Regulation of Ribosomal Protein Operons rplM-rpsI, rpmB-rpmG, and rplU-rpmA at the Transcriptional and Translational Levels

ABSTRACT

It is widely assumed that in the best-characterized model bacterium Escherichia coli, transcription units encoding ribosomal proteins (r-proteins) and regulation of their expression have been already well defined. However, transcription start sites for several E. coli r-protein operons have been established only very recently, so that information concerning the regulation of these operons at the transcriptional or posttranscriptional level is still missing. This paper describes for the first time the in vivo regulation of three r-protein operons, rplM-rpsI, rpmB-rpmG, and rplU-rpmA. The results demonstrate that transcription of all three operons is subject to ppGpp/DksA-dependent negative stringent control under amino acid starvation, in parallel with the rRNA operons. By using single-copy translational fusions with the chromosomal lacZ gene, we show here that at the translation level only one of these operons, rplM-rpsI, is regulated by the mechanism of autogenous repression involving the 5′ untranslated region (UTR) of the operon mRNA, while rpmB-rpmG and rplU-rpmA are not subject to this type of regulation. This may imply that translational feedback control is not a general rule for modulating the expression of E. coli r-protein operons. Finally, we report that L13, a primary protein in 50S ribosomal subunit assembly, serves as a repressor of rplM-rpsI expression in vivo, acting at a target within the rplM translation initiation region. Thus, L13 represents a novel example of regulatory r-proteins in bacteria.

IMPORTANCE It is important to obtain a deeper understanding of the regulatory mechanisms responsible for coordinated and balanced synthesis of ribosomal components. In this paper, we highlight the major role of a stringent response in regulating transcription of three previously unexplored r-protein operons, and we show that only one of them is subject to feedback regulation at the translational level. Improved knowledge of the regulatory pathways controlling ribosome biogenesis may promote the development of novel antibacterial agents.

INTRODUCTION

Ribosomal protein (r-protein) operons rplM-rpsI, rplU-rpmA, and rpmB-rpmG encode r-proteins L13-S9, L21-L27, and L28-L33, respectively. Despite long-term studies of the regulation of expression of ribosomal components (1–5), these three operons have never been explored and their regulation remains largely unexplained. At the same time, the products of the operons play noticeable roles in ribosome assembly and functioning, which stimulates investigation into the control mechanisms of their expression, given the importance of knowledge about ribosome biogenesis and its control.

Proteins L21 and L27 encoded by rplU-rpmA are bacterium specific (6). L27 was suggested to contribute to the function of the ribosomal peptidyl transferase center (PTC) via interactions of its N-terminal tail with both the A-site and P-site tRNAs, facilitating their accommodation in the PTC (7, 8). The L27-lacking strain has severe growth defects and is deficient in both the assembly and functioning of the 50S ribosomal subunits (7). However, a recent study of the impact of L27 on the activity of PTC argues against a key role of L27 in peptide bond formation on the ribosome, although the lack of L27 slightly reduced the average elongation rate in vitro (9). This implies that the slow-growth phenotype of the L27-deficient strain may be attributed to the impaired ribosome assembly, reducing the pool of active ribosomes inside the cell, rather than to defects in the PTC (9). As for the rpmA-encoded L21, it has been reported to interact with 23S rRNA (10), but we still do not know much about its functional importance.

The rplM-rpsI operon encodes universal ribosomal proteins L13 and S9 (6). An essential protein, L13 is an early 50S assembly component interacting with 23S RNA, and its incorporation in vivo requires the DEAD-box RNA helicase SrmB (11). It has been proposed that SrmB is necessary for organizing the L13 binding site on 23S rRNA by preventing formation of erroneous alternative structures (12). Protein S9 is also very important. The binding of S9 to 16S RNA during 30S assembly depends on S7 and additional assembly cofactors (13, 14). Though viable, the rpsI knockout strain exhibits a strong growth defect (15, 16), most likely because of the active role of S9 in maintaining the translation reading frame via contacts of its long C-terminal tail with the P-site tRNA (15, 17). In addition, the mutant producing reduced amounts of S9 displays an increased accumulation of immature 16S rRNA (17S rRNA), and thereby smaller amounts of 70S ribosomes and polysomes are formed (18).

