Skip to main content
NCBI home page
Genes & Development logoLink to Genes & Development
. 2007 Jan 1;21(1):124–136. doi: 10.1101/gad.1496707

Design principles of the proteolytic cascade governing the σE-mediated envelope stress response in Escherichia coli: keys to graded, buffered, and rapid signal transduction

Rachna Chaba 1,6, Irina L Grigorova 2,6,7, Julia M Flynn 4,5, Tania A Baker 4,5, Carol A Gross 1,3,8
PMCID: PMC1759897  PMID: 17210793

Abstract

Proteolytic cascades often transduce signals between cellular compartments, but the features of these cascades that permit efficient conversion of a biological signal into a transcriptional output are not well elucidated. σE mediates an envelope stress response in Escherichia coli, and its activity is controlled by regulated degradation of RseA, a membrane-spanning anti-σ factor. Examination of the individual steps in this protease cascade reveals that the initial, signal-sensing cleavage step is rate-limiting; that multiple ATP-dependent proteases degrade the cytoplasmic fragment of RseA and that dissociation of σE from RseA is so slow that most free σE must be generated by the active degradation of RseA. As a consequence, the degradation rate of RseA is set by the amount of inducing signal, and insulated from the “load” on and activity of the cytoplasmic proteases. Additionally, changes in RseA degradation rate are rapidly reflected in altered σE activity. These design features are attractive as general components of signal transduction pathways governed by unstable negative regulators.

Keywords: σE, stress response, proteolysis, ATP-dependent proteases, DegS, RseP

Proteolytic cascades are widely used to transduce signals across membranes to enable cells to respond to environmental stress and coordinate processes in different cellular compartments. However, the molecular properties of these cascades that facilitate the desired outputs have rarely been examined. In this work, we examine the individual steps in the protease cascade governing the activity of the σE-mediated envelope stress response in Escherichia coli to determine the construction features of this cascade that facilitate faithful transmission of signal and a rapid output.

σE directs RNA polymerase to transcribe genes encoding proteins that ensure the synthesis, assembly, and homeostasis of outer membrane porins and lipopolysaccharide, the two major components of the unique outer membrane of Gram-negative bacteria (Dartigalongue et al. 2001; Rezuchova et al. 2003; Rhodius et al. 2006). Envelope integrity is required under all growth conditions, and σE is an essential transcription factor (De Las Penas et al. 1997a). Perturbations in the integrity and protein-folding state of the envelope caused by temperature upshift, chaperone depletion, or accumulation of unassembled porins increase σE activity; conversely, temperature downshift and/or depletion of porins decrease σE activity (Mecsas et al. 1993; Hiratsu et al. 1995; Raina et al. 1995; Rouviere et al. 1995; Missiakas et al. 1996; Rouviere and Gross 1996; Ades et al. 2003).

The components of the signal transduction system that control σE activity are shown in Figure 1. RseA, a membrane-spanning anti-σ factor, inhibits σE activity. The cytoplasmic domain of RseA (RseA1–108) binds to σE and its periplasmic domain binds to RseB (De Las Penas et al. 1997b; Missiakas et al. 1997). Interaction of σE with RseA prevents σE from binding to RNA polymerase (Campbell et al. 2003). This inhibitory interaction is relieved when RseA is degraded via a proteolytic cascade. RseA is cleaved first in its periplasmic domain by DegS and then within its transmembrane region by the RIP protease, RseP, thereby releasing RseA1–108 still bound to σE (Ades et al. 1999; Alba et al. 2002; Kanehara et al. 2002; Walsh et al. 2003; Akiyama et al. 2004). σE is finally released from this inhibitory interaction when RseA1–108 is degraded by cytoplasmic proteases (Flynn et al. 2004).

Figure 1.

Figure 1.

Open in a new tab

A protease cascade degrades RseA to release σE.

The proteolytic cascade degrading RseA is activated by unassembled porins in the envelope compartment of the cell. Porins constitute 2% of all cellular proteins and form trimeric pores in the outer membrane (Nikaido 1996). When folding is inefficient, unassembled porins accumulate and their C termini activate DegS to initiate cleavage of RseA (Walsh et al. 2003). Exposed C termini are an excellent indicator of impaired folding as they are normally sequestered in the porin trimer interface (Cowan et al. 1995). Free DegS is inactive in cleaving RseA because its catalytic triad is inappropriately positioned. However, when porin C termini bind to the DegS PDZ domain, the catalytic triad is repositioned and cleavage of RseA ensues (Walsh et al. 2003; Wilken et al. 2004). DegS, RseB, RseA, and the PDZ domain of RseP itself all inhibit RseP cleavage of intact RseA, thereby ensuring that DegS always acts before RseP in the RseA-processing pathway (Kanehara et al. 2003; Bohn et al. 2004; Grigorova et al. 2004). Since DegS is the only protease that senses the porin signal, obligate sequential cleavage of RseA ensures that the initiation of the proteolytic cascade reflects the status of this inducing signal.

In the present study we dissect the individual steps in the proteolytic cascade degrading RseA. A kinetic description of the pathway reveals that the initial signal-sensing cleavage step is rate-limiting under all conditions, ensuring that the overall degradation rate of RseA reflects signal status. We find that RseA and σE interact with picomolar affinity, ensuring that free σE is generated almost exclusively by cytoplasmic ATP-dependent proteases that disassemble the complex and degrade RseA1–108. Indeed, enormous resources are invested into clearing RseA1–108 from the cell: Essentially, the entire network of cytoplasmic ATP-dependent proteases participates in this process. We will discuss why the organizational structure of this protease cascade facilitates rapid and sensitive transmission of signal to the output response. Proteolytic cascades govern a variety of processes in diverse organisms (Brown et al. 2000; Weihofen and Martoglio 2003). We believe that the investigation and subsequent cross-comparison of such pathways will significantly improve our understanding of the principles that underlie these signaling proteolytic cascades.

Results

A kinetic description of the degradation of RseA under inducing conditions

Initiation of RseA degradation by DegS is relatively slow during steady-state growth (T1/2 ∼ 8 min) (Ades et al. 2003), and is rate-limiting under these noninducing conditions, as no intermediates in degradation were observed (Ades et al. 1999; Alba et al. 2002; Kanehara et al. 2002; Akiyama et al. 2004). Here, we determined whether initiation of RseA degradation is rate-limiting when stress is generated by overexpressing the YYF peptide (Walsh et al. 2003). We measured the T1/2 of all three RseA species (full-length RseA, DegS cleavage product, and RseP cleavage product) using the pulse-chase protocol described in Ades et al. (2003) that is known not to disrupt the physiological state of the cells (see Materials and Methods).

A typical gel showing pulse chase data for degradation of RseA under inducing conditions is shown in Figure 2A. Quantification of the disappearance of full-length RseA (Fig. 1, Step 1) reveals that cleavage by DegS is initiated with a T1/2 of 1 min (Fig. 2B). An intermediate-sized band (termed RseA-I) (Fig. 2A), corresponds to the initial DegS cleavage product: It is identical in size to the RseA product generated in a strain unable to process DegS cleaved RseA further because it lacks RseP (data not shown). The level of RseA-I reflects both its rate of generation by DegS cleavage and its rate of disappearance as a consequence of cleavage by RseP (Fig. 1, Step 2). Knowing the rate of DegS cleavage, we were able to calculate that this intermediate has a T1/2 ≤ 20 sec, which reflects the rate of RseP cleavage (see Materials and Methods). Although a faint band corresponding to the cytoplasmic fragment of RseA (RseA1–108 generated by RseP cleavage) is visible on the gel, it was too weak to be quantified. Therefore, we isolated this step from the circuitry of the σE signal transduction pathway and then measured the degradation rate of RseA1–108 when in complex with σE (Fig. 1, Step 3). To do this, adjacent plasmid-located genes encoding RseA1–108 and σE were expressed from a modified (trc) promoter with low activity (see Materials and Methods). We estimate that each protein was overexpressed ∼10-fold (R. Chaba, unpubl.; see Materials and Methods). Under these conditions, the steady-state level of RseA1–108 is barely visible on a Western blot; however, disappearance of this protein could be reproducibly quantified using a pulse-chase protocol. These experiments revealed that the rate of degradation of RseA1–108 was very rapid (T1/2 ≤ 20 sec, Fig. 2C; Fig. 1, Step 3). Thus, we conclude that cleavage by DegS is the slowest step in the RseA degradation pathway under inducing conditions.

Figure 2.

Figure 2.

