01-Wick 89

Adv Biochem Engin/Biotechnol (2004) 89: 1– 45 DOI 10.1007/b93957

Molecular Components of Physiological Stress Responses in Escherichia coli Lukas M. Wick1, 2 · Thomas Egli1 1

2

Department of Environmental Microbiology and Molecular Ecotoxicology, Swiss Federal Institute for Environmental Science and Technology, PO Box 611, 8600 Dübendorf, Switzerland E-mail: [email protected] Present address: L.M. Wick, Microbial Evolution Laboratory, National Food Safety and Toxicology Center, Michigan State University, East Lansing MI, 48824, USA

Abstract In order to survive under and adapt to different conditions Escherichia coli has evolved elaborate systems that are able to sense and respond to environmental stimuli.Very often, different stresses act on a bacterium simultaneously and a variety of stresses have similar effects on cellular molecules and processes. Therefore, the various stress response systems have to interact (cross talk) with each other. A complex network of global regulatory systems with a multitude of molecular components ensures a coordinated and effective answer. Such regulatory components include DNA, mRNAs, sRNAs, proteins, such as DNA-and RNA binding proteins, alternative sigma factors and two-component systems, as well as small molecular weight molecules, as for example (p)ppGpp. These regulatory systems govern the expression of a plethora of further effectors that aim at maintaining stability of the cellular equilibrium under the various conditions. Using five of the most important stress response systems, we will discuss the roles and mechanisms of such regulatory and effector molecules in more detail. The heat shock response, controlled by the sigma factor s32, and the envelope stress response, controlled by the sigma factor sE and the Cpx two-component system, both result in an increased expression of chaperones and proteases in response to misfolded proteins. The cold shock response governs expression of RNA chaperones and ribosomal factors, ensuring accurate translation at low temperatures. The general stress response depends on the sigma factor sS, which controls the expression of more than 50 genes conferring resistance to many different stresses. The (p)ppGpp-dependent stringent response reduces the cellular protein synthesis capacity and controls further global responses upon nutritional downshift. A lot has been learned in recent years about the mechanisms of action of single components. However, the main challenge for the future is to achieve an understanding of the interactions of these components under different physiological conditions. Keywords Heat shock · Envelope stress · Cold shock · General stress response · Stringent

response

1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3

2

Molecular Components Involved in Stress Response Regulation

4

2.1 2.1.1 2.1.2 2.2 2.3

Nucleic Acids . . . . . . . . . . . DNA . . . . . . . . . . . . . . . . RNA . . . . . . . . . . . . . . . . Proteins . . . . . . . . . . . . . . Small Molecular Weight Effectors

. . . . .

4 4 4 5 7

3

The Heat Shock Response . . . . . . . . . . . . . . . . . . . . . .

7

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

© Springer-Verlag Berlin Heidelberg 2004

2

L.M. Wick · T. Egli

3.1 3.1.1 3.1.2 3.1.3 3.2 3.2.1 3.2.2 3.2.3

Regulation of the Heat Shock Response Transcriptional Regulation . . . . . . . Translational Regulation . . . . . . . . . Posttranslational Regulation . . . . . . . Protein Folding and Degradation Control Chaperones . . . . . . . . . . . . . . . . Proteases . . . . . . . . . . . . . . . . . . Posttranslational Quality Control . . . .

. . . .

. . . . . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

8 9 11 11 12 12 14 14

4

The Envelope Stress Response

4.1 4.2 4.3

The sE Response . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 The Cpx Response . . . . . . . . . . . . . . . . . . . . . . . . . . 17 The Bae Response . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

5

The Cold Shock Response . . . . . . . . . . . . . . . . . . . . . . 19

5.1 5.2 5.3 5.4 5.5

Cold Shock Induced Proteins . . . . . CspA – The Major Cold Shock Protein The CspA Family . . . . . . . . . . . . Sensing of Cold Shock . . . . . . . . . Changes in Membrane Composition .

6

The Stringent Response

6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3

Regulation of (p)ppGpp Synthesis and Decay RelA . . . . . . . . . . . . . . . . . . . . . . . SpoT . . . . . . . . . . . . . . . . . . . . . . . Effects and Mechanisms of (p)ppGpp . . . . . Effects of (p)ppGpp . . . . . . . . . . . . . . . Mechanisms of (p)ppGpp Regulation . . . . . Growth Rate Control by (p)ppGpp . . . . . .

7

The General Stress Response

. . . . . . . . . . . . . . . . . . . . 32

7.1 7.1.1 7.1.2 7.1.3 7.2 7.2.1 7.2.2 7.2.3

Regulation of sS . . . . . . . . Transcriptional Regulation . Translational Regulation . . . Posttranslational Regulation . Effects of sS . . . . . . . . . . Physiological Effects of sS . . sS-Dependent Promoters . . . Role of sS in Various Habitats

. . . . . . . .

8

Conclusions and Perspectives . . . . . . . . . . . . . . . . . . . . 37

9

References

. . . . . . . . . . . . . . . . . . . . 15

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

. . . . .

20 22 24 25 25

. . . . . . . . . . . . . . . . . . . . . . . 26

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

. . . . . . .

. . . . . . . .

26 27 28 30 30 31 31

32 32 34 35 35 35 36 36

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

Molecular Components of Physiological Stress Responses in Escherichia coli

3

List of Abbreviations AHL AI-2 cAMP CRP DnaK/DnaJ/GrpE GroEL/GroES Hsp LB NTP OMP ppGpp pppGpp (p)ppGpp RNAP sRNA

Acylated homoserine lactone Autoinducer 2 (furanosyl borate diester) Cyclic adenosine monophosphate cAMP receptor protein Chaperone system consisting of DnaK, DnaJ and GrpE Chaperone system consisting of GroEL and GroES Heat shock protein Luria-Bertani broth Nucleotide triphosphate Outer membrane protein Guanosine 3¢,5¢-bispyrophosphate Guanosine 3¢-diphosphate, 5¢-triphosphate pppGpp or ppGpp RNA polymerase Small untranslated RNA, also called noncoding RNA (ncRNA)

1 Introduction A bacterial cell has very limited abilities to choose and modify its environment actively. Therefore, it is of vital importance that it is equipped with mechanisms that allow it to respond rapidly and effectively to a variety of environmental changes that can threaten the cell’s integrity. Cellular homeostasis is achieved by a multitude of sensors and effectors, which are able to sense and respond to changes in temperature, pH, oxygen concentration, nutrient availability, osmolarity etc. Normally, a cell will be confronted with different stresses simultaneously and various stresses might affect the same cellular components. Therefore, the stress response systems have to communicate with each other in order to find an efficient answer to the stresses they are exposed to. The focus of this chapter will be on the basic regulation and mechanisms of stress responses and not their practical application. We will first discuss some general aspects of the different classes of molecules which serve as global stress sensors and regulators. Then we will review more specifically stress responses that are involved in two common stresses, namely in temperature (heat shock, envelope stress, cold shock) and in nutrient availability (general stress response, stringent response). Because these stress responses have been most intensively studied in Escherichia coli, the focus will be on this bacterium. However, most of the principles and molecular components discussed here are universal. In the bacterial world, two-component systems constitute the major way of signal transduction [1, 2].Also the use of different sigma factors in controlling gene expression [3] and the role of (p)ppGpp in stress responses is widespread [4]. The heat shock proteins and the major cold shock protein CspA are not only prevalent and well conserved throughout the prokaryotic but also the eukaryotic world [5–7].

4


2 Molecular Components Involved in Stress Response Regulation The major different classes of cellular molecules involved in the regulation of stress responses are nucleic acids, proteins, and small effector molecules (Table 1). 2.1 Nucleic Acids 2.1.1 DNA

The DNA topology influences gene expression and it can function both as a sensor and as a regulator. The degree of DNA supercoiling varies in response to environmental stresses such as starvation, transition from aerobiosis to anaerobiosis and shifts in temperature, pH, and osmolarity [8]. Upon stress the DNA conformation might change directly because of the altered physico-chemical conditions, or indirectly as a consequence of the modified activity of proteins (see Heat shock and Cold shock). Proteins involved in the control of DNA topology are DNA topoisomerases, such as gyrase and topoisomerase I [8, 9], and the histone-like DNA-binding proteins, the major four ones being HU, IHF, Fis, and H-NS [10, 11]. H-NS generally acts as a repressor and plays a role in a variety of stresses [12]. For example, during cold shock H-NS expression is essential for effective adaptation [13] whereas hns deletion mutants show increased cellular concentrations of sS and a better ability to survive low pH and high osmolarity than the wildtype [14]. 2.1.2 RNA

RNAs can also act as stress sensors and regulators. A change in the secondary structure of mRNAs in response to an altered environment can vary their stability or the efficiency of their translation, as seen for example for the rpoH, cspA, and rpoS RNA (see Heat shock, Cold shock, and General stress response, respectively). An interesting class of regulatory molecules that gained a lot of interest in recent years are the small untranslated RNAs, the sRNAs. The genome of E. coli might encode up to 50 sRNAs [15]. Several of them bind to complementary sequences in target mRNAs, and this binding can inhibit or activate translation of these mRNA. Bacterial antisense sRNAs are generally able to recognise several different target mRNAs. Not all sRNAs are antisense RNAs, some act as modulators of the activity of target proteins. One hypothesis is that sRNAs evolved as regulators not only because their production is economical but also because they are very versatile [16–19]. Examples of sRNAs are DsrA, RprA, and OxyS (see General stress response).


5

2.2 Proteins

In addition to the histone-like proteins, the major two protein families involved in transcriptional regulation of stress responses are alternative sigma factors and two-component regulatory systems. Escherichia coli has seven sigma factors: the household sigma factor s70, the heat shock sigma factor s32 (see Heat shock), the envelope stress sigma factor sE (see Envelope stress), the general stress response sigma factor sS (see General stress response), s28 (sF) involved in regulation of flagella and chemotaxis genes [20], s54 (sN) involved mainly in transcription of genes involved in nitrogen assimilation [21], and sFecI, which controls genes for uptake of ferric citrate [22] (Table 1). The activity of sigma factors can often be inhibited by their interaction with anti-sigma factors [23]. Examples of sigma factor anti-sigma factor pairs are: s32-DnaK, sE-RseA, and sS-RssB (see Heat shock, Envelope shock, and General stress response, respectively). Two-component systems consist of a sensor-transmitter and a response regulator protein. In response to environmental signals the typically membranebound sensor-transmitter autophosporylates at a conserved histidine residue. This phosphate group is subsequently transferred to an aspartate residue of the response regulator, which, in its phosphorylated form, generally acts as a transcription activator [1, 2]. Escherichia coli has about 30 of these systems [24], only some of them will be mentioned here. The KdpD/KdpE [25] and EnvZ/OmpR [1, 26] systems are involved in osmotic regulation, the NtrB/NtrC [27] and PhoR/PhoB [28] systems respond to nitrogen and phosphate limitation, respectively, and the CreC/CreB [28] system controls gene expression dependent on variations in carbon/energy sources. Other sensor-transmitter pairs are ArcB/ArcA [29, 30], which act as a repressor system of aerobic respiratory pathways in response to O2 deprivation, whereas the SoxR/SoxS [31, 32] system is activated upon oxidative stress. The CpxA/CpxR system will be discussed in more detail below (see Envelope stress). Besides the SoxR/SoxS two-component system, E. coli has a second regulator, OxyR, that plays a role in response to oxidative stress. OxyR combines both sensor and response regulator functions in one molecule. Oxidation of two cysteine residues in the OxyR molecule leads to the formation of a disulfide bond, which concomitantly results in transcriptional activity of OxyR [31, 32]. The DNA-binding protein Lrp is a global transcriptional regulator protein whose cellular role is still unclear. Since it mostly represses genes required for uptake and metabolisms of nutrients in rich medium and activates expression of several biosynthetic genes, a role in coordinating cellular metabolism with the nutritional state of the environment was suggested [33, 34]. Recent gene expression profiling experiments showed that Lrp represses the expression of many stationary phase induced proteins and several global regulators, including H-NS, HU, IHF, sS, and sE. Therefore, it still remains to clarify which genes are controlled directly by Lrp and which are controlled indirectly by one or more of these global regulators [35, 36]. The situation gets further complicated since it has been

6


Table 1 Major global regulatory proteins discussed in this review

Protein

Alternative names

Gene (alternative gene names) a

Role

CpxA

–

cpxA (ecfB, eup, ssd, rssE)

Envelope stress (sensor kinase)

CpxP

–

cpxP (yiiO)

Negative regulation of CpxA

CpxR

–

cpxR (yiiA)

Envelope stress (response regulator)

CspA

–

cspA

Major cold shock protein

DnaK

Hsp70

dnaK (grpC, grpF, seg, groPAB, groPC, groPF)

Heat shock, chaperone; also involved in regulation of sS

Fis

–

fis (nbp)

Histon-like protein

Hfq

HF-I

hfq

Interaction with DNA and RNA, especially sRNAs

H-NS

Protein H1, protein B1

hns (bglY, cur, drc, drdX, drs, irk, msyA, osmZ, pilG, topS, hnsA, topX, fimG, virR, H1, B1)

Histon-like protein

HU

–

HU is a heterodimer of HU-alpha (encoded by hupA) and HU-beta (encoded by hupB (hopD, dpeA))

Histon-like protein

IHF

–

IHF is a heterodimer of IhfA (encoded by ihfA (himA, hid)) and IhfB (encoded by ihfB (himD, hip))

Histon-like protein

Lrp

Methylation- Lrp (alsB, livR, lss, lstR, mbf, blocking oppI, ihb, lrs, rblA) factor

Response to nutrient availability

RelA

–

relA (RC)

Synthesis of (p)ppGpp in stringent response

RseA

–

rseA (yfiJ, mclA)

Anti-sigma factor of sE

RssB

–

rssB (sprE, mviA, hnr, ychL)

Anti-sigma factor of sS

SpoT

–

spoT

Synthesis and degradation of (p)ppGpp in stringent response

s32

sH, RpoH

rpoH (fam, hin, htpR)

Heat shock sigma factor

s70

sD, RpoD

rpoD (alt)

Housekeeping sigma factor

sE

s24, RpoE

rpoE (sigE)

Envelope stress sigma factor

sS

s38, RpoS

rpoS (appR, csi2, katF, nur, otsX, sigS, abrD, dpeB)

General stress response sigma factor

a

From [349].


7

reported that H-NS on its part represses Lrp [14], and that Lrp and sS activity are both up-regulated by (p)ppGpp [37–39]. Besides the global transcriptional regulators, there are also several regulatory proteins that act at the posttranscriptional level. Such a function has been shown for the RNA-binding proteins Hfq, CsrA, CspA, CspC, and CspE (for the latter three proteins see Cold shock). The global regulatory role of Hfq is largely due to its regulation of rpoS translation, but the expression of several other proteins are controlled by Hfq in a sS-independent manner [40]. Hfq interacts with most of the sRNAs [15]. CsrA is involved in the control of glycogen metabolism, glycolysis, acetate metabolism, motility, adherence, cell morphology and other functions [41]. Proteases are a class of proteins important in the posttranscriptional control of global regulators (see Heat shock, Envelope shock, and General stress response). 2.3 Small Molecular Weight Effectors

There are three major classes of small effector molecules that are involved in the regulation of multiple genes. These include cAMP, which together with its acceptor protein (CRP) controls the expression of many catabolic genes [42, 43], (p)ppGpp (see Stringent response), and autoinducers. Autoinducers mediate cell density-dependent regulation of gene expression, a phenomenon called quorum sensing. Two common types of autoinducers are the acylated homoserine lactones (AHL), which are generally specific for particular species of bacteria, and a furanosyl borate diester (AI-2 for autoinducer 2), which seems to be a universal signal for interspecies communication [44, 45]. So far, neither AHL-synthesizing activity nor genes homologous to known AHL synthetase genes have been found in any E. coli strain [46]. However, interestingly, the E. coli genome encodes a gene (sdiA) homologous to luxR-like genes, whose gene products generally are receptors for AHL and act as transcriptional activators after AHL-binding [45, 46]. SdiA is involved in the control of cell division and multidrug resistance ([47] and references therein). Escherichia coli produces the other common autoinducer AI-2 in a reaction that is dependent on the product of the luxS gene [48]. AI-2-mediated quorum sensing controls genes involved in flagella assembly, motility, chemotaxis, and virulence ([49] and references therein). However, as Winzer et al. [50] have pointed out, quorum sensing has not always been clearly proven in several studies that claim to have found quorum sensing effects.Artefacts due to released metabolites and toxic compounds must be taken into consideration as well [50].

