✕
Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2015 Jul 2;523(7558):106-10.
doi: 10.1038/nature14356. Epub 2015 Apr 27.

Structures of actin-like ParM filaments show architecture of plasmid-segregating spindles

Tanmay A M Bharat  1 , Garib N Murshudov  1 , Carsten Sachse  2 , Jan Löwe  1
Affiliations

Structures of actin-like ParM filaments show architecture of plasmid-segregating spindles

Tanmay A M Bharat et al. Nature. .

Abstract

Active segregation of Escherichia coli low-copy-number plasmid R1 involves formation of a bipolar spindle made of left-handed double-helical actin-like ParM filaments. ParR links the filaments with centromeric parC plasmid DNA, while facilitating the addition of subunits to ParM filaments. Growing ParMRC spindles push sister plasmids to the cell poles. Here, using modern electron cryomicroscopy methods, we investigate the structures and arrangements of ParM filaments in vitro and in cells, revealing at near-atomic resolution how subunits and filaments come together to produce the simplest known mitotic machinery. To understand the mechanism of dynamic instability, we determine structures of ParM filaments in different nucleotide states. The structure of filaments bound to the ATP analogue AMPPNP is determined at 4.3 Å resolution and refined. The ParM filament structure shows strong longitudinal interfaces and weaker lateral interactions. Also using electron cryomicroscopy, we reconstruct ParM doublets forming antiparallel spindles. Finally, with whole-cell electron cryotomography, we show that doublets are abundant in bacterial cells containing low-copy-number plasmids with the ParMRC locus, leading to an asynchronous model of R1 plasmid segregation.

