Genetic and Molecular Analysis of cglB, a Gene Essential for Single-Cell Gliding in Myxococcus xanthus

Abstract

Gliding movements of individual isolated Myxococcus xanthus cells depend on the genes of the A-motility system (agl and cgl genes). Mutants carrying defects in those genes are unable to translocate as isolated cells on solid surfaces. The motility defect of cgl mutants can be transiently restored to wild type by extracellular complementation upon mixing mutant cells with wild-type or other motility mutant cells. To develop a molecular understanding of the function of a Cgl protein in gliding motility, we cloned the cglB wild-type allele by genetic complementation of the mutant phenotype. The nucleotide sequence of a 2.85-kb fragment was determined and shown to encode two complete open reading frames. The CglB protein was determined to be a 416-amino-acid putative lipoprotein with an unusually high cysteine content. The CglB antigen localized to the membrane fraction. The swarming and gliding defects of a constructed ΔcglB mutant were fully restored upon complementation with the cglB wild-type allele. Experiments with a cglB allele encoding a CglB protein with a polyhistidine tag at the C terminus showed that this allele also promoted wild-type levels of swarming and single-cell gliding, but was unable to stimulate ΔcglB cells to move. Possible functions of CglB as a mechanical component or as a signal protein in single cell gliding are discussed.

Myxococcus xanthus is a rod-shaped, gram-negative soil bacterium that translocates on solid surfaces in the direction of the cell’s long axis, in a mode of movement called gliding motility (8, 30). Translocation by gliding motility is found in many phylogenetically unrelated prokaryotes. In M. xanthus, the microscopic gliding movements of individual cells are coordinated, leading to macroscopically visible expansion of swarms on an agar plate (14). These swarms originate from the edge of a colony and are observed as flare-like projections, with single cells moving ahead of groups. Thus, a wild-type colony exhibits an undefined edge. Genetic analyses of gliding motility, studies on swarm expansion, and single-cell tracking experiments have revealed that cellular movements of M. xanthus are controlled by two extensive gene systems, the A (adventurous)- and S (social)-motility systems as well as mgl and frz genes (3, 6, 11, 12, 14, 30, 32). The A-motility system controls gliding motility of individual cells (11), while the S-motility system is essential for movement of cells in groups (12). The A- and S-motility systems are complementary in the sense that colonies of cells deficient in both systems (A⁻ S⁻ double mutants) do not swarm (11). The systems differ in that close cell-to-cell proximity is required only for S-motility to operate. It has been suggested that the A- and S-motility systems encode for two different motors promoting surface translocation (11, 12, 31, 36, 38–40). Recent genetic and molecular studies on the S-motility system have shown the sglI region of M. xanthus to encode the genes required for structure, export, and function of type IV pili (13, 36–39). In a variety of diverse bacteria, type IV pili have been postulated to be involved in a mode of surface translocation known as twitching motility (5, 9, 27). Thus, S-motility in M. xanthus may be related to type IV pilus-dependent twitching movements.

Gliding movements of individual M. xanthus cells occur in the absence of any visible cell organelle. The molecular structure of the gliding motor and the physics of force generation remain unknown. Since the A-motility system controls gliding of individual, isolated cells (11) (see below), it is likely that the genes of the A-motility system include those genes that encode components of the gliding motor. More than 37 genes are known to affect A-motility, and mutations in these genes, in contrast to those in S-motility genes, cause defects primarily in gliding motility rather than in gliding as well as in fruiting body formation (11, 20, 21). Colonies of such A⁻ S⁺ cells exhibit a sharp, highly delineated edge, with no single cells present at the colony’s perimeter. This characteristic colony morphology is due to the inability of single cells to glide when separated from other cells by more than 2 μm (11) (see below). A-motility genes can be further divided into two subclasses, the agl genes and the cgl genes, based on an additional mutant phenotype (10, 11). If the defect in single-cell gliding of an A-motility mutant can be complemented by extracellular rescue, i.e., by mixing with wild-type cells or mutant cells of another motility class, then the gene is designated cgl (for contact or conditional gliding). Typically, this extracellular rescue is only transient though sufficient to allow microscopic observation of flare formation at swarm edges. A-motility mutants that cannot be rescued are called agl (adventurous gliding) mutants.

The peculiar phenotype of cgl mutants suggests that the cgl gene products may function either as mechanical elements of the gliding motor that are associated with the outer membrane, as regulators of its activity, or as both. Five cgl loci, cglB, cglC, cglD, cglE, and cglF, have been identified, and these loci have been mapped by using transposon mutagenesis in conjunction with Mx8 cotransduction experiments (11, 29). During preliminary experiments, we identified cglB mutants as having a strong motility defect and a clear stimulation response. We have exploited these properties to conduct a molecular analysis of cglB in M. xanthus.

MATERIALS AND METHODS

Bacterial strains, plasmids, phages, and growth media.

