Present address: 71 Sevilla Street, Urb. Vista Alegre, Aguadilla, PR 00603, USA.
Regulation and mode of action of the second small RNA activator of RpoS translation, RprA
Article first published online: 31 OCT 2002
Volume 46, Issue 3, pages 813–826, November 2002
How to Cite
Majdalani, N., Hernandez, D. and Gottesman, S. (2002), Regulation and mode of action of the second small RNA activator of RpoS translation, RprA. Molecular Microbiology, 46: 813–826. doi: 10.1046/j.1365-2958.2002.03203.x
- Issue published online: 31 OCT 2002
- Article first published online: 31 OCT 2002
- Accepted 5 August, 2002.
Translation of the stationary phase sigma factor RpoS is stimulated by at least two small RNAs, DsrA and RprA. DsrA disrupts an inhibitory secondary structure in the rpoS leader mRNA by pairing with the upstream RNA. Mutations in rprA and compensating mutations in the rpoS leader demonstrate that RprA interacts with the same region of the RpoS leader as DsrA. This is the first example of two different small RNAs regulating a common target. Regulation of these RNAs differs. DsrA synthesis is increased at low temperature. We find that RprA synthesis is regulated by the RcsC/RcsB phosphorelay system, previously found to regulate capsule synthesis and promoters of ftsZ and osmC. An rcsB null mutation abolishes the basal level, whereas mutations in rcsC that activate capsule synthesis also activate expression of the rprA promoter. An essential site with similarity to other RcsB-regulated promoters was defined in the rprA promoter. Activation of the RcsC/RcsB system leads to increased RpoS synthesis, in an RprA-dependent fashion. This work suggests a new signal for RpoS translation and extends the global regulation effected by the RcsC/RcsB system to coregulation of RpoS with capsule and FtsZ.
The stationary-phase sigma factor RpoS is a regulator of several genes that are expressed during stationary-phase or during growth under stress conditions in Escherichia coli and related bacteria. RpoS levels change dramatically during these conditions, a result of both changes in the synthesis of the protein, primarily at the level of translation, and changes in its degradation (Hengge-Aronis, 2000). A small RNA, DsrA, is known to regulate RpoS translation (Sledjeski et al., 1996) via a stretch of 17 nucleotides (nt) in the first stem–loop of DsrA; this region is complementary to nucleotides in the regulatory leader region of the RpoS mRNA. This allows DsrA to disrupt this regulatory structure and promote RpoS translation (Majdalani et al., 1998). Although many small RNAs have now been described in E. coli, and many have important regulatory roles, this is one of the rare examples of a small RNA acting to stimulate rather than inhibit translation (Wassarman, 2002).
In a dsrA mutant strain carrying an RpoS–lacZ translational fusion, residual β-galactosidase activity is still detectable (Majdalani et al., 2001). From a screen of a multicopy library for suppressors of the dsrA mutation, we identified a second small RNA that activates RpoS–lacZ translation. This 105-nt regulatory small RNA was called RprA (RpoS regulatory RNA A) (Fig. 1). RprA has only limited homology to DsrA (Majdalani et al., 2001); we demonstrate here that it can, nonetheless, interact with the RpoS messenger RNA to activate it in a manner parallel to that used by DsrA. The identification of two small RNAs that regulate a common target provides a unique opportunity to examine what is essential for small RNA action on this target.
At low temperatures, DsrA synthesis is high and RpoS synthesis is strongly dependent on DsrA (Sledjeski et al., 1996; Repoila and Gottesman, 2001); RprA is not made at elevated levels at low temperatures. Although we were able previously to demonstrate a modest effect of RprA on RpoS translation after osmotic shock (Majdalani et al., 2001), it seemed likely that understanding regulation of RprA would clarify the conditions under which it is important for cell growth. Northern blot analysis suggests that the pattern of DsrA and RprA accumulation is rather different, consistent with different signals regulating their synthesis and stability (Majdalani et al., 2001; Repoila and Gottesman, 2001; Wassarman et al., 2001). This strongly suggests that these RNAs are not redundant but, instead, are made under different conditions and therefore modulate RpoS expression in response to different environmental factors. Here we demonstrate that the regulation of rprA synthesis is quite different from that for DsrA and that RpoS, unexpectedly, is part, albeit indirectly, of the group of genes regulated by the RcsC/RcsB two-component system.
Mode of action of RprA
RpoS translation is negatively regulated by the formation of a hairpin stem–loop structure in the mRNA leader that occludes the Shine–Dalgarno (S–D) site just upstream of the rpoS translation start (Fig. 2A) (Brown and Elliott, 1997). DsrA, the first small RNA found to positively regulate RpoS translation, acts by pairing with the inhibitory stem of the hairpin upstream region (Majdalani et al., 1998) (Fig. 2B). Within a stretch of 23 nt in DsrA, from nt 9–31, 21 are complementary to the RpoS leader (Fig. 2B). We have previously demonstrated a role for the nucleotides shown in bold in Fig. 2B in direct RNA–RNA pairing to allow RpoS translation (Majdalani et al., 1998). In a linear alignment, RprA has no such extended complementarity. The best linear alignment gave an 11 out of 13 nt match between RprA and the RpoS leader (between A33 and U45 of RprA) (Fig. 2C). However, this complementarity is with the same RpoS leader region shown to be critical for DsrA activity. Therefore, we tested the role of this complementarity in RprA's ability to activate RpoS translation. The A37 nucleotide in RprA is predicted to pair with U124 of the RpoS leader. A set of U124 mutations described previously (Majdalani et al., 1998) allowed us to test the importance of the predicted pairing at this site. As a control, we used the A39 nucleotide in RprA, which is not predicted to pair with the RpoS leader mRNA (Fig. 2C). Mutant derivatives of RprA carrying C, G and U at either A37 or A39 were constructed, and cloned into pNM12 to place expression of the RNA under the control of the ParaBAD promoter and to allow comparison with the wild-type pBAD-rprA (pNM100). Mutant and wild-type plasmids, as well as the vector, were tested for their ability to stimulate expression of four different derivatives of an RpoS–lacZ translational fusion. These fusions differed in the nucleotides in the leader at the positions usually occupied by the predicted pair U124:G208 (see Fig. 2).
