Structural basis for the negative regulation of bacterial stress response by RseB

Abstract

The σE-dependent stress response in bacterial cells is initiated by the DegS- and RseP-regulated intramembrane proteolysis of a membrane-spanning antisigma factor, RseA. RseB binds to RseA and inhibits its sequential cleavage, thereby functioning as a negative modulator of this response. In the crystal structure of the periplasmic domain of RseA bound to RseB, the DegS cleavage site of RseA is unstructured, however, its P1 residue is buried in the hydrophobic pocket of RseB, which suggests that RseB binding blocks the access of DegS to the cleavage site.

Introduction

Regulated intramembrane proteolysis (RIP) is a control mechanism underlying transmembrane signal transfer and performs a key role in the initiation of the essential signal transduction pathways in diverse organisms.¹ For example, the Notch signaling pathway, which is critical for a variety of cell–cell communications in multicellular organisms, is controlled by the RIP of the Notch receptor by ADAM-family metalloprotease and gamma-secretase.¹ In Gram-negative bacteria, the sequential cleavage of RseA, a membrane-spanning anti-σE factor, modulates the initiation of the envelope-stress response.^2,3 RseA forms a tight complex with σE using its N-terminal cytoplasmic domain, thereby inhibiting the transcription of σE-dependent genes. Under stress conditions that include the misfolding of periplasmic proteins, two membrane proteases, DegS and RseP, sequentially degrade RseA to liberate σE (Supporting Information Fig. S1). DegS, which is activated when its PDZ domain is bound to the C-terminal peptide of unfolded outer membrane porins (OMPs), cleaves the C-terminal periplasmic domain of RseA.⁴ Subsequently, RseP cleavage within the membrane domain of RseA releases the cytoplasmic domain of RseA (associated with the σE) from the membrane. In the final step, the cytoplasmic domain of RseA is degraded such that the released σE can interact with RNA polymerase.⁵ RseB also participates in the regulation of σE-dependent envelope-stress response by inhibiting the intramembrane proteolysis of RseA.⁶ RseB has been previously demonstrated to suppress the proteolytic activity of DegS for RseA,^6,7 independently of the activation mechanism of DegS.⁸ Accordingly, either the deletion of the rseB gene or the release of RseB from RseA results in a more rapid degradation of RseA and increased activity of σE.^6,9 In this study, we attempted to characterize the regulatory role of RseB in the proteolytic cleavage of RseA and determined the crystal structure of RseB in complex with RseA_peri (the periplasmic domain of RseA, residues 121–216) at a resolution of 2.3 Å by molecular replacement using apo-RseB (PDB ID: 2P4B) as a template (Table I).

Table I.

Data Collection and Refinement Statistics

	RseAperi•RseB
Data collection
Space group	P212121
Cell dimensions
a, b, c (Å)	87.05, 119.58, 150.67
α, β, γ (°)	90.00, 90.00, 90.00
Resolution (Å)	30.00–2.30 (2.38–2.30)a
Rsym or Rmerge	6.0 (20.8)
I/σI	20.0 (3.5)
Completeness (%)	93.0 (81.5)
Redundancy	4.6
Refinement
Resolution (Å)	20.00–2.30
No. reflections, working/free	62725/3320
Rwork/Rfree	23.9/27.1
No. atoms
Protein	10108
Zn2+	6
Water	421
B-factors
Protein	52.1
Zn	82.3
Water	50.0
R.m.s. deviations
Bond lengths (Å)	0.008
Bond angles (°)	1.482
Ramachandran plot
Most favored (%)	87.2
Additionally allowed (%)	12.7
Generously allowed (%)	0.2
Disallowed (%)	0.0

^aValues in parentheses are for highest-resolution shell.

