Myxococcus xanthus fibril appendages are essential for excitation by a phospholipid attractant

Abstract

Isolated (A-motile) Myxococcus xanthus cells glide over solid surfaces and display excitation, a suppression of direction reversals, when presented with phosphatidylethanolamine (PE) purified from its own membranes or synthetic dilauroyl PE and dioleoyl PE. Although the mechanism of PE signal transduction is unknown, we hypothesized that M. xanthus might use surface-associated factors to detect exogenous PE to prevent endogenous lipids from self-stimulating the sensory system. Peritrichous protein and polysaccharide appendages called fibrils were correlated with dilauroyl PE excitation. Wild-type cells treated with Congo red, an inhibitor of fibril assembly, and mutants defective in fibril biosynthesis showed an elevated reversal period, which suggested that fibrils regulate the gliding motor. Furthermore, the loss of fibrils resulted in loss of excitation to dilauroyl PE but not dioleoyl PE. Restoration of fibril production to these mutants restored the dilauroyl PE response. In addition, the dif cytoplasmic signal transduction system and starvation conditions were required for dilauroyl PE excitation. The chemically specific nature of the response and the dependence on the dif system suggests that fibrils define a novel sensory organelle whose evolution may have been necessary to prevent autostimulation by endogenous membrane lipids. Because the hydrophobic nature of dilauroyl PE would be inaccessible to periplasmic chemosensors, we suggest that fibrils act as extracellular signal transducers to probe surfaces for insoluble chemical signals.

Myxococcus xanthus uses two distinct motility systems to glide over solid surfaces (1). The adventurous or A-motility system enables cells to move as individuals. The social or S-motility system requires cell-cell contact because S motile cells (A⁻S⁺) cannot move as single cells (2). At high cell densities both motility systems operate, and mutation of both systems is required to render the cells nonmotile. The mechanism of A motility is not understood. S motility bears extensive similarity to twitching motility in its dependence on type IV pili (3–5). M. xanthus displays chemotaxis toward phosphatidylethanolamine (PE) purified from its own cell membrane and chemically synthesized dilauroyl (di C12:0) and dioleoyl (di C18:1 ω9c) PE (6). PE is the first discrete chemoattractant for the gliding bacteria and is the first lipid attractant for any bacterium. How the PE chemotaxis sensory systems are coupled to the operation of the A and S motors is a new frontier.

Escherichia coli cells swim through liquid media by using flagella and are chemotactic toward several amino acids and sugars. These attractants are detected by integral membrane methyl-accepting chemotaxis proteins (MCPs) to regulate flagellar rotation (7). Chemotaxis is divided into two processes, excitation and adaptation. E. coli regularly punctuates smooth swimming with periods of erratic tumbling but during excitation by an attractant the tumbling behavior is suppressed. Excitation results in directed movement because cells run smoothly for a longer period while traveling up an attractant gradient (8). Chemotaxis ceases when a uniform concentration of attractant is attained because an adaptive response restores the basal frequency of tumbles (9).

Despite its motility differences with E. coli, the M. xanthus response toward PE demonstrates similar excitation and adaptation behaviors. Isolated A-motile M. xanthus cells move along their long axis (equivalent to a run) and reverse their direction of movement (equivalent to a tumble) once every 6.8 min (10). In the excitation assay the behavior of isolated A-motile cells is examined in response to a uniform concentration of an attractant. M. xanthus cells increase their reversal period, the time between reversal events, when presented with an attractant, and the basal reversal frequency is restored by adaptation after 45 min of PE exposure (6). This finding suggests that although E. coli and M. xanthus move differently, they use similar strategies to mediate chemotaxis.

In this work we examine the cellular components required for the response of A-motile cells to PE by using the excitation assay. We were puzzled by the fact that lipids extracted from the membranes of cells repress reversals and direct cell movements when presented exogenously, but do not autostimulate the producer cell. The use of an integral membrane MCP as the sole chemoreceptor does not seem a likely solution to the problem of autostimulation because it is surrounded by membrane lipid. We searched for cell surface components that might be required for PE chemotaxis and discovered that fibrils, flexible appendages that extend away from the cell surface, are required for excitation by dilauroyl PE. We suggest that the fibrils are extracellular signal receptors/transducers for detecting insoluble chemical signals.

