Media and growth conditions

B. subtilis, B. megaterium, and E. coli strains were grown at 37 °C in Luria-Bertani (LB) media, supplemented, when needed, with ampicillin (Amp, 100 μg/ml), spectinomycin (Spc, 100 μg/ml), chloramphenicol (Cm, 5 μg/ml), tetracycline (Tet, 10 μg/ml), kanamycin (Kan, 10 μg/ml), and MLS (lincomycin 25 μg/ml and erythromycin 1 μg/ml). One milimolar IPTG (isopropyl-d-thiogalactopyranoside, Amresco) was added where indicated. D. radiodurans was grown in TGY medium (1% Bacto Tryptone; 0.5% yeast extract; 0.1% glucose).

Bacterial strain construction

Bacterial strains are listed in Supplementary Table 1, and primers used for the validation of gene deletions are listed in Supplementary Table 2. Genomic DNAs (gDNAs) were isolated using the High Pure PCR Template Preparation Kit (Roche) and PCR reactions were carried out using an Expand High Fidelity PCR System (Roche).

SEM procedure

Sample preparation and processing were carried out essentially as in⁵⁶ but with some modifications. In brief, exponential cultures of B. subtilis strains (OD₆₀₀ ~ 0.6) were pre-fixed with 3% glutaraldehyde in culturing media at room temperature (RT), washed with cacodylate buffer and fixed with 3% glutaraldehyde in cacodylate buffer at 4 °C overnight. These extensively washed cells were then sedimented onto poly-l-lysine-treated circular glass coverslips at 4 °C in a Petri-dish moist chamber for 42 h. The cells attached to the coverslips were washed three times and postfixed in 1% OsO₄ for one hour at room temperature and again washed three times. The coverslips were dehydrated in a graded ethanol series (25, 50, 75, 90, 96 100, and 100%) followed by absolute acetone and critical point dried in a K850 Critical Point Dryer (Quorum Technologies Ltd, Ringmer, UK). The dried samples were sputter-coated with 3 nm of platinum in a Q150T Turbo-Pumped Sputter Coater (Quorum Technologies Ltd, Ringmer, UK). The final samples were examined in a FEI Nova NanoSEM scanning electron microscope (FEI, Brno, Czech Republic) at 3 or 5 kV using ETD, CBS, and TLD detectors and using SEM software Helios NanoLab. The beam deceleration mode was used when sample charging occurred⁵⁷.

SIM procedure

For membrane staining, Nile Red was added to 1 ml of exponentially growing culture at a final concentration of 10 μg/ml. After 10 min of incubation at RT, bacteria were pelleted and washed once with 1× PBS. For staining of flagella, Alexa Fluor 488 maleimide conjugate (5 μg/ml, final concentration) was added and samples were washed twice with 1× PBS, then stained with Nile Red. The bacteria were subsequently re-suspended in 1× PBS, spotted on a coverslip covered with a thin agarose pad (1.5% agarose in 1× PBS supplemented with 10× diluted LB medium). Alternatively, for pressure experiments, samples were spotted on poly-l-lysine glass slides and covered with a poly-l-lysine coated coverslip. The samples were observed using a DeltaVision OMX™ equipped with a 60× 1.42, PlanApo N, oil immersion objective and softWoRx™ Imaging Workstation software. ZsGreen and GFP-tagged protein or Alexa Fluor maleimide conjugate were imaged using 488 nm excitation; Nile Red was imaged using 568 nm excitation.

Time-lapse fluorescence microscopy

To visualize the membrane structures, 1 ml of exponentially growing bacterial culture was stained by Nile Red (10 μg/ml) or by FM4–64 (1 μg/ml), washed twice with 1× PBS and re-suspended in 1× PBS containing 1 μM SYTOX green. The sample was then immediately either spotted onto a poly-l-lysine glass slide and covered with a poly-l-lysine coverslip (GLG method or P-GLG method when pressure ~80 kPa was applied for 10 s) or spotted on a 1× PBS agarose pad covered with a non-coated coverslip (GAG method). Pictures were obtained at the indicated time points (t = 0 is the start of microscopy, typically 30 s after coverslip addition) using a Olympus BX63 fluorescence microscope equipped with a Andor Zyla 5.5 sCMOS camera (alternatively, an Olympus IX81 microscope equipped with Hamamatsu Orca/ER camera was also used). Olympus CellP imaging software or Olympus Image-Pro Plus 6.0 software was used for image acquisition and analysis.

