Transposon Insertions of magellan-4 That Impair Social Gliding Motility in Myxococcus xanthus

Abstract

Myxococcus xanthus has two different mechanisms of motility, adventurous (A) motility, which permits individual cells to glide over solid surfaces, and social (S) motility, which permits groups of cells to glide. To identify the genes involved in S-gliding motility, we mutagenized a ΔaglU (A⁻) strain with the defective transposon, magellan-4, and screened for S⁻ mutants that form nonmotile colonies. Sequence analysis of the sites of the magellan-4 insertions in these mutants and the alignment of these sites with the M. xanthus genome sequence show that two-thirds of these insertions lie within 27 of the 37 nonessential genes known to be required for social motility, including those necessary for the biogenesis of type IV pili, exopolysaccharide, and lipopolysaccharide. The remaining insertions also identify 31 new, nonessential genes predicted to encode both structural and regulatory determinants of S motility. These include three tetratricopeptide repeat proteins, several regulators of transcription that may control the expression of genes involved in pilus extension and retraction, and additional enzymes involved in polysaccharide metabolism. Three insertions that abolish S motility lie within genes predicted to encode glycolytic enzymes, suggesting that the signal for pilus retraction may be a simple product of exopolysaccharide catabolism.

MYXOCOCCUS xanthus is a soil bacterium that can glide over solid surfaces without the use of flagella. The genetic analysis of its ability to glide shows that M. xanthus has two different sets of genes involved in two very different mechanisms of motility. Most single mutations affect the ability of M. xanthus to glide either as single cells or as groups of cells, but not as both. General searches for mutants of M. xanthus with motility defects have shown that single-cell (adventurous, A) motility and group-dependent (social, S) motility are genetically separable and that each type of motility requires the functions of large, nonoverlapping sets of genes (Hodgkin and Kaiser 1979a,b; Macneil et al. 1994a,b; Youderian et al. 2003). Mutations in only three genes, mglA, which encodes a small GTPase (Stephens and Kaiser 1987; Stephens et al. 1989; Hartzell and Kaiser 1991a,b; Hartzell 1997), agmA, which is predicted to encode a critical amidase involved in cell wall biogenesis (Youderian et al. 2003), and epsI/nla24, predicted to encode an activator of transcription (Caberoy et al. 2003; Lancero et al. 2004; Lu et al. 2005), have been shown to abolish both mechanisms of motility simultaneously.

Mutants defective in A motility form colonies with groups of motile cells, but without individual motile cells, at their edges. Adventurous motility in M. xanthus requires the functions of at least 30 different genes, many of which have no known homologs in nongliding bacteria (Youderian et al. 2003). Loss-of-function mutations in each of these genes do not affect social motility. Although it has been proposed that nozzles that cluster at the cell poles may generate force by directional secretion of polymer (Wolgemuth et al. 2002), genetic and biochemical support for this model is forthcoming. In contrast, social motility in M. xanthus involves the interaction between two organelles, type IV pili and exopolysaccharide/fibril material, and is much better understood.

Mutants defective in S motility form colonies with individual motile cells, but without groups of motile cells, at their edges. Social motility results from a cycle in which the cell extends a type IV pilus, the pilus tip attaches to a receptor, and the pilus is retracted (Wu and Kaiser 1995; Merz et al. 2000; Sun et al. 2000; Skerker and Berg 2001; Li et al. 2003). This mechanism appears to be conserved among diverse Gram-negative bacteria, including M. xanthus, Pseudomonas aeruginosa, and Haemophilus influenzae (see Mattick 2002, for review). For M. xanthus, the substrate for pilus attachment is presented on the surface of a partner (receptor) cell within about one cell length of a pilus-producing cell (Hodgkin and Kaiser 1977; Wu and Kaiser 1995). Previous biochemical studies have shown that peritrichous fibrils constitute the receptor organelle required for S motility; these are composed of both protein and exopolysaccharide (EPS) components (Behmlander and Dworkin 1991; Behmlander and Dworkin 1994). The EPS component, but not the protein component, of fibrils appears to be essential for social motility and cell–cell adhesion. Mutants defective in EPS production are defective in S motility (Lu et al. 2005), and EPS stimulates pilus retraction (Li et al. 2003).

Our current understanding of S motility is missing several key elements. Mutants defective in lipopolysaccharide (LPS) production also are defective in S motility (Bowden and Kaplan 1998; Yang et al. 2000a); however, as yet, we do not know what role LPS plays in the mechanism of social motility. Also, we do not understand the molecular mechanism of pilus extension and retraction, nor have we identified the full set of genes encoding the cytoplasmic and membrane components of this molecular motor. Other important details missing from our picture of the mechanism of S motility in M. xanthus include an understanding of how the signals that trigger pilus retraction are generated and transduced upon the contact of a pilus-bearing cell with a receptor-bearing cell, how the next cycle of pilus extension and retraction is reinitiated after cell–cell contact, and what the signals are precisely for pilus extension and retraction. It is not known whether the cycle of pilus extension and retraction involves the polymerization and depolymerization of the pilus. To begin to define the genes encoding products involved in these processes, we have conducted an extensive screen for mutants with transposon insertions in the nonessential genes required for S motility.

MATERIALS AND METHODS

Bacterial strains and growth conditions:

The M. xanthus strains we have generated in this study are derivatives of the wild-type strain DK1622, and its A⁻ mutant derivative MxH1777 (ΔaglU) (White and Hartzell, 2000), and are listed in Table 1. M. xanthus was grown at 32° in CTPM liquid medium (1% casitone, 10 mm Tris pH 7.6, 1 mm potassium phosphate pH 7.5, 5 mm MgSO₄) and CTPM agar (1.5%) plates; CTPM was supplemented with kanamycin (Km; 40 μg/ml). Plasmids were introduced into M. xanthus by electroporation (Kashefi and Hartzell 1995; Youderian et al. 2003). Escherichia coli strain DH5α (λ pir) was used for the construction of plasmids and the preparation of plasmid DNA. Plasmids were introduced into this strain by electroporation, and derivatives with plasmids were grown in LB medium supplemented with Km (40 μg/ml). Plasmid pMycoMar, donor of the mini-mariner element magellan-4, has been described (Rubin et al. 1999). Restriction endonucleases and DNA modifying enzymes were from New England Biolabs (Beverly, MA) and were used under recommended conditions for plasmid constructions.

TABLE 1.

