Competition for iron shapes metabolic antagonism between Bacillus subtilis and Pseudomonas marginalis

Abstract Siderophores have long been implicated in sociomicrobiology as determinants of bacterial interrelations. For plant-associated genera, like Bacillus and Pseudomonas, siderophores are well known for their biocontrol functions. Here, we explored the functional role of the Bacillus subtilis siderophore bacillibactin (BB) in an antagonistic interaction with Pseudomonas marginalis. The presence of BB strongly influenced the outcome of the interaction in an iron-dependent manner. The BB producer B. subtilis restricts colony spreading of P. marginalis by repressing the transcription of histidine kinase-encoding gene gacS, thereby abolishing production of secondary metabolites such as pyoverdine and viscosin. By contrast, lack of BB restricted B. subtilis colony growth. To explore the specificity of the antagonism, we cocultured B. subtilis with a collection of fluorescent Pseudomonas spp. and found that the Bacillus–Pseudomonas interaction is conserved, expanding our understanding of the interplay between two of the most well-studied genera of soil bacteria.


Culturing
Growth was assessed by inoculating DK1042 or PS92 into the wells of a 96-well microtiter plate at a final OD600 = 0.001 and incubating the plate in a Synergy HTX multi-mode reader at 30 ℃ with shaking capturing optical density at 600 nm (OD600), msfGFP (Ex: 485/20 nm, Em: 528/20 nm), and mKate2 (Ex: 590/20 nm, Em: 635/32 nm) every 15 minutes.Colony forming units from liquid cultures were counted from cultures grown in 50 mL broth in 250 mL Erlenmeyer flasks. 1 mL culture was harvested from each sample, serially diluted, and spread on selective media.Four independent cultures were used as biological replicates, and three 1 mL volumes were taken from each flask as technical replicates.Iron limitation was achieved by adding 2,2-bypridine or by not adding FeCl2 to the media (e.g. with MSgg).Limitation of zinc in MSgg, was achieved by not adding ZnCl2.KBase consisted of 20 g/L peptone in deionized H2O.Glycerol, K2HPO4, MgSO4•7H2O, and NaCl was added into the media before autoclavation, in the concentrations specified for KB (or 5 g/L in the case of NaCl). .Pairwise interactions on agar DK1042 and PS92 were routinely spotted 5 mm apart on agar surfaces using 2 µL overnight culture adjusted to an optical density at 600 nm (OD600) of 1.0 in appropriate media.Prior to spotting, plates were dried for 30 min in a lateral flow hood and incubated at 30°C.This experimental setup was employed for all coculture colony assays with multiple types of media and strains.When spotting mixed cultures, each species was adjusted to an OD600 of 1.0 and the two strains were then mixed in equal volumes.For mixed B. subtilis cultures, wild-type (WT) and mutant strains were mixed 1:1, 1:10, or 1:100, based on OD600, before spotting 2 µL as detailed above.

Timelapse microscopy
Timelapse series were acquired with the Carl Zeiss Axiozoom V.16 stereomicroscope (see main text) at 15-minute intervals for 72 hours.Bacterial cultures were spotted on agar solidified in a 35 mm petri dish which was then placed into a Tokai Hit stage top incubator pre-warmed to 30 °C.Sterile MilliQ water was added to the incubator water bath to maintain constant humidity in the chamber.

Transcriptomics
Samples were bead-beaten at 4500 RPM in a FastPrep for 30 seconds, and otherwise subjected to the manufacturer's protocol.DNA was digested with TURBO DNase following the protocol for rigorous DNA digestion.RNA concentration was assessed on a Qubit fluorometer using the Qubit High Sensitivity kit (concentrations ranging from 20-100 ng µL -1 ), and RNA integrity was assessed on an Agilent Bioanalyzer 2100 using Agilent RNA 6000 nano kit.All samples had an RNA integrity number (RIN) > 7.2.