As shown in Fig. 3A, when the RpoS target RNA carried the wild-type U:G pair, the wild-type residue A at position 37 of RprA had a higher stimulatory effect than A37U or A37C (which should be unable to pair with the U in the target RNA); A37G was reasonably effective as well, consistent with the predicted pairing. However, when the U124 is mutated to create a tighter C124:G208 pairing for the RpoS leader, the wild-type A37 form of RprA was inactive, as were A37U and A37C. In contrast, A37G was now active, as predicted for a pairing between A37G and C124. These results strongly support direct pairing between RprA position 37 and position 124 of the RpoS leader RNA. Results for two other RpoS constructs, G124:C208 and G124:U208, were reasonably consistent with this model although less clear-cut (Fig. 3A). Both show the highest induction with the A37U or A37C mutants of RprA, expected to pair with the G at position 124 of the leader. Why U is more active than C is unclear, as is the reason for the high basal level activity of the G-U pair in the absence of RprA. This difference in behaviour of U:G and G:U was previously noted for DsrA (Majdalani et al., 1998). Presumably the G:U pair loosens the RpoS hairpin; A37U may cause changes in RprA folding that increase stability or activity.
A39 is not predicted to pair with the RpoS leader (Fig. 2C). Consistent with this prediction, changes in the A39 nucleotide gave only minor differences in activity on any of the RpoS leaders, except for an unexpectedly high activity for the RprA A39U mutation in an RpoS G124:U208 strain, possibly suggesting that U at this position also improves RprA activity.
A region of the RpoS leader upstream of the U124 region discussed above was shown to participate in pairing with DsrA in an essential fashion (Majdalani et al., 1998) (Fig. 2B). This region was also important for RprA activity; a strain carrying an RpoS–lacz fusion mutant for this region (RpoS–lacZ–NcoI′) was defective for stimulation by RprA (Fig. 3C, compare values for RprA in left and right set of bars). In RprA, the only region predicted to pair with this part of the RpoS leader is about 18 nt downstream of the A37 patch of complementarity, at nt 55–62 (see Fig. 2C). We made the compensating mutations for the RpoS leader NcoI′ mutants in this region of RprA (RprA*) to determine if this pairing is important for RprA activity (Fig. 3C). In a strain carrying the wild-type RpoS–lacZ fusion, activity of the wild-type RprA (pNM100) was two to threefold higher than that for the RprA* mutant (pNM109). In contrast, in a strain carrying the RpoS–lacZ–NcoI′ mutation, RprA* restored activity to the wild-type level (Fig. 3C). Therefore, both regions of DsrA previously shown to be important for its action are also present and required in RprA, albeit in non-contiguous parts of the RNA.
We had previously mutagenized rprA in the region around U69 to insert an XhoI site in the DNA; this 6 nt change had no effect on RprA activity on RpoS. Insertion of a kanamycin resistance cassette at that site did destroy RprA activity; however, excision of the kanamycin resistance gene, which leaves a 12 nt insertion in the XhoI mutant sequence, restored activity (Majdalani et al., 2001). Therefore, sequences in this region of the RNA are not critical for RpoS stimulation. Sequences near the 5′-end of RprA were also tested by constructing a derivative of the rprA gene encoding an inverted stem–loop 1 (nt 5–30) in the RNA; this mutation did not affect RprA activity (data not shown). Neither of these positions is predicted to interact with the RpoS leader (Fig. 2C), supporting the model further, and ruling out important additional interactions with these regions. We note that the position of the regions in DsrA and RprA that stimulate RpoS translation are quite different; in DsrA the initial stem–loop is critical for activity (Majdalani et al., 1998). Therefore, the position of complementary sequences within small RNAs regulating the same target need not be similar. This data further demonstrates that RprA pairs with the same regions of the RpoS leader to activate translation, presumably by a mechanism similar to the one used by DsrA (Majdalani et al., 1998).
DsrA has been shown to require the RNA binding protein Hfq for its action on two different substrates (Sledjeski et al., 2001). Hfq has also been shown to be necessary for RpoS translation (Brown and Elliott, 1996). Recently, RprA was found to bind to Hfq in immuno-precipitation experiments (Wassarman et al., 2001). Here, we tested the requirement for Hfq in RprA function by measuring expression of an RpoS–lacZ fusion in a ΔdsrA or a ΔdsrA hfq::kan host in the presence of pSAC1, a multicopy plasmid carrying rprA. In the hfq+ host, the RpoS–lacZ fusion strain carrying pSAC1 expressed an average of 6 units of activity. In the hfq host, this was reduced to 0.9 units. The vector gave 1.4 units in the hfq+ host and 0.7 units in the hfq– host; this relatively low level of activity and therefore modest effect of Hfq on activity is presumably because no DsrA is present. Therefore, Hfq is needed for RprA stimulation of RpoS translation, consistent with the observed binding of RprA to Hfq (Wassarman et al., 2001).
Identification of regulators of rprA
One advantage of having two (or more) different small RNAs, each activating RpoS translation by a similar mechanism, would be the possibility of regulating RpoS in response to different signals via differential regulation of synthesis of the RNAs. The dsrA promoter is regulated by low temperature (Repoila and Gottesman, 2001) and does not resemble the rprA upstream region. rprA is located between two predicted open reading frames (ORFs), ydiK and ydkI. There are 187 nucleotides between the predicted translation stop codon of ydiK and the +1 start of the rprA transcript. Both rprA itself and portions of the upstream region are well conserved in E. coli, Salmonella subsp. and Klebsiella pneumoniae (Majdalani et al., 2001). The pattern of upstream conservation suggested that all essential sites for the rprA promoter would lie within the 142 nt that precede the +1 of transcription (Majdalani et al., 2001). Four promoter fragments of varying lengths were cloned into plasmid pRS415 to construct transcriptional fusions with lacZ and crossed into λRS45 (Simons et al., 1987). Single copy lysogens were isolated in DJ480, a Δlac derivative of MG1655. The longest promoter fusion starts at nt −142; the shortest starts at nt −43 (Fig. 4A). All the fusions retain the predicted −35 and −10 boxes and all but the shortest fusion retain a stretch of high conservation upstream of the −35 box.
Strains carrying the fusions were grown in Luria–Bertani (LB) broth and the β-galactosidase activity was determined (Fig. 4A). No differences were detected between fusions carrying the −142, −79, and −55 nt promoters (Fig. 4A, lines 1, 2 and 3). However, the −43 promoter fusion had little or no activity (Fig. 4A, line 4). This suggests that the conserved region between −55 and −43 includes sequences essential for a positive regulator of the rprA promoter.
