(A) Schematic diagram of the general steps of plasmid partition. Plasmids are represented as large circles, and the partition complexes formed by centromere-binding proteins binding to the centromere are shown as small circles. The first step is partition complex formation. Subsequently, plasmid pairing (interacting ovals) occurs. Once the plasmids are paired, the motor protein is recruited to the partition complex and drives plasmid separation. (B) Genetic organizations for the two main families, type I and II, of par loci. Shown are representatives from each class (type Ia, F and P1; type Ib, pTP228; double loci, pB171; and type II, R1). Orange arrows represent genes that encode motor proteins (ATPases) and blue arrows represent genes encoding centromere-binding proteins. Centromere sites are indicated by black bars and are labelled (e.g. parS, sopC, parC1, parC2 and parC). Arcs represent DNA-binding properties of the indicated gene products, continuous arcs show regulation of promoter activity by the indicated gene products and broken arcs represent formation of partition complexes.
Well-characterized centromere-like sites are shown for two of each class of par systems (type Ia, type Ib and type II). The arrows represent repeat elements within each centromere. The arrows are schematic; the lengths of the arrows do not represent nucleotide length. The number of nucleotides in each repeat are indicated. For example, each sopC repeat is 43 bp, the P1 parS B-box elements are 6 bp (orange), and the A-box elements are 7 bp (blue). The central IHF site is 29 bp (grey). The TP228 parH repeats contain 19 bp repeats, whereas the pTAR parS site contains consensus repeats that are 7 bp. The type II pSK41 parC repeats are 20 bp, whereas the R1 repeats within parC are 11 bp in length.
Figure 3Structures of type Ib centromere-binding proteins
(A) Schematic diagram showing the domain structure of the type Ib centromere-binding proteins. The N-terminal domain (grey) is a flexible region that functions in ATPase stimulation, ATPase recruitment and pairing. In addition, a role in DNA binding has been suggested. The C-terminal domain (magenta) contains the DNA-binding RHH motif. In some cases, there is a short region following the RHH (shown in grey). The RHH motifs of the structures in (B) and (C) are shown in the same orientations for comparison purposes. (B) Structure of the ParG protein from the E. coli TP228 plasmid shown colour-coded as in (A) . The protein is a dimer. The N-terminal flexible region is grey, and the RHH, which mediates all the dimer contacts, is magenta. (C) Structure of Streptococcus pyogenes plasmid pSM19035 ω protein colour-coded as in (A) [110,111]. On the left is the apo structure, and on the right is the ω–DNA structure. The DNA is shown as sticks and is coloured blue. N- and C-termini regions are labelled for one subunit in each apo structure. (B) and (C) were modelled using PyMOL (DeLano Scientific).
Figure 4Structures of type II centromere-binding proteins
(A) Schematic diagram showing the domain structure of the type II centromere-binding proteins. The N-terminal domain (magenta) forms a RHH domain responsible for centromere binding. The C-terminal region is a partially flexible domain that functions in ParM binding and also contributes to the stabilization of the RHH dimer and dimer-of-dimers. The RHH motifs of the structures in (B) and (C) are shown in the same orientations for comparison purposes. (B) Structure of the E. coli pB171 ParR colour-coded as in (A) . The N-terminal RHH domain is coloured magenta, and the C-terminal region, which consists of a three-helix cap and additional disordered residues, is grey. N- and C-terminal regions are labelled. (C) Structure of the Staphylococcus aureus pSK41 ParR protein coloured as in (A) . On the left is the protein in the absence of DNA, and on the right is the ParR–centromere complex. N- and C-termini regions of one subunit are labelled in the left-hand panel. The DNA is shown as sticks and coloured blue. (B) and (C) were modelled using PyMOL (DeLano Scientific). An interactive three-dimensional version of the structure shown in (C) can be seen at http://www.BiochemJ.org/412/0001/bj4120001add.htm.
Figure 5Structures of type Ia centromere-binding proteins
(A) Schematic diagram showing the domain structure of the type Ia centromere-binding proteins. The N-terminal domain (grey) is a flexible region involved in ATPase activation and recruitment. Motif B (cyan) functions as a higher-order oligomerization domain. The HTH unit (magenta) is a three-helix unit containing an HTH motif. This unit is highly conserved in the type Ia structures solved to date. The linker domain (blue) is a helical domain that links the HTH unit to the dimer domain. The dimer domain (light grey) mediates dimerization and is not conserved among the type Ia centromere-binding proteins. The HTH units of each structure are shown in the same orientation with its three helices labelled H1–H3 (magenta). (B) Structure of the T. thermophilus Spo0J protein colour-coded as in (A) . Although included in crystallization, the N-terminal domain was not visible. Motif B (cyan) mediates dimerization between Spo0J subunits at high concentrations. The HTH unit is magenta and the linker domain is blue. N- and C-termini regions are labelled. (C) Structure of the P1 ParB-(142–333)–parS small complex . The HTH unit is shown as the same orientation as that of Spo0J and is coloured magenta. The linker domain is blue, the dimer domain is grey, and the N- and C-termini regions of one subunit are labelled. The DNA is shown as a surface and is coloured yellow. (D) Structure of the RP4 ParB homologue, KorB. The C-terminal dimer-domain (residues 297–358) and DNA-binding domains (residues 101–294) were solved as separate domains (not all residues were visible) [118,119]. The KorB dimer was modelled by linking the two domains by a flexible region (indicated by broken lines). Part of motif B present in the structure is coloured cyan. The HTH unit is magenta and the linker domain is blue. The dimer domain, solved separately, is grey and is shown as linked by flexible residues of the linker domain. The DNA is shown as a yellow surface. (B)–(D) were modelled using PyMOL (DeLano Scientific).
