Translational Regulation of the Escherichia coli Stress Factor RpoS: a Role for SsrA and Lon

Media.Bacteria were grown at 37°C or 25°C in Luria-Bertani (LB) rich medium (37). Chloramphenicol (30 μg/ml), ampicillin (50 μg/ml), kanamycin (50 μg/ml), and tetracycline (30 μg/ml) were included when appropriate. MacConkey agar plates with 1% lactose were used in analysis of strains carrying lacZ chromosomal fusions.

Genetic procedures, bacterial strains, and plasmids.All strains used in this paper were derivatives of E. coli K-12 MG1655 and are listed in Table 1, as are plasmids used in this study. Transductions with P1vir were done as described previously (25). Plasmid DNA was extracted using QIAGEN kits. Transformation of appropriate bacterial strains was performed as described previously (37). MG1655 ΔX74lac (DJ480; obtained from D. Jin, National Cancer Institute [NCI]) was lysogenized with a λ phage carrying an rpoS-lacZ translational fusion (38) to create strain SG30013 (41).

The chromosomal P_BAD-rpoS990::lacZ translational fusion was constructed by the linear DNA transformation method (45). The starting strain NC397 (from D. Court, NIH) carries a defective λ that provides the λ Red functions and a kan^r-Cat^r-sacB cassette inserted between lacI and lacZ that deletes the lacI and lacZ promoters. The forward and reverse primers to amplify the P_BAD-rpoS sequence were designed to carry homology to 42 nt at the 3′ end of the kanamycin resistance cassette and 51 nt at the 5′ end of lacZ, respectively. Primer sequences are 5′-GAGGGTGGCGGGCAGGACGCCCGCCATAAACTGCCAGGAATTCAAACCGGTAACCCCGCTTATTAAAAGC-3′ and 5′-ATTAAGTTGGGTAACGCCAGGGTTTTCCCAGTCACGACGTTGTAAAACGACCTCGCGGAACAGCGCTTCGATATTCAG-3′.

The P_BAD-RpoS plasmid (see Table 1) was used as a template, and the PCR product was purified and electroporated into NC397 cells as described in Yu et al. (45). After electroporation, cells were grown for 3 h in LB at 32°C and then washed and suspended in M63 salts. They were plated on minimal glycerol-sucrose plates (M63, 0.2% glycerol, 5% sucrose, B1, biotin) and incubated 2 to 3 days at 32°C. Sucrose and Kan^r colonies were purified and screened for loss of Cat^r. We obtained one such clone, which was white or light pink when plated on MacConkey lactose and red on MacConkey arabinose. The construct was checked by PCR and sequenced entirely before being transduced by P1 into MG1655, selecting for Kan^r.

β-Galactosidase assays.β-Galactosidase activity of the various lacZ fusions was assayed on a SpectraMax 250 (Molecular Devices) microtiter plate reader as described previously (23). Specific activities are represented as the V _max divided by the optical density at 600 nm (OD₆₀₀), and these units are about 25 times lower than standard Miller units.

Expression and purification of His-tagged RpoS. E. coli cells harboring plasmid pBAD-HisRpoS were grown in LB at 25°C until the OD₆₀₀ was 0.5. Overexpression of His-tagged RpoS was then induced by addition of 0.05% arabinose. After collecting the cells by centrifugation, the bacterial pellets were resuspended in buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl) and cells were broken by sonication (90 s; amplitude, 37%). The soluble protein fraction was obtained by centrifugation at 12,000 rpm at 4°C for 30 min. The obtained supernatant (cell extract) was passed over a nickel-loaded metal chelating Sepharose column (0.2 ml). The column was washed with 30 column volumes of buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl), and bound proteins were eluted using a 50 mM Tris-HCl (pH 7.5) buffer containing 0.5 M imidazole. The protein concentration was measured by the method of Bradford using bovine serum albumin as a standard (3). Eluted proteins were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (equal quantities of proteins were directly mixed in 10 μl of SDS-gel loading buffer).

