Polar Localization Hub Protein PopZ Restrains Adaptor-Dependent ClpXP Proteolysis in Caulobacter crescentus

Regulated proteolysis is critical for the cell cycle progression of bacteria, such as Caulobacter crescentus. According to one model, this regulated proteolysis requires localization of the ClpXP protease at the stalked pole for its subsequent degradation of substrates, such as CtrA. This study offers evidence that supports an alternative model to explain how localization might influence protein degradation. Using a delocalized in vivo system created by the deletion of a polar organizing protein, PopZ, we show that activation of the ClpXP protease is independent of its polar localization. The data point to a role for PopZ in restraining ClpXP activity, likely by controlling the activity of upstream regulators of protease activity, such as CckA, though changes in its localization.

ABSTRACT

In Caulobacter crescentus, timely degradation of several proteins by the ClpXP protease is critical for proper cell cycle progression. During the cell cycle, the ClpXP protease, the substrate CtrA, and many other proteins are localized to the stalked pole dependent on a polar interaction hub composed of PopZ protein oligomers. Prior work suggests that the localization of ClpXP, protease substrates, and cofactors is needed for recognition of substrates, such as CtrA, by ClpXP. Here, we formally test this hypothesis by examining the role of PopZ in ClpXP activity and find, surprisingly, that CtrA degradation is enhanced in cells lacking polar localization due to loss of PopZ. The ClpXP adaptor CpdR is required for this enhanced degradation of CtrA and other adaptor-dependent substrates, but adaptor-independent substrate degradation is not affected upon loss of PopZ. We find that overexpression of PopZ also leads to faster degradation of CtrA but is likely due to nonphysiologically relevant recognition of CtrA by ClpXP alone, as loss of CpdR does not affect this enhancement. Our main conclusion is that loss of PopZ, and therefore loss of polar localization, does not result in the loss of ClpXP-regulated proteolysis, as would be predicted from a model which requires polar localization of ClpXP for its activation. Rather, our data point to a model where PopZ normally restrains ClpXP proteolysis by promoting the inactivation of the CpdR adaptor, perhaps through the activity and localization of the CckA kinase.

IMPORTANCE Regulated proteolysis is critical for the cell cycle progression of bacteria, such as Caulobacter crescentus. According to one model, this regulated proteolysis requires localization of the ClpXP protease at the stalked pole for its subsequent degradation of substrates, such as CtrA. This study offers evidence that supports an alternative model to explain how localization might influence protein degradation. Using a delocalized in vivo system created by the deletion of a polar organizing protein, PopZ, we show that activation of the ClpXP protease is independent of its polar localization. The data point to a role for PopZ in restraining ClpXP activity, likely by controlling the activity of upstream regulators of protease activity, such as CckA, though changes in its localization.

INTRODUCTION

Proteolysis plays an important role in facilitating cell cycle progression and various developmental transitions in bacteria. Examples are cell cycle progression in Caulobacter crescentus (here Caulobacter) and cellular transition from vegetative to sporulation stage in Bacillus subtilis (1–4). The cell cycle of Caulobacter starts with a replication-incompetent motile swarmer cell (G1-like phase). In response to developmental cues, the swarmer (SW) cell differentiates into a stalked (ST) cell. The ST cell is capable of replication and division to give birth to a new SW cell. Following replication and asymmetric division, the mother ST cell immediately starts DNA replication and another round of cell division, whereas the newborn SW cell again has to differentiate into ST cells in order to continue its cell cycle (5, 6). To maintain such stringent control during the SW-to-ST transition, levels of many proteins, including that of a protein called CtrA, change dramatically (3, 7).

