Spatiotemporal organization of microbial cells by protein concentration gradients Daniela Kiekebusch1,2,3 and Martin Thanbichler1,2,3 1
Max Planck Institute for Terrestrial Microbiology, Marburg, Germany Faculty of Biology, Philipps-Universita¨t, Marburg, Germany 3 LOEWE Center for Synthetic Microbiology, Marburg, Germany 2
The formation of protein concentration gradients is an effective means to restrict the activity of regulatory factors in space, thereby critically contributing to the spatiotemporal organization of biological systems. Although widely observed for extracellular proteins involved in tissue patterning, the implementation of this regulatory strategy was thought to be impossible in single, micron-sized cells. Recently, however, several intracellular proteins were shown to establish gradient-like distribution patterns, thereby relaying positional information to their downstream targets. In this review, we discuss gradient-forming systems from different microbial species, with an emphasis on their mode of action and the common principles that underlie their function. Protein gradients on the micron scale Concentration gradients can be defined as a gradual decrease in the concentration of molecules with increasing distance from their source. In biology, the spatial information encoded by such gradients is typically used to direct or restrict cellular processes to particular subcellular positions or to a limited number of cells. For instance, protein gradients have emerged as a common strategy to control the spatial patterning of tissues in developmental biology. Among the best studied examples are gradients of extracellular morphogens . These proteins are synthesized by a defined subset of cells only to be spread by diffusion to neighboring cells, where they typically elicit a concentration-dependent change in gene expression, ultimately leading to the formation of tissues and organs [2–7]. As exemplified by the well-studied Bicoid morphogen controlling anterior–posterior patterning in the Drosophila embryo, these gradients can extend over several hundreds of microns from a localized source . Their shape is thought to Corresponding author: Thanbichler, M. ([email protected]
). Keywords: Pom1; MipZ; ParA; cell size homeostasis; division site placement; plasmid segregation; plasmid P1; nucleoid; Caulobacter crescentus. 0966-842X/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tim.2013.11.005
be determined mainly by protein diffusion and concomitant uniform degradation , although other mechanisms are likely to be involved [8–11]. Because diffusion is fast, especially within the continuous cytoplasm [12–14], it was long assumed that similar concentration gradients could not be maintained infinitely at the scale of a single micron-sized cell. Moreover, proteolysis is too slow a process to be employed as a shape determinant in gradient-forming systems at the single cell level, where many reactions are diffusion-limited . However, theoretical modeling predicted that the establishment of intracellular activity and total protein gradients should be possible, although the underlying principles needed to be different from those shaping morphogen gradients [16,17] (Box 1). Confirming this prediction, the past decade has revealed several eukaryotic systems in which steady-state intracellular gradients provide positional information within individual cells. For instance, morphogenesis of the mitotic spindle was found to depend on a gradient in the phosphorylation state of the tubulin regulator stathmin . In addition, a total protein concentration gradient of the serine/threonine kinase Pom1 emerged to participate in chromosome segregation and cell division in the fission yeast Schizosaccharomyces pombe [19,20]. Moreover, a mechanistically similar system was shown to control the gradient-like distribution of the cell fate determinant MEX-5 in the single-celled Caenorhabditis elegans embryo . Nowadays, it is appreciated that prokaryotes exhibit a complex internal organization as well. This change of view is attributable to the identification of cytoskeletal elements, microcompartments, and an ever increasing set of proteins that are localized to specific subcellular regions [22–25]. Thus, the discovery of protein concentration gradients in these microscopic organisms does not seem surprising. For instance, upon nitrogen starvation, filamentous cyanobacteria initiate a developmental program leading to the differentiation of vegetative cells into N2fixing heterocysts . These specialized cells are distributed at regular intervals within filaments of vegetative cells. Whether or not a cell becomes a heterocyst is decided by a gradient of two negative regulators of heterocyst formation. Reminiscent of eukaryotic morphogens, these factors are synthesized by pre-existing heterocysts and diffuse along the filament into neighboring cells. The Trends in Microbiology, February 2014, Vol. 22, No. 2
Review Box 1. Prerequisites for the establishment of a concentration gradient in a micron-sized cell The main determinants shaping morphogen gradients are diffusion from a localized source and constant nondirectional degradation. According to a simple mathematical model , a morphogen gradient can thus be described by: rﬃﬃﬃﬃ D l¼ [I] k whereby D is the diffusion coefficient of the morphogen, k the degradation rate of the morphogen, and l the decay length of the gradient (i.e., the distance from the source at which the concentration of the morphogen has decayed to 1/e of its initial concentration). Instead of the degradation rate, the stability of a protein is often expressed by its half-life t½, which is related to k by k = ln2/t½. Hence, substituting k in Equation [I] results in: pﬃﬃﬃﬃﬃﬃﬃﬃﬃﬃ l ¼ Dt 12 [II] For a typical soluble protein with an average diffusion coefficient of 10 mm2/s [12–14] and a half-life of 1 h [71,72], Equation [II] yields a decay length of 230 mm, which is in good agreement with the experimentally determined range of some morphogen gradients [5,7,73]. Thus, degradation coupled with diffusion appears to be a suitable mechanism for the establishment of a morphogen gradient. However, a decay length in this range does not allow formation of an intracellular gradient because most cells are at least one order of magnitude smaller. In order to reduce the length of the gradient to the dimensions of a single cell, l needs to be much smaller. How can this be achieved? Assuming again a diffusion coefficient of 10 mm2/s and a cell length of 2 mm, this condition will only be fulfilled if k 2.5 s1 (see Equation [I]). Conversely, for a protein with a half-life of 1 h, a diffusion coefficient D 8 104 mm2/s would be required to give rise to an intracellular gradient. Thus, a gradient at the single cell level can be established by reducing the diffusional mobility of the protein, by increasing the decay rate of the protein, or by a combination of both. In vivo, a significant decrease in the diffusion rate can be realized by association of a protein with immobile structures such as membranes or chromatin . In addition, it is possible to decrease the half-life of the gradient-forming species by limiting gradient formation to a specifically modified form of the protein whose inactivation relies on an enzymatic reaction that is much faster than proteolysis. An example of this strategy are cycles of protein phosphorylation and dephosphorylation, which can reach rates of up to 800/s [16,75] (see also Box 2).
