Present address: Max-Planck-Institute for Terrestrial Microbiology, Karl-von-Frisch Str., 35043 Marburg, Germany.
Coupling of multicellular morphogenesis and cellular differentiation by an unusual hybrid histidine protein kinase in Myxococcus xanthus
Article first published online: 8 APR 2005
Volume 56, Issue 5, pages 1358–1372, June 2005
How to Cite
Rasmussen, A. Aa., Porter, S. L., Armitage, J. P. and Søgaard-Andersen, L. (2005), Coupling of multicellular morphogenesis and cellular differentiation by an unusual hybrid histidine protein kinase in Myxococcus xanthus. Molecular Microbiology, 56: 1358–1372. doi: 10.1111/j.1365-2958.2005.04629.x
- Issue published online: 8 APR 2005
- Article first published online: 8 APR 2005
- Accepted 21 February, 2005.
We describe an unusual hybrid histidine protein kinase, which is important for spatially coupling cell aggregation and sporulation during fruiting body formation in Myxococcus xanthus. A rodK mutant makes abnormal fruiting bodies and spores develop outside the fruiting bodies. RodK is a soluble, cytoplasmic protein, which contains an N-terminal sensor domain, a histidine protein kinase domain and three receiver domains. In vitro phosphorylation assays showed that RodK possesses kinase activity. Kinase activity is essential for RodK function in vivo. RodK is present in vegetative cells and remains present until the late aggregation stage, after which the level decreases in a manner that depends on the intercellular A-signal. Genetic evidence suggests that RodK may regulate multiple temporally separated events during fruiting body formation including stimulation of early developmental gene expression, inhibition of A-signal production and inhibition of the intercellular C-signal transduction pathway. We speculate that RodK undergoes a change in activity during development, which is reflected in changes in phosphotransfer to the receiver domains.
A hallmark in many developmental processes is the spatial coupling of multicellular morphogenesis and cellular differentiation. Fruiting body formation in Myxococcus xanthus is a visible illustration of this coupling. The developmental programme that culminates in formation of the spore-filled fruiting bodies is initiated in response to starvation (Dworkin, 1996). The first morphological manifestations of fruiting body formation are evident after 6 h as cells begin to aggregate into aggregation centres. As more cells accumulate in these centres, they increase in size and become mound shaped. Eventually a mound holds 105 densely packed cells. By 24 h, the aggregation process is complete and cells inside the mounds differentiate to spores, resulting in mature fruiting bodies. Spore maturation is completed after 72–120 h. Cells that remain outside the fruiting bodies do not become spores (O’Connor and Zusman, 1991). The isolation of mutants that sporulate outside fruiting bodies (Morrison and Zusman, 1979; Cho and Zusman, 1999; Kruse et al., 2001) provide evidence that fruiting body formation per se is not a prerequisite for sporulation. These mutants also suggest that M. xanthus harbours mechanisms that allow cells to monitor their position and, thus, co-ordinate aggregation and sporulation.
Fruiting body formation is initiated by the RelA-dependent intracellular accumulation of guanosine penta- and tetraphosphate [(p)ppGpp] (Singer and Kaiser, 1995). Two intercellular signals, the A- and C-signals, have been shown to be important for fruiting body formation and have been characterized functionally and biochemically. The A-signal becomes important for development at 2 h, consists of a mixture of amino acids and peptides (Kuspa et al., 1992a) and is part of a system that monitors the density of starving cells (Kuspa et al., 1992b). If a threshold concentration is reached, fruiting body formation proceeds. The C-signal becomes important at 6 h and has a key role in co-ordinating aggregation and sporulation. The C-signal induces aggregation and sporulation at distinct thresholds (Kim and Kaiser, 1991; Li et al., 1992; Kruse et al., 2001). Moreover, the C-signal is required for full expression of genes that are turned on after 6 h (Kroos and Kaiser, 1987). The spatial co-ordination of aggregation and sporulation has been suggested to be a consequence of the contact-dependent C-signal transmission mechanism (Kim and Kaiser, 1990), which ensures that only those cells that are closely packed inside fruiting bodies reach the threshold level of C-signal (Søgaard-Andersen et al., 2003). The C-signal has been suggested to be a 17 kDa cell surface-associated protein, which is produced by proteolytic processing of the full-length, 25 kDa CsgA protein (Kim and Kaiser, 1990; Lobedanz and Søgaard-Andersen, 2003) (Fig. 1).
Genetic and biochemical data suggest that implementation of the C-signal-dependent responses involves a branched pathway (Fig. 1). One branch leads to stimulation of csgA expression (Kim and Kaiser, 1991). A second branch leads to activation of the DNA binding response regulator FruA (Ogawa et al., 1996; Ellehauge et al., 1998). Genetic evidence suggests that activation involves phosphorylation of a conserved Asp in the receiver domain of FruA (Ellehauge et al., 1998). Downstream from FruA the pathway branches again with the C-signal stimulating the Frz chemosensory signal transduction system in the aggregation branch, in a FruA-dependent manner by inducing methylation of the FrzCD protein (Søgaard-Andersen and Kaiser, 1996). In the sporulation branch, activated FruA induces C-signal-dependent gene expression including the devRS operon, which is required for sporulation (Ellehauge et al., 1998; Horiuchi et al., 2002). In addition, genetic evidence suggests that the pathway defined by the histidine protein kinase SdeK converges with the C-signalling pathway at or downstream from activation of FruA (Fig. 1). The SdeK pathway stimulates aggregation, sporulation and C-signal-dependent gene expression (Garza et al., 1998; Pollack and Singer, 2001).
Two-component signal transduction systems enable bacteria to adapt to internal and external cues (Robinson et al., 2000). Classical two-component systems consist of a sensor histidine protein kinase and a response regulator protein. In response to stimulation, the kinase autophosphorylates on a conserved His. Subsequently, the phospho-group is transferred to a conserved Asp in the receiver domain of the response regulator. The phosphorylated response regulator elicits the appropriate response. Increasingly, two-component regulatory systems consisting of four modules are being identified in bacteria and lower eukaryotes (Robinson et al., 2000). These modules may exist as discrete proteins or as domains in hybrid proteins often in the form of a histidine protein kinase, a receiver domain, a His-containing phosphotransmitter (Hpt) domain and a response regulator. In these systems, signalling is initiated by autophosphorylation of the kinase, followed by phosphotransfer to the receiver, then to the Hpt domain, and, finally, to the response regulator.
Here, we describe the structurally complex hybrid histidine protein kinase, RodK, which regulates multiple steps during fruiting body formation including the spatial coupling of aggregation and sporulation.
RodK is required for aggregation and coupling of aggregation and sporulation
To identify genes involved in fruiting body formation, we carried out a miniTn5(tet) transposon mutagenesis of the fully motile strain DK1622, and screened for mutants with developmental defects (Rasmussen and Søgaard-Andersen, 2003). DK1622-Ω8153 displayed abnormal aggregation (data not shown). Reintroduction of miniTn5(tet)Ω8153 into DK1622 by generalized transduction verified that the developmental defects in DK1622-Ω8153 were caused by the insertion (data not shown).
