|Home | About | Journals | Submit | Contact Us | Français|
Type IV pili (TFP) are membrane-anchored filaments with a number of important biological functions. In the model organism Myxococcus xanthus, TFP act as molecular engines that power social (S) motility through cycles of extension and retraction. TFP filaments consist of several thousand copies of a protein called PilA or pilin. PilA contains an N-terminal α-helix essential for TFP assembly and a C-terminal globular domain important for its activity. The role of the PilA sequence and its structure–function relationship in TFP-dependent S motility remain active areas of research. In this study, we identified an M. xanthus PilA mutant carrying an alanine to valine substitution at position 32 in the α-helix, which produced structurally intact but retraction-defective TFP. Characterization of this mutant and additional single-residue variants at this position in PilA demonstrated the critical role of alanine 32 in PilA stability, TFP assembly and retraction.
Type IV pili (TFP) are multifunctional membrane-anchored filaments present in both pathogenic and non-pathogenic bacteria. An astonishing number of cellular functions have been attributed to the TFP, including adhesion, surface motility, biofilm formation, natural transformation, immune escape and cell signalling (Craig et al., 2004). TFP-mediated motility enables movement of bacterial cells across semi-solid surfaces such as mucosal epithelia and soft agar surfaces. This behaviour is called twitching motility in Pseudomonas aeruginosa and Neisseria gonorrhoeae, and social gliding motility in Myxococcus xanthus (Bradley, 1980; Kaiser, 1979).
M. xanthus is a Gram-negative soil bacterium able to glide on solid surfaces and coordinate thousands of cells into multicellular fruiting bodies during starvation. M. xanthus cells utilize two genetically distinct systems for the gliding motility to achieve their complex cell behaviours: adventurous (A) motility and social (S) motility (Hodgkin & Kaiser, 1979). Single-cell movement via A motility is the preferred type of locomotion on dry surfaces, while coordinated movement in groups via S motility is mainly utilized on moist surfaces (Shi & Zusman, 1993). S motility is known to be important for cell aggregation, cohesion and fruiting body morphogenesis (Hodgkin & Kaiser, 1979; MacNeil et al., 1994). TFP in M. xanthus act as molecular engines that power S motility (Hodgkin & Kaiser, 1979; Wu & Kaiser, 1995). Pili are extended at one cell pole, attach to the extracellular polysaccharides (EPS) on the surface of the substratum or another cell, and retract to pull the cell forward (Li et al., 2003). TFP are composed of several thousand subunits of PilA or pilin. The cycles of TFP extension and retraction are achieved by assembling PilA into polar filaments with the assistance of the ATPase PilB (Jakovljevic et al., 2008), followed by disassembling the pili into single subunits during retraction, a process that is mediated by the ATPase PilT (Jakovljevic et al., 2008). PilA monomers form a membrane-associated pool which can be recycled to form new TFP (Yang et al., 2010).
Although the structure of M. xanthus PilA has not been experimentally elucidated, the crystal structures of closely related pilins in P. aeruginosa (PilA) and N. gonorrhoeae (PilE) have been solved either in their full-length or in their N-terminal-truncated forms (Craig et al., 2003, 2006; Hazes et al., 2000; Keizer et al., 2001; Parge et al., 1995). These studies have revealed a remarkably conserved structure for full-length pilin, which consists of an N-terminal α-helix domain followed by a C-terminal globular domain. Consistent with data derived from fibre diffraction and electron microscopy (Craig et al., 2006), a TFP assembly model has been proposed based on these crystal structures, in which pilins are suggested to assemble into the pilus in a helical manner. The N-terminal α-helix forms the hydrophobic core, and the C-terminal globular domain forms both lateral and longitudinal contacts at the outer surface (Craig et al., 2004). The N-terminal α-helix is also believed to be essential for the membrane trafficking of PilA (Aas et al., 2007; Yang et al., 2010). This N-terminal domain can be further separated into a highly conserved hydrophobic α1-N subdomain (residues 13–40, counting from the first amino acid in the signal peptide) and a less conserved amphipathic α1-C subdomain (residues 41–65) (Craig et al., 2004). α1-N is predicted to mediate both membrane localization and crucial contacts between the subunits for TFP assembly. The C terminus of PilA has been shown to mediate cell attachment, and binding to surfaces and extracellular sugars (Craig et al., 2006; Giltner et al., 2006; Lee et al., 1994; Li et al., 2003; Scheuerpflug et al., 1999; Sheth et al., 1995).