MS analysis of cytoplasmic membrane fractions

Cytoplasmic membrane fractions from the B. subtilis wt (LK1432) and ΔsigD (LK1873) strains were prepared following a previously described protocol with some modifications¹¹. Briefly, cells were gently shaken and grown until they reached exponential phase (OD₆₀₀ = 0.6) and then pelleted (6000 × g, 10 min at 4 °C). Samples were re-suspended in 20 ml of 1× P buffer (35% NaCl; 35% Na₂HPO₄; 10% glycerol; v/v) containing 3 mM 2-mercaptoethanol and 1 mM serine protease inhibitor PMSF (phenylmethylsulfonyl fluoride) and sonicated on ice for 10× 10 s, with probe amplitude 0.5, (Hielscher sonicator, UP 200 s). The lysates were then centrifuged (6000 × g, 20 min at 4 °C), the supernatants were transferred to clean ultracentrifuge tubes, and the membranes were collected by centrifugation at 100,000 × g for 45 min at 4 °C. The supernatants were subsequently discarded and the pellets containing the membrane fractions were dissolved in 1× P buffer and analyzed by MS.

The proteins were analyzed by trypsin filter digestion in the form of an eFASP (enhanced Filter-aided sample preparation) method⁵⁸. Briefly, the samples were reduced, alkylated and digested by trypsin on YM-10 Microcon filters (Merck). The resulting peptides were then desalted on a C18 SPE column (PepClean, Thermo). The peptides were separated during LC-MS/MS using a nano-LC system (Ultimate 3000 RSLC nano, Dionex) on an Acclaim PepMap C18 column (75 um Internal Diameter, 250 mm length) by applying a 125 min acetonitrile elution gradient in 0.1% formic acid. The chromatographic column was connected via nanoESI to a tandem mass spectrometer (TripleTOF 5600, Sciex). We collected data for the identification and the quantification of the proteins in one measurement sequence. Employing two methods using the same chromatographic parameters but different mass spectrometric data collection setups [the first: the data-dependent analysis (DDA) method; the second: the data independent method (DIA)—SWATH⁵⁹], we measured the samples in a sequence of two consecutive runs of each sample (the first run aimed at collecting the DDA data to identify the proteins; the second aimed at collecting the DIA data for to quantify them). All resulting DDA spectra were searched together using Protein Pilot 4.5 (Sciex) against the Uniprot B. subtilis reference protein database (downloaded 8th of October 2015), thus creating the library used for SWATH processing in PeakView 2.2 (Sciex). In the SWATH analysis the retention time was aligned in all the samples by selecting the peptides common across the retention time range for all the samples. For quantification purposes, we allowed up to 30 peptides per protein, 6 transitions per peptide, a peptide confidence of 95% and a false-discovery rate threshold of 1%. The processing steps produced an intensity of transitions, peptides and proteins. The final protein table was processed in MarkerView (Sciex) in order to create a statistical evaluation. The Student’s t-test was performed on the monitored groups.

Spotting assay—non-conjugative plasmid transfer

Donor [LK1925, wt containing the Cm resistance gene on plasmid pCPP31-Y1)³⁹] and recipient (LK1922, wt harboring the MLS resistance gene on chromosomal DNA) strains were grown separately overnight. Next day, the strains were mixed (equal numbers of cells), diluted to an initial OD₆₀₀ = 0.05 and cultivated for 4 h at 37 °C with shaking. The mixtures, serially diluted, were spotted on LB agar (without antibiotics) and LB agar containing appropriate antibiotics (Cm for donor selection, MLS for recipient selection, Cm+MLS for recipients that acquired the non-conjugative plasmid). After 18 h, the dishes were photographed and double-resistant colonies were inoculated into LB to OD₆₀₀ = 0.05. 100 μl of bacterial suspensions were plated on LB agar with or without antibiotics (the same as above). Double-resistant colonies were inoculated into LB medium with antibiotics (Cm + MLS) and grown overnight at 37 °C. On the following day, their genomic (High Pure PCR Template Preparation Kit) and plasmid (see: B. subtilis plasmid isolation) DNAs were extracted. The presence of the erm gene (MLS^r) in double-resistant strains was verified by PCR (Expand High Fidelity PCR System).

