In the past, bacterial cells were thought of as vessels without internal organization, with proteins, mRNA, chromosomes and other soluble compounds dispersed somewhat randomly inside of a membrane-enveloped cytoplasm. This view has gradually changed, thanks mainly to recent advancements in imaging technology and the development of better fluorescent proteins and dyes. It has become clear that a large amount of the cytoplasmic material is specifically localized in the cell and that this localization is crucial for the correct operation of different cellular processes. It has become apparent that membranes play an important role in establishing specific localization patterns of membrane proteins within the cell [1,2]. The earlier fluidic mosaic model [3] assumed that lipids are homogeneously distributed in the membrane. This assumption was recently called into question when it was discovered that many different proteins are localized to specific sites in the bacterial membrane and that the membrane contains domains that differ in their lipid composition. The localization of proteins to specific sites in the cell membrane may be governed by chemical factors, for example the phospholipid composition of the lipid domains [4,5] and rafts [6], by different physical characteristics of the membrane, e.g. its degree of negative or positive curvature [7,8] or its electrical potential [9], and by the different lipid composition of membranes belonging to differentiated cell compartments. Taken together, all these factors show that the formation of lipid domains in cell membranes allows cellular asymmetry to be developed; this asymmetry is crucial for the life cycle and for many different vital processes in bacteria.

The two main classes of bacteria, Gram-positive and Gram-negative, differ in the composition of their cell wall and membranes. The cell wall of Gram-positive bacteria consists of a thick layer of peptidoglycan and a single membrane. The organization of a Gram-negative bacterial cell wall is more complex. It is made up of two membranes, an outer membrane and an inner membrane; these two membranes sandwich a region known as the periplasmic space which has very important functions in the survival and operation of the cell. This region acts as a buffer between the very different conditions of the external environment and the interior of the bacterium. It contains specialized transport proteins and sensory proteins, which detect conditions in the bacterium’s surroundings and help to determine an appropriate response to them.

Bacterial membranes are primarily comprised of different phospholipids. Bacterial phospholipids consist of a polar, hydrophilic head and a hydrophobic tail. The hydrophilic head appears as an electron-dense region in transmission electron microscopy while the tail, which is buried within the membrane, forms an electron-transparent region. Thus a thin section of membrane, when viewed by an electron microscope, appears as two parallel, thickly stained lines separated by an almost transparent region. In the membrane of Gram-positive bacteria and in the inner membrane of Gram-negative bacteria, the phospholipids are arranged fairly evenly on either membrane leaflet, forming a symmetric lipid bilayer. The outer membrane of Gram-negative bacteria, in contrast, has an asymmetric arrangement of phospholipids: most of the phospholipids are located in the inner leaflet of the membrane while the outer leaflet contains some phospholipids, but also proteins and modified lipid molecules termed lipopolysaccharides (LPS). The transmembrane proteins found in the outer membrane predominantly have a β-barrel fold whereas the inner membrane harbors mainly α-helix-based transmembrane proteins. The outer membrane provides a physical barrier from the outside world, but it also must maintain selective permeability for the uptake of nutrients to support the cell’s physiological functions and to allow virulence factors to be released.

The manner in which proteins localize on cell membranes is not dependent only on the lipid composition of the membrane or membrane domain. It was recently discovered that bacteria have also evolved completely different mechanisms by which proteins are able to recognize specific sites on the cell membrane, based on the membrane’s different physical rather than chemical characteristics. Two cues have been discovered based on whether the membrane has a negative curvature (concave shape) or positive curvature (convex shape).

