Myxococcus xanthus is a Gram-negative gliding bacterium that aggregates and develops into multicellular fruiting bodies in response to starvation. Two chemosensory systems (frz and dif), both of which are homologous to known chemotaxis proteins, were previously identified through characterization of various developmental mutants. This study aims to examine the interaction between these two systems since both of them are required for fruiting body formation of M. xanthus. Through detailed phenotypic analyses of frz and dif double mutants, we found that both frz and dif are involved in cellular reversal and social motility; however, the frz genes are epistatic in controlling cellular reversal, whereas the dif genes are epistatic in controlling social motility. The study suggests that the integration of these two chemotaxis systems may play a central role in controlling the complicated social behaviors of M. xanthus.
Fruiting body formation
Myxococcus xanthus is an unusual Gram-negative bacterium that undergoes multicellular development and primitive differentiation [1,2]. When deprived of nutrients, cells aggregate to form fruiting bodies that contain approximately 100 000 cells. With continued starvation, the aggregated cells develop into metabolically dormant myxospores [1,2]. The bacterium also exhibits social motility in which hundreds of thousands of cells move together in a large group [1–3]. It is not surprising that multiple sensory systems are involved in controlling these complicated social behaviors. Indeed, through genetic analyses, two chemosensory systems (frz and dif), both of which are homologous to known chemotaxis proteins, have been identified and are required for these social behaviors [4,5].
The frz chemosensory system was identified through characterizing a group of mutants that formed tangled frizzy filaments under fruiting conditions instead of the normal fruiting bodies . Molecular cloning and sequence analysis revealed that most of the frz gene products are homologous to chemotaxis gene products from the enteric bacteria [7–9]. For example, FrzA is homologous to CheW, FrzE is homologous to both CheA and CheY, FrzF is homologous to CheR, FrzG is homologous to CheB, and FrzCD is homologous to the methyl-accepting chemotaxis proteins (MCPs). The frz mutants have altered cellular reversal frequency which leads to defective chemotaxis [10,11]. Whereas wild-type M. xanthus cells reverse moving direction every 4–6 min, most frz mutants rarely reverse. One particular frzCD mutant (called frzD), which has a Tn5 insertion at the C-terminal portion of the frzCD gene, reverses much more frequently than the wild-type . Many frz gene products show biochemical functions similar to the chemotaxis proteins of Escherichia coli. For example, FrzCD is homologous to the C-terminal part of the MCPs of enteric bacteria, especially Tar, the receptor for aspartate in E. coli. Like MCPs in enteric bacteria, FrzCD can be methylated by S-adenosylmethionine. Further studies indicated that methylation and demethylation of FrzCD corresponds to stimulation by attractants and repellents, respectively [11–13]. Thus, modification of FrzCD has often been used as an indicator for cellular response to chemotactic signals [11,14].
The dif locus was initially identified through the characterization of a mutant defective in fruiting body formation . Molecular cloning, DNA sequencing and sequence analyses indicate that the dif locus encodes a new set of chemosensory genes homologous with MCPs, CheW, CheY and CheA. Genetic analyses and phenotypic characterization indicate that the M. xanthus dif locus is required for social (S) motility and production of extracellular matrix called fibrils [5,15].
Since both the frz and dif genes are required for fruiting body formation, they may interact with each other. This study aims to investigate these possible interactions through double mutation analyses.
2 Materials and methods
The bacterial strains used in this study are listed in Table 1. M. xanthus was grown and maintained at 32°C in CYE medium . Other media used in this study include MOPS medium (10 mM MOPS, pH 7.6; 8 mM MgSO4), and CF medium . Myxophage Mx4 was used for generalized transduction  to construct the strains listed in Table 1. For fruiting body formation, cells at about 5×108 cells ml−1 were placed on MOPS or CF plates (1.5% agar) and incubated at 32°C for 2–3 days. Colony edge morphology and cell motility were assayed using video microscopy as described by Shi and Zusman . The assays for FrzCD methylation were performed according to the methods described previously . For the auto-agglutination assay, M. xanthus was grown and analyzed in CYE. As described by Shimkets  and Wu and Kaiser , agglutination was determined continuously at room temperature by a change with the optical density (OD) measured at 600 nm with a Shimadzu BioSpec-1601 spectrophotometer.
