The Predation Strategy of Myxococcus xanthus

1Predatory bacteria are found in different phyla and apply different mechanisms to kill bacteria of other species and consume their biomass. Endobiotic predators, like the Bdellovibrio species, enter the periplasm of a prey cell, utilize its biomass to multiply and lyse the prey cell from within. In contrast, epibiotic predators induce prey lysis from the outside and feed on the released biomass. Epibiotic predation has been described for different proteobacteria, such as Ensifer adhaerens and Micavibrio admirandus (α-proteobacteria), Cupriavidus necator (β-proteobacteria), Lysobacter sp. (γ-proteobacteria) and the Myxobacteria (δ-proteobacteria).

2Myxobacteria are ubiquitous in soil and known for their elaborate multicellular behaviors: single cells form large clusters that move slowly along surfaces and secrete enzymes at high concentrations to access nutrients. When nutrients become scarce, cells coordinate their motility and aggregate into complex multicellular structures, called fruiting bodies, where they differentiate into spores (Figure 1A, white arrowhead). In this mini-review, we summarize our current understanding of the epibiotic predation strategy displayed by the model myxobacterium Myxococcus xanthus.

3Predation is a facultative behavior of M. xanthus, since it can also grow saprophytically by degrading dead biomass derived from plants or microorganisms. Its preferred energy and carbon source are amino acids and lipids, which are catabolized via various pathways. Moreover, M. xanthus lacks the biosynthetic pathways for valine, leucine, and isoleucine and thus relies on external sources for these branched amino acids. Comparative genome analysis revealed that genes associated with secondary metabolites, digestive enzymes, protein secretion, and TonB-dependent receptors/metal transport, which may be utilized for killing other bacteria and consuming their biomass, are overrepresented in M. xanthus.

4Compared to other bacterial predators, which display a limited prey spectrum, M. xanthus is a generalist that is able to feed on a broad range of bacteria and some fungi: It has been observed to prey on soil bacteria, including plant pathogens, on cyanobacteria, clinically relevant species, such as Pseudomonas aeruginosa and Staphylococcus aureus, yeasts, and other fungi. Comparative analysis of predation performance indicates that, under laboratory conditions, most prey species significantly support M. xanthus growth. Considering the broad spectrum of prey that can be utilized by M. xanthus, it is likely that different molecular mechanisms, acting either in isolation or synergistically, are required to prey on different species. Moreover, numerous studies addressing different aspects of M. xanthus predation have corroborated that predation is a complex, multilayered process that requires a combination of different traits: prey cells are encountered, recognized, and lysed, while the predator remains unharmed. Biomass released from prey is degraded, taken up, and metabolized. These processes can require individuals or a large number of predator cells and may evoke a specific response by the prey.

5Regardless of the prey species used, M. xanthus must be in close proximity to prey cells in order to induce their lysis and to benefit from their biomass (Figure 1B;;; ). M. xanthus encounters prey cells by gliding slowly on surfaces, powered by two motility mechanisms: polar type IV pili are extended and retracted to move by “social” S-motility, where cells typically form large clusters that are embedded in exopolysaccharides. Alternatively, M. xanthus cells can move as individuals, which are powered by motility engines that are laterally embedded in the cell envelope (“adventurous” A-motility). The S- and A-motility mechanisms were both shown to contribute to efficient predation.

6By using time-lapse microscopy, it was demonstrated that an individual M. xanthus cell is able to induce lysis of an E. coli cell (Figure 1B;; ). Moreover, the M. xanthus cell was observed to change its behavior upon contact with E. coli, by halting next to a prey cell and/or repeatedly reversing its direction of movement. This presumably allows M. xanthus to remain within the area and ensure lysis of all available prey cells. Cell reversals in M. xanthus are regulated by a specialized chemosensory system, the frz pathway, and mutation of key regulators reduces predation efficiency. Regulated reversals are also essential for the characteristic rippling pattern, which can be observed in preying M. xanthus populations: upon entering and lysing a prey colony, M. xanthus cells synchronize their motility and reverse in oscillating waves (Figure 1A, yellow arrowhead). It was shown, both experimentally and by computational simulation, that rippling in presence of prey increases the rate of colony expansion by the predator. At the same time, rippling reduces the mean square displacement of individual cells, which would allow the predator cells to remain within the prey area longer.

7Overall, regulated cell motility determines how frequently M. xanthus encounters prey, and therefore mutations in the respective genes reduce predation efficiency. However, the ability to kill and consume prey cells is not affected, as motility mutants are still able to lyse and grow on prey.