The high degree of conservation observed among the PilA N-terminal α-helix domains from different bacteria (Li et al., 2005) suggests that M. xanthus pilin may assemble into TFP in a manner similar to that predicted for P. aeruginosa and N. gonorrhoeae. Previously, we have shown that specific pilA sequence changes in M. xanthus can affect TFP assembly and PilA localization, which in turn alter EPS production and fruiting body formation (Yang et al., 2010). Additional roles of the PilA sequence and its structure–function relationship in TFP-dependent S motility still remain to be elucidated. In this study, we used a pilA mutant library (Yang et al., 2010) to isolate a mutant that produced TFP but failed to swarm on 0.3% agar (Shi & Zusman, 1993). We subsequently mapped the mutation to A32 and generated additional single-residue mutant variants of PilA with alterations at this position for functional characterization. This allowed us to establish the critical role of amino acid A32 in PilA stability and TFP-related functions.
M. xanthus strains and plasmids used in this study are listed in Table 1. M. xanthus cells were grown in CYE medium (1%, w/v, casitone, 0.5%, w/v, yeast extract, 8 mM MgSO4 in 10 mM MOPS buffer, pH 7.6) (Campos et al., 1978) at 32 °C on a rotary shaker at 300 r.p.m. and were maintained on 1.5%, w/v, agar plates. The Escherichia coli strain DH5α used for cloning and plasmid construction was grown at 37 °C in Luria–Bertani medium (Sambrook & Russell, 2001). Medium and plates were supplemented with kanamycin at 100 µg ml−1 when needed.
Site-directed mutagenesis of pilA was conducted using two-step overlap PCR as previously described (Yang et al., 2010) and adapted from the standard protocol (Sambrook & Russell, 2001). The mutant pilA gene was cloned into the EcoRI and BamHI sites of pSWU19, and the resulting plasmids were electroporated (Kashefi & Hartzell, 1995) into DK10410 (ΔpilA) (Wu & Kaiser, 1996) and SW2017 (ΔpilA ΔdifA) to select for kanamycin resistance. All plasmids and mutants constructed in this study were confirmed by PCR and sequencing (data not shown). The sequences of the primers used for mutant construction are shown in Table 2.
The swarming assay was performed as previously described (Shi & Zusman, 1993). Briefly, M. xanthus cells were grown in CYE medium to the exponential growth phase and concentrated to OD600 10 (5×109 cells ml−1) in CYE medium. An aliquot of 5 µl concentrated culture was spotted onto 0.3% or 1.5% CYE agar plates and incubated for 3 days at 32 °C. S motility was analysed by observing the colonies for expansion on 0.3% CYE agar plates.
EPS production was visualized using a calcofluor white binding assay as described by Black & Yang (2004). Briefly, exponential growth phase cells were concentrated to OD600 10 in MOPS buffer (10 mM MOPS, 8 mM MgSO4, pH 7.6). An aliquot of 5 µl concentrated culture was spotted onto 0.3% CYE agar plates containing calcofluor white (CF agar) at 50 µg ml−1. The plates were incubated at 32 °C in the dark for 3 days before they were examined and photographed under the illumination of a long-wavelength (365 nm) UV light source.
