Trigger factor association with ribosomes decreases while DnaK association increases during cytoplasmic Mg²⁺ starvation

We used sedimentation over a sucrose cushion to examine the association of chaperones TF and DnaK with ribosomes in wild-type S. Typhimurium facing either cytoplasmic Mg²⁺ starvation (5 h in 10 μM Mg²⁺)—a condition that triggers expression of 2 Mg²⁺ importers and decreases ATP amounts [30,31]—or Mg²⁺ abundance (4.5 h in 10 mM Mg²⁺). The amount of TF bound to ribosomes was approximately 30 times lower in the former condition than in the latter condition (Fig 1B). By contrast, 3 times more DnaK was bound to ribosomes during cytoplasmic Mg²⁺ starvation than under Mg²⁺-abundant conditions (Fig 1B). (Similar results were obtained when ribosome association of TF and DnaK was investigated by polysome profiling over a sucrose gradient (S1A and S1B Fig), revealing that DnaK associates with both monosomes and polysomes.)

We determined that the molar ratio of ribosome-bound DnaK to ribosomes is 2.94 (±0.72) in bacteria experiencing cytoplasmic Mg²⁺ starvation but only 1.13 (±0.38) in bacteria experiencing Mg²⁺ abundance (mean ± SD). Control experiments showed no DnaK association with ribosomes that were fully dismantled due to treatment with the ion chelator ethylenediaminetetraacetic acid (S1C and S1D Fig).

To investigate how Mg²⁺ availability directly impacts chaperone binding to ribosomes, we incubated purified DnaK or TF with ribosomes from the PURExpress coupled transcription-translation system [32] (see later section) and diluted the reactions to 10 mM or 0.16 mM Mg²⁺. In agreement with findings obtained in vivo (Fig 1B), approximately 30 times more DnaK sedimented with ribosomes in low Mg²⁺ samples than in Mg²⁺-abundant samples (Fig 1C). By contrast, approximately 10 times less TF associated with ribosomes in low Mg²⁺ than in Mg²⁺-abundant samples (Fig 1D).

The Mg²⁺-regulated changes in the chaperone associated with ribosomes raised the question: What is the physiological consequence(s) of DnaK binding to ribosomes in lieu of TF?

DnaK reduces protein synthesis in bacteria facing low cytoplasmic Mg²⁺

To determine whether DnaK binding to ribosomes in place of TF alters protein synthesis in vivo, we compared the behavior of isogenic wild-type and dnaK mutant strains. To avoid the general growth defects (i.e., not specific to cytoplasmic Mg²⁺ starvation conditions) that result from deletion of the full dnaK coding region [33], we used a mutant isolated in our laboratory that harbors a transposon Tn10dCm insertion between nucleotides 1691 and 1692 of the 1917 nucleotide-long dnaK coding region. Designated dnaK14, this dnaK allele specifies a truncated DnaK protein (S2A Fig) that shares 100% amino acid identity in the first 564 residues with the 638 residue-long wild-type DnaK but lacks the 74 C-terminal residues because of a stop codon originating from the transposon Tn10dCm in frame with the dnaK coding region (S2B Fig). The truncated DnaK retains the full nucleotide-binding domain (NBD) and most of the substrate-binding domain (SBD) but is missing the 42 C-terminal residues within the a-helical lid subdomain of the SBD and the C-terminal 32 residues of the protein, which are predicted to be disordered and whose function is not well understood [34] (S2C Fig). Because the dnaK14 mutant grew like wild-type S. Typhimurium in LB broth and LB agar media, it was used for the in vivo experiments described below.

Pulse labeling of nascent proteins with ³⁵S-methionine followed by quantification of the radioactivity in the protein precipitate revealed that the dnaK14 mutant exhibited approximately 3-fold higher ³⁵S-methionine incorporation than the wild-type strain (Fig 2A) in bacteria experiencing cytoplasmic Mg²⁺ starvation, indicating that the dnaK14 mutant synthesized more protein than the wild-type strain. This novel phenotype is specific to cytoplasmic Mg²⁺ starvation conditions because wild-type and dnaK14 strains exhibited similar ³⁵S-methionine incorporation when Mg²⁺ was abundant (Fig 2B). Independent support for these findings was obtained by labeling nascent proteins with the click chemistry-compatible methionine analog L-azidohomoalanine: the dnaK14 mutant incorporated more methionine analog than wild-type S. Typhimurium in low Mg²⁺ (Fig 2C), but not when Mg²⁺ was abundant (Fig 2D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. DnaK inhibits protein synthesis in vivo and in vitro.

(A, B) ³⁵S-methionine labeling of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 5 h of growth in low (10 μM) Mg²⁺ (A) or 4 h of growth in high (10 mM) Mg²⁺ (B). (C, D) L-azidohomoalanine (AHA) labeling of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 5.5 h of growth in low (10 μM) Mg²⁺ (C) or 4.5 h of growth in high (10 mM) Mg²⁺(D). (E, F) ³⁵S-methionine labeling of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q) or the wild-type dnaK-expressing plasmid (pDnaK) following 5 h of growth in low (10 μM) Mg²⁺ (E) or of wild-type (14028s), dnaK14 (CC186), and dnaJ cbpA djlA (CC656) S. Typhimurium following 5 h of growth in low (10 μM) Mg²⁺ (F). (G) Western blot analysis of clarified cell lysates (left) and ribosome-bound fractions (right) of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 5 h of growth in low (10 μM) Mg²⁺ using antibodies recognizing DnaK or the ribosomal protein control S1. (H) Western blot analysis of purified DnaK protein or truncated DnaK (1–563) (5 μM) bound to translating ribosomes in vitro. Blot was developed with antibodies directed to the DnaK protein (ribosome-bound DnaK) or stained for total protein (supernatant fractions and ribosome-bound S1 protein; band above S1 represents the DnaK protein). (I) Western blot analysis of in vitro synthesized SsrB-FLAG and PmrA-FLAG proteins in the presence or absence of purified DnaK protein (5 μM). Blots were developed with antibodies directed to the FLAG tag or the DnaK protein. (J) Western blot analysis of in vitro synthesized PmrA-FLAG protein in the presence or absence of purified DnaK (5 μM) or purified HtpG (5 μM) proteins. (K) Western blot analysis of in vitro synthesized PmrA-FLAG protein in the presence or absence of purified DnaK (5 μM) using linear DNA or mRNA templates. Data in (A–B and E–F) represent mean ± SD of at least 3 independent biological replicates. Shown in (C and D) are the representatives of 3 independent biological replicates. Shown in (G) is the representative of 2 independent biological replicates. Total protein amounts in samples were quantified to verify equivalent sample loading. Numerical values represent mean ± SD. Shown in (H) is the representative of 4 assays. Numerical values represent mean ± SD. Shown in (I–K) are the representatives of 2 to 4 assays performed in buffer containing 9 mM Mg²⁺. Statistical analysis was performed using two-tailed Student’s t test comparing the indicated sample group to the wild-type sample group or comparing the bracketed sample groups (ns = not significant). MW = molecular weight. The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g002

