Most bacterial chromosomes and many low-copy number plasmids depend on the Type-I system for the segregation of DNA. The Type-I system is characterized by the presence of a P-loop ATPase superfamily of protein, known as the ParA/MinD family of ATPases, that have a deviant Walker A box motif KGGXXK[S/T] (Koonin, 1993; Lutkenhaus, 2012). The genetic loci important for partitioning also carry a centromeric repeat sequence, parS, and encode a second protein, the centromere binding protein (CBP) or adaptor protein ParB that binds parS sites and interacts with ParA as well. The Walker A motif in ParA is directly involved in interactions with the bound ATP molecule. Walker A motifs in the Type-I ParA family differ from the classical Walker A motif in having this additional signature lysine residue (Table 1), and thus the term deviant Walker A motif follows (Motallebi-Veshareh et al., 1990; Hayes, 2000; Lutkenhaus and Sundaramoorthy, 2003; Wendler et al., 2012). A second motif, the B box, characterized by conserved negatively charged residue (D/E), plays an important role in Mg²⁺ coordination and ATP hydrolysis (Fung et al., 2001; Schumacher, 2008). Mutations in the conserved Walker A box lead to the loss of plasmids and point to a key role for ATP hydrolysis in mediating segregation (Ebersbach and Gerdes, 2001; Fung et al., 2001; Libante et al., 2001; Barillà et al., 2005; Pratto et al., 2008). The Type-I system is further sub-divided into Type-Ia and Type-Ib based on the structure of ParA (Hayes, 2000; Schumacher, 2008).

Type-Ia/large ParA proteins (~300–450 amino acids) have an extended N-terminal helix-turn-helix (HTH) motif and can act as repressors of their gene expression (Figures 1A,B) (Abeles et al., 1985; Hirano et al., 1998; Libante et al., 2001). The HTH domains help in this sequence-specific DNA binding to operator regions near their promoters (Mori et al., 1989; Davis et al., 1992; Hayes et al., 1994; Ravin et al., 2003). Type-Ia is one of the first identified DNA segregation systems and is typically encoded by partitioning loci carried in plasmids (Austin and Abeles, 1983; Ogura and Hiraga, 1983; Mori et al., 1986). Specific examples of Type-Ia include the parABS and the sopABC (parABS_F) system of P1 bacteriophage and F plasmid, respectively. In the case of F plasmid (the one encoding for Fertility factor), the genes appear in the order of sopA, sopB, and sopC, hereafter referred to as parA_F, parB_F, and parS_F, respectively. The expression of ParA_F and ParB_F are driven by a single promoter located upstream of parA_F (Mori et al., 1986, 1989; Hirano et al., 1998). The parABS_F cluster serves as the partitioning system, wherein parS_F serves as the centromeric sequence and comprises twelve 43-bp base pair repeats (Figure 1A). Each 43-bp sequence contains a short 16-bp inverted repeat to which ParB_F binds as a dimer (Hayakawa et al., 1985; Lane et al., 1987; Mori et al., 1989; Hanai et al., 1996). This ParB_F-parS_F complex is then recruited to ParA_F, which is the NTPase and functions as the motor protein (Ogura and Hiraga, 1983).

On the contrary, members of the Type-Ib subfamily are smaller (~200–250 amino acids) and lack the N-terminal HTH domain. They are thus referred to as smaller ParAs and are found in the chromosomal loci of many bacteria. However, they are also found in a few plasmids, like in TP228 of Salmonella newport. The ParA ATPases in the Type-Ib systems do not have any repressor functions due to the lack of the N-terminal HTH domain required for site-specific DNA binding at their promoter regions (Figure 1A). Examples of the Type-Ib system on plasmids include parFGH loci on the S. newport plasmid TP228 and the parABS loci (δ/parA, ω/parB and parS) in the plasmid pSM19035 carried by Streptococcus pyrogenes (de la Hoz et al., 2004; Fothergill et al., 2005; Pratto et al., 2008). Type-Ib also includes ParA from Caulobacter crescentus (Mohl and Gober, 1997) and Soj from Bacillus subtilis (ParA_Bsu/BsSoj) (Leonard et al., 2005a; Lee and Grossman, 2006) and Helicobacter pylori (ParA_Hpy/HpSoj) (Lee et al., 2006) encoded by bacterial genomes. ParB members of the Type-Ib family carry an N-terminal protein-protein interaction domain, central HTH domain and a self-dimerization domain at the C-terminus. Further, residues near the N-terminus of ParB specify their interactions with the cognate ParA (Funnell, 2016; Kawalek et al., 2020). Similar to the Type-1a system, ParB binds to parS and interacts with ParA, thus linking the plasmid to the motor protein ParA.

The ParA proteins, both the larger Type-1a and the smaller Type-Ib, also bind DNA in a non-sequence specific manner and is critical for their DNA segregation functions (Ebersbach and Gerdes, 2005; Leonard et al., 2005a; Hatano et al., 2007; Hester and Lutkenhaus, 2007; Castaing et al., 2008; Vecchiarelli et al., 2010; Roberts et al., 2012; Le Gall et al., 2016). The currently accepted models for DNA segregation by ParA, i.e., the diffusion-ratchet, DNA-relay and DNA hitch-hiking mechanisms (described in detail below), emphasize the importance of bacterial nucleoids. Briefly, the DNA hitch-hiking model proposes that the ParA-ATP dimers localize on the HDRs of the nucleoid to which the ParB-parS complex binds and stimulates the ATPase activity of ParA. The hydrolysis event converts the nucleoid bound ParA-ATP into ParA-ADP, resulting in the release of ParA from the nucleoid. This creates a zone of depletion of DNA bound ParA, and the released ParB-parS complex undergoes diffusive motion until captured by ParA-ATP dimers within the HDRs, resulting in a directional motion toward the highest concentration of DNA bound ParA-ATP in the cell (Vecchiarelli et al., 2012; Le Gall et al., 2016; McLeod et al., 2017). The bacterial nucleoid thus forms a key substrate for the ParA in its function as a motor protein in the equipartitioning of plasmids and chromosomes into daughter cells (Vecchiarelli et al., 2012).