Insertions of magellan-4 that impair social gliding motility

Coordinates	Locus	Insertion	Strain	Gene
0421763–0421764	MXAN0358	mis-23	MxH1521	sgmA; ileS
0500448–0500449	MXAN0440	mis-66	MxH1561	sgmB
1287846–1287847	MXAN1106	mis-15	MxH1513	sgmC
1288161–1288162	MXAN1106	mis-77	MxH1572	sgmC
1946011–1946012	MXAN1641	mis-146	MxH1629	sgmD
2122234–2122235	MXAN1795	mis-17	MxH1515	sgmE
2259176–2259177	MXAN1926	mis-61	MxH1557	mglB
2462808–2462809	MXAN2128	mis-114	MxH1604	sgmF
2542123–2542124	MXAN2203	mis-7	MxH1505	sgmG
2939846–2939847	MXAN2526	mis-10	MxH1508	sgmH
2981508–2981509	MXAN2561	mis-96	MxH1589	sgmI
3422255–3422256	MXAN2921	mis-112	MxH1603	sgmJ
3422987–3422988	MXAN2922	mis-119	MxH1609	sgmK
3422987–3422988	MXAN2922	mis-129	MxH1616	sgmK
3589491–3589492	MXAN3060	mis-141	MxH1626	cglB
3589491–3589492	MXAN3060	mis-147	MxH1630	cglB
3611423–3611424	MXAN3084	mis-2	MxH1501	tgl
3611426–3611427	MXAN3084	mis-85	MxH1580	tgl
3611426–3611427	MXAN3084	mis-88	MxH1583	tgl
4055186–4055187	MXAN3474	mor-105	MxH1598	stk
4084951–4084952	MXAN3506	mis-19	MxH1517	sgmL
4084987–4084988	MXAN3506	mis-131	MxH1618	sgmL
4085119–4085120	MXAN3506	mis-26	MxH1524	sgmL
4085119–4085120	MXAN3506	mis-127	MxH1614	sgmL
4477791–4477792	MXAN3759	mis-142	MxH1627	sgmM
4563759–4563760	MXAN3797	mis-51	MxH1548	sgmN
4619415–4619416	MXAN3840	mis-121	MxH1611	scpB
4672427–4672428	MXAN3886	mis-116	MxH1606	agmA
5036166–5036167	MXAN4089	mor-115	MxH1605	carR
5096652–5096653	MXAN4149	mis-75	MxH1570	frzS
5098300–5098301	MXAN4150	mis-68	MxH1563	sgmO
5788164–5788165	MXAN4613	mis-135	MxH1622	sgmP
5791219–5791220	MXAN4616	mis-130	MxH1617	sgmQ
5792644–5792645	MXAN4616	mis-101	MxH1594	sgmQ
5792644–5792645	MXAN4616	mis-120	MxH1610	sgmQ
5800139–5800140	MXAN4620	mis-70	MxH1565	sgmR
5801403–5801404	MXAN4621	mis-43	MxH1540	rfbC
5804537–5804538	MXAN4621	mis-102	MxH1595	rfbC
5805964–5805965	MXAN4622	mis-41	MxH1538	rfbB
5807214–5807215	MXAN4623	mis-21	MxH1519	rfbA
5819677–5819678	MXAN4639	mis-28	MxH1526	sgmS
5822620–5822621	MXAN4640	mis-9	MxH1507	sgmT
5901872–5901873	MXAN4707	mis-55	MxH1551	sgmU
6413442–6413443	MXAN5134	mis-122	MxH1612	mutS
6640018–6640019	MXAN5333	mis-24	MxH1522	sgmV
6947398–6947399	MXAN5592	mis-42	MxH1539	sgmW
7140653–7140654	MXAN5766	mis-100	MxH1593	sgmX
7142162–7142163	MXAN5766	mis-13	MxH1511	sgmX
7145304–7145305	MXAN5770	mis-73	MxH1568	sgmY
7146533–7146534	MXAN5772	mis-97	MxH1590	pilQ
7147173–7147174	MXAN5772	mis-63	MxH1559	pilQ
7148427–7148428	MXAN5772	mis-50	MxH1547	pilQ
7148895–7148896	MXAN5772	mis-29	MxH1527	pilQ
7148895–7148896	MXAN5772	mis-30	MxH1528	pilQ
7149164–7149165	MXAN5773	mis-103	MxH1596	pilP
7149521–7149522	MXAN5772	mis-31	MxH1529	pilP
7149933–7149934	MXAN5774	mis-78	MxH1573	pilO
7149933–7149934	MXAN5774	mis-79	MxH1574	pilO
7151318–7151319	MXAN5776	mis-39	MxH1536	pilM
7151453–7151454	MXAN5776	mis-44	MxH1541	pilM
7152059–7152060	MXAN5776	mis-118	MxH1608	pilM
7152059–7152060	MXAN5776	mis-128	MxH1615	pilM
7152136–7152137	MXAN5776	mis-94	MxH1588	pilM
7152564–7152565	MXAN5777	mis-108	MxH1601	nla23
7152889–7152890	MXAN5777	mis-58	MxH1554	nla23
7152889–7152890	MXAN5777	mis-86	MxH1581	nla23
7152889–7152890	MXAN5777	mis-89	MxH1584	nla23
7156382–7156383	MXAN5780	mis-64	MxH1560	pilI
7156382–7156383	MXAN5780	mis-133	MxH1620	pilI
7156717–7156718	MXAN5780	mis-69	MxH1564	pilI
7156717–7156718	MXAN5780	mis-104	MxH1597	pilI
7157002–7157003	MXAN5781	mis-74	MxH1569	pilH
7157880–7157881	MXAN5782	mis-144	MxH1628	pilG
7157964–7157965	MXAN5782	mis-99	MxH1592	pilG
7158021–7158022	MXAN5782	mis-76	MxH1571	pilG
7158154–7158155	MXAN5782	mis-80	MxH1575	pilG
7158204–7158205	MXAN5782	mis-33	MxH1531	pilG
7158273–7158274	MXAN5782	mis-83	MxH1578	pilG
7158288–7158289	MXAN5782	mis-4	MxH1502	pilG
7158344–7158345	MXAN5782	mis-1	MxH1500	pilG
7158429–7158430	MXAN5782	mis-106	MxH1599	pilG
7158429–7158430	MXAN5782	mis-107	MxH1600	pilG
7159294–7159295	MXAN5783	mis-91	MxH1585	pilA
7159294–7159295	MXAN5783	mis-92	MxH1586	pilA
7159632–7159633	MXAN5783	mis-84	MxH1579	pilAa
7160010–7160011	MXAN5784	mis-132	MxH1619	pilR
7162816–7162817	MXAN5786	mis-14	MxH1512	pilC
7162902–7162903	MXAN5786	mis-32	MxH1530	pilC
7163164–7163165	MXAN5786	mis-134	MxH1621	pilC
7163494–7163495	MXAN5786	mis-11	MxH1509	pilC
7163494–7163495	MXAN5786	mis-60	MxH1556	pilC
7163494–7163495	MXAN5786	mis-123	MxH1613	pilC
7164417–7164418	MXAN5787	mis-34	MxH1532	pilT
7164529–7164530	MXAN5787	mis-82	MxH1577	pilT
7164529–7164530	MXAN5787	mis-87	MxH1582	pilT
7166434–7166435	MXAN5788	mis-36	MxH1534	pilB
7231003–7231004	MXAN5831	mis-148	MxH1631	sgmZ; glgP
7572031–7572032	MXAN6125	mis-109	MxH1602	sgnA
8039216–8039217	MXAN6518	mis-40	MxH1537	sgnB
8154610–8154611	MXAN6627	mis-52	MxH1549	sgnC
8200210–8200211	MXAN6671	mis-38	MxH1535	sglK
8200360–8200361	MXAN6671	mis-45	MxH1542	sglK
8200465–8200466	MXAN6671	mis-48	MxH1545	sglK
8200465–8200466	MXAN6671	mis-49	MxH1546	sglK
8461333–8461334	MXAN6908	mis-35	MxH1533	sgnD; pgi
8461447–8461448	MXAN6908	mis-71	MxH1566	sgnD; pgi
8461447–8461448	MXAN6908	mis-93	MxH1587	sgnD; pgi
8982735–8982736	MXAN7360	mis-59	MxH1555	sgnEa
9039673–9039674	MXAN7415	mis-18	MxH1516	epsZ
9039673–9039674	MXAN7415	mis-20	MxH1518	epsZ
9039673–9039674	MXAN7415	mis-47	MxH1544	epsZ
9040255–9040256	MXAN7415	mis-57	MxH1553	epsZ
9040369–9040370	MXAN7415	mis-72	MxH1567	epsZ
9042228–9042229	MXAN7417	mis-117	MxH1607	epsY
9042228–9042229	MXAN7417	mis-136	MxH1623	epsY
9042228–9042229	MXAN7417	mis-138	MxH1624	epsY
9045962–9045963	MXAN7421	mis-62	MxH1558	epsV
9046689–9046690	MXAN7422	mis-22	MxH1520	epsU
9047313–9047314	MXAN7422	mis-46	MxH1543	epsU
9063422–9063423	MXAN7441	mis-81	MxH1576	epsH
9065683–9065684	MXAN7442	mis-6	MxH1504	sgnF
9065986–9065987	MXAN7442	mis-27	MxH1525	sgnF
9066004–9066005	MXAN7442	mis-67	MxH1562	sgnF
9070550–9070551	MXAN7445	mis-25	MxH1523	epsE
9071704–9071705	MXAN7448	mis-56	MxH1552	epsD
9072040–9072041	MXAN7448	mis-53	MxH1550	epsD
9072433–9072434	MXAN7448	mis-8	MxH1506	epsD
9072508–9072509	MXAN7448	mis-16	MxH1514	epsD
9072734–9072735	MXAN7449	mis-5	MxH1503	epsC
9072805–9072806	MXAN7449	mis-12	MxH1510	epsC
9072805–9072806	MXAN7449	mis-139	MxH1625	epsC
9076128–9076129	MXAN7451	mis-98	MxH1591	epsA

