Tunable force transduction through the Escherichia coli cell envelope

Abstract

The outer membrane (OM) of Gram-negative bacteria is not energised and so processes requiring a driving force must connect to energy-transduction systems in the inner membrane (IM). Tol (Tol-Pal) and Ton are related, proton motive force- (PMF-) coupled assemblies that stabilise the OM and import essential nutrients, respectively. Both rely on proton-harvesting IM motor (stator) complexes, which are homologues of the flagellar stator unit Mot, to transduce force to the OM through elongated IM force transducer proteins, TolA and TonB, respectively. How PMF-driven motors in the IM generate mechanical work at the OM via force transducers is unknown. Here, using cryoelectron microscopy, we report the 4.3Å structure of the Escherichia coli TolQR motor complex. The structure reaffirms the 5:2 stoichiometry seen in Ton and Mot and, with motor subunits related to each other by 10 to 16° rotation, supports rotary motion as the default for these complexes. We probed the mechanism of force transduction to the OM through in vivo assays of chimeric TolA/TonB proteins where sections of their structurally divergent, periplasm-spanning domains were swapped or replaced by an intrinsically disordered sequence. We find that TolA mutants exhibit a spectrum of force output, which is reflected in their respective abilities to both stabilise the OM and import cytotoxic colicins across the OM. Our studies demonstrate that structural rigidity of force transducer proteins, rather than any particular structural form, drives the efficient conversion of PMF-driven rotary motions of 5:2 motor complexes into physiologically relevant force at the OM.

Keywords: cell envelope; force transduction; gram-negative bacteria; outer membrane.

Fig. 1.

Outward rotation relates motor complex structures. (A) Single-particle cryo-EM map of the TolQR complex solved to 4.27 Å. Monomers are coloured individually and the position in the IM is indicated by gray rectangle. (B) An atomic model of TolQR was determined through rigid-body refinement into the map in (A). TolQ forms a pentameric structure within which a TolR (gray) dimer is embedded. Only the TPHs of TolR were sufficiently resolved in the cryo-EM map. (C) Alignment of pentameric structures of ExbB and MotA to TolQ pentamer (transparent), in ChimeraX. A single transmembrane helix common to ExbB (green), TolQ (blue), and MotA (orange) is shown which is related through clockwise rotation along the plane of the membrane. (D) A single chain of MotA, ExbB, and TolQ was aligned and single transmembrane helices (transparent cartoon) containing conserved Thr residues (Sticks) are shown. Conserved Thr residues are predicted to play a role in the proton transduction pathway and are accordingly situated in the same region of TM helices in all three motor complexes. (E) A single monomer of the ExbB (transparent) pentamer was aligned with the best-fitting chain of TolQ (green) in ChimeraX. Fit of remaining TolQ monomers (coloured) to ExbB (gray) is shown sequentially without any further realignment. Between the structures, minor perturbations of accessory helices on the periplasmic side and transmembrane helices occur, with the most pronounced distortions between the two proteins observed in the position of accessory helices on the cytoplasmic side of the pentamers.

Fig. 2.

Colicin cytotoxicity assays demonstrate that domain II chimeras drive bacteriocin transport across the OM. (A) Domain II-swapped chimeras of TolA (green) and TonB (pink) were generated, as well as domain II knock-outs (ΔII) and substitution with an intrinsically disordered sequence of 127 amino acids (gray ribbons: AIA/BIB, respectively). Gray bars indicate domain swap connections. Group A constructs are defined by having a C-terminal TolAIII and group B constructs by TonBIII. The essential histidine residue of the SHLS motif within transmembrane helices of TolA and TonB (His22/20) is shown. (B) Controls confirming Tol and Ton dependence for colicins E9 and Ia, respectively. A fourfold dilution series of ColE9 (from 8 µM) or fivefold series of ColIa (from 10 µM) was applied to a bacterial lawn, where clearance zones indicate cell killing. (C) Plasmid complementation expressing wild-type TolA and TonB restores ColE9- and CoIA-mediated killing in ΔtolA (Top row) and ΔtonB (Bottom row) cells, respectively. Constructs with domain II deletions were unable to uptake colicins, but domain II swapped constructs conferred colicin sensitivity (ABA/BAB), as did constructs with disordered central domains to a lesser extent (AIA/BIB). Mutagenesis of the SHLS motif in TolA_H22A and TonB_H20A mutants abrogates killing (–56).

