AsgD, a new two-component regulator required for A-signalling and nutrient sensing during early development of Myxococcus xanthus

Authors


David R. Zusman. E-mail zusman@uclink4.berkeley.edu; Tel. (+1) 510 642 2293; Fax (+1) 510 643 6334.

Abstract

Myxococcus xanthus has a complex life cycle that includes fruiting body formation. One of the first stages in development has been called A-signalling. The asg (A-signalling) mutants have been proposed to be deficient in producing A-signal, resulting in development arresting at an early stage. In this paper, we report the identification of a new asg locus asgD. This locus appears to be involved in both environmental sensing and intercellular signalling. Expression of asgD was undetected during vegetative growth, but increased dramatically within 1 h of starvation. The AsgD protein is predicted to contain 773 amino acids and to be part of a two-component regulatory system because it has a receiver domain located at the N-terminus and a histidine protein kinase at the C-terminus. An asgD null mutant was defective in fruiting body formation and sporulation on CF medium. However, the defects of the mutant were complemented extracellularly when cells were mixed with wild-type strains or with bsgA, csgA, dsgA or esgA mutants, but were not complemented extracellularly by asgA, asgB or asgC mutants. In addition, the mutant was rescued by a subset of A-factor amino acids. Surprisingly, when the mutant was plated on stringent starvation medium rather than CF, cells were able to form fruiting bodies. Thus, it appears that AsgD is directly or indirectly involved in sensing nutritionally limiting conditions. The discovery of the asgD locus provides an important sensory transduction component of early development in M. xanthus.

Introduction

Myxococcus xanthus is a Gram-negative soil bacterium that displays a complex life cycle and social behaviours, including multicellular development. In a nutrient-rich environment, rod-shaped cells grow vegetatively. However, if nutrients are depleted, cells move into aggregation centres to construct fruiting bodies in which ≈ 100 000 individual cells undergo morphogenesis from rod-shaped vegetative cells to spherical myxospores (for reviews, see Shimkets, 1990; Dworkin and Kaiser, 1993). Development of M. xanthus is a multicellular event involving large numbers of cells which must co-ordinate their activities. It has been hypothesized that co-ordination is facilitated by extracellular signalling because some mutants cannot develop normally on their own but develop when they are mixed with wild-type cells (McVittie et al., 1962; Hagen et al., 1978). Hagen et al. (1978) isolated many such developmental mutants, which they classified into four extracellular complementation groups: asg, bsg, csg and dsg. Downward et al. (1993) later identified another extracellular complementation group, which is represented by the esg mutant.

One of the most studied of the ‘signalling’ steps is called A-signalling (for a review, see Kim et al., 1992). The asg (A-signalling) mutants have been proposed to be defective in producing A-signal, resulting in developmental arrest at an early stage (Hagen et al., 1978; Kuspa et al., 1986). A-signal activity can be satisfied by a mixture of amino acids and short peptides (A-factor) that are generated by proteolysis (Kuspa et al., 1992a). The asg mutants are thought to export little or no A-factor and therefore are not able to complement other asg mutants (Kuspa et al., 1986). A subgroup of individual amino acids has been identified that by themselves rescue developmental defects of the asgB mutant (Kuspa et al., 1992b). These A-factor amino acids have been proposed to also function as a cell density signal (Kuspa et al., 1992b). Three asg loci, asgA, asgB and asgC, were identified before this study (Kuspa and Kaiser, 1989; Mayo and Kaiser, 1989). DNA sequence analysis suggests that asgA encodes a histidine protein kinase (Plamann et al., 1995; Li and Plamann, 1996), asgB encodes a DNA-binding protein (Plamann et al., 1994), and asgC encodes a major vegetative sigma factor which presumably impacts transcription in M. xanthus (Davis et al., 1995). In this study, we report the discovery of a new member of the Asg family of developmental signalling loci. We call this mutant asgD. A null mutation in asgD causes developmental defects which can be complemented extracellularly by the wild type and by the bsg, csg, dsg and esg mutants. However, the defects of the mutant cannot be complemented extracellularly by the asg mutants.

Results

Mutant isolation and characterization of the mutant locus

We were interested in identifying additional genes involved in the development of M. xanthus strain DZ2. Our approach was to prepare a plasmid library in the pZErO-2 vector (KmR) carrying about 500 bp of random M. xanthus DNA fragments. These plasmids were then introduced into M. xanthus by electroporation. As these plasmids cannot replicate autonomously in M. xanthus, only strains which underwent homologous recombination between the cloned DNA fragment and chromosomal DNA could produce colonies on a medium containing kanamycin. However, recombination also created insertional mutations because the cloned DNA fragments were smaller than most genes. We were particularly interested in one isolate, DZ4212, as this mutant showed severe defects in early development. For example, when 20 μl of cells (5 × 109 cells ml−1) were spotted on CF plates, DZ4212 failed to form fruiting bodies and produced almost no spores (less than 50 spores per spot in 7 days). In contrast, under the same conditions, the wild-type strain DZ2 formed fruiting bodies and produced around 1 × 108 spores per spot. However, the mutant showed normal vegetative colony morphology and did not show any defect in motility and cohesiveness (data not shown).

