1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Escherichia coli endonuclease RNase E plays a key role in rRNA maturation and mRNA decay. In particular, it controls the decay of its own mRNA by cleaving it within the 5′-untranslated region (UTR), thereby autoregulating its synthesis. Here, we report that, when the synthesis of an RNase E substrate is artificially induced to high levels in vivo, both the rne mRNA concentration and RNase E synthesis increase abruptly and then decrease to a steady-state level that remains higher than in the absence of induction. Using rne–lacZ fusions that retain or lack the rne 5′UTR, we show that these variations reflect a transient mRNA stabilization mediated by the rne 5′UTR. Finally, by putting RNase E synthesis under the control of an IPTG-controlled promoter, we show that a similar, rne 5′UTR-mediated mRNA stabilization can result from a shortage of RNase E. We conclude that the burst in substrate synthesis has titrated RNase E, stabilizing the rne mRNA by protecting its 5′UTR. However, this stabilization is self-correcting, because it allows the RNase E pool to expand until its mRNA is destabilized again. Thus, autoregulation allows RNase E to adjust its synthesis to that of its substrates, a behaviour that may be common among autoregulated proteins. Incidentally, this adjustment cannot occur when translation is blocked, and we argue that the global mRNA stabilization observed under these conditions originates in part from this defect.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In Escherichia coli, proteins that play a general role in gene expression, e.g. in the synthesis or translation of mRNAs or in the folding, maturation or degradation of either RNAs or proteins, often exploit their activity for autoregulating their own expression. For instance, protein CRP, which activates many promoters by binding upstream of them, can also bind downstream of its own promoter, repressing its activity (Hanamura and Aiba, 1991). Likewise, many proteins that bind to specific sites on rRNA or tRNA, such as ribosomal proteins (r-proteins) and certain aminoacyl tRNA synthetases, can also bind to similar sites on their own mRNAs, inhibiting their translation (Nomura et al., 1980; Draper, 1987; Springer, 1996). Similarly, the heat shock proteins specifically downregulate the activity of σ32, a factor required for the transcription of their genes; in particular, the DnaK–DnaJ–GrpE chaperone team, which normally associates with denatured proteins, can also associate with σ32, inactivating this factor and perhaps promoting its degradation (Gross, 1996; Arsène et al., 2000). Finally, RNase III and RNase E, two endonucleases involved in rRNA maturation and mRNA decay, also repress their own synthesis by initiating the decay of their mRNAs (Bardwell et al., 1989; Jain and Belasco, 1995).

What is the benefit for the bacterial cell of these autoregulation loops? It is generally believed that they act as sensors of the metabolic demand for the corresponding proteins. Thus, the overproduction of rRNA or tRNAThr results in an increased synthesis of r-proteins or threonyl-tRNA synthetase respectively. Presumably, these proteins are titrated by the burst of their substrates so that the autoregulation loop is derepressed and their synthesis increases (Yamagishi and Nomura, 1988; Comer et al., 1996). Similarly, the misfolded or denatured polypeptides that accumulate during heat shock are thought to titrate the DnaK–DnaJ–GrpE team, thereby releasing free σ32 and contributing to the induction of heat shock protein synthesis (Craig and Gross, 1991). Following workers in the heat shock field, we use the term ‘homeostasy’ to designate this adjustment between the synthesis of a protein and that of its substrates.

RNase E is a key E. coli endonuclease that controls the production of p5S and 16.3S rRNAs, the immediate precursors of 5S and 16S rRNAs, as well as the decay of many or most mRNAs in the cell (Ghora and Apirion, 1978; Ono and Kuwano, 1979; Cohen and McDowall, 1997; Coburn and Mackie, 1999; Li et al., 1999). In particular, it initiates the decay of its own mRNA by cleaving it in the 5′ untranslated region (UTR), thereby autoregulating its own synthesis (Mudd and Higgins, 1993; Jain and Belasco, 1995; Diwa et al., 2000). Recently, we have observed that situations that result in a translational block also cause an apparent inhibition of RNase E; presumably, this inhibition contributes to the well-known stabilization of bulk mRNA observed under these conditions (Lopez et al., 1998). As an explanation, we noted that the pool of free RNase E in growing cells may be limited because of autoregulation and that, once translation is blocked, it cannot expand further. Under these conditions, the synthesis of rRNA is boosted, and this newly synthesized rRNA is unstable, presumably because it cannot assemble with r-proteins (Shen and Bremer, 1977). We therefore proposed that RNase E is then permanently titrated by excess substrates (Lopez et al., 1998). This model yields the simple prediction that, by inducing the synthesis of a highly expressed RNase E substrate in normally growing cells, it should be possible to titrate RNase E independently of any translation block. However, titration should be transient because the pool of RNase E is free to expand in response to the stabilization of its cognate mRNA. Accordingly, autoregulation would allow homeostasy of RNase E expression as for heat shock or ribosomal proteins. It is this critical point of the model that is tested here.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

A truncated rrnB transcript lacking most of the 16S and 23S sequence is a substrate for RNase E

Our first aim was to design a system that allows massive and inducible synthesis of a well-defined RNase E substrate in the cell. Ideally, this substrate should be devoid of biological activity in order to avoid perturbing growth; in particular, it should remain untranslated. On this basis, we chose plasmid pNO2681 (Gourse et al., 1985) as a source of substrate RNA. This pBR322 derivative carries a defective version of the rrnB operon lacking most of the 16S and 23S rRNA genes, but retaining the 5′ and 3′ ends of the operon, including the two RNase E cleavage sites flanking the 5S rRNA and the RNase E site located upstream of the 16S rRNA (Ghora and Apirion, 1978; Li et al., 1999; Fig. 1A and B). Its putative 1.8 kb transcript carries no known ribosome binding site, nor is it functional in ribosome biosynthesis (Gourse et al., 1985). In pNO2681, the defective rrnB operon lacks its own promoter but is fused downstream of the strong PL promoter from bacteriophage lambda. Whereas at 30°C, this promoter is tightly repressed in cells synthesizing the thermolabile cI857 lambda repressor, it can be switched on by shifting the cultures to 42°C, which inactivates the repressor (Gourse et al., 1985). Since the effect of blocking translation upon RNase E activity (Lopez et al., 1998) and upon rRNA synthesis and stability (Shen and Bremer, 1977), as well as most information on E. coli physiology (Bremer and Dennis, 1996), have been studied in the E. coli B (or B/r) background, the E. coli B strain BL21(DE3) (Studier et al., 1990) was chosen as a host for this study. In the derivatives used, the genuine lacZ gene has been inactivated in order to allow the use of various rne–lacZ fusions for testing RNase E activity (see below).


Figure 1. A. Outline of an experimental system designed for the inducible expression of a RNase E substrate in E. coli. Plasmid pNO2681 (Gourse et al., 1985) carries a truncated version of the rrnB operon (grey box) under the control of the PL promoter (thick arrow), inserted between the HindIII (H) and BamHI (B) sites of pBR322. At 30°C (but not at 42°C), the PL promoter is repressed (curved line) by the thermolabile cI857 repressor encoded by the plasmid pcI857 (circle).