minimum of the resulting gradient, which lies halfway between two heterocysts, then defines the vegetative cell that will transform into a new heterocyst [26–29]. In addition to this morphogen-like gradient, at the single cell level, total protein and protein phosphorylation gradients have been found to operate in bacteria as well. For instance, the virulence factor IcsA of Shigella flexneri, which is crucial for efficient transmission of the pathogen between host cells, forms a concentration gradient on the outer membrane, with a maximum at one of the cell poles . This pattern was proposed to be generated by polar targeting of IcsA, subsequent lateral movement along the outer membrane, and concurrent uniform degradation, but the precise mechanism and the potential impact of other factors on gradient formation still need to be explored. Another example is given by the developing bacterium Caulobacter crescentus, where a gradient in the phosphorylation of the transcriptional regulator CtrA contributes to replicative asymmetry  (Box 2). C. crescentus also provides an example for a total protein gradient, formed by the ATPase MipZ, which mediates the 66
Trends in Microbiology February 2014, Vol. 22, No. 2
spatial regulation of cell division in this species . Finally, recent studies have revealed a critical role for gradients of regulatory, ParA-like ATPases in the partitioning of low-copy number plasmids in Escherichia coli [33,34]. In this review, we discuss recently identified gradientforming systems in microorganisms. We highlight their biological roles and the current state of knowledge about their molecular function. In particular, we will focus on three systems that have already been characterized in more detail: the Pom1 gradient in the fission yeast S. pombe, the MipZ gradient in C. crescentus, and the parABS system of the E. coli P1 plasmid (Table 1). Based on these examples, we will then outline common regulatory principles that are at work during the establishment of stable protein gradients within cells. Regulation of cell size homeostasis in S. pombe The first microbial protein gradient to be analyzed in detail is formed by the protein kinase Pom1 in the fission yeast S. pombe. Pom1 localizes in a cortical bipolar gradient with maxima at the poles and a minimum in the cell center [35,36] (Figure 1A). Apart from its implication in the establishment of cell polarity and division site placement, Pom1 has also been associated with the regulation of mitotic entry of S. pombe cells [19,20,35,37]. The latter function is tightly coupled to its gradient-like distribution and relies on a negative effect of Pom1 on the signaling network controlling mitotic entry. More specifically, Pom1 directly inhibits the function of the protein kinase Cdr2, a positive regulator of mitosis that localizes in a band of cortical nodes at midcell [19,20]. In small cells, the Pom1 gradient overlaps with Cdr2, leading to a cell cycle arrest. As the cell grows and the poles move apart, the concentration of Pom1 at midcell decreases, thereby finally relieving the inhibition of Cdr2 and initiating a signaling cascade that culminates in cell division (Figure 1A,B). In accordance with this model, Pom1-deficient cells were shown to consistently divide at a smaller size than the wild type, indicating premature entry into mitosis. By contrast, slight overexpression of pom1 delayed cell division with cells being larger than the wild type at the onset of cytokinesis [19,20]. Thus, the Pom1 gradient controls cell size homeostasis by coupling cell division to cell elongation. Although these studies provided a comprehensive model for cell size control by Pom1, the molecular mechanism underlying formation of the Pom1 gradient remained elusive. Recently, however, evidence was provided that gradient formation depends on a dynamic localization cycle that is driven by reversible Pom1 phosphorylation  (Figure 1C). In a first step, phosphorylated Pom1 is recruited to the tips of the cell by the polar landmark protein Tea4 . The concomitant interaction of Tea4 with the protein phosphatase Dis2  then stimulates the dephosphorylation of Pom1. This change in the modification state enhances the accessibility of a basic lipidbinding region, resulting in the relocation of Pom1 to the cytoplasmic membrane. Starting at the cell poles, Pom1 slowly diffuses laterally along the membrane towards midcell. Simultaneously, it autophosphorylates at several sites within its lipid-binding region, which
Trends in Microbiology February 2014, Vol. 22, No. 2
Box 2. CtrA phosphorylation: an example of an intracellular activity gradient Caulobacter crescentus is well known for its asymmetric life cycle where cytokinesis produces two distinct daughter cells: a replicationcompetent stalked cell and a swarmer cell that does not duplicate its chromosome. This replicative asymmetry is governed by a gradient in the phosphorylation state of the transcriptional regulator CtrA [31,76]. In its phosphorylated form (CtrAP), CtrA inhibits initiation of DNA replication by binding to specific sites in the origin region . In predivisional cells, which possess two copies of the chromosome, CtrAP is concentrated at the swarmer pole whereas dephosphorylated CtrA is prevalently found at the stalked pole. Because it is not yet possible to easily visualize protein activity gradients in vivo, the skewed subcellular distribution of CtrAP was verified indirectly by showing that the origin at the stalked pole in predivisional cells is more active than the one at the swarmer pole . Thus, although CtrA is homogeneously distributed in predivisional cells, active CtrAP is asymmetrically localized (Figure I). This gradient in the phosphorylation state of CtrA is created by differential activities of the hybrid histidine kinase CckA at the two cell poles. Although serving as a kinase at the swarmer pole, CckA is active as a phosphatase at the stalked pole, thereby establishing a source and sink for CtrAP at opposite ends of the cell. The resulting asymmetry in the replicative competence of both origins of a predivisional cell is further pronounced by cell division, which confines the kinase and phosphatase activities of CckA to different daughter cells. As a result, the stalked cell, in which CtrAP levels are low, immediately starts a new round of DNA replication after cytokinesis, whereas the swarmer cell is arrested in G1 phase. In agreement with the experimental data, mathematical modeling predicted that a protein phosphorylation gradient can arise from spatially separating the phosphorylation and dephosphorylation reaction. Moreover, these activities need to be faster than the diffusion of the protein [16,31,78]. For CtrA, which has a diffusion
gradually weakens its association with the membrane, eventually causing its detachment from the cell cortex. Phosphorylated Pom1 then diffuses rapidly within the cytoplasm until it finally reassociates with Tea4 at the poles. In summary, the Pom1 gradient is initiated by the polar recruitment of Pom1 and shaped by the lateral movement of molecules along the plasma membrane and its concurrent, autocatalytic conversion into a membrane-detached state. Importantly, only the membrane-bound fraction of Pom1 is able to productively interact with Cdr2 to inhibit mitotic entry, which ensures proper interpretation of the Pom1 gradient by the cell cycle regulatory machinery. Interestingly, Pom1 is not entirely uniformly distributed in the polar regions but forms dynamic clusters at the membrane . Mathematical modeling showed that protein clustering is a mechanism employed by cells in order to buffer against noise [40,41]. In this context, noise is defined as fluctuations in the Pom1 gradient, for example, due to differences in protein levels between single cells or varying diffusional mobility of proteins within a given cell. Reducing the influence of these variations by dynamic protein clustering, the Pom1 gradient is able to provide precise spatial cues for the mitotic machinery, resulting in a population of evenly sized cells. Spatiotemporal regulation of cell division in C. crescentus In the asymmetric alpha-proteobacterium C. crescentus, DNA replication and cell division are closely coordinated in time and space. The coupling of both processes is mediated by a bipolar gradient of the protein MipZ, a representative
constant of approximately 1–10 mm2/s, the phosphorylation and dephosphorylation rates were estimated to be 50–100/s . These rates are in accordance with the requirements for the establishment of a stable intracellular phosphorylation gradient (see also Box 1). Thus, the CtrA system exemplifies that stable gradients in the activity of a protein can confer cellular asymmetry even if the protein itself exhibits spatial homogeneity. It also highlights the fact that the even distribution of most cytoplasmic proteins does not exclude the existence of underlying activity gradients, based on differential phosphorylation or, alternatively, distinct conformational states induced by the binding of nucleotides or other cofactors [79,80].