The 132 insertions of magellan-4 in the M. xanthus genome described in this report are listed in order of their sites within the genome sequence, which can be found at http://cmr.tigr.org/tigr-scripts/CMR/GenomePage.cgi?org=gmx. Given are the gene numbers in which the insertions are situated, the allele numbers of the insertions, the strain numbers of the derivatives of MxH1777 with each insertion, and the gene names.

^aAn insertion lies upstream of the coding sequence indicated.

Isolation and phenotypic screening of potential social motility mutants:

Electroporation of MxH1777 cells with plasmid pMycoMar was performed as described (Youderian et al. 2003). Electroporation mixes were plated on CTPM Km agar and incubated for 5 days at 32°. After incubation, plates were screened visually to identify small colonies with smooth edges, the phenotype of A⁻S⁻ double mutants. Mutants were purified twice, and the phenotypes of single colonies formed by each mutant were compared after each purification step. Colonies displaying reduced motility after the initial plating were chosen conservatively for purification, and, of these, 149/154 colonies displayed a stable A⁻S⁻ phenotype. We found that one of the 149 mutants, MxH1777 mor-105, formed colonies with an unusual morphology; these colonies were smaller in diameter and more than twice the height of the colonies formed by the parent or other mutant strains and have a wrinkled appearance. Only 5/154 colonies chosen after the initial plating were found to form colonies with the same size as the parental (A⁻S⁺) strain after purification. Four of these were not considered in our subsequent analyses, but the fifth mutant was passed through our sequence analysis, because it had a colony morphology strikingly different from that of the parent and formed bright orange colonies. This mutant was designated MxH1777 mor-115 (for “morphology”). The alleles of magellan-4 insertions in all other mutants were designated mis (“mutation in S”) followed by their isolation number. Five of the 149 mutants, with alleles mis-121, mis-122, mis-141, mis-146, and mis-147, were found to form colonies with sizes intermediate between those formed by the parental strain and mutants with more severe motility defects (see results).

Cloning of M. xanthus genomic DNA flanking magellan-4 insertions in S genes:

To subclone magellan-4 insertions in S genes, M. xanthus genomic DNA was isolated from vegetative cultures of MxH1777 S∷magellan-4 strains, using the method for the rapid isolation of myxophage Mx8 DNA (Magrini et al. 1997). Each strain was grown to a density of 5 × 10⁸/ml in CTPM Km medium, and 1 ml of cells were harvested by low speed centrifugation and resuspended in 0.4 ml distilled water. Buffer [50 μl of 2.0 m Tris–HCl (pH 8.0) and 0.1 m ethylenediamine tetraacetate (EDTA)], 2 μl diethylpyrocarbonate, and 10 μl 10% sodium dodecyl sulfate (SDS) were added. Mixtures were incubated at 75° for 5 min, and 50 μl of 5 m potassium acetate was added to precipitate SDS-protein complexes. After incubation at 4° for 2 hr, mixtures were microcentrifuged at 10,000 × g for 15 min. Supernatants were precipitated with 2.5 volume ethanol, and pellets were rinsed in 70% ethanol, 10 mm Tris–HCl (pH 8.0), 1 mm EDTA, and 10 mm MgCl₂ (DNA wash solution) for 5 min. DNA pellets were air dried, resuspended in 200 μl 10 mm Tris–HCl (pH 8.0) and 0.1 mm EDTA, and stored routinely at −20°. Genomic DNA (2 μl) was cleaved with BssHI or NotI for 8–12 hr in a total volume of 12 μl, enzymes were inactivated by incubation at 75° for 15 min, and the mixtures were placed on a 0.025-μm filter (Millipore, Bedford, MA) and dialyzed against 1000 volume distilled water for 30 min. Cleaved DNA (8 μl) was treated with T4 DNA ligase in a total volume of 10 μl at 25° for 4–12 hr, incubated at 65° for 20 min, and drop dialyzed prior to electroporation into E. coli DH5α (λ pir). Approximately 1–100 Km^R electroporants were obtained after electroporation of 1 μl of each ligated DNA and were recovered on LB Km plates after incubation at 37° for 24 hr. We were successful in cloning and sequencing the magellan-4 insertions present in 134/154 mutants. The MxH (M. xanthus Hartzell) strain names for mutants with mis or mor insertions are listed in Table 1.