B. Detail of the truncated rrnB operon. The corresponding transcript, which extends down to the natural rrnB terminators (rrnT), differs from the genuine rrnB transcript by a large internal deletion (symbolized by dotted lines), which removes most of the 16S and 23S sequences (open boxes). Note that, because of the deletion, the duplex regions that are cleaved by RNase III during rRNA maturation, (cleavage sites noted ‘III’) are no longer present so that the transcript is presumably not a substrate for RNase III. The RNase E sites flanking the 5S rRNA and the RNase E site located upstream of the 16S rRNA are arrowed (‘E’). Starred bars noted ‘5S’ and ‘9S’ show the positions of the oligonucleotides used in Figs 2 and 7 to probe 5S rRNA or its precursors, respectively.

Download figure to PowerPoint

Plasmids pNO2681 or pBR322 were introduced into cells carrying pcI857, a pBR322-compatible plasmid encoding the repressor (Remaut et al., 1983; Fig. 1A). The doubly transformed cells were grown at 30°C and then shifted to 42°C to induce the PL promoter. Cells carrying either pNO2681 or pBR322 grew at the same rate both before and after the temperature shift, indicating that this induction is not toxic. Total RNA was extracted from the doubly transformed cells before or after the shift and analysed on Northern blots, using oligonucleotide probes complementary to the 5S rRNA or the immediately upstream flanking sequence (‘5S’ and ‘9S’ probes respectively; Fig. 1B). With pNO2681-carrying cells growing at 30°C, a very faint signal corresponding to the 1.8 kb species was detected with both probes (cf. Fig. 2A; only the 9S probing is shown). As expected, this species was absent from pBR322-carrying cells. After the 42°C shift, the 1.8 kb species from pNO2681-carrying cells became much more prominent, reflecting the activation of the PL promoter (Fig. 2A). To assess whether this species is a substrate for RNase E, the same experiment was repeated with cells carrying the rne-1 mutation. The corresponding RNase E polypeptide is functional at 30°C but is rapidly inactivated at 42°C (Ono and Kuwano, 1979; Mudd et al., 1990). With these cells as with rne+ cells, a very faint, pNO2681-specific signal corresponding to the 1.8 kb species was detected at 30°C (Fig. 2A). After the 42°C shift, this species accumulated to much higher levels than in rne+ cells, suggesting that it is a substrate for RNase E. The unprocessed precursors of 5S rRNA originating from the genuine rrn operons (‘9S’ rRNA; see below) also accumulated at 42°C as a result of RNase E inactivation. As expected, this accumulation was observed in both pBR322- and pNO2681-carrying cells (Fig. 2A).


Figure 2. A. Induction of transcription from pNO2681 (cf. Fig. 1A). Cells harbouring plasmids pcI857 and either pNO2681 (pNO lanes) or pBR322 (pBR lanes) have been harvested either before or at various times after a shift from 30°C to 42°C, and total RNA has been probed on Northern blots with the 9S probe (cf. Fig. 1B). The cells used encode an RNase E activity that is either thermostable (wt) or thermolabile (rne-1) at 42°C. The position of the 1.8 kb transcript from the truncated rrnB operon is arrowed. The bracket marked ‘9S’ pinpoints the polydisperse precursors of the 5S rRNA that are synthesized from the seven chromosomal rrn operons and accumulate when RNase E is inactive (rne-1 strain, 42°C). Note that, beyond the 9S and 1.8 kb species, another species, presumably a degradation product from the 1.8 kb transcript, accumulates at 42°C in rne-1 cells harbouring plasmid pNO2681 (starred).

B. Effect of induction upon the accumulation of 5S rRNA and rne mRNA. Cells were grown and total RNA was extracted as in (A), except that only wild-type (rne+) cells were used. Northern blots were probed with probes specific for the 5S and 23S rRNA and for the rne mRNA (see Experimental procedures).

Download figure to PowerPoint

The 5S probing showed that, in rne+ cells carrying pNO2681, the abundance of 5S rRNA over other rRNA species increased progressively after shifting to 42°C (compare the 5S and 23S probings in Fig. 2B). No comparable increase was observed in pBR322-carrying cells. This result shows that the 1.8 kb transcript can finally be processed into 5S rRNA, i.e. that the RNase E cleavage sites flanking the 5S rRNA sequence (Fig. 1B) are functional. After 1 h (0.75 doubling times), the 5S rRNA expression has increased by 70%. This figure is close to the theoretical value that can be calculated using previous estimates for the rate of transcription from plasmids similar to pNO2681 (1.6- to 1.7-fold the rate of transcription from all seven chomosomal rrn operons; Yamagishi and Nomura, 1988), indicating that the processing of the 1.8 kb transcript into 5S rRNA is nearly quantitative. Whereas from the above figure the rate of synthesis for the 1.8 kb species may seem very high indeed, this rate is presumably actually less than that of mRNA molecules, many of which are also RNase E targets (see Discussion).

Altogether, these results indicate that, like the genuine 9S rRNA, the 1.8 kb transcript is a substrate for RNase E. When induced, it presumably accounts for a significant but not dominant fraction of total RNase E targets in the cell (see Discussion). In the following, it is referred to as ‘the substrate’.

The rne mRNA is stabilized transiently after induction of substrate synthesis

We next recorded the effect of substrate induction upon the steady-state concentration of the rne mRNA. To this end, Northern blots prepared with total RNA from the same cultures as above (rne+ cells) were hybridized with an rne-specific probe (Fig. 2B, rne probing). The major species detected with this probe is 3.6 kb in length, consistent with the known length of the rne transcript (Casarégola et al., 1992; Jain and Belasco, 1995). With pNO2681-carrying cells, the concentration of this species increased abruptly (by ≈ sixfold) after 10 min at 42°C, and decreased progressively thereafter; however, even after 60 min at 42°C, it remained ≈ twofold higher than before the temperature shift. In contrast, with cells carrying the control plasmid pBR322, the concentration of the rne mRNA changed more smoothly and over a more limited range (Fig. 2B).

The transient accumulation of the rne mRNA after substrate induction might reflect either an increased promoter activity or a stabilization of the rne mRNA. In the latter event, it would presumably reflect a transient shortage in RNase E activity, because the stability of the rne mRNA is controlled by the rate of cleavage of its 5′UTR, which itself responds to RNase E activity (Jain and Belasco, 1995). To test this issue, we used two in-phase rnelacZ fusions designed by Jain and Belasco (1995). These fusions consist of the rne promoter followed by the same rnelacZ hybrid coding sequences, but they differ in the presence or absence of most of the 361-nt-long rne 5′UTR in between. Only with the former fusion was mRNA stability and protein yield affected by RNase E activity as for the rne gene (Jain and Belasco, 1995). These fusions were then introduced onto the chromosome of BL21(DE3). For simplicity, the corresponding derivatives are qualified here as ‘+5′UTR’ and ‘−5′UTR’ respectively (Fig. 3A); by extension, the same qualifiers are used to designate the rne–lacZ fusions themselves or the mRNAs originating from them. The +5′UTR and −5′UTR cells carrying plasmids pcI857 and either pNO2681 or pBR322 were then grown as before. Total RNA was extracted before or after the 42°C shift and analysed on Northern blots using a lacZ probe. Whether pNO2681 or pBR322 was present, most of the signal detected before the shift consisted of an abundant smear arising from incomplete molecules: the prevalence of these species presumably reflects the fact that the time required for synthesizing the long rne–lacZ mRNA exceeds its lifespan (Yarchuk et al., 1992; Jain and Belasco, 1995). Only with the −5′UTR cells was a weak signal corresponding in size (3.7 kb) to the complete rnelacZ fusion mRNA observed, suggesting that, at 30°C, this mRNA is slightly more stable than its +5′UTR counterpart. After the shift, the fusion mRNA from +5′UTR cells (4.0 kb in this case) became apparent transiently in cells carrying pNO2681, whereas it remained almost undetectable in the same cells carrying pBR322. In contrast, with −5′UTR cells, no difference in mRNA pattern was detected whether pNO2681 or pBR322 was present (Fig. 3B). Thus, even though the +5′UTR and −5′UTR mRNAs are synthesized from the same rne promoter, only the former behaves like the genuine rne mRNA in accumulating after substrate induction. We conclude that this accumulation reflects mRNA stabilization, probably caused by a shortage in RNase E activity (see below for further results on this point).