Fig. 3.

OM stabilisation by Tol requires a structured TolA domain II. (A) Group A constructs demonstrate a continuum of OM stability. Full complementation (wild-type growth), on 2% SDS (w/v) graded “+++”, moderate growth on 0.4% SDS (w/v) rated “++”, and low growth on 0.4% SDS rated “+”, in accordance with additional mutants which demonstrated intermediate SDS-tolerance (SI Appendix, Fig. S8A). (B) Group A constructs exhibit a spectrum of Pal accumulation, which correlates with observed OM stability. ΔtolA Pal-mCherry cells were transformed with group A construct expression plasmids, grown to mid-log phase, and then imaged by transillumination (Top) and fluorescence microscopy (Centre). (The scale bar indicates 5 µm.) Fluorescence profiles of dividing cells were obtained by plotting medial fluorescence along the cell length axis, which was then averaged per construct type (Bottom). The central line indicates the mean fluorescence value along the cell length axis, and shaded bands indicate ± SD. Relative SDS tolerance levels from panel A are indicated Top Left, and number of dividing cells averaged per profile is also indicated Top Right. (Inset) Western blot bands indicate the expression levels of constructs (SI Appendix, Fig. S9), and micrographs indicate vesiculation of ΔtolA cells ± AIA.

Fig. 4.

Colicin transport across the OM correlates with the efficiency of force transduction mediated by TolA domain II. Time-kill assay shows rate of uptake varies with domain II structure and appears to decrease as a consequence of intrinsically disordered domain II. Cells were treated with 100 nM ColE9 and incubated in LB media at 37 °C with shaking; then, aliquots were incubated with trypsin at indicated time points. The number of colony-forming units (CFU) was used to assess the number of live cells in suspension, where the decrease in viable cells over time reflects the rate of cell killing. Mean ColE9 translocation half-lives for different constructs were calculated from the logarithmic slope, as 4.1 min (TolA), 7.8 min (ABA), and 50.4 min (AIA). 95% CIs of mean translocation half-life were 3.7 to 4.6 min (TolA), 7.2 to 8.5 min (ABA), and 44.3 to 58.5 min (AIA). N = 3 biological replicates per sample, each comprising 3 technical repeats. Error bars indicate the SEM.

Fig. 5.

Efficient force transduction to the OM for Pal-mediated OM stabilisation and invagination requires secondary structure in TolAII. (A) TolA rapidly imports colicins. I) TolA engages TolQR below the PG layer. 2) TolQR energises conformational rearrangements within TolAII to apply tension to the TolB-ColE9 complex (large arrow) and entry of unfolded ColE9 to the periplasm from where it refolds and translocates to the cytoplasm to cause cell death through nuclease action (51). The precise conformational changes that take place during force transduction are currently unknown but may, for example, involve a combination of TolA refolding and/or wrapping about the stem of TolR, together transducing a torsional force perpendicular to the OM. (B) AIA imports colicins slowly. 1) AIA interacts with TolQR. 2) Force transduction via AIA occurs less efficiently than via TolA (small arrow). This is sufficient to import some cytotoxic colicins, but their passage across the OM is significantly slower than in wild-type TolA cells. (C) TolA rapidly recruits Pal to the division site, stabilising and invaginating the OM at the septum. I) TolQRA is trafficked to the division septum to capture mobilised TolB-Pal complexes. 2) TolA efficiently dissociates TolB-Pal complexes, depositing many Pal monomers which stabilise the OM. (D) AIA exhibits poor Pal recruitment at the septum, which fails to stabilise the OM. 1) AIA is a weak mechanotransducer in the forced dissociation of the TolB-Pal complex. 2) The inefficiency of AIA force transduction slows Pal accumulation at the septum, resulting in OM destabilisation and blebbing.