Assay for RpoS degradation in vivo.Cells were grown in LB at 25°C until the OD₆₀₀ was 0.4. Cells were treated with chloramphenicol (100 μg/ml) or spectinomycin (100 μg/ml), and 1-ml samples were removed at the indicated time points and treated as described below.

Gel electrophoresis and Western blotting for RpoS.One-milliliter samples were removed from cultures to tubes containing 100 μl of cold 50% tricholoroacetic acid (TCA). After centrifugation, pellets were washed twice with 500 μl of cold 80% acetone, air dried, and resuspended in SDS-gel loading buffer. Equal quantities of protein were separated on precast SDS-12% PAGE acrylamide gels (Invitrogen) and transferred onto nitrocellulose membrane filters (Invitrogen). Filters were incubated with anti-RpoS antibodies at 1:4,000 (S. Wickner, NCI). Immunoblots were developed by using horseradish peroxidase-conjugated goat anti-rabbit antibody, followed by enhanced chemiluminescence (Amersham Pharmacia).

tmRNA (SsrA) is implicated in RpoS regulation.We noted that an ssrA::cat derivative of a strain carrying a translational fusion of RpoS, rpoS750::lacZ (38), was significantly less red (Lac⁺) on lactose MacConkey plates than the isogenic ssrA⁺ strain (Fig. 1A). This fusion contains the region from the promoter upstream of the nlpD gene to nucleotide 750 of rpoS, linked to lacZ, and includes the ClpXP degradation signal within RpoS; it is therefore subject to the same transcriptional, translational, and proteolytic degradation regulatory signals as RpoS itself. The experiments described here were carried out to determine the basis for this difference in RpoS expression.

• Open in new tab
• Download powerpoint

FIG. 1.

There is less RpoS in an ssrA-deficient strain. (A) Strains SG30013 (WT) and CRB200 (ssrA) were plated on a lactose MacConkey plate and incubated overnight at 37°C. (B) Strains MG1655 (WT) and CRB233 (MG1655ssrA) were grown in LB at 25°C. One-milliliter samples were taken at different OD₆₀₀s during growth and TCA precipitated. Proteins were separated by SDS-PAGE and RpoS was detected by Western blotting, as described in Materials and Methods. Numbers between the two panels give the relative amount of RpoS in the wild-type host compared to the ssrA mutant host. (C) Half-life of RpoS in different mutant strains. The half-life of RpoS was measured in cells grown at 25°C in LB. Protein synthesis was inhibited with chloramphenicol or spectinomycin at an OD₆₀₀ of 0.4. Samples were removed at specific time points and analyzed by Western blotting as described in Materials and Methods.

The results seen on plates for the fusion were confirmed for the native endogenous RpoS protein with a Western blot with anti-RpoS antibody; the differences were more reproducible for cells grown at 25°C and are shown in Fig. 1B. There was significantly less RpoS protein in the ssrA mutant even at lower cell densities. While liquid assays of the fusion showed no difference between ssrA ⁺ and ssrA::cat strains at 37°C, there was significantly lower β-galactosidase activity in the absence of SsrA in stationary phase at 25°C (Table 2, compare lines 1 and 2).

A transcriptional fusion of rpoS and lacZ carrying the same region of rpoS was unaffected by an ssrA mutant, suggesting that regulation of RpoS by SsrA is posttranscriptional (data not shown).

Degradation of RpoS is dependent upon ClpXP and the adaptor protein RssB (47) and is more rapid in exponential phase than in stationary phase. Two sets of experiments were done to investigate whether the ssrA effect was via effects on RpoS degradation. Direct measurement of the degradation rate of RpoS at 25°C in a wild-type (WT) or ssrA mutant strain gave a half-life of 6 min in both cases (Fig. 1C). This experiment was done at an OD₆₀₀ of 0.4, where we see a consistent twofold difference in RpoS accumulation (Fig. 1B). In a clpP mutant, the half-life of RpoS was >30 min (Fig. 1C), consistent with degradation of RpoS even at 25°C, depending entirely on ClpXP.

