Protein Dosage of the lldPRD Operon Is Correlated with RNase E-Dependent mRNA Processing

Adjustment of gene expression is critical for proper cell function. In the case of polycistronic transcripts, posttranscriptional regulatory mechanisms can be used to fine-tune the expression of individual cistrons.

ABSTRACT

The ability of Escherichia coli to grow on l-lactate as a sole carbon source depends on the expression of the lldPRD operon. A striking feature of this operon is that the gene encoding the transcriptional regulator (LldR) is located between the genes encoding the permease (LldP) and the dehydrogenase (LldD). In this study, we report that the dosages of the LldP, LldR, and LldD proteins are not modulated on the transcriptional level. Instead, modulation of the protein dosage is correlated primarily with RNase E-dependent mRNA-processing events that take place within the lldR mRNA, leading to the immediate inactivation of lldR, to differential segmental stabilities of the resulting cleavage products, and to differences in the translation efficiencies of the three cistrons. A model for the processing events controlling the molar quantities of the proteins in the lldPRD operon is presented and discussed.

IMPORTANCE Adjustment of gene expression is critical for proper cell function. In the case of polycistronic transcripts, posttranscriptional regulatory mechanisms can be used to fine-tune the expression of individual cistrons. Here, we elucidate how the protein dosage of the Escherichia coli lldPRD operon, which presents the paradox of having the gene encoding a regulator protein located between genes that code for a permease and an enzyme, is regulated. Our results demonstrate that the key event in this regulatory mechanism involves the RNase E-dependent cleavage of the primary lldPRD transcript at an internal site(s) located within the lldR cistron, resulting in a drastic decrease in the amount of intact lldR mRNA, in differential segmental stabilities of the resulting cleavage products, and in differences in the translation efficiencies of the three cistrons.

INTRODUCTION

Escherichia coli, a facultatively anaerobic organism, is capable of growing on l-lactate as a sole carbon source. This is achieved by the activity of the flavin mononucleotide-dependent l-lactate dehydrogenase, a membrane-associated enzyme, which oxidizes l-lactate to the central metabolite pyruvate. This enzymatic activity depends on the respiratory growth conditions; it reaches a maximum in aerobically growing cells in the presence of l-lactate (1) but is completely absent under fermentative growth conditions, even in the presence of the inducer (2).

The gene encoding l-lactate dehydrogenase is located on the lldPRD operon (formerly known as lctPRD) (3), which also includes the lldP and lldR genes, encoding an l-lactate permease and a transcriptional regulator, respectively (3). The expression of the lldPRD operon has been shown to be driven by a single promoter (4) and to be controlled by LldR and ArcA (4, 5). LldR, which belongs to the GntR family of helix-turn-helix proteins, has been proposed to activate the expression of the lldPRD operon in cells grown in the presence of l-lactate and to repress the expression of the operon in the absence of the inducer, serving as a dual transcriptional regulator (4). ArcA, on the other hand, is a constituent of the ArcBA two-component signal transduction system, which under anoxic or fermentative growth conditions initiates a phosphorelay cascade culminating in ArcA phosphorylation (ArcA-P) (6–9). ArcA-P, in turn, binds to the lldPRD promoter region (at positions –124 to –107 relative to the lldP translation initiation site) and represses lldPRD transcription (5).

An interesting feature of the lldPRD operon is that the stop codon of each upstream gene overlaps the start codon of the downstream gene in such a way that the three coding regions are arranged in three different reading frames (Fig. 1) (3). Another unusual feature of this operon is the order of the three overlapping genes; the gene encoding the regulator protein is located between the genes that code for a permease and an enzyme. This genetic organization constitutes a paradox, since the regulator protein is expected to be required in small amounts relative to those of the two other, nonregulatory proteins. Here, we address the question of how the dose of the product of each gene of the lldPRD operon is modulated. Our results demonstrate that mRNA processing, differential segmental stability, and differences in translation efficiencies ensure the fine-tuning of the expression of the individual genes of this operon. A regulatory model is discussed.

FIG 1.