Methods

Strains and Conditions.

M. xanthus DK1622 (wild type), DK3482 (tglA), LS1119 (tglA/stk), DK3470 (dsp), SW501 (difE), and SW504 (difA) were grown in CYE broth [10 g/liter Difco casitone, 5 g/liter yeast extract, 10 mM 3-(N-morpholino)propanesulfonic acid (Mops, pH 7.6), and 4 mM MgSO₄] with vigorous shaking at 32°C.

Excitation Assay.

The excitation assay measures changes in the reversal period in response to exposure to a uniform concentration of attractant (6). A TPM agar plate [10 mM Tris (hydroxymethyl) aminomethane HCl (pH 7.6), 8 mM MgSO₄, 1 mM KHPO₄-KH₂PO₄, and 1.5% Difco agar] was dried for 10 min at 37°C. Dioleoyl PE and dilauroyl PE were dissolved in chloroform and chloroform/methanol (1:1), respectively, to 0.005 mg/ml, 0.05 mg/ml, and 0.5 mg/ml concentrations. Four microliters of PE solution was applied to TPM agar in an area of about 0.4 cm² and dried for 20 min at 37°C. Five microliters of M. xanthus cells diluted to 5 × 10⁶ cells/ml in Mops buffer (10 mM Mops, pH 7.6/8 mM MgSO₄), was dried on top of the test compound for 15 min at 32°C to allow the liquid grown cells to adjust to the surface. Plate cultures were observed at room temperature with a Leitz Laborlux D microscope for 45 min at ×640 magnification. Stop-motion digital movies were produced by using a microscope-mounted Sony Power HAD 3CCD color video camera and a Macintosh 9500 with Adobe premier software (frame capture rate: 12 frames/min). To determine cellular reversal period, the paths of 20 isolated cells were followed, and reversals were manually enumerated.

In some experiments, purified fibrils were added to 5 × 10⁷ cells in 1 ml of cohesion buffer (10 mM Mops, pH 6.8/1 mM CaCl₂/1 mM MgCl₂) at a final concentration of 3.2 × 10⁻⁷ μg carbohydrate/cell and rotated at 32°C for 20 min before spotting cells on top of test compound.

Fibril Isolation and Quantitation.

Fibrils were purified from developmental M. xanthus cells by using a protocol modified from Clemans, Chance, and Dworkin (11). M. xanthus DK1622 was grown to mid-log phase in 200 ml of CYE broth, harvested, and resuspended in Mops buffer to 5 × 10⁹ cells/ml. A total of 4.2 ml of cell suspension was spread on each of two 500-ml TPM agar trays (17′′ × 11′′), covered, and incubated for 24 h at 32°C. The cells then were scraped in a beaker and stirred in 5 ml of ice-cold TNE buffer [10 mM Tris (hydroxymethyl) aminomethane HCl, pH 7.5/100 mM NaCl/5 mM EDTA] for 10 min. Five milliliters 1% SDS in TNE buffer was added and stirred for 30 min at room temperature. The solution was centrifuged at 12,000 × g for 10 min at 4°C, and the pellet was resuspended in 2.5 ml of ice-cold TNE and 2.5 ml of 1% SDS TNE and stirred for 1 h at room temperature. The pellet was recovered and then washed once in 5 ml of TNE buffer, twice in 5 ml of 10 mM Mops buffer (pH 6.8), and twice in 5 ml of cohesion buffer. Finally, the pellet was resuspended in 1 ml of cohesion buffer and stored at −20°C. Fibrils were quantified as a function of carbohydrate content (12).