To isolate potential activators of rprA, a pBR322-based E. coli genomic library (Ulbrandt et al., 1997) was introduced into strain DH300, carrying the 142 nt fusion. As DH300 carrying the control vector is weakly Lac+ on MacConkey Lactose Ampicillin (MacLacAmp) plates, we screened for clones that increased β-galactosidase activity. From 25 000 colonies, 81 candidate clones were isolated. DNA was extracted from the clones that appeared to have the highest activity on plates. The insert ends of these plasmids were sequenced and mapped against the E. coli genome using blast (Altschul et al., 1990) to determine insert size and genes carried. Most plasmids carried DNA fragments that mapped to different and independent regions of the chromosome, suggesting that many DNA fragments, in multicopy, can activate the fusion, and that we had not saturated the search for activating plasmids (Table 1). However, two isolates carried the rcsF gene; multicopy rcsF had been found previously to increase capsule synthesis (Gervais and Drapeau, 1992), and our isolates were mildly mucoid as well. One other isolate carried a truncated portion of rcsC, the sensor kinase of the capsule phosphorelay regulatory system. We have not confirmed expression of the truncated protein. The isolation of these plasmids suggested that rprA might be regulated by the proteins that regulate capsular polysaccharide synthesis.
|Fragment boundaries||Activity in wild typeb||Genes presentc||Description of genesd|
|2315498–2318854||5.9||< rcsC ′ > atoS′||RcsC, membrane protein, sensor kinase for colanic acid synthesis; AtoS, sensor kinase for ato operon|
|4351948–4354200||n/d||< lysU′ < yjdL < ‘cadA||LysU, t-RNA synthetase; YjdL, probable protein/oligopeptide symporter|
|3596465–3598252||n/d||< livJ ′ < ‘rpoH||LivJ, periplasmic binding protein, amino acid transporter|
|4044948–4047602||2.6||> ‘polA > spf||Spot 42, small RNA|
|2266375–2270968||n/d||> ‘yeiR > yeiU > spr > rtn > yejA′||YeiU, putative permease; Spr, putative lipoprotein suppressor of prc (periplasmic protease) mutation; Rtn, membrane protein conferring resistance to lambda and N4 phages|
|583999–588373||n/d||< ompT ′ < envY < ybcH < ‘nfrA||OmpT, outer membrane protease VIII; EnvY, regulator of envelope protein genes YbcH, unknown function|
|218585–220840||15.3||< proS′ < yaeB < rcsF < ‘yaeC||YaeB, unknown function; RcsF, capsule regulator|
|219459–221276||15.3||< yaeB ′ < rcsF < yaeC < ‘yaeE||RcsF, capsule regulator; YaeC, putative lipoprotein|
|11667–14313||ND||< yaaI ′ > dnaK > dnaJ ′||DnaK, chaperone; DnaJ, co-chaperone|
|3529293–3533985||2.2||< yhgE′ > pckA < envZ < ‘ompR||YhgE, putative transporter; PckA, PEP carboxykinase; EnvZ, IM osmosensor kinase|
|879521–881675||11.1||< yliJ′ > dacC < ‘deoR||YliJ, putative glutathione transferase; DacC, penicillin binding protein 6|
|2117666–2119621||3.0||< wzxC ′ < wcaJ||WzxC, probable capsule transport; WcaJ, UDP-glucose lipid carrier transferase|
|775655–779098||5.8||>tolA′ > tolB > pal > ybgF ′||TolB, periplasmic protein/tolerance to colicins; Pal, essential lipoprotein for bacterial envelope integrity|
|46513–49510||2.2||> ‘yaaU > yabF > kefC ′||YabF, putative oxidoreductase; KefC, Na+/K+ antiporter|
Capsular polysaccharide synthesis genes (cps) in E. coli are regulated at the level of transcription by the phosphorelay system, RcsB/YojN/RcsC (Takeda et al., 2001). In this system, RcsC is the transmembrane sensor kinase/phosphatase and RcsB is the response regulator. In addition, regulation of the cps genes requires RcsA, an unstable positive regulatory protein (Stout and Gottesman, 1991). To determine the role of these capsule regulators on rprA expression, various rcs mutations were introduced by P1 transduction into strain DH300. The strains were assayed for expression of the rprA–lacZ fusion at 37°C and levels of RprA RNA, with an RprA-specific biotinylated probe (Fig. 5).
As shown in Fig. 5A, rprA promoter activity, which is low in a wild-type strain, is further diminished in an rcsB null strain; the basal level of RprA RNA is also abolished (Fig. 5B, compare lanes 1 and 3). The activity of the rprA promoter was restored (and increased beyond the basal level) by complementing the null mutation with a multicopy plasmid carrying rcsB (data not shown). In contrast, an rcsA mutant had no effect on the expression of the fusion or on the level of RprA RNA (Fig. 5). RcsA is present in very small amounts in wild-type strains due to its degradation by the Lon protease. To confirm the lack of dependence of RprA synthesis on RcsA, we assayed rprA–lacZ activity in a lon strain or in DH300 carrying a multicopy rcsA plasmid (pATC400). In a lon strain, rprA–lacZ levels increase modestly, to three units. In the presence of multicopy rcsA, rprA–lacZ levels rose by three to fourfold, to nine units. The effect of a similar increase in RcsA, either by use of a lon– mutant or by introduction of a multicopy rcsA+ plasmid, on expression of the cps system, is to raise expression 50 to 100-fold (Stout and Gottesman, 1991). Therefore, RcsA is not needed and has only a slight stimulatory effect on rprA expression.
RcsB is regulated via RcsC-dependent phosphorylation and dephosphorylation. In an rcsC null mutant strain, there is about a 10-fold increase in activity of the rprA–lacZ fusion and in RprA RNA levels above wild-type levels (Fig. 5), although cps gene expression is not increased in this mutant (Parker et al., 1992). Another rcsC allele, rcsC137, is a mutation that was originally selected to significantly increase cps expression (Brill et al., 1988). rcsC137 is presumed to have lost its dephosphorylation activity. The mutation causes an induction of rprA–lacZ expression levels of more than 50-fold; RprA RNA is also increased, at least 50-fold (Fig. 5). An rcsC137 rcsA double mutant shows the same level of activity as an rcsC317 strain (55 units), further confirming that the activation of rprA does not require RcsA. Capsule synthesis is down to 5–10% of the rcsC137 levels in an rcsC137 rcsA double mutant (Gottesman and Stout, 1991). These results suggest that RcsC both positively and negatively regulates the activity of RcsB, presumably via phosphorylation. This data further supports the conclusion that RcsB but not RcsA is required for rprA expression.
All of the plasmids isolated in the original screening failed to stimulate the rprA–lacZ fusion in an rcsB mutant host (data not shown), suggesting that all regulation we can detect is dependent upon RcsB. The level of expression for cells carrying most of the plasmids was modest (see Table 1), and was not sufficient to make cells mucoid. For the rcsF plasmids, which do make cells mucoid, the increased level of rprA–lacZ expression was dependent on functional RcsC; in an rcsC mutant, the expression was reduced to that seen for the rcsC mutant alone (data not shown), suggesting that the rcsF plasmid acts upstream of RcsC. This result is somewhat different from a previous report of RcsF action (Gervais and Drapeau, 1992) and is discussed below.