Results and Discussion

Structure determination

The RseA_peri•RseB structure was determined at a resolution of 2.3 Å by molecular replacement using the large domain (residues 26–200) of E.coli RseB (PDB ID: 2P4B) as a template.¹⁰ Although the asymmetric unit of the crystal harbors four RseA_peri•RseB complexes (Supporting Information Fig. S2), the dimeric structure of the complex has been known from size exclusion and SAXS (Small Angle X-ray Scattering) data.¹¹ Each complex (Com1–Com 4 in Supporting Information Fig. S2) is composed of one RseB monomer and one RseA_peri monomer. Com1 is in contact with two other complexes, Com2 and Com3. The Com1:Com2 interaction, which is mediated by hydrogen bonds between relatively well-conserved residues, buries the 1190 Ų surface area of each complex. RseB:RseB, RseA_peri:RseA_peri, and RseA_peri:RseB interfaces contribute to this burial of surface area by 889 Ų, 57 Ų, and 244 Ų, respectively. The Com1:Com3 interaction results in the burial of 370 Ų, which is primarily a RseB:RseB contact and involves a zinc ion that was added for the purposes of crystallization (Supporting Information Fig. S3). Moreover, the dimeric interaction between the N-terminal regions of the two RseAs in the Com1:Com2 dimer and their proximity to the transmembrane region demonstrate the involvement of the transmembrane domain of RseA in dimeric contacts [Fig. 1(a) and Supporting Information Fig. S4]. There is no biologically relevant higher-order oligomeric form that can be generated by symmetry operations. Accordingly, the Com1:Com2 dimer (or Com3:Com4 dimer) was considered biologically relevant [Fig. 1(a) and Supporting Information Fig. S2]. In this manuscript, we used Com1 and Com1:Com2 to describe the monomeric and dimeric RseA_peri•RseB complexes, respectively. The RMSD between Com1:Com2 and Com3:Com4 complexes was 0.43 Å for 626 Cα atoms.

Figure 1.

Structure of the RseA_peri•RseB complex. (a) Dimer model of the RseA_peri•RseB complex. Each RseA_peri is depicted in magenta or yellow, and RseB is depicted in green or slate. Ribbon diagram (b) and surface model (c) of the monomeric RseA_peri•RseB complex. The regions important for their binding, RseB_25–209, RseB_210–216, RseB_217–315, RseA_132–151, and RseA_169–190, are colored slate, purple, red, green, and yellow, respectively. The DegS cleavage site is shown in blue.

Overall structure

We were able to model most residues in RseB, with the exception of the first N-terminal residue, residues 240–246, and three C-terminal residues. The loop connecting the large (RseB_25–209) and small (RseB_217–315) domains was disordered in the apo-RseB model^10,12; however, it was well ordered in the RseA_peri•RseB complex due to the interaction with the bound RseA_peri, which results in the stabilization of the loop (Fig. 1). By way of contrast, the 96-residue periplasmic domain of RseA (residues 121–216) was largely unstructured, and only two regions involved in RseB binding, RseA_132–151 and RseA_169–190, were modeled (Fig. 1 and Supporting Information Fig. S5). RseB evidenced similar structures in their apo- (Chain A of 2P4B or Chain A of 2V43) and RseA_peri-bound states, with an RMSD of 0.85 Å for 276 Cα atoms (2P4B) or 1.61 Å for 261 Cα atoms (2V43), thereby indicating that its overall conformation is largely maintained upon RseA binding. The major local conformational change was found in two β-strands (β5 and β6; residues 88–104) in the small domain of RseB (2P4B), which binds directly to the C-terminus of RseA_132–151 [Fig. 1 and Supporting Information Fig. S6(a)]. The conformational changes are more drastic when the RseA-bound RseB was compared with another crystal structure of apo-RseB (PDB ID: 2V43) where four β-strands (β3–β6; residues 68–104) exhibit large conformational changes after binding to the C-terminus of RseA [Fig. 1 and Supporting Information Fig. S6(b)].

Interaction between RseA_peri and RseB

RseA binds to a broad area of the RseB groove that is formed between the large and small domains of RseB (Figs. 1 and 2). RseA_132–151 mostly forms a random coil rather than a regular secondary structure and interacts with residues in the large domain of RseB. The residues in RseA_132–151 form hydrophobic interactions with the hydrophobic residues or aliphatic carbons of bulky residues of RseB, with the exception of Lys144, which forms a salt-bridge with Glu181 of RseB [Fig. 2(a)]. The results of the histidine pull-down assay verified that an RseA_peri mutant featuring Ala substitutions at Gly143, Lys144, and Pro147 was still capable of binding to RseB (data not shown), thereby indicating that the electrostatic interaction attributable to Lys144 is not critically important to the association between RseB and RseA_132–151. Consistent with this finding, Lys144 is not conserved among RseA homologues in Gram-negative bacteria (Fig. 3).