Key: CckA (kinase) CckA (phosphatase)
CtrA~P CtrA Cytokinesis
Stalked cell TRENDS in Microbiology
Figure I. Establishment of the CtrA phosphorylation gradient in Caulobacter crescentus.
of the so-called ParA-like ATPases . Similar to small G proteins, the members of this ATPase family act as molecular switches, alternating between a monomeric and dimeric form with distinct activities (Box 3). The same is true for MipZ, which depending on its oligomerization state interacts with components of either the DNA segregation or cell division machinery. Shortly after the initiation of DNA replication, which only occurs once during the C. crescentus division cycle, the duplicated origin regions are rapidly segregated to opposite cell poles. This process is mediated by a conserved tripartite segregation system, comprising a centromer-like DNA sequence ( parS), a DNA-binding protein (ParB), and a partition ATPase (ParA) . Several parS motifs cluster in the origin–proximal region of the chromosome and are specifically recognized by ParB [43,44]. ParA then interacts with the resulting ParB–parS nucleoprotein complex to direct the segregation process [45,46]. Interestingly, MipZ was found to be recruited to ParB as well. As a consequence, segregation and polar attachment of the origin-bound ParB– parS complexes gives rise to a bipolar gradient-like distribution of MipZ, with concentration maxima at the poles and a minimum at the cell center (Figure 1D). Moreover, MipZ directly interacts with the prokaryotic tubulin homolog FtsZ , a GTPase whose polymerization into the so-called Zring sets the basis for the assembly of the cell division machinery . In vitro experiments showed that MipZ inhibits the GTP-dependent polymerization of FtsZ, at least in part by stimulating its GTPase activity [32,48]. As a result, Z-ring formation and, hence, cell division are directed toward the subcellular site that displays the lowest MipZ 67
Trends in Microbiology February 2014, Vol. 22, No. 2
Table 1. Intracellular protein concentration gradients discussed in this review Gradient-forming protein Pom1
Serine/threonine protein kinase
In vivo function of gradient Regulation of cell size homeostasis in Schizosaccharomyces pombe Spatial regulation of cell division in Caulobacter crescentus P1 plasmid segregation in Escherichia coli
concentration, which is located halfway between the poles at approximately midcell (Figure 1E). Consistent with a key role of MipZ in division site placement, abolishment of the MipZ gradient by overexpression of mipZ prevents FtsZ assembly throughout the cell, leading to a complete and immediate block in cell division. Depletion of MipZ, by contrast, allows Z-ring formation at random cellular locations. A recent study has clarified the role of ATP binding and hydrolysis in the function of MipZ and established a model that describes the molecular mechanism underlying MipZ gradient formation . Collectively, the MipZ gradient was shown to reflect the distribution of nucleoid-associated MipZ dimers (Figure 1F). As a central component of the system, the ParB–parS complex not only recruits MipZ monomers to the cell poles but also stimulates their ATPdependent dimerization. MipZ dimers are released from ParB and bind nonspecifically to chromosomal DNA, which forms a dense meshwork extending throughout the C. crescentus cell. Importantly, DNA binding drastically reduces the diffusion rate of MipZ, thereby leading to the preferential retention of MipZ dimers in the pole–proximal cellular regions. Mutations in the catalytic site that stabilize the monomeric or dimeric state cause an increase or decrease in the diffusion rate of MipZ, respectively, and abolish its typical gradient-like distribution . Thus, reminiscent of the Pom1 system, both the polar source and reduced mobility of MipZ dimers are critical to the formation of the concentration gradient. Spontaneous ATP hydrolysis triggers dissociation of dimers into monomers, which diffuse freely in the cytoplasm until they undergo nucleotide exchange and are recaptured by the polar ParB–parS complexes. The intrinsic ATPase activity of MipZ thus functions as a pacemaker for the lifetime of the nucleoid-bound dimeric complex, similar to the autophosphorylation activity of Pom1 restricting the lifetime of membrane-associated Pom1. Importantly, only the dimeric form of MipZ productively interacts with FtsZ, ensuring that the positional information encoded in the gradient of MipZ dimers can be decoded accurately to faithfully position the division site. It is interesting to note that the MipZ gradient is slightly asymmetric, with its maximum shifted towards the new cell pole, in line with a similar bias in the positioning of FtsZ [45,46]. The mechanistic basis of this effect, however, is still unclear. Segregation of a low-copy plasmid in E. coli Similar to bacterial chromosomes, many low-copy number plasmids take advantage of a three-component parABS 68
Matrix for gradient Cytoplasmic membrane
Landmark complex Tea4–Dis2
Cellular target Cdr2
system to ensure their faithful transmission to daughter cells upon cell division. However, despite considerable effort, the precise mechanism underlying the function of these systems is still incompletely understood. As a model, the plasmid prophage P1 of E. coli has been the focus of intensive research for many years. P1 exhibits a dynamic localization pattern, which leads to an even distribution of its copies within the cell [49,50]. Recent work has provided evidence that this process is driven by nucleoid-associated concentration gradients of the partition ATPase ParA [33,51] (Figure 1G,H). ParA has nonspecific DNA binding activity and, in addition, interacts with a nucleoprotein complex formed by ParB and multiple plasmid-borne parS motifs [33,50,52,53]. It binds dynamically to the nucleoid and the P1 ParB–parS complex in vivo [50,54], thereby retaining plasmid copies in the nucleoid region [52,54,55]. Importantly, this process relies on a stimulatory effect of ParB on the ATPase activity of ParA [53,56–58]. Based on these observations, plasmid movement was proposed to be mediated by a dynamically changing, skewed distribution of ParA molecules on the nucleoid , but the mechanistic details remained controversial. In contrast to ParA proteins from other systems, which were hypothesized to form filaments that drive plasmid segregation through a mitotic-like pulling mechanism [59– 63], P1 ParA was not observed to polymerize [33,51]. Recent in vitro reconstitution studies of the P1 system suggest that it may rather control the partitioning process through a diffusion-ratchet mechanism in which the random diffusional motion of plasmids is constrained by dynamic nucleoid-bound ParA molecules  (Figure 1I). In this model, ParA–ATP associates nonspecifically with chromosomal DNA and concurrently interacts with the plasmid-bound ParB–parS partition complex, thereby attaching plasmid copies to the nucleoid surface. In these quaternary complexes, ParB stimulates the ATPase activity of ParA, generating ParA–ADP, which detaches from the DNA. This process gradually reduces the number of contacts between the ParB–parS complex and the remaining nucleoid-bound ParA–ATP dimers. As a consequence, the plasmid starts to move in a circular Brownian motion around its anchor point, leading to the ParB-stimulated detachment of additional ParA–ADP molecules in the nucleoid regions surrounding it. Importantly, after nucleotide exchange, ParA–ATP has to undergo a series of slow conformational changes before regaining the ability to bind nonspecifically to DNA , preventing its reassociation with the nucleoid in the immediate vicinity of the plasmid