Open in a new tab

Cleavage of RseA by DegS is rate-limiting under inducing conditions. (A) Degradation of RseA in wild-type cells after initiation of the stress response by overexpression of YYF peptide. CAG53077 was grown in supplemented M9 media at 30°C to O.D.450 ∼ 0.15, and the stress response was initiated by the addition of IPTG to induce expression of YYF peptide. After 10 min of induction, the cells were pulse-labeled with L-[35S] methionine, followed by a chase of unlabeled methionine. At various time points after the chase, cells were collected and processed as described in Materials and Methods. Bands corresponding to the full-length RseA (RseA), RseA intermediate generated by cleavage of RseA by DegS (RseA-I), cytoplasmic fragment of RseA generated upon cleavage by RseP (RseA1–108), and the standard (RseA*; see Materials and Methods) are indicated by arrows. (B) Cleavage of full-length RseA by DegS. The rate of cleavage of full-length RseA by DegS under inducing conditions was determined by measuring the T1/2 of RseA in wild-type cells (CAG53077) upon overexpression of YYF peptide as described in Materials and Methods. A representative data set is shown. (C) The stability of RseA1–108 in wild-type cells. CAG53209 was grown in supplemented M9 media at 30°C to O.D.450 ∼ 0.2, and induced with IPTG for 10 min to overexpress RseA1–108/σE. The stability of RseA1–108 was determined as described in Materials and Methods. A representative data set is shown.

The cytoplasmic domain of RseA binds very tightly to σE

In principle, σE could be released from RseA1–108 both by dissociation and by degradation. To investigate the relative contributions of these two processes, we characterized the stability of the complex using RseA1–100, a slightly truncated version of the normal cytoplasmic fragment (RseA1–108). This fragment is easy to purify because it is stable in E. coli. It is also sufficient for anti-σ activity both in vivo and in vitro (De Las Penas et al. 1997b; Missiakas et al. 1997; Campbell et al. 2003). We used fluorescence anisotropy to measure the dissociation of rhodamine-labeled RseA1–100 from σE. To permit rhodamine labeling, residue E28 was mutated to cysteine, chosen because it was far from the RseA–σE interface. The anisotropy of rhodamine-labeled RseA1–100E28C was considerably lower than that of the labeled RseA1–100E28C/σE complex (Fig. 3A). Therefore dissociation of labeled RseA1–100E28C from σE in the presence of excess unlabeled RseA can be measured by determining the kinetics of the decrease in anisotropy. These experiments revealed that in the presence of a 10-fold excess of unlabeled RseA1–100 there was no decrease in anisotropy during a 2-h time course, indicating that the dissociation rate constant of RseA1–100E28C from σE was much less than 10−4 sec−1 (Fig. 3B). This very slow dissociation rate is not an artifact of the E28C substitution and/or rhodamine labeling since equivalent results were obtained from the reverse experiment: A 10-fold excess of rhodamine-labeled RseA1–100E28C did not compete for unlabeled RseA1–100 bound to authentic σE during 2 h of incubation (data not shown). These experiments indicate that the RseA1–100/σE complex dissociates very slowly in vitro.

Figure 3.

Figure 3.

Open in a new tab

Kinetic measurements of the dissociation and association rate constants of σE binding to RseA1–100. (A) Fluorescent anisotropy of rhodamine-labeled his6-RseA1–100E28C (his6-RseA1–100E28C*) alone and in complex with his6-σE. (B) Changes in fluorescent anisotropy of the complex between his6-RseA1–100E28C* and his6-σE (100 nM) after the addition of excess unlabeled his6-RseA1–100 (1.5 μM). (C) Changes in fluorescence intensity of his6-RseA1–100E28C*, induced by binding of his6-σE. his6-RseA1–100E28C* (128 nM) was rapidly mixed with various concentrations of his6-σE (125, 240, and 370 nM), using a stopped flow technique.

To measure the association rate constant between rhodamine-labeled RseA1–100E28C and σE, we followed changes in fluorescence intensity after mixing reactants in a stopped-flow apparatus (Fig. 3C). The determined value of 1.5 ± 0.2 × 107 M−1 sec−1 is close to that expected for a diffusion-limited reaction. This rate constant is neither a consequence of the E28C substitution nor of the labeling: Binding of both unlabeled RseA1–100 and labeled RseA1–100E28C to σE was confirmed to be the same by measuring changes in anisotropy induced by addition of σE to labeled RseA1–100E28C premixed at different concentrations with unlabeled RseA1–100 (data not shown). Based on the measured values of the association and dissociation rate constants, the equilibrium dissociation constant (Kd) between RseA1–100 and σE in vitro is very tight (<10 pM). Taken together, the rapid degradation of RseA1–108 (Fig. 2C) compared with the very slow spontaneous dissociation of σE from RseA1–100 strongly suggests that free σE is predominantly generated by the ATP-dependent, enzyme-catalyzed process of RseA degradation in vivo.

Multiple ATP-dependent proteases degrade RseA1–108 in vitro

The ATP-dependent unfoldase/protease ClpXP is competent both to disassemble the RseA1–108/σE complex and to degrade RseA1–108; RseA1–108 alone and in complex with σE were degraded at approximately equivalent rates (Flynn et al. 2004). We determined whether any other ATP-dependent proteases could perform these reactions in vitro. Except for HslUV, all of the ATP-dependent proteases degraded free RseA1–108 with the following efficiencies: ClpAP ∼ ClpXP > Lon >> FtsH under standard in vitro conditions (Fig. 4). However, whereas ClpXP degrades RseA1–108 in the RseA1–108/σE complex as efficiently as free RseA1–108, ClpAP and Lon degrade complexed RseA1–108 more slowly than free RseA1–108 and FtsH hardly degrades complexed RseA1–108 at all. These results are in accord with the lower unfoldase activity of Lon (A. Matouschek, pers. comm.) and FtsH (Herman et al. 2003). We conclude that four of the five ATP-dependent proteases degrade RseA1–108 in vitro, and three are also capable of catalyzing the disassembly of the RseA1–108/σE complex relatively efficiently.

Figure 4.

Figure 4.

Open in a new tab

Multiple ATP-dependent proteases degrade RseA1–108 in vitro. Degradation of RseA1–108 alone or RseA1–108 in complex with σE (RseA1–108/σE) by ATP-dependent proteases was carried out as described in Materials and Methods. Samples were removed at the defined time intervals, separated by SDS-PAGE, and stained with Sypro Ruby protein gel stain. Bands were visualized by a FluorImager.

Multiple ATP-dependent proteases participate in RseA1–108 degradation in vivo

To identify the ATP-dependent proteases active against RseA1–108 in vivo, we determined whether singly deficient protease strains were defective in degradation of RseA1–108. Only disruption of clpX or clpP decreased the rate of degradation of RseA1–108 (Fig. 5A; data not shown), validating previous data that indicated ClpXP as the major protease degrading RseA1–108 (Flynn et al. 2004). However, as RseA1–108 was degraded very rapidly in the absence of ClpXP (T1/2 ∼ 1.6 min) (Fig. 5A), additional proteases must be active in RseA1–108 degradation in vivo.

Figure 5.

Figure 5.

Open in a new tab

Multiple proteases degrade RseA1–108 in vivo. (A) Multiple ATP-dependent proteases degrade RseA1–108 in vivo. The stability of RseA1–108 was measured upon overexpression of RseA1–108/σE in wild-type (WT, CAG53209), ΔclpX (CAG53193), ΔclpP (CAG53194), ΔclpXΔlon (CAG53195), ΔclpPΔclpXΔlon (CAG53217), ΔclpPΔclpXΔlonΔhslUV (CAG53220), and ΔclpPΔclpXΔlonΔhslUVΔclpA (CAG53228) as described in the legend for Figure 2C. Data shown are the average of at least two independent experiments. (B) RseP is involved in degrading RseA1–108 in vivo. The stability of RseA1–108 was measured upon overexpression of RseA1–108/σE in ΔclpXΔrseP (CAG53206) and ΔclpXΔlonΔrseP (CAG53215), and their isogenic strains with intact rseP (CAG53193 and CAG53195) as described in the legend for Figure 2C. Note that the σE supplied on the plasmid allows us to delete rseP, which is essential in wild-type strains. Data shown are the average of at least two independent experiments.

We tested strains lacking multiple ATP-dependent proteases to identify the additional proteases participating in RseA1–108 degradation (Fig. 5A; data not shown). The rate of degradation of RseA1–108 in a ΔclpAΔclpX strain was the same as that in a ΔclpX strain, even though formation of both the ClpAP and ClpXP proteases are prevented in this strain (data not shown). This result is consistent with the similar stabilization of RseA1–108 in a ΔclpX and a ΔclpP strain (Fig. 5A) and suggests that ClpAP does not play a major role in degrading RseA1–108 in vivo even though ClpAP was almost as efficient as ClpXP in degrading RseA1–108 in vitro. In contrast, a ΔclpXΔlon strain showed a reproducibly slower rate of degradation of RseA1–108 (T1/2 = 2.2 min) than the ΔclpX strain, suggesting that Lon is second to ClpXP in degrading RseA1–108. Although this could be an indirect effect, the fact that Lon degrades RseA1–108 in vitro suggests that the slower rate of degradation is likely to result from eliminating Lon-mediated turnover of RseA. A ΔclpPΔclpXΔlon strain showed a further and larger decrease in degradation of RseA1–108 (T1/2 = 3.5 min). One interpretation of these results is that when Lon is missing, ClpAP degrades RseA1–108; removal of ClpP would prevent this latter degradation. Degradation by ClpAP is consistent with its rapid degradation of RseA1–108 in vitro. Interestingly, a ΔclpPΔclpXΔlonΔhslUV strain further decreases degradation of RseA1–108 (T1/2 = 4.0 min) even though HslUV did not degrade RseA1–108 in vitro (see below). A ΔclpPΔclpXΔlonΔhslUVΔclpA strain degrades RseA1–108 somewhat more slowly (T1/2 = 5.5 min) (Fig. 5A), possibly because ClpA unfoldase function is removed. With the assistance of cytoplasmic protein unfoldases, FtsH may carry out this residual degradation, a proposition we did not test because ftsH is essential. In summary, successive removal of each ATP-dependent protease decreased the rate of degradation of RseA1–108 in small incremental steps. Some of these decreases could be an indirect reflection of increased protease load on the remaining proteases. However, the fact that four of the five ATP-dependent proteases degrade RseA1–108 in vitro suggests that the effect is direct. Taken together, these data make the key point that RseA1–108 has evolved to be a substrate for many proteolytic machines, and as a consequence, even when the cytoplasmic proteolytic machinery is significantly disabled, RseA1–108 is still degraded very rapidly.