The bacterial strains used in this study are listed in Table 1. Escherichia coli TG1 recO 1504::Tn5 and E. coli DH10B were used as hosts for subcloning, E. coli BL21(DE3) was used to express the CglB-His protein.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant genotype or characteristics	Motility phenotype	Reference or source
M. xanthus
ASX1	ΔcglB	A− S+	This study
ASX2	ΔcglB ΔpilA	A− S−	This study
ASX21	ΔcglB, complemented with cglB+	A+ S+	This study
ASX31	ΔcglB, complemented with cglB-His	A+ S+	This study
ASX34	ΔcglB, complemented with cglB-His, mgl-9 linked to Tn5-132	mgl	This study
ASX36	ΔcglB, complemented with cglB+, mgl-9 linked to Tn5-132	mgl	This study
DK307	cglB1 sglA1	A− S−	11
DK321	cglB2 sglA1	A− S−	11
DK331	cglB3 sglA1	A− S−	11
DK335	cglB5 sglA1	A− S−	11
DK344	cglB6 sglA1	A− S−	11
DK348	cglB8 sglA1	A− S−	11
DK352	cglB10 sglA1	A− S−	11
DK353	cglB11 sglA1	A− S−	11
DK355	cglB12 sglA1	A− S−	11
DK357	cglB13 sglA1	A− S−	11
DK377	cglB14 sglA1	A− S−	11
DK379	cglB15 sglA1	A− S−	11
DK382	cglB16 sglA1	A− S−	11
DK388	cglB17 sglA1	A− S−	11
DK1218	cglB2	A− S+	12
DK1622	Wild type	A+ S+	13
DK1932	sglA1 tgl-4, Tn5 Ω1932 linked to cglB+	A+ S−	29
DK3685	mgl-9, Tn5-132 linked to mgl	mgl	31a
DK10410	ΔpilA	A+ S−	38
JZ315	cglB::Tn5phoA Ω315	A− S+	16
E. coli
DH10B	F−mcrAΔ(mrr-hsdRMS-mcrBC) φ80dlacZΔM15 ΔlacX74 endA1 recA1 deoR Δ(ara leu)7697 araD139 galU galK nupG rpsL λ−	NAa	Bethesda Research Laboratories
TG1 recO 1504::Tn5	K12 supE thi Δ(lac-proAB) hsdΔ5 F′[traD36 proA+B+ lacIqlacZΔM15]	NA	Novagen
BL21 (DE3)	E. coli B F−ompT rB− mB− (λDE3 lysogen) (λDE3 = imm λ21; in the int gene of λ is cloned a fragment bearing lacI and the T7 RNA polymerase gene regulated by the lacUV5 promoter)	NA	Novagen
Plasmid
pBluescript II SK	ColE1 ori, lacZ, Apr		Stratagene
pPLH343	Mx8 attP site, Kmr		P. Hartzell (15)
pBJ113	KG cassette		B. Julien
pET-21b	T7 promoter, His tag		Novagen
pBSKS-1	13-kb SacI fragment of the M. xanthus chromosome in pBluescript II SK; orfA, cglB+, orfC::Tn5 Ω1932		This study
p343H7	2-kb SacI/HindIII fragment of 13-kb insert of pBSKS-1 (right side of Tn5 Ω1932) in pPLH343		This study
p343H3	5.1-kb SacI/HindIII fragment of 13-kb insert of pBSKS-1 in pPLH343; orfA, cglB+, orfC::Tn5 (partial)		This study
p343B4	4.1-kb BamHI/HindIII fragment of 13-kb insert of pBSKS-1 in pPLH343; orfA, cglB+, orfC::Tn5 (partial)		This study
p343B4R	As p343B4 but reverse orientation of 4.1-kb insert		This study
p343E3	3-kb EcoRI/HindIII fragment of 13-kb insert of pBSKS-1 in pPLH343; promoterless cglB+, orfC::Tn5 (partial)		This study
p343E3R	As p343E3 but reverse orientation of 3-kb insert		This study
p343BBG2	1.8-kb BglII/HindIII fragment of 13-kb insert of pBSKS-1 in pPLH343; 3′ end of cglB+ and orfC::Tn5 (partial)		This study
p343BBG3	2.2-kb BamHI/BglII fragment of 13-kb insert of pBSKS-1 in pPLH343; orfA and 3′ truncated cglB+		This study
p343B4.His	cglB allele encoding wild-type CglB protein with C-terminal 6-histidine tag in pPLH343		This study
pCB25	1.5-kb SalI/EcoRI fragment of 13-kb insert of pBSKS-1 in M13mp18; cglB+		This study
pCB26	0.6-kb ApaLI internal deletion of pCB25; ΔcglB+		This study
pBSK-BE1	1-kb EcoRI/BamHI fragment of 13-kb insert of pBSKS-1 in pBluescript-II SK; upstream region of cglB		This study
pBSK-BSΔcglB	2-kb BamHI/SalI fragment of 13-kb insert of pBSKS-1 lacking internal ApaLI fragment in pBluescript-II SK; ΔcglB+ plus upstream region		This study
p113ΔcglB	2-kb BamHI/SalI fragment of pBSK-BSΔcglB in pBJ113; ΔcglB+ plus upstream region		This study
pET21cglBΔ21	cglB allele encoding wild-type cglB with N-terminal 21-amino-acid deletion and C-terminal 6-histidine tag		This study

^aNA, not applicable. 

pBluescript II SK (Stratagene) was used for subcloning, and pPLH343 (constructed by P. Hartzell; described by Kalman et al. [15]) was used for ectopic expression of DNA in M. xanthus. This vector contains the Mx8 attP site, which directs site-specific recombination and insertion of the plasmid into the chromosomal Mx8 attachment site when electroporated into M. xanthus. pBJ113 (constructed by B. Julien) was used to replace the wild-type copy of cglB by an in-frame deleted copy in M. xanthus. It has a positive-negative KG cassette with a kanamycin resistance (Km^r) gene for positive selection and a galactokinase gene (galK) for negative selection (34). pET-21b (Novagen) was used to create a C-terminal CglB-polyhistidine fusion (CglB-His) and to express a recombinant CglB protein in E. coli. The M13 derivative phages M13mp18 and -19 (41) were used as vectors for DNA sequencing or for subcloning.

M. xanthus strains were grown in CTT medium (1% Casitone, 8 mM MgSO₄, 10 mM Tris-HCl, 1 mM KH₂PO₄ [pH 7.6]) at 32°C or on 1.5% agar CTT plates containing appropriate antibiotics (kanamycin at 40 μg/ml or oxytetracycline at 12.5 μg/ml) when required. E. coli was grown in Luria-Bertani broth (LB) or on 2% agar LB plates supplemented with ampicillin (100 μg/ml) or kanamycin (50 μg/ml) as needed.

DNA manipulation.

Myxococcal chromosomal DNA was prepared as described previously (2). Plasmid or single-stranded DNA preparations, alkaline phosphatase treatments, ligations, and other DNA manipulations were performed according to standard procedures for E. coli (25). Plasmids were introduced into E. coli by transformation (25) or electroporation. Plasmids were introduced into M. xanthus by electroporation according to the method of Kashefi and Hartzell (18).

DNA sequencing and sequence analysis.