M. xanthus cells were grown in CYE to exponential growth phase and concentrated to OD600 10 in Tris–phosphate–magnesium (TPM) buffer (10 mM Tris/HCl, pH 7.6, 1 mM KH2PO4, 8 mM MgSO4). Aliquots (5 µl) of concentrated cells were spotted onto CF agar (Hagen et al., 1978) and incubated for 3 days at 32 °C. Pictures of fruiting bodies were taken using a Nikon Eclipse TE200 inverted microscope fitted with a SPOT camera/software (Diagnostic Instruments).
Western blot analysis was performed as described previously (Yang et al., 2010) following standard protocols (Harlow & Lane, 1988). For whole-cell lysates, 108 M. xanthus cells were lysed by boiling in SDS-PAGE loading buffer for 10 min. For surface pili detection, TFP were sheared off from 1010 M. xanthus cells by vigorous vortexing for 20 min. The suspension was sedimented at 16000 g for 5 min and the supernatant was transferred to a clean tube. Pili/pilin in the solution were precipitated by adding TCA to a final concentration of 10%, incubating on ice overnight, and sedimenting at 16000 g for 30min at 4 °C. Pellets were washed once with acetone, resuspended in SDS-PAGE loading buffer and boiled for 10 min. Primary anti-PilA antibody (Li et al., 2005) was used at a 1:10000 dilution and anti-rabbit horseradish-peroxidase-conjugated secondary antibody (Pierce) was used at a 1:20000 dilution. Blots were developed using the Supersignal West Pico chemiluminescence reagent (Pierce). Images were taken with the ChemiDoc XRS system (Bio-Rad).
Pilin precipitation and mixing assays were performed as previously described (Li et al., 2003). For pilin precipitation, cell surface pili were sheared off from 1010 cells by vortexing as described above. The isolated pili were incubated with chitin suspension (final concentration 0.5 mg ml−1) or purified EPS (0.5 mg ml−1 carbohydrate) at 32 °C for 30 min. The mixtures were pelleted by centrifugation at 6000 g for 5 min for chitin and 10000 g for 10 min for EPS. Supernatants were discarded; pellets were resuspended in 80 µl SDS-PAGE loading buffer and boiled for 10 min before analysis by Western blotting. For the mixing/retraction assay, EPS and chitin suspensions were mixed with 5×109 cells and incubated at 32 °C for 30 min. Remaining cell surface pili were analysed by Western blotting.
To test for tethering of M. xanthus to solid surfaces via TFP, video microscopy assays were conducted as described previously (Sun et al., 2000). M. xanthus cells at the mid-exponential growth phase were serial diluted in MOPS buffer to obtain concentrations of about 50 isolated cells in the microscope field of view. Aliquots (10 µl) of each sample were deposited into each well of a 24-well cell culture plate containing 1%, w/v, methylcellulose solution in MOPS buffer. Plates were incubated at 32 °C for 30 min or longer to allow cells to settle on the bottom of the well. Gliding motility and tethering behaviour were observed using an inverted microscope (Nikon) with a ×40 objective lens. Serial digital images were taken at 5 s intervals using a Spot camera (Diagnostic Instruments).
Exponentially growing M. xanthus cells were washed once with MOPS buffer, stained using 2.5% uranyl acetate and examined with a CM120 transmission electron microscope (FEI) operated at 120 kV with a LaB6 filament. Images were recorded with a Tietz F224HD charge-coupled device (CCD) camera with the EMMENU4 digital camera software. Images were taken at ×13000 magnification.