Proteins L28 and L33 are bacterium specific (6). The gene rpmG (L33) can be deleted without any impact on growth rate, indicating the redundant role of L33 in ribosome synthesis and functions (19, 20). In contrast to L33, L28 encoded by rpmB plays an important role in ribosome assembly, as in its absence only a few 70S ribosomes are made, while 30S and 47S particles accumulate. The 47S intermediates are easily converted to mature 50S subunits when L28 is supplied from a plasmid (19, 20).

While the products of the rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons are important for ribosome assembly and/or functioning, little if anything is known about the mechanisms of their regulation. Moreover, their transcriptional start sites (TSSs) have been located only very recently by the differential RNA sequencing (RNA-seq) approach (21). In the present work, we define the features of promoter regions upstream of the TSSs. We show that transcription of the three operons is subject to negative stringent control in a ppGpp/DksA-dependent manner, most likely due to the features of the discriminator region separating the −10 promoter element from the TSS. At the same time, translational feedback regulation was observed only for the rplM-rpsI operon and was not found for rplU-rpmA or rpmB-rpmG. Finally, we show that protein L13 serves as a translational repressor that regulates rplM-rpsI expression in vivo.

MATERIALS AND METHODS

Strains and plasmids.

The strains and plasmids used in this study are listed in Table 1. Plasmids pL13/S9, pL21/L27, and pL28/L33 were constructed to express in trans the products of the three operons studied here; all of them were created by cloning the entire operon sequences flanked with their own promoters and terminators into BamHI/HindIII sites of pACYC184 (rplU-rpmA and rpmB-rpmG) or into BamHI/ClaI sites in the case of rplM-rpsI. The corresponding regions were amplified by PCR on Escherichia coli genomic DNA by using Phusion Hot Start II high-fidelity DNA polymerase (Thermo Scientific) and pairs of appropriate primers which comprised the restriction sites for subsequent cloning. The resulting plasmids were checked by sequencing. Expression of the wild-type protein S9 from plasmid pL13/S9 was also confirmed by its ability to suppress the slow-growth phenotype of the rpsI::kan strain (a generous gift of M. Bubunenko). To build the plasmid expressing only rplM, the rpsI gene was deleted from plasmid pL13/S9 by the “outward” PCR technique used to amplify the whole plasmid except for the rpsI coding sequence. The primer rpsI_left (5′-TCAGCCATTGCCTATAATCCCG) was complementary to positions −14 to +8 relative to the rpsI ATG start codon, and the primer rpsI_right (5′-ATTGGCTTCTGCTCCGGCAG) corresponded to the sequence separating the rpsI stop codon and the rho-independent terminator of the operon. The primers were phosphorylated by T4 polynucleotide kinase (New England BioLabs) and then used in PCR comprising Phusion Hot Start II high-fidelity DNA polymerase and pL13/S9 as a template. The PCR product was purified, circularized by ligation of blunt ends, and used for transformation of DH5α. The resulting plasmids were sequenced to choose the right clone.

TABLE 1.

Escherichia coli strains and plasmids used in this study

E. coli strain or plasmid	Relevant characteristic(s)	Reference or source
Strains
DH5α	Cloning host	Laboratory stock
ENS0	his, formerly HfrG6Δ12, Lac−	25
LABrplU::lacZ	ENS0 derivative bearing chromosomal rplU′-′lacZ reporter under the rplU promoter	This work
LABrplU-rpmA::lacZ	ENS0 bearing rplU-rpmA′-′lacZ reporter	This work
LABrplM::lacZ	ENS0 bearing rplM′-′lacZ	This work
LABrpmB(P2)::lacZ	ENS0 bearing rpmB′-′lacZ	This work
LABrplY::lacZ	ENS0 bearing rplY′-′lacZ	24
IBrpsO188::lacZ	ENS0 bearing rpsO′-′lacZ	40
CF1693	MG1655 relA251::kan spoT207::Cm	42
RLG8124	dksA::tet	R. L. Gourse
NB470	W3110 λred (exo-beta-gam) rpsI::kan	M. Bubunenko
Plasmids
pEMBLΔ46	pEMBL8+ derivative lacking lacZ RBS	25
pEL21TIR	pEMBLΔ46 derivative bearing the rplU′-′lacZ reporter under the rplU promoter	This work
pEL13TIR	pEMBLΔ46 derivative bearing rplM′-′lacZ under the rplM promoter	This work
pEL28TIR-P2	pEMBLΔ46 derivative bearing rpmB′-′lacZ under the rpmB P2 promoter	This work
pACYC184	Tetr Cmr; cloning vector	Laboratory stock
pL13/S9	pACYC184 derivative expressing rplM-rpsI	This work
pL13	pACYC184 derivative expressing rplM	This work
pL21/L27	pACYC184 derivative expressing rplU-rpmA	This work
pL28/L33	pACYC184 derivative expressing rpmB-rpmG under the rpmB P2 promoter	This work
pS1 (pSP261)	pACYC184 derivative expressing rpsA	43
pHfq (pTX381)	pACYC184 derivative expressing hfq	44
pL25	pACYC184 derivative expressing rplY	24
pS15	pACYC184 derivative expressing rpsO	40

Construction of fusions with the chromosomal lacZ gene.