3 The Heat Shock Response The heat shock response was first observed as a response to elevated temperatures, hence its name. However, this physiological response is a cell’s answer not only to elevated temperature but also to several other adverse conditions such as

8


exposure to certain chemicals (solvents, certain antibiotics), hyperosmotic shock, and as response to overproduction of foreign and own proteins. Such challenges frequently lead to misfolded, unfolded, or damaged proteins in the cell. These proteins may expose at the surface hydrophobic patches that are normally buried inside the folded protein. These exposed hydrophobic residues can result in aggregation of proteins, an event that constitutes a serious threat to the organization and functioning of all cytoplasmatic components in any cell. To prevent this the heat shock response is triggered. Characteristic for the heat shock response is the increased expression of chaperones and ATP-dependent proteases. These proteins are involved in protein folding and degradation under both stressed and unstressed conditions. The regulation of these genes is dependent on the sigma factor s32. A second heat shock sigma factor is sE, which is activated at higher temperatures. The latter sigma factor regulates protein folding and turnover in the envelope. Due to this compartmental difference of the two heat shock sigma factors, stresses eliciting increased transcription by sE are summarised as envelope stress and, therefore, will be treated separately from the s32 heat shock response (see below). Members of the heat shock regulon are not only important under stress conditions; their role in controlling the folding and degradation of proteins is of vital importance under all growth conditions. A failure of these protective mechanisms can lead to protein aggregates within a cell that can be directly visualised as the formation of inclusion bodies. The presence of heat shock proteins in prokaryotes as well as in eukaryotes and the high degree of conservation of these homologous proteins in all organisms further underscores their crucial importance for cellular survival. 3.1 Regulation of the Heat Shock Response

The heat shock response was first described in Drosophila melanogaster as a stress response to heat [51]. In E. coli the heat shock response is mediated by the alternative sigma factor s32. Increase in s32 is not only triggered by heat but also by ethanol, certain antibiotics (puromycin, nalidix acid), viral infection, methylating and alkylating agents, cadmium chloride, hydrogen peroxide, amino acid restriction [52] and carbon starvation [53–55]. More than 30 heat shock proteins are controlled by s32, most of them being chaperones (DnaK, DnaJ, GrpE, ClpB, GroEL, GroES, HtpG, IbpA, IbpB) and proteases (Lon, ClpAP, ClpXP, HslUV (=ClpYQ), FtsH) (see [56] for a recent list of heat-inducible proteins). At a temperature of 30 °C there are less than 50 molecules of s32 per cell compared to about 3000 molecules per cell of the household sigma factor s70. Upon a heat shock from 30 °C to 42 °C the level of s32 transiently increases about 17-fold within 5–6 min and after 15 min reaches a new steady state, which lies about five times higher than the pre-shift level [57]. Both the composition of the growth medium and growth conditions have an effect on the expression of heat shock proteins [58, 59]. Herendeen et al. [60] determined protein levels by two-dimensional gel electrophoresis in cells grown at seven temperatures between 13.5 °C and 46 °C in a glucose medium supple-


9

mented with amino acids, vitamins, adenine, guanine, cytosine, and thymine. In exponentially growing cells the levels of the heat shock proteins GroEL (B56.5), DnaK (B66), and HtpG (C62.5) were fairly constant between 13.5 °C and 30 °C. Between 30 °C and 46 °C a tenfold, a fourfold, and a fivefold increase in GroEL, DnaK, and HtpG levels, respectively, was found [60]. A similar result was found for cells growing on LB where expression of the htpG::lacZ reporter at 45 °C was twofold higher than at 37 °C. However, in batch cultures on glucose or pyruvate minimal media, no increase in htpG::lacZ expression was reported, neither during a transient heat shock nor between cells growing at 45 °C compared to cells growing at 37 °C [59]. In continuous cultures, on the other hand, changes in htpG::lacZ expression were comparable to the ones observed by Herendeen et al. [60], both in glucose-limited as well as in complex medium. It was also found that steady state levels of htpG::lacZ expression decreased by about 20% with an increase in dilution rate from 0.23 h–1 to 0.63 h–1 [58]. Similar to the regulation of the general stress response sigma factor sS, regulation of the heat shock response sigma factor s32 occurs also at the transcriptional, translational, and the protein stability/activity level (Fig. 1). 3.1.1 Transcriptional Regulation

Transcription of the heat shock sigma factor gene rpoH is controlled by at least four promoters, (P1, P3, P4, P5) [61, 62]; three of them are transcribed by RNAP s70, whereas sE is responsible for transcription from promoter P3 at very high temperatures (45–50 °C) when s70 becomes inactivated by the elevated temperature [63, 64]. Other inducers of s32 transcription are ethanol and DNA gyrase inhibitors [61, 62, 65]. DNA gyrase inhibitors such as nalidixic acid, oxolinic acid and novobiocin reduce the negative supercoiling activity of DNA gyrase and thus lead to relaxation of DNA. Induction of the heat shock response by DNA gyrase inhibitors supports the idea that DNA topology is involved in sensing environmental stresses and is also affecting transcription of genes during the stress response. The model described by Lopez-Garcia and Forterre [66] supposes that a heat shock affects the activities of topoisomerases in a way that leads to relaxation of the DNA and that this change in activity is enhanced by the thermally induced change in the DNA geometry. DNA relaxation, triggered either by DNA gyrase inhibitors or heat shock, then favours transcription of stress response genes, such as rpoH (Fig. 1). This model is based on the observations that rpoH promoter activity (P1 promoter) increases threefold upon novobiocin treatment [65], that amounts of s32 in cells increase after addition of nalidixic acid, oxolinic acid and novobiocin [67] and that in gyrA mutants the heat shock response is not transient anymore but continuous [68]. Moreover, as it is typical for most shock responses, relaxation of DNA upon heat shock is transient and the reversion to the normal state is mediated by a negative feedback loop through the interaction of the heat shock protein DnaK with DNA gyrase [69] (Fig. 1).

10



11

3.1.2 Translational Regulation

Control at the translational level seems to be dependent on the ambient temperature only, and no dependence of the cellular levels of proteases, chaperones, and misfolded proteins, which play a role in the control of s32 stability (see below), has been observed so far. In fact, the s32 mRNA acts as a cellular thermometer, and hence, this represents a first level of stress sensing.At lower temperatures the rpoH mRNA folds into a secondary structure, in which the ribosome-binding site and the initiation codon are involved, and this impedes efficient translation.At higher temperatures this secondary structure is unfolded leading to ribosome binding and increased s32 synthesis ([70] and references therein) (Fig. 1). Experiments with wt and mutated s32 mRNA revealed a clear correlation between temperature melting profiles of the mRNA secondary structure, formation of the mRNA-30S ribosome-initiator tRNA complex and expression of rpoH-lacZ translational fusions [70]. 3.1.3 Posttranslational Regulation

On the protein stability level, the heat shock response is regulated by DnaK/DnaJ/GrpE over the amount of damaged or misfolded proteins in the cell [71]. DnaK and its co-chaperones DnaJ and GrpE do not only recognise and bind misfolded proteins (see below) but also its own regulator s32. Thus DnaK/DnaJ/GrpE competes with RNAP in binding of s32.Whereas binding to the chaperone leads to degradation of s32, the binding to RNAP results in a stabilization of s32 and consequently to transcription activation of heat shock protein genes. Under normal growth conditions most of s32 is sequestered away from RNAP by free DnaK/DnaJ/GrpE, and shows a short half-life of less than 1 min (Fig. 1).After a temperature upshift or other stresses that result in misfolded proteins, these damaged proteins compete in binding to DnaK/DnaJ/GrpE with s32,

Fig. 1 The heat shock response. Under normal conditions little s32 is translated and cellular s32

levels are low. The translated s32 is bound mainly to the chaperone complex DnaK/DnaJ/GrpE and degraded by FtsH (situation top left). A heat shock (1.) leads to denaturation of proteins and binding of these misfolded proteins to the chaperone complex DnaK/DnaJ/GrpE. This frees s32 from DnaK/DnaJ/GrpE, thereby stabilizing s32 and allowing s32 to bind to RNAP, which results in the transcription of heat shock genes.A further increase in s32 levels after heat shock is due to a decrease in DNA supercoiling and melting of the rpoH mRNA secondary structure (1.). The induced heat shock proteins are mainly proteases and chaperones, which degrade or refold the misfolded proteins. This leads to a decrease in the concentration of misfolded proteins and concomitantly to an increase in free DnaK/DnaJ/GrpE (2.). As a result, free DnaK/DnaJ/GrpE then again binds to s32, leading to its inactivation and shut down of the heat shock response (3). DnaK/DnaJ/GrpE has a central regulatory role due to its interactions with s32, misfolded proteins, and DNA gyrase. The interaction with DNA gyrase is necessary for the re-supercoiling of DNA and the shut down of the heat shock response. P: protein; Hsp: heat shock proteins (chaperones and proteases (scissors)), the major protease involved in s32 degradation being FtsH; RNAP: RNA polymerase

12


leading to a transient stabilization of s32 and increased expression of heat shock proteins (Fig. 1). Due to the higher level of heat shock proteins, free DnaK/ DnaJ/GrpE destabilises s32 again and in turn shuts off the heat shock response [72, 73] (Fig. 1). The negative feedback control enables a fast regulation of the heat shock response. Originally, it was thought that the mere sequestration from RNAP of s32 by DnaK/DnaJ/GrpE keeps the s32 in a state sensitive to proteolysis [72, 73]. However, recent experiments have shown that s32 mutants defective in RNAP binding still need the DnaK/DnaJ/GrpE chaperone system for rapid degradation. This suggests a more active role of DnaK/DnaJ/GrpE in the s32 degradation [74]. In contrast to the increase in the rate of translation of rpoH mRNA, the stability of the s32 protein itself decreases with increasing temperature [75]. The increased translation, but at the same time also increased degradation of s32 at higher temperatures and the negative feedback control of s32 by its own transcription products allow a fast control and a fine tuning of the heat shock response. It appears that the membrane bound protease FtsH is primarily responsible for s32 degradation [76–78]. However, there are strong indications that the cytoplasmic proteases ClpQY (HslVU), ClpP and Lon are also involved in s32 proteolysis [79]. All these proteases are themselves regulated by s32, thus constituting a negative feedback loop. Recently, a novel sE-dependent membrane bound protease, EcfE, has been proposed to degrade s32 and also sE [80]. However, this activity has been questioned by Kanehara et al. [81], who showed that EcfE is involved in degradation of the sE anti-sigma factor RseA and, therefore, sE activation, rather than the earlier proposed inactivation (see also below, Envelope Shock). 3.2 Protein Folding and Degradation Control 3.2.1 Chaperones

Chaperones and proteases exert vital functions in all cells by ensuring correct folding and functionality of proteins. Nascent, newly synthesised and misfolded proteins may exhibit surface-exposed hydrophobic patches that can act as binding sites for chaperones [82, 83]. It is estimated that about 10–20% of newly synthesised proteins need the help of chaperones for correct folding [84, 85]. The danger of an aggregation is enhanced by molecular crowding, i.e., the high density of proteins and other molecules in the cytosol, which is estimated to be about 300 to 400 g/l [86, 87]. The chaperones thus prevent aggregation and catalyse protein folding. Proteins recalcitrant to proper folding are degraded by proteases. If both these mechanisms fail protein aggregates accumulate in the cell. The chaperones of E. coli are members of protein families that are highly conserved and named after their molecular weight. These families range from the small heat shock protein family (including IbpA and IbpB), over proteins with a molecular weight of ca. 10 kDa (Hsp10 with GroES), 40 kDa (Hsp40 with DnaJ), 60 kDa


13

(Hsp60 including GroEL), 70 kDa (Hsp70 with DnaK), 90 kDa (Hsp90 with HtpG), to 100 kDa (Hsp100 including ClpB) [7, 82, 83, 88]. The different chaperones vary in their function and act at different time points of the lifetime of proteins. During the synthesis of new proteins, it is necessary to prevent the building up of erroneous tertiary structures or aggregation with other proteins before the whole primary sequence of the domains has been achieved and these domains can fold properly. Nascent proteins interact with trigger factor (TF), which is – in contrast to all other major cytosolic chaperones – not controlled by s32. DnaK both interacts with nascent proteins and is involved in post-translational folding of newly synthesised and misfolded proteins [5, 89]. TF has chaperone and peptidyl-prolyl-cis/trans isomerase functions and it is associated with ribosomes in a 1:1 stoichiometry [90, 91]. Since it is situated near the exit channel for the polypeptides on the large ribosome unit, it probably interacts with all nascent proteins, and this much more efficiently than DnaK [92]. Most small proteins fold properly after synthesis and interaction with TF, whereas proteins larger than 20–30 kDa subsequently also need DnaK for folding [85]. TF and DnaK have overlapping functions in interacting with proteins co-translationally. Cells devoid of either TF or DnaK are viable at 37 °C whereas double mutants are not [85, 89]. DnaK acts in concert with DnaJ and GrpE. DnaK exhibits ATPase activity, a GrpE-binding site at the N-terminal part, as well as a DnaJ plus polypeptidebinding site at the C-terminal part. Binding of ATP to DnaK results in release of bound protein substrate (low affinity ATP bound state), binding of a new protein then stimulates ATPase activity, which stabilises the ADP/DnaK/DnaJ/GrpE/protein complex (high affinity ADP bound state). The exchange of ADP with ATP then again releases the protein [82, 93]. The high affinity state is achieved through catalysis by DnaJ. In fact, DnaJ seems to scan for hydrophobic patches in substrates to which it then binds. Subsequently, it targets DnaK to this hydrophobic binding site or sites nearby and mediates the high affinity binding state through ATP hydrolysis [94]. GrpE catalyses the exchange of ADP/ATP and the substrate-releasing step. Interestingly, GrpE undergoes reversible conformational changes upon increase in temperature. This change in conformation leads to a decreasing catalytic activity above 40 °C, thus stabilising the high affinity state of DnaK [95]. The catalytic mechanism of the GroEL/GroES chaperone complex is different from the one of the DnaK/DnaJ/GrpE chaperone system, since GroEL/GroES provide a cavity in which proteins can fold properly. This cavity is achieved by the stacking of two heptameric rings consisting of identical subunits. One end of such a ring is occupied by GroES forming a cage in which proteins up to about 60 kDa can be enclosed and have the chance to fold in this protected environment. However, the GroEL/GroES chaperone system also promotes folding of some proteins that are too large to fit into the cavity provided by GroEL/GroES. Two models have been proposed for the GroEL/GroES action, the Anfinsen cage (or confinement) model and the iterative annealing model (see [7, 82, 83, 96] for review). Besides helping the folding of newly synthesised and refolding of misfolded proteins, several experiments demonstrated that chaperones are able to solubilise