The increased protein synthesis exhibited by the dnaK14 mutant is due to the disruption of the dnaK gene rather than polarity on the downstream dnaJ gene [35,36] because plasmid pDnaK, which expresses a wild-type copy of the dnaK gene from a heterologous promoter, largely complemented the mutant, whereas the plasmid vector control did not (Fig 2E). Moreover, a S. Typhimurium triple mutant lacking the cbpA, djlA, and dnaJ genes, which encode the 3 J-domain cochaperones of DnaK characterized in the closely related species Escherichia coli [37], exhibited similar ³⁵S-methionine incorporation to the wild-type strain (Fig 2F), unlike the increased ³⁵S-methionine incorporation of the dnaK14 mutant (Fig 2A, 2E and 2F).

DnaK association with ribosomes (Fig 1B and 1C) appears to be responsible for the decreased protein synthesis exhibited by wild-type S. Typhimurium experiencing cytoplasmic Mg²⁺ starvation (Fig 2A–2D) because the truncated DnaK produced by the dnaK14 strain did not associate with ribosomes (Fig 2G) in vivo. Furthermore, purified truncated DnaK (consisting of the first 563 amino acids specified by native codons) bound translating ribosomes from the PURExpress coupled transcription-translation system [32] (see next section) approximately 10 times less than purified full-length DnaK (Fig 2H).

Because the remaining C-terminal tail portion of some truncated forms of DnaK can interact with its own SBD [38], we explored whether the lack of association between DnaK(1–563) and ribosomes was caused by the truncated DnaK occupying its own SBD binding pocket. Control experiments established that the basal in vitro ATPase activity (without J-domain cochaperones or nucleotide exchange factor) of purified truncated DnaK is similar to that of full-length DnaK (S2D Fig). These results indicate that the C-terminal portion remaining in the truncated DnaK does not mimic a peptide substrate and that truncated DnaK does not exist in a tail-bound state [38], arguing against the possibility that truncated DnaK fails to bind ribosomes due to a self-occupied substrate binding site. (Note that the ATPase activity of the truncated DnaK was stimulated less by J-domain cochaperones and nucleotide exchange factor than the ATPase activity of the full-length wild-type DnaK [S2D Fig].)

Cumulatively, the results in this section show that S. Typhimurium responds to cytoplasmic Mg²⁺ starvation by increasing DnaK association with ribosomes, which reduces protein synthesis.

Purified DnaK reduces protein synthesis in vitro

To establish whether DnaK is directly responsible for the reduction in protein synthesis observed in vivo (Fig 2A), we examined the effect of the purified DnaK protein on protein synthesis in vitro. We used the PURExpress in vitro protein synthesis system [32], which consists of His-tagged proteins necessary for coupled transcription/translation, untagged ribosomes, no added chaperones, and 9 mM Mg²⁺ [32], which is necessary for transcription and translation to take place, and verified that DnaK preparations lacked contaminating proteins (S3A Fig), DNase activity (S3B and S3C Fig), or RNase activity (S3D and S3E Fig). Reactions were programmed with DNA templates specifying C-terminally FLAG-tagged reporter proteins, selected because they are expressed by S. Typhimurium during low Mg²⁺ [39,40]. Purified DnaK decreased the amounts of synthesized proteins up to 5-fold relative to controls consisting of the buffer used to resuspend DnaK (Fig 2I) or the protein chaperone HtpG (Fig 2J), known as the bacterial Hsp90 [41]. The yield of synthesized proteins decreased correspondingly with increasing DnaK concentrations (S3F Fig).

DnaK also decreased the amounts of synthesized proteins when mRNA was used instead of DNA (Fig 2K), indicating that DnaK inhibits translation and arguing against the possibility that the phenotypes obtained with DNA templates resulted from DnaK inhibiting gene transcription of the coupled transcription/translation PURExpress system [32]. That the DnaK preparations appear to be free of DnaJ (S3A Fig) and that the truncated DnaK retains DnaJ’s binding domain [42] (S2A–S2C Fig) argue that DnaK’s inhibition of protein synthesis does not require cochaperone DnaJ.

Trigger factor interferes with DnaK’s reduction of protein synthesis

When the protein synthesis machinery was incubated with purified TF prior to addition of DnaK, the yield of the in vitro protein synthesis reaction was higher than when DnaK was added alone (Fig 3A). By contrast, TF had no effect when DnaK was added to the reaction prior to TF (Fig 3B). Though these results suggest the possibility of DnaK and TF sharing binding sites in the translation machinery, an alternative interpretation is that binding of one of these chaperones to the translation machinery results in conformational changes that hinder the ability of the other chaperone to reduce protein synthesis. Control experiments demonstrated that TF neither increases nor decreases protein synthesis when added alone (i.e., in the absence of DnaK) to the PURExpress system (Fig 3A and 3B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Trigger factor hinders DnaK-mediated reduction of protein synthesis.

(A) Western blot analysis of in vitro synthesized PmrA-FLAG protein in the absence or presence of purified DnaK, purified TF, or purified TF added before purified DnaK addition (5 μM each protein). Blot was developed with antibodies directed to the FLAG tag. Total protein staining was performed to visualize translation factor EF-Tu as a loading control. (B) Western blot analysis of in vitro synthesized PmrA-FLAG protein in the absence or presence of purified DnaK, purified TF, or purified DnaK added before purified TF (5 μM each protein). Blot was developed with antibodies directed to the FLAG tag. Total protein staining was performed to visualize translation factor EF-Tu as a loading control. (C) ³⁵S-methionine labeling of wild-type (14028s) dnaK14 (CC186) and tig (CC361) S. Typhimurium following 5 h of growth in low (10 μM) Mg²⁺. (D) ³⁵S-methionine labeling of wild-type (14028s) dnaK14 (CC186), tig (CC361), and dnaK tig (CC362) S. Typhimurium following 6 h of growth in low (10 μM) Mg²⁺ at 30°C. Shown in (A and B) are the representatives of 2 to 3 assays performed in buffer containing 9 mM Mg²⁺. MW = molecular weight. Data in (C and D) represent mean ± SD of at least 3 independent biological replicates. Statistical analysis was performed using two-tailed Student’s t test comparing the indicated sample group to the wild-type sample group or comparing the bracketed sample groups (ns = not significant). The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g003

A mutant lacking the tig coding region displayed wild-type protein synthesis during cytoplasmic Mg²⁺ starvation (Fig 3C), in agreement with the in vitro results discussed above (Fig 3A and 3B). By contrast, when the tig deletion was combined with the dnaK14 mutation, the resulting double mutant had the same elevated protein synthesis as the dnaK14 single mutant (Fig 3D). (The latter experiment was performed at 30°C because a tig dnaK mutant is not viable at higher temperatures [21].) Cumulatively, these results likely reflect that little TF associates with ribosomes in bacteria experiencing cytoplasmic Mg²⁺ starvation (Fig 1B and 1D) to interfere with the reduction of protein synthesis mediated by DnaK.