RprA levels increase with increasing cell density and are most abundant in stationary phase (data not shown) (Argaman et al., 2001; Wassarman et al., 2001). We estimate an increase of at least eightfold between mid-log and late stationary phase cells in the level of RprA. We examined the expression of the rprA–lacZ fusion for its behaviour on entry into stationary phase, and found, at the most, a 2.5-fold increase in promoter activity at similar points in the growth curve. Therefore, it seems likely that a substantial part of the increase in RprA accumulation in stationary phase is due to stabilization of RprA rather than increased synthesis.
Analysis of the rprA promoter
We had shown above that a site for positive activation of the rprA promoter lies between −55 and −43, a conserved part of the rprA promoter. Consistent with this site playing a critical role in regulation, expression of all but the shortest fusion is stimulated by the rcsF plasmid, and this stimulation is abolished in an rcsB mutant (data not shown). Thus the site between −43 and −55 behaves genetically like the site of RcsB action.
This sequence was compared with those previously defined as sites for RcsB action. In the case of cps synthesis, regulation requires an RcsA/RcsB heterodimer to bind at a site located far upstream of the −35 box (between −150 and −700) (Kelm et al., 1997; Wehland et al., 1999; Wehland and Bernhard, 2000). Recently, promoters where RcsB acts without RcsA have been described; in these cases the RcsB binding site has been found just upstream of the −35 region of the promoter. The first such promoter described was the ftsA1p promoter of ftsZ (Carballes et al., 1999). Mutagenesis studies of the RcsB binding site in the ftsZ promoter identified an imperfect inverted repeat sequence that is presumed to be the recognition sequence for RcsB (Carballes et al., 1999); this sequence bears some similarity to the RcsA/RcsB site identified in the cps promoters. Recently, one of two promoters for osmC was also found to be dependent on RcsB but not on RcsA (Davalos-Garcia et al., 2001); again, the site necessary for regulation is just upstream of the −35 region.
The rprA promoter is most similar in pattern of regulation and location of the critical sequences for regulation to the ftsA1p and osmC1p promoters (Fig. 4B). We noted a partial inverted repeat between −47 and −38, centred around an AGA - - - TCT motif (dotted underline, Fig. 4B line 1) found to be most critical in the ftsA1p promoter (Carballes et al., 1999). Furthermore, the G −46 A mutation described previously that curtails promoter activity destroys the AGA motif (Majdalani et al., 2001) (Fig. 4B, line 2). To further confirm the importance of these nucleotides in the regulation of rprA expression, we mutagenized this region as shown in Fig. 4B. All mutations that changed the AGA - - - TCT motif led to a complete loss of activity from the rprA promoter. Compared with wild type (Fig. 4B, line 1), duplication of the TCT motif (Fig. 4B, line 3), duplication of the AGA motif (Fig. 4B, line 4) or inversion of the motif (Fig. 4B, line 5), all had detrimental effects on activation. In addition, insertion of a single (Fig. 4B, line 6) or double nucleotide (Fig. 4B, line 7) displacing the location of the half-sites relative to each other also affected activation severely. Our results are fully consistent with the previous identification of this sequence as an RcsB site and demonstrate that this sequence is required for rprA promoter function.
RprA-dependent translation of RpoS–lacZ
Previously, we had shown that RprA-dependent expression of RpoS could be detected, in the absence of DsrA, under conditions of osmotic shock (Majdalani et al., 2001). We have shown above that activation of the cps regulators increases the accumulation of RprA significantly. We would predict that this increase in RprA ought to lead to a concomitant increase in RpoS protein levels.
In an rcsC137 strain, the levels of RprA are about 50-fold higher than in a wild-type strain (Fig. 5); in an rcsC– strain, the levels are about 10-fold higher. To determine if the higher RprA amounts expressed in these strains lead to increased RpoS translation, we constructed a set of isogenic strains that carry an RpoS–lacZ translational fusion in the chromosome and various combinations of mutations in rcsC and rprA (Table 2). Cells were grown at 37°C and aliquots were taken during exponential growth for a β-galactosidase assay and parallel RNA and protein extractions. The Northern blot analysis of RNA samples confirmed what had been observed in the previous section, that RprA levels increase by 10- and 50-fold in rcsC and rcsC137 strain respectively (Table 2, line 1 columns 3 and 5). Quantification of RpoS in a Western blot indicates that RpoS amounts increase by about twofold in an rcsC– strain but RpoS–lacZ activity does not show any increase (Table 2, column 3); we do not know the reason for this discrepancy, but the overall effect is not large in either case. A strain carrying null mutations in both rcsC and rprA shows a slight decrease in the amount of RpoS protein but β-galactosidase activity remains at the same low level as in the parental strain (Table 2, column 4). In an rcsC137 mutant, RpoS levels increase to about 35 times the amount in the wild-type strain; RpoS–lacZ activity increases 20-fold (Table 2, column 5). Unexpectedly, although deletion of rprA in the rcsC137 host causes a 10-fold drop in RpoS amounts and about a sixfold decrease in β-galactosidase activity, these levels are still higher than in the wild-type strain (compare columns 1 and 6). This suggests that the rcsC137 mutation has both a major RprA-dependent and a smaller, RprA-independent effect on RpoS. Deletion of dsrA in this strain indicates that the residual RpoS expression is not DsrA-dependent (data not shown). The residual effect of rcsC137 on RpoS expression could be due to increased transcription or decreased degradation of RpoS, in-creased transcription of other small RNAs, or more indirect effects on cell physiology.
|WT||rprA −||rcsC −||rcsC − rprA −||rcsC137||rcsC137 rprA −|
|RpoS–lacZ specific activity||1||0.7||1||1||22||3.5|
Comparison of the action of RprA, the second small RNA found to positively regulate RpoS translation, to DsrA, provides an excellent opportunity to start to define essential elements for small RNA action on a given target. In this work, we find that DsrA and RprA share two common target sites on the RpoS leader RNA: a 5′ unpaired region of the RpoS message (Fig. 2A, nt 107–114) and a second stretch believed to pair with the region near the site for initiation of translation, leading to sequestration of the S–D region of the RpoS mRNA (Fig. 2A, nt 115–128). Mutations that affect the pairing between the small RNA and the leader in either of these two stretches is sufficient to inhibit small RNA action; thus contact with both regions is necessary for either DsrA or RprA to stimulate RpoS translation.