Sequence analysis:

Plasmid DNAs with subcloned BssHII fragments were sequenced initially with primers Mar1 and Mar2 (Biosource, Camarillo, TX), complementary to the ends of magellan-4 (Youderian et al. 2003). In several cases, we extended these sequences with additional primers or by sequencing plasmids with NotI subclones from the same mutant. Sequences were assembled into contigs and then used in BLASTn searches (Altschul et al. 1990) against the DK1622 genome sequence to identify the TA target site for each magellan-4 insertion, given as the coordinates of the M. xanthus sequence (http://cmr.tigr.org/tigr-scripts/CMR/GenomePage.cgi?org=gmx) (Table 1). In all cases, these searches yielded a unique dinucleotide target site of insertion without accompanying deletion or rearrangement. Comparison of our sequences with the M. xanthus genome confirmed the identity of >50 kb of the assembled sequence (our unpublished results).

RESULTS

Genetic screen for magellan-4 insertions in S motility genes:

Transient cis-complementation of defective transposons (DNA elements flanked by transposon ends that are lacking an active transposase gene) by transposase has proven to be a powerful method for performing insertion mutagenesis with transposons. This is because stable insertions that do not undergo additional rounds of transposition are generated in the absence of the transposase gene (Hughes and Roth 1988).

Various strategies have been developed for the transient cis-complementation of defective transposons carried on either linear or circular DNA donor molecules, using transduction, conjugation, or transformation as mechanisms for their delivery. One such strategy involves the transformation of a recipient cell with a plasmid that cannot replicate in recipient cells and that expresses transposase in cis to a defective transposon carrying a selectable antibiotic-resistance determinant. Selection for antibiotic-resistant transformants enriches for recombinants in which the defective transposon has inserted into the target genome; their subsequent outgrowth results in the segregative loss of the donor plasmid molecule, giving rise to mutants with stable insertions.

Plasmid pMycoMar (Rubin et al. 1999) is one such plasmid donor for the transposition of the defective magellan-4 element. This minitransposon has the ends of a Himar element, a transposon in the mariner superfamily (Hartl et al. 1997) that moves from donor to recipient DNA molecules by a “cut and paste” mechanism, dependent only upon transposase (Lampe et al. 1996). These ends flank the npt gene from bacterial transposon Tn5, which confers resistance to kanamycin (Km^R), as well as the R6Kγ origin of replication, which can function only when the Pir initiation protein is supplied in trans. Adjacent to the defective magellan-4 element, pMycoMar can express mariner transposase from the mycobacterial T6 promoter (Barsom and Hatfull 1996). The G + C-rich T6 promoter is active in Mycobacterium tuberculosis but inactive in E. coli (Rubin et al. 1999). We have shown that this promoter also functions well in M. xanthus (Youderian et al. 2003). In an E. coli host that expresses the Pir protein, pMycoMar replicates autonomously, but does not express transposase. In M. tuberculosis and M. xanthus, pMycoMar cannot replicate autonomously and expresses transposase. After electroporation of M. xanthus with pMycoMar, transient cis-complementation results in the transposition of magellan-4 into the target genome and the generation of Km^R insertion mutants; the donor plasmid is lost upon cell division (Youderian et al. 2003).

Because magellan-4 has the R6Kγ origin of replication, it permits the rapid sequence analysis of insertion mutations. When the chromosomal DNA of a mutant with a magellan-4 insertion is restricted with an endonuclease that does not cleave within the element, ligated, and used to transform an E. coli pir+ recipient strain, Km^R recombinants are recovered. Such recombinants carry plasmids with the magellan-4 element and regions of the target genome flanking the site of magellan-4 insertion. Sequences of the sites of each insertion in each plasmid can be determined by using primers internal to, and pointing outward from, the ends of transposed magellan-4 elements.

In another article, we described the use of pMycoMar to generate insertions in adventurous (A) motility genes in M. xanthus. We electroporated a strain of M. xanthus deficient in S motility with pMycoMar, screened for Km^R recombinants that have a nonmotile colony morphology (resulting from an S–A∷magellan-4 genotype), and found such mutants with a frequency of ∼1% (115/12,000). We concluded from this result that magellan-4 has a much broader spectrum of targets in M. xanthus than transposon Tn5, because similar approaches with Tn5 yield only ∼0.2% mutants that form nonmotile colonies. This interpretation is supported by the results of sequence analysis, which enabled us to identify 34 different genes disrupted by 115 independent insertions of magellan-4. These 34 genes included the four genes previously known to be involved in A motility. Because 15 of the 30 new genes we identified in this search were represented only by single insertions of magellan-4, we concluded that we have yet to saturate the genes required for A motility using this transposon (Youderian et al. 2003).

In this report, we describe the results of the reciprocal experiment. We started with a mutant defective in A, but not in S, motility. We chose the ΔaglU-1777 mutant as our starting strain, because we have demonstrated that this allele, when paired with several different S mutant alleles, results in a nonmotile colony phenotype (White and Hartzell 2000). We electroporated this mutant with plasmid pMycoMar and screened for Km^R recombinants that have a nonmotile colony morphology (resulting from a ΔaglU S∷magellan-4 genotype). Again, we find such mutants with a frequency of ∼1% (149/15,000), a frequency about fivefold higher than that obtained with Tn5 in similar mutagenesis experiments (Macneil et al. 1994a). These mutants were purified twice, to confirm that their defects in motility breed true, and grown in liquid CTPM medium (or on plates) to prepare genomic DNA for sequence analysis. Chromosomal DNA from nonmotile, ΔaglU S∷magellan-4 double mutants was purified, cleaved, and ligated and used to transform an E. coli recipient strain that produces the Pir protein; plasmids were prepared from Km^R electroporants of E. coli and sequenced. As shown in Table 1, we determined the sequences of the sites of magellan-4 insertions in each of the plasmids derived from 132 of the 149 independent mutants. All of the insertions were found within the preferred TA dinucleotide target sequence duplicated by magellan-4 upon insertion. For the majority of these insertions, we determined that the TA target site had been duplicated by insertion, by sequencing both insertion junctions.

