doi:10.1016/j.jmb.2006.09.004 J. Mol. Biol. (2006) 364, 1–8

COMMUNICATION

Conserved Small Non-coding RNAs that belong to theσE Regulon: Role in Down-regulation of OuterMembrane Proteins

Jesper Johansen, Anders Aamann Rasmussen, Martin Overgaardand Poul Valentin-Hansen⁎

Department of Biochemistryand Molecular Biology,University of SouthernDenmark, Campusvej 55,DK-5230, Odense M, Denmark

Abbreviations used: sRNA, smallOM, outer membrane; OMP, outer mIPTG, isopropyl-β-d-thiogalactopyraguanosine 3′,5′-bispyrophosphate; σphase sigma factor in E. coli; σN, thecontrolling nitrogen use in E. coli.E-mail address of the correspondi

valentin@bmb.sdu.dk

0022-2836/$ - see front matter © 2006 E

Enteric bacteria respond to misfolded proteins by activating the transcrip-tion of “heat shock” genes. These genes are arranged in two major regulonscontrolled by the alternative sigma factors σH and σE. The two transcriptionfactors coordinate the stress response in different cellular compartments; theσH regulon is induced by stress in the cytoplasm whereas the σE regulon isactivated by stress signals in the cell envelope. In Escherichia coli σE plays acentral role in maintaining cell envelope integrity both under stressconditions and during normal growth. Previous work established that σE

is essential for viability of the bacterium and up-regulates expression ofapproximately 100 protein-encoding genes that influences nearly everyaspect of the cell envelope. Moreover, the expression of several outermembrane proteins is down-regulated upon σE activation. Here, we showthat two Hfq-binding small RNAs, MicA and RybB, are under positivecontrol of σE. Transient induction of RybB resulted in decreased levels of themRNAs encoding OmpC and OmpW. σE-mediated regulation of ompC andompW expression was abolished in strains lacking RybB or Hfq. RecentlyMicA was shown to act in destabilizing the ompA transcript when rapidlygrown cells entered the stationary phase of growth. Also, the alternativesigma factor down-regulates this message in a small non-coding RNA-dependent fashion. These findings add the σE regulon to the growing list ofstress induced regulatory circuits that include small regulatory RNAs andprovide insight in a homeostatic loop that prevent a build-up ofunassembled outer membrane proteins in the envelope.

© 2006 Elsevier Ltd. All rights reserved.

Keywords: RybB RNA; Hfq; outer membrane proteins; σE regulon; non-coding RNAs

*Corresponding author

Bacteria respond to changes in their environmentby global changes in transcription. These changes intranscription are often accomplished by the induc-tion of alternative sigma factors, which direct RNApolymerase to specific promoters. In enteric bacteriaone of the key pathways involved in maintaining

non-coding RNA;embrane protein;noside; ppGpp,S, the stationarysigma factor

ng author:

lsevier Ltd. All rights reserve

cell envelope integrity during stress and normalgrowth is controlled by the alternative sigma factorσE. The rpoE gene encoding σE is essential inEscherichia coli, and an unknown suppression hasto occur to allow growth of ΔrpoE strains.1 Activa-tion of σE is triggered by various stress signals,which are sensed in the envelope and communicatedto the cytoplasmic compartment by a complex signaltransduction pathway. The primary point of regula-tion of σE is its interaction with the inner membrane-spanning protein RseA and the regulated proteoly-sis of this anti-sigma-factor by the transmembraneproteases DegS and RseP (formerly known as YaeL).Under non-stress conditions σE is sequestered byRseA, thereby decreasing the cytoplasmic availabil-ity of σE for transcription initiation. DegS, the

d.

†http://www.genolist.pasteur.fr/Colibri/

2 σE-dependent Small Regulatory RNAs

primary sensor of the signal transduction pathway,becomes activated when its periplasmic PDZ do-main recognizes specific unfolded C-terminal motifsof outer membrane porins.2 Activated DegS initiatesthe cleavage of RseA which triggers a proteolysiscascade that leads to release of σE and in turn toinduction of the envelope stress response genes.3–5

Several lines of evidence, however, indicate thatthe σE system can be activated by other stresssignalling pathways. First, it has been observed thatoverexpression of full-length OmpC can still induceσE in a strain that should not be capable of sensingunfolded porins due to a deletion of the PDZdomain of DegS.3 Second, the periplasmic proteinRseB, encoded together with σE and RseA in therpoE-rseABCoperon, binds to RseA and the interac-tion appears to modulate both activity and stabilityof RseA. Since RseB also binds to misfoldedperiplasmic proteins, it has been proposed that thepresence of such proteins might titrate RseB awayfrom RseA and lead to an increase in active unboundσE.6 Third, σE activity increases upon entry intostationary phase.7 This growth phase dependentactivation of σE is dependent upon the metabolicsignalling molecule guanosine 3′,5′-bispyrophos-phate (ppGpp) and does not require RseA.8