Type-II partitioning system was first discovered in the resistance plasmid R1 (Jensen and Gerdes, 1997). The partitioning system in this group contains an actin-like ATPase called ParM, an adaptor protein, ParR and a centromere site, parC. Monomeric ParM has an actin-like fold, with its crystal structures bearing a close resemblance to the eukaryotic actin (Bork et al., 1992; van den Ent et al., 2002). A search of the microbial genomes revealed as many as 40 different families of actin-like proteins, mostly found on plasmids (Derman et al., 2009). A few have been shown to assemble into polymeric structures and filaments with varying architectures (Becker et al., 2006; Derman et al., 2009; Popp et al., 2012; Jiang et al., 2016; Koh et al., 2019). However, the ParMRC system is the most extensively studied among the DNA partitioning systems. ParM, like actin, assembles into a two-stranded helix but with an opposite twist. While actin forms a right-handed filament, ParM helices are mainly left-handed, with 12 subunits per turn as opposed to 13 monomers in actin (Orlova et al., 2007; Popp et al., 2008; Gayathri et al., 2012). Interestingly, in the ParR (C-terminal 17 residue peptide) bound form, domains IA and IIB show a rotation of 9.1° and 10.7° toward the nucleotide, reminiscent of the shift in actin structure from G-actin to F-actin. Further, the ParR peptide binding site overlaps with the ParM polymerisation interface and thus allows ParR to bind only to the barbed end or the polymerizing end (Gayathri et al., 2012). Also, ParM grows bidirectionally with similar monomer addition rates at both ends and exhibits dynamic instability, a characteristic feature of microtubules (Gerdes et al., 2000; Garner et al., 2004).

Moreover, in the presence of a non-hydrolysable analog of ATP, these in vitro filaments were longer suggesting that the growth and retraction of filaments were driven by ATP hydrolysis (Garner et al., 2007). Although ParM can bind both ATP and GTP, the most preferred substrate is ATP, for which ParM has a 10-fold higher affinity than GTP (Popp et al., 2008; Galkin et al., 2009). Cryo-electron microscopy of cells overexpressing ParM has shown that the protein assembles into closely packed filament bundles with an average of 3–5 ParM filaments per bundle (Salje et al., 2009). Further studies suggest that the filaments grow by insertional polymerisation, wherein new ParM subunits are inserted at the interface of the filaments with the ParR-parC complex (Møller-Jensen et al., 2003; Garner et al., 2004, 2007; Salje et al., 2010; Gayathri et al., 2012). In vitro reconstitution of the ParMRC segregation machinery using polystyrene beads for immobilizing parC, purified ParR and ParM proteins, and ATP resulted in the assembly of dynamic filaments of ParM. Surprisingly, the ParM filaments were seen to grow and shrink in a manner that was reminiscent of the dynamic instability of microtubules (Garner et al., 2004, 2007). However, in the presence of another parC coated bead in proximity, the filaments appeared to bundle and push the beads further away, consistent with the bipolar antiparallel filament structure (Gayathri et al., 2012, 2013). Thus, ParMRC is a minimalistic tripartite system that is sufficient for plasmid segregation and does not require any additional host factors (Garner et al., 2007). Based on cryo-electron microscopy, in vitro reconstitution experiments, and fluorescence imaging, a “search, capture, and push” model for plasmid segregation by ParM and other actin-like proteins has been proposed (Campbell and Mullins, 2007; Garner et al., 2007; Salje et al., 2009, 2010; Gerdes et al., 2010; Gayathri et al., 2012).

