OUP user menu

Extracellular biology of Myxococcus xanthus

Anna Konovalova , Tobias Petters , Lotte Søgaard-Andersen
DOI: http://dx.doi.org/10.1111/j.1574-6976.2009.00194.x 89-106 First published online: 1 March 2010


Myxococcus xanthus has a lifecycle characterized by several social interactions. In the presence of prey, M. xanthus is a predator forming cooperatively feeding colonies, and in the absence of nutrients, M. xanthus cells interact to form multicellular, spore-filled fruiting bodies. Formation of both cellular patterns depends on extracellular functions including the extracellular matrix and intercellular signals. Interestingly, the formation of these patterns also depends on several activities that involve direct cell–cell contacts between M. xanthus cells or direct contacts between M. xanthus cells and the substratum, suggesting that M. xanthus cells have a marked ability to distinguish self from nonself. Genome-wide analyses of the M. xanthus genome reveal a large potential for protein secretion. Myxococcus xanthus harbours all protein secretion systems required for translocation of unfolded and folded proteins across the cytoplasmic membrane and an intact type II secretion system. Moreover, M. xanthus contains 60 ATP-binding cassette transporters, two degenerate type III secretion systems, both of which lack the parts in the outer membrane and the needle structure, and an intact type VI secretion system for one-step translocation of proteins across the cell envelope. Also, analyses of the M. xanthus proteome reveal a large protein secretion potential including many proteins of unknown function.

  • Myxococcus xanthus
  • extracellular matrix
  • protein secretion
  • bacterial motility
  • secondary metabolism
  • contact-dependent activity


Myxobacteria have a lifestyle characterized by a rich social life. These social interactions play decisive roles in growth, motility and development and allow myxobacterial cells to organize into distinct spatial patterns as well as to feed cooperatively. Myxococcus xanthus has emerged as the pre-eminent model system to understand the genetics and molecular mechanisms involved in the social behaviour of myxobacteria. In the first part of this review, the focus is on experimental evidence demonstrating that the social behaviours of M. xanthus depend on extracellular functions. In the second part, genomic analyses illustrating the large genetic potential of M. xanthus to sense and respond to changes in the environment as well as the large potential for protein secretion are presented. Myxococcus xanthus has also emerged as an excellent model system to analyse the evolution of social behaviours at the molecular level (Velicer et al., 2000; Velicer & Yu, 2003; Fiegna et al., 2006). The topics presented in this review are discussed with special reference to extracellular functions.

The ecology and lifecycle of myxobacteria

Myxococcus xanthus– and myxobacteria in general – belongs to the delta subgroup of the proteobacteria. Myxobacteria are typically found in topsoil, with only a few marine and freshwater myxobacteria having been isolated (Reichenbach, 1999; Velicer & Hillesland, 2008). Soil myxobacteria grow as saphrophytes on dead organic matter by decomposing degradable polymers or by preying on other microorganisms including bacteria and fungi in what has been described a wolf-pack-like manner (Rosenberg & Varon, 1984; Reichenbach et al., 1999). Predation by M. xanthus is conveniently observed in laboratory experiments by placing M. xanthus cells adjacent to other microorganisms (Shi & Zusman, 1993a; Berleman et al., 2006; Berleman & Kirby, 2007). However, the importance of myxobacteria in predation in their natural habitats was illustrated only recently in a study in which living 13C-labelled Escherichia coli cells were added to soil samples and used as a proxy for living biomass (Lueders et al., 2006). rRNA stable isotope probing revealed that the 13C-labelled E. coli carbon pool was sequestered mainly by gliding bacteria including myxobacteria, thus placing myxobacteria near or at the top of the microbial food chain.

The social lifestyle of M. xanthus crucially depends on the ability of cells to display active movement. Myxococcus xanthus cells move by gliding motility, which is the movement of a rod-shaped cell in the direction of the cell's long axis on a surface and in the absence of a flagellum (Henrichsen, 1972). If present on a solid surface and at a high cell density, M. xanthus cells self-organize into three morphologically distinct spatial patterns, spreading colonies, ripples or fruiting bodies (Dworkin, 1996) (Fig. 1). The pattern formed largely depends on the nutritional status of the cells. In the presence of nutrients, the motile, rod-shaped cells grow and divide and form spreading colonies. Cells at the edge of a colony spread coordinately over the surface, forming a thin, film-like structure. In the absence of nutrients, the spreading behaviour is constrained and cells initiate a developmental programme that culminates in the formation of multicellular, spore-filled fruiting bodies. Fruiting body formation proceeds in distinct morphological stages that are separated in time and space. The first signs of fruiting body formation are evident after 4–6 h of starvation as cells aggregate to form small aggregation centres. As they accumulate more cells, the centres increase in size and eventually become mound-shaped. By 24 h, the aggregation process is complete and each nascent fruiting body contains approximately 105 densely packed cells. Inside the nascent fruiting bodies, the rod-shaped cells undergo morphological and physiological differentiation into spherical myxospores, resulting in mature fruiting bodies. Spore maturation is finished approximately 72 h after the onset of starvation. Only 10% of the total population undergoes sporulation and these cells are those that have accumulated inside the fruiting bodies. Up to 30% of the cells remain outside the fruiting bodies. These cells remain rod-shaped and differentiate to a cell type called peripheral rods (O'Connor & Zusman, 1991a, b). Finally, the remaining cells undergo lysis (Rosenbluh et al., 1989). Recently, developmental cell lysis was suggested to reflect programmed cell death and was shown to depend on an unconventional toxin–antitoxin system involving the MazF mRNA interferase (Nariya & Inouye, 2008).


The three cellular patterns formed by Myxococcus xanthus cells. Scale bars=1 mm.

Aggregation and sporulation are the two invariable morphological processes in fruiting body formation. Under less stringent starvation conditions (Shimkets & Kaiser, 1982) or in the presence of prey (Berleman et al., 2006; Berleman & Kirby, 2007), M. xanthus cells organize into a third pattern referred to as rippling. During rippling, cells accumulate in equispaced ridge-like structures separated by troughs of low cell density. The ridge-like structures move coordinately and synchronously as travelling waves over the surface (Reichenbach, 1965; Shimkets & Kaiser, 1982) (for a time lapse movie of rippling cells, see Welch & Kaiser, 2001). Microscopic examination of rippling cells has shown that individual cells essentially oscillate back and forth, suggesting that colliding waves reflect off of each other (Sager & Kaiser, 1994; Welch & Kaiser, 2001; Sliusarenko et al., 2006). Rippling is typically initiated before aggregation. Later, during the aggregation process, the wave structure disintegrates and cells aggregate into the nascent fruiting bodies.

The M. xanthus cell envelope

Myxococcus xanthus and other myxobacteria are rod-shaped gram-negative bacteria. Originally, the peptidoglycan sacculus of M. xanthus was proposed to consist of discontinuous peptidoglycan separated by trypsin and sodium dodecyl sulphate (SDS)-sensitive material (White et al., 1968). However, recent analyses of the M. xanthus peptidoglycan sacculus have shown that it forms a continuous, rod-shaped structure and that the peptidoglycan only deviates from that in other gram-negative bacteria by containing a significant fraction of ll-diaminopimelic acid in place of the more common meso-ll-diaminopimelic acid in the pentapeptide and by containing a so far unidentified modification of N-acetylmuramic acid (Bui et al., 2009).