Fragments for DNA sequencing were cloned into M13mp18 or M13mp19 and sequenced by the dideoxynucleotide chain termination method (26) with [α-³⁵S]dATP (10 mCi ml⁻¹; Amersham) and modified T7 DNA polymerase (Sequenase version 2.0; U.S. Biochemical). To resolve secondary structure, sequencing reactions were carried out with 7-deaza-dGTP (22) or dITP substituting for dGTP. Both strands were sequenced with primers supplied in the Sequenase kit or with internal oligonucleotide primers (17-mer). To verify genetic constructs (gene fusions and in-frame deletions) or to confirm the introduction of a mutation, sequencing was performed. Sequence analysis was performed with programs from the Wisconsin Package (version 9.1-Unix; Genetics Computer Group, University of Wisconsin, Madison, Wis.).

Plasmid constructions.

pBSKS-1 is a pBluescript II SK plasmid containing a 13-kb SacI chromosomal fragment from M. xanthus DK1932. Tn5 Ω1932 is inserted centrally in this fragment, and the cglB locus mapped to the left end of Ω1932 as indicated in Fig. 1. Subclones of this fragment were constructed in plasmid pPLH343, which contains the Mx8 attP site and integration genes. Plasmids p343H7 and p343H3 contain a SacI/HindIII fragment of the right and left sides, respectively, of Ω1932 (Fig. 1; Table 1). p343B4, p343B4R (same insert as in p343B4 but in the other orientation), p343E3, p343E3R (same insert as in p343E3 but in the other orientation), p343BBG2, and p343BBG3 are deletion subclones of p343H3 (Fig. 1). pCB25 is a M13mp18 derivative that contains the SalI/EcoRI fragment to the left of Ω1932 (Fig. 1). pBSK-BE1 is a pBluescript II SK derivative that contains the 1 kb EcoRI/BamHI fragment to the left of Ω1932 (Fig. 1).

FIG. 1.

Characterization of the cglB locus. The open box represents Tn5 Ω1932. (A) Restriction map of the 13-kb SacI fragment cloned into pBSKS-1. Only relevant sites are shown, and not all SalI sites were mapped. DNA fragments adjacent to Ω1932 were subcloned into plasmid pPLH343, which integrates at the chromosomal Mx8 attachment site attB. All of these fragments were tested for complementation of the swarming defect of cglB mutant colonies, and results are indicated as + (complementation), − (no complementation) or +/− (partial complementation). The arrow represents the Tn5 Km^r gene. (B) Predicted ORFs in the 2.85-kb sequence to the left of Tn5 Ω1932. The white line within the arrow indicates the fragment removed in the in-frame deletion to construct the cglB null mutant. Only relevant endonuclease restriction sites are shown.

Construction of an ORFB null mutant.

To construct the open reading frame B (ORFB) deletion, pCB25 was digested with ApaLI to release a 663-bp internal fragment of the insert. The digest was treated with Klenow fragment DNA polymerase in the presence of deoxynucleoside triphosphates to produce blunt DNA ends and religated with T4 DNA ligase. The plasmid with the deleted copy of ORFB is called pCB26. The insert of pCB26 was released as an EcoRI/SalI fragment and ligated into pBSK-BE1 (see above) digested with EcoRI/SalI to create plasmid pBSK-BSΔcglB. ΔcglB lacks DNA sequence encoding for amino acids from positions 11 to 231, which represent 53% of the amino acids of the protein including half of the signal peptide sequence. The ΔORFB mutation was sequenced across the deletion site in plasmid pCB26 to confirm that no frameshift was introduced. To construct a gene replacement of cglB with ΔcglB (ΔORFB) in the M. xanthus chromosome, the 2-kb EcoRI/SalI insert of pBSK-BSΔcglB was cloned into pBJ113 digested with EcoRI/SalI to create p113ΔcglB. Plasmid pBJ113 contains a positive/negative KG cassette for screening the two-step integration/excision events during the gene replacement (34). p113ΔcglB was introduced into M. xanthus by electroporation. This plasmid does not replicate in M. xanthus, and therefore in the first step, Km^r electroporants must contain a copy of the plasmid integrated into the chromosome by homologous recombination. In the second step, several Km^r colonies were plated directly on 1% galactose–CTT agar. The galK gene confers sensitivity of M. xanthus to galactose. Surviving galactose-resistant (Gal^r) cells must lack a functional galK gene, either by excision of the integrated plasmid by a second homologous recombination event or by mutation of galK. If an excision occurred, either the wild-type or the deletion allele remains in the M. xanthus chromosome. Southern blot analysis was used to distinguish between these two possibilities.

Construction of CglBΔ21-His expression vector for overexpression of cglB in E. coli.

Since the predicted CglB protein contains a signal peptide sequence that is typical of lipoproteins, it is most likely that the protein localizes to the membrane. We expressed CglB without the N-terminal 21 amino acids (CglBΔ21) to ensure that it localizes in the cytoplasm during expression in E. coli. The region of the cglB gene encoding CglBΔ21 was amplified by PCR using the primers TGGCGGATCCGACGTACGACTTC (N terminus) and TGGACTCGAGCTGACGGATGGCCC (C terminus), which contain BamHI and XhoI restriction sites (underlined), respectively. The resulting 1.2-kb product was digested with BamHI and XhoI and then ligated into similarly digested pET-21b (Novagen) to form plasmid pET21cglBΔ21, which was transformed into E. coli BL21(DE3). Transformed cells were selected by using ampicillin. The insert of this plasmid was sequenced to ensure error-proof amplification. The predicted amino acid sequence of the CglBΔ21 protein, expressed from pET21cglBΔ21, was MASMTGGQQMGRDPTYD-CglB-IRQLEHHHHHH. Residues 2 to 12 are a T7 tag comprising the N-terminal 11 amino acids of the T7 gene 10 protein that enhances overexpression of the protein, while the terminal 6 amino acids were histidine residues. Residues in bold indicate the first three and the last three amino acids of the CglB amino acid sequence included in CglBΔ21-His.

Construction of cglB-His allele for expression in M. xanthus.

To construct plasmid p343B4.His, containing the gene encoding the wild-type CglB with a C-terminal tag of six histidine residues, plasmid pET21cglBΔ21 was digested with StyI and treated with Klenow fragment to produce blunt ends. Subsequently, the linear plasmid was further digested with BglII to release a 0.2-kb fragment encoding the C-terminal part of CglB with the polyhistidine tag. From plasmid p343B4, a 12-kb fragment, containing plasmid pPLH343 plus a portion of cglB that encodes the N-terminal part CglB, was released after linearization with NdeI, blunt-end formation after treatment with Klenow fragment, and final digestion with BglII. The 0.2- and 12-kb fragments were ligated. Plasmid p343B4.His contained a gene encoding the wild-type CglB protein with a polyhistidine tag (IRQLEHHHHHH) at the C terminus including about 1 kb DNA upstream of cglB-His. This plasmid was introduced into ΔcglB strain ASX1 by electroporation to create strain ASX31.

