Stability of the Osmoregulated Promoter-Derived proP mRNA Is Posttranscriptionally Regulated by RNase III in Escherichia coli

ABSTRACT

The enzymatic activity of Escherichia coli endo-RNase III determines the stability of a subgroup of mRNA species, including bdm, betT, and proU, whose protein products are associated with the cellular response to osmotic stress. Here, we report that the stability of proP mRNA, which encodes a transporter of osmoprotectants, is controlled by RNase III in response to osmotic stress. We observed that steady-state levels of proP mRNA and ProP protein are inversely correlated with cellular RNase III activity and, in turn, affect the proline uptake capacity of the cell. In vitro and in vivo analyses of proP mRNA revealed RNase III cleavage sites in a stem-loop within the 5′ untranslated region present only in proP mRNA species synthesized from the osmoregulated P1 promoter. Introduction of nucleotide substitutions in the cleavage site identified inhibited the ribonucleolytic activity of RNase III on proP mRNA, increasing the steady-state levels and half-life of the mRNA. In addition, decreased RNase III activity coincided with a significant increase in both the half-life and abundance of proP mRNA under hyperosmotic stress conditions. Analysis of the RNA bound to RNase III via in vivo cross-linking and immunoprecipitation indicated that this phenomenon is related to the decreased RNA binding capacity of RNase III. Our findings suggest the existence of an RNase III-mediated osmoregulatory network that rapidly balances the expression levels of factors associated with the cellular response to osmotic stress in E. coli.

IMPORTANCE Our results demonstrate that RNase III activity on proP mRNA degradation is downregulated in Escherichia coli cells under osmotic stress. In addition, we show that the downregulation of RNase III activity is associated with decreased RNA binding capacity of RNase III under hyperosmotic conditions. In particular, our findings demonstrate a link between osmotic stress and RNase III activity, underscoring the growing importance of posttranscriptional regulation in modulating rapid physiological adjustment to environmental changes.

INTRODUCTION

The enzymatic properties and physiological roles of RNase III family enzymes are evolutionarily well conserved in both prokaryotes and eukaryotes (1–3). These enzymes are double-stranded-RNA-specific endo-RNases that create 5′-phosphate and 3′-hydroxyl termini with two-nucleotide overhangs. Genome-wide analyses of Escherichia coli transcripts indicated that the abundance of a large number of mRNA species is regulated by RNase III (4, 5). The abundance of several E. coli mRNA transcripts is dependent on the endoribonucleolytic activity of RNase III, including rnc (6), pnp (7), bdm (8), betT (9), corA (10), proU (11), mltD (12), and rng (4). Studies indicate that a subgroup of mRNA transcripts encoding factors associated with the cellular response of E. coli to osmotic stress is regulated by RNase III (8, 9, 11). The mechanisms of RNase III-mediated regulation of bdm (8), betT (9), and proU (11) have been identified. Specifically, RNase III controls the degradation of these mRNAs and their cleavage is significantly altered in E. coli cells exposed to osmotic stresses. The mRNA abundance of proP, another important osmoregulator, appeared to be dependent on the cellular RNase III concentration (8). However, the osmotic stress-induced RNase III-mediated regulation of proP has not been studied.

In E. coli, the proP gene encodes a low-affinity transporter of osmoprotectants, including proline and glycine betaine, which sense extracellular osmotic pressure and respond by maintaining membrane turgor pressure (13, 14). ProP is a member of the major facilitator superfamily and is an osmoprotectant proton symporter that is regulated by high osmotic pressure (15–17). proP is transcribed from two different promoters, an rpoD-dependent promoter, P1, and an rpoS-dependent promoter, P2. P1, which is similar to the proU promoter, is transiently induced upon subculture and is upregulated under hyperosmotic conditions, while P2 is induced by the stationary phase in the presence of Fis, a small, nucleoid-associated protein (18–20). Expression of proP is transcriptionally regulated by cyclic AMP receptor protein (CRP), an osmotic repressor of proP P1 transcription, and by Fis, an activator of rpoS-dependent promoter expression (15, 18, 20). Recent studies showed that lesions at proQ, which encodes an RNA chaperone, reduced the levels of ProP protein and activity through an unknown mechanism (21–24).

On the basis of previous findings that suggested an important role for RNase III in the posttranscriptional regulation of genes involved in the cellular response to osmotic stresses, we investigated the effect of RNase III activity on the expression of the proP gene, the protein product of which greatly contributes to the osmotic resistance of E. coli. Here, we provide direct evidence that RNase III controls the degradation of proP mRNA by cleaving a stem-loop in its 5′ untranslated region (UTR) present only in mRNA species synthesized from the osmoregulated P1 promoter, suggesting a physiological relationship between the regulation of RNase III activity and osmotic stress resistance.