CtrA is a transcriptional factor that controls transcription of ∼95 genes and also functions as an inhibitor of DNA replication (8, 9). In SW cells, CtrA is phosphorylated by a membrane-bound bifunctional kinase, CckA, via the phosphotransferase ChpT (10–12). The phosphorylated CtrA binds tightly to DNA at chromosomal origin of replication (oriC) to block the initiation of replication (3, 8, 12). The same CckA-ChpT kinase pathway also phosphorylates the ClpX adaptor and response regulator CpdR in the SW cell, which functionally inactivates it (12–15). During the SW-to-ST transition, ChpT-CckA dephosphorylates both CtrA and CpdR via shuttling of the phosphoryl groups back to CckA (16, 17). The dephosphorylated CpdR then drives localization of ClpXP protease to the stalk pole (13). In parallel with this, other cofactors, namely PopA and RcdA, were shown to localize with CtrA to the same stalk pole (18, 19). This convergent localization of ClpXP protease and the CtrA substrate at the stalk pole was postulated to increase the local concentration of the protease and the substrate, leading to the degradation of CtrA (19, 20). The destruction of CtrA then allows the assembly of replication machinery at oriC, which then initiates the replication process.

In Caulobacter, the polar organizing protein PopZ was found to function as a scaffold to recruit the accessory factors and the protease complex involved in CtrA degradation at the stalked pole (21). Besides serving as a scaffold, PopZ also anchors sister chromosomes at the stalked pole by directly binding to ParB, which in turn binds to parS sequences near the chromosomal origin to participate in chromosome segregation (22, 23). PopZ also mediates polar localization of proteins involved in cellular signaling, including both the transmembrane histidine kinase CckA and the DivJ protein, through SpmX binding (23, 24). It was proposed that PopZ functions as a switch between chromosome tethering and protein scaffolding during the SW-to-ST transition in Caulobacter to accommodate programmed asymmetry during the cell cycle (21). Further dissection of the PopZ protein revealed an N-terminal region, which is sufficient for binding all of its partner proteins, and a C-terminal region for homo-oligomerization (25). Together, these protein localization studies supported a model where spatial compartmentalization of the protease ClpXP and the substrate CtrA promotes removal of CtrA during SW-to-ST transition in Caulobacter.

Recent studies suggest that spatial localization might not be critical for CtrA degradation. For example, deletion of RcdA results in loss of CtrA degradation (18), but RcdA mutants that are unable to localize to the stalk pole can still sustain normal cell cycle-dependent degradation of CtrA (26). Localization of ClpXP to the stalk pole is not necessary for all of its activity, as some ClpXP substrates, such as FtsZ, were still degraded in a SW cell where ClpXP was shown to be delocalized (27). Finally, in vitro reconstitution experiments using purified proteins supported an adaptor hierarchy model where CpdR, RcdA, and PopA work in a coordinated fashion as adaptors to degrade many substrates, including CtrA (4, 15).

In this study, we find that CtrA degradation is enhanced in PopZ-lacking cells, suggesting that PopZ restrains CtrA degradation. The ClpXP adaptor CpdR is required for this enhanced degradation, as CtrA was stabilized in a ΔcpdR ΔpopZ strain. This degradation enhancement in ΔpopZ cells also extends to other CpdR- and RcdA-dependent substrates, namely, PdeA and TacA, respectively. These results indicate that ClpXP protease activity is enhanced at the level of the CpdR adaptor. However, ClpXP activity is not globally stimulated in ΔpopZ cells, as degradation of adaptor-independent substrates, such as a degron-appended green fluorescent protein (GFP), is not affected. Overexpressing PopZ also leads to enhanced degradation of CtrA, which we propose reflects a nonphysiological recognition of CtrA by ClpXP alone, as loss of the CpdR adaptor does not affect this enhancement. Together, these results support a model where PopZ-mediated localization of the protease ClpXP and its adaptors is not essential for its activation, but rather PopZ may affect protein degradation principally through regulation of upstream indirect regulators of the ClpXP protease, such as the CckA kinase.

RESULTS

Degradation of CtrA is enhanced in cells lacking the polar organizing protein PopZ.