The strategy to generate specialized strains in which expression of the chromosomal lacZ gene would be governed by the transcriptional and translational regions of the unrelated gene has been described previously (22–24). Based on this approach, we created rplU′-′lacZ, rplM′-′lacZ, and rpmB′-′lacZ chromosomal fusions. PCR products for cloning into the pEMBLΔ46 vector (25) in-frame with the lacZ coding sequence were obtained with E. coli genomic DNA by using the same forward primers as those used for constructing the above-mentioned plasmids expressing operon products, while the reverse primers were complementary to the coding regions of the first genes of the operons and comprised a restriction site appropriate for cloning. For the rplU-rpmA operon, we also created a rplU-rpmA′-′lacZ fusion, taking into account that a presumable regulatory site might be situated at the beginning of the second cistron. The resulting plasmids were checked by sequencing and then used for transformation of E. coli strain ENS0 to generate Lac⁺ clones by homologous recombination (22–25). The strains obtained were named LABrplU::lacZ, LABrplU-rpmA::lacZ, LABrplM::lacZ, and LABrpmB(P2)::lacZ (Table 1).

Site-directed mutagenesis to create a C-to-G substitution in the discriminator region of the rplU promoter.

The technique of introducing mutations in the discriminator region by using a two-step PCR was described previously (23). Here, we applied it for changing C(−5) to G in the discriminator region of the rplU promoter. In the first step, two overlapping PCR fragments were generated using pEL21TIR (Table 1) as a template and two pairs of primers. The overlapping “internal” primers comprised the desirable mutation (in bold), namely, rplU-C(−5)G-for (5′-TGGCGCCCTATTGTGAATATTTATAGC) and rplU-C(−5)G-rev (5′-CAATAGGGCGCCAATATTACGCAAAAC) (overlapping segments underlined), while the “external” primers (UPlac, 5′-GTTAGCTCACTCATTAGGCACCCC; DSlac, 5′-GGCGATTAAGTTGGGTAACGCCAGGG) corresponded to the invariant plasmid regions (26). In the second step, the two purified PCR products were mixed and amplified in the presence of UPlac and DSlac. The resulting product was treated with BamHI and HindIII and cloned in pEMBLΔ46. The presence of the desired mutation was checked by sequencing. The construction was transferred onto the chromosome of strain ENS0 as described above. The resulting strain was named LABPrplUmut::lacZ.

Induction of stringent response and isolation of total RNA.

The procedure for the induction of a stringent response and isolation of total RNA was carried out essentially as described previously (24). Strains were grown in LB medium at 37°C with vigorous shaking. At an optical density at 600 nm (OD₆₀₀) of ∼0.4 to 0.5, 2-ml aliquots of cell cultures were withdrawn and mixed with 4 ml RNAprotect bacterial reagent (Qiagen). To induce a stringent response, l-serine hydroxamate (SHX; Sigma) was added to the residual volume of the culture (0.5 mg/ml final concentration), and cultivation was continued for an additional 15 min. Aliquots of 2 ml were withdrawn at this time point and mixed with 4 ml RNAprotect bacterial reagent. Total RNA was isolated from treated and untreated probes by using the RNeasy minikit (Qiagen) according to the recommendations of the manufacturer. RNase-free DNase (Qiagen) was added to the columns during RNA extraction for 15 min to ensure the absence of DNA contamination in RNA samples. The amount of total RNA in the preparations was estimated by measuring the OD₂₆₀.

Quantification of the in vivo transcripts by RT-qPCR using LightCycler software.

