Bacterial Gliding Motility: Multiple Mechanisms for Cell Movement over Surfaces.

Key Words twitching, type IV pilus, Myxococcus, Flavobacterium, Mycoplasma

* Abstract The mechanisms responsible for bacterial gliding motility have been a mystery for almost 200 years. Gliding bacteria move actively over surfaces by a process that does not involve flagella. Gliding bacteria are phylogenetically diverse and are abundant in many environments. Recent results indicate that more than one mechanism is needed to explain all forms of bacterial gliding motility. Myxococcus xanthus "social gliding motility" and Synechocystis gliding are similar to bacterial "twitching motility" and rely on type IV pilus extension and retraction for cell movement. In contrast, gliding of filamentous cyanobacteria, mycoplasmas, members of the Cytophaga-Flavobacterium group, and "adventurous gliding" of M. xanthus do not appear to involve pili. The mechanisms of movement employed by these bacteria are still a matter of speculation. Genetic, biochemical, ultrastructural, and behavioral studies are providing insight into the machineries employed by these diverse bacteria that enable them to glid e over surfaces.

INTRODUCTION

Surfaces are important features of many environments. Nutrients are concentrated at surfaces, making them attractive sites for bacterial colonization, often resulting in the formation of complex biofilms (26,76). Bacteria have evolved efficient strategies to move over surfaces. Proteus mirabilis, Vibrio parahaemolyticus, Serratia marcescens, and many others employ numerous flagella to spread over moist surfaces in a process known as swarming motility (46). Other bacteria, such as Pseudomonas aeruginosa and Neisseria gonorrhoeae, use type IV pill to move in a process called twitching motility (49). Finally, diverse bacteria such as Myxococcus xanthus, Flavobacterium johnsoniae, Phormidium uncinatum, and many others slither over surfaces by a mysterious process known as gliding motility (54,77,103).

Bacterial gliding motility is defined as smooth translocation of cells over a surface by an active process that requires the expenditure of energy. Gliding does not require flagella, and cell movement generally follows the long axis of the cell. As a result of their movements, gliding bacteria often produce colonies that have thin spreading edges (Figure 1A).

Bacteria from many branches of the eubacterial phylogenetic tree exhibit gliding motility (Table 1). It is particularly common in three large groups: the myxobacteria (members of the [delta] proteobacteria), the cyanobacteria, and the Cytophaga-Flavobacterium group. Gliding bacteria live in environments as diverse as the human mouth, ocean sediments, and garden soil. Although they are abundant in many natural environments, they are less common as objects of experimental study in the laboratory. For this reason they are often mistakenly considered to be rare and exotic organisms.

Over the years, many different models have been proposed to explain bacterial gliding motility. Recent results suggest that it is unlikely that any single mechanism will be able to explain all forms of bacterial gliding. Instead, it appears that there are several different types of gliding motility "motors." Recent evidence suggests that type IV pili are required for some forms of gliding motility (11, 115). Because type IV pili are also required for bacterial twitching motility, we begin with a discussion of this phenomenon.

TWITCHING MOTILITY: TYPE IV PILL AND CELL MOVEMENT

Twitching motility occurs in a wide variety of bacteria, including P. aeruginosa, N. gonorrhoeae, and even some strains of Escherichia coli (49, 115). Movement of cells occurs in short, intermittent jerks of up to several micrometers. Twitching motility usually requires a very moist surface, and cells typically move only when they are within several micrometers of each other. The possible involvement of polar pil was suggested by Henrichsen (49), who observed a correlation between the presence of polar pili and ability of cells to exhibit twitching motility.

In 1980 Bradley proposed that retraction of polar pili was the driving force for P. aeruginosa twitching motility (17). This hypothesis was based on observations regarding bacteriophage infection of cells (15, 16). Wild-type cells were susceptible to infection, whereas mutants that lacked pili, or mutants that were hyperpiliated, were resistant to infection. Examination of wild-type cells that were exposed to bacteriophage revealed phage particles attached to the cell wall near the poles. Cells of nonpiliated mutants generally had no phage attached, whereas cells of hyperpiliated mutants had phage attached to their pili, but rarely to the cell wall. Pilus-specific antisera inhibited twitching motility and protected cells from bacteriophage infection (15, 17). To explain these observations Bradley suggested that wild-type pili were retractile and that phage that attached to these pili were brought in contact with the cell surface by pilus retraction. He also proposed that pilus retraction was responsible for c ell movement, because nonpiliated and hyperpiliated mutants were incapable of twitching motility (17).

Bradley's ideas languished in relative obscurity for more than a decade but were eventually resurrected and confirmed by the results of a number of different laboratories. It is now clear that type IV pili are required for twitching motility (5,83). These pili, which are about 6 nm in diameter and up to 4 [micro]m in length, are typically found at one or both cell poles. Like other pili, they play an important role in attachment of cells to surfaces. However, they are not merely passive structures. Active extension and retraction of type IV pili appear to be involved in a wide variety of processes such as cell movement (83), transformation (123), conjugation (133), bacteriophage infection (16,63), biofilm formation (85), activation of host cell responses (81), and cytotoxicity (25).