Construction of CglB donor strains.

A stock of phage Mx8 was prepared from strain DK3685 (mgl-9 linked by about 80% to Tn5-132 (tetracycline resistant [Tet^r]). The phage stock was used to infect strain ASX21 (ΔcglB, p343B4), and transductants were selected on CTT agar plates containing 12.5 μg of oxytetracycline per ml. A nonswarming, Tet^r colony, named ASX36 (ΔcglB, mgl-9 linked to Tn5-132, p343B4) was selected for further studies. Phage stock of strain DK3685 was also used to infect strain ASX31. A Tet^r, nonswarming colony was selected for further studies and named ASX34 (ΔcglB, mgl-9 linked to Tn5-132, p343B4.His).

Stimulation of ΔcglB mutants.

An assay for extracellular stimulation of movement of cglB cells was modified from the procedure of Hodgkin and Kaiser (10). Donor and recipient strains were grown in CTT liquid medium to a density of approximately 100 Klett units (Klett 100; 5 × 10⁸ cells/ml). Strain ASX1 was used as the recipient, and strains DK3685, ASX36, and ASX34 were used individually as donors. Suspensions of 50 μl of donor and recipient cells were prepared and mixed, and 2-μl droplets were placed on CTT (1.5% agar) plates that had been prepared the previous day. After incubation of the plates for 5 h at 25°C, the edges of the droplets were visualized in an inverted microscope, and the colony edges were recorded.

Expression of CglBΔ21-His in E. coli and antiserum production.

An overnight culture of E. coli BL21(DE3) harboring pET21cglBΔ21 was used to inoculate LB medium containing 50 μg of ampicillin per ml. The culture was incubated at 37°C until an optical density at 600 nm of 0.6 to 0.8 was reached, isopropylthiogalactoside (IPTG) was added to a final concentration of 0.4 mM, and growth was allowed to continue for 1 h. CglBΔ21-His protein was found to be present predominantly in inclusion bodies. Inclusion bodies were solubilized with 6 M urea, and the CglBΔ21-His protein was further purified under denaturing conditions according to the protocol recommended by Novagen, except that a wash buffer with higher imidazole concentration (38 mM) was used. To refold the protein, the fractions containing isolated CglBΔ21-His protein were dialyzed against successive changes of 20 mM Tris (pH 7.9) (buffer A) containing 4 M urea, buffer A containing 2 M urea, and buffer A with no denaturant agent. The protein was finally concentrated in a microconcentrator (Centricon). The purified CglBΔ21-His was used by Josman Laboratories (Napa, Calif.) for the preparation of polyclonal antiserum.

SDS-PAGE and Western immunoblot analysis.

To obtain M. xanthus protein extracts, cells were grown in liquid culture to Klett 100, harvested by centrifugation (10,000 × g, 10 min), and resuspended to a density of Klett 1,000 in B buffer (50 mM Tris-HCl [pH 8], 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride). Cells were disrupted by sonication in ice. Samples were centrifuged for 40 min at 30,000 × g (4°C) to remove cell wall components and debris. The supernatant was ultracentrifuged at 200,000 × g for 1 h to obtain the soluble (supernatant) and membrane (pellet) fractions. Membranes were resuspended to 1/10 the original volume in the same buffer. Samples solubilized with sodium dodecyl sulfate (SDS)–β-mercaptoethanol loading buffer (pH 6.8) were separated by polyacrylamide gel electrophoresis (PAGE) on SDS–12% polyacrylamide gels (19). Separated proteins were transferred to a polyvinylidene difluoride membrane (Bio-Rad), using a Bio-Rad mini-transblot cell. Antisera were diluted 1:5,000, and the blots were developed with chemiluminescence reagent (Renaissance; DuPont NEN).

Analysis of gliding movements of single cells.

Gliding movements of individual cells were recorded and quantified as described previously (30).

Nucleotide sequence accession number.

The cglB nucleotide sequence was deposited at GenBank under accession no. AF032467.

RESULTS

Characterization of the motility defect of cglB mutant cells.

Colonies of M. xanthus wild-type strain DK1622 expand as swarms on agar plates. The perimeter of a colony consists of isolated cells and of groups of cells (Fig. 2a). However, both single and small groups of cells are absent at the perimeter of A-motility mutant colonies (11), resulting in sharp, well-defined edges. This macroscopically visible phenotype is found in cglB mutant colonies (Fig. 2b).

FIG. 2.

Colony morphology of M. xanthus wild-type and cglB motility mutant strains. (a) Wild-type (DK1622); (b) ASX1 (ΔcglB S⁺); (c) ASX2 (ΔcglB ΔpilA) and (d) ASX21 (ASX1 containing p343B4). Note the absence of individual and small groups of cells at the perimeter of the ΔcglB strain ASX1. Genetically complemented strain ASX21 (ASX1 containing p343B4) exhibited a colony morphology that resembled that of DK1622 (wild type). All swarming was abolished in the A⁻ S⁻ double-deletion mutant ASX2. For strain construction, see Materials and Methods. Bars indicate 100 μm.

Microscopic analysis, using time-lapse videomicroscopy of single M. xanthus wild-type cells, has revealed that this microorganism moves in two distinguishable patterns (30). When wild-type cells are separated by 0.5 μm or less, the average gliding velocity is 5.0 μm/min (±2.6 μm/min). When cells are separated by more than 0.5 μm, the average gliding velocity decreases to 3.8 μm/min (±1.4 μm/min) and remains constant with increasing cell-cell distance (up to 8 μm measured). To examine the effect of the motility defect of a cglB mutation on individual cell movements, we performed motion analysis of single cells of strain JZ315, which carry a Tn5phoA insertion in the cglB gene (16). Gliding movements of 50 cells were recorded and analyzed as previously described (30). Gliding velocities and cell-cell distances were calculated and are shown in Fig. 3. No single-cell movement was detected when cglB mutant cells were separated by more than approximately 2.5 μm. Outside that range of cell-cell distance, the velocities recorded were below the detection limit of active movement (i.e., less than 1 μm/min [30]). The average velocity of cells in close proximity (cell-cell distance of 0 to 2.5 μm) was 3.2 μm/min (±2.3 μm/min). These microscopic observations, in conjunction with the colony edge phenotype, demonstrate that cglB mutant cells are defective in gliding movement of isolated single cells.