Various genomic strategies have been used tounravel the E. coli σE regulon.9–12 Taken together theapproaches have led to the identification of 47 σE-dependent promoters, which control the expressionof approximately 100 genes. The majority of thesegenes participate in the synthesis, assembly andhomeostasis of outer membrane proteins (OMPs)and lipopolysaccharide, the key constituents of theouter membrane. Also σE up-regulates the expres-sion of itself, its negative regulators RseA and RseBand the positive regulator of the cytoplamic stressresponse σH. An important feature of the rpoE-rseABC operon is the presence of two conserved σE

promoters: one upstream of rpoE and a secondupstream of rseA.11 The autoregulation of σE andRseA ensures a rapid increase in proteins needed forenvelope homeostasis under stress conditions and,moreover, provides an efficient mechanism forshutting off σE activation when stress signals arereduced or removed. In addition, the expression ofseveral OMPs (i.e. OmpA and the porins OmpC,OmpF, OmpWand OmpX) is down-regulated uponinduction of σE.11,12 How this down-regulation maybe accomplished is not well understood.Here we report that two conserved small non-

coding RNAs (sRNAs), MicA (alias SraD) and RybB,belong to the σE regulon. Transcription of the micAand rybB genes is strictly dependent on σE. We usedmicroarrays to identify putative targets for the twosRNAs and found that induction of RybB stronglydown-regulated ompC and ompW (yciD)13,14 and, inkeeping with previous work, that a transientexpression of MicA led to a strong reduction of theompAmRNA level. The effect of RybB expression onompC and ompW mRNA levels was confirmed byNorthern blot analysis. Moreover, σE-mediatedregulation of ompA and ompC/ompW expression

was dependent on MicA and RybB, respectively, aswell as on the Sm-like chaperone Hfq. Takentogether, these findings indicate that σE-dependentsRNAs are important players in a homeostaticregulatory loop that prevent a build-up of unas-sembled OMPs in the envelope during stressconditions.

Transcriptional regulation of MicA and RybBRNAs

The 72 nt MicA RNA was identified by acomputational approach that focused on promoterelements recognized by the E. coli house keepingRNA polymerase (Eσ70) and on Rho-independentterminators present in “empty” intergenic regions.The transcript of MicA turned out to be shorter thanpredicted and 5′ end mapping data assigned thetranscription initiation site to a promoter with a lessperfect match to the consensus sequences for σ70-RNA polymerase. The synthesis of MicA is tightlyregulated; the RNA transcript is present at lowlevels during exponential growth in LB medium at37 °C and exhibits a strong increase in abundanceupon entry into stationary phase.15 The regulatoryfactor(s) that mediates this growth phase variationin MicA expression has not been identified. Recentwork demonstrated that E. coli MicA RNA is anantisense regulator, which down-regulates the ompAmRNA level when rapidly grown cells enterstationary phase. The MicA-mediated decay of theompA mRNA depends on Hfq and in vitro studiessuggest that binding of MicA specifically interfereswith ribosome binding to the translational initiationregion of the messenger.16,17

In an attempt to gain insight into transcriptionalregulation of micA we inspected the DNA segmentupstream of the 5′ end of the transcript. No obvioussequences that resemble the σ70 consensus sequenceseem to be present in the immediate upstreamregion. Rather, this region contains −10 and −35sequences with a perfect match to σE-dependentpromoters (i.e. GAACTT-N16-TCTNA).18 This find-ing prompted us to search for additional sRNAs thatmight be under positive control of σE. For this weused pattern searches at the Colibri database† andfocused on DNA sequences inside intergenic regionsthat contain conserved σE and rho-independentterminator motifs separated by 60–140 base-pairs.The search resulted in an additional candidate, theRybB RNA, which was identified in an approachusing comparative genomics and microarrays.19 ThesRNAwas found to co-immunoprecipitate with Hfq,suggesting that it might act via base-pairing to targetmRNAs. The expression pattern of RybB is similar tothat found for MicA20 and the sequence of the twosRNA genes, as well as the immediate upstreamregion, is conserved in a wide range of Enterobac-teriaceae (sequence alignments of the promoterregions are presented in Figure 1). These findings

3σE-dependent Small Regulatory RNAs

suggested that expression of MicA and RybB mightbe under positive control of σE. Indeed, such aregulation would be consistent with the growthphase-dependent transcription activation of the twosRNAs.To test this possibility, we assayed the expression

pattern of the MicA and RybB RNAs in isogenicwild-type and ΔrpoE strains.21 Total RNA sampleswere isolated from the strains grown to log phaseand early stationary phase and subjected toNorthern analysis. The experiment is presented inFigure 2(a). In accordance with previous studies theRNA species increased in abundance during tran-sition from vegetative cells to stationary cells in thewild-type strain.15,17,20 In contrast, MicA and RybBexpression could not be detected in the ΔrpoEstrain.To verify this result, we constructed a low-copy-

number plasmid (pNDM-rpoE) that expressed σE

under the control of the inducible lac promoterderivative PA1/O4.