B. subtilis plasmid isolation

Two milliliter of overnight culture was pelleted (5 min, 13,200 × g at 4 °C) and washed twice with 500 µl of TES (0.02 M, Tris-HCl pH 8; 5 mM, EDTA; 0.1 M, NaCl). The supernatant was discarded and 40 µl of lysozyme-containing buffer (30 mM, Tris-HCl pH 8; 50 mM, EDTA; 50 mM, NaCl; 25%, sucrose; lysozyme 500 µg/ml) was added. After 15 min at 37 °C, the sample was placed on ice. 160 µl of SDS buffer (2 ml of 10% SDS; 2 ml of 0.5 M EDTA pH 8; 12 ml of TES buffer) was mixed with 50 µl of 5 M NaCl and added to the sample, gently mixed and left on ice for 60 min. The mixture was subsequently centrifuged (30 min, 13,200 × g at 4 °C) and the supernatant containing the plasmid DNA was transferred to a clean micro centrifuge tube. Finally, the plasmid DNA was precipitated with ethanol and analyzed on an agarose gel.

Non-conjugative plasmid transfer

Donor (harboring plasmid pCPP31-Y1, Cm^r) and recipient (containing MLS^r in genome) strains were grown separately overnight. The next day, cultures were pelleted (9000 × g, 3 min at RT) and dissolved in 1× DNAse I buffer diluted in 1× PBS. DNAse I was then added (or not for the negative control) and samples were incubated for 20 min at 37 °C. Donor and recipient strains were then mixed (or each strain was grown separately) at a 1:1 ratio at an initial OD₆₀₀ = 0.05 into fresh LB medium without antibiotics. Cultures were gently shaken for 4 h at 37 °C. Subsequently, bacteria were diluted to OD₆₀₀ = 0.5 and 100 μl were plated on LB agar dishes with appropriate antibiotics (MLS-recipient selection; Cm + MLS—selection for double-resistant bacteria). CFU/ml were counted on the following day. The efficiency of plasmid receipt by recipient cells was expressed as the ratio of double-resistant cells (those able to grow on Cm + MLS) to all recipient cells in the mixture that grew in the presence of MLS.

Quantitative PCR

Two milliliters of exponentially growing cells (LB medium with 200 µM IPTG, gentle shaking) [wt (LK1432), ΔsigD (LK1873), ΔcomK (LK2380), P_hs-comK (LK2317); OD₆₀₀ ~0.8] were treated with RNAprotect Bacteria reagent (QIAGEN), pelleted and immediately frozen. Their total RNA was isolated with the RNeasy Mini Kit (QIAGEN). Prior to RNA extraction, recovery marker RNA was added [a fragment of 16 s rRNA from M. smegmatis (amplified by primers nos. LK1281 and LK1282, see Supplementary Table 2)]. Finally, RNA was DNase treated (TURBO DNA-free™ Kit, Invitrogen). Five micrograms of total RNA was reverse transcribed to complementary DNA with reverse transcriptase using random hexamers as primers (SuperScript™ III Reverse Transcriptase, Invitrogen). This was followed by qPCR in a LightCycler 480 System (Roche Applied Science) containing LightCycler® 480 SYBR Green I Master and 0.5 μM of each primer. The primers were designed with Primer3 software and their sequences are given in Supplementary Table 2. The data were normalized to the recovery marker and the number of cells.

Growth of B. subtilis in the presence of antibiotics

One milliliter of B. subtilis exponential culture was stained with Nile Red (final concentration 10 μg/ml) for 10 min at RT and washed once with 1× PBS. LB agarose containing 1 µM SYTOX green and an appropriate antibiotic (Amp, 500 μg/ml; Cm, 5 μg/ml; Rif, 50 μg/ml) was prepared inside a gene frame (Invitrogen). For time-lapse experiments, it was necessary to create a narrow agarose strip (~2.5 mm) in the middle of the gene frame (by removing agar from its sides) to allow oxygen to efficiently diffuse through the sample. The Nile Red-stained culture was then spotted on agarose and covered with a coverslip that was held by the gene frame (no pressure) and placed in a chamber at a constant 37 °C temperature. Pictures were taken with an Olympus CellR IX81 detection and analysis system equipped with a Plan-Apochromat 100×/1.45 NA oil objective and an EMCCD Hamamatsu camera.