To study whether the rplM-rpsI, rpmB-rpmG, and rplU-rpmA operons are stringently regulated by ppGpp/DksA during amino acid starvation, we quantified the corresponding transcripts in total RNA preparations isolated before and 15 min after SHX treatment of the wild-type, ppGpp⁰ (relA::kan spoT::Cm), and dksA::tet cells. For this purpose, we exploited the reverse transcription-quantitative PCR (RT-qPCR) approach based on using the external RNA standard (24, 27). To generate RNA standards for the calibration curves, DNA templates were prepared by PCR with the transcript-specific primers (Table 2). The forward primer comprised the T7 promoter sequence fused to the beginning of the corresponding transcript, and the reverse primer was complementary to the coding region of the first gene in the operon. The purified PCR products were transcribed in vitro with T7 RNA polymerase by using the Riboprobe System-T7 (Promega); the transcription was followed by RQ DNase digestion. The resulting RNA products were purified according to the Promega protocol, and RNA concentrations were estimated by measuring the OD₂₆₀. Serial dilutions of the individual RNA standard (from 10 ng/μl to 100 fg/μl) were then prepared in RNase-free water mixed with MS2 RNA (0.5 μg/μl final concentration). These serial dilutions (2 μl each) were used in 20-μl RT reaction mixtures with the corresponding reverse primer. In parallel with RNA standards, 1 μg total RNA isolated from each strain before and 15 min after SHX treatment was reverse transcribed with the same transcript-specific primer. RT reactions were performed in a final volume of 20 μl for 1 h at 42°C. Real-time PCR was then run with the use of a LightCycler 480 II system (Roche). Each 25-μl reaction mixture contained 5 μl 5× qPCRmix HS SYBR (Evrogen), reverse and forward primers for the transcript under study (1 μl of 5 μM solution), and 2 μl of the corresponding RT mix.

TABLE 2.

Primers used in this work for promoter regulation studies

Primer	Sequence (5′→3′)	Position descriptiona
rplU_tss_for	GTGAATATTTATAGCGCACTCTGAATC	−60 to −34
rplU_RT_rev	CTTCACCGTTTGCGATCATCAGC	114 to 136
T7_rplUb	AGTAATACGACTCACTATAGGGTGAATATTTATAGCG	−60 to −45
rplM_tss_for	CCCCACGTTACAAGAAAG	−156 to −139
rplM_RT_rev	GACGAGCCAGTTCAGTAGC	85 to 103
T7_rplMb	AGTAATACGACTCACTATAGGGACCCCACGTTAC	−157 to −146
rpmBP2_tss_for	CCTTTGAGAATCTCGGGTTTGGC	−139 to −117
rpmB_RT_rev	GCGTTTAGTCGCGTTCAGTGC	61 to 81
T7_rpmBP2b	AGTAATACGACTCACTATAGGGCCTTTGAGAATCTCG	−139 to −125

^aThe positions of primers are indicated relative to the AUG start codon, where A is +1; all forward primers correspond to, and reverse primers are complementary to, the indicated regions.

^bFor this primer, the T7 promoter is underlined, and positions are indicated for the regions relating to the operon.

The amount of each transcript (synthesized from rplM, rplU, and rpmB promoters) was determined using the external standard curve for quantification with the second derivative maximum method (27). The RNA concentrations were calculated in transcript copies per microgram of total RNA. It should be mentioned that for quantification of the rplU′-′lacZ transcript, RT-qPCRs were done with primers rplU_tss_for (Table 2) and DS-lac (see above), and the corresponding standard curve was obtained using T7-rplU (Table 2) and DS-lac.

Cell growth and β-galactosidase assay.

Cell cultures were grown in LB medium at 37°C, harvested in exponential phase (OD₆₀₀, ∼0.4 to 0.5), and used for preparing clarified cell lysates as described previously (26). Protein concentration in a fraction of soluble proteins was determined by the Bradford assay (Bio-Rad). The specific β-galactosidase activities were measured as previously described (26) and expressed in nanomoles of ONPG (o-nitrophenyl-β-d-galactopyranoside) hydrolyzed per minute per milligram of total soluble cell proteins.

Western blot analysis.

Total soluble proteins from strains bearing the rplM′-′lacZ chromosomal reporter and either pACYC184 or its derivative, pL13, were prepared as described for the β-galactosidase assay and analyzed in a 12% Laemmli gel (5 μg/lane). Separated proteins were transferred onto a nitrocellulose membrane (Bio-Rad) and successively revealed with polyclonal goat antibodies against L13 and S9 (a generous gift from P. Sergiev, Lomonosov Moscow State University) and then with secondary donkey anti-goat IgG conjugated with horseradish peroxidase (HRP; Santa Cruz Biotechnology). Visualization was accomplished with the Immun-Star HRP chemiluminescence reagent (Bio-Rad) and a Bio-Rad VersaDoc MP4000 image station.