The M. xanthus lipopolysaccharide O-antigen is similar in overall structure to that of lipopolysaccharide in other gram-negative bacteria and the carbohydrate moiety consists of glucose, mannose, rhamnose, arabinose, xylose, galactosamine, glucosamine, 2-keto-3-deoxyoctulosonic acid, 3-O-methylpentose and 6-O-methylgalactosamine (Yang et al., 2008). Mutants deficient in lipopolysaccharide biosynthesis show defects in fruiting body formation (Fink & Zissler, 1989; Bowden & Kaplan, 1998). Lipopolysaccharide mutants were also proposed to have a defect in A-motility (Fink & Zissler, 1989); however, detailed genetic analyses have provided evidence that lipopolysaccharide mutants display defects in type IV pili (T4P)-dependent motility (Bowden & Kaplan, 1998).

The M. xanthus extracellular matrix (ECM)

Myxococcus xanthus cells are covered by an ECM composed of exopolysaccharides and proteins in a ratio of approximately 1: 1 (Behmlander & Dworkin, 1994). ECM has important functions in motility, cell–cell cohesion and fruiting body formation (Shimkets et al., 1986; Lux et al., 2004). Specifically, the polysaccharide portion of ECM triggers retraction of T4P (Li et al., 2003) and the ECM protein FibA has been proposed to be involved in the regulation of motility (Kearns et al., 2002). ECM was reported to have a fibrillar structure based on scanning electron microscopy studies (Behmlander & Dworkin, 1991). However, it has been suggested that the fibrillar structure may have been an experimental artefact caused by dehydration of the ECM during sample preparation (Yang et al., 2008). The polysaccharide fraction of ECM is composed of the five monosaccharides: galactose, glucosamine, glucose, rhamnose and xylose (Behmlander & Dworkin, 1994). The protein portion of ECM was analysed in a proteomics-based approach in which ECM was enriched from developing cells using a procedure involving 0.5% or 0.1% SDS extraction of cells (Curtis et al., 2007). Among the 41 proteins identified, 20 are likely integral inner or outer membrane proteins or cytoplasmic proteins and the remaining 21 have been suggested to be good ECM protein candidates. Among the 21 candidate ECM proteins, 13 were predicted to contain lipoprotein signal peptides, thus raising the question as to whether the method used is specific for ECM proteins or whether it also includes proteins from the inner and the outer membrane. Regardless, only five of the 21 candidate ECM proteins have predicted functions. These functions include protease activity, amidohydrolase activity and coating of the myxospores. Inactivation of several of the genes encoding putative ECM proteins caused no defects in fruiting body formation, with the exception of the fibA gene, which encodes the FibA zinc metalloprotease (Kearns et al., 2002).

Accumulation of ECM is a tightly regulated process and involves several different regulators (Fig. 2). The best-studied system involved in the regulation of ECM accumulation is the Dif chemosensory system (Yang et al., 1998b) (formerly Dsp; Lancero et al., 2002), which is one of eight chemosensory systems in M. xanthus (Kirby et al., 2008). The core proteins of Dif are DifA, DifC and DifE, which are orthologues of methyl-accepting chemosensory proteins (MCP), CheW adaptor proteins and CheA histidine protein kinases, respectively (Yang et al., 1998b). These three proteins are encoded in an operon that encodes three additional proteins (DifB, DifD and DifG). DifD is an orthologue of the CheY response regulator, DifG of the CheC phosphatase, and DifB is a hypothetical protein of unknown function (Black & Yang, 2004). A methyltransferase and a methylesterase are absent in the Dif system. DifA, DifC and DifE are required for ECM accumulation, including exopolysaccharides and FibA, which is generally used as a proxy for the protein content of the ECM (Yang et al., 2000; Bellenger et al., 2002), whereas DifD and DifG negatively regulate exopolysaccharide accumulation (Black & Yang, 2004); the effect of DifD and DifG on FibA accumulation is unknown (Black & Yang, 2004). A difB mutation has no obvious defects in exopolysaccharide or FibA accumulation (Black & Yang, 2004). On the basis of these findings, it has been proposed that the DifE kinase stimulates ECM accumulation by phosphorylation of a yet to be identified response regulator (DifX) (Black & Yang, 2004; Yang et al., 2008). Moreover, DifD would act as a phosphate sink, draining the phosphoryl group on DifE away from DifX and in this way inhibit exopolysaccharide accumulation (Black & Yang, 2004; Yang et al., 2008). According to this scheme, DifG would be involved in regulating exopolysaccharide accumulation by dephosphorylating DifD∼P (Black & Yang, 2004; Yang et al., 2008). Based on the observations that mutants lacking T4P accumulate reduced amounts of ECM, that a pilT mutant, which assembles nonfunctional T4P, accumulates increased amounts of ECM, and that dif mutations are epistatic to mutations affecting T4P assembly and function, it has been proposed that T4P may serve as a sensor for the Dif system (Black et al., 2006).


Model of the regulation of ECM accumulation in Myxococcus xanthus. A type IV pilus is shown spanning the cell envelope and potentially interacting with DifACE proteins (indicated in red). From the DifE histidine protein kinase, the phosphoryl group is hypothesized to be transferred to an unknown response regulator (DifX) to stimulate ECM accumulation or to DifD to inhibit ECM accumulation. The interaction between DifE and the response regulator Nla19 has only been demonstrated in a yeast two-hybrid system analysis. Response regulators and DnaK orthologues involved in the regulation of ECM accumulation are indicated in orange and green, respectively. Receiver domains are indicated by pentagons. Red arrows or inhibitions indicate positive or negative effects on exopolysaccharide accumulation. Blue arrows, * and ** indicate a positive effect on FibA accumulation, no effect on FibA accumulation or FibA accumulation not tested, respectively. FibA is the best-characterized protein in the ECM and is generally used as a proxy for the protein fraction of the ECM.

Other regulators of ECM accumulation (Fig. 2) include the NtrC-like transcriptional activator Nla24 (Caberoy et al., 2003) (also referred to as EpsI; Lu et al., 2005). An nla24 mutant accumulates reduced amounts of exopolysaccharides and normal levels of FibA (Lancero et al., 2004). The lack of the NtrC-like transcriptional activator Nla19 causes an increased accumulation of exopolysaccharides (Lancero et al., 2005). Interestingly, Nla19 interacts with DifE in the yeast-two-hybrid system, suggesting a possible link between Nla19 and Dif (Lancero et al., 2005). DigR is a third DNA-binding response regulator, which is also involved in the regulation of ECM accumulation (Overgaard et al., 2006). The lack of DigR causes increased accumulation of exopolysaccharides, while FibA accumulation is reduced. DigR is an orphan response regulator and its cognate histidine protein kinase is not known (Overgaard et al., 2006). For Nla24 and Nla19, it is not known how they affect ECM accumulation. In the case of DigR, global transcriptional profiling experiments using DNA microarrays showed that DigR is not likely to directly regulate the expression of genes coding for proteins involved in exopolysaccharide biosynthesis. Rather, DigR directly or indirectly regulates the expression of four genes encoding extracytoplasmic function (ECF) sigma factors (Overgaard et al., 2006). ECF sigma factors are generally involved in the expression of genes encoding proteins involved in transport, secretion and homeostasis (Helmann, 2002). Therefore, it was suggested that DigR has a function in cell envelope homeostasis. So far, the only genes encoding proteins likely to be directly involved in exopolysaccharide biosynthesis are those mapping to the eps locus (Lu et al., 2005). The eps locus encompasses 26 genes and encodes proteins with homology to glycosyltransferases, glucanases and exopolysaccharide transporters. It is currently not known whether expression of the eps genes is directly regulated by Nla24 and Nla19. In addition to the bona fide signal transduction proteins, two DnaK orthologues have been implicated in the regulation of ECM accumulation (Fig. 2). The lack of StkA causes an increased accumulation of exopolysaccharides (Dana & Shimkets, 1993), whereas the lack of SglK causes a decrease in exopolysaccharide accumulation (Weimer et al., 1998; Yang et al., 1998a).