MATERIALS AND METHODS

Strains and plasmids.

The E. coli strains and plasmids used in this study are listed in Table 1. E. coli strains BL2014 and BL20124 were constructed by deleting the proP gene (from the 5′ UTR to the 3′ end of the coding region) from the genomic DNA of MG1655 and BL2012, respectively, as described by Datsenko and Wanner (25). The PCR primers used in these experiments were proP H1P1 (5′-ATAAGACAGCGTCACATCAGGCCATCCGTTTCAGCTGTGTAGGCTGGAGCTGCTTC-3′) and proP H2P2 (5′-GGCCGTCGCGCTGATTTTTCTGGCGTTTGCGGAAATCATATGAATATCCTCCTTA-3′). pKD3 was used as the template. Plasmids pProP3 and pProP4 are derivatives of pPM30 (26) that direct the synthesis of unmodified and carboxy-terminally hexahistidine-tagged forms of ProP, respectively, under the control of the intact P1 and P2 promoters. These plasmids can also direct the conditional synthesis of ProP proteins under the control of the lacUV5 promoter. To construct pProP3 and pProP4, DNA fragments containing the proP gene were amplified with PCR primers prop-5-UTR-F (5′-ATGCGGCCGCGTCATTAACTGCCCAATT-3′) and prop-3-end-R (5′-ATTCTAGATTATTCATCAATTCGCGGATG-3′) for pProP3 and proP 5-UTR-F and prop-3-end-His-R (5′-ATTCTAGATTAATGATGATGATGATGATGTTCATCAATTCGCGGATG-3′) for pProP4. The products were then cloned into the NotI and XbaI restriction sites in pRNG3 (4). To construct the pProP4 derivatives, DNA fragments containing mutations in the RNase III cleavage site of proP mRNA were amplified by the overlap extension PCR method and cloned into pProP4 at the NotI and XbaI restriction sites. The PCR primers used were prop-mt-R (5′-ATCTCTGTAACAAGCGTAGCNNGTTTGCTTACACCCTCCGGT-3′), prop-108F (5′-GCTACGCTTGTTACAGAGAT-3′), proP 5-UTR-F, and proP 3-end-His-R.

TABLE 1.

Strains and plasmids used in this study

Strain or plasmid	Genotype or description	Source or reference
Strains
MG1655	F− lambda− ilvG-rfb-50 rph-1	41
BL2012	MG1655 rnc-14::ΔTn10	10
BL2014	MG1655 proP	This study
BL20124	MG1655 rnc-14::ΔTn10proP	This study
RM2	F− trp lacZ ΔputPA101 rpsL thi	42
WG1072	RM2 ΔproQ756::kan	22
Plasmids
pPM30	pSC101 ori, Ampr	26
pRNG3	pSC101 ori, Ampr, rng under PlacUV5 control	4
pProP3	pSC101 ori, Ampr, proP under PlacUV5 control	This study
pProP4	pSC101 ori, Ampr, His-tagged proP under PlacUV5 control	This study
pProP4-mutant clones	Same as pProP4 but with nucleotide substitutions at −111 and/or −110 region(s)	This study
pSD80-rnc-his	ColE1 ori, Ampr, His-tagged rnc under tac promoter control	This study

Plasmid pSD80-rnc-his is a derivative of pSD80 (27) that directs the synthesis of a carboxy-terminally hexahistidine-tagged form of RNase III under the control of the tac promoter. To construct pSD80-rnc-his, DNA fragment containing the rnc gene was amplified with PCR primers RNC 5′ (5′-AGAATTCATATGAACCCCATCGTAATTA-3′) and RNC cds-his R(pst1) (5′-ATCTGCAGTCAGTGGTGGTGGTGGTGGTGGCCATTGGTTAACTGCTC-3′) and cloned into the EcoRI and PstI restriction sites in pSD80.

Semiquantitative RT-PCR analysis.

Semiquantitative reverse transcription (RT)-PCR was performed, and the results were analyzed as previously described (9, 12). Total RNA was isolated with an RNeasy Miniprep kit (Qiagen, Valencia, CA, USA), and 1 μg of RNA was used for cDNA synthesis with a PrimeScript first-strand cDNA synthesis kit (TaKaRa, Otsu, Shiga, Japan). The primers used for RT-PCR were proP + 1 (5′-GTCATTAACTGCCCAATTCA-3′) and proP + 309 (5′-CACCAAAATCGAACCATTCC-3′) for proP, pnp-RT5′ (5′-TTTCGCCTACGGAGTACGGA-3′) and pnp-RT3′ (5′-TTGTTCGTAGCCGCGTACTG-3′) for pnp, rpsO-5′RT (5′-GTACACTGGGATCGCTGAATT-3′) and rpsO-3′RT (5′-GGCCCCCTTTTCTGAAACTCG-3′) for rpsO, and rnpB-F (5′-GAAGCTGACCAGACAGTCGC-3′) and rnpB-R (5′-AGGTGAAACTGACCGATAAG-3′) for M1.