In addition, the effect of the ssrA mutant on the translational rpoS-lacZ fusion was measured in a clpP mutant in which RpoS is normally stable. As expected, in a clpP mutant RpoS::LacZ levels were significantly higher, particularly in exponential phase (Table 2, compare lines 1 and 3). However, the effect of the ssrA mutation was still detected, both in exponential and stationary phases (Table 2, compare lines 3 and 4). Therefore, the ssrA effect is independent of ClpXP-dependent RpoS degradation. Taken together with the absence of a change in RpoS stability in ssrA mutants, we can rule out a role for ssrA in degradation by ClpXP or another protease.

A translational fusion containing only the first 477 nucleotides of the rpoS coding region, ending before the region required for ClpXP degradation, rpoS477::lacZ, was also tested. No differences in expression level between the wild type and an ssrA mutant were detected for this fusion (data not shown). Combined with our other results, these data suggest that the region necessary for the ssrA effect extends beyond nucleotide 477 and that this regulation may be at the level of RpoS translation.

Mode of action of SsrA.SsrA has two significant effects, relieving ribosome stalling and targeting incomplete proteins for degradation (for a review, see reference 11). Mutant derivatives of ssrA can distinguish these effects and were tested for their ability to complement the ssrA::cat mutation (Table 2). The pJW34 plasmid expresses the ssrA(UG) allele, a mutant unable to be charged with Ala, and hence does not interact with the ribosome (42) and should act like a null allele. As expected, this mutant was unable to complement the defect in ssrA mutant cells (Table 2, line 10). pJW28, encoding wild-type ssrA, fully complemented the ssrA defect (Table 2, line 7). However, two other ssrA derivatives, ssrA(DD) and ssrA(O), each allow ribosome release but tag translating proteins with amino acid sequences that are not recognized by cellular proteases. The DD derivative replaces the C-terminal AA amino acids that are recognized by ClpXP with DD (22); ssrA(O) adds a truncated three-amino-acid tag, which is also not recognized by the ClpXP protease (42). Both plasmids complemented the SsrA defect in rpoS::lac expression, although complementation by the DD allele was consistently less complete (Table 2, lines 8 and 9). The DD allele has been reported to be less efficient at tagging than the wild type (35), which may explain the lack of full complementation in this case. Therefore, tagging for degradation is not the crucial function involved in modulating RpoS levels.

We propose that the efficiency of RpoS mRNA translation is increased, directly or indirectly, by the ability of SsrA to interact with ribosomes and relieve ribosome stalling.

Lon protease also affects RpoS accumulation.For reasons discussed briefly below, we tested the effect of mutations in lon on RpoS expression and tested epistasis of lon and ssrA by looking at a double mutant. A lon::Tn10 mutation was transduced into wild-type and ssrA mutant strains carrying the translational rpoS750-lacZ fusion, and β-galactosidase was assayed. Table 2 shows that there is more RpoS-Lac in the lon mutant than in the wild-type strain in both exponential and stationary phases at 25°C (Table 2, compare lines 1 and 5). Significantly, the lon ssrA double mutant acts like the lon strain and has a higher level of RpoS than the wild-type or the ssrA mutant strain (Table 2, compare lines 1, 2, 5, and 6). These results suggest that the loss of RpoS expression in an ssrA deletion requires Lon, either directly or indirectly. Western blot analysis of RpoS levels confirmed this result in lon mutant and lon ssrA double mutant strains (Fig. 2A). There was significantly more RpoS in exponential phase at 25°C in these two strains; quantitation of the blot by scanning showed a threefold increase in both lon and lon ssrA derivatives relative to the wild-type strain. At higher temperatures (30° and 37°C), the lon mutant had little or no effect on RpoS levels (see Fig. S1 in the supplemental material). The half-life of RpoS in both a WT and a lon mutant strain at 25°C was approximately 6 min (Fig. 1C). Therefore, Lon is not directly degrading RpoS but must be acting indirectly to affect synthesis of RpoS.