14


and refold stable protein aggregates. This is achieved as a sequential action of the chaperones ClpB and DnaK/DnaJ/GrpE [97, 98]. In the proposed mechanism ClpB first binds to protein aggregates, an interaction of ATP with ClpB then leads to a structural change in ClpB, which in turn results in an increased exposure of hydrophobic regions in the aggregated protein. This allows DnaK/DnaJ/GrpE to bind and catalyse dissociation of the aggregates and refolding of the proteins into native solubilised proteins [98]. Besides these well-studied chaperones there are further chaperones whose function is still largely unknown.Among them are IbpA and IbpB, both of which belong to the family of the small heat shock proteins, which are found in a variety of organisms. The small heat shock proteins are supposed to be holdases, i.e. ATP-independent chaperones that bind to hydrophobic surfaces of proteins in order to prevent aggregation. Therefore, they are able to build up a reservoir of nonnative proteins for an extended period of time. Later, when conditions allow, these proteins can be passed on to the ATP-dependent DnaK/DnaJ/GrpE and GroEL/GroES chaperones for proper folding [99]. Experimental evidence for such a function of IbpA and IbpB in E. coli has been shown in several studies [100–103]. Also the E. coli HtpG protein, which is homologous to the eukaryotic Hsp90 family, belongs to the ATP-independent holdases [104]. A chaperone with an unusual property is Hsp33. Its activity is regulated posttranslationally by the redox conditions of the environment. In its inactive reduced state, Hsp33 has zinc bound to four conserved cysteine residues. Under oxidising conditions, the zinc is released, the cysteine residues form disulfide bridges, and the chaperone becomes activated [105]. 3.2.2 Proteases

Several of the ATP-dependent proteases are under s32 control, namely Lon, ClpP, HslV (ClpQ), FtsH [56]. Some of these proteases function as two-component ATPdependent proteases together with regulatory subunits that show chaperone activity. ClpP associates either with ClpA or ClpX. ClpA and ClpX are supposed to recognise, unfold and present protein substrates to the protease ClpP. Lon and FtsH are homo-oligomers, but their polypeptide chains fold into domains with different functions, one with chaperone and one with proteolytic activity (reviewed in [106]). 3.2.3 Posttranslational Quality Control

Chaperones and proteases work together in the posttranslational quality control. Gottesman et al. [107] have proposed a triage model for damaged proteins in prokaryotes. In this model, proteins with exposed hydrophobic residues are recognised either by a chaperone or a regulatory component of a protease. If the protein binds to the protease it is degraded (although sometimes a protein unfolded by the chaperone component of the protease (see previous paragraph) may be released without degradation [108]). If it is bound to the chaperone it


15

might be converted to its native conformation. A protein that is released from a chaperone without having folded into its native state then undergoes another cycle of either chaperone- or protease-binding until it finally folds correctly or is degraded. This triage pathway functions stochastically and the fate of a protein (folding or degradation) depends on the kinetics of its interaction (binding and release) with the chaperones and proteases [106, 107]. In an extensive study Bukau and co-workers [109] dissected the roles of different chaperones and proteases in the quality control in the E. coli cytosol.A central role was found for DnaK and the DnaK-ClpB bi-chaperone network in preventing and reversing protein aggregation and also in providing thermotolerance at 50 °C. Lon was found to be the most efficient protease; Lon and ClpXP were essential at 42 °C in cells with low DnaK levels [109]. An interesting link between starvation, oxidative stress, and posttranslational quality control has been suggested by Nyström and co-workers [110–114]. They found that during starvation the levels of oxidised proteins increase and that this is probably one reason for heat shock induction in stationary phase. Interestingly, some proteins are more susceptible to stasis-induced oxidation, among them DnaK and H-NS. This suggests that oxidation of specific proteins serves as a regulatory means [113, 114]. Moreover, they showed that starved cells exhibit decreased translational fidelity, which may be another cause for induction of heat shock proteins, and that mistranslated proteins are primarily prone to oxidation. They further suggested that this increased unrepairable oxidation (by carbonylation) of mistranslated proteins might serve as a system to tag proteins and shunt them through the degradation rather than the refolding pathway [111, 112]. In biotechnology an often-encountered problem is the aggregation of heterologously over-expressed proteins. In some cases the co-expression of chaperones can alleviate formation of protein aggregates. The recently discovered capability of the DnaK-ClpB bi-chaperone network to solubilise already formed protein aggregates might lead to further strategies in the use of chaperone systems during synthesis of heterologous proteins [115].

4 The Envelope Stress Response Environmental conditions disturbing the extracytoplasmic space of gram-negative bacteria (i.e. the envelope consisting of inner membrane, periplasm, and outer membrane), elicit the so-called envelope stress response. There are at least two envelope stress response systems in E. coli. One is dependent on the sigma factor sE, whereas the other response is regulated by the Cpx two-component system. The two stress response pathways are distinct, both with respect to the stresses they sense and the response they trigger. Nevertheless, they also show some overlapping elements. Generally, activation of the sE and Cpx systems lead to increased synthesis of extracytoplasmic proteases, foldases and chaperones, in order to restore the functionality of the membrane. The sE system uses a membrane bound anti-sigma factor (RseA) to transmit inducing extracytoplasmic signals to the sE factor in the cytoplasm and induce gene activation by sE-RNAP. In the Cpx system the cytoplasmic membrane-bound histidine kinase CpxA ac-

16


tivates gene transcription through phosphorylation of the cytoplasmic response regulator CpxR upon envelope stress. Regulation of both pathways is achieved by the same mechanistic principles, namely autoactivation and feedback inhibition. Recently, a third envelope stress system has been found, which is also controlled by a two-component system, the sensor kinase BaeS and the response regulator BaeR. 4.1 The s E Response

The major components known so far to be involved in the signal sensing and transduction of the sE response are the alternative sigma factor sE, its anti-sigma factor RseA and the additional regulatory proteins RseB and RseC (see below for details). sE of E. coli belongs to the subfamily of ECF (extracytoplasmic function) sigma factors found in a variety of bacteria [116, 117]. The first promoters identified to be recognised by sE-RNAP are one of the promoters (promoter 3) of the rpoH gene, encoding the cytoplasmic heat shock sigma factor, and the degP (htrA) promoter [64]. DegP is a periplasmic protease with a broad substrate specificity [64, 118]. Interestingly, it can function both as a chaperone or a protease in a temperature-dependent manner [119]. Generally, target genes of sE-RNAP play a role in polypeptide degradation and folding in the envelope, including the protease/chaperone DegP [64], the chaperones Skp, SurA [120], the disulfide bond oxidoreductase DsbC [120], and the peptidyl-prolyl-isomerases FkpA [121] and SurA [120]. But also proteins involved in lipopolysaccharide metabolism, regulatory and sensory proteins are activated by sE (see [122] for review and also [120]). In a genetic screen Dartigalongue et al. [120] were able to characterise probably most of the genes regulated by sE-RNAP. From their and previous studies they concluded that some 43 genes of about 20 operons are under sE control [120]. sE is essential at all growth temperatures, which shows its importance in maintaining a functional envelope also under non-stress conditions [123]. The sE response is triggered by heat and ethanol [64, 124, 125], cold shock [120], hyperosmotic shock [126] and entry into stationary phase [127]. Furthermore, overexpression of outer membrane proteins and expression of outer membrane proteins containing mutations that cause protein misfolding lead to an up to sixfold increased activity of sE as assayed by quantifying expression of reporter genes under control of sE-dependent promoters (degP, rpoHP3, rpoE promoters) [128–130]. In contrast, mutants exhibiting reduced levels of outer membrane proteins show up to fivefold lowered sE activity [129]. Similarly, deletion of surA, which codes for a periplasmic peptidyl-prolyl-isomerase, induces the sE response five- to sevenfold, whereas multiple copies of the folding factor genes surA, fkpA, and skp are able to suppress an activation of sE in the case of envelope stress[130, 131]. sE activates the transcription of its own gene [124, 125] together with the regulatory genes rseA, rseB, rseC, which are all located in the same operon [132, 133]. By its autoactivation sE levels quickly rise upon shock conditions and, therefore, induce a fast and strong response.After the cells have adapted, a quick shutdown of the response is brought about by feedback inhibition of sE by RseA and RseB.


17

RseA is a sE anti-sigma factor located in the inner membrane. It has a cytoplasmic N-terminal domain, one transmembrane domain and a periplasmic C-terminal domain. During normal growth conditions most sE is inactive because it is sequestered by the cytoplasmic part of RseA [132, 133] (Fig. 2). Upon envelope stress RseA is degraded by the sequential action of the two cytoplasmic membrane-bound proteases DegS and YaeL (EcfE) (Fig. 2).Whereas the proteolytic active site of DegS is located at the periplasmic site of the plasma membrane, the active site of YaeL is found at the cytoplasmic site [81, 134–137].YaeL is a part of the sE regulon and essential for cell growth [120, 134]. The proteolysis of RseA leads to the liberation of sE and induction of genes recognised by sE-RNAP (Fig. 2). Therefore, RseA deletion mutants show a constitutive expression of the sE regulon, whereas overexpression of RseA represses the sE response [132, 133]. RseB interacts with the periplasmic domain of RseA and has a negative regulatory function. It is supposed that binding of RseB to RseA stabilises the RseA-sE complex in the absence of envelope stress, hence, keeping sE activity low (Fig. 2). Upon stress the RseB-RseA interaction is weakened, maybe by titration of RseB by misfolded proteins [138] (Fig. 2). The role of RseC in envelope stress regulation is still unclear. 4.2 The Cpx Response

The Cpx response is controlled by the membrane-bound histidine kinase CpxA, its cognate cytoplasmic response regulator CpxR (CpxR-P in its phosphorylated active form) and the periplasmic inhibitor CpxP (see below). Like in the sE response, several genes controlled by the Cpx response encode proteases (DegP [118, 139], HtpX [140]), a disulfide bond oxidoreductase (DsbA [118, 121]), and peptidyl-prolyl-isomerases (PpiA [118], PpiD [141]), which play a role in protein folding and degradation in the envelope.Although similar in function, only a few of the response proteins are identical in the two pathways. The chaperone/protease DegP (HtrA) [139] and probably also the general heat shock response sigma factor s32 [118] can be upregulated not only by sE but also by CpxR-P. The two pathways show not only differences in their target genes but also in the inducing signals.Whereas the sE response is mainly triggered by misfolding of outer membrane proteins, the main signal for the Cpx response seems to be the accumulation of misfolded envelope proteins at the periplasmic site of the inner membrane [122] (Fig. 2). It becomes more and more apparent that the Cpx system is not only involved in envelope stress but in a variety of other cellular processes. These include motility and chemotaxis, biofilm formation, adaptation to or recovery from stationary phase and pathogenesis ([142] and references therein). It is estimated that about 100 operons are under direct control of CpxR-P. Moreover, the Cpx pathway interacts with several other pathways (sS, sE, s32) in a complex, yet unresolved, response network. Interestingly, CpxR-P negatively regulates the other envelope stress system, the sE system [142]. The Cpx response is induced by alterations in membrane structure, as has been shown with cells overexpressing the lipoprotein E, NlpE [143], resulting in the accumulation of the enterobacterial common antigen lipid II biosynthetic in-

two types of regulatory systems, namely the sE response, based on an alternative sigma factor, and the Cpx system, based on a two-component system. The inducing signals and the target genes of the two systems differ for the most part, but the systems show some similar regulatory principles. Misfolded proteins in the envelope are sensed by the anti-sigma factor RseA or the two-component sensor kinase CpxA, respectively. The sensing mechanisms are still unknown. However, the titration of the negative regulators RseB or CpxP, respectively, by misfolded proteins could be an inducing signal. In the sE response, the anti-sigma factor RseA is then degraded by the sequential action of the membrane-bound proteases DegS and YaeL. This liberates sE, which then binds to RNAP and activates transcription of several extracytoplasmic proteases and folding factor genes. In the Cpx response, the sensor kinase gets autophosphorylated at a histidine residue. This phosphate group is then transferred to an aspartate residue of the response regulator CpxR, which in the phosphorylated form activates transcription of its target genes. OMP: outer membrane proteins; -P: phosphate group; RNAP: RNA-polymerase; ?: supposed interaction; scissors: protease activity

Fig. 2 The envelope stress response. There are two major envelope stress responses in E. coli, which are based on two different types of regulatory systems,

18 L.M. Wick · T. Egli


19

termediate in the inner membrane [144], and in mutants lacking phosphatidyl ethanolamine [145]. In addition, the Cpx system is activated at the onset of stationary phase autogenously by CpxR-P and also by sS [146], at alkaline pH [147], upon spheroplast formation [148] and in the presence of misfolded proteins in the periplasm as demonstrated for the misfolded periplasmic domain of the integral membrane protein MalF [149] or misfolded pilin subunits [128]. Activation of the Cpx two-component response occurs through phosphorylation of the response regulator CpxR by the inner membrane histidine kinase CpxA. Upon envelope stress CpxA undergoes a conformational change, which leads to autophosporylation at a histidine residue. This phosphate is then transferred to an aspartate residue of CpxR. The phosphorylated CpxR now recognises specific upstream elements of its target genes and then positively or negatively influences their transcription [118, 142, 150–152] (Fig. 2). In the absence of stress inducing signals, CpxA may also function as a phosphatase and keep CpxR in a dephosphorylated inactive state [151] (Fig. 2). CpxP is a periplasmic protein, whose overexpression leads to an inhibition of the Cpx response. It is believed that CpxP binds to CpxA in the absence of stress and thus inhibits its autophosphorylation. Upon stress CpxP is freed from CpxA maybe by degradation or titration, which activates CpxA (Fig. 2). The Cpx response is not fully induced in a cpxP mutant, but can be further activated. This suggests other Cpx activating factors yet to be identified [147, 148, 153]. The Cpx pathway is subject to autoregulation and feedback inhibition in a similar way as observed for the sE pathway. Both the transcription of the cpxRA operon itself and the cpxP gene, which encodes a negative regulatory element of the Cpx response, are activated by CpxR-P [146, 147, 153] (Fig. 2). 4.3 The Bae Response

As mentioned above, spheroplast formation induces the Cpx response. One member of the Cpx regulon is Spy, which may be involved in outer membrane protein biogenesis. However, Spy is also induced upon spheroplast formation in a Cpxindependent manner [148]. Recently, it has been shown that the BaeS/BaeR two component system is responsible for this Cpx-independent induction of Spy. Spy is induced by the Bae response upon a variety of envelope stresses such as spheroplast formation, production of misfolded P pili subunits and addition of indole, which is expected to disrupt the bacterial envelope. Although spy is the only target gene of the Bae regulon identified so far, experiments with spy mutants suggested that there are other BaeR regulated genes which confer resistance to envelope stresses [154].

