AglZ regulates adventurous (A-) motility in Myxococcus xanthus through its interaction with the cytoplasmic receptor, FrzCD

Summary

Myxococcus xanthus moves by gliding motility powered by Type IV pili (S-motility) and distributed motor complexes (A-motility). The Frz chemosensory pathway controls reversals for both motility systems. However, it is unclear how the Frz pathway can communicate with these different systems. In this paper, we show that FrzCD, the Frz pathway receptor, interacts with AglZ, a protein associated with A-motility. Affinity chromatography and cross-linking experiments showed that the FrzCD-AglZ interaction occurs between the uncharacterized N-terminal region of FrzCD and the N-terminal pseudo-receiver domain of AglZ. Fluorescence microscopy showed AglZ-mCherry and FrzCD-GFP localized in clusters that occupy different positions in cells. To study the role of the Frz system in the regulation of A-motility, we constructed aglZ frzCD double mutants and aglZ frzCD pilA triple mutants. To our surprise, these mutants, predicted to show no A-motility (A⁻S⁺) or no motility at all (A⁻S⁻), respectively, showed restored A-motility. These results indicate that AglZ modulates a FrzCD activity that inhibits A-motility. We hypothesize that AglZ-FrzCD interactions are favored when cells are isolated and moving by A-motility and inhibited when S-motility predominates and A-motility is reduced.

Introduction

Myxococcus xanthus is a Gram negative soil bacterium that has a complex life cycle that includes vegetative swarming, predation, and fruiting body formation. These complex behaviors require coordinated movement on solid surfaces using two very different gliding motility systems (Hodgkin and Kaiser, 1979). The first motility system is called social (S-) motility, which is similar to twitching motility in Pseudomonas aeruginosa (Wall and Kaiser, 1999). S-motility is powered by Type IV pili, localized at the leading cell pole. The pili bind to polysaccharides on the substrate or on the surface of other cells and retract, pulling cells forward (Sun et al., 2000; Wall and Kaiser, 1999). In contrast, adventurous (A-) motility is not as well understood. It is thought to be powered by a combination of unidentified motor proteins, focal adhesion complexes, and slime secretion (Mignot et al., 2007b; Wolgemuth et al., 2002). Over 30 proteins have been identified as participating in the A-motility system, but their specific functions are mostly unknown (Leonardy et al., 2007; Rodriguez and Spormann, 1999; Youderian et al., 2003; Yu and Kaiser, 2007).

To achieve directed motility, M. xanthus cells periodically reverse. Reversals allow cells to change direction and re-orient themselves in much the same way that flagellar rotation reversals in enteric bacteria allow cells to periodically change direction and follow a biased random walk (Boyd and Simon, 1982; Zusman et al., 2007). During reversals, the lagging cell pole becomes the leading cell pole and several S- and A-motility proteins are redirected from one cell pole to the other (Leonardy et al., 2007; Mauriello and Zusman, 2007; Mignot et al., 2005; Mignot et al., 2007b). Cell reversals and polarity switching are regulated by the Frz (frizzy) signal transduction system, which is similar to enteric chemotaxis pathways (Zusman et al., 2007). The Frz system consists of a cytoplasmic receptor (FrzCD), a histidine kinase (FrzE), two CheW-like coupling proteins (FrzA and FrzB), a dual response regulator protein (FrzZ), a methyltransferase (FrzF), and a methylesterase (FrzG). Most mutations in the Frz pathway cause infrequent reversals, although some mutations in FrzCD cause hyper-reversals (Blackhart and Zusman, 1985; Bustamante et al., 2004). Frz mutants have intact A- and S-motility engines but due to the loss of regulation of both reversal frequency and coordination of groups of cells, these strains do not swarm, ripple or form fruiting bodies as well as wild type (Berleman et al., 2008; Bustamante et al., 2004).

In order to investigate how the FrzCD receptor receives stimuli from the external environment, we searched for FrzCD interacting proteins. We identified several proteins of interest. One of these, AglZ, is a previously characterized A-motility protein shown to be associated with distributed complexes implicated in A-motility (Mignot et al., 2007b). Mutant analyses showed that AglZ negatively regulates FrzCD activity and functions to couple A-motility to the Frz chemosensory pathway. This work reveals new levels of complexity in the function of chemotaxis proteins and gliding motility in M. xanthus.

Results

FrzCD interacts with A-motility proteins

In order to identify proteins that interact with FrzCD, we expressed glutathione-S-transferase (GST) tagged FrzCD and GST tagged FrzCD_2-280 (Fig. 1) in E. coli and used the two fusions as bait in affinity chromatography experiments (Sambrook, 2001). For these experiments, M. xanthus strain DZ2 (wild type) cell lysates were prepared and loaded onto columns containing pre-bound GST-FrzCD. The eluted fractions were analyzed by tandem mass spectrometry (MS/MS, Proteomics/Mass Spectrometry Laboratory, UC Berkeley) and the proteins identified by searching the proteomic database for M. xanthus. Proteins that were consistently found in the eluate of four different experiments are listed in Table 1. In control experiments, GST alone did not bind any of these proteins. The affinity chromatography experiments showed potential interactions with the CheA histidine kinase homologue, FrzE, and the two CheW proteins, FrzA and FrzB of the Frz pathway. These results were anticipated as CheA and CheW are known to interact in chemotaxis systems (Inclan et al., 2007; Kentner et al., 2006; Studdert and Parkinson, 2005). We did not detect FrzE, FrzA and FrzB when the N-terminal region of FrzCD was used as bait (GST-FrzCD_2-280, Fig. 1 and Table 1), suggesting that interactions with these proteins occur through the C-terminal signaling domain of FrzCD. In contrast, the N-terminal region of FrzCD (Fig. 1) did interact with two previously described A-motility proteins, AglZ (Mignot et al., 2007b; Yang et al., 2004) and AgmU (Youderian et al., 2003) (Table 1). The interaction between FrzCD and AglZ was also observed in two separate immunoprecipitation experiments in which wild type M. xanthus cell extracts were incubated with purified antibodies against FrzCD (data not shown). Additionally, FrzCD showed possible interactions with two proteins of unknown function (no BLAST hits) and a DbsA-like homologue (Table 1).