DnaK’s ATPase activity is stimulated by ribosomes and required for reduction of protein synthesis

DnaK requires ATP hydrolysis for substrate binding and release (reviewed in [43,44]). J-domain cochaperones stimulate DnaK’s ATPase activity [7]. We have now determined that a purified variant of full-length DnaK unable to hydrolyze ATP [45,46]—DnaK(T199A)—failed to reduce protein synthesis in vitro, unlike wild-type DnaK (Fig 4A). In vivo, a DnaK(T199A)-expressing plasmid failed to fully correct the abnormally high protein synthesis of the dnaK14 mutant, behaving like the vector control and unlike the isogenic plasmid pDnaK (Fig 4B). Thus, DnaK’s ability to reduce protein synthesis requires: its ability both to associate with ribosomes, which is dependent on its C-terminal 74 residues (Fig 2G and 2H), and to hydrolyze ATP (Fig 4A), which is conferred by the N-terminal domain of the protein [47].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. DnaK’s ATPase activity is stimulated by ribosomes and required to reduce protein synthesis.

(A) Western blot analysis of in vitro synthesized PmrA-FLAG protein in the presence or absence of purified DnaK or DnaK(T199A) (5 μM). Blot was developed with antibodies directed to the FLAG tag. Total protein staining was performed to visualize translation factor EF-Tu as a loading control. (B) ³⁵S-methionine labeling of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q), wild-type dnaK-expressing plasmid (pDnaK), or mutant dnaK(T199A)-expressing plasmid (pDnaKT199A) following 5 h of growth in low (10 μM) Mg²⁺. (C–E) ATP hydrolysis in the presence or absence of purified DnaK (2 μM) alone or in combination with cochaperones (DnaJ [0.4 μM] or CbpA [0.4 μM] and GrpE [0.2 μM]) or 70S ribosomes (0.5 μM) (C), in the presence or absence of purified DnaK(T199A) (2 μM) alone or in combination with 70S ribosomes (0.5 μM) (D) or in the presence or absence of purified DnaK (2 μM) alone or in combination with 30S ribosomal subunit (0.5 μM), 50S ribosomal subunit (0.5 μM), or 70S ribosomes (0.5 μM) (E). Shown in (A) is the representative of 3 assays. MW = molecular weight. Data in (B) represents mean ± SD of 3 independent biological replicates. Data in (C–E) represent mean ± SD of 2 to 4 independent assays performed in buffer containing 20 mM Mg²⁺. Absorbance at 360 nm reflects the production of 2-amino-6-mercapto-7-methylpurine stimulated by the presence of inorganic Pi, the product of the ATPase reaction. Statistical analysis in (B) was performed using two-tailed Student’s t test comparing the indicated sample group to the wild-type sample group or comparing the bracketed sample groups (ns = not significant). The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g004

We established that purified bacterial ribosomes (New England Biolabs) stimulate DnaK’s ATPase activity in the absence of J-domain cochaperones (Fig 4C), albeit approximately 45% less than J-domain cochaperones and GrpE (Fig 4C). ATP hydrolysis was not detected when ribosomes were incubated with the catalytic-defective mutant DnaK(T199A) [45] (Fig 4D), arguing against the possibility of wild-type DnaK stimulating a hypothetical ATPase contaminating the ribosome preparation. The large (50S) subunit of the ribosome, which harbors TF’s binding site [18], stimulated ATP hydrolysis as much as intact 70S ribosomes (Fig 4E), whereas the small (30S) subunit had no effect on DnaK’s ATPase activity (Fig 4E). (These assays were conducted in buffer containing 20 mM Mg²⁺.)

Control experiments indicate that the ribosome-stimulated ATP hydrolysis by DnaK is not due to sample contamination with a J-domain cochaperone because DnaK alone failed to refold heat-denatured luciferase in the presence or absence of ribosomes (S4 Fig), behaving like the no protein control (S4 Fig), and unlike the positive control consisting of DnaK, J-domain cochaperone CbpA, and GrpE, which efficiently refolded luciferase (S4 Fig).

Taken together, the results in this section demonstrate that ribosomes stimulate ATP hydrolysis by DnaK; that this stimulation is independent of J-domain chaperones; and that DnaK’s ATPase activity is required for reduction of protein synthesis.

DnaK prevents protein aggregation during cytoplasmic Mg²⁺ starvation

Cells must adjust their rate of protein synthesis to their protein folding capacity or face the toxic effects of accumulating misfolded protein aggregates [48–50]. We reasoned that DnaK coordinates these 2 activities when bacteria face cytoplasmic Mg²⁺ starvation because this stressor limits protein homeostatic capacity by decreasing the concentration of ATP [51], thereby reducing ATP-dependent proteolysis [25], ATP-dependent chaperoning [26], and ATP-dependent protein solubilization [27]. Moreover, DnaK reduces protein synthesis in vitro (Fig 2I) and during cytoplasmic Mg²⁺ starvation in vivo (Fig 2A and 2B). These findings suggested that DnaK helps maintain protein homeostasis in bacteria facing cytoplasmic Mg²⁺ starvation.

As hypothesized, the dnaK14 mutant had more protein aggregates than the wild-type strain following 6 h in low Mg²⁺ medium (Fig 5A), when bacteria face cytoplasmic Mg²⁺ starvation [51,52]. By contrast, protein aggregate amounts in Mg²⁺-abundant conditions were not elevated in the dnaK14 mutant (Fig 5B). Moreover, the fraction of newly synthesized proteins (determined by ³⁵S-methionine-labeling) present in the insoluble fraction of bacteria facing cytoplasmic Mg²⁺ starvation was twice as high in the dnaK14 mutant as in the wild-type strain (Fig 5C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. S. Typhimurium coordinates protein synthesis with protein folding capacity.