Because the mutant phenotype of DZ4212 was tentatively attributed to plasmid insertion, the region containing the insertion was analysed by cloning and DNA sequencing (see Experimental procedures). The DNA sequence showed that the plasmid insertion was located in an open reading frame (ORF), which is designated asgD (A-signal) (Fig. 1) because of the phenotype described below. To confirm this conclusion and to exclude the possibility that other mutations might have caused the mutant phenotype of DZ4212, a new asgD insertion mutant was constructed by inserting a plasmid pKY482, which contains a different internal DNA fragment of asgD, into the wild-type strain DZ2 (Fig. 1). The new mutant, DZ4232, showed the same phenotype as DZ4212, suggesting that insertional mutation of the asgD locus is indeed responsible for the mutant phenotype of the strain. We also sequenced 3 kb upstream and 2.5 kb downstream of asgD, but were unable to locate any ORFs with similarity to proteins of two-component systems. However, as the plasmid insertion mutation in asgD could have possible polar effects on downstream genes, we created an asgD in frame deletion mutation, which should not disturb the expression of downstream genes. The mutant was created using plasmid pKY568 (Fig. 1), as described in the Experimental procedures. The resulting mutant, DZ4235, encodes a protein missing amino acids 12–737 of AsgD, so that 94% of the AsgD ORF (773 amino acids) is deleted. However, DZ4235 showed the same mutant phenotype as DZ4212 and DZ4232 (Table 1 and Fig. 2[link]), suggesting that the mutation in asgD is solely responsible for the developmental defects of the mutants and that asgD is essential for the development of M. xanthus.

Figure 1.

. A physical map of the asgD locus and plasmids used for mutant construction. A. DNA regions inserted into plasmids and used for the construction of plasmid insertion mutants. B. The sequence of the DNA at the beginning of the asgD gene. The putative ribosome binding site is underlined.

Table 1. . Sporulation of the asgD mutant under different starvation conditions.a a. Twenty microlitres of 1000 Klett cells (5 × 109 cells ml−1) was placed on 1.5% agar plates containing the medium described in the table. After incubation at 34°C for 7 days, the numbers of sonication-resistant refractile spores were visually counted using a haemocytometer. The number of spores on CF is higher than the number of cells originally placed on the plates since CF medium supports some growth of the cells.b. Indicated as a number of germinated spores because there were too few spores to be visually counted with a haemocytometer.c. Less than 1 × 10−4%.Thumbnail image of
Figure 2.

. Developmental phenotype of the asgD mutant. Cells (1 × 108) of DZ2 (wild type), and DZ4235 (ΔasgD) were placed on CF plates as a 20 μl spot and incubated at 34°C for 4 days. The morphology of the cells was photographed using a dissecting microscope. Bar, 2 mm.

Similarity of AsgD to other proteins in the database

The deduced asgD gene product AsgD contains 773 amino acids. Sequence analysis indicates a receiver domain at the N-terminal region, a histidine protein kinase domain at the C-terminal region of AsgD, and an intermediate region of unknown function that separates these two domains (Fig. 3A). The N-terminal region from amino acid 53 to amino acid 170 of AsgD is 36.8%, 33.3% and 30.1% identical to the receiver domains of the response regulator PleD from Caulobacter crescentus (Hecht and Newton, 1995) and HydG and OmpR from Escherichia coli (Wurtzel et al., 1982; Stoker et al., 1989) respectively (Fig. 3B). The aspartate residue at amino acid 102 is predicted to be a phosphorylation site of the receiver domain. The C-terminal region, from amino acid 550 to amino acid 773 of AsgD, is 44.9%, 30.8% and 29.2% identical to the kinase domains of the histidine protein kinase SdeK from M. xanthus (Garza et al., 1998), KdpD from Escherichia coli (Walderhaug et al., 1992) and PilS from Pseudomonas aeruginosa (Hobbs et al., 1993) respectively (Fig. 3C). The histidine residue at amino acid 560 is predicted to be an autophosphorylation site. The AsgD protein has an unusual domain organization compared with other histidine protein kinases but is similar to that of AsgA from M. xanthus and CyaC from Anabaena sp. Like AsgA, AsgD has a receiver domain at the N-terminus and a histidine protein kinase domain at the C-terminus (Fig. 3A). The receiver domains of AsgD and AsgA are 31.5% identical and the kinase domains of AsgD and AsgA are 26.6% identical. However, AsgA does not have an intermediate region between the domains whereas AsgD has an ≈ 380-amino-acid-long intermediate region. AsgD is 24.6% identical to CyaC from Anabaena sp., which also has a receiver domain at the N-terminus, an intermediate region and a kinase domain (Katayama and Ohmori, 1997). However, unlike CyaC, AsgD does not have an adenylate cyclase domain at the C-terminus.

Figure 3.