FIG. 3.

Gliding motility of individual cglB mutant cells. Gliding movements of 50 cells of strain JZ315 were analyzed, and gliding speeds as well as the shortest distance of a moving cell to its nearest neighbor were measured (30). Of the 6,127 speed-distance pairs measured, 4,671 and 1,456 speed values were recorded for cell-cell distances of less than or equal to 0.5 μm and larger than 0.5 μm, respectively. The apparent gap in the speed values corresponding to the distance range of 0.1 to 0.5 μm is due to low spatial resolution. In that range, separated and touching cells cannot be distinguished well, and the distances were recorded as 0.0 μm (representing touching cells) in most cases. The dotted line indicates the threshold for detecting active movement in this assay.

Cloning of the cglB locus.

Previous results have shown that in merodiploids of M. xanthus, the A-motility phenotype of cglB⁺ is dominant over cglB mutations (1). Therefore, plasmids carrying a cglB⁺ allele should restore normal adventurous motility upon introduction into cglB mutants. We used rescue of the swarming defect of cglB mutants as a functional assay to clone the cglB wild-type allele.

M. xanthus DK1932 carries transposon Tn5 Ω1932, which is 90% linked to the wild-type cglB locus (29). The cglB locus was cloned directly from a chromosomal digest of DK1932 DNA (Fig. 1). Genomic DNA from strain DK1932 was digested with SacI, the pool of fragments was ligated into SacI-digested pBluescript II SK, the ligated plasmids were electroporated into E. coli DH10B, and transformants were selected for resistance to kanamycin. Plasmid pBSKS-1, conferring kanamycin and ampicillin resistance, was then electroporated into M. xanthus cglB mutant strains DK321 and DK1218. Since pBluescript does not replicate in M. xanthus, Km^r electroporants most likely resulted from a single crossover event where the SacI fragment of the plasmid recombined with the homologous region of the M. xanthus chromosome. This recombination created a tandem duplication of the cloned myxococcal DNA. The resulting Km^r electroporants were then scored for regained A-motility, and introduction of pBSKS-1 was found to restore A-motility in cglB2 mutants DK321 and DK1218 at frequencies of 61 and 68%, respectively.

A restriction analysis of the 13-kb SacI fragment cloned in pBSKS-1 showed that Tn5 Ω1932 was inserted centrally within the fragment (Fig. 1). To localize the cglB locus with respect to the Tn5 insertion, the chromosomal DNA flanking both right and left ends of the transposon was cloned into plasmid vector pPLH343 to generate the clones p343H3 (left end) and p343H7 (right end) (Fig. 1A). These plasmids were introduced separately into cglB mutant strains DK321 and DK1218 and integrated at the Mx8 attachment site, which maps at least 2 Mbp away from the cglB locus (4). The use of pPLH343-derived plasmids in this experiment ensured complementation of the cglB mutation, as opposed to gene reconstruction at the chromosomal cglB locus. Plasmid p343H3 was found to restore A-motility to both strains, whereas p343H7 had no effect (Fig. 1A). Smaller fragments of p343H3 were subcloned into pPLH343 (Fig. 1A) and introduced into strains DK321 and DK1218. p343B4 and p343B4R (which contained the same insert but in the opposite orientations) both completely restored A-motility. However, p343E3 only partially restored A-motility. Neither p343BBG2, p343BBG3, nor p343E3R (same insert as in p343E3 but in the opposite orientation [Fig. 1A]) restored A-motility to either strain. The p343B4 construct was then introduced into strains containing the cglB mutant alleles cglB1, cglB3, cglB5, cglB6, cglB8, cglB10, cglB11, cglB12, cglB13, cglB14, cglB15, cglB16, and cglB17 (Table 1) and was shown to complement the A-motility defect in all cases. These complementation experiments suggest that the functional cglB transcription unit lies within a 2.85-kb DNA fragment which extends from the left end of the Tn5 Ω1932 insertion site in DK1932 to the upstream BamHI site (Fig. 1A).

Nucleotide sequence of cglB.

The nucleotide sequence of the 2.85-kb fragment cloned in p343B4 was determined. An analysis of the nucleotide sequence for coding regions in combination with a codon usage table for Myxococcus genes (28) identified two complete ORFs (ORFA and ORFB) (Fig. 1B). A third incomplete ORF, truncated due to the insertion of Tn5 Ω1932, was identified downstream of ORFB. The three ORFs are transcribed in the same direction and may constitute an operon. All of the genes have appropriate codon usage for M. xanthus, the third-position G+C content for ORFA being 84% and that for ORFB being 78%, which is in the lower range of the characteristic third-position G+C content of myxococcal genes. The partial sequence of ORFC showed a third-position G+C content of 90%, which is characteristic of myxococcal genes (28).

ORFA is predicted to start with an ATG codon (nucleotides [nt] 452 to 455) and to end at a TGA stop codon (nt 785 to 787). The encoded protein is predicted to contain 111 amino acids and to have a mass of 12,049 Da. ORFB is predicted to start with an ATG start codon (nt 1063 to 1065), preceded by a potential ribosomal binding site (AAGGA) at nt 1047 to 1051, and ends in a TAG stop codon (nt 2311 to 2313). This ORF is predicted to encode a protein of 416 amino acids with a predicted mass of 44,221 Da (Fig. 4). ORFC is predicted to start with an ATG codon (nt 2395 to 2397). Searches in databases (GenBank release 107.0 [June 1998], EMBL release 48.0 [September 1996], and SWISS-PROT release 35.0 [April 1998] with FASTA, BLAST, and TFASTA (Genetics Computer Group) revealed no significant similarities between the deduced amino acid sequences of ORFA or ORFB and any other known proteins. However, a search for motifs of the deduced proteins encoded by ORFA and ORFB revealed that the ORFB product has an amino-terminal signal peptide sequence typical of lipoproteins (Fig. 4). The amino acid sequence deduced from the partial sequence of ORFC showed a strong similarity (62.5%) to the ribosomal protein S6 modification protein from E. coli (17), and the gene has tentatively been named rimK.