PopZ is a scaffolding protein that facilitates polar localization of a multitude of proteins, including those that are directly involved in CtrA degradation, such as CpdR, RcdA, PopA, and ClpX (21, 28). Cells lacking PopZ have morphological defects and fail to localize the aforementioned proteins to the stalked pole (21, 28). Since polar localization of the ClpXP protease and the substrate CtrA, facilitated by the cofactors CpdR/RcdA/PopA, was postulated to be critical for CtrA degradation (Fig. 1D), we hypothesized that CtrA degradation would be lost in cells lacking PopZ if this model was correct. Contrary to this hypothesis, we observed that CtrA was degraded even more rapidly in ΔpopZ cells than in wild-type cells (Fig. 1A and B). CtrA stability and cell morphology were restored to those of wild-type cells when this ΔpopZ strain was transformed with a plasmid expressing the protein PopZ, suggesting that the enhanced degradation of CtrA in the ΔpopZ strain is due to specifically to the loss of PopZ (Fig. 1A to C). Together, these results suggest that the scaffolding protein PopZ restrains CtrA degradation, and they support our in vitro work, where CpdR, RcdA, and PopA-cdG physically interact, even in the absence of subcellular localization, to stimulate ClpXP-mediated degradation of CtrA (Fig. 1D) (4).

FIG 1.

CtrA degradation is enhanced in cells lacking the polar organizing protein PopZ. (A) CtrA degradation in wild-type, ΔpopZ, and ΔpopZ complemented by a PopZ-expressing plasmid cells (n.b., the xylose promoter is leaky and PopZ is expressed even without the addition of inducer). Cells were grown to exponential phase in peptone-yeast extract (PYE), and then translation was blocked by adding kanamycin. CtrA stability was monitored by Western blotting analyses by probing the blots with anti-CtrA antibody. ClpP was used as a loading control. (B) Quantitation of Western blots. Bands corresponding to CtrA and ClpP were quantified using ImageJ (NIH, USA), CtrA band intensities were normalized to ClpP intensities and normalized to time zero. Data represent mean ± standard deviation (SD) of two independent experiments. (C) Expressing PopZ from a plasmid restored wild-type morphology of ΔpopZ cells. (D) A model depicting adaptor complex-mediated proteolysis of CtrA by ClpXP. CpdR, RcdA, and PopA assemble in a hierarchical manner to deliver many substrates, including CtrA, to ClpXP for degradation (4).

Adaptor-independent ClpXP and non-ClpXP proteolysis is not affected in cells lacking PopZ.

The loss of PopZ protein stimulated CtrA degradation. This could be explained by prolific activation of ClpXP or proteolysis in general or by activation of the adaptors needed to degrade CtrA in vivo. To address this, we monitored degradation of a ClpXP reporter substrate comprised of GFP fused to the C-terminal degron of CtrA (GFP∼CtrA₁₅), which does not require adaptors for degradation. Degradation of M2 epitope-tagged GFP∼CtrA₁₅ is not affected in ΔpopZ cells compared to that in wild-type cells, suggesting that the stimulation of CtrA degradation is specific to changes in adaptor activity (Fig. 2A and B). To examine whether protein degradation is globally stimulated in ΔpopZ cells, we monitored degradation of a ClpXP-independent substrate DnaA. Degradation of DnaA was unaffected in ΔpopZ cells compared to that in wild-type cells, suggesting that degradation enhancement in ΔpopZ cells is specific to ClpXP-dependent substrates (Fig. 2C and D). Therefore, we next set out to examine if changes in adaptor activity explained the increased CtrA degradation in ΔpopZ cells.

FIG 2.

Degradation of adaptor-independent or ClpXP-independent proteolysis is not affected by popZ. (A) Wild-type cells and ΔpopZ cells expressing an M2-tagged GFP∼CtrA₁₅ were grown to exponential phase and induced with 0.3% xylose for 2 h. Protein synthesis was then blocked by the addition of chloramphenicol. Lysates from equal volumes of cells were collected at indicated time points for SDS-PAGE gels. M2-GFP∼CtrA₁₅ stability was monitored by Western blotting analyses by probing the blots with an anti-M2 antibody. ClpP is used as a loading control. (B) Quantitation of Western blots. (C) Degradation of the ClpXP-independent substrate DnaA is not affected in ΔpopZ cells. Conditions used were similar to those described in the Fig. 1A legend, except that the blot was probed using an anti-DnaA antibody. ClpP was used as a loading control. (D) Quantitation of Western blots for DnaA. Data represent mean ± SD from two biological replicates.