Fig. 1.

Schematic representation of GST-FrzCD protein fusions used as bait in affinity chromatography experiments. The list of proteins that were detected is shown in Table 1.

Table 1.

Proteins detected by using GST-FrzCD fusions as bait in affinity chromatography analyses

Gene annotation, protein name	GST-FrzCD	GST-FrzCD2-280	GST
MXAN 2991, AglZ, Yang et al. (2004)	+	+	-
MXAN 4870, AgmU, Yourderian et al. (2003)	+	+	-
MXAN 3206, DsbA-like protein	+	+	-
MXAN 2736, hypothetical protein	+	+	-
MXAN 5512, hypothetical protein	+	+	-
MXAN 4143, FrzA, Bustamante et al. (2004)	+	-	-
MXAN 4144, FrzB, Bustamante et al. (2004)	+	-	-
MXAN 4146, FrzE, Bustamante et al. (2004)	+	-	-

FrzCD and AglZ interact through their N-terminal regions

To further characterize the interaction between FrzCD and AglZ, we performed in vitro protein cross-linking analyses. Full length AglZ, as well as the N-terminal pseudo-receiver domain (AglZ^Rec, residues 2-240), the two coiled-coil domains (AglZ^C-coil, residues 230-1384), and the C-terminal coiled-coil domain (AglZ^C-ter, residues 849-1384) of AglZ (Fig. 2C) (Mignot et al., 2007b) were cloned and expressed with His tags at their N-terminus, purified, and incubated with or without FrzCD in the presence of the cross-linking agent, formaldehyde. The samples were analyzed by SDS polyacrylamide gel electrophoresis (Fig. S1) and Western immunoblotting using purified anti-FrzCD antibodies. When FrzCD and either AglZ or AglZ^Rec were co-incubated in the presence of formaldehyde, we observed at least three additional bands of higher molecular weight (between 150 and 200 KDa) (Fig. S2A, lanes 11-12). This preliminary result suggests that FrzCD does not interact with the AglZ coiled-coil domains but does interact with the N-terminal pseudo-receiver domain of AglZ.

Fig. 2.

Formaldehyde cross-linking experiments with FrzCD and AglZ.

A. Western Blot with anti-FrzCD purified antibodies of purified FrzCD N-terminal (FrzCD^N-ter) or C-terminal (FrzCD^C-ter) regions or/and AglZ domains (roman numerals indicate AglZ domains shown in panel C) incubated in the presence of cross-linker (10 mM formaldehyde). Lane 1-6 FrzCD or AglZ proteins incubated alone; lane 7-10, FrzCD^N-ter co-incubated with different AglZ domains; lane 11-14, FrzCD^C-ter co-incubated with different AglZ domains. The arrowheads indicate bands only seen after co-incubating FrzCD^N-ter with AglZ full-length or AglZ^Rec and cross-linker.

B. Schematic representation of the N-terminal and C-terminal regions of FrzCD used in the in vitro cross-linking shown in A. Each protein is fused to a 6-His tag at the N-terminus. C. Schematic representation of the different AglZ domains used in the in vitro cross-linking shown in A. Each protein is fused to a 6-His tag at the N-terminus.

To assess whether the interaction between FrzCD and AglZ occurs directly through their respective N-terminal regions, we performed a second set of in vitro protein cross-linking experiments using various purified his-tagged domains of FrzCD and AglZ. For these experiments, we mixed the N-terminal (FrzCD^N-ter; amino acids 2-135) and C-terminal (FrzCD^C-ter; amino acids 133-417) domains of FrzCD (Fig. 2B) with, AglZ, AglZ^Rec, AglZ^C-coil and AglZ^C-ter (Fig. 2C) in the presence of the cross-linking agent formaldehyde. The results, presented in Fig. 2A, show that purified samples of FrzCD^N-ter, FrzCD^C-ter and AglZ^Rec produced two higher molecular weight products in the presence of the cross-linking agent (Fig. 2A, lanes 1-2, and 4). When FrzCD^N-ter was mixed with either AglZ or AglZ^Rec in the presence of formaldehyde, we observed additional bands at the expected size of 170 KDa and 50 KDa for a 1:1 FrzCD^N-ter/AglZ or 1:1 FrzCD^N-ter/AglZ^Rec, respectively (Fig. 2A, lane 7 and 8). In contrast, there were no additional bands when we incubated FrzCD^N-ter with AglZ^C-coil or AglZ^C-ter (Fig. 2A, lane 9 and 10). Because the FrzCD antibody cross-reacted with AglZ^Rec (Fig. 2, lane 4), as a control for the cross-linking experiment, FrzCD^N-ter dosage dependent assays were performed (Fig. S3). These experiments show that the new band appearing in lane 8 of Fig. 2 does not result from cross-reaction of the FrzCD antibody with AglZ^Rec as it increased in intensity as the concentration of FrzCD^N-ter increased. These results support our previous findings that only the AglZ^Rec is involved in the interaction with FrzCD and that this interaction occurs by way of the FrzCD N-terminus.