Partitioning machinery of this kind has been described in the pXO1 plasmid of B. anthracis and B. cereus (Tinsley and Khan, 2006; Hoshino and Hayashi, 2012), pBToxis plasmid of B. thuringiensis (Tang et al., 2006; Larsen et al., 2007) and more recently in bacteriophages c-st of Clostridium botulinum (Oliva et al., 2012) and 201φ2-1 of Pseudomonas chlororaphis (Kraemer et al., 2012). The segregation machinery comprises three components (tubZRC) similar to those found in Type-I and Type-II systems. However, unlike the others, the motor NTPase TubZ or PhuZ (for Phage Tubulin/FtsZ) belongs to the tubulin/FtsZ superfamily of cytoskeletal proteins (Tinsley and Khan, 2006; Larsen et al., 2007; Anand et al., 2008; Oliva et al., 2012). While, in general, the sequence similarity with eukaryotic tubulin amounts to even <15%, TubZ exhibits striking structural similarity with that of the bacterial cell division protein FtsZ and tubulin (Löwe and Amos, 1998; Nogales et al., 1998; Larsen et al., 2007; Aylett et al., 2010). TubZ undergoes polymerisation upon binding GTP (Anand et al., 2008; Chen and Erickson, 2008; Hoshino and Hayashi, 2012) and assembles into two or four-stranded filaments (Aylett et al., 2010; Montabana and Agard, 2014). Interestingly, the subunit interactions upon polymerisation leading to the coupling of GTP hydrolysis are closer to that of α/β-tubulin. Further, the protofilaments in the presence of GTPγS show a right-handed twist with 14 subunits over one complete turn (360°) and assemble into a parallel double-helical filament when expressed in E. coli as well (Aylett et al., 2010). The filaments undergo treadmilling, i.e., it undergoes directional polymerisation, where one end exhibits growth and the other shrinkage, which assists in DNA partitioning (Larsen et al., 2007; Aylett et al., 2010). The Type-III system is thus an example of the involvement of tubulin-like proteins in the DNA segregation in bacteria. TubR constitutes the CBP in this Type-III system that binds to the centromeric sequence tubC and recruits TubZ (Tang et al., 2007; Ni et al., 2010). The structure of the TubR-tubC complex shows that TubR binds tubC through an N-terminal winged helix-turn-helix motif and assembles into a DNA-nucleoprotein complex in vitro, forming a slight right-handed superhelix (Ni et al., 2010; Aylett and Löwe, 2012). Further, the TubR-tubC complex stabilizes the TubZ polymers and possibly exerts its effect by preventing the depolymerisation of TubZ filaments (Aylett and Löwe, 2012; Oliva et al., 2012). Reconstitution experiments reveal that the TubR-tubC complex tracks the depolymerizing minus-end of the TubZ filaments (Fink and Löwe, 2015). The bacteriophage tubulin-like protein PhuZ (TubZ_ΦKZ), on the other hand, resembles the eukaryotic microtubule in many aspects. It was the first prokaryotic tubulin that was shown to exhibit dynamic instability, a property dependent upon the energy derived from GTP hydrolysis (Kraemer et al., 2012; Aylett et al., 2013; Erb et al., 2014). However, unlike tubulin, it assembles into a triple-stranded helical filament (Zehr et al., 2014). One end of the PhuZ filaments is anchored to the cell poles, and they further assemble into bipolar spindle-like structures. These bipolar spindles of PhuZ resemble eukaryotic microtubules during mitosis and help place the viral nuclei at mid-cell and control viral reproduction (Erb et al., 2014).

ParA protein function in Type-I partitioning systems is essential for DNA segregation of the plasmids and the bacterial chromosome, as described above. ParA belongs to the Walker A type Cytoskeletal ATPases (WACA) family of proteins and contains a Walker A motif, a Walker A' motif, a Walker B motif and a ParA specific sequence (Walker et al., 1982; Motallebi-Veshareh et al., 1990; Koonin, 1993; Lutkenhaus, 2012) (Figure 1B). As described above, the larger ParA found in Type-Ia systems has an auto-regulatory function and thus directly regulates the transcription of the parAB operon from its promoter (P_par) and control levels of both ParA and ParB proteins (Mori et al., 1989; Davis et al., 1992; Hayes et al., 1994; Davey and Funnell, 1997; Hirano et al., 1998; Ravin et al., 2003; Komai et al., 2011). ParA exists in a monomer-dimer equilibrium in the cell wherein only the ATP-bound ParA dimers associate with the nucleoid, and ADP-bound dimers (and monomers) are free (Vecchiarelli et al., 2010, 2013a; Havey et al., 2012). The conformational change involves nucleotide-binding wherein the ATP bound form is dimeric (Schumacher et al., 2012; Vecchiarelli et al., 2013a). ParA exhibits a weak ATPase activity similar to the other members of the WACA superfamily (Davis et al., 1992; Watanabe et al., 1992; Libante et al., 2001; Leonard et al., 2005a; Barillà et al., 2007; Havey et al., 2012) and ATP hydrolysis activity can be stimulated 3-fold by ParB, and 1.5 fold by DNA. However, the DNA-ParB complex can exert a much stronger effect and stimulate the ATPase activity of ParA by 10–15-fold (Davis et al., 1992; Watanabe et al., 1992; Bouet et al., 2007; Ah-Seng et al., 2009). Moreover, mutants defective in ATPase activity, like ParA_P1 K122Q and ParF_TP228 D111A, or the mutants displaying hyper-ATPase activity, ParF_TP228 P104A, R169A, and G179A, are all impaired in plasmid stability. These studies highlight the crucial role played by ATP hydrolysis of ParA in plasmid segregation (Fung et al., 2001; Libante et al., 2001; Dobruk-Serkowska et al., 2012; Vecchiarelli et al., 2013a; McLeod et al., 2017; Caccamo et al., 2020). Further, ParA-ATP dimers themselves have been postulated to exist in two different states/conformations, an active conformation (ParA-ATP^*) that binds to non-specific DNA and an inactive state (ParA-ATP) that cannot bind DNA. Thus, it is the active state ATP-bound dimer (ParA-ATP^*)₂ that associates with the bacterial nucleoid (Vecchiarelli et al., 2010, 2013a). The ParA-ADP dimers produced soon after ATP hydrolysis can no longer associate with the bacterial nucleoid. ParA can rebind DNA only upon nucleotide exchange with ATP or reassociation with ATP and the conformational change to the ParA-ATP^* state. However, ParA-ADP can act as a transcriptional repressor of the ParA promoter (Davey and Funnell, 1997; Bouet and Funnell, 1999; Libante et al., 2001; Hao and Yarmolinsky, 2002; Baxter et al., 2020).

