Proximity-based defensive mutualism between Streptomyces and Mesorhizobium by sharing and sequestering iron

Abstract Microorganisms living in soil maintain intricate interactions among themselves, forming the soil microbiota that influences the rhizosphere microbiome and plant growth. However, the mechanisms underlying the soil microbial interactions remain unclear. Streptomyces and Mesorhizobium are commonly found in soil and serve as plant growth-promoting rhizobacteria (PGPR). Here, we identified an unprecedented interaction between the colonies of red-soil-derived Streptomyces sp. FXJ1.4098 and Mesorhizobium sp. BAC0120 and referred to it as “proximity-based defensive mutualism (PBDM).” We found that metabolite-mediated iron competition and sharing between the two microorganisms were responsible for PBDM. Streptomyces sp. FXJ1.4098 produced a highly diffusible siderophore, desferrioxamine, which made iron unavailable to co-cultured Mesorhizobium sp. BAC0120, thereby inhibiting its growth. Streptomyces sp. FXJ1.4098 also released poorly diffusible iron-porphyrin complexes, which could be utilized by Mesorhizobium sp. BAC0120, thereby restoring the growth of nearby Mesorhizobium sp. BAC0120. Furthermore, in ternary interactions, the PBDM strategy contributed to the protection of Mesorhizobium sp. BAC0120 close to Streptomyces sp. FXJ1.4098 from other microbial competitors, resulting in the coexistence of these two PGPR. A scale-up pairwise interaction screening suggested that the PBDM strategy may be common between Mesorhizobium and red-soil-derived Streptomyces. These results demonstrate the key role of iron in complex microbial interactions and provide novel insights into the coexistence of PGPR in soil.

A volume of 20 μL of DFOB (20 mM) was mixed with 20 μL of F1 (5 mM) or F2 (5 mM) to form a premix, which was then spotted at point A on the plate.After 4 h, M.
sp. BAC0120 was drop plated at points B to E and then further cultured at 28℃ for four days.Scale, 10 mm.In the co-culture plates, S. sp.FXJ1.4098 was inoculated four days ahead of M. sp.BAC0120.Sampling of S. sp.FXJ1.4098 was performed after its inoculation for six days (T1) and seven days (T2).Sampling of M. sp.BAC0120 was performed after its inoculation for three days (T2) and four days (T3).sp.BAC0120 inhibited by the plant pathogen was partially restored by S. sp.FXJ1.4098.Orange areas indicate that iron is chelated by siderophores.

Fig. S2 .
Fig. S2.Quantification of inhibitory activity on M. sp.BAC0120 of the extracts from interaction plate areas a, b, and c shown in Fig. 2B.The inhibitory activity was quantified by measuring the radius of the inhibition zone.Results are shown as dot plots and depicted with mean ± standard deviation.

Fig
Fig. S3.A representative HPLC profile showing a peak of inhibitor as the only remarkably differential metabolite distinguishing extracts from areas a−d in the co-culture plate of S. sp.FXJ1.4098 vs. M. sp.BAC0120 (n = 3).

Fig. S4 .
Fig. S4.Chemical elucidation and identification of the inhibitor.A. Mass spectrum of the inhibitor.B. Ultraviolet (UV) spectrum of the inhibitor.The mass signal of the inhibitor was at m/z 601.3566 [M + H] + and it only displayed an end UV absorbance.C. Annotation of the CID MS/MS spectrum for the [M + H] + ion of the inhibitor.Peaks underlined in red are consistent with the fragmentation pattern proposed for desferrioxamine E (DFOE) in the top panel, which assumes a macrocycle opening at the position of the red double dotted line.

Fig. S5 .
Fig. S5.Extracted ion chromatograms from UHPLC-HRMS analysis of culture extracts from S. sp.FXJ1.4098.The results of both the monoculture and coculture were the same.The m/z values of desferrioxamines (DFOs) used to generate each chromatogram are listed on the right.

Fig. S6 .
Fig. S6.A representative image showing the growth inhibition effect of desferrioxamine (DFO) on M. sp.BAC0120.Growth comparison of M. sp.BAC0120 on medium with or without 200 μM DFOB (n = 3).M. sp.BAC0120 was spotted onto GYM agar (left) and GYM agar containing 200 μM DFOB (right).The photograph was taken four days after inoculation of M. sp.BAC0120.The white dotted circle indicates the original inoculation sites of M. sp.BAC0120.

Fig. S9 .
Fig. S9.HPLC detection and identification of differential compounds in the extracts from areas a and b of co-culture plates of S. sp.FXJ1.4098 vs. M. sp.BAC0120 At the absorption value of 406 nm, F1 and F2 can be obviously detected in the extracts from areas a and b (Fig. 2A) of the co-culture plate, but not areas c and d (n = 3).

Fig. S10 .
Fig. S10.UV and mass spectra of Product F1 and F2. A. UV spectrum of product F1.The maximum UV absorbance was at 392 nm.B. Mass spectrum of product F1.The mass signal of product F1 was at m/z 655.2755 [M + H] + .C. UV spectrum of product F2.The maximum UV absorbance was at 406 nm.D. Mass spectrum of product F2.The mass signal of product F2 was at m/z 716.1807 [M + H] + .