The first protein, which was reported to recognize a negative curvature in bacteria was DivIVA. This B. subtilis protein is involved in two different functions, one during vegetative growth and a second during sporulation. During cell division, DivIVA recognizes the newly forming cell poles, localizes to and persists at these polar sites, and recruits the other Min system proteins, MinJ, MinD and MinC, which block premature septum formation [31]. DivIVA recruits a different set of proteins to the cell poles during sporulation, when it is required for the proper segregation of sister chromosomes anchored to opposite cell poles by structures close to their ori-regions [53–55]. In this sporulation-specific chromosomal arrangement, RacA acts as a bridge between the DivIVA at the cell pole and the ori region of the chromosome [56,57]. The apparent switching of binding partners by DivIVA may serve to couple relief of septum formation inhibition to faithful chromosome segregation during sporulation. Interestingly, DivIVA requires no additional proteins, cell division or otherwise, for its polar localization. It was quite a mystery how this protein was able to recognize these specific sites not only in B. subtilis but also when expressed heterologously E. coli and Schizosaccharomyces pombe, both organisms which do not possess a homologue of it [58]. One suggestion was that DivIVA recognized the CL domain at the poles of both B. subtilis and E. coli, but this failed to explain its polar localization in S. pombe. The correct explanation came from the pioneering work of Lenarcic et al.[7] and was elaborated further by Ramamurthi and Losick [47]. They found that DivIVA preferably binds to strongly curved cell membranes in cell shape mutants through its N-terminus (Figure 1B). In addition, DivIVA is able to distinguish between different degrees of negative membrane curvature and displays a hierarchical preference for the most concave surface available. Thus, it preferentially localizes at septation sites and then to cell poles. This was shown by blocking cytokinesis, causing a redistribution of DivIVA where it is enriched at the poles. In lysozyme treated, spherical cells (protoplasted cells), DivIVA is largely uniformly distributed [47]. DivIVA forms “doggy bone”-like structures of about 22 nm in length, which assemble into large lattices [59]. These doggy-bone structures appear to be composed of 6–8 DivIVA subunits and it was proposed that this relatively large size stabilizes them by “bridging” opposing membranes [7].

Taken together, all these results suggest that a concave surface alone is sufficient for the recruitment of DivIVA to specific sites in the cell. DivIVA homologues from different microorganisms sometimes have different functions and they can recruit different proteins to specific sites. In the non-sporulating Gram-positive cocci Streptococcus pneumoniae, which lacks a Min system, a DivIVA homologue is still important for cell division and it interacts with other cell division proteins [60]. In the filamentous actinomycete Streptomyces coelicolor, DivIVA is crucial for tip growth and branching [61,62], and interacts with a protein involved in tip growth [63]. In Corynebacterium glutamicum and Mycobacterium tuberculosis, DivIVA is required for polar growth [64,65]. Interestingly, so far DivIVA is the only protein that has been described in bacterial species, which is able to recognize negative curvature. Thus, it appears that negative curvature is not a widely used cue for membrane protein localization. In the rod shaped S. pombe, the Pom1 protein localizes to the cell poles and the unique geometry of this environment suggests that Pom1 may be a eukaryotic example of a protein that recognizes membrane curvature on a cellular scale [66].

In sporulating bacteria such as B. subtilis, both convex (positive curvature) and concave membrane shapes can be found (Figure 1C). The formation of convex-shaped membranes is linked to asymmetric cell division, when the cell is divided into a smaller forespore and a larger mother cell. The forespore is subsequently engulfed by the mother cell and its surface is formed by a positively curved membrane, which is recognized by a small, 26 amino-acid long, sporulation specific protein called SpoVM [8]. This protein is likely the first of a large set of coat proteins, which are synthesized in the mother cell and deposited on the forespore surface. SpoVM forms an amphiphatic helix with its hydrophobic face buried in the lipid bilayer [67]. In a mutant in which no positive curvature is formed and SpoVM is still expressed, it localizes to the inner membrane of the mother cell. SpoVM has also been shown to bind to the outer surface of liposomes [8].

There is an additional physical characteristic of the bacterial cell membrane, which can influence the spatial organization of not only the lipid domains but also the proteins which recognize them. Recently, Strahl and Hamoen showed that the proton motive force (pmf) generated in the membrane can largely delocalize many proteins, which normally recognize the cues described above [9]. They tested more than 20 proteins which show specific localization patterns and which are involved in cell division and its regulation, cell shape regulation, chromosome segregation, signal transduction, secretion and others. A pmf dissipation was achieved by adding the specific proton-ionophore CCCP (carbonyl cyanide m-chlorophenylhydrazone). Among the proteins that rapidly changed their localization patterns after incubation with CCCP were those which bind to the membrane transiently through an amphiphatic α-helix, including MinD and FtsA or cytoskeletal proteins that are predicted to be attached to the membrane through other proteins, including MreB, MreBH and Mbl (Table 1). Nevertheless, recently it has also been discovered that both E. coli and T. maritima MreB interact directly with the cell membrane [68] either through an N-terminal amphiphatic helix or via a membrane insertion loop. However, incubation with CCCP had no effect on the localization of proteins recognizing negative curvature, like DivIVA, or those with many transmembrane helices, such as MinJ.