Extracellular vesicles, zipper-like contact zones and contact-dependent activities

Recently, electron microscopic tomography revealed spectacular details about the extracellular space of M. xanthus cells (Palsdottir et al., 2009). In these analyses, M. xanthus cells that had formed a biofilm-like structure were analysed, thus allowing the careful analysis of features in the extracellular space and of cells in direct contact. The extracellular space was found to be completely filled with vesicles with a diameter of 30–60 nm (Fig. 3, red-framed panel). These vesicles were either in direct contact with the outer membrane of intact cells or were linked to the outer membrane of intact cells by proteinaceous tethers. Occasionally, individual vesicles were also connected by proteinaceous tethers. The vesicle interior showed features consistent with the presence of cargo or an internal organization. In areas in which two cells were in direct contact, zipper-like structures were observed connecting the outer membranes (Fig. 3, green-framed panel). These zipper-like structures often showed accumulation of material that is most likely of a proteinaceous nature. No direct outer membrane fusions were observed between two cells.


Extracellular vesicles and zipper-like contact zones in Myxococcus xanthus. Panels framed in red: the extracellular space between M. xanthus cells is filled with membrane vesicles. (a) Vesicle close to the membrane of an intact cell, (b and d) vesicles tethered to the cell membrane by proteinaceous tethers, (c) vesicle in contact with three cells and (e) vesicles in linear arrangement close to the cell membrane and connected by proteinaceous tethers. Scale bars=100 nm. Panel framed in green: (a and b) zipper-like structures in regions of direct cell–cell contact. Green arrows indicate the borders of the zipper-like structures. Scale bars=100 nm. Both panels were reproduced from Palsdottir et al. (2009) with permission from the American Society of Microbiology.

Palsdottir et al. (2009) suggested that the membrane vesicles and zipper-like structures in cell–cell contact regions are likely to play central roles in cell–cell interactions in M. xanthus. Indeed, M. xanthus cells display a number of contact-dependent cell–cell interactions that may now have found their structural foundations in these structures. For example, Tgl is an outer membrane lipoprotein (Rodriguez-Soto & Kaiser, 1997) required for assembly of the outer membrane secretin PilQ, which acts as a conduit for T4P in the outer membrane (Nudleman et al., 2006). tgl mutants are unable to assemble T4P (Wall et al., 1998); however, if tgl mutant cells are brought into contact with tgl+ cells, T4P assembly and T4P-dependent motility are stimulated in the tgl mutant cells (Hodgkin & Kaiser, 1977; Wall et al., 1998). Similarly, mutants of the A-motility system carrying mutations in any of the five genes cglBCDEF can be stimulated to move by means of the A-motility system by contact with cells containing a corresponding wild-type allele (Hodgkin & Kaiser, 1977). Of the five cglBCDEF genes, only cglB has been analysed and shown to also encode a lipoprotein (Rodriguez & Spormann, 1999). In the case of the tgl and cglB mutants, stimulation depends on the transfer of Tgl or CglB proteins, respectively, from donor to recipient cells (Nudleman et al., 2005). The transfer of proteins was suggested to depend on fusions of the outer membranes of donor and recipient cells (Nudleman et al., 2005). Given that Palsdottir and colleagues did not observe outer membrane fusions between two cells, an alternative explanation would be that the membrane vesicles transfer membrane proteins between cells (Palsdottir et al., 2009).

A different type of cell–cell contact-dependent phenomenon was observed in the case of the FrzCD protein (Mauriello et al., 2009a). The FrzCD protein is a cytoplasmic MCP orthologue and part of the Frz chemosensory system, which regulates the cellular reversal frequency (Zusman et al., 2007). FrzCD localizes to clusters that continuously change size, number and position. When moving cells make side-to-side contacts, FrzCD clusters in adjacent cells show a transient alignment. This phenomenon suggests that M. xanthus cells harbour a trans-envelope signalling system that allows the specific interaction between M. xanthus cells and in this way enabling the distinction between self (an adjacent M. xanthus cell) and nonself (a surface). The zipper-like structures observed by Palsdottir and colleagues would be a good candidate for containing such a system.

Gliding motility

Cell motility is essential for most – if not all – social activities of M. xanthus. Myxococcus xanthus cells move by gliding motility using two distinct motility systems referred to as the A- and the S-motility system, respectively (Hodgkin & Kaiser, 1979b) (for a recent review, see Leonardy et al., 2008). The two motility systems operate independent of each other (Hodgkin & Kaiser, 1979a, b) to generate motive force in the same direction (Kaiser & Crosby, 1983; Leonardy et al., 2007) and display different selective advantages on different surfaces (Shi & Zusman, 1993b); yet, their activities are coordinated. S-motility is generally only operational when cells are within contact distance of each other, while A-motility is operational in single cells as well as in groups of cells (Hodgkin & Kaiser, 1979a, b). Occasionally, cells undergo reversals in which cells stop and then resume gliding in the opposite direction, with the old leading pole becoming the new lagging cell pole (Blackhart & Zusman, 1985).

The T4P molecular machine provides the force for S-motility (Kaiser, 1979; Wu & Kaiser, 1995). T4P are polarly localized structures and are present only at the leading cell pole (Sun et al., 2000; Mignot et al., 2005). Force generation by T4P involves three steps: extension of T4P from the leading cell pole, attachment to a surface, followed by retraction. While extension does not generate a force sufficient to move a cell, a force exceeding 100 and 150 pN per T4P is generated during retractions in Neisseria gonorrhoeae (Maier et al., 2002) and M. xanthus (Clausen et al., 2009), respectively, and this force is sufficiently large to pull a bacterial cell forward (Merz et al., 2000; Sun et al., 2000; Skerker & Berg, 2001).

In addition to T4P, S-motility depends on lipopolysaccharide O-antigen (Bowden & Kaplan, 1998) and ECM (Arnold & Shimkets, 1988). It is currently not known how lipopolysaccharide O-antigen affects T4P motility. The connection between ECM and T4P-dependent motility was established by the observation that the polysaccharide portion of ECM triggers the retraction of T4P (Li et al., 2003). This finding may also explain the contact dependence of T4P-dependent motility because only cells within contact distance of each other would be able to retract T4P efficiently (Li et al., 2003).