The region flanking miniTn5(tet)Ω8153 contains several open reading frames (ORFs) recognized on the basis of a high GC content in the third position of codons typical of GC-rich organisms (Bibb et al., 1984) (Fig. 2A). miniTn5(tet)Ω8153 is inserted in an ORF designated rodK (for regulator of differentiation kinase). The deduced rodK protein consists of 981 amino acids, and contains four domains that show significant similarity to domains in two-component regulatory systems (Fig. 3A and B). The central part of the protein contains a histidine protein kinase domain that includes the conserved H-box, with the potential site of autophosphorylation (H360), as well as the conserved N-, D/F- and G-boxes (Robinson et al., 2000). The kinase domain is followed by three receiver domains (Fig. 3A and C). Receiver domains are characterized by several conserved residues (Robinson et al., 2000), which are all present in the three receiver domains. The only exception is the third receiver, which contains Glu instead of Asp at the position corresponding to position 12 in Escherichia coli CheY (Fig. 3C). The 350 amino acid N-terminus in RodK has no similarity to any entry in the database; moreover, a smart analysis (Schultz et al., 1998) did not reveal conserved domains or motifs with a significant score. tmhmm2 (Krogh et al., 2001) did not identify trans-membrane spanning regions, suggesting that RodK is localized to the cytoplasm.
To determine whether the loss of RodK function gives rise to developmental defects, two additional rodK mutants were constructed. The rodK::pAAR328 allele was constructed by integration of plasmid pAAR328, which contains an internal rodK fragment (Fig. 2B), in rodK by homologous recombination. Integration of pAAR328 disrupts rodK upstream of the putative site of autophosphorylation. The second mutation is a markerless in-frame deletion from position 307–2728 in rodK corresponding to codon 103–910. This deletion removes all four putatively phosphorylatable amino acid residues in RodK (Fig. 2B). To analyse the effect of the two rodK mutations on development, the two rodK strains were exposed to starvation in parallel with the wild-type (WT) strain. On CF starvation agar DK1622 WT cells had formed fruiting bodies at 24 h (Fig. 4A). The two rodK mutants had similar phenotypes and both formed abnormally shaped fruiting bodies at 48 h. At 120 h, these fruiting bodies were smaller, more irregular and more translucent than those formed by WT cells (Fig. 4A). After 72 h on CF agar, the two rodK mutants sporulated at a twofold higher level than the WT (Table 1) but after 120 h, the sporulation level was similar to that of the WT. Many of the spores formed by the two rodK strains were located outside the abnormally shaped fruiting bodies (Fig. 4A). Under the more stringent starvation conditions of submerged culture, the two rodK mutants were unable to form fruiting bodies (Fig. 4B) and after 120 h the level of sporulation was 10-fold lower than in WT (Table 1). The spores formed by the rodK mutants in submerged culture were formed in the absence of aggregation (Fig. 4B).
|CF agarb||Submerged culturec|
|48 h||72 h||120 h||168 h||120 h|
|DK1622||Wild type||<0.001||45 ± 15||85 ± 23||100 ± 9||100 ± 22|
|SA1706||rodK::pAAR328||<0.001||80 ± 9||110 ± 9||105 ± 15||7.1 ± 0.8|
|SA1708||ΔrodK||<0.001||105 ± 15||120 ± 15||115 ± 23||9.5 ± 1.7|
|SA1741||ΔrodK, rodK::pAAR359||ND||ND||ND||ND||105 ± 33|
|SA1742||ΔrodK, rodK::pAAR360||ND||ND||ND||ND||10.5 ± 4.4|
To demonstrate that the developmental defects in the ΔrodK strain were caused by loss of RodK function, the ΔrodK strain was complemented with a rodK+ allele. pAAR359, which contains the rodK+ allele (Fig. 2B), was integrated by a single homologous recombination event upstream of the deletion. The developmental defects in the ΔrodK strain were corrected by pAAR359 with respect to aggregation and sporulation, and under both starvation conditions (Fig. 4 and Table 1). These observations provide evidence that the defects in the ΔrodK strain are caused by loss of RodK function. Two lines of evidence suggest that the defects in the rodK::pAAR328 strain (SA1706) are also caused by lack of RodK rather than by polar effects on downstream genes. First, the ΔrodK (SA1708) and rodK::pAAR328 strains (SA1706) had similar developmental defects. Second, integration of the plasmid pAAR360 (which encodes the rodKH360A allele, cf. below) upstream of the ΔrodK allele did not cause additional developmental defects. From these experiments, we conclude that loss of RodK function causes an aggregation defect, a conditional sporulation defect, and uncouples these two events.
A rodK mutation does not affect vegetative motility or growth
Myxococcus xanthus harbours two gliding systems, the social (S)-system and adventurous (A)-system (Hodgkin and Kaiser, 1979). To determine whether a rodK mutation affects gliding motility in vegetative cells, the rodK::pAAR328 allele was introduced into an A–S+ mutant strain (DK1217) and an A+S– mutant strain (DK1300) resulting in the two strains SA1717 and SA1718 respectively. Subsequently, the effect of rodK::pAAR328 mutation on gliding motility was assessed. Strains that carry an A– mutation as well as an S– mutation are non-spreading and grow as small, smooth-edged colonies (Hodgkin and Kaiser, 1979). SA1717 and SA1718 displayed a motility behaviour that was indistinguishable from that observed in the parental strains (DK1217 and DK1300) (data not shown). These data suggest that the developmental defects in a rodK mutant are not secondary to an effect on gliding motility.
The rodK strains were indistinguishable from WT with respect to growth rate in rich Casitone medium and in the chemically defined A1 medium, with respect to pigmentation, and glycerol-induced sporulation (data not shown). To assess whether loss of RodK function causes an alteration in recognition of starvation, the two rodK strains and the WT strain were exposed to starvation in submerged culture in the presence of different concentrations of Casitone. Casitone consists of a mixture of peptides, which serves as a carbon source for M. xanthus. The three strains all displayed growth in the presence of 0.5% Casitone. In the presence of 0.25% Casitone, the three strains initiated development; the WT strain formed spore-filled fruiting bodies while the rodK mutants sporulated. These observations indicate that loss of RodK function does not interfere with recognition of starvation.
To test whether the sporulation defect caused by loss of RodK function is cell-autonomous, the rodK::pAAR328 strain was co-developed with the WT strain in submerged culture. Under these conditions, the level of sporulation in the rodK strain was not raised (data not shown). Thus, the sporulation defect caused by loss of RodK function is cell-autonomous, i.e. it cannot be complemented by extracellular secretions from WT cells.
RodK is a histidine protein kinase and kinase activity is essential for RodK function in vivo
To verify that RodK is a histidine protein kinase and is able to autophosphorylate using ATP as phospho-donor, two different RodK proteins were purified. His6-RodKD657N,D782N,D909N has an N-terminal His6-tag and the three potentially phosphorylated Asp residues in the receiver domains have been replaced with Asn residues to avoid possible intramolecular phosphotransfer reactions. His6-RodKH360A,D657N,D782N,D909N is identical to His6-RodKD657N,D782N,D909N except that the potential site of autophosphorylation, H360, has been substituted with an Ala. The two proteins were assayed for kinase activity in vitro using [γ-32P]-ATP as phospho-donor. Figure 5 shows that His6-RodKD657N,D782N,D909 was phosphorylated in the presence of [γ-32P]-ATP whereas phosphorylation of His6-RodKH360A,D657N,D782N,D909N was not detected.