5 The Cold Shock Response It is part of the life cycle of E. coli to get shed from the colon of warm-blooded animals into the environment. The shedding may be coupled with an abrupt drop off in temperature and an effective adaptation to cold shock is therefore required

20


to ensure survival. Refrigeration is also a common method to preserve food. Hence, knowledge of bacterial adaptation mechanisms to cold is of great interest. With decreasing temperature the specific growth rate of E. coli slows down without a considerable change in protein levels until about 20 °C. However, a more pronounced decrease in ambient temperature below 20 °C leads to several problems: translation at ribosomes is inhibited, RNA and DNA form stable secondary structures, and membrane fluidity decreases. The cold shock response is the answer to these challenges. The response comprises induction of a specific set of proteins: the cold shock proteins CspA, CspB, CspG and CspI. These proteins are DNA- and RNA-binding proteins and can act as transcriptional activators or mRNA chaperones.As mRNA chaperones they counteract the stabilisation of secondary mRNA structures. Such an activity has been shown to lead to transcription anti-termination of further cold shock genes and might also be important by resolving mRNA secondary structures that impede efficient translation at ribosomes. Further cold shock proteins turn the ribosomes into ribosomes that are functional at lower temperatures.After a transient growth arrest following a cold shock cells are able to resume growth after synthesis of these proteins. There are five additional proteins in E. coli that are named as cold shock proteins (CspC, CspD, CspE, CspF, CspH) because of their homology to CspA. However, these Csp proteins are not induced upon cold shock. In order to increase membrane fluidity cold shocked cells increase the amount of unsaturated fatty acids in the membrane. With respect to controlling membrane fluidity, only two enzymes, one involved in fatty acid and one in lipid A biosynthesis, are found so far to be under temperature control: the b-ketoacylacyl carrier protein synthetase II (FabF), which promotes the elongation of palmitoleic acid to cis-vaccenic acid, and the acyltransferase LpxP, which incorporates palmitoleic acid into lipid A. 5.1 Cold Shock Induced Proteins

Proteins induced by a cold shock from 37 to 15 °C have been classified into two categories: Class I proteins, which are expressed at a very low level at 37 °C and are induced more than tenfold after cold shock, and Class II proteins, which are present at 37 °C and are induced less than tenfold after a cold shock [155]. Besides the cold-inducible proteins of the CspA family CspA, CspB, CspG, and CspI (see below), also CsdA [156], RbfA [157, 158], NusA [159] and PNP [160] belong to class I proteins. Class II proteins include RecA [161], IF-2 [162], H-NS [13], the a-subunit of DNA gyrase [163], Hsc66 [164], HscB [164], trigger factor [165], dihydrolipoamide transferase and pyruvate dehydrogenase [166]. IF-2, CsdA (a member of the DEAD-box helicases) and RbfA (a 30S ribosome associated protein) are involved in translation. These proteins transform coldsensitive non-functional ribosomes into cold-adapted ribosomes, enabling growth at lower temperatures [157] (Fig. 3). H-NS and GyrA levels are increased because the major cold shock protein CspA (see below) interacts with the so-called Y-box motifs in their promoters leading to transcriptional activation [167, 168]. DNA gyrase negatively supercoils


21

Fig. 3 The cold shock response.After a cold shock most cellular mRNA is no more translatable by un-adapted ribosomes, except mRNA for the major cold shock protein CspA and probably other cold shock proteins (CspB, CspI, CspG, CsdA, RbfA). Moreover, cspA mRNA is stabilised at low temperatures. This leads to an increase in cellular CspA levels (1.). CspA functions as a transcriptional activator and a transcription antiterminator, which leads to induction of further cold shock proteins. CspA might also act as a general mRNA chaperone and increase the translatability of bulk mRNA by resolving impeding secondary structures. Some of the cold shock proteins turn the ribosome into a cold-adapted ribosome and thereby restore the cell’s ability to translate bulk mRNA (2.). After a burst in the acclimation phase the levels of cold shock proteins are reduced to a new basal level. This is achieved by a negative feedback loop mediated by the cold shock protein PNP, which selectively degrades mRNAs of cold shock proteins. In addition, CspA negatively regulates its own synthesis. RNAP: RNA-polymerase; ?: supposed interaction; scissors: RNase activity

22


DNA and this form of DNA transiently increases in cold-shocked cells [169]. HNS, a histone-like protein, and a global transcriptional regulator, is essential for effective adaptation to low temperatures [13]. Not only the control of mRNA synthesis but also the control of mRNA degradation is important in gene regulation. Mutants lacking the cold shock induced RNA-degrading enzyme polynucleotide phosphorylase (PNP) are cold sensitive probably because the high levels of cspA mRNA bind efficiently to ribosomes and hamper the translation of bulk mRNA [170, 171]. Trehalose levels were observed to be eight-fold increased in cells after cold shock from 37 °C to 16 °C. Although these trehalose levels did not have an effect on cell growth at 16 °C, they markedly increased cell viability if the temperature was further decreased to 4 °C. Kandror and co-workers [172] have shown that sS is essential for the synthesis of trehalose, and hence, it seems logical that levels of the general stress response sigma factor sS are also increased by cold shock. 5.2 CspA – The Major Cold Shock Protein

One of the most important problems cells encounter at low temperatures is efficient translation of bulk mRNA. One of the problems arising is that mRNAs build secondary structures at lower temperatures, and these structures impede with efficient translation [166, 173, 174] (Fig. 3). A second problem that has to be overcome is that the ribosome, in its state as it is during growth at 37 °C, seems nonfunctioning at 15 °C [157]. The cell solves the former problem by expressing RNA chaperones that are able to resolve the secondary RNA structures, and the latter problem is solved by synthesizing a set of proteins that turn the ribosomes into functional ones at low temperatures [157, 173, 174]. However, in order to express these cold shock genes there must be mRNAs that are efficiently translated at low temperatures by cold-unadapted ribosomes. The mRNA of cspA seems to be the main mRNA that exhibits such properties (for mechanisms see below). The cspA promoter is highly active at both high and low temperature with only a slight increase of cspA transcription after cold shock. Hence, cspA is transcribed at 37 °C, but its mRNA is very unstable (half-life less than 12 s); CspA levels are therefore low at 37 °C. A temperature downshift increases the stability of cspA mRNA drastically (half-life at 15 °C is 20 min). The stabilisation of mRNA plays the primary role in CspA induction. Stabilisation of cspA mRNA is transient until the cells are cold adapted and able to resume growth. As a result of cspA mRNA stability the level of CspA increases to more than 10% of the total cell’s protein. CspA is therefore referred to as the major cold shock protein [175–178]. However, research by Brandi and coworkers [179] showed that CspA is also present in cells that were not cold-shocked suggesting additional roles of this protein. At 37 °C CspA concentrations are growth phase dependent with highest amounts (1% of total cellular protein) during the early exponential phase. This induction was achieved either through dilution of cells from stationary phase into fresh medium or addition of nutrients to stationary phase cells. Therefore, it was suggested that CspA is not only a cold shock but also


23

a nutritional up-shift protein [180]. Fis and H-NS were shown to be important in the growth-phase dependent transcriptional regulation of cspA. Fis, whose levels are high during early exponential phase, activates cspA transcription, whereas H-NS, the levels of which increase during the late exponential phase of growth, represses cspA transcription [179]. Upon cold shock, the cspA mRNA is stabilised and translated by the cold-unadapted ribosomes (Fig. 3). Responsible for this extraordinary feature of the cspA mRNA is an exceptionally long (159 bp) 5¢ untranslated region (UTR) [176]. The first 25 bp of the UTR contain the so-called cold box sequence [181, 182]. Xia and coworkers [183] have shown that this cold box is able to form a stable stem-loop structure that stabilises cspA mRNA after a cold shock. Indeed, the 5¢UTR alone, without the AUG initiation codon and the coding region, directly associates with ribosomes [183]. Thus, the 5¢UTR promotes enhanced translation of cspA at low temperatures. Previous studies suggested that cold-inducible mRNAs contain another region with a characteristic sequence, the so-called down-stream box, which also contributes to the efficient translation initiation at low temperatures. It was proposed that this down-stream box is able to base pair with a complementary sequence of the 16S rRNA and by doing this to enhance translation initiation [184–188]. However, this interaction has been severely questioned by several recent studies [189–192]. CspA recognises the Y-box motif in DNA sequences, which can explain the transcriptional activation of hns and gyrA by CspA [167, 168] (Fig. 3). Besides that, CspA binds to RNA non-specifically and with low affinity. This led to the suggestion that CspA acts as an RNA chaperone and destabilises translation-impeding secondary RNA structures, the formation of which is enhanced at lower temperatures [173] (Fig. 3). At the surface of the CspA protein several aromatic side chains are found. They are proposed to intercalate with single stranded nucleic acids and therefore enable binding of CspA to RNA and DNA [193]. Whereas the proposed role of CspA as a general mRNA chaperone enabling translation of bulk mRNA at lower temperatures has still to be proven, Bae et al. [194] have shown that CspA also acts as a transcription antiterminator. Both, in vivo and in vitro, CspA and its homologues CspE and CspC function as transcription antiterminators at Ç-independent terminators. The prevention of secondary RNA structures in nascent RNA is the proposed mechanism for antitermination, since nucleic acid melting activity of CspE has been shown to be necessary for this process. Transcription antitermination is the reason for induction of several cold-shock inducible genes (nusA, infB, rbfA, and pnp) [194] (Fig. 3). In addition CspA negatively regulates its own synthesis on the transcriptional and translational level and the transient nature of CspA production in the acclimation phase is due to this autoregulation [195].Another negative feedback loop is mediated by the cold inducible RNase PNP, which degrades cspA mRNA [171] (Fig. 3).

24


5.3 The CspA Family

There are nine proteins in E. coli showing high homology to CspA, and they were named alphabetically CspA to CspI [196]. However, not all of these proteins are cold shock inducible. Only four of them, CspA [176], CspB [197], CspG [198], and CspI [181] are induced after a temperature downshift (see below). CspD is a nutritional stress protein. It is induced in early stationary phase, its cellular concentrations are inversely related to growth rate and it is supposed to play a role as an inhibitor of DNA replication in nutrient-depleted cells [199, 200]. CspC and CspE are highly induced at 37 °C [201]. The levels of CspE further increase transiently after dilution of stationary phase cells into fresh medium [202], reminiscent of the induction of CspA under the same conditions [179]. The importance of CspA and CspE in resuming growth after nutrition addition to starved cells is shown by prolonged lag-times of DcspE mutants and even longer lag-times of DcspEcspA double mutants [202]. The function of CspF and CspH is unclear [196]. Recently a novel cold inducible gene, the ves gene, has been found that shows sequence similarity to the genes of the CspA family [203]. The CspA family is related to the eukaryotic Y-box proteins. They all contain a region called cold-shock domain, which binds to a specific regulatory DNA sequence within a so-called Y-box motif [5, 197]. From all this information it is now assumed that the CspA family has evolved by repeated gene duplication and diversification of genes as an adaptive response to different environmental stresses [196]. The cold shock inducible proteins CspA, CspB, CspG, and CspI are regulated differently. CspA is maximally induced between 10 and 24 °C, CspB and CspG are found only below 20 °C with a maximum expression at 15 °C, and CspI is expressed between 10 and 15 °C [181, 204]. Despite these differences, the four proteins can functionally substitute for each other to some degree, because deletions of several csp genes are necessary to result in cold sensitivity. Interestingly, although normally not cold shock induced, CspE accumulates at low temperatures in strains having cspA, cspB and cspG deleted. CspE, therefore, can take over functions of the cold-induced CspA family members and simultaneous deletion of four csp genes including cspE (cspA, cspB, cspE, and cspG) is necessary to produce cold-sensitive mutants [195, 205]. CspE negatively regulates CspA at 37 °C. Therefore, CspA gets highly induced at 37 °C in DcspE mutants. There are indications that the regulation of CspA is caused by an interaction of CspE with the transcription elongation complex at the cspA cold box region [202]. CspE and CspC have been shown to be involved also in the regulation of sS and UspA levels. This function seems to be based solely on the nucleic acid binding properties and not on the nucleic acid melting ability of CspE and CspC. CspE mutants without nucleic acid melting activity but with RNA binding activity show the same effect on sS regulation as the wild type. It has been proposed that rpoS mRNA is stabilised upon binding of CspE and CspC [206, 207].


25

5.4 Sensing of Cold Shock

So far, two main mechanisms have been suggested for sensing a rapid decrease in ambient temperature. On one hand, there is the stabilisation of the cspA mRNA leading to an increased level of CspA. However, the mechanism for this mRNA stabilization is not yet clear. It was proposed that cold induced changes in the secondary structure of the cspA mRNA might change its susceptibility towards RNases [155]. On the other hand, ribosomes have been suggested to be cellular thermosensors [208]. In this model a cold shock reduces the translational efficiency and hence blocks the ribosomal A-site. This inhibits RelA activity and can explain the observed decrease in (p)ppGpp levels after cold shock (see also The stringent response). Moreover, (p)ppGpp0 mutants showed a higher induction of cold shock proteins, whereas mutants with artificially increased (p)ppGpp levels had decreased levels of cold shock proteins after a temperature downshift from 37 °C to 10 °C suggesting a negative role of (p)ppGpp in the control of cold shock genes [209]. The idea that inhibition of initiation of translation is a signal for cold shock induction is supported by the observation that cold-sensitive ribosomal mutations and certain antibiotics can also induce cold shock. One hypothesis is that upon blocking of ribosomes, the mRNA degradative machinery, which is normally tightly coupled to the ribosome, dissociates from the ribosome. This in turn could lead to the stabilisation of the cspA mRNA [155]. Also the topology of the chromosomal DNA might be involved in temperaturedependent control of some genes. Changes in temperature cause changes in DNA twist. The promoter activity of certain genes might be especially sensitive to the relative position of the –10 and –35 regions, which is dependent on DNA twist. It has been proposed that the increase in RecA synthesis upon cold shock is due to the temperature induced change in DNA twist [210]. 5.5 Changes in Membrane Composition

With decreasing temperature the fluidity of the membrane decreases, which has a negative effect on the functions of the membrane. Therefore, it is important that cells control the fluidity of their membranes to maintain it functional (homeoviscous adaptation) [211]. To this end E. coli increases its proportion of unsaturated fatty acids in the membrane with falling temperatures [212]. Palmitoleic acid (C16:1cisD9)and cis-vaccenic acid (C18:1cisD11) are practically the only two unsaturated fatty acids in E. coli. The key difference with respect to cold shock adaptation between these two fatty acids is that palmitoleic acid can be incorporated only in position two of the phospholipid backbone, whereas cis-vaccenic acid can be incorporated also in position one because it is able to compete with the saturated palmitic acid for this position [213]. Therefore, an increase in cis-vaccenic acid levels leads to a higher ratio of unsaturated fatty acids in phospholipids. Such an increase is mediated by the activity of b-ketoacyl-acyl carrier protein synthetase II (encoded by the fabF gene), which catalyses the elongation of palmi-

26


toleic acid to cis-vaccenic acid. This enzyme is present also at higher temperatures, but its activity is increased at lower temperatures [214, 215]. It seems to be the only enzyme involved in fatty acid biosynthesis that is controlled by temperature. The composition of the glycerophospholipids is then governed by the relative concentrations of palmitic, palmitoleic and cis-vaccenic acid [213]. In contrast to this, the composition of the outer membrane lipid A is controlled by the activity of different acyltransferases during lipid biosynthesis. Palmitoleic acid is not present in lipid A of E. coli grown at 30 °C. However, it comprises about 11% of fatty acyl chains of lipid A in cells grown at 12 °C, where it is incorporated into lipid A instead of lauric acid. The acyltransferase LpxP, which is responsible for the incorporation of palmitoleic acid, is induced about 30-fold after temperature downshift from 30 to 12 °C [216]. It is suggested that the LpxP activity makes the membrane a more effective barrier against harmful chemicals at low temperatures [217].