22 A wild-type strain was trans-formed with the plasmid, and as a control withplasmid pNDM220 (empty vector).23 The resultingstrains were grown to exponential phase (A600∼0.4),rpoE expression was induced with increasingamounts of IPTG for 10 min, and the RNA levelsof MicA and RybB were determined by Northernblotting of total RNA extracts. In Figure 2(b), theupper panel shows the induction of rpoE transcrip-tion as a function of added IPTG. The lower panelsshow that a short-term expression of rpoE induced

the expression of the two sRNAs and that the MicAand RybB levels correlated with the mRNA levels ofrpoE. Also we determined the ompA mRNA levelfrom total RNA isolated from wild-type and ΔmicAstrains expressing rpoE from the low–copy numberplasmid. The results are presented in Figure 2(c) andshow that the ompA mRNA steady-state level inwild-type cells dropped rapidly as rpoE synthesisincreased, whereas induction of rpoE synthesis in theΔmicA mutant strain had no detectable effect on theompA mRNA level. Thus, micA and rybB transcrip-tion requires σE and the alternative sigma factorindirectly down-regulates ompA mRNA in a MicA-dependent fashion.Hfq binding sRNAs are believed to act as anti-

sense regulators and have in many cases beenshown to affect the stability of target mRNAs.24,25

In an attempt to identify putative targets for RybB aswell as additional targets for MicA, we thereforeexamined the effects of controlled transient expres-sion of the two RNAs on mRNA profiles by DNAmicroarrays. To this end we constructed ΔmicA/pNDM-micA and ΔrybB/pNDM-rybB strains car-rying an ectopic copy of the sRNA gene under thecontrol of the inducible PA1/O4 promoter. Thestrains, along with control strains carrying theempty vector pNDM220, were grown in LBmediumto late log phase and induced with IPTG for 10 min.Total RNAwas extracted from the two set of strains,treated with DNase I and converted into fluorescentlabelled cDNA by reverse transcription. cDNA from

Figure 1. Sequence conserva-tion analysis of micA and rybBpromoters in different enterobac-teria. σE promoter elements areframed and arrows indicate initia-tion points of RNA synthesis.

Figure 2. Regulation ofmicA and rybB synthesis by thetranscriptional factor σE. (a) Effects of rpoE deletion onMicA and RybB RNA levels. Isogenic rpoE+ (BW25113)41

and ΔrpoE (ECA101)21 strains were grown in LB mediumand samples for total RNA extraction were removed at theindicated A600 values. For determination of amounts of

4 σE-dependent Small Regulatory RNAs

control samples and sRNA induced samples werelabelled with Cy3 (green) and Cy5(red), respective-ly, and relative mRNA levels were determined bytwo-colour hybridization to E. coli glass slide cDNAmicroarrays (purchased from Ocimum Biosolu-tions). The induction of MicA and RybB underthese conditions was six- to eightfold higher than theexpression from chromosomal copies. For RybBmarked changes were observed for two genesencoding the outer membrane proteins OmpC andOmpW (down-regulated 11- and sevenfold, respec-tively, upon rybB induction) whereas only one spotcorresponding to OmpA was strongly affected byMicA expression (down-regulated ∼12-fold).The microarray data suggested that RybB, like

MicA, is a regulatory RNA that acts in modulatingthe expression of outer membrane proteins (OMPs).To examine such a role of the sRNA we first testedwhether the ompC mRNA levels are affected bydeleting the rybB gene. Wild-type and ΔrybB strainswere cultivated in LB, cells were harvested atvarious time points after inoculation, RNA wasextracted, and Northern analysis was performed onequal amounts of total RNA to determine therelative ompC mRNA levels at each point duringgrowth. The results are presented in Figure 3(a) andshow that the ompC mRNA levels were significantlyelevated in the strain deleted for rybB (three- tofourfold during transition from exponential phase tostationary phase). Next we investigated by Northernanalysis the effect of transient RybB expression on

ompC mRNA levels. As for the arrays describedabove, RybB was induced from a low copy numberplasmid in a strain deleted for the chromosomal copyof the sRNA. In Figure 3(b), the upper panel showsthe induction of the RybB RNA as a function ofadded IPTG. The middle panel shows that theinduction of RybB led to a rapid decrease inompCmRNA and the RybB levels inversely correlatedwith the levels of ompCmRNA. Finally, we askedwhether RybB and Hfq are also required for σE-mediated regulation of ompC expression. The resultspresented in Figure 3(c) show that the steady-stateompC mRNA level remained unaffected of elevatedexpression of rpoE when RybB or Hfq was absentfrom the cell. Similar results were obtained with theompW mRNA (Figure 3). Consistent with the micro-arrays, however, the effect of RybB was stronger onthe ompC mRNA than on the ompW mRNA.

