nsd, a Locus That Affects the Myxococcus xanthus Cellular Response to Nutrient Concentration

Abstract

Expression of the previously reported Tn5lac Ω4469 insertion in Myxococcus xanthus cells is regulated by the starvation response. Interested in learning more about the starvation response, we cloned and sequenced the region containing the insertion. Our analysis shows that the gene fusion is located in an open reading frame that we have designated nsd (nutrient sensing/utilizing defective) and that its expression is driven by a σ⁷⁰-like promoter. Sequence analysis of the nsd gene product provides no information on the potential structure or function of the encoded protein. In a further effort to learn about the role of nsd in the starvation response, we closely examined the phenotype of cells carrying the nsd::Tn5lac Ω4469 mutation. Our analysis showed that these cells initiate development on medium that contains nutrients sufficient to sustain vegetative growth of wild-type cells. Furthermore, in liquid media these same nutrient concentrations elicit a severe impairment of growth of nsd cells. The data suggest that the nsd cells launch a starvation response when there are enough nutrients to prevent one. In support of this hypothesis, we found that, when grown in these nutrient concentrations, nsd cells accumulate guanosine tetraphosphate, the cellular starvation signal. Therefore, we propose that nsd is used by cells to respond to available nutrient levels.

When nutrients are depleted from the soil environment of a Myxococcus xanthus colony, the gram-negative bacteria initiate a complex multicellular developmental program that culminates in the formation of a fruiting body. This developmental response is part of a survival mechanism and is coupled to a cellular differentiation process that converts rod-shaped vegetative cells into stress-resistant, dormant myxospores. In response to this environmental signal (nutrient depletion), about 10⁵ cells undergo extensive changes in gene expression and coordinate their behaviors and movements in order to construct the spore-filled fruiting body (9).

Progression through the complex developmental process requires extensive cell-cell interaction and communication (13, 25). An important part of the intercellular exchange occurs through production and reception of cell-cell signals (27). The A-signal is thought to act very early in development because it is required for full expression of nearly all of the identified developmentally regulated genes (27). However, expression of a few known developmental genes is regulated as part of the early developmental response independent of A-signal. These genes are likely part of the system that individual cells use to detect starvation and to begin their planned response to this condition.

Nutrient availability governs the decision to leave vegetative growth and enter the developmental survival response. As such, the mechanisms for accurately sensing environmental and internal nutrient levels are central to the regulation of entry into development. The nutrient-sensing process relies on several general cellular processes, including extracellular detection of nutrients, transport of nutrients into the cell, proper metabolism of nutrients, and activation or repression of signal transduction pathways that result in developmental gene expression. A breakdown in any of these vital processes may cause faulty nutrient sensing, blocking the cell's ability to correctly detect starvation and, therefore, its ability to accurately regulate entry into development.

The primary source of nutrients for M. xanthus cells is amino acids (4, 8), and, when they become unavailable, M. xanthus cells launch an early starvation response that mimics the stringent response in Escherichia coli (for a review, see reference 5), triggering immediate accumulation of guanosine penta- and tetraphosphate [(p)ppGpp and ppGpp, respectively] (31). Based on analysis of mutants, we know that (p)ppGpp signaling is both necessary and sufficient for initiation of development (36). A mutation in the relA gene prevents synthesis of (p)ppGpp, and colonies of cells carrying this mutation do not undergo the distinct morphological changes that are characteristic of development (36). Furthermore, these mutants do not express known developmental genes, nor do they produce A-signal (15). Without the capability to synthesize (p)ppGpp, cells are unable to initiate development, and we propose that this is due to an inability to sense starvation.

In addition to relA, several other loci have been implicated in nutrient sensing and regulation of entry into development. Mutations in crdB, mcp3A, mcp3B, and cheA3, genes in the Che3 cluster, cause a temporal defect in initiation of development, causing the formation of fruiting bodies 12 to 24 h earlier than in wild-type cells and accelerating the expression profile of developmentally regulated genes (22). Furthermore, the mcp3B mutant initiates the developmental program in the presence of nutrients, forming early-stage fruiting bodies and some spores when grown on nutrient-rich agar medium (22). Another gene, socE, inhibits the stringent response and its associated accumulation of (p)ppGpp, preventing cells from entering development in the presence of nutrients. Even when grown in nutrient-rich broth, cells depleted of SocE stop growing, accumulate (p)ppGpp, and induce sporulation (7). It is likely that these identified genes are only a few of many genes involved in the early starvation response and regulation of entry into development.

One of the previously identified genes whose developmental expression is (p)ppGpp dependent and A-signal independent is the locus defined by the Tn5lac Ω4469 fusion (25, 26, 36). This gene fusion was initially characterized as part of a collection of developmentally regulated Tn5lac reporter fusions (26). Here, we report our further characterization of the expression of Ω4469. We cloned and sequenced the region surrounding Tn5lac Ω4469 and identified the open reading frame (ORF) carrying the insertion mutation, now designated nsd (nutrient sensing/utilizing defective). Here, we show the effect of the nsd::Tn5lac Ω4469 mutation on the cells' ability to appropriately respond to the level of available nutrients and we hypothesize that nsd is involved in regulating entry into development by governing how cells respond to nutrient availability. Furthermore, we propose that the phenotype resulting from the nsd::Tn5lac Ω4469 mutation is part of a starvation response mediated through production of (p)ppGpp.