The Frz system is downstream to AglZ in the regulation of A-motility

To explore the biological function of the AglZ-FrzCD interaction, we used a strain carrying a plasmid insertion in aglZ (see Experimental Procedures) (Mignot et al., 2007b) and an in frame deletion of frzCD (Bustamante et al., 2004). We used the plasmid insertion mutant for simplicity of strain construction but first verified that the phenotypes of the plasmid insertion and deletion mutants were comparable (Fig. S4). We then compared the phenotypes of the aglZ frzCD double mutant with the phenotypes of wild type, frzCD and aglZ strains. As described previously, frzCD mutants, like wild type, exhibit both A and S-motility (Bustamante et al., 2004) while aglZ mutants show only S-motility (Fig. 3, top panels) (Mignot et al., 2007b). frz mutants have a functional S-motility system, but these strains swarm less than wild type because they are defective in regulating cell reversals and chemotaxis. The integrity of the A- and S-motility systems can be easily determined as cells exhibiting A-motility on 1.5% agar show individual cell movements at a colony edge, whereas cells exhibiting S-motility show only group movements (Hodgkin and Kaiser, 1979). When we examined the aglZ frzCD double mutant, surprisingly, we found that cells showed restored A-motility (Fig. 3, top panel). To confirm that cells were exhibiting A-motility and not S-motility, we introduced a pilA mutation into the aglZ frzCD strain. Strains that lack PilA show a complete loss of S-motility, as PilA is the major component of Type IV pili and is essential for S-motility (Wall and Kaiser, 1999; Wu and Kaiser, 1997). A strain lacking both aglZ and pilA had a smooth colony edge indicating a complete absence of motility (Fig. 4B) which was verified by time-lapse video microscopy (Mignot et al., 2007b). However, the aglZ pilA frzCD triple mutant showed restored single cell motility (Fig. 4C), confirming that the A-motility system was functional in this strain. The restoration of A-motility in the aglZ frzCD double mutant indicates that AglZ is not a structural component of the A-motility motor but rather a regulatory protein, and that FrzCD is downstream of AglZ, functioning as a negative regulator of A-motility.

Fig. 3.

A-motility and developmental aggregation analysis of strains DZ2 (wild type), frzCD, aglZ and aglZ frzCD. Cells (10 μl), at a concentration of 4 × 10⁹ cfu ml⁻¹, were spotted on CYE or CF plates containing an agar concentration of 1.5%, incubated at 32°C and photographed after 48h with a WTI charge-coupled device (CCD)-72 camera, using a Nikon Labphot-2 or a Zeiss (model 476009-9901) microscopes. Scale bar is 20 μm for the top row and 10 mm for the bottom row.

Fig. 4.

A-motility analyses of strains pilA (A), aglZ pilA (B), aglZ pilA frzCD (C), aglZ pilA frzCD^H49A (D), aglZ pilA frzF (E), aglZ pilA frzG (F), aglZ pilA frzZ (G), aglZ pilA frzCD^Δ6-130(H) and aglZ pilA frzCD^Δ6-183 (I).Cells (10 μl), at a concentration of 4 × 10⁹ cfu ml⁻¹, were spotted CYE plates containing an agar concentration of 1.5%, incubated at 32°C and photographed after 48h with a WTI charge-coupled device (CCD)-72 camera, using a Nikon Labphot-2 or a Zeiss (model 476009-9901) microscopes. Scale bar, 20 μm.

To determine whether the negative regulation of A-motility by FrzCD occurs directly as a function of the receptor activity or as an output of a functional Frz pathway, we constructed mutants defective in other critical Frz pathway genes and analyzed their A-motility phenotypes. Fig. 4D, E and G show that the aglZ pilA frzE^H49A, aglZ pilA frzF, and aglZ pilA frzZ strains all displayed A-motility phenotypes similar to aglZ pilA frzCD. These results show that methylation of the receptor by the methyltransferase, FrzF (Scott et al., 2008), and phosphotransfer from the histidine kinase, FrzE, to the response regulator, FrzZ (Inclan et al., 2007; Inclan et al., 2008) are both required for the negative regulation of A-motility and that, therefore, this regulation occurs as an output of the Frz pathway. We did not observe restoration of A-motility in aglZ pilA frzG cells (Fig. 4F). This result is in agreement with previous data showing that FrzG is not critical for motility (Bustamante et al., 2004).

Surprisingly, under starvation conditions, aglZ frzCD cells aggregated to form mounds, even though neither frzCD nor aglZ cells are able to do this (Fig. 3, bottom panel). Compared to frz mutants, aglZ frzCD cells also showed an increased frequency of cell reversals (Table 2). The increased reversal frequency of aglZ frzCD mutants may suggest the presence of another chemosensory system compensating for the loss of the Frz system in aglZ mutants or, alternatively, random reversal events that might occur in the absence of both AglZ and Frz functions. It is interesting that Leonardy et al., also observed that a mutation in the A-motility gene, RomR, restored reversals in a frzE mutant (Leonardy et al., 2007). It should be noted that although aglZ frzCD cells formed aggregates reminiscent of fruiting bodies, this strain was defective in sporulation.

Table 2.