It is currently not known how the A-motility system generates mechanical force. However, two models have been proposed. Myxococcus xanthus cells moving by means of the A-motility system leave behind a slime trail. The composition of the slime trail is currently unknown. Wolgemuth et al. (2002) suggested that A-motility depends on secretion and the subsequent hydration of a polyelectrolyte gel from nozzle-like structures embedded in the cell envelope. In M. xanthus, nozzle-like structures were observed at both poles by electron microscopy; however, polyelectrolyte secretion was suggested to occur only at the lagging cell pole, i.e. the pole opposite to that containing T4P. This model is supported by the finding that numerous genes required for A-motility encode proteins involved in polymer synthesis and export (Youderian et al., 2003; Yu & Kaiser, 2007). Moreover, the RomR response regulator protein, which is required for A-motility, is localized in a bipolar, asymmetric pattern with a large cluster at the lagging pole (Leonardy et al., 2007). The precise function of RomR in A-motility is not known; however, the localization pattern of RomR suggests that at least part of the A-motility machinery is localized to the lagging cell pole.

In the second model for A-motility, force generation depends on multiple focal adhesion complexes distributed along the cell body (Mignot et al., 2007). These complexes are defined by the AglZ protein, which is required for A-motility (Yang et al., 2004). Using an active AglZ protein fused to the yellow fluorescent protein (YFP), Mignot et al. (2007) observed that AglZ–YFP forms clusters localized along the cell body. These clusters remain at fixed positions relative to the substratum when cells move. The clusters were proposed to assemble from a large cluster of AglZ at the leading cell pole and then disperse at the lagging cell pole. According to this model, motive force would be generated by a protein complex that includes AglZ, spans the cell envelope, adheres to the substratum and pulls on a cytoskeletal structure (Mignot et al., 2007). Recently, AglZ was suggested not to be part of the A-motility machinery, but rather to be a negative regulator of the Frz chemosensory system that regulates the cellular reversal frequency (Mauriello et al., 2009b). Thus, the precise function of the focal adhesion complexes defined by AglZ is currently unknown. A model for A-motility involving multiple motors located along the cell body is supported by the observations that cells that move only by means of the A-motility system move with the same speed irrespective of cell length (Sun et al., 1999), and that the leading part of a filamentous cell can move while the lagging part remains stationary (Sliusarenko et al., 2007).

Over the years, a large number of A-motility mutants have been isolated. Only a few of these mutants have been analysed at the molecular and cellular level. Future molecular and cellular analyses of these mutants, in combination with the identification of the genes encoding the proteins in the nozzle-like structures and in the focal adhesion complexes, will likely help resolve the mechanism(s) underlying A-motility.

Intercellular signalling during fruiting body formation

Fruiting body formation depends extensively on intercellular communication between M. xanthus cells. Five intercellular signals have been defined genetically (Hagen et al., 1978; Downard et al., 1993). However, only two of these signals, the A- and C-signals, have been characterized in some detail biochemically and functionally. These two systems have different functions. The A-signalling system functions to ensure that fruiting body formation does not initiate unless a sufficiently high number of cells are starving. The C-signalling system functions to ensure the correct temporal order of rippling, aggregation and sporulation. The C-signal also provides cells with positional information, ensuring the spatial coupling of aggregation of cells into fruiting bodies and sporulation of cells that have accumulated inside fruiting bodies.

The A-signal becomes important for development after 2 h of starvation (Kuspa et al., 1986). The A-signal consists of two fractions: a heat-stable and a heat-labile fraction (Kaplan & Plamann, 1996). The heat-stable fraction is a mixture of amino acids and peptides. The heat-labile fraction consists of at least two proteases. According to the current model for the mechanism of the A-signal, cells during the early stages of starvation secrete a mixture of proteases that digest surface proteins, causing the release of peptides and amino acids that have A-signal activity. If a threshold concentration of A-signal is reached, the A-signal is sensed and the expression of A-signal-dependent genes follows, resulting in the progression of the developmental programme. The A-signalling system is not a functional analogue of homoserine lactone-based quorum-sensing systems in gram-negative bacteria because the A-signal at a low concentration supports development and at a high concentration cell growth.

Six mutants (asgA to asgE and sigD) that fail to produce the A-signal have been isolated (Kaplan & Plamann, 1996; Cho & Zusman, 1999; Garza et al., 2000; Viswanathan et al., 2006). The corresponding proteins are thought to be components of a regulatory pathway(s) important for A-signal synthesis. Noticeably, none of these six genes encode proteases and the A-signal proteases are still unidentified. Several mutations have been identified that cause a bypass of the A-signal requirement for expression of A-signal-dependent genes (Kaplan et al., 1991; Bowden & Kaplan, 1998; Xu et al., 1998). However, a receptor for the A-signal remains to be identified. Recent data have also suggested a connection between asg genes – but not the A-signal – and predation, i.e. several asg mutants have a reduced predation efficiency (Pham et al., 2005), and in the presence of prey, the A-signal is dispensable for fruiting body formation, but not for sporulation (Berleman & Kirby, 2007).

The C-signal becomes important for fruiting body formation after 6 h of starvation (Kroos & Kaiser, 1987) and is absolutely required for rippling, aggregation and sporulation (Shimkets et al., 1983). The C-signal acts in a threshold-dependent manner to induce rippling, aggregation and sporulation (Kim & Kaiser, 1991; Li et al., 1992; Kruse et al., 2001), i.e. rippling is induced at a low threshold, aggregation at an intermediate threshold and at a high threshold level sporulation is induced. The C-signal is a 17-kDa protein (p17) and its synthesis depends on the csgA gene (Kim & Kaiser, 1990b, c; Lobedanz & Søgaard-Andersen, 2003). csgA codes for a 25-kDa protein (p25) (Kruse et al., 2001), which is proteolytically cleaved to generate p17 (Lobedanz & Søgaard-Andersen, 2003; Rolbetzki et al., 2008). During the proteolytic cleavage of p25, approximately 8 kDa are removed from the N-terminus and p17 corresponds to the C-terminus of p25 (Lobedanz & Søgaard-Andersen, 2003; Rolbetzki et al., 2008). The subcellular localization of p17 and p25 is controversial. Wild-type cells stained with antibodies against a peptide corresponding approximately to p17, followed by staining with secondary antibodies with colloidal gold and analysed by transmission electron microscopy showed gold particles uniformly associated with the ECM and the cytoplasm (Shimkets & Rafiee, 1990), whereas p17 and p25 were only detected in the outer membrane and not in the cytoplasm in biochemical fractionation experiments on whole-cell lysates (Lobedanz & Søgaard-Andersen, 2003). Moreover, p17 and p25 were detected in the outer membrane in a dif (dsp) mutant deficient in ECM biosynthesis (Lobedanz & Søgaard-Andersen, 2003). In total, these observations suggest that p17 as well as p25 are anchored in the outer membrane.

p25 accumulates in vegetative cells; however, it is only cleaved to generate p17 during starvation (Kruse et al., 2001). Recently, the regulatory mechanism restricting p25 cleavage to starving cells was elucidated (Fig. 4). As both p25 and p17 are anchored in the outer membrane and proteolysis of p25 to p17 is blocked by inhibitors of serine proteases (Lobedanz & Søgaard-Andersen, 2003), it was hypothesized that the protease responsible for p25 cleavage is a secreted serine protease. A candidate approach led to identification of the subtilisin-like protease PopC as the protease that directly cleaves p25 (Rolbetzki et al., 2008). The mechanism underlying the regulated proteolysis of p25 is based on the regulated secretion of PopC. PopC accumulates in the cytoplasm of vegetative cells and is only secreted by starving cells. Therefore, despite the fact that both PopC and p25 accumulate in vegetative cells, they are only present in the same cell compartment in starving cells, thus restricting p25 cleavage to starving cells. Once secreted, PopC is rapidly degraded and only acts in cis. The fast degradation, combined with the slow secretion of PopC, likely ensures the slow accumulation of p17 on the cell surface, which is necessary for the proper function of the C-signal as a developmental timer and morphogen. Neither PopC nor p25 has signal peptides and it is currently unknown how these two proteins are secreted.