Western blot analysis.

Western blot analysis was performed as previously described (28). An anti-His tag monoclonal antibody was used to detect ProP-His₆, and polyclonal antibodies against RNase III and ribosomal protein S1 were also used. Specific proteins were imaged with a VersaDoc 100 (Bio-Rad, Hercules, CA) and quantified with Quantity One software (Bio-Rad).

In vivo cross-linking and immunoprecipitation.

Cross-linking and immunoprecipitation were carried out as previously described (29, 30).

For in vivo cross-linking, E. coli strain BL2012 harboring pSD80-rnc-his was incubated at 37°C in M63 medium supplemented with 22 mM glucose and 0.17 M NaCl until an optical density at 600 nm (OD₆₀₀) of 0.5 was reached. The culture was further incubated for 4 h in the presence of 0.17 or 0.5 M NaCl. Samples were washed twice with lysis buffer (20 mM Tris-HCl [pH 8.0], 150 mM KCl, 1 mM MgCl₂, 1 mM dithiothreitol [DTT]) and resuspended in 10 ml of lysis buffer. Formaldehyde was added to a final concentration of 1% (vol/vol) (0.36 M), and the samples were rotated at room temperature for 10 min. Cross-linking reactions are quenched by the addition of glycine (pH 7.0) to a final concentration of 0.25 M, followed by incubation at room temperature for 5 min. Cells were washed twice with lysis buffer and resuspended in 2 ml of radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl [pH 7.5], 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.05% SDS, 1 mM EDTA, 150 mM NaCl).

For immunoprecipitation, cell extracts and 20 U of RNase Inhibitor (TaKaRa, Otsu, Shiga, Japan) were added to Ni-nitrilotriacetic acid beads, and the mixture was rotated at 4°C for 2 h. Samples were washed five times with 1 ml of high-stringency RIPA buffer (50 mM Tris-HCl [pH 7.5], 1% NP-40, 1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 1 M NaCl, 0.2 mM phenylmethylsulfonyl fluoride).

Northern blot analysis.

The procedure for Northern blot analysis has been described previously (10). The random hexamer probes used for proP mRNA detection were synthesized with a random primed DNA labeling kit (Roche, Pleasanton, CA). PCR products containing the proP 5′ UTR and open reading frame were synthesized with primers Prop-H1 (5′-ATAAGACAGCGTCACATCAG-3′) and T7-proP-R and used as the template for the synthesis of the random hexamer probes. The oligonucleotide probe used for M1 RNA and 16S rRNA were M1 (5′-GCTCTCTGTTGCACTGGTCG-3′) and 16S 1195R (5′-TAAGGGCCATGATGACTTG-3′), respectively.

In vitro cleavage analysis.

His-tagged RNase III purification and cleavage assays were performed as previously described (31, 32). The procedure used to synthesize 3′- or 5′-end-labeled transcripts was previously described (10). Primers T7-proP-F (5′-TAATACGACTCACTATAGGGTCATTAACTGCCCAATTCA-3′) and T7-proP-R (5′-TTATTCATCAATTCGCGGAT-3′) were used to synthesize the full-length proP mRNA transcript, and Prop-st-T7-F (5′-CTTAATACGACTCACTATAGGGCAGGCGTCAACTGGTTTGATTGTACATTCC-3′) and Prop-st-T7-R were used to synthesize the model hairpin transcript. One picomole of labeled model hairpin transcript was incubated with 0.05 pmol of purified RNase III in the presence of 0.25 μg ml⁻¹ yeast tRNA (Ambion, Austin, TX) and 20 U of RNase Inhibitor (TaKaRa, Otsu, Shiga, Japan) in cleavage buffer (30 mM Tris-HCl [pH 7.9], 160 mM NaCl, 0.1 mM DTT, 0.1 mM EDTA [pH 8.0]). For primer extension analyses of the in vitro cleavage products of the full-length proP mRNA transcript, 1 pmol of unlabeled proP mRNA transcript was incubated with or without 0.09 pmol of purified RNase III for 10 min in cleavage buffer containing 10 mM MgCl₂.