(A, B) Quantification of total, soluble, and insoluble protein amounts in wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 6 h of growth in low (10 μM) Mg²⁺ (A) or 4.5 h of growth in high (10 mM) Mg²⁺ (B). (C) Percent of ³⁵S-methionine-labeled proteins in soluble and insoluble fractions isolated from wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 5 h of growth in low (10 μM) Mg²⁺. (D) Percent of ³⁵S-methionine-labeled proteins remaining in wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 18 h of treatment with 62.5 μg/ml tetracycline to halt translation. Values were calculated as the amount of ³⁵S-methionine signal remaining divided by the ³⁵S-methionine signal of a sample collected immediately prior to tetracycline addition. Data in (A) and (B) represent mean ± SD of at least 2 independent biological replicates. Data in (C) and (D) represent mean ± SD of 3 independent biological replicates. Statistical analysis was performed using two-tailed Student’s t test comparing the bracketed sample groups (ns = not significant). The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g005

We then monitored the fate of ³⁵S-methionine-labeled proteins in bacteria grown in low (10 μM) Mg²⁺ medium before and after treatment with the protein synthesis inhibitor tetracycline (62.5 μg/ml). The amounts of radiolabeled proteins remained stable in the wild-type strain 18 h after tetracycline treatment but decreased by approximately 60% in the dnaK mutant relative to starting values (Fig 5D), indicating decreased global protein stability.

These results indicate that DnaK’s reduction in protein synthesis during cytoplasmic Mg²⁺ starvation aids protein homeostasis by preventing protein synthesis from exceeding protein folding capacity.

DnaK is necessary for survival against cytoplasmic Mg²⁺ starvation and hyperosmotic stress

DnaK is required for survival against cytoplasmic Mg²⁺ starvation because the dnaK14 mutant was approximately 10,000-fold less viable than wild-type S. Typhimurium following a 24 h incubation in low (10 μM) Mg²⁺ medium (Fig 6A). This survival defect is specific to low Mg²⁺ because the mutant exhibited wild-type survival in medium containing high (10 mM) Mg²⁺ (Fig 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. The dnaK gene is required for bacterial survival against Mg²⁺ starvation.

(A, B) Survival of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 24 h in low (10 μM) Mg²⁺ (A) or 24 h in high (10 mM) Mg²⁺ (B). (C) Survival of wild-type (14028s), dnaK14 (CC186), and dnaJ (EG16309) S. Typhimurium following 24 h in low (10 μM) Mg²⁺. (D) Survival of wild-type (14028s), dnaK14 (CC186), and dnaJ cbpA djlA (CC656) S. Typhimurium following 24 h in low (10 μM) Mg²⁺. (E–H) Survival of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q) or the dnaK-expressing plasmid (pDnaK) (E), of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q), wild-type dnaK-expressing plasmid (pDnaK), or dnaK(nucleotides 1–1691)-expressing plasmid following 24 h in low (10 μM) Mg²⁺ (F), of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q), wild-type dnaK-expressing plasmid (pDnaK), or dnaK(T199A)-expressing plasmid (pDnaKT199A) (G), or of wild-type (14028s), dnaK14 (CC186), and tig (CC361) S. Typhimurium (H) following 24 h in low (10 μM) Mg²⁺. (I) Survival of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium following 24 h in hyperosmotic (1 M NaCl) conditions. Data in (A–I) represent mean ± SD of at least 3 independent biological replicates. Statistical analysis was performed using two-tailed Student’s t test comparing the indicated sample group to the wild-type sample group or comparing the bracketed sample groups (ns = not significant). The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g006

The survival defect of the dnaK14 mutant is not due to polar effects on dnaJ because: first, a dnaJ mutant exhibited wild-type survival in low Mg²⁺ conditions (Fig 6C), as did the dnaJ cbpA djlA triple mutant (Fig 6D); and second, plasmid pDnaK rescued the dnaK14 mutant (Fig 6E). By contrast, neither the plasmid vector control nor plasmid pDnaK(1–563) carrying the native codons corresponding to the first 563 residues of DnaK, thus specifying the truncated DnaK, rescued the dnaK14 mutant for survival in low Mg²⁺ (Fig 6E and 6F). (Wild-type S. Typhimurium exhibited similar survival when harboring pDnaK, the vector control, or pDnaK(1–563), indicating that the truncated DnaK protein is not dominant negative on wild-type DnaK [Fig 6E]. In addition, wild-type and dnaK14 mutant S. Typhimurium harboring the plasmid vector or plasmid pDnaK exhibited similar survival during growth in high Mg²⁺. [S5A Fig])

DnaK’s ability to hydrolyze ATP is required for bacterial survival against cytoplasmic Mg²⁺ starvation because plasmid pDnaK(T199A) failed to restore wild-type survival to the dnaK14 mutant, behaving like the vector control and unlike pDnaK (Fig 6G). While DnaK is required for bacterial survival against cytoplasmic Mg²⁺ starvation (Fig 6A), TF is dispensable because the tig mutant behaved like the wild-type strain (Fig 6H). In agreement with the results of others [21], TF was also not required for growth in nutrient abundant conditions (S5B Fig).

Hyperosmotic stress triggers transient cytoplasmic Mg²⁺ starvation even when Mg²⁺ in the environment is plentiful [53]. This is because bacteria respond to hyperosmotic stress by importing large amounts of the osmoprotective ion K⁺ [54], which is countered by Mg²⁺ efflux from the cytoplasm. We therefore hypothesized that a functional dnaK gene would be required to survive hyperosmotic stress. As expected, viability of the dnaK14 mutant was approximately 100-fold lower than that of the wild-type strain following 1 day incubation in medium containing 1 M NaCl and 1 mM Mg²⁺ (which corresponds to the physiological Mg²⁺ concentration in the bacterial cytoplasm) [55] (Fig 6I).

Given the increased protein synthesis of the dnaK14 mutant (Fig 2A–2D) and wild-type protein synthesis of the tig mutant (Fig 3C and 3D), the results in this section suggest that bacterial survival against cytoplasmic Mg²⁺ starvation requires DnaK’s ability to reduce protein synthesis.

DnaK confers survival against cytoplasmic Mg²⁺ starvation by decreasing protein synthesis

If DnaK confers survival against cytoplasmic Mg²⁺ starvation by decreasing protein synthesis, a dnaK-independent decrease in protein synthesis should rescue the dnaK14 mutant. The results of 3 independent experimental approaches support this notion. First, we used plasmid pMgtC, which harbors a wild-type copy of the mgtC gene transcribed from a heterologous promoter. mgtC expression lowers ATP amounts and promotes accumulation of the second messenger (p)ppGpp (guanosine tetraphosphate), which decreases rRNA synthesis, reduces the number of ribosomes, and results in lower overall cellular protein synthesis [51]. When introduced into the dnaK14 mutant, plasmid pMgtC, but not the plasmid vector, decreased protein synthesis (S6 Fig). Remarkably, pMgtC restored wild-type viability to the dnaK14 mutant facing cytoplasmic Mg²⁺ starvation (Fig 7A), whereas the plasmid vector control did not (Fig 7A).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. A reduction in protein synthesis furthers bacterial survival against Mg²⁺ starvation.