Genetic structure of the lldPRD operon. Open arrows indicate the positions and directions of the three structural genes. Gene lengths are given in parentheses. Transcriptional elements in the upstream region of the structural genes (such as promoter –35 and –10 sequences), the transcription initiation site, and the ArcA-P and LldR binding sites are indicated. The relative position of the RBS of each gene is shown. In the text boxes, the ribosome binding sites and the start codons are underlined, and the stop codons of the overlapping genes are shown in boldface and labeled with asterisks.

RESULTS

The protein dosage of the lldPRD operon is not controlled at the transcriptional level.

The genetic organization of the lldPRD operon possesses the interesting feature of having a transcriptional regulator (lldR) located between genes encoding a permease (lldP) and an enzyme (lldD) (Fig. 1). However, this presents a paradox, because transcriptional regulator proteins are expected to be required in very small amounts in comparison to structural or enzymatically active proteins. To ensure that this was the case, chromosomal fusions leading to C-terminal 3×FLAG tags were generated for LldR and LldD, and the protein levels in cells growing under inducing conditions (i.e., aerobic growth and the presence of l-lactate) were compared by Western blot analysis using anti-Flag antibodies. It was found that the LldD protein was at least 500-fold more abundant than the LldR protein (Fig. 2A). The longevities of the LldD and LldR proteins were then probed (Fig. 2B) in order to find out whether this could explain the difference in protein amounts. No significant difference in stability was observed between LldD and LldR (Fig. 2B), discarding this possibility. Furthermore, previous proteomic analyses have revealed that LldD is about 13- to 15-fold more abundant than LldP (10, 11), providing further support to the uneven stoichiometry of the protein products of the lldPRD operon. We therefore speculated that the product doses of the lldPRD operon might be controlled at the transcriptional or translational level. The first possibility implies the presence of functional promoters within the lldPRD operon. Therefore, the number of promoters controlling the expression of the three lld genes was scrutinized. To this end, lacZ transcriptional fusions were constructed, one carrying the upstream region of the lldPRD operon that includes the promoter (strain ECL5002) (12) and another carrying the DNA sequence comprising the lldP gene (from the 10th codon), the entire lldR coding sequence, and the first 65 nucleotides (nt) of lldD (strain IFC5028). The lacZ fusion-carrying strains were grown aerobically in the absence or presence of l-lactate, and β-galactosidase activity was measured at the mid-exponential-growth phase. It was found that reporter expression was induced by l-lactate only in the strain carrying the construct that includes the promoter upstream of lldP (Fig. 2C), in accordance with previous studies (4, 5). Thus, transcription of all three lld genes relies on a single promoter located upstream of the lldP gene, and as a consequence, a single polycistronic lldPRD mRNA is to be expected.

FIG 2.