Insertions of magellan-4 in genes required for type IV pilin biogenesis:

Among the sequenced insertions of magellan-4 that affect S motility, 50 were found in one cluster of genes required for type IV pilin biosynthesis, located in a 27-kb region of the M. xanthus genome (Figure 1). These insertions are distributed among 40 different target sites, and insertions that occur at the same site are often found in opposite orientations. The most frequent target for insertion in this region was the pilG gene, in which 10 insertions were recovered at 9 different sites (with the pair of transposon insertions at the same site found in opposite orientations). These results confirm that the spectrum of insertions made by magellan-4 in M. xanthus is extremely broad and shows little preference in choosing AT target sites within the same gene.

Figure 1.

magellan-4 insertions identify 15 different genes required for S motility in the M. xanthus pil gene cluster. The 27-kb region of the M. xanthus genome with the pilA (pilin) and adjacent genes (boxes) is shown; the directions of transcription are indicated by arrows. Coordinates and gene numbers are those of the M. xanthus genome sequence ( http://cmr.tigr.org/tigr-scripts/CMR/GenomePage.cgi?org=gmx). Positions of magellan-4 insertions are indicated by triangles. Shaded boxes indicate genes shown previously to be required for S motility; solid boxes indicate the new genes identified in this study. The top and bottom represent contiguous sequence.

As shown in Figure 1, The 23 predicted coding sequences in this region of the genome are transcribed in the same direction, most likely as at least five separate operons. These include the pilA gene (MXAN5783), which encodes pilin (Wu and Kaiser 1996); the pilG (MXAN5782), pilH (MXAN5781), and pilI (MXAN5780) genes (Wu et al. 1998); and the pilB (MXAN5788), pilT (MXAN5787), pilC (MXAN5786), and pilQ (MXAN5772) genes, each of which has been shown to be essential for S motility (Wu et al. 1997; Wall et al. 1999). Insertions of magellan-4 were obtained in each of these genes, as well as in pilP (MXAN5773), pilO (MXAN5774), pilM (MXAN5776), pilR (MXAN5784), and nla23(MXAN5777), predicted to encode a response regulator homologous to PilR (Wall et al. 1999; Caberoy et al. 2003). No insertions were found in the large pilS (MXAN5785) gene, consistent with the finding that it is not essential for S motility (Wall et al. 1999).

In addition to genes in this cluster known or suspected to be required for S motility, insertions were recovered in gene MXAN5770, immediately downstream from MXAN5771, which is not required for S motility (Wu and Kaiser 1996). MXAN5770 is predicted to encode a conserved hypothetical protein. Two different insertions were recovered in nearby gene MXAN5766, predicted to encode a product with tetratricopeptide (TPR) repeats (see below).

Insertions of magellan-4 in genes required for exopolysaccharide biosynthesis:

Among the independent magellan-4 insertions that disrupt S motility, 24 disrupt genes known to be required for exopolysaccharide (EPS) biosynthesis. Recently, Lu et al. (2005) identified a large group of genes involved in EPS biosynthesis, which also is required for S motility. Their strategy to identify these genes was to mutagenize M. xanthus by electroporation with a plasmid library made with a Km^R vector that carries random, ∼500-bp inserts of M. xanthus DNA prepared by physical shearing. Because their plasmid vector cannot replicate in M. xanthus, Km^R recombinants must arise due to the integration of the plasmid, which occurs most frequently by homologous recombination in this host, as in the yeast Saccharomyces cerevisiae. Thus, this strategy is equivalent to that of inactivating a yeast gene by the integration of a plasmid that carries an insert missing the 5′ and 3′ ends of a gene; integration of a plasmid results in a merodiploid in which the integrated vector is sandwiched between two incomplete copies of the gene, one missing its 5′ end and one missing its 3′ end, required for function (Neff et al. 1983).

Using this strategy, they screened among 5000 “integrative disruption” mutants and found 68 with reduced motility. Four of these 68 mutants are unable to form fruiting bodies, a phenotype of many S motility mutants. Also, they do not bind calcofluor white, which is an indicator dye for EPS production (Ramaswamy et al. 1997). The integrated plasmids in three of these four integrative disruption mutants were found to flank a 37-kb region of the M. xanthus genome, which they designated the “eps” region.

Sequence analysis shows that this region includes genes predicted to encode products homologous to those of the genes known to be involved in EPS biosynthesis in other bacteria. In-frame deletion and insertion mutations in 11 genes within this region impair social motility (Figure 2). These are: epsA (MXAN7451), epsC (MXAN7449), epsD (MXAN7448), epsE (MXAN7445), epsF (MXAN7444), epsH (MXAN7441), epsI/nla24 (MXAN7440), epsU (MXAN7422), epsV (MXAN7421), epsY (MXAN7417), and epsZ (MXAN7415) (Lu et al. 2005). As shown in Table 1, 24/131 of the independent magellan-4 insertions that disrupt S motility fall within 9 of these 11 genes. Three insertions interrupt the coding sequence MXAN7442, which lies between epsG and epsH and is predicted to encode a membrane protein of unknown function. The predicted products of these 12 genes fall into four functional categories. The majority (7/12) of these are sugar transferases presumably required for EPS assembly, two have response regulator domains (EpsF and EpsI), one is presumably involved in EPS export (EpsY), and two (EpsV and the product of MXAN7442) have as yet unknown functions. We did not recover magellan-4 insertions in the easA (MXAN2293) gene, which encodes an accessory protein essential for EPS biogenesis (Lu et al. 2005).

Figure 2.

magellan-4 insertions identify 10 different genes required for S motility in the M. xanthus eps gene cluster. The details are as in Figure 1; note that the central portion of the eps gene cluster, which does not contain genes required for S motility, is not shown.

Insertions of magellan-4 in genes required for LPS biosynthesis:

Among the independent magellan-4 insertions that disrupt S motility, four fall into a cluster of five adjacent genes. Four of these five genes, rfaA (MXAN4623), rfaB (MXAN4622), rfaC (MXAN4621), and wbgB (MXAN4619), are known to be required for LPS biosynthesis (Bowden and Kaplan 1998; Yang et al. 2000a) (Figure 3, top). Also, one magellan-4 insertion interrupts MXAN4620, predicted to encode a hypothetical protein with unknown function. Four magellan-4 insertions lie within the nearby genes MXAN4616, predicted to encode a glycosyl transferase, and MXAN4613, predicted to encode a second dTDP-glucose 4,6-dehydrogenase similar to that of the rfbB (MXAN4622) gene. These adjacent genes also are likely involved in the biosynthesis of LPS.

Figure 3.

magellan-4 insertions identify six different genes required for S motility in the M. xanthus rfb gene cluster. (Top) Shown are the positions of magellan-4 insertions in the rfb gene cluster, and adjacent genes, as in Figures 1 and 2. (Bottom) Three additional regions of the M. xanthus genome have adjacent genes with multiple insertions of magellan-4 that confer the loss of S motility. The sgmJ and sgmK genes define a new cluster of genes presumably required for LPS biogenesis, sgmO is adjacent to frzS, and the sgmT (histidine kinase/response regulator) and sgmS genes lie immediately upstream of agmH, required for A motility (Youderian et al. 2003). One or both of these genes may be required for both S and A motility.