Figure S1 :
Figure S1: Interaction is constrained to solid King's B. a: Colonies grown on LB for 72h at 30 ℃. b: ΔdhbA spotted next to DK1042 on King's B and LB and grown at 30 ℃ for 72h.c: Non-normalized colony-areas visualized in Figure 1.p-values are from Student's t-test adjusted for multiple testing by the Benjamini-Hochberg method.d: Growth in liquid King's B and LB.Strains were mixed in 96-well plates in equal starting density and growth was measured on a microplate reader with fluorescence and optical density.Starting densities were adjusted to be identical in monoculture and coculture experiments RFP: mKate2 signal (Bacillus), GFP: msfGFP signal (Pseudomonas).e: Viable cell count from planktonic cultures grown in Erlenmeyer flasks.

Figure S2 :
Figure S2: Interaction depends on medium composition.a: Colony area of Bacillus and Pseudomonas in interactions on MSgg with and without supplemented metals.Interactions were cultivated at 30 ℃ for 72h.b: Relative colony size of Bacillus and PS92 cultivated on King's B with or without supplemented Mg 2+ , K + , or glycerol (G).KBase is water and 20 g/L peptone.Dashed lines and squares are median, 1 st , and 3 rd quantile, respectively for DK1042 + PS92 (left) and ΔdhbA + PS92 (right) on regular King's B. c and d: Stereomicroscopy of representative samples from a and b. e: Stereomicroscopy of representative samples grown on LB with either supplemented glycerol (1% v/v, LBG), Mn 2+ (100 µM, LBM) or both (LBGM).

Figure S3 :
Figure S3: LC-MS in solid and liquid King's B and LB.a: Base peak chromatogram of DK1042 grown at 30 ℃ for 72h.Base peak intensity is measured on all eluted molecules over time.Highlights are approximate elution times for the indicated molecules (other molecules can elute at the same time point).b: Integrated peak area of select secondary metabolites measured in their respective ion chromatograms extracted by m/z-value.p-values are from Student's t-test adjusted for multiple testing by the Benjamini-Hochberg method.c: Same as b, but with Bacillibactins measured from their UV chromatograms.BB: Bacillibactin, Bae: Bacillaene, Srf: Surfactin.

Figure S4 :
Figure S4: DK1042 can complement bacillibactin-deficient mutants.a: Stereomicroscopy of single gene deletion mutants deficient in genes encoded in the dhb operon cultivated on King's B at 30 ℃ for 72h either alone or next to PS92.See Figure 2b.b: Stereomicroscopy of dhb mutants mixed in equal ratios and their interactions as in a. See Figure 2c.c: Relative colony size of Bacillus and PS92 cultivated as a and b.Dashed lines and squares are median, 1 st , and 3 rd quantile, respectively for DK1042 + PS92 (grey) and ΔdhbA + PS92 (red).d: Stereomicroscopy of representative samples from c.

Figure S5 :
Figure S5: RNAseq quality checks.a and b: Principal component analysis of transcript reads in each sample for Bacillus (a) and Pseudomonas (b).c: Theoretical number of reads per coding sequence (CDS) in each genome for each sample.Calculated as nreads/nCDS.d and e: log-transformed read count for genes of interest highlighted in Figure 4a and 4b for Bacillus (d) and Pseudomonas (e).

Figure S6 :
Figure S6: Viscosin is not responsible for Bacillus antagonism.Close-proximity colonies of DK1042 and Pseudomonas fluorescens SBW25 wild type and ΔviscA deficient in viscosin production.Colonies were cultivated on King's B at 30 ℃ for 72h.Mass spectrometry imaging reveals the presence/absence of select metabolites as well as their spatial localization in the interactions.Grey value is root mean squared intensity across all samples, and lookup tables were scaled identically across each metabolite.NaN pixels were set to zero.Metabolites were annotated with metaspace using the 2019 Natural Product Atlas database.m/z = 1162.5336was manually annotated as a pyoverdine structure (this m/z-value is different from the pyoverdine structure in figure4).