Model of regulation of p25 cleavage by secretion-regulated proteolysis. In vegetative cells, p25 accumulates in the outer membrane and PopC in the cytoplasm. Upon starvation, PopC is proposed to be secreted by a secretion system (marked in blue) that remains to be identified. Subsequent to the secretion of PopC, p25 is cleaved to p17 and PopC rapidly degraded.

p17 is anchored in the outer membrane and is thus nondiffusible (Lobedanz & Søgaard-Andersen, 2003). In addition, C-signal transmission requires active motility and proper cell alignment (Kroos et al., 1988; Kim & Kaiser, 1990a, d). Based on these observations, it has been suggested that C-signal transmission is contact dependent and involves direct cell–cell contacts. According to this model, C-signal transmission depends on the interaction between p17 and a p17-receptor on an adjacent cell. This receptor remains to be identified. C-signal transmission thus adds to the list of activities in M. xanthus that are cell–cell contact dependent.

Genomic evidence for extracellular activities – signal transduction and regulatory proteins

The 9.1 Mb M. xanthus genome contains 7380 protein coding sequences, with gene duplications and divergence apparently being the major contributors to genome expansion (Goldman et al., 2006). Genome expansion did not occur randomly, but typically involved genes encoding proteins involved in sensing and responding to changes in the environment (Goldman et al., 2006). These proteins include 38 ECF sigma factors (Kroos & Inouye, 2008). Among the ECF sigma factors in M. xanthus, three have been analysed and are involved in the regulation of carotenoid biosynthesis, motility and development (Kroos & Inouye, 2008). The M. xanthus genome also encodes 251 proteins of two-component regulatory systems (Shi et al., 2008; Whitworth & Cock, 2008). Among these proteins, at least 40 have functions in motility or development (Shi et al., 2008). Interestingly, only 45 of the 118 histidine protein kinases are predicted to be integral membrane proteins, suggesting that only a subset of these kinases are directly involved in sensing external signals (Shi et al., 2008). The M. xanthus genome also encodes eight chemosensory systems (Kirby et al., 2008). Among these, four have been analysed and shown to have a function in the regulation of motility, development and synthesis of the ECM (Kirby et al., 2008). Finally, the M. xanthus genome encodes 102 protein Ser/Thr kinases (Inouye et al., 2008), many of which are involved in the regulation of motility or development (Inouye et al., 2008). Sixty-six of these kinases are predicted to be integral membrane proteins, suggesting that they are involved in sensing external signals (Inouye et al., 2008). Interestingly, regulatory proteins of one-component systems (Ulrich et al., 2005) are strikingly underrepresented in the M. xanthus genome (Goldman et al., 2006; Kroos & Inouye, 2008). The logic underlying the over- and underrepresentation of certain regulatory systems remains to be explored.

Genomic evidence for extracellular activities – protein secretion systems and ATP-binding cassette (ABC) transporters

Protein secretion has so far not been studied in detail in M. xanthus. To explore the genomic potential in M. xanthus for extracellular functions, we performed a genome-wide screen for protein secretion systems. This analysis suggests that M. xanthus has a large potential for protein secretion.

Myxococcus xanthus contains all the systems for the translocation of unfolded proteins across the cytoplasmic membrane to the periplasm and for the integration of membrane proteins into the cytoplasmic membrane (Table 1) (for a review, see Driessen & Nouwen, 2008). In the Sec system, M. xanthus is only missing the chaperone SecB; however, SecB is reportedly dispensable (Driessen & Nouwen, 2008). Moreover, M. xanthus contains a signal recognition particle and its receptor as well as the membrane protein insertase YidC.

View this table:

Proteins of Sec system, Sec-dependent systems and Tat system encoded by the Myxococcus xanthus genome

M. xanthus geneComponentFunction and comments
Sec system and Sec-dependent systems
Not foundSecBProtein export chaperone
MXAN_5345SecAProtein translocase, ATPase
MXAN_3319SecYProtein translocase
MXAN_3071SecEProtein translocase; only identified using hidden Markov model
MXAN_2818SecGProtein translocase
MXAN_4691SecDSecYEG protein translocase auxillary subunit
MXAN_4690SecFSecYEG protein translocase auxillary subunit
MXAN_4692YajCSecYEG protein translocase auxillary subunit
MXAN_4854FfhSignal recognition particle
Remains to be identified4.5S RNASignal recognition particle
MXAN_5735FtsYSignal recognition particle-docking protein
MXAN_7509YidCMembrane protein insertase
Tat system
MXAN_5905TatBProtein translocase
MXAN_5904TatCProtein translocase
MXAN_2960TatAProtein translocase
Signal peptidases
MXAN_3509Signal peptidase ICleavage of type I signal peptides
MXAN_0368Signal peptidase IICleavage of type II signal peptides
MXAN_0369Signal peptidase IICleavage of type II signal peptides
MXAN_3930Signal peptidase IICleavage of type II signal peptides
MXAN_3944Signal peptidase IICleavage of type II signal peptides
  • * Myxococcus xanthus genes were identified using three strategies: information in the JCVI M. xanthus database, blastp searches of the M. xanthus genome using either protein sequences from Escherichia coli, Bacillus subtilis, Pseudomonas aeruginosa and Streptomyces coelicolor or hidden Markov models.

  • The prepilin peptidase GspO is mentioned in the context of the T2SS and a T4P prepilin peptidase (PilD, MXAN_5779) is found in the T4P gene cluster.

The twin-arginine translocation (Tat) pathway is used to translocate folded proteins across the cytoplasmic membrane (Berks et al., 2003). Myxococcus xanthus contains all three components (TatA, TatB and TatC) of this system (Kimura et al., 2006) (Table 1). An M. xanthus tatBC deletion mutant is viable and displays a pleiotropic phenotype with a decreased growth rate, an inability to form spore-filled fruiting bodies and altered surface properties (Kimura et al., 2006). The majority of the predicted Tat substrates are hypothetical proteins, oxidoreductases, peptidases and lipoproteins (Kimura et al., 2006).

To help proteins cross the cytoplasmic membrane, M. xanthus contains a single gene encoding a signal peptidase I, which processes nonlipoprotein substrates that are exported by the SecYEG pathway or the Tat pathway, and four signal peptidase II, which cleave off signal peptides of lipoproteins exported by the Sec pathway (for a review, see Paetzel et al., 2002).