Reversal frequencies

Strain	Average reversals in 30 minutes (# cells)*
Wild type	1.9 (27)
frzCD	0.3 (25)
aglZ	N/D**
aglZ frzCD	9.48 (43)

^*P<5^10-6

^**single cells are not motile

Analysis of A-motility in FrzCD mutants lacking the N-terminal domain

Affinity chromatography and cross-linking experiments suggested that the interaction between FrzCD and AglZ involves the N-terminal region of FrzCD (Table 1 and Fig. 2). To further investigate this finding, we constructed and characterized AglZ mutants that contained deletions in the N-terminal region of FrzCD. Specifically, frzCD^Δ6-130 mutants showed near-normal Frz pathway functioning, as the N-terminal domain of FrzCD does not appear to be critical for swarming or fruiting (Bustamante et al., 2004). The aglZ pilA frzCD^Δ6-130 (Fig. 4H) strain showed recovery of A-motility with many single cells swarming out from the edge of the colony. Since the deletion of the N-terminal domain of FrzCD (Fig. 4H) had the same effect on A-motility as the deletion of the full-length protein (Fig. 3, top panel and 4C), we conclude that the N-terminal region of FrzCD is important to achieve inhibition of A-motility. From these results one might expect that the absence of the N-terminus of FrzCD and the consequent loss of the interaction between FrzCD and AglZ would result in an A^– phenotype similar to aglZ mutants, but this is not the case. These experiments, therefore, suggest that one or more additional unidentified proteins or factors interact with the N-terminal region of FrzCD and signal to FrzCD to cause the arrest of motility in the absence of AglZ.

It is interesting to note that the aglZ pilA frzCD^Δ6-183 triple mutant, which contains the larger frzCD^Δ6-183 deletion that confers a hyper-reversing non-swarming phenotype (Bustamante et al., 2004), hyper-reverses and fails to swarm (Fig. 4I). This result also suggests that FrzCD is downstream of AglZ.

FrzCD and AglZ form dynamic cytoplasmic clusters with different localization patterns

Since our experiments indicated that FrzCD and AglZ interact, we were interested in using fluorescence microscopy to investigate the sub-cellular localization of FrzCD and AglZ and their relationship to each other in vivo, anticipating that they would co-localize. Accordingly, we constructed a M. xanthus strain carrying both frzCD-gfp (Mauriello et al., 2009) and aglZ-mCherry fusions. The strain expressing the FrzCD-GFP fusion was proficient at responding to attractants/repellents, and showed normal A-motility swarming and fruiting; however, frzCD-gfp cells showed a small reduction in S-motility swarming and reduced cell reversals (half of wild type) (Mauriello et al., 2009). The AglZ-mCherry fusion was stably expressed and fully functional (Fig. S4 and S5). To visualize FrzCD and AglZ in moving cells, we spotted cells on a slide coated with a thin layer of nutrient agar and captured images every 30 s for 10 min by time-lapse fluorescence microscopy. Fig. 5 (first row) shows bright field images of the cells. Fig. 5 (second row) shows the FrzCD-GFP fluorescence. In these cells, the FrzCD-GFP fluorescence appeared as multiple clusters, as previously described (Mauriello et al., 2009). The position of the clusters was dynamic and appeared random, although clusters were not observed at the cell poles. Fig. 5 (third row) shows that in these same cells, AglZ-mCherry also localized in clusters. These clusters showed a different localization pattern than FrzCD-GFP. The AglZ-mCherry clusters were similar to the AglZ-YFP clusters described by Mignot et al. (Mignot et al., 2007b): they formed at the leading cell pole but became distributed along the cell length. The clusters remained fixed relative to the substratum as cells moved forward, suggesting that they might identify focal adhesion sites. In Fig. 5 (bottom row), merged images of the FrzCD-GFP and AglZ-mCherry fluorescence show that the clusters did not co-localize: AglZ-mCherry clusters localized primarily at the leading cell pole or along the cell in positions not occupied by FrzCDGFP clusters; in contrast, FrzCD-GFP clusters localized in distributed but non-polar positions. These results suggest that FrzCD and AglZ occupy mutually exclusive positions in the cell, perhaps limiting their interactions to the interfaces between clusters.

Fig. 5.

Analysis of FrzCD-GFP and AglZ-mCherry clusters in vivo. frzCD-gfp aglZ-mCherry cells were imaged at 1 min intervals. Pictures in the first row were taken in bright-field; pictures in the second row show FrzCD-GFP clusters; pictures in the third row show AglZ-YFP clusters; pictures in the last row show overlaid images from the second and third rows. Arrows indicate the positions of AglZ-YFP clusters. Scale bar, 5μm.

A. A field including several cells.

B. Time-lapse of a single cell moving forward.

C. A field showing a compact group of cells.

D. Isolated cell in 1% methylcellulose.

Interestingly, when we examined AglZ-mCherry localization in groups of cells or isolated cells in methylcellulose where A-motility is reduced or absent (Sun et al., 2000), we found that AglZ-mCherry appears diffuse or as a single polar cluster (Fig. 5C and D). In contrast, under these same conditions, FrzCD-GFP clusters remained unchanged. This suggests that the putative AglZ focal adhesion complexes are not assembled when cells are moving by S-motility and that AglZ interactions with FrzCD might be reduced when S-motility is predominant, so that the Frz pathway can inhibit A-motility.

We note that FrzCD-GFP localization did not change in an aglZ mutant (Fig. 6A), nor did AglZ-YFP localization change in a frzCD mutant (Fig. 6B). These results suggest that AglZ and FrzCD clusters form independently of each other.