Degradation of both CpdR- and RcdA-dependent substrates is enhanced in cells lacking PopZ.

Because CtrA requires CpdR, RcdA, and PopA for degradation (Fig. 1D), we explored the need for each tier of the adaptor hierarchy by using substrates specific to each level. PdeA is a phosphodiesterase that only requires CpdR for degradation, while TacA is a response regulator that requires both CpdR and RcdA for ClpXP-mediated degradation (4, 15, 29). In order to determine which tier of adaptor-dependent proteolysis is compromised in ΔpopZ cells, we monitored the degradation of both PdeA and TacA. Degradation of both PdeA and TacA is enhanced in cells lacking PopZ, suggesting that, along with CtrA degradation, degradation of other adaptor-dependent ClpXP substrates is also stimulated in ΔpopZ cells (Fig. 3A and B). These results point to the conclusion that in cells lacking PopZ, the CpdR adaptor is more active.

FIG 3.

Degradation of CpdR- and RcdA-dependent substrates is enhanced in cells lacking PopZ. (A) Strains expressing M2-tagged PdeA were grown to exponential phase and induced with 0.3% xylose for 3 h before inhibiting protein translation by the addition of chloramphenicol. (C) TacA stability was monitored in wild-type and ΔpopZ cells. Lysates from equal volumes of cells were collected at indicated time points (see Materials and Methods). PdeA and TacA stability was monitored by Western blotting analyses by probing the blots with anti-M2 and anti-TacA antibodies. Asterisks denote cross-reacting contaminant. (B and D) Quantitation of Western blots. Bands corresponding to M2PdeA, TacA, and ClpP were quantified. M2-PdeA levels were normalized to ClpP, TacA levels were normalized to the cross-reacting contaminant. Data represent mean ± SD from two biological replicates.

CpdR is epistatic to PopZ in the CtrA degradation pathway.

The adaptor CpdR is required for stimulation of CtrA degradation both in vivo and in vitro (4, 13, 14). We reasoned that since CpdR appears to be constitutively active in ΔpopZ cells, deletion of CpdR in this background should result in loss of CtrA degradation. To test this hypothesis, we made a ΔcpdR ΔpopZ double knockout strain by transducing ΔcpdR:tet into the ΔpopZ strain using the ϕCr30 phage. As expected, CtrA degradation was stabilized in the ΔcpdR ΔpopZ double knockout background, similarly to that in the ΔcpdR strain (Fig. 4A and B). Microscopy experiments confirmed delocalization of CpdR-YFP in the ΔpopZ and ΔcpdR ΔpopZ cells (Fig. 4C) (21). CtrA degradation was restored in these ΔcpdR ΔpopZ strains expressing CpdR-YFP, suggesting that the loss of CtrA degradation in the ΔcpdR ΔpopZ strain could be attributed solely to loss of CpdR (Fig. 4B).

FIG 4.

CpdR is epistatic to PopZ in the CtrA degradation pathway. (A) CtrA stability following translational shutoff was monitored in wild-type, ΔpopZ, ΔcpdR, and ΔcpdR ΔpopZ cells and ΔcpdR ΔpopZ cells expressing CpdR-YFP from a plasmid. (B) Quantification of CtrA levels from triplicate biological experiments; error bars are standard deviations. (C) Phase-contrast and fluorescence microscopy images of wild-type, ΔpopZ, and ΔcpdR ΔpopZ cells expressing CpdR-YFP from a plasmid. Arrows indicate polar CpdR-YFP foci.

Overexpression of the PopZ protein stimulates CtrA degradation in a CpdR-independent manner.