• Open in new tab
• Download powerpoint

FIG. 2.

There is more RpoS in a lon-deficient strain. (A) Western blot analysis of RpoS accumulation in a lon mutant strain in exponential phase at 25°C. Strains MG1655 (WT), CRB233 (ssrA), CRB296 (lon), and CRB298 (ssrA lon) were grown in LB at 25°C. One-milliliter samples were taken at an OD₆₀₀ of 0.3 and TCA precipitated. Proteins were separated by SDS-PAGE and RpoS was detected by Western blotting, as described in Materials and Methods. Different amounts of extract were loaded on the gel, as indicated, to improve quantitation. (B) The SsrA and Lon effects are observed when RpoS is expressed from a P_BAD vector. Strains CRB306 (WT), CRB307 (ssrA), CRB308 (lon), and CRB309 (ssrA lon) were grown in LB containing 0.005% arabinose at 25°C. One-milliliter samples were taken at OD₆₀₀s of 0.3 and 0.5 and TCA precipitated. Proteins were separated by SDS-PAGE and RpoS was detected by Western blotting, as described in Materials and Methods. Truncated RpoS polypeptides are indicated with a dot. (C) Effect of different mutations on the P_BAD-rpoS990::lacZ chromosomal fusion. Strains CRB316 (WT), CRB317 (ssrA), CRB319 (lon), and CRB318 (dsrA) were grown in LB containing 0.01% arabinose at 25°C. β-Galactosidase activity was measured during growth and is presented as specific activity as a function of OD₆₀₀. Results of a typical experiment are shown.

Both SsrA and Lon act independently of the RpoS leader and small RNAs.The major translational regulation of rpoS is dependent upon the upstream hairpin in the rpoS mRNA and its interaction with Hfq and small RNAs (reviewed in reference 33). A direct test of the importance of the upstream leader for the SsrA effect was carried out by measuring the effects of ssrA and lon in RpoS expression from constructs lacking all upstream sequences, as well as the natural promoters. A plasmid carrying the rpoS gene under the control of the P_BAD promoter (pBAD-RpoS) was initially used. The plasmid was introduced into strains carrying a mutation in the chromosomal copy of rpoS and also carrying mutations in ssrA, lon, or both. The accumulation of RpoS was determined after induction with a low concentration of arabinose (0.005%). As seen with RpoS encoded by the chromosome (Fig. 1B), there was less RpoS expressed from the leaderless rpoS plasmid in the ssrA mutant strain; this is most visible at an OD₆₀₀ of 0.3 (Fig. 2B, compare lanes 1 and 3). The lon mutant led to increased expression of RpoS (Fig. 2B, compare lanes 1 and 5), and, as before, the ssrA mutant did not have any measurable effect on RpoS levels in the lon mutant (Fig. 2B, lane 7) but may have had some effect on the accumulation of truncated versions of RpoS (Fig. 2B, lane 8; also see Fig. 3, described below). None of these products were observed in the absence of arabinose (data not shown). These results suggest that both the ssrA effect and the lon effect are independent of the upstream leader and therefore are not dependent upon regulation by small RNAs such as DsrA.

• Open in new tab
• Download powerpoint

FIG. 3.

Overexpression of proteins in an ssrA mutant background. (A) Truncated RpoS proteins appear when RpoS is overexpressed in an ssrA mutant strain (pBAD-RpoS). Strains all carried a pBAD-RpoS plasmid and a chromosomal rpoS mutation. Strains CRB250 (WT), CRB251 (ssrA), and CRB291 (clpP) were grown in LB at 25°C. At an OD₆₀₀ of 0.8, 0.1% arabinose was added for 30 min. Cells (1 ml) were then harvested, precipitated with TCA, and washed with acetone. Proteins were separated by SDS-PAGE, and RpoS was detected by Western blotting, as described in Materials and Methods. The size estimates for the shortest RpoS truncated protein bands are 16, 17, and 19 kDa. (B) Purification of His-tagged truncated RpoS proteins. Strains all carried a pBAD-HisRpoS plasmid and chromosomal rpoS and lon mutations. Strains CRB304His (WT) and CRB305His (ssrA) were grown in LB at 25°C until an OD₆₀₀ of 0.5. Overexpression of His-tagged RpoS was then induced by addition of 0.05% arabinose. After 3 h of induction, the cells were harvested and cellular extracts were realized by sonication. His-tagged proteins were then purified with Ni-nitrilotriacetic acid columns, and the same amount of proteins (1.25 μg) was loaded on an SDS-PAGE acrylamide gel (12%) and analyzed by Western blotting. Truncated His-tagged RpoS proteins specific to the ssrA host are indicated with a star.