8According to our current understanding, M. xanthus uses a combination of secondary metabolites and secreted enzymes to kill and lyse prey cells, and to consume their biomass (Figure 1B;;; ). M. xanthus encodes for 24 biosythetic clusters that produce secondary metabolites. In addition to several unknown products, these clusters synthesize pigments, siderophores, bacteriocins, and antibiotics that target bacteria or fungi. For two antibiotics, myxovirescin A and myxoprincomide, a role in predation has been demonstrated. Myxoprincomide is a linear peptide of nine amino acids that is synthesized by a hybrid biosynthetic module of non-ribosomal peptide/polyketide synthases, but its mode of action remains unknown. Lack of myxoprincomide reduces M. xanthus growth on B. subtilis and increases the number of surviving prey cells. However, myxoprincomide is not required for growth of M. xanthus on E. coli, and also not for growth on a B. subtilis mutant that is unable to produce the antibiotic bacillaene. It therefore does not appear to be a general killing factor, but might rather be required to specifically counteract antibiotics that are produced by prey in response to M. xanthus.

9Myxovirescin A (or “TA”) is a bactericidal antibiotic, which targets signal peptidase II (LspA) and thereby inhibits lipoprotein maturation, leading to the formation of lethal cross-links between the cell wall and the inner membrane in growing cells. In co-culture experiments, significantly more E. coli cells survive in the presence of a myxovirescin A mutant compared to M. xanthus wild type, which implies a role for myxovirescin A in prey killing. However, lack of myxovirescin makes no difference for predation under nutrient-free conditions, i.e., when prey cells are not actively growing. The resulting hypothesis is that bactericidal antibiotics, such as myxovirescin A, are secreted by M. xanthus to neutralize prey metabolism, which might facilitate the subsequent degradation of prey biomass by hydrolytic enzymes. Obviously, prey killing can be accomplished by other secondary metabolites or additional mechanisms. The production of myxovirescin is controlled by the alternative sigma factor σ54 in conjunction with specific enhancer binding proteins (EBPs). σ54-dependent regulators are abundant in M. xanthus and regulate the expression of many secondary metabolites and proteins, mainly during development (e.g., ). Transcriptional control by σ54/EBPs might therefore coordinate secondary metabolite production, development, and predation.

10The production and secretion of hydrolytic proteins by M. xanthus is well documented: bacteriolytic proteins with amidase and glucosaminidase activity have been isolated from M. xanthus culture supernatant early on. Sequencing of the M. xanthus genome revealed several proteins with an annotated function related to peptidoglycan degradation, as well as numerous secreted enzymes with putative proteolytic activity. Proteases and peptidases, such as MepA, presumably contribute to predation by degrading proteins that are released from prey cells. More importantly, enzymes with peptidoglycan-degrading activity could potentially serve as killing factors that induce prey cell lysis. Recently, GluM, an outer membrane β-1,6-glucanase of the myxobacterium Corallococcus, was shown to be required for lysing the chitinous cell wall of certain fungi. The homologous protein in M. xanthus, Oar, has glucanase activity as well, but was independently shown to be involved in exporting a protease that is required for induction of the cell differentiation program. The possibility of a dual role of Oar in the predation of fungi and the regulation of development remains to be investigated, but these observations emphasize a putative role of cell wall-lytic proteins as killing factors in myxobacterial predation. However, while the expression and secretion of bacteriolytic enzymes by M. xanthus is unquestionable, predation factors that might specifically mediate the lysis of prey cells have not been identified, yet.

11The requirement for close proximity to induce prey lysis implies that the delivery of killing factors by M. xanthus is locally restricted. For protein effectors, this could be achieved by targeted secretion via protein secretion systems of the types II, III, IV or VI, which are used in other bacteria to deliver lethal effectors in the proximity or directly into prokaryotic or eukaryotic cells. The M. xanthus genome encodes for protein secretion systems of type II, III (incomplete), and VI. The type VI secretion system (T6SS) was shown to mediate antagonism between different kin of myxobacteria, and even between physiologically challenged siblings of the same kin. One effector delivered via the T6SS is the toxin MXAN_0050, a nuclease, against which the secreting cell protects itself by a concomitantly expressed anti-toxin, MXAN_0049. However, it has not been reported, whether the T6SS and MXAN_0050 contribute to killing of other species for biomass acquisition during predation. A recent study reports that the survival of E. coli in prolonged co-culture with M. xanthus correlates with mutation of the E. coli outer membrane protease, OmpT. This led to the intriguing suggestion that pre-lytic factors secreted by M. xanthus might be activated by the prey itself, indicating a putative mechanism to limit lysis to the prey cells.