Conclusions

Over the past decade, it has become clear thatcells contain a wealth of small, non-coding RNAs

MicA and RybB RNAs, 5 μg of the total RNA sampleswere denatured and separated on urea −8% (w/v)polyacrylamide gels. Northern blot was performed withspecific uniformly labelled RNA probes.17 (b) Effects ofrpoE expression on MicA and RybB RNA levels. The rpoEgene was cloned downstream of the isopropyl-β-D-thiogalactopyranoside (IPTG)-inducible PA1/O4-promoterin the low–copy number expression vector pNDM220,which contains the lacIq gene.23 Wild-type cells (S∅928)harbouring pNDM220 and pNDM-rpoE, respectively,were grown in LB medium to an A600 of 0.3, and theninduced. A sample was removed 10 min after IPTGaddition and processed as for (a). The rpoE mRNA (upperpanel) as well as MicA and RybB sRNAs (lower panels)were analysed by Northern blot. (c) Effect of rpoEexpression on ompA mRNA levels in wild-type andΔmicA cells (experiments as in (b)). The strains usedwere S∅928 and S∅928ΔmicA carrying pNDM220 (con-trol) or pNDM-rpoE. The rpoE gene was amplified fromS∅928 by PCR using the primers pNDM220-rpoE-Kpn(5′-GGGGTACCTTTGGTTTGGGGAGACTTTACC-TCGG) and pNDM220-rpoE-rev (5′-CCGGAATTCCCGC-TATCGTCAACGCCTGATAAG). The PCR product wascloned into KpnI/EcoRI digested pNDM220. The 32P-labelled RNA probes used for detection of RybB RNA andrpoE mRNAwere generated by in vitro transcription withT7 RNA polymerase.42 DNA templates carrying a T7promoter were made by PCR using the primers: oRybB fw(5′-GCCACTGCTTTTCT TTGATGTCCCC)/oT7RybB rev( 5 ′ -GGGCCTAATACGACTCACTATAGGGCA-CAAAAAA CCCATCAACCTTGAACCGA) and orpoEfw (5′-GTTGAACGGGTCCAGAAGG)/oT7rpoE rev (5′-GGGCCTAATACGACTCACTATAGGGAGCGCACGA-TAGGCTTTAAT). The RNA probes used for detection ofompA mRNA and MicA sRNAwere as described.17

5σE-dependent Small Regulatory RNAs

that have important roles in regulating geneexpression at the post-transcriptional level.26 Inthe bacterium E. coli genome-wide searches haveled to the identification of more than 70 sRNAsand many of these are known or believed to actby base-pairing to modify the translation and/orthe stability of target mRNAs.27,28 In contrast toclassical antisense RNAs present in accessoryelements, the chromosomally encoded riboregula-tors are trans-encoded; i.e. they are transcribedfrom loci that are distinct from the genes encodingtheir targets. Consequently, antisense/target RNAcomplementarity is incomplete, and regulationmust be achieved via short stretches of base-pairing. Such mode of binding allows the recog-

nition of multiple target RNAs as well as therecognition of a single target by multiple sRNAs.Antisense RNAs acting in this manner require acofactor, the RNA chaperone Hfq, which has beenshown to promote RNA–RNA annealing in vitro.Moreover, Hfq protects many of the sRNAsagainst endonuclease cleavage.29–31 The protein isconserved in a wide range of bacteria and isrelated to the Sm-like proteins in eukaryotes andArchaea.32 Another feature of the chromosomallyencoded antisense RNAs is that they are tightlyregulated at the transcriptional level, and fre-quently expressed as part of global regulatorynetworks which function in response to environ-mental stress signals. This suggests that many of

Figure 3. (a) Effects of rybB deletion on ompC andompW mRNA levels. Strains S∅928 (wild-type) andS∅928ΔrybB were grown in LB medium and samplesfor total RNA extraction were removed at the indicatedA600 values. 5 μg of the total RNA samples were separatedon urea −5% polyacrylamide gels and probed withuniformly labelled RNA probes for ompC and ompWmRNAs. (b) Effects of controlled RybB expression on thecellular level of the ompC and ompW mRNAs. Cells ofS∅928/pNDM220, S∅928ΔrybB/pNDM220 andS∅928ΔrybB harbouring pNDM-rybB (derivative ofpNDM220 that express rybB under the control of theIPTG-inducible PA1/O4-promoter) were grown in LB to anA600 of 1.2, and then induced with varying concentrationsof IPTG for 10 min. Subsequently, samples were removed,total RNAwas extracted and RybB sRNA as well as ompCand ompW mRNAs were analysed by Northern blot anal-ysis. (c) Effects of a short-term ectopic expression of rpoEon RybB sRNA and omp mRNAs. Isogenic rybB+ (S∅928),ΔrybB and Δhfq strains, transformed with pNDM-rpoE orthe corresponding empty vector (pNDM220), were grownto an A600 of 1.2; σ