(A) Survival of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium harboring the plasmid vector (pUHE-21-2-lacI^q) or the mgtC-expressing plasmid (pMgtC) following 24 h of growth in low (10 μM) Mg²⁺. (B) Survival of wild-type (14028s), dnaK14 (CC186), mgtC (EL1), and dnaK mgtC (CC241) S. Typhimurium following 24 h of growth in low (10 μM) Mg²⁺. (C) Survival of wild-type (14028s) and dnaK14 (CC186) S. Typhimurium in the presence or absence of sublethal amounts of the protein synthesis inhibitors tetracycline or spectinomycin or the RNA polymerase inhibitor rifamycin following 24 h of growth in low (10 μM) Mg²⁺. Data in (A and B) represent mean ± SD of at least 3 independent biological replicates. Statistical analysis was performed using two-tailed Student’s t test comparing the indicated sample group to the wild-type sample group or comparing the bracketed sample groups (ns = not significant). Shown in (C) are the representatives of 3 assays. The numerical values underlying this figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3002560.g007

Second, inactivation of the mgtC gene in the dnaK14 mutant background further decreased survival against cytoplasmic Mg²⁺ starvation (Fig 7B). This result likely reflects that MgtC and DnaK reduce protein synthesis by 2 distinct mechanisms: decreasing ribosome amounts [51] and activity (Figs 2I–2K, 3A, 3B, and 4A), respectively. And third, the bacteriostatic protein synthesis inhibitors tetracycline and spectinomycin increased survival of the dnaK14 mutant when added at subinhibitory concentrations (Fig 7C). By contrast, a subinhibitory concentration of the RNA polymerase inhibitor rifamycin failed to rescue the mutant (Fig 7C), suggesting that survival is specifically restored by protein synthesis inhibitors.

Bacterial strains and growth conditions

Bacterial strains, plasmids, and oligonucleotides used in this study are presented in S1 and S2 Tables. S. enterica serovar Typhimurium strains are derived from wild-type strain 14028s. Strains were constructed using λ Red-mediated recombination [77] and P22-mediated phage transduction. To minimize the appearance of spontaneous mutations in strains harboring mutations in the dnaK, dnaJ, cbpA, djlA, or tig genes, the mutations (genetically marked with an antibiotic resistance marker) were regularly transduced into a fresh wild-type 14028s background and propagated at 30°C during strain construction.

S. Typhimurium strains were grown in N-minimal medium (pH 7.7) supplemented with 0.1% casamino acids, 38 mM glycerol, and the indicated concentrations of MgCl₂. Unless otherwise indicated, bacteria were grown at 37°C in a water bath with 250 rpm shaking.

Plasmids were generated by Gibson assembly, confirmed by Sanger sequencing, and introduced by electroporation into S. Typhimurium strains. IPTG concentrations used for plasmid induction in physiology experiments were 100 μM IPTG for pDnaK, pDnaK(T199A), and pDnaK(1–563), and 250 μM for pMgtC. Ampicillin for plasmid maintenance was used at 50 μg/ml in S. Typhimurium and 100 μg/ml in E. coli.

Investigating in vivo association of DnaK and TF with ribosomes

Bacterial cultures were grown overnight in 10 mM Mg²⁺, washed 3 times in medium lacking Mg²⁺, and subcultured 1:50 in 200 ml of N-minimal medium containing 10 μM or 10 mM Mg²⁺ at 37°C with 250 rpm shaking. At the indicated times, the cultures were rapidly chilled by pouring into polypropylene-copolymer centrifuge bottles (Nalgene) containing ice slurry. Cells were pelleted for 20 min (12,000 rpm, 4°C) in a Evolution RC Superspeed centrifuge (Sorvall) using a SLA-3000 fixed angle rotor (Sorvall), washed 1 time with 25 ml wash buffer (20 mM Tris (pH 7.5), 100 mM NaCl, 10 mM MgCl₂), and resuspended in 1 ml of ice-cold Buffer A (20 mM Tris (pH 7.5), 300 mM NH₄Cl, 10 mM MgCl₂, 0.5 mM EDTA, 6 mM β-mercaptoethanol) supplemented with 0.4 mg/ml lysozyme, 10 U/ml DNase I (New England Biolabs), and 10 U/ml SUPERase In RNAse inhibitor (Thermo Fisher Scientific). Following incubation on ice for 30 min, cells were further lysed by sonication at 4°C. The lysate was centrifuged at 4°C at 30,000×g for 30 min to obtain the clarified S30 fraction. The absorbance at 260 nm of the S30 fraction was measured with a Nanodrop using a 1:100 dilution. An equivalent of 60 A260 units in a total volume of 4 ml was layered on a 4-ml 38% sucrose cushion made with Tight-Couples Buffer (20 mM Tris-HCl (pH 7.5), 50 mM MgOAc, 100 mM NH₄Cl, 1 mm DTT, 0.5 mM EDTA). The preparation was ultracentrifuged for 16 h (37,500 rpm, 4°C) using a 90 Ti rotor (Beckman Coulter). Following ultracentrifugation, the supernatant was removed, and the ribosomal pellet was resuspended in 100 μl of Tight-Couples Buffer and shaken into solution overnight at 4°C. The absorbance at 260 nm of the ribosomal fraction was measured using a Nanodrop instrument (NanoDrop Machines).

To confirm the amount of input protein present in clarified cell lysates, an aliquot of each S30 extract was normalized for equivalent RNA content (5.25 mg/ml RNA = 5 μl extract used) and probed by western blot. To determine the presence of ribosome-associated DnaK and TF in ribosomal fractions, an aliquot of each fraction was normalized for equivalent RNA content (20 mg/ml RNA = 15 μl fraction used), brought up to 15 μl in Tight-Couples Buffer, and combined with 5 μl of 4× Laemmli sample buffer. The samples were analyzed by western blotting using anti-DnaK, anti-TF, and anti-S1 antibodies. Total ribosomal protein staining was performed with SYPRO Ruby Protein Gel Stain (Thermo Fisher Scientific).

Clarified lysates for polysome profiling were obtained as described above. An equivalent of 60 A260 units in a total volume of 400 μl was layered on 10 ml 10% to 40% linear sucrose gradients made with Buffer A and ultracentrifuged for 3 h and 15 min (40,000 rpm, 4°C) using an SW40 Ti rotor (Beckman Coulter). Following ultracentrifugation, the tubes were mounted onto a density gradient fractionator (Brandel) and fractionation was performed using a Minipuls peristaltic pump (Gilson). The absorbance at 260 nm of each fraction 300 μl fraction was measured using a Nanodrop instrument (NanoDrop Machines) and plotted to obtain the in vivo polysome profile.