Unequal protein production from the lldPRD operon is not controlled at the transcriptional or translational level. (A) Levels of the LldR–3×FLAG (31.9-kDa) and LldD–3×FLAG (45.4-kDa) proteins in E. coli cells grown aerobically on M9 minimal medium supplemented with 0.2% Casamino Acids and 20 mM l-lactate, as determined by Western blot analyses using anti-Flag antibodies. Cell extracts from strains IFC5032 and IFC5033 were loaded in lanes 1 and 2, respectively. The total protein content loaded in lane 1 is 100 times higher than that loaded in lane 2. The positions of standard protein markers are shown on the left. (B) Stabilities of the LldR–3×FLAG and LldD–3×FLAG proteins after inhibition of protein synthesis with chloramphenicol. (Left) Western blot analysis using anti-Flag antibodies. (Right) Quantitative determination of LldR and LldD from Western blot experiments plotted as a function of time. (C) Probing for possible internal promoters in the lldPRD operon. (Left) Schematic representation of the lldPRD operon and the fragments fused to lacZ to construct strains ECL5002 and IFC5028. (Right) Strains ECL5002 and IFC5028 were grown aerobically on M9 minimal medium supplemented with 0.2% Casamino Acids either in the presence (+) or in the absence (−) of 20 mM l-lactate. At the mid-exponential phase of growth, cells were harvested and assayed for β-galactosidase activity, expressed as a percentage of the activity of strain ECL5002 grown in the presence of the inducer. Averages from four independent experiments are presented, and standard deviations (error bars) are indicated. (D) Translation initiation rates of the three lldPRD genes as determined by lacZ translational fusions. (Left) Schematic representation of the lldPRD operon and the fragments fused to lacZ for strains IFC5029, IFC5030, IFC5031, and IFC5032. The relative position of the RBS of each gene or lacZ fusion is indicated by a triangle. (Right) Strains IFC5029, IFC5030, IFC5031, and IFC5032 were grown aerobically on M9 minimal medium supplemented with 0.2% Casamino Acids and 20 mM l-lactate as the inducer. At the mid-exponential phase of growth, cells were harvested and assayed for β-galactosidase activity, expressed as a percentage of the activity of strain IFC5031. Averages from four independent experiments are presented, and standard deviations (error bars) are indicated. (E) Analysis of the lld transcript by Northern blotting. Strain MC4100 was incubated on M9 minimal medium supplemented with 0.2% Casamino Acids in the presence (+) or absence (−) of 20 mM l-lactate, and total RNA isolated from samples that were harvested during mid-exponential growth was probed for the lld transcript using an lldD-specific DNA probe. A blot representative of results from three independent experiments is shown. The positions of standard RNA markers are shown to the left of the blot. The lower panel shows the ethidium bromide-stained rRNA bands of the corresponding Hybond membrane, used as a loading control.

The protein dosage of the lldPRD operon is partially controlled at the translational level.

To assess the possibility that the differential protein production of the lldPRD operon was mediated at the level of translation, the strengths of the three ribosome binding sites (RBS) were compared. To this end, three different ′lacZ translational fusions, lldP′-′lacZ, lldR′-′lacZ, and lldD′-′lacZ, were constructed and placed on the λ attachment site of the MC4100 chromosome, generating the corresponding strains IF5029, IFC5030, and IFC5031. All constructs contained the lldPRD native promoter and the entire sequence upstream of the gene of interest, including the RBS and the first two codons of the respective cistron (Fig. 2D). The strains were grown aerobically in the presence of the inducer l-lactate, and β-galactosidase activity was measured at the mid-exponential-growth phase (Fig. 2D). It was found that the expression of the reporter driven by the RBS of lldP was approximately 2-fold lower than the expression of those driven by the RBS of lldR and lldD. However, no significant difference was observed between the RBS of lldR and the RBS of lldD (Fig. 2D). As a possible alternative cause of the differential concentrations of the lldPRD operon protein products, we analyzed the bias in the preferential use of codons in their genes. To this end, we evaluated the codon adaptation index (CAI) (13), using the codon usage table of E. coli K-12 as a reference set (14). The CAI values obtained for lldD, lldP, and lldR were 0.790, 0.745, and 0.737, respectively. Interestingly, the CAI value of the first third of the lldR coding sequence was found to be lower than that of the rest of the lldR coding sequence (0.729 and 0.756, respectively), suggesting a lower translation elongation rate along the first section of lldR. Although the relative order of the CAI values for the genes of the lldPRD operon corresponds to the relative abundances of their respective protein products, and similar differences in CAI values have been shown to contribute to uneven stoichiometries of proteins from polycistronic transcripts (15), these are not sufficient to explain the vast difference observed between the LldR and LldD proteins (Fig. 2A). Thus, additional mechanisms must be involved in the regulation of the expression of the lldPRD operon.

RNase E-dependent processing of the lldPRD polycistronic mRNA.