. Domain organization and sequence alignment of AsgD with homologous proteins. A. Domain organization of AsgD. B. The N-terminal region of AsgD is aligned with the receiver domains of PleD from Caulobacter crescentus (Hecht and Newton, 1995) and HydG and OmpR from Escherichia coli (Wurtzel et al., 1982; Stoker et al., 1989). C. The C-terminal region of AsgD is aligned with the histidine kinase domains of SdeK from M. xanthus (Garza et al., 1998), KdpD from Escherichia coli (Walderhaug et al., 1992), and PilS from Pseudomonas aeruginosa (Hobbs et al., 1993). The stars indicate putative phosphorylation sites.

Developmental phenotype of the asgD mutant is partially rescued under stringent starvation conditions

The ΔasgD mutant DZ4235 is developmentally defective and did not sporulate on CF plates (Table 1). Under the same conditions, the wild-type strain produced ≈ 1 × 108 spores per spot. However, when the mutant was placed on a more stringent starvation medium, MMC, the mutant produced 8.4 × 106 spores per spot, 35% of the number of spores produced by the wild-type strain under the same conditions (Table 1). The CF medium contains 10 mM MOPS (pH 7.6), 0.015% Casitone, 8 mM MgSO4, 1 mM KH2PO4, 2% sodium citrate, 1% pyruvate and 1.5% Difco agar. In contrast, the MMC medium contains 10 mM MOPS, 4 mM MgSO4, 2 mM CaCl2 and 1.5% Difco agar. It should be noted that Difco agar may contain trace amounts of uncharacterized materials. The overall levels of nutrients in the CF medium, although higher than the MMC medium, are limited, perhaps barely enough to support a twofold increase in cell number. Thus, most wild-type strains undergo a gradual starvation and then fruit on this medium, forming larger fruiting bodies with higher numbers of spores than cells on the MMC medium. Because the development of the asgD mutant was inhibited on the CF medium, we deleted individual components of the medium in an attempt to identify the component responsible for inhibiting development. Table 1 shows that deleting KH2PO4 or (NH4)2SO4 from CF plates did not restore fruiting and sporulation in DZ4235. However, deleting citrate, pyruvate or Casitone from the CF medium increased the number of spores produced by DZ4235 to 0.25%, 0.96% and 44% of the wild type respectively (Table 1). As aggregation and sporulation of the mutant are stimulated by omitting different kinds of nutrients, i.e. citrate, pyruvate or Casitone (a proteolytic digest of casein), we conclude that the development of the asgD mutant is inhibited by the overall level of nutrients in the CF medium. We hypothesize that the asgD mutant is hypersensitive to nutrients and that its development is inhibited even by the limited amount of nutrients found in the CF medium, although low enough to induce development in the wild type. Although stringent starvation conditions partially rescue development of the asgD mutant, the rescue is not complete. For example, the fruiting bodies from the asgD mutant on CF medium lacking Casitone (Fig. 2) or on the MMC medium (which lacks citrate, pyruvate and Casitone) are irregular and are not as compact as fruiting bodies from the wild-type strain. In addition, the sporulation efficiency of the mutant on these media is always lower than that of wild type (Table 1). This indicates that asgD is required for normal development even under strict starvation conditions.

Extracellular complementation of the asgD mutant

McVittie et al. (1962) and Hagen et al. (1978) showed that many developmental mutants, although unable to form fruiting bodies on their own, can be rescued to fruit by mixing with wild-type cells before placing on starvation agar. We therefore tested whether the developmental defects of the asgD mutant on CF plates can be complemented extracellularly by mixing with the wild-type strain DZ2. When 5 × 107 cells of the kanamycin-resistant asgD mutant DZ4212 were mixed with 5 × 107 cells of the kanamycin-sensitive wild-type strain DZ2 and placed as 20 μl spots on CF plates, the mixed cells produced 3.3 × 104 kanamycin-resistant spores per spot (Table 2). Under the same conditions, the asgD mutant (DZ4212) produced only 37 kanamycin-resistant spores per spot. Thus, it appears that DZ2 stimulates the asgD mutant to increase sporulation about 1000-fold. Stimulatable mutants in M. xanthus are hypothesized to be ‘signal’ mutants and are organized into five extracellular complementation groups (Hagen et al., 1978; Downward et al., 1993). We tested whether the asgD mutant belongs to one of these groups. As shown in Table 2, when the asgD mutant DZ4212 was mixed with bsgA, csgA, dsgA and esgA mutants, the number of kanamycin-resistant spores from the asgD mutant increased 32- to 486-fold (Table 2). Aggregation of the asgD mutant was also partially rescued in these mixing experiments. For example, Fig. 4 shows that DZ4212 (asgD) and DK5209 (bsgA), each of which normally do not form fruiting bodies under developmental conditions, are able to form fruiting bodies when mixed. However, extracellular complementation of the asgD mutant DZ4212 was not achieved with the three A-signal mutants asgA, asgB and asgC. The number of kanamycin-resistant spores from the mixed cells was below that from DZ4212 alone (Table 2). In addition, the mixed cells of the DZ4212 and asg mutants did not form fruiting bodies. This suggests that the asgD mutant is a member of the asg mutant group.