MATERIALS AND METHODS

Bacterial strains and plasmids.

Strains and plasmids used in this study are listed in Table 1. Recombinant DNA work was performed by standard techniques (34). Unless otherwise stated, molecular biology reagents were purchased from Promega (Madison, Wis.). All plasmids were routinely maintained in DH5α.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Source or reference
E. coli strain DH5α	supE44 ΔlacU169 φ80ΔlacZM15 hsdR17 recA1 endA1 gyrA96 thi-1 relA1	14
M. xanthus strains
DK1622	Wild type	20
DK4469	DK1622 Tn5lac (Kanr) Ω4469	26
DK7825	DK1622 Tn5lac (Tetr) Ω4469	Kaiser laboratory
MS801	DK7825 lacZ::pTT1	This study
Plasmids
pBGS18	Kanr	37
pRS415	AmprlacZ	R. Simons
pTT1	Kanr (pBGS18); 2.6-kb BcII fragment from pRS415 (containing lacZ) combined with BamHI-digested pBGS18	This study
pTEA1	Kanr (pBGS18); 6.1-kb KpnI fragment from MS801	This study
pTEA5	Kanr (pBGS18); 2.1-kb KpnI-BamHI fragment from pTEA1	This study

Media for growth and development.

E. coli cells were grown at 37°C in Luria-Bertani (LB) medium or on plates containing LB medium and 1.5% agar (34), supplemented with 50 μg of ampicillin/ml or 50 μg of kanamycin sulfate/ml, as necessary. M. xanthus strains were grown at 32°C with vigorous shaking in nutrient-rich CTT broth (1% Casitone [Difco], 10 mM Tris-HCl [pH 7.6], 1 mM KH₂PO₄, 8 mM MgSO₄) or on CTT plates containing 1.5% agar (18). In some assays, the Casitone concentration was reduced from 1 to 0.5 or 0.25%. For developmental assays, either CF (0.015% Casitone [Difco], 10 mM Tris-HCl [pH 7.6], 1 mM KH₂PO₄, 8 mM MgSO₄, 0.02% [NH₄]₂SO₄, 0.1% pyruvate, 0.2% citrate) (13) or TPM (10 mM Tris-HCl [pH 7.6], 1 mM KH₂PO₄, 8 mM MgSO₄) (4) plates containing 1.5% agar were used and cells were incubated at 32°C. Myxospores were suspended in CTT soft agar (CTT broth containing 0.7% agar) before being placed onto CTT agar plates. The defined minimal medium used was A-1 plus 0.8% agarose for use as solid medium (4).

β-Galactosidase assays.

Vegetative samples were harvested from M. xanthus cell cultures grown in CTT broth at 32°C with vigorous swirling. At various times, cell density was measured with a Klett-Summerson photoelectric colorimeter and 500-μl aliquots were removed and quick-frozen in liquid nitrogen. For starvation of M. xanthus, cells were grown in CTT broth to a density of 5 × 10⁸ cells/ml, a 1-ml aliquot was removed and stored as the time-zero (t₀) sample, and the remaining cells were harvested by centrifugation. The resulting cell pellet was washed with an equal volume of TPM buffer and then resuspended to a final concentration of 5 × 10⁹ cells/ml in TPM buffer. This culture was incubated at 32°C with vigorous swirling. At various times, 100-μl aliquots were removed and added to 400 μl of TPM buffer and the mixture was quick-frozen in liquid nitrogen. β-Galactosidase assays were performed with quick-frozen extracts, as previously described (21). β-Galactosidase specific activity is defined as nanomoles of o-nitrophenol produced per minute per milligram of protein.

Cloning the nsd locus.

To clone the region upstream from Tn5lac, we first integrated the plasmid pTT1 at the locus, via homologous recombination with lacZ. pTT1 contained a 2.6-kb BclI fragment from pRS415 (containing lacZ) cloned into the BamHI site in pBGS18. We introduced pTT1 into DK7825 (Ω4469 Tn5lac Tet^r) by electroporation (33) and selected Km^r Tet^r colonies on CTT agar plates. We confirmed the single-copy integration event at the Tn5lac Ω4469 site by Southern blotting, using lacZ as the probe. We then used the resulting strain, MS801, to excise and clone the region upstream of Tn5lac. We isolated MS801 genomic DNA (29) and digested it with KpnI. This pool of genomic DNA fragments was self-ligated and introduced into E. coli by electroporation. We selected Km^r colonies and used restriction digests to screen for the appropriate 9.8-kb plasmid containing 3.7 kb of the vector (pBGS18), 4.0 kb of Tn5lac, and 2.1 kb of DNA from upstream of the Tn5lac insertion site. The resulting plasmid, pTEA1, was selected for sequencing. The 2.1-kb KpnI-BamHI fragment was subcloned into pBGS18 to form pTEA5. The cloned region spanned from the 3′ KpnI site to the BamHI site 50 bp downstream from the 5′ end of Tn5lac.

DNA sequencing analysis.

The DNA sequence was determined by the Division of Biological Sciences facility at the University of California, Davis. The sequence was analyzed with ABI Prism software and assembled with Deneba Sequencher software. Custom-designed oligonucleotide primers were synthesized by Operon Technologies, Inc. (Alameda, Calif.).

Analysis of RNA.

