

Strains and A. vinelandii ATCC 12837 sequencing

Bacillus subtilis NCIB 3610 ycbU-lmrB::specR was used since it allowed the use of spectinomycin as a selection marker for CFU counting in co-culture.

Azotobacter vinelandii ATCC 12837 WT was chosen for its relevance since this strain was isolated from the environment. This strain was sequenced for this study. Genomic DNA was extracted from 9 mL of 8-h cultures using Monarch Genomic DNA Purification Kit. Fragment reads were generated on an Illumina NextSeq sequencer using TG NextSeq 500/550 High Output Kit v.2 (300 cycles) (RNOmique Platform, UdeS). The genome was then assembled de novo using Unicycler with default parameters obtaining 376 contigs for a size of around 5.2 Mbp. Using RAST to construct a phylogenic tree, A. vinelandii DJ was identified as the closest strain to ATCC 12837. Based on that, the contigs were ordered (scaffold) using MeDuSa, which allowed us to obtain 1 scaffold (out of 17) containing most of the genome for convenient use.

Cell growth conditions

For both the mono- and co-culture experiments B. subtilis was precultured from glycerol stocks for 18 h on Luria-Bertani (LB; 1% wt/vol tryptone, 0.5% wt/vol yeast extract, and 0.5% wt/vol NaCl) agar (1.5%) plates at 37°C. A. vinelandii was precultured from glycerol stocks for 36 h on the defined medium Subtiburk agar (1.5%) [4.60 mM KH₂PO₄, 1.47 mM K₂HPO₄, 0.01 mM MOPS (3- (N-morpholino)propanesulfonic acid, pH 7), 0.811 mM MgSO₄, 0.955 mM CaSO₄, 5.37 µM FeCl₃, 1.23 µM Na₂MoO₄, 0.05 mM MnCl₂, 0.001 mM ZnCl₂, 0.002 mM thiamin, and 2% saccharose] at 30°C. Subtiburk was supplemented with increasing NH₄Cl from none to 37.4 mM, depending on the experiments. Colonies were collected and suspended in Subtiburk liquid medium to obtain an inoculum at an optical density of 0.5 at 600 nm (OD₆₀₀). Both strains were then inoculated in 24-well plates at a final OD₆₀₀ of 0.005 in a ratio of 1:1 for co-culture experiments. Plates were incubated at 30°C.

A. vinelandii and B. subtilis growth curves

To follow the growth of A. vinelandii and B. subtilis in mono- and co-culture, plate counting estimating the CFU was used. Aliquots of 10-µL dilutions ranging from 10⁻² to 10⁻⁶ of the different conditions were dropped on LB plates supplemented with spectinomycin (100 µg/mL) to select B. subtilis in the mono and co-culture; Subtiburk plates supplemented with crystal violet (2 µg/mL) were used to select A. vinelandii in the mono and co-cultures. From the dilution factor, we obtained the CFU per milliliter in each of the tested conditions. Growth curves correspond to the CFU per milliliter, depending on time (in days).

B. subtilis spore quantification

To assess for spores of B. subtilis, in the absence of added N (0-mM NH₄Cl) across growth conditions, aliquots of mono- and co-cultures of B. subtilis were incubated for 20 min at 80°C to kill vegetative cells. Then, the sporulation percentage was calculated using CFU per milliliter after and before the heat treatment.

Supernatant additions

In a low N Subtiburk (0.15 mM NH₄Cl), after 4 days, in order to achieve stationary phase and a high cell density (e.g., a potential good amount of secreted compounds in the growth supernatants), 1 mL of sample was filtered using a 0.22-µm Millipore membrane filter and concentrated to 500 µL using a Vacufuge. The samples were then completed with fresh Subtiburk medium to the original volume of 1 mL before reinoculation with mono- or co-cultures of B. subtilis and A. vinelandii at an OD₆₀₀ of 0.005 for 5 days.

Nitrogen content

Samples were withdrawn during exponential phase (days 1 and 2) and during the stationary phase (days 4–6) in low (0.15 mM NH₄Cl) and high (15 mM NH₄Cl) exogenous nitrogen concentrations. Cells were then washed with phosphate-buffered saline buffer. Supernatant-free cells were lyophilized for 48 h and then homogenized through bath sonication. Then, 1 mg ± 0.5 mg of dried sample material was encapsulated in tin (Sn) capsules. Total nitrogen was analyzed at the UC Davis Stable Isotope Facility using an elemental analyzer.