Insertions of magellan-4 in additional genes known to be required for S motility:

Many additional genes have been characterized that encode products essential for S motility. These include the three genes known to be required for both S and A motility, mglA (MXAN1925), which is the second gene in an operon with its interacting partner, mglB (MXAN1926) (Hartzell and Kaiser 1991a,b; Hartzell 1997), agmA (MXAN3886), which encodes one of five different N-acetylmuramoyl l-alanine transamidases predicted to be made by M. xanthus (Youderian et al. 2003), and epsI/nla24 (MXAN7440), which encodes a response regulator essential for EPS biosynthesis (Caberoy et al. 2003; Li et al. 2003). Among our nonmotile derivatives of parent strain MxH1777, we recovered one mutant with an insertion of magellan-4 in the mglB gene; such mutations also were recovered in our reciprocal hunt for magellan-4 insertions that reduce the motility of a parental S⁻ strain and are defective in A motility; these insertions have polar effects on the expression of mglA (Youderian et al. 2003). In the current hunt, we also recovered one magellan-4 insertion in the agmA gene, confirming that its function is required for both A and S motility (Youderian et al. 2003), but no insertions in the epsI/nla24 gene (Figure 2).

Genes identified previously that are essential for S motility also include tgl (MXAN3084), which encodes an outer membrane lipoprotein (Rodriguez-Soto and Kaiser 1997a,b) that can be transferred among cells upon contact (Nudleman et al. 2005); sglK (MXAN6671), which encodes a DnaK homolog essential for S motility (Weimer et al. 1998); frzS (MXAN4149), which encodes a cytoplasmic protein with a coiled-coil domain that can assemble into myosin-like lattices (Ward et al. 2000); nla1 (MXAN5853) and nla19 (MXAN1078), which encode two different response regulators (Caberoy et al. 2003); masK (MXAN1922), which encodes an essential tyrosine kinase (Thomasson et al. 2002); and three genes in the dif operon (MXAN6696–MXAN6691) encoding a methyl-accepting chemotaxis protein (difA, MXAN6696), a histidine kinase (difE, MXAN6692), and a CheW homolog (difC, MXAN6694) (Yang et al. 2000b; Bellenger et al. 2002). We recovered magellan-4 insertions in tgl, frzS, sglK, and nla23, but not in masK, nla1, nla19, difA, difC, or difE (Table 1).

Recently, Pham et al. (2005) showed that integrative disruption of the MXAN4150 (sgmO) gene, which they named rasA, results in defects in fibril production, S motility, and multicellular development. Our result that a magellan-4 insertion in this gene impairs S motility confirms their results independently. In addition, Kimura et al. (2004) found that the simultaneous disruption of the cotranscribed rppA (MXAN5907) and mmrA (MXAN5906) genes also results in an S motility defect. Therefore, we might predict that (polar) insertions of magellan-4 in the rppA gene would also disrupt S motility; however, we did not recover insertions in this possible target.

Insertions of magellan-4 in genes previously not known to be required for S motility:

Among the 132 magellan-4 insertions that we characterized, seven are from mutants that form colonies with distinct morphologies or display only partial S motility defects. One of these mutants, MxH1777 mor-115, forms bright orange colonies with a motility phenotype indistinguishable from that of the MxH1777 parent. Sequence analysis of this mutant shows that its magellan-4 insertion lies in the very 3′ end of the carR gene, which encodes a regulator of carotenoid biosynthetic gene expression, consistent with the observation that it overproduces the bright orange carotenoids made by M. xanthus (Hodgson 1993; Martinez-Laborda et al. 1990). A second morphology mutant forms colonies with a wrinkled appearance that are smaller than, and more than twice the height of, those formed by its parent. The insertion in this mutant, mor-105, lies within the stk (MXAN3474) gene, which encodes a DnaK homolog involved in cell-cell adhesion. Loss-of-function mutations in the stk gene suppress a subset of S motility defects (Dana and Shimkets 1993).

Five mutants were found to form colonies with sizes intermediate between those formed by their A⁻ parent and its A⁻S⁻ mutant derivatives. One of these carries an insertion in MXAN1641, predicted to encode a protein of unknown function. A second has an insertion in MXAN3840, predicted to encode a homolog of the ScpB protein, which regulates chromosome segregation in Gram-positive bacilli (Soppa et al. 2002). A third has an insertion in the M. xanthus homolog of the mutS gene (MXAN5134), required for mismatch repair. It is possible that defects in these three genes have pleiotropic effects on colony size and affect S motility indirectly by impairing central cell processes. In addition, we recovered two insertions in the cglB gene (MXAN3060), known to be involved in A motility, but not in S motility (Rodriguez and Spormann 1999), that confer partial motility defects. These results confirm that the ΔaglU cglB double mutants have a more severe defect than their ΔaglU parent (White and Hartzell 2000), consistent with our recent results that several other alleles in A motility genes, when paired with the ΔaglU mutation, result in more severe A motility defects (our unpublished results).

The remaining 41 insertions of magellan-4 confer profound defects in S motility; these likely affect the expression of genes whose products are directly involved in S motility. These insertions inactivate genes encoding proteins involved in central metabolic process, in both cytoplasmic and membrane proteins that may be components of the pilus motor, in regulators of transcription, and in polysaccharide metabolism. Three magellan-4 insertions inactivate genes MXAN0358, MXAN3759, and MXAN3797, predicted to encode central metabolic functions. The first of these genes encodes one of two isoleucine tRNA synthetases found on the M. xanthus genome, demonstrating that this ileS gene is not essential for growth. The effect of the magellan-4 insertion in this gene on S motility could be direct, but it is quite possible that this insertion affects the expression of the adjacent MXAN0359–MXAN0362 genes. These four genes are predicted to encode homologs of the Pseudomonas aeruginosa genes fimT, pilW, pilV, and pilY1, which contribute to the stability of type IV pili (Alm and Mattick 1995, 1997; Alm et al. 1996; Carbonnelle et al. 2005). The second of these is in a gene, MXAN3759, predicted to encode one of two propionyl-CoA carboxylases made by M. xanthus (Kimura et al. 1997). The third of these is in a gene, MXAN3797, predicted to encode a short-chain acyl-CoA dehydrogenase.