The role of ATP-independent proteases in RseA1–108 degradation

Given the requirement for disassembly, we have thus far solely considered the role of ATP-dependent proteases in degradation of RseA1–108. However, the crystal structure of RseA1–100/σE indicates that only the first 66 amino acids of RseA interact with σE; the remainder is predicted to be unstructured and has been found to have a binding site for the ClpX adaptor protein, SspB (Campbell et al. 2003; Flynn et al. 2004; Levchenko et al. 2005). This large unstructured tail might be a substrate for cleavage by ATP-independent proteases. The preferred cleavage site for RseP follows Ala 108; however, RseP might cleave other portions within the unstructured region at lower efficiency. We therefore tested whether removal of RseP, in addition to some ATP-dependent proteases, altered degradation of RseA1–108 (Fig. 5B). Indeed, a ΔclpXΔrseP strain degraded RseA1–108 more slowly than a ΔclpX strain (T1/2 = 2.0 min vs. T1/2 = 1.6 min); likewise a ΔclpXΔlonΔrseP strain exhibited more stabilization of RseA1–108 than the isogenic strain with intact rseP (T1/2 = 3.5 min vs. T1/2 = 2.2 min). Cleavage by RseP in the unstructured tail might generate a new cytoplasmic variant of RseA that is a substrate for additional proteases, such as HslUV, thereby rationalizing degradation of RseA1–108 by HslUV in vivo, but not in vitro.

The activity of σE is relatively insensitive to the flux of substrates through ClpXP

The substrate load for the cytoplasmic proteases undoubtedly varies with growth conditions. We asked whether σE activity was sensitive to substrate load on ClpXP, the major protease degrading RseA1–108. Major substrates of ClpXP are σS and ssrA-tagged proteins (Schweder et al. 1996; Gottesman et al. 1998). We tested whether decreasing the flux of these substrate proteins through ClpXP increases the rate of degradation of RseA1–108. As endogenous RseA1–108 is not observable on Western blots in wild-type cells due to its rapid degradation, we overproduced RseA1–108 and compared its accumulation in wild-type and mutant strains. Both a ΔssrA strain (which lacks ssrA-tagged substrates) and a ΔrssB strain (which lacks the protein that targets σS to ClpX) accumulated less RseA1–108 than the wild-type strain (Fig. 6A), indicating that degradation of RseA1–108 increased when competing substrates are eliminated.

Figure 6.

Figure 6.

Open in a new tab

σE activity is relatively insensitive to the flux of substrates through ClpXP (A,B) and fluctuations in the level of ClpX (C). (A) In order to compare the rate of degradation of RseA1–108 in wild-type and mutants, RseA1–108 was overexpressed in ΔssrA (CAG53162) and ΔrssB (CAG53144), and their isogenic wild-type strains (CAG53153 and CAG53092). Samples were separated by SDS-PAGE and probed with antibodies against the cytoplasmic domain of RseA. (B) Relative σE activity of wild-type (WT, CAG43450), ΔssrA (CAG53073), and ΔrssB (CAG53072) strains. Cells were grown in LB at 30°C to O.D.600 of 0.1–0.15 and expression of YYF peptide was induced with IPTG. β-Galactosidase activities were measured as described in Materials and Methods. Data shown are the average of three independent experiments. (C) Relative σE activity of wild-type (WT, CAG53077) and ΔclpX (CAG53078) strains was determined as described for B. (Inset) The rate of cleavage of full-length RseA by DegS in wild-type (WT, CAG53077) and ΔclpX (CAG53078) strains upon induction of YYF peptide was determined by measuring the stability of RseA as described in Materials and Methods. A representative data set is shown.

We next asked whether increased degradation of RseA1–108 observed in ΔssrA and ΔrssB strains resulted in altered σE activity. Note that in these experiments, strains do not contain the RseA1–108 overexpression construct: All forms of RseA are generated exclusively by the endogenous protease cascade. We find that with or without induction of the YYF peptide, ΔssrA (Fig. 6B, top) and ΔrssB (Fig. 6B, bottom) have σE activity equivalent to that of the wild-type strain even though RseA1–108 was degraded more rapidly in the mutant strain backgrounds (Fig. 6A). This result is in agreement with our finding that the initial cleavage step is the slowest in the pathway and shows that this kinetic arrangement makes σE activity insensitive to perturbations that result in increased degradation of RseA1–108.

The activity of σE is relatively insensitive to fluctuations in the level of ClpX

As ClpXP is the major cytoplasmic protease that degrades RseA1–108 to liberate σE, we tested the effect of removing ClpX on σE activity, which stabilizes RseA1–108 approximately fivefold (T1/2 = 1.6 min in a ΔclpX strain vs. 20 sec in wild type) (Fig. 5A). Under noninducing conditions, the σE activity of a ΔclpX strain was comparable to that of the wild-type (Fig. 6C), indicating that the increased stability of RseA1–108 does not affect transcriptional activity. This is expected, because initial cleavage by DegS is much slower than the rate of final degradation of RseA1–108 under noninducing conditions. Under inducing conditions, the σE activity of a ΔclpX strain is induced ∼1.5-fold to twofold less than that of the wild-type strain (Fig. 6C; Flynn et al. 2004). Two factors contribute to lower activity in the ΔclpX strain. First, initial DegS cleavage is slower than in the wild-type strain (Fig. 6C, inset: T1/2 = 1.1 min for wild type vs. 1.9 min for ΔclpX), possibly because physiological changes in ΔclpX cells might indirectly alter membrane proteolysis. Second, degradation of RseA1–108 in ΔclpX cells (T1/2 = 1.6 min) is now comparable to the rate of RseA cleavage by DegS (T1/2 = 1.9 min), so that DegS cleavage and degradation by cytoplasmic proteases become colimiting. The net result is a somewhat slower generation of free σE. We note that the magnitude of the defect in σE induction is significantly less than the magnitude of the proteolytic defect. Thus, the organization of the proteolytic cascade decreases the sensitivity of σE activity to alterations in the proteolytic environment of the cell that diminish degradation of RseA1–108.

Discussion

The goal of this study was to investigate the design principles that permit a protease cascade to convert a signal generated in the envelope to an immediate change in the activity of a transcription factor. To this end, we examined individual steps in the proteolytic pathway that controls the activity of the essential transcription factor σE by regulating the degradation of the anti-σ factor, RseA. Our major findings are that the signal-sensing cleavage step initiating the pathway is rate-limiting under all conditions, that this proteolysis pathway is robust to changes in the activity of cytoplasmic proteases, and that free σE is released almost exclusively by proteolysis. Below we discuss the construction principles that underlie each of these characteristics and their biological significance.

Organization of the proteolytic cascade

The proteolytic cascade consists of three sequential steps: a ligand-activated, initial cleavage step carried out by DegS, in which RseA is cleaved in the periplasmic domain; a highly regulated (possibly conformationally controlled) second step carried out by the RIP protease, RseP, which releases the RseA1–108/σE complex from the membrane; and a final disassembly-degradation step, carried out by multiple ATP-dependent proteases within the cytoplasmic compartment of the cell (Figs. 1, 4, 5A; Ades et al. 1999; Alba et al. 2002; Kanehara et al. 2002; Walsh et al. 2003; Akiyama et al. 2004; Flynn et al. 2004). This last step releases free σE, which binds to RNA polymerase and promotes transcription of the σE-regulon to combat envelope stress.

This work demonstrates that even under inducing conditions, the signal-sensitive DegS cleavage step is the slowest in the proteolytic cascade (T1/2 = 1 min); the two subsequent steps are each at least threefold faster than the initiating cleavage step (Fig. 2). This arrangement ensures that the overall rate of degradation of the protein reflects the status of the signal up to DegS cleavage rates as rapid as T1/2 ∼ 30 sec.