Primer extension.

The procedure used for primer extension analysis has been described previously (8, 10). The primer used was prop-125R (5′-CCGCAAAAGGGGAGCGGGAA-3′).

Determination of proline content.

E. coli strains BL2014 and BL20124 harboring either pPM30 or pProP3 were grown in modified Davis minimal medium (pH 7.5) containing 50 mM Na₂HPO₄-NaH₂PO₄ buffer, 0.1 mg ml⁻¹ MgSO₄, 1.0 mg ml⁻¹ (NH₄)₂SO₄, 0.5 mg ml⁻¹ sodium citrate, 10 mM glucose, 1 mM KCl, and 1 mM proline. When the cultures reached an OD₆₀₀ of 0.5, 1 M NaCl was added. Proline content was assessed before and 3 h after the addition of 1 M NaCl (free proline content) with a colorimetric assay as previously described (33). Briefly, cell extracts were reacted with an equal volume of 6% aqueous 5-sulfosalicylic acid, acid ninhydrin, and acetic acid at 100°C for 60 min. Samples were immediately placed on ice and mixed with 50% ethanol. The proline content of each sample was calculated by measuring its absorbance at 520 nm.

RESULTS

RNase III negatively regulates proP expression.

To investigate whether the absence of RNase III activity affects the abundance of proP mRNA, we measured the steady-state levels of proP mRNA in E. coli strains MG1655 and BL2012 by semiquantitative RT-PCR. Consistent with the microarray data from our previous study (8), rnc mutant cells showed a 2.0-fold increase in the amount of proP mRNA compared to that observed in wild-type cells (Fig. 1A). Western blot analysis also showed that the level of hexahistidine-tagged ProP expression from pProP4 in rnc mutant cells is approximately 3.5 times as high as that in wild-type cells (Fig. 1B). This enhanced ProP expression functioned as an adaptation reaction of E. coli cells exposed to high osmolarity, as indicated by the highest rate of proline uptake detected in rnc mutant cells in the presence of 0.5 M NaCl (Fig. 1C). Under high-osmolarity conditions, the external proline concentration of rnc-positive cells was approximately 4.24-fold higher than that observed in rnc mutant cells. These results imply that, in addition to transcriptional activation of proP expression under hyperosmotic conditions, RNase III is also associated with increased proP expression. Our previous finding of downregulated RNase III activity under hyperosmotic conditions also supports this notion (8, 9). Taken together, these results indicate that steady-state levels of proP mRNA are influenced by the cellular RNase III concentration, suggesting the involvement of RNase III in proP mRNA decay and subsequent regulation of proline uptake under high-osmolarity conditions.

FIG 1.

Downregulation of proP expression by RNase III. (A) Semiquantitative RT-PCR analysis of proP mRNA. The effects of rnc deletion on proP mRNA levels were measured by using total RNA prepared from E. coli strains MG1655 (wild type, WT) and BL2012 (rnc). The strains were grown in M63 medium supplemented with 22 mM glucose and 0.17 M NaCl until an OD₆₀₀ of 0.6 was reached. The abundance of M1, the RNA component of RNase P, was measured as an internal standard to evaluate the total amount of RNA in each reaction mixture. (B) Western blot analysis of the ProP protein. E. coli strains BL2014 and BL20124 harboring the pProP4 plasmid, which contains the gene encoding hexahistidine-tagged ProP, were grown as described above. The amounts of ProP, RNase III, and ribosomal protein S1 were analyzed by Western blot assay. The S1 protein was used as an internal standard to evaluate the amount of cell extract loaded in each lane. The same membrane that was probed for His-tagged ProP was also probed with polyclonal antibodies to RNase III and S1. (C) Effects of RNase III activity on proline uptake as determined by colorimetric analysis of external proline. Strains BL2012 and BL20124 harboring either pPM30 or pProP3 were incubated at 37°C in modified Davis minimal medium containing 1 mM proline. Proline uptake was measured in the absence of NaCl (0hr) and again after 3 h of incubation in the presence of 1 M NaCl, and the final cell density was adjusted to an OD₆₀₀ of 0.5. The external proline content was determined as described by Nagata et al. (33).

Identification of putative RNase III cleavage sites in proP mRNA.