There are two components of pmf, a transmembrane electric potential and a transmebrane proton gradient. It was shown that the electric potential is responsible for this change in protein localization. The change in membrane potential could cause a subtle conformational change in the amphiphatic helix or it may affect the membrane lipid properties, or both. What biological role could the regulation of protein localization through changes in membrane potential have? Bacteria living in different environmental niches, which can change rapidly, could use changes in membrane potential to respond more quickly to environmental challenges. An important environmental factor which can change the membrane potential is the availability of electron acceptors such as oxygen. B. subtilis is a soil bacterium and in its natural environment the oxygen level changes constantly. At low oxygen levels, the cells become elongated, possibly as a consequence of the delocalization of the MinD and FtsA proteins [9].

Some membrane proteins have the ability to change their topology in response to changes in the lipid composition of the membrane, for example, lactose permease LacY from E. coli[69]. This protein shows a dual bi-directional reversible topology, which is dependent on the membrane lipid composition. Another documented case of a membrane protein with a dual topology is the L envelope protein of the hepatitis B virus [70]. Many other proteins also show similar characteristics [71,72].

The lipid composition of the membranes can fluctuate greatly during the cell cycle in all organisms from bacteria to humans. A clear example of this can be seen during sporulation in B. subtilis. Sporulation is activated by complex regulatory circuits and begins when a threshold concentration of phosphorylated Spo0A is reached. The whole process is closely linked to the cell cycle and it involves an asymmetric cell division. The beginning of sporulation is marked by remarkable changes not only in chromosome partitioning but also in cell division. The sporulation septum bisects the axial filament leaving only one third of one chromosome in the forespore and creating a transient genetic asymmetry [73]. The remaining two-thirds of the chromosome are then transferred from the larger mother cell into the smaller forespore via a conjugation-like mechanism directed by the partitioning protein SpoIIIE [74].

After sporulation begins, Spo0A brings about global changes in gene expression in the cell [75]. During sporulation, gene expression is orchestrated through the activity of four compartment-specific sigma factors (σ^F, σ^E, σ^G and σ^K), activated either in the forespore or in the mother cell [76]. These sigma factors, subunits of RNA polymerase, recognize specific sets of promoter sequences and allow transcription of different genes. Recently, it was discovered that Spo0A reactivates de novo fatty acid and phospholipids synthesis for membrane biogenesis during transient genetic asymmetry and after asymmetric cell division [14]. These authors suggested that de novo lipid synthesis was asymmetrically excluded from the forespore (Figure 1D). This system would allow a different lipid composition to develop on the outer, mother side, of the forespore membrane in contrast to its inner membrane. This was shown to be crucial for the function of SpoIIR, which is produced in the forespore, secreted into the septal membrane, and directionally activates σ^E in the mother cell [14]. It is also possible that a difference in lipid composition between the inner and outer membranes of the forespore is important for the function of other membrane proteins. One of these proteins might be SpoIIE, which has a dual function in asymmetric septum formation and in activation of the first compartment specific sigma factor, σ^F[77,78]. Other examples of proteins, which might be affected are SpoIIQ and SpoIIIAH, which are important for forespore engulfment.

Even more dramatic rearrangements in the membranes were observed during sporulation of Acetonoma longum, an evolutionarily close relative of Clostrideae sp. [79]. These authors showed that during sporulation the inner membrane of the mother cell is inverted and transformed to become an outer membrane of the germinating cell. Their results point to sporulation as a mechanism by which the bacterial outer membrane could have arisen. All these factors, taken together, show that dynamic membrane remodeling events during sporulation, which involve differentiated membrane lipid synthesis, may have crucial structural and regulatory roles for the proper localization and function of the proteins involved.

Many proteins in bacterial cells are compartmentalized and localized in a way, which suggests that these activities are not random, but depend upon certain membrane cues. One of these cues is the presence of certain regions of different lipid composition called membrane domains and they were identified with the help of proteins, which are known to be not dependent on other proteins for their correct localization. However, localization of some of the GFP tagged proteins was recently questioned [80,81] and it will require reevaluation.

Research into the formation of these lipid domains is a relatively new field, but nevertheless it has already given important insights into how proteins recognize specific topological cues and perform their tasks during the cell cycle. Lipid domains, lipid rafts and different physical membrane characteristics have been studied mainly in the two model organisms E. coli and B. subtilis and these studies have clearly demonstrated that these membrane characteristics are crucial for several important cellular processes, including cell division, signal transduction, sporulation, biofilm formation, chromosome segregation, secretion, and cytoskeleton formation.

In the near future, many interesting revelations about membrane biogenesis in other bacterial species should also come to light. There are still many unanswered questions about the heterogeneity, size, half-lives and dynamics of the membrane cues already characterized. The answers will require the development of new technologies, especially in imaging, as well as the use of multidisciplinary approaches, including elements of biology, chemistry, mathematics and physics.