Figure 3. A. The BL21(DE3) derivatives used here, ENS401 (‘+5′UTR’) and ENS406 (‘−5′UTR’), harbour constructs consisting of the rne promoter followed by an in-phase fusion between the rne (dark grey box) and lacZ (open box) coding sequences. In between, the 5′UTR of the rne gene (light grey box) is either present (+5′UTR cells) or absent (−5′UTR cells). These fusions are inserted in the malT–malP intergenic region of the chromosome. The lacZ gene is followed by the beginning of the lacY gene (lacY′) and by two tandemly arranged terminators (Ter). Also represented is the rne gene under the control of its genuine promoter (Prne) and 5′UTR.

B. +5′UTR and −5′UTR cells harbouring plasmids pcI857 and either pNO2681 or pBR322 (cf. Fig. 1A) were grown and used for RNA extraction as in Fig. 2B. The transcript from the rnelacZ fusion was probed on Northern blots. Arrows indicate the position of the rnelacZ mRNA species from +5′UTR and −5′UTR cells (4.0 kb and 3.7 kb respectively); given their size, these species presumably result from a processing of the initial transcript downstream of the lacZ gene (Murakawa et al., 1991). The position of the 23S and 16S rRNA is indicated by ticks. To check for uniform RNA loading, the 23S rRNA has been visualized directly on the same membrane (bottom).

Download figure to PowerPoint

The accumulation of the rne or rne–lacZ mRNA after substrate induction is paralleled by an increased synthesis of the corresponding proteins

Next, we tested whether the stabilization of the rne or +5′UTR mRNAs that follows substrate induction results in an increased synthesis of the corresponding proteins (Fig. 4). To this end, we used the same cultures as for Fig. 3B (+5′UTR cells). At timed intervals after the temperature shift, aliquots of the cultures were pulse labelled with a [35S]-methionine–cysteine mixture. Extracts from the labelled cells were then incubated with anti-RNase E and anti-β-galactosidase antibodies, and immune complexes were analysed by SDS–PAGE (see Experimental procedures). Although several labelled proteins were apparent on the gels, presumably reflecting the limited specificity of the antibodies used, two polypeptides corresponding in size to RNase E (apparent molecular weight ≈ 180 kDa; see Casarégola et al., 1992) and RNase E–β-galactosidase hybrid protein (139 kDa) were detected. Their identity was ascertained by running in parallel extracts of cells that either overexpress or lack these proteins (see Experimental procedures for details). Before the shift, the rate of RNase E synthesis was essentially the same whether cells carried pNO2681 or pBR322 (see Fig. 4A and B for quantification). In contrast, after the shift, this synthesis was distinctly higher in pNO2681- than in pBR322-carrying cells; the ratio between the two went through a maximum at 20 min after induction (fourfold) but, even after 60 min, it still amounted to about twofold. The behaviour of the fusion protein was very similar (see the legend to Fig. 4 for comment on the 60 min point). Thus, the rate of synthesis of the two proteins closely followed the changes in the accumulation of the cognate mRNAs (compare Fig. 4B with Figs 2B and 3B).


Figure 4. A. Aliquots of the same cultures used in Fig. 3 (+5′UTR cells) were pulse labelled with [35S]-methionine and cysteine, and the rates of synthesis of labelled RNase E (E) and hybrid RNase E–β-galactosidase protein (E-Z) were detected by immunoprecipitation, using a mixture of antibodies directed against the two proteins. Lanes marked ‘E overexp.’, ‘E-Z overexp.’ and ‘E truncation’ correspond to extracts in which each of the RNase E or hybrid polypeptides is overexpressed or missing altogether. The star pinpoints a couple of polypeptides (presumably the β and β′ subunits of E. coli RNA polymerase) that cross-hybridize with these antibodies and are used as internal standards.

B. Histograms showing the relative rate of synthesis of RNase E (closed bars) or hybrid RNase E–β-galactosidase (hatched bars) in cells carrying either pNO2681 or pBR322. This figure is a quantification of the experiment shown in (A). In this particular experiment, the recovery of the E-Z protein in the (pBR, 60′) sample was low; hence, the high value in the corresponding histogram (dotted box). This low recovery was not seen in another experiment and is presumably artifactual.

Download figure to PowerPoint

The burst in the synthesis of RNase E–β-galactosidase hybrid protein after substrate induction in +5′UTR cells could also be confirmed directly by β-galactosidase activity measurements. Whereas before the temperature shift, this activity was similar in pNO2681- or pBR322-carrying cells, after 90 min at 42°C, it was 40% higher in pNO2681-carrying cells. It is noteworthy that β-galactosidase activity reflects the average rate of synthesis of the hybrid protein over the whole duration of the culture, hence the modesty of the difference observed.

The effect of varying RNase E expression on the stability of the +5′UTR and −5′UTR mRNAs

Jain and Belasco (1995) used the rne-1 mutation to show that the +5′UTR and −5′UTR mRNAs are stabilized differently by lowering the RNase E activity in the cell. In this section, we show that the same result can be achieved by decreasing the expression of the wild-type enzyme. This result supports the view that the differential stabilization of these mRNAs after substrate induction is caused by RNase E titration.

To gain experimental control over RNase E expression in +5′UTR or −5′UTR cells, we replaced the promoter and 5′UTR of the chromosomal rne gene with the promoter and 5′UTR from the lac operon (Fig. 5A). This change allows modulation of RNase E expression by the lac operon inducer IPTG. The downstream lac–rne junction is so tailored that the lacZ initiation codon is followed immediately by the second codon of the rne gene, i.e. the RNase E sequence is not modified (Fig. 5B). In the following, this engineered rne gene is called ‘Plac–rne’; cells harbouring it are referred to as ‘Plac’ cells. As expected since RNase E is an essential enzyme, Plac cells did not form colonies on LB plates lacking IPTG. Likewise, growth in liquid medium was IPTG dependent: at high IPTG concentrations (≥ 160 µM), it was equivalent to that of cells carrying the wild-type rne gene but, for lower IPTG concentrations, the growth rate declined progressively (at 40 µM IPTG, it has dropped to one-half its natural value). These variations correlated with the expression of RNase E in the cell. The latter was quantified on Western blots using an anti-RNase E antibody. Extracts of cells carrying the wild-type rne gene, as well as samples of pure RNase E, were run in parallel as controls (Fig. 5C). For high IPTG concentrations (≥ 320 µM), the expression of RNase E in Plac cells reached a plateau, presumably reflecting full induction of the Plac promoter; this plateau corresponds to ≈ 250% of the RNase E expression from cells carrying the wild-type rne gene (hereafter, the latter expression is referred to as ‘normal’). Below 320 µM IPTG, RNase E expression decreased progressively: for 160 µM IPTG, it matches closely the normal expression, whereas for 40 µM IPTG, it has dropped to only 10% of this expression (Fig. 5C). In summary, the Plac–rne construct allows manipulation of RNase E expression over a broad range, say from 1/10th to 2.5-fold the normal expression: the lower limit is a practical one, being set by the slow growth of the cultures.