Vegetative RNA was isolated from cells growing in exponential phase (2.5 × 10⁸ to 6 × 10⁸ cells/ml) in CTT broth at 32°C with vigorous swirling. Starvation RNA was isolated from cells grown in TPM buffer, as described above, for 12 h. Developmental RNA was isolated from cells after 12 h of development on TPM agar (21). RNA was extracted from quick-frozen cell extracts by the hot-phenol method (34). Total cellular RNA was analyzed by slot blotting (21), using a 309-bp SphI-HincII fragment from the 5′ end of the nsd ORF excised from pTEA5. The specificity of the probe was confirmed with Saccharomyces cerevisiae tRNA, which yielded almost no signal. The signal was imaged and quantified with the Storm 840 phosphorimager (Molecular Dynamics) and the ImageQuant, version 5.0, software package (Molecular Dynamics).

Primer extension analyses.

Primer extensions were carried out as described previously (12). Vegetative (mid-exponential phase; 5 × 10⁸ cells/ml) and developmental (12-h) RNA samples were prepared as described above. The primer used for this study, PE1 (5′-CGTTCGACATCCTCATCGACCAGCAT-3′), was custom designed and synthesized by Genosys (Sigma).

Development and sporulation assays.

To examine development and to collect myxospore samples, M. xanthus cells were grown to mid-exponential phase (5 × 10⁸ cells/ml), harvested by centrifugation (7,700 × g for 5 min), washed twice with TPM buffer, and spotted in 20-μl aliquots as previously described (26). Fruiting body formation was monitored visually with a dissecting microscope (Nikon; SMZ800), and images were recorded (Nikon; Coolpix995). Sporulation efficiencies were determined both by counting the number of phase-bright spores with a Petroff-Hausser counting chamber and phase-contrast microscopy and by determining the viability of these spores based on their ability to form cell colonies on CTT agar, as described previously (40).

Cell growth assays.

To monitor growth in CTT, 0.5% CTT, and 0.25% CTT, cell density was monitored with a Klett-Summerson photoelectric colorimeter and recorded in arbitrary Klett units. Cells were grown to mid-exponential phase (Klett 100; 5 × 10⁸ cells/ml) in CTT and then diluted to 10 Klett units in the appropriate medium. This culture was allowed to reach mid-exponential phase then diluted once more to 10 Klett units in the same medium. Growth was then monitored and recorded into stationary phase. To determine cell yield on agar plates, cells were grown and diluted as described above. Once they reached mid-exponential phase in the appropriate medium, the cells were spotted in 20-μl aliquots onto the corresponding agar plates and allowed to dry. Individual spots were scraped off the plates, and each was resuspended in 200 μl of water. Spots were collected immediately after spotting and drying for the t₀ sample and after 96 h for the t_F (final) sample. The protein concentration was assayed (3), and cell yield was determined by calculating the protein concentration by subtracting the concentration in the t₀ sample from that in the t_F sample. To measure swarm expansion, cells were spotted onto the appropriate agar plates as just described and the swarm diameter was measured at various time intervals with the aid of a dissecting microscope (Nikon; SMZ800).

Determination of ppGpp accumulation.

In vivo analysis of guanosine nucleotides was performed as described by Manoil and Kaiser (30), except as follows. Cells were grown to mid-exponential phase in CTT and then diluted to 10 Klett units in 0.5% CTT. After one cell doubling, 100 μCi of [³²P]orthophosphate (Perkin-Elmer)/ml was added to the cultures to radiolabel the cells for determination of ppGpp accumulation. The cells were allowed to grow overnight to equilibrate the intracellular phosphate pools. Cell growth was monitored, and samples were removed for determination of ppGpp concentrations. Nucleotides were extracted and analyzed by thin-layer chromatography as previously described (30). The signal generated in each nucleotide spot was detected with a Storm 840 PhosphorImager (Molecular Dynamics) and quantitated with the ImageQuant, version 5.0, software package (Molecular Dynamics).

Nucleotide sequence accession number.

Sequence data for nsd were submitted to GenBank under accession number AY325890.

RESULTS

Expression pattern of nsd.

Previous studies demonstrated that Tn5lac Ω4469 expression is increased in a development-dependent manner beginning about 4 h after removal of nutrients (26). We examined expression of Tn5lac Ω4469 in DK4469 cells during vegetative growth and development, using β-galactosidase specific activity as a transcription reporter. Under vegetative conditions with DK4469 cells grown in CTT broth, β-galactosidase specific activity remains relatively constant throughout exponential phase and into stationary phase (Fig. 1). As noted by Kroos et al., when DK4469 cells are shifted to developmental conditions, the β-galactosidase specific activity begins to increase within about 4 h and continues to increase until about 12 to 24 h, when the expression level peaks at a fourfold increase over the vegetative level (26). These data indicate that expression of Tn5lac Ω4469 occurs throughout vegetative growth and increases early in the developmental program.

FIG. 1.

Patterns of Tn5lac Ω4469 expression. The mean β-galactosidase specific activity was determined by at least three independent experiments. Error bars represent the standard deviations of the means. •, β-galactosidase specific activity from DK4469 cells during vegetative growth in CTT broth; ▴, cell density recorded in arbitrary Klett units.