Trends in Microbiology February 2014, Vol. 22, No. 2
0 5 30
Pom1 –tdTomato Cdr2 –GFP
Cell cycle delay
Cell cycle progression
P A ParA
FtsZ ring assembly Smulaon of ParA Stochasc reassociaon of plasmid with nucleoid ATPase acvity at high ParA concentraons
MipZ–ATP (dimer) MipZ–ADP (monomer) Nucleoid
Key: Pom1 phosphorylaon state
ParA–ATP ParA–ADP Nucleoid TRENDS in Microbiology
Figure 1. Comparison of different gradient-forming systems. (A) The subcellular localization of the protein kinase Pom1 and its regulatory target Cdr2 in Schizosaccharomyces pombe. (B) The mechanism of cell size homeostasis in S. pombe. In small cells, the concentration of Pom1 at midcell is high, leading to phosphorylation of the protein kinase Cdr2. Phosphorylated Cdr2 is unable to inhibit the protein kinase Wee1, which in turn leads to hyperphosphorylation and, hence, inhibition of the cyclin-dependent kinase Cdc2, thereby preventing entry into mitosis. In large cells, the concentration of Pom1 at midcell is low, which impedes phosphorylation of Cdr2, thus ultimately allowing progression of the cell cycle into mitosis. (C) The mechanism underlying the formation of the Pom1 gradient. Phosphorylated Pom1 (white spheres) is a soluble protein that diffuses freely in the cytoplasm. It is captured at the poles by the landmark protein Tea4 and dephosphorylated by the polar phosphatase Dis2, thereby gaining affinity for the cytoplasmic membrane. After binding to the pole–proximal regions of the membrane, dephosphorylated Pom1 (red spheres) slowly diffuses along the cell cortex towards midcell. On its way, it gradually autophosphorylates at multiple residues in its membrane targeting region. These modifications gradually weaken its membrane affinity and finally cause its release from the cortex, thereby restarting the cycle. (D) The subcellular localization of the cell division inhibitor MipZ and its regulatory target FtsZ in Caulobacter crescentus. (E) The regulation of division site placement by MipZ in C. crescentus. Assembly of the cell division apparatus requires polymerization of the tubulin homolog FtsZ into the so-called FtsZ ring. The site of FtsZ ring formation is regulated by the ATPase MipZ, which forms a bipolar gradient, with maxima at the cell poles and a minimum at the cell center. Because MipZ acts as an inhibitor of FtsZ polymerization, its gradient-like distribution prevents FtsZ ring assembly in the pole–proximal regions, thereby effectively directing cytokinesis to midcell. (F) The mechanism of MipZ gradient formation. MipZ monomers (white spheres) diffuse freely in the cytoplasm. They are captured by the polar ParB–parS complexes, where they are stimulated to dimerize in an ATP-dependent manner. MipZ dimers (red spheres), which have nonspecific DNA binding activity, dissociate from ParB and interact with the matrix of chromosomal DNA that extends throughout the cell, with their density decreasing as a function of distance from the poles. The slow intrinsic ATPase activity of MipZ eventually leads to dissociation of the dimer, generating monomers that detach from the DNA, undergo nucleotide exchange, and then restart the cycle. (G) Localization dynamics of the Escherichia coli P1 plasmid-partitioning ATPase ParA and the cognate ParB–parS complex in an in vitro reconstituted system. The ParB–parS complex is attached to the surface by ParA bound to a carpet of nonspecific DNA. The motion of the plasmid around its attachment site leads to progressive depletion of ParA in its surroundings. Once all ParA molecules associated with the plasmid have been removed, the plasmid dissociates from the surface (bottom rows). (H) Partitioning of the P1 plasmid through dynamic local gradients of the ATPase ParA. The ParB–parS complex interacts with ParA attached to the nucleoid surface, generating a ParA depletion zone. Once all tethers between the plasmid and the nucleoid have been cut, the plasmid moves along the gradient of ParA molecules in a randomly chosen direction until the density of ParA becomes high enough to attach it stably to the nucleoid again. (I) The mechanism underlying the formation of local ParA gradients. ATPbound ParA has nonspecific DNA binding activity and associates with the nucleoid (red spheres). The ParB–parS complex interacts with ParA–ATP and thus attaches to the nucleoid surface. Its stimulatory effect on the ATPase activity of ParA, combined with Brownian motion of the plasmid, leads to dissociation of ParA–ADP from the nucleoid in the vicinity of the plasmid anchor point. ParA–ADP molecules (white spheres) need to undergo nucleotide exchange and a series of conformational changes before they can bind DNA again, which prevents their immediate re-association with the nucleoid and thus facilitates the formation of the ParA depletion zone. Panels (A) and (B) adapted from ; panel (C) adapted from ; images in panel (D) by D. Kiekebusch; and panel (G) adapted from .
cluster. This lag results in the generation of a ParA depletion zone, in which the concentration of DNA-bound ParA gradually increases with increasing distance from the initial anchor point of the plasmid cluster. The spatial information encoded in the ParA gradient is translated
into directed motion once all contacts between the plasmid and the nucleoid have been removed. Initially, the plasmid moves in a stochastically chosen direction, where it encounters an increasing number of nucleoid-associated ParA–ATP molecules. Because transient local interactions 69
Trends in Microbiology February 2014, Vol. 22, No. 2
Box 3. P-loop ATPases that function as molecular switches MipZ and ParA are evolutionary related to P-loop GTPases (G proteins) . Within this large and diverse group of nucleotide hydrolases, the homologs of ParA and MipZ along with the closely related MinD proteins (Box 4) form distinct subfamilies, characterized by unique structural features. Recently, another classification of P-loop GTPases and related ATPases was proposed, which relies on their mode of action rather than on structural similarities . This approach separated P-loop proteins into conventional G proteins and G proteins activated upon dimerization (GADs). MipZ, ParA, and MinD unambiguously share common functional features with GADs. Similar to GADs, they dimerize in a nucleotide-dependent manner via the regions containing the ATP binding site (Figure I). The ATP-bound dimer represents the biologically active form of the protein that interacts with its target. Moreover, only the dimeric complex has catalytic activity, with nucleotide hydrolysis converting the protein into the nonactive, ADP-bound monomeric form. Similar to G proteins, the ATPase cycles of MipZ, ParA, and MinD are regulated by effectors that stimulate dimerization or ATP hydrolysis. However, there is no strict differentiation between targets and effectors, as observed for conventional G proteins, because the target can also act as an ATPase-activating factor (as is the case for ParA). Based on nucleotide binding and hydrolysis, MipZ, ParA, and MinD function as molecular switches that cycle between an ATPbound ‘on’ state and an ADP-bound ‘off’ state. In doing so, they can establish spatiotemporally organized interaction networks that relay spatial information to the cell division machinery and plasmid clusters, respectively.