Figure 5. A. Control of RNase E expression in-wild type (rne+) and Plac cells. In wild-type (WT) cells, RNase E (scissors) autoregulates its own expression by cleaving the 5′UTR of its mRNA, whereas in Plac cells, its expression is downregulated by the lac repressor (the lacI product). The lac and rne sequences are indicated by empty and grey boxes respectively. Negative regulation loops are indicated by curved lines. Note that, in Plac and wt cells, these loops operate at the transcriptional and post-transcriptional levels respectively.

B. Sequence encompassing the downstream lac–rne junction on the chromosome of Plac cells. ‘+1’ designates the lac transcription start point. The major lac operator site (nt 1–22) is underlined; the lacZ Shine–Dalgarno and initiation codon are in upper case, and the rne sequence is italicized.

C. Total proteins from Plac cells grown at different IPTG concentrations were analysed by Western blot using an antibody against RNase E. To check for uniform protein loading, the blot was also treated with an antibody against an unrelated protein, threonyl-tRNA synthetase (ThrS). Extracts of WT cells as well as various amounts of pure RNase E were also loaded as controls. Bound immunoglobulins were revealed with [125I]-protein A and autoradiography.

Download figure to PowerPoint

The +5′UTR and −5′UTR cells carrying the Plac–rne construct (Fig. 6A) were then grown in the presence of variable IPTG levels; total RNA was prepared from these cultures and probed on Northern blots with a lacZ internal probe. With +5′UTR cells, the full-length species predominated when RNase E expression was below normal, whereas for higher expression, only incomplete species were detected, indicating mRNA destabilization (Fig. 6B). In contrast, with −5′UTR cells, the mRNA pattern remained similar for all IPTG concentrations. These changes were paralleled by variations in β-galactosidase expression: with +5′UTR cells, increasing IPTG concentration from 40 to 320 µM caused this expression to decrease more than 10-fold, whereas in −5′UTR cells, no such effect was seen (variations, if any, were in the reverse direction; Fig. 6B). Altogether, these results show that the +5′UTR mRNA but not the −5′UTR mRNA is stabilized when RNase E expression falls below normal, as with the rne-1 mutation.


Figure 6. Effect of RNase E concentration upon the stability of the +5′UTR and −5′-UTR mRNAs.

A. The strains used here are identical to the +5′UTR and −5′UTR strains (Fig. 3), except that the genuine rne gene is replaced by the Placrne fusion (Fig. 5A).

B. The RNA from the above cells growing with different concentrations of IPTG (as indicated below each lane) has been probed on Northern blots with a lacZ internal probe. The histograms below show the β-galactosidase activities in the same cultures used for mRNA analysis. All symbols are as in Fig. 3A and B.

Download figure to PowerPoint

It is noteworthy that, under conditions of low RNase E expression, the β-galactosidase level is higher in +5′UTR than in −5′UTR cells. This difference, which has also been observed in the presence of the rne-1 mutation (Jain and Belasco, 1995), indicates that the +5′UTR mRNA is intrinsically more translatable and/or less susceptible to RNase E-independent decay than the −5′UTR mRNA.

Other effects of varying RNase E expression

We also assayed the effect of varying RNase E expression upon the maturation or decay of other RNase E substrates. A classical substrate is the so-called ‘9S’ rRNA. This RNase III-generated fragment encompasses the 3′ end of the rRNA precursor; it is further processed by RNase E into p5S, the immediate precursor of the 5S rRNA (Ghora and Apirion, 1978). 9S rRNA is in fact polydisperse, owing to heterogeneities at the 3′ end of the seven rRNA operons (Condon et al., 1995; Fig. 7A). To test the effect of RNase E expression upon 9S rRNA processing, total RNA from Plac cells growing with various concentrations of IPTG was probed on Northern blots with the 5S and 9S probes. To achieve a better resolution, acrylamide–urea electrophoresis was used here instead of agarose electrophoresis (Fig. 7B). For RNase E expression equal to or larger than normal, the signals originating from 9S rRNA were quite faint compared with that originating from the 5S rRNA, consistent with a rapid processing of the 9S rRNA. Quantitatively, raising RNase E expression from 100% to 250% of normal caused an ≈ twofold decrease in the 9S compared with the 5S rRNA signals. Conversely, for lower RNase E expressions, the 9S rRNA signals rose progressively until, together, they eventually became nearly as intense as the 5S rRNA signal (Fig. 7B). Thus, lowering the RNase E concentration below normal slows down 9S rRNA processing markedly.


Figure 7. The rate of 9S rRNA processing varies with RNase E expression.

A. Schematic representation of the most promoter-distal RNase III fragment from the E. coli rrn transcript (‘9S rRNA’; cf. Fig. 1B). The sequence heterogeneity among the seven rrn operons downstream of the 5S sequence is symbolized by the hatched box. All other symbols are as in Fig. 1B.

B. Total RNA from Plac cells growing with different IPTG concentrations (as indicated below each lane) was separated by urea–PAGE, blotted and probed with the 5S (left) or 9S (right) probes. The arrow points to 5S rRNA. Stars or ticks on the left indicate species that are detected with both probes. The starred species match in size the classical ‘9S’ RNA from the rrnB operon (243 nt; see Cormack and Mackie, 1992).

Download figure to PowerPoint

Finally, we assessed the effect of varying RNase E concentration upon the functional stability of bulk mRNA. Plac cells growing exponentially in the presence of various concentrations of IPTG were treated with the transcription inhibitor rifampicin, and the residual protein-synthesizing capacity of the cultures was subsequently recorded as a function of time by following the ability of aliquots to incorporate [35S]-methionine in proteins (Fig. 8A). This capacity decayed nearly exponentially with time, allowing the functional half-life of bulk mRNA to be determined (Fig. 8B). For RNase E expressions equal to or higher than normal (IPTG ≥ 160 µM), this half-life was nearly constant (1.7–1.9 min) and similar to that observed with cells carrying the wild-type rne gene. However, for lower RNase E expressions, the half-life increased significantly: when RNase E expression dropped to 1/10th of the normal value, the increase was twofold (Fig. 8A and B).


Figure 8. The functional stability of bulk mRNA varies with RNase E concentration.