The chemical stability of the phospho-linkage in His6-RodKD657N,D782N,D909 was examined under acidic and alkaline conditions to determine whether the protein was phosphorylated on a His. Incubation of the phosphorylated form of His6-RodKD657N,D782N,D909 under acidic conditions resulted in a complete loss of the phosphoryl group while incubation under alkaline conditions only reduced labelling by 10% (data not shown). The acid labile but alkaline stable phospho-linkage in His6-RodKD657N,D782N,D909 is characteristic of phosphoramidates (Duclos et al., 1991). These experiments show that RodK is a histidine protein kinase and that H360 is the site of autophosphorylation.
To test the importance of H360 in RodK in vivo, a rodK allele encoding a RodK protein in which H360 has been replaced with an Ala was constructed. The plasmid pAAR360, which carries this rodK allele, was inserted upstream of the ΔrodK allele in SA1708 (Fig. 2). The rodKH360A allele did not correct the developmental defects caused by the ΔrodK mutation (Fig. 4A and B; Table 1). Immunoblots of total-cell lysates from WT and ΔrodK/pAAR360 cells developed in submerged culture using anti-RodK antibodies showed that the accumulation profile of the two proteins was indistinguishable (data not shown). These experiments suggest that the histidine protein kinase activity of RodK is essential for RodK activity in vivo.
RodK accumulation decreases during development in response to the A-signal
To understand how RodK is regulated, the accumulation of RodK was examined during development using anti-RodK antibodies. In vegetative WT cells, the antibodies recognize a protein with a size similar to the His6-RodK protein, which has a calculated MW of 108.5 kDa (Fig. 6; data not shown). From the assigned start codon, the calculated MW of native RodK is 107.7 kDa. This protein is absent in a ΔrodK strain (Fig. 6). These observations suggest that the assigned start codon in rodK is correct and that the anti-RodK antibodies are specific to RodK. To analyse accumulation of RodK during development, cell lysates of exponentially growing or starving WT cells were examined in immunoblots (Fig. 6A). RodK was present in exponentially growing cells and the level remained constant until 12 h. After 12 h, the amount of RodK declined until it was no longer detectable at 24 h.
To test whether RodK accumulation is regulated by the A- or C-signals, immunoblots were carried out on cell lysates from starving asgA and csgA cells (Fig. 6B). The asgA mutant is strongly reduced in A-signal production (Kuspa and Kaiser, 1989), and the csgA mutant is unable to synthesize C-signal. In the asgA mutant, the level of RodK did not decrease after 12 h but remained constantly high for at least 48 h. Similar results were obtained with an asgB mutant (data not shown). In the csgA mutant, the RodK accumulation profile was similar to that observed in WT (Fig. 6A). These results provide evidence that the decrease in RodK accumulation late in development is regulated by the A-signal whereas it is not regulated by the C-signal.
A rodK mutation affects developmental gene expression
To determine the time of action of RodK during development and to clarify whether a rodK mutant is affected in the stringent response, A-signalling or C-signalling, the expression of three developmentally regulated lacZ reporter fusions was studied in WT and rodK::pAAR328 cells. In WT cells each reporter fusion increases its expression at a particular time point during starvation. Moreover, expression of these fusions depends in different combinations on (p)ppGpp and the intercellular A- and C-signals. As the developmental defects in a rodK mutant are most severe when cells are starved in submerged culture, isogenic rodK+ and rodK strains each containing a single lacZ fusion were starved under these conditions and the specific activity of β-galactosidase expressed from the fusions was measured (Fig. 7A). Tn5lacΩ4408 is inserted in the sdeK gene and is induced by (p)ppGpp accumulation early in development (Singer and Kaiser, 1995). In vegetative cells, the expression level of Tn5lacΩ4408 was identical in the two strains. However, during starvation the expression level in the rodK strain was twofold lower than in WT. Tn5lacΩ4521 is induced at approximately 2 h and depends on (p)ppGpp and A-signal (Kuspa et al., 1986; Singer and Kaiser, 1995). Tn5lacΩ4521 was induced with the same timing in the two strains. However, the level of expression in the rodK strain was two- to threefold higher than in WT. Tn5lacΩ4414 is inserted in the devR gene and has a polar effect on the downstream devS gene (Thöny-Meyer and Kaiser, 1993). Tn5lacΩ4414 is turned on after 6 h in an A- and C-signal-dependent manner (Kuspa et al., 1986; Kroos and Kaiser, 1987). The rodK mutation caused a twofold increase in the expression of this fusion without altering the timing of induction. In summary, none of the fusions were expressed at higher levels in the rodK mutant in vegetative cells suggesting that the rodK mutation does not cause an inappropriate induction of developmental gene expression in vegetative cells. Moreover, RodK is required for full induction of the sdeK gene early during development, inhibits expression of an A-signal-dependent fusion and inhibits expression of a C-signal-dependent fusion.
A rodK mutant produces increased levels of A-signal
The increased expression of the A-signal-dependent Tn5lacΩ4521 fusion leads us to examine A-signal synthesis in a rodK mutant. An asgB mutant synthesizes 5–10% of the normal level of A-signal (Kuspa et al., 1992a). To test A-signal synthesis in a rodK mutant, we carried out co-development experiments in which the asgB mutant DK9009, which also carries Tn5lacΩ4521, was co-developed with ΔrodK or WT cells in submerged culture. Subsequently, the level of β-galactosidase activity expressed from Tn5lacΩ4521 was analysed. In these experiments, cells co-developed with the asgB cells produce the bulk of the A-signal. The level of β-galactosidase activity expressed from DK9009 was similar in the two co-development experiments until 6 h (Fig. 8). However, from 6 to 24 h the level of β-galactosidase activity was 1.5- to 2.0-fold higher in the presence of ΔrodK cells than in the presence of WT cells. At all time points, the levels of β-galactosidase activity in the co-development experiments were significantly higher than that observed when DK9009 was developed in the absence of other cells. This result suggests that loss of RodK function results in increased A-signal synthesis.
A rodK mutation alters the FrzCD methylation pattern
The increased expression of the C-signal-dependent Tn5lacΩ4414 fusion prompted us to examine the accumulation of proteins in the C-signal transduction pathway in a rodK mutant (Fig. 1). Immunoblots were made on cell lysates from WT and ΔrodK cells starved in submerged culture and reacted with anti-CsgA, anti-FruA or anti-FrzCD antibodies respectively (Fig. 7B). The two CsgA proteins accumulated with similar profiles in the two strains. FruA also accumulated with the same profile in the two strains. In response to starvation, FrzCD initially demethylates and then re-methylates (McBride and Zusman, 1993). FrzCD methylation after 6 h is C-signal dependent (Søgaard-Andersen and Kaiser, 1996). The FrzCD methylation pattern was similar in the two strains until 3 h. However, at 6 h FrzCD was almost solely found in the methylated form in rodK cells whereas WT still contained a considerable amount of the non-methylated form. The FrzCD methylation pattern indicates that the C-signal transduction pathway is stimulated as a consequence of the rodK mutation.