Fig. 6.

Analysis of FrzCD-GFP and AglZ-YFP clusters in vivo.

A. FrzCD-GFP clusters in frzCD-gfp and frzCD-gfp aglZ cells.

B. AglZ-YFP clusters in aglZ-yfp and aglZ-yfp frzCD cells.

The N-terminal domain of FrzCD influences FrzCD and AglZ cluster localizations

To determine whether the N-terminal domain of FrzCD affects the sub-cellular localization of FrzCD and AglZ, we constructed a strain carrying frzCD^Δ6-130-gfp and aglZ-mCherry fusions. In this strain, FrzCD^Δ6-130-GFP was expressed at wild type levels (Fig. S6) and the cells responded to attractants/repellents, formed fruiting bodies, and showed normal A-motility swarming (Fig. S7). However, this strain showed a small reduction in S-motility swarming and cell reversals compared to the parental strain (Fig. S7). Analysis of FrzCD^Δ6-130-GFP localization in this strain showed that the N-terminal FrzCD deletion caused significant defects in cluster localization (Figures 5 and 7, second rows). For example, while FrzCD-GFP formed numerous and tight clusters localized preferentially in the center of cells, the FrzCD^Δ6-130-GFP clusters appeared diffuse and showed no preferential positions in cells. Nevertheless, the clusters were still dynamic, constantly changing their number, position and intensity (Fig. 7B, second row). Interestingly, in these cells, FrzCD^Δ6-130-GFP and AglZ-mCherry clusters frequently co-localized, as indicated by the appearance of yellow fluorescence. Co-localization was seen most frequently at the cell poles, but was also found in other regions of the cell (Fig. 7, bottom row). In contrast, FrzCD-GFP clusters did not co-localize with AglZ-mCherry clusters (Fig. 5). Our results suggest that the interaction between the N-terminal region of FrzCD and AglZ is important for FrzCD cluster formation and for the mutual localization patterns of FrzCD and AglZ, indicating that there is a link between protein interactions, sub-cellular localization and regulation of A-motility.

Fig. 7.

Analysis of FrzCD^Δ6-130-GFP and AglZ-mCherry in vivo. frzCD^Δ6-130-gfp aglZ-mCherry cells were imaged at 1 min intervals. Pictures in the first row were taken in bright-field; pictures in the second row show FrzCD^Δ6-130-GFP; pictures in the third row show AglZ-YFP clusters; pictures in the last row show overlaid images from the second and third rows. Arrows indicate the positions of AglZ-YFP clusters. Scale bar, 5μm. A. A field including several cells. B. Time-lapse of a single cell moving forward.

Discussion

The Frz chemosensory system controls cell reversals for both the A- and S- motility systems of M. xanthus. In this study, we searched for proteins that interact with FrzCD, a methyl-accepting chemotaxis protein (MCP) receptor for the pathway. One unexpected finding was that the uncharacterized N-terminal domain of FrzCD interacts with two A-motility proteins, AglZ and AgmU. AglZ was previously identified and characterized as an interacting partner of MglA, a Ras-like GTPase that is required for both A- and S-motility (Hartzell and Kaiser, 1991; Yang et al., 2004). AglZ has a pseudo-receiver domain at the N-terminus and two large coiled-coil domains at the C-terminus. Mignot et al. showed that AglZ localizes in clusters that remain fixed relative to the substratum as cells move forward, suggesting the existence of focal adhesion complexes powering cell movement during A-motility (Mignot et al., 2007b). AgmU was previously described as an A-motility protein based on mutant studies, but little is currently known about AgmU (Youderian et al., 2003).

Mutant analyses showed that AglZ is not a component of the engine powering A-motility, but rather a regulator of the A-motility system, since the aglZ frzCD double mutant shows restored A-motility. AglZ shares numerous similarities with the S-motility protein FrzS. They have similar physical structure (N-terminal pseudo-receiver domain and C-terminal coiled-coil domain) and they both localize mostly at the leading pole, switching polarity in a coordinate manner as cells reverse (Mignot et al., 2005, 2007a; Yang et al., 2004). It has been proposed that AglZ and FrzS are located downstream of the Frz pathway in the regulation of A- and S-motility respectively (Inclan et al., 2008; Mauriello and Zusman, 2007). However, our genetic analyses strongly suggest that AglZ is upstream to the Frz pathway providing an input by directly interacting with the N-terminal domain of FrzCD (Fig. 8). Since FrzS is so similar to AglZ, we were interested in determining if frzS frzCD double mutants would show restored S-motility. They did not: these mutants showed that FrzS is downstream to the Frz pathway (Y. Inclán and D. Zusman, unpublished data). This suggests that AglZ and FrzS, despite their structural similarities, might regulate A- and S-motility through different mechanisms.

Fig. 8.

A model describing roles for AglZ and FrzCD in regulating A-motility. AglZ is shown as acting upstream to the Frz pathway in regulating A-motility. According to this model, FrzCD is an inhibitor of A-motility, while AglZ is shown relieving this inhibition when it binds to the N-terminal domain of FrzCD. We speculate that the interaction between AglZ and FrzCD occurs primarily when cells are isolated and need to turn on the A-engine. Signals to the A-motility engine might be transferred via FrzZ, which is phosphorylated by FrzE. The bold arrow indicates the direction of the phosphotransfer. It is still not known how the phosphorylation states of FrzZ, a double response regulator, impacts motility.