Type II secretion systems (T2SS) represent a major pathway for translocation of proteins from the periplasm across the outer membrane (Filloux et al., 2004). T2SS consist of a core of 10–12 proteins (Filloux et al., 2004). The M. xanthus genome encodes an intact T2SS system, and most of the components of this system are encoded in a single gene cluster (MXAN_2504–MXAN_2515) (Table 2). Myxococcus xanthus does not contain genes encoding GspN and GspM; however, these two components are dispensable for the function of T2SS (Filloux et al., 2004). A GspO prepilin peptidase is encoded in a small gene cluster, which also encodes paralogues of GspD and GspE (MXAN_3105–MXAN_3107) (Table 2). In addition, the genome contains at least four genes encoding orphan paralogues of GspC, GspE and GspG (Table 2). Several of the proteins of T2SS are highly homologous to proteins of T4P (Planet et al., 2001; Peabody et al., 2003). All the genes for T4P are localized to a single gene cluster (MXAN_5771–MXAN_5788), with the exception of the tgl gene (MXAN_3084), which encodes an outer membrane lipoprotein required for assembly of the outer membrane secretin PilQ (Nudleman et al., 2006), and four orphan paralogues of the PilT secretion ATPase (MXAN_1995, MXAN_0415, MXAN_6705 and MXAN_6706) (Clausen et al., 2009).

View this table:

Proteins of T2SS encoded by Myxococcus xanthus genome

M. xanthus geneComponentFunction and comments
MXAN_2504GspLIntegral cytoplasmic membrane protein
MXAN_2505GspKMinor pseudopilin
MXAN_2506GspJMinor pseudopilin
MXAN_2507GspIMinor pseudopilin
MXAN_2508GspHMinor pseudopilin
MXAN_2509GspGMajor pseudopilin
MXAN_2510GspGMajor pseudopilin
MXAN_2512GspFIntegral cytoplasmic membrane protein
MXAN_2513GspESecretion ATPase
MXAN_2514GspDOuter membrane secretin
MXAN_2515GspCIntegral cytoplasmic membrane protein
MXAN_3105GspOPrepilin peptidase
MXAN_3106GspDOuter membrane secretin
MXAN_3107GspESecretion ATPase
MXAN_2332GspCIntegral cytoplasmic membrane protein; orphan
MXAN_2658GspESecretion ATPase; orphan
MXAN_3824GspGMajor pseudopilin; orphan
MXAN_7176GspESecretion ATPase; orphan
  • * Myxococcus xanthus genes were identified using two strategies: information in the JCVI M. xanthus database and blastp searches of the M. xanthus genome using either protein sequences from Escherichia coli, Bacillus subtilis, Pseudomonas aeruginosa and Streptomyces coelicolor.

  • MXAN_3105–MXAN_3107 make up a three-gene cluster outside the major T2SS gene cluster.

The alternatives to the two-step secretion of proteins to the outside of cells using the Sec/Tat pathway in combination with the T2SS are the type I, III, IV and VI secretion systems (Economou et al., 2006). These systems transfer proteins from the cytoplasm directly to the cell exterior or into the cytoplasm of a eukaryotic cell in one-step mechanisms. The M. xanthus genome encodes all these systems, with the exception of a type IV secretion system.

The type I secretion system (T1SS), often also referred to as ABC secretion systems or ABC protein exporters, generally consist of two inner membrane proteins, an ATPase (the ABC protein), a membrane fusion protein and the outer membrane TolC protein (Andersen et al., 2001). T1SS have structural similarity to ABC transporters, which are involved in import and export of a wide variety of different compounds (Young & Holland, 1999), thus making it difficult to predict how many T1SS M. xanthus encodes. The M. xanthus genome encodes at least 73 ABC proteins and 15 TolC-like proteins (Table 3). Based on the genetic organization of the genes encoding the ABC proteins, the M. xanthus genome likely codes for 60 ABC transporters.

View this table:

TolC paralogs encoded by the Myxococcus xanthus genome

M. xanthus genePrimary annotationGenetic organization
MXAN_0985Putative cobalt–zinc–cadmium resistance proteinEfflux transporter
MXAN_0990Cation efflux system protein CusCEfflux transporter
MXAN_1791Outer membrane efflux proteinABC transporter
MXAN_3424Outer membrane efflux proteinOrphan
MXAN_3431Outer membrane efflux proteinOrphan
MXAN_3447Hypothetical proteinEfflux transporter
MXAN_3744Hypothetical proteinOrphan
MXAN_3905Efflux transporter, outer membrane efflux proteinEfflux transporter and ABC transporter
MXAN_4176Outer membrane efflux proteinEfflux transporter and ABC transporter
MXAN_4198Putative outer membrane macrolide efflux protein, efflux transporter, HAE1 family, outer membraneABC transporter
MXAN_5030Efflux proteinEfflux transporter
MXAN_6176Outer membrane efflux proteinEfflux transporter
MXAN_6487Outer membrane efflux protein domain protein, efflux transporter, HAE1 family, outer membraneOrphan
MXAN_7238Efflux proteinEfflux transporter
MXAN_7436Outer membrane efflux proteinEfflux transporter
  • * Myxococcus xanthus genes were identified using two strategies: information in the JCVI M. xanthus database and blastp searches of the M. xanthus genome using the TolC protein sequences from Escherichia coli.

  • Indicates the presence of genes encoding components of an ABC-transporter or of an efflux transporter adjacent to the tolC gene.

Type III secretion systems (T3SS) support protein export from the bacterial cytoplasm across the periplasm and the outer membrane directly to either the extracellular space or the cytosol of eukaryotic cells (Cornelis, 2006). T3SS are often described as specific to pathogenic bacteria; however, bioinformatic studies have revealed the presence of T3SS genes in nonpathogenic bacteria, suggesting that T3SS may have functions not related to pathogenesis (Pallen et al., 2005). We found two gene clusters in the M. xanthus genome encoding subunits of T3SS. Both gene clusters are highly degenerate and none of them seem to encode an intact T3SS (Fig. 5). Cluster I (MXAN_2434–MXAN_2464) (Table 4; Fig. 5) encodes eight of the nine proteins that are highly conserved in T3SS – the missing protein being the outer membrane secretin (YscC in the Yersinia spp. nomenclature). The remaining eight proteins include those making up the cytoplasmic C ring, the inner membrane MS ring and the associated export apparatus. Thus, this system is missing the genes encoding the outer membrane ring and the needle structure. Cluster I encodes several proteins containing tetratricopeptide repeats and one containing a forkhead-associated domain. Proteins with these domains have been reported to function as T3SS chaperones and regulators (Pallen et al., 2002, 2003). Cluster II (MXAN_5643–MXAN_5654) (Table 5; Fig. 5) consists of 11 genes and also only encodes the proteins that would make up the C ring, the MS ring and the export apparatus. Importantly, this cluster does not encode the ATPase (YscN in the Yersinia spp. nomenclature) that energizes protein export. Thus, both clusters lack the proteins that would make up the secretion apparatus in the outer membrane and the needle structure. Consistently, there are no reports in the literature on M. xanthus cells containing needle structures protruding from the cell surface. It is not clear whether two gene clusters encode functional protein secretion/translocation systems. However, it is intriguing that despite the apparent degeneracy of the two gene clusters, the components present precisely match those making up the part of T3SS that allows protein translocation across the cytoplasmic membrane. Therefore, it remains a possibility that these two systems are specifically optimized to the needs of M. xanthus and are involved in protein translocation only across the cytoplasmic membrane. Alternatively, both systems could function with one or more of the three secretins of T2SS (Table 2) and T4P. The order of conserved genes in the two clusters is largely conserved (Fig. 5) and is similar to that of T3SS gene clusters in other organisms, thus making it difficult to determine whether the two clusters are connected by a gene duplication event. However, the very different identity/similarity values between paralogues in the two clusters (Fig. 5) suggest that the two clusters did not arise from a gene duplication event.