DegS-independent cleavage of RseA by RseP (a RIP protease) is highly controlled; whereas full-length RseA is cleaved very slowly by RseP, DegS-processed RseA is cleaved rapidly (Alba et al. 2002; Kanehara et al. 2002). The mechanistic details underlying this are unknown; however DegS, RseB, the periplasmic domain of RseA, and the PDZ domain of RseP participate in at least two independent mechanisms to repress activity of this protease toward full-length RseA (Kanehara et al. 2003; Bohn et al. 2004; Grigorova et al. 2004). We have suggested that RseP cleavage of full-length RseA is blocked so that DegS cleavage of RseA will be rate-limiting over a wide range of OMP signal. Interestingly, other RIP proteases that are widely used to transmit signals via a protease cascade between the compartments of a bacterial cell appear to occupy an intermediate position in every cascade: They transmit the signal rather than initiate the cascade and their activity is inhibited until the initial cleavage event is completed (Schobel et al. 2004; Chen et al. 2005; Matson and DiRita 2005). Cells use widely different proteins to accomplish inhibition; it remains an open question as to whether these inhibitory events are mechanistically similar or distinct (Resnekov and Losick 1998; Dong and Cutting 2003; Zhou and Kroos 2004, 2005). It will be interesting to examine how quantitative relationships between initial and RIP-mediated cleavages further the output goals of several different systems.

Free σE is released only after degradation of RseA1–108. This point in the pathway could provide an opportunity for a second regulatory step: RseA1–108 might be treated as a dead-end product to be eliminated at all costs; alternatively, its degradation (and σE activity) might be modulated in concert with other proteolytic events keyed to the stress status of the cell; for example, the rate of degradation of σS. To examine this issue, we performed a careful analysis of the proteases degrading RseA1–108 and investigated the cellular consequences of increasing or decreasing the rate of degradation of RseA1–108. These studies together with other data provide a clear answer: The cell uses multiple mechanisms to ensure that RseA1–108 is degraded regardless of the proteolytic status of the cell. First, RseA1–108 is the most rapidly degraded substrate of its major protease, ClpXP (T1/2 ≤ 20 sec) (Fig. 2C). Thus, DegS cleavage will remain rate-limiting even if competing substrates slightly decrease its rate of degradation by ClpXP. Second, in the unlikely event that ClpXP is overwhelmed by other substrates, RseA1–108 is degraded rapidly by other proteases (T1/2 = 1.6 min), leaving DegS cleavage of RseA rate-limiting under a broad range of conditions and thereby ensuring minimal effects on σE activity (Fig. 6C). Finally, the two major proteases that degrade RseA1–108, ClpXP and Lon, are both in the σE-regulon (Rhodius et al. 2006). Hence, under conditions inducing σE, the amount of these proteases will increase to further ensure that DegS cleavage remains rate-limiting. It will be interesting to determine whether uncoupling of degradation rate from the proteolytic status of cytoplasmic proteases is a general characteristic of systems regulated by unstable anti-σs.

RseA1–108 is the first protein identified to be a substrate of all of the ATP-dependent proteases. This allowed us to examine the interconnections between these proteases. Our most important finding was that ClpAP degrades RseA1–108 very poorly in vivo even though it degrades both free and σE-bound RseA1–108 almost as well as ClpXP in vitro. The low activity of ClpAP in vivo cannot be explained by ClpS or SspB inhibition of ClpAP (Flynn et al. 2001; Dougan et al. 2002; Farrell et al. 2005) as ClpAP still has very little activity against RseA1–108 in ΔclpS, ΔsspB and ΔclpSΔsspB strains (R. Chaba, unpubl.; data not shown). Possibly, a currently unidentified inhibitor may prevent ClpAP from degrading RseA1–108 in vivo. Additionally, our data raises the possibility that HslUV degrades RseA1–108 in vivo, but not in vitro. The simplest explanation for this discrepancy is that ATP-independent protease(s) expose an HslUV-degradation tag; more complex explanations, such as an unidentified HslUV-adaptor protein needed for substrate recognition, can be envisioned. Our study suggests that RseP might be a candidate protease to expose an HslUV recognition signal. Previous studies indicated that RseP cleaves RseA only at a single position, both in vivo and in vitro, which resulted in generation of RseA1–108. However, these studies were performed on an artificial substrate, with both cytoplasmic and periplasmic domains replaced by entities more amenable to analysis (Akiyama et al. 2004). RseA1–108 has a long extended tail; it is possible that RseP cleaves at additional positions within this tail or at other unstructured regions of the protein under conditions where degradation of RseA1–108 is delayed (Campbell et al. 2003). We plan to test this idea as it would be the first indication that a RIP protease can act on the cytoplasmic portion of a protein.

Transducing the degradation signal to a change in activity of σE

σE and RseA1–100 bind extremely tightly; their picomolar affinity is comparable to that of an antigen–antibody interaction. This exceptionally tight binding is explained by the extensive interface between the two proteins: RseA1–100 is sandwiched between the two domains of σE, making extensive interactions with each domain, and burying a total surface area of 3805 Å2 (Campbell et al. 2003). To dissociate spontaneously, σE would need to break all of these interactions, a situation that rarely occurs. Instead, σE is released by an extremely efficient disassembly-degradation reaction catalyzed by the cytoplasmic ATP-dependent proteases, which occurs with a T1/2 of <20 sec.

Why does the cell use the energy intensive process of disassembly and degradation of RseA1–108 to set the rate of σE release directly rather than relying solely on indirectly regulating free σE by changing the relative levels of RseA and σE? This latter scenario avoids the high-energy cost of dissociating the two proteins because it does not require very tight binding between σE and RseA. Modeling kinetics of down-regulation of σE activity in each scenario, as demonstrated in Figure 7, provides a possible answer. σE down-regulation results from a decrease in the DegS activating signal, which translates into decreased DegS cleavage of RseA, such as occurs after temperature downshift (Ades et al. 2003). In a dissociation-dependent scenario, free σE (and thus σE activity) is set by the relative levels of RseA and σE. RseA levels rise slowly in response to its decreased degradation (Fig. 7A, top); free σE drops in concert with the rise in RseA level (Fig. 7A, bottom), resulting in a slow decrease in σE activity. In contrast, when free σE is generated solely by proteolysis, RseA stabilization results in a rapid decrease in free σE (Fig. 7B, bottom) prior to any change in the relative levels of RseA and σE (Fig. 7B, top) provided that RseA is at least stoichiometric with σE. Intuitively, the rapid drop in free σE is explained by the fact that the rate of σE release would slow immediately upon a decrease in RseA degradation, while any preexisting σE rapidly reassociates with RseA (Fig. 3C). Published data and our unpublished results suggest that RseA is in excess over σE under basal conditions (Ades et al. 2003) and the high rate of synthesis of RseA relative to σE under inducing conditions (Ades et al. 2003) suggests that this scenario is possible under inducing conditions as well. Results from this scenario fit well with experimental determinations, indicating that σE exhibits a 10-fold decrease in activity within a few minutes after temperature downshift, at a time when the RseA∷σE ratio is virtually unchanged from its value at high temperature (Ades et al. 2003). This situation obtains because RseA stabilization is compensated by decreased synthesis from its σE-dependent promoter. Likewise, in accordance with the proteolysis-controlled model, upon temperature upshift, increase in σE activity occurs prior to change in the RseA∷σE ratio (I. Grigorova, unpubl.; data not shown). Only extreme inducing signals, such as continuous overexpression of outer membrane porins, lead to a significant change in the RseA∷σE ratio, possibly because degradation overwhelms synthesis (Ades et al. 1999). In conclusion, physiological inducing signals that trigger DegS-mediated cleavage of RseA, work by increasing the dynamics of RseA production and degradation and therefore the dynamics of binding and release of σE, rather than by rapidly changing the level of RseA.

Figure 7.

Figure 7.

Open in a new tab

Two scenarios by which stabilization of RseA could decrease σE activity. (A) An increase in the level of RseA relative to that of σE decreases σE activity. (Top) Analytical calculation of the change in RseA level resulting from changing its rapid degradation at 43°C (T1/2 = 2 min) to its slow degradation at 30°C (T1/2 > 50 min) (Ades et al. 2003), reveals that the characteristic time of RseA increase upon temperature downshift is on the order of tens of minutes. (Bottom) Thus, a decrease in the level of free σE (σE not bound to RseA), which determines σE activity, occurs very slowly in concert with the rise in RseA. (B) A decrease in the rate of σE release from RseA decreases σE activity. This scenario requires that the degradation rate of RseA >> the dissociation rate of RseA from σE. Computer simulation of a quantitative model of our system using parameters determined experimentally in this work and Ades et al. (2003) (RseA degradation, σE and RseA synthesis, dilution by cell growth, and binding) (data not shown) demonstrated that decreased release of σE into the cytoplasm as a consequence of decreased initiation of RseA degradation by DegS would result in rapid down-regulation of free σE (bottom) on a time scale much faster than changes in the relative levels of the two proteins (top). This scenario does not require changes in the relative levels of σE and RseA.