FIG. 4.

Predicted amino acid sequence of CglB. Note the similarity of the N-terminal CglB region to a typical signal sequence of prokaryotic lipoproteins (indicated above the sequence). The three regions of the signature sequence consist of a 1- to 3-amino-acid N-terminal, positively charged region (n-region), followed by a 9- to 15-amino-acid hydrophobic region (h-region) and a polar carboxy-terminal region that starts at the conserved cysteine at position +1 of the predicted mature lipoprotein (not shown) (7, 35). The consensus of a signal peptidase II cleavage site includes L, M, V, A, S, T, I, or F at position −4; L, V, A, T, F, or I at −3; A, S, I, T, V, G, or Q at −2; G, A, or S at −1, and C at +1 (7). Arrow indicates the predicted start of the mature CglB protein after posttranslational modification.

Construction of an ORFB null mutant.

An analysis of the sequencing data showed plasmid p343E3 to contain the complete ORFB gene but without the upstream ORFA gene and without its promoter region (Fig. 1A). Since p343E3 partially rescued the A-motility defect of cglB mutants, it is likely that ORFB is the cglB gene (Fig. 1A). To prove that mutations in ORFB can cause the same A-motility defect as in cglB mutants, we constructed a null mutant of ORFB in M. xanthus. An in-frame deletion of ORFB was constructed to generate plasmid p113ΔcglB as described in Materials and Methods. Plasmid p113ΔcglB, containing the ΔORFB allele and KG cassette, was introduced into M. xanthus strains DK1622 (A⁺ S⁺) and DK10410 (A⁺ S⁻) by electroporation. The Gal^r Km^s colonies were then scored visually for the loss of A-motility. Of all the Gal^r electroporants in both DK1622 and DK10410, 25 and 20%, respectively, were defective in A-motility. Southern blot analysis of the chromosomal DNA of these strains showed that the wild-type copy of ORFB was replaced by the in-frame deleted ORFB copy (data not shown). In contrast, the wild-type copy of ORFB was retained in those Gal^r electroporants that retained A-motility. The strains with the deleted copy of ORFB were designated ASX1 (derived from DK1622) and ASX2 (derived from DK10410) (Fig. 2b and c).

Complementation of the in-frame ORFB null mutation.

To demonstrate that the previously described functional transcription unit of cglB was sufficient to complement the in-frame ORFB deletion, plasmids p343B4, p343B4R, and p343E3 were introduced into the chromosomal Mx8 prophage attachment sites of ASX1 (ΔcglB S⁺) and ASX2 (ΔcglB S⁻). Plasmids p343B4 and p343B4R completely restored A-motility swarming (Fig. 2d), whereas rescue by p343E3 was again only partial. p343B4 and p343B4R contain the complete ORFB plus 1 kb upstream including the putative promoter region. This promoter region is absent in p343E3. Thus, it is possible that in the p343E3 construct, cglB is expressed from a weak promoter within the plasmid or at the Mx8 attachment site. However, since ORFB was sufficient to rescue the A-motility defect of the null mutant, this result demonstrates that ORFB is the cglB gene.

Gliding movements of individual cells of the complemented null mutant strain ASX21 were quantified by time-lapse videomicroscopy. Movements of 35 cells were recorded, and the gliding speeds were calculated. As shown in Fig. 5, gliding motility of individual cells that were separated from each other by more than 2.5 μm was restored. The average gliding speed was 3.83 μm/min (±2.2 μm/min), which is equivalent to the speed of wild-type cells (3.8 μm/min). These observations demonstrate that reconstruction of the cglB gene restores wild-type gliding of individual cells of ASX1.

FIG. 5.

Gliding motility of single cells of M. xanthus ASX21 (ASX1 complemented by p343B4). Gliding movements of 35 individual ASX21 cells were analyzed; 1,476 speed-distance value pairs were plotted. Speeds corresponding to cell-cell distances of less than 0.5 μm are not shown. The dotted line indicates the threshold for detecting active movement in this assay (30).

Construction and properties of the cglB-His allele.

To facilitate biochemical studies on the function of CglB, we constructed plasmid p343B4.His. This construct allowed expression of the wild-type CglB protein with a C-terminal amino acid sequence changed from RQR to RQLEHHHHHH. This cglB-His allele was tested for expressing an active CglB protein by examining the colony morphology and gliding movements of single cells. Plasmid p343B4.His was introduced into the chromosomal Mx8 prophage attachment site of cglB null mutant strain ASX1 to form strain ASX31. Figure 6A shows a typical edge of an ASX31 colony. Single cells and groups of cells are visible at the perimeter to an extent which is indistinguishable from that of wild-type DK1622 or ASX21 colonies (Fig. 2a and d). Gliding movements of 10 individual cells of strain ASX31 was also investigated by time-lapse videomicroscopy. Single cells were observed to glide when well separated from other cells (≥2 μm) and to translocate at an average speed of 3.75 μm/min (±2.06 μm/min) (Fig. 6B). Thus, with respect to colony morphology and single-cell gliding, no difference was discernible between the wild-type and His-tagged alleles of cglB, indicating that the modification of the C terminus of CglB does not affect its function in gliding motility.

FIG. 6.

Complementation of ΔcglB mutant phenotype by the cglB-His allele. (A) Colony morphology of ASX31 (ΔcglB, p343B4.His). Bar indicates 100 μm. Note the similarity of colony edge to that of DK1622 and ASX21 (Fig. 1). (B) Gliding motility of individual cells of M. xanthus ASX31. Gliding movements of 10 individual cells were analyzed; 493 speed-distance value pairs were measured and plotted. Speeds corresponding to cell-cell distances of less than 0.5 μm are not shown. The dotted line indicates the threshold for detecting active movement in this assay.