Overexpression of PopZ leads to the enlargement of the polar region and overrecruitment of proteins such as CtrA, CpdR, RcdA, and ClpX to these polar zones (21). To determine if prolific polar recruitment affects substrate degradation, we monitored CtrA degradation in strains overexpressing PopZ. Because we had found that the loss of PopZ results in faster CtrA degradation, we were surprised to find that overexpression of PopZ also results in faster CtrA degradation compared to that in wild-type cells (Fig. 5A and B). This stimulation was specific to CtrA; PdeA degradation was not affected (Fig. 5A; see also Fig. S1 in the supplemental material). Overexpressing PopZ did not globally compromise protein degradation, as degradation of a ClpXP-independent substrate, DnaA, remained unaffected (Fig. 5A and S1).

FIG 5.

Overexpression of PopZ specifically stimulated CtrA degradation in wild-type and ΔcpdR cells. (A) Degradation of CtrA, PdeA, and DnaA was monitored in wild-type (WT) cells overexpressing PopZ from a high-copy-number xylose-inducible plasmid. Cultures were induced in exponential phase for 8 h using 0.3% xylose. After inhibiting protein synthesis by the addition of kanamycin, lysates from equal volumes of cells were collected at indicated time points and loaded onto SDS-PAGE gels. Asterisks denote cross-reacting bands. (B) Bands corresponding to CtrA and ClpP were quantified using ImageJ (NIH, USA), and normalized band intensities over time are shown. Data represent mean ± SD from three independent experiments. (C) Degradation of CtrA, PdeA, and DnaA was monitored in ΔcpdR cells or ΔcpdR cells overexpressing PopZ. (D) Bands corresponding to CtrA and ClpP were quantified, and normalized intensities are shown. Data represent mean ± SD from three independent experiments.

Why is CtrA degradation stimulated upon PopZ overexpression but that of PdeA is not? We considered this result in light of the fact that PdeA absolutely requires CpdR for degradation both in vivo and in vitro (15, 30), whereas ClpXP alone can degrade purified CtrA in vitro (31). Our working model is that overexpression of PopZ forces ClpX and CtrA to the poles, which could increase concentration or restrict diffusion sufficiently to directly drive recognition of CtrA by ClpXP. If this is true, then this overexpression should bypass the need for CpdR. Indeed, we found that CtrA, which is stable in ΔcpdR cells as expected, was degraded in ΔcpdR cells overproducing PopZ (Fig. 5C and D). Importantly, PdeA was stable in the absence of CpdR, regardless of PopZ overexpression. Together, these results suggest that forcing recruitment of the protease and the substrate by overexpressing PopZ is sufficient to bypass the need for adaptors when the substrate can be directly recognized by the protease.

DISCUSSION

In this work, we determined if polar localization of the ClpXP protease was critical for its activation in Caulobacter. Prior genetic experiments showed that the adaptors CpdR, RcdA, and PopA are necessary for normal CtrA degradation in vivo, and microscopy-based experiments showed that these proteins facilitated localization of the ClpXP protease and the CtrA substrate to the stalked pole, presumably to aid in degradation (13, 18, 19). Recent in vitro reconstitution experiments showed that all of these accessory factors work together as biochemical adaptors that enhance the affinity of CtrA for ClpXP (4, 14). Surprisingly, we found that CtrA is degraded more rapidly in cells lacking the PopZ protein than in wild-type cells. The fact that the degradation of other CpdR- and RcdA-dependent substrates is also enhanced in ΔpopZ cells supports a model where polar localization mediated by the PopZ protein might be critical for restraining adaptor-mediated protein degradation rather than driving degradation.

Why does degradation of ClpXP substrates increase in cells lacking PopZ? PopZ is needed for driving localization of a number of proteins to the stalked pole (Fig. 6). Given the epistatic relationship between CpdR and PopZ (Fig. 4), one possible explanation is that activity of the adaptor CpdR is elevated in the absence of PopZ. Indeed, levels of CpdR are increased in cells lacking PopZ (21). It is also possible that specific activity of CpdR is increased because of changes in its phosphorylation state. CckA is a membrane-bound bifunctional histidine kinase that phosphorylates both CtrA and CpdR via the phosphotransferase ChpT (10–12). When bound to cdG, CckA switches to a phosphatase-preferring state and dephosphorylates CtrA and CpdR (16, 17). Because increased local density of CckA at the stalked pole appears to drive its kinase activity (32) and PopZ is needed for CckA localization (23), a working model for our results is that loss of CckA localization forces it into a phosphatase state, resulting in constitutive dephosphorylation of CpdR and resulting activation of adaptor-dependent ClpXP degradation of substrates. Consistent with this model, cells expressing mutant CckA alleles with disrupted polar localization result in more rapid CtrA degradation (33).