The tight coupling of proteolysis and activity found in the RseA/σE pathway might be important for other Group 4 (ECF) σs in addition to σE. Group 4 σs comprise the majority of the alternate σs, and they control a wide range of responses to environmental stress in diverse bacteria. A predominant paradigm is these ECF σs positively control their own transcription and are located in operons together with their anti-σ regulators (Missiakas and Raina 1998; Raivio and Silhavy 2001; Bashyam and Hasnain 2004). These anti-σs are often regulated by proteolysis and show increased degradation when stress increases and decreased degradation when stress decreases (Browning et al. 2003; Schobel et al. 2004; Reiling et al. 2005). We suspect that the ability to immediately down-regulate activity in response to decreased proteolysis will be an important design principle for the proteolytic cascades governing the activity of many Group 4 σs and possibly in other systems governed by unstable negative regulators.

Materials and methods

Media, antibiotics, strains, and plasmids

Luria-Bertani (LB) and M9 minimal medium were prepared as described (Sambrook et al. 1989). M9 was supplemented with 0.2% glucose, 1 mM MgSO4, 2 μg/mL thiamine, and all amino acids (40 μg/mL), except methionine. When required, the medium was supplemented with 100 μg/mL ampicillin (Ap), 20 μg/mL chloramphenicol (Cm), 30 μg/mL kanamycin (Kan), 50 μg/mL spectinomycin (Spec), and/or 10 μg/mL tetracycline (Tet). Expression of proteins from T7, Ptrc, or mutated Ptrc promoters was induced by 1 mM isopropyl-β-D-thiogalactopyranoside (IPTG).

Strains and plasmids are listed in Table 1. Most experiments used MG1655; comparable results were obtained with MC1061. Details of plasmid construction are available on request.

Table 1.

Strains and plasmids used in this study

graphic file with name 124tbl1.jpg

Open in a new tab

Buffers

Buffer I comprises 10 mM Tris-HCl (pH 9.0), 5% glycerol, 0.1 mM aminoethylbenzenesulfonic acid (AEBSF), and 0.5 mM β-mercaptoethanol (BME).

PD buffer contains 25 mM HEPES-KOH (pH 7.5), 5 mM KCl, 5 mM MgCl2, 0.02% NP-40, and 10% glycerol. HO buffer contains 50 mM HEPES-KOH (pH 7.5), 20 mM MgCl2, 300 mM NaCl, 10% glycerol, and 0.5 mM DTT. Buffer H consists of 50 mM Tris-HCl (pH 7.6), 100 mM KCl, 25 μM zinc acetate, 0.1% NP-40, 10 mM MgCl2, and 1 mM DTT. Lon buffer contains 50 mM Tris-HCl (pH 8.0), 5 mM KCl, 15 mM MgCl2, and 2 mM DTT.

Proteins

ClpA (Maurizi et al. 1994), ClpP (Kim et al. 2000), ClpX (Levchenko et al. 1997), FtsH (Herman et al. 2003), HslU, HslV (Burton et al. 2005), Lon (Goldberg et al. 1994), RseA1–108, and σE (Flynn et al. 2004) were purified as described.

For the purification of his6-σE, his6-RseA1–100, and his6-RseA1–100E28C, plasmids pPER76, pJT6, and pRseA1–100E28C were transformed in BL21(DE3). Cultures were grown at 37°C in LB containing appropriate antibiotics to an O.D.600 ∼ 0.5 and induced with IPTG for 1–2 h; harvested cells were subjected to three freeze/thaw cycles in liquid nitrogen and ice water. Cells were resuspended in Buffer I/500 mM NaCl/10 mM imidazole (pH 7.9) containing 200 μg/mL lysozyme and lysed by sonication. After centrifugation, the supernatant was applied to Co2+ metal ion affinity column (Talon resin, Clontech) equilibrated in Buffer I/500 mM NaCl/10 mM imidazole. The column was washed with Buffer I/500 mM NaCl/20 mM imidazole and step-eluted with Buffer I/150 mM NaCl/40, 60, 100, 200 mM imidazole. The eluted protein was dialyzed against Buffer I/50 mM NaCl/25% glycerol/0.1% Tween 20. The Co2+ elution pool was further loaded on HiTrap Q (Pharmacia) anion exchange column equilibrated in Buffer I/50 mM NaCl. The Q column was washed with equilibration buffer and step-eluted with Buffer I/100, 250 mM NaCl. Fractions containing protein were pooled, dialyzed against the dialysis buffer, flash-frozen in liquid nitrogen, and stored at −80°C.

Fluorescent labeling

Prior to fluorescent labeling with rhodamine on cysteine 28, his6-RseA1–100E28C purified on HiTrap Q anion exchange column was loaded on PD10 desalting column. The protein was eluted using dialysis buffer without BME. The protein was incubated with 10-fold molar excess of rhodamine maleimide (Molecular Probes) at 4°C for 2 h The reaction was quenched with 1000-fold excess of DTT and the sample was centrifuged at 4°C to remove precipitated dye. In order to remove unreacted dye, the supernatant was chromatographed on PD10 desalting column prewashed with dialysis buffer containing 2 mM DTT. Fractions were analyzed by Coomassie staining of SDS-PAGE gels and by fluorescent imaging using Alpha Imager 2000.

Polarization anisotropy

His6-RseA1–100E28C freshly labeled with rhodamine, unlabeled his6-σE, and his6-RseA1–100 were diluted to appropriate concentration using dialysis buffer without glycerol. The samples were degassed under vacuum at room temperature before data collection. Data was collected with a K2 Multifrequency Fluorimeter (ISS) set to λex = 543 nm and λem = 575 nm.

The binding curve of labeled his6-RseA1–100E28C and his6-σE was obtained by measuring polarization anisotropy for 100 nM labeled his6-RseA1–100E28C in the presence of various concentrations of his6-σE. Relative affinities of labeled his6-RseA1–100E28C and unlabeled his6-RseA1–100 were measured by adding different concentrations of his6-σE to the mixture of labeled and unlabeled proteins (100 nM each protein). Both the proteins bound his6-σE with the same affinity. The dissociation rate of labeled his6-RseA1–100E28C/his6-σE complex (100 nM) was measured by monitoring anisotropy of the complex upon addition of excess unlabeled his6-RseA1–100 (1.5 μM). The dissociation rate of unlabeled his6-RseA1–100/his6-σE complex was measured in a similar manner in the presence of excess labeled his6-RseA1–100E28C.

Kinetic assay

The time course of his6-σE binding to labeled his6-RseA1–100E28C was obtained by following the increase in rhodamine fluorescence intensity (λex = 543 nm, λem = 575 nm) upon formation of the complex, using a stopped flow hand-operated mixing system (SFA-20; HiTech Scientific) and fluorimeter (K2; ISS). To calculate the association rate constant of his6-σE and labeled his6-RseA1–100E28C, the time course of their binding was measured for various concentrations of the two proteins (125 nM and 128 nM; 247 nM and 128 nM; 250 nM and 100 nM; 128 nM and 370 nM of his6-σE and labeled his6-RseA1–100E28C). The data was analyzed using Kaleidagraph (Synergy Software) by fitting to the function shown in Equation 1.

graphic file with name 124equ1.jpg

where I0 is the fluorescence intensity of labeled his6-RseA1–100 E28C not in complex with his6-σE, ΔI is the increase in fluorescence intensity upon his6-σE binding, C is the fraction of labeled his6-RseA1–100 E28C in complex with his6-σE, σE and RseAcyto are the concentrations of his6-σE and labeled his6-RseA1–100 E28C mixed together, and k is the association rate constant.

Determination of RseA, RseA-I, and RseA1–108 stability by pulse-chase immunoprecipitation

Cells were grown in supplemented M9 minimal medium lacking methionine (with 100 μg/mL Ap) at 30°C to an O.D.450 of 0.15–0.25. Cultures were induced with IPTG for 10 min to overexpress YYF peptide or RseA1–108/σE. Cells were pulse-labeled for 15 sec with L-[35S] methionine followed by a chase of 0.1% unlabeled methionine. Samples were processed as described (Ades et al. 2003). As an internal standard, an equal amount of extract from L-[35S] methionine-labeled cells overexpressing the truncated form of RseA, HA-RseA140 (also designated as RseA* in the text) (Kanehara et al. 2002) was added to each sample prior to immunoprecipitation. Samples were immunoprecipitated with antibodies against the cytoplasmic domain of RseA. T1/2 of RseA and RseA1–108 was determined as described (Ades et al. 2003). To estimate the T1/2 of RseA-I, both accumulation and disappearance of this intermediate on the gel was quantified. The net generation of RseA-I resulting from RseA cleavage by DegS and its disappearance upon cleavage by RseP can be described by Equation 2, where kDegS and kRseP are the rate constants of RseA cleavage by DegS and RseA-I cleavage by RseP, respectively.

graphic file with name 124equ2.jpg

 Maximal accumulation of RseA-I was measured to be between 25 and 50 sec after the chase. Using the value for kDegS determined as described (Ades et al. 2003) and Equation 3, we have estimated kRseP and thus the T1/2 of RseA-I.

graphic file with name 124equ3.jpg

 T1/2 of RseA-I can be also estimated from the ratio of RseA-I to RseA that does not change with time at longer time points. For this, we quantified the ratio of RseA-I to RseA at each time point and calculated T1/2 of RseA-I using Equation 4.

graphic file with name 124equ4.jpg

 The T1/2 of RseA-I measured using Equation 4 was in good agreement with that measured using Equations 2 and 3.