To further study the interaction between FrzCD and AglZ, we investigated the localization of AglZ and FrzCD by fluorescence microscopy. Interestingly, in living cells, AglZ-mCherry and FrzCD-GFP appear as punctate clusters, as previously described (Mignot et al., 2007b; Mauriello et al., 2009). While affinity chromatography and cross-linking experiments suggest that the proteins interact, fluorescence microscopy showed that FrzCD and AglZ clusters appear to occupy different positions in cells. This suggests that AglZ and FrzCD have a stronger affinity for some other organizing element or cytoplasmic filaments than for each other and that any interactions between AglZ and FrzCD must be limited, perhaps to some cluster interfaces. This hypothesis is consistent with the in vitro cross-linking data, in which only a small population of FrzCD molecules participates in the formation of FrzCD-AglZ complexes, while a much larger portion of FrzCD participates in the formation of homo-oligomers. Interestingly, the localization of FrzCD and AglZ is changed in the absence of the N-terminal region of FrzCD, suggesting that the interaction between the N-terminal region of FrzCD and AglZ is important for their localization with respect to each other and also that localization patterns and regulation of A-motility are linked.

Another indication of the link that exists between protein localization and the regulation of motility is the observation that when cells are placed under conditions that favor S-motility, for example in cell groups or isolated cells in methylcellulose, AglZ-mCherry appears diffuse or as a single polar cluster (Fig. 5C and D). This suggests that the putative AglZ-focal adhesion complexes are not assembled when cells are moving by S-motility. These results also suggest that AglZ interactions with FrzCD might be reduced in cell groups, so that the Frz pathway can inhibit A-motility when S-motility is predominant.

Figure 8 shows a model for the regulation of A- and S-motility by the Frz pathway based on the data presented in this paper. In this model, AglZ is shown as acting upstream to the Frz pathway in regulating A-motility. FrzCD is viewed as an inhibitor of A-motility, while AglZ is shown relieving this inhibition when it binds to the N-terminal domain of FrzCD since mutations in AglZ show no A-motility. We speculate that the interaction between AglZ and FrzCD occurs primarily when cells are isolated and need to turn on the A-engine, but is inhibited when cells are in groups, a situation that favors S-motility. Additionally, it is possible that the inhibition of A-motility is sensitive to the presence of neighboring cells or objects in its path. This would be in agreement with the recent discovery that FrzCD cluster distributions are sensitive to cell-cell contacts (Mauriello et al., 2009). Clearly, more work needs to be done to explore the implications of this regulatory circuit.

In conclusion, our examination of FrzCD interacting proteins has revealed a role for AglZ as a modulator of the inhibitory activity of the Frz pathway on A-motility. The localization of FrzCD and AglZ in living cells suggests that they are organized as clusters that interact directly with distributed A-motility motors. Clearly, a link exists between protein interactions, localization and regulation of motility. The complexities of motility and its regulation in M. xanthus are still far from being well understood. However, unraveling these mysteries should yield new insights into the dynamics of the bacterial cell and its ability to interact with its environment.

Experimental Procedures

Strains and growth conditions

Bacterial strains and plasmids are listed in Table 3. All M. xanthus strains were cultured in CYE medium, which contains 10 mM MOPS pH 7.6, 1% (w/v) Bacto Casitone (BD Biosciences), 0.5% Bacto yeast extract and 4 mM MgSO₄ (Campos et al., 1978). For phenotypic assays, cells (10 μl), at a concentration of 4 × 10⁹ cfu ml⁻¹, were spotted on CF-agar plates or CYE plates containing an agar concentration of 0.5 or 1.5 %, incubated at 32°C and photographed after 48h with a WTI charge-coupled device (CCD)-72 camera, using a Nikon Labphot-2 or a Zeiss (model 476009-9901) microscopes.

Table 3.

Strains and plasmids used in this study

M. xanthus strains	Genotype	Reference source
DZ2	Wild type	Campos and Zusman (1975)
DZ4469	pilA::tet	Vlamakis et al. (2004)
TM7	aglZ::kan	Mignot et al. (2007))
DZ4480	ΔfrzCD	Bustamante et al. (2004)
DZ4629	frzEH49E	Inclan et al. (2008)
DZ4482	ΔfrzG	Bustamante et al. (2004)
DZ4483	ΔfrzF	Bustamante et al. (2004)
DZ4484	ΔfrzZ	Bustamante et al. (2004)
DZ4487	frzCDΔ6-130	Bustamante et al. (2004)
DZ4484	frzCDΔ6-183	Bustamante et al. (2004)
DZ4725	pilA::tet aglZ::kan	This study
DZ4726	ΔfrzCD aglZ::tet	This study
DZ4727	ΔfrzCD pilA::tet aglZ::kan	This study
DZ4728	frzEH49E pilA::tet aglZ::kan	This study
DZ4729	ΔfrzG pilA::tet aglZ::kan	This study
DZ4730	ΔfrzF pilA::tet aglZ::kan	This study
DZ4731	ΔfrzZ pilA::tet aglZ::kan	This study
DZ4732	frzCDΔ6-130pilA::tet aglZ::kan	This study
DZ4733	frzCDΔ6-183 pilA::tet aglZ::kan	This study
DZ4734	frzCD-gfp aglZ-mCherry::kan	This study
DZ4770	ΔaglZ	This study
DZ4725	ΔfrzCD aglZ-yfp	This study
DZ4726	frzCD-gfp aglZ::kan	This study
DZ4727	frzCDΔ6-130-gfp	This study
DZ4728	frzCDΔ6-130-gfp aglZ-mCherry	This study
E. coli strains
DH5α	Host for cloning	Invitrogen
tuner	Host for overexpression	Novagen
Plasmids
pET28a	Protein expression vector	Novagen
pGEX-2TK	Protein expression vector	Invitrogen
pDPA161	pET28a with his6::frzCD	Astling (2003)
pBN1	pGEX-2TK with gst::frzCD (2-417AA)	This study
pBN2	pGEX-2TK with gst::frzCD (2-280 AA)	This study
pBN3	pET28a with his6::frzCD (2-135 AA)	This study
pBN4	pET28a with his6::frzCD (132-417 AA)	This study
pBN5	pET28a with his6::aglZ (2-1384 AA)	This study
pBN6	pET28a with his6::aglZ (2-240 AA)	This study
pBN7	pET28a with his6::aglZ (238-1384 AA)	This study
pBN8	pET28a with his6::aglZ(849-1384 AA)	This study
pEM147	pBJ114 with aglZ::m-cherry	This study
PEM124	pBJ114 with frzCDΔ6-130 deletion cassette	This study