ParA has a weak auto-repression activity, and its full repressor function depends on the co-repressor ParB, together with which it strongly represses transcription of its promoter P_par (Friedman and Austin, 1988; Hayes et al., 1994; Libante et al., 2001). Using surface-plasmon-resonance, Bouet and group have revealed that three ParA_F dimers, i.e., a trimer of dimers, bind to the promoter region to suppress gene expression (Boudsocq et al., 2021). The ParB-parS (SopA-sopC) complex further enhances this auto-regulatory function. It was initially thought that only ParA-ATP and ParA-ADP, but not ParA-ATP^* states, were competent in binding to the operator sites parOP at the promoter (Davey and Funnell, 1997; Fung et al., 2001; Ravin et al., 2003; Vecchiarelli et al., 2013a). However, recent studies using the non-specific DNA binding mutant ParA_P1 R351A suggest that ParA-ATP^* state is also proficient in binding parOP. The abrogated ns-DNA binding results in a free pool of excess ParA-ATP^* dimers, resulting in repression of transcription from parOP. This auto-repression activity of ParA thus can solely be attributed to its specific DNA binding activity mediated by the HTH domain of the protein (Baxter et al., 2020).

The active state ATP-bound dimeric conformation (ParA-ATP^*)₂, enables the binding of ParA molecules to the nucleoid (Vecchiarelli et al., 2010, 2013a). ParA has an affinity for DNA in a non-sequence-specific manner in this conformation. Since bacterial cells have a lot of nsDNA in the form of the nucleoid, ParA molecules are found localized to the nucleoid. Work from laboratories around the world on several ParA family proteins has shown that ParA is predominantly nucleoid bound, and its nucleoid binding function is essential for plasmid maintenance (Leonard et al., 2005a; Hayes and Barillà, 2006; Hatano et al., 2007; Hester and Lutkenhaus, 2007; Castaing et al., 2008; Vecchiarelli et al., 2010; Roberts et al., 2012; Lim et al., 2014; Volante and Alonso, 2015; Le Gall et al., 2016; McLeod et al., 2017; Caccamo et al., 2020). The interaction of ParA with nsDNA has been probed in vitro by several studies and visualized in vivo by using sophisticated fluorescence microscopy techniques (Lim et al., 2005, 2014; Hatano et al., 2007; Castaing et al., 2008; Hatano and Niki, 2010; Roberts et al., 2012; Le Gall et al., 2016; McLeod et al., 2017). Further, nsDNA binding defective mutants, ParA_F K340A and ParA_P1 R351A have segregation defects suggestive of the critical role of ns-DNA binding in the process of plasmid maintenance (Castaing et al., 2008; Baxter et al., 2020). In addition, fluorescence microscopy and time-lapse imaging of ParA_F have suggested that the protein assembles into helical polymeric structures that undergo oscillatory behavior on the nucleoid. These observations had earlier led to suggestions that the polymerisation dynamics of ParA drove plasmid segregation (Barillà et al., 2005; Lim et al., 2005; Hatano et al., 2007; Ringgaard et al., 2009). However, as described in detail below, more recent studies on ParA dynamics in vivo and in vitro biochemical assays have favored a polymerisation-independent mechanism of DNA partitioning by ParA.

ParB is a DNA binding protein that binds to specific repeat sequences found in plasmids or genomes (parS) and forms an active component of the bacterial DNA segregation machinery. The role of chromosomally encoded ParB in various cellular functions and DNA partitioning has been recently reviewed in detail by Kawalek et al. (2020). The crystal structure of the full-length ParB protein from B. subtilis (21–218 aa), also known as SpoOJ, provides details about its domain organization (Soh et al., 2019). ParB comprises of different domains: the N-terminal domain, a central DNA binding HTH motif and a C-terminal domain, all connected by flexible linkers (Funnell, 1991, 2016; Schumacher and Funnell, 2005; Badrinarayanan et al., 2015; Soh et al., 2019) (Figure 2A). The C-terminal domain plays a pivotal role in the homo-dimerization of ParB (Khare et al., 2004; Leonard et al., 2005a). Mutation R149G within the HTH motif affects parS binding (Autret et al., 2001; Gruber and Errington, 2009; Fisher et al., 2017) and the central DNA binding domain thus is critical for the recognition and binding of ParB to parS sites (Leonard et al., 2005a; Schumacher and Funnell, 2005). The DNA binding domain or HTH motif plays a vital role in specific DNA interaction with parS sites and spreading to the adjacent DNA after binding to the parS sites. Spreading is a crucial feature of ParB, and it involves the formation of higher-ordered complexes. ParB is known to initiate binding to DNA at parS sites and spread over parS flanking regions, often covering a large span of the nsDNA (reviewed in Jalal and Le, 2020; Kawalek et al., 2020). The N-terminal stretch is necessary for protein oligomerisation as well as interaction with ParA. It is defined by a conserved stretch of Arginine residues GERRxRA, referred to as the arginine patch. Mutations within the arginine patch impair ParB spreading as well as nucleoid segregation functions (Yamaichi and Niki, 2000; Ah-Seng et al., 2013; Chen et al., 2015). This patch is thus essential for the spreading of ParB, foci formation and DNA partitioning (Rodionov et al., 1999; Autret et al., 2001; Bartosik et al., 2004; Breier and Grossman, 2007; Kusiak et al., 2011; Graham et al., 2014; Funnell, 2016). Further, ParB is thought to stimulate the ATPase activity of ParA via an arginine finger (R-finger) motif contained within the N-terminal region (Ah-Seng et al., 2009; Zhang and Schumacher, 2017).