A rodK mutation partially alleviates the requirement for csgA
To further analyse how RodK interacts with the C-signal transduction pathway, the rodK::pAAR328 allele was introduced into csgA, fruA, devRS and sdeK single mutants creating four double mutants. Intriguingly, the csgA, rodK double mutant formed normal fruiting bodies on CF agar and sporulated at WT levels in submerged culture (Fig. 9). However, the double mutant was still unable to form fruiting bodies in submerged culture (data not shown). The fruA, rodK double mutant had developmental defects similar to the fruA mutant on CF agar and in submerged culture and was unable to aggregate and sporulate (Fig. 9). The devRS, rodK double mutant had a more severe aggregation defect than the two single mutants whereas the level of sporulation was similar to that in the two single mutants (Fig. 9). The sdeK, rodK double mutant had a more severe aggregation defect than the two single mutants on CF agar and sporulated at the same low level as the sdeK mutant (Fig. 9). Thus, the absence of RodK partially compensates for a C-signal deficiency. However, an absence of RodK does not compensate FruA, DevRS or SdeK deficiencies.
Here we report the identification of an unusual hybrid histidine protein kinase, RodK, which is required for normal fruiting body development and the spatial coupling of aggregation and sporulation in M. xanthus.
In its central part RodK has similarity to histidine protein kinases. In vitro phosphorylation experiments biochemically confirmed that RodK has kinase activity, and that phosphorylation occurs on H360. Moreover, a RodK protein carrying an H360A substitution was unable to complement a ΔrodK mutant in vivo. These data provide evidence that RodK is a histidine protein kinase and that the kinase activity is important for normal fruiting body formation. In its N-terminus, RodK contains an unusual 350-amino-acid sensor domain. Sequence analysis of the sensor domain did not identify domains or motifs that could provide a clue to the input signal that regulates RodK activity. RodK is likely to be located to the cytoplasm, as sequence analyses did not identify trans-membrane spanning regions.
In its C-terminal part, RodK contains three receiver domains and can therefore be classified as a hybrid histidine protein kinase. The domain architecture of RodK is highly unusual. Among 3162 histidine protein kinase homologues identified in the databases, 747 have a kinase/receiver, 63 a kinase/receiver/receiver and only 11 a kinase/receiver/receiver/receiver architecture. The latter 11 hybrid kinases have not been characterized functionally or biochemically. The role of the three receiver domains could be to receive signals from or transmit signals to other two component regulatory systems. However, the observation that the RodK protein carrying the H360A substitution gives rise to a phenotype, which is similar to that of a ΔrodK strain, seems to argue against this possibility and suggests that the phosphoryl group on H360 in RodK is the major phosphodonor for the three receiver domains. Intramolecular phosphotransfer in hybrid kinases has been reported, e.g. for BvgS (Uhl and Miller, 1996) and ArcB (Georgellis et al., 1998). However, in these hybrid kinases the intramolecular phosphotransfer from the kinase domain only involves transfer to a single receiver domain. Thus, the unusual domain architecture in RodK may be paralleled by unusual intramolecular phosphotransfer reactions. Experiments are in progress to analyse in detail the intramolecular phosphotransfer in RodK.
RodK is present in vegetative cells and remains detectable in starving cells for at least 18 h. We have been unable to detect defects in rodK cells during vegetative growth. This suggests that RodK does not regulate initiation of fruiting body formation but rather regulates steps in fruiting body formation. The decrease in RodK accumulation during development is the result of a regulated event as a mutant unable to synthesize the A-signal continues to accumulate RodK. The decrease in RodK accumulation was not evident until 18 h of starvation suggesting that RodK accumulation is not directly regulated by the A-signal but by a factor in the A-signal regulatory cascade. Whether this factor regulates RodK accumulation at the level of synthesis or at the level of degradation remains to be elucidated.
The complex architecture of RodK is paralleled in its complex effects on developmental events. The negative effect of a rodK mutation on sdeK expression suggests that RodK normally stimulates sdeK expression. RodK stimulating the stringent response could result in this effect. However, three lines of evidence argue against this mechanism. First, a rodK mutation does not interfere with recognition of starvation. Second, a relA mutation results in decreased A-signal production (Harris et al., 1998) whereas a rodK mutation results in increased A-signal production. Third, a relA mutation results in decreased csgA expression (Crawford and Shimkets, 2000) whereas a rodK mutation does not interfere with accumulation of the csgA proteins. On the basis of these arguments, we suggest that RodK stimulates sdeK expression and that this stimulation is independent of the stringent response.
A rodK mutant displays increased A-signal production. The observation that an sdeK mutation does not result in increased A-signal production (Pollack and Singer, 2001) argues that RodK, independently of its stimulatory effect on sdeK expression, inhibits A-signal synthesis. The increased A-signal production in a rodK mutant could explain the increased level of transcription of the A-signal-dependent Tn5lacΩ4521 fusion in a rodK mutant. To our knowledge RodK is the first protein to be identified that inhibits A-signal production.
The A- and C-signalling systems are connected by two observations: C-signal-dependent genes depend on the A-signal for full expression (Kroos and Kaiser, 1987). Moreover, the A-signal is required for transcription of fruA (Ogawa et al., 1996; Ellehauge et al., 1998). The increased A-signal production in a rodK mutant could potentially result in increased expression of Tn5lacΩ4414 and early methylation of FrzCD if it resulted in increased accumulation of the FruA protein. However, FruA accumulation is normal in a rodK mutant. In addition, as an sdeK mutation results in decreased expression of the C-signal-dependent Tn5lacΩ4414 fusion (Pollack and Singer, 2001), the effect of a rodK mutation on sdeK expression cannot explain the increased expression of this fusion in a rodK mutant. On the basis of these arguments, we suggest that a rodK mutation, independently of its negative effect on sdeK expression and its stimulatory effect on A-signal production, interferes with the normal function of the C-signalling system. The increased transcription of the C-signal-dependent Tn5lacΩ4414 fusion and the accelerated methylation of the FrzCD protein in a rodK mutant suggest that RodK normally inhibits the C-signal transduction pathway upstream from these two events. The site of RodK inhibition is downstream from the branchpoint in the C-signal transduction pathway that leads to accumulation of the two csgA proteins because accumulation of these two proteins is normal in a rodK mutant. These observations suggest that the inhibitory effect of RodK on the C-signal transduction pathway is at or upstream from activation of FruA. In summary, we suggest that RodK has three distinct functions during fruiting body formation: stimulation of sdeK expression, inhibition of A-signal production, and inhibition of the C-signal transduction pathway.