12Another mode of protein secretion and delivery may be provided by outer membrane vesicles (OMVs, Figure 1B), which are produced by many Gram-negative bacteria. Indeed, M. xanthus OMVs were shown to contain lytic components and to be involved in predation. OMVs emerge after fission from the secreting cell and are released into the environment. When OMVs reach a recipient cell via diffusion, they could deliver their cargo by fusing with the outer membrane of Gram-negative bacteria, or by disintegrating next to a Gram-positive cell. The packaging of lytic compounds into a vesicle instead of freely secreting them into the extracellular space might provide several advantages: The lytic compounds remain at high local concentration, and the slow passive diffusion of a large vesicle increases the likelihood of a secreting cell to be in close proximity for nutrient uptake after prey killing, reducing the risk of exploitation by non-secreting cells. Moreover, delivering a cocktail of various lytic factors would increase efficiency and prevent resistance formation by prey. With OMVs, membrane-associated lytic components could also be transported to the prey cell, and soluble secreted cargo would be protected from the environment.

13The proteome of M. xanthus OMVs has been shown to differ from that of the periplasm and outer membrane, hinting to the presence of a dedicated protein sorting mechanism to load OMVs. OMVs contain several proteins with predicted hydrolytic functions, as well as antibiotics, such as myxovirescin A and the antifungal myxalamide. Indeed, the lytic potential of isolated M. xanthus OMVs toward E. coli was experimentally demonstrated. However, as in the case of freely secreted enzymes, it is not clear, which of the OMV cargo components mediate prey killing or just degradation of dead biomass. Furthermore, the lytic effect of M. xanthus OMVs toward Gram-negative bacteria other than E. coli or Gram-positive bacteria has not been been demonstrated, so far. In multicellular swarms, individual M. xanthus cells were observed to be connected via a network of membrane tubes and OMV chains. It was hypothesized that these appendices may play a role for inter-cellular signaling and outer membrane exchange among myxobacteria. However, membrane tubes were also observed under experimental conditions that prevent outer membrane exchange, which suggests that both processes are functionally separate. It remains to be investigated, how outer membrane structures within a myxococcal swarm relate to the proposed function of OMVs in the delivery of lytic factors during predation, and how M. xanthus cells might protect themselves against their lytic cargo.

14The synthesis and release of putative predation-specific lytic factors by M. xanthus consumes energy and resources, and can only be productive when prey cells are available. A transcriptome study singled out only a few M. xanthus genes that are differentially expressed when starved cells are supplied with live E. coli prey, suggesting the constitutive expression of the predatory arsenal under the tested conditions. However, M. xanthus gene transcription did respond both to the addition of pre-killed bacteria and of hydrolyzed peptides. These observations indicate that lysed bacteria, but not live prey cells, are perceived as nutrients. If the production of lytic factors is not regulated on the level of transcription, their release from the cell might be triggered upon prey encounter to reduce the loss of costly molecules. Indeed, individual M. xanthus cells alter their motility behavior when in contact with live prey, but not with dead cells. This indicates that M. xanthus can recognize cells of other bacterial species and can discriminate live from dead cells, but the underlying molecular mechanisms are unknown. When competing with other myxobacterial kin groups, M. xanthus is able to identify its own kin and eliminate cells of relatives. This form of self/non-self discrimination is mediated by a modular surface receptor, TraA, that differs among M. xanthus kin groups. However, a possible function for this system during predation of other bacterial species has not been described, yet.

15While the recognition and response to live prey cells remains elusive, it is clear that M. xanthus can sense and react to the presence of lysed prey: peptidoglycan (or its break-down products) induce coordinated cell reversals that induce rippling patterns. Although this behavior has been termed “predataxis” and requires the frz chemosensory pathway, it does not represent a directed, chemotactic movement of M. xanthus toward prey. In fact, several studies have shown that M. xanthus is not chemotactic toward prey, and it remains controversial, whether M. xanthus shows chemotactic behaviors in general. However, Shi and Zusman observed chemotactic behavior of E. coli prey toward M. xanthus when conducting predation assays under conditions where both M. xanthus and E. coli are motile. This led to the striking proposal that M. xanthus releases chemoattractants as a primitive form of prey ensnarement, which still awaits further investigation.