Pi TRENDS in Microbiology
Figure I. Switch-like behavior of ParA-like P-loop ATPases.
with the freely diffusible partition complex do not remove ParA from the DNA and, thus, do not influence the ParA gradient, the direction of movement is reinforced by the increasing number of contacts between the plasmid and DNA-bound ParA. Finally, the plasmid becomes tethered to the nucleoid again, while the ParA depletion zone in its wake is slowly refilled with ParA–ATP. Although this model for P1 partitioning agrees well with the in vitro observations, it has so far not been possible to visualize a ParA concentration gradient in live cells [50,54]. This might be attributable to the limited size of plasmid clusters in vivo, which contain on average only one or two copies . Therefore, the depletion zones may normally be too small to be visualized by conventional fluorescence microscopy, consistent with the fact that they are most readily observed for bigger clusters in vitro . It was, however, possible to detect in vivo depletion zones for a homologous partition system, encoded by the E. coli F plasmid . The localization dynamics of the F plasmid are similar to those of P1, and in vitro reconstitution 70
studies suggested that the two plasmids use similar segregation mechanisms . However, there are intriguing questions that remain to be answered. It is still incompletely understood how the diffusion-ratchet mechanism can achieve the equipositioning of multiple plasmids within the cell and how individual plasmids are released from plasmid clusters to enable their subsequent segregation. We anticipate that mathematical modeling of the P1 system might help address these questions and provide directions for future research. Besides, it will be interesting to see whether the segregation of bacterial chromosomal origins , carboxysomes  or chemotaxis clusters , which are driven by ATPases closely related to P1 ParA, could be based on the same mechanistic principles as P1 partitioning. A common scheme for intracellular concentration gradients The intracellular concentration gradients discussed here have different functions in phylogenetically distinct organisms, including a single-celled eukaryote and two bacterial species. Nevertheless, all three systems appear to rely on the same mechanistic principles. Their function is based on three basic components: (i) a regulator (Pom1, MipZ, ParA) relaying positional information to a downstream target by means of its gradient-like distribution; (ii) a cellular scaffold (cytoplasmic membrane, nucleoid) on which the gradient is established; and (iii) a landmark complex (Tea4–Dis2, ParB–parS) that locally changes the diffusional properties of the regulator (fast vs. slow diffusing) (Figure 1C,F,I). The slow diffusing form of the regulator, generated by a switch in its conformational or modification state, spreads over the scaffold in a gradient-like pattern and mediates the specific biological function of the system. The lifetime of this species is limited by the enzymatic activity of the regulator that, either spontaneously or upon stimulation, triggers its conversion into the nonactive, fast-diffusing form. Because this activity is rather low, most of the regulator molecules will be in the active state at any given time. The shape of the gradient is determined by the combined action of the localized landmark complex and the enzymatic activity of the regulator. The landmark complex sets a limit to the gradient by either defining the source or the sink of the gradientforming regulator species. In the case of the Pom1 and MipZ systems, it defines the maximum of the gradient by catalyzing the formation of the slow-diffusing, active form, whereas in the P1 ParA system it sets the minimum by stimulating the generation of the fast-diffusing, non-DNA binding form. The induced or endogenous catalytic activity of the regulator, by contrast, determines the dimensions of the gradient, acting as timing device that defines the lifetime of the gradient-forming species. Collectively, the gradients thus reflect the distribution of the slow diffusing, matrix-bound, active regulatory species, which is shaped by the concerted action of a landmark complex and an enzymatic activity inherent to the regulator. Despite these overall similarities, there are also variations to this scheme that reflect the specific biological function of each gradient. For example, both Pom1 and MipZ are negative regulators that form steady-state bipolar gradients, restricting the activity of their target to
Review midcell. These gradients emerge from a polar source that is defined by a cognate landmark protein, with the concentration of the regulator decreasing with increasing distance from the landmark protein. By contrast, the P1 ParA gradient is transient and approximately one order of magnitude shorter than the Pom1 and MipZ gradients. Moreover, it has an inverse orientation, with the concentration of ParA gradually rising with increasing distance from the ParB–parS landmark complex, which acts as a sink for scaffold-attached regulator molecules. Intriguingly, in the P1 system, the ParB–parS complex represents both the landmark complex and the downstream regulatory target. As a consequence, the ParA gradient provides relative positional information, whereas the Pom1 and MipZ systems relay absolute spatial information to their targets. Thus, despite the existence of common principles, variations in the interactions between the different players lead to distinct dynamics and thus varying, system-specific outputs. Concluding remarks Within cells, positional information has long been thought to be primarily communicated by cytoskeletal elements or localized proteins. However, in recent years, intracellular protein concentration gradients have emerged as another versatile means to control the arrangement of cellular components. Although evolutionarily and functionally distinct, the three examples presented in this review suggest the existence of a general scheme for the establishment of intracellular gradients. The mechanistic principles involved apply to both prokaryotic and eukaryotic cells, underscoring the generality and robustness of this regulatory strategy. Notably, mathematical modeling indicates that gradient-forming systems are able to gauge distances
Trends in Microbiology February 2014, Vol. 22, No. 2