To determine the presence of ribosome-associated DnaK in sucrose fractions, 16.5 μl of each fraction was combined with 5.5 μl of 4× Laemmli sample buffer. The samples were analyzed by western blotting using anti-DnaK and anti-TF antibodies.

Dismantling of ribosomes was performed by sedimenting the lysate through a sucrose cushion or gradient supplemented with 10 mM EDTA.

The DnaK to ribosome ratio was estimated as follows: First, the number of molecules of DnaK in samples was compared to a standard curve of purified protein. Second, the number of ribosomes in the same samples was estimated based on RNA concentration measured by Nanodrop (Nanodrop Machines). Approximately 85% of total RNA was assumed to be ribosomal RNA (rRNA) [78]. Each ribosome was assumed to be 2.7 mDa in mass [79], and 60% of ribosomal mass was assumed to be made up of rRNA [79].

Investigating in vivo protein synthesis

For ³⁵S-methionine labeling, cells were grown overnight in 10 mM Mg²⁺, washed 3 times in medium lacking Mg²⁺, and subcultured 1:50 in 2 ml of N-minimal medium with 10 μM or 10 mM Mg²⁺. At the indicated times (5 h for low Mg²⁺ and 4 h for high Mg²⁺), 1 ml of the culture was transferred to a new tube for labeling. Each 1 ml of culture was labeled with 4.33 μCi of ³⁵S-methionine (PerkinElmer). Following 1 h of ³⁵S-methionine incorporation, 1 ml aliquots of each culture were added to microcentrifuge tubes pre-aliquoted with 200 μl ice-cold 50% trichloroacetic acid (TCA). (This moderately longer pulse labeling time was used to eliminate the need for growing bacteria in medium lacking methionine, which is necessary for short labeling times and would have caused cells to experience both Mg²⁺ starvation and methionine starvation as opposed to solely Mg²⁺ starvation as achieved in this study.) Proteins were precipitated overnight at 4°C.

Precipitated protein mixtures were collected by vacuum filtration on 25 mm glass microfiber filters, grade GF/C (Whatman) prewashed with 5 ml ice-cold 5% TCA. Filters were washed 2 times with 5 ml ice-cold 5% TCA and 1 time with 5 ml ice-cold ethanol. Filters were then removed and dried for at least 2 h at room temperature. Dried filters were soaked in 5 ml Ultima Gold scintillation fluid (PerkinElmer) for 15 min. Signal from each filter was measured using an LS 6500 liquid scintillation counter (Beckman Coulter). Scintillation values from a blank filter, which harbored minimal scintillation signal, were subtracted from sample readings. An unlabeled control with bacterial culture not pulse labeled with ³⁵S-methionine was included in every experiment.

L-AHA labeling was performed as described [51]. Briefly, cells were grown in 2 ml N-minimal medium lacking casamino acids and instead supplemented with all individual amino acids except methionine. After 5.5 h growth, 2 μl of 40 mM L-AHA (Click Chemistry Tools) dissolved in DMSO was added to 2 ml cultures. Cultures were returned to the water bath and L-AHA incorporation was allowed to proceed for 30 min. Subsequently, cultures were rapidly chilled on ice, pelleted, and washed 3 times with an equivalent volume of ice-cold phosphate-buffered saline (PBS). Cell pellets were frozen and stored at −80°C. The pellets were thawed, resuspended in a buffer of 50 mM Tris (pH 8.0) with 0.5% sodium dodecyl sulfate (SDS), and lysed by sonication. Lysates were centrifuged for 10 min (30,000×g, 4°C) to separate insoluble components. The soluble lysate was transferred to a clean microcentrifuge tube and covalent attachment of tetramethylrhodamine azide (TAMRA) (Click Chemistry Tools) was carried out using the Click-&-Go Protein Reaction Buffer Kit (Click Chemistry Tools). Total protein was isolated by methanol–chloroform extraction and the protein pellet was washed 3 times with 450 μl methanol to remove unattached TAMRA. The washed pellet was resuspended in 200 μl of Laemmli sample buffer lacking β-mercaptoethanol. The total protein content of the solution was measured using a Micro BCA Protein Assay Kit (Thermo Fisher Scientific) and 7.5 μg of total protein per sample was electrophoresed on SDS-PAGE gels. Nascent protein amounts were analyzed by western blotting using anti-TAMRA antibodies (Novus Biologicals).

To investigate the presence of ³⁵S-methionine-labeled proteins in insoluble and soluble cellular fractions, bacteria were grown in 2 ml low (10 μM) Mg²⁺ medium for 5 h and labeled with ³⁵S-methionine for 1 h, and 1 ml aliquots of each culture were added to microcentrifuge tubes pre-aliquoted with 200 μl ice-cold 50% TCA. Following 10 min of incubation at 4°C, the samples were centrifuged at low speed (1,000×g, 4°C) to separate insoluble inclusion bodies. The supernatant was transferred to a fresh microcentrifuge tube, further acidified by addition of 200 μl of ice-cold 50% TCA, and incubated overnight at 4°C. The insoluble pellets were washed 1 time with 100 μl of PBS with 0.01% Triton to remove membrane proteins and resuspended in 750 μl PBS prior to loading onto filters. Proteins were collected on filters and radioactivity was quantified by scintillation counting as described above.

To investigate global protein stability, bacteria were grown for 5 h in low (10 μM) Mg²⁺ medium and labeled with ³⁵S-methionine for 1 h. A 0.5 ml sample (denoted as pre-tetracycline) was collected, combined with 100 μl of ice-cold 50% TCA, and incubated overnight 4°C. ³⁵S-methionine signal of the precipitated protein in this sample represents the starting pool of newly synthesized protein to be examined in the stability experiment. Tetracycline was added to the remaining culture at a final concentration of 62.5 μg/ml to halt translation and the fate of the protein pool was monitored; if proteins degrade, loss of ³⁵S-methionine signal in future precipitated samples is expected. Following 18 h, a 0.5 ml post-tetracycline was collected as described above. Proteins were collected on filters and radioactivity was quantified by scintillation counting as described above. Protein amounts remaining following tetracycline treatment were calculated as the amount of ³⁵S-methionine signal in the post-tetracycline sample divided by the ³⁵S-methionine signal of the pre-tetracycline sample.