The distinctive phenotype of cgl mutants is that their motility defect can be rescued by extracellular complementation in cell mixing experiments using donor strains that contain the appropriate cgl wild-type allele (10, 11). This rescue results in transient gliding movements of cgl mutant cells, as indicated by flare formation at the spot edge of the mixed cells. To test whether the CglB-His protein was active in stimulating ΔcglB mutant strain ASX1, a modified stimulation assay was developed (10). Recipients were cells of ASX1, and donors were strains DK3685, ASX36, and ASX34 which were rendered nonswarming by an mglA mutation. Cells of mutants defective in mglA do not exhibit macroscopic movements, and colonies do not show any swarming activity (data not shown). Recipient and donor cells were mixed, and 2-μl spots were placed on CTT (1.5% agar) plates. After an incubation period of 5 h, the edges of the dried droplets were inspected by microscopy. Cell movement as indicated by swarming flares that expand from the edge of the droplet, resulted only from stimulated cells of strain ASX1, because these recipient cells alone are nonswarming, and donor cells do not swarm because of the mglA defect.

As evident in Fig. 7A, donor cells with the cglB wild-type allele, which was expressed from the chromosomal locus, were able to stimulate ΔcglB cells. When strain ASX36, where the wild-type allele was expressed from the chromosomal Mx8 prophage attachment site, was used as donor, a reduced but still noticeable stimulation of ASX1 was observed (Fig. 7B). However, no stimulation was visible when cglB-His served as the donor allele in strain ASX34 (Fig. 7C). These results suggest that ectopic expression of cglB-His in donor cells is insufficient for stimulation of gliding, although cells carrying the cglB-His allele exhibit wild-type motility behaviour.

FIG. 7.

Stimulation of gliding movements in M. xanthus ΔcglB strain ASX1. For details of the stimulation assay, see Materials and Methods. Donor and recipient cells were mixed, and droplets were placed on CTT (1.5%) agar plates. After 5 h of incubation, edges of the droplets were examined by microscopy. Donor strains were DK3685 (A), ASX36 (B), and ASX34 (C). Bars indicate 20 μm.

Expression of CglB in E. coli and generation of anti-CglB polyclonal antibodies.

CglB was expressed in E. coli without the signal peptide (21 N-terminal amino acids) and with a polyhistidine tag at the C terminus to facilitate single-step purification by metal-chelating chromatography using a nickel column. For this purpose, we constructed plasmid pET21cglBΔ21 (see Materials and Methods), which encodes the CglB-polyhistidine fusion protein CglBΔ21-His. Plasmid pET21cglBΔ21 was introduced into E. coli BL21(DE3). Expression of CglBΔ21-His in pET21cglBΔ21 is regulated by the T7 promoter, which is recognized by an IPTG-inducible T7 RNA polymerase encoded by the lysogen of bacteriophage DE3 (33). Expression is also under control of a lac operator immediately downstream of the T7 promoter. After induction of cells containing pET21cglBΔ21 by IPTG, the expression of a protein with the expected molecular mass for CglBΔ21-His (44.4 kDa) was observed on Coomassie blue-stained SDS-polyacrylamide gels (data not shown). Purified protein was injected into rabbits to generate polyclonal anti-CglB antibodies.

Detection and localization of CglB in M. xanthus.

Cell extracts of M. xanthus strains containing the wild-type allele or a mutated allele of cglB were examined for the presence of CglB protein with anti-CglB antibodies. A single band having a positive reaction with the antiserum was detected in cell extracts of DK1622 (wild type). This band had an electrophoretic mobility equivalent to that corresponding to the estimated molecular mass of CglB (44 kDa) (Fig. 8A, lane 2). Extracts of several different cglB mutant strains were also tested for the presence of CglB antigen by Western blot analysis. The strains tested included three different classes: (i) those derived by UV or chemical agent mutagenesis (DK307, DK321, DK331, DK335, DK344, DK348, DK352, DK353, DK355, DK357, DK377, DK379, DK382, and DK388); (ii) a null mutant generated by in-frame deletion of cglB (ASX1); and (iii) a knockout mutant that contains a Tn5phoA transposon insertion in the cglB gene (JZ315). Of the cglB mutants tested, extracts of strains ASX1, DK321, DK331, DK335, DK352, DK355, DK357, DK377, DK379, DK382, DK388, and JZ315 showed no reaction with anti-CglB antiserum (data not shown). However, strains DK307, DK344, DK348, and DK353 still contained the CglB protein as detected by Western blot analysis (data not shown). Since these strains lack A-motility, the CglB protein in these strains is presumably inactive.

FIG. 8.

Localization and expression levels of CglB antigen determined by PAGE and Western blot analysis using polyclonal anti-CglB antiserum. (A) Cell extract (lane 2) of DK1622 was separated into cytoplasmic (lane 3) and membrane (lane 4) fractions. Only a single 44-kDa band cross-reacted with the anti-CglB antibody in the total-cell extract and in the membrane fraction. Lanes 2 to 4 contain about 20 μg of protein. Lane 1, size markers. (B) Expression levels of CglB in DK1622 (lane 1), ASX21 (lane 2), DK1622 (lane 3), and ASX31 (lane 4). Each lane contains about 12 μg of cellular protein.

To localize the CglB protein, cell extract of M. xanthus DK1622 was separated into membrane and cytoplasmic fractions (see Materials and Methods). CglB protein was detected in total-cell extracts and in the membrane fraction (Fig. 8A, lanes 2 and 4) but not in the cytoplasmic fraction (Fig. 8A, lane 3). Subcellular fractions of the mutant strains DK307, DK344, DK348, and DK353 were prepared and analyzed by Western blotting. In these mutants, CglB protein was also detected in the membrane fraction (data not shown).

Cellular levels of CglB were investigated in M. xanthus strains where cglB or cglB-His was expressed ectopically. As evident in Fig. 8B, when cglB was expressed from the chromosomal Mx8 prophage attachment site, the amount of the protein detected by Western blot analysis was slightly less than in the wild type. More noticeably, when the cglB-His allele was expressed ectopically, the level of CglB antigen was reduced. These results suggest that (i) expression of cglB from the chromosomal Mx8 prophage attachment site results in slightly reduced level of CglB protein and (ii) the C-terminal modification in CglB-His contributes to a further reduction in cellular CglB level.