Table 2. . Extracellular complementation of the sporulation of the asgD mutant by other strains.aThumbnail image of
Figure 4.

. Extracellular complementation of the asgD mutation by the wild type and by the signal mutants. Each strain was cultured in CYE to exponential phase, harvested and resuspended to 5 × 109 cells ml−1. DZ4212 (asgD) was mixed with the same volume of DZ2 (wt), DZ4236 (asgB), or DK5209 (bsgA). Twenty microlitres of each strain or mixed strains was placed on CF plates and incubated at 34°C for 4 days. A. DZ2. B. DZ4212 (asgD). C. DZ4236 (asgB). D. DZ4212 (asgD) and DZ4236 (asgB). E. DK5209 (bsgA). F. DZ4212 (asgD) and DK5209 (bsgA). Bar, 2 mm.

Developmental defects of the asgD mutant are rescued by some l-amino acids

The developmental defects of the asg mutants can be rescued by a mixture of amino acids and short peptides known as A-factor (Kuspa et al., 1992a, b). Furthermore, some amino acids found in A-factor can, by themselves, rescue the developmental defects of the asgB mutant (Kuspa et al., 1992b). We therefore tested whether the developmental defects of the asgD mutant can be rescued by individual l-amino acids on CF plates. As shown in Table 3 and Fig. 5, the results indicate that the developmental defects of the asgD mutant are also rescued by some individual amino acids. Leucine was the most effective amino acid at rescuing the asgD mutants; sporulation was almost identical to that of wild-type cells (95% of the wild-type level of spores). Leucine is one of the A-factor amino acids. However, proline, which is also one of the A-factor amino acids, had no effect. In contrast, valine, which is reported to have no detectable A-factor activity, partially restored the development of the asgD mutant. Isoleucine is also reported to be one of the A-factor amino acids. However, isoleucine inhibited the development of the wild-type strain DZ2 and the asgD mutant. Under the same conditions, isoleucine did not inhibit development of FB strains and its derivatives such as DZF1 and DK1622. Methionine also inhibited the development of the wild type as reported previously (Rosenberg et al., 1973; Shi and Zusman, 1995). However, methionine partially restored the development of the asgD mutant (3.3 × 105 spores per spot).

Table 3. . Effects of amino acids on sporulation of DZ4235 on CF plates a. 1 mM of each l-amino acid was added to CF plates.b. A-factor activity (U ml−1) of each amino acid (100 μM) reported by Kuspa et al. (1992a).c. Number of sonication-resistant refractile spores in a spot, which was originally placed as 20 μl spots of 1000 Klett cells (5 × 109 cells ml−1) on CF plates and incubated at 34°C for 7 days.d. Indicated as a number of germinated spores because there were too few spores to be counted with a haemocytometer.e. Below 1 × 10−4%.f. Sporulation of the wild type is inhibited.Thumbnail image of
Figure 5.

. Effects of amino acids on the development of the asgD mutant DZ4235. Cells (1 × 108) of DZ2 or DZ4235 were placed on CF plates containing 1 mM of each l-amino acid and were incubated at 34°C for 4 days. Bar, 5 mm.

Because of the unexpected rescue of the asgD mutant phenotype by some amino acids but not others, we became concerned that the asgD mutation might cause defects in amino acid metabolism and that these defects might be influencing the amino acids rescue experiments. We therefore tested whether the asgD mutant may require additional amino acids for growth by culturing the cells in a defined medium, Al (Bretscher and Kaiser, 1978). We found that the mutant grew in this medium and that the growth rate was the same as the wild-type strain DZ2 (data not shown). This indicates that DZ4235 does not require additional amino acids besides those already present in the A1 medium. The A1 medium contains many amino acids but does not contain alanine, aspartate, cysteine, glycine, histidine, serine and tryptophan (Bretscher and Kaiser, 1978). However, as shown in Table 3, all of these amino acids rescue the developmental defects of the asgD mutant. Therefore, it is unlikely that the effects of exogenous amino acids on the development of the asgD mutant are caused by simply supplying amino acids to cells that are defective in amino acid biosynthesis.

Some A-factor amino acids inhibit development of the asgD mutant

As shown in Table 3, some A-factor amino acids, such as proline, phenylalanine and tyrosine, did not rescue the developmental defects of the mutant in CF medium. Under stringent starvation conditions (CF plates lacking Casitone), in which the developmental phenotype of the mutant is partially rescued, the asgD mutant DZ4235 produced 1.8 × 107 spores from a 20 μl spot of cells (5 × 109 cells ml−1). However, when 1 mM of arginine, asparagine, glutamine, phenylalanine, proline or tyrosine was added to these plates, DZ4235 failed to sporulate (Table 4). In contrast, the sporulation of DZ4235 was not affected by lysine (Table 4), but lysine is not an A-factor amino acid (Kuspa et al., 1992a). These results suggest that the A-factor amino acids can be divided into two groups: one group is able to rescue the developmental defects of the asgD mutant and the other group inhibits the development of the asgD mutant. Although proline inhibits development of the asgD mutant, development of the wild-type strain DZ2 is stimulated by proline. This further suggests that the asgD mutation has redirected M. xanthus to recognize proline, which is originally a stimulatory signal for the wild type, as an inhibitory signal.