Bioinformatics tools.

Nucleotide sequences corresponding to the regulatory regions of the bacterial r-protein operons rplU-rpmA, rplM-rpsI, and rpmB-rpmG were obtained from the NCBI Gene database. Multiple promoter regions of the operons from several gammaproteobacterial families were aligned by using WebLogo (28). A list of species for alignment included representatives of Enterobacteriaceae (E. coli, Yersinia pestis, Salmonella enterica, Morganella morganii, Erwinia amylovora, Edwardsiella ictaluri, Hafnia alvei), Pasteurellaceae (Haemophilus influenzae, Mannheimia haemolytica, Pasteurella multocida), Vibrionaceae (Vibrio cholerae, Vibrio fischeri, Vibrio parahaemolyticus), Shewanellaceae (Shewanella oneidensis), and Aeromonadaceae (Aeromonas hydrophila).

RESULTS AND DISCUSSION

Localization of the rplM-rpsI, rplU-rpmA, and rpmB-rpmG promoters and their regulation by alarmone ppGpp and transcription factor DksA under amino acid starvation.

Ribosome biogenesis requires the coordinated synthesis of all ribosomal components in stoichiometric amounts (16S, 23S, and 5S rRNAs and more than 50 r-proteins) and is tightly controlled. Transcription of rRNA operons proceeds at a high level during growth on rich media, but it dramatically decreases upon starvation, the phenomenon known as a ppGpp/DksA-dependent negative stringent response (4). Recent works have shown that promoters of many r-protein operons are also subject to negative stringent control mediated by ppGpp and its cofactor DksA, in parallel with the rRNA operon promoters (5, 23, 24). This indicates that the stringent response plays an important role in regulating the synthesis of all ribosomal components (both rRNAs and r-proteins) to provide rapid reallocation of scarce resources from costly ribosome biogenesis to processes necessary for stress resistance and amino acid synthesis. It seems important to know whether stringent control works for all r-protein operons. Here, we studied three previously unexplored operons, rplM-rpsI, rplU-rpmA, and rpmB-rpmG. Recent localization of the transcription start sites (TSSs) for all E. coli transcription units by RNA-seq (21; E. Hajnsdorf, personal communication) allowed us to outline the promoter regions for these operons and to study transcriptional control.

As expected for promoters that drive expression of ribosomal components, the rplU, rplM, and rpmB P2 promoters bear features typical of the σ⁷⁰-dependent promoters (29); in particular, the divergence from the consensus −10 element is marginal (Fig. 1A). The promoter patterns of the three operons are well conserved in gammaproteobacteria (Fig. 1B). Importantly, all of the de novo localized promoters possess a GC-rich discriminator region downstream from the −10 element, suggesting that transcription might be stringently regulated by ppGpp and DksA under conditions of nutrient limitation (4, 5, 23, 24). To test this possibility, we evaluated the changes in the amounts of the transcripts in total RNA isolated just before and 15 min after induction of serine starvation by SHX treatment of wild-type cells, as well as of relA spoT (ppGpp⁰) and dksA mutants (see Materials and Methods). For this purpose, RT-qPCR with an external RNA standard was used (24, 27). In each case, the RNA standard was identical to the part of the individual transcript analyzed, and all the reactions with the individual transcripts and the corresponding RNA standards were done in parallel with the same primers. We took into account that unlike rplU-rpmA and rplM-rpsI, which are governed by single promoters, the rpmB-rpmG operon has two promoters, namely, rpmB P1 at the end of the preceding radC (yicR) gene (30) and rpmB P2 in the intergenic radC-rpmB region (21). In this case, we evaluated transcription from both promoters by choosing the RNA standard that corresponds to the beginning of the transcript from the intergenic promoter P2 (Table 2).

FIG 1.