Western blot analysis

Samples in Laemmli sample buffer were boiled at 95°C for 10 min. The indicated amount of each mixture was electrophoresed on NuPAGE 4% to 12% Bis-Tris gels (Thermo Fisher Scientific) and transferred to nitrocellulose membranes using the iBlot 2 transfer device (Thermo Fisher Scientific). Following transfer, membranes were blocked in 3% skim milk at room temperature for 2 h, incubated with the indicated primary antibodies for 1 h, washed 3 times, incubated with secondary antibodies (GE Healthcare and Promega) for 1 h, and washed 3 times. Rabbit anti-FLAG (Thermo Fisher Scientific) and rabbit anti-DnaK (Thermo Fisher Scientific) antibodies were used at 1:5,000 dilution in TBST. Mouse anti-TF (Takara Bio) antibodies were used at 1:50,000 dilution. Rabbit anti-S1 ribosomal protein (Agrisera) antibodies were used at 1:1,000 dilution. Horseradish peroxidase-conjugated anti-rabbit (GE Healthcare) and anti-mouse (Promega) secondary antibodies were used at 1:5,000 dilution. Chemiluminescent signal was developed with SuperSignal West Femto Maximum Sensitivity Substrate (Thermo Fisher Scientific) and captured with an ImageQuant LAS 4000 imager (Fujifilm).

Protein purification

Overnight cultures of E. coli BL21(DE3) carrying the plasmids for recombinant protein expression (C-terminal His tags for DnaK, TF, DnaK(T199A), DnaK(1–563), CbpA, GrpE, and HtpG and N-terminal His tag for DnaJ; presented in S1 Table) were diluted 1:100 in 500 ml LB broth containing 100 μg/ml ampicillin. After 3 h of growth at 37°C with 250 rpm shaking, protein production was induced by addition of IPTG (1 mM). The temperature was lowered to 16°C and induction was allowed to proceed overnight with shaking at 150 rpm. The cultures were pelleted by centrifugation for 30 min (8,500 rpm, 4°C) in an Evolution RC Superspeed centrifuge (Sorvall) using an SLC-6000 fixed angle rotor (Sorvall). The pellet was resuspended in 15 ml of Buffer A (50 mM Tris (pH 8.0), 150 mM NaCl, 1 mM MgCl₂) supplemented with 10 U/ml DNase I (New England Biolabs) and 150 μg/ml lysozyme. Following 30 min of incubation on ice, cells were further lysed by sonication. The lysate was clarified by ultracentrifugation for 30 min at 35,000 rpm in a 90 Ti rotor (Beckman Coulter).

The clarified lysate was applied to a column containing a 1 ml bed volume of Ni-NTA agarose (Qiagen) pre-equilibrated with Buffer B (50 mM Tris (pH 8.0), 150 mM NaCl). The agarose beads were washed 8 times with 8 ml of Buffer B to remove unbound proteins. Captured proteins were eluted 4 times, each with 1 ml Buffer C (50 mM Tris (pH 8.0), 150 mM NaCl, 250 mM imidazole). Following SDS-PAGE and Coomassie blue staining analysis of all fractions obtained throughout the purification process, the eluates were pooled, buffer exchanged with Buffer D (50 mM Tris (pH 8.0), 150 mM NaCl, 10% glycerol), and concentrated using Amicon Ultra-15 Centrifugal Filter Units (Millipore).

Protein concentration was determined using a Nanodrop (Nanodrop Machines) and confirmed using a Micro BCA Protein Assay Kit (Thermo Fisher Scientific). Proteins were stored at −80°C long-term, thawed on ice when ready to use, subsequently stored at 4°C, and used within 2 weeks of thawing.

Investigating DnaK association with ribosomes in vitro

In vitro protein synthesis reactions using the PURExpress system (New England Biolabs) based on components from E. coli, a species closely related to S. Typhimurium, were set up in 10 μl reaction volumes and supplemented with 1 U SUPERase-In RNase inhibitor (Thermo Fisher Scientific). Purified DnaK or TF was added at a final concentration of 5 μM. A total of 500 ng of plasmid template pDHFR supplied in the PURExpress kit was added as the last component to initiate protein synthesis. The reactions were incubated at 37°C for 1 h to allow protein synthesis to proceed. After the 1 h incubation, the reactions were immediately placed on ice to halt protein synthesis, and 140 μl of TAKM7 buffer (50 mM Tris-HCl (pH 7.5), 70 mM NH₄Cl, 30 mM KCl) with either 10 mM or no MgCl₂ was added to bring the volume to 150 μl. A 2.5 μl sample was reserved for analysis of the input sample.

The reaction combined with TAKM7 buffer was layered onto 400 μl of a 10% sucrose cushion made with TAKM7 buffer with either 10 mM or 0 MgCl₂ in 1 ml open-top thick wall polypropylene tubes (Beckman Coulter), bringing the final sample Mg²⁺ concentration to 10 mM or 0.16 mM, respectively. Samples were ultracentrifuged for 35 min at 70,000 rpm, 4°C in a TLA-120.2 rotor (Beckman Coulter) to separate ribosomes from other components.

Following sedimentation by ultracentrifugation, the supernatant was removed and saved. Ribosome pellets were resuspended in 25 μl TAKM7 buffer. The RNA concentration of the resuspended ribosomes was measured using a Nanodrop instrument (Nanodrop Machines). Non ribosome-bound supernatants (10 μl) and normalized ribosome samples (100 ng/μl RNA = 20 μl sample loaded) were electrophoresed on NuPAGE 4% to 12% Bis-Tris gels (Thermo Fisher Scientific) and transferred to a nitrocellulose membrane as described above. The blot was developed with Revert 520 Total Protein Stain (LI-COR Biosciences), followed by probing with anti-DnaK (Thermo Fisher Scientific) or anti-TF (Takara Bio) antibodies.

To investigate ribosome association of full-length and truncated DnaK, 10 μl PURExpress reactions were set up as described above. Purified DnaK or its truncated form, DnaK(1–563), was added at a final concentration of 5 μM. Following 1 h of protein synthesis at 37°C, 140 μl TAKM7 buffer (50 mM Tris-HCl (pH 7.5), 70 mM NH₄Cl, 30 mM KCl, 7 mM MgCl₂) was added to bring sample volumes to 150 μl. Samples were ultracentrifuged, electrophoresed, and transferred to nitrocellulose membranes as described above, followed by probing with anti-DnaK (Thermo Fisher Scientific) antibodies.

Samples lacking purified DnaK or TF were included as negative controls.

In vitro protein synthesis assay

DNA templates were prepared by PCR amplification of wild-type S. Typhimurium genomic DNA using primer pairs presented in S2 Table, which contain the requisite T7 promoter and terminator sequences for use with the PURExpress system.

In vitro protein synthesis using the PURExpress system (New England Biolabs) was carried out following the manufacturer’s instructions. Where indicated, purified proteins were added to the following concentrations: DnaK or DnaK(T199A) (5 μM), HtpG (5 μM), or TF (5 μM). An equivalent volume of empty buffer was added to controls lacking added purified protein. Reactions were initiated by addition of 250 ng DNA template and incubated at 37°C in a thermomixer with 300 rpm agitation. Following 45 min of incubation, 1 μl aliquots were removed and immediately quenched with pre-aliquoted Laemmli sample buffer (Bio-Rad) to a final volume of 10 μl (Bio-Rad). Samples were analyzed by western blotting using anti-FLAG antibodies.