PubMed Disclaimer

Figures

Extended Data Figure 1
Extended Data Figure 1. Resolution estimate of the ParM+AMPPNP reconstruction
(a) Resolution of the ParM+AMPPNP reconstruction was estimated using ResMap and this estimate was plotted back onto the cryo-EM density. Blue indicates high-resolution while red indicates lower resolution. (b) The power spectrum of the aligned segments (left) compared to the power spectrum of the re-projection of the cryo-EM reconstruction (right). A reflection is observed in both cases at 4.8 Å−1 indicating that the resolution extends beyond this shell. See Fig. 2e for Fourier Shell Correlation (FSC) curves.
Extended Data Figure 2
Extended Data Figure 2. Intra- and inter-protofilament interactions in ParM filaments
(a) Atomic model of one protofilament (strand) of ParM is shown with the residues at the protein:protein interface highlighted in red. See ED Table 2 for a detailed list. (b) A magnified view of the interface. Three residues at the interface have been labelled. (c) The complete ParM filament (i.e. both protofilaments/strands) shown end-on. (d) Atomic model of the ParM filament with the inter-protofilament residues at the protein:protein interface highlighted in orange. (e) A magnified view of d). Salt bridging residues are labelled. (f) An orthogonal view of d). See ED Table 2 for a detailed list of interacting residues.
Extended Data Figure 3
Extended Data Figure 3. The ParM inter-protofilament interface is small but important
(a) Cryo-EM density for the ParM+AMPPNP filament is shown at an isosurface contour level of 2.0 σ from the mean value. Overlaid on the density, refined atomic coordinates from REFMAC are additionally displayed as grey ribbons. Residues forming salt bridges at the inter-protofilament interface are highlighted. (b) The same figure as a), except the cryo-EM density shown at an isosurface contour level of 1.5 σ from the mean. (c) 1.0 σ from the mean. (d-f) A magnified view of the primary salt-bridged interface consisting of charged residues that form the ParM inter-protofilament interface. The cryo-EM density is shown as a mesh at three different contour levels to demonstrate resolved side chain densities. Positively charged residues are highlighted in red while negatively charged residues are highlighted in orange. (g) Two residues (K258 and R262) that were the best resolved (marked with an * in d), were mutated to aspartic acid to test the importance of this inter-protofilament interface. A cryo-EM image of this mutant protein assembled with AMPPNP is shown. A much higher concentration of the protein was required to obtain filaments on cryo-EM grids (Supplementary Methods). This experiment was repeated four times. (h) Randomly selected cryo-EM images of ParM+AMPPNP and ParM(K258D, R262D)+AMPPNP were used to count occurrences of straight and bent filaments by visual inspection. The results of this quantification are shown as a percentage bar diagram. For the ParM protein, 82% of all filaments were classified as straight, while 18% were bent (n=345). Using exactly the same classification criteria, only 15% of the filaments were found to be straight and 85% of the filaments were bent (n=45) for the ParM(K258D, R262D) mutant protein. (i) Reference-free class averages show that most of the ParM(K258D, R262D) filaments are made up of double protofilaments like wild-type ParM. Some class averages show evidence of bending. A few class averages show that single protofilaments were present in the sample (lower panels). However, the double mutation destabilises the entire ParM filament, making filament formation an unfavourable reaction, illustrating that even though the inter-protofilament interface is small, it is critical for ParM filament formation.
Extended Data Figure 4
Extended Data Figure 4. ParM adopts a compact conformation until ATP is hydrolysed to ADP or until phosphate is released
(a) ParM protein (10 μM) was incubated with ATP (2 mM) and cryo-EM samples were prepared after 5 minutes. Many filaments were observed on the grid. This experiment was repeated ten times. (b) After two hours, no filaments were seen in the same reaction. Presumably, ATP had been hydrolysed and ParM had returned to monomeric form. This experiment was repeated three times. (c) When sodium orthovanadate (4 mM) was included in the reaction, filaments could be observed even after two hours. This experiment was repeated three times. (d) The same reaction as a), except ATP was replaced by ADP. No filaments were observed in this reaction. This experiment was repeated four times. (e, f) We performed real-space helical reconstruction of the ParM+ATP filaments (red) and ParM+ATP+vanadate filaments (yellow), and compared them with the ParM+AMPPNP filament structure (green). Comparison shows that ParM is held in a very similar conformation until hydrolysis of ATP is complete or until phosphate is released since we currently cannot distinguish these two possible effects of vanadate. See Fig. 2e for resolution estimates and ED Table 1 for image processing statistics.
Extended Data Figure 5
Extended Data Figure 5. Model of the ParM doublet
(a) A cryo-EM image of ParM+AMPPNP + 2% PEG 6000. Instances of doublets are marked with yellow arrowheads. This experiment was repeated 15 times. (b) More examples of ParM doublets observed in cryo-EM. (c) Class averages of the doublets. (d) Directionality assignment of the filaments in the doublet. Individual sub-segments and their assigned directionality are indicated by triangles, coloured based on the cross-correlation score in the alignment procedure: red indicates a poor cross correlation score while green indicates a good score. (e) A schematic model of the anti-parallel ParM doublet. Directionality is indicated with a yellow arrow. (f) The thickest parts of ParM filaments of the doublet (as they appear in projection) are marked with black arrowheads.
Extended Data Figure 6
Extended Data Figure 6. Validation of the doublet model
(a) Two ParM monomers arranged in an anti-parallel orientation, as obtained from the ParM cryo-EM doublet model. (b) Two ParM monomers arranged in an anti-parallel orientation, obtained from crystal packing of a monomeric ParM X-ray structure (PDB 4A62) . (c) Two residues at the interface of the doublet (see ED Table 2), S19 and G21 were mutated to arginine to improve affinity of ParM filaments to each other. Cryo-EM images of the mutant protein with AMPPNP show spontaneous doublet formation and filament bundling without crowding agent, validating the doublet model. This experiment was repeated six times.