16Several prey bacteria have been observed to induce specific counteractions to protect themselves against predation by M. xanthus, with the predominant mechanisms being the formation of a mechanical barrier or the secretion of antibiotics. For example, Bacillus subtilis builds biofilm megastructures, which are filled with spores that cannot be attacked by M. xanthus. Also, biofilm-forming E. coli with a matrix of curli are less sensitive to predation by M. xanthus than non-biofilm producing mutants. Recently, co-evolution experiments with E. coli and M. xanthus demonstrated that predation, in fact, selects toward a mucoid prey phenotype. Similarly, a galactoglucane-producing Sinorhizobium meliloti strain is more resistant toward M. xanthus predation compared to the non-mucoid mutant. The exact mechanism behind the resistance provided by biofilm barriers is not clear, but they might simply obstruct access of predator cells and their lytic factors to prey. Another protective strategy is the production of antibiotics: B. subtilis was reported to escape killing by M. xanthus by production of the antimicrobial peptide bacillaene. Although this compound does not impede predator growth, it seems to transiently inhibit predation, providing B. subtilis with enough time to differentiate into predation-resistant spores. Streptomyces coelicolor produces an antibiotic, actinorhodin, upon co-cultivation with M. xanthus, which however does not confer a survival advantage. Recently, an enzyme released by Bacillus licheniformis was found to glycosylate myxovirescin A, which inactivates the antibiotic and renders the producing strain less susceptible to predation by M. xanthus.

17The M. xanthus life cycle is largely directed to maintain multicellular interactions, for example manifested in social motility and coordinated cell differentiation. Vegetative growth and predation have been discussed in the context of multicellularity, since Rosenberg et al. observed that the growth rate of M. xanthus growing in liquid culture increases depending on cell density, and suggested that nutrient acquisition during predation might involve cooperative behavior similar to a “wolf pack,” where the cooperation between individuals with specific tasks enhances the efficiency of the group: hydrolytic enzymes to degrade prey biomass are secreted by the members of a M. xanthus swarm as common goods and increase in concentration with the number of secreting cells, allowing the efficient degradation of (prey) biomass. This indicated that predation by a M. xanthus swarm might be more efficient than by individual cells. Indeed, Zhang et al. recently reported that individual M. xanthus cells most often leave E. coli prey they have killed without degrading the biomass, presumably because they do not produce a sufficient amount of degradative enzymes.

18The “wolf pack” analogy has been used frequently to describe and discuss the predation behavior of myxobacteria. Recently, the concept of wolf pack predation by M. xanthus was revisited, and a new interpretation of the M. xanthus predation strategy was proposed, which highlights that the secretion of lytic factors and nutrient release are in fact proportionate to cell density. Additional arguments were that the secretion of hydrolytic factors during predation likely is constitutive, and that cell density-dependent nutrient hydrolysis in liquid culture cannot be extrapolated to the native soil habitat of M. xanthus, in which diffusion parameters of hydrolytic compounds and predator motility play a different role. Indeed, on a solid agar surface an individual M. xanthus cell is able to lyse several E. coli cells. Moreover, transposon mutagenesis revealed that mutants that lack exopolysaccharides, and therefore are unable to form stable multicellular swarms, show increased predation. These observations may point out the impact of individual M. xanthus cells, rather than of EPS-coated swarms, for prey killing.

19Predation behavior of M. xanthus is a process of astonishing complexity: various prey are specifically recognized and killed, while the predator remains undamaged. Some prey species actively react by initiating counter-mechanisms, which implies elaborate signaling and regulatory pathways. Therefore, M. xanthus predation offers fascinating insights into the interspecies interactions that shape bacterial communities. However, addressing the ecological significance of predation in soil environments will require a more detailed understanding of the molecular mechanisms that mediate predation, namely the recognition and killing of prey. A key goal should be to identify additional predation factors that specifically kill bacteria, rather than degrade dead biomass, and to analyze their mode of action, which may also hold information on resistance formation by prey. Understanding the delivery of killing factors might point toward the mechanisms the predator uses for prey recognition and self-protection. However, dissecting the contribution of individual components might be a difficult task, considering the various different compounds that already have been described to be involved in the process, and which possibly have redundant functions. To address the complexity of predation, it might be useful to analyze the behavior of mutants on different temporal and spatial scales, which discriminate between prey killing and biomass acquisition. Finally, a detailed knowledge of the molecular mechanisms that M. xanthus uses to kill various prey species might also spur the development of novel antibacterial strategies to control pathogenic bacteria in medicine or agriculture.

20All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

21The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.