Degradation assays

In order to monitor the degradation of RseA1–108, ATP-dependent proteases (ClpX6: 0.3 μM and ClpP14: 0.8 μM for ClpXP; ClpA6: 0.3 μM and ClpP14: 0.8 μM for ClpAP; HslU6: 0.3 μM and HslV12: 0.8 μM for HslUV; Lon6: 0.3 μM; FtsH6: 0.3 μM), ATP (4 mM) and an ATP regeneration system (50 μg/mL creatine kinase and 2.5 mM creatine phosphate) were mixed in the appropriate buffer and incubated for 2 min at either 30°C or 37°C. RseA1–108 (2 μM) or RseA1–108/σE (the complex was formed by incubating 2 μM of RseA1–108 with 4 μM σE at 30°C for 10 min) was added to initiate the reaction. Samples were removed at various intervals, electrophoresed on SDS-PAGE, and visualized using Sypro Ruby protein gel stain (Molecular Probes) on a FluorImager 595 (Molecular Dynamics).

Degradation assays with ClpAP and ClpXP were performed at 30°C, whereas for FtsH, HslUV, and Lon the reactions were carried out at 37°C. Buffer condition was HO buffer for ClpAP, PD buffer for ClpXP and HslUV, Buffer H for FtsH, and Lon buffer for Lon.

Western blotting

The level of RseA1–108 and σE was monitored by Western blot analysis. Cultures were grown in LB (with 100 μg/mL Ap) at 30°C to an O.D.600 ∼ 0.3 and induced with IPTG for 1 h. Samples were processed as described (Alba et al. 2001) and transferred to PVDF membrane. The following dilutions of primary antisera (all rabbit) were used: anti-RseA cytoplasmic domain (1:5000) and anti-σE (1:10,000). The secondary antibody (used at 1:10,000) was an anti-rabbit HRP (Amersham Life Sciences). Blots were developed with SuperSignal West Dura Extended Duration Substrate (Pierce). Epi Chemi II Darkroom (UVP Laboratory Products) was used to capture the light emitted from the blots. The intensity of the bands was quantified using associated software (Labworks).

β-Galactosidase assays

Overnight cultures were diluted to an O.D.600 ∼ 0.015 in LB with 100 μg/mL Ap and grown at 30°C to an O.D.600 of 0.1–0.15. Cultures were induced with IPTG to overexpress YYF peptide. σE activity was measured by monitoring β-galactosidase expression from a single-copy σE-dependent lacZ reporter gene as described (Miller 1972; Mecsas et al. 1993; Ades et al. 1999). All assays were performed three times to ensure reproducibility of the data.

Acknowledgments

We thank Mark Dayle and Dyche Mullins for help with the fluorescent anisotropy experiments and Jennifer Hou for help with SspB and ClpS/ClpAP experiments. We also thank members of the Gross laboratory for critically reading the manuscript. This work was supported by National Institutes of Health (NIH) Grants GM-32678 (to C.A.G) and GM-049224 (to T.A.B). I.L.G was supported by Burroughs Wellcome Predoctoral Fellowship. T.A.B and J.M.F are employees of HHMI.

