The bacterial chromosome encodes information at multiple levels. Beyond direct protein coding, genomes encode regulatory information required to orchestrate the proper timing and levels of gene expression and protein synthesis, and contain binding sites and regulatory sequences to co-ordinate the activities of proteins involved in chromosome repair and maintenance. In addition, it is becoming increasingly clear that yet another level of information is encoded by the bacterial chromosome – the three-dimensional packaging of the chromosomal DNA molecule itself and its positioning relative to the cell. This vast structural blueprint of specific positional information is manifested in various ways, directing chromosome compaction, accessibility, attachment to the cell envelope, supercoiling, and general architecture and arrangement of the chromosome relative to the cell body. Recent studies have begun to identify and characterize novel systems that utilize the three-dimensional spatial information encoded by chromosomal architecture to co-ordinate and direct fundamental cellular processes within the cytoplasm, providing large-scale order within the complex clutter of the cytoplasmic compartment.
Bacteria are considered to be the most basal of life forms. Despite the implied simplicity of the bacterial cell, unexpected levels of organization of the molecular components within these tiny organisms exist. For instance, a recent study discovered that at least 10% of all proteins in the bacterium Caulobacter crescentus (Caulobacter) display specific and reproducible subcellular localizations (Werner et al., 2009). In addition to protein localization, the chromosomal molecule and its cadre of bound proteins, or nucleoid, is highly compacted to fit into the relatively small cytoplasmic space, but surprisingly displays an elegant and reproducible large-scale organization. In Caulobacter, loci along the chromosome occupy specific and reproducible spatial co-ordinates within the cytoplasm that are arrayed linearly within the cell volume with respect to their genomic position (Fig. 1) (Viollier et al., 2004; Umbarger et al., 2011). This organization is maintained over successive cycles of replication, segregation and cell division, and may therefore be considered a dynamic and heritable epigenetic feature of the species. Other bacteria display drastically distinct chromosome organizational schemes. The Escherichia coli chromosome, for example, adopts a transversally folded structure that positions the right and left arms of the chromosome towards the cell poles, and origin and terminus of replication near the cell centre (Nielsen et al., 2006; Wang et al., 2006) (Fig. 1).
Recent studies have revealed unprecedented insight into the structural organization of bacterial chromosomes and have characterized fundamental cellular processes that require the dynamic spatial organizational information coded in the chromosome for their function. Together, these results are beginning to reveal that three-dimensional chromosome structure is a key architectural feature that is central to bacterial cellular organization and functional physiology. Here we will discuss several recently described bacterial systems that utilize the architectural arrangement of the chromosome within the cytoplasm to facilitate the spatial regulation of key developmental events.
Chromosome architecture can facilitate the regulation of chromosome replication initiation
Although the reproducible subcellular positioning of specific chromosomal loci has been described in various bacteria, the effect of this large-scale chromosomal organization on the bacterial cell cycle for the most part is unclear. One notable exception is the dimorphic alphaproteobacterium Caulobacter crescentus (Caulobacter), which has integrated large-scale chromosomal organization into the basic workings of its cell cycle regulatory system. In the pre-replicative swarmer cell, the chromosomal origin of replication (Cori) is positioned near the old cell pole (Mohl and Gober, 1997; Jensen et al., 2001), where it is physically tethered via interactions between a site-specific DNA binding protein ParB, bound to Cori-proximal DNA sequences (parS), and the polar scaffolding protein PopZ (Bowman et al., 2008; Ebersbach et al., 2008) (Fig. 2). The master cell cycle regulator CtrA is bound to specific sequences in Cori, ensuring that the initiator protein DnaA cannot initiate replication (Quon et al., 1998; Marczynski and Shapiro, 2002) (Fig. 2). To initiate the cell cycle, a phospho-relay signal causes the ClpXP protein degradation machinery to accumulate at the old cell pole where it rapidly and specifically degrades the CtrA master regulator, licensing initiation (McGrath et al., 2006). Meanwhile, an unknown signal stimulates the separation of the ParB/parS complex from the polar PopZ network, releasing the licensed origin from its tether and allowing DNA replication to initiate (Fig. 2) (Bowman et al., 2010). Thus, Caulobacter utilizes the subcellular positioning of the chromosome origin region to efficiently and robustly regulate cell cycle progression.