E synthesis was induced with IPTG for10 min and RNA from equal amounts of cells wereelectrophoretically separated and probed as describedabove. (d) Predicted pairing of RybB RNA to the trans-lational initiation region of the ompC and ompWmessages.The start codon is indicated in bold. The predicted ompCregion overlaps with the target for the MicC RNA.35 TheΔrybB deletion strain was constructed as described for theΔmicA strain;17 the antibiotic resistance cassette wasamplified from plasmid pKD341 with the primers oRybBdel-1 (5′CACAACCGCAGAACTTTTCCGCAGGGCAT-CAGTCTTAATT gtgtaggctggagctgcttc) and oRybB del-2(5′-TGGTTGAGAGGGTTGCAGGGTAGTAGATAAGTTTTAGATAcatatgaatatcctccttag) (sequences from the E. coligenome are in capital letters). For construction of pNDM-rybB, the rybB gene was amplified from S∅928 by PCRusing the primers. AatII pA1O4 rybB (5′-GCCTGACGTC-GGCAAAAAGAGTGTTGACTTGTGAGCGGATAACAATGATACT TAGATT CGCCACTGCTTTTCTTTGATG)and BamHI 3′ rybB (5′-CCCCGGATC CGTTGAGAGGG-TTGCAGGGTA). The PCR product was cloned intoAatII/BamHI digested pNDM220. DNA templates forthe RNA probes used for detection of ompC and ompWmRNAs were generated by PCR using the primers:oompC F (5′-GCAAACGCTGCTGAAGTT TA) plusoompC R T7 (5′-GGGCCTAATACGACTCACTATAGGG-TGCCCTGGATCTGATATTCC) and oompW F (5′-AACC-TACGCCCAGTTTGTTG) plus oompW R T7 (5′-GGGCCTAATACGACTC ACTATAGGGGCGCAATAC-CAGTTCGATTT).

Figure 4. Schematic representation of σE-dependentsmall sRNAs involved in regulation of outer membraneproteins. The mode of action of RybB on its targets is stillunclear, but since Hfq is required for the control of OmpCand OmpWexpression, the sRNA is predicted to act by anantisense mechanism.

6 σE-dependent Small Regulatory RNAs

these riboregulators are involved in adaptiveresponses.24

An increasing number of Hfq-binding sRNAshave been implicated in down-regulation of thesynthesis of OMPs, suggesting that one importantcellular role of regulatory RNAs is to modulate andfine-tune the composition of the outer membrane.MicF, the first chromosome-encoded antisense RNAto be discovered in E. coli,33 is encoded divergentlyfrom the ompC gene and negatively regulates ompFexpression. A number of different external stressfactors stimulate micF transcription and four tran-scriptional activators are known to bind the micFpromoter. The ompC and micF genes shares regula-tory sites for the OmpR osmoregulator, and expres-sion of both genes is turned on at high osmolarity viathe EnvZ-OmpR two-component system. The relat-ed MarA, SoxS and Rob regulators bind to the samepromoter element and activate micF transcription inresponse to various stimuli, including weak acids,oxidative stress and certain antibiotics. Expressionof micF also is induced at high temperature and byother toxic compounds such as ethanol and Cu2+;the mechanism(s) by which this occurs is notknown.34 More recently, five additional chromo-somally encoded riboregulators of OMPs, have beenidentified, MicA, MicC, OmrA, OmrB and RseX. TheMicC RNA down-regulates ompC expression in anHfq-dependent manner and base-pair with theShine–Dalgarno region of the ompC mRNA. Inagreement with reciprocal expression of ompC andompF under many different environmental condi-tions, MicC exhibits an expression pattern that isopposite to MicF. However, the molecular basis ofmicC regulation still remains to be elucidated.35

OmrA and OmrB are encoded adjacent on thechromosome, possess sequence similarity at their 5′and 3′ ends, and are, like micF, under the control ofthe EnvZ/OmpR two-component system. Whenoverexpressed they negatively regulate at leastfour OMP genes, ompT, cirA, fecA and fepA. OmpTis a protease, whereas CirA, FecA and FepA areTonB-dependent receptors involved in iron trans-port.36 The RseX RNAwas identified as a multicopysuppressor that alleviated the requirement for theessential DegS and RseP proteases, which asdescribed are involved in releasing σE from theanti-sigma factor RseA. When overexpressed, RseXRNA down-regulates the expression of ompA andompC. Natural conditions of rseX induction havenot yet been found.37 Another fascinating exampleof sRNA regulation of porin expression involvesthe lysogenization of E. coli cells by the lamdoidphage, PA-2. Here, infection leads to the expressionof a phage-encoded porin as well as a phageregulatory RNA, IpeX, which inhibits ompC andompF expression.38