Double and triple mutants were constructed by electroporating M. xanthus cells with 4 μg plasmid DNA or 1 μg of chromosomal DNA. Cells were plating on 100μg/ml kanamycin to select the plasmid (pZerO carrying from nucleotide position 101 to 601 of aglZ) insertion in aglZ (Mignot et al., 2007) or tetracycline to select the insertion of the tet in pilA (Vlamakis et al., 2004). BJAglZY (Mignot et al., 2007) was used to electroporate ΔfrzCD cells to obtain strain DZ4725.

To construct the ΔaglZ strain, and in frame deletion cassette was produced with primers A1 (AAGCTTCCGTCAGCCCGCAGGCGTTCA), A2 (GGTACCATCATGCTCGCTTTCGACGAT), A3 (GGTACCGGACCCTGGTCATCCAGT), A4 (GAATTCGTAGCACTGGAGCCGGATGA) and cloned into pBJ114 (Bustamante et al., 2004). Transformants were obtained by homologous recombination based on a previously reported method (Bustamante et al., 2004). To construct the frzCD-gfp aglZ-mCherry strain, the yfp gene in plasmid pBJaglZ-yfp (Mignot et al., 2007b) was replaced with the m-cherry gene previously amplified with primers 1 (GGATCCACCGGTCGCCACCATGGTGAGCAAGGGCGAGGAG) and 2 (AAGCTTTACTTGTACAGCTCGTCCATGCC). pBJaglZ-yfp and m-cherry were digested with BamHI and HindIII. 4 μg of the resulting plasmid named pEM147, were used to electroporate strain DZ4620 (DZ2 with frzCD-gfp) (Mauriello et al., 2009). A clone was selected for Kan resistance and named DZ4734.

A frzCD^Δ6-130 -gfp gene fusion was produced by constructing a cassette to delete from aminoacid 6 to 130 of FrzCD. The in frame deletion cassette was constructed with primers FrzCD7 (GAGCTGTACAAGTCCGGTACCATGTCCCTGGACACCCCCAA), FrzCD11 (GTCGACGACGATGGAGAAGCCCTT), Frzup (GAGCTCCCAGCATCGGGCTGTCAT), GFP9 (GGTACCGTCGACTGCAGAATT) and cloned into pBJ114 (Bustamante et al., 2004). The resulting plasmid was named pEM124 and used to transform DZ4620 (DZ2 with frzCD-gfp) (Mauriello et al., 2009). Transformants were obtained by homologous recombination based on a previously reported method (Bustamante et al., 2004). The resulting strain named DZ4727 was transformed with plasmid pEM147. The resulting strain, carrying frzCD^Δ6-130 -gfp and aglZ-mCherry fusions, was named DZ4728.

Protein expression and purification

The coding sequence of each protein or protein fragment was amplified by PCR from genomic DNA with the following primers: full-length FrzCD (for GST-FrzCD fusion), GGATCCTCCCTGGACACCCCCAACGAG and GAATTCTAGTCGGCCTTGAACCGCTTGA, FrzCD_2-280 (for GST-FrzCD_2-280 fusion), GGATCCTCCCTGGACACCCCCAACGAG and GAATTCTCACCCGCGCGGCTGCCTTCC; FrzCD^N-ter (amino acids 2-135), GGATCCTCCCTGGACACCCCCAACGAG and AAGCTTAGCGCAGCGTCTCGATGACCT; FrzCD^C-ter (amino acids 132-417), GGATCCGAGACGCTGCGCACCTTCGT and AAGCTTAGTCGGCCTTGAACCGCT; AglZ (amino acids 2-1384), GGATCCAGCGTCGGGTCCTCATCGTCG and AAGCTTAGTCGTCGTCTTCCTTGGCCGT; AglZ^Rec (amino acids 2-240), GGATCCAGCGTCGGGTCCTCATCGTCG and AAGCTTAGCGCACGTTCCAGATTTCAGAGA; AglZ^C-coil (amino acids 238-1384), GGATCCAGCGCGAGCTGCTCTCCGGCG and AAGCTTAGTCGTCGTCTTCCTTGGCCGT; AglZ^C-ter (amino acids 849-1384), GGATCCATTGGTCATCTCCGCAGCGAGCT and AAGCTTAGTCGTCGTCTTCCTTGGCCGT. The underlined sequences and bold letters represent restriction sites and translation stop codons, respectively. PCR products were digested and inserted into pGEX-2TK (Invitrogen) or pET28a (Novagen). All constructs were confirmed by DNA sequencing. Expression of the recombinant proteins was induced by 1 mM IPTG (isopropyl-h-d-thiogalactopyranoside). Cells were then cultured at 18 °C for 20 h and harvested by centrifugation at 8,000 rpm for 10 min.