Because cells must be cultured on a solid surface for the developmental program to proceed, the process of starvation can be uncoupled from that of development (24). When grown in liquid starvation medium (TPM buffer), cells undergo a starvation response and the genes required for this process are activated. Without the presence of a solid surface, however, development cannot follow starvation, and many developmentally regulated genes are not activated in this situation (24). Developmental genes that are regulated by starvation in TPM buffer are regulated by the starvation response, and this regulation occurs independently of the developmental program. To determine whether nsd expression is regulated by starvation, we assayed β-galactosidase specific activity of DK4469 cells placed on solid TPM agar plates and compared this to the activity of cells placed in liquid TPM buffer. The β-galactosidase specific activity induced by development on a solid surface mirrors that induced by liquid starvation conditions (data not shown). These findings demonstrate that a starvation response induces a fourfold increase in Tn5lac Ω4469 expression, and this increase occurs independently of development on a solid surface.

Cloning and analysis of the nsd locus.

The expression pattern of nsd indicates that it is part of the starvation response, and, because we are interested in further defining the starvation response, we would like to investigate the function of nsd. Often clues to molecular function can be found by evaluating gene sequence, so we cloned a 2.1-kb region 5′ to nsd and analyzed its sequence for any detectable functional motifs or characterized gene homologues, as described in Materials and Methods. This region spans from the upstream KpnI site to the BamHI site 50 bp downstream from the 5′ end of Tn5lac (Fig. 2). The Tn5lac insertion site is located 131 bp from the 3′ end of a 0.5-kb ORF that is predicted to be transcribed in the same direction as the lacZ gene in the Tn5lac Ω4469 insertion; we designated this ORF nsd. The amino acid sequence deduced from the ORF is similar (40 to 47%) to that of a hypothetical protein found in Novosphingobium aromaticivorans, Corynebacterium efficiens, Azobacter vinelandii, and Nostoc sp. strain PCC 7120. In each case, the protein was identified as part of a genome sequencing effort and has yet to be experimentally characterized, so we cannot assign any sequence-based putative functions to Nsd.

FIG. 2.

Physical map of the nsd locus. Boxes, indicated ORFs (arrows inside indicate the predicted direction of transcription). The site and direction of the Tn5lac (large triangle) insertion were determined by DNA sequencing and Southern blotting, as described in Materials and Methods. Bent arrow, location of the vegA promoter (23); bar labeled pTEA5, cloned region; stippled bar, nsd probe fragment used in RNA slot blot analysis.

Transcribed in the opposite orientation to nsd is orf1, and its putative start codon is located 120 to 140 bp upstream from that of nsd (Fig. 2). There are no known homologues or functional domains identified for orf1, based on BLAST analysis (National Center for Biotechnology Information, National Library of Medicine).

We found that the downstream region contains the previously characterized vegA gene (23). This gene has its own well-characterized σ⁷⁰-like promoter (23), which lies between the predicted 3′ end of nsd and the 5′ end of vegA. In contrast to that of nsd, vegA expression is notably higher in cells growing vegetatively than in developing cells, and a null mutation in vegA is lethal (23). Because their expression profiles differ significantly and because the Tn5lac insertion into nsd is not lethal, we reasoned that the Tn5lac insertion into nsd does not result in loss of vegA expression and that the nsd promoter does not drive expression of vegA. Collectively, the sequence analysis indicates that nsd is the gene containing the Tn5lac Ω4469 insertion and that nsd comprises a one gene transcriptional unit.

Localizing the 5′ end of the nsd transcript.

Because nsd is expressed vegetatively and because its expression changes as part of a starvation response, it is possible that there are two promoters for nsd: one that operates under vegetative conditions and one that operates under starvation conditions. Conversely, there may be a single promoter whose activity is modulated in response to starvation. To determine which of these possibilities is the case, we examined the 5′ end of the nsd mRNA from cells grown under both vegetative and starvation conditions. We performed primer extension using RNA isolated from DK1622 cells grown to mid-exponential phase in CTT broth and from DK1622 cells that had been allowed to develop for 12 h (the time when peak β-galactosidase specific activity is observed). The primer PE1 (Fig. 3A) detected a 5′ mRNA end that maps to a guanosine nucleotide (Fig. 3B). The same 5′ end was detected with both the vegetative and developmental template RNA, and we have designated this nucleotide as the transcriptional start site (TSS) (Fig. 3A and B). These data suggest that there is a single nsd promoter that is regulated in order to achieve a constant low level of expression during vegetative growth and an increased level during a starvation response.

FIG. 3.

Sequence analysis of the region upstream from nsd. (A) Sequence of the region upstream of nsd. Bent arrow, predicted start site of the nsd transcript. The putative promoter elements are underlined at the indicated −10 and −35 regions. The predicted start codon for nsd (+30) is in boldface. The sequence complementary to PE1, the primer used to identify the 5′ end of the transcript, is underlined (+40 to +65). (B) Mapping the 5′ end of the nsd transcript by primer extension analysis. A, G, C, and T show the DNA sequencing ladders. Arrows at PE_v and PE_d lanes indicate each product resulting from primer extension with primer PE1 and total RNA prepared from either vegetative (PE_v) or developing (PE_d) DK1622 cells, as described in Materials and Methods. (C) Comparison of the proposed nsd promoter with the E. coli σ⁷⁰consensus sequence and with three known M. xanthus σ⁷⁰ promoters: those for vegA (23), tps (19), and ops (19).