To explore the possibility that the individual protein dosage is controlled by posttranscriptional events such as mRNA processing, we constructed a fourth ′lacZ translational fusion. This translational fusion, which is an lldR′-′lacZ fusion similar to that of IFC5030 except that it includes almost the entire lldR coding sequence (positions +1 to +759 relative to the lldR start codon), was placed on the λ attachment site of the MC4100 chromosome, generating strain IFC5032 (Fig. 2D). IFC5032 and IFC5030 were grown aerobically in the absence or presence of l-lactate, and β-galactosidase activity was measured at the mid-exponential-growth phase. It was found that reporter expression of strain IFC5032 was 10-fold lower than that of IFC5030 (Fig. 2D). It was therefore tempting to speculate that the lldR cistron may contain sites of regulation at the posttranscriptional level. This is supported by a previous report, where Northern blot analysis revealed that although the lldD gene was induced by l-lactate, the full-length lldPRD mRNA was not detected (3). To confirm this result, total RNA was extracted from MC4100 cells grown aerobically in M9 minimal medium supplemented with 0.2% Casamino Acids in the absence or presence of the inducer, and Northern blot analysis was carried out using an lldD-specific probe (Fig. 2E). In agreement with the previous report (3), the lldD gene was strongly induced at the presence of l-lactate, but the full-length lldPRD mRNA, expected to be about 3,600 nt, was not detected under any growth condition (Fig. 2E). Thus, it seems reasonable to suggest that posttranscriptional mRNA processing might be involved in regulating the individual transcription product doses of the lldPRD operon.

In E. coli, mRNA processing and stability rely on the action of numerous RNases (16). Of these, RNase E, RNase III, and RNase P are responsible for most of the primary endoribonucleolytic RNA-processing events. Therefore, the fate of the lldPRD messenger was examined in cells deficient in RNase E [rne(Ts)], RNase P [rnp(Ts)], RNase III (rnc::Tn10), or RNase G (rng::cat) (17–22). The rne(Ts) and rnp(Ts) mutant alleles, producing conditional temperature-sensitive enzymes, were employed because RNase E and RNase P are essential for cell viability. Also, the rng mutant was used because its product, RNase G, is a paralog of RNase E (23), and the two enzymes have been reported to have overlapping functions (23). Northern blot analyses using the specific lldP and lldD probes and RNA isolated from the RNase-deficient mutants and their isogenic strains were carried out. It was observed that after a shift to a nonpermissive temperature (43°C), a band of ∼ 3,600 nt, corresponding to the full length lldPRD messenger, accumulated with time in the rne(Ts) mutant but not in the wild-type strain (Fig. 3). In contrast, no effect was observed in the rnc and rng mutant strains or in the rnp(Ts) mutant strain after a shift to the nonpermissive temperature (see Fig. S1 in the supplemental material). Thus, it seems reasonable to conclude that RNase E is responsible for processing the lldPRD mRNA into two mRNA species, carrying, respectively, the intact lldP and lldD mRNAs, by endoribonucleolytic cleavage(s) within the lldR mRNA. As a consequence, the amount of intact lldR mRNA to be translated decreases drastically, and therefore, only small amounts of LldR are produced.

FIG 3.

Analysis of lldPRD mRNA in an rne⁺ and an rne(Ts) strain by Northern blotting. Strains CH1827 (rne⁺) and CH1828 [rne(Ts)] were grown aerobically at 30°C. At an OD₆₀₀ of 0.4, cells were shifted to 43°C, samples were withdrawn at the indicated times, and total RNA was isolated. Shown are autoradiograms of blots with the specific lldP probe (A) and the lldD probe (B). The positions of standard RNA markers are shown on the left. Arrowheads indicate the positions of the lldP, lldD, and lldPRD transcripts. Lower panels show the ethidium bromide-stained rRNA bands of the corresponding Hybond membranes, used as loading controls.