DISCUSSION

This is the first report of a genetic and molecular characterization of a locus that is specifically involved in single-cell gliding (A-motility) of M. xanthus. By genetic complementation of the gliding defect of a cglB mutant, we isolated the functional transcriptional unit of the cglB gene from a 2.85-kb chromosomal DNA fragment. We also report the identification of CglB protein in cell extracts, its subcellular localization, and initial experiments to elucidate its function in single cell gliding.

The nucleotide sequence of the 2.85-kb fragment contains two complete ORFs (ORFA and ORFB) and one incomplete ORF (ORFC). Several lines of evidences indicate that ORFB is the cglB gene. (i) The DNA fragment extending from the left end of Tn5 to the EcoRI site (Fig. 1A) contains only ORFB and 13 bp upstream of the ATG start codon. This fragment partially complements the A-motility defect of cglB mutants when cloned in one orientation but not in the other. Since the 13-bp upstream of the start codon are insufficient to accommodate a promoter region, ORFB is presumably being expressed from a promoter within the plasmid or chromosomal region (attB) where the plasmid inserted. (ii) A 663-bp in-frame deletion of ORFB was constructed and used to replace the wild-type copy in M. xanthus. By constructing this in-frame deletion mutation, we observed that the motility phenotype of the null mutant was due solely to the deletion of ORFB (Fig. 2). (iii) The defects in swarming and in single cell gliding of the ORFB null mutant ASX1 could be complemented with plasmid p343B4 that contains ORFB (Fig. 2 and 5). Taken together, these observations demonstrate that ORFB is the cglB gene.

The 416-amino-acid protein encoded by ORFB (cglB) did not show significant similarity to any known protein in the databases. However, a search for protein motifs revealed that it has a 19-amino-acid N-terminal sequence that is typical of prokaryotic lipoproteins (Fig. 4) (7, 35). This signal peptide contains two positively charged residues, arginine and lysine (at positions 2 and 4, respectively), followed by a 12-amino-acid hydrophobic region (LPLLSALSVGAV, positions 5 to 16). Immediately flanking this region is a signal peptidase II recognition site (VLAC, positions 17 to 20). Thus, the mature form of CglB is predicted to be 397 amino acids in length and to start at amino acid 20.

The mature CglB protein is predicted to contain 17 cysteine residues, an unusually high number for a lipoprotein (Fig. 4). These cysteines cluster in regions that resemble to some degree an EGF (epidermal growth factor)-like domain. EGF-like domains include six cysteine residues over a length of 29 amino acids with a consensus pattern for the last three cysteines of CxCx₅Gx₂C. These six cysteines in EGF-like domains have been shown to form disulfide bonds. In the CglB protein, the spacing of cysteines is not conserved compared to an EGF-like domain; however, the protein has three regions, amino acids 71 to 93, amino acids 216 to 248, and amino acids 305 to 348, with a high proportion of cysteine residues (four to five). Although the overall similarity to an EGF motif is weak, and a functional significance is not yet clear, mutations of the cysteine residues in these regions should reveal whether they are important for CglB function.

CglB has been detected in the membrane fraction of M. xanthus by Western blot using anti-CglB antibodies (Fig. 8). In wild-type M. xanthus, the anti-CglB antibodies reacted specifically with a protein of 44 kDa, the expected mass for CglB. In extracts of the cglB null mutant and certain cglB mutants, no positive reaction was detected. During maturation of CglB, a 19-amino-acid N-terminal peptide sequence is presumably removed by signal peptidase II, and the cysteine at position 20 covalently bound to a diacylglycerol moiety (Fig. 4). These two modifications (excision and addition) may compensate for each other, explaining why the CglB protein detected in M. xanthus has the same molecular mass as the one predicted from its amino acid sequence. Interestingly, the M. xanthus Tg1 protein, which is encoded by an S-motility gene, appears also to be a lipoprotein (23, 24). Similar to the case for CglB, the motility phenotype of tgl mutants can be rescued by extracellular stimulation in mixing experiments with cells that contain a wild-type tgl allele.

Individual isolated cells of cglB mutants (ASX1 and JZ315) are completely defective in gliding motility (Fig. 2 and 3). The movement retained (Fig. 2 and 3; cell-cell distance of 0 to 2 μm) is probably due to type IV pilus-dependent S-motility (30, 36–39). The gliding defect of isolated single cglB cells can be rescued by extracellular complementation upon mixing of mutants with other cells that contain a cglB wild-type allele (reference 10 and Fig. 7). This interesting phenotype raises the possibility that CglB functions (i) as an essential component in the assembly or in the mechanics of the gliding apparatus that can be transferred between cells or (ii) as a signaling molecule which signals single cells to glide away from groups of cells. To help distinguish between the two models, the motility experiments conducted with the cglB-His allele may provide useful insights because they show that the two functions of cglB, in single-cell gliding and in stimulation of cells to move, can be separated. As shown in this study (Fig. 2 and 6), the wild-type and histidine-tagged alleles of cglB showed virtually identical levels of rescue of the ΔcglB mutant phenotype with respect to swarming movements and single-cell gliding. The C-terminal addition of a glutamate and six histidine residues and the change of the last amino acid, Arg 416, to leucine do not affect the essential function of CglB in gliding (Fig. 6). In all of these experiments, the cglB alleles were expressed from the chromosomal Mx8 prophage attachment site. However, when the ability to stimulate ΔcglB cells to move was investigated, a noticeable difference between these two alleles was observed. The wild-type but not the cglB-His allele was able to stimulate movements of ΔcglB mutant cells (Fig. 7). This differential stimulation correlates with the relative expression levels of the respective CglB proteins; when both are expressed ectopically, the cellular protein level of CglB-His is lower than that of CglB (Fig. 8B). Although the reason for the reduced expression level is unknown, the results may suggest that extracellular stimulation by CglB, as examined by this laboratory assay, is not essential for single-cell gliding, because this cglB-His allele promotes wild-type gliding (Fig. 6).

A mechanical role of CglB in single-cell gliding may be suggested by its relative abundance in the membrane, presumably the outer membrane. On the basis of rough estimates from Western blot analysis, a typical M. xanthus wild-type cell contains between 10⁴ and 10⁵ molecules of CglB. Such a density of CglB in the outer membrane is consistent with a structural role for this protein in gliding motility. Future structure-function analysis of CglB should provide insight into the mode of CglB action.