A hallmark in fruiting body formation is the spatial coupling of aggregation and sporulation. espA (Cho and Zusman, 1999) and frz (Morrison and Zusman, 1979) mutations by unknown mechanisms result in sporulation outside fruiting bodies. In addition, overexpression of the C-signal (Kruse et al., 2001) results in sporulation outside fruiting bodies presumably by allowing cells to reach the threshold level of C-signalling that induces sporulation before they have accumulated inside fruiting bodies. In the case of rodK mutants, we suggest that the mechanism underlying uncoupling of aggregation and sporulation is the lack of inhibition of the C-signal transduction pathway. The absence of this inhibition would allow cells to reach the threshold level of C-signalling that induces sporulation even though they have not accumulated at the high cell density inside fruiting bodies. The aggregation defect in rodK mutants could also be a consequence of the increased activity in the C-signal transduction pathway, as this may induce the cellular changes involved in sporulation prematurely and in that way interfere with C-signal transmission and aggregation. The conditional sporulation defect in rodK mutants, with sporulation levels being normal on CF agar and 10-fold reduced in submerged culture, suggests that the primary defect in rodK mutants is the lack of co-ordination of aggregation and sporulation rather than sporulation per se.
The notion that RodK inhibits the C-signal transduction pathway at or upstream from activation of FruA is supported by the observations that a rodK mutation partially compensates a csgA mutation whereas a rodK mutation does not compensate fruA, devRS or sdeK mutations. Presumably, in the csgA, rodK double mutant, the absence of the stimulating input from the C-signal to the pathway is compensated for by the absence of inhibition by RodK. In the fruA, rodK, devRS, rodK and devRS, rodK double mutants this compensation would not be possible due to the lack of FruA, DevRS and SdeK.
How may RodK regulate three temporally separated events during fruiting body formation? To date only hybrid histidine protein kinases with a single receiver domain have been shown to be part of a phosphorelay system (Uhl and Miller, 1996; Georgellis et al., 1998; Takeda et al., 2001). The role of multiple receiver domains in a hybrid kinase is unknown. Activated response regulators elicit the appropriate response by altering gene expression or by engaging in protein–protein interactions (Robinson et al., 2000). RodK is unlikely to directly regulate gene expression suggesting that it acts by engaging in protein–protein interactions. The kinase activity of RodK may well change during development, in response to signals sensed through the N-terminal sensory domain. The RodK kinase domain may have differential affinities for the three receiver domains and, thus, phosphorylate the three receivers in a pattern that depends on input activity. If this were the case, different patterns of phosphorylated receivers would be established during development. If the protein–protein interactions that RodK engages in depend on the phosphorylation pattern of the receiver domains, these interactions would change during development. Alternatively, the three receiver domains may act as phosphate sinks and modulate the rate of signalling through RodK to other response regulators in response to input activity.
In addition to SdeK (Pollack and Singer, 2001) and TodK (Rasmussen and Søgaard-Andersen, 2003), RodK is the third cytoplasmic histidine protein kinase for which genetic evidence suggests that it defines a signalling pathway that converges with the C-signal transduction pathway to regulate aggregation and sporulation. The SdeK pathway stimulates aggregation and sporulation while the TodK and RodK pathways have inhibitory effects. It has been suggested that SdeK and TodK sense signals related to the metabolic state of individual cells (Rasmussen and Søgaard-Andersen, 2003). Thus, the emerging picture of the design of the regulatory network that governs aggregation and sporulation has two features. First, the activity of signal transduction pathways that are modulated by intercellular and intracellular signals are integrated. Second, positively and negatively acting pathways are integrated to determine the overall outcome. This design is analogous to that of the phosphorelay that governs initiation of sporulation in Bacillus subtilis (Perego, 1998). The fundamental difference between the two networks is the time point at which signal integration occur. In B. subtilis signal integration influences the initiation of the developmental programme that results in sporulation whereas in M. xanthus signal integration occurs during fruiting body formation and influences the timing and spatial coupling of aggregation and sporulation.
Cell growth, development, motility assays and measurements of β-galactosidase activity
Escherichia coli strains were grown in LB broth in the presence of relevant antibiotics (Sambrook et al., 1989). M. xanthus cells were grown in CTT medium in liquid cultures or on CTT agar plates (Hodgkin and Kaiser, 1979). Kanamycin or oxytetracycline was used for selective growth at concentrations of 40 µg ml−1 and 10 µg ml−1 respectively. Aggregation was monitored on CF agar (Shimkets and Kaiser, 1982) or in submerged culture (Kuner and Kaiser, 1982). Aggregation was followed visually using a Leica MZ8 stereomicroscope or a Leica IMB/E inverted microscope. Cells were photographed using a Sony 3CCD camera. Spore titres were determined as the number of sonication- and heat-resistant colony-forming units (Søgaard-Andersen et al., 1996). Activities of β-galactosidase were quantified as described (Kroos et al., 1986). Glycerol-induced sporulation was tested as described (Licking et al., 2000). Strains to be tested for motility were grown in CTT to a density of 5 × 108 cells ml−1, harvested and resuspended in MC7 (10 mM MOPS pH 7.0, 1 mM CaCl2) to a calculated density of 5 × 09 cells ml−1. Cell suspension (5 µl) was spotted on a thin layer of 0.5% or 1.5% agar supplemented with 0.5% CTT prepared on a sterile microscope slide. The colony edge morphology was inspected after 24 h (Hodgkin and Kaiser, 1979; Shi and Zusman, 1993).
Myxococcus xanthus and E. coli strains used in this work are listed in Table 2. Plasmids were propagated in E. coli TOP10 unless otherwise stated. DK1622 was used as the WT strain throughout. All strains constructed were confirmed either by Southern blot analysis or by polymerase chain reaction (PCR). DK9007 was constructed by generalized transduction using Mx8clp2ts3 (Martin et al., 1978) propagated on DK4324 to infect DK1622. DK9009 was constructed by generalized transduction using Mx8clp2ts3 propagated on DK9521 to infect DK9007. SA1706, SA1717, SA1718, SA1730, SA1732, SA1734, SA1739 and SA1740 were constructed by integration of pAAR328 in rodK+ in DK1622, DK1217, DK1300, SA1000, SA1465, DK7809, LS202 and DK7862 respectively. SA1708 carries a markerless rodK in-frame deletion generated by integration of pAAR348 in rodK in DK1622 followed by excision of the WT gene as described (Julien et al., 2000). SA1705 carries pAAR300 integrated in orf10. SA1741 was constructed by homologous integration of pAAR359 upstream of ΔrodK in SA1708. SA1742 was constructed by homologous integration of pAAR360 upstream of ΔrodK in SA1708.