In an effort to probe possible cleavage sites within the lldR mRNA, primer extension mapping was employed. RNA isolated from the wild-type strain or the rne(Ts) mutant was used together with a primer complementary to positions 528 to 549 of the lldR transcript with respect to the start of lldR translation. Two reverse transcriptase products whose intensities were significantly lower in the RNase E mutant at the nonpermissive temperature (43°C) than in the wild-type strain or the RNase E mutant at 30°C were identified within a portion of the coding region of the lldR mRNA (at positions +381 and +389 with respect to the start of lldR translation) (Fig. 4A). Notably, the sequences of these sites are reminiscent of the proposed RNase E recognition/cleavage site, which comprises single-stranded AU-rich regions (Fig. 4B) (24–30). Additionally, a larger primer extension product was detected only in the RNase E mutant at the nonpermissive temperature, at position +341 relative to the lldR start codon (Fig. 4). However, we were not able to discern whether this product is the site for the cleavage of an RNase different from RNase E or whether it represents an incomplete reverse transcriptase product. Nevertheless, it can be concluded that RNAse E is responsible for processing the lldPRD mRNA within the lldR open reading frame, thereby hampering the translation of lldR.

FIG 4.

Primer extension analysis of sites of RNase E cleavage into the lldR transcript. (A) Total RNAs from strains CH1827 (rne⁺) and CH1828 [rne(Ts)], grown aerobically in the presence of 20 mM l-lactate at 30°C and 15 min after a shift to a nonpermissive temperature (43°C), were prepared and used for primer extension analysis. A 5′-end-labeled primer specific for the lldR transcript was annealed to total RNA and extended by reverse transcriptase. The resulting cDNA was resolved in an 8% polyacrylamide gel alongside a DNA sequencing ladder. Positions of relevant primer extension products are marked on the right. The relevant portion of the mRNA sequence (corresponding to the complementary sequence of the DNA sequencing ladder) is presented on the left, and the identified 5′ ends are shown in boldface. (B) Schematic representation of the lldPRD operon and RNase E cleavage sites identified by primer extension analysis. Sites of cleavage into the lldR transcript at positions +381 and +389 with respect to the start codon of lldR are shown (▴), and predicted RNase E recognition sequences, which fulfill the proposed RNase E target motif RN↓WUU (where R stands for G or A, W stands for A or U, and N stands for any nucleotide) (30), are marked in boldface. The 5′ end corresponding to the larger primer extension product detected in the RNase E mutant, at position +341 with respect to the lldR start codon, is also indicated (Δ).

Differential mRNA stability of lldP and lldD mRNA species.

As mentioned above, endoribonucleolytic cleavage(s) within lldR should result in a drastic decrease in the amount of intact lldR mRNA, and therefore, only a small fraction of the transcribed lldR will be translated to a functional LldR protein. Moreover, the two cleavage products, carrying the intact lldP and lldD, are expected to be present at equivalent concentrations and to produce similar amounts of the LldP and LldD proteins. However, previous proteomic analyses have indicated that LldD is 13- to 15-fold more abundant than LldP (10, 11), a difference that cannot be fully explained by the rate of initiation of translation (Fig. 2C) or by the bias in the preferential use of codons. Therefore, we hypothesized that differential stability of the two mRNA species may also contribute to the modulation of the LldP and LldD protein doses. To test this, the longevities of lldP and lldD mRNAs were determined. E. coli MC4100 cells were grown in M9 minimal medium supplemented with Casamino Acids and l-lactate as the inducer. At mid-exponential growth, transcription was blocked by rifampin, which inhibits the action of the DNA-dependent RNA polymerase, and aliquots were withdrawn at various time intervals. Total RNA was isolated, and the half-life of the mRNA was analyzed by Northern blotting using lldP- and lldD-specific probes (Fig. 5). The half-life of the lldP mRNA was found to be ∼2.5 min, whereas the half-life of the lldD mRNA was about 7 min (Fig. 5). Thus, differential transcript stability should also contribute to the modulation of the intracellular amounts of the LldP and LldD proteins.

FIG 5.

Northern blot analysis of lldP and lldD mRNA stabilities. E. coli MC4100 was grown in LB medium supplemented with 20 mM l-lactate at 37°C to the mid-exponential phase prior to the addition of rifampin. Samples were then harvested at the indicated times, and RNA was extracted as described in Materials and Methods. Ten micrograms of total RNA from each time point was used for Northern blot analysis. (Left) Autoradiograms of blots using the specific lldP (top) or lldD (bottom) probe. Lower panels show the ethidium bromide-stained rRNA bands of the corresponding Hybond membranes. (Right) Semilogarithmic plot of lldP and lldD mRNA decay. The calculated half-lives were 6.8 min for lldD and 2.5 min for lldP.