The region at nt 107–114 of the RpoS leader is not predicted to pair with the downstream message of RpoS. The absence of a pairing partner within RpoS mRNA is supported by the observation that changes in this sequence are well tolerated, as long as a complementary small RNA activator is available. We propose that interaction at this sequence (or possibly anywhere in this upstream single-stranded RNA leader sequence) can serve to initiate small RNA:mRNA interactions. The second region of pairing between the small RNAs and the RpoS leader is the business end. This region of the RpoS upstream mRNA (nt 115–128 in Fig. 2A) is predicted to pair with the region of the RpoS leader near the start of translation (Brown and Elliott, 1997). Although the S–D sequence may not itself be paired, double-stranded RNA just upstream and downstream of the S–D sequence might be expected to block ribosome binding. Thus, it seems likely that translation will require the combined disruption of the majority of the four G:C pairs that the RpoS leader makes with itself, at positions 117/219, 118/218, 119/216 and 120/215 as well as the UG and CG pairs at positions 124/208 and 125/207. Complementarity with the inhibitory region of the RpoS leader at these five positions is provided by both DsrA and RprA, and we have now demonstrated a pairing requirement between both DsrA (Majdalani et al., 1998) and RprA (Fig. 2) to positions U107-U112 and to U124 of the RpoS leader.
How important is the position of the sequences within the small RNAs? From an examination of these two cases, we note a number of similarities and differences: First, although the region of RpoS complementarity is primarily within the first stem–loop of DsrA (Majdalani et al., 1998), it lies in a second short stem–loop in RprA. Thus, ‘positive regulation’ need not be encoded in the initial stem–loop, as has been suggested (Lease et al., 1998). Second, in both molecules, the A of DsrA and RprA that pairs with U124 in the RpoS leader is in the first position in a predicted 5 nt loop structure; both of these loops also end with a UC, also predicted to pair with the RpoS leader. Mutagenesis experiments with the DsrA loop sequence support the importance of the C in the fifth position, although changes in U were tolerated, as were changes at the second and third positions in the loop (N. Majdalani, unpublished data). In RprA, the second and third positions of the loop are predicted to remain unpaired; mutations at position three (A39) were relatively free of effect on RprA function (Fig. 3B). What the functional significance of these similarities is will require a detailed examination of both RNA structures and interactions. Third, small RNA sequences involved in RpoS pairing are found in predicted stems in both small RNAs; thus possible pairing partners within the same RNA do not preclude a role for these sequences in interacting in trans with another RNA. Whether Hfq, the RNA binding protein that is necessary for the action of both DsrA and RprA, mediates melting and/or interactions of the small RNAs with their target is thus far unknown. It stimulates interactions between complementary RNAs, but it is not known if it specifically promotes pairing or disrupts alternative structures, allowing pairing to proceed more efficiently (Moller et al., 2002; Zhang et al., 2002). Fourth, it is not necessary that small RNA sequences interacting with the RpoS leader be contiguous. Although they are contiguous for DsrA, in RprA, 9 nt intervene between the region interacting with the 5′ unpaired RpoS leader and the region interacting with the S–D sequestration region. The interaction of DsrA with another target, hns, appears to involve sequences at both the beginning and the end of the hns target (Lease and Belfort, 2000). Therefore, in searching for complementarity for a given small RNA and possible targets, separated stretches of complementarity should be considered ac-ceptable; unfortunately, this degree of freedom makes it more difficult to identify unique matches.
Comparing the regions of complementarity between RpoS mRNA and either DsrA or RprA suggests possible minimum requirements needed for an sRNA to positively regulate RpoS translation. In a recent genome-wide search for small RNAs in E. coli, two more small RNAs were found that can upregulate RpoS–lacZ when present on a multicopy plasmid (Wassarman et al., 2001). Both contain sequences capable of pairing with the RpoS translation inhibitory region, but would suggest further flexibility in the minimum pairing requirements derived from RprA and DsrA alone. Further study and mutation of these small RNAs will be necessary to confirm this complementarity as significant and further define what variation in the sequence is acceptable, as well as to determine if these additional small RNAs have a biological role in regulating RpoS translation, as DsrA and RprA do.
Although DsrA and RprA share sequence complementarity to RpoS, Hfq dependence, and most likely share a general mechanism of action, one reason for two (and possibly more) such small RNAs to have evolved in E. coli and related bacteria is revealed by the very different conditions under which they are synthesized. DsrA has been shown to increase in synthesis and stability at low temperature, and the action of DsrA on RpoS translation is seen most dramatically at low temperatures. Temperature dependence of synthesis depends on the core promoter of DsrA (F. Repoila and S. Gottesman, in preparation).
Regulation of the rprA promoter is dependent on a site located just upstream of the −35 box and depends upon RcsB. Neither rcsC137 nor any of the plasmids activate the rprA promoter if the strain is mutant for RcsB or the fusion is missing or mutant for this upstream regulatory region. Although we have not directly demonstrated RcsB binding to this site in the rprA promoter, the genetics of RcsB dependence and similarity of this site to sites in the osmC1p and ftsA1p promoters (Carballes et al., 1999; Davalos-Garcia et al., 2001) that are also regulated by RcsB independently of RcsA supports the model that RcsB acts directly on the rprA promoter. RcsB was shown to stimulate the binding of RNA polymerase to the osmCp1 promoter at the RcsB site around the −35 box (Davalos-Garcia et al., 2001).
Our data suggests that RcsC both positively and negatively regulates RcsB activity. We believe the negative regulation (demonstrated by the higher basal level of RprA synthesis in an rcsC mutant) reflects the phosphatase activity of RcsC. Positive regulation can be seen in the rcsC137 host, in previous work on the cps system (Gottesman, 1995) and in the dependence on RcsC for stimulation of rprA by a number of the plasmids we isolated, including the rcsF plasmid. Our observation that RcsF acts via RcsC is somewhat different from previous proposals (Gervais and Drapeau, 1992), although the relationship of RcsC and RcsF was not extensively examined in that work. This issue is a subject of ongoing experiments, but we note that RcsF and its homologues contain sequences found in lipoproteins, consistent with an extracytoplasmic location that might act to transmit signals to RcsC. We interpret the failure of an rcsC– mutant to stimulate cps synthesis as a reflection of a requirement for much higher RcsB activation to stimulate the promoters dependent on an RcsA/RcsB heterodimer. This model remains to be proven, but we note that only the most active of the plasmids identified as stimulating rprA synthesis, the ones carrying rcsF, make cells mucoid. Although RprA synthesis does increase in the rcsC null mutant, the increase does not lead to increased RpoS expression under the conditions we tested. Therefore, the threshold for stimulation of the cps operon by activated RcsB may not be that different from the threshold needed for RcsB to, indirectly, stimulate RpoS synthesis.