Previous S1 nuclease mapping and primer extension analyses of proP mRNA showed several uncharacterized proP mRNA species with 5′ ends that were mapped between the proP P1 and P2 promoters (Fig. 2) (18, 19). These proP mRNA species were thought to result from either primer extension artifacts or transcription from other start sites. However, on the basis of our observation of RNase III-dependent proP mRNA abundance, we hypothesized that they may represent mRNA species that were cleaved by RNase III. To test whether the region between the P1 and P2 promoters contains cis-acting elements that are responsive to RNase III, we performed a primer extension analysis with a 5′-end ³²P-labeled primer (proP-125R) that was designed to hybridize to a region downstream of the P2 promoter. Total RNA was purified from wild-type and rnc mutant cells grown in Luria-Bertani medium. In the lane loaded with the reaction mixture containing total RNA from wild-type cells, we observed one cDNA band, which we designated the B site (Fig. 3A). For a higher-resolution analysis of the cDNA bands, the reaction was performed with total RNA prepared from E. coli cells overexpressing proP mRNA. In this reaction mixture, we observed two major cDNA bands extended from proP mRNA that appeared to be RNase III dependent (Fig. 3B). One band corresponded to the B site, while the other was designated the A site. These bands were not present in the lane loaded with the reaction mixture containing total RNA extracted from cells that overexpressed proP mRNA in the absence of RNase III expression. To ensure that these sites were generated by RNase III cleavage, we synthesized a full-length proP transcript in vitro and incubated it with purified RNase III. RNA samples were subsequently purified from the cleavage reactions and analyzed by primer extension with the 5′-end ³²P-labeled primer (proP-125R). Consistent with the results of the in vivo primer extension analyses of proP mRNA (Fig. 3A and B), we observed two major cDNA bands that corresponded to cleavage sites A and B (Fig. 3C).

FIG 2.

Schematic representation of the proP promoter region. The locations of the transcription initiation sites of the P1 and P2 promoters relative to the start of the proP coding sequence, as proposed by Xu and Johnson (18) and Mellies et al. (19), are illustrated. An RNase III-targeted hairpin loop identified in this study is illustrated between the P1 and P2 promoters in the sequence at the bottom. Potential RNA polymerase recognition sequences are denoted with horizontal brackets.

FIG 3.

Identification of putative RNase III cleavage sites in proP mRNA in vitro and in vivo. (A) Primer extension analysis of endogenously expressed proP mRNA. (B) Primer extension analysis of heterologously overexpressed proP mRNA. Total RNA was prepared from MG1655 and BL2012 cells (A) or those harboring pProP3 (B), which were grown at 37°C in Luria-Bertani (LB) medium supplemented with 0.3 M NaCl. Total RNA (40 [B] or 80 [A] μg) was hybridized with a 5′-end ³²P-labeled primer (proP-125R). Synthesized cDNA products were analyzed by 12% polyacrylamide gel electrophoresis (PAGE). Sequencing ladders were produced with the same primer used for cDNA synthesis, and a PCR product encompassing the proP gene was used as the template. Putative transcription initiation sites derived from the P1 and P2 promoters are identified as TIS (P1) and TIS (P2), respectively. −, no expression; +, endogenous expression; +++, overexpression; **, cDNA bands synthesized in a proP mRNA-independent manner under high-osmolarity conditions. (C) Primer extension analysis of a proP transcript synthesized in vitro. One picomole of full-length proP transcript synthesized from the P1 transcript start site was incubated with or without 0.09 pmol of purified RNase III for 10 min in cleavage buffer with MgCl₂. RNA samples were purified by phenol-chloroform extraction and ethanol precipitation and then hybridized with a 5′-end ³²P-labeled primer (proP-125R). Lane m contains the reaction mixture used in the last lane of Fig. 3B and was used to compare cDNA products synthesized from the RNase III-cleaved synthetic proP transcript in vitro. Synthesized cDNA products were analyzed by 12% PAGE. (D) Predicted secondary structure of proP mRNA. The secondary structure was deduced with the M-fold program (40). The model hairpin RNA used for the in vitro cleavage assays presented in Fig. 2E and F is shown in the right panel. (E, F) In vitro cleavage of the model proP hairpin RNA. One picomole of 5′-end ³²P-labeled (E) or 3′-end ³²P-labeled (F) proP model hairpin was incubated with 0.05 pmol of purified RNase III in a cleavage buffer with (III + Mg²⁺) or without (III) MgCl₂. Samples were withdrawn at the indicated time intervals and separated by 12% PAGE in gels containing 8 M urea. Cleavage products (A and B) were identified by using size markers generated by alkaline hydrolysis (Hydrolysis) and RNase T1 digestion. Positions of G residues are numbered 1 to 17 in the secondary structure and gel picture. Other minor cleavage products indicated by asterisks might have been produced by RNase III digestion of RNA transcripts containing an incomplete 3′ or 5′ end.