|Strain or plasmid||Relevant characteristic(s)||Source or reference|
|JM109||F′traD36 proA+B+lacI qΔlacZM15/Δ(lac-proAB) glnV44 e14 –gyrA96 recA1 relA1 endA1 thi hsdR17||New England Biolabs|
|TOP10||F–mcrAΔ(mrr-hsdRMS-mcrBC) φ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(ara-leu)7679 galU galK rpsL (StrR) endA1 nupG||Invitrogen|
|DK1622||Wild type||Kaiser (1979)|
|DK1217||aglB1||Hodgkin and Kaiser (1979)|
|DK1300||sglG1||Hodgkin and Kaiser (1979)|
|DK4324||Tn5lacΩ4521, sglA1, asgB480||Kuspa et al. (1986)|
|DK4398||asgB480 Tn5lacΩ4411||Kuspa et al. (1986)|
|DK5057||asgA476 Tn5Ω4560||Kuspa and Kaiser (1989)|
|DK7809||Tn5lacΩ4521(tet)||Kruse et al. (2001)|
|DK7862||sdeK::Tn5lacΩ4408(tet)||Y. Cheng and D. Kaiser|
|DK9009||Tn5lacΩ4521, asgB480 Tn5-132Ω4679||This work|
|DK9521||asgB480 Tn5-132Ω4679||I. Keseler and D. Kaiser|
|LS202||csg741||Shimkets and Asher (1988)|
|SA1000||devR::Tn5lacΩ4414(tet)||Ellehauge et al. (1998)|
|SA1465||fruA::tet||Kruse et al. (2001)|
|SA1717||aglB1, rodK::pAAR328||This work|
|SA1718||sglG1, rodK::pAAR328||This work|
|SA1730||devR::Tn5lacΩ4414(tet), rodK::pAAR328||This work|
|SA1732||fruA::tet, rodK::pAAR328||This work|
|SA1734||Tn5lacΩ4521(tet), rodK::pAAR328||This work|
|SA1739||csg741, rodK::pAAR328||This work|
|SA1740||sdeK::Tn5lacΩ4408(tet), rodK::pAAR328||This work|
|SA1741||ΔrodK, rodK::pAAR359||This work|
|SA1742||ΔrodK, rodK::pAAR360||This work|
|pBGS18||Cloning vector, KmR||Spratt et al. (1986)|
|pUHE24-2||Expression vector, ApR||Lutz and Bujard (1997)|
|pBJ113||Vector for constructing gene replacements, KmR and galK +||Julien et al. (2000)|
|pAAR300||706 bp (pos. 6394–7100)aBamHI–HindIII fragment in pBGS18||This work|
|pAAR306||9846 bp (pos. −2746–7100)aBamHI–HindIII fragment in pBGS18||This work|
|pAAR327||pUHE24-2 derivative for expression of His6-RodKWT||This work|
|pAAR328||659 bp (pos. 318–977)aHindIII–HindIII fragment in pBGS18||This work|
|pAAR348||969 bp [pos. −208–3143 (pos. 307–2728 deleted)]EcoRI–HindIIIb in pBJ113||This work|
|pAAR353||pUHE24-2 derivative for expression of His6-RodKD657N,D782N,D909N||This work|
|pAAR355||pUHE24-2 derivative for expression of His6-RodKH360A,657N,D782N,D909N||This work|
|pAAR359||3351 bp (pos. −208–3143)aKpnI–HincII fragment encoding rodK + in pBGS18||This work|
|pAAR360||3351 bp (pos. −208–3143)aKpnI–HincII fragment encoding rodKH360A allele in pBGS18||This work|
Plasmids used to generate mutant rodK alleles or for protein purification are listed in Table 2. PCR fragments used for cloning were obtained using Pfx polymerase (Invitrogen). DNA sequences of all fragments obtained by PCR were verified by DNA sequencing. pAAR328 was constructed by cloning a HindIII-digested PCR product extending from position 318–977 in rodK in the HindIII site of pBGS18 (Spratt et al., 1986). The PCR fragment was generated using M. xanthus chromosomal DNA as template, and the primers 5′-atcggAAGCTTcgccggagcgctcgactac (bases in capital provide a HindIII site) and 5′-atcggAAGCTTgccggtccaggtc ctcgttc (bases in capital provide a HindIII site). pAAR306 was generated by in situ cloning. First, pAAR300 was integrated in orf10 (Fig. 1A). pAAR300 contains a fragment that extends from position 6394–7100 in orf10 (Fig. 1A). This fragment was generated by PCR using M. xanthus chromosomal DNA as template, and the primers 5′-gcagctGGATCCgcctgacgc ccaggcacc (bases in capital provide a BamHI site) and 5′-atcggAAGCTTcaggtcttccgggaggccct (bases in capital provide a HindIII site). The PCR fragment was digested with BamHI and HindIII and cloned in the same sites in pBGS18. pAAR300 was electroporated into DK1622 and integrated in orf10. Chromosomal DNA from the resulting strain, SA1705, was extracted and digested with BamHI. The BamHI-digested chromosomal DNA was ligated and introduced by electroporation into E. coli strain TOP10 followed by selection for kanamycin resistance. The resulting plasmid, pAAR306, contains 9846 bp of the rod locus starting from the BamHI site at position −2746 (Fig. 1A). For pAAR325 construction, a WT rodK allele on a 3351 bp KpnI–HincII fragment obtained from pAAR306 was cloned in the same sites in pUC18 (Yanisch-Perron et al., 1985). pAAR342 contains a PCR fragment, which was generated using pAAR325 as template, and the primers 5′-cgGGATCCctgtcgcttccccgcctgg (bases in capital provide a BamHI site) and 5′-cgGGATC Cctccagcgtggcctcgcac (bases in capital provide a BamHI site). The two primers anneal to the 5′- and 3′-ends of rodK, respectively, and ‘point away from rodK’. Therefore, the PCR product generated contains the pUC18 vector sequence and the 5′- and 3′-ends of rodK. The PCR product was digested with BamHI (sites provided by the primers), ligated and transformed into TOP10. The plasmid generated, pAAR342, contains an in-frame deletion in the rodK reading frame. pAAR348 (Fig. 1B) was constructed by cloning a 969 bp EcoRI–HindIII fragment containing the rodK deletion allele from pAAR342 into pBJ113 (Julien et al., 2000) digested with the same enzymes. pAAR359 (Fig. 1B) was constructed by cloning a SalI–EcoRI fragment containing the WT rodK allele from pAAR325 in the same sites in pBGS18. pAAR334 was constructed in a two-step procedure. First, a PCR product was generated using pAAR325 as template, and the primers o8153(H360A) (5′-gcggatttcgGCgctgaagttgg (bases in capital generate the H360A mutation) and o8153-1 (5′-gacatcac cgagcgcaagcatc). Second, a PCR product was generated using pAAR325 as template, and the primers o8153-3 (5′-ctggtggaggtgccagaggactg) and the PCR fragment from the first round of PCR. The product from the second round of PCR was digested with SmaI and ApaI and cloned into pAAR325 digested with the same enzymes generating pAAR334. pAAR360 is identical to pAAR359 with the exception that the H360 codon in rodK was changed to an Ala codon (cac changed to gcc). pAAR360 was constructed