Nucleoid architecture is a structural scaffold for molecular partitioning machines, including the origin segregation apparatus
To ensure organizational inheritance over generations, many bacteria have evolved dedicated mechanisms to initiate chromosome segregation at specific loci, and to actively transport these segments to specific and reproducible subcellular addresses. Many bacteria utilize DNA partitioning systems (Par systems) related to those found on plasmids to actively segregate a chromosomal centromere-like sequence, parS, which in Caulobacter has been shown to be both the site of force generation during initial chromosome segregation and the first chromosomal locus to be partitioned (Toro et al., 2008). In addition to the parS centromere, bacterial partitioning systems consist of two additional core components: a site-specific DNA binding protein ParB, which binds to parS, and an ATPase ParA (Fig. 2). Seminal studies in Vibrio cholerae and in the E. coli plasmid pB171 showed that actively segregating parS loci follow a retracting cloud-like structure of the ATPase ParA, suggesting a conserved mitotic ‘pulling’ mechanism for segregation (Fogel and Waldor, 2006; Ringgaard et al., 2009). Further investigations revealed that ParB/parS complexes interact with ParA structures, stimulating ATP hydrolysis in ParA and release of ADP-containing subunits. In this way, multimeric ParB complexes follow along a shortening ParA structure via a burnt bridge Brownian ratchet-like mechanism (Fig. 2) (Leonard et al., 2005; Fogel and Waldor, 2006; Ringgaard et al., 2009; Ptacin et al., 2010; Schofield et al., 2010; Vecchiarelli et al., 2010). Importantly, several studies demonstrate an essential role for cooperative DNA binding by ParA in formation and function of the ParA mitotic structure (Leonard et al., 2005; Ringgaard et al., 2009; Hui et al., 2010; Ptacin et al., 2010; Vecchiarelli et al., 2010), suggesting that the nucleoid itself forms a structural matrix for the assembly of a track-like structure of ParA that guides ParB movement. In organisms such as Caulobacter, which partition centromeres unidirectionally from one pole to the other, in addition to polarity determinants such as TipN (Ptacin et al., 2010; Schofield et al., 2010), the initial asymmetric positioning of the ParB/parS complex relative to the ParA/nucleoid structure may facilitate the establishment of unidirectionality in segregation (Fig. 2).
In addition to chromosomal centromere and plasmid partitioning, recent studies have identified other ParA-based molecular machines that partition diverse cargoes throughout the cytoplasm, such as carbon-fixing carboxysome organelles in Synechococcus elongatus (Savage et al., 2010) and cytoplasmic chemosensory complexes in Rhodobacter sphaeroides (Roberts et al., 2012). The latter study demonstrated a requirement for chromosomal DNA binding by the ParA component, similar to ParA-mediated DNA partitioning machines, suggesting a common mechanism that utilizes the nucleoid DNA as a structural matrix for ParA-mediated molecular partitioning systems. How the global packaging of the chromosome affects ParA-mediated cargo segregation remains an open and intriguing question.