6 The Stringent Response The stringent response tunes anabolism with available resources. It serves as a control mechanism that reduces the cellular protein synthesis capacity, when substrates for protein synthesis get scarce, and a high level of protein synthesis machinery would be a waste of energy. It was first recognised as an inhibition of stable RNA (rRNA, tRNA) synthesis when cells were starved for amino acids. A mutant was isolated that did not show this stringent coupling of stoppage of stable RNA synthesis upon amino acid starvation, but continued to accumulate stable RNA for about one generation time. This mutant was termed relaxed and the genetic locus responsible was therefore called relA. The stringent response is triggered not only by amino acid starvation but also other nutritional limitations, such as carbon-, nitrogen-, and phosphorus-limitation, and a variety of other stresses. All these events lead to an increase in the cellular concentration of the mediator of the stringent response (p)ppGpp. The main two enzymes controlling (p)ppGpp levels are RelA and SpoT. RelA is bound to the ribosome and catalyses (p)ppGpp synthesis when ribosomal elongation stalls due to the presence of uncharged tRNAs upon amino acid starvation. SpoT is located in the cytoplasm and is responsible for an increase in (p)ppGpp levels in response to a variety of stresses, which influence the (p)ppGpp synthesis and degradation activity of SpoT. The exact mechanism by which SpoT responds to inducing stresses is unknown, but the cellular concentration of uncharged tRNAs might be one signal. Further enzymes controlling the concentration of (p)ppGpp are Gpp, and Ndk. 6.1 Regulation of (p)ppGpp Synthesis and Decay

The cellular concentrations of the mediators of the stringent response, the alarmones pppGpp and ppGpp (summarised as (p)ppGpp), are regulated by the enzymes RelA, SpoT, Gpp, Ndk. It seems that pppGpp and ppGpp have the same effects in a cell; however, this has not been systematically explored [218]. Both RelA


27

((p)ppGpp synthetase I) and SpoT ((p)ppGpp synthetase II) can synthesise (p)ppGpp from GTP (or GDP) and ATP. Whereas RelA has only synthetase activity and seems to be the main synthesizing enzyme, SpoT has only a weak synthetase activity but mostly acts as a hydrolase and degrades (p)ppGpp to GDP or GTP. Gpp hydrolyses pppGpp to ppGpp, and Ndk finally restores the pool of GTP by phosphorylating GDP to GTP (reviewed in [218]). (p)ppGpp levels not only increase upon amino acid starvation [219] but also upon starvation for several nutrients (C, N or P) [220–222], increase in osmolarity [223], oxidative stress (H2O2), 10% ethanol, CdCl2, 6-amino-7-chloro-5,8-dioxoquinoline [224] and with decreasing growth rate [225–227]. A temperature upshift leads to an increase in (p)ppGpp levels [228], a temperature downshift results in a decrease [229] (see also Cold shock). Whereas RelA is mainly responsible for (p)ppGpp synthesis during amino acid starvation, (p)ppGpp levels during balanced growth and in response to most other environmental conditions seem to be regulated by the synthesis and degradation activity of SpoT ([230, 231]; Fig. 4; see also below). 6.1.1 RelA

RelA, which is found in cellular extracts associated with the ribosome (see below) is responsible for the main synthesis of (p)ppGpp, by transferring pyrophosphate from ATP to GTP or GDP, respectively [232]. The Km of RelA for GTP and GDP is similar (0.5 mmol/l) and this affinity is in the range of the cellular GTP concentration. Because in the early phase of the stringent response GDP concentrations are usually an order of magnitude lower than those of GTP, pppGpp is the main product of the stringent response [233–235]. pppGpp is subsequently converted to ppGpp by Gpp [236]. RelA consists of two domains, the catalytic N-terminal domain and the regulatory C-terminal domain [237, 238]. It seems that RelA-RelA interactions, mediated by the C-terminal domain, have an important function in the regulation of RelA activity [239]. All cellular RelA is found associated with the 50S ribosomal unit [235, 240] (Fig. 4). Due to the low abundance of RelA only about 0.5%–1% of ribosomes have bound RelA. An overexpression of RelA therefore increases the amount of ribosomes with bound RelA and also the chance of an uncharged tRNA encountering a RelA-carrying ribosome. This elevates the (p)ppGpp levels in such cells and immediately lowers the growth rate due to inhibition of stable RNA synthesis [237]. Because interaction of the ribosome and RelA is necessary to elicit the stringent response, not only mutations in the relA gene but also mutations in the ribosomal protein L11 (encoded by the rplK (=relC) gene) can lead to a relaxed phenotype similar to relA mutants [240, 241]. It has been shown recently that a proline rich helix in the N-terminal portion of the ribosomal protein L11 is necessary for the activation of RelA activity [242]. Binding of uncharged tRNAs, which increases during amino acid starvation, to the A-site of the ribosome triggers the synthesis of (p)ppGpp by RelA [243] (Fig. 4). At each cycle of uncharged tRNA binding and release, (p)ppGpp is syn-

28


Fig. 4 The major triggers and consequences of the stringent response. RelA is only active in as-

sociation with the ribosome and it synthesises (p)ppGpp when free tRNA is encountered at the A-site of the 50S ribosome. SpoT is located in the cytoplasm and shows both (p)ppGpp synthesizing and degradation activity. Several stresses influence these two activities resulting in a net increase of (p)ppGpp. The sensing mechanisms for the SpoT-mediated (p)ppGpp increase are still unknown. RMF: Ribosome modulation factor

thesised, and thus (p)ppGpp levels are increasing with the fraction of uncharged tRNA [244]. 6.1.2 SpoT

SpoT, a protein homologous to RelA [245], was first recognised to be a (p)ppGpp 3¢-pyrophosphohydrolase involved in degradation of (p)ppGpp [246, 247]. relA mutants are still able to accumulate (p)ppGpp in response to C-, N-, and P-star-


29

vation, and increase in osmolarity, but not during starvation for less than 10 amino acids [220–222, 231, 238]. They also maintain basal levels of (p)ppGpp in balanced growth [227, 238]. The search for a second enzyme involved in (p)ppGpp synthesis led to the discovery that SpoT not only has a (p)ppGpp 3¢pyrophosphohydrolase but also a (p)ppGpp synthetase activity [248, 249]. In relA spoT double mutants, (p)ppGpp production is abolished and these mutants are therefore designated as (p)ppGpp0 strains [249]. The catalytic sites involved in the degradation and synthesis of (p)ppGpp are found in distinct but overlapping regions in the SpoT protein [250]. In contrast to RelA, SpoT is not bound to ribosomes but is free in the cytosol [251] and SpoT activity is not controlled by the ribosomal protein L11 [231]. Accumulation of (p)ppGpp mediated by RelA and SpoT is triggered by different signals (Fig. 4). Whereas the increase in RelA activity upon emergence of uncharged tRNA at the A-site of translating ribosomes is well understood, the control of (p)ppGpp levels due to synthetic or degradation activity of SpoT is less elucidated. Using a relA mutant, Murray and Bremer [231] determined changes in SpoT mediated synthesis and degradation during different stresses. On one hand, multiple amino acid starvation and energy deprivation (upon addition of sodium azide) both led to an increased rate of synthesis and a reduced degradation rate of (p)ppGpp. On the other hand, both, synthesis and degradation rates were reduced due to carbon source starvation; however, rates of (p)ppGpp degradation were decreased more than those for the synthesis explaining the net increase of (p)ppGpp. Based on these observations the following control mechanisms were suggested: The synthetase activity of SpoT is unstable (average lifetime 40 s) and its activity is generated during or shortly after spoT mRNA translation depending on amino acid availability. Due to this instability continuous protein synthesis is required to maintain (p)ppGpp synthesis by SpoT. The environmental stresses, i.e. multiple amino acid starvation and carbon/energy deprivation, control SpoT hydrolysis activity [231]. It has been demonstrated in in vitro experiments that SpoT (p)ppGpp hydrolase activity is inhibited by uncharged tRNAs [252], hence the in vivo signal sensed by SpoT might be the general accumulation of uncharged tRNA [231]. Therefore, starvation of a single amino acid triggers the RelA response due to idling ribosomes, but not the SpoT stringent response since in this model SpoT senses overall free tRNA [231]. Overall free tRNA levels are high only in cells starved for multiple amino acids, not in cells starved for a single amino acid, since in the latter case charging of all other tRNAs increases due to the slowed protein synthesis [253]. Although RelA is mainly involved in the response to amino acid starvation and SpoT in the response to other stresses and in controlling the steady state levels of (p)ppGpp during balanced growth, it remains difficult to assign the exact contribution of each of these two enzymes to the (p)ppGpp levels found under different cultivation conditions.

30


6.2 Effects and Mechanisms of (p)ppGpp 6.2.1 Effects of (p)ppGpp

The effects of (p)ppGpp are pleiotropic [218] (Fig. 4). The primary characteristic of the stringent response is the decrease in stable RNA levels. Induction of sS, inhibition of active transport of several metabolites and especially the enhanced transcription of some amino acid biosynthesis enzymes are further effects of (p)ppGpp accumulation. Furthermore, (p)ppGpp is probably involved in bacterial cell cycle regulation (reviewed in [218]). In fact, (p)ppGpp0 mutants are auxotrophic for several amino acids showing the need of basal (p)ppGpp levels for transcription of some enzymes involved in amino acid biosynthesis [249]. Therefore, mutants able to confer prototrophy to relA/spoT double mutants are important in the research for the mechanisms and sites of action of (p)ppGpp [254]. A major effect of the stringent response resulting from the quick shutdown of stable RNA synthesis is a reduction of the number of ribosomes during nutritional downshift. However, (p)ppGpp additionally modulates the cellular translation capacity over the ribosomal modulation factor (RMF), whose expression requires (p)ppGpp. RMF reversibly promotes dimerisation of active 70S ribosomes to inactive 100S ribosomes and thus controls the cellular protein synthesis capacity without influencing the total number of ribosomal proteins [255]. Recently, a mechanism based on the toxin-antitoxin system RelBE was described, in which RelE acts as a global inhibitor of translation during amino acid starvation in a (p)ppGpp-independent manner [256]. Another toxin-antitoxin system, the MazEF system, is under control of (p)ppGpp. It has been suggested that this system is responsible for programmed cell death, which can be triggered by some antibiotics and maybe other stresses such as starvation [257, 258]. (p)ppGpp is also important in positive control of the sS-mediated general stress response. It has been shown that (p)ppGpp is not only necessary for the production of sS [37] but also for the transcription of sS-dependent promoters [38]. This requirement of (p)ppGpp is due to its role as a regulator of sigma factor competition. Both sS and the heat shock sigma factor s32 compete more successfully for RNA core polymerase with s70 in the presence of elevated (p)ppGpp levels [259]. In order to synthesise proteins needed in response to starvation and to readjust the cellular composition of the cell, it is necessary to degrade intracellular proteins for the supply of amino acids.A major source of amino acids upon starvation are the ribosomes, with Lon and Clp being the major proteases involved in their degradation. Regulation of Lon-dependent ribosome degradation is linked to the stringent response via polyphosphate. As a result of a competitive inhibition by (p)ppGpp of exopolyphosphatase Ppx, the enzyme responsible for PolyP degradation, polyphosphate accumulates in cells growing at low levels of phosphate (0.1 mmol/l Pi) and amino acids (2 mg/l, each) dependent on elevated (p)ppGpp levels [260]. Recent experiments suggest that the binding of polyphosphate to certain free ribosomal proteins and to the Lon protease stimulates Lon-


31

dependent degradation of ribosomal proteins [261]. Interestingly, PolyP also seems to play a role in inducing sS expression [262]. Besides sS the expression of other global regulators is influenced by cellular (p)ppGpp concentrations, namely the cAMP receptor protein (CRP) (negatively) and Lrp (positively) [39]. In fact, in their recent expanded stringent response model, Chang et al. [263] place (p)ppGpp as a global regulator at the top of a network, which is triggered by any environmental condition leading to growth arrest. Transcription in the cell is then reprogrammed by (p)ppGpp leading to induction of stress survival genes and repression of transcription, translation, and DNA replication. It should be noted that this model is based on gene expression profiling of wt cells under different conditions and needs to be tested in further experiments with mutant strains. 6.2.2 Mechanisms of (p)ppGpp Regulation

The complexity of the stringent response makes it difficult to distinguish between direct and indirect effects of (p)ppGpp. Several studies have suggested a direct interaction of (p)ppGpp with RNAP. The interaction seems to include the b,b¢ and s subunits of RNAP [254, 264–267]. Barker et al. [268, 269] studied transcription of positively (for amino acid synthesis) and negatively (for rRNA transcription) regulated promoters in vivo and in vitro to elucidate the mechanisms of (p)ppGpp regulation. Based on their results and previous findings [267, 270] they proposed the following model: Binding of (p)ppGpp to RNAP decreases the lifetime of the open complex at all promoters. For the rRNA promoters whose open complex lifetimes are short and rate-limiting for transcription initiation such a decrease leads actively to a decrease in transcription rate. In contrast, the positive regulation on some amino acid promoters seems to be an indirect one. Since the open complex of these promoters are long-lived, the decrease in open complex lifetime upon (p)ppGpp binding does not affect the transcription rate. But transcription from these promoters is limited by RNAP-binding and, therefore, they are very sensitive to levels of free RNAP. The increase of free RNAP, which results from the decreased expression of stable rRNA genes (which can comprise up to 80% of the total RNA and, therefore, a change in transcription of these operons has an immediate influence on free RNAP levels) enhances the expression of those genes with low RNAP-binding efficiency. However, the possibility that the positive regulation is not only achieved passively by an increase in free RNAP, but is also actively stimulated by additional unknown factors, as suggested by Choy [271], cannot be ruled out completely. 6.2.3 Growth Rate Control by (p)ppGpp

The cellular content of ribosomes has to meet the demands of protein synthesis, which is between 0.3 h–1 and 1.7 h–1 directly dependent on specific growth rate. The growth rate dependent control takes place at the level of rrn transcription, where the control of ribosomal protein synthesis is coupled to the production

32


of rRNA through feedback mechanisms [272]. Since upon nutritional downshift (p)ppGpp mediates a decrease in rrn operon transcription and (p)ppGpp levels in the cell are inversely related to the growth rate, this alarmone could be responsible for growth rate dependent control of ribosome synthesis. However, also in (p)ppGpp0 strains rRNA transcription is directly dependent on growth rate, therefore (p)ppGpp is not essential for the growth rate dependent regulation [273]. A promising candidate for the regulation of rRNA transcription is the intracellular level of the initiating NTP (ATP or GTP). Gourse and co-workers [274, 275] suggested that cytosolic NTP concentrations indicate the translational capacity of a cell and that rRNA transcription is regulated by changing NTP concentrations, a process called NTP sensing. However, NTP pools in growing cells are difficult to measure and results contradicting the NTP sensing model have been reported [276]. Therefore, the view on how this “growth rate control” of ribosome content is achieved is still not understood and, depending on growth conditions, rRNA regulation might depend on the interplay of several mechanisms.