Table 4. . Inhibitory effects of some A-factor amino acids on the developmental rescue of DZ4235 under stringent starvation conditions. a. Each l-amino acid (1 mM) was added to CF plates lacking Casitone.b. A-factor activity (U ml−1) of each amino acid (100 μM) reported by Kuspa et al. (1992a).c. Number of sonication-resistant refractile spores in a spot, which were originally placed as 20 μl spots of 1000 Klett cells (5 × 109 cells ml−1) on plates and incubated at 34°C for 7 days.d. Indicated as a number of germinated spores because there were too few spores to be counted with a haemocytometer.Thumbnail image of

Expression of an asg-dependent lac fusion Tn5 lac-Ω4521 is not affected by the asgD mutation

Tn5 lac-Ω4521 is one of the representative asg-dependent lac fusions created by Kroos et al. (1986), and its expression is commonly used as an A-factor assay. Expression of Tn5 lac-Ω4521 is normally induced within 2 h of starvation. We introduced an asgD mutation into strain DK6620 carrying Tn5 lac-Ω4521 (this strain is called DZ4275) and tested the effect of the asgD mutation on expression of the fusion in the cell after spotting onto CF plates. The expression of Tn5 lac-Ω4521 was found to be the same as that of the asgD+ strain (data not shown). Under the same conditions, expression of Tn5 lac-Ω4521 in the asgB mutant DK6600 was defective (data not shown). This suggests that the expression of Tn5 lac-Ω4521 is asgD independent. This predicts that the asgD mutant would produce wild-type levels of A-factor, at least as measured by Tn5 lac-Ω4521.

Expression of asgD is developmentally regulated

To study the expression of asgD during development, we created an asgD–lacZ translational fusion between codon 13 of asgD and codon 8 of lacZ. The fusion was first created in vitro as a plasmid, as described in the Experimental procedures. The resulting plasmid, pKY565, was then introduced into M. xanthus DZ2 (asgB+), DK1622 (asgB+) and DZ4236 (asgB) by electroporation, where it recombined into the chromosome to create DZ4233, DZ4238 and DZ4239 respectively (Fig. 6A). The plasmid contained a region between 653 bp upstream and 37 bp downstream from the asgD translation start codon and the phenotype of the resultant strains was indistinguishable from wild-type DZ2 under developmental conditions. Thus, although these strains contain a lacZ reporter fusion in asgD, they also contain a wild-type copy of asgD that is expressed normally. During vegetative growth and glycerol-induced sporulation, no β-galactosidase activity was detected from DZ4233, which carries the asgD–lacZ fusion in the DZ2 background (data not shown). However, under starvation conditions, expression of the fusion was induced within 1 h and β-galactosidase activity increased until after 24 h (Fig. 6B). Expression of the fusion in the asgB mutant was also induced by starvation (Fig. 6B). However, the expression level was lowered by 50% compared with the asgB+ strain (Fig. 6B). Proline or leucine (1 mM) did not rescue the reduced expression of the fusion in the asgB mutant (data not shown). These results indicate that asgD is a developmentally regulated gene and its expression is only partially affected by AsgB.

Figure 6.

. Developmental expression of the asgDlacZ fusion. A. Construction of the asgD–lacZ fusion in M. xanthus. An asgD–lacZ fusion was generated in vitro and fused into the asgD locus by homologous recombination, resulting in an asgD–lacZ fusion in an asgD+ background. B. Expression of the asgD–lacZ fusion under developmental conditions. DZ4238 (○; asgB+) and DZ4239 (●; asgB) carrying the asgD–lacZ fusion were placed on CF plates and incubated at 34°C. The cells were harvested at different time points and β-galactosidase activity was measured as described previously (Kroos et al., 1986). Every point represents the average of two experiments.

Discussion

Multicellular behaviours of M. xanthus such as swarming, group feeding and fruiting body formation require cell–cell co-ordination and intercellular signalling. A-signalling has been proposed to be an important early signalling step in M. xanthus based on extracellular complementation studies with a group of mutants called asg mutants. In this paper, we describe the isolation and characterization of a new asg locus asgD. This locus fits the formal definition of an asg locus because an asgD null mutant is unable to undergo development by itself but can undergo development when complemented extracellularly by the wild-type strain or by the bsg, csg, dsg and esg mutants. However, the asgD mutant cannot be complemented by the three other asg mutants. Furthermore, the developmental defects of the asgD mutant can be rescued by some of the same individual amino acids responsible for the rescue of the asgB mutant.