Recent work on the RcsC/B system suggests that RcsC may act indirectly, through a phosphorelay, in which phosphate moves from the conserved histidine on RcsC to a conserved aspartate in a response regulator domain at the C-terminus of RcsC. From there, it can be transferred to the conserved histidine of YojN, a protein with homology to other phospho-transfer proteins, and from YojN to RcsB (Takeda et al., 2001). None of the experiments presented here address this downstream part of the phosphorelay. In preliminary experiments, YojN does participate in signalling for RprA synthesis, as predicted by this phosphorelay; the role of YojN in both rprA and cps synthesis will be discussed in more detail in future work (N. M. and M. Heck, unpublished results).
What are the inducing signals for the RcsC/RcsB system? It has been found that osmotic shock leads to a transient, RcsC-dependent induction of cps (Sledjeski and Gottesman, 1996). RpoS synthesis is also induced by osmotic shock, and, in the absence of DsrA, this induction is RprA dependent (Majdalani et al., 2001). Mutations that can be interpreted as perturbing the cell surface have been found to induce cps synthesis, including rfa mutations, disrupting the core of LPS synthesis (Parker et al., 1992) and mutations in synthesis of other small oligosaccharides (mdo) (Ebel et al., 1997). Overproduction of a membrane protein with a domain homologous to DnaJ, DjlA, also induces cps (Clarke et al., 1997; Kelley and Georgopoulos, 1997; (Genevaux et al., 2001); whether the appearance of dnaK and part of dnaJ on one of our stimulatory plasmids (Table 1) is part of the same pathway is not clear. Other membrane proteins have also been found to induce the cps system (Chen et al., 2001), and we found a number of membrane proteins in our screen for plasmids that activate rprA–lacZ expression (Table 1). Thus, our results are consistent with the idea that perturbation of the cell surface, including the membrane, activates the RcsC phosphorelay (Conter et al., 2002). All of these inducing situations, including those described here, require RcsB and RcsC for activity. However, the exact nature of the signal or how it is sensed remains to be discovered, and the explanation for some of our isolates (for instance, spf, encoding Spot 42 RNA) remains unknown.
One might also ask what the physiologic result of RcsC activation is. We have shown here that RprA and RpoS increase in an rcsC137 host, when capsule synthesis also increases. Other targets identified thus far include one of the ftsZ promoters and osmC, an envelope protein of unknown function (Gervais et al., 1992; Carballes et al., 1999; Davalos-Garcia et al., 2001). It is noteworthy that osmC has two promoters, one regulated by RcsB and the other regulated by RpoS (Davalos-Garcia et al., 2001); thus, one might predict that at low levels of activation of the RcsC system, osmC will be made from the first promoter, and at higher levels, it will be made from both, by virtue of RpoS activation. Thus two operons encoding components of the envelope and cell surface, osmC and cps, are coregulated via RcsC with an essential cell division protein and a major stress response regulator. What growth condition unites these various targets is not yet known, but it is tempting to imagine some condition where cell surface stress is relieved by capsule synthesis and by OsmC as well as RpoS and possibly other, thus far unknown, RprA target(s). Identification of other RprA targets, if they exist, will help in understanding the nature of the regulatory circuit controlled by RcsC and RcsB.
Strains and plasmids
Strains used in the experiments of Fig. 3, bearing the RpoS–lacZ translational fusion, were as described elsewhere (Majdalani et al., 1998), except that an ΔrprA5::cat sacB mutation (see below) was introduced into a subset of the strains. This RpoS–lacZ fusion does not carry the region of RpoS necessary for RpoS degradation. The following strains are ΔdsrAΔrprA5::cat sacB and carry either a wild-type RpoS–lacZ (U124:G208 in the mRNA leader region; NM22215, derived from NM22186) or one of three mutant versions; C124:U208 (NM22216, derived from NM22201), G124:C208 (NM22217, derived from NM22203) or G124:U208 (NM22219, derived from NM22207). A fourth mutant carries the NcoI′ mutation in its leader region (NM22205) and is rprA+ (Majdalani et al., 1998).
Strains used in the experiments in Table 2 carried the RpoS750–lacZ fusion described previously (Sledjeski et al., 1996); this fusion does carry the degradation signal for RpoS. All strains were derived from DJ480, a Δlac derivative of MG1655, obtained from D. Jin, NCI. The strains were lysogenized with the λGN272 phage carrying the fusion to create EM1050, a gift from E. Massé, NIH; rcs mutant alleles were introduced into either the ΔrprA10 (described below) or rprA+ derivative by P1 transduction. Specific derivatives are de-scribed in the legend to Table 2. rcsC137 was most easily transduced by lysogenizing the strain with P1 kan clr100 (from M. Yarmolinsky, NCI), selecting mucoid kanamycin-resistant lysogens, and inducing the lysogen by a temperature shift. This minimized the large number of second site, non-mucoid revertants seen when P1 infection was used to obtain a lysate for transduction; a linked chloramphenicol resistance insertion was used to select for transductants.
Strain NM22553 is a derivative, by P1 transduction, of NM22506 carrying ΔdsrA and hfq1::kan, a KanR insert near the 5′-end of hfq that is phenotypically Hfq– (Tsui et al., 1994). NM22554 is the isogenic strain that carries ΔdsrA and hfq2::kan, a KanR near the 3′-end of hfq that is phenotypically Hfq+ (Tsui et al., 1994).
The DH100 (rprA43–lacZ ), DH200 (rprA79–lacZ ), DH 300 (rprA142–lacZ ), DH400 (rprA55–lacZ ), DH401(rprA55 (G46A)–lacZ ), DH403 (rprA55(AGA to TCT)–lacZ, DH406 (rprA55(TCT to AGA)–lacZ), DH407 (rprA55(AGA-TCT to TCT-AGA)–lacZ ), DH408 (rprA55(G insertion 43)–lacZ ), and DH409 (rprA55(GG insertion 43)–lacZ ) strains were all constructed by crossing λRS45 with pDH100, pDH200, pDH300, pDH400, pDH401, pDH403, pDH406, pDH407, pDH408 and pDH409, respectively, to isolate λ lac transducing phage; these phage were used to lysogenize DJ480. Single-copy lysogens were confirmed by polymerase chain reaction (PCR) according to Powell and colleagues (Powell et al., 1994). The fusions are depicted in Fig. 4.
Derivatives of DH300 carrying various mutations in the Rcs pathway were made by P1 transduction; strain names are given in the figure legends. Specific strains are described in the legend to Fig. 5, except for DH314 (lon) and DH315 (rcsC137, rcsA), and plasmid pATC400 (pRcsA) (Torres-Cabassa and Gottesman, 1987), discussed in the text but not in the figure.