A. Plac cells growing with 40 (top) or 160 (bottom) µM IPTG were treated with rifampicin (zero time) and, at the indicated times, the residual protein-synthesizing capacity of the culture was measured by incubating an aliquot with [35S]-methionine until incorporation had ceased. Labelled proteins were separated by SDS–PAGE and visualized by autoradiography.

B. Total radioactivity incorporated into proteins is plotted versus time, assuming exponential decay kinetics; bulk mRNA half-lives are estimated from the slopes. Open circles, 40 µM IPTG (10% normal RNase E expression; cf. Fig. 5C). Closed diamonds, 80 µM IPTG (50% normal). Open triangles, 320 µM IPTG (250% normal).

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Here, we have examined the effect of overexpressing an RNase E substrate upon the decay of mRNAs that carry the rne 5′UTR, the target for RNase E autoregulation. To facilitate the interpretation of results, we have also studied the effect of varying RNase E expression upon the stability of the same mRNAs and, more generally, of other RNase E substrates. For clarity, the latter results are discussed first.

Experimental control of RNase E expression

To gain experimental control over RNase E expression, we have replaced the promoter and 5′UTR from the rne gene with the promoter and 5′UTR from the lac operon. As the lac fragment used in this replacement encompasses not only the main operator site (cf. Fig. 5B), but also a minor one (OIII) located upstream of the promoter, the transcription of the engineered rne gene is expected to be very tightly repressed by the lac repressor in the absence of IPTG (by ≈ 440-fold; Oehler et al., 1990). Under conditions of maximal IPTG induction, RNase E expression was 2.5-fold the normal level. Interestingly, this modest overexpression has no effect on growth, whereas overexpression of RNase E from high-copy-number plasmids bearing the rne gene has been found to be toxic (Claverie-Martin et al., 1991). In contrast, growth was impaired when the RNase E expression fell below normal and was completely abolished in the absence of IPTG, confirming the essential nature of RNase E. However, surprisingly, as little as 1/10th of the normal expression is enough to sustain growth, albeit at a twofold reduced rate. The same observation has been made by Belasco and colleagues (cited in Jiang et al., 2000).

Besides these effects on growth, variations in RNase E expression also affected the stability of RNase E substrates. In particular, the functional decay of bulk mRNA, the processing of 9S rRNA and the cleavage of the rne 5′UTR (as judged by the compared stabilities and β-galactosidase yields of the ‘+5′UTR’ and ‘−5′UTR’ mRNAs; see Fig. 6) were all slowed when RNase E expression dropped below normal. These effects mimic those observed with the rne-1 mutation (Mudd et al., 1990; Jain and Belasco, 1995), confirming that the phenotype associated with this mutation results from a shortage of RNase E activity and not from a property of the mutated polypeptide per se. Conversely, increasing RNase E expression from 100% to 250% of normal appears to affect individual substrates differently. Thus, bulk mRNA is not destabilized further, and the processing of 9S rRNA is only modestly accelerated, as judged by the magnitude of the 9S signal on Northern blots (Fig. 7B). In contrast, the rne 5′UTR is destabilized more significantly, as seen by the fact that the β-galactosidase yield from the +5′UTR mRNA decreases fourfold in this range of RNase E expression (Fig. 6B). Whereas the molecular basis for this differential behaviour is unknown, the particular sensitivity of the rne 5′UTR to changes in RNase E expression seems reasonable given its role as a ‘sensor’ of RNase E demand (see below).

Homeostasy of RNase E expression in wild-type cells

Turning back to cells carrying the wild-type rne gene, we have observed that, when the synthesis of an artificial RNase E substrate is induced to relatively high levels (see below), the concentration of the rne mRNA and the synthesis of the RNase E polypeptide first increase rapidly and then decrease to a new steady-state level, which nevertheless remains higher than in the absence of induction. These changes are not the result of changes in the activity of the rne promoter, but rather of variations in the rate of cleavage of the rne 5′UTR. Thus, whereas the +5′UTR fusion mRNA accumulates after substrate induction much as the rne mRNA itself, no accumulation was seen with the −5′UTR mRNA (Fig. 3B). Given the inverse correlation between the stability of the +5′UTR mRNA and the availability of RNase E (Fig. 6B), these observations suggest that RNase E is transiently titrated after substrate induction. A straightforward interpretation is that, because of autoregulation, the pool of free RNase E in the cell is limited: a burst in substrate synthesis causes its titration, stabilizing RNase E substrates including the rne mRNA. However, this stabilization is transient, because it allows the RNase E pool to expand and correct the titration. Thereby, autoregulation allows RNase E to adjust its expression to that of its substrates.

As noted in the Introduction, this behaviour is not unique to RNase E, but is also observed with several other autoregulated proteins. Particularly significant in this respect is the case of heat shock proteins. Upon heat shock, their synthesis increases and then decreases to a new steady state, which remains higher than before the heat shock. These variations parallel the activity of the sigma factor σ32. Although the mechanisms modulating σ32 activity are multiple, the sequestration of σ32 by the chaperone team DnaK–DnaJ–GrpE clearly plays an important role. According to current models, the misfolded or denatured polypeptides that accumulate during heat shock titrate the chaperones, releasing free σ32 and causing heat shock protein synthesis to increase. However, this increase is only transient because, once the pools of DnaK, DnaJ and GprE have expanded enough, σ32 will become sequestrated again (Craig and Gross, 1991). The parallel between this ‘homeostatic’ model of heat shock response and our present interpretation is striking. Accordingly, we regard the response of RNase E expression to changes in the concentration of its substrates as another example of homeostasy.

Quantitative considerations

The notion that RNase E can be titrated by excess substrate leads to speculative but intriguing questions about the turnover of this enzyme. We rely here on the physiological data compiled by Bremer and Dennis (1996).

At 37°C and in the medium used here (0.75 doublings h−1), the rate of stable RNA synthesis is about 5.6 × 105 nt min−1 cell−1 (extrapolated for the values listed in the above reference for 0.6 and 1 doublings h−1). About 85% of this synthesis corresponds to the 5.5 kb rRNAs precursors; these precursors are therefore made at a rate of (5.6 × 105 × 0.85)/5500, or ≈ 90 copies min−1 cell−1. After RNase III cleavage, each precursor will yield two RNase E substrates (cf. Fig. 1B). Therefore, rRNA synthesis will contribute a total of 180 (90 × 2) new RNase E substrates min−1. As for mRNA synthesis, it corresponds to ≈ 6.1 × 105 nt min−1 cell−1 (Bremer and Dennis, 1996). Assuming, for the sake of this discussion, that the average size of an mRNA is 2000 nt, then about 300 (6.1 × 105/2000) new mRNA molecules are synthesized per minute and per cell. Altogether, the total number of RNase E substrates synthesized per minute will range from 180 to 480, depending on how many mRNA molecules are RNase E substrates. When induced, the artificial substrate used here is synthesized at a rate 1.7-fold higher than the rRNA precursors, or ≈ 150 (90 × 1.7) molecules per min. According to the above estimates, its synthesis represents an increase of 24% [150/(150 + 480)] to 45% [150/(150 + 180)] of the total RNase E substrates in the cell. That this relatively modest increase is enough to titrate the enzyme suggests that RNase E is already close to substrate saturation under normal growth conditions.