Although the asgD mutant shares several similarities with the other asg mutants asgA, asgB and asgC, it also shows several important differences. (1) Phase variation: M. xanthus shows two colony types, called ‘yellow’ and ‘tan’, which are generally attributed to phase variation (Burchard and Dworkin, 1966). The two colony types differ in many traits which include colony colour, motility, fruiting and sporulation proficiency and gene expression. The vegetative colonies of the three asg mutants are locked in the ‘tan’ phase and are morphologically different from the wild type during vegetative growth (Kuspa and Kaiser, 1989). In contrast, the colonies of the asgD mutant are ‘yellow’ and appear to be morphologically similar to the wild type. (2) Cell cohesiveness: the cells of the three asg mutants are less cohesive than the wild-type cells (Kuspa and Kaiser, 1989). In contrast, the asgD mutant has wild-type cohesiveness (data not shown). (3) Effect of nutrition on the ability of cells to undergo development: the asgD mutant does not sporulate on CF medium but does sporulate on more stringent starvation media such as CF medium lacking Casitone. In contrast, the other asg mutants show more severe developmental defects on the stringent starvation media (Kuspa and Kaiser, 1989). (4) Effect of amino acid addition on rescue of sporulation phenotype: the asgD mutant is also different from the asgB mutant in the ways in which it responds to individual amino acids under starvation conditions. For example, on CF medium, the defects of the asgD mutant are not rescued by the addition of phenylalanine, proline or tyrosine, which are found in A-factor and are able to rescue the developmental defects of the asgB mutant (Kuspa et al., 1992b). However, on CF medium, the developmental defects of the asgD mutant are rescued by alanine, leucine and serine, as are those of the asgB mutant. It should be noted that the asgA and asgC mutants also differ from the asgB and asgD mutants. The developmental defects of the asgA mutant are not rescued by the A-factor amino acids at all and the rescue of the asgC mutant by these amino acids is slight (Kuspa et al., 1992b). (5) Gene expression: the asgD gene is not expressed during vegetative growth but expression begins within 1 h of cells being placed under developmental conditions. In contrast, the asgA–C genes are expressed under both vegetative and developmental conditions (Plamann et al., 1994; 1995; Davis et al., 1995). (6) Expression of asg-dependent lac fusions: Tn5 lac-Ω4521 is one of the asg-dependent lac fusions whose expression is defective in the asgA–C mutants (Kuspa et al., 1986). However, Tn5 lac-Ω4521 was expressed normally in the asgD mutant.

The A-signalling hypothesis suggests that asg mutants are defective in A-factor production, resulting in the arrest of development at an early stage. However, the asgD mutant presumably produces enough A-factor to induce expression of Tn5 lac-Ω4521 at the wild-type level. Recently, Mitch Singer and co-workers have also isolated a mutant that appears to belong to the Asg complementation group but still produces enough A-factor to induce expression of Tn5 lac-Ω4521 (M. Singer, personal communication). This suggests that the asgD mutation might not abolish all of A-signalling, but rather interfere with only a part of what is called A-signalling. This suggestion is supported by the rescue experiments described in this paper. The developmental defects of the asgD mutant are rescued by only a subset of the A-factor amino acids, including alanine, leucine and serine (Table 3). The other A-factor amino acids, including phenylalanine, proline and tyrosine, fail to rescue the defects (Table 3). This suggests that A-signalling consists of several separate steps and that the asgD mutant has a defect in only one of these steps (probably a later step as asgD is expressed only under starvation conditions). Tn5 lac-Ω4521 is expected to be regulated by the intermediate steps that are intact in the mutant. Tn5 lac-Ω4521 is expressed normally at 2 h (Kroos et al., 1986). Interestingly, expression of some asg-dependent lac fusions in asgB mutant cells is not rescued by addition of A-factor (i.e. supernatant from starved asg+ cells or 1 mM proline, either of which rescues Tn5 lac-Ω4521 expression in the asgB mutant), although expression of one of these fusions (Tn5 lac-Ω4411) can be rescued extracellulary by wild-type cells (Bowden and Kaplan, 1996). This further suggests the presence of other A-signalling steps besides those defined by the expression of Tn5 lac-Ω4521.

Fruiting body development of M. xanthus is initiated by starvation (Dworkin, 1963; Hemphill and Zahler, 1968) but also requires a solid surface and high cell density (Wireman and Dworkin, 1975). CF agar is a useful starvation medium for M. xanthus because wild-type strains recognize this medium as nutritionally limiting and develop into large mature fruiting bodies; this occurs even though CF agar contains enough nutrients for cells to perhaps double their numbers. As development is blocked for the asgD mutant on the CF medium, the mutant must recognize this medium as suitable for vegetative activities and inhibitory for development. The asgD mutant does not sporulate on CF medium, but does sporulate on the more stringent starvation media such as MMC or CF lacking Casitone, pyruvate or citrate. This suggests that the asgD mutant requires more stringent starvation conditions than does the wild type. We hypothesize that the asgD null mutation causes the threshold for perceived starvation to be lowered (Fig. 7A). Thus, the protein encoded by asgD, either directly or indirectly, must be involved in sensing nutritionally limiting conditions (Fig. 7B). If the levels of nutrients in the environment are higher than the threshold level, the AsgD nutrient-sensing system would block development and the cells would remain as arrested vegetative cells, as are stationary-phase cells. In contrast, if the levels of nutrients are lower than the threshold level, the cells would enter the developmental pathway and form fruiting bodies. We propose that this system is also involved in A-signalling. The complementation experiment with individual amino acids suggests that the A-factor amino acids can be divided into at least two groups. The group 1 amino acids are predicted to function upstream of the AsgD system whereas the group 2 amino acids are predicted to function downstream of the AsgD system (Fig. 7B). Because AsgD is not expressed under vegetative growth conditions but is only expressed just after the cells are placed under starvation conditions, this AsgD function may be part of a gate-keeper pathway for the initiation of development. However, because AsgD is itself regulated by starvation, M. xanthus must have other nutrient-sensing pathways that act at earlier steps and lead to the expression of AsgD. It is likely that M. xanthus has multiple ways to monitor starvation to ensure that cells only enter development at the appropriate time.