FIG 6.

Cartoon of the stalked pole. Depiction of the stalked pole region during the transition to the stalked cell state. The ClpXP protease, CtrA substrate, and various adaptors are recruited to this pole at this time through PopZ. The recruitment of CckA results in phosphorylation of CpdR in swarmer cells, but CckA can also act as a phosphatase to dephosphorylate CpdR in stalked cells. Loss of PopZ results in increased CpdR activity, which could result from (i) less phosphorylated CpdR due to delocalized CckA, (ii) higher levels of CpdR, (iii) defects in the cell cycle, or (iv) other possibilities.

Although there are several mechanistic models to explain our findings, our major conclusion is that loss of subcellular localization of ClpXP, its adaptors, and its substrates does not result in substrate stabilization. In fact, we generally see an acceleration of ClpXP substrate degradation, which is particularly evident with CtrA. These results conflict with the original model for ClpXP-mediated protein degradation, wherein colocalization of substrates and protease were predicted to drive degradation (18). We favor a model where coordinated localization to the stalked pole (Fig. 6) creates a permissive environment for the tight coordination of adaptor activation (such as CckA-dependent dephosphorylation of CpdR) with cell cycle events. Imbalances in this coordination can misregulate protease activity, resulting in more rapid degradation in the absence of PopZ. However, this imbalance can also be driven by overexpression of PopZ, which also results in increased degradation of some ClpXP substrates. Thus, the tight coordination of stalked pole proteins has many consequences for protein degradation, but not simply through increasing local concentrations of protease and substrates to accelerate degradation.

MATERIALS AND METHODS

Bacterial culture conditions and plasmid construction.

Escherichia coli and Caulobacter strains and plasmids used in this study are listed in Table 1. E. coli strains were grown in LB medium at 37°C with the appropriate antibiotic (50 μg/ml kanamycin, 15 μg/ml tetracycline, or 30 μg/ml chloramphenicol). Caulobacter strains were grown in peptone-yeast extract (PYE) medium at 30°C with the appropriate antibiotic for plasmid maintenance, as needed (1 μg/ml tetracycline, 1 μg/ml chloramphenicol, or 5 μg/ml kanamycin). Construction of plasmids can be found in the cited references in the table, with the exception of pLXM-gfp∼CtrA15, which was generated by using the Gateway system as described previously (34). Briefly, this plasmid was generated by amplifying enhanced GFP (eGFP) with appropriate primers to append the last 15 residues of CtrA to the C terminus of eGFP. This product was cloned into a pENTR/D-TOPO plasmid using TOPO PCR cloning and then transferred to the pLXM plasmid by Gateway cloning (Invitrogen). This resulted in a xylose-inducible, N-terminally M2-tagged eGFP appended with the last 15 residues of CtrA. All plasmids were transformed into Caulobacter strains using electroporation (35). Microscopy was performed as described previously (4, 15). Phage transduction was performed using ϕCr30 (36).

TABLE 1.