RNase E is a relatively abundant protein, being present at several hundred copies per cell (Kido et al., 1996). That this rather large pool is nearly saturated by a flow of only 175–475 new substrate molecules per min suggests that the turnover of this enzyme is rather low. Conceivably, the cleavages themselves are slow or the enzyme remains durably associated with substrate before or after cleavage; alternatively, the degradation of most mRNA molecules involves multiple cleavages, delaying enzyme release. According to another interpretation, RNase E may be compartmentalized in vivo so that only a fraction of its pool is actually available for RNA decay/processing. Interestingly, in this respect, whereas mRNA decay is usually regarded as mostly cytoplasmic, most of the RNase E is known to be localized in the vicinity of the inner membrane (Liou et al., 2001).

The stabilization of mRNA after a translation block may reflect RNase E titration

In E. coli B cells growing under the conditions used here (minimal or amino acid-supplemented glycerol medium), the total rate of RNA synthesis increases ≈ twofold after a block in translation. In particular, for rRNA, the increase is two- to 2.5-fold. Moreover, the newly synthesized rRNA is then unstable (Shen and Bremer, 1977), and RNase E presumably participates in its decay, as suggested by the presence of rRNA fragments within degradosome preparations (Bessarab et al., 1998). We recently proposed that this burst of substrate synthesis under conditions in which the RNase E pool cannot expand causes permanent RNase E titration, accounting for the stabilization of mRNAs observed when translation is blocked. The artificial substrate used here retains the three well-characterized RNase E sites of the rRNA precursor, and its rate of synthesis is similar to the extra rate of rRNA synthesis observed after a translation block. Since RNase E can be titrated by inducing the synthesis of this substrate, it is reasonable to assume that it can also be titrated after a translational block. Thus, RNase E titration is probably one of the causes for the bulk mRNA stabilization under these conditions; whether it is the only one remains to be seen.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plasmids and strains

Plasmids pEZ201 and pEZ206, which carry in phase rnelacZ fusions retaining or lacking most of the rne 5′UTR, respectively, were kindly donated by Drs Jain and Belasco, as was plasmid pRNE101, a pACYC177 derivative carrying the rne gene and flanking sequences (Jain and Belasco, 1995). Plasmid pNO2681 was a gift from Dr R. Gourse (Gourse et al., 1985). Plasmid pcI857 (Remaut et al., 1983) was obtained from Dr M. Springer.

Construction of strains carrying the +5′UTR and −5′UTR fusions

We started from plasmid pTlacZ-Arg5 (Lopez et al., 1994). Within its polylinker, this pUC18 derivative carries sequentially: (i) a truncated lac operon encompassing the lacZ gene and the beginning of lacY; and (ii) two tandemly arranged transcriptional terminators. The KpnI–DraIII fragment extending from the polylinker sequence down to within the lacZ coding sequence was then replaced by the KpnI–DraIII fragments from pEZ201 and pEZ206, which extend from upstream of the rne promoter down to the same DraIII site within lacZ. The whole KpnI–XbaI inserts extending from upstream of the rne promoter to downstream of the terminators were then subcloned from the resulting plasmids into the shuttle plasmid pOM43 (Chevrier-Miller et al., 1990), yielding plasmids pSS201 and pSS206 respectively. The inserts were then transferred into the malT–malP intergenic region of the chromosome of MO20 [a Lac derivative of BL21(DE3)] as described by Lopez et al. (1994), yielding strains ENS401 and ENS406.

Construction of the Plac strains

Using the fusion polymerase chain reaction (PCR) technique (Ho et al., 1989), we constructed a 1.8 kb chimeric BamHI–SalI fragment carrying sequentially: (i) the region extending from nt −853 to −39 with respect to the rne transcription start; (ii) the region extending from nt −126 to +41 with respect to the lac transcription start; and (iii) the region extending from nt +365 to +1194 with respect to the rne transcription start. The resulting fragment carries the lac promoter and 5′UTR, flanked by rne homology regions. The rne and lac fragments were amplified from plasmids pRNE101 and pUC19 respectively. The chimeric fragment was then cloned between the BamHI and SalI sites of the shuttle plasmid pKO3 (Link et al., 1997). After sequencing, the fragment was transferred into the rne region of the E. coli chromosome using homologous recombination between the plasmid- and chromosome-borne rne sequences (Link et al., 1997), yielding the Placrne construct (Fig. 5A and B). As a recipient strain, we used a derivative of MG1655 carrying a Tn10 transposon near the rne gene. The latter was transferred from CH1828 [ams (ts), zce-726::Tn10;Mudd et al., 1990] by transduction, selecting for TetR and ability to grow at 42°C. The Placrne construct was then transduced from the MG1655 background into strains ENS401 and ENS406, selecting for TetR and IPTG-dependent growth. The presence of the Placrne construct in the final strains was checked by PCR.

Growth of cells

Cells were grown in MOPS medium (Neidhardt et al., 1974) containing glycerol (0.2% w/v) as the carbon source. For substrate induction studies, the above medium was supplemented with kanamycin (50 µg ml−1) and ampicillin (100 µg ml−1). Cells were grown to an OD600 of 0.4 at 30°C, then shifted to 42°C for the time indicated. For experiments with Plac cells, this medium was supplemented with all amino acids, nucleic acid bases and vitamins, together with variable concentrations of IPTG, as indicated. Growth temperature was 37°C in this case. With Plac cells, the use of starter cultures yielded inconsistent results: we suspect that the expression of RNase E differs widely in exponential and saturated cultures, generating ‘memory’ effects when the latter are diluted. Therefore, cells were routinely reisolated on M63B1 plates (Miller, 1972) containing 0.5% casamino acids and a low concentration (32 µM) of IPTG. After 20 h at 37°C, the very small colonies that appeared were used directly to inoculate liquid cultures (one colony per 5 ml). Cells were harvested at an OD600 of 0.3–0.4.

RNA analysis

RNA was extracted as described by Yarchuk et al. (1992). Conditions for agarose or acrylamide–urea electrophoresis, RNA blotting and membrane hybridization with randomly 32P-labelled DNA fragments or 5′32P-labelled oligonucleotides have been described before, as have the 1.8 kb lac fragment used to probe the rne–lacZ mRNA and the ‘5S’ and ‘9S’ oligonucleotides used to probe the 5S and 9S rRNA (Yarchuk et al., 1992; Lopez et al., 1994; 1999). The oligonucleotide probe used to detect 23S rRNA was 5′-AAGGTTAAGCCTCACGGTTC-3′; the probe used to detect the rne mRNA was an 816 bp BamHI fragment from plasmid pGM102 (Cormack et al., 1993), which encompasses the 3′ region of the rne gene. Radioactive signals were quantified with a BAS1000 imager (Fuji).

Protein analysis

Purification of RNase E

RNase E was overexpressed from BL21(DE3) cells harbouring plasmid pGM102 (Cormack et al., 1993). The overexpressed protein was eluted from a preparative 7.5% SDS–polyacrylamide gel. Its concentration was estimated by running aliquots on an analytical gel in parallel with known amounts of bovine serum albumin (BSA), staining with Coomassie blue and comparing the intensities of the stained bands by densitometry.