Yet another striking but less understood exampleof OMP regulation is provided by the σE regulatorysystem. Thus, σE not only up-regulates expression ofmembers of the σE regulon but also negativelyimpacts expression of several OMPs.11 Since the σE-inducing signal is misfolded or unfolded proteins

that possess OMP-like C termini (e.g. the porinsOmpC, OmpF, OmpW and OmpX)(2) the σE-mediated regulation of OMP expression may haveevolved in order to damp activation and/or to limitthe duration of the stress response. However, recentwork demonstrated that σE activity can be modu-lated by internal signals as well, suggesting that σE

has additional roles in the cell beyond monitoringextra-cytoplasmic stress.8 Thus, the signalling path-way that activates σE during entry into stationaryphase is dependent upon the nutritional stress signalppGpp and occurs in cells lacking RseA. Strikingly,also the activity of the alternative sigma factors σS

and σN increases upon entry into stationaryphase.39,40 These findings have let to the hypothesisthat the ppGpp-signalling pathway provides amechanism by which alternative sigma factors canbe activated in concert to provide a coordinate re-sponse to nutritional depletion.8 The mechanism(s)by which ppGpp modulates σE activity and the roleof this sigma factor in stationary phase are not fullyunderstood. However, expression of OMPs must betightly controlled under all growth conditions andthe ppGpp pathway may be important for balancingOMP levels during growth arrest. Furthermore, thispathway may be critical for proper remodelling ofthe cell surface during stationary phase. Thus, the E.coli cell undergoes remarkable developmentalchanges during stationary phase, including exten-sive changes in the envelope, which make stationaryphase cells resistant to a variety of environmentalinsults.In studies aimed at understanding the regulation

of MicA expression we realized that micA and rybBtranscription could be regulated by the alternativesigma factor, σE. The data presented in Figure 2(a)and (b) proved that the two sRNAs belong to the σE

7σE-dependent Small Regulatory RNAs

regulon and that transcription of the micA and rybBgenes seems to be completely dependent upon σE.Also genetic screens have resulted in the identifica-tion of σE as the critical regulator of rybB transcrip-tion (K. Thompson and S. Gottesman, personalcommunication). Furthermore, the data presented inFigure 2(c) revealed that σE-mediated regulation ofompA expression requires MicA RNA.The analysis of changes in global gene expression

profiles upon a short-term etopic expression of rybBidentified two candidate targets for the RybB RNA,ompC and ompW, whose expression also is down-regulated by σE. The two candidates were con-firmed to be negative targets of RybB (Figure 3).Thus, the cellular levels of ompC and ompW mRNAswere significantly elevated in a strain deleted forRybB (Figure 3(a)), while a transient induction ofrybB expression caused a decline in the twomessages (Figure 3(b)). Further analysis revealedthat σE-mediated control of ompC and ompWexpression is dependent on RybB and Hfq (Figure3(c)). It is noteworthy that a decrease in ompC andompW mRNA signals upon entry into stationaryphase is still observed in the absence of the RybBRNA (Figure 3(a)), whereas a ΔmicA strain failed toexhibit a significant ompA down-regulation instationary phase.16,17 We do not yet understandthe molecular nature of the additional regulation ofompC and ompW. However, the regulation isindependent of σE; moreover, it does not requireHfq (data not shown), suggesting that it is indepen-dent of small Hfq-binding regulatory RNAs.The exact mode of action of RybB on its targets has

not yet been defined. However, since Hfq binds toRybB and is required for its activity on OMPexpression, it is reasonable to assume that it actsby an antisense mechanism. Negatively acting Hfq-binding RNAs usually bind at or near the transla-tional initiation region of target RNAs. We thussearched for possible regions around the start site ofthe two mRNAs that might be targets for antisenseregulation. In each case putative pairing regionscould be found between the 5′ segment of RybB andthe ribosome binding site. Moreover, the loop of the3′ hairpin of RybB possesses complementarity to aregion located adjacent to the ompC Shine–Dalgarnosequence (Figure 3(d)). The sequences of thesesegments of RybB are conserved and also the regionimmediately upstream of the ompC gene is very wellconserved in E. coli, Salmonella, Klebsiella, Citrobacterand Enterobacter. Thus, RybB might act in a fashionsimilar to that described for MicA.The results presented here have added the σE

regulon to the growing list of stress-inducedregulatory circuits that include small regulatoryRNAs. Moreover, coupled with recent work on σE

regulated genes and σE activation,8,11 they haveprovided insight into a regulatory network thatdown-regulates OMP expression and thereby pre-vents a “flow” of OMPs to the envelope duringstress conditions. Such a down-regulation of OMPexpression can, as illustrated by ompA, ompC andompW, be accomplished by small trans-encoded

regulatory RNAs that are under positive control ofσE (Figure 4). As micF expression is stimulated byseveral environmental factors that also activate σE

(e.g. high temperature, ethanol and Cu2+) it ispossible that a similar strategy accounts for down-regulation of ompF expression upon σE activation.However,micF is not amember of theσE regulon andhow micF activation is brought about under thesestress conditions needs to be better understood.34

Acknowledgements

We are grateful to B.H. Kallipolitis for commentson the manuscript and D.H. Nies for bacterialstrains. This work was supported by grants fromthe Danish Natural Science Research Council andthe Danish National Research Foundation.