Figure 2.

RseA_peri•RseB interaction. (a) Schematic drawing of RseA_peri•RseB interaction. RseA residues located in random coils and helices are shown as orange circles and green pentagons, respectively. RseB residues in the same complex and from the other RseA_peri•RseB complex are marked as white and cyan boxes, respectively. Charge interactions, hydrogen bonds, and hydrophobic contacts are shown as red, blue, and black lines, respectively. (b) Binding interface between RseA and RseB. RseB, RseA_132–151, and RseA_169–190 are colored purple, green, and yellow, respectively. Dotted lines indicate charge interactions and the involved residues are depicted by stick models. (c) The DegS cleavage site (Val148-Ser149) bound to the RseB groove in the RseA_peri•RseB complex is drawn in a ribbon model with the same color scheme as in Fig. 1(b). The residues near the cleavage site are drawn as stick models and labeled. The cleavage site is indicated by a black arrow and labeled.

Figure 3.

Multiple sequence alignment of the periplasmic domain of RseAs from Gram-negative bacteria. Identical and similar residues are boxed in blue and yellow, respectively. Species abbreviations are as follows: Ec, Escherichia coli; Sf, Shigella flexneri; Se, Salmonella enterica; Yp, Yersinia pestis; Eca, Erwinia carotovora; Vc, Vibrio cholerae; So, Shewanella oneidensis; Hi, Haemophilus influenzae; Ms, Mannheimia succiniciproducens.

RseA_169–190 exhibits a helical conformation and binds principally to the small domain of RseB. The charged residues in RseA_169–190 are well-conserved in the RseA homologues and are important to RseB binding (Figs. 2 and 3). Most notably, Arg172, Asp179, Glu181, and Arg184/Arg185 in RseA_169–190 form salt bridges with Glu293, Arg239, Arg282, and Asp109 of RseB, respectively [Figs. 2(a,b)]. It has been demonstrated that RseA_169–185 is the minimum fragment necessary for RseB binding^8,10; additionally, the mutation of the conserved Arg residues in this fragment (Arg172, Arg184, and Arg185) abolishes RseB binding activity^10,13 (Fig. 3). Therefore, RseA_132–151 does not appear to be the primary determinant in RseB binding, but it may perform other additional functions, such as recruiting RseB or sterically inhibiting the access of proteases. Two RseA_peri•RseB complexes in a dimer (Com1:Com2) are also stabilized via intercomplex interactions [Fig. 1(a) and Supporting Information Fig. S2]. The Val135, Phe136, and Thr138 residues of RseA in Com1 are in contact with Ile50, Asn51/Thr179/Gln182, and Arg169/Arg184 of RseB in Com2, respectively, and vice versa [Figs. 1(a),2(b)].

Structural implication of the binding of RseA to RseB

The DegS cleavage site (Val148-Ser149; P1-P1′) at the C-terminal end of RseA_132–151, is located within the RseB groove in the RseA_peri•RseB complex [Figs. 1,2(c)]. Val148 is buried in the hydrophobic pocket formed by Phe100 and Leu102 of RseB and Leu182 of RseA_169–190. Ser149, which is located near the helix in RseA_169–190, forms a hydrogen bond with Gln178 of RseA. As a result, the DegS cleavage site is almost completely hidden by RseB and RseA_169–190, such that DegS access is restricted in the RseA_peri•RseB complex, and probably also in the RseA•RseB complex (Fig. 1). From this perspective, it has been theorized that the binding of RseA_132–151 to RseB contributes to locating the cleavage site deep inside of the RseB groove, thereby rendering it resistant to DegS cleavage. This mechanism is consistent with the model proposed in Ref.8.

In the proteolytic cascade of RseA, the cleavage by RseP requires prior periplasmic cleavage by DegS and the release of RseA_149–216.⁵ It was reported recently that the interaction of the newly exposed C-terminal residue of RseA_1–148, Val 148, with the second PDZ domain of RseP is critically important for the cleavage.¹⁴ It is expected that the interaction between RseB and RseA_1–148 is not very strong due to the lower binding affinity of RseA_121–173 for RseB detected in previous biochemical studies.¹⁰ These findings suggest that RseB in complex with RseA_149–216 will dissociate from RseA_1–148 after DegS cleavage and that RseB is unlikely to reassociate with RseA_1–148. Therefore, RseP will interact with the C-terminal end of the DegS-cleaved RseA (RseA_1–148).