Quantification of RNase E in Plac cells

Pellets of exponentially growing cells (equivalent to 0.1 OD600) were resuspended directly in Laemmli sample buffer and separated by SDS–PAGE (7.5% acrylamide). The proteins were transferred to a nitrocellulose membrane (Amersham), which was incubated for 1 h at 37°C in buffer A (20 mM Tris-HCl, pH 7.5, 0.9% NaCl, 0.1% Triton X-100) containing 5% BSA. The membrane was incubated overnight with anti-RNase E and anti-ThrS antibodies [gifts from Dr A. J. Carpousis (Toulouse) and Drs H. Putzer and C. Condon (Paris); antibodies were used at 1:5000 in buffer A with 0.5% (BSA)], then washed (3 × 15 min) with buffer A and, finally, incubated for 2 h with buffer A containing 0.1 µCi ml−1[125I]-protein A (Amersham). After washing again, the membrane was exposed to a Fuji imager screen.

Functional half-life of bulk mRNA (Plac cells)

Rifampicin (500 µg ml−1 final) was added to exponentially growing cultures (OD600≈ 0.4) of Plac cells (or control wild-type cells). Aliquots (500 µl) were subsequently collected at timed intervals and incubated with 10 μCi of [35S]-methionine–[35S]-cysteine mix (Amersham) until 30 min after rifampicin addition. As methionine and cysteine are present in excess in the growth medium (see above), only a small fraction of the label is incorporated in these experiments, so that the incorporated radioactivity reflects the protein-synthesizing activity of the culture. The labelled samples were then separated by SDS–PAGE (12% acrylamide); after drying the gels, the total incorporated radioactivity was quantified with the Fuji imager.

Rate of synthesis of RNase E and hybrid RNase E–β-galactosidase polypeptides after substrate induction

Cells growing exponentially at 30°C were shifted to 42°C as described above. At timed intervals after the shift, aliquots of the cultures (0.5 ml) were removed, incubated for 1 min at 42°C with 10 µCi of [35S]-methionine–cysteine mix and then for 5 min with 10 mM cold methionine−0.5 mM cold cysteine (final concentration). Cells were collected by centrifugation, resuspended in SDS–PAGE loading buffer without thiols and heated (100°C for 4 min). Samples were diluted 20-fold in the immunoprecipitation buffer described by Carpousis et al. (1994), except that Triton X-100 was used in place of Genapol and BSA (0.2 mg ml−1) was added. Extracts were incubated with anti-RNase E and anti-β-galactosidase antibodies (2.5 µl each for 0.5 ml of culture) for 1 h at 4°C and treated with 100 µl of a 20% (v/v) suspension of protein A–sepharose beads (Pharmacia). After centrifugation, the immune complexes were washed with the immunoprecipitation buffer (without BSA), analysed by SDS–PAGE (7.5% acrylamide) and, finally, quantified with the Fuji imager. The anti-β-galactosidase antibody was from Rockland. As controls, cultures overexpressing RNase E or RNase E–β-galactosidase fusion polypeptides or lacking them altogether were grown and processed as above. For this purpose, we used BL21(DE3) rne-1 cells that had been incubated at 42°C (at this temperature, the inactive RNase E polypeptide encoded by the rne-1 allele accumulates), BL21(DE3) cells carrying pSS206, a multicopy plasmid encoding the fusion protein (see above) and BL21(DE3)rne131 cells, which synthesize a truncated RNase E (Lopez et al., 1999).


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We are much indebted to Drs C. Jain, J. G. Belasco and R. Gourse for plasmids, and to Dr A. J. Carpousis for antibodies. We thank Dr P. J. Lopez for his participation in early experiments, and Drs I. Iost and M. Springer for fruitful discussions. This work was supported by CNRS, ENS and by grants from ARC (no. 5474) and MENRT (programme ‘Microbiologie’) to M.D. S.S. was supported by the Fundaçao para a Ciência e Tecnologia (fellowship PRAXIS XXI/BM/19113/99), and I.M. by the Fondation pour la Recherche Médicale.