The ΔrprA null strains were constructed by the method of Yu and colleagues (Yu et al., 2000) in two steps. In the first step, primers were designed to create a linear PCR product encoding a chloramphenicol acetyl transferase–sucrase (cat-sacB) cassette flanked by stretches of DNA homologous to the regions upstream and downstream of the rprA reading frame. This PCR product was introduced by electroporation into strain DY330 to allow recombination onto the chromosome. Stable recombinants were selected on LB chloramphenicol plates (LB-Cm, 10 µg ml−1 of chloramphenicol), purified once and screened for sucrose sensitivity on M63 minimal medium (MM) plates (0.2% glycerol; 5% sucrose); this strain is NM3; the mutation is ΔrprA5::cat-sacB. This deletion/insertion was then linked to aroD by P1 transduction of the cassette into strain AB2848 (aroD–; from ATCC). Sucrose-sensitive, CmR cells were purified and tested for lack of growth on unsupplemented MM (Aro–); one such isolate was strain NM109. A P1 lysate from this strain was used to transduce the deletion into recipient hosts, selecting for CmR and screening for aroD–. In a second step, an unmarked deletion of rprA was created by electroporation of a single-stranded oligonucleotide carrying homology to the regions upstream and downstream of the insertion site into NM10, carrying the rprA5::cat-sacB insertion. Cells were plated on M63 MM plates with glycerol and 5% sucrose to select for sucrose-resistant growth. Colonies were screened for CmS; the resulting unmarked deletion (ΔrprA10) was transduced with P1 into the rprA5::cat-sacB aroD hosts, selecting for growth on M63 minimal glucose medium. AroD+ cells were screened for CmS and checked by PCR.
pSAC1 carries rprA and neighbouring sequences in a pBR322-related vector (Majdalani et al., 2001); RprA is expressed from its own promoter on this plasmid. To construct the pBAD-rprA plasmid derivatives (pNM series), PCR was used to generate promoter-less DNA fragments with the forward primer (AAGTCCGTATGCCTACTATGGCCACACG GTTATAAATC) carrying an MscI site and the reverse primer (ACGTACGTGAATTCGAAGAGAGTTCACAGTATC) containing an EcoRI site. The PCR fragments were then cloned into the same sites in pNM12, a derivative of pBAD24 (Guzman et al., 1995) that contains an MscI site at the −7 to −2 region upstream of the +1 transcription start site (Majdalani et al., 1998; 2001). The resulting plasmids express rprA and derivatives under the control of the araBAD promoter (ParaBAD). All constructs were confirmed by sequencing using the pBAD 5′-primer (CTGTTTCTCCATACCCGTT) (Guzman et al., 1995).
Transformations were carried out either by electroporation using a GenePulsar II (Bio-Rad) or by the transformation and storage solution method (TSS) (Chung et al., 1989).
The QuickChange (Stratagene) mutagenesis kit was used according to manufacturer's specifications. It was used to generate all the rprA mutants using pSAC1 as template; mutants were cloned into the MscI–EcoRI sites of pNM12 as described above, to place them under ParaBAD control. Mutations in rprA were confirmed by sequencing all plasmids.
Construction of rprA–lacZ fusions
Polymerase chain reaction fragments corresponding to the promoter regions from −142 to −1, or −79 to −1 or −43 to −1 were amplified from pSAC1 and cloned into the EcoRI and BamHI sites of pRS415 (Simons et al., 1987) to yield plasmids pDH 100, 200 and 300 respectively. A PCR fragment was amplified from pDH300 using a 5′-primer corresponding to the −55 region of the promoter and a 3′-primer corresponding to an EcoRV site region internal to lacZ. This fragment was cloned into the EcoRI–EcoRV sites of pRS415 to yield pDH400. All subsequent pDH series plasmids were constructed in a similar manner with the 5′-primer carrying the mutation to be introduced.
The β-galactosidase activity of the lacZ fusion was assayed on a SpectraMax 250 (Molecular Devices) microtitre plate reader as described previously (Majdalani et al., 1998). Specific activities are represented as the Vmax divided by the OD600 and these units are about 25 times lower than standard Miller units.
Northern blot analysis
Northern blot analyses were performed as described elsewhere (Majdalani et al., 1998; 2001). The RprA.probe, a 5′-end biotinylated DNA oligonucleotide probe (GGGGATTTC CATGCTTATAAATCAATATGT) was used at 100 ng ml−1 and developed with Ambion's BrightStar Biodetect non-isotopic kit (Majdalani et al., 2001).
Western blot analysis
Western blots were performed as described by Majdalani and colleagues (Majdalani et al., 2001). From exponentially growing cells, a 1 ml aliquot was taken and mixed with 110 µl of 50% tricarboxylic acid (TCA). Samples were set on ice for 10 min, centrifuged for 10 min, and pellets were washed once with 50% acetone. Pellets were allowed to air-dry before resuspension in 100 µl of sample buffer. For stationary phase cells, 0.25 ml of cells was mixed with 110 µl of 50% TCA and treated as mentioned. The volume of sample loaded was adjusted according to OD600 readings.
Transformations were carried out by electroporation into DH300 and plated on MacConkey Lactose plates containing 50 µg ml−1 of ampicillin (MacLacAmp) to give about 2000 colonies per plate, and incubated at 37°C. We screened about 25 000 colonies and isolated 81 red colonies. Plasmid DNA was purified using Promega's Wizard Plus Minipreps kit and insert junctions were sequenced using pBRlib.for (CCT GACGTCTAAGAAACCATTATTATC) and pBRlib.rev (AACGA CAGGAGCACGATCATGCG) primers. The UW-GCG blast search program (http:www.ncbi.nlm.nih.govblast; version 2.0) was used to map these sequences in the E. coli genome (Altschul et al., 1990).
We thank members of the laboratory for suggestions throughout this work. We thank D. Jin for DJ480, M. Yarmolinsky for P1 kan clr100 and E. Massé for EM1050.