7 The General Stress Response The general stress response is mediated by the product of the rpoS gene and confers resistance to a variety of stresses to the cell. sS is a second major s-factor besides s70 and its target sequences show a high degree of homology to the ones of s70. Activation of sS is triggered by a number of environmental conditions, including nutrient starvation (carbon, nitrogen, phosphorus, amino acids), growth under nutrient limitation (carbon, nitrogen), high osmolarity, shift to high or low temperature, and shift to acidic pH. First studies on sS were done in starving cultures, and this sigma factor was therefore initially called the starvation s factor (sS). However, due to its role as a major regulator during a variety of stresses it nowadays seems more appropriate to call it the general stress response s factor. 7.1 Regulation of s S

Induction of sS is triggered by different environmental factors such as entry into stationary phase, heat and cold shock, acid pH, osmotic shock, and oxidative stress (for reviews see [277–281]). Depending on the stress inducing the response, an increase of sS levels in cells is due to regulation at the transcriptional, the translational or the protein stability level (Fig. 5). However, regulation by translational and posttranslational control seems to be much more important than regulation by transcriptional control. 7.1.1 Transcriptional Regulation

Three promoters are responsible for rpoS transcription. From two promoters, rpoS is transcribed together with the gene nlpD, which is located upstream of


33

Fig. 5 The general stress response is regulated at the transcriptional, translational, and post-

translational level with different stresses acting at different levels. This figure is reproduced from [279] with the permission of the publisher

rpoS and codes for a lipoprotein of unknown function. These two promoters are weak and account for the basal low-level expression of sS in growing non-stressed cells [282–284]. The third promoter lies within the nlpD gene and seems to be responsible for the major part of rpoS transcription. Reduced growth rate is thought to further increase transcription of rpoS from this third promoter [283, 285–287] (Fig. 5). Several molecules were proposed to influence transcription of the rpoS gene including cAMP [285, 286, 288], (p)ppGpp [37, 283, 289], polyphosphate [262], BarA [290], and weak acids [291, 292]. Contradicting results have been found for the role of cAMP, both a negative [285, 286] and a positive [288] effect on transcription were reported. (p)ppGpp has a double role in controlling the sS response. In addition to the need of (p)ppGpp for transcription of the rpoS gene itself [37, 283, 289], (p)ppGpp is also required for the transcription of sS-dependent genes [38]. Recent findings suggest that (p)ppGpp increases the ability of sS and s32 to compete with s70 for binding to core polymerase [259]. Those studies [37, 38, 259, 283, 289] have shown that (p)ppGpp is essential for efficient sS expression and also has a role in regulation of sS-dependent genes. However, whether or not variations in cellular (p)ppGpp concentrations also have a regulatory role (besides the need of (p)ppGpp for basal expression of sS) on sS levels in the cell is still unclear. On one hand, Brown et al. [293] have found a regulatory effect of (p)ppGpp on sS expression. However, the major effect was not found on the transcriptional level, but the translational level. The same study also showed that DksA affected this translational induction of sS. On the other hand, a regulatory function of (p)ppGpp has been questioned recently because the induction ratio of rpoS-lacZ fusions upon entry into stationary phase was the same in wt and ppGpp0 mutants [294]. The main difference was the much lower absolute level of sS in ppGpp0 mutants at all times. There are many open questions regarding transcriptional control of sS and in the published studies contradicting results are reported. Comparison of results from different laboratories is hampered by the use of different strains, media, and experimental conditions. Several studies on sS for example have been done with

34


E. coli MC4100, a strain with a relA mutation, which affects control of (p)ppGpp levels. 7.1.2 Translational Regulation

Enhanced translation of the rpoS mRNA is mediated by stresses such as low temperature [295], high osmolarity [40, 286] and late stationary phase [286] (Fig. 5). It is assumed that under normal conditions the rpoS mRNA builds a secondary stem-loop structure, which impedes efficient translation by occluding the Shine Dalgarno sequence [296]. At high osmolarity or low temperature this secondary structure is supposed to undergo a conformational change and that this change increases rpoS translation. Factors influencing translation of rpoS mRNA are the small regulatory RNAs DsrA, RprA, and OxyS. DsrA is able to bind to the 5¢ untranslated region of rpoS mRNA. This binding opens the inhibitory secondary structure, thus enables access of ribosomes and as a consequence enhances translation [297–299]. Both the synthesis and the stability of DsrA are increased at lower temperature making DsrA a thermometer for rpoS control and explaining the increase in sS levels upon temperature downshift [300]. RprA, another small untranslated RNA, has been shown to be responsible for stimulation of sS translation in DsrA negative strains after osmotic shock. Since, in contrast to DsrA, RprA does not have an extensive region of complementarity to the sS leader, its mode of action is unclear [301]. Whereas DsrA and RprA stimulate rpoS translation, the small RNA OxyS inhibits this process. OxyS regulates about 40 genes as a response to oxidative stress. The repression of sS by OxyS is assumed to prevent activation of the general stress response genes under conditions where oxidative stress response mediated by OxyR is sufficient [302, 303]. An important factor involved in the regulation of rpoS expression through all of these sRNAs is the small RNA-binding protein Hfq. Hfq is necessary for rpoS translation [40, 304–306], and it binds to DsrA, OxyS, RprA [15] and also rpoS mRNA [302]. It is assumed that Hfq mediates RNA-RNA interactions and thus enables regulation of mRNA translation through small regulatory RNAs [307, 308]. The RNA binding proteins CspC and CspE, which belong to the CspA family (see also “Cold shock response”), also seem to have a regulatory effect on sS levels. In strains overexpressing either CspC or CspE the rpoS mRNA was more abundant and more stable than in the wt [206]. Further proteins influencing rpoS mRNA translation are the DNA-binding, histone-like proteins HU and H-NS, both of which are also able to bind to RNA. Whereas HU stimulates sS expression [309], H-NS regulates it down [310, 311]. Further factors that have been proposed to affect translation of rpoS are the heat shock protein DnaK (stimulating) [312], and the metabolite UDP-Glucose (inhibiting) [313]. Despite the apparent wealth of information there are still many open questions concerning the mechanisms by which these factors influence sS translation, and even less is known about the regulation of these factors by environmental conditions.


35

7.1.3 Posttranslational Regulation

Control of proteolysis is the main mechanism ensuring a quick rise in sS levels after sudden environmental changes. Despite all the regulatory elements in transcriptional and translational control, sS is expressed constitutively in fast growing cells, but it is degraded at a high rate (half-life of about 1.4 min) and its cellular levels are, therefore, very low [286]. The protease ClpXP is responsible for sS turnover [314]. Carbon starvation [286], shifts to high temperature [312], low pH [315], or high osmolarity [316] dramatically decrease the turnover rate of sS, and let its cellular levels rise quickly following stress (Fig. 5). RssB, a protein homologous to response regulators of two component systems, plays a central role in the degradation of sS. RssB forms a complex with sS in a 1:1 stoichiometry and phosphorylation of RssB enhances this reaction. The presence of RssB is essential for sS degradation; hence, sS is stable in rssB null mutants. The RssB-sS complex is bound by ClpXP, which then unfolds and degrades sS concomitant with the hydrolysis of ATP. RssB acts catalytically in this process, i.e. it is not degraded together with sS. The factors and conditions influencing the phosphorylation state of RssB and, therefore, the turnover rate of sS are still unknown [317–322]. One lysine residue in sS (K173) is absolutely necessary for binding of RssB and sS turnover. Interestingly, this very amino acid is also important in recognition of rpoS-dependent promoters. When K173 is mutated to glutamate (glutamate is the amino acid present at the corresponding position in s70) the characteristics of the mutated sS are more like the one of s70, both in promoter recognition and turnover [317, 318, 323]. DnaK has a double function in that it not only stimulates rpoS translation but also stabilises the sS protein. DnaK mediates a decrease of sS turnover upon nutrient starvation and heat shock.Whether DnaK directly protects sS from degradation or diminishes the activity of the components involved in sS degradation is not known [312, 324]. Although a heat shock leads to enhanced sS levels, the sS-regulated genes do not seem to play a direct role in heat adaptation upon heat shock, but the cross-protection towards other stresses might be of advantage under those natural conditions that elicit the heat shock response [312]. 7.2 Effects of s S 7.2.1 Physiological Effects of s S

sS controls resistance to a variety of stresses to the cell, not only to the stress that elicits the response, i.e. leads to a cross protection. Under environmental conditions that lead to slow growth or growth arrest de novo protein synthesis is hampered, which makes it difficult for the cell to respond to an additional stress quickly and effectively. Therefore, it makes sense when cells are ready for a variety of stresses to follow, with sS preparing them for this situation [279]. The morphological and physiological characteristics of cells expressing sS are sphere-

36


shaped cells, showing higher resistance against hydrogen peroxide and oxygen radicals, desiccation, acid and basic pH, osmotic stress, ethanol, heat and cold [172, 278]. Trehalose is an interesting disaccharide that protects the cell against various stresses, such as high osmolarity, desiccation, cold, and partly to heat [172, 325, 326]. The operon otsBA encoding the trehalose-synthesizing enzymes is sS-dependent [327]. sS has also been reported to play a role in biofilm formation [328–330], in control of virulence factors in pathogenic E. coli [331] and in programmed cell death [332]. Because some of the proteins regulated by sS have regulatory functions themselves, complex regulatory networks may be triggered by conditions that lead to increased sS levels [279]. 7.2.2 s S-Dependent Promoters

The sS-dependent promoter sequences are very similar to the s70-dependent ones. In vitro, most promoters can be activated by both s factors [333–335]. In spite of this similarity there are several sequence specific factors that make a promoter optimal for sS recognition. Such an optimal promoter sequence was suggested by Becker and Hengge-Aronis [323] to have no –35 region, a TC motif at –14/–13 positions and a TATACT hexamer at –10 position. In addition to these sequence specific factors, there are several other factors that contribute to the s factor specificity of promoters. Such factors are salt concentration, DNA supercoiling, and the regulators H-NS, Lrp, cAMP-Crp, IHF, and Fis (summarised in [336]). 7.2.3 Role of s S in Various Habitats

Interestingly, although needed as a response to many stresses encountered in the environment, the expression of sS is of disadvantage when heterotrophic cells have to cope with one stress which is very common in the environment, namely carbon/energy limitation. sS levels in E. coli increase with decreasing growth rate and it seems that E. coli already prepares for stationary phase at growth rates around 0.1–0.2 h–1 [337]. sS is believed to compete with s70 for the RNA core polymerase leading to decreased transcription of s70-dependent genes when sS levels are high [338]. Since genes for high affinity glucose uptake are s70-dependent, rpoS positive strains have lower levels of high affinity uptake systems at low growth rates [339, 340] and, hence, a higher Ks than rpoS negative strains [341]. Such a negative influence on Ks might not only occur during growth with glucose but also for a variety of other carbon substrates. This is especially puzzling because ecologically speaking specific growth rates around 0.1 h–1 are not very low when one considers that the estimated average specific growth rate of E. coli in the intestine is 0.07 h–1 or less [342, 343] (this is of course an average value and growth rate may vary with fluctuating nutrient inputs). In aquatic environments growth rates are certainly lower. However, it is exactly in such environments where the ability to take up nutrients with high affinity is important. This might explain why mutants with rpoS mutations are found in environmental isolates


37

[344], and that they are selected for in the stationary phase of C-limited batch cultures (so-called GASP mutants) [344, 345] and carbon- and nitrogen-limited chemostats [339]. In this context it is worth mentioning that the long-term starvation cultures as they have been used to select for rpoS mutants that take over such starving populations, resemble ill-defined continuous cultures. In such cultures, a significant part of the cells dies in the stationary phase and the mutant sub-population growths on lysis products of their deceased siblings. Hence, the conditions are characterised by cryptic growth on a mixture of different cell lysis and hydrolytic products of cell debris. The continuously released and produced nutrients are efficiently removed by the surviving sub-population. However, the vital role of sS in protection against different stresses is indisputable. For example, cells re-entering their primary habitat, the colon of warmblooded animals, benefit from the sS-mediated protection against the low pH in the stomach [346]. This is probably the reason why in clinical isolates attenuated rpoS alleles are found less frequently than in environmental isolates (discussed in [347]).

8 Conclusions and Perspectives Escherichia coli has evolved a multitude of sensing and effector mechanisms in order to respond to the variety of conditions to which it might get exposed during its lifetime. With increasing knowledge of the molecular components involved in these mechanisms it becomes more and more clear that the stress responses are linked to each other in multiple ways. This means that when we turn at one knob of the cellular regulation by manipulating a strain genetically, there might be a response at an unexpected end of the cell. On the other hand, by taking alternative routes the cellular system might be able to buffer the genetic manipulation and the intended effect will stay away. Therefore, the major future challenge will be to understand the whole network of stress responses and its functionality under different environmental conditions. However, we still are far from having all the pieces of the puzzle to get the whole picture. Recent methods such as gene expression profiling with DNA arrays will considerably speed up this process, allowing us to gather enormous sets of data in just a few experiments. Cluster analysis of such data obtained under different stress conditions are useful to create and test hypotheses about the role of global stress regulators (see for example [14, 263]). Besides that, other approaches will be needed in order to elucidated the full mechanisms of stress sensing, which for several stresses is still poorly understood (e.g. SpoT-dependent stringent response; sS), but also large parts of the stress response regulation, which often occurs at the posttranscriptional level. The interaction between stress responses also might explain some of the contradicting results that are reported from different groups. Although the same main stress might have been applied, there might be some hidden second stresses influencing the experimental outcome. One example is the use of badly designed media, lacking for example trace elements, which can lead to an unintended nutritional stress besides the stress applied deliberately [348]. This makes it important that experimental conditions such as

38


media, cultivation conditions, and the genetic background of the strains, are carefully chosen, controlled, and reported in order to prevent post hoc, ergo propter hoc fallacies. Acknowledgement We thank Thomas Whittam for helpful comments on the manuscript and for

financial support of L.M.W. during part of this project. The work was also supported partly by a grant from the Swiss National Science Foundation (NF 31-50885.97). Note added in proof The role of quorum sensing on RpoS leves has now been ruled out [350]. Increasing RpoS levels correlated strongly with decreasing specific growth rate but not with cell density.

9 References 1. Egger LA, Park H, Inouye M (1997) Genes Cells 2:167 2. Hoch JA, Silhavy TJ (1995) Two-component signal transduction.ASM Press,Washington, DC 3. Wosten MMSM (1998) FEMS Microbiol Rev 22:127 4. Godfrey HP, Bugrysheva JV, Cabello FC (2002) Trends Microbiol 10:349 5. Wolffe AP (1994) Bioessays 16:245 6. Matsumoto K, Wolffe AP (1998) Trends Cell Biol 8:318 7. Hartl FU, Hayer-Hartl M (2002) Science 295:1852 8. Drlica K (1992) Mol Microbiol 6:425 9. Luttinger A (1995) Mol Microbiol 15:601 10. Ali Azam T, Iwata A, Nishimura A, Ueda S, Ishihama A (1999) J Bacteriol 181:6361 11. Pettijohn DE (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, p 158 12. Atlung T, Ingmer H (1997) Mol Microbiol 24:7 13. Dersch P, Kneip S, Bremer E (1994) Mol Gen Genet 245:255 14. Hommais F, Krin E, Laurent-Winter C, Soutourina O, Malpertuy A, Le Caer JP, Danchin A, Bertin P (2001) Mol Microbiol 40:20 15. Wassarman KM, Repoila F, Rosenow C, Storz G, Gottesman S (2001) Genes Dev 15:1637 16. Storz G (2002) Science 296:1260 17. Wassarman KM, Zhang A, Storz G (1999) Trends Microbiol 7:37 18. Eddy SR (2001) Nat Rev Genet 2:919 19. Altuvia S, Wagner EGH (2000) PNAS 97:9824 20. Helmann JD (1991) Mol Microbiol 5:2875 21. Reitzer L, Schneider BL (2001) Microbiol Mol Biol Rev 65:422 22. Visca P, Leoni L, Wilson MJ, Lamont IL (2002) Mol Microbiol 45:1177 23. Hughes KT, Mathee K (1998) Annu Rev Microbiol 52:231 24. Mizuno T (1997) DNA Res 4:161 25. Jung K, Altendorf K (2002) J Mol Microbiol Biotechnol 4:223 26. Cai SJ, Inouye M (2002) J Biol Chem 277:24155 27. Magasanik B (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, pp 1344– 1356 28. Wanner BL (1995) In: Hoch JA, Silhavy TJ (eds) Two-component signal transduction.ASM Press, Washington, DC, p 203 29. Patschkowski T, Bates DM, Kiley PJ (2000) In: Storz G, Hengge-Aronis R (eds) Bacterial stress responses. ASM Press, Washington, DC, p 61