By screening a previously generated M. xanthus pilA mutant library (Yang et al., 2010), one mutant was identified that was deficient in S motility, despite its ability to produce surface pili. This mutant, which was designated A32V (Fig. 1a), exhibited impaired S motility when its swarming zones were compared with those of the wild-type strain on 0.3% agar (Fig. 1b). However, in contrast to the smooth edge produced by ΔpilA, the colony edge of mutant A32V was rough, indicating residual S motility. As visualized by the calcofluor white binding assay (Black & Yang, 2004), A32V was found to produce detectable amounts of EPS and retain the ability to form fruiting bodies (Fig. 1b). Since surface pili act as a positive signal for EPS production (Black et al., 2006) and A32V produces more EPS than ΔpilA (Fig. 1b), we speculated that this mutant may still possess surface pili. Western blot analysis revealed that whole-cell PilA expression levels in A32V were similar to those in the wild-type, while the amount of surface pili appeared to be elevated in the mutant strain (Fig. 1c). Electron microscopy confirmed the existence of TFP on the A32V cell surface (Fig. 1d). Based on these results, we hypothesize that the PilA in A32V can assemble into surface pili but are unable to retract. Comparison of A32V with a ΔpilT mutant strain that is defective in the M. xanthus ATPase responsible for TFP retraction (Nudleman & Kaiser, 2004; Wu et al., 1997) showed that both strains were similarly impaired in S motility, had rough colony edges/surfaces and produced EPS (Fig. 1b). However, in ΔpilT, the amount of whole-cell PilA was significantly higher than in both the wild-type and A32V, and more pili were present on the cell surface (Fig. 1c).
The apparent overpiliation and S motility-deficient phenotypes in A32V suggested that the mutated PilA is capable of assembling into TFP that are unable to retract. This could be due to two different mechanisms. First, the mutagenized TFP in A32V may have lost the ability to bind EPS and is therefore unable to recognize the retraction trigger (Li et al., 2003); second, the TFP in A32V still bind to their target EPS but are unable to retract. To distinguish between these two possibilities, PilA-A32V was introduced into a ΔpilA ΔdifA double mutation background. In the ΔdifA background, all assembled pili remain on the cell surface because the cells are defective in EPS production and therefore lack the trigger for retraction (Yang et al., 1998, 2000). The addition of purified EPS from wild-type cells or the EPS analogue chitin was found to trigger TFP retraction and abolish the overpiliation phenotype (Li et al., 2003). Using the ΔdifA mutant as a control, the pilA-A32V ΔdifA mutant allowed us to verify TFP assembly in A32V and to test the binding and retracting ability of the mutated TFP. Pili were purified from the cell surface by shearing and were analysed using Western blots. pilA-A32V ΔdifA produced amounts of surface pili similar to ΔdifA (Fig. 2a, lane 1), demonstrating that the mutated A32V pilin can assemble into TFP in the ΔdifA background.
To test the binding ability of the TFP, sheared pili were incubated with either chitin (Fig. 2a, lane 2) or EPS isolated from the wild-type strain DK1622 (Fig. 2a, lane 3) and subjected to the precipitation assay. Similar to wild-type pili, the mutated A32V pili were precipitated by EPS and chitin at the same levels, indicating that the mutated TFP still have binding ability. To test if the mutated TFP can retract, cells were subjected to the mixing assay to trigger TFP retraction using either chitin (Fig. 2a, lane 5) or EPS (Fig. 2a, lane 6). While ΔdifA retracted its pili after incubation with both chitin and EPS, the pilA-A32V ΔdifA mutant showed little reduction in recoverable surface TFP, indicating that retraction of TFP was not triggered by the presence of EPS or chitin. Therefore, the mutated A32V pilin/pili still recognized its/their target polysaccharides although the mutated TFP were unable to retract.
To further confirm our findings, a video microscopy-based assay was used to provide direct visual evidence that the mutant A32V pili still have binding ability but are unable to retract. Sun et al. (2000) developed an assay capable of measuring S motility at the single-cell level. When deposited in 1% methylcellulose in MOPS buffer on polystyrene surfaces, wild-type M. xanthus cells glide over the surface and occasionally tether to the surface with their TFP, resulting in the detection of cells that ‘stand up’ (Fig. 2b). Cells lacking TFP (ΔpilA) were non-motile in this assay and unable to ‘stand up’. We analysed 500 A32V mutant cells under these experimental conditions and did not find motile cells; however, we did observe tethering behaviour (Fig. 2b). The phenotype of A32V in this assay is similar to that of the TFP retraction-deficient mutant pilT (Sun et al., 2000). These results further confirmed that A32V produces TFP, which allows the cells to tether, although the pili are unable to retract. As a consequence, S motility is impaired in the mutant strain.