more robustly than spatially uniform systems . This accuracy and the unique ability to provide spatial information about the cell may explain the common use of regulatory protein gradients in biology. To date, the known examples of intracellular protein concentration gradients are limited to relatively small, single-celled organisms (Pom1, MipZ, ParA) or one-cell embryos (MEX-5). However, in general, the mechanistic principles at work in these systems should also be implementable in cells that function as part of a tissue in multicellular species. Similar to protein activity gradients, which are a common theme in the regulation of mitosis , gradients in the concentration of regulatory proteins are thus likely to be identified in such more complex settings in the future. Importantly, the design principles underlying intracellular protein gradients differ significantly from those shaping extracellular morphogen gradients. Whereas the lifetime of a morphogen critically depends on the rate of protein degradation, intracellular gradient-forming systems take advantage of regulatory components whose activity is controlled by reversible modification, involving an autocatalytic enzymatic reaction. Intriguingly, the two prokaryotic examples of protein concentration gradients both involve a ParA-like (P-loop) ATPase. Considering that another member of this ATPase family, MinD, is used to establish a timeaveraged concentration gradient that places the division site in E. coli (Box 4), this type of enzyme appears to be especially suitable as the central regulatory component of gradient-forming systems in prokaryotes. Interestingly, there are recent indications for the establishment of concentration gradients by intrinsically stable cytoskeletal structures . It will be interesting to unravel the mechanistic concepts underlying these
Box 4. The MinCDE system: a time-integrated concentration gradient In Escherichia coli, polar cell division events are prevented by the MinCDE system . Central to this system is MinD, a P-loop ATPase closely related to MipZ and ParA [81,84]. MinD binds to the cytoplasmic membrane in the polar regions of the cell to which it recruits the cell division inhibitor MinC. The resulting MinCD complex shows a remarkable, rapid pole-to-pole oscillation, which gives rise to a timeaveraged MinCD concentration gradient that has a defined minimum at midcell . As a result, the cell division machinery is assembled in the cell center (Figure I). The establishment of this time-integrated gradient is dependent on the ATPase activity of MinD and on MinE, which is a topological specificity factor for MinD. Similar to MipZ and ParA, MinD operates by a molecular switch mechanism (see also Box 3). ATP binding induces dimerization of MinD and subsequent association with the cytoplasmic membrane [86,87]. Membrane-bound MinD then recruits the cell division inhibitor MinC, thereby preventing cell division. The switch in the activity of MinD is regulated by its ATPase activity, which in turn is modulated by MinE and phospholipids [87–89]. Interestingly, MinE preferentially interacts with the MinD molecules that are closest to the center of the cell. When interacting with MinD, MinE displaces MinC from MinD. At the same time, it stimulates the ATPase activity of MinD, thereby leading to the release of MinD monomers from the membrane. It is thought that MinE can immediately rebind to adjacent membrane-associated MinD molecules and, thus, progressively disassemble the polar MinCD cap. By contrast, MinD first has to exchange ADP for ATP followed by ATP-dependent dimerization before being able to reassociate with the membrane. Although the precise mechanism is not yet clear, membrane association occurs preferentially at the cell pole that is devoid of MinCD.
Repetition of this cycle gives rise to the characteristic oscillatory behavior of MinCD and, consequently, its time-averaged gradient-like distribution. Interestingly, there are striking analogies between the MinCDE and P1 ParA systems. It is therefore tempting to speculate that the MinE-driven retraction of MinCD might also function by a diffusionratchet mechanism, driven by a transient short-range spatial gradient in MinCD concentration at the membrane .
Key: MinD–ATP MinD–ADP MinC MinE FtsZ
TRENDS in Microbiology
Figure I. Division site placement by the Escherichia coli Min system.
Review Box 5. Future challenges Although protein concentration gradients are often relatively shallow, the spatial information they encode is read out with high accuracy, resulting in the restriction of target activity to a precisely defined position within the cell. It will be interesting to investigate how this precision is achieved at the molecular level. To elucidate the function of gradient-forming systems, it is essential to identify the minimal set of components required for their establishment. Only in vitro reconstitution of the systems in a close-to-native environment will be able to address this issue. Mathematical modeling is necessary to obtain a comprehensive, quantitative understanding of gradient-forming systems. To verify the models, it will be necessary to determine the reaction and interaction kinetics of their components using both in vitro and live-cell approaches. Protein concentration gradients may often act at a sub-micrometer scale and thus escape detection by conventional microscopic methods. Advances in super-resolution microscopy may help to identify new gradient-based regulatory systems in vivo. Protein activity gradients cannot be readily detected by fluorescence microscopy unless diffusion of the protein is affected by the modification. Efforts have been made to directly visualize activity gradients by means of immunocytochemistry using phosphorylation state-specific antibodies or by the application of fluorescence resonance energy transfer (FRET)-based sensors. Future endeavors should aim at developing additional sensitive in vivo probes to monitor protein activity. In the systems characterized so far, the mobility of the gradientforming protein species is limited through interaction with membranes or DNA. It will be interesting to see whether there are gradient-forming systems that employ different mechanisms to reduce the diffusion rate of proteins within the cell.
gradients (Box 5) because they have to be different from those shaping the gradients of mobile proteins. In any way, the gradient-forming systems analyzed thus far may illustrate only the beginning of what might emerge to be the rule rather than the exception. Acknowledgments We thank Sophie Martin for providing figures and Andreas Schramm and Carina Holkenbrink for helpful comments on the manuscript. This work was supported by the Max Planck Society, the LOEWE Center for Synthetic Microbiology (SYNMIKRO), and the Human Frontier Science Program (RGY0076/2013).