To investigate in vitro protein synthesis supplied with mRNA template, mRNA corresponding to the appropriate gene was transcribed in vitro using T7 RNA polymerase (New England Biolabs) at 37°C for 1 h. The reaction was treated with DNase I digestion (New England Biolabs), purified by ethanol precipitation, and resuspended in nuclease-free water. In vitro protein synthesis was performed as described using 2 μg mRNA template in place of 250 ng DNA template.

To investigate in vitro protein synthesis in the presence of purified DnaK and TF, the reaction was carried out as described above following preincubation of DnaK or TF with the protein synthesis machinery (PURExpress Solutions A and B) for 30 min at room temperature. Following preincubation, the second chaperone was added along with DNA template. The reaction was allowed to proceed for 45 min and reporter protein yield was assessed by western blot, both as described above.

ATP hydrolysis assay

Release of inorganic phosphate (Pi) from ATP hydrolysis was measured using the EnzChek Phosphate Assay kit (Thermo Fisher Scientific) according to the manufacturer’s instructions, with the exception that reactions were scaled down to 100 μl and performed in clear-bottom 96-well plates. Where indicated, purified proteins were added to the following final concentrations: DnaK (2 μM), DnaJ (0.4 μM), CbpA (0.4 μM), or GrpE (0.2 μM). 70S ribosomes (New England Biolabs) or 30S or 50S ribosomal subunits were added to a concentration of 0.5 μM. Following preincubation at room temperature for 10 min, 4 μl of 10 mM ATP was added to initiate the reaction. Reactions proceeded at 22°C. Absorbance at 360 nm, which reflects the production of 2-amino-6-mercapto-7-methylpurine stimulated by the presence of inorganic Pi [80] (the product of the ATPase reaction), was recorded every 5 min using a Spectramax 340PC-384 plate reader. The A360 value at t = 0 was subtracted from subsequent measurements. Controls reactions performed without the addition of ATP confirmed that no contaminating Pi was present in reagents and labware.

Insoluble protein isolation and analysis

Bacterial cultures were grown overnight in 10 mM Mg²⁺, washed 3 times in medium lacking Mg²⁺, and subcultured 1:50 in 10 ml of low or high Mg²⁺ medium for the indicated times. Insoluble protein isolation was performed as described [81]. Total, soluble, and insoluble protein samples were electrophoresed on NuPAGE 4% to 12% Bis-Tris gels (Thermo Fisher Scientific) and stained with Coomassie blue. Quantification of protein amounts was performed using the gel analysis feature of ImageJ to measure intensities and abundances of protein bands present in each lane.

Bacterial survival assay

Bacterial cultures were grown overnight in 10 mM Mg²⁺, washed 3 times in medium lacking Mg²⁺, and subcultured 1:50 in 2 ml N-minimal medium with 10 μM or 10 mM Mg²⁺. For hyperosmotic stress conditions, an overnight culture grown in medium with high (10 mM MgCl₂) was washed 3 times in medium with 1 mM MgCl₂ and diluted 1:50 into LB medium with 1 M NaCl and 1 mM MgCl_2. After 24 h of growth, cells were serially diluted in sterile PBS and plated on LB agar plates. After overnight incubation at 37°C, the number of colony forming units (CFUs) was enumerated and corrected for the dilution factor and plating volume to calculate the number of CFUs in the culture.

Sublethal chemical translation inhibitor assay

Overnight bacterial cultures grown in high Mg²⁺ were washed 3 times in medium lacking Mg²⁺ as described. Upon subculture into 2 ml low Mg²⁺ medium at a 1:50 dilution, antibiotics were added to the concentrations indicated above. Following 24 h of growth, the cultures were serially diluted in sterile PBS and plated on LB agar plates.

Luciferase reactivation

Luciferase reactivation was performed as described [82,83] with slight modifications and scaled down to a 40 μl reaction volume. Briefly, 80 nM firefly luciferase (Sigma Aldrich) was denatured at 42°C for 5 min in a buffer of 25 mM HEPES (pH 7.5), 50 mM KCl, 15 mM MgCl₂, and 2 mM DTT, then placed on ice. BSA (0.05 mg/ml), ATP (1 mM), and an ATP regeneration system (20 mM creatine phosphate and 0.06 mg/ml creatine kinase) were added to the reaction mix, followed by addition of chaperones, cochaperones, and/or ribosomes. DnaK, CbpA, and GrpE were used at concentrations of 2 μM, 0.4 μM, and 0.2 μM, respectively, and 70S ribosomes were used at a concentration of 0.5 μM.

Renaturation of luciferase was allowed to proceed at 30°C. At the indicated times, 5 μl aliquots were removed. Luciferase activity was determined with the addition of 120 μl of substrate containing 200 μM D-luciferin (Sigma Aldrich), 0.5 mM ATP, and 10 mM MgCl₂. Luminescence was measured in a Spark Multimode plate reader (TECAN) with an integration time of 10,000 msec. The luminescence value at t = 0 was subtracted from subsequent measurements to determine the amount of luminescent activity regained.

Nuclease activity assay

To rule out the presence of DNases or RNases in our preparations of purified DnaK protein, DNA and mRNA corresponding to pmrA-FLAG (the same template used in in vitro protein synthesis assays) were produced by PCR and T7-mediated in vitro transcription, respectively. The resulting DNA and mRNA fragments were purified using a QIAquick PCR purification kit (Qiagen) and ethanol precipitation, respectively. Stability reactions were performed in a buffer of 25 mM Tris (pH 8.0), 100 mM KCl, 12 mM MgCl₂, 2 mM ATP, 4 mM phosphoenolpyruvate, and 20 μg/ml pyruvate kinase in the presence of either empty protein storage buffer or 5 μM DnaK, and 250 ng DNA substrate or 1 μg RNA substrate was added to a reaction volume of 10 μl. Samples were incubated at 37°C. At the indicated times, 4 μl aliquots (for DNA) or 3.5 μl aliquots (RNA) were removed and frozen in dry ice. Aliquots were combined with 6X DNA loading dye (Thermo Fisher Scientific) and electrophoresed on 1% Tris-acetate-EDTA agarose gels, then stained with SYBR Gold nucleic acid gel stain (Thermo Fisher Scientific) and visualized in an Amersham ImageQuant 800 imager (Cytiva).

Raw data accessibility

Images of uncropped gels, strain verification, additional replicates, quantifications, and all other raw data can be found in: DOI: 10.17632/m7788yc9sf.1.