Recent structural and biochemical studies on ParB have revealed that ParB is a CTPase with the arginine patch forming the catalytic center (Soh et al., 2019; Jalal et al., 2020; Osorio-Valeriano et al., 2021). This surprising finding was revealed by crystal structures of ParB of B. subtilis and PadC, a ParB member from M. xanthus, which were found to be associated with CDP and CTP, respectively (Soh et al., 2019; Osorio-Valeriano et al., 2021). The revelation of these ParB structures with CTP as a cofactor has opened up new avenues of research and questions on the role of CTP in regulating the process of ParB spreading and DNA partitioning (Jalal et al., 2020, 2021). CTP binding results in dimerization of the N-terminal domains of two ParBs in the presence of Mg²⁺, which acts as a cofactor in the process. Such self-dimerization is sufficient to cause the formation of a clamp that entraps the DNA (Soh et al., 2019; Osorio-Valeriano et al., 2021). Recent studies using optical tweezers and microscopy have shown that binding of CTP or the non-hydrolysable analog CTPγS enhances ParB binding to parS sites and results in DNA condensation (Balaguer et al., 2021). A conformational change mediates ParB dimerization upon binding to CTP and parS. Upon CTP hydrolysis, CDP has a low affinity for ParB, and thus, it dissociates from the NBD of ParB and results in a switch to its apo-form (Figure 2B). Moreover, studies have shown that ParB networks show a continuous exchange and turnover of ParB at both protein-protein and protein-DNA interfaces (Madariaga-Marcos et al., 2019). Consistently, CTP hydrolysis defective mutants, Q52A and E93A in M. xanthus ParB, exhibit abnormal chromosome segregation (Osorio-Valeriano et al., 2021). Further, ParB spreading to distal regions from parS sites could be prevented by the presence of protein roadblocks, suggesting that ParB spreading on nsDNA was one-dimensional. CTP binding thus plays an essential role in ParB spreading and DNA condensation (Balaguer et al., 2021).

ParB has been recently shown to undergo Liquid-liquid phase separation (LLPS), favoring condensate formation in the presence of CTP (Babl et al., 2022). Liquid-liquid phase separation is a thermodynamic process wherein a homogeneous mixture of two or more molecules de-mixes into distinct phases and thus helps compartmentalize membrane-less organelles (Hyman et al., 2014; Guilhas et al., 2020; Babl et al., 2022). In most eukaryotic subcellular structures, like germline P-bodies (Brangwynne et al., 2009), stress granules (Protter and Parker, 2016) etc., compartmentalization depends upon LLPS. In bacteria, LLPS has been implicated in the assembly of carboxysomes and divisome protein FtsZ (Monterroso et al., 2019; Wang et al., 2019; MacCready et al., 2020). Recently, ParB has also been shown to exist in two different phases, a gas phase and a liquid phase. Further, in a process similar to the LLPS in eukaryotic proteins, ParB from Corynebacterium glutamicum also forms spherical assemblies in the presence of crowding agents, is affected by alterations in ionic strength and is stabilized by potassium glutamate (Babl et al., 2022). The condensation of ParB is further favored by the presence of parS containing DNA and the motor protein ParA helps in the separation of these condensates (Figure 2C). Moreover, in the case of the ParA_F K120Q mutant, the condensates merge and do not segregate suggesting that the ParB stimulated ATPase activity of ParA is required to separate these condensates (Guilhas et al., 2020; Babl et al., 2022).

The partitioning locus in the case of B. subtilis consists of two proteins, Soj and SpoOJ and a centromere-like sequence parS (Ireton et al., 1994; Leonard et al., 2005a; Lee and Grossman, 2006). While Soj belongs to the ParA family, SpoOJ is a ParB-like protein, both of which were initially isolated as sporulation genes and hence the names Soj-SpoOJ. We refer to Soj and SpoOJ proteins as ParA_Bsu and ParB_Bsu, respectively. Overexpression of ParA_Bsu or deletion of ParB_Bsu leads to the formation of aberrant nucleoid morphology and anucleate cells (Lee and Grossman, 2006). ParA_Bsu is a smaller Type-Ib Walker A ATPase with only 21% sequence similarity to ParA_F protein. ParA_Bsu works cooperatively with ParB_Bsu to enable chromosome segregation. ParB_Bsu binds to the eight repetitive sequences on parS located on both sides of the origin of replication oriC. When visualized by fluorescence microscopy, this complex appears as discrete spots on binding ParA_Bsu proteins. A green fluorescent protein (GFP) fusion of ParA_Bsu shows nucleoid localization of the protein. However, when co-expressed with ParA_Bsu, it leads to the formation of discrete spots that colocalise with the ParB-parS complex in the cell (Leonard et al., 2005a; Scholefield et al., 2011). GFP fusions of ParA_Bsu seem to undergo oscillatory movement (Leonard et al., 2005a), a characteristic feature of the ParA superfamily (Raskin and de Boer, 1999; Hatano et al., 2007; McLeod et al., 2017). While ParA_Bsu controls the condensation of ParB_Bsu foci, ParB_Bsu regulates the dynamic behavior of ParA_Bsu, and thus both the proteins are functionally interlinked. Biochemical studies show that ATP bound ParA_Bsu exists as a dimer, and an ATPase mutant D44A was crystallized in dimeric form (Leonard et al., 2005a). Non-specific DNA binding and nucleoid association of ATP-bound ParA_Bsu is mainly mediated by surface-exposed arginine residues (Hester and Lutkenhaus, 2007). Further, electron microscopy data suggest that, like Thermus thermophilus ParA (ParA_Tth/TthSoj), ParA_Bsu also forms a nucleoprotein filament, showing that the dimeric form of ParA_Bsu recruits other ParA_Bsu dimers facilitating the formation of a polymeric structure (Leonard et al., 2004, 2005a). Interaction with ParB_Bsu stimulates the ATPase activity of ParA_Bsu, and ADP-bound forms dissociate from DNA (Scholefield et al., 2011). In addition to its role in DNA segregation, the ParA_BSu is also a transcriptional regulator that controls the expression of many sporulation genes, including spo0A, spoIIA, spoIIG, and spoIIE (McLeod and Spiegelman, 2005).