References 1 Wolpert, L. (1969) Positional information and the spatial pattern of cellular differentiation. J. Theor. Biol. 25, 1–47 2 Dessaud, E. et al. (2010) Dynamic assignment and maintenance of positional identity in the ventral neural tube by the morphogen sonic hedgehog. PLoS Biol. 8, e1000382 3 Dessaud, E. et al. (2007) Interpretation of the sonic hedgehog morphogen gradient by a temporal adaptation mechanism. Nature 450, 717–720 4 Driever, W. and Nusslein-Volhard, C. (1988) A gradient of bicoid protein in Drosophila embryos. Cell 54, 83–93 5 Gregor, T. et al. (2007) Stability and nuclear dynamics of the bicoid morphogen gradient. Cell 130, 141–152 6 Struhl, G. et al. (1989) The gradient morphogen bicoid is a concentration-dependent transcriptional activator. Cell 57, 1259–1273 7 Yu, S.R. et al. (2009) Fgf8 morphogen gradient forms by a source-sink mechanism with freely diffusing molecules. Nature 461, 533–536 8 Wartlick, O. et al. (2009) Morphogen gradient formation. Cold Spring Harb. Perspect. Biol. 1, a001255 9 Kicheva, A. et al. (2007) Kinetics of morphogen gradient formation. Science 315, 521–525 10 Vuilleumier, R. et al. (2010) Control of Dpp morphogen signalling by a secreted feedback regulator. Nat. Cell Biol. 12, 611–617 72
Trends in Microbiology February 2014, Vol. 22, No. 2
11 Wolpert, L. (2011) Positional information and patterning revisited. J. Theor. Biol. 269, 359–365 12 Deich, J. et al. (2004) Visualization of the movement of single histidine kinase molecules in live Caulobacter cells. Proc. Natl. Acad. Sci. U.S.A. 101, 15921–15926 13 Elowitz, M.B. et al. (1999) Protein mobility in the cytoplasm of Escherichia coli. J. Bacteriol. 181, 197–203 14 Meacci, G. et al. (2006) Mobility of Min-proteins in Escherichia coli measured by fluorescence correlation spectroscopy. Phys. Biol. 3, 255–263 15 Tropini, C. et al. (2012) Physical constraints on the establishment of intracellular spatial gradients in bacteria. BMC Biophys. 5, 17 16 Brown, G.C. and Kholodenko, B.N. (1999) Spatial gradients of cellular phospho-proteins. FEBS Lett. 457, 452–454 17 Lipkow, K. and Odde, D.J. (2008) Model for protein concentration gradients in the cytoplasm. Cell. Mol. Bioeng. 1, 84–92 18 Niethammer, P. et al. (2004) Stathmin-tubulin interaction gradients in motile and mitotic cells. Science 303, 1862–1866 19 Martin, S.G. and Berthelot-Grosjean, M. (2009) Polar gradients of the DYRK-family kinase Pom1 couple cell length with the cell cycle. Nature 459, 852–856 20 Moseley, J.B. et al. (2009) A spatial gradient coordinates cell size and mitotic entry in fission yeast. Nature 459, 857–860 21 Griffin, E.E. et al. (2011) Regulation of the MEX-5 gradient by a spatially segregated kinase/phosphatase cycle. Cell 146, 955–968 22 Cabeen, M.T. and Jacobs-Wagner, C. (2010) The bacterial cytoskeleton. Annu. Rev. Genet. 44, 365–392 23 Kerfeld, C.A. et al. (2010) Bacterial microcompartments. Annu. Rev. Microbiol. 64, 391–408 24 Pilhofer, M. and Jensen, G.J. (2013) The bacterial cytoskeleton: more than twisted filaments. Curr. Opin. Cell Biol. 25, 125–133 25 Shapiro, L. et al. (2009) Why and how bacteria localize proteins. Science 326, 1225–1228 26 Flores, E. and Herrero, A. (2010) Compartmentalized function through cell differentiation in filamentous cyanobacteria. Nat. Rev. Microbiol. 8, 39–50 27 Callahan, S.M. and Buikema, W.J. (2001) The role of HetN in maintenance of the heterocyst pattern in Anabaena sp. PCC 7120. Mol. Microbiol. 40, 941–950 28 Corrales-Guerrero, L. et al. (2013) Functional dissection and evidence for intercellular transfer of the heterocyst-differentiation PatS morphogen. Mol. Microbiol. 88, 1093–1105 29 Yoon, H.S. and Golden, J.W. (1998) Heterocyst pattern formation controlled by a diffusible peptide. Science 282, 935–938 30 Robbins, J.R. et al. (2001) The making of a gradient: IcsA (VirG) polarity in Shigella flexneri. Mol. Microbiol. 41, 861–872 31 Chen, Y.E. et al. (2011) Spatial gradient of protein phosphorylation underlies replicative asymmetry in a bacterium. Proc. Natl. Acad. Sci. U.S.A. 108, 1052–1057 32 Thanbichler, M. and Shapiro, L. (2006) MipZ, a spatial regulator coordinating chromosome segregation with cell division in Caulobacter. Cell 126, 147–162 33 Hwang, L.C. et al. (2013) ParA-mediated plasmid partition driven by protein pattern self-organization. EMBO J. 32, 1238–1249 34 Vecchiarelli, A.G. et al. (2013) Cell-free study of F plasmid partition provides evidence for cargo transport by a diffusion-ratchet mechanism. Proc. Natl. Acad. Sci. U.S.A. 110, E1390–E1397 35 Ba¨hler, J. and Pringle, J.R. (1998) Pom1p, a fission yeast protein kinase that provides positional information for both polarized growth and cytokinesis. Genes Dev. 12, 1356–1370 36 Padte, N.N. et al. (2006) The cell-end factor Pom1p inhibits Mid1p in specification of the cell division plane in fission yeast. Curr. Biol. 16, 2480–2487 37 Celton-Morizur, S. et al. (2006) Pom1 kinase links division plane position to cell polarity by regulating Mid1p cortical distribution. J. Cell Sci. 119, 4710–4718 38 Hachet, O. et al. (2011) A phosphorylation cycle shapes gradients of the DYRK family kinase Pom1 at the plasma membrane. Cell 145, 1116–1128 39 Alvarez-Tabares, I. et al. (2007) Schizosaccharomyces pombe protein phosphatase 1 in mitosis, endocytosis and a partnership with Wsh3/Tea4 to control polarised growth. J. Cell Sci. 120, 3589–3601
Review 40 Howard, M. (2012) How to build a robust intracellular concentration gradient. Trends Cell Biol. 22, 311–317 41 Saunders, T.E. et al. (2012) Noise reduction in the intracellular Pom1p gradient by a dynamic clustering mechanism. Dev. Cell 22, 558–572 42 Gerdes, K. et al. (2010) Pushing and pulling in prokaryotic DNA segregation. Cell 141, 927–942 43 Mohl, D.A. and Gober, J.W. (1997) Cell cycle-dependent polar localization of chromosome partitioning proteins in Caulobacter crescentus. Cell 88, 675–684 44 Toro, E. et al. (2008) Caulobacter requires a dedicated mechanism to initiate chromosome segregation. Proc. Natl. Acad. Sci. U.S.A. 105, 15435–15440 45 Ptacin, J.L. et al. (2010) A spindle-like apparatus guides bacterial chromosome segregation. Nat. Cell Biol. 12, 791–798 46 Schofield, W.B. et al. (2010) Cell cycle coordination and regulation of bacterial chromosome segregation dynamics by polarly localized proteins. EMBO J. 29, 3068–3081 47 Erickson, H.P. et al. (2010) FtsZ in bacterial cytokinesis: cytoskeleton and force generator all in one. Microbiol. Mol. Biol. Rev. 74, 504–528 48 Kiekebusch, D. et al. (2012) Localized dimerization and nucleoid binding drive gradient formation by the bacterial cell division inhibitor MipZ. Mol. Cell 46, 245–259 49 Gordon, S. et al. (2004) Kinetics of plasmid segregation in Escherichia coli. Mol. Microbiol. 