Studies with P. aeruginosa have identified at least 35 genes that are involved in type IV pilus biogenesis and twitching motility (5,97). Five of these genes, pilS pilR, fimS, algR, and rpoN, are involved in transcriptional regulation of genes required for pilus biogenesis and function. At least 20 genes are involved specifically in pilus biogenesis. Of these, PilA codes for the primary structural subunit of the pilus, whereas pilB codes for a protein with a putative NTP binding domain that is thought to be required for pilus polymerization and extension. PilD is a peptidase/methylase that processes PilA and many of the other Pil proteins. The pilus extends through the outer membrane of the cell with the help of PilQ, which is thought to function as a secretin. Two additional genes, pilT and pilU, are not required for pilus biogenesis but are required for pilus-mediated cell movement. Cells with mutations in either of these genes are hyperpiliated, presumably because they produce pili but fail to retract them . PilT and PilU exhibit sequence similarity to each other and to PilB, and all have putative nucleotide binding domains. ATP hydrolysis by PilT and PilU may power pilus disassembly at the cell membrane, causing retraction (83, 121, 124). At least eight other genes that are similar in sequence to chemotaxis (che) genes of other bacteria are also required for normal twitching motility (27,28). Typically, chemotaxis proteins sense environmental stimuli and alter the behavior of the bacterial flagellum to control the directed movement of cells (13). The P. aeruginosa che homologs that affect twitching motility are thought to be involved in both production of pili and regulation of pilus function to result in directed movement. Mutations in some of these genes result in lack of pili, whereas mutations in others result in cells that exhibit twitching motility but migrate in an apparently uncontrolled fashion.

Cells of N. gonorrhoeae and Neisseria meningitidis also have type IV pili and exhibit twitching motility, moving over wet glass at up to 1 [micro]m/second (49, 83, 123). Pili are important virulence factors of these bacteria and are involved in attachment to host cells and possibly invasion (82). Many of the genes involved in pilus biogenesis and function are similar in sequence to those of P. aeruginosa described above (37, 113). In particular, homologs to pilA (encoding pilin), pilB (thought to be involved in pilus extension), and pilT (involved in pilus retraction) are present. Homologs to chemotaxis genes are notably lacking from the N. meningitidis genome (86, 110).

The pilus retraction model for twitching motility has been widely accepted, but until recently direct evidence linking pilus retraction to cell movement was lacking. Studies of N. gonorrhoeae twitching motility by Merz et al (83) provided this evidence. Wild-type cells held with laser tweezers near a microcolony were actively pulled toward the colony. Cells of a pilT mutant, which produced pili that were apparently nonretractile, were not pulled toward a pilT microcolony. Furthermore, latex spheres that were coated with monoclonal antibodies that recognized a surface epitope of pili were also pulled toward wild-type N. gonorrhoeae cells but not toward cells of a pilT mutant. Movement of cells or latex spheres occurred at about the same speed as twitching movement over wet surfaces (approximately 1 [micro]m/second). These results demonstrate that pilus retraction provides sufficient force to propel cells of N. gonorrhoeae, providing strong evidence that retraction of pili results in twitching motility. Pilus e xtension may also result in cell movement, but this has not been demonstrated.

MYXOBACTERIAL GLIDING MOTILITY: MULTIPLE SYSTEMS FOR CELL MOVEMENT

The myxobacteria are arguably the most complex of the prokaryotes (31). Myxobacterial cells travel in multicellular swarms that cooperatively digest macromolecular food sources such as prey cells of other bacteria. When nutrients are scarce, thousands of myxobacterial cells aggregate to form complex, multicellular fruiting bodies. Within the fruiting bodies, vegetative cells differentiate to form dormant myxospores. Myxobacteria also produce numerous secondary metabolites including antibiotics and antitumor drugs (91a). Given this complexity, it is not surprising that myxobacterial genomes are considerably larger than those of most prokaryotes. The genome of Myxococcus xanthus, the best studied of the myxobacteria, consists of approximately 9.5 megabase pairs (22).

The early experiments of Hodgkin and Kaiser suggested that M. xanthus has two independent systems for movement over surfaces, the "S" system, for "social gliding" and the "A" system, for "adventurous gliding" (52, 53). Wild-type colonies spread over an agar surface, with cells moving in multicellular peninsulas and rafts, and individual cells migrating "adventurously" beyond the confines of these "social" groups. Strains with mutations in S-motility genes ([A.sup.+][S.sup.-]) move well as individual cells but are less efficient at group movements, move poorly on very moist surfaces, and are usually defective in fruiting body formation (53,62,99). In contrast, cells with mutations in A-motility genes ([A.sup.-][S.sup.+] move well in groups but are unable to glide as individual cells on media solidified with 1.5% agar, unless they are within several micrometers of another cell (62, 103). These [A.sup.-][S.sup.+] mutants form spreading colonies with no individual cells beyond the edge of the colony. Wild-type ce lls, and cells of [A.sup.-][S.sup.+] or [A.sup.+][S.sup.-] mutants, move slowly over 1.5% agar surfaces, at rates of approximately 0.025 to 0.1 [micro]m per second (104). Disruption of both motility systems (AS) results in complete loss of motility (53). …