In C. crescentus, two WACA family members coordinate to promote segregation (Mohl and Gober, 1997; Thanbichler and Shapiro, 2006; Toro et al., 2008; Corrales-Guerrero et al., 2020). Both proteins can interact with the centromere binding protein, ParB, which binds to the parS DNA sites that lie close to the origin of replication (oriC) in C. crescentus. At the beginning of the cell cycle, parS sites are anchored to the cell poles by the interaction of ParB with PopZ, a polarly localized protein. However, at the onset of the S-phase, the chromosome duplicates and one copy of the ParB-parS complex moves to the other pole of the cell driven by the nucleoid bound ParA-ATP gradient (Thanbichler and Shapiro, 2006). MipZ, another ParA member, binds to ParB at both the cell poles and forms a concentration gradient along the cell length, with the highest concentration being on the poles and the lowest at the mid-cell site (Easter and Gober, 2002; Thanbichler and Shapiro, 2006). MipZ is an antagonist of FtsZ, and a high concentration of MipZ prevents the formation of Z-ring at the poles but promotes its assembly at the mid-cell site (Thanbichler and Shapiro, 2006; Du and Lutkenhaus, 2012). Although ParB can bind both ParA and MipZ, the roles played by ParB at the two poles are different. At the old pole, where ParB is bound to parS and interacts with ParA, ParB acts to stimulate the ATP hydrolysis of ParA-ATP. This stimulation of the ATPase activity of ParA by the ParB-parS complex aids in generating the ParA gradient for the effective movement of the ParB-parS complex to the new pole of the cell (Corrales-Guerrero et al., 2020).

In contrast, MipZ binds ParB at both the old and new poles; ParB plays a role of a catalyst and promotes dimerization and binding of MipZ to DNA by recruiting monomers. However, ATP hydrolysis due to the intrinsic ATPase activity of MipZ results in a release of the MipZ monomers from the nucleoids at ParB distal regions. Since ParB is localized at cell poles, the association of MipZ with the nucleoid is at its lowest concentration at the mid-cell position (Kiekebusch et al., 2012). Both MipZ and ParA have conserved residues that bind nsDNA and have nucleoid binding activities that help DNA segregation (Du and Lutkenhaus, 2012). Thus, both MipZ and ParA work synergistically in C. crescentus DNA segregation, and both are indispensable for accurate chromosome partitioning. Further, MipZ serves another critical function of precisely positioning the Z-ring for cell division in C. crescentus.

Bactofilins belong to a group of polymeric cytoskeletal proteins defined by a conserved DUF583 domain, also known as the bactofilin domain (Kühn et al., 2010; Punta et al., 2012). They are characterized by extended beta-sheet structures and are ubiquitous in both Gram-positive and Gram-negative bacteria (Lin and Thanbichler, 2013; Zuckerman et al., 2015). In the rod-shaped bacterium Myxococcus xanthus, four bactofilin paralogs have been identified. They serve diverse functions like cell-shape maintenance (Koch et al., 2011), cell motility via polar localization of a small GTPase, SofG (Bulyha et al., 2013) and chromosome segregation (Lin et al., 2017; Anand et al., 2020). Three paralogs of bactofilin, BacNOP assemble into a polymeric structure and form a complex with the ParB-like protein, PadC, to restrict ParABS machinery to a well-defined position in the subpolar region of the cell. The ParA_Mxa ATPase decorates the entire length of the bactofilin filament using the ParB-like adaptor protein PadC. The pole-distal ends of the bactofilin filaments are bound by the centromeric DNA (parS) binding protein ParB_Mxa ensuring that the genome is equitably transferred after each round of cell division (Lin et al., 2017; Anand et al., 2020).

Discovered in B. subtilis, DivIVA was named so due to the defects in septum placement in its absence (Edwards et al., 2000; Perry and Edwards, 2006; Oliva et al., 2010; van Baarle et al., 2013). It is a tetrameric coiled-coil protein that binds to the membrane by sensing negative curvature (Lenarcic et al., 2009; Ramamurthi and Losick, 2009; Oliva et al., 2010) and is mainly found in Gram-positive bacteria (Lin and Thanbichler, 2013). During sporulation, DivIVA localizes to the poles of the cell and binds to RacA and Soj/SpoOJ (ParAB_Bsu) complex (Wu and Errington, 2003; van Baarle et al., 2013). In addition, DivIVA associates with the peptidoglycan synthesis proteins in Streptomyces, Corynebacterium and Mycobacterium, organizes the growth directing tip-complex and interacts with ParB (Hempel et al., 2008, 2012; Donovan et al., 2012; Ginda et al., 2013; Holmes et al., 2013). Interestingly, a DivIVA bound protein, Scy, associates with the ParA ATPase in S. coelicolor (Ditkowski et al., 2013).