To biochemically demonstrate the cleavage of proP mRNA by RNase III, an in vitro cleavage assay was performed with a model hairpin RNA containing putative RNase III cleavage sites A and B in the proP mRNA (Fig. 3D). The model hairpin RNA contains the nucleotide sequence between nucleotides −164 and −100 from the start codon of proP. RNase III cleavage of a 5′-end ³²P-labeled model hairpin RNA in vitro generated one major product and one minor product, the lengths of which corresponded to cleavage at sites A and B, respectively (Fig. 3E). The predicted secondary structure of the hairpin was confirmed by analyzing the cleavage patterns of the model hairpin RNA after RNase T1 digestion. The other minor cleavage products observed might have resulted from the intrinsic ability of RNase III to randomly cleave RNA transcripts at high RNase III concentrations (8–10). We also synthesized a 3′-end ³²P-labeled model hairpin and performed an in vitro cleavage assay. The results confirmed that RNase III cleaves the RNA at sites A and B (Fig. 3F).

RNase III cleavage determines proP mRNA stability in vivo.

To test whether RNase III cleavage regulates the stability of proP mRNA, we randomly introduced nucleotide substitutions at cleavage site B in the proP overexpression plasmid (pProP4) and analyzed five of these clones (Fig. 4A). Various levels of ProP protein expression were observed among the clones (Fig. 4B). To examine the relationship between the steady-state levels of proP mRNA and ProP protein expression, total RNA was isolated from these clones and a Northern blot assay was performed to measure the steady-state levels of proP mRNA (Fig. 4C). The increase in proP mRNA abundance resulting from nucleotide substitutions at cleavage site B accordingly affected ProP protein expression levels. Clone 5 showed the most dramatic increase in proP mRNA abundance (6.65-fold increase). This clone contained two nucleotide substitutions (C−111A and C−110G) that completely disrupted the base pairing between cleavage sites A and B. To test whether the increased steady-state levels of mutant proP mRNA were a consequence of decreased RNase III cleavage, we performed a primer extension analysis. The results demonstrated alterations in the abundance and cleavage patterns of the cDNAs produced from the mutant proP mRNA to those produced from the wild-type proP mRNA (Fig. 4D). cDNA bands corresponding to cleavage site B in the mutant proP mRNA of clones 1, 2, and 3, which contain C−110U, C−111U C−110A, and C−110G, respectively, exhibited a shift of one or two nucleotides. Clones 4 and 5, which contain the C−111U C−110G and C−111A C−110G mutations, respectively, did not show cleavage products from the B site. Cleavage products from the A site in mutant proP mRNA were absent from clone 5. Although the effects of these mutations on RNase III-mediated cleavage of proP mRNA differed, the ratios of the total intensity of the cDNA bands corresponding to the RNase III cleavage products to that of the putative P1 transcriptional initiation site were inversely correlated with the steady-state levels of proP mRNA and ProP protein expression levels. In addition, analysis of mutant proP mRNA from clones 4 and 5 indicated that RNase III cleavage at site B (not A and B) is the rate-limiting step for proP mRNA degradation in vivo, as abolishment of RNase III cleavage activity at site B was sufficient to stabilize proP mRNA (Fig. 4D). Next, the half-lives of these proP mRNAs were measured to test whether the increased steady-state levels of proP mRNA are a consequence of increased proP mRNA stability. Wild-type proP mRNA and a mutant proP mRNA transcribed from clone 4 were used for these experiments since this mutant mRNA appears to form a stem-loop with a stability similar to that of the wild-type mRNA (ΔG = −70.36 versus −77.81 kcal/mol), and it exhibited a significant increase in mRNA abundance. The half-life of the mutant mRNA was 2.7-fold higher than that of the wild type (∼0.75 min versus ∼2.00 min), indicating a good correlation between proP mRNA abundance and stability. These results indicated that RNase III cleavage activity at the stem-loop is largely responsible for proP mRNA abundance.

FIG 4.