Promoters of the rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons. (A) E. coli sequences upstream of the transcription start sites (TSSs) reported by Thomason et al. (21) with classic promoter elements (boxed) and their match with consensus hexamers (in parentheses) indicated. (B) Conservation of promoter patterns in gammaproteobacteria according to WebLogo analysis (26). (C) Transcription of the rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons is subject to ppGpp/DksA-dependent negative stringent control under amino acid starvation. Quantification of transcripts (molecules per microgram of total RNA) from the rplU, rplM, and rpmB promoters in wild-type (wt), ppGpp⁰, and dksA strains before and 15 min after SHX addition by RT-qPCR with external RNA standards using LightCycler software. The average results of two independent RT-qPCR experiments (each probe in triplicate) and standard deviations are shown. (D) Effect of serine starvation on transcription of the rplU′-′lacZ reporter expressed under wild-type rplU promoter (Pwt) or its variant with a C(−5)G point mutation in the discriminator region.

The results (Fig. 1C) indicate that transcription of all three operons is subject to a ppGpp/DksA-mediated negative stringent response. Indeed, a significant decrease in transcript abundance was observed for 15 min after induction of serine starvation in wild-type cultures, ranging from a 3.3-fold downregulation in the case of the rplM promoter to a 5.6-fold decline for rplU (Fig. 1C). In the relA spoT strain (ppGpp⁰), expression of all three operons was higher, downregulation upon induction of serine starvation disappeared (relax response), and moreover, a further increase in transcript amounts after SHX treatment took place in the cases of the rplM and rpmB promoters. An analogous situation was recently described for rplY transcription (24). Since significant changes in profiles of expression under serine starvation have been reported not only for wild-type cells but also for the relA mutant (31), the increase in transcript amounts under starvation, observed for some r-protein operons in the ppGpp⁰ mutant, may be related to as-yet-undefined factors. One of the possible factors may be an elevated stability of the transcripts, given that expression of the rne gene that encodes RNase E, the major endoribonuclease involved in mRNA degradation, is downregulated under these conditions (31).

Our data also indicate an important role for DksA in negative regulation of the r-protein promoters during the stringent response. Indeed, no decrease in transcript abundance under serine starvation was observed for the rplM and rpmB promoters in the dksA mutant (Fig. 1C). In the case of rplU, some reduction took place but to a much lesser extent: 2-fold in the dksA mutant versus 5.6-fold in wild-type cultures (Fig. 1C). These results most likely reflect divergent DksA dependence of the promoters, as the possibility of a DksA-independent impact of ppGpp on the kinetic features of individual promoters cannot be excluded (32).

Taken together, the data obtained demonstrate that transcription of the rplM-rpsI, rplU-rpmA, and rpmB-rpmG operons is downregulated by ppGpp/DksA during the stringent response, which is undoubtedly important for coordinated synthesis of ribosomal components.

Stringent regulation of the rplU-rpmA operon requires specific promoter features.

To show that the changes in transcript levels during a stringent response can be attributed to the regulation of r-protein promoter activities and not to the posttranscriptional events like transcript decay, we changed the discriminator region in the rplU promoter within the rplU′-′lacZ fusion. Earlier, Haugen et al. (33) proposed that rRNA promoters as well as other stringently regulated promoters had evolved to make weak contacts between the discriminator region and σ region1.2, resulting in short-lived open complexes that are susceptible to ppGpp/DksA regulation. In particular, when the base two positions downstream from the −10 element was a C, complexes with RNA polymerase (RNAP) were much shorter lived than complexes with the same promoters bearing G in this position. It was demonstrated that the rrnB P1 with a C-to-G substitution was not regulated during a stringent response (33). We introduced the C(−5)G mutation in the rplU promoter within the rplU′-′lacZ reporter initially on plasmid pEL21 (Table 1) and then transferred this mutation onto the chromosome by homologous recombination. The resulting strain, LABPrplUmut::lacZ, was used for total RNA isolation from nontreated and SHX-treated cells. We then compared the rplU′-′lacZ transcript abundances in the two RNA preparations. The results indicate that the rplU promoter with a C(−5)-to-G substitution is no longer susceptible to a stringent control. Moreover, at an elevated concentration of ppGpp caused by starvation, the transcript level from the mutated promoter even increased, while the promoter activity under normal conditions was lower than that with the wild-type promoter (Fig. 1D). The following explanation may be proposed. The C(−5)G substitution strengthens the RNAP-promoter complex due to the optimal contact with σ1.2. This may slow down promoter clearance, leading to less-efficient transcription. The elevated ppGpp concentration upon starvation reduced the longevity of the open complex, thus facilitating promoter clearance. These experiments demonstrate that a decrease in the rplU transcript level upon starvation is due mainly to promoter features (suboptimal interactions of σ1.2 with the discriminator). We suppose that the same is true for the rplM-rpsI and rpmB-rpmG operons, whose promoters also have a C in the −5 position (Fig. 1A). At the same time, we cannot completely rule out possible effects of the mRNA leader features on the transcript level during the stringent response.