Results

Lipids extracted from the membranes of M. xanthus cells repress direction reversals when presented exogenously, but do not autostimulate the producer cell (6). We reasoned that to prevent autostimulation the detection system must not be in direct contact with the membrane lipids. M. xanthus has two types of appendages that extend away from the cell surface: type IV pili and fibrils. Pili are rigid filaments of repeating pilin monomers that extend from one cell pole and are thought to provide the power for S motility (3, 4). Fibrils are long, flexible appendages composed of a polysaccharide matrix decorated with proteins (13–15). Although the genes involved in fibril biogenesis have not been identified and sequenced, there are several ways to remove fibrils from the cell surface. One involves treating cells with the dye Congo red, which appears to prevent crystallization of the fibril polysaccharide matrix. Congo red-treated cells lack fibrils when examined by electron microscopy (16). Treatment of wild-type cells with 57 μM Congo red elevated the M. xanthus basal reversal period from 6 to 15 min, suggesting that fibrils influence the operation of the A-motility motor. More important, Congo red inhibited excitation to dilauroyl PE over a wide concentration range (Fig. 1a) but had no effect on the response to dioleoyl PE (Fig. 1b). It appeared that fibrils regulate the excitation of A motility in a chemically specific manner.

Figure 1.

Elimination of fibrils inhibits excitation by dilauroyl PE. Wild-type DK1622 M. xanthus reversal periods with increasing amounts of dilauroyl PE (a) and dioleoyl PE (b). Reversal periods were determined with (●) and without (○) 57 μM Congo red dye added to the agar. Error bars are the standard deviation of three replicates.

tglA cells are defective in the assembly of pili and consequently the ability to move by S motility (17, 18). The tglA mutation is pleiotropic, as tglA mutants are also defective in fibril production (19). tglA cells retain A motility but lack excitation to both dilauroyl and dioleoyl PE (Fig. 2). tglA can be partially complemented by a second mutation in the stk locus that restores fibril production, but not piliation or S motility (19). Suppression of the fibril defect restored excitation to dilauroyl PE (Fig. 2a), which argues that fibrils, and not pili, are necessary for dilauroyl PE excitation. The stk mutation did not restore excitation to dioleoyl PE (Fig. 2b), suggesting that tglA must lack additional elements necessary for dioleoyl PE excitation. The loss of fibrils by Congo red treatment increased the basal reversal period 2- to 3-fold (Fig. 1a) but tglA did not exhibit a similar behavior. We suspect this loss is caused by the pleiotropic nature of tglA; the loss of pili, components involved in dioleoyl PE perception, or other unknown defects may suppress an elevated reversal period caused by the lack of fibrils. Interestingly, complementation of fibrils by the stk second site suppressor decreased the tglA reversal period by 30%, suggesting that fibrils are involved in maintaining a lower reversal period (Fig. 2).

Figure 2.

Genetic complementation of the fibril defect restores excitation to dilauroyl but not dioleoyl PE in a tglA mutant. DK3482 tglA (▵) and LS1119 tglA/stk (▴) M. xanthus reversal periods with increasing amounts of dilauroyl PE (a) and dioleoyl PE (b). Error bars are the standard deviation of three replicates.

Another S-motility mutant, dsp, lacks fibrils but retains type IV pili (16). As with the Congo red treatment, the lack of fibrils in the dsp mutant resulted in an increase in the basal reversal period from 7 min to at least 16 min (Fig. 3). More significantly, dsp cells did not suppress reversals in the presence of dilauroyl PE but displayed normal excitation to dioleoyl PE (Fig. 3a). Unlike tglA cells, dsp cells are unable to form fibrils with the stk mutation (19) and lack the fibril proteins (20). However, dsp can be complemented by the addition of purified wild-type fibrils, which presumably self-assemble on the dsp cell surface (21). Using a similar approach, we incubated dsp cells with purified wild-type fibrils for 20 min before initiating the excitation assay. The addition of purified fibrils lowered the unstimulated reversal period of dsp closer to that of the wild type and restored excitation to dilauroyl PE (Fig. 3b).

Figure 3.

Purified fibrils restore dilauroyl PE excitation in an M. xanthus dsp mutant. (a) DK3470 dsp M. xanthus reversal periods with increasing amounts of dilauroyl (□) and dioleoyl PE (■). (b) Reversal periods of dsp cells with increasing amounts of dilauroyl PE (◊). Purified fibrils were added exogenously to dsp before observation (⧫). Error bars are the standard deviation of three replicates.