Are there other transcription factors arranged in a genomic organization similar to that of the lldPRD operon?

The above observations raised the question of whether the genomic organization of the lldPRD operon is unique or whether other transcription factor (TF) cistrons are also present in the middle of their respective operons. To answer this question, we used RegulonDB (31) to obtain a list of the E. coli K-12 genes that encode TFs (301 annotated TFs) and ProOpDB (32) to obtain the set of predicted E. coli K-12 operons. By employing a Perl program that was developed ad hoc, the relative positions of the genes encoding TFs within the operons were located and accounted for. The complete list of E. coli operons encoding TFs is presented in Fig. S2 and Table S3, and the results are summarized in Fig. 6. Our analysis revealed that 143 of the TFs are encoded by monocistronic mRNAs, and 158 are encoded by mRNAs located on polycistronic units. In the latter group, 78 TF genes were located at the beginning of the transcript, and 61 were located at the end; only in 19 cases was the TF located internally, in a manner similar to that of the lldPRD operon (Fig. 6; also Fig. S2). Among these, six TFs are constituents of the family of two-component signal transduction systems and are cotranscribed with their cognate partners (Fig. S2), and two operons, gutM-srlR-gutQ and marR-marA-marB, contain TF-encoding genes in both the first and the second positions. Finally, in six operons the presence of internal promoters has been reported (Fig. S2). Thus, of the 301 TFs that exist in E. coli K-12, only 5 have a setting similar to that of the lldPRD operon. It would be interesting to find out whether posttranscriptional mechanisms, similar to those described here for the lldPRD mRNA, control the expression of these TFs.

FIG 6.

Gene organization of the 301 predicted TFs encoded in the E. coli genome. Open arrows represent genes encoding predicted TFs, and shaded arrows represent other genes. “(n)” stands for a number from 1 to 15. The number of TFs in each organizational group (TFs in monocistronic mRNA units or at the beginning, end, or middle of polycistronic mRNA units) is given on the left, and the number of TFs that are part of a two-component regulatory system (TCS), and share an operon with the cognate histidine kinase, is given in parentheses.

DISCUSSION

The rate of synthesis of a protein depends on the amount of its mRNA, which, in turn, depends on its rates of synthesis and decay. However, numerous prokaryotic genes that encode related functions are organized in polycistronic operons. Such an organization is expected to facilitate the coordinated expression of the different cistrons of the operon. Nonetheless, the protein products of different genes of an operon may be produced in different molar quantities. In these cases, additional steps of regulation are needed to fine-tune the expression of specific genes within an operon. For example, the presence of internal promoters allows differential transcription levels of the genes of an operon (33, 34). In other cases, the translation efficiencies of individual cistrons in a polycistronic transcript can determine the unequal protein dosage (15, 35). Alternatively, differential protein production from the genes of a polycistronic mRNA can occur after the processing of the primary transcript by RNase-dependent cleavages within intergenic regions that give rise to mRNA molecules with different segmental stabilities (36–45).

Yet another mode of modulation of gene expression emerged from our study on the lldPRD operon. The model proposed here for the processing events controlling the molar quantities of the proteins encoded in the lld operon can be described as follows. RNase E promptly cleaves the primary lldPRD transcript at an internal site(s) located within the lldR cistron. This event results in a drastic decrease in the amount of intact lldR mRNA, and therefore, only small amounts of functional LldR are produced. Endoribonucleolytic cleavages within lldR result in two mRNA species with differential stabilities, carrying the intact lldP (half-life [t_1/2], ≈2.5 min) and lldD (t_1/2, ≈7 min) cistrons. The difference in longevity between the lldP and lldD mRNA species is reflected in the steady-state amounts of the two messengers.