39

30. Lynch AS, Lin ECC (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, pp 1526–1538 31. Storz G, Imlay JA (1999) Curr Opin Microbiol 2:188 32. Storz G, Zheng M (2000) In: Storz G, Hengge-Aronis R (eds) Bacterial stress responses. ASM Press, Washington, DC, p 47 33. Newman EB, Lin RT, D’Ari R (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, p 1513 34. Calvo JM, Matthews RG (1994) Microbiol Rev 58:466 35. Tani TH, Khodursky A, Blumenthal RM, Brown PO, Matthews RG (2002) PNAS 99:13471 36. Hung S-P, Baldi P, Hatfield GW (2002) J Biol Chem 277:40309 37. Gentry DR, Hernandez VJ, Nguyen LH, Jensen DB, Cashel M (1993) J Bacteriol 175:7982 38. Kvint K, Farewell A, Nystrom T (2000) J Biol Chem 275:14795 39. Landgraf J, Wu J, Calvo J (1996) J Bacteriol 178:6930 40. Muffler A, Traulsen D, Fischer D, Lange R, Hengge-Aronis R (1997) J Bacteriol 179:297 41. Romeo T (1998) Mol Microbiol 29:1321 42. Bruckner R, Titgemeyer F (2002) FEMS Microbiol Lett 209:141 43. Botsford JL, Harman JG (1992) Microbiol Rev 56:100 44. Chen X, Schauder S, Potier N,Van Dorsselaer A, Pelczer I, Bassler BL, Hughson FM (2002) Nature 415:545 45. Miller MB, Bassler BL (2001) Annu Rev Microbiol 55:165 46. Michael B, Smith JN, Swift S, Heffron F, Ahmer BMM (2001) J Bacteriol 183:5733 47. Rahmati S, Yang S, Davidson AL, Zechiedrich EL (2002) Mol Microbiol 43:677 48. Schauder S, Shokat K, Surette MG, Bassler BL (2001) Mol Microbiol 41:463 49. Sperandio V, Torres AG, Kaper JB (2002) Mol Microbiol 43:809 50. Winzer K, Hardie KR, Williams P (2002) Curr Opin Microbiol 5:216 51. Tissieres A, Mitchell HK, Tracy UM (1974) J Mol Biol 84:389 52. Neidhardt FC,VanBogelen RA (1987) In: Neidhardt FC, Ingraham JL, Low KB, Magasanik B, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol 2. ASM Press, Washington, DC, p 1334 53. Jenkins DE, Auger EA, Matin A (1991) J Bacteriol 173:1992 54. Jenkins DE, Schultz JE, Matin A (1988) J Bacteriol 170:3910 55. Nyström T (1994) Mol Microbiol 12:833 56. Yura T, Kanemori M, Morita MT (2000) In: Storz G, Hengge-Aronis R (eds) Bacterial stress responses. ASM Press, Washington, DC, p 3 57. Straus DB, Walter WA, Gross CA (1987) Nature 329:348 58. Heitzer A, Mason CA, Snozzi M, Hamer G (1990) Arch Microbiol 155:7 59. Mason CA, Dünner J, Indra P, Colangelo T (1999) Appl Environ Microbiol 65:3433 60. Herendeen SL, VanBogelen RA, Neidhardt FC (1979) J Bacteriol 139:185 61. Nagai H, Yano R, Erickson JW, Yura T (1990) J Bacteriol 172:2710 62. Erickson JW, Vaughn V, Walter WA, Neidhardt FC, Gross CA (1987) Genes Dev 1:419 63. Wang QP, Kaguni JM (1989) J Bacteriol 171:4248 64. Erickson JW, Gross CA (1989) Genes Dev 3:1462 65. Lopez-Sanchez F, Ramirez-Santos J, Gomez-Eichelmann MC (1997) Biochim Biophys Acta 1353:79 66. Lopez-Garcia P, Forterre P (2000) Bioessays 22:738 67. Mizushima T, Ohtsuka Y, Mori H, Miki T, Sekimizu K (1996) Mol Gen Genet 253:297 68. Mizushima T, Natori S, Sekimizu K (1993) Mol Microbiol 238:1 69. Ogata Y, Mizushima T, Kataoka K, Kita K, Miki T, Sekimizu K (1996) J Biol Chem 271: 29407 70. Morita MT, Tanaka Y, Kodama TS, Kyogoku Y, Yanagi H, Yura T (1999) Genes Dev 13:655 71. Bukau B (1993) Mol Microbiol 9:671

40 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. 98. 99. 100. 101. 102. 103. 104. 105. 106. 107. 108. 109. 110. 111. 112. 113. 114. 115. 116. 117. 118. 119. 120.

L.M. Wick · T. Egli Tomoyasu T, Ogura T, Tatsuta T, Bukau B (1998) Mol Microbiol 30:567 Blaszczak A, Georgopoulos C, Liberek K (1999) Mol Microbiol 31:157 Tatsuta T, Joo DM, Calendar R, Akiyama Y, Ogura T (2000) FEBS Lett 478:271 Kanemori M, Yanagi H, Yura T (1999) J Biol Chem 274:22002 Herman C, Thevenet D, D’Ari R, Bouloc P (1995) PNAS 92:3516 Tatsuta T, Tomoyasu T, Bukau B, Kitagawa M, Mori H, Karata K, Ogura T (1998) Mol Microbiol 30:583 Tomoyasu T, Gamer J, Bukau B, Kanemori M, Mori H, Rutman AJ, Oppenheim AB,Yura T, Yamanaka K, Niki H, Hiraga S, Ogura T (1995) EMBO 14:2551 Kanemori M, Nishihara K, Yanagi H, Yura T (1997) J Bacteriol 179:7219 Dartigalongue C, Loferer H, Raina S (2001) EMBO J 20:5908 Kanehara K, Ito K, Akiyama Y (2002) Genes Dev 16:2147 Bukau B, Horwich AL (1998) Cell 92:351 Sigler PB, Xu Z, Rye HS, Burston SG, Fenton WA, Horwich AL (1998) Annu Rev Biochem 67:581 Ewalt KL, Hendrick JP, Houry WA, Hartl FU (1997) Cell 90:491 Teter SA, Houry WA, Ang D, Tradler T, Rockabrand D, Fischer G, Blum P, Georgopoulos C, Hartl FU (1999) Cell 97:755 van den Berg B, Ellis RJ, Dobson CM (1999) EMBO J 18:6927 Ellis RJ (2001) Curr Opin Struct Biol 11:114 Lund PA (2001) Adv Microb Physiol 44:93 Deuerling E, Schulze-Specking A, Tomoyasu T, Mogk A, Bukau B (1999) Nature 400:693 Patzelt H, Rudiger S, Brehmer D, Kramer G,Vorderwulbecke S, Schaffitzel E,Waitz A, Hesterkamp T, Dong L, Schneider-Mergener J, Bukau B, Deuerling E (2001) PNAS 98:14244 Hesterkamp T, Hauser S, Lutcke H, Bukau B (1996) PNAS 93:4437 Schaffitzel E, Rudiger S, Bukau B, Deuerling E (2001) Biol Chem 382:1235 Beissinger M, Buchner J (1998) Biol Chem 379:245 Rudiger S, Schneider-Mergener J, Bukau B (2001) EMBO J 20:1042 Grimshaw JPA, Jelesarov I, Schonfeld H-J, Christen P (2001) J Biol Chem 276:6098 Ellis RJ (2001) Curr Biol 11:R1038 Zolkiewski M (1999) J Biol Chem 274:28083 Goloubinoff P, Mogk A, Zvi APB, Tomoyasu T, Bukau B (1999) PNAS 96:13732 Ehrnsperger M, Graber S, Gaestel M, Buchner J (1997) EMBO J 16:221 Veinger L, Diamant S, Buchner J, Goloubinoff P (1998) J Biol Chem 273:11032 Shearstone JR, Baneyx F (1999) J Biol Chem 274:9937 Kuczynska-Wisnik D, Kcdzierska S, Matuszewska E, Lund P, Taylor A, Lipinska B, Laskowska E (2002) Microbiology 148:1757 Kitagawa M, Miyakawa M, Matsumura Y, Tsuchido T (2002) Eur J Biochem 269:2907 Jakob U, Lilie H, Meyer I, Buchner J (1995) J Biol Chem 270:7288 Jakob U, Muse W, Eser M, Bardwell JC (1999) Cell 96:341 Wickner S, Maurizi MR, Gottesman S (1999) Science 286:1888 Gottesman S, Wickner S, Maurizi MR (1997) Genes Dev 11:815 Pak M, Hoskins JR, Singh SK, Maurizi MR, Wickner S (1999) J Biol Chem 274:19316 Tomoyasu T, Mogk A, Langen H, Goloubinoff P, Bukau B (2001) Mol Microbiol 40:397 Nyström T (2001) Arch Microbiol 176:159 Ballesteros M, Fredriksson A, Henriksson J, Nystrom T (2001) EMBO J 20:5280 Dukan S, Farewell A, Ballesteros M, Taddei F, Radman M, Nystrom T (2000) PNAS 97:5746 Dukan S, Nystrom T (1998) Genes Dev 12:3431 Dukan S, Nystrom T (1999) J Biol Chem 274:26027 Schlieker C, Bukau B, Mogk A (2002) J Biotechnol 96:13 Helmann JD (2002) Adv Microb Physiol 46:47 Missiakas D, Raina S (1998) Mol Microbiol 28:1059 Pogliano J, Lynch A, Belin D, Lin E, Beckwith J (1997) Genes Dev 11:1169 Spiess C, Beil A, Ehrmann M (1999) Cell 97:339 Dartigalongue C, Missiakas D, Raina S (2001) J Biol Chem 276:20866


121. 122. 123. 124. 125. 126. 127. 128. 129. 130. 131. 132. 133. 134. 135. 136. 137. 138. 139. 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151. 152. 153. 154. 155. 156. 157. 158. 159. 160. 161. 162. 163. 164. 165. 166. 167. 168. 169. 170. 171.

41

Danese P, Silhavy T (1997) Genes Dev 11:1183 Raivio TL, Silhavy TJ (2001) Annu Rev Microbiol 55:591 De Las Penas A, Connolly L, Gross C (1997) J Bacteriol 179:6862 Raina S, Missiakas D, Georgopoulos C (1995) EMBO 14:1043 Rouviere PE, De Las Penas A, Mecsas J, Lu CZ, Rudd KE, Gross CA (1995) EMBO 14: 1032 Bianchi AA, Baneyx F (1999) Mol Microbiol 34:1029 Nitta T, Nagamitsu H, Murata M, Izu H, Yamada M (2000) J Bacteriol 182:5231 Jones CH, Danese PN, Pinkner JS, Silhavy TJ, Hultgren SJ (1997) EMBO J 16:6394 Mecsas J, Rouviere PE, Erickson JW, Donohue TJ, Gross CA (1993) Genes Dev 7:2618 Missiakas D, Betton JM, Raina S (1996) Mol Microbiol 21:871 Rouviere PE, Gross CA (1996) Genes Dev 10:3170 Missiakas D, Mayer MP, Lemaire M, Georgopoulos C, Raina S (1997) Mol Microbiol 24:355 De Las Penas A, Connolly L, Gross CA (1997) Mol Microbiol 24:373 Kanehara K, Akiyama Y, Ito K (2001) Gene 281:71 Ades SE, Connolly LE, Alba BM, Gross CA (1999) Genes Dev 13:2449 Alba BM, Leeds JA, Onufryk C, Lu CZ, Gross CA (2002) Genes Dev 16:2156 Alba BM, Zhong HJ, Pelayo JC, Gross CA (2001) Mol Microbiol 40:1323 Collinet B, Yuzawa H, Chen T, Herrera C, Missiakas D (2000) J Biol Chem 275:33898 Danese P, Snyder W, Cosma C, Davis L, Silhavy T (1995) Genes Dev 9:387 Shimohata N, Chiba S, Saikawa N, Ito K, Akiyama Y (2002) Genes Cells 7:653 Dartigalongue C, Raina S (1998) EMBO J 17:3968 De Wulf P, McGuire AM, Liu X, Lin ECC (2002) J Biol Chem 277:26652 Snyder W, Davis L, Danese P, Cosma C, Silhavy T (1995) J Bacteriol 177:4216 Danese PN, Oliver GR, Barr K, Bowman GD, Rick PD, Silhavy TJ (1998) J Bacteriol 180:5875 Mileykovskaya E, Dowhan W (1997) J Bacteriol 179:1029 De Wulf P, Kwon O, Lin ECC (1999) J Bacteriol 181:6772 Danese PN, Silhavy TJ (1998) J Bacteriol 180:831 Raivio TL, Laird MW, Joly JC, Silhavy TJ (2000) Mol Microbiol 37:1186 Mourez M, Skouloubris S, Betton JM, Dassa E (1997) Mol Microbiol 26:821 Dong J, Iuchi S, Kwan HS, Lu Z, Lin EC (1993) Gene 136:227 Raivio T, Silhavy T (1997) J Bacteriol 179:7724 Weber RF, Silverman PM (1988) J Mol Biol 203:467 Raivio TL, Popkin DL, Silhavy TJ (1999) J Bacteriol 181:5263 Raffa RG, Raivio TL (2002) Mol Microbiol 45:1599 Thieringer HA, Jones PG, Inouye M (1998) Bioessays 20:49 Jones PG, Mitta M, Kim Y, Jiang W, Inouye M (1996) PNAS 93:76 Jones JG, Inouye M (1996) Mol Microbiol 21:1207 Dammel CS, Noller HF (1995) Genes Dev 9:626 Friedman D, Olson E, Georgopoulos C, Tilly K, Herskowitz I, Banuett F (1984) Microbiol Rev 48:299 Donovan W, Kushner S (1986) PNAS 83:120 Walker GC (1984) Microbiol Rev 48:60 Gualerzi CO, Pon CL (1990) Biochemistry 29:5881 Sugino A, Peebles C, Kreuzer KN, Cozzarelli NR (1977) PNAS 74:4767 Lelivelt M, Kawula T (1995) J Bacteriol 177:4900 Kandror O, Goldberg AL (1997) PNAS 94:4978 Jones P, Inouye M (1994) Mol Microbiol 11:811 La Teana A, Brandi A, Falconi M, Spurio R, Pon C (1991) PNAS 88:10907 Brandi A, Pon C, Gualerzi C (1994) Biochimie 76:1090 Goldstein E, Drlica K (1984) PNAS 81:4046 Xia B, Etchegaray J-P, Inouye M (2001) J Biol Chem 276:35581 Yamanaka K, Inouye M (2001) J Bacteriol 183:2808

42


172. Kandror O, DeLeon A, Goldberg AL (2002) PNAS 99:9727 173. Jiang W, Hou Y, Inouye M (1997) J Biol Chem 272:196 174. Phadtare S, Yamanaka K, Inouye M (2000) In: Storz G, Hengge-Aronis R (eds) Bacterial stress responses. ASM Press, Washington, DC, p 33 175. Fang L, Jiang W, Bae W, Inouye M (1997) Mol Microbiol 23:355 176. Goldstein J, Pollitt N, Inouye M (1990) PNAS 87:283 177. Goldenberg D, Azar I, Oppenheim AB, Brandi A, Pon CL, Gualerzi CO (1997) Mol Gen Genet 256:282 178. Goldenberg D, Azar I, Oppenheim AB (1996) Mol Microbiol 19:241 179. Brandi A, Spurio R, Gualerzi CO, Pon CL (1999) EMBO J 18:1653 180. Yamanaka K, Inouye M (2001) Genes Cells 6:279 181. Wang N, Yamanaka K, Inouye M (1999) J Bacteriol 181:1603 182. Jiang W, Fang L, Inouye M (1996) J Bacteriol 178:4919 183. Xia B, Ke H, Jiang W, Inouye M (2002) J Biol Chem 277:6005 184. Sprengart ML, Fuchs E, Porter AG (1996) EMBO 15:665 185. Etchegaray J-P, Inouye M (1999) J Bacteriol 181:5852 186. Etchegaray J-P, Inouye M (1999) Mol Microbiol 33:438 187. Etchegaray J-P, Inouye M (1999) J Biol Chem 274:10079 188. Mironova RS, Xu J, AbouHaidar MG, Ivanov IG (1999) Microbiol Res 154:35 189. Resch A, Tedin K, Grundling A, Mundlein A, Blasi U (1996) EMBO 15:4740 190. Blasi U, O’Connor M, Squires CL, Dahlberg AE (1999) Mol Microbiol 33:439 191. Moll I, Huber M, Grill S, Sairafi P, Mueller F, Brimacombe R, Londei P, Blasi U (2001) J Bacteriol 183:3499 192. La Teana A, Brandi A, O’Connor M, Freddi S, Pon C (2000) RNA 6:1393 193. Newkirk K, Feng W, Jiang W, Tejero R, Emerson S, Iouye M, Montelione G (1994) PNAS 91:5114 194. Bae W, Xia B, Inouye M, Severinov K (2000) PNAS 97:7784 195. Bae W, Jones P, Inouye M (1997) J Bacteriol 179:7081 196. Yamanaka K, Fang L, Inouye M (1998) Mol Microbiol 27:247 197. Lee SJ, Xie A, Jiang W, Etchegaray JP, Jones PG, Inouye M (1994) Mol Microbiol 11:833 198. Nakashima K, Kanamaru K, Mizuno T, Horikoshi K (1996) J Bacteriol 178:2994 199. Yamanaka K, Inouye M (1997) J Bacteriol 179:5126 200. Yamanaka K, Crooke E, Wang Y-H, Inouye M (2001) Mol Microbiol 39:1572 201. Yamanaka K, Mitani T, Ogura T, Niki H, Hiraga S (1994) Mol Microbiol 13:301 202. Bae W, Phadtare S, Severinov K, Inouye M (1999) Mol Microbiol 31:1429 203. Yamada M, Nagamitsu H, Izu H, Nakamura K, Talukder AA (2002) J Mol Microbiol Biotechnol 4:163 204. Etchegaray JP, Jones JG, Inouye M (1996) Genes Cells 1:171 205. Xia B, Ke H, Inouye M (2001) Mol Microbiol 40:179 206. Phadtare S, Inouye M (2001) J Bacteriol 183:1205 207. Phadtare S, Inouye M, Severinov K (2002) J Biol Chem 277:7239 208. VanBogelen RA, Neidhardt FC (1990) PNAS 87:5589 209. Jones P, Cashel M, Glaser G, Neidhardt F (1992) J Bacteriol 174:3903 210. Wang J-Y, Syvanen M (1992) Mol Microbiol 6:1861 211. Sinensky M (1974) PNAS 71:522 212. Marr AG, Ingraham JL (1962) J Bacteriol 84:1260 213. Cronan JEJ, Rock CO (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, p 612 214. Garwin JL, Klages AL, Cronan JEJ (1980) J Biol Chem 255:3263 215. Garwin JL, Cronan JEJ (1980) J Bacteriol 141:1457 216. Carty SM, Sreekumar KR, Raetz CRH (1999) J Biol Chem 274:9677 217. Vorachek-Warren MK, Carty SM, Lin S, Cotter RJ, Raetz CRH (2002) J Biol Chem 277:14186