Introduction of mutations at the putative RNase III cleavage sites inhibits RNase III-mediated cleavage of proP mRNA. (A) Secondary structures of the proP model hairpin RNAs containing the nucleotide substitutions of each mutant proP mRNA are shown. The mutated nucleotides are those within the rectangles. WT, wild type. (B) Effects of RNase III cleavage site mutations on ProP protein levels. E. coli BL2014 strains harboring either pProP4 or one of five mutant clones with the indicated substitution(s) (mutant clone 1, C−111U; mutant clone 2, C−110U and C−111A; mutant clone 3, C−111G; mutant clone 4, C−110U and C−111G; or mutant clone 5, C−110A and C−111G) were grown as described in the legend to Fig. 2B. The amounts of ProP, RNase III, and ribosomal protein S1 were analyzed by Western blot assay. (C) Effects of RNase III cleavage site mutations on steady-state levels of proP mRNA, as determined by Northern blot assay. Total RNAs were prepared from strain MG1655 cells harboring either pProP4 or one of five mutant clones grown as described above. The abundance of proP mRNA was measured by Northern blotting with a random probe internally radiolabeled with [α-³²P]dCTP. The abundance of M1 RNA, the RNA component of RNase P, was measured with the 5′-end ³²P-labeled M1 primer and served as an internal standard for evaluation of the total amount of RNA in each lane. (D) Effects of substitutions in the RNase III cleavage site of proP mRNA on RNase III cleavage in vivo. Total RNA was prepared from the cultures described above and analyzed as described in the legend to Fig. 3. (E) Effects of mutations in the RNase III cleavage sites on proP mRNA decay. Plasmid pProP4-mutant clone 4 (carrying C−111U and C−110G) was chosen since it had a wild-type-like hairpin loop and increased steady-state levels. BL2014 strains harboring either pProP4 or pProP4-mutant clone 4 were grown in LB medium supplemented with 0.3 M NaCl at 37°C to an OD₆₀₀ of 0.6. Total RNA samples were prepared from the cultures at 0, 30, 60, 120, and 240 s after the addition of rifampin (1 mg ml⁻¹) and separated on 1.2 M agarose gels containing 0.6 M formaldehyde. The abundance of proP mRNA was measured as described for panel C. The abundance of 16S rRNA was measured with a 5′-end ³²P-labeled 16S-1195R primer and served as an internal standard for evaluation of the total amount of RNA in each lane. Two (left panel) or 7 (right panel) ng of the synthetic proP mRNA used in Fig. 3C was loaded into the first lane of the blot.

Osmoregulation of proP expression by RNase III.

Previous studies have demonstrated the RNase III activity-mediated osmoregulation of the expression of several osmosensing factors, including Bdm, ProU, and BetT (8, 9, 11). On the basis of these results, we investigated the relationship between the osmoregulation of RNase III activity and proP expression levels by measuring the half-lives of proP mRNAs under high-osmolarity conditions. The results of a semiquantitative RT-PCR analysis showed that in wild-type E. coli, the half-life of proP mRNA under high-osmolarity conditions was approximately twice as long as that under normal conditions (1.5 min versus 3.2 min) (Fig. 5A). Steady-state levels of proP mRNA were also approximately 1.4 times as high under high-osmolarity conditions, showing a correlation between the half-life and abundance of proP mRNA. In rnc mutant cells, the half-life of proP mRNA also differed slightly under high-osmolarity conditions (3.4 min versus 4.4 min under normal and high-osmolarity conditions, respectively), and the steady-state levels of proP mRNA were approximately 1.1 times as high under high-osmolarity conditions (Fig. 5A). The increased steady-state levels of proP mRNA observed in rnc mutant cells under high-osmolarity conditions may reflect the activation of proP mRNA transcription from P2, which does not contain RNase III cleavage sites, as has been previously shown (15, 18, 19). A small decrease in the half-life of proP mRNA in rnc mutant cells under high-osmolarity conditions suggests that there may be some rnc-independent components of the regulation of proP mRNA degradation. One such candidate is ProQ, an RNA chaperone protein that affects the expression and activity of ProP via an unknown mechanism (21–24). We tested this possibility by measuring the half-life of proP mRNA in response to hyperosmotic stress in the presence and absence of functional ProQ. The results did not indicate ProQ involvement in the RNase III-mediated posttranscriptional regulation of proP expression, as ProQ-dependent changes in the half-life of proP mRNA were not observed (Fig. 5B).

FIG 5.

Effects of RNase III activity on proP under conditions of high osmotic stress. (A) Effects of osmotic stress on the half-life of proP mRNA. Total RNA was isolated from MG1655 and BL2012 cultures grown under the conditions described in the legend to Fig. 1A, except that they were grown until they reached an OD₆₀₀ of 0.6, following which rifampin (1 mg ml⁻¹) was added. WT, wild type. (B) Effects of proQ expression on proP mRNA abundance. The effects of proQ on the half-life of proP mRNA under osmotic stress are shown. Total RNA was isolated from the cultures of RM2 and WG1072 (RM2ΔproQ756::kan) grown under the conditions described above. (C) Effects of osmotic stress on the RNA binding capacity of RNase III. The BL2012 strain harboring pSD80-rnc-his was incubated at 37°C in M63 medium supplemented with 22 mM glucose and 0.17 M NaCl until an OD₆₀₀ of 0.5 was reached and then grown for an additional 4 h in the presence of 0.17 or 0.5 M NaCl. Subsequently, the cells were cross-linked with formaldehyde and harvested to immunoprecipitate (IP) RNase III with an anti-His tag monoclonal antibody. RNA samples were isolated and analyzed by semiquantitative RT-PCR with proP-, pnp-, and rpsO-specific primers. The amounts of RNase III were analyzed by Western blot assay. IB, immunoblot.