In contrast to wild-type M. xanthus, which forms fruiting bodies in response to starvation conditions (Fig. 1A), the frzE mutant formed ‘frizzy’ filaments when starved (Fig. 1B). The difE mutant, known to be defective in social motility, also failed to form fruiting bodies, but did not form ‘frizzy’ filaments (Fig. 1C). Because the formation of the ‘frizzy’ filaments is thought to be related to abnormal FrzCD modifications in response to various environmental stimuli and altered cellular reversal frequency [12,22], we first tested the FrzCD methylation response of the wild-type and the frzE and difE mutants to various stimuli using Western blotting analyses. The wild-type and difE mutant methylated FrzCD in the presence of CYE (Fig. 2, lanes 1 and 9) and in the late phase of fruiting body formation (Fig. 2, lanes 3 and 11). In response to isoamyl alcohol, these strains demethylated FrzCD (Fig. 2, lanes 4 and 12). In contrast, the frzE mutant showed no response to these stimuli (Fig. 2, lanes 5–8). These results suggest that, although the frzE mutation abrogates the FrzCD modification response, the difE mutation has a minimal effect of the frz sensory transduction system.
Fruiting body formation and morphology of colony edges. Fruiting body formation (top panel) was examined and photographed under light microscopy with a 4× objective lens. The edges of colonies (bottom panel) were photographed under phase contrast microscopy with a 40× objective lens. A and E, DK1622 (wild-type); B and F, SW600 (frzE); C and G, SW501 (difE); D and H, SW604 (frzE difE). SW602, SW603 and SW605 exhibited similar phenotypes as SW604 (data not shown).
FrzCD modification in response to various stimuli. FrzCD methylation was analyzed with Western blotting using anti-FrzCD antibodies as previously described [4,20]. The lower bands are methylated FrzCD. Lanes 1, 5, and 9: DK1622, SW600, and SW501 cells, respectively, treated with fresh CYE medium for 1 h; lanes 2, 6, and 10: DK1622, SW600, and SW501 cells, respectively, treated with MOPS medium for 1 h; lanes 3, 7, and 11: DK1622, SW600, and SW501 cells, respectively, treated with MOPS medium for 24 h; lanes 4, 8, and 12: DK1622, SW600, and SW501 cells, respectively, treated with 0.1% isoamyl alcohol. SW601 had similar phenotypes as SW600 and SW504 had similar phenotypes as SW501 (data not shown).
Secondly, we examined the cellular reversal frequency of the wild-type and various frz and dif mutant strains of M. xanthus (Table 2). The frzE mutant showed an inability to reverse direction, whereas the frzD mutant reversed more frequently than the wild-type. The cellular reversal frequency of the dif mutants was also affected, but not to the same degree as the frz mutants. Both the difA and difE mutants reversed their direction approximately half as frequently than did the wild-type, suggesting that the dif chemosensory system does have some effect on ‘cellular motors’.
Cellular reversal interval of various M. xanthus strains
Cellular reversal interval (min per reversal)
Cellular reversal was assayed with time-lapse video microscopy as described by Shi and Zusman . The data presented are the mean of 50 cells randomly chosen for motility analyses.