While DivIVA is restricted to Gram-positive bacteria, PopZ (polar organizer protein Z), an evolutionarily unrelated small, acidic protein with an alpha-helical structure present in many Gram-negative species of α-proteobacteria, serves a similar function (Bowman et al., 2008; Ebersbach et al., 2008; Holmes et al., 2016). PopZ has a highly conserved N-terminal and a C-terminal domain, analogous to DivIVA, required for higher-order assembly (Oliva et al., 2010; Bowman et al., 2013). PopZ assembles into polymers that form a network with gel-like properties, which is impermeable to macromolecules (Bowman et al., 2010). However, this 3D network of PopZ at the cell pole captures the ParA released from the nucleoid by the migrating ParB-parS complex. This capture of released ParA is mediated by direct interactions with PopZ (Ptacin et al., 2014). The increased concentration of ParA at the PopZ proximal new pole triggers the ParA DNA binding activity, which reinforces the movement of the ParB-parS complex in the same direction, prevents reversals and immobilizes the origins at cell poles via interactions with ParB. In C. crescentus, cells lacking PopZ have chromosome segregation defects, highlighting the important role played by PopZ in regulating ParA mediated chromosome segregation (Bowman et al., 2008, 2010; Ebersbach et al., 2008; Schofield et al., 2010; Ptacin et al., 2014; Holmes et al., 2016).

In C. crescentus, TipN (Tip of New Pole) is a polarly localized landmark protein and contains a large C-terminal coiled-coiled domain and two transmembrane domains at its N-terminus. TipN relocalises at the cell division site in an FtsZ and FtsI dependent manner late during the pre-divisional stage and is thus present at the new pole in both the daughter cells (Huitema et al., 2006; Lam et al., 2006). At the new pole, TipN recruits TipF, a receptor for the second messenger c-di-GMP [bis-(3′-5′)-cyclic dimeric guanosine monophosphate] and a positive regulator of flagella assembly. However, this interaction between TipN and TipF is restricted in the swarmer cells due to the low level of c-di-GMP but allows the formation of a “flagella organizing center” at the new pole of the stalked cell due to the high levels of c-di-GMP in stalked cells (Davis et al., 2013). Interestingly, while ΔtipN cells exhibit mild chromosome segregation defects (Ptacin et al., 2010), deletion of popZ leads to synthetic lethality in ΔtipN cells due to severe DNA segregation and cell division defects (Schofield et al., 2010). TipN directly interacts with ParA_Ccr at the new pole and regulates the movement and overall speed of the ParB-parS complex (Ptacin et al., 2010; Schofield et al., 2010).

MinD, a component of the min system, is a Walker A ATPase member majorly recruited for regulating cell division in bacteria (Lutkenhaus and Sundaramoorthy, 2003; Lutkenhaus, 2012). The structure of the MinD dimer reveals its similarity to the ParA proteins such as Soj (ParA_Bsu) (Wu et al., 2011). The min locus consists of two other genes, minC and minE, wherein MinC functions as an inhibitor of FtsZ ring assembly via its interaction with MinD (Hu et al., 1999; Hu and Lutkenhaus, 2000; Dajkovic et al., 2008). MinD, along with MinE, undergoes pole-to-pole oscillations and produces a gradient of MinCD division inhibitory complex (Hu and Lutkenhaus, 1999; Raskin and de Boer, 1999; Hale et al., 2001; Meinhardt and de Boer, 2001) with the maximum time-averaged concentration at the cell poles (Lutkenhaus, 2012). MinD, in its ATP bound dimeric form, binds to the membrane via a C-terminal amphipathic helix (Hu et al., 2002; Szeto et al., 2002; Hu and Lutkenhaus, 2003; Zhou and Lutkenhaus, 2003). Following this membrane association, MinE is recruited, which stimulates the ATPase activity of MinD and thus releases it from the bacterial membrane (Hu and Lutkenhaus, 2001; Hu et al., 2002; Park et al., 2011; Wu et al., 2011). Crystal structures of the MinD-MinE complexes reveal the conformational changes in MinE upon interaction with MinD. The structures also show how two released β-sheets in a four-stranded β-sheet MinE dimer convert into α-helices and suggests how MinE remains associated with the MinD-membrane complex (Park et al., 2011). Repeated binding and release cycles of MinD driving oscillation of MinCD complex that acts as a spatial regulator of FtsZ ring in bacteria. The period of the oscillations is dictated by the built-in delays in the system, such as nucleotide hydrolysis and exchange rates in MinD (Howard et al., 2001; Meinhardt and de Boer, 2001; Huang et al., 2003; Kretschmer and Schwille, 2016). Numerous studies have reconstituted the dynamics of the MinDE system in vitro on supported membrane bilayers and observed standing waves of MinD being chased by MinE and recapitulate oscillations within confinement (Loose et al., 2008, 2011; Ivanov and Mizuuchi, 2010; Zieske and Schwille, 2013, 2014; Vecchiarelli et al., 2016).

Further, the deletion of MinD results in the production of anucleate cells suggesting that MinD plays a critical role in chromosome segregation (Di Ventura et al., 2013). Numerical computer simulations have proposed that the min system may contribute to the movement of the chromosome from the mid-cell to the poles and might be mediated by the binding of MinD to DNA (Di Ventura et al., 2013). However, recent in vitro reconstitution experiments of MinDE on membrane bilayers failed to detect any direct recruitment of DNA to MinD, suggesting that the chromosome segregation defects in the absence of Min proteins might be due to indirect effects (Ramm et al., 2018).