Dynamic chromosome architecture couples chromosome replication and segregation to the spatial and temporal initiation of cell division
Thus far, we have discussed mechanisms by which chromosome architecture facilitates the initiation of chromosome replication as well as the subsequent segregation of chromosomal origin regions. To ensure proper cell cycle progression, most bacteria also contain systems that couple chromosome replication and segregation to the initiation of cell division. Several spatial sensory systems have been described that sense changing chromosome architecture during segregation and couple these signals to the regulation of the cell division machinery (Fig. 3). During chromosome segregation in Caulobacter, the centromere bound ParB complex also recruits the cell division inhibitor MipZ, a ParA/MinD-like ATPase that specifically inhibits FtsZ polymerization (Fig. 3A) (Thanbichler and Shapiro, 2006). ParB binding to MipZ activates MipZ DNA binding activity, creating a nucleoid bound gradient of this FtsZ polymerization inhibitor near the ParB/parS complex (Fig. 3A) (Kiekebusch et al., 2012). After ParB/parS segregation to opposite poles, the bipolar positioning of MipZ gradients along the nucleoid prevents FtsZ from initiating cell division near the cell poles and neighbouring nucleoid and drives FtsZ assembly near midcell (Fig. 3A). Therefore, in Caulobacter, the spatio-temporal signal of dynamically changing chromosome organization co-ordinates chromosome segregation with the proper positioning of the cell division machinery.
Escherichia coli and Bacillus subtilis have evolved distinct systems to co-ordinate changing chromosome architecture during segregation with the positioning of the cell division apparatus, but which function using similar principles to the Caulobacter MipZ system. One such system is the three-component Min system that prevents FtsZ from forming pole-proximal division septa (reviewed in Bramkamp and van Baarle, 2009). Min systems work in concert with another mechanism, termed nucleoid occlusion (NO), characterized thus far in B. subtilis, E. coli and Staphylococcus aureus (Wu and Errington, 2004; Bernhardt and de Boer, 2005; Veiga et al., 2011). NO systems utilize spatial information about nucleoid organization to position an inhibitor of divisome formation over the chromosome, driving the assembly of the division apparatus towards nucleoid-free regions near midcell (Fig. 3B). Decoration of the nucleoid by NO proteins, however, is not random. The B. subtilis and S. aureus NO proteins (termed Noc) are site-specific DNA binding proteins related to ParB (Wu et al., 2009; Tonthat et al., 2011), whereas the E. coli NO protein (called SlmA) is predicted to adopt a fold related to TetR (Bernhardt and de Boer, 2005). These factors bind specific DNA sequences and spread to neighbouring DNA to generate large nucleoprotein clusters (Wu et al., 2009; Cho et al., 2011; Tonthat et al., 2011). NO protein binding sites are distributed across the origin-proximal 2/3 of the chromosome, but are conspicuously absent from the terminus region that is positioned at midcell (Wu et al., 2009; Cho et al., 2011; Tonthat et al., 2011). DNA replication and segregation therefore create a post-replicative chromosomal architecture that organizes NO nucleoprotein complexes away from the cell centre, and in concert with the polar Min system shepherd divisome formation to midcell between the newly segregated nucleoids as DNA replication is completed (Fig. 3B).
Structural changes in nucleoid architecture and positioning are required for proper sporulation in B. subtilis
In addition to regulating proper cell division during vegetative growth in most bacteria, the architectural remodelling of nucleoid structure is required for the complex chromosomal gymnastics during sporulation in Bacillus species. The common theme is that the remodelling process involves the co-ordinated activities of site-specific DNA binding proteins that are recruited to repeated non-coding DNA sequences specifically arrayed within the nucleoid structure, marking distinct 3D positions in the cytoplasmic space.