by cloning a SalI–EcoRI fragment from pAAR334 in pBGS18 digested with the same enzymes. pAAR327 was constructed as follows: a PCR product was generated using pAAR325 as template, and the primers 5′-cccGAATTCattaaaga ggagaaattaactatgcatcaccatcaccatcacgcctcctcacagcctccctctg (bases in capital provide an EcoRI site, bases underlined add six His codons after the ATG start codon of rodK) and 5′-agcgtggcctcgcaccagtc. The PCR product was digested with EcoRI and EagI yielding a fragment of 187 bp and cloned in pAAR325 digested with EcoRI and NotI generating pAAR326. The His6-tagged rodK construct was moved from pAAR326 on a EcoRI–HindIII fragment to the expression vector pUHE24-2 (Lutz and Bujard, 1997) digested with the same enzymes generating pAAR327. To construct pAAR353, several intermediate plasmids were used. pAAR341 carries a rodK allele with all three D to N substitutions (D657N, D782N, D909N). The His6-tag was added to this construct by moving a 187 bp EcoRI–NotI fragment from pAAR326 to pAAR341 digested with the same enzymes generating pAAR347. Subsequently, the His6-rodKD657N,D782N,D909N construct was moved to pUHE24-2 digested with EcoRI and HindIII on an EcoRI–HindIII fragment from pAAR347 giving rise to pAAR353. For pAAR335 construction, a PCR product was generated using pAAR325 as template, and the primers o8153(D657N) (5′-caggtggatgtTcaggaggatg, mismatch in capital gives rise to the D657N substitution) and o8153-2 (5′-aaggtggtggtggccgaggac). In a second round of PCR using pAAR325 as template, and the primers o8153-3 (5′-ctggtg gaggtgccagaggactg) and the PCR product from the first round of PCR. The PCR product from the second round of PCR was digested with AgeI and ApaI and ligated into pAAR325 digested with the same enzymes giving rise to pAAR335. pAAR335 contains a rodK allele encoding the D657N substitution. To construct pAAR339, pAAR325 was used as template in three PCR reactions. In the first PCR reaction the primers o8153(D782N) (5′-gcatcatcaaGtTcag caccacc, the mismatch bases in capital gives rise to the D782N substitution) and o8153-2 (5′-aaggtggtggtggccgag gac) were used. In the second PCR reaction, the primers were o8153(D909N) (5′-catcctgatgAacctgtcgcttc, mismatch in capital gives rise to the D909N substitution) and −47 seq primer (5′-cgccagggttttcccagtcacgac). In the third PCR reaction, the primers were the PCR products from the first two rounds of PCR. The PCR product from the third PCR reaction was digested with ApaI and XhoI and ligated into pAAR325 digested with the same enzymes generating pAAR339. pAAR339 contains a rodK allele encoding the D782N and D909N substitutions. pAAR341 was constructed by moving a 180 bp AgeI–ApaI fragment with the D657N mutation from pAAR335 to pAAR339, which contain the D782N and D909N mutations, digested with AgeI and ApaI. pAAR341 contains a rodK allele encoding the D657N, D782N and D909N substitutions. For pAAR355 construction, a 187 bp EcoRI–NotI fragment from pAAR326 (contains His6-tag codons) and a 3015 bp NotI–HindIII fragment, which contains a rodK allele with the H360A, D657N, D782N, D909N mutations, from pAAR354 were ligated in one step into pUHE24-2 digested with EcoRI and HindIII. pAAR354 construction: an ApaI–XhoI fragment that carries the D657N, D782N, D909N substitutions from pAAR341 was cloned into the same sites in pAAR334 giving rise to pAAR354.
DNA sequencing of region around miniTn5(tet)W8153
The sequences flanking miniTn5(tet)Ω8153 were obtained by arbitrary PCR (Rasmussen and Søgaard-Andersen, 2003). These sequences were subjected to a blast search (Altschul et al., 1990) against the M. xanthus genome released by The Institute for Genomic Research at http://www.tigr.org and approximately 5 kb on either side of the insertion was retrieved. The sequence reported in this article has been deposited in the GenBank database (Accession No. AY779624).
Protein purification, in vitro phosphorylation, generation of antibodies and immunoblot analyses
His6-RodKWT, His6-RodKD657N,D782N,D909N and His6-RodKH360A, D657N, D782N, D909N proteins were expressed from pAAR327, pAAR353 and pAAR355, respectively, in E. coli JM109 and purified as described (Martin et al., 2001). In vitro phosphorylation of proteins was carried out in TGMNKD buffer [50 mM TrisHCl (pH 8), 10% (v/v) glycerol, 5 mM MgCl2, 150 mM NaCl, 50 mM KCl, 1 mM DTT] containing 0.5 mM [γ-32P]-ATP (14.8 GBq mmol−1; Amersham) and RodK (5 µM) for 5 h at 20°C. Aliquots of 10 µl were quenched in 5 µl of 3× SDS/EDTA loading dye [7.5% (w/v) SDS, 90 mM EDTA, 37.5 mM TrisHCl (pH 6.8), 37.5% glycerol, 0.3 M DTT], loaded without prior heating on a standard 8% polyacrylamide gel and separated by SDS-PAGE at 4°C (Porter and Armitage, 2002).
To test the chemical stability of the phosphoryl group in His6-RodKD657N,D782N,D909N, quenched samples of 10 µl were transferred to 2 µl of 1 M HCl, H2O or 3 M NaOH, incubated at 42°C for 1 h and separated by SDS-PAGE (Zhou and Wolk, 2003). The gel was quantified by phosphorimaging (Molecular Dynamics) and analysed using ImageQuant (v5.0, Molecular Dynamics).
To generate anti-RodK antibodies, His6-RodKWT was purified and used to immunize a rabbit using standard procedures (Sambrook et al., 1989). CsgA and FrzCD immunoblot analyses were carried out as described (Kruse et al., 2001). FruA and RodK immunoblots were carried out using standard procedures (Sambrook et al., 1989) and using peroxidase-conjugated goat anti-rabbit immunoglobulin G as secondary antibodies as recommended by the manufacturer (Roche). Blots were developed using Renaissance Plus Chemiluminescence reagent (NEN Life Sciences).
We thank Y. Cheng and D. Kaiser for the gift of strains. Sequencing of M. xanthus at TIGR was accomplished with support from NSF. The Danish Natural Science Research Council, the FREJA programme and the Carlsberg Foundation supported this work.