Figure 6Structure of the P1 ParB-(142–333)–parS double B-box interaction: P1 ParB is a bridging factor
(A) Sequence of the P1 parS centromere site. The site is divided into right and left ParB-binding sites, with an IHF site located in the centre. The A- and B-boxes are labelled. The leftside B1 and A1 boxes are coloured blue and green, whereas the rightside A2, A3 and B2 boxes are coloured yellow, green and blue respectively. Note that the third A-box, A4, which is located to the left of B2, is not necessary for partition and is not shown. The rightside parS site, called parS small, encompasses A2–A3–B2 and is the minimal site required for partition. (B) Structure of P1 ParB-(142–333) bound to an A3–B2-containing DNA site in which the protein dimer–DNA duplex stoichiometry was 1:2 . One ParB-(142–333) subunit is coloured magenta and the other is yellow. The DNA is shown as surfaces with box elements coloured as in (A). Modelled using PyMOL (DeLano Scientific). An interactive three-dimensional version of the structure shown in (B) can be seen at http://www.BiochemJ.org/412/0001/bj4120001add.htm.
On the left (red) is the structure of apo Soj. On the right is the structure of Soj(D44A), which forms a nucleotide sandwich dimer (one subunit is red and the other green) upon binding ATP. The ATP molecules are shown as CPK (Corey–Pauling–Koltun) with carbon, nitrogen, oxygen and phosphorous atoms coloured cyan, blue, red and magenta respectively . Modelled using PyMOL (DeLano Scientific).
Figure 8Structures of actin-like ParA homologue, E. coli R1 ParM
(A) Crystal structure of R1 ParM in its apo state . (B) Structure of the R1 ParM ADP-bound state. Comparison with the apo structure reveals that ADP binding induces a 25° relative closure of the two domains . (C) Structure of eukaryotic actin bound to ADP . In all structures, helices are coloured cyan, β-strands are coloured magenta and loop regions are coloured light pink. Also labelled are subdomains Ia, Ib, IIa and IIb. Modelled using PyMOL (DeLano Scientific).
Figure 9EM reconstructions of R1 ParM filaments and type II insertional polymerization model
Top: comparison of EM reconstructions from Orlova et al.  (left) and Popp et al.  (right). Left: reprinted by permission from Macmillan Publishers Ltd: Nature Structural and Molecular Biology. Orlova, A., Garner, E., Galkin, V. E., Heuser, J., Mullins, R. D. and Egelman, E. H. (2007) The structure of bacterial ParM filaments. Nat. Struct. Mol. Biol. 14, 921–922. Copyright 2007. http://www.nature.com/nsmb/. Right: reprinted by permission from Macmillan Publishers Ltd: EMBO Journal. Popp, D., Narita, A., Oda, T., Fujisawa, T., Matsuo, H., Nitanai, Y., Iwasa, M., Maeda, K., Onishi, H. and Maeda, Y. (2008) Molecular structure of the ParM polymer and the mechanism leading to its nucleotide-driven dynamic instability. EMBO J. 27, 570–579. Copyright 2008. http://www.nature.com/emboj/ ParM molecules fitted into the model are numbered in the Orlova model. Bottom: model of R1 plasmid partition (from Campbell and Mullins ). Reproduced from The Journal of Cell Biology, 2007, 179: 1059–1066. Copyright 2007 The Rockefeller University Press. The proposed steps in the model are numbered 1–5. Step 1 involves nucleation of filaments. After nucleation, one end of the filaments attaches to one partition complex, and the other end of the filament searches for the partition complex located on another plasmid. In step 2, the plasmids diffuse in the cell until they encounter a second plasmid. Once in close proximity, in step 3, both ends of the filaments are bound by a plasmid, preventing catastrophic collapse. In step 4, because the filaments are now stable, they continue to grow and push the plasmids to opposite cell poles. In step 5, the plasmids are finally pushed into the cell poles, the force of which destabilizes the interactions of the ParM filaments with the plasmid. The plasmid is then released, at which point the free end of the filament is destabilized and rapidly depolymerizes.
(A) Model of pSK41 segrosome assembly. To construct this model, the pSK41 partition complex structure and the Orlova (for which co-ordinates were deposited) ParM filament model were used [90,130]. As can be seen from this model, the ParM filament can be embraced within the partition complex pore without steric clash and can favourably interact with the C-terminal flexible domains of the ParR molecules, found at high concentrations within the pore. (B) In this pSK41 segregation model, the ParM filament is captured between two partition complexes on different plasmids. The filament and partition complexes are drawn roughly to scale. Centromere DNA is coloured brown, the ParR molecules are magenta, green, blue and yellow, and ParR molecules participating in spreading are cyan. ParR flexible C-domains, which interact with the ParM filament, are represented as half circles. The ParM filament is represented as a yellow surface rendering. Modelled using PyMOL (DeLano Scientific).