We analyzed the region upstream of the putative TSS and identified a sequence similar to that of σ⁷⁰ promoters (Fig. 3A and C). There are four out of six matches to the E. coli σ⁷⁰ consensus sequence in both the −10 and −35 regions (16). We did not observe similarities to other known types of bacterial promoters. This σ⁷⁰-like promoter appears to drive expression of nsd under vegetative conditions and is activated to increase expression during the starvation response.

nsd expression is not autoregulated.

Our sequence analysis showed that Tn5lac Ω4469 is inserted into the nsd ORF and may, therefore, alter the expression of nsd if the gene is autoregulated. To determine whether the insertion into nsd results in a change in expression of nsd from that observed in a wild-type strain, we directly examined nsd mRNA levels in the Tn5lac Ω4469 strain, DK4469, and in the wild-type parent strain, DK1622.

The RNA slot blots demonstrated that, under vegetative and developmental conditions, the nsd message present in DK4469 (nsd) cells is approximately equal to that in DK1622 (nsd⁺) cells (Fig. 4). This shows that the insertion mutation does not alter the regulation of gene expression of the native locus. Furthermore, our data confirm that the β-galactosidase specific activity levels correlate with nsd mRNA levels and that nsd::Tn5lac is, therefore, a reliable reporter of nsd expression.

FIG. 4.

Levels of nsd mRNA present in wild-type and Ω4469 cells. The mean RNA levels were determined by at least three independent experiments. Error bars represent standard deviations of the means. Total RNA was isolated from DK1622 (solid bars) and DK4469 (open bars) cells grown under growth, starvation, or developmental conditions. The levels of nsd mRNA were normalized to the amount present in the vegetative DK1622 sample, which was set equal to 1.

nsd initiates the developmental program on nutrient agar plates.

It was previously reported that nsd cells respond to standard laboratory starvation conditions by forming fruiting bodies and sporulating similarly to the wild type (26). We assayed DK4469 and its wild-type parent strain, DK1622, for the ability to form fruiting bodies on TPM (data not shown) and CF agar plates (Fig. 5) and confirmed these findings. However, there is a subtle phenotype in cells carrying the nsd mutation: these cells appear to form a greater number of fruiting bodies that are slightly smaller in size than those formed by wild-type cells (Fig. 5, insets).

FIG. 5.

Development and growth of wild-type and nsd cells. Wild-type DK1622 (A) and nsd::Tn5lac Ω4469 DK4469 (B) cells were compared for development on CF agar. Cells were photographed under a dissecting microscope after 72 h of incubation. Images shown were recorded at 10× magnification; the inset at ×50 shows mature, darkened fruiting bodies. (Bottom) Vegetatively growing cells were photographed at 10× magnification after 24 h of growth on CTT agar.

When grown on nutrient-rich 1% CTT for up to 96 h, nsd cells behave as the wild type, with respect to colony morphology and rate of swarm expansion (Fig. 6A). However, by 96 h the flares of nsd cells reaching out from the spot appear very textured and loose aggregates and early mounds of cells form throughout the periphery of the original spot, presumably where the colony has used up most of the available nutrients (data not shown). This phenotype suggests that these cells are attempting to initiate the developmental program while growing on 1% Casitone, nutrient conditions that do not normally induce development. Instead of initiating development, wild-type colonies continue to expand outward, presumably in search of more nutrients. We also noticed that when streaked onto rich medium nsd cells often formed ripples, which are patterns of coordinated cell movements often observed in the early stages of fruiting body formation (35). Based on these observations, we examined the effects of varying nutrient (Casitone) levels on the growth and the development of nsd cells, compared to the wild type.

FIG. 6.

Swarm expansion on nutrient agar plates. Cells were placed in 20-μl aliquots at a density of 5 × 10⁹ cells/ml onto agar plates containing 1% (A), 0.5% (B), or 0.25% (C) Casitone. We measured the diameter of each spot at the indicated time intervals with the aid of a dissecting microscope. Solid bars, sizes of wild-type swarms; open bars, sizes of nsd::Tn5lac Ω4469 swarms. The swarm diameter is given as the average of four independent experiments, and error bars represent the standard deviations of the means.

Growth on 0.5% and 0.25% Casitone agar.

We reduced the Casitone concentration twofold and grew both strains on agar plates containing 0.5% Casitone (Fig. 7). Wild-type cells continue to grow. After 24 h, the spot containing DK1622 cells appears flat, indicating that cells have not initiated the developmental program. They enter the mound stage of development at about 96 h (Fig. 7A). Cells continue to swarm out from the edge of the spot, and mound formation moves outward from the center of the spot over the course of time (Fig. 6B). This pattern of initiation of the developmental program indicates that, as growing cells use up the available nutrients, they begin to form fruiting bodies. The mounds in the central spot develop into large, slightly less dense fruiting bodies. The morphology of the fruiting bodies that form out from the edge of this spot more closely resembles that seen when cells are starved under standard conditions.

FIG. 7.

Growth and development on 0.5% CTT agar. Wild-type DK1622 (A) and nsd::Tn5lac Ω4469 DK4469 (B) cells were plated and cultured on 0.5% CTT agar. Growth and development were monitored, and cells were photographed every 24 h at 10× magnification. Insets show mound and fruiting body morphology after 96 h at 50× magnification.