- 1990) Basic local alignment search tool. J Mol Biol 215: 403–410. , , , , and (
- 1984) The relationship between base composition and codon usage in bacterial genes and its use for the simple and reliable identification of protein-coding sequences. Gene 30: 157–166. , , and (
- 1999) Sporulation timing in Myxococcus xanthus is controlled by the espAB locus. Mol Microbiol 34: 714–725. , and (
- 2000) The Myxococcus xanthus socE and csgA genes are regulated by the stringent response. Mol Microbiol 37: 788–799. , and (
- 1991) Chemical properties and separation of phosphoamino acids by thin-layer chromatography and/or electrophoresis. Methods Enzymol 201: 10–21. , , and (
- 1996) Recent advances in the social and developmental biology of the Myxobacteria. Microbiol Rev 60: 70–102. (
- 1998) The FruA signal transduction protein provides a checkpoint for the temporal co-ordination of intercellular signals in M. xanthus development. Mol Microbiol 30: 807–817. , , and (
- 1998) SdeK is required for early fruiting body development in Myxococcus xanthus. J Bacteriol 180: 4628–4637. , , , , , , and (
- 1998) Signal decay through a reverse phosphorelay in the Arc two-component signal transduction system. J Biol Chem 273: 32864–32869. , , , and (
- 1998) The guanosine nucleotide (p)ppGpp initiates development and A-factor production in Myxococcus xanthus. Genes Dev 12: 1022–1035. , , and (
- 1979) Genetics of gliding motility in Myxococcus xanthus (Myxobacterales): two gene systems control movement. Mol Gen Genet 171: 177–191. , and (
- 2002) Role of fruA and csgA genes in gene expression during development of Myxococcus xanthus. Analysis by two-dimensional gel electrophoresis. J Biol Chem 277: 26753–26760. , , , , and (
- 2000) Spatial control of cell differentiation in Myxococcus xanthus. Proc Natl Acad Sci USA 97: 9098–9103. , , and (
- 1979) Social gliding is correlated with the presence of pili in Myxococcus xanthus. Proc Natl Acad Sci USA 76: 5952–5956. (
- 1990) C-factor: a cell–cell signaling protein required for fruiting body morphogenesis of M. xanthus. Cell 61: 19–26. , and (
- 1991) C-factor has distinct aggregation and sporulation thresholds during Myxococcus development. J Bacteriol 173: 1722–1728. , and (
- 2001) Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305: 567–580. , , , and (
- 1987) Expression of many developmentally regulated genes in Myxococcus depends on a sequence of cell interactions. Genes Dev 1: 840–854. , and (
- 1986) A global analysis of developmentally regulated genes in Myxococcus xanthus. Dev Biol 117: 252–266. , , and (
- 2001) C-signal: a cell surface-associated morphogen that induces and coordinates multicellular fruiting body morphogenesis and sporulation in M. xanthus. Mol Microbiol 40: 156–168. , , , and (
- 1982) Fruiting body morphogenesis in submerged cultures of Myxococcus xanthus. J Bacteriol 151: 458–461. , and (
- 1989) Genes required for developmental signalling in Myxococcus xanthus: three asg loci. J Bacteriol 171: 2762–2772. , and (
- 1986) Intercellular signaling is required for developmental gene expression in Myxococcus xanthus. Dev Biol 117: 267–276. , , and (
- 1992a) Identification of heat-stable A-factor from Myxococcus xanthus. J Bacteriol 174: 3319–3326. , , and (
- 1992b) A-signalling and the cell density requirement for Myxococcus xanthus development. J Bacteriol 174: 7360–7369. , , and (
- 1992) csgA expression entrains Myxococcus xanthus development. Genes Dev 6: 401–410. , , and (
- 2000) A common step for changing cell shape in fruiting body and starvation-independent sporulation in Myxococcus xanthus. J Bacteriol 182: 3553–3558. , , and (
- 2003) Identification of the C-signal, a contact-dependent morphogen coordinating multiple developmental responses in Myxococcus xanthus. Genes Dev 17: 2151–2161. , and (
- 1997) Independent and tight regulation of transcriptional units in Escherichia coli via LacR/O, the TetR/O and AraC/I1-2 regulatory elements. Nucleic Acids Res 25: 1203–1210. , and (
- 1993) FrzCD, a methyl-accepting taxis protein from Myxococcus xanthus, shows modulated methylation during fruiting body formation. J Bacteriol 175: 4936–4940. , and (
- 1978) Systematic isolation of transducing phages for Myxococcus xanthus. Virology 88: 44–53. , , , and (
- 2001) CheR- and CheB-dependent chemosensory adaptation system of Rhodobacter sphaeroides. J Bacteriol 183: 7135–7144. , , , , , , et al. (
- 1979) Myxococcus xanthus mutants with temperature-sensitive, stage-specific defects: evidence for independent pathways in development. J Bacteriol 140: 1036–1042. , and (
- 1991) Development in Myxococcus xanthus involves differentiation into two cell types, peripheral rods and spores. J Bacteriol 173: 3318–3333. , and (
- 1996) FruA, a putative transcription factor essential for the development of Myxococus xanthus. Mol Microbiol 22: 757–767. , , , , and (
- 1998) Kinase-phosphatase competition regulates Bacillus subtilis development. Trends Microbiol 6: 366–370. (
- 2001) SdeK, a histidine kinase required for Myxococcus xanthus development. J Bacteriol 183: 3589–3596. , and (
- 2002) Phosphotransfer in Rhodobacter sphaeroides chemotaxis. J Mol Biol 324: 35–45. , and (
- 2003) TodK, a putative histidine protein kinase, regulates timing of fruiting body morphogenesis in Myxococcus xanthus. J Bacteriol 185: 5452–5464. , and (
- 2000) A tale of two components: a novel kinase and a regulatory switch. Nat Struct Biol 7: 626–633. , , and (
- 1989) Molecular Cloning. A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. , , and (
- 1998) SMART, a simple modular architecture research tool: identification of signaling domains. Proc Natl Acad Sci USA 95: 5857–5864. , , , and (
- 1993) The two motility systems of Myxococcus xanthus show different selective advantages on various surfaces. Proc Natl Acad Sci USA 90: 3378–3382. , and (
- 1988) Use of recombination techniques to examine the structure of the csg locus of Myxococcus xanthus. Mol Gen Genet 211: 63–71. , and (
- 1982) Induction of coordinated movement of Myxococcus xanthus cells. J Bacteriol 152: 451–461. , and (
- 1995) Ectopic production of guanosine penta- and teraphosphate can initiate early developmental gene expression in Myxococcus xanthus. Genes Dev 9: 1633–1644. , and (
- 1996) C factor, a cell-surface-associated intercellular signaling protein, stimulates the cytoplasmic Frz signal transduction system in Myxococcus xanthus. Proc Natl Acad Sci USA 93: 2675–2679. , and (
- 1996) Intercellular C-signaling in Myxococcus xanthus involves a branched signal transduction pathway. Genes Dev 10: 740–754. , , , and (
- 2003) Coupling gene expression and multicellular morphogenesis during fruiting body formation in Myxococcus xanthus. Mol Microbiol 48: 1–8. , , , , , and (
- 1986) Kanamycin-resistant vectors that are analogues of plasmids pUC8, pUC9, pEMBL8 and pEMBL9. Gene 41: 337–342. , , , , and (
- 2001) A novel feature of the multistep phosphorelay in Escherichia coli: a revised model of the RcsC to YojN to RcsB signalling pathway implicated in capsular synthesis and swarming behaviour. Mol Microbiol 40: 440–450. , , , , and (
- 1993) devRS, an autoregulated and essential genetic locus for fruiting body development in Myxococcus xanthus. J Bacteriol 175: 7450–7462. , and (
- 1996) Central role of the BvgS receiver as a phosphorylated intermediate in a complex two-component regulatory phosphorelay. J Biol Chem 271: 33176–33180. , and (
- 1985) Improved M13 phage cloning vectors and host strains: nucleotide sequence of the M13mp18 and pUC19 vectors. Gene 33: 103–119. , , and (
- 2003) A two-component system mediates developmental regulation of biosynthesis of a heterocyst polysaccharide. J Biol Chem 278: 19939–19946. , and (