51, 461–469 50 Sengupta, M. et al. (2010) P1 plasmid segregation: accurate redistribution by dynamic plasmid pairing and separation. J. Bacteriol. 192, 1175–1183 51 Vecchiarelli, A.G. et al. (2010) ATP control of dynamic P1 ParA-DNA interactions: a key role for the nucleoid in plasmid partition. Mol. Microbiol. 78, 78–91 52 Havey, J.C. et al. (2012) ATP-regulated interactions between P1 ParA, ParB and non-specific DNA that are stabilized by the plasmid partition site, parS. Nucleic Acids Res. 40, 801–812 53 Vecchiarelli, A.G. et al. (2013) Dissection of the ATPase active site of P1 ParA reveals multiple active forms essential for plasmid partition. J. Biol. Chem. 288, 17823–17831 54 Hatano, T. and Niki, H. (2010) Partitioning of P1 plasmids by gradual distribution of the ATPase ParA. Mol. Microbiol. 78, 1182–1198 55 Bouet, J.Y. and Funnell, B.E. (1999) P1 ParA interacts with the P1 partition complex at parS and an ATP-ADP switch controls ParA activities. EMBO J. 18, 1415–1424 56 Davis, M.A. et al. (1992) Biochemical activities of the ParA partition protein of the P1 plasmid. Mol. Microbiol. 6, 1141–1147 57 Davis, M.A. et al. (1996) The P1 ParA protein and its ATPase activity play a direct role in the segregation of plasmid copies to daughter cells. Mol. Microbiol. 21, 1029–1036 58 Fung, E. et al. (2001) Probing the ATP-binding site of P1 ParA: partition and repression have different requirements for ATP binding and hydrolysis. EMBO J. 20, 4901–4911 59 Barilla, D. et al. (2005) Bacterial DNA segregation dynamics mediated by the polymerizing protein ParF. EMBO J. 24, 1453–1464 60 Hui, M.P. et al. (2010) ParA2, a Vibrio cholerae chromosome partitioning protein, forms left-handed helical filaments on DNA. Proc. Natl. Acad. Sci. U.S.A. 107, 4590–4595 61 Leonard, T.A. et al. (2005) Bacterial chromosome segregation: structure and DNA binding of the Soj dimer–a conserved biological switch. EMBO J. 24, 270–282 62 Pratto, F. et al. (2008) Streptococcus pyogenes pSM19035 requires dynamic assembly of ATP-bound ParA and ParB on parS DNA during plasmid segregation. Nucleic Acids Res. 36, 3676–3689 63 Ringgaard, S. et al. (2009) Movement and equipositioning of plasmids by ParA filament disassembly. Proc. Natl. Acad. Sci. U.S.A. 106, 19369–19374 64 Hatano, T. et al. (2007) Oscillating focus of SopA associated with filamentous structure guides partitioning of F plasmid. Mol. Microbiol. 64, 1198–1213 65 Wang, X. et al. (2013) Organization and segregation of bacterial chromosomes. Nat. Rev. Genet. 14, 191–203
Trends in Microbiology February 2014, Vol. 22, No. 2
66 Savage, D.F. et al. (2010) Spatially ordered dynamics of the bacterial carbon fixation machinery. Science 327, 1258–1261 67 Roberts, M.A. et al. (2012) ParA-like protein uses nonspecific chromosomal DNA binding to partition protein complexes. Proc. Natl. Acad. Sci. U.S.A. 109, 6698–6703 68 Tostevin, F. (2011) Precision of sensing cell length via concentration gradients. Biophys. J. 100, 294–303 69 Fuller, B.G. (2010) Self-organization of intracellular gradients during mitosis. Cell Div. 5, 5 70 Fuchino, K. et al. (2013) Dynamic gradients of an intermediate filament-like cytoskeleton are recruited by a polarity landmark during apical growth. Proc. Natl. Acad. Sci. U.S.A. 110, E1889– E1897 71 Goldberg, A.L. and St John, A.C. (1976) Intracellular protein degradation in mammalian and bacterial cells: Part 2. Annu. Rev. Biochem. 45, 747–803 72 Nath, K. and Koch, A.L. (1971) Protein degradation in Escherichia coli. II. Strain differences in the degradation of protein and nucleic acid resulting from starvation. J. Biol. Chem. 246, 6956–6967 73 Grimm, O. et al. (2010) Modelling the Bicoid gradient. Development 137, 2253–2264 74 Kumar, M. et al. (2010) Mobility of cytoplasmic, membrane, and DNAbinding proteins in Escherichia coli. Biophys. J. 98, 552–559 75 Mayover, T.L. et al. (1999) Kinetic characterization of CheY phosphorylation reactions: comparison of P-CheA and smallmolecule phosphodonors. Biochemistry 38, 2259–2271 76 Tsokos, C.G. and Laub, M.T. (2012) Polarity and cell fate asymmetry in Caulobacter crescentus. Curr. Opin. Microbiol. 15, 744–750 77 Quon, K.C. et al. (1998) Negative control of bacterial DNA replication by a cell cycle regulatory protein that binds at the chromosome origin. Proc. Natl. Acad. Sci. U.S.A. 95, 120–125 78 Kholodenko, B.N. et al. (2000) Diffusion control of protein phosphorylation in signal transduction pathways. Biochem. J. 350 (Pt 3), 901–907 79 Kala´b, P. et al. (2006) Analysis of a RanGTP-regulated gradient in mitotic somatic cells. Nature 440, 697–701 80 Kala´b, P. et al. (2002) Visualization of a Ran-GTP gradient in interphase and mitotic Xenopus egg extracts. Science 295, 2452–2456 81 Leipe, D.D. et al. (2002) Classification and evolution of P-loop GTPases and related ATPases. J. Mol. Biol. 317, 41–72 82 Gasper, R. et al. (2009) It takes two to tango: regulation of G proteins by dimerization. Nat. Rev. Mol. Cell Biol. 10, 423–429 83 Lutkenhaus, J. (2007) Assembly dynamics of the bacterial MinCDE system and spatial regulation of the Z ring. Annu. Rev. Biochem. 76, 539–562 84 de Boer, P.A. et al. (1989) A division inhibitor and a topological specificity factor coded for by the minicell locus determine proper placement of the division septum in E. coli. Cell 56, 641–649 85 Raskin, D.M. and de Boer, P.A. (1999) Rapid pole-to-pole oscillation of a protein required for directing division to the middle of Escherichia coli. Proc. Natl. Acad. Sci. U.S.A. 96, 4971–4976 86 Szeto, T.H. et al. (2002) Membrane localization of MinD is mediated by a C-terminal motif that is conserved across eubacteria, archaea, and chloroplasts. Proc. Natl. Acad. Sci. U.S.A. 99, 15693–15698 87 Wu, W. et al. (2011) Determination of the structure of the MinD-ATP complex reveals the orientation of MinD on the membrane and the relative location of the binding sites for MinE and MinC. Mol. Microbiol. 79, 1515–1528 88 Hu, Z. and Lutkenhaus, J. (2001) Topological regulation of cell division in E. coli. spatiotemporal oscillation of MinD requires stimulation of its ATPase by MinE and phospholipid. Mol. Cell 7, 1337–1343 89 Hu, Z. and Lutkenhaus, J. (2003) A conserved sequence at the Cterminus of MinD is required for binding to the membrane and targeting MinC to the septum. Mol. Microbiol. 47, 345–355 90 Vecchiarelli, A.G. et al. (2012) Surfing biological surfaces: exploiting the nucleoid for partition and transport in bacteria. Mol. Microbiol. 86, 513–523