Table 1.
Bacterial strains used in this study.
Table 2.
Plasmids and primers used in this study.
Fig 1.
Generation of an in silico BacM model.
(A) The CD spectrum of isolated purified BacM fibers indicates a β-structure-only protein devoid of α-helices. (B) The observed peak at 1631 cm-1 of the amide I region of the FTIR spectrum of purified BacM fibers is likewise characteristic of β-sheets. No indications of α-helices are found as the peak between 1650 and 1655 cm-1 in the Fourier self-deconvoluted (FSD) spectrum disappeared upon deuteration. This behavior is indicative for unstructured regions. The shift of the major β-structure peak from 1631.9 cm-1 to 1625 cm-1 upon deuteration is characteristic of an “extended β-sheet” which is the result of intermolecular β-structure-like interactions of the polymerized BacM. (C) Ribbon diagrams of various BacM models generated by five different molecular modeling servers. All models predict that nearly the entire protein is formed by a left-handed β-helix with the non-structured flexible C-terminus variously oriented with respect to the barrel-shaped molecule. No α-helices are present in the models of the protein.
Fig 2.
Bactofilins contain repeat regions with conserved hydrophobic residues within β-strands.
(A) Alignment of the repeat regions in the 4 bactofilin paralogs of M. xanthus identified by HHrepID reveal that each paralog contains 6 repeats (top); Sequence logo showing conservation of residues at each position in the repeat, generated using the server at weblogo.berkeley.edu (bottom). (B) Repeat regions of BacM are aligned, revealing the highly conserved hydrophobic residues. S: β-strand; Orange box: highly conserved Gly residue; Red box: long second-strand anchor. (C) Conserved hydrophobic residues align along the β-sheet sides of the solenoid structure. Side and top views of the ribbon diagram of the I-TASSER BacM model with the first two hydrophobic residues of each repeat highlighted in purple. (D) Table of parameters used to judge the quality of the predicted models by the five template-based servers. Models were evaluated based on the number of repeats and their alignment as well as ordered N- and C-termini. Checks and minuses denote that a given model possesses or lacks a given quality, respectively.
Fig 3.
Structure of BacM fibers, ribbons, and filaments after dialysis against various buffers.
Exogenously expressed and purified BacM polymerizes into different structures upon dialysis against different buffers. (A) In a phosphate-citric acid buffer (pH 6.0, top), or at high concentrations (100 μM, bottom), large ribbon-like structures were observed. (B) In Tris, CHES or CAPS buffers (pH 7.5, 9.5 or 10.5, respectively), 10 nm fibers resembling those isolated from M. xanthus cells predominate (top row). The addition of 20 mM glycine to the buffers (bottom row) favors the formation of 3 nm filaments at increasing abundance with increasing pH. Scale bar = 50 nm.
Fig 4.
Modeling of the dimerization domain of bactofilin predicts hydrophobic interface and charged surfaces.
(A) The bactofilin domain from the I-TASSER model was modeled using ClusPro 2.0 and predicts head-to-tail dimerization. Individual BacM monomers are represented as blue and green ribbon structures, and the dimer is presented as rotated 180o. (B) The interface between two bactofilin-domains is predicted to contain continuous stacking of hydrophobic subunits within β-strands, with highly conserved residues at key positions to form a hydrophobic pocket. The residues mutated in this study are highlighted with red boxes. (C) An electrostatic surface map of the dimer model reveals patches of charged residues that are solvent exposed. Negatively charged areas are shown in red, positively charged areas in blue, and areas that are charge neutral are shown in white.
Fig 5.
Evaluation of polymerization and fiber formation of recombinant mutant forms of BacM.
While the wild type (A) and the C-terminal truncation mutant (D) are able to polymerize, the L35E (B) and the I124D/F125R (C) mutants are no longer able to form fibers and aggregate instead. (D) Despite their ability to polymerize, the C-terminal truncation forms fibers that show a distinct aberrant “braided” morphology when compared with the smooth wild type fibers. The scale bar in the large field is 0.5 μm, while the scale bar in the inset is 100 nm.
Fig 6.
Mutations to putative BacM interface fail to rescue the morphological phenotype of a deletion mutant.
(A) Images of ΔbacM cells or the indicated rescue strains were taken by phase contrast light microscopy. (B) The same cell lines were fixed and imaged by immunofluorescence microscopy with an anti-BacM antibody (red) and stained with DAPI (blue). Top panel, anti-BacM. Bottom panel, merge with phase contrast. Scale bar = 5 μm.
Fig 7.
Evaluation of polymerization and fiber formation of recombinant wild type BacM in the electron microscope after various treatments.
Fiber formation is robust and occurs under a wide variety of conditions. In the presence of chaotropic agents, such as urea, the wild type protein polymerizes at concentrations of up to 1 M (upper row), while polymerization is more sensitive to low than to high pH values (middle row) and does not occur at NaCl concentrations of 0.5 M and higher (lower row) indicating the importance of charge for the lateral association of BacM filaments. The scale bar in the large field is 0.5 μm, while the scale bar in the inset is 100 nm.