Putative T3SS in Myxococcus xanthus. (a) Generic T3SS system with the localization of YscL and the nine core proteins conserved in all T3SS. Nomenclature of the proteins is the indicated letter preceded by Ysc (Yersinia spp. nomenclature). The inner and outer membranes are indicated. (b) Potential structure of the T3SSs encoded by gene clusters I and II in M. xanthus; colour code and orientation is as in (a). (c) Genetic organization of the two T3SS gene clusters in M. xanthus; colour code is as in (a). The connections list the % identity/similarity between paralogues.

View this table:

Proteins encoded by the type III secretion system gene cluster I in Myxococcus xanthus

M. xanthus geneComponentFunction and comments
MXAN_2434Hypothetical conserved in Myxococcales
MXAN_2435TPR repeat containing protein
MXAN_2436Hypothetical conserved in Stigmatella aurantiaca
MXAN_2437RNA polymerase sigma factor, FliA/WhiG/SigD family
MXAN_2438YscI?Hypothetical conserved in Myxococcales with low homology to YscI
MXAN_2439YscJLipoprotein in cytoplasmic membrane; component of MS ring
MXAN_2440Hypothetical conserved in Stigmatella aurantiaca
MXAN_2441YscLMay tether YscN ATPase to export channel
MXAN_2442YscNType III secretion system ATPase
MXAN_2443Hypothetical conserved in Stigmatella aurantiaca
MXAN_2444YscP?Low homology to flagellar hook-length control protein FliK, which is related to YscP
MXAN_2445YscQLikely makes up cytoplasmic C-ring
MXAN_2447YscRBasal structure, cytoplasmic membrane protein
MXAN_2448YscSBasal structure, cytoplasmic membrane protein
MXAN_2449YscTBasal structure, cytoplasmic membrane protein
MXAN_2450YscUBasal structure, cytoplasmic membrane protein
MXAN_2451YscUBasal structure, cytoplasmic membrane protein
MXAN_2452Hypothetical conserved in Stigmatella aurantiaca
MXAN_2453TPR repeat containing protein
MXAN_2454Hypothetical conserved in Stigmatella aurantiaca
MXAN_2458Hypothetical conserved in Stigmatella aurantiaca
MXAN_2459TPR repeat containing protein
MXAN_2461TPR repeat containing protein
MXAN_2463LcrD/YscVBasal structure, cytoplasmic membrane protein
MXAN_2464FHA domain containing protein
  • * Myxococcus xanthus genes were identified using two strategies: information in the JCVI M. xanthus database and blastp searches of the M. xanthus genome using protein sequences from Yersinia spp..

  • Nomenclature is based on that of Yersinia spp. (Cornelis et al., 2006).

  • Included in the type III secretion system gene cluster I because a paralogous gene is also present in the type III secretion system gene cluster II.

  • § Proteins that are universally conserved in T3SS (Cornelis et al., 2006).?, homology between the M. xanthus protein and the Yersinia protein is low.

View this table:

Proteins encoded by the type III secretion system gene cluster II in Myxococcus xanthus

M. xanthus geneComponentFunction and comments
MXAN_5643LcrD/YscVBasal structure, cytoplasmic membrane protein
MXAN_5644YscUBasal structure, cytoplasmic membrane protein
MXAN_5645YscTBasal structure, cytoplasmic membrane protein
MXAN_5646YscSBasal structure, cytoplasmic membrane protein
MXAN_5647YscRBasal structure, cytoplasmic membrane protein
MXAN_5649YscQLikely makes up cytoplasmic C-ring
MXAN_5651YscJLipoprotein in cytoplasmic membrane; component of MS ring
MXAN_5652YscI?Hypothetical conserved in Myxococcales with low homology to YscI
MXAN_5654Hypothetical conserved in Myxococcales
  • * Myxococcus xanthus genes were identified using two strategies: information in the JCVI M. xanthus database and blastp searches of the M. xanthus genome using protein sequences from Yersinia spp.

  • Nomenclature is based on that of Yersinia spp. (Cornelis et al., 2006).

  • Proteins that are universally conserved in T3SS (Cornelis et al., 2006).?, homology between the M. xanthus protein and the Yersinia protein is low.

Type VI secretion systems (T6SS) also allow the direct translocation of proteins from the bacterial cytoplasm to the extracellular space or to the cytosol of eukaryotic cells (Cascales, 2008; Filloux et al., 2008). These systems have mostly been analysed in pathogenic bacteria and are also commonly referred to as virulence factors. Bioinformatic analyses have, however, shown that T6SS are also widespread in nonpathogenic bacteria (Bingle et al., 2008; Boyer et al., 2009). The structure of T6SS is unknown. However, bioinformatic analyses have shown that 13 proteins constitute the core of T6SS (Bingle et al., 2008; Boyer et al., 2009). It was previously noticed that the M. xanthus genome contains a single gene cluster for T6SS and that this gene cluster encodes all 13 core proteins (Bingle et al., 2008; Boyer et al., 2009) (Table 6), suggesting that this gene cluster codes for a functional T6SS. Interestingly, a deletion covering MXAN_4807–MXAN_4813 shows no growth defects or motility defects; however, the mutant is unable to form mature fruiting bodies (A. Konovalova, unpublished data).

View this table:

Proteins of the type VI secretion system encoded in the Myxococcus xanthus genome

M. xanthus geneCore proteins of T6SS/Vibrio cholerae O1 El tor N16961 geneTrivial name and comments
MXAN_4805Conserved hypothetical
MXAN_4813COG0542/VC_A0116ClpV ATPase
MXAN_4815Conserved hypothetical
  • * Core proteins of T6SS are indicated by their respective COG identity (cluster of orthologous groups of proteins) and the orthologous gene in Vibrio cholerae O1 El tor N16961 (Boyer et al., 2009).

Genomic evidence for extracellular activities – secreted proteins

Given that T4P systems have also been implicated in the secretion of proteins other than the T4P subunit (Hager et al., 2006), it is clear that M. xanthus has a large potential for protein secretion. Which proteins are then secreted? Myxococcus xanthus has been shown to secrete protease(s), lysozyme(s), amidase(s) and glucosamidase(s) (Hart & Zahler, 1966; Sudo & Dworkin, 1972); however, the corresponding genes are not known. Moreover, no experimental analyses have addressed the composition of the secretome in M. xanthus. To learn more about the secretome of M. xanthus, we performed a genome-wide screen for secreted proteins encoded by the M. xanthus genome. For comparison, we included the genomes of E. coli, Bacillus subtilis, Pseudomonas aeruginosa, Caulobacter crescentus and Streptomyces coelicolor in these analyses. As shown in Table 7, this in silico analysis illustrates that M. xanthus likely secretes a large number of proteins. However, when corrected for genome size, the relative number of potentially secreted proteins is not significantly different from those of the five other model organisms, with the exception of lipoproteins, which are predicted to make up 6% of the total M. xanthus proteome (Table 7). In line with this finding, it is of interest that the M. xanthus genome encodes four signal peptidase II (Table 1). It should be emphasized that due to the presence of Sec- and Tat-independent secretion systems, the number of secreted proteins in Table 7 likely represents an underestimate. For instance, two of the best-studied secreted proteins in M. xanthus, PopC and p25, do not contain signal peptides. Among potentially secreted proteins in M. xanthus, approximately 10% are predicted to have hydrolase activity. Forty per cent of all genes in M. xanthus are annotated as conserved hypothetical, hypothetical or protein of unknown function. However, 58% of all the proteins predicted to be secreted have unknown functions. Thus, many surprises are likely concerning the extracellular functions of these proteins.