Materials and Methods

Protein expression and purification

E.coli RseA_peri (periplasmic domain containing residues 121–216) and RseB (residues 24–318) were expressed separately in E. coli BL21(DE3) as previously described.^10,11 His-Trx-RseA_peri- and RseB-expressing cells were harvested and mixed at a ratio of 1:3 (wet weight) to ensure the formation of the complex. The cells were then sonicated in buffer A (20 mM Tris-HCl pH 7.5 and 0.1M NaCl). The RseA_peri•RseB complex was then purified by nickel-affinity chromatography and size exclusion chromatography. The cleared lysates were loaded onto a metal-chelating column (GE Healthcare, Princeton, NJ) and the proteins were eluted with 50–500 mM imidazole gradient. The fractions containing His-Trx-RseA_peri•RseB were pooled and dialyzed twice against buffer A. The His-Trx tag was removed with thrombin at room temperature, and RseA_peri•RseB was purified further using a Superdex-200 column (GE Healthcare, Princeton, NJ) pre-equilibrated with buffer A, then concentrated to 15 mg/mL.

Crystallization and data collection

The crystallization of the RseA_peri•RseB complex was performed using the microbatch method at 14°C. The crystallization drop was prepared by mixing 1 μL protein solution (8–10 mg/mL) and 1 μL crystallization reagent (28% PEG550MME, 10 mM ZnSO₄, and 100 mM MES pH 6.5) under a layer of Al's oil (Hampton Research, Aliso Viejo, CA). The crystals in a drop were flash-frozen in a cold nitrogen stream at 100 K without the addition of a cryoprotectant, and the diffraction data were collected at PLS-BL4A (Beam line 4A, Pohang Light Source, South Korea; wavelength 1.0000 Å). The diffraction images were recorded to an ADSC Quantum 210 CCD detector. The diffraction data were indexed and integrated using HKL2000 and scaled using SCALEPACK.¹⁵

Structure determination

The RseA_peri•RseB structure was determined by molecular replacement using PHASER¹⁶ with the large domain (residues 26–200) of E.coli RseB¹⁰ (PDB 2P4B) as a template. Three large domains were initially identified, and the small domains were added to the model. The fourth RseB was generated by a noncrystallographic symmetry operation. Several cycles of rigid body, positional, simulated annealing and B-factor refinements, and model rebuilding were conducted at a resolution of 2.3 Å using the CNS and COOT programs.^17,18 The RseA model was placed on the additional electron density. The RseA_peri•RseB structure was refined further using REFMAC.¹⁹ The final refinement with solvents resulted in R and R_free values of 23.8% and 27.0%, respectively. The data collection and refinement statistics are summarized in Table I. The figures were drawn using PyMOL.²⁰ The protein–protein interface was calculated using Protorp.²¹ The coordinates and structure factors for the RseA_peri•RseB complex have been deposited under accession code 3M4W.

Summary

In this study, we characterized the inhibitory mechanisms of RseB in the regulated proteolysis of RseA. The C-terminal helix in RseA_169–190 is a major contributor to the formation of a stable complex with RseB, whereas the N-terminal region of the periplasmic domain of RseA is necessary for the burial of the DegS cleavage site within an inaccessible pocket in the RseA•RseB complex. In the regulation of the envelope-stress response, RseB functions by blocking the access of DegS protease rather than converting RseA into a compact structure that is resistant to proteolysis, as the random coil structure around the cleavage site is maintained in the complex. Accordingly, in this study we explain why the release of RseB is a prerequisite for the degradation of RseA and the activation of the σE-dependent envelope stress response at the atomic level. In the crystal structure of the RseA•RseB complex, RseB is unlikely to interfere with the further degradation of the periplasmically cleaved RseA. Therefore, the newly exposed Val148 will be readily recognized by RseP for subsequent cleavage, which results in the activation of the envelope-stress response.