During sporulation in B. subtilis, the newly replicated chromosomes undergo a large-scale rearrangement into an elongated and compacted structure called the axial filament (Ryter et al., 1966), in which the origin regions of each chromosome form the ends of the rod-like structure and are tethered to opposite cell poles (Fig. 3C) (Ben-Yehuda et al., 2003a). Recent studies have identified repeated origin-proximal non-coding DNA sequences that are involved in the compaction of the origin region of the axial filament and tethering of this complex to the cell pole. One important factor is the ParB homologue Spo0J, which binds to repeated origin-proximal parS sequences and forms a large and compact oligomeric complex (Lin and Grossman, 1998; Leonard et al., 2004; Rodionov and Yarmolinsky, 2004; Murray et al., 2006; Breier and Grossman, 2007). Spo0J specifically recruits the bacterial condensin complex SMC to these regions (Gruber and Errington, 2009; Sullivan et al., 2009), likely facilitating the compaction of the ends of the axial filament near the cell poles. Another site-specific DNA binding protein, RacA, binds to a different subset of repeated DNA sequences arrayed along this region of the chromosome and tethers this compacted region to the pole via interactions with the polar DivIVA complex (Fig. 3C) (Ben-Yehuda et al., 2003a; Wu and Errington, 2003; Ben-Yehuda et al., 2005; Lenarcic et al., 2009). Finally, the recently identified RefZ protein binds to yet another repeated non-coding DNA sequence near the origin and terminus regions of the chromosome (Wagner-Herman et al., 2012). Although the molecular mechanism has not been established, RefZ complexes have been postulated to direct FtsZ ring formation over a neighbouring chromosomal segment to facilitate proper chromosome trapping during prespore development (Fig. 3C) (Wagner-Herman et al., 2012). Thus, these systems collaborate to organize the architecture of the sporulating chromosome and couple this organization to proper asymmetric division site placement. This specific nucleoid architecture during division traps important origin-proximal developmental genes within the early prespore and allows the establishment of compartment-specific gene expression programmes required for spore development (Losick and Stragier, 1992).
Remarkably, the large-scale architectural rearrangements of the chromosome that facilitate entry into sporulation are transient and rapidly undone during the packaging of the chromosome into the developing prespore compartment. The organization of the chromosomal axial filament structure during asymmetric division, in addition to facilitating compartmentalized gene expression, also presents a specific region of the chromosome to be trapped by the closing septum (Ben-Yehuda et al., 2003b). As the septum encircles this trapped DNA region, the septal DNA-translocating ATPase SpoIIIE recognizes specific non-coding DNA sequence elements to export the chromosome into the prespore compartment (Becker and Pogliano, 2007; Ptacin et al., 2008). SpoIIIE acts like a molecular wire stripper, transferring naked chromosomal DNA into the developing spore while excluding DNA bound proteins in the mother cell (Marquis et al., 2008). Finally, during spore germination, the DNA is transiently packaged de novo into a ring-shaped structure that may protect the chromosomal DNA from damage before transitioning back into the vegetative organization to begin growth anew (Ragkousi et al., 2000). These intricate rearrangements and de novo organizational events illustrate that the chromosomal molecule contains all of the necessary information to orchestrate these highly dynamic and reversible architectural changes. The multitude of systems that have evolved to utilize non-coding information in the chromosome architecture to generate overall morphological and developmental changes in the cell underlines the importance and utility of a reproducible nucleoid organization.
Chromosome architecture may organize the cytoplasm by spatially regulating gene expression and protein synthesis
By definition, a reproducible chromosome architecture places specific chromosomal loci into discrete regions within the cytoplasm, and genes themselves are no exception. Until recently, it was believed that after transcription, mRNAs were released from transcription complexes to diffuse throughout the cytoplasm.
However, recent work from the Jacobs-Wagner lab has discovered that mRNA molecules in both Caulobacter and E. coli remain constrained close to their template locations on the nucleoid (Montero Llopis et al., 2010). These authors further reported that localized mRNAs prevent the diffusion of active ribosomes, thereby spatially constraining the translation of the coded polypeptide (Montero Llopis et al., 2010). These results suggest that the spatial clustering of related genes may facilitate the formation of the protein complexes for which they code. On a larger scale, it remains unclear but tempting to speculate whether the chromosomal location of a gene may bias the final destination of its coded product. However, a concurrent study identified a contrasting phenomenon in E. coli that appears to target certain mRNAs to the final location of the polypeptides for which they code in a translation-independent manner (Nevo-Dinur et al., 2011), indicating that various mechanisms may impact the spatial ordering of newly synthesized protein products.