Finally, another important issue that can be included in our model is the accessibility to RNase E of its target sites within the lldR mRNA. It is well known that while an mRNA is being translated, it is covered by ribosomes and is thereby protected from endoribonucleolytic cleavages by different RNases (46, 47). In the case of the lldPRD operon, our results from translation initiation experiments did not show that the translation initiation efficiency of lldR was impaired relative to that of the lldP or lldD gene. Notwithstanding, the codon biases of the three cistrons, determined by the codon adaptation index (CAI) (13), revealed that the translation elongation rate of lldR may be lower than those of lldP and lldD. Most importantly, the first third of the lldR coding sequence was found to have a lower CAI value than the rest of the operon. This would result in a low rate of translation elongation and would contribute to ribosome crowding in this mRNA region. Indeed, the correlation between low translation elongation efficiencies, inferred from lower CAI values, and elevated ribosome densities has been demonstrated previously (15, 48). As a consequence, one can speculate that the fragment corresponding to the second third of lldR in the polycistronic unit should have a lower ribosome density and therefore remain accessible to RNase E.

Thus, a mechanism that is able to efficiently modulate the molar quantities of the three proteins encoded on the lldPRD operon is proposed. This mechanism involves (i) the intrinsically lower rate of translation and the ribosome crowding of the first third of lldR due to its lower biased codon usage compared to the rest of the operon (this step is supported by the CAI values but remains to be demonstrated experimentally), (ii) concomitant exposure of the mRNA segment corresponding to the second third of the lldR cistron to RNase E cleavage, an event that leads to a nearly immediate inactivation of the lldR mRNA, (iii) differential stabilities of the two cleavage products generated, which carry, respectively, the lldP and the lldD mRNAs, and (iv) lower translation efficiency of lldP than of lldD, due to lower translation initiation and elongation rates.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

The bacterial strains used in this study are in listed in Table S1 in the supplemental material. Strains were routinely grown in lysogeny broth (LB) medium or in M9 defined minimal medium (49) supplemented with 0.2% (wt/vol) Casamino Acids (casein acid hydrolysate). When appropriate, the medium was supplemented with 20 mM l-lactate as an inducer, and antibiotics were added to the medium at the following concentrations: kanamycin, 50 μg/ml; ampicillin, 100 μg/ml; and tetracycline, 10 μg/ml. For mRNA half-life measurements, the transcriptional inhibitor rifampin was used at a final concentration of 200 μg/ml. For assays to determine protein stability, chloramphenicol (10 μg/ml) was used to inhibit translation. Strains IFC5033 and IFC5034, expressing FLAG-tagged LldR and LldD, respectively, were created using the λ Red recombinase method, as described previously (50) (the primers used are listed in Table S2). The transcriptional lacZ fusion in strain IFC5028 was generated using primers LacZPDfw and LacZDrev (Table S2). The PCR product was digested with the restriction enzymes EcoRI and BamHI and was ligated to the corresponding sites of plasmid pRS415 (51), and the construct was transferred to the att(λ) site of the chromosome of MC4100 via phage λRS45 as described elsewhere (51). Strains IFC5029, IFC5030, IFC5031, and IFC5032, carrying different lacZ translational fusions (Fig. 2C), were obtained using the CRIM system (52). Briefly, plasmid-borne translational lacZ fusions were constructed by cloning the PCR-amplified fragments (the primers used are listed in Table S2) into the PstI/BamHI site (for strains IFC5029 and IFC5032) or the PstI/EcoRI site (for strains IFC5030 and IFC5031) of plasmid pLFT (53). Constructs were integrated into the chromosome at the att(λ) site using the helper plasmid pINT-ts (54) as described previously (52).

Western blotting.