To purify GST-tagged proteins, pellets were suspended in lysis buffer (PBS with 8% (v/v) Glycerol and 0.1% (w/v) CHAPS) and cells lysed by sonication on ice. The lysates were centrifuged twice (18,000 rpm, 4°C, 30 min) to remove debris prior to the purification. Supernatants were loaded into 1 ml GSTrap™ HP column (GE Healthcare) via a ÄKTA FPLC (GE Healthcare). Proteins were washed with the same buffer, and then eluted with elution buffer (PBS with 8% (v/v) Glycerol and 0.1% (w/v) CHAPS, 10 mM reduced glutathione).

To purify His-tagged proteins, cell pellets were resuspended in a different lysis buffer (50 mM HEPES pH 8.0, 500 mM NaCl), and 5 ml HisTrap™ nickel columns (GE Healthcare) were used. The eluition was performed by using a buffer containing 50 mM HEPES pH 8.0, 500 mM NaCl, 250 mM imidazole. For the purification of His-tagged FrzCD full-length protein, 8% (v/v) Glycerol and 0.1% (w/v) CHAPS were added in both lysis and elution buffer. Eluted proteins were concentrated using Millipore centrifugal filter devices with a 10 kDa cut-off. Protein concentrations were determined by Bradford assays (Bio-Rad).

GST affinity chromatography

0.1 mg of purified GST-FrzCD or GST-FrzCD_2-280 protein was injected into a new 1 ml GSTrap™ HP column, which was pre-washed with PBS lysis buffer. M. xanthus strain DZ2 was cultured in CYE medium up to 100 Klett. Cells were harvest from 50 ml culture and suspended into 5ml PBS, 8% (v/v) Glycerol and 0.1% (w/v) CHAPS. Cells were lysed by sonication on ice and centrifuged. Supernatants were injected into GST-FrzCD (or GST-FrzCD_2-280) pre-bound columns. Washing was performed in the presence of PBS, 8% (v/v) Glycerol and 0.1% (w/v) CHAPS for 50 column volumes, elution in the presence of PBS, 8% (v/v) Glycerol and 0.1% (w/v) CHAPS, 10 mM reduced glutathione. The peak fractions were collected and digested overnight with Trypsin (sequencing grade, Roche) at 1:100 w/w rate (Trypsin: protein) as previously described (Scott et al., 2008). Tandem mass spectrometry (MS/MS, Proteomics/Mass Spectrometry Laboratory, UC Berkeley) was performed; spectra data were generated by Bioworks software (ThermoFinnigan) and converted to MASCOT format. The proteins in the eluted sample were identified by searching the proteomic database for M. xanthus. As a negative control, a similar experiment was performed in parallel using GST protein as bait.

In vitro protein cross-linking

In vitro protein cross-linking reactions were performed in 20 mM HEPES pH 8.0, 100 mM NaCl, 0.5 mM EDTA, 8% (v/v) glycerol in 20 μl final volumes. Proteins were diluted to the following concentrations to keep them at a 1:1 molar ratio: FrzCD 5 μg/ml, FrzCD^N-ter 2.5 μg/ml, FrzCD^C-ter 2.5 μg/ml, AglZ 25 μg/ml, AglZ^Rec 5 μg/ml, AglZ^C-coil 15 μg/ml, AglZ^C-ter 10 μg/ml. In the titration experiment shown in Fig. S3, while the concentration of AglZ^Rec was kept constant (2.5 μg/ml), we used different concentrations of FrzCD (from 0.25 to 25 μg/ml). 10 mM formaldehyde was added into each pre-mixed reaction. Reactions were stopped by adding 50 mM Tris-HCl pH 8.0 after 20 min of incubation at room temperature. 1 μl sample of each reaction was used for western blot.

Immunoblot analysis

Sample were ran on 10% SDS PAGE. After electrophoresis, gels were transferred to Trans-Blot™ nitrocellulose membrane (Bio-Rad). Blots were probed with anti-FrzCD antibody as previously described (Scott et al., 2008) and with ImmunoPure™ Goat anti-rabbit secondary antibody (Thermo Scientific). The blots were developed with Western lightning™ chemiluminescence reagent plus (PerkinElmer) and X-OMAT 2000A processor (KODAK).

Time-course and time-lapse fluorescence microscopy

For motion analysis by time-lapse microscopy, 10 μl of cells from 4 × 10⁸ cfu ml⁻¹ vegetative CYE cultures were spotted on a thin fresh 0.5 CTT agar (Wu and Kaiser, 1997) pad atop a slide. A cover slip was added immediately on the top of the pad, and the obtained slide was analyzed by microscopy using a Delta-Vision microscope. To look at cells in methylcellulose, 10 μl of cells were spotted on a slide, previously supplied of a silicon gasket, submerged with 100 μl of 1% methylcellulose and covered with a cover slip. Typical time-lapse movies were shot for 10 min with frames captured every 30 s. For fluorescence microscopy, the images were captured with the FITC-filter or Rhodamine-filter. To avoid blue light toxicity, the time of illumination was limited to 0.5 s in the presence of a 50% neutral density filter. Movies were obtained by processing the series of images collected with the Quick Pro Time software.