The dif locus maps close to dsp on the M. xanthus chromosome, and dif mutants have similar defects in S motility (22). The dif locus consists of at least five genes homologous to the enteric chemotaxis proteins (23). Recent results suggest that dif mutants also lack fibrils, suggesting a cognate relationship between fibrils and a cytoplasmic signal transduction system (22). We tested mutants in difA, an MCP homolog, and difE, a cheA homolog, for their excitation response to the two attractants. Both mutants displayed elevated basal reversal periods, which is consistent with the observation that they lack fibrils (Fig. 4). The difA mutant lacked excitation to dilauroyl PE but the difE mutant failed to increase its reversal period in the presence of either attractant. The addition of purified wild-type fibrils to difA cells resulted in a decrease in the basal reversal period as with dsp. Unlike dsp, exogenous fibrils did not restore excitation to dilauroyl PE (Fig. 4a). These results suggest that both DifA and fibrils are essential for the response.

Figure 4.

The Dif cytoplasmic sensory transducers are essential for dilauroyl PE excitation. (a) The reversal periods of SW504 difA in response to increasing amounts of dioleoyl PE (○) and dilauroyl PE with (■) and without (●) the addition of purified wild-type fibrils. (b) The reversal periods of SW501 difE in the presence of increasing amounts dioleoyl PE (○) and dilauroyl PE (●). Error bars are the standard deviation of three replicates.

Environmental conditions are also known to regulate fibril biogenesis. At low cell densities, starvation in the presence of divalent cations is necessary for fibril formation (24). Under similar conditions, nutrient-rich media inhibits fibril production (20). The reversal period assay was performed on CTT agar, a nutrient-rich growth medium (1), to suppress fibril biogenesis, and under these conditions the cells did not demonstrate excitation to dilauroyl PE but retained excitation to dioleoyl PE (Fig. 5). Therefore, it appears that although the dilauroyl PE response is starvation-dependent, the dioleoyl PE response is starvation-independent. CTT-grown cells are another case in which fibrils are absent but the basal reversal period is similar to wild type (Fig. 5). The physiology of M. xanthus differs greatly between growing and starving cells, and perhaps growing M. xanthus has an alternate means of suppressing the fibril-deficient reversal defect. dif mutants have a wild-type reversal period of 6–8 min when observed on CTT media, and the dif basal reversal defect only becomes apparent under starvation conditions (23) (Fig. 4). Thus, nutrient-rich conditions can suppress the elevated reversal periods we observe.

Figure 5.

Starvation is required for excitation to dilauroyl PE. Wild-type DK1622 M. xanthus reversal periods on nutrient-rich CTT agar in response to increasing amounts of dilauroyl (●) and dioleoyl PE (○). Error bars are the standard deviation of three replicates.

In summary, Congo red- and CTT-treated wild-type cells and the tgl, dsp, and dif mutants all lack excitation to dilauroyl PE and all lack fibrils. Furthermore, restoration of fibril production to tglA with the stk mutation restored excitation to dilauroyl PE as did addition of fibrils to the dsp mutant. Except for tglA, all display a normal response to dioleoyl PE. These results suggest that fibrils are essential for excitation by dilauroyl PE but not dioleoyl PE. To determine whether type IV pili are necessary for the dioleoyl PE response, we examined the excitation of DK10410, which lacks the gene encoding the principle pilin subunit PilA (4). The pilA mutant displayed excitation to dioleoyl PE (data not shown), indicating that pili are not required for this response.

Discussion

In E. coli, chemotaxis sensory transduction begins when a hydrophilic chemoattractant or an attractant/protein-complex binds to the periplasmic domain of a transmembrane MCP (7). MCP binding then inhibits a cytoplasmic phosphotransfer relay to control the direction of flagellar rotation (25). However, recognition of a hydrophobic ligand such as PE by the hydrophilic periplasmic domain of an MCP is unlikely and suggests that the M. xanthus PE sensory transduction pathway involves novel elements. Here we suggest that appendages called fibrils may be instrumental in carrying information about insoluble chemicals to the A-motility motor.

M. xanthus chemotaxis is the first bacterial system known to use a lipid chemoattractant. Both dilauroyl and dioleoyl PE induce similar responses, but until now it was unclear whether these species were detected separately or whether both attractants stimulated one system with loose specificity. It appears that each species has a unique detection method because fibril appendages are required for excitation by one species but not the other.