Autogenous regulation of rplM-rpsI at the translation level and the lack of autogenous control of the rplU-rpmA and rpmB-rpmG operons.

Genes encoding dozens of bacterial r-proteins form operons (21 in E. coli) which may include only one (e.g., rpsT, rpsO, rplY, or rpmE), two (e.g., rplM-rpsI or rplU-rpmA), or many genes (e.g., spc or S10 operons encode 11 r-proteins). It has been suggested that r-protein operons of E. coli are commonly regulated at the translation level by the mechanism of autogenous repression (1–3, 34, 35). Such a feedback control is based on the capability of some r-proteins to act not only as structural ribosomal components but also as highly specific repressors of their own mRNAs when they are produced in excess over the rRNA available for de novo ribosome assembly. In E. coli, the repressor r-protein may inhibit translation either by hampering ribosome binding to mRNA or by entrapment of the ribosome in the nonproductive complex (3, 34). A link connecting r-protein synthesis and rRNA level has been reported for many E. coli r-protein operons but not for all, so that about half of the operons are still waiting for their control mechanisms to be unraveled (35). Although recent studies have extended a list of autogenously regulated E. coli operons by adding rplY (24) and rpsF-priB-rpsR-rplL (36–38), several operons still represent an open field for investigations.

We examined the potentiality of rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons for autogenous control in vivo. The experimental approach used here is based on chromosomally integrated fusions of the operon regulatory regions with the lacZ reporter (see Materials and Methods). This approach has already demonstrated its effectiveness in studies of regulatory loops of rpsA (39), rpsB-tsf (22), rpsO (40), and rplY (24). Most frequently, the target for autogenous repression is located in the structured 5′ region of the operon mRNA, although there are two exceptions, i.e., when the repressor protein binds within the intercistronic region (S7 in regulation of the str operon) or even when it binds at the beginning of the third cistron in the case of S8-mediated regulation of the spc operon (reviewed in references 3, 34, and 35). We first generated chromosomal lacZ fusions (rplU′-′lacZ, rplM′-′lacZ, and rpmB′-′lacZ) in which the lacZ expression was governed by the operon promoter and translation initiation region (TIR) of the first gene (Fig. 2A). The β-galactosidase synthesis was measured in the presence of the empty vector pACYC184 (pCtr) or its derivatives bearing active rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons. We also used plasmids expressing r-protein S1 and Hfq because both of the proteins possess high sequence-nonspecific RNA-binding activities and thus may serve as specificity controls (41).

FIG 2.

Analysis of the feedback regulation of the rplU-rpmA, rplM-rpsI, and rpmB-rpmG operons. (A) Structures of the operons. P, promoter; t, terminator. Arrows designate the primers used for cloning the whole operons in pACYC184 or the promoters and translation initiation regions in pEMBLΔ46 (see Materials and Methods). Numbers beside arrows indicating primers represent the 5′ positions relative to the initiation codon of the first operon gene; numbers below operons correspond to the length (in base pairs) of the operon parts. (B) Effects of expression of operon products in trans on activities of the rplU′-′lacZ, rplU-rpmA′-′lacZ, rplM′-′lacZ, and rpmB(P2)′-′lacZ reporters in the β-galactosidase assay. Average results for at least four independent assays and standard deviations are shown. An empty vector, pACYC184 (pCtr), and its derivatives, pS1 and pHfq, were used as specificity controls.

The results of the β-galactosidase assays showed that autogenous repression that involved the 5′ untranslated region (UTR) and TIR of the first operon gene took place only for the rplM′-′lacZ reporter, whereas neither rplU′-′lacZ nor rpmB′-′lacZ reacted to the overexpression of their operon products (Fig. 2B). However, we did not completely exclude the possibility that a mechanism similar to the S8-mediated repression of the spc mRNA may operate in other r-protein operons. To test such a possibility, we constructed a rplU-rpmA′-′lacZ reporter and integrated it into the chromosome. The β-galactosidase activity in the cell was measured in the presence of either the control vector or its derivative, pL21/L27. We observed only a small decrease, about 23%, in the β-galactosidase level (Fig. 2B, rplU-rpmA′-′lacZ construct). In our opinion, this marginal difference cannot be designated feedback regulation based on the accepted definition. As for rpmB-rpmG, it was reported earlier that overexpression of L28 from a plasmid had no impact on the level of L33 in a cell and vice versa (20). Together with our observations, these data support our proposal that rpmB-rpmG is not feedback regulated.