  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • 1
    Arsène, F., Tomoyasu, T., Bukau, B. (2000) The heat shock response of Escherichia coli. Int J Food Microbiol 55: 39DOI: 10.1016/s0168-1605(00)00206-3
  • 2
    Bardwell, J.C., Régnier, P., Chen, S.M., Nakamura, Y., Grunberg-Manago, M., Court, D.L. (1989) Autoregulation of RNase III operon by mRNA processing. EMBO J 8: 34013407
  • 3
    Bessarab, D.A., Kaberdin, V.R., Wei, C.-L., Liou, G.-G., Lin-Chao, S. (1998) RNA component of Escherichia coli degradosome: evidence for rRNA decay. Proc Natl Acad Sci USA 95: 31573161DOI: 10.1073/pnas.95.6.3157
  • 4
    Bremer, H. & Dennis, P.P. (1996) Modulation of chemical composition and other parameters of the cell by growth rate. In Escherichia coli and Salmonella: Cellular and Molecular Biology. Neidhardt, F.C. (ed.). Washington, DC: American Society for Microbiology Press, pp. 15531569
  • 5
    Carpousis, A.J., Van Houwe, G., Ehretsmann, C., Krisch, H.M. (1994) Copurification of E. coli RNase E and PNPase: evidence for a specific association between two enzymes important in RNA processing and degradation. Cell 76: 889900
  • 6
    Casarégola, S., Jacq, A., Laoudj, D., McGurk, G., Margarson, S., Tempête, M. , et al. (1992) Cloning and analysis of the entire Escherichia coli ams gene. ams is identical to hmp1 and encodes a 114 kDa protein that migrates as a 180 kDa protein. J Mol Biol 228: 3040
  • 7
    Chevrier-Miller, M., Jacques, N., Raibaud, O., Dreyfus, M. (1990) Transcription of single-copy hybrid lacZ genes by T7 RNA polymerase in Escherichia coli: mRNA synthesis and degradation can be uncoupled from translation. Nucleic Acids Res 18: 57875792
  • 8
    Claverie-Martin, F., Diaz-Torres, M.R., Yancey, S.D., Kushner, S.R. (1991) Analysis of the altered mRNA stability (ams) gene from Escherichia coli. Nucleotide sequence, transcriptional analysis, and homology of its product to MRP3, a mitochondrial ribosomal protein from Neurospora crassa. J Biol Chem 266: 28432851
  • 9
    Coburn, G.A. & Mackie, G.A. (1999) Degradation of mRNA in Escherichia coli: an old problem with some new twists. Prog Nucleic Acid Res Mol Biol 62: 55108
  • 10
    Cohen, S.N. & McDowall, K.J. (1997) RNase E: still a wonderfully mysterious enzyme. Mol Microbiol 23: 10991106
  • 11
    Comer, M.M., Dondon, J., Graffe, M., Yarchuk, O., Springer, M. (1996) Growth rate-dependent control, feedback regulation and steady-state mRNA levels of the threonyl-tRNA synthetase gene of Escherichia coli. J Mol Biol 261: 108124
  • 12
    Condon, C., Squires, C., Squires, C.L. (1995) Control of rRNA transcription in Escherichia coli. Microbiol Rev 59: 623645
  • 13
    Cormack, R.S. & Mackie, G.A. (1992) Structural requirements for the processing of Escherichia coli 5S ribosomal RNA by RNase E in vitro. J Mol Biol 228: 10781090
  • 14
    Cormack, R.S., Genereaux, J.L., Mackie, G.A. (1993) RNase E activity is conferred by a single polypeptide: overexpression, purification, and properties of the ams/rne/hmp1 gene product. Proc Natl Acad Sci USA 90: 90069010
  • 15
    Craig, E.A. & Gross, C.A. (1991) Is hsp70 the cellular thermometer? Trends Biochem Sci 16: 135140
  • 16
    Diwa, A., Bricker, A.L., Jain, C., Belasco, J.G. (2000) An evolutionary conserved RNA stem–loop functions as a sensor that directs feedback regulation of RNase E gene expression. Genes Dev 14: 12491260
  • 17
    Draper, D.E. (1987) Translational regulation of ribosomal proteins in E. coli: molecular mechanisms. In Translational Regulation of Gene Expression. Ilan, J. (ed.). New York: Plenum Press, pp. 126
  • 18
    Ghora, B.K. & Apirion, D. (1978) Structural analysis and in vitro processing to p5 rRNA of a 9S RNA molecule isolated from an rne mutant of E. coli. Cell 15: 10551066
  • 19
    Gourse, R.L., Takebe, Y., Sharrock, R.A., Nomura, M. (1985) Feedback regulation of rRNA and tRNA synthesis and accumulation of free ribosomes after conditional expression of rRNA genes. Proc Natl Acad Sci USA 82: 10691073
  • 20
    Gross, C. (1996) Function and regulation of the heat-shock proteins. In Escherichia coli and Salmonella: Cellular and Molecular Biology. Neidhardt, F.C. (ed.). Washington, DC: American Society for Microbiology press, pp. 13821399
  • 21
    Hanamura, A. & Aiba, H. (1991) Molecular mechanism of negative autoregulation of Escherichia coli crp gene. Nucleic Acids Res 19: 44134419
  • 22
    Ho, S.N., Hunt, H.D., Horton, R.M., Pullen, J.K., Pease, L.R. (1989) Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77: 5159
  • 23
    Jain, C. & Belasco, J.G. (1995) RNase E autoregulates its synthesis by controlling the degradation rate of its own mRNA in E. coli: unusual sensitivity of the rne transcript to RNase E activity. Genes Dev 9: 8496
  • 24
    Jiang, X., Diwa, A., Belasco, J.G. (2000) Regions of RNase E important for 5′-end dependent RNA cleavage and autoregulated synthesis. J Bacteriol 182: 24682475
  • 25
    Kido, M., Yamanaka, K., Mitani, T., Niki, H., Ogura, T., Hiraga, S. (1996) RNase E polypeptides lacking a carboxyl-terminal half suppress a mukB mutation in Escherichia coli. J Bacteriol 178: 39173925
  • 26
    Li, Z., Pandit, S., Deutscher, M.P. (1999) RNase G (CafA protein) and RNase E are both required for the 5′ maturation of 16S ribosomal RNA. EMBO J 18: 28782885
  • 27
    Link, A.J., Phillips, D., Church, G.M. (1997) Methods for generating precise deletions and insertions in the genome of wild-type Escherichia coli: application to open reading frame characterization. J Bacteriol 179: 62286237
  • 28
    Liou, G.G., Jane, W.N., Cohen, S.N., Lin, N.S., Lin-Chao, S. (2001) RNA degradosomes exist in vivo in Escherichia coli as multicomponent complexes associated with the cytoplasmic membrane via the N-terminal region of ribonuclease E. Proc Natl Acad Sci USA 98: 6368
  • 29
    Lopez, P.J., Iost, I., Dreyfus, M. (1994) The use of a tRNA as a transcriptional reporter: the T7 late promoter is extremely efficient in Escherichia coli but its transcripts are poorly expressed. Nucleic Acids Res 22: 11861193
  • 30
    Lopez, P.J., Marchand, I., Yarchuk, O., Dreyfus, M. (1998) Translation inhibitors stabilize Escherichia coli mRNAs independently of ribosome protection. Proc Natl Acad Sci USA 95: 60676072
  • 31
    Lopez, P.J., Marchand, I., Joyce, S.A., Dreyfus, M. (1999) The C-terminal half of RNase E, which organizes the Escherichia coli degradosome, participates in mRNA degradation but not rRNA processing in vivo. Mol Microbiol 33: 188199
  • 32
    Miller, J.H. (1972) Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, pp. 431435
  • 33
    Mudd, E.A. & Higgins, C.F. (1993) Escherichia coli endoribonuclease RNase E: autoregulation of expression and site-specific cleavage of mRNA. Mol Microbiol 9: 557568
  • 34
    Mudd, E.A., Krisch, H.M., Higgins, C.F. (1990) RNase E, an endoribonuclease, has a general role in the chemical decay of Escherichia coli mRNA: evidence that rne and ams are the same genetic locus. Mol Microbiol 4: 21272135
  • 35
    Murakawa, G.J., Kwan, C., Yamashita, J., Nierlich, D.P. (1991) Transcription and decay of the lac messenger: role of an intergenic terminator. J Bacteriol 173: 2836
  • 36
    Neidhardt, F.C., Bloch, P.L., Smith, D.F. (1974) Culture medium for Enterobacteria. J Bacteriol 119: 736747
  • 37
    Nomura, M., Yates, J.L., Dean, D., Post, L.E. (1980) Feedback regulation of ribosomal protein gene expression in Escherichia coli: structural homology of ribosomal RNA and ribosomal protein mRNA. Proc Natl Acad Sci USA 77: 70847088
  • 38
    Oehler, S., Eismann, E.R., Krämer, H., Müller-Hill, B. (1990) The three operators of the lac operon cooperate in repression. EMBO J 9: 973979
  • 39
    Ono, M. & Kuwano, M. (1979) A conditional lethal mutation in an E. coli strain with a longer chemical lifetime of messenger RNA. J Mol Biol 129: 343357
  • 40
    Remaut, E., Tsao, H., Fiers, W. (1983) Improved plasmid vector with a thermoinducible expression and temperature-regulated runaway replication. Gene 22: 103113
  • 41
    Shen, V. & Bremer, H. (1977) Chloramphenicol-induced changes in the synthesis of ribosomal, transfer, and messenger RNA in Escherichia coli B/r. J Bacteriol 130: 10981108
  • 42
    Springer, M. (1996) Translational control of gene expression in Escherichia coli and bacteriophages. In Regulation of Gene Expression in E. coli. Lin, E.C.C., and Lynch, A.S. (eds). Austin, TX: R.G. Landes Company, pp. 85126
  • 43
    Studier, F.W., Rosenberg, A.H., Dunn, J.J., Dubendorff, J.W. (1990) Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol 185: 6089
  • 44
    Yamagishi, M. & Nomura, M. (1988) Effects of induction of rRNA overproduction on ribosomal protein synthesis and ribosome subunit assembly in Escherichia coli. J Bacteriol 170: 50425050
  • 45
    Yarchuk, O., Jacques, N., Guillerez, J., Dreyfus, M. (1992) Interdependence of translation, transcription and mRNA degradation in the lacZ gene. J Mol Biol 226: 581596