Figure 7.

. Proposed model for the function of AsgD. A. The asgD mutation decreases the threshold nutrient level that determines whether the cells continue to develop. The threshold nutrient level of the wild type is marked with ‘wt’ on the slope and the threshold level of the asgD mutant is marked with ‘asgD’. The nutrient level in the CF medium would be below the threshold level of the wild type but above the threshold level of the mutant. B. The AsgD protein is a part of the system that senses nutritional conditions in the environment. This system is also involved in A-signalling. The A-factor amino acids can be divided into at least two groups. The group 1 amino acids are predicted to function upstream of the AsgD system whereas the group 2 amino acids are predicted to function downstream of the AsgD system.

AsgD contains a receiver domain and a histidine protein kinase domain and is predicted to be part of a two-component regulatory system. The AsgD protein has an unusual domain organization compared with other histidine protein kinases but is similar to that of AsgA from M. xanthus, which also has a receiver domain at the N-terminus and a histidine protein kinase domain at the C-terminus. Interestingly, AsgA is also a part of the A-signalling system. Yang and Kaplan (1997) identified another histidine protein kinase, SasS, that is essential for A-signalling. Unlike AsgA and AsgD, SasS does not have a receiver domain but has two transmembrane domains at the N-terminus as well as a kinase domain in the C-terminal half. SasS has been proposed to be involved in responding to extracellular A-signal. It is easy to imagine that the A-signalling system itself consists of several signal transduction pathways as the cells need to sense different environmental factors and integrate them together to produce signals and then respond to the new signals. AsgD, identified in this study, is one of the necessary components involved in A-signalling and nutrient sensing. Study of AsgD should give further insights into the mechanism of this extracellular signalling and environmental sensing system.

Experimental procedures

Bacterial strains, plasmids and culture conditions

The bacterial strains used in this study are listed in Table 5. E. coli TOP10 (Invitrogen) was used for in vitro DNA manipulation. Luria–Bertani broth (LB) was used for growth of E. coli (Sambrook et al., 1989). M. xanthus was cultured vegetatively in CYE, which contains 1% (w/v) Casitone, 0.5% yeast extract, 10 mM MOPS (pH 7.6) and 8 mM MgSO4 (Campos et al., 1978). Development of M. xanthus was initiated by placing concentrated cells on 1.5% agar plates containing CF or MMC medium. The CF medium contains 10 mM MOPS (pH 7.6), 0.015% Casitone, 8 mM MgSO4, 1 mM KH2PO4, 2% sodium citrate and 1% pyruvate (Hagen et al., 1978). The MMC medium contains 10 mM MOPS, 4 mM MgSO4 and 2 mM CaCl2 (Rosenbluh and Rosenberg, 1989). The A1 medium was used as a defined medium (Bretscher and Kaiser, 1978). Liquid cultures were incubated at 32°C with shaking at 275 r.p.m. Solid culture plates were incubated at 34°C.

DNA manipulations and sequence analysis

DNA manipulations were performed using standard protocols (Sambrook et al., 1989). Oligonucleotides were synthesized at Operon Technologies. PCR reactions were carried out with Taq DNA polymerase (Promega) or Pfu polymerase (Stratagene) in the presence of 5% glycerol. Most of the DNA sequencing was carried out at the DBS sequencing facility at University of California-Davis. Gapped blast was used for homology searches (Altschul et al., 1997). DNA and amino acid sequences were analysed using gene inspector (Textco) and lasergene computer software (Dnastar). The asgD nucleotide sequence has been deposited in the GenBank DNA sequence database (accession no. AF16633).

Plasmid construction

Construction of a library of plasmids containing about 500 bp DNA fragments from M. xanthus has been described elsewhere (Cho and Zusman, 1999). pKY482 is a derivative of pZErO-2 and carries a 1.1 kb internal DNA fragment of asgD. To construct pKY482, genomic DNA isolated from DZ4212 was digested with ApaI. The resultant DNA was self-ligated and used to transform E. coli to a kanamycin-resistant strain. pKY483 carries a 7 kb EcoRI DNA fragment from DZ4212, which contains a DNA sequence 6 kb downstream from the C-terminal end of asgD. pKY485 carries a 5 kb NotI DNA fragment from DZ4212, which contains a DNA sequence 3 kb upstream from the N-terminal end of asgD. pKY483 and pKY485 were constructed by a method similar to the pKY482 construction except EcoRI and NotI were used, respectively, instead of ApaI to digest the genomic DNA. pKY482, pKY483 and pKY485 were used to determine the DNA sequence of asgD and flanking regions.