The only mutation in A32V is the alanine to valine mutation at position 32 in the α-helix of PilA, indicating the importance of this amino acid for PilA function. To further investigate the role of A32, we conducted site-specific mutagenesis of pilA to obtain a series of single-residue variants at the same position in PilA. The original alanine was changed to amino acids with different R groups (side chains) and different chemical properties (Mathews et al., 2000), resulting in A32G, A32V, A32L, A32S, A32P, A32K, A32D and A32E. All mutants were tested for S motility, EPS production, fruiting body formation, PilA production and surface TFP assembly. Changing alanine into amino acids within the same R group (A32G, A32V, A32L) or into another structurally similar and uncharged residue (A32S) resulted in PilA production at wild-type levels (Fig. 3). Replacement of alanine with negatively or positively charged amino acids (A32K, A32D, A32E) or proline (A32P) led to a dramatic reduction in PilA levels (Fig. 3). For the mutants with little or no PilA expression, no surface pili were detected and little EPS were observed in the calcofluor white binding assay. Of the mutants that still produced wild-type levels of PilA, the surface pili of A32S were similar to those of the wild-type. Interestingly, however, the mutants A32G, A32V and A32L exhibited progressive phenotypes corresponding to the increase in side-chain length of the R group. Glycine (G) is one methyl group shorter than alanine, and the phenotype of A32G was similar to that of the wild-type. For this mutant, the same amount of surface pili was detected, and the TFP were fully functional, as indicated by its ability to swarm on 0.3% agar. In the mutant PilA of A32V, alanine at position 32 is replaced by valine (V), which carries a side chain that is two methyl groups longer than the original alanine. PiA-A32V was able to assemble into TFP, although the mutated TFP was not able to retract, as indicated by the lack of detectable swarming despite the presence of EPS (Fig. 3). Extension of the side chain to leucine in PilA-A32L resulted in a more dramatic phenotype. Although A32L produced whole-cell PilA at wild-type levels, no surface pili were detected. The structural changes in A32L PilA may impair the ability of the mutated pilin to assemble into TFP. Therefore, PilA monomers in A32L accumulated in the cell, likely in the membrane, leading to dramatic EPS reductions and a non-fruiting phenotype (Fig. 3) (Yang et al., 2010). These findings for A32 variants demonstrated that the uncharged nature of A32 is important for the stability of PilA, while side-chain length is important for the proper assembly of pilin subunits into functional pili.
The PilA sequence contains a proline at residue 34 that is close to A32 and is believed to be essential for the formation of a kink in the α1-N subdomain (Craig et al., 2006) (Fig. 1a). The curved structure created by this kink has been implicated as assisting in the tight packing of pilin subunits into TFP. It is possible that the phenotypes we observed in the A32 mutants were due to structural changes affecting this kink. In order to verify the importance of P34 in PilA, we replaced P34 with alanine. In the resulting mutant, P34A, both whole-cell PilA and surface pili were expressed at reduced levels compared with the wild-type and resulted in a corresponding reduction in S motility on 0.3% agar (Fig. 4). These results indicated that although P34A produced less pilin, the mutated pilins still assembled into functional TFP. Therefore, the absence of the kink due to substitution of the proline with alanine affected PilA production but not TFP assembly or its function. To further study the role of the α-helix structure in PilA, we generated three additional mutants by replacing amino acids near the kink region with prolines (A30P, I31P, I33P). These three proline mutants showed phenotypes similar to that of A32P in that none of the mutants produced PilA or EPS, and no swarming was detected on 0.3% agar (Fig. 4). These data demonstrated that dramatic structural changes such as a turning produced by a proline residue in the α1-N subdomain of PilA are not tolerated and may lead to instability of PilA in M. xanthus.