Interplay between chromosome immobilization and organization
The reproducible organization of the bacterial chromosome is faithfully maintained across generations of dynamically orchestrated cycles of replication and segregation. Surprisingly, however, the mechanisms that guide and shape this process remain poorly understood. The myriad of nucleoid-associated proteins (NAPs) involved in nucleoid landscaping and their proposed mechanisms of action are outside the scope of this review, and have been recently reviewed elsewhere (Browning et al., 2010; Rimsky and Travers, 2011). However, one compelling new paradigm for large-scale chromosome organization appears to be the specific immobilization of certain chromosome segments to cellular landmarks, and subsequent compaction and organization of the rest of the chromosome around these points. Studies on the organization of the Caulobacter chromosome identified an interaction between the origin-proximal centromere parS region and the new cell pole via interactions of ParB with the polar polymeric network of PopZ (described above; Bowman et al., 2008; Ebersbach et al., 2008). Strikingly, when the parS region of the chromosome is relocated relative to the rest of the chromosome, the chromosome is concomitantly reorganized around the new parS location (Umbarger et al., 2011), implying that the organization of chromosomal loci is ‘clocked’ relative to this tethering point.
Escherichia coli appears to have taken a distinct strategy to coupling chromosome locus tethering and large-scale organization. As mentioned above, the E. coli chromosome is packaged into a dramatically different arrangement than that of Caulobacter, placing the origin region at midcell, the terminus region generally within this area, and the left and right chromosome arms splayed out towards opposite cell poles (Fig. 1) (Nielsen et al., 2006; Wang et al., 2006). Work by the Boccard and Espeli labs has characterized a novel system that organizes the terminus region of the E. coli chromosome into a micron-scale organizational domain, or ‘macrodomain’. This system utilizes a repeated palindromic binding site (matS) found only in the terminus (ter) macrodomain of the chromosome to recruit a site-specific DNA binding protein, MatP, required for the structural organization of the region (Mercier et al., 2008). In addition to organizing the ter macrodomain, MatP interacts with ZapB, a component of the cell division apparatus, facilitating the maintenance of the ter region at the cell centre during cell division (Espeli et al., 2012; Thiel et al., 2012). Therefore, the E. coli MatP/matS system and Caulobacter ParB/PopZ system may promote chromosome organization relative to a reproducible cellular landmark such as the cell pole or developing cell division apparatus. Intriguingly, the deletion of the bacterial condensin subunit MukB in E. coli produces cells that place replication origins adjacent to the cell pole (Danilova et al., 2007), suggesting a possible link between origin-proximal condensin complexes and division site placement as well.
As we have seen, bacteria maintain specific, reproducible and genetically encoded links between chromosome organization and the processes of DNA replication, DNA segregation, cell division and sporulation. While the importance of nucleoid structure as an architectural feature in bacteria has become increasingly clear, the actual underlying structure of bacterial nucleoids is a puzzle that is only beginning to be pieced together. Novel technical approaches are uncovering surprising results. Recent studies using fluorescence deconvolution and high-throughput interaction mapping and localization have described the Bacillus and Caulobacter chromosomes as ellipsoidal with two compacted arms gently twisted upon one other (Fig. 1) (Berlatzky et al., 2008; Umbarger et al., 2011). Electron microscopy studies in other organisms have reported similar structures (Butan et al., 2011), as well as bizarre and astonishing chromatin configurations such as whirled nucleoid arrangements (Robinow and Kellenberger, 1994) and nucleoids contained within membrane bound compartments (Fuerst, 2005). Similar studies in other organisms will undoubtedly yield both similarities and surprising differences in chromosome structures across bacteria, and facilitate the discovery of novel systems that utilize nucleoid architecture to organize and orchestrate cellular development.
We thank Grant Bowman, Esteban Toro and Seth Childers for critical reading of the manuscript and thoughtful suggestions. This work is supported in part by National Institutes of Health Grants R01 GM51426 R24 and GM073011-04d to L.S., and NIH/NIGMS Fellowship F32GM088966-3 to J.P. The authors declare no conflict of interest.