Footnotes

References

  1. Ades S.E., Connolly L.E., Alba B.M., Gross C.A., Connolly L.E., Alba B.M., Gross C.A., Alba B.M., Gross C.A., Gross C.A. The Escherichia coli σE-dependent extracytoplasmic stress response is controlled by the regulated proteolysis of an anti-σ factor. Genes & Dev. 1999;13:2449–2461. doi: 10.1101/gad.13.18.2449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Ades S.E., Grigorova I.L., Gross C.A., Grigorova I.L., Gross C.A., Gross C.A. Regulation of the alternative σ factor σE during initiation, adaptation, and shutoff of the extracytoplasmic heat shock response in Escherichia coli. J. Bacteriol. 2003;185:2512–2519. doi: 10.1128/JB.185.8.2512-2519.2003. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Akiyama Y., Kanehara K., Ito K., Kanehara K., Ito K., Ito K. RseP (YaeL), an Escherichia coli RIP protease, cleaves transmembrane sequences. EMBO J. 2004;23:4434–4442. doi: 10.1038/sj.emboj.7600449. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Alba B.M., Zhong H.J., Pelayo J.C., Gross C.A., Zhong H.J., Pelayo J.C., Gross C.A., Pelayo J.C., Gross C.A., Gross C.A. degS (hhoB) is an essential Escherichia coli gene whose indispensable function is to provide σE activity. Mol. Microbiol. 2001;40:1323–1333. doi: 10.1046/j.1365-2958.2001.02475.x. [DOI] [PubMed] [Google Scholar]
  5. Alba B.M., Leeds J.A., Onufryk C., Lu C.Z., Gross C.A., Leeds J.A., Onufryk C., Lu C.Z., Gross C.A., Onufryk C., Lu C.Z., Gross C.A., Lu C.Z., Gross C.A., Gross C.A. DegS and YaeL participate sequentially in the cleavage of RseA to activate the σE-dependent extracytoplasmic stress response. Genes & Dev. 2002;16:2156–2168. doi: 10.1101/gad.1008902. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bashyam M.D., Hasnain S.E., Hasnain S.E. The extracytoplasmic function σ factors: Role in bacterial pathogenesis. Infect. Genet. Evol. 2004;4:301–308. doi: 10.1016/j.meegid.2004.04.003. [DOI] [PubMed] [Google Scholar]
  7. Bohn C., Collier J., Bouloc P., Collier J., Bouloc P., Bouloc P. Dispensable PDZ domain of Escherichia coli YaeL essential protease. Mol. Microbiol. 2004;52:427–435. doi: 10.1111/j.1365-2958.2004.03985.x. [DOI] [PubMed] [Google Scholar]
  8. Brown M.S., Ye J., Rawson R.B., Goldstein J.L., Ye J., Rawson R.B., Goldstein J.L., Rawson R.B., Goldstein J.L., Goldstein J.L. Regulated intramembrane proteolysis: A control mechanism conserved from bacteria to humans. Cell. 2000;100:391–398. doi: 10.1016/s0092-8674(00)80675-3. [DOI] [PubMed] [Google Scholar]
  9. Browning D.F., Whitworth D.E., Hodgson D.A., Whitworth D.E., Hodgson D.A., Hodgson D.A. Light-induced carotenogenesis in Myxococcus xanthus: Functional characterization of the ECF σ factor CarQ and anti σ factor CarR. Mol. Microbiol. 2003;48:237–251. doi: 10.1046/j.1365-2958.2003.03431.x. [DOI] [PubMed] [Google Scholar]
  10. Burton R.E., Baker T.A., Sauer R.T., Baker T.A., Sauer R.T., Sauer R.T. Nucleotide-dependent substrate recognition by the AAA+ HslUV protease. Nat. Struct. Mol. Biol. 2005;12:245–251. doi: 10.1038/nsmb898. [DOI] [PubMed] [Google Scholar]
  11. Campbell E.A., Tupy J.L., Gruber T.M., Wang S., Sharp M.M., Gross C.A., Darst S.A., Tupy J.L., Gruber T.M., Wang S., Sharp M.M., Gross C.A., Darst S.A., Gruber T.M., Wang S., Sharp M.M., Gross C.A., Darst S.A., Wang S., Sharp M.M., Gross C.A., Darst S.A., Sharp M.M., Gross C.A., Darst S.A., Gross C.A., Darst S.A., Darst S.A. Crystal structure of Escherichia coli σE with the cytoplasmic domain of its anti-σ RseA. Mol. Cell. 2003;11:1067–1078. doi: 10.1016/s1097-2765(03)00148-5. [DOI] [PubMed] [Google Scholar]
  12. Casadaban M.J., Cohen S.N., Cohen S.N. Analysis of gene control signals by DNA fusion and cloning in Escherichia coli. J. Mol. Biol. 1980;138:179–207. doi: 10.1016/0022-2836(80)90283-1. [DOI] [PubMed] [Google Scholar]
  13. Chen J.C., Viollier P.H., Shapiro L., Viollier P.H., Shapiro L., Shapiro L. A membrane metalloprotease participates in the sequential degradation of a Caulobacter polarity determinant. Mol. Microbiol. 2005;55:1085–1103. doi: 10.1111/j.1365-2958.2004.04443.x. [DOI] [PubMed] [Google Scholar]
  14. Cowan S.W., Garavito R.M., Jansonius J.N., Jenkins J.A., Karlsson R., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Garavito R.M., Jansonius J.N., Jenkins J.A., Karlsson R., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Jansonius J.N., Jenkins J.A., Karlsson R., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Jenkins J.A., Karlsson R., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Karlsson R., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Konig N., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Pai E.F., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Pauptit R.A., Rizkallah P.J., Rosenbusch J.P., Rizkallah P.J., Rosenbusch J.P., Rosenbusch J.P., et al. The structure of OmpF porin in a tetragonal crystal form. Structure. 1995;3:1041–1050. doi: 10.1016/s0969-2126(01)00240-4. [DOI] [PubMed] [Google Scholar]
  15. Dartigalongue C., Missiakas D., Raina S., Missiakas D., Raina S., Raina S. Characterization of the Escherichia coli σE regulon. J. Biol. Chem. 2001;276:20866–20875. doi: 10.1074/jbc.M100464200. [DOI] [PubMed] [Google Scholar]
  16. De Las Penas A., Connolly L., Gross C.A., Connolly L., Gross C.A., Gross C.A. σE is an essential σ factor in Escherichia coli. J. Bacteriol. 1997a;179:6862–6864. doi: 10.1128/jb.179.21.6862-6864.1997. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. De Las Penas A., Connolly L., Gross C.A., Connolly L., Gross C.A., Gross C.A. The σE-mediated response to extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two negative regulators of σE. Mol. Microbiol. 1997b;24:373–385. doi: 10.1046/j.1365-2958.1997.3611718.x. [DOI] [PubMed] [Google Scholar]
  18. Dong T.C., Cutting S.M., Cutting S.M. SpoIVB-mediated cleavage of SpoIVFA could provide the intercellular signal to activate processing of Pro-σK in Bacillus subtilis. Mol. Microbiol. 2003;49:1425–1434. doi: 10.1046/j.1365-2958.2003.03651.x. [DOI] [PubMed] [Google Scholar]
  19. Dougan D.A., Reid B.G., Horwich A.L., Bukau B., Reid B.G., Horwich A.L., Bukau B., Horwich A.L., Bukau B., Bukau B. ClpS, a substrate modulator of the ClpAP machine. Mol. Cell. 2002;9:673–683. doi: 10.1016/s1097-2765(02)00485-9. [DOI] [PubMed] [Google Scholar]
  20. Farrell C.M., Grossman A.D., Sauer R.T., Grossman A.D., Sauer R.T., Sauer R.T. Cytoplasmic degradation of ssrA-tagged proteins. Mol. Microbiol. 2005;57:1750–1761. doi: 10.1111/j.1365-2958.2005.04798.x. [DOI] [PubMed] [Google Scholar]
  21. Flynn J.M., Levchenko I., Seidel M., Wickner S.H., Sauer R.T., Baker T.A., Levchenko I., Seidel M., Wickner S.H., Sauer R.T., Baker T.A., Seidel M., Wickner S.H., Sauer R.T., Baker T.A., Wickner S.H., Sauer R.T., Baker T.A., Sauer R.T., Baker T.A., Baker T.A. Overlapping recognition determinants within the ssrA degradation tag allow modulation of proteolysis. Proc. Natl. Acad. Sci. 2001;98:10584–10589. doi: 10.1073/pnas.191375298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Flynn J.M., Levchenko I., Sauer R.T., Baker T.A., Levchenko I., Sauer R.T., Baker T.A., Sauer R.T., Baker T.A., Baker T.A. Modulating substrate choice: The SspB adaptor delivers a regulator of the extracytoplasmic-stress response to the AAA+ protease ClpXP for degradation. Genes & Dev. 2004;18:2292–2301. doi: 10.1101/gad.1240104. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Goldberg A.L., Moerschell R.P., Chung C.H., Maurizi M.R., Moerschell R.P., Chung C.H., Maurizi M.R., Chung C.H., Maurizi M.R., Maurizi M.R. ATP-dependent protease La (lon) from Escherichia coli. Methods Enzymol. 1994;244:350–375. doi: 10.1016/0076-6879(94)44027-1. [DOI] [PubMed] [Google Scholar]
  24. Gottesman S., Roche E., Zhou Y., Sauer R.T., Roche E., Zhou Y., Sauer R.T., Zhou Y., Sauer R.T., Sauer R.T. The ClpXP and ClpAP proteases degrade proteins with carboxy-terminal peptide tails added by the SsrA-tagging system. Genes & Dev. 1998;12:1338–1347. doi: 10.1101/gad.12.9.1338. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Grigorova I.L., Chaba R., Zhong H.J., Alba B.M., Rhodius V., Herman C., Gross C.A., Chaba R., Zhong H.J., Alba B.M., Rhodius V., Herman C., Gross C.A., Zhong H.J., Alba B.M., Rhodius V., Herman C., Gross C.A., Alba B.M., Rhodius V., Herman C., Gross C.A., Rhodius V., Herman C., Gross C.A., Herman C., Gross C.A., Gross C.A. Fine-tuning of the Escherichia coli σE envelope stress response relies on multiple mechanisms to inhibit signal-independent proteolysis of the transmembrane anti-σ factor, RseA. Genes & Dev. 2004;18:2686–2697. doi: 10.1101/gad.1238604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Guyer M.S., Reed R.R., Steitz J.A., Low K.B., Reed R.R., Steitz J.A., Low K.B., Steitz J.A., Low K.B., Low K.B. Identification of a sex-factor-affinity site in E. coli as γδ. Cold Spring Harb. Symp. Quant. Biol. 1981;45:135–140. doi: 10.1101/sqb.1981.045.01.022. [DOI] [PubMed] [Google Scholar]
  27. Herman C., Prakash S., Lu C.Z., Matouschek A., Gross C.A., Prakash S., Lu C.Z., Matouschek A., Gross C.A., Lu C.Z., Matouschek A., Gross C.A., Matouschek A., Gross C.A., Gross C.A. Lack of a robust unfoldase activity confers a unique level of substrate specificity to the universal AAA protease FtsH. Mol. Cell. 2003;11:659–669. doi: 10.1016/s1097-2765(03)00068-6. [DOI] [PubMed] [Google Scholar]