These results indicate that the increased steady-state levels and half-life of proP mRNA may be associated with decreased RNase III activity, as previously reported (8, 9). In order to determine the basis of the decreased RNase III activity observed under hyperosmotic conditions, we analyzed the RNA bound to RNase III by in vivo cross-linking and immunoprecipitation of RNase III (Fig. 5C). The amount of the RNase III substrates, such as the proP and pnp mRNAs, which were cross-linked to RNase III under hyperosmotic conditions was approximately 1.4- to 4-fold lower than that precipitated under normal conditions. The amount of rpsO mRNA, which is not an RNase III substrate (8, 34, 35), was not significantly different. These results indicate that the decreased RNase III cleavage activity detected under hyperosmotic conditions is related to the decreased RNA binding capacity of RNase III.

DISCUSSION

The regulation of proP expression at the transcriptional level under different osmolarity conditions has been well studied (15, 16, 18, 21, 22, 36, 37). These studies showed that the proP P1 promoter rapidly and transiently responds to alterations in the osmolarity of the culture medium due to the activities of CRP and osmoprotectants (15, 19, 38). In this study, we identified an additional proP expression regulatory pathway involving RNase III at the posttranscriptional level. Here, we demonstrated that RNase III cleavage of the stem-loop in the 5′ UTR determines the decay rate of proP mRNA synthesized from the osmoregulated P1 promoter and affects ProP protein levels and proline uptake capacity (Fig. 2 and 3). This observation led us to speculate that subtle conformational changes in the RNase III-targeted stem-loop region could be induced by regulatory factors under various osmotic stress conditions, which may alter the RNase III ribonucleolytic activity on proP mRNA. Though ProQ is known to affect the expression and activity of ProP via an unknown mechanism, it does not seem to play a role in this process (Fig. 5B). Another possibility is that the increased cellular levels of proline under conditions of hyperosmotic stress alter RNase III activity on proP mRNA. However, we were not able to detect significant changes in the in vitro RNase III activity on the model proP hairpin RNA (see Fig. S1 in the supplemental material). It is still possible that ProQ and/or proline play a role in the RNase III-mediated posttranscriptional regulation of proP expression under other conditions. When E. coli cells are exposed to high-salt conditions, transcription from the proP P1 promoter is induced for a brief period of time (∼5 min) and is then dramatically repressed after 15 min, the time point that coincides with the osmolarity-dependent binding of the CRP to a site within the proP P1 promoter (15). Our present study explains how the increased steady-state levels of proP mRNA resulting from this burst of proP transcription are maintained under hyperosmotic conditions, as RNase III cleavage of proP mRNA is downregulated in hyperosmotic media. Our analysis of the RNA bound to RNase III suggested the basis for decreased RNase III activity under hyperosmotic conditions (Fig. 5C), which did not appear to be specific to the mRNA species involved in osmoregulation. Previous studies indicated that RNase III activity in E. coli could also be regulated under other conditions, such as bacteriophage infection, cold stress, and aminoglycoside antibiotic stress (4, 32, 39). These studies identified various protein regulators of RNase III, including T7 protein kinase and YmdB. However, these regulators do not appear to be involved in the osmoregulation of RNase III activity (8) and we are currently investigating the factors involved in osmoregulation.

The importance of the regulation of RNase III activity in the expression of osmosensing genes such as proU (11) and betT (9) under various osmolarity conditions has been recently recognized. Similarly, we showed that the expression of one of these important osmosensing genes, proP, is posttranscriptionally regulated by RNase III activity. Because the regulation of osmosensing factors needs to be rapidly controlled, posttranscriptional regulation of gene expression by RNase III activity has advantages over the de novo synthesis of factors that regulate the transcription of osmosensing genes. In addition, the downregulated RNase III activity observed under high osmolarity maintains the induced expression levels of osmoresponsive genes, resulting in the survival of E. coli cells upon osmotic shock. Taken together, the results suggest that the RNase III-mediated osmoregulatory pathway is largely responsible for the modulation of the expression of typical osmosensing and osmoresponsive genes, including bdm, proU, betT, and proP. Consequently, this pathway regulates cell survival upon sudden changes in environmental osmotic pressure.