The rplM-rpsI operon is regulated by L13 at the level of translation.

Thus, only one of the three studied operons, rplM-rpsI, appeared to be autogenously regulated. This raises the question of which one of the encoded r-proteins, L13 or S9, is a repressor of rplM′-′lacZ mRNA expression. To resolve this, we deleted the rpsI coding sequence from pL13/S9 to obtain a pACYC184 derivative, pL13, expressing only the rplM gene. The β-galactosidase assay revealed downregulation of the rplM′-′lacZ reporter in the presence of pL13 (Fig. 3A), indicating that L13 acted as an autogenous repressor. This effect is specific only for the rplM-lacZ fusion, as neither rplY-lacZ nor rpsO-lacZ reporters changed their activities in the presence of L13 in trans (Fig. 3A). In E. coli, most r-proteins acting as autogenous regulators block translation initiation, except for L4, which affects both transcription and translation of the S10 operon (3). Transcription regulation was found to result from the L4-stimulated premature termination within the S10 mRNA leader (3). To test such a possibility for the L13-mediated autogenous control, we determined whether the level of the rplM′-′lacZ transcript could change in the presence of pL13 in comparison with the control vector. No alteration in the transcript amount has been revealed by RT-PCR analysis (Fig. 3B), allowing us to conclude that L13-mediated autogenous regulation occurs at the translation level.

FIG 3.

The rplM-rpsI operon is regulated by r-protein L13 at the translation level. (A) Specific downregulation of the rplM′-′lacZ reporter by L13 in trans. An empty vector, pACYC184 (pCtr), and its derivatives, pL13, pL25, and pS15, were used to transform strains carrying rplM′-′lacZ, rplY′-′lacZ, and rpsO′-′lacZ fusions. The average results of three independent β-galactosidase assays and standard deviations are shown. (B) Overexpression of L13 does not affect the rplM′-′lacZ transcript level. Results of RT-PCR analysis are shown. RT-PCR products obtained with primers rplM_tss_for and DSlac were separated in a 2% agarose gel. Lane M, 100-bp DNA ladder (Fermentas). (C) Toxic effect of pL13 on E. coli growth rate. Growth curves in the presence of pACYC184 and pL13 are shown. (D) Effects of L13 in trans on cellular amounts of S9 as revealed by Western blotting. Two independent samples of total proteins isolated from pL13-transformed cells were tested.

It should be mentioned that pL13 slowed cell growth, revealing a toxic effect of excessive L13 amounts on cell metabolism (Fig. 3C). This is most likely because L13 in excess downregulates the expression of chromosomally encoded S9, which is very important for ribosomal activity (see the introduction). Western blot analysis argued in favor of this assumption, showing that in the presence of pL13 the level of S9 was reduced by about 1.5-fold (Fig. 3D). Thus, L13 represents a novel bifunctional r-protein involved not only in ribosome assembly but also in regulation of expression of its own mRNA

Concluding remarks and perspectives.

In this work, we studied for the first time transcriptional and translational regulation of the rplM-rpsI (encoding L13 and S9), rpmB-rpmG (L28 and L33), and rplU-rpmA (L21 and L27) operons. Transcription of these operons was found to be negatively regulated by ppGpp/DksA in response to amino acid starvation. Translation regulation has been examined in vivo with the use of the operon-specific chromosomal lacZ fusions under normal versus augmented synthesis of the operon products. Only the rplM-lacZ reporter appeared to be feedback regulated, whereas rplU-lacZ and rpmB-lacZ did not exhibit any autogenous control. The data obtained suggest that not all E. coli r-protein operons are autogenously regulated at the level of translation and highlight an important role of transcription regulation in a cascade of mechanisms controlling balanced and coordinated synthesis of ribosomal components.

Studies of the regulation of the rplM-rpsI translation revealed a novel r-protein regulator. We found that L13, a primary protein in 50S ribosomal subunit assembly, served as an autogenous repressor of the operon mRNA. Our preliminary data have shown that a 157-nt-long 5′ UTR of the rplM-rpsI mRNA folds in a developed secondary structure and comprises a number of highly conserved sequence/structure features, including an unusual Shine-Dalgarno sequence (GGU), thus encouraging further studies of the regulatory mechanism.