In contrast, after 24 h on 0.5% Casitone agar plates, the nsd cells appear to have entered the early mound stage of development (Fig. 7B). Notably, the nsd spot has a uniform appearance, indicating that the entire population of cells has initiated the developmental program at about the same time and the same rate. The swarm does not appear to increase in size significantly over the course of the experiment, unlike wild-type cells, which continue to expand the swarm (Fig. 6B). These data suggest that the entire population of nsd cells respond to the nutrient concentration at about the same time and cease to grow very soon after being plated onto the medium. Over time, the nsd mounds become more compact and darken (Fig. 7B, inset), but they do not mature into fruiting bodies with the same morphology as fruiting bodies formed on standard starvation agar medium (Fig. 5B, inset).

A further twofold reduction in the Casitone concentration, to 0.25% (Fig. 8), caused wild-type cells to form mounds at about 24 h, but these are present only in the center of the spot (Fig. 8A). The cells at the edge of the spot continue to swarm outward for the duration of the experiment (Fig. 6C). Over time, the region of fruiting body formation moves out from the center as the expanding swarm uses up the nutrients, similar to the observed behavior of wild-type cells on 0.5% Casitone. In contrast, by 24 h, the nsd cells have developed to the late mound stage (Fig. 8B). As observed when grown on 0.5% Casitone agar plates, the mounds form simultaneously throughout the entire spot and there is little swarming out from the initial edge of the spot (Fig. 6C). This pattern of fruiting body formation mirrors that seen when cells are grown on starvation media, such as CF, except that here the resulting fruiting bodies are irregular in size and shape (Fig. 8B, inset).

FIG. 8.

Growth and development on 0.25% CTT agar. Wild-type DK1622 (A) and nsd::Tn5lac Ω4469 DK4469 (B) cells were plated and cultured on 0.25% CTT agar. Growth and development were monitored, and cells were photographed as for Fig. 7.

nsd is altered in spore formation.

We have shown that nsd cells initiate development under conditions that normally support growth. To determine whether these cells are able to complete the developmental process under these conditions, we examined their ability to form viable spores. nsd spores isolated from starvation agar were present in about the same numbers as wild-type spores, although the viability of nsd spores was somewhat reduced from that of wild-type spores (Table 2).

TABLE 2.

Sporulation efficiencies of wild-type and nsd::Tn5lacΩ4469 strains^a

Medium	Strain	% of wild-type:
		Sporesb	Viable sporesc
CF (0.015% Casitone)	DK1622	100.0 ± 8.8	100.0 ± 21.9
	DK4469	111.0 ± 7.0	56.7 ± 19.4
0.25% CTT	DK1622	100.0 ± 3.4	100.0 ± 19.7
	DK4469	87.4 ± 5.8	10.1 ± 10.6

^aCells were placed on the indicated medium and incubated for 4 days. Sporulation efficiency was measured by direct spore counts and viability assays. Spore assays were performed three times for each strain. Each mean value (+/− standard deviation) is expressed as a percentage of that of the wild type (DK1622).

^bNumber of heat- and sonication-resistant spores counted on a Petroff-Hausser chamber under phase-contrast microscopy.

^cNumber of colonies that arose from heat- and sonication-resistant spores transferred to CTT agar plates after 5 days of incubation.

When spores were isolated from agar plates containing 0.25% Casitone, both nsd and wild-type cells made spores. The number of nsd spores was slightly lower than that of wild-type spores, but the viability of these nsd spores was significantly reduced (Table 2). nsd cells grown on 0.25% Casitone agar produced heat- and sonication-resistant ovoid spores with a 90% reduction in viability from that of wild-type spores. The data suggest that, while the nsd mutation causes cells to initiate development in the presence of nutrients, the cells are unable to successfully complete the developmental program. When grown on 1% and 0.5% CTT, neither wild-type nor nsd cells produced viable spores but the nsd cells produced heat- and sonication-resistant cells. Unlike myxospores, these cells were not phase bright, their shapes were varied, and they were unable to germinate into cell colonies.

nsd does not alter the cell density requirement for development.

Cell density is an important component of the developmental process, and it has been shown that high cell density is necessary for development to proceed (42). For this reason, we investigated whether the nsd mutation affects the developmental requirement for high cell density. We spotted cells onto agar plates containing various nutrient concentrations, from 0 to 1%, as described above, but we varied the concentration of cells spotted onto each agar plate in a twofold dilution series from 5 × 10⁹ (used in standard developmental assays) to 5 × 10⁶ cells/ml. We found that cell density affected the development of nsd cells analogously to wild-type cells: at high cell density on starvation media, both strains develop normally; as cell density decreased, development was increasingly more impaired in both strains (data not shown).

nsd affects the doubling time and peak biomass under low-nutrient conditions.

To investigate the effect of nutrient concentration on cell growth, we examined cells' ability to grow in liquid media with various Casitone concentrations. The developmental program requires a solid surface with which cells interact. By growing cells in liquid media, we uncouple the developmental process from the starvation process (24). This assay examines the cells' response to nutrient concentration in the absence of the developmental program.

When grown in 1% CTT broth, both wild-type and nsd cells grow at about the same rate, doubling approximately every 5 h but the nsd peak biomass is only about 80% of wild type (Fig. 9A). However, when the nutrient concentration is reduced twofold, the doubling times of the two strains vary considerably (Fig. 9B): DK1622 cells double approximately every 8.5 h, while nsd cells double every 17 h. In addition, the nsd cells' peak biomass is only about 64% of that of wild-type cells.