Strains and plasmids used in this study

Plasmid or strain	Description	Source or reference
Strains
TOP10	E. coli cloning strain	Invitrogen
NA1000	Synchronizable Caulobacter crescentus	Laub Lab
CPC104	NA1000 ΔpopZ (Specr) (CJW2238)	23
CPC165	NA1000 ΔcpdR (Tetr)	34
CPC164	NA1000 ΔtacA (Tetr)	34
CPC204	NA1000 ΔcpdR ΔpopZ (Tetr Specr)	This study
CPC107	NA1000 pJS14 -PxylX-popZ (Cmr)	This study
CPC201	CPC204 pJS14-PxylX-popZ (Cmr)	This study
CPC313	NA1000 pLXM-pdeA (Tetr)	29
CPC314	CPC104 pLXM-pdeA (Tetr)	This study
CPC233	NA1000 pcpdR-cpdR-yfp (Kanr)	15
CPC234	CPC104 pcpdR-cpdR-yfp (Kanr)	This study
CPC235	CPC204 pcpdR-cpdR-yfp (Kanr)	This study
CPC118	CPC165 pJS14-PxylX-popZ (Cmr)	This study
CPC221	NA1000 pLXM-gfp∼ctrA15 (Kanr)	This study
CPC207	CPC104 pLXM-gfp∼ctrA15 (Kanr)	This study
Plasmids
pENTR/D-TOPO	Entry vector for Gateway cloning (Kanr)	Invitrogen
pJS14-xylX-PopZ	Medium-copy-no. vector for expression of PopZ (cmR)	23
pLXM-DEST(kan)	pMR10-PxylX-M2 destination vector; broad host range, low copy no., xylose inducible, N-terminal M2 tag (Kanr)	34
pcpdR-cpdR-yfp	Low-copy-no. plasmid expressing CpdR-YFP from the cpdR promoter (Kanr)	15
pLXM-gfp∼ctrA15	pLXM plasmid expressing M2-GFP appended with last 15 residues of CtrA (Kanr)	This study

In vivo protein degradation assays and Western blotting.

Overnight Caulobacter cultures in PYE containing the appropriate antibiotics were back diluted to an optical density at 600 nm (OD₆₀₀) of ∼0.1 and allowed to grow to an OD₆₀₀ of ∼0.3 in PYE medium, with appropriate antibiotics as needed. When needed, 0.2% xylose was added to induce tagged protein expression during this outgrowth. Protein synthesis was blocked by the addition of 50 μg/ml kanamycin or 30 μg/ml chloramphenicol. Following antibiotic addition, 1 ml of cells at an OD₆₀₀ of 0.3 was collected at indicated time points and centrifuged at 15,000 rpm for 1 min. Supernatants were removed and pellets were resuspended in 50 μl of 2× SDS sample buffer (12% glycerol, 4% SDS, 100 mM Tris HCl [pH 6.8], trace bromophenol blue, and 40 mM dithiothreitol [DTT]) and then snap-frozen in liquid nitrogen.

For Western blotting, thawed samples were boiled at 95°C for 10 min and centrifuged at 15,000 rpm for 10 min to remove cellular debris. After centrifugation, 10-μl samples of supernatant were resolved on 10% or 12% SDS-PAGE gels. Proteins from the gel were then transferred to a polyvinylidene difluoride (PVDF) membrane using semidry electrophoresis. After the membranes were blocked with TBST (Tris-buffered saline plus 0.1% Tween) containing 3% dry milk, they were probed overnight at 4°C with primary antibodies. Antibodies used were polyclonal rabbit anti-CtrA (1:5,000 dilution), anti-DnaA (1:10,000 dilution), anti-TacA (1:10,000 dilution), affinity-purified anti-PdeA (1:1,000 dilution), anti-ClpP (1:5,000 dilution), or monoclonal mouse anti-FLAG M2 (1:5,000 dilution; Sigma). For luminescence detection (all except Fig. 4), after washing off the excess primary antibody, the membranes were probed with goat anti-rabbit (Millipore, USA) or goat anti-mouse (Millipore, USA) secondary antibodies conjugated to horseradish peroxidase (HRP) enzyme. Proteins were visualized by the luminescence from the HRP substrate, using the chemiluminescence detection system G:Box (Syngene, UK). For infrared (IR) dye detection (Fig. 4), membranes were probed with goat anti-rabbit or anti-mouse secondary antibodies directly conjugated with IRDye 680 or 800 (Li-Cor Biosciences) and detected using an Odyssey CLx imager. In all cases, quantification of band intensities was performed using ImageJ software (NIH).