We confirmed these results by constructing a new rpoS::lac chromosomal fusion, a P_BAD-rpoS::lacZ translational fusion. As with the pBAD-RpoS plasmid, the rpoS gene was cloned downstream of the P_BAD promoter, directly at the initiator ATG; the ribosome binding site was derived from the P_BAD vector. In this construct, lacZ was fused to the 3′ end of the rpoS coding region at nucleotide 990. This fusion was created in the chromosome at the site of the lac operon, deleting lacI and the lac promoter, which were replaced by sequences from P_BAD and rpoS (see Materials and Methods). Thus, all the upstream regions necessary for normal transcriptional regulation and translational regulation by small RNAs were absent. We induced the expression of rpoS-lacZ with 0.01% arabinose and assayed β-galactosidase produced in wild-type and mutant strains. Figure 2C shows that the ssrA::cat mutation led to a defect in rpoS::lacZ expression and a lon::Tn10 mutation led to increased expression, consistent with other results. A dsrA::cat mutation had no effect on rpoS::lacZ expression, confirming that regions important for the pairing of the small RNA were absent.

Truncated RpoS proteins are produced in an ssrA mutant strain.In the experiments described above, we noted the presence of a number of truncated derivatives of RpoS, both in the ssrA mutant (Fig. 2B, lane 4) and the lon ssrA mutant (Fig. 2B, lane 8). To look at these further, cells were grown with a higher level of arabinose (0.1%), and samples were taken at a somewhat higher optical density (OD₆₀₀ = 0.8). In the ssrA mutant strain, a series of truncated RpoS proteins were produced (Fig. 3A). These proteins are only seen after arabinose induction and depend on the presence of the plasmid (data not shown), strongly supporting the idea that they are fragments of RpoS and not cross-reacting proteins.

One possible explanation for the accumulation of truncated proteins in an ssrA mutant is that translational pause/arrest sites within RpoS are normally sites for ssrA tagging and subsequent degradation of the tagged proteins. In the absence of ssrA, tagging cannot occur, and the resulting truncated proteins are therefore more stable. In this model, we would expect similar truncated species to accumulate in a clpP mutant, where most SsrA-tagged proteins are stable (9, 12). That was not the case (Fig. 3A). It does seem possible that these truncated proteins are somewhat unstable and that the Lon protease contributes to their degradation, since they accumulate to a greater extent in lon ssrA hosts (Fig. 2B, lane 8 compared to lane 4 and lane 6).

In a parallel set of experiments, a plasmid encoding a His-tagged RpoS (His-RpoS) was also induced in a lon and lon ssrA host, and the proteins were purified using an affinity column. The purified proteins are shown in Fig. 3B and again show an enrichment of truncated species in the ssrA mutant. Therefore, at least some of these shorter species are truncated at the C terminus of RpoS, consistent with translation arrest and release of the polypeptide.

The truncated RpoS proteins quite likely reflect defective translation of RpoS (premature stops). Our results are consistent with an effect of SsrA in promoting translation of the rpoS open reading frame (ORF) independently of the native promoter, the leader, and the ribosome binding site. In a control experiment, a plasmid expressing LacZ was examined in the wild type and ssrA mutant. While we detected low levels of truncated LacZ proteins from this plasmid, the amounts were not increased in an ssrA mutant strain (see Fig. S2 in the supplemental material), suggesting that the effect on RpoS is specific.