Extended Data Figure 7
Extended Data Figure 7. ParM bundles and doublets observed in vivo
(a-k) E. coli B/R266 cells were transformed with a high-copy (pDD19), or medium-copy (pKG321) plasmid containing the ParMRC locus. Transformed cells were grown to log phase and then prepared for cryo-EM. This figure shows a gallery of ParM bundles (blue arrows) and doublets (yellow arrows) observed in these cells. Panels a, c, e and h show cells transformed with the high-copy number plasmid while panels b, d, f, g, i, j and k show cells transformed with the medium copy number plasmid. Each experiment with different copy number plasmids was performed only once due to low-throughput nature of cryo-ET.
Figure 1
Figure 1. 4.3 Å cryo-EM reconstruction of ParM+AMPPNP filaments
(a) Cryo-EM image of ParM+AMPPNP filaments. Inset: class average. This experiment was repeated nine times. (b) A 4.3 Å reconstruction of the filaments, isosurface contoured at 2 σ away from the mean (see ED Fig. 1 and Video 1). (c) The same reconstruction as b), overlaid with the refined atomic model with individual ParM subunits coloured differently. (d-g) Enlarged regions of the cryo-EM map showing resolved secondary structure elements and side chain densities, contoured at 1 σ. (h) Density for the nucleotide is stronger than that of the protein (contoured at 3 σ).
Figure 2
Figure 2. ParM filaments are made up of two protofilaments held together by salt bridges, which are perturbed when ParM is bound to ADP
(a) The refined atomic model of ParM+AMPPNP filaments shows that the protofilaments are held together laterally by salt bridges. Basic residues at the interface are highlighted in red and acidic residues in orange (see ED Table 2). Within the protofilaments’ longitudinal interfaces, more extensive hydrophobic interactions are observed (see ED Fig. 2). (b) A magnified view of a). The charges of two basic residues at the interface were inverted by mutation for panel c) (K258D, R262D). (c) The resulting protein formed filaments inefficiently. Cryo-EM image showing filaments of ParM(K258D, R262D) assembled with AMPPNP. This experiment was repeated four times. (d) In addition to normal double-helical filaments, some single-helical filaments were observed by image classification and averaging. (e) Fourier shell correlation (FSC) curves for the four cryo-EM structures presented in this study (see ED Table 1). (f) Cryo-EM image of ParM+ADP filaments. High protein concentrations were required to obtain these filaments and monomeric proteins can be seen. This experiment was repeated six times. (g) Comparison of filtered class averages of ParM+ATP and ParM+ADP filaments. Compared to the ATP bound state, the pitch of the ParM+ADP filaments reduced by ~ 3 Å (see Video 2). (h) Cryo-EM reconstruction of ParM+ADP filaments at 11 Å resolution with 5 copies of the ParM+ADP X-ray structure fitted. (i) The same pseudo-atomic fit without the cryo-EM density. (j) A magnified view of the perturbed inter-protofilament interface in the ParM+ADP filaments.
Figure 3
Figure 3. ParM doublets formed in vitro
(a) Cryo-EM images of ParM doublets formed in vitro with crowding agent PEG 6000. This experiment was repeated 15 times. (b) Slice through an electron cryotomogram (cryo-ET) showing clear lack of super-helicity in the doublets (see Video 3). (c) A 2D class average of the ParM doublet. The thickest parts of double helical ParM filaments have been indicated with yellow arrowheads (see ED Fig. 5). (d) Model of the doublet, shown in the same orientation as the class average in c) (see Video 4). (e) An orthogonal, magnified view of the doublet cut at the plane shown as a dashed line in d). (f) Atomic model of the doublet. Residues shown in red in one ParM filament interact with residues in orange in the other filament (see ED Table 2). (g) An orthogonal view of the doublet, with the filament axes going into the plane of the paper. One of the residues (S19) that forms the doublet interface has been highlighted (see ED Fig. 6).
Figure 4
Figure 4. ParM doublets in E. coli cells, imaged by cryo-ET
(a) A mutant of ParM that hydrolyses ATP more slowly (D170A) was over-expressed in E. coli cells. Tomographic slices show large bundles of ParM blocking cell division. This experiment was performed two times. (b) The ParMRC operon driven from high-copy number plasmid pDD19. Tomographic slice showing an example of observed doublets. (c) Tomographic slice for a medium-copy number plasmid (pKG321). (d) Tomographic slice for a low-copy number plasmid, emulating the native low-copy number R1 plasmids (pKG491, ‘mini-R1’ replicon) in E. coli (see Videos 5-6 to view entire tomograms). Each experiment with different copy number plasmids was performed once. (e) Schematic depicting proposed asynchronous plasmid DNA segregation. Bipolar ParM spindles are seeded when replication has produced two parC centromeric regions, still in close proximity. Each seeds one unipolar ParM filament that then come together in an antiparallel fashion to form the segregating bipolar spindle. Non-productive unipolar filaments or spindles that lack plasmid attachment will be destroyed through ParM’s dynamic instability. This is in contrast to earlier ideas in which all sister plasmids would be segregated through one bundle of filaments, containing double the number of unipolar filaments as the copy number of the plasmid in the cell .
See this image and copyright information in PMC

References

    1. Møller-Jensen J, Jensen RB, Löwe J, Gerdes K. Prokaryotic DNA segregation by an actin-like filament. EMBO J. 2002;21:3119–27. - PMC - PubMed
    1. Gerdes K, Howard M, Szardenings F. Pushing and pulling in prokaryotic DNA segregation. Cell. 2010;141:927–42. - PubMed
    1. Gayathri P, et al. A bipolar spindle of antiparallel ParM filaments drives bacterial plasmid segregation. Science. 2012;338:1334–7. - PMC - PubMed
    1. Orlova A, et al. The structure of bacterial ParM filaments. Nat Struct Mol Biol. 2007;14:921–6. - PMC - PubMed
    1. Popp D, et al. Molecular structure of the ParM polymer and the mechanism leading to its nucleotide-driven dynamic instability. EMBO J. 2008;27:570–9. - PMC - PubMed

Publication types

MeSH terms

LinkOut - more resources