Diverse methods for high-resolution bacterial culture observation

Exponential bacterial culture (OD₆₀₀ = 0.6) was diluted to low density (OD₆₀₀ = 0.05) and spotted on cellophane lying on an LB agar dish and grown for 4 h. In all, 4 × 2 cm cellophane strips containing bacteria were carefully taken and mounted onto glass slides using Scotch tape. The slides were placed into a glass desiccator with a small container of 2% OsO₄ in double-distilled water (ddH₂O) and the cells were allowed to fix at room temperature for several days⁶⁰. The fixed cellophane strips were then cut into pieces in size of standard SEM mounts (12.5 mm) and mounted with conductive tape onto aluminum mounts (SPI). The samples were then sputter-coated with 3 nm of platinum. The SEM analysis was essentially done as described above.

As an alternative to the cellophane SEM experiment, we used a SIM-based approach. Here, instead of cellophane, a thin ~1 mm 1× PBS agarose pad was used. The pad was positioned on an LB agar dish containing 10 μg/ml Nile Red. After 4 h, the agarose pad was removed, covered with a coverslip and observed by SIM (see: SIM procedure).

Subsequently, two previously published protocols for observing NTs were used with minor changes^1,9,11. First, bacterial cells were grown directly on LB agar dishes for 4 h at 37 °C and then imprinted onto glow-discharge activated EM grids⁶¹. The grids were fixed with drops of 2.5% glutaraldehyde in sodium cacodylate buffer in Petri-dish for 20 min¹. The fixed grids were washed three times with ddH₂O and dehydrated in a graded ethanol series (25, 50, 75, 90, 96, 100, and 100%), and air-dried directly from 100% ethanol. Finally, the grids were sputter-coated with 3 nm of platinum and mounted into a transmission electron microscopy grid table for SEM examination. A FEI Nova NanoSEM scanning electron microscope (FEI, Brno, Czech Republic) at 5 kV using ETD, CBS, and TLD detectors in beam deceleration mode was used for SEM analysis.

Second, exponentially growing cells were spotted at OD₆₀₀ = 0.05 onto glow-discharge activated formvar/carbon film coated EM grids placed on a nitrocellulose membrane¹¹. These grids were taken after 4 h of cultivation (at 37 °C) on an LB agar dish and fixed with drops of 3% glutaraldehyde in a petri-dish most chamber. The fixed grids were subsequently processed as described above.

Image analysis

The final adjustment of the fluorescent images was done using Fiji ImageJ or the Analysis 3.2 Pro software suite (Sis GmbH; Olympus, now EMSIS GmbH) for SEM images. The free software Gimp (https://www.gimp.org/) and Inkscape (https://inkscape.org/) were also used for final image plate setup.

NTs vs. cell membrane signal calculation

Measurement of the membrane signal in the time-lapse images was done following bleaching correction. NT and cell membrane signals were measured at each time point. The background value was subtracted from the NT and cell membrane signal, yielding the final signal intensity.

SYTOX green penetration

The SYTOX green signal was measured at each time point in the time-lapse images. The background was subtracted and the highest value was normalized to 1. Combination with the membrane staining signal allowed the NT formation time point to be detected.

Quantification and statistical analysis

For each experiment we had at least three biological replicates. Averages of individual cells from different replicas are reported. The number of analyzed cells is given in the charts and figures. NT quantification was done manually. MS Excel and SigmaPlot were used for all statistical analyses, data processing, and presentation. For testing uniformity of distribution of NTs over cell we assumed the auxiliary hypothesis that poles cover <50% of cell surface. p-value for this joint hypothesis was computed using quantiles of the Beta distribution. Comparison of cell counts was performed with a binomial generalized linear model using the glm function in R⁶² and—as an additional check—the stan_glm function from the RStanArm package⁶³. Analysis of plasmid transfer and relative expression was performed on the log scale using a linear model with the lm and cor.test functions in R. Where applicable, contrasts and multiple testing corrections were performed with the TukeyHSD function. The difference of spatial distributions of NTs between wild type and Ampicilin-treated bacteria was assessed with Chi-squared test using the chisq.test function in R. Where applicable, we used models with full interactions. Whenever multiple model formulations were considered, we have reported the p-value least favorable to our conclusions. See “Code availability” for details regarding code for the statistical analysis.

Reporting summary

Further information on research design is available in the Nature Research Reporting Summary linked to this article.