The loss of fibrils, either genetically or biochemically, alters the behavior of A-motile M. xanthus and inhibits excitation to dilauroyl PE. We propose that fibrils initiate a sensory transduction pathway that communicates information about insoluble materials across the cell membrane. In support of this hypothesis, Congo red-treated wild-type cells, the dsp mutant and dif mutants demonstrated an elevated reversal period in the absence of PE, which indicates a role for fibrils in motility control. Furthermore, the loss of fibrils resulted in the loss of excitation to dilauroyl PE but not dioleoyl PE as one might expect for a chemoreceptor. When considered together, these data suggest that M. xanthus fibrils serve as chemosensors for lipid recognition. Fibrils also mediate intercellular cohesion (16) although cohesion and excitation are not obligately interdependent as fibril-mediated excitation is observed in isolated cells. Isolated M. xanthus cells are not stimulated by the endogenous attractant in their own membrane, and thus localization of receptors away from the cell surface may have evolved to prevent autoexcitation.

It is not known how fibrils transmit information to the inside of the cell. It is unlikely that fibrils simply degrade PE into soluble components that serve as chemoattractants as the minimum structure with activity is diacylglycerol (6). Fibrils could be physical channels that siphon signal lipids to cytoplasmic membrane receptors. They might be mechanosensors whose freedom of motion is restricted when binding PE. Although these are possible mechanisms, the overriding theme of bacterial chemoreceptors is one of conformational change. In E. coli, ligand recognition alters the conformation of MCPs, and this physical change inhibits a cytoplasmic chemical modification relay (25, 26). We therefore favor the notion that fibrils bind PE and recognition is propagated along the length of the fibril by recursive conformational change via either steric or covalent modification. Covalent modification is particularly intriguing as a 29-kDa fibril protein becomes ADP-ribosylated and could play a role in signal transduction (27).

Fibrils may interact with cytoplasmic components to control the A-motility motor. M. xanthus motility is regulated by at least two distinct cytoplasmic signal transduction systems, Frz and Dif, homologous to the enteric Che proteins (23, 28). Mutants in the Frz system demonstrate excitation to dilauroyl PE but are defective in chemotaxis because they lack the adaptive response (6). Fibrils therefore are unlikely to interact directly with the Frz system because fibril mutants are defective in excitation and frz mutants are not. The Dif system is essential for social motility and dif mutants lack fibrils (22). It is not known why signal transduction mutants produce cell surface defects but the phenotype might indicate that the Dif proteins form the cognate cytoplasmic transduction system for fibril-mediated signals. This idea is supported by the fact that the difA mutant specifically lost excitation to dilauroyl PE and could not be rescued by exogenous fibrils, thus difA is likely the MCP that receives the dilauroyl PE stimulus. The difE mutant was defective in excitation to both dilauroyl and dioleoyl PE and suggests that both attractants feed into a common cytoplasmic signal transduction pathway. The MCP for dioleoyl PE is not known. That fibril assembly is coupled to the Dif cytoplasmic signal transduction system further supports the idea that the fibrils may transduce chemical signals.

M. xanthus fibrils are not only responsible for cell cohesion but also, like antennae, appear to probe surfaces for insoluble chemical signals. But how widespread is this type of appendage-mediated perception system? Cell-contact dependent alteration of cell physiology is being observed with increasing frequency in pathogenic bacteria (29). In Neisseria meningitidis, adhesion to host cells by type IV pili is required for enhanced gene expression, colony dispersal, effacing lesions and ultimately, virulence (30, 31). Fibril-like appendages are induced after contact of Salmonella typhimurium to host cells and are required for subsequent cell invasion (32). P pilus-mediated adherence of E. coli to host globoside receptors directly results in induction of virulence-essential iron acquisition genes (33), and E. coli induces filamentous appendages before type III-mediated secretion of virulence factors (34). In each case an extracellular appendage, albeit of different types, is required for attachment and, like the fibrils of M. xanthus, may transduce a ligand-specific signal to change the physiology of the pathogen. The unification of adhesion and signaling within an intercellular appendage may be the hallmark of a new and diverse class of bacterial extracellular signal transducers.

Abbreviations

PE

phosphatidylethanolamine

MCP

methyl-accepting chemotaxis protein