FIG. 9.

Cell growth in liquid nutrient media. Cells were cultured in liquid medium with 1% (A), 0.5% (B), or 0.25% (C) Casitone, as described in Materials and Methods. •, wild-type cell density; ▴, nsd::Tn5lac Ω4469 cell density.

With another twofold decrease in nutrient concentration, to 0.25% Casitone, the generation times of both wild-type and nsd cells are further lengthened. Wild-type cells double in almost 9 h, while the nsd cells double in 22.4 h (Fig. 9C). This 2.5-fold increase in doubling time is accompanied by a 50% reduction in peak biomass, compared to wild-type cells grown in the same medium.

In liquid media, nutrient concentration appears to affect nsd peak biomass, and we also observed a nutrient-dependent effect on the cell yield of nsd cells on solid media (Fig. 10). We placed the same number of nsd and wild-type cells onto agar plates, allowed the cells to grow for 4 days, and then determined the increase in biomass of each strain, as reflected by their protein concentrations. When grown on agar medium containing either 0.5% or 0.25% Casitone, nsd cells achieved only about 65% of the wild-type cell yield. On 1% CTT agar medium, however, the two strains reached about the same yield (Fig. 10). This demonstrates that nutrient concentration affects nsd cell growth on solid media, as we have also shown that it does in liquid media.

FIG. 10.

Change in cell yield on solid media. Cells were grown on the indicated agar medium for 3 days. The increase in protein concentration was calculated by subtracting the initial protein concentration (cells removed from the plate immediately after spotting) from the final protein concentration (cells removed from the plate after 3 days of growth). Solid bars, wild type; open bars, nsd::Tn5lac Ω4469. Each data point is the average of three independent experiments; error bars show the standard deviations of the means.

nsd does not cause auxotrophy.

The illicit response to nutrient concentration that is seen in nsd cells could be due to an auxotrophic metabolic disorder. To test this possibility, we grew cells on defined minimal medium, A-1, which contains only those nutrients known to be necessary for growth of wild-type M. xanthus cells (4). nsd cells grow on A-1 medium (data not shown), demonstrating that the defective response to nutrient concentration is not due to a metabolic auxotrophy. The nsd cells grew more slowly than wild-type cells on this medium, consistent with the result that nsd cells have a slower growth rate when nutrients are not in excess. Notably, however, neither nsd nor wild-type cells formed fruiting bodies on this defined minimal medium.

Cells accumulate the starvation signal ppGpp when grown in nutrients.

We found that nsd causes a reduction in growth rate when cells are grown in 0.5% CTT, and we hypothesize that this occurs because cells detect starvation. We know that the cellular response to starvation is initiated by an increase in ppGpp levels within the cell (36). Therefore, we examined the guanosine nucleotide levels of both wild-type and nsd cells grown in 0.5% CTT medium to determine whether the effect of nutrient concentration on cell growth is accompanied by an increase in ppGpp. We found that, during steady-state growth, nsd cells accumulate about twofold more ppGpp than wild-type cells when grown in 0.5% CTT. The level of ppGpp stays relatively constant throughout the exponential phase and into early stationary phase. These data demonstrate a twofold increase in steady-state ppGpp levels, compared to wild-type levels, in nsd cells grown in 0.5% Casitone, consistent with a twofold reduction in the growth rate.

DISCUSSION

We previously proposed a dual model for initiation of development in M. xanthus cells whereby two signaling pathways interact to couple the processes of starvation sensing and quorum sensing, resulting in progression through the early stages of development (36). This model asserts that an individual cell senses starvation and responds initially by accumulating (p)ppGpp, which in turn changes the cell's gene expression profile and initiates the cellular starvation pathway. Accumulation of (p)ppGpp contributes to cellular release of A-signal, initiating the quorum-sensing system in M. xanthus (15). Release and reception of A signal further alter the gene expression profile, propagating the population signaling pathway and ushering cells through the early stages of development.

We know that elevated levels of (p)ppGpp are necessary and sufficient to initiate development in M. xanthus (15, 30, 36), but we have little direct information on the mechanisms responsible for starvation sensing, (p)ppGpp production, and the effect that (p)ppGpp accumulation has on transcription in M. xanthus. We can make inferences based on what is known about the E. coli stringent response, but, given that M. xanthus couples a stringent response to its developmental program, there are likely to be aspects of its (p)ppGpp pathway that are unique to this organism.

To determine the pathways responsible for (p)ppGpp production, we must learn how M. xanthus cells sense and respond to starvation. Here, we have used nsd as a tool for examining these processes in M. xanthus. In M. xanthus, we know of at least four developmental genes whose regulation requires (p)ppGpp but has no absolute requirements for other known developmental signals (25, 36). Of these, nsd, sdeK, and asgE have been studied in detail. All three are activated in response to starvation, independent of development on a solid surface, and are activated early, within the first few hours of initiation of development (12). When cells are grown in rich medium, nsd is constitutively expressed throughout vegetative growth, while sdeK expression displays growth phase dependence (12). Expression of asgE is more complex, with several promoters driving its expression, but expression from the asgE developmental promoter (P_dev) shows a regulatory pattern similar to expression from the sdeK promoter (10, 11). Developmental transcription of both asgE and sdeK is driven by a putative σ⁵⁴-like promoter (10, 11, 12). Interestingly, in the case of nsd expression, a single 5′ transcript is detected during both vegetative growth and development and transcription appears to be controlled by a single σ⁷⁰-like promoter, not a σ⁵⁴-like promoter, as observed for sdeK and asgE P_dev.