In order to analyze the effect of the frz and dif mutations on cell motility, the colony edge morphology of the wild-type and various mutants on 1.5% agar was examined. As expected, wild-type M. xanthus exhibited both individual cell motility (A motility) and cell group motility (S motility) (Fig. 1E). The difE mutant, known to lack S motility, showed only A motility (Fig. 1G). Furthermore, A−dif mutants (such as SW506 and SW610) were nonmotile (; data not shown). Similar to the wild-type strain, the frzE mutant demonstrated both A and S motility (Fig. 1F), suggesting that the frz mutation has little effect on motility. We also found that both the A−frzE and S−frzE mutants retained S and A motility, respectively (Fig. 3D–F), further confirming that the frz genes are not directly involved in either type of motility. However, it is interesting to note that the colony edges of the A−frzE strains are unique. Unlike the A−frz+ strains, which are packed with cell groups at the colony edges (Fig. 3B and C), the A−frzE mutant strains have only a limited number of small cell groups at the colony edges (Fig. 3E and F). We don't know what the cause of this interesting phenotype is as yet; however, it is possible that the frz genes play a role in directing large cell groups to move from the center of the colony (less nutrients) to the edge of the colony (more nutrients). Perhaps the failed directional movement of the A−frzE mutants caused fewer cell groups to move to the colony edge.
Analyses of A and S motility of the frz mutants with colony edge morphology. The edges of colonies were photographed under phase contrast microscopy with a 40× objective lens. A, DK1300 (sglG1); B, DK1217 (aglB1); C, DK1218 (cglB2); D, SW607 (frzE sglG1); E, SW608 (frzE aglB1); F, SW609 (frzE cglB2). SW611 had similar phenotypes as SW607 (data not shown).
We also examined the ability of various frz and dif mutants to self-agglutinate (Fig. 4). The wild-type and the frzE mutant show marked agglutination within 3 h. However, like other social motility mutants, the difE mutant did not agglutinate in suspension. This suggests that the frz mutations have minimal effect on cellular agglutination.
Cellular agglutination of various M. xanthus strains. wt: wild-type DK1622; frzE, SW600; frzE difE, SW604. The experiments were performed according to Wu et al. . 2.5×108 cells ml−1 cell suspensions were used for analyses. Absorbance at 600 nm was used as the index for cellular agglutination. SW601 had similar phenotypes as SW600, whereas the SW504, SW602, SW603 and SW605 had similar phenotypes as SW501 and SW604 (data not shown).
To further study the interactions between frz and dif, we constructed various double mutants (Table 1). The dif mutations had an epistatic effect on the frz mutations in terms of social motility and cellular agglutination. The dif frz double mutants formed neither fruiting bodies nor the ‘frizzy’ filaments characteristic of the frzE mutant (Fig. 1D). Furthermore, the frzE difE double mutant showed only A motility (Fig. 1H). Like the dif mutants, the dif frz double mutants also failed to self-agglutinate (Fig. 4). These data suggest that social motility is required for expression of the ‘frizzy’ filaments and lack of agglutination. In contrast, the frz mutation was epistatic over the dif mutation in terms of cellular reversal and FrzCD modification. All double mutants with the frzE mutation were unable to reverse direction, whereas the double mutants with the frzD reversed more frequently than the wild-type strain (Table 2). FrzCD methylation of the double mutants was similar to the frz mutant in that there was no response to the different stimuli (data not shown).
The above phenotypic analyses suggest that the frz and dif systems function independently of each other to carry out different cellular functions (directional movement and social motility), yet do have some overlapping functions in controlling cellular reversal and coordinated cell movement. The double mutation analyses indicate that the frz genes are epistatic in controlling cellular reversal and directional movement, whereas the dif genes are epistatic in controlling coordinated social motility. Based on these observations, we would like to propose that fruiting body formation in M. xanthus requires two parallel chemosensory systems defined by the dif and frz genes. The dif system is responsible for organizing cells into social groups and for coordinating cell movements within a social group. The function of the frz system is to confer chemotactic capability that directs cells into aggregation centers. These dual chemosensory systems, independent of and yet complementary to each other, constitute an integrated signaling network that enables the bacterium to perform various social behaviors.
We thank Dr. Sharon Hunt Gerardo for careful editing of the manuscript. This work is supported by National Institutes of Health Grant GM54666 to W.S. and National Institutes of Health Training Grants AI07323 and DE07296 to Z.Y.
(1995) The frz signal transduction system controls multicellular behavior in Myxococcus xanthus. In: Two-Component Signal Transduction (HochJ.A., SilhavyT., Eds.), pp. 419–430. American Society for Microbiology, Washington, DC.