Iron Homeostasis in Bacillus subtilis Requires Siderophore Production and Biofilm Formation

Iron acquisition is of fundamental importance for microorganisms, since this metal is generally poorly bioavailable under natural conditions. In the environment, most bacteria are found tightly packed within multicellular communities named biofilms. Here, using the soil Gram-positive bacterium Bacillus subtilis, we show that biofilm formation and the production of siderophores, i.e., small molecules specifically binding metals, are both essential to ensure Fe uptake from the medium and maintain cellular Fe homeostasis. The biofilm matrix appears to play an important role favoring the efficient usage of siderophores. Taken together, our results demonstrate a close link between biofilm formation and iron acquisition in B. subtilis, allowing a better comprehension of how bacteria can cope with metal limitation under environmental conditions.

2. Cells need to be isolated from the biofilm matrix, since exopolysaccharides from the matrix are likely metal chelators that would impact the evaluation of metal content. 3. Cells need to be submitted to repeated oxalate-EDTA washes in order to get rid of Fe precipitated on the cell surface (1).
Consequently, the following protocol was established: At regular time intervals during growth in multiwall plates, cells were harvested in 15mL vials and pelleted by centrifugation (Beckman CoulterTM Adventi centrifuge J-25I with JLA 16,250 rotor, 6500 xg, 20°C, 7 min). The cell-containing pellets are then immediately suspended in 1mL of a 4% paraformaldehyde (PF) solution (4% paraformaldehyde in PBS) and incubated 7 minutes at room temperature. This step allows fixation of the cells, and conservation of the fluorescent proteins integrity. The fixed cells were then pelleted by centrifugation (6500 xg, 20°C, 7 min). The pellet was re-suspended in 1mL oxalate/EDTA solution (0.1 M/ 0.05 M) and incubated at room temperature for 7 minutes, both to remove external metals that could be trapped in the biofilm matrix and to wash the leftover paraformaldehyde. Cells were again pelleted by centrifugation (6500 xg, 20°C, 7 min), and then re-suspended in 5 mL of NaCl 0.5 M. The NaCl suspension of cells was separated in two, in order to do parallel analysis on the same sample: flow cytometry (A) and elemental analysis (B). In the case where only the elemental analysis is performed, the whole 5mL of cells in NaCl is treated according to (B).
(A) 1 mL of cells of the cells in NaCl suspension was centrifuged on a Fisher Scientific AccuSpin Micro17 (13,300 xg, 2 min). The pellet was the suspended in 1 mL of GTE solution (50 mM Glucose, 10 mM EDTA, 20 mM Tris pH8) and stored at 4°C for a maximum of 3 days before flow-analysis.
(B) 4 mL of cells in NaCl was sonicated on a Q125-Sonicator (power 20%, 10 pulses of 1 sec with 1 sec pause) to detach the clumped cells. Then, NaOH was added to a final 0.1M concentration to help solubilize the biofilm matrix at RT for 5min (adapted from (2). Cells were then separated from the biofilm matrix, now soluble, by centrifugation (6500 xg, 20°C, 7 min). Pelleted cells were then washed once with 1mL oxalate/EDTA (0.1 M/ 0.05 M), pelleted (6500 xg, 20°C, 7 min), and washed again with 1mL of a diluted oxalate/EDTA solution (0.025M/ 0.0125M) solution. Cells were finally pelleted by centrifugation (6500 xg, RT, 7 min), and the pellets were then store at 4°C until elemental analysis.

Controls performed for method validation: cell lysis
Controls were performed to validate that the various steps during cells isolation and preparation did not affect the cellular metal content. First, we evaluated if sonication and/or NaOH treatment caused cell lysis and leaking of metals. These controls were performed with the biofilm mutant epsA-O tasA to avoid a possible metal-chelating effect of the biofilm matrix, which would bias the results. Briefly, 9 replicates were treated as followed (A) 3 replicates were fixed with PF, and washed twice with oxalate-EDTA.
(B) 3 replicates were fixed with PF, washed once with oxalate-EDTA, resuspended in NaCl and sonicated as describe above, and washed again with oxalate-EDTA.
(C) 3 replicates were fixed with PF, and washed once with oxalate-EDTA, resuspended in NaCl, treated with NaOH, and washed again with oxalate-EDTA All 9 replicates were then digested on a SCP Science Digiprep Jr with 1mL of nitric acid at 65°C for 45min. After digestion, each tube was filled at 10mL with Milli-Q water, and samples were analyzed for phosphorus, manganese, magnesium and iron content on an inductively-coupledplasma mass spectrometer (ICP-MS; Thermo Scientific XSerie2). All samples showed similar amount of metals ( Fig S10). Since there was no difference between the fixed (A), fixed and sonicated (B), fixed, sonicated and NaOH treated cells (C), we conclude that the treatments do not cause cell lysis.

Controls performed for method validation: matrix solubilization
Controls were performed to examine if our treatments efficiently get rid of the biofilm matrix and of a putative biofilm matrix-chelator effect. 9 replicates of WT cells were grown in MSgg during 25h to obtain a mature biofilm, and collected by centrifugation. At this step, 3 supernatants were kept as is (M1). Cells were then fixed with PF and pelleted; after this 3 supernatants (S1) and 3 pellets were kept (T1). Remaining 6 cell pellets were resuspended in NaCl and sonicated as described, after which 3 replicates were kept (T2). 2 mL NaCl and 500uL NaOH 1M were added to the 3 last replicates, and cells were collected by centrifugation. After this centrifugation step, supernatants (S3) and cell pellets (T3) were kept. Each sample was observed by microscopy for cell-clumping (indicative of leftover biofilm) and tested for chelating properties by CAS and at 65°C for 45 min. After digestion, each tube was filled at 10 mL with Milli-Q water, and analyzed for phosphorus on an inductively-coupled-plasma mass spectrometer (ICP-MS; Thermo Scientific, XSeries2) as previously described (3). Cell number was plotted in function of intracellular phosphorus ( Figure S1), and shows a strong linear correlation.  Table   S2, S3, and S4.      Phosphorus quantification was used as a proxy for cell quantification (Fig S1). Results are the mean of the three biological replicates and error bars represent standard deviation.