Cells of strains IFC5033 and IFC5034 were grown aerobically at 37°C in LB medium supplemented with l-lactate and were harvested by centrifugation at an optical density at 600 nm (OD₆₀₀) of 0.6. Cell pellets were resuspended in 60 μl of lysis buffer (50 mM Tris-HCl, 4% SDS [pH 6.8]) and were boiled for 5 min. Cell extracts from strain IFC5034 were diluted 1:100 with lysis buffer. Aliquots of 10 μl were separated by SDS-PAGE (12% polyacrylamide gel), and the proteins were transferred to a Hybond-ECL filter (Amersham Biosciences). The filter was equilibrated in TTBS buffer (25 mM Tris, 150 mM NaCl, and 0.05% Tween 20) for 10 min and was incubated in blocking buffer (1% milk in TTBS) for 1 h at room temperature. Monoclonal antibodies against the FLAG epitope were added at a dilution of 1:10,000 and were incubated for 1 h at room temperature. The bound antibody was detected by using an anti-mouse IgG antibody conjugated to horseradish peroxidase (1:10,000 dilution) and the Immobilon Western detection system (Millipore). Protein bands were quantified using ImageQuant software (Molecular Dynamics).

β-Galactosidase activity assay.

For the β-galactosidase activity assays, cells were cultured aerobically in 10 ml of M9 mineral medium supplemented with 0.2% Casamino Acids, contained in 250-ml baffled flasks at 37°C with shaking (300 rpm), and harvested at an OD₆₀₀ of 0.6. Where indicated, l-lactate was added at a final concentration of 20 mM. β-Galactosidase activity was assayed and calculated in Miller units as described previously (55).

RNA extraction and Northern blot analysis.

RNA was obtained by the hot phenol extraction method as described previously (56). For Northern blots, 10 μg of RNA was fractionated on denaturing 1.2% agarose-formaldehyde gels and transferred to nitrocellulose membranes (Amersham XL) by capillary transfer by using 20× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate). Membranes were cross-linked using a UV-cross-linking device (Stratalinker; Stratagene) and were prehybridized for 3 h at 42°C in a buffer containing 5× Denhardt’s solution (57), 5× SSC, 0.2% SDS, 50% formamide, and 250 mg of sheared salmon sperm DNA per ml. Probes were denatured at 90°C for 5 min, placed on ice, added to the blocking solution, and incubated at 42°C overnight. For generating the lldP- and lldD-specific probes, primer pairs lldP1/lldP2 and lldD1/lldD2 (Table S2) were used in PCRs with plasmid pLCT2 (3) as the template. The PCR products were introduced by T/A cloning into the pGEM-T Easy vector (Promega), generating plasmids pGEM-ProbeP (for lldP) and pGEM-ProbeD (for lldD). Each probe was obtained by digesting either plasmid pGEM-ProbeP or plasmid pGEM-ProbeD with EcoRI. The fragments were separated on agarose gels and purified using the Qiagen Agarose purification kit. Probe labeling was performed by using [α-³²P]dCTP and the RadPrime kit (Invitrogen) according to the manufacturer’s instructions. Membranes were washed twice with 50 ml of 2× SSC and 0.1% SDS at 37°C and twice with 0.2× SSC and 0.1% SDS at 42°C. Images were obtained using phosphorimager screens and the Typhoon image scanner (Amersham), and radioactive bands were quantified using ImageQuant software (Molecular Dynamics).

Primer extension analysis to identify RNase E cleavage sites.

Strains CH1827 (rne⁺) and CH1828 [rne(Ts)] were grown aerobically in LB medium supplemented with 20 mM l-lactate at 30°C to an OD₆₀₀ of 0.4. Cultures were then shifted to 43°C for 15 min before cells were collected and total RNA was isolated by the hot phenol extraction method (56). Forty micrograms of RNA was annealed to 50 pmol of primer PE2-lldPRD labeled with ³²P at its 5′ end (Table S2), complementary to positions +528 to +549 of the lldR transcript with respect to the start of lldR translation, by heating to 95°C for 3 min and then slowly cooling to 45°C. The primer was extended with Maxima H Minus reverse transcriptase (Thermo Scientific) at 55°C for 30 min, and the products were then purified by 1-butanol precipitation and analyzed by electrophoresis in 8% polyacrylamide–8 M urea gels alongside sequencing ladders. Sequencing ladders were generated by extending the same primer (PE2-lldPRD) with plasmid pLCT2 (3) as the template.