Antibiotic-induced degradation of FtsZ reveals distinct stages of Bacillus subtilis FtsZ ring assembly and constriction

ADEP antibiotics induce the degradation of the cell division protein FtsZ, thereby primarily depleting the cytoplasmic FtsZ pool that is needed for treadmilling FtsZ rings. We here studied the effect of ADEP on FtsZ ring formation. Our data reveal the disintegration of early FtsZ rings during ADEP treatment, while progressed FtsZ rings finalize cytokinesis, thus indicating different roles for FtsZ treadmilling during distinct stages of divisome assembly and constriction.

an initial step involving FtsZ treadmilling and a second step which increasingly depends on peptidoglycan synthesis 10 .
To investigate the role of treadmilling in B. subtilis FtsZ ring formation further, we employed antibiotics of the ADEP class as tools to modulate the cytoplasmic pool of FtsZ. ADEP deregulates the bacterial caseinolytic protease, activating its dormant core ClpP for the untimely degradation of FtsZ 11,12 . ADEP incubation thus leads to an impressive filamentation phenotype of B. subtilis at concentrations close to the minimal inhibitory concentration (MIC) 12,13 . Very recently, we showed that ADEP-ClpP preferably targets the N terminus of monomeric FtsZ, leading to unfolding and degradation of the FtsZ N-terminal domain 14 . Intriguingly, N-terminal degradation was prevented upon nucleotide binding to FtsZ, most probably due to a stabilization of the FtsZ protein fold.
Therefore, at ADEP concentrations resulting in a filamentation phenotype, ADEP primarily leads to a depletion of the cytoplasmic pool of nucleotide-free FtsZ in the bacterial cell 14 , thus reducing the FtsZ concentration below the critical level needed for FtsZ ring formation 15,16 and continuously sequestering available FtsZ needed for treadmilling (Fig. 1a). Hence, ADEP is instrumental to investigate the role of the cytoplasmic FtsZ pool and treadmilling in FtsZ ring formation and dynamics. To do so, we first tested the effect of ADEP on polymerized FtsZ (in the presence of GTP) (Fig. 1b). Once assembled into protofilaments, FtsZ substantially resisted the degradation by ADEP-ClpP. It may thus be hypothesized that, if FtsZ ring assembly and constriction fully depend on FtsZ treadmilling, ADEP treatment should result in the disintegration of early as well as late stage FtsZ rings, or at least, the progression of late stage FtsZ rings should be halted. To test this hypothesis, we conducted time-lapse fluorescence and super-resolution microscopy experiments with ADEP-treated B. subtilis strain 2020, which expresses FtsZ fused to GFP using filamentation concentrations of the antibiotic. By following FtsZ ring formation over time, we observed that ADEP inhibited the initiation of FtsZ ring assembly, and early FtsZ rings that had just been formed disintegrated. In contrast, more progressed FtsZ rings constricted and finished septum formation, apparently being unaffected by ADEP treatment, finally yielding two separated daughter cells (Fig. 1cd, Supplementary Fig. S1 and S2). Following up on this, we investigated whether the arrival of the late-stage cell division protein PBP2b 17 , a septal peptidoglycan synthase, would coincide with a successful constriction of progressed FtsZ rings during ADEP-treatment. By using B. subtilis strain CM03, which allows for the concomitant expression of mCherry-FtsZ and GFP-PBP2b, we observed that the divisome consistently finalized cell division after PBP2b had substantially arrived at the septum area and had formed clear foci. Contrariwise, earlier FtsZ rings disintegrated prior to the observed arrival of PBP2b (Fig. 1e, Supplementary Fig. S3). Hence, our data imply distinct stages during FtsZ ring initiation, maturation and constriction. Obviously, during ADEP treatment it is distinguished between early and more progressed FtsZ rings, thereby adding another level of complexity to the elaborate mechanism of ADEP action. Also, our data suggest a two-step model of cell division in B. subtilis (Fig. 2), in which more progressed FtsZ rings are significantly less sensitive to a depletion of the cytoplasmic pool of nucleotide-free FtsZ, and thus less dependent on FtsZ treadmilling, in contrast to initial assembly and early stage FtsZ rings.

Protein purification
FtsZ and ClpP proteins were derived of B. subtilis 168 (trpC2; wild-type strain; NC_000964.3) and were expressed as C-terminally His6-tagged proteins in E. coli BL21(DE3) harboring the respective expression plasmid as described earlier 12 . Quantity and quality of the purified proteins were verified by Bradford assay (using bovine serum albumin as control), Nanodrop spectrophotometry (Nanodrop Technologies) and SDS-PAGE. We have shown previously that purification-tags have no effect on FtsZ degradation or enzyme activity of both proteins 14 .
After initial incubation for 4 min at 37 °C allowing baseline correction, GTP was added to the respective reaction mixture to a final concentration of 1 mM. Then, reaction mixtures were transferred into a photometer cuvette for monitoring light transmission at 400 nm and 37 °C over time. For in vitro degradation of polymerized FtsZ, 12 µM ADEP2 (or equal volume of DMSO as a control) and 1.5 µM purified ClpP protein (monomer concentration) were added to the polymerization reaction after 16 min, as indicated, and light transmission was further monitored for 34 min. As an independent control, FtsZ (4 µM) was incubated with 1 mM GTP in activity buffer (50 mM Tris/HCl pH 8, 25 mM MgCl2, 100 mM KCl, 2 mM DTT) at 37 °C for 30 min. Then 1.5 µM ClpP (monomer concentration) and 3.75 µM ADEP (or equal volume of DMSO as a control) were added to the reaction mixtures that were further incubated at 37 °C. Samples were taken after 120 min and were analyzed via SDS-PAGE using standard techniques as previously described 12,14 .

Cloning strategy
For colocalization studies of FtsZ with PBP2b (mCherry-FtsZ/GFP-PBP2b), strain B. subtilis CM03 was constructed as follows. Plasmid pJCM02 was generated using the coding sequence of pbpB that was amplified from chromosomal DNA of B. subtilis 168 (trpC2; wild-type strain;