ARAs

The nitrogen fixation activity of A. vinelandii was assayed using an ARA (56). Briefly, 1 mL of bacterial culture was transferred in a 25-mL vial. The vial was closed with a rubber septum, and 10% (vol/vol) of the headspace was replaced by acetylene (C₂H₂) produced from the reaction of calcium carbide (CaC₂) with water in a Tedlar bag. After a 2-h incubation at 30°C, 3 mL headspace was withdrawn and transferred to a vacuumed 3-mL vial. Ethylene C₂H₄ production was quantified by gas chromatography equipped with a Flame ionization detector (GC-8A; Shimadzu, Kyoto, Japan) and a Supelco stainless-steel column, 80/100 HAYESEP N (2.44 m × 3.2 mm × 2.1 mm; Sigma-Aldrich Canada, Oakville, ON). A conversion ratio of 3:1 was used to convert acetylene reduction rates to N reduction rates (Hardy et al., 1968).

Peptide extraction

To prepare the samples for proteomic analysis, 24 mL of cells growing in low N (0.15 mM NH₄Cl) was collected after 1, 2, 5, and 6 days post-inoculation. The 48 cell pellets (biological triplicate and technical duplicate for each time point) were sonicated (12 cycles of 5-s pulse/5 s off, intensity of 20%–25%) on ice in a lysis buffer composed of 8 M urea, 10 mM HEPES (pH 8) to extract and solubilize total proteins. The samples were then centrifuged at 16,000 × g for 10 min at 4°C, and the supernatants were transferred to new 1-mL low bind tubes. Proteins were then quantified using a colorimetric method (Pierce BCA kit). Each sample was then normalized to obtain a concentration of 45 µg of proteins in 100 µL. Proteins were reduced with 5 mM dithiothreitol (DTT), boiled at 95°C for 2 min, and then incubated at room temperature (RT) for 30 min. Proteins were then alkylated with 7.5 mM of chloroacetamide and incubated again at RT for 20 min in the dark. One hundred fifty microliters of 50 mM NH₄HCO₃ was added to the mixture for a final concentration of urea of 2 M, and trypsin digestion (1:50) was then performed at 30°C overnight. The samples were then acidified using 0.2% trifluoroacetic acid (TFA). Peptides obtained were concentrated and desalted using a ZipTip 10-µL micropipette tips with a C18 column. The columns were washed and equilibrated before loading with the total of 30 µL of sample. Peptides were then eluted in 1% formic acid and 50% acetonitrile, transferred to a glass vial, quantified by nanodrop at 205 nm, and stored at −20°C before the liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis.

Sample analysis by LC-MS/MS

For the protein quantification, aliquots of 250 ng of samples were analyzed by high-performance liquid-chromatography nanoElute system (Bruker Daltonics). They were loaded onto a trap column with a flow of 4 µL/min (Acclaim PepMap100 C18 column, 0.3 mm id × 5 mm; Dionex Corporation, catalog # 164567) and eluted onto a C18 column (1.9 μm bead size, 75 μm × 25 cm; PepSep). The peptides were eluted with a linear gradient of 5%–37% acetonitrile in 0.1% formic acid over 120 min at 500 nL/min and injected into a TimsTOF Pro ion mobility mass spectrometer equipped with a Captive Spray nanoelectrospray source (Bruker Daltonics). Mass spectra were acquired using a data-dependent acquisition auto-MS/MS with a 100–1,700 m/z mass range.

Quantitative protein analysis

Raw data were analyzed using the MaxQuant software v.1.6.17.0, and spectra were matched against the UniProt reference of B. subtilis strain 168 and A. vinelandii DJ. The parameters used for the MaxQuant analysis were two miscleavages; fixed modification was carbamidomethylation on cysteine; enzymes were trypsin (K/R not before P); variable modifications included in the analysis were methionine oxidation, protein N-terminal acetylation, and protein carbamylation (K, N-terminal). The mass search tolerance was 10 ppm for precursor ions, and a tolerance of 20 ppm was used for fragment ions. Identification values “protein FDR” were set to 0.05. Minimum peptide count was set to 1. Label-free quantification (LFQ) was also selected with an LFQ minimal ratio count of 2. The “second peptides” and “match between runs” options were also used.

Perseus v.2.0.11 was then used for data processing. A filtering step for contamination in samples such as keratin and normalization was performed using the median of the median intensities of each condition. Missing peptide intensity values were replaced by a noise value.

Statistical analysis

Statistical analyses were performed using GraphPad Prism v.9.

DATA AVAILABILITY

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE partner repository with the data set identifier PXD056293.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/aem.01528-24.

Supplemental material. aem.01528-24-s0001.pdf.

Tables S1 and S2; Fig. S1 and S2.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.