pKY468 is a 2.17 kb plasmid vector containing a colE1 replication origin, a kanamycin-resistant gene and a multicloning site from pBend3 (Zwieb and Brown, 1990). pKY480 is a 4.8 kb plasmid vector that is similar to pKY468 except that it contains a sacB gene from pKNG101 (Kaniga et al., 1991). pKY481 is a 8.4 kb plasmid vector that is similar to pKY468 except that it contains the lacZY genes from pMC1403 (Casadaban et al., 1980).

pKY565 is a derivative of pKY481 carrying a translational fusion between codon 13 of asgD and codon 8 of lacZ. The fusion was first created in vitro by cloning a PCR fragment containing a region between 653 bp upstream and 37 bp downstream from the asgD translation start codon into XhoI and BamHI sites of pKY481. The PCR fragment was amplified using two oligonucleotides 5′-GCACTCGAGCACGCGCTTC TACCGCTACGA-3′ and 5′-GACGGATCCTGCAACGGCAC CATCGTCGG-3′ as primers and pKY485 as a DNA template. pKY568, which was used to create an asgD in frame deletion, was constructed by the reported method (Link et al., 1997). Four oligonucleotides were designed to amplify a PCR fragment carrying an asgD in frame deletion: 5′-GACGGATCCACGCGCTTCTACCGCTACGAC-3′ (N1), 5′-ACGGTCGC AGTACGATCCTGTACGGCACCATCGTCGGAATCC-3′ (N2), 5′-ACAGGATCGTACTGCGACCGTGGCTCTACATCGTGC AGGACA-3′ (C1), and 5′-GACCTGCAGTGGAGGCAGGAC CCAATCGTG-3′ (C2). Primers N1 and N2 were used to amplify an N-terminal PCR fragment, whereas primers C1 and C2 were used to amplify a C-terminal PCR fragment. The resulting two PCR products and primers N1 and C2 were then used to amplify a final PCR product carrying an asgD in frame deletion. The final PCR fragment was then ligated into pKY480 after it was digested with BamHI and XhoI to generate pKY568.

Mutation of M. xanthus

Plasmid DNAs were introduced into M. xanthus by electroporation with the conditions described previously (Kashefi and Hartzell, 1995). All the plasmids used in this study cannot replicate autonomously in M. xanthus. Thus, selection of transformants against antibiotics allowed growth of only cells carrying a plasmid that has recombined into the chromosome. Because M. xanthus DZ2 has a tendency to clump after electroporation, the transformed cells were selected by growing in a CYE broth containing 100 μg ml−1 kanamycin for 3 days. The cells were then diluted and plated on CYE agar plates containing kanamycin. The insertion mutations were characterized by PCR amplification or by restriction mapping after DNA fragments containing the insertions were cloned. Replacement of Tn5-wt (KmR) with Tn5-132 (TetR) was carried out as previously described (Avery and Kaiser, 1983).

The asgD in frame deletion mutant was created by counterselecting DZ4234 against 5% sucrose as described previously (Wu and Kaiser, 1996). DZ4234 contains an insertion of the plasmid pKY568, which carries a sacB gene, a kanamycin-resistant gene and an asgD in frame deletion. Because of the sacB gene, DZ4234 cannot grow in the presence of sucrose. As a result, the counterselection against sucrose allows only cells that have lost the sacB gene by excision of the plasmid. When the plasmid is excised from the chromosome, about half of the cells excise the wild-type asgD gene and retain the asgD in frame deletion as a result of homologous recombination, resulting in asgD in frame deletion mutants. After counterselection against sucrose, such in frame deletion mutants were screened based on their developmental phenotype and kanamycin sensitivity. Then, the deletion was confirmed by PCR amplification.

Mutant characterization

Fruiting bodies and individual cells were observed with a Zeiss microscope (Model 47 60 09-9901) and a Nikon Labphot-2 microscope respectively. Images were captured with a Dage-MTI CCD-72 series camera and saved as a computer file using NIH Image (Research Services Branch of the National Institute of Mental Health, USA). β-Galactosidase activity was assayed according to a protocol described previously (Kroos et al., 1986). Sonication-resistant spores were visually counted using a haemocytometer (Hausser Scientific) after the harvested cells were sonicated to disperse clumps and disrupt vegetative cells. Heat-resistant spores were counted based on their viability on CYE plates after treatment with heat (50°C for 2 h) and sonication in water. Crude A-factor was prepared with a method reported by Kuspa et al. (1986). Cell–cell cohesion was measured according to the method of Shimkets (1986).

Acknowledgements

We thank Kathleen A. O'Connor, Mandy J. Ward and other members of the Zusman laboratory for helpful discussions and suggestions, and Mitch Singer (University of California at Davis) for sharing his unpublished data with us. We also thank Heidi B. Kaplan and Gabriela Bowden (University of Texas at Houston) for providing strains. This work was supported by NIH grant GM20509.

Ancillary