Fourteen magellan-4 insertions define 13 new genes predicted to encode proteins of unknown function (see Table 2). These include MXAN0440, predicted to encode an integral membrane exporter, and MXAN1795 and MXAN7360, predicted to encode lipoproteins; the insertion that affects MXAN7360 lies upstream of its potential start codons and likely inactivates its promoter. Three genes, MXAN1106, MXAN4639, and MXAN5766, are predicted to encode proteins with TPR motifs; the first among these is a protein that also has a DnaJ motif adjacent to its C-terminal TPR motif; the latter two have pairs of adjacent TPR motifs. These TPR repeat domain pairs compose a larger domain similar to that of the Pseudomonas aeruginosa PilF protein, which is required for the assembly of type IV pili (fimbrae) (Watson et al. 1996), a domain shared by the Tgl lipoprotein (Rodriguez-Soto and Kaiser 1997a,b). Four magellan-4 insertions define four new genes predicted to encode transcription regulators; three of these have response-regulator domains.

TABLE 2.

magellan-4 insertions identify 31 new genes required for S motility

Gene	Name	Putative function
Central metabolic functions
MXAN0358 (1)	sgmA; ileS	Isoleucyl tRNA synthetase; downstream of fimT, pilW
MXAN3759 (1)	sgmM; pccB2	Propionyl-CoA carboxylase
MXAN3797 (1)	sgmN	Short-chain acyl-CoA dehydrogenase
Cytoplasmic and membrane proteins of unknown function
MXAN0440 (1)	sgmB	Inner membrane protein; RND exporter
MXAN1106 (2)	sgmC	DnaJ/TPR repeat domain protein
MXAN1641 (1)a	sgmD	Hypothetical protein
MXAN1795 (1)	sgmE	Lipoprotein
MXAN2526 (1)	sgmH	Hypothetical protein
MXAN4150 (1)	sgmO	Conserved hypothetical protein; downstream of frzS
MXAN4620 (1)	sgmR	Hypothetical protein; downstream of rfbC; methyltransferase
MXAN4639 (1)	sgmS	TPR repeat protein; PilF domain
MXAN5766 (1)	sgmX	TPR repeat protein; PilF domain
MXAN5770 (1)	sgmY	Conserved hypothetical protein
MXAN6125 (1)	sgnA	Hypothetical protein; helicase?
MXAN6518 (1)	sgnB	Conserved hypothetical protein; ABC transporter
MXAN7360 (1)	sgnE	Lipoprotein
MXAN7442 (2)	sgnF	Membrane protein
Regulators of transcription
MXAN2128 (1)	sgmF	LysR family activator
MXAN4640 (1)	sgmT	Histidine kinase/response regulator
MXAN5592 (1)	sgmW	Response regulator
MXAN6627 (1)	sgnC	Response regulator
Polysaccharide metabolism
MXAN2203 (1)	sgmG	Asp box motif protein; glycosyl hydrolase?
MXAN2561 (1)	sgmI	Fibronectin type III domain protein; hydrolase?
MXAN2921 (1)	sgmJ; wbaZ	Glycosyl transferase group I domain
MXAN2922 (2)	sgmK	LPS biosynthesis; WcaJ C-terminal domain
MXAN3506 (4)	sgmL	Polysaccharide biosynthesis
MXAN4613 (1)	sgmP; rfbB2	dTDP-glucose 4,6-dehydrogenase
MXAN4616 (3)	sgmQ	Glycosyl transferase group I and II domains
MXAN4707 (1)	sgmU; rfaF	Heptosyl transferase 9 domain
MXAN5333 (1)	sgmV; rfaG	Glycosyl transferase group I domains (2)
MXAN5831 (1)	sgmZ; glgP	Glycogen phosphorylase
MXAN6908 (3)	sgnD; pgi	Glucose-6-phosphate isomerase

Genes are divided into four functional categories as shown and are numbered in accord with the M. xanthus genome sequence available at http://cmr.tigr.org/tigr-scripts/CMR/GenomePage.cgi?org=gmx. After each gene number, in parentheses is shown the number of magellan-4 insertions characterized in each gene; new genes involved in S motility were designated sgmA through sgnF; alternative suggestions for gene names are also listed. The predicted function of each gene is listed.

^aWe have included MXAN1641 as an S gene although its direct role in motility has yet to be determined (see results).

Finally, the magellan-4 insertions define 12 new genes predicted to encode proteins involved in polysaccharide metabolism, 9 of which are likely critical for EPS biosynthesis, LPS biosynthesis, or both. Three magellan-4 insertions affect genes likely to encode products involved in polysaccharide catabolism; MXAN2203 is predicted to encode a glycosyl hydrolase with an Asp box motif (β hairpin), suggesting that its activity may be regulated by a complex phosphorelay system (Zhang and Shi 2005). The product of MXAN2561 has a fibronectin type III domain, found among several bacterial polysaccharide hydrolases (Folders et al. 2001; Kataeva et al. 2002). The glgP gene (MXAN5831) encodes glycogen phosphorylase.

DISCUSSION

Among 128 independent insertions of the defective transposon magellan-4 in genes required for S motility in M. xanthus, two-thirds lie within 27 of the 37 nonessential genes whose functions have been shown to be required for S motility. These include genes involved in the biogenesis of type IV pili, EPS and LPS. Our results show that the mechanism of S motility is more complex genetically than that of flagellar motility, which requires the products of <60 genes in Salmonella and E. coli (Chilcott and Hughes 2000). If we estimate that we have identified ∼75% of the nonessential genes involved in S motility by our mutant hunt, then we predict that there should be at least 80 nonessential genes whose products are required for S motility.

This number likely is an underestimate. Although the spectrum of magellan-4 insertions in S genes is very broad, it is not random and likely shows both site and regional preferences in M. xanthus, as does any other transposon and as is suggested by the results shown in Figures 1–3. Another indication that we have yet to reach saturation of the S motility genes with insertions of magellan-4 is that previously we found 12/115 nonmotile derivatives of an S mutant with insertions of magellan-4 in the mglBA genes (among 36 target genes) (Youderian et al. 2003), but only 1 of the 128 insertions in S motility genes described in this report falls within the mglBA operon (among 58 target genes). We note that some of the genes we have identified with insertions of magellan-4 may not be involved directly in S motility, due to polar and other effects that insertions may have on the expression of neighboring genes. This is certainly the case for insertions of magellan-4 in the mglB gene, which are polar in the expression of mglA (Youderian et al. 2003), and may also be the case for insertions in MXAN4620, upstream of wbgB (Figure 3, top), and in MXAN0358 (ileS), adjacent to homologs of the P. aeruginosa fimT, pilW, pilV, and pilY1 genes. These homologs may represent a new pil gene cluster.