View this table:

Predicted integral cytoplasmic membrane proteins and secreted proteins

Protein characteristicE. coli B. subtilis P. aeruginosa C. crescentus S. coelicolor M. xanthus
Contains trans-membrane helices
Total number10291133133483618911358
(% of total CDS)(25%)(28%)(24%)(20%)(23%)(20%)
Contains signal peptide
Total number828606130670111981383
(% of total CDS)(20%)(15%)(23%)(19%)(15%)(20%)
Contains type II signal peptide
Total number12510318676178385
(% of total CDS)(3%)(3%)(3%)(2%)(2%)(6%)
Contains Tat signal peptide
Total number3328458319161
(% of total CDS)(0.8%)(1.2%)(0.8%)(2.2%)(2.3%)(0.8%)
  • * The data are from the following strains: Escherichia coli K-12 MG1655, Bacillus subtilis ssp. subtilis str. 168, Caulobacter crescentus CB15, Pseudomonas aeruginosa PAO1, Streptomyces coelicolor A3(2) and Myxococcus xanthus DK1622.

  • Predicted using tmhmm (Krogh et al., 2001).

  • Predicted using signalp3 (Emanuelsson et al., 2007).

  • § Predicted using lipop (Juncker et al., 2003).

  • Predicted using tatp (Bendtsen et al., 2005).

Predation, secreted proteins and secondary metabolites

Myxobacteria are saprophytes and predators of other microorganisms. How then do myxobacteria prey? Laboratory experiments have shown that colonies of M. xanthus cells by an unknown mechanism can attract E. coli cells and subsequently lyse these cells (Shi & Zusman, 1993a). Conversely, the finding that the carbon of living E. coli cells in soil samples is predominantly recovered from gliding bacteria (Lueders et al., 2006) suggests that myxobacteria also use cell motility to actively track down prey. Once prey has been tracked down, killing and polymer degradation follow. Little is known about these two processes; however, killing is thought to depend on secreted hydrolytic enzymes and secondary metabolites with antimicrobial activity, while polymer degradation would depend on secreted hydrolytic enzymes.

Likewise, the function and importance of the many potentially secreted proteins in predation are currently unknown. The degradation of polymers is thought to be a cooperative process (Rosenberg et al., 1977). Thus, if casein is the sole carbon source, then M. xanthus cells only show significant growth above a certain cell density and the growth rate increases with increasing cell density. Moreover, cooperative growth is correlated with increases in extracellular protease activity and in the concentration of hydrolysed casein in the medium, suggesting the cooperative hydrolysis of casein. Consistent with this idea, if casein is replaced with hydrolysed casein, M. xanthus growth is independent of cell density.

Genome mining has shown that myxobacteria have a large potential for secondary metabolite synthesis. Thus, genes encoding proteins involved in secondary metabolite synthesis are highly enriched in the M. xanthus genome and make up 8.6% of the genome (Goldman et al., 2006). In total, the M. xanthus genome contains at least 18 gene clusters coding for nonribosomal peptide synthetases and polyketide synthases (Bode & Müller, 2008). For five of these biosynthetic gene clusters, a link has been established to the compound families synthesized (Wenzel & Müller, 2009). Specifically, the biosynthetic gene clusters involved in the biosynthesis of myxovirescins (antibiotics), myxalamides (antibiotics), myxochelines (siderophores), myxochromides and DKxanthenes (the dyes that give M. xanthus colonies their characteristic yellow colour) have been identified. Experimental evidence supports the notion that the remaining 13 gene clusters are expressed; however, the compounds synthesized remain to be identified. The function and importance of these compounds in predation are currently unknown. However, DKxanthenes are essential for spore formation (Meiser et al., 2006).

Interestingly, the secondary metabolites synthesized by various M. xanthus isolates from locations worldwide seem to vary at a high level (Krug et al., 2008). The analysis of the secondary metabolome of 98 M. xanthus isolates revealed that myxalamides, myxochelines, myxochromides and DKxanthenes were produced by all isolates, whereas other compounds such as althiomycin were only produced by two of the isolates. Moreover, the number of nonubiquitous compounds produced per strain varied from 6 to 24, with a mean of 16. These nonubiquitous compounds also included unknown secondary metabolites. A subset of the same strains was also tested in different pair-wise combinations for their ability to form mixed fruiting bodies (Fiegna & Velicer, 2005). The majority of the pair-wise interactions were negative with bidirectional inhibition of sporulation and a few combinations revealed social exploitation, i.e. a strain performed better in a mixing experiment than alone. The molecular mechanism(s) underlying these ‘antisocial’ interactions are unknown; however, the differences in the secondary metabolomes of the strains tested suggest that the secondary metabolites produced by M. xanthus are not only involved in killing other microorganisms but sometimes also their close relatives.

Generally, predation by secretion of secondary metabolites and proteins, i.e. remote predation as opposed to cell–cell contact-dependent predation as in Bdellovibrio bacteriovosus (Velicer & Mendes-Soares, 2009), suggests the existence of mechanisms that allow the predator to distinguish self (M. xanthus) from nonself (prey). How M. xanthus accomplishes this task is not known.

Concluding remarks

The extracellular biology of M. xanthus poses many interesting questions for the future: first, M. xanthus cells display a number of activities that are contact dependent. These contacts are of two kinds. There are activities that depend on direct cell–cell contacts between M. xanthus cells, i.e. transfer of the Tgl and CglB proteins, alignment of FrzCD clusters, T4P-dependent motility and C-signal transmission. Second, one activity has been described to depend on contacts between M. xanthus cells and the substratum on which they move, i.e. A-motility involving focal adhesion complexes. These contact-dependent activities suggest that M. xanthus cells contain systems that allow the discrimination between self and nonself. Likewise, predation implies the existence of such systems. A challenge for the future will be to elucidate the molecular and cellular mechanisms of these systems.

A second challenge for the future is the elucidation of the function of the T3SS and T6SS in M. xanthus. These systems are generally connected to virulence; however, M. xanthus has not been reported to be a pathogen. Then what is the function of these secretion systems? And what are the functions of the many secreted proteins for which the function is currently unknown?


We would like to thank Stuart Huntley and Chris van der Does for help with the bioinformatic analyses and Anke Treuner-Lange and Iryna Bulyha for help with Fig. 1. Work in the authors' laboratory is supported by the Max Planck Society.


  • Editor: Keith Chater


View Abstract