Fig. S14 .
Fig. S14.Transcriptional analyses of heme biosynthetic genes in S. sp.FXJ1.4098 by RT-qPCR.Total RNAs were isolated from S. sp.FXJ1.4098 monocultured or co-cultured for six and seven days, and used for synthesizing cDNA.The 16S rRNA gene was used as an internal reference to normalize the RNA concentration.The relative transcription levels of 10 genes (tRNAglu, hemA, hemL, hemB, hemC, hemD, hemE, hemY, hemH, and hemQ) were obtained after normalization against the internal reference at corresponding time points.Error bars show the standard deviation of three independent experiments.Statistical significance was determined by Student's t-test.*, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Fig. S15 .
Fig. S15.Transcriptional analyses of heme uptake cluster genes in M. sp.BAC0120 by RT-qPCR.Total RNAs were isolated from M. sp.BAC0120 monocultured or co-cultured for three and four days, and used for synthesizing cDNA.The 16S rRNA gene was used as an internal reference to normalize the RNA concentration.The relative transcription levels of 5 genes (TBDR gene, hmuS, hmuT, hmuU, and hmuV) were obtained after normalization against the internal reference at corresponding time points.Error bars show the standard deviation of three independent experiments.Statistical significance was determined by Student's t-test.*, p < 0.05; **, p < 0.01; ***, p < 0.001; ns, not significant.

Fig. S16 .
Fig. S16.A dynamic model of hemin diffusivity.A. Representative images showing growth recovery effect on M. sp.BAC0120 by a volume of 20 μL of DFO (20 mM) mixed with 20 μL of hemin in different concentrations.The mixture was spotted at point A. The growth of M. sp.BAC0120 was recorded after 4 days of culture.B. The radius of hemin recovery (rh, equaled to the radius of recovery zone where M. sp.BAC0120 can grow around hemin) and the concentrations of hemin (chemin) conform to an exponential fitting function: chemin=0.004374e0.3521rh ."chemin" represents the concentration of hemin spotted at point A; "e" indicates Euler number; and "rh" indicates the radius of hemin diffusion.

Fig. S17 .
Fig. S17.Interactions among S. sp.FXJ1.4098,M. sp.BAC0120, and co-isolated bacteria, or plant pathogens.A. Colonies of co-isolated bacteria (left panel) and their pairwise interactions with M. sp.BAC0120 (right panel) (n = 3).The growth of M. sp.BAC0120 was inhibited or not affected by the co-isolated bacteria.B. Pairwise interactions between co-isolated bacteria and S. sp.FXJ1.4098 or FXJ1.4098ΔdesD (n = 3).The growth of co-isolated bacteria was either inhibited by S. sp.FXJ1.4098 due to its production of DFO or not affected by S. sp.FXJ1.4098.C. Tripartite interactions among M. sp.BAC0120, coisolated bacteria, and S. sp.FXJ1.4098 or FXJ1.4098ΔdesD (n = 3).The growth of M. sp.BAC0120 inhibited by the co-isolated bacteria was partially restored by S. sp.FXJ1.4098, and the growth of M. sp.BAC0120 not affected by the co-isolated bacteria was partially inhibited by S. sp.FXJ1.4098.D. The colony of a plant pathogen (left panel) and its pairwise interaction with M. sp.BAC0120 (right panel) (n = 3).The growth of M. sp.BAC0120 was inhibited by the plant pathogen.E. Pairwise interactions between a plant pathogen and S. sp.FXJ1.4098 or FXJ1.4098ΔdesD (n = 3).The growth of the plant pathogen was partially inhibited by S. sp.FXJ1.4098 attributed to its production of DFO.F. Tripartite interactions among a plant pathogen, M. sp.BAC0120, and S. sp.FXJ1.4098 or FXJ1.4098ΔdesD (n = 3).The growth of M.

Fig. S18 .
Fig. S18.Characterization of different pairwise interaction types between streptomycetes and rhizobia.A. Representative pictures for three interaction types: competition (Streptomyces inhibited the growth of rhizobia), PBDM (Streptomyces inhibited the growth of rhizobia at a certain distance while not affecting those in the vicinity), and neutrality (both Streptomyces and rhizobia grew as well as their monocultures).Rhizobia were inoculated onto the culture plates four days after the inoculation of Streptomyces strains, and pictures were taken after four days of co-culture.A total of 77 pairs of interactions (n = 3) were tested, among which 38 pairs showed competition, 27 pairs showed PBDM, and 12 pairs showed neutrality.Representative pictures are shown for three pairs of each interaction type.B. Quantitative and statistical analyses of the radius of inhibition zone in competitive interactions.The radius of inhibition zone by Streptomyces ranged from 15.17 ± 0.55 to 39.97 ± 0.76 mm (mean ± standard deviation), and no recovery zone was detected in these interactions.C. Quantitative and statistical analyses of the radii of inhibition and recovery zones in PBDM interactions.The radius of inhibition zone by Streptomyces ranged from 20.33 ± 0.25 to 40.37 ± 0.40 mm (mean ± standard deviation) and the radius of recovery zone by Streptomyces ranged from 10.77 ± 1.45 to 14.20 ± 2.05 mm (mean ± standard deviation).Grey shadows (error bars) show the standard deviation from three biologically independent experiments.

Fig. S19 .
Fig. S19.Detection of siderophore distribution in the pairwise interaction plates of 11 Streptomyces strains and M. sp.BAC0120 by an overlay of chrome azurol S (CAS) agar (n = 3).
overlapping sequences with the pJQ200SK plasmid for Gibson 209 assembly of linear DNA fragments.

Table S6 . Relative expression levels of genes involved in heme biosynthesis in co-cultured and monocultured S. sp. FXJ1.4098 in transcriptome data
Numbers in red highlight the up-regulated expression of the corresponding genes in the transcriptome analysis of co-cultured S. sp.FXJ1.4098 compared to the monoculture.FC, fold change.Fpkm, fragments per kilobase of transcript per million mapped reads.