Two of the new genes we have identified as social motility genes are likely involved in fatty acid metabolism. One of the magellan-4 insertions identifies a second M. xanthus gene, MXAN3759, predicted to encode a propionyl-CoA acyl carboxylase. A defect in its paralogue, pccB (MXAN1113), results in a reduction in the relative amount of long-chain fatty acids present in the membranes of vegetative and developing cells, but not in their abolition. Thus, although a pccB mutant can aggregate, it cannot form heat-resistant spores in response to carbon starvation (Kimura et al. 1997), a phenotype that can be masked partially by the addition of methylmalonyl-CoA to developing cells. The fact that a second propionyl-CoA acyl carboxylase isozyme also is required for social motility underscores the importance of long-chain fatty acids in the integrity of the M. xanthus membrane. The gene MXAN3797 is predicted to encode one of the >20 short-chain acyl-CoA dehydrogenases (SCADs) made by M. xanthus. Whether fatty acid metabolism plays a direct role in S motility or whether the magellan-4 insertions in MXAN3759 and MXAN3797 result in pleiotropic effects on membrane biogenesis remains to be determined.

magellan-4 insertions that impair S motility identify four new genes predicted to encode regulators of transcription. Three of these genes are predicted to encode products with response regulator domains. Among these, the product of one (MXAN4640) also is predicted to have a histidine kinase domain. This gene may be cotranscribed with MXAN4639, which also is required for S motility (Figure 3, bottom). The M. xanthus genome includes genes for 150 different proteins with response regulator domains. An earlier mutant hunt using integrative disruption to probe the functions of 27 of these genes showed that four are required for S motility (Caberoy et al. 2003). That we obtained magellan-4 insertions in only 2 of these 4 genes among the 27 again demonstrates that our magellan-4 insertions have not saturated the genes required for S motility. The fact that we obtained magellan-4 insertions in 3 additional genes among the 123 predicted to encode additional response regulators and that have yet to be tested for function shows that at least seven different response regulators are required for S motility. The systematic functional analysis of the 120 remaining genes predicted to encode response regulators may reveal additional genes required for S motility. We also obtained an insertion in gene MXAN2128, which encodes 1 of the 24 proteins in the LysR family of transcription regulators predicted to be encoded by genes present on the M. xanthus genome.

Among the genes of unknown function identified as social motility genes by magellan-4 insertions are three new genes predicted to encode proteins with TPR repeat motifs. We suspect that these proteins will play central roles in the mechanism of S motility for several reasons. The M. xanthus genome has >60 genes predicted to encode proteins with TPR repeat domains, at least three of which, agmK (MXAN4863), aglT (MXAN4869), and agmU (MXAN4870), are required for A motility (Youderian et al. 2003). The Tgl lipoprotein, which is transferred from cell to cell upon contact (Nudleman et al. 2005), also has a TPR repeat domain. Such domains are critical for the function of kinesin-dependent eukaryotic motors (Gindhart and Goldstein 1996). One or more of these proteins may play a structural role in the mechanism of pilus movement in M. xanthus.

During the mechanics of pilus extension and retraction, ATP is predicted to be hydrolyzed to energize the pilus motor, as is the case for the variety of motors that drive the mechanisms of eukaryotic motility involving the cycle of macromolecular sliding and/or polymerization/depolymerization. The analysis of S mutants suggests that the proteins involved in ATP hydrolysis may include not only PilT and PilB, the ATPases thought to be involved in pilus extrusion, but also the homologs of DnaJ and DnaK involved in S motility. Two of the M. xanthus genes predicted to encode >12 homologs of HSP70 (DnaK), sglK (Weimer et al. 1998; Yang et al. 1998), and stk (Dana and Shimkets 1993), play central roles in S motility. Also, one of the genes we have identified by magellan-4 insertions (MXAN1106) encodes a TPR repeat protein with a DnaJ domain.

Many of the new S motility genes that we have identified are involved in LPS and/or EPS polysaccharide biosynthesis. Three of these, MXAN4613, MXAN4616, and MXAN4620, are clustered near three genes previously shown to be involved in S motility (Figure 3, top); additional genes within this cluster are predicted to be involved in LPS production and may be required for S motility. The two adjacent genes MXAN2921 and MXAN2922 lie within an additional cluster of genes likely required for LPS biosynthesis and S motility (Figure 3, bottom). The role that LPS plays in S motility could be direct or indirect, because many bacteria regulate the balance between LPS and EPS biogenesis, and defects in LPS biosynthesis may cause dramatic changes in the structure or amount of EPS produced by cells. LPS mutants of M. xanthus produce more EPS antigen than do wild-type cells (Bowden and Kaplan 1998).

It is clear that EPS plays a direct role in the mechanism of S motility. Mutants defective in EPS production are hyperpiliated, a phenotype that can be rescued by the addition of EPS or chitin, a polymer of N-acetylglucosamine, but not by cellulose, a polymer of glucose (Li et al 2003). N-acetylglucosamine is a major component of M. xanthus EPS. The addition of chitin to wild-type M. xanthus cells results in pilus retraction. In contrast, the addition of N-acetylglucosamine to wild-type M. xanthus cells has the opposite effect and results in hyperpiliation. Glucose and galactose have no effect on pilus extension (Li et al. 2003). Our results suggest that the signals that govern pilus extension and retraction are the products of polysaccharide catabolism. The magellan-4 insertions identified three genes potentially involved in polysaccharide catabolism that are essential for S motility. These three genes include MXAN2561, predicted to encode a protein with a fibronectin type III domain, implicated in the hydrolysis of polysaccharides; MXAN2203, predicted to encode a glycosyl hydrolase; and, most important, the glgP gene (MXAN5831), predicted to encode glycogen phosphorylase.

Although social motility appears to be a mechanism conserved among diverse Gram-negative bacteria, very few genetic determinants of this mechanism are conserved in sequence among the bacteria that display social (or “twitching”) motility, M. xanthus and P. aeruginosa (Peabody et al. 2003). The gene encoding glycogen phosphorylase (EC 2.4.1.1), which cleaves polysaccharides with α(1 → 4) glucan linkages, is conserved among M. xanthus, P. aeruginosa, and Haemophilus influenzae, which has been shown recently to display S motility (Bakaletz et al. 2005). Glycogen phosphorylase is an economical enzyme that uses the cofactor PLP to carry out the phosphorylation of monosaccharides concomitant with their hydrolysis from polysaccharides, without consuming ATP to generate its monosaccharide phosphate products. We propose that the signal for the retraction of the M. xanthus pilus is likely N-acetylglucosamine phosphate, the predicted product of glycogen phosphorylase acting on chitin as a substrate. N-acetylglucosamine phosphate is an ideal signal for pilus retraction, because it can be cleaved to form N-acetylglucosamine by the phosphatases made by M. xanthus (Weinberg and Zusman 1990). Indeed, a mutation in the pph1 gene, encoding a Ser/Thr protein phosphatase, affects vegetative swarming on 0.3% agar but not on 1.5% agar (Treuner-Lange et al. 2001), a phenotype indicative of a defect in S motility. Currently, we are testing the hypothesis that glycogen phosphorylase generates N-acetylglucosamine phosphate from EPS as a signal for pilus retraction.