  28. Hiratsu K., Amemura M., Nashimoto H., Shinagawa H., Makino K., Amemura M., Nashimoto H., Shinagawa H., Makino K., Nashimoto H., Shinagawa H., Makino K., Shinagawa H., Makino K., Makino K. The rpoE gene of Escherichia coli, which encodes σE, is essential for bacterial growth at high temperature. J. Bacteriol. 1995;177:2918–2922. doi: 10.1128/jb.177.10.2918-2922.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Jensen K.F. The Escherichia coli K-12 ‘wild types’ W3110 and MG1655 have an rph frameshift mutation that leads to pyrimidine starvation due to low pyrE expression levels. J. Bacteriol. 1993;175:3401–3407. doi: 10.1128/jb.175.11.3401-3407.1993. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kanehara K., Ito K., Akiyama Y., Ito K., Akiyama Y., Akiyama Y. YaeL (EcfE) activates the σE pathway of stress response through a site-2 cleavage of anti-σE, RseA. Genes & Dev. 2002;16:2147–2155. doi: 10.1101/gad.1002302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kanehara K., Ito K., Akiyama Y., Ito K., Akiyama Y., Akiyama Y. YaeL proteolysis of RseA is controlled by the PDZ domain of YaeL and a Gln-rich region of RseA. EMBO J. 2003;22:6389–6398. doi: 10.1093/emboj/cdg602. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kim Y.I., Burton R.E., Burton B.M., Sauer R.T., Baker T.A., Burton R.E., Burton B.M., Sauer R.T., Baker T.A., Burton B.M., Sauer R.T., Baker T.A., Sauer R.T., Baker T.A., Baker T.A. Dynamics of substrate denaturation and translocation by the ClpXP degradation machine. Mol. Cell. 2000;5:639–648. doi: 10.1016/s1097-2765(00)80243-9. [DOI] [PubMed] [Google Scholar]
  33. Levchenko I., Yamauchi M., Baker T.A., Yamauchi M., Baker T.A., Baker T.A. ClpX and MuB interact with overlapping regions of Mu transposase: Implications for control of the transposition pathway. Genes & Dev. 1997;11:1561–1572. doi: 10.1101/gad.11.12.1561. [DOI] [PubMed] [Google Scholar]
  34. Levchenko I., Grant R.A., Flynn J.M., Sauer R.T., Baker T.A., Grant R.A., Flynn J.M., Sauer R.T., Baker T.A., Flynn J.M., Sauer R.T., Baker T.A., Sauer R.T., Baker T.A., Baker T.A. Versatile modes of peptide recognition by the AAA+ adaptor protein SspB. Nat. Struct. Mol. Biol. 2005;12:520–525. doi: 10.1038/nsmb934. [DOI] [PubMed] [Google Scholar]
  35. Matson J.S., DiRita V.J., DiRita V.J. Degradation of the membrane-localized virulence activator TcpP by the YaeL protease in Vibrio cholerae. Proc. Natl. Acad. Sci. 2005;102:16403–16408. doi: 10.1073/pnas.0505818102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Maurizi M.R., Thompson M.W., Singh S.K., Kim S.H., Thompson M.W., Singh S.K., Kim S.H., Singh S.K., Kim S.H., Kim S.H. Endopeptidase Clp: ATP-dependent Clp protease from Escherichia coli. Methods Enzymol. 1994;244:314–331. doi: 10.1016/0076-6879(94)44025-5. [DOI] [PubMed] [Google Scholar]
  37. Mecsas J., Rouviere P.E., Erickson J.W., Donohue T.J., Gross C.A., Rouviere P.E., Erickson J.W., Donohue T.J., Gross C.A., Erickson J.W., Donohue T.J., Gross C.A., Donohue T.J., Gross C.A., Gross C.A. The activity of σE, an Escherichia coli heat-inducible σ-factor, is modulated by expression of outer membrane proteins. Genes & Dev. 1993;7:2618–2628. doi: 10.1101/gad.7.12b.2618. [DOI] [PubMed] [Google Scholar]
  38. Miller J.H. Experiments in molecular genetics. Cold Spring Harbor Laboratory; Cold Spring Harbor, NY: 1972. [Google Scholar]
  39. Missiakas D., Raina S., Raina S. The extracytoplasmic function σ factors: Role and regulation. Mol. Microbiol. 1998;28:1059–1066. doi: 10.1046/j.1365-2958.1998.00865.x. [DOI] [PubMed] [Google Scholar]
  40. Missiakas D., Betton J.M., Raina S., Betton J.M., Raina S., Raina S. New components of protein folding in extracytoplasmic compartments of Escherichia coli SurA, FkpA and Skp/OmpH. Mol. Microbiol. 1996;21:871–884. doi: 10.1046/j.1365-2958.1996.561412.x. [DOI] [PubMed] [Google Scholar]
  41. Missiakas D., Mayer M.P., Lemaire M., Georgopoulos C., Raina S., Mayer M.P., Lemaire M., Georgopoulos C., Raina S., Lemaire M., Georgopoulos C., Raina S., Georgopoulos C., Raina S., Raina S. Modulation of the Escherichia coli σE (RpoE) heat-shock transcription-factor activity by the RseA, RseB and RseC proteins. Mol. Microbiol. 1997;24:355–371. doi: 10.1046/j.1365-2958.1997.3601713.x. [DOI] [PubMed] [Google Scholar]
  42. Nikaido H. Outer membrane. In: Neidhardt F.C., et al., editors. Escherichia coli andSalmonella: Cellular and molecular biology. ASM Press; Washington, DC: 1996. pp. 29–47. [Google Scholar]
  43. Raina S., Missiakas D., Georgopoulos C., Missiakas D., Georgopoulos C., Georgopoulos C. The rpoE gene encoding the σE (σ24) heat shock σ factor of Escherichia coli. EMBO J. 1995;14:1043–1055. doi: 10.1002/j.1460-2075.1995.tb07085.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Raivio T.L., Silhavy T.J., Silhavy T.J. Periplasmic stress and ECF σ factors. Annu. Rev. Microbiol. 2001;55:591–624. doi: 10.1146/annurev.micro.55.1.591. [DOI] [PubMed] [Google Scholar]
  45. Reiling S.A., Jansen J.A., Henley B.J., Singh S., Chattin C., Chandler M., Rowen D.W., Jansen J.A., Henley B.J., Singh S., Chattin C., Chandler M., Rowen D.W., Henley B.J., Singh S., Chattin C., Chandler M., Rowen D.W., Singh S., Chattin C., Chandler M., Rowen D.W., Chattin C., Chandler M., Rowen D.W., Chandler M., Rowen D.W., Rowen D.W. Prc protease promotes mucoidy in mucA mutants of Pseudomonas aeruginosa. Microbiology. 2005;151:2251–2261. doi: 10.1099/mic.0.27772-0. [DOI] [PubMed] [Google Scholar]
  46. Resnekov O., Losick R., Losick R. Negative regulation of the proteolytic activation of a developmental transcription factor in Bacillus subtilis. Proc. Natl. Acad. Sci. 1998;95:3162–3167. doi: 10.1073/pnas.95.6.3162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Rezuchova B., Miticka H., Homerova D., Roberts M., Kormanec J., Miticka H., Homerova D., Roberts M., Kormanec J., Homerova D., Roberts M., Kormanec J., Roberts M., Kormanec J., Kormanec J. New members of the Escherichia coli σE regulon identified by a two-plasmid system. FEMS Microbiol. Lett. 2003;225:1–7. doi: 10.1016/S0378-1097(03)00480-4. [DOI] [PubMed] [Google Scholar]
  48. Rhodius V.A., Suh W.C., Nonaka G., West J., Gross C.A., Suh W.C., Nonaka G., West J., Gross C.A., Nonaka G., West J., Gross C.A., West J., Gross C.A., Gross C.A. Conserved and variable functions of the σE stress response in related genomes. PLoS Biol. 2006;4:43–59. doi: 10.1371/journal.pbio.0040002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rouviere P.E., Gross C.A., Gross C.A. SurA, a periplasmic protein with peptidyl-prolyl isomerase activity, participates in the assembly of outer membrane porins. Genes & Dev. 1996;10:3170–3182. doi: 10.1101/gad.10.24.3170. [DOI] [PubMed] [Google Scholar]
  50. Rouviere P.E., De Las Penas A., Mecsas J., Lu C.Z., Rudd K.E., Gross C.A., De Las Penas A., Mecsas J., Lu C.Z., Rudd K.E., Gross C.A., Mecsas J., Lu C.Z., Rudd K.E., Gross C.A., Lu C.Z., Rudd K.E., Gross C.A., Rudd K.E., Gross C.A., Gross C.A. rpoE, the gene encoding the second heat-shock σ factor, σE, in Escherichia coli. EMBO J. 1995;14:1032–1042. doi: 10.1002/j.1460-2075.1995.tb07084.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Sambrook J., Fritsch E.F., Maniatis T., Fritsch E.F., Maniatis T., Maniatis T. Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press; Cold Spring Harbor, NY: 1989. [Google Scholar]
  52. Schobel S., Zellmeier S., Schumann W., Wiegert T., Zellmeier S., Schumann W., Wiegert T., Schumann W., Wiegert T., Wiegert T. The Bacillus subtilis σW anti-σ factor RsiW is degraded by intramembrane proteolysis through YluC. Mol. Microbiol. 2004;52:1091–1105. doi: 10.1111/j.1365-2958.2004.04031.x. [DOI] [PubMed] [Google Scholar]
  53. Schweder T., Lee K.H., Lomovskaya O., Matin A., Lee K.H., Lomovskaya O., Matin A., Lomovskaya O., Matin A., Matin A. Regulation of Escherichia coli starvation σ factor (σS) by ClpXP protease. J. Bacteriol. 1996;178:470–476. doi: 10.1128/jb.178.2.470-476.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Studier F.W., Moffatt B.A., Moffatt B.A. Use of bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. J. Mol. Biol. 1986;189:113–130. doi: 10.1016/0022-2836(86)90385-2. [DOI] [PubMed] [Google Scholar]
  55. Walsh N.P., Alba B.M., Bose B., Gross C.A., Sauer R.T., Alba B.M., Bose B., Gross C.A., Sauer R.T., Bose B., Gross C.A., Sauer R.T., Gross C.A., Sauer R.T., Sauer R.T. OMP Peptide signals initiate the envelope-stress response by activating DegS protease via relief of inhibition mediated by its PDZ domain. Cell. 2003;113:61–71. doi: 10.1016/s0092-8674(03)00203-4. [DOI] [PubMed] [Google Scholar]
  56. Weihofen A., Martoglio B., Martoglio B. Intramembrane-cleaving proteases: Controlled liberation of proteins and bioactive peptides. Trends Cell Biol. 2003;13:71–78. doi: 10.1016/s0962-8924(02)00041-7. [DOI] [PubMed] [Google Scholar]
  57. Wilken C., Kitzing K., Kurzbauer R., Ehrmann M., Clausen T., Kitzing K., Kurzbauer R., Ehrmann M., Clausen T., Kurzbauer R., Ehrmann M., Clausen T., Ehrmann M., Clausen T., Clausen T. Crystal structure of the DegS stress sensor: How a PDZ domain recognizes misfolded protein and activates a protease. Cell. 2004;117:483–494. doi: 10.1016/s0092-8674(04)00454-4. [DOI] [PubMed] [Google Scholar]
  58. Zhou R., Kroos L., Kroos L. BofA protein inhibits intramembrane proteolysis of pro-σK in an intercompartmental signaling pathway during Bacillus subtilis sporulation. Proc. Natl. Acad. Sci. 2004;101:6385–6390. doi: 10.1073/pnas.0307709101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Zhou R., Kroos L., Kroos L. Serine proteases from two cell types target different components of a complex that governs regulated intramembrane proteolysis of pro-σK during Bacillus subtilis development. Mol. Microbiol. 2005;58:835–846. doi: 10.1111/j.1365-2958.2005.04870. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Genes & Development are provided here courtesy of Cold Spring Harbor Laboratory Press

RESOURCES