Carboxysomes are membrane-bound organelles in Cyanobacteria that help in carbon fixation. These proteinaceous micro-compartments exist in low-copy numbers in the cells and thus depend on partitioning machinery to ensure their transmission during cell division (MacCready et al., 2018, 2020). Fluorescent labeling of the nucleoid, carboxysomes and McdA inside these cells has enabled the tracking of carboxysomes in live cells. Fluorescence imaging of carboxysomes has shown that these organelles are arranged linearly and segregate equally during cell division that was dependent upon a ParA homolog in Synechococcus elongatus (Savage et al., 2010). Recent studies have further shown that the ParA-like Walker A ATPase partitioning machinery, now named McdA (for Maintenance of Carboxysome Distribution), mediates Carboxysome maintenance (MacCready et al., 2018). Although McdA lacks the signature amino-terminal lysine residue (Table 1), it has a strong ATPase and non-specific DNA binding activity. Moreover, another small protein McdB, unrelated to ParB, has been shown to stimulate the ATPase activity of McdA and regulate carboxysome positioning. Although McdB shows no sequence similarity to ParB, it can form higher-order oligomers like ParB (Schumacher et al., 2019). McdB (like ParB) stimulates ATPase activity of McdA driving the directed movement of carboxysome toward a higher concentration of McdA on the nucleoid by a diffusion ratchet mechanism. Thus, Carboxysomes also employ a McdAB protein complex in a manner very similar to the ParAB complex (MacCready et al., 2018).

Certain par loci in bacterial chromosomes contain only parA sequences and lack parB and centromeric sequence parS. These additional par loci are located outside the parAB operon (found close to the oriC) and are referred to as orphan ParA systems. These orphan ParAs can be found in many metabolic operons of bacterial genomes. Also, reports suggest that bacterial genomes encode multiple orphan ParAs. One such orphan ParA, called PpfA, is involved in chemotactic signaling in Rhodobacter sphaeroides. The chemotaxis protein cluster is formed by the partner proteins TlpT and CheW proteins. Mutants in the conserved Walker A motif are known to affect cluster formation. Thus, PpfA helps in the dispersion and segregation of the chemoreceptor clusters (Thompson et al., 2006; Roberts et al., 2012). Remarkably, a distinct clade of ParA-like ATPase is encoded within the chemotaxis operon in V. cholerae (named ParC) and many γ-proteobacteria that have polar flagella (Ringgaard et al., 2011). In V. cholerae, the ParA-like ATPase, ParC, regulates the polar localization of the chemotaxis proteins CheW1 and CheY3 (Ringgaard et al., 2011). The subcellular localization dynamics of ParC are altered by its binding partner, ParP, which promotes the polar assembly of the chemotaxis array (Ringgaard et al., 2011, 2018; Alvarado et al., 2017).

ParI is an Orphan ParA member of the Walker A ATPase family present in the genomic island of Pseudomonas putida (Miyakoshi et al., 2012). The expression of ParI negatively regulates the maintenance of IncP-7 plasmids and results in their loss from cells. Studies on ParI have revealed that mutations in the conserved Walker A motif region (mainly the ATPase domain) of ParI fail to destabilize IncP-7 plasmids. ParI is an example of plasmid-mediated incompatibility residing within a genomic island.

Bacteria produce cellulose as a biofilm matrix polymer to enable the cohesion of biofilms. BcsQ (bacterial cellulose synthesis) proteins help produce cellulose in enterobacteria. BcsQ is a homolog of the ParA/MinD family of ATPase and is activated by cyclic di-GMP mediated signalling. Using fluorescently tagged BcsQ, it has been confirmed that this protein mainly localizes to the cell poles in bacteria and cell-cell adhesion mainly occurs via cellulose production at the cell poles. Thus, a ParA/MinD family of ATPase controls cell-cell adhesion and biofilm formation by regulating cellulose biosynthesis (Le Quéré and Ghigo, 2009).

Flagella serves as a locomotory organ in many organisms. The bacterial flagella is a complex structure requiring ~40 genes for its assembly (Dasgupta et al., 2003). The flagellar genes in bacteria are involved in locomotion, biofilm formation, and pathogenesis (O'Toole and Kolter, 1998; Gellatly and Hancock, 2013; Guttenplan and Kearns, 2013; Mukherjee and Kearns, 2014). The flagella number and distribution are characteristic features of each organism. In P. aeruginosa, the number of flagella is regulated by FleN (Dasgupta et al., 2000, 2003; Köhler et al., 2000; Dasgupta and Ramphal, 2001; van Ditmarsch et al., 2013). FleN, also called FlhG or MinD2, is a ParA/MinD superfamily protein (Table 1), whose absence leads to the upregulation of genes involved in the synthesis of the flagellar motor, basal body, hook proteins etc. and conferring the multi-flagellate phenotype (Dasgupta et al., 2003; Chanchal et al., 2017). FleN acts antagonistic to FleQ and is known to inhibit FleQ dependent transcription. Such effect is essential for maintaining the correct number of flagella in the cell (Dasgupta and Ramphal, 2001; Chanchal et al., 2017). Homologs of FleN are also present in many other bacteria, including B. subtilis, V. cholerae, Geobacillus thermodenitrificans, and Campylobacter jejuni (Correa et al., 2005; Guttenplan and Kearns, 2013; Schuhmacher et al., 2015; Gulbronson et al., 2016).