- 1990) A basic local alignment search tool. J Mol Biol 215: 403–410. , , , , and (
- 2001) Novel small RNA-encoding genes in the intergenic region of Escherichia coli. Current Biol 11: 941–950. , , , , , , and (
- 1988) Fine-structure mapping and identification of two regulators of capsule synthesis in Escherichia coli K-12. J Bacteriol 170: 2599–2611. , , and (
- 1996) Efficient translation of the RpoS sigma factor in Salmonella typhimurium requires Host Factor I, an RNA-binding protein encoded by the hfq gene. J Bacteriol 178: 3763–3770. , and (
- 1997) Mutations that increase expression of the rpoS gene and decrease its dependence on hfq function in Salmonella typhimurium. J Bacteriol 179: 656–662. , and (
- 1999) Regulation of Escherichia coli cell division genes ftsA and ftsZ by the two-component system rcsC-rcsB. Mol Microbiol 34: 442–450. , , , and (
- 2001) Characterization of the RcsC YojNRcsB phosphorelay signaling pathway involved in capsular synthesis in Escherichia coli. Biosci Biotechnol Biochem 65: 2364–2367. , , , , , and (
- 1989) One-step preparation of competent Escherichia coli: transformation and storage of bacterial cells in the same solution. Proc Natl Acad Sci USA 86: 2172–2175. , , and (
- 1997) Point mutations in the transmembrane domain of DjlA, a membrane-linked DnaJ-like protein, abolish its function in promoting colanic acid production via the Rcs signal transduction pathway. Mol Microbiol 25: 933–944. , , and (
- 2002) The RcsCB His-Asp phosphorelay system is essential to overcome chlorpromazine-induced stress in Escherichia coli. J Bacteriol 184: 2850–2853. , , , and (
- 2001) Regulation of osmC gene expression by the two-component system rcsB-rcsC in Escherichia coli. J Bacteriol 183: 5870–5876. , , , , and (
- 1997) Inactivation of mdoH leads to increased expression of colanic acid capsular polysaccharide in Escherichia coli. J Bacteriol 179: 6858–6861. , , , and (
- 2001) DjlA is a third DnaK co-chaperone of Escherichia coli, and DjlA-mediated induction of colanic acid capsule requires DjlA–DnaK interaction. J Biol Chem 276: 7906–7912. , , , , and (
- 1992) Identification, cloning, and characterization of rcsF, a new regulator gene for exopolysaccharide synthesis that suppresses the division mutation ftsZ84 in Escherichia coli K-12. J Bacteriol 174: 8016–8022. , and (
- 1992) The rcsB gene, a positive regulator of colanic acid biosynthesis in Escherichia coli, is also an activator of ftsZ expression. J Bacteriol 174: 3964–3971. , , and (
- 1995) Regulation of capsule synthesis: Modification of the two component paradigm by an accessory unstable regulator. In Two-Component Signal Transduction. Hoch, J.A. and Silhavy, T.J. (eds). Washington, DC: American Society for Microbiology, pp. 253–262. (
- 1991) Regulation of capsular polysaccharide synthesis in Escherichia coli K12. Mol Microbiol 5: 1599–1606. , and (
- 1995) Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol 177: 4121–4130. , , , and (
- 2000) The general stress response in Escherichia coli. In Bacterial Stress Responses. Storz, G. and Hengge-Aronis, R. (eds). Washington: D.C. American Society for Microbiology Press, pp. 161–178. (
- 1997) Positive control of the two-component RcsC/B signal transduction network by DjlA: a member of the DnaJ family of molecular chaperones in Escherichia coli. Mol Microbiol 25: 913–931. , and (
- 1997) Interaction of the regulator proteins RcsA and RcsB with the promoter of the operon for amylovoran biosynthesis in Erwinia amylovora. Mol Gen Genet 256: 72–83. , , , and (
- 2000) A trans-acting RNA as a control switch in Escherichia coli: DsrA modulates function by forming alternative structures. Proc Natl Acad Sci USA 97: 9919–9924. , and (
- 1998) Riboregulation in Escherichia coli: DsrA RNA acts by RNA: RNA interactions at multiple loci. Proc Natl Acad Sci USA 95: 12456–12461. , , and (
- 1998) DsrA RNA regulates translation of RpoS message by an anti-antisense mechanism, independent of its action as an antisilencer of transcription. Proc Natl Acad Sci USA 95: 12462–12467. , , , , and (
- 2001) Regulation of RpoS by a novel small RNA: the characterization of RprA. Mol Microbiol 39: 1382–1394. , , , , and (
- 2002) Hfq: a bacterial Sm-like protein that mediates RNA–RNA interaction. Molec Cell 9: 23–30. , , , , , , and (
- 1992) Role of the rfaG and rfaP genes in determining the lipopolysaccharide core structure and cell surface properties of Escherichia coli K-12. J Bacteriol 174: 2525–2538. , , , , , and (
- 1994) Rapid confirmation of single copy lambda prophage integration by PCR. Nucleic Acids Res 22: 5765–5766. , , , , , and (
- 2001) Signal transduction cascade for regulation of RpoS: temperature regulation of DsrA. J Bacteriol 183: 4012–4023. , and (
- 1987) Improved single and multicopy lac-based cloning vectors for protein and operon fusions. Gene 53: 85–96. , , and (
- 1996) Osmotic shock induction of capsule synthesis in Escherichia coli K-12. J Bacteriol 178: 1204–1206. , and (
- 1996) The small RNA, DsrA, is essential for the low temperature expression of RpoS during exponential growth in Escherichia coli. EMBO J 15: 3993–4000. , , and (
- 2001) Hfq is necessary for regulation by the untranslated RNA DsrA. J Bacteriol 183: 1997–2005. , , and (
- 1991) Regulation of capsular polysaccharide synthesis in Escherichia coli K12. Mol Microbiol 5: 1599–1606. , and (
- 2001) A novel feature of the multistep phosphorelay in Escherichia coli: a revised model of the RcsCYojNRcsB signalling pathway implicated in capsular synthesis and swarming behavior. Mol Microbiol 40: 440–450. , , , , and (
- 1987) Capsule synthesis in Escherichia coli K-12 is regulated by proteolysis. J Bacteriol 169: 981–989. , and (
- 1994) Characterization of broadly pleiotropic phenotypes caused by an hfq insertion mutation in Escherichia coli K-12. Mol Microbiol 13: 35–49. , , and (
- 1997) The E. coli signal recognition particle is required for the insertion of a subset of inner membrane proteins. Cell 88: 187–196. , , and (
- 2002) Small RNAs in bacteria: diverse regulators of gene expression in response to environmental changes. Cell 109: 141–144. (
- 2001) Identification of novel small RNAs using comparative genomics and microarrays. Genes Dev 15: 1637–1651. , , , , and (
- 2000) The RcsAB box. J Biol Chem 275: 7013–7020. , and (
- 1999) Identification of an RcsA/RcsB recognition motif in the promoters of expopolysaccharide biosynthetic operons from Erwinia amylovora and Pantoea stewartii subspecies Stewartii. J Biol Chem 274: 3300–3307. , , , , , and (
- 2000) An efficient recombination system for chromosome engineering in Escherichia coli. Proc Natl Acad Sci USA 97: 5978–5983. , , , , , and (
- 2002) The Sm-like Hfq protein increases OxyS RNA interaction with target mRNAs. Molec Cell 9: 11–22. , , , , and (