References

1. Las Penas, A., Connolly, L. & Gross, C. A. (1997). σE isan essential sigma factor in Escherichia coli. J. Bacteriol.179, 6862–6864.

2. Walsh, N. P., Alba, B. M., Bose, B., Gross, C. A. &Sauer, R. T. (2003). OMP peptide signals initiate theenvelope-stress response by activating DegS proteasevia relief of inhibition mediated by its PDZ domain.Cell, 113, 61–71.

3. Alba, B. M. & Gross, C. A. (2004). Regulation of theEscherichia coli σE-dependent envelope stress re-sponse. Mol. Microbiol. 52, 613–619.

4. Ades, S. E. (2004). Control of the alternative sigmafactor σE in Escherichia coli. Curr. Opin. Microbiol. 7,157–162.

5. Ruiz, N. & Silhavy, T. J. (2005). Sensing external stress:watchdogs of the Escherichia coli cell envelope. Curr.Opin. Microbiol. 8, 122–126.

6. Collinet, B., Yuzawa, H., Chen, T., Herrera, C. &Missiakas, D. (2000). RseB binding to the periplasmicdomain of RseAmodulates the RseA: σE interaction inthe cytoplasm and the availability of σE. RNA poly-merase. J. Biol. Chem. 275, 33898–33904.

7. Miticka, H., Rowley, G., Rezuchova, B., Homerova, D.,Humphreys, S., Farn, J. et al. (2003). Transcriptionalanalysis of the rpoE gene encoding extracytoplasmicstress response sigma factor σE in Salmonella entericaserovar Typhimurium. FEMS Microbiol. Letters, 226,307–314.

8. Costanzo, A. & Ades, S. E. (2006). Growth phase-dependent regulation of the extracytoplasmic stressfactor, σE, by guanosine 3′,5′-bispyrophosphate(ppGpp). J. Bacteriol. 188, 4627–4634.

9. Rezuchova, B., Miticka, H., Homerova, D., Roberts, M.& Kormanec, J. (2003). New members of the Escher-ichia coli σE regulon identified by a two-plasmidsystem. FEMS Microbiol. Letters, 225, 1–7.

10. Dartigalongue, C., Missiakas, D. & Raina, S. (2001).Characterization of the Escherichia coli σE regulon.J. Biol. Chem. 276, 20866–20875.

11. Rhodius, V.A., Suh,W.C.,Nonaka,G.,West, J. &Gross,C.A. (2006). Conserved andvariable functions of theσE

stress response in related genomes. PLoS Biol. 4, 43–59.12. Kabir, M., Yamashita, D., Koyama, S., Oshima, T.,

8 σE-dependent Small Regulatory RNAs

Kurokawa, K., Maeda, M. et al. (2005). Cell lysisdirected by σE in early stationary phase and effect ofinduction of the rpoE gene on global gene expressionin Escherichia coli. Microbiol. 151, 2721–2735.

13. Pilsl, H., Smajs, D. & Braun, V. (1999). Characterizationof colicin S4 and its receptor, OmpW, a minor proteinof the Escherichia coli outer membrane. J. Bacteriol. 181,3578–3581.

14. Hong, H., Patel, D. R., Tamm, L. K. & van den Berg, B.(2006). The outer membrane protein OmpW forms aneight-stranded β-barrel with a hydrophobic channel.J. Biol. Chem. 281, 7568–7577.

15. Argaman, L., Hershberg, R., Vogel, J., Bejerano, G.,Wagner, E. G., Margalit, H. & Altuvia, S. (2001). Novelsmall RNA-encoding genes in the intergenic regionsof Escherichia coli. Curr. Biol. 11, 941–950.

16. Udekwu, K. I., Darfeuille, F., Vogel, J., Reimegård, J.,Holmqvist, E. & Wagner, E. G. (2005). Hfq-dependentregulation of OmpA synthesis is mediated by anantisense RNA. Genes Dev. 19, 2355–2366.

17. Rasmussen, A. A., Eriksen, M., Gilany, K., Udesen, C.,Franch, T., Petersen, C. & Valentin-Hansen, P. (2005).Regulation of ompA mRNA stability: the role of asmall regulatory RNA in growth phase-dependentcontrol. Mol. Microbiol. 58, 1421–1429.

18. Miticka, H., Rezuchova, B., Homerova, D., Roberts,M. & Kormanec, J. (2004). Identification of nucleotidescritical for activity of the σE-dependent rpoEp3promoter in Salmonella enterica serovar Typhimurium.FEMS Microbiol. Letters, 238, 227–233.

19. Wassarman, K. M., Repoila, F., Rosenow, C., Storz, G.& Gottesman, S. (2001). Identification of novel smallRNAs using comparative genomics and microarrays.Genes Dev. 15, 1637–1651.