How do the starvation response and the intracellular (p)ppGpp level regulate nsd expression? In E. coli, during a stringent response (p)ppGpp directly interacts with RNA polymerase, affecting its promoter affinity (1, 2, 6, 17, 39). This results in activation of gene expression based on promoter structure, and thus (p)ppGpp appears to stimulate transcription of many σ⁷⁰-like promoter consensus sequences (28, 32). A model for indirect (p)ppGpp gene activation would include activation, either directly or indirectly, of expression of another regulator protein, which would then activate expression of other genes, such as nsd. For example, nsd could be under the control of SigD (41), the putative M. xanthus homologue to the E. coli starvation response sigma factor, σ³⁸ (rpoS) (17, 38).

The nsd insertion mutation affects a cell's ability to detect and/or utilize available nutrients, leading to unwarranted induction of the starvation response. Unfortunately, the predicted Nsd sequence does not provide any clues as to its structure or function, so we turned to genetics and physiology in a further attempt to investigate its potential roles in the starvation sense-and-response pathways in M. xanthus.

We closely examined the effects of the nsd::Tn5lac Ω4469 mutation on a cell's response to nutrient concentrations. Based on our data, nsd cells are unable to perceive when nutrient levels are sufficient to sustain growth, causing an inappropriate induction of the M. xanthus starvation pathway. This conclusion is based on several lines of evidence. First, with a modest reduction in nutrient levels, the growth rate of nsd cells is reduced about twofold compared to the wild-type rate. Second, this aberrant reduction in growth rate by nsd cells coincides with a starvation response: nsd cells produce relatively high levels of (p)ppGpp compared to their wild-type counterparts. As with E. coli (5), we observed an inverse correlation between growth rate and steady-state (p)ppGpp levels. Third, on solid media nsd cells stop growing and initiate development in the presence of nutrients that are sufficient to support continued vegetative growth of wild-type cells.

It is clear from our data that nsd cells do not respond to external nutrient levels in the same way as wild-type cells, and there are several possible explanations for this phenomenon. Our working hypothesis is that cells are unable to accurately detect the concentration of available nutrients, either because they are unable to efficiently take up nutrients or because they are unable to effectively utilize the available nutrients. This effect may be a general defect in development, transport, or metabolism, or it may be specific to the nutrient source used in nutrient-rich media, Casitone. Though we show that nsd cells grow on minimal A-1 media, the growth defect and the subsequent developmental phenotype could be due to a problem in the transport of a peptide specific to Casitone. While the cause of the nutrient sensing defect is unclear, the nsd gene product must be at least partially required, directly or indirectly, for one or more of these functions. In its absence, cells do not receive sufficient nutrients to sustain a wild-type growth rate and initiate a starvation response and initiate the developmental program.

We have concluded, based on the phenotype of nsd::Tn5lac Ω4469 cells, that nsd is involved in the cell's response to nutrient concentration and the concomitant initiation of development. This conclusion is based on the phenotypic effects on cell growth, (p)ppGpp accumulation, and the characteristic colony morphology associated with development. All of these indicators of early development suggest that nsd cells initiate development in the presence of nutrients. However, the fact that sporulation is defective in nsd cells indicates that the later stages of development are impaired. Thus, it appears that the mutation in nsd causes cells to initiate the developmental program in the presence of nutrients but that these cells are unable to complete the process.

Why is fruiting body development defective in nsd cells? There may be other nutrient-sensing pathways that are required to feed into the starvation response. Perhaps these accessory pathways are functional and do not detect starvation when the nsd mutant is grown in the presence of reduced nutrients. In this situation, the initiation event may be defective or the later events may not receive all the proper input to proceed normally to fruiting body maturation and sporulation. In contrast, when these mutant cells undergo a strong starvation on CF or TPM media, all the necessary components from the contributing pathways allow development to proceed normally, yielding functional spores.

We know that many developmental genes receive input from multiple signaling pathways, so perhaps in the nsd mutant some developmental pathways are activated while others are not, and the result is an incomplete and not fully functional developmental program. The nsd mechanism of nutrient sensing activates (p)ppGpp synthesis, but we do not yet know how this activation occurs. This may be part of the M. xanthus stringent response in which (p)ppGpp accumulates in response to a deficiency in charged tRNA, or perhaps it is part of a mechanism for sensing starvation that is unique to M. xanthus.

Because the stringent response in M. xanthus is coupled to initiation of development, it may require unique mechanisms for integrating these processes. For M. xanthus cells, a deficiency of amino acids causes starvation for carbon, nitrogen, and energy, so the stringent response is linked to a general starvation response. The decision to initiate development must be tightly monitored and controlled in order to assure that the cost of fruiting body formation is justified and that the resulting mature fruiting body is functional. For these reasons, it is likely that cells require unique components to modify the M. xanthus stringent response and integrate it into a developmental control point. The SocE/CsgA system for modulating intracellular (p)ppGpp levels has been previously described (7), and now Nsd may be another of these components used by M. xanthus to modify and/or regulate its stringent response.