43

218. Cashel M, Gentry DR, Hernandez VJ,Vinella D (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, p 1458 219. Cashel M, Gallant J (1969) Nature 221:838 220. Irr JD (1972) J Bacteriol 110:554 221. Lazzarini RA, Cashel M, Gallant J (1971) J Biol Chem 246:4381 222. Spira B, Silberstein N, Yagil E (1995) J Bacteriol 177:4053 223. Harshman RB, Yamazaki H (1972) Biochemistry 11:615 224. VanBogelen RA, Kelley PM, Neidhardt FC (1987) J Bacteriol 169:26 225. Andersson L, Shaojun Y, Neubauer P, Enfors S-O (1996) J Bacteriol 46:255 226. Teich A, Meyer S, Lin HY, Andersson L, Enfors S-O, Neubauer P (1999) Biotechnol Prog 15:123 227. Ryals J, Little R, Bremer H (1982) J Bacteriol 151:1261 228. Gallant J, Palmer L, Pao CC (1977) Cell 11:181 229. Pao CC, Dyess BT (1981) J Biol Chem 256:2252 230. Condon C, Squires C, Squires CL (1995) Microbiol Rev 59:623 231. Murray KD, Bremer H (1996) J Mol Biol 259:41 232. Lipmann F, Sy J (1976) Prog Nucleic Acid Res Mol Biol 17:1 233. Weyer WJ, de Boer HA, de Boer JG, Gruber M (1976) Biochim Biophys Acta 442:123 234. Pedersen FS, Kjeldgaard NO (1977) Eur J Biochem 76:91 235. Cochran JW, Byrne RW (1974) J Biol Chem 249:353 236. Somerville CR, Ahmed A (1979) Mol Gen Genet 169:315 237. Schreiber G, Metzger S, Aizenman E, Roza S, Cashel M, Glaser G (1991) J Biol Chem 266:3760 238. Metzger S, Schreiber G, Aizenman E, Cashel M, Glaser G (1989) J Biol Chem 264:21146 239. Gropp M, Strausz Y, Gross M, Glaser G (2001) J Bacteriol 183:570 240. Ramagopal S, Davis BD (1974) PNAS 71:820 241. Friesen JD, Fiil NP, Parker JM, Haseltine WA (1974) PNAS 71:3465 242. Yang X, Ishiguro EE (2001) J Bacteriol 183:6532 243. Haseltine WA, Block R (1973) PNAS 70:1564 244. Richter D (1976) PNAS 73:707 245. Metzger S, Sarubbi E, Glaser G, Cashel M (1989) J Biol Chem 264:9122 246. Laffler T, Gallant JA (1974) Cell 3:47 247. An G, Justesen J, Watson RJ, Friesen JD (1979) J Bacteriol 173:1100 248. Hernandez VJ, Bremer H (1991) J Biol Chem 266:5991 249. Xiao H, Kalman M, Ikehara K, Zemel S, Glaser G, Cashel M (1991) J Biol Chem 266:5980 250. Gentry D, Cashel M (1996) Mol Microbiol 19:1373 251. Gentry D, Cashel M (1995) J Bacteriol 177:3890 252. Richter D (1980) Mol Gen Genet 178:325 253. Yegian CD, Stent GS, Martin EM (1966) PNAS 55:839 254. Hernandez VJ, Cashel M (1995) J Mol Biol 252:536 255. Izutsu K, Wada A, Wada C (2001) Genes Cells 6:665 256. Christensen SK, Mikkelsen M, Pedersen K, Gerdes K (2001) PNAS 98:14328 257. Aizenman E, Engelberg-Kulka H, Glaser G (1996) PNAS 93:6059 258. Sat B, Hazan R, Fisher T, Khaner H, Glaser G, Engelberg-Kulka H (2001) J Bacteriol 183:2041 259. Jishage M, Kvint K, Shingler V, Nystrom T (2002) Genes Dev 16:1260 260. Rao NN, Liu S, Kornberg A (1998) J Bacteriol 180:2186 261. Kuroda A, Nomura K, Ohtomo R, Kato J, Ikeda T, Takiguchi N, Ohtake H, Kornberg A (2001) Science 293:705 262. Shiba T, Tsutsumi K,Yano H, Ihara Y, Kameda A, Tanaka K, Takahashi H, Munekata M, Rao NN, Kornberg A (1997) PNAS 94:11210 263. Chang DE, Smalley DJ, Conway T (2002) Mol Microbiol 45:289 264. Chatterji D, Fujita N, Ishihama A (1998) Genes Cells 3:279

44 265. 266. 267. 268. 269. 270. 271. 272. 273. 274. 275. 276. 277. 278. 279. 280. 281. 282. 283. 284. 285. 286. 287. 288. 289. 290. 291. 292. 293. 294. 295. 296. 297. 298. 299. 300. 301. 302. 303. 304. 305. 306. 307. 308. 309. 310. 311. 312.

L.M. Wick · T. Egli Reddy PS, Raghavan A, Chatterji D (1995) Mol Microbiol 15:255 Toulokhonov II, Shulgina I, Hernandez VJ (2001) J Biol Chem 276:1220 Zhou YN, Jin DJ (1998) PNAS 95:2908 Barker MM, Gaal T, Gourse RL (2001) J Mol Biol 305:689 Barker MM, Gaal T, Josaitis CA, Gourse RL (2001) J Mol Biol 305:673 Bartlett MS, Gaal T, Ross W, Gourse RL (1998) J Mol Biol 279:331 Choy HE (2000) J Biol Chem 275:6783 Keener J, Masayasu N (1996) In: Neidhardt FC, Curtiss R III, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology. ASM Press, Washington, DC, p 1417 Gaal T, Gourse R (1990) PNAS 87:5533 Gaal T, Bartlett MS, Ross W, Turnbough CL Jr, Gourse RL (1997) Science 278:2092 Barker MM, Gourse RL (2001) J Bacteriol 183:6315 Petersen C, Moller LB (2000) J Biol Chem 275:3931 Loewen PC, Hu B, Strutinsky J, Sparling R (1998) Can J Microbiol 44:707 Hengge-Aronis R (1996) In: Neidhardt FC, Curtiss RIII, Ingraham JL, Lin ECC, Low KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE (eds) Escherichia coli and Salmonella: cellular and molecular biology, vol 1.ASM Press,Washington, DC, p 1497 Hengge-Aronis R (2000) In: Storz G, Hengge-Aronis R (eds) Bacterial stress responses. ASM Press, Washington, DC, p 161 Eisenstark A, Calcutt MJ, Becker-Hapak M, Ivanova A (1996) Free Radic Biol Med 21:975 Hengge-Aronis R (2002) Microbiol Mol Biol Rev 66:373 Lange R, Hengge-Aronis R (1994) Mol Microbiol 13:733 Lange R, Fischer D, Hengge-Aronis R (1995) J Bacteriol 177:4676 Ichikawa JK, Li C, Fu J, Clarke S (1994) J Bacteriol 176:1630 Lange R, Hengge-Aronis R (1991) Mol Microbiol 5:49 Lange R, Hengge-Aronis R (1994) Genes Dev 8:1600 Takayanagi Y, Tanaka K, Takahashi H (1994) Mol Gen Genet 243:525 McCann MP, Fraley CD, Matin A (1993) J Bacteriol 175:2143 Zgurskaya H, Keyhan M, Matin A (1997) Mol Microbiol 24:643 Mukhopadhyay S, Audia JP, Roy RN, Schellhorn HE (2000) Mol Microbiol 37:371 Mulvey MR, Switala J, Borys A, Loewen PC (1990) J Bacteriol 172:6713 Schellhorn HE, Stones VL (1992) J Bacteriol 174:4769 Brown L, Gentry D, Elliott T, Cashel M (2002) J Bacteriol 184:4455 Hirsch M, Elliott T (2002) J Bacteriol 184:5077 Sledjeski DD, Gupta A, Gottesman S (1996) EMBO 15:3993 Brown L, Elliott T (1997) J Bacteriol 179:656 Lease RA, Belfort M (2000) PNAS 97:9919 Lease RA, Cusick ME, Belfort M (1998) PNAS 95:12456 Majdalani N, Cunning C, Sledjeski D, Elliott T, Gottesman S (1998) PNAS 95:12462 Repoila F, Gottesman S (2001) J Bacteriol 183:4012 Majdalani N, Chen S, Murrow J, St John K, Gottesman S (2001) Mol Microbiol 39:1382 Zhang A, Altuvia S, Tiwari A, Argaman L, Hengge-Aronis R, Storz G (1998) EMBO J 17:6061 Altuvia S, Weinstein-Fischer D, Zhang A, Postow L, Storz G (1997) Cell 90:43 Muffler A, Fischer D, Hengge-Aronis R (1996) Genes Dev 10:1143 Brown L, Elliott T (1996) J Bacteriol 178:3763 Sledjeski DD, Whitman C, Zhang A (2001) J Bacteriol 183:1997 Moller T, Franch T, Hojrup P, Keene DR, Bachinger HP, Brennan RG, Valentin-Hansen P (2002) Mol Cell 9:23 Zhang A, Wassarman KM, Ortega J, Steven AC, Storz G (2002) Mol Cell 9:11 Balandina A, Claret L, Hengge-Aronis R, Rouviere-Yaniv J (2001) Mol Microbiol 39:1069 Yamashino T, Ueguchi C, Mizuno T (1995) EMBO 14:594 Barth M, Marschall C, Muffler A, Fischer D, Hengge-Aronis R (1995) J Bacteriol 177:3455 Muffler A, Barth M, Marschall C, Hengge-Aronis R (1997) J Bacteriol 179:445


313. 314. 315. 316. 317. 318. 319. 320. 321. 322. 323. 324. 325. 326. 327. 328. 329. 330. 331. 332. 333. 334. 335. 336. 337. 338. 339. 340. 341. 342. 343. 344. 345. 346. 347. 348. 349. 350.

45

Bohringer J, Fischer D, Mosler G, Hengge-Aronis R (1995) J Bacteriol 177:413 Schweder T, Lee K, Lomovskaya O, Matin A (1996) J Bacteriol 178:470 Bearson S, Benjamin W Jr, Swords W, Foster J (1996) J Bacteriol 178:2572 Muffler, A, Traulsen D, Lange R, Hengge-Aronis R (1996) J Bacteriol 178:1607 Becker G, Klauck E, Hengge-Aronis R (2000) Mol Microbiol 35:657 Becker G, Klauck E, Hengge-Aronis R (1999) PNAS 96:6439 Muffler A, Fischer D, Altuvia S, Storz G, Hengge-Aronis R (1996) EMBO 15:1333 Pratt LA, Silhavy TJ (1996) PNAS 93:2488 Zhou Y, Gottesman S (1998) J Bacteriol 180:1154 Zhou Y, Gottesman S, Hoskins JR, Maurizi MR, Wickner S (2001) Genes Dev 15:627 Becker G, Hengge-Aronis R (2001) Mol Microbiol 39:1153 Rockabrand D, Livers K, Austin T, Kaiser R, Jensen D, Burgess R, Blum P (1998) J Bacteriol 180:846 Welsh DT, Herbert RA (1999) FEMS Microbiol Lett 174:57 Strom AR, Kaasen I (1993) Mol Microbiol 8:205 Hengge-Aronis R, Klein W, Lange R, Rimmele M, Boos W (1991) J Bacteriol 173:7918 Adams JL, McLean RJC (1999) Appl Environ Microbiol 65:4285 Corona-Izquierdo FP, Membrillo-Hernandez J (2002) FEMS Microbiol Lett 211:105 Prigent-Combaret C, Brombacher E, Vidal O, Ambert A, Lejeune P, Landini P, Dorel C (2001) J Bacteriol 183:7213 Beltrametti F, Kresse AU, Guzman CA (1999) J Bacteriol 181:3409 Bishop RE, Leskiw BK, Hodges RS, Kay CM, Weiner JH (1998) J Mol Biol 280:583 Tanaka K, Takayanagi Y, Fujita N, Ishihama A, Takahashi H (1993) PNAS 90:3511 Tanaka K, Kusano S, Fujita N, Ishihama A, Takahashi H (1995) Nucleic Acids Res 23:827 Nguyen LH, Jensen DB, Thompson NE, Gentry DR, Burgess RR (1993) Biochemistry 32:11112 Hengge-Aronis R (1999) Curr Opin Microbiol 2:148 Notley L, Ferenci T (1996) J Bacteriol 178:1465 Farewell A, Kvint K, Nystrom T (1998) Mol Microbiol 29:1039 Notley-McRobb L, King T, Ferenci T (2002) J Bacteriol 184:806 Ferenci T (1999) Res Microbiol 150:431 Wick LM, Weilenmann H, Egli T (2002) Microbiology 148:2889 Savage DC (1977) Ann Rev Microbiol 31:107 Savageau MA (1983) Am Nat 122:732 Finkel SE, Zinser E, Gupta S, Kolter R (1998) In: Busby SJW, Thomas MC, Brown NL (eds) Molecular microbiology. Springer, Berlin Heidelberg New York, p 3 Zambrano MM, Siegele DA, Almiron M, Tormo A, Kolter R (1993) Science 259:1757 Lin J, Smith M, Chapin K, Baik H, Bennett G, Foster J (1996) Appl Environ Microbiol 62:3094 Vulic M, Kolter R (2001) Genetics 158:519 Egli T (2000) In: Lederberg J (ed) Encyclopedia of microbiology, vol 3.Academic Press, San Diego, p 431 Rudd KE (2000) Nucl Acids Res 28:60 Ihssen J, Egli T (2004) Microbiology (in press)

Received: August 2003