20. Vogel, J., Bartels, V., Tang, T. H., Churakov, G., Slagter-Jager, J. G., Huttenhofer, A. & Wagner, E. G. (2003).RNomics in Escherichia coli detects new sRNA speciesand indicates parallel transcriptional output in bacte-ria. Nucl. Acids Res. 31, 6435–6443.

21. Egler, M., Grosse, C., Grass, G. & Nies, D. H. (2005).Role of the extracytoplasmic function protein familysigma factor RpoE in metal resistance of Escherichiacoli. J. Bacteriol. 187, 2297–2307.

22. Lanzer, M. & Bujard, H. (1988). Promoters largelydetermine the efficiency of repressor action. Proc. NatlAcad. Sci. USA, 85, 8973–8977.

23. Gotfredsen, M. &Gerdes, K. (1998). The Escherichia colirelBE genes belong to a new toxin-antitoxin genefamily. Mol. Microbiol. 29, 1065–1076.

24. Gottesman, S. (2005). Micros for microbes: non-codingregulatory RNAs in bacteria. Trends Genet. 21, 399–404.

25. Storz, G., Opdyke, J. A. & Zhang, A. (2004).Controlling mRNA stability and translation withsmall, noncoding RNAs. Curr. Opin. Microbiol. 7,140–144.

26. Storz, G. (2002). An expanding universe of noncodingRNAs. Science, 296, 1260–1263.

27. Gottesman, S. (2004). The small RNA regulators of

Escherichia coli: roles and mechanisms. Annu. Rev.Microbiol. 58, 303–328.

28. Storz, G., Altuvia, S. & Wassarman, K. M. (2005). Anabundance of RNA regulators. Annu. Rev. Biochem. 74,199–217.

29. Zhang, A. X., Wassarman, K. M., Ortega, J., Steven,A. C. & Storz, G. (2002). The Sm-like Hfq proteinincreases OxyS RNA interaction with target mRNAs.Mol. Cell, 9, 11–22.

30. Møller, T., Franch, T., Højrup, P., Keene, D. R.,Bächinger, H. P., Brennan, R. G. & Valentin-Hansen,P. (2002). Hfq: a bacterial Sm-like protein thatmediates RNA-RNA interaction. Mol. Cell, 9, 23–30.

31. Valentin-Hansen, P., Eriksen, M. & Udesen, C. (2004).The bacterial Sm-like protein Hfq: a key player inRNA transactions. Mol. Microbiol. 51, 1525–1533.

32. Schumacher, M. A., Pearson, R. F., Møller, T., Valentin-Hansen, P. & Brennan, R. G. (2002). Structures of thepleiotropic translational regulator Hfq and an Hfq-RNA complex: a bacterial Sm-like protein. EMBO J.21, 3546–3556.

33. Mizuno, T., Chou, M. Y. & Inouye, M. (1984). A uniquemechanism regulating gene expression: translationalinhibition by a complementary RNA transcript(micRNA). Proc. Natl Acad. Sci. USA, 81, 1966–1970.

34. Delihas, N. & Forst, S. (2001). MicF: an antisense RNAgene involved in response of Escherichia coli to globalstress factors. J. Mol. Biol. 313, 1–12.

35. Chen, S., Zhang, A., Blyn, L. B. & Storz, G. (2004).MicC, a second small-RNA regulator of Omp proteinexpression in Escherichia coli. J. Bacteriol. 186,6689–6697.

36. Guillier, M. & Gottesman, S. (2006). Remodelling ofthe Escherichia coli outer membrane by two smallregulatory RNAs. Mol. Microbiol. 59, 231–247.

37. Douchin, V., Bohn, C. & Bouloc, P. (2006). Down-regulation of porins by a small RNA bypasses theessentiality of the regulated intramembrane proteol-ysis protease RseP in Escherichia coli. J. Biol. Chem. 281,12253–12259.

38. Castillo-Keller, M., Vuong, P. & Misra, R. (2006).Novel mechanism of Escherichia coli porin regulation.J. Bacteriol. 188, 576–586.

39. Gourse, R. L., Ross, W. & Rutherford, S. T. (2006).General pathway for turning on promoters tran-scribed by RNA polymerases containing alternativeσ factors. J. Bacteriol. 188, 4589–4591.

40. Magnusson, L. U., Farewell, A. & Nyström, T. (2005).ppGpp: a global regulator in Escherichia coli. TrendsMicrobiol. 13, 236–242.

41. Datsenko, K. A. & Wanner, B. L. (2000). One-stepinactivation of chromosomal genes in Escherichia coliK-12 using PCR products. Proc. Natl Acad. Sci. USA,97, 6640–6645.

42. Møller, T., Franch, T., Udesen, C., Gerdes, K. &Valentin-Hansen, P. (2002). Spot 42 RNA mediatesdiscoordinate expression of the E. coli galactoseoperon. Genes Dev. 16, 1696–1706.

Edited by I. B. Holland

(Received 31 May 2006; received in revised form 27 August 2006; accepted 1 September 2006)Available online 8 September 2006