In this study, we characterized a pilA mutant that was deficient in S motility despite its ability to produce surface pili. The mutation mapped to A32 of PilA, which was found to be important in a previous study (Yang et al., 2010). By further characterizing this mutant as well as additional pilA point mutants at this position, we concluded that A32 is important for PilA stability, TFP assembly and related functions. We also showed that different amino acid structures at position 32 in PilA have different effects on these features.
Through genetic, biochemical and video microscopic assays, we found that PilA-A32V is produced at wild-type levels and assembles into TFP (Fig. 1). These TFP, however, are unable to retract. In the TFP assembly model, the N-terminal α-helix of PilA is packed in the TFP core, where it interacts via both hydrophobic and electrostatic interactions (Craig et al., 2006). The α1-N subdomain (amino acids 13–40) is almost entirely hydrophobic and thus provides an optimal environment for tight packing. In addition, the negatively charged residue 17E and positively charged N terminus of residue 13F point toward the filament centre and form an inter-subunit salt bridge which strengthens the binding between pilin subunits. The sequence of this PilA subdomain is optimized to maximize the contacts between neighbouring subunits. The valine replacement at the alanine 32 site may have altered the structure of the α1-N subdomain of PilA and consequently altered the inter-subunit interaction. This possible structural change in the assembled TFP could then interfere with the retraction process.
How do structural changes in A32V TFP affect retraction and S motility? We suggest the following three possibilities that would be consistent with the increased level of surface pili detected using Western blotting in the TFP shear-off assay (Fig. 1c) and the lack of S motility (Figs 1b and and2b).2b). (1) The stability of the mutated TFP is reduced due to the lack of tight packing and close interactions at the core of the TFP filament. Although PilA-A32V can be assembled into TFP by PilB, the reduced stability of the mutated TFP leads to filament disintegration upon the pulling force exerted during the PilT-mediated retraction process. The pili would break apart before the cells move forward, leading to a drastic reduction in S motility. (2) The alterations in PilA affect its interaction with the retraction motor PilT. PilT forms hexameric polymers (Satyshur et al., 2007) that are essential for TFP retraction, which is powered by the ATPase activity of PilT (Jakovljevic et al., 2008). The structural changes in the mutated TFP may prevent correct interaction of PilA with PilT or affect the force between TFP and the PilT machinery necessary for retraction. Therefore, retraction through the PilT apparatus is impaired. Initially we hypothesized that the mutated TFP may change the turning angle in the kink region of the α-helix and thus affect the structure of the TFP and the interaction with PilT. However, the P34A mutant constructed in this study showed substantial S motility (Fig. 4), demonstrating that the kink in the α-helix of PilA is less likely to be essential for TFP retraction. (3) Alterations in TFP affect signal transduction. TFP act as sensors for EPS production (Black et al., 2006), and the binding of EPS to TFP triggers pilus retraction (Li et al., 2003). Although the mutated TFP produced by A32V can still bind to EPS, the downstream signal transduction pathway to trigger the retraction process may be defective. Even if A32V does not totally block the retraction signal, it is possible that the altered TFP produce a weakened signal for retraction which may alter the balance between extension and retraction, leading to a net increase in extension.
The α-helix, especially the α1-N domain in PilA, is highly conserved among different bacteria, including non-pathogenic and pathogenic species such as P. aeruginosa and N. gonorrhoeae. Our findings for M. xanthus PilA will aid in our understanding of the minimal structural requirements for successful assembly of TFP and shed light on the structure–function relationship between PilA and TFP-dependent motility. Finally, further studies to solve the detailed structure of M. xanthus PilA are needed in order to better understand the complexity of TFP.
This work was supported by the National Institutes of Health (NIH) grant GM54666 to W.S. We acknowledge the use of cryoEM facilities at the Electron Imaging Center for NanoMachines supported in part by the NIH (1S10RR23057 to Z.H.Z.).